Almond shells as a sustainable alternative to commercial insoluble fiber sources in piglets: effects on performance and digestive health

Abstract

Nutritional strategies have been evaluated to alleviate the negative effects of weaning, including the incorporation of dietary fiber (DF). Dietary fiber comprises soluble and insoluble fractions, both important for intestinal health. This study evaluated the effects of three insoluble fiber (IF) sources—coarse almond shells (AS), wheat bran (WB), and a commercial IF source (CIF)—included in high-NDF diets (12.0%) for weaned piglets, compared with a lower-NDF control diet (NC; 10.2%). All diets were isoenergetic (net energy: 2,550 kcal/kg) and isonitrogenous (crude protein: 16.3%). Growth performance, dry matter digestibility, diarrhea incidence, jejunal gene expression, and jejunal and colonic microbiota composition were assessed 14 d post-weaning. Neither growth performance nor digestibility was affected by treatment (P > 0.05). Piglets fed AS showed lower diarrhea incidence than NC-fed piglets (0.67 vs 1.67 d; P < 0.05). Jejunal gene expression revealed higher expression of growth factors (EGF, IGF1) in AS-fed piglets than in WB-fed piglets (1.86 vs 0.63, 1.20 vs 0.86; P < 0.05). Dietary treatments also influenced intestinal microbiota composition (P < 0.05). In the jejunum, IF (WB, CIF, and AS) inclusion reduced the abundance of Proteobacteria compared with NC (0.31, 0.37, and 0.32 vs 7.23%; P < 0.05). At family level, WB increased Lactobacillaceae abundance relative to NC (89.24 vs 42.51%; P < 0.05), whereas CIF and AS showed intermediate values (59.67 and 57.88%). In the colon, microbial communities were affected by both IF inclusion and fiber physicochemical characteristics (P < 0.05). CIF-fed piglets exhibited higher Proteobacteria abundance (5.35%) and lower richness (P < 0.05), while AS (1.81%) produced microbial profiles similar to NC and WB (1.22 and 0.79%). Overall, ASs, characterized by their high IF content and coarse particle size, reduced diarrhea incidence, increased jejunal growth factor expression, and promoted a favorable intestinal microbial profile without affecting growth performance or digestibility. These findings support the use of ASs as a sustainable alternative fiber source in piglet nutrition.

The source of insoluble dietary fiber influences gut health in weaned piglets. Almond shells improved indicators of intestinal health while maintaining growth performance.

Introduction

In intensive swine production systems, early weaning is commonly practiced to improve sow productivity and production efficiency. Nevertheless, weaning represents a major physiological and nutritional stressor for piglets, causing changes in intestinal structure, epithelial barrier integrity, microbial populations, and growth performance (Upadhaya and Kim 2021). Among the nutritional approaches evaluated to reduce weaning-associated stress, dietary fiber (DF) supplementation has gained increasing interest (Han et al. 2024), as it may support intestinal health through multiple complementary mechanisms (Hedemann et al. 2006; Sun et al. 2021). However, although moderate DF levels may exert positive effects, excessive dietary inclusion can negatively affect productive performance (Montagne et al. 2012).

Post-weaning diarrhea is considered one of the major challenges during the weaning period, as it causes substantial economic losses due to decreased feed intake, impaired growth, increased mortality, and elevated medication costs (Rhouma et al. 2017). Enterotoxigenic Escherichia coli (ETEC) is one of the main bacterial causes of post-weaning diarrhea, with the F4-fimbriated strain being one of the most prevalent ETEC pathotypes worldwide (Fairbrother et al. 2005). Dietary inclusion of DF has been associated with a reduced incidence of this disorder (Uddin et al. 2023), although the response is strongly influenced by the type of fiber used. DF can be classified into soluble fiber (SF) and insoluble fiber (IF).

IF is particularly relevant in post-weaning diets because it can support intestinal function by regulating digesta flow, stimulating intestinal development and digestive enzyme activity, improving nutrient digestibility, and promoting beneficial microbial populations and volatile fatty acid production (Han et al. 2024). Through these mechanisms, the inclusion of low levels of IF in post-weaning piglets has been shown to modulate the gut microbiota and, consequently, reduce the incidence of diarrhea (Gerritsen et al. 2012; Chen et al. 2020).

Among the sources of IF studied, wheat bran (WB) is one of the most widely used, and its inclusion in commercial piglet diets is well established in the European Union, especially given increasing legal restrictions on the use of medicinal substances. Its incorporation has been associated with several beneficial effects on intestinal function and productive performance, including a reduced incidence of diarrhea, increased expression of tight junction proteins, enhanced barrier integrity, promotion of beneficial bacterial populations, and increased volatile fatty acid production (Flis et al. 2017; Han et al. 2024).

The growing interest in IF supplementation has also driven the rapid expansion of commercial lignocellulosic fiber products for piglet nutrition, most of which are derived from processed wood sources (Ardana et al. 2024). Flis et al. (2017) recommend the inclusion of 1.5–2% lignocellulosic concentrates to support gastrointestinal health; however, available evidence suggests that their effects are highly dependent on the origin and physicochemical characteristics of the fiber source.

Almond shells (ASs), by contrast, represent a lignocellulosic by-product of the almond industry with limited industrial utilization. Given that almonds are the most extensively produced tree nut worldwide (FAOSTAT 2025), large quantities of ASs are generated annually. Consequently, there is an increasing interest in their valorization within circular-economy approaches (Maaloul et al. 2017). In this line, Ruiz et al. (2026) assessed the inclusion of ASs (2.5%) at different particle sizes as an IF source in weaned piglet diets compared with WB, reporting improvements in intestinal morphology and performance-related traits. Nevertheless, comparative studies involving ASs, commercial IF sources, and low-fiber diets are still scarce, indicating a limited understanding of their potential as alternative fiber ingredients.

Accordingly, we hypothesized that ASs, owing to their IF composition and particle size, could influence intestinal function by modulating microbial communities and epithelial responses. As a result, ASs were expected to improve growth performance and positively affect intestinal gene expression and microbial communities compared with an IF-free diet, while eliciting responses comparable to those of conventional IF sources such as WB or a commercial IF. Therefore, the aim of this study was to assess the effects of dietary ASs as an IF source in weaned piglets, compared with an IF-free control diet and other IF sources (WB and a commercial IF), on growth performance, nutrient digestibility, gene expression, and the gut microbial community during the first 14 d post-weaning.

Material and methods

Insoluble fiber sources

The IF sources employed in the present study were WB, a commercial IF source and ASs. WB was sourced from a local milling facility in southeastern Spain. The commercial IF used was Jelucel®, a widely utilized IF source in food and feed technology, supplied by JELU-WERK J. Ehrler GmbH & Co. KG (Germany). Milled ASs, a lignified by-product derived from the thick endocarp of the almond fruit (Prunus amygdalus L.), were included at coarse particle size based on previous findings reporting beneficial effects of larger particle sizes of this fiber source (Ruiz et al. 2026). The coarse-milled ASs were supplied by a local company, Hermen Europa SL (Spain). The particle size distribution of the IF sources is presented in Table 1.

