Resistant starch (RS) is defined as a component of dietary starch that is not absorbed in the small intestine of healthy individuals and thus reaches the colon undigested, similar to dietary fibre 1. Evidence is mounting to suggest that RS is a protective agent against several serious gastrointestinal disorders 2 including that of colorectal cancer 34. The mechanism for protection may be associated with metabolic products of anaerobic bacterial fermentation of RS. Fermentation of RS produces short-chain fatty acids (SCFA) which lower luminal pH; increase bacterial biomass and faecal bulk; and modify the composition of the microbiota, especially by stimulating the growth of beneficial bacteria including bifidobacteria and lactobacilli 5. Butyrate, one of the principal colonic SCFA has generated the most interest and may be protective against colorectal cancer 67. Butyrate has been shown to be the primary energy source for the colonic epithelium 8; it inhibits the growth of cancer cells in vitro and forces a more normal differentiated phenotype 6. In addition, it is a potent pro-apoptotic agent 9 as it removes genetically disordered cells which help to counteract the biological consequences of genomic instability. Colonic production of butyrate by fermentation is associated with reduced tumour mass in an animal model, provided that fermentation is active in distal colon 10. More recently in an experimental animal studies, RS feeding was associated with the generation of increased butyrate levels in the colon as well as protecting against carcinogen-induced colon tumourigenesis 1112.

Further clarification of the health benefits of RS is, however, made difficult by the fact that it is heterogeneous in nature with the many different forms varying in their physicochemistry, digestibility, fermentability and fibre content. Types of RS have been traditionally classified into four main types based upon structural considerations and bacterial fermentability 1. RS1 includes physically entrapped starch within whole plant cells and food matrices (e.g. coarsely milled grain). RS2 consists of native starch granules that are highly resistant to digestion by α-amylases (e.g. green banana, HAMS). RS3 comprises retrograded starches, formed when starchy foods are cooked and cooled. RS4 comprises chemically modified starches (e.g. esterified starches) where the modification interferes with the amylolytic activity of digestive enzymes. Apart from these structural differences, different sources of RS may vary according to proportion of nondigestible starch, dietary fibre content, starch granule size and other physico-chemical differences that might result in different levels of digestibility and fermentation patterns 13. This might lead to different luminal conditions and hence variations in epithelial biological responsiveness. It has also been recently shown that there can be an important synbiotic interaction with colonic microbiota resulting in altered epithelial response to carcinogen-induced genomic lesions in the colon 5. Because of the diversity and forms of RS they cannot be expected to all perform in the same manner physiologically 14. As a consequence, different forms of RS might vary considerably in their effects on colonic biology and the consequences of that for colonic disorders such as cancer.

The aim of the current study was to investigate the effects of feeding high amylose maize starches (HAMS) that differed in their RS content, dietary fibre content and degree of processing on colonic fermentation and implications for colonic biology. We have explored their effects on luminal conditions, namely SCFA concentrations and pH, and the relationships to certain epithelial events potentially regulated by luminal conditions: specifically colonic epithelial proliferation and apoptotic response to a DNA-damaging agent. The study focused on RS2 and RS3 derived from HAMS and included native and hydrothermally (or heat-moisture) treated forms. Hydrothermal processing is a means of significantly increasing the dietary fibre content of HAMS. The starches examined were as follows: Hylon VII which is a RS2 type starch and is the base starch for the preparation of Hi-maize 240 and Hi-maize 260. Hi-maize 1043 is a RS2 type starch that has been hydrothermally prepared from HAMS, as reported by Bird et al. 15; although made from a different base HAMS, the Hi-maize 1043 and Hi-maize 260 have been made using similar hydrothermal methodology. Novelose 330 which is a retrograded RS3 generated from the hydrolysed products of corn starch, 16 and has not been previously reported in terms of apoptotic response to a DNA-damaging agent.

Our findings show that fermentation-dependent luminal events do differ between forms of HAMS, even between starches of the same physico-chemical type, namely RS2. The findings also show that epithelial events are dependent on RS-induced changes in the luminal environment. These significant consequences for epithelial biology might be relevant to risk for colorectal disease including neoplasia.

Increased fermentation, reflected in lowered pH and increased concentrations of SCFA in the digesta, has previously been reported with RS in rats 2021, pigs 15 and humans 3. We found high concentrations of SCFA, particularly butyrate with selected RS forms (i.e. Hi-maize 260 and HYLON VII); furthermore both of these maintained the highest butyrate levels along the length of the colon. We have previously shown that maintenance of active fermentation along the length of the colon, including high concentrations of butyrate, is significant in protecting against colorectal cancer development in animal models where, as in humans, cancers predominate in the distal colon 101122.

Plausible explanations that could account for maintenance of high butyrate levels throughout the colon include the RS type, RS amount and/or the dietary fibre content of individual sources of RS and the degree of treatment (specifically hydrothermal treatment). Unfortunately, the present study was unable explain what the critical characteristic of HAMS was that maintained high butyrate levels. It did not appear to be dietary fibre as Hi-maize 260 and HYLON VII which were the best diets for butyrate production are both RS2-type starches, contain similar amounts of RS (46% and 48% respectively) 23 yet the total dietary fibre levels are quite different (60% and 18% respectively) 23. The hydrothermal treatment also does not seem to have significantly improved fermentation, because HYLON VII is the base starch for the hydrothermal preparation of Hi-maize 260, and it was both of these RS forms that performed the best in terms of SCFA production including butyrate. Quite significant differences in effects on butyrate production were observed between the different RS2-type starches, however, on balance the RS2-type starches tended to perform much better than the single RS3-type starch (ie. Novelose 330). Other factors not measured in the current study that might account for the higher butyrate production could be related to alterations in the colonic microbiota. Certain dietary carbohydrates have been shown to have the capacity to stimulate the growth of particular butyrate producing bacteria directly, or indirectly through stimulation of non-butyrate producing bacteria via metabolic cross-feeding 24

