Review Article
Gut Microbiota in Obesity and Metabolic Abnormalities: A Matter of Composition or Functionality?

https://doi.org/10.1016/j.arcmed.2017.11.003Get rights and content

The obesity pandemic and the metabolic complications derived from it represent a major public health challenge worldwide. Although obesity is a multifactorial disease, research from the past decade suggests that the gut microbiota interacts with host genetics and diet, as well as with other environmental factors, and thus contributes to the development of obesity and related complications. Despite abundant research on animal models, substantial evidence from humans has only started to accumulate over the past few years. Thus, the aim of the present review is to discuss structural and functional characteristics of the gut microbiome in human obesity, challenges associated with multi-omic technologies, and advances in identifying microbial metabolites with a direct link to obesity and metabolic complications.

To date, studies suggests that obesity is related to low microbial diversity and taxon depletion sometimes resulting from an interaction with host dietary habits and genotype. These findings support the idea that the depletion or absence of certain taxa leaves an empty niche, likely leading to compromised functionality and thus promoting dysbiosis. Although the role of altered gut microbiota as cause or consequence of obesity remains controversial, research on microbial genomes and metabolites points towards an increased extraction of energy from the diet in obesity and suggests that metabolites, such as trimethylamine-N-oxide or branched-chain amino acids, participate in metabolic complications. Future research should be focused on structural and functional levels to unravel the mechanism linking gut microbiota and obesity.

Introduction

Over the past decades, obesity has emerged as a leading health problem worldwide. By 2014, more than 1.9 billion adults were overweight. Of these, over 600 million were obese, across both developed and developing countries (1). In obesity, the accumulation of excess body fat adversely affects health and predisposes to metabolic diseases, such as type 2 diabetes, coronary heart disease and certain forms of cancer, among others (2). As economic transition advances, it has been predicted that overweight and obesity trends in developing nations will surpass the maximum values reported in developed countries (3). In fact, according to the Organization for Economic Cooperation and Development (OECD), Mexico is ranked among countries with the highest prevalence of overweight and obesity, in both adults and children (Obesity Update, OECD, 2014) and one of the leading countries in Latin America. This situation is alarming, particularly considering that between a third and half of obese children will grow up to be obese adults (4). Obesity thus carries the threat of a reduced length and quality of life for current and future generations, and efforts to better understand the factors contributing to the pathophysiology of the disease are of utmost importance (5).

Fundamentally, overweight and obesity result from a positive energy balance. However, a complex interaction between genetic and environmental factors is involved in weight gain. Although it has been shown that obesity is highly heritable, the genetic variants associated with obesity explain only a small proportion of the overall obesity risk (<5% for BMI) (6), and environmental factors further modify this genetic contribution to obesity. An obesogenic environment includes factors that promote an overconsumption of energy through the availability of low-cost and high-energy foods and the habit of consuming these in large quantities. At the same time, a reduction in physical activity and an increase in sedentary behaviors decreases energy expenditure (7). However, the combination of these environmental factors and genetic susceptibility only provides a partial explanation for the development of obesity, and other factors have been suggested to explain its high prevalence.

The gut microbiota has been put forward as a relevant factor connecting genes, environment and immune system (8), thus playing a role in the pathogenesis of obesity and related metabolic complications. While animal and human studies have demonstrated that a gut microbiota imbalance, or dysbiosis, is associated with obesity (9), it remains controversial whether the role of gut microbiota in the development of this disease is causal. Given our increasing knowledge in the field, it is likely that the mechanistic role of gut microbes and their metabolites in obesity will be uncovered. This will lead to the development of personalized, precise therapies to improve or prevent obesity and derived metabolic complications. Thus, the aim of this review is to discuss the structural and functional capacity of the gut microbiome in the context of human obesity, the challenges associated with multi-omic technologies, and the advances in the identification of microbial metabolites that have been directly linked to obesity and metabolic complications.

The human colon is inhabited by bacteria, archaea, eukaryotes and viruses. The bacterial community comprises up to 1000 species and is dominated by the two phyla Firmicutes and Bacteroidetes, while Proteobacteria, Actinobacteria, Fusobacteria and Verrucomicrobia are less abundant (10). This community jointly encodes more than five million genes that complement human genome functions and provide a wide range of additional metabolic capabilities (11).

The colonization of the gastrointestinal tract begins at birth. Type of delivery, feeding practices, hygiene and antibiotic treatment are the main factors shaping the composition of the infant microbiota 12, 13. After birth, the human gut is colonized by strict anaerobes, establishing a microbial community characterized by low diversity and dominated by the phyla Proteobacteria and Actinobacteria. As the neonate grows, the microbiota becomes more diverse, and the phyla Firmicutes and Bacteroidetes become dominant. An individually distinct microbial profile is reached at one year (14). Although it has been postulated that around three years of age, the gut microbiota becomes more stable and adult-like (15), the differences observed between children and adults 16, 17 suggest that modifications continue over many years, and that maturity and higher levels of stability are reached until early adulthood. Thus, throughout life, multiple factors, such as host genetics, lifestyle, diet and medication, will influence gut microbiota composition and functionality, to either maintain a symbiotic relationship with the host or induce an imbalance, leading to dysbiosis. Interestingly, some of these factors have also been implicated in the development of obesity, reinforcing that gut microbiota perturbations by endogenous and environmental factors could contribute to the development of diseases.

During natural birth, infants are colonized by taxa originating from the mother's vagina, whereas infants born by C-section are initially colonized by microbes from the mother's skin. Thus, infant gut microbiota composition varies with the type of delivery (18). Although it is unclear whether these signatures are maintained later in life, the association of C-section delivery with an increased risk of obesity in childhood (19) suggests a contribution of the gut microbiota to this risk.

