Resistant starch (RS), the fraction of ingested starch that escapes from the small intestine (Asp, Reference Asp1992), is emerging as a protective agent against several serious pathologies of the large bowel (Topping & Clifton, Reference Topping and Clifton2001). There is also some evidence that RS may be more effective than dietary fibre in reducing risk of colorectal cancer (Cassidy et al. Reference Cassidy, Bingham and Cummings1994). Although the mechanisms by which RS could promote colonic health are not understood fully, the general consensus is that it acts indirectly through products of its fermentation by the gut microflora. Starch that enters the large bowel is normally actively metabolised by the numerous saccharolytic bacteria inhabiting that region of the gut (Pryde et al. Reference Pryde, Duncan, Hold, Stewart and Flint2003), resulting in greater SCFA production and a concomitant reduction in luminal pH (Cummings et al. Reference Cummings, Beatty, Kingman, Bingham and Englyst1996). Acidification of the intracolonic environment is considered important for suppressing the production and activity of a range of bacterial metabolites implicated in colonic disease (Topping & Bird, Reference Topping and Bird1999; Bird et al. Reference Bird, Brown and Topping2000a). SCFA also have important nutritional and physiological roles in maintaining the integrity of the bowel wall. Butyrate, one of the principal colonic SCFA, is attracting considerable interest because of its potential to lessen the risk of carcinogenesis (Brouns et al. Reference Brouns, Kettlitz and Arrigoni2002). In addition to modulating microbial metabolic activity, certain RS have a selective action on the bacterial composition of the large bowel microflora which also may have health implications for the host. Studies in pigs (Brown et al. Reference Brown, Warhurst, Arcot, Playne, Illman and Topping1997) and mice (Wang et al. Reference Wang, Brown, Khaled, Mahoney, Evans and Conway2002) have shown that high-amylose maize starches (HAMS), the RS investigated most extensively thus far, are very effective in raising colonic bifidobacteria numbers. These prebiotic actions are comparable to those of established prebiotic oligosaccharides, notably the inulin-type fructans (Roberfroid, Reference Roberfroid2001).
Small intestinal starch digestibility is altered by the relative proportions of amylose and amylopectin in the granule and also by partial hydration at elevated temperature for a period of time (Sievert & Pomeranz, Reference Sievert and Pomeranz1989; Kweon & Shin, Reference Kweon and Shin1997; Perera et al. Reference Perera, Hoover and Martib1997). HAMS is particularly susceptible to this hydrothermal (or heat-moisture) process with suggestions that changes are occurring in the associations between starch chains in the amorphous regions and the manipulations in the degree of crystalline order. The changes are intragranular and after hydrothermal treatment the intact starch granules remain. For HAMS it has been observed that hydrothermal treatment leads to increases in RS (Sievert & Pomeranz, Reference Sievert and Pomeranz1989) and dietary fibre (Crosby, Reference Crosby2003).
The aim of the current experiment was to compare the effects of untreated HAMS with the same starch after it had been subjected to hydrothermal treatment (designated HTHAMS) on large bowel fermentation events and microbiology in pigs. This HAMS is a commercially rich source of dietary fibre and RS, with an amylose content of 85 %, and it was expected that the hydrothermal treatment would lower its small intestinal digestibility. A conventional waxy maize starch (amylose content 0 %) was used as the reference starch because of its very high small-intestinal digestibility and low RS content (Bird et al. Reference Bird, Hayakawa, Marsono, Gooden, Record, Correll and Topping2000b; Brown et al. Reference Brown, Storlien, Brown and Higgins2003). Pigs were used for the experiment as they are considered to be a good model of gut function in man (Topping & Clifton, Reference Topping and Clifton2001).
Methods
Animals
Twenty-four, 4-week-old Large White crossbred pigs (males n 13, females n 11) were used, with an initial live weight of 23·7 (sem 0·6) kg when acquired from a local commercial piggery. They were housed individually in galvanised-steel pens with concrete floors and maintained at a controlled temperature (25 °C) for the duration of the study. Before the study commenced, the pigs were habituated to their new environment and feeding routine over 10 d during which time they were all fed the control diet (see Table 1). The study protocol was approved by the Animal Experimentation Ethics Committee of the CSIRO Health Sciences and Nutrition and conformed to published guidelines (National Health & Medical Research Council, CSIRO & Australian Agricultural Council, 1985).
Table 1 Composition (g/kg) of the experimental diets
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
* Mazaca 3401x cornstarch supplied by Starch Australasia Ltd, Lane Cove, NSW, Australia.
† Casein obtained from Cottee Nutritionals Pty Ltd, Gordon, NSW, Australia.
‡ Wheat bran (‘Clean Bran’) obtained from Goodman Fielder Mills (Allied Flour and Bakery Services), Mile End, SA.
§ Standard pig grower vitamin and mineral mix obtained from Roche Vitamins Australia Ltd, Wagga Wagga, NSW, Australia.
