Exopolysaccharide Producing Bifidobacterium animalis subsp. lactis Strains Modify the Intestinal Microbiota and the Plasmatic Cytokine Levels of BALB/c Mice According to the Type of Polymer Synthesized

micb-11-601233 No embe 21, 2020 Time: 13:27 # 1 ORIGINAL RESEARCH published: 26 No embe 2020 doi: 10.3389/ micb.2020.601233 Edi ed by: Sonia G. Sáyago-Aye di, Ins i u o Tecnológico de Tepic, Mexico Re iewed by: Hugo Ga cia, Tecnologico Nacional de Mexico/IT Ve ac uz, Mexico Pa icia Bu ns, Na ional Uni e si y o he Li o al, A gen ina Sil ia Tu oni, Uni e si y o Bologna, I aly *Co espondence: Pa icia Ruas-Madiedo [email p o ec ed] †These au ho s ha e con ibu ed equally o his wo k Special y sec ion: This a icle was submi ed o Food Mic obiology, a sec ion o he jou nal F on ie s in Mic obiology Recei ed: 31 Augus 2020 Accep ed: 05 No embe 2020 Published: 26 No embe 2020 Ci a ion: Saba e C, Moline o-Ga cía N, Cas o-B a o N, Diez-Echa e P, Hidalgo-Ga cía L, Delgado S, Sánchez B, Gál ez J, Ma golles A and Ruas-Madiedo P (2020) Exopolysaccha ide P oducing Bi idobac e ium animalis subsp. lac is S ains Modi y he In es inal Mic obio a and he Plasma ic Cy okine Le els o BALB/c Mice Acco ding o he Type o Polyme Syn hesized. F on . Mic obiol. 11:601233. doi: 10.3389/ micb.2020.601233 Exopolysaccha ide P oducing Bi idobac e ium animalis subsp. lac is S ains Modi y he In es inal Mic obio a and he Plasma ic Cy okine Le els o BALB/c Mice Acco ding o he Type o Polyme Syn hesized Ca los Saba e 1,2†, Na alia Moline o-Ga cía1,2†, Nu ia Cas o-B a o1,2, Pa icia Diez-Echa e3,4, Lau a Hidalgo-Ga cía3,4, Susana Delgado1,2, Bo ja Sánchez1,2, Julio Gál ez3,4, Abela do Ma golles1,2 and Pa icia Ruas-Madiedo1,2* 1Depa men o Mic obiology and Biochemis y o Dai y P oduc s, Ins i u o de P oduc os Lác eos de As u ias – Consejo Supe io de In es igaciones Cien í icas (IPLA-CSIC), Villa iciosa, Spain, 2Mic oheal h G oup, Ins i u o de In es igación Sani a ia del P incipado de As u ias (ISPA), O iedo, Spain, 3CIBER-EHD, Depa men o Pha macology, Cen e o Biomedical Resea ch (CIBM), Uni e si y o G anada, G anada, Spain, 4Ins i u o de In es igación Biosani a ia de G anada (ibs.GRANADA), G anada, Spain Bac e ia-hos in e ac ions a e media ed by di e en mic obial associa ed molecula pa e ns which a e mos o en su ace s uc u es such as, among o he s, exopolysaccha ides (EPSs). In his wo k, he capabili y o wo isogenic EPS-p oducing Bi idobac e ium animalis subsp. lac is s ains o modula e he gu mic obio a o heal hy mice, was assessed. Each s ain p oduces a di e en ype o polyme ; he opy s ain S89L syn hesized a hamnose- ich, high-molecula weigh EPS in highes abundance han he non- opy DMS10140 one. BALB/c mice we e o ally ed o 10 days wi h milk-bi idobac e ial suspensions and ollowed a e wa d o 7 pos -in e en ion days (wash-ou pe iod). The colonic con en o mice was collec ed in se e al sampling poin s o pe o m a me a axonomic analysis. In addi ion, he in luence o speci ic mic obial clades, appa en ly s imula ed by he opy and non- opy s ains, on mouse plasma ic cy okine le els was in es iga ed h ough hie a chical associa ion es ing. Analysis o 16S RNA gene sequences showed ha he abundance o Fi micu es phylum signi ican ly inc eased 7 days a e cessing he ea men wi h bo h s ains. The ela i e abundance o Allop e o ella genus also ose, bu a e sho e pos - ea men imes (3 days o bo h DMS10140 and S89L s ains). Some bac e ial clades we e speci ically modula ed by one o ano he s ain. As such, he non- opy DMS10140 s ain exe ed a signi ican in luence on In es inomonas genus, which inc eased a e 4 pos -adminis a ion days. On he o he hand, eeding wi h he opy s ain S89L led o an inc ease in sequences o Faecalibaculum genus a 4 pos - ea men days, while he abundance o E ysipelo ichaceae and Lac obacillaceae amilies inc eased o F on ie s in Mic obiology | www. on ie sin.o g 1No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 2 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide p olonged imes. Associa ion es ing e ealed ha se e al lac obacilli and bi idobac e ial signi ican ly s imula ed by opy S89L s ain we e posi i ely associa ed wi h he le els o ce ain cy okines, including IL-5 and IL-27. These esul s highligh ele an changes in mice gu mic obio a p oduced a e adminis