Beta-Glucans Supplementation Associates with Reduction in P-Cresyl Sulfate Levels and Improved Endothelial Vascular Reactivity in Healthy Individuals

Background Oat and barley beta-glucans are prebiotic fibers known for their cholesterol-lowering activity, but their action on the human gut microbiota metabolism is still under research. Although the induction of short-chain fatty acids (SCFA) following their ingestion has previously been reported, no study has investigated their effects on proteolytic uremic toxins p-cresyl sulfate (pCS) and indoxyl sulfate (IS) levels, while others have failed to demonstrate an effect on the endothelial function measured through flow-mediated dilation (FMD). Objective The aim of our study was to evaluate whether a nutritional intervention with a functional pasta enriched with beta-glucans could promote a saccharolytic shift on the gut microbial metabolism and improve FMD. Methods We carried out a pilot study on 26 healthy volunteers who underwent a 2-month dietary treatment including a daily administration of Granoro “Cuore Mio” pasta enriched with barley beta-glucans (3g/100g). Blood and urine routine parameters, serum pCS/IS and FMD were evaluated before and after the dietary treatment. Results The nutritional treatment significantly reduced LDL and total cholesterol, as expected. Moreover, following beta-glucans supplementation we observed a reduction of serum pCS levels and an increase of FMD, while IS serum levels remained unchanged. Conclusions We demonstrated that a beta-glucans dietary intervention in healthy volunteers correlates with a saccharolytic shift on the gut microbiota metabolism, as suggested by the decrease of pCS and the increase of SCFA, and associates with an improved endothelial reactivity. Our pilot study suggests, in addition to cholesterol, novel pCS-lowering properties of beta-glucans, worthy to be confirmed in large-scale trials and particularly in contexts where the reduction of the microbial-derived uremic toxin pCS is of critical importance, such as in chronic kidney disease.

W hole-grain barley and oats and some dry-milled bran grain products were authorized for use according to the health claim made by the Food and Drug Administration that they "decrease the risk of coronary heart disease" (1,2). Whole-grain barley and oat flours contain ␤-glucans, which are soluble dietary fibers. The European Food Safety Authority recognized that the "regular consumption of oat ␤-glucans can actively lower/reduce blood LDL-cholesterol and total cholesterol" (3). In addition to reducing cholesterol, the positive associations between the consumption of ␤-glucans and the reduction/prevention of cardiovascular diseases, as well as the reduction of glycemia, insulin resistance, and metabolic syndrome, have been well documented (4,5). The minimum dose of 3 g/day of ␤-glucans is recommended to achieve positive effects on human health (1,3,6).
Several mechanisms were described to explain the hypocholesterolemic effect of ␤-glucans: (i) increased viscosity at the level of the small intestine and, consequently, slowed gastric emptying, digestion, and absorption of molecules, including glucose, dietary cholesterol, and bile acids (7,8), (ii) decreased enterohepatic bile acids by binding at the intestinal level with subsequent increased use of cholesterol for bile acid synthesis (9,10), (iii) reduced synthesis of hepatic cholesterol due to improved insulin sensitivity (10,11), and (iv) inhibited hepatic synthesis of cholesterol by acetate and propionate, which are produced by colonic fermentation of ␤-glucans (12,13). Saccharolytic and proteolytic fermentations are the major fermentation processes, which are carried out by metabolically active microbes at the colon level (14). Saccharolytic fermentation was associated with the synthesis of shortchain fatty acids (SCFA) (acetate, propionate, and butyrate), intermediate metabolites, such as succinate, acrylate, lactate, formate, and ethanol, and small final molecules (hydrogen, methane, and carbon dioxide) (15). Proteolytic fermentation was asso-ciated with the synthesis of SCFA (acetate, propionate, and butyrate) and branched-chain fatty acids (BCFA) (isobutyric, isovaleric, and 2-methylbutyric acids), free amino acids (FAA), and some potentially toxic metabolites (phenols, indoles, ammonia, and amines) (14,16,17). The type of colonic fermentation mainly depends on the type of microbiota and fermentable substrate availability (17). Based on the key role in human health, the intestinal microbiota was acknowledged as a metabolic organ (18). Most current research on novel functional foods is focused on the selection and characterization of prebiotics (e.g., inulin, fructooligosaccharides and galactooligosaccharides), which are not digested by human gastrointestinal enzymes but selectively stimulate the growth and/or activity of generally regarded as safe (GRAS) bacteria that may improve host health (19). Indeed, a prebiotic effect of ␤-glucans on the intestinal microbiota was also hypothesized (5). In particular, the positive effects of ␤-glucans on the growth of beneficial intestinal lactobacilli and bifidobacteria were shown in in vitro studies (20,21) and animal experiments (22,23). Nevertheless, other studies that also used animal models did not show significant effects of ␤-glucans on lactobacilli and/or bifidobacteria (24,25). Unfortunately, human clinical challenges that have dealt with the prebiotic effect of ␤-glucans are rather scarce. In one study, barley ␤-glucans increased the cell density of colonic bifidobacteria in older healthy subjects (26). A pilot study with polypectomized patients showed no significant effect of ␤-glucans on the fecal microbiota and concentration of SCFA (27). The evidence that ␤-glucans positively influence the human intestinal microbiota is still insufficient for interpretation or difficult to interpret, and additional studies are needed to fill this gap (5).
This study compared the fecal microbiota and metabolomes of healthy subjects before (HS) and after (HSB) 2 months of daily administration of durum wheat flour and whole-grain barley pasta containing the minimum recommended intake (3 g) of ␤-glucans. The fecal microbiota was characterized through integrated approaches, which were based on culture-independent and -dependent methods.

Subjects.
The study was carried out in accordance with the Helsinki Declaration (IV adaptation) and European Guidelines for Good Clinical Practice. The protocol of the study was approved by the institutional review board of the Azienda Ospedaliero-Universitaria Consorziale Policlinico of Bari, Italy (authorization no. 1570/2014). Written consents were obtained from all volunteers. One group of Caucasian HS (15 female and 11 male), aged between 28 and 57 years, were enrolled in the study (Table 1). Exclusion criteria were history of gastrointestinal disease, diabetes, cardiovascular diseases, hyperlipidemia, and consumption of alcohol. Volunteers were not treated with antibiotics and/or functional foods (probiotics and/or prebiotics) for Ն3 months before recruitment and sampling.
Feeding regime. Pasta Granoro Cuore Mio (Granoro s.r.l., Corato, Bari, Italy) was used in this study. Pasta was made by using a mixture of durum wheat flour (75%) and whole-grain barley flour (25%). The gross composition of pasta Granoro Cuore Mio was as follows: moisture, 11%; protein, 11% of dry matter; carbohydrate, 69% of dry matter; fat, 2% of dry matter; total fibers and ␤-glucans, 7% and 3% of dry matter, respectively. Before the administration of 100 g of pasta Granoro Cuore Mio containing 3 g of barley ␤-glucans (Granoro s.r.l.), each volunteer was instructed to follow the usual diet, including 100 g of pasta every day for 2 months. Before and after the dietary treatment, a food frequency questionnaire and a 24-h recall questionnaire were administered to each volunteer by a dietitian. These two tools allowed extrapolation of the weekly cumulative frequency of food and the component intake data, respectively (Tables 2 and 3 ). All volunteers confirmed that no remarkable changes occurred in their meals and medication during the 2 months of treatment. Additionally, a questionnaire aimed at evaluating intestinal effects (bloating, meteorism, constipation, satiety, diarrhea, reflux, and flatulence) was administered (26). Gastrointestinal symptoms were graded from 0 (no symptoms) to 3 (severe symptoms) (26). Before the treatment with durum wheat and whole-grain barley pasta, cholesterolemia, body mass index, and fecal microbiota and metabolome composition were determined.