Table 1.

Particle size distribution of wheat bran, commercial IF, and almond shells.

	Wheat bran	Commercial IF	Almond shell
dgw 1 , µm	499.2	55.5	1,221.2
S gw, dimensionless	5.4	3.0	1.5
d 0.1, µm	34.0	15.0	839.4
d 0.5, µm	927.3	53.5	1,209.0
d 0.9, µm	1,511.8	209.6	1,627.0
D 3,2, µm	59.1	24.1	864.7
D 4,3, µm	864.7	88.2	1,218.3

¹ d_gw, geometric mean diameter; S_gw, geometric standard deviation; d_0.1, 10th percentile diameter; d_0.5, median diameter; d_0.9, 90th percentile diameter; D_3,2, surface-weighted mean diameter; D_4,3, volume-weighted mean diameter.

Experimental design and diets

To evaluate the effects of different fibrous sources in diets for weaned piglets, four experimental diets were formulated (Table 2). A control diet (NC), prepared without the inclusion of any fiber source, contained 10.2% neutral detergent fiber (NDF). The remaining three diets were formulated to achieve a similar NDF concentration of 12% by incorporating different IF sources: 8% WB, 2.5% commercial IF (CIF; Jelucel®), or 2.5% AS. All diets were isoenergetic and isonitrogenous and designed to meet the nutritional requirements of piglets with a weaning body weight of 7–12 kg during the prestarter phase (days 1–14 post-weaning), including the target NDF level, in accordance with Molist and Mateos (2023).

Table 2.

Ingredients (g/kg) and calculated composition (g/kg) of the experimental diets.

	Experimental diets1
Ingredients (g/kg)	NC	WB	CIF	AS
Barley	150	150	150	150
Corn	200	200	200	200
Wheat	390	297	346	346
NUTRIN CP2	50.0	50.0	50.0	50.0
Premix 3	50.0	50.0	50.0	50.0
Whey	45.0	45.0	45.0	45.0
HP 3004	30.0	30.0	35.0	35.0
Plasma	35.0	35.0	35.0	35.0
Lard	23.1	35.8	35.9	35.9
Wheat bran	…	80.0	…	…
Jelucel®5	…	…	25.0	–
Almond shell	…	…	…	25.0
Diatomaceous earth	8.00	8.00	8.00	8.00
Monocalcium phosphate	5.69	5.50	5.85	5.85
Salt	3.00	3.00	3.00	3.00
Calcium carbonate	2.00	2.00	2.00	2.00
L-Lysine HCl , 780 g/kg Lys	4.20	4.04	4.13	4.13
L-Threonine, 980 g/kg Thr	1.83	1.79	1.83	1.83
DL-Methionine, 880 g/kg Met	1.60	1.59	1.68	1.68
L-Tryptophan, 980 g/kg Trp	0.57	0.54	0.57	0.57
L-Valine, 970 g/kg Val	0.43	0.37	0.46	0.46
Calculated nutrient composition (g/kg as fed)6
CP	164	166	163	163
Standardized ileal digestible lysine	10.8	10.8	10.8	10.8
Standardized ileal digestible threonine	7.34	7.34	7.34	7.34
Standardized ileal digestible methionine	3.42	3.42	3.42	3.42
Standardized ileal digestible tryptophan	2.38	2.38	2.38	2.38
Standardized ileal digestible valine	7.45	7.45	7.45	7.45
Standardized ileal digestible isoleucine	5.65	5.65	5.65	5.65
Lactose	31.63	31.63	31.63	31.63
NE (kcal/kg)	2,550	2,550	2,550	2,550
NDF	102	120	120	120
ADF	37.9	43.9	51.9	51.9
ADL	9.5	11.1	15.1	15.1

CP, crude protein; NE, net energy; NDF, neutral detergent fiber; ADF, acid detergent fiber; ADL, acid detergent lignin.

¹NC, diet with no fiber source added; WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% almond shells.

²NUTRIN CP (Industrial Técnica Ganadera S.L., Murcia, Spain) complementary feeds. Composition (g/kg): 510 g protein, 67 g ash, 40 g crude fiber, 20 g fat, <50 g calcium, <20 g phosphorus, 9 g sodium, 32 g lysine, and 8 g methionine.

³Premixes are provided per kg of the diet: 12,000 IU vitamin A, 1,900 IU vitamin D3, 75.0 mg vitamin E, 2.20 mg vitamin K3, 125 mg vitamin C, 140 mg biotin, 0.80 mg folic acid, 2.00 mg vitamin B1, 6.00 mg vitamin B2, 3.00 mg vitamin B6, 120 mg iron (iron chelate of glycine hydrate), 0.70 mg iodine (potassium iodide), 120 mg copper (copper chelate of solid glycine hydrate), 22.5 mg manganese (glycine hydrate manganese chelate), 96.9 mg zinc (zinc sulfate monohydrate), 23.1 mg zinc (zinc oxide), 0.10 mg selenium (sodium selenite), and 0.20 mg selenium (hydroxylated analog of selenomethionine). 6-phytase (EC 3.1.3.26), 1250 FYT. 25,000 FTU 6-Phytase EC 3.1.3.26, 360,000 U endo-1,4-beta-glucanase EC 3.2.1.4, 47,000 U endo-1,3(4)-beta-glucanase EC 3.2.1.6, 700,000 U endo-1,4-beta-xylanase EC 3.2.1.8, 8,000 U alpha-amylase EC 3.2.1.1, 34,000 U Bacillolysin EC 3.4.24.28, 20 × 10⁹ CFU Saccharomyces cerevisiae CNCM I-1079, 400 × 10⁷ CFU Bacillus subtilis (ATCC PTA-6737).

⁴Product of Hamlet Protein (Horsens, Denmark). Composition (g/kg): 920 g dry matter, 560 g protein, 232 g carbohydrates, 68.0 g ash, 35.0 g crude fiber, and 25 g fat. Essential amino acids (g/16 g of N): 6.10 g lysine, 1.30 g methionine, 1.40 g cysteine, 3.90 g threonine, 1.35 g tryptophan, 7.70 g leucine, 4.60 g isoleucine, 5.00 g phenylalanine, 3.70 g tyrosine, 4.80 g valine, 2.60 g histidine and 7.20 g arginine.

⁵Supplied by JELU-WERK J. Ehrler GmbH & Co. KG (Germany).

⁶According to Molist and Mateos (2023).

Before the fiber sources were incorporated, all feed ingredients were ground through a 2-mm screen to standardize particle size distribution. Diatomaceous earth was incorporated as an indigestible marker for the determination of the coefficient of apparent total tract digestibility (CATTD). Experimental diets were offered ad libitum in mash form. The analyzed chemical composition of the fiber sources and experimental diets is presented in Table 3.

