Correlations of mutations in katG, oxyR-ahpC and inhA genes and in vitro susceptibility in Mycobacterium tuberculosis clinical strains segregated by spoligotype families from tuberculosis prevalent countries in South America

Background Mutations associated with resistance to rifampin or streptomycin have been reported for W/Beijing and Latin American Mediterranean (LAM) strain families of Mycobacterium tuberculosis. A few studies with limited sample sizes have separately evaluated mutations in katG, ahpC and inhA genes that are associated with isoniazid (INH) resistance. Increasing prevalence of INH resistance, especially in high tuberculosis (TB) prevalent countries is worsening the burden of TB control programs, since similar transmission rates are noted for INH susceptible and resistant M. tuberculosis strains. Results We, therefore, conducted a comprehensive evaluation of INH resistant M. tuberculosis strains (n = 224) from three South American countries with high burden of drug resistant TB to characterize mutations in katG, ahpC and inhA gene loci and correlate with minimal inhibitory concentrations (MIC) levels and spoligotype strain family. Mutations in katG were observed in 181 (80.8%) of the isolates of which 178 (98.3%) was contributed by the katG S315T mutation. Additional mutations seen included oxyR-ahpC; inhA regulatory region and inhA structural gene. The S315T katG mutation was significantly more likely to be associated with MIC for INH ≥2 μg/mL. The S315T katG mutation was also more frequent in Haarlem family strains than LAM (n = 81) and T strain families. Conclusion Our data suggests that genetic screening for the S315T katG mutation may provide rapid information for anti-TB regimen selection, epidemiological monitoring of INH resistance and, possibly, to track transmission of INH resistant strains.

characterize mutations in katG, ahpC and inhA gene loci and correlate with minimal inhibitory concentrations (MIC) levels and spoligotype strain family. Mutations in katG were observed in 181 (80.8%) of the isolates of which 178 (98.3%) was contributed by the katG S315T mutation. Additional mutations seen included oxyR-ahpC; inhA regulatory region and inhA structural gene. The S315T katG mutation was significantly more likely to be associated with MIC for INH ≥2 μg/ mL. The S315T katG mutation was also more frequent in Haarlem family strains than LAM (n = 81) and T strain families.

Conclusion:
Our data suggests that genetic screening for the S315T katG mutation may provide rapid information for anti-TB regimen selection, epidemiological monitoring of INH resistance and, possibly, to track transmission of INH resistant strains.

Background
Tuberculosis (TB), a curable disease caused by M. tuberculosis, has never been adequately controlled in high prevalence countries because of inadequate funding of public health programs and limited access to health care caused by poverty. In the last several decades, the concurrent HIV epidemic has further accentuated the magnitude of the global TB burden. Further complicating the TB resurgence is the recent increase in the occurrence of simultaneous resistance to first line drugs, isoniazid (INH) and rifampin (RIF), that defines multidrug resistance (MDR), as well as, to second line drugs, resulting in extensive drug resistance (XDR) [1,2]. Although current control measures and short-term treatment schemes address the problem of drug resistance, knowledge on individual drug resistance profiles is needed for targeted intervention [3]. Global surveillance of M. tuberculosis drug resistance has been proposed to guide appropriate treatment policies [4]. Brazil and Peru are responsible for approximately 50% of the new TB cases in the Americas [5,2]. Moreover, 2,443 and 2,760 MDR-TB cases were reported respectively for Brazil from 2000 to 2006 [6] and Peru in just 2005 [7].
In the last years, molecular epidemiological approaches have shown that certain emerging M. tuberculosis strains, that induce more severe forms of TB, manifest higher failure/relapse than others. These features of certain isolates of M. tuberculosis strains, therefore, accentuate TB burden even in countries with good TB control programs, such as Vietnam [8][9][10]. Strains of the Beijing/W and Haarlem strain families of M. tuberculosis are emerging in certain global regions and are associated with drug resistance [11,12]. Importantly, specific mutations have been described in M. tuberculosis genes that are associated with resistance to rifampin or streptomycin and noted particularly in W/Beijing and Latin-American & Mediterranean (LAM) strain families [13].
The current view, since Middlebrook's original description, is that INH resistant strains of M. tuberculosis are less virulent; whether INH resistant and catalase-negative strains are indeed attenuated has been recently ques-tioned [14]. The mechanism for INH resistance is only partly elucidated. Resistance to INH is associated with mutations in several genes that include at least katG, inhA and ahpC. The katG gene encodes the enzyme catalase-peroxidase that functions to convert INH, which lacks antimycobactericidal activity, into an active compound [15]. The inhA (ORF) gene encodes an enoyl acyl carrier protein reductase involved in fatty acid synthesis. These fatty acids are the target of the active derivative of INH [4]. The inhA promoter gene region regulates the expression of an enoyl acyl carrier protein reductase. Mutations of this region may decrease the level of protein expression. The ahpC gene encodes alkyl-hydroperoxide reducatse involved in cellular regulation of oxidative stress [16]; mutations in the intergenic region oxyR-ahpC may also reduce the level of expression. The substitution of a single nucleotide of the amino acid at position 315 of katG (S→T), vary from 53% to 96% of INH resistant isolates [17,18]. Importantly, it was shown that the katG S315T mutation is associated with INH resistance without diminishing the virulence or transmissibility of M. tuberculosis strains [3,19]. The lack of attenuation associated with the katG S315T substitution and its high frequency among INH resistant clinical isolates suggests that the majority of these isolates will be virulent, and this premise was supported by a recent population-based molecular epidemiological study carried out in The Netherlands [20]. In this study, DNA fingerprinting demonstrated that, although INH resistant strains in general were less often transmitted between humans, the transmission of katG S315T mutants was similar to drug susceptible strains [20,18].
There is a paucity of information regarding the frequency and types of gene mutations associated with INH resistance among M. tuberculosis strains from South America. Moreover, studies of mutations associated with INH resistance have been limited in the scope of the genes assessed, the number of isolates evaluated, and lacked correlation with in vitro INH levels determined by minimal inhibitory concentration. Thus, we conducted a comprehensive characterization of mutations in the katG, oxyR-ahpC, and inhA genes in over 200 INH resistant M. tuber-culosis isolates from three MDR high prevalence countries from South America, namely, Argentina, Peru and Brazil and correlated the mutational data with minimal inhibitory concentration (MIC) level for INH and strain families as determined by spoligotyping.

