Molecular epidemiology of multidrug-resistant strains of Mycobacterium tuberculosis

REVIEW 10.1111/j.1469-0691.2011.03577.x Molecular epidemiology of multidrug-resistant strains of Mycobacterium tuberculosis W. Sougakoff National Reference Centre for Mycobacteria (CNR-MyRMA), Laboratoire de Bactériologie-Hygiène, UPMC, ER5, Faculté de Médecine Pitié-Salpêtrière, Paris, France Abstract Genotyping of Mycobacterium tuberculosis has been extensively used for investigating epidemics of multidrug-resistant strains of M. tuberculosis, in order to identify the factors involved in the transmission of such strains and determine effective control programmes to limit their expansion at both the individual and population levels. Here, we review the methods currently used to study the molecular epidemiology of multidrug-resistant M. tuberculosis strains, and the insights provided by these techniques regarding global trends and the transmission dynamics of multidrug-resistant tuberculosis at a world scale. Keywords: Molecular epidemiology, multidrug-resistant, Mycobacterium, resistance, tuberculosis Article published online: 7 May 2011 Clin Microbiol Infect 2011; 17: 800 805 Corresponding author: W. Sougakoff, Laboratoire de Bactériologie-Hygiène, UPMC ER5, Faculté de Médecine Pitié-Salpêtrière, 91 boulevard de l Hôpital, 75634 Paris Cedex 13, France E-mail: wladimir.sougakoff@upmc.fr Introduction Multidrug resistance, i.e. resistance to at least rifampin and isoniazid, in Mycobacterium tuberculosis is caused by the sequential accumulation of mutations in the genes encoding the targets of rifampin and isoniazid. Extensively drug-resistant (XDR) tuberculosis (TB) is defined as multidrug resistance (resistance to isoniazid and rifampin), plus resistance to a fluoroquinolone and any of the injectable second-line drugs (amikacin, capreomycin, or kanamycin). The emergence and spread of multidrug-resistant (MDR) and XDR TB is hampering efforts to control and manage TB. Since the emergence of MDR strains in the 1990s, the prevalence of MDR TB has slowly but constantly increased, such that it now accounts for approximately 5% of global TB cases. According to the Multidrug and extensively drug-resistant TB 2010 Global report on surveillance and response, it can be estimated that 440 000 cases of MDR TB emerged globally in 2008, causing an estimated 150 000 deaths [1]. More worrisome is the emergence of XDR TB, a nearly untreatable form of the disease with an estimated rate of 15% among the MDR strains [2]. In this review, the insights provided by molecular epidemiology with regard to global trends of transmission of MDR M. tuberculosis strains are presented and discussed in light of the actual expansion of MDR/XDR TB. Current Methods used for the Molecular Typing of MDR/XDR Strains of M. tuberculosis Restriction fragment length polymorphism (RFLP) analysis of IS6110 insertion sequence relies on the determination of the number of copies and location of insertion on the chromosome of the IS6110 element [3]. This technique is generally recognized as a reference standard for strain differentiation in molecular epidemiological studies, and has been widely used for the identification and investigation of outbreaks, the identification of laboratory cross-contamination, and distinction between re-infection and relapse [3]. However, the method has several disadvantages that seriously limit its use in routine practice: it has poor discriminatory power when applied to isolates with a low IS6110 copy number, it is labour-intensive, it requires significant amounts of high-qual- Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases

CMI Sougakoff Molecular epidemiology of MDR M. tuberculosis 801 TABLE 1. Major phylogenetic lineages within the Mycobacterium tuberculosis complex (MTBC) strains adapted to humans MTBC species LSP-based lineage nomenclature Spoligotype-based families M. tuberculosis Indo-ic EAI M. tuberculosis East Asian Beijing M. tuberculosis East African Indian CAS M. tuberculosis Euro-American H, LAM, X, T, S, others M. africanum West African I AFRI2 M. africanum West African II AFRI1 AFRI, M. africanum; CAS, Central Asian; EAI, East African Indian; H, Haarlem; LAM, Latin American Mediterranean; LSP, large sequence polymorphism. ity DNA, and it generates band patterns that are difficult to share between laboratories. Spoligotyping, which is based on polymorphisms in the clustered regularly interspaced short palindromic