Microbes in the City

The metagenomic approach led investigators to look at ecosystems as collections of genes rather than as collections of organisms. For example, the gut microbiomes of obese and non-obese people show consistent differences in the abundance of genes for energy production and carbohydrate transport, regardless of the specific strains of organisms present. Similarly, genes for antibiotic resistance often appear in multiple species in different microbial ecosystems.

It's not just scientists who find microbiomes compelling: President Obama has taken an interest, and Handelsman described several White House efforts to support the field. The President's Budget for Fiscal Year 2016 includes $1.2 billion for studying antibiotic resistance, with an emphasis on using microbiome data to track resistance genes and to identify new sources of antibiotics.

The government also requested input from scientists, and received a huge number of suggestions about where the field should go next. "We need tools to be able to look at microbes and the chemicals they produce," said Handelsman, adding that more sophisticated mathematical models will also be essential for analyzing the deluge of microbiome data.

One pill makes you larger

Martin Blaser of New York University began his presentation with a familiar set of maps documenting the skyrocketing rates of obesity in the U.S., with the greatest and fastest weight gains in the southeastern states. Maps showing outpatient antibiotic prescriptions reveal a similar pattern; states where doctors hand out the most antibiotics have the highest obesity rates. "This is not a difference based on the rate of serious bacterial infections, this reflects culture and practice," Blaser said.

Rates of obesity (right) and antibiotic prescription (left) overlap geographically. (Images presented by Martin Blaser. Sources: Obesity data – U.S. CDC; Antibiotic-use data – Hicks et al. 2013.)

Though the observations do not prove causality, they do point to a possible mechanism for rising obesity. Farmers have known for decades that low doses of antibiotics cause animals to grow faster and fatter. This use of sub-therapeutic antibiotics as livestock growth promoters now accounts for over 70% of U.S. antibiotic use.

Blaser and his colleagues found that livestock-like antibiotic regimens and high-fat diets have additive effects in mice, prompting the animals that received both to get much fatter than controls. In a series of additional experiments, the researchers demonstrated that short courses of antibiotics at a crucial stage of early development—similar to what many children now experience—also led to weight gain and changes in gut microbiota in mice.

Going viral

Rumi Chunara of New York University shifted the focus from bacteria to viruses, describing her work on disease tracking. Chunara's group wants to improve influenza surveillance, which currently relies on reports from clinics. While those data are crucial for public health planning, they miss the bulk of flu cases in the community. "A lot of times if you have influenza or an acute respiratory infection, you might not go in to see your physician, so that data doesn't get captured," Chunara explained.

Collecting illness reports directly from individuals would be a more sensitive way to measure disease rates, but previous efforts to do so have fallen short. The Google Flu Trends tool, for example, relied on Internet search data to predict influenza rates, but search habits turned out to be a poor proxy for illness rates.

Chunara decided to take a direct approach instead, recruiting a broad sample of people through multiple media channels and then giving them clinical sampling kits. On experiencing flu symptoms, participants send the research team nasal swabs and saliva samples. The researchers test the samples for 17 common respiratory pathogens and tell participants which illness is present and which other pathogens have been detected in their area. The data correlate well with clinical surveillance results, and Chunara now hopes to expand the program to include foodborne illnesses.

It takes a village

Ilana Brito of the Massachusetts Institute of Technology gave a hot topic talk to close the first session. She has studied a different type of microbial transmission question: How do microbes move between people and their environments? In a large city such as New York, these transfers occur millions of times a day, making them difficult to track. Instead, Brito focused on a more tractable environment by visiting and sampling small, isolated villages in the Fiji Islands.

Brito enlisted 300 people from five villages and collected extensive data on their family structures, social contacts, and behaviors. She then performed two types of sequencing analyses on more than 500 samples from the people and their environments, tracking both gene and microorganism distribution. Comparing the data to sequences from Americans in the Human Microbiome Project, Brito found some genes were common to both cohorts, while others were unique to one group or the other.

Fijian gut microbiota contain more genes for breaking down plant fibers while American gut microbiota contain more for animal protein degradation, consistent with the different diets of the two groups. Individual villages within Fiji also show distinct microbial gene patterns. "It suggests that there are place-based ... mobile gene pools that might have specific functions for different populations even at the very, very local scale," Brito said.

