Inclusion of Chicory (Cichorium intybus L.) in Pigs' Diets Affects the Intestinal Microenvironment and the Gut Microbiota

Abstract

The content and composition of prebiotic plant fiber in the diet is important in promoting gut-related health. This study investigated the effects of the dietary inclusion of chicory forage and roots on the intestinal microenvironment of pigs. Thirty-seven-week-old pigs were fed 1 of 5 diets for 18 days, including a cereal-based control diet and 4 diets with the inclusion of 80 and 160 g kg⁻¹ of body weight chicory forage (CF80 and CF160), 80 g kg⁻¹ chicory root (CR80), and a mix of 80 g kg⁻¹ forage and 80 g kg⁻¹ chicory root (CFR). The animals maintained good performance and health irrespective of diet. Bacterial community structure and diversity in ileal and colonic samples was assessed using terminal restriction fragment length polymorphism (T-RFLP), combined with cloning and sequencing. Samples clustered perfectly according to gut segment with a higher bacterial diversity in colon than ileum. Distal ileum was dominated by lactic acid bacteria (LAB), and the relative amount of this group was increased by the CF160 and CFR diets. The colonic bacterial community was dominated by butyrate-producing bacteria and Prevotella. The increased relative abundance of butyrate-producing bacteria in the colon was positively correlated with the molar proportion of acetic acid and furthermore linked to the chicory forage diets (CF80 and CF160). Diets including chicory roots (CR80 and CFR) were correlated with a higher colonic abundance of Megasphaera elsdenii. The fermentation products and pH in digesta responded to diet type and were correlated with shifts in the microbiota, showing that chicory influences the intestinal microenvironment of pigs.

INTRODUCTION

The application of 16S rRNA gene-based analyses has revealed a tremendously high diversity of gut microbiota at bacterial species and phylotype levels (22). During the past decade, there has been an intense focus on microbial ecology, with the aim of understanding the links between the gut microbiota, health, and diseases. However, compared to human subjects, limited information is available on the porcine gut, and more than 80% of the bacterial phylotypes identified may represent unknown species (21).

Pigs are known to have naturally occurring and economically important enteric diseases, such as swine dysentery and postweaning diarrhea (9, 24). Outbreaks of bacterial diseases in the gut can impose significant constraints on pig production by reducing animal welfare and productivity. The contaminated products can also pose threats to the human food chain (9). To uncover the effect of the gut microbiota upon pig performance and health, the dominant bacterial species that colonize the gut must be identified, and the extent to which these are influenced by diet composition must be determined.

Certain substrates, such as dietary fiber, escape digestion in the foregut and reach the hindgut of animals and are now a central issue in nutrition application. These substrates, commonly referred to as prebiotics, can alter gut bacterial composition, modify intestinal fermentation processes, promote gut development, and possibly improve host health (9). Chicory (Cichorium intybus L.) is a perennial herb that can be used as a fiber source in pig diets (17). Chicory root, which has a high content of inulin-type fructan and oligofructose, is a prototype prebiotic for monogastrics (32). Previous studies have reported that inulin-type fructan and oligofructose can promote the growth of beneficial microbes, such as bifidobacteria and lactic acid bacteria (LAB), along the gastrointestinal (GI) tract (26, 33). In contrast, little is known about the effect of chicory forage inclusion upon the composition of porcine gut microbiota. Chicory forage has a high content of uronic acids, which in dicotyledonous plants derive from galactouronic acid and are the building blocks of pectin (34). Plant-origin pectin has been used as a soluble fiber feed ingredient in pig diets previously. For example, feeding pigs fermentable carbohydrates, including sugar beet pulp pectin, is reported to increase colon microbiota diversity and Lactobacillus amylovorus population (20). Diets with a high content of sugar beet pulp pectin resulted in increased numbers of lactobacilli in the small intestine and short-chain fatty acid (SCFA) products in the colon of weaned piglets (6), whereas apple pectin diets did not affect pig small intestine motility but altered intestinal morphology (7). In addition, a recent study from our group showed that different levels of chicory forage inclusion increased the digestibility of uronic acids and maintained pig performance compared to a cereal-based control diet (17).

This study investigates the effects of changes in the composition of the fiber fraction in a cereal-based diet on gut environment, morphology, and intestinal microbiota in pigs. Our hypothesis was that the inclusion of chicory forage and chicory root in a cereal-based diet would modify the intestinal microbiota, and that these changes would be associated with changes in gut morphology, fermentation pattern, and pH in ileal and colonic digesta of pigs.

