Spoligotype-Based Comparative Population Structure Analysis of Multidrug-Resistant and Isoniazid-Monoresistant Mycobacterium tuberculosis Complex Clinical Isolates in Poland

MATERIALS AND METHODS

Patients and specimens.This study included a total of 46 MDR strains of M. tuberculosis isolated from as many TB patients in Poland. The patients (40 men and 6 women; age range, 31 to 79 years; median age, 50.5 years) were recruited in several TB clinics across Poland during a 1-year period from 1 January 2004 to 31 December 2004 (Fig. 1). Primary isolation, species identification, and drug susceptibility testing (DST) were done at regional mycobacteriology laboratories. The isolates were subcultured and sent to the National TB Reference Laboratory (NTRL) at the National Tuberculosis and Lung Diseases Research Institute in Warsaw, where confirmatory identification and DST were performed. Testing of susceptibility to the first-line drugs was carried out according to the 1% proportion method in Löwenstein-Jensen medium, with the following critical concentrations: streptomycin (SM), 4 μg/ml; INH, 0.2 μg/ml; RMP, 40 μg/ml; and ethambutol (EMB), 2 μg/ml. Basic demographic data were collected for each patient, using a standard questionnaire prepared by the NTRL, and sent to the local TB laboratories.

• Open in new tab
• Download powerpoint

FIG. 1.

Map of Poland depicting the distribution of MDR/INH-monoresistant M. tuberculosis isolates within the 16 provinces of the country. The shaded provinces are the ones from which at least one MDR-TB strain was obtained (administrative map of Poland drawn by the authors; note that boundaries in the map do not attempt to portray exact frontiers between administrative regions in Poland or between Poland and its neighbors).

RESULTS AND DISCUSSION

This study included all bacteriologically confirmed MDR-TB cases in Poland over a 1-year period (2004; n = 46 isolates from as many TB patients). These MDR-TB cases represented 18.7% of all drug-resistant-TB cases reported in Poland throughout 2004 (smear-positive TB cases registered [total], n = 9,493; any drug resistance, n = 246; INH-monoresistant isolates, n = 82; MDR-TB isolates, n = 46). The results obtained by spoligotyping of the MDR-TB isolates compared to the INH-monoresistant isolates (n = 71) are summarized in Tables 1 and 2. Whenever available, a phylogenetic clade description for each SIT was also provided, as well as their geographical distribution among neighboring countries, limited to Germany (n = 704), Czech Republic (n = 530), Sweden (n = 379), and Kaliningrad, Russia (n = 90), as well as Poland (strains from previous studies; n = 317). For the 46 clinical isolates, spoligotyping produced a total of 27 different patterns (genetic diversity, 58.7%), 4 of which corresponded to orphan patterns (not found among more than 70,000 strains recorded in the SITVIT2 database). The remaining 42 (91.3%) strains from this study belonged to 23 SITs, including the 3 newly created SITs. The latter were new SITs created after either a match within this study, as in the case of SIT2538, comprising 2 MDR isolates, or a match with another orphan strain recorded in SpolDB4 (SIT3093 matched a strain from Italy, whereas SIT3094 matched a strain from previous studies in Poland). Regarding clustering, a total of 26 (56.5%) clinical isolates were clustered (7 clusters containing 2 to 6 isolates per cluster), and 20 (43.5%) isolates were unique within this study (including the 4 orphan strains).

The clade assignment demonstrated that the highest number of MDR strains belonged to the ill-defined T family (13 strains, or 28.3% of all strains). The most prevalent shared spoligotype within this family was SIT53, representing the T1 sublineage (6 strains, or 13% of all strains and 46% of the T-clade strains). The second-largest group, comprising 8 (17.4%) of the analyzed strains, displayed spoligotypes related to those of the Haarlem family, most of them being representative of the H1 sublineage (SIT47, SIT1557, and SIT2333). The third-largest group comprised members of the LAM family (6 strains, or 13% of all the strains). All these strains displayed spoligotypes that belonged to the LAM9 subclade (SIT42 and SIT891). Four (8.7%) of the MDR isolates harbored the spoligotyping pattern (SIT1) characteristic of the Beijing family of strains. Two minor clades, namely, X and S, were represented by one and two isolates, respectively. An important proportion of the strains (12 strains, or 26.1% of all the strains) could not be assigned to any of the major clades described within the spoligotype database. A comparison with INH-monoresistant strains (n = 71) (Table 2) showed a lower clustering rate (but not significantly lower; 56.5% versus 70.4%) (Table 3) for the MDR strains, considered a proxy for recent transmission. The results also underlined a higher genotypic diversity (58.7% versus 42.2%) and a higher number of unique profiles (43.5% versus 29.6%) of the MDR isolates compared to the INH-monoresistant isolates (4).

