|Home | About | Journals | Submit | Contact Us | Français|
We used 909 strains to compare the population structures of the Mycobacterium tuberculosis Beijing family between different birth-year cohorts in Japan. The results revealed that the spread of a modern sublineage that has high transmissibility is currently increasing, while the spread of an ancient sublineage, STK, has significantly decreased in younger generations.
Comparative studies of the Mycobacterium tuberculosis population structure have generated interest in understanding its epidemiological relevance to human disease (2, 3, 9, 15, 21). One of the most extensively studied clades with respect to host-pathogen compatibility is the Beijing family (5, 7, 8, 16, 23), which is highly prevalent in East Asia (26). Recent studies have demonstrated the worldwide dissemination of the modern Beijing family strains (having IS6110 insertions in the NTF region) (1, 14, 17, 18) and have led to speculation about the hypervirulent features of this sublineage (4, 7, 13, 16). In contrast to the worldwide prevalence of modern Beijing strains, the ancient Beijing strains (having an intact NTF region) are highly diverse and dominant in Japan (11, 27). Studying this singularity of the Beijing strains in Japan would expand our understanding of the heterogeneity in the fitness of different sublineages. In this study, we used data from a population-based study lasting 5 years and 8 months to investigate the probable shift in the M. tuberculosis population structure during the previous decades and determine the future trend in Japan, with special attention to the modern Beijing strains.
We obtained 909 M. tuberculosis isolates from newly diagnosed pulmonary tuberculosis (TB) patients between April 2002 and December 2007. These isolates accounted for approximately 70% of the new culture-positive pulmonary TB cases detected during the collection period in Kobe City. The isolates were phylogenetically classified as belonging to the Beijing clade or other clades by spoligotyping (12) and to the ancient and modern Beijing sublineages on the basis of the presence of IS6110 in the NTF region (19, 27). The ancient and modern Beijing strains were subclassified by analyzing 10 synonymous single-nucleotide polymorphisms (6, 8, 11). Molecular typing based on the variable number of tandem repeats (VNTR) method with 19 loci, which comprised the 15 loci of Supply et al. (24) and 4 hypervariable loci (QUB-11a, QUB-3232, VNTR-3820, and VNTR-4120) (10), was performed for all strains. Recent transmission was suggested by the clustering of identical VNTR profiles. The presence or absence of region of difference 181 (25) in all the Beijing strains was analyzed.
Of the 909 isolates, 714 (78.5%) were classified as Beijing family strains (Table (Table1),1), and these included 44 Beijing-like strains. Except for the sequence type 11 (ST11) and ST26 sublineages, all the Beijing strains contained deletions in region of difference 181. All the modern strains harbored only one IS6110 insertion element. The non-Beijing family strains comprised 32 spoligotypes. The details of the genetic data of all isolates are summarized in Table S1 in the supplemental material.
The average patient age, patient gender, cluster rate (number of clustered isolates/total number of isolates), and proportion of multidrug-resistant strains did not differ significantly between the Beijing and non-Beijing strains (Table (Table1).1). However, further classification of the Beijing family strains revealed a significantly high cluster rate (42.9% versus 31.3%, P = 0.022, Pearson's chi-square test) in the case of the modern Beijing strains. This indicates that the transmissibility of the modern Beijing strains is higher than that of the non-Beijing strains. Further, the cluster rate of the STK strains was observed to be low (12.6% versus 31.3%, P < 0.01). Moreover, the average age of patients affected by the modern strains was significantly younger than that of those affected by the non-Beijing strains (58.7 versus 64.9 years, P < 0.01, Welch's t test) (Table (Table1).1). These results imply that the transmission of various sublineages is different, which in turn implies that the population structure of the Beijing M. tuberculosis strains that are prevalent in Japan is more dynamic than stable. Human immunodeficiency virus infection and introduction by foreigners remain minor factors in the epidemiology of TB in Japan (22), and these factors have negligible effects on the population dynamics of M. tuberculosis strains there.
