Introduction

Antibiotics have been used widely and massively as growth promoters and therapeutic agents to stop infectious diseases in livestock [1]. Indeed, so-called growth promotion is a sparing effect on the animal immune system, which would otherwise be a metabolic burden on livestock performance. This practice has been profitable for farmers, allowing production with huge economies of scale and minimal animal death due to pathogenic bacteria. The price paid for this boon has been the development of a reservoir of bacterial resistance to antibiotics, now called the “resistome,” which includes multi-resistant strains that are transmissible throughout the production chain [2]. Even more troubling, residues of antibiotics can end up in human foods, especially meat and its derivatives. Consequently, the use of antibiotics as growth promoters has been prohibited in the European Union (under directive 2001/82/CE) and in other countries [3]. Total replacement of antibiotics in livestock production has bolstered the development of new strategies of pathogen control, including the use of probiotics [4]. According to the United Nations Food and Agriculture Organization, probiotics are “live microorganisms, which when administered in adequate amounts, confer a health benefit to the host” [5]. The key criteria for selecting a microbial strain as a probiotic are safety, survival under gastrointestinal conditions, and ability to inhibit pathogens [6]. Most probiotics approved so far are species of lactic acid bacteria [3, 7, 8]. However, other bacteria have been evaluated, such as Escherichia coli Nissle 1917, a potent inhibitor of Salmonella spp. as well as other enteric pathogens and now used in the first approved probiotic formulation [9]. Studies of potential pathogenic properties are required in order to qualify strains of E. coli or Enterococcus spp. as probiotic candidates. These non-GRAS species then have to be evaluated for bacteriocin production and survival in the gastrointestinal tract, in particular in the presence of bile salts and at low pH [10,11,12]. Their responses to unabsorbed dietary starch, non-caloric polysaccharides such as fiber and oligosaccharides (major sources of carbon and energy for some probiotics) and mucin (an inexhaustible source of oligosaccharides and glycoproteins in the gastrointestinal tract) also have to be evaluated [13, 14]. Strains of non-GRAS genera such Entercoccus sometimes meet these requirements [15]. Notwithstanding non-GRAS status and even potential antibiotic resistance, Enterococcus-based probiotic formulations have been developed and marketed. Examples are Symbioflor 1 (SymbioPharm, Herborn, Germany) and E. faecium SF68® or NCIMB 10415 (Cerbios-Pharma SA, Barbengo, Switzerland) [16], which are currently used for preventing or treating diarrhea in pigs [17]. Probiotic strains of E. coli produce bacteriocins called colicins [18], which were in fact the first bacteriocins to be characterized [19]. Of note, the antimicrobial spectrum of colicins includes pathogenic vancomycin-resistant enterococci [20], and pathogenic strains of E. coli [21, 22]. They bind to specific receptors on targeted cells and form voltage-dependent channels into the inner membrane [23].

Pediococcus has received GRAS status, and strains of P. pentosaceus and P. acidilactici offer ideal probiotic profiles and are already being used as probiotic supplements in animal feed [8]. As for enterococci, one of notable advantages of pediococci is the production of bacteriocins that tolerate gastrointestinal conditions [24]. These bacteriocins are pH-stable and temperature-stable [25,26,27,28]. They bear the N-terminal YGNGV(L) sequence that typifies class IIa bacteriocins and they act through membrane permeabilization or by binding to specific receptors [29, 30].

Probiotic formulations usually contain a single well-characterized strain. However, recent studies show benefits of probiotics based on multiple strains [31, 32]. To develop new therapeutic options for infections by Gram-negative or Gram-positive bacilli in farming and veterinary medicine, we recently isolated strains of E. coli, E. faecalis, and P. pentosaceus from livestock (swine) and claimed their bacteriocin-producing features [33].

In this study, we gained more insights and characterize these strains for their potential health risk, mutual compatibility, and survival under gastrointestinal conditions (TIM-1 dynamic model), allowing to evaluate their suitability as probiotic strain as standalone or mixed in a consortium.

