Effects of grazing and housing system on growth performance, serum indices, rumen fermentation parameters, and microbial community in Gangba sheep

Abstract

This study investigated the effects of grazing and housing systems on growth performance, serum indices, rumen fermentation, and microbial composition of Gangba sheep in the high-altitude environment at the foot of the Himalayas. Thirty-six 4-month-old male Gangba sheep were assigned to grazing or housing systems, arranged as 9 replicates per treatment with 2 sheep per replicate, and the trial lasted 75 days (15 days of adaptation and 60 days of feeding). Compared with the grazing group, the housing system significantly increased final body weight and average daily gain (P < 0.05). Serum triglycerides, total antioxidant capacity, and glutathione peroxidase were higher, whereas total cholesterol, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol decreased (P < 0.05). Immune activation was reduced, as shown by lower immunoglobulins (IgA, IgG), interleukins (IL-2, IL-4, IL-6), tumor necrosis factor-α, and interferon-γ levels (P < 0.05), while growth hormone, growth hormone releasing hormone, and insulin-like growth factor 1 also declined (P < 0.05), indicating a nutritional feedback regulation of the growth hormone–insulin-like growth factor 1 axis. Rumen fermentation intensity increased under the housing system, with higher concentrations of ammonia nitrogen, acetate, propionate, butyrate, and total volatile fatty acids but lower pH (P < 0.05). Lipase and α-amylase activities increased, whereas cellulase activity declined (P < 0.05), suggesting a metabolic shift toward starch and lipid digestion. 16 S rRNA sequencing revealed decreased microbial α-diversity and an increased Firmicutes/Bacteroidetes ratio in the housing sheep, along with enrichment of UCG-001 and norank_o_RF39. In summary, the housing system improved growth performance, antioxidant status, and rumen fermentation efficiency while reducing immune activation and reshaping the rumen microbiota toward more energy-efficient metabolism.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-04936-6.

Introduction

Gangba sheep, a local Tibetan breed originating from Gangba County at the foot of the Himalayas, have been domesticated since the 7th century [1, 2]. Similar to most Tibetan sheep production systems [3, 4], traditional grazing remains the dominant feeding practice. However, as the demand for sustainable livestock production and animal health management increases, the comparison between traditional grazing and modern housing systems has gained attention, although studies conducted under high-altitude conditions are still limited.

The Qinghai–Tibet Plateau (QTP) is characterized by extremely cold temperatures, hypoxia, strong ultraviolet (UV) radiation, and seasonal forage shortages. These environmental stresses often result in reduced growth rates and productivity in ruminants compared with lowland regions [5–7]. Such disadvantages are largely attributed to multiple physiological constraints involving growth, metabolism, immune function, and endocrine regulation [8, 9]. Compared with lowland sheep, ruminants raised in high-altitude Himalayan regions face distinct physiological and nutritional constraints [10, 11]. Chronic hypoxia and harsh climatic conditions may increase maintenance energy expenditure, intensify oxidative stress, and enhance immune activation, thereby limiting nutrient allocation toward growth and production [12, 13]. Consequently, the growth performance, metabolic responses, and rumen microbial adaptation of plateau sheep cannot be interpreted solely based on evidence from lowland systems. Understanding how feeding systems interact with these altitude-specific constraints is therefore critical for optimizing nutritional management in high-altitude sheep production. Transitioning from grazing to housing systems alters not only feed composition and intake patterns but also animal activity, population density, and management conditions. These changes can markedly influence growth performance, rumen fermentation, antioxidant status, and immune responses [14, 15]. Previous studies have shown that diet structure strongly affects rumen microbial community composition, thereby regulating nutrient utilization efficiency and host health [16, 17]. Moreover, the rumen microbiota contributes to energy harvesting, volatile fatty acids (VFAs) production, nitrogen metabolism, immune modulation, and oxidative stress balance, which may differ substantially between feeding systems [18–20].

Gangba sheep possess a unique genetic background and strong adaptability to high-altitude environments [1]. This makes them an ideal model for evaluating how feeding systems influence the physiological and microbial mechanisms of adaptation in plateau ruminants [21, 22]. Therefore, the objective of this study was to compare the effects of grazing and housing feeding systems on growth performance, serum antioxidant and immune indices, hormone levels, and rumen microbial community structure and function in Gangba sheep. Furthermore, we aimed to elucidate the potential host–microbiome interactions associated with their adaptive responses to different feeding systems. The results provide new insights into the physiological adaptation of Gangba sheep and offer a scientific basis for improving sustainable livestock production in high-altitude pastoral regions.