Determination of blood cholesterol. The level of total cholesterol was measured on the Siemens Dimension RxL Max by using the Siemens enzymatic methods (Siemens Medical Solution Diagnostics, Tarrytown, NY). Low-density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterol levels were estimated by using the Friedewald equation (28) for values of Ͻ300 mg/dl.   . Fecal samples suspended in RNAlater were stored at Ϫ80°C for further RNA and metabolomic analyses. Samples diluted with Amies transport medium were immediately subjected to plate counts and analysis by the Biolog system. RNA extraction from fecal samples. An aliquot of ca. 200 mg of fecal sample diluted in RNAlater was used for RNA extraction with the stool total RNA purification kit (Norgen Biotek Corp., Ontario, Canada). The quality and concentration of RNA extracts were determined using 1% agarose-0.5ϫ Tris buffer-EDTA (TBE) gels and spectrophotometric measurements at 260, 280, and 230 nm through the NanoDrop ND-1000 spectrophotometer. Total RNA extracted (ca. 2.5 g) was transcribed to cDNA using random examers and the Tetro cDNA synthesis kit (Bioline USA, Inc., Taunton, MA, USA), according to the manufacturer's instructions (29).
bTEFAP and data analyses. For each volunteer, the three cDNA samples were pooled and used for bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP) analysis. Pooled samples of cDNA were analyzed for each subject. bTEFAP was performed by Research and Testing Laboratories (Lubbock, TX), according to standard laboratory procedures and using the 454 FLX sequencer (454 Life Sciences, Branford, CT, USA). The primers were forward 28F (GAGTTTGATCNTGGCTCAG) and reverse 519R (GTNTTACNGCGGCKGCTG), based upon the V1 to V3 region (Escherichia coli positions 27 to 519) of the 16S rRNA gene (30). The bTEFAP procedures were performed based upon Research and Testing Laboratories (RTL) protocols (Research and Testing Laboratories, Lubbock, TX). Raw sequence data were screened, trimmed, and filtered with default settings, using the QIIME pipeline version 1.4.0 (http://qiime. org/). Chimeras were excluded by using B2C2 (Research and Testing) (31). Sequences of Ͻ250 bp were removed. The average length of the sequences was 484 bp. The sequences are available on the Research and Testing Laboratories website. The FASTA sequences for each sample, without chimeras, were evaluated using BLASTn against the database derived from GenBank (http://ncbi.nlm.nih.gov) (32).
Bioinformatics and data analysis. The sequences were first grouped into operational taxonomic unit (OTU) clusters with 97% identity (3% divergence), using USEARCH (33). To determine the identities of bacteria, sequences were first queried using a distributed BLASTn.NET algorithm (33) against the database of high-quality 16S bacterial sequences that were derived from the NCBI. Database sequences were characterized as high quality based on criteria originally described by Ribosomal Database Project (RDP) version 10.28 (34).
Alpha diversity (determined by rarefaction, Good's coverage, Chao1 richness, and Shannon diversity indices) and beta diversity measures were calculated and plotted using QIIME. Diversity was examined from two perspectives. First, overall richness (i.e., number of distinct organisms present within the microbiome) (alpha diversity) was expressed as the number of OTUs and was quantified using the Chao1 richness estimator: , where n1 is the number of singletons (species captured once) and n2 is the number of doubletons (species captured twice).
Second, overall diversity (which is determined by both richness and evenness and the distribution of abundance among distinct taxa) was expressed using the Shannon diversity index. The Shannon diversity index (H=) is calculated using the equation where R is richness and p i is the relative abundance of the ith OTU.
