Fibre and body weight

One food macro-component that differentiates current Westernised diets from those of the hunter–gatherer is the quantity and quality of dietary fibre. It has been estimated that the hunter–gatherer Palaeolithic diet delivered more than 100 g fibre per d⁽Reference Eaton and Leeds¹⁴⁾. Yet, the current average American diet contains only 15 g fibre per d, just half of what current US recommendations deem adequate⁽¹⁵⁾. Moreover, the type of fibres forming the larger proportion of current dietary intake is likely to be different too. The diet of the hunter–gatherer contained a greater proportion of fibre from fruits and vegetables which ferments to a greater extent than the cereal grain fibres which prevail in current Western diets⁽Reference Eaton¹⁶⁾. Agricultural, animal husbandry and industrial reforms have all contributed towards this dietary shift⁽Reference Eaton and Konner¹⁷⁾.

There is evidence to suggest that dietary manipulation to increase dietary fibre intake may be protective against obesity⁽¹⁸⁾. Epidemiological studies have highlighted that diets rich in fibre are inversely correlated with body weight, BMI⁽Reference Appleby, Thorogood and Mann^19–Reference Maskarinec, Takata and Pagano²¹⁾ and adiposity⁽Reference Nelson and Tucker^22, Reference Kromhout, Bloemberg and Seidell²³⁾. Results from weight-management intervention studies using high-fibre foods and supplements are also encouraging⁽Reference Ryttig, Tellnes and Haegh^24–Reference Tucker and Thomas²⁶⁾. It appears that fibre acts via numerous mechanisms to increase levels of both satiation (the satisfaction of appetite that develops during the course of eating which leads to meal termination) and satiety (the state in which further eating is inhibited and occurs as a consequence of having eaten). A classic human study showed that different apple preparations had different effects on appetite despite a constant rate of ingestion and isoenergetic portions. Subjects reported that a whole apple made them feel significantly fuller and suppressed future hunger more than apple purée, which in turn was more satiating than apple juice⁽Reference Haber, Heaton and Murphy²⁷⁾. The increased physical disruption and removal of fibre from whole fruit to juice may be linked to a reduced satiating action. Reduced mastication, reduced gastric distension and a more rapid glycaemic response may also be contributing factors. Processes such as grain milling and carbohydrate refinement have become an intrinsic part of food manufacturing since the Industrial Revolution. This perhaps explains the positive correlation between intake of refined products and weight gain in observational studies⁽Reference Liu, Willett and Manson²⁸⁾.

Not all fibre intervention studies have shown a positive correlation with weight loss or perceived satiety⁽Reference Howarth, Saltzman and McCrory²⁹⁾. However, as study protocols vary considerably and rely heavily on accurate food intake reporting, this is perhaps not surprising. Howarth summarised in 2001 that the majority of fibre-supplementation studies in energy-restricted subjects showed beneficial outcomes on satiety, reduced energy intake and body weight⁽Reference Howarth³⁰⁾. Furthermore, these effects seem to be preserved with ad libitum energy intake. Howarth calculated from ad libitum studies available that an additional fibre intake of 14 g/d (dietary fibre or fibre supplements) was associated with a 10 % reduction in energy intake and a mean body-weight loss of 1·9 kg over 3·8 months, with the greatest weight loss seen in obese individuals⁽Reference Howarth³⁰⁾. Most studies also investigated weight change as a primary outcome.

Fibre has been proposed to inhibit excess dietary energy intake by a number of mechanisms⁽Reference Heaton³¹⁾ including: (1) displacement of energy-dense foods from the diet; (2) inhibition of absorption of macronutrients at the intestinal surface; (3) an altered luminal environment; (4) an increased gastric distension. Further purported physical and hormonal effects of fibre have been reviewed in detail elsewhere⁽Reference Howarth^30, Reference Slavin³²⁾ and are summarised in Fig. 1. However, it is perhaps worth remembering that the term ‘fibre’ covers a vast array of substances and their differing chemical properties dictate their physiological effects. Non-fermentable fibres such us cellulose are believed to play a limited role in inhibiting excess energy intake. Whilst their non-nutritive nature allows them to cause distension and dilute the energy density of foods, in rodent studies at least, it is apparent that following a period of adaptation, rodents simply consume an increased food volume to match their previous energy intake⁽Reference Roy, Keenan and Zablah-Pimentel³³⁾. Fermentable carbohydrates (FC) appear to also have a role in reducing food intake and therefore aiding the control of body weight and body fat as shown in multiple animal FC-supplementation studies⁽Reference Keenan, Zhou and McCutcheon^34–Reference Perrigue, Monsivais and Drewnowski³⁶⁾. Whilst the evidence in human subjects is more controversial⁽Reference Howarth, Saltzman and McCrory²⁹⁾, FC or perhaps more specifically the products of colonic fermentation process do appear to play some part in the control of food intake and body weight.