Table 3.

Chemical composition (g/kg of dry matter) of the IF sources and experimental diets.

	Fibre source			Experimental diets1
Item	Wheat bran	Commercial IF	Almond shell	NC	WB	CIF	AS
DM	901	960	913	903	905	904	905
Chemical composition (g/kg of DM)
Ash	3.85	0.23	1.82	39.8	43.2	43.0	42.2
CP	155	…	9.98	156	151	154	147
NDF	359	920	901	107	132	131	133
ADF	118	746	563	37.1	45.8	57.3	56.9
ADL	31.3	387	202	5.32	6.38	11.21	11.15
SF	50.6	6.1	13.9	16.7	39.7	26.2	27.1
IF	422	993	956	151	166	170	174
TDF	477	999	970	168	206	196	201

¹NC, diet with no fiber source added; WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel); AS, diet containing 2.5% milled almond shells; DM, dry matter; CP, crude protein; NDF, neutral detergent fiber; ADF, acid detergent fiber; ADL, acid detergent lignin; SF, soluble fiber; IF, insoluble fiber; TDF, total dietary fiber.

Housing and animal management

All animal-related procedures were conducted under project authorization No. A13220603, approved by the Regional Ministry of Water, Agriculture, Livestock, and Fisheries of the Region of Murcia, as well as by the Ethics and Biosecurity Committee for Animal Experimentation of the University of Murcia, Spain. Animal handling and experimental use were carried out in compliance with the Spanish Policy for Animal Protection (RD 118/2021) and European Union Directive 2010/63/EU concerning the protection of animals used for scientific purposes.

The experiment was performed on a commercial pig farm operating under an intensive production system (Porcisan, Cuevas del Reyllo, Murcia, Spain). A total of 480 Landrace × Large White piglets weaned at 26 ± 2 d of age, with an average initial body weight of 7.37 ± 0.15 kg, were first distributed into homogeneous groups according to weaning body weight. The resulting pens were then randomly assigned to one of the four experimental diets. Piglets were housed on fully slatted floors (3.21 × 2.76 m) in groups of 20 piglets (six replicates per treatment). Each pen was equipped with a feeder and two nipple drinkers, and feed and water were supplied ad libitum throughout the experimental period. Room temperature was maintained between 22 and 25 °C.

Fecal consistency was assessed daily using the scoring system described by Pedersen et al. (2011) through visual evaluation of five fresh fecal samples per pen (grade 1 = hard, dry, and crumbly feces; grade 2 = soft feces; grade 3 = watery diarrhea). Diarrhea incidence was calculated per pen as the number of days presenting clinical signs of diarrhea relative to the total duration of the trial (14 d) (Mateos et al. 2006).

Feed intake was recorded daily at the pen level, whereas individual body weight was measured using a Gram Next Xtrem F electronic scale (Flintec, Spain). After 14 d on the experimental diets, piglets were individually weighed, and average daily gain (ADG), average daily feed intake (ADFI) and gain-to-feed ratio (G/F) were calculated. Water consumption was not measured during the experimental period.

Sample collection

To determine CATTD of the experimental diets, fecal samples were collected on days 12 and 13 of the trial from eight randomly selected piglets per pen by gentle rectal massage (25 g per piglet). Fecal samples from the eight piglets within each pen were pooled and thoroughly homogenized to obtain one representative sample per pen and sampling day. Samples were stored at –20 °C until subsequent analysis.

On day 14, piglets were fasted overnight and weighed individually, after which one animal per replicate (six piglets per treatment) was selected for slaughter based on the mean body weight of its replicate group. Selected piglets were anesthetized and humanely euthanized by intravenous administration of sodium pentobarbital (0.15 mg/kg body weight; VetViva Richter GmbH, Austria). Following exsanguination, the abdominal cavity was immediately opened from the sternum to the pubis. Ligatures were applied at the distal esophagus and at the rectum, and the entire gastrointestinal tract was removed. The major intestinal segments were subsequently identified and sampled. Jejunal samples were collected at 3.5 m distal to the ligament of Treitz. A jejunal tissue sample (approximately 1 cm) was collected for gene expression analyses, rinsed three times with phosphate-buffered saline (PBS), placed into cryotubes (Freezing Vials, JetBiofil, Guangzhou, China), rapidly frozen in liquid nitrogen, and stored at –80 °C until analysis. Jejunal and colonic digesta samples were collected in cryotubes (Freezing Vials, JetBiofil, Guangzhou, China) after carefully opening the intestinal segments longitudinally. The luminal contents were carefully collected and thoroughly homogenized to ensure sample homogeneity, rapidly frozen in liquid nitrogen, and stored at −80 °C until microbial analysis.

Chemical and particle size characterization

Feed samples were ground using a 1-mm screen (Retsch ZM200, Haan, Germany) before chemical analysis. Dry matter (DM), crude protein (CP) and ash contents were determined in accordance with the procedures of the Association of Official Analytical Chemists (AOAC) (2006). NDF, acid detergent fiber (ADF) and acid detergent lignin (ADL) were analyzed following the methodology described by Van Soest et al. (1991). NDF was determined with heat-stable amylase, and values were reported inclusive of residual ash. ADF was similarly expressed on an ash-inclusive basis, and ADL was quantified after cellulose solubilization with sulfuric acid (72% w/w). SF, IF, and total DF (TDF) were determined using enzymatic–gravimetric methods according to AOAC (2006), employing a Megazyme K-TDFR assay kit (Megazyme, Wicklow, Ireland).

Frozen fecal samples were oven-dried at 60 °C until constant weight and ground through a 1-mm sieve (Retsch ZM200, Haan, Germany). Feed and fecal samples were analyzed for acid-insoluble ash (AIA) as described by Sales and Janssens (2003). The coefficient of apparent total tract digestibility (CATTD) of DM was calculated using AIA as an internal marker, according to the following equation: CATTD (%) = [1 − (AIA_diet/AIA_feces)] × 100, where AIA_diet and AIA_feces correspond to the marker concentration in the diet and fecal DM, respectively.

Particle size distribution of each IF source was determined in 1-g samples by laser diffraction using a Mastersizer 2000LF (Malvern Panalytical Ltd, Spectris, USA) across five different lots per IF source. Geometric mean diameter (d_gw), geometric standard deviation (S_gw), surface‑weighted mean diameter (D_3,2), and volume‑weighted mean diameter (D_4,3) were calculated based on the measured fraction of each particle size, following the formulas proposed by Lyu et al. (2020). The 10th percentile diameter (d_0.1), median diameter (d_0.5), and 90th percentile diameter (d_0.9) were obtained according to the detected percentile particle size.