Measures of diversity were screened for group differences by using analysis of variance (ANOVA). Multivariate differences among groups were evaluated with permutational multivariate analysis of variance using the distance matrices function in Adonis (35). For Adonis, distances between samples were calculated using unweighted or weighted UniFrac, and then an ANOVA-like simulation was conducted to test for group differences. In addition, multivariate differences were assessed using the analysis of similarities function in anosim (35). Sequence data were processed using a Research and Testing pipeline that is described at http: //www.researchandtesting.com/docs/Data_Analysis_Methodology .pdf. Spearman correlations between OTU and metabolite concentration were computed. All analyses were conducted in R, using the vegan (35), labdsv (36), DESeq2 (37), and phyloseq (38) packages.
Community-level catabolic profiles. Carbon source utilization patterns of the fecal microbiota were assessed using Biolog 96-well Eco microplates (Biolog, Inc., Hayward, CA, USA) (39). The microplates contained 31 different carbon sources (carbohydrates, carboxylic acids, polymers, amino acids, amines, and miscellaneous substrates) in triplicate. Five grams of feces diluted with Amies transport medium (1:1) was homogenized in a filter bag with 45 ml of 0.9% sterile sodium chloride (wt/vol) solution (Classic Blender) to remove fecal solid particulate. The homogenized feces were centrifuged at 11,000 rpm for 15 min at 4°C. The pellet was washed with 50 mM Tris-HCl (pH 7.0) and then with sterile sodium chloride solution and centrifuged at 11,000 rpm for 15 min at 4°C. The cellular suspension was diluted (1:10) into the sterile sodium chloride solution and subsequently centrifuged at 2,000 rpm for 2 min at 4°C. The cellular suspension was diluted (1:20) into sterile chloride solution and dispensed (150 l) into each of the 96 wells of the Biolog Eco microplates. The microplates were incubated at 37°C in the dark under anaerobic conditions and with slow stirring. The color development was measured at 590 nm every 24 h with a microplate reader (MicroStation; Biolog, Inc.). Three indices were determined (40). Shannon's diversity (H=), indicating the substrate utilization pattern, was calculated as follows: H= ϭ Ϫ ⌺ p i ϫ ln(p i ), where p i is the ratio of the activity of a particular substrate to the sums of the activities of all substrates at 120 h. Substrate richness (S), measuring the number of different substrates used, was calculated as the number of wells with a corrected absorbance of Ͼ0. 25 Fecal concentration of free amino acids. FAA of fecal samples were analyzed through the Biochrom 30 series amino acid analyzer (Biochrom Ltd., Cambridge Science Park, England) with a sodium cation-exchange column (20 cm long by 0.46 cm inner diameter). A mixture of amino acids at known concentrations (Sigma Chemical Co., Milan, Italy) was added to cysteic acid, methionine sulfoxide, methionine sulfone, tryptophan, ornithine, glutamic acid, and ␥-aminobutyric acid and used as the standard. Proteins and peptides in the fecal samples were precipitated by adding 5% (vol/vol) cold solid sulfosalicylic acid, holding the samples at 4°C for 1 h, and centrifuging at 15,000 ϫ g for 15 min. The supernatant was filtered through a 0.22-m-pore-size filter and diluted when necessary with sodium citrate loading buffer (0.2 M [pH 2.2]). Amino acids were postcolumn derivatized with ninhydrin reagent and detected by absorbance at 440 nm (for proline and hydroxyproline) or 570 nm (for all other amino acids).
