Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study investigated the effects of two feline probiotic strains (L. agilis ZY25 and L. salivarius ZY35) on EPEC-induced intestinal damage in a murine model. It includes comprehensive parameters such as barrier integrity (D-LA, DAO, LPS), humoral immunity (IgA, IgG, IgM), cytokine profile, histomorphometers, 16S rRNA microbiota analysis, LEfSe, and correlation analysis.The multilayered analysis (barrier-immune-microbiota axis), the comparison of three different probiotic application strategies, the inclusion of antibiotic-positive controls, and the microbiota-inflammation correlation analysis all demonstrate that this study is robust and of high quality. 

Key Points Requiring Major Revision

Murine Model – Feline Translation Problem
Although the article emphasizes feline-derived probiotics:
Model Motor: C57BL/6 mouse
Target application: Feline gastrointestinal health
This translational gap has not been sufficiently discussed.
The following heading should be added to the Discussion section:
“Limitations of the Mouse Model for Probiotics for Cats”
Also, the differences between feline and mouse microbiota should be discussed with the literature.

The article does not investigate molecular mechanisms such as tight junction protein expression (ZO-1, Occludin, Claudin-1), NF-κB, Nrf2, MAPK, and TLR4 signaling.
Barrier damage markers were measured, but Western blot, RT-qPCR, and immunohistochemistry were not observed.

At a minimum, ZO-1/Occludin immunofluorescence staining and TLR4 or NF-κB activation should be included or explicitly indicated as deficient.

Statistical Methodology Duncan multiple range test was used. Multiple testing correction is unclear except for the microbiota. Power analysis should be presented. Effect size (η² or Cohen’s d) should be added. Tukey test should be preferred after ANOVA.

16S Analysis

The Firmicutes/Bacteroidota ratio should be reported using statistical analysis. Functional prediction analysis should be included.

Continuous Supplementation

The article states that the PRO group is the most effective, however; There is no dose-response relationship, colonization has not been confirmed, there is no strain-specific colonization data, and strain tracking has not been performed using qPCR.

The abstract is too superficial – the mechanistic emphasis should be increased.

The introduction should include more feline-microbiome literature.

Figures 7 and 8 should be higher resolution.

The LDA cutoff value should be specified.

The ethical approval number is present, but the ARRIVE guideline checklist is not included.

The randomization and blinding procedure should be detailed.

Major Revision

The article has publication potential, however: Reinforcement of mechanistics, improvement of statistics, translational discussion, and depth of microbiota analysis are required.

 

 

Author Response

Reviewer 1

Comments and Suggestions for Authors

Point 1: This study investigated the effects of two feline probiotic strains (L. agilis ZY25 and L. salivarius ZY35) on EPEC-induced intestinal damage in a murine model. It includes comprehensive parameters such as barrier integrity (D-LA, DAO, LPS), humoral immunity (IgA, IgG, IgM), cytokine profile, histomorphometers, 16S rRNA microbiota analysis, LEfSe, and correlation analysis. The multilayered analysis (barrier-immune-microbiota axis), the comparison of three different probiotic application strategies, the inclusion of antibiotic-positive controls, and the microbiota-inflammation correlation analysis all demonstrate that this study is robust and of high quality.

We sincerely thank the reviewer for the positive and encouraging evaluation of our work. We are grateful that the reviewer recognized the comprehensiveness of our study design, including the integrated assessment of the barrier-immune-microbiota axis, the comparison of different probiotic intervention strategies, the inclusion of an antibiotic-positive control, and the correlation analysis between microbiota and inflammatory parameters.

We appreciate the reviewer’s recognition that these aspects enhance the robustness and overall quality of the study. Encouraged by these comments, we have carefully revised the manuscript according to the subsequent suggestions to further improve its clarity, rigor, and translational relevance.

Point 2: Murine Model – Feline Translation Problem, Although the article emphasizes feline-derived probiotics: Model Motor: C57BL/6 mouse, Target application: Feline gastrointestinal health, This translational gap has not been sufficiently discussed. The following heading should be added to the Discussion section:“Limitations of the Mouse Model for Probiotics for Cats” Also, the differences between feline and mouse microbiota should be discussed with the literature.

We sincerely thank the reviewer for this important comment. In response, we have added a dedicated subsection in the Discussion. In this revised section, we further discuss the translational gap between the murine EPEC challenge model used in this study and the intended feline application. We also added relevant discussion, supported by the literature, on the differences between feline and mouse gut microbiota and their potential implications for probiotic colonization and host responses. These additions have helped to improve the translational interpretation and scientific rigor of the manuscript.

Although the present study provides in vivo evidence that feline-derived L. agilis ZY25 and L. salivarius ZY35 can alleviate EPEC-induced intestinal injury in C57BL/6 mice, caution is required when translating these findings to cats. The murine EPEC challenge model offers a highly controlled platform for evaluating anti-infective and gut-protective effects, but it does not fully reproduce the intestinal ecosystem, dietary physiology, immune environment, or microbial community structure of the feline host. In particular, cats are obligate carnivores, and their gut microbiota is strongly influenced by age, diet composition, and gastrointestinal disease status, whereas the gut microbiota of laboratory mice is shaped by rodent-specific ecology, standardized chow, coprophagy, housing conditions, and cage effects. These interspecies differences may influence both probiotic colonization dynamics and the magnitude or mechanism of host responses[1]. Experimental studies have shown that native gut bacteria display a competitive “home-site advantage” within their original host background, indicating that microbial fitness and host responsiveness are partly species-dependent. Therefore, although feline-derived strains may show beneficial effects in mice, their ecological performance, persistence, and functional interactions may differ in the feline intestine. Notably, feline-origin probiotic administration has already shown preventive benefit in kittens under field-relevant conditions, supporting the importance of validation in the target species itself. Taken together, the current findings should be interpreted as proof-of-concept evidence supporting the biological potential of these feline-derived probiotics, rather than as definitive evidence of clinical efficacy in cats. Future studies should therefore verify strain colonization, host responses, and microbiota remodeling in feline models or naturally occurring feline gastrointestinal disorders[2].

1. Drut, A.; Mkaouar, H.; Kriaa, A.; Mariaule, V.; Akermi, N.; Méric, T.; Sénécat, O.; Maguin, E.; Hernandez, J.; Rhimi, M. Gut microbiota in cats with inflammatory bowel disease and low-grade intestinal T-cell lymphoma. Front. Microbiol. 2024, 15, 1346639. https://doi.org/10.3389/fmicb.2024.1346639.
2. Sprockett, D.D.; Price, J.D.; Juritsch, A.F.; Schmaltz, R.J.; Real, M.V.F.; Goldman, S.L.; Sheehan, M.; Ramer-Tait, A.E.; Moeller, A.H. Home-site advantage for host species-specific gut microbiota. Sci. Adv. 2023, 9, eadf5499. https://doi.org/10.1126/sciadv.adf5499.

 

Point 3: The article does not investigate molecular mechanisms such as tight junction protein expression (ZO-1, Occludin, Claudin-1), NF-κB, Nrf2, MAPK, and TLR4 signaling. Barrier damage markers were measured, but Western blot, RT-qPCR, and immunohistochemistry were not observed. At a minimum, ZO-1/Occludin immunofluorescence staining and TLR4 or NF-κB activation should be included or explicitly indicated as deficient.

We thank the reviewer for this important and constructive comment. We agree that the present study did not directly assess tight junction proteins or signaling pathways such as ZO-1, Occludin, Claudin-1, NF-κB, Nrf2, MAPK, or TLR4 using Western blot, RT-qPCR, immunohistochemistry, or immunofluorescence. In response, we have revised the Discussion to explicitly acknowledge this as a limitation of the current work. We now clarify that our interpretation of barrier protection and immune modulation is based on indirect but biologically relevant evidence, including serum D-lactate, DAO, and LPS levels, intestinal histopathology, cytokine profiles, immunoglobulin responses, and gut microbiota alterations. We have also added a discussion noting that previous studies have shown that probiotic-mediated barrier protection may involve the regulation of tight junction proteins and inflammatory signaling pathways such as TLR4/NF-κB and MAPK, but these mechanisms were not directly verified in the present study and should be addressed in future work using molecular and histological assays.