Materials and methods

Experimental site, animals, and management

The experiment was conducted from May to July at the Gangba Sheep Cooperative in Mendecun Village, Gangba County, Tibet, China (elevation: 4,700 m; 88.418°E, 28.262°N), located on the northern slope of the central Himalayas. The dominant forage species in the alpine pasture include Kobresia pygmaea, Carex moorcroftii, Festuca ovina, and Stipa purpurea, typical perennial grasses and sedges of the Qinghai–Tibet Plateau. The experiment was conducted during the growing season, when pasture biomass and nutritional value are relatively stable. During this period, alpine pasture generally contains approximately 8–12% crude protein (CP), 55–65% neutral detergent fiber (NDF), and 30–35% acid detergent fiber (ADF), representing a high-fiber, moderate-protein forage structure [15, 23, 24]. A total of 36 healthy, four-month-old male Gangba sheep (initial body weight = 16.44 ± 0.61 kg) were purchased from the Gangba Sheep Cooperative in Mendecun Village and individually identified with ear tags. All animals were clinically healthy at the beginning of the experiment and were subjected to routine veterinary supervision throughout the trial. Health status was monitored daily by trained personnel to detect any signs of illness or abnormal behavior. No clinical disease symptoms were observed during the experimental period. The animals were randomly assigned to two feeding systems. In the grazing group (9 replicates), every two sheep sharing the same identification number were treated as one replicate and grazed on natural pasture daily from 09:00 to 18:00. Due to the extensive and nomadic grazing pattern in Gangba County, characterized by large grazing areas and uneven forage distribution, individual daily feed intake of grazing sheep could not be accurately quantified under practical field conditions. In the housing group (9 replicates), every two sheep kept in the same 1 m × 1 m pen were considered one replicate. For the housing group, diets were provided twice daily at 09:00 and 18:00, and all animals had ad libitum access to forage and water. The ingredient composition and nutrient levels of the experimental diets are presented in Table 1. The experiment lasted 75 days, consisting of a 15-day adaptation period followed by 60 days of feeding. Daily feed intake was recorded to calculate the average daily feed intake (ADFI) for the housing group.

Table 1.

Composition and nutrient levels of the experimental diets (air-dry basis, %)

Feed Composition		Nutrient Levels2
Oat Hay	52.00	Crude protein	12.77
Corn	28.00	Crude fat	1.57
Soybean meal	13.00	Neutral detergent fiber	47.22
Wheat bran	5.00	Acid detergent fiber	69.64
Limestone	0.50	Organic matter	90.91
Premix1	1.00	Ash	9.09
Salt	0.50
Total	100

¹ Premix (per kilogram content): VA 700,000–1,000,000 IU, VE 4,000 IU, VK3 100 mg, VB1 100 mg, VB2 260 mg, VB12 1,000 µg, VB6 100 mg, VD3 200,000–500,000 IU, niacin 1,600 mg, pantothenic acid 500 mg, folic acid 50 mg, copper 0.2–1.5 g, iron 3–20 g, zinc 2–12 g, biotin 10 mg, manganese 2–14 g, iodine 10–100 mg, selenium 10–50 mg, cobalt 5–40 mg

² Nutrient levels were measured

Growth performance measurement

Body weight, height, body length, and chest circumference of all sheep were measured on day 1 and day 60. Average daily gain (ADG) was calculated as the difference between initial and final body weight divided by feeding days.

Sample collection

At the end of the trial, 9 sheep in each group were fasted for 24 h prior to blood and rumen fluid collection to minimize short-term postprandial variation and standardize metabolic conditions before sampling. Specifically, blood samples were collected via jugular venipuncture using vacuum tubes, centrifuged at 3,000 × g for 10 min at 4 °C, and serum samples were stored at − 80 °C. Rumen fluid was collected using an oral stomach tube before morning feeding to allow non-invasive sampling under practical production conditions. To minimize saliva contamination, the initial portion of rumen fluid was discarded, and samples were collected from the midstream fraction as previously described [25, 26]. The collected rumen fluid was filtered through four layers of gauze and immediately used for pH measurement with a portable pH meter (HORIBA Advanced Techno Co., Ltd., Osaka, Japan). Approximately 20 mL was stored at − 80 °C for later analyses of fermentation characteristics, enzyme activities, and microbial composition. After homogenization, 10 mL was used for fermentation parameter and enzyme activity analyses, 5 mL was used for 16 S rRNA gene sequencing, and the remaining volume was stored as backup for potential further analyses.

Serum biochemical, antioxidant, immune, and hormonal analyses

Serum biochemical, antioxidant, immune, and hormonal indices were determined using commercial assay kits following manufacturer instructions. Biochemical parameters included alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C), measured using an automatic biochemical analyzer (AU680, Beckman Coulter, CA, USA). Antioxidant parameters included total antioxidant capacity (T-AOC), glutathione peroxidase (GSH-Px), malondialdehyde (MDA), and superoxide dismutase (SOD). Immune indices, including immunoglobulins (IgA, IgG, and IgM), tumor necrosis factor-α (TNF-α), interleukins (IL-2, IL-4, and IL-6), and interferon-γ (IFN-γ), as well as endocrine parameters, including growth hormone (GH), growth hormone–releasing hormone (GHRH), insulin-like growth factor 1 (IGF-1), and myostatin (MSTN), were determined using enzyme-linked immunosorbent assay (ELISA) kits specific for ovine species.

Rumen fermentation characteristics and enzyme activities

Rumen ammonia nitrogen (NH₃-N) concentration was determined by colorimetric assay using a UV spectrophotometer (UV-2450, Shimadzu, Japan). VFAs, including, acetate, propionate, isobutyrate, butyrate, valerate, and isovalerate, were quantified by gas chromatography (7890 A, Agilent Technologies, Santa Clara, CA, USA) as previously described [15]. The activities of cellulase, lysozyme, neutral xylanase, pectinase, neutral protease, lipase, and α-amylase were determined using commercial microplate colorimetric assay kits according to the manufacturer’s instructions (Shanghai Jining Bioengineering Institute, Shanghai, China).