With appropriate skepticism, though, the ability to sequence entire microbial ecosystems provides unprecedented insight into pathogenesis and epidemiology. As recent outbreaks of West Nile virus, severe acute respiratory syndrome (SARS), and Middle East respiratory syndrome (MERS) have demonstrated, this new capability has arrived just in time.

Rather than waiting to react to the next emerging disease, Lipkin and his colleagues are hoping to anticipate it by surveying likely reservoir hosts for new pathogens. It's a big job. Looking just at flying foxes (Pteropus giganteus) in Bangladesh, the team found 58 viruses in these bats' feces; 50 of them were new to science. "Our estimate is that there's a minimum of 320 000 viruses yet to be identified" on Earth, Lipkin said.

Closer to home, Lipkin's team also investigated the case of a woman in the New York area who became acutely ill immediately after Hurricane Sandy. After sequencing non-human DNA from infarcts that appeared on her scalp, the researchers identified a polyomavirus that had never been seen before. They still do not know where it came from, or whether it was the chief cause of her illness.

Many emerging infectious diseases are zoonotic, originating in animals and moving into people. Diseases highlighted in orange are zoonotic; diseases highlighted in white spread from person to person. (Image courtesy of Ian Lipkin)

Some of the team's discoveries are also helping scientists study previously known viruses. For example, Lipkin's lab has found viruses in dogs that are closely related to human hepatitis C virus, providing a new laboratory model for that disease.

Home is where the microbiome is

Jack Gilbert of Argonne National Laboratory presented his group's work on microbial distribution in homes and hospitals. Because every person brings a distinct microbiome wherever he or she goes, buildings hold enormous biodiversity. "The bacteria, the fungi, the archaea, the viruses, and the way we build these ecosystems could ... potentially have profound implications for everything from autism and Alzheimer's and Parkinson's ... to digest[ion]," Gilbert said.

To study these interactions, Gilbert and his colleagues enlisted volunteers who were moving from one house to another. The team sampled the houses and occupants before and after the move. Each family had a unique microbial fingerprint that colonized the new house within 24 hours, displacing the building's former microbiota. Furthermore, close family members shared more microbes with one another than with lodgers living in the same house, tracing a microbial map of relationships. Adding a pet such as a dog to the house dramatically increased the spread of microbes between individuals, so couples living with a dog became more microbially similar than were couples living without one.

Microbial transfers trace relationships between individuals and via the built environment. (Image courtesy of Jack Gilbert)

In another project, Gilbert is studying a large new hospital building in Chicago, monitoring environmental conditions within the building as well as its microbial distribution. The first surprise was how quickly the microbiome of the entire structure changed when the facility first opened. "As soon as the hospital opened, literally the very next day, the entire system shifted to an entirely new ecosystem state, and this new ecosystem state is an occupied building environment," Gilbert said. The researchers now hope to find ways to increase beneficial microbial populations while reducing harmful ones in the hospital.

The gift of feces

Eric Alm of the Massachusetts Institute of Technology began his talk with a description of one of the best-known examples of microbiome engineering—treatment of Clostridium difficile infections with fecal transplants. Antibiotics sometimes allow C. difficile to overrun patients' normal intestinal microbiomes, causing serious digestive problems. Transferring feces from healthy donors into these patients cures about 90% of these infections.

However, Alm and his colleagues found that the transplant patients' gut microbiota are distinct from both their original microbiota and those of their donors. "It looked nothing like either, it was just very distinct, and it was very difficult ... to predict which of these particular species were going to engraft," Alm said.

The researchers built computer models to predict which species and strains would take up residence after a fecal transplant. Most species contained three to five strains, which tended to engraft or vanish as a group; patients either retained all or none of the strains of that species. When a species persisted, it also tended to promote the acquisition of more strains of the same species later. "We're learning that transmission to some extent depends on whether you've already got something in your microbiome," Alm explained.

Individual genes revealed another level of activity. After finding that antibiotic-resistance genes travel between farm animals and human guts, Alm's team discovered that the promoter regions upstream of the genes were the most mobile parts. This transfer of regulatory regions can drive significant changes in phenotypes, possibly switching some bacteria between pathogenic and nonpathogenic forms.