However, the present study did not directly assess tight junction proteins or related signaling pathways, such as ZO-1, occludin, Claudin-1, NF-κB, TLR4, Nrf2, or MAPK. Therefore, the proposed mechanisms should be interpreted with caution and remain inferential. Our conclusions are mainly supported by indirect evidence, including reduced serum D-lactate, DAO, and LPS levels, improved intestinal morphology, and modulation of inflammatory and immune indices. Future studies should incorporate molecular and histological analyses to further validate the mechanistic basis of the protective effects observed here.

Point 4: Statistical Methodology Duncan multiple range test was used. Multiple testing correction is unclear except for the microbiota. Power analysis should be presented. Effect size (η² or Cohen’s d) should be added. Tukey test should be preferred after ANOVA.

We thank the reviewer for this important comment. In response, we revised the statistical analysis section to improve rigor and transparency. Specifically, for endpoint measurements, we replaced Duncan’s multiple range test with Tukey’s honestly significant difference (HSD) test as the post hoc procedure following one-way ANOVA. We also clarified the handling of multiple comparisons by explicitly stating that Tukey-adjusted comparisons were used for non-microbiota data, while multiple-testing correction for microbiota-related differential analyses was performed using the Benjamini–Hochberg false discovery rate (FDR) procedure where applicable.

In addition, we added effect size reporting (η²) for one-way ANOVA analyses to improve the interpretability of group differences. Regarding power analysis, a formal a priori power calculation was not performed at the study design stage; therefore, we have now explicitly stated this in the revised manuscript and clarified that sample size was determined based on previous comparable animal studies, practical feasibility, and ethical considerations. We believe these revisions have substantially strengthened the statistical rigor and reporting transparency of the manuscript.

Due to sample availability and completeness, longitudinal outcomes and endpoint serum/histology/organ-weight analyses were performed with n = 6/group, whereas n = 5/group were used for cecal 16S rRNA sequencing; microbiome samples were selected without reference to outcome measurements. The exact n for each analysis is indicated in the corresponding figure legends. All statistical analyses were performed using IBM SPSS Statistics (version 24.0; IBM Corp., Armonk, NY, USA). Data are presented as mean ± standard error of the mean (SEM) unless otherwise stated. Prior to parametric testing, normality and homogeneity of variance were evaluated. For endpoint measurements (e.g., serum indices, histological morphometry, and organ indices), comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) test for post hoc multiple comparisons. For one-way ANOVA models, effect size was estimated as η². In figures and tables, different lowercase letters indicate significant differences among groups (p < 0.05), whereas groups sharing at least one letter are not significantly different. Body weight and fecal scoring index plots were generated using OriginPro 2024 (OriginLab Corp., Northampton, MA, USA), and all other graphs were generated using GraphPad Prism (version 10.1.2; GraphPad Software, San Diego, CA, USA). For 16S rRNA gene sequencing data, downstream statistical analyses and visualizations were performed based on the OTU abundance table, including alpha-diversity and beta-diversity analyses; group differences in community structure were evaluated using permutational multivariate analysis of variance (PERMANOVA), and differential taxa were identified using Kruskal–Wallis tests or LEfSe. Where applicable, p values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) procedure. A formal a priori power analysis was not performed; sample size was determined based on previous comparable animal studies, practical feasibility, and ethical considerations.

Point 5: 16S Analysis, The Firmicutes/Bacteroidota ratio should be reported using statistical analysis. Functional prediction analysis should be included.

Thank you for your thoughtful comments on our microbiota analysis. We fully agree that the Firmicutes/Bacteroidota (F/B) ratio is often used as a classical indicator of gut microbial balance. However, after careful consideration, we chose not to add a statistical analysis of the F/B ratio in the current manuscript. Instead, we present the phylum-level data as original relative abundances in Figure 7E. We believe this approach is more transparent and statistically robust, because the F/B ratio is derived from two compositional variables and may amplify technical variation or generate distributions that are difficult to interpret reliably. Moreover, the biological trend is already clear from the raw data: EPEC infection decreased Firmicutes and increased Bacteroidota, whereas probiotic treatments, particularly PRO and PRO-P, reversed this dysbiosis and restored the microbial composition toward that of the healthy control group. In addition, the downstream genus-level analyses provide more functionally meaningful and statistically supported evidence for these changes, such as the restoration of Lachnospiraceae_NK4A136_group and the suppression of Desulfovibrio, which more directly reflect the microbial shifts associated with probiotic protection.

We also appreciate the reviewer’s suggestion to include functional prediction analysis, such as PICRUSt2, to infer the metagenomic potential of the microbiota. While such tools are useful for hypothesis generation, we deliberately chose not to include predictive functional profiling in this study because of its inherent limitations. Functional prediction based on 16S rRNA sequencing is an indirect inference rather than a direct measurement, and its accuracy depends heavily on the availability and relevance of reference genomes in public databases. Given that our study involves a mouse model and feline-derived probiotic strains, we are concerned that these predictions may introduce uncertainty and speculative interpretations that cannot be fully validated by the current dataset. In particular, predicted pathway changes do not necessarily reflect actual transcriptional activity, protein expression, or metabolite production. Therefore, to avoid overinterpretation, we have chosen to base our conclusions on directly measured parameters, including cytokine levels, histopathology, intestinal barrier markers, and observed taxonomic changes.

Importantly, the main purpose of our microbiome analysis was to characterize taxonomic alterations associated with probiotic intervention and to relate these microbial changes to host physiological outcomes. In this context, we believe that the correlations between specific bacterial taxa and host phenotypes, as shown in the Spearman correlation analysis in Figure 9, provide more direct and biologically relevant support for our conclusions than broad functional predictions. The observed associations between beneficial genera, such as Lachnospiraceae-related taxa, and improvements in barrier integrity and inflammation already offer meaningful mechanistic insight into microbiota-mediated protection. At the same time, we agree that metagenomic and metabolomic approaches would be valuable next steps for clarifying the functional consequences of these microbial shifts. Accordingly, we have now explicitly acknowledged this point in the Discussion section and stated that future studies should integrate metagenomic and metabolomic analyses to further elucidate host-microbe interactions.

Point 6: Continuous Supplementation

The article states that the PRO group is the most effective, however; There is no dose-response relationship, colonization has not been confirmed, there is no strain-specific colonization data, and strain tracking has not been performed using qPCR.

We thank the reviewer for raising these critical methodological points. We fully acknowledge that our study does not include dose-response data, strain-specific quantification (e.g., by qPCR), or direct evidence of probiotic colonization in the murine gut. After careful re-evaluation, we have revised the manuscript to ensure our interpretations are accurate and do not overreach.

1. Regarding "Dose-Response" and Experimental Design

We wish to clarify that our experimental design was not intended to test a dose-response relationship (i.e., different concentrations of probiotics), but rather to compare different timing regimens (pre-treatment, therapeutic, and continuous). The "superior efficacy" of the PRO group refers specifically to the regimen of continuous administration compared to the single-phase regimens (PRO-P and PRO-T). We have revised the text throughout the manuscript to avoid any implication of a dose-response effect where none was measured, and to emphasize that our comparisons are between intervention strategies rather than doses.

2. Regarding Colonization and Strain Tracking

We accept the reviewer's criticism that we lack direct evidence of strain colonization. Our original discussion used terms such as "sustained colonization" or "sustained exposure," which may have inadvertently suggested a mechanistic link we did not directly prove. We have systematically revised the manuscript to replace speculative language regarding colonization with more accurate descriptions of our intervention regimen. Specifically, we have modified key sentences in the Abstract, Discussion, and Conclusions sections, and have added a new paragraph to the Limitations section explicitly acknowledging these methodological constraints.

