Dietary Ellagic Acid Ameliorated Clostridium Perfringens-Induced Subclinical Necrotizing Enteritis In Broilers Via Regulating In ammation Signaling Pathways And Cecal Microbiota to Inhibit Intestinal Barrier Damage

Yu Tang (  2356771229@qq.com ) China Agricultural University https://orcid.org/0000-0002-2620-877X Xinyue Zhang China Agricultural University College of Animal Science and Technology Yanan Wang China Agricultural University College of Animal Science and Technology Yongpeng Guo China Agricultural University College of Animal Science and Technology Peiqi Zhu JIANGSU LIHUA ANIMAL HUSBANDRY CO., LTD. Guiguan Li COFCO feed Co., Ltd. Jianyun Zhang China Agricultural University College of Animal Science and Technology Qiugang Ma China Agricultural University College of Animal Science and Technology Lihong Zhao China Agricultural University College of Animal Science and Technology


Introduction
Necrotizing enteritis (NE) is a common in ammatory disease of small intestine caused by Clostridium perfringens. NE poses an important threat for various animals, including chickens, pigs, sheep, and goats. C. perfringens is a spore-forming, anaerobic, gram-positive bacterium and an opportunistic pathogen found in the environment and the intestinal microbiota of humans and animals [1]. C. perfringens strains vary signi cantly in toxin production and has been divided into types A-G based on the presence of encoding genes for alpha (CPA), beta (CPB), epsilon (ETX), iota (ITX), NetB, and CPE toxins [2]. In poultry, NE is caused mainly by type A strains, which lead to economic losses of $6 billion annually in the global poultry industry [3].
NE usually occurs in broiler at 2-to 6-week-old, including acute clinical NE and subclinical NE (SNE) [4]. Acute clinical NE is characterized by diarrhea, bloody feces, intestinal ulcer erosion, peracute course, and high mortality [5]. Whereas the ock suffering from SNE presents no overt clinical signs and low mortality [4], in most cases, even only an overall reduction in broiler performance is observed [6]. As a result, SNE is di cult to diagnose and control timely, leading to more widespread infections and greater economic losses than acute clinical NE [6]. In 2015, a study reported that in drug-free broiler ocks of eight commercial farms in Canada, 27.45% of the ocks suffered from acute clinical NE, and 49.02% of the ocks suffered from SNE [1]. In addition, previous studies [7,8] have demonstrated that SNE usually results in pathological changes of intestinal structure, damages of intestinal barrier function, activation of intestinal in ammatory pathways, disorders of intestinal micro ora, poor digestion and absorption, and depressing growth performance of broilers. Therefore, modulation in intestinal health may be a potential strategy to prevent and treat SNE in broiler.
With banning of antibiotics, various strategies have been used against SNE in broiler; apart from organic acids [9], polysaccharides [10], vaccines [11], prebiotics [8], and probiotics [7], plant extracts have been demonstrated to be effective for its protection on broiler health [12]. Ellagic acid (EA) is a chromene-dione derivative (2,3,7,8tetrahydroxy-chromeno [5,4,3-cde]chromene-5,10-dione; C14H6O8) extracted from various fruits, nuts, vegetables, and herbs [13]. EA possesses numerous pharmacological activities, including antioxidant activities [14], anti-in ammatory activities [15,16], anti-cancer activities [13], and anti-metabolic syndrome activities [17]. In rats, EA exerted anti-in ammatory and antioxidant functions against streptozotocin -induced diabetic nephropathy via reduced the activation of NF-κB and increased the nuclear translocation of Nrf2 to up-regulate GSH, γ-GCL, and SOD activities [18]. Meanwhile, the alleviating effect of EA on in ammatory mediators has also been widely reported in mice or rats, including TNF-α, IL-1β, IL-6, IL-8, and iNOS, through TLRs, NF-κB, and STAT signaling pathways [15,16,19]. In addition, EA can alter intestinal microbiota composition and be transformed to urolithins, possessing the potential protection against oxidative stresses and in ammatory diseases in gastrointestine [20]. Ellagitannins (ETs), which can be hydrolyzed to EA in digestive tract, show a prebiotic effect on promoting the growth of Lactobacillus and Bi dobacterium [21]. In human, pomegranate ETs can increase the counts of Akkermansia mucini la [22], improving the host metabolic functions and immune responses. However, no study has investigated the protective in uences and mechanisms of EA against intestinal diseases (especially SNE) in poultry to date. Therefore, this study was undertaken to explore the preventing effects and mechanisms of EA on growth performance, immune response, intestinal barrier function, antioxidant capacity, and intestinal micro ora of broilers suffered with SNE induced by C. perfringens.

