Natural Killer cell subsets in hematological diseases : learning for immunotherapy Dang Nghiem Vo

To cite this version:

Dang Nghiem Vo. Natural Killer cell subsets in hematological diseases : learning for immunotherapy. Human health and pathology. Université Montpellier, 2018. English. ￿NNT : 2018MONTT013￿. ￿tel- 01868212￿

HAL Id: tel-01868212 https://tel.archives-ouvertes.fr/tel-01868212 Submitted on 5 Sep 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés.

THÈSE POUR OBTENIR LE GRADE DE DOCTEUR DE L’UNIVERSITÉ DE M ONTPELLIER

En Biologie Santé

École doctorale CBS2 n°168 - Sciences Chimiques et Biologiques pour la Santé

Unité de recherche INSERM U1183 - Institut de Médecine Régénératrice et de Biothérapie (IRMB)

Natural Killer cell subsets in hematological diseases : learning for immunotherapy

Présentée par Dang Nghiem VO Le 03 Juillet 2018

Sous la direction de Dr. Martin VILLALBA-GONZALEZ

Devant le jury composé de

Prof. Marie-Alix POUL, PU, Institut de Recherche en Cancerologie de Montpellier (IRCM) Président du Jury

Dr. Julian PARDO, DR, Biomedical Research Center of Aragon (CIBA), Zaragoza, Espagne Rapporteur

Dr. Cyril FAURIAT, CR, Centre de Recherche en Cancérologie de Marseille (CRCM) Rapporteur

Dr. Laurent GROS, CR, Institut de Recherche en Cancerologie de Montpellier (IRCM) Examinateur

Dr. Brigitte KERFELEC, CR, Centre de Recherche en Cancérologie de Marseille (CRCM) Examinateur

Dr. Martin VILLALBA-GONZALEZ, DR, Institut de Médecine Régénératrice et de Biothérapie (IRMB) Directeur de thèse

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! "6CFGOD87=8E8FLK! ! ! $GJ=EGKLg) ') OGMD<) DAC=) LG) L@9FC) EQ) L@=KAK) 9GJ) @AK) ;GFLAFMGMK)KMHHGJL)L@JGM?@GML)L@=)LAE=)G>)EQ).@")KLMGJ)HJGNA==<:9;C)O@=F=N=J)F=;;=KK9JQ)>GJ) L@=)HJG?J=KK)G>)EQ)OGJCK)O@AD=)=F;GMJ9?AF?)E=)>GJ):=AF?)9F)AF<=H=F<=FL)J=K=9J;@=Jj) +GJ=GN=Jg)@=)K@GO=<)E=)@GO)LG):=)HJG9J)AF)L@=).@") BGMJF=Q)OAL@GML)@AK)?MA<9F;=j) ) 1=;GFj)+9JA=kFDAP).GMDg) "Jj)*9MJ=FL)%JGKg)"Jj)(MDA9F).9J=D=;g)>GJ)L@=AJ) CAF)EQ).@")OGJCj)5AL@)KH=;A9D)L@9FCK)LG)"J) *9MJ=FL)%JGK)9F<)"Jj)(MDA9F).9JGJ)L@=AJ)N9DM9:D=)GDDGOF?)FGL=g)')OGMD<)DAC=)LG)=PHJ=KK)EQ)?J9LLALM<=)>GJ)9DD)EQ);GDD=9?M=K)9F<) K=FAGJK)AF)GMJ)L=9Eg)'0+ k#IMAH=)Çg)AF;DMGJ) L@=AJ) @=DH>MD) ==<:9;C) :ML) 9DKG) AF) L=;@FA;9D) 9F<) <9ADQ) DA>=) KMHHGJLK) 9JGMF<) L@=) :=F;@) ) EQ) .@") AF) L@=) D9:j) ') 9E) 9DKG) AF<=:L)LG)9DD)L@=)GL@=J);GDD=9?M=K)9L)L@=)AFKLALML=)9F<)D9:E9L=K)>GJ)@=DHAF?)E=)AF)E9FQ) >=J=FL)9KH=;LK)AF)9F<)GMLKA<=)L@=)D9:j) ) $AF9DDQg)')OGMD<)DAC=)LG)K9N=)L@=)D9KL)OGJGJ)EQ)EGE)9F<)<9<)>GJ)HJGNA>A;MDLA=Kj) 2G) EQ) <9GJ) 9DO9QK) :=DA=NAF?) 9F<) =F;GMJ9?AF?) >GJ) EQ) HMJKMAL) G>) J=K=9J;@) ;9J==Jj) FF<) EGEg) >GJ) :=DA=NAF?)AF)E=)J=?9J)EQ)<=;AKAGF)9F<)@=J)=F

! }! "86D4J4L?GF!G) L@AK) L@=KAK) OGJC) O9K) KMHHGJL=<) :Q) 9) >=DDGOK@H) HJGNA<=<) :Q) +AFAKL=J=) <=) Dn#FK=A?F=E=FL)1MH=JA=MJ)=L)<=)D9)0=;@=J;@=)t+#10uj)F)K=;GF<)H9JL)G>)L@AK)OGJC)O9K) CAFMF<=<) :Q) !&3k+GFLH=DDA=J) ;GFLJ9;LK) NA9) J=;GEE=F<9LAGF) >JGE) EQ) L@=KAK) 9MF)L@=)OGJCK)HJ=K=FL=<)AF)L@AK)L@=KAKj) ) ) ) ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ~! *?KL!G87!4JL?6D8K!?F!H88J\J8N?8O87!BGMJF4DK! ) ~j) )JRQOAFKC9) #jg) FDD=F<=k4=?9) ,jg) !GJFADDGF) Fjg) .:# "D(C g) !9QJ=>GMJ;I) *jg) .9F9:A=J=K)!jg)4AD;@=K)!jg)"=;@9F=Lk+=JNADD=)(jg)&A;@=JA)7jg)0GKKA)(j$jg)!9JLJGF)%jg) 4ADD9D:9)+j)'<=FLA>A;9LAGF)G>)FFLAkLMEGJ)!=DDK)!9JJQAF?),9LMJ9D))ADD=J)t,)u)!=DD) FFLA?=FK) AF) .9LA=FLK) 5AL@) &=E9LGDG?A;9D) !9F;=JKj) # >C+:9>8>B: ) Äg) ~ÅÑÇs~ÅÖÑ) tÄ}~Éuj) ) Äj) .:B# "D(C g) FD=PA9) !jg) FDD=F<=k4=?9) ,jg) +GJK;@9MK=J) $jg) &GMGL) 0jg) +=F9J<) !jg) 29JL=) )jg) !9JLJGF) %jg) 4ADD9D:9) +j) ,)) ;=DD) 9;LAN9LAGF) 9F<) J=;GN=JQ) G>) ,)) ;=DD) KM:K=LK)AF)DQEH@GE9)H9LA=FLK)9>L=J)G:AFMLMRME9:)9F<)D=F9DAAAHBC@C) 9DDG?=F=A;) ,)) ;=DDK) OAL@) =>>A;A=FL) 9FLA:G8F j)tÄ}~Üuj)tAF)HJ=KKu) ) Çj) )@9F)Fj)3k&jg)FDD=F<=k4=?9),jg)%AL=F9Q)"jg)%=J:9Dk!@9DGAF)1jg)%GF<=9M)!jg).:# "D(C g) =DC9@D9)1jg)-J=;;@AGFA)1jg)29D9JA;G)%jg) =JLGDAFA)$jg) GRA;)+jg)49D:BG>;>88/:DCEGF j)Ög)~s~É)tÄ}~Öuj) ) Éj) =DC9@D9) 1jg) )@9F) Fj) 3k&jg) %AL=F9Q) "jg) FD=PA9) !jg) %GF<=9M) !jg) .:# "D(C g) -J=;;@AGFA) 1jg) 29D9JA;G) %jg) =JLGDAFA) $jg) !9JLJGF) %jg) &=JF9F<=R) (jg) "9MB9Lk !@9N9FA=M) +jg) FDD=F<=k4=?9) ,jg) 4ADD9D:9) +j) !@9F?=K) AF) E=L9:GDAKE) 9>>=;L) =PHJ=KKAGF) G>) F !) LJ9FKHGJL=JK) L@JGM?@) #0)É) 9F<) <=H=F:BG>;>88/:DCEGF j)tÄ}~Üuj)tAF)HJ=KKu) ! ! ! ! ! !

! Ä! 034=7!@8!!@?D7?D! ! ,4CDB35D!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WY ! .KCE>K!/E4CD3?D;7==7!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W[ ! !&,-0#.!WP!%#+#.,)!'+0.,"1!0',+!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!XX ! WPW ! %7?7B3=!@F7BF;7G!34@ED!D:7!;>>E?7!CICD7>!3?6!;DC!5@>A@?7?D!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!XY ! WPX ! +(!57==!@?D@97?I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!XZ ! WPY ! +(!57==!76E53D;@?!3?6!D@=7B3?57!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!X_ ! WPZ ! +(!57==!B757AD@B!B7A7BD@;B7!3?6!C;9?3=;?9!A3D:G3IC!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!YY ! WP[ ! &@G!6@7C!3!+(!57==!<;==!;DC!D3B97DR!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!ZX ! WP\ ! &E>3?!+(!57==!3?6!D:7!;>>E?7!5@?DB@=!@8!F;B3=!;?875D;@?!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!ZZ ! WP] ! 0E>@B!;>>E?@CEBF7;==3?57!8B@>!D:7!+(!57==!A7BCA75D;F7O!DE>@B!;>>E?7!7F3C;@?! CDB3D79;7C!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!Z[ ! WP^ ! !2#1(1, !7HA3?C;@?!@8!3==@97?7;5!+(!57==!8@B!53?57B!;>>E?@D:7B3AI!PPPPPPPPPPPPPPPPPPPPPPPPPP!Z^ ! WP_ ! !"Z[!7HAB7CC;@?!3?6!3=D7B?3D;F7!CA=;5;?9!;C@8@B>C!@?!+(!57==!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP![W ! WPWV ! ,;>C!3?6!CDBE5DEB7!@8!D:7!D:7C;C!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP![Z ! !&,-0#.!XP!*,0#.',)/!,+"!*#0&,"/!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP![^ ! XPW!%,)#+!CDE6IO!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP![_! XPX!,*)!CDE6IO!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!\W ! XPY!&'2!CDE6I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!\X ! XPZ! !+#1(/., !+(!57==!7HA3?C;@?!8B@>!:73=D:I!6@?@B!1! !PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!\] ! !&,-0#.!YP!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!\_ ! '67?D;8;53D;@?!@8!3?D;QDE>@B!57==C!53BBI;?9!+3DEB3=!(;==7B!T+(U!57==!3?D;97?C!;?! A3D;7?DC!G;D:!:7>3D@=@9;53=!53?57BC!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!\_ ! !&,-0#.!ZP!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!^Z ! +(!57==!35D;F3D;@?!3?6!B75@F7BI!@8!+(!57==!CE4C7DC!;?!=I>A:@>3!A3D;7?DC!38D7B! @4;?EDEJE>34!3?6!=7?3=;6@>;67!DB73D>7?D!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!^Z ! !&,-0#.![P!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!_] ! /DE6I!@8!+(!57==!A@AE=3D;@?C!;?!,5ED7!*I7=@;6!)7E<7>;3!T,*)U!A3D;7?DC!6EB;?9! 5:7>@D:7B3AI!DB73D>7?D!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!_] ! !&,-0#.!\P!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WWV! ";CD;?5D!A:7?@DIA;5!;>AB;?DC!;?!+(!57==!8B@>!A3D;7?DC!G;D:!&E>3?! '>>E?@678;5;7?D!2;BEC!T&'2U!;?875D;@?!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WWV! !&,-0#.!]P!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WXY! #HA3?C;@?!@8!3==@97?7;5!+(!57==C!G;D:!788;5;7?D!3?D;4@6IQ67A7?67?D!57==! 5ID@D@H;5;DI!393;?CD!>E=D;A=7!DE>@BC!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WXY! !&,-0#.!^P!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WY_! 0:7!-"(W!'?:;4;D@B!";5:=@B@357D3D7!!@?DB@=C!!:@=7CD7B@=!&@>7@CD3C;C!0:B@E9:! D:7!#.([S*#$X!-3D:G3I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!WY_! !&,-0#.!_P!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W[\! !:3?97C!;?!>7D34@=;C>!38875D!7HAB7CC;@?!@8!, !!DB3?CA@BD7BC!D:B@E9:!#.([!3?6! 67A7?6;?9!@?!A[Y!CD3DEC!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W[\!

! Å! !&,-0#.!WVP!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W]Z! *;D@5:@?6B;3=!!@>A=7H!'!35D;F;DI!C;9?3=C!3?D;@H;63?D!B7CA@?C7!D:B@E9:!#.([!PPP!W]Z! !&,-0#.!WWP!%#+#.,)!"'/!1//',+!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_V! ^PW!%,)#+!CDE6I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_W ! ^PX!,*)!CDE6I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_Y ! ^PY!&'2!CDE6I!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_Z ! ^PZ! !+#1(/., !+(!57==!7HA3?C;@?!3?6!;>A35D!@8!>@6E=3D;?9!53?57B!57==!>7D34@=;C>!;?!+(Q 43C76!;>>E?@D:7B3AI!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_[ ! ^P[!/E>>3BI!L!8;?3=!5@?5=EC;@?C!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_] ! ^P\!$EDEB7!A7BCA75D;F7C!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!W_^ ! .787B7?57!PPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPP!XVV! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! Ç! );CD!@8!8;9EB7C! ! $'%1.#!WPW )1;@=E9LA;)EG<=D)G>)@=E9LGHGA=KAK)9F<)DAF=9?=);GEEALE=FLj ! }Ä ! ! $'%1.#!WPX )2OG)E9AF)HGHMD9LAGFK)G>),));=DDK)AF);AJ;MD9LAF?)H=JAH@=J9D):DGG<)OAL@)L@=AJ)>=J=F;=)AF) !"ÉÑ)9F<)!"~Ñ)=PHJ=KKAGFj ! }Å ! ! $'%1.#!WPY !.A>GHMRIBK>G<>L!;>MP>>G!!"ÅÇ;K!:G=!!"ÅÇ=BF!,)!<>EE!IHINE:MBHGLi! }É ! ! $'%1.#!WPZ !1NFF:KBS>=!LM:@>L!H?!,)!<>EE!=>O>EHIF>GMLi! }Ñ ! ! $'%1.#!WP[ !2A>!FBLLBG@jL>E?!ARIHMA>LBL!BG!,)!<>EEi! }Ö ! ! $'%1.#!WP\ !!"B??>K>GM!FH=>EL!H?!,)!<>EE!>=N<:MBHGi! ~| ! ! $'%1.#!WP] !!'2'+jF>=B:M>=!BGAB;BMHKR!LB@G:EBG@!I:MAP:Ri! ~Ç ! ! $'%1.#!WP^ !!,)!<>EE!:=!K><>IMHKLi! ~Ñ ! ! $'%1.#!WP_ !,)!<>EE!:Qi! ~Ö ! ! $'%1.#!WPWV !!!,)!<>EE!<:G!;>!:=!HK!BGAB;BM>=!OB:!MA>!!"}ÄÄ!K><>IMHK!=>I>G=BG@!HG!=j:=:IM>K! FHE>Li! Ä{ ! ! $'%1.#!WPWW !!!1NFF:KR!H?!,)!<>EEjF>=B:M>=!!>QH:MA!K><>IMHKo=>:MA!EB@:G=!I:MAP:RLi! Ä~ ! ! $'%1.#!WPWX!! !2NFHK!<>EEL!<:G!=>O>EHI!O:KBHNL!BFFNG>!>O:LBHG!LMK:M>@B>L!MH!:OHB=!,)!<>EE!F>=B:M>=! :MM:G>BEE!:=HIMBO>!MK:GL?>K!<:G!;>!:O>=!;R!=B??>K>GM!:IIKH:L! H?!>Q!OBOH!>QI:GLBHGo:@:BG!MA>!:GMBjMNFHK!IHM>GMB:E!H?!,)!<>EE!p;:L>=!BFFNGHMA>K:IRi ! ÄÖ ! ! $'%1.#!WPWZ!! !"B!KHE>!BG!!q1$)r!:KG:MBO>!LIEBK>GM!BLH?HKF!>QIK>LLBHG!HG!ERFIAHLi! Å~ ! ! $'%1.#![PW!!"B:@K:F!H?!@:MBG@!LMK:M>@R!?HK!,)!<>EE!:G=!HMA>K!ERFIAHL!IHINE:MBHGL!BG!L+*i! |{| ! ! $'%1.#![PX !!!.KHIHKMBHG!H?!ERFIAH!:G=!FHGH!IHINE:MBHGLi! |{} ! ! $'%1.#![PY !!,)!<>EE!:KLi!! |{~ ! ! $'%1.#![PZ!!,)!<>EE!>Q!OBOH!=>@K:GNE:MBHG!:G=!MKH@HEE!E:M>L!PBMA!!"~~á!"|Äá!FHGH!IHINE:MBHG!;NM! GHM!!"~~á!"|Äj!FHGH!IHINE:MBHGi! r!IKHIHKMBHG!H?!!"ÄÅ0L0-á!,)!<>EE!E:M>L!PBMA! !"~~á!"|Äá!MKH@H=!,)!<>EEi! |{Ç ! ! $'%1.#![P]!! !!"ÄÅ0L0-á!LMK:MB?B<:MBHGh!:G:ERLBL!H?!O:KBHNL!LNK?:<>!F:KD>KL!;:L>=!HG!L!BG! !"ÄÅ0L0-á!,)!<>EEL!I>K<>GM:@>!:?M>K!MK>:F>GMi! |{É ! ! $'%1.#!\PW!!%:MBG@!LMK:M>@R!?HK!,)!<>EE!:G=!HMA>K!ERFIAHL!IHINE:MBHGL!BG!A>:EMAR!=HGHKL!:G=!&'4á! I:MB>GMLi! ||Å ! !

