Background

Tuberculosis (TB) is a common life-threatening infectious disease caused by Mycobacterium tuberculosis (MTB) [1]. The standard new therapy includes a six-month treatment of four recommended first-line drugs, i.e., isoniazid, rifampin, pyrazinamide and ethambutol [2]. However, the misuse of these antibiotics has led to the emergence of multidrug-resistant (MDR) strains of MTB [3]. According to the WHO report in 2016, Pakistan is ranked among the top five countries, with a prevalence of 56% of global TB and among the high-burden drug-resistant countries [4]. The incidence of drug resistance (MDR/RR-TB) is twenty-six thousand (16–36) with 14 (8.5–19) per hundred thousand individuals in a population. This situation leaves a high challenge to the TB control program in the country.

Khyber Pakhtunkhwa (KPK), a region in Pashtun, is the third largest province of Pakistan, with an area of 74,521 km2 and a population of approximately 30,523,371 individuals. Tuberculosis remains a major public health problem and one of the neglected health areas in the past. Recently, a TB control program has been launched at Hayatabad Medical Complex Peshawar that regularly monitors the incidence of TB as well as drug susceptibility testing (DST) in the population.

Pyrazinamide (PZA), a key first-line antibiotic used for the short-course treatment of 6 months, kills dormant tubercle bacilli at an acidic pH, whereas other drugs fail to work during the early severe stages of chemotherapy [5]. Due to some technical and buffering issues of drug susceptibility testing (DST) for PZA, the results of phenotypic resistance are not always reliable [6]. The MGIT 960 system is the most reliable method to perform PZA-DST [7, 8] and is currently the only available phenotypic test method to explore PZA susceptibility. MTB isolates are cultured in the presence of PZA under acidic conditions as required for conversion into pyrazinoic acid (POA) [8], the active form of pyrazinamide in vivo, activated by MTB PZase (PncA). These issues have prompted efforts towards molecular methods of PZA resistance [9].

Mutations occurring in the pncA gene are most commonly associated with pyrazinamide (PZA) resistance [9,10,11,12].

In Khyber Pakhtunkhwa, there are no molecular studies to explore pncA mutations in the KPK Pashtun-dominant region. Here, we aimed to compare phenotypic resistance to PZA to genotype and identify mutations in pncA among PZA-resistant isolates circulating in this epidemiologically distinct area in a Pashtun-dominant population, which may be useful in tracing the transmission in patients.

Material and methods

Ethical considerations

The present investigation was approved by the Institutional Ethics Committee of CUST Islamabad and Provincial Tuberculosis Reference Laboratory (PTRL) KPK under reference number PTP/PTRL-402/16. Prior to the study, informed consent was obtained from each TB patient, however, the results were not linked back to individual patients.

Study samples

All samples were processed at the BSL-III facility of PTRL, Hayatabad Medical Complex (HMC). The lab receives TB cases from the entire province, which is facilitated by the MGIT 960 system of drug susceptibility testing. The data for TB patients were collected from their guardians or caretakers.

Sample processing, isolation and mycobacterial culture

The samples were processed using the N-acetyl-L-cysteine–sodium hydroxide (NALC–NaOH) concentration method [13] in a Falcon tube containing an equal volume of the NaOH/N-acetyl-L-cysteine (NALC), which was subsequently vortexed and incubated at room temperature for 15 min for decontamination and digestion. Next, 50 ml of phosphate buffer was transferred to each tube, followed by centrifugation at 3000 rpm for 15 min. The supernatant was transferred to a fresh tube containing 5% phenol, while the pellet was mixed with phosphate buffer and cultured on Lowenstein–Jensen medium (LJ) in MGIT tubes containing 7H9 media.

Drug susceptibility testing (DST)

Drug susceptibility testing of PZA was performed through the automated BACTEC MGIT 960 system (BD Diagnostic Systems, NJ, USA) [14]. Mycobacterium tuberculosis H37Rv and Mycobacterium bovis were used as susceptible and resistant controls, respectively. A sample was marked as PZA resistant if growth was found at 100 μg/ml of the PZA critical concentration. DST for resistant isolates was repeated for confirmation of drug resistance. The PZA-resistant samples were further subjected to DST with isoniazid (INH), rifampin (RIF), ethambutol (EMB), amikacin (AMK), streptomycin (SM), capreomycin (CAP), ofloxacin (OFX) and kanamycin (KM) through the BACTEC MGIT 960 system, with critical concentrations of drugs as per the policy guidelines of the WHO (WHO 2014) [15]. The resistant samples were further manually assessed to confirm the growth of MTB against the critical drug concentration.