MATERIALS AND METHODS

Animals and experimental design.

Thirty-seven-week-old Yorkshire piglets of mixed gender (initial body weight [BW], 15.5 ± 1.6 kg) were selected from 6 different litters and used in an 18-day experiment. The experiment had a split-litter design, where one pig per litter was assigned to 1 of 5 diets, resulting in 6 observations per dietary treatment. Piglets had ad libitum access to feed and water throughout the experiment. They were weighed on days 0, 7, 14, and 18, and feed intake was recorded daily. The study was performed at the Swedish University of Agricultural Sciences (SLU) in Uppsala and was approved by the ethical committee for the Uppsala region.

Housing and diets.

The piglets were housed individually in pens fitted with rubber matting and had no access to straw during the experiment. Room temperature was maintained at 20 ± 1°C, with an additional heat lamp in each pen to increase piglet comfort during the first week of the experiment. The following 5 diets were studied in the experiment: a cereal-based control diet (C), two diets with 80 or 160 g kg⁻¹ of body weight chicory forage meal (CF80 and CF160, respectively), one diet with 80 g kg⁻¹ of chicory root meal (CR80) (Inu60; Inter-Harz Gmbh, Germany), and one diet with a mix of 80 g kg⁻¹ chicory forage meal and 80 g kg⁻¹ chicory root meal (CFR). All diets were formulated to meet the nutritional requirements of piglets (25). Diet C was composed of cereals (wheat and barley) and supplemented with protein, amino acids, minerals, and vitamins. In diets with chicory forage and root meal inclusion, the cereal mixture was substituted on an air-dry basis (dried ingredients with a moisture content of 10 to 12%). The diet ingredients are shown in Table S1 in the supplemental material, together with chemical composition and gross energy analyzed as previously described (18). No antibiotics were administered to the pigs during the experiment.

Sampling.

On the last day of the experiment, pigs were sedated using a mix of Domitor (1 mg ml⁻¹ medetomidine HCl; Orion Pharma, Espoo, Finland) and Zoletil (25 mg ml⁻¹ tiletamine and 25 mg ml⁻¹ zolazepam; Vibrac S.A., Carros, France) in a dose of 0.05 ml kg⁻¹ BW and killed by a lethal dose of pentobarbital sodium (60 mg ml⁻¹; Apoteket, Umeå, Sweden) at 100 mg kg⁻¹ BW. The abdominal cavity was opened and the entire GI tract removed. For morphological analysis, one segment of tissue (3 cm) from the distal ileum (50 cm cranial to the ileocecal valve) and one segment from the proximal colon (20 cm from the cecum) were collected. The pH of digesta was measured (PHM 210; Radiometer, Cedex, France), and intestinal contents were sampled. One sample was stored at −80°C for terminal restriction fragment length polymorphism (T-RFLP) analysis, and another sample was stored at −20°C for the analysis of organic acids. Fecal samples were collected as previously described (18).

DNA extraction and PCR conditions.

DNA was extracted from 200-mg digesta samples from either ileum or colon using a QIAamp DNA Stool Minikit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. To ensure the proper lysis of bacteria, the initial heating step was supplemented with 3 cycles of 5 min of heating (at 95°C) and snap-freezing (in liquid nitrogen). Bacterial 16S rRNA genes were amplified from each DNA extraction (two technical replicates per extract) with general primers Bact-8F (5′AGAGTTTGATCCTGGCTCAG-3′; 5′ end labeled with 6-carboxyfluorescein) and 926r (5′-CCGTCAATTCCTTTRAGTTT-3′). The PCRs were run under conditions described elsewhere (11).

T-RFLP analysis.

The PCR products were digested with the restriction enzyme HaeIII (GE Healthcare, Uppsala, Sweden), the digested products were separated on a capillary sequencer (ABI 3730), and T-RFLP profiles were processed using Peak Scanner v1.0. The relative peak area of each terminal restriction fragment (TRF) was calculated by dividing individual peak area by total peak area within the following size limits: lower limit, 50 bp; upper limit, 500 bp. TRFs with a relative abundance of ≤0.5% were excluded from the remaining analysis, while from the replicate T-RFLP profiles only peaks that occurred in both replicates were included in the analysis.

Cloning and sequencing.

To identify bacteria that corresponded to TRFs of interest, eight clone libraries from digesta and fecal samples of pigs were constructed according to Dicksved et al. (11). A total of 48 inserts from each library were PCR amplified using vector primers M13f and M13r, and the resulting PCR products were analyzed by T-RFLP analysis as described above. Clones with TRF sizes of interest were chosen for sequencing, giving a total of 144 different clones. The sequences then were aligned to the GenBank database using standard nucleotide BLAST at NCBI (http://www.ncbi.nlm.nih.gov). Pintail v1.0 was used to identify chimeric sequences which were subsequently removed from further analyses.

