Estimation of the mutation rate of Mycobacterium tuberculosis in cases with recurrent tuberculosis using whole genome sequencing | Scientific Reports…

Studying M. tuberculosis latency in humans is harsh due to the difficulty of isolating the dormant bacteria, which is not possible until the active disease. Much has been published regarding latent TB and the percentages of reactivation and disease, but the latency data in patients who have already passed the disease have not been studied. Different approaches were used to mimic this process10,11,12. This work shows, for the first time, results obtained using isolates of patients with recurrent TB. Aragon, a region in the North of Spain, has a low incidence of TB. Thanks to the surveillance protocol carried out in this region since 2004, all M. tuberculosis isolates are genotyped and registered, allowing to trace the clinical TB history of the patients. Around 5% of the TB cases in our population correspond to recurrent TB. Of them, 89.8% were TB cases with isolates showing identical IS6110-RFLP patterns, indicating a potential relapse. Most of them (71%) were later considered as fail of treatment. In contrast, 10.2% of the patients had isolates with different genotypes, considered as reinfections. Among the total of TB cases in our community, reinfection occurs in 0.5% of the TB cases, reflecting that reinfection is uncommon among our population. These data are in agreement with a previous study in Madrid population, which showed an 87.5% of relapses and 12.5% of reinfections among the cases with recurrent TB13. However, in a study in the Canary Islands, the results showed a higher reinfection percentage (44%) versus the 55% of relapses14. A more extreme result was obtained in a study in London, in which 72.6% of the repeated patients were classified as reinfections against a 27.4% of relapses15. The large variation of the results among the different studies suggests that they largely depend on the population sample studied. It would be very interesting to analyse the reinfection cases in each of the studies to understand the reasons for these differences. Regarding endemic TB regions, a higher percentage of recurrent TB was found. Around 9.5% of TB patients had recurrent TB in Malawi (39.6% had relapse and 14.4% reinfection, the rest was undetermined)16 and a study carried out in India demonstrated that the majority of relapses they had were among HIV negative people (95% of TB recurrences) while the majority of reinfections were among HIV positive people (75% of TB recurrences)17.

Regarding the epidemiological and risk factors of the relapsed TB cases studied, we found that relapse was significantly earlier in HIV positive patients (in the first two years since the first episode) when compared to HIV negative patients (p value=0.041), what would be in accordance with a compromised immune system. We also found a trend that males suffered relapse earlier than females, which could be linked to other risk factors such as the use of IV drugs, smoking and the HIV status, which were more frequent in males in our study population. Any risk factor was found as significant for causing an earlier reactivation by Colangeli et al. 12, however they recognized that the clinical cases studied did not have in general any comorbidity.

The number of SNPs between the pairs ranged from 0 to 8. Remarkably, three among the 18 pairs had more than 5 SNPs between the first and the relapsed isolate, interpreted as not recent transmission18, even though the bacteria were isolated from the same patient. This could be related to clinical characteristics of the patients, as immunosuppression, HIV status or the treatment adherence. Surprisingly, several SNPs were found in the first isolates that were absent in the relapsed isolates, as if they had reverted. This phenomenon was extreme in P12, in which six out of the seven SNPs found were absent in the relapsed isolate. The explanation could be the presence of different clones in the patient19,20. In this way, in the different disease episodes a different clone was isolated, resulting from different bottlenecks and selective pressures of the original strain21,22. The reinfection with an identical strain has been described as a limitation of these kind of studies, but in our case, it can be discarded as only one of the pairs belonged to a large endemic cluster (P4, with 0 SNPs). The rest of the pairs were infected with orphan or small-outbreak strains of up to four cases, differently from other studies with large endemic clusters and high TB prevalence22.