Gas chromatography-mass spectrometry solid-phase microextraction analysis of fecal volatile compounds. After preconditioning, according to the manufacturer's instructions, a polydimethylsiloxane/divinylbenzene fiber (65 m) and a manual solid-phase microextraction (SPME) holder (Supelco, Inc., Bellefonte, PA, USA) were used. Before headspace sampling, the fiber was exposed to a gas chromatography (GC) inlet for 1 h for thermal desorption at 250°C. Three grams of fecal sample was placed into 10-ml glass vials and added to 10 l of 4-methyl-2-pentanol (final concentration, 33 mg/liter) as the internal standard. The samples were then equilibrated for 10 min at 40°C. SPME fiber was exposed to each sample for 40 min. Both the equilibration and absorption phases were carried out with stirring. The fiber was then inserted into the injection port of the gas chromatograph for 10 min of sample desorption. Gas chromatography-mass spectrometry (GC-MS) analyses were carried out with an Agilent 7890A gas chromatograph (Agilent Technologies, Palo Alto, CA) coupled to an Agilent 5975C mass selective detector operating in electron impact mode (ionization voltage, 70 eV). A Supelcowax 10capillary column (length, 60 m; inside diameter, 0.32 mm; Supelco, Inc.) was used. The temperature program was 50°C for 1 min, followed by an increase at a rate of 4.5°C/min to 65°C, an increase at a rate of 10°C/min to 230°C, and then 230°C for 25 min. The injector, interface, and ion source temperatures were 250, 250, and 230°C, respectively. The mass-to-charge ratio interval was 30 to 350 Da at a rate of 2.9 scans per s. Injection was carried out in splitless mode, and helium (flow rate, 1 ml/min) was used as the carrier gas. Molecules were identified based on a comparison of their retention times with those of pure compounds (Sigma-Aldrich, Milan, Italy). Identities were confirmed by searching mass spectra in the available databases (NIST version 2005 and Wiley version 1996). Quantitative data for the identified compounds were obtained by interpolation of the relative areas versus the internal standard area (42). All GC-MS/SPME raw files were converted to netCDF format via ChemStation (Agilent Technologies, USA) and subsequently processed by the XCMS toolbox (https: //xcmsonline.scripps.edu/). XCMS software allows automatic and simultaneous retention time alignment, matched filtration, peak detection, and peak matching. GC-MS/SPME data were organized into a matrix for subsequent statistical analysis.
Statistical analysis. Culture-dependent data were obtained at least in triplicate. The analysis of variance (ANOVA) was carried out on transformed data, followed by the separation of means with Tukey's honestly significant difference (HSD) test, using the statistical software Statistica 6.0 for Windows 1998 (StatSoft, Vigonza, Italy). In order to identify differences between HS and HSB samples, GC-MS/SPME data were analyzed by canonical discriminant analysis of principal coordinates (CAP) (41). The total variance obtained in the principal coordinates used to perform the CAP was Ͻ80% for all samples. Significance testing was carried out using 999 permutations. The correctly performed permutation test assigns ca. 90% of the samples. Moreover, models based on projection to latent structures (PLS) in its discriminant (DA) version were built based on the normalized concentration of the significant molecules that were identified (43). Metabolically active bacterial families/genera and fecal volatile compounds were analyzed by principal component analysis (PCA) using the statistical software Statistica 6.0 for Windows 1998 (Stat-Soft).
Nucleotide sequence accession number. The sequence data were submitted to the Sequence Read Archive database of the National Center for Biotechnology Information under accession no. PRJNA290897.

Diet and clinical evaluation.
In this study, the diet of the volunteers was monitored for 2 months before the administration of durum wheat flour and whole-grain barley pasta containing the minimum recommended intake (3 g) of ␤-glucans. This was done to ascertain that every volunteer consumed equivalent amounts of carbohydrates, total protein, and fat. The only statistically signif- icant (P Ͻ 0.05) difference in the diet, before and after administration of durum wheat flour and whole-grain barley pasta, was the amount of total fiber (Table 3). Barley flour, including that with ␤-glucans, was well tolerated by all volunteers, and no statistically significant (P Ͼ 0.05) effects regarding bloating, abdominal pain score, flatulence, or frequency or consistency of evacuations of feces were observed (see Table S1 in the supplemental material). The observed reduction in non-HDL cholesterol confirmed the European Food Safety Authority healthy claim (3). The estimation of cholesterol amount was used as an internal control to validate the adherence of each volunteer to the diet and, consequently, the consistency of the study.
