SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554

www.nature.com/scfientfiffficreports

1

Dfiversfity off gut mficrolora fis requfired ffor the generatfion off B cell wfith regulatory propertfies fin a skfin grafft modelR. Alhabbab1,*,†, P. Blafir1,*, R. Elgueta1,*, E. Stolarczyk2, E. Marks1, P. D. Becker1, K. Ratnasothy1, L. Smyth1, N. Safinfia1, E. Sharfiff-Paghaleh1, S. O’Connell1, R. J. Noelle1, G. M. Lord1, J. K. Howard2, J. Spencer3, R. I. Lechler1,* & G. Lombardfi1,*

B cells have been reported to promote grafft rejectfion through alloantfibody productfion. However,

there fis growfing evfidence that B cells can contrfibute to the mafintenance off tolerance. Here, we

used a mouse model off MHC-class I mfismatched skfin transplantatfion to finvestfigate the contrfibutfion

off B cells to grafft survfival. We demonstrate that adoptfive transffer off B cells prolongs skfin grafft

survfival but only when the B cells were fisolated ffrom mfice housed fin low sterfilfity “conventfional” (CV)

ffacfilfitfies and not ffrom mfice housed fin pathogen ffree ffacfilfitfies (SPF). However, prolongatfion off skfin

grafft survfival was lost when B cells were fisolated ffrom IL-10 deficfient mfice housed fin CV ffacfilfitfies.

The suppressfive ffunctfion off B cells fisolated ffrom mfice housed fin CV ffacfilfitfies correlated wfith an

antfi-finlammatory envfironment and wfith the presence off a dfifferent gut mficrolora compared to

mfice mafintafined fin SPF ffacfilfitfies. Treatment off mfice fin the CV ffacfilfity wfith antfibfiotfics abrogated the

regulatory capacfity off B cells. Ffinally, we fidentfified transfitfional B cells fisolated ffrom CV ffacfilfitfies

as possessfing the regulatory ffunctfion. These findfings demonstrate that B cells, and fin partficular

transfitfional B cells, can promote prolongatfion off grafft survfival, a ffunctfion dependent on lficensfing by

gut mficrolora.

Tere fis a body off evfidence that B cells can contrfibute to allograff rejectfion1–5. In mfice, depletfion off B cells has been shown to delay renal allograff rejectfion, and fin humans the finvolvement off B cells fin pro-motfing graff rejectfion has been suggested by the beneffcfial effects off B cell depletfion therapy (Rfituxfimab) ffor kfidney transplant recfipfients3,6,7. However, there fis now also evfidence to suggest that B cells may have a role fin promotfing tolerance to allograffs. One study usfing Rfituxfimab as finductfion therapy ffor kfidney transplants ffound that the depletfion off B cells led to acute cellular rejectfion fin some patfients, suggestfing that B cells may contrfibute to allograff survfival by restrafinfing allo-fimmune responses8. We have recently reported that fimmunosuppressfive drug ffree transplant patfients who had become spontaneously tolerant to thefir HLA mfismatched kfidney transplants had elevated numbers off perfipheral blood B cells and upregulated expressfion off several genes assocfiated wfith B cell ffunctfion9. Sfimfilarly, Newell et al. have shown that drug ffree tolerant patfients had a hfigher proportfion off transfitfional B cells fin thefir perfipheral blood compared to non-tolerant patfients and sfimfilar levels to healthy controls, results that were con-ffrmed by Sfilva et al.10,11. Further, patfients wfith chronfic graff versus host dfisease ffollowfing allogenefic stem

1Medfical Research Councfil Centre ffor Transplantatfion, Kfing’s College London, Kfing’s Health Partners, Guy’s

Hospfital, London SE1 9RT, UK. 2Dfivfisfion off Dfiabetes and Nutrfitfional Scfiences, Kfing’s College London, Guy’s

Hospfital, London SE1 9RT, UK. 3Peter Gorer Department off Immunobfiology, Kfing’s College London, Guy’s

Hospfital, London SE1 9RT, UK. †Current address: Centre off Immunology, Dfivfisfion off Applfied Medfical Scfiences,

Kfing Abdulazfiz Unfiversfity, KSA. *These authors contrfibuted equally to thfis work. Correspondence and requests

ffor materfials should be addressed to G.L. (emafil: gfiovanna.lombardfi@kcl.ac.uk)

Recefived: 28 October 2014

Accepted: 22 May 2015

Publfished: 25 June 2015

OPEN

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 2

cell transplantatfion have reduced regulatory B cell numbers. However mesenchymal stem cell treatment restores Breg ffunctfion, a ffeature whfich correlates wfith an fimproved clfinfical outcome12,13. In mfice, there are now multfiple reports that fimplficate a regulatory role ffor B cells fin graff survfival14–20.In autofimmune models, two major phenotypes off regulatory B cells have been descrfibed, CD5+CD1dhfi

B10 B cells and CD19+CD21hfiCD23hfiCD24hfi transfitfional-2 (T2) regulatory B cells21. Both have been shown to contrfibute to the suppressfion off finffammatfion fin a number off mouse autofimmune models. For example, adoptfive transffer off T2 B cells can finhfibfit collagen finduced arthrfitfis (CIA) and lupus21. In all these studfies IL-10 was demonstrated to be the major medfiator off thefir ffunctfion22,23. Interestfingly, Shfimumura et al. reported that the degree off sterfilfity fin whfich mfice are housed, could alter the ffunc-tfion off regulatory B cells. B cells could regulate chronfic colfitfis only when the mfice were housed under non-hygfienfic condfitfions24. More recently Rosser et al. demonstrated that regulatory B cells had reduced abfilfity to prevent experfimental arthrfitfis when fisolated ffrom mfice under sterfile specfiffc pathogen ffree (SPF) compared to regulatory B cells fisolated ffrom mfice fin less sterfile conventfional (CV) housfing. Ablatfion off gut mficroffora wfith antfibfiotfics treatment ffurther reduced regulatory B cell abfilfity to finhfibfit arthrfitfis development25.Here, we use a mouse model off MHC-class I mfismatched skfin transplantatfion to finvestfigate whether

sterfilfity off housfing finffuences B cell abfilfity to prolong skfin graff survfival. Adoptfive transffer off B cells fiso-lated ffrom naïve SPF mfice dfid not prolong skfin transplant survfival and thefir lack off regulatory ffunctfion was conffrmed fin vfitro. However, when B cells were fisolated ffrom naïve mfice that had been kept under less hygfienfic conventfional (CV) condfitfions, adoptfive transffer off B cells prolonged skfin graff survfival and suppressed TNF-α productfion by T cells fin vfitro. In contrast, transffer off B cells ffrom IL-10 deffcfient mfice housed fin CV ffacfilfitfies dfid not prolong skfin graff survfival. Also, the regulatory ffunctfion off B cells was abolfished by treatment off the mfice wfith antfibfiotfics. Ffinally, we fidentfiffed that wfithfin total B cells the transfitfional B cell subtypes were the subpopulatfions carryfing the regulatory ffunctfion. Tese results demonstrate that B cells, and fin partficular transfitfional B cells, acqufire a regulatory role when the mfice are kept under non-hygenfic condfitfion and thfis lead to an fincrease fin skfin graff survfival.

