Mutations of rpoB, katG, inhA and ahp genes in rifampicin and isoniazid-resistant Mycobacterium tuberculosis in Kyrgyz Republic

The aim of this study was to identify mutations of rpoB, katG, inhA and ahp-genes associated Mycobacterium tuberculosis resistance to rifampicin (RIF) and isoniazid (INH) in Kyrgyz Republic. We studied 633 smear samples from the primary pulmonary tuberculosis (TB) patients. We verified Mycobacterium tuberculosis susceptibility to RIF and INH using culture method of absolute concentrations, and commercially available test named “TB-BIOCHIP” (Biochip-IMB, Moscow, Russian Federation). For RIF-resistance, TB-BIOCHIP’s sensitivity and specificity were 88% and 97%, 84% and 95% for INH-resistance, and 90% and 97% for multi-drug resistance (MDR). In RIF-resistant strains, TB-BIOCHIP showed mutations in codons 531 (64.8%), 526 (17.3%), 516 (8.1%), 511 (5.4%), 533 (3.2%), 522 (0.6%) and 513 (0.6%) of rpoB gene. The most prevalent was Ser531 > Leu mutation (63.7%). 91.2% of mutations entailing resistance to INH were in katG gene, 7% in inhA gene, and 1.8% in ahpC gene. Ser315→Thr (88.6%) was the most prevalent mutation leading to resistance to INH. In Kyrgyz Republic, the most prevalent mutation in RIF-resistant strains was Ser531 → Leu in rpoB gene, as opposed to Ser315 → Thr in katG gene in INH-resistant Mycobacterium tuberculosis. In Kyrgyz Republic, the major reservoir of MDR Mycobacterium tuberculosis were strains with combined mutations Ser531 → Leu in rpoB gene and Ser315 → Thr in katG gene. TB-BIOCHIP has shown moderate sensitivity with the advantage of obtaining results in only two days.


Background
Kyrgyzstan is a country in Central Asia with the total population of 6 million, whereas mountainous terrains with poor access to medical care occupy 90% of its territory. As reported by the WHO, Kyrgyzstan shows high incidence and mortality from TB, and the former peaked in 2001 with 167.8 cases per 100,000 population including penitentiary system patients. TB incidence in Kyrgyzstan that year was 16.8 times the worldwide threshold incidence of 10.0 per 100,000. The corresponding mortality rate that year was 27.0 per 100,000 including penitentiary system deaths. Since 2002, there was a consistent reduction in both incidence and mortality from TB in the country. By 2012, the incidence dropped by 37.8% to 104.3 per 100,000, whereas 68.1% reduction in mortality was registered, and the latter equaled 8.6 per 100,000, including penitentiary system patients. During the period of 2012-2016, both incidence and mortality continued to drop. In 2016, 93.1 new cases and 5.8 deaths per 100,000 were reported [1].
Parallel to the overall persistent reduction in TB incidence, the number of multidrug resistant (MDR) and extensive drug resistant (XDR) tuberculosis cases increased. Among all new primary smear-positive pulmonary TB cases in 2015, 26% were MDR, whereas among previously treated patients this number equaled 54.6% [1].
Treatment success in MDR TB remains low, reaching only 47-62.7% [2]. MDR TB is not only a serious clinical and epidemiological problem, but also entails substantial economic costs of management. Thus, treating patients with resistance to the main anti-TB, such as rifampicin (RIF) and isoniazid (INH), may be times more expensive compared to treatment costs incurred by the management of TB susceptible to the main medications panel.
Quite much is known about the mechanisms of resistance to RIF and INH. Thus, Mycobacterium tuberculosis resistance to RIF can be explained by mutations in rpoB gene, whereas katG, inhA, ahpC, kasA and ndh genes mutations are associated with INH resistance [3,4]. Presentations from around the world report varying geographic distribution of selected mutations of this kind [5], when the most prevalent strains exhibit mutations in codons 531, 526 and 516. In countries with poor epidemiological situation around TB, including former Soviet countries, strains with mutation in codon 531 (Ser531→Leu) are the most prevalent (50-70%) ones [6,7]. On the contrary, this codon mutation in other countries does not exceed 30-48% [8][9][10].
We have previously addressed and published the outcomes of our studies addressing genes rpoB, katG, inhA, and ahp mutations [11,12]. The current analysis presents data on these genes from a larger Mycobacterium tuberculosis DNA sample, as well as the most and the least prevalent RIF-and INH-resistant Mycobacterium tuberculosis strains in Kyrgyz Republic. This updated analysis also includes TB-BIOCHIP specificity and sensitivity data in detecting RIF-and INH-resistant/susceptible Mycobacterium tuberculosis in comparison with a standard culture method.
Therefore, local distribution of drug resistance found in circulating strains should be described based on the molecular studies. In Kyrgyzstan, mutations of rpoB, katG, inhA, and ahpC genes have not been studied in local Mycobacterium tuberculosis strains in detail, and major mutations leading to Mycobacteria resistance to RIF and INH have yet to be identified. Therefore, the aim of this study was to describe mutations in rpoB, katG, inhA and ahpС genes and to select the dominating population of RIF and INH-resistant Mycobacterium tuberculosis strains in TB patients living in Kyrgyzstan.