Richness and diversity of the fecal microbiota based on 16S rRNA gene sequencing data analysis. Total metabolically active bacteria from fecal samples of healthy subjects before (HS) and after (HSB) the administration of durum wheat flour and whole-grain barley pasta were analyzed by pyrosequencing of the 16S rRNA gene. After quality control, pyrosequencing analysis yielded 156,563 total bacterial read sequences, with a mean of 3,010 reads (range, 1,146 to 5,621 reads) per sample. The   index, respectively) between HS and HSB samples (see Fig. S2 in the supplemental material). According to the alpha-diversity values, the three phylogenybased beta-diversity measures did not show a clear separation between the microbiota compositions of HS and HSB samples in weighted and unweighted UniFrac distance principal coordinate analysis plots ( Fig. 1; see also Fig. S3 in the supplemental material). Also, the Adonis statistical test indicated no significant (P ϭ 0.199) differences between the microbial diversities of HS and HSB samples. However, anosim results, which were based on weighted UniFrac testing for multivariate differences, differed (P ϭ 0.04; TestStat, 0.05) between HS and HSB samples. Overall, seven phyla (Actinobacteria, Bacteroidetes, Firmicutes, Fusobacteria, Lentisphaerae, Proteobacteria, and Verrucomicrobia) and one candidate division (TM7) were identified (see Fig. S4 in the supplemental material). However, Firmicutes and Bacteroidetes represented Ͼ85% of all 16S rRNA sequences. No significant (P Ͼ 0.05) differences were found for the relative abundances of the phyla between HS and HSB samples. The only exception was the phylum Fusobacteria, which was higher in HS than in HSB samples (average relative abundance, 3.06 versus 1.85%, respectively; P ϭ 0.019).
According to alpha and beta diversity results and considering the 35 most dominant OTUs (representing 95% of the total microbiota) in all fecal samples, HS and HSB volunteers were variously distributed (Fig. 2). Few statistically significant (P Ͻ 0.05) differences were detected between the fecal microbiota of HS samples and that of HSB samples (Table 4) OTU correlation. OTU correlations were investigated considering family-level ( Fig. 3A and B) and genus-level ( Fig. 3C and D) taxonomic assignments, with significant correlations considered at a false discovery rate (FDR) of Ͻ0.050. Several family-related positive correlations were always found (e.g., Coriobacteriaceae with Streptococcaceae and Porphyromonadaceae with Clostridiaceae). Other positive correlations found in HS samples (e.g., Eubacteriaceae with Coriobacteriaceae, Oscillospiraceae, and Streptococcaceae) were not confirmed in HSB samples. A similar trend was detected also at the genus level. The only exception was for Bacteroides, which showed an increased number of positive correlations in HSB samples compared to that in HS samples.
Community-level catabolic profiles and cultivable bacteria. The substrate utilization pattern (H= index) and substrate richness (S index) values were calculated (Fig. 4). The H= and S indices of the HSB samples were lower than those of the HS samples. The E index, which measures the statistical significance (equitability) of the H= and S index values, confirmed the above-described significant (P Ͻ 0.05) differences.
Selective media were used to enumerate cultivable bacteria (Table 5). HSB samples had fewer total anaerobes than HS samples (P ϭ 0.05). The median numbers of presumptive Lactobacillus in HS samples were lower than those in HSB samples. Other significant (P Ͻ 0.05) differences concerned the numbers of Enterobacteriaceae, total coliforms, and presumptive Bacteroides, Porphyromonas, Prevotella, Pseudomonas, Alcaligenes, and Aeromonas bacteria, which were the lowest in the HSB fecal samples.
Fecal metabolome. Fecal samples from HSB had lower (P Ͻ 0.05) levels of some FAA (Pro, Trp, Thr, and Arg) and metabolites from the catabolism of FAA (e.g., ␥-aminobutyric acid)  (Fig. 5B). Pentanoic acid was the only SCFA found at the highest level in HS samples. The other statistically significant differences for VOC regarded the levels of phenylethyl alcohol, benzaldehyde, indole, 2,3-butanedione, 6-methyl-5-hepten-2-one, and acetophenone, which were the highest in HSB samples. FAA and GC-MS/SPME data were also analyzed using the principal component analysis (PCA). The discrimination of fecal samples between HS and HSB was evident (see Fig. S7 in the supplemental material).