Gene expression

RNA extraction and cDNA analysis

Total RNA was isolated from jejunal tissue using the Maxwell® RSC SimplyRNA Tissue Kit (Promega, USA). The purified RNA was subsequently used as a template for complementary DNA (cDNA) synthesis using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA). Quantitative real-time PCR was carried out in a 96-well plate to quantify the expression of target genes. The specific primer sequences used for each gene are detailed in Table 4.

Table 4.

Forward and reverse primers used for the genes analyzed.

Gene	Primer Forward (5’-3’)	Primer reverse (5’-3’)	Reference
IL-1B	GGTGACAACAATAATGACCTGTTATTTG	GCTCCCATTTCTCAGAGAACCA	1
IL-6	CCAATCTGGGTTCAATCAGGAG	ACAGCCTCGACATTTCCCTTATT	1
TNF-α	CACCACGCTCTTCTGCCTACT	GACGGGCTTATCTGAGGTTTGA	1
IFN-γ	TGACTTTGTGTTTTTCTGGCTCTT	CACTCTCCTCTTTCCAATTCTTCAA	1
IL-10	TGAGGCTGCGGCGCT	GAGCTTGCTAAAGGCACTCTTCA	1
TLR-4	CATCCCCACATCAGTCAAGATACT	TCAATTGTCTGAATTTCACATCTGG	1
TGFB1	GCGGCAGCTCTACATTGACTT	GACCTTGCTGTACTGAGTGTCTAGG	1
IL-8	GGAAAAGTGGGTGCAGAAGGT	GAGAATGGGTTTTTGCTTGTTGT	1
IL17A	CCAGACGGCCCTCAGATTAC	ATCTTCCTTCCCTTCAGCATTG	1
IL22	TGTTCCCCAACTCTGATAGATTCC	GTTGTTCACATTTCTCTGGATATGCT	1
Claudin1	CTTCGACTCCTTGCTGAATCTGA	CTTCCATGCACTTCATACACTTCAT	1
Occludin	CAGGTGCACCCTCCAGATTG	CAGGCCTATAAGGAGGTGGACTT	1
ZO-1	GCTATGTCCAGAATCTCGGAAAA	TGCTTCTTTCAATGCTCCATACC	1
SLC5A1	GGCCATCTTTCTCTTACTGGCA	TCCCACTTCATGAAAAGCAAAC	1
SLC15A1	GGTTATCCCTTGAGCATCTTCTTC	AGTGCTCTCATTCCATAGTAGGAAAA	1
ALPI	ATGTCTTCTCTTTTGGTGGCTACA	GGAGGTATATGGCTTGAGATCCA	1
SI	CGACCCCTTTTGCATGAGTT	AAGGCTGGACCCCATAGGAA	1
CCK	CAGCAGGCTCGAAAAGCAC	AATCCATCCAGCCCATGTAGTC	1
ACTB	CAAGGACCTCTACGCCAACAC	TGGAGGCGCGATGATCTT	1
GAPDH	TTCGTCAAGCTCATTTCCTGGTA	TCCTCGCGTGCTCTTGCT	1
B2M	TCACTCCTAACGCTGTGGATCA	CGGTTAGTGGTCTCGATCCC	1
IGF1R	ATGGAGGAAGTGACAGGGACTA	GTGGTGGTGGAGGTGAAGTG	2
IGF1	AACAAGCCCACAGGGTACG	TCTCCAGCCTCCTCAGATCA	2
EGF	ATGTCAACGAGTGTGCCTTC	TCAGGGAGCAGAACAAATCC	2
EGFR	ATCTGTAACCCGCTGTGCTC	GGCATTCTCCACGAACTCTC	2
SLC6A19	TCTGTCCACAACAACTGCGAG	CAGCGAAGTTCTCCTGCGTC	3
IL-12p35	GGCCTGCTTACCACTTGAAC	GCATTCATGGCCTGGAACTC	4
IL-12p40	CTGAAGAAGACGGCATCACG	AGGAGTGACTGGCTCAGAAC	4

¹ González-Solé et al. (2020).

² Xing et al. (2017).

³ Chen et al. (2020).

⁴ Levast et al. (2010).

Quantitative PCR

Quantitative real-time PCR data were analyzed using the 2^–ΔΔCq method. The housekeeping genes ACTB, GAPDH, and B2M were used as internal reference genes. Relative gene expression was calculated from cycle threshold values and primer efficiencies of the target genes, normalized to the reference genes and expressed relative to the NC treatment.

Microbiological analyses

Microbial DNA extraction

Microbial DNA was extracted from 100 mg of jejunal and colonic digesta using the MagMAX Microbiome Ultra Nucleic Acid Isolation Kit (Thermo Fisher Scientific, Warrington, UK) in combination with the KingFisher Duo Prime Purification System (Thermo Fisher Scientific, Warrington, UK). Bacterial community composition was characterized by sequencing the V3–V4 region of the 16S rRNA gene following amplification with the barcoded primer pair 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′ GGACTACHVGGGTWTCTAAT 3′). DNA concentration and purity were evaluated using the Qubit™ dsDNA HS Assay Kit (Invitrogen, Merelbeke, Belgium) on a Qubit™ 2.0 Fluorometer.

Illumina sequencing

Sequencing of the amplified libraries was performed on an Illumina MiSeq platform (San Diego, CA, USA) using the MiSeq Reagent Kit, generating paired-end reads of 2 × 300 base pairs. The quality of the sequencing libraries was evaluated using the Bioanalyzer High Sensitivity DNA Analysis Kit (Agilent Technologies, California, USA).

Bioinformatics

Overall sequencing quality of the 16S V4 libraries was initially evaluated using FastQC (Andrews et al. 2010). Sequence data were processed using the R package DADA2 (v1.8.0; Callahan et al. 2016). The ‘filterAndTrim’ function was applied using a maximum expected error threshold of 2.5, and reads shorter than 165 base pairs after trimming were discarded. Dataset-specific error rates were learned and used by DADA2 to infer true biological sequences. Forward and reverse reads were subsequently merged to construct the sequence table. Amplicon sequence variants (ASVs) were screened for chimeras using the DADA2 chimera removal function, and the detected artifacts were eliminated.

Taxonomic classification of ASVs was performed using the SILVA reference database (v132) through the IDTAXA classifier (Murali et al. 2018), as implemented in the R package DECIPHER (Wright 2016). Non-bacterial sequences were excluded from the dataset. Rarefaction was not performed, in accordance with the recommendations of McMurdie and Holmes (2013), to avoid introducing bias in relative abundance estimates. To improve species-level resolution, the Species-level IdentificatioN of metaGenOmic amplicons (SPINGO, v1.3; Allard et al. 2015) algorithm was applied to the 16S ITGDB database (Hsieh et al. 2022), using similarity and bootstrap cutoff thresholds of 0.5 and 0.8, respectively.