Fig. 1

Schematical summary of the hormonal, intrinsic and colonic properties of dietary fibre and how they may reduce body weight by modulation of appetite⁽Reference Slavin³²⁾. (Reproduced by permission.)

SCFA

Recently, the colonic effects of fibre have come under fresh scrutiny. Accumulating evidence has suggested that the satiating effects of fibre may be due to SCFA, the major fermentation product of fibre⁽Reference Hamer, Jonkers and Venema³⁷⁾. By definition, all dietary fibre passes through the small intestine unaffected by human digestive processes. However, upon reaching the colon, anaerobic bacteria are able to degrade some dietary fibres via a fermentation process to yield SCFA. In general, the fermentability of soluble fibres by the colonic microbiota is much greater than that of insoluble fibres. Pectin, resistant starches, gums and polyfructans (for example, inulin) are the most highly fermented substrates, and for the purpose of the present review shall be collectively known as FC. FC yield metabolisable energy (about 2 kcal/g; 8 kJ/g), gases (CO₂, H₂ and CH₄) and SCFA from fermentation reactions. However, the production of SCFA is subject to vast inter-individual variation, largely dependent upon the type of FC, the host's microbiota and gut transit time (for a review of factors regulating SCFA production, see Macfarlane & Macfarlane⁽Reference Macfarlane and Macfarlane³⁸⁾).

SCFA consist of carbon chain length between C2 and C6. With insignificant quantities of SCFA coming directly from the diet, bacterial fermentation and endogenous synthesis are considered the major sources of SCFA in non-ruminant animals⁽Reference Ballard^39, Reference Pouteau, Nguyen and Ballevre⁴⁰⁾. Over 80 % of SCFA present in the human colonic lumen are in the form of acetate (C2), propionate (C3) and butyrate (C4) in the approximate molar concentrations of 57:22:21⁽Reference Cummings, Pomare and Branch⁴¹⁾. About 90 % of SCFA are rapidly absorbed in the colon; butyrate is almost entirely utilised by colonocytes as their preferred energy substrate⁽Reference Roediger⁴²⁾. Propionate is primarily removed by the liver⁽Reference Cummings, Pomare and Branch⁴¹⁾. Peripheral blood levels are therefore low under normal physiological conditions at 1–3 μm-butyrate and 4–5 μm-propionate. Acetate, however, passes more freely into the peripheral circulation and therefore higher plasma levels of up to 100–150 μm are seen⁽Reference Hong, Nishimura and Hishikawa⁴³⁾. This figure can be increased up to 250 % following the hepatic breakdown of ingested ethanol⁽Reference Siler, Neese and Hellerstein⁴⁴⁾.

It is well documented that butyrate plays a critical role in regulating colonocyte cell proliferation, differentiation⁽Reference Roediger^42, Reference Topping and Clifton⁴⁵⁾ and inflammatory response⁽Reference Topping and Clifton⁴⁵⁾. Acetate and propionate, however, are less desirable colonic substrates⁽Reference Roediger⁴⁶⁾, so their physiological roles are proposed to be more linked to carbohydrate and lipid metabolism⁽Reference Wolever, Spadafora and Eshuis⁴⁷⁾. Their complex relationship with hepatic lipid processing has been outlined elsewhere⁽Reference Wolever, Cummings and Rombeau^48, Reference Topping, Pant, Cummings, Rombeau and Sakata⁴⁹⁾. SCFA have been shown in animal models to be important mediators of colonic blood flow⁽Reference Kvietys and Granger⁵⁰⁾, fluid and electrolyte balance⁽Reference Binder and Mehta⁵¹⁾ and gastrointestinal motility⁽Reference Yajima⁵²⁾. More recently, however, they have also been proposed as key energy homeostasis signalling molecules, feeding back important information on the nutrient milieu of the colon.