Organic acid analysis.

The digesta samples were centrifuged for 5 min at 13,000 × g. A volume of 200 μl of the clear solution was mixed with 800 μl of the internal standard solution (0.1% of pivalic acid in 5 mM sulfuric acid). The analysis of organic acids (SCFA and lactic acid) in the samples was performed according to Andersson and Hedlund (1) by a high-performance liquid chromatography (HPLC) system consisting of the following: Alliance 2795 Separations module with a temperature control module II range of 40 to 70°C, a 2414 RI detector (Waters Association), a separation column (ReproGel H; 9-μm volume; 300 by 8 mm) and a precolumn packet (ReproGel H; 9-μm volume; 30 by 8 mm; Dr. Maisch HPLC GmbH, Ammerbuch, Germany). The mobile phase was 0.005 M sulfuric acid at a flow rate of 0.8 ml min⁻¹, and the column temperature was kept at 60°C.

Morphological examination.

Samples for morphological evaluation were rinsed in phosphate-buffered saline (0.01 M, pH 7.4) to remove digesta and placed in 4% phosphate-buffered paraformaldehyde fixative (1/15 M, pH 7.2) for 48 h. Segments then were trimmed and embedded in paraffin according to standard procedures. Embedded tissue samples were cut into pieces of 4-μm thickness and stained with hematoxylin and eosin, and one slide from each segment of an individual pig was examined (6, 23) using a transmitted light microscope (Carl Zeiss, Germany) equipped with a Canon A640 digital camera using image analysis software (AxioVision release 4.8.2).

To avoid subjective bias, the slides were blindly coded and all measurements were performed by one individual. Villus height, crypt depth, and the thickness of muscularis externa at counterpart locations were evaluated as previously described (23) on 10 well-oriented villus and crypt preparations on each slide from ileum samples. Measurements were made only where the villus-crypt unit was perpendicular to the muscularis mucosa. Villus height was depicted as the distance from the top of the villus to the villus-crypt junction. Total mucosal thickness was measured from the top of the villus to the border over the muscularis mucosa. The depth of crypt was then defined as the difference between total mucosal thickness and villus height. The thickness of the muscularis externa, which consists of an inner circular layer and a longitudinal outer muscular layer, was measured. For colon, the thickness of the mucosa and the muscularis externa were measured.

Statistical analysis.

Simpson's index of diversity was calculated based on T-RFLP data to assess bacterial species richness and evenness (4). Data analysis to identify links between the relative abundance of TRFs and diet type only included TRFs that were present in more than 3 pigs.

Diet response in terms of performance, gut morphology, pH value in digesta, fermentation products, and microbiota was assessed using the procedure Mixed in SAS (version 9.1; SAS Institute, Cary, NC). The model included diet (C, CF80, CF160, CR80, and CFR) and sex as fixed factors and litter as a random factor. Two-way interactions between treatment and segment were tested but were not statistically significant, therefore they were excluded from the model. For morphological evaluation, means of 10 villus-crypt unit values from each pig were used; the model included individual pig as a random factor.

Cluster analysis of TRF data was performed using Spearman rank correlation, and the heat map was visualized using MultiExperiment Viewer (MeV 4.5.1). To examine the relationships between variables, Pearson correlations of fermentation products and specific bacterial group within animal were estimated with PROC CORR in SAS (SAS Institute, Cary, NC), and the Pearson correlation (r) of the means of each measurement was calculated. Significance was set at P < 0.05.

Nucleotide sequence accession numbers.

Unique DNA sequences were deposited in GenBank at NCBI under the following accession numbers: JF794786 to JF794852.

RESULTS

Growth performance and gut morphology.

The pigs remained healthy throughout the experiment. The daily feed intake (DFI), daily weight gain (DWG), and feed conversion ratio (FCR) did not differ between diets. Details of growth performance are shown in Table S2 in the supplemental material.

The gut morphology of various regions was affected by diet (Table 1). In ileum, shorter villus height was observed with diet CF160 than with diets C and CR80 (P = 0.03), whereas no diet effect was seen for crypt depth, the villus/crypt ratio, or the thickness of muscularis externa. In colon, the muscularis externa was thicker with diet CF80 than with diets C, CF160, and CR80 (P = 0.04), whereas animals on diet CFR had thicker muscularis externa than those on diet CF160. Mucosal thickness in colon was not affected by diet.