a ion o he opy S89L s ain ha we e associa ed o a po en ial immune modula ion e ec . Keywo ds: exopolysaccha ide, bi idobac e ia, mic obio a, mice, cy okine, immune esponse INTRODUCTION Bi idobac e ium is pa o he human in es inal mic obio a being one o he mos abundan gene a in his ecosys em, mainly in in an s. Fu he mo e, some bi idobac e ial s ains a e conside ed o ha e p obio ic p ope ies and, hus, hey a e commonly used in comme cial p obio ic p oduc s. Some o hei heal h p ope ies could be ela ed o he p esence o speci ic su ace molecules, such as exopolysaccha ides (EPS). EPS a e ca bohyd a e polyme s syn hesized by some bac e ia ha can be o ally libe a ed o he ex acellula milieu, o can emain loosely a ached o he bac e ial su ace. EPS play an impo an ole in he in es inal ecosys em by media ing bac e ial-hos in e ac ions, modula ing he immune sys em esponse, and also ac ing as a e men able ca bon sou ce by o he membe s o he mic obio a (Cas o-B a o e al., 2018). Besides, hese polyme s a e o g ea impo ance o he p oducing bac e ia due o hei p o ec i e ole agains ad e se condi ions, hus allowing hei pe sis ence in he gu o a longe ime (Fanning e al., 2012). EPS s uc u es con ain di e en monome s, mainly D-glucose, D-galac ose, and L- hamnose, al hough o he s can be ound such as N-ace yl glucosamine, D- glucose, D- ibose, o ucose. Some o hese bac e ial EPS can be used by he human gu mic obio a leading o a high p oduc ion o sho chain a y acids (Salaza e al., 2016;Liu e al., 2019). I has been demons a ed ha some EPS-p oducing bi idobac e ia can modula e he in es inal mic obio a di e si y and unc ion (by modi ying he p o ile o eleased me aboli es), as de e mined in i o by means o pH-con olled ecal ba ch e men a ions o in i o using animal expe imen al s udies (Salaza e al., 2016;Yan e al., 2019). Mic obio a modula ion in in lamma o y p ocesses can be o g ea impo ance. Recen s udies desc ibe he e olu ion o gu mic obiome du ing in lamma o y bowel disease and de ec ele an ela ionships be ween ce ain axa and se um le els o an ibodies h ough hie a chical associa ion models (Lloyd-P ice e al., 2019). Among he biological ac i i ies o EPS desc ibed, also he s uc u e-immuni y ela ionships ha e gained g ea a en ion (Hidalgo-Can ab ana e al., 2012;Xu e al., 2019). In ac , wo ecen s udies ca ied ou wi h an EPS-p oducing Bi idobac e ium longum s ain in a mu ine model o DSS-induced coli is showed ha he s ain was able o alle ia e he in lamma o y symp oms h ough he mic obio a modula ion and he main enance o he mucosal ba ie (Yan e al., 2019, 2020). Se e al s udies dealing wi h EPS p oducing Bi idobac e ium animalis subsp. lac is s ains, speci ically hose ha ing a “ opy” pheno ype due o he p oduc ion o a hamnose- ich high molecula weigh (HMW)-EPS, ha e been ca ied ou in ou esea ch g oup. I has been demons a ed ha s ains p oducing his opy EPS can in i o elici di e en immune esponses when co-incuba ed wi h PBMC (pe iphe al blood mononuclea cells) om humans (López e al., 2012) o wi h GALT (gu associa ed lymphoid issue) isola ed om a s (Hidalgo-Can ab ana e al., 2014), being also able o amelio a e in lamma o y symp oms in a DSS-induced coli is mice model (Hidalgo-Can ab ana e al., 2016). Howe e , so a , no s udies demons a ing whe he he in es inal mic obio a could be di e en ially modula ed by opy and non- opy s ains in a heal hy animal model ha e been epo ed. The e o e, he aim o his s udy was o de e mine he capabili y o di e en EPS-p oducing B. animalis subsp. lac is s ains o modi y he in es inal mic obio a in heal hy BALB/c mice and o explain he in luence o ele an clades s imula ed by EPS on he p oduc ion o se um cy okines in animals wi h a non- dis u bed mucosal ba ie . To achie e his, a ecen ly de eloped bac e ial model, based on he wild- ype s ain B. animalis subsp. lac is DSM10140 (non- opy) and i s opy isogenic mu an S89L ha p oduces he HMW-EPS in highe abundance, was selec ed. Mu an S89L was ob ained using gene eplacemen echniques o subs i u e he gene bala _1410 ( esponsible o he de e mina ion o he EPS chain leng h) o DSM10140 o he mu a ed one wi h a single nucleo ide change (Cas o-B a o e al., 2017). The e o e, di e ences be ween bo h s ains on mic obial and immune modula ion capabili ies could be only a ibu ed o a single