Interaction between early in ovo stimulation of the gut microbiota and chicken host – splenic changes in gene expression and methylation

Background Epigenetic regulation of the gene expression results from interaction between the external environment and transcription of the genetic information encoded in DNA. Methylated CpG regions within the gene promoters lead to silencing of the gene expression in most cases. Factors contributing to epigenetic regulation include intestinal microbiota, which in chicken can be potently modified by in ovo stimulation. The main aim of this study was to determine global and specific methylation patterns of the spleen under the influence of host-microbiome interaction. Results Fertilized eggs of two genotypes: Ross 308 and Green-legged Partridgelike were in ovo stimulated on d 12 of incubation. The injected compounds were as follows: probiotic – Lactococcus lactis subsp. cremoris IBB477, prebiotic – galactooligosaccharides, and synbiotic – combination of both. Chickens were sacrificed on d 42 post-hatching. Spleen was collected, RNA and DNA were isolated and intended to gene expression, gene methylation and global methylation analysis. We have proved that negative regulation of gene expression after administration of bioactive substances in ovo might have epigenetic character. Epigenetic changes depend on the genotype and the substance administered in ovo. Conclusion Epigenetic nature of microbial reprogramming in poultry and extension of issues related to host-microbiome interaction is a new direction of this research.


Introduction
Epigenetic regulation of the gene expression is an interaction of the external environment with the genetic information. They are potentially heritable changes in the gene expression which does not involve alteration in DNA sequence. Changes in the gene expression may be triggered by factors such as: nutrition, health, climate or stress 1 . Epigenetic mechanisms responding to these environmental factors include expression of microRNA (miRNA) sequences, histone modi cation and DNA methylation. The methylation process might be in uenced by many components present in the chicken diet such as probiotic bacteria 2 and also, the prebiotic fermentation product: the short-chain fatty acid, called butyrate 3,4 . These components mainly modulate the microbiota of the gastrointestinal tract (GIT), which affects the host organism by regulating its immune response, metabolism, digestive processes, or nutrient absorption 5 . Supplementation by bioactive substances can directly modulate composition of the host microbiota, which indirectly acts on the whole host organism 6 . Modulation effects depend on the time and mode of administration of the bioactive substance 7 . In chickens prebiotics, probiotics or synbiotics are routinely delivered in-feed and in-water directly after hatching 6 . Supplementation with bioactive substances is continued for another two weeks. During this period GIT is colonized by bene cial bacteria 8 . In poultry, an alternative route of bioactive substances administration is provided by in ovo technology. It is based on single dose of prebiotics, probiotics or synbiotics injected on the day 12 of embryonic development into the air chamber of the egg [9][10][11][12] . Such a route of delivery ensures contact of GIT with bioactive substances as early as possible. In the previous studies we proved e ciency of in ovo technology for reprogramming the chicken microbiome at an early stage of embryonic development by administering prebiotics (e.g. inulin, GOS, RFO) or synbiotics (e.g. inulin + Lactococcus lactis subsp. lactis, GOS + Lactobacillus salivarius, RFO + Lactobacillus plantarum) [10][11][12] .
DNA methylation involves the addition of methyl residues to the cysteine, present within the CpG island, what blocks the enzymes accession (including transcriptases) thereby inhibiting the DNA transcription. DNA methylation is characterized by the stability of cytosine modi cations within the CpG dinucleotides 13 . Remodeling of DNA methylation in genome can be global or locus-speci c 14 . During embryogenesis, global demethylation of DNA occurs, and a de novo methylation pro le appears at the blastocyst stage 15 . Therefore, the chicken environment, egg composition and conditions of fertilized eggs incubation might signi cantly in uence an embryo DNA methylation. DNA requires donors of methyl groups and cofactors, originating from the outside environment (e.g. food). Studies have repeatedly con rmed DNA methylation as a conservative epigenetic mechanism in organism development 16,17 .
However, little is still known about tissue-speci c DNA methylation patterns.
It has been generally recognized that the overall pattern of the DNA methylation is conserved in many species 18,19 . The knowledge about these patterns and the epigenetic regulation of gene expression in poultry is still scarce. We hypothesize that in chickens stimulated with bioactive substances in ovo on the day 12 of embryonic development, an epigenetic mechanism of the gene expression was developed.
Therefore, the main aim of this study was to determine global and speci c methylation patterns for the spleen as a main immune organ under the in uence of host-microbiome interaction. We carried out this analysis in a few steps: (1) meta-analysis of the whole transcriptome data of immune organs, to select candidate genes for DNA methylation analysis; (2) estimation of the global methylation level in spleen stimulated in ovo; (3) estimation of the expression and methylation level of candidate genes CpG islands' in spleen. The innovativeness of the study is expressed by acquiring new knowledge about the mechanisms of gene silencing, where the gene expression regulation resulted from the stimulation of chicken microbiome by bioactive substances injected in ovo.

