Definition

The American Association of Cereal Chemists (AACC) gave the following scientific and botanical definition in 1999: ‘Whole grains shall consist of the intact, ground, cracked or flaked caryopsis, whose principal anatomical components – the starchy endosperm, germ and bran – are present in the same relative proportions as they exist in the intact caryopsis’⁽²⁰⁾. The definition given by the Whole Grains Council in May 2004 includes processed food products: ‘Whole grains or foods made from them contain all the essential parts and naturally-occurring nutrients of the entire grain seed. If the grain has been processed (e.g. cracked, crushed, rolled, extruded, and/or cooked), the food product should deliver approximately the same rich balance of nutrients that are found in the original grain seed’⁽²¹⁾. The US Food and Drug Administration published a Draft Guidance on Whole-grain Label Statements in 2006 that adopted the international AACC definition and included amaranth, barley, buckwheat, bulgur, maize (including popcorn), millet, quinoa, rice, rye, oats, sorghum, teff, triticale, wheat and wild rice; pearled barley was not included because some outer layers of the bran fraction are removed⁽²²⁾. Pseudocereals such as amaranth, buckwheat and quinoa have similar macronutrient compositions (carbohydrates, proteins and lipids), and are used in the same traditional ways as cereals^(23, Reference Jones²⁴⁾. The response to the US Food and Drug Administration Draft Guidance by the AACC International recommended that some traditional cereals such as ‘lightly pearled barley, grano (lightly pearled wheat), nixtimalized corn and bulgur that has been minimally processed be also classified as whole grains’⁽²³⁾, making allowance for small losses of components that occur through traditional processing. The Whole Grain Task Force stated in 2008 that it ‘supports the use of the term whole-grain for products of milling operations that divide the grain into germ, bran and endosperm, but then recombine the parts into their original proportions before the flour leaves the mill’⁽Reference Jones²⁴⁾. However, as I will explain later, most of the products defined as whole-grain foods in studies showing the health benefits of whole-grain cereals are made of recombined whole-grain flours⁽Reference Jones²⁴⁾, which rarely contain the same proportions of bran, germ and endosperm as the intact grain before milling. Thus, the germ fraction is almost always removed because its high lipid content (about 9 %) may go rancid upon storage⁽Reference Srivastava, Sudha and Baskaran²⁵⁾. Processing whole-grain cereals also leads to losses of bioactive compounds so they cannot really deliver ‘approximately the same rich balance of nutrients that are found in the original grain seed’⁽²¹⁾. Thus, if researchers had referred strictly to the definitions given above, few studies could have concluded that whole-grain cereal foods protect human health. Alternative definitions have therefore been proposed by the Whole Grain Task Force in which ‘as they exist in the intact caryopsis’ in the AACC definition is replaced by ‘as found in the least-processed, traditional forms of the edible grain kernels’ or completed by adding ‘as they exist in the intact caryopsis to the extent feasible by the best modern milling technology’⁽Reference Jones²⁴⁾. This last definition is probably the best adapted to our Western country technologies. But none of these alternative definitions has been adopted to date and there is still no official international definition of whole-grain cereal products in Europe.

What proportions?

Finally, the proportion of whole grains that must be present in a cereal product needs to be defined for it to be considered a whole-grain product. The issue is still debated. The definition given by the American Food and Drug Administration⁽²⁶⁾ in 1999 was: ‘For purposes of bearing the prospective claim, the notification defined ‘whole grain foods’ as foods that contain 51 percent of total weight or more whole grain ingredient(s) by weight’ (extract). This definition was debated and contested by the European Whole Grain Task Force in 2008. They explained that: ‘Using total weight gives advantage to products sold by dry weight such as crackers and ready-to-eat cereal. Because foods like breads have a proportionally high water content, even some breads made with all whole grain flours but containing significant amounts of nuts, seeds and fruit would fail to meet the 51 % by weight rule’⁽Reference Jones²⁴⁾. Apparently, there is still no international consensus as to the right proportion of whole grain by dry weight (DW) in a product in order for it to be called a whole-grain product. Each country has its own definition and standards⁽²¹⁾. However, most research and observational studies, particularly those on breakfast cereals, estimate the whole-grain intake from products containing at least 25 % whole grains or bran by weight⁽Reference de Munter, Hu and Spiegelman^5, Reference Adom, Sorrells and Liu^14, Reference Jacobs, Meyer and Kushi^27, Reference Liu, Manson and Stampfer²⁸⁾. Thus, a study on young individuals aged 4–18 years found that using a 51 %-based definition underestimated the whole-grain intake by 28 %, breakfast cereals (56 %) and bread (25 %) being the major sources of whole-grain cereals⁽Reference Thane, Jones and Stephen²⁹⁾. In another study on adiposity among two cohorts of British adults, the same research team assumed that whole-grain foods contained ≥ 10 % whole grains and found little or no association between the whole-grain intake and anthropometric indices⁽Reference Thane, Stephen and Jebb³⁰⁾. This suggests that the threshold of 10 % is probably too low and emphasises the need to harmonise how the whole-grain cereal food intake is calculated. In these studies, generally carried out in Western countries, whole-grain cereal foods considered are, for the most cited, whole-grain breads (for example, dark, brown, wholemeal and rye bread), whole-grain breakfast cereals (for example, muesli), popcorn, cooked porridges (oatmeal or whole wheat), wheat germ, brown rice, bran, cooked grains (for example, wheat, millet and roasted buckwheat) and other grain-based foods such as bulgur and couscous. A complete list of food ingredients classified as whole grains in the US Department of Agriculture (USDA) pyramid servings database is reported by Cleveland et al. ⁽Reference Cleveland, Moshfegh and Albertson³¹⁾. Refined grain foods generally include white breads (for example, French baguette), sweet rolls, noodles, pasta, cakes, biscuits, viennoiseries, muffins, refined grain breakfast cereals, white rice, pancakes, waffles and pizza.

The importance of whole-grain cereal product consumption

There are far fewer whole-grain cereal products on the market than there are refined products, at least in Western countries. The major sources of whole-grain cereals are breads, breakfast cereals and whole-grain cereals consumed as such (for example, brown rice or quick-cooking whole-grain barley and wheat). Epidemiological data show that the consumption of two to three servings of whole-grain cereal per d is sufficient to get beneficial health effects⁽Reference Lang and Jebb³²⁾. The recommended consumption of whole-grain cereal products differs from one country to another, but most recommend increased whole-grain cereal product consumption^(21, Reference Lang and Jebb³²⁾. For example, at least three servings daily are recommended in the USA, that is, about 48 g of whole-grain cereals⁽Reference Welsh, Shaw and Davis³³⁾; between six and twelve servings daily are recommended in Australia and four servings daily in Denmark⁽²¹⁾. Other countries such as Canada, UK, Greece, Germany, Austria and Switzerland are not so precise and generally recommend an increase in cereal consumption with emphasis on whole-grain products⁽²¹⁾. Surveys carried out in the USA and the UK showed that most individuals consume less than one serving per d and about 30 % any, and that only 0·8 to 8 % of those surveyed in the USA consumed the recommended three servings per d⁽Reference Cleveland, Moshfegh and Albertson^31, Reference Lang and Jebb^32, Reference Albertson and Tobelmann³⁴⁾. The situation is quite different in Scandinavian countries, where individuals consume more whole-grain cereal products, particularly rye-based⁽Reference Lang and Jebb³²⁾. For example, Norwegians consume an estimated four times more whole-grain products than do Americans⁽³⁵⁾, but less than the Finns, 40 % of whom may consume four or more slices of dark bread per d⁽Reference Prättälä, Helasoja and Mykkänen³⁶⁾. Why is consumption so low in other Western countries? There are probably several reasons. First, unlike fruits and vegetables, individuals do not know about the benefits of whole-grain cereal products. Second, individuals tend to think that whole-grain cereal products are not very tasty. And third, whole-grain cereal products are less common and many are difficult to identify as being whole-grain (problem of labelling). Last, time and money have been cited as obstacles to eating more nutritiously⁽Reference Adams and Engstrom³⁷⁾.