ASV count tables, taxonomic annotations, and sample metadata were combined using the phyloseq package (v3.18; McMurdie and Holmes 2013). Sequencing depth was normalized to the median read count per sample, and taxa bar plots were generated using the R package microViz (v0.12.0; Barnett et al. 2021). Taxa accounting for less than 0.01% of total relative abundance or present in fewer than 60% of samples were excluded from further analyses.

Alpha diversity metrics (Chao1, Fisher, Shannon, Simpson, and Berger–Parker) were calculated using the R package vegan. Beta diversity was assessed using qualitative Jaccard and quantitative Bray–Curtis dissimilarity matrices (Lozupone and Knight 2005; Lozupone et al. 2007), and differences in microbial community structure were visualized using principal coordinates analysis (PC). Additionally, beta diversity distances between jejunal and colonic samples from the same piglet were calculated to evaluate dietary effects on the dynamics of microbial community composition.

Statistical analyses

Data from the four experimental treatments were analyzed using one-way analysis of variance (ANOVA) according to following formula:

$$Y_{ij}=\mu +T_i+e_{ij}$$

where $Y_{ij}$ is the analyzed variable, $\mu$ is the overall mean, $T_i$ is the effect of dietary treatment ($i=\mathrm{NC},\mathrm{ }\mathrm{WB},\mathrm{ }\mathrm{CIF},$ or $\mathrm{AS}$), and $e_{ij}$ is the residual error. The analyzed variables included particle size distribution, performance, digestibility, relative gene expression, microbial relative abundance, and alpha diversity. The pen was considered the experimental unit for performance and digestibility parameters, whereas the individual piglet was regarded as the experimental unit for gene expression and microbiota analyses. Relative gene expression data were log₂-transformed before statistical analysis, whereas microbial relative abundance data were transformed using a centered log-ratio approach. Differences among least-squares means were evaluated using Tukey’s post hoc test. All statistical analyses were conducted using SPSS (version 28.0; SPSS Inc., Chicago, IL, USA).

Differences in beta diversity among dietary treatments were assessed in R using the vegan package (Oksanen et al. 2019) by performing analysis of similarities (ANOSIM) and permutational multivariate analysis of variance (PERMANOVA; adonis2’ function), both based on 999 permutations. Statistical significance was set at P < 0.05, and statistical trends were considered when 0.05 ≤ P < 0.10.

Results

Growth performance and digestibility

Growth performance parameters and digestibility coefficients were not affected by dietary treatments (Table 5; P > 0.05). Furthermore, AS-fed piglets had a lower diarrhea incidence than piglets fed the NC diet (P < 0.05), whereas WB- and CIF-fed piglets did not differ from either group.

Table 5.

Effect of including different sources of IF in the diets of weaned piglets on performance, CATTD of DM, fecal score, and diarrhea incidence from weaning to 14 d post-weaning.

	Experimental diets1				SEM2	P-value
	NC	WB	CIF	AS
Initial body weight (kg)	7.37	7.35	7.39	7.36	0.032	0.972
Final body weight (kg)	10.04	9.71	10.11	9.99	0.086	0.409
ADG (kg/d)3	0.19	0.17	0.19	0.18	0.007	0.633
ADFI (kg/d)	0.29	0.29	0.30	0.30	0.006	0.742
G/F	0.65	0.58	0.64	0.60	0.015	0.450
CATTD of DM (%)	85.20	84.07	83.90	83.91	0.237	0.172
Faecal score	1.60	1.51	1.54	1.51	0.036	0.821
Diarrhoea incidence (days)	1.67a	1.17ab	1.41ab	0.67b	0.128	0.048

¹NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells.

²Standard error of the mean (n = 6 per parameter).

³ADG, average daily gain; ADFI, average daily feed intake; G/F, gain-to-feed ratio; CATTD of DM, coefficient of apparent total tract digestibility of dry matter; Fecal score, grade 1 = hard, dry, and crumbly feces; grade 2 = soft feces; grade 3 = watery diarrhea; Diarrhea incidence, calculated per pen as the number of days with clinical signs of diarrhea.

^a,bMeans within a row without a common letter are statistically different (P < 0.05).

Intestinal gene expression

Relative gene expression in jejunal tissue from piglets fed ASs, WB, a commercial IF source, or a low-fiber control diet is shown in Table 6.

Table 6.

Effect of including different sources of IF in the diets of weaned piglets on relative jejunal gene expression (relative to NC) after 14 d of feeding.

	Experimental diets1				SEM2	P-value
Gene	NC	WB	CIF	AS
ALPI	1.00	0.82	0.95	0.97	0.055	0.692
CCK	1.00	0.88	1.07	1.01	0.077	0.853
Claudin 1	1.00	0.89	1.13	1.07	0.178	0.951
EGF	1.00ab	0.63b	0.92ab	1.86a	0.161	0.007
EGFR	1.00	1.21	1.07	1.02	0.072	0.667
IFN-γ	1.00	1.17	1.36	0.90	0.166	0.740
IGF1	1.00ab	0.86b	1.03ab	1.20a	0.044	0.010
IGF1R	1.00	0.90	0.93	1.02	0.042	0.736
IL-10	1.00	1.11	1.23	0.82	0.073	0.242
IL-12p35	1.00	0.99	0.57	0.81	0.086	0.375
IL-12p40	1.00	1.17	0.68	0.91	0.074	0.194
IL17A	1.00	1.20	1.72	1.02	0.278	0.634
IL1B	1.00	1.17	1.02	0.70	0.153	0.394
IL-22	1.00b	3.27a	2.10ab	0.91b	0.566	0.011
IL-6	1.00	1.33	1.21	0.95	0.143	0.681
IL-8	1.00	1.16	1.20	0.80	0.096	0.242
Occludin	1.00	0.94	1.15	1.02	0.032	0.185
SI	1.00	0.86	0.93	0.64	0.066	0.191
PepT1	1.00	0.96	0.93	0.83	0.037	0.504
SGLT1	1.00	0.89	0.82	0.51	0.063	0.212
BOAT1	1.00	1.29	1.00	0.85	0.070	0.149
TGF-B1	1.00	0.86	0.87	0.99	0.030	0.294
TLR-4	1.00	0.96	1.17	0.96	0.068	0.640
TNF-α	1.00	1.07	0.95	0.92	0.056	0.821
ZO-1	1.00	0.83	0.99	1.07	0.053	0.509

¹NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells.

²Standard error of the mean (n = 6 per parameter).

^a,bMeans within a row without a common letter are statistically different (P < 0.05).

Dietary effects on relative gene expression were observed for EGF, IGF1, and IL-22 (P < 0.05). EGF and IGF1 expression levels were higher in AS-fed piglets than in WB-fed piglets, whereas CIF- and NC-fed piglets did not differ from either group (P < 0.05). IL-22 expression was higher in WB-fed piglets than in piglets fed the NC or AS diets, whereas the CIF group did not differ from either group (P < 0.05).