gene which de e mines he chain size o he polyme . Then, he main di e ence be ween bo h s ains is ha S89L p esen s in i s su ace a bigge amoun o he HMW-EPS han i s pa en al DSM10114 s ain. MATERIALS AND METHODS Bac e ial G ow h Condi ions The DSM101410 and S89L s ains we e cul i a ed in MRSc [MRS (Bioka Diagnos ics) supplemen ed wi h 0.25% L-cys eine-HCl (Sigma-Chemical Co.)] a 37◦C, o 24 h, in a ja unde anae obic condi ions (Anae ocul A, Me ck). Cul u es we e washed wi h PBS and esuspended in hea - ea ed 11% skimmed milk (BD Di co, The mo Fishe Scien i ic Inc.). Bi idobac e ial suspensions in milk, con aining on a e age 8.9 ±0.4 Log CFU/ml, o each s ain we e daily p epa ed o be adminis e ed (dose o 100 µl) o he mice by means o a gas ic ube. Expe imen al Design The animal expe imen al p ocedu e was app o ed by he E hical Commi ee o Labo a o y Animals o he Uni e si y o G anada (Spain) (Re . No. CEEA-2010-286). Female BALB/c mice (7– 9 weeks old, app oxima ely 20 g) we e ob ained om Jan ie Labs (S Be he in Cedex) and kep unde con en ional condi ions F on ie s in Mic obiology | www. on ie sin.o g 2No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 3 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide wi h a s anda d pelle ed die and s e ilized wa e o 1 week be o e beginning he expe imen s. A o al o 114 animals we e andomly dis ibu ed in h ee g oups: 48 mice ecei ing he non- opy s ain (DSM10140 g oup), 48 mice ecei ing he opy s ain (S89L g oup) and he 18 emaining mice did no ecei e any ea men ( e e ence g oup). The weigh a he beginning o he expe imen al p ocedu e o each animal g oup was (mean ±SD): 22.28 ±1.45 g o he con ol g oup, 22.82 ±1.56 g o DSM10140- ea ed g oup and 22.14 ±1.65 o g he S89L- ea ed g oup. The s a is ical analysis pe o med (ANOVA) showed no s a is ical di e ences in he ini ial weigh among he h ee g oups (p>0.05). The expe imen al design was as ollows: 10 days o in e en ion wi h daily o al adminis a ion o he milk-bi idobac e ia suspensions, ollowed by 7 days o pos - in e en ion wi hou bi idobac e ia in ake (wash-ou pe iod). No a ia ions in beha io o heal h s a us we e obse ed in he h ee g oups o mice du ing he expe imen a ion pe iod. In di e en sampling poin s (5 and 10 days o in e en ion, and 1, 3, 4, and 7 pos -in e en ion days), 8 mice om each bi idobac e ial g oup, and 3 mice om he e e ence g oup ( hus, 18 con ol animals in o al), we e sac i iced in o de o collec he colonic con en and he blood se um. Fo ha , each mouse was anes he ized wi h an o e dose o halo hane and blood was ex ac ed om he hea using hepa inized ubes. A e ha , animals we e sac i iced by ce ical disloca ion and he gu was excised o collec i s con en . Samples we e s o ed a −80◦C a e hei collec ion. Cy okine Analysis The P oca aPlex Mul iplex Immunoassay o mouse (The mo Fishe Scien i ic Inc.) was used o quan i y he le els o di e en cy okines (IL-1β, IL-2, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12p70, IL-13, IL-17a, IL-18, IL-22, IL-23, IL-27, GM-CSF, IFN-γ, and TNF-α) by means o he FACS Can o II low cy ome e (BD Biosciences). The de ec ion limi s (pg/mL) o hose de ec ed in ou samples we e: 2.21 ±0.36 (IL-5), 5.48 ±0.72 (IL-6), 15.08 ±0.62 (IL-9), 2.27 ±0.01 (IL-10), 1.65 ±0.36 (IL-17a), 12.30 ±0.80 (IL-23), 2.68 ±0.01 (IL-27), and 1.12 ±0.19 (IFN-γ). 16S RNA and 16S-23S In e nal T ansc ibed Space (ITS) Gene Sequencing To al DNA was isola ed om he colonic con en o he 114 samples using he QiaEz DNA-ex ac ion p o ocol p e iously op imized in ou esea ch g oup (Milani e al., 2013), consis ing in a mechanical cell dis up ion s ep, ollowed by enzyma ic lysis and combined wi h an ex ac ion wi h he QIAamp S ool DNA ki (Qiagen). Using he p ime s P obio_Uni and P obio_Re (Milani e al., 2013), he V3 egion o he 16S RNA gene was ampli ied. The 16S-23S In e nal T ansc ibed Space s (ITS) we e ampli ied om ex ac ed DNA using he speci ic p ime pai P obioBi -ITS_Fw and P obioBi -ITS_Re , which a ge s he a iable egion be ween he 16S RNA and 23S RNA gene sequences (Milani e al., 2014). Sequencing was pe o med using an Illumina MiSeq machine a GenP obio S.R.L. (Pa ma, I aly). Sequence eads we e il e ed and he esul ing ones we e p ocessed using a pe sonalized sc ip o QIIME so wa e (Capo aso e al., 2010) ma ched by pai -ends. Quali y con ol il e ing was pe o med, keeping sequences wi h a mean sequence quali y sco e >20 and a leng h be ween 140 and 400 bp. Da a Analysis Shapi o–Wilk’s No mali y es (p<0.05) and Le ene’s es (p<0.05) o de e mine he homogenei y o a iances we e calcula ed o all da a gene a ed. S a is ically signi ican di e ences be ween samples we e calcula ed h ough K uskal– Wallis and Mann–Whi ney s a is ical es s o non-pa ame ic independen samples ollowed by False Disco e y Ra e (FDR) pos hoc using Benjamini–Hochbe g me hod wi h a alue o 0.25. All s a is ical analyses we e compu ed on R 3.5.0 and Mo hu so wa e. In o de o calcula e di e si y measu es, he 16S RNA eads we e clus e ed in Ope a ional Taxonomic Uni s (OTUs) de ined a ≥99% sequence homology by meanings o UCLUST so wa e (Edga , 2010). All eads we e classi ied o he lowes possible axonomic ank using he QIIME and he e e ence da abase SILVA (Quas e al., 2013). Simila i y o he mic obial communi ies be ween he samples was calcula ed by UniF ac me hod (Lozupone and Knigh , 2005). Phyloseq (McMu die and Holmes, 2013) and Mic obiome (Lah i and She y, 2017) packages we e also used in he analysis o sequencing da a. A e assignmen o eads o phylum, amily, genus and species le els, wo di e en ial analyses we e pe o med: Me as as algo i hm o Mo hu so wa e, and DESeq2 di e en ial abundance es ing o sequencing da a (Lo e e al., 2014). To de e mine s a is ical di e ences in axonomic da a acco ding o he s ain, ea men ime o bo h padj lowe han 0.05 and log2FoldChange g ea e han 1.5 we e conside ed. Rele an axonomic clades we e associa ed o cy okine plasma ic le els h ough hie a chical all-agains -all associa ion es ing (HAIIA)1(Lloyd-P ice e al., 2019) conside ing q- alues and a Bon e oni False Disco e y Ra e o 0.25. In addi ion, a g aphical co ela ion ne wo k be ween axa and cy okines was compu ed using cc epe (Schwage e al., 2019) and qg aph (Epskamp e al., 2012) packages. RESULTS In luence o Ropy and Non- opy S ains on Cy okine Le els The adminis a ion o non- opy DSM10140 and opy S89L B. animalis subsp. lac is s ains o BALB/c mice exe ed a ele an in luence on he le els o some se um cy okines. S a is ically signi ican di e ences in hei p oduc ion acco ding o he s ain selec ed ( opy o non- opy) as well as he ea men ime (5 and 10 days o in e en ion, and 1, 3, 4, and 7 pos -in e en ion days) we e ound (Figu e 1, he s a is ical 1hu enhowe .sph.ha a d.edu/halla F on ie s in Mic obiology | www. on ie sin.o g 3No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 4 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 1 | Box plo (median, in e qua ile ange, minimum and maximum alues) o he le el (pg/ml) o IL-10, IFN-γ, IL-17A quan i ied in he blood se um o con ol mice g oup, as wells as in he g oups ed wi h he non- opy DSM10140 o wi h he opy S89L Bi idobac e ium animalis subsp. lac is s ains a di e en sampling poin s (5i and 10i days o in e en ion, and 1p, 3p, 4p, and 7p days pos -in e en ion). Wi hin each s ain- ea men g oup, samples (days) ha do no sha e a common le e a e s a is ically di e en (p<0.05). Wi hin each ea men day, he s a is ical di e ences be ween he DSM10140 and S89L g oups a e ma ked wi h an as e isk (*p<0.05). Fo in o ma ion abou o he cy okines, see Supplemen a y Table 1 in he Supplemen a y Ma e ial. analysis is shown in Supplemen a y Table 1). Rega ding he an i-in lamma o y IL-10, al hough no s a is ical di e ences we e de ec ed (wi h he excep ion o sample 3p), le els we e highe a e S89L adminis a ion compa ed o DMS1040 s ain. On he o he hand, signi ican ly lowe le els o IFN-γand IL-17a we e ound in mice ea ed wi h DSM10140 han in ea ed wi h S89L du ing he in e en ion pe iod and un il he i s 3 pos -in e en ion days. Howe e , no di e ences we e de ec ed be ween bo h mice g oups when he a ios IL-10/INF-γo IL- 10/IL-17a we e calcula ed, al hough a endency o ha e highe alues (i.e., an i-in lamma o y p o ile) o he S89L ea ed g oup was de ec ed, p obably due o he highes p oduc ion o IL-10 (Supplemen a y Table 1). The elease o o he cy okines, such as IL-9, IL-23, and IL-27, was also enhanced in S89L ea men s (Supplemen a y Table 1), al hough a e 4 pos - ea men days no di e ences be ween bo h mice g oups we e de ec ed. In bo h mice g oups, simila concen a