Whole-grain and wholemeal

The terms ‘whole-grain’ and ‘wholemeal’ are mostly used synonymously. It is generally believed that whole-grain products are made with wholemeal flour, and that they may secondarily also contain intact grains. But the form in which grain is incorporated into food, intact or milled, is nutritionally significant. Thus ‘wholemeal’ (made of milled whole-grain flour) and ‘whole-grain’ (made with intact cereal grains) breads have different effects on postprandial glycaemia. The whole-grain breads produce a significantly lower glycaemic response than the wholemeal breads⁽Reference Jenkins, Wesson and Wolever³⁸⁾. This underlines the importance of food structure on physiology. Thus, for clarity, the term ‘whole-grain’ should be used for cereal products containing more or less intact cereal kernels, and ‘wholemeal’ for cereal products made of more or less refined flour, in which bran, germ and endosperm are first separated, and then reassembled, in proportions that rarely correspond to those of intact grains, as the germ fraction is generally removed.

Food structure

The structure of food has long been recognised as an important parameter governing the health benefit of whole-grain cereal products. The first study was performed in 1977 by Haber et al. on the influence of apple structure (intact apples v. apple purée v. fibre-free apple juice) on satiety, plasma glucose and serum insulin. The removal of fibre and/or the disruption of the physical food structure was accompanied by reduced satiety, disturbed glucose homeostasis and an inappropriate insulin response⁽Reference Haber, Heaton and Murphy⁴⁷⁾. Almost 10 years later, it was shown that simply swallowing carbohydrate-rich foods (rice, apple, potato and sweetcorn) without chewing was sufficient to significantly decrease postprandial glycaemia⁽Reference Read, Welch and Austen⁴⁸⁾. This was the simplest way to emphasise the importance of food structure (chewing v. no chewing) on digestion. Then, Jenkins et al. studied the effects of wholemeal and wholegrain breads and showed that the glycaemic index (GI) of wholemeal breads (wheat or barley flour-based) without intact grains was the same as that of white bread made of refined flour (>90), and that increasing the intact barley kernel or cracked wheat grain content of the bread (50 and 75 %) resulted in a significantly large decrease in the GI from 92–96 to 39⁽Reference Jenkins, Wesson and Wolever³⁸⁾. Thus, an intact botanical food structure is more important than the composition of the food (the presence of fibre in wholemeal bread and absence from white bread) for influencing physiological responses like those related to satiety and glucose metabolism. Many later studies have confirmed these results, emphasising the importance of preserving the natural initial fibrous network, particularly in more or less intact wheat, barley, rye and oat kernels⁽Reference Fardet, Leenhardt and Lioger^49–Reference Nilsson, Ostman and Granfeldt⁵²⁾.

Whole-grain cereals as a rich source of fibre

Dietary fibre is defined by the AACC as ‘the edible parts of plants or analogous carbohydrates that are resistant to digestion and absorption in the human small intestine with complete or partial fermentation in the large intestine. Dietary fibre includes polysaccharides, oligosaccharides, lignin and associated plant substances. It promotes beneficial physiological effects including laxation and/or blood cholesterol attenuation and/or blood glucose attenuation’⁽⁵³⁾. This definition includes that fraction of starch not digested in the small intestine, resistant starch (RS). Whole-grain wheat may contain from 9 to 17 g total fibre per 100 g edible portion (Table 2), which is more than in most vegetables (generally < 6 g/100 g edible portion). Thus, consuming whole-grain cereal products is undoubtedly a good way of increasing the fibre intake from the 10–15 g/d eaten by most Western populations to the recommended level of about 30–35 g/d.

Wheat is relatively poor in soluble fibre. It has been found that the soluble:insoluble fibre ratio is about 1:5 for whole-grain wheat, 1:10 for wheat bran and 1:3 for wheat germ (Table 2). Whole-grain wheat therefore provides large quantities of insoluble fibre (up to 11 g/100 g) and RS (up to 22 % for certain high-amylose barley varieties⁽Reference Nilsson, Ostman and Holst⁵⁴⁾). Cereal fibre is now recognised to be beneficial for bowel health. Wheat has a great diversity of fermentable carbohydrates. Except for lignin, whose nutritional benefits are not really known, all the types of fibre compounds, including soluble and insoluble fibre, oligosaccharides and RS, have important physiological properties and provide significant health benefits⁽Reference Liu^18, Reference Topping⁵⁵⁾. For example, soluble fibre increases viscosity, which delays gastric emptying and limits glucose diffusion towards the enterocytes for absorption. This leads to a lower glucose response when sufficient quantities are ingested⁽Reference Wood⁵⁶⁾.

Cereal fibres also increase satiety and help control body weight⁽Reference Koh-Banerjee, Franz and Sampson⁵⁷⁾. The mechanisms by which dietary fibre positively affect body weight have been previously described: briefly, they involve hormonal effects via reduction of the insulin secretion, metabolic effects via increased fat oxidation and decreased fat storage due to greater satiety, and colonic effects via SCFA production⁽Reference Slavin⁵⁸⁾. Thus, the consumption of highly viscous fibre such as β-glucans, found mainly in barley and oats, is now recommended for the management of glucose homeostasis in type 2 diabetic subjects⁽Reference Jenkins, Jenkins and Zdravkovic⁵⁹⁾. Soluble fibre has also been shown to reduce cholesterolaemia in ileostomy subjects⁽Reference Zhang, Hallmans and Andersson⁶⁰⁾ by probably favouring an increase in bile acid excretion as shown in ileostomates following oat β-glucans consumption⁽Reference Lia, Hallmans and Sandberg⁶¹⁾. Increased bile acid excretion stimulates bile acid synthesis from serum cholesterol, so reducing cholesterolaemia⁽Reference Lia, Hallmans and Sandberg⁶¹⁾.

The fermentation of fibre and RS within the colon produces SCFA that are associated with a lower risk of cancer⁽Reference Scheppach, Bartram and Richter^62, Reference Slavin⁶³⁾, favouring the development of a healthy colonic microbiota (i.e. prebiotic effect)⁽Reference Costabile, Klinder and Fava⁶⁴⁾. These SCFA also reduce the proliferation of human colon cancer cell lines in vitro ⁽Reference Scheppach, Bartram and Richter^62, Reference Slavin⁶³⁾. RS is known to produce large quantities of butyrate⁽Reference Brouns, Kettlitz and Arrigoni⁶⁵⁾. The increased butyrate production by rats fed wheat bran is negatively associated with the proliferation of colon crypt cells that are involved in the development of colorectal cancer⁽Reference Boffa, Lupton and Mariani⁶⁶⁾. RS also significantly increases fat oxidation in humans, probably by increased SCFA production that inhibits glycolysis in the liver, so rendering it more dependent on fat-derived acetyl CoA as fuel, this effect being associated with a concomitant decrease in carbohydrate oxidation and fat storage⁽Reference Higgins, Higbee and Donahoo⁶⁷⁾.

In contrast, insoluble fibre, which is poorly fermented in the colon, favours an increased transit time and greater faecal bulking⁽Reference McIntosh, Noakes and Royle⁶⁸⁾, two parameters that probably prevent colon cancer by diluting carcinogens and reducing their time in contact with epithelial cells⁽Reference Ferguson and Harris⁶⁹⁾. The fermentation of some fibre also increases mineral absorption in rats, mainly by increasing the surface area available for absorption (epithelial cell hypertrophy) and/or by favouring better hydrolysis of phytic acid via enhanced fermentation, as was shown with RS⁽Reference Lopez, Coudray and Bellanger^70, Reference Lopez, Levrat-Verny and Coudray⁷¹⁾ and inulin (a fructan-type compound)⁽Reference Coudray, Bellanger and Castiglia-Delavaud^72, Reference Lopez, Coudray and Levrat-Verny⁷³⁾.

Whole-grain cereals and butyrate production

Whole-grain cereal products are an important indirect source of butyrate, produced notably through RS fermentation⁽Reference Brouns, Kettlitz and Arrigoni⁶⁵⁾. Butyrate has cancer-preventing properties in rats by inducing apoptosis⁽Reference Reddy, Hamid and Rao⁷⁴⁾ or reducing tumour mass⁽Reference McIntyre, Gibson and Young⁷⁵⁾. But its positive physiological action may not be restricted to these two effects. The precise mechanisms involved in the anti-colon cancer effect of butyrate have been reviewed from in vitro, animal and human studies and they mainly include a combination of several physiological modifications in relation to abnormal cell growth inhibition, immune system stimulation and modulation of DNA repair and synthesis⁽Reference Brouns, Kettlitz and Arrigoni⁶⁵⁾. Butyrate might also protect against breast and prostate cancers, as shown by in vitro studies on mammary⁽Reference Heerdt, Houston and Anthony⁷⁶⁾ and prostate⁽Reference Ellerhorst, Nguyen and Cooper⁷⁷⁾ cancer cell lines⁽Reference Brouns, Kettlitz and Arrigoni⁶⁵⁾. The RS content of whole-grain cereal products depends on the proportion of the different types of RS: RS1 which is physically inaccessible to α-amylase, RS2 which is raw starch granules, and RS3 which is recrystallised/retrograded amylose that is formed when cooked food cools. It is therefore difficult to obtain precise data on the RS content of whole-grain cereal products, but some products are enriched in RS by selecting high-amylose varieties of cereal. Nevertheless, products containing whole grains or made from high-amylose cereal varieties will have proportionally higher RS contents and produce more butyrate, as was shown in human subjects fed various breads, breakfast cereals and crackers⁽Reference Bird, Vuaran and King^78, Reference Liljeberg and Bjorck⁷⁹⁾. Whole-grain cereal products with an intact botanical structure, that is with intact kernels, will have a higher RS1 content, since it is inaccessible to α-amylase, and butyrate production. The relationship between the consumption of whole-grain cereals and/or their bran and germ fractions, butyrate production and long-term health effects deserve to be studied more thoroughly in human subjects, particularly because of the effects in rats of butyrate on fat oxidation and of total SCFA production on cholesterol synthesis reduction⁽Reference Hara, Haga and Aoyama⁸⁰⁾.