Materials And Methods
Experimental animals and treatments A 2 × 2 factorial randomised complete block design was used to investigate the effects of dietary EA level (0 or 500 mg/kg), C. perfringens challenge (challenged or unchallenged), and their interactions on broiler. A total of 240 1-day-old male Arbor Acres broilers with an average weight at 40.4 g (SD 1.57) were purchased from Beijing Arbor Acres Poultry Breeding Co., Ltd (Beijing, China). Upon arrival, birds were weighed and randomly assigned to four groups. Each group had six replicates with ten birds per replicate. Each replicate was reared in a cage (1.0m × 1.0m × 0.6m, length × width × height) with a raised wire-netted oor. The treatments of groups were as follows: (1) control group (Control, basal diet); (2) C. perfringens challenge group (CP, basal diet + C. perfringens challenge); (3) ellagic acid and C. perfringens challenge group (EAXCP, basal diet extra 500mg/kg ellagic acid + C. perfringens challenge); (4) ellagic acid group (EA, basal diet extra 500mg/kg ellagic acid). Ellagic acid (99%, extracted from pomegranate peel) was purchased from Shaanxi Pioneer Biotech Co., Ltd (Shaanxi, China).
Corn-soybean meal basal diets were formulated according to the nutrient requirements for broilers as recommended by the National Research Council (NRC, 1994) [23]. The composition and nutrient levels of the basal diets are presented in Table S1. All diets were crumbled, and neither antibiotics nor antibacterial supplements were added. To avoid cross-contamination, the unchallenged birds and C. perfringens-challenged birds were reared in two separate parts in a light and climate controlled room at a 23-hours light/1-hour dark cycle, and provided with feed and water ad libitum. Room temperature was maintained at 33°C during the rst 5 days and then gradually decreased by 5°C weekly until 23±1°C. In addition, birds were vaccinated against Newcastle disease virus and infectious bronchitis virus vaccines on day 7 and against bursa disease virus via drinking water on day 12 and 26 according to the routine immunization program.

Clostridium perfringens Challenge
Avian C. perfringens type A eld strain (CVCC2030) was obtained from State Key Laboratory of Animal Nutrition (Beijing, China). C. perfringens culture and challenge was performed on the basis of the previous reports [24][25][26], with some modi cations. Brie y, C. perfringens was anaerobically cultured in cooked meat medium with dried meat particles (CM605, CM607; Beijing Land Bridge Technology Co., Ltd) for 24 h at 37°C, then aseptically transferred into thioglycolate broth (70157, Millipore) and incubated anaerobically for 18 h at 37°C. Birds from CP and EAXCP groups were challenged with 1.0 mL of actively growing culture of C. perfringens at 2~3 × 10 8 CFU/mL by oral gavage each day from day 16 to 20, while those from Control and EA groups received an equal volume of thioglycolate broth.

Growth performance
On day 21 and 42, the birds were feed-deprived for 8 hours, and then the feed intake and body weight (BW) of the birds in each replicate were measured. The average daily feed intake (ADFI), average daily gain (ADG), and feed conversion ratios (FCR, feed intake/BW gain.) of the birds were calculated for days 1-21, 22-42, and 1-42, respectively.

Sample collection
At day 42, one bird per replicate was randomly selected, for blood samples collected by wing vein puncture. Serum was separated by centrifugation at 3000 rpm for 10 min at 4°C. After the birds were euthanized by jugular exsanguination, approximately 1 cm long jejunal samples between Meckel's diverticulum to the proximal of jejunum were collected and snap-frozen in liquid nitrogen; approximately 2 cm long jejunal samples in length midway between the endpoint of the duodenal loop and Meckel's diverticulum were collected, ushed, and xed with 10% neutral buffered formalin solution for morphological analysis [26]. Jejunal mucosa were scraped from the posterior part half of jejunum. Cecal content samples were aseptically collected and snap-frozen in liquid nitrogen. Serum, cecal content samples, and all of the tissues were stored at -80°C until analysis.