Although spoligotyping per se is not an appropriate method for assessing the burden of recently transmitted TB (it may overestimate the proportion of clustered cases [22-24]), it serves as an efficient preliminary procedure in determining the genetic relatedness of M. tuberculosis isolates (8, 9, 15, 26, 30, 31, 34). Furthermore, the interpretation of the clustering results in the context of recent transmission is fraught with several ambiguities. This is because the amount of transmission represented by molecular clustering depends on a number of methodological factors, such as the methodology employed, as mentioned above; sample size; or study duration. For the last two, the larger the number of TB cases included in the study, the higher the proportion of clustered or recently transmitted cases. Likewise, the observation time is positively associated with clustering, and it is considered that studies based on less than 2 years of sampling underestimate recent transmission (20, 35). Given the small pool of analyzed strains and fact that the study period lasted only 1 year, the rate of clustering for MDR-TB and INH-monoresistant isolates (56.5% and 70.4%, respectively; the difference was not statistically significant; P = 0.1252) in our study seems to be high. Nevertheless, there are conflicting results in the literature on the relative transmissibility of drug-resistant and drug-sensitive M. tuberculosis strains. While some studies showed that drug-resistant isolates were less likely to be associated with clustering (46, 48), others identified drug resistance as a strong predictor of being in a cluster (1, 10).

In spite of a high degree of genetic diversity, an important part (41.3%), of the analyzed MDR strains belonged to one of the four clusters SIT53, SIT891, SIT1, and SIT1557. The compactness of the MDR M. tuberculosis population structure was also apparent at the phylogenetic level, as four major lineages, i.e., the T, Haarlem, LAM, and Beijing families, compristed more than 67% of the isolates studied (Table 1). With 30% of the isolates in the SpolDB4 database (9), the T family contains poorly informative, ubiquitous spoligotypes with obscure genealogies. It is supposed, however, that most of the spoligotypes within the T clade represent relatively old genotypes prevalent in Europe (9). The Haarlem family, of European origin, comprises nearly 1/4 of the M. tuberculosis population in Europe, although it also accounts for a similar proportion of strains in the Caribbean and Central America (9, 17). In both these regions, the Haarlem family is believed to represent a remnant of the post-Columbian European colonization (17, 41). The finding of the Haarlem strains in our MDR population assumes particular significance in light of previous studies demonstrating the ability of the Haarlem genotypes to cause outbreaks of MDR-TB, as reported in Argentina (37), the Czech Republic (30), and Tunisia (34). The association between drug resistance and the Haarlem family has also been observed in other studies. Strains of this family predominated among MDR-TB cases in Tehran, Iran, and drug-resistant TB cases in Hungary; the rates of infection by the Haarlem genotype were 33.5% (15) and 66.2% (28), respectively. The LAM family, which was the third-largest genogroup found in this study, has a geographical distribution concentrated in Latin America and the Mediterranean region. Strains of the LAM family are hypothesized to be of European descent, and their high rate of occurrence in Latin American countries is supposed to be linked to the Portuguese and Spanish colonization of the New World in the 15th century (8). It is noteworthy that the LAM family of strains in our study was represented exclusively by the LAM9 genotype, which is considered to be ancestral to the entire LAM lineage (9). This finding may speak in favor of the historical presence of the LAM genotype strains in this part of Europe. Also, the fact that an important proportion of Polish MDR strains belonged to the LAM family may be somewhat related to the situation observed in Poland's eastern neighbor, Russia, where the LAM genotypes represent as much as 30 to 50% of the MDR strains (26, 33).