DISCUSSION
This is one of the few studies showing the effects of the consumption of durum wheat flour and whole-grain barley pasta on the human fecal microbiota and metabolome, using a metaomics approach. As shown through pyrosequencing analysis, the intervention with durum flour wheat and whole-grain barley pasta, resulting in the ingestion of 3 g/day of ␤-glucans, did not affect the values of alpha and beta diversity. Previously, high-throughput sequencing techniques revealed that alpha diversity decreased with dietary supplementation of ␤-glucans from sea cucumber (Apostichopus japonicus) (44) and mirror carp (Cyprinus carpio) (45). The phylogenetic composition of the analyzed samples confirms that the Firmicutes and Bacteroidetes phyla constitute the most abundant bacterial OTUs of human intestinal microbiota. The composition of the main bacterial phyla (Firmicutes and Bacteroidetes) within the enrolled volunteers was variously affected, without a unique statistically significant trend. According to the 16S rRNA gene-based high-throughput sequencing, a wide variation was found between individuals (46,47). Levels of Clostridi-  aceae (Clostridium orbiscindens and Clostridium sp.), Roseburia hominis, and Ruminococcus sp. increased following the intervention. Previously, it was shown that barley and oat ␤-glucans induced clostridial cluster IX populations and the Clostridium histolyticum subgroup during in vitro fermentation by human fecal microbiota (48). On the contrary, oat ␤-glucans did not favor the growth of Clostridium sp. in SHIME and C57BL/6J mouse models (21,24,49). However, several discrepancies between the different studies might be due to the different models and methods used. Levels of F. prausnitzii, Faecalibacterium sp., and D. invisus decreased following the diet intervention with barley ␤-glucans. A similar trend was found for members of the genus Fusobacterium, belonging to Fusobacteria. Based on OTU correlations, ␤-glucans negatively impacted bacterial interactions. The positive bacterial interactions also decreased during dietary supplementation with ␤-glucans from sea cucumber (44). In agreement, the community-level catabolic profiles showed decreased substrate utilization pattern (H= index) and Shannon index after diet intervention with barley ␤-glucans. All these results indicated low metabolic diversity. Culture-dependent methods showed that diet intervention with whole-grain barley markedly decreased the total number of fecal anaerobic cultivable bacteria. Decreased levels of cultivable anaerobes were also found during in vitro fermentation of oat ␤-glucans by human fecal slurry (20). According to the in vitro study conducted by Hughes et al. (48), levels of cultivable presumptive thermophilic and mesophilic Lactobacillus spp. increased following the diet intervention with barley ␤-glucans. No positive effects were observed regarding organisms of the genus Bifidobacterium. A randomized, placebo-controlled, and double-blind human intervention trial showed that cultivable fecal lactobacilli isolated from 26 healthy volunteers were not affected by 0.75 g of barley ␤-glucans (26). The same study showed a statistically significant increase in the cell density of bifidobacteria. After grouping volunteers according to age, a significant bifidogenic effect was detected only in subjects Ͼ50 years. The administration of 3 g/day of ␤-glucans to 33 polypectomized patients showed no effects on the fecal cultivable lactobacilli and bifidobacteria (27). An increased level of lactobacilli was found in animal models after a barley-or oat-containing diet, especially using high-viscosity ␤-glucans (22,23,24). At the same time, an oat product-based diet or ␤-glucans from Laminaria digitata, Laminaria hyperborea, and Saccharomyces cerevisiae did not increase the levels of lactobacilli in pigs and rats (25,50). Taken together, these results suggested that the prebiotic potential of whole-grain barley/␤-glucan was not always reproducible in vivo.
