2017-Bel-Paneth cells secrete lysozyme via sec.pdf

INTESTINAL STRESSPaneth cells secrete lysozyme viasecretory autophagy during bacterialinfection of the intestineShai Bel,1Mihir Pendse,1Yuhao Wang,1Yun Li,1Kelly A.Ruhn,1Brian Hassell,1Tess Leal,1Sebastian E.Winter,2Ramnik J.Xavier,3,4,5Lora V.Hooper1,6*Intestinal Paneth cells limit bacterial invasion by secreting antimicrobial proteins,includinglysozyme.However,invasive pathogens can disrupt the Golgi apparatus,interfering withsecretion and compromising intestinal antimicrobial defense.Here we show that duringbacterial infection,lysozyme is rerouted via secretory autophagy,an autophagy-basedalternative secretion pathway.Secretory autophagy was triggered in Paneth cells bybacteria-induced endoplasmic reticulum(ER)stress,required extrinsic signals from innatelymphoid cells,and limited bacterial dissemination.Secretory autophagy was disruptedin Paneth cells of mice harboring a mutation in autophagy gene Atg16L1 that confersincreased risk for Crohns disease in humans.Our findings identify a role for secretoryautophagy in intestinal defense and suggest why Crohns disease is associated withgenetic mutations that affect both the ER stress response and autophagy.Themammalianintestineishometoadiversepopulationofbacteria,whichincludespatho-gensthatcandisrupthostcellularfunctions.The intestinalepitheliumdefendsagainstbacterial encroachment through multiplemechanisms,including antimicrobial proteinsecretion and destruction of invading bacteriathrough autophagy(1).Paneth cells arespecial-ized intestinal epithelial cells that secrete abun-dant antimicrobial proteins,including lysozyme;thus,disrupting Paneth cell secretion can lead toinflammatorydisease(24).Pathogenicmicrobescan trigger endoplasmic reticulum(ER)stressthat interferes with protein secretion(5,6)andcompromisesantimicrobialproteindelivery,rais-ing the question of how Paneth cells preservetheir antimicrobial function during pathogen-induced stress.Invasivebacteria,includingSalmonellaentericaserovar Typhimurium(S.Typhimurium),triggerautophagy in intestinal enterocytes.This is indi-cated by abundant epithelial cell autophagosomes,marked by LC3(microtubule-associated protein1 light chain 3),that capture and eliminate invad-ing bacteria(7).S.Typhimurium also invadedPanethcells(fig.S1),andinvasionwasassociatedwith elevated numbers of LC3+puncta in Panethcells(Fig.1,A and B).The puncta numbers werecomparable to those inmice subjected to fasting,atriggerofcanonicalautophagy(8)(Fig.1,AandB).However,many of the LC3+structures ininfected Paneth cells were markedly larger(0.2to 7 mm in diameter)than the LC3+puncta inS.Typhimuriuminfected enterocytes(1mm)orin Paneth cells of fasted(0.5 mm)mice(Fig.1,Cand D).TocharacterizethecontentsoftheLC3+vesicles,weperformedimmunofluorescence,transmissionelectron microscopy,and coimmunoprecipitationassays.These assays revealed that the large LC3+vesicles contained lysozyme(Fig.1,E to G)andwere absent in Paneth cells of uninfected andfasted mice,where lysozyme was packaged intoLC3vesicles(fig.S2).Ultrastructure analysisshowed thatthe large granules in Paneth cells ofinfected mice contained lysozyme(fig.S3)andwere surrounded by a double membrane(Fig.1H),a hallmark of autophagosomes(8).Gran-ules from uninfected mice were surrounded bya single membrane(Fig.1H).The LC3+vesiclesdid not contain bacteria(fig.S4,A and B)orthe antimicrobial proteins REG3g or cryptdin 5(fig.S5).Also,cryptdin 5 was not packaged insecretory granules and was excluded from theLC3+vesicles ininfected mice(fig.S6,Aand B).This suggested that infection interferes withpackaging and secretion of Paneth cell antimicro-bialproteinsandthatlysozymemightbereroutedthrough an alternative secretion pathwayinvolv-ing