Gastrointestinal tract (GIT) microbial fermentation is an important process in humans and monogastric animals. During fermentation, dietary fibre and non-dietary material, such as mucin, are degraded by the GIT microbiota. The end products of fermentation are mainly SCFA, organic acids, such as lactate, and gasses like CO2, CH4 and H2(Reference Cummings and Englyst1). The SCFA have beneficial health effects for the host, both locally within the GIT (e.g. as an energy source for epithelial cells)(Reference Roediger2) and systemically (e.g. regulation of glucose homoeostasis)(Reference Morrison and Preston3). A prevailing paradigm is that fermentation in humans and monogastric animals occurs predominantly in the hindgut with little fermentation in the foregut(Reference Cummings and Englyst1).
There is a considerable number of microbes present in the foregut of the human and growing pig(Reference Wang, Ahrne and Jeppsson4–Reference Montoya, Henare and Zhu7). These microbes may ferment dietary fibre, and several studies have reported significant disappearance (i.e. fermentability) of dietary fibre at the end of the small intestine in both human ileostomates and pigs(Reference Englyst and Cummings8–Reference Bach Knudsen and Canibe11). For example, 13 % of the NSP in potato(Reference Englyst and Cummings8) and 15–46 % of the dietary pectin(Reference Holloway, Tasman-Jones and Maher9) disappeared in the foregut of human ileostomates. Results of several experiments also indicate low ileal digestibility values for dietary fibre in human ileostomates and ileal-cannulated pigs(Reference Bach Knudsen and Canibe11–Reference Englyst and Cummings13). Montoya et al. (Reference Montoya, Henare and Rutherfurd14) discussed that non-dietary gut materials may interfere with dietary fibre determination in ileal digesta and thus lead to underestimation of dietary fibre ileal digestibility. For instance, an estimate of the ileal digestibility of soluble fibre in kiwifruit increased 50 % units when it was corrected for interfering non-dietary materials(Reference Montoya, Rutherfurd and Moughan15). Consequently, dietary fibre fermentation in the foregut may be greater than is commonly believed.
The greater number and more diverse population of microbes in the ileum(Reference Wang, Ahrne and Jeppsson4) coupled with a longer transit time of digesta in the lower small intestine indicate that fermentation in the foregut may occur mainly within the ileum. Moreover, and based on functional genome analysis, the human ileal microbiota appears to be able to take up and metabolise simple carbohydrates (i.e. low-molecular-weight carbohydrates such as oligosaccharides) rapidly(Reference Zoetendal, Raes and van den Bogert5). This is important as transit time in the ileum is considerably shorter than in the hindgut(Reference Graff, Brinch and Madsen16). In the growing pig, the adenylate energy charge in the last third of the small intestine was similar to that of the caecum, despite the lower number of anaerobic bacteria(Reference Jensen and Jørgensen17). These observations indicate the potential for a quantitatively significant amount of fermentation occurring in the small intestine of humans and pigs. However, there are no reported studies that have been able to quantify the ileal fermentation due to the lack of a methodology to do so. It is hypothesised that ileal fermentation is as important as hindgut fermentation in terms of organic matter (OM) disappearance (i.e. fermentability) and synthesis of SCFA despite difference in microbial population.
In this study, growing pigs were fed a human-type diet and ileal and hindgut OM fermentations were determined in the same animal using optimised in vivo/in vitro fermentation assays(Reference Montoya, de Haas and Moughan18,Reference Coles, Moughan and Awati19) . This is the first study able to quantify OM fermentability, OM fermented, synthesis and absorption of SCFA in the ileum. The quantitative importance of ileal fermentation was then assessed by comparing the ileal fermentation parameters with those obtained in the hindgut fermentation. The growing pig was used as an animal model for the adult human foregut digestion(20,Reference Deglaire and Moughan21) , as collection of digesta in different locations of the small intestine is difficult and invasive in humans.
Materials and methods
In vivo assay
Dietary treatment
A high-fibre diet comprising foods commonly consumed by humans(Reference Baer, Rumpler and Miles22) was formulated to meet the nutrient requirements of the growing pig (National Research Council(23)) (online Supplementary Table S1). Titanium dioxide (TiO2) was added to the diet as an indigestible marker.
Animals housing and experimental design
Ethics approval for the animal trial was obtained from the Massey University Animal Ethics Committee (Palmerston North, New Zealand). The animal housing and experimental design were as described in detail by Montoya et al. (Reference Montoya, de Haas and Moughan18). Briefly, five 9-week-old entire male pigs (Hampshire × (Landrace × Large white), 23 (sd 1·6) kg body weight) were housed individually in metabolism pens (1·5 × 0·5 m) in a room maintained at 24 (sd 2·4)°C with a 10 h light–14 h dark cycle. Pigs received the experimental diet for 14 d, with a gradual adaptation from commercial diet to human-type diet during the first 3 d. The daily ration was 100 g DM/kg metabolic body weight (BW0·75) per d and given as two equal meals at 08.00 and 16.00 hours. Pigs had free access to water during the study. Pigs were monitored during feeding. After feeding, cages were thoroughly washed and toys were provided to the pigs. On day 15, pigs were fed half their daily ration as one meal and euthanised 5 h postprandial by intracardial injection of sodium pentobarbitone (0·3 ml Pentobarb 300 per kg body weight; Provet). The small intestine was dissected out immediately and ligated into three equal parts. Digesta from the last 50 cm of the second (approximate terminal jejunum) and last (approximate terminal ileum) thirds of the small intestine were collected and used as substrates for the in vitro ileal and hindgut fermentation, respectively (Fig. 1). Digesta from the remaining final third (i.e. last third minus terminal ileum) of the small intestine were collected along with caecal digesta for preparing microbial inocula, for the in vitro ileal and hindgut fermentation, respectively. All digesta were collected in plastic bags containing carbon dioxide before being stored in insulated containers at 4°C to minimise bacterial activity during the time required to weigh the fresh digesta substrates. Representative samples of terminal jejunal digesta, terminal ileal digesta and faeces were collected in Eppendorf tubes and stored at –20°C for determination of concentration of SCFA. For the microbial analysis, aliquots from the ileal (i.e. last third minus last 50 cm) and caecal digesta were collected in Eppendorf tubes and stored at –80°C. Terminal jejunal and terminal ileal digesta and faeces were also collected, stored at –20°C, freeze-dried and finely ground for the determination of DM, OM and TiO2.