Patients and study design
The study commenced in 2009 and ended in 2011. Patients were recruited from the various inpatient departments of the National Centre of Tuberculosis (NCT), located in Bishkek. Six hundred and thirty-three samples of Mycobacterium tuberculosis were isolated from 633 patients from eight provinces of Kyrgyz Republic, including 192 from Bishkek‚ 150 from Chuy, 51 from Issyk-Kul, 66 from Naryn, 36 from Talas, 36 from Osh‚ 69 from Jalalabat, and 33 from Batken.
Tuberculosis diagnosis in recruited patients was confirmed by microscopy and culturing. We considered the following inclusion criteria: TB diagnosed for the first time in life, smear positive microscopic analysis of sputum, age 18 to 70, inpatient treatment and negative human immunodeficiency virus (HIV) test.
All patients underwent chest X-ray, smear microscopy, supplemented by the bacteriological verification of Mycobacterium tuberculosis at the NCT in Bishkek. This study protocol was approved by the local committee on bioethics of the NCT and the Institute of Molecular Biology and Medicine (Bishkek). All patients in this study provided written informed consent to participate.
In Kyrgyz Republic, on average, there were 1500 new cases of smear-positive TB within the period described in the manuscript. We, therefore, consider the sample size of this study (N = 633) sufficient for generalization of the molecular and genetic background of TB identified in Kyrgyzstan.

Data collection
Sputum samples were collected from patients using conventional methods on the first day of admission to the clinic. Following decontamination, each sputum specimen was divided into two parts. The first portion was used to test Mycobacterium resistance to RIF and INH using classical phenotypic method, whereas the second portion was used to test resistance to RIF and INH using genetic approach, i.e. TB-BIOCHIP.
Bacteriological study and phenotypic drug susceptibility testing (DSTs) We used World Health Organization (WHO) recommendations for agent culturing in Lowenstein-Jensen culture medium and first line drug resistance testing via absolute concentrations method [13]. In Kyrgyz Republic, along with the conventional culture method of absolute concentrations with its results available 2-3 months after, they also use automatic «Bactec MGIT 960» test with liquid medium, which shortens result anticipation period to 2-3 weeks.
Whenever we detected the growth of Mycobacterium tuberculosis, Kyrgyz national TB guidelines mandate DSTs using the absolute concentration method on Löwenstein-Jensen medium in the NCT. The corresponding INH and RIF concentrations were 0.2 μg/mL and 40 μg/mL.

