The highly virulent G. parasuis strain XX0306 (serotype 5) was preserved by the Laboratory of the College of Animal Science and Veterinary Medicine, Shenyang Agricultural University. The strain was cultured overnight at 37°C in tryptic soy broth (TSB, Solarbio, Beijing, China) supplemented with 5% fetal bovine serum (Dingguo, Beijing, China) and 2% nicotinamide adenine dinucleotide (NAD, Solarbio).

Twelve four-week-old piglets were purchased from Shuoyang Breeding Cooperative (Xinmin, China). The experimental protocol was approved by the Ethics Committee of Shenyang Agricultural University (No. 202106034). The piglets were randomly divided into two groups of six each and housed under identical conditions. The infection group was intranasally inoculated with 1 × 10¹⁰ CFU of bacterial suspension, while the uninfected group was administered an equal volume of TSB. All piglets were anesthetized with sodium pentobarbital (30 mg/kg, intravenous injection; Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) and euthanized by exsanguination on day 19 post-infection. A complete necropsy was performed immediately. Gross lesions in major organs (heart, liver, spleen, lungs, kidneys, and intestines) were documented photographically.

Blood samples were collected from the anterior vena cava and left at room temperature for 20 minutes to allow natural coagulation, followed by centrifugation at 3,500 rpm for 15 minutes to separate the serum. Serum levels of inflammatory cytokines interleukin 1 beta (IL-1β), interleukin 8 (IL-8), and tumor necrosis factor (TNF-α) were measured using ELISA kits (Enzyme-linked Biotechnology Co. Ltd., Shanghai, China), strictly following the manufacturer’s protocols. Absorbance was measured at 450 nanometers using a microplate reader (Servicebio, Wuhan, China), and a standard curve was generated using the Gen5 software (version 3.11) provided with the instrument to calculate the sample concentration.

Serum markers for intestinal barrier function, diamine oxidase (DAO) and D-lactate, were measured using corresponding ELISA kits (Enzyme-linked Biotechnology Co. Ltd.) following the provided instructions.

Tissue samples from the heart, liver, spleen, lung, kidney, duodenum, jejunum, ileum, cecum, and colon were collected. Solid organs were trimmed into approximately 1.5 cm × 1.5 cm × 0.5 cm blocks, and intestinal segments were cut to 2 cm in length. All tissues were rinsed with sterile phosphate-buffered saline (PBS, Solarbio), fixed in 4% neutral buffered formalin (Solarbio) for over 24 hours, and processed for paraffin embedding.

The fixed tissues were dehydrated through a graded ethanol series, cleared in xylene (Sinopharm Chemical Reagent Co., Ltd.), embedded in paraffin (Solarbio), and sectioned at 5 μm thickness. Sections were stained with hematoxylin and eosin (H&E) using a commercial kit (Solarbio) and observed under a light microscope (Leica Microsystems, Wetzlar, Germany). Images were captured using Leica Application Suite X software (version 3.7.4).

For IFA, deparaffinized and rehydrated sections underwent antigen retrieval by microwave heating in 0.01 M sodium citrate buffer (pH 6.0; Servicebio). Non-specific binding was blocked with 10% goat serum (Solarbio) for 1 h at room temperature. Sections were incubated overnight at 4°C with a rabbit polyclonal anti-G. parasuis primary antibody (produced in-house by immunizing rabbits with whole-cell antigens of strain XX0306; diluted 1:500). Following three washes, sections were incubated for 45 minutes at 37°C with a FITC-conjugated goat anti-rabbit IgG secondary antibody (Epizyme, Shanghai, China; diluted 1:200). For effusion samples (pleural, pericardial, and joint fluid), smears were prepared on glass slides, flame-fixed, and processed identically from the blocking step onward. Slides were mounted with 90% glycerol in PBS and visualized under a fluorescence microscope (Leica Microsystems).

Alveolar lavage fluid (ALF): The entire lung was excised, and bronchoalveolar lavage was performed by instilling and gently kneading with 50 mL of ice-cold sterile PBS. The recovered lavage fluid was centrifuged at 8,000 rpm for 10 min at 4°C. The pellet was snap-frozen in liquid nitrogen and stored at -80°C for subsequent DNA extraction.

Cecal contents: The cecum was exteriorized, and a segment was ligated at both ends. The outer surface was disinfected with 75% ethanol and rinsed with PBS. The segment was opened aseptically, and the luminal contents were collected using a sterile scalpel, placed into 2 mL cryovials, snap-frozen in liquid nitrogen, and stored at -80°C.

Microbial DNA was extracted from ALF pellets and cecal contents using the HiPure Stool DNA Kit (Magen, Guangzhou, China) according to the manufacturer’s protocol. The V3-V4 hypervariable region of the bacterial 16S rRNA gene was amplified with universal primers (341F: CCTACGGGNGGCWGCAG-3’; 806R: GGACTACHVGGGTATCTAAT-5’). Amplicons were purified, quantified, and sequenced on an Illumina NovaSeq 6000 platform (PE250) by GeneDenovo Biotechnology Co., Ltd. (Guangzhou, China).

