Time-Dependent Stress Response to Force-Feeding Is Associated with Dynamic Gut Microbiota Changes in Mule Ducks

Abstract

This study aimed to investigate the temporal dynamics of physiological and gut microbial responses in Mule ducks (M-D) during force-feeding (F-F), with the goal of identifying potential regulatory targets to reduce feeding stress. Male M-Ds were subjected to either F-F or ad libitum feeding. We conducted longitudinal analysis at 72, 78, and 84 days of age to assess growth performance, serum biochemical profiles, and intestinal inflammatory markers, while assessing gut microbiota composition through 16S rDNA sequencing. The F-F group exhibited superior growth performance. Initial physiological responses at day 72 included significantly reduced serum corticotropin-releasing hormone (CRH) and jejunal tumor necrosis factor-alpha (TNF−α). Conversely, F-F induced a persistent and profound alteration in the gut microbiome by day 84, characterized by reduced alpha diversity and a significant enrichment of the genus Limosilactobacillus. Correlation analysis identified Limosilactobacillus as a keystone taxon, strongly associated with intestinal metabolites. Our findings demonstrate that M-Ds undergo time-dependent metabolic and immunological adaptations in response to F-F stress, which correlates with distinct alterations in gut microbiota composition, particularly the enrichment of Limosilactobacillus. These findings provide a theoretical basis for developing microbiota-targeted strategies to alleviate F-F stress in foie gras production.

1. Introduction

Foie gras production involves the controversial practice of force-feeding (F-F) ducks and geese to promote liver enlargement [1,2,3], with Mule ducks (M-D) being the predominant breed in the industry owing to their superior hepatic hypertrophy capacity. However, this F-F stress raises substantial animal welfare concerns as it often leads to adverse health consequences in the animals [4,5,6,7]. Current efforts to improve animal welfare through dietary modifications face challenges, primarily due to the limited understanding of M-Ds’ physiological tolerance thresholds under F-F conditions. This knowledge gap substantially impedes the development of evidence-based welfare optimization strategies in foie gras production.

F-F induces metabolic and physiological stress in waterfowl. Studies on Pekin ducks and Landes geese demonstrate that this practice leads to hepatic lipid accumulation accompanied by oxidative damage, inflammation, and elevated serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) [6,8,9]. Furthermore, as the gut serves as a vital immune organ, it exhibits immediate responses to stress [10]. Research on these waterfowl species has revealed that F-F stress induces jejunal pathology characterized by villus atrophy, crypt hyperplasia, and dysregulation of digestive enzymes [6]. These intestinal lesions collectively impair nutrient absorption and are accompanied by concurrent gut microbiota dysbiosis [6,11,12,13], while systemic manifestations include corticosterone imbalance and abnormal levels of high-density lipoprotein (HDL) and low-density lipoprotein (LDL) [6,14,15,16]. Notably, overfeeding studies in M-Ds demonstrate that high-starch diets increase the abundance of Lactobacillus and its phylogenetically related genus Limosilactobacillus in ileal/cecal microbiota. This microbial shift is associated with enhanced mucosal barrier integrity and reduced pro-inflammatory markers [12,17,18]. However, it remains unclear whether M-Ds experience excessive physiological stress during F-F, and how this process may influence serum and intestinal biomarkers, and their potential interplay with gut microbiota dynamics [19].

Given the limited understanding of M-Ds’ physiological tolerance thresholds under F-F conditions and the knowledge gap that impedes evidence-based welfare optimization strategies, this study aimed to investigate the temporal dynamics of physiological responses, intestinal health parameters, and gut microbiota alterations in M-Ds during F-F, with particular focus on identifying potential adaptive mechanisms that may inform welfare improvement strategies. Therefore, we hypothesize that F-F stress affects serum hormones, intestinal morphology, and stress-related biomarkers in M-Ds, but their tolerance threshold to F-F stress is relatively high, which may be associated with adaptive changes in the gut microbiota. In this study, 65-day-old M-Ds were selected for F-F to assess the impact of feeding stress on serum markers of oxidation, inflammation, and liver damage, as well as the integrity of intestinal morphology and barrier function and the composition of gut microbiota and their metabolic products. These findings provide a theoretical foundation for nutritional modulation targets to alleviate the adverse effects of F-F in foie gras production.