ions o o he quan i iable cy okines, such as IL-5 and IL-6, we e de ec ed, bu he las one end o be eleased in lowe concen a ions in mice ed wi h he opy S89L s ain. The le els o he o he s analyzed (IL-1β, IL-2, IL-4, IL-10, IL-12p70, IL-13, IL-18, IL-22, GM-CSF, and TNF-α) we e below he limi o de ec ion o he me hod used (Supplemen a y Table 1). A e a p incipal componen analysis (PCA) o his da a, no clea pa e ns could be in e ed in he gene al cy okine p o ile ob ained, since he PCA explains a low pe cen age (below 50%) o a iance (Supplemen a y Figu e 1). These esul s could be ela ed wi h he high in e -indi idual a iabili y ound among mice. In luence o Ropy and Non- opy S ains on Gu Mic obio a The in luence o he opy and non- opy EPS-p oducing s ains on mice gu mic obio a composi ion was s udied. Fi s , he alpha di e si y measu ing a iabili y o species wi hin a sample was calcula ed using di e en indices (Shannon, Simpson, and In e se Simpson) which p o ide complemen a y in o ma ion, e lec ing di e en pa e ns in he co e mic obio a. These indexes showed a g ea e dispe sion o mice ed wi h non- opy DSM10140 and opy S89L s ains in compa ison o he con ol g oup, which was mo e accen ua ed du ing he pos -in e en ion pe iod, especially in S89L g oup (Supplemen a y Figu e 2). The global Chao1 index, an es ima ion o he numbe o species ep esen ed by only one indi idual in he sample, was 90.3, anging om 89.6 o 93.0 in DSM10140 g oup and om 82.0 o 91.4 in S89L- ea ed mice. On he o he hand, he be a-di e si y (B ay–Cu is dis ance) measu ing di e ences in composi ion among samples, e ealed ha DSM10140 and S89L g oups had a mo e di e se mic obio a han he con ol g oup (Figu e 2A), especially a he end o he pos -in e en ion pe iod (Figu e 2B), achie ing a highe numbe o species a he ou h day o pos - in e en ion o he opy S89L ea men (Figu e 2C). These esul s sugges ha he adminis a ion o non- opy DSM10140 F on ie s in Mic obiology | www. on ie sin.o g 4No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 5 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 2 | Be a-di e si y (B ay–Cu is me ics) o he ela i e abundance o colonic mic obio a samples om mice g ouped acco ding o he EPS-s ain used o eeding (A), he in e en ion ime (B) and a combina ion o bo h EPS-s ain and sampling (days) poin s (C). Di e en le e s indica e s a is ical di e ences among he di e en EPS-s ain g oups (A) o sampling days (B); in he combina ion o bo h ac o s (C), wi hin each EPS-s ain g oup samples (days) hose ha do no sha e a common le e a e s a is ically di e en (p<0.05), whe eas wi hin each ea men day, he s a is ical di e ences be ween he DSM10140 and S89L g oups a e ma ked wi h an as e isk (*p<0.05). and opy S89L s ains s imula e a signi ican numbe o species ha a e p esen in low abundances in he con ol g oup. In addi ion, he opy EPS seems o exe a s onge e ec on gu mic obio a ha akes place mainly a he inal days o he pos - ea men pe iod. The ela i e abundance o he main (≥1%) bac e ial phyla de ec ed in he colon con en o he di e en BALB/c mice g oups is ep esen ed in Figu e 3 ( he comple e s a is ical analysis is p o ided in Supplemen a y Table 2). In gene al, he e olu ion o abundances along he in e en ion and pos - in e en ion pe iod was simila in bo h bi idobac e ial- ea ed mice g oups. Speci ically, pe cen ages emained wi hou changes (p>0.05) be ween he wo sampling poin s (5i and 10i) o he in e en ion pe iod and hey esemble ha o he e e ence g oup. Howe e , a 1 and 3 pos -in e en ion days s a is ical di e ences (p<0.05) o some phyla we e de ec ed wi h espec o he in e en ion pe iod and also wi h espec o longe imes o he pos -in e en ion. In ac , i seems ha he mo e ime elapsed since he end o in e en ion (4 and 7 days, pos -in e en ion), he mo e esembles he p o ile o he mic obio a phyla o he ini ial s a e and o he e e ence (non- ea ed) g oup. Thus, he g ea es changes we e de ec ed a e 1 and 3 days o cessa ion o bi idobac e ial in ake; an inc ease in Ac inobac e ia and Bac e oide es abundance, o de imen o he Fi micu es phylum, was ound. This indica es ha he bigges in luence o ou EPS-p oducing bi idobac e ia on he colonic mic obio a occu ed du ing he wash-ou pe iod, a e inishing he o al adminis a ion o bo h s ains, sugges ing ha his in e en ion wi h ou EPS-p