No diet effects (P > 0.05) were observed for ALPI, CCK, Claudin 1, EGFR, IFN-γ, IGF1R, IL-10, IL-12p35, IL-12p40, IL-17A, IL-1B, IL-6, IL-8, Occludin, SI, PepT1, SGLT1, BOAT1, TLR-4, TNF-α, and ZO-1 expression.

Microbial analysis

Relative abundance

The effects of dietary treatment on microbial relative abundance in the jejunum and colon of piglets are represented in Supplementary Figures 1 and 2, respectively.

In the jejunum content, at the phylum level, WB, CIF, and AS increased the relative abundance of Firmicutes and decreased Proteobacteria compared with the NC group (P < 0.05). At the family and genus levels, WB increased the relative abundance of Lactobacillaceae and Lactobacillus, respectively, compared with NC (P < 0.05), whereas CIF and AS showed no significant differences from the other dietary treatments.

In the colonic content, at the phylum level, CIF-fed piglets had a higher relative abundance of Proteobacteria than those in the other treatment groups (P < 0.05). At the genus level, a trend toward a lower Escherichia–Shigella abundance was observed in NC, WB, and AS compared with CIF (P < 0.10). At the family level, no significant differences were detected among dietary treatments (P > 0.05).

Bacterial richness and evenness

Figures 1 and 2 show the alpha-diversity indices of the jejunal and colonic microbiota, respectively, in piglets fed the experimental diets.

Figure 1.

Effect of including different sources of IF in the diets of weaned piglets on jejunal alpha diversity indices. NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells (n = 6 per diet).

Figure 2.

Effect of including different sources of IF in the diets of weaned piglets on colonic alpha diversity indices. NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells (n = 6 per diet).

In the jejunal content, WB-fed piglets exhibited reduced Observed, Chao1, Fisher, Shannon, and Simpson indices, while showing an increased Berger–Parker index compared with the NC, CIF, and AS groups (P < 0.05).

In the colonic content, CIF-fed piglets displayed lower Observed and Chao1 indices (P < 0.05) and tended to have reduced Fisher index values (P < 0.10) compared with the NC, WB, and AS groups. No dietary effects were observed for the Shannon, Simpson, or Berger–Parker indices (P > 0.05).

Bacterial diversity

Beta-diversity of the jejunal and colonic microbiota in piglets was assessed using Bray–Curtis and Jaccard dissimilarity metrics and visualized using PCoA in Figures 3 and 4, respectively. Beta-diversity distances (Bray–Curtis and Jaccard) between jejunum and colon samples for each piglet and diet are presented in Figure 5.

Figure 3.

Effect of including different sources of IF in the diets of weaned piglets on jejunal beta diversity indices (Bray-Curtis, left; Jaccard, right). Each plot includes 95% confidence ellipses and colored centroids for each dietary group. NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells (n = 6 per diet).

Figure 4.

Effect of including different sources of IF in the diets of weaned piglets on colonic beta-diversity indices (Bray-Curtis, left; Jaccard, right). Each plot includes 95% confidence ellipses and colored centroids for each dietary group. NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells (n = 6 per diet).

Figure 5.

Effect of including different sources of IF in the diets of weaned piglets on beta-diversity distances per piglet between the jejunal and colonic content (Bray-Curtis, left; Jaccard, right) of weaned piglets on day 14. NC, diet with no fiber source added (NDF 10.2%); IF-supplemented diets (NDF 12.0%): WB, diet containing 8% wheat bran; CIF, diet containing 2.5% commercial IF source (Jelucel®); AS, diet containing 2.5% milled almond shells. Stars indicate significant differences between dietary treatments (P < 0.05) (n = 6 per diet).

Analysis of jejunal beta-diversity revealed dietary effects on both Bray–Curtis and Jaccard dissimilarity matrices, as detected by ANOSIM and adonis (P < 0.05). Across both statistical approaches and distance metrics, WB-fed piglets exhibited distinct microbial community structures (P < 0.05) compared with NC, CIF, and AS-fed piglets. Furthermore, CIF and AS piglets tended to differ from NC piglets in microbial composition (P < 0.10).

Analysis of colonic beta-diversity indicated dietary effects on both Bray–Curtis and Jaccard dissimilarity matrices, as determined by ANOSIM and adonis (P < 0.01). Thus, for both metrics, WB-fed piglets exhibited distinct microbial communities compared with those fed the NC, CIF, and AS diets (P < 0.05), and microbial communities in NC-fed piglets differed from those in CIF- and AS-fed piglets (P < 0.05). Colonic community composition differed between the CIF and AS treatments according to the Jaccard metric using adonis (P < 0.05), with a trend detected using ANOSIM (P < 0.1); additional trends were observed for Bray–Curtis dissimilarities (P < 0.10).

Beta-diversity distances between jejunal and colonic samples within each piglet revealed no dietary effect on Bray–Curtis metrics (P > 0.05). In contrast, Jaccard distances were greater (P < 0.05) in WB-fed piglets compared with those fed NC, whereas CIF- and AS-fed piglets did not differ from either dietary treatment.

Discussion

The inclusion of DF in pig nutrition has gained attention due to its capacity to improve intestinal health through physiological effects on the digestive tract and modulation of gut microbiota (Grześkowiak et al. 2022). In weaned piglets, maintaining intestinal integrity and function is particularly important because early weaning predisposes piglets to post-weaning syndrome (Upadhaya and Kim 2021). The incorporation of IF is one strategy to mitigate these effects, promoting epithelial integrity and microbial homeostasis and contributing to the prevention and reduction of post-weaning diarrhea (Han et al. 2024). Moreover, the physical and chemical characteristics of IF sources may influence their effects on intestinal function and health.

Traditionally, DF has been considered detrimental due to impaired nutrient utilization, reduced palatability, and compromised performance. However, recent studies indicate that IF inclusion does not necessarily negatively affect growth performance (Grześkowiak et al. 2022). Similarly, in the present study, IF sources (8% WB and 2.5% CIF or AS) did not impair performance compared with the NC group. This suggests that, under iso-energetic and iso-nitrogenous conditions, increasing dietary NDF by 1.8% does not necessarily compromise productive performance.

These findings are consistent with previous research. Slama et al. (2020) reported that the inclusion of soy hulls (2.5%) and a lignocellulosic product (1.5%) in diets for weaned piglets did not affect performance compared with a control diet without IF sources. Similarly, Pascoal et al. (2012) observed no impairment in growth when piglets were fed diets containing soy hulls (3%) and lignocellulosic material (1.5%). Other studies have evaluated cereal by-products as IF sources; for instance, Kraler et al. (2015) found that incorporating 15% WB had no negative impact on productive performance, whereas Yu et al. (2016) reported that diets containing WB, oat hulls, and soy hulls maintained performance compared with IF-free diets. Together with the present results, these findings indicate that low-to-moderate inclusion levels of IF-rich sources do not compromise growth performance in weaned piglets.

