Study setup
Six adults and six elderly, who were included in a previously conducted in vivo GOS intervention study [11], donated their faecal material for the current study (Fig. S1) at their first visit or at least 4 weeks after the intervention period. Each participant defecated into a stool collector (Excretas Medical BV, Enschede, the Netherlands). Directly after defecation, faecal material was divided into two portions. A small portion (~0.5 g) was frozen immediately. The remaining faeces was anoxically cryo-conserved and used as inoculum for the in vitro incubations. The viability of different microbial groups in the anoxically cryo-conserved faecal material was determined with propidium monoazide (PMA) dye. The in vitro incubations lasted for 24 h with samples collected in duplicate to compare microbiota composition, carbohydrate degradation and metabolite production between age groups (adults vs elderly). The degrading capacity for two typical bifidogenic carbohydrates, i.e., GOS and 2′-FL, was determined for the microbiota of all six adults and six elderly and compared to a non-carbohydrate control. To further extend these experiments, we also studied the degradation of other typical bifidogenic carbohydrates, i.e. FOS, inulin, and IMMP, using the faecal inocula of three adults and three elderly for which sufficient material was still available.
Participants
The six adults (20–30 yrs) and six elderly participants (70–85 yrs) of the intervention study [11] were randomly contacted and participated in the current study, who differed significantly in age, but not in sex, BMI, alcohol consumption, smoking, medication use or dietary fibre intake (Table 1). None of the participants took acid inhibitors (e.g., proton pump inhibitors), nor antibiotics 90 days prior to the study, nor did any of the participants have a chronic disorder or major surgery, as these factors potentially could have limited participation, completion of the study, or interfered with the study outcomes. Detailed description of the inclusion and exclusion criteria has been provided previously [11]. Subject codes as shown in the results were randomly assigned in the data analysis phase and cannot be traced back to individual subjects without the specific randomization key. The study was approved by the medical Ethics Committee of the Maastricht University Medical Center+ and registered in the US National Library of Medicine (http://www.clinicaltrials.gov) with the registration number NCT03077529 [11].
Dietary intake
Participants in the current study completed the dietary records on 3 consecutive days, after instructed to record their food, beverage and dietary supplement intake based on standard household units. Their nutrient intake was analyzed using the online dietary assessment tool of The Netherlands Nutrition Centre (www.voedingcentrum.nl).
Carbohydrates
Five different carbohydrates, i.e., GOS, 2′-FL, FOS, inulin and IMMP were used as sole carbon sources in this study. GOS and the human milk oligosaccharide 2′-FL (Fucα1-2Galβ1-4Glc) were kindly provided by Friesland Campina (Amersfoort, The Netherlands). In order to mimic the actual portion of GOS utilized by intestinal microbiota, purified GOS with <3% monomers and lactose was used. Size distribution of mono- and oligomers was as follows: 2.4% degree of polymerization (DP)1, 11.3% DP2, 41.8% DP3, 25.6% DP4, 12.1% DP5, 4.6% DP6, 1.4% DP7, 0.34% DP8, 0.11% DP9. FOS and inulin were kindly provided by Sensus (Roosendaal, the Netherlands). FOS or oligofructose (Frutalose® OFP) is derived from partial enzymatic hydrolysis of inulin from chicory and consisted for 92 ± 2 % of FOS (DP2-10) and for 8 ± 2 % of a mixture of fructose, glucose and sucrose. Long-chain inulin (Frutafit® TEX!), termed as inulin in the current study, is also derived from chicory, comprising ≥99.5% inulin (DP2 to 60, average chain length ≥22 monomers), and ≤0.5% mixture of fructose, glucose and sucrose. The IMMP is IMMP-92 (AVEBE, Groningen, the Netherlands) which is a novel indigestible α-glycan derived from starch, with 92% of α-(1 → 6) glycosidic linkages [21].