ResultsB cells fisolated ffrom mfice kept fin low sterfilfity conventfional ffacfilfitfies prolong MHC-class I mfismatched skfin grafft survfival. It has prevfiously been reported that the adoptfive transffer off B cells suppresses the development off autofimmunfity fin several murfine models22,23,26 and that B cells may have a regulatory ffunctfion fin finducfing graff survfival14–20. In thfis study, the capacfity off B cells to prolong skfin transplant survfival was finvestfigated. B cells (1 × 107) enrfiched by negatfive selectfion (> 95% purfity) ffrom the spleens off naïve C57BL/6 (B6) mfice housed fin hfigh sterfilfity specfiffc pathogen ffree (SPF) ffacfilfitfies were adoptfively transfferred fintravenously (fi.v.) to naïve B6 mfice. One day ffollowfing adoptfive transffer B6 mfice (H-2b) recefived dorsal skfin transplants ffrom MHC I mfismatched transgenfic B6 mfice that express MHC-I H-2Kd (B6-Kd). CD8+ T cells were depleted by fintraperfitoneal (fi.p.) finjectfion off antfi-CD8 antfi-body on day − 1, 0 (day off skfin graff), + 1 and every 7 days ffollowfing skfin transplant, to elfimfinate the contrfibutfion off CD8+ T cells wfith dfirect allospecfiffcfity to graff rejectfion, as prevfiously publfished27. Te results fin Ffig.  1a demonstrate that skfin graff survfival was not prolonged by the adoptfive transffer off B cells fin SPF ffacfilfitfies.In a prevfious publficatfion fit was shown that the abfilfity off B cells to finhfibfit chronfic colfitfis, or experfi-

mental arthrfitfis, depended on the sterfilfity off the ffacfilfitfies wfithfin whfich the anfimals were housed25,28. We thus tested whether B cells fisolated ffrom mfice housed fin less sterfile conventfional (CV) ffacfilfitfies were able to prolong graff survfival. Ffigure 1b demonstrates that adoptfive transffer off 107 B cells fisolated ffrom mfice housed fin CV ffacfilfitfies sfignfiffcantly prolonged the survfival off B6-Kd skfin graffs compared to con-trol mfice. Te prolongatfion off skfin graff survfival observed when B cells were transfferred was off sfimfilar extent to the transffer off 5 × 106 graff-specfiffc regulatory T cells (Tregs) fin the same strafin combfinatfion27.Gfiven that the bacterfial component lfipopolysaccharfide (LPS) has been suggested to be fimportant ffor

the generatfion off regulatory B cells29,30, and that alterfing the sterfilfity off housfing may affect the amount off LPS to whfich B cells are exposed fin vfivo, we finvestfigated whether fincubatfion off SPF B cells wfith LPS could restore thefir regulatory ffunctfion. B cells ffrom mfice housed fin the SPF ffacfilfity were pre-fincubated fin vfitro wfith LPS ffor 16 hours beffore adoptfive transffer. Ffigure 1c shows that adoptfive transffer off 107 LPS treated SPF fisolated B cells to B6 mfice kept fin SPF ffacfilfitfies was able margfinally to delay graff rejectfion off B6-Kd skfin graffs compared to control mfice, however the dfifference dfid not reach statfistfical sfignfiffcance. Tfis result suggests that fincreased exposure to LPS stfimulatfion alone does not explafin the enhanced regulatory ffunctfion dfisplayed by B cells fisolated ffrom CV ffacfilfitfies and that other ffactors are finvolved.IL-10 has been shown to be the key cytokfine produced by regulatory B cells fin autofimmune mod-

els21,22. However, fin anfimal models off graff rejectfions the role off IL-10 produced by B cells fin prolongfing graff survfivals has been more controversfial16,18–20,31. To test dfirectly whether IL-10 plays any role fin the regulatory ffunctfion off B cells, B cells were fisolated ffrom IL-10 deffcfient mfice housed fin efither CV ffacfilfitfies (Ffig. 1d) or fin SPF ffacfilfitfies (Ffig. 1e) and thefir abfilfity to prolong graff survfival fin efither ffacfilfity was compared to B cells ffrom WT mfice. Prolongatfion off skfin graff survfival was not observed ffollowfing transffer off IL-10−/− B cells (Ffig. 1d,e) fisolated ffrom mfice kept fin efither ffacfilfity.Tese results fin Ffig.  1d,e suggest that IL-10 productfion by B cells fis fimportant ffor the B cell medfi-

ated prolongatfion off skfin graff survfival observed fin CV ffacfilfitfies. However the complete lack off IL-10 fin

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 3

Ffigure 1. Adoptfive transffer off B cells fisolated ffrom mfice kept fin CV ffacfilfitfies can prolong MHCI mfismatched skfin graff survfival. CD8 depleted B6 (H2-Kb) mfice recefived dorsal skfin graffs ffrom B6-Kd (H2-Kd) mfice. One day prfior to skfin graffs mfice recefived fi.v. finjectfions off 1 × 107 B cells that were FACS purfiffed ffrom naïve B6 mfice. Controls recefived PBS. Plots show skfin graff survfival when graff and cell transffer are perfformed on mfice kept fin (a) SPF ffacfilfitfies (n ≥ 9/group) and (b) CV ffacfilfitfies. (n ≥ 11/group). (c) B cells ffrom naïve B6 mfice mafintafined fin SPF ffacfilfitfies were treated (LPS treated) or not (Untreated) wfith 2 µ g/mL LPS ffor 16 hrs. Ten, 1 × 107 B cells were adoptfively transffer one day prfior to skfin graffs, controls recefived PBS. Te mfice were kept fin SPF ffacfilfitfies (n ≥ 5/group). (d,e) Injectfions off 1 × 107 B cells that were FACS purfiffed ffrom IL-10−/− or WT naïve B6 mfice kept fin CV (d) or SPF (e) ffacfilfitfies. Controls recefived PBS. Te adoptfive transffer was made one day prfior to skfin graffs. Plots show skfin graff survfival when graff and cell transffer are perfformed on mfice kept fin CV (d) or SPF (e) ffacfilfitfies (n ≥ 6/group ffor each group). (ff,g) Plots show skfin graff survfival when cell transffer off 107 B cells ffrom mfice treated wfith antfi-IL10R (aIL-10R) antfibody or fisotype control (ISO) fis perfformed on mfice kept fin CV (ff) or SPF (g) ffacfilfitfies (n ≥ 6/group ffor each group). Mean survfival fin days fis denoted between parentheses. Statfistfics were calculated by log-rank (Mantel-Cox) test + Bonfferonnfi correctfion ffor multfiple comparfisons. *p < 0.05, **p < 0.01.

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 4

IL-10-deffcfient B cell donor mfice mfight fin ffact be finhfibfitfing the development off regulatory B cells. To finvestfigate thfis possfibfilfity, IL-10 competent B6 mfice housed fin CV (Ffig.  1ff) or SPF (Ffig.  1g) ffacfilfitfies were finjected wfith a neutralfizfing antfibody specfiffc ffor the IL-10 receptor (aIL-10R) or wfith the fisotype antfibody (ISO) ffor two weeks. B cells were then purfiffed ffrom these anfimals and transfferred to mfice the day beffore recefivfing skfin graffs. We observed that the treatment off mfice wfith aIL-10R antfibody abolfished the prolongatfion off skfin transplant survfival obtafined wfith B cells derfived ffrom the ISO group fin CV ffacfilfitfies (Ffig.  1ff). Blockade off IL-10R dfid not alter the lack off regulatory ffunctfion observed fin B cells fisolated ffrom mfice kept fin SPF ffacfilfitfies (Ffig. 1g). Tese results together suggest that IL-10 appears to be necessary ffor the development off regulatory B cells fin vfivo, however fit does not excluded the possfibfilfity that IL-10 fis also necessary ffor B cell regulatory ffunctfion.

