OUP user menu

Impact of drug resistance on fitness of Mycobacterium tuberculosis strains of the W-Beijing genotype

Olga S. Toungoussova, Dominique A. Caugant, Per Sandven, Andrey O. Mariandyshev, Gunnar Bjune
DOI: http://dx.doi.org/10.1016/j.femsim.2004.05.012 281-290 First published online: 1 November 2004

Abstract

Mycobacterium tuberculosis strains of the W-Beijing genotype became a common cause of tuberculosis during the past years and they are often associated with drug resistance. The biological factors facilitating the selection and wide dissemination of these strains are not known. To determine how acquisition of drug resistance affected growth of strains of the W-Beijing genotype, the growth of 55 M. tuberculosis isolates were studied using the BBL MGIT Mycobacteria Growth Indicator Tube and the BACTEC MGIT 960 System. Susceptible strains of non-Beijing genotypes were found to be the most fit strains. Drug-resistant strains of non-Beijing genotypes were more likely to grow slower than susceptible strains (P=0.001). Drug-resistant strains of the W-Beijing genotype had two tendencies of growth: some of them showed reduced growth compared to susceptible strains, while others did not show loss of fitness measured as growth.

Keywords
  • Mycobacterium tuberculosis
  • Fitness
  • W-Beijing family

1 Introduction

The incidence of tuberculosis has been rising throughout the world [1] and has been accompanied by increasing anti-tuberculosis drug resistance. Surveillance data obtained from developing countries indicate that nearly 50 million persons worldwide may be infected with drug-resistant Mycobacterium tuberculosis strains [2]. During the past years, the proportion of M. tuberculosis strains having multidrug resistance (MDR), i.e. resistance to at least rifampin and isoniazid [3], has increased dramatically. It is obvious that the evolution and spread of anti-tuberculosis drug resistance is due to improper use of antibiotics. The reversion of this situation and combating tuberculosis caused by drug-resistant strains depend on several measures such as diagnosis, proper use of antibiotics, prescription of adequate treatment regimens, prevention of non-compliance and introduction of new anti-tuberculosis drugs.

Many pathogens develop genotypes that have highly effective virulence including more efficient transmission [4]. M. tuberculosis strains belonging to the W-Beijing family is a good example. These isolates exhibit closely related restriction fragment length polymorphism (RFLP) patterns and contain the last 9 of the 43 polymorphic spacer sequences in the chromosomal direct repeat (DR) locus by spoligotyping [5]. Such MDR isolates were identified in the USA in 1990s and designated the W strains [6]. In 1995, M. tuberculosis isolates with similar characteristics were found in the Beijing province, China, and were designated the Beijing genotype [5,7]. These genetically related isolates, variously called W or Beijing, are now referred as the W-Beijing family [6].

Scientific publications indicate that the W-Beijing strains are prevalent in population based studies [812]; however most of these studies have bias collections. Strains of the W-Beijing family have been shown to acquire resistance to anti-tuberculosis drugs more frequently than strains of other genotypes [812]. Positive selection of the W-Beijing M. tuberculosis strains accompanied by rising resistance to anti-tuberculosis drugs might explain their dissemination worldwide.

Mutations leading to drug resistance development may influence the fitness of the microorganism. The definition of fitness includes a microorganism's ability to survive, reproduce and to be transmitted [13]. Evidence has been provided that drug-resistant mutants of M. tuberculosis have reduced fitness compared to their parent susceptible strains [1416]. The relative fitness of the W-Beijing M. tuberculosis strains after they acquire resistance to anti-tuberculosis drugs has, however, not been studied. Drug-resistant W-Beijing strains, like ordinary resistant mutants, may loose fitness when they acquire resistance to anti-tuberculosis drugs. If they do, then they should not be competitive towards susceptible M. tuberculosis strains unless antibiotics are present. The reality shows that the W-Beijing strains often constitute the bulk of primary MDR [1012].

Growth in vitro is one of several fitness indicators. The aim of the present study was to measure and compare growth of drug-susceptible and resistant W-Beijing and non-Beijing M. tuberculosis strains. To determine the impact of drug resistance on fitness, the growth of susceptible and drug-resistant strains belonging to both genotypes was analysed using the BBL MGIT Mycobacteria Growth Indicator Tube and the BACTEC MGIT 960 System. The growth of nine pairs of isolates obtained from the same patients before and after acquisition of resistance to additional drugs was also analysed.

