NIS expression in thyroid tumors, relation with prognosis clinicopathological and molecular features

7:1 78–90C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s RESEARCH NIS exp ession in hy oid umo s, ela ion wi h p ognosis clinicopa hological and molecula ea u es Ca a ina Ta a es1,2,3, Ma iaJoão Coelho1,2,4, Ca a ina Eloy1,2,3, Miguel Melo1,2,5,6, Ad ianaGaspa daRocha1,2,7, Ana Pes ana1,2,3, Rui Ba is a1,2,3, LucianaBueno Fe ei a1,2,3, Elisabe e Rios1,2,3,8,9, Samia Selmi-Ruby10, B unoCa adas1,2,4, Luísa Pe ei a1,2,3, ManuelSob inho Simões1,2,3,8,9 and Paula Soa es1,2,3,8 1Ins i u o de In es igação e Ino ação em Saúde (i3S), Po o, Po ugal 2Ins i u e o Molecula Pa hology and Immunology o he Uni e si y o Po o (IPATIMUP), Po o, Po ugal 3Medical Facul y o he Uni e si y o Po o, Po o, Po ugal 4Ins i u e o Biomedical Sciences o Abel Salaza (ICBAS), Po o, Po ugal 5Depa men o Endoc inology, Diabe es and Me abolism, Uni e si y and Hospi al Cen e o Coimb a, Coimb a, Po ugal 6Medical Facul y, Uni e si y o Coimb a, Coimb a, Po ugal 7Public Heal h Uni , ACeS Baixo Mondego, Coimb a, Po ugal 8Depa men o Pa hology, Medical Facul y o he Uni e si y o Po o, Po o, Po ugal 9Depa men o Pa hology, Hospi al de S. João, Po o, Po ugal 10Inse m UMR-S1052, CNRS UMR5286, Cen e de Reche che en Cancé ologie de Lyon, Lyon, F ance Co espondence should be add essed o P Soa es: [email p o ec ed] Abs ac Thy oid cance he apy is based on su ge y ollowed by adioiodine ea men . Theinco po a ion o adioiodine by cance cells is media ed by sodium iodide sympo e (NIS) (codi ied by he SLC5A5 gene), ha is unc ional only when a ge ed o he cell memb ane. We aimed o e alua e i NIS exp ession in hy oid p ima y umo s would be help ul in p edic ing umo beha io , esponse o he apy and p ognosis. NIS exp ession was add essed by qPCR and immunohis ochemis y. In o de o alida e ou da a, we also s udied SLC5A5 exp ession on 378 p ima y papilla y hy oid ca cinomas om The Cance Genome A las (TCGA) da abase. In ou se ies, SLC5A5 exp ession was lowe in ca cinomas wi h ascula in asion and wi h ex a hy oidal ex ension and in hose ha bo ing BRAFV600E mu a ion. Analysis o SLC5A5 exp ession om TCGA da abase con i med ou esul s. Fu he mo e, i showed ha la ge umo s, wi h loco egional ecu ences and/ o dis an me as ases o ha bo ing RAS, BRAF and/o TERT p omo e (TERTp) mu a ions p esen ed signi ican ly less SLC5A5 exp ession. Rega ding immunohis ochemis y, 12/211 o he cases demons a ed NIS in he memb ane o umo cells, hose cases showed a iable ou comes conce ning he apy success, p ognosis and all bu one we e wild ype o BRAF, NRAS and TERTp mu a ions. SLC5A5 mRNA lowe exp ession is associa ed wi h ea u es o agg essi eness and wi h key gene ic al e a ions in ol ing BRAF, RAS and TERTp. Mu a ions in hese genes seem o dec ease p o ein exp ession and i s a ge ing o he cell memb ane. SLC5A5 mRNA exp ession is mo e in o ma i e han NIS immunohis ochemical exp ession ega ding umo agg essi eness and p ognos ic ea u es. 10.1530/EC-17-0302 Key Wo ds hy oid cance NIS SLC5A5 immunohis ochemis y Endoc ine Connec ions (2018) 7, 78–90 ID: 17-0302 71 h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 797:1 In oduc ion Sodium iodide sympo e is a ansmemb ane glycop o ein (codi ied by he SLC5A5 gene) exp essed almos exclusi ely in he basola e al plasma memb ane o hy oid ollicula cells. I plays a cen al ole in hy oid me abolism, media ing he ac i e anspo o iodine om he bloods eam in o he ollicula cells, he i s s ep o hy oid ho mones’ syn hesis. NIS plays an essen ial ole in he ea men o di e en ia ed hy oid ca cinomas (DTC), which usually main ain NIS exp ession, allowing he ecogni ion and he ea men o ecu ences and me as ases wi h adioac i e iodine (RAI) (1). None heless, a signi ican subg oup o DTC pa ien s wi h ad anced disease lose NIS exp ession and become e ac o y o 131I; some o hese pa ien s die wi hin 3–5 yea s (2). Fu he mo e, a s udy pe o med by Yild i im-Poy az and cowo ke s (3) demons a ed ha NIS exp ession