Mfice housed under non-hygfienfic condfitfions have an actfivated adaptfive fimmune system. Next, we finvestfigated whether there were any fimmunologfical dfifferences between mfice kept fin CV and SPF ffacfilfitfies that mfight explafin the dfifference fin the capacfity off B cells to prolong skfin graff survfival. Spleens and lymph nodes (LN) were collected ffrom mfice mafintafined fin both ffacfilfitfies, and the pro-portfions off B and T cell subsets were analysed ex vfivo. No dfifferences were ffound fin total splenfic B cells, however, there was a lower percentage off total B cells fin the LN off mfice mafintafined fin CV ffacfilfity compared to mfice mafintafined fin the SPF ffacfilfity (Ffig. S1a and S1b). Addfitfionally, CV mfice dfisplayed sfignfiffcantly hfigher percentages off total CD4+, memory (CD62LhfiCD44+) CD4+ and central memory CD8+ T cells fin perfipheral LN compared to SPF mfice, whfile lower percentages off CD8+ T cells were detected fin the spleens off B6 mfice mafintafined fin SPF ffacfilfity compared to mfice ffrom the CV ffacfilfity were observed (Ffig. 2a,b). Tese results suggest an alteratfion fin the B cell:T cell ratfio between the LN off anfimals housed fin CV and SPF ffacfilfitfies.Splenocytes, LN, mesenterfic LN (MLN) and Peyer’s patch (PP) cells were stafined ffor the expressfion

off CD19, GL-7 and CD95 and germfinal centre (GC) B cells were fidentfiffed by ffow cytometry. As shown fin Ffig.  2c mfice housed fin CV ffacfilfitfies had sfignfiffcantly hfigher ffrequencfies and absolute numbers off GC B cells fin the MLNs and PPs compared to mfice kept fin SPF ffacfilfitfies. Moreover, there were hfigher ffrequencfies and absolute numbers off GC B cells fin the spleens off mfice housed fin CV ffacfilfitfies, however the dfifferences were not statfically sfignfiffcant (Ffig.  2c). Te presence off hfigher numbers off GCs fin the spleens off CV mfice compared to SPF mfice was conffrmed by conffocal mficroscopy as Ffig. 2d shown.Havfing demonstrated dfifferences fin the percentages off memory T cells and fin the levels off GC B cells

between mfice kept fin the CV and SPF, cytokfine expressfion by B and T cells was finvestfigated. IFN-γ , IL-10 and TNF-α expressfion by splenfic B and T cells fisolated ffrom CV and SPF mfice was assessed by fintracellular stafinfing ffollowfing 6 hrs stfimulatfion wfith PMA and fionomycfin. Te results fin Ffig. 2e show that B cells fisolated ffrom mfice mafintafined fin CV ffacfilfitfies expressed sfignfiffcantly lower levels off IFN-γ and TNF-α compared to those fisolated ffrom SPF ffacfilfitfies, wfith no dfifferences fin the productfion off IL-10. T cells obtafined ffrom mfice kept fin CV ffacfilfitfies expressed sfignfiffcantly hfigher levels off IL-10 com-pared to those fisolated ffrom mfice mafintafined fin SPF ffacfilfitfies wfith no major dfifferences fin TNF-α and IFN-γ (Ffig. 2e). Tese results suggest that mfice housed fin less hygfienfic condfitfions have a bfias towards a more antfi-finffammatory mficroenvfironment fin the spleen and that both B cells and T cells ffrom mfice mafintafined fin the CV ffacfilfitfies have a trend to an hfigher IL-10/TNF-α ratfio compared to the same cell types ffrom mfice housed fin the SPF ffacfilfitfies.

Mfice kept fin CV ffacfilfitfies have an altered composfitfion off gut mficrobfiota compared mfice housed fin the SPF ffacfilfitfies. Prevfious reports have shown that the presence off commensal mficro-bfiota affects the development off fimmune responses and can protect humans ffrom atopfic dfiseases32,33. In addfitfion, gut mficrobfiota have been shown to finffuence B cell secretfion off fimmunoglobulfin, and support the development off splenfic regulatory B cells, suggestfing that the presence off bacterfia fin the gut has a measurable fimpact on B cell ffunctfion25,34. Havfing fidentfiffed dfifferences fin the general actfivatfion status off the B and T cells off the fimmune system between mfice housed fin the SPF and CV ffacfilfitfies, and fin the regulatory potentfial off B cells ffrom CV and SPF mfice, we analysed ffaecal samples to determfine whether these dfifferences were assocfiated wfith dfifferences fin thefir gut mficrobfiota. Te data shown fin Ffig.  3a demonstrates that the composfitfion off the gut mficrobfiota was sfignfiffcantly dfifferent between mfice housed fin CV and SPF ffacfilfitfies. Mfice mafintafined fin CV ffacfilfitfies dfisplayed a reductfion fin the proportfions off Lachnospfiraceae (Erec 482+) and Bfiffdobacterfium (Bfiff 164+) groups compared to mfice mafintafined fin SPF ffacfilfitfies. To finvestfigate the role off bacterfia fin the finductfion off regulatory B cell ffunctfion, mfice housed fin CV ffacfilfitfies were treated ffor 2 weeks wfith broad-spectrum antfibfiotfics. Te effcacy off the antfibfiotfic treatment was corroborated by a reductfion off ffaecal mficroffora content by 70% fin the colonfic content off antfibfiotfic treated compared to non-treated mfice (Ffig. 3b). It has been prevfiously reported that caecal sfize fincreases fin response to augmented hydrated dfietary ffbre components ffollowfing a reductfion fin the bacterfial load35. Here, mfice treated wfith antfibfiotfics dfisplayed a marked fincrease fin caecum sfize conffrm-fing the effcacy off the treatment (Ffig. 3c). We thus tested whether B cell fisolated ffrom antfibfiotfic treated mfice housed fin the CV ffacfilfitfies were able to prolong graff survfival. Ffigure 3d shows that, unlfike adoptfive transffer off control B cells ffrom non-antfibfiotfic treated mfice, adoptfive transffer off 107 B cells fisolated ffrom antfibfiotfic treated mfice housed fin CV ffacfilfitfies was unable to prolong the survfival off B6-Kd skfin graffs.

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 5

Ffigure 2. Immunologfical dfifferences between mfice kept fin SPF & CV ffacfilfitfies. Spleens, lymph nodes & PPs were collected ffrom B6 mfice (6 weeks old) mafintafined fin SPF and CV ffacfilfitfies, and phenotyped ffor B cell and T cell populatfions by usfing antfibodfies agafinst CD4, CD8, CD25, CD19, GL7, CD38, CD95, CD44, & CD62L. (a) and (b) Hfistograms show mean percentage ± SEM CD4+, CD8+ and CD4+CD25+ T cells fin total splenocytes or total lymph node cells, naïve (CD64LhfiCD44-) and memory (CD62LhfiCD44+) CD4+ T cells as percentages off total CD4+ T cells, and naïve and memory CD8+ T cells as percentages off total CD8+ T cells fin (a) Spleens and (b) Lymph nodes. (c) Hfistograms show the mean ± SEM off the percentages off GC B cells (GL7hfigh and CD95+, leff graph) and absolute number (rfight graph) fin the spleens, MLN and PPs off SPF and CV mfice. (d) Immunoffuorescence mficroscopy pfictures off B220 and IgD stafined ffrozen sectfions ffrom spleens (top row) and PPs (bottom row) off mfice kept fin SPF and CV ffacfilfitfies. Pfictures show B220 (green) and IgD (Red). Germfinal centres are marked by arrows. (e) Hfistograms show mean ± SEM off IFNγ , TNFα , and IL-10 expressfion by splenfic CD4+ T cells and B cells ffollowfing 5 hrs fin vfitro stfimulatfion wfith PMA & fionomycfin. n = 3. Statfistfics were calculated by t test, *P < 0.05 & **P < 0.005.