2. Materials and Methods

2.1. Experiment Design and Sample Collection

A total of 36 healthy male M-Ds, aged 65 ± 2 days (mean ± SD), were selected. The study employed a two-factor design to evaluate the physiological changes in M-Ds during the F-F period, comparing two dietary treatment groups at three different time points. The specific details are as follows: ad libitum feeding (CON) versus progressive F-F. Blood and intestinal tissue samples were collected at three developmental stages (72, 78, and 84 days of age) to assess physiological parameters and evaluate both temporal dynamics and dietary effects.

Ducks in the ad libitum feeding group had unrestricted access to feed throughout the experimental period. The F-F protocol began at 500 g/day on Day 1 and was increased by 20 g per session each day until reaching 1 kg/day. A pneumatic F-F machine was used for the F-F group. All ducks were housed in individual cages measuring 20 cm × 45 cm; six replicates were used per treatment, each containing one M-D. The room was kept well ventilated and maintained at 22 ± 2 °C (mean ± SD). Manure on the floor was cleaned and rinsed away daily to ensure proper environmental hygiene. The incremental feeding schedule for the F-F group is shown in Supplementary Table S1. The dietary nutrient concentrations were designed to comply with or surpass the nutritional guidelines established by the National Research Council (1994) (

Table 1. Composition and nutrient levels of basal diets for M-Ds ¹.

Ingredient	(%)
Corn	96.9
Duck oil	2
Salt	0.5
Limestone	0.3
Premix 2	0.2
Dicalcium phosphate	0.1
Total	100
Nutrient levels 3
Apparent metabolizable energy (MJ/kg)	13.72
Crude protein (%)	14.33
Ether extract (%)	9.67
Total phosphorus (%)	0.47
Total calcium (%)	0.27

¹ Values in Table 1 are expressed on a dry-matter (DM) basis. The force-feeding slurry was prepared at 60% dry feed and 40% water (w/w). ² The premix includes ferrous sulfate monohydrate 175 mg, copper sulfate pentahydrate 25 mg, manganese sulfate monohydrate 150 mg, zinc sulfate monohydrate 120 mg, 1% potassium iodide 80 mg, 1% sodium selenite 65 mg, multi-dimensional meat duck 120 mg, niacinamide 10 mg, vitamin C 100 mg, VE 75 mg, VB2 5 mg, tannic acid 300 mg, Yisuanbaoling 100 mg, composite enzyme 200 mg, probiotics (Nuowekang) 200 mg, stress-resistant sodium 200 mg. ³ Nutrients are calculated values.

).

On days 72, 78, and 84 of age, the M-Ds were individually weighed following a 12 h fast at the end of the feeding cycle. To minimize the potential influence of circadian rhythms on the results, blood samples (10 mL) were collected at 09:00 h. Following the procedure described in reference [20], blood was collected from the right brachial (wing) vein; serum was separated and stored at −20 °C.

According to the procedure described previously in [21], the M-Ds were euthanized by severing the neck vessels after blood collection and subsequently dissected. Fresh intestinal tissue samples were collected, including approximately 1 cm segments of jejunum and cecum that were fixed in 20 mL of 4% paraformaldehyde. Concurrently, both luminal contents and tissue samples from jejunum and cecum were separately collected, immediately flash-frozen in liquid nitrogen, and stored at −80 °C for subsequent analysis.

2.2. Histopathological Analyses

Intestinal morphology was examined through histological analysis using Hematoxylin and Eosin (H&E) staining, following the protocol described in our previous work [22]. In brief, tissue samples from the jejunum and cecum were preserved in 4% neutral buffered paraformaldehyde for 24 h. These samples were subsequently processed by dehydration, embedded in paraffin wax, and sliced into sections of 4 μm thickness. The sections were stained with H&E (Servicebio, Wuhan, China) according to standard histological techniques. Additionally, measurements of villus height and crypt depth in the jejunum were obtained using Image-Pro Plus 6.0 software (Media Cybernetics, Bethesda, MD, USA).

