static-curis.ku.dk · Københavns Universitet Early microbial colonization affects DNA methylation of genes related to intestinal immunity and metabolism in preterm pigs

u n i ve r s i t y o f co pe n h ag e n

Koslashbenhavns Universitet

Early microbial colonization affects DNA methylation of genes related to intestinalimmunity and metabolism in preterm pigsPan Xiaoyu Gong Desheng Nguyen Duc Ninh Zhang Xinxin Hu Qi Lu HanlinFredholm Merete Sangild Per T Gao FeiPublished inD N A Research

DOI101093dnaresdsy001

Publication date2018

Document VersionPublishers PDF also known as Version of record

Citation for published version (APA)Pan X Gong D Nguyen D N Zhang X Hu Q Lu H Gao F (2018) Early microbial colonizationaffects DNA methylation of genes related to intestinal immunity and metabolism in preterm pigs D N AResearch 25(3) 287ndash296 httpsdoiorg101093dnaresdsy001

Download date 06 aug 2019

Full Paper

Early microbial colonization affects DNA

methylation of genes related to intestinal

immunity and metabolism in preterm pigs

Xiaoyu Pan1dagger Desheng Gong2dagger Duc Ninh Nguyen1 Xinxin Zhang2

Qi Hu2 Hanlin Lu2 Merete Fredholm3 Per T Sangild1 and Fei Gao2

1Comparative Pediatrics and Nutrition Department of Veterinary and Animal Sciences Faculty of Health and Medical

Sciences University of Copenhagen Frederiksberg DK 1870 C Denmark 2Genome Analysis Laboratory of the

Ministry of Agriculture Agricultural Genomics Institute at Shenzhen Chinese Academy of Agricultural Sciences

Shenzhen 518120 China and 3Animal Genetics Bioinformatics and Breeding Department of Veterinary and Animal

Sciences Faculty of Health and Medical Sciences University of Copenhagen Frederiksberg DK 1870 C Denmark

To whom correspondence should be addressed Tel thorn45 35332698 Email ptssundkudk (PTS) TelFaxthorn86 755

23251432 Email flys828gmailcom (FG)daggerThe authors wish it to be known that in their opinion the first two authors should be regarded as joint first authors

Edited by Prof Takashi Ito

Received 13 July 2017 Editorial decision 3 January 2018 Accepted 8 January 2018

Abstract

Epigenetic regulation may play an important role in mediating microbendashhost interactions and

adaptation of intestinal gene expression to bacterial colonization just after birth This is particu-

larly important after preterm birth because the immature intestine is hypersensitive to invading

bacteria We compared the intestinal DNA methylome and microbiome between conventional

(CON) and antibiotics-treated (AB) preterm pigs used as a model for preterm infants Oral AB

treatment reduced bacterial density (100-fold) diversity and fermentation improved the resist-

ance to necrotizing enterocolitis (NEC) and changed the genome-wide DNA methylation in the

distal small intestine Integration of epigenome data with previously obtained proteome data

showed that intestinal immunendashmetabolic pathways were affected by the AB-induced delay in

bacterial colonization DNA methylation and expression of intestinal genes related to innate

immune response phagocytosis endothelial homeostasis and tissue metabolism (eg CPN1

C3 LBP HIF1A MicroRNA-126 PTPRE) differed between AB and CON pigs even before any

evidence of NEC lesions Our findings document that the newborn immature intestine is influ-

enced by bacterial colonization via DNA methylation changes Microbiota-dependent epigenetic

programming of genes related to gut immunity vascular integrity and metabolism may be crit-

ical for short- and long-term intestinal health in preterm neonates

Key words epigenetics DNA methylation microbiota immunity metabolism

VC The Author(s) 2018 Published by Oxford University Press on behalf of Kazusa DNA Research Institute

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (httpcreativecommonsorglicensesby-nc40)

which permits non-commercial re-use distribution and reproduction in any medium provided the original work is properly cited For commercial re-use please contact

journalspermissionsoupcom 287

DNA Research 2018 25(3) 287ndash296

doi 101093dnaresdsy001

Advance Access Publication Date 19 January 2018

Full Paper

nloaded from httpsacadem

icoupcomdnaresearcharticle-abstract2532874818260 by guest on 14 Septem

ber 2018

1 Introduction

Development of balanced hostndashmicrobe interactions is critical foradaptation of newborn infants to their external environment1

Newborn infants are normally protected from pathogens but allowexposure to a high density of beneficial (commensal) microbes thatfacilitate normal development of the immune system in accord withthe ldquohygiene hypothesisrdquo2 It is not clear how bacterial colonizationaffects microbendashhost interactions in preterm infants that have animmature innate immune system and limited access to protectionfrom motherrsquos own milk just after birth3 Up to 10 of all hospital-ized preterm infants suffer from necrotizing enterocolitis (NEC) anintestinal inflammatory disease associated with formula feeding andgut microbiota dysbiosis4 To prevent NEC different interventionsto manipulate bacterial colonization have been used including anti-biotics (AB) probiotics and prebiotics Delaying bacterial coloniza-tion via oral administration of AB for the first few days after pretermbirth decreases NEC incidence in both infants and pigs5ndash8 ThisNEC-preventive treatment is not used in clinical practice mainly dueto concerns of increased antimicrobial resistance Nevertheless it isimportant to investigate the molecular mechanisms responsible forthe enhanced adaptation of the immature intestine when gut bacte-rial colonization is delayed by AB treatment just after preterm birthThis is critical to identify intestinal pathways that are affected bybacterial colonization and new ways to prevent microbe-inducedintestinal dysfunction and NEC

Recent research has documented direct connections betweenmicrobial metabolites and epigenetic modifications910 For instancemicrobiota-derived folate choline and methionine are involved inthe one-carbon metabolism that contributes to the methyl donorS-adenosylmethionine for DNA methylation DNA methylation mayregulate gene transcription cellular differentiation and normal devel-opment11 Therefore DNA methylation may play an important rolein mediating microbendashhost interactions in early life when rapidadaptation of the immature intestine is critical for survival especiallyfollowing preterm birth In newborn mice germ-free conditions alterthe DNA methylation of the colonic epithelium and affect geneactivation and thereby intestinal functions12 Correspondingly theintestinal DNA methylome may be affected by the timing densityand diversity of bacterial colonization just after birth in pretermneonates

