Clinical and Public Health Microbiology

NIH Workshop:
The Human Microbiome – Conference Round-Up for August 18, 2017

CONFERENCE ROUND-UP  August 18, 2017

The Human Microbiome: Emerging Themes at the Horizon of the 21st Century

Microbes inhabit just about every part of the human body, outnumbering human cells by ten to one. The ten-year, National Institutes of Health (NIH) Common Fund Human Microbiome Project was established to understand how microbial communities impact human development, physiology, immunity, brain development, and behavior, and to create research resources for this emerging field.

This 2017 NIH-wide microbiome workshop was organized by a planning committee of the trans-NIH Microbiome Working Group (TMWG)1, which includes program staff from the 19 NIH Institutes, Centers and Offices that support human microbiome research through their extramural portfolios. The TMWG is interested in taking stock of where the microbiome field stands after NIH’s ten-year investment in the Human Microbiome Project and evaluating what is needed for this field to advance over the next decade. This meeting will cover advances that reveal specific ways that microbiota influence the physiology of the host, in both health and disease, and how the microbiota may be manipulated at the community, population, organismal, or molecular level to maintain and/or improve host health.

Below are summaries of only a few of the sessions that occurred today.  To hear all the sessions, go to the NHGRI web site ( and access the livestream on demand recordings for each day of the conference.  Each session lasted 15 minutes to give more speakers a chance to contribute and give attendees as large an overview of the field as possible.


Keynote III: Microbiome Interventions: From Fecal Transplants to Synthetic Microbial Therapeutics

(Eric Alm, presenter – Massachusetts Institute of Technology)

Alm is an environmental microbiologist and became interested in the microbiome after an attack of food poisoning.  During the salmonella infection, most of his normal microbiome bacteria went away and never came back.  The new bacteria were very similar, so the functional capabilities of the microbiome went on, but he began to think about how to fix broken or unbalanced microbial systems, especially in the gut.

Antibiotics try to reprogram the microbiome but fail and leave people open to infection afterward.  After three Clostridium difficile infections, the FDA will allow clinicians to try fecal transplants.  In one study, four healthy donors treated 20 patients with a 90% success rate.  One donor treated 12 patients.

Alm and his team decided to model whether a species would engraft, and if it did, what its abundance would be.  He would like to get away from fecal transplants, which he considers crude, and find a way to replace the gut microbiome through the oral route with synthetic treatments.  This will involve finding ways to get bacteria past the acid conditions of the stomach and determining what dosages produce effective results.

At the species level, the team uses metagenomic data to infer strain.  If the number of samples is greater than the number of strains in the samples, then you can often predict how many strains, how abundant, and what their genotypes are.  In one fecal transplant study the team performed, both donor strains transferred to the recipient, but in different proportions.  A new strain for the recipient can thrive if the recipient receives a critical mass of it.  If the donor and the recipient have the same strain, the recipient strain will usually prevail.

If a fecal transplant engrafts in a patient, the patient gets 100% of the strains from the donor.  A third of the strains are environmental (not from donor or recipient) and won’t engraft until all the donor strains engraft

Alm and his team are making a collection of strains they can mail out to people who need them and have very good strain coverage in their freezers, including less abundant strains and strains that were difficult to culture earlier.  They have B. fragilis from 12 different donors.  The team calculates the number of organisms they send by the biomass of the recipient.  Alm said that, when he is designing these therapeutics, it doesn’t matter to him whether the patient hosted the bacteria before their infection or not: if the patient needs that particular strain of bacteria, it goes in.

All of the genetic diversity in each strain evolved during that person’s lifetime.  Each donor has their own tree.  The team cultured a number of isolates from different people and performed genomic and metagenomic sequencing to see how they competed with each other.  The more snips from each isolate means the isolate has colonized the donor longer.  If the strain mutates easily, it often stays on the same tree branch.

The researchers tried to determine which bacterial strains were present because of positive selection, so they measured the hits on the gene that codes for them.  If there were multiple hits on the same gene, it meant that the bacteria was enriched in non-synonymous ways and was there because of positive selection.