an LC3+vesicle.CanonicalautophagytargetsthecargoinLC3+autophagosomes for degradation in lysosomes(8).Immunofluorescence analysis indicated thatthelysozyme-filledLC3+vesiclesdonotfusewithlysosomes,as the vesicles did not colocalize withthe lysosomal marker cathepsin D(Fig.1,I andJ),implying that lysozyme is not targeted fordegradation.The lysozyme-filled LC3+vesiclesalso did not colocalize with p62/sequestesome1(SQSTM1),which selects autophagosome cargofordegradation(8).SQSTM1wasassociated withthe smaller LC3+puncta in Paneth cellsbutwasnot associated with the larger LC3+lysozyme+vesicles in infected mice(fig.S7),implying thatlysozyme is not selected for degradation in theautolysosome through the canonical selective auto-phagy pathway.LC3 also accumulated atthe apicalsurfaceofPanethcellsfrominfectedmice,andLC3+vesicles had fused with the apical surface and dis-chargedlysozymeintotheintestinallumen(Fig.1K).Thiswasnotobserved in uninfected mice(fig.S7),which suggests that the LC3+vesicles might beinvolved in lysozyme secretion.Duringconventionalproteinsecretion,proteinsare transported through the ERGolgi complex,packaged in secretory granules,and released tothe extracellular space.There are various alter-native secretory pathways,including one thatutilizes components of the autophagy pathwayand is known as secretory autophagy(9).In se-cretory autophagy,cargo is transported in anLC3+vesicleanddischargedattheplasmamem-brane,thus bypassing the ER-Golgi complex(9).Rab8a,a marker of secretory autophagy ves-icles(10),colocalized with the lysozyme-filledLC3+vesicles.Rab8a also coimmunoprecipi-tated with LC3 only in infected mice but did notcolocalize with cryptdin-5(fig.S8).This sug-gested that lysozymemightbeselectivelysecretedthroughthesecretoryautophagypathwayduringinfection.To further test this idea,we isolated Panethcellcontaining crypts,infected them in vitrowhile treating with chemical inhibitors of con-ventional secretion and autophagy,and analyzedthe supernatants for lysozyme secretion.Inhibit-ing ER-Golgi trafficking with brefeldin A(BFA)did not affect lysozyme secretion in infected oruninfected crypts(Fig.2,A and B),indicatingthat lysozyme secretion can bypass the ER-Golgipathway.Lysozyme secretion was also not alteredby treatment with chloroquine(Fig.2,A and B),which prevents lysosome acidification(8),imply-ing that inhibiting autophagic degradation doesnot affect lysozyme secretion.However,treat-ment with 3-methyladenine(3-MA),which in-hibits autophagosome nucleation(8),impairedlysozyme secretion and caused an accumula-tion of intracellular lysozyme in infected crypts(Fig.2,A and B).Accordingly,secretions fromBFA-treated but not 3-MAtreated crypts killedbacteria(Fig.2C),indicating that secretory auto-phagy is essential for antibacterial defense ininfected crypts.We next studied mice in which autophagy isperturbed by a mutation in the Atg16L1(autoph-agy related 16like 1)gene.A mutation in theAtg16L1 gene Thr300Ala300(T300A)confersan increased risk of developing Crohns diseasein humans(11).Mice harboring this mutation(Atg16L1T300A)exhibit decreased antibacterialautophagy and abnormal Paneth cell lysozymedistribution(12).Whereas crypts from uninfectedwild-type and Atg16L1T300Amice secreted similaramounts of lysozyme,lysozyme secretion was im-paired in crypts from S.TyphimuriuminfectedRESEARCHBel et al.,Science 357,10471052(2017)8 September 20171 of 51Department of Immunology,The University of TexasSouthwestern Medical Center,Dallas,TX 75390,USA.2Department of Microbiology,The University of TexasSouthwestern Medical Center,Dallas,TX 75390,USA.3Broad Institute,Cambridge,MA 02142,USA.4Center forComputational and Integrative Biology,MassachusettsGeneral Hospital,Boston,MA 02114,USA.5GastrointestinalUnit and Center for the Study of Inflammatory BowelDisease,Massachusetts General Hospital,Harvard MedicalSchool,Boston,MA 02142,USA.6The Howard HughesMedical Institute,The University of Texas SouthwesternMedical Center,Dallas,TX 75390,USA.