Placental Microbiome’s Existence Challenged

“What they have done is extremely good,” Andrew Onderdonk, a microbiologist at Brigham and Women’s Hospital and Harvard Medical School, tells The Scientist, “I just don’t accept their conclusions.” His group has used both molecular and culture-based techniques to investigate the bacteria in the placentas of preterm infants. “The inconsistency resides in what they consider to be contaminants and what other investigators, including myself, have shown by a number of methods . . . are not contaminants.” 

See “The Maternal Microbiome”

The authors of the new study biopsied villi from 537 placentas that had been delivered vaginally or by Caesarean section. They spiked the tissue with a bacterium not found in humans, Salmonella bongori, to act as a positive control, then extracted DNA. For 80 of the samples, the researchers used both 16S rRNA gene sequencing and metagenomic analysis. For the remaining samples, they used two different DNA extraction kits in parallel and then performed 16S sequencing on each sample in order to compare the results from the two kits.

Because 16S rRNA sequencing and metagenomic sequencing work in different ways, the investigators predicted that any artifacts present in one should not show up in the other. They could detect S. bongori, their positive control, using both methods, as well as group B Streptococcus in about five percent of the placentas collected via Caesarean section before labor started, but otherwise the 16S and metagenomic results didn’t match up. And although the researchers found other bacterial DNA in some of the placental samples from vaginal and C-section deliveries, they could pinpoint the origin of much of the DNA to contamination in the DNA extraction kits.

“Most of the bacterial signals we detect don’t come from poor handling of the tissue,” says coauthor Steve Charnock-Jones, a biologist at the University of Cambridge. “It’s an unavoidable problem that you have to use reagents to get the DNA out of the tissue, and those reagents themselves have bacterial DNA in them.”

See “DNA Extraction Kits Contaminated”

“If you’re analyzing a sample from soil or stool where there is a huge number of bacteria, the contribution of what’s in the reagents is completely irrelevant,” he says. Anywhere that there are very few microbes, “the contribution from the kit is a much greater proportion so that’s what you detect,” he adds.

Charnock-Jones and colleagues also identified other potential sources of contamination that could explain the bacterial signals they detected. Among them were the mode of delivery—that is, vaginal bugs that could have been deposited on the placenta as it came through the birth canal—as well as handling during the biopsy, reagents used to amplify the samples, or the sequencing equipment or materials. And they did not find a link between birth outcomes, such as preterm birth or babies born small for their gestational age, and the bacterial signatures. The authors interpreted all their findings to mean that there is no resident microbiome in the human placenta.

Kjersti Aagaard, a physician and microbiologist at Baylor College of Medicine, questions that conclusion. She coauthored a 2014 study using molecular methods to characterize placental microbiota. Aagaard tells The Scientist that depending upon how deep the sequencing coverage was, bacteria could have been missed because they are present in such low numbers. She adds that just because there are shared bacteria between the placenta and those that others have shown to be present in the vagina doesn’t mean that the overlap was caused by contamination. The authors of the new paper didn’t assess the vaginal microbiota of their subjects.

Onderdonk also cautions about drawing conclusions based on 16S sequencing in the placenta. In some of his group’s prior work, “the 16S signal—even when we knew there were lots of bacteria present in the sample—was extremely hard to detect,” he says. “There are likely inhibitors that prevent proper detection of the DNA that’s present.”

Despite these and other open questions, the families and genera that the new study found were similar to what Aagaard’s team showed in 2014. “If you take the data at face level and you disregard some of their speculations and conclusions, it is highly consistent with what we and others have published,” she says.

“A missed opportunity for the paper is that they didn’t leave anything open,” Indira Mysorekar, a biologist at Washington University School of Medicine who did not participate in the work, tells The Scientist. She explains that the placental location from which the authors took their samples could have affected their findings, but regardless, whether there are microbes present or not is not the whole story. For one thing, if the placenta is truly sterile, it likely has far-reaching implications for fetal immune system development, she says. “These studies, which focus on the sequencing and contamination, technically are extremely important and critical and necessary, but they don’t address the biology. There are so many exciting questions to ask.”

M.C. de Goffau et al., “Human placenta has no microbiome but can contain potential pathogens,” Nature, doi:10.1038/s41586-019-1451-5, 2019.

Abby Olena is a freelance journalist based in North Carolina. Find her on Twitter @abbyolena.