Using preterm pigs as a model for preterm infants we appliedreduced representation bisulphite sequencing (RRBS) for intestinalDNA methylome profiling and 16S rRNA sequencing for micro-biome profiling in animals with or without oral AB treatment for5 days after birth We observed marked DNA methylation differen-ces between the two groups of pigs associated with differences in thetotal load and community structure of bacteria The epigenome datawere integrated with proteome data from a previous study using thesame animal model to gain further insight into the biological path-ways that shape the interface between the gut microbiota and intesti-nal cells

2 Materials and methods

21 Animals and their treatment

All animal procedures were approved by the Danish NationalCommittee on Animal Experimentation Fourteen preterm pigs fromthree sows (Danish LandraceLarge WhiteDuroc) were selectedfrom a previous larger study where the phenotypic characteristicshave been described in detail8 All pigs were caesarean-delivered at

day 106 (90 of gestation) Preterm pigs delivered at 90 gesta-tion show intestinal characteristics of human infants delivered at70 gestation Fourteen pigs showing limited or no severe clinicalsigns of NEC before tissue collection were chosen for this study onDNA methylation patterns The newborn preterm pigs were immedi-ately transferred to a piglet intensive care unit and reared in tempera-ture- and oxygen-regulated incubators The pigs were weighed andfitted with umbilical arterial and orogastric catheters and then sub-ject to passive immunization with maternal plasma as previouslydescribed8 Pigs were subjected to delayed bacterial colonizationusing daily oral boluses of broad-spectrum AB (AB group nfrac147)and they were compared with pigs raised conventionally (CONgroup nfrac147) with spontaneous microbial colonization and dailyboluses of saline For the AB group the selection of AB and doseswere adapted from the current use of AB (intravenous) to preterminfants in Denmark78 Thus a combination of ampicillin (30 mgkgBW thrice daily) gentamicin (25 mgkg BW twice daily) and metro-nidazole (10 mgkg BW thrice daily) was used to target a broad rangeof microorganisms

For nutrition all pigs were initially provided with parenteralnutrition (PN) via the umbilical catheter supplemented with minimalenteral nutrition (MEN) with formula via the orogastric tubeThe PN solution was based on a commercially-available product(Kabiven Fresenius Kabi Bad Homburg Germany) and its compo-sition was adjusted to meet nutritional requirements of preterm pigsas previously described7 The formula diet consisted of three prod-ucts commonly used for infants (per litre of water 75 g LiquigenMCT 80 g Pepdite and 70 g Arla DI-9224 from Nutricia AlleroslashdDenmark and Arla Food Ingredients Viby J Denmark respec-tively) The provision of MEN was initiated within 5 h of delivery asboluses of 3 mlkg every 3 h on days 1ndash2 On day 3 PN supply wasstopped and total enteral nutrition with formula was provided asboluses of 15 mlkg every 3 h until euthanasia and tissue collectionon day 5

22 Tissue collection intestinal morphology and

inflammatory cytokine analysis

The pigs were euthanized with sodium pentobarbital (200 mgkgia) The gastrointestinal tract was immediately removed and thesmall intestine was carefully emptied of its contents and weighedPieces of the distal small intestine (83 along the length of the smallintestine) were snap frozen in liquid nitrogen and kept at 80C forsubsequent analysis of the DNA methylome and inflammatory cyto-kines IL-8 levels in the distal small intestine were analysed withELISA (RampD Systems Abingdon Oxfordshire UK) following themanufacturerrsquos instructions and measured as picograms per milli-gram of wet tissue In addition two 1-cm pieces were collectedand fixed in 4 paraformaldehyde for histological analysis Theparaformaldehyde-fixed samples were embedded in paraffin sec-tioned (5 lm) mounted on slides and stained with haematoxylin andeosin Representative cross-sections were selected from each pig andat least 10 well-oriented crypts and villi were measured using anAxiophot microscope (Carl Zeiss Oberkochen Germany) and NIHimage software version 160 (softWoRx Explorer version 11Applied Precision Issaquah WA) To estimate the epithelial cell pro-portion in the small intestine tissue from each group we examinedcross sections of the distal small intestine by immunohistochemistryusing the epithelial cell marker cytokeratin (DAKO M3515)Staining was developed with UltraVision LP Detection System(ThermoFisher Scientific) The sections were counterstained with

288 DNA methylation in microbendashhost interaction

ber 2018

Mayerrsquos haematoxylin Images were acquired using the OlyVIA soft-ware (OLYMPUS version 29) and the proportion () of the posi-tive staining for cytokeratin in cross-sectional area was analysed bythe IHC toolbox in ImageJ

23 Organic acids and microbiota analysis

The luminal contents of the colon were subjected to organic acid(short-chain fatty acid SCFA) concentration measurements usinggas chromatography and total bacterial load was quantified usingRT-qPCR as previously described13 The luminal contents of the dis-tal small intestine were collected for the analysis of gut microbiotaIn brief total DNA was extracted from the distal small intestinalcontents and the v3-v4 hypervariable regions of the 16S rRNAsequence were amplified with PCR The resultant amplicons weresequenced using the Illumina MiSeq system (Illumina San DiegoCA) producing paired-end reads The microbiota community struc-ture was analysed with Mothur based on the official protocol devel-oped by the Mothur developers14 The optimized reads wereclustered into operational taxonomic units (OTUs) and representa-tive sequences were aligned to the SILVA reference (versionSILVA123) The OTU annotation results were employed to deter-mine the microorganism composition in each sample

24 Reduced representation bisulphite sequencing

(RRBS)