Table 1.

Impact of diet on morphology parameters in ileum and colon of pigs

Sample source and parameter	Dieta					SE	P value
	Control	CF80	CF160	CR80	CFR
Ileum
Villus height (μm)	614 a	562 ab	550 b	633 a	586 ab	18.9	0.03
Crypt depth (μm)	235	195	211	229	215	20.8	0.7
V:Cb	2.9	3.1	3.0	3.0	3.0	0.32	1.0
Thickness of muscularis externa (μm)	518	442	393	537	454	51.4	0.3
Colon
Mucosal thickness (μm)	538	545	521	561	538	22.9	0.8
Thickness of muscularis externa (μm)	406 ac	545 b	348 a	361 ac	508 bc	50.0	0.04

^aCF80, 80 g kg⁻¹ chicory forage; CF160, 160 g kg⁻¹ chicory forage; CR80, 80 g kg⁻¹ chicory root meal (Inu60; Inter-Harz Gmbh, Germany); CFR, mix of 80 g kg⁻¹ chicory forage and root meal. Data are presented as least-square means ± SE (n = 6). Different letters within rows indicate significant difference (P < 0.05).

^bV:C, ratio of villus height to crypt depth.

Organic acids and pH in the gut.

Diet modulated fermentation in both small and large intestine (Fig. 1; also see Table S3 in the supplemental material). In ileum, the lactic acid concentration was higher with diet CR80 than with diets C and CF80 (P = 0.05) (Fig. 1A), and no differences were seen for the other diets. In contrast, a lower total SCFA concentration was found with diets CR80 and CFR than with diets C and CF80 (P = 0.047). This was primarily due to changes in acetic acid concentration (P = 0.045) (Fig. 1B). The pH value in ileal digesta was not affected by diet (see Table S3).

Fig 1.

Effect of diet on the concentration of organic acids (mmol liter⁻¹) in digesta of pigs. (A) Concentration of lactic acid (mmol liter⁻¹) in ileum; (B) concentration of short-chain fatty acids (SCFA mmol liter⁻¹) in ileum; (C) concentration of SCFA (mmol liter⁻¹) in colon. Data are presented as least-square means ± standard errors (SE) (n = 5 in ileum; n = 6 in colon). Different superscript letters indicate significant difference (P < 0.05). CF80, 80 g kg⁻¹ chicory forage; CF160, 160 g kg⁻¹ chicory forage; CR80, 80 g kg⁻¹ chicory root meal (Inu60; Inter-Harz Gmbh, Germany); CFR, mix of 80 g kg⁻¹ chicory forage and root meal.

In colon, no dietary response for total SCFA concentration was found (see Table S3 in the supplemental material). Acetic acid concentration was higher for diets CF80 and CF160 than diets CR80 and CFR (P = 0.03) (Fig. 1C), whereas diet C had intermediate values. The opposite pattern was seen for n-valeric acid, with significantly higher concentration with diets CR80 and CFR than with diets CF80 and CF160 (P = 0.003) (Fig. 1C). The changes in ratios of acetic acid (P = 0.02) and n-valeric acid (P = 0.0001) relative to SCFA followed the same pattern as that of changes in acid concentration. Colonic pH was lower with the CFR diet than with all other diets except CF80 (P = 0.05) (see Table S3).

Gut bacterial community structure and phylotypes.

T-RFLP profiles from ileal and colon digesta samples were analyzed and yielded a total of 122 distinct TRFs. Some TRFs, such as TRF85 and TRF331 in ileum and TRF262 and TRF414 in colon, had a universal distribution and were present in all pigs on all diets (Fig. 2). Cluster analysis combined with a heat map was used to get an overview of the bacterial community structure in the samples. Two distinct clusters were generated (Fig. 2), which were perfectly matched with the gut segment. Despite interindividual variation, all ileal samples clustered together and were clearly separated from colonic samples. These 2 clusters had different predominant TRFs, with TRF331 being the most abundant for the ileum cluster and TRF264 being the most abundant in colon. The richness and evenness of the ileal and colonic microbiota was estimated based on T-RFLP data using Simpson's diversity index. No difference was found between diets. However, the colonic microbiota had a much higher diversity (15.2 ± 0.89) than the ileal microbiota (8.5 ± 0.89) (P < 0.0001).

Fig 2.