Material and Methods

Bacterial Strains and Growth Conditions

Escherichia coli ICVB442 and ICVB443, E. faecalis ICVB497 and ICVB501, and P. pentosaceus ICVB491 strains were isolated and characterized for their probiotic characters, mainly their antimicrobial activity against Gram-positive and Gram-negative pathogens in a previous report focused on drops of animals living in captivity at the zoological garden of Lille (France) [33]. E. coli ICVB442 and ICVB443 as well as E. faecalis ICVB497 and ICVB501 strains were grown in brain heart infusion (BHI) medium (Sigma Aldrich, St Louis, MO, USA) at 37 °C for 24 h. P. pentosaceus ICVB491 strain was grown at 37 °C for 24 h in de Man, Rogosa, and Sharpe (MRS) [34] broth (Sigma Aldrich). In addition, the probiotic Lactobacillus rhamnosus GG (ATCC 53103) was used as control strain to assess biofilm formation, whereas E. faecalis Symbioflor®1 probiotic strain, Clostridium perfringens DSM 756 pathogenic strain, E. faecalis ATCC 700802, and E. coli ATCC 25922 reference strains were also used as controls for cytotoxicity assays. Staphylococcus aureus ATCC43300 and C. perfringens DSM 756 were used as control strains for hemolysis, antibiotics resistance, and assessment of gelatinase activity.

Evaluation of Pathogenic Properties of the Bacteriocin-Producing Strains

Resistance to Antibiotics

Overnight broth culture of each strain was diluted in peptone water (5 μL in 5 mL). The P. pentosaceus ICVB491 suspension was spread on MRS agar, whereas E. coli ICVB442, ICVB443, and E. faecalis ICVB497, ICVB501 strain suspensions were spread on Mueller-Hinton agar (Sigma Aldrich). Antibiotic discs (Becton-Dickenson, Mississauga, Canada) were deposited on the agar surface, and the plates were incubated overnight at 37 °C. E. coli was tested with gentamycin (10 μg), ampicillin (10 μg), ciprofloxacin (5 μg), cefoxitin (30 μg), nalidixic acid (30 μg), ceftazidime (30 μg), or cefotaxime (30 μg). P. pentosaceus was tested with penicillin (10 μg), gentamycin (10 μg), erythromycin (15 μg), cefotaxime (30 μg), ciprofloxacin (5 μg), trimethoprim-sulfamethoxazole (1.25/23.75 μg), and vancomycin (30 μg). E. faecalis was tested with ciprofloxacin (5 μg), ampicillin (10 μg), gentamycin (10 μg), vancomycin (30 μg), streptomycin (10 μg), chloramphenicol (30 μg), erythromycin (15 μg), and trimethoprim-sulfamethoxazole (1.25/23.75 μg). Antibiotic susceptibility was interpreted in accordance with Clinical and Laboratory Standards Institute (CLSI) recommendations [35], based on the diameter of the zone of inhibition formed around each disc.

Hemolytic Activity

Hemolytic activity was evaluated on Mueller-Hinton agar into which defibrinated sheep blood (5% v/v, Nutri-Bact, Terrebonne QC, Canada) was blended just before solidification (at 41 °C). Then, 18 h cultures at 37 °C in BHI broth (for E. coli ICVB442, ICVB443 and E. faecalis ICVB497, ICVB501 strains) or in MRS broth (for P. Pentosaceus ICVB 491) were used. The broth cultures were streaked onto the blood agar plates and incubated for 36 h at 37 °C. Based on the zone of clearing, hemolytic activity was described as β-hemolysis (complete) or α-hemolysis (partial).

Gelatinase Production

Gelatinase activity was determined using Petri plates containing nutrient medium agar supplemented with 5% (w/v) gelatin (BDH Chemicals, Ltd., Poole, UK). The agar surface was inoculated by streaking the tested strain grown in its usual culture broth. After incubation for 2 to 5 days at 37 °C, Frazier reagent (HgCl2 12 g, HCl 20 mL, distilled water 20 mL) was added to reveal the hydrolysis. Formation of a halo around a colony indicated gelatinase activity.