Drug susceptibility testing
All isolates previously shown to be INH resistant by the proportion method were retested to determine the MIC levels. All isolates retested by MIC were INH resistant defined as ≥ 0.2 μg/mL. The majority of the isolates were resistant to ≥ 0.5 μg/mL INH.

Mutation frequency
We next characterized mutations in katG, ahpC and inhA (ORF or regulatory regions) gene loci. Among the 224 INH resistant M. tuberculosis isolates, the katG gene was the most frequently mutated gene (80.8%; 181/224). A mutation in codon 315 of the katG gene was present in 178 isolates. At this codon, the substitution from AGC to ACC leading to the amino acid change serine to threonine (S to T), seen in 166 (74.1%) isolates. In addition, a single nucleotide polymorphism (SNP) from AGC (S) to AAC (N) was seen in 9 isolates; and from AGC (S) to ACG (L) was noted for 3 isolates. In other regions of the katG gene, substitution SNPs were identified at codons 258, 299 and 300 (Table 1). We also screened for mutations in oxyR-ahpC and inhA (ORF and regulatory) gene loci previously reported to be associated with INH resistance. Mutations were also identified including in oxyR-ahpC (8.9%, n = 20 isolates), inhA regulatory gene region (9.8%, n = 22 isolates), and inhA ORF gene region (1.3%, n = 3 isolates) (see Table 1). Figure 1 depicts correlation of MIC level with frequencies of individual mutations and cumulative mutations. As shown, 99.8% of isolates with MIC ≤ 8 μg/ mL present at least one mutation. The data suggest that with increasing MIC levels, the assessed mutations could account for or is associated with an increasingly greater proportion of isolates having the quantified resistance MIC level.

Association between MIC levels, characterized mutations and spoligotype strain families
Higher level INH resistance (≥2 μg/mL) was significantly associated with the S315T katG mutation, as shown by a greater odds ratio of 1.97 (Table 3). Of note, in isolates with MIC ≥16 μg/mL (83.0%, n = 38) a mutation was found one or more of the studied genes. We next evaluated for potential the relationship between MIC levels and mutations and strain families. The S315T katG mutation was found in LAM isolates (77.9%, n = 81), Haarlem isolates (94.4%, n = 34), and in T isolates (68.7%, n = 22). Of the Beijing strains (n = 4), 3 presented with the S315T katG mutation. We noted a statistical association between Haarlem strain family with the S315T katG mutation (p = 0.01) ( Table 3). When the specific S315T katG mutation was considered, the Haarlem genotype occurred more frequently among those M. tuberculosis strains with MIC ≥2 μg/mL (p = 0.02). The most frequent Haarlem spoligotype pattern was the shared international type (SIT) 50, which was found in 19 (52.7%) isolates and only one of these did not possess the S315T katG mutation. LAM strain family showed the highest frequency (46.2%, n = 104) among the 224 isolates which were distributed among 20 different SITs according to spolDB4 (Table 4). The LAM9 lineage was the most frequent LAM lineage (29.8%) identified in all three countries studied. In contrast to Haarlem, the LAM strain family was not associated with the S315T katG mutation (p = 0.58) nor with higher MIC values (p = 0.79) ( Table 3). Among T family strains, 6 (18.8%) isolates were related to sub-clades T2, T3, T4 and T5; 22 (68.7%) isolates had the S315T mutation.