Point 7: The abstract is too superficial – the mechanistic emphasis should be increased.

The introduction should include more feline-microbiome literature.

Thank you for this valuable comment. We fully agree that the abstract should place greater emphasis on the mechanistic insights provided by our study. In the original version, the abstract focused mainly on phenotypic outcomes, such as weight loss and diarrhea, while the underlying mechanisms were only briefly mentioned. In response, we have substantially revised the abstract to strengthen its mechanistic focus. Specifically, we now state more clearly that the study evaluates the protective effects of the probiotics with particular emphasis on barrier function, immune modulation, and microbial homeostasis. We have also expanded the Results section of the abstract by including representative mechanistic markers, such as reduced serum indicators of intestinal permeability (D-lactate, DAO, and LPS), enhanced systemic immunoglobulins (IgA, IgG, and IgM), a more balanced cytokine profile (increased IL-4 and IL-10; decreased TNF-α, IL-6, IL-1β, IFN-γ, and CRP), as well as enrichment of beneficial taxa such as Lachnospiraceae_NK4A136_group and suppression of the pro-inflammatory genus Desulfovibrio. In addition, the concluding sentence has been revised to explicitly state that the protective effects involve integrated mechanisms of barrier reinforcement, immune rebalancing, and microbial stabilization.

We also appreciate the reviewer’s suggestion to include more feline-microbiome literature in the Introduction. After careful consideration, however, we believe that the current Introduction already provides sufficient feline-specific background to frame the knowledge gap addressed in this study. In paragraph 4, we cite four references directly related to feline gut microbiota and probiotics, including studies on dietary modulation of the gut microbiota in dogs and cats, probiotic effects on feline gut health through fecal microbiota and short-chain fatty acids, a randomized placebo-controlled trial of feline-origin probiotics in shelter kittens, and a recent review on probiotics and feline health. Together, these references establish several key points relevant to our study: the feline gut microbiome shows strong host specificity shaped by obligate carnivory, this microbial ecosystem is sensitive to external perturbations, the available evidence supporting probiotic use in cats remains limited, and systematic mechanistic studies on feline-derived lactobacilli are still lacking. In our view, this level of coverage is sufficient to contextualize the present work and define the specific gap our study aims to address.

Importantly, the Introduction was intentionally structured to emphasize the mechanistic rationale of the study rather than to provide an extensive overview of feline microbiome ecology. The first paragraphs introduce the general importance of gut microbiota and probiotics, the pathogenesis of EPEC infection and the need for alternatives to antibiotics, and the known mechanisms by which probiotics counteract pathogenic Escherichia coli. Within this framework, feline microbiome specificity and the current lack of mechanistic studies on feline-derived probiotics are introduced as the relevant species-specific context, before the final paragraph presents the two feline-derived strains investigated here and the objectives of the study. We believe this structure provides a logical progression from general background to the specific scientific question, while maintaining a clear mechanistic focus. Expanding the feline-microbiome discussion further in the Introduction would, in our view, shift the emphasis away from the central rationale of the manuscript.

At the same time, we recognize the importance of addressing feline-specific considerations in interpreting our findings. Rather than broadening the Introduction, we have chosen to discuss these issues in the Discussion section, particularly in the limitations paragraph, where we address the challenges of translating results from a murine model to cats, the potential host-specific advantages of feline-derived probiotics, and the need for validation in the target species. This placement allows us to acknowledge interspecies differences appropriately while preserving the focus and coherence of the Introduction. Moreover, as highlighted in the cited literature, feline microbiome research remains an emerging field, with relatively few mechanistic studies and a literature base still dominated by descriptive reports, reviews, and limited clinical trials. In this context, we believe our study makes a meaningful contribution by providing an in vivo mechanistic evaluation of feline-derived Ligilactobacillus strains against EPEC, while the current Introduction appropriately reflects both the existing knowledge and the unresolved questions in the field.

Point 8: Figures 7 and 8 should be higher resolution. The LDA cutoff value should be specified. The ethical approval number is present, but the ARRIVE guideline checklist is not included. The randomization and blinding procedure should be detailed.

We thank the reviewer for pointing out that the LDA cutoff value was not specified in the original manuscript. We agree that this parameter should be clearly stated to ensure transparency and reproducibility of the LEfSe analysis. In response, we have now added this information to the Materials and Methods section, under the subsection “16S rRNA Gene Sequencing and Microbial Bioinformatics Analysis.” The following sentence has been included: “For LEfSe (Linear Discriminant Analysis Effect Size) analysis, the threshold on the logarithmic LDA score was set to 2.0 to identify significant discriminatory taxa among groups.” This cutoff value is commonly used in microbiome studies and allows the identification of taxa that show not only statistical significance (p < 0.05, Kruskal-Wallis test) but also a meaningful biological effect size.

We also agree with the reviewer that a more detailed description of the randomization and blinding procedures is essential for evaluating the rigor of the study and the potential risk of bias. Although the original manuscript stated that the mice were “randomly assigned,” the specific procedures were not described in sufficient detail. In response, we have expanded the Animals and Experimental Design subsection to provide a clearer account of how randomization was performed and how blinding was implemented during the experiment and data analysis. We believe that these revisions improve the methodological transparency of the manuscript and allow readers to better assess the reliability and internal validity of the study.

Point 9:  Major Revision

The article has publication potential, however: Reinforcement of mechanistics, improvement of statistics, translational discussion, and depth of microbiota analysis are required.

In response to the major revision request, we have carefully addressed all comments raised by the reviewers and substantially revised the manuscript accordingly. Guided by the editorial suggestions, our revisions focused on four major aspects: strengthening mechanistic interpretation, improving statistical reporting and methodological transparency, enhancing translational relevance, and deepening the microbiota analysis. Overall, these revisions have significantly improved the clarity, rigor, and scientific impact of the study. Below, we summarize the principal changes made in the revised manuscript.

First, we reinforced the mechanistic framework of the study at multiple levels. The Abstract has been substantially revised to highlight the specific mechanistic indicators supporting our conclusions, including D-lactate, DAO, LPS, cytokines, immunoglobulins, and key microbial taxa. It now explicitly states that the protective effects of the probiotics are mediated through “integrated mechanisms involving barrier reinforcement, immune rebalancing, and microbial stabilization.” In the Discussion, we refined the wording throughout to avoid overinterpretation of mechanisms that were not directly tested, such as definitive colonization, while strengthening the discussion of biologically plausible pathways supported by our data and the literature, including barrier protection, regulation of inflammatory responses, and microbiota-associated metabolic functions. We also expanded the translational discussion by clarifying important study limitations, such as the absence of dose-response evaluation, strain-specific quantification, and direct colonization evidence. In addition, we revised the interpretation of the superior efficacy observed in the continuous supplementation group, attributing it more cautiously to the administration regimen itself rather than to proven colonization, and we outlined future directions involving molecular tracking approaches and feline validation models. To further improve the contextual relevance of the work, we also added a synthesizing statement integrating the current feline microbiome literature to better position the study within the existing field.

Second, we improved the statistical and methodological transparency of the manuscript and further strengthened the microbiota presentation. The LDA score threshold used for LEfSe analysis has now been explicitly specified in the Materials and Methods section as ≥2.0 to ensure reproducibility. We also expanded the description of the experimental design by detailing the randomization procedure, which was based on a computer-generated random number table, and the blinding procedure, under which investigators remained unaware of group allocation until completion of data analysis. In addition, a statement confirming compliance with the ARRIVE guidelines has been added, and the completed ARRIVE Essential 10 checklist is now included as Supplementary File S1. Regarding the microbiota data, all relevant figures, especially Figures 7 and 8, have been replaced with high-resolution versions to improve readability and visualization of taxonomic features. We further emphasized the biological significance of the Spearman correlation analysis in both the Results and Discussion sections, as these correlations provide strong links between specific microbial taxa and host physiological parameters. Finally, although we did not add a separate statistical analysis of the Firmicutes/Bacteroidota ratio, we clarified our rationale in both the rebuttal and the revised text, emphasizing that the phylum-level shifts are already clearly shown in Figure 7E and that the genus-level analyses offer more informative mechanistic resolution. We believe that, taken together, these revisions have substantially strengthened the manuscript in terms of mechanistic depth, statistical rigor, translational relevance, and microbiological insight. All modifications have been clearly marked in bold in the revised manuscript for ease of review, and a detailed point-by-point response to each reviewer comment has been provided separately. We are sincerely grateful for the opportunity to improve our work and hope that the revised version now meets the standards for publication in your esteemed journal.