Heat map showing the relative abundance of terminal restriction fragments (TRFs) of microbiota in digesta from ileum and colon in pigs. The abundance intervals are presented with a progressive color scale at the bottom of the figure, where 0.0 is the lowest and 0.5 is the highest abundance. Each row represents a TRF size, and each column corresponds to one pig. The samples were clustered into 2 main groups using Spearman rank correlation, with green branches representing ileum samples and blue branches representing colon samples. TRF331 is the most abundant phylotype in ileal microbiota; TRF264 is the most abundant phylotype in colonic microbiota; TRF85 and TRF331 in the ileum cluster and TRF262 and TRF414 in the colon cluster had universal appearances.

Clone libraries were created to identify the dominant TRFs and to further define the composition of the microbiota. A total of 384 clones from digesta and fecal samples of pigs were generated, and 144 of those with TRF sizes of interest were sequenced. These clones were matched with bacteria belonging to the 3 phyla Firmicutes, Bacteroidetes, and Actinobacteria. Members of the Firmicutes were highly diverse and were dominated by Lactobacillus spp. and bacteria that fall into clostridial cluster XIVa and cluster IV, while the phylum Bacteroidetes was mainly represented by different species of Prevotella. In addition, Bifidobacterium sp., a member of the phylum Actinobacteria, was found in both ileal and colonic samples but in low abundance.

Diet-responsive bacteria in ileum.

Although individual variation was evident, we were able to identify diet-associated changes in specific bacterial groups. The accumulated relative abundance of TRFs identified as LAB (TRF62 to TRF64, TRF244, TRF253, and TRF331) was higher (P = 0.049) (Fig. 3A) with diets CF160 and CFR than with diets C and CF80. Moreover, butyrate-producing bacteria comprising bacteria that fall into clostridial cluster XIVa and cluster IV were grouped for their major role as lactate and butyrate producers in the gut, and no diet-related difference was seen in the relative abundance of the TRFs (TRF223, TRF258, TRF274, TRF317, and TRF318) (Fig. 3B). However, there was a marked increase (P = 0.01) in the relative abundance of TRFs identified as Prevotella spp. (TRF130, TRF161, TRF163, TRF264, and TRF411) with diet CFR compared to that for the other diets (Fig. 3C).

Fig 3.

Effect of diet on the relative abundance (percent) of phylotypes corresponding to dominant groups of bacteria in digesta of pigs. (A) Relative abundances (percent) of phylotypes corresponding to the sum of lactic acid bacteria (LAB). (B) Relative abundances (percent) of phylotypes corresponding to the sum of butyrate-producing bacteria. (C) Relative abundances (percent) of phylotypes corresponding to the sum of Prevotella spp. Values are means ± SE; n = 5 in ileum, n = 6 in colon. Within a segment, different letters indicate statistical difference (P < 0.05). CF80, 80 g kg⁻¹ chicory forage; CF160, 160 g kg⁻¹ chicory forage; CR80, 80 g kg⁻¹ chicory root meal (Inu60; Inter-Harz Gmbh, Germany); CFR, mix of 80 g kg⁻¹ chicory forage and root meal.

The abundance of specific TRFs in ileum was also found to be correlated with diet type (Table 2). Diet CFR was especially correlated with higher relative abundances of Lactobacillus mucosae (TRF62; P = 0.008) and Ruminococcus flavefaciens (TRF260; P = 0.04). With diet CR80, the abundance of Mitsuokella multacida (TRF205; P = 0.045) was higher, but this diet had lower numbers of Clostridiales bacterium (TRF297; P = 0.05) than diet C. In addition, diet CF80 was associated with higher relative abundance of Clostridium butyricum (TRF222; P = 0.02) than diets CF160 and CFR.

Table 2.

Impact of diet on the relative abundance (%) of TRF phylotypes detected in ileal digesta

Putative identityb	No. of TRFs	% maximum identity	Dieta					SE	P value
			Control	CF80	CF160	CR80	CFR
Lactobacillus mucosae/F	62	99	0.8 b	0.2 b	2.0 b	1.6 b	4.6 a	0.84	0.008
Lactobacillus reuteri/F	64	99	3.4	0.5	0.9	1.2	0.7	0.75	0.06
Mitsuokella multacida/F	205	93	0.4 b	0.8 ab	0 b	1.8 a	0.1 b	0.43	0.045
Clostridium butyricum/F	222	96	4.3 ab	9.1 a	1.0 b	6.6 a	1.3 b	1.70	0.02
Faecalibacterium prausnitzii/F	223	99	1.6	3.3	0.4	2.2	0.1	0.84	0.08
Ruminococcus flavefaciens/F	260	95	0.1 a	0.1 a	0 a	0.5 ab	1.7 b	0.44	0.04
Clostridiales bacterium/F	297	91	4.4 a	2.4 ab	2.7 ab	0.01 b	0.9 b	0.99	0.05

^aCF80, 80 g kg⁻¹ chicory forage; CF160, 160 g kg⁻¹ chicory forage; CR80, 80 g kg⁻¹ chicory root meal (Inu60; Inter-Harz Gmbh, Germany); CFR, mix of 80 g kg⁻¹ chicory forage and root meal. Data are presented as least-square means ± SE (n = 5). Different letters within rows indicate significant difference (P < 0.05).

