Detection and localization of viral infection in the pancreas of patients with type 1 diabetes using short fluorescently-labelled oligonucleotide probes

Onco a ge 1 www.impac jou nals.com/onco a ge www.impac jou nals.com/onco a ge / Onco a ge , Ad ance Publica ions 2017 De ec ion and localiza ion o i al in ec ion in he panc eas o pa ien s wi h ype 1 diabe es using sho luo escen ly-labelled oligonucleo ide p obes Niels Busse1,*, Fede ico Pa oni1,*, Sa ah J. Richa dson2, Ju a E. Laiho3, Maa i Oika inen3, Gun F isk4, Heikki Hyö y3,5, Eelco de Koning6,7, Noel G. Mo gan2 and Ka h in Maedle 1 1 Isle Biology Labo a o y, Uni e si y o B emen, Ge many 2 Isle Biology Exe e , Uni e si y o Exe e Medical School, UK 3 Depa men o Vi ology, School o Medicine, Uni e si y o Tampe e, Tampe e, Finland 4 Depa men o Immunology, Gene ics and Pa hology, Uppsala Uni e si y, Sweden 5 Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , Tampe e, Finland 6 Depa men o In e nal Medicine, Leiden Uni e si y Medical Cen e , Leiden, The Ne he lands 7 Hub ech Ins i u e/Uni e si y Medical Cen e U ech , U ech , The Ne he lands * Sha ed i s au ho s Co espondence o: Fede ico Pa oni, email: [email p o ec ed] Co espondence o: Ka h in Maedle , email: [email p o ec ed] Keywo ds: en e o i uses, ype 1 diabe es, panc eas, isle s, oligonucleo ide p obes, Pa hology Sec ion Recei ed: No embe 25, 2016 Accep ed: Janua y 19, 2017 Published: Janua y 29, 2017 ABSTRACT En e o i uses, speci ically o he Coxsackie B i us amily, ha e been implica ed in igge ing isle au oimmuni y and ype 1 diabe es, bu hei p esence in panc ea a o pa ien s wi h diabe es has no been ully con i med. To de ec he p esence o e y low copies o he i us genome in issue samples om T1D pa ien s, we designed a panel o luo escen ly labeled oligonucleo ide p obes, each o 17-22 nucleo ides in leng h wi h a unique sequence o speci ically bind o he en e o i al genome o he pico na i idae amily. Wi h hese p obes en e o i al RNA was de ec ed wi h high sensi i i y and speci ici y in in ec ed cells and issues, including in FFPE panc eas sec ions om pa ien s wi h T1D. De ec ion was no impeded by a ia ions in sample p ocessing and s o age he eby o e coming he po en ial limi a ions o agmen ed RNA. Co-s aining o small RNA p obes in pa allel wi h classical immuns aining enabled i us de ec ion in a cell-speci ic manne and mo e sensi i ely han by i al p o ein. INTRODUCTION Type 1 diabe es (T1D) is a ch onic mul i ace ed diso de ha esul s om selec i e au oimmune- media ed des uc ion o he insulin p oducing β-cells. En i onmen al ac o s [1], oge he wi h gene ic p edisposi ion [2], in e ac coope a i ely o ini ia e ch onic isle au oimmuni y [3]. Vi uses ha e been p oposed as possible ini ia o s o isle au oimmuni y and we e i s implica ed as long ago as he nine een h cen u y al hough i was no un il much la e ha a clea associa ion was es ablished be ween mumps and diabe es [4-6]. Imp o emen s in molecula biology subsequen ly b oadened he panel o i uses which a e implica ed in causing diabe es [7, 8] and he weigh o e idence now sugges s ha coxsackie i uses [9] play a ole. In suppo o his, a clea co ela ion be ween en e o i us in ec ion and he onse o T1D was e ealed in associa ion s udies [10] and ia a comp ehensi e me a- analysis [11]. Coxsackie i uses belong o he Pico na i idiae, and a e small posi i e-sense single s anded RNA i uses, which ha e been shown ecen ly o induce a pe sis en , slowly- eplica ing in ec ion in bo h myoca dium and panc eas. This may esul om al e a ion o he i al genome du ing he p og ess o in ec ion including he gene a ion o na u ally occu ing 5’-dele ions [12-14]. Se e al di ec (immunohis ochemis y) and Onco a ge 2 www.impac jou nals.com/onco a ge indi ec (se ology, isola ion o i uses om pa ien s) app oaches ha e con i med he p esence o en e o i us bo h in he ci cula ion and in he isle s o T1D pa ien s [15-20]. En e o i uses in he panc eas we e de ec ed by immunos aining o i al p o ein (VP1) [17, 21], which is highly