Biofilm Formation Ability

Biofilm formation was quantified according to a method published previously [36]. E. coli ICVB442, ICVB443 and E. faecalis ICVB497, ICVB501 strains were grown overnight in 3 mL of BHI broth, whereas P. pentosaceus ICVB491 and Lb. rhamnosus GG (ATCC 53103) were grown in MRS broth. Polystyrene micro-assay plates were loaded with broth (230 μL per well) and 20 μL of overnight culture were added in triplicate. Wells containing medium only were negative controls, whereas those containing Lb. rhamnosus GG (ATCC 53103) were positive controls. After incubation for 24 h at 37 °C, the culture (or broth) was poured off, and the wells were washed three times with 300 μL of sterile ultrapure water. Methanol (250 μL) was added to fix adherent bacteria for 15 min. The wells were then emptied, air-dried, stained for 5 min with 250 μL of crystal violet, and the assay plate was placed under running tap water to remove excess stain. After air-drying, 250 μL of 33% (v/v) glacial acetic acid (Sigma Aldrich) were added to each well to resolubilize the dye bound to adherent cells. The OD570 nm was determined for each well using a PowerWave XS2 automated 96-well assay plate reader (Bio-Tek Instruments, Winooski, VT, USA). The cutoff OD (ODc) was defined as the mean OD570 nm of the negative control values plus three standard deviations. Strains were classified on the following basis: OD570 nm ≤ ODc = non-adherent, ODc < OD570 nm ≤ 2 × ODc = weakly adherent and 2 × ODc < OD570 nm = strongly adherent. The ODc was determined for BHI and MRS broths.

Cytotoxicity Assay on the Porcine Intestinal Epithelial Cell Line IPEC-1

The cytotoxicity of bacteriocin-producing strains was determined using a porcine intestinal epithelial cell line-1 called IPEC-1 (DSMZ, Braunschweig, Germany) grown to confluence in the wells of micro-assay plates. The assay is based on measurement of lactic dehydrogenase activity released into the culture medium by lysed cells. Each strain was tested at a concentration 105 cfu/mL or 107 cfu/mL for multiplicity of infection (MOI) equal to 1 or 100. Cultures were centrifuged (8228g, 10 min, 4 °C), washed twice with 4 mL of phosphate-buffered saline (NaCl 137 mmol/L, KCl 2.7 mmol/L, Na2HPO4 10 mmol/L, KH2PO4 1.8 mmol/L), and resuspended in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA) with 10% fetal bovine serum and no antibiotics. The IPEC-1 cells were washed twice to remove the culture medium completely then covered with 100 μL of bacterial suspension. DMEM was used as a negative control. The plates were then incubated for 24 h at 37 °C and 5% (v/v) CO2. Probiotic Symbioflor®1, C. perfringens DSM 756, Enterococcus faecalis ATCC 700802, and Escherichia coli ATCC 25922 were also used as controls. After incubation, 50 μL from each well were transferred to another micro-assay plate for testing with the lactate dehydrogenase substrate, dye, and cofactor (Sigma Aldrich) added as solutions at a 1:1:1 (v/v/v) ratio (100 μL total). The plate was returned to the incubator for 20–30 min. The absorbance in each well was then measured at 490 nm and subtracted from the absorbance at 690 nm. Epithelial cell viability was calculated from the absorbance relative to the negative control.

Mutual Compatibility of the Bacteriocin-Producing Strains

A cross-inhibition assay was performed using the deep-well diffusion method [37]. Overnight culture at 37 °C in 3 mL of BHI (E. coli ICVB442; ICVB443 and E. faecalis ICVB497; ICVB501) or MRS (P. pentosaceus ICVB491) was centrifuged (8334g, 8 min, 4 °C). Half of the supernatant was adjusted to pH 7.0. Both portions were tested in wells in BHI or MRS agar seeded with one of the five strains [38]. Plates were incubated overnight at 37 °C. The presence of an inhibition zone indicated inhibitory activity.

The compatibility assay was performed by re-suspending the centrifugal pellet in 200 μL of supernatant of another strain in micro-assay wells. BHI or MRS broth was used as a positive control and uninoculated broth was used as a negative control. The OD600 was measured every 15 min during 36 h of incubation at 37 °C. The plate was agitated gently before each measurement. The Tlag (time elapsed upon the initial 0.01 increase in optical density) and μmax (specific growth rate at the inflection point) were calculated from the OD600 graph. Peak optical density (ODmax) was noted.