! É! $'%1.#!\PX!!%>G>K:E!IKHIHKMBHG!H?!ERFIAH!IHINE:MBHGL!BG!&"!:G=!&'4á!I:MB>GMLi! ||Ç ! ! $'%1.#!\PY!! !0>=N<>=!,)%}"!>QIK>LLBHG!BG!,)!<>EEL!BG!BG?>=!I:MB>GMLi! ||Ç ! ! $'%1.#!\PZ !!,)!LN;IHINE:MBHG!:G:ERLBL!HG!&"!:G=!&'4á!I:MB>GMLi! ||É ! ! $'%1.#!\P[!! !0>=N<>=!!"|Ç!>QIK>LLBHG!BG!!"ÄÅ0L=BF!:G=!!"ÄÅ0-á!,)!LN;IHINE:MBHGL!BG!&'4á!I:MB>GMLi ! ||Ö ! ! $'%1.#!\P\!!2K:GL?>K!H?!&'4!K><>IMHK!!"Ä!:G=!!!0Å!?KHF!M:K@>M!<>EE!MH!,)!<>EE!;R!MKH@HEE!F:CHK!:<>IMHKL!:G=!MA>!EB@:G=L!MA:M!MA>R! K>i! ~Å ! ! 0, )#!XP!! +NEMBI:K:F>M>K!?EHP!MKR!I:G>E!?HK!L+*!L:FIE>L!HG! "!$HKM>LL:!BGLMKNF>GMi! Ç~ ! ! 0, )#!YP! !+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!?HK!&'4!L:FIE>L!HG! >GMi! ÇÄ ! ! 0, )#!ZP !+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!q>QM>G=>=r!?HK!&'4!L:FIE>L!HG! "!$HKM>LL:!BGLMKNF>GMiÇÅ ! ! 0, )#![P !+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!q>QM>G=>=r!?HK!&'4!L:FIE>L!HG! >GMi! ÇÇ ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! Ñ! );CD!@8!35B@?I>C! ! ! L"!!! :GMB;H=Rj=>I>G=>GM!<>EEjF>=B:M>=!!ERFIAH!!FR>EHB=!E>ND>FB:! ! L.j|! :BG!|! ! !0! j<>EE!K><>IMHK! ! j,&*!! ! j<>EE!GHGj&H=@DBG!ERFIAHF:! ! !*.! GBMHK! ! !+.! EHB=!IKH@>GBMHK! ! !2! FHMA>K:IR! ! !2*! ! ! "L.j|{o|}! ",L6!:BG!|{o|}! ! "!! =>G=KBMBEE! ! "!L! =BM:M>! ! "* !*!! ! =B??NL>!E:K@>! j<>EE!ERFIAHF:! ! ",L+j|! ",L6!:<<>LLHKR!FHE>!|! ! #L2j}! #PBG@lL!L:K=!MK:GLBGj;:KK!OBKNL! ! #0)j|o}oÅ! >QMK:<>EENE:K!LB@G:EjK>@NE:M>=!DBG:L>!|o}oÅ! ! $L!1! ?ENHK>L<>G<>j:=!<>EE!LHKMBG@! ! $:L*! $:L!EB@:G=! ! %K;j}! @KHPMA!?:<>IMHKj;HNG=!IKHM>BG!}! ! &!+4!! ! ANF:G!@:EHOBKNL! ! &"! A>:EMAR!=HGHK! !

! Ö! A#1!! ANF:G!>F;KRHGBF!<>EE! ! &*L! ANF:G!E>NDH!:GMB@>G! ! &1!! A>F:MHIHB>MBF!<>EE! ! &1!2! A>F:MHIHB>MBF!<>EE!MK:GLIE:M:MBHG! ! '!L+j|! BGM>K<>EENE:K!:=A>LBHG!FHE>!|! ! '$,w! BGM>K?>KHG!@:FF:! ! '@%! BFFNGH@EH;NEBG!%! ! '+B"! BFFNGHFH=NE:MHKR!BFB=>!=KN@! ! B.1!! BG=N<>=jIENKBIHM>GM!LM>F!<>EE! ! '21+! BFFNGHK><>IMHK!MRKHLBG>j;:L>=!LPBM<>IMHK!MRKHLBG>j;:L>=!:<>IMHK!MRKHLBG>j;:L>=!BGAB;BMHKR!FHMB?! ! )'0! DBEE>Kj<>EE!BFFNGH@EH;NEBGjEBD>!K><>IMHK! ! )*0%|!! ! DBEE>K!<>EE!E>!K><>IMHK!LN;?:FBER!%!F>F;>K!|! ! *L)! ERFIAHDBG>j:=!DBEE>K!<>EE! ! *L+.j|! ERLHLHF:Ej:LLH=!F>F;K:G>!IKHM>BG!|! ! *L2! EBGD>K!?HK!:EE! ! *#,! E>G:EB=HFB=>! ! *$Lj|! ERFIAH!?NG=!:GMB@>G!|! ! FL;! FHGH!<>EE!ERFIAHF:! ! +!*j|!! ! FR>EHB=!<>EE!E>ND>FB:!|! ! +!+4!! ! FNKBG>!@:EHOBKNL! ! +"1!! FR>EHB=j=>KBO>=!LNIIK>LLHK!<>EE! ! +#)! FBMH@>Gj:=!IKHM>BG!DBG:L>!DBG:L>!

! |{! ! +&!j'o''! F:CHK!ABLMHQ!'o''! ! +'!jLo ! +&!!IMB=>jK>E:M>=!L>JN>G<>!Lo ! ! ,!L+j|! G>NK:E!<>EE!:=A>LBHG!FHE>!|! ! ,!0! G:MNK:E!<>IMHK! ! ,$jx ! GN:K!?::K!?:=!2!<>EE! ! ,)! G:MNK:E!DBEE>K!<>EE! ! ,)"'! G:MNK:E!DBEE>K!=>O>EHIF>GM!BGM>KF>=B:M>! ! ,)%}jLo!o"! G:MNK:E!DBEE>K!@KHNI!}!F>F;>K!Lo!o"! ! - 8! H;BGNMNSNF:;! ! -6.&-1! HQB=:MBO>!IAHLIAHKRE:MBHG! ! .L+.! I:MAH@>Gj:LLH=!FHE>KG! ! . +!! I>KBIA>K:E!;EHH=!FHGHGN:K!<>EE! ! ."j|! IKH@K:FF>=!<>EE!=>:MA!IKHM>BG!|! ! .'~)! IAHLIAHBGHLBMB=>!~jDBG:L>! ! .*!w! IAHLIAHEBI:L>!!!@:FF:! ! .00! I:MM>KG!K><>IMHK! ! .2.:L>! ! MRKHLBG>jLI>BG!IAHLIA:M:L>! ! .40! IHEBHOBKNL!K><>IMHK! ! 1L.! L>KNF!:FREHB=!IKHM>BG! ! ! ! 1&'.! 1&}!=HF:BGj! ! 1&.j|o}! Src homology region 2 domain-containing phosphatase-1o}! ! LB0,L! LF:EE!BGM>K?>KBG@!0,L! ! 1*L+! LB@G:EBG@!ERFIAH!! !

! ||! 1*2! L>! ! 1RD! LIE>>G!MRKHLBG>!DBG:L>! ! 2L+! MNFHK!:LLH=!F:! ! 2!0! 2!<>EE!K><>IMHK! ! 2'%'2! 2!<>E!BFFNGHK><>IMHK!PBMA!'@!:G=!'2'+!=HF:BGL! ! 2'+j~! 2!<>EE!BFFNGH@EH;NEBG!FNE:M>=!:IHIMHLBLjBG=N@! K>@NE:MHKR!2!<>EE! ! 3! ! NF;BEB<:E!BG! ! 4:Oj|! @N:GBG>!GNHMB=>!>Q!?:M:j=!IKHM>BG!DBG:L>!É{! ! ! !

!

!

!

! ! ! ! ! ! !

! |}! "5KLJ46L! ) ) ,9LMJ9D) )ADD=J) t,)u) ;=DDK) 9J=) AFF9L=) ;QLGLGPA;) DQEH@G;QL=K) L@9L) HD9Q) 9F) AEHGJL9FL) JGD=) AF) AEEMF=) ;GFLJGD) G>) LMEGJ) ;=DD) >GJE9LAGF) 9F<) NAJMK) AF>=;LAGFj) 'F) @=9DL@Q) H=GHD=g) ,)) ;=DDK) J=HJ=K=FLK) @=L=JG?=F=GMK) HGHMD9LAGFK) <=>AF=<) :Q) >=J=FL) H@=FGLQHA;9D)E9JC=JK)9F<)H=J>GJEAF?)KH=;A>A;)>MF;LAGFKj),));=DDK)>JGE)H9LA=FLK)OAL@) F=GHD9KLA;) E9DA?F9F;A=K) 9F<) NAJ9D) AF>=;LAGF) 9J=) @GO=N=J) LQHA;9DDQ) JGE) @=9DL@Q)H=GHD=):Q)L@=)9HH=9J9F;=)G>),));=DD)KM:K=LKg)O@A;@)9J=)>=J=FLA9L=<):Q)L@=AJ) !"ÇÉ)AKG>GJE)HJG>AD=j)!"ÇÉ)AK)9);GEEGFkD=MCG;QL=)LQJGKAF=)H@GKH@9L9K=)9:MF<9FLDQ) =PHJ=KK=<) GF) 9DD) FM;D=9L=<) @=E9LGHGA=LA;) AEEMF=) ;=DDKj) FDL=JF9LAN=) KHDA;AF?) N9JA9FL) J=KMDL=<) AF) ?=F=J9LAGF) G>) L@=) DGF?kAKG>GJE) !"ÇÉ0F) 9F<) L@=) K@GJLkAKG>GJE) !"ÇÉ0-g) O@A;@) =PHJ=KK) >=J=FLDQ) GF) F9\N=) 9F<) =>>=;LGJrE=EGJQ) 2) ;=DDKj) #PHJ=KKAGF) G>) !"ÇÉ) AKG>GJEK)GF),));=DDK)AK)D9J?=DQ)MFCFGOFj)5=)@9N=)HJ=NAGMKDQ)K@GOF)L@9L)>=J=FLA9D) =PHJ=KKAGF) G>) !"ÇÉ0F) 9F<) !"ÇÉ0-) AKG>GJEK) A<=FLA>A=<) KH=;A>A;) ,)) ;=DD) KM:K=LK) AF) @=E9LGDG?A;9D) JGE)H9LA=FLK) LJ=9L=<)OAL@)*=F9DA) LMEGJ);=DDK)t!@9L=J)Çuj)5=)G:K=JN=<)L@=)K9E=)AF)F+*)H9LA=FLK)9>L=J);@=EGL@=J9HQj)'F) L@AK);9K=)L@=)!"ÇÉ0F0-ã),));=DD)KM:K=L)KLJGF?DQ);GJJ=D9L=K)OAL@)LJG?G;QLGKAK)G>)L@=) EGFG;QL=rE9;JGH@9?=) E9JC=J) !"~Ç) t!@9HL=J) Éuj) 'EEMFGH@=FGLQHAF?) G>) ,)) ;=DDK) >JGE)&'4kAF>=;L=<)H9LA=FLK)J=N=9D=<)L@=)HJ=K=F;=)G>)!"ÇÉ0FJGE)@=E9LGDG?A;9D);9F;=JK)9F<)&'4)AF>=;LAGF)MF;LAGF9D)@9DDE9JCK) 9F<) 9F9DQRAF?) !"ÇÉ) AKG>GJE) HJG>AD=) AF) L@=K=) H9L@GDG?A;9D) ;GF>A;A=FL)E=L@G<)>GJ)=PH9FKAGF)G>)@A?@DQ)9;LAN9L=<),));=DDK)>JGE)ME:ADA;9D) ;GJ<):DGG<)t3! u) >B8 I>GEC j)2@=K=),));=DDK)HJGN=)KM:KL9FLA9D)9FLA:G

! |~! :9K=<)AEEMFGL@=J9HA=Kj) ) )=QOGJ

! |Ä! /:KME:!0M5KL4FL?8DD8! #

*=K) ;=DDMD=K) ,9LMJ9D) )ADD=J) t,)u) KGFL) <=K) DQEH@G;QL=K) ;QLGLGPAIM=K) AFFWK) IMA) BGM=FL) MF) J`D=) AEHGJL9FL) <9FK) D=) ;GFLJ`D=) AEEMFAL9AJ=) <=) D9) >GJE9LAGF) <=) ;=DDMD=K) LMEGJ9D=K)=L)<=)D|AF>=;LAGF)NAJ9D=j)!@=R)D=K)H=JKGFF=K)=F):GFF=)K9FLWg)D=K);=DDMD=K),)) J=HJWK=FL=FL) <=K) KGMKk=FK=E:D=K) @WLWJG?VF=K) AFAK) H9J) >WJ=FLK) E9JIM=MJK) H@WFGLQHAIM=K)=L)=PW;ML9FL)HDMKA=MJK)>GF;LAGFK)KHW;A>AIM=Kj)*=K);=DDMD=K),))HJGN=F9FL) <=) H9LA=FLK) 9LL=AFLK) <=) LME=MJK) E9DA?F=K) FWGHD9KAIM=K) =L) <|AF>=;LAGF) NAJ9D=) KGFL) ;=H=F<9FL)AD) ;GEHGKW) <=) DnAKG>GJE=) !"ÇÉj) !"ÇÉ)=KL)D9)LQJGKAF=)H@GKH@9L9K=)S)D=M;G;QL=K);GEEMFWE=FL)=PHJAEW=)H9JEA)LGML=K) D=K) HGHMD9LAGFK) <=) ;=DDMD=K) AEEMFAL9AJ=K) @WE9LGHG\WLAIM=K) FM;DWW=Kj) 3F) N9JA9FL) <|WHAKK9?=) 9DL=JF9LA>) 9) =FLJ9[FW) D9) ?WFWJ9LAGF) <=) D|AKG>GJE=) !"ÇÉ0F) DGF?M=) =L) <=) D|AKG>GJE=);GMJL=)!"ÇÉ0-)IMA)K|=PHJAE=FL)>WJ=EE=FL)<9FK)D=K);=DDMD=K)2)F9\N=K)=L) =>>=;LJA;=KrEWEGAJ=Kj) !=H=F<9FLg) D|=PHJ=KKAGF) <=K) AKG>GJE=K) <=) ,)) !"ÇÉ) =KL) D9J?=E=FL) AF;GFFM=j) ,GMK) 9NGFK) HJW;W<=EE=FL) EGFLJW) IM=) D9) ;Gk=PHJ=KKAGF) <=K) AKG>GJE=K) !"ÇÉ0F) =L) !"ÇÉ0-) A<=FLA>A9AL) <=K) KGMKk=FK=E:D=K) <=) ;=DDMD=K) ,)) KHW;A>AIM=K)<9FK)D=K)E9D9<3# J uj)3F=)IM=KLAGF)J=KL=)>DGM=i) ;GEE=FL) ;=K) ;=DDMD=K) !"ÇÉ0F0-ã) ,)) ;@9F?=FLk=DD=K) DGJKIM=) D=MJK) ;=DDMD=K) ;A:D=K) GJE=K) !"ÇÉ) <9FK) >WJ=FLK) LQH=K) <=) LME=MJK) E9DA?F=K) @WE9LGDG?AIM=K)t !5/;6><3# JDLu)=L)D|AF>=;LAGF)NAJ9D=)t4'&)k)!5/;6><3# M uj)*9)<=MPAVE=) H9JLA=) <=) ;=LL=) L@VK=) 9) >GMJFA) MF=) EWL@G<=) =>>A;9;=) <=) ?WFWJ9LAGF) <=) ;=DDMD=K) ,)) 9DDG?WFAIM=K) 9N=;) MF) HGL=FLA=D) @9ML=E=FL) 9;LANW) HGMJ) D|AEEMFGL@WJ9HA=) 9FLA;9F;WJ=MK=j) !=K) ;=DDMD=K) ,)) =PH9FKW=K) H=MN=FL) XLJ=) MLADAKW=K) =F) L9FL) IMn9?=FL) MFAIM=) =F) EGFGL@WJ9HA=) GM) =F) ;GE:AF9AKGF) 9N=;) <=K) 9FLA;GJHK) EGFG;DGF9MP) ;DAFAIM=E=FL) 9HHJGMNWK) GM) S) <=K) EWA=FL)D|AEEMFG?WFA;ALW)<=K);=DDMD=K)LMEGJ9D=K)HGMJ)J=F>GJ;=J)D|=>>=L)9FLAkLMEGJ9D) ,)k

! |Å! ! !'".(1/#!gY!"6L?N4L?GF!78K!68DDMD8K!,)!8L!J:6MH:J4L?GF!78K!KGMK\HGHMD4L?GFK!78!68DDMD8K!,)! 6>8R! D8K! H4L?8FLK! 4LL8?FLK! 78! DQEH>GE8! 4HJ9K! LJ4?L8E8FL! H4J! G5?FMLMRME45! 8L! D:F4D?7GE?78! ! *=) 15/;6><3#K )9)=P9EAFW)D=K)JWKMDL9LK)<=)D|9F9DQK=)<=K)KGMKk=FK=E:D=K)<=);=DDMD=K) ,))>=LK) <=) D|G:AFMLMRME9:) t- 8u) 9KKG;AW) 9M) DWF9DAJ9;L9AJ=K) =L) 9?J=KKA>K) t"* !*) =L) +!*uj) *n- 8)EGFLJ=)MF=);QLGLGPA;ALW);=DDMD9AJ=)GJL=)tF"!!u) H9J) J9HHGJL) 9M) JALMPAE9:) =L) MF=) 9;LANALW) ;DAFAIM=) 9EWDAGJW=) HGMJ) D=) LJ9AL=E=FL) <=) ;=JL9AF=K) FWGHD9KA=K) !"Ä}ãj) !=H=F<9FLg) D|=>>A;9;ALW) <=) D|9FLA;GJHK) EGFG;DGF9D) tEF:u) =F)EGFGL@WJ9HA=)=KL)DAEALW=j)*=K);=DDMD=K),9LMJ9D))ADD=J)t,)u)KGFL)<=K)EWGF;LAGFF=DD=Kj) *=) EWGF;LAGFK);QLGLGPAIM=K)WHMAKW=K)<=K);=DDMD=K),)j) .=F<9FL) D=) LJ9AL=E=FLg) FGMK) 9NGFK) 9F9DQKW) <=K) 9KH=;LK) KHW;A>AIM=K) <=) D9) :AGDG?A=) <=K) ;=DDMD=K) ,)j) *=) LJ9AL=E=FL) 9) AFN=JKW) D=) H@WFGLQH=) ,)) AEE9LMJ=) <=K) H9LA=FLK) =L) 9M?E=FLW) D|=PHJ=KKAGF) <=K) JW;=HL=MJK) <|9;LAN9LAGF) <=K) ,)j) *=K) JW;=HL=MJK) AF@A:AL=MJK) WL9A=FL)KGAL)AF;@9F?WKg)KGAL)GJL=)JWHGFK=)<=K),))S)D9)>AF)AF) GJL=E=FLj)!=H=F<9FLg)H9J;=)IM=) LGMK) D=K) H9LA=FLK) 9N9A=FL) J=UM) >=LK)

