Frequency of mutations in rifampicin and isoniazid resistant isolates of M. tuberculosis: an analysis from Central India

Abstract

Background

The spread of drug-resistant tuberculosis has challenged tuberculosis control strategies globally. The present study aims to analyze the frequency of mutations in rpoB, katG and inhA genes in strains of M. tuberculosis complex (MTBC) circulating in Central India. It is anticipated that the findings may provide a starting point to understand the evolutionary success of drug-resistant strains of MTBC in this region.

Methods

Line probe assay was carried out on 720 consecutive sputum samples of MDR suspects from June 2012 to May 2013. Mutation frequencies in the rpoB, katG and inhA genes were analyzed.

Results

Mutations were identified in 269 (37.6%) samples, as follows: 55 (7.6%) samples had mutations conferring resistance to only isoniazid, 84 (11.6%) had mutations conferring resistance to only rifampicin and 130 (18%) isolates had mutations conferring resistance to both isoniazid and rifampicin. The most frequent mutation in the rpoB gene was at codon S531L, seen in 141 (19.5%) isolates. The most frequent mutation in the katG gene was at codon S315T1, seen in 151 (20.9%) isolates; and in the inhA gene at codon C15T, seen in 21 (2.9%) isolates. Some unidentified mutations were also observed.

Conclusion

The patterns and the frequency of the mutations identified in this study indicate the most frequent mutations at S531L codon in the rpoB gene, S315T1 codon in the katG gene and C15T codon in the promoter region of the inhA gene. Controlling the emergence and spread of MDR TB requires an understanding of the evolution of these mutations.

Background

The most recent report of the World Health Organization (WHO) on drug-resistant tuberculosis clearly indicates that the fight against tuberculosis is far from over.¹ The spread of drug-resistant tuberculosis has challenged tuberculosis control strategies globally.² The WHO-endorsed line probe assay (LPA), based on multiplex polymerase chain reactions, is used worldwide as a part of tuberculosis control programmes for simultaneous identification of M. tuberculosis complex (MTBC) and detection of drug resistance to rifampicin (RIF) and isoniazid (INH) among these strains.^3-5 It gives results within 48 hours and provides information on the most common mutations and levels of resistance.⁶

We have attempted to analyze the frequency of mutations in the rpoB, katG and inhA genes that code for RIF and INH sensitivity in strains of MTBC circulating in the state of Madhya Pradesh (India) by LPA. Madhya Pradesh, in Central India, is a densely populated region with 72.6 million people.⁷ Currently, there is no data from this region on the prevalent mutation patterns of these genes.

Methods

Ethical approval was provided by the Institutional Ethics Committee (IEC) as per approval document IEC/15/Micro/14, dated 01.01.2015, of the Bhopal Memorial Hospital & Research Center (BMHRC), Bhopal. We analyzed the results generated from the LPA tests from the mandated diagnostic workflow of the two accredited Mycobacteriology laboratories in the state of Madhya Pradesh in Central India (National Reference Laboratory, BMHRC, Bhopal and Intermediate Reference Laboratory, Indore). LPA was carried out on sputum samples of MDR suspects defined as per criteria ‘C’ of the Revised National Tuberculosis Control Programme (RNTCP) (MDR suspect criteria detailed in Table 1). Samples were received from all districts of Madhya Pradesh under the RNTCP from June 2012 to May 2013. A total of 720 sputum samples were processed within 48 hours of collection by the NALC-NaOH decontamination method.⁸ The sediments were suspended in 1 mL sterile phosphate buffer (pH 6.8). Smears of the sediments were stained Ziehl-Neelsen. DNA extraction of all smear-positive decontaminated samples was carried out using the GenoLyse® kit (Hain Lifescience GmbH, Nehren, Germany) as per manufacturer’s instructions.⁹ The GenoType MTBDRplus (Hain Lifescience GmbH, Nehren, Germany) line probe assay for detection of rifampicin and/or isoniazid resistance was carried out according to the manufacturer’s instructions.¹⁰

Table 1. MDR suspect criteria as defined by the Revised National Tuberculosis Control Programme, India.

Criteria A

• All failures of new TB cases

• Smear positive previously treated cases who remain smear positive from the 4th month onwards

• All pulmonary TB cases who are contacts of known MDR TB cases

Criteria B – in addition to Criteria A:

• All smear positive previously treated pulmonary TB cases at diagnosis

• Any smear positive follow up result in new or previously treated cases

Criteria C – in addition to Criteria B

• All smear negative previously treated pulmonary TB cases at diagnosis,

• HIV TB coinfected cases at diagnosis

Mutations in codons 505 to 533 of the rpoB gene were analyzed. The four mutation probes for the rpoB gene on the LPA strip detected the mutations at the D516V (rpoB MUT1), H526Y (rpoB MUT2A), H526D (rpoB MUT2B) and S531L (rpoB MUT3). The two mutation probes for the katG gene detected mutations at the codons S315T1 (katG MUT1) and S315T2 (katG MUT2). The four mutations probes for the inhA gene detected mutations at the codons C15T (inhA MUT1), A16G (inhA MUT2), T8C (inhA MUT3A) and T8A (inhA MUT3B). Missing wild type bands represented known/unknown mutation/s.