DNA extraction, PCR amplification and sequencing

Genomic DNA from PZA-resistant isolates were extracted by sonication [16, 17]. One microliter of fresh culture was transferred from a Mycobacterium Growth Indicator Tube (MGIT) to a microcentrifuge tube and boiled at 86 °C for 30 min using an Echotherm™ IC22 Digital, Chilling/Heating Dry Bath followed by 15 min sonication using a sonicator (ELMASONIC S30). All the samples were centrifuged for 5 min at 10,000 rpm. The supernatant containing DNA was stored at − 20 °C. The fragments containing pncA were amplified using previously reported primers (pncA-F = 5GCGTCATGGACCCTATATC-3 and pncA-R = 5 AACAGTTCATCCCGGTTC-3=) [18]. Each 50-μl PCR contained 0.1 μl of each DNTs, 0.8 μl of Taq (New England Biolabs, UK)), 5 μl of PCR buffer, 3 μl of MgCl2, 1 μl of each forward and reverse primer, 34.8 μl of molecular grade water and 4 μl of genomic DNA. The PCR conditions were 5 min at 94 °C for denaturation, followed by 30 cycles of 30 s at 94 °C, 30 s at 56 °C, and 72 °C for 1 min, with an extension step at 72 °C for 5 min, as previously described. The PCR product was analyzed by 6 Applied Biosystems 3730xl (Macrogen Korea).

Data analysis

The sequence data obtained were loaded into Mutation Surveyor V5.0.1 software [19]. The data were analyzed and compared with the PncA (Rv2043c) gene of RefSeq database of NCBI (NC_000962), while the patient data were entered by using Epi-Data entry version 3.1 software and analyzed through Epi-Data analysis software.

Results

Socio-demographic characteristics

A total of 4518 samples were collected from TB subjects from all districts of KPK. Among these individuals, 753 subjects were culture positive, with ages ranging from 8 to 76 years (median age = 34.34). A majority of the cases were never treated (diagnostics) (44/69). All patients were KPK residents with Pushto as the main language (Table 1).

Table 1 Drug susceptibility profile and socio-demographic data of 69 PZA resistant patients
Full size table

Drug susceptibility pattern

By using the BACTEC MGIT 960 system, 69 (14.8%) isolates were classified as PZA resistant. All PZA resistant isolates, 26 PZA-sensitive isolates and one H37Rv isolate were sequenced to analyze the mutations in the 561 bp region of pncA. Multidrug drug resistance (MDR) and extreme drug resistance (XDR) were detected in 52/69 (75.35%) and 6/69 (8.69%) isolates, respectively.

Risk factors such as age, gender, history, reason, disease type and resistance type are presented in Table 1. The presence of a high number of MDR isolates (52/69) in PZA-resistant isolates shows the major risk factor in transmission and treatment failure.

Mutation in PZA-resistant and PZA-susceptible isolates

Mutations were investigated in both resistance and susceptible isolates in the coding region (561 bp) of pncA (Table 2). Among the 69 PZA-resistant isolates 51 (74%), thirty-six different mutations with fifteen novel mutations, including 194_203delCCTCGTCGTG and 317_318delTC, were detected; but these variations were not found in TBDRM and GMTV Databases [20, 21]. The most common mutations detected at positions 287 and 423 were Lys96Thr (n = 4) and Ser179Gly (n = 4), respectively. A majority of these variations were substitution mutations, except for three deletions, 194_203 del CCTCGTCGTG (n = 1), 317_318delTC (n = 1), 530 del C (n = 3). A common synonymous mutation at position 195 C > T (Ser65Ser) was observed in both resistant (n = 22) and susceptible isolates (n = 16).

Table 2 Mutations in pncA gene of M. tuberculosis in 51 isolates
Full size table

Comparing phenotypic PZA-resistant with DNA sequencing

Among sensitive isolates we did not detect any nonsynonymous mutations in the coding region of pncA. To estimate the performance of DST compared with the pncA sequencing result, the genotypic data and phenotypic data for all 69 resistance isolates were evaluated. Considering phenotype as a reference, among the 69 resistant isolates, 51 (74%) isolates showed mutations, with sensitivity of 79.31% (95% CI, 69.29 to 87.25%) and specificity of 86.67% (95% CI, 69.28 to 96.24%).

