Inactivation of Sox9 in fibroblasts reduces cardiac fibrosis and inflammation

Inac i a ion o Sox9 in ib oblas s educes ca diac ib osis and in lamma ion Gesine M. Scha , … , Johann Baue sachs, Joe g Heineke JCI Insigh . 2019;4(15):e126721. h ps://doi.o g/10.1172/jci.insigh .126721. Fib o ic sca ing d i es he p og ession o hea ailu e a e myoca dial in a c ion (MI). The e o e, he de elopmen o speci ic ea men egimens o coun e ac ib osis is o high clinical ele ance. The ansc ip ion ac o sex-de e mining egion Y box 9 (SOX9) unc ions as an impo an egula o du ing emb yogenesis, bu ecen da a poin owa d an addi ional causal ole in o gan ib osis. We show he e ha SOX9 is up egula ed in he sca a e MI in mice. Fib oblas -speci ic dele ion o Sox9 amelio a ed MI-induced le en icula dys unc ion, dila a ion, and myoca dial sca ing in i o. Unexpec edly, dele ion o Sox9 also po en ly elimina ed pe sis ing leukocy e in il a ion o he sca in he ch onic phase a e MI. RNA-Seq om he in a c sca e ealed ha Sox9 dele ion in ib oblas s esul ed in s ongly down egula ed exp ession o genes ela ed o ex acellula ma ix, p o eolysis, and in lamma ion. Impo an ly, Sox9 dele ion in isola ed ca diac ib oblas s in i o simila ly a ec ed gene exp ession as in he ca diac sca and educed ib oblas p oli e a ion, mig a ion, and con ac ion capaci y. Toge he , ou da a demons a e ha ib oblas SOX9 unc ions as a mas e egula o o ca diac ib osis and in lamma ion and migh cons i u e a no el he apeu ic a ge du ing MI. Resea ch A icle Ca diology Find he la es e sion: h p://jci.me/126721/pd 1insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE Au ho ship no e: GMS and KK con ibu ed equally o his wo k. Con lic o in e es : The au ho s ha e decla ed ha no con lic o in e es exis s. Copy igh : © 2019, Ame ican Socie y o Clinical In es iga ion. Submi ed: Decembe 10, 2018 Accep ed: July 9, 2019 Published: July 16, 2019. Re e ence in o ma ion: JCI Insigh . 2019;4(15):e126721. h ps://doi.o g/10.1172/jci. insigh .126721. Inac i a ion o Sox9 in ib oblas s educes ca diac ib osis and in lamma ion Gesine M. Scha ,1,2 Ka ja Kilian,1 Julio Co de o,3 Yong Wang,1,4 And ea G und,1,2 Melanie Ho mann,1 Na ali F oese,1 Xue Wang,1 And eas Kispe ,5 Ral Kis ,6 Simon J. Conway,7 Robe Ge e s,8 Kai C. Wolle ,1,4 Ge gana Dob e a,3,9 Johann Baue sachs,1 and Joe g Heineke1,2,9 1Depa men o Ca diology and Angiology, Hanno e Medical School, Hanno e , Ge many. 2Depa men o Ca dio ascula Resea ch and 3Depa men o Ana omy and De elopmen al Biology Cen e o Biomedicine and Medical Technology Mannheim, Eu opean Cen e o Angioscience, Medical Facul y Mannheim o Heidelbe g Uni e si y, Mannheim, Ge many. 4Di ision o Molecula and T ansla ional Ca diology, Depa men o Ca diology and Angiology, and 5Ins i u e o Molecula Biology, Hanno e Medical School, Hanno e , Ge many. 6Ins i u e o Gene ic Medicine, Facul y o Medical Sciences, and Cen e o O al Heal h Resea ch, School o Den al Sciences, Newcas le Uni e si y, Newcas le upon Tyne, Uni ed Kingdom. 7HB Wells Cen e o Pedia ic Resea ch, Indiana Uni e si y School o Medicine, Indianapolis, Indiana, USA. 8Helmhol z Cen e o In ec ion Resea ch, B aunschweig, Ge many. 9Ge man Cen e o Ca dio ascula Resea ch pa ne si e Mannheim/Heidelbe g, Heidelbe g, Ge many. In oduc ion Ch onic hea ailu e (CHF) eme ges mos commonly a e myoca dial in a c ion (MI) in he Wes - e n wo ld (1). Despi e some ecen ad ances in ea men s a egies, he mo ali y o pa ien s wi h CHF emains unaccep ably high. Ca diac ib osis is a majo d i e o disease p og ession in CHF (2). Al hough he deposi ion o ex acellula ma ix (ECM) is needed o eplace dead ca diomyocy es a e MI and o p e en en icula up u e, excessi e ib osis causes la ge in a c sca s esul ing in ca diac dila a ion and educed ca diac unc ion (2–4). Wi hin he i s days a e MI, sca o ma ion coincides wi h he educ ion o leukocy e in il a ion and in lamma ion, which is ini ially needed o clea nec o ic ca diomyocy es (5). Adequa e e mina ion o he in lamma o y esponse, howe e , is equi ed in he myoca dium o p e en le en icula dila a ion, which (besides la ge MI size) is he main con ibu o o hea ailu e de elopmen and poo p ognosis (6). Al hough ini ial in a c size a e MI can be he apeu ically educed by immedia e pe cu aneous co ona y in e en ions o es o e pa ency o he in a c ed co ona y essel, ECM deposi ion and ela ed sca expansion as well as pe sis ing myoca dial in lamma ion canno be a ge ed by cu en he apies (6, 7). In his ega d, i would be especially desi able o modula e myoca dial sca issue in a way ha educes ma ix deposi ion and in lamma o y signals. Howe e , nei he he p edominan ly esponsible cell ype(s) no lineage- es ic ed nodal molecula egula o s ha could be add essed by u u e he apies a e well de ined. Fib o ic sca ing d i es he p og ession o hea ailu e a e myoca dial in a c ion (MI). The e o e, he de elopmen o speci ic ea men egimens o coun e ac ib osis is o high clinical ele ance. The ansc ip ion ac o sex-de e mining egion Y box 9 (SOX9) unc ions as an impo an egula o du ing emb yogenesis, bu ecen da a poin owa d an addi ional causal ole in o gan ib osis. We show he e ha SOX9 is up egula ed in he sca a e MI in mice. Fib oblas -speci ic dele ion o Sox9 amelio a ed MI-induced le en icula dys unc ion, dila a ion, and myoca dial sca ing in i o. Unexpec edly, dele ion o Sox9 also po en ly elimina ed pe sis ing leukocy e in il a ion o he sca in he ch onic phase a e MI. RNA-Seq om he in a c sca e ealed ha Sox9 dele ion in ib oblas s esul ed in s ongly down egula ed exp ession o genes ela ed o ex acellula ma ix, p o eolysis, and in lamma ion. Impo an ly, Sox9 dele ion in isola ed ca diac ib oblas s in i o simila ly a ec ed gene exp ession as in he ca diac sca and educed ib oblas p oli e a ion, mig a ion, and con ac ion capaci y. Toge he , ou da a demons a e ha ib oblas SOX9 unc ions as a mas e egula o o ca diac ib osis and in lamma ion and migh cons i u e a no el he apeu ic a ge du ing MI. 2 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE Residen ca diac ib oblas s a e conside ed he main homeos a ic p oduce o ECM in he hea , al hough o he cells, o example ca diomyocy es, a e also able o make and sec e e collagens (8). In esponse o pa hological o e load o ischemic inju y, quiescen ib oblas s become ac i a ed, p oli e a e, mig a e, and di e en ia e owa d a pa hological myo ib oblas pheno ype, which deposi s excess ECM and exe s con ac ile p ope ies (8). Sex-de e mining egion Y box 9 (SOX9) is a ansc ip ion ac o ha plays an essen ial ole du ing mammalian de elopmen , whe e i c ucially egula es chond ogenesis and sex di e en ia ion (9–12). He - e ozygous mu a ions o he SOX9 gene lead o he pheno ype o campomelic dysplasia, a human skele al mal o ma ion synd ome wi h XY sex e e sal ha is usually le hal wi hin he neona al pe iod (13, 14). As an impo an egula o o ECM genes, SOX9 plays a signi ican ole in he pa hogenesis o a ious ib o ic diseases, such as li e ib osis, glome uloscle osis, and hea al e calci ica ion (15–17). Recen ly, SOX9 was sugges ed o he i s ime as a po en ial egula o o ca diac ib osis in mice a e ischemia/ epe u- sion inju y (18). The unc ional ele ance o SOX9 in pa hological ca diac ib osis, he ca diac cell ype i ac s in o p omo e ECM deposi ion, as well as i s e ec on he en i e ib oblas ansc ip ome, howe e , emain cu en ly unknown. He e, we desc ibe ha ib oblas -speci ic down egula ion o Sox9 educes ECM deposi ion in a mouse model o MI in i o and ha i p e en s he ac i a ion owa d a p oli e a i e and mig a ing ib oblas pheno ype in cell cul u e in i o. In addi ion, we demons a e ha educed Sox9 exp ession in ib oblas s amelio a es pe sis ing in lamma ion wi hin