 

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study evaluated the protective effects of Ligilactobacillus (L.) agilis ZY25 and L. salivarius ZY35, isolated from healthy cats, against EPEC-induced intestinal injury in C57BL/6 mice. Forty-eight mice were assigned to six groups. The results showed that the EPEC challenge resulted in negative effects and microbial dysbiosis. Probiotic administration alleviated these alterations, improving weight gain and epithelial integrity, reducing serum barrier injury markers, enhancing immunoglobulin and anti-inflammatory cytokine levels, and decreasing pro-inflammatory mediators, and restored diversity, enrichment of beneficial genera, and suppression of potentially harmful taxa. The authors concluded that feline-derived probiotics mitigate EPEC-induced intestinal dysfunction by strengthening barrier function, modulating immune responses, and rebalancing gut microbiota, supporting their potential application in feline gastrointestinal health management.

Abstract: Breifley describe the treatments, when was the EPEC challenge done? briefly, describe the experimental design, provide P values for significant terms.

L51 and everywhere else, don't start the sentense with an abbreviation

L90: avoid "we" and "our"

L94-95: Provide a better objectives of the study.

L118: Provide replicate number per treatment. what was the experimental design?

More details about the addminestration of the probiotic is needed.

L130: provide more informationan about the housing and environmental conditions.

Did you check the animals E. coli level before starting the experiment?

L232 and everywhere else in the result section: Provide P value. The numbers are presented in the figures, provide percentages of increase or decrease compared to the control.

Fig 1b and all figures: provide P value and letters for mean separation, provide sample size as (n=   ).

L237, Fig 1c: a fecal score should be analyzed using non-linear method or chi square 

Some figures could be improved for calrity, you may increase font size.

Discussion and conclusion: well-written

Provide the limitation and the implication of the study

 

 

Author Response

Reviewer 2

Comments and Suggestions for Authors

1. Abstract: Breifley describe the treatments, when was the EPEC challenge done? briefly, describe the experimental design, provide P values for significant terms.

We thank the reviewer for this suggestion. The Abstract has been revised to include a brief description of the experimental design, the timing of EPEC challenge, and P values for key outcomes.

Revised text:

"In this 21-day experiment, 48 mice were randomly assigned to six groups (n=8/group): control, EPEC model, chlortetracycline treatment, probiotic treatment (post-infection), probiotic pre-treatment (pre-infection only), and continuous probiotic supplementation (pre- and post-infection). EPEC challenge was performed daily during experimental days 8-14. EPEC challenge resulted in significant weight loss, increased diarrhea incidence, elevated serum D-lactate, diamine oxidase, and lipopolysaccharide levels, impaired intestinal morphology, immune imbalance, and microbial dysbiosis. Probiotic administration alleviated these alterations, as evidenced by restored intestinal morphology, reduced serum markers of barrier permeability, enhanced systemic immunoglobulins, a balanced cytokine profile (increased IL-4, IL-10; decreased TNF-α, IL-6, IL-1β, IFN-γ, CRP; all p < 0.05 vs. MOD), and modulation of the gut microbiota."

2. L51 and everywhere else, don't start the sentense with an abbreviation

We have carefully reviewed the entire manuscript and revised all instances where sentences began with abbreviations. For example, L51 has been changed from "EPEC is a major cause..." to "Enteropathogenic Escherichia coli (EPEC) is a major cause..." All similar cases have been corrected throughout the text.

3. L90: avoid "we" and "our"

We have removed all instances of "we" and "our" throughout the manuscript, replacing them with passive constructions or "the present study" as appropriate. For example, L90 has been changed from "Building on the our previous isolation..." to "Building on previous isolation...".

4. L94-95: Provide a better objectives of the study.

We have expanded the study objectives to be more specific and comprehensive.

5. L118: Provide replicate number per treatment. what was the experimental design?

We have added this information to the Materials and Methods section. The revised text now clearly states that the study employed a completely randomized design with six treatment groups (n = 8 mice per group), and specifies the technical replicates used for histological measurements, serum assays, and 16S rRNA sequencing.

6. More details about the addminestration of the probiotic is needed.

We have added detailed information about probiotic preparation and administration, including that the suspension was freshly prepared each day, gavage was performed between 9:00–11:00 AM to minimize circadian variations, and for groups receiving both probiotic and EPEC on the same day, probiotics were administered 2 h prior to EPEC gavage to allow for potential competitive exclusion effects.

7. L130: provide more informationan about the housing and environmental conditions.

We have added comprehensive details about housing conditions, including cage type (individually ventilated cages), density (4 mice per cage), dimensions, temperature (22 ± 2°C), humidity (50 ± 10%), light/dark cycle (12 h), bedding type, diet, water, and cage enrichment. This ensures full transparency and reproducibility.

8. Did you check the animals E. coli level before starting the experiment?

We acknowledge that baseline E. coli levels were not specifically quantified. However, all mice were SPF-grade and purchased from a single vendor, and a 7-day acclimatization period was employed to minimize inter-individual microbial variability. We have added this information to the manuscript to provide full transparency regarding this methodological consideration.

9. L232 and everywhere else in the result section: Provide P value. The numbers are presented in the figures, provide percentages of increase or decrease compared to the control.

We have revised the entire Results section to include both P values and percentage changes relative to the control group where appropriate. Each key finding is now accompanied by both statistical significance and meaningful percentage changes, making the results more interpretable for readers.

10. Fig 1b and all figures: provide P value and letters for mean separation, provide sample size as (n=   ).

We have updated all figure legends to include sample size (n= ) and clear indication of statistical significance using lowercase letters to indicate significant differences among groups. This standardization ensures that all figures are self-explanatory and meet journal formatting requirements.

11. L237, Fig 1c: a fecal score should be analyzed using non-linear method or chi square 

We agree with the reviewer that ordinal data such as fecal scores require appropriate non-parametric analysis. In the original submission, we had already used the Mann-Whitney U test for pairwise comparisons between each treatment group and the MOD group at individual time points. This is clearly stated in the Statistical Analyses section and has now been explicitly noted in the Figure 1C legend. We believe this approach is appropriate for the ordinal nature of the data.

12. Some figures could be improved for calrity, you may increase font size.

We have increased font sizes for all taxonomic labels in Figures 7 and 8 to improve readability. These improvements will enhance the clarity of the figures for all readers.

13. Discussion and conclusion: well-written

We thank the reviewer for this positive feedback. We have carefully reviewed these sections and made only minor adjustments for consistency with the revisions elsewhere. We appreciate your recognition of the quality of our discussion and conclusions.

14. Provide the limitation and the implication of the study

Limitations are comprehensively discussed in the final paragraphs of the Discussion section, including: the use of a murine model rather than feline subjects, the lack of strain-specific quantification, the absence of dose-response data, and the need for validation in feline models. To further strengthen the implications, we have added a sentence to the Conclusions highlighting the practical importance of sustained probiotic administration for clinical applications and the foundation this provides for developing evidence-based probiotic protocols for feline gastrointestinal health.

 

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Major Revision
The article presents promising findings but requires significant clarification and additional methodological explanations before publication.

The authors reported that probiotics were obtained from cat feces, but the experiment on live organisms was performed on mice.