Discussion

PZA is a distinctive antituberculosis drug that plays a key role in shortening TB treatment. PZA kills nonreplicating persistent MTB and is prescribed in both susceptible and MDR-TB treatment. After conversion into its active form POA by pyrazinamidase (PZase), PZA remains active at low pH during acidic stress [11]. However, in a large number of cases, MTB patients develop resistance against PZA that led to the survival of persistent bacteria. Conventional methods of PZA DST increase the level of false resistance that may result from media buffering issues and large inoculum sizes, where the acidic environment is required for drug action but inhibits the growth of MTB [6, 22, 23]. Under such conditions, the most reliable method in the present scenario is the molecular detection of PZA-resistance, which involves sequencing the pncA gene to assess mutations in the 561 bp coding region and upstream regulatory region. In the present study, more than half of the tested PZA-resistant isolates were also MDR-TB, 52/69 (75.35%) isolates, consistent with the results of previous studies [12, 24]. Previous studies [25, 26] have also shown a correlation between mutations in the pncA gene and phenotypic PZA-resistance. Based on these findings, we report mutations in 51 (74%) resistant isolates that harbor 36 mutations in the coding region of pncA, with sensitivity and specificity of pncA sequencing of 79.31% (95% CI, 69.29 to 87.25%) and 86.67% (95% CI, 69.28 to 96.24%), respectively. Streicher et al. and Whitfield reported better sensitivity and specificity for pncA sequencing compared with that of MGIT 960 DST of 90.9, 100% [27] and 95.0 (95% CI 92.1–98.0), 99.1 (95% CI 98.4–99.9) [28]. Miotto et al. (2014) identified 280 mutations in 1950 clinical strains [9], which were categorized into four groups, 1) very high confidence resistance mutations, 2) high-confidence resistance mutations, 3) mutations with an unclear role and 4) mutations not associated with phenotypic resistance based on confidence level. We detected 12 mutations with very high confidence resistance, while the rest of mutations detected have been found in Miotto unclear category (Table 2). The mutations 211C > T, 212A > G, 226A > C, 286A > C and 422A > C in the present study (Table 2) were previously shown as very high confidence resistance mutations [9, 12, 29]. Molecular biomarkers that could specifically target the first two categories should be developed [9].

Tan et al. (2014) reported that each geographical region has a diverse type of variations in pncA. Isolates from Southern China exhibited a scattered type of mutations in 561 bp region, which remains a complex target in the development of diagnostic biomarkers in identification of all resistance conferring mutations [26]. Some strains, which were PZA resistant by conventional DST, lack any mutations in PncA and its regulatory gene, suggesting other targets of drug and issues concerning DST.

The residues Cys138, Asp8, Lys96 and Asp49, His51, His57, and His71 are present in the active and metal binding sites [30, 31] of the pncA-encoded enzyme pyrazinamidase (PZase). We identified mutations dispersed throughout the pncA gene (35A > C---538G > T) nearby the area of metal binding and active site amino acids (46–76 and 133–146). We detected mutations that are important for enzyme catalysis and metal binding (Table 2). However, we did not detect any mutations in the 18 PZA-resistant MTB isolates, suggesting the involvement of other genes RpsA and PanD (aspartate decarboxylase) [32, 33]. A potential new target of PZA, the clpC1 (Unfoldase) gene, which encodes a family of ATPases, was identified in PZA-resistance isolates in addition to the previously identified genes pncA, rpsA and panD [34]. However, the role of these genes (rpsA, panD and clpC1) in PZA resistance is small compared to that of PncA. In a more recent study four new efflux proteins Rv0191, Rv1667c, Rv3756c and Rv3008 were implicated in PZA/POA resistance [35]. These findings suggest a new mechanism for PZA resistance in MTB. Further investigations are needed to identify the quantitative role of all these targets and mechanisms in PZA-resistant MTB for better management of drug-resistant TB.

In conclusion, considering phenotype as a reference, among the 69 PZA resistant isolates, 51 (74%) showed mutations with sensitivity of 79.31% (95% CI, 69.29 to 87.25%) and specificity of 86.67% (95% CI, 69.28 to 96.24%). The mutations 33C > A, 53C > A, 194_203 Del CCTCGTCGTG, 205C > A, 317–18 Del TC, 331G > T, 376G > A, 419G > A, 430G > A, 449G > C, 508G > C, 519G > A, 522G > A, 530DEL C and 535A > G were not found in the GMTV and TBDRM databases and neither in previous studies suggesting the need of some more studies from distinct geographical regions must be carried out for some novel mutations confined to that specific areas. Majority of mutations were of high confidence intervals and uncharacterized category in resistance. Molecular methods to investigate PZA resistance by screening mutations in pncA gene in distinct epidemiological regions offer a much more rapid alternative compared to that of conventional bacteriology. Mutations in pncA gene are highly linked with resistance to PZA and scattered in the entire coding region of pncA. Further, we found an association between PZA resistance and resistance to other important first line drugs, INH and RIF, which is a major hurdle in the treatment management of MDR TB. This high frequency of pncA mutations from geographically distinct regions recommends the WHO guidelines to empirically use the pyrazinamide in drug-resistant TB should be considered. Further studies with large sample size may strengthen these findings to identify the mutations in PZA-resistant isolates specific to certain geographical areas for the better treatment and development of geographically specific biomarkers.