exp essed unde acu e i al in ec ion and diminishes in pe sis en in ec ion and may no be de ec ed unde ci cums ances whe e i al eplica ion is comp omised [22]. Also, non-speci ic in e ac ion o he VP1 an ibody wi h o he cellula p o eins has been epo ed [23, 24]. A well-cha ac e ized coho o human panc ea ic dono issue has been es ablished by nPOD (Ne wo k o Panc ea ic O gan Dono s wi h Diabe es) and is a ailable, simila o o he coho s, mainly as o malin ixed, pa a in embedded samples (FFPE) [25]. This me hod ensu es p ese a ion o he samples o many yea s, bu a limi ing ac o is he ela i ely poo RNA in eg i y o en associa ed wi h FFPE p ese a ion which means ha analysis o issue samples by PCR can be di icul [26]. Also, de ec ion o he i al genome in he panc eas has been challenging and mo e sensi i e and eliable me hods a e equi ed. This has c ea ed p essu e o de elop al e na i e (accu a e and equally sensi i e) me hods o allow he de ec ion o RNA in single cells wi hin FFPE issue samples. An ob ious candida e is in si u hyb idiza ion, whe e labeled oligonucleo ide p obes speci ically pai wi h a ge nucleic acids ia Wa son-C ick base pai ing and some success has been gained using long p obes (~50-100n s) labeled wi h ei he adioiso opes o enzymes ca alyzing ch omogenic eac ions. Howe e , his app oach also has limi a ions in he ace o samples wi h poo RNA in eg i y o e y low a ge RNA copy numbe s. Thus, imp o ed RNA FISH me hods ha e been de eloped o o e come hese hu dles. En e o i al RNA was de ec ed by a new gene a ion o RNA p obes (Quan iGene ® ViewRNA Assay), which depend on signal ampli ica ion [25]. This app oach o e s he ad an age o signal ampli ica ion ia he use o b anched seconda y p obes bu , heo e ically, may be a ec ed by se e al condi ions. To c ea e a docking si e o he b anched p obes, wo p obes mus si adjacen o one-ano he on he a ge sequence, e ec i ely leng hening he de ac o sho p obes. Howe e , RNA deg ada ion can s ill a ec he docking o p obes and hence signal enhancemen . Flexibili y in choosing he binding si e wi h he bes he modynamic cha ac e is ics may also be comp omised. Because o such limi a ions o he de ec ion o en e o i uses wi h high accu acy and sensi i i y, we p esen he e an adap ed me hod o a ge single RNA molecules wi h sho (~20n s) luo escen ly labeled oligonucleo ides in si u. These oligonucleo ides anneal o common egions o he RNA genome o membe s o he coxsackie i us amily. Such sho singly labeled oligo RNA p obes a e esis an o RNAse and RNA de ec ion is less a ec ed by a ge RNA deg ada ion, making hese p obes mo e e sa ile while e aining sensi i i y and speci ici y. Because hei binding o he a ge sequences occu s independen ly o each p obe, his gi es p obe combina ions he ad an age ha he e is a deg ee o eedom in hei posi ioning wi hou isk o he loss o s ingency, e iciency o speci ici y. To gene a e a dis inc luo escen signal abo e backg ound noise, a su icien amoun o labeled p obes mus bind in close p oximi y [27-29], he eby ensu ing high speci ici y and conside able lexibili y o de ec ion a e hyb idisa ion. The use o small con iguous RNA species also o e comes he po en ial limi a ion o agmen ed a ge RNA. Wi h ou newly es ablished p o ocol, we success ully de ec i al RNA in bo h cell cul u e and FFPE issue sec ions, in combina ion wi h classical immunos aining. Sho p obes we e able o de ec i al in ec ion a lowe i al loads han classical immunos aining and he me hod is compa able in sensi i i y o ha o semi-nes ed PCR [30]. RESULTS An es ablished p o ocol o sho RNA-oligop obe labeling in FFPE issue sec ions In o de o de elop a obus p o ocol o RNA- oligop obe labeling o FFPE panc ea ic issue sec ions, we ini ially es ed a comme cially a ailable p obe se a ge ing he housekeeping gene (GAPDH) in cul u ed HEK293 cells and in isola ed FFPE human isle s (Figu e 1). Na u al, as well as ixa i e-de i ed signal noise, is a majo p oblem when employing luo escence mic oscopy o de ec p obes a ge ing RNA molecules, especially