M. tuberculosis Strains and Phenotypes | CDC EID

EID Journal Home > Volume 16, Number 2–February 2010

Volume 16, Number 2–February 2010
Research
Associations between Mycobacterium tuberculosis Strains and Phenotypes
Timothy Brown,1 Vladyslav Nikolayevskyy,1 Preya Velji, and Francis Drobniewski
Author affiliations: United Kingdom Health Protection Agency, London UK (T. Brown, F. Drobniewski); and Queen Mary College, University of London, London (V. Nikolayevskyy, P. Velji, F. Drobniewski)

Suggested citation for this article

Abstract
To inform development of tuberculosis (TB) control strategies, we characterized a total of 2,261 Mycobacterium tuberculosis complex isolates by using multiple phenotypic and molecular markers, including polymorphisms in repetitive sequences (spoligotyping and variable-number tandem repeats [VNTRs]) and large sequence and single-nucleotide polymorphisms. The Beijing family was strongly associated with multidrug resistance (p = 0.0001), and VNTR allelic variants showed strong associations with spoligotyping families: >5 copies at exact tandem repeat (ETR) A, >2 at mycobacterial interspersed repetitive unit 24, and >3 at ETR-B associated with the East African–Indian and M. bovis strains. All M. tuberculosis isolates were differentiated into 4 major lineages, and a maximum parsimony tree was constructed suggesting a more complex phylogeny for M. africanum. These findings can be used as a model of pathogen global diversity.

Tuberculosis (TB), caused by bacteria of the Mycobacterium tuberculosis complex (MTBC), remains a global threat to human health, which causes an estimated 2 million deaths annually (1). No horizontal gene transfer has been reported in MTBC, and the genome is more highly conserved than other pathogenic bacteria (2). Nevertheless, genotyping tools have recently identified several polymorphisms in the MTBC genome that have provided insight into its evolution. Three major groups of MTBC genome alterations have been reported: single nucleotide polymorphisms (SNPs), large sequence polymorphisms (LSPs), and polymorphisms within repetitive sequences such as variable number tandem repeats (VNTRs). The first 2 groups mark irreversible genetic events and can be used to construct phylogenies for M. tuberculosis (2–6). An association between geographic region and M. tuberculosis families, defined by specific polymorphisms, has been demonstrated. This geographic structuring producing genetically, and perhaps phenotypically, distinct MTBC populations may contribute to differences in clinical features such as severity of disease or prevalence of extrapulmonary disease (6–8) and should be considered during the development of new drugs and vaccines.

Sreevatsan et al. divided MTBC strains into 3 principal genetic groups (PGG1–PGG3) based on SNPs in codon 463 of katG and codon 95 of gyrA (2). More recently, on the basis of polymorphisms in the oxyR, katG, and rpoB genes, strains have been divided into 5 lineages (I–IV and M. bovis); lineages I, III, and IV represent subgroups within PGG1, and lineage II corresponds to PGG 2 and 3 (7). By combining these markers with LSPs RD239, RD105, RD750, RD711, and RD702, a small 7bp deletion in the pks15/1 gene and other SNPs, Gagneaux and Small were able to confirm these M. tuberculosis lineages and 2 lineages of M. africanum (6). The deletions RD9 and TbD1 are useful phylogenetic markers for other members of MTBC complex and ancestral M. tuberculosis strains (3). The loss and acquisition of repeats or spacers in the direct repeats region (9) does not appear to limit their value in biogeographic and phylogenetic studies (10,11).

Genotypic variation of MTBC strains at various geographic settings and significant associations between certain allelic variants at VNTR loci, MTBC lineages, and spoligotyping families have been reported (7,12–15). However, most studies used single genotyping methods on small populations or convenience samples. Population-based studies have focused primarily on areas of low- to middle-TB incidence, and it is unclear whether the results are universally applicable (16–18). Larger population-based studies on geographically diverse populations are needed to establish the phylogenetic, epidemiologic, and clinical relevance of such associations.

London accounts for nearly half of all TB cases in the United Kingdom (≈3,300 cases in 2006; incidence rate 44.8/100,000). Because 75% of these TB patients were born abroad (19), (Health Protection Agency update; www.hpa.org.uk), and clinical signs of disease develop within 2 to 3 years of arrival, we believe that the multicultural and diverse community in London provides a unique setting for studying the global biodiversity of MTBC. We aimed to establish whether MTBC isolates circulating in the London population are a useful model of global diversity, to determine the phylogenetic relevance of polymorphisms in repetitive regions of the MTBC genome, especially for M. africanum and its position in TB evolution, and to investigate associations between lineage and phenotype.