Molecular and genetic methods of testing resistance to RIF and INH
Drug resistance to both RIF and INH was tested using commercially available "TB-BIOCHIP" set (JSC «BIO-CHIP-IMB», Moscow, Russian Federation), which was developed by the Institute of molecular biology named after V.A. Engelhard of the Russian Academy of Science (Moscow, Russian Federation). Testing Mycobacterium tuberculosis resistance to RIF and INH using TB-BIOCHIP included the stages of clinical specimen decontamination, germ lysis, two consecutive stage of multiplex PCR, PCR-product (amplicon) hybridization on a biologic microchip, registration, and interpretation. All these stages are described in the product manual [13].

Сlinical specimen decontamination and germ lysis
We used N-acetyl L-cysteine and 3% NaOH solution to decontaminate sputum specimens for 40 min at room temperature. We then resuspended cell extracts in TE buffer and centrifuged them for 10 min, and then exposed the pellet to 30 μl of lyses buffer, which contained 1% TritonX-100, and then incubated it for 20 min at 95°C. After that we cooled the extract and centrifuged it for 10 min. Two μl aliquots of supernatant were used for polymerase chain reaction (PCR).
In order to check for possible contamination‚ we also tested negative control sets of all reagents except target DNA along with the samples in every set of amplification reactions.

On-chip hybridization and microchip analysis
The corresponding procedure has been described elsewhere in detail [16].

Hybridization outcomes registration
We tested mutations in rpoB, katG, inhA, and ahpС genes using the fluorescence intensity pattern analysis on the chip with Chipdetector portable fluorescence analyzer. This device works with Imageware software (BIOCHIP-IMB, Moscow, Russian Federation). TB-BIO-CHIP can identify mutations in rроВ gene (27 types), katG (11 types), inhA (5 types) and ahpС (5 types), which result in Mycobacterium tuberculosis resistance to RIF and INH.
DNA extraction from clinical samples‚ PCR amplification of the rpoB, katG, inhA and ahpС genes‚ on-chip hybridization and detection of mutations were carried out by the Institute of Molecular Biology and Medicine‚ Bishkek‚ Kyrgyz Republic.

Statistical methods
We compared phenotypic and genetic methods of detecting Mycobacterium tuberculosis resistance and susceptibility to RIF and INH using 2*2 tables. We calculated TB-BIOCHIP sensitivity (ability to detect true resistance) and specificity (ability to detect true sensitivity) using the formulae: specificity = Tn/(Tn + Fp) × 100%, sensitivity = Tp/(Tp + Fn) × 100%, where Tp is a true positive outcome, Tn is a true negative outcome, Fp is a false positive outcome, and Fn is a false negative outcome [17].

Results
Patients' age ranged from 18 to 70 years, of whom 60% were patients in the age group below 40. In total, there were 386 males (61%) and 247 (39%) females, with the corresponding male to female ratio 1.5:1.

Identification of RIF-resistant and susceptible Mycobacterium tuberculosis using phenotypic DST and biochip
Cultural method yielded 209 RIF-resistant specimens out of the overall number N = 633, whereas 424 specimens were susceptible to RIF. Biochip analysis confirmed resistance to RIF in 185 (88%) specimens out of 209, whereas RIF-susceptibility was confirmed in 411 (97%) specimens out of 424. Therefore, TB-BIOCHIP resulted in 88% sensitivity and 97% specificity. The overall agreement of two methods in RIF-resistant/susceptible strains identification was 94% (596 out of 633) ( Table 1).
We found nucleotide replacement in position 15 from translation initiation site of inhA gene (inhA T_15). There were also two other mutations found in positions 9 and 12 from translation initiation site in the promotor locus of ahpC gene (ahpC_9 and ahpC_12). In general, INH-resistant strains in Kyrgyz Republic is quite homogenous and is predominantly represented by Ser → Thr replacement in codon 315 of katG gene ( Table 4).

