Effects on physiology and metabolic health

A renewed focus to understand the relationship between diet and metabolic health has arisen in part due to the increased prevalence of obesity and associated comorbidities over the past 100 years, mostly due to high calorie intake, particularly an increased intake of dietary fat, which affects metabolic health [Reference Hu, Wang, Yang, Li, Togo and Wu10–Reference Wang, Wang, Zhang, Popkin and Du15]. A particular interest in protein intake has emerged with many weight loss or weight maintenance recommendations promoting increased protein intake, generally above 20% of total energy intake within the 10–30% acceptable macronutrient range for proteins [Reference Wolfe, Cifelli, Kostas and Kim16]. Data show that HPD intake reduced body weight gain or cause weight loss up to 10%, with a reduction in fat mass and an increased lean mass in overweight and obese individuals of both sexes (Table 1) [Reference Santesso, Akl, Bianchi, Mente, Mustafa, Heels-Ansdell and Schunemann14,Reference Moon and Koh30–Reference Sacks, Bray, Carey, Smith, Ryan and Anton32]. The effects extended to include reduction in plasma insulin levels, triacylglycerol, high-density lipoproteins and blood pressure (Table 1). Notably, whilst these effects have been shown relative to baseline measurements or in comparison with carbohydrate intake (Table 1), there is evidence that the quality of the protein also impacts metabolic health (Table 1). For instance, whey protein (WP) intake reduced waist circumference and circulating ghrelin and insulin-like growth factor (IGF)1 levels in obese humans in comparison with soy intake (Table 1). Relative to collagen, WP caused a reduction in visceral fat, with similar effects shown for milk proteins, containing both WP and casein, compared with controls fed milk proteins and soy (Table 1). These effects reported for ad libitum intake have been extended to include calorie restriction, with WP showing a greater improvement of metabolic health than other protein sources (Table 1), but there are few exceptions (Table 1) [Reference Piccolo, Comerford, Karakas, Knotts, Fiehn and Adams26,Reference Kjolbaek, Sorensen, Sondertoft, Rasmussen, Lorenzen, Serena, Astrup and Larsen29]. It is also important to highlight that there are data showing unhealthy outcomes of HPDs. For instance, red meat intake has been associated with increased risk of colorectal cancers and kidney disease [Reference Aykan33,Reference Lew, Jafar, Koh, Jin, Chow, Yuan and Koh34]. The different effects of proteins on metabolic health can be related to the quantity and composition of the amino acids, how the proteins are digested and absorbed through the gut and the impact on the gut microbiota and their functional capacity to produce metabolites, which ultimately affects host health [Reference Agus, Clement and Sokol35–Reference Gorissen, Crombag, Senden, Waterval, Bierau, Verdijk and van Loon38]. For this review, we focused attention on dairy, meat or plant protein intake and their impact on the abundance of metabolites produced in the gut (and, hence, detected in faeces) as well as that emerge in circulation/urine to better understand how different proteins affect host metabolic health. Our focus was on data related to human studies, but in a few cases we have mentioned rodent studies to draw conclusions. The search includes effects of HPD, where the protein content was equal or greater than 20% of total energy intake.

Table 1. Impact of protein quantity and quality on body weight and metabolic health in humans

The direction of change is shown by arrows, as increase (↑), decrease (↓) or no change (↔).Ad lib, ad libitum; CHO, carbohydrate; CR, calorie restriction; Dur, duration; EAA, essential amino acids; F, female; HDL, high-density lipoproteins; IGF, insulin-like growth factor; M, male; MP, milk proteins; WMD, weight maintenance diet; WP, whey proteins; WPH, whey protein hydrolysate; W, weeks.

