Detección, caracterización y análisis funcional de la complejidad clonal en la tuberculosis humana y bovina
Yurena Navarro García defended the PhD Thesis at the Faculty of Veterinary Medicine of the Complutense University of Madrid
December 17th, 2015
Genotyping tools in tuberculosis were initially developed for epidemiological studies. However, they have also made it possible to reveal the existence of clonally complex infections caused by Mycobacterium tuberculosis (MTB) and Mycobacterium bovis (M. bovis), thus calling into question the assumption that each episode of tuberculosis (TB) is caused by a single strain. Coinfection by more than one strain (mixed infection) or simultaneous presence of clonal variants (polyclonal infection) were described. Heterogeneous distribution of strains or clonal variants in different infected tissues (compartmentalized infection) has also been described.
However, there are few studies focused on these phenomena, and those available correspond to the description of anecdotal cases or to the analysis of these events in populations with high TB incidence. Thus, the first objective of this thesis focused on defining the magnitude in which clonally complex infections by MTB occur in an unselected population, in a context with moderate incidence, where the expectations of detecting this kind of infections are limited. By applying IS6110-RFLP and MIRU-VNTR analysis, clonally complex infections were detected in 11 patients with pulmonary TB (1.6%) and in 10 patients with pulmonary and extrapulmonary TB (14.1%). Nine out of these 21 cases were mixed infections and the remaining 12 corresponded to polyclonal infections. Finally, in 9 cases (5 patients with mixed infection and 4 patients with polyclonal infection) compartmentalization of the infection was documented.
Once the systematic description of clonally complex infections was done, we focused on the study of polyclonal infections. Specifically, to evaluate the potential functional meaning associated to the acquisition of subtle genetic rearrangements by microevolution phenomena.
First, we evaluated the potential in silico meaning of the variations in the number of IS6110 copies and in the number of repetitions in MIRU-VNTR loci. The analysis detected that some of these variations are located intragenically, which means either a potential impact on the structure and function of the encoded proteins, or even a loss of function when the modification truncated the reading frame of the gene. In addition, variations were also located in intergenic MIRU-VNTR loci, which could mean a potential effect on the expression of adjacent genes.
After the in silico analysis, we evaluated the effect on gene expression that the variability in the number of repetitions mapping in intergenic loci could cause. By applying quantitative RT-PCR on four pairs of related clonal variants, we showed that in three of the four analyzed loci, variations in 1 or 2 repetitions led to a subtle change in the expression of the adjacent genes.
Finally, evaluation of the functional significance of microevolution should include the study of the infectivity of the involved clonal variants. With the premise that subtle variations in infectivity might not be detected by applying standard infection models, we performed four different versions of in vitro infection models (based on: macrophages differentiated from the THP-1 human cell line, activated with or without IFN-γ; individual infections with each clonal variant and competitive coinfections), which allowed us to detect the effect that the occurance of subtle genetic rearrangements had on infectivity.
Once analyzed the proportion of clonally complex infections, and revealed the functional impact of microevolution, it was neccesary to study these events on a selection of patients, representative of different clinical circumstances. These studies allowed us to assess the way in which complex infections can interfere with diagnostic and therapeutic aspects, as well as to get information on the dynamics of the microevolution phenomena, that cannot be obtained from the descriptive population-based studies.
The first study corresponded to a patient with suspected MDR-TB who met difficulties on her diagnosis and treatment due to the involvement of a non-suspected complex infection. An integrative analysis putting together clonal analysis, molecular epidemiology strategies, clinical and epidemiological data, allowed us to reveal that the patient actually had a mixed infection with two MTB strains, one susceptible and the other MDR. The susceptible strain had been recently acquired, it was circulating in the epidemiological context of the patient, and coincided in our patient with the reactivation of a MDR strain, three years after exposure to the index case infected with the same strain. The MDR strain was underrepresented in the patient with mixed infection and it showed lower fitness than the susceptible strain.