The ‘second-meal effect’

The ‘second-meal effect’ is characterised by an improved carbohydrate tolerance at a meal (either lunch or breakfast, called the ‘second meal’) about 4–5 or 10–12 h after the consumption of a low-GI meal (i.e. the ‘first meal’), an effect which may contribute to the long-term metabolic benefits of low-GI diets. It was first described by Jenkins et al. who used viscous guar gum⁽Reference Jenkins, Wolever and Nineham⁸¹⁾, and thereafter for low-GI carbohydrate foods such as lentils⁽Reference Jenkins, Wolever and Taylor⁸²⁾. Recently, mechanisms have been proposed to explain the sustained positive effect of low-GI whole-grain products composed of intact barley or rye kernels consumed at diner or breakfast on the glycaemic response at the following meal, breakfast or lunch⁽Reference Nilsson, Ostman and Granfeldt^52, Reference Nilsson, Ostman and Holst^54, Reference Nilsson, Granfeldt and Ostman⁸³⁾.

The physiological mechanisms involved appear to differ according to the interval between the two meals, dinner to breakfast (about 10–12 h) or breakfast to lunch (about 4–5 h). The shorter period seems to be sufficient for the low-GI feature of the cereal product consumed at breakfast to reduce the glucose response at lunch, probably by improving blood sugar regulation and insulin sensitivity⁽Reference Nilsson, Ostman and Holst⁵⁴⁾. The longer interval between dinner and breakfast involved the fermentation of indigestible carbohydrates in the colon, reduced plasma NEFA and modified glucose metabolism. This indicates that the presence of specific dietary fibre (soluble or insoluble or RS) in boiled barley kernels is more significant in this ‘second-meal effect’ than is its low GI.

SCFA produced during the fermentation of fibre in the colon might be particularly involved⁽Reference Nilsson, Granfeldt and Ostman⁸³⁾ through at least three potential processes: a possible decrease of the gastric emptying rate by SCFA as reviewed in rats and humans⁽Reference Cherbut⁸⁴⁾, notably through an increased level of the polypeptide YY in blood by SCFA, that may lead to a reduced rate of glucose entry into the bloodstream; the ability of propionate and acetate to reduce serum NEFA in humans⁽Reference Wolever, Spadafora and Eshuis⁸⁵⁾, circulating fatty acids being able to induce peripheral and hepatic insulin resistance in humans⁽Reference Homko, Cheung and Boden⁸⁶⁾; and, finally, the possible specific action of propionate on glucose metabolism by increasing hepatic glycolysis and decreasing hepatic glucose production as shown in isolated rat hepatocytes⁽Reference Anderson and Bridges⁸⁷⁾. A later study on healthy subjects⁽Reference Nilsson, Ostman and Holst⁵⁴⁾ confirmed that the low-GI feature of the products consumed in the evening meal was not per se involved in the improved glucose response at breakfast, and that the lower plasma NEFA concentration combined with the high plasma propionate content (from fermentation in the colon) contributed to the overnight benefits in terms of glucose tolerance⁽Reference Nilsson, Granfeldt and Ostman⁸³⁾. The quantity and quality of the indigestible carbohydrates (for example, barley fibre and RS) are most important. There is also an important relationship between gut microbial metabolism and insulin resistance⁽Reference Nilsson, Ostman and Holst⁵⁴⁾.

These results suggest that the influence of carbohydrates on glucose tolerance over a longer time (semi-acute) is optimal when the food structure is preserved (i.e. a low-GI feature) and content of RS and/or fibre is high (i.e. production of specific SCFA). Eating barley or rye kernels for breakfast resulted in lower cumulative postprandial increases in blood glucose after breakfast, lunch and dinner (a total of 9·5 h) than did a breakfast of white-wheat bread⁽Reference Nilsson, Ostman and Granfeldt⁵²⁾. From a technological point of view, the quantity and quality of the indigestible carbohydrates is therefore particularly important, in addition to preserving a more or less intact botanical food structure, for a better control of glucose metabolism, especially to prevent type 2 diabetes.