Genomic DNA was extracted from the 14 rinsed intact pieces of dis-tal intestine using the DNeasy Blood amp Tissue Kit (Qiagen) and sub-jected to RRBS library preparation as previously described15 Weused whole intestinal tissue samples because whole tissue responseslikely better represented the in vivo state of the intestine with all theinteracting cell types than just isolated cell types (eg enterocytesgoblet cells enteroendocrine cells) In brief 15mg of genomic DNAwas digested with the MspI enzyme (NEB) followed by end repairA-base tailing and 5-methylcytosine-modified adapter ligation Sizeselection was performed to obtain DNA fractions of MspI-digestedproducts in the range of 40ndash250 bp Subsequently bisulphite treat-ment was conducted using the ZYMO EZ DNA Methylation-GoldKit following the manufacturerrsquos instructions Twelve cycles of PCRwere performed to enrich the DNA fragments in which each librarywas integrated with the DNA index The libraries were analysedusing an Agilent 2100 Bioanalyzer and qPCR for quality controlThe libraries were then subjected to paired-end 125 bp multiplexsequencing on the HiSeq 2500 platform Raw sequencing data wereprocessed via the Illumina base-calling pipeline Low-quality readsthat contained more than 30 lsquoNrsquos or showed a low-quality value(quality value lt20) in over 10 of the sequence were omitted fromthe data analysis The bisulphite sequence MAPping program(BSMAP)16 was used for sequence alignment to the Ensembl pigreference genome (Sscrofa102) The methylation levels of individualcytosines were calculated as the ratio of the sequenced depth of theascertained methylated CpG cytosines to the total sequenced depthof individual CpG cytosines

25 Illumina hiseq sequencing-based bisulphite

sequencing PCR (BSP)

Gene-specific DNA methylation was assessed by BSP according to apreviously published method17 IN brief BSP primers were designedusing the online MethPrimer software and listed in SupplementaryTable S1 Genomic DNA (500 ng) was converted using the ZYMO

EZ DNA Methylation-Gold Kit (ZYMO) and one-tenth of the elu-tion products were used as templates for PCR amplification Foreach sample BSP products of multiple genes were generated pooledequally and subjected to adaptor ligation Barcoded libraries from all14 samples were sequenced on the Hiseq platform using paired-end250 bp strategy Data were processed and analysed using BSMAP asdescribed above

26 Statistical analysis

For phenotypic values bacterial load SCFA and cytokine data com-parisons between the two groups were conducted using Studentrsquost-test and a two-tailed P-value lt005 was considered statisticallysignificant The correlation between the log-transformed relativeabundance of the microbiota and concentration of SCFAs wasassessed using Spearmanrsquos rank correlation test A two-sided P-valuelt005 was regarded as statistically significant Differentially methy-lated positions (DMPs) were identified according to the methylationlevels of cytosines between samples from the two groups using theMannndashWhitney U-test DMPs showing a mean methylation differ-ence gt02 and a P-value lt001 were considered statistically signifi-cant To identify differentially methylated regions (DMRs) pairs ofsignificant DMPs were used to delimit regions exhibiting homogene-ous methylation changes and the interval methylation levels betweenthe two groups were tested using the MannndashWhitney U-test with afalse discovery rate of lt005 For gene expression analysis RT-qPCR analysis was performed using QuantiTect SYBR Green PCRKit (Qiagen) on LightCycler 480 (Roche) Relative quantification oftarget genes was normalized to housekeeping gene HPRT1(Supplementary Table S2) Comparisons were made using theStudentrsquos t-test and a two-tailed P-value lt005 was considered asstatistically significant

3 Results

31 Phenotypic effects of AB treatment in preterm pigs

Pigs from the two groups in this study were subgroups of pigs from alarger study where detailed clinical characteristics and phenotypicvariables have been published previously8 At autopsy none of theAB pigs had any NEC lesions in their intestines while two of sevenCON pigs were diagnosed as NEC according to our macroscopicNEC evaluation system18 One of these NEC pigs showed colonlesions (haemorrhage local necrosis pneumatosis intestinalis NECscore 5) while the other NEC pig showed both colon and distal intes-tine lesions (severe extensive pneumatosis intestinalis haemorrhagenecrosis NEC score 6) All the AB pigs were given a NEC score of 1reflecting no visible lesions or abnormalities (Fig 1a) Histologicalanalysis of the distal small intestine revealed clear villus atrophy inthe pig diagnosed as NEC in the distal intestine (Fig 1b) Comparedwith CON pigs the AB pigs showed a higher growth velocityand lower crypt depth (both Plt005) but no change in villusheight (Fig 1c) To assess the approximate proportion of epithelialcells immunohistochemical analyses of the epithelial marker cyto-keratin was performed on cross sections of distal small intestineNo significant difference in the proportion of epithelial cells wasobserved between AB and CON pigs (441 vs 422 Pfrac14038Supplementary Fig S1)

Relative to CON pigs the AB pigs had a lower mean total bacte-rial load (106 versus 108 in bacterial counts Fig 1d) with two of theAB pigs having values similar to that in CON pigs We then analysedthe microbial composition from their intestinal contents using 16S

289X Pan et al

ber 2018

Figure 1 Phenotypic and microbial characteristics of preterm pigs (a) Representative photographs and (b) histopathology of the small intestine in preterm pigs

receiving AB (AB pigs NEC score 1) or CON-raised pigs and with NEC lesions (CON-NEC pigs NEC scores 5ndash6) (c) Growth velocity of preterm pigs and villus

height and crypt depth in the distal intestine (d) Bacterial load and (e) dendrogram illustrating the microbial composition in the distal small intestine of preterm

pigs (f) IL-8 expression in the distal intestine (g) Lactate concentration in the colon Bar charts were presented as the mean values 6 SEM (Plt 005