when hese a e p esen in low abundance. Thus, while ou es p obes ga e a e y speci ic signal in cul u ed HEK293 cells (Figu e 1A), p obing o FFPE human isle s gene a ed high backg ound noise bo h in he absence o p obes (Figu e 1B) and wi h he GAPDH p obes (Figu e 1C), when ollowing he s anda d p o ocol. Reduc ion in he FFPE-de i ed backg ound signal was achie ed by he emo al o any emaining pa a in wax c ys als (see ma e ial and me hods). This in ol ed he use o a p o ocol in which xylene washes we e unde aken a high empe a u es p io o a s ep u ilizing pepsin/HCl o sepa a e p o eins om nucleic acids. In addi ion, Sudan black was included o educe he o e all FFPE-de i ed backg ound luo escence. Using his modi ied p o ocol, single posi i e do s ( ep esen ing hyb idiza ion o as ew as one RNA molecule) could be de ec ed wi hin he Abbe- di ac ion limi and we e easily disc imina ed om any esidual backg ound noise (Figu e 1D). Onco a ge 3 www.impac jou nals.com/onco a ge Figu e 1: An es ablished p o ocol o sho RNA-oligop obe labeling in FFPE issue sec ions. A. GAPDH oligonucleo ides we e es ed in he cell line HEK 293. Single luo escen spo s, each ep esen ing one single RNA molecule, a e clea ly isible. B. No mal depa a iniza ion p ocedu e shows high backg ound noise in FFPE isle sec ions e en wi hou p obes (No P obes) and does no allow clea dis inc ion o p obe signal and backg ound, when GAPDH oligonucleo ides we e added C.; S anda d P o ocol. Modi ica ion o he depa a iniza ion and pos -hyb idisa ion p o ocol leads o backg ound educ ion and inc eased signal in ensi y o GAPDH oligonucleo ides D.; Modi ied P o ocol. RNA P obes a e labeled wi h Quasa 570 ( ed) and nuclei we e s ained wi h DAPI (blue), scale ba depic s 10µm. Onco a ge 4 www.impac jou nals.com/onco a ge RNA-oligonucleo ide p obe design and speci ici y P obe se CVB_1 included a wide ange o g oup B coxsackie i uses; i consis ed o a mix u e o 40 sho oligonucleo ides, each comp ising 17-22 nucleo ides co e ing he whole i al genome (see ma e ial and me hods and Figu e 2A). This enabled us o a ge single RNA molecules. To de ec a posi i e signal, i was de e mined ha a leas 17 o he p obes in a gi en se mus bind o hei a ge sequence wi h only one misma ch allowed wi h espec o he s ingency pa ame e s [25]. To es he speci ici y o he p obe se CVB_1, HEK293 cells we e ixed and p ocessed a e ei he cul u e wi hou i uses o ollowing in ec ion wi h coxsackie i uses CVB3 and CVB2 (Figu e 2B) which sha e a sequence simila i y o abou 79%. Cells we e in ec ed o 2h a an MOI o 5 and he i uses allowed o eplica e o 24h p io o ixa ion and analysis. P obes e icien ly de ec ed i al genomes wi hin in ec ed cells (Figu e 2B, CVB3 and CVB2) whe eas signal was absen om unin ec ed cells (Figu e 2B, non-in ec ed), con i ming ha he i us-speci ic p obe se had no o - a ge e ec s. The speci ici y o he p obe se was u he es ed in HEK293 cells in ec ed wi h cy omegalo i us (CMV), a DNA i us o he he pes i idae amily (Figu e 2B, CMV) o hepa i is A i us (HAV), a posi i e ssRNA i us o he pico na i idae amily (Figu e 2B, HAV). Following in ec ion a an MOI o 5 and incuba ion o 4 days o ensu e i al eplica ion, he e was no isible cy opa hic e ec . The p esence o i us was con i med by RT-PCR (Suppl.Figu e 1A). CMV appea ed no o be in an ac i e phase o eplica ion as no signal was gene a ed by RT- PCR o DNAse- ea ed samples. In nei he HAV no CMV in ec ed cells we e p obe-speci ic hyb idiza ion spo s de ec ed, con i ming he p obe speci ici y. HAV sha es pa ial sequence simila i y wi h CVB3 (<45%) while CMV is a DNA i us; o each, he numbe o “on a ge ” p obes was below he de ec ion limi o he assay. The likelihood ha he signals de ec ed in samples we e non-speci ic was u he excluded by s aining CVB3 in ec ed isle s in he absence o he p obe se and by s aining unin ec ed cells wi h he p obe se . In each case, he nega i e con ols deli e ed no s aining, whe eas i ally in ec ed cells yielded posi i e signals (Suppl.Figu e 1B). RNAse A ea men