Food: a new form of personalised (gut microbiome) medicine for chronic diseases?

Abstract

Filling in the knowledge gaps between what we eat and the diseases we develop may lie in our guts, literally. The human large intestine houses the largest reservoir of microorganisms in or on the human body. With a 100-fold greater gene count than humans, the gut microbiome has huge potential to place a large metabolic burden (or advantage) on its host. The number of diverse gut microbial species is diminished in nearly all modern chronic conditions studied. The ‘Western diet’, rich in animal protein, fats and artificial additives, and lacking in fibre, beneficial microbes, plant phytochemicals, vitamins and minerals, is thought to drive these conditions by encouraging gut dysbiosis. Evidence from recent dietary intervention studies suggest adopting a plant-based, minimally processed high-fibre diet may rapidly reverse the effects of meat-based diets on the gut microbiome. However, recent work has shown that individual diet responses may be complicated by host genetics and the wide variation in the gut microbiome. Now that we measure genes and microbes more accurately, we are embarking on an exciting era of using both food and microbes as potential therapies.

Introduction

For decades, we have believed that improving our diets may reduce our risk of dying of heart disease and cancer. However, the details have eluded us. The answer to this problem may lie in our guts, literally. With interest surging and large-scale crowd-sourced projects burgeoning (www.AmericanGut.org, www.BritishGut.org), we are finding that the presence of certain microorganisms and how many different species (‘richness’) within our guts are modified in a range of conditions. At a macrolevel, reduced gut bacterial ‘diversity’ (like richness refers to the species count though also takes into account the abundances of the species in the sample), a marker of ‘dysbiosis’ (microbial imbalance), has been found in autoimmune diseases and cardio-metabolic conditions as well as accelerated aging. It has been recognised that the largest influence on our gut microbial profile comes back to what we are habitually eating ourselves and essentially feeding to our microscopic friends. Here, we discuss the current literature surrounding the dietary manipulation of this largely ignored ‘organ’ – the microbiome.

Methods

We examined key papers and reviews on the influence of diet on the microbiome. We identified important primary studies (through PubMed, Medline, Google and Google Scholar) which related to the topics supporting the main theme of our paper: (1) early dietary influences on the infant gut microbiome; (2) patterns in diet and their components which drive differences in the gut microbiome and their benefit or detriment to the host; (3) the importance of inter-individual variation in influencing response to dietary interventions. Although this is not a systematic review, we summarised findings to provide a narrative accessible to those with limited pre-existing knowledge of the topic.

Microbiome: what is it?

The ‘microbiome’ refers to the collective genomes of the microorganisms within an environmental niche, whereas ‘microbiota’ are the microorganisms themselves. Quantifying the microbiome most commonly involves genome-based methods including targeted amplicon (primarily 16S ribosomal ribonucleic acid sequencing) or shotgun metagenomics. In terms of our human genomes, we are practically identical – up to 99.5%, comparatively our microbiomes contain a 100-fold greater gene count¹ though we share only an estimated 10–20% of our microbial species.

By far the largest source of microorganisms present on the human body is within the large intestine. No characteristic ‘normal’ gut microbiome has been established, though trends have been found from infancy to young childhood.

Diet and the gut microbiome from infancy to young childhood

Colonisation of the gastrointestinal tract is essential for regulating the development of gut mucosal immunity. Seeding occurs during delivery through vaginal-faecal-oral microbial transfer with vaginal birth and transfer from the mother and environment. Infants delivered via Caesarean section have been found to have gut microbiomes very different to their mothers, with skin, oral and environmental microbes present during delivery being the first colonisers.² Even prior to birth, emerging evidence suggests that the maternal diet³ and probiotics⁴ given during pregnancy may influence the fetal microbiome.

In addition to key beneficial bacteria (believed to come from the mother’s gut), human milk provides the infant with bioactive molecules such as proteins, antibodies, growth factors and glycans, which modulate the composition and function of the developing gut microbiota. Bifidobacterium comprises 70–80% of the fecal microbiota of breast-fed infants and appears to be responsible for many of the promoted benefits.⁵ An important linkage between humans and Bifidobacterium is indicated by the presence of complex carbohydrate structures (glycans) in human milk which are indigestible by humans though Bifidobacterium contain specific genes encoding enzymes which break down glycans as a primary food source.⁶ Caesarean babies lack key Bifidobacterium and may therefore not fully exploit the benefits of breast milk. The cessation of breastfeeding, co-occurring with food introduction, appears to drive the stabilisation into an adult gut microbiome profile² (dominated by the phyla Bacteriodetes and Firmicutes) within 18 to 36 months. Changes into young childhood following this period have not yet been well characterised but are likely not as pronounced and dependent on the introduction of different foods.