Rumen fluid microbiota

Genomic DNA was extracted from rumen fluid using the E.Z.N.A.^® Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA). DNA integrity was verified by 1% agarose gel electrophoresis. The V3–V4 hypervariable region of the bacterial 16 S rRNA gene was amplified using primers 338 F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) [15]. PCR amplicons were purified and subjected to paired-end sequencing on the Illumina MiSeq PE300 platform (Illumina, San Diego, USA). Raw reads were processed using UPARSE (v7.0) to cluster operational taxonomic units (OTUs) at 97% similarity after chimera removal. Taxonomic classification was performed with the ribosomal database project (RDP) Classifier (v2.11) against the 16 S rRNA reference database using a confidence threshold of 0.7.

Bioinformatic and statistical analyses were conducted on the Majorbio Cloud Platform (https://cloud.majorbio.com). Alpha-diversity indices (Ace, Chao1, Shannon) were calculated using Mothur (v1.30.2). Beta-diversity was assessed by principal coordinate analysis (PCoA) based on Bray–Curtis distances, and group differences were evaluated by PERMANOVA. Differential taxa were identified at the phylum and genus levels using the Wilcoxon rank-sum test. Functional phenotypes were predicted using BugBase, and microbial–metabolite correlations were analyzed by Spearman’s correlation (|R| > 0.6, P < 0.05).

Statistical analysis

Treatment differences were analyzed using the test procedures of JMP software (JMP, version 10; SAS Institute Inc., Cary, NC). Differences were considered significant if P ≤ 0.05 and as a trend if 0.05 < P ≤ 0.10 (Table 1).

Results

Growth performance

Feeding system had a significant effect on the growth performance of Gangba sheep (Table 2). Compared with the grazing group, sheep in the housing group had higher final body weight, body height, body length, chest circumference, and average daily gain (ADG) (P < 0.05). The final body weight of the Gangba sheep in the housing group increased by 28.46%, and ADG increased by 156.81% relative to grazing sheep.

Table 2.

Effects of different feeding systems on growth performance of Gangba sheep

Items	Feeding system (Mean ± SD)		P-value
	Grazing system	Housing system
Initial Body weight, kg	16.28 ± 0.61	16.21 ± 0.94	0.852
Initial Body height, cm	50.30 ± 3.46	50.25 ± 4.03	0.905
Initial Body length, cm	51.11 ± 4.28	51.44 ± 2.92	0.850
Initial Chest circumference, cm	65.78 ± 2.17	65.89 ± 1.96	0.911
Final Body weight, kg	19.96 ± 1.06	25.64 ± 0.82	< 0.001
Final Body height, cm	57.67 ± 2.74	60.78 ± 2.11	0.016
Final Body length, cm	65.22 ± 4.82	70.33 ± 3.28	0.003
Final Chest circumference, cm	71.22 ± 3.93	76.22 ± 2.05	0.004
Average daily gain, g/d	61.22 ± 13.98	157.22 ± 11.64	< 0.001
Average Daily Feed Intake, kg/d	--	1287.04 ± 48.32

Differences are considered significant when P < 0.05, show a tendency when 0.05 < P < 0.10, and are not significant when P > 0.10

Serum biochemical and antioxidant parameters

Differences in feeding systems markedly influenced serum biochemical and antioxidant indicators (Table 3). Sheep in the housing group showed higher serum concentrations of TG, T-AOC, and GSH-Px (P < 0.05), whereas TC, HDL-C, LDL-C, AST, and ALP were significantly lower than those in the grazing group (P < 0.05).

Table 3.

Effects of different feeding systems on serum biochemical and antioxidant parameters in Gangba sheep

Items	Feeding system (Mean ± SD)		P-value
	Grazing system	Housing system
Total Cholesterol, mmol/L	2.05 ± 0.53	1.59 ± 0.24	0.032
Triglyceride, mmol/L	0.16 ± 0.06	0.36 ± 0.08	<0.001
High density lipoprotein cholesterol, mmol/L	1.10 ± 0.27	0.58 ± 0.24	<0.001
Low density lipoprotein cholesterol, mmol/L	0.61 ± 0.17	0.44 ± 0.12	0.035
Aspartate Aminotransferase, U/L	36.67 ± 8.92	26.60 ± 7.30	0.021
Alkaline Phosphatase, U/L	142.03 ± 13.27	102.27 ± 8.86	<0.001
Total Superoxide Dismutase, U/mL	1.49 ± 0.23	1.42 ± 0.15	0.468
Total Antioxidant Capacity, U/mL	9.77 ± 2.13	20.66 ± 5.22	<0.001
Malondialdehyde, nmol/mL	1.02 ± 0.11	0.93 ± 0.43	0.575
Glutathione Peroxidase, U/mL	7.85 ± 1.73	12.95 ± 2.21	<0.001

Differences are considered significant when P < 0.05, show a tendency when 0.05 < P < 0.10, and are not significant when P > 0.10

Serum immune and hormonal indices

Feeding system significantly affected immune function and endocrine status of Gangba sheep (Table 4). Compared with grazing, the housing system resulted in lower serum levels of IgA, IgG, IL-2, IL-4, IL-6, TNF-α, and IFN-γ (P < 0.05). IgM tended to be lower in the housing group (P = 0.0562). In addition, serum concentrations of GH, GHRH, IGF-1, and MSTN were all significantly lower in the housing group compared with grazing sheep (P < 0.05).