Upon review of the literature, the three genotype families characterized above (Haarlem, LAM, and T) appeared to be dominant in many European countries. They represent as much as 60% of M. tuberculosis strains isolated in Italy (31), 70% of the strains from Sweden (10), and 80% of the strains from France and Portugal (6, 11). Thus, the population structure revealed in our study is characteristic of a European country. Nonetheless, the finding in this study of four strains belonging to the Beijing genotype requires special attention. This is because strains of this genotype possess high epidemiological potential due to their enhanced transmissibility, particular propensity for developing drug resistance, and increased virulence. The Beijing genotype strains were initially found in China but have spread globally in recent years. The prevalence of the Beijing family is remarkably high on the Asian continent, and in many countries (China, South Korea, Hong Kong, and Thailand) it far exceeds 50% (21). In other parts of the world, the proportion of TB attributable to the Beijing genotype is much more variable: low in parts of Africa (0 to 4%), Latin America (<1%), and Western Europe (<6%); intermediate in North America and the Caribbean (8 to 27%); and high in the countries of the former Soviet Union (29 to 56%) (13). Several studies from across the world have documented the association between the Beijing genotype and drug resistance, including multidrug resistance (2, 15, 29, 39, 44, 45). This association is most clearly evidenced in the post-Soviet countries. A study performed in the Archangel Oblast in Russia showed that Beijing strains accounted for 77% of all drug-resistant TB cases and of all MDR-TB cases (44). In Estonia and Latvia, the Beijing strains were the cause of 87% and 58% of the MDR-TB cases, respectively (29, 45). In our study, the proportion of MDR strains belonging to the Beijing family (9%) was similar to those reported in Spain (6%) (39) and Denmark (8%) (32). At the same time, it was about half of those reported in the Netherlands (17%) (7) and Sweden (22%) (19). In all these countries, most of the patients with Beijing isolates were immigrants from Asia and Eastern Europe. Of the four Beijing strains identified in our study, only one was obtained from a foreign-born (Mongolian) patient. It is noteworthy that the number of MDR-TB cases due to the Beijing genotype in Poland remained stable. In a previous study on the molecular epidemiology of drug-resistant M. tuberculosis strains in Poland in 2000, four MDR and three non-MDR isolates were found to belong to the Beijing genotype. Three of those isolates (2 MDR and 1 non-MDR) were obtained from patients born in countries where the genotype is prevalent (38). All this suggests that even if foreign-born persons constitute the main reservoir for the Beijing family strains, these strains are being transmitted in Poland, although at a very low rate. Whether previously proposed variable host-pathogen compatibility in M. tuberculosis (18) could explain the propensity of specific mycobacterial lineages (such as the T or Haarlem clade) to spread within the Polish population to the detriment of the Beijing genotype remains purely speculative and open to discussion.

When comparing the genetic structure of the MDR strains with that of INH-monoresistant strains, some important features were observed. First, a total of 9 shared types (SIT37, SIT40, SIT47, SIT50, SIT53, SIT253, SIT280, SIT775, and SIT2538) described for the MDR strains were also found among INH-monoresistant strains. These spoligotypes comprised 34.8% of the former and 60.6% of the latter. This percentage difference was due to the fact that, for INH-resistant TB, spoligotype SIT53 accounted for 33.8% of all the cases (4). No such predominance of a single genotype was observed among MDR-TB cases. This finding might be indicative of a greater ability of drug-monoresistant M. tuberculosis strains to spread within the population compared with that of MDR strains. Second, 76.1% of the MDR isolates and 71.8% of the INH-resistant isolates yielded spoligotypes that had been reported in Poland previously (including SIT2538, shared by 2 MDR strains and 2 INH-resistant strains). This observation, together with the fact that nearly half of the spoligotypes identified among both MDR (48.1%) and INH-monoresistant (43.3%) M. tuberculosis isolates are present in Poland's neighboring countries, indicates that the vast majority of MDR- and INH-resistant-TB cases are due to infections with strains actively circulating in Poland and/or its neighbors. The fact that all but two patients (one Mongolian, with MDR-TB, and one Ukrainian, with INH-resistant TB) were Polish underlines the importance of in-country transmission. Third, the phylogenetic analysis revealed that the two most prevalent clades among MDR and INH-resistant M. tuberculosis strains were T and Haarlem, representing 46% and 80% of these groups, respectively. These results corroborate previous observations on the predominance of the T and Haarlem genotypes in the epidemiology of drug-resistant TB in Poland (5, 38). Furthermore, the presence of the T and Haarlem genotype isolates in such high proportions is consistent with the general features of the European population of M. tuberculosis.