2.3. Serum Hormone and Biomarker Quantification

Serum hormones and intestinal stress markers in M-Ds were quantified using a one-step double-antibody sandwich ELISA according to the manufacturer’s protocol (Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China): Corticosterone (cat. no. ml037352), corticotropin-releasing hormone (CRH, cat. no. ml037489), adrenocorticotropic hormone (ACTH, cat. no. ml064226), trypsin (cat. no. ml064310), amylase (AMS, cat. no. ml064308), lipase (cat. no. ml064323), superoxide dismutase (SOD, cat. no. ml063948), malondialdehyde (MDA, cat. no. ml064218), glutathione peroxidase (GSH−Px, cat. no. ml037685), catalase (CAT, cat. no. ml064325), alanine aminotransferase (ALT, cat. no. ml061167), aspartate aminotransferase (AST, cat. no. ml036895), glucose (GLU, cat. no. ml037598), alkaline phosphatase (ALP, cat. no. ml064320), interleukin−1β (IL−1β, cat. no. ml061217), tumor necrosis factor−alpha (TNF−α, cat. no. ml036890), interleukin−22 (IL−22, cat. no. ml064352), interleukin−17 (IL−17, cat. no. ml064350), interleukin−6 (IL−6, cat. no. ml061135), interleukin−4 (IL−4, cat. no. ml061194), and transforming growth factor−beta 1 (TGF−β1, cat. no. ml061273).

The detection ranges were corticosterone (1–25 ng/mL), CRH and ACTH (2.5–80 pg/mL), trypsin (1–400 U/mL), AMS (1–400 IU/L), lipase (1–800 U/L), SOD (1–320 U/mL), MDA (0.1–8 nmol/mL), GSH−Px (1–480 U/mL), CAT (1–80 U/mL), ALT (1–64 U/L), AST (1–320 U/L), GLU (1–48 mg/L), ALP (0.1–8 IU/L), IL−1β (1–160 pg/mL), TNF−α (1–80 pg/mL), IL−22 and IL−6 (0.1–32 pg/mL), IL−17 (1–48 pg/mL), IL−1β, IL−4, and TGF−β1 (1–160 pg/mL). The minimum detectable concentrations for these kits ranged from <10 pg/mL to 10 μIU/mL. The intra-assay coefficient of variation was less than 10%, and the inter-assay coefficient of variation was less than 15%.

2.4. Short-Chain Fatty Acid Detection

Concentrations of the main short-chain fatty acids (SCFAs) in the digesta samples were analyzed using gas chromatography (7890A, Agilent, Santa Clara, CA, USA), in accordance with the methodology outlined in our previous study [23]. Briefly, thaw 0.7 g of digesta and suspend it in 1.5 mL of ultrapure water. Vortex the mixture and let it sit on ice for 30 min. Centrifuge each sample at 10,000× g and 4 °C for 15 min. Transfer 1 mL of the supernatant into a 1.5 mL centrifuge tube and mix it with 0.2 mL of crotonic acid-metaphosphate acid. Let it sit at 4 °C for 30 min, followed by centrifugation at 10,000× g and 4 °C for 10 min. Collect 0.3 mL of the supernatant and add it to 0.9 mL of methanol. Vortex the mixture and centrifuge it at 8000× g and 4 °C for 5 min. Finally, analyze the supernatant using gas chromatography with a flame ionization detector. The oven temperature should be set to increase from 100 °C to 190 °C, with nitrogen “N₂” used as the carrier gas at a flow rate of 1 mL/min.

2.5. Bacterial 16S rDNA Gene Sequencing

The procedure of 16S rDNA sequencing was described in detail according to a previous study [24]. The simplified workflow is presented here. For each group, six biological replicates were collected, and every two samples were pooled as one composite sample, resulting in three pooled replicates per group. The simplified workflow is presented here. The bacterial DNA in digesta were extracted by using the Stool DNA Kit (D4015-01, Omega Bio-tek, Norcross, GA, USA), then amplified with the primers of V3-V4 region in bacterial 16S rDNA. The amplicons were purified and subsequently sequenced using the Illumina MiSeq system (Illumina, San Diego, CA, USA) with the MiSeq Reagent Kit v3 followed by quality control and assembly.

2.6. Statistical Analyses

All statistical analyses were performed in IBM SPSS Statistics, version 27.0 (IBM Corp., Armonk, NY, USA) and the Majorbio Cloud Platform (https://cloud.majorbio.com; accessed on 19 July 2025). Data were tested for normality and homogeneity of variance before analysis, confirming that they met the assumptions required for parametric tests.