Plasma D-xylose concentration
Plasma D-xylose concentration was measured as Zhang et al. [24]. Brie y, at day 42, the other one feeddeprived bird per replicate was randomly selected, weighed, and administered D-xylose (X1500; Sigma-Aldrich) solution at a dose of 0.1 g/kg body weight (infused with 10 % D-xylose) by oral gavage. After 1 hour, blood samples were collected into heparinized vacuum tubes by wing vein puncture. Plasma was separated by centrifugation at 3000 rpm for 10 min at 4°C and stored at -80°C. D-xylose concentration in plasma was determined using the D-xylose assay kit (A035; Nanjing Jiancheng Bioengineering Institute) according to the manufacturer's instructions.

Intestinal morphology
Fixed jejunal tissues were embedded in para n, then tissue rings were sliced into 5-µm thickness, depara nized in xylene, rehydrated, and mounted on glass slides [25,26]. Sections were stained by haematoxylin and eosin (H&E). The slides were photographed on a microscope slide scanner (3D HISTECH Ltd, Budapest, Hungary, Model Pannoramic MIDI). At least nine villi per section and two sections each sample were measured to evaluate villus height (VH), and crypt depth (CD) using CaseViewer (V 2.43). The means of villus height and crypt depth were calculated and used to obtain the villus height/crypt depth (V/C).
Intestinal immune and tight junction-related genes expression Total RNA was extracted from jejunal tissues using Eastep® Super Total RNA Extraction Kit (15596018; Promega Bingjing Biotech Co., Ltd) according to the manufacturer's instructions. The concentration and purity of total RNA were determined on an Ultra-micro spectrophotometer (IMPLEN, NanoPhotometer® N60). Total RNA from each sample was reverse-transcribed into complementary DNA using a TRUEscript RT Kit (+gDNA eraser) (PC5402; Aidlab Biotechnologies CO. Ltd). Two-step quantitative real-time PCR was performed with a Sybr Green qPCR Mix (PC3302; Aidlab Biotechnologies CO. Ltd) on a Real-Time PCR Detection Systems (Bio-Rad, CFX Connect™) according to the manufacturer's instructions. Oligonucleotide primers of in ammatory mediator genes (TNF-α, IL-1β, IL-8, IFN-γ, TGF-β, and iNOS), in ammation-related signaling pathway genes (TLR2, TLR4, MyD88, NF-κB, JAK1, JAK2, JAK3, STAT1, and STAT6), and tight junction-related genes (ZO-1, occludin, and claudin-2) for chicken were designed based on databases of National Center for Biotechnology Information (NCBI) using Oligo (V 7.0) and synthesized by Sango Biotech Co., Ltd (Shanghai, China). Table S2 lists the quantitative real-time PCR primers used in this study. The relative mRNA expression levels of each target gene were calculated based on the expression of the housekeeping gene β-actin using the 2 −ΔΔCt method [27].
Cecal microbiota pyrosequencing and analysis Bacterial DNA was extracted from cecal content samples using a QIAamp DNA Stool Mini Kit (51504; Qiagen Inc.) according to the manufacturer's instructions. The concentration and purity of total DNA were determined on an Ultra-micro spectrophotometer (IMPLEN, NanoPhotometer® N60). V4 region of bacterial 16S rRNA gene was ampli ed with the barcoded primer pair 515F/806R (515F: 5′-GTG CCA GCM GCC GCG GTA A-3′, 806R: 5′-GGA CTA CHV GGG TWT CTA AT-3′) using PCR, then PCR products run on a 2% agarose gel and were puri ed using a QIAquick Gel Extraction Kit (28706; Qiagen Inc.) according to the manufacturer's instructions. Pyrosequencing for 16S rDNA was performed on a high-throughput sequencing platform (Illumina, HiSeq® 2500 Miseq PE250).
Sequencing results were merged using FLASH (V 1.2.7), ltered using QIIME (V 1.9.1), and the chimera sequences were excluded based on Silva database using UCHIME (V 4.1) to obtain effective tags nally. The effective tags with ≥ 97% similarity were assigned to the same OTUs using Uparse (V 7.0.1001), and the taxonomic information of each operational taxonomic unit (OUT) was annotated based on Silva Database using Mothur (V 1.35.1). Multiple sequence alignment was conducted using MUSCLE (V 3.8.31) to analyse the phylogenetic relationship of different OTUs and the difference of the dominant species among different treatment groups. OTUs abundance information was normalized based on the sample with the most minimal sequences for subsequent analysis.
Venn diagram, rarefaction curve, box plot analysis, principal co-ordinates analysis (PCoA), and bacteria relative abundance were created with R software (V 2.15.3). Alpha diversity, including ACE, Chao1, Simpson, and Shannon index, were calculated using QIIME (V 1.9.1). Beta Diversity was calculated from bray_curtis distance using QIIME (V 1.9.1). Line discriminant analysis effect size (LEfSe) analysis was used to determine the signi cance of the difference between treatments.
All of the procedures were conducted by Novogene Bioinformatics Technology Co. Ltd (Beijing, China).