Frequency of INH resistance associated mutation in spoligotype strain families
To evaluate for genetic correlation of strains with the same spoligopatterns, DRE-PCR was performed on isolates presenting the same INH conferring mutation and the same spoligotype. DRE-PCR has previously been used to genetically classify strains with the same spoligotyped as being genetically related (or clustered isolates Correlation or MIC levels and percentage of strains bearing the studied mutations in KatG, ahpC and inhA gene loci   (Table 4).
In total, 62 (27.6%) of S315T katG mutated isolates appeared distributed in 29 clusters, most of them with just two isolates per cluster. Of the INH resistant strains that did not have the S315T katG mutation, 19 (27.9%) were in clusters. The proportion of clustering was higher among LAM lineage M. tuberculosis isolates (40.7%; 33/ 81) carrying the S315T katG mutation than in LAM isolates without the S315T katG mutation (26%; 7/23). A higher proportion of clustering in which the S315T katG mutation was also noted for the few W/Beijing strains (50% (2/4). In contrast, the proportion of clustering in S315T katG mutated was lower for Haarlem isolates (23.5%, 8 of 34), T (18%, 4 of 22).

Discussion
Identification of markers for rapid determination of TB drug resistance is needed to combat the increasing prevalence of MDR TB. Mutations in select genes of M. tuberculosis have been used as correlates for anti-TB drug resistance. Prior reports have evaluated in a limited setting one or more of the gene loci evaluated by this report including, katG, ahpC, regulatory region of inhA, and the ORF region of inhA. However, none of these studies have comprehensively catalogued mutations in all of these loci in a single study and testing large numbers of clinical samples from TB prevalent regions such as, South America, nor have they correlated the identified mutations with INH MIC levels.
In this study, each clinical isolate was characterized for mutations not only in katG gene, but also in ahpC, regulatory region of inhA, and ORF region of inhA. Frequencies of katG mutation among INH resistant M. tuberculosis isolates in three South American countries was: Brazil (81.3%), Peru (82.4%) and, Argentina (71.4%). Our study does not aim to provide a profile of the involved sites, but to characterize mutations from the available strains during the period. The frequency for the katG S315T mutation in INH resistant M. tuberculosis isolates was comparable to the previously reported rate for patients diagnosed in Kuwait, Brazil and The Netherlands (65% and 55%, respectively) but was lower than described in Russia (95%) [13,20,22,23].
In this study, we also correlated MIC levels with the katG S315T mutation in INH resistant M. tuberculosis isolates. We demonstrated that 83.0% (n = 127) of the INH resistant strains with the katG S315T mutation possessed a MIC for INH ≥2 μg/mL (p = 0.05). These data are in accordance with The Netherlands report, where 95% of the INH strains with this mutation had a MIC for INH of > 2 μg/L (20). The mutation AGC to ACC at codon 315 tended also to be associated with MIC ≥2 μg/mL (p = 0.06; OR = 1.79 [confidence interval (CI): 0.92-3.49]). Part of the success of the katG S315T mutated isolates in the community is probably because the catalase-peroxidase enzyme is still active in these mutants; indeed, 30% to 40% of the initial catalase activity remains when this mutation is introduced into the katG gene by site-directed mutagenesis [19,24].    Mutations in coding or regulatory regions of other genes such as the oxyR-ahpC region have also been associated with INH resistance, but occur less frequently [1]. Mutations of the oxyR-ahpC region have been described in 4.8% to 24.2% of INH resistant M. tuberculosis isolates [25][26][27]15]. Usually, higher levels of INH resistance and/or loss of catalase activity are associated with mutations in inhA and ahpC genes [28,29]. In the present study, few isolates had mutations in more than one gene. Eight isolates (3.6%) had mutations in both katG and oxyR-ahpC; 5 from Peru and 3 from Brazil (Table 1). Of note, M. tuberculosis isolates with the katG S315T mutation and inhA or ahpC, or inhA and ahpC genes tended to occur more frequently in isolates with a MIC for INH of ≥2 μg/mL, appearing in 22 isolates (p = 0.06; OR 0.95-4.8). After the katG gene, the inhA promoter gene was the second most frequently mutated gene, with mutation in 10% of the M. tuberculosis isolates. This frequency is in accordance to others, varying from 10% to 34.2%, described elsewhere [30,31]. All mutations occurred in the regulatory region of the mabA-inhA operon with a C to T change at position -15, reported to be associated with INH resistance [32,28]. Similarly as has been previously described by others, few mutations were identified in the inhA ORF [4,23].
Frequencies of M. tuberculosis lineage found in our study were in range with frequencies described in recently published population-based studies performed in other South American countries [33,34]. LAM family was the most frequent lineage found by this study, occurring among 46.4% of the INH resistant M. tuberculosis isolates in our South American study population. This proportion is virtually identical to that found among INH resistant M. tuberculosis isolates from Russia [13]. The Haarlem family was the second most frequent family, with a similar proportion of isolates belonging to the Haarlem family as reported in in Russia (10%) [12]. A high frequency of the katG S315T mutation in INH resistant M. tuberculosis isolates of the Haarlem strain family was also described in South Africa [12] and Tunisia [35]. As with the W/Beijing family, the Haarlem family is widespread [36], and has mutations within putative mutator genes [37,38]. Mutation in such genes may afford these strains a higher adaptability to hostile environments, following challenge by anti-TB drugs or engulfment within macrophages [38].
The Haarlem family appears to favor the emergence of MDR-TB strains, and was associated with outbreaks in Argentina [39], the Czech Republic [40] and Tunisia [35]. W/Beijing family strains, which are often associated with drug resistance, although prevalent in many regions of the world, are mostly localized in Asia and Eastern European countries [11,8,41,42], and, at present, uncommon in Latin American countries [33,34,43,44], which was confirmed by this study (only five W/Beijing isolates were identified). The T family occurred in 14.3% of our INH resistant M. tuberculosis isolates, which is similar to the proportion reported in Paraguay (8.6%) and in Venezuela (13%) [22,34]. As a descriptive study on selected M. tuberculosis isolates that were provided by the reference TB laboratories from different regions in Latin America, its limitation rely on the lack of generazibility. The available M. tuberculosis isolates included in the project have no aiming to be a representative from each country on the mutations profiles of INH resistant M. tuberculosis isolates.
The second phase of this study is underway: the evaluation of same techniques using randomly INH sensitive and INH resistant M. tuberculosis isolates isolated at National Drug Resistant Surveillance carried out in those countries in the last years.
Even though the application of DOTS has stabilized the prevalence of TB or has led to decline in some countries, drug-resistant TB is rapidly emerging in a significant number of areas in the world [2]. Under standard treatment regimens it is often not possible to identify primary drug-resistant cases and these regimens are therefore unsuitable for the control of drug-resistant strains. TB control thus relies on improving current TB diagnosis and early detection of drug-resistant TB, preferably using rapid and accurate screening tools other than the sole reliance on AFB smear and culture identification and susceptibility testing.