Experimental design, diets and feeding procedures
The pigs were weighed at the end of the 10-d adaptation period and assigned to one of four dietary treatment groups of six animals each on the basis of gender and live weight. It was necessary to allot unequal numbers of male and female pigs to one group selected at random (four and two respectively; HAMS+HTHAMS diet). Experimental diets were formulated to provide (by weight, as is basis): 150 g/kg protein, 540 g/kg starch, 200 g/kg lipid, and a total of 75 g/kg dietary fibre (as NSP). Diets were identical in composition except for the source of maize starch, which was either a conventional low-amylose cornstarch (Mazaca 3401C, Control) or one of two different high-amylose maize starches (HAMS or HTHAMS). The diet for the remaining treatment group contained both high-RS ingredients, which were present in equal proportions (HAMS+HTHAMS). All three starch products were commercial food ingredients manufactured by Starch Australasia (Lane Cove, NSW, Australia). HTHAMS was prepared using HAMS under conditions where the starch had a moisture content of 25 % and it was heated to 125 °C for 120 min. HAMS and HTHAMS contain 85 % and either 30 % or 60 % total dietary fibre, respectively (Brown, Reference Brown2004), whereas the starch used in the control diet had an amylose content of about 0 %. When formulating the diets allowance was made for the small amount of wheat starch (about 7 %) contributed to the whole diet by the fibre source (wheat bran, which contained 220 g starch/kg). Control and high-RS treatment diets were fed as two meals of identical weight, at 09.00 and 16.00 hours, at a rate proportional to the metabolic weight of the pigs (70 g/kg live weight0·75). Powdered rations were placed in feed hoppers and an equivalent amount of water added immediately. Pigs were weighed twice weekly prior to their morning feeding and food intake adjusted accordingly. Diets were manufactured on a weekly basis and stored in large bins at room temperature ( < 20 °C) until fed. Pigs had unrestricted access to tap water for the entire period of study. Chromic oxide, a particulate digesta marker, was incorporated (0·1 g/100 g) into each diet for the final 4 d of the treatment period to determine starch digestibility. Experimental diets were fed for 21 days and the pigs then anaesthetised and killed to enable collection of tissues and digesta.
Sampling procedures
Faeces (24-h collections) were removed from the pen floor periodically during the adaptation period and at the end of the first and second week of the treatment period and total daily faecal output was recorded. Freshly voided stools were used for bacterial enumeration and the remaining portions, after recording pH, were stored frozen to await determination of SCFA and other analytes. At the completion of the feeding period and approximately 16 h after the pigs had been fed the previous evening, they were weighed, sedated with ketamine (Ketapex; Apex Laboratories, St Mary's, NSW, Australia) and anaesthetised by inhalation of halothane (Rhone Merieux, West Footscray, VIC, Australia); mixed with O2 at 4 % v/v. The procedures used for collection of plasma and digesta have been described in detail previously (Bird et al. Reference Bird, Brown and Topping2000b). In brief, the abdominal cavity was opened by midline laparotomy and an aliquot of urine withdrawn from the bladder, transferred to a labelled vial and stored frozen ( − 20 °C) until analysis. The pigs were then killed by barbiturate overdose (Lethabarb, Virbac, Peakhurst, NSW, Australia) and the gastrointestinal tract immediately ligated at the oesophagus and rectum and the entire gastrointestinal tract together with the attached liver excised. The large and small bowel were isolated by means of appropriately positioned ligatures then separated and their respective unstretched lengths measured. Samples of ileal and mid colonic tissue were excised immediately and processed for histomorphometry. Small intestinal contents were removed from the distal 1 m, weighed and stored frozen ( − 20 °C) for subsequent analysis. The caecum and colon were separated, and the latter divided into three segments of equal length (proximal, mid and distal colon). Digesta from each of the segments and from the caecum were extruded and weighed. Using aseptic techniques, aliquots of digesta from the proximal colon were serially diluted, the suspensions plated on appropriate differential and selective media and subsequently enumerated for key enteric bacteria. pH was recorded for the remaining portion of the digesta samples which were prepared for later analysis and stored frozen ( − 20 °C). The colon and caecum, devoid of contents, were blotted dry and weighed. Frozen samples were thawed at 4 °C prior to processing for analysis. Four pigs were killed each day, one from each treatment. The experiment was staggered to ensure that the experimental period (21 d) was the same for all pigs. Representative samples of meals fed to each pig over the last 4 d of study were collected and stored to await analysis of starch and chromium.
Biochemical analyses
Digesta, faeces and food samples were thawed at 4 °C, homogenised, freeze-dried and then milled (1 mm screen) for measurement of dry matter, starch and chromium. Moisture content was calculated as the loss in mass of digesta after freeze-drying. Chromium content of digesta was determined by atomic absorption spectroscopy (Pearton & Mallett, Reference Pearton and Mallett1972). In brief, samples were accurately weighed into an acid-washed beaker and ashed in a muffle furnace at 250 °C for 1 h, followed by 4 h at 450 °C. Residues were cooled overnight before adding 2·4 ml of conc. HNO3 and heating to about 120 °C. Then 300 mg of potassium chlorate was added to each sample and heated until the volume had been reduced to approximately 1·6 ml. Residues were allowed to cool, diluted appropriately and analysed using an atomic absorption spectrophotometer (Varian Spectra Model AA 400, Varian, Inc.). A reducing nitrous oxide–acetylene flame in a matrix of 4 % HNO3 (v/v) was used to eliminate possible interferences with Cr detection.
Total starch was analysed as free glucose after α-amylase/amyloglucosidase digestion using a commercial procedure (Total Starch Assay Kit, Megazyme Ltd, Melbourne, Australia) that was based on the method of McCleary et al. (Reference McCleary, Solah and Gibson1994).
