Multi-omics analysis reveals mechanisms of Qingying granules in treating porcine warm disease: dosage optimization and systems biology insights

Abstract

Background

Warm diseases in swine, particularly those associated with highly pathogenic viruses such as African swine fever virus (ASFV) and porcine reproductive and respiratory syndrome virus (PRRSV), inflict devastating economic losses on global pork production. Current control strategies rely heavily on antibiotics and vaccines; however, emerging antimicrobial resistance and vaccine efficacy variability necessitate alternative therapeutic approaches. Traditional Chinese Veterinary Medicine offers potential solutions, yet rigorous evidence-based studies are required to meet modern regulatory standards. This study aimed to investigate Qingying granules, a novel veterinary formulation derived from the classic Qingying decoction, for the treatment of warm disease in pigs and to elucidate its molecular mechanisms through multi-omics analysis.

Methods

A natural disease model was established using pigs presenting with warm disease. A single production batch of Qingying granules was chemically characterized by UPLC-QE-MS and quantified by HPLC for marker compounds (chlorogenic acid and berberine hydrochloride). Animals were randomly allocated to six dosage groups (0.25–2.00 g/kg/day), a positive control (Qingwen Baidu Powder), and a healthy control group (n = 8). Clinical efficacy was assessed by quantitative syndrome scores, body temperature normalization, weight gain, and lung histopathology. Lung tissues from the optimal-dose group, disease control, and healthy control were subjected to transcriptomic sequencing, untargeted metabolomics, and RT-qPCR validation.

Results

The optimal therapeutic dose of the Qingying granules was 0.75 g/kg/day, which significantly alleviated fever, reduced clinical signs, and inhibited morphological changes of lung tissue. HPLC analysis of the Qingying granules determined the concentrations of chlorogenic acid and berberine hydrochloride as 0.70 mg/g and 0.77 mg/g, respectively. Multi-omics analysis identified 1,955 differentially expressed genes (DEGs) and 23 altered metabolites between the untreated Disease Model and Healthy Control groups. Furthermore, Qingying granules treatment (0.75 g/kg/day) reversed the expression of key genes (LCN2, LRRC18, PDK4, IL1RL1, S100A4) and normalized critical metabolites (7-hydroxyetodolac). Integrated pathway analysis revealed modulation of the PI3K-Akt signaling pathway, cytokine-cytokine receptor interactions, and glycolysis/gluconeogenesis.

Conclusion

Qingying granules at 0.75 g/kg/day exerts potent therapeutic effects on porcine warm disease through multi-targeted regulation of inflammation, immune homeostasis, and metabolic balance. These findings establish a scientific foundation for evidence-based clinical application of traditional herbal formulations in antibiotic-free swine production and demonstrate the utility of multi-omics approaches in veterinary pharmacology research.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12917-026-05664-9.

Background

China is the global leader in swine production and pork consumption, making the swine industry a strategic pillar of agricultural economic stability [1]. Porcine warm disease, a syndrome pattern recognized in traditional Chinese veterinary medicine (TCVM), manifests as acute febrile illness with high mortality, particularly affecting weaned piglets. However, porcine disease prevention and control—particularly febrile conditions associated with highly pathogenic and contagious pathogens such as African swine fever (ASF) and porcine reproductive and respiratory syndrome (PRRS)—represents a critical bottleneck hampering sustainable industry development [2, 3]. These conditions are characterized by abrupt onset, rapid progression, high transmissibility, and significant mortality, affecting pigs regardless of breed or age [4, 5].

Current prevention and treatment strategies for swine warm disease mainly rely on antiviral drugs, antibacterial agents and vaccines [6, 7]. However, the limited availability of specific medicines for this syndrome, coupled with drug residues, emerging drug resistance and mixed infections, greatly restricts the clinical application of these pharmaceuticals and vaccines [8]. Traditional Chinese medicine (TCM) features multiple components and multi-target effects, making it well-suited for the treatment of warm-heat disease.Many formulas are common for clinic use like Lianhuaqingwen Capsule, Fufang Xiangru Liquid, and Shengjiang Powder. Their efficacy in viral elimination and inflammation regulation has been evaluated [9].

TCVM offers a unique theoretical framework for understanding febrile diseases through the “Wei-Qi-Ying-Xue” differentiation system, representing progressive stages of febrile diseases in TCM, from superficial to deep and mild to severe. In TCVM theory, the ‘Yingfen’ stage describes the progression of severe systemic inflammation where pathogenic ‘heat’ penetrates the nutrient level (Yingfen). This stage, characterized by endothelial injury and fluid consumption, clinically manifests as sustained hyperthermia, cutaneous macule, and lethargy [10]. In contemporary swine production, infectious diseases such as classical swine fever, African swine fever, pseudorabies, porcine reproductive and respiratory syndrome, and porcine contagious pleuropneumonia manifest clinical signs that highly coincide with the Yingfen syndrome of Traditional Chinese Veterinary Medicine, typically presenting as hyperthermia and cutaneous erythema with petechiae [11, 12]. Additionally, certain TCVM preparations have demonstrated therapeutic efficacy in the treatment of these infectious diseases [13, 14].

Qingying Decoction, a classic formula for treating Yingfen syndrome, comprises Buffalo Horn (Cornu Bubali), Honeysuckle Flower (Lonicerae Japonicae Flos), Adhesive Rehmannia Fresh Root (Rehmannia Glutinosa), and Dwarf Lilyturf Tube (Ophiopogonis Radix). Historically indicated for “clearing heat and cooling blood”, modern pharmacological studies have demonstrated anti-inflammatory, antimicrobial, antioxidant, antipyretic, and immunomodulatory effects of this formulation [15, 16]. Previous studies have demonstrated that Qingying Decoction can inhibit the inflammatory response by regulating the PI3K/AKT/FoxO pathway [17].

Building on this body of evidence, our research team has developed a novel veterinary formulation, Qingying Oral Liquid (New Veterinary Drug Certificate No. 24, 2016; Category Ⅲ), a novel veterinary therapeutic formulation for the management of febrile syndromes in poultry. However, the inherent bitter taste of oral liquid formulations presents palatability challenges in swine due to their sensitive olfactory system. To address this limitation, we formulated Qingying granules, maintaining the original herbal composition while improving stability, storage, transportation, and administration via feed mixing.