Plt001 Plt0001 Plt 00001)

ber 2018

rRNA microbiome sequencing Data from two pigs in the AB groupwere omitted due to insufficient reads for OTU annotation Based onthe remaining 12 samples there was lower inter-individual diversityof the intestinal microbiota in AB versus CON pigs The two NECpigs from the CON group was dominated by Clostridium andshowed marked compositional difference from the other pigs asindicated by hierarchical clustering (Fig 1e) The five healthypigs from CON group were dominated by Enterococcus andEnterobacter while most AB pigs were dominated by Enterobacterexcept for one pig that was dominated by Staphylococcus (Fig 1e)Hence in addition to a decreased total bacterial count the AB pigshad less abundance of Gram-positive bacteria (Clostridium andEnterococcus) than the NEC and healthy pigs from the CON groupie CONndashNEC and CONndashHealthy pigs respectively Previous stud-ies show that stool samples from preterm infants that develop NECare dominated by Firmicutes especially Enterococcus in the earlypostnatal period1920 Therefore the dominant Enterococcus inCONndashHealthy pigs might imply a risk of developing NEC althoughnormal intestinal morphology was observed in these pigs at autopsyThe level of interleukin 8 (IL-8) a biomarker of NEC in both humaninfants and rats21ndash23 was elevated in the distal intestine of theseCONndashHealthy pigs (Plt005 Fig 1f) Finally colonic lactate con-centrations known to be elevated in association with NEC18 wasmarkedly increased in the CONndashHealthy pigs (Plt00001 Fig 1g)The lactate concentration was positively correlated with the relativeabundance of Enterococcus (qfrac14061 Plt005)

32 DNA methylation profile in preterm pig intestine

Next we evaluated the DNA methylome of the small intestine in theAB and CON groups via RRBS which was developed to measurethe DNA methylation of high-CG regions at a single base-pair resolu-tion As we applied a 125-bp paired-end sequencing strategy MspI-digested fragments of the RRBS library were expanded to 40ndash250 bpAs a result we generated a total of 5765 gigabases (Gb) of cleanbases from 14 libraries after quality control analyses Using theBSMAP we found that 680 of the clean reads could be mapped tothe pig reference genome reaching an average read depth of 871ndash1359 per strand for each sample The bisulphite conversion rate ofCndashto-T reached 996 as calculated based on the methylated level oflambda DNA (Supplementary Table S3) As RRBS enriches high-CpGregions in the genome and mammalian DNA methylation occursalmost exclusively at CGs we focused on analyses of CpG methyla-tion Only the CpG sites showing 4 or more coverage per strandwere analysed to maintain a high accuracy level for methylation lev-els resulting in an average of 466 million CpG sites being covered inthe analyses for each of the samples The pig genome contains 5606million CpGs and we therefore managed to examine 831 of allCpGs in the pig genome using this representation strategy

DNA methylation might be biased towards specific alleles in spe-cific genomic regions owing to gene imprinting24 To address thisissue we applied two methods to analyse allele-specific methylation(ASM) across 14 samples including an AMR-based method25 andan SNP calling-based method referred to as ldquoSMAPrdquo26 By integrat-ing the results from these two methods we sought to reveal genomicregions that most likely showed differential methylation betweenpaternal and maternal genomes and we checked whether theseregions of genomic imprinting could be affected by AB usageHowever no significant differences in ASM levels were observedbetween the two groups indicating that the administered oral AB

and the different colonization levels had no effect on ASM(Supplementary Fig S2 data not shown)

33 Genome-wide DNA methylation changes

Based on the above analyses we next sought to infer the changes inintestinal genomic methylation in response to delayed microbial col-onization by AB treatment just after preterm birth To avoid thepotentially confounding influence of X chromosome inactivation onDNA methylation patterns between male and female neonates onlyautosomal data were used in this study hereafter We first examinedthe global pattern of genome-wide methylation Hierarchical cluster-ing and principal component analysis (PCA) of the methylation levelsof all CpG sites were performed to examine the whole-genome meth-ylation status of these samples Both clustering and PCA results indi-cated that the two groups were not clearly separated into twoclusters owing to individual epigenomic variation (Fig 2a andSupplementary Fig S3) However based on the average methylationlevels observed across all samples genic regions showed clear diver-gence between the two groups suggesting differences in specificgenomic regions or CpG sites (Supplementary Fig S4) Thereby wenext carried out pair-wise comparisons to screen for DMPs betweenthe two groups across the whole genome This revealed 3850 CpGsites to be DMPs exhibiting methylation level differences of greaterthan 20 (Fig 2b) Among these DMPs 466 were distributedwithin genic regions including gene body regions or putative regula-tory elements of promoters (2 kb upstream and 500 bp downstreamfrom the transcription start sitesmdashTSS)

Based on the identification of DMPs we further screened for keygenome-wide DMRs that can potentially affect gene functions (seeMaterials and methods) A total of 87 DMRs were identified in the ABversus CON groups (Supplementary Table S4) Compared with theCON group 47 DMRs were hypomethylated in the AB group TheseDMRs were distributed across all the autosomes with an averagelength of 47 base pairs Based on the current gene annotation for thepig genome 48 DMRs were associated with putative promoters orintragenic regions (Supplementary Table S4 and Fig 2c) There were39 genes with gene symbols that were ascertainable from the Ensembldatabase Two genes (NNAT and MEST) were recognized as imprintedgenes according to an online imprinting gene database (httpwwwgeneimprintcom (11 January 2018 date last accessed)) and were excludedfrom the subsequent analyses To confirm the observed DNA methyla-tion variations BSP was performed on five genes with DMRs in eitherputative promoter or gene body Examination on all individual cyto-sines within the five genes revealed significant correlation between theRRBS and BSP data (Plt00001 R2 frac14 078 Fig 2d andSupplementary Fig S5) According to hierarchical clustering by individ-ual methylation level of the DMRs in these 37 genes the AB and CONpigs were grouped into two distinct clusters (Fig 2e) Within the CONgroup the two NEC pigs and five healthy pigs were not clearly sepa-rated This suggests that NEC lesions observed on day 5 did not mark-edly affect the intestinal methylation status

34 Integrated methylomendashproteome analyses

We next sought to evaluate the functional significance of the identi-fied DMRs Considering DNA methylation plays an important rolefor transcription regulation it is expected that divergent DNA meth-ylation will ultimately lead to changes of protein expression and fur-ther trigger changes of cellular functions via proteinndashproteininteractions To address this possibility we applied a previously pub-lished proteome data set with a series of differentially expressed