The normality of the data distribution for serum parameters, intestinal morphology and mucosal factors, digestive enzyme activity, and SCFA of intestinal chyme was assessed using Q–Q plots. Homogeneity of variances was evaluated using Levene’s test. When the assumptions of normality and homogeneity were satisfied, the data were analyzed by two-way analysis of variance (ANOVA) using the General Linear Model (GLM) procedure in IBM SPSS Statistics, version 27.0 (IBM Corp., Armonk, NY, USA) to evaluate the effects of different treatment groups (CON and F-F) and feeding periods (72 d, 78 d, and 84 d). For variables that did not meet the assumptions of normality, appropriate transformations were applied prior to analysis. Duncan’s multiple range test was used for post hoc analysis when significant main or interaction effects, or a trend toward interaction, were observed. The results are presented as means ± SEM. Statistical significance was defined as p ≤ 0.05, and tendencies were considered when 0.05 < p < 0.10.

For microbiome analyses, 16S rDNA data were processed on the Majorbio Cloud Platform. Alpha diversity (Chao1 and Shannon) was compared between feeding regimens at each time point using the two-sample Wilcoxon rank-sum test. Beta diversity was summarized by principal coordinate analysis (PCoA) based on Bray–Curtis dissimilarities, and group separation was assessed with both ANOSIM and PERMANOVA (Adonis), each with 999 permutations. Wilcoxon rank-sum test was used to compare variation in jejunal and cecal digesta samples on taxonomic composition. Correlations between microbial taxa (after arcsine–square-root transformation) and host markers (TNF−α, lipase, CRH, propionic acid, buryric acid, IL1 and IL17) are computed within each intestinal segment and within each age using Pearson’s r, with two-sided tests and BH-FDR control over the correlation matrix. Heatmaps annotate * (P_FDR < 0.05), and the legend indicates which associations remain noteworthy after FDR control.

3. Results

3.1. The Growth Performance and Intestinal Morphology of M-Ds

Analysis of M-D growth performance revealed interactions between feeding regimen and age on average body weight (p < 0.01). Compared to the CON group, the body weight in the F-F group increased on d 72, d 78, and d 84 of age (p < 0.01; Supplementary Figure S1). Additionally, no significant cellular damage was observed in jejunal and cecal tissue sections under different rearing conditions. The overall intestinal tissue structure was well-maintained, with no evident signs of damage (Figure 1A–D). There was no feeding regimen × age interaction for villus length, crypt depth, or villus-to-crypt ratio (V/C) in the jejunum of M-Ds (p > 0.05; Figure 1E–G).

A feeding regimen × age interaction showed a tendency for jejunal lipase activity (p = 0.079, Figure 2). Compared to day 72, trypsin activity increased by day 78 in all groups (p < 0.05). However, by day 84, activity declined from the peak observed on day 78 (p < 0.05, Figure 2). Throughout the trial period, no significant differences were detected between the CON and F-F groups at any time point (p > 0.05). There was no feeding regimen × age interaction for jejunal trypsin and AMS activity (p > 0.05).

3.2. Hormone, Inflammatory, and Oxidative Indicators, Along with Liver Function Parameters in Serum and Intestinal Tissue

To assess the impact of F-F on stress responses in M-Ds, serum levels of CORT, CRH, and ACTH were measured (Figure 3). There was a feeding regimen × age interaction found for serum CRH (p < 0.05). Compared to the CON group, serum CRH levels were significantly lower in the F-F group on day 72 (p < 0.05). However, subsequent measurements at days 78 and 84 showed no intergroup differences (p > 0.05). There was no feeding regimen × age interaction found for serum CORT and ACTH levels (p > 0.05).

Further statistical analysis identified that there was a feeding regimen × age interaction found for cecal IL−1β levels (p < 0.05), while jejunal TNF−α (p = 0.051) and cecal IL−17 levels (p = 0.087) demonstrated a tendency for interaction. The F-F group demonstrated reduced jejunal TNF−α concentrations and elevated cecal IL−1β content compared to CON controls at d 72 (p < 0.05), but not at d 78 and d 84 (Figure 4).