Big microbes on campus

Juan Maestre of the University of Texas at Austin started his presentation with an ambitious vision of the future. Maestre described a world in which researchers could build a three-dimensional map tracking all the microbes and genes in a city in real time, with the information available to a highly educated population able to understand the meaning of the data. "There is a long road to get there, but in the meanwhile we can do something," Maestre said.

Maestre's project, UTBiome, focuses on the University of Texas campus in Austin, TX. The project includes several sampling campaigns that cover building renovations, creeks, laboratories, and other features of the area. With the aid of dozens of students using online forms and user-friendly sampling kits, Maestre has collected over 400 samples.

The program has illuminated several microbial phenomena on the campus, revealing, for example, the effects of aging sewer infrastructure on biofilms in creeks and dramatic changes in indoor air when people enter a room. The data are stored online but only downloadable by request. "We are trying to find ... the best way to put it in a way that scientists can access it and the general public can access it but nobody freaks out," Maestre said.

The researchers plan to study the microbiomes of local homes to compare apartments and detached homes. They will also study how different ventilation systems and building materials, as well as the presence of pets, affect the microbiomes of buildings.

Like Boston's transit system, New York's subway looks a lot like skin at the microbial level. The Bronx had the most biodiversity, with Brooklyn second. One station, however, stood out. At South Ferry on Manhattan's southern tip, the researchers found 10 species not seen anywhere else in the subway system. That station flooded in 2012 during Hurricane Sandy, and some of the distinctive sequences matched species previously associated with fish. "What happened on the walls [was] an ocean or marine environment that was left behind," Mason said.

The samples from across the subway system also contained human skin cells and DNA, which revealed local ethnic differences. More Asian-associated markers appeared in Chinatown, while in northern Manhattan around Harlem there was more African DNA. "For all kingdoms, [there's] a molecular echo that's left behind of the DNA that was transiting through that system," Mason said.

Sex and the sewer

Jane Carlton of New York University took the discussion into the sewer, but she got there indirectly. Initially, her group wanted to study Trichomonas vaginalis, the most common sexually transmitted pathogen. In an effort to determine the genetic diversity of T. vaginalis in New York City, Carlton and her colleagues worked with public health clinics to collect vaginal swab samples. That project was a success, but the team could only sample one trichomonad in a relatively small number of individuals. To broaden their analysis to other trichomonads and expand it citywide, they looked at sewage.

Trichomonads are a diverse group of organisms, including many pathogens. (Image courtesy of Jane Carlton)

The NYC sewer system pumps 1.5 billion gallons of waste through some seven thousand miles of pipes each day. Carlton's group started by sampling a much smaller system, the on-site sewage treatment plant at the Visionaire, a recently constructed luxury condominium. The plant's aerated sewage tank "is full of cockroaches, which is great, because cockroaches have trichomonads in them," Carlton said.

After developing their analytical techniques in the Visionaire plant, the investigators collaborated with the NYC Department of Environmental Protection to sample the citywide system. Those specimens contained diverse taxa of protists, including T. vaginalis and other trichomonads. The team is now trying to work backward through the pipelines to map protists to their sources. Carlton closed with a call for more standardized protocols and analytical techniques across microbiome studies. "In my dream I'd love that we could compare our data with ... sewage data in [other cities]," she said.

Going with the flow

Rachel Poretsky of the University of Illinois at Chicago also studies a sewer, but one that doubles as a navigable waterway—the Chicago River. "It's estimated that the Chicago River flow is about 70% wastewater effluent," she said. Although the effluent is mostly treated, heavy rainfall sometimes overloads the city's combined sewer system and forces authorities to open floodgates, dumping potentially contaminated effluent into Lake Michigan.

Poretsky sampled the river at points upstream and downstream of the major waste treatment plants and near combined sewer overflows. To look at how the discharge from the plants affects aquatic microbes, she and her colleagues performed both ribosomal DNA sequencing and metagenomic analyses. "In terms of community composition, the river is distinct before effluent input and after," she said, adding that the overall biodiversity of the samples declines downstream of the treatment plants. The metagenomic analyses also revealed that the sewage plants deliver numerous antibiotic-resistance genes to the river's microbiome.