! |Ç! !'".(1/#!hY!$LM78!78K!HGHMD4L?GFK!78!68DDMD8K!,)!6>8R!78K!H4L?8FLK!4LL8?FLK!78!D8M6:E?8! EQ:DG@78!4?=M;!]*"+^!4M!6GMJK!7_MF8!6>?E?GL>:J4H?8 ! ! *>L! I:MB>GML! :MM>BGML! =>! E>N! FRXEH]=>! :B@NZ! q*L+r! LHGM! MRIBJN>F>GM! <:K:!:<M!NG>!IKHEB?XK:MBHG!K:IB=>!=>!IKXNKL!FRXEH]=>L! AXF:MHIH]XMBJN>L! BFF:MNK>L! :II>EXL! FRXEH;E:LM>Li! *>! MK:BM>F>GM! LM:G=:K=! IHNK! <>L! F:EB@GBMXL! >LM! @XGXK:E>F>GM! NG>! ! :O>L! :GMAK:L! >M! =>! E:! !>G!MK:BM>F>GM!=>!IK>FBWK>!BGM>GMBHGf!LHNO>GM!LNBOB>!=zNG>!@K>??>!=>!<>EENE>L! LHNL! AXF:MHIH]XMBJN>L! q2!1&r! S! =>L! I:MB>GML! >G! K>i! !>I>G=:GMf! E:! 2!1&! Gz>LM! I:L! MHNCHNKL! KX:EBL:;E>i! *>L! <>EENE>L! ,)! =>L! I:MB>GML! :MM>BGML! =>! *L+! IKXL>GM>GM! @XGXK:E>F>GM!=>L!=RL?HGF>GML!LXOWK>L!q!HLM>EEH!>M!:Ei!}{{Äg!$:K:@!:G=!!:EB@BNKB! }{{Çri!.HNK!FB>NQ!G=K>!E>!IAXGHMRI>!=>L!<>EENE>L!,)!S!E>L!I:MB>GML!:MM>BGML! =>!*L+!:IKWL!NG!MK:BM>F>GM!=>!!q!2rf!GHNL!:OHGL!:G:ERLX!E>L!IAXGHMRI>L! =>!<>EENE>L!,)!S!É!I:MB>GML!T!}!M>FILh!:N!=B:@GHLMBM!~{!CHNKL!:IKWL!E:!!2i!~{!CHNKL! :IKWL!!2f!E>!LN!>LM!I:KMB>EE>F>GM!K>LM:NKXi!!>LM!!?HKM>! =BFBGNMBHG!=>L!FRXEH;E:LM>L!!"~~á!"|Äj!BFF:MNK>Li!$:BM!BGMXK>LL:GMf!E:!<>EENE>!,)!=N! I:MB>GM!=>!GHMK>!XMN=>!:!FHGMKX!NG>!IKHIHKMBHG!:<!=>!E:!IHINE:MBHG!!"Ç};KB@AM!,)! !"Ç}*! MHNM! >G! KX=NBL:GM! E>L! <>EENE>L! >QIKBF:GM! !"ÅÉi! !>! NG! IAXGHMRI>! BFF:MNK>! :IKWL! !2i! ">! IENLf! GHNL! :OHGL! X@:E>F>GM! LB@G:EX! NG>! =BFBGNMBHG! =>! E:! =X@K:GNE:MBHG!>Q!OBOH!>M!NG>!;:BLL>!=N!GBO>:N!=>!."j|!=:GL!Ez>GL>F;E>!=N!GM! =>L!<>EENE>L!,)f!LN@@XK:GM!NG!IAXGHMRI>!FHBGL!:!E:!! =>L!<>EENE>L!MNFHK:E>Li!#G?BGf!GHNL!:OHGL!!E>L!<>EENE>L!,)!!"ÄÅ0L0-á!LHGM! >G!!MNFHK:E>!!=:GL!GHMK>!XMN=>!IKXGM>!q !:3A;DB7! Zri!!NKB>NL>F>GMf!!"|Äf!F:BL!I:L!!"~~f!Lz:LLH!?HKM>F>GM!T!E:!JN:GMBMX!=>!<>EENE>L!,)! !"ÄÅ0L0-ái! 9M:GM! =HGGX! JN>! !"|Ä! >LM! NG! F:KJN>NK! LIX! =>L! IHINE:MBHGL! =>! FHGHLoF:Lf! GM! E>L! <>EENE>L! ,)! !"ÄÅ0L0-á! BGM>K:@BLL>GM! IKX?XK>GMB>EE>F>GM!:O>L!IHINE:MBHGL!F:BL!I:L!E>L!IKXNKL!FRXEH]=>L!>QIKBF:GM! !"~~!K>LM>!T!>QIEHK>Ki! ! !'".(1/#!iY!#EHJ8?FL8K!H>:FGLQH?IM8K!7?KL?F6L8K!74FK!D8K!68DDMD8K!,)!78!H4L?8FLK!?F<86L:K! H4J!D8!N?JMK!?EEMFG7:ME4?F!]3('^ ! !

! |É! *>! K_E>! =>L! <>EENE>L! ,)! =:GL! EzBG?>! 4'&! :! XMX! ;B>G! XM:;EBi! ">L! XMN=>L! KX<>GM>L! LN@@WK>GM! JN>! EzBG?>! 4'&! <:NL>! NG>! AHFXHLM:LB>! =>L! <>EENE>L! ,)! :MMXGNX>!>M!NG!KXI>KMHBK>!:;>KK:GM!=>!E>NKL!KX<>IM>NKL!q+BDNE:D!>M!:Ei!}{|Ég!1K! }{|Çri! *>L! <>EENE>L! ,)! =>! I:MB>GML! >QIHLXL! T! EHG@! M>KF>! T! E:! OBKXFB>! 4'&! IKXL>GM>GM! NG! IAXGHMRI>! =BLMBG! =BFBGNMBHG! =N! M:NQ! =>! !"ÅÇ! M:G=BL! JN>! E>! !"|Ç!BGM:@!"|Çári!!>L!<>EENE>L!,)!!"ÅÇG>@!=RL?HGEE>L!OHB>GM! E>NK!!I:K!K:IIHKM!T!E>NKL!AHFHEH@N>L!!"ÅÇ=BF!"|Çái!">!IENLf! BEL!LHGM!X@:E>F>GM!=X?BGML!=:GL!E:!IKH=N!Lf!M>EL!JN>!Ez'$,w!>M!E>!2,$jvf! EHKL!=>!E>NK!LMBFNE:MBHG!qLEM>K!>M!:Ei!}{{Åg!+:OBEBH!>M!:Ei!}{{Åri!.HNK!XMN=B>K!EzBFI:! EzBG?>!4'&!LNK!E:!IHINE:MBHG!=>!<>EENE>L!,)!>G!@XGXK:E!>M!LNK!Ez>QIK>LLBHG!=>L! BLH?HKF>L! =>! !"ÄÅ! >G! I:KMBKf! GHNL! :OHGL! :G:ERLX! =>L! X! L:G@! =>! I:MB>GML!:MM>BGML!=>!4'&!T!A:NM>!OBKXFB>>M!OBKXFB>!:;L>GM>!q:OBKXFB>ri!!A>S!<>L!I:MB>GMLf! GHNL! :OHGL! H;L>KOX! NG>! ;:BLL>! =N! IHNK<>GM:@>! =>L! GML! ,)! >M! ,)2! I:K! K:IIHKM!:NQ!&"!q=HGG>NKL!L:BGLrf!<>!JNB!LN@@WK>!JN>!EzBG?>!4'&!I>NM!:OHBK!NG! BFI:!=>L!<>EENE>L!,)i!">!?:VHG!BGMXK>LL:GM>f!E:!IHINE:MBHG!=>!<>EENE>L! ,)!=>!GHMK>!!=>!I:MB>GML!:!FHGMKX!=>L!F:KJN>L!=>!KX=N!,)%}"i!9M:GM! =HGGX! JN>! E:! =BFBGNMBHG! =>! ,)%}"! >LM! OK:BL>F;E:;E>F>GM! EBX>! :N! EB@:G=f! GHNL! :OHGL! LNIIHLX!JN>!EzBG?>!4'&!>GMK:\G:BM!=>L!M:NQ!IENL!XE>OXL!=>!,)%}"*!S!E>L! I:MB>GML!F>GM!:NQ!&"i!*>!IKH?BE!=zBLH?HKF>L!!"ÄÅ!=>L!<>EENE>L!,)!=XKBOX>L! =>!GHL!I:MB>GML!>LM!=BLMBG!<>ENB!=>L!&"h!:O>L!IKHIHKMBHGL!IENL!XE>OX>L!=>!LHNLj IHINE:MBHGL!,)!!"ÄÅ0L=BF!>M!!"ÄÅ0-á!,)!S!E>L!I:MB>GMLi!*:!IENI:KM!=>L!<>EENE>L!,)! =>L!&"!IKXL>GM:B>GM!NG!IAXGHMRI>!!"ÄÅ0Lái!'E!>LM!BFIHKM:GM!=>!GHM>K!JN>!<>L!<>EENE>L! ,)! !"ÄÅ0L=BF! >M! !"ÄÅ0-áLIXL! :N! 4'&! >QIKBF:B>GM! IKH@K>LLBO>F>GM! NG! M:NQ! BG?XKB>NK! =>! !"|Ç! =:GL! =>L! <>EENE>L! ,)! !"ÄÅ0-á! IKHO>G:GM! T! E:! ?HBL! =>L! I:MB>GML! :OBKXFBJN>L!>M!OBKXFBJN>L!=HGM!E:!IENI:KM!LHGM!!"ÅÇ=BF!"|Çji!*:!I>KM>!=>!!"|Ç!LNK!E:! IHINE:MBHG!=>!,)!!"ÅÇ=BF!"|Çá!EHKL!=>!Ez:!L"L+|É! :!XMX!=XFHGMKX>!!<:K:!=>!E:!LMBFNE:MBHG!IKHEHG@X>!=>L!,)!q0HF>>!>M! :Ei! }{|~ri! .:K! GMf! GHNL! :OHGL! XFBL! EzARIHMAWL>! JN>! El:N@F>GM:MBHG! =>L! IHINE:MBHGL! =>! ,)! !"ÄÅ0L=BF! >M! !"ÄÅ0-á! :R:GM! NG! ?:B;E>! GBO>:N! =>! !"|Ç! S! E>L! I:MB>GML!BG?>!4'&f!KXLNEM:BM!=zBGM>K:EENE:BK>L!>GMK>!E>L!,)!>M!E>L!<>EENE>L! BG?>L!I:K!E>!OBKNL!! "##("($ i!! !

! |Ñ! ">!IENLf!GHNL!:OHGL!=XFHGMKX!BG!OBMKH!JN>!E>L!<>EENE>L!,)!IKBF:BK>L!:EEH@XGBJN>L! LHGM! <:I:;E>L! =z>??>K! NG>! MKH@H! LNK! =>L! EB@GX>L! =>! ERFIAHL! 2! !"Äái! 0XLNEM:GM!=>!E:!MKH@Hf!E>L!<>EENE>L!,)!HGM!:L!KX<>IM>NKL!=z>GMKX>!=N!4'&f! !!"Ä!>M!!!0Åf!IHNO:GM!L>GLB;BEBL>K!T!NG>!BG?>NK>!I:K!E>!4'&i!*>!4'&! I>NM! BG?>K! GHG! L>NE>F>GM! E>L! ERFIAHL! 2! !"Äá! F:BL! :NLLB! E>L! F:L! q&HG>RM!:Ei!}{|Çri!,HL!KXLNEM:ML!LN@@WK>GM!JN>!E>L!<>EENE>L!,)!IHNKK:B>GM!=>O>GBK! NG>!!=N!4'&!:IKWL!E:!MKH@H!=>L!:GMB@WG>L!=>L!<>EENE>L!2!IKHO>G:GM!=>!<>EENE>L! BG?>L!I:K!E>!4'&f!<>I>G=:GM!=>L!IK>NO>L!=BK>L!IHNK!KXIHG=K>!T!<>MM>!JN>LMBHG!LHGM! >G!>G!LMB@:MBHGi! ! !'".(1/#! jY! #PH4FK?GF! 78K! 68DDMD8K! ,)! 4DDG=:F?IM8K! 4N86! MF8! 6QLGLGP?6?L:! 68DDMD4?J8! 8F>GM! :FXEBHKX! E>! MK:BM>F>GM! =>! <>KM:BGL! <:G<>KLi! !>I>G=:GMf! >G! @XGXK:Ef! E>L! FL;! L>NEL! HGM! NG>! :NMBJN>! EBFBMX>i! *zNG! =>! E>NKL! IKBGL! =z:LM! =zBG=NBK>! NG>! L"!!f! JNB! >LM! FX=BX>!I:K!=>L!<>EENE>L!,)i!+:EA>NK>NL>F>GMf!E:!IENI:KM!=>L!I:MB>GML!:MM>BGML!=>!<:G<>K! HGM!=>L!=RL?HGF>GML!BFFNGBM:BK>L!LXOWK>L!:??>E:!I>NM!YMK>! ! <>EENE>L! ,)! :EEH@XGBJN>L! >M! >QI:GLX>Lf! BGH??>GLB?i! ,X:GFHBGLf! F:E@KX! E>NK! ?HKM! IHM>GMB>E! ! ! =B??XK>GM>L! MNF>NKLf! E>L! KXLNEM:ML!L!HGM!XMX!FX=BHLi!! *KD:@67CO !,HNL!:OHGL!L!<>EENE>L!,)!:EEH@WGBJN>L!>M!=>L!FL;L!IHNK! :FXEBHK>K!E>!MK:BM>F>GM!=N!<:G<>Ki!,HNL!:OHGL!@XGXKX!=>L!<>EENE>L!,)!>QI:GLX>L!q>j,)r! :O>!?HKM>L!KXIHGL>L!L"!!! "##("'&$ !>M! "##("($ !!=B??XK>GM>L!MNF>NKL!>M!>G!NMBEBL:GM! =B??XK>GML!FL;!MAXK:I>NMBJN>Lf!T!L:OHBK!E>!KBMNQBF:;f!EzH;BGNMNSNF:;f!E>!=:K:MNFNF:;f! E>!<>MNQBF:;!>M!E>!MK:LMNSNF:;i!! .KCE=D3DCO ! 0>F:KJN:;E>F>GMf! E>L! <>EENE>L! >j,)! I>NO>GM! YMK>! KOX>L! EX>L! >Mf! :IKWL! =XL! =>! FL;Li! 'EL! FX=B>GM! EzL"!!! I:K! =>L! FX<:GBLF>L!=XI>G=:GML!>M!BG=XI>G=:GML!=>!E:!=X@K:GNE:MBHGi!#G!HNMK>f!BEL!LNKFHGM>GM! <>KM:BGL!FX<:GBLF>L!:GMBj:IHIMHMBJN>L!IKXL>GML!=:GL!E>L!<>EENE>L!E>NLi!! !@?5=EC;@?O !,HNL!:OHGL!XM:;EB!NG!GHNO>:N!IKHMH!=z:QI:GLBHG!=>L! <>EENE>L! ,)! :O>! :OX>i! *zNMBEBL:MBHG! =>! FL;L! >G! L!<>EENE>L!>j,)!IHNKK:BM!IHM>GMB>EE>F>GM!:FXEBHK>K!E>!MK:BM>F>GM!=N!<:G<>Ki!

! |Ö! !-,!*20(-,0!&$,$/"*#0 ! ! #G!KXLNFXf!E>!MK:O:BE!LNK!Ez:G:ERL>!=>L!BLH?HKF>L!!"ÄÅ!=:GL!E:!LHNLjIHINE:MBHG!=>! <>EENE>L!,)!BLLN>L!=>!=BO>KL!<:G<>KL!AXF:MHEH@BJN>L!M>EL!JN>! j,&*!q!Är!>M!*L+! q!År!:!KXOXEX!=>L!IK>NO>L!=>!=RL?HGF>GM!=>!<>L!<>EENE>L!:N!FHF>GM!=N! =B:@GHLMB!IENLf!E>L!IKH?BEL!=zBLH?HKF>L!!"ÄÅ!LNK!E>L!<>EENE>L!,)!=>!<>L!I:MAHEH@B>L! E>L! =B??XK>GGM! =>L! I>KLHGG>L! >G! ;HGG>! L:GMXi! LBGLBf! E>L! <>EENE>L! ,)! !"ÄÅ0-á! LHGM! I>KMBG>GM>L! <:K! :LLHL! T! ElXEBFBG:MBHG! =>L! <>EENE>L! MNFHK:E>L! qLMjTj=BK>! *#,! IENL! - 8f!!A:IBMK>!Äri ! ! *>L!<>EENE>L!,)!IKHO>G:GM!=>!I:MB>GML!BG?>!4'&!GG>GM!=>L!LHNL! IHINE:MBHGL!<:K:L!I:K!!"ÄÅ0L=BF!>M!!"ÄÅ0-á!=HGM!E>!GBO>:N!=>!!"|Ç!=BFBGN>! IKH@K>LLBO>F>GMf!LN@@XK:GM!NG>!:GM>!BG!OBOHi!!>L!<>EENE>L!,)!IKHO>G:GM!=>! I:MB>GML! 4'&! á! IKXL>GM>GM! X@:E>F>GM! NG! KX<>IM>NK! :NK! ,)%}"! =BFBGNXf! <>! JNB! IHNKK:BM!YMK>!NG>!F:KJN>!=z:EMXK:MBHG!?HGEE>!BG=NBM>!I:K!NG!OBKNLi!! ! #G?BGf! Ez>QI:GLBHG! =>L! <>EENE>L! ,)! =z3! ! Lz>LM! :OXKX>! >??B<:<>! >G! NKL! :GMB!=>L! :GMB@WG>L! MNFHK:NQi! !>E:! KXLNEM>! >G! NG>! L"!!! ?HKM>! BG! OBMKH! >M! BG! OBOHi! !>! MK:O:BE! LHNMB>GM! ElARIHMAWL>! =>! EzNMBEBL:MBHG! =>! <>EENE>L!>,)!!!>G!BFFNGHMAXK:IB>i! ! +HMLjEENE>!MN>NL>!G:MNK>EE>f!<:G<>K!AXF:MHEH@BJN>f!4'&f!!"ÄÅf!BFFNGHMAXK:IB>! :GMBjMNFHK:E>i ) ) ! ! ! ! ! ! ! ! ! ! ! ! !

! }{! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! }|! ! ! ! ! ! ! ! ! ! ! ! !