Notably, no feeding regimen × age interactions were found for the measured parameters, including inflammatory cytokines such as IL−22, IL−6, IL−4, and TGF−β1 in both jejunum and cecum, along with IL−1β and IL−17 specifically in the jejunum and TNF−α in the cecum (Figure 4), as well as oxidative indicators comprising SOD, MDA, GSH−Px, and CAT in both intestinal sites, and metabolic biomarkers including ALT, AST, GLU, and ALP similarly in both locations (p > 0.05; Figure 5).

3.3. Effects of F-F on Gut Microbiota Composition in M-Ds

Alpha diversity results (Figure 6) showed that in the jejunum, the F-F group had a lower Shannon index compared to the CON group at d 84 (p < 0.01).

In the cecum, the Chao1 and Shannon indices of the F-F group were lower than those of the CON group on d 84 (p < 0.01, Figure 7). However, no significant differences in alpha diversity were observed between the groups in either intestinal segment at earlier time points (p > 0.05).

Beta diversity results (Figure 8) were assessed using principal coordinate analysis (PCoA) based on Bray–Curtis dissimilarities, with statistical significance evaluated by both PERMANOVA and ANOSIM (999 permutations). No significant separation (p > 0.050) of microbial communities was observed in the jejunum or cecum at d 72 or d 78. However, by d 84, statistically significant separation emerged in both the cecum (PERMANOVA: R² = 0.770, p = 0.041; ANOSIM: R = 1.000, p = 0.041) and jejunum (PERMANOVA: R² = 0.637, p = 0.050; ANOSIM: R = 1.000, p = 0.050), indicating substantial diet-driven divergence in community composition at this later time point.

To further elucidate the taxonomic composition changes underlying these diversity shifts, microbial community structure was analyzed at both the class and genus levels on d 72, d 78, and d 84 (Figure 9). In the jejunum (

Table 2. Effects of feeding regimen and age on bacterial abundance in jejunal microbiome at class and genus levels ¹.

Microbial Taxa	72 d		78 d		84 d		p-Value
	CON	F-F	CON	F-F	CON	F-F	D	R	R × D
Class level
Bacteroidia	0.1 ± 0.07 B	0.1 ± 0.10 B	0.0 ± 0.02 B	0.1 ± 0.05 B	0.3 ± 0.24 A	0.0 ± 0.01 B	0.151	0.085	0.050
Campylobacteria	37.2 ± 19.96 A	1.0 ± 0.99 B	0.0 ± 0.01 B	0.0 ± 0.01 B	0.0 ± 0.00 B	0.0 ± 0.01 B	0.057	0.095	0.073
Genus level
Helicobacter	37.2 ± 34.57 A	1.0 ± 1.71 B	0.0 ± 0.02 B	0.0 ± 0.03 B	0.0 ± 0.00 B	0.0 ± 0.01 B	0.057	0.095	0.073
Lactobacillus	0.0 ± 0.01 B	0.0 ± 0.00 B	0.4 ± 0.49 B	0.2 ± 0.11 B	0.1 ± 0.06 B	42.4 ± 22.71 A	0.002	0.008	0.002
Lactococcus	0.1 ± 0.17 B	0.0 ± 0.00 B	1.0 ± 1.74 B	0.0 ± 0.00 B	4.9 ± 3.53 A	0.0 ± 0.00 B	0.051	0.021	0.051
Limosilactobacillus	0.1 ± 0.17 B	0.0 ± 0.00 B	0.1 ± 0.05 B	0.1 ± 0.03 B	0.2 ± 0.16 B	32.0 ± 8.79 A	<0.001	<0.001	<0.001
Enterococcus	1.2 ± 0.75 B	8.7 ± 4.12 A	5.8 ± 2.06 A	1.8 ± 0.46 B	13.9 ± 9.51 A	0.3 ± 0.27 B	0.743	0.358	0.089

¹ Data are presented as mean ± SEM (n = 3, relative abundance, %). Different uppercase letters denote significant differences (p < 0.05). CON, control group, where ducks were provided with sufficient feed, allowing free access to feed; F-F, conventional force-feeding group; 72 d, 78 d, and 84 d, M-D age (days); R, feeding regimen; D, age of M-Ds; R × D, interaction between feeding regimen and age.