Preliminary studies confirmed therapeutic potential against porcine febrile syndromes [18], yet optimal dosage and molecular mechanisms remained undefined. To ensure reproducibility and quality control, we conducted comprehensive chemical characterization of a single production batch of the formulation using UPLC-QE-MS for profiling and HPLC for quantitative marker analysis. This study aimed to establish the optimal clinical dose through systematic dose-ranging, and elucidate therapeutic mechanisms through integrated transcriptomic and metabolomic analysis, with validation by reverse transcriptase quantitative PCR (RT-qPCR). Collectively, these results lay a scientific foundation for the rational and evidence-based clinical application of Qingying granules in the management of porcine febrile syndromes.

Methods

Preparation and component analysis of Qingying granules

Qingying granules are composed of nine kinds of Chinese medicinal materials, including Forsythiae Fructus, Radix Rehmanniae Recens, and Lonicerae Japonicae Flos. The specific proportions of these medicinal herbs are consistent with those in the original formula (Qingying Decoction), as detailed in Table 1. The formulation used in the experiment was provided by Shandong Dezhou Shenniu Pharmaceutical Co., Ltd. (Batch No. 20220301), with each 1 g of product equivalent to 1 g of crude medicinal materials. Qingwen Baidu Powder (positive control), composed of 14 herbal components including Gypsum Fibrosum and Coptidis Rhizoma, was provided by Shandong Mingfa Animal Pharmaceutical Co., Ltd. (Batch No. 22081601).

Table 1.

Composition of the Qingying granules

Lation name	Chinese name	English name	Proportion of each component herb (%)
Cornu Bubali	Shuiniujiao	Buffalo Horn	33
Lonicerae Japonicae Flos	Jinyinhua	Honeysuckle Flower	10
Rehmannia Glutinosa	Shengdihuang	Adhesive Rehmannia Fresh Root	16
Ophiopogonis Radix	Maidong	Dwarf Lilyturf Tube	10
Radix et Rhizoma Salviae Miltiorrhizae	Danshen	Danshen Root	6
Radix Scrophulariae	Yuanshen	Figwort Root	10
Coptidis Rhizoma	Huanglian	Rhizome of Chinese Goldthread	6
Bambusa Emeiensis	Zhuyexin	/	3
Fructus Forsythiae	Lianqiao	Weeping Forsythia Capsule	6

Chemical profiling (UPLC-QE-MS)

For chemical profiling, 1 g of Qingying granules was extracted with 1 mL water: acetonitrile: isopropanol (1:1:1, v/v/v) via ultrasonic extraction at 4 °C for 30 min. Following centrifugation (15,300 g, 4 °C, 20 min), supernatants were analyzed using UPLC-Orbitrap-MS (Vanquish UPLC coupled with Q Exactive HF-X MS, Thermo Fisher Scientific). Compounds were identified by matching against the Sanshu Biotechnology secondary mass spectrometry database and confirmed using authentic standards where available.

Quantitative analysis (HPLC)

Chlorogenic acid (from Lonicerae Japonicae Flos) and berberine hydrochloride (from Coptidis Rhizoma) were quantified using a validated HPLC–UV method (335 nm) on an Agilent XDB-C18 column with gradient elution. This established method was applied to the experimental batch to determine marker compound contents for correlation with therapeutic effects.

Animal recruitment and disease model

This trial was conducted at a large-scale commercial swine farm in Qingdao, China. A total of 82 Duroc-Landrace-Yorkshire hybrid pigs (5–6 weeks old, mixed sex, initial body weight 7.2 ± 1.2 kg) were recruited. Pigs were housed in individual pens (1.2 m × 0.8 m) with libitum access to water, and standard commercial feed.

Porcine warm disease was diagnosed based on our previously established protocol (Patent No.: CN116210638A). This represents a natural disease model where pigs exhibiting clinical manifestations consistent with TCVM-defined warm disease were enrolled. The diagnostic criteria integrate TCVM syndrome differentiation with modern clinical parameters, as detailed in Table 2. This represents a natural disease model where pigs exhibiting clinical manifestations consistent with warm disease were enrolled. Clinically, eligible pigs presented with ≥ 3 primary signs (with hyperthermia mandatory) and ≥ 2 secondary signs (Table 2), achieving a syndrome score of 16–27 (Table 3). Specific clinical differentiation was performed by a senior expert certified in both TCVM and clinical veterinary diagnosis. Healthy Control pigs exhibited normal body temperature (38.0–39.5 °C), normal feed intake, and absence of clinical abnormalities (Table 4).

Table 2.

Qualitative diagnostic and enrollment criteria for swine warm disease

Symptoms	Clinical signs
Primary clinical signs	1. Hyperthermia
	2. Cutaneous macule
	3. Reduced feed consumption
	4. Lethargy
Secondary signs	1. Polydipsia with reduced drinking
	2. Dry or loose stools
	3. Rough and unkempt hair coat
	4. Tachypnea
	5. Dry snout

Table 3.

Quantitative syndrome scoring system for disease severity and efficacy evaluation

	Category	Evaluation index	Score
Primary clinical signs	Temperature	38.0℃ ≤ T < 39.5℃	0
		39.5℃ ≤ T < 40.5℃	2
		40.5℃ ≤ T < 41.0℃	4
		T ≥ 41.0℃	6
	Feed intake	Normal feed intake	0
		Slight decrease in feed intake	2
		Significant decrease in feed intake	4
		Near-cessation of feed intake	6
	Mental state	Alert responsiveness: moves immediately in response to verbal prompting	0
		Sluggish responsiveness: fails to move to verbal prompting, only moves when physically prodded	2
		Lethargic: fails to move when physically prodded, only moves when lightly tapped	4
		Somnolent: remains immobile even when lightly tapped	6
	Macules	No macules	0
		Slightly erythematous skin	2
		Faint petechiae on the skin or around hair follicles	4
		Extensive petechiae covering large areas of the skin	6
Secondary signs	Water intake	Normal	0
		Slight decrease	1
		Significant decrease	2
		Near-cessation	3
	Feces	Normal	0
		Slightly dry or unformed	1
		Dry and hard or loose	2
		Hard and pellet-like or watery	3
	Hair coat	Normal	0
		Dull	1
		Rough and unkempt	2
		Disheveled	3
	Respiration	Normal	0
		Increased respiratory rate	1
		Tachypnea	2
		Dyspnea with costal flaring	3
	Snout	Moist	0
		Dry, without cracks	1
		Dry, with a few cracks	2
		Dry, with deep fissures	3

Table 4.