The digestibility results were consistent with the performance data, indicating that the inclusion of IF sources and the associated increase in dietary NDF did not negatively affect organic matter utilization under isoenergetic and isonitrogenous conditions. This observation agrees with Yu et al. (2016) and Slama et al. (2020), who similarly reported no differences in DM digestibility between diets containing IF sources and IF-free control diets.

Furthermore, previous studies have reported that IF inclusion reduces diarrhea incidence in weaned piglets (Flis et al. 2017). In the present study, AS reduced diarrhea incidence compared with the IF-free diet, whereas WB and CIF did not produce similarly pronounced effects. These findings may be partly related to differences in DF composition and particle size among the fiber sources. In particular, AS combined a higher IF content than WB with a coarser particle size than CIF, potentially integrating characteristics that contributed to the lower diarrhea incidence observed compared with the IF-free diet. Further support for this interpretation is provided by the other intestinal health indicators evaluated in the present study. Supporting the plausibility of this interpretation, Molist et al. (2012) observed that coarse WB (1,088 μm) at 4%, but not fine WB (445 μm), reduced diarrhea incidence (0.6 vs. 1.5) in weaned piglets challenged with enterotoxigenic E. coli.

One of the main indicators of intestinal function assessed in this study was jejunal gene expression. The greatest variation was observed within the growth factor family, where IGF1 (insulin-like growth factor) and EGF (epidermal growth factor) were upregulated in AS-fed piglets compared with WB-fed piglets, whereas IGF1R and EGFR remained unchanged. IGF1 acts as an endocrine, paracrine, and autocrine regulator of epithelial cell proliferation and survival in the gastrointestinal tract (Kuemmerle 2012; Zheng et al. 2018), intestinal IGF1 expression is largely influenced by luminal compounds (Bortvedt and Lund 2012). Likewise, EGF regulates cell growth, proliferation, differentiation, and survival (Arda-Pirincci and Bolkent 2014; Alexander et al. 2015) and contributes to intestinal maturation and epithelial homeostasis (Duh et al. 2000). Following receptor binding, EGF signaling has also been associated with apoptosis inhibition, inflammation modulation, tight junction regulation, and reduced enteric pathogen colonization (Lamb-Rosteski et al. 2008; Arda-Pirincci and Bolkent 2014; Chen et al. 2014; García‐Hernández et al. 2015).

As noted above, luminal compounds are among the main factors that may influence the expression of these growth factors. The IF supplied by AS, together with its larger particle size compared with WB and CIF, may explain the observed gene expression response. This was reflected in the particle size distribution: WB exhibited a D_3,2 value, indicating smaller particles, similar to CIF, but an intermediate D_4,3 value, indicating larger particles. Together with the S_gw value, these parameters reflected a high degree of heterogeneity, although WB still consisted of overall smaller particles than AS. The combined characteristics of AS, namely its higher IF content and coarse particle size, could therefore have exerted greater mechanical stimulation on the intestinal wall than WB. This mechanical stimulation of epithelial cells may have promoted IGF1 and EGF synthesis through autocrine or paracrine signaling, enhancing epithelial differentiation, proliferation, and maturation. This response may also be related to the greater SF content of WB, which is mainly composed of arabinoxylan (70%) and β-glucans (6%) (Merali et al. 2015). Supporting this interpretation, Hedemann et al. (2006) found that pectin inclusion, compared with barley hulls, reduced mitotic counts in the small intestine.

After 14 d of exposure to AS particles, intestinal adaptation may have required activation of signaling pathways involved in epithelial development. The absence of changes in IGF1R and EGFR expression suggests that regulation occurred primarily through ligand production rather than receptor abundance, consistent with the complex regulation of growth factor receptor signaling (Hernández-Sánchez et al. 1997).

Consistent with this, IL-22 expression was lower in AS-fed piglets than in WB-fed piglets, whereas no changes were observed in other inflammatory (IFN-γ, IL-10, IL-12p35, IL-12p40, IL-17A, IL-1β, IL-6, IL-8, TGF-β1, TLR-4, and TNF-α) or barrier-related genes (Occludin, Claudin 1, and ZO-1). As a key mediator of immune–epithelial communication (Keir et al. 2020), the reduction in IL-22 together with increased IGF1 and EGF expression suggests selective activation of growth and developmental pathways rather than enhanced immune or barrier responses, indicating a shift toward growth-related adaptation (Alexander et al. 2015; Zheng et al. 2018).

Another functional category analyzed included nutrient transport (PePT1, SGLT1, and BOAT1) and metabolism (ALPI, SI, and CCK), which remained unaffected by dietary treatment. These findings partially agree with Chen et al. (2020), who reported no effects of IF inclusion on PePT1 or GLUT2 expression, although SGLT1 was upregulated compared with diets differing in DF content.

Beneficial bacterial communities are closely associated with DF, which serves as an important substrate for microbial metabolism and modulates microbial community structure and activity (Jha and Berrocoso 2015). In the present study, microbial characterization based on relative abundance, alpha diversity, and beta diversity was performed in jejunal and colonic contents, revealing site-specific dietary effects.

In the jejunum, the inclusion of IF sources such as WB, CIF, and AS promoted favorable microbial shifts by reducing Proteobacteria abundance and increasing Firmicutes abundance. At the family and genus levels, WB showed the highest relative abundances of Lactobacillaceae and Lactobacillus compared with the NC diet, whereas CIF and AS showed intermediate values. These findings indicate that, although most microbial effects of DF have traditionally been described in the colon, substantial microbial modulation also occurs in the small intestine, a key ecosystem during the weaning period (Mahmud et al. 2023). Similar observations have been reported by Liu et al. (2018), who identified Firmicutes as the predominant phylum in the jejunum of piglets fed WB, corn bran, alfalfa meal, or a commercial lignocellulosic product. The increased abundance of Firmicutes is particularly relevant because this phylum includes beneficial bacteria involved in microbial metabolism, including butyrate producers associated with improved intestinal health and productive performance (Han et al. 2017). Furthermore, the shifts observed at the family and genus levels appeared to depend on fiber composition. The greater hemicellulose and SF content of WB, mainly arabinoxylans and β-glucans (Merali et al. 2015), may have promoted earlier fermentation in the small intestine and favored the proliferation of specific bacterial groups. In contrast, CIF and AS, characterized by higher IF and lower SF contents, produced intermediate microbial profiles between WB and NC.

Conversely, Proteobacteria abundance decreased when IF sources were included. This effect may reflect the ability of IF to modify the intestinal environment and limit the proliferation of potentially pathogenic bacteria. Similar findings have been reported by Gerritsen et al. (2012), who observed reduced ileal E. coli counts in piglets fed diets containing wheat straw and oat hulls, and by Li et al. (2020), who reported that IF inclusion in piglets challenged with enterotoxigenic E. coli promoted microbial homeostasis through reductions in proteolytic bacteria such as Enterobacteriaceae.

