First Baseline of Circulating Genotypic Lineages of Mycobacterium tuberculosis in Patients from the Brazilian Borders with Argentina and Paraguay

Background At the triple border Brazil/Paraguay/Argentina there is easy mobility from one city to another for economic and tourism activities. This constant and fast population mobility is mainly to visit Iguazu Falls, in the Iguazu River, on the border of the Brazilian state of Paraná and the Argentina. As the incidence of tuberculosis is high in this setting, our study aimed to establish a first baseline of circulating genotypic lineages of Mycobacterium tuberculosis. Methodology/Principal Findings This study included 120 patients from 10 cities in southwestern Paraná, Brazil with pulmonary symptoms, from July 2009 to July 2011. Information about sex, age, clinical features and address was collected by reviewing the national tuberculosis notification database. Of these, 96 (80%) isolates were identified as M. tuberculosis and 22 (22.9%) were drug resistant (20, 20.8% INH mono-resistant and 2, 2.1% multidrug-resistant). All isolates were subjected to genotyping by Spoligotyping and MIRU-VNTR typing. The distribution of the isolates analyzed by spoligotyping revealed 30 distinct patterns. The four mainly detected clades were Latin American and Mediterranean (LAM), ill-defined T, Haarlem (H) and S. The MIRU-VNTR showed 85 distinct patterns. Spoligotyping combined to MIRU-VNTR allowed 90 distinct patterns. Conclusions/Significance Our study demonstrated that there is significant molecular diversity in circulating M. tuberculosis, with predominance of the LAM and T clades in cities of southwestern Paraná, Brazil, bordering Argentina and Paraguay.


Introduction
Tuberculosis (TB) is an infectious disease that remains a major public health problem in the world. Brazil is among 22 countries with major number of cases of TB in the world, with an incidence of 46/100,000 populations in 2012 and the distribution of TB cases varies greatly according to distinct geographic regions [1]. In the state of Paraná (PR), south of Brazil, 2,350 new TB cases (incidence 22.4/100,000) were detected in 2011 [2]. However, these rates are higher in certain regions in the state of Paraná, such as in Paranaguá City, a port area, and on the triple border of Brazil/Paraguay/Argentina. The three municipalities concentrated near the border, which are Foz do Iguaçu in Brazil, Ciudad del Este in Paraguay and Puerto Iguazu in Argentina, have an incidence of 56.6/100,000, 42.5/100,000, and 23.4/100,000, respectively [3]. These high rates are related to the living conditions of populations, lack of access to health services in some regions and high population mobility. Economic activities and tourism to see one of the New Seven Wonders of Nature, Iguazu Falls (Cataratas do Iguaçu), a waterfall on the Iguazu River on the border of the Brazilian state of Paraná and Misiones in Argentina, have provided strong population growth in Brazilian cities near to the border. In this setting, the constant and fast population mobility from one city to another can favors the spread of TB. In the two major municipalities at the Brazilian border, Foz do Igauçu and Cascavel, which have laboratory to perform culture for Mycobacterium tuberculosis and attend many others small surrounding cities for TB diagnosis had an incidence of TB 41.8/100,000 and 25.5/100,000 respectively. Molecular typing methods have been used for a variety of epidemiological investigations and fast tracking geographic distribution of M. tuberculosis clusters [4]. Identical strains fingerprints are called clusters and are usually associated with recent transmission, while strains presenting unique fingerprint profile suggest remote transmission [5,6]. A PCR-based method, Mycobacterial Interspersed Repetitive Unit -Variable Number Tandem Repeat (MIRU-VNTR) typing allows the high throughput and discriminatory analysis of M. tuberculosis clinical isolates [7,8]. The 15-and 24-MIRU-VNTR have been indicated for epidemiological and phylogenetical studies with M. tuberculosis. However, the 12-MIRU-VNTR added of Spoligotyping, another PCR based method, has been widely used in the molecular epidemiology of TB. Spoligotyping investigates the population structure of M. tuberculosis, focusing on the identification of genotypic lineages, spoligotype clades and their geographic distribution [9]. Spoligotyping has proven its usefulness for studying phylogeographical aspects of prevailing M. tuberculosis genotypic lineages and in combination with 12-MIRU-VNTR, it is highly discriminatory allowing the study of both the epidemiological and phylogeographical aspect of tubercle bacilli, including large-scale studies [10,11].
This study aimed to establish a first molecular insight of circulating genotypic lineages of M. tuberculosis in cities in the southwestern Paraná State, south of Brazil, which borders the countries of Paraguay and Argentina countries, based on Spoligotyping and 12-MIRU-VNTR typing.
12 loci-MIRU-VNTR typing conducted in the 96 clinical isolates showed a total of 85 distinct MIRU patterns (File S1).

