We analysed the safety, immunomodulatory effects, antibiotic resistance, and capacity to inhibit Salmonella spp. and C.J. of B.S. strains isolated from diverse sources.
The evaluation of haemolytic activity is recommended if the isolated bacteria are intended to be used in food products [27]. Lack of haemolytic activity is essential during the selection of probiotic strains because a lack of haemolysin ensures that virulence will not appear among the bacterial strains [1]. Strains A, B, C, D, F, and G were γ haemolytic, while strains E and DSM 17299 were α haemolytic, and strains H and I were β haemolytic (Table 1). Although the E strain showed alpha haemolysis, this haemolysis was very discreet. Thus, as the DSM 17299 strain was also α haemolytic, the E strain was not excluded only by this attribute.
For mucin degradation, all strains were positive except strains C and I (Table 1). Since the commercial strain (DSM 17299) used as a control was positive for mucin degradation (Table 1) and we found an absence of necrosis and apoptosis in the primary culture of fibroblasts in strain C (Fig. 1), we considered that these strains should be further analysed. Commensal bacteria may penetrate the intestinal mucus barrier without harming the host [28]. However, severe changes in the intestinal barrier structure can affect its function. Commensal strains of Bifidobacterium used in the food industry for decades have the potential to degrade mucin in vitro [29]. Thus, as an isolated feature, mucin degradation is not a risk indicator in some cases.
Assessment of the apoptosis and necrosis index is appropriate for candidate strains for probiotics. We inoculated high doses of BS (5 log CFU/well) in chicken fibroblasts and evaluated the total number of adherent cells, necrosis, and apoptosis index. Interestingly, 18 hours postinoculation, strain I led to a decrease in cells marked with DAP (Fig. 1), indicating that many dead cells had detached from the plate. The YP/DAP and IP/DAP increased (Fig. 1), showing an increase in apoptosis and necrosis, respectively. This result indicates that strain I is not safe because it kills cells by mechanical necrosis and apoptosis. Strain H increased necrosis but not apoptosis at 36 hours after inoculation (Fig. 1). As strains H and I were β haemolytic, leading to more significant cell death 18 and/or 36 hours after inoculation in chicken fibroblasts, these strains were not considered safe and were excluded from further analysis.
Strains A, D, and F led to higher mortality in C.Es. The C.Es. inoculated with strain G presented no deaths (Fig. 2). As C.Es. in early and intermediate incubation stages are more sensitive, even nonpathogenic bacteria in high doses can probably lead to death. Therefore, we considered an acceptable mortality rate similar to that of the commercial control. In this way, we evaluated mortality and the subsequent results to assess the selection of strains.
Regarding the biochemical results, the selected strains had a lower albumin-to-protein ratio but a C-reactive protein concentration similar to the P.C. (Fig. 3). The level of C-reactive protein increases in blood in response to inflammation, infection, or tissue damage [30] and is an important marker of inflammation in dogs and humans. However, few recent studies have evaluated C-reactive protein in birds or C.Es. Although previous studies have shown that C-reactive protein can be an inflammation marker in chickens [31] C-reactive protein does not rise in chickens as quickly as it does in humans [32].
Chicken embryos infected with infectious bronchitis virus do not have increased C-reactive protein levels[33] Our study did not find an increase in C-reactive protein even in the positive control. The dynamics of the release of this marker from inflammation may be different in C.Es., and other acute phase proteins should be indicated for study in chickens [34]
In inflammatory processes, there is an increase in total plasma proteins because globulins rise and occasionally decrease albumin, causing a decrease in the albumin/globulin ratio. Often, the total proteins may be in normal ranges, although the albumin/globulin ratio decreased, so this relationship has greater clinical significance. In our study, there was no decrease in the albumin/protein ratio, perhaps because the phase of acute inflammation had passed. This hypothesis should be considered since there was an increase in IL-4 and IL-10 in the positive control (Table 2), showing a phase of the immune system's resilience. On the other hand, in dehydrated birds, an increase in albumin is evident because albumin increases, while total protein can be low. The increased dehydration can be explained by the fact that the injured C.E. has increased energy requirements or respiratory rate, losing more water than the others[35]
We quantified Il-6, a proinflammatory cytokine generated by innate and adaptive responses. It is interesting to study this cytokine because, in the intestine, it modifies the expression of different tight junction proteins and increases tight junction permeability [36–39]. Even as a proinflammatory cytokine, IL-6 may indicate an immunomodulatory effect when it increases concomitantly with anti-inflammatory cytokines such as IL-10 [40]. Our study only found IL-6 in the C.Es. inoculated with S.P. (positive control), indicating that inflammation was not induced in the strains tested.