For SCFA analysis, aliquots of thawed large bowel contents and faeces (approximately 2 g) were mixed with 3 volumes of internal standard (3·52 mmol/l oenanthic acid) and the pH of the suspension determined using an appropriate glass electrode. The suspension was then centrifuged for 15 min at 3000g and 4 °C and an aliquot of the supernatant distilled at low temperature (Vreman et al. Reference Vreman, Dowling, Raubach and Weiner1978). Approximately 10 μl distillate was assayed for individual SCFA by gas chromatography using a Hewlett-Packard gas chromatograph (Model HP5710) equipped with a Zebron ZB FFAP glass capillary column (0·53 mm i.d. × 30 m; Phenomenex, Pennant Hills, NSW, Australia). Individual SCFA were quantified using flame-ionisation.
Histomorphometry
The lumen of intestinal segments were flushed free of residual digesta using phosphate-buffered saline (4 °C, pH 7·4) and cut longitudinally along the wall adjacent to where the mesenteric border was attached. Ileal and colonic segments (about 1 cm) were excised immediately and fixed in neutral buffered formalin for 24 h and then placed in 70 % ethanol (v/v) to await histological assessment. Tissue slices (approximately 1–2 mm) were cut and encased in tissue cassettes before being prepared using an automated process that involved dehydrating in ethanol, clearing with histolene, and then permeating and embedding with paraffin. Sections were cut and stained with haematoxylin and counterstained with eosin yellow. Dimensions of intestinal anatomical components were measured using light microscopy combined with computer-aided image analysis (Video Pro, Leading Edge, Bedford Park, South Australia).
Bacteriology
Aliquots (10 g) of freshly collected digesta from the proximal colon and faeces were serially diluted (10-fold dilutions), and 0·1 ml of each of the suspensions plated onto the following semiselective agar plates: Columbia blood medium, chromogenic E. coli/coliform medium and Rogosa medium (all from Oxoid, West Heidelberg, Victoria), and bifidus blood medium (Reuter, Reference Reuter1985), as modified by Pachenari et al. (Reference Pachenari, Conway and Playne2002) to enumerate total anaerobes, total aerobes, E. coli and coliforms, lactobacilli, and bifidobacteria, respectively. Plates were then incubated aerobically, or anaerobically using the Anaerocult A mini procedure (Merck Pty Ltd, Kilsyth, VIC, Australia), at 37°C for 1 to 6 d, depending on the assay. Relevant bacterial colonies on culture dilution plates were counted visually and the number of organisms calculated as colony forming units (CFU) per gram of wet digesta or faeces.
Calculations and statistics
Digesta SCFA pools were calculated as the product of concentration (mmol/l) of individual acids, or their sum (total SCFA), × volume of digesta water (ml) in respective intestinal compartments. Faecal SCFA excretion was calculated similarly. Intestinal starch digestibility was calculated according to a standard formula (marker ratio method) using values for Cr content of feed, ileal and distal colonic digesta, and faeces. Statistical evaluations were performed using the General Linear Model (GLM) procedure of SAS version 8.02 (SAS Institute, Cary, NC, USA). Gender and block (day 0 weight class) were not significant effects and were subsequently removed from the model. Dietary treatment effects were analysed by ANOVA. If a significant F test was observed (P < 0·05), differences among means were then determined by the protected difference (PDIFF) procedure. Repeated measures ANOVA was used to analyse differences in SCFA and other fermentation-dependent measures among sampling sites in the large bowel. Differences between treatment means were considered significant at P < 0·05. For digesta and faecal counts of bacteria, the data were transformed (log10) before statistical analysis of treatment effects. For samples that contained bacterial populations below the assay level of detection, an appropriate arbitrary value (1 log unit lower than the specified detection level for the respective assay) was assigned and used in subsequent statistical calculations. Data are shown as least squares mean and pooled standard error of the least squares mean (sem). For bacterial counts, results are expressed as log10 CFU/g. For most variables, means are for six observations per group; however, on some occasions samples (e.g. digesta) were either unavailable or present in insufficient quantities to permit analysis. For these variables, the number of observations is shown in parentheses. For tabulated results, values within a column with different superscripts differ significantly from the control.
Results
Food intake and body weight gain
All diets were well accepted and meals were consumed promptly. The pigs grew with an average gain of 391 (pooled sem 13, n 24) g/d for all groups across the treatment period. Final live weight (34·8 (pooled sem 0·7) kg) and feed conversion efficiency (436 (pooled sem 12) g/kg) were unaffected by diet but live weight gain for the treatment period was lower (P < 0·05) for pigs fed HAMS compared to the control group (7·6 v. 9·3 (pooled sem 0·5) kg).
Small and large intestinal starch digestibilities
There was almost complete disappearance of starch in the gastrointestinal tract of pigs fed the control starch as evidenced by the small amount of starch in faeces ( < 2 g/100 g) over the whole experiment. Replacement of the control maize starch with HAMS in the treatment groups resulted in a significant increase in faecal starch concentration and excretion. For instance, in week 2, faecal starch output was 1, 30, 70 and 50 g/d for control, HAMS, HTHAMS and the HAMS+HTHAMS groups, respectively (pooled sem 10 g/d; values all significantly different from one another).
Only very small amounts of starch were detected in ileal digesta of control pigs (Table 2). By contrast, feeding HAMS diets produced significantly higher starch levels in ileal digesta, particularly for pigs fed HTHAMS either alone or in combination with HAMS (P < 0·001 and P < 0·05, respectively). Ileal starch digestibility to the terminal ileum was >96 % in pigs fed the control diet but more than 12 % and 30 % of starch in the HAMS and HTHAMS diets, respectively, escaped into the large bowel (Table 3). Although the difference in ileal starch digestibility between control and groups fed high-amylose starches was large, it was not statistically significant for HAMS (P>0·05). These data are based on a small sample size because the ileum of some pigs at slaughter was devoid of contents (see Table 3).
