Synbiotics for Broiler Chickens—In Vitro Design and Evaluation of the Influence on Host and Selected Microbiota Populations following In Ovo Delivery

Synbiotics are synergistic combinations of prebiotics and probiotics. In chickens, synbiotics can be delivered in ovo to expedite colonization of the gut by beneficial bacteria. We therefore aimed to design synbiotics in vitro and validate them in broiler chickens upon in ovo delivery. The probiotic components of the synbiotics were Lactobacillus salivarius and Lactobacillus plantarum. Their growth was assessed in MRS medium supplemented with different prebiotics. Based on in vitro results (hatchability and growth curve), two synbiotics were designed: S1 –Lactobacillus salivarius with galactooligosaccarides (GOS) and S2 –Lactobacillus plantarum with raffinose family oligosaccharides (RFO). These synbiotics were delivered to Cobb broiler chicken embryos on day 12 of incubation at optimized doses (105 cfu egg-1 of probiotic, 2 mg egg-1 of prebiotic). Post hatching, 2,400 roosters were reared (600 individuals group-1 divided into eight replicate pens). Microbial communities were analyzed in ileal and cecal digesta on day 21 using FISH. Gene expression analysis (IL1β, IL4, IL6, IL8, IL12, IL18, IFNβ, and IFNγ) was performed on days 7, 14, 21, and 42 for the spleen and cecal tonsils with RT-qPCR. Body weight and feed intake of the roosters did not differ by the treatments. Microbial populations of Lactobacillus spp. and Enterococcus spp. in the ileum were higher in S1 and S2 than in the control. In the cecum, the control had the highest bacterial counts. S1 caused significant up-regulation of IL6, IL18, IL1β, IFNγ, and IFNβ in the spleen on day 21 and IL1β on day 7 (P < 0.05). In cecal tonsils, S1 caused significant down-regulation of IL12, IL8, and IL1β on day 42 and IFNβ on day 14 (P < 0.05). S2 did not elicit such patterns in any tissues investigated. Thus, we demonstrate that divergent effects of synbiotics in broiler chickens were reflected in in vitro tests.

Introduction parameters of meat quality [25], digestive potency of the pancreas [26], and molecular changes in the cecal tonsils, spleen [17,18], and liver (A. Dunislawska, personal communication). However, the reported influence of in ovo treatment on the animal was strongly dependent on properties of the given prebiotic (e.g., inulin, GOS, RFO, or beta-glucans). These effects are more complex for synbiotics containing at least two biologically active compounds, which exert different effects on the organism when they are delivered alone than when they are delivered together. Based on numerous studies, careful pre-selection of the synbiotic and its validation for animal studies is crucial. Because in vivo examination of bioactive properties of prebiotics, probiotics, and synbiotics is time-consuming, labor-intensive, and require large numbers of animals, in vitro assays should be preferentially used [27].
In this study, we hypothesized that in vitro screening of synbiotics prior to in vivo study could be indicative of the biological properties synbiotics and their impact on the host. The goal of this study was to design Lactobacillus synbiotics based on in vitro testing and validate their impact on broiler chickens after in ovo injection into 12-day-old embryos. In our experiment, we assessed the impact of treatment with specific synbiotics on the microbiome and host. The animals were studied from hatching to market age under commercial conditions using performance and molecular responses as indicators of the effectiveness of the treatments.