2 Materials and methods

2.1 M. tuberculosis strains

Clinical M. tuberculosis strains belonging to the W-Beijing and non-Beijing genotypes were included in the study. A total of 55 M. tuberculosis strains (9 susceptible of the W-Beijing genotype, 15 resistant of the W-Beijing genotype, 11 susceptible of non-Beijing genotype, and 20 resistant of non-Beijing genotype) were selected from a collection of clinical M. tuberculosis isolates (Table 1), made during 1998–2001 from patients with pulmonary tuberculosis in Archangel, Russia [12,17]. Selection of strains involved clustered strains and strains having different mutations in the gene encoding the β-subunit of RNA polymerase (rpoB), leading to rifampin resistance. A cluster of M. tuberculosis strains was defined as two or more strains exhibiting 100% identical IS6110 RFLP patterns [5,18,19]. In addition, nine pairs of M. tuberculosis strains isolated from the same patients with identical IS6110 RFLP patterns, but showing different susceptibility patterns over time were included. Two of the nine pairs were isolated from patients originally from former Yugoslavia (patient 1) and Somalia (patient 2) living in Norway, the remaining seven pairs were obtained from Russian patients living in Archangel, Russia (Table 1).

View this table:
Table 1

Characteristics of the 55 M. tuberculosis strains from Archangel, Russia

Strain no.Resistance toCluster IS6110Mutation in rpoB geneComment
Non-Beijing
418/98SusceptibleIReproducibility
421/98SusceptibleIReproducibility
853/00SusceptibleIReproducibility
5830/00SusceptibleIReproducibility
216/01SusceptibleYReproducibility
736/02SusceptibleYReproducibility
7004/01SusceptibleZReproducibility
722/02SusceptibleZReproducibility
5871/00SusceptibleK
350/98SusceptibleK
420/98SusceptibleM
419/98HI
5882/00HK
193/01HM
396/98HSJ
400/98HSJ
6964/01HSJ
3633/02HSJ
398/98HSEM
6963/01HRSJ531 TCG→TTG
7028/01HRS531 TCG→TTG
83/99HRSEP531 TCG→TTG
5835/00HRSEP526 CAC→TAC
5849/00HRSEP533 CTG→CCG
5849/01HRSEP533 CTG→CCG
6965/01HRSEQ516 GAC→GTC
231/01HRSEQ516 GAC→GTC
219/01HRSEQ516 GAC→GTC
7000/01HRSEJ531 TCG→TTG
7014/01HRSED531 TCG→TTG
721/02HRSE531 TCG→TTG
368/95SusceptiblePatient 1
258/96HR531 TCG→TTGPatient 1
205/94HPatient 2
352/95HR531 TCG→TTGPatient 2
396/98HSPatient 3
68/99HRS533 CTG→CCGPatient 3
446/97HSPatient 4
158/98HRS531 TCG→TTGPatient 4
W-Beijing
5824/00SusceptibleA
3639/02SusceptibleA
249/01SusceptibleB
6997/01SusceptibleB
724/02SusceptibleB
729/02SusceptibleB
423/98SusceptibleC
7012/01SusceptibleG
7013/01SusceptibleG
7027/01HRB516 GAC→GGC
215/98HRSC516 GAC→GTC
5828/00HRSEA531 TCG→TTG
5868/00HRSEA531 TCG→TTG
3639/02HRSEA531 TCG→TTG
423/99HRSEA531 TCG→TGG
848/00HRSEB513 CAA→CTA
5866/00HRSEB513 CAA→CTA
206/01HRSEB513 CAA→CTA
7002/01HRSEB513 CAA→CTA
3600/02HRSEB531 TCG→TTG
360/98HRSEC531 TCG→TTG
422/99HRSEC531 TCG→TTG
7019/01HRSEG526 CAC→TAC
192/01HRSEX531 TCG→TTG
503/97SCPatient 5
441/98HRSC516 GAC→GTCPatient 5
205/97SPatient 6
160/98HRSE526 CAC→CTCPatient 6
395/98HSAPatient 7
78/99HRSA531 TCG→TTGPatient 7
346/98HSBPatient 8
5836/00HRSEB513 CAA→CTAPatient 8
443/97HSEAPatient 9
5823/00HRSEA531 TCG→TGGPatient 9
  • H, isoniazid; R, rifampin; S, streptomycin; E, ethambutol.