The authors need to clearly explain:

Why were mice chosen instead of cats?
Whether host-specific probiotic effects can be generalized to mice, and the limitations regarding host-microbiome specificity should be written more clearly.

Statistics
Initial group size: n = 8  
Analysis size: n = 6
Microbiota analysis: n = 5
No disease, death, etc. were reported in the animals during the experiment.

No criteria have been established for selecting samples for sequencing.

The article does not provide a detailed characterization of probiotic strains. Antimicrobial resistance profile, functional properties of probiotics (acid tolerance, bile tolerance), and adhesion capacity are crucial aspects for verifying the safety and functionality of probiotics.

Tight junction proteins (ZO-1, occludin, claudin-1), mucin genes (MUC2), and epithelial integrity markers are incomplete without this data, making mechanistic interpretation incomplete.

The reasons for using OTU analysis. Whether ASV analysis increases taxonomic resolution. The authors should explain the above points.

The experimental timeline (Figure 1) is somewhat complex. A simplified experimental scheme is presented.

The three phases need to be explained more clearly: prevention, infection, and treatment.
The discussion of cytokines (IL-4, IL-10, TNF-α, IL-6, IFN-γ) can be improved by relating them to: Th1/Th2 immune balance and mechanisms of intestinal inflammation.

More recent references (2023–2025) should be added.
Additional literature on host-adapted probiotics in pets should be provided.

Minor grammatical corrections are suggested.

Examples: inconsistent hyphens, repeated statements in results, and minor formatting inconsistencies.

Explain the rationale for the mouse model for feline probiotics.

Provide a detailed characterization of the probiotic strains.

Include additional barrier function markers.

Clarify sample selection and statistical power.

Expand the discussion on improving microbiome analysis methodology.

 

Author Response

Reviewer 1

Comments and Suggestions for Authors

We would like to express our sincere gratitude for your careful evaluation of our manuscript and for the constructive comments you provided. We greatly appreciate the time and effort you invested in reviewing our work. Your insightful suggestions have helped us identify important shortcomings in the current manuscript and have provided valuable guidance for improving its scientific quality, methodological clarity, and overall presentation. In order to respond more clearly and systematically, we have organized your comments into the following major issues:

Point 1: The authors reported that probiotics were obtained from cat feces, but the experiment on live organisms was performed on mice.

The authors need to clearly explain:

Why were mice chosen instead of cats?

 Whether host-specific probiotic effects can be generalized to mice, and the limitations regarding host-microbiome specificity should be written more clearly.

Response: We thank the reviewer for this important and insightful comment. We fully agree that the rationale for using mice instead of cats, as well as the limitations regarding host specificity, should be more clearly explained.

In the present study, mice were selected as a controlled in vivo proof-of-concept model, rather than as a direct surrogate for cats. Compared with feline experiments, the murine EPEC challenge model allows more standardized pathogen exposure, more uniform genetic and environmental backgrounds, and more controlled evaluation of intestinal injury, inflammatory responses, and microbiota alterations. This makes it suitable for an initial mechanistic assessment of whether the feline-derived strains possess biological activity against EPEC-induced intestinal dysfunction under controlled experimental conditions.

At the same time, we agree that host-specific probiotic effects cannot be fully generalized from mice to cats. Cats are obligate carnivores with distinct gastrointestinal physiology, dietary patterns, immune environment, and gut microbial ecology, all of which may influence probiotic colonization, persistence, and functional outcomes. Recent evidence also supports the concept of a host-specific “home-site advantage” of native gut microbiota, indicating that microbial fitness and host responsiveness may differ across species. Therefore, we have revised the manuscript to clarify that the present findings should be interpreted as proof-of-concept evidence supporting the biological potential of these feline-derived strains, rather than as definitive evidence of clinical efficacy in cats. We have also strengthened the limitation statement and emphasized that validation in feline models or clinical feline studies is still required.

These revisions have been made in the Introduction, Materials and Methods, Discussion, and Conclusions. Please see P3. L98-106; P3. L129-133; P19-20. L708-742

Using a murine EPEC challenge model with chlortetracycline as a positive control, we sought to establish a controlled first-step in vivo proof-of-concept system to evaluate the biological efficacy of these feline-derived strains and to compare different intervention regimens under standardized experimental conditions. Although this model does not replicate the full physiological and microbial context of the feline intestine, it provides a tractable mammalian platform for assessing barrier-related injury, immune responses, and microbiota alterations before target-species validation. These data are therefore intended to support further development and subsequent evaluation of feline-oriented probiotic formulations, rather than to serve as definitive evidence of efficacy in cats.

Mice were used as a controlled mammalian model for initial in vivo evaluation because the EPEC challenge system is well established and allows standardized infection conditions, reduced inter-individual variability, and systematic assessment of intestinal barrier injury, inflammatory responses, and microbial changes prior to validation in the target species.

Although the present study provides in vivo evidence that feline-derived L. agilis ZY25 and L. salivarius ZY35 can alleviate EPEC-induced intestinal injury in C57BL/6 mice, caution is required when translating these findings to cats. In this study, the murine EPEC challenge model was used as a controlled proof-of-concept platform for evaluating anti-infective and gut-protective effects under standardized experimental conditions. However, this model does not fully reproduce the intestinal ecosystem, dietary physiology, immune environment, or microbial community structure of the feline host. In particular, cats are obligate carnivores, and their gut microbiota is strongly influenced by age, diet composition, and gastrointestinal disease status, whereas the gut microbiota of laboratory mice is shaped by rodent-specific ecology, standardized chow, coprophagy, housing conditions, and cage effects. These interspecies differences may influence probiotic colonization dynamics, ecological fitness, and the magnitude or mechanism of host responses [51]. Experimental studies have further shown that native gut bacteria may exhibit a competitive “home-site advantage” in their original host background, indicating that microbial adaptation and host responsiveness are partly species dependent. Therefore, the present findings should be interpreted as proof-of-concept evidence of biological activity in a controlled mammalian system, rather than as definitive evidence of clinical efficacy in cats. Notably, feline-origin probiotic administration has already shown preventive benefits in kittens under field-relevant conditions, further underscoring the importance of validation in the target species itself [52].

Furthermore, while the continuous supplementation regimen (PRO) showed the greatest efficacy among the three intervention strategies tested in this study, several methodological limitations should be acknowledged. Our experimental design was not intended to establish a dose-response relationship, nor did it include strain-specific quantification (e.g., qPCR) to confirm the active colonization or persistence of L. agilis ZY25 and L. salivarius ZY35 in the murine gut. Therefore, the superior protective effects observed in the PRO group should be attributed to the continuous administration regimen itself, namely the longer and uninterrupted exposure of the host to the probiotics, rather than to proven long-term colonization or mucosal association. The precise mechanisms by which sustained intake confers cumulative benefits, whether through transient metabolic effects, immune modulation, or ecological interactions, remain to be elucidated. Future studies should combine molecular tracking approaches with metagenomic and metabolomic analyses to verify strain colonization dynamics, optimize dosing regimens, and assess host responses and microbiota remodeling in feline models or naturally occurring feline gastrointestinal disorders.

Point 2: Statistics, Initial group size: n = 8, Analysis size: n = 6, Microbiota analysis: n = 5
No disease, death, etc. were reported in the animals during the experiment. No criteria have been established for selecting samples for sequencing.

Response: We thank the reviewer for this important comment. We agree that the description of sample size and sample selection was not sufficiently transparent in the original manuscript.

In the present study, the initial group size was n = 8 per group, and no animals died or were excluded because of disease, treatment failure, or other clinical events during the experiment. However, this was not clearly stated in the original manuscript, and we appreciate the opportunity to clarify it. As this study was conceived as a staged proof-of-concept investigation, we also prioritized preserving sufficient resources for subsequent validation in the target species, which requires substantially more complex and costly recruitment and follow-up.