Although the distributions of the four major clades (T, Haarlem, LAM, and Beijing) were significantly different in the MDR and INH-monoresistant strains (P = 0.0003), the most striking feature was the exclusive presence of Beijing and LAM lineages in the MDR group (8.7% and 13%, respectively). This suggests that, in addition to the possible emergence of MDR strains from a pool of INH-monoresistant strains of the T, H1, H3, and S lineages, a proportion of MDR-TB cases could also arise due to ongoing transmission of specific genotypes, such as LAM and Beijing. To better recognize the population genetic relationships between INH-monoresistant (n = 71) and MDR (n = 46) M. tuberculosis isolates, we drew and compared MSTs of these 2 data sets independently (Fig. 2A and B, respectively), as well as a combined tree for all the isolates together (n = 117) (Fig. 2C). As mentioned earlier, the greatest contributor to INH-monoresistant strains of M. tuberculosis in Poland is the cluster made up of SIT53 strains belonging to the T1 sublineage (33.8% of all INH-monoresistant strains). The T1 strains, represented by its prototype, SIT53, constitute the central node of the MST shown in Fig. 2A. Nonetheless, there are known difficulties associated with attempts to draw epidemiological inferences from ill-defined T-family spoligotypes (representing evolutionarily recent TB strains), which corresponded to about 30% of all entries in the SpolDB4 database present in all continents (9). Based on single spacer differences, they were initially stratified into 5 sublineages (T1 to T5), out of which 8 nested clades with robust spoligotyping signatures have been extracted so far (T-Tuscany, T1-Russia, T2-Uganda, T3-Ethiopia, T3-Osaka, T4-Central Europe, T5-Madrid, and T5-Russia). Consequently, further molecular characterization of the isolates designated “T” by default (17) is needed before phylogenetic inferences are drawn from the central node made up of spoligotype SIT53. In the MST shown in Fig. 2A, it is immediately followed by SIT50 (9.9%), SIT47 (7%), and SIT1253 (2.8%), which are successively placed on the MST within the well-defined H3, H1, and S subgroups. The spoligotyping results and clade definitions were also linked to the distribution of clinical isolates within PGG1 versus PGG2/3 (characterized by the lack of spacers 33 to 36). With the exception of a single orphan strain with an unknown signature (the 5th strain from the top in Table 2), isolates causing the bulk of INH monoresistance in Poland were made up of evolutionarily modern lineages of tubercle bacilli (T, Haarlem, and S). The T and Haarlem lineages/sublineages were duly represented by prototypes that characterized their description in the SpolDB4 database (9): T1 (SIT53), T1_RUS (SIT280), T3 (SIT37), T4 (SIT40), T5 (SIT44), H1 (SIT47), and H3 (SIT50).

• Open in new tab
• Download powerpoint

FIG. 2.

MST illustrating the potential evolutionary relationships of M. tuberculosis spoligotypes in Poland. The tree summarizes the phylogenetic links between two spoligotypes differing by genetic changes. The lengths of the branches indicate the levels of changes induced by loss or gain of spoligotype spacers in the 43-oligonucleotide format to induce a shift from one allele to another. Solid lines represent a single spacer change, while dotted lines represent 2 (black) or more (gray) spacer changes. The colors of the circles are proportional to the number of clinical isolates in our study: sky blue, 1 or 2 isolates; marine blue, 3 to 5 isolates; deep blue, 6 or 7 strains; and red, >8 strains. The remaining colors in the tree drawing denote strain lineages/sublineages for closely related strains such as T1_RUS2, T, H1, S, etc. The tree drawing was set to integrate data based on both single and multiple DVR deletions and to further differentiate between the loss of contiguous and noncontiguous multiple direct variable repeats (DVRs). (A) MST of isoniazid-monoresistant strains. (B) MST of MDR strains. (C) Combined MST for all the isolates together (n = 117). Note that the central node of the unrooted trees is represented by SIT53, which is the prototype of the T1 lineage.

The MST of the MDR strains involved five major branches (Fig. 2B). The largest essentially contained genotypes from the ill-defined T family. The other four branches exclusively or predominantly included genotypes from the Haarlem (H1 and H3), LAM, Beijing, and S families (Fig. 2B). The genotype SIT53 had a central position in this MST similar to that of the INH-monoresistant strains. However, as stressed above, further molecular characterization of SIT53 node strains is needed before phylogenetic inferences are drawn. Thus, a distinctive feature of the MDR population was the presence of the Beijing and LAM family strains and a significantly lower proportion of Haarlem family strains compared with the INH-monoresistant population. This observation suggests that, in addition to acquired resistance to RMP occurring within a pool of INH-monoresistant strains, a nonnegligible proportion of MDR-TB in Poland is due both to active transmission within the autochthonous population and to frequent contacts with neighboring countries. A second-line typing using MIRU-VNTRs and detailed contact tracing, in conjunction with characterization of specific katG alleles shared among INH-monoresistant and MDR strains (33, 38, 46), might be helpful to answer this question.