We attempted to determine the shift that had occurred in the population of the Beijing family strains during previous decades by comparing the population structures of strains isolated from elderly TB patients (these strains represent the population structure that existed decades ago) and young TB patients (these strains reflect the population structure of currently prevalent strains) (Table (Table2).2). The major cause of TB in the elderly is the reactivation of the M. tuberculosis strains acquired prior to World War II, when TB was highly prevalent in Japan (20, 22). The fact that the cluster rate in the cohorts born in or before 1925 is lower than that of cohorts born later (Table (Table2)2) indicates that the elderly are not actively involved in recent TB transmission. A comparison of the M. tuberculosis populations isolated from cohorts born in different years (Table (Table2)2) suggested that the population structure of the M. tuberculosis Beijing family strains in Japan before World War II—when TB was highly prevalent—was significantly different from that of the currently prevalent strains. A probable artifact bias from an unrecognized large-scale outbreak could be ruled out since no large cluster formation was found in a particular generation (see Table S2 in the supplemental material). Notably, the incidence of disease due to the modern Beijing strains was low in the elderly (18% in the cohort born in 1925 or earlier and 15.3% in the cohort born from 1926 to 1935) but highest in the young (31.1% in the cohort born in 1965 or later) (Table (Table2)2) (P = 0.016, Z test for the proportions between the cohort born in 1925 or earlier and the cohort born in 1965 or later).
The same trend was observed when we rechecked the data from our previous study using 355 Beijing family strains obtained from Osaka (175 isolates) and Kobe (180 isolates) (27). The data from Osaka showed that 2 out of 21 (9.5%) isolates from patients more than 75 years old were modern strains, while this number was 9 out of 35 (26.5%) in patients less than 35 years old. These data, which were collected from different cities in Japan, strengthened the conclusion of this study that the modern Beijing strains show a higher rate of occurrence in the young. On the other hand, the strains belonging to the STK sublineage, which accounted for 23.3% of the strains isolated from the cohorts born in 1925 or earlier, exhibited a significantly low incidence in the cohorts born later (Table (Table2).2). The cluster rates of these two sublineages were quite different (42.9% versus 12.6%) (Table (Table1).1). From these results, we assumed that the modern Beijing strains, with a high degree of transmissibility, are currently spreading in Japan, while there is a continuous shrinkage of the STK strains, with low transmissibility. Multidrug resistance does not appear to be the reason (Table (Table1)1) for the prevalence of the modern strains, as reported in other studies (13, 16). The main reason for the occurrence of STK would be the endogenous reactivation of TB in elderly individuals. There were no significant differences in the incidence of the other ancient Beijing sublineages among the various cohorts (Table (Table2).2). It was reported that the modern Beijing sublineage shows a significantly higher transmissibility than the ancient Beijing sublineage among homeless people in Japan (28). Taken together, these data indicate that the population structure of M. tuberculosis in Japan may undergo changes and eventually resemble the typical worldwide situation, with a predominance of the modern Beijing sublineage. Further studies analyzing paleopathological samples, such as paraffin-embedded lung biopsy specimens or old cultures preserved for several decades, would be required to prove this hypothesis.
It is interesting to assume that the observed trends in the case of the modern and STK strains may be associated with the Mycobacterium bovis bacillus Calmette-Guérin (BCG) vaccination. In Japan, mass vaccination with BCG was initiated in 1942. The individuals in the cohort born in the year 1925 and earlier were mostly infected with TB without receiving the BCG vaccination since 81% of the people in the cohort born from 1921 to 1925 were infected before they reached the age of 30 (22). On the other hand, it is highly probable that the cohort born in and after the year 1965 were administered BCG vaccines. It has been previously reported that the BCG vaccination favors the positive selection of modern Beijing strains (13). Our results support this finding, and we further demonstrate the possible negative selection of STK strains.
In conclusion, the population structures of the Beijing family strains in Japan were different for different birth cohorts. We believe that the modern Beijing strains, with a high degree of transmissibility, are currently spreading in Japan, while the spread of the STK strains, with low transmissibility, will decrease in the future. It is essential to continuously monitor the population shift for a long period in order to evaluate the effectiveness of current TB control measures and achieve better TB control in Japan.
This work was supported by grants from JSPS Grant-in-Aid for Scientific Research (A) (20249007) and the United States-Japan Cooperative Medical Science Program (TB and leprosy panel).
Published ahead of print on 6 August 2009.
†Supplemental material for this article may be found at http://jcm.asm.org/.