[ Pobierz całość w formacie PDF ]
//-->Bergströmet al. BMC Research Notes2012,5:402RESEARCH ARTICLEOpen AccessNature of bacterial colonization influencestranscription of mucin genes in mice duringthe first week of lifeAnders Bergström1*, Matilde B Kristensen1, Martin I Bahl1, Stine B Metzdorff2, Lisbeth N Fink3, Hanne Frøkiær2and Tine R Licht1AbstractBackground:Postnatal regulation of the small intestinal mucus layer is potentially important in the developmentof adult gut functionality. We hypothesized that the nature of bacterial colonization affects mucus gene regulationin early life.We thus analyzed the influence of the presence of a conventional microbiota as well as two selectedmonocolonizing bacterial strains on the transcription of murine genes involved in mucus layer development duringthe first week of life.Mouse pups (N = 8/group) from differently colonized dams: Germ-free (GF), conventional specific pathogen free(SPF), monocolonized with eitherLactobacillus acidophilusNCFM (Lb) orEscherichia coliNissle (Ec) were analyzedby qPCR on isolated ileal tissue sections from postnatal days 1 and 6 (PND1, PND6) after birth with respect to:(i) transcription of specific genes involved in mucus production (Muc1-4,Tff3)and (ii) amounts of 16S rRNA ofLactobacillusandE. coli.Quantification of 16S rRNA genes was performed to obtain a measure for amounts ofcolonized bacteria.Results:We found a microbiota-independent transcriptional increase of all five mucus genes from PND1 to PND6.Furthermore, the relative level of transcription of certain mucus genes on PND1 was increased by the presence ofbacteria. This was observed forTff3in the SPF,Ec,andLbgroups; forMuc2in SPF; and forMuc3andMuc4inEcandLb,respectively.Detection of bacterial 16S rRNA genes levels above the qPCR detection level occurred only on PND6 and only forsome of the colonized animals. On PND6, we found significantly lower levels ofMuc1, Muc2andMuc4genetranscription forLbanimals with detectableLactobacilluslevels as compared to animals withLactobacilluslevelsbelow the detection limit.Conclusions:In summary, our data show that development of the expression of genes encoding secreted(Muc2/Tff3) and membrane-bound (Muc1/Muc3/Muc4) mucus regulatory proteins, respectively, is distinct andthat the onset of this development may be accelerated by specific groups of bacteria present or absent at themucosal site.Keywords:Germ free mice, Monocolonized, qPCR, LinRegPCR, Postnatal transcription onset, Probiotics,Lactobacillus acidophilusNCFM,Escherichia coliNissle, 16S rRNA* Correspondence:adbe@food.dtu.dk1Gut Ecology Group, Department of Food Microbiology, National FoodInstitute, Technical University of Denmark, Mørkhøj Bygade 19, Søborg 2860,DenmarkFull list of author information is available at the end of the article© 2012 Bergström et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creativereproduction in any medium, provided the original work is properly cited.Bergströmet al. BMC Research Notes2012,5:402Page 2 of 7BackgroundThe interplay between the microbiota of the gut and theintestinal mucus layer in early life is important in thedevelopment of the epithelial barrier as part of theinnate immune defense [1]. The first weeks and monthsafter birth are believed to be crucial for establishment ofthe gut microbiota and consequently for the health andintegrity of the epithelium throughout life [2,3]. In thisperiod, a development regulated by endogenous factorssuch as hormones, in parallel with gene regulationcaused by the microorganisms present in the gut, takesplace [4,5]. The presence and composition of the micro-biota has been shown to be directly involved in theregulation of gene transcription in the intestinal epithe-lium, including the mucin genes,Muc1-4and the trefoilfactorTff3[4,6].In the human intestines, MUC1-4 are the most preva-lent [6] of the different mucin gene transcripts describedto date [1,7,8]. In the gastrointestinal tract, specificmucins show coordinated expression and localizationwith the viscosity regulating trefoil factors (TFF’s), inparticular TFF3 [1]. Epithelial linings contain bothmembrane-bound (MUC1, MUC3, MUC4) and secretedgel-forming