Results from a cross-sectional observational study were recently published in mBio describing the relationship between diet and the microbiome, more specifically, with the degree of AMR. A total of 290 generally healthy adults were recruited in Davis, California, as part of the USDA Nutritional Phenotyping Study, a cohort study designed to elicit multiple connections between dietary/lifestyle/environmental factors and metabolic health. Dietary intake was assessed by both self-administered 24-hour recalls as well as by food frequency questionnaire, and whole-genome shotgun sequencing was performed to describe the gut microbiome, with traditional statistics and machine learning used to analyze the data.
The primary analysis examined the association between specific dietary components and the resistome, the collection of antibiotic resistant genes (ARGs) found in the microbiome. As the authors point out, ARGs are found in ancient humans and people living in non-industrialized communities with no access to modern antibiotics. For example, characterization of the microbiome from Yanomami subjects in the Amazon, people with no Western contact or exposure to antibiotics, revealed unprecedented bacterial and functional diversity of the fecal and skin microbiomes (though not greater oral diversity) – the highest of any human population recorded. But ARGs were found even in this population, including ARGs against multiple antibiotics, even third and fourth-generation cephalosporins.
Recognition that the modern industrial lifestyle is associated with a lower diversity of the gut microbiome prompted the study USDA’s Nutritional Phenotypying Study, an attempt to better understand the relationship diet may play in these changes over time. There were several important takeaways. First, individuals consuming higher levels of fiber, lower levels of animal protein, and more diverse diets had a lower abundance of ARGs. When categorized by low, medium, and high levels of ARG prevalence, individuals with the highest (calorie-adjusted) fiber intake were in the lowest ARG cluster, as were individuals with the lowest intake of protein, especially from beef and pork. Specifically, the gene aminoglycoside-O-phosphotransferase (aph3-dprime) had an inverse association with fiber intake, as did genes involved in multi-metal resistance. Aph3-dprime provides aminoglycoside resistance, the most common mechanism of resistance found in this cohort.
Second, a machine learning-based analysis of this data found that the number of different types of food consumed (phylogenetic diversity) was the largest dietary predictor of difference between the low and medium ARG clusters, with the greatest diversity found in the low ARG cluster. Additionally, plasma dehydroepiandrosterone sulfate (DHEA-S) levels also were associated with cluster differences, though more weakly than dietary intake.
An attempt was made to determine if specific bacterial families could differentiate ARG clusters, perhaps serving as a proxy for ARG prevalence. This was also fruitful, as the abundance of specific taxa, including Enterobacteriaceae and Streptococcaceae, had a linear association with ARG clusters, with more predictive value than any lifestyle or dietary data. This is consistent with data from models of the human gut microbiota, which indicate a bloom in Enterobacteriaceae following exposure to antibiotics, as well as the general understanding that antibiotic use depletes more anaerobic bacteria (obligate anaerobes), which are responsible for producing short-chain fatty acids (SCFAs). A drop in SCFAs, as well as an increase in inflammatory mediators produced by the facultative anaerobes (such as Enterobacteriaceae), may partly explain the association between lower microbial diversity and systemic inflammation (as well as metabolic dysfunction) which has previously been observed in other cohort studies. Facultative anaerobes are also much more likely to be responsible for infection and harbor antibiotic resistant genes.
Other studies have provided important insights relating to dietary intake and the microbiome. PREDICT 1, an intervention study of diet-microbiome-cardiometabolic interactions, was published in Nature Medicine in 2021. This trial involved cohorts in both the US and the UK, with over 1000 participants in total. The study found that diet quality and diversity play a role in the composition of the gut microbiota; on a very broad level, unrefined whole plant foods (e.g., spinach, tomatoes, broccoli) were associated with a greater prevalence of SCFA producing taxa, while less healthy plant foods (juices, refined grains, sweetened beverages) as well as animal-based foods were associated with multiple Clostridium species, though there were distinctions between “healthy” and “less-healthy” animal-based foods. Several dietary indices of diet quality and diversity, such as the Healthy Food Diversity index, were closely correlated with the composition of the gut microbiota. There were several important findings in this study, including associations between the microbiota and several cardiovascular biomarkers and risk prediction scores; for example, the profound observation that the gut microbiome was strongly associated with postprandial triglyceride, C-peptide, and insulin levels, as well as some postprandial lipid measures. These types of studies continue to provide key insights into the relationship between diet, metabolic health, the gut microbiome, and other important determinants of health, such as AMR.