  • Strains of the same RFLP were assigned to a cluster designed by a capital letter.

The identification of the isolates was performed using 16S rRNA hybridisation technique (AccuProbe; GenProbe Inc., San Diego, CA, USA) and standard microbiological tests (niacin accumulation test and nitrate reduction test). RFLP analysis of M. tuberculosis DNA was performed according to the internationally standardised methodology [20,21]. Spoligotyping was performed by using a commercially available kit (Isogen Bioscience BV, Maarssen, The Netherlands) according to the instructions supplied by the manufacturer, as previously described [22]. M. tuberculosis strains were defined as belonging to the W-Beijing family if they showed the following characteristics: the strains harboured a high number of IS6110 copies (from 13 to 18) and more than two-thirds of them were present at the same genomic sites (they clustered within 60% similarity); and they had identical spoligotyping results, showing hybridisation only with the last 9 of the 43 possible spacers [5,7,12]. M. tuberculosis strains of the W-Beijing and non-Beijing genotypes included isolates susceptible and resistant to the first line anti-tuberculosis drugs (ethambutol, isoniazid, rifampin and streptomycin). Drug susceptibility testing of all strains was performed using the radiometric broth method (BACTEC, Becton Dickinson Diagnostic Systems, Sparks, MD, USA) [2325]. Mutations associated with rifampin resistance in the rpoB gene were identified using the Inno-LiPA Rif. TB test (Innogenetics N.V., Ghent, Belgium) [26,27].

2.2 Strain cultivation and growth

The strains were taken from the −70 °C freezer, defrosted, inoculated to Lowenstein–Jensen media and cultivated for three weeks at 37 °C. A suspension of bacilli was then prepared in 4 ml Middlebrook 7H9 broth (BACTEC, Becton Dickinson Diagnostic Systems, Sparks, MD, USA) containing glass beads. The suspension was vortexed for 20 s and then allowed to sediment for 20 min. The supernatant was transferred to another sterile tube and allowed to sediment for another 15 min. The supernatant was transferred to a new sterile tube, adjusted to turbidity comparable to a McFarland No. 0.5 standard (1.5×108 bacterial cells/ml), and then diluted 1:500. A volume of 0.5 ml of the dilution (1.5×105 bacterial cells) was added to the BBL MGIT Mycobacteria Growth Indicator Tube (BACTEC, Becton Dickinson Diagnostic Systems, Sparks, MD, USA). The tubes were entered into the BACTEC MGIT 960 System (BACTEC, Becton Dickinson Diagnostic Systems, Sparks, MD, USA) [28], incubated at 37 °C, and monitored for increasing fluorescence. The BACTEC MGIT 960 System performs monitoring for fluorescence (in units) every hour. It does not, however, automatically provide growth curves. The results of strain growth therefore had to be obtained manually by printing out results about every 24th hour during the first 96 h and every 6th hour after the 96th hour. This methodology was tested in a pilot study on four susceptible M. tuberculosis strains (two W-Beijing and two non-Beijing). The reproducibility of the method was demonstrated by testing eight susceptible M. tuberculosis strains of non-Beijing genotypes in three independent experiments starting from cultivation on Lowenstein–Jensen media each time. Standardization of inoculum was achieved by adjusting to turbidity comparable to a McFarland No. 0.5 standard 1 ml of which contains 1.5×108 bacterial cells. The standard provides approximate estimate of number of bacterial cells in a liquid suspension. Measurements of growth for nine pairs of strains obtained from the same patients were performed twice in independent experiments. The remaining strains were tested once.

The prepared dilutions were also cultivated on blood agar plates and incubated at 37 °C for 48 h to detect bacterial contamination. None of the dilutions used were contaminated.

2.3 Statistical analysis

The main continuous variable in statistical analysis was the time when a strain reached the growth level of 200 units. This variable was calculated using graphics of growth curves and mathematical formula: tt1/t2t1=yy1/y2y1, where t= time, and y= growth in units. The lag phase was defined as the time from the start of cultivation to the beginning of detectable growth. In order to test for differences in growth rate, the mean time from the beginning of growth to 200 units of growth was also calculated.

