Effects of Ligilactobacillus salivarius on the control of pullorum disease and cecal microbiota in red-feathered native chickens

Abstract

Pullorum disease (PD), caused by Salmonella Pullorum (SP), remains a persistent challenge in native chicken production in Asia. Recurrent outbreaks and reliance on antibiotics have raised concerns about antimicrobial resistance. This study established a reproducible clinical PD model in red-feathered native chickens (RFCs) and evaluated Ligilactobacillus salivarius (LS) as a potential alternative to antibiotic. Oral administration of a field SP isolate (SPB6) at 1 × 10⁸ CFU per chick for four consecutive days induced typical PD signs and persistent bacterial colonization, whereas a single-dose challenge failed to produce consistent disease. Using this model, 100 SP-free RFCs were randomly assigned to five groups of 20 RFCs each: SP challenge only (A), SP + amoxicillin treatment (B), LS prophylaxis + SP (C), SP + nine-day LS treatment (D), and an unchallenged control group (E). Both amoxicillin and LS treatments reduced SP shedding and tissue colonization; notably, nine-day LS regimen achieved sustained suppression of SP isolation rates and bacterial loads comparable to those observed with amoxicillin on days 7, 10, and 17 after infection. Metagenomic analysis in cecal microbiota revealed that nine-day LS treatment enriched the abundance of short-chain fatty acid-producing species, such as Faecalicatena contorta and Lacrimispora saccharolytica, which are associated with intestinal integrity and immune resilience. In conclusion, LS reduced SP shedding and intestinal colonization, with greater efficacy following prolonged administration. LS also modulated the cecal microbiota in PD-affected RFCs by increasing the relative abundance of beneficial taxa. These findings provide experimental support for the evaluation of LS as a potential alternative to antibiotics for PD control. Further studies that extend the duration of LS administration are warranted and are likely to enhance its protective effects.

Introduction

Pullorum disease (PD) is a highly contagious infection in chickens caused by Salmonella Pullorum (SP). The pathogen affects both breeding and meat-type chickens and can trigger severe outbreaks in chicks under three weeks of age, with high mortality resulting from vertical and horizontal transmission (El-Saadony, et al., 2022). PD outbreaks cause considerable economic losses in poultry production worldwide (Xu et al., 2018). In many countries, control programs have successfully reduced the prevalence of PD or even eliminated it from commercial flocks, such as broiler populations (Barrow and Freitas Neto, 2011). For example, routine screening and the culling of infected breeding stocks are regularly implemented to maintain flock health. However, PD continues to pose a significant threat to native chicken populations in parts of Asia and other regions. The persistence of infection is largely attributed to the recurrent introduction of infected native chickens into breeder flocks and the lack of consistent test-and-removal programs for these populations.

Antimicrobials are commonly used to prevent and control PD in chickens. However, such treatments rarely achieve complete elimination of SP from infected hosts. The pathogen can invade and persist within intestinal epithelial cells and macrophages, allowing replication while evading host immune defenses (Wigley et al., 2001). Consequently, SP often survives antimicrobial therapy, and infected chickens remain asymptomatic carriers capable of transmitting the pathogen to susceptible flocks. Recurring outbreaks compel farmers to rely on repeated antibiotic use, which not only increases production costs but also contributes to the development of antimicrobial resistance (Sun et al., 2021). To address these concerns, many countries have restricted the use of antibiotics in animal production to mitigate the spread of resistant bacteria and safeguard public health (Gao, et al., 2017; Lettini, et al., 2016). These challenges highlight the urgent need for effective antibiotic alternatives that can strengthen host defenses against SP infection and reduce dependence on conventional antimicrobials.

Members of the Lactobacillaceae family, widely used as probiotics in poultry production, have been shown to promote growth (Yin et al., 2023), suppress intestinal pathogens (Chen et al., 2020; Tabashsum et al., 2020; Wang et al., 2021b), and enhance immune responses (Wang et al., 2018; Xu et al., 2020). Within this family, Ligilactobacillus salivarius (LS) is a natural commensal organism in the poultry gastrointestinal tract (Dunne et al., 2001; Guinane et al., 2015). Recent studies have reported that LS can inhibit Salmonella through competitive exclusion and the production of antimicrobial compounds (Saint-Cyr et al., 2017; Wang et al., 2020). Beyond its direct antibacterial effects, LS also supports intestinal health by reinforcing the epithelial barrier and promoting a balanced gut microbiota (Sun et al., 2020; Wang et al., 2023). These attributes make LS a promising candidate as an antibiotic alternative in poultry production.

Since PD remains a major problem in native chicken populations across Asia, antimicrobial treatments continue to be the primary means of control, increasing the risk of drug residues and the emergence of resistant strains. Based on previous findings and supporting evidence from the literature, LS shows strong potential as an antibiotic substitute for managing bacterial infections. Therefore, this study aimed to reproduce clinical PD in red-feathered native chickens (RFCs) as a disease model and to evaluate the prophylactic and therapeutic effects of LS. The efficacy of LS was assessed by quantifying bacterial loads in intestinal and systemic tissues and analyzing the cecal microbiota composition. We anticipate that the use of LS could offer poultry farmers a practical and sustainable strategy to mitigate SP infection and shedding in native chicken populations.

Materials and methods

Ethical statement

All animal experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of National Taiwan University (NTU) and were approved under protocol No. NTU-111-EL-00105.

Chicken selection and housing conditions

One-day-old, Salmonella-free RFCs were obtained from a commercial hatchery in southern Taiwan. All chicks had been vaccinated against Marek’s disease, Newcastle disease, and infectious bronchitis. The absence of Salmonella was confirmed by bacteriological examination of cloacal swabs before placement into the animal biosafety level 2 poultry facility at the Animal Resource Center, NTU. Chicks were randomly assigned to experimental groups and housed in separate acid-resistant plastic pens (180 × 88 × 71 cm) with fresh litter. Environmental conditions were maintained at 50–60% relative humidity, and a 10 h light/14 h dark cycle. Each group of chickens was provided with heating lamps to maintain an appropriate ambient temperature, ensuring that all chicks remained active and in normal physiological condition. Broiler starter feed (Uni-President, Tainan, Taiwan) free of antibiotics and anticoccidials, along with water, was provided ad libitum throughout the experimental period.