Statistical analysis
Data was analysed using GraphPad Prism (V 8.0.1). As a 2×2 factorial arrangement, two-way ANOVA was used to determine the main effects of dietary EA level and C. perfringens challenge, and their interaction, Tukey's multiple comparison was used to separate means when interactive effects signi cantly different or had a trend of difference [24,26]. Results are presented as the means ± SEMs. All statements of signi cance were based on P<0·05, and P value between 0.05 and 0.10 was classi ed as a tendency [26].

Growth performance
The growth performances of broilers on BW, ADG, ADFI, and FCR were shown in Table 1  All values are expressed as the means (n = 6). * Signi cant main effect (P<0.05) of C. perfringens challenge or dietary EA level. BW, body weight. ADFI, average daily feed intake. ADG, average daily gain. FCR, feed conversion ratios = g of feed intake / g of BW gain, g/g.

Intestinal morphology
As depicted in Fig. 1 Intestinal mucosa integrity and barrier-related enzyme activities Serum DAO activity was measured to re ect the intestinal mucosa integrity of broilers, as well as the activities of iNOS and LZM in jejunal mucosa to evaluate barrier function. As described in Fig. 3(A, B, and C), the infection of C. perfringens enormously increased the activities of serum DAO, as well as iNOS and LZM (P<0.01) in jejunal mucosa of broilers, however, dietary EA supplement decreased iNOS and LZM activities (P<0.01). Furthermore, a highly interacting effect on iNOS activity (P<0.01) was observed between dietary EA levels and C. perfringens challenge. As the results of multiple comparisons, birds in CP group showed higher iNOS activity than those in the other three groups (P<0.05).
Tight junction-related gene expression in jejunal mucosa    S2 (A)). The Good's coverage estimators (Table S3) and the rarefaction curves (Fig. S2 (B)) indicated that su cient sequencing coverage was achieved.
The alpha diversity of cecal microbiota was shown in Table S4 (Fig. 5E).