Starting in infancy, the microbial community within the gut is influenced by the diet. Studies report that the gut microbiota composition varies between breastfed and formula-fed infants, an effect which is mainly attributed to the presence of human milk oligosaccharides 20, 21. Weaning and the introduction of solid foods produce important changes in gut microbiota composition, thus during infancy, inter-individual similarities are mainly given by the type of diet 15, 22. Interestingly, research in humans and animal models has shown that the gut microbiota can be rapidly altered by a change in diet. Switching to a high-fat/high-sugar diet or a diet based on animal products can modify the structure of the gut microbiota within a few days 23, 24. In addition, habitual food intake seems to have a long-term influence on gut microbiota composition (22). A comparison between children living in Italy and Burkina Faso found a higher abundance of Bacteroidetes, specifically of the genera Prevotella and Xylanibacter, in the gut microbiota of African children. Since Prevotella and Xylanibacter are known to degrade cellulose and xylans, this could be interpreted as an adaptation of the gut microbiota of African children to maximize the energy extraction from their high-fiber diet (25). Similarly, enterotypes, described as a gut microbiota configurations mainly driven by a specific genus (Bacteroides, Prevotella or Ruminococcus) (26), have been linked to long-term dietary patterns, particularly with the consumption of a diet rich in protein and animal fat (western diet) and with a high consumption of carbohydrates (27). Lastly, a healthy diverse diet has a positive impact on gut microbiota diversity (28). Dietary pattern analysis, which assesses the diet as a whole, rather than individual foods or nutrients (29), has shown, for example, that subjects with a dietary pattern characterized by lower consumption of confectionary and sugary drinks and high consumption of fruits, yogurts and soups (prepared mainly with water and vegetables) had a richer and more diverse gut microbiota (29). Thus, the association of certain dietary patterns with a reduced obesity risk (30) may be at least partly mediated by a beneficial modulation of gut microbiota composition.

Medication, especially antibiotics, can have a profound effect on the gut microbiota (31). In a large-scale study of gut microbiota variation, medications used by participants over the six months preceding the study had the largest explanatory value (32). The effect of antibiotics ranges from fairly mild to drastic, depending on its antimicrobial spectrum (33), and the restoration to the initial gut microbiota structure (also termed resilience) varies among individuals (34). Interestingly, prospective studies in infants have shown that early antibiotic exposure increases the risk of overweight or obesity later in life 35, 36, 37. While some authors have suggested that this could be explained by a weak stability and dysbiosis-prone gut microbiota, long-term cohort studies are needed to corroborate this hypothesis (38). Other medications besides antibiotics can also affect gut microbiota composition. Metformin, a widely-prescribed drug for the treatment of diabetes, has been found to increase butyrate-producing bacteria, which has been linked to its beneficial effect (39). Long-term use of proton pump inhibitors (PPIs), frequently taken for gastroesophageal reflux and peptic ulcer disease, is also able to modify the composition of the gut microbiota and diminish its diversity 40, 41, 42. Interestingly, it has been proposed that gut microbiota alterations caused by the use of certain medications, such as antidepressants, could potentially be at the root of the weight gain often seen in the patients taking them 43, 44. Together, these findings highlight the importance of considering medications in gut microbiome studies: without doing so, it will be even harder to distinguish whether any alteration in the gut microbiota is caused by the medications used or by the underlying disease (45).

As discussed in a comment published earlier this year by Peter Turnbaugh, a pioneer of gut microbiome-obesity research, even after a decade of research on animal models demonstrating the contribution of gut microbiota to host adiposity, the field struggles to provide answers on the role and signature of the gut microbiota in human obesity (46). Despite the enormous amount of information provided by an exponential increase of studies on animal models in the last decade, studies in humans have only gathered pace over the last couple of years. More importantly, the consistency of the evidence coming from humans is still poor, and a specific microbial signature linked with obesity has not been identified. Among the factors potentially underlying discrepancies between studies are differences in technical procedures (e.g. use of qPCR, 16S microarrays, 16S rRNA sequencing and metagenomics) and processing methodologies, as well as environmental factors that may influence the gut microbiota, such as diet, medication and geographic origin, among others.

Most studies addressing the involvement of the microbiome in obesity have relied on the use of the 16S rRNA marker gene. While this approach has expanded our knowledge of gut microbial diversity, and provided a relatively accurate fingerprint of phylogenetic community membership and structure, it reveals little about the functional properties of the community. Although tools predicting function from 16S rRNA sequences have been developed such as PICRUSt, (47), their use is limited by the level of accuracy with which the reference genome database used in the analysis reflects gene functions for the microbial community of interest. Thus, to gain a more accurate insight into the microbial community at the species and strain levels, as well as into microbiome gene function, high throughput shotgun metagenomic sequencing has become an important tool, as it avoids biases introduced by amplicon sequencing.

Therefore, in the next paragraphs, we will critically examine the evidence for the relationship of the gut microbiota with overweight and obesity on the one hand, and with metabolic complications on the other. Such an analysis can be done at different levels (Figure 1). We will cover the association of the gut microbiota with obesity and metabolic complications at a structural and functional level by discussing metagenomics and metabolites. In addition, we will emphasize the difference between children and adults, for which there are clear evidences in the literature (16).

The present review includes case-control studies for overweight, obesity and metabolic complications. We also included cross-sectional and longitudinal studies assessing an association of BMI with gut microbiota. In addition, to enhance the clarity of the review, we will briefly discuss studies using qPCR and microarrays, before focusing on 16S rRNA and shotgun sequencing (metagenomics).

Most of the case-control studies for overweight and obesity have been performed by qPCR, likely due to its low cost, but with the main limitation that this approach does not provide a global evaluation of the sample bacterial community. In Table 1, we summarize the major findings that include only those taxa previously suggested to be associated with obesity. Interestingly, in adults, the most consistent finding is a higher abundance of Escherichia coli in normal-weight vs. overweight or obese individuals, and a higher abundance of Lactobacillus in obese vs. normal-weight 51, 54, 56, 57, 59, 61. This is somehow intriguing, since there are multiple pathogenic strains of E coli, which are associated with deleterious health effects, such as intestinal infections and irritable bowel syndrome (66). In contrast, Lactobacillus has traditionally been thought of as “beneficial bacteria” and is widely used as a probiotic (67). In this context, results from probiotic interventions have shown contrasting effects depending on the species or strain used (68), and a recent meta-analysis has shown that the effects can also be age dependent. In adults, Lactobacillus administration was found to have overall a significantly positive effect on weight loss, while in infants and children, it was followed by significant weight gain (69). The association of Lactobacillus with overweight and obesity in children has also been replicated in qPCR studies 51, 54.