Schematic overview of the in vivo/in vitro ileal and hindgut fermentation methodology.
In vitro fermentation assays
A combined in vivo/in vitro methodology was used based on the growing pig. The pig provided both the substrate (terminal jejunal and terminal ileal digesta) entering each of the fermentation sections (ileum and hindgut) and the microbial inocula (ileal and caecal digesta) for the ileal and hindgut fermentation, respectively(Reference Montoya, de Haas and Moughan18). The substrate and inoculum of each fermentation section are then fermented in vitro to be able to determine OM fermentability and synthesis of SCFA. The in vivo and in vitro results were then combined to predict amounts of OM fermented and synthesis and absorption of SCFA.
The combined in vivo/in vitro ileal and hindgut fermentation assays were optimised for different parameters such as incubation time, amount of digesta and pH medium as reported elsewhere(Reference Montoya, de Haas and Moughan18,Reference Coles, Moughan and Awati19) . For example, there was no effect of the incubation time (1–7 h) on the ileal OM fermentability using the in vivo/in vitro methodology(Reference Montoya, de Haas and Moughan18).
Ileal fermentation
The in vitro ileal fermentation was performed according to Montoya et al. (Reference Montoya, de Haas and Moughan18). The inoculum was prepared by pooling ileal digesta (digesta from the final third of small intestine minus the last 50 cm) from all pigs and mixing with a sterilised anaerobic 0·1m phosphate buffer saline (PBS) solution (4·1 mm l-cysteine, pH 7). The ratio digesta:PBS was 0·22:1, w/v. Bottles contained either 1 g of fresh substrate (i.e. terminal jejunal digesta for each of the five pigs) suspended in 5 ml of PBS or PBS alone (blanks). These bottles were then inoculated with 5 ml of the ileal inoculum. A total of six bottles per pig were used. The ileal fermentation was conducted anaerobically at 37°C for 2 h.
Hindgut fermentation
The in vitro hindgut fermentation was performed according to Coles et al. (Reference Coles, Moughan and Awati19). The inoculum was prepared by pooling caecal digesta from all pigs and mixing them with a sterilised anaerobic 0·1m PBS solution (4·1 mm l-cysteine, pH 7). The ratio digesta:PBS was 0·33:1, w/v. Bottles contained either 1 g of fresh substrate (i.e. terminal ileal digesta for each of the five pigs) suspended in 5 ml of PBS or PBS alone (blanks). These bottles were then inoculated with 5 ml of the caecal inoculum. A total of six bottles per pig were used. Hindgut fermentation was conducted anaerobically at 37°C for 24 h.
After ileal and hindgut fermentation, the contents of three bottles were analysed to determine the concentration of SCFA. The remaining three bottles were autoclaved (121°C for 20 min) to inactivate the bacteria and remove fermentation products prior to OM determination. The values of the three bottles for concentration of SCFA and OM determination were averaged per pig. Thus, the number of replicates was five for both ileal and hindgut fermentation.
Chemical analysis
The diet and substrate materials were analysed in duplicate for DM, ash, OM (DM – ash), TiO2(Reference Short, Gorton and Wiseman24), starch (Kit AA/AMG; Megazyme), crude protein (N × 6·25; using a LECO elemental analyser)(25) and lipids (by Soxhlet extraction using petroleum diethyl ether)(25). The diet was also analysed for gross energy (using a LECO AC-350 Automatic Calorimeter) and soluble and insoluble dietary fibre(Reference Prosky, Asp and Schweizer26). DM, ash and OM contents were also determined on material remaining after the in vitro fermentation. The concentration of SCFA was determined in the terminal jejunal digesta, terminal ileal digesta and faeces, and in the samples after in vitro fermentation, as described previously(Reference Montoya, Rutherfurd and Moughan27) with iso-caproic acid as an internal standard.
Microbial analysis
DNA extraction
DNA was extracted from ileal and caecal digesta (0·25 g) using the DNeasy Powersoil kit (QIAGEN), with alterations described by Healey et al. (Reference Healey, Murphy and Butts28). Prior to extraction, the sample was homogenised in bead tubes (0·1 mm and 0·5 mm mix in bead solution) using a FastPrep-24 5 G instrument (MP Biomedicals) at 5·5 m/s for three 60 s cycles with 5 min rest on ice in between. Extracted DNA was quantified and quality-checked on a Qubit fluorometer (Invitrogen) and QIAxpert spectrophotometer (QIAGEN), respectively.
