1. Introduction
Improving the sustainability of pig livestock is a priority of global policies, in line with the agroecology principles, that imposed a re-thinking of the farming system, including the reduction in chemical dependency and the development of antimicrobial alternatives [
1,
2,
3,
4].
The most critical phase of pig production is weaning, characterized by gastrointestinal disorders, such as post-weaning diarrhea (PWD), mainly due to
Escherichia coli O138, O139, and O141 serogroups, which causes important economic losses (mortality, morbidity and stunted growth) [
5,
6]. Two important bacterial virulence factors are F18 adhesive fimbriae and verocytoxin (VT2e), often found in association in the β-haemolytic O138
E. coli strain [
7,
8]. The fimbriae are responsible for the attachment to porcine intestinal villi, while the toxin destroys endothelial cells of small vessels, resulting in blood clots, hemorrhage, ischemic necrosis, and oedema in vital organs, including the brain [
9,
10,
11,
12,
13].
After the European ban of antibiotics [
14], traditionally used to control PWD, an increased use of Zinc oxide (ZnO) was observed to control enteric disorders, with consequent negative impacts on the environment and enhanced spread of antibiotic-resistant strains [
15,
16,
17]. Therefore, the development of effective alternative tools is needed to reduce disease incidence and guarantee the performance, health and welfare of pigs. A number of innovative replacements have been thoroughly discussed, such as edible vaccines, immunotherapeutics, bacteriophages, antimicrobial peptides, gut microbiota modulators, plant extracts, and integrated approaches [
18,
19,
20,
21].
PWD is a multifactorial disease in which the interaction of host-infectious agent, environment, management strategies, age and genetics and immunologic factors play critical roles in the outcome [
22,
23]. Therefore, a more comprehensive understanding of these aspects is needed, and is achievable through the experimental reproduction of the disease in controlled and repeatable conditions. Although knowledge has advanced substantially in recent decades, much more needs to be learned about the pathogenesis and the control of PWD outbreaks, which are increasingly spreading worldwide.
The aim of the study was to set up experimental conditions to simulate the simultaneous outbreak of PWD and enterotoxaemia in weaned piglets, through verocytotoxic O138 Escherichia coli (F18 positive) challenge and apply a multidisciplinary investigative approach. In particular, the animals were evaluated for nutritional traits, zootechnical performance, clinical signs, immunohistochemical and immunological parameters.
2. Materials and Methods
2.1. Animal Housing and Experimental Challenge
The animals susceptible to verocytotoxic O138
E. coli were selected by polymerase chain reaction for the polymorphism on α 1,2 fucosyltransferase (FUT1) gene [
24,
25]. Genomic DNA of each pig was extracted from duodenum tissue using the Wizard
® Genomic DNA Purification Kit (Promega Italia S.r.l, Milan, Italy).
PCR reaction for the FUT1 gene was performed using 0.3 μM of each specific primer (F: CTTCCTGAACGTCTATCAAGACC; R: CTTCAGCCAGGGCTCCTTTAAG) in a volume of 20 μL containing 2 μL of 10× standard buffer, 0.75 mM MgCl
2, 160 μM dNTPs, 1 U Taq polymerase (Fisher Molecular Biology, Rome, Italy), and 1 μL containing 50–100 ng template DNA according to the following steps: denaturation at 95 °C for 3 min, 30 cycles of denaturation at 95 °C for 30 s, annealing at 56 °C for 30 s, and extension at 72 °C for 30 s in a thermocycler (Applied Biosystems 2720, Waltham, MA, USA) [
26]. The FUT1 amplicon underwent restriction analysis with 7 U HhaI (Thermo Fisher Scientific, Waltham, MA, USA) in a final volume of 25 µL overnight at 37 °C, and electrophoresis in 10% polyacrylamide gel for visualization of the restriction fragments by GelRed (Olerup SSP AB, Stockholm, Sweden).
