In situ characterization of stem cells-like biomarkers in meningiomas

Alami e al. Cance Cell In (2018) 18:77 h ps://doi.o g/10.1186/s12935-018-0571-6 PRIMARY RESEARCH In si u cha ac e iza ion o s em cells-like bioma ke s inmeningiomas Hanin Alami 1†, Mona Aloma i2†, Abdulla Ahmed A. Salwa i2, Mohamad Saka2, Mohammed Bangash3, Saleh Baeesa3, Fahad Alghamdi4, Angel Ca acedo5,6, Hans‑Jue gen Schul en6, Adeel Chaudha y1,6,7, Adel Abuzenadah1,2,7 and Deema Hussein2* Abs ac Backg ound: Meningioma cance s em cells (MCSCs) con ibu e o umo agg essi eness and d ug esis ance. Suc‑ cess ul he apies de eloped o inope able, ecu en , o me as a ic umo s mus a ge hese cells and es ic hei con ibu ion o umo p og ession. Un o una ely, he iden i y o MCSCs emains elusi e, and MSCSs’ in si u spa ial dis ibu ion, he e ogenei y, and ela ionship wi h umo g ade, emain unclea . Me hods: Se en umo s classi ied as g ade II o g ade III, including one case o me as a ic g ade III, and eigh g ade I meningioma umo s, we e analyzed o combina ions o en s em cell (SC)‑ ela ed ma ke s using immuno luo es‑ cence o consecu i e sec ions. The co ela ion o exp ession o all ma ke s we e in es iga ed. Th ee dimensional spa‑ ial dis ibu ion o ma ke s we e quali a i ely analyzed using a g id, designed as a eposi o y o in o ma ion o posi i e s aining. All s a is ical analyses we e comple ed using S a is ical Analysis So wa e Package. Resul s: The pa e ns o exp ession o SC‑ ela ed ma ke s we e de e mined in he con ex o wo dimensional dis i‑ bu ion and cellula ea u es. All ma ke s could be de ec ed in all umo s, howe e , F izzled 9 and GFAP had di e en ial exp ession in g ade II/III compa ed wi h g ade I meningioma issues. Co ela ion analysis showed signi ican ela ion‑ ships be ween he exp ession o GFAP and CD133 as well as SSEA4 and Vimen in. Da a om h ee dimensional analy‑ sis showed a complex dis ibu ion o SC ma ke s, wi h inc eased gene he e o‑exp ession being associa ed wi h g ade II/III umo s. Sub egions ha showed mul iple co‑s aining o ma ke s including CD133, F izzled 9, GFAP, Vimen in, and SSEA4, bu no necessa ily he p oli e a ion ma ke Ki67, we e highly associa ed wi h g ade II/III meningiomas. Conclusion: The dis ibu ion and le el o exp ession o CSCs ma ke s in meningiomas a e a iable and show he e o‑ exp ession pa e ns ha ha e a complex spa ial na u e, pa icula ly in g ade II/III meningiomas. Thus, esul s s ongly suppo he no ion o he e ogeneous popula ions o CSCs, e en in g ade I meningiomas, and call o he use o mul iple ma ke s o he accu a e iden i ica ion o indi idual CSC subg oups. Such iden i ica ion will lead o p ac ical clinical diagnos ic p o ocols ha can quan i a e CSCs, p edic umo ecu ence, assis in guiding ea men selec ion o inope able umo s, and imp o e ollow up o he apy. Keywo ds: Meningioma, Cance s em cells, Immuno luo escence, CD133, SOX2, Nes in, F izzled 9, GFAP, SSEA4, Olig2 © The Au ho (s) 2018. 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 eco mmons .o g/licen ses/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. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i eco mmons .o g/ publi cdoma in/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Open Access Cance Cell In e na ional *Co espondence: [email p o ec ed]; [email p o ec ed] †Hanin Alami and Mona Aloma i a e i s co‑au ho s 2 King Fahd Medical Resea ch Cen e , King Abdulaziz Uni e si y, P.O. Box. 80216, Jeddah 21589, Saudi A abia Full lis o au ho in o ma ion is a ailable a he end o he a icle Page 2 o 14 Alami e al. Cance Cell In (2018) 18:77 Backg ound Meningiomas occu in mul iple ex a-axial loca ions wi hin a achnoid memb anes and a e highly equen compa ed wi h o he ypes o cen al ne ous sys em umo s (CNSTs) [1–3]. Gene ic analyses o bulk menin- gioma issues iden i ied mu a ions in se e al pa hways including he phosphoinosi ide 3-kinase (PI3K) and he G p o ein-coupled ecep o smoo hened (SMO) signaling pa hways [4–6]. His opa hologically, hese umo s a e