Study of the Ecological and Evolutionary Dynamics of Escherichia Coli in the Digestive Commensal Flora

Escherichia Coli poses a major public health problem. E. Coli is not only a commensal of the digestive tract but also a major opportunistic pathogen, first cause of urinary tract infections, first cause of bacteremia. However, little is known about the dynamics of intestinal colonization of the human host. Understanding the dynamics of colonization is crucial because selection for the major traits of E. coli is antibiotic resistance and virulence (the propensity to cause infection) works in commensalism, in the gut, not in infections. This study will make possible for the first time to study the colonization dynamics of E. coli in a large healthy host population. In particular, we will quantify how the loss or gain of certain strains in the intestine depends on the properties of the host and the bacteria. This study will thus provide a better understanding of the E. coli's epidemiological dynamics and the development of certain traits such as antibiotic resistance.

Main objective and criterion The main objective of this research is to quantify the succession (gain, loss or replacement) of E. coli strains in the gut microbiota in healthy volunteers.

Secondary objectives and criteria The secondary objective is to study the intra-host evolution of E. coli by focusing on the evolution of clones colonizing the same host for long periods of time. This makes possible to understand whether the intra-host evolution is dominated by mutations conferring a selective advantage (such as resistance to antibiotics) or on the contrary neutral mutations (having no effect on the proteins produced).

Population The two phases of this study will be conducted in healthy adults of both sexes. Healthy volunteers will be recruited from the file of healthy volunteers for biomedical research at the Clinical Investigation Center (CIC) of the Bichat - Claude Bernard hospital group.

Practical procedure This research will take place in two successive phases: A first so-called pilot phase will first be conducted with healthy volunteers. During this pilot phase, 50 healthy volunteers will be selected and followed every 2 weeks for a stool sample. The pilot phase will include 10 visits for a total duration of 18 weeks per volunteer. The strains isolated during this pilot phase will be sequenced, making possible to characterize the strains present and the possible changes of strains between two stool samples. Succession rates will be estimated. The results of this phase identify the optimal sampling rate that maximizes accuracy in estimating succession rates. This will be retained for the second phase.

The participation of the volunteers included in the pilot study will be broken down into 2 stages :

1. a stage of volunteers'selection at the CIC: each volunteer will have a clinical examination to ensure their eligibility for participation in the research. He must meet all the selection criteria on the day of this visit and sign the collection consent;
2. an 18-week follow-up period during which the volunteers go to the CIC to deposit stool samples every two weeks for 18 weeks (corresponding to 10 visits).

A second phase, a prospective cohort study will also be conducted in healthy volunteers. During this second phase, healthy volunteers who participated in the first phase will be able to continue their participation and new volunteers will be selected.The healthy volunteers included in the prospective cohort study will be followed up with visits and stool samples which will be defined according to the results of the pilot phase for a period of 8 to 144 weeks maximum.

The participation of volunteers included in the cohort study will be broken down into 2 stages:

1. a stage for the selection of volunteers at the CIC (V1): all participants will have a clinical examination to ensure their eligibility for participation in the research. They must meet all the inclusion criteria available on the day of this visit and sign the collection consent.
2. a follow-up period of 144 weeks maximum: 8 follow-up visits with a rhythm of 1 to 18 weeks, ie 8 to 144 weeks max. The volunteers will go to the CIC to deposit stool samples at each of the 8 planned visits. The schedule for these 8 visits will be determined based on the results of the pilot phase.

Number of subjects selected First phase: 50 subjects Second phase: 200 subjects, ie 0 to 50 from the pilot phase + 150 to 200 new subjects, depending on the number of volunteers in the pilot phase who wish to continue.

Number of centers Monocentric national study: Clinical Investigation Center of the Bichat - Claude Bernard hospital group.

Research calendar

1. First phase:

• duration of inclusion: 6 months (26 weeks)
• duration of participation (treatment + follow-up): 18 weeks
• total duration: 44 weeks

1. Second phase:

• duration of inclusion: 18 months (78 weeks)
• duration of participation (treatment + follow-up): 8 to 144 weeks
• total duration: 86 to 222 weeks

Statistical analysis To estimate the rates of gain, loss and replacement of E. coli strains depending on host and bacterial factors, statistical analyzes will be performed using R software within the French Institute for Health and Medical Research U722: first on the 50 volunteers in the pilot phase, which will allow a first estimate of these rates and define thus the optimal sampling schedule for the prospective cohort phase, then on the 200 volunteers of this second phase.

The statistical model used for the two analyzes will describe the dynamics of host colonization. We will assume that the host colonization state follows a discrete-state Markov chain in continuous time. This mathematical model describes the transitions between the different possible states, by strain gain, strain loss, and strain replacement within the host. It calculates the probability of the data obtained knowing the parameters (strain gains, strain loss, strain replacement) and therefore the likelihood (the probability of the data knowing the model). From the likelihood, the parameters can be estimated by maximum likelihood methods.

Quantitatively, the parameters will be estimated by maximum likelihood and the precision will be quantified by confidence intervals obtained by the likelihood profile method.

The dependence of the gain, loss and replacement rates on several host or bacterial variables will also be estimated. The significance of the effect of a variable on the gain, loss and replacement rates will be tested by model comparison: the models will be adjusted by maximum likelihood. The comparison of models will be done by likelihood ratio test.

First, the factors that are suspected a priori to be important for transition rates will be tested. The level of significance of these factors will be corrected for multiple tests (Bonferroni correction). In a second step, the impact of all the elements of the genome will be analyzed in an exploratory way (genome-wide association study).