Chicago is the only major city in the U.S. whose sewage plants treat, but do not fully disinfect, sewage. That situation is about to change, at least for most of the city's plants. Under pressure from the U.S. Environmental Protection Agency, Chicago is now adding a disinfection step at its smaller sewage facilities, and Poretsky's team is poised to study the effects of this change. The group is already performing laboratory-scale experiments to predict the effects of different disinfection regimens. "What'll be interesting is when we start looking at the metagenomic data and determining ... if the same sorts of patterns hold in the environment," Poretsky said.

In cities, American cockroaches specialize in sewer life. To date, the only study of the microbiome of a field population of these roaches focused on insects in a natural cave. That work revealed a surprisingly depauperate microbiota. Schal hopes to study sewer roaches soon to see if their microbiomes are more diverse.

To determine the functions of insect microbiomes, Schal compared feces from microbially sterile German cockroaches and from nonsterile controls. Roaches preferred to aggregate near the nonsterile feces, "suggesting that bacteria associated with these feces are producing agents ... that cause these cockroaches to aggregate," Schal said. His team found several volatile carboxilic acid compounds in the nonsterile feces that were not in the sterile feces. A purified cocktail of these compounds proved to be a powerful roach attractant.

Besides looking at the effects of microbes on pests, the researchers are now doing the reverse, examining how large infestations change the microbiomes of homes. An infestation of ten thousand insects—not uncommon in some types of housing—consumes and generates large quantities of organic material that should alter the local microbiome. To measure this effect, Schal plans to analyze the microbiomes of infested homes before and after extermination efforts.

Selective media

Laurie Garrett of the Council on Foreign Relations led a panel discussion to end the meeting. Garrett began by asking panelist Joel Ackelsberg about the experience of explaining nuanced microbiome data to the public. Ackelsberg was the public face of the New York City Department of Health and Mental Hygiene when researchers published the first data surveying subway microbiomes. That project included some findings that could be considered frightening when taken out of context, and indeed city newspapers were quick to trumpet some of the more striking results, such as the identification of potential Bacillus anthracis (anthrax) sequences on trains.

Ackelsberg emphasized the importance of transparency in such situations. "If we're not transparent with what we know and what we don't know and what we are doing in order to learn more, not only is it bad risk communication, but I think the public sees through it very quickly," he said, adding that "we all live one bad quote away from the abyss when we're in the public [eye]."

Garrett next asked Jo Handelsman about the White House's position on the potential use of microbiome evidence in court. Handelsman began by explaining that with the exception of DNA testing, no forensic technique has been held to truly rigorous standards. Crime scene investigators match bite marks and hair, for example, by eye, and experts commonly score matches and mismatches inconsistently. "It's a little bit hard to hold the microbial forensics that we know must be coming down the pike to some high standard of science when there's currently very little in forensics that can even be called science," Handelsman said, echoing the conclusion of a 2009 National Academy of Sciences report on the field.

Paula Olsiewski talked about the role of the Sloan Foundation, which currently funds far more research on the microbiology of the built environment than the U.S. government does. Garrett asked whether researchers are expected to get informed consent from cities or building owners. Olsiewski explained that while there are no rigorously defined standards yet, she makes a point of meeting with researchers seeking Sloan funding to discuss how they will interact with property owners. She added that the research could help improve architecture: "We're inadvertently designing ecosystems; maybe with some of what we learn we can design them so that they're better for us."

Audience members asked panelists about the relative contributions of both antibiotic use on farms and human overprescription in the growing problem of antibiotic resistance. "The microbes that are in us are a lot more likely to affect us and become pathogens, [but antibiotic] use on farms as you know is vastly higher," Handelsman said, explaining that the available evidence does not show which use drives more clinically relevant antibiotic resistance.

Handelsman also addressed the question of regulating the rapidly growing probiotics industry. "The market has already gotten ahead of the regulation, no question about it; the claims on many of the probiotics are unfounded," she said. It is unclear, she added, whether the U.S. Food and Drug Administration will be able to implement new regulations before the end of the Obama administration, particularly because the agency still lacks a permanent commissioner.