Butyrate, the physiologically most important SCFA, is produced by anaerobic fermentation of carbohydrate and other substrates in the colonic lumen and is considered to be protective against CRC 25. In addition to our rodent studies linking butyrate to protection in rodent models 101122 others have also shown that high butyrate producing substrates 26 and delivery of butyrate directly to the distal colonic mucosa have been linked to protection against the initial stages of colon carcinogenesis 27. Butyrate production has been also been linked to reversing the genetic damage and loss of mucosal barrier in rats induced by feeding higher protein diets 2829. In addition to these anti-carcinogenic effects butyrate may also exert anti-inflammatory effects in vivo 3031.

Other luminal consequences of fermentation were also significantly altered by feeding these HAMS. A lower pH throughout the large bowel was observed. An acidified colon is associated with a decreases risk for CRC cancer 32. Lower pH values are believed to prevent the overgrowth of pH-sensitive pathogenic bacteria and lower the production of potentially harmful toxic or carcinogenic products in the colon, including secondary bile acids and protein fermentation products (ammonia and phenols) 233. Faecal output was increased by feeding cellulose and all the RS forms. The RS3 'NOVELOSE 330' was the most effective diet at increasing large bowel digesta weight and faecal output; this may be a characteristic of being a retrograded starch.

Luminal events are clearly linked to changes in epithelial biology. Apoptosis is an important innate cellular event for protection against the development of colorectal cancer. This includes removal of cells with genomic instability that have developed during oncogenesis 34 and deletion of cells suffering DNA insult from genotoxic agents such as carcinogens 35. Enhancement of apoptosis during initiation events increases elimination of mutated cells that might otherwise progress to malignancy 36 and defective apoptotic response is associated with increased risk 37. In the present study we have examined the relationship between butyrate and the apoptotic response to carcinogen by Spearman correlation procedures using data from individual animals performed regardless of dietary group. We found that butyrate concentration in both the caecum and faeces was positively correlated with AOM-induced apoptosis in the distal colon. This is consistent with earlier reports and confirms an important facilitatory role of butyrate for apoptosis in epithelial homeostasis 1820. Production of butyrate by fermentation of RS seems a likely explanation by which RS enhances the apoptotic response to carcinogen-induced genomic lesions. The present study showed that the apoptotic response to the carcinogen was significantly higher in the Hi-maize 260 group compared to the Control group, however Hi-maize 260 was not significantly different from the other RS containing groups. It appears that neither the degree of processing or dietary fibre content was directly related to the apoptotic response rather it was butyrate levels that predicted the apoptotic response

Colonic cell mass is dependent on the presence of RS or dietary fibre as crypt atrophy was observed in the rats fed a diet deprived of RS and dietary fibre. This effect is consistent with our previous studies 3839 and that of others 4041. All of the RS diets (ie. both RS2 and RS3) and the Cellulose diet reversed the atrophy which was observed with the fibre/RS free Control diet. Fermentative production of SCFA is considered to have a trophic affect on the colonic epithelium 2, in the present study cellulose reversed the colonic atrophy independent of SCFA production. Reversal of atrophy by a non-fermentable fibre such as this cellulose is not exactly the same mechanism as with RS, however, because cell turnover and proliferation were affected somewhat differently between cellulose and the RS forms used. Our results also showed reduced colonic cell proliferation (as measured by Ki-67 labelling index) in the rats fed any of the RS containing diets. Increased cell turnover may enhance the risk of mutations which can lead to an increased risk of developing colorectal cancer 42. Similar reductions in cell proliferation have also been observed in rats fed fermentative substrates like RS 121628 and the carbohydrate oligofructose 43. It is likely that the increased SCFA resulting from fermentation of starch in the colon contributed to the observed effects on colonic epithelial proliferation.

In conclusion, the present results show that consumption of high amylose maize starches which are rich sources of RS elicit a range of effects in the luminal environment and epithelial biology of the large bowel of rats. These effects differ significantly between different forms of RS and are not necessarily consistent either within a specific physico-chemical type (namely RS2) nor are they obviously predictable from the known characteristic of a RS. On the other hand, generation of high concentrations of distal colonic butyrate is associated with reduced cancer risk, particularly in experimental models and with facilitation of the epithelial apoptotic response to genomic damage. As a consequence, faecal butyrate concentrations might be a useful biomarker of risk for CRC and might provide more information than attempts to ascertain types and amount of RS in the diet (and fibre for that matter). This has implications for human epidemiological and interventional studies trying to further explore the usefulness of RS for protection against CRC. Predicting benefit from apparent dietary composition might not be useful. Furthermore, designing specific interventions with an RS might require a careful consideration of its impact on fermentation events including butyrate in the distal colon.

RL and GY declare they are recipients of a research grant from National Starch and Food Innovation, but otherwise all authors declare no other potential conflicts of interest.

RL, IB and GY conceived and designed the study. RL and YU carried out animal treatments, biochemical analysis, and data analysis. RL, YU and GY participated in interpretation of results and preparation of the article. All authors read and approved the final version of this manuscript.