Experimental designs

Trial 1: Establishment of the Pullorum Disease Model To establish a PD challenge model, 25 one-day-old Salmonella-free RFCs were randomly assigned to 4 groups: a high-dose challenge group (Group At1; n = 6), a low-dose challenge group (Group Bt1; n = 6), a consecutive low-dose challenge group (Group Ct1; n = 7), and an unchallenged control group (Group Dt1; n = 6). On day 1, chicks in Group At1 and Group Bt1 were orally inoculated with 0.3 mL of the SPB6 strain containing 1 × 10⁹ CFU and 1 × 10⁸ CFU, respectively. Chicks in Group Ct1 received 0.3 mL of SPB6 at 1 × 10⁸ CFU once daily for four consecutive days, whereas Group Dt1 received 0.3 mL of phosphate-buffered saline (PBS) on day 1 as a negative control. Cloacal swabs were collected on days 1, 3, 7, and 14 post-challenge to monitor SP shedding. On day 14, all RFCs were humanely euthanized by CO₂ inhalation, and tissue samples (spleen, liver, and cecum) were collected for bacterial enumeration by viable counts on brilliant green phenol-red lactose sucrose (BPLS) agar (Himedia, Mumbai, India) agar. The experimental design is summarized in Table 1.

Table 1.

Experimental design in trial 1.

Group	N	SP challenge (day and dose)		Cloacal sampling (dpc)	Tissue1 collection (dpc)
At1	6	d1	1 × 109 CFU/chick	1, 3, 7, 14	14
Bt1	6	d1	1 × 108 CFU/chick	1, 3, 7, 14	14
Ct1	7	d1-d4	1 × 108 CFU/chick	1, 3, 7, 14	14
Dt1	6	-	-	1, 3, 7, 14	14

All chickens were confirmed negative for Salmonella by cloacal swab culture prior to the trials.

N: numbers of samples; SP: Salmonella Pullorum; dpc: days post-challenge.

A_t1: high-dose challenge group; B_t1: low-dose challenge group; C_t1: consecutive low-dose challenge group; D_t1: unchallenged control group.

¹The collected tissues include liver, spleen, and cecum.

Trial 2: Evaluation of Preventive and Therapeutic Effects of LS The PD challenge model described in Trial 1 was applied in Trial 2 to assess the preventive and therapeutic efficacy of LS against SP infection. A total of 100 one-day-old Salmonella-free RFCs were randomly assigned to five treatment groups (n = 20 per group): SP-infected (Group A), SP + amoxicillin treatment (Group B), LS prophylaxis + SP (Group C), SP + LS treatment (Group D), and unchallenged control (Group E). On day 1, chicks in Group C received 0.3 mL of freshly prepared LS inoculum containing 6.6 × 10⁸ CFU. Beginning on day 2, chicks in Groups A-D were orally challenged with 2.17 × 10⁸ CFU of SP inoculum once daily for four consecutive days. Following SP challenge, amoxicillin was administered to Group B via drinking water at 50 mg/L (50 ppm) for three days. In contrast, Group D received 0.3 mL of LS inoculum containing 8.18 × 10⁸ CFU orally for nine consecutive days. Meanwhile, cloacal swabs were collected on days 4, 7, 10, and 17 post-challenge (P4D, P7D, P10D, and P17D) to monitor SP shedding. On P17D, all RFCs were humanely euthanized by CO₂ inhalation. Tissue samples (spleen, liver, and cecum) were collected for SP quantification by viable counts on BPLS agar. The experimental design is summarized in Table 2. The dosing schedules in the present study were designed to reflect distinct preventive and therapeutic strategies under practical field conditions. The prophylaxis group received a single LS dose before infection to reflect the limited intervention window available immediately after hatch and before exposure to SP in PD endemic regions. In contrast, the treatment group received LS for nine consecutive days after infection to represent a post-exposure intervention strategy. The three-day amoxicillin regimen followed label recommendations and served as a practical comparator. Differences in treatment duration were intentional and reflected strategy-driven designs aligned with field management practices. During the rearing period, all chickens were cared for by trained personnel, and their health status, incidence of diarrhea, and mortality were monitored and recorded daily. Statistical analyses were performed at the end of the experiment.

Table 2.

Experimental design in trial 2.

Group	N	SP challenge (day)	Treatment (day)		Cloacal sampling (dpc)	Tissue1 collection (dpc)
			Amo	LS
A (SP)	20	2-5	-	-	4, 7, 10, 17	17
B (SP + Amo)	20	2-5	6-8	-	4, 7, 10, 17	17
C (LS + SP)	20	2-5	-	1	4, 7, 10, 17	17
D (SP + LS)	20	2-5	-	6-14	4, 7, 10, 17	17
E (CTL)	20	-	-	-	4, 7, 10, 17	17

All chickens were confirmed negative for Salmonella by cloacal swab culture prior to the trials.

N: numbers of samples; SP: Salmonella Pullorum; Amo: Amoxicillin; LS: Ligilactobacillus salivarius; dpc: days post-challenge; CTL: control.

A: SP-infected group; B: SP infected and amoxicillin-treated group; C: LS prophylaxis followed by SP infection; D: SP-infected followed by LS treatment; E: uninfected control group.

¹The collected tissues include liver, spleen, and cecum.

