Involvement of Ubiquitin-Editing Protein A20 in Modulating Inflammation in Rat Cochlea Associated with Silver Nanoparticle-Induced CD68 Upregulation and TLR4 Activation

NANO EXPRESS Open Access In ol emen o Ubiqui in-Edi ing P o ein A20 in Modula ing In lamma ion in Ra Cochlea Associa ed wi h Sil e Nanopa icle-Induced CD68 Up egula ion and TLR4 Ac i a ion Hao Feng 1 , Ilma i Pyykkö 1 and Jing Zou 1,2* Abs ac Sil e nanopa icles (AgNPs) we e shown o empo a ily impai he biological ba ie s in he skin o he ex e nal ea canal, mucosa o he middle ea , and inne ea , causing pa ially e e sible hea ing loss a e deli e y in o he middle ea . The cu en s udy aimed o elucida e he molecula mechanism, emphasizing he TLR signaling pa hways in associa ion wi h he po en ial ec ui men o mac ophages in he cochlea and he modula ion o in lamma ion by ubiqui in-edi ing p o ein A20. Molecules po en ially in ol ed in hese signaling pa hways we e ho oughly analysed using immunohis ochemis y in he a cochlea exposed o AgNPs a a ious concen a ions h ough in a ympanic injec ion. The esul s showed ha 0.4 % AgNPs bu no 0.02 % AgNPs up egula ed he exp essions o CD68, TLR4, MCP1, A20, and RNF11 in he s ial basal cells, spi al ligamen ib ocy es, and non-senso y suppo ing cells o Co i’s o gan. 0.4 % AgNPs had no e ec on CD44, TLR2, MCP2, Rac1, myosin ligh chain, VCAM1, E k1/2, JNK, p38, IL-1β,TNF-α, TNFR1, TNFR2, IL-10, o TGF-β.This s udy sugges ed ha AgNPs migh con e mac ophage-like unc ions on he s ial basal cells and spi al ligamen ib ocy es and enhance he immune ac i i ies o non-senso y suppo ing cells o Co i’so gan h ough he up egula ion o CD68, which migh be in ol ed in TLR4 ac i a ion. A20 and RNF11 played oles in main aining cochlea homeos asis ia nega i e egula ion o he exp essions o in lamma o y cy okines. Backg ound Wi h he apid de elopmen o nano echnology and inc easing applica ions o enginee ed nanoma e ials in ou daily li es, hei po en ial sa e y issues ha e be- come a se ious conce n in public heal h. The a ea model has been applied o in es iga e he impac o sil e nanopa icles (AgNPs) on he pe meabili y o biological ba ie s in he skin, mucosa, and inne ea ha is analogous o he ne ous sys em (e.g. he b ain and spinal co d) [1]. P e ious esea ch showed ha AgNPs led o hyalu onan accumula ion in he cochlea, impai ed biological ba ie s in he skin o he ex e nal ea canal, mucosa o he middle ea , and inne ea , and conse- quen ly caused hea ing loss a e deli e y in o he middle ea [1–3]. Hyalu onan ac s as an endogenous pa hogen- associa ed molecula pa e n (PAMP) in esponse o haza dous signals h ough binding hyalu onan-binding p o eins (hyaladhe ins) including oll-like ecep o s 2/4 (TLR2/4), CD44, ecep o o hyalu onan-media ed mo ili y, and umou nec osis ac o -α(TNF-α)-s imula ed glycop o ein-6 [4–7]. Among hem, TLR2/4 is a ca ego y o mammalian homologues o D osophila Toll p o eins ha a e o g ea impo ance o inna e hos de ence. They belong o he pa e n ecogni ion ecep o s (PRRs) ha speci ically ecognize and espond o an expansi e a - ie y o PAMPs [8]. Mo eo e , TLR4 is esponsible o sensing dange /damage-associa ed molecula pa e ns * Co espondence: [email p o ec ed] 1 Hea ing and Balance Resea ch Uni , Field o O o-la yngology, School o Medicine, Uni e si y o Tampe e, Medisiina inka u 3, 33520 Tampe e, Finland 2 Depa men o O ola yngology-Head and Neck Su ge y, Cen e o O ola yngology-Head and Neck Su ge y o Chinese PLA, Changhai Hospi al, Second Mili a y Medical Uni e si y, Shanghai, China © 2016 Feng e al. Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 DOI 10.1186/s11671-016-1430-9 (DAMPs) and con e ing immunos imula o y ac i i y [9]. The ac i a ion o TLRs ini ia es he up egula ion o ansc ip ion ac o s (e.g. nuclea ac o -κB(NF-κB) and ac i a o p o ein-1) ha play pi o al oles in p o- ducing in lamma o y molecules (e.g. in e leukin-1β(IL-1β), in e leukin-6 (IL-6), and TNF-α oge he wi h i s ecep o s TNFRs), chemokines (e.g. monocy e chemoa ac an p o- eins (MCPs)), and eac i e oxygen/ni ogen species, lead- ing o in lamma o y diseases [10–12]. Se e al p o eins ha a e implica ed in media ing TLR signaling a enua ion ha e been iden i ied such as he ubiqui in-edi ing p o ein A20 [13–15]. A20 ac s as a nega i e e ec o in egula ing TLR-media ed in lamma- o y esponse, and i s o e exp ession inhibi s TLR2- and TLR4-media ed IL-8 syn heses in ai way epi helial cells [16]. A20 loss ele a es he le els o NF-κB- egula ed in lamma o y cy okines and causes spon aneous ce eb al in lamma ion [17]. RING inge p o ein 11 (RNF11), a c i ical componen o A20, is indica ed as one o he key nega i e egula o s in con olling he NF-κB signaling pa hway. RNF11 was shown o p o ec mic oglia i i- a ed by lipopolysaccha ide h ough manipula ing he NF-κB signaling pa hway [18]. RNF11 knockdown in he monocy es led o pe sis en TNF- and lipopolysaccha ide- media ed NF-κB signaling ac i a ion and up egula ed NF- κB-associa ed in lamma o y gene ansc ip s [18, 19]. As ano he impo an hyaladhe in, CD44 is capable o ec ui ing monocy es om he pe iphe al blood upon hyalu onan binding [20]. Fu he s udy has e ealed ha weakened in e ac ion be ween CD44 and hyalu onan dec eases he p oduc ion o MCPs and consequen ly unde mines he ec ui men o mononuclea cells [21]. MCPs a e a amily o small hepa in-binding, posi i ely cha ged chemokines ha play an indispensable ole in con olling cell beha iou in esponse o exogenous s imula ion. They a e c ucial in igge ing he mobiliza ion and mig a ion o immunocompe en cells such as mono- cy es, neu ophils, lymphocy es, and dend i ic cells along he bone ma ow sinusoids ha equen ly anas omose wi h capilla ies and in di ec ing hem in o he in lamed issues [22]. In he inne ea , spi al ligamen ib ocy es ac as he p ima y immune senso s in esponse o lipopoly- saccha ide, in ol ing TLR2-dependen NF-κB signaling ac i a ion and MCP1 up egula ion and esul ing in mono- cy e mig a ion and consequen ial in il a ion [23, 24]. Adhesion molecules play a c i ical ole in media ing leukocy e immobiliza ion as a esul o ancho ing [25]. Among hem, ascula cell adhesion molecule 1 (VCAM1) enables olling monocy es along he mic o ascula wall a a a slowe eloci y o adhe e o he endo helial cells [26]. Rac1, a membe o Rho-like small GTPase, media ed by he phospho yla ion o myosin ligh chain p o ein, a- cili a es ac in cy oskele al emodelling and modula es igh junc ional p o eins (e.g. occludin and claudin). The b eakdown o igh junc ion in he mic o ascula wall enables he leukocy es o in il a e in o he a ge ing si e [27–29]. The ex acellula signal- egula ed kinases 1/2 (E k1/2), c-Jun N- e minal kinases 1/2/3 (JNK1/2/3) (also known as s ess-ac i a ed p o ein kinases), and p38 iso- o ms (α,β,γ,andδ) ha belong o heMAPKs amilya e conside ed o be he elemen a y componen s o cellula signaling ansduc ion unde lying leukocy e locomo ion and endo helial cell ac i i ies [30, 31]. Mig a ed monocy es can di e en ia e in o mac ophages. Plas ici y and lexibili y a e he key ea u es o mac o- phages and e lec hei ac i a ion s a es [32]. Ac i a ed mac ophages ha e dis inc i e unc ional pheno ypes ha a e simila o he Th1/Th2 pola iza ion pa adigm o T lymphocy es and can be de ined as M1 and M2. M1 in- duced by Th1 signa u e cy okines (e.g. in e e on-γ(IFN-γ) and TNF-α), which a e associa ed wi h he TLR-dependen signaling pa hway, has he abili y o up egula ing genes in ol ed in cell-biased immuni y, enhancing an igen p es- en a ion, and p oducing a dis inc i e a ay o in lamma o y cy okines (e.g. IL-1β,IL-6,andTNF-α). M2 induced by Th2 signa u e cy okines (e.g. IL-4 and IL-13) plays an im- po an ole in immune supp ession, an i-in lamma ion (e.g. in e leukin-10 (IL-10)), issue egene a ion, and wound