oducing bi idobac e ia had a delayed e ec on mic obio a dynamics in ag eemen wi h di e si y analyses p e iously desc ibed. The o he (mos abundan ) phylum p esen ing a ia ions along he in e en ion pe iod was P o eobac e ia (Figu e 3 and Supplemen a y Table 2); a educ ion in he ela i e abundance was obse ed om he beginning o he in e en ion pe iod o he end o he expe imen al p ocedu e in bo h bi idobac e ial- ea men g oups, bu his e ec only emained in he g oup S89L ed wi h he opy-EPS bi idobac e ia. Indeed a e 4 and 7 days pos -in e en ion he g oup ed wi h he s ain DSM10140 showed highe (p<0.05) ela i e abundance o his phylum han he S89L- ed g oup (Supplemen a y Table 2). In bo h cases, he le els o P o eobac e ia we e signi ican ly lowe han in he con ol g oup. To co obo a e hese s a is ically signi ican di e ences, DESeq2 di e en ial abundance es ing was applied o all phyla de e mined (Supplemen a y Figu e 3). I was also ound ha Tene icu es we e signi ican ly highe in mice ed wi h opy and non- opy s ains compa ed o con ol g oups (Supplemen a y Figu e 3A). P o eobac e ia abundance dec eased du ing pos -in e en ion pe iod when combined da a om bo h ea men s (Supplemen a y Figu e 3B) in ag eemen wi h he esul s p esen ed in Figu e 3. Mo eo e , signi ican di e ences we e obse ed o Fi micu es conside ing bo h he F on ie s in Mic obiology | www. on ie sin.o g 5No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 6 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 3 | Rela i e abundance o he main (≥1%) bac e ial phyla de ec ed in he colonic con en o di e en BALB/c mice g oups (n≥6 pe g oup and sampling day). The e e ence g oup was cons i u ed by non- ea ed mice o he same li e ; ea men g oups we e ed o 10 days wi h a daily dose o 8.9 ±0.4 Log CFU/ml, suspended in skimmed milk, o he EPS-p oducing s ains B. animalis subsp. lac is DSM10140 (pa en al, non- opy s ain) o S89L (mu an , opy s ain p oducing a HMW-EPS). Colonic con en samples we e ob ained om mice sac i iced a wo poin s du ing he in e en ion pe iod (5 and 10 days) and a ou poin s pos -in e en ion (wi hou adminis a ion o bi idobac e ia: 1, 3, 4, and 7 days). Fo s a is ical analysis see Supplemen a y Table S2 in Supplemen a y Ma e ial. selec ed s ain and in e en ion ime sepa a ely (Supplemen a y Figu e 3C); highe abundances a he end o he in e en ion (10i poin ) in S89L g oup we e obse ed. Howe e , in bo h g oups o mice, a ele an inc ease in he ela i e abundance o his phylum was achie ed a e 7 pos - ea men days in ag eemen wi h he non-pa ame ic s a is ical es s p esen ed in Figu e 3. Rega ding amily le el analysis, di e ences o ela i e abundances along he expe imen al poin s (Figu e 4), as well as be ween bo h ea men mice g oups a a gi en poin (Supplemen a y Table 3) we e obse ed. Among he mos abundan (>2%) amilies, he no iceable change a he end o he in e en ion pe iod (10i) was he signi ican (p<0.05) inc ease in Lac obacillaceae, which was also co obo a ed by DESeq2 di e en ial abundance es (Supplemen a y Figu e 4A). A dec ease in Bac e oidaceae and Ruminococcaceae (Clos idia class) in he S89L g oup was also deno ed (Figu e 4 and Supplemen a y Table 3) hen, concomi an ly, he e was an inc ease o Fi micu es/Bac e oide es a io a his 10i poin (Supplemen a y Figu e 5). Finally, he S89L ea men g oup also showed a signi ican dec ease o Rikenellaceae (Bac e oide es phylum) wi h espec o he DSM10140- ed g oup (Supplemen a y Table 3). Du ing he pos -in e en ion (wash-ou ) pe iod, a ema kable inc ease o Bac e oidales S24-7 amily was de ec ed a day 1 by bo h con en ional s a is ical es s (Figu e 4) and DESeq2 di e en ial abundance es ing (Supplemen a y Figu e 4B) in bo h g oups o mice, which ended o decline a e wa d wi h a concomi an long- e m inc ease o Lac obacillaceae. Cu iously, in he i s pos - in e en ion day he Bi idobac e iae amily inc eased i s ela i e abundance in bo h g oups o mice (Figu e 4), and he pe cen ages luc ua ed along wash-ou pe iod de ec ing signi ican di e ences be ween DSM10140 and S89L g oups a he 4 h pos -in e en ion day (Supplemen a y Table 3). A his day, he e we e also signi ican (p<0.05) di e ences on he ela i e abundance o Desul o ib ionaceae amily be ween ea men g oups, being lowe he pe cen age in S89L- ea ed mice. I should be no ed ha