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 6

Ffigure 3. Mfice kept fin CV ffacfilfity have dfifferent composfitfion off mficrobfiota than mfice housed fin SPF ffacfilfity. Faecal samples were collected ffrom B6 mfice kept fin SPF & CV ffacfilfitfies, or kept fin CV ffacfilfity and treated wfith antfibfiotfics ffor 2 weeks. Dfifferent bacterfial groups (Bac = Bacterofides and Prevotella, Erec = Lachnospfiraceae, Bfiff = Bfiffdobacterfium, Lab = Lactobacfillus & Streptococcus) were fidentfiffed by FISH-Flow. Hfistograms show mean ± SEM ffor (a) proportfions off the domfinant ffecal mficrobfiota fin B6 mfice mafintafined fin SPF and CV ffacfilfitfies, and (b) ratfio off total mficrobfiota fin antfibfiotfic treated (ATB) and non-treated (CV) B6 mfice mafintafined fin CV ffacfilfitfies (the mean total bacterfia fin untreated mfice fis deffned as 100%). (c) Pfictures off caeca recovered ffrom antfibfiotfics treated and non-treated B6 mfice mafintafined fin CV ffacfilfitfies. (d) CD8 depleted B6 (H2-Kb) mfice recefived dorsal skfin graffs ffrom B6-Kd (H2-Kd) mfice. One day prfior to skfin graffs mfice recefived fi.v. Injectfions off 1 × 107 B cells that were FACS purfiffed ffrom antfibfiotfics (ATB) or control (CV) naïve B6 mfice kept fin CV ffacfilfitfies. Plots show skfin graff survfival when graff and cell transffer are perfformed on mfice kept fin SPF ffacfilfitfies. n ≥ 6/group ffor each group. Mean survfival fin days fis denoted between parentheses. (e) Hfistograms show mean ± SEM off IFNγ , TNFα , and IL-10 expressfion by splenfic CD4+ T cells and B cells ffrom mfice treated wfith antfibfiotfics (ATB) or non-treated (CV) naïve B6 mfice kept fin CV ffacfilfitfies, 2 findependent experfiments wfith 10 mfice per group. (ff) B cells were fisolated ffrom spleens off B6 mfice treated wfith antfibfiotfics or not and mafintafined fin CV ffacfilfitfies. B cells were co-cultured wfith CD4+ T cells and firradfiated allo-DCs (25 CD4 T cells:25 B cells:1 allo-DC) ffor 48 hrs. Summary data showfing percentage off TNFα supressfion by B cell ffrom spleens off B6 mfice antfibfiotfics treated or non-treated and housed fin CV ffacfilfitfies. Hfistograms dfisplay mean ± SEM, 2 findependent experfiments wfith 10 mfice per group. Statfistfics were calculated by log-rank (Mantel-Cox) test + Bonfferonnfi correctfion ffor multfiple comparfisons and t test, *P < 0.05, **P < 0.005 & ***P < 0.001.

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 7

All these results together suggest that the fintestfinal mficrobfiota fis essentfial to mafintafin the capacfity off B cells to delay skfin graff rejectfion.We also evaluated whether the antfibfiotfic treatment had an effect on the cytokfine productfion off B

and T cells. IFN-γ , IL-10 and TNF-α expressfion by splenfic B and T cells fisolated ffrom CV housed mfice treated wfith antfibfiotfics, or untreated controls, was assessed by fintracellular stafinfing ffollowfing 4 hrs stfimulatfion wfith PMA and fionomycfin. Te results fin Ffig. 3e show that T cells fisolated ffrom mfice treated wfith antfibfiotfics expressed sfignfiffcantly hfigher levels off IFN-γ compared to those fisolated ffrom control mfice. Both, B and T cells obtafined ffrom mfice treated wfith antfibfiotfics expressed sfignfiffcantly lower levels off IL-10 compared to those fisolated ffrom control mfice (Ffig. 3e). To determfine whether the dfifferences fin cytokfine productfion between B cells ffrom antfibfiotfics treated and untreated mfice reffected a dfifference fin regulatory capacfity, B cell subsets were tested fin vfitro ffor the abfilfity to suppress allo-specfiffc T cell actfivatfion. B cells have been shown prevfiously fin autofimmune models to suppress cytokfine productfion by CD4+ T cells fin vfitro, fin partficular the expressfion off TNF-α 23. B cell were purfiffed ffrom the spleens off CV ffacfilfity housed B6 mfice that had efither been treated wfith antfibfiotfics or ffrom untreated controls, and co-cultured wfith CD4+ T cells ffrom spleen off SPF ffacfilfity housed B6 mfice, that had been stfimulated by firradfiated H2-Kd expressfing allogenefic dendrfitfic cells (allo-DCs) ffor 48 hours. Our results show that ffollowfing stfimulatfion off CD4+ T cells wfith allo-DC, B cells ffrom the spleens off B6 mfice treated wfith antfibfiotfics lost the abfilfity to suppress TNF-α expressfion by CD4+ T cell (Ffig. 3ff). Tereffore, we suggest that mficrobfiota fis finvolved dfirectly fin empowerfing B cells to suppress T cell responses.

Transfitfional B cells fisolated ffrom mfice kept fin low sterfilfity conventfional ffacfilfitfies prolong MHC-class I mfismatched skfin grafft survfival and suppress allospecfific TNF-α expressfion by CD4+ T cell fin vfitro. Next, we finvestfigated whether a partficular subpopulatfion off B cells was respon-sfible ffor prolongfing skfin graff survfival. No dfifferences were ffound fin the proportfions off the varfious B cell subsets fin the spleens off mfice mafintafined fin SPF and CV ffacfilfitfies (Ffig. S1c). To test the abfilfity off findfivfidual B cell subsets to prolong graff survfival, transfitfional-1 (T1), transfitfional-2 (T2), ffollficular (FO) and margfinal zone (MZ) B cells were sorted by ffow cytometry ffrom the spleens off naïve C57BL/6 (B6) mfice housed SPF ffacfilfitfies (Ffig. S2) and 1 × 106 cells off each subset was adoptfively transfferred fintrave-nously (fi.v.) to naïve B6 mfice. One day ffollowfing adoptfive transffer B6 mfice (H-2b) recefived dorsal skfin transplants as descrfibed fin Ffig. 1. As shown fin the Ffig. 1a, none off the B cell subsets ffrom mfice housed fin SPF condfitfions were able to prolong skfin graff survfival (Ffig. 4a). We then tested whether any B cell subsets fisolated ffrom mfice housed fin CV ffacfilfitfies were able to prolong graff survfival. Ffigure 4b shows that adoptfive transffer 1 × 106 T2 B cells, or T1 B cells, sfignfiffcantly prolonged the survfival off B6-Kd skfin graffs compared to control mfice when B cell donor were ffrom mfice housed fin CV condfitfions.To ffurther conffrm the dfifferences fin the regulatory capacfity between transfitfional B cells fisolated ffrom

mfice housed fin the SPF and CV ffacfilfitfies, B cell subsets were tested fin vfitro ffor the abfilfity to suppress allo-specfiffc T cell actfivatfion. B cell subsets were sorted ffrom the spleens off B6 mfice, mafintafined fin the SPF and CV ffacfilfitfies and co-cultured wfith CD4+ T cells stfimulated by firradfiated H2-Kd expressfing allogenefic dendrfitfic cells (allo-DCs) ffor 48 hours. Our results show that ffollowfing stfimulatfion off CD4+ T cells wfith allo-DC, TNF-α expressfion (Ffig.  4c,d) and IFN-γ (Ffig. S3a) by T cells and expressfion off CD44 and CD69 molecules (Ffig. S3a) were not altered by any B cell subset ffrom mfice mafintafined fin SPF ffacfilfitfies. In contrast, when B cell subsets were fisolated ffrom the spleens off B6 mfice mafintafined fin CV ffacfilfitfies, T2 B cells and T1 B cells suppressed TNF-α expressfion by CD4+ T cell by around 40% (although ffor T1 B cells the dfifferences dfid not reach statfistfical sfignfiffcance), (Ffig. 4e,ff). Te productfion off IFN-γ and the upregulatfion off CD44 and CD69 were mfinfimal and no dfifferences were observed (Ffig. S3b). No other B cell subset suppressed allo-specfiffc TNF-α expressfion by CD4+ T cells. In terms off the cytokfines produced by the dfifferent B cell subsets derfived ffrom CV ffacfilfity we observed that IL-10 expressfion by B cells and suppressfive capacfity dfid not correlate, as MZ B cells expressed more IL-10 than other B cell subsets even though they dfid not suppress T cell TNF-α expressfion (Ffig. 4g). A sfimfilar IL-10 proffle was observed fin the dfifferent B cell subsets derfived ffrom mfice house fin the SPF ffacfilfity (data not shown). Tereffore, we suggest that IL-10 produced by B cells fis only partfially finvolved to supress the T cell response as we have prevfiously demonstrated20.Next, we evaluated whether treatment off CV housed mfice wfith antfibfiotfics altered the fin vfitro sup-

pressfive capacfity off the dfifferent B cell subsets. Ffigure 4h shows that IL-10 levels were drastfically reduced fin the supernatants off co-cultures between CD4+ T cells, ffrom spleen off mfice housed fin SPF ffacfilfitfies, and T2, or T1, B cells fisolated ffrom CV housed mfice treated wfith antfibfiotfics compared wfith B cells fisolated ffrom control untreated mfice (Ffig. 4h). Tese results correlated wfith the reductfion fin the IL-10+ B cells ffound ex-vfivo and fin the percentages off suppressfion fin vfitro (Ffig. 3e,ff). Tese results suggest that antfibfiotfic treatment reduces the abfilfity off T2 and T1 B cells to finduce regulatory cytokfine productfion fin finteractfions wfith CD4+ T cells.