SCFA receptors

Until 2003, the G-protein-coupled receptors GPR41 and GPR43 were considered orphaned. During routine screening with bioactive compounds or ‘ligand fishing’ it emerged that free fatty acid receptor 3 (FFA3; formerly GPR41) and FFA2 (formerly GPR43) were activated by acetate using Ca²⁺ mobilisation assays in transfected human embryonic kidney (HEK293T) and CHO-K1 cells⁽Reference Brown, Goldsworthy and Barnes^53–Reference Nilsson, Kotarsky and Owman⁵⁵⁾. In vitro activation also occurred at physiological doses with other SCFA including propionate and butyrate. FFA2 was found to have similar micromolar activation potencies for acetate, propionate and butyrate whilst FFA3 had activation in the following potency order: propionate>butyrate≫acetate⁽Reference Brown, Goldsworthy and Barnes^53, Reference Le Poul, Loison and Struyf⁵⁴⁾. The 100-fold lower activation potency of FFA3 to acetate can be used diagnostically to distinguish FFA3 from FFA2⁽Reference Brown, Goldsworthy and Barnes⁵³⁾. Whilst the receptors share endogenous ligands, it is apparent that their G-protein signalling mechanism differs. FFA2 and FFA3 can both signal using the pertussis toxin-sensitive G_i/o pathway, whilst data suggest that FFA2 also possesses the ability to signal using the G_q/11 pathway⁽Reference Le Poul, Loison and Struyf⁵⁴⁾. The physiological significance of this dual-coupled signalling mechanism present in FFA2 but not in FFA3 is not yet understood.

Free fatty acid receptor 2 and appetite regulation

The first suggestion that FFA2 may be implicated in appetite-regulatory processes was the identification of FFA2 mRNA expression in both rat whole-wall and separated mucosa samples from the distal ileum and colon⁽Reference Karaki, Mitsui and Hayashi⁵⁶⁾. These findings have since been confirmed⁽Reference Dass, John and Bassil⁵⁷⁾ and also reported in whole-wall samples of the human colon⁽Reference Karaki, Tazoe and Hayashi⁵⁸⁾. Recent quantitative real-time PCR anatomical profiling in mouse tissue by Regard et al. provides further comparative expression data highlighting the greatest expression levels of FFA2 in the bone marrow and islet cells of the pancreas. Considerable quantities were also found in the white and brown adipose tissue, spleen and the large intestine⁽Reference Regard, Sato and Coughlin⁵⁹⁾.

Peptide YY

In rats, FFA2 immunoreactive (IR) cells are completely co-localised with peptide YY (PYY) IR enteroendocrine L-cells of the gastrointestinal tract, with all IR FFA2-expressing cells also expressing PYY⁽Reference Karaki, Mitsui and Hayashi⁵⁶⁾. Although both neural and hormonal factors stimulate the release of PYY it is also likely that nutrients in the luminal environment induce L-cell peptide secretion of the peptide and others encoded by the proglucagon gene: glucagon-like-peptides (GLP) 1 and 2 (GLP-1, GLP-2) and oxyntomodulin (OXM)⁽Reference Greeley, Jeng and Gomez^60, Reference Adrian, Ferri and Bacarese-Hamilton⁶¹⁾. This co-localisation discovery in both the rat ileum and human colon may suggest that activation of FFA2 by SCFA ligands facilitates or modifies PYY secretion⁽Reference Karaki, Mitsui and Hayashi⁵⁶⁾.

PYY release is directly proportional to the energy content of an ingested meal⁽Reference Adrian, Ferri and Bacarese-Hamilton^61, Reference Pedersen-Bjergaard, Host and Kelbaek⁶²⁾ and plasma levels are reduced by fasting. The full length (PYY_1-36) and truncated (PYY_3-36) peptide forms are synthesised and secreted from L-cells throughout the entire length of the gastrointestinal tract with the highest concentrations found in the colon and rectum⁽Reference Adrian, Ferri and Bacarese-Hamilton⁶¹⁾. Lower concentrations are also seen in the CNS⁽Reference Broome, Hokfelt and Terenius⁶³⁾. PYY_3-36 preferentially binds to the Y2 receptor which has the highest levels of tissue expression in the arcuate nucleus of the hypothalamus, an important neuronal site in the central integration of appetite control⁽Reference Broberger, Landry and Wong⁶⁴⁾. Significant reductions in food intake have been demonstrated following peripheral administration of physiological levels of PYY_3-36 in rodents⁽Reference Batterham, Cowley and Small^65, Reference Chelikani, Haver and Reidelberger⁶⁶⁾ and in both normal-weight and obese human subjects⁽Reference Batterham, Cowley and Small⁶⁵⁾. Rodent data have suggested that PYY is likely to exert its hypophagic effect by increasing the expression of the anorexigenic pro-opiomelanocortin (POMC) neuropeptide populations and reducing the expression of neuropeptide Y within the arcuate nucleus⁽Reference Batterham, Cowley and Small⁶⁵⁾. Rodent vagotomy experiments have also highlighted the importance of the brain stem and the vagus nerve in the appetite-reducing effects of PYY⁽Reference Koda, Date and Murakami⁶⁷⁾.