Testing trial
Lactobacilli synbiotics selection in vitro. For in vitro trials, two Lactobacillus strains were chosen, based on their proven probiotic activities in chickens: Lactobacillus (L.) salivarius IBB3154 and L. plantarum IBB3036. Both strains were derived from the collection of the Institute of Biochemistry and Biophysics (IBB), Polish Academy of Sciences (PAS) in Warsaw, Poland. For synbiotic design, several prebiotics were tested: inulin (Sigma-Aldrich GmbH, Schnelldorf, Germany), RFO [28], GOS (trade name: Bi 2 tos; Clasado Biosciences, Ltd., Jersey, UK), and beta-glucans (trade name: LactoShield; BioAtlantis, Ltd., Tralee, Ireland). Growth curves of these bacteria in the presence of different prebiotics were estimated using an automated growth analyzer Bioscreen C (Oy Growth Curves Ab Ltd). The total volume of the well in the Bioscreen plate was 400 μL and was filled with a test solution comprising 360 μL of MRS medium [29] and 40 μL of inoculum of each Lactobacillus bacteria. Several variants of MRS medium were used. These consisted of a standard with glucose and one of four modified MRS media in which glucose was replaced with inulin, RFO, GOS, or beta-glucans [30,31]. Inocula were prepared as follows: tested strains (L. salivarius IBB3154 and L. plantarum IBB3036) were inoculated in tubes with standard MRS medium with glucose and incubated for 16 h at 37˚C [32,33]. Next, they were centrifuged (4000× g for 15 min) and suspended in peptone water. Suspensions were mixed thoroughly and diluted to obtain OD 1 at 600 nm corresponding to 10 6 cfu cm -3 . Growth of the tested strains, L. salivarius IBB3154 and L. plantarum IBB3036, was measured turbidimetrically hourly for 72 h using a wide-range filter (420 nm-580 nm) with 10 s mixing before each measurement. All measurements were performed in triplicate.
Doses optimization in vivo. To evaluate doses of synbiotics for the in vivo study, trials were performed in which effects of synbiotics delivered in ovo on egg hatchability was determined for L. salivarius IBB3154 with GOS (S1) and L. plantarum IBB3036 combined with RFO (S2). A sample of 4,200 Cobb500FF eggs was randomly distributed into 14 experimental groups. On day 12 of embryo development, a control group was mock injected in ovo with 0.2 mM physiological saline (0.9%). Another group was used as a negative control without in ovo injection. The remaining 12 experimental groups were used for synbiotic injection in ovo with six different doses of each synbiotic. Doses were prepared by combining the three levels of bacterial counts (10 3 , 10 4 , 10 5 bacteria cfu egg -1 ) with two levels of prebiotic concentrations (2 or 5 mg egg -1 ). The experimental setup is presented in the Tab A in S1 File. Application of 0.2 mL of aqueous solution into the air cell of each egg was performed with a syringe and 4-mm-long needle. After injection, the hole was sealed with non-toxic glue to avoid embryo contamination and prevent moisture loss. After in ovo injection, eggs were further incubated until hatching. The analysis of hatchability was performed in a commercial hatchery (DrobexAgro, Solec Kujawski, Poland). Statistical analyses of hatchability were conducted using a one-way analysis of variance (ANOVA) and Duncan's post hoc test to compare means between experimental groups.