Another interesting observation is the lower abundance of the genus Akkermansia in obesity, which has been observed mainly in children 49, 53. In adults, less consistent findings indicate a diminished abundance of Bifidobacterium, Methanobrevibacter and Faecalibacterium prausnitzii in obesity 55, 57, 60. Interestingly, the negative association of F. prausnitzii with BMI observed in qPCR studies was replicated in microarray studies. These latter ones also consistently report a higher abundance of Roseburia, a member of Lachnospiraceae family, in subjects with high BMI 63, 64. These studies were performed in small samples of European populations (Belgium and Netherlands) with likely similar environmental patterns, and the Dutch study evaluated monozygotic twins discordant for obesity, to diminish the genetic effect. Roseburia has the ability to hydrolyze and ferment polysaccharides into short chain fatty acids (SCFA), thereby increasing energy harvest from the diet (70). While the latter could represent a possible mechanism for the observed association between Roseburia and an increased BMI, intervention studies with high protein/low carbohydrate diets have observed a decrease in the abundance of this genus (71). These findings could rather exemplify an interaction between the diet and the microbiota. Given the pronounced influence of diet on gut microbiota composition, it is an important confounder variable in the observed associations with BMI, and future studies may need to take this into account. So far, only a few studies have addressed this issue and they indeed observed that certain associations were diet dependent. One of these studies found that after controlling for diet, only Eubacterium dolichum was positively associated with visceral fat mass as a surrogate of obesity, while Bifidobacterium correlated negatively (72).

While 16S rRNA and shotgun sequencing allow a global composition analysis, case-control studies performed over the past decade have not made extensive use of these techniques, likely due to the high costs involved. Indeed, a recent meta-analysis reported that only 5% of the studies aimed at evaluating BMI and gut microbiota in humans used 16S rRNA sequencing (73). Since the number of studies on gut microbiota differences between normal-weight, overweight and obese humans is still low, in this section, we also discuss studies testing for an association with BMI, particularly those with a large sample size. Table 2 provides a summary of the studies using next generation sequencing. In humans, one of the first findings was an increased Firmicutes to Bacteroidetes ratio in obese vs. lean subjects (80). Interestingly, this increased ratio has not been observed consistently in subsequent studies performed in either adults or children. This increased ratio has been extensively replicated in animal models, though it is not clear whether it represents a signature of obesity in humans. In fact, two meta-analyses found that the differences in Firmicutes/Bacteroidetes ratio between normal-weight and obese individuals were not statistically significant 73, 92. Furthermore, more recent studies with a large sample size (>500 subjects) fail to detect an association of the Firmicutes/Bacteroidetes ratio with BMI or obesity 32, 86. Overall, the Firmicutes/Bacteroidetes ratio might not function as a feature to distinguish between normal and obese human gut microbiota.

Some of the studies in children replicated the negative association between F. prausnitzii and BMI reported in adults, while others observed the opposite effect 53, 78. F. prausnitzii is a butyrate producer, which has been associated with a beneficial metabolic effect in animal models (93). A negative association with obesity is certainly in line with the results in animal models; however, the authors of studies observing the opposite effect argue that butyrate production is strain specific 62, 78. In any case, more comprehensive analyses are needed to elucidate the precise role of this taxon in obesity.

Interestingly, a meta-analysis in adults, which included the results of MetaHIT and the Human Microbiome Project, reported no significant effect of gut microbiome composition on BMI (83). This might be explained by increased inter-individual variability, which can probably outweigh the effect of the phenotype in an insufficiently large sample. In line with this, a large cross-sectional study aimed at evaluating the influence of different factors on gut microbiota composition showed that the effect of the BMI was relatively small; it suggested that in order to observe the phenotype effect, a large sample size should be considered (32). In fact, studies with a large sample size have identified significant associations between gut microbiota composition and BMI. In one such study including mono- and dizygotic twins to control for genetic background, Oscillospira, M. smithii, and the Christensenellacea family were significantly associated with a lower BMI. Interestingly, the last two taxa were found to be highly heritable (94), highlighting the need to control for genetic background as an important factor determining gut microbiome composition. The effect of Christensenellacea in lower weight gain was confirmed in gnotobiotic mice receiving donor feces from human individuals enriched with this bacterial family. Moreover, this association has been replicated by other groups (84) including our own observations in children (95). However, the precise mechanism underlying this association is still unknown. Another large-scale study found a negative association of the Rikenellaceae family with BMI (84). In a follow up study, based on shotgun sequencing, the authors extended their finding, in which they identified two species within the Rikenellaceae family that correlated negatively with BMI: Alistipes finegoldii and Alistipes senegalensis (89). This highlights the relevance of a deeper analysis for uncovering not only the species, but the specific strains involved in the physiopathology of the disease. Interestingly, a weight-loss intervention study in German women of the same enterotype (Bacteroides-enterotype) and based on a very low calorie diet, found Alistipes as a marker of persistent weight-loss success (96). Although more studies are needed to confirm these findings, it may be worth performing mechanistic studies to determine if this taxon participates in weight regulation.

A consistent finding in shotgun sequencing studies of different populations is the previously reported depletion of genus Akkermansia in subjects with a high BMI 32, 84, 87, 88. Akkermansia is a mucin-degrading bacterium, that inhabits the outer mucus layer of the intestinal barrier, with a median abundance >1% in the intestinal tract of Europeans 97, 98. In animal studies, Akkermansia muciniphila has been widely associated with a better metabolic profile. In humans, its abundance and gene richness have been positively associated with a healthier metabolic status, including a better body fat distribution and the absence of metabolic syndrome (MetS) 99, 100. Interestingly, a study in rodents suggested that the effect of A. muciniphila is diet dependent, since under a fiber-free diet, its mucin degrading capabilities may compromise the intestinal barrier and allow for pathogen susceptibility (101). A study in an Asian population showed a higher abundance of this taxon in subjects with type 2 diabetes (T2D), and authors suggested that the thinning of the intestinal barrier by Akkermansia may allow for bacterial translocation and thus contribute to the pathogenesis of T2D (102). In conclusion, while most studies suggest a beneficial role for Akkermansia in metabolic profile, it may have a dual role depending on dietary patterns. Therefore, in studies aimed to elucidate the role of gut microbiota in obesity, the environmental factors influencing the microbiome need to be considered. Similarly, an approach that includes the ecosystem surrounding gut microbiota should be considered for mechanistic studies.

Another consistent finding is the reported lower abundance of Bifidobacterium in subjects with high BMI, which was first reported in a qPCR-based study in children 72, 75. This is consistent with the literature, since Bifidobacterium has been associated with a better metabolic profile and has been considered as a probiotic for years. Furthermore, its beneficial effects are supported by studies in animal models, where administration of different Bifidobacterium species prevents excessive weight gain under a high-fat diet (103).