repeats (CRISPRs), can be used for both epidemiology and evolutionary analysis of M. tuberculosis [4]. Technically, the method is based on the amplification and detection of the presence or absence of non-repetitive sequences called spacers found between direct repeat elements in the CRISPR region [5]. Because it is based on PCR amplification, spoligotyping can be performed with very small quantities of DNA, so the method does not require the handling of large amounts of resistant isolates. Over the years, the signatures given by the 43 spacer-spoligotyping patterns have been used to define strain families and create a very useful nomenclature with which to describe the circulating genotypes of tubercle bacilli worldwide (Table 1) [6]. As compared with the other molecular methods, the main limitation of spoligotyping is its inferior discriminatory power, especially in some lineages such as the Beijing strains, which nearly all share a single spoligotype. Variable numbers of tandem repeats (VNTR) analysis is a powerful method that can provide adequate discrimination of M. tuberculosis strains, in terms of both identification of genetic lineages and estimation of M. tuberculosis transmission [7]. The method, which is PCR-based, relies on the analysis of tandem repeats present in 41 genetic elements (called mycobacterial interspersed repetitive units (MIRUs)) scattered throughout the M. tuberculosis genome, with a variable number of copies of the repeat unit in each locus. MIRU- VNTR analysis, which is a PCR-based approach, is relatively safe when applied to MDR strains, and generates numerical values that can easily be compared in interlaboratory studies. This method has other major advantages, such as its use for phylogenetic analysis supported by a dedicated website (MIRU-VNTRplus, available at http://www.miru-vntrplus.org/), and a discriminatory power that is generally regarded as being similar to that of IS6110-RFLP. Owing to its acceptable discriminatory power and the exchangeable format of the data generated by the method, MIRU-VNTR typing is considered to be the new reference standard for molecular epidemiological studies. Phylogeographical Analysis of MDR/XDR M. tuberculosis Strains An overall picture of the epidemiology of MDR TB in the world can be drawn from the multiple studies that have analysed, by applying genetic approaches such as IS6110-RFLP, spoligotyping, and MIRU-VNTR, the specific situations encountered in representative areas of the world. Such studies are summarized hereafter, with the goal of presenting the landscape of MDR/XDR TB diversity in the world. In China, the rate of MDR TB (8% in 2007) is significantly higher than the global average rate, and represents a major public health challenge [8]. In this country, and more generally on the Asian continent, the Beijing family is very common, with a prevalence that exceeds 50% in many countries (China, South Korea, Hong Kong, and Thailand) [9]. A very recent study based on the use of spoligotyping and MIRU- VNTR determined the transmission characteristics of M. tuberculosis isolates in a province of China where the Beijing family prevails [8]. Strikingly, 90% of the strains were found to have the Beijing family genotype, and, among these, 27% were MDR. The clustering rate among the MDR strains was rather low (10%), indicating that these strains are characterized by a significant level of genomic diversity. In Japan, the frequencies of MDR TB and XDR TB are approximately 1.9% and 0.5%, respectively [10]. The transmission dynamics of MDR TB and XDR TB were analysed by IS6110-RFLP, spoligotyping and MIRU-VNTR on a set of 38 MDR and 17 XDR isolates [10]. The genetic family that dominated was found to be Beijing (62%), followed by T (13%), Latin American Mediterranean (LAM) (5%), U (2%), East African Indian (EAI) (2%) and X (2%). Interestingly, the proportion of Beijing genotype did not differ significantly between the MDR non-xdr and the XDR strains, but 70% of the XDR strains were involved in clusters, indicating that the XDR TB cases in Japan are more likely to be associated with clustering than the non-xdr TB cases (38% of clustered strains among the non-xdr cases) [10]. One possible explanation that could account for the high clustering rate of XDR TB reported in this study is that new cases of XDR TB are more likely to be caused by transmission than by acquisition of multidrug resistance resulting from treatment failure.