291X Pan et al

ber 2018

proteins following antibiotic treatment of preterm pigs for five daysafter birth27 (Supplementary Table S5) This study used an identicaltreatment protocol except that the AB were provided both orally andsystemically As DNA methylation at TSS is associated with tran-scription silencing while the mechanism of methylation in othergenomic regions is less known we selected genes (ARAP1 CPN1DHCR7 HOXA2 TRMU and ssc-mir-126) containing DMRs adja-cent to TSSs and searched BioGRID database for their potentialinteractions with the divergently expressed proteins obtained fromthe proteome data We found that four genes (ARAP1 CPN1DHCR7 and HOXA2) were involved in extensive proteinndashproteininteraction networks in which 78 proteins showed either direct orindirect interaction with each other (Fig 3a) Among these fourgenes CPN1 showed direct interaction with complement component3 (C3) which is an important constituent of the innate immune sys-tem that enhances phagocytosis to clear microbes As CPN1 func-tions as a suppressor of C328 and was hypomethylated in itspromoter region (Fig 3b) it could be therefore up-regulated leadingto decreased C3 expression in the AB group To test this we exam-ined the mRNA expression levels of both CPN1 and C3 from alarger set of samples from the original study8 including 14 pigs fromthe AB group and 14 pigs from the CON group (of which 8 hadNEC and 6 were healthy) Relative to the CON pigs (both CONndashhealthy and CONndashNEC) the AB pigs showed significantly decreasedC3 level (Fig 3c) However in agreement with the gene expressiondatabase29 the qPCR results showed very low CPN1 expression lev-els in the small intestine preventing detailed comparisons between

groups (Supplementary Fig S6) Nevertheless consistent with thatC3 was reduced in the AB pigs lipopolysaccharide binding protein(LBP) and neutrophil chemotactic factor IL-8 were also reduced inthe AB pigs (Fig 3d) We examined the proportion of neutrophils byimmunohistochemical analyses on cross sections of distal small intes-tine using the neutrophil marker myeloperoxidase (MPO) Thenumber of MPO-positive cells among the overall cell population wassmall and 36 versus 47 per villus in the AB and CON groupsrespectively (Pfrac14008 Supplementary Fig S7) Together with thereduced IL-8 levels this suggests that that innate immune responsewas less activated in the AB pigs

Furthermore we employed the genes that were differentiallyexpressed in protein level together with all the DMR-associatedgenes to perform functional enrichment analysis In agreement withthe above results we found that 2 of the 11 significantly enrichedKEGG pathways (Supplementary Table S6) were related to lsquobacterialinfectionrsquo including lsquolegionellosisrsquo and lsquopathogenic Escherichia coliinfectionrsquo These two pathways were both related to Gram-negativebacteria supporting the change in LBP expression (Fig 3d)Interestingly we also revealed five metabolism-related pathways andfour pathways related to vascular functions Previous studies indi-cated that the recruitment of immune cells and phagocytosis mightresult in local depletion of oxygen which could further trigger astress response that increases angiogenesis and induces metabolicchanges to compensate for oxygen deficits30ndash32 We found thatwithin the glycolysis and gluconeogenesis pathways the glycolyticenzyme aldolase A (encoded by ALDOA) was down-regulated in the

Figure 2 Changes in DNA methylation in response to microbial colonization (a) PCA using genome-wide DNA methylation data (b) Scatter plot of the methyla-

tion levels of DMPs in the CON and AB groups showing the density at each point (c) Genomic distribution of DMRs (d) Correlation between RRBS and BSP

data (e) Heat map depicting the hierarchical clustering of DMR-associated genes for each preterm pig

ber 2018

AB group In contrast the gluconeogenesis regulatory enzymesfructose-1 6-bisphosphatase 1 (encoded by FBP1) and mitochondrialPEP-carboxykinase (encoded by PCK2) were up-regulated in the ABgroup (Supplementary Fig S8) Together the enriched KEGG path-ways indicate that delayed bacterial colonization following AB treat-ment may affect intestinal oxygenation vascular function and tissuemetabolism

35 Hypoxia-associated vascular endothelial functions

Based on above results we further studied in greater detail somegenes related to possible hypoxia and altered vascular functions ofthe immature intestine of AB and CON pigs This is relevant asinnate immunity is linked with hypoxia and hypoxia inducesangiogenesis3334 Under hypoxic conditions the expression ofhypoxia-inducible factor 1-alpha (HIF1A) will be triggered to medi-ate the hypoxia response of cells3334 We therefore quantified themRNA expression of HIF1A in the small intestines Consistent withthis HIF1A expression was reduced in AB versus CON pigs(Fig 4a) Increased HIF1A expression may stimulates vascular endo-thelial growth factor receptor 2 (VEGFR2) signalling in order to trig-ger angiogenesis and compensate for an oxygen deficit31 In thisstudy we did not observe any differences in VEGFA and VEGFR2expression between CON and AB pigs (Fig 4a) Previous studies sug-gested that lack of VEGFR2 signalling facilitated NEC35 This resultmight indicate these CON pigs also failed to activate the VEGFR2signalling pathway under hypoxic condition Another effect of hypo-xia is on the vascular tone pathway Our KEGG analyses had

indicated several DMR-associated genes involved in thevasodilation-related pathways including GNAS and MRVI1(Supplementary Table S6) which belong to the G protein-coupledreceptor (GPCR) signalling pathway GNAS encodes for stimulatoryG-protein alpha subunit (Gs-a) while MRVI1 (also called IRAG)had been shown to prevent calcium release within the GPCR signal-ling and thus contributes to vasodilation3637 RT-qPCR resultsshowed the mRNA expression of GNAS and MRVI1 also tended tobe decreased in the AB pigs versus NEC pigs though not significantly(Supplementary Fig S6)

Finally we investigated in detail three DMR-associated genespotentially related to endothelial homeostasis MicroRNA-126 isessential for vascular integrity and inhibits haemorrhage38 PTPREnegatively regulates endothelial cell proliferation39 and LRP8 ini-tiates endothelial antiapoptosis40 The DMRs in these genes were allhypomethylated in the AB versus CON pigs (Supplementary TableS4) The DMR within the putative promoter region of MicroRNA-126 was located in the intron of EGFL7 and contained four CpGs(Fig 4b) Hypomethylation of this region may increase the transcrip-tion of MicroRNA-126 and there was a tendency to up-regulation ofMicroRNA-126 in the AB pigs (Pfrac14029 Fig 4a) In contrastPTPRE and LRP8 contained the DMRs within their intron and exonrespectively (Fig 4b) The DMR in LRP8 also co-localized with aCpG island The expression of PTPRE was decreased in AB versusCON pigs especially when compared with NEC pigs (Plt005Fig 4a) The expression of LRP8 was too low to be accurately meas-ured in the small intestine in both groups (Supplementary Fig S6) asconfirmed by the gene database29