Whole-grain cereals as rich sources of anti-carcinogenic compounds

A survey of 61 433 women found that a high consumption of whole grains (hard whole-grain rye bread, soft whole-grain bread, porridge, and cold breakfast cereals) was associated with a lower risk of colon cancer⁽Reference Larsson, Giovannucci and Bergkvist¹¹⁾. An inverse association between cereal fibre and whole-grain cereal consumption and small-intestinal cancer incidence has also been reported⁽Reference Schatzkin, Park and Leitzmann¹²⁾. The roles played by dietary fibre and phytochemicals in preventing intestinal cancer in humans and animals have been reviewed and discussed for both human intervention and animal studies⁽Reference Slavin^45, Reference Ferguson and Harris^69, Reference Liu⁸⁸⁾. The positive action of the wheat bran oil on colon tumour incidence in rats (azoxymethane-induced cancer)⁽Reference Reddy, Hirose and Cohen⁸⁹⁾ and mice (Min cancer model)⁽Reference Sang, Ju and Lambert⁹⁰⁾ has also been demonstrated. This anti-carcinogenic effect is mainly attributed to the antioxidant and anti-inflammatory properties of several bioactive compounds, as increased oxidative stress and inflammation are involved in cancer aetiology⁽Reference Bartsch and Nair⁹¹⁾. Phenolic acids, flavonoids, carotenoids, vitamin E, n-3 fatty acids, lignan phyto-oestrogens, steroid saponins (found mainly in oats), phytic acid and Se are all potential suppressors of tumour growth, but human, animal and/or in vitro cell studies indicate that their mechanisms of action may differ (Tables 3 and 4)⁽Reference Slavin, Martini and Jacobs^46, Reference Ferguson and Harris^69, Reference Graf and Eaton^92–Reference Shamsuddin⁹⁵⁾. For example, cereal lignans are converted by fermentation into mammalian lignans or phyto-oestrogens (enterodiol and enterolactone). These may have a weak oestrogenic activity, and may protect against hormone-dependent cancers (prostate and breast cancers) and/or colon cancer⁽Reference Adlercreutz⁹⁶⁾. Studies on postmenopausal women, ovariectomised rats and liver and breast cancer cell cultures indicate that phyto-oestrogens inhibit cell proliferation by competing with oestradiol for type II oestrogen binding sites⁽Reference Adlercreutz, Mousavi and Clark^97, Reference Markaverich, Webb and Densmore⁹⁸⁾. Phytic acid would help reduce the rate of cell proliferation during the initiation and post-initiation stages (for example, decreased incidence of aberrant colon crypt foci) by complex mechanisms that involve its antioxidant properties, signal transduction pathways, gene regulation and immune response through enhancing the activity of natural killer cells⁽Reference Reddy⁹⁹⁾, and its anti-carcinogenic effect seems to be dose-dependent⁽Reference Ullah and Shamsuddin¹⁰⁰⁾. The high phytic acid content of whole-grain cereals (up to 6 % in wheat bran) has led to questions about whether the anti-cancer activity of wheat bran should be attributed more to phytic acid than to dietary fibre⁽Reference Ferguson and Harris^69, Reference Graf and Eaton⁹²⁾. Indeed, pure phytic acid is more efficient at reducing the incidence and multiplicity of mammary tumours in rats than is the bran fraction (All Bran; Kellogg^®)⁽Reference Vucenik, Yang and Shamsuddin¹⁰¹⁾. The many anti-carcinogenic actions of flavonoids include their ability to inhibit various stages of tumour development in animals⁽Reference Hollman and Katan¹⁰²⁾ and to reduce the mutagenicity of several dietary carcinogens in Salmonella typhimurium TA98NR⁽Reference Edenharder, Rauscher and Platt¹⁰³⁾. The anti-carcinogenic activity of ferulic acid is mainly attributed to its antioxidant capacity; it scavenges the free oxidative radicals that are involved in the aetiology of cancer, and to its ability to stimulate cytoprotective enzymes⁽Reference Barone, Calabrese and Mancuso^104, Reference Kawabata, Yamamoto and Hara¹⁰⁵⁾. Studies on azoxymethane-treated rats indicate that vitamin E and β-carotene inhibit the progression of aberrant crypt foci to colon cancer, especially the later stages of carcinogenesis, while wheat bran is better at inhibiting earlier stages⁽Reference Alabaster, Tang and Shivapurkar¹⁰⁶⁾. Lignins, by hydrophobically binding bile salts, might reduce the formation of carcinogens from them⁽Reference Eastwood and Girdwood^107, Reference Eastwood and Hamilton¹⁰⁸⁾. Their adsorptive ability would increase with increased methylation of the hydroxyl moieties on the phenyl-propane units⁽Reference Eastwood and Girdwood^107, Reference Eastwood and Hamilton¹⁰⁸⁾. Lignins also reduce DNA lesions in rat testicular cells and lymphocytes both in vitro and ex vivo ⁽Reference Labaj, Slamenova and Lazarova¹⁰⁹⁾. Se inhibits the occurrence of neoplasia in rats and mice, suggesting that an Se-poor diet is associated with an increased prevalence of neoplasia in specific human populations⁽Reference Wattenberg¹¹⁰⁾. This probably depends on the activity of the selenoprotein glutathione peroxidase, which is involved in the development of cancers⁽Reference Jablonska, Gromadzinska and Sobala¹¹¹⁾. Cereal bioactive compounds act via several other anti-mutagenic and anti-carcinogenic mechanisms⁽Reference Stavric¹¹²⁾. Important ones are the adsorption and dilution of carcinogens by insoluble dietary fibre and lignins⁽Reference Ferguson and Harris^69, Reference Alabaster, Tang and Shivapurkar^106, Reference Harris and Ferguson^113, Reference Harris, Roberton and Watson¹¹⁴⁾, and the action of SCFA produced by fibre fermentation⁽Reference Morita, Tanabe and Sugiyama¹¹⁵⁾. Butyrate is a major factor, as more is produced in the presence of RS, and favours apoptosis in human cancer cell lines⁽Reference Scheppach, Bartram and Richter⁶²⁾ and DNA repair in rats⁽Reference Toden, Bird and Topping¹¹⁶⁾. Interestingly, contrary to what was believed since the works of Burkitt emphasising the preponderant role of fibre in the prevention of Western diseases, notably colon cancer observed in Western countries and not in African rural population consuming high levels of dietary fibre⁽Reference Story and Kritchevsky¹¹⁷⁾, it is more and more believed today that the effect against colon cancer development might be before all attributed to RS⁽Reference Bauer-Marinovic, Florian and Muller-Schmehl¹¹⁸⁾, since a lower risk of colon cancer was recently observed in populations with a low level of fibre consumption but with a high intake of RS⁽Reference Bingham^119, Reference O'Keefe, Kidd and Espitalier-Noel¹²⁰⁾. This reinforces the idea that specific products of RS fermentation within the colon, such as butyric acid, are the active components. Betaine⁽Reference Cho, Willett and Colditz¹²¹⁾ may be added to the list of anti-carcinogenic compounds, as its concentration can reach 0·3 % in whole-grain wheat and 1·5 % in wheat bran (Table 2).

To summarise, the anti-carcinogenic effects of insoluble fibre (including lignin), phytochemicals and wheat bran oil can be distinguished. Insoluble fibre may act directly by adsorbing or diluting carcinogens (through increased faecal bulk by water absorption), or indirectly by decreasing colon pH (through SCFA production) and increasing butyrate production. The role of phytochemicals is complex and multi-factorial, and notably involves their antioxidant properties since increased oxidative stress is a major factor in the aetiology of cancers⁽Reference Bartsch and Nair^91, Reference Klaunig, Xu and Isenberg¹²²⁾. The exact components of wheat bran oil that reduce the development of colon tumours are still to be identified⁽Reference Reddy, Hirose and Cohen^89, Reference Sang, Ju and Lambert⁹⁰⁾. However, animal experiments indicate that dietary fibre, particularly soluble fibre, may not protect against or even enhance carcinogenesis. This may be due to the abrasive property of insoluble fibre, a too low pH ( < 6·5) reached within the colon following soluble fibre and RS fermentation, the enhanced colon glucuronidase activity (that converts conjugated carcinogens to free carcinogens) and the increased production of secondary bile acids (tumour promoters) within the colon due to the increased viscosity of some soluble fibre which reduces the reabsorption of bile salt in the small intestine⁽Reference Harris and Ferguson¹²³⁾.

Whole-grain cereals as a rich source of antioxidants

Whole-grain cereals can protect the body against the increased oxidative stress that is involved and/or associated with all the major chronic diseases: metabolic syndrome⁽Reference Ford, Mokdad and Giles¹²⁴⁾, obesity⁽Reference Higdon and Frei^125, Reference Keaney, Larson and Vasan¹²⁶⁾, diabetes⁽Reference Evans, Goldfine and Maddux^127, Reference Maiese, Morhan and Chong¹²⁸⁾, cancers⁽Reference Bartsch and Nair⁹¹⁾ and CVD⁽Reference Cai and Harrison^129, Reference Castelao and Gago-Dominguez¹³⁰⁾. Whole-grain cereals are good sources of antioxidants (thirty-one compounds or groups of compounds are listed in Table 4), as shown by measurements made in vitro of the antioxidant capacity of whole-grain, bran and germ fractions⁽Reference Martinez-Tome, Murcia and Frega^131–Reference Zielinski and Kozlowska¹³⁵⁾. However, this may not be the same in vivo ⁽Reference Fardet, Rock and Rémésy¹³⁶⁾, and up to today, to my knowledge, the number of studies exploring the in vivo antioxidant effect of whole-grain cereals and/or their fractions in human subjects does not exceed eleven⁽Reference Andersson, Tengblad and Karlstrom^137–Reference Wang, Han and Zhang¹⁴⁷⁾. The antioxidants in cereals differ in their structure and mode of action⁽Reference Slavin, Martini and Jacobs^46, Reference Fardet, Rock and Rémésy¹³⁶⁾. There are indirect antioxidants, such as Fe, Zn, Cu and Se, which act as cofactors of antioxidant enzymes, and direct radical scavengers such as ferulic acid, other polyphenols (lignans, anthocyanins and alkylresorcinols), carotenoids, vitamin E and compounds specific to cereals other than wheat, such as γ-oryzanol in rice and avenanthramides in oats. These can neutralise free radicals and/or stop the chain reactions that lead to the production of oxidative radical compounds (for example, the lipid chain peroxidation stopped by vitamin E within cell membranes). Another antioxidant mechanism involves phytic acid, which can chelate Fe and thus stop the Fenton reaction producing the highly oxidative and damaging free radical OH^∙, ultimately reducing lipid peroxidation⁽Reference Graf, Empson and Eaton¹⁴⁸⁾. Lignins are also considered to be antioxidants in vitro (radical-scavenging activity)⁽Reference Dizhbite, Telysheva and Jurkjane¹⁴⁹⁾, but precisely how they act in vivo is not known: they may adsorb oxidative damaging compounds within the digestive tract in a way similar to bile salts adsorption⁽Reference Eastwood and Girdwood^107, Reference Eastwood and Hamilton¹⁰⁸⁾. While the action of cereal antioxidants is not well characterised once the epithelial barrier has been crossed, there is a growing belief that cereal antioxidants protect the intestinal epithelium cells from oxygen-derived free radicals⁽Reference Fardet, Rock and Rémésy^136, Reference Vitaglione, Napolitano and Fogliano¹⁵⁰⁾, particularly those produced by bacteria that may help form active carcinogens by oxidising procarcinogens or those that may result from increased stool Fe content (Fenton reaction) due to a diet high in red meat⁽Reference Babbs¹⁵¹⁾. The concept of ‘dietary fibre-bound phytochemicals/phenolic compounds’ was proposed recently⁽Reference Liu^18, Reference Vitaglione, Napolitano and Fogliano¹⁵⁰⁾. The authors suggest that the antioxidant polyphenols survive digestion in the small intestine because most of them are bound to fibre (for example, esterification of phenolic acids to arabinoxylans) in the cereal food matrix. They reach the colon where the fibre is fermented and some of the antioxidants are released⁽Reference Vitaglione, Napolitano and Fogliano¹⁵⁰⁾. Vitaglione et al. hypothesised ‘the slow and continuous release in the gut of the dietary fibre bound antioxidants’, such as that of ferulic acid, which will determine the effects of these antioxidants, and considered dietary fibre to be a ‘natural functional ingredient to deliver phenolic compounds into the gut’⁽Reference Vitaglione, Napolitano and Fogliano¹⁵⁰⁾. For example, only 0·5–5 % of the ferulic acid is absorbed within the small intestine, mainly the soluble free fraction⁽Reference Adam, Crespy and Levrat-Verny^152–Reference Rondini, Peyrat-Maillard and Marsset-Baglieri¹⁵⁴⁾, and this typical whole-grain wheat phenolic acid (about 90 % of total phenolic acids) would probably exert a major action in the protection of the colon from cancer. Thus, bound antioxidant phenolic acids might act along the whole length of the digestive tract by trapping oxidative compounds. This fraction of bound polyphenols has often led to an important underestimation of the real antioxidant capacity of whole-grain cereals – and of their fractions – as measured in vitro and generally based on the measurement of the easily extractable polyphenol fraction⁽Reference Perez-Jimenez and Saura-Calixto^133, Reference Pellegrini, Serafini and Salvatore¹⁵⁵⁾. In vivo studies are now needed to examine this hypothesis, and to characterise and quantify this potential antioxidant effect within the digestive tract.