Survival Under Gastrointestinal Conditions

The tenacity of each strain in situ was evaluated using the TIM-1 dynamic model of the human digestive system (TNO Nutrition and Food Research Institute, Zeist, Netherlands), which consists of four flexible compartments connected in series mimicking the stomach, duodenum, jejunum, and ileum. Computer-controlled pumps produce peristaltic movement using external water pressure. Two independent sensors connected to the stomach and small intestine compartments allow temperature control (set at 37 °C). A sensor mounted in each compartment allows monitoring and control of pH. The gastric pH was started at 6.5 and decreased gradually to 3.5 at 30 min, 1.8 at 60 min, and 1.7 from 120 min to the end of digestion at 300 min. The duodenal pH was kept at 6.0, the jejunal pH at 6.5, and the ileal pH at 7.2. Hollow-fiber modules connected to the jejunal and ileal compartments allow dialysis against small intestinal electrolyte solution to simulate absorption. Gastric secretions were composed of 150 U/mg lipase and 3200 U/mg pepsin in gastric electrolyte solution (NaCl 6.2 g/L, KCl 2.2 g/L, CaCl2 0.3 g/L, NaHCO3 1.5 g/L). Porcine bile extract and 8xUSP pancreatin dissolved in small intestine electrolyte solution (NaCl 5.0 g/L, KCl 0.60 g/L, CaCl2 0.30 g/L) were injected into the duodenal compartment. The jejunal and ileal compartments were filled initially with small intestine electrolyte. A test protocol described previously [39] was used with minor modifications. Briefly, 310 mL of skim milk inoculated with 3 mL of culture were added to the gastric compartment. Aliquots of 500 μL were removed therefrom at 0, 30, 60, 90, and 120 min, after 30, 60, 90, 120, and 150 min from the duodenal compartment, and after 60, 120, 180, and 240 min from jejunal and ileal compartments. The effluent was collected, weighed, and aliquoted at 60, 120, 180, 240, and 300 min. The semifluid mass that passed from the stomach to the duodenum intestine at 300 min (chyme) was also weighed.

Viable counts were determined from serial dilutions (to 10−6) in sterile 1% (v/v) peptone water (Thermo Fisher Scientific Inc.). E. faecalis ICVB501 was counted on bile esculin agar (selective for group D streptococci), E. coli ICVB443 on eosin methylene blue agar (selective for coliforms), and P. pentosaceus ICVB491 on MRS agar. Plates were spread in duplicate and incubated at 37 °C for 24 h. Counts were expressed as decimal logarithm (log10) colony forming units (cfu) per mL of skim milk.

Statistical Analyses

The IPEC-1 viability test was performed in triplicate. The mutual compatibility experiment was performed twice, independently for each strain. Data were analyzed using SAS software (SAS Institution, Cary, NC, USA) and expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) with the Tukey multiple comparisons test was used to identify significant differences between treatments (P < 0.05).

Results

Antibiotic Susceptibility Was Strain-Dependent

Based on CLSI interpretation [35], E. coli ICVB442, ICVB443, and E. faecalis ICVB497, ICVB501 were sensitive to all of the antibiotics tested. However, E. faecalis showed intermediate resistance to trimethoprim-sulfamethoxazole. P. pentosaceus ICVB491 was also sensitive, except to ciprofloxacin, trimethoprim-sulfamethoxazole, and vancomycin, to which pediococci are generally resistant [40, 41].

Hemolysis and Gelatinase

E. coli ICVB442, ICVB443, and E. faecalis ICVB497 colonies produced clear zones on blood agar, indicating β-hemolysis. Neither E. faecalis ICVB501 nor P. pentosaceus ICVB491 displayed this undesirable trait (results not shown). All five strains were gelatinase-negative.

Biofilm Formation

The control optical density was 0.145 in BHI broth and 0.149 in MRS broth. E. coli ICVB442 was non-adherent (OD less than half the ODc), in other words, little or no tendency to form biofilm [42]. E. coli ICVB443 (OD = 0.511), E. faecalis ICVB497 (0.680), and P. pentosaceus ICVB491 (0.794) were strongly adherent (OD more than twice the ODc). E. faecalis ICVB497 (0.373) was weakly adherent [42]. The OD obtained for Lb. rhamnosus GG (used as a strong positive control) was 0.658.

Cytotoxicity at Low Concentrations

Cytotoxicity is deemed significant when the IPEC-1 cell layer viability is around 50%, moderate at 60–80%, low at 80–99%, and negative at 100% [43]. As shown in Fig. 1 A and B, C. perfringens was highly cytotoxic at both MOI. Symbioflor® 1, E. faecalis (ICVB497, ICVB501), and P. pentosaceus ICVB491 strains displayed low cytotoxicity at both MOI as no significant difference (P > 0.05) was noticed compared with the untreated IPEC-1 cells, used as controls (Fig. 1A, B). The cytotoxicity of E. faecalis ATCC 700802 was low at a MOI of 1 and moderate at 100. E. coli ATCC 25922 and E. coli ICVB442 showed low cytotoxicity at 1 and high at 100 with percentage of IPEC-1 cells viability significantly (less than 50%) lower (P < 0.05) than the control (Fig. 1B). E. coli ICVB443 was non-toxic at 1 but highly toxic at 100.