There is a special class of proteins that is most enriched in a single person and is associated with wholesale gene gain and loss plus point mutations.  The point mutations occur at interfaces and the hinge regions.  They are so important that they are good targets for phages so they constantly mutate.  Some mutations take over a tree quickly once they emerge, but most come up and either die off or take over on only one branch of the tree.

The team decided to study horizontal gene transfer and found a good example in the Japanese mobile gene pool where people acquired a gene from a marine organism that feeds on the seaweed used to make sushi wrappers.  They took the study to rural villages in Fiji, analyzed 300 individuals, wrote a paper, and got it published in 2012.  After publication, another researcher called them and told them their data was bad because they had only analyzed 500 base pairs and, even if the gene doesn’t get transferred, its regulatory information often does.  When regulatory information gets picked up by another bacteria it changes the bacteria’s function and has an effect on gene fitness.  This phenomenon is pretty widespread across the tree of life and is also species-dependent.

Sometimes it’s impossible to tell whether a horizontal gene transfer took place last week or 10,000 years ago.  The team wondered whether a new transfer species can pick up our adapted bugs, and whether horizontal transfer produces acquisition or multiple losses.  For instance, E. coli doesn’t accumulate diversity over time, it loses previous iterations as it evolves.  Microbes come into a genome and pick up new functions during the recipient’s lifetime.  On Day 0 of one of their experiments, none of the isolates had the prophage but the metagenome had the genes.  By Day 240, the isolates had one copy of the prophage and it was also in the metagenome, with more copies than anyone expected from looking at the isolates.

Species can change their fitness from point mutations and deletions.  They change morphology by making regulatory choices.  The team extracted and isolated endospores with 16s sequencing to find out what percentage of endospores are present as vegetative cells rather than spores at any point in time.  If 24 out of 24 people share a bacteria, the bacteria is considered spore-forming.  If the bacteria only exists in a subset of people, it is considered non-spore-forming.

These concentrations and states are highly variable, not permanent, and easily transmissible.


Gut Microbial Metabolites and Cancer Prevention

(Scott J. Bultman, presenter – University of North Carolina at Chapel Hill)

Dietary components metabolize into substances that influence the development of cancer through a number of different pathways, including altering the intestinal epithelial cells and the immune responses.

Butyrate is a very important metabolite in oncology.  It is a short chain with four carbons.  It is acquired from dietary fiber, and the complex carbohydrates get broken down in the intestine into fatty acids.  Butyrate can go to the mitochondria where it helps make energy, or it can go to the nucleus where it is important for genetics.  Intestinal microbes use butyrate, not glucose, for 70% of their energy production.

Butyrate-producing bacteria are less abundant in colon cancer, but no one knows whether it is a cause or an effect of the disease.  Bultman likes to use gnotobiotic mouse models that he can colonize with commensals that create microbiomes.  In one experiment, he fed these mice high-fiber diets, 60% of which metabolized efficiently into butyrate.  When a colorectal procarcinogen was added, mice on the high-fiber, butyrate-producing diet had a reduced tumor burden.

He performed the same experiment with a mutant strain of bacteria that produces less butyrate, and the protective effect was much reduced because it is dependent on the butyrate level.  When there is a lot of butyrate, mitochondrial energy production is high and the organism switches to a glucose metabolism.  This allows butyrate to build up in the system and affect gene expression.  Butyrate can also stop the process of H3 acetylation.  The team is currently looking at the role of butyrate in inflammation using diet.

Because using gnotobiotic mouse models is expensive and low-throughput, Bultman and his team are exploring mini-gut and organoid technology.  Their limitation is that bacteria need to be injected inside the steroid scaffolds that form the organoids.  The team is making a scaffold the same size and dimension as a human crypt and seeding it with human intestinal stem cells.  The scaffolds are open so chemicals can be applied, and the cells can be co-cultured with microbiota.  They produce a good oxygen gradient, even though the luminal part is anaerobic.

The team wants to move from identifying metabolites to truly understanding their function and there are a lot of interesting candidates to study.  Since butyrate is pleotropic, they expect the new cells probably will be, too, and are excited to move on to the next phase of their research.