*Corresponding author.Email:lora.hooperutsouthwestern.eduon December 27,2018 http:/science.sciencemag.org/Downloaded from Atg16L1T300Amice(Fig.2,DandE).TreatmentofinfectedcryptsfromAtg16L1T300Amicewith3-MAdidnotfurtherhinderlysozymesecretion(Fig.2,DandE),which suggests that3-MAinhibitionoflysozyme secretion in wild-type crypts was notdue to off-target effects.Accordingly,crypts fromAtg16L1T300Amiceshowedreducedbacterialkilling.and 3-MA did not further affect this reduction(Fig.2F).Furthermore,infection of Atg16L1T300Amicedidnotproducelysozyme-filledLC3+vesiclesas observed in wild-type mice(Fig.2,G and H).These results support a role for autophagy in ly-sozyme secretion during infection and suggesthow ATG16L1 gene mutations could lead to theaberrant lysozyme packaging and secretion thatBel et al.,Science 357,10471052(2017)8 September 20172 of 50.00.20.40.60.81.0Colocalization of lysozyme and LC3PearsonsR ValueCostesP ValueControlFastedLC3 DAPI+S.TmLYZ+S.TmLC3Cathepsin DMergeLC3LYZMerge+S.Tm-0.50.00.51.0Lysozyme&LC3Lysozyme&Cathepsin D+S.TmColocalization(Pearsons R value)+S.Tm-S.Tm*02040*LC3+puncta/cryptControlFasted+S.Tm6002468Vesicle diameter(m)*CryptVillustipFasted(-S.Tm)Crypt+S.Tm+S.TmCrypt+S.TmVillus tipFasted-S.TmLC3 DAPICryptLC3 LYZDAPI-S.Tm+S.TmIP:InputLC3LYZIgGInputLC3LYZIgGIB:LYZ+S.Tm-S.TmFig.1.Large LC3+vesicles in S.Typhimuriuminfected mice containlysozyme.(A)Immunofluorescence of LC3 in intestinal crypts.Nuclei arestained with DAPI(4,6-diamidino-2-phenylindole).Scale bars,10 mm.(B)Quantification of LC3+puncta.Each data point represents one mouse.(C)Immunofluorescence of LC3 in intestinal crypts.Scale bars,5 mm.(D)LC3+vesicle diameter measurements.(E)Immunofluorescence of LC3and lysozyme in S.Typhimuriuminfected intestinal crypts.A Paneth cell isoutlined.Arrows indicate a lysozyme-filled LC3+vesicle;arrowheadsindicate an autophagosome that does not contain lysozyme.Scale bars,5 mm.(F)Colocalization of LC3 and lysozyme in intestinal crypts fromS.Typhimuriuminfected mice.Each point represents one lysozymegranule.(G)Coimmunoprecipitation(IP)of intestinal lysates using theindicated antibodies.Immunoblot(IB)was performed with anti-lysozymeantibody.IgG,immunoglobulin G.(H)Transmission electron microscopy ofPaneth cells from uninfected(-S.Tm)and infected(+S.Tm)mice.Asterisks indicate secretory granules;arrowheads indicate surroundingmembranes.(I)Immunofluorescence of lysosomes(cathepsin D+),LC3,and lysozyme in S.Typhimuriuminfected intestinal crypts.Arrowsindicate a lysozyme-filled LC3+vesicle with no lysosome(cathepsin D)signal;arrowheads indicate lysosomes that are not coincident withlysozyme-filled LC3+vesicles.Scale bars,5 mm.(J)Quantification oflysosome(cathepsin D),LC3,and lysozyme colocalization in(I).Each datapoint represents one lysozyme-containing granule.Two points connectedby a line represent the same granule.The dashed line denotes the limit ofstrong colocalization.(K)Immunofluorescence of LC3 and lysozyme inintestinal crypts.Scale bars,10 mm.