!'".1#/!d Z!

! &#,#/"*!(,1/-"2!1(-,! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! }}! ! ! ! ! ! ! dZd !&8F8J4D!GN8JN?8O!45GML!L>8!?EEMF8!KQKL8E!4F7!?LK!6GEHGF8FL! ! 'EEMFALQ) AK) GF=) AEHGJL9FL) 9KH=;L) G>) DA>=j) 'F<==) DA>=LAE=) AF)>JGE)L@=)L@J=9LK)G>)>GJ=A?F)AFN9<=J) H9L@G?=FKj)'F)GJ<=J)LG)>)9?9AFKL)>=J=FL)AFN9) L@=) @GKL) AEEMF=) KQKL=E) ;GGH=J9L=) AF) >MF;LAGF) LG) HJGNA<=) HJGL=;LAGF)>JGE)N9JA=LQ)G>)H9L@G?=FA;)EA;JGGJ?9FAKEKg)NAJMK=K)9F<);9F;=Jj) ) 2@=) AFF9L=) AEEMF=) ;GEHGF=FLK) 9J=) L@=) >AJKL) DAF=) G>) <=>=FK=) L@9L) J=KHGFK=) LG) HJ=K=F;=) G>) H9L@G?=FK) AF) 9) :JG9<) 9F<) FGFkKH=;A>A;) E9FF=Jj) 2@=) ;=DDK) G>) AFF9L=) AEEMFALQ) ?=F=J9DDQ) AF;DM<=) ?J9FMDG;QL=Kg) E9;JGH@9?=Kg) F=MLJGH@ADKg) EGFG;QL=g) <=FA;)F9LMJ=)G>)L@=) AFF9L=) AEEMFALQ) L@9L) 9DDGOK) AL) LG) IMA;CDQ) K;9F) >GJ) H9L@G?=FA;) AFN99;LGJK) L@9L) D9LL=J) K@9H=K) L@=) <=N=DGHE=FL)G>)9F)9FLA?=FkKH=;A>A;)9<9HLAN=)AEEMF=)J=KHGFK=)?ANAF?):Q)2)9F<) );=DD)G>) L@=) 9<9HLAN=) AEEMFALQj) 2@=) KM:K=IM=FL) 9<9HLAN=) AEEMFALQ) J=IMAJ=K) DGF?=J) LAE=) LG) <=N=DGH)AF)L=JE)G>)HJGL=;LAGF):Q)AEEMFG?DG:MDAF)tAj=j)9FLA:G>A;A=FLDQ) :ML) 9DKG) HJGNA<=) DGF?) L=JE) HJGL=;LAGF):Q)E=9F)G>)AEEMF=);=DD)E=EGJQj)'L)O9K)?=F=J9DDQ)9HHJ=;A9L=<)L@9L)AEEMF=) E=EGJQ) ;=DDK) GFDQ) G;;MJ) AF) 9<9HLAN=) 2) 9F<) ) ;=DD) J=KHGFK=K) :ML) J=;=FL) =E=J?AF?) =NA<=F;=K) 9:GML) 9<9HLAN=) >=9LMJ=K) AF) EMJAF=) 9F<) @ME9F) ,)) ;=DD) AK) ;@9DD=F?AF?) L@AK) DGF?kKL9F) 9F<) G>>=J) 9) F=O) NA=O) LG) L@=) AFF9L=k9<9HLAN=) )L@=)LOG)H9JLK)G>)L@=)AEEMF=)KQKL=E)AF)9);GGJ

! }~! ) ('0.ﲩ@=E9LA;)EG<=D)G>)@=E9LGHGA=KAK)9F<)DAF=9?=);GEEALE=FLj )

L@=) GML;GE=K) O@=L@=J) D=9) L@=) AFN9=;LAGFKj) )) FDL@GM?@) LG) KGE=) =PL=FL) ;=DDK) G>) FGFk@=E9LGHGA=LA;) GJA?AF) @9N=) L@=AJ) JGD=) AF) AEEMF=)J=KHGFK=)LG)H9L@G?=FKg);=DDK)G>)L@=)AEEMF=)KQKL=E)>JGE):GL@)L@=)AFF9L=)9F<) 9<9HLAN=) ;GEHGF=FLK) 9J=) L@GK=) <=JAN=<) >JGE) @=E9LGHGA=LA;) GJA?AFj) &=E9LGHGA=LA;) KL=E) ;=DDK) t&1!Ku) 9J=) K=D>kJ=F=O9:D=) !"ÅÇã) KL=E) ;=DDK) L@9L) 9J=) DG;9L=<) AF) L@=) :GF=) E9JJGOg) O@A;@) ?AN=) JAK=K) LG) L@=) ;GEEGF) EQ=DGA<) HJG?=FALGJK) t!+.u) 9F<) ;GEEGF) DQEH@GA<)HJG?=FALGJK)t!*.u)J=KH=;LAN=DQj)2@=K=)DAF=9?=K);GEEALL=<)HJG?=FALGJK)OADDg) AF) LMJFg) ?AN=) JAK=) LG) L@=) KM:K=IM=FL) ;=DD) LQH=K) G>) L@=) EQ=DGA<) DAF=9?=) >JGE) L@=) !+.) GJA?AF)GJ)DQEH@GA<)DAF=9?=)>JGE)L@=)!*.)GJA?AF)J=KH=;LAN=DQj)2@=)K;@=E9LA;)>=J=FLA9LAGF) ;9K;9<=K) G>) L@=) &1!K) AK) <=EGFKLJ9L=<) AF) $64?<3# HCH j) 'F) L@=) K;GH=) G>) L@AK) L@=KAK) OGJCg) N9JAGMK) 9KH=;LK) G>) ,)) ;=DD) :AGDG?Q) 9F<) H@QKAGDG?Q) AF) ;9F;=JK) 9F<) NAJMK) AF>=;LAGF) OADD) :=) ;DGK=DQ) =P9EAF=<) 9F<) GJ=)AL)AK)=KK=FLA9DDQ)AEHGJL9FL)LG)J=NA=O)O@9L)O=)@9N=)CFGOF)9:GML),));=DD)>JGE) L@=)DAL=J9LMJ=j) dZe !,)!68DD!GFLG=8FQ! ! 'F)L@=)EA<)~áÖ}Kg)L@=)A<=FLALQ)G>),))O9K)KL9JL=<)LG):=)J=;G?FAR=<):Q)ALK)HGL=FLA9D)LG) 9MLGFGEGMK)DQKAK)G>);9F;=J);=DDK)OAL@GML)HJAGJ)9FLA?=F)AEEMFAR9LAGF)@=F;=)L@=)L=JE)

! }Ä! o,9LMJ9D))ADD=Jp)9JGK=)t&=J:=JE9F)=L)9Dj)~áÖÉh))A=KKDAF?)=L)9Dj)~áÖÉuj)1GGF)9>L=J)AL)O9K) J=9DAR=<) L@=) F==<) >GJ) 9) KMJ>9;=) E9JC=J) LG) =KK=FLA9DDQ) A<=FLA>Q) L@AK) o,9LMJ9D) )ADD=Jp) HGHMD9LAGF) G>) ;=DDK) >JGE) 2) 9F<) ) ;=DDK) G>) L@=) DQEH@GA<) DAF=9?=j) !GFL=EHGJ9JQ) <=N=DGHE=FL) G>) @Q:JAGJ) E9CAF?) EGFG;DGF9D) 9FLA:GDGO) ;QLGE=LJQ) 9DDGOAF?) 9FKO=J) LG) L@AK) HJG:D=E) :=;9E=) HGKKA:D=j) 2@=) >AJKL) EGFG;DGF9D) 9FLA:GMDDQ) A<=FLA>Q) ,)) ;=DD) >JGE) L@=) J=KL) G>) DQEH@GA<) HGHMD9LAGFK) O9K) *=Mk~~) tFGO) CFGOF)9K)!"~Ñu)t*9FA=J)=L)9Dj)~áÜÅuj)1M:K=IM=FLDQ)9FGL@=J)KMJ>9;=)E9JC=Jg)!"ÉÑ)t*=Mk ~áu) @9K) :==F) A<=FLA>A=<) LG) KM;;=KK>MDDQ) <=K;JA:=) @ME9F) ,)) ;=DD) t*9FA=J) =L) 9Dj) ~áÜáuj) !"ÉÑ) AK) 9) ~Ç}C"9) AKG>GJE) G>) L@=) F=MJ9D) ;=DD) 9<@=KAGF) EGD=;MD=) ~) t,!F+k~u) L@9L) 9:MF<9FLDQ)=PHJ=KK=<)AF);AJ;MD9LAF?),));=DDK)L@9L)HJ=NAGMKDQ)HJGHGK=<)LG)HD9Q)9)JGD=)AF) E=MJL@=J)KM??=KL=<)L@9L) !"ÉÑ)9DKG)H9JLA;AH9L=K)AF),));=DD)<=N=DGHE=FL)NA9)>9;ADAL9LAF?)AFL=J9;LAGF)OAL@)KLJGE9) ;=DDK)9F<)K=JN=K)9K)H9L@G?=F)J=;G?FALAGF)J=;=HLGJ)t.00u)AF)9FLAk>MF?9D)AFF9L=)AEEMF=) J=KHGFK=)t+9;=)=L)9Dj)Ä}~Ñh)8A=?D=J)=L)9Dj)Ä}~Öuj)'FL=J=KLAF?DQ)EGMK=),));=DD)) !"ÉÑ) @=F;=) ,)HÇÑg) 9) @A?@DQ) ;GFK=JN=<) F9LMJ9D) ;QLGLGPA;ALQ) J=;=HLGJ)t,!0u)AK)9)E9JC=J)>GJ)A<=FLA>Q),));=DD)9;JGKK)KH=;A=K)t59DR=J)=L)9Dj)Ä}}Öuj)'F) ;AJ;MD9LAF?)H=JAH@=J9D):DGG)LGL9D)DQEH@G;QL=)) ) )

('0.ﴎOG)E9AF)HGHMD9LAGFK)G>),));=DDK)AF);AJ;MD9LAF?)H=JAH@=J9D):DGG<)OAL@)L@=AJ) >=J=F;=)AF)!"ÉÑ)9F<)!"~Ñ)=PHJ=KKAGF)t!GGH=Jg)$=@FA?=Jg)9F<)!9DA?AMJA)Ä}}~u)

! }Å! HGHMD9LAGFKj) 9K=<) GF) L@=) KMJ>9;=) =PHJ=KKAGF) G>) !"ÉÑ) 9F<) !"~Ñg) LGL9D) ,)) ;=DD) HGHMD9LAGF) ;9F) :=) >MJL@=J) >=J=FL) AF) E9FQ) KMJ>9;=) E9JC=JK) =PHJ=KKAGF) 9K) O=DD) 9K) ;QLGCAF=) K=;J=LAGF)9F<);QLGLGPA;ALQ)HGL=FLA9D)t $64?<3#HCIu) ! 'F)H=JAH@=J9D):DGG9;L)L@9L)!"ÉÑ:J)=PHJ=KK=K)@A?@)D=N=D) G>)L@=);@=EGCAF=)J=;=HLGJK)!!0Ö)9F<)!6!0Å)9K)O=DD)9K)L@=)9<@=KAGF)EGD=;MD=)!"ÑÄ*) L@9L) ?ANAF?) @GEAF?) 9F<) =FLJQ) KA?F9D) LG) L@=) 12*) =FNAJGFE=FL) O@=J=) L@=) J=;AHJG;9D) DA?9FGJ) L@=K=) J=;=HLGJK) 9J=) 9:MF<9FLDQ) =PHJ=KK=<) t$J=M<) =L) 9Dj) Ä}~Öuj) 'F) ;GFLJ9KL) !"ÉÑGJ=) L@=Q) 9J=) EGJ=) HJ=) ;QLGCAF=) J=;=HLGJKg) AL) @9K) :==F) <=EGFKLJ9L=<) L@9L) !"ÉÑ:J) ,)) ;=DD) ;GFKLALMLAN=DQ) =PHJ=KK=K) ;kCAL) t!"~~Öug) '*Ö0x) t!"~ÄÖu) 9F<) '*Ä0x) t!"ÄÉu) L@9L) =F9:D=) L@=E) LG) J=KHGFK=) LG) DGO) ) ;kCAL) DA?9F<) t)*ug)'*Ö)9F<)'*Ä)J=KH=;LAN=DQj) GL@)!"ÉÑ:J)9F<)!"ÉÑ>AFALQ)'*kÄ0)KM:MFALK)y)t!"~ÄÄug)9F<)z)t!"~ÅÄu):ML)D9;CAF?)L@=)x)KM:kMFAL)G>) L@=)'*kÄ0);GEHD=P)J=F<=JK)!"ÉÑ)'*kÄj)'F)LMJFg)!"ÉÑGJE)9FLA:GMF;LAGF) >GJ) =DAEAF9LAF?) 9FLA:G)!"~Ñ) 9F<);QLGLGPA;)?J9FMD=K)9J=)EGJ=);9H9:D=)G>)H=J>GJEAF?)F"!!)GJ)F9LMJ9D);QLGLGPA;ALQ) 9;LANALQ) LGO9JMD) ;=DDKj) 29C=F) LG?=L@=Jg) AF@=J=FL) >=J=F;=) :=LO==F) !"ÉÑ:J) 9F<) !"ÉÑMD) AF) AEEMFGEGGJ)!"ÉÑ:Ju)GJ)GJ)!"ÉÑMF;LAGF) G>) L@=) AEEMF=) J=KHGFK=j) $64?<3# HCJj) KMEE9JAR=<) H@=FGLQHA;) 9F<) >MF;LAGF9D)>=J=F;=K):=LO==F)L@=)LOG)HGHMD9LAGFK)G>),));=DDKj)

! }Ç! ) ('0.ʈ#.A>GHMRIBK>G<>L!;>MP>>G! !"ÅÇ;K!:G=!!"ÅÇ=BF!,)!<>EE! IHINE:MBHGLi!q!HHI>K!>M!:Ei!}{{|r!

) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) =KA<=K)L@=)LOG)O=DD);@9J9;L=JAR=<);GFN=FLAGF9D),))t;,)u)HGHMD9LAGFKg)?JGOAF?) =NA<=F;=) 9DKG) KM??=KL) L@=) HJ=K=F;=) G>) LAKKM=kKH=;A>A;) J=KA<=FL) ,)) tLJ,)u) ;=DD) HGHMD9LAGFK)L@9L)9J=)DG;9DAR=<)AF)DAN=Jg)ML=JMKg)DMF?g)CAMF;LAGF9DDQ) 9F<) AEEMFGH@=FGLQHA;9DDQ) ;GEH9J=<) LG) L@=) ;,)) HGHMD9LAGFK)) t BaJCKLJaEg)*BMF??J=Fg)9F<)+A;@9YDKKGF)Ä}~Ñh)+=DK=F)=L)9Dj)Ä}~Ñuj) ) 4=:E:8 9C:F8 ,)8 8:@@8 9:I:@CDR ) #9JDQ) KLM>=J=FLA9LAGF);9F):=)AFB8 I>GEC );MDLMJ=)G>)&1!)OAL@)'*k~É)GJ) +k<=JAN=<)KLGE9);=DDK)t!GDM;;Ag)!9DA?AMJAg)9F<)"A) 19FLG)Ä}}Åh)+ADD=Jg)FDD=Qg)9F<)+;%D9N=)~ááÇh)+J_R=Cg)FF<=JKGFg)9F<)!9DA?AMJA)~ááÑuj)

! }É! &GO=N=J) HJG?J=KKAF?) KLMJGE) N9JAGMK) KGMJ;=K) t#AKK=FK) =L) 9Dj) Ä}~Äh) $J=M<) =L) 9Dj) Ä}}Ég) Ä}}Ñh) 1;GNADD=)=L)9Dj)Ä}~Ñuj)F;;GJ)<=N=DGHE=FLg)E9LMJ=),));=DDK)9J=) ?=F=J9L=<) >JGE) &1!) ?GAF?) L@JGM?@) É) KL9?=K) G>) <=N=DGHE=FL9D) AFL=JE=MJL@=J);GEEALL=<)AFLG) ,));=DD)DAF=9?=)t1;GNADD=g)$J=MMJL@=J);GEEALL=<)AFLG),)) ;=DD) DAF=9?=) >=J=FLA9LAGF) :Q) L@=) D9;C) G>) 9:ADALQ) LG) ?AN=) JAK=) LG) 2r"!) HJ=;MJKGJK) 9DL@GM?@)KL9?=)Å),)"')KLADD)D9;C)2:=L)9F<)#-+#1)=PHJ=KKAGF)L@9L)KH=;A>A;)>GJ)E9LMJ=)) ) ) ) ) ) ) ) )

) ) ('0.ʉ#1NFF:KBS>=!LM:@>L!H?!,)!<>EE!=>O>EHIF>GMLi!q7N!>M!:Ei!}{|~r!

)

! }Ñ! ,));=DDj)$AF9DDQ)KL9?=)Ç)9F<)KL9?=)É),)"'K)9J=)H@=FGLQHA;9DDQ)A<=FLA;9D)LG)!"ÉÑ:J)9F<) !"ÉÑGJ=) J=?9J<=<) 9K) L@=) ) L@=) EGJ=) E9LMJ=) !"ÉÑB8 I>GEC )KLM)L@=)AF>=J=FLA9L=<) t*M=LC=k#N=JKDG@g) )ADDA?g) 9F<) 0GE9?F9FA) Ä}~Åh)7Mg)$J=MGJ=)MFLAD)>MJL@=J)=NA<=FL)=E=J?=K)LG)N9DA<9L=) GJ) ) @Q:JA<) J=KAKL9F;=) AF) :GF=) E9JJGO) LJ9FKHD9FL9LAGF) D=9)mEAKKAF?kK=D>p)J=;G?FALAGF)t*BMF??J=F)9F<))9JJ=)~áá}uj)2@AK) H@=FGE=FGF)O9K)D9L=J)GF)HJGN=F)L@QEA;kAF<=H=F<=FL)@=F;=)FGL)AFNGDN=<)AF)2);=DD):ML) ,)) ;=DDkE=p) @QHGL@=KAKg)AL)O9K)HJGHGK=<)L@9L),));=DDK);GFKL9FLDQ)K=FK=)L@=)HJ=K=F;=)G>)+&!);D9KK)'))

) ('0.@:!2A>!FBLLBG@jL>E?!ARIHMA>LBL!BG!,)!<>EE!q*CNG@@K>G!:G=!+:EF;>K@!}{{Ér!