802 Clinical Microbiology and Infection, Volume 17 Number 6, June 2011 CMI A very similar study was conducted in Taiwan by investigating, with spoligotyping and MIRU-VNTR, the transmission and predominant genotypes involved in MDR TB [11]. In this region, where the prevalence of MDR TB was 4% in 2004 2008, the Beijing family genotype was still found to be the predominant genotype among the MDR strains (48%), followed by the Haarlem (15%) and T1 (4%) genotypes. Approximately 62% of the strains were clustered into ten clusters, and the largest cluster belonged to the Beijing genotype, a point that confirms the significant role played by this lineage in the extensive transmission of MDR TB in Taiwan. India has one of the highest estimated rates of TB in the world, and accounts for 21% of all TB cases [12]. In this country, it is estimated that 1 3.4% of new patients are infected by an MDR M. tuberculosis strain [13]. An analysis of MDR isolates by spoligotyping indicated that the most prevalent strain family is represented by the Beijing lineage (41%), followed by the EAI (27%), Manu1 (20%) and Central Asian (CAS) (17%) lineages [14]. Thus, despite the fact that the proportion of Beijing strains remains quite low in India (predominant spoligotypes belong to the Central Asian Delhi family), it seems that there is an association between the Beijing lineage and MDR TB in this country [14,15]. The epidemiological situation that prevails in eastern Europe and in the Russian Federation and the former Soviet Union countries is exemplified by the paper published by Niemann et al. [16], who analysed the population structure of M. tuberculosis complex strains from the Republic of Georgia by 24-locus MIRU-VNTR and spoligotyping. In this country, the frequency of MDR strains can reach levels of up to 40% among previously treated patients. Strikingly, MDR TB seems to be mostly associated with the diffusion of the Beijing lineage (77% of the MDR isolates), thus confirming that Beijing strains represent a major factor driving the MDR TB epidemic in eastern Europe [16]. Regarding the Russian Federation, a recent study has applied molecular typing to a panel of 76 MDR M. tuberculosis strains circulating in Moscow; it showed that 48% of the strains were grouped into 12 clusters, ten of them including strains belonging to the Beijing family, which represented 78% of the MDR strains [17]. Interestingly, the level of clustered strains was especially high (40%) for Beijing strains in the group of patients who had not been treated previously, an observation that was interpreted as evidence for the active transmission of such strains [17]. It is interesting to note that, in spite of the geographical proximity of countries where the Beijing sublineages are highly prevalent, the phylogenetic distribution observed in Sweden is very different from that reported for Russia and the Baltic countries. According to Ghebremichael et al. [18], only 13% of the patients with MDR TB were found to have Beijing strains, and the majority of these patients came from Asia (75%). Thus, in spite of the geographical proximity of Russia and the Baltic states, there is no extensive spreading of Beijing strains from these regions within any of the Scandinavian countries (in Finland and Denmark, the incidence of Beijing strains remains very low) [18]. A study published in 2010 presented a spoligotype-based analysis of MDR M. tuberculosis isolates in circulation in Poland, which is representative of the population structure characteristic of a European country [19]. Globally, the MDR strains were characterized by high genotypic diversity and a high number of unique profiles. Nevertheless, four major lineages, i.e. the T, Haarlem, LAM and Beijing families, constituted more than 67% of the strains studied, showing the compactness at the phylogenetic level of the MDR M. tuberculosis population structure in Poland. The highest number of MDR strains belonged to the ill-defined T family (28%), followed by the Haarlem (17%) and LAM (13%) families, the Beijing genotype accounting only for 9% of the MDR strains [19]. The three major genotype families found in this study (Haarlem, LAM, and T) are dominant in many European countries, where they represent up to 60% of M. tuberculosis strains found in Italy, 70% of those found in Sweden, and 80% of those found in Portugal [19]. In Poland, the low proportion of Beijing MDR strains (9%) contrasts with the high proportion of TB cases attributable to this lineage in the countries of the former Soviet Union (approximately 30 50%). Nonetheless, the presence of the Beijing lineage exclusively in the MDR group suggests that the ongoing transmission of Beijing strains is a significant cause of MDR TB in Poland. Interestingly, most of these strains were isolated from