On an assay strip, 21 reaction zones representing various wild types (WT) and mutation (MUT) bands of the rpoB, katG and inhA genes were probed. The presence of mutation band/s at rpoB, katG or InhA gene regions or lack of at least one of the wild-type bands was considered indicative of resistance to the respective drug. The presence of all wild-type bands and absence of mutation bands was indicative of susceptibility to the drug. Each individual test was verified for the presence of five control zones that included a conjugate control, an amplification control, and three locus control zones (rpoB, katG and inhA). The assay was considered valid if bands for master mix and extraction negative controls were absent, and bands for conjugate and amplification controls were present.

Results

Out of a total of 720 DNA isolates from 720 samples, mutations were identified in 269 (37.6%) isolates. Fifty five (7.6%) isolates had mutations conferring resistance to only INH (INH monoresistant), 84 (11.6%) had mutations conferring resistance to only RIF (RIF monoresistant) and 130 (18%) isolates had mutations conferring resistance to both INH and RIF (multidrug resistant or MDR isolates).

Mutations in the rpoB gene

There were 29 different band patterns detected in the rpoB gene, which codes for the β subunit of RNA polymerase and sensitivity to RIF. The most frequent mutation was detected at codon S531L, which was seen in 141 (19.5%) isolates. The second most prevalent mutation in the rpoB gene was at codon D516V, which was seen in 21 (2.9%) isolates. The frequency of mutations at codons H526Y and H526D were found to be 1.3% (10 isolates) and 0.9% (7 isolates) respectively.

Forty three (5.9%) isolates had one or more missing WT bands while none of them were accompanied by MUT band in the hot spot region (81 base pair region), which is the rifampicin resistance-determining region of the rpoB gene, indicating unknown mutations. There were 23 isolates (3.1%) in which a single WT band deletion was found. Single WT2 band deletion was seen in 12 (1.6%) isolates, single WT3 band deletion in 3 (0.4%) isolates, single WT4 band deletion in 1 (0.1%) isolate, single WT7 band deletion in 5 (0.6%) isolates and single WT8 band deletion in 2 (0.2%) isolates. Two simultaneous WT band deletions with absence of MUT band were found in 19 (2.6%) isolates. WT 1 and 7 deletion in 2 (0.2%) isolates, WT 2 and 3 deletion in 3 (0.4%) isolates, WT 2 and 7 deletion in 1 (0.1%) isolate, WT 3 and 4 deletion in 9 (1.2%) isolates, WT 4 and 5 deletion in 2 (0.2%) isolates, WT 5 and 6 deletion and WT 7 and 8 deletion in 1 (0.1%) isolate each were seen. There was only 1 (0.1%) isolate in which 3 WT bands were found to be deleted simultaneously along with absence of MUT band. This isolate showed deletion at 2, 3 and 4 WT band positions.

All WT bands (1 to 8) of the rpoB gene were present in 596 (82.7%) isolates. Five hundred and six (70.2%) isolates did not show any mutant DNA sequences, thus indicating sensitivity to RIF. Ninety (12.5%) isolates displayed varied mutations along with presence of all WT bands. There were 70 (9.7%) isolates with MUT 3 band representing mutation at codon S531L, 6 (0.8%) isolates with MUT1 band indicating mutation at codon D516V, 5 (0.6%) isolates with MUT2A band indicating mutation at codon H526Y, 3 (0.41%) isolates with MUT2B representing mutation at codon H526D. Four (0.5%) isolates having both MUT1 and MUT3 bands representing mutations at codons D516V and S531L respectively; 1 (0.1%) isolate with presence of both MUT2A and MUT3 bands representing mutations at codons H526Y and S531L and 1 (0.1%) isolate with presence of all 4 mutation bands; MUT1, MUT2A, MUT2B and MUT3, representing mutations at codons D516V, H526Y, H526D and S531L respectively were seen.

Eighty one isolates (11.2%) had one or more missing WT bands accompanied by a mutation band. There were 61 isolates (8.4%) with absence of WT8 band and presence of MUT3 band (indicating mutation at codon S531L); 10 (1.3%) isolates with absence of both WT3 and WT4 bands and presence of MUT1 band (indicating mutation at codon D516V); 3 (0.4%) isolates with absence of both WT2 and WT8 bands along with presence of MUT3 probe (indicating mutation at codon S531L); 3 (0.4%) isolates with absence of WT7 along with the presence of MUT2B (indicating mutation at codon H526D) band; 2 (0.2%) isolates with missing WT7 band and presence of MUT2A band (indicating mutation at codon H526Y); 1 (0.1%) isolate with absence of both WT2 and WT8 band accompanied by presence of MUT2A band (indicating mutation at codon H526Y) and 1 (0.1%) isolate with absence of both WT3 and WT4 with presence of MUT3 band (indicating mutation at codon S531L) – Table 2.