Piglets (Large White × Landrace) aged 22 ± 2 weaned at 21 ± 2 days from a conventional farm free from diseases, without history of verocytotoxic
E. coli infections and negative for the presence of haemolytic
E. coli in faeces, were transported to the Experimental Animal Research and Application Centre of the University of Milan in Lodi (Italy), allocated in individual pens (100 × 50 cm) and maintained under controlled conditions (27–29 °C; relative humidity 60%), with plastic slatted floor, nipple drinker, environmentally safe eco and easily chewable enrichment material; water and feed were administered ad libitum. This in vivo trial complied with Italian law on animal experimentation and ethics (Italian Health Ministry authorization number 102/2015-PR) in accordance with European regulation [
27].
The animals, checked daily for health status, were divided in two experimental groups: non-infected controls (C;
n = 6) and infected ones (I;
n = 12). The infection was performed at day 20 post arrival with verocytotoxic
E. coli O138 strain previously tested positive for F18 and Vt2eB genes [
28]. One hour before infection, piglets were sedated intramuscularly with azaperon (2 mL/each, Stresnil
TM, Janssen Cilag SpA, Milan, Italy), thereafter 30 mL of 10% bicarbonate solution (Sigma-Aldrich, Saint Louis, MO, USA) was orally administered [
29]. After 10–15 min, the bacteria inoculum was given (5 mL containing 10
10 O138 F18
E. coli CFU) through intragastric catheter according to Rossi et al. [
19].
The piglets were fed a basal diet with zinc content 50 ppm, without antibiotics (Ferraroni S.p.A, Bonemerse, Italy), except for three days after the infection, when they received a high protein content diet (25.42 g/kg dry matter) (
Table 1).
2.2. Zootechnical Evaluation and Sample Collection
The rectal temperature was measured daily along with the faecal consistency, which was recorded according to a four-level scale: 0 = normal (faeces firm and well formed), 1 = soft consistency (faeces soft and formed), 2 = mild diarrhoea (loose faeces, usually yellowish), 3 = severe diarrhoea (faeces watery and projectile).
Faecal samples were collected from rectum every day after the infection for microbiological assays and immunoglobulin titration.
The typical clinical signs of VTEC infection (palpebral oedema, epiphora, vitality, depression, hair, perineal area, respiratory and neurological problems) were registered daily by the same veterinary surgeon, according to a three-level points scale score described by Rossi et al. [
19]. Palpebral oedema is defined as a pathological condition (related to the vasopermeabilization caused by verocitotoxins) characterized by the accumulation of fluid in the tissues of the inner part of the eyelids with different degree of severity: 0 = normal; 1 = mild bilateral oedema (puffy eyelids giving a sleepy appearance); 2 = severe (prominent and closed eyelids). Epiphora is an overflow of tears due to the inflammation and the reduced motility of eyelid or of the lacrimal pump. It was classified with the following score: 0 = normal; 1 = mild (moderate presence of eye discharge material in the corner of the eye); 2 = severe (abundant presence of brown discharge material in the corner of the eye); for the vitality and depression scores a disturbed behaviour was defined as slow reactions, an unsteady and slow gait whilst walking and an inattentive response when encouraged to move. Vitality score: 0 = good; 1 = loose (failure to react to stimulus); 2 = bad (slow response to stimuli); depression score: 0 = normal status; 1 = mild (slow reactions); 2 = high (lethargy); perineal area: 0 = clean; 1 = smear; 3 = smear with flogosis (red area caused by inflammation); respiratory score: 0 = normal; 1 = slightly quick respiratory rate; 2 = quick respiratory rate with open mouth; neurological score: 0 = normal; 1 = mild symptoms (incoordination); 2 = severe symptoms (lateral position, paddling limbs, central nervous symptoms); hair condition was evaluated on the dorsal of porcine skin. Hair score: 0 = regular (smooth, clean, flat and uniform); 1 = slightly irregular (fuzzy hair coat and/or scaly skin); 2 = irregular (bald patches, or a rough, dull uneven coat and reddened skin).
All piglets were individually weighed on day 0, 7, 13, 20, 22, 27 and 29 post infection while the feed intake (FI) was daily measured weighing the residual feed at the pen level (experimental unit for FI evaluation). Then, individual average daily gain (ADG) and gain to feed ratio (G:F) were calculated.