The second case corresponded to a patient with two sequential TB episodes, as a result of reactivation, with different clonal variants isolated from each episode. This case was used as a model of analysis integrating all the strategies developed in this thesis. Thus, clonal variants were genotyped by IS6110-RFLP and MIRU-VNTR, showing variations in both markers. WGS analysis revealed the presence of SNPs between the variants. The integration of these data allowed us to define precisely the microevolution dynamics, showing that each variant had evolved independently from a common parental strain. Expression assays demonstrated that the acquired variability led to differences in the expression between variants. Finally, the systematic application of several models of infection identified a higher infectivity for the variant isolated in the second episode, as shown by applying a competitive infection model in Balb/c mice. Additionally, a new variant emerged by microevolution during an in vitro infection assay and this new variant had higher fitness than the parental strain. All data from the integrated study of this patient allowed us to consider a high ability to acquire variability for the involved strain. This suggests that the microevolution events, in addition to be modulated by clinical and epidemiological circumstances, also lie on bacterial factors.
The third case analyzed reinforced the hypothesis of the existence of a role in the microevolution events for bacterial factors, because involved a strain with a low tendency to acquire variability. This case corresponded to a patient persistently infected, along 8 years, with a Beijing strain that had been responsible for a large outbreak in the past. Prolonged infection in this case was a result of poor adherence to therapy, which meant an intermittent treatment. Surprisingly, despite the huge opportunities offered by this scenario for the detection of microevolution, no resistance mutations or any variations were identified by analyzing the standard genotypic markers. Even whole genome sequencing failed to identify SNPs throughout the infection period.
Once fulfilled the study of clonal complexity in human tuberculosis, we decided to transfer equivalent analytical strategies and the knowledge acquired to the study of clonal complex in bovine TB (bTB). Due to the delay experienced in the implementation of highly-discriminating genotyping strategies for the characterization of Mycobacterium bovis (M. bovis) isolates, the knowledge gaps in relation to clonal complexity were wide.
The experience with MTB indicated that MIRU-VNTR offers the best choice to an accurate identification of polyclonal and mixed infections, and therefore, equivalent studies in M. bovis should apply it. Our first objective was to adapt and optimize this technique to the characterization of M. bovis. We developed a multiplex PCR followed by an analysis of PCR products by capillary electrophoresis and automatic alellic assignment. A panel of 9 loci, suitable to offer a high discrimination, was selected base don data from a study of molecular epidemiology in bTB performed in our country. The application of our new methodology on a sample of M. bovis isolates showed its efficiency and discriminative power.
When the methodology to adequately address the phenomenon of clonal complexity in M. bovis infections was optimized, we decided to perform a first analysis by focusing on the most extreme and uncommon version of clonal complexity, the compartmentalized infection. All the animals in a population with M. bovis isolation from two or more different anatomical sites were selected, and their spoligotypes were compared. Six (10.9%) of them were infected by different strains. The additional analysis by MIRU-VNTR confirmed that the compartmentalization was strict, with a single strain at each location. Further analysis of infected animals in the same herds allowed us to trace the presence, in independent animals, of the strains involved in the compartmentalization, indicating that it was caused by overexposure.
Finally, the existence of farms chronically infected with M. bovis offered the opportunity to address a chronological study of the dynamics of microevolution, an issue that cannot be easily addressed in human TB. This study could be considered as a model for studying the variability that can be acquired by microevolution. Eight farms that had been infected along more than one year by the same strain were selected. The MIRU-VNTR analysis of sequential isolates obtained during the infection revealed the existence of microevolution in half of the farms studied. Microevolution involved different lineages and different MIRU-VNTR loci. We observed that the emergence of clonal variants did not depend on the length of the infection and on the total number of infected animals. Furthermore, it was observed how the clonal variants, generated by microevolution, replaced the initial strain in 2 out of 3 farms, suggesting a higher infective ability for these emerged variants.
In summary, this thesis means an advance regarding the analysis of clonal complexity in human and animal TB; it developes metodology and analytical strategies to optimize its detection and characterization and to allow a more in depth knowledge of its functional meaning. All these aspects support a new knowledge status which will facilitate the progress in this field.