), a significant feeding regimen × age interaction was observed for the relative abundances of Bacteroidia and Campylobacteria at the class level and for Helicobacter, Lactobacillus, Lactococcus, Limosilactobacillus, and Enterococcus at the genus level (p < 0.05). Compared to the CON group, the F-F group had lower Campylobacteria and Helicobacter but higher Enterococcus on day 72 (p < 0.05). By day 84, Bacteroidia decreased while Lactobacillus increased (p < 0.05). Additionally, Enterococcus levels declined on both days 72 and 84 (p < 0.05).

In the cecum (

Table 3. Effects of feeding regimen and age on bacterial abundance in cecal microbiome at class and genus levels ¹.

Microbial Taxa	72 d		78 d		84 d		p-Value
	CON	F-F	CON	F-F	CON	F-F	D	R	R × D
Class level
Actinobacteria	0.1 ± 0.05 B	0.1 ± 0.09 B	0.1 ± 0.08 B	1.1 ± 0.63 A	0.0 ± 0.01 B	0.2 ± 0.05 B	0.007	0.004	0.011
Deltaproteobacteria	7.8 ± 6.99 AB	8.8 ± 8.16 AB	16.9 ± 7.06 A	2.1 ± 2.70 B	11.4 ± 3.12 AB	3.3 ± 2.78 B	0.804	0.018	0.090
Genus level
unclassified_Ruminococcaceae	25.1 ± 7.69 AB	18.7 ± 3.77 BC	17.2 ± 3.95 BC	13.9 ± 4.33 C	16.8 ± 3.37 BC	33.1 ± 2.81 A	0.020	0.368	0.004
unclassified_Bacteroidales	0.9 ± 0.18 B	1.7 ± 1.17 B	0.7 ± 0.12 B	2.4 ± 0.44 B	3.6 ± 3.84 B	15.9 ± 1.00 A	<0.001	<0.001	<0.001
Anaerotignum	0.6 ± 0.38 B	0.4 ± 0.06 B	0.3 ± 0.16 B	0.3 ± 0.10 B	0.3 ± 0.09 B	8.3 ± 0.65 A	<0.001	<0.001	<0.001
unclassified_Firmicutes	0.9 ± 0.63 C	1.9 ± 1.16 BC	4.3 ± 1.39 A	1.2 ± 0.12 C	3.1 ± 1.45 AB	0.2 ± 0.04 C	0.084	0.004	0.006
Bifidobacterium	0.1 ± 0.05 B	0.1 ± 0.09 B	0.1 ± 0.05 B	0.1 ± 0.09 B	0.0 ± 0.00 B	1.1 ± 0.63 A	0.007	0.004	0.006
Limosilactobacillus	0.0 ± 0.00 B	0.00 ± 0.00 B	0.0 ± 0.00 B	0.0 ± 0.02 B	0.0 ± 0.00 B	0.1 ± 0.02 A	0.009	0.002	0.014
unclassified_Lachnospiraceae	10.6 ± 1.93 A	9.8 ± 0.32 A	11.4 ± 0.98 A	8.7 ± 2.05 A	12.6 ± 3.23 A	2.6 ± 0.23 B	0.305	0.010	0.059

¹ Data are presented as mean ± SEM (n = 3, relative abundance, %). Different uppercase letters denote significant differences (p < 0.05). Abbreviations: CON, control group, where ducks were provided with sufficient feed, allowing free access to feed; F-F, conventional force-feeding group; 72 d, 78 d, and 84 d, M-D age (days); R, feeding regimen; D, age of M-Ds; R × D, interaction between feeding regimen and age.

), a significant feeding regimen × age interaction was observed in the relative abundances of Actinobacteria and Deltaproteobacteria at the class level, as well as unclassified Ruminococcaceae, unclassified Bacteroidales, Anaerotignum, unclassified Firmicutes, Bifidobacterium, Limosilactobacillus, and unclassified Lachnospiraceae at the genus level (p < 0.05). Compared to CON, the F-F group showed distinct temporal microbial changes: increased Actinobacteria and decreased Deltaproteobacteria (class level) on day 78, with persistently reduced unclassified Firmicutes through day 84 (p < 0.05). By day 84, unclassified Ruminococcaceae, Bacteroidales, Anaerotignum, Bifidobacterium, and Limosilactobacillus increased, while unclassified Lachnospiraceae decreased (p < 0.05).