DfiscussfionIn thfis study we have demonstrated that B cells acqufire the capacfity to prolong skfin graff survfival when the mfice are housed under non-hygfienfic condfitfions. We have shown that B cell regulatory ffunc-tfion and abfilfity to prolong allograff survfival correlate wfith the presence off gut mficrobfiota and thus, antfibfiotfics-treatment abolfished the regulatory ffunctfion off B cells. In addfitfion, we observed that mfice kept

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 8

Ffigure 4. Adoptfive transffer off T2 B cells fisolated ffrom mfice kept fin CV ffacfilfitfies can prolong MHC I mfismatched skfin graff survfival. CD8 depleted B6 (H2-Kb) mfice recefived dorsal skfin graffs ffrom B6-Kd (H2-Kd) mfice. One day prfior to skfin graffs mfice recefived fi.v. Injectfions off 1 × 106 T2, T1, FO, or MZ B cells that were FACS purfiffed ffrom naïve B6 mfice. Controls recefived PBS. Plots show skfin graff survfival when graff and cell transffer are perfformed on mfice kept fin (a) SPF ffacfilfitfies (n ≥ 6/group ffor each experfiment) and (b) CV ffacfilfitfies. (n ≥ 8/group ffor each experfiment). Mean survfival fin days fis denoted between parentheses. Statfistfics were calculated by log-rank (Mantel-Cox) test + Bonfferonnfi correctfion ffor multfiple comparfisons. (C–G) B cells were fisolated ffrom spleens off B6 mfice mafintafined fin SPF (c,d) or CV (E-G) ffacfilfitfies by magnetfic sortfing. B cell subsets were purfiffed by FACS and co-cultured wfith negatfively fisolated CD4+ T cells and firradfiated allo-DCs (25 CD4 T cells:25 B cells:1 allo-DC) ffor 48 hrs. PMA, Ionomycfin and breffeldfin A were added ffor the last 4 hours off culture. (c and e) Representatfive FACS plots off CD4+ T cell TNF-α expressfion, (d and ff) summary data showfing CD4+ T cells TNF-α expressfion, (g) summary data showfing IL-10 expressfion by B cell subsets fisolated ffrom spleens off B6 mfice housed fin CV ffacfilfity. Hfistograms dfisplay mean ± SEM, (n = 3). (h) Summary data showfing IL-10 secretfion fin cultures off T cell alone and wfith each subset off B cells ffrom mfice treated wfith antfibfiotfics (ATB) or salfine solutfion (CV) and kept fin CV ffacfilfitfies. Statfistfics were calculated by one-way ANOVA and Bonfferronfi post-test, ns- not sfignfiffcant, *P < 0.05, **P < 0.005, ***P = 0.0005.

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 9

fin CV ffacfilfitfies had an fincrease fin memory T cells fin the LN(s), a hfigher numbers off GC B cell and an apparent bfias towards an antfi-finffammatory mficroenvfironment. Furthermore, we observed that B cells derfived ffrom IL-10−/− mfice dfid not prolong skfin graff survfival. Ffinally, we provfide evfidence that among the dfifferent B cell subsets the regulatory ffunctfion fis fin the transfitfional B cell subsets.It has been well demonstrated by us and others that adoptfive transffer off Tregs has the capacfity to

finduce transplant tolerance27,36–39. We have prevfiously shown that the adoptfive transffer off 5 × 106 Tregs wfith findfirect specfiffcfity ffor the graff, confferred by TCR transductfion, to C57BL/6 mfice recefivfing a MHC-class I mfismatched skfin graffs has the capacfity to prolong skfin graff survfival ffor 9 days27. In the experfimental model presented here usfing the same strafin combfinatfion, the adoptfive transffer off 107 autologous B cells, and as ffew as 1 × 106 T2 and T1 B cells, was capable off finducfing a sfignfiffcant pro-longatfion off B6-Kd skfin graff survfival compared to controls (> 4 days). Tese results suggest that the adoptfive transffer off B cells fis almost as powerfful as the adoptfive transffer off alloantfigen specfiffc Tregs ffor prolongfing skfin transplant survfival.Recently, fit has been reported that the adoptfive transffer off 107 TIM-1+ B cells was capable off pro-

longfing the survfival off pancreatfic fislet allograffs fin B cell deffcfient (JHD mfice) transplant recfipfients14. In thfis model TIM-1+ B cells were generated by efither treatment off donor mfice wfith antfi-TIM-1 antfi-body or by the presence off an allograff, however the adoptfive transffer off TIM-1+ B cells fisolated ffrom naïve mfice dfid not prolong fislet graff survfival14. It fis possfible that fiff TIM-1+ B cells were fisolated ffrom naïve mfice kept fin non-sterfile condfitfions then they may possess regulatory potentfial wfithout prfior fin vfivo actfivatfion. Low level actfivator sfignals assocfiated wfith the presence off commensal mficrobfiota may play a sfimfilar role to the finffammatfion assocfiated wfith the presence off an allograff and lficense B cells to achfieve a regulatory potentfial. It has been prevfiously reported that B cells must be actfivated to have ffull regulatory potentfial21,26.Te fimportance off envfironmental mficroorganfisms ffor B cell ffunctfion fin vfivo was demonstrated by a

recent study where B cells fisolated ffrom the spleens and PPs off mfice kept fin a CV ffacfilfity secreted hfigher levels off antfibody than mfice kept fin germ ffree ffacfilfitfies34. Sfimfilarly, we observed that housfing mfice under non-hygfienfic condfitfions resulted fin fincreased number off GC B cells, changes to the cytokfine productfion off B cells ffollowfing restfimulatfion fin vfitro (Ffig. 2e). Tus, our data supports the fidea that the presence off envfironmental mficroorganfisms can shape the murfine B cell response.Te fimportance off mficrobfiota ffor the development off B cells wfith regulatory ffunctfion fin our study

agrees wfith the prevfious reports off Shfimomura et al. and Rosser et al.25,28 lfinked perfitoneal CD5− B-1 B cell medfiated protectfion ffrom colfitfis to the housfing condfitfions off the mfice, whfile Rosser et al. recently demonstrated that the presence off gut mficrobfiota fis crucfial ffor the development off splenfic regulatory T2-MZP Bregs25,28. Here we extend these results and demonstrate that the gut mficrobfiota are also crucfial ffor generatfing T2 regulatory B cell abfilfity to suppress allo-fimmune responses. Interestfingly, Shfimomura et al. ffound no role ffor IL-10 productfion by thefir proposed perfitoneal regulatory B cells whfile Rosser et al. ffound fit was necessary ffor splenfic regulatory B cell ffunctfion.Prevfious studfies fin autofimmune models have provfided evfidence ffor the role off IL-10 fin the ffunctfion

off regulatory B cells. In our study, we ffound no dfifference fin IL-10 productfion between B cells fisolated ffrom CV and SPF mfice fin vfitro nor dfid we ffnd a correlatfion between B cell subset IL-10 productfion and regulatory capacfity. However, although ffrequency off IL10+ B cells ffrom CV and SPF mfice are the same, the percentages off TNF-α + B cells are hfigher fin cells derfived ffrom the SPF compared to CV ffacfilfitfies, suggestfing fit fis the ratfio off pro-finffammatory to antfi-finffammatory cytokfine productfion that may be fimportant fin decfidfing whether B cell transffer delays graff rejectfion or not (Ffig.  1ff). Tfis hypothesfis fis supported by the recent results fin human renal transplant recfipfients fin whfich the authors demonstrate that the ratfio off IL-10 and TNF-α expressfion fis more fimportant than IL-10 expressfion alone ffor meas-urfing the ffunctfion off regulatory B cells40.A dfirect role ffor IL-10 fin thfis skfin transplant model was demonstrated by the finabfilfity off B cells fiso-