Inocula preparation

A clinically virulent SP isolate (SPB6) was obtained from a chick showing omphalitis and diarrhea on a commercial RFC farm. The isolate was recovered from cloacal swabs of 2-day-old chicks with white diarrhea. Initial isolation was performed on xylose lysine deoxycholate agar (Himedia, Mumbai, India) and incubated aerobically at 37°C for 24 h. Small, dome-shaped, translucent colonies were selected and confirmed as Salmonella by PCR (Kang et al., 2011). Serotyping was conducted using Salmonella O antiserum group D1, factors 1, 9, and 12 (BD Biosciences, San Jose, CA, USA), confirming the isolate as SP. For challenge inoculum preparation, SPB6 was cultured on tryptic soy agar (Neogen, Lansing, MI, USA) at 37°C overnight. A single colony was then transferred to 10 mL tryptic soy broth (Neogen, Lansing, MI, USA) and incubated at 37°C with shaking at 120 rpm for 24 h to reach approximately 1 × 10⁹ CFU/mL. Viable counts were determined by serial dilution plating on TSA in triplicate. The LS strain (BCRC 14759; Food Industry Research and Development Institute, Hsinchu, Taiwan) was prepared for both prophylactic and therapeutic use. Frozen glycerol stocks were streaked onto deMan, Rogosa and Sharpe agar (Neogen, Lansing, MI, USA) and incubated for activation. LS was then inoculated into 20 mL of MRS broth and incubated at 37°C with shaking at 120 rpm for 18 h to reach approximately 1 × 10⁹ CFU/mL. Viable counts were determined by serial dilution plating on MRS agar in triplicate.

Sample collections and bacterial enumeration

Cloacal swabs were collected from all RFCs in each group to assess the detection rate and viable counts of SP. Swabs were immersed in 10 mL of buffered peptone water (Neogen, Lansing, MI, USA) supplemented with 20 µg/mL gentamicin and incubated at 37°C for 18 h. For selective enrichment, 100 µL of the BPW culture was transferred into 10 mL of Rappaport-Vassiliadis broth (Neogen, Lansing, MI, USA) containing 20 µg/mL gentamicin and incubated at 41.5°C for 24 h. Enriched cultures were serially diluted 10-fold, and 100 µL of each dilution was plated onto BPLS agar supplemented with 20 µg/mL gentamicin. Plates were incubated at 37°C for 24 h. SP colonies appeared as smooth, reddish colonies on BPLS agar, and colony counts were determined using the plate count method and expressed as log₁₀ CFU/mL.

After euthanasia, cecal contents were collected and immediately stored at −80°C for microbiota analysis. Tissue samples of the liver, spleen, and cecum were aseptically excised from each RFC. Tissues from the same chicken were pooled, homogenized in PBS, serially diluted 10-fold, and plated on BPLS agar. Plates were incubated at 37°C for 24 h, and bacterial enumeration was performed as described above.

Analysis of cecal microbiota

DNA extraction and 16S rRNA amplicon generation Total genomic DNA was extracted from approximately 200 mg of cecal contents using the CatchGene Stool DNA Kit (QIAGEN, Valencia, CA, USA) according to the manufacturer’s protocol. DNA quality was evaluated by 1% agarose gel electrophoresis, and samples showing sharp, high-molecular-weight bands without smearing were considered desirable for sequencing. DNA concentration was measured with a Qubit 4.0 fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and adjusted to 1 ng/µL. The full-length 16S rRNA gene (V1–V9 regions) was amplified using barcoded primers (forward: 5′-GCATC/barcode/AGRGTTYGA-TYMTGGCTCAG-3′; reverse: 5′-GCATC/barcode/RGYTACCTTGTTACGACTT-3′) with the KAPA HiFi HotStart ReadyMix PCR kit (Roche, Basel, Switzerland). Each reaction contained 2 ng of genomic DNA. The PCR conditions were as follows: initial denaturation at 95°C for 3 min; 25 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, extension at 72°C for 30 s; and a final extension at 72°C for 5 min. Amplicons were verified on 1% agarose gels, and products of approximately 1.5 kb were excised and purified using AMPure PB Beads (Pacific Biosciences, Menlo Park, CA, USA) for library preparation.

SMRTbell library construction and sequencing Equal molar amounts of barcoded PCR products were pooled, and 500–1,000 ng of the pooled amplicons was used for library construction. SMRTbell libraries were prepared using the SMRTbell Prep Kit 3.0 (Pacific Biosciences, Menlo Park, CA, USA) following the manufacturer’s instructions. Briefly, pooled amplicons underwent DNA repair, A-tailing, and adapter ligation, followed by bead-based purification steps to remove unligated fragments and adapter dimers. Purified libraries were incubated with sequencing primer v4 and polymerase using the Sequel II Binding Kit 2.1 (Pacific Biosciences, Menlo Park, CA, USA). Sequencing was performed on the PacBio Sequel IIe System in circular consensus sequencing mode using a SMRT Cell 8M and the Sequel Sequencing Kit 2.0. Runs were conducted for 30 h, and high-fidelity long reads (HiFi reads) with a minimum Phred quality score of 30 were retained for analysis. After demultiplexing, reads were processed using the Divisive Amplicon Denoising Algorithm 2 (DADA2, v1.14) (Callahan et al., 2016) for quality filtering, de-replication, amplicon sequence variant (ASV) inference, and chimera removal. Reads were trimmed and filtered with a maximum threshold of two expected errors per read, yielding ASVs at single-nucleotide resolution across the full-length 16S rRNA gene. Taxonomic classification of ASVs was performed against the NCBI database (Balvočiūtė and Huson, 2017) using the QIIME2 feature-classifier (Bokulich et al., 2018) and consensus BLAST (Camacho et al., 2009) algorithms. Multiple sequence alignment of ASVs was conducted with MAFFT (Katoh and Standley, 2013) within QIIME2, using the NCBI reference database for phylogenetic comparisons. This analytical pipeline provided accurate taxonomic resolution and robust characterization of the cecal microbiota.

