Assessing prebiotic effects of resistant starch on modulating gut microbiota with an in vivo animal model and an in vitro semi-continuous fermentation model

Resistant starch (RS) is the starch that escapes human digestion system and will be delivered to the lower gut for fermentation by gut microbiota. Five types of resistant starch have been described, including 1) physically amylase-inaccessible type 1 RS, 2) native high amylose type 2 RS, 3) retrograded type 3 RS, 4) chemically modified type 4 RS, and 5) fatty acid complexed type 5 RS. RS has been shown to lower postprandial glycemic index (including plasma glucose and insulin levels), indicating effects of RS in reducing risk of type II diabetes and possible obesity as well, In addition, RS fermentation in the lower gut has been shown to maintain colon health and prevent colon carcinogenesis, mainly through production of SCFA. Physiological significances of SCFA, especially butyrate, include providing an energy source for colonocytes, maintaining the colonic epithelia cell layer, stimulating basal crypt cell proliferation, and inducing apoptosis of hyper-proliferative cancer cells. Furthermore, RS has also been considered as a prebiotic to modulate gut microbiota favoring RS degradation and SCFA production. The long-term objective of our study is to understand the underlying mechanisms of RS in improving human health from two correlated aspects, modulating gut microbiota and producing SCFA. We hypothesized that when using an in vivo animal model and an in vitro fermentation model, resistant starch consumption would shift gut microbiota pattern towards favoring RS fermentation, especially SCFA production,\nIn respect to the physiological significance of resistant starch implicated in improving colon health, we hypothesized that HA7 (high amylose starch VII, type 2 RS), and HA7-SA (palmitic acid complexed HA7, type 5 RS) would modulate colonic microbiota into different patterns. The bacterial pattern shifts can explain the differential efficacy of two RS suppressing colon carcinogenesis. Previous studies showed that pre-neoplasm, aberrant crypt foci (ACF) induced by carcinogen Azoxymethane (AOM) were only decreased by 8-wk feeding of HA7 (21 y 9) in male Fisher 344 rats, but not by HA7-SA (38 y 18), when compared to highly digested control starch (48 y 14). We used PCR-DGGE (Denaturing Gradient Gel Electrophoresis) to analyze colonic microbiota with bacteria universal 16S rRNA gene primers. Our results showed two RSs induced significant shifts of colonic microbiota compared to control digestible starch-fed rats. Moreover, differential bacterial patterns were observed between two RSs such as the specific enrichment of putative Bifidobacterium pseudolongum by HA7-SA, but not by HA7. More importantly, a significant correlation was observed between gut microbiota patterns and ACF numbers developed in AOM treated animals. Further analysis of starch fermentation-associated bacteria groups showed similar shifts of Clostridium cluster IV by two RSs with increased putative Ruminococcus bromii. By contrast, the bacterial pattern of Clostridium cluster XIVa showed correlation with ACF number. Specific enrichment of putative Ruminococcus obeum R. sp. SR1/5 was observed in HA7-fed animals, whereas specific decrease of Bacteroides sp. ASF519 or Parabacteroides goldsteinii (T) by HA7-SA was observed in the Bacteroides fragilis group. Our results suggested that gut microbiota patterns modulated by RS were related to differential efficacy of HA7 and HA7-SA in decreasing colonic carcinogenesis.\nWe established an in vitro semi-continuous anaerobic incubation model to compare fermentability, i.e. SCFA production, of four high amylose starches: HAV, HAVI, HAVII and GEMS-067 from four maize lines with different genetic background and amylose contents (55%, 65%, 70% and 70% respectively). This was done to evaluate prebiotic effects of resistant starch on modulating gut microbiota hypothetically towards favoring RS fermentation. The digested starch residues (SR), obtained from in vitro digestion with the AOAC 991.43 method, were incubated with fecal microbiota to simulate human digestion of cooked starchy food. In our study a total of 32 individuals were recruited: 17 lean, 9 overweight and 6 obese individuals. The study was conducted in two phases 5 months apart. Each phase consisted of a 3-wk incubation period with a frequency of changing BHI (Brain Heart Infusion) medium and SR substrates every 3.5 d. We observed significantly decreased pH, increased gas production, increased butyrate and total SCFA concentration in incubations with the four SR compared to blank BHI medium control starting at wk 1. There was no difference between SR. Molar proportions of butyrate was increased by SR with decreased acetate proportion, both of which achieved stability starting at wk 2. Additionally, propionate concentration was only increased by SRV at the end of the 3-wk incubation compared to BHI medium, but not by other SR. Large inter-individual variation was observed in the proportional increase of butyrate by SR compared to blank BHI medium control. No significant difference was found between lean and overweight individuals from fermentation indicators measured. We concluded that a stable long-term semi-continuous in vitro fermentation model was established to simulate carbohydrate fermentation in human lower gut. We also showed significant increase of butyrate production by RS fermentation with human fecal microbiota.\nThe modulation of microbiota by amylase digestion residues of HAV, VI and VII in our semi-continuous fermentation model was further analyzed with PCR-DGGE by examining total bacterial pattern and that of RS fermentation associated bacteria groups, including Clostridium cluster IV, XIVa and Bacteroides fragilis group. We hypothesized that bacterial patterns obtained after 3 wk fermentation will be selected by three SR incubation and lean/overweight microbiota as well. Our results showed that total bacterial patterns were shifted by three SR incubation with fecal microbiota from 30 donors at the end of 3-wk fermentation compared to control BHI medium. Moreover, bacterial pattern in SRV fermentation samples differed from that of SRVI and SRVII, which shared a certain degree of similarity. However, bacterial pattern of Clostridium cluster IV and XIVa in SRVI fermentation samples differed from SRV and SRVII, which shared similar patterns, whereas no shifts were observed by any SR in Bacteroides fragilis group bacterial pattern compared to blank BHI medium control. Most important, we observed the putative Ruminococcus bromii was specifically selected during SR incubation by microbiota from lean individuals, but not by microbiota from overweight and obese individuals. We concluded that our in vitro semi-continuous fermentation model can be used to assess prebiotic effects of RS by simulating long-term RS consumption in human. We also showed that Ruminococcus bromii, belonging to Clostridium cluster IV, was selectively enriched by SR and microbiota of lean individuals.\nIn summary, our studies with both in vivo animal model and in vitro fermentation model supported previous recognition of resistant starch acting as a prebiotic to modulate gut microbiota, especially on Clostridium clusters IV and XIVa. Ruminococcus bromii was specifically induced in rat model fed RS as well as in vitro fermentation model using lean microbiota. Moreover, different RS may have different fermentation outcomes. Our findings provided solid evidence to answer the fundamental question of how RS exerted effects on shifting bacterial pattern. In addition, we also showed that the physiological significance of RS might be affected by physical-chemical properties of starch and pre-existing microbiota as well.