he in a c sca , p e en s ca diac dila a ion, and imp o es ca diac unc ion a e MI. Resul s Sox9 is exp essed in ca diomyocy es and ca diac ib oblas s and is up egula ed a e MI in mice. Immuno luo escence s aining o SOX9, pla ele -de i ed g ow h ac o ecep o –α (PDGFR-α, a ma ke o ca diac ib oblas s) (19), and ca diac oponin T (as a ca diomyocy e ma ke ) showed exp ession o SOX9 in nuclei o ca diac ib oblas s and ca diomyocy es in adul mouse en icles (Figu e 1A). Quan i ica ion o mRNA le els in isola ed ca diomyocy es and isola ion o ca diac ib oblas s by quan i a i e PCR (qPCR) om uninju ed mouse hea s con i med Sox9 exp ession in bo h cell ypes, al hough mRNA le els o Sox9 we e ma ked- ly highe in ib oblas s compa ed wi h ca diomyocy es (Supplemen al Figu e 1A; supplemen al ma e ial a ailable online wi h his a icle; h ps://doi.o g/10.1172/jci.insigh .126721DS1). Analysis o SOX9 by immunoblo ing showed a signi ican up egula ion o SOX9 in he in a c zone ea ly a e MI (day 3), which pa ially dec eased o e ime in he sca egion 7 and 28 days a e MI bu emained signi ican ly ele a ed compa ed wi h uninju ed sham hea s o he emo e zone a e MI. In con as , en icula samples om he emo e a ea o he same MI-inju ed mice showed no signi ican di e ence in SOX9 p o ein le els compa ed o sham-ope a ed mice (Figu e 1, B–F). To in es iga e he unc ional ele ance o Sox9 in ac i a ed ib oblas s in i o, B6;129S -Sox9 m1Gs (Sox9 l/ l) mice we e b ed wi h mice exp essing C e ecombinase linked o he pe ios in (Pos n) p omo e o ob ain mice wi h speci ic Sox9 knockou in ac i a ed ib oblas s (Pos nC e/+ Sox9 l/ l; o sho , Sox9 l-Pe – C e; Figu e 2A). Pos n is exp essed in a ious s uc u es unde no mal condi ions du ing ca diac de el- opmen (e.g., in he de eloping ca diac cushions), bu i s exp ession in quiescen ib oblas s o ca diomy- ocy es o he uninju ed adul hea is negligible. Howe e , he Pos n-C e allele is obus ly up egula ed in adul ac i a ed ib oblas s ollowing ca diac inju y (20–24). Acco dingly, immuno luo escence s aining o SOX9 in he in a c a ea o Sox9 l-Pe –C e mice 7 days a e MI e ealed a ma ked educ ion o SOX9+ ib oblas s (s ained wi h PDGFR-α) compa ed wi h li e ma e con ols (Sox9 l/ l; o sho , Sox9 l/ l; see Figu e 2, B and C). Mo eo e , immunoblo analysis o Sox9 l-Pe –C e mice a e MI e ealed a ma ked educ ion o SOX9 p o ein le els compa ed wi h con ol mice only in he ca diac sca bu no in he emo e egion o he myoca dium (Figu e 2, D–F). Because he sca is composed mainly o ib oblas s wi h only a ew esidual myocy es, hese esul s sugges ed dec eased exp ession o SOX9 in ib oblas s wi hin he sca egion a e MI, bu no in he emo e a ea, which p edominan ly (a leas wi h ega d o cell olume) consis s o ca diomyocy es. Hence, Sox9 l-Pe –C e mice exe ed educed ib oblas SOX9 le els wi hin he sca bu no in he ca diomyocy e compa men in he emo e myoca dium. Fib oblas -speci ic Sox9 dele ion imp o es ca diac unc ion a e MI. To analyze he impac o ib oblas SOX9 a e MI, Sox9 l-Pe –C e and Sox9 l/ l (con ol mice) unde wen le an e io descending co ona y a e y (LAD) liga ion o sham su ge y. A e 42 days, we ound no di e ence in hea weigh /body 3 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE weigh (HW/BW) a io be ween Sox9 l-Pe –C e and con ol mice, while i signi ican ly inc eased in bo h geno ypes a e MI e sus sham-ope a ed mice (Figu e 3A). The lung weigh /body weigh (LW/BW) a io was no di e en be ween Sox9 l-Pe –C e and con ol mice, and he e was no di e ence be ween he MI and sham coho s, indica ing ha he animals we e in a compensa ed s a e o hea ailu e wi hou pulmona y edema (Figu e 3B). T ans ho acic echoca diog aphy e ealed a signi ican ly be e sys olic ca diac unc ion (measu ed as le en icula ejec ion ac ion) o he Sox9 l-Pe –C e compa ed wi h con ol mice 7 days a e MI, which was main ained un il 42 