abolished he signal om in ec ed cells and con i med ha he p obe se is speci ic o i al RNA (Suppl.Figu e 1B). We u he es ed he CVB_1 p obe se on an a ay o cell lines p e iously gene a ed o use wi h Quan igene® ViewRNA i us p obes [21]. G een monkey kidney cells (GMK and Ve o), he human ce ix (HeLa), al eola (A549) epi helial ca cinomic and habdomyosa coma (RD) muscle cells we e in ec ed wi h i uses om he en e o i us g oups A and B o adeno i us (DNA i us) [21] (Table 1 and Suppl.Figu e 2). In line wi h he esul s ob ained wi h Quan igene® ViewRNA by Laiho e al. [21] he CVB_1 p obe se yielded posi i e s aining o i uses o bo h g oups A and B, while i did no s ain cells in ec ed wi h adeno i us and human pa echo i us 1 (HPeV1) (0/ 40 p obes binding). Also, he e was no binding o sequences om coxsackie i us A5 (11/40 p obes heo e ically ma ch he i us sequence). Table 1: RNA oligonucleo ide s aining o di e en pico na i idae and con ol i uses Vi us S ain Resul Vi us S ain Resul EV71 PB-EV71Hus ++ Echo3 PB-E3DiT23 ++ CVB1 ATCC ++ Echo4 ATCC ++ CVB2 ATCC ++ Echo6 ATCC + CVB3 ATCC ++ Echo9 ATCC + CVB4 ATCC ++ Echo11 ATCC + CVB5 ATCC ++ Echo30 ATCC ++ CVB6 ATCC ++ PV3 Sabin + CVA2 PB-CVA2V38 ++ HPeV1 ATCC - CVA4 PB-CVA4V36 ++ Adeno i us C VR846 - CVA5 PB-CVA5V43 - A549 cells -- CVA6 PB-CVA6V303V + RD cells -- CVA9 ATCC ++ Ve o cells -- CVA10 PB-CVA10V2530 ++ HeLa cells -- CVA16 PB-CVA16V55 ++ GMK cells -- P obe speci ici y was es ed on a cell a ay (FFPE); di e en cell lines we e spo ed ei he as unin ec ed con ols o in ec ed wi h di e en i uses. The esul s ob ained wi h RNA p obes se CVB_1 a e displayed. Rep esen a i e images a e shown in Fig. S2. Onco a ge 5 www.impac jou nals.com/onco a ge Figu e 2: RNA-oligonucleo ide p obe design and speci ici y. A. Scheme o cus om-designed oligonucleo ide (CVB_1) annealing h oughou he i al genome. B. Vi al RNA p obes we e es ed agains non-in ec ed, CVB3 and CVB2 (100% and 79% simila i y o consensus sequence, espec i ely; posi i e con ol), CMV, (DNA i us; nega i e con ol) and HAV (<45% simila i y o consensus sequence) in ec ed HEK 293 cells. Cells we e in ec ed wi h an MOI o 5 and ha es ed a e 24h (con ol, CVB3, CVB2) o 4 days (CMV, HAV) pos - in ec ion. RNA P obes a e labeled wi h Quasa 570 ( ed) and nuclei we e s ained wi h DAPI (blue), scale ba depic s 10µm. Onco a ge 6 www.impac jou nals.com/onco a ge RNA-oligonucleo ide p obe sensi i i y and consis ency Nex , we e alua ed sensi i i y o he CVB_1 p obe se . Ou RNA-FISH sys em was compa ed wi h wo o he mos widely used and well es ablished echniques o i us de ec ion: RT-PCR and immunohis ochemis y using an an ibody agains he i al capsid p o ein 1 (VP1). RT-PCR is he mos powe ul and speci ic ool o RNA de ec ion; bu RNA accessibili y and deg ada ion a e wo impo an limi ing ac o s. To compa e he sensi i i y o he RNA p obes wi h analysis by PCR, he panc ea ic line CM9 [31] and HEK293 cells we e in ec ed wi h CVB3. Since bo h RT-PCR and RNA-FISH can, in p inciple, de ec he p esence o a single RNA molecule, we used successi e 10- old se ial dilu ions o i us p io o in ec ion o cells and o compa e he sensi i i y o he me hods. Cells we e pla ed in duplica es and he i us (CVB3; s a ing MOI o 100) was cen i ugally inocula ed a 16°C o 1h o synch onize he in ec ion. A e inocula ion, any unbound i uses we e emo ed by washing and he cells incuba ed a 37°C o an addi ional hou , o allow i us in e naliza ion and genome elease om he capsid, be o e ixa ion o cell lysis o RNA ex ac ion. Bo h RNA-FISH (Figu e 3A, 3B and Suppl.Table 1) and RT-PCR (Figu e 3C and Suppl.Table 2) we e able o de ec he p esence o i al genomes, e en a he highes dilu ion (10-8) (Figu e 3B, 3C). Using RNA-FISH p obes, single RNA molecules we e de ec ed and a pla eau eached a a dilu ion o 10-3 i us by bo h isual coun ing and RT-PCR (Figu e 3B-3C). The speci ici y o he RT-PCR esul s we e con i med by examina ion o dissocia ion cu es, which we e iden ical in all cases while samples om non-in ec ed cells