Allogeneic Hematopoetic Stem Cell Transplantation, Clinical Outcomes, Intestinal Bacteria, and Potential Mechanisms

(Robert Jenq, presenter – The University of Texas M.D. Anderson Cancer Center)

Jenq told the attendees that he was a clinician, not a microbiologist, and would be presenting from that standpoint.  He talked about the need to get federal funds to explore the microbiome because the therapeutic procedures associated with microbiome research cannot be patented so they won’t attract biotech money.  His example was that fecal transplants were first reported on in 1958, but were not pursued until 55 years later in Holland when researchers decided to explore whether the microbiome could be used as a curative therapy.

Bone marrow transplants were first developed to rescue patients after high-dose cytotoxic chemotherapy.  Clinicians removed the patient’s marrow, gave the patient chemo, then gave them their marrow back.  The field moved to using marrow from patients who were cancer free to increase long-term success.  These allogeneic transplants produced a higher cure rate but came with graft vs. host disease (GVHD) that continues to be a problem.  Disease relapse is higher if GVHD occurs.

During a transplant, the patient’s bone marrow is destroyed by irradiation so the immune system won’t reject the graft.  In this vulnerable state, bacteria can enter and lead to an inflammatory cascade that affects the skin, the GI tract, and the internal organs.  Mice irradiated in germ-free settings had better results, and gut leak antibiotics also helped.  In the beginning, clinicians didn’t know how to restore the gut microbiome after this huge assault, or how critical restoring it was to patient survival.

Researchers started collecting stool samples from bone marrow transplant patients, sequencing the bacteria they found with 16s, and equating the results to clinical outcomes.  Some patients maintained their microbiome pretty well through the transplant process, and some did not.  There was great variability.  The presence of B. lautya bacteria in the gut microbiome predicted good microbiome maintenance and good overall outcomes and its absence predicted poor microbiome maintenance and poor overall outcomes.  The team also found a urinary metabolite that correlated with microbiome health that they started using because it was a quick measure that did not rely on sequencing.

The team also looked at which antibiotics produced better overall outcomes by studying 850 allogeneic bone marrow transplant patient over 20 years.  Piperocillin-tazobactam and imipenen-cilastatin IV increased mortality, but antibiotics that were used to treat penicillin-intolerant patients that were narrow spectrum, anaerobe-sparing drugs produced much better outcomes.  The best results occurred when antibiotics were not used at all.  The key seemed to be keeping the mucin layer intact in the intestine.


Question and Answer Period – Early Morning Sessions

Several interesting points were made during this time.

  • Maria Gloria Dominguez-Bello, who spoke on Microbiomes and Phenotypes: Impact, Restoration, and Transfer, told the conference that she and her team have created a human microbiome seed vault using fecal samples from Indians in her country who live deep in the jungle and have been isolated for 24,000 years.
  • Peter Turnbaugh thought microbiome research could be speeded up by putting things on microchips but said the process was expensive and needed more funding.
  • Someone asked the team what the microbiomes of centenarians looked like and no one knew. Everyone thought this would be a good exploration.


Microbiota-Mediated Defense Against Antibiotic-Resistant Infections

(Eric G. Pamer, presenter – Memorial Sloan Kettering Cancer Center)

The microbiome changes in allogeneic bone marrow transplant patients.  The first study his team performed involved just five patients.  Samples were taken once a week.  Some patients had dramatic changes in their microbiome from one week to the next and some maintained their microbiome well.  The team now collects daily samples.  Before the transplant, patient microbiomes have high bacterial complexity and very high density.  After treatment starts, density can decrease by six logs and there is much less diversity.

In another study of 94 patients, sampling started when the immune system began to recover at the beginning of engraftment.  The patients’ microbiomes contained lots of B. lautya at the start but the levels went down after treatment.  If microbial diversity remained high, the mortality rate from GVHD was around 8% over two years; if it was low, mortality could be as high as 52% over the same time period.

The composition of the gut microbiome at engraftment may have long-term effects.  In another randomized trial, the team extracted and froze patient stool samples and gave patients back their original microbiome one month after engraftment.  They classified the samples according to which bacteria was dominant: entero, staph, or lacto.  A high percentage of patients reacquired their original microbiome.  At the midpoint of the trial there were 26 bloodstream infections, one GVHD death, one infection death, and one death due to disease progression.