Raw sequencing reads were processed using USEARCH (Version 11.0). After merging, quality filtering, and chimera removal, sequences were clustered into operational taxonomic units (OTUs) at 97% similarity. Taxonomic annotation was performed against the SILVA database (Release 138). Microbial community analysis and identification of differential taxa were performed using the LEfSe method (LDA score > 2.0, P < 0.05). Functional potential was predicted from the 16S data using the Tax4Fun2 package in R (based on KEGG database).

Metabolites were extracted from cecal content samples by homogenization in 80% methanol-water solution, followed by sonication in an ice-water bath for 6 minutes and centrifugation at 5,000 rpm for 1 minute at 4°C. The supernatant was collected, lyophilized, and reconstituted in 10% methanol. Quality control (QC) samples were prepared by pooling equal volumes from all samples.

Untargeted metabolomic profiling was performed by GeneDenovo Biotechnology Co., Ltd. Specifically, analysis was conducted using ultra-high-performance liquid chromatography coupled with quadrupole-Exactive mass spectrometry (UHPLC-QE-MS) on a Thermo Fisher Vanquish UHPLC system equipped with a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Chromatographic separation was achieved on a Hypersil Gold C18 column (100 × 2.1 mm, 1.9 μm; Thermo Fisher Scientific) at 40 °C using a methanol/water gradient in both positive (0.1% formic acid) and negative (5 mM ammonium acetate) ionization modes. Full MS scans ranged from m/z 100 to 1,500. Raw data were processed using Compound Discoverer software (Version 3.3; Thermo Fisher Scientific). Orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed using the ropls package, and the model was validated by permutation tests. Differential metabolites were identified based on a variable importance in projection (VIP) ≥ 1 and a Student’s t-test p-value < 0.05. Metabolite annotation and pathway enrichment analysis were conducted using the KEGG database (https://www.genome.jp/kegg).

Correlations between the relative abundance of differentially abundant intestinal microbial genera and the intensity of differential metabolites were assessed using Pearson’s correlation coefficient, calculated with the cor.test function in R (Version 4.3.1).

Data are presented as mean ± standard deviation (SD). Statistical analysis and visualization were conducted using SPSS 15.0 software (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 8.0 software (GraphPad Software, San Diego, CA, USA). Comparisons between two groups were made using an independent samples t-test, with a significance level set at P < 0.05, indicating a statistically significant difference.

Piglets inoculated with Glaesserella parasuis strain XX0306 developed clinical signs including growth retardation, coughing, and joint swelling (Figure 1A). At necropsy, infected piglets exhibited characteristic lesions of Glässer’s disease: fibrinous polyserositis manifesting as pleural and pericardial effusions with velvet heart appearance, hepatic congestion, and pulmonary consolidation with surface hemorrhages (Figures 1B, C). Histopathological analysis confirmed myocardial fiber disruption, alveolar congestion, and extensive inflammatory cell infiltration in the lungs (Figure 1D).

Consistent with the observed pathology, serum levels of pro-inflammatory cytokines (TNF-α, IL-1β, and IL-8) were significantly elevated in the infected group compared to the uninfected control group (Figure 2A). Immunofluorescence assay (IFA) directly detected the presence of G. parasuis in pleural effusions, pericardial effusions, and lung tissues of infected piglets (Figures 2B, C). In contrast, G. parasuis was not detected by IFA in joint effusions from the same animals.

The composition of the lung microbiota was characterized using 16S rRNA gene sequencing. A total of 1,622 operational taxonomic units (OTUs) were identified, with 340 OTUs shared between the infected and uninfected groups (Figure 3A). Alpha diversity indices (Shannon, Simpson, and Pielou) were significantly lower in infected piglets, confirming a reduction in lung microbial diversity post-infection (Figure 3B). Beta diversity analysis (ANOSIM, R = 0.117) suggested a separation between groups, although this did not reach statistical significance (P = 0.106) (Figure 3C).

At the phylum level, infection led to an increase in the relative abundance of Proteobacteria and slight decreases in Firmicutes, Cyanobacteria, and Bacteroidota (Table 1). At the genus level, significant increases in the relative abundances of Actinobacillus and Glaesserella were observed in infected lungs (Figure 3D). Concurrently, the number of detectable genera was reduced post-infection. The genus Sandarakinorhabdus was markedly enriched in the lungs of uninfected piglets (Figure 3E).

Functional prediction based on 16S data revealed that the abundance of genes associated with microbial transport and catabolism was significantly lower in the lung microbiota of infected piglets (Figure 3F).