Table 2 Starch content (g/100 g dry matter) of diets and of ileal and distal colonic digesta of pigs fed the experimental diets*(Values are least squares means and pooled sem of six animals per group except ileum control and HTHAMS (n 4) and ileum HAMS and HAMS+HTHAMS (n 2))
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
*Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01, adP < 0·001.
For distal colonic digesta starch content, control v. HAMS, P = 0·091 and HAMS+HTHAMS v. HAMS, P = 0·027.
Table 3 Ileal and distal colonic starch digestibility (%) of pigs fed the experimental diets*(Values are least squares means and pooled sem of six animals per group except ileum control and HTHAMS (n 4) and ileum HAMS and HAMS+HTHAMS (n 2))
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01.
For ileal starch digestibility, HTHAMS v. HAMS, P = 0·048.
For distal colonic starch digestibility, HTHAMS v. HAMS, P = 0·027 and HAMS+HTHAMS v. HAMS, P = 0·055.
*Digestibility coefficients were calculated as:
Very small amounts of starch were recovered from the distal colon of control pigs in contrast to those fed the RS diets (Table 2). For pigs fed HTHAMS and HAMS+HTHAMS, starch accounted for about 30 % of dry matter in the distal colon, whereas for pigs fed HAMS, only about half this amount of starch was present (P>0·05). Nearly all of the starch that entered the colon of control pigs was fermented to completion (99·6 %) whereas there was proportionately less microbial breakdown of starch in the colon of pigs fed HTHAMS or HAMS+HTHAMS (both P < 0·01).
Large bowel morphology
Consumption of the HAMS diets had a pronounced trophic effect on the large bowel. Although caecal weight of pigs fed the various high-RS diets was greater than that of control pigs, only in the case of HAMS were the differences statistically significant (P < 0·05; Table 4). The colon of pigs fed the HAMS diets was also substantially heavier as well as longer relative to controls, regardless of whether the data are expressed in absolute terms (Table 4) or normalised to live weight (data not shown). The increase in tissue mass over that of controls was similar in the proximal, mid and distal colon, irrespective of the source of the RS diet (Table 4). Similarly, colon length was increased (by as much as 35 %) by all three high-RS diets compared to control (Table 4). Pigs fed diets containing HTHAMS had a longer colon (P < 0·01) and tended to have a heavier distal colon (P = 0·10) than those fed HAMS. There was a highly significant relationship between estimated RS intake (feed intake × ileal starch digestibility) and colon length (r 2 0·69, P < 0·001).
Table 4 Weight (g) and length (m) of the large bowel of pigs fed experimental diets *(values are least square means and pooled sem of six animals per group)
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01, adP < 0·001.
For distal colon weight, control v. HAMS, P = 0·052 and HTHAMS v. HAMS, P = 0·096.
For colon length, HTHAMS v. HAMS, P = 0·007.
* Empty wet weight of various compartments and unstretched length of colon.
Despite the distinct dietary effects on gross colonic morphology, the microscopic structure of the mid-colonic wall was not markedly affected by feeding diets high in RS (data not shown). Generally, the only linear dimension for mid-colon tissue for which treatment effects were statistically significant was total thickness of the colonic wall, which was greater for HTHAMS than HAMS or HAMS+HTHAMS (2326 v. 1721 and 1887 μm, P < 0·001 and P < 0·05, respectively; pooled sem116). These values were not significantly different from the controls (1999 μm).
Faecal and large bowel digesta mass and moisture content
Faecal output was significantly greater for pigs fed diets high in RS starch compared to controls (average of 300 g/d in week 2 for the three high-RS diets compared to 103 g/d for controls; pooled sem23, P < 0·01). Similar differences (approximately twofold) were observed for the amounts of caecal and colonic digesta in pigs fed the control and high-RS diets (Table 5). Values for digesta mass were greatest for HTHAMS, intermediate for HAMS+HTHAMS, and least for HAMS at each large bowel sampling site (same ranking for faecal mass). However, differences between the individual high-RS groups were not uniformly significant across sampling sites. Water content of digesta declined progressively from the ileum to the distal colon (Table 6). Moisture content of ileal digesta was similar among treatment groups (P>0·05) but caecal and proximal colonic digesta of control pigs was drier than that of pigs fed RS diets (Table 6). In addition, digesta in the proximal, mid and distal colon of pigs fed HAMS contained more water per unit mass than that of pigs fed either HTHAMS or HAMS+HTHAMS, and water content of distal colonic digesta was greater for HAMS than controls (P < 0·05). Moisture content of faeces collected in week 1 from pigs fed HAMS was greater (P < 0·01) than that of controls (67·7/100 g v. 61·4 g/100 g); however, by week 2, faecal water content was reasonably uniform (data not shown) and averaged 65 % (range 58–74 %) for all groups.
Table 5 Wet weight (g) of digesta in the intestinal tract of pigs fed experimental diets(Values are least squares means and pooled sem of six animals per group except ileum HAMS and HAMS+HTHAMS (n 2) and ileum HTAMS (n 4))
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01, adP < 0·001.
For proximal colon, HTHAMS v. HAMS, P = 0·001 and HTHAMS v. HAMS+HTHAMS, P = 0·01.