Alpha- and beta-diversity indices were also evaluated. Regarding alpha diversity, WB reduced richness (Observed, Chao1, and Fisher) and evenness (Shannon and Simpson) while increasing dominance (Berger–Parker). This pattern was likely driven by the greater abundance of Lactobacillaceae and Lactobacillus, suggesting that WB favored these taxa at the expense of others. Although greater microbial richness and evenness are generally associated with improved ecosystem stability through functional redundancy (Louca et al. 2018), the changes observed with WB reflected an enrichment of beneficial bacterial populations and may therefore still be considered positive.

Regarding beta diversity, Bray–Curtis and Jaccard analyses identified three distinct jejunal microbial clusters according to fiber inclusion and composition. The absence of a fiber source resulted in microbial communities that differed from those observed in the high-IF diets, likely due to differences in dietary NDF content (10.2 vs. 12.0%). Moreover, the higher SF and hemicellulose contents of WB compared with CIF and AS (SF: 3.97 vs. 2.62 and 2.71%; hemicellulose: 8.62 vs. 7.37 and 7.61%, respectively) further contributed to the distinct microbial communities observed among IF-supplemented diets.

The particle size of fiber sources plays a fundamental role in modulating the physiological effects exerted in the large intestine and its microbiota (Lannuzel et al. 2024). In the colonic content, CIF increased Proteobacteria abundance compared with NC, WB, and AS diets. Furthermore, CIF reduced bacterial richness (Observed, Chao1, and Fisher) without affecting evenness (Shannon and Simpson) or dominance (Berger–Parker). In line with this, McRorie and McKeown (2017) suggested that fine/smooth IF may exert a constipating effect in the large intestine, increasing the retention time available for nutrient fermentation, including protein fermentation, and thereby promoting the growth of Proteobacteria. Similarly, Koval et al. (2022) reported that fine cellulose concentrate worsened the colonic microbiome of rats, increasing Enterobacteriaceae, Proteus vulgaris, and Staphylococcus aureus. However, beneficial effects of fine cellulose concentrates have also been reported in weaned piglets (Uddin et al. 2023), suggesting that their effects on microbial composition and diversity are context dependent.

Regarding colonic microbial communities assessed through Bray–Curtis and Jaccard dissimilarity matrices, four distinct microbial clusters were identified. These differences appeared to be driven by IF inclusion (NC vs. IF-supplemented diets), fiber source (WB vs. the other IF sources), and particle size (CIF vs. AS). The lower NDF content of the NC diet generated a distinct microbial community, whereas the higher hemicellulose and SF contents of WB likely contributed to its own microbiota clustering from the other IF-supplemented diets. In agreement, Huang et al. (2024) reported that different fiber sources can differentially shape gastrointestinal microbial communities. Although CIF and AS were chemically similar, their marked differences in particle size likely contributed to the divergence of their microbial profiles (Tuncil et al. 2018).

Lastly, jejunal–colonic distances were evaluated to determine the similarity of microbial communities within the same animal. The absence of dietary effects on Bray–Curtis distances suggests that relative abundance patterns were comparable among treatments. However, Jaccard distances differed between WB- and NC-fed piglets, indicating that WB mainly affected the presence or absence of taxa rather than their relative abundances. These results suggest that WB promoted greater taxonomic turnover between the jejunum and colon, whereas AS and CIF showed intermediate values, likely reflecting their predominantly IF composition and lower capacity to induce differential taxon replacement than WB.

Conclusion

In conclusion, increasing DF intake through supplementation with different IF sources, including WB, commercial IF, or coarse ASs, did not affect growth performance or feed digestibility in weaned piglets but exerted beneficial effects on digestive health. Specifically, the inclusion of ASs reduced diarrhea incidence compared with an IF-free diet and improved jejunal gene expression by enhancing growth factor expression relative to WB inclusion. The type of IF source also influenced microbial modulation, likely owing not only to differences in SF content, as observed with WB but also to structural characteristics such as particle size, particularly in the case of ASs. These findings highlight the potential of ASs as a functional fiber source that modulates intestinal microbial communities and gene expression while maintaining growth performance, offering a sustainable alternative for piglet nutrition.

Glossary

Abbreviations

ADF

acid detergent fiber

ADFI

average daily feed intake

ADG

average daily gain

AIA

acid insoluble ash

ANOVA

one-way analysis of variance

AS

diet containing 2.5% milled almond shells

CATTD

coefficient of apparent total tract digestibility

CIF

diet containing 2.5% commercial IF source

d _gw

geometric mean diameter

d _0.1

10th percentile diameter

d _0.5

median diameter

d _0.9

90th percentile diameter

D _3,2

surface-weighted mean diameter

D _4,3

volume-weighted mean diameter

DF

dietary fiber

NDF

neutral detergent fiber

NE

net energy

G/F

gain-to-feed ratio

IF

insoluble fiber

IL

interleukin

NC

diet without fiber source added

PERMANOVA

permutational analysis of variance

SF

soluble fiber

S _gw

geometric standard deviation

TDF

total dietary fiber

WB

diet containing 8% wheat bran

Author contributions

Eduardo Ruiz (Conceptualization-True, Data Curation [true], Formal Analysis [true], Investigation [true], Methodology [true], Resources [true], Software [true], Validation [true], Visualization [true], Writing Original Draft [true], Writing Review Editing [true]), Silvia Martínez-Miró (Funding Acquisition [true], Investigation [true], Methodology [true], Resources [true], Supervision [true], Validation [true], Writing Review Editing [true]), Cristian Jesús Sánchez (Investigation [true], Methodology [true], Software [true], Validation [true], Visualization [true]), Lucía Ayala (Investigation [true], Methodology [true], Validation [true]), Ana Montalbán (Investigation [true], Methodology [true], Validation [true]), Josefa Madrid (Funding Acquisition [true], Investigation [true], Methodology [true], Resources [true], Validation [true], Writing Review Editing [true]), Nieves Ortega (Visualization [true], Writing Review Editing [true]), and Fuensanta Hernández (Funding Acquisition [true], Investigation [true], Methodology [true], Project Administration [true], Resources [true], Supervision [true], Validation [true], Visualization [true], Writing Review Editing [true])

Supplementary data

Supplementary data is available at Journal of Animal Science online.

Funding

This work was funded by the Centro para el Desarrollo Tecnológico Industrial, Ministry of Science and Innovation (Spain; IDI: 20220616).

Conflict of interest statement. The authors declare no known competing financial interests or personal relationships that could have influenced the work reported in this paper.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.

Ethical statement

All animal procedures were carried out under project authorization No. A13220603, approved by the Regional Ministry of Water, Agriculture, Livestock, and Fisheries of the Region of Murcia, and by the Ethics and Biosecurity Committee for Animal Experimentation at the University of Murcia, Spain.