in non umo al hy oid issues associa es wi h highe a es o delayed s uc u al esponse. NIS exp ession has been widely s udied in no mal hy oid and umo issues, on one hand o e i y i i s down egula ion could be he molecula cause o he dec ease o RAI up ake and on he o he hand o unde s and he impai ing mechanisms o NIS exp ession and unc ion. Howe e , no clea answe eme ged om he esul s ob ained in he p e ious s udies. Despi e he cen al ole o NIS in diagnosis, ea men and ollow-up o hy oid cance pa ien s, eliable me hods o asce aining NIS exp ession and unc ionali y in clinical samples a e no a ailable. In he majo i y o he s udies, SLC5A5 mRNA le els a e lowe in hy oid ca cinomas han in adenomas (4) and no mal adjacen hy oid (5, 6, 7); u he mo e, SLC5A5 exp ession p esen s some limi a ions in p edic ing NIS exp ession and unc ionali y: whe eas a nega i e o low mRNA le el may lead o educed p o ein exp ession, a posi i e o high mRNA exp ession does no always co espond o highe p o ein le els o highe unc ionali y(7, 8). These obse a ions sugges ha in hy oid ca cinomas, besides ansc ip ion egula ion, NIS exp ession appea s o be modula ed by pos - ansc ip ional e en s. The e o e, s udies o NIS exp ession by immunohis ochemis y (IHC) (1, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25), may be, heo e ically, mo e in o ma i e since hey ‘g ab’ NIS a s ep o wa d in i s biological p ocessing and allow he e alua ion o he localiza ion o NIS in he basola e al plasma memb ane o ollicula cells ( he unc ional anspo e ). Acco ding o he published da a, NIS exp ession (e alua ed by IHC) a ies in di e en hy oid issues. In no mal hy oid, i is low and e y he e ogeneous; only a ew ollicula cells wi hin some ollicles exp ess NIS in he basola e al plasma memb ane (10, 14, 17, 21, 26), sugges ing ha , NIS exp ession is igh ly egula ed in hy oid gland. In ca cinomas, when NIS is p esen , i is usually exp essed in a highe numbe o cells han in no mal issue and he exp ession is mainly in acy oplasmic, poo ly a ge ed o he basola e al plasma memb ane (1, 11, 12, 13, 14, 17, 21, 22, 23). The inc eased in acy oplasmic NIS s aining in hy oid umo s compa ed o no mal issue has been poin ed ou as a eason o he dec eased RAI up ake in umo s, e lec ing a mislocaliza ion o NIS om he basola e al memb ane, which would impai i s ac i i y (17). This assump ion has been ques ioned, because he eal signi icance o in acy oplasmic NIS de ec ed by immunos aining emains uncla i ied (21). The molecula mechanisms esponsible o he down egula ion and/o no a ge ing o he basola e al memb ane o NIS in hy oid umo s emain poo ly unde s ood, bu some s udies demons a ed ha bo h mRNA and p o ein a e di e en ially exp essed acco ding o he gene ic backg ound o he umo . In ac , papilla y hy oid ca cinomas (PTCs) ha bo ing he BRAFV600E mu a ion p esen lowe SLC5A5 mRNA and NIS p o ein exp ession as well as less a ge ing o he basola e al memb ane compa ed o PTCs BRAFWT (19, 24, 27). None heless, a ecen s udy epo ed di e en associa ion be ween NIS memb ane exp ession and BRAFV600E mu a ion in a se ies o 96 cPTCs, demons a ing ha he ones ha bo ing BRAFV600E mu a ion exp essed mo e o en NIS in he cell memb ane o umo cells compa ed o BRAFWT cPTCs (28). Less is known abou he impac o mu a ions in o he genes (i.e. RAS and TERTp) on SLC5A5 and NIS exp ession/ a ge ing o he basola e al memb ane. NIS being he cen al molecule o DTC ea men , i is logical o s udy i i s exp ession in he p ima y umo would be help ul in p edic ing he apy esponse as well as umo beha io and p ognosis. Some s udies ied o unde s and i NIS immunohis ochemical exp ession in hy oid p ima y umo s would be help ul in p edic ing 131I up ake in ecu ences and dis an me as ases. Al hough posi i e NIS immunos aining in p ima y umo s seemed o be p edic i e o posi i e ecu ences and me as ases on 131I scans, o he s udies did no dis inguish whe he NIS was exp essed in he cell basola e al memb ane, and This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 