Field trial
In ovo injection and tissue collection. For the field trial with optimized doses of synbiotics, 5850 Cobb500FF eggs were incubated in a commercial hatchery using an automated incubator at 37.8˚C and a relative humidity of 61%-63%. Egg weight was approximately 65 grams. Eggs used for experiment originated from a herd at the age of 42 weeks. On day 12 of incubation, eggs were randomly distributed into two experimental groups: S1, S2 and the control group. Eggs were injected in ovo with 0.2 mL of either S1 or S2 aqueous solution. Synbiotics consisted of 10 5 bacteria cfu egg -1 and 2 mg egg -1 of prebiotic. The control group (C) was mock-injected with 0.2 mM physiological saline (0.9%). Additionally, an uninjected control (U) was incubated in parallel. Eggs were incubated until hatching. At hatching, hatchability was scored to determine mortality of the embryos resulting from in ovo injection. A total of 2,400 roosters was moved to a commercial farm (Piast, Olszowa Experimental Unit 0161, Poland) for rearing. Animals from each experimental group (600 individuals group -1 ) were split into 8 pens (75 individuals pen -1 ). Each pen was considered a replicate.
During the entire rearing period, chickens had free access to feed and fresh drinking water. Mortality, body weight, fodder consumption, and feed conversion efficiency (FCE) were recorded on a regular basis. FCE was determined using residual feed intake (RFI) and residual growth (RIG). RFI was defined as the difference between observed and predicted feed intake. On days 7, 14, 21, and 42 post hatching, five randomly selected individuals (one from each experimental group) were euthanized. Tissues (spleen, cecal tonsils, and jejunum) were collected and stabilized in fix RNA solution (EURx, Gdansk, Poland). Overall, 60 broiler chickens (4 collection times × 3 groups × 5 individuals per group) were used for gene expression analysis. The remaining 1,980 roosters were used to estimate production parameters. For the analyses of gastrointestinal microbial communities, the contents of the ileum and ceca collected from two birds per pen were pooled (eight replicates of approximately 10 g) on 21 day of life. Samples were immediately frozen and stored at -80˚C for the microbiota composition analysis by fluorescent in situ hybridization of single bacterial cells (FISH).
During the entire rearing period, all experimental broilers were under veterinary care. Chickens exhibiting any signs of severe illness or moribundity were euthanized. Animal use was approved by the Local Ethical Committee for Animal Experimentation, University of Sciences and Technology, Bydgoszcz, Poland 36/2012 on July 12, 2012.
Microbial community analysis by fluorescent in situ hybridization (FISH). From all treatments 2 birds from 9 pens was sacrificed (n = 18) by cervical dislocation. The samples of the digesta from different gastrointestinal segments (ileum, caeca) were collected and randomly pooled into six replicates per treatment, three birds per sample. For FISH analysis, 100 μL of the ileal digesta was diluted in PBS and pipetted onto 0.22 μm polycarbonate filters (Frisenette, Knebel, Denmark) and vacuumed with a vacuum pump (KNF Neuberger GmbH, Freiburg, Germany). After vacuuming, filters were transferred onto cellulose discs for dehydration in growing dilutions of ethanol (50,80, and 96%, 3 min each). For each sample, a series of identical filters was prepared to determine optimal hybridization. Oligonucleotide probes used for this study (Tab B in S1 File) were selected based on the results of previous studies [69][70][71][72]. Hybridizations were conducted in 50 μL of hybridization buffer (0.9 M NaCl; 20 mM Tris/HCl, pH 7.2; 0.01% SDS) containing the oligonucleotides probes. After hybridization, filters were washed with a washing buffer (20 mM Tris/HCl, pH 7.2; 0.01% SDS; 5 mM EDTA) for 20 min at 48˚C. After washing, filters were rinsed gently in distilled water, air-dried, and mounted on slides with VectaShield (Vector Laboratories Inc., Burlingame, CA, US), an antifading agent containing DAPI (4',6-diamidino-2-phenylindole). To distinguish bacteria from other particles in the ileal samples (DAPI), filters were left at 4˚C for 1 h in the dark until visualization with a Carl Zeiss Microscope Axio Imager M2 (Carl Zeiss Jena GmbH, Jena, Germany), as described elsewhere [34,35]. A detailed FISH protocol is presented in the Fig A in S1 File. From one replicate (polycarbonated filter) 50 areas were measured and used as means in calculation. Statistical analyses of the selected microbiota populations represent 6 pooled replicates from 18 birds per group; each visualised by fluorescent microscopy 50 times.
RNA isolation and reverse transcription quantitative PCR (RT-qPCR). Prior to total RNA isolation, tissues were homogenized with the TissueRuptor homogenizer (Qiagen GmbH, Hilden, Germany) in TRIzol1 LS Reagent (Ambion/Thermo Fisher Scientific, Valtham, USA). Further steps of isolation were performed with a commercial kit (Universal RNA Purification Kit, EURx, Gdansk, Poland). RNA quality and quantity was controlled by electrophoresis on 2% agarose gel and NanoDrop 2000 (Scientific Nanodrop Products, Wilmington, USA). In addition, 10% of samples were used to control the integrity of RNA on automated electrophoresis with the Experion Automated Electrophoresis System (BioRad, Hercules, California, USA).
Relative quantification of gene expression and statistical analysis. Relative gene expression analysis was conducted separately for each experimental group by the ΔΔCt method [37] using UB and G6PD as reference genes. Geometric means of Ct (cycle threshold) values of reference genes were used in the analysis. For each of the samples, Ct differences between target and reference genes were calculated. Control (C) samples were used as calibrators. ΔΔCt was calculated by deducting the ΔCt value of the calibrator from ΔCt of the unknown sample. For the calculation of the normalized expression level of the gene, the following formula was used: R = 2 -ΔΔCt [37]. Statistical analyses were performed by comparing the Ct value of each experimental group with that of the control group by Student's t-test (P < 0.05). Standard errors of the mean (SEM) were applied as a parameter of variation within the groups.