The antioxidants in whole-grain cereals act via different, complex, and synergetic mechanisms in vivo. However, the antioxidant action of whole-grain cereals has not yet been convincingly validated in human subjects and requires further exploration.

Whole-grain cereals as rich sources of magnesium

Among plant-based foods, whole-grain cereals, together with legumes, nuts and seeds, are one of the best sources of Mg: whole-grain wheat contains 104 mg Mg/100 g, wheat bran 515 mg, and wheat germ 245 mg (Table 2). The high Mg content of whole-grain cereals may explain its favourable impact on insulin sensitivity and diabetes risk (Fig. 2)⁽Reference McCarty¹⁵⁶⁾, diabetes being otherwise frequently associated with Mg deficiency⁽Reference Durlach and Collery¹⁵⁷⁾. Mg can increase insulin secretion and the rate of glucose clearance from the blood in humans⁽Reference Paolisso, Sgambato and Gambardella^158, Reference Paolisso, Sgambato and Pizza¹⁵⁹⁾. This was also proposed to explain the lower insulin response in obese and overweight adults following the consumption of a whole-grain-based diet as compared with those on a refined cereal-based diet⁽Reference Pereira, Jacobs and Pins¹⁶⁰⁾. High-Mg diets reduce insulin resistance in rats fed a high-fructose diet⁽Reference Balon, Jasman and Scott¹⁶¹⁾; they also reduce the development of spontaneous diabetes in obese Zucker rats, a model of non-insulin-dependent diabetes mellitus, but these rats had to be given Mg before the onset of diabetes to obtain protection⁽Reference Balon, Gu and Tokuyama¹⁶²⁾. Most explanations of the prevention of type 2 diabetes by Mg are based on the finding that Mg stimulates insulin-dependent glucose uptake in elderly subjects⁽Reference Paolisso, Sgambato and Gambardella^158, Reference Gould and Chaudry¹⁶³⁾. It also protects Mg-deficient animals from the production of reactive oxygen species⁽Reference Weglicki, Mak and Kramer¹⁶⁴⁾. Reactive oxygen species are partly responsible for the increased hyperglycaemia-mediated oxidative stress in diabetic subjects⁽Reference Ceriello, Bortolotti and Crescentini^165, Reference Pereira, Ferderbar and Bertolami¹⁶⁶⁾. Mg also acts as a mild physiological Ca antagonist⁽Reference Iseri and French¹⁶⁷⁾. Obese and diabetic patients with insulin resistance have excess free intracellular Ca and these two clinical conditions are associated with hypertension⁽Reference Resnick¹⁶⁸⁾. In addition, Mg helps keep the concentration of intracellular Ca optimal through various complex cellular mechanisms involving Ca channels, Ca sequestration/extrusion by the endoplasmic reticulum and Ca binding sites on proteins and membranes⁽Reference McCarty¹⁵⁶⁾. Finally, low serum plasma Mg has been positively associated with a higher risk of coronary atherosclerosis or acute thrombosis⁽Reference Liao, Folsom and Brancati¹⁶⁹⁾, suggesting that whole-grain cereal Mg might also contribute to the prevention of CVD. This may also involve the inhibition of platelet-dependent thrombosis by Mg supplementation in patients with coronary artery disease⁽Reference Shechter, Merz and Paul-Labrador¹⁷⁰⁾ and the positive effect of Mg upon blood pressure regulation in hypertensive patients⁽Reference Kawano, Matsuoka and Takishita¹⁷¹⁾. The capacity of a regular prolonged consumption of whole-grain cereals to sustain a high plasma Mg concentration therefore deserves to be investigated in the context of type 2 diabetes prevention.

Fig. 2 Current accepted mechanisms for how whole grain protects against major chronic diseases (modified with permission from Professor I. Björck (University of Lund, Sweden); see the HealthGrain brochure for original diagram: ‘Progress in HEALTHGRAIN 2008’, a project from the European Community's Sixth Framework Programme, FOOD-CT-2005-514008, 2005–2010; see Poutanen et al. ⁽Reference Poutanen, Shepherd and Shewry⁴⁷⁸⁾ for more details about the Project). GI, glycaemic index; II, insulinaemic index.

The action of some anti-nutrients on starch hydrolysis and glycaemia

Whole-grain cereals are also a source of antinutrients with both adverse and positive health effects. The most important are phytic acid, lectins, tannins, saponins and inhibitors of enzymes such as proteases and α-amylases. Their main negative effect is their ability to reduce the bioavailability and the absorption of some nutrients (for example, the chelation of minerals by phytic acid and tannins), the binding of lectins to epithelial cells that damages the intestinal microvillae, and inhibition of digestive enzymes by tannins, which inhibits growth in animals⁽Reference Al-Mamary, Al-Habori and Al-Aghbari^172, Reference Thompson¹⁷³⁾. Cereal products in the human diet are cooked; this leads to losses of antinutrients such as lectins and enzyme inhibitors, and the major health outcome appears to be the low dietary Fe bioavailability in African populations that consume sorghum or finger millet-based beverages, gruels and porridges, both cereals containing phytic acid and a high tannin content⁽Reference Gillooly, Bothwell and Charlton^174, Reference Tatala, Svanberg and Mduma¹⁷⁵⁾. For example, the phytate and Fe-binding phenolic compounds in whole-grain millet flour may reach 0·6 g/100 g (DW)⁽Reference Lestienne, Besancon and Caporiccio¹⁷⁶⁾. This is one of the key factors responsible for Fe-deficiency anaemia in developing countries⁽Reference Tatala, Svanberg and Mduma¹⁷⁵⁾. On the other hand, the use of traditional processing such as germination, soaking, pre-fermentation and cooking may help to decrease the tannin and phytic acid contents, so improving Fe bioavailability⁽Reference Hassan and El Tinay^177–Reference Towo, Matuschek and Svanberg¹⁸⁰⁾.

However, phytic acid, lectins, protease inhibitors and tannins also contribute to the low-GI property of whole-grain foods⁽Reference Thompson^181, Reference Yoon, Thompson and Jenkins¹⁸²⁾. In wheat and derived whole-grain food products, since lectins and enzyme inhibitors are inactivated by cooking processes, this is primarily phytic acid which would reduce glycaemia through several potential mechanisms: thus, binding with proteins closely associated with starch, association with digestive enzymes, chelation of Ca required for α-amylase activity, direct binding with starch, effect on starch gelatinisation during cooking processes and slowing of gastric emptying rate might be involved⁽Reference Thompson¹⁸¹⁾.