Faecal sample collection and storage
To store and transport freshly defecated faeces under anoxic conditions, Anaerocult® A mini (Merck KGaA, Darmstadt, Germany) was activated with 10 ml nuclease-free water (Promega, Madison, WI, US), and placed next to the faeces in the stool collector before the lid was closed to create an anoxic atmosphere. Afterwards, the stool collector and two open bags of AnaeroGen (AnaeroGenTM 3.5 L Sachet, Thermo Scientific, Waltham, Massachusetts, US) were put into an anoxic box (AnaeroPack™ 7.0 L Rectangular Jar, Thermo Scientific) and stored at 4 °C until transportation. Samples were transported on ice from Maastricht to Wageningen University & Research within 9 h. After arrival, the anoxic box was transported immediately into the anaerobic chamber (MK3 Workstation, Don Whitley, UK), filled with an atmosphere of 4% H2 and 96% N2. For each donated sample, every 17.5 g faeces were mixed with 7.5 g dialysate (Tritium Microbiologie, Eindhoven, the Netherlands), 35.7 g nuclease-free water and 9.8 ml glycerol. The mixed faecal slurry was transferred into a serum bottle and sealed with a butyl rubber stopper and metal crimp cap inside the anaerobic chamber, and afterwards stored at −80 °C.
In vitro incubations
Anoxically cryo-conserved faecal inoculum was defrosted and transferred to an anaerobic chamber filled with an atmosphere of 96% N2 and 4% H2 (BACTRON 300, Shel Lab, Cornelius, Oregon, US). Using standard ileal efflux medium (Tritium Microbiologie) [20], incubation was done with one of the carbohydrates (10 mg/ml) and 10% (v/v) faecal inoculum in duplicate, while incubations without faecal inoculum or without carbohydrates, respectively, served as controls. Specifically, every 1 L of medium comprised 400 ml BCO (60 g/L casein, 60 g/L bacto peptone and 1 g/L ox bile), 16 ml salts solution (156.3 g/L di-potassium hydrogen phosphate, 281.3 g/L sodium chloride, 28.13 g/L calcium chloride dihydrate, 0.31 g/L iron (II) sulfate heptahydrate,0.63 g/L hemin porcine), 4 ml cystein.HCl solution, 0.8 ml vitamin mix (1 mg/L menadion, 2 mg/L D(+)biotine, 0.5 mg/L Vitamin B12, 10 mg/L D(+)pantothenate, 5 mg/L aminobenzoic acid, 4 mg/L thiamine HCL and 5 mg/L nicotinamide adenine dinucleotide free acid) and 100 ml MES (1 M pH 6.0). Except for the to-be-studied carbohydrates (i.e., IMMP, short chain & long chain inulin, 2′-FL and GOS), no additional carbohydrates were added in the fermentation medium. Ten ml batch incubation bottles were used in the current study, and filled with 6 ml fermentation medium. Cultures were incubated at 37 °C on a rotary shaker at 200 rpm for 24 h.
Sample collection
Samples were collected 0, 4, 10 and 24 h after inoculation (Fig. S2). Specifically, at each time point, two incubation bottles (the duplicate) per treatment were sacrificed for sample collection. The headspace gas was sampled first to determine H2 and CH4 production. Three aliquots of 1 ml culture were then distributed into 1.5 ml Eppendorf tubes. One of these aliquots was heated at 100 °C for 5 min to determine carbohydrates in the supernatant. Afterwards, all aliquots were centrifuged at 4 °C at 18,600 rcf for 10 min. The supernatants from the other two unheated tubes were stored at −20 °C for metabolite measurement, while the remaining pellets were stored at −80 °C for microbiota analysis.
Carbohydrate, gas, and metabolite measurements
Degradation of GOS, 2′-FL, FOS, inulin and IMMP was determined using High-Performance Anion Exchange Chromatography (HPAEC) with Pulsed Amperometric Detection (PAD). Specifically, samples taken during the incubation were diluted and centrifuged for 15 min at 18,600 rcf. Ten microlitres of supernatant was injected to an ISC5000 HPLC system (Dionex, Sunnyvale, CA, US), which was composed of a CarboPac PA‐1 column (250 mm × 2 mm ID), a CarboPac PA guard column (25 mm × 2 mm ID) and an ISC5000 ED detector (Dionex) in the PAD mode. Detailed description of gradients and dilution factors are provided in the supplementary information. The degradation and size of the large IMMP molecules was also determined using High Performance Size Exclusion Chromatography as described previously [20].
Headspace gas composition was measured using a CompactGC gas chromatograph (Global Analyser Solutions, Breda, The Netherlands), equipped with a Carboxen PDD precolumn (pressure: 200 kPa, split flow: 20 ml/min, column oven: 90 °C, valve oven: 80 °C) with a carrier gas flow of 20 ml/min and a TCD column (pressure: 200 kPa, split flow: 10 ml/min, column oven: 80 °C, valve oven: 80 °C).