Increased plasma PYY levels have been consistently seen in rats⁽Reference Keenan, Zhou and McCutcheon^34, Reference Delzenne, Cani and Daubioul^68, Reference Zhou, Martin and Tulley⁶⁹⁾ and human subjects⁽Reference Greenway, O'Neil and Stewart⁷⁰⁾ following FC supplementation. Moreover, direct infusion of SCFA into rabbit and rat colons has been shown to increase PYY secretion⁽Reference Longo, Ballantyne and Savoca^71, Reference Cherbut, Ferrier and Roze⁷²⁾, suggesting that SCFA derived from the fermentation process may be responsible. Whilst therapeutically encouraging, perhaps the more clinically relevant question is: are elevated plasma PYY levels resulting from FC supplementation sufficient in magnitude to bring about a change in appetite or body-weight regulation? Ingestion of lactitol (a fermentable sugar alcohol) significantly increased postprandial PYY plasma level response in rats, and chronic consumption decreased weight gain compared with normal chow controls. Acute supplementation of lactitol in human subjects was associated with a significant attenuation of postprandial PYY decline at 5 h and a trend towards reduced appetite⁽Reference Gee and Johnson⁷³⁾, suggesting that a physiological effect may be present after acute supplementation of relatively small doses (15 g) of FC.

Pre-proglucagon gene products

Enteroendocrine L-cells are also one of the major sites for expression of the proglucagon gene which encodes for GLP-1, GLP-2 and OXM. These are co-stored and co-secreted with PYY from enteroendocrine L-cells⁽Reference Kim, Carlson and Jang⁷⁴⁾. Therefore, it is possible that luminal SCFA also stimulate L-cell secretion of GLP-1 and other products of the pre-proglucagon gene via FFA2. GLP-1 is a potent anorexigenic hormone and incretin. It is released following nutrient ingestion in proportion to the energetic content of the meal⁽Reference Elliott, Morgan and Tredger⁷⁵⁾. Acute intracerebroventricular administration (icv) of GLP-1 to rodents is associated with a decline in short-term energy intake⁽Reference Turton, O'Shea and Gunn⁷⁶⁾ and a reduced body weight following repeated administration⁽Reference Davis, Mullins and Pines⁷⁷⁾. Intravenous injection of GLP-1 in both lean and obese human subjects produces a dose-dependent hypophagic effect⁽Reference Verdich, Flint and Gutzwiller⁷⁸⁾. The GLP-1 receptor is expressed in the pancreatic islets⁽Reference Regard, Sato and Coughlin^59, Reference Bullock, Heller and Habener⁷⁹⁾, reflecting its importance as an incretin. Appreciable quantities are also seen in the hypothalamus and brainstem⁽Reference Shimizu, Hirota and Ohboshi^80, Reference Goke, Larsen and Mikkelsen⁸¹⁾. The anorectic effects of GLP-1 are abolished following vagotomy in rodents, suggesting that the interplay between the vagus and hypothalamus may be mediating the peptide's anorexigenic effect⁽Reference Abbott, Monteiro and Small⁸²⁾.

OXM is co-secreted with PYY and GLP-1 from the entoendocrine L-cells following nutrient ingestion. Like GLP-1, peripheral and icv administration of OXM has also been shown to reduce food intake in both human subjects and rodents⁽Reference Cohen, Ellis and Le Roux^83–Reference Baggio, Huang and Brown⁸⁵⁾ and act as an incretin⁽Reference Schjoldager, Baldissera and Mortensen⁸⁶⁾. OXM is thought to signal via the GLP-1 receptor⁽Reference Baggio, Huang and Brown⁸⁵⁾, but the anorexigenic mechanism of OXM has yet to be fully characterised. Increased expression of the anorexigenic hypothalamic peptide α-melanocyte-stimulating hormone⁽Reference Dakin, Small and Batterham⁸⁴⁾ and suppression of the orexigenic hormone ghrelin⁽Reference Cohen, Ellis and Le Roux⁸³⁾ have been shown to result from peripheral OXM administration.

FC-supplementation studies have consistently been associated with increased colonic proglucagon mRNA expression⁽Reference Keenan, Zhou and McCutcheon^34, Reference Delzenne, Cani and Daubioul^68, Reference Zhou, Martin and Tulley^69, Reference Delmee, Cani and Gual^87–Reference Zhou, Hegsted and McCutcheon⁹⁰⁾. However, only some studies were able to corroborate these findings with increased circulating plasma GLP-1 levels⁽Reference Keenan, Zhou and McCutcheon^34, Reference Delzenne, Cani and Daubioul^68–Reference Greenway, O'Neil and Stewart^70, Reference Cani, Dewever and Delzenne^88, Reference Piche, des Varannes and Sacher-Huvelin⁹¹⁾. Others have shown no effect⁽Reference Gee and Johnson^73, Reference Robertson, Bickerton and Dennis^92, Reference Frost, Brynes and Leeds⁹³⁾. Nevertheless, the duration of supplementation is important to consider in the interpretation of results. Sufficient time must be given to allow for the adaptation of the gut microbiota to the additional FC within the diet for maximal fermentation⁽Reference May, Mackie and Fahey⁹⁴⁾. Few studies have investigated the effect of SCFA on plasma OXM levels. As suggested for PYY, if GLP-1 and OXM secretion could be enhanced via FFA2 activation, the potential for dietary or pharmacological manipulation to increase satiety could be used in the treatment of obesity. Equally, with specific relevance to GLP-1 and OXM, it may be useful in potentiating insulin secretion in patients with type 2 diabetes via their incretin action.