Conclusion

The proposed mechanisms by which whole-grain cereals may protect the body are shown in Fig. 2. The most important ones are the preservation of food structure, fibre fermentation in the colon, the hypoglycaemic and hypoinsulinaemic, antioxidant, anti-inflammatory and anti-carcinogenic properties of several bioactive compounds, improved insulin sensitivity by Mg and reduced hyperhomocysteinaemia by betaine, a significant CVD risk factor (for details about betaine, see the ‘New hypotheses’ section below). However, an extensive list of all the bioactive compounds in whole-grain wheat and its fractions (Table 2), the ways they act and their health effects as isolated free compounds (Tables 3 and 4) makes it possible to formulate new hypotheses to explain the protective role of whole-grain cereals. Whole-grain cereals, particularly wheat and/or wheat bran and germ, are also a source of n-3 fatty acids (especially α-linolenic acid), sulfur compounds (reduced glutathione (GSH), oxidised glutathione (GSSG), methionine and cystine), oligosaccharides (fructans, raffinose and stachyose), P, Ca, Na, K, B vitamins, flavonoids (for example, anthocyanins and isoflavonoids), alkylresorcinols, betaine, choline, phytosterols, inositols, policosanol and melatonin. The actions of these compounds will be described in the next ‘New hypotheses’ section. The antioxidant hypothesis will be discussed with a broader perspective, as well as the health benefits of active compounds from whole-grain cereals that are less often studied, such as B vitamins, sulfur compounds, methyl donors and lipotropes, α-linolenic acid, lignins, oligosaccharides, policosanol and melatonin.

The antioxidant hypothesis must not be reduced to free radical scavenging and antioxidant enzyme activation

There is more and more evidence that the primary effect of antioxidants from whole-grain cereals is in the digestive tract, where they protect intestinal epithelial cells from attack by free radicals⁽Reference Fardet, Rock and Rémésy^136, Reference Vitaglione, Napolitano and Fogliano¹⁵⁰⁾. However, the mechanisms by which antioxidants that cross the intestinal barrier protect the body remain uncertain. Published studies on animals and human subjects fed the free compounds give rise to new explanations of the antioxidant protection by whole-grain cereals. The antioxidant action of whole-grain cereals might be multi-factorial and much more complex than it first appears. There are at least four new mechanisms to be studied in the context of whole-grain cereals: the action of polyphenols on cell signalling and gene regulation modifying the redox status of tissues and cells, the action of sulfur amino acids on glutathione synthesis, the possible stimulation of endogenous antioxidants by whole-grain cereal bioactive compounds, and the underestimated antioxidant properties of phytic acid and lignin.

Whole-grain cereals as a source of bioactive compounds with underestimated physiological effects

The nutrigenomic approach

Nutrigenomics in nutrition is devoted to the study of the influence of dietary interventions on gene transcription (transcriptome), protein synthesis (proteome) and metabolites (metabolome, the whole set of metabolites) in cells, body fluids and tissues⁽Reference Elliott, Pico and Dommels^322–Reference Zeisel³²⁶⁾. One of the most important objectives of nutrigenomics is to detect and identify early metabolic disturbances and their regulation (for example, in relation to oxidative stress or inflammation) that can lead to more serious chronic diseases. The possibility of detecting some diseases early could change clinical nutrition and public health practices⁽Reference Zeisel³²⁶⁾. This implies studying the effects of bioactive compounds in whole-grain cereals on gene expression, protein synthesis and the metabolome. In the field of nutritional studies, besides the measurement of usual biomarkers such as plasma glucose (for example, GI) or urinary lipid peroxides (oxidative stress index), it seems particularly important to focus on the metabolome, which reflects both the endproducts of metabolism and the changes over time of metabolism following food consumption. While many metabolomic studies have been done with isolated compounds, notably in pharmacology for drug toxicity⁽Reference Keun³²⁷⁾, very few have been done with complex food products. In metabolomics and nutrition, only a few studies have been performed⁽Reference Rezzi, Ramadan and Fay³²⁸⁾: to characterise the metabolic effect of energy restriction⁽Reference Selman, Kerrison and Cooray³²⁹⁾, vitamin deficiency⁽Reference Griffin, Muller and Woograsingh³³⁰⁾ or of intake of PUFA-rich oils⁽Reference Mutch, Grigorov and Berger³³¹⁾, antioxidant-rich foods such as soya⁽Reference Solanky, Bailey and Beckwith-Hall³³²⁾, chamomile⁽Reference Wang, Tang and Nicholson³³³⁾ and tea⁽Reference Van Dorsten, Daykin and Mulder³³⁴⁾, or of pure dietary antioxidants such as epicatechin⁽Reference Solanky, Bailey and Holmes³³⁵⁾, catechin⁽Reference Fardet, Llorach and Martin³³⁶⁾ or ferulic and sinapic acids and lignins⁽Reference Fardet, Llorach and Orsoni²⁷¹⁾. Studies on rats have been carried out using the metabolomic approach to explore the metabolic fate and the effect on endogenous metabolism of whole-grain and refined wheat flours⁽Reference Fardet, Canlet and Gottardi²³⁰⁾ and of lignin-enriched wheat bran lignins⁽Reference Fardet, Llorach and Orsoni²⁷¹⁾. It has thus been shown that whole-grain wheat flour consumption leads to significant increases in liver betaine and GSH and decreases in some liver lipids, but has no effect on conventional lipid and oxidative stress biomarkers. It also causes a greater urinary excretion of tricarboxylic acid cycle intermediates, aromatic amino acids and hippurate (from phenolic acid degradation in the colon). When the diet was changed to refined wheat flour, a new metabolic balance was reached within 48 h, and conversely from refined to whole-grain flour (Fig. 3)⁽Reference Fardet, Canlet and Gottardi²³⁰⁾. The metabolomic approach also showed that rats did not appear to metabolise lignins from wheat bran within 2 d of the regimen, but they are likely to affect endogenous metabolism through mechanisms which need to be elucidated⁽Reference Fardet, Llorach and Orsoni²⁷¹⁾. Results are convincing in that new metabolic effects have been unravelled using this new open approach, for example, the role of symbiotic microbiota in triggering diet-induced mechanisms of steatosis⁽Reference Dumas, Barton and Toye³³⁷⁾ or some specific metabolic pathway disturbances in diabetic rats⁽Reference Zhang, Nagana Gowda and Asiago³³⁸⁾, thus improving our understanding of diseases and the mechanisms responsible for them. However, more significant conclusions could be drawn once the databases for compound identification are completed and distributed. To my knowledge, few if any studies have investigated the effect of consuming complex whole-grain cereals and their fractions on gene expression. The tools are now available to study this, which would provide important information about which gene-regulated metabolic pathways are stimulated by the synergetic action of the bioactive compounds in whole-grain cereals, not the restricted action of isolated compounds. Thus, nutrigenomics should enable us to better characterise the metabolic pathways affected in vivo by the antioxidants in whole-grain cereals.

Fig. 3 Linear discriminant (LD) analysis score plot of the ¹H NMR urinary spectra highlighting the separation before, between and after the diet change (days 14–15) and between the urine sampling times (postprandial (PP) and post-absorptive (PA)). (- - - -), Refined flour followed by whole-grain flour consumption (RF-WGF) group; (—), whole-grain flour followed by refined flour consumption (WGF-RF) group. Each polygon represents the limits of the metabolic profile obtained for the ten rats of a given group at a given day and urine sampling time. Urine samples were collected from days 13 to 28 (for details, see Fardet et al. ⁽Reference Fardet, Canlet and Gottardi²³⁰⁾).