This work was designed as an initial proof-of-concept murine study prior to target-species validation. Because terminal biochemical/histological assays and 16S rRNA sequencing required additional resources, a randomly selected subset of animals was used for some downstream analyses. Specifically, body weight and fecal score were analyzed in all animals (n = 8/group), whereas terminal serum/histology/organ-weight analyses were performed in n = 6/group / longitudinal and terminal analyses were performed in n = 6/group, and cecal microbiota sequencing was performed in n = 5/group. Importantly, the sequencing samples were selected randomly and without reference to outcome measurements.

We also acknowledge that the original manuscript did not explicitly state whether a priori inclusion/exclusion criteria had been defined. In the revised manuscript, we now clarify that no animals were excluded due to mortality or clinical events, and that the exact n for each analysis is now stated more explicitly in the Methods and figure legends. We also revised the wording to avoid any implication of outcome-based sample selection.

These revisions were made in the Materials and Methods (Sections 2.2, 2.7, and 2.8) and in the relevant figure legends. Please see P3-4. L136-141; Please see P5-6. L227-232; Please see 6. L262-271

After acclimatization, mice were randomly assigned to six groups (n = 8 per group) using a computer-generated random number table (Random Allocation Software, version 1.0). All animals completed the study, and no mice were excluded because of death, disease, or treatment-related complications during the experiment. Body weight and fecal score were monitored in all animals throughout the study. For terminal analyses, randomly selected subsets were used for some downstream assays, as detailed below.

Genomic DNA was extracted from cecal content samples collected from five randomly selected mice per group (n = 5 per group). Sample selection for 16S rRNA gene sequencing was performed without reference to clinical, biochemical, histological, or microbiota outcome data. DNA extraction was conducted using the MagPure Soil DNA LQ Kit (Shanghai Magen Biotechnology Co., Ltd., Shanghai, China) according to the manufacturer’s instructions.

The initial group size was n = 8 per group, and all animals completed the study without death or clinical exclusion. Body weight and fecal score were analyzed in all animals (n = 8/group). Because this study was designed as an initial proof-of-concept experiment and terminal biochemical/histological assays were constrained by assay resources, serum, organ index, and histological analyses were performed in a randomly selected subset of six mice per group (n = 6/group). Cecal microbiota analysis by 16S rRNA sequencing was performed in five randomly selected mice per group (n = 5/group) because of sequencing-resource limitations. Sample selection for downstream analyses was performed without reference to outcome measurements. The exact n for each analysis is indicated in the corresponding figure legends.

Point 3: The article does not provide a detailed characterization of probiotic strains. Antimicrobial resistance profile, functional properties of probiotics (acid tolerance, bile tolerance), and adhesion capacity are crucial aspects for verifying the safety and functionality of probiotics.

Response: We thank the reviewer for this important comment. We agree that the original manuscript did not provide sufficiently clear information regarding the characterization of the probiotic strains used in this study.

The two strains used here, L. agilis ZY25 and L. salivarius ZY35, had been systematically characterized in our previously published study [21], including evaluation of their safety- and probiotic-related properties, such as antimicrobial susceptibility/safety assessment, acid tolerance, bile tolerance, and adhesion-related characteristics, as well as their anti-EPEC activity. In the original manuscript, this information was only briefly referred to, which may have caused confusion. In the revised manuscript, we have now clarified that these strain-characterization assays had already been performed and published previously, and we have cited the corresponding article more explicitly in the Introduction and Materials and Methods sections. Because these data have already been reported in detail elsewhere, we did not repeat the numerical results in the current manuscript. Please see P3. L119-125.

21. Wang, W.; Dong, H.; Chen, Q.; Chang, X.; Wang, L.; Miao, C.; Chen, S.; Chen, L.; Wang, R.; Ge, S.; Xiong, W. Antibacterial Efficacy of Feline-Derived Lactic Acid Bacteria against Enteropathogenic Escherichia coli: A Comprehensive In Vitro Analysis. Fermentation 2024, 10, 514. https://doi.org/10.3390/fermentation10100514.

Table 2. Antimicrobial activity against EPEC of representative LAB isolates.

Isolates	Antimicrobial Activity	Isolates	Antimicrobial Activity
ZY1	++++	ZY26	+++
ZY2	++++	ZY27	+++
ZY3	++++	ZY28	+++
ZY4	++++	ZY29	++++
ZY5	+++	ZY30	+++
ZY6	+++	ZY31	+++
ZY7	++++	ZY32	++++
ZY8	+++	ZY33	++++
ZY9	+++	ZY34	+++
ZY10	++++	ZY35	++++
ZY11	+++	ZY36	+++
ZY12	++++	ZY37	++++
ZY13	++++	ZY38	+++
ZY14	++++	ZY39	++++
ZY15	+++	ZY40	+++
ZY16	++++	ZY41	+++
ZY17	++++	ZY42	++++
ZY18	+++	ZY43	+++
ZY19	++++	ZY44	++++
ZY20	+++	ZY45	+++
ZY21	++++	ZY46	++++
ZY22	+++	ZY47	++++
ZY23	++++	ZY48	+++
ZY24	++++	ZY49	++++
ZY25	++++	ZY50	+++

Notes: +++, diameter of the inhibition zone: 18.00–22.00 mm; ++++, more than 22.00 mm; the diameter of the inhibition zone included that of the hole puncher (10.00 mm).

Figure 2. Cell surface hydrophobicity and auto-aggregation ability of LAB isolates: (A) hydrophobicity of selected LAB isolates; (B) auto-aggregation ability of selected LAB isolates. Different lowercase letters denote significant difference (p < 0.05).

Figure 4. Survival of selected LAB strains in the simulated gastrointestinal fluids. Viable count (log CFU/mL) of the selected LAB strains after simulated gastrointestinal tract (GIT) conditions. Gastric juice T0: viability at the beginning of the gastric juice treatment; gastric juice T1: viability after the simulation of gastric conditions; intestinal juice T2: viability at the beginning of the gastric juice treatment; intestinal juice T3: viability after the simulation of enteric conditions. Different lowercase letters on the same row denote significant differences (p < 0.05) during the assay.

Figure 5. Hemolytic activity of selected LAB isolates: (A) positive control—Staphylococcus aureus ATCC 29213^T, (B) ZY33, (C) ZY25, and (D) ZY35.

Table 4. Antibiotic susceptibility of isolates ZY25 and ZY35.

Isolate	GEN	CIP	CTR	E	AMP	TET	SXT	C	MY	PEN
ZY25	R	R	I	S	S	I	R	S	I	S
ZY35	I	R	I	S	S	I	S	S	R	S

Notes: S—susceptible; I—intermediate resistant; R—resistant. The concentrations of antibiotics are expressed in micrograms per disc (μg/disc): genmalicin (GEN, 10), ciprofloxacin (CIP, 5), ceftriaxone (CTR, 30), erythromycin (E, 15), ampicillin (AMP, 10), telracyclin (TET, 30), compound sulfamethoxa (SXT, 25), chloramphenicol (C, 30), lincomycin (MY, 2), and penicillin (PEN, 10).

Table 6. Antibacterial spectra of ZY25 and ZY35.

Isolates	Indicator Bacteria
	P. aeruginosa	S. aureus	L. monocytogenes	E. coli	B. subtilis	S. dysenteriae
ZY25	−	+++	++++	+++	+++	++++
ZY35	+++	+++	+++	+++	+++	+++

Notes: 1. −, no inhibition; +++, diameter of the inhibition zone: 18.00–22.00 mm; ++++, more than 22.00 mm; the diameter of the inhibition zone included that of the hole puncher (10.00 mm). 2. P. aeruginosa: Pseudomonas aeruginosa CICC 23694^T, S. aureus: Staphylococcus aureus ATCC 29213^T, L. monocytogenes: Listeria monocytogenes CICC 23929^T, E. coli: Escherichia coli CICC 24189^T, B. subtilis: Bacillus subtilis CICC 10275^T, S. dysenteriae: Shigella dysenteriae CICC 23829^T.