Microbiome and Statistical Analyses ASV profiles and their relative abundances were used to characterize the composition and dominant taxa of the cecal microbiota. Taxonomic profiles were exported as CSV files, and the relative abundances of the top 10 taxa at each taxonomic level were visualized using stacked bar plots. Alpha diversity was assessed using Pielou’s evenness, Shannon entropy, Margalef, and Simpson indices to evaluate community richness, evenness, and overall diversity. Beta diversity was analyzed using normalized ASV abundances to compare microbial community structures among groups. Principal component analysis (PCA) was applied to identify the top five species contributing most to group differences at the species level. Significant dissimilarities in beta diversity between treatments were tested using the multiple response permutation procedure (MRPP). Differential abundance analyses of the top contributory taxa were conducted in QIIME2 using the metagenomeSeq package after normalization of taxonomic counts. The linear discriminant analysis effect size (LEfSe) method (Segata et al., 2011) was used to identify statistically and biologically significant biomarkers between groups, applying a Kruskal–Wallis test (P ≤ 0.05) and an linear discriminant analysis (LDA) score threshold of 4.0. Pairwise comparisons of taxa at the species level between the SP-challenged group (Group A) and other treatment groups (Groups B-E) were performed using STAMP with Welch’s t-test. Additional significance testing of alpha diversity indices was carried out using Wilcoxon rank-sum test for pairwise comparisons. Differences in SP isolation rates among groups were analyzed using the Chi-squared test and Fisher’s exact test was applied when expected cell counts were insufficient for chi-squared analysis. Bacterial cell counts were log-transformed prior to statistical comparison. The Wilcoxon rank-sum test and Kruskal-Wallis test was then used for two-group and multiple-group comparisons. Statistical significance was defined as P ≤ 0.05.

Results

PD model in RFCs

Following oral challenge with the SPB6, chicks in Groups A_t1–C_t1 developed clinical signs of pullorum disease characterized by white diarrhea. Mortality occurred in groups receiving 1 × 10⁸ CFU of SPB6, either as a single dose (Group B_t1) or for four consecutive days (Group C_t1), with Group C_t1 showing the highest incidence of diarrhea and mortality (Table 3). Cloacal swab analysis revealed that Groups A_t1 and B_t1 had significantly higher SP detection rates than the control group (Group D_t1) on day 1 post-challenge (P1D), whereas no significant differences were found at P3D, P7D, or P14D. In contrast, Group C_t1 maintained significantly higher SP detection rates at P1D, P3D, and P14D compared with Group D_t1 (P < 0.05). A similar trend was observed in bacterial load analysis. Group C_t1 exhibited persistently higher levels of SP shedding across multiple time points and significantly greater bacterial loads than controls (Table 4). SP was also recovered at significantly higher rates from the liver, spleen, and cecum in Group C_t1 than in Group D_t1. Continuous oral inoculation with 1 × 10⁸ CFU of SPB6 for four consecutive days successfully induced clinical PD, characterized by consistent bacterial persistence in the cloaca and systemic organs. These findings confirm the successful establishment of a reproducible PD model, which was subsequently applied in Trial 2.

Table 3.

SP isolation rates from cloacal and tissue samples in Trial 1.

Group	N	Death	Mortality	SP isolation rate
				Cloacal swabs				Tissues
				P1D	P3D	P7D	P14D	P14D
At1	6	0	0%	83.3%a (5/6)	16.7%ab (1/6)	0% (0/6)	50%ab (3/6)	66.7%ab (4/6)
Bt1	6	1	16.7%	83.3%a (5/6)	60%ab (3/5)	20% (1/5)	20%b (1/5)	60%ab (3/5)
Ct1	7	2	28.6%	71.4%a (5/7)	71.4%a (5/7)	66.7% (4/6)	100%a (5/5)	80%a (4/5)
Dt1	6	0	0%	0%b (0/6)	0%b (0/6)	0% (0/6)	0%b (0/6)	0%b (0/6)
Total	25	3	-	-	-	-	-	-

N: sample size; SP: Salmonella Pullorum; P1D, P3D, P7D, and P14D: 1-, 3-, 7-, and 14-days post-challenge; SD: standard deviation.

Tissues represent pooled liver, spleen, and cecum.

A_t1: high-dose challenge group; B_t1: low-dose challenge group; C_t1: consecutive low-dose challenge group; D_t1: unchallenged control group.

Multiple comparisons were performed. Different superscript letters indicate significant differences among groups within the same time point (P < 0.05, Fisher’s exact test). Groups sharing at least one letter are not significantly different.

Table 4.

Mean log₁₀ CFU/g of SP recovered from cloacal and tissue samples in Trial 1.

Group	N	SP counts (Mean±SD)
		Cloacal swabs (Log CFU/ml)				Tissues (Log CFU/g)
		P1D	P3D	P7D	P14D	P14D
At1	6	4.19±3.25ab	0.94±2.29ab	0	1.84±2.86ab	1.86±1.49b
Bt1	6	3.01±2.57a	2.45±2.52ab	0	0.82±1.82b	1.59±1.47b
Ct1	7	4.42±3.02a	3.36±2.37a	2.14±2.37	4.91±0.64a	3.36±1.96a
Dt1	6	0b	0b	0	0b	0b

N: sample size; SP: Salmonella Pullorum; CFU: colony-forming unit; P1D, P3D, P7D, and P14D: 1-, 3-, 7-, and 14-days post-challenge; SD: standard deviation.

Tissues represent pooled liver, spleen, and cecum.

A_t1: high-dose challenge group; B_t1: low-dose challenge group; C_t1: consecutive low-dose challenge group; D_t1: unchallenged control group.

Multiple comparisons were performed. Different superscript letters indicate significant differences among groups within the same time point (P < 0.05, Wilcoxon rank-sum test). Groups sharing at least one letter are not significantly different.

Effects of LS against SP infection

In Trial 2, SP shedding from cloacal swabs was monitored on P4D, P7D, P10D, and P17D. Across all time points, RFCs receiving treatments (Groups B, C, and D) showed lower SP detection rates and bacterial loads than the positive control group (Group A). At P4D, amoxicillin treatment (Group B) showed the greatest reduction in SP detection rate, while LS treatment (Group D) showed the most pronounced reductions at P7D. At P10D and P17D, Group D maintained the lowest SP detection rates among all challenged groups (Table 5). A similar trend was observed for bacterial load reduction. At P4D, the strongest inhibition was observed in Group B, whereas at P7D, P10D, and P17D, bacterial loads were lowest in Group D. In addition, SP detection rates and bacterial loads in tissues (liver, spleen, and cecum) were lower in Groups B, C, and D than in Group A, with Group D again showing the greatest overall reduction (Table 6). A nine-day LS treatment reduced SP isolation rate and bacterial loads in cloacal swabs at P17D to levels comparable to those on uninfected control. In contrast, three-day amoxicillin treatment through drinking water effectively suppressed SP shedding in the short term (P4D), but this effect declined thereafter. From P7D onward, LS treatment demonstrated superior suppression of SP shedding compared with amoxicillin, and this effect persisted until the end of the trial (P17D).