Quantitative PCR
Escherichia coli (Nissle) was used as a representative bacterium for the total bacteria and was grown in tryptic soya broth (Oxoid) at 37°C aerobically. Cell density was determined using a haemocytometer (Neubauer), and the culture was concentrated as required to 1·0 × 109 cells/ml. DNA was then extracted as described above. A standard curve was constructed using 1:10 dilutions of the extracted standard DNA. Samples and standards were run in triplicate by absolute quantification on the Light Cycler 480 real-time PCR instrument (Roche). SYBR Green I Master Mix (Roche) detection chemistry was used to detect double-stranded DNA amplification. The total reaction volume was 20 µl, consisting of 10 µl SyBr Green I Master mix, 4 µl forward primer (2·5 µm), 4 µl reverse primer (2·5 µm) and 2 µl DNA template or sterile water (blank). Each quantitative PCR run included one activation cycle (95°C, 5 min), 32–40 run cycles (including denaturation (95°C, 30 s), annealing (60°C, 60 s) and extension (72°C, 60 s)), and one melt curve cycle (60–95°C at 0·1°C/s with continuous fluorescence acquisition) followed by a cooling cycle (40°C). The melt curve cycle enabled the differentiation between target product and non-specific double-stranded products such as primer-dimers. The universal primers used were forward (5′- TCCTACGGGAGGCAGCAGT) and reverse (5′ – GGACTACCAGGGTATCTAATCCTGTT)(Reference Nadkarni, Martin and Jacques29).
16S ribosomal RNA gene sequencing and bioinformatics
Purified DNA from each sample was sent to the Massey Genome Service (Massey University). The samples underwent library preparation as previously described(Reference Kozich, Westcott and Baxter30) using primers that amplified the V3–V4 hypervariable region of the 16S rRNA gene, 16SF_V3 (5′ – AATGATACGGCGACCACCGAGATCTACAC-index-TATGGTAATTGGCCTACGGGAGGCAGCAG) and 16SR_V4 (5′ – CAAGCAGAAGACGGCATACGAGAT-index-AGTCAGTCAGCCGGACTACHVGGGTWTCTAAT). The library was pooled at equal concentrations and run on one lane of an Illumina MiSeq instrument using 2 × 250 bp paired-end chemistry. Quantitative Insights Into Microbial Ecology software version 1.8.0 was used to analyse the Illumina MiSeq sequencing data(Reference Caporaso, Kuczynski and Stombaugh31). To assemble the forward and reverse reads into a continuous sequence, PANDASeq was used with parameters of at least 40 bp overlap, a minimum of 350 bp length and a maximum of 500 bp length. Chimeras were filtered from the sequences and the reads clustered into operational taxonomic units (OTU) based on a 97 % identity threshold using USEARCH (-cluster_fast command with default parameters)(Reference Edgar32,Reference Edgar, Haas and Clemente33) . Sequence alignment was carried out using PyNAST with reference to the Greengenes database (version 13_8)(Reference DeSantis, Hugenholtz and Larsen34). The resultant OTU table was denoised by removing taxa with fewer than five total sequences across all samples. Alpha rarefaction was calculated using the Faith’s Phylogenetic Diversity metric(Reference Faith35) to a rarefaction depth of 10 000 sequences. Beta diversity was determined using Euclidean distances as input to generate principal coordinate plots. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States was used to predict the functional profiling of ileal and caecal microbial communities.
Calculations
Due to different amounts of DM entering the ileum and hindgut, data were normalised for dietary DM intake based on the ratio of the indigestible marker in the diet and digesta (i.e. data were expressed per kg diet DM intake). Normalising of parameters allowed comparing gut locations, as measures were expressed in the same unit. The calculations used for determining in vitro OM fermentability(Reference Montoya, Rutherfurd and Moughan27) and in vivo/in vitro fermented OM for either ileal or hindgut fermentation were as follows:
$${\rm{OM}}{\mkern 1mu} {\rm{fermentabilit}}{{\rm{y}}_{in{\kern 1pt} vitro}}{\mkern 1mu} \left( \% \right) = ({\rm{O}}{{\rm{M}}_{{\rm{before}}{\kern 1pt} in{\kern 1pt} vitro{\kern 1pt} {\rm{fermentation}}}}{\rm{ }} - ({\rm{O}}{{\rm{M}}_{{\rm{after}}{\kern 1pt} in{\kern 1pt} vitro{\kern 1pt} {\rm{fermentation}}}}{\rm{ }} - (({\rm{O}}{{\rm{M}}_{{\rm{blank}}{\kern 1pt} {\rm{initial}}}} + {\rm{O}}{{\rm{M}}_{{\rm{blank}}{\kern 1pt} {\rm{final}}}}){\rm{/2}}))){\rm{/}}{\mkern 1mu} {\rm{O}}{{\rm{M}}_{{\rm{before}}{\kern 1pt} in{\kern 1pt} vitro{\kern 1pt} {\rm{fermentation}}}} \times {\rm{100}}$$
$$\hskip -23pt{\rm{Fermented}}\,{\rm{O}}{{\rm{M}}_{in\,vivo{\rm{/}}in\,vitro}}\,({\rm{g/kg}}\,{\rm{diet}}\,{\rm{DM}}\,{\rm{intake}}) = {\rm{OM}}\,{\rm{fermentabilit}}{{\rm{y}}_{in\,vitro}}\,{\rm{/\,100}} \times \left( {{\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm\,{/\,Ti}}{{\rm{O}}_{{\rm{2 - terminal}}\,{\rm{jejunal}}\,{\rm{or}}\,{\rm{terminal}}\,{\rm{ileal}}\,{\rm{digesta}}}}} \right)$$
where OMblank initial and OMblank final are the amount of OM in the blanks prior (initial) and after (final) in vitro fermentation, respectively. In vitro ileal and hindgut fermentation had their own blanks. TiO2-diet and TiO2-terminal jejunal or terminal ileal digesta are the TiO2 concentrations (g/kg DM) in the diet and digesta, respectively. The TiO2-terminal jejunal digesta was used for the in vitro ileal fermentation, whereas TiO2-terminal ileal digesta was used for the in vitro hindgut fermentation.