Discussion
According to our results, C. perfringens challenge caused damages on jejunal barrier of broilers and increased the permeability of jejunal mucosa, allowing antigenic substances (LPS, etc.) to enter the blood and internal environment, which in turn triggered jejunal in ammation and oxidative stress, as well as systemic in ammation, reducing the ability of intestinal digestion and absorption, nally impaired the growth performance of broilers. However, dietary EA supplementation exerted anti-in ammatory and antioxidant effects in the jejunal mucosa, which protected and improved the intestinal barrier, preventing the invasion of antigenic substances, and nally improved the growth performance of broilers. Meanwhile, the supplementation of dietary EA also relieved the cecal microbiota imbalance caused by the C. perfringens challenge, protecting the health of broilers (Fig. 6).
Toll-Like Receptors (TLRs) are important members of pattern recognition receptors, TLR4 could recognize LPS, which is unique to Gram-negative bacteria, and TLR2 could recognize peptidoglycans (PGN), which is abundant in Gram-positive bacteria [28]. TLRs can trigger subsequent in ammatory responses through MyD88 dependent or independent signaling pathways that activate NF-κB and nally lead to the release of proin ammatory mediators, including TNF-α, IL-1β, IL-6, IL-8, and iNOS [28]. In in ammatory bowel disease, LPS or cytokines (E.g IL-6 and IFN-γ) can activate JAK/STAT signaling pathway to regulate the expression of proin ammatory mediators, including Claudin-2 and iNOS [29]. In our results, C. perfringens challenge increased the mRNA expression of TLR4, TLR2, NF-κB, JAK3, and STAT6, while dietary EA supplement relieved these adverse effects. Due to the de ciency of appropriate antibodies available for use in studies of chickens, we did not determine the protein levels and phosphorylation status of components in these signaling pathways. A series of studies [7,12,30] have reported the activation process of TLR/NF-κB or JAK/STAT signaling pathways in broilers with C. perfringens challenge. Similar to our results, EA was proved to possess a protective effect on concanavalin A-induced hepatitis in mice via decreasing the expressions of TLR2 and TLR4 and suppressing NF-κB signaling pathway [19]. C. perfringens challenge in this study has no obvious effect on the mRNA expression of MyD88, which may indicate TLRs activate NF-κB through MyD88 independent signaling pathways. Some studies [29,31] have reported EA inhibited the phosphorylation of JAK1, JAK2, STAT1, and STAT3 to exert anti-in ammatory effects in keratinocytes or rats, but no report has been found on the impact of EA on JAK3/STAT6 in any animals. In human and mice, the activation of JAK3/STAT6 pathway was related to the differentiation of monocytes and the enhancement of Th2 in ammatory response (the release of IL-4, IL-5, and IL-13) [32]. It means that C. perfringens challenge may trigger the Th2 in ammatory response related to the JAK3/STAT6 pathway in jejunal mucosa of broilers, while EA relieves this hazard in this pathway.
During the in ammatory response, the activation of TLR/NF-κB and JAK/STAT signaling pathways can induce the release of a variety of pro-in ammatory cytokines, which will lead to the activation of immune cells and the production of more cytokines [33]. TNF-α and IL-1β are pleiotropic pro-in ammatory cytokines, whose dysregulations are linked with a wide range of pathological conditions, such as infection, metabolic syndrome, and in ammatory bowel disease [33]. IL-8 is a very potent trigger to immune cells' migration and proliferation, which guides neutrophils to the direction of in ammation [33]. TGF-β and IFN-γ also play an important role in a variety of in ammation-related diseases; C-reactive protein promotes the in ammatory response of atrial brillation through the overexpression of TGF-β related to the TLR4/NF-κB/TGF-β pathway in HL-1 cells, which is related to heart arrhythmia [34], while IFN-γ was reported to contribute to the hepatic in ammation in HFDinduced nonalcoholic steatohepatitis by STAT1β/TLR2 signaling pathway in mice [35]. iNOS is related to immune response via macrophage defence mechanism, its expression and the increase of NO levels can cause various in ammation-related pathophysiological conditions, the cell wall components of bacteria (mainly through LPS) can activate the JAK/STAT signaling pathway and subsequently activate NF-κB to initiate iNOS transcription [36]. In our study, C. perfringens challenge caused up-regulations on the mRNA expressions of proin ammatory mediator genes TNF-α, IL-1β, and iNOS in jejunal mucosa, while the EA diets down-regulated the mRNA abundances of TNF-α, IL-1β, IL-8, and iNOS. A series of studies [12,24,30] have proved that C. perfringens challenge can cause an up-regulation on pro-in ammatory mediator genes in the intestine of broilers, including TNF-α, IL-1β, IL-8, TGF-β, IFN-γ, and iNOS. Meanwhile, the alleviating effect of EA on in ammatory mediators, including TNF-α, IL-1β, IL-8, and iNOS has been widely reported in mice or rats [15,16], which are in line with our results and further indicated that EA reduced in ammatory mediators in broilers may be through NF-κB and STAT signaling pathways. However, the C. perfringens challenge or dietary EA levels show no signi cant effect on TGF-β and IFN-γ in our results, which may be related to the difference on C. perfringens strains and frequency of the challenge.
The activation of in ammatory pathways and the release of in ammatory mediators can affect the antioxidant, barrier, and absorption functions of the jejunum, which are vital to the growth performance and health of broilers.  [14], which may jointly explain the antioxidant mechanism of EA in SNE induced by C. perfringens. Moreover, in the oxidized sh oil-induced oxidative stress of mice, the supplementation of EA in diet increased the total antioxidant capacity (T-AOC) and the activities of the glutathione peroxidase (GSH-Px) and SOD, while decreased the MDA concentration in the intestine [38].
Another report demonstrated that EA exerted anti-in ammatory and antioxidant functions against streptozotocin-induced diabetic nephropathy in rats via reducing the activation of NF-κB and increasing the nuclear translocation of Nrf2 to up-regulate GSH, γ-GCL, and SOD activities [18].