non-migrant patients born in Poland, which suggests that active transmission of MDR Beijing strains within the autochthonous population occurs in parallel with the more commonly observed transmission from immigrants from Asia and eastern Europe. When compared to that in Poland, the population structure of the MDR isolates circulating in Portugal presents interesting characteristics. This country has the fourth highest TB incidence rate in the European Union, with high levels of MDR TB (primary rate, 1.5%; retreatment cases, 2.4%) [20]. A study based on MIRU-VNTR typing indicates that a large proportion (55%) of the MDR strains from the Lisbon area belong to a single family related to the LAM genotype (referred to as the Lisboa family in the Perdigao s report), which caused a large MDR TB outbreak in the Lisbon region at the beginning of the 1990s. More alarming is the fact that 53% of these strains were XDR TB isolates, which repre-

CMI Sougakoff Molecular epidemiology of MDR M. tuberculosis 803 sents the highest XDR TB rate ever reported [21]. Thus, the development of XDR TB observed in the Lisbon region probably results from the circulation in the community of Lisboa MDR strains for more than 10 years, which has permitted the progressive acquisition of resistance to secondline drugs [21]. Finally, the population structure of the MDR isolates circulating in France has been evaluated by applying spoligotyping and 24-locus MIRU-VNTR to a panel of 98 MDR strains collected by the French National Reference Centre (CNR MyRMA) during the period 2009 2010 (Sougakoff W, Veziris N, Millot G, Bastian S, Jarlier V, unpublished data). The overall distribution of the lineages on a map according to the patients country of birth confirms the predominance of the Beijing family among the MDR strains obtained from patients born in central Europe, eastern Europe, and Asia (approximately 75% of the MDR strains in this group of patients) (Fig. 1). Regarding the group of patients born in Africa, one can note a greater diversity of lineages, with a majority of strains being assigned to the ill-defined T lineage (52%), and the remainder to the T1-Ghana, Haarlem, LAM, Cameroon, TUR and Beijing families (Fig. 1). The MDR strains obtained from the patients born in France (10% of the MDR TB cases) were also found to belong to different lineages (Beijing, Ghana, T, LAM, and S) (Fig. 1). The minimum spanning tree shown in Fig. 2 reveals the high genetic heterogeneity and the complexity of the population structure for the MDR strains, with 13 clusters of strains sharing identical MIRU codes and representing 41% of the tested isolates. Six of the 13 clusters were attached to the Beijing genotype family within a single clonal complex encompassing 19% of the MDR strains (clonal complex 1 in Fig. 2). It is noticeable that no apparent epidemiological link was found for a majority (67%) of the MDR strains found in the six Beijing clusters, emphasizing the diversity of the Beijing strains circulating in France. Conclusions and Projections for the Future Molecular fingerprinting methods have undoubtedly had a very significant impact on ascertainment of the phylogeny of the bacilli that account for the worldwide diffusion of MDR TB and XDR TB, and elucidatation of the route of transmission of MDR TB. The report published in 2010 by Brown et al. [22] clearly confirms the tendencies highlighted by the various reports presented in this review. First, there is a strong association between a patient s country of origin and the family of the MDR isolates: Beijing dominates in patients Greenland Canada Suomi Finland Iceland Sverige Sweden Nor Nor United Kingdom Poiska Poland and ykpaïha aine Kazakhstan POCCRN Russia Monga ia South Pacific United States México Venezuela Colombia Brasil Brazil Bolivia Chile Argentina North Atlantic Turkey Turkey Iraq Iran Afghanistan Algeria Pakistan Libya Egypt Saudi Arabia India Mauritania Mali Niger Chad Sudan Nigeria Eth opia South Atlantic Kenya Congo Tanzania gola Namibia Botswana South Africa Madagascar Indian China Thailand Indonesia S korea Japan Australia Papua New Guinea FIG. 1. Geographical distribution of the lineages of multidrug-resistant strains (n = 98) received in 2009 2010 by the French National Reference Centre for mycobacteria (CNR MyRMA), according to the country of birth of the patients. The diameters of coloured circles are proportional to the numbers of strains belonging to a given lineage. The colour indicates the mycobacterial interspersed repetitive unit (MIRU)-24-defined lineage to which each strain is related: Beijing is red, T is dark blue, T1-Ghana is light green, Haarlem is green, LAM is dark green, Cameroon is yellow orange, TUR is blue, S is light blue, Dehli/CAS is green khaki, and undetermined types are pink and violet.