Table 5.

SP isolation rates from cloacal and tissue samples in Trial 2.

Group	N	Death	Mortality	SP isolation rate
				Cloacal swabs					Tissues
				P1D	P4D	P7D	P10D	P17D	P17D
A (SP)	20	3	15.0%	66.7%a (12/18)	70.6%a (12/17)	58.8%a (10/17)	47.1%a (8/17)	41.2%a (7/17)	52.9%a (9/17)
B (SP+Amo)	20	2	10.0%	60.0%a (12/20)	45.0%a (9/20)	50%a (10/20)	35.0%a (7/20)	27.8%a (5/18)	44.4%a (8/18)
C (LS+SP)	20	2	10.0%	63.2%a (12/19)	66.7%a (12/18)	61.1%a (11/18)	44.4%a (8/18)	22.2%a (4/18)	38.9%a (7/18)
D (SP+LS)	20	2	10.0%	70.0%a (14/20)	61.1%a (11/18)	50.0% a (9/18)	27.8%a (5/18)	16.7%ab (3/18)	27.8%a (5/18)
E (CTL)	20	0	0%	0%b	0%b	0%b	0%b	0%b	0%b
Total	100	91	-	-	-	-	-	-	-

N: sample size; SP: Salmonella Pullorum; P1D, P4D, P7D, P10D and P17D: 1-, 4-, 7-, 10-, and 17-days post-challenge; SD: standard deviation.

Tissues represent pooled liver, spleen, and cecum.

A: SP-infected group; B: SP infected and amoxicillin-treated group; C: LS prophylaxis followed by SP infection; D: SP-infected followed by LS treatment; E: uninfected control group.

Multiple comparisons were performed. Different superscript letters indicate significant differences among groups within the same time point (P < 0.05, Fisher’s exact test). Groups sharing at least one letter are not significantly different.

Table 6.

Mean log₁₀ CFU/g of SP recovered from cloacal and tissue samples in Trial 2.

Group	N			SP counts (Mean±SD)
		Cloacal swabs (Log CFU/ml)						Tissues (Log CFU/g)
		P1D	P4D		P7D	P10D	P17D	P17D
A	20	3.15±2.44a	3.24±2.37a		3.00±2.65a	1.81±2.05a	1.39±1.80a	1.45±1.42a
B	20	2.82±2.51a	1.92±2.56a		2.22±2.41a	1.33±1.93a	0.97±1.69a	1.30±1.54a
C	20	2.85±2.43a	2.84±2.24a		2.36±2.14a	1.60±1.90a	0.75±1.49a	1.05±1.38a
D	20	3.29±2.43a	2.63±2.32a		1.81±2.02a	0.96±1.67a	0.63±1.51ab	0.94±1.50 a
E	20	0b	0b		0b	0b	0b	0b

N: sample size; SP: Salmonella Pullorum; SD: standard deviation; CFU: colony-forming unit; P1D, P4D, P7D, P10D and P17D: 1-, 4-, 7-, 10-, and 17-days post-challenge.

Tissues represent pooled liver, spleen, and cecum.

A: SP-infected group; B: SP infected and amoxicillin-treated group; C: LS prophylaxis followed by SP infection; D: SP-infected followed by LS treatment; E: uninfected control group.

Multiple comparisons were performed. Different superscript letters indicate significant differences among groups within the same time point (P < 0.05, Wilcoxon rank-sum test). Groups sharing at least one letter are not significantly different.

Microbial structure and shifts in response to treatments

In healthy RFCs, the predominant cecal genera were Lachnoclostridium (36.96%), Eisenbergiella (30.99%), and Blautia (8.88%) (Fig. 1A). At the species level, Lachnoclostridium edouardi (36.94%) was the most abundant, followed by Eisenbergiella massiliensis (31.00%), Blautia pseudococcoides (7.08%), Blautia hominis (1.48%), Mediterraneibacter glycyrrhizinilyticus (4.74%), and Mediterraneibacter massiliensis (2.02%) (Fig. 1B). In the SP-challenged group (Group A), the dominant species shifted to Eisenbergiella massiliensis (48.54%), Mediterraneibacter glycyrrhizinilyticus (7.99%), and Vescimonas fastidiosa (7.97%). Compared with the control group (Group E), the relative abundances of Vescimonas fastidiosa and Lacrimispora amygdalina were significantly higher (Welch’s t-test, P = 0.0102 and P = 0.0186, respectively), whereas Lachnoclostridium edouardi markedly decreased from 36.94% to 5.68% (P = 0.0315). In the amoxicillin-treated group (Group B), Eisenbergiella massiliensis (49.45%), Blautia pseudococcoides (16.78%), and Blautia hominis (5.25%) were the top three species. Compared with Group A, Fusicatenibacter saccharivorans and Blautia glucerasea were significantly enriched (P = 0.0484 and P = 0.00581, respectively), while Lachnoclostridium edouardi decreased significantly (P = 0.0467). In the LS-treated groups, both prophylactic (Group C) and therapeutic (Group D), the dominant genera were Eisenbergiella massiliensis (39.34% and 39.65%), Blautia pseudococcoides (15.30% and 11.66%), and Vescimonas fastidiosa (8.76% and 9.61%). Compared with Group A, Group C showed significant increases in Fusicatenibacter saccharivorans and Eubacterium coprostanoligenes (P = 0.0464 and P = 0.0191, respectively), accompanied by a significant reduction in Lacrimispora amygdalina (P = 0.0435). In Group D, the relative abundances of Lacrimispora saccharolytica and Faecalicatena contorta were significantly higher than those in Group A (P = 0.0043 and P = 0.0261, respectively).

Fig. 1.

Relative abundance of the top ten dominant taxa in the cecum of red-feathered native chickens (RFCs) under different treatments. Each bar represents the mean relative abundance of each bacterial taxon within a group. (A) Genus level. (B) Species level. Taxa with relative abundances below the top ten are classified as “Other.”.