Concentration of organic acids was determined by High Performance Liquid Chromatography (HPLC), using a SUGAR SH1821 column (SHODEX, Tokyo, Japan). The column was operated at 54 °C with a flow rate of 0.8 ml/min, using 0.01 N H2SO4 as eluent. The compounds were detected by an RID-20A (Shimadzu, Kyoto, Japan) refractive index detector at a temperature of 40 °C. Four hundred µl of collected supernatant was mixed with 600 µl of 10 mM DMSO in 0.01 N H2SO4, and 10 µl of this mixture was injected for analysis. All analytical measurement data were processed using Chromeleon ™ Chromatography Data System (CDS) Software (Thermo Scientific).
Microbiota composition analysis
The microbiota composition in faecal- and batch incubation samples was determined by sequencing of barcoded 16S ribosomal RNA (rRNA) gene amplicons (details in Supplementary information). In short, total DNA was obtained from the collected pellet by repeated bead beating and purification with a Maxwell® 16 Instrument (Promega, Leiden, The Netherlands). The V4 region of the 16S rRNA gene was amplified in triplicate using barcoded 515F [22] − 806R [23] primers and total bacterial DNA as template as described previously [24]. An equimolar mix of purified PCR products was sent for Illumina Hiseq2500 (2 × 150 bp) sequencing (Eurofins Genomics, Konstanz, Germany). Raw sequence data were processed using NG-Tax 1.0 with default settings [25]. Taxonomy was assigned based on SILVA database version 128 [26, 27]. A detailed description is provided in the supplementary information. The raw sequence data has been uploaded to the European Nucleotide Archive with accession number PEJEB41341. In addition, the total genomic DNA was used for total bacterial quantification as it has been described earlier [11].
Viability measurements
The fraction of viable microbes in the anoxically cryo-conserved faeces was determined with PMA dye, a photoreactive dsDNA-binding dye that only penetrates the envelop of dead cells [28]. Briefly, 1 ml of anoxically cryo-conserved faecal inoculum from each donor was mixed with 2.5 µl of 20 mM PMA dye (Biotium, Inc., Fremont, CA, USA), and incubated at room temperature for 5 min in the dark, followed by treatment with a PMA-LiteTM LED photolysis device (Biotium, Inc.) for 15 min. Subsequently, samples were centrifuged at 4 °C for 10 min (1500 rcf). The pellet was used for microbiota composition analysis.
Statistical analysis
All statistical analyses were conducted in R (R-3.6.3). Relative abundance of microbial taxa was calculated based on 16S rRNA gene sequence read counts. The microbial diversity (Inverse Simpson) and richness (Phylogenetic Diversity) were calculated based on amplicon sequence variants, which were also used to calculate distance matrices. Permutational multivariate analysis of variance (PERMANOVA) was performed based on weighted and unweighted UniFrac distance matrices. Principal coordinate analysis was used to visualize the microbiota composition variation between samples [29]. Microbiota variation partitioning was assessed by fitting environmental variables (i.e., age group, sampling time point and type of carbohydrate) to weighted and unweighted UniFrac distance matrices, using the adonis function in the vegan package [30]. To compare and contrast alterations in microbiota composition with different carbohydrates versus non-carbohydrate control during the incubation, we used principal response curve analysis to identify genera which fit best (weights > 0.05) to explain the observed difference, using the prc function in the vegan package [30]. As for the metabolite data, redundancy analysis (RDA) in combination with Monte Carlo permutation was performed to assess to what extent explanatory variables, i.e., incubation time, subject- and carbohydrate-specificity, could explain the overall variation in metabolite data, using the rda function in the vegan package [30]. To assess the effect of age group (adult vs elderly) on the degradation of carbohydrates/concentration of metabolites during incubation, we analyzed the data using two-way mixed ANOVA, with one between-subjects factor (age group) and one within-subjects factor (incubation time), using the anova_test function in the rstatix package [31]. False discovery rate (FDR) correction according to the Benjamini–Hochberg procedure was applied for multiple testing when applicable. A corrected P value < 0.05 was considered to indicate significant difference.
Source: Ecology - nature.com