Analysis of short-chain fatty acid (SCFA) concentrations in intestinal chyme revealed a feeding regimen × age interaction showing a tendency for cecal propionic acid (p = 0.071) and butyric acid (p = 0.059; Figure 10). Compared to the CON group, the F-F group showed significantly higher concentrations of propionic acid and butyric acid in cecal contents at both d 78 and d 84 (p < 0.05). No significant feeding regimen × age interactions were detected for other SCFAs in either jejunal or cecal contents (p > 0.05).

3.4. Correlation Analysis Between Gut Microbiota Composition and Serum/Intestinal Markers in M-Ds

Pearson correlation analysis with Benjamini–Hochberg FDR correction revealed that at 72 d and 78 d, no significant correlations were observed between cecal microbiota and the measured physiological and biochemical parameters (P_FDR > 0.05, Figure 11). By 84 d, significant associations emerged between cecal microbial taxa and metabolic products (Figure 11): Limosilactobacillus exhibited significant positive correlations with both propionic acid and butyric acid levels (P_FDR < 0.05), Anaerotignum showed a significant positive correlation with butyric acid (P_FDR = 0.0443), and Bifidobacterium demonstrated a significant positive correlation with propionic acid (P_FDR = 0.0433). In contrast, jejunal microbiota showed no significant correlations at any time point (P_FDR > 0.05; Figure 11).

4. Discussion

F-F is a critical technique for producing foie gras [2], where it promotes high energy intake, markedly increasing hepatic fat deposition and metabolic efficiency. The M-D, renowned for its exceptional capacity for liver hypertrophy, has become the predominant species used in foie gras production. However, the intensive metabolic load imposed by F-F also triggers complex physiological and intestinal adaptations, the characterization of which is vital for balancing production performance with animal welfare. The use of 65-day-old M-Ds was driven by both previous research and practical evidence [25,26,27], which collectively recommend initiating F-F at a body weight of 3.5 kg.

Overfeeding with high-energy diets markedly promotes weight gain in pigs, chickens, cattle, and humans [28,29,30,31]. In this study, the average daily gain (ADG) in the force-fed group was significantly higher than that of the CON group, which aligns with previous findings [32]. The observed growth enhancement stems from F-F-induced high energy intake, mediated by maintained intestinal barrier integrity and upregulated digestive enzymes, ultimately boosting fat deposition and energy utilization efficiency [33,34].

The integrity of intestinal function is characterized by the morphological integrity of the gut, with villus height, crypt depth, and the villus height-to-crypt depth ratio serving as representative in calves, pigs, chickens, and Pekin ducks [6,35,36,37]. Intestinal villi absorb nutrients in the small intestine. Increased villus height enhances absorption surface area, promoting growth [37]. Crypt depth indicates epithelial cell production rate—shallower crypts suggest reduced cell renewal [38]. The villus height-to-crypt depth ratio (V/C) effectively assesses intestinal function. Digestive enzyme activity reflects the ability of M-Ds to digest nutrients [39,40]. Previous research on geese has demonstrated that trypsin activity increases with F-F [41]. In this study, the findings of histopathological examination in M-D tissues preliminarily suggest that the intestinal barrier of the experimental M-Ds remained intact and functioned normally. Furthermore, the observed increase in trypsin activity appears to be mediated by F-F-induced hyperphagia, which enhances digestive enzyme secretion.