Lactobacilli-synbiotic selection in vitro
Results of turbidimetric measurements of tested bacteria in the presence of five prebiotics are presented in Fig 1A (L. salivarius) and Fig 1B (L. plantarum). For L. salivarius, the optimal growth curve was observed in the presence of GOS and the MRS standard medium with glucose. Bacterial growth in the standard MRS medium supplemented with RFO was much weaker than that in MRS medium with glucose. Maximum optical density was achieved in L. salivarius cultures with GOS and was 0.06 OD higher than in the MRS culture with glucose. The optimal growth curve of L. plantarum was detected for the MRS standard medium with glucose. Comparably high growth curves of L. plantarum were observed in the presence of GOS and RFO in the MRS medium. Growth of L. plantarum in the MRS standard medium with glucose was 0.74 OD higher than that in MRS medium with RFO. Bacterial growth in the MRS standard medium supplemented with either inulin or beta-glucans was much weaker than that in the MRS standard medium with glucose.

Dose optimization in testing trial
Significant differences in hatchability of eggs injected in ovo were observed between two doses of prebiotics, 2 mg egg -1 and 5 mg egg -1 . In this experiment 4,200 eggs were distributed to 14 groups. Hatchability was significantly higher after administration of 2 mg egg -1 GOS (Fig 2A) or RFO (Fig 2B). Higher prebiotic concentration (5 mg) caused significant decline in egg hatchability. A hatchability of approximately 90% was observed after S1 injection containing 10 5 of L. salivarius and 2 mg of GOS and after S2 injection of 10 3 or 10 5 of L. plantarum and 2 mg of RFO. Hatchability in the control group (C) was 90%, whereas in the uninjected (U) group was 89%. Based on these results, the optimized dose of both S1 and S2 synbiotics was 2 mg of prebiotic and 10 5 of bacteria. The aim of the selection process was to determine the highest level of bioactive compound that did not impede hatchability.

Broiler chicken performance in the field trial
For the field trial, based on 5,850 eggs, hatchability was 89.1%, 91.6%, and 91.9% in the S1, S2, and C groups, respectively. Body weight differed only on the first day after hatching. Chickens from the S2 group (41.6 g) were significantly heavier than those in the S1 (40.3 g) and C (40.7 g) groups (P < 0.05). We did not find any differences in body weight gain (BWG) among the groups over the entire rearing period. The lowest feed intake (FI) between day 1 and 10 was observed in the S1 group (247 g) and differed significantly from the C (254 g) and S2 (258 g) groups. For the entire rearing period, FI in the S1 group (4,930 g) was comparable to the C (4,940 g) group. The lowest feed consumption (4,898 g) was observed in the S2 group. Feed conversion efficiency (FCE) calculated for the entire rearing period (day 1 to 42) was almost identical for all groups and equal to 1.60 g g -1 in the synbiotic-treated groups and 1.59 g g -1 in the C group. Mortality was numerically lower in the synbiotic-treated groups and equal to 0.83% (S1) and 1.17% (S2). Numerically, the highest chicken mortality was detected in the C group (1.83%). Chicks performance (BWG, FI, FCE and mortality) carried out on 1,980 broilers to estimate production parameters in response to different synbiotics delivered in ovo is presented in the Tab D in S1 File.

Microbial community analysis by fluorescent in situ hybridization (FISH)
In the ileal digesta, irrespective of the treatment, application of the synbiotics lowered Lactobacillus sp./Enterococcus sp. Other microbiota populations exhibited different responses. S1 ( Fig  3A) decreased the total number of bacteria, as well as the Bacteroides-Prevotella cluster, Clostridium leptum subgroup, and the Eubacterium rectale cluster (P = 0.0001). In contrast, in the S2 group, higher counts of the Bacteroides-Prevotella cluster, Clostridium leptum subgroup, and the Eubacterium rectale cluster (P = 0.001) were determined. In cecal digesta (Fig 3B), the highest number of total bacteria was detected in the C group (P = 0.0001). Both synbiotics lowered the Bacteroides-Prevotella cluster and increased the Eubacterium rectale cluster. Moreover, in the S1 group, higher counts of Lactobacillus sp. /Enterococcus sp. were observed (P = 0.0001).