Tight junction proteins are vital structures of the physical barrier in jejunal mucosa, which form a seal between intestinal epithelial cells and prevent the transmission of macromolecules [7]. In the present study, C. perfringens challenge decreased the jejunal mRNA expressions of ZO-1 and occludin in broilers and increased the mRNA expression of claudin-2, the dietary supplementation of EA relieved these adverse effects. ZO-1 and occludin are barrier-forming proteins, whose reduction mean damage of tight junctions; whereas claudin-2 is a pore-forming protein, whose increase can increase the permeability of intestinal barrier [24]. As many studies reported [9,12], the infection of C. perfringens can reduce the mRNA expression of ZO-1 and occludin in broilers through the activation of NF-κB. While pomegranate and pomegranate leaf, which rich in EA, can relieve the decrease of ZO-1 and occludin caused by alcoholic liver disease or hyperlipidemia in the intestine of mice [17].
The infection of C. perfringens can increase the expression of claudin-2 in intestine of broilers [24], which may be explained as a result of 'cross-talk' caused by IL-6 between JAK/STAT, SAP/MAPK, and PI3K signaling pathways [39]. Interestingly, the mRNA expression of occludin was increased in the broilers only fed the diet with EA supplementation; in another study [12], thymol and carvacrol supplementation demonstrated a similar effect on the mRNA expression of occludin in broilers challenged with C. perfringens.
D-xylose crosses the intestinal mucosa via a Na + -dependent mobile-carrier mechanism, in the case of malabsorption syndrome, the entry of D-xylose from the gut lumen to the portal vein is damaged, resulting in reduced concentrations of D-xylose in blood [24]. DAO is an intracellular enzyme in the small intestinal epithelia and released into the peripheral circulation as a result of intestinal villi damage, so the level of serum DAO could re ect the severity of intestinal mucosal injury [40]. In our study, the decrease of plasma D-xylose concentration indicated that C. perfringens challenge had impaired the intestinal absorption function, while the increase of DAO activity in serum may imply the relation to the impaired intestinal epithelium. The supplement of dietary EA alleviated the decrease of plasma D-xylose concentration, but had no effect on DAO activity in serum. Similar to our results, the arginine additive alleviated an increase on plasma D-xylose concentration caused by the C. perfringens challenge [24]. LZM can cleave peptidoglycan of the cell wall in Gram-positive bacteria, resulting in the loss of cellular membrane integrity and cell death [26]. In our results, C. perfringens infection increased the activities of iNOS and LZM in jejunal mucosa, while the supplement of EA in diet relieved these adverse effects. LZM was up-regulated in the gastrointestinal tract of patients affected by chronic in ammation, which was related to the LZM-mediated processing of luminal bacteria in the colon that triggered the pro-in ammatory response [41]. These up-regulations of iNOS and LZM in our result further explained the mechanism of chronic in ammation caused by SNE.
Damage of the intestinal barrier and absorption function was also re ected in the microstructure of jejunum. The results of jejunal morphology, including VH, CD, and V/C ratio by HE staining, were important indexes that intuitively re ected jejunal health and absorption surface. C. perfringens challenge seriously destroyed the villi structure and reduced the absorption surface of nutrient, which is in line with the results reported previously [12,25]. On the contrary, the dietary EA supplementation alleviated the jejunal lesions, indicated the good condition of enterocytes and e cient ability of nutrient absorption. In the mice model [38], EA effectively alleviated the intestinal damage caused by oxidized sh oil via signi cantly increasing the VH and V/C, while improving the mucous epithelium injury. Also, thymol and carvacrol alleviated the ileal lesion and improved V/C ratio in broilers with C. perfringens infection [12]. Furthermore, the antioxidant and anti-in ammatory effects of EA may explain the mechanism that is bene cial to the health of intestine and villus-crypt architecture.
Intestinal NE lesions and mucosal atrophy greatly compromises epithelial permeability and mucosal barrier function, which may result in adverse effects on internal environment homeostasis and production performance of broilers, therefore, these serum in ammation biomarkers were used to evaluate the systemic in ammatory response intensity of broilers. LPS is an endotoxin produced by Gram-negative bacteria, its increase in blood re ected the bacteria translocation to liver, spleen, and blood [42]. IL-6 is an important cytokine of in ammatory bowel diseases, which can activate the JAK/STAT signaling pathway and promote the release of various in ammatory factors [29]. CRP is synthesized in liver, mainly in response to IL-6, and can be combined with the pathogen LPS to activate the classical complement pathway [43]. PCT is a diagnostic marker of bacterial infection, which is produced by LPS, TNF-α, and IL-6 acting on neuroendocrine cells or special cells in the liver and spleen [44]. MPO is a sign of neutrophil aggregation and in ammation, its activity is a marker of neutrophil in ltration into the intestine [44]. C. perfringens infection increased the concentrations or activities of LPS, IL-6, CRP, PCT, and MPO, causing a higher stress of systemic in ammatory response in broilers, while the supplement of EA in diet relieved these adverse effects. In line with our results, Lactobacillus acidophilus supplementation signi cantly decreased the serum LPS content in broilers with C. perfringens challenge [29], while EA treatment can decrease the mRNA expressions of TNF-α and IL-6 in the liver and intestine of oxidative stress mice [38]. In NE model caused by C. perfringens, probiotic powder containing Lactobacillus plantarum decreased the MPO activity in the ileum mucosa of broilers [37]. Overall, our study re ected that EA alleviated the systemic in ammatory response caused by C. perfringens challenge, possibly by protecting the integrity of intestinal mucosa and reducing the expression of in ammatory factors.