days a e su ge y (Figu e 3, C and D). Mo eo e , he le en icula chambe s o Sox9 l-Pe –C e mice we e signi ican ly less dila ed (i.e., exe ed a educed LVEDA) 7 and 42 days a e MI compa ed wi h con ol mice (Figu e 3, E and F). No di e ence occu ed be ween bo h geno ypes a any ime poin a e sham su ge y, and no ana omical abno mali ies we e obse ed in Sox9 l-Pe –C e mice. Dele ion o Sox9 in ib oblas s in i o educes myoca dial sca ing and ib oblas p oli e a ion a e MI. To iden i y he unde lying cause o he p ese a ion o ca diac unc ion a e MI upon Sox9 dele ion in ib oblas s, we quan i ied he ex en o myoca dial sca o ma ion in hea s 7 days a e MI. Masson’s ich ome s aining e ealed signi ican ly smalle myoca dial sca a eas in Sox9 l-Pe –C e compa ed wi h con ol mice (Figu e 4, A and B). Immuno luo escence s aining o he p oli e a ion ma ke Ki67 wi h PDGFR-α as a ib oblas ma ke sugges ed educed ib oblas p oli e a ion in he sca s o Sox9 l-Pe –C e compa ed wi h con ol mice 7 days a e MI (Figu e 4, C and D). To in es iga e whe he Figu e 1. Ca diac SOX9 p o ein le el a e MI. Adul wild- ype mice unde wen MI by pe manen liga ion o he le an e io descending a e y. (A) Rep e- sen a i e immuno luo escence images show SOX9 (g een) in ca diac ib oblas s (s ained wi h PDGFR-α, ed) and ca diomyocy es (s ained wi h oponin T, ed) in he emo e a ea; nuclei a e s ained wi h DAPI (blue). A ows indica e SOX9+ ca diac ib oblas s; a owheads indica e SOX9+ ca diomyocy es. Scale ba : 50 μm. (B–F) Wes e n blo analysis (B) and densi ome ic quan i ica ion (C–F) o SOX9 p o ein le el in hea issue o he emo e o in a c a ea and sham con ols isola ed om wild- ype mice 3, 7, o 28 days a e sham ope a ion o MI. Ponceau S ( ange 40–55 kDa) was used as a loading con ol. Fo compa ison o SOX9 p o ein le el in he in a c a ea a di e en ime poin s (F), da a we e no malized o 3-day sham as a e e ence. Da a a e shown as mean ± SEM. n = 3–4 mice/g oup; 1-way ANOVA wi h Tukey’s mul iple-compa isons es ; *P < 0.05, ***P < 0.001, ****P < 0.0001. 4 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE Sox9 dele ion p e en s he o ma ion o myo ib oblas s, we pe o med immuno luo escence s aining o smoo h muscle ac in (α-SMA) in in a c hea s o Sox9 l-Pe –C e compa ed wi h con ol mice. We ound a clea inc ease o α-SMA deposi ion in he in a c a ea o Sox9 l-Pe –C e mice as well as li e - ma e con ols compa ed wi h he emo e zone, wi h a end owa d highe α-SMA deposi ion in he in a c a ea o con ol mice, al hough a high a iabili y o s aining in ensi y exis ed wi hin he sca s and be ween mice (Figu e 4E). Quan i ica ion o Ac a2 mRNA le el in he emo e and in a c a ea o Sox9 l-Pe –C e and con ol mice 7 days a e MI con i med he inc ease o Ac a2 in he in a c a ea com- pa ed wi h con ol. Al hough he e was a end owa d lowe Ac a2 mRNA le el in he in a c issue o Sox9 l-Pe –C e compa ed wi h con ol mice, we did no ind a signi ican educ ion o Ac a2 mRNA le els upon Sox9 dele ion in ou in i o model (Figu e 4F). Dele ion o Sox9 in i o educes p oli e a ion, mig a ion, and gel con ac ion o ca diac ib oblas s bu has no di ec e ec on α-SMA o ma ion. To in es iga e he di ec e ec s o Sox9 dele ion on he pheno ype o ca diac ib oblas s, we ex ac ed ib oblas s om hea s o adul Sox9 l/ l mice and ea ed hem wi h adeno i uses exp essing C e ecombinase (AdC e) o con ol i us (Adβ-gal). qPCR and Wes e n blo - ing e ealed highly e icien down egula ion o Sox9 mRNA (by abou 90%) and p o ein exp ession (by abou 60%) in AdC e- ea ed ib oblas s in i o (Supplemen al Figu e 1B and Figu e 5, A and B). Cell p oli e a ion was assessed wi h a B dU inco po a ion immunoassay, which e ealed signi ican ly less p oli e a ion in Sox9-de icien ib oblas s compa ed wi h con ol cells (Figu e 5C), he eby con i ming he in i o esul s shown in Figu e 4, C and D. To measu e he ib oblas con ac ile