)

! }Ö! EGD=;MD=K) =PHJ=KK) GF) @=9DL@Q) ;=DDK) 9F<) J=E9AF) LGD=J9FLj) 'F) ;GFLJ9KL) O@=F) NAJ9D) AF>=;LAGF) GJ) LMEGJ) LJ9FK>GJE9LAGF) A=<) @=F;=) )'0) DG;MK) 9J=) MF<=J?GAF?) =PL=FKAN=) HGDQEGJH@AKE) t.9J@9E) Ä}}Éuj) 2@=) KAEAD9J) KALM9LAGF) G;;MJK) AF) EA;=) O@=J=) )'0k @GEGDG?GMK) AF) EGMK=) AK) *QÇá) J=;=HLGJ) >9EADQ) L@9L) J=;G?FAR=K) &kÄ) ?=F=K) =F;GGJ)L@=)FGFk ;D9KKA;9D) &*Fk#) DA?9F<) AF) @ME9F) 9F<) /9k~:) AF) EA;=j) F;;GJp) @QHGL@=KAK)L@=J=>GJ=)AL)J=IMAJ=K)9);GFk+&!);D9KK)')LG)E9AFL9AF)LGD=J9F;=g)L@AK)FGLAGF)AK)KMHHGJL=<) :Q)=9JDQ)OGJCK)MKAF?),));=DD);DGF=K)t49DA9FL=)=L)9Dj)~ááÖuj),=N=JL@=D=KKg)J=;=FL)=NA<=F;=) KM??=KL=<)L@=)HJ=K=F;=)G>),));=DD)KM:K=LK)L@9L)k+&!) ;D9KK) ') 9F<) L@=K=) ,)) ;=DDK) 9J=) @QHGJ=KHGFKAN=) MHGF) KLAEMD9LAGF)tFF>GKKA)=L)9Dj)Ä}}Ñh)$=JF9F<=R)=L)9Dj)Ä}}Éh))AE)=L)9Dj)Ä}}Éuj)2@AK)J=IMAJ=K) 9F)=PH9FKAGF)G>)L@=)mEAKKAF?kK=D>p);GF;=HL)AFLG),));=DD)=MDDQ)>MF;LAGF9D):Q)=PHJ=KKAF?)9L)D=9KL)GF=))'0)GJ) L@=) AF@A:ALGJQ) ,)%ÄF) J=;G?FARAF?) L@=) ;D9KKA;9D) +&!) ;D9KK) ') t&*FkFr r!u) 9F<) FGFk ;D9KKA;9D) &*Fk#j) 2@=) MFk=9AD) LG) =PHJ=KK) 9FQ) K=D>kDA?9F<) )'0) 9F<) J=E9AF=<)@QHGJ=KHGFKAN=g)@=F;=)LGD=J9FL)LG)FGJE9D)@=9DL@Q);=DDKj)1=N=J9D)EG<=DK)@9N=) :==F)HJGHGK=<)LG)=PHD9AF)L@=)HJG;=KK)G>),));=DD)=B<8 AC9:@P ) ,)) ;=DD) =PHJ=KKAF?) K=D>k+&!) DA?9F<) J=;=HLGJK) HJGEGL=) >MF;LAGF9D) E9LMJ9LAGF) D=9F6EA>B<8 AC9:@P ) 2@AK) EG<=D) KM??=KL) ,)) ;=DD) 9J=) >MDDQ) >MF;LAGF9D) :Q) <=>9MDL)9F<)o)>9AD)LG)=PHJ=KK)GF=)G>)L@=))'0)>GJ)K=D>k+&!);D9KK)'))

! ~{! ) ) ) ) ) ) ) ) ) • ) ) • ) ) • ) • ) • ) • ) • ) • ) • ) • ) • ) ) ('0.@##"B??>K>GM!FH=>EL!H?!,)!<>EE!>=N<:MBHG!q&`@ENG=!:G=! KH=BG!}{|{r #

) • LG) J=L9AF) LGD=J9F;=g) ,)) ;=DDK) L@9L) 9;IMAJ=) AF@A:ALGJQ) )'0) 9J=) L@=J=>GJ=) 9NGA<) L@AK) =>>=;L) @=F;=) J=E9AF=<) >MF;LAGF9Dj) !GEH9JAF?) LG) L@=) o9JEAF?p) EG<=Dg) L@AK) EG<=D) GJ) )'0) =P;=HL) AF@A:ALAF?)9;LAN9LAGF)KA?F9Dj) )

! ~|! • !>FQ>BG:E68G>CB8 AC9:@P ) 2@AK) EG<=D) =E=J?=<) :Q) L@=) G:K=JN9LAGF) L@9L) K=D>) +&!) ;D9KK) ') EGD=;MD=K) ;9F) AFL=J9;L) OAL@) L@=AJ) ;GJJ=KHGFF g) L@=J=>GJ=) J=L9AF) L@=) AF@A:ALGJQ) )'0) >JGE) J=;G?FARAF?) K=D>k +&!) ;D9KK) ') EGD=;MD=K) HJ=K=FL=<) GF) L9J?=L) ;=DDKj) &GO=N=J) L@=) IM=KLAGF) O@=L@=J) 9DD) )'0K) ;9F) AFL=J9;L) OAL@) +&!k') EGD=;MD=K) AF) 8>F8 E9C=K) AL) MF;D=9J)O@=L@=J)L@AK)EG<=D)>MDDQ)=PHD9AFK)L@=),))=9J) L@=) o9JEAF?p) 9F<) oAFAL=) KL9L=) G>) >MF;LAGF9Dr@QHGJ=KHGFKAN=) KL9LMK) :ML) ,)) ;=DD) =AF>=J=FL)+&!) ;D9KK) ') 9DD=K) HGKK=KK) FGFk=IM9D) m=>=;Lp) GF) L@=AJ) ;GJJ=KHGFGJ=) KM??=KLK) 9) A=<) KH=;LJME) G>) ,)) ;=DD) HGHMD9LAGFK) OAL@) >=J=FL) >MF;LAGF9D) HGL=FLA9D) <=H=F) L@=) )'0) t9)'0u) J=;=HLGJ) >9EADQ) J=N=9D=<) 9<) ,)) ;=DD) =))'0Ä"1~),));=DD)KM:K=LK)AF)&*Fk!Äã)kAF@A:ALGJQ))'0) GJ),)%ÄF)=PHJ=KKAGF)KM??=KLAF?)L@=K=))'0Ä"1~),));=DDK)OGMD<):=)@A?@DQ)9MLGJ=9;LAN=) LGO9J) +&!k') =PHJ=KKAF?) ;=DDK) t!G?F=L) =L) 9Dj) Ä}~}uj) &GO=N=Jg) KM:K=IM=FL) ;@9J9;L=JAR9LAGF)KLMGJ)L@=)J=KH=;LAN=)9)'0u)@GEGRQ?GMK)kJ=9;LAF?) 9)'0k=PHJ=KKAF?) ,)) ;=DDK) D9;CAF?) 9FQ) AF@A:ALGJQ)J=;=HLGJK)O=J=)LMF=<)9EADQ)J=;=HLGJK)Ä Ç)t!"ÄÇÇug)1*F+$Ñ)9F<)GL@=J)AF@A:ALGJQ) J=;=HLGJK) KM;@) 9K) 2'%'2g) !"~Ñ~) t,)0k.~Fu) ;9F) =

! ~}! DA?9F) L@=K=)FGFk;GFN=FLAGF9D)=B8 >AAHB>GLR )2@=)HJ=K=F;=)G>):GL@),)) HGHMD9LAGFK) AF) FGJE9D) AF>A;A=F;Q) H=JKH=;LAN=j) 2@=J=>GJ=) L@=J=) EMKL) :=) 9) >=J=FL) JGD=) G>) :GL@) HGHMD9LAGFKi) =) HJGL=;LAGF) AF) L@=) AEEMF=) KQKL=Ej) 0=;=FL) KLM) EMJAF=) ;QLGE=?9DGNAJMK) t+!+4u) AF>=;LAGFK) KM??=KLAF?)L@9L)MFk=kJ=;=HLGJKg)AK)EGJ=)HGL=FLA9DDQ)9;LAN=)9?9AFKL);=DDK)L@9L)GJE=<)LMEGJ);=DDK)9F<)GJ)NAJMK)L@9L)MK=<)L@=K=)AEEMF=k =N9KAGF)KLJ9L=?A=Kj) ) dZg !,)!68DD!J868HLGJ!J8H8JLG?J8!4F7!K?=F4D?F=!H4L>O4QK! ) )'F) KHAL=) L@=) D9;C) G>) J=9JJ9?F=<) 9FLA?=FkKH=;A>A;) J=;=HLGJ) DAC=) GL@=J) DQEH@G;QL=) ;GMFL=JH9JLK)KM;@)9K)2);=DD)9F<) );=DDg),));=DDK)=PHJ=KK)9)OA<=)KH=LJME)G>)?=JEDAF=k =F;G<=<):GL@)AF@A:ALGJQ)9F<)9;LAN9LAF?)J=;=HLGJK)L@9L)OGJCK)AF)KQF=J?Q)LG)=P=JL)N9JA=LQ) G>) >MF;LAGFK) AF) @GKL) <=>=FK=) E=;@9FAKE) 9F<) @GE=GKL9KAK) t*9FA=J) Ä}}Åuj) 'L) AK) FGO) OA<=DQ)9;;=HL=<)L@9L)L@=)AFL=?J9LAGF)G>):GL@)AF@A:ALGJQ)9F<)9;LAN9LAF?)KA?F9DAF?)AFHML) ?GN=JFK) L@=) GML;GE=) G>) ,)) ;=DDK) 9F<) L@=AJ) L9J?=L) AFL=J9;LAGF) D=9GJ=)AL)AK)F=;=KK9JQ)LG):JA=>DQ)KMEE9JAR=)L@=)),)) ;=DD)J=;=HLGJK)9F<)L@=AJ)DM=FL),));=DD) 9;LANALQj) ) 'B=>7>GCEL8 E:8:DGCEF j) .=J@9HK) L@=) EGKL) O=DD) ;@9J9;L=JAR=<) ;D9KK) G>) ,)) ;=DD) AF@A:ALAF?)J=;=HLGJK)9J=)E=E:=JK)G>)L@=))'0)>9EADQ)9K)GJEKi) )'0k*) J=;=HLGJK) OAL@) DGF?) ;QLGHD9EA;) L9AD) ;GFL9AFAF?) AEEMFGJ=;=HLGJ) LQJGKAF=k:9K=<) AF@A:ALAGF) EGLA>) t'2'+u) 9J=) AF@A:ALGJQ) O@=J=9K)K@GJL);QLGHD9KEA;)L9AD);GMFL=JH9JLK))'0k1)D9;CAF?)'2'+):ML);9F);GMHD=<)OAL@))

! ~~! +#!#*-)+, ) '&%D(", ) !:8839>= ) )) *$Fk~) '!F+k~rÄrÅ) ) 'EHDA;9L=<)AF)9<9HLAN=) !"Ä) !"ÇÜg)!"ÉÜ) ,));=DD)KM:K=LK)AF)&!+4) AF>=;LAGF) D253=6:9# >A:JAFG?=Fg)'!F+k~g) !"~~:r;) ) +313;>:<= ) '!F+kÇ) ,!F+)t!"ÉÑu) $%$0~) ) *kK=D=;LAF) %DQ!F+k~g)+9

'*2Ä)t*'*0 ~u) &*Fk'g)3*k~Ü)t!+4u) )

)'0Ä"*~) &*Fk!Ä) ) )'0Ä"*ÄrÅ) &*Fk!~) ) )'0Ä"*Ç) &*Fk%) ) )'0Å"*~) &*Fk OÇ) ) )'0Å"*Ä) &*FkFvÅk~~g)!H%k-",) ) ,)%ÄFr!"áÇ) &*Fk#) ) !"~Ñ}) &*Fk!) ) 1A?D=;)Å) ) &95606>::<= ) 1A?D=;)á) ) *F'0~) !GDD9?=F) ) )*0%~) !9<@=JAFK) ) AEEMF=)=P@9MKLAGF) 2'+kÅ) %9D=;LAFká) 9KKG;A9L=<)E9JC=JK) AEEMF=)=P@9MKLAGF) ."k~) ."k*~g)."k*Ä) 9KKG;A9L=<)E9JC=JK) !"áÑ) ,=;LAFk~g).40)t!"~ÉÉu) ,=;LAFkÄ)t!"~~Äug) K@9J=)DA?9F

$;?0'''9)t!"~Ñu) '?%) J=KHGFKA:D=)>GJ)F"!!)

)'0Ä"1~) &*Fk!Ä) ) D1>6@/>694# )'0Ä"1ÄrÅ) MFCFGOF) ) 9;LAN9LAF?)J=;=HLGJ)LG) +313;>:<= ) )'0Ä"*Ç) &*Fk%) KGDM:D=)&*Fk%) )'0Å"1~) &*Fk OÇ) ) &*FkFv~~g)KGE=)&*Fk!) )'0Ä"1Ç) ) 9DD=D=K)

! ~Ä! #PH9FKAGF)G>) ,)%Ä!r!"áÇ) &*Fk#) ,)%Ä!ã!"ÉÖã)AF) N9JAGMK)NAJMK)AF>=;LAGFK) +'!kFr g)3* .~kÑ) &kÑ}g)+MDLk~g)09=k~) ,)%Ä") t@ME9Fu) tEGMK=u) ,)0.k~F)t!"~Ñ~u) **2k~) ) NAJ9D)&Fg)HJGH=J9EADQ)J=;=HLGJK) !0F!!r1*F+$Ö) !0F!!r1*F+$Ö) !"ÄÖ) !"Ö}) 2,$)KMH=J>9EADQ) !"~ÅÖ)tÇk~ u) Çk~ *) J=;=HLGJK) "3/>5# $9K*) $9K*) 2,$)KMH=J>9EADQ) <33;>:<= ) 20F'*) 20F'*0) J=;=HLGJK) ) ) #$%)@#1NFF:KR!H?!,)!<>EE!F:CHK!:<>IMHKL!:G=!MA>!EB@:G=L! MA:M!MA>R!K>i!! ####################################

) "F.~Äg) 9F) 9;LAN9LAF?k9KKG;A9L=<) J=;=HLGJ) L@=J=>GJ=) HJGEGL=) ,)) ;=DD) 9;LAN9LAF?) KA?F9DAF?j) )'0Ä"*Ç) AK) L@=) GFDQ) =P;=HLAGF) 9EGF?) )'0) >9EADQ) AF) O@A;@) L@AK) J=;=HLGJ) ;GFL9AFK) 9F) '2'+) ) ;GMHDAF?) LG) '2F+k9<9HL=J) EGD=;MD=Kg) )'0Ä"*Ç)E9AFDQ)AFNGDN=<))ALK)K=D=;LAN=DQ)KGDM:D=) DA?9F<)&*Fk%)AF)L@=)ML=JMKj)'F)9<GJ) FGFk+&!) ;D9KK) ') DA?9F9EADQ) J=;=HLGJK) 9DKG) HJ=K=FL) GF) ,)) ;=DDK) :ML) L@=AJ) JGD=K) AF) J=?MD9LAF?) ,)) ;=DD) AF@A:ALAF?) KA?F9DK)J=E9AF=<)D9J?=DQ)MF=PHDGJ=<)t -/073#H uj) ) ) &CJ8 '1'+QE:8:DGCEF8 A:9>6G:8 ,)8 8:@@8 >B=>7>G>CBR ) 'L) AK) KH=;MD9L=<) L@9L) MHGF) =F?9?AF?)OAL@)L@=)DA?9F)L@=)AF@A:ALGJQ)J=;=HLGJK)AK)H@GKH@GJQD9L=<)

! ~Å! :Q) HGKKA:DQ) 9F) 1J;) >9EADQ) CAF9K=) J=KMDLAF?) J=;JMALE=FL) G>) K=N=J9D) H@GKH@9L9K=K) AF;DM9EADQ) E=E:=JK) 9F<) ,)%ÄFk!"áÇ) HJ=>=J=FLA9DDQ) J=;JMAL) LQJGKAF=) H@GKH@9L9K=) 1&.~r1&.Ä) O@=J=9K) )*0%~) AK) EGJ=) 9KKG;A9L=<) OAL@) L@=) DAHA<) H@GKH@9L9K=) 1&'.) t2=KKE=J) =L) 9Dj) Ä}}Öuj) 'F) LMJFg) L@=K=) H@GKH@9L9K=K) HJ=)L@=)!JCg);k!:D)9F<)!Å%) ;GEHD=Pj) 'F) :9K9D) ;GF>=;L)G>)49Nk~)<=H@GKH@GJQD9LAGF):Q);k!:D)t $64?<3#HCNFj)

) ) $;9EB7!WP]!!'2'+jF>=B:M>=!BGAB;BMHKR!LB@G:EBG@!I:MAP:Riq&`@ENG=!:G=! KH=BG!}{|{r )