healing (e.g. ans o ming g ow h ac o -β(TGF-β)and ascula endo helial g ow h ac o (VEGF)) [33, 34]. The cu en s udy aimed o elucida e he exac mechan- ism o AgNP-induced biological ba ie unc ional changes in he inne ea . We exposed he a inne ea o AgNPs and hypo hesized ha TLR signaling pa hways we e in- ol ed in AgNP-induced hea ing loss in associa ion wi h he po en ial ec ui men o mac ophages in he a coch- lea. A20 migh play a ole in egula ing he downs eam signaling o TLR pa hways. Molecules po en ially in- ol ed in hese signaling pa hways we e ho oughly analysed using immunohis ochemis y in he a coch- lea a e AgNP exposu e. Me hods Animal and AgNPs Ten albino male Sp ague-Dawley a s weighing be ween 250 and 300 g we e kep a an ambien empe a u e o 20–22 °C wi h a ela i e humidi y o 50 ± 5 % unde a 12/12-h ligh /da k cycle in he expe imen al animal uni , Uni e si y o Tampe e. The expe imen s we e pe o med unde gene al anaes hesia wi h a mix u e o 0.5 mg/kg mede omidine hyd ochlo ide (Domi o ®, O ion, Espoo, Finland) and 75 mg/kg ke amine hyd ochlo ide (Ke ala ®, P ize , Helsinki, Finland) adminis e ed ia in ape i oneal injec ion, ollowed by in amuscula injec ion o en o- loxacin (Bay il® e , O ion, Tu ku, Finland) a a dose o 10 mg/kg o p e en po en ial in ec ion. The animals’ eyes we e p o ec ed by ca bome (Visco ea s®, No a is Heal hca e A/S, Denma k). All p ocedu es in he s udy Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 2 o 13 complied wi h he local e hics commi ee s anda ds (pe mission numbe : ESAVI/3033/04.10.03/2011) and we e conduc ed in acco dance wi h Eu opean Legisla- ion. The AgNPs (Colo obbia, Fi enze, I aly) used in his s udy we e highly ace ed wi h a mean size o 21 ± 8 nm using ansmission elec on mic oscope. The mean hyd odynamic size o he pa icles was 117 ± 24 nm when suspended in deionized wa e (dH 2 O) using dynamic ligh sca e ing, and he ze a po en ial was measu ed o be −20 ± 9 mV [2]. Mo e esul s o he cha ac e iza ion could be e e ed o ou p e ious s udy [2]. AgNP Adminis a ion A e anaes he iza ion, 40 μl o ei he 0.4 (n=5) o 0.02 % (n= 5) AgNPs we e injec ed in o he middle ea ca i y unde an ope a ing mic oscope (OPMI1-F, Ca l Zeiss, Jena, Ge many) acco ding o a p e iously epo ed p ocedu e [1–3]. The es ed concen a ions we e selec ed acco ding o he audi o y b ains em esponse esul s showing ha 0.4 % AgNPs caused e e sible hea ing loss ha pa ially eco e ed a he se en h day, while 0.02 % AgNPs only induced hea ing loss a 32 kHz ha e u ned o he baseline a he se en h day. Mo eo e , mic o-CT scanning displayed ha 0.4 % AgNPs caused an ob ious middle ea in il a ion ha was absen in he a s exposed o 0.02 % AgNPs [1–3]. The con ala e al ea (n= 10) e- cei ed dH 2 O unde he same ci cums ances and was used as a nega i e con ol. Sample P epa a ion On he se en h day pos -injec ion, he anaes he ized a s we e pe used wi h 0.01 M pH 7.4 phospha e-bu e ed saline (PBS) con aining 0.6 % ( / ) hepa in (LEO Pha ma A/S, Balle up, Denma k) ia a ca diac app oach ollowed by 4 % pa a o maldehyde (Me ck, Espoo, Finland) o ix he head. The bullae we e isola ed a e decapi a ion and decalci ied using 10 % EDTA (Sigma-Ald ich, S einheim, Ge many) in he ollowing 4 weeks wi h weekly solu ion changes. A s anda d p ocedu e o pa a in embedding and issue block was conduc ed in acco dance wi h he p o ocol in a p e ious s udy [3]. Immuno luo escence S aining The p ocedu e o immuno luo escence s aining was in acco dance wi h he p o ocol in a p e ious s udy [3]. The p ima y an ibodies used in he assay we e hos ed in abbi and we e an i-CD68 (1:200, Abcam, UK), an i- CD44 (1:400, Abcam, UK), an i-TLR2 (1:250, No us Biologicals, UK), an i-TLR4 (1:200, No us Biologicals, UK), an i-MCP1 (1:4000, No us Biologicals, UK), an i- MCP2 (1:200, GeneTex, USA), an i-Rac1 (1:800, Abcam, UK), an i-myosin ligh chain (1:100, Cell Signaling