Regardless of the presence or abundance of specific bacterial families or species, obesity has been associated with a lower diversity and richness of the gut microbiota 74, 78, 84. This is particularly true for animal models, where this effect has been observed in several studies. In humans, the picture is less clear and a meta-analysis concluded that the effect was not consistent across studies, likely because of differences in technical or bioinformatics approaches. Only de novo clustering, which retains OTUs that do not match the database, resulted in significant differences between obese and lean individuals (92). Furthermore, another meta-analysis, where sequences were re-analyzed by de novo clustering, showed a lower diversity and richness in obese subjects, and this result was obtained when analyzing both, individual studies and pooled data (73). These findings suggest that using de novo clustering, where unidentified taxa contribute, allows the identification of diversity differences among study groups. However, Sze MA, et al. also noted that the effect of diversity in obesity is relatively small, and that many of the published studies might be underpowered to observe or even reproduce this effect (73). In fact, the significant association of obesity with low microbial diversity was confirmed in two large-scale studies of European populations, thus supporting the need of a large sample size to detect any effect 32, 89. Using an alternative approach, two studies analyzed diversity based on the number of genes present in the metagenome, rather than the number of OTUs or species. Le Chatelier E, et al. (87) clustered subjects by low and high gene count and compared phenotype differences between these clusters. Interestingly, subjects in the high gene count cluster had a lower BMI and a better metabolic profile. Another recent study, in Chinese subjects confirmed a lower gene count in obese subjects as compared with healthy controls (91). The above studies suggest that a low microbial diversity and richness is associated with an increased relative risk of obesity (73). Longitudinal studies are needed to determine whether the reduced microbial diversity and richness are a cause or consequence of obesity, and eventually intervention studies aimed at increasing gut microbiota diversity might aid in the treatment of the disease.

In conclusion, the most consistent results among gut microbiota compositional studies indicate that obesity is characterized by a low abundance of certain taxa, such as Rikenellaceae and Oscillospira, as well as a decreased diversity of bacterial communities (Figure 2). These findings suggest that it is the absence or decrease of certain bacteria that contributes to obesity. Also, the reduced abundance of Christensenellaceae, Akkermansia and Bifidobacterium in obesity might be mediated by dietary or genetic effects. Although, some of these taxa have been extensively studied in animal models, it is worth investigating which specific functions they might be doing within the human gut community.

As stated previously, the obesity pandemic has a significant impact on the prevalence of metabolic complications, such as insulin resistance, dyslipidemias, hypertension and of MetS. Worldwide, it has been estimated that approximately 20–30% of adults suffer from MetS (104). Because of the impact of gut microbiota on host metabolism, many studies have searched for specific microbial signatures associated with the presence of these disorders.

In a study of Korean twins, Lactobacillus, Suterella and Methanobrevibacter were enriched in adults with MetS. In an analysis of individual traits, the main positive associations found were of Suterella and Lactobacillus with fasting blood glucose, and of Lactobacillus, with waist circumference as previously observed, (as a probable surrogate marker for a high BMI). In contrast, Parabacteroides, Bifidobacterium, Odoribacter, Akkermansia, Rikenellaceae and Christensenellaceae were over-represented in healthy individuals. Most of these taxa also correlated with individual traits; serum triglycerides, blood pressure and waist circumference (100). Unfortunately, since the authors did not control the correlations for BMI, it is difficult to dissect whether the latter effects are directly driven by the metabolic traits or confounded by waist circumference. In fact, except for Parabacteroides, all of the above taxa have been previously found to be associated with BMI 82, 84.

Metagenomics analyses allow a deeper insight into the species or strains involved in the presence of metabolic complications. In one such study in obese Danish women, the species A. muciniphila, B. cellulosilyticus, B. faecis, B. nordii, B. pectinophilus, I. bartlettii, D. longicatena, O. splanchnicus, B longum and R. inulinivorans were all negatively associated with insulin resistance or dyslipidemia, and most of these associations remained significant after adjustment for confounding factors such as age, body fat and habitual dietary intake (105). However, an example of a negative association was A. muciniphila with serum triglycerides, which was lost after correcting for dietary fat intake. These results support the notion that Akkermansia may protect against a high BMI, but its effect on host physiology is likely diet dependent. Similarly, the association of B. longum with insulin resistance was lost after an adjustment for carbohydrate intake. Together, these findings stress the importance of evaluating dietary intake when studying the link between the gut microbiota and metabolic markers. Interestingly, this study found contrasting associations of different F. prausnitzii strains with markers of inflammation and insulin resistance, emphasizing the importance of analyzing down to strain level (105).

Other studies have evaluated the association of gut microbiota structure and lipid profile independent of BMI. In the Dutch LifeLines-DEEP study, Clostridiaceae and Lachnospiraceae families were associated with high LDL cholesterol levels. Similarly, Collinsella stercoris showed a strong positive association with triglyceride levels, while the family Pasteurellaceae and the genus Coprococcus correlated negatively with serum triglycerides (84). Again, and consistent with several reports in the literature, higher diversity was associated with a healthier metabolic profile, since it correlated negatively with triglyceride levels and with the presence of MetS 84, 100.

It is clear that more controlled case-control studies with a large sample size and which account for confounding factors, are needed to corroborate current findings and recognize further taxa contributing to obesity and metabolic complications. In addition, it is not clear whether variation in composition affects the metabolic activity of the community and, consequently, of the host. Shotgun sequencing studies will provide more information about the functional changes related to obesity and metabolic complications. Finally, most of these studies have been performed in adults, however, considering the increasing prevalence of metabolic complications in children, large-scale studies in are also needed in this age group.

The genes of the gut microbiome encode functions that cannot be performed by the host. For example, it has been observed that the functional profile of the microbiome of breastfed babies shows characteristics related to carbohydrate and vitamin metabolism, which may represent a response to the content of breastmilk (15). Post-weaning, the microbiome is also involved in the metabolism of amino acids and the biotransformation of bile acids and various xenobiotics (106).