classi ied by he Wo ld Heal h O ganiza ion (WHO) in o 15 a ian s wi hin g ades I o III. Un o u- na ely, up o 20% o g ade I umo s eoccu , and apa om Mib-1, molecula ma ke s ha enable p edic ion o ecu ence ha e no been es ablished [3, 7, 8]. Meningiomas ha e been shown o ha bo cance s em cells (CSCs), highly esilien cance cells ha employ de egula ed s em cell (SC) exp ession p o iles and a e capable o causing eoccu ence [9–14]. Ta ge - ing CSCs is p edic ed o enhance he apy ou comes [3]. A ange o genes and hei p o eins ha e been associ- a ed wi h he iden i y o CNST CSCs. CD133/P om- inin-1, a i e- ansmemb ane glycop o ein, is no mally exp essed in emb yonic neu al SC adial glial/ependy- mal cells and in ependymal cells in he adul b ain [15]. The p o ein is hough o in e ac wi h selec ed gangliosides o modula e cell- o-cell con ac in a cell cycle- ela ed manne [16, 17]. In CNSTs, high CD133 exp ession has been associa ed wi h poo su i al [18– 21]. In meningioma cell lines, highe CD133 exp ession co ela es posi i ely wi h cell p oli e a ion and d ug esis ance [9, 13, 22, 23]. The exp ession o Nes in, a ype VI in e media e ila- men , has been shown o be impo an CSC ma ke o CNST g ow h, mig a ion, and in asion [24–26], pos- sibly by in luencing he cell cycle [27]. Highe exp es- sion o Nes in has been de ec ed in g ades II and III meningiomas compa ed o g ade I [28]. The de egu- la ed exp ession o he ansc ip ion ac o SOX2 has also been obse ed in se e al CNST CSCs [29–32]. The knockdown o SOX2 was shown o slow he g ow h and p oli e a ion o GBM CSCs [33]. In GBM cells posi i e o CD133, silencing SOX2 impai ed umo ini ia ion and d ug esis ance [34]. F izzled 9 (FZD9) belongs o he izzled p o ein amily, ans-memb ane signaling molecules ha ac as ecep o s o he WNT p o ein, and plays a key ole in cell de elopmen by main aining plane cell pola i y [35]. Mu a ions in FZD/WNT genes a e linked o se e al malignancies [36]. In as ocy oma and glioblas oma, FZD9 is p edominan ly exp essed by neoplas ic cells, and i s exp ession is posi i ely co ela ed wi h WHO g ading and Ki-67 posi i i y [37]. Inhibi - ing he FZD amily in glioblas oma cell lines leads o inc eased di e en ia ion [38]. S age-speci ic emb yonic an igen-4 (SSEA4), also known as FUT4 and CD15, is a glycosphingolipid (GSL) con aining a e minal sialic acid esidue (N-ace ylneu- aminic acid) and is in ol ed in he globo-se ies gan- glioside syn hesis. SSEA4 is highly exp essed du ing he p eimplan a ion s age in ge m cells in he es is and o a- ies, and is down- egula ed upon di e en ia ion [39–41]. Ta ge ing SSEA4 in i o supp essed he g ow h o GBM cell lines [42], and cells posi i e o SSEA4 ha e a highe capabili y o me as asis and in asion [43–47]. Olig2 is a basic helix–loop–helix (bHLH) ansc ip ion ac o ha is exp essed in oligodend ocy es and in oligodend i ic p ogeni o cells [48–50]. The p o ein was shown o medi- a e he p oli e a ion, mig a ion, and in asion o bo h no - mal as ocy es and malignan GBM cells [50–53]. P o eins associa ed wi h he di e en ia ion o SCs include Vimen in, glial ib illa y acidic p o ein (GFAP), and be a III ubulin (βIII- ubulin/βIIIT). Vimen in is a class III in e media e p o ein ha is exp essed in mes- enchymal cells. The p o ein’s main unc ion is o sup- po he cy oskele on [54], and i is highly associa ed wi h meningiomas [55]. GFAP is a class III in e media e ilamen p o ein, wi h i e di e en iso o ms (GFAPα, GFAPβ, GFAP gamma γ, GFAP δ, and GFAP k), and was shown o be exp essed in he as ocy e lineage du ing he de elopmen o he CNS [56, 57]. βIII-Tubulin is a neu on-speci ic mic o ubule equi ed o neu onal axon guidance, main enance, and de elopmen [58]. Mu a ions in he βIII- ubulin gene esul in mul iple diso de s o he CNS [59], and high p o ein exp ession is equen ly de ec ed in se e al CNSTs [60]. Al hough no limi ed o he iden i y o CSCs, hese ma ke s a e equen ly asso- cia ed wi h i , and hei exp essions a y acco ding o umo ype and p og ession [61]. Impo an ly, ecen e idence has indica ed ha he he e o-iden i y o CSCs can be de ec ed e en wi hin a single umo de eloped in a pa ien [62, 63]. P e iously, we published gene exp ession p o iles o mos o he meningioma pa ien s’ issues collec ed o ou coho [64, 65], as well as o hei co esponding cell lines [22]. Fo his wo k, we aimed o de e mine he he e o-dynamic cha ac e is ics o MCSCs in si u and iden i y di e en ial pa e ns associa ed wi h g ades II/III umo s. Me hods Sample collec ion Meningioma specimens collec ed be ween Feb ua y 2013 and Decembe 2015 we e ob ained wi hin 30min o umo emo al and ozen immedia ely a − 80°C. Neu- opa hologis s diagnosed su gical specimens acco ding o WHO classi ica ion. The clinical p o iles o he included pa ien s and hei umo s’ his opa hological ea u es a e Page 3 o 14 Alami e al. Cance Cell In (2018) 18:77 shown in Addi ional ile1: TableS1. Addi ional ile2: Fig- u e S1 shows H&E ep esen a i e sec ions o his ological a ian s o meningiomas included in his wo k, as well as a ypical ea u es. The exp ession p o iles o p e alen cance d i e genes [66], ex ac ed om a o emen ioned publica ions, a e shown in Addi ional ile3: TableS2. Cy o ial sec ioning Each ozen issue was c yosec ioned o gene a e 10 consecu i e sec ions a a hickness o 4 µm. Slides o sec ions we e s o ed a − 20 °C un il p ocessed o immuno luo escence. Immuno luo escence s aining Sec ions we e le a oom empe a u e o 5 min o de os , and issues we e enclosed wi h wax o e ain solu ions. Then, hey we e washed i e imes o 5min in phospha e bu e ed saline (PBS). Sec ions we e ixed wi h 4% o malin o 10min, hen washed h ee imes o 5in wi h PBS. Sec ions we e pe meabilized, blocked o non-speci ic an igens wi h eshly made blocking eagen (5% no mal goa se um, 0.25% T i on X-100 in PBS), and incuba ed o 1h a oom empe a u e. Sin- gle o double p ima y an ibodies solu ions (An ibod- ies, 2% NGS, 0.25% T i on X-100 in PBS) we e added o each sec ion, and sec ions we e incuba ed in a humidi y chambe o e nigh a 4°C. The ollowing day, sec ions we e washed h ee imes o 10 min wi h 0.25% T i- on X-100 in PBS (PBST) be o e incuba ing hem wi h a seconda y an ibodies solu ion (488 goa an i-mouse (1:300, ab150105, abcam) and 555 goa an i- abbi (1:700, ab150074, abcam) o 1h in he da k a oom empe a- u e. Sec ions we e hen washed i e imes o 5min wi h PBST. PBST was emo ed, and a d op o Vec ashield wi h DAPI was added o each sec ion o s ain nuclei. Fo each issue, sec ions we e s ained in he ollowing o de : sec- onda y only (nega i e con ol); mouse an i-Nes in (1:50, ab6142, abcam) wi h abbi an i-Ki67 (1:200, ab16667, abcam); mouse an i-CD133 (1:100, 130-092-395, Mil e- nyi) wi h abbi an i-SOX2 (1:200, 09-0024, S emgen ); mouse an i-Vimen in (1:100, ab8978, abcam) wi h ab- bi an i-F izzled 9 (1:100, ab150515, abcam); abbi an i-GFAP (1:500, ab7260, abcam); abbi an i-be a III Tubulin (1:500, ab18207, abcam), mouse an i-SSEA4 (1:100, ab16287, abcam) wi h abbi an i-SOX2 (1:200, 130-095-636, Mil enyi); and mouse an i-SSEA4 (1:100, ab16287, abcam) wi h abbi an i-Olig2 (1:500, Ab42453, abcam). P ocessed slides we e s o ed a 4°C. Image acquisi ion, enhancemen , andcoun ing All images we e aken wi hin he i s 2weeks a e s ain- ing. Fo each sec ion, i e coo dina e- ixed dispe sed egions we e selec ed o image. Pic u es we e aken a 20× magni ica ions using a Leica DMI6000 mic oscope and Leica DFC425 came a. Pho os o indi idual channels we e combined in Pho oshop 7.0.1. Enhancemen s o he images we e cons ained by signal le els o nega i e con ols o seconda y an ibodies only. Due o he complexi y o s ain- ing ea u es, co-posi i e, mono-posi i e, and nega i e cells we e manually coun ed o each egion wi hin each sec ion using Pho oshop 7.0.1. Manual coun ing was pe o med wice by wo independen scien is s, and indica ions o posi i i y o each ma ke and inal coun s we e con i med wi h a neu opa hologis . Images o Ki67 s ained sec- ions we e also coun ed by an independen hi d pe