Dietary patterns as a primary driver of differences in the adult microbiome

Species diversity and richness have been found to be around one-third lower in Americans than Malawians and Amerindians.⁷ The ‘Western diet’-containing foods abundant in animal protein, fats, artificial additives and reduced fibre, naturally present beneficial microbes, plant phytochemicals, vitamins and minerals are thought to drive differences. Dietary fibre consumption, which is high on non-Western plant-based diets, is particularly important and promotes stability in richness in humans.⁸

At the phylum level, the adult microbiome is dominated by either the Firmicutes or Bacteroidetes. Early studies have suggested a high Firmicutes to Bacteroidetes ratio was a marker for obesity – though this has not been confirmed by meta-analyses.⁹ At the genus level, three basic subgroups (enterotypes) have been described in healthy adults, independently dominated by either Bacteroides (type 1), Prevotella (type 2) or Ruminococcus (type 3) genera, although others dispute these subtypes. One study showed groups that consume long-term diets high in carbohydrates and plant polysaccharides (a type of fibre; including vegetarians and vegans) show high levels of Prevotella, whereas individuals consuming animal-derived fats clustered into the Bacteroides enterotype.¹⁰ Analyses at lower taxonomic levels (examining species or operational taxonomic units (OTUs)) have confirmed microbiome shifts following radical dietary changes based primarily on plants or meat.

Primary differences in the microbiomes found between cultures consuming plant- versus animal-based diets are being confirmed by dietary intervention studies.^11,12 Following a five-day dietary intervention, one showed increased abundances of bacteria that ferment fibre to produce butyrate such as Roseburia, Eubacterium rectale and Ruminococcus bromii.¹¹ Whereas on the animal-based diet where increased bile salts are required, the abundance of microorganisms which survive bile, including Alistipes, Bilophila and Bacteroides was increased.¹¹ Furthermore, some foodborne microbes derived from animal (e.g. lactic acid bacteria for curing meats) or plant-based (e.g. Rubus chlorotic mottle virus which infects spinach) diets in humans appeared to colonise the gut.¹¹ Likewise, there are anecdotal reports of soil microbes in human stool samples from people ingesting organic vegetables.

Functional studies into the gut microbiomes of mammalian carnivores and herbivores have shown clear differences.¹³ For instance, glutamate synthase enzyme commissions (ECs) were increased in herbivores, a trend found in Malawian and Amerindian adult microbiomes,⁷ whereas glutamine degradation enzyme commissions had increased representation in carnivores,¹³ a trend corresponding to US samples.¹³ Moreover, in mammalian carnivores, gut microbiomes have increased capacity for degrading amino acids likely due to an abundance of protein, whereas herbivores have increased phosphoenolpyruvate activity to produce aromatic amino acids when protein supply is limited,¹³ both confirmed via dietary intervention in humans.¹¹

Dietary factors that modulate the gut microbiota

Gut bacteria benefit the host by improving immune function, fermenting dietary fibres to short-chain fatty acids (SCFAs) that provide energy to colonocytes and producing essential nutrients including vitamin K and B vitamins (biotin, B₁₂, thiamin and riboflavin) and antioxidants. Beneficial bacteria also keep harmful organisms and gut dysbiosis at bay, protecting individuals from illness.

Probiotic foods

The term ‘probiotic’ refers to a colonic commensal known to have beneficial effects on health (such as Bifidobacterium and Lactobacillus acidophilus added to yoghurt). Probiotics are commonly produced as freeze-dried microbe supplements or provided live in food. Probiotics have been postulated to improve a whole host of maladies although many claims come from rodent studies and have yet to be fully supported by evidence in humans. A recent systematic review has highlighted that probiotics may have limited efficacy on body weight reduction.¹⁴ The evidence is best for the very young, elderly or those on antibiotics where the microbiome is already disrupted. Other natural probiotic foods include fermented products such as cheese, kefir and pickles (like kimchi or sauerkraut).