804 Clinical Microbiology and Infection, Volume 17 Number 6, June 2011 CMI studies is that the Beijing family could be strongly associated with multidrug resistance [22], and has spread globally in recent years, now representing a major factor driving the MDR TB epidemic in the world [16]. In this context, molecular epidemiology of MDR M. tuberculosis should progress to improve our capacity to detect chains of transmission and enhance TB control activities. Transparency Declaration The author does not have any conflict of interest to declare. References FIG. 2. Minimum spanning tree based on the diversity of the 24- locus mycobacterial interspersed repetitive unit (MIRU)-variable numbers of tandem repeats (VNTR) results for a representative sample of multidrug-resistant strains received at the French National Reference Centre (CNR MyRMA) (2009 2010). The main clonal complexes (CCs) are indicated (a CC consists of strains that differ from each other by a single locus change): CC1, Beijing; CC2, probably Haarlem; CC3, T1-Ghana; CC4, T2 T3; CC5, S. Clusters are indicated by filled circles. The tree was drawn by using the MIRU- VNTRplus website. born in East Asia, whereas CAS, EAI and Beijing dominate in patients born on the Indian subcontinent [22]. In contrast, the strains grouped in the Euro-American superfamily (encompassing the LAM, Haarlem and T lineages) are infrequent in these regions, but are globally equally distributed in all other regions of the world. The other important and worrying conclusion that can be drawn from genotyping 1. Prasad R. Multidrug and extensively drug-resistant TB (M/XDR-TB): problems and solutions. Indian J Tuberc 2010; 57: 180 191. 2. Shah NS, Wright A, Bai GH et al. Worldwide emergence of extensively drug-resistant tuberculosis. Emerg Infect Dis 2007; 13: 380 387. 3. McEvoy CR, Falmer AA, Gey van Pittius NC, Victor TC, van Helden PD, Warren RM. The role of IS6110 in the evolution of Mycobacterium tuberculosis. Tuberculosis (Edinb) 2007; 87: 393 404. 4. van Embden JD, van Gorkom T, Kremer K, Jansen R, van Der Zeijst BA, Schouls LM. Genetic variation and evolutionary origin of the direct repeat locus of Mycobacterium tuberculosis complex bacteria. J Bacteriol 2000; 182: 2393 2401. 5. Kato-Maeda M, Metcalfe JZ, Flores L. Genotyping of Mycobacterium tuberculosis: application in epidemiologic studies. Future Microbiol 2011; 6: 203 216. 6. Brudey K, Driscoll JR, Rigouts L et al. Mycobacterium tuberculosis complex genetic diversity: mining the fourth international spoligotyping database (SpolDB4) for classification, population genetics and epidemiology. BMC Microbiol 2006; 6: 23. 7. Supply P, Mazars E, Lesjean S, Vincent V, Gicquel B, Locht C. Variable human minisatellite-like regions in the Mycobacterium tuberculosis genome. Mol Microbiol 2000; 36: 762 771. 8. Wang J, Liu Y, Zhang CL et al. Genotypes and characteristics of clustering and drug susceptibility of Mycobacterium tuberculosis isolates collected in Heilongjiang Province, China. J Clin Microbiol 2011; 49: 1354 1362. 