Impact of Prebiotic and Probiotic Diets on Gut Bacterial Microbiota and Susceptibility to <scp><i>Edwardsiella ictaluri</i></scp> Challenge in Oxytetracycline‐Treated Channel Catfish

Antibiotics are often used in response to bacterial disease outbreaks in aquaculture. The nature of aquaculture necessitates a population-based approach to manage disease, which subjects both sick and unaffected fish to antibiotics. Broad-spectrum antibiotics indiscriminately kill or inhibit bacterial growth, including otherwise innocuous commensal or transient microbes. For a sick individual, the benefits of clearing the infection outweigh the potential adverse effects of antibiotic-induced dysbiosis. However, for unaffected individuals, antibiotic treatment can disrupt existing microbial gut communities, which may put the individual at risk for subsequent enteric infection due to reduced competition with commensal microbes (Fishbein, Mahmud, and Dantas 2023). Pre and probiotics are thought to enhance recovery of the gut microbiota following antibiotic disruption (Merrifield et al. 2010; Vargas-Albores et al. 2021). Oxytetracycline is one of three antibiotics approved for use in US catfish aquaculture. This research examines the impact of oxytetracycline on the gut microbiota of channel catfish, along with the effectiveness of prebiotics and probiotics in facilitating recovery of the gut microbiome and enhancing survival during a subsequent Edwardsiella ictaluri challenge. Feed was formulated as an all-plant protein feed similar to that used in commercial catfish practices and prepared as previously described (Older et al. 2024). Feed formulations and proximate composition analyses are provided in Table S1. Oxytetracycline (OTC) was included following FDA recommendations to provide 41 mg/kg of body weight per day. After drying, samples of the feed were ground and analysed for their nutrient composition (AOAC, 2005). Thirty channel catfish juveniles (mean weight: 21.9 g) were stocked in each of 20 aquaria (110-L, 4 treatments, 5 aquaria/treatment) supplied with flow-through groundwater (2 L/min, ~25 C) and supplemental aeration. Fish were fed the reference diet 1 week prior to the feeding trial and received rations of ~3% body weight daily. Positive control (PC), prebiotic and probiotic groups received the oxytetracycline-containing feed for 10 days, while the negative control (NC) group received the reference diet throughout the trial. After the 10-day antibiotic course, fish were fed either the reference (PC), prebiotic (GroBiotic-A; International Ingredient Corporation, St. Louis, MO, USA), or probiotic (Bactocell; Lallemand, Montreal, Canada) diet for the remainder of the study (Figure 1). Digesta sampling occurred on days 0, 4, 8 and 12 after the antibiotic course, the remaining fish were immersion-challenged with Edwardsiella ictaluri S97-773 as previously described (Older et al. 2024). In addition to the aquaria described, an additional five aquaria were set up identically as the NC group for use as a sham-challenged control; no fish from these aquaria were used for digesta sampling. One day prior to each sampling, fish were fed their respective diets to apparent satiation, with feeding staggered at 5-min/tank intervals to ensure consistent digesta transit time between tanks. On sampling days, three fish per tank were euthanised with an overdose of buffered tricaine methanesulphonate (MS-222 at ~300 mg/L). Digesta was pooled by tank in a sterile 15-mL centrifuge tube and held on ice until transfer to −80°C. DNA was extracted from digesta and diet samples (250 mg) using the QIAGEN DNeasy PowerSoil Pro Kit (QIAGEN, Hilden, Germany). Sequencing libraries (16S rRNA V4 region; primers 515F and 806R) were prepared as previously described (Gohl et al. 2016) and sequenced at the University of Minnesota Genomics Center on an Illumina MiSeq. Sequencing data were processed in QIIME2 v2022.8 (Bolyen et al. 2019). Adapters were removed with cutadapt (Martin 2011) and sequences quality filtered (forward and reverse reads trimmed to 240 and 150 nt respectively) and merged with 50 nt overlap using DADA2 (Callahan et al. 2016). Taxonomic classification was performed with a Naïve Bayes classifier trained on the SILVA SSURef NR99 database v138.1 (Quast et al. 2013) and trimmed to the V4 region with RESCRIPt (Robeson et al. 2021). After filtering out ‘Unassigned’, ‘Archaea’, ‘Eukaryota’, ‘Mitochondria’ and ‘Chloroplast’ ASVs, ASVs were inserted into the SILVA reference phylogeny using SEPP (Janssen et al. 2018) prior to exporting relative abundance tables and performing alpha and beta diversity analyses. Digesta samples (n = 80) were rarefied to 22,527 sequences prior to diversity analyses, which led to seven samples being discarded, leaving 73 for diversity analyses. The discarded samples included four samples from day 0 (OXY-PC-0-4, OXY-PC-0-9, OXY-PRE-0-12, OXY-PRE-0-15) and one digesta sample each from the other sampling timepoints (OXY-NC-4-16, OXY-PC-8-14 and OXY-NC-12-16). All statistical analyses, except for differential abundance testing, were performed in R v4.3.0 (R Core Team 2023). Alpha diversity (measured with Pielou's evenness, Faith's phylogenetic diversity, observed features and Shannon entropy) was compared among treatment groups at each timepoint using Kruskal–Wallis tests, followed by post hoc Dunn's tests with p-value adjustment via the Benjamini–Hochberg's method. Beta diversity (calculated with Bray–Curtis, Jaccard, unweighted UniFrac and weighted UniFrac) results were subjected to permutational multivariate analysis of variance (PERMANOVA) tests using the vegan package (Dixon 2003). Testing for differentially abundant taxa was performed using the linear discriminant analysis effect size (LEfSe) algorithm in the one-against-all mode, with LDA score (log 10) > 3 and p < 0.01 considered significant (Segata et al. 2011). Kaplan–Meier survival curves and Cox proportional hazards model were used to analyse the survival data using packages survival (Therneau 2024) and survminer (Kassambara, 2021). There was no oxytetracycline-induced dysbiosis in terms of alpha diversity (Figure 2) or beta diversity. Fish fed the reference diet without oxytetracycline (NC) had significantly higher relative abundance (average relative abundance = 13.75%) of Turicibacter compared to oxytetracycline treatments (4.28%; Figure 3). The family Erysipelotrichaceae (NC = 13.77%, Oxy = 4.31%) and order Erysipelotrichales (NC = 13.77%, Oxy = 4.32%) were also found in significantly higher relative abundances in the NC group. There were no differences in alpha or beta diversity among the four dietary treatments at any timepoint. Only a few taxa of low relative abundance on day 12 had significant differential relative abundance among treatment groups. Pediococcus spp. were found in significantly higher relative abundance in the probiotic group (3.27%) compared to all other groups (< 0.02%), likely due to Pediococcus spp. being the bacteria in the probiotic feed additive. Likewise, Epulopiscium spp. was found in higher abundance (1.3%) in the probiotics group compared to the prebiotic, NC and PC groups (all < 0.2%). Prebiotic samples collected on day 12 had significantly higher relative abundances of three taxa within the same lineage: Aneurinibacillus, Aneurinibacillaceae and Aneurinibacillales. The genus Aneurinibacillus was only present in prebiotic samples, accounting for an average of only 0.027% relative abundance. A lack of oxytetracycline-induced dysbiosis contradicts the generally accepted dogma that antibiotics cause significant gut microbiome disruption. However, these data are consistent with some previous studies. For example, Nile tilapia (Oreochromis niloticus) exhibited similarly low oxytetracycline-induced dysbiosis, with similar diversity and community structure between treated (100 mg OTC/kg body weight/day) and untreated groups (Payne et al. 2021). Likewise, zebrafish (Danio rerio) exposed to a low dose of oxytetracycline (5000 ng/L) for a month showed no changes in diversity, although several taxa were linked to the duration of oxytetracycline exposure (Kayani et al. 2022). In contrast, rainbow trout (Oncorhynchus mykiss) gut microbiota were significantly affected after 8 days of dietary oxytetracycline administration (35 mg OTC/kg body weight/day), which became more pronounced over time (Payne et al. 2022). Atlantic salmon (Salmo salar) exhibited less diverse communities dominated by Aeromonas following oxytetracycline administration (75 mg OTC/kg body weight/day; Navarrete et al. 2008). Although the presented results did not indicate a strong impact of oxytetracycline on catfish intestinal bacterial communities, antibiotic-induced dysbiosis has been reported in multiple studies in catfish fed florfenicol, another antibiotic approved for use in catfish (Abdelhamed