Immune-related gene expression signatures
Based on biological material collected from 60 broilers for expression analysis we can infer that in ovo stimulation with S1 caused significant up-regulation of IL6, IL18, IL1β, IFNγ, and IFNβ in the spleen on day 21 and the IL1β on day 7 (P < 0.05) (Fig 4A). S1 stimulation induced a pattern of up-regulation of gene expression in the spleen for the panel of genes at the four time points: 7, 14, 21, and 42 days of life. Administration of S2 did not cause any significant pattern of immune-related gene modulation in the spleen (Fig 4B). Significant up-regulation was detected for IL8 on day 14 and down-regulation for IL12 on day 42 (P < 0.05). Analysis was performed based on 5 biological repeats, and every in 2 technical repeats.
In the cecal tonsils, S1 caused a significant down-regulation of IL12, IL8, and IL1β on day 42 and IFNβ on day 14 (P < 0.05) (Fig 5A). In this tissue, there was a clear down-regulatory pattern of the gene expression induced by the S1 treatment. Conversely, S2 did not cause a pattern of immune-related gene expression in the cecal tonsils. Up-regulation of gene expression was detected on day 7 for the genes IL12 and IL18 and on day 42 for genes IL8 and IL18. IL12 was down-regulated on days 14 and 21 (P < 0.05) (Fig 5B). Immune-related gene expression in the jejunum in the S1 and S2 groups did not differ significantly from that in the C group.

Lactobacilli-synbiotic selection in vitro
During in vitro selection of different compositions of lactobacilli and prebiotics, growth curves were estimated to determine potency of the given LAB strain to metabolize prebiotic oligosaccharides. Growth of the two tested LAB strains in the presence of four prebiotics was compared to the reference medium, standard MRS medium with glucose. Bacterial growth in the reference medium was defined as having the optimal logarithmic growth curve, and as such, it was treated as the reference growth curve for further comparisons. In our analysis, the logarithmic growth curve was constructed for both LAB strains (L. salivarius and L. plantarum) in the presence of GOS and RFO. The growth curve of L. salivarius in the MRS medium with GOS exhibited a comparable slope and height to that of growth curve in the MRS medium with glucose. This confirmed that L. salivarius utilizes GOS as effectively as glucose.  Conversely, growth of L. plantarum in the MRS medium with RFO was lower than that in the MRS medium with glucose, although the slope of the curve was similar. This could be explained by the more complex structure of the saccharides present in RFO (i.e., trisaccharideraffinose, tetrasaccharide-stachyose, and pentasaccharide-verbascose [22,28]. These results are in accordance with findings of Gulewicz et al. [31], who reported that growth of LAB in MRS medium with RFO was less potent than in that in MRS medium with glucose. There are other reports that confirm the variation in utilization of prebiotics by bacteria, which largely depends on the LAB strain [38,39]. It could be concluded that L. salivarius metabolization of GOS is almost as effective as a basic source of carbon (i.e., glucose), but the L. plantarum RFO prebiotic provided a less efficiently metabolizable source of carbon.
The best combination of prebiotic and probiotic (i.e., synbiotic) for in ovo injection can be seen from two different perspectives. First, the most suitable prebiotic for the probiotic bacteria injected in ovo is the one that supports growth of the probiotic component. In this case, an optimally combined synbiotic allows for more efficient use of both components because the viability of the probiotic bacteria is improved by using the prebiotic as a substrate for fermentation [40]. We define such a synbiotic as exerting synergistic effects towards the probiotic. Second, if the prebiotic is less efficiently used by the probiotic bacteria, it remains available to other indigenous stains of intestinal microbiota. In this situation, the prebiotic has a positive influence on the host organism through improvement of microbial balance in the intestines [41]. In our opinion, such synbiotic can be defined as synergistic toward the host. In the next steps of the in vivo experiments, two synbiotics (i.e., L. salivarius + GOS and L. plantarum + RFO) were designed based on in vitro results. These synbiotics were expected to present the two types of synergism explained above (i.e., synergism between the prebiotic and probiotic components [S1] and synergism between the two independent bioactive compounds and the host [S2].