! ~Ç! ,)88:@@868G>I6G>B<8E:8:DGCEFO8'1@+Q6FFC8>6G:98E:8:DGCE8A:A7:EF j)'2F+kE=) '2F+k9KKG;A9L=<) 9;LAN9LAF?) J=;=HLGJK) AF;DM<=) L@=) DGO) 9>>AFALQ) $;z0'''9) t!"~Ñug) L@=) @=L=JG)L@=))'0)>9EADQ)t)'0k1u)9F<) L@=) >9EADQ) G>) F9LMJ9D) ;QLGLGPA;ALQ) J=;=HLGJK) t,!0Kuj) 'F) ,)) ;=DDg) !"~Ñ) ;GF>=JK) E9BGJ) >MF;LAGF) AF) F"!!) LG) =DAEAF9L=<) '?%~k;G9L=<) L9J?=L) ;=DDKj) 2@=) AEHGJL9F;=) G>) !"~Ñ) >MF;LAGF) AF) ,)) ;=DD) H=J@9HK) <=EGFKLJ9L=<) :Q) L@=) >9;L) L@9L) AL) AK) L@=) GFDQ) 9;LAN9LAF?) J=;=HLGJ);9H9:D=)G>)E=J=K@DQ)AKGD9L=<) ,)) ;=DD) t7j) 2j) JQ;=KGF) =L) 9Dj) Ä}}Ñuj) !"~Ñ) ;9F) H9AJ) OAL@) :GL@) !"Å ζ) 9F<) $; ε0' γ) @GEG) L@9L) ;9F) 9KKG;A9L=) OAL@) ,)) ;=DD) 9;LAN9LAF?) J=;=HLGJKg) "F.~Äg) O@A;@) O9K) A<=FLA>A=<) AF) ~ááÖ) t*9FA=J) =L) 9Dj) ~ááÜuj) "F.~Ä) O9K) K@GOF) LG) H9AJ) OAL@) +&!k'k:AFkDA?9FGJ) L@=K=) +&!k'kKH=;A>A;) 9;LAN9LAF?) J=;=HLGJK) 9J=) KLADD) HGGJDQ) A<=FLA>A=<)@GO=N=J)GFk?GAF?)J=K=9J;@=K)HGAFL=<)GML)L@=)JGD=K)G>)L@=K=)J=;=HLGJK)AF),)) ;=DD)J=KHGFK=K)AF)&'4)9F<)!+4)NAJ9D)AF>=;LAGFK)9F<)9DKG)AEH9;L)GF),));=DD)=) '2F+k9KKG;A9L=<) J=;=HLGJK) 9J=) L@=) Å) E=E:=JK) G>) L@=) ,!0) >9EADQ) AF;DM<=K) ,)HÇÑ)t,!0~ug),)HÇÇ)t,!0Äu)9F<),)HÅ})t,!0Åu)OAL@),)HÇÑ)O9K)EGKL)O=DD)CFGOF) 9K) 9) H9Fk,)) ;=DD) E9JC=J) 9;JGKK) KH=;A=Kj) FDL@GM?@) ,)HÇÑ) 9F<) ,)HÅ}) ;GFKLALMLAN=DQ) =PHJ=KK=<) GF) 9DD) ,)) ;=DDKg) ,)HÇÇ) GFDQ) MHJ=?MD9L=<) 9>L=J) ,)) ;=DD) 9;LAN9LAGF) tFJFGFg) +9JC=Dg)9F<)+9F<=D:GAE)Ä}}Ñh)4AL9D=)=L)9Dj)~ááÜuj)!QLGHD9KEA;)),)HÇÑ)9F<) ,)HÅ}) ;9F) H9AJ) OAL@) :GL@) !"Å ζ) 9F<) $; ε0' γ) ;@9AFK) :ML) ,)HÇÇ) ;9F) GFDQ) H9AJ) OAL@) "F.~Äj) ,)HÇÑ) 9F<) ,)HÇÇ) 9J=) J=HGJL=<) LG) J=;G?FAR=) NAJ9D) HJGL=AF) DA?9F) DA?9F) HJ=K=FLu) ;9F) :AF<) LG) L@=K=) J=;=HLGJK) OADD) J=IMAJ=) >MJL@=J)KLM)L@=) Ö)KMH=J>9EADQg) Ök&Ñ)HJ=K=FLK) GF)K=N=J9D)LMEGJ);=DD)DAF=K)@9K):==F)K@GOF)LG)K=JN=)9K)9)DA?9F<)>GJ),)HÅ})t J9F6G:98 9CJBFGE:6A8 F>B<8 86F869: j) FDL@GM?@) ?J=9LDQ)

! ~É! EGD=;MD=K) !"Å ζg $; ε0' γ 9F<) "F.~Ä) AF) ,)) ;=DD) AK) ?J=9LDQ) J=K=E:D= G>) !0) 9F<) 2!0) KA?F9DAF?)L@9L)9J=)O=DD)MF<=JKLGG<)AF) )9F<)2);=DDj))1AEHDQ)KH=9CAF?g)=F?9?=E=FL)G>)L@=) ;GJJ=KHGF) L@=) E=E:=J) G>) 1J;) CAF9K=g) L@AK) OADD) AF) LMJF) J=KMDLK) AF) J=;JMALE=FL) G>) L@=) LQJGKAF=) CAF9K=) 1QC) GJ) 8F.kÖ}) LG) L@=) AFLJ9;QLGHD9KEA;) J=;=HLGJ) ;GEHD=Pj) 'F) J=KHGFK=) LG) L@AKg) K=N=J9D)>MJL@=J)) L@=K=) KA?F9DAF?) ;GEHD=P=K) 9;LAN9LAGF) >AF9DDQ) D=9) C=Q) LJ9FK;JAHLAGF) >9;LGJK) G>) L@=) >9EADQ) ,$kF2g) ,$kκ ) 9F<) F.k~) J=KHGFKA:D=) >GJ) ;QLGCAF=) K=;J=LAGF):ML)9DKG)HJ=H9J=)>GJ);QLGLGPA;)?J9FMD=)=PG;QLGKAK):Q)J=9JJ9F?=E=FL)G>)9;LAFk ;QLGKC=D=LGFj)'2F+kE=

) ('0.@<##,)!<>EE!:=!K><>IMHKL!q*:GB>K! }{{Ñr#

,)8 8:@@8 68G>I6G>B<8 E:8:DGCEO8 ,)%\"Q"@.[Z8 8CAD@:K8 6B98 G=:8 U>B9H8:9QF:@;V8 E:8CG>CBP )2@=)!kLQH=)D=;LAF)9;LAN9LAF?)J=;=HLGJ),)%Ä"g)O@A;@):=DGF?)LG)L@=),)%Ä)

! ~Ñ! ?=F=)>9EADQg);9F):=);GFKA<=J=<)GF=)G>)tA>)FGLu)L@=)EGKL)AEHGJL9FL)J=;=HLGJK)AF),));=DD) >MF;LAGF)AF)=DAEAF9LAF?)<9F?=J)L9J?=L);=DDKj)FDL@GM?@)ALK)>MF;LAGF)EGKL)KLMGMF<)AF) γδ2);=DDK)9F<) !"Üã)2);=DDK)t 9M=J)~áááh)(9EA=KGF)=L)9Dj)Ä}}Äuj)'L)@9K)DGF?):==F)KH=;MD9L=<)L@9L),)) ;=DDK) 9J=) 9:D=) G>) J=;G?FAR=) 9F<) CADD) ;=DDK) L@9L) 9J=) MF<=J?GAF?) KLJ=KK) KM;@) 9K) NAJMK) AF>=;L=<) ;=DDK) GJ) LJ9FK>GJE=<) LMEGJ) ;=DDKg) L@AK) @QHGL@=KAK) O9K) J=?9J<=<) 9K) L@=) oAFp) @QHGL@=KAKj) 2@=) >AF) ,)%Ä") 9F<) ALK) J=KH=;LAN=) DA?9FAJE=<) L@AK) L@=GJQ) t%9KK=J) =L) 9Dj) Ä}}Éuj) 'F) @ME9Fg) ,)%Ä")J=;G?FAR=K)Ü)DA?9F),)%Ä") :ML) AF) EGMK=) 9DL=JF9LAN=) KHDA;AF?) J=KMDL=<) AF) Ä) N=JKAGFK) G>) L@=) EGD=;MD=) <=FGL=<) 9K) ,)%Ä"k*)t>GJ)DGF?)AKG>GJEu)=PHJ=KK=<)E9AFDQ)GF)J=KLAF?),));=DDK)9F<),)%Ä"k1)t>GJ) K@GJLu) O@A;@) AFL=J) ,)) ;=DD) 9;LAN9LAGF) t"A=>=F:9;@) =L) 9Dj) Ä}}Äh) %AD>ADD9F) =L) 9Dj) Ä}}Äuj),)%Ä")O9K)FGL)>GMF<)LG):=)9KKG;A9L=<)OAL@)'2F+k;GFL9AFAF?)EGD=;MD=K):ML)) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ('0.@=#,)!<>EE!:Q!q*:GB>K! }{{Ñr!

! ~Ö! AFKL=9<) :AFk;GFL9AFAF?) "F.~}) OAL@) 9F) =P;=HLAGF) G>) ,)%Ä"k1) AKG>GJEK) L@9L) :AF) J=;JMALK) L@=):AF)=AL@=J)HÜÉ)G>).'Å))GJ)%J:Ä)9F<)>AF9DDQ)D=9)K=N=J9D) I6G>B<8 E:8:DGCEO8 0*@+8 ;6A>@L8 E:8:DGCEF j) 1*F+) >9EADQ) J=;=HLGJK) =PHJ=KK=<)GF),));=DDK)AF;DM<=)Ä Ç)t!"ÄÇÇug),2 kF)9F<)!0F!!j)"A>>=J=FL)>JGE)'2F+k 9KKG;A9L=<) J=;=HLGJK) 9F<) ,)%Ä"g) ;QLGHD9KEA;) L9AD) G>) L@=K=) J=;=HLGJK) ;GFL9AFK) 9F) AEEMFGJ=;=HLGJ)LQJGKAF=k:9K=<)KOAL;@)EGLA>)t'21+u)O@A;@);9F):=):GL@)AF@A:ALGJQ)GJ) 9;LAN9LAF?) J=;=HLGJKj) .=J@9HK) L@=) EGKL) O=DD) KLM) L@AK) ?JGMH) AK) !"ÄÇÇ) t9DKG)CFGOF)9K)Ä Çug)KAF;=)Ä Ç)J=;=HLGJ);9F)K=JN=)9)MF;LAGF)GF),));=DDKg)=P9;L) E=;@9FAKE) GF) @GO) Ä Ç) LJ9FK>=J) KA?F9DAF?) LG)

) ('0.᥁ ### ,)!<>EE!<:G!;>!:=!HK!BGAB;BM>=!OB:!MA>!!"}ÄÄ!K><>IMHK!=>I>G=BG@!HG! =j:=:IM>K!FHE>L!q*:GB>K!}{{Ñr!

! Ä{! :AF<) LG) :GL@) 1F.) 9F<) #F2kÄ) tGFDQ) AF) @ME9Fu) AF) O@A;@) 1F.k<=H=F<=FL) 9;LAN9LAGF) J=IMAJ=K)$QF)J=;JMALE=FL)9F<)AFNGDN=<)>=J=FL)9;LAN9LAF?);GEHD=P=K)KM;@)9K)%J:Ä)9F<) *F2k) .*!kγ) 9F<) 49Nk~g) O@=J=9K) AF@A:ALAF?) KA?F9D) J=IMAJ=<) L@=) H@GKH@9L9K=K) 1&.k~rÄg) 1&'.g) !KC) 9F<) L@AK) HJG;=KK) AFNGDN=<) #F2kÄ) 9KKG;A9LAGF) t $64?<3# HCHG u) t GLLAFG) =L) 9Dj) Ä}}}h)!@=F)=L)9Dj)Ä}}Ñh)4=ADD=LL=)Ä}}Ñh)59LRDg)1L=::AFKg)9F<)*GF?)Ä}}}uj) ) ) ,)8 8:@@8 68G>I6G>B<8 E:8:DGCEO8 ",@+Q[j) ",F6) 9;;=KKGJQ) EGD=;MD=k~) tCFGOF) 9K) ",F+k~g) !"ÄÄÑu) AK) 9F) AEEMFG?DG:MDAF) t'?ukDAC=) EGD=;MD=) AF) ,)) ;=DD) 9F<) GL@=J) ;=DD) LQH=K) O@A;@) J=;G?FAR=) KH=;A>A;9DDQ) !"~~Ä) tF=;LAFkÄu) 9F<) !"~ÉÉ) tHGDAGNAJMK) J=;=HLGJg) .40u)t GLLAFG)=L)9Dj)Ä}}Åh)1@A:MQ9)=L)9Dj)~ááÑuj)0=;=FLDQ)HM:DAK@=<)J=KMDL)AF)1J;)CAF9K=)>9EADQ)9F<) J=;JMALE=FL) G>) %J:Ä) 9F<) GL@=J) ) ALK) KA?F9DAF?) >MF;LAGFg) ",F+k~) 9DKG) HJGEGL=K) ;=DD) 9<@=KAGF) :Q) H9JLA;AH9LAF?) OAL@) 9FGL@=J) 9<@=KAGF) EGD=;MD=) *$Fk~) L@=J=>GJ=) HJGEGL=) AEEMF=) KQF9HK=) >GJE9LAGF) :=LO==F) ,)) ;=DD) LG) L9J?=L) ;=DDK) t1@A:MQ9) =L) 9Dj) Ä}}Åuj) #PHJ=KKAGF) G>) ",F+k~) GF) ,)) ;=DD) MF<=JDAF=<) 9F) 9DL=JF9LAN=) E9LMJ9LAGF) H9L@O9QK) ?GN=JFK)@ME9F),));=DD)>MF;LAGFK)t+9JLAF=L)=L)9Dj)Ä}~Éuj) ) ) *$@Q[8 6B98 CG=:E8 8C68G>I6G>B<8 E:8:DGCEFP8 FK) E=FLAGF=<) 9:GN=g) ,)) ;=DDK) J=IMAJ=) >GJE9LAGF) G>) L@=) AEEMFGDG?A;9D) KQF9HK=) OAL@) L@=AJ) L9J?=LK) >GJ) =>>A;A=FLDQ) H=J>GJE) ;QLGLGPA;ALQ) 9F<) 9NGAGJE9LAGF) HJGEGL=) :Q) *$Fk~) t9DKG) CFGOF) 9K) !"~Üg) AFL=?JAF) βkÄu) t&KM) =L) 9Dj) Ä}~Ñuj) ,)) ;=DD) AK) MFAIM=) AF) ALK) ;9H9:ADALQ) LG) HJGEGL=) 9<@=KAGF) :Q) *$Fk~) :AFGDDGO=<) 9>L=J) GL@=J) ;Gk 9;LAN9LAF?)J=;=HLGJK)9K)K==F)AF)GL@=J);=DD)LQH=K)KM;@)9K)2);=DD)t59LRD)9F<)*GF?)Ä}~}uj) $MJL@=JEGJ=g) :AF) *$Fk~) LG) '!F+k~) 9DGF=) AK) KM>>A;A=FL) LG) AF=Og)9DL@GM?@)L@=)DAKL) ?G=K) GF) :ML) L@=) E=;@9FAKE) G>) L@=K=) J=;=HLGJK) GF) ,)) ;=DD) >MF;LAGFK) 9J=) KLADD) D9J?=DQ) MF=PHDGJ=<)9F<)L@=J=>GJ=)J=DQ)GF)L@=)F==<)>GJ)>MLMJ=)OGJCKj) )

! Ä|! dZh !'GO!7G8K!4!,)!68DD!C?DD!?LK!L4J=8LX! ) =AF?)<=N=DGH=<)>JGE)L@=)K9E=)HJG?=FALGJg)<=KHAL=)L@=)N9KL)>=J=F;=)GF)L@=) L9J?=L) ;=DD) J=;=HLGJ) J=;G?FALAGFg) ;QLGLGPA;) 2) DQEH@G;QL=) t!2*u) 9F<) ,)) ;=DD) 9J=) KLJACAF?DQ)KAEAD9J)GF)@GO)L@=Q)LJA??=J)L9J?=L);=DD)9HGHLGKAKj)+9AFDQg),));=DDK)9F<)!2*) E=>=J=FL) E=;@9FAKEKi) ;QLGLGPA;) ?J9FMD=) =PG;QLGKAKg) 9F<) <=9L@)DA?9FJGE)L@=)2,$)KMH=J>9EADQ)E=E:=JKj)FDL@GM?@) GF=)OGMD<)FGL)=P;DM<=)L@=)AFLJAFKA;)HJGk9HGHLGLA;)F9LMJ=)G>)K=;J=L=<);QLGCAF=K)KM;@)9K) '$, γ)9F<)2,$ α):Q),));=DD)9F<)!2*g)H=J@9HK)AF)L@=)J=9D)H@QKAGDG?A;9D);GFL=PLg)9DD)L@=K=) E=FLAGF=<)E=;@9FAKEK);GFLJA:ML=<)AFLG)L9J?=L);=DD)<=9L@j) ) !LGCGCK>88 B9H8>B<8 8:@@8 9:6G=P ) +9LMJ=) ,)) ;=DDK) 9J=) >=9LMJ=<) OAL@) HJ=>GJE=<) K=;J=LGJQ) ?J9FMD=K) ;GFL9AFAF?) L@=) HGJ=k>GJEAF?) HJGL=AF) tH=J>GJAFg) .0$~u) K=JAF=) HJGL=9K=K)CFGOF)9K)?J9FRQE=K)t%JRuj)2@=J=)9J=)É)LQH=K)G>)%JR)AF)@ME9F):ML)GFDQ)%JRk F)9F<) )9J=)EGKLDQ)KLMGJ) AF) >MJL@=J) AFN=KLA?9LAGF) t GN=FK;@=F) 9F<) )MEE=J) Ä}~}h) .9J9;=) tAj=g) AEEMFGDG?A;9D) KQF9HK=u) :Q) 9;LAF) J=9JJ9F?=E=FL) 9F<) J=D=9K=) =>>=;LGJ) EGD=;MD=Kj) 'F) LMJFg) H=J>GJAF) >GJEAF?) @GD=K) AF) L@=) L9J?=L) ;=DD) E=E:J9F=K) O@A;@) =F9:D=) =FL=JAF?) G>) %JR ) AFLG) L@=) ;QLGKGD) O@=J=) AL) ;9F) ;D=9N=) HJGk9HGHLGLA;) ;9KH9K=K) D=9) L@=) ;9KH9K=kE=) L@=) ?J9FMD=K) 9J=) =PHGK=<) LG) L@=) HD9KE9) E=E:J9F=) O@A;@) ;9F) :=) K=JN=<) 9K) 9) E9JC=J) G>) J=;=FL)<=?J9FMD9LAGFj) ) ":6G=Q@><6B98A:9>6G:988:@@89:6G= j) =KA<=K)<=?J9FMD9LAGF)E=;@9FAKEg),));=DDK)9DKG) =PHJ=KK) <=9L@) DA?9F) L@=) 2,$) >9EADQ) FGL9:DQ) $9KkDA?9F<) t$9K*u) 9F<) 2,$kJ=D9L=<) 9HGHLGKAK)AFGJ)AF

! Ä}! ) ('0.᥂ ###1NFF:KR!H?!,)!<>EEjF>=B:M>=!! >QH:MA!K><>IMHKo=>:MA!EB@:G=!I:MAP:RL!q+:KM[G>Sj*HLM:Hf!LG>Ef!:G=!.:K=H! }{|År!