Analysis of 16S rRNA gene sequencing data revealed a total of 1,025 OTUs in cecal contents, with 398 OTUs shared between infected and uninfected piglets (Figure 4A). Alpha diversity indices (Chao and ACE) showed an increasing trend in infected piglets, indicating elevated microbial richness, though these changes were not statistically significant. Goods coverage values confirmed sufficient sequencing depth. Beta diversity analysis (PCoA) revealed distinct clustering between groups, which was supported by a significant ANOSIM result (R = 0.356, P<0.01) (Figures 4B, C).

At the phylum level, the gut microbiota was dominated by Firmicutes, Bacteroidota, Proteobacteria, and Actinobacteria. Infection led to a significant decrease in the relative abundance of Firmicutes (uninfected: 84.41%; infected: 76.10%) and Bacteroidota (9.22% to 3.93%), accompanied by an increase in Proteobacteria (2.40% to 9.08%) and Actinobacteria (2.98% to 4.54%) (Figure 4D; Table 2).

LEfSe analysis (LDA > 2, P<0.05) identified a pronounced dysbiosis at the genus level, characterized by a marked increase in potentially pathogenic genera and a concurrent decrease in several beneficial taxa. Specifically, infected piglets showed a significant increase (P<0.01) in the relative abundances of Streptococcus, Campylobacter, and Desulfovibrio genera, which are often associated with intestinal inflammation or dysfunction. In contrast, the abundances of beneficial genera such as Prevotella_9 and Lachnospira, which are involved in fiber fermentation and short-chain fatty acid production, were significantly reduced (P<0.05). Interestingly, the abundance of another beneficial genus, Bifidobacterium, was significantly higher in the infected group (P<0.01) (Figure 4E; Table 3).

Functional prediction (Tax4Fun2) indicated significant alterations in 12 KEGG pathways. Pathways for Amino Acid Metabolism, Metabolism of Cofactors and Vitamins, and Cell Growth and Death were significantly downregulated in infected piglets (P<0.01). In contrast, pathways for Carbohydrate Metabolism, Lipid Metabolism, and Xenobiotics Biodegradation and Metabolism were significantly upregulated (P<0.05 or P<0.05) (Figure 4F; Table 4).

Untargeted metabolomic analysis of cecal contents using both positive and negative ionization modes showed clear separation between infected and uninfected piglets in the OPLS-DA score plots, with robust model validation (Figures 5A–C).

A total of 119 differential metabolites were identified (VIP ≥ 1, P<0.05). Among these, 30 metabolites (22 in positive and 8 in negative mode) were successfully annotated in the KEGG database (Figure 5D). Fifteen metabolites, including liquiritigenin (P<0.01), citraconic acid (P<0.05), and argininosuccinic acid (P<0.05), were significantly decreased in infected piglets. Conversely, fifteen metabolites, including citrulline (P<0.01) and N-acetyl-L-phenylalanine (P<0.05), were significantly increased (Table 5).

KEGG pathway enrichment analysis of the 30 annotated differential metabolites revealed significant alterations in specific metabolic pathways. The top 20 most enriched pathways are listed in Table 6, with Metabolism of xenobiotics by cytochrome P450 and Arginine biosynthesis showing the most significant enrichment (both P<0.05, Figure 5E). Notably, three key intermediates of the arginine biosynthesis pathway (argininosuccinate, N-acetylornithine, and citrulline) were identified among the differential metabolites.

Pearson correlation analysis between the 15 significantly altered gut microbial genera and the 30 differential metabolites identified a network of 71 statistically significant associations (P<0.05, Figure 6). Notably, Desulfovibrio, Prevotella_7, Prevotella_9, and Lachnospiraceae_XPB1014_group each exhibited significant correlations with more than eight differential metabolites, while Streptococcus and Leucobacter correlated with seven and six metabolites, respectively.

Of particular relevance to the perturbed arginine biosynthesis pathway, two of its key intermediates showed specific microbial correlations. Argininosuccinate levels were negatively correlated with Leucobacter (P<0.05), Streptococcus (P<0.05), and Desulfovibrio (P<0.05), but positively correlated with Prevotella_9 (P<0.01). N-Acetylornithine exhibited a negative correlation specifically with Desulfovibrio (P<0.05).

Histopathological analysis of the small intestine revealed disorganized intestinal epithelial cells, mild villus damage, and glandular atrophy in infected piglets compared to the well-structured mucosa of uninfected controls (Figure 7A). Morphometric measurements of the intestinal architecture confirmed these observations. In infected piglets, crypt depth in the jejunum was significantly increased (P<0.05), while villus length and the villus-to-crypt (V/C) ratio in the duodenum were significantly reduced (P<0.05). Similar decreasing trends in villus length and villus-to-crypt (V/C) ratio were observed in the ileum, although these did not reach statistical significance (Figure 7B).

Consistent with the observed structural damage, serum biomarkers of intestinal barrier integrity were significantly altered. Serum levels of diamine oxidase (DAO) and D-lactate were both significantly elevated in infected piglets (P<0.05) (Figure 7C), indicating increased intestinal permeability.