Materials and Methods
Study Design and Bacterial Isolates
One isolate from each of the 2,261 MTBC culture-positive patients was included in this prospectively designed population study. These isolates were collected from patients in all 30 London National Health Service hospitals between April 1, 2005, and March 31, 2006. Demographic data, including gender, date of birth, and country of birth were assigned to world regions according to an existing United Nations classification (20).

Identification
Cultures were identified by using standard phenotypic identification tests (21) and molecular methods (Genotype Mycobacterium CM, AS, and MTBC kits; Hain Lifescience GmbH, Nehren, Germany) and the INNO LiPA Rif TB assay (Innogenetics, Ghent, Belgium) performed as recommended by the manufacturer. DNA was extracted from cultures using chloroform extraction as described (22). Isoniazid, rifampin, ethambutol, streptomycin, pyrazinamide, and ciprofloxacin susceptibilities were determined by using the resistance ratio method (21).

Genotyping
All extracts were typed by using automated 15 mycobacterial interspersed repetitive unit–VNTR (MIRU-VNTR) fragment analysis (23–26). Clustered isolates were further genotyped by using an extended panel of 7 hypervariable VNTR loci (27). Data were exported to BioNumerics (Applied Maths, Sint-Martens-Latem, Belgium) for cluster analysis.

Spoligotyping was performed according to the manufacturer's instructions (Isogen Lifescience, IJsselstein, the Netherlands) (9). Images were digitized and entered into BioNumerics software by using the BNIMA module (Applied Maths). Spoligotypes were assigned to families and subfamilies by using the online tools at http://cgi2.cs.rpi.edu/~bennek/SPOTCLUST.html (10). We have used the established spoligotyping families Beijing, Central Asian (CAS), East African–Indian (EAI), and M. bovis as lineage designations, as well as European American (EuroAm) (13,28) for the M. tuberculosis lineage, which includes the X, T, LAM, S, and Haarlem families.

Other Methods
Detection of TbD1 and RD9 (3,13) was conducted by PCR fragment analysis (3). Reverse hybridization methods were used to analyze the 4 lineage-defining SNPs in 3 genes (oxyRC37T, katGC87A, rpoBT2646G, and rpoBC3243T) reported by Baker et al. (7) for selected isolates (n = 259) (12) and mutations in katG, inhA, and rpoB genes associated with drug resistance (22).

Data were analyzed by using Excel, BioNumerics (Applied Maths), SPSS 12.0 (SPSS Inc, Chicago, IL, USA) software and online interactive statistical tools (www.quantitativeskills.com/sisa). Categorical variables were analyzed by using relative risks (RRs), odds ratios (ORs), and the χ2 test. Discrimination power of genotyping methods was assessed using the Hunter-Gaston index (29).

Results
Diversity within the Study Population
We studied 2,261 isolates, representing 95.7% of all the bacteriologically confirmed TB cases reported in London from April 1, 2005, through March 31, 2006. Using routine phenotypic and genotypic methods, we identified 99.1% (2,241) as MTBC; the remaining 20 were too heavily contaminated for analysis.

Spoligotypes were generated for 98.8% (2,233) of the isolates; 656 types were identified, of which 458 were unique and 198 were shared by groups of 2–221 isolates. Isolates were assigned to families and subfamilies on the basis of their spoligotype by using the online tools at http://cgi2.cs.rpi.edu/~bennek/SPOTCLUST.html. All but 4 spoligotypes were assigned to >1 of 36 groups; 88.4% of isolates were assigned to a single spoligotyping family or subfamily. The remaining 11.6% were assigned to 2 families, albeit with given probabilities of <0.9. All the main spoligofamilies seen globally were represented within this population (Table 1).

Isolates were cultured from a variety of body sites; 57% were of pulmonary origin. Where known, 60% of isolates were cultured from male patients and 40% from female patients; median age was 33 years. The COB was available for 1,381 (61.0%) patients; 1,157 (83.8%) were born in 89 countries outside the United Kingdom (Appendix Table). The population included representatives from all regions of the world (20).

abrir aquí para acceder al extenso documento CDC EID completo:
M. tuberculosis Strains and Phenotypes | CDC EID