Fermentable carbohydrates and central appetite regulation

More recently it has been suggested that FC may not only affect acute satiety signals such as PYY and GLP-1, but may also exert a more chronic action on body-weight control by indirectly altering expression of hypothalamic neuropeptides. The arcuate nucleus contains two neuronal populations, which as components of the gut–brain axis act to regulate energy balance: the anorexigenic POMC and cocaine- and amphetamine-regulated transcript (CART) neurones and the orexigenic neuropeptide Y and agouti-related peptide neurones. In a recent study, Shen et al. fed one group of rats with a high-resistant starch diet (30 %, w/w) and another group was fed standard chow for 65 d. It was found that POMC mRNA expression was significantly increased in the resistant starch group compared with controls⁽Reference Shen, Keenan and Martin⁹⁵⁾. This effect was preserved in capsaicin-treated rats, suggesting that this effect was not vagally mediated. Although it is likely that SCFA pass across the blood–brain barrier via a carrier-mediated mechanism⁽Reference Oldendorf⁹⁶⁾, there is little or no apparent expression of FFA2 in the hypothalamus⁽Reference Regard, Sato and Coughlin⁵⁹⁾. Therefore, it may perhaps be concluded that any effects seen on hypothalamic neuropeptide expression are occurring as a consequence of increased levels of plasma PYY and GLP-1.

Free fatty acid receptor 2 activation as an energy homeostasis mechanism

Before the displacement of fibre-rich foods during the Industrial Revolution⁽Reference Cordain, Eaton and Sebastian⁹⁷⁾, excess food intake, i.e. a positive energy balance, would have been predictably associated with a proportional rise in fibre intake. Should our hypothesis regarding the functional role of FFA2 be correct, this resultant increase in SCFA produced in the colon leads to amplified FFA2 signalling. Up-regulation of PYY and GLP-1 expression and secretion and possibly other products of the proglucagon gene, for example OXM, would ensue. Fig. 2 illustrates that anorexigenic neural circuits are subsequently activated to reduce food intake and increase energy expenditure to restore the body back to neutral energy balance. Projection of this schematic to the current mode of Westernised, energy-dense diets disrupts this homeostatic cycle, as the inevitable parallel between energy and fibre density no longer exists. Although an increase in PYY and GLP-1 levels follows an increased energy intake, an enhanced secretion profile associated with the SCFA activation of FFA2 may be absent.

Fig. 2

A proposed role of fibre and free fatty acid receptor 2 (FFA2) activation in maintaining energy homeostasis. This projected negative-feedback-type mechanism assumes that an increase in food and energy intake within a traditional (fibre-rich) diet is associated with increased SCFA production. Thus, FFA2 is activated to increase anorexigenic signalling, increasing the secretion of peptide YY (PYY) and glucagon-like-peptide-1 (GLP-1). When a reduction in food is brought about, fibre intake consequently falls, and FFA2 activation is reduced.

SCFA and gastrointestinal motility

A negative-feedback mechanism known as the ‘ileal brake’ can occur following nutrient digestion⁽Reference Van Citters and Lin⁹⁸⁾; this is particularly true of fats. Various neural and hormonal signals act to slow gastrointestinal motility to enhance digestion and nutrient absorption whilst eliciting an anorexigenic effect. Yet it is apparent that SCFA may have a role in gastrointestinal motility beyond that of other fats as part of the ileal brake. The majority of work linking gastrointestinal motility and SCFA stems from ruminant animal studies⁽Reference Kendall and McLeay⁹⁹⁾ where, due to differences in gut physiology, SCFA production is much greater⁽Reference Bergman¹⁰⁰⁾. However, there is some evidence in non-ruminants that SCFA may regulate overall transit time of digesta through the large intestine⁽Reference Dass, John and Bassil^57, Reference Tazoe, Otomo and Kaji¹⁰¹⁾ and potentially contribute to appetite regulation. Cherbut hypothesises that there are three possible mechanisms eliciting this response: (1) SCFA stimulation of vagal nerves in the gut; (2) a direct effect of SCFA on intestinal smooth muscle tone; or (3) as a consequence of the indirect changes in secretion of PYY and other regulatory peptides⁽Reference Cherbut, Cummings, Rombeau and Sakata¹⁰²⁾.