capaci y, which is a hallma k o cell di e en ia ion in o myo ib oblas s, he cells we e in ec ed wi h AdC e o Adβ-gal and Figu e 2. Fib oblas -speci ic Sox9 dele ion in i o. (A) Schema ic illus a ion showing he in e b eeding o B6;129S -Sox9 m1Gs (Sox9 l/ l) wi h Tg(Pos n-C e)1Sjc (Pe -C e) mice o ob ain mice wi h speci ic Sox9 dele ion in ac i a ed ca diac ib oblas s (Sox9 l-Pe –C e). (B and C) Rep esen a i e IHC images (B) and quan i ica ion (C) o SOX9+ ib oblas s om he sca a ea o Sox9 l/ l and Sox9 l-Pe –C e hea issue 7 days a e MI by pe manen liga ion o he le an e io descending a e y showing SOX9 (g een) in ca diac ib oblas s (s ained o PDGFR-α, ed); nuclei a e s ained wi h DAPI (blue). Scale ba s: 50 μm. A o al o n = 6 hea s pe g oup we e s ained, and a leas 4 images pe hea we e analyzed o C. (D–F) Wes e n blo analysis (D) and den- si ome ic quan i ica ion (E and F) illus a ing a highly e icien dele ion o SOX9 p o ein om Sox9 l-Pe –C e mouse ca diac sca issue 42 days a e MI. SOX9 p o ein le els in emo e samples a e MI show no signi ican di e ences. GAPDH was used as a loading con ol. Da a a e shown as mean ± SEM. n = 3 mice/g oup. Two- ailed S uden ’s es was used o compa ison o 2 g oups; *P < 0.05, **P < 0.01. 5 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE pou ed in o a collagen gel ma ix. A e 48 hou s, con ac ion was quan i ied as he di e ence be ween he gel a ea a e 48 hou s and 0 hou s, which showed ha gel ma ices wi h Sox9-de icien ib oblas s we e signi ican ly less con ac ed (indica ing less myo ib oblas abundance) han he con ol samples (Figu e 5D). Mo eo e , cell mig a ion was analyzed unde baseline condi ions o a e s imula ion wi h angio ensin II, TGF-β1, o FBS. The mig a ion le el, quan i ied as closu e o a de ined sc a ch a ea wi h- in 24 hou s, was educed in Sox9-de icien ca diac ib oblas s unde baseline condi ions as well as a e s imula ion wi h angio ensin II, TGF-β1, o FBS (Figu e 5, E–H). O e all, we ound ha Sox9 de iciency causes less ac i e ca diac ib oblas s compa ed wi h con ol cells in i o. Nex , we analyzed he o ma ion o s ess ibe s as a key ma ke o myo ib oblas di e en ia ion in iso- la ed ca diac ib oblas s wi h Sox9 dele ion compa ed wi h con ol cells, unde baseline condi ions o a e s imula ion wi h TGF-β1 o 24 hou s o induce myo ib oblas di e en ia ion. Unexpec edly, luo escence s aining o ilamen ous ac in wi h phalloidin and consecu i e quan i ica ion o s ess ibe –posi i e ib o- blas s ailed o e eal di e ences be ween Sox9-de icien ca diac ib oblas s and con ol cells (Supplemen al Figu e 1, C and D). Mo eo e , mRNA le els o key myo ib oblas di e en ia ion ma ke s, pa icula ly Ac a2 and o he s (Fmn1, Cnn1, Tagln), we e simila ly exp essed be ween Sox9-de icien ca diac ib oblas s and con ol cells (Supplemen al Figu e 1, E–H). Fib oblas SOX9 con ols ECM o ma ion and in lamma ion in he ch onic s age a e MI in i o. Nex , we in es- iga ed he e ec o ib oblas -speci ic Sox9 dele ion in he la e phase a e MI when he o ma ion o he ch onic sca is comple ed and myo ib oblas s a e u he di e en ia ed in o quiescen ma i ib ocy es (25). Figu e 3. Imp o ed ca diac unc ion o mice wi h ib oblas -speci ic Sox9 dele ion a e MI. Mo phological analysis and unc ional e alua ion o Sox9 l/ l and Sox9 l-Pe –C e mice 7 days and 42 days a e MI by pe manen liga ion o he le an e io descending a e y. (A and B) Quan i ica ion o he HW/BW a io (A) shows an inc ease in Sox9 l/ l and Sox9 l-Pe –C e mice 42 days a e MI; he LW/BW a io (B) shows no signi ican di e ences. (C–F) Echoca diog aphy showed an imp o ed ejec ion ac ion in Sox9 l-Pe –C e mice compa ed wi h Sox9 l/ l mice 7 (C) and 42 (D) days a e MI. Mo eo e , Sox9 l-Pe –C e mice exe ed less le en icula dila a ion, i.e., inc eased le en icula end-dias olic a ea (LVEDA), in ela ion o Sox9 l/ l mice 7 (E) and 42 (F) days a e MI. Sham-ope a ed mice had no impai men o ca diac unc ion. Da a show mean ± SEM. n = 4 (Sox9 l/ l and Sox9 l-Pe –C e; sham), n = 7 (Sox9 l/ l; MI), and n = 11 (Sox9 l-Pe – C e; MI). Two-way ANOVA wi h Holm-Šídák’s mul iple-compa isons es ; *P < 0.05, **P < 0.01, ***P < 0.001. 