Table 6 Water content (g/100 g) of digesta in the intestinal tract of pigs fed experimental diets(Values are least squares means and pooled sem of six animals per group except ileum HAMS and HAMS+HTHAMS (n 2), ileum control (n 5) and ileum HTAMS (n 4))
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01, adP < 0·001.
For proximal colon, HAMS v. HTHAMS, P = 0·03 and HAMS v. HAMS+HTHAMS, P = 0·02.
For mid colon, HAMS v. HTHAMS, P = 0·02 and HAMS v. HAMS+HTHAMS, P = 0·004.
For distal colon, HAMS v. HTHAMS, P = 0·009 and HAMS v. HAMS+HTHAMS, P = 0·005.
Large bowel and faecal pH, and SCFA
Faecal pH was similar (6·39 (pooled sem0·05); P>0·05) among the four dietary groups during the adaptive period when all pigs were fed the control diet. pH remained essentially constant in the control group during the subsequent two weeks of study (6·30 and 6·40, respectively, for weeks 1 and 2 of the treatment phase). Feeding HAMS elicited an abrupt and substantial fall of more than 1 pH unit in faecal pH at week 1 (data not shown), and this difference was still evident at week 2 (Table 7). Compared to the control diet, those high in RS resulted in substantially lower digesta pH values throughout the entire large bowel, including the distal colon (Table 7).
Table 7 pH of digesta in the large bowel and faeces of pigs fed experimental diets(Values are least squares means and pooled sem of six animals per group)
HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Effects were significant for diet (P < 0·001) but not site or their interaction (P>0·05).
Mean values for treatment were significantly different from control: adP < 0·001.
For proximal colon, HAMS v. HTHAMS, P = 0·03.
* Faecal samples were collected on day 14 whereas digesta was sampled on day 21.
Large bowel concentrations of total and major individual SCFA are presented in Fig. 1. Generally, total SCFA concentrations were higher throughout the large bowel, especially in the colon, for pigs fed RS diets compared to controls (Fig. 1(a)). There were few differences in digesta SCFA concentrations between groups fed different types of high-amylose starch. A relative increase in digesta concentrations of acetate and propionate largely accounted for the higher concentrations of total SCFA induced by feeding diets high in RS (Figs. 1(b) and 1(c), respectively). Diets high in RS usually raised digesta butyrate concentration throughout the colon relative to the control diet, but the difference did not always attain statistical significance (Fig. 1(d)). There was no effect of diet on caecal levels of butyrate; however, in the proximal and mid colon, the HAMS treatment elicited a 60 % and 120 % increase (P < 0·05 and P < 0·01), respectively, in digesta butyrate concentrations relative to controls. In the distal colon, only HTHAMS and HAMS+HTHAMS treatments raised butyrate concentration (by 83 % and 69 %, respectively; both P < 0·05).
Fig. 1 Concentrations of (a) total and individual SCFA (b–d, acetate, propionate and butyrate, respectively) in caecal, and proximal, mid and distal colonic digesta of pigs fed diets containing low- (Control) or various high-amylose resistant starches (–○–, Control; –■–, high-amylose maize starch (HAMS); –▲–, hydrothermally treated HAMS (HTHAMS); –△–, HAMS + HTHAMS). Values are means for six pigs; pooled sem values for total, acetate, propionate and butyrate concentrations were 9·1, 5·3, 4·9 and 1·2 mmol/l, respectively. Effect of diet and sampling site were significant for total and each individual SCFA (all P < 0·001). Mean values with different letters within a compartment are significantly different from control (abP < 0·05, acP < 0·01, adP < 0·001).
Large bowel pools of total and individual SCFA were substantially larger in pigs fed diets high in RS (Fig. 2) because of a greater large bowel digesta mass (see Table 5) and higher SCFA concentration (see Fig. 1). Each of the RS diets increased the size of the butyrate pool throughout the entire large bowel, including the distal colon (Fig. 2(d)). The SCFA profile for faeces was similar to that for large bowel digesta although, unlike digesta, faecal butyrate concentration and molar ratio were not affected by diet (P>0·05; data not shown). Daily faecal excretion of total and major individual SCFA was substantially greater in pigs fed RS diets compared to controls, largely due to a greater faecal mass in the former treatment groups (data not shown). Of the various high-RS dietary treatments, the combination of HAMS and HTHAMS was the most effective at raising faecal SCFA excretion (data not shown).
Fig. 2 Pools of total (a) and individual SCFA (b–d, acetate, propionate and butyrate, respectively) in the caecum, and proximal, mid and distal colon of pigs fed diets containing low- (Control) or various high-amylose resistant starches (–○–, Control; –■–, high-amylose maize starch (HAMS); –▲–, hydrothermally treated HAMS (HTHAMS); –△–, HAMS + HTHAMS). Values are means for six pigs; pooled sem values for total, acetate, propionate and butyrate pools were 3·4, 1·8, 1·6 and 0·4 mmol, respectively. Effect of diet and site were significant for total and each individual SCFA (all P < 0·001). Diet × site interaction for total SCFA pool (P < 0·01). Mean values with different letters within a compartment are significantly different from control (abP < 0·05, acP < 0·01, adP < 0·001).
Colonic and faecal microflora
Total anaerobic bacterial density (CFU/g) was greater in proximal colonic digesta in pigs fed the high-RS diets compared to controls (Table 8). The numbers of pigs with colonic bifidobacterial counts below the level of detection of the assay (104 CFU/g) was greater for controls than groups on diets high in RS (data not shown). There were no effects of diet on numbers of total aerobic bacteria, although the total anaerobe:aerobe ratio was greater (P < 0·05) for pigs fed HAMS compared to controls (data not shown).