et al. 2019; Older et al. 2024; Wang et al. 2019; Zeng et al. 2019). A potential reason florfenicol, but not oxytetracycline, produces significant dysbiosis in channel catfish could be a longer history of oxytetracycline use compared to florfenicol. Oxytetracycline was first approved for human administration in 1950 (EPA 1993) and was approved for use in catfish in 1970 (FDA 1970). Comparably, florfenicol was first approved for catfish in 2005 (FDA 2005). Thus, there has been more time for various bacterial species to develop tolerance to oxytetracycline and/or acquire resistance genes. The likelihood of the fish in the present study having antibiotic-resistant gut microbes is relatively low, given they were obtained as fry, reared indoors for research purposes and never administered antibiotics. Further research into the frequency of antibiotic resistance in hatcheries, broodstock populations and groundwater sources would be valuable to understand potential reservoirs for antibiotic resistance. Moreover, research directly comparing dysbiosis induced by florfenicol and oxytetracycline is warranted to confirm these putative differences in antibiotic-induced effects. There were no differences in survival rates between fish fed the probiotic after oxytetracycline administration (75%) and those fed the control diet after antibiotic administration (PC, 73.3%), neither of which differed from the negative control fish administered the reference diet throughout the study (80%; Figure 4). Interestingly, there appeared to be an additive effect of the antibiotic and the prebiotic, resulting in the lowest survival rate (62.5%), which was significantly lower than that of the NC group (p = 0.02). This is consistent with previous work by Older et al. (2024), which reported that channel catfish fed a prebiotic diet after florfenicol administration also had the lowest survival upon E. ictaluri challenge (Older et al. 2024). No mortalities were observed within the sham-challenged control group. The present study did not identify strong oxytetracycline-induced dysbiosis in catfish receiving a 10-day course of oxytetracycline. Further research is needed to understand why a lack of dysbiosis was observed and to evaluate if the same results occur when fish are cultured in the pond environment, wherein they are exposed to a larger quantity and diversity of microbes. Additional investigations into other microbiome features, for example, functional output, would be valuable in obtaining a broader view of potential effects of oxytetracycline and subsequent prebiotic or probiotic supplementation on gut microbiome disruption and recovery respectively. Caitlin E. Older: methodology, writing – original draft, visualization, writing – review and editing, investigation, formal analysis, data curation, conceptualization, software. Bradley M. Richardson: investigation, software, writing – review and editing, formal analysis. Matt J. Griffin: resources, writing – review and editing, conceptualization. J. Grant Reifers: investigation, writing – review and editing. Penelope M. Goodman: investigation, writing – review and editing. Cynthia Ware: investigation, writing – review and editing. Delbert M. Gatlin: resources, writing – review and editing. David J. Wise: resources, writing – review and editing, conceptualization. Fernando Y. Yamamoto: resources, conceptualization, writing – review and editing, investigation, methodology, visualization. Ingredients for the experimental diets were kindly provided by Delta Western Grains Inc. and Fish Belt Feeds. The authors gratefully acknowledge Mrs. Sandra Philips for her technical assistance throughout the feeding trial and laboratorial analyses. This research used resources provided by the SCINet project and the AI Center of Excellence of the USDA Agricultural Research Service, ARS project number 0500-00093-001-00-D. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer. Animal procedures were carried out in compliance with the Institutional Animal Care and Use Committee at Mississippi State University under the protocols IACUC-22-020 and IACUC-22-068. The authors declare no conflicts of interest. Sequencing data are available from NCBI under BioProject PRJNA1152613. TableS1 Feed formulation and analysed proximate composition of the diets. Values for proximate composition are expressed on a dry-matter basis. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.