Some strains of Klebsiella pneumoniae in the gut can turn starchy and sugary foods into alcohol.

Ami Images/Science Source

Microbe that got man drunk could help explain common liver disease

A man in China who, after eating high-carbohydrate or sugary meals, became so intoxicated that he blacked out, has led researchers to discover strains of bacteria in the human gut that could be an important driver of the world’s most common liver disease.

That condition, nonalcoholic fatty liver disease (NAFLD), affects an estimated 1 billion people worldwide, and nearly one in three Americans. The excess fat in liver cells that is characteristic of the disease usually does not cause any symptoms, but in about 25% of people with NAFLD, the accumulation progresses and sometimes causes life-threatening cirrhosis or liver cancer. Scientists found that that the Chinese man’s odd malady stems from gut bacteria that synthesize alcohol from his meals. Researchers say the finding could lead to better ways of predicting who will develop severe forms of NAFLD and may even suggest ways to thwart its progression.

Obesity, diabetes, and other conditions are associated with NAFLD, but no single underlying mechanism explains why fat builds up in the liver of so many people. Some studies have tied gut bacteria to NAFLD, but the idea has remained controversial. The complexity of intestinal flora makes it difficult to sort the contributions of single species. The new finding, published in Cell Metabolism today, focuses on a novel strain of Klebsiella pneumoniae that predominates in a small cohort of NAFLD patients and also caused liver damage in mouse experiments. “I have to admit this is pretty impressive,” says infectious disease specialist David Haslam of Cincinnati Children’s Hospital Medical Center in Ohio, who has been circumspect about earlier attempts to tie intestinal microbes to NAFLD.

Jing Yuan, a microbiologist at the Capital Institute of Pediatrics in Beijing and lead author of the paper, says she and her colleagues became intrigued in June 2014, when a 27-year-old man sought care in Beijing for bouts of unexplained intoxication that dated back 10 years and were getting worse. Some thought he must be a closet drinker, and his mother regularly had him use a breathalyzer. That showed high blood alcohol levels even when she was certain he had not had alcohol. Even odder, when he drank several colas, he sometimes became drunk.

Doctors previously had diagnosed the man’s intoxication problem as autobrewery syndrome, a rarely reported condition in which people become drunk from starchy or sugary foods. It is thought to be caused by gut fermentation, aided by an abundance of yeast. But antifungal treatment had no effect on the man. Liver biopsies showed he had nonalcoholic steatohepatitis (NASH), the severe form of NAFLD. He was moved to the intensive care unit and closely observed. Doctors noted that after he ate a meal high in sugar, his blood alcohol level rose to as high as 400 milligrams per deciliter. “That’s equivalent to 15 shots of 40% [80-proof] whisky,” Yuan says.

Because some other microbes can metabolize sugars into alcohol, Yuan and colleagues analyzed 14 of the man’s stool samples taken at different times for species-specific bacterial DNA fragments. They found that when he was most intoxicated, 18.8% of the bacteria in a sample were K. pneumoniae, a 900-fold increase over normal. When they put these bacteria in a medium of yeast and sugar, they could isolate strains of the bacterium that produced high, medium, or low levels of alcohol.

That’s equivalent to 15 shots of 40% [80-proof] whisky.

Jing Yuan, Capital Institute of Pediatrics

The researchers next analyzed feces from 43 patients with NAFLD, 32 of whom had the severe form, and compared them with 48 healthy people. The team found high levels of high-alcohol- (HiAlc) or medium-alcohol-producing K. pneumoniae strains in samples from 61% of the patients versus 6% of the controls.

To probe further whether these gut microbes might explain the liver disease, the researchers fed mice HiAlc K. pneumoniae, alcohol, or a mixture of yeast and sugars as a control. By 4 weeks, mouse given the bacteria or alcohol had evidence of liver damage, but the others did not. The researchers also put microbes from the original patient into mice raised germ free. Again, the animals suffered liver damage. “The studies are carefully done, and the results are quite convincing,” says gastroenterologist Anna Mae Diehl, whose lab at Duke University in Durham, North Carolina, specializes in NAFLD.

Viruses known as phages can kill specific bacteria. In HiAlc-infected mice pretreated with the phages that target those bacteria, the rodents did not suffer any detectable liver abnormalities. “This raises the possibility that phage might be used to treat NASH,” Haslam says.

The medical literature offers no evidence that people with NAFLD feel drunk when they don’t drink alcohol, and mice with HiAlc did not develop measurable blood alcohol levels. So the researchers decided to also give those rodents high doses of glucose. Blood alcohol levels skyrocketed in the mice, which behaved as if they were inebriated. Giving glucose or fructose to the NAFLD patients similarly produced big jumps in blood alcohol levels that healthy controls didn’t experience. This result suggests giving oral glucose coupled with a blood alcohol test could lead to a diagnostic for HiAlc K. pneumoniae and possibly even predict which NAFLD patients will progress to NASH. “That’s very intriguing and exciting if confirmed in larger human trials,” Haslam says.

Yuan and colleagues report that the initial patient they studied recovered from his bacteria-driven autobrewery syndrome after he began to take antibiotics and changed his diet. His NASH has abated, too. Her team is now planning to study the gut microbes of a large group of people, including children, over time. “We want to investigate why some people have high-alcohol-producing strains of K. pneumoniae in their gut while others don’t,” she says.

Diehl cautions that the new study speaks only to a subset of NAFLD patients. But she predicts “this work will attract a lot of attention.”