Figure 3 Methylome-proteome network analyses (a) Visualization of potential interactions between genes associated with DMRs and genes encoding differen-

tially expressed proteins according to the BioGRID interaction data set (b) Individual methylation level of each CpG cytosine within the DMR at CPN1 promoter

(c d) Relative expression of genes in the distal intestine presented as mean values 6 SEM (Plt005 Plt001 Plt0001 Plt00001)

293X Pan et al

ber 2018

4 Discussion

The initial bacterial colonization of gut in early life of infants is crit-ical because the microbiota shapes development of immunity and haseffects on metabolism1 The molecular mechanisms whereby thisoccurs remain obscure but bacterial products such as fermentationmetabolites may affect host cells via epigenetic modifications41

Using preterm pigs as a model for preterm infants we have demon-strated that differences in bacterial colonization of the immaturesmall intestine induce marked changes in gene expression that areregulated by epigenetic mechanisms A delay in bacterial coloniza-tion resulting from a relatively modest AB-induced reduction in bac-terial density in the small intestine over the first 5 days causedchanges to DNA methylation for intestinal genes related to improvedinnate immune response hypoxia-related vascular function and tis-sue metabolism These apparent beneficial effects are targets to helpprevent the immature intestine from detrimental responses to invad-ing bacteria after preterm birth Among the CON-reared pigs theobserved intestinal methylation and gene expression levels weresimilar for pigs with and without NEC lesions indicating that evenmoderate differences in intestinal bacterial colonization and density(eg 106 versus 108 bacteria) not NEC lesions was the main factoraffecting methylation of genes in the immature intestine just afterbirth

Preliminary evidence for a dynamic crosstalk between the intesti-nal methylome and bacterial colonization after birth has beenderived from rodents1242 The preterm pig is the only model of pre-term infants that combines a high sensitivity to intestinal disorders(NEC feeding intolerance) with many other physiological signs ofpreterm birth (eg respiratory dysfunction cardiovascular impair-ments metabolic dysfunction)18 Recent studies have demonstrated ahigh similarity of both DNA methylation patterns and gut micro-biomes between pigs and humans supporting that pigs are highly rel-evant biomedical models for study of human diseases43ndash45 In thisstudy we investigated sections of the distal small intestine because

this intestinal region is most commonly affected by NEC lesionshighly populated with bacteria and critical for immune developmentand bacterial tolerance Maladaptation to feeding and bacterial colo-nization in preterm neonates (eg NEC) involves all cell types andlayers of the small intestine hence we investigated whole tissuerather than isolated cell populations Five days after preterm birthwhen intestinal lesions were relatively mild in control pigs the pro-portion of epithelial cells was similar in the two groups as indicatedby our staining of intestinal cross sections The short-term AB treat-ment did not induce any marked change in the already low propor-tion of neutrophils in the epithelium and in a previous study asimilar AB treatment did not affect the proportion of intestinal gobletcells 7 The observed DNA methylation changes are thereforeunlikely to result from microbiota-dependent changes in the relativecell proportions in the immature intestinal mucosa during the first5 days after birth

Microbial community dysbiosis is an important factor for neona-tal gut health and preterm infants with a gut dominated byFirmicutes in the first days after birth may later develop NEC1920 Inthis study the CON-reared pigs were also dominated by Firmicutes(Enterococcus and Clostridium) on day 5 while all AB-treated pigswere protected from NEC lesions and dominated by EnterobacterCompared with CON pigs LBP IL-8 and C3 were reduced in thedistal intestine of AB-treated pigs LBP is a pattern recognition recep-tor transferring a variety of ligands from both Gram-positive andGram-negative bacteria to the host through toll-like receptors(eg TLR2 and TLR4)46 The chemokine IL-8 can be induced bothby TLR247 and TLR448 and recruits immune cells towards the site ofinfection C3 labels the pathogen to facilitate phagocytosis to clearbacteria including Enterococcus49 Reduced expression of thesegenes indicates less need to mount an innate immune response in theAB-treated pigs

Our results demonstrated clear effects of an AB-induced delay inbacterial colonization on intestinal DNA methylation and expressionof selected genes The integrative methylomendashproteome analysis

Figure 4 Expression of genes related to hypoxia and vascular function (a) The relative expression of HIF1A (related to hypoxia) and VEGFA VEGFR2

ssc-mir-126 and PTPRE (related to vascular function) Values are presented as mean 6 SEM (Plt005 Plt 001 Plt0001 Plt00001) (b) Three genes

(ssc-mir-126 PTPRE LRP8) related to endothelial homeostasis showed hypomethylation in the promoters or gene bodies in the AB pigs

ber 2018

revealed that genes involved in bacterial infection vasodilation-related pathways and metabolic pathways were most affected Theseaffected functions may be closely associated with differential degreeof intestinal hypoxia induced by variable innate immune response toinvading bacteria Normally immune cell recruitment to sites ofinfection and induction of phagocytosis consume excessive oxygenand induce angiogenesis and glycolysis to compensate for the oxygendeficit30ndash32 Accordingly we observed decreases in C3 and HIF1Atranscription in the AB versus CON group The suppressor of C3CPN1 might be more expressed due to promoter hypomethylationin response to AB treatment Similarly MicroRNA-126 expressiontended to be increased in the AB group probably due to promoterhypomethylation In mice and zebra fish loss of MicroRNA-126impacts endothelial cell proliferation and vascular integrity resultingin fragile and leaky vessels5051 On the other hand the intragenicDMR of PTPRE was hypomethylated in the AB group together withreduced expression levels This may be explained by the general posi-tive correlation between gene-body methylation and gene expressionobserved previously52 PTPRE is highly abundant in endothelial cellsand its down-regulation may indicate increased endothelial prolifera-tion in AB versus CON pigs39