The relative abundance of LS in the cecal contents was highest in Group C and exceeded the levels observed in Group B (P = 0.0411) and the unchallenged control group (Group E; P = 0.005). No significant differences were detected between Group C and Groups A or Group D (Supplementary Fig. 1).

Microbial diversities in response to treatments

Alpha and beta diversity analyses were conducted to evaluate the effects of different treatments on the composition, richness, and similarity of the cecal microbiota. Alpha diversity results showed that SP challenge groups (A-D) significantly increased phylogenetic diversity and species richness compared with the control group, whereas species evenness and overall diversity were not significantly affected (Fig. 2).

Fig. 2.

Alpha diversity indices of cecal microbiota among treatment groups. (A) Faith’s phylogenetic diversity index; (B) Margalef richness index; (C) Pielou’s evenness index; and (D) Shannon entropy. Data are expressed as mean ± SEM. Multiple comparisons were performed, and different superscript letters indicate statistically significant differences among groups (P < 0.05, Kruskal-Wallis test).

Beta diversity analysis using principal coordinates analysis (PCoA) and partial least squares discriminant analysis (PLS-DA) revealed that SP challenge groups (A-D) significantly altered microbial community composition relative to the control group (Fig. 3A). Neither amoxicillin treatment nor LS interventions markedly changed the overall beta diversity patterns induced by SP infection, as Groups A, B, C, and D displayed similar microbial profiles, as shown by both constrained PCoA and PLS-DA plots (Fig. 3B, C). In contrast, consecutive LS treatment (Group D) exhibited a distinct microbial structure

Fig. 3.

Beta diversity analysis of cecal microbiota composition under different treatments. (A) Principal coordinates analysis (PCoA); (B) Constrained PCoA; and (C) Partial least squares discriminant analysis (PLS-DA).

Furthermore, MRPP analysis confirmed that SP challenge groups (A-D) displayed significantly greater microbial community heterogeneity than the control group (all P < 0.05; Table 7). In addition, Group D showed higher community heterogeneity than Group A (P = 0.05) and Group C (P = 0.042).

Table 7.

Pairwise comparison of the species composition between groups by multiple response permutation procedure (MRPP).

Group	A	Observed-delta	Expected-delta	P value
A - B	0.0424	0.4457	0.4654	0.1220
A - C	0.0429	0.4324	0.4518	0.1070
A - D	0.0612	0.4202	0.4476	0.0500
A - E	0.1937	0.4766	0.5910	0.0080
B - C	0.0536	0.4342	0.4588	0.0890
B - D	0.0672	0.4220	0.4524	0.0640
B - E	0.1953	0.4783	0.5945	0.0050
C - D	0.0671	0.4087	0.4381	0.0420
C - E	0.2141	0.4651	0.5917	0.0020
D - E	0.2359	0.4529	0.5927	0.0030

A: SP-infected group; B: SP infected and amoxicillin-treated group; C: LS prophylaxis followed by SP infection; D: SP-infected followed by LS treatment; E: uninfected control group.

A represents the effect size of within-group homogeneity relative to random expectation.

Observed-delta and expected-delta represent the levels of within-group and between-group differences.

Contributory taxa in cecal microbiota

PCA identified Lachnoclostridium edouardi, Blautia hominis, Blautia pseudococcoides, Vescimonas fastidiosa, and Eisenbergiella massiliensis as the top contributory species shaping the cecal microbiota of RFCs (Fig. 4). Subsequent analysis using the metagenomeSeq package was performed to determine the group distribution of these key taxa. Lachnoclostridium edouardi was predominantly derived from healthy controls (Group E), with significantly higher normalized abundance compared with Groups B, C, and D (P < 0.05; Fig. 5). In contrast, Vescimonas fastidiosa was mainly associated with the SP-challenged groups (Groups A–D) and was significantly more abundant than in the control group (P < 0.05). The relative abundances of Blautia hominis, Blautia pseudococcoides, and Eisenbergiella massiliensis were comparable across all groups, with no significant differences detected (P > 0.05). Notably, Blautia glucerasea and Fusicatenibacter saccharivorans were predominantly enriched in the amoxicillin-treated group (Group B), showing significantly higher abundances than in the other groups. Conversely, Faecalicatena contorta and Lacrimispora saccharolytica were mainly derived from the LS treatment group (Group D) and were significantly more abundant than in the remaining groups.

Fig. 4.

Principal component analysis (PCA) showing the top five microbial taxa contributing to variance in cecal microbiota composition among Groups A–E. The size of each marker represents the degree of contribution to total variance.

Fig. 5.

MetagenomeSeq analysis of contributory and featured species in the cecal microbiota among treatment groups. The x-axis indicates treatment groups, and the y-axis represents normalized species counts. Data are shown as mean ± SD. Multiple comparisons were performed, and different superscript letters indicate statistically significant differences among groups (P < 0.05, Tukey test).

Biomarker taxa identified by LEfSe analysis

LEfSe analysis (P < 0.05, LDA > 4.0) was performed to identify group-specific bacterial biomarkers among the five treatment groups. In the overall comparison, Mediterraneibacter glycyrrhizinilyticus and Lacrimispora amygdalina were identified as biomarkers for the SP-challenged group (Group A). The amoxicillin-treated group (Group B) was characterized by enrichment of Fusicatenibacter saccharivorans and Blautia glucerasea. The LS prophylaxis group (Group C) was defined by Mediterraneibacter ruminococcustorques and Negativibacillus massiliensis, whereas the LS treatment group (Group D) was characterized by Vescimonas fastidiosa, Lacrimispora saccharolytica, Agathobaculum desmolans, Faecalicatena contorta, and Gorbachella massiliensis. The control group (Group E) was distinguished by Lachnoclostridium edouardi (Fig. 6A). The relative abundances of these biomarkers across groups are shown in the heatmap (Fig. 6B). The phylogenetic cladogram further illustrated distinct evolutionary branching patterns in the cecal microbiota associated with different LS applications (prophylactic versus therapeutic). Partial overlap was observed between Groups A and D, represented by Mediterraneibacter glycyrrhizinilyticus (Fig. 6C). Pairwise comparisons were conducted to further identify differentially abundant taxa between specific groups (Fig. 6D–E). In both A vs. B and A vs. D comparisons, Mediterraneibacter glycyrrhizinilyticus was identified as the primary biomarker for Group A. For Group B, Fusicatenibacter saccharivorans and Blautia glucerasea consistently served as biomarkers when compared with Groups A or D. Conversely, in the D vs. B comparison, Faecalicatena contorta, Lacrimispora saccharolytica, and Mediterraneibacter glycyrrhizinilyticus were the key discriminant taxa for Group D.