The normalised total number of bacteria and archaea and number per phyla/genus, and the predicted metabolic activity in ileal and caecal digesta were calculated as follows:
$$\eqalign{
\\ {\rm{Normalised}}{\mkern 1mu} {\rm{total}}{\mkern 1mu} {\rm{number}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} {\rm{bacteria}}{\mkern 1mu} {\rm{and}}{\mkern 1mu} {\rm{archae}}{{\rm{a}}_{{\rm{ileal}}{\kern 1pt} {\kern 1pt} {\rm{or}}{\kern 1pt} {\kern 1pt} {\rm{caecal}}{\kern 1pt} {\kern 1pt} {\rm{digesta}}}}\quad \cr
\\ {\mkern 1mu} {\mkern 1mu} \left( {{\rm{16S}}{\mkern 1mu} {\rm{rRNA}}{\mkern 1mu} {\rm{gene}}{\mkern 1mu} {\rm{copy}}{\mkern 1mu} {\rm{number/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ = {\rm{number}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} {\rm{bacteria}}{\mkern 1mu} {\rm{and}}{\mkern 1mu} {\rm{archae}}{{\rm{a}}_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} {\mkern 1mu} \left( {{\rm{16S}}{\mkern 1mu} {\rm{rRNA}}{\mkern 1mu} {\rm{gene}}{\mkern 1mu} {\rm{copy}}{\mkern 1mu} {\rm{number/kg}}{\mkern 1mu} {\rm{DM}}} \right) \cr
\\ \times ({\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm{/Ti}}{{\rm{O}}_{{\rm{2 - terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}}) \cr} $$
$$\eqalign{
\\ {\rm{Normalised}}{\mkern 1mu} {\rm{number}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} \,{\rm{bacteria}}{\mkern 1mu} {\rm{or}} \cr
\\ {\mkern 1mu} {\mkern 1mu} {\rm{archaea}}{\mkern 1mu} {\rm{per}}{\mkern 1mu} {\rm{phylum/genu}}{{\rm{s}}_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} {\mkern 1mu} \left( {{\rm{16S}}{\mkern 1mu} {\rm{rRNA}}{\mkern 1mu} {\rm{gene}}{\mkern 1mu} {\rm{copy}}{\mkern 1mu} {\rm{number/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ = {\rm{normalised}}{\mkern 1mu} {\rm{total}}{\mkern 1mu} {\rm{number}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} \, \cr
\\ {\mkern 1mu} {\mkern 1mu} {\rm{bacteria}}{\mkern 1mu} {\rm{and}}{\mkern 1mu} {\rm{archae}}{{\rm{a}}_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} {\mkern 1mu} \left( {{\rm{16S}}{\mkern 1mu} {\rm{rRNA}}{\mkern 1mu} {\rm{gene}}{\mkern 1mu} {\rm{copy}}{\mkern 1mu} {\rm{number/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ \times {\rm{relative}}{\mkern 1mu} {\rm{abundanc}}{{\rm{e}}_{{\rm{phylum/genus}}}}\left( \% \right){\mkern 1mu} {\rm{/}}{\mkern 1mu} {\rm{100}} \cr} $$
$$\eqalign{
\\ {\rm{Normalised}}{\mkern 1mu} {\rm{predicted}}{\mkern 1mu} {\rm{metabolic}}{\mkern 1mu} {\rm{activit}}{{\rm{y}}_{{\rm{ileal}}{\kern 1pt} {\rm{and}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} \left( {{\rm{activity/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) = {\rm{relative}}{\mkern 1mu} {\rm{activity/kg}}{\mkern 1mu} {\rm{DM}} \cr
\\ \times ({\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm{/Ti}}{{\rm{O}}_{{\rm{2 - terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{caecal}}{\kern 1pt} {\rm{digesta}}}}) \cr} $$
The synthesis of SCFA during in vitro ileal and hindgut fermentation, estimated in vivo/in vitro synthesis of SCFA (i.e. estimated ileal or hindgut synthesis based on the amount of DM entering either the ileum or hindgut per kg DM diet intake), normalised concentration of SCFA in terminal jejunal digesta, terminal ileal digesta and faeces and estimated in vivo/in vitro disappearance (absorption) of SCFA in ileum and hindgut were determined as described previously(Reference Montoya, Rutherfurd and Moughan27), using the following equations:
$$\eqalign{
\\ {\rm{Synthesis}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} {\rm{SCF}}{{\rm{A}}_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{hindgut}}{\kern 1pt} in{\kern 1pt} vitro}}{\mkern 1mu} \cr
\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,({\rm{mmol/kg}}{\mkern 1mu} {\rm{substrate}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{incubated}}) \cr
\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = ({\rm{SCF}}{{\rm{A}}_{{\rm{after}}{\kern 1pt} in{\kern 1pt} vitro{\kern 1pt} {\rm{fermentation}}}}\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{DM}}} \right) \cr
\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, - {\rm{SCF}}{{\rm{A}}_{{\rm{jejunum}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileum}}{\kern 1pt} {\rm{digesta}}}}\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{DM}}} \right) \cr
\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\, - (({\rm{SCF}}{{\rm{A}}_{{\rm{blank}}{\kern 1pt} {\rm{initial}}}} + {\rm{SCF}}{{\rm{A}}_{{\rm{blank}}{\kern 1pt} {\rm{final}}}}){\rm{/2}})) \cr} $$
