Abstract
The human gastrointestinal tract (GIT) is inhabited by a vast number of
microorganisms collectively called gut microbiota (GM). Among many functions assigned to the GM, its ability to stimulate and develop the host’s immune system has become a subject of intensive studies of many research groups worldwide. Recent studies show that imbalances in the microbial composition known as dysbiosis can cause irregularities in the function of the immune system. This in consequence has been demonstrated to have
indirect implications on development of multiple inflammatory diseases including: diabetes (type 1 and 2), obesity, inflammatory bowel disease (IBD), eczema, atherosclerosis (ATS), or rheumatic arthritis (RA). Grasping the complex relation between bacterial and immune disease markers into a mathematical model would be of great prophylactic and diagnostic value in treatment and prevention of many autoimmune diseases. This work presents the current knowledge on optimized methods for GM screening, data analysis tools, furthermore it presents approaches to tackle the challenging task of mathematical model development for the correlation of GM composition with
inflammatory disease parameters based on animal models.
In order to evaluate the usefulness of a murine model for accomplishing this
venture, a meta-analysis comparing GM consortia between humans and laboratory mice strains was conducted. More than 1.5 million high quality tag-encoded partial 16S rRNA gene sequences determined with pyro-sequencing reflecting the GM composition of 32 human and 88 mouse samples were used. We reported a high level of quantitative differences between the GM composition of mice and humans, but at the same time we were able to present a large qualitative similarity core, which in conclusion vindicated
mice as a human experimental model.
An additional task of this thesis was to develop a fast screening method and to investigate the distribution of two bacterial species namely: Akkermansia muciniphila and Candidatus Savagella in detail. These two members of the gut microbial community were previously reported, including our own study, to be directly linked with the host’s health/disease status. In the same manuscript we also demonstrated an alternative method of constructing DNA standards for uncultivable bacteria by cloning 16S rRNA gene into a
competent E.coli strai.
A sizeable part of this thesis was devoted to establish an optimal window of time capturing the crosstalk between the GM and inflammatory parameters. We demonstrated that both C-section and cross-fostering with a genetically distinct mouse strain influence the GM composition and immune markers in mice, and that this period during early life is particularly crucial for the first encounter with microorganisms, not only for the development of final GM composition, but also for the maturation of the immune system. In light of this work we hypothesized that in order to find a link between GM and disease markers an attempt to conjoin the events of early life with those taking place in the adulthood is needed.
Consequently, the following study investigated the correlation between the early and adult GM and immune parameters of non-obese diabetic mice (NOD) in order to select alleged bacterial markers of Type 1 diabetes (T1D). The GM composition was analyzed with pyro-sequencing and correlated with immune cell populations measured at the onset of diabetes and in non-diabetic mice at 30 weeks of age. With this study we have proposed a model in which four bacterial taxa act in favor of diabetes protection while two others
promote pathogenesis.
In the last two manuscripts presented here we have inspected ways of introducing dysbiosis using antibiotic treatment and diet, to subsequently evaluate potential impacts on the immune system and diabetes development. Treatment of NOD mice with vancomycin during infancy and adulthood resulted in dramatic changes in relative distribution of bacterial taxa in both groups, with the dominance of A. muciniphila to up to 90%.
However, only neonatally treated mice had a significantly lower cumulative diabetes incidence.We also tested how a gluten-free diet, known to protect against T1D, modulated the bacterial composition in a way that attenuates the inflammatory state. Pregnant NOD mice were fed a gluten-free or standard gluten-containing diet, but all pups were weaned to the standard diet only. The microbial composition of pups was verified after weaning to be different between the two categories and was clearly characterized by their high resemblance to the mothers. Interestingly, even though those differences have faded away with time, the development of diabetes was shown to be significantly delayed in the gluten-free born group.