Same as Colangeli et at.12, we did not find a significant correlation between the number of SNPs and the time between episodes. However, it is possible that P8 (160months between episodes and 0 SNPs) is altering the trend of SNP accumulation when the time between episodes increases. This is one limitation when working with small sample size, that a single point could have a great impact in the results. None of the SNPs found seemed related with recurrence as all were unique and therefore not common to more than one pair of isolates. It has been described that 0.5 SNPs per genome, per year is the standard mutation rate for M. tuberculosis10. Some studies, where multiple MDR/XDR isolates coming from the same patients were sequenced, have reported that selective pressure and antibiotic resistance can increase this mutation rate as high as more than 3 SNPs17,21. Despite all strains had been under the selective pressure of treatment, they did not achieve such a higher rate, maybe because they were drug susceptible. The mean mutation rate found in our study was 0.64 SNPs, slightly above the standard, due to the high mutation rate found in L4.1, almost double than the standard.

The correlation between the mutation rate and the relapsing period was found just marginally significant (p value=0.0613), differently to Colangeli et al.12, who found it significant. It is important to remark that the approaches were completely different: they used transmission events to mimic the latency period as the time between the diagnosis of the two cases, while we used isolates from the same patient who had a previous TB episode. We eliminated all patients with less than one year between the diagnosis of the episodes, as this was considered as a treatment failure, while Colangeli et al. 2020 had latency periods from one month, which was not possible in our clinical cases as a minimum of 6months of treatment was required. We did not find a significant correlation between the mutation rate along the variable generation times analysed when we split the data into [12years] and (214years), we observed just a small difference. This difference was much smaller than that found by Colangeli et al. 2020 (as high as 81010 for early latency), suggesting that mutation rate was constant during the relapsing period in recurrent TB cases. The mutation rate found in our study, 2.71010, was similar to that found by Ford et al. 2011 (21010)10, therefore both more distant from the one found by Colangeli et al. 2020. The reason why our results are similar to those of Ford et al. 2011 could be due to the similarity of the approaches applied, as they used lesions of the same macaques for studying latency and we used relapsed isolates from the same patients.

The analysis of the IS6110 element showed differences in the number of IS6110 copies in six of the pairs studied, affecting more than one IS copy in several pairs. It has been observed that IS6110 transposed more in great starvation conditions23, which could be similar to the conditions the mycobacteria found in the granuloma4. It was surprising that in four of the pairs studied, the relapsed isolates had lost 1 to 3 copies that were present in the first isolates. Noteworthy, the number of reads obtained in the fastQ files for these copies was considerably lower than for the rest of the IS copies. This suggests that those lost copies were not still fixed in the complete bacteria population, therefore a selection among the different clones present in the same patient had taken place24. It could be that the lost copies in the relapsed isolates had some deleterious effect for the mycobacteria as the relapsed bacteria were the ones without that IS copies. The fact that five out of the six pairs with IS6110 movements had more than 2years of relapsing period supports the idea of IS transposing more during the asymptomatic state of the patient23.

The main limitation to analyse the evolution of the bacteria during the dormancy period is the approach used for resembling this state. There is not a perfect approach, as it is impossible to reproduce what is happening inside the granuloma of a concrete patient, but we think that using isolates of the same patient is the closest way to do it. The difficulty to obtain the complete epidemiological information of the patients is another limitation because it does not allow to determine the accurate development of the diseases episodes. Another limitation is that some of the SNPs could be the result of a sequencing error or due to laboratory management, what would have a huge impact on the mutation rate. In addition, although there were more cases of potential relapses in our records, DNA of the isolates was not available. We decided not to re-cultivate these stored isolates to avoid more manipulation that could introduce errors such as additional SNPs that were not present in the original strains.

As a conclusion, the patients with HIV seemed to suffer reactivation in the first two years after the initial episode of TB more frequently than HIV negative patients. Besides, IS6110 movements occurred more frequently in patients with more than two years between episodes and it seems that different clones of the original strain could be responsible for the first and the following episodes. No correlation was found between the number of SNPs and the time between episodes, neither between the mutation rate and the relapsing period, just a trend of diminishing in longer time periods. Finally, the mutation rate seemed to be constant along all the period between episodes.

Here is the original post:
Estimation of the mutation rate of Mycobacterium tuberculosis in cases with recurrent tuberculosis using whole genome sequencing | Scientific Reports...