Animals housing and management
Seventy-two 1-day-old piglets (Danbred (Duroc x Danish Landrace x Yorkshire)) from 21 sows in parities 2-5 (1-5 piglets from each sow with equal gender distribution) were purchased from a commercial sow herd (Holbæk, Denmark). Upon arrival at the animal facility, the piglets were initially housed together in groups of three until they had learned to drink milk from a trough. Thereafter, all piglets were housed individually in cages (90 cm x 74 cm). Cages were cleaned once per day and equipped with enrichment material, heating lamps, and ad libitum access to water. Troughs and water supply were cleaned twice per day to ensure hygiene. The room temperature was kept constant at 26°C from day two until the end of the experiment (day 29). All piglets received a subcutaneous injection of iron dextran complex (Uniferon, 1 ml/pig, Unitron a/s, Kolding, Denmark) for prevention of anemia and a single oral treatment with toltrazuril (Baycoxine Vet. 0.4 ml/kg, Elanco ApS, Ballerup, Denmark) for prevention of coccidiosis on days three and four of life, respectively.
Experimental diet and feeding
During the initial 20 days of the experiment (i.e. 21 days of age), all piglets received a milk replacer diet consisting of a mixture of bovine milk enriched with whey protein (Bulk Powder Performance, Bulk, Essex, United Kingdom and WPI/WPC 90, Arla Foods Ingredients P/S, Viby, DK) and whey permeate (Variolac 836, Arla Foods Ingredients P/S, Viby, DK (Table S1). Daily, all piglets received 180 ml per metabolic body weight (kg0.75) from day one to three, 210 ml from day four to seven, 222 ml from day eight to 14, and 240 ml from day 15 to 19. Milk bolus feeding was provided every second hour, i.e. 12 meals per day, using an automatic milk feeding system (Big Dutchman, Vejen, Denmark). From day 20 to 21 the piglets were weaned and received 100 g of solid creep feed (Table S2), and from day 21 onwards, they were given ad libitum access to solid feed.
Experimental design
The animals were stratified according to sex and body weight and randomly allocated to one of four groups receiving oral administrations of colonic microbiota transplantation (CMT, n =18), colonic content filtrate transplantation (CcFT, n =18), gastric microbiota transplantation (GMT n = 18) or saline (CON, n = 18). Participants of this study were blinded to the treatment groups. The study was conducted over two experimental rounds with nine piglets per treatment group in each round.
Inoculum preparation and administration of transplants
Colon luminal and gastric contents were collected from fifteen 11-day-old healthy suckling piglets from the same farm as the recipient piglets. All donor piglets were healthy with no signs of diarrhea and raised without use of antibiotics or pharmaceutical levels of zinc oxide in the diet. After euthanization, the colon and gastric luminal content respectively was pooled across the animals, homogenized and diluted 1:1 with 20% sterile glycerol, and stored at -80°C until use. Before inoculation, colon and gastric content were further diluted 1:3 in sterile saline to a working concentration of 0.17 g content/ml (CMT and GMT). Furthermore, half of the CMT was centrifuged at 5000xg, 4°C for 30 min., and the supernatant was filtered through a 0.45 µm syringe filter (Filtropur S, PES, Sarstedt, Germany) after which it was ready to administer (CcFT). The inoculum was administered via a feeding tube into the stomach on the day of arrival (day two) and day three. The CMT, CcFT, and GMT groups received 6 ml of working solution per treatment (corresponding to 1 gram of original material diluted in 3.3% glycerol) and the CON group received equivalent volumes of sterile saline with 3.3 % glycerol. The donor material was screened for the following pathogens: ETEC, rotavirus, Lawsonia, Salmonella, and B. pilosicoli (Kjellerup laboratory, Landbrug & Fødevarer F.m.b.A. SEGES Laboratory for Swine diseases) and only donors free of these pathogens were used in the final inoculum.
ETEC challenge
On day 24, all piglets received an oral single dose of ET10 (O149:H10, F4ac, STb, LT) gelatine capsule containing 105 CFU/capsule. The procedure of the inoculum has been previously described in Rydal et al., [29]. The capsules were kept at -20°C until use and the CFU was determined by cultivation after thawing.
Clinical and performance endpoints
All piglets were weighed daily throughout the study (Bjerringbro vægte, model no. APM-60, Bjerringbro, Denmark). The feed conversion ratio (FCR) was measured as the weight of feed intake over a period divided by the weight gained during the period. From day one to 23, clinical and fecal scores were assessed visually once daily. From day 24 to 29 clinical score was assessed three times per day. The clinical status was scored according to a clinical system (1 = normal, 2 = mild symptoms, 3 = moderate symptoms, 4 = severe symptoms). The fecal score was recorded as either normal or diarrhoeic and daily prevalence in each group was calculated as (total cases of diarrhea on a specific day/piglet group size)*100. Leftovers of both milk and solid feed were collected and weighed twice per day during the experiment. On days 2, 7, 14, 19, and 23-28, stool samples were collected from each piglet for later qPCR analysis.