On the other hand, intestinal microbiota is involved in intestinal nutrition, defense, and immunity. The high diversity of intestinal microbiota is bene cial to maintain the stability of the intestinal microenvironment and defend against the invasion of pathogenic microorganisms [45]. In our study, only the dietary EA supplement increased the alpha diversity, including observed species and Shannon index, which may mean an improvement in intestinal health; but the beta diversity analysis showed no difference of the microbial community structure among groups, which may be related to the microbiota from different parts of the intestine and the time of sample collection. In broilers challenged by C. perfringens and Eimeria [46], the effects of dietary lauric acid supplement or the challenge on microbiota in the jejunum were distinct from those in the cecum, as well as the change of microbiota was more signi cant in jejunum; however, these treatments did not promote signi cant difference of taxa abundance and diversity in cecum, which may explain our results. In terms of microbial abundance, C. perfringens challenge increased the relative abundance of Firmicutes and decreased the relative abundance of Desulfobacterota. Similarly, EA increased Firmicutes relative abundance and showed a trend of lowering Desulfobacterota and Campilobacterota relative abundances. It has been reported that Firmicutes improved the utilization of energy in the diet and the ratio of Firmicutes to Bacteroides was often positively associated with weight gain [25]. However, both EA and C. perfringens challenge resulted in an increase in Firmicutes abundance and a decrease in Desulfobacterota abundance, which may be related to the longer time interval between challenge and sample collection, as well as the immune regulation of broilers, especially in the EAXCP group, which may mean that the challenge has played an immune-stimulating effect with the presence of EA. In rats with stress-induced depressive-like behavior [47], fecal microbiota transplantation ameliorates gut microbiota imbalance and intestinal barrier damage through increasing Firmicutes and decreasing Desulfobacterota and Bacteroidetes at phylum levels; this treatment also reduced the loss of villi and epithelial cells, suppressed the in ammatory cell in ltration, and increased the expression of ZO-1 and occludin in the ileum. These results were amazingly similar to ours, which may indicate that the microbiota displayed a similar mechanism in the intestinal protection of broilers. Campylobacter was believed to be closely related to the zoonotic campylobacter disease [48], the EA-induced decrease of Campilobacterota relative abundance may have a protective effect on the health of broilers. At the genus level, only the main effect of C. perfringens challenge showed a trend of heightening Ruminococcus]_ torques_group abundance. The increase of [Ruminococcus]_ torques_group abundance was reported in irritable bowel syndrome, which phylotype was associated with severity of bowel symptoms [49]. Another study showed that the Ruminococcus]_ torques_group seemed to be especially involved in controlling paracellular permeability [50], which may be another factor that SNE affects intestinal permeability of broilers in our result.
LEfSe analysis revealed the different phylotypes of cecal microbiota between groups. Compared with cecal microbiota in Control group, the increased abundance of Oscillospiraceae in CP group was thought to be linked to intestinal in ammation [51]. The effect of Butyricicoccaceae abundance on in ammation was lacking in reports, but it was thought to be an important butyrate producer [51], which may be bene cial to the recovery of the intestines. Gordonibacter_pamelaeae has been reported to have the function of transforming EA into urolithin [20], its high abundance was also observed in the EAXCP group, which may mean the transfer of the microbiota between the CP and EAXCP groups, because they were kept in the same pheasantry room.
Compared with cecal microbiota in Control group, a main increase on the abundances of Turicibacter_sp_H121 and Romboutsia_ilealis in EA group was found in our results. The increase on Turicibacter_sp_H121 was also observed in cecal microbiota of EAXCP group compared with CP group, which may indicate that EA is bene cial for Turicibacter_sp_H121, but the effect of the increase is unclear. Romboutsia ilealis is a bene cial intestine bacterium, whose decrease in the response to Streptococcus agalactiae infection of zebra sh was considered harmful [52]. The cecal microbiota between EAXCP and other three groups were quite different. In EAXCP group, the increased abundance of Sellimonas has been reported as a potential biomarker of homeostasis gut recovery after dysbiosis events [53], Bacteroidales was thought to be involved in the synthesis of fatty acids and are bene cial to the health of the host [54], Erysipelotrichaceae was highly abundant in good FCR broilers [55], and mice fed with normal diet possessed more abundant than those fed with high fat diet on Monoglobaceae and RF39 [56]. Rhodobacteraceae is widely reported in aquatic animals or marine environments and has no adverse effects on host health. In Control group, the abundance of Synergistes was reported to be negatively correlated with the levels of IL-1β, IL-6, and TNF-α in serum samples from piglets [57], but Phascolarctobacterium predominated among the Clostridia in low FCR birds [58]. Dietary supplementation with medium-chain a-monoglycerides can decrease the abundances of Cerasicoccus, and improve productive performance and egg quality in aged hens [58]. Comparing with birds in CP group, those in EAXCP group had higher abundance of Faecalibacterium, which was enriched in chickens with the higher BW [59]. It was speculated that Clostridiales_vadinBB60_group might also be bene cial bacteria in intestinal tract [7]. Bi dobacterium_breve was probiotic which has been veri ed [60]. Comparing with birds in EA group, those in EAXCP group had a higher abundance of Subdoligranulum, which was negatively correlated with CRP and IL-6 in human [61]. Fusobacterium mortiferum was often reported in clinical infections of human, but its strains, which were isolated from poultry caeca, also can produce bacteriocin-like substances inhibiting Salmonella enteritidis [62]. For another, birds in EA group had higher abundance of Muribaculaceae, which negatively correlated with in ammatory markers in high fat-high sucrose diet-induced insulin resistant mice [63]. Eubacterium_xylanophilum_group was thought to be lactic acid-and SCFA-producing bacteria, which enhanced intestinal homeostasis and ameliorated weaning stress in piglets [64]. Escherichia coli showed higher levels in broilers with smaller BW [59]. In rats, Elusimicrobium was thought to be bene cial bacteria, whose increase can protect the intestinal barrier [65]. Overall, C. perfringens challenge caused an adverse effect on the cecal microbiota of broiler chickens, dietary EA supplementation led to a small bene cial effect, while the simultaneous effect of dietary EA and challenge seems to stimulate the immune function of broilers and made them possess a better cecal microbiota. Furthermore, the cecal microbiota of the EAXCP group seems very different from other groups, which may explain the signi cant interaction between dietary EA level and C. perfringens challenge in our results.
Finally, growth performance is the most comprehensive indicator of commercial broiler quality. SNE induced by C. perfringens usually reduces the performance of broilers without serious clinical symptoms and high mortality [4,6].