Perhaps given less consideration is that FFA2 may be partly controlling this effect on gastrointestinal motility. Recent findings have indicated aside from co-localisation with PYY, SCFA stimulation of FFA2 may also affect gastrointestinal 5-hydroxytryptamine (5-HT) release. 5-HT, or serotonin, is a neurotransmitter in the CNS known to modulate mood, behaviour and appetite⁽Reference Berger, Gray and Roth¹⁰³⁾. Whilst the CNS actions of 5-HT are the most documented, 95 % of endogenous 5-HT is found peripherally in the gastrointestinal tract, primarily in enterochromaffin cells but also in 5-HT-containing mucosal mast cells⁽Reference Kim and Camilleri¹⁰⁴⁾. Mechanical and chemical stimuli during nutrient ingestion (including by SCFA) in the gut cause 5-HT to be dose-dependently released⁽Reference Zhu, Wu and Owyang^105–Reference Fukumoto, Tatewaki and Yamada¹⁰⁷⁾. Activation of the various 5-HT receptor sub-types has opposing effects on gastrointestinal motility and gastric distension. Karaki et al. have suggested that FFA2-IR cells share immunoreactivity with 5-HT-containing mucosal mast cells of the rat distal ileum and colon⁽Reference Karaki, Mitsui and Hayashi⁵⁶⁾. Therefore, SCFA activation of FFA2 may mediate the release of gut 5-HT and herald a role for FFA2 in gastric motility-mediated appetite regulation, independent of that of PYY and products of the proglucagon gene. The co-localisation of FFA2-IR cells with 5-HT-releasing cells in the human colon has yet to be confirmed⁽Reference Karaki, Tazoe and Hayashi⁵⁸⁾.

Free fatty acid receptor 2 actions and the metabolic syndrome

Independent of the effect on appetite-regulatory pathways, FFA2 may also improve certain metabolic derangements seen as part of the metabolic syndrome. The metabolic syndrome is used to describe a cluster of cardiovascular risk factors including central obesity, impaired glucose tolerance, dyslipidaemia and hypertension⁽Reference Grundy, Brewer and Cleeman¹⁰⁸⁾. Excess fat mass, characterised by both adipocyte hyperplasia and hypertrophy, is thought to partly explain why obese individuals have a greater disposition to exhibit the various components of the metabolic syndrome.

Adipose tissue is not just an energy reservoir for TAG; instead, it is now considered a complex endocrine organ. It is the site of secretion for a host of hormones, inflammatory markers and cytokines, some of which act as peripheral indicators of long-term energy balance, for example leptin⁽Reference Tilg and Moschen¹⁰⁹⁾. Moreover, the morphology of adipocytes seems to be key factor in physiological adaption to positive energy balance. The lipotoxicity theory suggests that obesity is a state whereby adipocytes become engorged with TAG, thereby reducing the buffering capacity of adipocytes to uptake TAG in the postprandial state⁽Reference Goossens¹¹⁰⁾. Consequently, TAG remain elevated and TAG deposition begins to occur in ectopic functional tissue, for example, the liver, pancreas or skeletal muscle. This is detrimental to functionality, glucose handling and insulin sensitivity of the cells.

Following the discovery of FFA2 expression in adipocytes, Hong et al. conducted a series of studies to try and elucidate a possible role of FFA2 in adipocyte functioning⁽Reference Hong, Nishimura and Hishikawa⁴³⁾. It was first shown that adipocyte expression of FFA2 in mice fed high-fat diets was significantly greater than that of their standard chow-fed counterparts. In 3T3-L1 cells it was found that FFA2 mRNA and PPARγ2 (a marker of preadipocyte differentiation) were increased following treatment with the SCFA acetate and propionate. Increased oil red O staining demonstrated increased fat accumulation in the cells, which was abolished with the down-regulation of FFA2 with antisense small interfering RNA. Taken together, this suggests that FFA2 and activation by SCFA may be critical in the differentiation of adipocytes, their capacity to store TAG and perhaps prevent the deposition of ectopic fat⁽Reference Robertson¹¹¹⁾.