Inclusion and exclusion criteria for healthy control pigs

Indexes	Inclusion criteria	Exclusion criteria
Mobility	Actively mobile	Slight depression
Body temperature	Normal	Over 39.5 °C
Appetite	Good	Slightly poor
Stool form	Formed stool	Unformed and abnormal stool
Age	5- to 6-week-old	More than 6-week-old or less than 5-week-old

All experimental procedures involving animals were reviewed and approved by the Laboratory Animal Ethics Commission of the Poultry Research Institute, Shandong Academy of Agricultural Sciences (Approval No. JQS-2024–16). The study was complied with the Regulations for the Administration of Affairs Concerning Experimental Animals of China.

Experimental design and group allocation

Phase 1: dose-ranging study (n = 64)

To determine the optimal therapeutic dose, 56 pigs diagnosed with warm disease were randomly allocated (random number table method) to seven groups (n = 8 per group): six Qingying Granule (QYG) groups receiving 0.25, 0.50, 0.75, 1.00, 1.50, or 2.00 g/kg body weight (BW)/day via oral gavage; and a Positive Drug Control (PDC) group receiving Qingwen Baidu Powder (50 g/day). A Healthy Control (HC) group (n = 8) comprised clinically normal pigs receiving no treatment. Treatment duration was 7 consecutive days, followed by a 7-day post-treatment observation phase. The optimal dose was determined based on clinical signs, body weight changes, temperature normalization, and histopathology.

Phase 2: mechanistic study (n = 18)

To elucidate molecular mechanisms, 12 additional pigs with warm disease were allocated to two groups (n = 6 per group): Untreated Disease Model (DM, euthanized at day 0 to establish baseline pathology), Qingying granules optimal-dose group (QYG, 0.75 g/kg/day for 7 days, euthanized at day 7). Six healthy pigs were included in the healthy control group (HC, euthanized at day 7). Lung tissues were harvested for transcriptomics, metabolomics, and RT-qPCR.

Clinical observations and efficacy evaluation

Clinical examinations were performed daily at 08:00 and 16:00 by blinded observers. Syndrome scores were calculated using Table 3 criteria. Rectal temperatures were recorded twice daily (08:30 and 16:30) using a calibrated digital thermometer. Individual body weights were measured on days 0, 8, and 15 (fasted overnight). Daily Average Weight Gain (DAWG) was calculated as: DAWG = (Final BW – Initial BW)/Days. Feed intake was measured individually (feed supplied minus residual, weighed using an electronic balance precision 0.1 kg). Average Daily Feed Intake (ADFI) = Total Feed Intake/Number of days. Feed Conversion Ratio (FCR) = Total Feed Intake/Total Weight Gain.

Efficacy Judgment Criteria modified from Compilation of Technical Guidelines for Veterinary Drug Research, China, (2006–2011) (Table 5):

Table 5.

Therapeutic efficacy evaluation criteria

Efficacy Grade	Judgment Criteria
Recovery	1. Syndrome disappears, with 100% reduction in syndrome score; 2. After treatment: fever subsides, body temperature returns to normal; feed intake and skin rashes vanish; all other accompanying clinical signs are completely eliminated; 3. No recurrence during the observation period
Markedly Effective	1. Syndrome is significantly improved, with 70% ≤ reduction in syndrome score < 100%; 2. After treatment: body temperature is basically normal; mental state and appetite improve; skin rashes and thirst (without drinking) are markedly alleviated; other accompanying clinical signs mostly disappear; 3. No clinical signs aggravation during the observation period
Effective	1. Syndrome is partially improved, with 30% ≤ reduction in syndrome score < 70%; 2. After treatment: body temperature is slightly elevated; mental state and appetite are slightly better; skin rashes, listlessness and thirst (without drinking) are somewhat relieved; other accompanying clinical signs are reduced; 3. No clinical signs aggravation during the observation period
Ineffective	1. Syndrome shows no obvious improvement or even aggravation, with reduction in syndrome score < 30%; 2. After treatment: body temperature does not decrease or even rises; skin rashes, listlessness and thirst (without drinking) are not relieved; other accompanying clinical signs are unimproved or aggravated

Histopathological examination

After the end of the observation period, all pigs were euthanized by intravenous injection of sodium pentobarbital at a dose of 100 mg/kg. Lung tissues were immediately harvested from the right caudal lobe (consistent anatomical location to minimize variation), fixed in 4% paraformaldehyde for 48 h, embedded in paraffin, sectioned (5 μm), and stained with hematoxylin–eosin (H&E). Sections were examined under an Olympus BX43F light microscope (× 100 and × 400 magnification) by a pathologist blinded to group allocation.

Transcriptomic analysis

To minimize individual biological variation while maintaining statistical power, lung tissue samples were pooled before RNA extraction (n = 3 pools per group, each pool comprising equal weights of tissue from 2 individual samples). This pooling strategy is consistent with previous veterinary omics studies [19, 20] and effectively reduces within-group variation for pathway analysis while conserving biological replicates. Total RNA was extracted using TRIzol reagent (Invitrogen) and purified with the RNeasy Mini Kit (Qiagen). RNA concentration and purity were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific); samples with A260/A280 ratios between 1.8 and 2.1 were included. RNA quantity was accurately measured using a Qubit RNA HS Assay Kit (Thermo Fisher Scientific). Raw sequencing reads were initially assessed for quality using FastQC (v0.11.9). Subsequently, adapter sequences, poly-G tails (resulting from the two-color chemistry of the NovaSeq platform), and low-quality bases (quality score < Q20) were trimmed using fastp (v0.23.2). Clean reads were aligned to the porcine reference genome Sus scrofa 11.1 (Ensembl release 115) using STAR (2.7.11a) with default parameters. Alignment statistics, including uniquely mapped reads, multi-mapped reads, and unmapped reads, were summarized using SAMtools (v1.15). The overall mapping rate ranged from 98.25% to 99.20%. The resulting BAM files were then sorted and indexed. Gene-level read quantification was performed using featureCounts (v2.0.3) from the Subread package, utilizing the corresponding GTF annotation file.