mucins (MUC2) expressing highly specificoligosaccharide side chains, which are important in rela-tion to filtering the entry of various moieties e.g. bacteriaand food to the underlying tissue. The membrane-boundmucins act as cell-surface receptors and sensors, mediat-ing signals to trigger cell proliferation, apoptosis, differ-entiation and specific secretions, when relevant [1]. Thefour human mucin genes (MUC1-4) all share a fairlyhigh degree of sequence, distribution and functionalhomology to the mouse mucin genesMuc1-4[9-12].As facultative anaerobes, lactobacilli andE. colistrainshave been recognized as successful early life colonizersof the sterile gastro-intestinal tract [13,14]. Strains ofLactobacillus acidophilusare known to stimulate tran-scription of mucin genesin vitro[15,16]. Moreoveradministrations of probiotic lactobacilli and bifidobac-teria have been shown to increase ileal gene and proteinlevels ofMuc3in adult rats [17] and cell cultures [16],respectively. CertainE. colistrains have been associatedwith increased production ofMUC2, MUC3andMUC4in human ileal cells [18].In order to elucidate the role of microbial colonizationin the postnatal regulation ofMuc1-4andTff3,we inves-tigated the expression of these genes in mouse ileal seg-ments isolated at the first day after birth (PND1) and sixdays after birth (PND6), respectively, from specificpathogen free, conventionally bred mice (SPF), micemonocolonized with eitherLactobacillus acidophilusNCFM (Lb) [19] orE. coliNissle (Ec) [20,21], and fromgerm free (GF) mice [15,22]. Specifically, samples werecollected and analyzed at PND1 and PND6 to examinethe immediate postnatal effects, which are relevant forimmune system priming [22,23]. Quantification of bac-terial 16S rRNA gene levels was performed to obtain ameasure of bacterial colonization levels in the differentanimal groups on PND1 and PND6.Results and discussionqPCRWe introduced several new primers in this study, allscoring successfully on our validation criteria. Lin-RegPCR [24,25] was utilized for qPCR analysis, as itenables individual PCR efficiencies to be calculated. Thestandard curve assumption, that in all samples the PCRefficiency/amplicon, based on one“representative”DNAsample is constant, is replaced by an assumption-freemethod based on linear regression in the exponentialphase of the fluorescence of the actual individualsamples analyzed [24]. Further, by including in the subse-quent calculation of average efficiency/amplicon, onlysuccessful samples within 5 % of the mean efficiency/amplicon, contributions from diverging samples to thefinal results are excluded.We tested the choice of reference gene, but interestinglyfound no significant difference in the results between beta-actin [26,27], neuroplastin (Genevestigator recommen-dation) nor the geometrical mean of them both.Effect of time and bacterial colonization on regulationof Muc1-4 and Tff3 transcriptionIn GF mice,Muc1-4andTff3,all showed statistically sig-nificant increases in transcription from PND1 to PND6,indicating that this event occurs during the first post-natal week independently of the presence of microbes(Figure 1). For certain mucin genes, presence of bacteriain the colonized animals correlated with an increasedrelative abundance of transcripts on PND1 compared totranscription levels of the same genes in GF mice. Thiswas particularly evident for the genesMuc2andTff3.Increased transcription on PND1 ofTff3was observed inconventional pups (SPF) as well as in pups of dams’monocolonized with eitherLborEc,while forMuc2,thiswas observed only in presence of a full microbiota (SPF).ForMuc3inEcandMuc4inLb,a higher level of tran-scription was observed on PND1 than in GF pups, indi-cating thatE. coliandLactobacillusmay specificallystimulate transcription of these genes immediately afterbirth (Figure 1).The higher level ofMuc2andTff3transcriptions atPND1, both encoding secreted proteins with goblet cellorigin [28], indicates that the presence of bacteria affectsgene transcription onset in these exocytotic cells. Whileboth gene products play protective roles during gutinflammatory conditions, at sites of epithelial damage[18,29-34] and during postnatal development [35,36],Bergströmet al. BMC Research Notes2012,5:402Page 3 of 7Figure 1Comparison of mucin gene expression between the four animal groups on PND1 and PND6.Transcription ofMuc1, Muc2, Muc3,Muc4andTff3on postnatal days (PND) 1 and 6 for pups in groups: GF, SPF,LbandEc.Each column represents the average relative abundance(Ra) (See Methods) of 8 animals. Error bars show SEM values.