Table 2. Banding patterns obtained by line probe assay in the rpoB, katG and inhA genes.

Gene	Δ WT	WT	Mutation	No (%)
rpoB	Δ WT2,8	WT1,3,4,5,6,7	S531L	3 (0.4%)
	Δ WT2,8	WT1,3,4,5,6,7	H526Y	1 (0.1%)
	Δ WT3,4	WT1,2,5,6,7,8	D516V	10 (1.3%)
	Δ WT3,4	WT1,2,5,6,7,8	S531L	1 (0.1%)
	Δ WT7	WT1,2,3,4,5,6,8	H526D	3 (0.4%)
	Δ WT7	WT1,2,3,4,5,6,8	H526Y	2 (0.2%)
	Δ WT8	WT1,2,3,4,5,6,7	S531L	61 (8.4%)
	…	WT1,2,3,4,5,6,7,8	D516V	6 (0.8%)
	…	WT1,2,3,4,5,6,7,8	H526Y	5 (0.6%)
	…	WT1,2,3,4,5,6,7,8	H526D	3 (0.4%)
	…	WT1,2,3,4,5,6,7,8	S531L	70 (9.7%)
	…	WT1,2,3,4,5,6,7,8	D516V & S531L	4 (0.5%)
	…	WT1,2,3,4,5,6,7,8	H526Y & S531L	1 (0.1%)
	…	WT1,2,3,4,5,6,7,8	D516V, H526Y, H526D & S531L	1 (0.1%)
	Δ WT2	WT1,3,4,5,6,7,8	…	12 (1.6%)
	Δ WT3	WT1,2,4,5,6,7,8	…	3 (0.4%)
	Δ WT4	WT1,2,3,5,6,7,8	…	1 (0.1%)
	Δ WT7	WT1,2,3,4,5,6,8	…	5 (0.6%)
	Δ WT8	WT1,2,3,4,5,6,7	…	2 (0.2%)
	Δ WT1,7	WT2,3,4,5,6,8	…	2 (0.2%)
	Δ WT2,3	WT1,4,5,6,7,8	…	3 (0.4%)
	Δ WT2,7	WT1,3,4,5,6,8	…	1 (0.1%)
	Δ WT3,4	WT1,2,5,6,7,8	…	9 (1.2%)
	Δ WT4,5	WT1,2,3,6,7,8	…	2 (0.2%)
	Δ W5,6	WT1,2,3,4,7,8	…	1 (0.1%)
	Δ WT7,8	WT1,2,3,4,5,6	…	1 (0.1%)
	Δ WT1,2,3	WT4,5,6,7,8	…	1 (0.1%)
	…	WT1,2,3,4,5,6,7,8	…	506 (70.2%)
katG	Δ WT	…	S315T1	142 (19.7%)
	Δ WT	…	…	18 (2.5%)
	…	WT	S315T1	8 (1.1%)
	…	WT	S315T1 & S315T2	1 (0.1%)
	…	WT	…	551 (76.5%)
inhA	Δ WT1	WT2	C15T	18 (2.5%)
	Δ WT1,2	…	…	1 (0.1%)
	Δ WT1	WT2	…	1 (0.1%)
	…	WT1,2	C15T	3 (0.4%)
	…	WT1,2	T8C	3 (0.4%)
	…	WT1,2	…	694 (96.3%)

WT – wild type, Δ WT – absence of wild type band.

Mutations in the katG gene

There were six different band patterns detected in the katG gene, which codes for catalase peroxidase enzyme and sensitivity to INH. Five hundred and fifty one (76.5%) isolates were sensitive to INH, and showed a WT band without any MUT band. Mutation in the katG gene at the codon S315T1 was found to be the most frequent, with 151 (20.9%) isolates showing this mutation. Nine (1.2%) isolates displayed at least one MUT band along with presence of the WT band. Eight (1.1%) isolates had MUT1 band indicating mutation at codon S315T1 along with presence of the WT band, and a single (0.1%) isolate had both MUT1 and MUT2 bands indicating mutations at codon S315T1 and S315T2 respectively along with presence of the WT band.

There were 142 (19.7%) isolates that did not have the WT band, and had only MUT1 band (indicating mutation at codon S315T1). The entire katG gene was absent in 18 (2.5%) isolates. There were a total of 151 (20.9%) isolates with MUT 1 band (with or without the WT band), indicating mutation at codon S315T1 (Table 2).