Fig. 6.

Differentially abundant taxa among groups identified by linear discriminant analysis effect size (LEfSe) with an LDA score threshold of 4. The most discriminative clades are shown across multiple taxonomic levels. (A) Biomarkers identified by comparisons among all five groups; (B) Heat map showing the relative abundances of identified biomarkers across groups; (C) Phylogenetic cladogram illustrating the taxonomic hierarchy from phylum to species; (D) Biomarkers identified by LEfSe comparisons among Groups A, B, and D; (E) Biomarkers identified from the comparison between Groups B and D.

Discussion

Indigenous chickens have been reported to exhibit greater resistance to Salmonella infection than commercial broilers (Ahmad et al., 2023; Dar et al., 2022; Weerasooriya et al., 2017). Native breeds also show reduced susceptibility to SP infection and enhanced bacterial clearance in the spleen compared with commercial chickens (Li et al., 2018). However, in the absence of systematic breeding programs and with the continued introduction of untested breeder stocks into native chicken flocks, PD outbreaks remain frequent. The common practice of relying on antibiotics for prevention and treatment is increasingly unsustainable due to consumer demand for antibiotic-reduced or -free production. To reproduce clinical manifestations such as white diarrhea and mortality, three field isolates (SP1, SPOF, and SPB6) were initially screened for virulence in chicks. Oral inoculation was employed to mimic natural exposure and horizontal transmission (Tie et al., 2018; Yang et al., 2022). Preliminary trials demonstrated that a single oral dose of 8 × 10⁷ CFU of the SPB6 isolate induced clinical symptoms with 9.1% mortality, whereas the other isolates caused no apparent disease or mortality (data not shown). Consequently, SPB6 was selected as the challenge strain for subsequent experiments. The dose optimization trial revealed that a single oral inoculation of 1 × 10⁹ or 1 × 10⁸ CFU of SPB6 resulted in a transient increase in SP shedding, which was significantly higher than controls at P1D but declined by P7D. In the 1 × 10⁹ CFU group, bacterial shedding slightly rebounded by P14D, whereas the 1 × 10⁸ CFU group maintained low detection rates. These observations indicate a short-term peak of SP shedding during the first week post-inoculation, followed by rapid suppression mediated by host immune responses, including serum antibody production that limits bacterial excretion through the gastrointestinal tract (Shen et al., 2022; Sun et al., 2022). In contrast, repeated oral administration of SPB6 at 1 × 10⁸ CFU for four consecutive days established a sustained infection. Cloacal shedding remained consistently high, and significant bacterial colonization was detected in the liver, spleen, and cecum, accompanied by clinical symptoms typical of PD. Similar infection models have been used to reproduce S. Enteritidis infection in broilers through repeated inoculations (Liu et al., 2023). These findings demonstrate that repeated exposure to SP is essential for establishing a reproducible clinical PD model in RFCs. Reducing recurrent transmission events within flocks may therefore represent an effective strategy to mitigate the occurrence and severity of PD in native chicken production systems.

SP can evade host immune defenses and persist within the spleen and reproductive tracts for extended periods (Shen et al., 2022). The infection process begins with intestinal invasion and local shedding, followed by systemic dissemination through the lymphatic system, tissue colonization, and eventual host mortality (Barrow and Freitas Neto, 2011; Chappell et al., 2009; Xian et al., 2020; Xie et al., 2020). To assess the effects of different treatments on SP shedding and persistence, bacterial prevalence and loads in cloacal swabs and internal organs were monitored. Amoxicillin, a broad-spectrum β-lactam antibiotic widely used in poultry production, is known to reduce bacterial shedding and organ colonization (Gharaibeh et al., 2021; Liu et al., 2019). It was selected as a reference treatment because it is regularly used in field flocks to treat PD. In our trial, three-day amoxicillin treatment following label recommendations reduced SP isolation rates and bacterial loads in cloacal swabs on P4D. However, this effect declined after P7D and was not sustained through P17D. Moreover, amoxicillin treatment did not significantly reduce bacterial colonization in the liver, spleen, and cecum. This limited bactericidal persistence may be attributed to the relatively short metabolic duration of amoxicillin, which is typically degraded within six days in chicken tissues (Zhang et al., 2019). Hu, et al. reported the similar results and suggested that amoxicillin is ineffective against Salmonella residing within macrophages during systemic infection (Hu et al., 2021). The amoxicillin formulation used in this study contained 500 mg of amoxicillin trihydrate per gram of powder. The manufacturer recommends administration at 100 g of powder per 1,000 L of drinking water for 2 to 3 consecutive days. In clinical settings, veterinarians may adjust treatment duration or dosage based on disease severity. Different durations or higher therapeutic dosing regimens may result in different efficacy outcomes.

LS is widely recognized as a probiotic with multiple beneficial properties. Previous studies have demonstrated its antioxidant activity (Yang et al., 2023), capacity to modulate immune responses (Zhang et al., 2011), and role in maintaining intestinal microbial balance (Shokryazdan et al., 2017). In the present study, LS was evaluated for both prophylactic and therapeutic applications against SP infection. Administration of LS reduced SP prevalence and bacterial loads in cloacal swabs and internal organs throughout the trial. Notably, a nine-day LS treatment achieved greater control of PD than a single prophylactic dose. Evidence from other Lactobacillus species showed that pre-infection administration of Lactobacillus paracasei and Lactobacillus plantarum enhanced immune responses, reduced Salmonella and Escherichia coli counts, and increased Lactobacillus abundance more effectively than post-infection treatment (Wang et al., 2020). Similarly, preventive supplementation with Lactobacillus casei promoted improved probiotic colonization and a lower incidence of diarrhea compared with therapeutic use (Deng et al., 2021). These discrepancies may reflect differences among Lactobacillus strains or the relatively short duration of LS prophylaxis, which may have not allowed stable intestinal colonization or full inhibitory activity. The prophylactic design in this study was deliberately restricted to a short pre-infection period to reflect field conditions in PD endemic regions. As a result, the duration of LS administration before SP exposure was limited. Recent studies show that sustained LS administration is often necessary to achieve reliable protective or therapeutic effects against Salmonella infection (Castillo et al., 2011; He et al., 2024). These findings suggest that, in PD endemic settings, preventive LS use should be combined with strict biosecurity and temporary isolation practices. Such measures would allow chicks to avoid early SP exposure while receiving LS for a sufficient period before placement into the main flock.