Meta-analysis based on published microarray data and gene selection
The meta-analysis of high-throughput transcriptome data (expression microarrays) allowed the identi cation of down -regulated genes in spleen. The gene expression was silenced after synbiotics and prebiotics administration in ovo. The comparison of down regulated genes is presented in Venn diagram (Fig. 1).
Analysis showed 108 down-regulated genes in spleen for ve analyzed experimental groups. The selected gene sequences and their connection are presented in Figure 2. A connection was shown between 6 genes. These gene were intended for further analysis using the qMSP reaction.

Global methylation
The results of global methylation for spleen after in ovo stimulation with prebiotic, probiotic and synbiotic in ovo are presented in the Table 1. There are statistically signi cant differences between the analyzed genotypes (GP and Ross) and between the probiotic and prebiotic/synbiotic injected groups in Ross genotype. Gene expression analysis showed a statistically signi cant decrease in the expression of 11 selected genes after the administration of a prebiotic and a synbiotic in Ross. The administration of the probiotic increased gene expression in Ross. There was a signi cant increase in: CYR61 expression after administration of the prebiotic; in CXCR5, NFATC1, CYR61, IKAZ1, ANGPTL4 after administration of synbiotic, and an increase in NR4A3 after administration of all substances in GP genotype. There was a signi cant decrease in the expression of the CD72 gene after the administration of the probiotic. The results are presented in the Table 2. DNA methylation analysis shows statistically signi cant changes in the level of methylation in the experimental groups relative to control. The differences were detected for 6 genes: SYK, ANGPTL4, TNFRSF14, IKZF1, NR4A3 and NFATC1. A statistically signi cant increase in the methylation level in SYK gene equals: 2% in C group, 24% in SYN, 26% in PRE and 28% in PRO group in GP genotype. The methylation level of ANGPTL4 was between 60 and 64% in Ross and 61% in C, 67% in PRO, 72.5% in PRE and 68% in SYN of GP. The methylation level of TNFRSF14 gene, in the spleen of Ross increased from 69% (C) to 83% after PRE administration. On the contrary, in GL genotype methylation levels decreased from 81% in C group to 72% in PRO and 68% in PRE group. IKZF1 methylation decreases in both genotypes after PRO administration by approximately 20%. However, after administration of the PRE in Ross, IKZF1 methylation level increases from 50% to 87%. The administration of bioactive substances signi cantly increased the methylation of the NR4A3 from 15% in the C group in the spleen of Ross, to 47% in PRO and SYN groups and 67% in PRE group. The similar picture is observed for the NFATC gene. Its methylation increased from 2% in C group to 39% in PRE group. The results are presented in the Figure  3.

Discussion
This research aims to verify the epigenetic character of spleen gene expression stimulated by the administration of bioactive substances (prebiotic, probiotic, synbiotic) during embryo development on day 12 of egg incubation in two different chicken genotypes.
Meta-analysis based on published microarray data and gene selection The effects of the in ovo administration of bioactive substances on broiler chickens were already proved and published elsewhere 11 . A signi cant, long-term effects of in ovo delivery of bioactive substances were also determined at the molecular level, using transcriptomic approach 10  Hereby, we present the rst and comprehensive study of gene expression and methylation levels after administration of individual components of the synbiotic, as well as the synbiotic itself, in two extremely different genotypes of chickens. Earlier studies on which the meta-analysis was based concerned only transcriptome changes after administration of prebiotic and synbiotic in chicken broilers 10,12 .
Some differentially methylated genes have been relatively hypomethylated, suggesting that administration of bioactive substances may be associated with lower level or reduction of methylation, which consequently leads to changes in gene expression. However, the effect of methylation on gene expression also depends on many factors, such as the location of CpG, which does not allow a clear conclusion to be made regarding the increase or decrease of gene expression 26 . Little is known about the spleen DNA methylome and its potential cause-and-effect role in shaping the immune response. Due to the fact that methylation is tissue-dependent, it was necessary to select the tissue which plays a key role in the context of poultry production. It is unknown if transcriptional silencing of gene expression is associated with epigenetic mechanisms such as DNA methylation or histone acetylation . Identifying mechanisms that cause gene silencing would be key to understanding the molecular basis of environmental impact 27 .
Literature provides the evidence that microbiome has a signi cant impact on the regulation of epigenetic mechanisms in mammals 28 . It has also been shown that there is a change in the microbiological pro le in chicken intestines due to the in ovo administration of the GOS prebiotic 29 and the synbiotic consisting of GOS and lactic acid bacteria 11 . In this study, GOS was used as a prebiotic, lactic acid bacteria as a probiotic, and their combination in the form of a synbiotic, proving that their in ovo administration cause transcriptomic and epigenetic changes at the level of DNA methylation of genes in spleen. Despite the well-known intestinal microbiota and the growing knowledge of epigenetic regulation such as methylation, there are only few studies combining both issues 30,31 .