Effect on the gut microbiota

The gastrointestinal tract is inhabited by the microbiota, with the colon containing a much higher density of microbiota (10¹⁰–10¹¹ cells per millilitre of contents), as well as a much more diverse microbial composition compared with other parts of the intestine [Reference Eckburg, Bik, Bernstein, Purdom, Dethlefsen and Sargent39,Reference Vuik, Dicksved, Lam, Fuhler, van der Laan and van de Winkel40]. This complexity in microbial communities is further supported by the colonic structure and functions. Notably, the colon is made up of a number of colonic epithelial cells (Fig. 1), with many of these cells, particularly goblet cells, capable of producing enzymes contributing to the metabolism of nutrients in the intestinal mucosa before they reach the circulation [Reference Husted, Trauelsen, Rudenko, Hjorth and Schwartz41]. The colonic microbiome also acts as a vital part of the digestive process by breaking down complex carbohydrates, proteins and fats, which are not broken down enzymatically in the preceding parts of the digestive system [Reference Tremaroli and Bäckhed42]. The transit rate in the colon is much slower compared with the small intestine, allowing increased microbial action on the food material [Reference Roager, Hansen, Bahl, Frandsen, Carvalho and Gobel43]. Indeed, microbial metabolism of digested dietary proteins results in the production of a range of nutrients and metabolites, which have diverse physiological functions (see below).

The microbiota composition plays an important role in metabolic health [Reference Cunningham, Stephens and Harris37,Reference Ley44,Reference Ley, Peterson and Gordon45]. Notably, the alpha and beta microbial diversity, measured by the richness of diversity and evenness and relative differences in the overall diversity of taxa, respectively, highlight the similarities and differences in the microbiota across the different interventions, and associated metabolic states. A rich and diverse gut microbiota composition generally reflects a microbiota that is more resilient and capable of functioning better, with a loss in species diversity a common finding in several disease states [Reference Manor, Dai, Kornilov, Smith, Price, Lovejoy, Gibbons and Magis46]. The importance of the gut microbiota in mediating protein effects was highlighted by recent work showing that WP reduced body weight gain in high-fat-fed mice and that this effect can be transferred via faecal matter onto mice fed casein [Reference Nychyk, Barton, Rudolf, Boscaini, Walsh and Bastiaanssen6,Reference Boscaini47,Reference Boscaini, Cabrera-Rubio, Golubeva, Nychyk, Fulling and Speakman48]. In contrast to animal studies [Reference Wu, Bhat, Gounder, Mohamed Ahmed, Al-Juhaimi, Ding and Bekhit49], only few studies show an impact of dietary proteins on the gut microbiota in humans in the overweight and obese categories (Table 2). Of note, subjects ingesting varied quantities of dietary fats, whilst co-ingesting proteins at 25% energy from various sources (red and white meat and plants), show no effect on the alpha or beta diversities [Reference Lang, Pan, Cantor, Tang, Garcia-Garcia and Kurtz50]. However, in the latter study, when the main effect of dietary fat on the gut microbiota was removed, an effect of dietary proteins can be seen on these micro-organisms, which were largely due to any source of protein rather than the quality of the protein consumed (Table 2). Similar data have been generated to show an impact of protein quantity on the composition of the gut microbiota (e.g. with or without fish intake or high and low gluten intake; Table 2). Where the impact of the source of proteins was investigated (pork versus chicken intake), the only changes in the gut microbiota was seen relative to baseline intake for each protein type [Reference Dhakal, Moazzami, Perry and Dey54] (Table 2). Imposing a calorie restriction for 8 weeks also did not affect the gut microbiota regardless of the hydrolysed state of the WP proteins [Reference Sun, Ling, Liu, Zhang, Wang and Tong28,Reference Beaumont, Portune, Steuer, Lan, Cerrudo and Audebert53] (Table 2). In contrast to the above studies, which used 16S rRNA sequencing to uncover microbial changes, a study by Bel Lassen et al. [Reference Bel Lassen, Attaye, Adriouch, Nicolaou, Aron-Wisnewsky and Nielsen25] used Metagenomics sequencing to explore the impact of an extended calorie restriction (12 weeks) on subjects consuming milk proteins supplemented with amino acids. The latter intervention was found to increase the microbial potential to produce amino acids compared with pea and casein intake (Table 2). This suggests that the interaction between the quality of the protein and the gut microbiota may be more subtle (at a functional level), requiring a greater depth of sequencing to uncover, but with the potential to influence the luminal pool of amino acids and their derivatives that are accessible by the host.