6 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE Masson’s ich ome s aining o he in a c ed hea s 42 days a e MI e ealed ha he sca a ea o Sox9 l- Pe –C e mice was s ill signi ican ly smalle compa ed wi h con ols a his la e s age (Figu e 6, A and B). To iden i y SOX9-dependen gene egula ion in he ch onic myoca dial sca , we pe o med a genome-wide ansc ip omic analysis o ca diac sca issue om Sox9 l-Pe –C e and Sox9 l/ l mice 42 days a e MI by RNA-Seq. Genes wi h signi ican di e en ial egula ion be ween bo h condi ions (a leas 50% up- o down egula ed) a e demons a ed in he hea map (Figu e 6C and Supplemen al Figu e 2A). In he ch onic in a c sca o Sox9 l-Pe –C e mice, 2611 genes we e down egula ed and belonged o he GO (biological p ocess) classes ca ilage de elopmen , cell chemo axis, ex acellula ma ix o ganiza ion, immune esponse, and p o eolysis (Figu e 6C). Highly down egula ed genes in he myoca dial sca o Sox9 l-Pe –C e mice included mainly collagen genes (Col2a1, Col9a2, Col11a1), in lamma o y media o s (Cxcl13, Il6, Tn s 15), and p o eases (Adam32, P ss23, Capn6; o he example genes a e lis ed in Figu e 6C and in Supplemen al Figu e 2, B–D). In he sca o Sox9 l-Pe –C e mice, 1218 genes we e up egula ed, which we e ela ed mainly o oxida i e cellula me abolism (GO classes ATP me abolic p ocess, ica boxylic acid cycle, and a y acid me abolic p ocess), o example, A p5b, Idh3a, and Acadl, bu we also ound genes o he ca ego y muscle con ac ion, o example, Myh6, Tnni3, and T n (mo e example genes a e lis ed in Figu e 6C) wi hin his g oup. We subsequen ly Figu e 4. Fib oblas -speci ic Sox9 dele ion educes he sca a ea and a enua es he ac i a ion o ca diac ib oblas s a e MI. (A and B) Masson’s ich ome s aining o ep esen a i e mid en icula sec ions (A) and quan i ica ion o he sca a ea (B) o Sox9 l-Pe –C e (n = 9) compa ed wi h Sox9 l/ l (n = 7) mouse hea s 7 days a e MI. Scale ba s: 1 mm. (C and D) Rep esen a i e IHC images showing Ki67 (g een) and PDGFR-α ( ed) as ma ke s o ib oblas p oli e a ion in he in a c egion o Sox9 l/ l and Sox9 l-Pe –C e mouse hea s 7 days a e MI (C) and quan i ica ion (D). Nuclei a e shown wi h DAPI (blue). A ows indica e exempla y Ki67+ ib oblas s. Scale ba s: 100 μm. A o al o 5 hea s pe g oup we e s ained, and 4 images pe hea we e analyzed o D. (E) Rep esen a i e IHC images showing SOX9 (g een) and α-SMA ( ed) in he emo e and in a c a ea o Sox9 l/ l and Sox9 l-Pe –C e mouse hea s 7 days a e MI. Nuclei a e shown wi h DAPI (blue). Rep esen a i e images o 5 hea s pe g oup a e shown. Scale ba s: 100 μm. (F) mRNA exp ession o Ac a2 measu ed wi h e e se ansc ip ion PCR in samples o he emo e and in a c a ea o Sox9 l/ l and Sox9 l-Pe –C e mouse hea s 7 days a e MI. Da a we e no malized o Ac a2 in he emo e a ea o Sox9 l/ l mice 7 days a e MI. n = 5 (Sox9 l/ l) and 6 (Sox9 l-Pe –C e). Da a a e shown as mean ± SEM. S uden ’s es was used o compa ison o 2 g oups (B and D); 2-way ANOVA wi h Holm-Šídák’s mul iple-compa isons es was used o F. P alue o Ac a2 in he in a c a ea o Sox9 l/ l and Sox9 l-Pe –C e was 0.11 (indica ed as n.s. in F). **P < 0.01, ****P < 0.0001. 