showed no signal (da a no shown). We nex compa ed he e iciency o he CVB_1 p obe se wi h he widely used VP1 (clone 5-D8/1, Dako cy oma ions) an ibody, using a cell a ay o he human al eola basal epi helial cell line A549 in ec ed wi h CVB1 o 2, 4 and 6 h o gene a e a popula ion ep esen ing di e en s ages o in ec ion [27]. A e in ec ion, cells we e se ially dilu ed wi h unin ec ed A549 cells o achie e a ange om undilu ed o 10-8. The cells we e hen ixed and pa a in embedded. When employed a a dilu ion o 1:2000 (which ensu es speci ici y and minimizes he possibili y o alse posi i es [23, 24]), he VP1 an ibody yielded posi i e signals only a dilu ions o 10 -1 o lowe , whe eas he RNA-FISH p obes we e able o de ec i al RNA e en a he highes dilu ion o 10-8 (Figu e 3D, 3E). The design o he RNA-FISH sys em should ci cum en he p oblem o RNA deg ada ion since i employs mul iple p obes o de ec he a ge RNA. RNA agmen a ion equen ly occu s in, o example, au opsy samples, whe e p ocessing and s o age unde RNAse ee condi ions is unlikely. We he e o e es ed FFPE CVB1- in ec ed GMK cell sec ions, which had been in ec ed and p ocessed a he same ime, bu hen cu and p ocessed a di e en imes (co e ing pe iods be ween 2012 and 2015) and s o ed a oom empe a u e. The p obes showed simila signals in all h ee samples ega dless o he p ocessing and s o age ime. Impo an ly, unin ec ed con ols we e nega i e (Suppl.Figu e 3). Coupling RNA-FISH and immunohis ochemis y Impo an ad an ages o he use o RNA- FISH p obes ela e no only o hei high sensi i i y and speci ici y bu also hei abili y o localize RNA molecules wi hin speci ic cells o issues such as he panc eas. We, he e o e, in es iga ed he localiza ion and dis ibu ion o he signal emana ing om he i al p obes wi hin he panc eas and compa ed his wi h de ec ion by immunohis ochemis y wi h he widely used VP1 an ibody. Figu e 4A shows a schema ic ep esen a ion o he expec ed p o ile o i us-s aining: RNA p obes ( ed) a e no expec ed o anneal o he i al RNA while i is packaged wi hin he capsid. Howe e , once eleased wi hin he cell, he p obes should bind. Con e sely, he VP1 an ibody (g een) should always bind o he capsid su ace o , concei ably, o ee VP1 which has no been inco po a ed in o capsids. To isualize any di e en ial labeling, human CM9 cells we e in ec ed wi h cen i ugally inocula ed CVB3 a an MOI o 1000. (Figu e 4B). An i-VP1 was de ec ed a he plasma memb ane and in he cy oplasm while RNA p obes we e localized exclusi ely in he cy oplasmic a ea (Figu e 4B, middle panel and Suppl.Figu e 4 o enhanced signals). As depic ed a la ge magni ica ion, signals om he RNA molecules co-localized, o we e in close p oximi y wi h hose a ising om he capsid p o ein, bu he wo we e no supe imposed. To e i y he s aining pa e ns, we also pe o med VP1 immunohis ochemis y p io o cell pe meabiliza ion and subsequen RNA-FISH, so ha he signals should no co-localize. As shown in Figu e 4B ( igh panel), unde hese condi ions he signals a ising om each me hod o de ec ion we e clea ly sepa a e. Unin ec ed CM9 cells we e used as con ols and showed no nonspeci ic signal o ei he an i-VP1 o he RNA- p obes (Figu e 4B). Success ul double s aining o i al RNA and VP1 was con i med in FFPE in ec ed mouse spleen. Spleen sec ions om mice in ec ed wi h CVB1 we e s ained wi h he CV_1 p obe se o de ec he i al genome and subsequen ly wi h an i-VP1 in an addi ional ound o s aining o he same sec ion. Bo h echniques showed posi i e s aining wi hin he same egions o he samples (Suppl.Figu e 5A). Double s aining o FFPE CVB3 in ec ed cul u ed a INS1-E cells o insulin and i us p obes showed co- localiza ion o i al RNA wi hin insulin posi i e INS- 1E cells (Figu e 4C). Close p oximi y o i al RNA and insulin was also ound in a Coxsackie in ec ed neona al Onco a ge 7 www.impac jou nals.com/onco a ge Figu e 3: RNA-oligonucleo ide p obe sensi i i y. A. CM 9 and HEK 293 we e in ec ed wi h a dilu ion se ies (MOI 102-10-8) o CVB3 and s ained wi h cus om-designed oligonucleo ides (CVB_1). Rep esen a i e images o CM9 and HEK293 cells in ec ed wi h ei he he highes o lowes dilu ion o CVB3 o he se ies a e shown. Whi e a ows highligh single i al spo s. B. Ten single images we e acqui ed o each dilu ion and single luo escen spo s we e manually coun ed. Resul s o HEK 293 (diamonds) and CM 9 (squa es) a e displayed as single spo s pe cell in loga i hmic scale. In o al, i al pa icles we e coun ed in 4470 HEK293 cells and CM9 cells. C. Vi al RNA om a pa allel expe imen was ex ac ed and analyzed by PCR; MOI o 100 (102) was se as 100%. D. Vi al RNA ( ed) and VP1 (1/2000; g een) s aining on a CVB1-dilu ion a ay o FFPE in ec ed GMK cells mixed wi h unin ec ed cells. RNA P obes we e labeled wi h Quasa 570 ( ed) and nuclei we e s ained wi h DAPI (blue), scale ba depic s 10µm. E. Summa y o he i al RNA and VP1 signals ob ained om a CVB1-dilu ion a ay o FFPE in ec ed GMK cells. Onco a ge 8 www.impac jou nals.com/onco a ge Figu e 4: Coupling RNA-FISH and Immunohis ochemis y. A. Theo e ical scheme o i al RNA and VP1 co-s aining. Ini ially, labeled oligonucleo ides canno bind i al RNA wi hin he capsid, bu only when he i us is eleased. On he o he hand, VP1 an ibody can bind o he capsid su ace. O e ime as mo e RNA is eleased, mo e p obes can anneal o hei a ge sequence. When a su icien amoun o oligonucleo ides is bound, g een (VP1) and ed (RNA) signal a e isible in close p oximi y. B. CM9 cells we e in ec ed wi h CVB3 (MOI 1000) o 1h a 16°C, ixed and p obed o i al RNA and VP1. VP1 (g een) was ound in close p oximi y wi h i al RNA ( ed), when capsid and RNA we e p esen in he cy osol a ea (B; middle panel). When VP1 s aining was pe o med be o e cell pe meabiliza ion no colocaliza ion o i al RNA/VP1 was de ec ed (B; igh panel). No c oss- eac i i y was ound wi h cellula p o eins o RNAs (B; le panel). C. Co-s aining and localiza ion o i al RNA ( ed) and insulin (g een) in CVB3-in ec ed INS-1E cells and HgCl2 ixed pa a in embedded CVB-in ec ed neona al panc eas. Samples we e s ained wi h RNA FISH p obes i s , analyzed and hen s ained o insulin. The non-g anula insulin s aining in his slide appea ed also by classical single insulin immunohis ochemis y and is caused by he issue condi ion. Nuclei we e isualized by DAPI s aining (blue); scale ba depic s 10µm. Onco a ge 9 www.impac jou nals.com/onco a ge mouse panc eas (Figu e 4C), con i ming he es ablished double-s aining p o ocol also in HgCl 2 ixed and pa a in embedded cells. To demons a e e icacy o i us s aining in a ious issues, we s ained a sec ion om a coxsackie- in ec ed neona al hea and ound cells displaying a s ong signal co esponding o i al RNA (Suppl.Figu e 5B). Vi al RNA localiza ion in he panc eas in T1D Ha ing es ablished he alidi y o ou app oach, we hen used an en e o i us genome alignmen (Suppl. Figu e 6) o design wo addi ional p obe se s (CVB_2 and CVB_3; Suppl.Tables 4,5), which would complemen he i s se and allow de ec ion o he en i e ange o g oup B en e o i uses. A scheme o he binding posi ions o he newly designed p obe se s is shown in Figu e 5A. We hen used a combina ion o he h ee p obe se s o a blinded analysis o au opsy panc ea ic issue eco e ed om nondiabe ic and T1D pa ien s ha ing emaining esidual β-cells om a UK coho [31, 32]. As shown in Table 2, we we e able o de ec he p esence o en e o i uses in 7 o 8 panc eas samples om T1D pa ien s and in 2 o 8 nondiabe ic con ols. While i al RNA was ound wi hin he insulin-posi i e isle a ea in 6 o 8 T1D pa ien s (Table 2 and Figu e 5B), 5 o 8 panc ea a also yielded posi i e signals o en e o i al RNA in he exoc ine a ea. A single T1D pa ien om he UK collec ion had i al RNA exclusi ely in he exoc ine a ea o he panc eas and no in he isle s (al hough in his pa ien , only 1 agmen ed and 2 no mal isle s we e ound h oughou he whole sec ion). Simila ly, his dis ibu ion o i al RNA was also obse ed in a panc eas ansplan biopsy om a pa ien wi h T1D who had de eloped ecu en disease om he nPOD-T ansplan a ion coho [32] (Figu e 5C). In