Table 4.

Effects of different feeding systems on serum immune and hormonal parameters in Gangba sheep

Items	Feeding system (Mean ± SD)		P-value
	Grazing system	Housing system
Immunoglobulin A, g/L	4.76 ± 1.25	3.29 ± 0.76	< 0.001
Immunoglobulin M, g/L	4.38 ± 0.73	3.65 ± 0.74	0.056
Immunoglobulin G, g/L	3.30 ± 0.78	2.55 ± 0.35	0.019
Interleukin 2, pg/mL	190.11 ± 37.75	142.75 ± 38.62	0.018
Interleukin 4, pg/mL	26.82 ± 5.43	15.00 ± 2.38	< 0.001
Interleukin 6, pg/mL	861.72 ± 98.68	630.35 ± 87.75	0.006
Tumor Necrosis Factor α, pg/mL	498.29 ± 85.18	398.36 ± 64.74	0.014
Interferon γ, pg/mL	231.09 ± 46.38	123.33 ± 24.44	< 0.001
Growth Hormone, ng/mL	12.14 ± 3.00	6.92 ± 2.03	< 0.001
Growth Hormone Releasing Hormone, pg/mL	20.50 ± 3.75	10.92 ± 2.38	< 0.001
Insulin-like Growth Factor 1, ng/mL	217.60 ± 43.24	146.19 ± 33.22	0.002
Myostatin, ng/mL	29.19 ± 8.65	20.80 ± 4.35	0.023

Differences are considered significant when P < 0.05, show a tendency when 0.05 < P < 0.10, and are not significant when P > 0.10

Rumen fermentation parameters

Feeding system significantly affected rumen fermentation parameters (Table 5). The housing system decreased ruminal pH (P < 0.05) but increased concentrations of NH₃–N, acetic acid, propionic acid, butyric acid, and total VFAs (P < 0.05). The molar ratio of acetate to propionate (A/P) and the proportion of acetate were higher in the housing group (P < 0.05). Conversely, the proportions of iso-butyric acid, isovaleric acid, and valeric acid were lower (P < 0.05), and the proportion of butyric acid tended to decrease (P = 0.0940).

Table 5.

Effects of different feeding systems on rumen fermentation parameters in Gangba sheep

Items	Feeding system (Mean ± SD)		P-value
	Grazing system	Housing system
pH	6.74 ± 0.17	6.17 ± 0.29	< 0.001
NH₃-N1, mg/g	62.33 ± 18.09	132.07 ± 10.46	< 0.001
Acetic acid, mmol/L	22.76 ± 1.96	52.83 ± 6.96	< 0.001
Propionic acid, mmol/L	6.12 ± 1.02	11.84 ± 1.84	< 0.001
iso-butyric acid, mmol/L	1.26 ± 0.30	1.27 ± 0.19	0.956
Butyric acid, mmol/L	3.54 ± 0.68	6.58 ± 1.43	< 0.001
iso-valeric acid, mmol/L	1.90 ± 0.26	1.84 ± 0.31	0.643
Valeric acid, mmol/L	0.55 ± 0.10	0.57 ± 0.12	0.611
Total VFAs, mmol/L	36.15 ± 3.11	74.93 ± 10.12	< 0.001
A/P2	3.78 ± 0.46	4.49 ± 0.027	< 0.001
Acetic acid, %	62.99 ± 1.37	70.55 ± 1.24	< 0.001
Propionic acid, %	16.90 ± 2.05	15.77 ± 0.78	0.143
iso-butyric acid, %	3.54 ± 0.95	1.72 ± 0.28	< 0.001
Butyric acid, %	9.78 ± 1.48	8.72 ± 1.03	0.094
iso-valeric acid, %	5.28 ± 0.70	2.49 ± 0.50	< 0.001
Valeric acid, %	1.51 ± 0.24	0.76 ± 0.08	< 0.001

Differences are considered significant when P < 0.05, show a tendency when 0.05 < P < 0.10, and are not significant when P > 0.10

¹ NH₃-N, ammonia nitrogen

² A/P, the molar ratio of acetate to propionate

Rumen enzyme activities

Rumen enzyme activities also varied with feeding system (Table 6). The housing system significantly decreased the activity of cellulase (P < 0.05), while the activity of lipase and α-amylase increased markedly (P < 0.05). Neutral protease activity tended to increase (P = 0.0595).

Table 6.

Effects of different feeding systems on rumen enzyme activities in Gangba sheep

Items	Feeding system (Mean ± SD)		P-value
	Grazing system	Housing system
Cellulase, µg/min/mL	15.73 ± 3.50	11.67 ± 2.27	0.001
Lysozyme, µg/mL	28.02 ± 11.23	21.65 ± 7.58	0.177
Neutral protease, nmol/min/mL	3.44 ± 0.87	4.64 ± 1.55	0.060
Lipase, U/L	1.27 ± 0.22	2.01 ± 0.28	< 0.001
α-Amylase, µU/L	10.40 ± 2.75	41.04 ± 14.72	< 0.001

Differences are considered significant when P < 0.05, show a tendency when 0.05 < P < 0.10, and are not significant when P > 0.10