In humans, in contrast with the wide differences in taxonomic composition between individuals, there are fewer interpersonal differences regarding the abundance of metabolic pathways, and the core metagenome is established early in life and remains stable thereafter (107). However, in contrast to the human genome, the gut microbiome is quite variable, with only a third of its constituent genes found in most individuals (108). The functions encoded by this core microbiome include microbial housekeeping functions, needed for their interaction with the host (i.e. epithelial adhesion and production of immunomodulatory compounds), and specialized functions, such as biosynthesis of SCFA, vitamins and amino acids or the enrichment of certain lipopolysaccharides (108). As with taxonomic composition, variations in genomic potential are influenced by lifestyle factors, such as host diet, and other environmental conditions. Indeed, some studies that evaluated the bacterial genome profiles of people from different geographic regions (Japan, Russia, China and Denmark), or of rural vs. urban populations, found significant differences in the abundance of genes involved in nutrient metabolism and xenobiotic detoxification. This suggests that diet or environmental factors, including the exposure to carcinogens or nitrogen oxides, may be shaping the gut microbiome 11, 109, 110.

Although several authors have suggested that a set of microbial functions is differentially present in obesity, the evidence is still scarce. In animal models of high-fat/high-sugar diet-induced obesity, an enrichment of pathways related to membrane transporters and carbohydrate metabolism has been observed; specifically involving the phosphotransferase system, which plays a role in the uptake of sugars and the phosphorylation of a variety of carbohydrates (111). The few existing studies dedicated to the bacterial metabolic pathways enriched or depleted in human obesity have found an enrichment in phosphotransferase system pathways and in genes involved in the production of NO2 91, 112, similar to reports in animal models. In addition, it has been suggested that obese microbiomes are capable of using a more diverse set of energy sources, thus contributing to an increased energy harvest (112). Other studies have observed an enrichment of folate, riboflavin, nicotinamide adenine dinucleotide and lipopolysaccharide biosynthesis in the microbiome of subjects with a high BMI (91). The observed enrichment of pathways involved in glutathione production might represent a mechanism to balance the oxidative stress present in obesity (88). Alternatively, researchers have focused on finding a healthy “functional core”, understood as a complement of metabolic and other molecular functions performed by the microbiome within a specific habitat, but not necessarily provided by the same organisms in different people (113). It is thought that the absence of some of these functions could lead to a dysbiotic state and thus contribute to obesity. Although a healthy “functional core” has not been established yet, it has been suggested that the relationship of microbial diversity with a healthy status represents an increased functional redundancy, as a result of a more diverse set of microbes (108). Of note, a lower diversity has been associated with other diseases besides obesity (114), suggesting that compromised microbial functionality might lead to disease.

Other studies have focused on metabolic complications, such as insulin resistance and MetS. A study in obese women showed that the microbiome of those with MetS was enriched in bacterial chemotaxis and flagellar assembly; this also correlated with higher fasting blood glucose and HbA1c (96). Although mechanistic studies are still lacking, the higher release of bacterial flagellar products might activate pro-inflammatory pathways through their interaction with the immune system and thus contribute to metabolic endotoxemia. A study that assessed both metabolic pathways and taxonomic groups present in metabolically compromised subjects found a reduction of butyrate-producing bacteria, as well as an increased potential for mucus degradation, hydrogen sulfide formation and oxidative stress management in these patients (87). In addition, a recent study aimed at identifying metabolic pathways differentiated in insulin resistance showed that the enriched pathways were mainly related to the biosynthesis of branched chain amino acids (BCAAs), vitamins, lipopolysaccharides as well as various transport systems, independent of BMI. In contrast, the microbiome negatively associated with HOMA-IR included modules related to methanogenesis, pyruvate oxidation and transport systems (including inward transport of BCAAs) (115). Overall, these studies support the idea that metabolically compromised individuals harbor an inflammation-associated microbiota.

Metagenomics studies in human obesity and metabolic complications are still scarce. While it is not yet clear which metabolic pathways contribute to obesity, the gut microbiome of subjects with metabolic complications is enriched in inflammatory pathways 87, 96. The interaction of some of these pathways with the host are mediated by various bacterial components and metabolites and a more detailed investigation on which of these metabolites are absorbed by the host might help elucidate whether they act at intestinal or systemic level. Interestingly, a metagenomics study of T2D was able to identify women with glucose intolerance based on their fecal microbiome (116). Although longitudinal studies are warranted, this emphasizes the potential of the gut microbiome to identify individuals at risk for metabolic complications. Thus, future intervention studies should aim at modulating not only the composition, but also the metabolic pathways of the gut microbiota.

While metagenome data provide information on the community's potential metabolic pathways, they do not necessarily reflect the actual metabolic pathways used by the gut microbiota (117). Thus, in an effort to go beyond potential pathways, an increasingly used approach is to analyze, within the host, the metabolites that are either produced or regulated by the gut microbiota, and to investigate how they contribute to host health. It has been estimated that ∼10% of all circulating metabolites in the human body are microbial-derived metabolites that participate in human metabolism (118). Some of the compounds that have been linked to either obesity or its metabolic complications will be reviewed in the following paragraphs.

Short-chain fatty acids (SCFAs) – i.e. butyrate, propionate and acetate – are the end products of polysaccharide fermentation, a metabolic pathway that has been implicated in diverse functions in obesity and metabolic complications (24). In general, SCFAs can be used as source of calories for the host. Butyrate functions as energy substrate for colonocytes, which absorb up to 95% of the produced butyrate (119). Acetate and propionate are absorbed through portal circulation and take part in de novo lipogenesis or gluconeogenesis in the liver (120), thus providing extra energy to the host (121). However, individual SCFAs can also have a positive metabolic effect by reducing inflammation, improving insulin sensitivity and contributing to satiety, (9). Although the mechanisms behind these effects go beyond the purpose of the present review 119, 122, these findings highlight the dichotomous role of SCFAs in the pathogenesis of obesity (121).