one nondiabe ic con ol, we also ound i al RNA in isle s. Compa ison o ou esul s wi h an i-VP1 s aining on he same samples pe o med sepa a ely in Exe e , 66% conco dance was achie ed. Thus, while i us was de ec ed using i al p obes in 6 ou o 8 T1D panc ea a wi hin he insulin-posi i e isle a ea, VP1 s aining was p esen in only 4/8 panc ea a. DISCUSSION In his s udy we p esen a obus me hod o iden i y and localize en e o i al RNA in FFPE issue, which was es ablished o he analysis o i al in ec ion in panc ea a om pa ien s wi h T1D. The use o RNA p obes allowed he disc imina ion o i ally-encoded RNA om i us eplica ion loci wi h high ideli y and sensi i i y. Adap ing a me hod om Raj e al. [28, 29], we c ea ed a se o sho luo escen ly labeled oligonucleo ide p obes ha anneal o common egions o he coxsackie i us amily in si u o a ge single RNA molecules. P obe se s a e easily de eloped using he online S ella is® RNA FISH P obe Designe . Wi h h ee di e en se s o p obes we we e able o localize i al RNA in all es ed issue including FFPE issues om pa ien s wi h p o en coxsackie i us in ec ion and in he panc eas o 7 Table 2: RNA oligonucleo ide s aining o human panc eases Sample G oup ID Age Sex Fixa ion Vi al RNA +* Isle i al RNA + VP1+ T1D E560 42 F BF ++++ + VP1+ T1D 11746 6M HgCl2++ + VP1+ T1D E375 11 F FS + + VP1+ T1D E554A 7MBouin - § + VP1 – T1D E428 5M BF ++ + VP1 – T1D E514 23 M BF - + VP1 – T1D E235 6M HgCl2- - § VP1 – T1D 8869 8M HgCl2+ - § Non-diabe ic con ol 8579 7 - HgCl2- - Non-diabe ic con ol 12054 7 - HgCl2- - Non-diabe ic con ol 330/71 47 M BF - - Non-diabe ic con ol 21/89 4F BF ++ - Non-diabe ic con ol 274/91 6M BF - - Non-diabe ic con ol 191/67 25 M BF - - Non-diabe ic con ol 315/89 9M BF + + Non-diabe ic con ol 540/91 11 M BF - - Summa y o he compa a i e s udy o panc ea a om T1D and non-diabe ic con ol pa ien s. Sec ions we e i s s ained o i al RNA, wi h a combina ion RNA p obe se s CVB_1, CVB_2 and CVB_3, ollowed by insulin s aining. - = less han 10 ully-in ec ed cells, + = 10 o 30 ully-in ec ed cells, ++ = 30 o 100 ully-in ec ed cells, +++ = 100 o 200 ully-in ec ed cells, ++++ = mo e han 200 ully-in ec ed cells in he exoc ine. §E554A showed en ully-in ec ed cells, E235 con ained jus 5 isle s on he sec ion and 8869 jus 2 ull and 1 agmen ed isle s ( e y small sec ions). Onco a ge 16 www.impac jou nals.com/onco a ge M, Mall G and Kandol R. Ongoing en e o i us-induced myoca di is is associa ed wi h pe sis en hea muscle in ec ion: quan i a i e analysis o i us eplica ion, issue damage, and in lamma ion. P oc Na l Acad Sci U S A. 1992; 89(1):314-318. 23. Richa dson SJ, Lee e P, Dhayal S, Russell MA, Oika inen M, Laiho JE, S edin E, Lind K, Rosenling T, Chapman N, Bone AJ, n PODVC, Foulis AK, e al. E alua ion o he ideli y o immunolabelling ob ained wi h clone 5D8/1, a monoclonal an ibody di ec ed agains he en e o i al capsid p o ein, VP1, in human panc eas. Diabe ologia. 2014; 57(2):392-401. 24. Hansson SF, Ko sg en S, Pon en F and Ko sg en O. En e o i uses and he pa hogenesis o ype 1 diabe es e isi ed: c oss- eac i i y o en e o i us capsid p o ein (VP1) an ibodies wi h human mi ochond ial p o eins. J Pa hol. 2013; 229(5):719-728. 25. Laiho JE, Oika inen S, Oika inen M, La sson PG, S one VM, Hobe D, Obe s e S, Flods om-Tullbe g M, Isola J and Hyo y H. Applica ion o bioin o ma ics in p obe design enables de ec ion o en e o i uses on di e en axonomic le els by ad anced in si u hyb idiza ion echnology. J Clin Vi ol. 2015; 69:165-171. 26. Skog O, Ing as S and Ko sg en O. E alua ion o RT- PCR and immunohis ochemis y as ools o de ec ion o en e o i us in he human panc eas and isle s o Lange hans. J Clin Vi ol. 2014; 61(2):242-247. 27. Femino AM, Fay FS, Foga y K and Singe RH. Visualiza ion o single RNA ansc ip s in si u. Science. 1998; 280(5363):585-590. 28. Raj A and Tyagi S. De ec ion o indi idual endogenous RNA ansc ip s in si u using mul iple singly labeled p obes. Me hods Enzymol. 2010; 472:365-386. 