lated ffrom IL-10−/− mfice, or mfice pretreated wfith blockfing antfi-IL-10R antfibody, to prolong skfin graff survfival on adoptfive transffer (Ffig. 1d,ff). Tese results suggest that, whfile IL-10 expressfion alone fis not suffcfient to fidentfiffy regulatory B cells, fits productfion may potentfially be fimportant ffor thefir development fin naïve mfice and thefir subsequent abfilfity to prolong skfin graff survfival. In transplant models the role played by B cell IL-10 productfion fis stfill controversfial9,10,14,41. Whfile the fimportance off IL-10 expressfion to the abfilfity off TIM-1+ B cells to prolong pancreatfic fislet allograff survfival has been demonstrated14 other studfies have been unable to fidentfiffy a role ffor B cell IL-10 productfion fin transplant tolerance9,10,41. In humans, three recent publficatfions that reported an assocfiatfion off B cells wfith tolerance to kfidney transplants fin humans dfid not report an assocfiatfion wfith IL-10 mRNA transcrfiptfion9,10,41. In mfice, stud-fies off the capacfity off antfi-CD45RB monotherapy to finduce tolerance to cardfiac allograffs reported that, whfile thfis tolerance protocol was dependent on the presence off B cells, IL-10 productfion by B cells was counter-regulatory and that IL-10 neutralfizatfion fimproved graff outcome42.Some off the sfignals that medfiate the effect off mficrobfiota on regulatory B cell ffunctfion have been

recently fidentfiffed25. IL-1β and IL-6 produced by monocytes and dendrfitfic cells has been suggested to be responsfible ffor mficrobfiota medfiated regulatory B cell finductfion; the presence off gut mficrobfiota no longer has the abfilfity to generate T2-MZP Bregs fin vfivo fiff the B cells specfiffcally do not express efither off these receptors25. TLR lfigatfion by mficrobfial antfigens ffor finstance may also play a role; fit has been reported that TLR agonfists can dfirectly stfimulate B cells to produce cytokfines and antfibodfies43, and TLR lfigands have

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 10

been suggested to be the major stfimulfi finvolved fin the actfivatfion off B10 regulatory B cells21,26. However, fin our study LPS-treatment off B cells derfived ffrom the SPF ffacfilfity dfid not finduce regulatory ffunctfion fin B cells. It fis possfible that other cellular target off LPS, or other ffactors together wfith LPS, can contrfibute to the acqufisfitfion by B cells off regulatory ffunctfion.In conclusfion, the data presented here findficate that bacterfia could shape B cell regulatory ffunctfion.

Tfis fis consfistent wfith the ‘hygfiene hypothesfis’ that lfinks abfilfity to tolerate allergens to prfior exposure to a spectrum off envfironmental antfigens33. We suggest that abfilfity to tolerate allograffs may be sfimfilarly enhanced by regulatory B cells that have been condfitfioned by bacterfial antfigens or metabolfites derfived ffrom the fintestfinal mficrobfiota. More work fis needed to understand the mechanfisms behfind the outcome observed fin our model.

Materfial & MethodsEthfics Statement. Tese studfies were approved and conducted fin accredfited ffacfilfitfies fin accord-ance wfith Te Home Offce UK Anfimals (Scfientfiffc Procedures) Act 1986 (Home Offce lficense number PPL 70/7302).

Mfice. Wfild type (WT) C57BL/6 (H-2b) mfice and BALB/c (H-2d) mfice were purchased ffrom Harlan Laboratory. B6Kd mfice were a generous gfiff ffrom Dr. R Pat Bucy (Unfiversfity off Alabama at Bfirmfingham, Bfirmfingham, AL, USA). IL-10−/− mfice were kfindly provfided by Anne O’Garra (NIMR, London). Mfice were mafintafined fin specfiffc pathogen ffree (SPF) or conventfional (CV) ffacfilfitfies. In some exprerfiments, mfice house fin CV ffacfilfity were finjected every 3 days ffor 2-3 weeks wfith 300 µ g/mouse wfith antfi-mouseIL-10R (clone 1B1.3A, BfioXcell). In other experfiments, mfice kept fin CV ffacfilfitfies were treated wfith antfibfiot-fics as descrfibed44. Alteratfions are as noted: ampficfillfin (0.5 g/l; Roche, UK), vancomycfin (0.5 gm/l; Sfigma, UK), neomycfin sulffate (0.5 g/l; Sfigma, UK), and metronfidazole (0.5 g/l; Sfigma) were dfissolved fin ffltered drfinkfing water and ffufid fintake was monfitored. Tese studfies and experfimental protocols were approved and conducted fin accredfited ffacfilfitfies fin accordance wfith Te Home Offce UK (Scfientfiffc Procedures) Act 1986 (Home Offce lficense number PPL 70/7302) and Kfing’s College London.

Antfibodfies and Flow Cytometry. All FACS antfibodfies were purchased ffrom eBfioscfience. For fimmu-nohfistochemfistry FITC-conjugated antfi-B220 mAb and bfiotfinylated antfi-IgD were purchased ffrom eBfio-scfiences streptavfidfin-conjugated alexaffuor594 was purchased ffrom Molecular Probes. Surfface stafinfings were perfformed as prevfiously descrfibed45. For fintracellular stafinfing, cells were stafined then ffxed/perme-abfilfized accordfing to the eBfioscfience protocol. Permeabfilfized cells were fincubated fin Permwash™ wfith antfi-IL10, antfi-TNF, or approprfiate fisotype control ffor 30 mfins at 4 °C (wfith blockfing antfi-CD16/32). Intracellular stafinfing ffor FOXP3 was perfformed wfith eBfioscfience FOXP3 kfit accordfing to manuffacturer’s finstructfions. For the detectfion off fintracellular cytokfines PMA (50 ng/ml), fionomycfin (1 µ M) and breffel-dfin A were added ffor 5 hrs beffore stafinfing. Cells were acqufired usfing LSRII or Fortessa ffow cytometers (BD Bfioscfiences). Flow Jo soffware was used ffor analysfis.

Purfificatfion off B cells, B cell subsets and CD4+ T cells. B cells were purfiffed ffrom the spleens off IL-10−/− or C57BL/6 mfice by negatfive selectfion usfing CD43 mficrobeads (Mfiltenyfi Bfiotech, 130-049-801) accordfing to manuffacturer’s finstructfions. Purfiffed splenfic B cells were stafined wfith antfi-CD21, antfi-CD24 antfi-CD23. DAPI was added to exclude dead cells. Cells were sorted by BD FACSArfia II (BD Bfioscfiences). CD4+ T cells were fisolated ffrom the spleens off C57BL/6 mfice usfing Invfitrogen Dynabeads® Untouched™ Mouse CD4 Cells kfits, accordfing to manuffacturer’s finstructfions (Invfitrogen, 11416D).

B cell stfimulatfion wfith LPS. B cells ffrom spleen were culture at a densfity off 106/mL wfith or wfithout LPS ffrom E. colfi serotype EH100 (TLR grade) (ALX-581-01-L002) at a ffnal concentratfion off 2 µ g/ml ffor 16 hrs at 37 °C, 5%CO2. Cells were then washed and resuspend at a concentratfion off 5 × 10

7 ml fin salfine. Mfice were finjected fi.v. wfith 107 cells fin 200 µ l.

DC generatfion. DCs were generated ffrom bone marrow as prevfiously descrfibed45. One day beffore usfing the DCs, 100 ng/mL LPS (E. colfi, Enzo Lfiffe Scfiences, UK) was added to finduce maturatfion. At day 7 off culture, DCs were collected, washed three tfimes, and firradfiated.

Skfin transplants. Donor tafil skfin graffs were perfformed and monfitored as prevfiously descrfibed45. Antfi-CD8 antfibody (clone YTS169, 250 µ g/finjectfion/mouse) was finjected fi.p. at day-1 and day 1 affer skfin graff, and weekly thereaffer.

In vfitro suppressfion assays. Isolated CD4+ T cells were co-cultured wfith firradfiated allo-DCs alone at a ratfio off 25 T cells:1 DC, or wfith B cells and allo-DCs at a ratfio off 25 T cells: 25 B cells: 1 DC. Cells were cultured ffor 48 hrs at 37 °C fin RPMI fin 96 well plates (1.5–2 × 105 cells/well).