Table 8 Viable bacterial counts (log10 CFU/g) in the proximal colonic digesta and faeces(Values are least squares means and pooled sem of six animals per group)
CFU, colony forming units; HAMS, high-amylose maize starch; HTHAMS, hydrothermally treated HAMS.
Mean values for treatment were significantly different from control: abP < 0·05, acP < 0·01, adP < 0·001.
* Samples collected on week 2 of the treatment period.
In general, there was a progressive decline in faecal counts of viable total aerobes, coliforms and E. coli, and an incremental increase in anaerobe:aerobe ratio, as the experiment progressed (data not shown). The bacterial population profile of faeces and proximal colonic digesta was similar for each of the treatment groups (Table 8). Density of faecal anaerobes was unaffected by diet, but feeding HAMS or HTHAMS increased (P < 0·05) faecal numbers of lactobacilli and bifidobacteria compared to the control diet (Table 8). Faecal counts of coliforms and E. coli were lower after 2 weeks of feeding HTHAMS, either alone or in combination with HAMS, compared to controls. However, HAMS as the sole RS in the diet did not affect faecal numbers of these bacterial groups.
Discussion
All diets used in the present study contained equal amounts of starch. However, as expected, the amylomaize starches HTHAMS and HAMS were less digestible than the control starch and were therefore much higher in RS. About 12 % and 30 % of starch present in HAMS and HTHAMS, respectively, resisted small intestinal digestion while the control diet contained < 4 % RS. The digestibility values for HAMS and also HAMS+HTHAMS are of lesser reliability because of the limited number of observations. It was not possible (for logistic reasons) to collect intestinal contents any earlier than 16 h after the last feeding. This may have contributed to the absence of adequate quantities of terminal ileal digesta in some pigs so the ileal digestibility data should be interpreted with caution. Nevertheless, interpolation of data obtained for HAMS and the mixed RS treatment (HAMS+HTHAMS) indicates that the RS content of HAMS may be in the order of 20 %, which is near to the value (30 %) reported for the same powdered product incorporated into various cooked foods and eaten by volunteer ileostomates (Muir et al. Reference Muir, Birkett, Brown, Jones and O'Dea1994). Govers et al. (Reference Govers, Gannon, Dunshea, Gibson and Muir1999) report that for pigs fed a diet containing coarsely ground high-amylose maize kernels, about 60 % of RS was detected in ileal digesta.
It seems that most types of RS tested to date are fermented more or less to completion by the porcine large bowel microflora (Bach Knudsen & Hansen, Reference Bach Knudsen and Hansen1991; Bird et al. Reference Bird, Hayakawa, Marsono, Gooden, Record, Correll and Topping2000b; Martinez-Puig et al. Reference Martinez-Puig, Perez, Castillo, Andaluz, Anguita, Morales and Gasa2003). Both HAMS and HTHAMS were fermented extensively in this experiment but about 7 % and 17 %, respectively, appeared in the faecal stream. Faecal starch excretion was greater in this experiment than in a previous study in which pigs were fed diets containing a similar level of HAMS (Brown et al. Reference Brown, Warhurst, Arcot, Playne, Illman and Topping1997). Pluske et al. (Reference Pluske, Hampson and Williams1997) have shown that small intestinal starch digestibility of growing pigs may increase with advancing age and the fact that our pigs were younger than those used by Brown et al. (Reference Brown, Warhurst, Arcot, Playne, Illman and Topping1997) may explain this difference. High-RS diets have also been shown to increase faecal starch excretion in man (Phillips et al. Reference Phillips, Muir, Birkett, Lu, Jones, O'Dea and Young1995), supporting the suggestion that RS fermentation can be less than complete. Although HTHAMS was less digestible in the small bowel than HAMS, these starches exhibited similar large bowel fermentability, with about 45 % of the starch that passed the terminal ileum subsequently disappearing in the large bowel. Accordingly, about 30 g more starch was degraded by the large bowel microflora of pigs fed HTHAMS than HAMS (96 g/d v. 67 g/d, respectively, at the end of experiment). Presumably the much longer retention time of digesta in the large compared to the small intestine (more than about tenfold) explains why whole tract digestibility was similar for the two amylomaize starches but small intestinal digestibility was not. As mentioned previously, comparisons should be interpreted cautiously given the small number of estimates of ileal digestibility for HAMS (and the combined maize starch treatment).