In normal health, postprandial insulin release should herald a shift in substrate utilisation from NEFA to glucose and therefore inhibit lipolysis and NEFA release⁽Reference Frayn and Frayn¹¹²⁾. However, cells that have become insulin resistant (possibly due to ectopic fat deposition) do not respond to this cue appropriately and continue to preferentially use NEFA as their energy source, causing a cluster of metabolic disturbances. FFA2 may be implicated in improving this metabolic disruption by reducing the flux of NEFA from adipocytes. In vitro, acetate and propionate were able to suppress isoproterenol-induced lipolysis in a dose-dependent manner⁽Reference Hong, Nishimura and Hishikawa⁴³⁾. In agreement, arterio-venous sampling by Robertson et al. demonstrated in human subjects that acetate and propionate act directly on subcutaneous abdominal adipocytes to reduce NEFA flux from adipocytes⁽Reference Robertson, Bickerton and Dennis⁹²⁾. More recently it has been suggested that FFA2 mediates this effect. Acetate dose-dependently reduced the release of NEFA from wild-type mouse adipocytes in vitro whilst, in vivo, intra-peritoneal (IP) injection of sodium acetate at 500 mg/kg to wild-type mice elicited a simultaneous peak of plasma acetate concentration and reduced plasma NEFA at 15 min post-injection⁽Reference Ge, Li and Weiszmann¹¹³⁾. This effect was abolished in FFA2 knockout mice, suggesting that the effects seen are occurring as a result of the direct activation of FFA2 by acetate.

The potential ability of FFA2 activation to stimulate adipogenesis and reduce lipolysis heralds a wider role for SCFA and FC in the control of the metabolic syndrome. It is intriguing that the beneficial effects of fibre supplementation on glucose control and dyslipidaemia observed in some studies⁽Reference Chandalia, Garg and Lutjohann¹¹⁴⁾ may be partly attributable to the previously unconsidered effect of SCFA on adipocyte FFA2. Demonstration of relatively high tissue expression levels of FFA2 on pancreatic islet cells⁽Reference Regard, Sato and Coughlin^59, Reference Leonard and Hakak¹¹⁵⁾ and the effect their activation has also warrants further investigation.

It is beyond the scope of the present paper to cover in detail the purported actions of the other SCFA-activated receptor FFA3. Readers are directed towards Stoddart's excellent review⁽Reference Stoddart, Smith and Milligan¹¹⁶⁾. Whilst FFA2 and FFA3 may share endogenous ligands, there is currently insufficient data to suggest if there is an overlap in the function of the receptors. FFA3 has also been found to be expressed in the human colon⁽Reference Tazoe, Otomo and Karaki¹¹⁷⁾. Although few data are currently available on its functionality, knockout models have suggested that it may be important in mediating the secretion of PYY⁽Reference Samuel, Shaito and Motoike¹¹⁸⁾. Another significant finding worth mentioning in brief is that of Xiong et al. ⁽Reference Xiong, Miyamoto and Shibata¹¹⁹⁾. They have shown that adipocyte FFA3 activation causes an increase in leptin secretion from mouse adipocytes in vitro and in vivo following SCFA (and most potently propionate) oral administration⁽Reference Xiong, Miyamoto and Shibata¹¹⁹⁾. Leptin, principally secreted from white adipose tissue, is a potent anorexigenic hormone and is a long-term and dynamic marker of body adiposity⁽Reference Maffei, Halaas and Ravussin¹²⁰⁾. This suggests that SCFA supplementation may also act on appetite via an FFA3-mediated response.

Future perspectives for free fatty acid receptor 2

Activation of colonic FFA2 by FC, SCFA, or pharmacological mimetics is a promising candidate in the fight against the current obesity onslaught. The novel synthetic allosteric agonists for FFA2 (phenylacetamides 1 and 2) have provided agents of possible pharmaceutical promise. Lee et al. demonstrated a greater activation potency of phenylacetamides 1 and 2 compared with endogenous SCFA and the ability of these compounds to activate both the G_q-coupled aequorin and G_i-coupled cyclic AMP inhibition pathways⁽Reference Lee, Schwandner and Swaminath¹²¹⁾. In vivo studies utilising these compounds are now eagerly awaited.

The co-localisation of FFA2 with PYY-releasing L-cells offers a further potential mechanism to link epidemiological observations of high-fibre diets and lower body weight. Diets high in FC are associated with increased PYY mRNA expression, PYY secretion and increased satiety. Whether PYY or GLP-1 could reach sufficient therapeutic levels through dietary supplementation alone, and if these physiological effects can be extrapolated directly to a reduction in food intake and body weight in humans require further exploration.