807:1 nega i e NIS s aining did no p edic 131I scan-nega i e me as ases (13, 15, 18). E en when only NIS memb ane s aining was conside ed, a nega i e NIS s aining in he p ima y umo was s ill no p edic i e o a nega i e 131scan o subsequen ecu ences (29). To he bes o ou knowledge, he e is only one s udy ha add essed possible associa ions be ween NIS exp ession, e alua ed by immunohis ochemis y (IHC), and clinicopa hological ea u es and p ognosis in a la ge se ies o hy oid p ima y umo s (1), epo ing a signi ican ly lowe NIS exp ession in olde pa ien s (≥45 yea s) and also ha NIS exp ession in he p ima y umo was no use ul as a p ognos icma ke . So, in ou opinion, mo e e ospec i e s udies in la ge se ies o p ima y umo s a e s ill necessa y, o unde s and he ole o NIS exp ession in he apy esponse, umo beha io and p ognosis, and also i o he ac o s besides BRAFV600E mu a ion can con ibu e o NIS down egula ion and/o misdi ec ing o he basola e al memb ane. Fu he mo e, i is also impo an o unde s and he ad an ages and limi a ions o he analysis o SLC5A5 and NIS exp ession and e alua e wha is he be e /mo e in o ma i e me hod o s udy NIS exp ession. Ha ing his in mind, we add essed SLC5A5 exp ession by qPCR and NIS exp ession by IHC analysis, in a la ge se ies o p ima y hy oid ca cinomas and looked o possible associa ions wi h some clinicopa hological and molecula ea u es, as well as o he esponse o RAI he apy and ou come. In o de o alida e ou esul s o SLC5A5 mRNA exp ession associa ions’ wi h clinicopa hological and molecula ea u es and also o ge new e idences, we used he da a a ailable abou SLC5A5 in TCGA Resea ch Ne wo k ha comple ed an in eg a ed genomic analysis o 496 PTCs using NGS and o he pan- genomic echnologies, oge he wi h de ailed pa hologic and clinical da a (30). Ma e ials and me hods Pa ien samples Ou se ies was composed by 255 hy oid samples om 229 pa ien s. Cases we e collec ed om he iles o he Ins i u e o Molecula Pa hology and Immunology o he Uni e si y o Po o (IPATIMUP, Po o, Po ugal), co esponding o pa ien s wi h hy oid umo s (n = 229) ope a ed and ollowed in wo uni e si y hospi als. Samples om no mal hy oid (n = 25) and G a es’ disease (n = 1) we e ob ained om he con ala e al lobe o he su gical specimens. Ca cinomas se ies was composed by 193 PTCs (123 cases o classical PTC (cPTC), 47 cases o ollicula a ian o PTC ( PTC) and 23 cases o o he PTC a ian s), 23 ollicula hy oid ca cinomas (FTC) and 13 poo ly di e en ia ed hy oid ca cinomas (PDTC). In 166 o he cases, he e was only o malin- ixed pa a in- embedded (FFPE) ep esen a i e issue; in 45 cases, he e we e FFPE samples and co esponden ozen issue ( he umo s we e di ided a he ime o su ge y) and in 18 cases, he e was only ozen issue a ailable. F ozen ma e ial was collec ed a he ime o su ge y and conse ed a −80°C. The his ology o all umo samples was e iewed by h ee pa hologis s (CE, ER, MSS) acco ding o he c i e ia o he Wo ld Heal h O ganiza ion (31). Clinicopa hological and molecula da a o he 229 pa ien s wi h ca cinoma a e summa ized in Supplemen a y Table1 (see sec ion on supplemen a y da a gi en a he end o his a icle). In 141 cases, ollow-up da a we e a ailable. The numbe o 131I ea men s a ied om 1 o 5 ea men s (mean 1.9), and he cumula i e o al dose o RAI was be ween 30 and 1146 mCi (mean 251 mCi). This wo k was app o ed by he E hic Commi ee o Heal h (CES) o he Hospi al Cen e o São João (CHSJ)/Facul y o Medicine o he Uni e si y o Po o (FMUP) (CES 137 284-13) and by he E hic Commi ee o he Facul y o Medicine o he Uni e si y o Coimb a (nº 1309). All he p ocedu es desc ibed in his s udy we e in acco dance wi h na ional e hical s anda ds (Law nº 12/2005) and Helsinki decla a ion. Pa ien s signed an in o med consen o m. Pa ien ollow-up Pa ien s we e ea ed and ollowed in acco dance wi h he in e na ional p o ocols a ailable a he ime. Da a ega ding he numbe o adioiodine ea men s and cumula i e ac i i y we e e ie ed om hospi al eco ds. Pa ien s we e conside ed as being disease- ee a he end o ollow-up i hey had unde ec able s imula ed hy oglobulin (in he absence o hy oglobulin an ibodies), and no e idence o he disease on adiog aphic o adionuclide imaging. Follow-up ime (in yea s) a ied om 0.3 up o 38.9yea s, wi h a mean alue o 8.0 ± 6.6yea s. Fo s a is ical analysis, we de ined he ca ego y ‘addi ional ea men s’, in which we included o he ea men modali ies in addi ion o adioiodine, including ex a su ge y, ex e nal beam i adia ion and ea men wi h y osine kinase inhibi o s. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 817:1 Da ase PTC in TCGA The e we e 378 umo cases o which he e was in o ma ion o he main d i e soma ic mu a ions (TERTp, BRAF and RAS), gende and SLC5A5 exp ession. O hese, we elimina ed 4 cases, o which he SLC5A5 exp ession was abo e he 99 pe cen ile, being ou lie s. A o al o 353 o he cases had in o ma ion abou umo size, 362 had in o ma ion o ex a hy oidal ex ension, 282 had in o ma ion o lymph node me as ases (a he ime o diagnosis) and all 374 had in o ma ion abou new umo e en (lymph node me as ases o local ecu ence (g ouped in loco egional ecu ence) and dis an me as ases). The SLC5A5 exp ession was in e ed om RNA-seq da a and quan i ica ion e lec s eads pe kilobase pe million mapped eads (RPKM). The e we e also 58 SLC5A5 exp ession measu es in adjacen issue o he PTC cases, and wo o hem we e no conside ed o u he analyses as alues we e abo e he 99 pe cen ile. DNA ex ac ion, PCR and Sange sequencing DNA ex ac ion om FFPE issues was pe o med om 10 μm sec ions a e ca e ul mic odissec ion. DNA ex ac ion was pe o med using Ul ap ep issue DNA ki (AHN Bio echnologie, No dhausen, Ge many) ollowing he manu ac u e ’s ins uc ions. The gene ic cha ac e iza ion o pa o he umo s ega ding BRAF, NRAS and TERT p omo e mu a ions (TERTp) had been epo ed p e iously; mu a ions we e sc eened as p e iously desc ibed (32, 33, 34). RNA ex ac ion and e e se ansc ip ion To al RNA was ex ac ed om umo s and om con ala e al no mal adjacen hy oid, om which ozen samples we e a ailable (n = 84), using a TRIzol comme cial ki (The mo Scien i ic/GIBCO) acco ding o he manu ac u e ’s p o ocol. RNA was quan i ied by spec opho ome y, and i s quali y was checked by analysis o 260/280 nm and 260/230 nm a ios. Fo cDNA p epa a ion, 1 μg o o al RNA was e e se- ansc ibed using he Re e Aid i s - s and cDNA syn hesis ki (The mo Scien i ic/Fe men as). Real- ime PCR Re e se ansc ip ion p oduc s we e ampli ied o he SLC5A5 gene and de ec ed by a p obe (IDT: In eg a ed DNA Technologies, Leu en, Belgium; no. HS.PT.56a.40789288), as p e iously desc ibed (35). Immunohis ochemis y Immunohis ochemis y was pe o med in no mal hy oid and in 211 ca cinomas. B ie ly, depa a inized and ehyd a ed sec ions we e subjec ed o hea -induced an igen e ie al in 10 mM sodium ci a e bu e (pH6.0). Endogenous pe oxidase ac i i y was blocked wi h 3% o hyd ogen pe oxide and nonspeci ic binding wi h La ge Volume Ul a V Block eagen (The mo Scien i ic/ Lab Vision). Sec ions we e hen incuba ed o e nigh a 4°C wi h an i-NIS an ibody (1:400) clone FP5A (The mo Scien i ic/Lab Vision) and in 24 ca cinomas wi h an i-NIS pAb 795 IgG (20 µg/mL) (kindly supplied by D Ruby) (36). Addi ionally, Ty amide Signal Ampli ica ion (TSA) Bio in Sys em (Pe kin-Elme ) was used o signal ampli ica ion in 44 ca cinomas, acco ding o manu ac u e ’s ins uc ions. The de ec ion was pe o med wi h a labeled, s ep a idin– bio in immunope oxidase de ec ion sys em (The mo Scien i ic/Lab Vision) ollowed by 3,3′-diaminobenzidine (Dako) and coun e s ained wi h hema oxylin. G a es’ disease sample was used as a posi i e con ol and he nega i e con ol consis ed in omission o he p ima yan ibody. Slides we e e alua ed by wo obse e s and we e analyzed acco ding o he pe cen age o umo -s ained cells, he in ensi y and he cellula localiza ion o he s aining. In o de o compa e ou esul s o he li e a u e, we conside ed cases wi h >5% o s ained umo cells ( ega dless o he cellula localiza ion) as posi i e. Ne e heless, all