^bF, Firmicutes.

Diet-responsive bacteria in colon.

TRFs correlated with LAB (TRF62, TRF244, TRF253, TRF257, and TRF331) in colon (Fig. 3A) were more than twice as abundant with diet CF160 (19.2%) than with diets C (7.4%) and CF80 (7.6%). However, this difference was not significant due to large individual variation (P = 0.08). The relative abundance of TRFs correlated with butyrate-producing bacteria (TRF223, TRF258, TRF271, TRF274, TRF317, and TRF318; Fig. 3B) and was markedly higher in pigs on diets CF160 and CF80 than in pigs on all other diets (P = 0.0002). The relative abundance of TRFs identified as Prevotella sp. group (TRF130, TRF160, TRF161, TRF163, TRF264, and TRF411 to TRF412; Fig. 3C) was lower (P = 0.02) with diet CF160 than with all other diets.

We also found correlations between diet and specific TRFs in the colon (Table 3). Diets CR80 and CFR were correlated with higher relative abundance of Dialister succinatiphilus (TRF211; P = 0.0003) and Megasphaera elsdenii (TRF275; P < 0.0001), respectively, than with diets C, CF80, and CF160. In addition, a TRF identified as Prevotella sp. (TRF163; P = 0.007) had higher abundance in pigs fed diet CR80 than in pigs fed all other diets, whereas the relative abundance of Prevotella copri was lower with diet CF160 (TRF264; P = 0.008), which corresponded to the lower numbers of the Prevotella group in pigs on this diet. The relative abundance of Faecalibacterium prausnitzii (TRF223) was lowest with diet CR80 (P = 0.03). As observed for the accumulated abundance of butyrate-producing bacteria, diets CF80 and CF160 were associated with the increased relative abundance of Roseburia sp. (TRF274; P = 0.0007).

Table 3.

Effect of diet on the relative abundance (%) of terminal restriction fragment (TRF) phylotypes detected in colonic digesta

Putative identityb	No. of TRFs	% maximum identity	Dieta					SE	P value
			Control	CF80	CF160	CR80	CFR
Lactobacillus mucosae/F	62	99	1.2	0.3	0.4	1.2	0.3	0.30	0.08
Prevotella sp./B	161	96	1.6	1.9	1.1	2.5	3.1	0.53	0.09
Prevotella sp./B	163	99	1.1 a	0.6 a	0.4 a	2.1 b	1.1 a	0.38	0.007
Veillonellaceae bacterium/F	207	97	0.8 ab	1.2 a	0.7 ab	0.1 b	0 b	0.30	0.046
Dialister succinatiphilus/F	208	98	0.01 a	0.01 a	0.2 a	0.9 b	0.5 ab	0.22	0.03
Dialister succinatiphilus/F	211	97	3.0 ac	2.1 bc	0.3 b	4.7 a	4.9 a	0.68	0.0003
Faecalibacterium prausnitzii/F	223	99	2.3 a	2.3 a	2.5 a	0.1 b	2.3 a	0.57	0.03
Clostridium cellulolyticum/F	261	88	3.3 ab	4.6 a	1.9 bc	0.3 c	2.4 abc	0.92	0.03
Prevotella copri/B	264	99	28.3 a	21.8 a	7.9 b	25.2 a	20.8 a	3.63	0.008
Roseburia sp./F	274	99	5.2 ac	7.5 c	11.8 b	4.7 ac	3.2 a	1.30	0.0007
Megasphaera elsdenii/F	275	100	2.8 a	1.8 ab	0.3 b	5.4 c	6.0 c	0.76	<0.0001
Phascolarctobacterium sp./F	302	99	1.1 a	0.5 ab	0.2 b	0.2 b	0 b	0.24	0.02
Eubacterium rectale/F	318	98	2.3	4.2	4.4	1.9	2.0	0.77	0.06
Prevotellaceae bacterium/B	411	98	0.2 a	0.3 ab	0.8 b	0 a	0 a	0.17	0.02
Prevotella sp./B	412	99	2.0 ab	3.3 a	3.6 a	0.3 b	2.0 ab	0.72	0.03

^aCF80, 80 g kg⁻¹ chicory forage; CF160, 160 g kg⁻¹ chicory forage; CR80, 80 g kg⁻¹ chicory root meal (Inu60; Inter-Harz Gmbh, Germany); CFR, mix of 80 g kg⁻¹ chicory forage and root meal. Data are presented as least-square means ± SE (n = 6). Different letters within rows indicate significant difference (P < 0.05).