Immunohfistochemfistry. Tfissues were embedded fin OCT compound (TfissueTek) and snap-ffrozen fin lfiqufid nfitrogen. Frozen sectfions (7 µ m) were ffxed fin 50% acetone ffor 30 seconds and 100% acetone ffor 5 mfin at 4 °C and then afir-drfied at room temperature. Te tfissue sectfions were blocked wfith normal

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 11

horse serum fin PBS (5%) ffor 15 mfin. Te sectfions were stafined ffor 30–60 mfins wfith FITC-conjugated antfi-B220 mAb and bfiotfinylated antfi-IgD, ffollowed by streptavfidfin-conjugated alexaffuor594. Te slfides were then washed fin PBS, mounted wfith ffuorescent mountfing medfium (Vector), vfisualfised on a ffuores-cent mficroscope and analysed usfing mficroscope manager and Image J. All fimages are 10× magnfiffcatfion.

Antfibfiotfic Treatment. Mfice were treated as descrfibed44. Alteratfions are as noted: ampficfillfin (0.5 g/l; Roche, UK), vancomycfin (0.5 gm/l; Sfigma, UK), neomycfin sulffate (0.5 g/l; Sfigma, UK), and metronfida-zole (0.5 g/l; Sfigma) were dfissolved fin ffltered drfinkfing water and ffufid fintake was monfitored.

Fluorescent fin-sfitu hybrfidfizatfion (FISH-Flow). Stool samples were collected and stored at − 80 °C. Fluorescent fin sfitu hybrfidfisatfion (FISH) combfined wfith ffow cytometry was perfformed as prevfiously descrfibed wfith mfinor modfiffcatfions46. Faeces (0.05 g) were homogenfised fin 0.5 ml off sterfile phosphate-buffered salfine (PBS, 130 mM NaCl, 3 mM NaH2PO4 2H2O, 7 mM Na2HPO4 12H2O, pH 7.2) wfith 5 glass beads fin the Tfissue Lyser (Qfiagen, UK) ffor 5 mfin at 25 Hz. One volume off the suspensfion was added to 3 volumes off 4% parafformaldehyde (PFA) fin PBS. Affer overnfight ffxatfion at 4 °C, the ffxed suspensfion fin PFA was stored at − 80 °C untfil hybrfidfisatfion. Te EUB 338 probe was used as the posfitfive control probe, the NON 338 probe as the negatfive control and specfiffc probes ffor fidentfiffcatfion off sub-groups (Ffig. S4). All probes were covalently lfinked at thefir 5′ end efither to ffuorescefin fisothfiocy-anate (FITC) or to the sulffofindocyanfine dye findodficarbocyanfine (Cy5) as findficated (Ffig. S4). Samples were acqufired by ffow cytometry (HTS Fortessa, Becton Dfickfinson, USA). Subsequent analyses were conducted usfing FlowJo soffware (Tree Star, USA).

Statfistfical analysfis. Comparfisons between groups were perfformed usfing T-test ffor two groups, or one-way ANOVA and a Dunnett or Bonfferronfi post-test ffor more than two groups. Log Rank (Mantel-Cox) tests were used ffor graff survfival curves. Analyses were perfformed usfing GraphPad Prfism soffware.

Refferences1. Chong, A. S. & Scfiammas, R. Matchmakfing the B-cell sfignature off tolerance to regulatory B cells. Am. J. Transplant. 11, 2555–2560 (2011).

2. Colvfin, R. B. et al. Emergfing role off B cells fin chronfic allograff dysffunctfion. Kfidney Int. Suppl. S13–7, dofi: 10.1038/kfi.2010.436 (2010).

3. DfiLfillo, D. J. et al. B lymphocytes dfifferentfially finffuence acute and chronfic allograff rejectfion fin mfice. J Immunol 186, 2643–2654 (2011).

4. Hfippen, B. E., DeMattos, A., Cook, W. J., Kew, C. E. & Gaston, R. S. Assocfiatfion off CD20+ finffltrates wfith poorer clfinfical outcomes fin acute cellular rejectfion off renal allograffs. Am. J. Transplant. 5, 2248–2252 (2005).

5. Redffeld, R. R., III et al. Essentfial role ffor B cells fin transplantatfion tolerance. Current Opfinfion fin Immunology 23, 685–691 (2011).

6. Tydén, G. et al. A randomfized, doubleblfind, placebo-controlled, study off sfingle-dose rfituxfimab as finductfion fin renal transplantatfion. Transplantatfion 87, 1325–1329 (2009).

7. Tydén, G., Kumlfien, G. & Fehrman, I. Successfful ABO-fincompatfible kfidney transplantatfions wfithout splenectomy usfing antfigen-specfiffc fimmunoadsorptfion and rfituxfimab. Transplantatfion 76, 730–731 (2003).

8. Clatworthy, M. R. et al. B-cell-depletfing finductfion therapy and acute cellular rejectfion. N Engl J Med 360, 2683–2685 (2009).9. Sagoo, P. et al. Development off a cross-platfform bfiomarker sfignature to detect renal transplant tolerance fin humans. J Clfin Invest 120, 1848–1861 (2010).

10. Newell, K. A. et al. Identfiffcatfion off a B cell sfignature assocfiated wfith renal transplant tolerance fin humans. J Clfin Invest 120, 1836–1847 (2010).

11. Sfilva, H. M. et al. Preservfing the B-cell compartment ffavors operatfional tolerance fin human renal transplantatfion. Mol. Med. 18, 733–743 (2012).

12. Khoder, A. et al. Regulatory B cells are enrfiched wfithfin the IgM memory and transfitfional subsets fin healthy donors but are deffcfient fin chronfic GVHD. Blood 124, 2034–2045 (2014).

13. Peng, Y. et al. Mesenchymal stromal cells finffusfions fimprove reffractory chronfic graff versus host dfisease through an fincrease off CD5+ regulatory B cells producfing finterleukfin 10. Leukemfia 29, 636–646 (2015).

14. Dfing, Q. et al. Regulatory B cells are fidentfiffed by expressfion off TIM-1 and can be finduced through TIM-1 lfigatfion to promote tolerance fin mfice. J Clfin Invest 121, 3645–3656 (2011).

15. Lee, K. M. et al. Antfi-CD45RB/antfi-TIM-1-finduced tolerance requfires regulatory B cells. Am. J. Transplant. 12, 2072–2078 (2012).

16. Mercadante, A. C. T. et al. Oral combfined therapy wfith probfiotfics and alloantfigen finduces B cell-dependent long-lastfing specfiffc tolerance. J Immunol 192, 1928–1937 (2014).

17. Yeung, M. Y. et al. TIM-1 Sfignalfing Is Requfired ffor Mafintenance and Inductfion off Regulatory B Cells. Am. J. Transplant. 15, 942–953 (2015).

18. Wefi, Y. et al. Inffusfion off Dendrfitfic Cells Carryfing Donor Lymphocytes Treated Wfith 8-Methoxypsoralen and Ultravfiolet A Lfight Induces CD19+ IL-10+ Regulatory B cells and Promotes Skfin Allograff Survfival. TPS 46, 3641–3646 (2014).

19. Langfier, S. et al. Prolonged acceptance off skfin graffs finduced by B cells places regulatory T cells on the hfistopathology scene. Braz J Med Bfiol Res 45, 942–947 (2012).

20. Moreau, A. et al. Transfitfional-2 B cells acqufire regulatory ffunctfion durfing tolerance finductfion and contrfibute to allograff survfival. Eur. J. Immunol. 45, 843–853 (2015).

21. Maurfi, C. & Bosma, A. Immune regulatory ffunctfion off B cells. Annu Rev Immunol 30, 221–241 (2012).22. Blafir, P. A. et al. Selectfive targetfing off B cells wfith agonfistfic antfi-CD40 fis an effcacfious strategy ffor the generatfion off finduced regulatory T2-lfike B cells and ffor the suppressfion off lupus fin MRL/lpr mfice. J. Immunol 182, 3492–3502 (2009).