Blood was collected from the jugular vein of each animal on days 20, 27 and 29 to determine microhematocrit and serum antibodies. Two animals of C (C3) and four animals of I (I3) groups were sacrificed at day 3 post infection and all the others were euthanized at day 29 (C9/I9). Samples (approximately 0.5 cm3) of jejunum, mesenteric lymph nodes and intestinal scrapes were obtained immediately after euthanasia and placed in RNA Later (Qiagen, Germany) for gene expression analysis; another part of the same excised samples was frozen for immunochemistry, and another one was fixed in 10% neutral buffered formalin for immunohistochemistry/histology. Samples of the ileum (one for each euthanized animal) were collected immediately after the euthanasia and carefully sealed with two sterile cotton ligatures, 3 cm apart from the proximal to the distal one and carried to the microbiology laboratory at 4 °C for microbiological analyses.
2.3. Microbiological Analysis
Faecal samples were analysed for the animal selection and the confirmation of experimental infection procedures. Briefly, faeces (1 g/animal) homogenized in 1 mL of sterile saline buffer were plated onto 5% sheep blood agar plates (Blood Agar Base No. 2, Oxoid, Hampshire, UK) and incubated overnight at 37 °C to evaluate the presence of hemolytic colonies. Up to 5 hemolytic colonies were selected from each plate and grown on MacConkey agar (Oxoid, Hampshire, UK), Triple Sugar Iron agar (Oxoid, Hampshire, UK), Simmons Citrate agar (Oxoid, Hampshire, UK), and Buffered Peptone Water Broth (Oxoid, Hampshire, UK). Colonies positive for glucose oxidation–fermentation, fermentation of lactose (MacConkey agar), indole production (peptone water + reagent Kovacs—Kovac’s reagent for indoles—Fluka, Saint Louis, MO, USA) and sodium citrate-negative were then subjected to biochemical tests with the API system® (API 20 NE—BioMerieux, Marcy-l’Étoile, France) for a more precise identification. The identification of the Lactobacillus spp. and Enterobacteriaceae was performed on plates specific for each type of bacteria and incubated under adequate conditions.
A semi-quantitative approach, based on serial dilutions in medium allowed for the determination of the bacteria concentration (detection limit 1 CFU/g).
E. coli strains were isolated from faecal samples of each pig, cultured (37 °C for 24 h) on Blood Agar and MacConkey agar (Oxoid, Hampshire, UK), further isolated (37 °C for 24 h) on Trypticase Soy agar (Oxoid, Hampshire, UK) and biochemically identified with API-20E method (BioMerieux, Marcy-l’Étoile, France). Besides the evaluation of hemolytic activity, they were serotyped by using monospecific antisera for the somatic antigen O138 [
30]. Finally, the genetic characterization was performed by PCR on isolated
E. coli strains for VT, VT1, VT2 and VT2e; the VT2e positive strains were also tested for F18 [
31,
32].
2.4. Histopathology and Immunohistochemistry
The organs fixed in 10% neutral buffered formalin were paraffin wax embedded, sliced with a microtome into five μm thick histological sections and stained with hematoxylin and eosin. Samples were examined for presence of inflammation both in villi and in lamina propria (infiltrates of lymphocytes, plasma cells, histiocytes and eosinophils), epithelial regeneration (enterocytes with high nucleus/cytoplasm ratio on the intestinal layer), fusion of villi, oedema in deep lamina propria, and T atrophy (T-dependent ipocellularity areas), stroma (fibroconnective and histiocytes) and follicular hyperplasia. All histological parameters were semi-quantitatively scored: 0 = absent, 1 = slight, 2 = moderate; 3 = strong.