Conclusion
The present data indicate that screening for the katG S315T mutation may be useful in South America for an early detection of INH resistance and, hence, provide rapid information for selection of appropriate anti-TB therapy. This information may also be used as a marker to evaluate the transmissibility of INH resistant TB in the community. Our study also demonstrated an association between a high MIC and katG S315T mutation, as well as an association between the katG S315T mutation, and Haarlem strain family that may in part explain the successful spread of Haarlem strains in South America.

Methods
The present experimental research that is reported in the manuscript has been performed with the approval of an appropriate ethics committee and carried out within an ethical framework. A 100 μL aliquot of this dilution was added into each well. The microplates were covered, sealed in plastic bags, and incubated at 37°C in the normal atmosphere. After 7 days of incubation, 30 μL of resazurin solution was added to each well, incubated overnight at 37°C, and assessed for color development. Resazurin sodium salt powder (Acros Organic N.V.) prepared at 0.01% (wt/ vol) in distilled water was used as a general indicator of cellular growth and viability. A change from blue to pink indicates reduction of resazurin and therefore bacterial growth. The MIC was defined as the lowest drug concentration that presented no color change. The cut off value for resistance was ≥ 0.2 μg/mL according Palomino et al, 2002 [32]. Growth controls containing no INH and sterility controls without M. tuberculosis were included in each MIC testing.

Nucleic acid extraction
Chromosomal DNA was extracted from cultures on Löwenstein-Jensen medium, using the CTAB method as described by van Embden et al., 1993 [48].

Sequence analysis
The genes were amplified with the following primers (KatG 1.

Spoligotyping
Spoligotyping was performed as described by Kamerbeek et al [49,21]. To determine the spoligotype family, patterns were compared to those in the international database of spoligo patterns (SpolDB4). The double repetitive element (DRE) PCR was performed in accordance to Friedman, 1995 [50]. The term 'cluster' was used for two or more M. tuberculosis isolates with identical spoligotype and DRE-PCR patterns.

Statistical analysis
Data were analyzed using Epi Info (version 6.03, CDC, Atlanta, GA, US; public domain). Categorical variables were compared by the Fisher exact or chi-squared test. A confidence interval (CI) of 95% was used in all odds ratio (OR) calculations.