Dose optimization in vivo
Use of in ovo technology for delivering microbiota-promoting bioactive compounds in chickens has several advantages over in-feed supplementation, including uniformity and precision of bioactive compound delivery to each embryo, low usage of those compounds, optimal timing for stimulating gut colonization with beneficial bacteria, as well as encouraging maturation of the immune system. The associated risks, however, includes potential impediment of embryo viability and hatchability because of improper injection technique or dose of the bioactive compounds. In our case, the injection technique had been optimized and, in every study conducted to date, including the current one, the hatchability score was over 90%. The injection site is an air cell within the egg and the timing, was at day 12 of incubation. This approach is least invasive for the embryo and allows for easy automatization without compromising egg hatchability. The other component that influences hatchability (i.e., doses of bioactive compounds for in ovo injection) must be optimized prior to field trials. In our studies, we assumed the criterion of hatchability as the major indicator of the optimal dosage [42]. By definition, we selected the highest dose of a given bioactive compound that did not cause significant reduction in hatchability. In our case, increasing bacterial count in the synbiotic did not influence the hatchability score. However, there was a considerable decrease in hatchability of the eggs that had been injected in ovo with synbiotics containing 2 mg as against those injected group by Student's t-test; * for P < 0.05. Each bar represents the mean (n = 5) ± standard error of the mean (SEM).
doi:10.1371/journal.pone.0168587.g004 with 5 mg of GOS or RFO prebiotic. Such a difference can be explained by overstimulation of the bacterial population in the embryonic gut, which interfered with embryo development.

Broiler chicken performance in the field trial
In the current experiment, synbiotics composed of Lactobacillus LAB species combined with GOS (S1) or RFO (S2) injected in ovo did not influence final BWG and FCE. In the previous trials, we observed changes in broiler performance traits (BWG and FCE) introduced by in ovo treatment with prebiotics and synbiotics. For example, in ovo delivery of two synbiotics composed of inulin + Lactococcus lactis subsp. lactis or GOS + Lactococcus lactis subsp. cremoris, significantly increased final BWG with unchanged FCE [21]. In another experiment, in ovo injection of GOS and RFO prebiotics alone also increased BWG, but with higher FI and FCE [42]. Changes in BWG and FCE in animals supplemented (in ovo or in-feed) is quite natural and it results from competition for nutrients between the host and its intestinal microbiota, which usually results in higher energy uptake [43]. The lack of changes in performance traits observed in this study was considered beneficial, given that synbiotics delivered in ovo did not increase FCE with consistent high final body weight. The absence of major effects of the treatment on performance traits can be explained by the fact that the current experiment was conducted with males originating from a highly selected broiler line. These chickens had been genetically previously selected for top performance in the field trials. The goal of synbiotic injection in ovo is to maintain the health of the organism, rather than vastly increase its performance. Thus, we achieved numerically improved survivability of the chickens, in both experimental groups (S1, S2).
To date, only a few studies on performance and immunomodulatory effects of synbiotics administered in ovo have been conducted [19,20]. Most of the studies on synbiotic administration examined the classical route of delivery, in-feed or in-water. Mookiah et al. [44] showed that in-feed supplementation of broiler diet with isomalto-oligosaccharides prebiotic, commercial lactobacilli probiotic, or their combination as a synbiotic resulted in significant improvement of BWG, FI, and FCE. Ghasemi et al. [45] observed a similar effect of improved FCE after administration of synbiotics in broiler chicken diets. In contrast, Murshed and Abudabos [46] reported poorer FCE in chickens fed diets with a synbiotic consisting of MOS, beta-glucans, and a highly active strain of Bacillus subtilis (DM17299). Moreover, high doses of dietary probiotic tended to increase FCE in later stages of broiler rearing [47]. Mousavi et al. reported that in-feed supplementation of broiler chickens with a synbiotic (Enterococcus faecium and oligosaccharides) improved FCE in the starting and growing phase, but not in finishing phase [48]. It might be concluded that despite of the mode of delivery in-feed, in-water, or in ovo, the observed effects of synbiotics differ because of biological properties of the bioactive compounds used.