Sample collection
All piglets were anesthetized on day 29 with an injection of a mix of zolazepam (25 g/ml, Virbac, Kolding, Denmark), tiletamine (25 g/ml, Virbac, Kolding, Denmark), ketamine (100 g/ml, MSD Animal Health, Copenhagen, Denmark), xylazine (20 mg/ml, ScanVet Animal Health A/S, Fredensborg, Denmark), and butorphanol (10 mg/ml, Biovet ApS, Fredensborg, Denmark). When full anesthesia was achieved, blood samples were drawn by cardiac puncture into heparinized vacutainers. Afterward, the piglets were euthanized with an intra-cardiac injection of sodium-pentobarbital (400 mg/ml, ScanVet Animal Health A/S, Fredensborg, Denmark). The spleen, liver, and kidney were harvested and weighed. The stomach and colon were weighed before and after it was emptied and rinsed with tap water. Furthermore, the small intestine (SI) was measured in length and weighed full and empty. Tissue from jejunum was taken for brush border enzyme activities, morphology, and histology. Colon luminal content was collected for 16s rRNA amplicon sequencing analysis and ETEC qPCR analysis, as described below.
F4-ETEC quantification
ETEC concentration was estimated by quantitative PCR (qPCR) on 1:10 dilution of stool samples using primers F4-F CACTGGCAATTGCTGCATCT and F4-R ACCACCGATATCGACCGAAC [30] amplifying the faeG gene (F4ac). Real-time PCR assay was performed on the LightCycler 96 System (Roche Life Science, Copenhagen, Denmark) in 20 μl reactions with FastStart Essential DNA Green Master mix (Roche Life Science, Copenhagen, Denmark) with the additions of each primer at a concentration of 0.5 μM. The cycling conditions were as follows: 2 min at 95 °C; 40 cycles of 15 s at 95 C and 60 s at 60 °C; and a melt curve step from 60 to 95 °C. A qPCR standard curve was created with tenfold dilutions of ETEC strain ET10 ranging from 107 to 101 CFU/ml. The DNA from both diluted stool samples and ETEC culture of calibration curve was extracted with the boiling method. The number of target copies in each sample was then calculated using the equation: copy number = [10^(−1/S)]^(I−Ct), where S is the slope of the log-linear part of the standard curve, I the intercept of the standard curve, and Ct is the cycle threshold of the sample. Copies were normalized for gram of feces and the limit of detection was set to 36 Ct, equal to ~100 F4ac copies/reaction. Daily ETEC prevalence within each group was calculated as (total cases of >100 F4ac on a specific day/piglet group size)*100.
Health indices measures in blood
Serum cytokine and chemokine concentrations (pg/ml) were measured using a ProcartaPlex Porcine kit (Affymetrix, eBIOscience, Vienna, Austria). Calibration curves from recombinant cytokine and chemokine standards were prepared for the 8-point standard dilution set with fourfold dilution steps in sterile PBS. The samples were measured using a Bio-Plex MagPix Multiplex Reader (Bio-Rad Laboratories Inc. by the Luminex Corporation, The Netherlands). The Bio-Plex Manager software's five-parameter logistic curve fitting (5PL) method was used for raw data analysis and calculation of cytokine concentrations. Using the manufacturer’s protocol of commercial ELISA kits, we quantitatively measured the levels of pig haptoglobin (Abcam, ab205091, Cambridge, United Kingdom), pig C-reactive proteins (CRP; Abcam, ab205089, Cambridge, United Kingdom), and pig major acute phase protein (MAP, ACUVET, Acuvet Biotech, Zaragoza, Spain) in blood serum.
Clinical biochemistry and hematology were measured in EDTA-stabilized (BD-Plymuth, PL6, 7BP, UK) blood. Upon centrifugation, plasma was isolated from EDTA-stabilized blood, and biochemical profiles were determined using an Advia 1800 chemistry system (Siemens Healthcare Diagnostics, Tarrytown, NY, USA).
Histomorphology of the intestine and measurement of enzyme activity formalin-fixed tissue samples from distal jejunum were dehydrated in ethanol and embedded in paraffin, and stained with hematoxylin and eosin. Morphometric analysis of villus height (µm), crypt depth (µm), enterocyte height (µm), number of infiltrating epithelium lymphocytes per 100 enterocytes (IEL/100E), number of goblet cells per 100 enterocytes (goblet cells/100E) were done with Zen Blue 3.0 software at ALAB Weterynaria (Warsaw, Poland). Histopathological lesions (infiltration of the stromal mucosa, mucosal epithelium, brush border, intestinal blunting, cell detritus rich in eosinophils, eosin, edema of stromal mucosa, vessels dilation in stromal mucosa of villi, infiltration of submucosa, edema of submucosa, hyperplasia of enterocytes, cell detritus in the villi surface, number of mitoses in intestinal crypts, hyperemia, vacuolization of neurons) were assigned to a 5-point scale (0 = no pathological changes, 1 = minimal, 2 = mild, 3 = moderate and 4 = marked). The gut-associated lymphoid tissue evaluation included the number of lymphoid follicles visible in intestinal sections.