*p < 0.05,**p < 0.01,***p < 0.001 relative to PND1 (for the same group)+p < 0.05,++p < 0.01,+++p < 0.001 relative to GF (for the same day).Muc2, unlike Tff3, polymerizes into a protective gel-likestructure [1]. Based on the obtained results, it is howevernot possible to determine whether there is a connectionbetween this difference in functionality and the corre-sponding gene regulation.Previously, we demonstrated how microbiota affectsileal gene expression of a number of immune relatedgenes (specific cytokines and chemokines) during thefirst week after birth [23]. As seen forMuc2in thepresent study, and also for a number of Toll-like recep-tor signaling pathway related genes such asTlr2/4, Irak1and the chemokineCxcl2,encoding MIP-2, the presenceof a full microbiota was required to influence geneexpression on PND1, which was only to a limited degreeaffected by monocolonization with eitherLactobacilliorE. coli[23].Increased transcription ofMuc3andMuc4on PND1was observed inEcandLbpups, respectively, but not inSPF (Figure 1). Although specific probiotic bacteria, in-cludingLactobacillus acidophilusNCFM [15],Lactoba-cillus rhamnosus[17],Bifidobacterium bifidum[17],Lactobacillus plantarum[16,17] as well as two atypical,enteropathogenicE.colistrains [18], have previously beenshown to stimulate mucin gene expression, this study isto our knowledge the first to address such effects at avery early stage of life.Muc1transcription levels were inthis study apparently not affected by the presence ofbacteria.Bacterial 16S rRNA abundance on PND1 and PND6bacterial levels in the ileal sections increased betweenPND1 and PND6 after birth, although the employed pro-cedure did not allow detection of bacterial 16S rRNA inall pups. Culture-based techniques have shown that thegut mucosal surfaces in newborn mice follow a ratherconserved colonization pattern [37]. In particular, lacto-bacilli colonize within the first 1–2 days after birth,whereas coliforms are normally not quantifiable in themucosal layers until approximately 9 days after birth[14]. These results are thus consistent with findings inthe SPF group in this study. It is however important tonote, that the current analysis was performed on wholeintestinal sections, including both luminal contents andmucosal surfaces, whereas the other studies referred towere based on analysis of mucosal surfaces only.There was a significantly lower level of transcripts(p < 0.05) ofMuc1, Muc2andMuc4in the pups with de-tectable amounts of lactobacilli on PND6 in theLbgroup than in pups with colonization below the detec-tion limit (Figure 2). In other words, colonization withrelatively high levels ofLactobacilliin the pups had anegative effect on mucin gene transcription on PND6.ForMuc2,pups colonized withLactobacillusbelow thedetection limit in theLbgroup were indeed comparableto GF pups.It has been established by others that degradation ofmucin in adult rats [38] as well as gene expression ofMuc1-4andTff3in six week old mice [6], is differentNone of the PND1 samples containedLactobacillusorE.coliin amounts above the qPCR detection limit (DL), inany of the four animal groups (Table 1). This wasexpected, since only partial bacterial colonization isachieved so short after birth. On PND6, 5/8 pups in boththeLband SPF groups, respectively, were positivelyabove theLactobacillus16S rRNA DL, while 8/8 and 0/8in theEcand SPF groups, respectively, were colonizedabove theE.coliDL. These observations corresponded todifferences in Nvalues (See Methods) of >300-fold forLactobacillusand >160-fold forE.coli.This shows thatTable 1 16S rRNA measured presence vs. absence of all4 animal groups on each of days PND1 and PND6PND1Lactobacillus16SGFSPFEcLb0/80/80/80/8E. coli16S0/80/80/80/8PND6Lactobacillus16S0/85/80/85/8E. coli16S0/80/88/80/8The