Prebiotic foods and fibre

A ‘prebiotic’ refers to food components which are not digested fully by the human body but nourish beneficial colonic microorganisms (‘microbial fertilisers’). Prebiotics generally consist of undigestable dietary fibre that may be added to foods (‘supplemented’) or naturally occurring in plant foods. Fibre is categorised into two forms: ‘soluble fibre’ (e.g. pectin in apple skin, raffinose in legumes and beta-glucan in oats), which dissolves in water to become viscous and is readily fermented to short-chain fatty acids (butyrate, propionate and acetate) and gases by colonic bacteria; and ‘insoluble fibre’ (e.g. cellulose in all plants, hemicellulose in grains and lignin in dry land plants), which does not dissolve in water, most types absorb water and provide bulk to stool, easing defecation. Only some forms of insoluble fibre are fermented to short-chain fatty acids by bacteria. Major short-chain fatty acids include butyrate (the primary fuel for colonocytes), propionate (used by the liver) and acetate (metabolised by peripheral tissues). Short-chain fatty acids have been found to be protective of diet-induced obesity and insulin resistance,¹⁵ and butyrate and propionate, but not acetate, have been found to control gut hormones and reduce food intake in mice.¹⁵ Short-chain fatty acids are key for microbial interactions with the immune system and provide around 6% of our energy. As well as most high-fibre foods being good for microbes – vegetables containing high levels of the soluble fibre inulin (e.g. artichokes, chicory, leeks, onion and garlic) appear particularly beneficial.

Polyphenols in food

Chemicals found in plants appear beneficial once metabolised by host bacteria¹⁶ – the most well-known are polyphenols. Many of these are contained in foods integral to the Mediterranean diet (extra virgin olive oil, red wine, onions, garlic, coffee, nuts and seeds, fruit and berries) which may explain some of its benefits. Polyphenols have long been recognised as having high antioxidant potential in vitro, though in vivo large differences in response and bioactivity are seen, potentially as a result of the host gut microbial profile.¹⁶ By modulating the gut microbiome profile, polyphenols such as those from pomegranate¹⁷ have been shown to be protective of obesity and associated inflammation in animals fed high-fat diets. Identifying key polyphenols that specifically nourish and increase certain beneficial microbes is a key future area for research.

Bile acids

Decades-old animal models have shown that a chronic high animal fat/low-fibre diet requires increased bile acids for digestion. Microorganisms that tolerate bile well are abundant in those consuming an animal-based diet.¹¹ A key role of dietary fibre, in particular soluble fibre such as beta-glucan found in oats and pectin, is to bind bile acids preventing their reabsorption and indirectly lower cholesterol. The gut bacteria play a central role in the deconjugation of bile acids, and it has recently been shown that the gut microbiota interact with the farnesoid X receptor (FXR) by metabolising bile acids to promote diet-induced obesity and metabolic disorder using Fxr−/− mice.¹⁸ Some tailored pro- and prebiotics, often in combination (‘synbiotics’), are being trialled in humans to reduce cholesterol.

Food additives and chemicals

Processed food additives (like emulsifiers and artificial sweeteners) may contribute to gut dysbiosis and metabolic disease development. Emulsifiers are used in processed foods to create oil-in-water emulsions preventing the fat and liquid components from separating. Commonly found emulsifiers include soy and egg yolk lecithin used in salad dressings. Emulsifiers have the potential to emulsify and erode the protective mucosal layer of the gastrointestinal tract, increasing inflammation and promoting dysbiosis.¹⁹ Most non-caloric artificial sweeteners are not digested by humans and therefore pass through to the gut. Feeding saccharin to mice increased the abundance of Bacteroides and under-representation of Clostridiales and induced glucose intolerance,²⁰ suggesting that sweeteners are not inert or harmless for the gut microbiota.

Meat components: Choline and carnitine

Choline and carnitine are components found in high quantities in animal products which have essential roles in lipid metabolism. The differential microbial metabolism of choline and carnitine to trimethylamine N-oxide has emerged as a factor in why some but not all meat consumers develop atherosclerosis.²¹ Moreover, the development of non-alcoholic fatty liver disease under choline deficiency, a long known mode of non-alcoholic fatty liver disease development, has been shown to occur by the conversion of choline to methylamines by gut microbiota, which was dependent on subject abundances of pathogenic Gammaproteobacteria and Erysipelotrichi.²²

Contribution of gut microbial dysbiosis to obesity and metabolic disease

The gut microbiome now appears to play a core role in the development, aggravation and sustainability of obesity, supported by the transplantation of gnotobiotic (germ free or with known strains of microorganisms) mice with faecal microbiomes from obese humans.²³ The mechanisms by which this occurs are still unclear, and much of the evidence comes from rodent models treated with a high-fat, low-fibre diet. Briefly, subjects with a microbiome predisposed to weight gain house increased abundances of microbes with the capacity for producing short-chain fatty acids, increasing caloric availability to the liver and tissues (‘energy harvest’).²³ Elevated levels of faecal short-chain fatty acids have been confirmed in obese compared to lean subjects coupled with an increase in the ratio of Firmicutes to Bacteroides/Prevotella.²⁴

It also appears gut microbiota dysbiosis may promote Western diet-induced obesity and metabolic complications such as the metabolic syndrome and non-alcoholic fatty liver disease through encouraging a pro-inflammatory and permeable gut mucosa by producing lipopolysaccharides (endotoxins), which may then enter portal circulation.²⁵ Though how much food itself contributes to the inflammatory state rather than the gut microbiome is as yet unclear.