$$\eqalign{{\rm{Estimated}}\,{\rm{synthesis}}\,{\rm{of}}\,{\rm{SCFA}}_{{\rm{ileal\,or\,hindgut}}}{\,_{in\,vivo/in\,vitro}} \\ \quad\,\,( {{\rm{mmol/kg}}\,{\rm{diet}}\,{\rm{DM}}\,{\rm{intake}}} ) \\ = {\rm{synthesis}}\,{\rm{of}}\,{\rm{SCFA}}_{{\rm{ileal}}\,{\rm{or}}\,{\rm{hindgut}}\, {in\,vitro}}\\ \quad\,\,( {{\rm{mmol/kg}}\,{\rm{substrate}}\,{\rm{DM}}\,{\rm{incubated}}} ) \\ \times ({\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm{/Ti}}{{\rm{O}}_{{\rm{2 - terminal}}\,{\rm{jejunum}}\,{\rm{or}}\,{\rm{terminal}}\,{\rm{ileal}}\,{\rm{digesta}}}})}$$
$$\eqalign{
\\ {\rm{Normalised}}{\mkern 1mu} {\rm{concentration}}{\mkern 1mu} {\mkern 1mu} {\rm{of}} \cr
\\ {\mkern 1mu} {\rm{SCF}}{{\rm{A}}_{{\rm{terminaljejunum}},{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileum}}{\kern 1pt} {\rm{digesta}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{faces}}}}{\mkern 1mu} \cr
\\ \left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ = {\rm{SCF}}{{\rm{A}}_{{\rm{terminal}}{\kern 1pt} {\rm{jejunum}},{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileum}}{\kern 1pt} {\rm{digesta}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{faces}}}}\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{DM}}} \right) \cr
\\ \times ({\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm{/Ti}}{{\rm{O}}_{{\rm{2 - terminal}}{\kern 1pt} {\rm{jejunal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{digesta}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{faces}}}}) \cr} $$
$$\eqalign{
\\ {\rm{Estimated}}{\mkern 1mu} {\rm{disappearance}}{\mkern 1mu} \cr
\\ {\mkern 1mu} {\mkern 1mu} {\rm{of}}{\mkern 1mu} {\rm{SCFA}}{\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right)_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{hindgut}}{\kern 1pt} in{\kern 1pt} vivo}} \cr
\\ = {\rm{Normalised}}{\mkern 1mu} {\rm{concentration}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} \cr
\\ {\mkern 1mu} {\mkern 1mu} {\rm{SCF}}{{\rm{A}}_{{\rm{terminal}}{\kern 1pt} {\rm{jejunum}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileum}}{\kern 1pt} {\rm{digesta}}}}\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ + {\rm{estimated}}{\mkern 1mu} {\rm{synthesis}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} {\rm{SCF}}{{\rm{A}}_{{\rm{ileal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{hindgut}}{\kern 1pt} {\kern 1pt} in{\kern 1pt} vivo/in{\kern 1pt} vitro}}{\mkern 1mu} \cr
\\ \left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) - {\rm{Normalised}}{\mkern 1mu} {\rm{concentration}}{\mkern 1mu} {\rm{of}}{\mkern 1mu} \cr
\\ {\mkern 1mu} {\rm{SCF}}{{\rm{A}}_{{\rm{termina}}{\kern 1pt} {\rm{lileum}}{\kern 1pt} {\kern 1pt} {\rm{digesta}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{faces}}}}\left( {{\rm{mmol/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr} $$
where SCFAblank initial and SCFAblank final are the SCFA (mmol/kg DM) in the blanks prior (initial) and after (final) in vitro fermentation, respectively. In vitro ileal and hindgut fermentation had their own blanks. SCFAjejunum or terminal ileum digesta are the SCFA (mmol/kg DM) in fresh terminal jejunal (ileal fermentation) or terminal ileal (hindgut fermentation) digesta, which represents the SCFA present in the digesta prior to being fermented.
The calculation used to determine the normalised nutrient content was as follows:
$$\eqalign{
\\ {\rm{Normalised}}{\mkern 1mu} {\rm{nutrient}}{\mkern 1mu} {\rm{conten}}{{\rm{t}}_{{\rm{terminal}}{\kern 1pt} {\rm{jejunal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} {\mkern 1mu} \left( {{\rm{g/kg}}{\mkern 1mu} {\rm{diet}}{\mkern 1mu} {\rm{DM}}{\mkern 1mu} {\rm{intake}}} \right) \cr
\\ = {\rm{nutrient}}{\mkern 1mu} {\rm{concentratio}}{{\rm{n}}_{{\rm{terminal}}{\kern 1pt} {\rm{jejunal}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{digesta}}}} \cr
\\ {\mkern 1mu} \left( {{\rm{g/kg}}{\mkern 1mu} {\rm{DM}}} \right) \times \left( {{\rm{Ti}}{{\rm{O}}_{{\rm{2 - diet}}}}{\rm{/Ti}}{{\rm{O}}_{{\rm{2 - terminal}}{\kern 1pt} {\rm{jejunum}}{\kern 1pt} {\rm{or}}{\kern 1pt} {\rm{terminal}}{\kern 1pt} {\rm{ileal}}{\kern 1pt} {\rm{digesta}}}}} \right) \cr} $$
Statistical analysis
For this study, a sample size of five replicates was required to detect a statistical difference (5 %) between GIT locations, with a power >80 % at a two-tailed 5 % significance level, based on variance (sd 2·4 %) and means reported in previous studies(Reference Montoya, Henare and Zhu7,Reference Bach Knudsen and Canibe11,Reference Awati, Williams and Bosch36) .