Table 2. Impact of protein quantity and quality on the composition and functional potential of the gut microbiota in humans

The direction of change is shown by arrows, as increase (↑), decrease (↓) or no change (↔).Ad lib, ad libitum; CR, calorie restriction; Dur, duration; F, female; M, male; MP, milk proteins; WP, whey proteins; WPH, whey protein hydrolysate W, weeks.

Effects on nutrients and metabolites

Most digested proteins are absorbed in the small intestine as amino acids, but some undigested proteins, especially following HPD intake, reach the colon where they are further broken down by proteolytic bacteria for the synthesis of other amino acids and/or into amino acid derivatives that have been associated with numerous health outcomes, including regulating digestion and absorption (Table 3). Of note, lysine, arginine, glycine and the branched-chain amino acids (BCAA), namely leucine, iso-leucine and valine, are the most preferred amino acid (AA) substrates of colonic microbiota [Reference Dai, Wu and Zhu83].

Table 3. Metabolic effects of dietary protein or microbial-derived amino acids and their metabolites

AA, amino acids; BCAA, branched-chain amino acids; BCFA, branched-chain fatty acid; GABA, gamma-aminobutyric acid; SCFA, short-chain fatty acid.

BCAA and their derivatives

BCAA are building blocks of lean tissue and are capable of modulating gene expression and signalling pathways, including regulating dietary nutrient absorption, partake in lipolysis, lipogenesis, glucose metabolism and intestinal barrier function [Reference Doi, Yamaoka, Fukunaga and Nakayama84–Reference Zhang, Zeng, Ren, Mao and Qiao86]. However, BCAA have also been shown to have negative effects on metabolism, with increased consumption of BCAA correlated with a more unhealthy metabolic state [Reference Orozco-Ruiz, Anesi, Mattivi and Breteler87], although these effects may be mediated somewhat by changing the levels of individual BCAAs [Reference Yu, Richardson, Green, Spicer, Murphy and Flores88]. Increased BCAA intake has been shown to result in increased insulin resistance [Reference Bishop, Machate, Henning, Henkel, Püschel and Weber89–Reference Vanweert, Schrauwen and Phielix91]. Negative effects of BCAA intake may also include an increased risk of cancer [Reference Rossi, Turati, Strikoudi, Ferraroni, Parpinel, Serraino, Negri and La Vecchia92]. Conversely, a reduction in BCAA intake has been shown to have positive effects on metabolic health [Reference Cummings, Williams, Kasza, Konon, Schaid and Schmidt93].

Intake of HPD increased faecal levels of BCAA, specifically following dietary casein and red and white meat intake (Table 4). By contrast, circulating levels of BCAA increased regardless of the type of protein consumed in both fasted and non-fasted states after prolonged intake (3–4 weeks) as well as after acutely challenges, where the post-prandial plasma increase was higher after milk protein consumption compared with plant protein intake (within 5 h), WP intake compared with casein intake (3 h) and following red meat intake compared with baseline measurements (within 4 h) (Table 5). It is interesting that HPD and BCAA have both positive and negative outcomes on metabolic health. Whilst this suggests a potential functional relationship in the way dietary proteins affect metabolic health, it is important to highlight the role of the gut microbiota as a modulator of the effects. This is because these micro-organisms can convert BCAA into short-chain fatty acids (SCFA) and branched-chain fatty acids (BCFA) (Table 3), which have diverse metabolic health effects (discussed below). Indeed, in agreement with the BCAA availability in the faeces and circulation, HPD intake also increases BCFA in faeces with some reaching the urine (Tables 4 and 5). By contrast, the availability of SCFA in faeces and urine was either unaffected or decreased, with the exception of acetate, which was increased in faeces and urine following casein intake (Tables 4 and 5 and further detailed below). The data suggest a potential microbial preference for conversion of amino acids into BCFA over SCFA in a background of HPD intake, which generally accompanies a low carbohydrate intake [Reference Beaumont, Portune, Steuer, Lan, Cerrudo and Audebert53,Reference Andriamihaja, Davila, Eklou-Lawson, Petit, Delpal and Allek106,Reference Sattari Najafabadi, Skau Nielsen and Skou Hedemann107].