Fig. 1
figure 1

Viability of IPEC-1 cells, expressed in percentage, in the presence of E. coli ICVB442, E. coli ICVB443, E. faecalis ICVB497, E. faecalis ICVB501, P. pentosaceus ICVB491, E. faecalis Symbioflor® 1, E. coli ATCC 25922, E. faecalis ATCC 700802, and C. perfringens DSM 756 with a 1:1 MOI (multiplicity of infection) bacterial cell/IPEC-1 cell of 1 (□); and MOI of 100 (■). Values are mean ± standard deviation; different letters indicate significant difference (P < 0.05)

Full size image

Compatibility of the Probiotic Candidate Strains

P. pentosaceus ICVB491 culture supernatant inhibited the growth of the four other bacteriocin-producing strains (Fig. 2A). No E. facealis or E. coli strain reached an ODmax of even 0.2 in its presence. Both E. faecalis strain culture supernatants had a limited effect on the other strains as no significant difference was noticed compared with the untreated control in most cases. Figure 2B shows that each strain, and particularly P. pentosaceus, produced something that slowed the growth rate (μmax) of the other strains significantly (P < 0.05). The P. pentosaceus ICVB491 supernatant also increased the lag time (Tlag) of all strains significantly reaching two- to tenfolds the control (Fig. 2C). Supernatant from E. faecalis ICVB497 and ICVB501 culture increased the Tlag of other strains except E. coli ICVB442, whereas E. coli ICVB443 supernatant had only a slight effect on the growth of the other strains.

Fig. 2
figure 2

Effect of culture supernatant of each bacteriocin-producing strain on the growth of the others: on E. coli ICVB442 (1) and ICVB443 (2); E. faecalis ICVB497 (3) and ICVB501(4); and P. pentosaceus ICVB491 (5). Left to right for each strain: control, CFS of E. coli ICVB442, of E. coli ICVB443, of E. faecalis ICVB497, of E. faecalis ICVB501, and of P. pentosaceus ICVB491. A On ODmax. B On μmax. C On Tlag. Values are mean ± standard deviation; different letters indicate significant difference (P < 0.05)

Full size image

Tolerance of Gastrointestinal Conditions

Based on hemolysis, biofilm formation, and cytotoxicity, two strains (E. coli ICVB442 and E. faecalis ICVB49) were excluded from the gastrointestinal simulation experiment. E. coli ICVB443, E. faecalis ICVB501, and P. pentosaceus ICVB491 were retained for survival test in the TIM-1 model. Figure 3a shows the distribution of the skim milk among the compartments during the 5-h digestive cycle. The viable counts are shown in Fig. 3b, c, d. The initial counts of E. coli ICVB443, E. faecalis ICVB501, and P. pentosaceus ICVB491 in the gastric compartment were, respectively, 9.13 ± 0.01, 9.19 ± 0.08, and 9.20 ± 0.10 log10 cfu/mL. However, after remaining stable for 1 h, the E. coli count fell to 7.79 log10 cfu/mL while E. faecalis dropped to the 5.51 log10 cfu/mL range and P. pentosaceus to the 4 log10 cfu/mL range. In the duodenal compartment, E. coli and P. pentosaceus viable counts decreased by 1 log after 2 h. Viable E. faecalis increased by nearly 1 log during the first hour, then dropped to 8 log10 cfu/mL and remained at that level. The viable counts of the 3 strains remained relatively stable during the passage through the jejunum and the ileum until the end of the simulated digestion. Survival of the proposed probiotic bacteria in the TIM-1 was 139% for E. coli ICVB443, 32% for E. faecalis ICVB501, and 46% for P. pentosaceus ICVB491.

Fig. 3
figure 3

Distribution of components in the TIM-1 compartments during a 5-h digestion. a Carrier medium (skim milk). b Viable E. coli ICVB443. c Viable E. faecalis ICVB501. d Viable P. pentosaceus ICVB491

Full size image

Discussion

The use of probiotics to replace antibiotics is now widespread [4, 44, 45]. Prior to filing any claim about the probiotic value of a bacterial strain, its potential pathogenic risk and its tenacity under conditions as close as possible to those normally prevailing in the gastrointestinal tract must be evaluated [3, 8]. Strains that meet all safety, efficacy, and viability criteria can now be used in consortium formulations, which are being proposed as the next generation of probiotics [7].