Genotype-dependent methylation
Global methylation analysis showed differences between two distinct chicken genotypes stimulated in ovo at the early stage of embryo development. The analyzes were based on two genotypes of different origin and selection history -the broiler chicken line and the native breed of dual purpose hens. The experimental conditions were exactly the same for both genotypes. Broiler chicken -Ross 308 is a meattype which was created as a result of the ongoing genetic intensive selection program. It is characterized by high resistance to diseases, an excellent pace of weight gain, excellent production parameters, namely the growth rate and feed e ciency in large-scale production. Breeding of chicken broilers usually take place in commercial hatcheries. The post -hatching process due to chicken's vaccination and transport cause so called hatching window, i.e. a gap in the access to food and water. At this time, the possibility of inoculation of the microbiota is reduced, which results in microbiota development disorders 32 . Greenlegged Partridgelike (GP) is a dual purpose Polish native breed. It is characterized by high disease resistance and low environmental and food requirements 33 . The GP was included in the genetic resources conservation program in a state where no selection was carried out. This may result in differentiation in response to the stimulation of the intestinal microbiota directly, and indirectly of the immune system, by external factors. Earlier research shows that in ovo administration of a synbiotic affects the development of the immune organs: bursa of Fabricius (in broiler chickens) and spleen (in dual purpose chickens). Improvement of GP spleen development by in ovo synbiotics improves the sensitivity of the immune system to immunomodulating environmental factors 34 . These two genotypes (Ross and GP) are characterized by a diverse gene methylation level as a response to bioactive substances given during embryo development. It can be assumed that the difference is due to the genotype. GP is more resistant and more environmentally adapted, therefore environmental factors do not show such strong effects as in chicken broilers. GP native chickens and Ross broilers were the subject of the analysis in our previous studies. We have determined the morphological differences and immune responses to environmental antigens -lipoteichoic acid (LTA) and lipopolysaccharide (LPS) in broilers (Ross) and native chickens (GP) stimulated in ovo with bioactive compounds (prebiotic, probiotic, and synbiotic) 35 . In ovo stimulation enhanced colonization of the immune organs with lymphocytes in broilers to higher extent than in native chickens. It might be speculated that there was not much potential for further stimulation in GP. However, the signi cant changes in the immune system morphology expressed by the increased number of germinal centers (GC) in spleen was determined in native chickens (but not in broilers) 35 . The number of GC in spleen indicates the activation of humoral immunity in animals by T-dependent antigens. Immune responses triggered in chickens stimulated in ovo are breed-dependent. The splenic (systemic) immune responses of broilers and native chickens challenged with LTA and LPS also showed distinct patterns. Genotype in uenced gene expression signatures of all immune-related genes analyzed in spleen of broilers and native chickens (P < 0.001) (Slawinska -personal communication). The higher number of cytokines was up-regulated in broiler chickens in comparison to native chickens. We conclude that the GP has more potent immune system than Ross, assessed by higher proportion of immune cells in spleen. GP is also less sensitive to environmental changes such as external stimulation of the microbiota, which can be indirectly observed in gene methylation levels.