Corticosterone (CORT) is recognized as a practical physiological indicator of stress response in ducks and is highly sensitive to stress in these animals [42]. When ducks experience stress, the hypothalamus releases corticotropin-releasing hormone (CRH), which stimulates the pituitary gland to secrete adrenocorticotropic hormone (ACTH). ACTH then acts on the adrenal glands, prompting the release of corticosterone into the bloodstream. This process is highly conserved across avian species [43,44]. It should be recognized, however, that CORT functions as a multifaceted hormone involved not only in stress signaling but also in various metabolic processes including energy mobilization, glucose homeostasis, and nutrient partitioning [45]. Prolonged exposure to stressors can lead to persistently elevated CORT levels and chronic stress responses in poultry [46]. During chronic stress responses, cells produce reactive oxygen species (ROS) that disturb redox homeostasis by interfering with key antioxidant enzymes, including superoxide dismutase (SOD), glutathione peroxidase (GSH−Px), and catalase (CAT), while simultaneously promoting the upregulation of pro-inflammatory cytokines [16,47]. This study suggests that despite hormonal activation, systemic oxidative disturbances were not evident during the study period. Notably, pro-inflammatory IL−17 levels in the cecum were reduced by the end of the experimental period, potentially indicating the activation of compensatory anti-inflammatory regulatory mechanisms. While these oxidative and inflammatory findings should be interpreted with appropriate caution given the sample size and potential for delayed or tissue-specific effects, they suggest that acute systemic stress impact was limited in the current experimental context. Interestingly, these results contrast with findings in Pekin ducks by Liu et al. [6], where F-F caused intestinal epithelial damage. This discrepancy highlights notable interspecies differences in physiological responses to F-F, suggesting that M-Ds may possess distinct adaptive mechanisms that warrant further investigation combining both physiological and behavioral assessment approaches.

Gut microbiota and their short-chain fatty acid (SCFA) metabolites mediate adaptive responses to exogenous stressors through dynamic adjustments in community composition. Studies in M-Ds, pigs, and geese demonstrate that intestinal microbiota sense environmental stressors and restructure their communities, enhancing SCFA synthesis. These metabolites, particularly butyrate, are transported into intestinal epithelial cells via Na⁺/H⁺-coupled monocarboxylate transporters, subsequently modulating host energy metabolism, gene expression, and immune signaling pathways [48,49]. This process promotes Muc2 gene expression and goblet cell proliferation, strengthening intestinal barrier integrity and ultimately improving the organism’s adaptation to external stressors [48,49,50,51].

Studies have shown that the gut microbiota can rapidly change in response to dietary alterations [52]. Such microbial shifts can influence the host via two principal mechanisms: interactions mediated by microbial metabolites [53] and the direct disruption of the intestinal barrier by pathogenic bacteria through toxins or other virulence factors [54]. Because F-F entails the rapid intake of large amounts of feed, it markedly reshapes the gut microbiome [6] and, in certain duck breeds, can compromise intestinal barrier integrity [6,34]. In this study, F-F induced gut microbiota restructuring in M-Ds, and the reduction in microbial diversity aligns with previous findings in stress-exposed animals, where physiological challenges often lead to less diverse but more specialized microbial communities [13,55]. The genus of Limosilactobacillus encompasses several species previously classified within Lactobacillus, demonstrating enhanced stress tolerance and metabolic versatility compared to traditional lactobacilli [56]. Members of this genus possess enhanced carbohydrate utilization pathways and stress response genes and superior epithelial adhesion via specialized surface proteins and produce antimicrobial bacteriocins, exopolysaccharides, and immunomodulatory peptides that suppress pro-inflammatory cytokines (TNF−α, IL−1β, IL−6), enhance IL−10 and TGF−β production, and modulate NF−κB signaling [56,57,58,59,60]. Moreover, previous studies indicate that M-Ds tolerate F-F relatively well, with the intestinal barrier often remaining intact after the procedure [8,25]. Enrichment of Limosilactobacillus helps preserve the intestinal barrier in M-Ds; additionally, M-Ds’ strong tolerance to F-F may explain why barrier integrity was maintained despite microbiota alterations. In the present study, the enrichment of Limosilactobacillus in force-fed groups, together with the correlation analysis results, suggests that the proliferation of this beneficial bacterium may represent an adaptive response that helps M-Ds withstand the physiological stress associated with F-F. Because two individual samples were pooled into one biological replicate, the effective sample size was limited to n = 3 per group and time point. Therefore, the current findings should be considered exploratory, and future studies with a larger number of individual samples are required to validate and expand these observations.

5. Conclusions

The combined data on serum factors, intestinal morphology, and microbial dynamics point to a model wherein M-Ds exhibit time-dependent metabolic and immunological adaptations in response to F-F stress. Central to this process is a characteristic shift in the gut microbiota, particularly a significant enrichment of Limosilactobacillus. These findings uncover a key mechanism and provide a theoretical foundation for developing microbiota-targeted therapies to combat F-F stress in foie gras production.