ou s a is ical analyses we e pe o med conside ing wo g oups; cases ha p esen ed memb ane s aining in umo cells and all he o he cases. Pho og aphs we e acqui ed using Nikon DS-L1 came a in 100× and 400× magni ica ions. S a is ical analysis S a is ical analysis was pe o med using 21.0 SPSS S a is ical Package (SPSS, 2003). Fishe ’s exac es and independen - samples - es we e pe o med o co ela e NIS and SLC5A5 mRNA exp ession wi h clinicopa hological and molecula ea u es. When pa ame ic es s we e no applicable, we used al e na i e es s, speci ically Mann– Whi ney (independen samples). Wilcoxon ( ela ed samples) was used o compa e SLC5A5 exp ession be ween umo samples and hei adjacen no mal coun e pa s. K uskal–Wallis es was used o co ela e SLC5A5 exp ession ( e ie ed om TCGA and da abase) wi h clinicopa hological and molecula ea u es. Values o P ˂ 0.05 we e conside ed s a is ically signi ican . This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 827:1 Resul s SLC5A5 mRNA exp ession SLC5A5 exp ession was signi ican ly lowe in ca cinomas han ha in no mal adjacen coun e pa s (Fig. 1). No signi ican di e ence was obse ed be ween he h ee di e en ca cinoma his o ypes (PTC, FTC and PDTC). Conside ing he analysis in DTC, SLC5A5 exp ession was signi ican ly lowe in males and in cases wi h ascula in asion (P = 0.003 and P = 0.03, espec i ely) (Table 1). SLC5A5 exp ession in no mal hy oid om males was no signi ican ly di e en om ha o emales (da a no shown). In addi ion, he e was a endency o lowe SLC5A5 le els in cases wi h ex a hy oidal ex ension (P = 0.06) and in PTCs ha bo ing BRAFV600E mu a ion (P = 0.07). When he s a is ical analysis was pe o med only in he PTC g oup, all he signi ican associa ions desc ibed in he DTC g oup we e main ained. SLC5A5 mRNA exp ession (TCGA da abase) The SLC5A5 exp ession was a ound 200 imes highe in no mal issue han in umo issue in bo h gende s, bu no di e ences in umo and in adjacen issue be ween gende s we e ound (Fig. 2A and B). SLC5A5 exp ession was signi ican ly highe in smalle umo s ≤2 cm (median = 5.85) compa ed o hose wi h >2 cm (median = 2.51) (P = 0.028; Fig.2C). The e was no s a is ical di e ence in SLC5A5 exp ession in p ima y umo s wi h (median = 3.0) o wi hou (median = 5.4) lymph node me as ases a he ime o diagnosis (P = 0.253) (Fig.2D). The SLC5A5 exp ession was educed wi h he le el o he ex a hy oidal ex ension (median alues: 5.4 o ‘none’; 2.8 o ‘minimal (T3)’ and 0.9 o ‘mode a e/ ad anced (T4a + b)’), eaching s a is ical signi icance o compa isons be ween ‘none’ s he ‘mode a e/ad anced (T4a + b)’ class and ‘minimal (T3)’ s ‘mode a e/ad anced (T4a + b)’ (P = 0.018 and P = 0.039, espec i ely Fig. 2E). We also obse ed a s a is ical signi ican dec ease ( om a median o 3.8 o 0.8; P = 0.002) o he SLC5A5 exp ession in cases wi h new umo e en s (Fig.2F), lumping oge he 12 cases o dis an me as asis (6 lung; 1 lung + bone; 1 lymph node only; 1 lung + emu + neck + pleu a + li e ; 1 bone; 2 unknown) and 14 loco egional ecu ences (10 lymph node only; 2 le hy oid; 1 lymph node + so issue; 1 unknown). Finally, SLC5A5 exp ession was signi ican ly highe in he absence (median = 21.77) o he e alua ed mu a ions: RAS (P = 0.034), TERTp (P = 0.0072) and BRAFV600E (P = 3.1 × 10−8). The PTCs ha ha bo ed only TERTp, only BRAF o simul aneous TERTp and BRAF mu a ions displayed signi ican ly lowe exp ession o SLC5A5 han he WT umo s. The g oup wi h RAS mu a ions displayed he second highes exp ession alue (median = 7.50), eaching s a is ical signi icance when compa ed wi h he g oups including BRAF mu a ion only and BRAF + TERTp mu a ions (median = 2.27 in BRAF (P = 0.042); median = 1.89 in TERT+BRAF (P = 0.027))(Fig.2G). NIS exp ession In no mal hy oid issues, NIS immunohis ochemical exp ession was mainly localized in he basola e al plasma memb ane o ollicula cells. NIS posi i i y was de ec ed Figu e1 SLC5A5 mRNA exp ession in hy oid ca cinomas and pai ed no mal adjacen coun e pa s. Table 1 Associa ions be ween SLC5A5 