Libraries were constructed with the NEBNext Ultra RNA Library Prep Kit and sequenced on the Illumina NovaSeq 6000 platform (150 bp paired-end reads, > 6 Gb data per sample). Raw reads were filtered to remove adapters and low-quality sequences. Raw count matrices were imported into R (v4.3.1) for downstream analysis using DESeq2 (v1.38.0). The DESeq function was applied using default parameters, which employ the median-of-ratios method for normalization and the Wald test for statistical significance. The DESeq2 design formula used for differential expression analysis was group. Genes with |log2FC|≥ 0.58 and a false discovery rate (FDR) ≤ 0.05 (Benjamini–Hochberg correction) were defined as differentially expressed genes (DEGs).” Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted using clusterProfiler (v4.6.0), with the significance threshold set at FDR < 0.05.

Metabolomic analysis

Lung tissue metabolites were extracted using methanol: water (4:1, v/v) containing internal standards. The extracts were analyzed by UPLC-Q Exactive MS (Thermo Fisher) in both positive and negative ion modes. Data preprocessing (peak picking, peak alignment) was performed using Progenesis QI software. Metabolites were identified using the KEGG, and LIPID MAPS databases. Multivariate analysis, including principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA), was conducted using SIMCA-P software (v14.1). Differentially abundant metabolites (DAMs) were identified by combining variable importance in projection (VIP) > 1.0 (from OPLS-DA) and p < 0.05 (Student's t-test, FDR corrected). Pathway enrichment analysis was performed using MetaboAnalyst 5.0.

Integration of transcriptomic and metabolomic data

To identify key multi-omics biomarkers, a Random Forest (RF) classification model was constructed using the randomForest package (v4.7–1.1) in R (v4.3.3). The input variables comprised significant differentially expressed genes (FDR < 0.05) and differentially abundant metabolites (p < 0.05). Data preprocessing included removing features with missing values or zero variance, log2-transforming the metabolite data, and Z-score scaling the merged matrix. The RF model was built with ntree = 500 and mtry = sqrt(p). Internal validation and model performance were evaluated using the Out-Of-Bag (OOB) error rate and confusion matrix. Feature importance was assessed using Mean Decrease in Accuracy (MDA) and Mean Decrease in Gini (MDG), with the top 30 features ranked by MDG extracted for visualization.

RT-qPCR validation

Lung tissues from different groups were homogenized, and total RNA was extracted and purified using a Quick RNA Kit (Accurate Biotechnology Co., Ltd., Hunan, China) in accordance with the manufacturer’s instructions. Total RNA (1 μg) was reverse-transcribed using the Evo M-MLV RT Kit (Accurate Biotechnology). Quantitative real-time PCR (qPCR) was performed with SYBR Green Premix Pro Taq HS on a QuantStudio 3 system (Applied Biosystems). Primers for target genes (LCN2, LRRC18, PDK4, IL1RL1, S100A4) and the reference gene (GAPDH) were designed and synthesized by Sangon Biotech (Shanghai) Co., Ltd. (Table 6). The relative expression levels were calculated using the 2^−ΔΔCt method [21], with normalization to the HC group.

Table 6.

Primer sequences used for quantitative real-time PCR

Gene name	Forward	Reverse
LCN2 (Lipocalin-2)	CCTAAGTCTCCTGTGGCTGGG	TGAGCTCGTAGGTGGTGGTG
PDK4 (Pyruvate dehydrogenase kinase 4)	ATCCCCGCTGTCCATAAAGC	CCAACTCCATCAGGCTCTGG
LRRC18 (Leucine-rich repeat-containing protein 18)	AAAAACGCCTCGACTTGAGC	GCTGACGTTGAGGTAGAGCA
S100A4 (S100 calcium-binding protein A4)	GCTAAAGGAGTTGCTGACCCG	CCACCTCGTTGTCCCTGTTG
IL1RL1 (Interleukin-1 receptor-like 1)	TCCCCGACAAGACCACTCTC	CAGTGTCGCTGACTTTGGC
GAPDH (Glyceraldehyde-3-phosphate dehydrogenase)	AGTGGACATTGTCGCCATCA	TACGTAGCACCAGCATCACC

Statistical analysis

Data were analyzed using SPSS Statistics (Version 26.0). All data are presented as mean ± standard deviation (SD). Normality was assessed using the Shapiro–Wilk test. One-way analysis of variance (ANOVA) followed by Tukey's post-hoc test was used for normally distributed data. A P value < 0.05 was considered statistically significant.

Results

Chemical components of Qingying granules

To identify the main chemical constituents, the components of Qingying granules were determined by UPLC-QE-MS. The total positive and negative ion chromatograms of Qingying granules are shown in Fig. 1A and B, and the chemical compositions of all compounds are presented in the supplementary materials (supplementary file 2). Adenosine and danshensu may be derived from Radix et Rhizoma Salviae Miltiorrhizae; quercetin, forsythiaside A, and cynaroside may be from Fructus Forsythiae; rehmannioside may be from Rehmannia glutinosa; berberrubine and coptisine chloride may be derived from Coptidis Rhizoma; sinapic acid and isoacteoside may be from Radix Scrophulariae; vitexin and homoorientin may be from Bambusa emeiensis; biochanin-7-O-glucoside and sucrose may be from Ophiopogonis Radix; and chlorogenic acid and rutin may be from Flos Lonicerae Japonicae. These chemical constituents were preliminarily detected by mass spectrometry using the database of Sanshu Biotechnology Co., Ltd. (Jiangsu, China), which may be the active components of Qingying granules.