Finally our results indicated that AB treatment reduced glycolysisand increased gluconeogenesis-related gene expression in the ABgroup (Supplementary Fig S8) These effects may be a direct conse-quence of the decreased tissue hypoxia and less need for activation ofinnate immune response to combat invading bacteria in the ABgroup Effects on tissue metabolism were also supported by DNAmethylation differences in genes related to metabolism For examplethe putative promoters of DHCR7 and TRMU showed hypomethy-lation in the AB group indicating up-regulation of these two genesDHCR7 encodes for 7-dehydrocholesterol reductase that producescholesterol using NADPH a cofactor used in anabolic reactionsSimilarly TRMU encodes for the mitochondrial tRNA-specific 2-thi-ouridylase 1 that is closely related with mitochondrial function53

Thus the potential up-regulation of DHCR7 and TRMU could beassociated with active energy consumption in the intestine of AB pigswithout hypoxic stress In conclusion a delay in bacterial coloniza-tion by oral AB treatment just after preterm birth may provide lowerinnate immune response less hypoxic stress better vascular integrityand increased metabolism in the immature intestine via epigeneticmechanisms

Acknowledgements

We thank Thomas Thymann Elin Skytte Kristina Moslashller Jane Povlsen andKarina Ryom for their technical support with animal procedures and labora-tory analyses

Funding

This study was supported by the Danish Strategic Research Council[NEOMUNE program 12-132401] the Agricultural Science and TechnologyInnovation Program (ASTIP) and the China Scholarship Council [ScholarshipNo 201406150073 to XP]

Data availability

All RRBS sequencing and processed data were deposited in the GeneExpression Omnibus (GEO) with accession GSE88697 Microbiome datahave been submitted to DNA Data Bank of Japan (Accession numberLC333600 to LC333727)

Conflict of interest

None declared

Supplementary data

Supplementary data are available at DNARES online

References

1 Castanys-Munoz E Martin MJ and Vazquez E 2016 Building a ben-eficial microbiome from birth Adv Nutr 7 323ndash30

2 Rook GA 2012 Hygiene hypothesis and autoimmune diseases Clin

Rev Allerg Immunol 42 5ndash153 Beck S Wojdyla D Say L et al 2010 The worldwide incidence of

preterm birth a systematic review of maternal mortality and morbidityBull World Health Org 88 31ndash8

4 Neu J and Pammi M 2017 Pathogenesis of NEC impact of an alteredintestinal microbiome Semin Perinatol 41 29ndash35

5 Grylack LJ and Scanlon JW 1978 Oral gentamicin therapy in the pre-vention of neonatal necrotizing enterocolitis A controlled double-blindtrial Am J Dis Child 132 1192ndash4

6 Egan EA Mantilla G Nelson RM and Eitzman DV 1976 A pro-spective controlled trial of oral kanamycin in the prevention of neonatalnecrotizing enterocolitis J Pediatr 89 467ndash70

7 Jensen ML Thymann T Cilieborg MS et al 2014 Antibioticsmodulate intestinal immunity and prevent necrotizing enterocolitis in pre-term neonatal piglets Am J Physiol Gastrointest Liver Physiol 306G59ndash71

8 Birck MM Nguyen DN Cilieborg MS et al 2016 Enteral but not

parenteral antibiotics enhance gut function and prevent necrotizingenterocolitis in formula-fed newborn preterm pigs Am J Physiol

Gastrointest Liver Physiol 310 G323ndash339 Paul B Barnes S Demark-Wahnefried W et al 2015 Influences of

diet and the gut microbiome on epigenetic modulation in cancer and other

diseases Clin Epigenet 7 11210 Ye J Wu W Li Y and Li L 2017 Influences of the gut microbiota on

DNA methylation and histone modification Dig Dis Sci 62 1155ndash6411 Reik W 2007 Stability and flexibility of epigenetic gene regulation in

mammalian development Nature 447 425ndash3212 Yu DH Gadkari M Zhou Q et al 2015 Postnatal epigenetic regula-

tion of intestinal stem cells requires DNA methylation and is guided bythe microbiome Genome Biol 16 211

13 Hansen CH Holm TL Krych L et al 2013 Gut microbiotaregulates NKG2D ligand expression on intestinal epithelial cells Eur J

Immunol 43 447ndash5714 Schloss PD Westcott SL Ryabin T et al 2009 Introducing mothur

open-source platform-independent community-supported softwarefor describing and comparing microbial communities Appl Environ

Microbiol 75 7537ndash4115 Gao F Zhang J Jiang P et al 2014 Marked methylation changes in

intestinal genes during the perinatal period of preterm neonates BMC

Genomics 15 71616 Xi Y and Li W 2009 BSMAP whole genome bisulfite sequence

MAPping program BMC Bioinformatics 10 23217 Gao F Liang H Lu H et al 2015 Global analysis of DNA methyla-

tion in hepatocellular carcinoma by a liquid hybridization capture-basedbisulfite sequencing approach Clin Epigenet 7 86

18 Sangild PT Thymann T Schmidt M Stoll B Burrin DG andBuddington RK 2013 Invited review the preterm pig as a model inpediatric gastroenterology J Anim Sci 91 4713ndash29

19 Mai V Young CM Ukhanova M et al 2011 Fecal microbiota in pre-mature infants prior to necrotizing enterocolitis PLoS ONE 6 e20647

20 Morrow AL Lagomarcino AJ Schibler KR et al 2013 Earlymicrobial and metabolomic signatures predict later onset of necrotizing

enterocolitis in preterm infants Microbiome 1 13

295X Pan et al

ber 2018

21 Benkoe TM Mechtler TP Weninger M Pones M Rebhandl Wand Kasper DC 2014 Serum levels of interleukin-8 and gut-associatedbiomarkers in diagnosing necrotizing enterocolitis in preterm infantsJ Pediatr Surg 49 1446ndash51

22 Bergholz R Zschiegner M Eschenburg G et al 2013 Mucosal losswith increased expression of IL-6 IL-8 and COX-2 in a formula-feedingonly neonatal rat model of necrotizing enterocolitis J Pediatr Surg 482301ndash7