Several mechanisms may explain the observed reduction in SP loads following LS treatment. LS produces antimicrobial compounds, such as bacteriocins and exopolysaccharides, which directly inhibit pathogenic bacteria (Svetoch et al., 2011; Wayah and Philip, 2018; Yildiz et al., 2023). LS also stimulates host immune responses against enteric pathogens (Quilodrán-Vega, et al., 2020; Wang et al., 2020). Consecutive oral administration of LS for several days further enhances Th1-mediated immune responses in splenic and intestinal lymphocytes, thereby strengthening cell-mediated immunity (Smelt et al., 2013; Yun et al., 2015). Studies that extended LS administration to four weeks reported marked reductions in SP loads in the liver, spleen, and cecum (He et al., 2024). These effects support sustained suppression of SP in both cloacal and tissue samples. In addition to direct antimicrobial and immunological effects, LS supplementation improves intestinal microbial diversity and composition, which further promotes host immunity and resistance to infection (Wang, et al., 2021a; Wei et al., 2023; Zhai et al., 2020). Together, these findings support the potential of LS as an alternative to antibiotics for PD control in chickens and indicate that longer administration periods provide progressively greater protective benefits.

The gut microbiota is increasingly recognized as an organ-like system that plays a critical role in host physiology, immunity, and resistance to pathogens (Khan et al., 2020; Rychlik, 2020). Numerous studies have demonstrated that probiotic modulation of the gut microbiota enhances growth performance, strengthens immune responses, and improves disease resistance in poultry (Rubio, 2019). In the present study, SP challenge induced marked alterations in the cecal microbiota, including increased species richness and distinct compositional shifts. These findings are consistent with the observations of Huang et al. (Huang et al., 2022). Taxonomic analysis revealed that SP challenge reduced the abundance of Lachnoclostridium edouardi, a short-chain fatty acid (SCFA)-producing bacterium associated with improved growth efficiency and anti-inflammatory activity (Li et al., 2020; Stanley et al., 2016; Wang et al., 2019), while increasing the relative abundance of Vescimonas fastidiosa. Although Vescimonas fastidiosa has been associated with host metabolism and leanness in mammals (Yang et al., 2021), its role in poultry health remains unclear. LEfSe analysis identified Mediterraneibacter glycyrrhizinilyticus as a key biomarker in SP-challenged RFCs. This species has been linked to Salmonella infection and metabolic disorders (Jan et al., 2023; Zhu et al., 2023), and is not typically abundant in the normal cecal microbiota of healthy chickens.

This study applied MRPP analysis to evaluate differences in cecal microbial community structure among treatment groups. Neither a three-day amoxicillin treatment nor a single prophylactic LS produced marked changes in overall microbial composition or internal diversity. In both groups, the cecal microbiota remained similar to that of the SP-challenged group. A nine-day LS treatment resulted in a marginal difference in community structure when compared with the SP-challenged group, with the statistical value approaching the 0.05 threshold. However, PCoA did not display clear separation between these groups, which suggests that the difference was unlikely to be biologically meaningful. LEfSe analysis identified Blautia glucerasea and Fusicatenibacter saccharivorans as biomarkers in the amoxicillin-treated group. Blautia species are recognized for their anti-inflammatory properties and association with age-related gut stability (Liu et al., 2021), while Fusicatenibacter saccharivorans is an SCFA producer associated with reduced inflammation and improved gut health (He et al., 2021; Jin et al., 2019; Pang et al., 2021). These findings indicate that amoxicillin treatment selectively favored bacterial taxa with potential anti-inflammatory functions, despite limited effects on overall microbial structure. In LS-treated RFCs, Faecalicatena contorta and Lacrimispora saccharolytica were identified as dominant biomarkers. Faecalicatena contorta is an SCFA-producing bacterium that has been associated with lower prevalence of chronic inflammatory diseases in humans (Shan et al., 2021; Zhu, et al., 2021). Lacrimispora saccharolytica exhibits extensive carbohydrate fermentation and polyamine metabolism, contributing to epithelial repair and maintenance of gut barrier integrity (Matsumoto and Benno, 2007; Rao et al., 2007). Previous poultry studies have reported positive associations between Lacrimispora saccharolytica abundance, improved growth performance, and probiotic supplementation (De Cesare, et al., 2020; Luo, et al., 2013). These findings indicated that consecutive LS treatment exerted a beneficial modulatory effect on the cecal microbiota of infected chickens, increasing the relative abundance of beneficial SCFA-producing taxa.

In summary, this study established a reproducible PD model in RFCs through consecutive oral inoculations with the SPB6 isolate. The model reflected clinical symptoms and persistent bacterial colonization observed under field conditions. Using this model, LS treatment after SP exposure reduced SP shedding, bacterial loads, and tissue colonization to levels comparable with amoxicillin treatment. LS treatment also modulated cecal microbiota by enriching beneficial SCFA-producing taxa associated with intestinal health and immune resilience. These findings provide an experimental support for evaluating LS as a potential alternative to antibiotics for the control of pullorum disease.

CRediT authorship contribution statement

Cheng-En Wu: Writing – original draft, Investigation, Formal analysis, Data curation. Sheng-Yao Wang: Writing – review & editing, Validation, Methodology, Formal analysis, Conceptualization. Jr-Wei Chen: Writing – review & editing, Validation, Supervision, Resources, Conceptualization. Wen-Yuan Yang: Writing – review & editing, Writing – original draft, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.