Interestingly, a combined tree for all the isolates together (Fig. 2C) (n = 117 isolates) reproduced most of the phylogenetic characteristics mentioned above for the MDR group. For example, similar to the MDR tree, the branches leading to the (i) T1_RUS, (ii) H1, and (iii) H3 sublineages were perfectly maintained, with the same succession of intermediary spoligotypes (Fig. 2C). The lack of insertion of INH-monoresistant strains within these well-defined branches, in conjunction with the exclusive presence of LAM and Beijing lineage strains within the MDR subset, underlines an active transmission of MDR strains belonging to the T, H1, T1-RUS2, and H3 lineages. On the other hand, strains belonging to the LAM and Beijing genotypic families are apparently transmitted independently of the prevailing INH-monoresistant pool. At the level of PGG classification, the only PGG1 strains included 4 Beijing strains in the MDR-TB sample. Added to the single PGG1 orphan strain in the INH-monoresistant group (see above), the total study sample (n = 117 strains) contained only 5, or 4.3%, of the PGG1 (ancient-lineage) strains compared to a very high proportion of PGG2/3 modern-lineage strains (112, or 95.7%). In conjunction with the noted absence of the CAS and EAI lineages in our study, this observation undeniably shows that circulating M. tuberculosis strains causing the bulk of drug resistance in Poland are made up of evolutionarily modern tubercle bacilli of the T, Haarlem, and S genotypic lineages.

Finally, we compared some of the basic epidemiological characteristics of the MDR-TB and INH-monoresistant-TB patients (Table 3). In both these groups, male patients predominated. The male-to-female ratios among MDR-TB patients and INH-resistant-TB patients were 6.7:1 and 2.7:1, respectively. Patients with MDR-TB were only slightly older than those with INH-resistant TB (median age, 50.5 versus 48 years). All of the patients studied were Polish, except two who were of Mongolian (MDR-TB group) and Ukrainian (INH-resistant-TB group) origin. There were no differences between MDR-TB cases and INH-resistant cases in terms of prevalence by region of residence. The prevalence of TB, regardless of the resistance phenotype, was highest in the western region of Poland, followed by the central and eastern regions. The proportion of patients who reported a history of TB was significantly higher among MDR-TB patients than among INH-resistant TB patients (84.8% versus 22.5%; P < 0.0001). Patients with MDR-TB were also more likely to have acid-fast bacillus (AFB) smear-positive sputa (P = 0.0014). The epidemiological status of MDR-TB in Poland, as inferred from patient demographic and clinical features, fits generally into the epidemiological picture of this infection described for other European countries. In a systemic review based on studies from 12 countries, being male, being less than 65 years old, and having a history of previous TB treatment were among major risk factors for MDR-TB in Europe (16). A reported history of anti-TB medication, which, according to the cited study, was the strongest determinant of MDR-TB, may suggest the acquired nature of drug resistance. A strong correlation between MDR-TB and sputum smear positivity in our study is also noteworthy. Such a correlation, very seldom encountered in similar investigations (25), may reflect the fact that patients with MDR-TB often did not receive adequate treatment at the outset (until the drug susceptibility results were available), thus remaining infectious for prolonged periods. A separate paper in preparation, devoted to the treatment of TB patients, shows that most, if not all, MDR-TB patients in Poland belong to the poorest members of the society; often having associated physical or mental handicaps, they were also characterized with poorer follow-up and longer delays to diagnosis and treatment (results not shown). Interestingly, there were virtually no MDR-TB cases among foreign-born persons in our study. This is in contrast to what has been observed in Western European countries, where the association between MDR-TB and being foreign born is remarkably consistent. The near absence of immigrants among MDR-TB cases can be explained by the relatively small immigrant population living in Poland.

In conclusion, our study is the first to provide a detailed population-based phylogenetic analysis of the MDR M. tuberculosis isolates compared to prevailing INH-monoresistant strains in Poland. Although spoligotyping, with its limited discriminatory power, may have overlooked finer intraspecies relationships within the M. tuberculosis complex, our results emphasize that the persistence of MDR-TB in Poland seemingly results from active transmission of well-established clones of M. tuberculosis historically linked to Poland within the autochthonous population. Some of the spoligotype-defined lineages and sublineages of the MDR group were shared by the INH-monoresistant group, which indicates that further contact-tracing epidemiology and a more discriminatory molecular methodology are warranted to pinpoint the major contributor to the development and persistence of MDR-TB in Poland. This is important in the context of the known instances of conversion reported in the literature from INH monoresistance to MDR (12), in which 10 out of 12 INH-monoresistant strains converted to MDR after treatment failure during therapy. If confirmed in extended studies, it will constitute a data-supported instance of INH monoresistance conversion to MDR status in Poland. Nonetheless, the exclusive presence of LAM and Beijing family strains within the MDR group suggests a particular role of these genotypes in the etiology of MDR-TB emergence in Poland.