SPSS for Windows version 9.0.1 (SPSS, Inc., Chicago, IL) was used for the statistical analysis. Differences between groups were tested by parametric ANOVA test. Difference between two means was used to test for reproducibility of growth of the same strains measured during three independent experiments. A P value of <0.05 was considered significant.

3 Results

3.1 Pilot study

In order to test the methodology and to select the interval between monitoring for increasing fluorescence as a measurement of growth, a pilot study was performed using four M. tuberculosis strains susceptible to the first line anti-tuberculosis drugs (Fig. 1). Two of these strains belonged to the W-Beijing genotype. The growth of the strains was measured about every 24th hour during the first 96 h. The growth of one of the strains began at 93 h. Then, measurements of growth every hour during 6 h revealed that the observed difference of growth from 1 to 6 h was not significant. Thus, the interval between two measurements of growth was selected to be 6 h.

Figure 1

Growth curves of four M. tuberculosis strains susceptible to the first line anti-tuberculosis drugs belonging to the W-Beijing and non-Beijing genotypes in the pilot study. Strain nos: •=249/01; ▄=724/02; ▲=7004/01; ▀=722/02.

3.2 Reproducibility

After the pilot study, the growth of eight M. tuberculosis strains susceptible to anti-tuberculosis drugs and belonging to non-Beijing genotypes was tested at three different times (Fig. 2). Two of these strains had also been tested during the pilot study. Testing for differences between two means of the main continuous variable for eight susceptible M. tuberculosis strains belonging to non-Beijing genotypes showed no significant difference between measurements performed during the first and the second experiments (P=0.83), between the second and the third experiments (P=0.59) and between the first and the third experiments (P=0.64). The growth of nine pairs of M. tuberculosis strains obtained from the same patients before and after acquisition of resistance to additional drugs was tested twice (Table 2). Comparison of means of the main variable for nine pairs of strains obtained from the same patients revealed no statistical difference between measurements performed during the two experiments (P=0.94).

Figure 2

Comparison of time necessary to reach 200 units of growth for 8 susceptible M. tuberculosis strains belonging to non-Beijing genotypes measured during the first (▲), the second (▄) and the third (×) experiments. Time used to reach 200 units of growth of two susceptible strains measured during the pilot study (•) is indicated.

View this table:
Table 2

Comparison of the time necessary to reach 200 units of growth for 9 pairs of strains with different susceptibility patterns isolated from the same patients at different time

GenotypePatient no.First isolateSecond isolate
Resistance toMean time to 200 units of growth (95% CI)Resistance toMutation in the rpoB geneMean time to 200 units of growth (95% CI)
Non-BeijingPatient 1Susceptible127.2 (102.4–151.9)HR531 TCG→TTG156.4 (115.1–197.6)
Patient 2H130.3 (117.6–143.0)HR531 TCG→TTG149.1 (118.6–179.6)
Patient 3HS155.4 (75.4–235.4)HRS533 CTG→CCG185.9 (67.0–438.8)
Patient 4HS155.8 (65.6–246.0)HRS531 TCG→TTG174.3 (104.0–452.6)
W-BeijingPatient 5S191.3 (17.8–364.7)HRS516 GAC→GTC158.4 (68.8–247.9)
Patient 6S154.4 (69.2–378.0)HRSE526 CAC→CTC183.0 (139.8–226.2)
Patient 7HS183.2 (132.4–234.0)HRS531 TCG→TTG178.0 (47.1–308.9)
Patient 8HS198.7 (133.9–263.5)HRSE513 CAA→CTA189.7 (132.5–246.9)
Patient 9HSE161.9 (5.8–329.6)HRSE531 TCG→TGG149.5 (76.4–222.5)
  • H, isoniazid; S, streptomycin; R, rifampin; E, ethambutol.