Fig. 1.

Identification of chemical components of Qingying granules by UPLC-QE-MS. A Positive ion mode chromatogram; B Negative ion mode chromatogram

HPLC analysis with validated methodology determined the contents of the two marker compounds in Qingying granules (Fig. 2). The content of chlorogenic acid was 0.70 mg/g and berberine hydrochloride was 0.77 mg/g. These quantitative data provide essential quality control baseline for the experimental batch.

Fig. 2.

HPLC chromatograms of marker compounds in Qingying granules. A Mixed reference standards (chlorogenic acid and berberine hydrochloride); B Qingying granules sample; C Negative control without Coptidis Rhizoma (berberine hydrochloride absent); D Negative control without Lonicerae Japonicae Flos (chlorogenic acid absent)

Clinical efficacy of Qingying granules

Body Temperature and Weight Gain: Body temperatures of the experimental pigs were monitored and recorded throughout the study, and the results are presented in Fig. 3. At 24 h post-administration, a proportion of pigs in all QYG groups returned to physiological normothermia. By 48 h post-administration, 75% of pigs in QYG Group 3 (0.75 g/kg) had regained normal body temperature, while 50% of those in QYG Groups 2 (0.50 g/kg) and 4 (1.00 g/kg) achieved similar recovery. In contrast, fewer than 50% of pigs in the remaining treatment groups reached normothermia at this time point. Following the 7-day treatment, 100% of pigs in QYG 0.75–2.00 g/kg groups achieved normothermia, demonstrating superior efficacy compared to lower doses and positive control.

Fig. 3.

Mean body temperature changes over time in different treatment groups. QYG1 to QYG6 groups received Qingying granules at doses of 0.25, 0.50, 0.75, 1.00, 1.50, and 2.00 g/kg/day, respectively. The PDC group received Qingwen Baidu Powder (50 g/day), and the HC group was the healthy control. Each data point represents the mean ± standard deviation (SD) of body temperature for each group (n = 8) at the indicated time point

Body weight data of the experimental pigs across all groups throughout the study are summarized in Table 7. On day 8, after seven days of treatment, the weight gain of pigs in QYG Groups 1 and 2 was significantly lower than that in the HC group (P < 0.05). By contrast, QYG Groups 3–6 and the PDC group showed no significant differences in weight gain relative to the HC group (P > 0.05). Following the 7-day observation period (day 15), significant reductions in body weight gain were detected in QYG Groups 1–2 and the PDC group when compared with the HC group. In addition, QYG Groups 3–6 achieved significantly higher weight gain than the PDC group, supporting the earlier conclusion that their cumulative weight gain outperformed that of QYG Groups 1–2 and the PDC group.

Table 7.

Body weight in experimental pigs during the study period

Treatment Group (n = 8)	QYG concentration (g/kg)	Body Weight at D0 (kg)	Body Weight at D8 (kg)	Body Weight at D15 (kg)
QYG 1 group	0.25	7.4 ± 1.0	8.1 ± 1.2*	10.6 ± 1.4*
QYG 2 group	0.50	7.4 ± 1.2	8.3 ± 1.3*	10.4 ± 1.5*
QYG 3 group	0.75	7.3 ± 1.4	8.7 ± 1.3	10.8 ± 1.2#
QYG 4 group	1.00	7.2 ± 1.0	8.5 ± 1.3	10.6 ± 1.5#
QYG 5 group	1.50	7.2 ± 1.3	8.7 ± 1.0	10.7 ± 1.3#
QYG 6 group	2.00	7.2 ± 1.1	8.7 ± 1.0	11.0 ± 0.9#
PDC group	-	7.4 ± 1.2	8.7 ± 1.4	10.8 ± 0.9*
HC group	-	7.4 ± 1.1	9.2 ± 0.9	11.7 ± 0.8#

HC Healthy Control, PDC Positive Drug Control, QYG Qingying granules

^*Indicates a significant difference (P < 0.05) compared to the HC group

^#Indicates a significant difference (P < 0.05) compared to the PDC group

Syndrome Scores and Clinical Signs: Throughout the experiment, pigs in the Control group remained healthy, maintaining normal mental status, regular feed intake, no skin rashes, and unremarkable clinical parameters (e.g., body temperature, fecal consistency). All pigs in the remaining groups (QYG Groups and the PDC group) initially presented with clinical abnormalities, including skin rashes, lethargy, and elevated body temperature.

On day 0 (baseline), the clinical syndrome scores of all QYG groups and the PDC group were significantly higher than those of the HC group, which was attributed to the enrollment of diseased pigs in these groups. On day 7 (at the end of treatment), scores in QYG Groups 1–2 and the PDC group remained significantly elevated relative to the HC group, whereas no significant difference was observed between QYG Groups 3–6 and the HC group. On day 14 (upon completion of the observation period), QYG Groups 1–2 and the PDC group still had markedly higher scores than the HC group, and the scores of QYG Groups 3–6 returned to normal levels. Syndrome scores are summarized in Table 8.

Table 8.

Changes in clinical syndrome scores across experimental groups

Treatment Group (n = 8)	QYG concentration (g/kg)	Score at D0	Score after Administration (D7)	Score at the End of Observation Period (D14)
QYG 1 group	0.25	21.75 ± 2.12*	6.67 ± 2.42*	4.17 ± 1.17*
QYG 2 group	0.50	21.38 ± 1.85*	4.29 ± 2.93*	2.43 ± 1.90*
QYG 3 group	0.75	21.88 ± 2.59*	2.50 ± 2.27#	0.88 ± 1.13
QYG 4 group	1.00	21.50 ± 2.73*	2.38 ± 2.07#	0.75 ± 1.16
QYG 5 group	1.50	21.25 ± 2.38*	2.13 ± 1.89#	0.63 ± 0.92
QYG 6 group	2.00	21.75 ± 1.67*	2.29 ± 1.89#	0.86 ± 1.21
PDC group	-	21.13 ± 3.18*	5.43 ± 2.76*	2.86 ± 2.04*
HC group	-	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00