7 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE ocused on he genes ha we e down egula ed because o Sox9 dele ion in Sox9 l-Pe –C e mice because o hese genes di ec egula ion by SOX9 was mo e likely. To in es iga e whe he hese down egu- la ed genes migh be di ec a ge s o SOX9, we in e sec ed ou RNA-Seq da a wi h SOX9 ChIP-Seq esul s om de eloping mammalian chond ocy es (26). In ha s udy, SOX9 binding egions we e cha ac e ized h oughou he genome, and SOX9 was ound o bind close o he p omo e in an indi- ec manne (“class I engagemen ,” ound mos ly in genes wi h gene al cellula unc ions) o o bind di ec ly o DNA in enhance egions o ca ilage- ela ed genes, which we e in la ge pa iden ical o he ECM genes in ou s udy. In e sec ion wi h he down egula ed genes om ou da a se e ealed ha 50.5% o he down egula ed genes exe ed di ec SOX9 binding, whe eby 43% showed class II engage- men o bo h class I and class II engagemen . GO e m analysis in hese 43% o down egula ed genes e lec ed e y simila GO classes (leukocy e mig a ion, cell chemo axis, ex acellula ma ix o gani- za ion; Supplemen al Figu e 3A) as obse ed among all down egula ed genes in he myoca dial sca o Sox9 l-Pe –C e mice. KEGG pa hway analysis e ealed sligh ly mo e class I engagemen alone bu Figu e 5. In i o Sox9 dele ion in ca diac ib oblas s educes mig a ion, p oli e a ion, and con ac ili y. Isola ed ca diac ib oblas s om uninju ed Sox9 l/ l mouse hea s we e in ec ed wi h adeno i us exp essing C e ecombinase (AdC e) o gene a e in i o Sox9 loss o unc ion o con ol adeno i- us (Adβ-gal). (A and B) E iciency o SOX9 dele ion is demons a ed on he p o ein le el by Wes e n blo (A) and densi ome ic quan i ica ion (B). (C) P oli e a ion o ca diac ib oblas s wi h Sox9 dele ion compa ed wi h con ol cells was measu ed by B dU inco po a ion (n = 24 samples/g oup). (D) Con- ac ion capaci y o ca diac ib oblas s quan i ied as a ea change o a collagen ma ix con aining ib oblas s wi h Sox9 dele ion o con ol cells a e 48 hou s compa ed wi h baseline a 0 (n = 6 samples/g oup). (E–H) Mig a ion o Sox9 l/ l ib oblas s ea ed wi h ei he Adβ-gal (β-gal) o AdC e (C e) unde baseline condi ions (uns imula ed, E) o in he p esence o p o ib o ic s imuli angio ensin II (Ang II, F), TGF-β1 (G), o 10% FBS (H), quan i ied as sc a ch closu e a e 24 hou s. Rep esen a i e pic u es a e shown di ec ly a e sc a ch and 24 hou s la e ; n = 12 samples/g oup. Scale ba s: 400 μm. Da a a e shown as mean ± SEM. Two- ailed S uden ’s es was used o compa ison o 2 g oups; **P < 0.01; ***P < 0.001. 8 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE iden i ied o he wise simila unc ional classes (cy okine-cy okine ecep o in e ac ion, ECM- ecep o in e ac ion, p o ein diges ion and abso p ion; Supplemen al Figu e 3B). Supplemen al Figu e 4 shows exempla y esul s o class I and class II binding o SOX9 o selec ed ECM genes (Col2a1, Col9a2, and Fn1; Supplemen al Figu e 4, A–C), in lamma o y genes (Cxcl13 and Il6; Supplemen al Figu e 4; D and E), and a p o eolysis gene (Mmp2; Supplemen al Figu e 4F) in conjunc ion wi h ou RNA-Seq esul s. These da a imply a di ec in e ac ion o SOX9 wi h he p omo e o enhance egions o ECM, p o e- olysis, o in lamma o y genes iden i ied in his s udy. Because RNA-Seq showed dec eased exp ession o in lamma o y media o s in he sca egion o Sox- 9 l-Pe –C e mice, we in es iga ed leukocy e abundance 6 weeks a e MI, which was s ongly diminished upon educed ib oblas SOX9 le els bo h in he sca as well as in he emo e a ea, al hough in he sca leukocy e abundance was gene ally highe in bo h con ol and Sox9 l-Pe –C e mice (Figu e 7, A and B). As one example o he ECM- ela ed gene exp ession, we con i med educed COL2A1 p o ein abundance in he sca o Sox9 l-Pe –C e mice by Wes e n blo and immuno luo escence s aining (Figu e 7, C–E). Figu e 6. Fib oblas -speci ic Sox9 dele ion educes he sca a ea and inhibi s exp ession o p oin lamma o y and ECM genes a e long- e m MI. (A and B) Masson’s ich ome s aining o ep esen a i e mid en icula sec ions (A) and quan i ica ion o he sca a ea (B) (n = 6–11) o Sox9 l-Pe –C e compa ed wi h Sox9 l/ l mouse hea s 42 days a e MI. Scale ba s: 1 mm. (C) Gene exp ession in sca a ea issue o Sox9 l-Pe –C e compa ed wi h Sox9 l/ l mouse hea s 42 days a e MI was analyzed wi h RNA-Seq. Hea map o di e en ially egula ed genes in he MI sca o Sox9 l/ l and Sox9 l-Pe –C e mice (n = 3) 42 days a e MI is shown on he le o C. Gene On ology (GO, biological p ocess) classi ica- ion o highly egula ed p ocesses and selec ed example genes a e demons a ed on he igh . Bioin o ma ic analysis o RNA-Seq da a is desc ibed in he Me hods sec ion. Da a a e shown as mean ± SEM. S uden ’s es was used o compa ison o 2 g oups; *P < 0.05. 