Sections of formalin-fixed paraffin-embedded tissues were dewaxed and unmasked by heat induced epitope retrieval and Buffer H (Bio-Optica, Milan, Italy); endogenous peroxidase activity was blocked by incubation with 3% H
2O
2. After rinsing and treatment with PBS, containing normal serum to reduce nonspecific background staining, sections were incubated with the primary antibodies specific for Iba1 (Wako, Neuss, Germany), CD3 (Dako, Næstved, Denmark), CD20 (Thermo Scientific, Rome, Italy), IgG (Vector Laboratories, Burlingame, CA, USA) and IgA (Abcam, Cambridge, UK) at the suggested dilutions. Sections were then incubated with biotinylated secondary antibody (rabbit anti-goat, Vector Laboratories, Burlingame, CA, USA) and labelled by the avidin-biotin-peroxidase (ABC) procedure (commercial immunoperoxidase kit VECTASTAIN
® Elite ABC-Peroxidase Kit Standard, Vector Laboratories, Burlingame, CA, USA). The immunoreaction was visualized with 3,3′-diaminobenzidine substrate (Peroxidase DAB Substrate Kit, Vector Laboratories, Burlingame, CA, USA) after counterstaining with Mayer’s haematoxylin (Merck, Kenilworth, NJ, USA) [
33] and blind evaluated with a light microscopy by a veterinary pathologist. Known positive control sections were included in each assay. CD3, CD20 and Iba1 immunoreactions were semi-quantitatively scored in
lamina propria (0 = absent, 1 = rare cells, 2 = some cells; 3 = numerous cells; 4 = very numerous cells), while IgG and IgA immunoreaction was semi-quantitatively scored on the epithelial luminal surface and
lamina propria (0 = absent, 1 = slight, 2 = moderate; 3 = strong).
2.5. Immunoenzmatic Assay
Intestinal scrapes were homogenized using a rotor-stator system (Ultra Turrax T25, Staufen im Breisgau, Germany) in lysis buffer (50 mM Tris–HCl, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, Sigma-Aldrich, Saint Louis, MO, USA) with protease inhibitor cocktail (1 mM PMSF, 5 μg/mL Complete Protease Inhibitor Cocktail, Roche diagnostics, Mannheim, Germany) and centrifuged at 470× g for 15 min at 4 °C. The supernatants were collected and the protein content was quantified by direct absorbance measurement at 280 nm in quartz cuvette.
The titers of serum and intestinal mucosa total IgA (as mg/mL and ug/mL, respectively) were determined using a sandwich ELISA (Swine IgA ELISA Quantitation Set, Bethyl Laboratories Inc., Montgomery, TX, USA) according to the manufacturer’s instructions. Intestinal mucosa TNF-α, IL-8 and CXCL9 (MIG) levels (ng/mg of total proteins) were detected using three specific Swine Do-it-Yourself sandwich ELISA kits (Kingfisher Biotech Inc., St. Paul, MN, USA), while intestinal mucosa IL-1β was determined using Porcine Interleukin 1β ELISA kit (Cusabio Life Science, Houston, TX, USA). Negative and blank (buffer) samples were included on each plate along with duplicates of samples.
2.6. RNA Extraction, Reverse Transcription and Real-Time PCR Assays
Total RNA was isolated from the samples homogenized by a rotor-stator system (Ultra Turrax T25, Staufen im Breisgau, Germany) in 2 mL of TRI
®Reagent (Sigma-Aldrich, Saint Louis, MO, USA). Purity and concentration of total RNA were spectrophotometrically determined (BioPhotometer Eppendorf, Hamburg, Germany) and 1 µg of total RNA from each sample was reverse transcribed to cDNA in a 25 µL final volume using the High-Capacity cDNA Archive kit (10 min at 25 °C, 60 min at 37 °C and 5 min at 95 °C) (Applied Biosystems, Waltham, MA, USA) [
34]. The resulting cDNA was assayed in Real-Time PCR (7000 Sequence Detection System, Applied Biosystem, Waltham, MA, USA) [
35] to quantify the expression of the best-characterized innate immunity receptors TLRs (2 and 4) in swine and pro-inflammatory cytokines IFN-γ and IL-1β in jejunum, and the expression of the antigen-presenting molecules MHC (type I and II) in mesenteric lymph node cells. Specific primer pairs were designed on the NCBI nucleotide sequences database. The porcine housekeeping gene beta-actin was used as endogenous control. All the primers were custom synthesized (Invitrogen, Waltham, MA, USA) (
Table 2).