Studies in humans have mostly evaluated fecal levels of SCFA. In obesity-focused studies of adults and children, the findings are highly consistent; overweight or obese individuals show higher levels of total SCFAs than normal-weight individuals 48, 55, 59, 78, 123, 124. This increase seems to arise from higher levels of all SCFAs, rather than just a single one. Since higher SCFA levels can potentially result from either increased production or reduced intestinal absorption, both conditions have been evaluated. A study aimed at evaluating rectal SCFA absorption found that reduced intestinal absorption was an unlikely cause of elevated SCFA levels, since no differences were observed between the obese and lean phenotype (125). Another study reported that, in vitro, the microbiota of obese children produced greater levels of SCFAs than the microbiota of normal-weight children (123). This conclusion is supported by metagenomic analyses in obese mice, where pathways capable of degrading dietary polysaccharides and producing SCFA were found enriched (126). Likewise, in a prebiotic intervention with inulin-type fructans, the SCFAs acetate and propionate, that correlated positively with BMI, decreased significantly after treatment (127), further supporting a role for SCFA in obesity. Interestingly, an intervention study aimed at delivering propionate to the colon through propionate-esterified carbohydrates reduced weight gain in a randomized 24 week study of overweight adults (128). These results suggest that modulation of specific rather than total SCFAs could help improve the metabolic profile. Whereas the role of SCFAs in metabolic complications has been extensively documented in animal models 129, 130. In humans, experimental evidence is mainly derived from dietary interventions using fermentable polysaccharides. These studies have found increased SCFA gut levels and an improvement in metabolic traits after treatment, although sometimes with conflicting results (129). There is thus a clear need for more studies specifically aimed at investigating the role of SCFAs in human metabolic complications.

Despite the well-known role of some bacterial species in the production of some of these SCFAs, studies assessing which bacteria within the gut contribute to SCFA production in humans are still scarce. Available data include correlations between OTUs of the genus Prevotella and propionate, as well as between Faecalibacterium and butyrate levels (78). In addition, since it is known that bacteria can adapt their metabolism to the prevailing conditions, it may be worth using metatranscriptomics to investigate which species are contributing to the increased energy extraction capacity. This type of analysis could also aid in the understanding of microbial community behavior, e.g. by uncovering whether increased SCFA production results from a metabolic adaptation of bacterial groups already present in the community, or from an increase in abundance of some of their members.

Primary bile acids (cholic and chenodeoxycholic acids) are synthesized in the liver from cholesterol and released into the small intestine, where they are conjugated to glycine or taurine. In the ileum, these bile acids are deconjugated by gut bacteria and converted into secondary bile acids, which in turn serve as signaling molecules (131). Secondary bile acids can bind to cellular receptors such as the farnesoid X receptor (FXR) and the G-protein-coupled bile acid receptor 1 (also known as TGR5), thus regulating a variety of physiological processes including energy expenditure, insulin sensitivity, and cholesterol balance (132). Through FXR, bile acids modulate serum triglyceride levels and lipid accumulation in the liver and contribute to the maintenance of glucose homeostasis (133). Animal models show that, by activating TGR5, secondary bile acids influence GLP-1 secretion by L-cells, thereby improving liver and pancreatic function and enhancing glucose tolerance (134). Furthermore, activation of TGR5 in brown adipose tissue increases energy expenditure, preventing obesity and insulin resistance 135, 136. Thus, the participation of the gut microbiota in the metabolism of bile acids strengthen its role as an important regulator of lipid and glucose metabolism, which in turn is linked to the development of obesity and metabolic complications.

There is a paucity of studies in animal models describing the role of bile acids in weight regulation and obesity (137). In humans, the few existing studies are focused on BMI rather than obesity and only consider adults. Some report a positive correlation of total serum bile acids with BMI (138), but correlations between specific bile acids – e.g. chenodeoxycholic acid (CDCA), deoxycholic acid (DCA) and other derivatives tauro-λ-muricholic acid and glyco-λ-muricholic acid–with BMI have also been observed 139, 140. In addition, certain bile acids have been positively associated with insulin resistance and serum triglycerides, independent of BMI 139, 141. While it is unclear whether the latter is cause or consequence, the ability of the microbiota to modify the pool of bile acids suggests the participation of some microorganisms in these pathways. Unfortunately, the specific contribution of the gut microbiota to bile acid metabolism has been poorly explored in human studies. Recently, a large-scale study found that serum bile acids explained a significant proportion of microbiota variance, and that the genus Parabacteroides was significantly associated with the levels of lithocholic acid (LCA). An exploration of Parabacteroides genomic pathways found genes involved in secondary bile acid metabolism, suggesting a possible role for this genus in the modulation of circulating levels of bile acids (142). The concept of a microbiota-bile acids-host physiology axis in humans, is very attractive and carries great potential. Nevertheless, additional studies are warranted to identify the role of the gut microbiota in modulating bile acid pools in humans, and thus explore their modulation as a mechanism underlying the development of obesity and metabolic complications.

Trimethyl-amine-N-oxide (TMAO) is a product derived from bacterial metabolism as well as from food sources (i.e. fish) that has been strongly associated with cardiovascular disease risk, (143) and extensively reviewed elsewhere (144). In animal models, it has shown to contribute to a pro-atherogenic state by inhibiting bile acid synthesis and increasing cholesterol transport by macrophages in the endothelial lining of arteries 145, 146. In addition, there is some evidence that it may contribute to impaired glucose homeostasis and to the development of a fatty liver 147, 148, 149. In circulation, TMAO results from the production of trimethylamine (TMA) from specific dietary components such as phosphatidylcholine (found in cheese, eggs, liver and peanuts) or L-carnitine (found in red meat). Microbial conversion of these dietary nutrients to TMA is performed by specific enzymes (TMA lyase) via a wide variety of metabolic pathways. Then TMA is absorbed into portal circulation and transported to the liver, where it is oxidized by the flavin monooxygenase 3 to TMAO and excreted by the kidneys (150).

So far, it is unclear whether TMAO is associated to obesity or BMI itself, and studies performed in humans have shown contrasting results 151, 152, 153, 154, 155. However, some reports have shown a positive association with glycemic state indicators (HbA1c values and fasting glucose), cholesterol levels, inflammation markers and the presence of MetS, while a negative correlation was observed with the estimated glomerular filtration rate 147, 153, 154. This suggests that TMAO is involved in the development of metabolic complications, rather than obesity per se. For instance, TMAO was shown to be a strong predictor of T2DM risk (156), while in adults with non-alcoholic fatty liver disease, TMAO was also associated with the steatosis score (147). These findings are supported by animal studies which found that, by altering the bile acid pool, TMAO can indirectly increase hepatic triglyceride levels and serum cholesterol levels. Besides bacterial production, TMAO can also originate directly from seafood consumption; however, most human studies have found no, or only a mild, correlation between meat, egg or fish intake and circulating TMAO levels 152, 155, 157. In contrast, TMAO has been found increased in individuals with a western-type dietary pattern (rich in animal food sources and processed sugar) compared to subjects with a high intake of a “prudent diet” (rich in fruits, vegetables, nuts, poultry and fish) (158). In fact, it has been suggested that gut microbiota composition and functionality may respond to dietary patterns rather than to single nutrients or foods (159).