3.3 Growth of non-Beijing strains

A total of 31 M. tuberculosis strains (11 susceptible and 20 resistant to at least one drug) of non-Beijing genotypes were analysed. The mean lag phase was 113.2 h (95% CI 103.9–122.5) for susceptible strains, and 138.0 h (95% CI 128.8–147.2) for drug-resistant strains. The observed difference was significant (F-statistic=13.7, P=0.001). The mean time necessary to reach 200 units of growth was 136.1 h (95% CI 126.9–145.3) for susceptible strains, and 162.6 h (95% CI 152.9–172.2) for drug-resistant strains (Fig. 3(a)). The observed difference was statistically significant (F-statistics=14.5, P=0.001) and indicated that resistance to anti-tuberculosis drugs resulted in a slower growth of the non-Beijing bacilli. In order to test for differences in growth rate, the time from the beginning of growth to 200 units of growth was determined. Drug-resistant strains (mean 24.6 h, 95% CI 23.4–25.7) were in average more likely to have slower growth rate than susceptible strains (mean 23.4 h, 95% CI 22.0–24.9), but the observed difference was not significant (F-statistics=1.6, P=0.21).

Figure 3

Time necessary to reach 200 units of growth for (a) 11 susceptible and 20 resistant M. tuberculosis strains belonging to non-Beijing genotypes and (b) 9 susceptible and 15 resistant M. tuberculosis strains belonging to the W-Beijing genotype.

The MDR strains had different mutations in the rpoB gene responsible for rifampin resistance (Table 3). Half of the strains harboured the 531 TCG→TTG mutation; the 516 GAC→GTC, 533 CTG→CCG and 526 CAC→TAC mutations were identified in three, two and one strain each. Significant difference in the mean time to reach 200 units of growth was observed for strains harbouring different mutations in the rpoB gene (F-statistics=8.7, P=0.01). The fastest growth was observed for strains with the 516 GAC→GTC mutation (144.6 h); strains with the 533 CTG→CCG mutation (200.6 h) showed the slowest growth; and strains with the 531 TCG→TTG mutation (165.9 h) revealed intermediate growth (Fig. 3(a), Table 3).

View this table:
Table 3

Time necessary to reach 200 units of growth according to the type of mutation in the rpoB gene

Mutation in rpoB geneNo. strainsMean time to 200 units of growth95% CI
Non-Beijing12167.6153.9–181.4
516 GAC→GTC3144.6128.9–160.3
526 CAC→TAC1181.2
531 TCG→TTG6165.9150.3–181.6
533 CTG→CCG2200.6155.4–245.7
W-Beijing15154.8143.4–166.3
513 CAA→CTA4179.0158.2–199.8
516 GAC→GGC1161.6
516 GAC→GTC1165.4
526 CAC→TAC1132.9
531 TCG→TGG1134.8
531 TCG→TTG7144.6130.9–158.2

3.4 Growth of the W-Beijing strains

The growth of 24 M. tuberculosis strains (9 susceptible and 15 resistant) of the W-Beijing genotype was analysed. The mean lag phase for susceptible strains was 119.0 h (95% CI 109.2–128.8), and 130.6 h (95% CI 118.9–142.3) for drug-resistant strains. The mean time to reach 200 units of growth was 143.9 h (95% CI 133.9–153.9) for susceptible strains compared to 154.8 h (95% CI 143.4–166.3) for drug-resistant strains (Fig. 3(b)), but the observed difference was not statistically significant (F-statistics=2.0, P=0.17). These strains also did not show significant difference in growth rate (F-statistics=0.5, P=0.50). The mean growth rate of susceptible strains was 24.9 h (95% CI 23.7–26.2); while the mean growth rate of drug-resistant strains was 24.2 h (95% CI 22.7–25.8).

Fig. 3 suggests that susceptible strains of the W-Beijing genotype (mean time to reach 200 units of growth 143.9 h) grow somewhat slower than susceptible strains belonging to non-Beijing genotypes (mean time to reach 200 units of growth 136.1 h), but the observed difference was not significant (F-statistics=1.7, P=0.21).

Mutations in the rpoB gene responsible for rifampin resistance of the W-Beijing strains showed polymorphism (Table 3). The majority (46.7%) of strains had the 531 TCGTTG mutation, four strains harboured the 513 CAA→CTA mutation. The 516 GAC→GGC, 516 GAC→GTC, 526 CAC→TAC and 531 TCG→TGG were found in one strain each. Strains with the predominant mutation 531 TCGTTG used 144.6 h as a mean to reach 200 units of growth, this was similar to the growth of susceptible strains belonging to the W-Beijing genotype (Fig. 3(b)). Drug-resistant strains with other mutations had slower growth (163.8 h), the observed difference was not significant (F-statistics=4.0, P=0.07).