) ) :GL@) $9K*) 9F<) <=?J9FMD9LAGF) 9J=) EGJ=) AEHGJL9FL) >GJ) E9LMJ=) ,)) ;=DDkE=) !9Äã) AF>DMP) tO@A;@) AK) =KK=FLA9D) >GJ) ?J9FMD=K) EG:ADAR9LAGFu) MF<=JDAF=<) L@=) JGD=) G>) <=9L@) J=;=HLGJK) H9L@O9Q) 9K) 9F) 9DL=JF9LAN=) E=;@9FAKE)G>),));=DD);QLGLGPA;ALQj)"GOFKLJ=9E)KA?F9DAF?):Q)<=9L@)J=;=HLGJ)D=9>) =L) 9Dj) ~ááÜuj) 'F) KMEE9JQg) MHGF) L9J?=L) ;=DD) J=;G?FALAGF) :Q) 9HHJGHJA9L=) 9;LAN9LAF?) J=;=HLGJg) ,)) ;=DD) ;9F) E=

! Ä~! <=?J9FMD9LAGF) 9F<) <=9L@) DA?9F8!?EEMF8!6GFLJGD!GA=<) MF;LAGF)AK)9);JM;A9D)H9JL)AF)9FLANAJ9D) AEEMF=) J=KHGFK=j) 2@AK) O9K) ;D=9JDQ) <=EGFKLJ9L=<) AF) ;9K=) KLMA;A=F;Q) 9>>=;L=<) K=N=J=DQ) :Q) NAJMK) AF>=;LAGFKg) H9JLA;MD9JDQ) L@GK=) G>) @=JH=K)>9EADQ)t AJGFg) QJGFg)9F<)1MDDAN9F)~áÜáh)+9;=)9F<)-J9F?=)Ä}~Ñuj),));=DD)9;LANALQ) =;LAGF);DGK=DQ)DAFC=<)LG)"!)AF)J=;AHJG;9D)E9FF=Jg)O@=J=)LQH=)~)'$,)9F<) HJGkAF>D9EE9LGJQ);QLGCAF=K)t'*k~Äg)'*k~Ég)'*k~Üu)HJGGJ)"!)E9LMJ9LAGF)9F<)>MF;LAGFKj)'*k~É) AF)KGDM:D=)>GJE)GJ) GE6BF kHJ=K=FL=<):Q)"!k<=JAN=<)'*k~É0x);@9AFK)?J=9LDQ)=F@9F;=K),)) ;=DD)KMJNAN9D)9F<)HJGDA>=J9LAGFj)) ) .=J@9HK)L@=)EGKL)O=DD)KLM),));=DDkE=JGE)EGMK=);QLGE=?9DGNAJMK)t+!+4u)G>)L@=)@=JH=KNAJMK)>9EADQj)'F)L@AK)EG<=D)AL)O9K) >AJKLDQ) <=EGFKLJ9L=<) L@9L) EGMK=) ,)) ;=DDK) =PHJ=KK) 9) NAJ9D) KH=;A>A;) J=;=HLGJ) tF9E=<) *QÇá&u)L@9L)J=;G?FAR=)L@=)+!+4)=F;G<=<)HJGL=AF)E~ÉÖ)tFJ9K=)=L)9Dj)Ä}}Äh)1EAL@)=L)9Dj) Ä}}Äuj)5@Q)+!+4)9>>GJ=;LAGF) EG<=D) AL) O9K) D9L=J) HJ=K=FL=<) >AJKL) =NA<=FL) >GJ) ,)) ;=DD) 9<9HLAN=) >=9LMJ=) L@9L) OA<=DQ) 9;;=HL=<)AK)9);@9J9;L=JAKLA;)9LLJA:ML=<)GFDQ)LG);=DDK)G>)L@=)9<9HLAN=)AEEMF=)J=KHGFK=) KM;@)9K)2)9F<) );=DD)t1MFg) =ADC=g)9F<)*9FA=J)Ä}}áuj)2@=J=>GJ=g),));=DD)J=E=E:=JKj)2@AK) J9AK=<) L@=) F=PL) IM=KLAGF) O@=L@=J) L@=J=) AK) 9F) =IMAN9D=FL) :=@9NAGJ) G>) @ME9F) ,)) ;=DD) 9?9AFKL)&!+4g)L@=)9FKO=J)AKi)HGKKA:D=j)1LMA;) HGHMD9LAGF) G>) @ME9F) ,)) ;=DDK) OAL@) !"áÇk,)%Ä!ã) H@=FGLQH=) AK) HJ=>=J=FLA9DDQ) =PH9F<=<)9F<)H=JKAKL)t%MET)=L)9Dj)Ä}}Çg)Ä}}Ñh)*GH=Rk4=J?=K)=L)9Dj)Ä}~~uj)3*~Ñk:AFGMF<)LG):=)9)DA?9F<)>GJ),)%Ä")AK)GF=)G>)L@=)L9J?=L)G>) &!+4) =F;G<=<) HJGL=AF) t3*~Ñug) L@AK) HJGL=AF) O9K) MK=<) :Q) L@=) NAJMK) 9K) 9F) =N9KAGF)

! ÄÄ! KLJ9L=?Q)>GJ)9NGAGMF<)LG)L9J?=L)LG) +'!kFg)9FGL@=J),)%Ä")DA?9F>=J=FL)9DL=JF9LAN=),)k <=H=F<=FL) AEEMF=) =N9KAGF) KLJ9L=?A=K) >GJ) KM;;=KK>MD) @GKL) AF>=;LAGF) t!@9DMHFQ) =L) 9Dj) Ä}}Ñh)1L=JFk%AFGKK9J)=L)9Dj)Ä}}Öuj)FK)E=FLAGF=<)9:GN=g)L@=)9FLANAJ9D)JGD=)G>),));=DD)AK) FGL)GFDQ)J=KLJA;L=<)LG)&!+4):ML)9DKG)=PH9F<)LG)GL@=J)NAJMK=K)G>)L@=)@=JH=K)>9EADQ)DAC=) &14)9F<)# 4g)=L;j)) ) -F=) G>) L@=) J=HGJL=<) DA?9F<) >GJ) L@=) ,)) ;=DD) J=;=HLGJ) ,)HÇÑ) 9F<) ,)HÇÇ) AK) AF>DM=FR9)&F)L@=J=>GJ=)AL)AK)J9LAGF9D)LG)9J?M=),));=DD)@9K)9)JGD=)AF)AF>DM=FR9)AF>=;LAGF) ;GFLJGDj)F;;GJ)AF>DM=FR9kAF>=;L=<)"!)9F<)L@AK)9;LAN9LAGF)>MJL@=J)HJ=N=FL=<):Q),)%Ä")9F<),)HÇÑ) :DG;CAF?) t"J9?@A) =L) 9Dj) Ä}}Öuj) $MJL@=JEGJ=g) ,)) ;=DDK) 9J=) >GMF<) LG) :=) J=;JMAL=<) LG) AF>DM=FR9kAF>=;L=<) EA;=nK) DMF?) 9F<) ,)HÇÑkrk) ,)) ;=DDK) ;9FFGL) J=K;M=) AF>=;L=<) EA;=) t%9RAL)=L)9Dj)Ä}}Ñuj) ) 'EEMFG?=F=LA;) KLM=;L=<)t$DGJ=Kk4ADD9FM=N9)=L)9Dj)Ä}}~h))@9CGG)=L)9Dj)Ä}}Çuj),GL9:DQg))'0Å"1~)9F<) &*Fk OÇ) 9DD=D=Kk;9JJQAF?) H9LA=FLK) 9J=) :=LL=J) 9L) <=D9QAF?) F'"1) KQEHLGEKg) )'0Å"1~ã) ,))AK)9DKG)K@GOF)LG);GFLJGD)&'4)J=HDA;9LAGF)AF)&*Fk OÇ)AF>=;L=<)L9J?=L);=DDK)AF)NALJG) tFDL=J)=L)9Dj)Ä}}Öh)!G@=F)=L)9Dj)~áááuj)$AF9DDQg)&'4k=F;G<=<)HJGL=AF),=>)O9K)>GMF<)LG) :DG;C),)%Ä")DA?9F),)) ;=DD)9K)9)K=D=;LAN=)HJ=KKMJ=)>GJ)&'4)AEEMF=)=N9KAGF)t!=J:GFA)=L)9Dj)Ä}}Öuj) ) dZj !1MEGJ!?EEMFGKMJN8?DD4F68!8!,)!68DD!H8JKH86L?N8Y!LMEGJ!?EEMF8! 8N4K?GF!KLJ4L8=?8K! ) ) 2@=) ;GF;=HL) G>) LMEGJ) AEEMFGKMJN=ADD9F;=) O9K) >AJKL) HJGHGK=<) :Q) .9MD) #JDA;@g) KM??=KLAF?)9)FGN=D)JGD=)G>)L@=)AEEMF=)KQKL=E)FGL)GFDQ)AF);GFLJGD)G>)AF>=;LAGF):ML)9DKG) AF) ;GFLJGD) G>) KHGFL9F=GMK) LMEGJ) >GJE9LAGFj) 2@AK) A<=9) O9K) D9L=J) >GJE9DAR=<) AFLG) 1AJ) +9;>9JD9F=) MJF=LnK) @QHGL@=KAK) G>) o;9F;=J) AEEMFGKMJN=ADD9F;=p) :Q) @AK) GOF) OGJ

! ÄÅ! )

U'B8 @6E<:N8 @CBI:98 6B>A6@FN8 @>?:8 ACFG8 C;8 G=:8 J6EAQ7@CC9:98 I:EG:7E6G:FN8 >B=:E>G67@:8 <:B:G>888=6B<:F8AHFG87:88CAACB8>B8FCA6G>888:@@F86B9868DECDCEG>CB8C;8G=:F:88=6B<:F8J>@@8 E:DE:F:BG868FG:D8GCJ6E98A6@>F86B8:IC@HG>CB6EL8B:8:FF6EL8G=6G8G=:E:8F=CH@987:8 FCA:8A:8=6B>FA8;CE8:@>A>B6G>B<8CE8>B68G>I6G>B<8FH8=8DCG:BG>6@@L896B<:ECHF8AHG6BG88:@@F8 6B98>G8>F8DCFGH@6G:98G=6G8G=>F8A:8=6B>FA8>F8C;8>AAHBC@C<>86@88=6E68G:EPV8 ) ) 2@AK) =PLJ9GJJGE)L@=)J=KMDLK)G>) '$, γkrk) 9F<) 0F%krk) CFG;CGML) EA;=) EG<=DK) 9F<) L@=) ) LMEGJk9KKG;A9L=<) 9FLA?=FK) t2FFuj) 1AF;=) ,)) ;=DD) 9J=) CFGOF) EM;@) D9L=J) L@9F) 2) ;=DDKg) L@=) E9AF) =>>=;LGJ) LQH=)G>);=DDK)L@9L)O=J=)J=?9J<=<)9K)9FLAkLMEGJ);=DDK)O=J=)2);=DDKj)&GO=N=Jg)AL)AK)KGGF) J=9DAR=<)L@9L),));=DD)9F<)!"Üã)2);=DDK)9J=)L@=)Ä)9JEK)G>)L@=)AEEMF=)KQKL=E)AF)9FLAk LMEGJ) =>>=;Lj) 'F) :GL@) @ME9F) 9F<) EGMK=g) ,)) ;=DD) <=>A;A=F;A=K) 9J=) D9J?=DQ) 9L) L@=) HGL=FLA9D) =PHGK=) G>) MF;GFLJGDD=<) NAJMK) AF>=;LAGF) :ML) 9DKG) 9J=) 9) @A?@=J) JAKC) G>) LMEGJ) <=N=DGHE=FL) t-J9F?=) Ä}~Åuj) 2J9FK?=FA;) EA;=) OAL@) ,)%Ä") <=HD=LAGF) K@GO) @A?@=J) ;@9F;=) G>) HJAE9JQ) LMEGJ) <=N=DGHE=FL) t%M=JJ9) =L) 9Dj) Ä}}Üuj) 'F) L@=K=) EA;=g) HJAE9JQ) LMEGJ);=DDK)O=J=)>GMF<)LG)=PHJ=KK),)%Ä")DA?9F),)%Ä"k=PHJ=KKAF?);=DDK)AF);GFLJGDDAF?)LMEGJ)?JGOL@) tE9BGJALQ) G>) ,)) ;=DDK) 9F<) KE9DD) KM:K=LK) G>) 2) ;=DDKuj) 2MEGJ) E=JGE),)%Ä"r,)%Ä"kDA?9F<)t,)%Ä"*u):Q)EMDLAHD=)E=;@9FAKEK)AF)O@A;@)FGL9:DQ):Q) K@=<),)%Ä"*)>JGE)L@=AJ)E=E:J9F=)KMJ>9;=):Q)E9LJAP)E=L9DDGHJGL=9K=K)t++.u) GJ)L@JGM?@)=PGKGE=K)t19DA@g)09EE=FK==g)9F<)1L=AFD=)Ä}}Äh)5M)=L)9Dj)Ä}}Çuj)'F)DAF=K) OAL@)L@=K=g)KGDM:D=),)%Ä"*K)O=J=)<=L=;L=<)GF)K=J9)G>)H9LA=FLK)OAL@)N9JAGMK);9F;=JK) t$=JFTF<=Rk+=KKAF9) =L) 9Dj) Ä}~}uj) #F;GMFL=JAF?) OAL@) KGDM:D=) ,)%Ä"*) ;9MK=K) ,)%Ä") GJ=) ;GMD<) >MJL@=J) KMHHJ=KK) ,)%Ä"k E=9;L)HJGEGL=),));=DDkE=>=;L) G>) GL@=J) K@=<

! ÄÇ! 2MEGJ)EA;JG=FNAJGFE=FL)AK)OA<=DQ)9;;=HL=<)LG)KMHHJ=KK)L@=)=>>=;LGJ)>MF;LAGFK)G>):GL@) ,));=DD)9F<)!2*)L@JGM?@)N9JAGMK)E=;@9FAKEKj)2MEGJ)EA;JG=FNAJGFE=FL)AK)LQHA;9DDQ) ;@9J9;L=JAR=<):Q)@9NAF?)KLJGF?)AEEMFGKMHJJ=KKAN=)=>>=;L):Q);QLGCAF=K)9F<)E=L9:GDAL=K) K=;J=L=<) :Q) L@=) LMEGJ) ALK=D>) 9F<) LMEGJk9KKG;A9L=<) HGHMD9LAGFK) KM;@) 9K) J=?MD9LGJQ) 2) ;=DDK)t2J=?Kug)EQ=DGA)L@=)59J:MJ?)=>>=;Lg)L@AK)AF)LMJF) KMHHJ=KK)LMEGJkAF>ADLJ9LAF?),));=DDK)9F<)!2*)=>>=;LGJ)>MF;LAGFK):Q)@QHGPA9)9F<)D9;L9L=k E=>=;L)t4ADD9D:9)=L)9Dj)Ä}~Åuj)-L@=J)LMEGJk<=JAN=<)E=L9:GDAL=)HJG>=;LGJ)>MF;LAGFK)t aLL;@=J)=L)9Dj)Ä}~Üh)4ADD9D:9)=L)9Dj)Ä}~Åuj)) ) 2MEGJ) ;=DDK) HJ=>=J=FLA9DDQ) J=;JMAL=<) AEEMF=) ;=DD) HGHMD9LAGFK) OAL@) KMHHJ=KKAN=) H@=FGLQH=Kg)AF)O@A;@)2J=?K)9J=);GEEGFDQ)9KKG;A9L=<)OAL@)LMEGJ)HJG?J=KKAGFj),GL9:DQ) 2J=?K)9J=)FGL)9>>=;L=<):Q)DGO)?DM;GK=);GF;=FLJ9LAGF)AF)L@=)LMEGJ)EA;JG=FNAJGFE=FL) KAF;=)L@=Q)E9AFDQ)MLADAR=)>9LLQ)9;A)=F=J?Q)t+A;@9D=C)=L)9Dj) Ä}~~uj) 2J=?K) 9J=) H@=FGLQHA;9DDQ) ;@9J9;L=JAR=<) :Q) !"Çã!"ÄÉ @A $GP.Åã) AF) O@A;@) !"ÄÉg) L@=) @A?@) 9>>AFALQ) '*kÄ) J=;=HLGJ) 9DH@9) ;@9AF) AK) ;GFKLALMLAN=DQ) =PHJ=KK=<) L@=J=>GJ=) ;9F) AFL=J>=J=) L@=) !"Çã) =>>=;LGJ) ;=DDK) 9F<) ,)) ;=DDK) ;JGKKL9DC) NA9) '*kÄ) K=IM=KLJ9LAGFj) F;;MEMD9LAGF)9F<)=PH9FKAGF)G>)2J=?K)9J=)>GMF<)LG):=)9KKG;A9L=<)OAL@)>=J=FL)LQH=K)G>) ;9F;=J) 9F<) ;GJJ=D9L=) OAL@) HGGJ=J) HJG?FGKLA;) t$M) =L) 9Dj) Ä}}Öh) -JE9F>=;L)NA9)L@=)2%$ky) H9L@O9Q) 9DL@GM?@) 2%$ky) AK) 9DKG) OA<=DQ) K=;J=L=<) :Q) GL@=J) AEEMF=KMHHJ=KKAN=) HGHMD9LAGFK) KM;@) 9K) +"1!) 9F<) LMEGJk9KKG;A9L=<) E9;JGH@9?=K) t2F+uj) !GDD=;LAN=DQg) L@=K=) LMEGJk9KKG;A9L=<) HGHMD9LAGFK) 9J=) ;GEEGFDQ) J=;JMAL=<) OAL@AF) L@=) LMEGJ) EA;JG=FNAJGFE=FL) L@9L) NA9) L@=AJ) N9JAGMK) Q=L) GN=JD9HHAF?) AEEMFKMHHJ=KKAN=) E=;@9FAKEK) ;GFLJA:ML=) LG) L@=) AEEMF=) =K;9H=) >JGE) =>>=;LGJ) ;QLGLGPA;) DQEH@G;QL=K) AF;DM>=;LK) K@GO=<) LG) :=) HJGEAKAF?) AF) ;9F;=J) AEEMFGL@=J9HQg)GF=)G>)O@A;@)AK)L@=)=P)NANG)=PH9FKAGF)HJGLG;GDK)G>)9;LAN9L=<),));=DDKj) )

! ÄÉ! ) ) ) ) ) ) # # # # # # # # # # # # # # ('0.᥃## #2NFHK!<>EEL!<:G!=>O>EHI!O:KBHNL!BFFNG>!>O:LBHG!LMK:M>@B>L!MH!:OHB=!,)! <>EE!F>=B:M>=!:MM:K!}{|Çri!