Identification of MDR Mycobacterium tuberculosis using phenotypic DST and biochip
Cultural method revealed 174 MDR specimens out of 633, which are resistant to both RIF and INH. Biochip method confirmed MDR in 157 (90%) out of 174 (Table 5). Cultural method yielded 459 non-MDR specimens out of the overall number N = 633. TB-BIOCHIP confirmed non-MDR in 446 (97%) cases out of 459, and such susceptibility was classified as true susceptibility to RIF and INH. Therefore, TB-BIOCHIP's sensitivity in MDR Mycobacterium tuberculosis identification was 90%, whereas the specificity was 97%, and the agreement rate was 95% (603 out of 633) ( Table 5).

Mutations in rpoB, katG, inhA and ahpC genes in MDR strains
MDR is a simultaneous resistance to RIF and INH, usually in rpoB gene leading to RIF-resistance, and katG, inhA or ahpC genes, associated with INH-resistance. 157 (25%) specimens in our study were suggestive of MDR. Table 6 shows that 86.6% (136/157) of all MDR cases was a combined mutation in rpoB and katG genes; 2.5% (4/157) cases was a combination of rpoB and inhA genes mutation, whereas combined mutation in three genes (rpoB, katG and inhA) was found in 10.9% (17/157) specimens. Dominating mutations were Ser531 → Leu replacement in rpoB gene combined with Ser315 → Thr replacement in katG gene (50.3%) ( Table 6).
In INH-resistant Mycobacterium tuberculosis, most common mutations occurred in katG gene (91.2%), inhA gene (7%) and ahpC gene (1.8%). Amongst all katG gene mutations, Ser315 → Thr replacement in codon 315 was likely the most prevalent. Moreover, we confirmed TB-BIOCHIP system has comparable sensitivity and specificity with culture method. Unlike the latter, TB-BIOCHIP can make drug resistance profile available for a clinician within two days, timely guiding epidemiological decision to isolate patients with MDR and correct their treatment almost immediately after the diagnosis.
The majority of RIF-resistant strains in fact have mutation in the so-called "hot" locus of rpoB gene (81 b. p., codon 507-533rifampicin resistance determining region, RRDR). rpoB gene codes beta-subunit of DNA-dependent RNA-polymerase which act as a major target for RIF [3,18]. Up to 95-98% of RIFresistant strains exhibit mutations in rpoB gene, whereas 90-95% of these mutations are located in RRDR [3,18]. Despite such miscellaneous mutation locations, most of them are located in three rpoB codons: 531, 526 and 516.
In general, our findings of most prevalent codon 531 with Ser531→Leu (64.8%) replacement in RIF-resistant strains in Kyrgyzstan TB patients somewhat similar to other former Soviet countries and territories with high burden of MDR TB in the prevalence and the profile of rpoB gene mutations [14,16,[19][20][21][22][23][24]. In general, Ser531 → Leu mutation is the prevalent among RIF-resistant strains, fund all over the world [3].
As opposed to RIF, Mycobacterium tuberculosis resistance to INH is associated with mutations in a number of genes. INH is a pro-drug, and its activation in a cell is performed by catalase-peroxidase, coded by katG gene in Mycobacterium tuberculosis [25]. INH activation entails a few forms of active oxygen, including superoxide, peroxide and hydroxyl radicals, nitrogen oxide isonicotinic acid radical that act at various targets in Mycobacterium tuberculosis cells, however enol-acidic phosphate-reductase is the principal link in mycolic acids synthesis.
MDR, which is a combined resistance to RIF and INH, is a serious issue given the economic cost of treatment. Of note, spontaneous chromosome mutations leading to resistance to RIF and INH in a given population have a probability of 3.5 × 10 − 6 and 3.1 × 10 − 8 [3,4], whereas the probability of MDR is times smaller and equals 9 × 10 − 14 , since responsible genes are located in different sites. This explains acceptable efficacy of this combination to treat TB in cases when the agent was initially susceptible to these medications [30]. Moreover, resistance to RIF develops the latest, therefore, RIF-resistance is a surrogate marker of MDR [30,34,35]. Resistance to RIF alone, however, develops exceptionally seldom, and 90% of RIFresistance is combined with resistance to INH. Our TB-BIOCHIP tests demonstrated such pattern. Thus, 28 (15%) specimens out of 185 RIF-resistant strains were resistant to RIF only, whereas 157 (85%) cases of rpoB gene mutations coincided with katG, inhA and ahpC genes mutations, usually responsible for INH-resistance.