Tech- nology, USA), an i-VCAM1 (1:50, P o ein ech, USA), an i-E k1/2 (1:400, Abcam, UK), an i-JNK (1:100, Cell Signaling Technology, USA), an i-p38 (1:100, Cell Signal- ing Technology, USA), an i-TNF-α(1:800, Abcam, UK), an i-TNFR1 (1:500, Abcam, UK), an i-TNFR2 (1:50, Abcam,UK),an i-IL-1β(1:400, No us Biologicals, UK), an i-IL-10 (1:400, Abbio ec, USA), an i-TGF-β(1:500, Abcam, UK), an i-A20 (1:200, Sigma-Ald ich, USA), and an i-RNF11 (1:100, Abcam, UK). B ie ly, he slices we e incuba ed wi h he p ima y an ibodies lis ed abo e a 4 °C o e nigh ollowed by Alexa Fluo ® 488 Goa An i-Rabbi IgG (1:200, dilu ed wi h 0.1 % BSA, Li e Technologies™, New Yo k, USA) as a seconda y an i- body a oom empe a u e o 1 h in a da k en i onmen . The nuclei we e coun e s ained wi h 10 μg/ml DAPI (Li e Technologies™, New Yo k, USA) a oom empe a u e o 10 min, and he slides we e moun ed o con ocal mic os- copy wi h an i-quenching luo omoun (Sigma-Ald ich, S . Louis, USA). In he nega i e con ol slices, he p ima y an ibodies we e eplaced wi h 0.1 % BSA (dissol ed in 0.01 M PBS pH 7.4; Sigma-Ald ich, S . Louis, USA). Immunos aining Visualized by 3,3′-Diaminobenzidine A e depa a iniza ion and hyd a ion, he slices we e imme sed in 3 % H 2 O 2 -me hanol a oom empe a u e o 30 min. A e insing wi h PBS o 2 × 2 min, he slices we e diges ed wi h 0.1 % ypsin (dissol ed in 0.01 M PBS pH 7.4; Sigma-Ald ich, S . Louis, USA) a 37 °C o 30 min. A e insing wi h 0.1 % PBS-Tween® 20 (dilu ed in 0.01 M PBS pH 7.4; Sigma-Ald ich, S . Louis, USA) o 3 × 2 min, he slices we e incuba ed wi h 10 % no mal goa se um (In i ogen, Paisley, UK) a oom empe a u e o 30 min ollowed by he p ima y an ibodies lis ed abo e a 4 °C o e nigh . A e insing wi h 0.1 % PBS-Tween® 20 o 3 × 2 min, he slices we e incuba ed wi h bio inyla ed goa an i- abbi IgG a a di- lu ion o 1:100 (Vec o Labo a o ies L d., Pe e bo ough, UK) a oom empe a u e o 1 h. A e insing wi h 0.1 % PBS-Tween® 20 o 3 × 2 min, he slices we e incu- ba ed wi h he s ep a idin-bio in-pe oxidase complex (Vec o Labo a o ies L d., Pe e bo ough, UK) a 37 °C o 1h.A e insingwi h0.1%PBS-Tween®20 o 3 × 5 min, an ibody binding was isualized by 3,3′-di- aminobenzidine using he DAB Pe oxidase Subs a e Ki (Vec o Labo a o ies L d., Pe e bo ough, UK) a oom empe a u e o 5 min. Al e na i ely, he nuclei we e coun e s ained using Ha is’s solu ion (Me ck, Da ms ad , Ge many). Dehyd a ion and i i ica ion we e comple ed by a s anda d p o ocol (70 % alcohol 10 s, 94 % alcohol 2 × 10 s, absolu e alcohol 2 × 1 min, and xylene 3 × 3 min). The slides we e moun ed o ligh mic oscopy wi h Cla ion™Moun ing Medium (Sigma-Ald ich, S . Louis, USA). Slices o nega i e con ols we e p epa ed a e he eplacemen o p ima y an ibodies wi h 0.1 % BSA (dis- sol ed in 0.01 M PBS pH 7.4; Sigma-Ald ich, S . Louis, USA). The s aining in ensi ies (shown by he g eyscale Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 3 o 13 alue ha was in e sely co ela ed wi h he s aining in en- si y) in he s ial basal cells, spi al ligamen ib ocy es, and spi al ganglion cells we e measu ed and semi-quan i ied using ImageJ 1.45S so wa e (NIH, Be hesda, USA). Con ocal and Ligh Mic oscopies The samples om immuno luo escence s aining we e obse ed and images ob ained unde a Nikon mic o- scope (ECLIPSE Ti) combined wi h an Ando con ocal sys em ins alled wi h Ando iQ 2.8 so wa e (Ando Technology, Bel as , UK). The exci a ion lase s we e 405 (blue exci a ion) and 488 nm (g een exci a ion) om an Ando Lase Combine sys em, and he co esponding emission il e s we e 450–465 (DAPI) and 525/50 nm (FITC), espec i ely. The immunos ained samples isual- ized by 3,3′-diaminobenzidine we e obse ed unde a ligh mic oscope (LEICA DM 2000, Espoo, Finland), and images we e digi ally pho og aphed using a came a ideo (Olympus DP 25, Tokyo, Japan) wi h he cellSens Di- mension 1.6 Olympus so wa e (Olympus Co po a ion, Tokyo, Japan) ins alled. Analysis and S a is ics S a is ical analyses we e pe o med using he IBM® SPSS® S a is ics Ve sion 20 so wa e package (SPSS Inc., Chicago, USA). One-way ANOVA was used o compa e he s ain- ing in ensi ies o CD68, TLR2, TLR4, MCP1, MCP2, A20, and RNF11 in he designa ed s uc u es o di e en coch- lea u ns among he cochleae exposed o 0.4 % AgNPs, 0.02 % AgNPs, and dH 2 O. The LSD pos hoc es was used o e alua e he pai wise di e ence. The independen sample es was used o compa e he s aining in ensi ies o CD44, Rac1, E k1/2, IL-1β,IL-10,andTGF-βin he designa ed s uc u es o di e en cochlea u ns be ween he cochleae exposed o 0.4 % AgNPs and dH 2 O. A alue o p< 0.05 indica ed ha he di e ence was s a is ically signi ican . Resul s AgNPs Augmen he Sensi i i y and Chemo ac ic P o eins o Cochlea Cells In he cochleae exposed o dH 2 O, he inne hai cells and pilla cells o Co i’s o gan showed mode a e s ain- ing o CD68, while he ou e hai cells and Dei e s’ cells demons a ed ex emely weak s aining o CD68 (Fig. 1h). The s ial basal cells, spi al ligamen ib o- cy es, and spi al ganglion cells exhibi ed mild s aining o CD68 (Fig. 1d, ). In he cochlea la e al wall, 0.4 % AgNPs in ensi ied CD68 s aining ema kably in he s ial basal cells (p< 0.01, pos hoc es ) and spi al ligamen ib ocy es (mainly ype III) (p< 0.01, pos hoc es ) in he i s u n (Fig. 1a). Howe e , no en- hanced s aining was obse ed in cells in he second and hi d u ns (Fig. 1b, c) (p>0.05, pos hoc es ). In he CD68 + cell popula ion, spa se ami ied cells and mononuclea cells we e iden i ied in he spi al ligamen and he modiolus, espec i ely (Fig. 1c, i). In Co i’s o gan, 0.4 % AgNPs inc eased CD68 s aining in he inne hai cells and pilla cells bu no in he ou e hai cells and Dei e s’cells (Fig. 1g). In he spi al ganglion cells and capilla y endo helial cells, 0.4 % AgNPs did no al e CD68 s aining in all u ns (Fig. 1e) (p> 0.05, pos hoc es ). The 0.02 % AgNPs had no in luence on CD68 s aining in he a o emen- ioned cells in all u ns (images no shown) (p> 0.05, pos hoc es ). In he cochleae exposed o dH 2 O, he s ial in e media e cells, s ial basal cells, spi al ligamen ib ocy es, spi al ganglion cells, and ou e hai cells, pilla cells, and Dei e s’ cells o Co i’s o gan showed in ensi e s aining o CD44 (Addi ional ile 1: Figu e S1B, S1D, and S1F), while he inne hai cells demons a ed mild s aining o CD44 (Addi ional ile 1: Figu e S1F). 0.4 % AgNPs had no in luence on he s aining in he a o emen ioned cells in all u ns (Addi ional ile 1: Figu e S1A, S1C, and S1E) (p> 0.05, independen sample es ). In he cochleae exposed o dH 2 O, he s ial basal cells, spi al ligamen ib ocy es (mainly ype II), spi al ganglion cells, and inne hai cells and pilla cells o Co i’s o gan showed in ensi e s aining o TLR2 (Addi ional ile 2: Figu e S2B, S2D, and S2F), while he ou e hai cells and Dei e s’cells displayed ex emely weak s aining o TLR2 (Addi ional ile 2: Figu e S2F). The s ial basal cells and spi al ligamen ib ocy es demons a ed mild s aining o TLR4 (Fig. 2d), while he spi al ganglion cells and hai cells, pilla cells, and Dei e s’cells o Co i’s o gan exhib- i ed ex emely weak s aining o TLR4 (Fig. 2 , h). In he cochleae exposed o 0.4 % AgNPs, he ou e hai cells and Dei e s’cells o Co i’s o gan showed mo e in ensi e s aining o TLR2 (Addi ional ile 2: Figu e S2E). How- e e , he s ial basal cells, spi al ligamen ib ocy es, and spi al ganglion cells did no show any changes in he s aining o TLR2 in all u ns (Addi ional ile 2: Figu e S2A and S2C) (p> 0.05, one-way ANOVA) no in he inne hai cells and pilla cells (Addi ional ile 2: Figu e S2E). The s ial basal cells (p< 0.05 in he i s and second u ns and p< 0.01 in he hi d u n, one-way ANOVA) and spi al ligamen ib ocy es (Fig. 2a–c) (p<0.05 in he i s , second, and hi d u ns, one-way ANOVA) demon- s a ed mo e in ensi e s aining o TLR4 ha was inde- penden o he cochlea u n (p> 0.05, one-way ANOVA). The inne hai cells, pilla cells, and Dei e s’cells displayed mo e in ensi e s aining o TLR4, bu he ou e hai cells did no (Fig. 2g). Howe e , he spi al ganglion cells did no show any changes (Fig. 2e). The 0.02 % AgNPs had no in luence on he s aining o TLR2 and TLR4 in he a o e- men ioned cells in all u ns (images no shown) (p>0.05, one-way ANOVA). Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 4 o 13 In he cochleae exposed o dH 2 O, he Dei e s’cells o Co i’s o gan showed in ensi e s aining o MCP1, while he inne hai cells and inne pilla cells exhib- i ed mode a e s aining o MCP1 (Fig. 3h). The s ial in e media e cells, s ial basal cells, spi al ganglion cells, ou e hai cells, and ou e pilla cells demon- s a ed mild s aining o MCP1 (Fig. 3d, , h), while he spi al ligamen ib ocy es displayed ex emely weak s aining o MCP1 (Fig. 3d). Unexpec edly, he s ial basal cells, spi al ligamen ib ocy es, spi al ganglion cells, and he hai cells, pilla cells, and Dei e s’cells o Co i’so gan showed in ensi e s aining o MCP2 (Addi ional ile 3: Figu e S3B, S3D, and S3F). In he cochleae exposed o 0.4 % (Fig. 3a) and 0.02 % AgNPs (image no shown), he s ial in e media e cells, capilla y endo helial cells, and s ial basal cells (p< 0.01, one-way ANOVA) in he i s Fig. 1 CD68 + cells in he a cochlea 7 days pos -in a ympanic injec ion o 0.4 % AgNPs shown by immuno luo escence con ocal mic oscopy o immunohis ochemis y. In he cochleae exposed o dH 2 O, he inne hai cells (IHCs) and pilla cells (PCs) o Co i’s o gan (CO) showed mode a e s aining, while he ou e hai cells (OHCs) and Dei e s’cells (DCs) demons a ed ex emely weak s aining (h). The s ial basal cells (SBCs), spi al ligamen ib ocy es (SLFs), and spi al ganglion cells (SGCs) exhibi ed mild s aining (d, ). In he cochleae exposed o 0.4 % AgNPs, heSBCsandSLFs(mainly ypeIII)in he i s u n(a)and heIHCsandPCso CO(g) displayed mo e in ensi e s aining. Spa se ami ied cells (c) and mononuclea cells (i) wi h CD68 s aining we e iden i ied in he spi al ligamen and he modiolus, espec i ely. Howe e , he SBCs and SLFs in he second and hi d u ns (b,c), SGCs (e), capilla y endo helial cells (CaECs) (e), OHCs, and DCs (g) did no show any changes. Compa isons o s aining in ensi y a e shown in jand k.Scaleba =30μmina– ,20μming,h, and he magni ied image in i, and 80 μmini Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 5 o 13 u n demons a ed mo e in ensi e s aining o MCP1. Howe e , he spi al ligamen ib ocy es (mainly ype III) in he cochleae exposed o 0.4 % AgNPs (Fig. 3a–c) (p<0.01 in he i s and hi d u ns and p< 0.05 in he second u n, one-way ANOVA) showed mo e in ensi e s aining o MCP1 ha was independen o he cochlea u n (p> 0.05, one-way ANOVA). In addi ion, 0.4 % AgNPs inc eased MCP1 s aining in he inne pilla cells and Dei e s’cells o Co i’so gan(Fig.3g).Howe e , he spi al ganglion cells did no show any changes (Fig. 3e) (p> 0.05, one-way ANOVA). Nei he 0.4 % no 0.02 % AgNPs a ec ed he s aining o MCP2 in he a o emen- ioned cells in all u ns (images no shown) (p> 0.05, one-way ANOVA). Fig. 2 TLR4 + cells in he a cochlea 7 days pos -in a ympanic injec ion o 0.4 % AgNPs shown by immuno luo escence con ocal mic oscopy o immunohis ochemis y. In he cochleae exposed o dH 2 O, he s ial basal cells (SBCs) and spi al ligamen ib ocy es (SLFs) showed mild s aining (d), while he spi al ganglion cells (SGCs), hai cells (HCs), pilla cells (PCs), and Dei e s’cells (DCs) o Co i’s o gan (CO) demons a ed ex emely weak s aining ( ,h). In he cochleae exposed o 0.4 % AgNPs, he SBCs and SLFs exhibi ed mo e in ensi e s aining ha was independen o he cochlea u n (a–c). In CO, he inne hai cells (IHCs), PCs, and DCs displayed mo e in ensi e s aining, bu he ou e hai cells (OHCs) did no (g). Howe e , he SGCs did no