son using au oma ed coun ing in Image J so wa e o analysis. Images we e masked o coun nuclei posi i e o Ki67, and coun s we e p oduced using ICTN plugin. S a is ical analysis o  heda a The esul s we e analyzed using SPSS e sion 21.0 o gen- e a e desc ip i e and in e en ial s a is ics. The di e ences be ween he manual and au oma ed coun s o Ki67 we e analyzed using - es s. The di e ences o he coun s o exp essions be ween g ades and he di e ences in he numbe o iden i ied unique sub- egions be ween indi- idual umo s we e explo ed using analysis o a iance (ANOVA) obus es s o equali y o means, and P- alues o Welch and B own–Fo sy he we e indica ed. Co ela- ions o ma ke s’ exp essions ac oss consecu i e umo sec ions we e analyzed using Spea man’s Rho co ela ion. Chiχ2 was used o es o he signi icance be ween g ades o indi idual sub- egions. Resul s In si u ea u es o SC associa ed ma ke s inmeningiomas The pa e ns o exp essions o all u ilized ma ke s we e obse ed in meningioma issues (Fig.1). Posi i ely s ained cells o nuclea Ki67 we e consis en ly dispe sed as single cells wi hin indi idual umo sec ions. Cells posi i e o nuclea SOX2 and cy oplasmic FZD9 we e consis en ly seen in niche-s ained oci, while cells posi i e o cy o- plasmic Vimen in we e de ec ed in la ge posi i e egions and had homo-exp ession pa e ns. Cells posi i e o Nes- in, CD133, GFAP, BIIIT, SSEA4, and Olig2 had a umo - dependen pa e n o exp ession, which did no ha e a dicho omous associa ion wi h g ade. Memb anous CD133 was de ec ed in 12 umo s, and Olig2 could be seen a he nuclea en elope, as well as he nucleus, in all umo s. E alua ion o  hea e age exp essions o single p o eins ing ade I andg ade II/III meningiomas iden i ied GFAP andFZD9 assigni ican di e en ial ma ke s Da a o Ki67 coun s showed no signi ican di e ence be ween he manual and au oma ed me hod (T es , P = 0.5), Addi ional ile4: Figu e S2, suppo ing he use Page 4 o 14 Alami e al. Cance Cell In (2018) 18:77 o manual coun ing o o he ma ke s ha we e complex o assess using au oma e me hods. The analysis o a e - age coun s o each single ma ke ’s posi i e s aining o g ade I and g ade II/III umo s indica ed Ki67+, Vimen- in+, BIIITubulin+ as di e en ial ma ke s (B own– Fo sy he ANOVA, P < 0.05), espec i ely, as shown in Table1 and Fig.2. Fo highly signi ican g ade- ela ed di e en ial ma ke s, single posi i e s aining o FZD9+ o GFAP+ was s a is ically signi ican ly highe in g ade II/III meningiomas (B own–Fo sy he ANOVA, P < 0.01). Fo double-s aining analysis (Table1 and Fig.3), he mos signi ican a e age coun inc ease in g ade II/III meningi- omas was seen o Vimen in+FZD9+ (B own–Fo sy he ANOVA, P < 0.01). The a e ages o cell coun s aining SSEA4+Olig2+, Nes in−Ki67+, o CD133−Sox+ we e also highe in g ade II/III meningiomas (B own–Fo sy he ANOVA, P < 0.05), while he a e age o he numbe o CD133+Sox+ cells dec eased in g ade II/III compa ed o g ade I meningiomas (B own–Fo sy he ANOVA, P < 0.05). Consecu i e sec ions ha e simila exp essions o asingle ma ke To de e mine he na u e o he posi i e spa ial dis ibu- ion o a single ma ke h oughou he dep h o a umo , he exp ession p o ile o bo h SSEA4 and SOX2 was de e mined in adjacen and dis al consecu i ely sec- ioned immuno luo escence-p ocessed issues. Adjacen sec ions six and se en we e s ained o de ec SSEA4, while dis al sec ions wo and six we e s ained o de ec SOX2 (Fig.4). The pe cen ages o cells posi i e o SSEA4 in sec ion six co ela ed wi h posi i e cells o SSEA4 in he adjacen sec ion se en (Spea man’s Rho co ela ion coe icien = 0.687, P < 0.001). Simila ly, he pe cen - ages o cells posi i e o SOX2 in sec ion wo co e- la ed wi h posi i e cells o SOX2 in he dis al sec ion 50µm a b Ma ke Pa e n o Exp ession (Maximum GI:8 , GII/III: 7) se u aeF alulleCeussiTnih iWnoi ubi siD Homo-exp ession He e o-exp essionNuclea Cy oplasmic Nuclea en elope Memb ane Niche Dispe sed Ki67 AllAll Nes in llA6:III/IIG,8:IG1:III/IIG Sox2 llAllAllA CD133 