29. Raj A, an den Bogaa d P, Ri kin SA, an Oudenaa den A and Tyagi S. Imaging indi idual mRNA molecules using mul iple singly labeled p obes. Na Me hods. 2008; 5(10):877-879. 30. Laiho JE, Oika inen M, Richa dson SJ, F isk G, Nyalwidhe J, Bu ch TC, Mo is MA, Oika inen S, Pugliese A, Do a F, Campbell-Thompson M, Nadle J, Mo gan NG, Hyo y H and G oup Jn-V. Rela i e sensi i i y o immunohis ochemis y, mul iple eac ion moni o ing mass spec ome y, in si u hyb idiza ion and PCR o de ec Coxsackie i us B1 in A549 cells. J Clin Vi ol. 2016; 77:21- 28. 31. G agnoli C. The CM cell line de i ed om li e me as asis o malignan human insulinoma is no a alid be a cell model o in i o s udies. Jou nal o cellula physiology. 2008; 216(2):569-570. 32. Pugliese A, Vend ame F, Reijonen H, A kinson MA, Campbell-Thompson M and Bu ke GW. New insigh on human ype 1 diabe es biology: nPOD and nPOD- ansplan a ion. Cu Diab Rep. 2014; 14(10):530. 33. Do a F, Galle i L, Sebas iani G and Vend ame F. Vi us In ec ions: Lessons om Panc eas His ology. Cu Diabe es Rep. 2010; 10(5):357-361. 34. Mo gan NG and Richa dson SJ. En e o i uses as causa i e agen s in ype 1 diabe es: loose ends o los cause? T ends Endoc inol Me ab. 2014; 25(12):611-619. 35. KOCH R. Uebe bak e iologische Fo schung. . Ve handlungen des X In e na ionalen Medichinischen Cong esses, Be lin 1890, Augus Hi shwald, Be lin, Ge many, 10 h In e na ional Cong ess o Medicine, Be lin, Ge many 1. 1891. 36. Chapman NM and Kim KS. Pe sis en coxsackie i us in ec ion: en e o i us pe sis ence in ch onic myoca di is and dila ed ca diomyopa hy. Cu Top Mic obiol Immunol. 2008; 323:275-292. 37. Viegas MS, Ma ins TC, Seco F and do Ca mo A. An imp o ed and cos -e ec i e me hodology o he educ ion o au o luo escence in di ec immuno luo escence s udies on o malin- ixed pa a in-embedded issues. Eu J His ochem. 2007; 51(1):59-66. 38. E ben T, Ossig R, Naim HY and Schnekenbu ge J. Wha o do wi h high au o luo escence backg ound in panc ea ic issues - an e icien Sudan black B quenching me hod o speci ic immuno luo escence labelling. His opa hology. 2016; 69(3):406-422. 39. Van de Winkle M, Maes E and Pipelee s D. Isle cell analysis and pu i ica ion by ligh sca e and au o luo escence. Biochem Biophys Res Commun. 1982; 107(2):525-532. 40. Willcox A, Richa dson SJ, Bone AJ, Foulis AK and Mo gan NG. Analysis o isle in lamma ion in human ype 1 diabe es. Clin Exp Immunol. 2009; 155(2):173-181. 41. Foulis AK, Liddle CN, Fa quha son MA, Richmond JA and Wei RS. The his opa hology o he panc eas in ype 1 (insulin-dependen ) diabe es melli us: a 25-yea e iew o dea hs in pa ien s unde 20 yea s o age in he Uni ed Kingdom. Diabe ologia. 1986; 29(5):267-274. 42. A des ani A, Pa oni F, Azizi Z, Kau S, Khob agade V, Yuan T, F ogne T, Tao W, Obe holze J, Pa ou F, Ke Con e J and Maedle K. MST1 is a key egula o o be a cell apop osis and dys unc ion in diabe es. Na Med. 2014; 20(4):385-397. 43. B ack K, F ings W, Do zaue A and Vallb ach A. A cy opa hogenic, apop osis-inducing a ian o hepa i is A i us. J Vi ol. 1998; 72(4):3370-3376. 44. Abha A. Low empe a u e la en hea he mal ene gy s o age: hea s o age ma e ials. Sola ene gy. 1983; 30(4):313-332. 45. Ole J, Wiedo n KH, Goldmann T, Kuhl H, Meh aein Y, Sche han H, Nike eghad F, Vollme E, Mulle AM and Mulle -Na ia J. HOPE ixa ion: a no el ixing me hod and pa a in-embedding echnique o human so issues. Pa hol Res P ac . 2001; 197(12):823-826. Onco a ge 17 www.impac jou nals.com/onco a ge 46. Ru ledge RG and S ewa D. Assessing he pe o mance capabili ies o LRE-based assays o absolu e quan i a i e eal- ime PCR. PLoS One. 2010; 5(3):e9731. 47. Ausubel FM, B en R and Kings on RE. Sho P o ocols in Molecula Biology. John Wiley & Sons. 2003. 48. Kea se M, Moi R, Wilson A, S ones-Ha as S, Cheung M, S u ock S, Bux on S, Coope A, Ma kowi z S, Du an C, Thie e T, Ash on B, Mein jes P and D ummond A. Geneious Basic: an in eg a ed and ex endable desk op so wa e pla o m o he o ganiza ion and analysis o sequence da a. Bioin o ma ics. 2012; 28(12):1647-1649.