Does the human placenta have a microbiome?

A few years ago, the medical community was in a bit of a tizzy over a scientific report that the human placenta has its own microbiome–a complex mixture of bacteria that maybe, just maybe, affected the health of newborn babies.

According to the New York Times' rather breathless reporting at the time, 

"the placenta ... harbors a world of bacteria that may influence the course of pregnancy and help shape an infant’s health and the bacterial makeup of its gut."

This news was very surprising to many scientists, who had long assumed that the placenta was sterile. The 2014 study, titled "The placenta harbors a unique microbiome," found hundreds of bacterial species in the placentas of 320 women. The Times report suggested that the "wrong mix" of bacteria might cause premature births, and it further suggested that the placental microbiome might seed the intestinal microbiome that babies develop later.

Turns out it was all wrong.

Many scientists were skeptical at the time. Those of us working in the microbiome field know that bacterial contamination is everywhere, and it's all too easy to "discover" microbes that came from other sources besides the tissue you thought you were studying. My colleague Jonathan Eisen (at UC Davis) called the 2014 paper and the accompanying discussion "serious overselling of the microbiome."

One good thing about science is that it corrects itself, although sometimes it takes a while. In this case, it took about 5 years. Two studies, both published in mid-2019, looked at hundreds more samples, and carefully screened out contaminants, and found: nothing.

In the first of the newer studies, a group of scientists led by Marcus de Goffau and Gordon Smith at the Sanger Institute in the UK looked at placentas from over 500 newborns. The looked very hard for any evidence of bacteria, but–unlike the scientists in the 2014 study–they took a much more rigorous view of contamination. Like the first study, they found hundreds of species of bacteria, but unlike the first study, the recognized that all of them (with one interesting exception, likely an infection) were contamination. As they explained: 

"samples of placental tissue become contaminated during labor and delivery, even when they were dissected from within the placenta."

In the second study, published just a few weeks after the UK study, a group at Bar Ilan University in Israel looked at 28 human placentas using 5 different techniques, including the techniques from the 2014 study. They found no evidence for bacteria in any of the placentas. Their conclusion was simple and stark: 

"the fetal environment in the womb is sterile."

Unfortunately, the leader of the original study, Kjersti Aagard from Baylor College of Medicine, refuses to admit that her original results were wrong. She claimed, in an interview with The Atlantic's Ed Yong last year, that the UK group were "too strict" about removing bacteria as contaminants, and that they "are not recognizing, or are naive to, other evidence for colonization" by bacteria.

To someone who works in the field, this kind of denial is all too familiar. Microbes are invisible, which means we never actually see the contamination happening–but it does happen, all the time, and many scientific results have evaporated upon closer scrutiny. Even the most careful processing of samples cannot get rid of all the bacterial DNA, because the laboratory kits that we use for sequencing themselves have bacterial DNA in them (as has been shown in several studies). 

So no, there's no placental microbiome, and that's likely a good thing.

One final note: in just the past two months, two major studies (here and here) have appeared that each claimed to find hundreds of bacterial species associated with many types of cancer. Sound familiar? Both studies tried hard to control for contamination, and both involved a massive amount of work. Despite their efforts to exclude contamination, though, I suspect we'll eventually find that these results, which are biologically quite implausible, will evaporate. It might just take a while.