Conclusion
In summary, we found that dietary ellagic acid ameliorated C. perfringens-induced SNE in broilers via regulating jejunal in ammation signaling pathways TLR/NF-κB and JAK3/STAT6, relieving jejunal oxidative stress, and balancing cecal microbiota to inhibit intestinal barrier damage, prevent systemic in ammatory response, and improve nutrient absorption capacity, nally protect and enhance growth performance. Declarations Ethics approval and consent to participate Broilers were cared for in accordance with the guidelines for the care and use of laboratory animals presented in the guide issued by the National Institute of Health and by China's Ministry of Agriculture. All experimental procedures were approved by the Animal Care and Use Committee of China Agricultural University (Approval NO.AW13301202-1-13).

Consent for publication
Not applicable.
Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Con icts of Interest
All of the authors declare that they have no con ict of interest.     between EA and EAXCP groups. These Fig.s showed the bacteria of which the LDA Score is greater than the set value (the default setting was 3.0) between groups. The length of the histogram represents the size of the difference species (i.e., LDA Score), and the different colors represent the different groups.

Figure 6
Dietary ellagic acid ameliorated C. perfringens-induced SNE in broilers via regulating in ammation signaling pathways TLR/NF-κB and JAK3/STAT6, as well as cecal microbiota to inhibit intestinal barrier damage. ↑ Arrows indicate the effect of stimulation; ┬ Arrows indicate the effect of suppression.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.