This analysis has a number of strengths. Firstly, we had an opportunity to take advantage of the use of TB-BIOCHIP method to verify mutations in rpoB, katG, inhA and ahpС gene. The method has been shown to confirm 88% RIF-resistant and more than 80% INHresistant strains within a day. Secondly, this study, to our best knowledge, is the first attempt to describe mutations in rpoB, katG, inhA and ahpС genes in Kyrgyz Republic, which can explain MDR to TB. We should also name the opportunity to include patients and specimens from all locations around the country, including low-and high-altitude towns, as another strength of this analysis. Finally, such large sample size may be one of very few reports from this region. As for limitations, we were not able to repeat TB-BIOCHIP analysis after 2 months of treatment once cultural method results were obtained due to technical reasons. TB-BIOCHIP has been shown to have moderate sensitivity. Its main advantage is the opportunity to obtain fast results, almost within 2 days. The timing and accuracy are essentially important for primary TB patients, when relevant treatment is pivotal to ensure treatment success and immediate segregation of patient groups in order to block the spread of resistant strains and formation of MDR. TB-BIOCHIP helps not only obtain fast results, but also enables to identify genetically opposing RIF and INH-resistant Mycobacterium tuberculosis variants. All these advantages justify the implementation and wide use of TB-BIOCHIP for express detection of RIF-and INH-resistant strains.
With currently moderate sensitivity, the method should be further improved and adapted to clinical practice, since missing 10% of cases in high TB/MDRTB incidence settings is challenging. Given that the biochip's efficiency depends on the number of available cells for DNA test with at least 500 cells needed, better specimens' collection and preparation may be the way to improve TB-BIOCHIP performance as it may alleviate DNA loss. Our study did not imply TB-BIOCHIP cost effectiveness analysis, however we consider the ability of TB-BIOCHIP to detect MDR Mycobacterium tuberculosis within 2 days as its main advantage. Saving time to the result will  probably substantially contribute to reducing the cost of patient's management. At present, a number of molecular and genetic methods of resistance verification are either recommended or in use in various settings. WHO recommends Xpert MBT/ RIF and Genotype MTBDRplus (Hain Lifescience GmbH, Nehren, Germany). Xpert MBT/RIF can verify RIFmutations and, in fact, can only detect Mycobacterium tuberculosis, resistant to RIF only. Genotype MTBDRplus can confirm resistance to two medications, such as RIF and INH, as a result of only three genes mutations, including rpoB, katG and inhA. As opposed to these two options, TB-BIOCHIP can confirm resistance to both RIF and INH resulting from mutations in four genes, including rpoB, katG, inhA and ahp.
We explain the discrepancies between TB-BIOCHIP and Phenotypic DST in detecting susceptibility with a number of reasons including rare mutations, which cannot be verified with the current biochip version; mixed population in a specimen, which may comprise both resistant and susceptible strains, which may show non-homogenous growth in Lowenstein-Jensen culture medium. Moreover, not all mutations will reflect phenotypic changes when cultivating strains in a solid culture medium with set concentrations of medications. Finally, TB-BIOCHIP can detect mutations associated with RIFand INH-mutations in both alive and dead Mycobacterium tuberculosis.

Conclusions
In Kyrgyz Republic, Ser531 → Leu replacement in rpoB gene was the commonest variant of RIF-resistance, whereas Ser315 → Thr mutation in katG gene could likely explain most of INH-resistance using TB-BIOCHIP analysis. MDR strains in our country likely developed most of their resistance because of combined mutations Ser531 → Leu in rpoB gene and Ser315 → Thr in katG gene. Taken together, this analysis will help guide treatment of patients with resistant strains and reduce the overall burden of the disease in the country.