Rumen microbial diversity and community composition

A total of 1,215,104 high-quality sequences were obtained from 18 rumen fluid samples, with an average read length of 417 bp. As shown in Fig. 1A–C, α-diversity indices (Ace, Chao1, and Shannon) were significantly lower in the housing group than in the grazing group (P < 0.05), indicating reduced microbial richness and diversity under the housing system. Principal coordinate analysis (PCoA) based on Bray–Curtis distances showed clear separation between groups (R = 0.8066, P = 0.0010; Fig. 1D), confirming that feeding system significantly altered the rumen microbial community structure. At the phylum level, the rumen microbiota was dominated by Bacteroidetes and Firmicutes in both groups. The relative abundance of Firmicutes increased, while Bacteroidetes decreased in the Gangba sheep of the housing system, leading to a higher Firmicutes/Bacteroidetes (F/B) ratio (P < 0.05; Figure S1A–C). At the genus level, four genera differed significantly between treatments (Figure S2A–D). The housing system decreased the relative abundances of Prevotella and Selenomonas but increased UCG-001 and norank_o_RF39 (P < 0.05). Linear discriminant analysis effect size (LEfSe) analysis identified a total of 74 taxa with significant differences linear discriminant analysis (LDA > 3), of which 38 were enriched in grazing and 36 in the sheep of the housing system (Fig. 2A–B).

Fig. 1.

Effects of feeding system on the rumen microbiota of Gangba sheep. (A) Ace index; (B) Chao1 index; (C) Shannon index; (D) PCoA chart; (E) Phylum level bar chart; (F) Genus level bar chart

Fig. 2.

Lefse multi-level species differential discrimination analysis chart; (A) LefSe cladogram; (B) LEfSe Bar

Predicted microbial phenotypes

According to BugBase functional prediction (Fig. 3A), the housing system reduced the proportion of anaerobic phenotypes but increased aerobic and facultative taxa (P < 0.05). In the grazing group, anaerobic phenotypes were mainly contributed by Prevotella, Selenomonas, and norank_o_RF39, while aerobic phenotypes in the housing group were dominated by norank_p_WPS-2, Acinetobacter, norank_o_Coriobacteriales, and norank_f_vadinBE97 (Fig. 3B–C).

Fig. 3.

BugBase phenotype prediction. (A) Comparison of predicted phenotypic profiles between groups; (B) Taxonomic contribution to the Anaerobic phenotype; (C) Taxonomic contribution to the Aerobic phenotype

Correlations between key rumen microbes, metabolites, and enzyme activities

Spearman correlation analysis revealed several significant associations between dominant rumen genera, fermentation parameters, and enzyme activities (Fig. 4). Norank_o_RF39 was positively correlated with NH₃–N (R = 0.68, P = 0.0018), butyrate (R = 0.63, P = 0.0055), and α-amylase activity (R = 0.69, P = 0.0017). Selenomonas showed a positive correlation with cellulase activity (R = 0.69, P = 0.0017) but negative correlations with NH₃–N (R = − 0.65, P = 0.0032), lipase (R = − 0.75, P = 0.0004), and α-amylase activity (R = − 0.65, P = 0.0035). Candidatus_Saccharimonas was positively associated with acetate (R = 0.63, P = 0.0047), propionate (R = 0.67, P = 0.0022), and neutral protease activity (R = 0.61, P = 0.0070). UCG-001 correlated positively with propionate (R = 0.63, P = 0.0050) and butyrate (R = 0.63, P = 0.0055), and norank_o_Clostridia_UCG-014 showed a similar positive correlation with butyrate (R = 0.63, P = 0.0050). In contrast, Prevotella was negatively correlated with acetate (R = − 0.75, P = 0.0003), propionate (R = − 0.74, P = 0.0004), butyrate (R = − 0.78, P = 0.0001), lipase (R = − 0.60, P = 0.0080), and α-amylase activity (R = − 0.67, P = 0.0022).

Fig. 4.

Spearman correlation analysis between rumen microorganisms and key fermentation parameters and enzyme activities

Discussion

Feeding system markedly influenced the growth performance, physiological metabolism, and rumen microbial ecology of Gangba sheep. This study compared animals raised under natural grazing versus housing feeding systems in a high-altitude environment. The results demonstrated that the feeding system not only altered nutrient intake and fermentation efficiency but also modulated immune and endocrine responses as well as microbial community structure, reflecting a complex host–microbiome adaptation to environmental and nutritional changes. These responses should be interpreted as integrated adaptive outcomes shaped by both feeding management and altitude-associated environmental constraints characteristic of plateau production systems.

Growth performance

Feeding systems were first reflected in differences in growth performance. Many studies have reported that, compared with traditional grazing, indoor feeding generally provides diets with higher levels of energy and protein, thereby promoting faster growth and higher weight gain in sheep and yak [27]. This is consistent with the significantly higher ADG of Gangba sheep observed under the housing system in our study. Normally, diets in the housing system are always rich in nutrients and more palatable, leading to increased feed intake and energy absorption that support faster weight gain. Meanwhile, indoor housing avoids energy expenditure during grazing, allowing more nutrients to be used for growth. However, growth advantages also depend on pasture conditions and seasons. During periods of abundant herbage, grazing sheep can obtain diverse nutrients from natural forages, and their growth rate is not necessarily lower than that of housing animals. For example, a study on yaks found that during the summer grazing season, grazing yaks had significantly higher final body weight and ADG than stall-fed yaks [15]. This may be because summer pastures at high altitude have higher nutritional value, and grazing activity increases feed intake and digestibility. Overall, under sufficient nutrient supply, the housing system promotes faster growth of Gangba sheep, but well-managed grazing can also maintain high weight gain, with the differences jointly influenced by pasture quality and management. These findings should be interpreted within the context of practical production systems under plateau conditions rather than as the isolated effect of diet composition alone. In Gangba County, natural pastures are unevenly distributed and grazing areas are extensive, requiring sheep to travel considerable distances daily to obtain forage. Under high-altitude conditions characterized by hypoxia and low ambient temperatures, such locomotion-associated energy expenditure may increase maintenance energy requirements, thereby limiting nutrient allocation toward growth performance. Such variation further indicates that growth responses in plateau sheep are jointly determined by nutrient availability, locomotion-related energy expenditure, and seasonal forage dynamics under high-altitude grazing conditions.