The positive control exhibited increased IL-6, IL-10, and IL-4 (Table 2) because S.P. caused inflammation, and the immune system tried to modulate the inflammation caused by S.P., similar to what occurs in the newborn animal[41, 42]. None of the strains studied increased IL-6, strain A increased IL-10, and the IL-4 level was similar to that in P.C. (Table 2). The increased concentration of IL-4 in strains A, C, D, E, G, and DSM 17299 at similar levels of P.C. may indicate an immunomodulatory effect of the strains inoculated. However, strains C, D, E and G showed levels of the cytokine IL-4 similar to those in the positive control but also in the negative control.
The safe and probable immunomodulatory effects of the B.S. strains should be interpreted in conjunction with several results. Since strains A, B, D, and F increased the embryo mortality rate and/or the albumin-to-protein ratio, we excluded these strains from our work; our objective justifies using B.S. with high safety in both newborn animals and C.Es.
C.Es. are a valuable in vivo model to evaluate probiotic safety, and the results obtained with the C.Es. inoculated with strain G indicate that this strain could bring benefits to C.E. growth (Fig. 2) in addition to other beneficial effects. Moreover, the villus height of C.Es. inoculated with strains G and E was higher than that of the commercial strain and the negative control (Fig. 4).
Previous studies have shown that whether B.S. causes beneficial or detrimental effects in C.Es. is strain dependent. Seeking to understand the beneficial effects of B.S. in hatchability, chick performance, and intestinal microflora, studies have shown that early B.S. probiotics inoculated in ovo can colonize the small intestine and create a deleterious environment for pathogenic bacteria that could impair chick health. As beneficial effects are obtained when probiotics are added to the feed, early inoculation in ovo could induce earlier stimulation of the immune system to confer protection as soon as the chicks reach the poultry houses [43, 44].
From the nine initial strains, we selected three strains (C, E, and G) to test the inhibitory effect against S.E., S.H., SI, S.M., S.T., S.V.M, APEC and C.J. In this study, we found that the B.S. strains tested have diverse degrees of inhibitory effects, and the effect is strain dependent. For the trial with 12 Salmonella spp., APEC and C.J., B.S. strains C and E had some degree of inhibition for 75% (9/12) and 91,6% (11/12), respectively, and strain G had moderate, strong, or very strong inhibitory effects in all pathogenic strains.
The first and second most commonly reported zoonoses in humans in the European Union in 2018 were campylobacteriosis and salmonellosis, respectively. C.J., S.T., S.V.M, S.E., and S.I. are among the most common species and serovars associated with disease and are prevalent and associated in poultry meat [45]. Several studies have tested the dietary effect of B.S. in chickens challenged with distinct Salmonella spp. and found an exclusion effect [46–50]. Similarly, the B.S. anti-Campylobacter effects in poultry are well documented but are variable and strain specific [51], as confirmed in the results found in this study. Strain G presented strong inhibitory effect on APEC that is an important pathogen to poultry production [52]. Our work makes it clear that the selected strains have action on different Salmonella serotypes, APEC and C.J. in addition to being safe. Since antibiotics are considered harmful chemicals and lead to increased antibiotic-resistant bacteria, dysbacteriosis, and drug residues in food products, the use of probiotics in the poultry industry has become popular in recent years. A probiotic included in commercial formulations, such as GalliPro (Chr Hansen) and Alterion (Novozymes), can improve chicken feed conversion and body weight, reduce lesions caused by Clostridium perfringens, elongate intestinal villi and modulate the microbiota to improve intestinal Lactobacillus concentration and reduce pathogens such as Salmonella and Campylobacter by competitive exclusion and other mechanisms. In this study, we have found secure and efficient strains of B.S. to inhibit Salmonella and Campylobacter in vitro. Further studies must be performed to understand the in vivo effects of the selected strains, either in feed or in ovo.
The European Food Safety Authority (EFSA) establishes specific parameters for testing antimicrobial resistance in all microorganisms used as food additives for humans and animals through the MIC and tetracycline, erythromycin, gentamicin, and vancomycin[27] antimicrobials of choice. Our results showed that the selected strains did not resist these antibiotics (Table 4), increasing the safety of inserting these strains as additives for animal production. In addition, we analysed the antibiotic sensitivity of the main antibiotic classes, and there was no resistance (Table 3).