It remains a scientific challenge to study the effects of SCFA on FFA2 in isolation from the other physiological effects of FC and dietary fibre supplementation. Individual variation in host bacterial colonisation and transit time can lead to an unpredictable conversion of FC to SCFA⁽Reference Macfarlane and Macfarlane³⁸⁾, so standardised FC doses may not necessarily produce equivalent levels of FFA2 activation. A chronic resistant starch-supplementation study in rats demonstrated increasing caecal concentrations of butyrate and dynamic shifts in the molar concentrations between the different SCFA even after 6 months of supplementation⁽Reference Le Blay, Michel and Blottière¹²²⁾. Measurement of SCFA in vivo is also complex, low concentrations are found in peripheral blood⁽Reference Pomare, Branch and Cummings¹²³⁾ and appear to have diurnal variation related to meal pattern⁽Reference Wolever, Josse and Leiter¹²⁴⁾. Whilst direct supplementation with SCFA may appear to ease these methodological troubles, oral SCFA supplements are absorbed in the small intestine and therefore will not reach the colon, the site of greatest L-cell abundance.

The quantity and molar ratio of the SCFA produced from bacterial fermentation, as mentioned previously, is dependent on many factors including the host's microbiota. Interestingly, the relative contribution of different bacterial strains to the host's overall microflora has also been independently associated with obesity. Controversial studies utilising the leptin-deficient ob/ob mouse have found that the abundance of the Bacteroidetes and the Firmicutes strains of an obese mouse differs to that of lean, wild-type littermates⁽Reference Turnbaugh, Ley and Mahowald¹²⁵⁾. Furthermore, the inoculation of a germ-free mouse with ob/ob microflora induces fat gain and obesity⁽Reference Turnbaugh, Ley and Mahowald¹²⁵⁾. This is perhaps occurring as a result of increased energy yield from the digestive processes of obese mice and humans and this at least in part appears to be mediated via an increase in the production of the SCFA acetate and butyrate⁽Reference Samuel and Gordon¹²⁶⁾. Obesity-associated changes in host colonic bacteria have also been demonstrated in human subjects⁽Reference Turnbaugh, Hamady and Yatsunenko¹²⁷⁾ but it is unclear whether this is a cause or effect relationship. This may be a physiological attempt to prevent future weight gain, using the possible FFA2-mediated mechanism highlighted in this paper or alternatively, a paradoxical hypothesis could be proposed whereby the obese state may be caused by a bacterial profile which has a greater efficiency to generate SCFA. Further research is still required to test both hypotheses.

The evidence in the present review outlines a case for the implication of FFA2 in the regulation of appetite and energy homeostasis through a variety of mechanisms. Whilst the purported secular trends of increasing energy intake and declining physical activity levels are the most commonly cited environmental factors in obesity development, the significance of the decline in fibre intake in the aetiology of obesity may be underestimated. The marked rise in refined product availability and subsequent fall in the consumption of fermentable fibre-rich foods since the Industrial Revolution have possibly rendered FFA2 control of appetite regulatory circuits ineffective. The prospect of using a readily accessible dietary constituent of FC to up-regulate this appetite-regulatory loop is promising of its own accord, but the possibility that there may also be adipocyte FFA2 action on improving components of the metabolic syndrome makes FFA2 targeting a high therapeutic priority.

It would be negligent not to consider the counter side to this argument that if there is a reduction in food intake brought about by FFA2 activation, this may be a host defence mechanism in response to increased colonic SCFA concentration or reduced colonic pH to prevent pathological changes in the gut. Common side effects experienced in human FC studies have included bloating, cramping, flatulence and soft stools⁽Reference Davidson and Maki¹²⁸⁾. It is perhaps important to remember that the colon is an important constituent of the digestive, endocrine and immunological systems and, therefore, changing the colonic environment to favour one such system may not necessarily benefit another. Chronic in vivo studies measuring histological and immunological parameters are required to investigate this further. It is also worth considering that a lack of the enzyme propionyl CoA carboxylase in the newborn allows propionic acid levels to build to fatal levels in the blood⁽Reference Hsia, Scully and Rosenberg¹²⁹⁾. This recessive condition leads to the development of hypophagia, vomiting, dehydration, lethargy, metabolic acidosis, encephalopathy and death. Therefore the potential clinical dangers of purposefully elevating circulating SCFA should be carefully considered.

Future research needs

FFA2 research must now be aimed at proving whether the links proposed between FFA2 activation and PYY, GLP-1 and 5-HT exert appreciable effects on satiety and body weight and whether they are therapeutically exploitable. Dose-finding experiments in human subjects are essential for elucidating a dose which exerts physiological benefits on energy homeostasis but not at the expense of gastrointestinal discomfort or pathological changes in gut microstructure. While the current evidence base is insufficient to warrant SCFA supplementation, strategies to promote traditional diets based on whole grains, fruit, vegetables and unrefined products appear to offer some protection from the development of obesity.