Biochemical and antioxidant indices

Changes in feeding systems can lead to alterations in metabolic level and antioxidant status of Gangba sheep. First, regarding energy metabolism, housing diets are rich in rapidly fermentable carbohydrates, which increase blood glucose and related metabolites [27, 28]. High-concentrate diets produce more propionate, which after ruminal absorption enhances hepatic gluconeogenesis and increases blood glucose concentration [29]. These findings explain the relatively higher serum glucose observed in the Gangba sheep of the housing group in our study. Secondly, protein metabolism also differed. In this experiment, grazing sheep had slightly higher serum total protein and albumin concentrations than housing sheep, possibly because grazing animals were leaner and plasma proteins were more concentrated, or because exposure to environmental antigens stimulated immunoglobulin synthesis, elevating total protein. This agrees with reports that summer grazing in yaks reduced plasma urea nitrogen and improved protein utilization efficiency [15].

In addition, antioxidant indices differed significantly between systems. Gangba sheep in housing group exhibited stronger antioxidant capacity, likely due to higher intake of antioxidant nutrients (selenium, vitamins, adequate protein), more stable feeding rhythm and diet quality, and lower oxidative stress. These metabolic adjustments may represent adaptive energy allocation strategies that allow plateau sheep to balance oxidative stress resistance with growth demands under contrasting management systems. Collectively, these responses contributed to the improved antioxidant capacity observed in the housing group [13, 30]. In summary, feeding systems regulate metabolic level and antioxidant defense in Gangba sheep through effects on nutrient intake and physical stress.

Serum immune and hormone levels

Grazing and housing system also induced adaptive changes in immune function and endocrine regulation. Regarding immune factors, we observed differences in serum immunoglobulin concentrations between the two groups. The housing group, living in cleaner environments and exposed to fewer pathogens, had lower baseline immunity, while the grazing group, exposed to more environmental antigens, exhibited moderate immune activation and higher immunoglobulin levels [11, 31]. However, differences in immune indices may also be influenced by environmental antigen exposure and hygiene conditions inherent to grazing systems, rather than nutritional factors alone [12]. All animals remained clinically healthy throughout the experimental period, suggesting that the observed immune differences were more likely associated with environmental exposure and feeding management rather than disease occurrence. Previous studies confirmed that sheep raised exclusively indoors had significantly lower serum IgA, IgG, and IgM levels than those allowed several hours of grazing each day [31]. In our experiment, grazing Gangba sheep had relatively higher total Ig levels, suggesting that moderate pathogen stimulation and exercise enhance humoral immunity. Similarly, in yaks under different feeding regimes, the grazing group showed higher IgA, IgG, and IgM concentrations than the housed group [15]. However, changes in immune factors also involve shifts in immune response patterns. In grazing yaks, not only were immunoglobulins elevated, but proinflammatory cytokines such as IL-2, IL-6, TNF-α, and IFN-γ also increased, while anti-inflammatory cytokines (IL-4, IL-10) decreased, indicating that grazing activated cellular immunity and enhanced stress and infection resistance, likely biasing immune responses toward a Th1-type pattern. This response benefits the clearance of environmental pathogens during grazing but may also impose higher metabolic cost, requiring sufficient nutrition to prevent excessive immune stress [32, 33].

Although high-energy housing diets typically promote anabolic metabolism through the insulin–IGF-1 pathway, in this study the serum GH, GHRH, and IGF-1 concentrations of Gangba sheep in the housing group were significantly lower than those of grazing animals, suggesting that nutrient abundance did not further activate the growth axis but may have induced a nutritional GH–IGF-1 feedback inhibition [34–36]. Grazing sheep, facing relative energy deficiency and environmental stress, instead increased GH secretion through activation of the pituitary–liver axis to mobilize energy and maintain homeostasis [29]. Meanwhile, lower MSTN in the Gangba sheep of the housing system suggests a metabolic tendency toward fat deposition and energy storage, while grazing sheep maintained stronger protein synthesis signaling [37]. Overall, feeding system jointly shapes the bidirectional regulation of the GH–IGF-1–MSTN endocrine network through energy intake and physical activity. In conclusion, feeding systems affect antigen exposure and nutritional balance, thereby shaping the immune strategies of Gangba sheep: grazing enhances immune vigilance and pathogen resistance, while the housing system provides high nutrition with a relatively low immune activation state. Such immune modulation may reflect an adaptive trade-off between pathogen exposure and metabolic efficiency under different plateau management environments.