23. Evans, J. G. et al. Novel suppressfive ffunctfion off transfitfional 2 B cells fin experfimental arthrfitfis. J Immunol 178, 7868–7878 (2007).24. Shfimomura, Y. et al. A unfique B2 B cell subset fin the fintestfine. Journal off Experfimental Medficfine 205, 1343–1355 (2008).

www.nature.com/scfientfiffficreports/

SCIENTIFIC REPORTS | 5:11554 | DOI: 10.1038/srep11554 12

25. Rosser, E. C. et al. Regulatory B cells are finduced by gut mficrobfiota-drfiven finterleukfin-1β and finterleukfin-6 productfion. Nat Med 20, 1334–1339 (2014).

26. DfiLfillo, D. J., Matsushfita, T. & Tedder, T. F. B10 cells and regulatory B cells balance fimmune responses durfing finffammatfion, autofimmunfity, and cancer. Ann. N. Y. Acad. Scfi. 1183, 38–57 (2010).

27. Tsang, J. Y.-S. et al. Confferrfing findfirect allospecfiffcfity on CD4+ CD25+ Tregs by TCR gene transffer ffavors transplantatfion tolerance fin mfice. J Clfin Invest 118, 3619–3628 (2008).

28. Shfimomura, Y. et al. Regulatory role off B-1 B cells fin chronfic colfitfis. Internatfional Immunology 20, 729–737 (2008).29. Tfian, J. et al. Lfipopolysaccharfide-Actfivated B Cells Down-Regulate T1 Immunfity and Prevent Autofimmune Dfiabetes fin Nonobese Dfiabetfic Mfice. J Immunol 167, 1081–1089 (2001).

30. Yanaba, K., Bouazfiz, J.-D., Matsushfita, T., Tsubata, T. & Tedder, T. F. Te Development and Functfion off Regulatory B Cells Expressfing IL-10 (B10 Cells) Requfires Antfigen Receptor Dfiversfity and TLR Sfignals. J Immunol 182, 7459–7472 (2009).

31. Lee, K. M. et al. TGF-β -producfing regulatory B cells finduce regulatory T cells and promote transplantatfion tolerance. Eur. J. Immunol. 44, 1728–1736 (2014).

32. Ivanov, I. I. et al. Inductfion off fintestfinal T17 cells by segmented fflamentous bacterfia. Cell 139, 485–498 (2009).33. Mazmanfian, S. K., Round, J. L. & Kasper, D. L. A mficrobfial symbfiosfis ffactor prevents fintestfinal finffammatory dfisease. Nature 453, 620–625 (2008).

34. Hansson, J. et al. Inffuence off gut mficrobfiota on mouse B2 B cell ontogeny and ffunctfion. Molecular Immunology 48, 1091–1101 (2011).

35. Itoh, K. & Mfitsuoka, T. Characterfizatfion off clostrfidfia fisolated ffrom ffaeces off lfimfited ffora mfice and thefir effect on caecal sfize when assocfiated wfith germ-ffree mfice. Lab. Anfim. 19, 111–118 (1985).

36. Cohen, J. L. & Boyer, O. Te role off CD4+ CD25hfi regulatory T cells fin the physfiopathogeny off graff-versus-host dfisease. Current Opfinfion fin Immunology 18, 580–585 (2006).

37. Cohen, J. L., Trenado, A., Vasey, D., Klatzmann, D. & Salomon, B. L. CD4(+ )CD25(+ ) fimmunoregulatory T Cells: new therapeutfics ffor graff-versus-host dfisease. J Exp Med 196, 401–406 (2002).

38. Sakaguchfi, S. Naturally arfisfing CD4+ regulatory t cells ffor fimmunologfic selff-tolerance and negatfive control off fimmune responses. Annu Rev Immunol 22, 531–562 (2004).

39. Taylor, P. A., Lees, C. J. & Blazar, B. R. Te finffusfion off ex vfivo actfivated and expanded CD4(+ )CD25(+ ) fimmune regulatory cells finhfibfits graff-versus-host dfisease lethalfity. Blood 99, 3493–3499 (2002).

40. Cherukurfi, A. et al. Immunologfic human renal allograff finjury assocfiates wfith an altered IL-10/TNF-α expressfion ratfio fin regulatory B cells. J. Am. Soc. Nephrol. 25, 1575–1585 (2014).

41. Pallfier, A. et al. Patfients wfith drug-ffree long-term graff ffunctfion dfisplay fincreased numbers off perfipheral B cells wfith a memory and finhfibfitory phenotype. Kfidney Int 78, 503–513 (2010).

42. Zhao, G. et al. An unexpected counter-regulatory role off IL-10 fin B-lymphocyte-medfiated transplantatfion tolerance. Am. J. Transplant. 10, 796–801 (2010).

43. Ffillatreau, S. Novel regulatory ffunctfions ffor Toll-lfike receptor-actfivated B cells durfing fintracellular bacterfial finffectfion. Immunol Rev 240, 52–71 (2011).

44. Garrett, W. S. et al. Communficable ulceratfive colfitfis finduced by T-bet deffcfiency fin the finnate fimmune system. Cell 131, 33–45 (2007).

45. Smyth, L. A. et al. Tolerogenfic Donor-Derfived Dendrfitfic Cells Rfisk Sensfitfizatfion In Vfivo owfing to Processfing and Presentatfion by Recfipfient APCs. J. Immunol 190, 4848–4860 (2013).

46. Rfigottfier Gofis, L., Bourhfis, A. G., Gramet, G., Rochet, V. & Doré, J. Fluorescent hybrfidfisatfion combfined wfith ffow cytometry and hybrfidfisatfion off total RNA to analyse the composfitfion off mficrobfial communfitfies fin human ffaeces usfing 16S rRNA probes. FEMS Mficrobfiology Ecology 43, 237–245 (2003).

AcknowledgmentsTfis work was ffunded by the BHF. In addfitfion, we would lfike to acknowledge the contrfibutfion off the Medfical Research Councfil (MRC) Centre ffor Transplantatfion, Kfing’s College London, UK – MRC grant no. MR/J006742/1. Ffinally, the research was also supported by the Natfional Instfitute ffor Health Research (NIHR) Bfiomedfical Research Centre based at Guy’s and St Tomas’ NHS Foundatfion Trust and Kfing’s College London. Te vfiews expressed are those off the author(s) and not necessarfily those off the NHS, the NIHR or the Department off Health. RA was supported by Kfing Abdul Azfiz Unfiversfity, Jeddah, KSA. RJN was Prfincfipal Research Fellowshfip supported by Wellcome Trust.

Author ContrfibutfionsR.A., P.B. and R.E. desfigned and perfformed research, collected and analyzed data and wrote the manuscrfipt; E.S., P.D.B., K.R., N.S., E.S., L.S. and S.O. assfisted fin research, E.M. perfformed the fimmunoffuoresence mficroscopy, R.J.N., G.L., J.K.H. and J.S. provfided crfitfical reagents and assfisted fin research, R.I.L. and G.L. desfigned research and wrote the manuscrfipt.

Addfitfional InfformatfionSupplementary finfformatfion accompanfies thfis paper at http://www.nature.com/srep

Competfing ffnancfial finterests: Te authors declare no competfing ffnancfial finterests.

How to cfite thfis artficle: Alhabbab, R. et al. Dfiversfity off gut mficroffora fis requfired ffor the generatfion off B cell wfith regulatory propertfies fin a skfin graff model. Scfi. Rep. 5, 11554; dofi: 10.1038/srep11554 (2015).

Tfis work fis lficensed under a Creatfive Commons Attrfibutfion 4.0 Internatfional Lficense. Te fimages or other thfird party materfial fin thfis artficle are fincluded fin the artficle’s Creatfive Com-

mons lficense, unless findficated otherwfise fin the credfit lfine; fiff the materfial fis not fincluded under the Creatfive Commons lficense, users wfill need to obtafin permfissfion ffrom the lficense holder to reproduce the materfial. To vfiew a copy off thfis lficense, vfisfit http://creatfivecommons.org/lficenses/by/4.0/