The statistical analyses were performed in SAS version 9.4 (SAS Institute). The difference between ileal and hindgut fermentation within each pig for the OM fermentability, fermented OM, estimated synthesis of SCFA, estimated disappearance of SCFA, the normalised number of bacteria and archaea (total, phyla and genus) and the normalised predicted metabolic activity was tested using a paired t test. The normal distribution of the difference for the t test was evaluated with the use of the Output Delivery System graphics and univariate procedure of SAS. Probability values of P < 0·05 were considered statistically significant, and a trend when 0·05 < P < 0·10.
The non-parametric two-sample t test (Monte Carlo permutation) from Quantitative Insights Into Microbial Ecology was used to determine alpha diversity significance. The Euclidean distance Principal Coordinates Analysis groupings were tested for significant separation using the adonis test (9999 permutations).
Results
With the exception of a pig that had loose stools during the first experimental days, all pigs were healthy.
Ileal and hindgut organic matter fermentability were not different, but the quantity of ileal fermented organic matter was greater than the organic matter fermented in the hindgut
There was no difference (P = 0·12) between ileal and hindgut in vitro OM fermentability, 28 (sem 1·9) and 35 (sem 2·1) %, respectively (Fig. 2). However, there was 66 % more fermented OM in the ileum than in the hindgut (P = 0·04).
Ileal and hindgut in vitro organic matter (OM) fermentability values (a) and in vivo/in vitro fermented OM (b) of pigs fed a human-type diet (n 5). The line for each gastrointestinal tract location represents the mean value.
Ileal and caecal microbiota have different taxonomic composition, diversity and predicted metabolic activity
The normalised total number of bacteria and archaea in the caecum tended to be greater than in the ileum (P = 0·09; Table 1). Ileal digesta contained greater (P < 0·05) numbers of the family Micrococcaceae (42-fold greater), and the genera Enterococcus (32-fold greater) and Leuconostoc (55-fold greater). In contrast, caecal digesta had greater (P < 0·05) numbers of the class Clostridiales (57-fold greater), the families Coriobacteriaceae (91-fold greater), Lachnospiraceae (443-fold greater), Tenericutes (24-fold greater) and Ruminococcaceae (227-fold greater), and the genera Methanosphaera (15-fold greater), Blautia (65-fold greater), Coprococcus (17-fold greater), and Ruminococcus (82-fold greater). Some bacteria observed in considerable numbers in the caecal digesta (e.g. Bacteriodales, Prevotella, Dorea, Lachnospira, Roseburia, Erysipelotrichaceae, Treponema and TM7-3_F16) were not detected in ileal digesta. Similar conclusions can be drawn from the relative abundance data (non-normalised data) (online Supplementary Figs. S1 and S2). The alpha diversity tended to differ between the ileal and caecal microbiota (P = 0·09; Fig. 3). Based on the normalised data, the microbiota in the caecal digesta resembled a more closely related community than the microbiota in the ileal digesta (Fig. 4). The differences in the ileal and caecal microbiota composition are reflected in differences in their predicted metabolic activity according to Phylogenetic Investigation of Communities by Reconstruction of Unobserved States analysis of pathways related to carbohydrate and protein metabolism (online Supplementary Table S2). For example, the caecal predicted metabolic activity for pyruvate metabolism was 1·2-fold greater (P = 0·02) than the ileal pyruvate metabolism.
Taxonomic composition (×109 16S ribosomal RNA gene copy number/kg diet DM intake) in ileal and caecal digesta of pigs fed a human-type diet (n 5)*
(Mean values with their standard errors)
ND, not detected.
* Only bacteria phyla/genera with >1 % abundance in at least one of the samples are reported.
† Bacteria could only be classified as far as class, order or family level.
‡ Statistical analysis was only conducted when bacteria were detected in a minimum of three pigs.
Alpha diversity numbers showing Faith’s phylogenetic diversity of microbial communities in ileal and caecal digesta of pigs fed a human-type diet (n 5) based on the normalised number of bacteria and archaea. The line for each gastrointestinal tract location represents the mean value.
Euclidean principal coordinates (PC) analysis of the distances of normalised relative abundance data (16S ribosomal RNA gene copy number/kg diet DM intake) in caecal (
) and ileal digesta (
) for pigs fed a human-type diet (n 5). Groupings exhibited a significant difference (P = 0·01, R 2 = 0·43) as determined by the non-parametric adonis test (9999 permutations). Most prevalent taxa responsible for variation in the plot are displayed.
Estimated synthesis of SCFA and disappearance differs between ileal and hindgut fermentation
The estimated synthesis of acetate and iso-valerate was 6·1- and 1·3-fold greater (P < 0·05), respectively, during ileal fermentation compared with hindgut fermentation (Fig. 5). In contrast, the estimated synthesis of propionate, butyrate and valerate was greater (4·6-, 8·3- and 4·5-fold, respectively; P < 0·05) during hindgut fermentation compared with ileal fermentation. Similar trends were observed for the in vitro synthesis of SCFA (online Supplemenatry Table S3). Based on the estimated disappearance data (Fig. 6), most SCFA disappeared in the GIT location where they were synthesised and the statistical differences were as for the synthesis of SCFA values.