^bF, Firmicutes; B, Bacteroidetes.

Correlations between intestinal microbiota and SCFA products.

A negative relationship between the relative abundance of LAB and acetic acid concentration was observed (r = −0.650; P = 0.001) in ileal samples. In colon, the relative abundance of butyrate-producing bacteria was positively correlated with the proportional acetic acid concentration (r = 0.576; P = 0.0007) as well as acetic acid and butyric acid concentrations (r = 0.372; P = 0.04). In addition, there was a strong positive relationship between the relative abundance of M. elsdenii (TRF275) and n-valeric acid concentration (r = 0.693; P < 0.0001).

DISCUSSION

Our results demonstrated that the inclusion of chicory (Cichorium intybus L.) in a cereal-based diet results in changes in the gut microenvironment and gut morphology of pigs. Within each diet type, these changes followed a similar pattern in the small and large intestine. However, the dietary responses were different between those with the inclusion of chicory root and those with chicory forage. This could be related to the chemical composition of the dietary fiber fraction, where the root is characterized by a high content of inulin-type fructan and oligofructose while the forage is characterized by a high content of pectin.

To facilitate the identification of a diet-related pattern in the occurrence of major bacterial groups, bacteria falling into clostridial cluster XIVa and cluster IV were grouped for their major role as lactate and butyrate producers in the gut. There was a marked increase in the relative abundance of TRFs identified as butyrate-producing bacteria with the increasing inclusion of chicory forage in the diet (Fig. 3B). No such response could be seen when chicory root was included in the diet. The shifts in this functional group of bacteria were correlated with acetic acid production in the colon (r = 0.5755; P = 0.0007). We also found a positive correlation between the concentration of acetic acid and butyric acid in the colon (r = 0.372; P = 0.04). Butyrate-producing bacteria are capable of utilizing acetate, and some of them are even net consumers (3), which makes acetic acid a key substrate in cross-feeding interactions between colonic bacteria (10). In the present study, chicory forage pectin appeared to be the preferred dietary substrate for butyrate-producing bacteria in the porcine colon instead of chicory inulin-type fructan. This is supported by high microbial pectin fermentation in the porcine colon, with acetate being the major end product (13). In contrast, in vitro cross-feeding studies have shown that inulin-type fructan and oligofructose are favorable substrates for butyrate-producing bacteria (14, 15). Scott and coworkers (30) also showed that Roseburia inulinivorans, a butyrate-producing bacterium, can utilize inulin-type fructan directly. This in vitro effect of inulin could not be verified in the present in vivo study. One reason could be that in vivo, the microbiota of the large intestine is affected not only directly by inulin itself but also indirectly by the metabolites produced and by cross-feeding (24, 26). Nevertheless, our results indicated that butyrate-producing bacteria were responsive to alternative dietary substrates.

The relative abundance of butyrate-producing F. prausnitzii (TRF223) was lower in pigs fed diet CR80 in this study. A previous study showed that F. prausnitzii can be promoted by inulin (27), whereas ingesting diets with pea fiber and fructooligosaccharides had no effect on the occurrence of F. prausnitzii in human fecal microbiota (5). We suspect that in our study, the lower prevalence of F. prausnitzii with diet CR80 could be due to the more complete digestion of chicory inulin-type fructans prior to the colon in pigs compared to that in humans (26, 33).