3.5 Growth of the paired strains

The nine pairs of M. tuberculosis strains isolated from the same patients at different time period after the strains had acquired resistance to additional anti-tuberculosis drugs were analysed. In all nine cases, RFLP patterns of the first and second isolates were identical. All the second isolates were MDR having different mutations in the rpoB gene (Table 2). The main continuous variable was used to describe the growth of the paired strains. The lag phase of the paired strains was also analysed; however, this variable did not demonstrate difference. The growth parameters of the strains showed difference according to a strain genotype, susceptibility pattern and mutation in the rpoB gene. Genetic mutations in genes responsible for development of resistance to other drugs than rifampin that were not analysed, may also explain the observed differences in growth. In particular, strains belonging to non-Beijing genotypes revealed significantly slower growth after developing resistance to additional drugs (Table 2). Four of the five strains belonging to the W-Beijing genotype grew slightly faster after developing of drug resistance than did the first isolates of the same strains. One isolate of the W-Beijing genotype having 526 CAC→CTC mutation grew slower after developing resistance to additional three drugs.

4 Discussion

Clinical M. tuberculosis strains isolated in the Archangel oblast showed phenotypic and genotypic polymorphism, i.e. different resistance patterns, various mutations in the rpoB gene and different genotypes. These characteristics may influence the fitness of strains measured as growth. Mycobacterial populations may be composed of a mixture of different genotypes with different susceptibility to anti-tuberculosis drugs, and various metabolic rates and fitness characteristics resulting in more or less successful dissemination of particular genotypes. Variation in the biological fitness of the strains may have significant impact on the epidemiology of tuberculosis [29].

Laboratory experiments have provided evidence that drug-resistant strains have reduced fitness in comparison to susceptible strains. For example, M. tuberculosis strains resistant to streptomycin [14], isoniazid [15] and rifampin [16] grow slower that their wild-type parents. Newer studies have shown that drug-resistant mutants might have different degrees of reduced fitness. A study of induced rifampin resistance evaluated by viable cell counts on rifampin containing and rifampin free plates using the reference strain of M. tuberculosis H37Rv revealed that some drug-resistant mutants had only modest decrease in fitness resulting in bacilli survival in the environment without drugs [16]. The study of relative fitness using growth competition in vitro as measurement documented that fitness of genetically similar drug-resistant M. tuberculosis isolates differed significantly [30].

Our study was performed using clinical isolates of M. tuberculosis and incubation in the closed system BACTEC MGIT, which implied safer methodology excluding cultivation of M. tuberculosis strains on plates. Standardisation of inoculum was achieved by adjusting the suspension of M. tuberculosis to turbidity comparable to a McFarland No. 0.5 standard 1 ml of which contains 1.5×108 bacterial cells. The standard provides approximate estimate of number of bacterial cells in a liquid suspension. A standardised inoculum is obviously important since its variations may influence the growth rate. In our study, the experiment was performed three times using eight susceptible M. tuberculosis strains of non-Beijing genotypes, and no statistical differences in growth were revealed. Measurements of growth for nine pairs of M. tuberculosis strains that were performed twice showed no statistical difference. Some strains (216/01 and 736/02) showed variation in growth measurements. Individual variations of growth for these strains were almost equal to the difference found on average between susceptible and resistant non-Beijing strains. However, statistical analysis for all strains tested several times revealed no difference between three experiments, suggesting that the results were reproducible. In addition, all strains belonging to non-Beijing genotypes having identical mutations in the rpoB gene and belonging to the same RFLP cluster showed similar growth, further supporting the reproducibility of the method employed.

In our study, three different parameters (time necessary to reach 200 units of growth, the lag phase and the growth rate) showed relative importance and concordance of the results. All three parameters can be used as indicators of M. tuberculosis growth and fitness.

Strains susceptible to anti-tuberculosis drugs and belonging to non-Beijing genotypes were found to be the most fit strains; that is not always consistent with molecular epidemiological studies. Molecular epidemiological studies evaluate fitness of M. tuberculosis clinical isolates using frequency and size of clusters. Comparison of these indicators for drug-resistant and drug-susceptible strains indicates that drug-resistant strains should be more fit [12,13,17].