Discussion
This is the first study conducted to establish the circulating genotypic lineages of M. tuberculosis isolated from TB patients in southwestern Paraná, Brazil that borders the countries of Paraguay and Argentina. Economic activities and tourism in the triple border have provided strong population growth in some cities from the Brazil border, mainly in Cascavel and Foz do Iguaçu. Constant population mobility and the rapid migration for commercial purpose, from one city to another one, at the triple border, may favor the spread of TB in this setting.
The age and sex distribution of patients in our study reflects the predominance of TB among adult male people (81.4% of all patients) as observed in all other regions in the word [1]. The 2:1 male to female sex-ratio may be explained by the local sociocultural barriers. In this setting, women stay mainly at home and men are engaged in external activity, which leads the latter to a higher exposition to the bacillus. In the present study, the prevalent isolates came from Cascavel, Foz de Iguaçu and other surrounding small cities.
The LAM, T and H were the largest clades observed in the present study, the three genotypic clades most frequent in Africa, Central America, Europe and South America [10]. The present results are similar to a previous one, carried out in our laboratory, by Noguti et al. [12], which had greater variability in SITs, but LAM, H and T were the largest clades in the northwestern Paraná. In Araraquara, São Paulo State, which borders northern Paraná, Mendes et al. [13] identified 73% of spoligopatterns in SpolDB4, and LAM, T and H clades were the most prevalent too. More recently, Santos et al. [11], in a comparative study between 1998-2001 and 2002-2006, observed replacement of ill-defined T family for LAM in Araraquara, São Paulo. Furthermore, there are a number of lineages found in Northern Paraná and São Paulo States studies, including Beijing, which was not detected in the present study.
The most prevalent SITs in southwestern Paraná were 60 (prototype of LAM4), 42 (prototype of LAM9), 53 (prototype of T1) and 177 (prototype of LAM9). These four most prevalent SITs observed in our study were found in Argentina too, as reported in SITVITWEB and SpolDB4, but no reference to Paraguay was made [10].
The SITs 334 and 713 (both susceptible and prototype of T1), characterized by one isolate for each one, have not been described      in Paraguay and Argentina yet. The first one has been described in Brazil only once, but the second hasn't been yet (http://www. pasteur-guadeloup.fr:8081/SITVITDemo/). The other spoligopatterns, which were described in SITVITWEB database, were already identified in Brazil and some of them in Argentina and Paraguay. The SITs 52 (prototype of T2), 60 (prototype of LAM4), 196 (prototype of T1), 822 (prototype of LAM9) and 866 (prototype of LAM9) were described for the first time in Paraná, Brazil (http://www.pasteur-guadeloup.fr:8081/SITVITDemo/). Of interest, INH mono-resistant was predominant in isolates genotyped as SIT60 (LAM4, 9 isolates), which is frequent in the Americas and some countries in Europe (http://www.pasteurguadeloup.fr:8081/SITVITDemo/). The two MDR were genotyped as SIT 60 (LAM4) but had different MIRU patterns. The spoligoforest tree indicates that there are conserved genotypes as the SIT53 (prototype of T1), which demonstrated the ability to generate 7 other genotypes, by the loss of 1 to 6 spacers, which predominate T clade. We could observe with SIT42 (prototype of LAM9) that the loss of only one spacer, in a different section of DR region, yielded 6 different SITs (20,60,64,177,822,866) belonging to the LAM clade (figure 1).
We observed a total of 9 single locus variation (SLV) events within the 12-locus subset. The best discriminatory power was obtained considering that a unique SLV in any one of the loci analyzed, remains strongly predictive of the absence of a link between the isolates. According to Hunter and Gaston [22] the most allelic diversity observed in the M. tuberculosis isolates studied was in MIRU loci 40 and 26 and moderated polymorphism were found in MIRU loci 2, 10, 16, 20, 23, 27 and 31. These allelic diversities also differs from the obtained in northwestern Paraná by Noguti et al. [12] that observed most allelic diversity in MIRU loci 40, 23, 10 and 16 and moderate polymorphism in MIRU loci 26, 20 and 31. Sharma et al. [14] working with only 6 loci-MIRU (MIRU 4, 10, 16, 26, 39 and 40) observed the most discriminatory loci, in order of diversity, 26, 10, 16 and 40 in isolates from Kanpur, India. Kovalev et al. [15] observed most allelic diversity in MIRU loci 26, 31 and 10 in M. tuberculosis isolated in Ural region, Russian Federation. On the other hand these authors refer to MIRU locus 23 as having the lowest discriminatory power, differing from our results, which was MIRU 39.
Despite of high discriminatory power of MIRU typing observed in our study, there were isolates clustered by this method. In these cases, some clustered isolates (25, 109, 29, 30, 1, 17, 11, 13 and 22) were discriminated by spoligotyping, demonstrating the need of using these two techniques combined to provide a higher discriminatory power. In this sense, Spoligotyping combined to 12 loci-MIRU-VNTR allowed 90 different molecular patterns for all 96 M. tuberculosis isolates analyzed. In these patterns there were 12 isolates that weren't possible to be differentiated as unique and were included in six patterns with two isolates each one. Four these clustered isolates are of particular concern to be discussed. The isolates 43 and 77, which were INH mono-resistant, were Table 3. Clusters by Spoligotyping/MIRU combination and epidemiological data of 12 Mycobacterium tuberculosis isolates from patients in cities in southwestern Paraná, Brazil. obtained from patients, who lived in the same neighborhood and nearby streets and were referred to the central laboratory in Cascavel for TB laboratorial diagnosis in the same year. Two other isolates (64 and 66) with the same spoligopattern and MIRU, were obtained from patients that lived in neighboring cities (Cascavel and Campo Bonito) and were referred to the central laboratory in Cascavel for TB laboratorial diagnosis. Thus, to overcome the observed limitation, additional study should be considered by using optimized 15-and 24-loci-MIRU typing to differentiate the isolates and provide additional data about any isolates relationship.
In conclusion, our study offers the first baseline about the genetic diversity of M. tuberculosis isolated from patients with TB in cities in southwestern Paraná State, Brazil, which borders two other countries, Paraguay and Argentina. This kind of study is important for tracing relationships among the strains and recognition of the mainly clades responsible for spreading the disease. One limitation of the present study is the small number of isolates from some cities in the samples. However this issue will be addressed by additional and better designed studies in this setting, with a higher timescale and number of isolates conducted in the future to understand the chain of TB transmission in this region.

