Tuberculosis (TB) remains a major cause of morbidity and mortality worldwide, causing more than 2 million deaths a year 12. Pakistan with a population of 140 million and a growth rate of 3.5% 3 has the seventh highest tuberculosis rate despite widespread BCG vaccination. TB prevalence in Pakistan can be attributed to poor health care systems and limited diagnostic and treatment modalities for TB4. The TB problem is also compounded by multi-drug resistance (MDR, resistance to at least rifampicin and isoniazid), WHO report suggests that globally 3% of M. tuberculosis isolates are MDR-TB5. Resistance to TB drugs is recognized in Pakistan67. While community based information is lacking, laboratory data suggests an increasing frequency of MDR from 14% in 1999 to 28% in 20046 and 47% in 20068.

International databases such as the SpolDB4.0 have revealed the clonal structure of M. tuberculosis isolates in different geographical settings. SpolDB4.0 data base further defines super families specific to certain locations9. Genotypic information has further expanded our understanding of strain prevalence and transmission 10111213. A few predominant genotypes circulating throughout the world e.g. Beijing, Haarlem, and African clusters have been associated with a number of major outbreaks 141516. These major strain groups have been described as being predominant pathotypes in the world 17. The abundance of polymorphism indicates that transposition and homologous recombination are the major events contributing to the diversity of M. tuberculosis strains 18. In addition, polymorphism seen with different molecular markers also describes mutual association. This supports the hypothesis that M. tuberculosis has a strong clonal population structure 18. In support of phylogeographical population structure of M. tuberculosis, differences in strain genetics may be responsible for the variation in BCG efficacy 19202122.

Predominant M. tuberculosis clades from the Indian sub-continent include Central Asian strain (CAS) 923 and Beijing strains 1124252627. Central Asian strain 1 (CAS1) are defined by absence of spacers 4–7 and 23–34 28. While, Beijing strains were characterized with the absence of 1–34 spacers in direct repeat region (DR). Beijing strains are reported to constitute about 50% of strains in far East-Asia and 13% of isolates globally29. East African-Indian strains, the T clade and Haarlem strains have also reported from India, Afghanistan and Iran3031. In Pakistan predominance of CAS1 (39%) with a 6% prevalence of Beijing isolates has previously been reported 32.

Globally, MDR-TB outbreaks have been associated with Beijing and Haarlem families 3334. In order to understand the population structure of M. tuberculosis in Pakistan, strains from the four provinces, Punjab, Sindh, Balochistan and NWFP were spoligotyped. Genotypic information was correlated with drug resistance to determine association between strain types and MDR. Predominant clades obtained were further analyzed to distinguish between "ancestral" versus "modern" lineages of tubercle bacilli based on the presence or absence of the TbD1 region.

This study presents novel information regarding genotypic diversity and drug resistance of M. tuberculosis strains in Pakistan. We found that 411 (44%) of 926 study isolates to be CAS1 or ST26 strains 25, confirming previous reports of 39% prevalence of CAS1 amongst M. tuberculosis strains 32. The results further showed that in addition to ST26 the majority of spoligotypes belonged to the CAS genotype. CAS1 has been also identified by recent studies as a predominant strain in Delhi25 and Mumbai44. Prevalence of Beijing strains in our study at 3% (n = 25) compares well with data from Delhi, where 8% of 105 isolates are reported to be of Beijing family25. We further identified 36 clusters (Pak clusters) not identical with any of the STs described within SpolDB4.0.

Comparison of pulmonary and extra pulmonary sources showed a significant association of 'Orphan' spoligotypes with extra pulmonary disease. This difference between pulmonary and extrapulmonary spoligotypes may be attributed to the greater ease with which pulmonary strains may be transmitted via aerosol routes, leading to greater transmission and strain clustering.

The global population structure of M. tuberculosis is reportedly defined by six phylogeographical lineages, each associated with specific human populations http://www.pnas.org/cgi/doi/10.1073/pnas.0525482992945. It has been suggested that particular lineages of M. tuberculosis might be adapted to specific human populations and maladapted to others45. Strain differences in different geographical regions may be linked to different ethnic subpopulations in these regions, and their migration histories 46.

