Details of the Sox2 Reporter Construct Used in This Study. Panelaprovidesthesox2

Supplementalfigurelegends

Supplementalfigure1

Details of the Sox2 reporter construct used in this study. PanelAprovidestheSox2

bindingconsensussequence. PanelBprovidesinformationregardingtheSox2

transcriptional response element(TRE). PanelCindicatesthepresenceofthree

tandem Sox2 TREs included in the reporter construct. TheSox2consensussequences

arehighlightedinboldandunderlied.

Supplementalfigure2

ALK+ALCLcelllines,Karpas299andSUP-M2,showednoevidenceofGFPsignal

detectable by flow cytometry (left panel). Similarly,cellsstablytransducedbythe

negativecontrollentiviralvector(i.e. containing mCMV but not the Sox2 transcriptional

response element) also did not show evidence of GFP expression(right panel).

Supplementalfigure3

Immunohistochemistrywasperformedusing10casesofT-cellmalignantlymphomas.

All5classifiedasperipheralT-celllymphoma,notfurtherclassified,showedno

definitiveSox2nuclearstaining(twocasesareillustratedinAandB,respectively).

NeoplasticcellsfromonecaseofT-celllymphomaarisinginthepost-transplantation

setting also showed strongSox2nuclear staining (illustrated in C and D).

Supplementalfigure4

AandB) The levels of Sox2 mRNA in Karpas 299 cells treated withsiRNAtoknock-

down STAT3 or NPM-ALK were significantly decreased, as compared to cells treated

withscrambledsiRNA(ScrsiRNA). QuantitativeRT-PCRusinganEppendorfRealplex

apparatusandtheQiagenQuantiTectSYBRGreenPCRkitwasemployed. The

efficiencyofSTAT3andNPM-ALKknock-downisillustratedinfigure2. CandD) Two

lymphomacelllines,JurkatandDG75,weretransfectedwiththeNPM-ALK. While

NPM-ALKexpressionandSTAT3activationwasobserved,noSox2expressionwas

detectableinthesecells. Cellstransfectedwiththeemptyvectorsservedasthe

negativecontrols. CelllysatesfromKarpas299servedasthepositivecontrol.

Supplementalfigure5

InadditiontotheexpressionofGFP,isolatedSox2activecellsubsetsinKarpas299(A)

and SUP-M2 (B) expressed significantly higher levels of luciferase than cells transduced

withthenegativecontrollentiviralvector(i.e. mCMVwithouttheSox2transcriptional

responseelement). Theluciferaseenzymaticactivitywasassessedusinga

commerciallyavailablekitfromPromega(Madison,WI,USA).

Supplementalfigure6

To assess if the Sox2 activity is a stable phenotype, we assessed the proportions of

GFP-positivecellsinpurifiedSox2activeandSox2inactivecellsinlong-termtissueculture.

Byflowcytometry,wefoundthattheproportionsofGFP-positivecellsintheSox2active

subsets gradually decreased in both Karpas 299cellsand SUP-M2 cells (right panel). In

contrast,isolatedSox2inactivecellsdidnotshowanydetectableincreaseinthe

proportionsofGFP-positivecells.

Supplementalfigure7

TofurthersupportthattheabsenceofGFPexpressioninSox2inactivecellswasnotdue

toalossofthereporterconstruct,weperformedPCRandfoundthatboththeGFP and

luciferasegenesweredetectableinbothSox2activeandSox2inactivecellsubsets. The

primer sets for this experiment aredescribed in supplemental table 1.

Supplementalfigure8

WeperformedexperimentstodetermineifthedifferenceintheSox2transcriptional

activityinALK+ALCLcellsisduetodifferencesintheSox2expressionornuclear

localization. AsshowninAandB,bywesternblots,theexpressionlevelsofSox2

proteinweresimilarbetweenthetwocellssubsetsinbothcelllines. As shown in C and

D,therewasnoappreciabledifferenceinthesubcellularlocalizationofSox2inbothcell

subsets. AsshowninE,EMSAwasperformed,andwefoundthatprobe/protein

complexeswereidenticalbetweentheSox2activeandinactivecells. TheLightShift

chemiluminescentEMSAKit(Thermoscientific,SouthLogan,USA)wasusedto

performEMSA. Theexperimentalconditionsusedforeachlanearedescribedas

follows: lane 1) free Sox2 probe only; lane 2) Sox2 probe and nuclear extract from SUP-

M2Sox2activecells;lane3)Sox2probeandnuclearextractfromSUP-M2Sox2inactive

cells;lane4)Sox2probeandnuclearextractfromSox2inactiveSUP-M2cellsand100X

non-labeledSox2probe;lane5)Sox2probeandnuclearextractfromSox2inactiveSUP-

M2 cells + 100X non-labeled Sox2mutantprobe.

Supplemental figure 9

ExaminationofpotentialSox2downstreamtargetsinpurifiedSox2activeandSox2inactive

cellsubsetsfromSUP-M2cellsusingquantitativeRT-PCRrevealedthattheexpression

ofnotch 1andPDGFR-alpha were higher in Sox2activecells, whereas BCL2 was higher

in Sox2inactivecells. Validation of these results was performed using Western blots in

bothKarpas299andSUP-M2. Of note, PDGFR-alpha was not expressed in SUP-M2

andacomparisonbetweenthetwocellsubsetsinthiscelllinecouldnotbemade. Anti-

BCL2 and anti-PDGFR-alpha were purchased from Santa Cruz, whereas anti-notch1

waspurchasedfromCellSignaling.

Supplementalfigure10

SerinephosphorylationofSox2inSox2activeandinactivecellsubsetswasexamined. In

order to confirm that the transcriptional activityof Sox2 is not related toadifferencein

serinephosphorylationofSox2,weperformedimmunoprecipitation. No difference in the

level of serine phosphorylation of Sox2 was seen between the two subsets. Asimilar

experimentwasdonetoinvestigatethestatusofSox2tyrosinephosphorylation;no

detectabledifferenceinSox2tyrosinephosphorylationwasdetectablebetweenthetwo

cellsubsets(notshown).