Patients, bacterial isolates and drug susceptibility test
The study was approved by the Ethics Committee of the State University of Maringa, Parana (protocol No. 018/2011, CAAE No. 0375.0.093.000-10). The information about sex, age, clinical features and address of patients was collected by reviewing the national TB notification database. As the study was conducted with M. tuberculosis isolated from sputa that were disposed of, in accordance with internal biosecurity rules, after test results release to the patient, and no additional patients' sputa samples and information, which could to identify them were requested (identification was carried out by a number given to each patient), the document participant consents achievement were exempted by the ethics committee. The study included 120 patients (66.7% male patients, age range from 17 to 82 years old) from 10 cities in southwestern Paraná, Brazil, which borders Paraguay and Argentina countries (Figure 2), with confirmed pulmonary TB by microscopy (Ziehl-Neelsen -ZN) and culture (Ogawa medium -OK). All patients were attended at two central laboratories in Cascavel and Foz do Iguaçu cities, southwestern Paraná, from July 2009 to July 2011. Of these, 96 isolates (80%) were identified as M. tuberculosis by biochemical methods [16] and drug susceptibility tests (DSTs) were carried out for Isoniazid (INH), Rifampicin (RMP) and Pyrazinamide (PZA) by Löwenstein-Jensen (L-J) Proportion Method [17].
BioNumerics software (version 4.45; Applied Maths, Sint-Martens-Latem, Belgium) was used for analysis of Spoligotyping and MIRU-VNTR patterns. Dendrogram was constructed for Spoligotyping and MIRU-VNTR combined. The genetic distance was built employing the UPGMA algorithm (Unweighted Pair Group Method with Arithmetic Mean).

Supporting Information
File S1 Sex and age of patients and drug susceptibility profile, Spoligotyping and MIRU typing of 96 Mycobacterium tuberculosis clinical isolates from cities in southwestern Paraná , Brazilian borders with Argentina and Paraguay.

Author Contributions
Conceived and designed the experiments: RFC LNCM CQFL FRP NRM.