mRNA exp ession wi h clinicopa hological and molecula ea u es in DTCs. SLC5A5 exp ession P alue Gende F (n = 47) 1.2 ± 2.2 M (n = 12) 0.2 ± 0.2 0.003 Age <45yea s (n = 30) 1.0 ± 1.5 ≥45yea s (n = 29) 1.1 ± 2.4 0.8 Tumo capsule P esen (n = 27) 1.1 ± 1.6 Absen (n = 30) 0.7 ± 1.6 0.4 Tumo capsule in asion Yes (n = 17) 0.9 ± 1.6 No (n = 11) 1.4 ± 1.4 0.4 Ex a hy oidal ex ension Yes (n = 17) 0.5 ± 1.1 No (n = 37) 1.4 ± 2.4 0.06 Lymphocy ic in il a ion P esen (n = 19) 0.9 ± 1.9 Absen (n = 37) 1.2 ± 2.2 0.7 Vascula in asion P esen (n = 28) 0.4 ± 0.8 Absen (n = 29) 1.5 ± 2.6 0.03 Lymph node me as ases P esen (n = 13) 0.5 ± 0.8 Absen (n = 18) 0.4 ± 0.7 0.8 BRAF* WT n = (27) 1.6 ± 2.7 V600E (n = 20) 0.5 ± 1.0 0.07 NRAS WT (n = 54) 1.0 ± 2.0 Mu (n = 6) 1.3 ± 1.7 0.7 *PTC only. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 837:1 in a ew oci o isola ed ollicles h oughou he issue and wi hin he posi i e ollicles mos o he cells we e posi i e. Posi i i y was mo e equen ly de ec ed in small ollicles composed by cuboidal and columna cells and a ely de ec ed in la ge ollicles limi ed by la ened cells (Fig.3A). In G a es’ disease, NIS was widely exp essed and p esen in he basola e al plasma memb ane o he g ea majo i y o ollicula cells (Fig. 3B). In ca cinomas, NIS s aining was obse ed in 71.6% o he cases (74.8% o cPTCs, 69.8% o PTCs, 80.9% o o he PTC a ian s, 55% o FTC and 67% o PDTC). I s loca ion was p edominan ly in he cy oplasm (124/211) (Fig.3C) and nucleus (15/211) and only 12/211 o he cases p esen ed NIS in he basola e al plasma memb ane o umo cells (Fig.3D). Since we obse ed a low pe cen age o ca cinomas wi h NIS s aining in he basola e al memb ane, Figu e2 SLC5A5 exp ession in p ima y PTCs (RPKM), da a e ie ed om TCGA da abase. Compa a i e analysis o SLC5A5 exp ession. (A)Be ween gende s in umo (TT) and no mal issue (NT); (B)be ween gende s only in umo issue (TT); (C)in umo s wi h ≤2 cm and >2 cm; (D) in cases wi h o wi hou lymph node me as ases a he ime o diagnosis; (E) in cases wi hou , wi h minimal (T3) and wi h mode a e/ad anced ex a hy oidal ex ension (T4a+b); (F)incases wi h and wi hou ecu ence and (G)be ween cases wi h di e en gene ic backg ounds (WT, RAS mu a ion, TERTp mu a ion, BRAF mu a ion, BRAF + TERTp mu a ion). The boxes ep esen he in e qua ile ange; he whiske s a e he 5% and 95% qua iles; he small open boxes a e he mean alues and he lines a e he median alues. Signi ican alues o he K uskal–Wallis es a e indica ed. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 847:1 wehypo hesized ha ou IHC app oach was no being sensi i e enough o de ec small amoun s o NIS. To cla i y his issue, we used wo s a egies: a TSA signal ampli ica ion me hod and he use o ano he NIS an ibody cha ac e ized by a di e en speci ici y compa ed o he comme cial an ibody (36). The TSA signal ampli ica ion me hod was applied in a subse o 44 ca cinomas wi h di e en s aining Figu e3 NIS immunoexp ession in di e en hy oid issues. (A) No mal hy oid; (B) G a es’ disease; (C)cy oplasma ic s aining in an oncocy ic PTC; (D)memb ane s aining in a PTC; (E and F) NIS immunoexp ession in a PTC wi hou and wi h TSA ampli ica ion signal, espec i ely; (GandH) NIS immunoexp ession in a FTC wi hou and wi h TSA ampli ica ion signal, espec i ely; (I)nega i e s aining in a cPTC and s ong memb ane s aining in he su ounding G a es’ disease. In F and H, no ice he loss o cy oplasma ic s aining a e he use o TSA ampli ica ion sys em. Ba 100 μm. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 857:1 pa e ns (16 wi h cy oplasmic s aining in he umo and memb ane s aining in adjacen hy oid; 3 wi h memb ane s aining in he umo ; 5 nega i e bo h in he in umo and he adjacen hy oid and, inally, 20 wi h only cy oplasmic s aining in umo and adjacen hy oid). When we compa ed he slides pe o med wi h and wi hou TSA signal ampli ica ion, we e i ied ha only he memb ane s aining emained and appea ed mo e in ense wi h he ampli ica ion me hod. In hese cases, he s aining in ol ed almos always he same oci o cells ha al eady p esen ed memb ane s aining (Fig. 3E, F, G and H) i.e. i did no s ain addi ional cells. The in a- cy oplasmic s aining anished bo h in cance and in no mal issues. Fu he mo e, we pe o med IHC using a homemade an ibody o human NIS, pAb 795 agains a pep ide co esponding o he C- e minal sequence o hNIS pAb 795 (36) in 24 ca cinomas (12 cPTC, 4 PTC, 2 mic o PTC, 2 all cell PTC, 2 FTC and 2 PDTC). The esul s we e simila o hose ob ained wi h clone FP5A (The mo Scien i ic/Lab Vision). Since some doub s emained abou he speci ici y o he cy oplasmic s aining, and because NIS is only ac i e when p esen in he basola e al memb ane o he cells, we pe o med s a is ical analysis di iding ou se ies in wo g oups: wi h and wi hou memb ane s aining. We did no ind any signi ican associa ion be ween NIS exp ession in he memb ane and age, umo size, umo capsule, mul i ocali y, lymphocy ic in il a ion, ascula in asion, lymph node me as ases, umo ma gins, dis an me as ases, s aging, BRAF, NRAS and TERTp s a us, addi ional ea men s, disease- ee s a us a one yea , disease- ee s a us a he end o ollow-up o disease-speci ic su i al in he DTC g oup. When we analyzed NIS exp ession be ween WT PTCs and hose ha bo ing any o he s udied mu a ions, we e i ied ha NIS-posi i e exp ession was signi ican ly mo e equen in WT PTCs (P = 0.01) (Table 2). The numbe o RAI he apies, as well as he cumula i e dose o RAI, did no di e signi ican ly be ween pa ien s wi h o wi hou NIS exp ession in he basola e al memb ane o p ima y umo ’s cells. The ho ough analysis o he ew cases wi h memb ane s aining (n = 12) e ealed ha all bu one ca cinoma we e wild ype o he s udied mu a ions (NRAS, BRAF o TERTp). These cases p esen ed a iable ou comes i.e. p esence o dis an me as ases, numbe o RAI he apies, cumula i e dose o RAI, he need o addi ional ea men s, disease- ee s a us and dea h (disease caused), ha a e appa en ly un ela ed wi h he p esence o NIS memb ane exp ession(Table3). Discussion In his wo k, we ied o cla i y he impac o NIS exp ession (mRNA and p o ein) on hy oid umo s’ agg essi eness and he apy success and, as a esul o he abo e, he pu a i e p ognos ic signi icance o SLC5A5 mRNA and NIS p o ein exp ession. Mo eo e , we also add essed he impac o he gene ic backg ound o he umo on SLC5A5 and NIS exp ession as well as i s a ge ing o he basola e al cell memb ane. We ound ha SLC5A5 exp ession was always lowe in umo s han in no mal adjacen coun e pa s as epo ed by o he g oups (6, 7, 37). We obse ed a signi ican ly lowe SLC5A5 exp ession in male gende pa ien s, and in cases wi h ascula in asion, as well as a endency o lowe SLC5A5 exp ession in cases wi h ex a hy oidal ex ension, bu no di e ences we e ound in cases wi h and wi hou lymph node me as ases (Table 1). When we compa e he esul s om ou se ies o hose om TCGA da a, we con i med ha umo s exp ess signi ican ly less SLC5A5 compa ed o no mal adjacen issue, ha SLC5A5 was no di e en ly exp essed in he p esence o absence o lymph node me as ases (a he ime o diagnosis) and a signi ican lowe SLC5A5 exp ession was ound in umo s wi h ex a hy oidal ex ension (mode a e/ad anced) compa ed o hose wi hou ex a hy oidal ex ension (Fig. 2D). Howe e , he di e en ial exp ession o SLC5A5 be ween gende s was no con i med (Fig. 2B). Un o una ely, in TCGA da abase, he e was no in o ma ion abou ascula in asion, so we could no alida e his esul in his la gese ies. The signi ican ly lowe SLC5A5 exp ession in cases p esen ing ascula in asion and ex a hy oidal ex ension sugges s ha a dec eased SLC5A5 exp ession may be associa ed o an agg essi e umo beha io and hus may help o cha ac e ize pa ien s a isk o poo he apy esponse. Fu he analysis o TCGA da a demons a ed ha SLC5A5 exp ession is signi ican ly lowe in cases ha had loco egional ecu ences and/o dis an me as ases (Fig.2E). Gi en he high p ognos ic impac o ecu ences and dis an me as ases (38), hese esul s sugges ha a Table 2 Associa ions be ween NIS exp ession and clinicopa hological and molecula ea u es in PTCs. NIS immunoexp ession P alue Nega i e Posi i e Gene ic backg ound n = 118 WT 45 (41.3%) 8 (88.9%) Mu a ed#64 (58.7%) 1 (11.1%) 0.011 #BRAF, NRAS o TERTp mu a ions. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access C Ta a es e al. SLC5A5/NIS exp ession in hy oid umo s 867:1 lowe exp ession o SLC5A5 in hy oid p ima y umo seems o be associa ed wi h ea u es o highe agg essi eness o he p ima y umo and also wi h a wo se p ognosis and wi h poo esponse o he apy. Two g oups epo ed ha SLC5A5 was signi ican ly less exp essed in DTCs la ge han 2 cm and PTCs la ge han 1 cm (in compa ison o ≤2 cm and <1 cm, espec i ely) (1, 39), TCGA esul s co obo a ed he li e a u e by showing ha la ge PTCs (>2 cm) exp essed signi ican ly less SLC5A5 compa ed o hose wi h ≤2 cm. Since la ge umo s a e associa ed wi h highe ecu ence a es and wo se p ognosis (40,41), he signi ican ly lowe SLC5A5 exp ession in umo s la ge han 2 cm may be conside ed as an addi ional ac linking lowe SLC5A5 mRNA exp ession wi h highe umo agg essi eness. Ne e heless, we mus in e p e his in o ma ion ca e ully: la ge umo s may p esen highe le els o nec osis/ ib osis and also addi ional non- ca cinoma ous issue as pa o he nodule, which may con ibu e o a lowe SLC5A5 exp ession. In ou se ies, we did no include mic oca cinomas, so he g oup o umo s wi h ≤2 cm was e y small, p ecluding any meaning ul analysis (da a no shown). P e ious s udies epo ed a lowe SLC5A5 exp ession in cases ha bo ing BRAFV600E, and he e is expe imen al e idence showing ha BRAFV600E can impai SLC5A5 exp ession (1, 19, 27, 39), ne e heless he impac o o he ele an mu a ions ound in hy oid umo s on SLC5A5 exp ession emained unknown. In ou se ies, SLC5A5 exp ession was lowe bu did no each s a is ical signi icance in he BRAFV600E PTC compa ed o ha o BRAF wild- ype g oup. The lack o signi icance in ou se ies may be due o di e ences in size and composi ion o he se ies, since he abo e men ioned s udies add essing SLC5A5 exp ession and BRAF V600E (1, 19, 27, 39) used la ge se ies o PTC. When we compa ed SLC5A5 exp ession ( e ie ed om TCGA da abase) be ween PTCs ha bo ing di e en mu a ions (BRAFV600E, TERTp and RAS) and WT PTCs, we obse ed ha independen ly o he mu a ion, SLC5A5 exp ession was always signi ican ly lowe compa ed o WT PTCs. Mo eo e , we also obse ed ha RAS mu a ion was he one wi h lowe impac on SLC5A5 exp ession. PTCs wi h RAS mu a ion displayed signi ican ly highe le els o SLC5A5 compa ed o BRAFV600E and BRAF + TERTp- mu a ed PTCs. In ac , i has been p e iously epo ed ha a dis inc p o ile o exp ession o genes in ol ed in hy oid ho mone biosyn hesis (being SLC5A5 one o hese genes) be ween BRAFV600E and RAS-d i en PTCs, wi h RAS-like PTCs ha ing ela i ely high hy oid di e en ia ion sco e (30). Table 3 Clinicopa hological and molecula da a o cases p esen ing NIS memb ane s aining. Diagnosis BRAF NRAS TERTp Lymph node me as ases Dis an me as ases Numbe o 131I he apies Cumula i e dose (mCi) Adi ional ea men s One yea DFS* DFS*,# Dea hs Case 1 cPTC WT WT WT No Bone 3 457.5 No No No No Case 2 cPTC WT WT WT No No 1 63 No Yes Yes No Case 3 cPTC WT WT WT Yes No 3 459 2 su ge ies No No No Case 4 PTC WT WT WT Yes No 1 37 No Yes Yes No Case 5 PTC WT WT WT No No 2 382 No No No No Case 6 PDTC WT WT WT No Lung + bone 5 798 2 su ge ies Yes No No Case 7 cPTC WT WT 124G>A Yes No 4 527 U/I No No No Case 8 sclPTC WT WT WT Yes No 3 400 U/I No No No Case 9 FTC WT WT WT No No 1 102 U/I U/I Yes No Case 10 cPTC WT WT WT Yes U/I U/I U/I U/I U/I U/I U/I Case 11 PTC WT WT WT U/I U/I U/I U/I U/I U/I U/I U/I Case 12 cPTC WT WT WT Yes U/I U/I U/I U/I U/I U/I U/I *DFS disease- ee su i al; #a he end o ollow-up. sclPTC, scle osing a ian o PTC; U/I, una ailable in o ma ion. This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. h ps://doi.o g/10.1530/EC-17-0302 h p://www.endoc ineconnec ions.o g © 2018 The au ho s Published by Bioscien i ica L d Downloaded om Bioscien i ica.com a 01/10/2020 04:08:30PM ia ee access