show any changes (e). Compa isons o s aining in ensi y a e shown in iand j. Scale ba = 30 μm Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 6 o 13 AgNPs had no E ec on he Exp essions o Tigh Junc ion-Associa ed P o eins Including Rac1, Myosin Ligh Chain, VCAM1, and MAPK Signaling P o eins In he cochleae exposed o dH 2 O, he s ial in e media e cells, s ial basal cells, spi al ganglion cells, and hai cells, pilla cells, and Dei e s’cells o Co i’s o gan showed in ensi e s aining o Rac1 (Addi ional ile 4: Figu e S4B, S4D, and S4F), while he spi al ligamen ib ocy es (mainly ype II) demons a ed mode a e s ain- ing o Rac1 (Addi ional ile 4: Figu e S4B). The spi al ganglion cells and inne pilla cells o Co i’s o gan exhib- i ed mode a e s aining o myosin ligh chain (Addi ional Fig. 3 MCP1 + cells in he a cochlea 7 days pos -in a ympanic injec ion o 0.4 % AgNPs shown by immuno luo escence con ocal mic oscopy o immunohis ochemis y. In he cochleae exposed o dH 2 O, he Dei e s’cells (DCs) o Co i’s o gan (CO) showed in ensi e s aining, while he inne hai cells (IHCs) and inne pilla cells (IPCs) exhibi ed mode a e s aining (h). The s ial in e media e cells (SIMCs), s ial basal cells (SBCs), spi al ganglion cells (SGCs), and ou e hai cells (OHCs) and ou e pilla cells (OPCs) o CO demons a ed mild s aining, while he spi al ligamen ib ocy es (SLFs) displayed ex emely weak s aining (d, ,h). In he cochleae exposed o 0.4 % AgNPs, he SLFs showed mo e in ensi e s aining ha was independen o he cochlea u n, while he SIMCs, SBCs, and capilla y endo helial cells (CaECs) demons a ed mo e in ensi e s aining in he i s u n (a–c). In CO, he IPCs and DCs exhibi ed mo e in ensi e s aining, bu he hai cells (HCs) and OPCs did no (g). Howe e , he SGCs did no show any changes (e). Compa isons o s aining in ensi y a e shown in iand j.Scaleba =30μm Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 7 o 13 ile 5: Figu e S5D and S5F), while he hai cells, ou e pilla cells, and Dei e s’cells displayed mild s aining o myosin ligh chain (Addi ional ile 5: Figu e S5F). The s ial basal cells and spi al ligamen ib ocy es showed ex emely weak s aining o myosin ligh chain (Addi ional ile 5: Figu e S5B). The s ial basal cells, spi al ligamen ib ocy es, spi al ganglion cells, and hai cells, pilla cells, and Dei e s’cells o Co i’s o gan showed ex emely weak s aining o VCAM1 (Addi ional ile 6: Figu e S6B, S6D, and S6F), JNK (Addi ional ile 7: Figu e S8B, S8F, and S8J), and p38 (Addi ional ile 7: Figu e S8D, S8H, and S8L). Howe e , he s ial in e media e cells, s ial basal cells, spi al ligamen ib ocy es, spi al ganglion cells, and hai cells, pilla cells, and Dei e s’cells o Co i’s o gan showed in ensi e s aining o E k1/2 (Addi ional ile 8: Figu e S7B, S7D, and S7F). 0.4 % AgNPs had no in luence on he s aining o Rac1 (Addi ional ile 4: Figu e S4A, S4C, and S4E) (p> 0.05, independen sample es ), myosin ligh chain (Addi ional ile 5: Figu e S5A, S5C, and S5E), VCAM1 (Addi ional ile 6: Figu e S6A, S6C, and S6E), E k1/2 (Addi ional ile 8: Figu e S7A, S7C, and S7E) (p> 0.05, independen sample es ), JNK (Addi ional ile 7: Figu e S8A, S8E, and S8I), and p38 (Addi ional ile 7: Figu e S8C, S8G, and S8K) in he a o emen ioned cells in all u ns. AgNPs Up egula ed he Exp essions o Ubiqui in-Edi ing P o eins A20 and RNF11 Wi hou A ec ing he Exp es- sions o In lamma o y Cy okines In he cochleae exposed o dH 2 O, he spi al ganglion cells, inne hai cells, and inne pilla cells o Co i’s o gan showed mild s aining o TNF-α(Addi ional ile 9: Figu e S9H and S9N), while he s ial basal cells, spi al ligamen ib ocy es, ou e pilla cells, ou e hai cells, and Dei e s’cells demons a ed ex emely weak s aining o TNF-α(Addi ional ile 9: Figu e S9B and S9N). The s ial in e media e cells, s ial basal cells, and spi al gan- glion cells exhibi ed mild s aining o TNFR1 (Addi ional ile 9: Figu e S9D and S9J), while he spi al ligamen ib o- cy es, hai cells, pilla cells, and Dei e s’cells displayed ex emely weak s aining o TNFR1 (Addi ional ile 9: Figu e S9D and S9P). The s ial in e media e cells and s ial basal cells showed mild s aining o TNFR2 (Addi ional ile 9: Figu e S9F), while he spi al liga- men ib ocy es, spi al ganglion cells, hai cells, pilla cells, and Dei e s’cells demons a ed ex emely weak s aining o TNFR2 (Addi ional ile 9: Figu e S9F, S9L, and S9R). The s ial basal cells, spi al ganglion cells, and pilla cells o Co i’s o gan exhibi ed in ensi e s aining o IL-1β, while he spi al ligamen ib ocy es (mainly ype II) and inne hai cells displayed mild s ain- ing o IL-1β(Addi ional ile 10: Figu e S10B, S10D, and S10F). The ou e hai cells and Dei e s’cells showed ex emely weak s aining o IL-1β(Addi ional ile 10: Figu e S10F). 0.4 % AgNPs had no in luence on he s ain- ing o TNF-α(Addi ional ile 9: Figu e S9A, S9G, and S9M), TNFR1 (Addi ional ile 9: Figu e S9C, S9I, and S9O), TNFR2 (Addi ional ile 9: Figu e S9E, S9K, and S9Q), and IL-1β(Addi ional ile 10: Figu e S10A, S10C, and S10E) (p>0.05, independen sample es ) in he a o emen ioned cells in all u ns. In he cochleae exposed o dH 2 O, he spi al ganglion cells showed in ensi e s aining o IL-10 (Addi ional ile 11: Figu e S11F), while he pilla cells o Co i’s o gan demons a ed mild s aining o IL-10 (Addi ional ile 11: Figu e S11J). The s ial basal cells, spi al ligamen ib o- cy es, hai cells, and Dei e s’cells exhibi ed ex emely weak s aining o IL-10 (Addi ional ile 11: Figu e S11B and S11J). The spi al ganglion cells and pilla cells o Co i’s o gan displayed in ensi e s aining o TGF-β(Addi ional ile 11: Figu e S11H and S11L), while he s ial basal cells, spi al ligamen ib ocy es, and inne hai cells demons a ed mild s aining o TGF-β(Addi ional ile 11: Figu e S11D and S11L). The ou e hai cells and Dei e s’cells showed ex emely weak s aining o TGF-β(Addi ional ile 11: Figu e S11L). 0.4 % AgNPs had no in luence on he s aining o IL-10 (Addi ional ile 11: Figu e S11A, S11E, and S11I) (p> 0.05, independen sample es ) and TGF-β(Addi ional ile 11: Figu e S11C, S11G, and S11K) (p> 0.05, independen sample es ) in he a o e- men ioned cells in all u ns. In he cochleae exposed o dH 2 O, he spi al ganglion cells, inne hai cells, pilla cells, and Dei e s’cells o Co i’s o gan showed in ensi e s aining o A20 (Fig. 4j, n), while he s ial basal cells, spi al ligamen ib ocy es, and ou e hai cells demons a ed mild s aining o A20 (Fig. 4d, n). The s ial basal cells, spi al ganglion cells, and inne pilla cells o Co i’s o gan exhibi ed in ensi e s ain- ing o RNF11, while he spi al ligamen ib ocy es, hai cells, and ou e pilla cells displayed mild s aining o RNF11 (Fig. 4h, l, p). The Dei e s’cells showed ex emely weak s aining o RNF11 (Fig. 4p). In he cochlea la e al wall, 0.4 % AgNPs enhanced he s aining o A20 (p<0.05 in he i s and second u ns and p> 0.05 in he hi d u n a he s ial basal cells, p< 0.05 in he i s and hi d u ns and p< 0.01 in he second u n a he spi al ligamen ib ocy es, one-way ANOVA) and RNF11 (p>0.05 in he i s and hi d u ns and p< 0.05 in he second u n a he s ial basal cells, p< 0.01 in he i s u n and p<0.05 in he second and hi d u ns a he spi al ligamen ib o- cy es, one-way ANOVA) ema kably in he s ial basal cells and spi al ligamen ib ocy es ha we e independen o he cochlea u n (Fig. 4a–c, e–g) (p> 0.05, one-way ANOVA). In Co i’s o gan, 0.4 % AgNPs inc eased A20 s aining in he ou e hai cells and Dei e s’cells (Fig. 4m) and RNF11 s aining in he ou e pilla cells and Dei e s’ cells (Fig. 4o). In he spi al ganglion cells and capilla y endo helial cells, 0.4 % AgNPs did no al e he s aining o Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 8 o 13 Fig. 4 (See legend on nex page.) Feng e al. Nanoscale Resea ch Le e s (2016) 11:240 Page 9 o 13