M. tuberculosis drug-resistant strains belonging to non-Beijing genotypes showed significantly reduced growth compared to drug-susceptible strains. The significance of the observed differences in growth parameters of non-Beijing strains and strains with different mutations in the rpoB gene can be argued as these strains were tested only once. However, three independent tests performed with the same eight strains suggested reproducibility. Thus, the observed difference in growth parameters should be valid. Drug-resistant strains were phenotypically and genotypically heterogeneous. MDR M. tuberculosis strains showed significant difference in the mean time to reach 200 units of growth depending on type of mutation in the rpoB gene. The number of strains examined was limited and the results must therefore be interpreted with caution. These strains are genetically heterogeneous, characterised by different RFLP patterns and may harbour various mutations in the katG, rpsL and embB genes responsible for the development of isoniazid, streptomycin and ethambutol resistance, respectively. Thus, possible genetic differences beyond the point mutations in the rpoB gene can explain the growth parameters.

The situation with the W-Beijing strains was different. The drug-resistant strains showed two tendencies in growth in comparison with the susceptible strains. Some drug-resistant strains of the W-Beijing genotype showed reduced growth compared to drug-sensitive strains, while other drug-resistant strains grew as well as susceptible strains. This difference between the W-Beijing and non-Beijing genotypes was also clearly demonstrated in experiments with nine pairs of M. tuberculosis isolates obtained from the same patients before and after MDR development. Four of the five pairs belonging to the W-Beijing genotype grew slightly faster after developing drug resistance; while the growth of one pair was reduced. All pairs of strains belonging to non-Beijing genotypes revealed significantly slower growth following MDR development.

M. tuberculosis strains of the W-Beijing genotype with the predominant 531 TCG→TTG mutation had growth parameters similar to susceptible strains indicating that drug resistance did not change the fitness in vitro of these strains. Strains having the 526 CAC→TAC and 531 TCG→TGG mutations grew faster than drug-susceptible strains. However, the number of strains examined was few and other mutations than that in the rpoB gene might also play a role in the difference of fitness. Mutations responsible for drug resistance development that decrease growth and initial fitness can be compensated by later mutations that restore bacteria's capacity for growth [13,31]. The degree of restoration of fitness by the compensatory mutations varies: in some cases restoration is complete, in others it can be only partial [32]. This can explain why some of the drug-resistant mutants of the W-Beijing genotype in our study had faster growth than susceptible strains; and some mutants had growth similar to susceptible strains.

RFLP patterns of rifampin resistant strains of the W-Beijing genotype were not identical and the strains might harbour various mutations in the katG, rpsL and embB genes responsible for the development of isoniazid, streptomycin and ethambutol resistance, respectively. Thus, heterogeneous growth of these strains can be explained by other differences among the W-Beijing strains.

Previously, we approached the evaluation of fitness of M. tuberculosis clinical strains using frequency of clustering and cluster sizes as an indicator of tuberculosis transmission and measure of adaptation [12,17]. The finding of relatively large clusters of drug-resistant M. tuberculosis strains and significant association between resistance and clustering indicated active transmission of drug-resistant strains. As a result, rates of drug resistance, especially MDR were high among new cases. The rate of transmission of the W-Beijing strains was higher than that of strains belonging to other genotypes, indicating their selective advantage [12,17].

The selective advantages and least physiological cost of rifampin resistance development may partly explain the spread of MDR tuberculosis throughout the world. Without these advantages, rifampin resistant strains would not be competitive towards drug-susceptible strains and the problem of primary MDR tuberculosis would be manageable by exclusion of failing drugs from the environment.

We hypothesise that selective advantages such as least physiological cost of rifampin resistance and compensatory mutations restoring fitness of M. tuberculosis may be responsible for the wide dissemination of the W-Beijing strains.

Acknowledgements

We thank Elisabet Rønnild, Anne Klem, Ann Christine Øvrevik, and Kjersti Haugum for skilful technical assistance. Funding was provided by grant 49711 from the Norwegian Ministry of Health and Social Affairs to P.S. and by the Norwegian Research Council grant 12083/730 to D.A.C.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
View Abstract