HC Healthy Control, PDC Positive Drug Control, QYG Qingying granules

^*Indicates a significant difference (P < 0.05) compared to the HC group

^#Indicates a significant difference (P < 0.05) compared to the PDC group

The syndrome score reduction rates and therapeutic efficacy outcomes of all groups are summarized in Supplementary file 1 (Table S4). According to the predefined efficacy criteria, the cure rates at the end of the 7-day treatment period were 0% in QYG Group 1, 25% in both QYG Groups 2 and 6, 37.5% in QYG Groups 3, 4, and 5, and 12.5% in the PDC group. By the end of the post-treatment observation period, the cure rates had increased to 37.5% in QYG Group 2, 50% in QYG Groups 3 and 6, and 62.5% in QYG Groups 4 and 5, with the PDC group reaching 25%; no cures were observed in QYG Group 1 at either time point. Furthermore, the overall effective rate of QYG Groups 3–6 reached 100% at the end of the treatment period, which was significantly higher than those of QYG Groups 1–2 and the PDC group.

Lung histopathology

The HC group exhibited normal pulmonary histological features, with neatly arranged lung cells and uniform morphology (Fig. 4A). In QYG Groups 1–2, alveolar wall thickening, hemorrhage, and inflammatory cell infiltration were still observed (Fig. 4B-C). In QYG Groups 3–6, these pulmonary injuries were significantly alleviated, with basically intact alveolar structures, no obvious alveolar fusion, and only scattered mild inflammatory cell infiltration (Fig. 4D-G). In the PDC group, only mild inflammatory cell infiltration and mild alveolar wall thickening were observed (Fig. 4H).

Fig. 4.

Histopathological changes in lung tissue across experimental groups. A Healthy control group; B-G Qingying granules groups 1–6 (0.25, 0.50, 0.75, 1.00, 1.50, and 2.00 g/kg/day); H Positive drug control group. Hematoxylin and eosin staining; original magnification, × 100. Black arrows indicate inflammatory cell infiltration; Red arrows indicate alveolar wall thickening and hemorrhage

Based on clinical efficacy, histopathological improvement, and cost-effectiveness considerations, 0.75 g/kg/day was determined as the optimal therapeutic dose for subsequent mechanistic studies.

Transcriptomics analysis

PCA demonstrated distinct clustering among DM, QYG, and HC groups, indicating significant transcriptomic divergence (Fig. 5A). Compared to the HC group, the DM group exhibited 1,955 DEGs (965 upregulated, 990 downregulated), indicating substantial disease-induced transcriptional perturbation. Critically, comparison between the QYG (0.75 g/kg) and DM groups revealed 178 DEGs (98 downregulated, 80 upregulated) that were modulated toward healthy levels, demonstrating the restorative effect of Qingying granules (Fig. 5B).

Fig. 5.

Transcriptomic analysis of lung tissue samples. A Principal component analysis (PCA) plot showing the global transcriptomic separation among the Healthy Control (HC), Untreated Disease Model (DM), and Qingying granules treatment (QYG) groups. Each data point represents an independent pooled sample (n = 3 pools per group). Ellipses indicate the 95% confidence intervals for each group. The x-axis (PC1) and y-axis (PC2) represent the first and second principal components, with the percentages indicating the proportion of total variance explained by each component. B Bar plot illustrating the number of differentially expressed genes (DEGs) identified in the comparisons of DM vs. HC, HC vs. QYG and QYG vs. DM. Red and green colors indicate upregulated and downregulated genes, respectively

In the comparison between the DM group and the HC group, GO enrichment analysis revealed significant enrichment of differentially expressed genes (DEGs) in three functional categories: molecular function, including immunoglobulin receptor binding and extracellular matrix structural constituent; cellular component, encompassing extracellular matrix, immunoglobulin complex, and circulating immunoglobulin complex; and biological process, involving extracellular matrix organization and blood vascular development (Fig. 6A). In the comparison between the QYG group and the DM group, DEGs were significantly enriched in molecular functions related to immune and signal transduction activities, such as MHC class I binding and G protein-coupled receptor (GPCR) activity; cellular components overlapping with extracellular structures, including extracellular matrix and plasma membrane; and biological processes involving glucocorticoid/steroid hormone response and negative regulation of cell proliferation (Fig. 6B), suggesting immune modulation and anti-inflammatory mechanisms.

Fig. 6.

Functional enrichment analysis of differentially expressed genes. A, B Bubble plots of significantly enriched Gene Ontology (GO) terms for biological process, molecular function, and cellular component categories. A DM vs HC; B QYG vs DM. C, D Bubble plots of significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. C DM vs HC; D QYG vs DM

KEGG pathway analysis of DM vs HC DEGs revealed enrichment in immune dysregulation pathways (Primary Immunodeficiency, Intestinal IgA network), inflammatory signaling (Cytokine-cytokine receptor interaction), and metabolic disruption (Glycolysis/Gluconeogenesis) (Fig. 6C). Comparison of QYG vs DM demonstrated that treatment significantly reversed these pathological pathways, particularly enriching in immune regulation (Antigen processing and presentation, Graft-versus-host disease), metabolic homeostasis (Linoleic acid metabolism), and anti-inflammatory processes (Fig. 6D), consistent with immune restoration and metabolic normalization.

RT-qPCR validation of key genes

To validate transcriptomic findings, five pivotal DEGs were quantified using RT-qPCR (Fig. 7). Compared to HC group, the DM group showed significant downregulation of IL1RL1 and S100A4 (P < 0.01), and upregulation of LCN2, LRRC18 and PDK4 (P < 0.01). QYG treatment significantly reversed these aberrations, restoring expression toward HC levels (P < 0.01 for all comparisons vs DM), confirming the multi-omics results.

Fig. 7.