The CAS strains have been shown to be predominant 3047 in this region, the distribution of CAS1 strains was greater in the province of Punjab (P < 0.01, Pearsons Chi-square test) as compared with Sindh, NWFP, and Balochistan. This could be due to the shared border and similar population mix between Punjab now in Pakistan (West Punjab prior to the 1947 partition of India), and East Punjab, (now Punjab, North India).

Although not statistically significant, East African-Indian strains were more prevalent in Sindh. The fact that 4% of our isolates were East African-Indian is comparable to an earlier study from Delhi, reporting 8% EAI strains in their population 25. Both these figures are in contrast to a study from Southern India indicating an 80% prevalence of TbD1+/EAI isolates amongst their samples 43. It has been hypothesized that lineages that are rare in specific human population are not adapted to spreading within these populations and show a significantly lower case-rate ratio. This hypothesis is supported by a recent report30 suggesting that TB in India is essentially caused by historic clones of tubercle bacilli which circulate with geographic predilection.

Deletion analysis demonstrated the absence of TbD1 region in CAS1 and its sub-families. A recent study 43 also suggests that CAS1 family evolved as a result of an evolutionary event causing TbD1 deletion from a common ancestor1841. The study further suggests that strains within CAS genotype are related through minor genetic changes. The similarity seen in our study between Pak clusters 1–12 and CAS genotype suggests an evolution from a common ancestor belonging to the same phylogenetic lineage CAS.

Haarlem4 and T1 were also identified amongst our isolates. It is reported that more than 60% of ST127 (H4, modern type) are localized in Iran and Russia. Their presence in our population suggests the role of traditional migratory routes from central Asia in the history of tuberculosis. In addition to the predominant groups, we also found clusters of rare/localized shared types listed in SpolDB4.0 that have previously been found only in North America, Australia and Europe (CAS1_KILI) as well as those found in neighboring Iran and India (LAM 9, LAM 6).

In the absence of national surveillance for drug resistance in Pakistan, the majority of published reports rely upon laboratory data based on passive specimen collection. There is thus little information about the burden of MDR at a community level. WHO estimates suggest an MDR rate of under 3% amongst new cases and 20–40% amongst previously treated TB cases48. A recent report of 1.8% MDR-TB, in untreated cases in Pakistan corroborates the WHO estimates 49. Since the samples in our study were not collected through active case finding, the high MDR rate noted in our specimens is likely to reflect the large number of previously treated patients included in this study. This however, cannot be confirmed due to absence of prior treatment history for our patients.

In agreement with an earlier report 32 CAS1 and related CAS sub-family spoligotypes in this study showed no correlation with MDR-TB. However, a higher relative risk of MDR amongst Beijing strains was noted. Association between Beijing strains and MDR varies worldwide 5051, A recent study from Mumbai, India reported a higher frequency of Beijing strains (35%) amongst MDR isolates 44. The distinctive feature of Beijing MDR-TB outbreak is accelerated transmission as compared with other MDR-TB outbreaks52. Further, East African-Indian strain was also found to be associated with MDR supported by a study from Iran 53. Recent studies moreover suggests that MDR-TB strains may be responsible for emergence of XDR-TB cases 5354.

The higher occurrence of streptomycin resistance in Beijing strains as compared to CAS genotype has importance in that particular lineage of M. tuberculosis might harbor polymorphisms which make them resistant to certain anti-tuberculosis drugs.

In our study population, association of Beijing strains with a younger age group (<15 years) is concerning. It suggests that Beijing strains may be an emerging strain type and is likely to increasingly contribute to the burden of drug resistant tuberculosis in this region.

The identification of a dominant spoligotype, CAS, similar to previously identified isolates in India and Bangladesh illustrates an important trend in the M. tuberculosis infection pattern in the South Asian region. All predominant clusters apart from CAS1 strains were equally distributed in the country suggest a continual transmission of strains. This data presents a comprehensive evaluation of the strain-to-strain variability in M. tuberculosis, important in phenotypic consequences, also these phylogeographical strain variation may affect the development of new diagnostic tools, drugs, and vaccines for treatment in the endemic region.

The authors declare that they have no competing interests.

MT conducted the experimental work and prepared the manuscript. ZH supervised the laboratory work and study design. AK and AA provided technical help with spoligotyping. RS provided statistical data analysis support. SG provided advice regarding MTB strain typing. RH planned the idea and supervised the study. All authors read and approved the final manuscript.