Gene expression of each sample was quantified in relationship to an animal chosen as calibrator (relative quantification) and normalized using the calculated beta-actin cDNA expression (mean) of the same sample and run [
34] and results were reported as relative expression compared to beta-actin.
2.7. Statistical Analysis
The results were analysed using the Wilcoxon Rank Sum test for unpaired samples using JMP® Pro 15 (SAS Inst. Inc., Cary, NC, USA). The statistical differences were considered when p ≤ 0.05.
Differences were first tested between C and I animals. Furthermore, the differences between 3-day and 9-day sacrificed animals were verified within C and I groups, respectively. Finally, the differences between C and I animals within 3-day and 9-day sacrificed animal groups were also tested. All measurements were assessed using the individual pig as an experimental unit. The results are presented as medians and range (minimum–maximum).
4. Discussion
An appropriate disease model is necessary to study the pathogenesis of the multifactorial PWD and to develop control strategies alternative to the use of antibiotics. The disease depends not only on virulence or pathogen–host interaction, but also on pig resistance and immune response. In this experimental model, in order to reduce the number of animals, piglets were chosen after polymorphism analysis of the FUT1 gene, correlated with F18 receptor expression [
36], susceptible to infection with
E. coli O138.
In order to bypass the first line of body’s defence against the invading microorganism, prior to experimental infection, piglets received a solution of bicarbonate, which reduces the neutralization of ingested bacteria by gastric acids.
The high-CP diet administered during the first three days after the infection was intended to be a “high-risk” diet, since several pathogens preferentially ferment proteins and high-CP diet in newly weaned piglets is indicated as one of the predisposing factors of VTEC infection. Moreover, the high-CP was due to soybean, which seems to favour the occurrence of
E. coli infection [
19].
The occurrence of VTEC infection was confirmed by bacteria faecal shedding in all the I group piglets and none of the C group ones. It peaked at day one post infection for the majority of animals and subsequently decreased in the following 7 days. A broad spectrum of clinical outcomes, including diarrhoea and enterotoxiaemia, which typically do not occur in all infected piglets, was also observed.
Zootechnical performances were impacted after challenge. The ADG reduction (days 20–22) was linearly related to the FI reduction in I group. Despite this significant reduction, the final BW was not statistically different in the two groups, probably due to compensatory FI starting 4 days after challenge, normally observed in E. coli infection. The feed efficiency was lower in the infected animals; the total recovery may take far longer than this experimental period.
The clinical symptoms of E. coli O138 infections are mainly related to verotoxins and F18 adhesins, which damage the vascular endothelium of the small intestine, subcutis, and brain and ultimately lead to subcutaneous oedema and neurological disorders. The animals showed, after infection, a broad spectrum of clinical outcomes, including specific (palpebral oedema and epiphora) and aspecific (depression, reduction in vitality) symptoms. In particular, the medians of sum of the daily scores of palpebral oedema, epiphora and vitality resulted significantly higher in the infected group; this variation could be used to monitor the infection in the acute phase (Σ3). A higher depression level and a reduction in the vitality were also measured in infected animals, both in acute and subacute phases, highlighting the importance of considering the role of behavioural alteration on disease outbreaks that often contribute to a reluctance to move.
The faecal scores were higher in infected animals, but not statistically significant. This may be explained with the limited number of animals enrolled in this study and possibly also for the increased intestinal fermentation due to high-CP diet administered to both experimental groups. The effect of the infection on the hair score was observed 9 days post bacterial inoculation (hairs appeared normal in the first few days). The development of fuzzy hair coat, lumps, scaly skin and bald patches is typically used in field as a parameter to identify compromised pigs and/or to evaluate their welfare. It is not considered a specific symptom of stress reaction that can be considered in the general evaluation of health status. The absence of mortality may depend on the adoption of a unique infection model, in line with the 3Rs guiding principles for more ethical use of animals in testing, and on environmental condition. The group housing of the animals contributes to increase the spread of the bacteria in the environment and the transmission.