The obese phenotype is probably aggravated from a nutrient imbalanced diet (high-fat, low-fibre, highly processed) allowing pathogenic bacteria to flourish combined with the avoidance of whole, minimally processed fruits, vegetables and grains high in fibre, phytochemicals and natural beneficial microflora preventing beneficial microorganism growth. However, how this diet achieves dysbiosis in some people and not others is still unclear and genetic influences on the host microbes²⁶ and metabolism may still be important.

Changing microbes via diet

Until recently, it was thought it took months to slightly alter gut microbiota via diet; however, recent studies have shown this can occur within days. For instance, when African Americans and rural Africans switched diets for only two weeks (the Americans adopted an African diet low in fat and high in plants and the Africans adopted an American diet high in animal fats), remarkable differences were found.¹² Following ‘Africanisation’, butyrate production increased 2.5 times with increased abundances of known butyrate-producing bacteria, and reduced secondary bile acid synthesis. ‘Westernisation’ reduced butyrate-producing microbes which accompanied a 50% reduction in butyrate production, increased urinary choline and pro-inflammatory bacterium Bilophila wadsworthia.¹² Another intervention comparing extreme plant- versus animal protein-based shifts mirrored these changes after only five days.¹¹

The timing of food may also be important. In mice, restricting high-fat feeding to an 8-h window per day increased microbiome diversity and abundances of protective bacteria compared to ad libidum feeding although total intakes were similar.²⁷ In particular, the abundance of Akkermansia increased.²⁷ Akkermansia muciniphila breaks down mucin and as a result may be beneficial for intestinal barrier integrity and associated inflammation. Inreased abundances of Akkermansia muciniphila have been associated with improved metabolic status following a caloric-restricted diet in humans.²⁸ Where the microbiome community is stable, it may be difficult to remedy damages through diet alone. Longer-term diet changes may be needed but will reduce compliance, it is therefore essential to promote preventative diets.

Microbiome and diet genetics: the value of the co-twin control method in microbiome research

Using data from 416 twin pairs, our group found gut microbiome phenotypes ranged from 0 to 39% heritable (referring to the degree to which genetic makeup contributes to the variance in a trait),²⁶ providing evidence that genetic variation influences gut microbial profiles. Examining identical twins that are matched for age, sex and the baseline genetic sequence and discordant for a behaviour or condition allows for a perfect case–control design; deemed the ‘co-twin control method’. In one such study of Finnish identical twin pairs concordant and discordant for BMI, intakes of energy and nutrients (monounsaturated and n-3 and n-6 polyunsaturated fatty acids, and soluble fibre) were more influential on microbial populations than the BMI group,²⁹ confirming the importance of diet, independent of genetics. Our group has performed pilot studies of high-fibre diets in identical twins, the preliminary results suggesting we can increase levels of diversity and beneficial microbes like Christensenella in a few weeks but larger studies are needed.

Can the microbiome inform personalised medicine and diets for the future?

Clearly a lifetime diet high in fruits, vegetables and whole grains enriched with probiotics, polyphenol rich foods and fibres is healthy for our gut microbes and is in line with longstanding dietary advice. However, it is becoming clear that we need to emphasise diversity of foods more as well as a personalised approach. Recently, Zeevi et al.³⁰ used machine-learning models including 16S ribosomal ribonucleic acid and shotgun metagenomic gut microbiome profiling to predict the different postprandial glucose response of 800 individuals to the same foods. Their algorithms significantly predicted individual postprandial glucose response better than carbohydrate counting or total caloric content of meals consumed. The authors hope to personalise meal choices based on gut microbiome profiles to reduce diabetes risk. With large populations (www.AmericanGut.org, www.BritishGut.org) now funding the study of their own microbiomes via crowd-funding combined with new discoveries, we are entering an exciting era of being able to modify health with food and measure the effects via our microbes.

Declarations

Competing Interests

TDS is an NIHR Senior Investigator.

Funding

TwinsUK was funded by the Wellcome Trust; European Community’s Seventh Framework Programme (FP7/2007-2013). The study also receives support from the National Institute for Health Research (NIHR) BioResource Clinical Research Facility and Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London.

Guarantor

TDS

Ethical approval

Not applicable

Contributorship

All authors contributed, read and approved the final version.

Acknowledgements

The authors would like to acknowledge the reviewer of this manuscript for their valuable comments.

Provenance

Not commissioned; peer-reviewed by Philiip Marsh.