Validation of differentially expressed genes by quantitative real-time PCR. Expression levels of (A) IL1RL1, B S100A4, C LCN2, D LRRC18, and E PDK4 in lung tissue relative to healthy control (set as 1.0). Data are presented as mean ± standard deviation (n = 6 per group). *P < 0.05, **P < 0.01 versus healthy control group; #P < 0.05, ##P < 0.01 versus model control group. HC, healthy control; DM, Untreated Disease Model; QYG, Qingying granules (0.75 g/kg/day)

Metabolomic profiling and pathway analysis

PCA revealed distinct metabolic clustering among groups (Fig. 8A). Untargeted metabolomics identified 23 differential metabolites between DM and HC (11 increased, 12 decreased), primarily involving carboxylic acids, flavonoids, and organooxygen compounds (Fig. 8B). Key disease-associated metabolites included elevated inflammatory mediators and reduced antioxidant compounds. Furthermore, Qingying granules treatment reversed this metabolic perturbation by regulating 10 metabolites (4 upregulated and 6 downregulated; Fig. 8C). These regulated metabolites were classified into lipids and lipid derivatives, amino acids and carbohydrates. Representative metabolites included tetradecyl sulfuric acid, decanoyl acetaldehyde, 7-hydroxyetodolac, and S-propargylcysteine.

Fig. 8.

Metabolomic profiling of lung tissue samples. A Principal component analysis (PCA) score plot showing metabolic separation among groups. B Volcano plot of differential metabolites between DM group and HC groups. C Volcano plot of differential metabolites between QYG group and DM groups. D Bubble plot of significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for differential metabolites

KEGG analysis revealed that disease-altered metabolites were enriched in Glycolysis/Gluconeogenesis, Glycerolipid metabolism, and Glycine-Serine-Threonine metabolism (Fig. 8D). Notably, these pathological metabolic pathways were normalized in the QYG group, with no significant difference from the HC group.

Integration of transcriptomic and metabolomic data

Random forest analysis identified key biomarkers distinguishing disease from health, including LCN2, PDK4, LRRC18 and 7-hydroxyetodolac (Fig. 9A-B). Correlation analysis revealed coordinated regulation between DEGs and metabolites: NABP1 expression negatively correlated with glycolytic intermediates (pyruvate, lactate), while LCN2 negatively correlated with 7-hydroxyetodolac levels.

Fig. 9.

Integration of transcriptomic and metabolomic data. A, B Random forest analysis identifying the top 30 most important biomarkers distinguishing (A) DM vs HC and B QYG vs DM. C Top 10 pathways enriched in differential genes and metabolites between DM group and HC groups

Integrated pathway analysis highlighted the PI3K-Akt signaling pathway and Cytokine-cytokine receptor interaction as core networks targeted by Qingying granules, linking metabolic reprogramming (glycolysis) with immune regulation (Fig. 9C). These findings establish a systems-level mechanism connecting herbal multi-component activity to disease resolution.

Discussion

This study validates Qingying granules (0.75 g/kg/day) as a potentially effective therapy for porcine warm disease, a severe febrile condition in swine production. Using a rigorously characterized natural disease model and integrated multi-omics approach, we demonstrate that this herbal formulation alleviates clinical signs, normalizes body temperature, improves growth performance, and inhibits morphological changes of lung tissue. Mechanistically, these benefits are mediated through multi-targeted modulation of inflammation (LCN2, IL1RL1), immune homeostasis (LRRC18), and energy metabolism (PDK4), establishing a molecular basis for traditional herbal therapy in veterinary medicine.

The chemical characterization of the single experimental batch used in this study represents a significant advancement in quality control for traditional Chinese veterinary medicine studies. The qualitative UPLC-QE-MS analysis identified multiple bioactive constituents from the nine component herbs, while the quantitative HPLC determination established baseline contents for two key marker compounds: chlorogenic acid (0.70 mg/g) and berberine hydrochloride (0.77 mg/g). This approach mitigates a common limitation in herbal medicine research by ensuring that the observed therapeutic effects in this study can be directly attributed to a defined chemical profile of the experimental batch. While quantification of a single batch does not resolve the root causes of batch-to-batch variability—which is inherently influenced by variations in the active compounds of raw herbal materials—comprehensive quality control measures are in place. In addition to HPLC quantification, Thin-Layer Chromatography (TLC) and other identification methods are utilized to verify the herbal components, ensuring that each production batch meets qualification standards. Although minor fluctuations in specific marker contents may occur, these multi-method controls ensure the overall quality controllability and consistency of Qingying granules. While this study focused on a single batch, the validated HPLC method provides a reproducible framework for subsequent batch release testing and multi-batch consistency evaluation.

The quantitative determination of chlorogenic acid and berberine hydrochloride provides additional mechanistic context. Chlorogenic acid, is a well-established anti-inflammatory compound that inhibits NF-κB signaling and reduces pro-inflammatory cytokine production [22]. Berberine hydrochloride possesses significant antipyretic, antimicrobial, and metabolic regulatory properties, including modulation of glycolytic metabolism through AMPK pathway activation [23]. The calculated intake levels of active ingredients at the optimal dose offer key dose references for future in vivo pharmacology research.

The dose-ranging study identified 0.75 g/kg/day as the optimal therapeutic dose, achieving comparable efficacy to higher doses (1.0–2.0 g/kg/day) while requiring less raw medicinal material per animal, thereby reducing direct medication costs in clinical practice. This dose–response relationship demonstrates the importance of systematic dose optimization in herbal medicine, where multi-component formulations may exhibit non-linear pharmacokinetics. The chemical characterization of the experimental batch provides a quality control benchmark for clinical application, ensuring that future batches can be standardized to deliver equivalent amounts of marker compounds. Compared to traditional liquid decoctions, the granular formulation improves stability, storage, and transportation convenience.

The transcriptomic data revealed massive transcriptional dysregulation in warm disease (1,955 DEGs), particularly affecting immune and inflammatory pathways. The upregulation of LCN2 (lipocalin-2) aligns with its established role as a pro-inflammatory mediator in acute lung injury and sepsis, where it promotes neutrophil activation and tissue damage [24]. PDK4 upregulation indicates metabolic shift toward glycolysis, consistent with inflammatory states and mitochondrial dysfunction in severe infection [25].

Qingying granules inhibit these pathological changes through coordinated multi-target modulation. The normalization of IL1RL1 (IL-33 receptor) suggests restored regulatory capacity in type 2 immune responses [26], while S100A4 restoration indicates improved tissue repair mechanisms [27]. Metabolically, the reversal of 7-hydroxyetodolac (a COX inhibitor) aligns with the antipyretic and anti-inflammatory properties documented for component herbs [28].