The probiotic preparation consisted of two feline-derived strains, L. agilis ZY25 and L. salivarius ZY35, isolated from feces of healthy cats. These strains had been previously characterized in our published study [21] for anti-EPEC activity as well as key safety- and probiotic-related properties, including acid tolerance, bile tolerance, antimicrobial susceptibility/safety assessment, and adhesion-related characteristics. Based on those prior evaluations, they were selected for the present in vivo study.

Point 4: Tight junction proteins (ZO-1, occludin, claudin-1), mucin genes (MUC2), and epithelial integrity markers are incomplete without this data, making mechanistic interpretation incomplete.

Response: We thank the reviewer for this important and constructive comment. We agree that the present study did not directly assess tight junction proteins (e.g., ZO-1, occludin, claudin-1), mucin-related markers such as MUC2, or other molecular indices of epithelial integrity, and therefore the mechanistic interpretation in the original manuscript was stronger than warranted by the available data.

   In the current study, our conclusions regarding barrier protection were based primarily on indirect but commonly used barrier-related readouts, including serum D-lactate, diamine oxidase (DAO), lipopolysaccharide (LPS), and intestinal histomorphology. We agree, however, that these measurements do not substitute for direct molecular evidence of tight junction restoration or mucin regulation. Accordingly, in the revised manuscript, we have carefully softened the mechanistic language throughout the title, abstract, Introduction, Discussion, and Conclusions. The manuscript now presents these findings as evidence of improved barrier-related injury indices and histological recovery, rather than direct proof of tight junction or mucin regulation.

We have also revised the Discussion to state explicitly that the proposed barrier-related mechanisms remain inferential and require future validation using direct epithelial markers such as ZO-1, occludin, claudin-1, and MUC2. We believe that this revision provides a more cautious and scientifically appropriate interpretation of the present results. Please see P1. L2-4; Please see P1. L38-41; Please see P3. L93-97; Please see P8. L322; Please see P17. L606; Please see P20. L754-763.

Feline-Derived Ligilactobacillus agilis ZY25 and Ligilactobacillus salivarius ZY35 Alleviate Enteropathogenic Escherichia coli–Induced Intestinal Injury and Microbial Dysbiosis in Mice

These findings suggest that feline-derived probiotics mitigate EPEC-induced intestinal dysfunction, accompanied by improved barrier-related indices, immune rebalancing, and microbial stabilization, thereby providing proof-of-concept evidence for their further evaluation in feline gastrointestinal health.

The primary objectives were: (1) to determine whether the timing and duration of probiotic administration (preventive, therapeutic, or continuous) influence efficacy against EPEC-induced intestinal injury; and (2) to assess barrier-related injury, systemic immune responses, intestinal histopathology, and cecal microbiota alterations in a murine EPEC challenge model.

3.2. Probiotics Improve Serum Biomarkers Associated with Barrier Disruption

Administration of L. agilis ZY25 and L. salivarius ZY35 effectively alleviated EPEC-induced intestinal dysfunction, as reflected by reduced serum D-lactate, DAO, and LPS levels, together with improved intestinal histomorphology and attenuation of inflammatory responses. These findings support a protective effect on barrier-related injury and endotoxin translocation in EPEC-challenged mice. However, the present study did not directly assess tight junction proteins (e.g., ZO-1, occludin, or claudin-1), mucin-related markers such as MUC2, or other molecular indices of epithelial integrity. Rather than demonstrating direct restoration of tight junction architecture, the current data indicate that probiotic supplementation was associated with improved permeability-related serum biomarkers and histological recovery. Previous studies have reported that L. salivarius strains can modulate epithelial tight junctions and inflammatory signaling, but such mechanisms were not directly verified in the present work [33-36]. Future studies should incorporate direct epithelial and mucosal markers, including ZO-1, occludin, claudin-1, and MUC2, to clarify the molecular basis of the barrier-related protection observed here.

Collectively, these findings indicate that ZY25 and ZY35 alleviate EPEC-induced intestinal injury in mice and are associated with improved barrier-related biomarkers, immune balance, and microbial homeostasis. Among the three intervention strategies tested, continuous supplementation showed the most consistent protective effects. However, direct molecular evidence for tight junction or mucin regulation was not obtained in the present study, and the underlying epithelial mechanisms therefore remain to be further clarified. In addition, because this work was conducted in a murine challenge model, further validation in cats is required before efficacy in feline gastrointestinal health management can be concluded.

Point 5: The reasons for using OTU analysis. Whether ASV analysis increases taxonomic resolution. The authors should explain the above points.

Response: We thank the reviewer for this important methodological comment. We agree that the rationale for using OTU-based analysis and the implications of ASV-based analysis should be more clearly explained.

In the present study, the 16S rRNA sequencing data were processed using an established UPARSE-based 97% OTU clustering workflow, which was the conventional pipeline available through our sequencing/analysis platform at the time of analysis. We chose this workflow because it is a validated and widely used approach for amplicon-based community profiling and is suitable for community-level comparisons such as alpha-diversity, beta-diversity, and dominant taxonomic shifts.

At the same time, we agree that ASV-based denoising methods (e.g., DADA2/deblur) generally provide higher sequence resolution, improved reproducibility, and better cross-study comparability than traditional OTU clustering. Therefore, the use of OTU analysis in the present study represents a more conventional but lower-resolution strategy. We have now revised the manuscript to explicitly acknowledge this point and to state that future reanalysis of the raw reads using an ASV-based pipeline, or validation by shotgun metagenomics, may further improve taxonomic resolution and functional interpretation.

Because the main objective of the present microbiota analysis was to evaluate overall community shifts associated with EPEC challenge and probiotic intervention, we believe that the OTU-based workflow remains acceptable for the current dataset; however, we fully acknowledge that ASV analysis would strengthen methodological rigor and resolution. These points have now been added to the Materials and Methods and Discussion sections. Please see P6. L254-258; P19. L681-684

Because the sequencing dataset in this study was originally processed using an established UPARSE-based 97% OTU workflow, all downstream community analyses were performed at the OTU level. Although ASV-based methods generally provide higher taxonomic resolution, the OTU-based approach used here was considered appropriate for overall community-level comparison in this proof-of-concept study.

It should also be noted that the present microbiota analysis was based on OTU clustering rather than an ASV-based denoising pipeline; therefore, the current findings should be interpreted as community-level evidence, and future ASV-based reanalysis may further improve taxonomic resolution.

Point 6: The experimental timeline (Figure 1) is somewhat complex. A simplified experimental scheme is presented. The three phases need to be explained more clearly: prevention, infection, and treatment.

Response: We thank the reviewer for this helpful suggestion. We agree that the original experimental timeline in Figure 1A was visually complex and that the three phases of the intervention (prevention, infection, and treatment) were not presented as clearly as they could be.

In the revised manuscript, we have therefore removed the original schematic Figure 1A and replaced it with a new Table 1, which presents the experimental design in a more concise and readable format. The table now clearly summarizes the three phases of the study—prevention (days 1–7), infection/challenge (days 8–14), and treatment/recovery (days 15–21)—as well as the corresponding intervention schedule for each experimental group. We believe that this tabular presentation makes the design easier to follow and more transparent for readers.

We also revised the corresponding description in the Materials and Methods (Section 2.2) to explain the three phases more explicitly. In addition, the figure legend for Figure 1 was updated accordingly so that Figure 1 now focuses only on body weight and fecal score outcomes. Please see P3. L134-136; Please see P4. L154-165; Please see P4. L180-181; Please see P7-8. L315-321

The study lasted 28 days, including a 7-day acclimatization period followed by a 21-day experimental period (experimental days 1–21; see Table 1 for the intervention schedule).