23 Neunhoeffer F Jansen H Goelz R et al 2015 Combination of exces-sive weight gain and interleukin-8 a possible predictor of necrotisingenterocolitis in neonates Z Geburtshilfe Neonatol 219 221ndash5

24 Ferguson-Smith AC 2011 Genomic imprinting the emergence of an epi-genetic paradigm Nat Rev Genet 12 565ndash75

25 Fang F Hodges E Molaro A Dean M Hannon GJ and SmithAD 2012 Genomic landscape of human allele-specific DNA methyla-tion Proc Natl Acad Sci USA 109 7332ndash7

26 Gao S Zou D Mao L et al 2015 SMAP a streamlined methylationanalysis pipeline for bisulfite sequencing Gigascience 4 29

27 Jiang P Jensen ML Cilieborg MS et al 2012 Antibiotics increasegut metabolism and antioxidant proteins and decrease acute phaseresponse and necrotizing enterocolitis in preterm neonates PLoS One 7e44929

28 Campbell WD Lazoura E Okada N and Okada H 2002Inactivation of C3a and C5a octapeptides by carboxypeptidase R and car-boxypeptidase N Microbiol Immunol 46 131ndash4

29 Kapushesky M Emam I Holloway E et al 2010 Gene expression atlasat the European bioinformatics institute Nucleic Acids Res 38 D690ndash8

30 Colgan SP and Taylor CT 2010 Hypoxia an alarm signal duringintestinal inflammation Nat Rev Gastroenterol Hepatol 7 281ndash7

31 Krock BL Skuli N and Simon MC 2011 Hypoxia-induced angio-genesis good and evil Genes Cancer 2 1117ndash33

32 Zeitouni NE Chotikatum S von Kockritz-Blickwede M and Naim HY2016 The impact of hypoxia on intestinal epithelial cell functions consequen-ces for invasion by bacterial pathogens Mol Cell Pediatr 3 14

33 Rius J Guma M Schachtrup C et al 2008 NF-kappaB links innateimmunity to the hypoxic response through transcriptional regulation ofHIF-1alpha Nature 453 807ndash11

34 Pugh CW and Ratcliffe PJ 2003 Regulation of angiogenesis by hypo-xia role of the HIF system Nat Med 9 677ndash84

35 Yan X Managlia E Liu SX et al 2016 Lack of VEGFR2 signalingcauses maldevelopment of the intestinal microvasculature and facilitatesnecrotizing enterocolitis in neonatal mice Am J Physiol GastrointestLiver Physiol 310 G716ndash25

36 Dudzinski DM and Michel T 2007 Life history of eNOS partners andpathways Cardiovasc Res 75 247ndash60

37 Schlossmann J and Desch M 2011 IRAG and novel PKG targeting inthe cardiovascular system Am J Physiol Heart Circ Physiol 301H672ndash82

38 Herbert SP and Stainier DY 2011 Molecular control of endothelialcell behaviour during blood vessel morphogenesis Nat Rev Mol CellBiol 12 551ndash64

39 Thompson LJ Jiang J Madamanchi N Runge MS and PattersonC 2001 PTP-epsilon a tyrosine phosphatase expressed in endotheliumnegatively regulates endothelial cell proliferation Am J Physiol HeartCirc Physiol 281 H396ndash403

40 Sinha RK Yang XV Fernandez JA Xu X Mosnier LO andGriffin JH 2016 Apolipoprotein E receptor 2 mediates activated proteinC-induced endothelial Akt activation and endothelial barrier stabilizationArterioscler Thromb Vasc Biol 36 518ndash24

41 Nagy-Szakal D and Kellermayer R 2011 The remarkable capacity forgut microbial and host interactions Gut Microbes 2 178ndash82

42 Cortese R Lu L Yu Y Ruden D and Claud EC 2016Epigenome-microbiome crosstalk a potential new paradigm influencingneonatal susceptibility to disease Epigenetics 11 205ndash15

43 Xiao L Estelle J Kiilerich P et al 2016 A reference gene catalogue ofthe pig gut microbiome Nat Microbiol 1 16161

44 Choi M Lee J Le MT et al 2015 Genome-wide analysis of DNAmethylation in pigs using reduced representation bisulfite sequencingDNA Res 22 343ndash55

45 Schachtschneider KM Madsen O Park C Rund LA GroenenMA and Schook LB 2015 Adult porcine genome-wide DNAmethylation patterns support pigs as a biomedical model BMCGenomics 16 743

46 Schroder NW Heine H Alexander C et al 2004Lipopolysaccharide binding protein binds to triacylated and diacylatedlipopeptides and mediates innate immune responses J Immunol 1732683ndash91

47 Thornton NL Cody MJ and Yost CC 2012 Toll-like receptor12 stimulation induces elevated interleukin-8 secretion in polymorphonu-clear leukocytes isolated from preterm and term newborn infantsNeonatology 101 140ndash6

48 Wheeler DS Chase MA Senft AP Poynter SE Wong HR andPage K 2009 Extracellular Hsp72 an endogenous DAMP is released byvirally infected airway epithelial cells and activates neutrophils viaToll-like receptor (TLR)-4 Respir Res 10 31

49 Leendertse M Willems RJ Flierman R de Vos AF Bonten MJand van der Poll T 2010 The complement system facilitates clearance ofEnterococcus faecium during murine peritonitis J Infect Dis 201544ndash52

50 Fish JE Santoro MM Morton SU et al 2008 miR-126 regulatesangiogenic signaling and vascular integrity Dev Cell 15 272ndash84

51 Wang S Aurora AB Johnson BA et al 2008 The endothelial-specificmicroRNA miR-126 governs vascular integrity and angiogenesis Dev Cell15 261ndash71

52 Ball MP Li JB Gao Y et al 2009 Targeted and genome-scale strat-egies reveal gene-body methylation signatures in human cells NatBiotechnol 27 361ndash8

53 Armengod ME Meseguer S Villarroya M et al 2014 Modificationof the wobble uridine in bacterial and mitochondrial tRNAs readingNNANNG triplets of 2-codon boxes RNA Biology 11 1495ndash507