As mentioned above, TMA lyases are responsible for the production of TMA. This pathway is widely, but discontinuously, distributed among bacteria (160), indicating that the phylogenetic composition of the microbiota is likely a poor predictor of anaerobic choline metabolism, and that an evaluation of the gene clusters associated with TMA production should be performed. In this context, it was recently shown that modulation of the microbial TMA lyase activity by 3,3-dimethyl-1-butanol, rather than changes in gut microbial composition, significantly inhibited microbial TMA formation, decreasing circulating TMAO levels and reducing atherosclerotic lesions in an animal model (161).

Higher serum levels of branched-chain amino acids (BCAAs) and aromatic amino acids have been related to obesity and insulin resistance in animal models and in humans (162). More recently, they have also been associated with the development of hypertriglyceridemia in children and adults 163, 164, 165. The source of these amino acids, originally postulated to originate from dietary sources, is still controversial. However, it was recently suggested that gut microbiota could be contributing to higher circulation levels of these amino acids, since up to 20% of circulating levels of amino acids can be the result of bacterial metabolism (166). In a multi-omics approach, it was shown for the first time that insulin resistant individuals with higher BCAA levels also showed an increase in biosynthetic pathways of these amino acids in their gut bacterial community (115). It was further demonstrated that the species B. vulgatus and Prevotella copri contributed to this increased biosynthesis, and increased BCAA production was confirmed in a murine model (115). Two additional studies have evaluated the association between these amino acids and gut microbiota composition in humans. The first study showed a significant positive correlation between BCAA serum levels and OTUs from the genus Blautia, and a negative correlation with Christensenellaceae (86). The second observed a negative correlation of Bacteroides including B. thetaiotaomicron, B. intestinalis, B. ovatus and B. uniformis with BCAAs (91). While it remains to be determined whether this negative association corresponds to a causal relationship, it has been suggested that by increasing amino acid degradation, bacteria could contribute to lower BCAA circulating levels 91, 115. It is still unknown whether bacterial metabolic pathways, including BCAA synthesis, also contribute to hypertriglyceridemia. Thus, more studies are needed to corroborate the role of gut microbiota in the circulating levels of amino acids, and to dissect their relationship not only with insulin resistance and obesity, but also with the development of hypertriglyceridemia.

Given the role of microbiota in human obesity, there has been an increasing interest in interventions targeted at modulating microbiome composition and functionality and thus improving host metabolic phenotype. Although there are many ways to achieve such a modulation, including dietary components (specific diets, probiotics, prebiotics or functional ingredients) and procedures such as fecal microbiota transplantation and bariatric surgery, it remains unclear which strategy will perform better in each case.

A large number of dietary interventions, including the use of functional foods, have shown changes in the structure of gut microbiota, concomitant to metabolic improvements (167). While the particular results of those dietary interventions extend beyond the scope of the present review, interestingly, some restrictive diets have been shown to decrease gut microbiota diversity, despite the improvement in metabolic phenotype (168). Since a higher microbial diversity has consistently been associated with a healthier profile, and it may be worth investigating whether, despite the decrease in microbial diversity with these restrictive diets, the functional core and its metabolic capabilities are maintained.

Although there is no consensus on which type of gut microbiome-modulating diet will induce a better metabolic profile, studies in obesity and metabolic complications show that the baseline microbiota composition can significantly influence or predict individual responsiveness to metabolic improvements 169, 170. These findings are compatible with the emerging concepts of precision health, aimed not only at unraveling the structure and function driving an individual's obesity, but also at the design of future dietary interventions.

Studies in animal models, where fecal microbiota transplantation has been shown to modify phenotypes, including adiposity or metabolic complications, suggest a promising value for clinical application via directed manipulation of the microbiome (112). Unfortunately, in humans, there is only one small clinical trial on fecal microbiota transplantation from lean to obese subjects with MetS; transplantation resulted in improved insulin sensitivity, increased gut microbial diversity, and butyrate-producing bacteria (R. intestinalis) in obese recipients (171). Hopefully, other ongoing clinical trials will be able to corroborate whether fecal transplantation could be a successful strategy (172).

The worldwide use of bariatric surgery as a treatment for morbid obesity has increased dramatically in recent years, due to its long-term benefits (173). The Roux-en-Y gastric bypass (RYGB) is one of the most effective procedures for long-term weight loss and metabolic improvement in these patients 174, 175. The precise mechanisms by which bariatric surgery causes sustained weight loss and resolves obesity-related metabolic comorbidities have not been fully understood, however it has recently been suggested that gut microbiota changes could play a role 176, 177, 178. Interestingly, an increase in bacterial diversity, an acknowledged biomarker of the host's health (87), has been consistently observed after RYGB surgery and associated with an improved metabolic profile 179, 180. In addition, specific changes at different taxonomic levels have been observed, e.g. an increased abundance of Proteobacteria and Escherichia, among others 176, 177, 178, 179, 180, 181, 182, 183. So far, it is unclear whether these changes are caused by the direct effects of the surgery or mediated by surgery-induced weight loss. Interestingly, germ-free mice receiving gut microbiota transplants from post RYGB individuals, with successful weight loss showed less fat accumulation than mice transplanted with gut microbiota from obese individuals (177). While these findings support a causal relationship between gut microbiota and weight loss in individuals after RYGB (176), a direct link between the surgical process and gut microbiota changes has been proposed, regardless of weight loss. For instance, reduced availability of nutrients could induce bacteria to develop a more diverse set of genes to satisfy energy demands, explaining why calorie restriction could lead to metabolic improvement, even before significant weight loss occurs 55, 80. However, animal models have shown that calorie restriction itself does not replicate gut microbiota changes induced by bariatric surgery (176). Therefore, changes in stomach pH, bile acid composition and gastrointestinal motility could also play a role in surgery-induced microbiome variation (184). The mechanisms involved in metabolic improvement associated with gut microbiota function remain unclear. A limited number of studies have assessed the functional effects of gut microbiota and their metabolome in the host, reporting increased functional annotations related with the use of amino acids, carbohydrate and fatty acid metabolites 176, 177, 183.

Hence, the generation of further knowledge on the composition and functionality of the gut microbiota related to weight-loss and remission of metabolic complications will be key in the development of therapeutic strategies that allow the modulation of the intestinal microbiota. It will help identify individuals who could benefit from specific interventions, such as nutritional, pharmacological and/or surgical therapy, to reduce weight and improve metabolic complications. This will allow the development of precision medicine for the treatment and management of patients with obesity.