Estimated synthesis of SCFA (mmol/kg diet DM intake) in the ileum and hindgut of pigs fed a human-type diet (n 5). The line for each gastrointestinal tract location represents the mean value.
Estimated disappearance of SCFA (mmol/kg diet DM intake) in the ileum and hindgut of pigs fed a human-type diet (n 5). The line for each gastrointestinal tract location represents the mean value.
Discussion
The ileal fermentation results confirm earlier observations in pigs(Reference Jaworski and Stein10,Reference Bach Knudsen and Canibe11,Reference Montoya, de Haas and Moughan18) and establishes that such fermentation is observed for a human-type diet. Indeed, one of the main results of this study was that the amount of fermented OM was 1·5-fold greater in the ileum compared with the hindgut, which is explained by a greater amount of OM entering the ileum than the hindgut (246 v. 141 g OM/kg diet DM intake) (online Supplementary Table S4). A greater amount of acetate and iso-valerate were synthesised during ileal fermentation compared with hindgut fermentation, but for hindgut fermentation, the synthesis of butyrate, propionate and valerate was greater. The present results are the first to demonstrate that SCFA disappeared in the same GIT location where they were synthesised.
The normalised total number of bacteria and archaea (16S rRNA gene copy number/kg diet DM intake) in the caecal digesta (19·5 × 1011) tended to be greater than in the ileal digesta (8·3 × 1011). Rowan et al. (Reference Rowan, Moughan and Wilson37) reported that, when taking the dietary DM into account, concentration of DNA (i.e. microbial marker) in fresh ileal digesta and faeces of pigs fed a human-type diet was not different. Recently, Montoya et al. (Reference Montoya, Henare and Zhu7) reported two times greater normalised total number of bacteria (per kg diet DM intake) in ileal digesta compared with faeces of pigs fed diets containing kiwifruit as the sole dietary fibre source. When the concentration data are normalised for diet DM intake, different conclusions are drawn and the potential role of the ileal microbiota in fermenting undigested material is highlighted. In the pig ileal digesta, the predominant bacterial genus was Streptococcus, which is also the predominant bacterial genus in ileal effluent from human ileostomates(Reference Zoetendal, Raes and van den Bogert5). Streptococcus is well adapted to the ileum because it has the ability to rapidly ferment simple carbohydrates (i.e. mono-, di- and oligosaccharides), which is important in the ileum as the retention time is shorter than in the hindgut(Reference Zoetendal, Raes and van den Bogert5,Reference Graff, Brinch and Madsen16) . A greater number of Streptococcus was observed in the ileal digesta of the pigs fed the human-type diet compared with caecal digesta. This may be related to a tending (P = 0·07) towards a greater amount of starch, a rapidly fermentable carbohydrate, entering the ileum (31 g/kg diet DM intake) compared with the hindgut (14 g/kg diet DM intake) (online Supplementary Table S4). In caecal digesta, the main bacteria present belonged to the class Clostridiales, the family Ruminococcaceae and the genera Ruminococcus, Streptococcus and Prevotella. Both Clostridium sp. and Ruminococcus sp. have the ability to ferment cellulose(Reference Wedekind, Mansfield and Montgomery38), a fibre that needs a longer time to ferment. Cellulose is expected to be a main component of the insoluble dietary fibre fraction of the human-type diet (online Supplementary Table S1). The predicted metabolic activity related to carbohydrate and protein metabolism demonstrated that the ileal and caecal microbiota of the pigs fed the human-type diet had different metabolic activity profiles. The overall ileal predicted metabolic activity was 84 % of the overall caecal predicted metabolic activity. Despite the differences in predicted metabolic activity, similar degrees of OM fermentability were observed during ileal and hindgut fermentation. The tendency towards greater microbial diversity in caecal digesta compared with ileal digesta coincides with a longer transit time in the hindgut, which gives bacteria a greater opportunity to grow, and for cross-feeding to occur(Reference Falony, Joossens and Vieira-Silva39). The differences in the microbial community indicate that the ileal microbial community has evolved aligned to ferment rapidly fermentable substrates in accordance with the faster transit time, whereas the hindgut microbial community has evolved to ferment more slowly fermentable substrates in line with the slower transit time. The co-existence of these microbial communities may result in more efficient and effective fermentation of diets as humans eat diets that are complex in nature and composition.