Table 4. Impact of dietary proteins on the metabolite profiles in the faeces in humans

The direction of change is shown by arrows, as increase (↑), decrease (↓) or no change (↔) of metabolites. The length of the dietary challenge is shown in subscript in weeks (W). BCAA, branched-chain amino acids; BCFA, branched-chain fatty acids; CAS, casein; EAA, essential amino acids; MP, milk proteins; RED, red meat; SCFA, short-chain fatty acids; GLTN, gluten; WHT, white meat.

Table 5. Impact of dietary proteins on the metabolite profiles in circulation or urine in humans

The direction of change is shown by arrows, as increase (↑), decrease (↓) or no change (↔) of metabolites. The length of the dietary challenge is shown in subscript in weeks (W) or hours (H) along with the medium in which the metabolite was detected and whether the subjects were fasted or non-fasted. BCAA, branched-chain amino acids; BCFA, branched-chain fatty acids; CAS, casein; EAA, essential amino acids; MP, milk proteins; RED, red meat; SCFA, short-chain fatty acids; GLTN, gluten; SCFA, short-chain fatty acids; WHT, white meat.

Exploring the potential mechanisms

All protein sources increased BCFA in faecal content, probably from the increased gut availability of BCAA (Fig. 2A), suggesting a greater bacterial conversion of BCAA to BCFA with the intake of different proteins, but we cannot exclude the contribution of other amino acids for this process. Dairy protein intake specifically increased faecal levels of indole and acetate (Fig. 2A). Alongside these health-promoting metabolites, several other metabolites emerge in faecal matter with known unhealthy outcomes. These were phenol (dairy), p-cresol derivatives (dairy and plant), ammonia (all protein sources) and tyramine (plant). In circulation, and regardless of the source of proteins, amino acids, including BCAA, increased (Fig. 2B). In addition, for dairy proteins, the impact on the faecal availability of acetate, indole, p-cresol, BCFA and butyrate mirrored availability in the circulatory system or urine (Fig. 2B), suggesting both gastro-intestinal and systemic effects of these metabolites. This contrasts with metabolites produced following plant and meat protein intake, which show fewer common responses in faecal matter and circulation (and urine) (Fig. 2). The contrasting levels of metabolites in the gut and circulation/urine following dairy, meat or plant protein intake could be related to the differences in the amino acid composition and the three-dimensional structure of the proteins accessible for enzymatic digestion and how the resulting digested peptides and amino acids are utilised by the dietary protein-sensitive gut microbiota to produce metabolites, which ultimately reach the gut and/or enter the circulatory system [Reference Agus, Clement and Sokol35–Reference Gorissen, Crombag, Senden, Waterval, Bierau, Verdijk and van Loon38].

Figure 2. Impact of protein quality on the metabolites in (A) faeces and (B) circulation or urine. The direction of change is shown by arrows, as increased (↑) or decreased (↓). Metabolites highlighted in red colour are known to cause unhealthy outcomes in humans. BCAA; branched-chain amino acids; BCFA, branched-chain fatty acids; EAA; essential amino acids.