9. Glynn JR, Whiteley J, Bifani PJ, Kremer K, van Soolingen D. Worldwide occurrence of Beijing/W strains of Mycobacterium tuberculosis: a systematic review. Emerg Infect Dis 2002; 8: 843 849. 10. Murase Y, Maeda S, Yamada H et al. Clonal expansion of multidrugresistant and extensively drug-resistant tuberculosis, Japan. Emerg Infect Dis 2010; 16: 948 954. 11. Hsu AH, Lin CB, Lee YS et al. Molecular epidemiology of multidrugresistant Mycobacterium tuberculosis in Eastern Taiwan. Int J Tuberc Lung Dis 2010; 14: 924 926. 12. Stavrum R, Myneedu VP, Arora VK, Ahmed N, Grewal HM. In-depth molecular characterization of Mycobacterium tuberculosis from New Delhi predominance of drug resistant isolates of the modern (TbD1) type. PLoS ONE 2009; 4: e4540. 13. Kant S, Maurya AK, Kushwaha RA, Nag VL, Prasad R. Multi-drug resistant tuberculosis: an iatrogenic problem. Biosci Trends 2010; 4: 48 55. 14. Chatterjee A, D Souza D, Vira T et al. Strains of Mycobacterium tuberculosis from western Maharashtra, India, exhibit a high degree of diversity and strain-specific associations with drug resistance,

CMI Sougakoff Molecular epidemiology of MDR M. tuberculosis 805 cavitary disease, and treatment failure. J Clin Microbiol 2010; 48: 3593 3599. 15. Purwar S, Chaudhari S, Katoch VM et al. Determination of drug susceptibility patterns and genotypes of Mycobacterium tuberculosis isolates from Kanpur district, North India. Infect Genet Evol 2011; 11: 469 475. 16. Niemann S, Diel R, Khechinashvili G, Gegia M, Mdivani N, Tang YW. Mycobacterium tuberculosis Beijing lineage favors the spread of multidrug-resistant tuberculosis in the Republic of Georgia. J Clin Microbiol 2010; 48: 3544 3550. 17. Afanas ev MV, Ikryannikova LN, Il ina EN et al. Molecular typing of Mycobacterium tuberculosis circulated in Moscow, Russian Federation. Eur J Clin Microbiol Infect Dis 2011; 30: 181 191. 18. Ghebremichael S, Groenheit R, Pennhag A et al. Drug resistant Mycobacterium tuberculosis of the Beijing genotype does not spread in Sweden. PLoS ONE 2010; 5: e10893. 19. Jagielski T, Augustynowicz-Kopec E, Zozio T, Rastogi N, Zwolska Z. Spoligotype-based comparative population structure analysis of multidrug-resistant and isoniazid-monoresistant Mycobacterium tuberculosis complex clinical isolates in Poland. J Clin Microbiol 2010; 48: 3899 3909. 20. Perdigao J, Macedo R, Joao I, Fernandes E, Brum L, Portugal I. Multidrug-resistant tuberculosis in Lisbon, Portugal: a molecular epidemiological perspective. Microb Drug Resist 2008; 14: 133 143. 21. Perdigao J, Macedo R, Malaquias A, Ferreira A, Brum L, Portugal I. Genetic analysis of extensively drug-resistant Mycobacterium tuberculosis strains in Lisbon, Portugal. J Antimicrob Chemother 2010; 65: 224 227. 22. Brown T, Nikolayevskyy V, Velji P, Drobniewski F. Associations between Mycobacterium tuberculosis strains and phenotypes. Emerg Infect Dis 2010; 16: 272 280.