Substance-dependent methylation
Analysis of the global methylation in spleen showed statistically signi cant differences between the substances administered in ovo in Ross broiler chicken. The obtained results indicate that the response after administration of the probiotic is similar to the control. However, the prebiotic and synbiotic differ signi cantly from the probiotic, with no differences between the prebiotic and the synbiotic. It could be speculated that the administration of an exogenous dose of bacteria into the egg does not constitute such a strong environmental signal as the administration of a prebiotic or synbiotic. It can also be assumed that in the case of synbiotic, the prebiotic component plays a key role in modulation of methylation and expression pro les. Analysis of the expression of single genes con rmed statistically signi cant negative regulation of all analyzed genes after in ovo administration of the prebiotic and synbiotic in Ross. In most analyzed genes, negative regulation of gene expression can be related to methylation level of the gene. Our analysis showed that in ovo administration of various substances differentiated the level of gene methylation. The perinatal period is crucial in the reprogramming of the microbiota, enabling colonization of the gastrointestinal tract of the embryo with bene cial bacteria before hatching 36 . In ovo stimulation performed on the day 12 of egg incubation assumes the administration of bioactive substances during embryonic development and stimulation of the native intestinal microbiota of the embryo before hatching. The prebiotic can penetrate the subcutaneous membrane and penetrate the embryonic circulatory system; while the probiotic becomes available during hatching, when the membrane is broken 32 . GOS effect on the modulation of gene expression after in ovo administration to chicken broilers has been extensively described 10,29 . Relative analysis of the number of bacteria in the intestinal contents after GOS administration showed that its effect depends on the segment of the intestine. It mainly affects the number of Bi dobacterium spp. and Lactobacillus spp. 29 .
In summary, negative regulation of the gene expression after administration of bioactive substances in ovo on day 12 of egg incubation of broiler chicken and native polish chicken may have epigenetic character. Epigenetic mechanisms depend on the genotype being analyzed and the substance administered in ovo. Epigenetic nature of this research is a new direction of microbial reprogramming in poultry and extension of issues related to host-microbiome interaction. This study indicates that there is potential in bioactive substances administered in ovo to target silencing gene expression in spleen that is behind DNA methylation.

Experimental outline
In the rst stage of the study, (1) a meta-analysis was performed on the basis of whole-transcriptome microarray data generated in previous experiments 10,12 in order to select immune-related genes that were silenced. Bioactive compounds used in those experiments for in ovo delivery were two prebiotics (P1 -GOS; P2 -inulin) and synbiotics ( S1 -Lactococcus lactis subsp. lactis + inulin; S2 -Lactobacillus salivarius + GOS, S3 -Lactococcus lactis subsp. cremoris + GOS in ovo to broiler chickens. The present experiment follows the same route of the substance delivery and focuses on single synbiotic the synbiotic and its individual components: prebiotic (GOS) and probiotic (Lactococcus lactis subsp. cremoris). These three bioactive substances were in ovo administered into chicken embryos of two contrasting chicken genotypes: Ross 308 and Green-legged partridgelike.
Based on spleen isolated from in ovo stimulated chickens a following analysis were performed: (2) global methylation to verify the epigenetic nature of the changes, and then (3) gene expression analysis (to con rm expression regulation under the in uence of various substances) and (4) methylation analysis of individual selected genes.

Meta-analysis of published microarray data for gene selection
Gene selection for methylation analysis was based on two sets of microarray data. These data sets contained broiler chicken transcripts generated from individuals which received prebiotic and synbiotic in ovo on day 12 of egg incubation 10,12 . Both projects carried out whole-genome microarray analyzes (Affymetrix, Santa Clara, US), based on genetic material isolated from spleen in experiment with in ovo injection of S1, S2 and S3. Meta-analysis aimed to select genes which were silenced at mRNA level in spleen. The analysis was carried out based on gene lists generated by Affymetrix Expression Console software. In silico selection of gene sequences was based on the following criteria: p-value (p<0.05) and fold change (down-regulation; FC< -1.0). Subsequently, selected gene groups were compared with each other based on the Venn diagrams. Connection of the selected genes was analyzed with STRING 37 Table 3. Relative gene expression analysis was conducted separately for each experimental group by the ΔΔCt method 41 using ACTB 42 and G6PDH 43 as reference genes. Geometric means of cycle threshold (Ct) values of reference genes were used in the analysis 44 . 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). Table 3. Primer sequences used in the RT-qPCR reaction modi cation, unmethylated cytosine undergoes deamination and uracil is formed. In contrast, 5methylcytosine (which is a methylation product) is resistant to this modi cation, making it possible to make it visible in the qPCR reaction with the help of primers speci c for methylated and unmethylated DNA. The mechanism of methylation concerns only cytosines that are part of the cytidine-phosphateguanosine (CpG) dinucleotide sequence. Therefore the primers for qMSP reactions were designed within CpG islands. Based on the selected gene list in meta-analysis, gene primers in two variants were designed: methylated and unmethylated (   Analysis of the relationship between proteins encoded by down-regulated genes selected based on microarray data.