*P 0.05;*P 0.01;*P 0.001;*P 0.0001;one-way analysis of variance(ANOVA)(B)and(D).S.Tm,Salmonella Typhimurium;LYZ,lysozyme.RESEARCH|REPORTon December 27,2018 http:/science.sciencemag.org/Downloaded from characterizes Paneth cellsfromCrohnsdiseasepatients(2).We next sought to identify the cellular signalsthattrigger secretory autophagy.S.Typhimuriumdisrupts the ER-Golgi complex in infected cellsand thus interferes with conventional secretion(13).Accordingly,S.Typhimurium induced Golgifragmentation in Paneth cells(fig.S9),whereasnoninvasive bacteria,including the commensalB.thetaiotaomicronandthemutantS.TyphimuriumDinvA,did not cause Golgi breakdown(fig.S9)or triggersecretoryautophagyoflysozyme(Fig.3,AandB).Disruption of the ER-Golgi complexactivates ER stress pathways that maintain cel-lular functions(14).This response includes ele-vated expression of CHOP(C/EBP homologousprotein)(15).CHOP levels were elevated intheintestinesofS.Typhimuriuminfectedmice(Fig.3C),indicating activation of the ER stress re-sponse.This suggested that ER stress mighttrigger secretory autophagy of lysozyme to cir-cumvent the secretion block caused by ER-Golgidisruption.Supporting this idea,the ER stressinhibitortauroursodeoxycholicacid(TUDCA)(6)reduced ER stress in infected mice,as indicatedby lowered CHOP expression(Fig.3C),and alsoreduced secretory autophagy of lysozyme(Fig.3,D and E).Conversely,treatment of uninfectedmice with the ER stress inducer thapsigargin(6)elevated CHOP expression(Fig.3C)and inducedsecretory autophagy of lysozyme(Fig.3,D andE).Thus,ER stress triggers secretory autophagyof lysozyme.We next investigatedthe intracellular signalingpathways that link ER stress to secretory autoph-agy in Paneth cells.When cells sense ER stress,PERK(protein kinase RNA-like endoplasmic re-ticulum kinase)is activated by phosphorylation.p-PERK then phosphorylates elongation initia-tion factor 2a(eIF2a),which inactivates eIF2aand attenuatestranslation(16).ThispathwaylinksER stress to compensatory autophagy activationin Paneth cells(17).S.Typhimurium infection in-creased phosphorylation of intestinal PERK andeIF2a(Fig.3F and fig.S10),consistent with itsactivation of ER stress pathways.eIF2a was notactivated by infection with the noninvasive bac-teria B.thetaiotaomicron and S.TyphimuriumDinvA(fig.S11),which indicates that activationdepends upon bacterial invasion.Treatment ofuninfectedmicewithsalubrinal,aselectiveinhib-itor of p-eIF2a dephosphorylation(18),promotedsecretory autophagy of lysozyme in Paneth cells(Fig.3,G and H),supporting the idea that ERstresstriggerssecretoryautophagythrough thePERK-eIF2a pathway.We next tested whether inhibiting secretoryautophagy would compromise intestinal defenseagainst oral S.Typhimurium infection.Inhibit-ing secretory autophagy by TUDCA treatment ofS.Typhimuriuminfected mice led to increasednumbersofS.Typhimuriumintheintestine,mes-enteric lymph nodes(MLNs),liver,and spleen(Fig.3I).Lysozyme gavage of TUDCA-treated in-fectedmicerescuedtheincreasedbacterialburden(Fig.3I),suggesting that the increased bacterialnumberswerenotduetoothereffectsofTUDCA.Thus,secretory autophagy is essential for hostdefense against invasive bacteria.Activation of antibacterial autophagy in intes-tinalenterocytesrequiresepithelialcellexpressionof the Toll-like receptor(TLR)signaling adap-torMyD88(7).Secretoryautophagywasinhibitedin Paneth cells of Myd88-deficient mice(Fig.4,Aand B),producing a diffuse distribution of lyso-zyme similar to that seen in mice hypomorphicfor Atg16L1(3)(Fig.4,A and C).Secretory auto-phagy was still evident in infected mice with anepithelialcellspecificdeletionofMyd88(Myd88DIEC)(Fig.4,A and B),which indicates that epithelialcell Myd88 is dispensable.In contrast,infectedmiceharboringadendriticcell(DC)specificMyd88deletion(Myd88DDC)failed to show secretoryautophagy and exhibited a diffuse distributionBel et al.,Science 357,10471052(2017)8 September 20173 of 5WT+S.TmLC3LYZMergeT300ASecreted LYZIntracellular LYZIntracellular Tubulin+S.Tm-S.TmControlBFAChloro3MAControlBFAChloro3MALYZ(arbitrary units)secretedIntracellular+S.Tm-S.Tm0.00.51.01.50.00.51.01.50.00.51.01.52.02.50.00.51.01.5ControlBFAChloro3-MAControlBFAChloro3-MA+S.Tm-S.Tm024860.00.51.0+S.Tm-S.TmsecretedIntracellularLYZ(arbitrary