ad anced DESeq2 di e en ial abundance es also de ec ed signi ican changes in mino amilies p esen in mice mic obio a (>0.1%). Fo example, Lachnospi aceae inc eased du ing pos - in e en ion pe iod as well as E ysipelo ichaceae, which achie ed F on ie s in Mic obiology | www. on ie sin.o g 6No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 7 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 4 | Mean (n≥6 mice pe g oup and sampling day) o he bac e ial amilies, ha ing a ela i e abundance pe cen age highe han 1%, p esen in he colon con en o mice ea ed wi h he EPS-p oducing s ains B. animalis subsp. lac is DSM10140 o S89L. The ea men p ocedu e is desc ibed in he legend o Figu e 3. Fo s a is ical analysis see Supplemen a y Table S3 in Supplemen a y Ma e ial. highe abundances a he ou h day a e opy S89L ea men . Simila ly, Co iobac e iaceae inc eased du ing he i s days o pos -in e en ion wi h bo h s ains and dec eased a he se en h day p obably indica ing a pa ial loss o he modula o y ac i i y (Supplemen a y Figu e 4B). In gene al, Ve ucomic obiaceae showed highe abundances in mice adminis e ed wi h S89L, ega dless ime (Supplemen a y Figu e 4A). Taxonomic analysis o gene a p esen in samples ( ela i e abundance >0.1%) using DESeq2 di e en ial abundance es e ealed a gene al inc ease in Ruminococcaceae UCG-014 genus compa ed o he con ol g oup (Supplemen a y Figu e 6A). In addi ion, a s imula ion o Ruminococcus gna us g oup du ing pos -in e en ion was obse ed, which was mo e accen ua ed in S89L ea men (Supplemen a y Figu e 6B) al hough no di e ences we e ound among he h ee g oups o mice (Supplemen a y Figu e 6A). An inc ease in he abundance o his bac e ium has been associa ed wi h p o-in lamma o y s a es, such as C ohn disease (Henke e al., 2019); howe e , as we ha e indica ed in p e ious sen ences, he sligh inc ease in he ela i e abundance o R. gan us did no co ela e wi h a p o-in lamma o y s a e in o expe imen al model. In addi ion, S89L adminis a ion led o a high abundance o Eubac e ium issica ena g oup and Faecalibaculum a e 3 and 4 pos - ea men days, al hough his modula o y ac i i y dec eased a longe imes. Simila ly, Allop e o ella and In es inimonas achie ed highe abundances a e 2–3 pos -in e en ion days wi h bo h s ains (Supplemen a y Figu e 6B). Gi en ha his in e en ion s udy was ca ied ou wi h wo s ains o Bi idobac e ium, his genus was analyzed in mo e de ail. Figu e 5 shows he “box and whiske ” plo ep esen ing he ela i e abundance o he sequences iden i ied as bi idobac e ia. In iguingly, du ing he in e en ion p ocedu e a educ ion in he ela i e abundance o Bi idobac e ium was obse ed in bo h bi idobac e ial- ea ed g oups wi h espec he le els ound in he e e ence (non- in e en ion) mice g oup. In ac , his is coinciden wi h he lowe p opo ion o Ac inobac e ia phylum obse ed in he wo ea ed mice g oups in compa ison o he e e ence one (Figu e 3). I seems ha his was he phylum educing i s ela i e abundance in highe ex en du ing he in e en ion pe iod and Bi idobac e ium genus migh accoun o his beha io . Indeed, he eco e ing o Ac inobac e ia phylum and Bi idobac e iaceae amily, deno ed a he i s pos -in e en ion day, also ma ch wi h he inc ease o his genus in bo h ea men g oups (Figu e 5 and Supplemen a y Figu e 7). Besides, he s a is ical di e ences (Supplemen a y Table 3) obse ed in he amily a he 10 h day o in e en ion, and he 4 h pos -in e en ion day be ween F on ie s in Mic obiology | www. on ie sin.o g 7No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 8 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 5 | Rela i e abundance (%) o sequences co esponding o Bi idobac e ium spp. in he colonic con en o mice ea ed wi h he EPS-p oducing s ains B. animalis subsp. lac is DSM10140 o S89L, and non- ea ed e e ence g oup. The ea men p ocedu e is desc ibed in he legend o Figu e 3. Fo each combina ion ea men day/s ain, he “box and whiske s” plo ep esen s median, in e qua ile ange and minimum and maximum alues, calcula ed om de sequences ob ained o a leas 6 mice; he a e age is ep esen ed wi h he whi e s a . Wi hin each ea men day, he non-pa ame ic Mann–Whi ney es o 2-independen samples was used o assess di e ences be ween bo h s ain (*p<0.05). The IBM SPSS S a is ic s25 package was used o he non-pa ame ic s a