As a marker of gut mucosal function, we measured the activity of aminopeptidase N, aminopeptidase A, dipeptidyl IV, maltase, sucrase, and lactase in distal jejunum SI, using the assay as described in Sangild et al., [31].
Intestinal microbiota
Total DNA from recipient colon content samples and two replicates of each inoculum was extracted using the QIAamp UCP Pathogen MiniKit (QIAGEN, Copenhagen, Denmark), according to the manufacturer’s instructions, with the addition of a bead-beating step using the Pathogen Lysis tube S (QIAGEN, Copenhagen, Denmark). A blank extraction control was included in the DNA extraction protocol. Partial 16S rRNA gene sequences were amplified using the Quick-16S NGS Library Prep Kit (Zymo Research, CA USA), which targets the V3-V4 region of the 16S rRNA gene, as previously described [32]. Negative and positive control (ZymoBIOMICS DNase/RNase Free Water and ZymoBIOMICS Microbial Community DNA Standard, respectively) were included in library preparation. Sequencing was performed on an Illumina MiSeq platform (2 × 300 bp paired-end reads) using the MiSeq Reagent Kit v3 (600 cycles; Illumina), according to manufacturer’s instructions. The 16S rRNA sequencing data have been submitted to the NCBI Sequence Read Archive (SRA) under BioProject PRJNA981444.
16S rRNA sequencing data were processed using DADA2 v1.14.1 [33] as implemented in R v4.2.1. Optimal filtering and trimming parameters were identified using FIGARO v3.0 [34]. A taxonomy table was assembled by assigning taxonomy to each amplicon sequence variant (ASV) using the Silva taxonomic database v.138.1 for DADA2 [35]. Potential contaminants were identified using control samples and removed using decontam v.1.12.0 [36]. Sequences matching mitochondria or chloroplast were also removed, along with any sequences not assigned to Bacteria. A phyloseq object was constructed from the ASV and taxonomy tables using phyloseq v1.30.0 [37] for subsequent analysis.
To enumerate culturable aerobic bacteria, 100 μL of 10-fold dilutions of each inoculum were spotted in triplicates on blood agar plates. Colonies were counted after incubaction at 37 °C for 24 h under aerobic conditions.
Data calculations and statistical analysis
Data analysis was performed using software R (version 2022.02.1 + 461, R Foundation for Statistical Computing, Vienna, Austria), and illustrations were done in GraphPad Prism (Version 9.3.1 (471), GraphPad Software, La Jolla CA, USA). Repeated measurements over time for continuous variables (growth, feed intake, quantitative ETEC counts) were analyzed using linear mixed-effect models. The prevalence of diarrhea and ETEC on each day were analyzed with pairwise logistic regressions. Average daily gain (ADG), FCR, the first incidence of diarrhea, health indices (cytokines and chemokines, APP, haptoglobin, CRP, and MAP)), clinical biochemistry, hematology, relative organ weights, morphology, histopathology, and enzyme activity were analyzed using linear models. All models included the following fixed effects: treatment, sex, experimental round, and either birth weight (diarrhea and ETEC prevalence, FCR) or sacrificed weight (biochemistry, hematology, morphology, health indices, relative organ weight, and enzyme activity). Validation of the linear models was done by testing the normality and homoscedasticity of the residuals and fitted values. If data did not meet the assumptions, data were log-transformed or transformed by reciprocal to meet the criteria. Validated linear models were analyzed with an ANOVA on the treatment level followed by a Tukey post-hoc test. The ordinal histopathology data was analyzed using the non-parametric Kruskal-Wallis and Dunn’s post hoc test analysis with Benjamin-Hochberg correction. Data are presented as means and standard deviations, except histological parameters which are presented as median and interquartile range (IQR). P-values below 0.10 were regarded as a tendency and p-values below 0.05 were regarded as statistically significant.
For 16S rRNA sequencing data, alpha-diversity (Shannon and Chao1) and beta-diversity (Bray–Curtis dissimilarity metric) indexes were calculated using R package vegan after rarefication with a depth of 90% of the minimum sample depth in the dataset. Multiple comparison of alpha-diversity indexes was performed using the Wilcoxon Rank Sum test and p-values were corrected for multiple comparisons using Holm’s correction. Beta-diversity was visualized using a Principal Coordinates Analysis (PCoA) plot, and differences in beta-diversity were estimated by permutational multivariate analysis of variance (PERMANOVA) using the Adonis function. Differential abundance analysis between treatments and control was performed using DESeq2 and contrasts were corrected for multiple comparisons using the Benjamini-Hochberg’s correction. Only ASVs with adjusted p-values < 0.05 and estimated fold change > 5 were considered significantly differentially abundant and were visualized in a heatmap from R package complex-heatmap. Code for data analysis can be accessed in the Github repository https://github.com/mpirolo/AVANT-WP1-FMT-trial.