Considering the immune parameters, proinflammatory cytokines and chemokines, such as TNF-α, IL-1β, IL-8, CXCL9 and IFN-γ, directly or indirectly mediate inflammatory response and are essential in initiating effective protection [
37]. Pig weaning has been reported to upregulate the expression of inflammatory cytokines in the gut during acute stress [
38]. The decrease in cytokines and chemokines (TNF-α and CXCL9) from day 3 to 9, observed in C group, may be explained with a diminished acute stress. In humans, it is known that the proinflammatory cytokine production increases with acute stress and decreases with chronic stress.
Immunohistochemistry confirmed the increase in inflammatory parameters in all infected animals.
Once the invading pathogen enters the host through mucous membranes, local rather than systemic immunity is essential. Interestingly we observed a dramatic statistically significant decrease in CD3+ T cells at day 3 in the intestinal epithelium. This may be explained with the ability of the pathogen to downregulate the expansion of newly recruited T cells into the epithelium, impairing the immune response. At day 3, B lymphocytes were absent in both C and I groups in the epithelium and they slightly increased in the I group in lamina propria, as indicated by CD20 staining. On the contrary, at day 9, a statistically significant increase in CD3+ T cells was observed in the I group in the intestinal epithelium and a slight increase (not statistically significant) in the lamina propria, probably as result of the effective host immune response, triggered by the cytokines. No differences in B lymphocytes were observed at day 9 between C and I groups, either in the intestinal epithelium nor in lamina propria. Overall, these results indicate a predominant cellular immune response at intestinal level, both in acute and chronic phases, that can be considered for the in vivo evaluation of alternative to antibiotic compounds.
T cells play a major role in pathogenesis of E. coli diseases; they respond to antigens in a selective and balanced fashion that allows for mounting an effective cell-mediated immune response.
IgA, secreted by plasma cells in intestinal
lamina propria, are a major component of the local immune barrier of the gut and play a crucial role in intestinal homeostasis and protection; therefore, they are used as indicator of intestinal mucosal immunity [
39]. At day 3 post infection no statistical difference was observed between C and I groups, while at day 9, the intestinal scrapes of I pigs displayed a significant lower amount of IgA than C pigs as consequence of the bacterial infection [
40]. It is unlikely that maternal IgA could be adsorbed by small intestinal enterocytes of few-week-old pigs due to intestinal closure and gut maturation [
41]; in contrast, it is possible that maternal IgA interact with intestinal microflora in the lumen and subsequently regulate the host immune response. Since one of the roles of IgA in the intestine is to bind the pathogenic bacteria and promote the internalization of the complex by M cells, it is likely that the amount of IgA present in the intestinal lumen substantially decreases after bacterial invasion. In addition, IgA undergo strong conformational changes upon antigen engagement and binding, which may explain the lower amount of IgA measurable by immunoassays. This occurs in the intestine only, as confirmed by our data, which show that levels of IgA in serum are about the same in C vs. I groups at both considered time points. Thus, on one side, the intestinal barrier heavily affects the microorganism; on the other side, the infectious agent impairs the gut homeostasis by destabilizing intestinal mucosa.
The kinetic of serum IgA, similar in C and I groups, showed a decrease in titer from 3 to 9 days, (statistical tendency only for the C group) as expected in accordance with the normal decline in maternally derived antibodies in the post-weaning piglets [
42].
The assessed panel of clinical symptoms can represent a useful tool in order to evaluate the effectiveness of a further intervention that mitigates or abate the severity of the E. coli O138 challenge. The clinical scores related to the acute symptoms (oedema, vitality, and depression) were discriminant after both 3 days and 9 days post-challenge and represent a consistent and valuable pattern of observation on a daily basis that can offer important information. On the contrary, the effect of the experimental infection on hair status was more indicative after 9 days post-challenge. Even if the observed inflammatory and immunity parameters provided interesting data, due to the high individual variability, a large sample numerosity could be necessary to better underline the differences.