Section snippets

Conclusions

The gut microbiota has an enormous metabolic capacity able to influence multiple metabolic pathways in the host, and to thus contribute to obesity. In recent years, the number of large-scale studies of gut microbiota composition has grown significantly, and recent reports are increasingly accounting for the confounders between gut microbiota and host physiology. This has allowed the identification of microbial features associated with obesity across different populations, and of interactions

Acknowledgments

We are thankful to Alejandro Rodriguez for his assistance to elaborate the figures. This work was supported by CONACYT grant #248765 to SM-R.

References (185)

  • D.A. Russell et al.

    Metabolic activities and probiotic potential of bifidobacteria

    Int J Food Microbiol

    (2011)
  • F. Drissi et al.

    Metabolic role of lactobacilli in weight modification in humans and animals

    Microb Pathog

    (2017)
  • T. Dror et al.

    Microbiota manipulation for weight change

    Microb Pathog

    (2017)
  • W.R. Russell et al.

    High-protein, reduced-carbohydrate weight-loss diets promote metabolite profiles likely to be detrimental to colonic health

    Am J Clin Nutr

    (2011)
  • J.K. Goodrich et al.

    Human genetics shape the gut microbiome

    Cell

    (2014)
  • H. Xie et al.

    Shotgun metagenomics of 250 adult twins reveals genetic and environmental impacts on the gut microbiome

    Cell Syst

    (2016)
  • W.A. Walters et al.

    Meta-analyses of human gut microbes associated with obesity and IBD

    FEBS Lett

    (2014)
  • J.K. Goodrich et al.

    Genetic determinants of the gut microbiome in UK Twins

    Cell Host Microbe

    (2016)
  • Organization WH

    Fact sheet: Obesity and Overweight 2016

  • P.G. Kopelman

    Obesity as a medical problem

    Nature

    (2000)
  • T. Bhurosy et al.

    Overweight and obesity epidemic in developing countries: a problem with diet, physical activity, or socioeconomic status?

    ScientificWorldJournal

    (2014)
  • S.J. van Dijk et al.

    Epigenetics and human obesity

    Int J Obes (Lond)

    (2015)
  • P.W. Franks et al.

    Exposing the exposures responsible for type 2 diabetes and obesity

    Science

    (2016)
  • J.O. Hill et al.

    Obesity and the environment: where do we go from here?

    Science

    (2003)
  • G. Musso et al.

    Obesity, diabetes, and gut microbiota: the hygiene hypothesis expanded?

    Diabetes Care

    (2010)
  • V. Nehra et al.

    Gut microbiota: modulation of host physiology in obesity

    Physiology (Bethesda)

    (2016)
  • P.B. Eckburg et al.

    Diversity of the human intestinal microbial flora

    Science

    (2005)
  • J. Li et al.

    An integrated catalog of reference genes in the human gut microbiome

    Nat Biotechnol

    (2014)
  • R. Wall et al.

    Role of gut microbiota in early infant development

    Clin Med Pediatr

    (2009)
  • J.M. Rodriguez et al.

    The composition of the gut microbiota throughout life, with an emphasis on early life

    Microb Ecol Health Dis

    (2015)
  • T. Yatsunenko et al.

    Human gut microbiome viewed across age and geography

    Nature

    (2012)
  • E.B. Hollister et al.

    Structure and function of the healthy pre-adolescent pediatric gut microbiome

    Microbiome

    (2015)
  • T. Ringel-Kulka et al.

    Intestinal microbiota in healthy U.S. young children and adults–a high throughput microarray analysis

    PLoS One

    (2013)
  • M.G. Dominguez-Bello et al.

    Delivery mode shapes the acquisition and structure of the initial microbiota across multiple body habitats in newborns

    Proc Natl Acad Sci USA

    (2010)
  • S. Kuhle et al.

    Association between caesarean section and childhood obesity: a systematic review and meta-analysis

    Obes Rev

    (2015)
  • M.B. Azad et al.

    Gut microbiota of healthy Canadian infants: profiles by mode of delivery and infant diet at 4 months

    CMAJ

    (2013)
  • J. Penders et al.

    Factors influencing the composition of the intestinal microbiota in early infancy

    Pediatrics

    (2006)
  • L.A. David et al.

    Diet rapidly and reproducibly alters the human gut microbiome

    Nature

    (2014)
  • P.J. Turnbaugh et al.

    The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice

    Sci Transl Med

    (2009)
  • C. De Filippo et al.

    Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa

    Proc Natl Acad Sci USA

    (2010)
  • M. Arumugam et al.

    Enterotypes of the human gut microbiome

    Nature

    (2011)
  • G.D. Wu et al.

    Linking long-term dietary patterns with gut microbial enterotypes

    Science

    (2011)
  • M.J. Claesson et al.

    Gut microbiota composition correlates with diet and health in the elderly

    Nature

    (2012)
  • L.C. Kong et al.

    Dietary patterns differently associate with inflammation and gut microbiota in overweight and obese subjects

    PLoS One

    (2014)
  • G. Falony et al.

    Population-level analysis of gut microbiome variation

    Science

    (2016)
  • D. Graf et al.

    Contribution of diet to the composition of the human gut microbiota

    Microb Ecol Health Dis

    (2015)
  • L. Dethlefsen et al.

    Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation

    Proc Natl Acad Sci USA

    (2011)
  • L.M. Cox et al.

    Antibiotics in early life and obesity

    Nat Rev Endocrinol

    (2015)
  • L. Trasande et al.

    Infant antibiotic exposures and early-life body mass

    Int J Obes (Lond)

    (2013)
  • A.P. Ville et al.

    Early Antibiotic Exposure and Risk of Childhood Obesity in Latinos

    Child Obes

    (2017)
  • Cited by (61)

    • A Tour-d’Horizon of microbiota therapeutics for metabolic disorders

      2023, Microbiome Therapeutics: Personalized Therapy Beyond Conventional Approaches
    • Beneficial role of gut microbiome in metabolic syndrome, obesity, and cardiovascular diseases

      2022, Microbiome, Immunity, Digestive Health and Nutrition: Epidemiology, Pathophysiology, Prevention and Treatment
    View all citing articles on Scopus
    View full text