In exploring the mechanisms for the physiological outcomes of HPD intake, the post-prandial increase in circulatory levels of AA including BCAA is notable because their increase has been associated with increased satiety in human subjects, in particular following WP consumption compared with casein intake. The effect can be related to increased circulatory levels of satiety related hormones, namely cholecystokinin, (GLP)-1 and glucose-dependent insulinotropic polypeptide [Reference Hall, Millward, Long and Morgan101]. Whilst these data have emerged from acute challenges, the long-term intake of HPD does not appear to cause changes in energy intake in humans, suggesting that there are other mechanisms at play [Reference Nilaweera and Cotter135]. In this regard, the increased availability of indole in the gut lumen by dairy protein intake is interesting because this metabolite and its derivatives are known to increase the release of GLP-1 from enteroendocrine cells [Reference Hu, Yan, Ding, Cai, Zhang, Zhao, Lei and Zhu69], and the activity of GLP-1 has been linked to roles beyond the reduction in food intake to include effects on adiposity [Reference Nogueiras, Perez-Tilve, Veyrat-Durebex, Morgan, Varela and Haynes136]. In addition, and separate from effects on GLP-1, indoles and their derivatives have a direct impact on adiposity, reducing adipogenesis and increasing thermogenesis [Reference Hu, Yan, Ding, Cai, Zhang, Zhao, Lei and Zhu69], and their detection in urine following dairy (and plant) protein, presumably by the crossover from circulation, supports circulatory effects. Given that BCFA have been associated with a reduction in body weight [Reference Taormina, Unger, Schiksnis, Torres-Gonzalez and Kraft75], it is not surprising that these should also increase in circulation following HPD intake (Fig. 2). The data suggest that indole and BCFA generated by HPD intake may contribute to the reduction in body weight through effects on intake, effects on energy expenditure and/or direct effects on lipid metabolism. These metabolites could account, at least in part, for the greater impact of dairy proteins on metabolic health compared with other sources of proteins. In contrast to body weight and adiposity, the effect of HPD on insulin sensitivity is inconstantly reported [Reference Santesso, Akl, Bianchi, Mente, Mustafa, Heels-Ansdell and Schunemann14,Reference Clifton, Keogh and Noakes21,Reference Drummen, Tischmann, Gatta-Cherifi, Adam and Westerterp-Plantenga137,Reference Gannon and Nuttall138]. This may be due in part to increased circulatory levels of BCAA and increased faecal and circulatory levels of p-cresol and its derivatives, which are known to reduce insulin sensitivity [Reference Bishop, Machate, Henning, Henkel, Püschel and Weber89,Reference De Bandt, Coumoul and Barouki90], counterbalanced by the increased levels of acetate detected mainly following dairy protein intake. Whilst HPD intake is known to reduce or cause no change in butyrate and propionate levels, the increased acetate level in faecal matter is significant because of important roles of these SCFA in energy balance regulation and insulin sensitivity [Reference Hernandez, Canfora, Jocken and Blaak139–Reference Pham, Joglekar, Wong, Nassif, Simpson and Hardikar141]. Overall, it is clear that proteins from different sources produce common and distinct metabolites, and their unique mechanisms of actions in the gut and/or via circulation presumably underlie the differences in physiological and metabolic outcomes of HPD.

Future directions

There are limited data on the effect of high protein intake on the composition and functional potential of the gut microbiota. Further studies are also needed to ascertain how plant proteins other than soy and gluten influence the nutrients and metabolite profiles in the gut, given the increased focus on these proteins as a sustainable production source for human consumption [Reference Langyan, Yadava, Khan, Dar, Singh and Kumar142]. Extending these lines of investigation, work is also needed to clarify the role of sex, since this parameter influences the protein quantity consumed, the composition and the functional potential of gut microbiota and physiological and metabolic parameters [Reference Bennett, Peters and Woodward143–Reference Santos-Marcos, Haro, Vega-Rojas, Alcala-Diaz, Molina-Abril and Leon-Acuna145]. While the focus of our review was on human data, cross-species investigations can provide a greater understanding of the role played by nutrients and metabolites identified here as potential mediators of metabolic health effects of HPD. These studies could involve the transfer of faecal matter from humans to other species such as rodents within each sex and/or supplementation or depletion of the nutrients or metabolites in the diet to ascertaining their biological significance.