Skip to main content
European Commission logo print header

Insights from within-host dynamics on the coexistence of antibiotic resistant and sensitive pathogens

Periodic Reporting for period 1 - PolyPath (Insights from within-host dynamics on the coexistence of antibiotic resistant and sensitive pathogens)

Período documentado: 2019-11-01 hasta 2021-10-31

The project PolyPath aimed at understanding the coexistence-maintaining mechanisms of antimicrobial resistant and sensitive strains within a pathogen by accounting for randomness on the scale of within-host bacterial growth and the scale of between-host disease transmission.

The new insights that are generated both on within-host dynamics of antimicrobial resistance dynamics as well as their impact on the between-host transmission of resistant strains are useful to tackle the societal problem of antibiotic drug resistance of multiple pathogens. In addition, the results on within-host dynamics might help identify favorable treatment strategies to reduce the risk of resistance evolution during antibiotic treatment.

Overall, the scientific objectives were as follows:
(1) Understanding the within-host dynamics of the pathogen under antibiotic treatment and during bacterial colonization of uninfected hosts.
(2) Resolving the problem of coexistence of antibiotic sensitive and resistant strains in a single pathogen by coupling within- and between-host dynamics to account for selective advantages of strains on different scales.
(3) Simulating the within-between-host model and confronting it to epidemiological datasets.

In view of the ongoing COVID-19 pandemic, the focus of the project shifted in March 2020 (4 months into the funding period) to address questions related to the propagation of COVID-19. The investigations that were conducted were diverse and ranged from the viral replication mechanism within an infected individual to transmission dynamics of this novel disease in the population. Specifically, I have worked on the following objectives:

(A) Assessing the potential of prophylactic antiviral treatment as a tool to prevent infection of people at high risk, e.g. health care workers.
(B) Analysis of an epidemiological model that accounts for a time-varying infectiousness of infected individuals over the course of their infection.
(C) Study of early epidemic dynamics and public health related consequences, e.g. an estimate for the minimal testing frequency in a population to detect infection clusters before they exceed a certain size.

These research questions provided important information on the potential of repurposing existing antiviral drugs as a potential treatment against COVID-19 (objective (A)) and an accurate description of infection cluster dynamics, which can be used to inform policy making with respect to COVID-19 containment and detection strategies (objective (C)).
Objective 1: I have studied the dynamics of antibiotic resistant and sensitive strains within a host during and after antibiotic treatment. I was particularly interested whether different types of antibiotics (biostatic antibiotics inhibit cell growth, biocidic antibiotics actively kill bacterial cells) have a different effect on the process of resistance evolution and proliferation. It turns out that they indeed affect the process differently, e.g. the carriage time of resistant bacteria in a host after treatment is much increased by treatment with bactericidal antibiotics. This work is still ongoing and I aim to finish a manuscript by the end of Summer 2021.

Objectives 2 and 3: Due to the shift in research questions, I did not have the time to work on these objectives.

(A) We studied a within-host dynamical model of virus dynamics and estimated the effect of prophylactic treatment with different types of antiviral drugs that differ in their mechanism of action, e.g. antivirals that attack the virus before it has infected susceptible target cells or antivirals that inhibit the production of new virus particles in infected target cells. We computed the efficacy for the different drug types to completely inhibit the establishment of a viral infection. Even if infection cannot be prevented entirely, prophylactic treatment can still delay infection establishment and by that "flatten the viral load curve" within hosts. The results of this work are published in the scientific journal PLoS Computational Biology.

(B) We have developed an epidemiological model that reflects the disease progression of COVID-19 in a population. Specifically, we accounted for a transmission rate that varies over the age of infection of infected individuals. With this model, we then estimated the effective reproduction number in France in Spring 2020. A manuscript summarising the results is published on the arXiv preprint server.

(C) We modeled the early epidemic dynamics of a local infection cluster. With this model, a special case of the model developed in (B), we for example estimated that the SARS-CoV-2 variant of concern B.1.1.7 detected (retrospectively) in September 2020 in the UK first appeared in the UK early August 2020. Last, we estimated a minimal testing frequency to detect clusters before they exceed a certain threshold size.

As a result of my COVID-19-related work, I was invited to seminars at the Frankfurt Institute of Advanced Sciences to talk about project (A), at the University of Helsinki and to the Biohasard workshop in Grenoble to talk about project (C) (all online seminars).
The methods I use to investigate the question of the emergence of antimicrobial resistance in response to different forms of treatment are state of the art tools of stochastic processes (objective 1). The application of these tools to the question of antimicrobial resistance are new, which is why I expect to obtain new insights and more detailed answers about this particular question. Since antimicrobial resistance is a global health problem that is already responsible for nearly a million deaths per year, any scientific advance that increases our understanding about the emergence and maintenance of antimicrobial resistant strains is valuable information that can be used to optimize treatment recommendations.

Early during the COVID-19 pandemic, we aimed to assess the potential of repurposed existing antiviral treatments to prevent infection by SARS-CoV-2 (project A). While we could contribute to the information about SARS-CoV-2 dynamics within infected individuals, unfortunately our analysis, which was one of the first analyses about antiviral treatment effects (first preprint in May 2020), showed that the efficacy threshold for antiviral treatments is at around 90% to prevent infection with SARS-CoV-2. This efficacy threshold is not reached by any existing antiviral treatment against other viral infections. Still, the more general results about within-host viral dynamics of SARS-CoV-2 have provided insights about the large variability of viral loads in infected individuals and numerical estimates of within-host parameters.

The analysis of the early dynamics of local outbreaks (project C) is in contrast to the country- or region-wide analyses performed by most epidemiological researchers. Our approach was novel because it combined two well-studied processes: (i) the probability of a local epidemic outbreak and (ii) the description of the epidemic size. With this new approach, we for example estimated that the variant of concern B.1.1.7 first detected in the UK in September 2020, appeared in the UK in early August 2020. Our results can also be used to estimate the minimal testing frequency in a population to detect clusters before they exceed a certain size.
(A) Within-host model of viral dynamics with different types of antiviral treatments.
(C) No. of infections until detection of an infection cluster depending on the testing frequency.