7:III/IIG,5:IGllA7:III/IIG,7:IG1:IG Vimen in llAllA FZD9 llAllA GFAP llA3:III/IIG,6:IG4:III/IIG,2:IG BIIIT llA5:III/IIG,7:IG2:III/IIG,1:IG SSEA4llA1:IG7:III/IIG,7:IG Olig2 llA7:III/IIG,6:IGllA1:III/IIG,6:IG6:III/IIG,2:IG DapiKi67 DapiNes in DapiSox2 DapiCD133DapiVimen in DapiFZD9 DapiGFAP DapiBIIIT DapiSSEA4 DapiOlig2 Fig. 1 Cellula ea u es and pa e ns o exp ession o all he ma ke s used o s ain meningioma issues. a Immuno luo escence ep esen a i e images showing Ki67 (Red), Nes in (g een), SOX2 ( ed), CD133 (g een), Vimen in (g een), FZD9 ( ed), GFAP ( ed), BIIIT ( ed), SSEA4 (g een), and Olig2 ( ed), each wi h DAPI (blue). b A able summa izing pa e ns o exp ession in e ms o he dis ibu ion wi hin issue and obse ed cellula ea u es. G g ade. All images we e aken a ×20 Page 5 o 14 Alami e al. Cance Cell In (2018) 18:77 six (Spea man’s Rho co ela ion coe icien = 0.749, P < 0.001). The e a e signi ican co ela ions be ween heexp essions o di e en SC associa ed ma ke s ac ossconsecu i e issues Since he exp ession p o iles o each o SOX2 and SSEA4 we e equi alen ly spa ially dis ibu ed h oughou con- secu i e sec ions o a umo mass, co ela ions be ween he exp essions o di e en single ma ke s ac oss all con- secu i e sec ions we e in es iga ed (Fig.5). Exp ession da a indica ed a highly signi ican co ela ion be ween he exp essions o Vimen in and SSEA4 and he exp es- sions o CD133 and GFAP. Signi ican co ela ions we e obse ed o he exp essions o SSEA4 wi h CD133 o Nes in, and SOX2 wi h BIIIT. FZD9 also had signi ican co ela ions wi h Vimen in, SOX2 o wi h Olig2. The p esence o Nes in-posi i e p oli e a ing cells co ela ed wi h he p esence o Vimen in+FZD9+ cells. Quali a i e analysis o sub‑a eas ac ossconsecu i e sec ions show inc eased he e o‑ egional exp ession ing ades II/III meningiomas To in es iga e he ela ionship be ween mul iple ma k- e s ac oss consecu i e sec ions, images o a coo di- na e- ixed egion wi hin s ained sec ions we e sco ed using a g id wi h 96 sub- egions, each co e ing an a ea o 0.0037mm2. The g id was used as a eposi o y shee o quali a i e in o ma ion o posi i e s aining in each sub-a ea o all consecu i e sec ions o each umo , as exempli ied in Fig.6a, Addi ional ile5: Figu e S3, and Addi ional ile6: Figu e S4. Collec i ely, he da a showed a complex dis ibu ion o he sco ing o he combined SC associa ed ma ke s, ac oss indi idual issues (208 unique combina ions, Addi ional ile7: TableS3), wi h inc eased he e o- egional exp ession being associa ed wi h g ade II/III meningiomas (ANOVA, P < 0.01, Fig.6b). In e es - ingly, he le el o he e o- egional exp ession sepa a ed umo s in o h ee signi ican ly di e en g oups (ANOVA, P < 0.01), wi h all umo s in g oup 1 (R1) being g ade I and all meningiomas in g oup 3 (R3) being g ade II/III, while umo s in g oup 2 (R2) had mixed g ades o I and II. Regions ha we e signi ican ly equen ly occu ing in g ade II/III bu ne e in g ade I meningiomas included hose ha we e posi i e o CD133+SOX2±Vimen in+ FZD9+GFAP+BTIII+SSEA4+Olig2+, and Nes in+Ki6 Table 1 The means o  exp essions, s anda d e o s, andANOVA P alues o g ade I e susg ade II/III umo s o single anddouble-s ained ma ke s Ma ke (s) G ade Mean STD e o P alue Nes in+GI 29.45 5.70 0.231 GII/III 39.22 5.73 Ki67+GI 0.77 0.17 0.019* GII/III 2.72 0.78 CD133+GI 36.60 6.22 0.770 GII/III 39.11 5.86 Sox2+GI 12.45 2.76 0.929 GII/III 12.82 3.22 Vimen in+GI 83.13 4.31 0.016* GII/III 94.56 1.57 F izzled9+GI 11.82 2.35 0.000** GII/III 31.23 4.10 GFAP+GI 49.11 5.80 0.000** GII/III 78.03 3.28 BIIITubulin+GI 30.65 5.16 0.033* GII/III 47.01 5.49 SSEA4+GI 75.70 4.57 0.053 GII/III 87.11 3.57 Olig2+GI 55.33 5.02 0.072 GII/III 67.55 4.43 Nes in+Ki67+GI 0.51 0.16 0.080 GII/III 1.23 0.37 Nes in+Ki67−GI 28.94 5.59 0.258 GII/III 37.98 5.63 Nes in−Ki67+GI 0.26 0.08 0.037* GII/III 1.49 0.56 CD133+Sox2+GI 11.73 2.71 0.039* GII/III 5.48 1.20 CD133+Sox2−GI 24.87 4.41 0.224 GII/III 33.63 5.61 CD133−Sox2+GI 0.72 0.31 0.023* GII/III 7.35 2.77 Vimen in+FZD9+GI 11.80 