Large bowel digesta mass and faecal output were increased by feeding amylomaize starches. The contents of all large bowel compartments were numerically greater for HTHAMS than HAMS but were not significant statistically. The actual amount of RS was insufficient to account fully for the large bowel digesta mass in pigs fed RS. Presumably, the increase in digesta weight must come mostly from biomass (Shetty & Kurpad, Reference Shetty and Kurpad1986; Cummings et al. Reference Cummings, Beatty, Kingman, Bingham and Englyst1996), which is consistent with the greater number of culturable bacteria in the colon of pigs on the amylomaize diets. The amylomaize diets also may have increased digesta mass by sparing some NSP from fermentation (Phillips et al. Reference Phillips, Muir, Birkett, Lu, Jones, O'Dea and Young1995; Cummings et al. Reference Cummings, Beatty, Kingman, Bingham and Englyst1996), possibly by down-regulating microbial β-glucosidase activity (Hylla et al. Reference Hylla, Gostner, Dusel, Anger, Bartram, Christl, Kasper and Scheppach1998). However, assuming that this mechanism operates in pigs, its contribution is likely to be small (Govers et al. Reference Govers, Gannon, Dunshea, Gibson and Muir1999). Previous studies have shown increases in digesta and faecal weight in pigs with feeding of amylomaize (e.g. Brown et al. Reference Brown, Warhurst, Arcot, Playne, Illman and Topping1997) and other (e.g. Bird et al. Reference Bird, Hayakawa, Marsono, Gooden, Record, Correll and Topping2000b) types of RS. RS consumption has been shown to increase faecal weight in most human studies (Phillips et al. Reference Phillips, Muir, Birkett, Lu, Jones, O'Dea and Young1995; Cummings et al. Reference Cummings, Beatty, Kingman, Bingham and Englyst1996; Noakes et al. Reference Noakes, Clifton, Nestel, Le Leu and McIntosh1996), but the effect is relatively modest. In the present study, the increase in digesta and faecal mass depended on the type of starch, with HTHAMS more effective than HAMS but only in the proximal colon (see Table 5).
Fermentation was increased markedly by inclusion of amylomaize starch in the diet as evidenced by the greater concentrations and pools of SCFA throughout the large bowel and faeces. Similar data have been reported previously for pigs (Govers et al. Reference Govers, Gannon, Dunshea, Gibson and Muir1999) and other species, including rats (Coleman et al. Reference Coleman, Landstrom, Royle, Bird and McIntosh2002). Other forms of RS (e.g. RS in cooked and cooled brown rice; Bird et al. Reference Bird, Hayakawa, Marsono, Gooden, Record, Correll and Topping2000b) have given similar outcomes. Amylomaize also stimulates colonic fermentation in human subjects as evidenced by increased faecal SCFA levels (Noakes et al. Reference Noakes, Clifton, Nestel, Le Leu and McIntosh1996). As expected, SCFA concentrations were greatest in the proximal colon and declined towards the distal region. This spatial distribution is similar to that seen in human subjects (Mitchell et al. Reference Mitchell, Lawson, Davies, Kerr-Grant, Roediger, Illman and Topping1985) and is thought to reflect the progressive depletion of fermentative substrate and the rapid absorption of SCFA by the epithelium on passage of the faecal stream (Cummings, Reference Cummings1981).
The amylomaize starch diets altered the pattern of fermentation with a higher molar contribution by propionate that occurred largely at the expense of acetate. These changes have been noted previously with HAMS in the large bowel and faeces of pigs (Topping et al. Reference Topping, Gooden, Brown, Biebrick, McGrath, Trimble, Choct and Illman1997) and with other sources of RS, such as legumes (Fleming et al. Reference Fleming, Fitch and Chansler1989) and cooked brown rice (Bird et al. Reference Bird, Hayakawa, Marsono, Gooden, Record, Correll and Topping2000b). Butyrate levels were also higher in pigs fed RS but the effects were less in the distal colon. There is a body of in vitro and in vivo experimentation to support the view that starch fermentation favours butyrate production compared with other carbohydrate substrates (Englyst et al. Reference Englyst, Trowell, Southgate and Cummings1987; Topping & Clifton, Reference Topping and Clifton2001). Butyrate is a major metabolic fuel for normal colonocytes and is a key regulator of their proliferation, differentiation and apoptosis (Sengupta et al. Reference Sengupta, Muir and Gibson2006). The lower level in the distal bowel may reflect preferential butyrate metabolism by the epithelium, which could be of benefit in lowering the risk of serious non-communicable diseases such as colorectal neoplasia, especially in the distal colon, where the incidence of intestinal neoplasms is highest (Rabeneck et al. Reference Rabeneck, Davila and El-Serag2003). Although HTHAMS supplied more substrate to the colonic microflora than HAMS, there were no statistically significant differences between these amylomaizes, either alone or in combination, in increasing butyrate pools (see Fig. 1).
In vivo (Topping et al. Reference Topping, Gooden, Brown, Biebrick, McGrath, Trimble, Choct and Illman1997; Govers et al. Reference Govers, Gannon, Dunshea, Gibson and Muir1999) and in vitro experiments (Christl et al. Reference Christl, Katzenmaier, Hylla, Kasper and Scheppach1997) indicate that microbial HAMS fermentation is reasonably rapid. However, in the present study, both HTHAMS and HAMS elicited luminal changes that spanned the complete length of the large bowel and modest quantities of unfermented starch were excreted in faeces. These data indicate that fermentation of these particular starches may not be particularly fast. The RS source, as well as various other factors, including the dietary amount, age of the animal and dietary level, may be important. Other dietary constituents, notably insoluble NSP, can also influence the rate and site of starch fermentation in situ (Govers et al. Reference Govers, Gannon, Dunshea, Gibson and Muir1999). RS types that are not readily fermented might afford a distinct advantage over more susceptible substrates, such as oligoscacccharides and soluble NSP, in maintaining the health of the distal bowel. Combinations of RS with various NSP sources have proved effective in moving the site of maximal carbohydrate breakdown further along the gastrointestinal tract of rats (Le Leu et al. Reference Le Leu, Hu and Young2002), pigs (Govers et al. Reference Govers, Gannon, Dunshea, Gibson and Muir1999) and man (Muir et al. Reference Muir, Yeow, Keogh, Pizzey, Bird, Sharpe, O'Dea and Macrae2004).