More importantly, our integrated multi-omics analysis reveals a systems-level mechanism characterized by immuno-metabolic coupling during porcine warm disease, which Qingying granules effectively reverse. In severe systemic inflammation, immune cells undergo profound metabolic reprogramming, shifting from oxidative phosphorylation to aerobic glycolysis to meet the rapid biosynthetic and energy demands of inflammation [29]. Our transcriptomic data identified the upregulation of PDK4, a key gatekeeper that inhibits pyruvate entry into the TCA cycle, which directly correlates with the enrichment of the glycolysis/gluconeogenesis pathway in the metabolomic data. Concurrently, the upregulation of pro-inflammatory mediators like LCN2 and the dysregulation of IL1RL1 reflect an overactive immune cascade.

These two systems are not independent; rather, they form a pathological positive feedback loop. The inflammatory signaling (e.g., via PI3K-Akt and cytokine–cytokine receptor interactions) drives the glycolytic shift, while the accumulating glycolytic intermediates and disturbed lipid metabolism further provide the metabolic substrate to sustain chronic inflammation [29]. Recent studies have highlighted that PDK4 functions as a metabolic checkpoint for M1 macrophage polarization, linking enhanced glycolysis to pro-inflammatory phenotype commitment [29]. Similarly, LCN2 has been recognized as a node bridging inflammatory and metabolic pathways in obesity and related metabolic disorders [30].

The therapeutic action of Qingying granules cannot be fully explained by a single-target model. Instead, its multi-component nature enables it to act as a system regulator: by concurrently normalizing PDK4-mediated metabolic reprogramming and resolving LCN2/IL1RL1-driven inflammation, Qingying granules break the pathological immuno-metabolic feedback loop. This simultaneous restoration of both immune homeostasis and metabolic balance provides a robust systems-level mechanistic basis for the therapeutic effects of traditional herbal formulations [29].

Notably, the metabolic pathway analysis from our multi-omics data revealed modulation of glycolysis/gluconeogenesis, which aligns precisely with the known metabolic regulatory effects of berberine [31]. This convergence between chemical quantification and systems biology findings strengthens the mechanistic interpretation of the hypothesized mode of therapeutic action of the Qingying granules and supports the selection of these compounds as quality control markers for the Qingying granules formulation, ensuring both chemical consistency and biological relevance across production batches. With global pressure to reduce antibiotic use in livestock production, Qingying granules offer a scientifically validated alternative strategy for managing warm diseases through multi-target immune regulation and anti-inflammatory effects. This reduced the clinical reliance on prophylactic and therapeutic antibiotics in swine production. The established molecular mechanisms support regulatory approval pathways in jurisdictions requiring mechanistic evidence for herbal veterinary products.

Limitations and Future Directions: This study has several limitations. First, due to the life-threatening nature of porcine warm disease and ethical considerations to minimize animal suffering, a Model Control group (untreated diseased animals) was not included in the dose-ranging study (Phase 1). All diseased pigs received either Qingying granules treatment or positive drug control to ensure therapeutic intervention. Consequently, the optimal dose determination relied on comparisons across treatment groups rather than direct comparison with untreated disease progression. This design limitation was partially mitigated in Phase 2 (mechanistic study), where a Model Control group was included for transcriptomic and metabolomic comparisons. Second, while we employed a natural disease model with rigorous clinical criteria, the absence of comprehensive pathogen identification limits etiological precision. The observed syndrome complex may represent mixed infections common in field conditions, which enhances clinical relevance but complicates mechanistic interpretation. Future studies should incorporate metagenomic screening to characterize the microbial landscape, including the pathogenic and commensal microbiomes. Third, while the single-batch approach ensures internal consistency, future studies must validate these findings across multiple production batches. Given the inherent variability of active compounds in herbal materials due to environmental factors (e.g., soil, light), conducting compositional analyses on multiple batches is essential. This approach will allow us to establish the acceptable effective range of marker compounds required to reproduce the desired therapeutic effects, which is critical for robust quality control. Fourth, functional validation (e.g., siRNA knockdown or overexpression of LCN4 or PDK4) is required to establish definitive causality between gene expression changes and clinical outcomes. Finally, multi-tissue analysis (spleen, lymph nodes, liver) would provide comprehensive immune profiling.

Conclusion

Qingying granules at 0.75 g/kg/day exert potent therapeutic effects on porcine warm disease through multi-targeted regulation of inflammation, immune homeostasis, and metabolic balance. Integrated transcriptomic and metabolomic analyses identified key pathogenic genes (IL1RL1, S100A4, LCN2, LRRC18, PDK4) and metabolic perturbations that were normalized by treatment. These findings provide scientific validation for traditional herbal therapy in veterinary medicine, establish quality control benchmarks, and support clinical application in antibiotic-free swine production systems.

Authors’ contributions

Y. L conceived and designed the work. S. L coordinated technical support and funding. Z. Y wrote the original draft. Z. Y, H. Q, Z. Z, B. Y, S. Y and K. L performed the experiments and collected the samples. Y. L and H. Q reviewed the manuscript. All authors have read and approved to the published version of the manuscript.

Funding

This work was supported by the National Key Research and Development Program of China (2023YFD1800804-06), the National Natural Science Foundation of China (Nos. 32302918 and 32302919) and Shandong Provincial Key Research and Development Program (Science and Technology Innovation Boosting Action Plan for Rural Revitalization, No. 2023TZXD083).

Data availability

The raw sequencing data from this study was uploaded to the Genome Sequence Archive (GSA) database with the submission identifier CRA040839 (https://ngdc.cncb.ac.cn/gsa/). All other data supporting the findings are included within the article and its supplementary files.

Declarations

Ethics approval and consent to participate

All experimental procedures involving animals were reviewed and approved by the Laboratory Animal Ethics Commission of the Poultry Research Institute, Shandong Academy of Agricultural Sciences (Approval No. JQS-2024–16). The study was carried out on a commercial swine farm with informed consent from the owner.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.