2.36 0.000** GII/III 31.08 4.12 Vimen in+FZD9−GI 71.34 3.87 0.159 GII/III 63.48 3.95 Vimen in−FZD9+GI 0.03 0.02 0.125 GII/III 0.15 0.08 SSEA4+SOX2+GI 12.57 2.50 0.565 GII/III 10.59 2.36 SSEA4+SOX2−GI 63.13 4.12 0.021* GII/III 76.52 3.91 SSEA4−SOX2+GI 0.07 0.05 0.539 GII/III 0.03 0.03 SSEA4+Olig2+GI 49.81 5.42 0.035* GII/III 64.64 4.29 SSEA4+Olig2‑ GI 29.15 3.51 0.324 GII/III 23.97 3.85 SSEA4−Olig2+GI 5.52 2.95 0.405 GII/III 2.91 0.95 Table 1 (con inued) P alues o ANOVA Welch and B own–Fo sy he a e indica ed *P signi ican a he 0.05 le el (2- ailed) **P is signi ican a he 0.01 le el (2- ailed) Page 6 o 14 Alami e al. Cance Cell In (2018) 18:77 7+CD133+Vimen in+FZD9+GFAP+BTIII+SSEA4+O lig2+ (Fig.6c, d). Discussion Collec i ely, meningiomas p esen a unique model o explo ing umo p og ession in CNSTs, as hey encom- pass umo s wi h a a ie y o agg essi eness and g ades. Ou s udy sheds a ligh in o he p o ein exp ession and co-localiza ion o c i ical SC and de elopmen al ma ke s ha a e implica ed in modula ing malignancy. In pa icu- la , we p esen a comp ehensi e di e en ial analysis o he h ee dimensional spa ial dis ibu ion o SC ma ke s insi u, hei co-exp ession, and hei co ela ion in ela- ion o g ade. The ea u es obse ed o indi idual p o eins in he meningioma samples we e consis en wi h hei manu- ac u ing da a and p e ious publica ions in o he issue ypes [42, 57, 67–73]. Ki67-posi i e cells we e clea ly dispe sed, indica ing ha di iding cells we e no pa icu- la ly g ouped oge he . Bo h SOX2 and FZD9 we e less equen and occu ed in niches, which is in conco dan wi h niche-o ganized CSCs. All o he s udied ma ke s had a iable cha ac e is ics ha had ei he niche, he e o-, o homo-exp ession, in a umo -dependen manne . O pa icula in e es is he localiza ion o Olig2. The exclu- sion o his p o ein om he nucleus has been epo ed o be associa ed wi h as ocy e di e en ia ion, while nuclea Olig2 was shown o a ge ch oma in emodele s, p io di e en ia ion in oligodend ocy e p ogeni o s [49, 53, 74]. In his coho , Olig2 was p edominan ly obse ed in he nucleus, a he nuclea en elope, and only occa- sionally in he cy oplasm, hus implying ha meningi- oma cells may beha e like oligodend ocy e p ogeni o s. Howe e , u he de ailed wo k is equi ed o cla i y his * * ** * ** 0100 200300 400500 600700 Jed29_MN Jed13_MN Jed49_MN Jed79_MN Jed45_MN Jed58_MN Jed72_MN Jed70_MN Jed39_MN Jed38_MN Jed61_MN Jed40_MN Jed43_MN Jed64_MN Jed62_MN A e age P o ein Exp ession % Vimen in+ SSEA4+ Olig2+ GFAP+ BIIITubulin+ Nes in+ CD133+ F izzled9+ Sox2+ Ki67+ Jed61_MN Jed40_MN Jed49_MN Jed64_MN Jed79_MN Jed62_MN Jed58_MN Jed13_MN G ade IG ade IG ade II G ade II DAPIFZD9 DAPIGFAP b 100µm a G ade IG ade II/III Fig. 2 The le el o exp ession o he selec ed ma ke s in g ade I and g ade II/III meningioma samples. a The a e age pe cen ages o cells posi i e o each make in g ade I and g ade II/III meningiomas. Signi ican changes a 0.05 a e indica ed by * and a 0.01 a e indica ed by **. b Immuno luo escence images o FZD9 and GFAP in a selec ion o g ade I and g ade II/III meningiomas. DAPI (blue) FZD9 ( ed), GFAP ( ed). Fi e independen egions we e sco ed o each ma ke wi hin a s ained umo sec ion. All images we e aken a ×20 Page 7 o 14 Alami e al. Cance Cell In (2018) 18:77 obse a ion and u u e s udies will need o be comple ed on a la ge scale. No ably, he exp ession o all indi idual p o eins was no dicho omous o g ade. Cells posi i e o all SC ma ke s we e de ec ed in g ade I meningiomas, sug- ges ing ha ei he he es ablishmen o CSC clones occu s ea ly in umo de elopmen , o ha by he ime umo s become clinically e iden , CSCs a e al eady es ablished. Howe e , consis en wi h published da a, a highe numbe o posi i e cells s ained o Ki67 and Vimen in we e de ec ed in g ade II/III compa ed wi h g ade I meningiomas [13, 69]. To he bes o ou knowledge, his s udy is he i s o p esen in si u analysis o he exp ession o SSEA4, OLIG2 and FZD9 in meningiomas. Cells posi i e o SSEA4 and OLIG2 we e