Consumption of RS had a significant trophic effect on the large bowel, manifest partly as a lengthening of the colon. This action was greater for HTHAMS than HAMS but only HAMS provoked an increase in the weight of the empty caecum. Others have reported increases in gut weight in pigs (Martinez-Puig et al. Reference Martinez-Puig, Perez, Castillo, Andaluz, Anguita, Morales and Gasa2003) and rats ingesting RS (Lopez et al. Reference Lopez, Coudray, Bellanger, Levrat-Verny, Demigne, Rayssiguier and Remesy2000; Perrin et al. Reference Perrin, Pierre, Patry, Champ, Berreur, Pradal, Bornet, Meflah and Menanteau2001) or other fermentable polysaccharides (Perrin et al. Reference Perrin, Pierre, Patry, Champ, Berreur, Pradal, Bornet, Meflah and Menanteau2001; Verghese et al. Reference Verghese, Rao, Chawan and Shackelford2002). Raw potato starch (Martinez-Puig et al. Reference Martinez-Puig, Perez, Castillo, Andaluz, Anguita, Morales and Gasa2003) and a specific amylomaize starch (Topping et al. Reference Topping, Gooden, Brown, Biebrick, McGrath, Trimble, Choct and Illman1997) have been shown to increase colon length of pigs. In the latter study, colonic elongation was proportional to the amylose content of the diet. The present data confirm and extend those findings and show that the elongation is related positively to the amount of starch reaching the terminal ileum. Presumably, the various SCFA, which are the principal products of RS (and fibre) fermentation in the large bowel, are the trophic stimuli (Topping & Clifton, Reference Topping and Clifton2001). However, total wet digesta mass was correlated more strongly to the length of the colon (r 2 0·827, P < 0·001) than to the quantity of SCFA present in the lumen (r 2 0·494, P < 0·001), which suggests that the mechanical actions of RS may have a greater influence on large bowel size and structure than fermentation-mediated events. However, micromorphometric analysis (of the mid colon) did not demonstrate a thickening of the musculature in response to RS ingestion.
The present investigation provides further evidence that amylomaize starches can act as prebiotics in that they selectively promote the proliferation and/or activity of one or a limited range of colonic bacteria considered beneficial to the health of the host (Gibson & Roberfroid, Reference Gibson and Roberfroid1995). HAMS and HTHAMS were equally effective and both increased bifidobacteria counts in the proximal colon by more than 3 log units compared to the control group. The prebiotic effect was not as remarkable when measured in faeces, largely because of the higher bifidobacterial concentrations of the control pigs (see Table 8). A variety of factors have been shown to influence the magnitude of the prebiotic effect, including the initial size of the microbial population targeted by the intervention (Tuohy et al. Reference Tuohy, Finlay, Wynne and Gibson2001). In the present study, the faecal bifidobacteria population of the control group is comparable to previous reports for pigs fed various diets (Moore et al. Reference Moore, Moore, Cato, Wilkins and Kornegay1987). Amylomaize has also been shown to be a bifidogenic prebiotic in mice (Wang et al. Reference Wang, Brown, Khaled, Mahoney, Evans and Conway2002). In vitro studies have demonstrated that bifidobacteria is a major starch utiliser (Macfarlane & Englyst, Reference Macfarlane and Englyst1986), which is consistent with the in vivo effects just mentioned. For most variables, effects were greatest for HTHAMS, intermediate when the two amylomaizes were given together, and least for HAMS. However, in terms of bifidogenicity, the amylomaize combination was not as effective as either of the individual RS sources given alone in raising bifidobacteria counts. The reason for the diminished response is not apparent and further investigation is required to understand the interaction between different starches on the large bowel microbiota.
HAMS and HTHAMS also increased the size of the intestinal lactobacilli population, which agrees with previous reports for HAMS and other types of RS (Kleessen et al. Reference Kleessen, Stoof, Proll, Schmiedl, Noack and Blaut1997; Wang et al. Reference Wang, Brown, Khaled, Mahoney, Evans and Conway2002). However, the result was not as impressive as that seen for bifidobacteria. Feeding HTHAMS either alone or in combination with HAMS was shown to suppress faecal E. coli counts relative to controls, whereas HAMS, which had similar physico-chemical effects in the large bowel, was without effect. We have found that a high-RS diet based on cooked and cooled white rice (combination of RS1 and RS3) dramatically reduced the E. coli population in the proximal colon of young pigs (A Bird, M Jackson, R Rankin and D Topping, unpublished results).
The current investigation has shown that both HAMS and HTHAMS are rich sources of RS and their consumption profoundly alters the morphology of the porcine large bowel and the physico-chemical and microbial properties of its contents. Of the amylomaize starches, HTHAMS delivered substantially more fermentable carbohydrate to the large bowel than HAMS but this did not necessarily equate to a more active fermentation. Both starches elicited changes in the luminal environment that encompassed the entire large bowel, but there were indications that HTHAMS acted more distally than HAMS. If these results in pigs can be extrapolated to man, foods containing HTHAMS and/or HAMS may be useful for reducing risk of infectious enteric disease and protecting against degenerative large bowel disorders, particularly in the distal colon.
Acknowledgements
We thank Rodney Trimble and Sylvia Usher for sample processing and carbohydrate analysis of diet ingredients and animal samples, Ben Scherer for chromium determinations, Peter Royle for protein analysis of feedstuffs, Debbie Davies for assistance with care of pigs, and Drs Reza Zarrinkalam and David Tivey for histology and micromorphometric analysis of intestinal tissue. Financial support for this research was received from Starch Australasia Ltd and CSIRO.