In the present study, five bacterial strains were evaluated as probiotic candidates: E. coli ICVB442 and ICVB443, E. faecalis ICVB497 and ICVB501, and P. pentosaceus ICVB491. While the latter species is not a surprising choice, the two other might still raise a few brows. All were tested for antibiotic susceptibility, hemolytic and gelatinase activities, mutual compatibility, and cytotoxicity to a porcine intestinal epithelial cell line. None of these five strains exhibited a gelatinase activity, a property considered as a detrimental virulence factor for pathogenic E. coli and E. faecalis strains. This protease hydrolyse gelatin, casein, and hemoglobin, as well as other active compounds, can compromise the integrity of soft tissues [46, 47].

E. coli ICVB442 and E. coli ICVB443 strains were not resistant to the antibiotics tested. However, they were hemolytic and had a cytotoxic effect on IPEC-1 cells, albeit only at high multiplicity of infection (Fig. 1). E. coli ICVB443 was more inclined to form biofilm. Both E. faecalis ICVB497 and ICVB501 strain formed biofilm, ICVB497 somewhat more. Of note, these promising in vitro results need to be confirmed in vivo as a biofilm formation capability can help a probiotic strain to defy harsh conditions in the gastrointestinal tract, and possibly block colonization of the intestinal mucosa by pathogens [48].

Remarkably, the probiotic grade is attributed for some Enterococcus strains based taking into consideration their global safety profile and the absence of virulence factors. Neither E. faecalis ICVB497 nor ICVB501 strain displayed any antibiotic resistance, and noticeably both of them were devoid of gelatinase activity. However, E. faecalis ICVB497 has a β-hemolytic activity, which practically rules out its possible use as a probiotic. Notwithstanding some controversial aspects attributed to Enterococcus, some strains have benefitted of probiotic grade [49]. In fact, they have been tested for human use [50, 51], for example, to limit diarrheas and inflammatory bowel diseases as well as in livestock [2, 52].

Compatibility assays showed that both E. coli ICVB442, ICVB443 strains and both E. faecalis ICVB497, ICVB501 strains could function in the same probiotic formulation (Fig. 2). However, P. pentosaceus ICVB491 inhibited the other strains, likely because of the acidity of its culture supernatant (based on agar diffusion tests, results not shown), but possibly also by producing an inhibitory substance such as a bacteriocin [33].

Survival during gastrointestinal transit, a key factor for selecting new probiotic strains [53, 54], has been tested successfully in static models [53] and dynamic models [54]. The dynamic model designed 25 years ago by Minekus et al. [54] has proven its effectiveness for evaluating survival of beneficial microorganisms [55, 56] and the persistence of bacteriocins [39, 57], drugs [58,59,60], and other substances [61,62,63] in human and animal digestive tracts.

Of the three strains deemed worthy of testing in the TIM-1 in this study, E. coli ICVB443, which was retained as its non-toxic at 1 MOI, was the most robust, its final viability value reaching 139% of the initial load into the system (Fig. 3), which is consistent with previous similar studies [64]. The P. pentosaceus ICVB491 strain has interestingly revealed comparable features with P. acidilactici UL5 tested by Kheadr et al. [39], using a skim milk culture in a TIM-1 model. Pediococcus strains usually survive better than other lactic acid bacteria in the TIM-1 model [65].

Finally, at 32%, the survival of E. faecalis ICVB501 was the lowest measured in the present study, which has been observed elsewhere for enterococci under such conditions [66, 67].

Conclusion

In summary, E. coli (ICVB442 and ICVB443), E. faecalis (ICVB497 and ICVB501), and P. pentosaceus ICVB491 were tested for their suitability as parts of a consortium-type probiotic formulation. Pathogenicity and cytotoxicity tests including their antibiotic resistance, hemolysis, biofilm forming ability, and their compatibility assays enabled us to select three strains that presented suitable attributes for probiotic use. These strains are P. pentosaceus ICVB491, which belongs to species possessing GRAS status, E. coli ICVB443, and E. faecalis ICVB501 that exhibited worthy features that deserve to be tested under a dynamic model. These strains will be the focus of future research regarding the use of these colicin-producing, enterocin-producing, and pediocin-producing commensal strains, for simultaneous use as a novel consortium probiotic supplement to promote the health of piglets.