mo e equen in g ade II/III meningiomas and he numbe o FZD9-posi i e cells was signi ican ly highe in g ade II/III meningiomas, al hough he o e - all le els emained ela i ely low, implying ha g ow h o FZD9-posi i e cells in meningiomas is es ic ed. Su p isingly, and in con as o o he s udies, mo e cells posi i e o GFAP o BIIIT we e de ec ed in g ade II/III meningiomas [75]. A o m o GFAP ha di e s in he C- e minal domain was de ec ed in he sub- en icula zone (SVZ) o he b ain, sugges ing ha GFAP may no be an exclusi e as ocy ic di e en ia- ion ma ke [56, 57]. Indeed, i is impo an o conside ha o p o eins wi h mul iple o ms, he de ec ion o a p o ein’s exp ession using immunos aining will depend DAPI Vimen inFZD9 Jed29_MN DAPINes inKi67 Jed45_MN DAPICD133Sox2 Jed49_MN DAPISSEA4Sox2 Jed79_MN Nega i e con ol Jed29_MN DAPISSEA4Olig2 Jed72_MN 0100 200300 400500 Jed29_MN Jed13_MN Jed49_MN Jed79_MN Jed45_MN Jed58_MN Jed72_MN Jed70_MN Jed39_MN Jed38_MN Jed61_MN Jed40_MN Jed43_MN Jed64_MN Jed62_MN A e age Exp ession % SSEA4+SOX2- Vimen in+F izzled9- SSEA4+Olig2+ Nes in+Ki67- SSEA4+Olig2- CD133+ Sox2- Vimen in+F izzled9+ SSEA4+SOX2+ CD133+ Sox2+ CD133- Sox2+ SSEA4-Olig2+ Nes in+Ki67+ Nes in-Ki67+ Vimen in-F izzled9+ SSEA4-SOX2+ * * * * * * a b 100µm G ade IG ade II/III Fig. 3 The le el o exp ession o double s ained issues o g ade I and g ade II/III meningioma samples. a The a e age pe cen ages o cells posi i e o co‑s ained ma ke s. Signi ican changes a 0.05 a e indica ed by As e isk. b Rep esen a i e immuno luo escence images o double s ained ma ke s o Ki67 ( ed) wi h Nes in (g een), SOX2 ( ed) wi h CD133 (g een), Vimen in (g een) wi h FZD9 ( ed), SSEA4 (g een) wi h SOX2 ( ed), and SSEA4 (g een) wi h Olig2 (Red), each wi h DAPI (blue). Fi e independen egions we e sco ed o each double ma ke wi hin a s ained umo sec ion. All images we e aken a ×20 Page 8 o 14 Alami e al. Cance Cell In (2018) 18:77 on he u ilized an ibody [76]. Acco ding o he manu- ac u ing in o ma ion shee , he GFAP an ibody used in his wo k was aised agains he ull leng h o a pu i ied na i e p o ein co esponding o human GFAP. Compa ed o p e ious s udies [10, 13, 28, 67, 68, 77, 78], co-s aining o SOX2, CD133 and Nes in ac oss a single sec ion also p o ided a ew unexpec ed obse - a ions. In pa icula , he a e age numbe o cells posi- i e o bo h SOX2 and CD133 was lowe in g ade II/III meningiomas, while cells posi i e o SOX2 and nega i e CD133 inc eased in equency. The inc ease in he la e was pa icula ly no ed in he ecu en umo Jed49_MN. The ac ion o Ki67+ cells ha we e Nes in nega i e we e mo e equen in g ade II/III meningiomas, e en hough Nes in exp ession ended o sligh ly inc ease wi h g ade [28]. Toge he , hese obse a ions may be explained by he CSC clonal e olu ion heo y, whe e o example, cells posi i e o SOX2 and CD133 could occu a ea ly de elopmen and di e ge la e o pa ne wi h o he SC- ela ed genes [79]. In addi ion, hey highligh in i o and insi u di e ences in he exp ession o CSCs ma ke s ha may e lec epigene ic changes, in luenced by he mic oen i onmen . The analysis o a single ma ke h oughou he consec- u i e sec ions along a dep h o 32μm indica ed a s ong co ela ion o exp ession o bo h adjacen and dis al sec- ions o meningioma issues. Basic analysis loca ing CSC niches ac oss consecu i e sec ions has been a emp ed p e iously in b eas cance issues [80, 81]; howe e , no co ela ion o exp ession was s udied. Spea man’s Rho ac o indica ed a highly signi ican co ela ion be ween he exp essions o Vimen in and SSEA4, and he exp es- sions o CD133 and GFAP. The co-exp ession o SSEA4 and Vimen in has been obse ed in mul ipo en mes- enchymal SCs and in pos na al pe iodon al ligamen (PDL)-de i ed SCs (PDLSC) [11, 82]. CD133 and GFAP 0 10 20 30 40 50 60 70 80 90 100 02040608 01 00 Pe cen age o Posi i e Cells in Sec ion 6 (%) Pe cen age o Posi i e Cells in Sec ion 2 (%) 0 10 20 30 40 50 60 70 80 90 100 020406080 100 Pe cen age o Posi i e Cells in Sec ion 7 (%) Pe cen age o