Conclusion

The metabolic fate and health effects of major compounds such as lignin (up to 9 % in wheat bran), ferulic acid (up to 0·6 % in wheat bran), phytic acid (up to 6 % in wheat bran) and betaine (up to 1·5 % in wheat bran) (Table 2) have been little studied when originating from whole-grain cereals. Yet, these three compounds may account for about 11 % of wheat bran (Table 1), and therefore deserve to be studied more. Wheat germ also merits greater attention since it contains quite significant levels of bioactive compounds such as α-linolenic acid (about 530 mg/100 g), GSH (about 133 mg/100 g), GSSG (about 69 mg/100 g), thiamin (about 1·75 mg/100 g), vitamin E (about 27·1 mg total tocols/100 g), flavonoids (about 300 mg/100 g), betaine (about 851 mg/100 g), choline (about 223 mg/100 g), myo-inositol (>11 mg/100 g) and phytosterols (about 430 mg/100 g) (Table 2). It thus contains 2·5 % of vitamins and minerals, at least 1·6 % of lipotropic compounds and 1·2 % of sulfur compounds. All these compounds are involved in the new hypotheses proposed here and their corresponding physiological mechanisms. Based on past and new hypotheses, a synthetic view of the mechanisms underlying the health benefits of whole-grain cereals and their fractions can be proposed (Fig. 4). The diagram purposefully illustrates the complexity of the mechanisms involved and their obvious synergy and interconnection in vivo. Due to this complexity, whole-grain cereal bioactive compounds are listed in Table 4, ranking according to the five major health outcomes generally considered in the literature: body-weight regulation, CVD, diabetes, cancers, and gut health; mental, brain and skeleton health being new proposed ways to explore. One important question remains: do bioactive compounds exert the same effects when they are free compounds and when they are in whole-grain cereals? This is notable, because their bioavailability in whole-grain cereals is probably lower than the free compounds (Table 2) and because the quantities in whole-grain cereal products do not match the daily human needs. Again, it is probably the summed and combined action of all the bioactive compounds on a particular physiological function (as illustrated in Fig. 4 and Table 4) which leads to improved specific physiological functions such as antioxidant status and glucose homeostasis, especially when whole-grain products are consumed daily, generating long-term health benefits. This is why it is urgent to carry out further in vivo studies both in rats and human subjects, to unravel the complex mechanisms activated by the consumption of highly complex foods such as whole-grain cereal products. Intervention studies on human subjects consuming whole-grain cereals are so rare that they should be carried out first. The non-invasive characteristic and high potential of the metabolomic approach for unravelling new metabolites and metabolic pathways affected by a given diet and its ability to explore the complexity inherent in metabolism means that it should accompany the measurement of the usual biomarkers in order to describe the metabolic actions of whole-grain cereals in all their complexity. The mechanisms described in Fig. 4 are complex, but are above all interconnected as in the whole organism. Metabolomics therefore seems to be the most appropriate tool for studying such an interconnectedness, and so provide a more realistic view of how whole-grain cereal bioactive compounds act in synergy. For example, inflammation, oxidative stress and immune system-related metabolic pathways are generally all involved in cancers, as is the case for other metabolic diseases in which there is a progressive metabolic imbalance following an unhealthy diet. Finally, genomic studies are needed on the action of whole-grain cereals on gene regulation, as bioactive compounds really exert their physiological effects within the cell. While isolated free bioactive compounds may be used for in vitro studies on cell cultures, studies in animals and human subjects should use an integrated ‘complex food approach’.

Fig. 4 Current and new proposed physiological mechanisms involved in protection by whole-grain cereals (adapted from Table 3). The dotted thin arrows () indicate the link between whole-grain bioactive compounds and protective physiological mechanisms, while the plain arrows () indicate the relationship between physiological mechanisms and health outcomes.

The bran fraction

The proportion of the bran fraction varies with the cereal type: for wheat, rice and maize, it is 10–16 % of the whole grain. The bran fraction in rice contains about 15–20 % oil⁽Reference Souci, Fachmann and Kraut^215, Reference Britz, Prasad and Moreau³³⁹⁾. This oil is rich in bioactive compounds and contains more than 100 different antioxidants, such as lipoic acid, a powerful antioxidant⁽Reference Packer, Witt and Tritschler^340, Reference Roy, Sen and Tritschler³⁴¹⁾ that helps prevent cognitive deficits, is beneficial in the treatment of Alzheimer's disease⁽Reference Maczurek, Hager and Kenklies³⁴²⁾, and may protect against risk factors of CVD⁽Reference Wollin and Jones³⁴³⁾. Rice bran contains tocotrienols (10·6 mg/100 g)⁽Reference Yu, Nehus and Badger³⁴⁴⁾, γ-oryzanol (281 mg/100 g)⁽Reference Yu, Nehus and Badger³⁴⁴⁾ and up to 1·2 % phytosterols⁽Reference Emmons, Peterson and Paul³⁴⁵⁾ such as β-sitosterol, all of which may help improve the blood lipid profile and reduce the risk of CVD⁽Reference Heinemann, Kullak-Ublick and Pietruck^346–Reference Wilson, Nicolosi and Woolfrey³⁴⁸⁾. Rice bran also contains up to 21 % dietary fibres⁽Reference Emmons, Peterson and Paul³⁴⁵⁾. Maize bran has more dietary fibre than wheat and rice bran, about 74–79 %⁽Reference Souci, Fachmann and Kraut^215, Reference Kahlon and Chow^349, Reference Saulnier, Vigouroux and Thibault³⁵⁰⁾. It contains about 4 % phenolic acids, about 50 % heteroxylans and about 20 % cellulose, and is almost devoid of lignins⁽Reference Saulnier, Vigouroux and Thibault³⁵⁰⁾. It is particularly rich in ferulic acid (up to 3 %), mainly in a very resistant (to enzymes) bound form⁽Reference Saulnier, Marot and Elgorriaga³⁵¹⁾. And, contrary to wheat for which phytate is essentially in the bran fraction, 90 % of maize phytate is in the germ fraction⁽Reference O'Dell, De Boland and Koityonann³⁵²⁾.

Some specific compounds

Some bioactive compounds are quite specific to certain cereals: γ-oryzanol in rice, avenanthramides and saponins in oats, and, although present in other cereals such as wheat, β(1 → 3)(1 → 4)-glucans in oats and barley, and alkylresorcinols in rye. Their mechanisms of action and health effects are shown in Table 3.

Optimising and controlling the delivery of bioactive compounds for improving health

There are great differences between the food content in a defined nutrient and the percentage really metabolised, or even absorbed. This is especially true for cereal products where numerous factors linked to the food matrix may limit the release of macro- and micronutrients. There is increasing evidence that the physical structure of natural cereal food matrices (for example, intact cereal kernels) or the artificial microstructure of processed cereal products may either favour or limit the bioavailability of nutrients, and thus their nutritional effects. However, differences in bioaccessibility–bioavailability of nutrients, particularly micronutrients, at present cannot be correlated with differences in long-term health effects, except for the positive health effects of starch and its so-called slowly digestible fraction⁽Reference Englyst, Kingman and Cummings^405, Reference Lehmann and Robin⁴⁰⁶⁾. The question is therefore: is there a positive correlation between increased or decreased bioaccessibility of a given nutrient and its health effect? This probably depends on the nutrient considered and on the health status of the subject. For example, the rapid release of glucose from starch digestion into the bloodstream is advantageous in some situations (for example, the urgent need for glucose for brain or muscles to function, as for immediate intellectual and physical efforts), and harmful in other situations (for example, type 2 diabetes). The same approach is now being developed for proteins (slow v. rapid proteins) and lipids for which their physical state and/or their physico-chemical properties may influence the release of amino acids and fatty acids, respectively, into the bloodstream. The resulting significant metabolic impact could be used in some situations such as diabetic subjects⁽Reference Marangoni, Idziak and Rush⁴⁰⁷⁾, the elderly⁽Reference Remond, Machebeuf and Yven⁴⁰⁸⁾ and for patients on enteral nutrition suffering from pancreatic insufficiency to adequately hydrolyse lipids⁽Reference Armand, Pasquier and Andre⁴⁰⁹⁾.

In vitro bioaccessibility and in vivo bioavailability studies with vegetables and whole-grain cereals and/or their fractions have clearly shown that food structure affects the bioavailability of polyphenols, carotenoids, minerals, trace elements and vitamins (Table 2)⁽Reference Parada and Aguilera⁴⁰³⁾. Table 2 shows the results of bioavailability studies on whole-grain wheat products and wheat bran. Much data are still lacking: studies exploring the bioavailability of compounds in whole-grain cereals are scarce and the products are often consumed as part of a complex diet that also supplies the same bioactive compounds from other foods. For example, studies on mineral or trace element bioavailability in rats often included mineral mixtures that made it difficult to determine the exact apparent absorption of the mineral supplied by the cereals. Thus, radiolabelled cereal products should be used more frequently to answer such questions. The few data obtained show that bioactive compounds are far from being 100 % bioavailable within the small intestine. No more than 5 % of the ferulic acid in wheat bran is released into the small intestine, so that most reaches the colon where it can exert an antioxidant protective action on the gut epithelium. On the other hand, there is convincing evidence that the small proportion absorbed in the small intestine can affect cell signalling and the activation or repression of some genes. Thus, in a way similarly to starch, it seems that two fractions of ferulic acid can be defined: the rapidly available ferulic acid released and absorbed in the small intestine (i.e. free and soluble-conjugated), and slowly available ferulic acid gradually released mainly in the colon (i.e. ester-linked)⁽Reference Kroon, Faulds and Ryden²⁶⁴⁾, each fraction having its own health benefits.