Microbial community analysis by fluorescent in situ hybridization (FISH)
It is well documented that all segments of the poultry GIT are colonized by different populations of microbiota [49,50]. For instance, ceca are the main fermentative chambers with the

Fig 5. Changes in relative expression of immune-related panel of genes in the cecal tonsils of broiler chickens injected in ovo with synbiotics (A) Lactobacillus salivarius with GOS and (B)
Lactobacillus plantarum with RFO. Analysis was performed using the ΔΔ Ct method to determine fold induction. Synbiotics were injected on day 12 of embryonic development. Sampling days were 7, 14, 21, and 42 days post hatching. Statistical analysis consisted of comparing the experimental groups with the control group by Student's t-test; * for P < 0.05. Each bar represents the mean (n = 5) ± standard error of the mean (SEM).
doi:10.1371/journal.pone.0168587.g005 highest activity and density of strict anaerobes. For increased performance of broiler chickens, microbiota density and activity (i.e., fermentation) should be minimized in the upper parts of the GIT (i.e., ileum) and increased in the lower GIT segments (i.e., ceca). In the current study, in ovo application of S1 and S2 synbiotics triggered different microbial colonization of the ileum and ceca and S1 appeared more effective in modulating a beneficial shift in the GIT microbiota. Supplementation with S1 reduced total microbiota content in the ileum, including the Bacteroides-Prevotella, Clostridium leptum, the Eubacterium rectale cluster, as well as Lactobacillus spp./Enterococcus spp. The same treatment increased the Clostridium leptum subgroup, Eubacterium rectale cluster, and Lactobacillus spp./Enterococcus spp. in the ceca. The Clostridium leptum subgroup includes many species that belong to Clostridium, Eubacterium, and Ruminococcus genera. These are mostly butyrate-producing and fibrolytic species, which have significant effects on intestinal health. Changes in the Clostridium leptum subgroup are indicative of the health status of the gut. A lower number of Clostridium leptum subgroup in the ileum and increased number in the ceca could also play a beneficial role in GIT microecology because of higher production of short-chain fatty acids in the ceca and lower competition for nutrients with the host in the ileum. For S1, similar effects were also observed in the Eubacterium rectale cluster. Species of this genus produce organic acids, including butyrate, acetate, lactate, or formate, but not propionic and succinic acids, as major products of dietary fiber fermentation [51,52]. Phylogenetic analysis based on 16S rRNA sequences confirmed that E. rectale belongs to the clostridial cluster XIVa, as defined by Collins et al. [53] within the phylum Firmicutes [54,55]. Therefore, similar to the Clostridium leptum subgroup, abundance of E. rectale is a good indicator of the butyrate producing microbiota, which indirectly affects epithelial cell structure and function, particularly in the lower parts of the GIT.
Microbiota belonging to the Bacteroides-Prevotella cluster are one the most frequently isolated pathogens from clinical specimens, from almost all anatomic sites [56]. Moreover, it has been demonstrated that the Bacteroides-Prevotella cluster can be used to detect fecal contamination in water environments [57]. In present study, application of the S1 synbiotic decreased the number of the Bacteroides-Prevotella subgroup in both the ileal and cecal digesta, S2 only in ileum.
Lactobacilli are one of the most abundant bacteria groups in GIT of broiler chickens. Lu et al. [58] reported that approximately 70% of the sequences in an ileum-derived 16S rRNA gene library is belonged to Lactobacillus genus. However in the ceca, inhabited mostly by strict anaerobes, lactobacilli accounted for only 8% of the library sequences [35,58]. In human and animal nutrition, LAB including Lactobacillus genus are usually considered beneficial for the host, mainly because they produce lactic and acetic acids, which leads to pH reduction [59,60] and certain bacteriocins, which help fighting pathogenic bacteria [61,62]. On the other hand, there is also an evidence that some small intestinal LAB, e.g. L. salivarius or Enterococcus faecium might have negative effects on broiler performance [63,64] due to deconjugation of bile salts [65]. Therefore, lower counts of Lactobacillus sp./Enterococcus spp. determined in the current study may explain that there were no adverse effects of applied synbiotics on broiler chicken performance, particularly FCE.