The complex biochemical interactions and antagonistic activities within the intestinal microbiota prior to treatment might be responsible for the different responses to the ␤-glucan-enriched diet. Levels of presumptive cultivable Bacteroides, Porphyromonas, and Prevotella spp. decreased following the diet intervention with barley ␤-glucans. Overall, Prevotella abundance is higher in humans who consume more of a vegetable-based diet (14). During in vitro fermentation, Prevotella and Bacteroides levels increased with low-molecular-weight ␤-glucans only (48). Levels of cultivable Bacteroides sp. decreased in the SHIME model after oat bran feeding (21), while no significant differences were found in rats after feeding with products enriched with oat fiber (50). A barley-rich diet was associated with a reduced level of cultivable Bacteroides sp. in the intestinal tract of rats (22). According to this study, cultivable Bacteroides was negatively affected in polypectomized patients after intake of barley ␤-glucans (27). In vitro and animal trials support that Enterobacteriaceae could not utilize ␤-glucans (20,25,50). According to in vitro and animal trials, this study showed a significant reduction in cultivable Enterobacteriaceae and total coliforms. A similar trend was found in the administration of 3 g/day of ␤-glucans to 33 polypectomized patients (27). On the contrary, small amounts of ␤-glucans (0.75 g) were not effective in decreasing the level of total cultivable coliforms or E. coli (26). First, this study showed that the cell density of cultivable Pseudomonas, Alcaligenes, and Aeromonas markedly decreased after the diet intervention with pasta enriched with barley ␤-glucans.
In vitro and animal trials support that high barley ␤-glucan consumption is associated with a significant increase in the synthesis of SCFA (46,49,50). This study confirmed the previous GC-MS data showing a noticeable increase in 2-methyl-propanoic acid, acetic acid, butanoic (butyric) acid, and propanoic (propionic) acid after the diet intervention with durum wheat flour and whole-grain barley pasta. SCFA induce positive gastrointestinal and systemic effects (51). SCFA are also important modulators of the host immune function (52). Acetate is used as an energy source for liver and peripheral tissues and acts as a signaling molecule in gluconeogenesis and lipogenesis (53). Propionate serves as a precursor for gluconeogenesis and reduces the synthesis of hepatic cholesterol (54). Propionate was also associated with a decrease in insulin secretion in pancreatic islet cells of rats (55). Butyric acid serves as the main energy source for colonocytes and protects against inflammation (56,57). Butyrate also affects the regulation of apoptosis and cellular proliferation, resulting in a reduced risk of colon cancer (58). BCFA (e.g., isobutyrate and isovalerate) were not affected by whole-grain barley, indicating that the greatest synthesis of SCFA was obtained without increasing undesirable protein fermentation (44). Previously, it was demonstrated that the intake of dietary fibers or symbiotic foods (e.g., fructooligosaccharides, Lactobacillus helveticus, and Bifidobacterium longum) leads to a modulation of the gut metabolic activity, with an increase of SCFA (59)(60)(61). Previously, in vitro data showed that barley and oat ␤-glucans or other oligosaccharides support the growth of Clostridia strains, which synthesize acetate and butyrate (62,63). Bacteroides strains are propionate producers via the succinate pathway (64), and they were positively correlated with propionic acid.
This study highlights some in vivo effects of whole-grain barley on the fecal microbiota and metabolome. Whole-grain barley appeared to be effective in the modulation of the composition and the metabolic pathways of the intestinal microbiota, leading to an increased level of SCFA. We thank Maria Piccolo (Department of Soil, Plant and Food Sciences, University of Bari Aldo Moro, Bari, Italy) and Giuseppe Dalfino and Stella Diamante (DETO, Nephrology Unit, University of Bari Aldo Moro, Bari, Italy) for technical assistance.

ACKNOWLEDGMENTS
We declare no conflicts of interest.