Mutation in the inhA gene

Six different band patterns were identified at the promoter region of inhA gene, which codes for the NADH enoyl ACP reductase enzyme and sensitivity to INH. There were 21 isolates (2.9%) with MUT1 band, indicating mutation at codon C15T and 3 isolates (0.4%) with MUT3A band, indicating mutation at codon T8C. Six hundred ninety four isolates (96.3%) with presence of all WT bands did not show any MUT band representing sensitivity to INH. Six (0.8%) isolates had all WT and at least one MUT band in the inhA gene. Three (0.4%) isolates had MUT1 band indicating mutation at codon C15T, along with presence of all WT bands and 3 (0.4%) had MUT3A band indicating mutation at T8C, along with simultaneous presence of both WT1 and WT2 bands. Eighteen (2.5%) isolates had only a WT2 band and a MUT1 band. The entire inhA gene was deleted in 1 (0.1%) isolate. Unknown mutation was represented by the absence of WT1 band accompanied by the presence of WT2 and MUT1 band (indicative of mutation at codon C15T) and was found in 1 (0.1%) isolate (Table 2).

Discussion

M. tuberculosis exhibits multiple mechanisms to inactivate anti-TB drugs owing to point mutations, insertions or deletions in the genes coding for sensitivity to the respective drug.¹¹ LPA detects common mutations in the rpoB, katG and inhA genes that result in resistance to RIF and INH. Earlier studies have shown that the frequency of mutation is highest (96%) in the core, hot spot region of the rifampicin resistance-determining region of the rpoB gene.^12-14 Van Deun A et al.¹⁵ have reported 46.9% and 63%, S531L mutation frequencies from Bangladesh and Kinshasa respectively. These studies indicate that the most frequent mutation occurs at codon 531 of the rpoB gene that results in amino acid exchange from serine to leucine. Our data are consistent with the above findings, exhibiting the most frequent mutation in S531L codon in the rpoB gene. We also found mutations at codons D516V, H526D and H526Y most frequently. Van Deun A et al.¹⁵ however, have reported D516V mutation frequencies of 3.4% and 7.1%; H526D mutation frequencies of 6.9% and 2%; and H526Y mutation frequencies of 9.7% and 2.4% from Bangladesh and Kinshasa respectively. It appears that these mutation frequencies vary widely across geographical areas.

The predominant band patterns in RIF-resistant strains in our study were absence of WT3 and WT4 accompanied by the presence of MUT 1 probe; deletion of WT 7 band accompanied by the presence of MUT2A or MUT2B band and presence of MUT 3 band with absence of WT 8 band. This pattern of absence of specific WT band/s accompanied by presence of specific MUT band/s was consistent with the description in the laboratory manual for LPA.¹⁶ However, two unusual patterns were observed in this study. One was absence of WT2 and WT8 bands with presence of MUT 2A band (instead of MUT3 band), and the other was absence of WT3 and WT4 bands with presence of MUT3 probe band (instead of MUT2A or MUT2B band). These unique patterns may indicate that there are uncommon mutations in strains of MTB circulating in this geographical area. This emphasizes the need for phenotypic confirmation of atypical LPA results, that is, by culture and drug sensitivity testing.¹³ The presence of WT along with MUT bands in RIF-resistant isolates expressed either heteroresistance or mixed infection. The assay could not differentiate between the two.

Mutation of the S315T codon was the most common mutation of the katG gene in our study. This is similar to results from earlier studies.^13,17-19 An unusual pattern was observed in one isolate in which WT deletion band was not accompanied by a concomitant presence of a MUT band. This may be due to mutation of one or more genes other than katG and inhA, notably ahpC, oxyR, kasA, furA and ndh.^20,21

There were 6.11% unknown mutations in our study. These mutations may or may not have been present in this region earlier. Van Deun A et al.¹⁵ classified 511Pro, 516Tyr, 526Asn, 526Leu, 526Ser, 533Pro, and 572Phe rpoB mutations as “disputed” resistance, based on the discordant RIF drug susceptibility testing results. The unknown mutation patterns in the rpoB gene in our study may belong to the “disputed” resistance patterns as described by them. Further studies, using DNA sequencing, are needed to identify the nature of these mutations.

A limitation of the present study is that clinical data was not available for the patients.

Conclusion

In the rpoB gene, the frequency of mutation was the highest at the S531L codon in the hot spot core region. Mutation frequency was the highest at the S315T1 codon in the katG gene, and at the C15T codon in the promoter region of the inhA gene. Varied banding patterns were recognized by LPA indicating heteroresistance in the population. A considerable number of unidentified mutations were also observed. Further studies to elucidate these mutations would enhance our understanding of the epidemiology and transmission dynamics of drug resistant-strains in this geographical area and help provide the basis for effective strategies for control of drug-resistant TB in this region.