Prevotella spp. are strictly anaerobic bacteria that have been identified as a dominant genus in the large intestine of pigs (21). In this study, Prevotella spp. comprised the dominant phylotype in pig colon, but this genus was also prevalent and abundant in the ileum and showed a similar response pattern to different diets (Fig. 3C). Arumugam and coworkers (2) recently reported that 1 of the 3 enterotypes that appears in human gut microbiota is driven by Prevotella, and that this enterotype can be responsive to the dietary regimen (35). De Filippo et al. (8) demonstrated that rural African children living on a diet consisting of millet grain, sorghum, legumes, and vegetables harbor a gut microbiota that was particularly dominated by the Prevotella enterotype. This indicates that the abundance of Prevotella is influenced by the type of dietary fiber. In the present study, changes in Prevotella spp. responded to different fiber fractions in the diet (see Table S1 in the supplemental material). A gradual reduction in Prevotella spp. in colon microbiota was seen when more chicory-forage pectin was included than when the cereal-based control diet was used (Fig. 3A). According to Dodd et al. (12), various species of Prevotella in human colonic microbiota can degrade dietary xylan from cereal grains. Moreover, Prevotella bryantii has been suggested to be able to break down soluble xylans by producing xylanases, mannanases, and β-glucanase in nonruminants (16). In our study, the cereal-based control diet had the highest content of arabinoxylan (from wheat and barley), while in the chicory forage diets there was a gradual decrease in the content of arabinoxylan and a gradual increase in the dietary content of pectin. In contrast, in the chicory root diet (CF80) arabinoxylan was replaced with inulin and in the mixed diet (CFR) arabinoxylan was replaced with both pectin and inulin.

Physiological conditions differ between locations in the gut and determine bacterial community structure and diversity (22). Our results showed that the microbiota clustered strongly by gut segment rather than by individual pig, with one distinct cluster in ileum and one in colon (Fig. 2). A clear dominance of Lactobacillus spp. in ileum was identified, confirming earlier findings in pig small intestine (20, 21). The LAB group is known to contain beneficial bacteria and includes many strains that are used as probiotics. For instance, Lactobacillus johnsonii has been found to be involved in the regulation of interleukin-12 production mediated by its cell wall component, which might help to maintain host homeostasis (31), and Lactobacillus mucosae was first identified from pig small intestine and reported to possess probiotic mucus-binding capacity (28). Interestingly, in our study, these two species were correlated with chicory inclusion but not with the control diet. In addition, the accumulated abundance of TRFs matched with LAB was higher in the ileum of pigs fed diets CF160 and CFR (mixture of chicory forage pectin and inulin-type fructan) (Fig. 3A). Similar results have been reported for pigs fed a mixture of fermentable carbohydrates, including sugar beet pulp and inulin, compared to pigs fed a maize starch basal diet (20). In our study, the LAB group was promoted by both inulin-type fructan and chicory-forage pectin. Earlier studies have shown that 20 to 100% of inulin (25, 33) can be degraded in the porcine small intestine, while for pectin the prececal digestibility in pigs varies from 2 (13) to 30% (E. Ivarsson and J. E. Lindberg, unpublished data).

Colonization by specific bacteria may be important and can affect the host (2, 22). One example is M. elsdenii, which is abundant in colon microbiota of pigs fed a swine dysentery-preventive diet with the inclusion of chicory roots (24). M. elsdenii is not a primary fructan degrader, but this phylotype was present in higher abundance in the colon of pigs fed chicory inulin diets than in pigs fed the control diet in the present study. This could be related to increased colonic n-valeric acid concentration in pigs on inulin-associated diets (see Table S3 in the supplemental material). An important trait of M. elsdenii is its ability to utilize lactate and to convert it to propionic and/or valeric acid (24). Accordingly, a strong positive correlation between the relative abundance of M. elsdenii (TRF275) and the concentration of n-valeric acid in the colon (r = 0.693; P < 0.0001) was established in this study.

Diet type can influence gut anatomy indirectly or directly via the local microbiota composition. However, the impact of dietary fiber on gut morphology is variable and depends on type of fiber, level of inclusion, and digestion site in the GI tract (6, 7, 19, 32). For instance, inulin-type fructan has been shown to promote gut development, and this is suggested to be mediated by the stimulation of LAB in the pig GI tract (32). In our study, the greatest villus height in distal ileum was observed in pigs on diet CR80, which could be related to the highest lactic acid concentration on this diet (Fig. 1A). Moreover, a thinner muscle layer in colon was found in pigs fed diet CF160, which contains the highest content of chicory forage pectin. This could be explained by increased gut distension due to the water-holding and gel-forming properties of pectin. Roth and coworkers (29) showed that a citrus pectin diet can improve intestinal conditions in rats with short-bowel syndrome.

In conclusion, we were able to identify responses in gut environment and microbiota composition that were correlated with functional bacterial groups and the dietary inclusion of chicory (Cichorium intybus L.). In ileum, the inclusion of chicory roots was linked with lactic acid concentration in digesta and the relative abundance of LAB. In colon, the inclusion of chicory forage was associated with the relative abundance of butyrate-producing bacteria and colonic acetate concentration.