During ileal fermentation, acetate was the main SCFA synthesised, whereas butyrate and propionate were synthesised in greater amounts during hindgut fermentation. Similar trends were reported for concentrations of these SCFA in the ileal and caecal digesta of adult humans suffering sudden death(Reference Cummings, Pomare and Branch40). Concentration data need to be interpreted carefully, however, as concentration of SCFA represents only the amount of SCFA that has not been absorbed at the time of collection. Synthesis data are more meaningful. The differences seen here for synthesis of SCFA related to the ileal and hindgut fermentation may be related to several factors: (i) the incubation time of ileal fermentation (2 h) compared with hindgut fermentation (24 h), (ii) the microbial composition and (iii) the amount and chemical composition of the substrate available. A longer fermentation time may be one of the factors explaining the greater butyric acid synthesis. Longer fermentation time allows cross-feeding to occur whereby SCFA like acetate can be converted into other SCFA, like butyrate(Reference Boets, Gomand and Deroover41–Reference Walker, Duncan and Carol McWilliam Leitch43). This would lower the concentration of acetate while increasing the concentration of butyrate. Another factor explaining the greater hindgut butyrate synthesis is higher numbers of butyrate-synthesising bacteria reported in caecal digesta. Similarly, the greater caecal propionate synthesis may be related to higher numbers of propionate-synthesising bacteria such as Prevotella (Reference Long, Zhang and Liu44) reported in the caecal digesta. The tendency to a lower phylogenetic diversity in the ileum compared with the caecum may explain the greater concentration of acetate observed after ileal fermentation compared with hindgut fermentation. Almost all GIT bacteria have the ability to synthesise acetate, whereas butyrate and propionate synthesis pathways are highly conserved in a limited number of GIT bacteria(Reference Louis and Flint45). The amount of crude protein entering the ileum was 2-fold greater than the amount entering the hindgut (online Supplementary Table S4), which may explain the greater iso-valerate synthesis during ileal fermentation since iso-valerate is a product of protein fermentation(Reference Macfarlane, Gibson and Cummings46). The SCFA synthesised during ileal fermentation are expected to have similar effects on the ileal microbiota as has been reported for the caecal microbiota. For example, the synthesis of SCFA reduces the pH and promotes the growth of different bacteria, like Roseburia (Reference Walker, Duncan and Carol McWilliam Leitch43).
Based on the estimated disappearance of SCFA, the SCFA were absorbed and/or metabolised in the same GIT location as they were synthesised. Previously, human studies have shown that SCFA can be absorbed in both the ileum and hindgut(Reference Schmitt, Soergel and Wood47,Reference McNeil, Cummings and James48) . These results indicate that SCFA synthesised in the ileum may have a local effect at the ileal epithelium and/or be absorbed to serve systemically in the host. For example, human ileal epithelial cells contain NEFA receptors, which, upon binding with SCFA, can stimulate the production of satiety hormones, such as peptide YY and glucagon-like peptide-1(Reference Kaji, Karaki and Tanaka49), which then increase ileal motility(Reference Kamath, Phillips and Zinsmeister50). Dietary intervention may be a strategy to modulate the ileal synthesis of SCFA and therefore the delivery of SCFA both locally and systemically. Further studies to investigate the influence of diet on ileal microbiota, their fermentation capacity and synthesis of SCFA, and how ileal fermentation affects the host are warranted.
To compare estimated ileal and hindgut fermentation in the present work, a combined in vivo/in vitro methodology was used. Limitations and advantages of this methodology have been described previously(Reference Montoya, de Haas and Moughan18). One of the limitations of this in vivo/in vitro methodology is that in vivo absorption and fermentation of dietary and non-dietary nutrients occur simultaneously in the ileum. During the in vitro fermentation, this absorption is not simulated. The presently described study involves only one human-type diet, and the work needs to be extended to more diverse diets and sources of dietary fibre. That the human-type diet was highly fermentable in the ileum, however, demonstrates that ileal fermentability may be a hitherto largely under-recognised yet important characteristic of foods and diets consumed by humans. Some studies have reported important ileal digestibility values of dietary fibre in human ileostomates(Reference Englyst and Cummings8,Reference Holloway, Tasman-Jones and Maher9) . Both the ileal and hindgut fermentation of foods need to be better understood. To apply the combined in vivo/in vitro methodology in humans, however, requires an animal model to allow for the sampling of terminal jejunal digesta and ileal digesta. A faecal inoculum sourced from adult humans can replace the caecal inoculum as used here for determining hindgut fermentation(Reference Coles, Moughan and Awati19). The growing pig is a valid animal model for the adult human for the foregut digestion of food(20,Reference Miller and Ullrey51) , and it can be expected that pig terminal ileal digesta samples can be used to provide a suitable substrate for the hindgut fermentation methodology (faecal inoculum). The question remains, however, as to whether pig ileal digesta provide a suitable inoculum for a fermentation assay related to human ileal fermentation. There may be differences in the microbiota between the two species, and controlled comparisons in this respect are needed. The second question is whether the species differences in the microbiota composition if they do occur have a significant effect on fermentation. As demonstrated in this experiment, despite differences between ileal and hindgut microbial composition, OM fermentability was not different between the two sections of the intestinal tract. The pig-based in vivo/in vitro ileal fermentation model needs to be fully validated, as does its extension to evaluating human foods.
In conclusion, results of this experiment demonstrate that a diet consisting of foods commonly consumed by humans was well fermented in the ileum of the growing pig. Indeed, the amount of fermented OM was greater during ileal than hindgut fermentation. The ileal fermentation synthesised important amount of SCFA (mainly acetic acid), which are absorbed or metabolised in the ileum. Given similarities in the foregut of humans and pigs, such fermentation may also occur in humans, but the suitability of the growing pig as an animal model for the adult human to provide inocula and substrate for the fermentation methodology remains to be assessed.
Acknowledgements
We acknowledge Patrick Morel for assisting with the Animal Ethics application.
This study was funded by the Centre of Research Excellence Fund from the Tertiary Education Commission and the New Zealand Ministry of Education
C. A. M. and P. J. M. designed the study. C. A. M. and E. S. H. conducted the in vivo/in vitro fermentation experiments. P. B. conducted the microbial analysis and analysed the microbial data. A. M. E. H. analysed data, performed statistical analysis and wrote the paper. C. A. M., P. J. M., P. B. and W. C. M. critically reviewed the manuscript. All authors have read and approved the final manuscript.
The authors declare that there are no conflicts of interest.
Supplementary material
For supplementary material referred to in this article, please visit https://doi.org/10.1017/S0007114520001385