Rumen fermentation parameters

Feeding systems alter dietary composition and thus significantly affect rumen fermentation patterns. Sheep in the housing system usually consume a higher proportion of concentrates, and the increased starch content leads to higher total VFAs and lower rumen pH [27, 29, 38]. In contrast, grazing sheep primarily consume high-fiber forage, resulting in lower total VFA but near-neutral pH, which favors the activity of fibrolytic bacteria and rumen stability [29]. Given that grazing intake could not be precisely quantified, variations in substrate availability between systems may also have contributed to the observed differences in fermentation profiles. Therefore, fermentation responses likely reflect the combined influence of diet composition, intake pattern, and activity-associated energy demand rather than feeding system alone.

Propionate and acetate play distinct roles in host metabolism. Propionate serves as the main precursor for hepatic gluconeogenesis in ruminants, providing energy via glucose synthesis, while acetate is mainly used for fatty acid and cholesterol synthesis [39, 40]. Moreover, butyrate is the primary energy source for rumen epithelium and contributes to epithelial barrier integrity and anti-inflammatory homeostasis [41, 42]. In this study, feeding system also altered the molar ratios of VFAs: grazing diets rich in structural carbohydrates led to higher acetate proportion, while high-concentrate diets in the housing system increased propionate content and the acetate/propionate ratio, facilitating gluconeogenesis and energy deposition [43]. The increase of butyrate further benefited energy storage and immune function. Nitrogen metabolism was also affected. Diets in the housing system generally have higher protein levels, leading to increased ruminal NH₃–N. In contrast, low-protein grazing diets resulted in lower NH₃–N levels, possibly limiting microbial growth and protein synthesis efficiency [27, 29].

Rumen enzyme activities

Feeding systems also indirectly affected digestive enzyme activity profiles by altering rumen microbial composition. Fibrolytic bacteria such as Butyrivibrio, Rikenellaceae_RC9, Prevotella, and Selenomonas were enriched in grazing sheep; these microorganisms degrade plant cell-wall polysaccharides—especially hemicellulose, pectin, and soluble sugars—through secretion of xylanase, pectinase, and polysaccharide hydrolases, thereby enhancing fiber utilization efficiency [44, 45]. UCG-001 and norank_o_RF39 were enriched in housing group and positively correlated with the activity of digestive enzymes such as neutral protease, lipase, and α-amylase, suggesting key roles in nutrient enzyme expression in the rumen [46, 47]. Overall, grazing and housing system reshaped the rumen microbial community and digestive physiology, leading to redistribution of enzyme activity and changes in nutrient utilization efficiency, representing adaptive responses to distinct feeding environments.

Rumen microbial community

In this study, grazing and housing system significantly affected the rumen bacterial community of Gangba sheep. Rumen bacterial α-diversity was higher in grazing sheep, indicating more abundant microbial populations under natural conditions. In both systems, the dominant phyla were Firmicutes and Bacteroidetes. Notably, the F/B ratio was significantly higher in the housing group, which may be associated with enhanced energy capture and metabolic adaptation under housing conditions, promoting better growth performance [48, 49]. Such microbial restructuring may contribute to differential energy-harvesting strategies supporting physiological adaptation under plateau production environments. At the genus level, Prevotella and Selenomonas were enriched in grazing sheep and involved in the degradation of hemicellulose, pectin, and soluble carbohydrates, cooperating with cellulolytic bacteria to enhance total fiber-degrading enzyme activity and fiber utilization efficiency [44, 45]. UCG-001 and norank_o_RF39 were enriched in the housing group and positively correlated with digestive enzyme activities such as protease, lipase, and α-amylase, indicating their participation in efficient nutrient metabolism [46, 47].

From the perspective of functional phenotypes, different feeding systems significantly changed oxygen demand and metabolic traits of rumen bacteria. The grazing rumen, with a stable and low redox potential environment, was enriched in typical strict anaerobes such as Christensenellaceae_R-7_group, Ruminococcaceae, Oscillospiraceae, Clostridia_UCG-014, and Prevotellaceae_UCG-004, which are responsible for cellulose and hemicellulose degradation and butyrate and acetate production—key functional groups maintaining anaerobic stability [44, 50, 51]. The dominance of these anaerobic fermenters was consistent with the BugBase prediction showing higher anaerobic phenotypes in the grazing group, suggesting a more reductive and stable rumen environment. In contrast, the rumen of housing group enriched some facultative or microaerophilic taxa (such as Lactobacillus and Treponema), though the main groups remained anaerobic fermenters from Firmicutes and Bacteroidetes, including Phascolarctobacterium, Lachnospiraceae_AC2044_group, Clostridium_sensu_stricto_1, and Ruminococcus. These bacteria utilize lactate and succinate as intermediates to produce acetate, propionate, and butyrate, reflecting a more active and efficient fermentative metabolism [17, 42]. Thus, the functional differences mainly represented a microbial transition from anaerobic stability to facultative adaptation: grazing rumens were dominated by fiber fermentation and acetate/butyrate metabolism, while housing rumen favored lactate and succinate pathways for rapid propionate production [42, 52, 53]. This metabolic divergence not only affected rumen pH and VFA profiles but also regulated host energy metabolism and hormonal levels through VFA signaling [39, 54, 55].