Posi i e Cells in Sec ion 6 (%) Jed40_MN Jed38_MNJed61_MN Jed70_MN Sec ion 2Sec ion 6Sec ion 6Sec ion 7 100µm ab DAPISox2 DAPISSEA4 SSEA4Sox2 Fig. 4 The co ela ion o he exp ession o SSEA4 and SOX2 in adjacen and dis al consecu i ely sec ioned immuno luo escence‑p ocessed issues. a Rep esen a i e immuno luo escence images o adjacen sec ions 6 and 7 s ained o SSEA4 (g een), and o dis al sec ions 2 and 6 s ained o SOX2 ( ed). All images we e aken a ×20. b G aphs showing Spea man’s Rho co ela ions be ween posi i e exp ession o SSEA4 in sec ions 6 and 7 o SOX2 in sec ions 2 and 7, o all samples Page 9 o 14 Alami e al. Cance Cell In (2018) 18:77 co-exp ession has been de ec ed in glioneu onal umo s [83], glioblas oma cells [84], and ac i a ed B1 as ocy es [85, 86]. Such co ela ion implica es ac i a ed B1 as o- cy es’ exp ession-like p og am in a leas a ac ion o meningioma cells. Signi ican co ela ions we e also obse ed o he exp essions o SSEA4 wi h CD133 o Nes in, FZD9 wi h Vimen in o SOX2 o Olig2, and SOX2 wi h BIIIT. En ichmen o SSEA4 and CD133-posi i e cells om co d blood ma ked e y small emb yonic-like s em cells (VSELs) ha ha e high elome ase ac i i y and exp ess plu ipo en SC ma ke s OCT4, SSEA4, NANOG, and SOX2 [87]. Simila ly, he co-exp ession o SSEA4 and Nes in has been obse ed in human umbilical co d ma ix-de i ed mesenchymal SCs [88]. The p esence o Nes in-posi i e p oli e a ing cells also co ela es wi h he p esence o Vimen in+FZD9+ cells. Co-exp ession o FZD9 and Nes in has been obse ed in neu al s em p o- geni o , de i ed om pa ien s wi h Williams synd ome, a de elopmen al diso de caused by mu a ions in ch omo- some 7 [89]. The co ela ion o FZD9 wi h SOX2 is pe - haps no su p ising, gi ing ha hey a e bo h pa o he WNT signaling pa hway, a pa hway ha is ac i a ed in some meningiomas [37]. Pe haps mo e su p ising is he co ela ion be ween SOX2 and BIIIT. This combina ion has been implica ed in axane esis ance o pa ien s wi h s age III o a ian epi helial cance [90] and obse ed in GBM cell lines [91]. In e es ingly, he exp ession o Ki67 alone does no co ela e wi h any pa icula ma ke , sug- ges ing ha p oli e a ing cells belong o a he e ogeneous popula ion o clones. Al e na i ely, cells may be exi ing SC-like s a us o di ide. An inc ease in he umo he e ogenei y o CNSTs has long been associa ed wi h agg essi eness, esis ance, and eoccu ence [79, 92–96]. Recen s udies ha e add essed he e ogenei y using no el and challenging app oaches [62, 97]; howe e , e y ew a e documen ed CD133+ Sox2+ SSEA4+ Sox2+ Ma ke Nes in+Ki67+ CD133+ Sox2+ Vimen in+FZD9+ SSEA4+ Olig2+ GFAP+ BIIITubulin+ Nes in+ Ki67+ CD133+ Sox2+ Vimen in+ FZD9+ SSEA4+ Sox2+ SSEA4+ Olig2+ Nes in+1.000 .379 .504 .289 .500 .504 .636*.075 .457 .229 .692** .356 .504 .171 .164 Ki67+ 1.000 .011 .239 .025 .429 .146 .236 .311 .307 .842** .107 .429 .371 .146 CD133+ 1.000 .496 .239 .286 .561*.168 .657** .364 .190 .727** .286 .464 .261 Sox2+ 1.000 .196 .589*.343 .425 .232 .557*.405 .894** .589*.932** .518* Vimen in+1.000.529*.689** .489 .282 .271 .222 .202 .529*.068 .611* FZD9+1.000 .282 .532*.361 .404 .516*.411 1.000** .568*.568* SSEA4+1.000.357.350.489.409.540*.282 .254 .514* Olig2+ 1.000 .429 .396 .258 .463 .532*.496 .957** GFAP+1.000 .350 .344 .356 .361 .279 .396 BIIITubulin+1.000 .444 .574*.404 .532*.443 Nes in+Ki67+ 1.000 .335 .516*.466 .276 CD133+Sox2+ 1.000.411 .847** .556* Vimen in+FZD9+ 1.000.568*.568* SSEA4+Sox2+ 1.000.561* SSEA4+Olig2+ 1.000 Nes in+ CD133+ BIIIT+ GFAP+ Olig2+ Vimen in+ SSEA4+ FZD9+ Sox2+ Nes in+ Ki67+ SSEA4+ Olig2+ Vimen in+ FZD9+ Co ela ion, P<0.05 Weak Co ela ion Co ela ion, P<0.01 a bc Fig. 5 Co ela ion ends be ween he exp essions o di e en ma ke s ac oss consecu i e issues. a A lis showing Spea man’s Rho co ela ion coe icien s. *Co ela ion is signi ican a he 0.05 le el (2‑ ailed). **Co ela ion is signi ican a he 0.01 le el (2‑ ailed). b Illus a ions o he s eng h o co ela ions be ween di e en single ma ke s, and c co‑s ained ma ke s