Modeling species interaction in the gut microbiota using data from patients

Joao Xavier

Program for Computational and Systems Biology

Sloan Kettering institute for Cancer Research

“Mathematical models to study the biology of pathogens and the infectious diseases they cause”

Joao B Xavier, Jonathan M Monk, Saugat Poudel, Charles J Norsigian, Anand V Sastry, Chen Liao, Jose Bento, Marc A Suchard, Mario L Arrieta-Ortiz, Eliza JR Peterson, Nitin S Baliga, Thomas Stoeger, Felicia Ruffin, Reese AK Richardson, Catherine A Gao, Thomas D Horvath, Anthony M Haag, Qinglong Wu, Tor Savidge, Michael R Yeaman

iScience, 2022

Feces of hospitalized patients: non-invasive samples to study microbiome ecology

MSKCC has >200 bone marrow transplants each year�

>10,000 samples from >1,000 patients �map patients’ microbiome states

Data compiled in

Liao et al (2021) Scientific Data

What did we learn? �Loss of diversity decreases survival

Peled et al. N Engl J Med 2020

A mirage of data?

Problems:

• Steep learning curve to analyze the data limits reuse by others.
• Data submitted to SRA in batches makes data hard to navigate.
• Data tables produced from processing sequencing data are often more useful to researchers than the sequencing data.
• Clinical metadata distributed across supplemental materials of different papers makes to integrate data.

Reverse translational research: �clinical data for basic research

Randomized trial tested efficacy of �auto-FMT after allo-HCT

Taur et al Science Translational Medicine 2018

Auto-FMT reconstitutes gut microbiota�to patients after allo-HCT

Taur et al Science Translational Medicine 2018

Pilot randomized trial �had 10 control and 15 treated patients

Taur et al Science Translational Medicine 2018

Each patient—treated or control—had �unique white blood cell dynamics

But on average auto-FMT seemed to boost blood levels of neutrophils, lymphocytes and monocytes

Schluter et al Nature 2020

Does the gut mi

Expanding ecological inference to study the impact of gut microbiome on the immune system dynamics

Generalized

Lotka-Volterra

Equation:

Schluter et al Nature 2020

The response variable is the white blood cell dynamics�The covariates are drugs, microbiome and others

Schluter et al Nature 2020

Inference quantifies impact of immunomodulatory drug (G-CSF) and of white blood cells on themselves

mycophenolate mofetil

Cetirizine

GCSF

Schluter et al Nature 2020

`

Schluter et al Nature 2020

Faecalibacterium, Ruminococcus associated with better responses to anti–PD-1 immunotherapy of melanoma

Effect of gut microbes smaller than some drugs (G-CSF) �but could accumulate over time

Schluter et al Nature 2020

Microbiome influences immune dynamics in humans

• Gut microbes influence immunity: widely accepted but no evidence in humans.
• Bone marrow transplantation is the strongest deliberate perturbation to the immune system.
• Antibiotic prophylaxis perturbs the microbiome too.
• Joint recovery of immune and microbiome enable study of the interaction between the two.
• Effect of gut microbes smaller than some drugs (G-CSF) but could accumulate over time.
• Microbiome engineering could potentiate immune modulatory therapies (cancer immunotherapy).

​

Xavier lab

Vanni Bucci (now at U Mass Medical)

Katharine Coyte (now at U Manchester)

Jonas Schluter (now at NYU Medical)

Jinyuan Yan

Bradford Taylor

Ana Djukovic

Chen Liao

Wendy Lin

U01 AI124275 (Pamer/Xavier)

R01 AI137269 (Xavier/Taur)

Collaborators

Eric Pamer (now at U Chicago)

Eric Littmann

Ying Taur

Marcel van den Brink

Tsoni Peled

Itsik Pe’er (Columbia)

Castori Center

Emily Fontana

Luigi Ammoretti

Roberta Wright