Pamer and his colleagues want to identify bacterial species that confer resistance to infections and find out which species are resistant to the antibiotics they use in their hospital.  They have explored mouse microbiomes to answer these questions and found that the presence of Clostridium synbans reduced susceptibility to C. difficile, and the greater the amount of Clostridium, the higher the survival rate.  They treated mice with ampicillin to reduce BRE domination and the results were variable, so they created a seven-bacteria cocktail to enhance survival and tested it in vitro.  Co-culturing B. lautya products created high resistance to BRE.  They created a special B. lautya strain and created their own treatment that inhibits BRE.

They also discovered that an assembly of obligate anaerobes can block infections.

The researchers want to create a bio bank of genome-sequenced, mouse-tested bacteria that come with descriptions of their metabolic dependencies.



Engineering Biological Computers for Human Health Applications

(Timothy Lu, presenter – MIT)

The gut microbiome affects digestion, the immune system, and mood.  Lu would like to move away from broad-spectrum antimicrobials to narrow-spectrum agents, but said the community will need to develop rapid diagnostic tests that can identify bacteria in the office before narrow-spectrum antimicrobials can be used in a clinical setting.

Lu would also like to use CRISPR to engineer micropeptides and bacteriophages that can target microbe DNA.  He and his team discovered that two phagemids could drop bacterial viability fourfold without antibiotics.  If they can target locations on plasmids, they can cure cells of bacterial contamination without compromising their viability.  He wants to develop a library of phages that deliver better, conjugated plasma transfer systems and edit bacteria communities using subtractive techniques.

Lu’s team is exploring whether they can genetically modify bacteria so they can be used in oral treatments for microbiome restoration.  Patients could swallow a pill with a front-end sensor that tells clinicians what is happening in their gut.  Such a sensor was tested in a pig model of bleeding and detected bleeding in the GI tract in 45 minutes.  The team considers this a promising path.

The team would also like to learn how to domesticate “really cool” bacteria and build large libraries of gene regulatory parts, promotors, and inducible systems that can be turned off by various metabolites.  They want to learn how to control gene expression in orthogonal ways so they can switch memory circuits on and off and shut off reporter genes as well.

The researchers are currently taking genetic circuits and putting them into probiotic strains to treat metabolic disorders such as urea cycle disorder and phenylketonuria with non-invasive oral therapies.  They are taking engineered probiotics, genetically modifying them, and adding an anaerobically inducible switch to make sure it’s on.  The switch keeps the probiotic off during manufacture but is turned on before the patient receives it.  The process is in Phase I clinical trials.

Lu and his colleagues are trying to create as many tools for the community as they can and get them into the clinic.  They have created a company called Symlogic to help them do this.

The gaps Lu sees are the need for better sensors to record functional interactions between microbes in the host in real time and in specific locations, and better organism collections so the entire scientific community can work on a problem together. .


Connecting the Gut Microbiome and Cancer Through Circadian Rhythms

(Eugene B. Chang, presenter – University of Chicago)

Chang said there were three parts to his story: relating cancer to metabolism, relating the microbiome to metabolism, and relating cancer to the microbiome.

The intestinal microbiome can be altered by altering circadian rhythms.  Two-thirds of the genes in the intestinal microbiome drive metabolism and one-third control immune mechanisms.  Circadian rhythms are regulated by clock genes and produce oscillatory behavior throughout the day.  The clock genes determine when the body burns and when it stores energy, and are driven by light and dark signals.  In the absence of adequate gut microbiota, these light and dark signals don’t get through.  The oscillatory circadian patterns regulate behavior, weight gain, sleep, and cancer biology through key proteins that regulate cell cycles.  If you disrupt circadian function (sleep disorders, shift work), you disrupt cell cycles and promote tumorigenesis.

High-fat diets alter circadian behavior and also ablate the production of butyrate, which produces phasic shifts in metabolism.  Low fat diets ablate the production of hydrogen sulfide and acetate.  The metabolites that microbes make are mostly fermenting organisms that vary their activity throughout the day.  Intraperitoneal butyrate at a particular time of day can correct the effects of a high-fat diet.

The mechanisms that connect circadian rhythms and metabolism in human beings are hard to study.  Chang and his team would like to develop microbiome-based predictive markers that will help researchers develop effective and sustainable therapeutics.