Betaine (about 0·9 % of wheat bran; Table 1), unlike ferulic acid, is probably much more bioavailable since it is not bound to other constituents: is there a need to slow down its release and to favour a fraction reaching the colon, for example, for improving its anticancer effect⁽Reference Giovannucci, Rimm and Ascherio⁴¹⁰⁾? The same issue, that is the optimal bioavailability to reach, might be questioned for polyphenols such as lignans and alkylresorcinols, vitamins and minerals, and phytosterols. The problem for phytic acid is slightly different; we need to know the extent to which it is reasonable to pre-hydrolyse it in order to combine a maximum mineral bioavailability with its antioxidant effect in the gut against free radicals produced by microbiota, and from its potential hypoglycaemic effect as well.

Otherwise, the case of fibre is not yet resolved for whole-grain wheat which contains more insoluble fibre than soluble fibre (soluble:total fibre ratio is about 0·16; calculated from Table 2): what would be the optimum ratio of soluble:total fibre to reach? It is not known to what extent it would be beneficial to increase the soluble fibre content, for example, by pre-hydrolysing insoluble arabinoxylans to soluble arabinoxylans (soluble:total arabinoxylans ratio is about 0·18; calculated from Table 2). Soluble fibres may be beneficial to health by reducing the postprandial glucose response through increased viscosity⁽Reference Lu, Walker and Muir⁴¹¹⁾ (see Tables 3 and 4), but they may also be harmful, by, for example, increasing the risk of colon cancer⁽Reference Moore, Park and Tsuda⁴¹²⁾.

Provided it has positive health benefits, the range by which industrial processes can improve the bioaccessibility and bioavailability of cereal bioactive compounds is therefore large. This approach has been applied to starch with success⁽Reference Björck and Asp⁴¹³⁾, by controlling its delivery in the gut by rendering it more slowly hydrolysed (i.e. slowly digestible starch) within the small intestine, or by making it inaccessible to α-amylase (i.e. RS), so that a fraction of starch reaches the colon where it is fermented to the anti-carcinogenic molecule butyrate, the preferred fuel for colonocytes (see Whole-grain cereals and butyrate production section). Technologists know how to modulate the proportions of these three fractions in cereal products, i.e. rapidly, slowly and indigestible starch. RS is representative of the different ways it can be used by breeders and technologists to control the delivering of a compound, i.e. starch, within the digestive tract. It has been seen that the RS content of whole-grain products may be very high, up to 12 % in ordinary barley kernels and even 22 % by combining intact botanical structure with a high-amylose barley variety⁽Reference Nilsson, Ostman and Holst⁵⁴⁾. The formation of RS can be technologically favoured through starch encapsulation within the cereal food matrix by protein or fibre networks (RS1), restricting starch granule gelatinisation (RS2), the use of high-amylose cereal varieties with a high content of retrograded starch (RS3) and/or chemical modification such as acylation (RS4). RS is now considered to be a prebiotic compound that can positively modify microbiota growth in quality and quantity within the colon⁽Reference Dongowski, Jacobasch and Schmiedl^414, Reference Topping, Fukushima and Bird⁴¹⁵⁾. If technologists may be able to modify processing parameters such as temperature, extrusion pressure, retrogradation and/or chemical modification to increase the RS content, breeders can select high-amylose cereal varieties⁽Reference Chanvrier, Appelqvist and Bird^294, Reference Rahman, Bird and Regina⁴¹⁶⁾, amylose being more slowly digested than amylopectin⁽Reference Goddard, Young and Marcus^417, Reference Hallfrisch and Behall⁴¹⁸⁾.

The traditional use of fermentation and the development of new technologies

Fig. 5 shows the ways in which the nutritional quality of whole-grain cereals can be improved. There are mainly three: the growing conditions, the genetic approach and through technological processes.

Fig. 5 Ways for improving cereal product nutritional quality. RS, resistant starch.

The importance of preserving bran and germ fractions

The bioactive compounds in whole-grain cereals are unevenly distributed (Fig. 1). Some (mainly soluble fibre, Se, some B vitamins, carotenoids and flavonoids) are present in significant quantities in the endosperm, but most are in the bran (especially the aleurone layer) and germ fractions. This fact alone shows the importance of preserving these fractions in cereal products, at least in the most currently consumed forms of breads and breakfast cereals, and to a lesser extent pasta, crackers and biscuits. Some products consumed on special occasions (i.e. generally not at breakfast, lunch or dinner), such as cakes, pastries and viennoiseries, use very refined flours (extraction rate of 70–82 %), and it is probably not meaningful to use less refined flours. To preserve the bran and germ fractions means either reincorporating fractions later in the recipe or using the whole-grain cereal so as to maintain its botanical structure relatively intact during processing. However, reincorporation of the bran and germ fractions implies destroying the botanical structure with the loss of its health benefits (for example, increased satiety or RS content), unless technological processes can yield a cereal product with an artificial compact food structure as for pasta⁽Reference Fardet, Hoebler and Baldwin⁴⁷²⁾ or breads with decreased loaf volume⁽Reference Burton and Lightowler⁴⁶¹⁾.

The concept of the ‘whole-grain package’

The content of individual bioactive compounds in whole grain often seems too low for them to have any significant or lasting physiological effects. It is becoming more and more evident that the synergetic action of several bioactive compounds contributes to health protection and/or the maintenance of one physiological function, not just one compound. Fig. 1 and Table 4 illustrate this concept of the ‘whole-grain package’: thus, obesity/body-weight regulation, CVD, type 2 diabetes, cancers, gut, mental/nervous system and skeleton health may be potentially protected by at least, respectively, ten, thirty-four, seventeen, thirty-two, ten, twenty-six and sixteen different bioactive compounds and/or groups of compounds (i.e. oligosaccharides, tocols, phenolic acids, flavonoids, saponins, inositols, γ-oryzanol, lignans and alkylresorcinols). Because of their many protective bioactive compounds (at least twenty-six), whole-grain cereals are particularly suitable for protecting the body from CVD, cancers and mental/nervous system disorders. The long-term protection against mental or nervous system disorders by consuming whole-grain cereal products therefore deserves to be studied in human subjects, notably because depression ranks among the major causes of mortality and disability with an overall prevalence of 5–8 %⁽Reference Coppen and Bolander-Gouaille²⁷⁴⁾. It is also remarkable that at least thirty compounds and/or groups of compounds may participate in antioxidant protection through different mechanisms (Tables 3 and 4), which approximately corresponds to a total of at least 3·9, 13·4 and 6·3 % of the whole-grain wheat, wheat bran and germ fractions (Tables 1 and 2). As most age-related and chronic diseases are associated with increased oxidative stress, the regular consumption of whole-grain cereal products should benefit all of us, but particularly the elderly.

The importance of pesticides and mycotoxins

Since whole-grain cereals include by definition the outer parts of the grain, they may contain pesticides and mycotoxins (for example, zearalenone and deoxynivalenol in wheat or fumonisin in maize). Their presence should not decrease the benefits of bioactive compounds also mainly contained in the outer layers. For example, there may be a relationship between the consumption of fumonisin-contaminated maize in some regions of the world (for example, China and South Africa) and the occurrence of oesophageal cancers⁽Reference Chu and Li^473, Reference Rheeder, Marasas and Thiel⁴⁷⁴⁾. However, more generally, the consequences of long-term consumption of high quantities of mycotoxin-contaminated cereal grains for human health (i.e. toxicological effects) are not well known. The link between some cancers and exposure to pesticides has been well established, particularly among farmers⁽Reference Lebailly, Niez and Baldi⁴⁷⁵⁾. It is therefore particularly relevant that recommendations for the consumption of more whole-grain cereal products should be accompanied by the production of less contaminated cereals, such as those from organic agriculture devoid of pesticides.

Perspectives

It is surprising to note that, although numerous epidemiological surveys have shown a significant and positive association between whole-grain cereal consumption and the prevention of several chronic diseases, fewer studies have been performed on the mechanisms involved. For example, to my knowledge, no more than eleven studies have examined the antioxidant hypothesis by postprandial or intervention studies in human subjects to investigate the antioxidant effect of whole-grain cereals, bran or germ⁽Reference Fardet, Rock and Rémésy¹³⁶⁾, with only a recent postprandial study on human subjects consuming wheat bran⁽Reference Price, Welch and Lee-Manion¹⁴⁶⁾. Therefore, there is a real gap between observational studies and the elucidation of the mechanisms involved. The mechanisms are certainly complex, as has been seen. But more data are needed on the mechanisms involved so as to prepare strong, convincing arguments for an increased consumption of whole-grain cereal products by the public, to better inform health professionals about their health benefits, to favour their marketing by the food industry and to develop new health claims in the near future.