Mice were gavaged with 0.2 mL per administration according to a three-phase schedule, which is summarized in Table 1. Phase I (days 1–7) was the prevention/pre-feeding phase, during which probiotic pre-exposure was established in the relevant groups. Phase II (days 8–14) was the infection/challenge phase, during which mice in the challenged groups received EPEC daily. Phase III (days 15–21) was the treatment/recovery phase, during which post-challenge interventions were continued according to group assignment. The probiotic suspension was freshly prepared each day by mixing the two strains at a 1:1 ratio in sterile PBS to achieve a final concentration of 2 × 10⁹ CFU/mL. Gavage was performed using a sterile 20-gauge curved feeding needle between 9:00 and 11:00 AM each day to minimize circadian variations. For groups receiving both probiotic and EPEC on the same day (PRO-P and PRO during the challenge phase), probiotics were administered 2 h prior to EPEC gavage to allow for potential competitive exclusion effects.

Table 1. Three-phase experimental design and intervention schedule for the murine EPEC challenge study.

Group	Day1-7	Day8-14	Day15-21
CON	PBS	PBS	PBS
MOD	PBS	PBS+1×109 CFU EPEC	PBS
CTC	PBS	PBS+1×109 CFU EPEC	PBS+2 mg Chlorotetracycline
PRO-T	PBS	PBS+1×109 CFU EPEC	PBS+1×109 CFU probiotics
PRO-P	PBS+1×109 CFU probiotics	PBS+1×109 CFU EPEC+1×109 CFU probiotics	PBS
PRO	PBS+1×109 CFU probiotics	PBS+1×109 CFU EPEC+1×109 CFU probiotics	PBS+1×109 CFU probiotics

Figure 1. Effects of probiotic intervention on body weight and fecal score in EPEC-challenged mice. (A) Body weight changes (g) during the experimental period, (B) Time course of the fecal scoring index reflecting diarrhea severity (scoring criteria are described in the Methods). Data are presented as mean ± SEM (n = 8). CON, control; MOD, model; CTC, chlortetracycline; PRO-T, probiotic treatment; PRO-P, probiotic pre-treatment; PRO, probiotic prevention plus treatment; EPEC, enteropathogenic Escherichia coli.

Point 7: The discussion of cytokines (IL-4, IL-10, TNF-α, IL-6, IFN-γ) can be improved by relating them to: Th1/Th2 immune balance and mechanisms of intestinal inflammation. More recent references (2023–2025) should be added. Additional literature on host-adapted probiotics in pets should be provided.

Response: We thank the reviewer for this valuable suggestion. We agree that the original Discussion did not sufficiently interpret the cytokine findings within the framework of Th1/Th2 immune balance and intestinal inflammation. In the revised manuscript, we have substantially rewritten the cytokine-related Discussion to clarify that the EPEC-induced decrease in IL-4 and IL-10, together with the increase in TNF-α, IL-6, and IFN-γ, is consistent with a shift toward a more pro-inflammatory Th1-skewed immune milieu, whereas probiotic supplementation partially restored a more regulated immune balance. We also linked these cytokine changes more explicitly to intestinal inflammatory injury and epithelial dysfunction. In addition, we added more recent references (2023–2025) to strengthen the mechanistic discussion and incorporated additional literature on host-adapted probiotics in companion animals, particularly feline-origin and dog-origin probiotic studies, to better support the translational relevance of the present work. Please see P17-18. L620-.647

Beyond local intestinal protection, L. agilis and L. salivarius supplementation also influenced systemic immune responses, as reflected by increased IL-4, IL-10, and immunoglobulin levels and decreased TNF-α, IL-6, IFN-γ, and CRP. From an immunological perspective, this pattern is consistent with partial correction of an EPEC-induced pro-inflammatory shift toward a more Th1-dominant milieu. In intestinal inflammation, TNF-α and IFN-γ are closely associated with Th1-driven immune activation and can aggravate epithelial injury, macrophage activation, and barrier dysfunction, whereas IL-4 and IL-10 are more closely linked to Th2/regulatory responses that help counterbalance excessive Th1-associated inflammation. IL-6 further amplifies inflammatory cascades and contributes to sustained mucosal immune activation [37–39]. Accordingly, the cytokine profile observed in the MOD group suggests a disturbed inflammatory equilibrium, while probiotic administration—particularly in the continuous supplementation group—shifted this profile toward a more regulated immune state rather than simply suppressing inflammation non-specifically.

This interpretation is also relevant in the context of host-adapted probiotics for companion animals. Recent feline research has emphasized that many probiotic candidates used in cats are not cat-derived and may therefore show reduced host adaptation, whereas feline-origin strains are increasingly considered more relevant for target-species gastrointestinal applications [40]. Consistent with this concept, a randomized placebo-controlled trial demonstrated that a feline-origin Enterococcus hirae probiotic reduced diarrhea risk in shelter kittens, supporting the translational value of host-adapted probiotic strategies in cats [19]. In addition, recent feline studies and companion-animal reviews continue to highlight the growing interest in probiotic applications in cats and dogs [18,20,41]. Collectively, these data strengthen the interpretation that ZY25 and ZY35 may alleviate EPEC-associated intestinal inflammation not only through antimicrobial and microbiota-mediated effects, but also through modulation of cytokine networks linked to Th1/Th2 immune balance. Future studies in cats should further verify whether these feline-derived strains produce comparable immunoregulatory effects under target-species conditions.

37. Meyer, F.; Wendling, D.; Demougeot, C.; Prati, C.; Verhoeven, F. Cytokines and intestinal epithelial permeability: A systematic review. Autoimmun. Rev. 2023, 22, 103331. https://doi.org/10.1016/j.autrev.2023.103331
38. Baum, S.; Hamedi, K.; Loftus, C.; Loftus, G.; Zhou, E.-R.; Arce, S. From Cytokines to Biomarkers: Mapping the Immunopathology of Inflammatory Bowel Disease. Cells 2025, 14, 1589. https://doi.org/10.3390/cells14201589
39. Martínez-García, M.; Hernández-Lemus, E. Pro-Inflammatory and Anti-Inflammatory Interleukins in Periodontitis: Molecular Roles, Immune Crosstalk, and Therapeutic Perspectives. Int. J. Mol. Sci. 2025, 26, 10094. https://doi.org/10.3390/ijms262010094
40. Wang, J.; Zhou, Z.; Lv, C.; Wang, T.; Han, T.; Zhang, C.; Wang, H.; Hu, J. Isolation and in vitro investigation on lactic acid bacteria for potential probiotic properties from cat feces. Front. Vet. Sci. 2024, 11, 1495745. https://doi.org/10.3389/fvets.2024.1495745
41. Sun, J.; Gu, X.; Zhang, H.; Zhao, L.; Wang, J.; Wang, X.; Tao, H.; Wang, Z.; Han, B. Application of Probiotics in Cats and Dogs: Benefits and Mechanisms. Vet. Sci. 2025, 12, 1008. https://doi.org/10.3390/vetsci12101008

Point 8: Minor grammatical corrections are suggested. Examples: inconsistent hyphens, repeated statements in results, and minor formatting inconsistencies.

Response: We thank the reviewer for this careful and helpful comment. We agree that the original manuscript contained several minor language and formatting inconsistencies, including inconsistent hyphenation, occasional repetitive wording in the Results section, and minor typographical or formatting issues.

In the revised manuscript, we have carefully proofread the entire text and performed a comprehensive language and formatting revision. Specifically, we standardized hyphenation and compound modifiers throughout the manuscript (e.g., feline-derived, host-adapted, barrier-related, EPEC-induced, pre-treatment, proof-of-concept), corrected minor typographical errors and punctuation inconsistencies, unified terminology and formatting in figure legends, and reduced repetitive phrasing in the Results section where appropriate. We also corrected several small formatting issues, such as inconsistent figure references, capitalization, and spacing.

These revisions were implemented throughout the manuscript, including the Introduction, Materials and Methods, Results, Discussion, figure legends, and reference formatting, to improve the overall clarity, consistency, and readability of the text. All changes have been highlighted in red in the revised manuscript for the reviewer’s and editor’s convenience. Please refer to the full revised manuscript.

Author Response File: Author Response.pdf