A comprehensive understanding of the challenges associated with the modeling of unique biological interactions, including strengths and weaknesses of state of the art methods used to fulfill this task, as well as future prospects of such models has been presented in this thesis. The current work is the first step towards developing a full-blown mathematical model with a predictive value of host fitness based on subsampling bidirectional interactions between GM and inflammatory disease parameters, with a special emphasis on T1D.
microorganisms collectively called gut microbiota (GM). Among many functions assigned to the GM, its ability to stimulate and develop the host’s immune system has become a subject of intensive studies of many research groups worldwide. Recent studies show that imbalances in the microbial composition known as dysbiosis can cause irregularities in the function of the immune system. This in consequence has been demonstrated to have
indirect implications on development of multiple inflammatory diseases including: diabetes (type 1 and 2), obesity, inflammatory bowel disease (IBD), eczema, atherosclerosis (ATS), or rheumatic arthritis (RA). Grasping the complex relation between bacterial and immune disease markers into a mathematical model would be of great prophylactic and diagnostic value in treatment and prevention of many autoimmune diseases. This work presents the current knowledge on optimized methods for GM screening, data analysis tools, furthermore it presents approaches to tackle the challenging task of mathematical model development for the correlation of GM composition with
inflammatory disease parameters based on animal models.
In order to evaluate the usefulness of a murine model for accomplishing this
venture, a meta-analysis comparing GM consortia between humans and laboratory mice strains was conducted. More than 1.5 million high quality tag-encoded partial 16S rRNA gene sequences determined with pyro-sequencing reflecting the GM composition of 32 human and 88 mouse samples were used. We reported a high level of quantitative differences between the GM composition of mice and humans, but at the same time we were able to present a large qualitative similarity core, which in conclusion vindicated
mice as a human experimental model.
An additional task of this thesis was to develop a fast screening method and to investigate the distribution of two bacterial species namely: Akkermansia muciniphila and Candidatus Savagella in detail. These two members of the gut microbial community were previously reported, including our own study, to be directly linked with the host’s health/disease status. In the same manuscript we also demonstrated an alternative method of constructing DNA standards for uncultivable bacteria by cloning 16S rRNA gene into a
competent E.coli strai.
A sizeable part of this thesis was devoted to establish an optimal window of time capturing the crosstalk between the GM and inflammatory parameters. We demonstrated that both C-section and cross-fostering with a genetically distinct mouse strain influence the GM composition and immune markers in mice, and that this period during early life is particularly crucial for the first encounter with microorganisms, not only for the development of final GM composition, but also for the maturation of the immune system. In light of this work we hypothesized that in order to find a link between GM and disease markers an attempt to conjoin the events of early life with those taking place in the adulthood is needed.
Consequently, the following study investigated the correlation between the early and adult GM and immune parameters of non-obese diabetic mice (NOD) in order to select alleged bacterial markers of Type 1 diabetes (T1D). The GM composition was analyzed with pyro-sequencing and correlated with immune cell populations measured at the onset of diabetes and in non-diabetic mice at 30 weeks of age. With this study we have proposed a model in which four bacterial taxa act in favor of diabetes protection while two others
promote pathogenesis.
In the last two manuscripts presented here we have inspected ways of introducing dysbiosis using antibiotic treatment and diet, to subsequently evaluate potential impacts on the immune system and diabetes development. Treatment of NOD mice with vancomycin during infancy and adulthood resulted in dramatic changes in relative distribution of bacterial taxa in both groups, with the dominance of A. muciniphila to up to 90%.
However, only neonatally treated mice had a significantly lower cumulative diabetes incidence.We also tested how a gluten-free diet, known to protect against T1D, modulated the bacterial composition in a way that attenuates the inflammatory state. Pregnant NOD mice were fed a gluten-free or standard gluten-containing diet, but all pups were weaned to the standard diet only. The microbial composition of pups was verified after weaning to be different between the two categories and was clearly characterized by their high resemblance to the mothers. Interestingly, even though those differences have faded away with time, the development of diabetes was shown to be significantly delayed in the gluten-free born group.
A comprehensive understanding of the challenges associated with the modeling of unique biological interactions, including strengths and weaknesses of state of the art methods used to fulfill this task, as well as future prospects of such models has been presented in this thesis. The current work is the first step towards developing a full-blown mathematical model with a predictive value of host fitness based on subsampling bidirectional interactions between GM and inflammatory disease parameters, with a special emphasis on T1D.
| Originalsprog | Engelsk |
|---|
| Forlag | Department of Food Science, Faculty of Science, University of Copenhagen |
|---|---|
| Antal sider | 170 |
| Status | Udgivet - 2014 |