**3.2.2 Recurrent tuberculosis: Relapse or reinfection?**

The frequency and determinants of exogenous reinfection and of endogenous reactivation of TB in patients previously treated are poorly understood. The importance of reinfection as a cause for recurrence of TB is unclear and has potential public-health implications. Different studies have used IS*6110* genotyping to answer this question. The possibility of genotyping the isolates from initial and recurrent disease episodes allows to differentiate an episode of reinfection from that of relapse of TB.

At this respect, differences are shown depending on the incidences of TB and of the HIV status of the patients. In Spain, a country with a low incidence rate of TB, two studies on this issue were conducted. In the Gran Canaria Island, 2.4% of the cases had recurrent TB in a 5 years-period. Up to 44% of them corresponded to exogenous reinfection proved by IS*6110* genotypes (Caminero et al., 2001). In a second study conducted in Madrid extended twelve years, up to 3.1% of the patients had a second episode of TB. Only one recurrent case showed different genotypes, suggesting exogenous re-infection. Re-infection is possible among people in low-risk areas, but the rates are lower than those occurring in high-risk areas (Cacho et al., 2007). On the other hand, in countries with high incidence as India, most of the recurrences after successful treatment of TB are due to exogenous reinfection in HIVinfected persons, in contrast to endogenous reactivation in HIV-uninfected persons. Strategies for prevention and treatment of TB infection must take these findings into consideration (Narayanan et al., 2010). Conversely, one study carried out in Karinga Malawi, concluded that HIV increases the rate of recurrent TB by increasing the rate of reinfection disease (Crampin et al., 2010). Other authors reviewed different studies on recurrence and argued that, apart from extreme situations, the problem of recurrence due to reinfection has few implications for TB-control programmes (Lambert et al., 2003).

#### **3.2.3 Limits of IS***6110* **as epidemiological tool**

A common dilemma of the different markers used for typing tuberculosis, including IS*6110*, is how to interpret the variability of the patterns. If two *M. tuberculosis* isolates from 2

IS*6110* the Double-Edged Passenger 69

All those changes could be a risky to the bacteria's genomes integrity, being the carriage of mobile IS either a potential enemy with deadly influence on the bacterial fitness or a helpful ally contributing to the improvement of that fitness. Our current knowledge on how the IS*6110*-mediated mutations influence in the genome plasticity of the *M. tuberculosis* genome

The numerous studies published on IS*6110*-RFLP with epidemiological purposes showed a high level of variability in the locations of this IS along the *M. tuberculosis* genome (see part 3.2). On the basis of those results the rate of transposition of IS*6110* was estimated to be about 18% over a period of 5-6 years. However it seems evident that the events of transposition are related to changes in the environment in which the bacteria are involved. It was suggested that transpositional events occur following mutational burst instead of following a constant mutation rate; this can explain the observation that changes in RFLP patterns would occurred more frequently during transmission and before diagnosis (soon after the bacilli enter inside the host) or after relapses or any other main event during the course of the infection (Schürch et al., 2010). In agreement to this consideration, two rather different half-life times were calculated for the IS*6110*-RFLP patterns stability in serial patient's isolates: 0.6 and 10.7 years; this most probably be due to changes in the patient's management or to the course of the infection in the different settings compared (Schürch et

Independently of why, how or when its transpositions occurred, IS*6110* mediates genome plasticity of members of the MTBC, and that plasticity is ongoing both under controlled

To confirm the last assertion, some papers described changes in the RFLP pattern during infection. This is showing that microevolution of the bacilli mediated by IS could occur not only during transmission between patients but also during the course of the disease in a single patient (Al-Hajoj et al., 2010). Besides, the comparison of the whole-genomes of six different H37Rv strains, collected from several laboratories, showed that multiple IS*6110* transposition events have occurred in the genome even under *in vitro* "controlled"

Since late nineties, several methods have been applied to identify and sequence the loci in which the IS was integrated inside the genome. The methods applied for the identification and sequence of the flaking-regions included cloning of the agarose-excised hybridizing bands (Beggs et al., 2000); reverse dot blot assay (Steinlein & Crawford, 2001); wholegenome microarrays (Kivi et al., 2002), ligation-mediated PCR (Otal et al., 2008) and construction of BACs libraries (Alonso et al., 2011) among others. All these procedures are usually cumbersome and show difficulties to detect all the insertions present, particularly in

The development of high throughput whole-genome sequencing procedures has allowed the overcome of some of those difficulties, however this procedure is so far not of general

environment *in vitro* and during infection *in vivo* (Fang et al., 1999b).

will be reviewed herewith.

al., 2010).

**4.1 Moving along the genome** 

environments (Ioerger et al., 2010).

**4.1.1 How to identify the IS***6110* **insertion sites** 

those strains carrying high IS*6110* copy number.

different patients present the same genotype, transmission may have occurred between them. However, once transmission has occurred, the genotypes may change, resulting in divergent fingerprints. The advantage of IS*6110* as marker is that the clock of change of the IS*6110* patterns was determined in serial isolates; the half-life was extrapolated to be 3.2 years. These changes were predicted more common for persons with extrapulmonary disease and for those who had both pulmonary and extrapulmonary isolates. This fact supported the use of IS*6110* typing in epidemiologic studies of recent transmission of TB (de Boer et al., 1999). The results of a study carried out to estimate the recent transmission based on IS*6110*-RFLP suggested that the interpretation of the recent transmission index, and the resulting necessary public health interventions, will vary according to how researchers account for spontaneous mutation when estimating transmission from the genotyping data (Benedetti et al., 2010).

In spite of all the studies carried out with this genomic element, some limitations have been found. Besides the technical difficulties that IS*6110* typing presents for some laboratories (the long time that the mycobacteria requires to growth, the equipment and the software required for the analysis), this method have also demonstrated difficulties for differentiating LCS*,* including *M. bovis* strains and is unable to identify strains with cero copies. Some studies have solved this problem by applying a second technique for these cases (Thong-On et al., 2010). Other studies with high prevalence of strains with LCS do not recommend this technique in their settings (Asgharzadeh et al., 2011). Mixed infections represent another limitation, which could be underestimated using IS*6110*-RFLP and could be confused with exogenous reinfection (Shamputa et al., 2006). The mixed tuberculosis infection suspected as a result of the IS*6110-*RFLP method could be clearly identified by MIRU-VNTR typing, which is more sensitive for the detection of multiple *M. tuberculosis* strains (Allix et al., 2004).