15 insigh .jci.o g h ps://doi.o g/10.1172/jci.insigh .126721 RESEARCH ARTICLE 26. Ohba S, He X, Hojo H, McMahon AP. Dis inc ansc ip ional p og ams unde lie Sox9 egula ion o he mammalian chond ocy e. Cell Rep. 2015;12(2):229–243. 27. F angogiannis NG. The mechanis ic basis o in a c healing. An ioxid Redox Signal. 2006;8(11–12):1907–1939. 28. Shinde AV, F angogiannis NG. Fib oblas s in myoca dial in a c ion: a ole in in lamma ion and epai . J Mol Cell Ca diol. 2014;70:74–82. 29. Lindne D, e al. Ca diac ib oblas s suppo ca diac in lamma ion in hea ailu e. Basic Res Ca diol. 2014;109(5):428. 30. Sa o Y, Yanagi a M. Residen ib oblas s in he kidney: a majo d i e o ib osis and in lamma ion. In lamm Regen. 2017;37:17. 31. Kanisicak O, e al. Gene ic lineage acing de ines myo ib oblas o igin and unc ion in he inju ed hea . Na Commun. 2016;7:12260. 32. Kamachi Y, Kondoh H. Sox p o eins: egula o s o cell a e speci ica ion and di e en ia ion. De elopmen . 2013;140(20):4129–4144. 33. P i che J, A hwal V, Robe s N, Hanley NA, Hanley KP. Unde s anding he ole o SOX9 in acqui ed diseases: lessons om de elopmen . T ends Mol Med. 2011;17(3):166–174. 34. Fu ado MB, e al. Ca diogenic genes exp essed in ca diac ib oblas s con ibu e o hea de elopmen and epai . Ci c Res. 2014;114(9):1422–1434. 35. C eeme s EE, Cleu jens JP, Smi s JF, Daemen MJ. Ma ix me allop o einase inhibi ion a e myoca dial in a c ion: a new app oach o p e en hea ailu e? Ci c Res. 2001;89(3):201–210. 36. Fon ana L, e al. Fib onec in is equi ed o in eg in α β6-media ed ac i a ion o la en TGF-β complexes con aining LTBP-1. FASEB J. 2005;19(13):1798–1808. 37. Ge G, G eenspan DS. BMP1 con ols TGFβ1 ac i a ion ia clea age o la en TGFβ-binding p o ein. J Cell Biol. 2006;175(1):111–120. 38. Ammi an e M, Shalapou S, Kang Y, Jamieson CA, Ka in M. Tissue inju y and hypoxia p omo e malignan p og ession o p os a e cance by inducing CXCL13 exp ession in umo myo ib oblas s. P oc Na l Acad Sci U S A. 2014;111(41):14776–14781. 39. Kis R, Sch ewe H, Balling R, Sche e G. Condi ional inac i a ion o Sox9: a mouse model o campomelic dysplasia. Genesis. 2002;32(2):121–123. 40. Lindsley A, e al. Iden i ica ion and cha ac e iza ion o a no el Schwann and ou low ac endoca dial cushion lineage- es ic - ed pe ios in enhance . De Biol. 2007;307(2):340–355. 41. F oese N, e al. An i-and ogenic he apy wi h inas e ide imp o es ca diac unc ion, a enua es emodeling and e e s pa holog- ic gene-exp ession a e myoca dial in a c ion in mice. J Mol Cell Ca diol. 2018;122:114–124. 42. Heinz S, e al. Simple combina ions o lineage-de e mining ansc ip ion ac o s p ime cis- egula o y elemen s equi ed o mac- ophage and B cell iden i ies. Mol Cell. 2010;38(4):576–589. 43. Ta azona S, Fu ió-Ta í P, Tu à D, Pie o AD, Nueda MJ, Fe e A, Conesa A. Da a quali y awa e analysis o di e en ial exp ession in RNA-seq wi h NOISeq R/Bioc package. Nucleic Acids Res. 2015;43(21):e140. 44. Ramí ez F, Dünda F, Diehl S, G üning BA, Manke T. deepTools: a lexible pla o m o explo ing deep-sequencing da a. Nucleic Acids Res. 2014;42(Web Se e issue):W187–W191.. 45. Appa i M, e al. C1q-TNF- ela ed p o ein-9 p omo es ca diac hype ophy and ailu e. Ci c Res. 2017;120(1):66–77. 46. Da is J, Bu AR, Da is GF, Bi nbaume L, Molken in JD. A TRPC6-dependen pa hway o myo ib oblas ansdi e en ia ion and wound healing in i o. De Cell. 2012;23(4):705–715.