fraction denotes number of samples significantly above detection limit(DL) of the total number (N = 8 in each group).Bergströmet al. BMC Research Notes2012,5:402Page 4 of 7Figure 2Difference in mucin gene expression between pups with detectable and undetectableLactobacillus16S inLb.Comparison ofMuc1, Muc2, Muc3, Muc4andTff3transcription on PND6 between pups in theLbgroup, whereLactobacillus16S could be detected and could notbe detected, respectively. Each column represents the average relative abundance (Ra) of 5/8 (above DL) or 3/8 (below DL) animals. Error barsshow SEM values. DL = Detection Limit.*p < 0.05, denotes significant difference between detectably colonized and not colonized.when comparing GF and conventional animals. Clearly,gene regulation induced by the colonizing microbiota isa complex and continuous process occurring throughoutthe first weeks of life, and as a more stable and adult-likemicrobiota is probably not achieved until the end ofweaning process at approximately 21 days after birth[39], the expression of the mucus regulating genes maychange not only in newborn animals, but also later in lifein response to periodic changes in the microbiotaConclusionsIn this manuscript, we show distinct differences betweenthe expression patterns of secreted (Muc2/Tff3) andmembrane-bound (Muc1/Muc3/Muc4) mucus-regulatorygenes in the very first days after birth. Presence of a fullmicrobiota (SPF) increased the relative level of transcrip-tion ofMuc2andTff3,which implies the two corre-sponding secreted gene products, Muc2 and Tff3, to playprotective roles in the postnatal intestinal layer develop-ment. The immediate activation ofMuc2/Tff3transcrip-tion may provide a coating of the new born ilealepithelial layer, allowing only passage of certain sub-stances or organisms.MethodsAnimal experimentsperformed 7 days after mating by applying 5x108CFU ml-1in 0.5 ml PBS suspension orally using a pipetteand 0.5 ml to the abdominal skin.Lbwas grown anaer-obically in de Man, Rogosa, and Sharpe broth (MRS,Merck, Darmstadt, Germany) andEcaerobically inLuria-Bertani broth (LB, Merck) overnight at 37°C. Thecultures were harvested, washed twice in sterilephosphate-buffered saline (PBS) (Lonza, Basel, Switzer-land), re-suspended in 1/50 of the original culturevolume and frozen at−80°C until use. Prior to adminis-tering bacteria to the mice,Ecsuspensions were dilutedtenfold in PBS immediately to obtain 5x108CFU ml-1.Lbsuspensions were not diluted. Four litters spontan-eously delivered from 4 different mothers in each group;SPF, GF,LbandEc,were used for the experiment. Atpost-natal days 1 and 6, the pups were put down andthe distal ileum (segment from cecum and 3 cm up) wasremoved from the small intestine of two pups per litter andfrozen in RNAlater (Qiagen, Hilden, Germany). No sep-aration of mucosal from luminal content was performed.EthicsThe mouse experiment was performed under a license toDepartment of Microbiology, National Food Institute,from the Danish Council for Animal Experimentation.RNA isolationGF Swiss Webster mice and SPF mice, containing con-ventional microbiota, were purchased from Taconic (LilleSkensved, Denmark) and kept in GF isolators or underspecific pathogen-free conditions, respectively [22]. Fecalsamples from GF mice, taken at sampling i.e. once aweek, were cultivated on non-selective LB medium andunder aerobic and anaerobic conditions to confirm steril-ity of isolators. For breeding, pairs of female mice werehoused with one male until plugs were observed. Mono-colonization of pregnant mice withEcandLbwasTissues were removed from RNAlater and homogenizedby a rotor strator in RLT buffer (Qiagen). RNA fromtissue homogenate was extracted using RNeasy Mini Kitfrom Qiagen following the supplier's protocol.Primer design and validationA list of all primers used in this study is presented inTable 2. All primers found in references were initiallyBergströmet al. BMC Research Notes2012,5:402Page 5 of 7Table 2 Primers used for qPCRPrimer nameMuc1Muc2Muc3Muc4Tff3NeuroplastinBeta-actinLactobacillus16S rRNAE. coli16S rRNAFwd (5´-3´)TCGTCTATTTCCTTGCCCTGCCCAGAAGGGACTGTGTATGTGGTCAACTGCGAGAATGGAGTCTCCCATCACGGTTCAGTCTCTGTCACATCGGAGCAGTGTCGCTGCTCAGAACGAACCAAGAAGTCCACCTTCCAGCAGATGTAGCAGTAGGGAATCTTCCAaCATGCCGCGTGTATGAAGAARev (5´-3´)ATTACCTGCCGAAACCTCCTTTGTGTTCGCTCTTGGTCAGTACGCTCTCCACCAGTTCCTTGTCATTCCACACTCCCAGATGAAGCACCAGGGCACATTCTTACGGGTGGCAGTGAGTTGAAAGGGTGTAAAACGCAGCCACCGCTACACATGGAGbCGGGTAACGTCAATGAGCAAAAmpliconsize185276982807716011734196ReferenceThis studyModified from [44]Modified from [6]This study[45]Modified from [46]This studya[47]b[48][49]premierbiosoft.com/netprimer/index.html). Primers notscoring a rating of at least 90 % were not accepted andnew primers were then designed with NCBI’s primerblast/) and the quality was again verified until satisfactionwith the Net Primer Software. All newly designed primerswere designed to span exon junctions to avoid amplifica-tion of genomic DNA. The specificity of all primers wasevaluatedin siliconcbi.nlm.nih.gov/Blast.cgi).qPCR setupThree separate qPCR experiments on ileal cDNA wereperformed; 1) and 2) were separate replications of relativequantifications on mucus gene transcription (Muc1-4 andTff3)with selected reference genes (see next paragraph)and 3) on presence or absence of specific bacterial 16SrRNA analysis (Lactobacillus,E.coli).qPCR data analysisQuantitative PCR (qPCR)Isolated ileal RNA was reverse transcribed into cDNAusing SuperScriptWVILO™ cDNA Synthesis Kit fromInvitrogen, Denmark. After verifying the quality of thecDNA by spectroscopy (A260/A280= 1.8 ± 10 %) measuredon a NanoDrop ND-1000 Spectrophotometer (SaveenWerner, Limhamn, Sweden), it was used as template inquantitative real-time PCR using the ABI prism 7900HTfrom Applied Biosystems. All cDNA concentrations werewithin the range of 90-100 ngμl-1. The amplification reac-tions were carried out in a total volume of 20μlcontaining10μlmaster mix (2x PerfeCTaTMSYBRWGreen Super-Mix, ROX from Quanta BiosciencesTM), 0.4μlof eachprimer (10μM),2μltemplate cDNA, and 7.2μlnuclease-free water (Qiagen GmbH, Germany) purified for PCR.The amplification program consisted of one cycle of 50 °Cfor 2 min; one cycle of 95 °C for 10 min; 40 cycles of 95 °C for 15 s and 60 °C for 1 min; and finally one cycle ofdissociation curve analysis for assessing the amplificationproducts (95 °C for 15 s, 60 °C for 20s and increasingramp rate by 2 % until 95 °C for 15 s). These conditionswere selected based on preliminary qPCR experiments ontarget DNA with similar concentrations (100 ngμl-1).Samples of all amplification products were furthersubjected to gel electrophoresis in 2 % agarose, followed byethidium bromide staining in order to verify ampliconsizes.All qPCR analysis was performed with the freely avail-able LinRegPCR tool developed by Ruijter et al. [24,25].The raw fluorescence data were exported from theABI prism 7900HT SDS-software, and the LinRegPCRprogram was used to estimate baselines and individualPCR efficiencies in order to calculate output as targetstarting concentration, expressed in arbitrary fluores-cence units N, for each PCR sample by the formulaN= threshold / (EffCt), where Effmeandenotes themeanoptimal PCR mean efficiency/amplicon, threshold theoptimal“cutoff”in the exponential region and Ctisthe number, where each PCR sample exceeds this thresh-old. Samples with no amplification, baseline error, toomuch noise or without plateau were automaticallyexcluded by the LinRegPCR software. Subsequently, foreach amplicon the average of all remaining, successfulsamples within 5 % of the mean value of all successfulsamples/amplicon were used in the calculation of Effmeanfor each amplicon. All N-values were calculated asmeans of double qPCR determinations.For relative quantification of mucus gene transcripts,two different eukaryotic reference genes were used namelybeta-actin [40] and neuroplastin, the latter suggested bythe Genevestigator software (https://www.genevestigator.com) [41,42] based on microarray data on similar or-ganism (M.musculus)and tissue (ileum). We used thegeometric mean of the two reference genes as previouslysuggested [43]. Normalization to relevant reference geneexpression was then calculated according to the formula: [ Pobierz całość w formacie PDF ]