Gene expression. Modulation of the intestinal microflora has indirect impact on immune system. In this study we demonstrated that S1 triggered up-regulation of immune-related genes in spleen and their down-regulation in cecal tonsils. Our previous in vivo experiments proved that synbiotics applied in ovo significantly regulated immune-related genes (IL-1β, IL-12p40,  in spleen and cecal tonsils of adult broiler chickens [19,20]. It was shown that the level of gene expression is time-and tissue-dependent [19,20]. Slawinska et al. determined similar tendency of up-regulation of the immune-related gene expression in spleen followed by clear down-regulation in cecal tonsils after in ovo injection of synbiotics [19]. This opposite tendency might be related to different level of exposure both tissues to luminal antigens [20]. Cecal tonsils are part of GALT, thus they are in the close proximity to the intestinal microflora and continuously exposed to microbe-associated molecular patterns (MAMPs). GALT is specialized in anti-inflammatory mechanisms used to eliminate or tolerate microbiota [66]. This mechanism controls the responses of host and develops tolerance towards pathogens, which leads to recognition of commensal bacteria and activation transient and non-inflammatory immune response [67]. In ovo administration of synbiotics ensures an early contact between GALT and beneficial bacteria, which promotes development of tolerance mechanisms [19]. This phenomenon explains down-regulation of the immune-related genes in cecal tonsils, that we determined in this study. S1 and S2 synbiotics delivered in ovo exerted a different gene expression pattern of immune-related genes. This difference might be explained by the various composition of both bioactive compounds (S1 and S2). Prebiotic compound in S1 was defined as synergistic to probiotic, while S2 was defined as synergistic to the host. The other explanation could be as follows: bioactive compounds (e.g. prebiotics or probiotics) differ in their affinity to membrane receptors, and as such, they trigger different level of downstream responses in the host cells. In this study, S1 (containing L. salivarius IBB3154 and GOS) and S2 (containing L. plantarum IBB3036 and RFO) expressed different immunomodulatory effects in chickens as measured by mRNA level of immune-related genes. To assess the specific properties of lactobacilli to stimulate immune responses, an in vitro test had been carried out using chicken macrophage-like HD11 cell line. It confirmed higher stimulatory potential of Lb. salivarius IBB3154 (~3000 of DEG) vs. L. plantarum IBB3036 (~500 DEG) in activating gene expression upon direct interaction between macrophages and bacteria (A. Slawinska, personal communication). Synergistic effects of different combinations of prebiotics and probiotics were also evaluated in vitro using DT40 cell line. Results from this study indicated that the same bacteria strain combined with a different prebiotic triggered various levels of immune-related gene expression in the host cells [68]. This confirms, that not only a more potent probiotic is needed to express a proper immunomodulatory effect on the host, but it also should be combined with prebiotic.

Conclusions
The expected role of microflora-stimulating bioactive compounds, such as prebiotics, probiotics or synbiotics is to modulate beneficial changes in intestinal microbiota of the host. In poultry, those bioactive compounds can be supplemented in-feed/in-water or injected in ovo. Biological effects of synbiotics, irrespective of the route of delivery, depend solely on careful selection of bioactive compounds (prebiotic and probiotic). In this study, we presented a workflow associated with in vitro selection of the synbiotics and its consequences for the downstream animal study, including abundance of intestinal microbial communities, performance parameters, and molecular responses of the immune system. Synbiotic selection in vitro provides some indication of their biological potential in vivo. Both synbiotics had beneficial effects on the overall status of the organisms defined by low mortality and high production parameters. However, we showed that synbiotic composition affected gene expression in cecal tonsils and the spleen, as well as microflora of the GIT. Out of the two synbiotics tested in this experiment, S1 composed of L. salivarius and GOS appeared to be more potent in establishing a down-regulatory pattern in the immune-related gene expression in GALT and a beneficial shift in the microbiota composition in the GIT.