In 1984, the microbiologist Barry Marshall notoriously used himself as an experimental subject for his research, and drank the contents of a flask containing the bacterium Helicobacter pylori as part of his efforts to demonstrate that bacteria cause stomach ulcers1. Writing in Nature, Duan et al.2 do not report taking such drastic action to investigate a bacterial connection to disease. Nevertheless, their careful analysis of a liver disease called alcoholic hepatitis, in studies of mice and analysis of samples from people who have the disease, also provide attention-grabbing evidence for the involvement of a suspected bacterial culprit.
Alcoholic hepatitis is a poorly understood condition related to high alcohol intake, and is difficult to treat. Previous experiments in mice have hinted that the gut-dwelling bacterium Enterococcus faecalis might be involved3. However, E. faecalis is usually thought of as an old friend that inhabits the guts of many animals across the evolutionary tree, from humans to nematode worms4. This species usually represents less than 0.1% of all the bacteria in faecal samples from healthy people5. However, after antibiotic treatment, bacteria of the genus Enterococcus increase in prevalence to become one of the most common types of microbe in the gut6. E. faecalis can infect the blood, heart, bladder and brain, and teeth that have undergone root-canal surgery7,8.
Duan and colleagues analysed human faecal samples. They identified E. faecalis in the stools of about 80% of people with alcoholic hepatitis that they tested, and about 30% of the strains of E. faecalis present had genes that encode a toxin called cytolysin. Furthermore, people with the disease had almost 3,000 times more E. faecalis in their stool samples than did people who did not have alcoholic hepatitis. That isn’t concrete proof that the disease is caused by this bacterium. However, the authors’ data also show that the presence of cytolysin in stools correlates with mortality — 89% of the people whose faecal samples contained cytolysin died within 180 days of hospitalization, compared with only 3.8% of the people who had alcoholic hepatitis but whose stool samples lacked the toxin.
The authors next examined the connection between E. faecalis and liver disease in mice. The animals were colonized with strains of E. faecalis that either did or didn’t make cytolysin, and some were then fed a high-alcohol diet, with others given an alcohol-free diet. Only the mice on the high-alcohol diet and that had been colonized with cytolysin-producing E. faecalis developed liver damage (Fig. 1a).
Then, using germ-free mice (which had no natural microorganisms), the authors transplanted stool samples from people with alcoholic hepatitis that contained E. faecalis strains in which cytolysin was either present or absent. Mice on a high-alcohol diet that were colonized with stools containing cytolysin displayed a range of signs indicating liver damage and the death of liver cells, whereas animals on such a diet and colonized with stools lacking cytolysin showed no major signs of liver damage.
To understand the disease-causing mechanisms, the authors isolated liver cells from the animals, and found that cell death in response to cytolysin exposure was dose-dependent. The response to cytolysin was the same whether or not the mice had received a high-alcohol diet. This suggests that, rather than alcohol causing alcoholic hepatitis by damaging the liver cells, damage arises because alcohol increases the permeability of the gut lining to allow cytolysin-producing E. faecalis to reach the liver and cause disease symptoms (Fig. 1a).
Given the limited treatment options for alcoholic hepatitis, the authors investigated whether steps might be taken to develop a therapy that exploits bacterium-targeting viruses called bacteriophages, or phages for short. Phages have the advantage over antibiotics of being highly specific, and so avoid also killing beneficial bacteria. Furthermore, because the surface of a human cell differs substantially from that of a bacterial cell, phages aren’t thought to infect animal or human cells9.
Phages have been used to remove Salmonella and Shigella bacteria from infected human intestines for almost 100 years10. They have also been used to remove the disease-causing bacterium Clostridium difficile from artificial intestines, and from hamsters infected with this bacterium11,12. It has been suggested that they might one day be used in humans or animals to remodel the composition of the community of gut microorganisms (the microbiota), to produce a healthier microbiota consisting of more bacteria associated with good health and fewer associated with disease13. The potential of E. faecalis-targeting phages to tackle human diseases is already being discussed7, and phages can kill antibiotic-resistant strains of E. faecalis associated with human bone and wound infections14,15 and dental cavities16. Furthermore, phages are being developed for use in the food industry to remove E. faecalis from cheese cultures to prevent the production of toxic waste products17.
To test whether a method could be developed to specifically remove cytolysin-producing E. faecalis from mice, the authors identified some phages that target these bacteria (Fig.1b) but leave other gut bacteria unaffected. Mice that received human stool samples and a high-alcohol diet and that were given E. faecalis-targeting phages had less liver damage than did mice given phages that killed a different bacterium not usually found in animals.
This study demonstrates the advantages of using phages in detective work to investigate the contributions of microbes to disease. The authors show that phages can be used to identify disease-causing bacterial components, and also raise the possibility that phages might offer potential treatment options. Further tests, including clinical trials, would be required to assess whether a phage approach would be useful in a human context. For example, phage treatment might help to target E. faecalis in the gut before a person receives a liver transplant.
In Duan and colleagues’ study, the phages could treat a disease in which a causal component is a bacterium that normally resides in the gut, even though the disease site is elsewhere in the body. Although much phage research focuses on the use of these viruses to treat diseases associated with antibiotic-resistant bacteria, the work by Duan et al. raises the possibility of a much wider clinical role for them. There is growing evidence that gut microbes can affect the function of certain cells in the brain, and studies are ongoing to determine whether such microbes have a role in human brain diseases (see go.nature.com/2cp1kfk). Perhaps phages could become part of the next generation of targeted antimicrobial therapies for diseases that are currently difficult to treat. Indeed, there might be many diseases that we currently don’t realize have a microbial component, and which could be tackled by phages.
Nature 575, 451-453 (2019)