Correlations between rumen microbes and metabolites

Spearman correlation analysis further revealed potential interactions between rumen core bacteria and key metabolites. Norank_o_RF39 was positively correlated with NH₃–N, butyrate, and α-amylase activity, indicating multiple substrate metabolism capabilities related to protein degradation and starch utilization. Selenomonas showed positive correlation with cellulase activity, but negative correlations with NH₃–N, lipase, and α-amylase, reflecting its preference for fibrous substrates and limited function in concentrate-rich diets [56, 57]. Candidatus_Saccharimonas showed positive correlations with acetate, propionate, and neutral protease activity, suggesting involvement in synergistic carbohydrate–protein metabolism [58]. UCG-001 and norank_o_Clostridia_UCG-014 were positively correlated with butyrate, indicating typical butyrate-producing activity.

Conversely, Prevotella showed negative correlations with acetate, propionate, butyrate, lipase, and α-amylase activities, implying that high Prevotella abundance may suppress other VFA-producing bacteria, reducing total fermentation efficiency. These correlations indicate that microbial taxa have distinct substrate preferences and that their metabolic products (VFAs, enzymes) interact through community networks to regulate rumen fermentation [44, 59]. This ecological process provides functional evidence for understanding microbe–metabolite interactions and highlights the potential value of manipulating rumen structure to optimize energy metabolism.

Overall, compared with grazing group, Gangba sheep in housing system exhibited remarkable advantages in several aspects. They showed higher body weight and ADG, enhanced antioxidant capacity, and lower immune activation, suggesting improved health status. Rumen fermentation was more efficient, characterized by higher total VFAs and a more stable pH, while digestive enzyme activity was redistributed to optimize nutrient utilization efficiency. The rumen microbiota also shifted toward a structure favoring high energy use efficiency, typified by a higher F/B ratio [49, 60]. Overall, compared with grazing conditions, the housing system was associated with improved growth performance, altered physiological metabolism, and reshaped rumen microecology under the present experimental conditions. It should be noted that daily feed intake of grazing sheep was not directly quantified due to the extensive nomadic management system in the study region. Therefore, the observed differences between feeding systems may partly reflect variation in nutrient intake in addition to environmental and management factors. Future studies integrating precise intake monitoring and controlled supplementation strategies would further clarify the relative contributions of diet composition and management factors.

Conclusion

This study systematically evaluated the effects of grazing and housing systems on the growth performance, serum biochemical and immune indices, rumen fermentation characteristics, and microbial community of Gangba sheep. The results demonstrated that housing system significantly increased final body weight and average daily gain, enhanced antioxidant capacity, and reduced immune activation compared with grazing. Sheep in housing system exhibited improved rumen fermentation efficiency, characterized by higher total VFAs concentrations, optimized enzyme activity profiles, and a microbial community structure favoring energy-efficient metabolism, reflected by an increased Firmicutes/Bacteroidetes ratio. These findings suggest that housing system promotes more efficient nutrient utilization and metabolic adaptation in Gangba sheep. In conclusion, housing system substantially improved the physiological and microbial adaptability of Gangba sheep to high-altitude environments, providing a theoretical basis and practical guidance for optimizing feeding systems and promoting sustainable sheep production on the Tibetan Plateau.

Abbreviations

ADFI

average daily feed intake

ADG

Average daily gain

ALT

alanine aminotransferase

A/P

the molar ratio of acetate to propionate

AST

aspartate aminotransferase

ELISA

enzyme-linked immunosorbent assay

F/B

Firmicutes/Bacteroidetes

GH

growth hormone

GHRH

growth hormone–releasing hormone

GSH-Px

glutathione peroxidase

HDL-C

high-density lipoprotein cholesterol

IFN-γ

interferon-γ

IgA IgG, and IgM

immunoglobulins

IGF-1

insulin-like growth factor 1

IL-2, IL-4, and IL-6

interleukins

LDL-C

low-density lipoprotein cholesterol

LEfSe

linear discriminant analysis effect size

LDA

linear discriminant analysis

MDA

malondialdehyde

MSTN

myostatin

NH₃-N

ammonia nitrogen

OTU

operational taxonomic unit

PCoA

principal coordinate analysis

QTP

Qinghai-Tibet Plateau

RDP

Ribosomal Database Project

SOD

superoxide dismutase

T-AOC

total antioxidant capacity

TC

total cholesterol

TG

triglyceride

TNF-α

tumor necrosis factor-α

UV

ultraviolet

VFA

volatile fatty acid

Authors’ contributions

Yining Xie : writing – original draft, validation, project administration. Xiaokang Jing : methodology, data curation. Zhaohan Zhan : visualization, data curation. Yongqi Tan: software, methodology. Liang Chen : supervision, funding acquisition, writing – review & editing. Hongfu Zhang : conceptualization, resources, funding acquisition. All authors read and approved the final manuscript.

Funding

This research was supported by the National Key Research and Development Program (2022YFD1302102) and Agricultural Science and Technology Innovation Program (ASTIPIAS07).

Data availability

The raw sequencing data of rumen microorganisms generated in this study have been archived in the NCBI Sequence Read Archive (SRA) under accession number PRJNA1221220.

Declarations

Ethics approval and consent to participate

All experimental procedures were approved by the Animal Welfare and Ethics Committee of the Institute of Animal Science, Chinese Academy of Agricultural Sciences (approval number: IAS2024-126), and were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals issued by the Ministry of Agriculture and Rural Affairs of China in 2019. No slaughtering or injection procedures were performed on the animals in this study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.