is ical analysis. he DSM10140 and S89L g oups a e also coinciden wi h he endency obse ed o Bi idobac e ium spp. To comple e axonomic cha ac e iza ion, a en a i e species-le el analysis was also pe o med. Fo his pu pose, s a is ically ele an lac obacilli and bi idobac e ia (>0.1%) de e mined by DESeq2 di e en ial abundance es was ca ied ou wi h he 16S RNA gene sequences. Abundances o B. animalis we e inc eased in bo h ea men s wi h espec o he con ol g oup (Supplemen a y Figu e 8A), bu no s a is ical di e ences we e ound. Cu iously, a g ea inc ease in Lac obacillus eu e i popula ions was achie ed a he end o pos -in e en ion pe iod wi h bo h s ains (Supplemen a y Figu e 8B), co obo a ing he delayed e ec o he mic obio a modula o y ac i i y p e iously obse ed o o he clades. Fu he , he ITS egions we e sequenced o s udy he species belonging o Bi idobac e ium genus (Figu e 6). Adminis a ion o bo h s ains enhanced B. bi idum g ow h compa ed o he con ol g oup, ega dless in e en ion ime (Figu e 6A). On he o he hand, he abundances o B. pseudolongum subsp. globosum and B. pseudolongum subsp. pseudolongum showed a g ea abundance du ing pos -in e en ion in bo h ea men s while B. pseudolongum spp. g ow h signi ican inc eased a 4 days o pos - ea men in S89L g oup (Figu e 6B). I was p e iously epo ed ha animals ed wi h EPS-p oducing bi idobac e ia a e able o inc ease he popula ions o o he bi idobac e ial species (Salaza e al., 2011), which could be explained by he use o he polyme s as e men able subs a es o o he mic obio a inhabi an s, including bi idobac e ia (Salaza e al., 2016). Associa ions Be ween Se um Cy okine Le els and Gu Mic obio a To be e unde s and he po en ial biological e ec o he adminis a ion o S89L opy s ain in mice, se um cy okine ma ke s and mic obial axonomic da a we e in eg a ed h ough co ela ion ne wo ks (Figu e 7) and hie a chical all-agains -all associa ion es ing (HAIIA; Figu e 8). Posi i e and nega i e associa ions be ween axonomic da a and cy okine p o iles we e i s ep esen ed as a co ela ion ne wo k (Figu e 7). As expec ed, Lac obacillaceae amily was posi i e associa ed o Lac obacillus species while di e en species o bi idobac e ia posi i ely associa ed o each o he . In his sense, posi i e ela ionships F on ie s in Mic obiology | www. on ie sin.o g 8No embe 2020 | Volume 11 | A icle 601233 micb-11-601233 No embe 21, 2020 Time: 13:27 # 9 Saba e e al. Mic obio a and Cy okine Modula ion by Bi idobac e ium Exopolysaccha ide FIGURE 6 | Abundance pe cen ages o s a is ically signi ican di e en Bi idobac e ium species a e In e nal T ansc ibed Space (ITS) sequencing acco ding o he EPS-s ain selec ed (A) and bo h EPS-s ain and in e en ion ime (B). be ween Co iobac e iaceae amily and En e o habdus genus a e due o he sha ed class Co iobac e ia. Posi i e ela ionship be ween Tene icu es and Rumicococcaceae UCG-014 as well as be ween Lachnospi aceae amily and In es inimonas we e obse ed. As expec ed, abundances o Fi micu es and Bac e oidales S-24 g oup we e co ela ed in a nega i e way. Wi h ega d o cy okine p o iles, in gene al, hey showed posi i e co ela ions o each o he . The HAIIA analysis (Figu e 8) e ealed ha Lac obacillaceae amily and speci ically L. eu e i we e associa ed o high le els o IL-5. Simila ly, popula ions o B. pseudolongum subspecies and L. eu e i we e associa ed o highe se um le els o IL-27. I is wo h emembe ha he S89L s ain ( opy- EPS) s imula ed he g ow h o hese mic oo ganisms, exe ing a delayed e ec ha ook place du ing he pos -in e en ion ime. Ve ucomic obiaceae amily sligh ly con ibu ed o he elease o IL-17a cy okine and he dec ease o IL-10/IL-17a a io. I has also been obse ed ha he IFN-γ was posi i ely associa ed o B. bi idum, and nega i ely associa ed o B. pseudolongum subsp. globosum, indica ing a di e en modula o y pa e n. Wi h ega d o o he clades, P o eobac e ia, educed du ing pos -in e en ion, is associa ed o a lowe elease o in e leukins IL-9 and IL-27. DISCUSSION A di e en ial modula o y e ec o wo isogenic EPS-p oducing B. animalis subsp. lac is s ains on bo h gu mic obio a and cy okine p oduc ion in a heal hy mice model has been ound. Po en ial ela ionships be ween he di e en mic obial clades F on ie s in Mic obiology | www. on ie sin.o g 9No embe 2020 | Volume 11 | A icle 601233