Dr. Gary Gibbons, Director, NHLBI:  Hi, I'm Gary Gibbons, Director of the National Heart, Lung and Blood Institute at the National Institutes of Health. And with me today is Dr. Stanley Hazen to talk more about his study, recently published in Nature Medicine, that relates to cardiovascular disease and the way, and the potential role of gut bacteria in our microbiome. So, Stan, welcome to Director's Corner here.
Dr. Stan Hazen, Cleveland Clinic:  Thank you.
Dr. Gibbons:  Your research offers some new insights into the relationship between diet and cardiovascular health that clinicians and patients have wondered about for clearly more than 50 years. How would you briefly summarize the major findings of your research most recently published in Nature Medicine?
Dr. Hazen:  Well, what we were testing was whether or not a nutrient that's abundant in red meat, carnitine, it's an amino acid that is, it is not essential. What I mean by that is that we synthesize all that we need from our diet. We don't need to take any exogenous amount in. And it plays an important role in energy metabolism by shuttling fatty acids into mitochondria. What we were testing is whether or not carnitine could serve as a substrate for gut flora to generate a metabolite called trimethylamine N-oxide that we had previously shown was linked to cardiovascular disease a couple of years ago. And the reason why we suspected that this might be the case is because of our work a couple of years ago that showed a link between gut microbiota and cardiovascular disease by its metabolism of a structurally related compound called choline, which our major dietary source of is phosphatidylcholine the, or lecithin, the building block of cell membranes.
So, what we began with was a study in humans, and used either a steak or a capsule containing isotope-labeled carnitine and asked a simple question: Will it get metabolized into trimethylamine N-oxide, a compound that we had previously shown when fed to animals would promote atherosclerosis and a marker in blood levels of a large number of patients that tracks with cardiovascular risk. And what we found is that by either ingesting a steak, which is a major source of carnitine, or by taking the capsule with the isotope labeled carnitine, the t-mano or TMAO, I call it t-mano for short, appeared in the blood stream. But then, if we gave the subjects a cocktail of antibiotics and suppressed their intestinal flora, and then gave them that same kind of carnitine challenge, no TMAO appeared in the bloodstream. So that told us, for the first time, that what we're measuring in the blood following eating a steak, in the form of this compound TMAO, was, in fact, coming from gut flora or gut microbes. 
And so the next thing we wanted to know is whether or not carnitine was associated with cardiovascular risk. Previous studies had shown that a higher level of carnitine in the blood associated with higher dietary carnitine. And, of course, higher dietary carnitine associates with a diet rich in red meat, which is a known epidemiologic risk factor for heart disease. So we looked in a large sequential sampling of subjects who came to cardiology for evaluation and were then followed over time to see who went on and developed a heart attack, stroke, or death. This was over 2,500 people and we measured plasma carnitine initially in their blood. And we saw that carnitine levels vary strongly and predicted future risk of heart attack, stroke, or death in the individuals over the ensuing three-year period after collection of the blood sample. And this was independent of existing risk factors, such as age, gender, diabetes, smoking, hypertension, cholesterol levels, renal function, etc. 
But because we thought it wasn't carnitine that was the actual culprit that was proatherogenic but rather, it was this TMAO, the metabolite that's liberated by the gut flora, we then measured TMAO in the same samples, and asked, when we include that in the statistical model, what happens to the prognostic value of carnitine. And it completely eliminated any association with cardiac risk. And only the TMAO remained a strong predictor of risk, consistent with the carnitine not being the atherogenic piece, per se, but the precursor that gave rise to the atherogenic piece, which is TMAO. 
So then we went from the human studies to animals and confirmed, in germ-free animals, that carnitine would get converted into TMAO in a gut flora dependent fashion. The germ-free animals couldn't do it when you gavage them with carnitine. But if you put them in conventional cages and let their gut recolonize with microbes, now you give them the carnitine and now they made the TMAO. We then looked at these animals in an atherosclerosis animal model and saw that the germ-free animals, or the animals in which the gut flora was suppressed, if they were fed a carnitine-rich diet, they did not develop enhanced atherosclerosis. However, the animals that had intact flora and made the TMAO, carnitine in the diet doubled the extent of atherosclerotic plaque burden in the animal models.
And then the more exiting and interesting, actually, findings came up. If we actually looked at the ability to convert carnitine into TMAO in the animals that were on a chronic carnitine diet, it was ten to 100-fold more on the animals that were chronically on a carnitine diet. And what this suggested to us is that they had shifted the composition of their microbes. And the types of microbes that like carnitine for fuel actually became more populus in their intestine and then subsequently were able to use carnitine and make TMAO or TMA, which gets converted into TMAO. And that was actually found to be the case, that the microbial composition in the animals on a chronic carnitine diet had been shifted to types of microbes that we were able to show preferred carnitine and would associate with forming TMAO. 
And so then, again, bouncing back and forth between animals and humans, we then said okay, let's go back to the human situation. Where would you have a situation with a chronic exposure to carnitine or not? So we started comparing carnitine metabolism in omnivores as compared to vegetarian and vegans. And those were very interesting findings because we found that how an omnivore metabolizes carnitine was dramatically different than a vegetarian or a vegan. These were all long-term vegetarian or vegans, over a year of vegetarian or vegan lifestyle. And what we found is that the omnivore would readily produce TMAO appearing in the bloodstream following ingestion of carnitine, whether it be a steak or a capsule with the isotope labeled carnitine. And the vegetarian or vegan would generate very, very little TMAO after ingesting a capsule of carnitine. And we even had one vegan that agreed to eat a steak in the name of science and saw that no TMAO was produced, despite eating a large filet mignon. And so he had a very nice steak, but did not generate any TMAO. 
And then we collected stool from the subjects, from individuals who were vegetarian or vegan versus omnivores and then performed analysis on the microbial composition of the stool. And compared that to the plasma levels of the TMAO, and were able to identify specific microbial taxa that co-associated with the chronic dietary pattern of the individual, as well as the plasma TMAO. So there were specific genus of microbes that seemed to be more prevalent in the omnivore and associated with forming TMA and TMAO, and found at much lower levels in the vegetarian or the vegan.
Dr. Gibbons and Dr. Hazen Transcript
Segment 1