! dZk !!'#&!&#!8PH4FK?GF!G8J4HQ! ! FDL@GM?@) ,)) ;=DD) HGKK=KK=K) AFLJAFKA;) 9FLAkLMEGJ) HGL=FLA9Dg) Q=L) L@=Q) >9AD) LG) ;GFLJGD) LMEGJ) ?JGOL@) AF) ;9K=K) G>) ;9F;=J) H9LA=FLKj) 'F<==ADLJ9LAF?) ,)) ;=DDK) >JGE) H9LA=FLK) OAL@) ;9F;=J) HJG?J=KKAGF) MF;LAGF9D) ;@9J9;L=JAKLA;K) 9F<) 9:FGJE9D) H@=FGLQH=K) 9K) 9) ;GFK=IM=FL)G>)N9JAGMK)LMEGJkAFJ=K@DQ) AKGD9L=<)F9\N=),));=DDK)=P@A:AL)DGO)F9LMJ9D);QLGLGPA;ALQ)9F<)GFDQ)J=HJ=K=FL)9)EAFGJ))

! ÄÑ! ) ('0.᥄####LNMHEH@HNL!HK!:EEH@>G>BEE!:=HIMBO>!MK:GL?>K!<:G!;>!:O>=!;R! =B??>K>GM!:IIKH:L!H?!>Q!OBOH!>QI:GLBHGo:@:BG!MA>!:GMBjMNFHK!IHM>GMB:E! H?!,)!<>EE!p;:L>=!BFFNGHMA>K:IRf!:=:IM>=!?KHF!q%NBEE>K>Rf!&NGMBG@MHGf!:G=!1FRMA! }{|Çri!

KM:K=L) OAL@AF) LGL9D) HGHMD9LAGF) G>) H=JAH@=J9D) :DGG<) EGFGFM;D=9J) ;=DDK) t. +!uj) 2@=J=>GJ=g) :K8 I>IC ) =PH9FKAGF) KLJ9L=?A=K) 9J=) J=IMAJ=<) AF) GJ<=J) LG) J=k?9AF) 9FLAkLMEGJ) 9;LANALQ)G>),));=DDK)AF),));=DD):9K=<)AEEMFGL@=J9HQ)9HHJG9;@=K)t$64?<3#HCHJ uj)) ) 0CHE8:F8 C;8 6@@C<:B:>88 ,)8 8:@@F j) !AJ;MD9LAF?) ,)) ;=DDK) ;9F) :=) =>>=;LAN=DQ) AKGD9L=<) >JGE) $A;GDD) :M>>=J) HJ=H9J9LAGF) G>) @=9DL@Q) ),));=DD)>GJ) :K8I>IC )=PH9FKAGFj)!GEH9J=<)LG) ,)) ;=DDK) <=JAN=<) >JGE) . +!g) 3! ) EGFGFM;D=9J) HGHMD9LAGFK) ;GFL9AFK) EGJ=) ,)) ;=DD) OAL@)AEE9LMJ=)H@=FGLQH=K)@=F;=)?J=9L=J)HJGDA>=J9LAGF)HGL=FLA9D)t!GFMJL@=J):=)>=J=FLA9L=<)AFLG)>MDDQ)E9LMJ=),));=DDK)AF)NALJGj)$AF9DDQg)N9JAGMK)KL=E);=DDK) AKGD9LAGF) 9F<) ?=F=J9LAGF) E=L@G>A;A=F;Q) 9F<) EMDLAHD=) AF) NALJG)

! ÄÖ! E9FAHMD9LAGF) KL=HK) AFNGDN=<) AF) ?=F=J9LAF?) L@=K=) ;=DDK) ;GMD<) E9C=) L@=E) D=KK) 9HHJG9;@9:D=)L@9F). +!)9F<)3! k<=JAN=<),));=DDKj) ) !LGC?>B:F8;CE8,)88:@@8>B8I>GEC8:KD6BF>CB j)2@=)>AJKL)MK=)G>);QLGCAF=)LG)?=F=J9L=),)) ;=DDK)>GJ);DAFA;9D)HMJHGK=)O9K)9HHDA=<)N=JQ)=9JDQ)KAF;=)~áÜ}):Q)MKAF?)G>)@A?@)JGE)9MLGDG?GMK). +!)t0GK=F:=J?) =L) 9Dj) ~áÜÉuj) 'L) O9K) KGGF) D9L=J) J=9DAR=<) L@9L) L@=) LMEGJ) J=B=;LAGF) 9;LANALQ) :Q) L@AK) '*kÄ) 9;LAN9LAGF)E=L@G<)O9K)E9AFDQ)9LLJA:ML=<)LG),));=DD)t.@ADDAHK)=L)9Dj)~áÜÖuj)&GO=N=J)L@=) E=L@G<)LG)?=F=J9L=)*F));=DDK)9J=)HJ=LLQ)EM;@)@AF<=J=<):Q)2);=DD)HGHMD9LAGFK)O@A;@) 9DKG)J=IMAJ=)'*kÄ)>GJ)L@=AJ)=PH9FKAGF)9F<)KMJNAN9Dg)=KH=;A9DDQ)>GJ)2J=?K))!"ÄÉj)FDL=JF9LAN=DQg)2);=DD)HGHMD9LAGF);9F):=)<=HD=L=<)>JGE)L@=). +!) EAP)LG)9;IMAJ=)@A?@=J)HMJALQ)G>)=PH9F<=<),));=DDK)OAL@GML);QLGCAF=);GEH=LALAGFj)'*k~É) AK) 9FGL@=J) AEHGJL9FL) HJGEGL=J) G>) ,)) ;=DD) >=J=FLA9LAGFg) KMJNAN9D) 9F<) @GE=GKL9KAK) t!GGH=J) =L) 9Dj) Ä}}Äuj) ,GL9:DQg) '*k~É) HJGEGL=) ,)) ;=DD) KMJNAN9D) NA9) L@=) 9FLAk9HGHLGLA;) +!*k~)9F<)L9J?=LAF?)'*k~ÉkE=MF;LAGF) G>) ,)) ;=DD) AF) NANG) t!9JKGF) =L) 9Dj) ~ááÖh) "=D;GFL=) =L) 9Dj) Ä}~Ñh) &MFLAF?LGF) =L) 9Dj) Ä}}Öuj) !GE:AF9LAGF) LJ=9LE=FL) G>) '*k~Äg) '*k~É) 9F<) '*k~Ü) @9K) :==F) K@GOF) LG) ?=F=J9L=) 9) DGF?kDAN=<) HGHMD9LAGF) G>) ,)) ;=DDK) OAL@) E=EGJQkDAC=) HJGH=JLA=K) t*=GF?) =L) 9Dj) Ä}~Çuj) 0=;=FL) KLM) F=ODQ) A<=FLA>A=<) E=E:=J)G>);GEEGFk?9EE9);@9AF)J=;=HLGJ);QLGCAF=Kg)'*kÄ~)AF),));=DD)9;LAN9LAGF)9F<) L@=)HGL=FLA9D)MK=)G>)'*kÄ~)AF)KQF=J?Q)OAL@)'*kÄ)9F<)'*k~É)t-R9CA)=L)9Dj)Ä}}}h)<=)0@9E)=L) 9Dj) Ä}}Öuj) &GO=N=Jg) ?=F=J9DDQ) ,)) ;=DD) ?=F=J9L=<) :Q) ;QLGCAF=) KLAEMD9LAGF) 9DGF=) ?AN=K) DGO=J) QA=D<) G>) =PH9FKAGF) AF) ;GEH9JAKGF) OAL@) =PH9FKAGF) E=L@G<) MKAF?) :GL@) ;QLGCAF=) 9F<)>==<=J);=DDj) ) /C@:8C;8;::9:E88:@@F8>B8,)88:@@8:KD6BF>CB8DECGC8C@F j) =KA<=K);QLGCAF=Kg),));=DD)9DKG) J=IMAJ=K) ;GkKLAEMD9LGJQ) KA?F9DK) >JGE) KMJJGMFGJ) KMJNAN9D) 9F<) HJGDA>=J9LAGFj) !GE:AF9LAGF)9HHJG9;@=K)MKAF?):GL@)KLAEMD9LGJQ);QLGCAF=K)KM;@)9K)'*kÄr'*k~É)LG?=L@=J) OAL@)AJJ9==<=J);=DDK)K==E)LG)=F@9F;=),));=DD)=PH9FKAGF)J9L=);GEH9J=K)LG)=9;@) G>) L@=) KAF?D=) >9;LGJ) 9DGF=j) 2@=J=) 9J=) >=J=FL) O9QK) G>) MKAF?) >==<=J) ;=DDK) AF) ,)) =PH9FKAGF) HJGLG;GDKg) GF=) OGMD<) :=) AJJ9>=J=FL) ,)) =PH9FKAGF)

! Å{! 9HHJG9;@=Kj) 2@=J=) 9J=) EMDLAHD=) 9) MKAF?) AJJ9==<=J) ;=DDKi) MFDAC=) 9MLGDG?GMK) . +!g) L@=K=) ;=DD) DAF=K) 9J=) N=JQ) J=99;LGJK) t!@ADGJE=<) DQEH@G:D9KLGA<) ;=DD) DAF=K) t# 4k*!*u) ;9F) 9DKG) HJ=>=J=FLA9DDQ) HJGEGL=),));=DD)=PH9FKAGF)9F<)KMJNAN9D)AF)NALJG)t.=JMKKA9)=L)9Dj)~áÜÖuj)# 4k*!*):9K=<) =PH9FKAGF) HJGLG;GDK) @9K) :==F) <=N=DGH=<) LG) HJG) ,)) ;=DDK) 9F<) ;GEH9LA:D=)OAL@)%+.)J=?MD9LAGFK)t =J?)=L)9Dj)Ä}}áh)49KM)=L)9Dj)Ä}~Éuj) ! dZl !!"gh!8PHJ8KK?GF!4F7!4DL8JF4L?N8!KHD?6?F=!?KG);=DDKg)!"ÇÉ)K=JN=<)9K)9) H@GKH@GLQJGKAF=) H@GKH@9L9K=) t.2.9K=u) L@9L) ) 2!0) 9F<) !0) AF) 2) 9F<) ) DQEH@G;QL=K) J=KH=;LAN=DQj) 'F) ,)) ;=DDKg)!"ÇÉ)AK)9DKG)K@GOF)LG);GFLJGD)'2F+)KA?F9DAF?)HJG:9:DQ)L@JGM?@)'2F+k;GMHD=<) 9;LAN9LAF?) J=;=HLGJK) KM;@) 9K) !"~Ñg) ,!0K) GJ) ,)%Ä!r)'0k1) E=E:=JK) G>) +&!k;D9KK) ') :AFA;A=FL) EA;=) J=N=9D=<) ) L@AK) EGD=;MD=) AF) J=?MD9LAF?) ,)) ;=DD) '2F+kKA?F9DAF?i) AFGJ) '2F+kE=A;A=FL),));=DDK)9J=)MF9:D=)LG);GFLJGD)+!+4)AF>=;LAGFg)HJG:9:DQ)NA9)'2F+)AFALA9L=<) KA?F9DAF?) t&=KKD=AF) =L) 9Dj) Ä}~~uj) .JAE9JQ) KA?F9DAF?) J=?MD9LGJQ) >MF;LAGF) G>) !"ÇÉ) ;QLGHD9KEA;) H@GKH@9L9K=) QF) E=E:=JK)G>)L@=)1J;)>9EADQg)L@=J=:Q)9;LAN9LAF?)1QC)9F<)8F.Ö})KA?F9DAF?)) '2F+k;GMHD=<)J=;=HLGJKj)) ) !"ÇÉ) AK) 9) EMDLA>MF;LAGF9D) HJGL=AFg) KAF;=) :=KA<=K) ALK) HGKALAN=) JGD=) AF) J=?MD9LAF?) '2F+) KA?F9DAF?) AF) DQEH@G;QL=) HGHMD9LAGFKg) !"ÇÉ) AK) 9DKG) J=HGJL=<) LG) :=) 9) F=?9LAN=) J=?MD9LGJ)G>)L@=)(9Ck12F2)H9L@O9QK)t19K9CAg)19K9CAk'JA=g)9F<).=FFAF?=J)Ä}}~u)t$64?<3# HCHK uj)2@=)mKH=M)L@=)EGD=;MD=)O9K)>GMF<)LG)9KKG;A9L=<))

! Å|! ) ('0.᥅## #"B!KHE>!BG!!q1$)r! :M!:Ei!}{{|r!

) OAL@)(9Cg)L@AK)J=KMDL=<)AF)<=H@GKH@GJQ9DLAGF)G>)(9C)9F<)AF9;LAN9L=)) !"ÇÉ) D=9) ;QLGCAF=K)KM;@)9K)'*kÅ)9F<)'$,kx)t.=FFAF?=J)=L)9Dj)Ä}}~uj)) ) -F=) G>) L@=) J=E9JC9:D=) >=9LMJ=K) G>) !"ÇÉ) AK) L@=) 9:ADALQ) LG) ?=F=J9L=) EMDLAHD=) AKG>GJEK) L@JGM?@) 9DL=JF9LAN=) KHDA;AF?) G>) !"ÇÉ) HJ=kE0,Fj) 'F) H9JLA;MD9Jg) =PGFK) G>)

! Å}! DG;9LAGF) ÇgÉ) 9F<) Ñ) tLJ9GJE)EMDLAHD=)N=JKAGFK)G>)L@=)!"ÇÉ)EGD=;MD=K)OAL@)>=J=FL);GE:AF9LAGF)AF) L@=) Fg) ) 9F<) !) ;GEHGF=FLKj) FK) 9) J=KMDL) G>) L@AK) =>>=;Lg) HGL=FLA9DDQ) L@=J=) 9J=) LGL9D) Ü) HGKKA:D=) ;GE:AF9LAGFK) :=LO==F) L@=) Fg) ) 9F<) !) J=?AGFK) @GO=N=J) AF) J=9DALQ) L@=J=) 9J=) GFDQ)É)AKG>GJEK)L@9L)9J=)<=L=;L=<)9L)L@=);=DDMD9J)KMJ>9;=)D=N=D)G>)DQEH@G;QL=Kg)L@GK=)9J=i) !"ÇÉ0F !g)!"ÇÉ0F g)!"ÇÉ0 !g)!"ÇÉ0 )9F<)!"ÇÉ0-)tO@A;@)AK)D9;CAF?)G>)9DD)L@=)Fg) ) 9F<) !) J=?AGFKu) t $64?<3# HCHL uj) 5@9L) ;GMD<) :=) L@=) AEHDA;9LAGF) AF) L=JE) G>) >MF;LAGF) :=LO==F) L@=K=) >=J=FL) AKG>GJEKl) 'F) =;LGHA;9DDQ) =PHJ=KKAGF) G>) ;@AE=JA;) EGD=;MD=K) G>) L@=)<=>AF=<)AKG>GJEK)G>)!"ÇÉ)OAL@)L@=);QLGHD9KEA;))!"ÇÉg)2);=DD)9;LAN9LAGF) 9F<)<=N=DGHE=FL)9J=)FGL)9>>=;L=<)KM??=KLAF?)L@9L)L@=).2.9K=))!"ÇÉ)AK)EGKL) =KK=FLA9D)>GJ)2);=DD)>MF;LAGF)>MF;LAGF)t)GRA=J9

) ) ('0.ʆ:## #LEM>KG:MBO>!LIEBK>GM!BLH?HKF!>QIK>LLBHG!HG! ERFIAHLi!:r!@>G>!LMKN!H?!.2.0!!q!"ÄÅr!:G=!>QHGjLIEBjF0,L!

! Å~! :M!EH<:MBHG!ÄfÅ!:G=!Ç!@>G>K:M>!Ñ!IHLLB;E>!BLH?HKFLi!;r!LF:MB!H?!!"ÄÅ! IKHM>BG!PBMA!MA>!EHG@>LM!:G=!LAHKM>LM!BLH?HKFL!:G=!MA>BK!=B??>K>G<>!BG!-jEBGD>=! @ERKGi!K>GM!>QIK>LLBHG!BG!!"ÄÅ!BLH?HKFL!BG!G:]O>o>??>EEi! L=:IM>=!?KHF!q2O>KE>R!}{{Çri!

) O@9L)AK)L@=)>=J=F;=)AF)N9JAGMK)=PLJ9;=DDMD9J))L@=)>=J=FL)AKG>GJEKl)-F=) @QHGL@=KAK) HJGHGK=<) L@=) ) !"ÇÉ) AK) >=J=FLDQ) J=?MD9L=<) )=PLJ9;=DDMD9J)AKG>GJEKg)@GO=N=J)L@=)AEHDA;9LAGF)G>)L@AK)@QHGL@=KAK)KLADD)FGL) Q=L);GEHDQ)OAL@);GFL=EHGJ9JQ)J=KMDLK)AF)2);=DD)KLM)!"ÇÉ)AK)EGKLDQ)KLM=J=FLA9DDQ) =PHJ=KK) L@=) DGF?) AKG>GJE) !"ÇÉ0Fg) MHGF) 9FLA?=FA;) KLAEMD9LAGF) NA9) 2!0) 2) ;=DD) J9HAAD=j) 9K=<)GF)L@AK)H@=FGE=FGFg)!"ÇÉ0F)NK)!"ÇÉ0-)9J=)?=F=J9DDQ)MK=<)LG)>=J=FLA9L=)2) ;=DD)oF9\N=p)GJ)o=>>=;LGJrE=EGJQp)H@=FGLQH=K)AF)L@=)2);=DD)HGGDj)&GO=N=Jg)AF),));=DDKg) L@=) >=J=LA9D) =PHJ=KKAGF) G>) !"ÇÉ) AKG>GJEK) 9F<) ALK) ;GJJ=D9LAGF) OAL@) ,)) ;=DD) >B8 I>IC ) 9;LANALA=K)AF)>=J=FL)H9L@GDG?A;9D);GF8!L>8K?K! ! ,)) ;=DDK) 9J=) ;GFKA<=J=<) >JGFLDAF=) AF) L@=) 9FLAkLMEGJ) >MF;LAGF) G>) L@=) AEEMF=) KQKL=E) Q=L) AF) H9LA=FLK) OAL@) @=E9LGDG?A;9D) E9DA?F9F;A=K) 9F<) KGDA<) LMEGJK) L@=Q) 9J=) MF9:D=) LG) C==H) L@=K=) LJ9FK>GJE=<) LMEGJ) ;=DDK) AF) ;@=;Cg) H9JLDQ) :=;9MK=) G>) LMEGJk AFMF;LAGFKj)5AL@AF)L@=)K;GH=)G>)EQ)L@=KAK)OGJCg)')K=L)LG)AFN=KLA?9L=) L@=)@9DDE9JCK)AF),));=DD)H@=FGLQH=K)9F<)>MF;LAGF9D)9LLJA:ML=K)AF)>=J=FL)H9L@GDG?A;9D) ;GFG;MKg) :ML) FGL) =P;DMKAN=DQg) GF) L@=) =PHJ=KKAGF) G>) !"ÇÉ) AKG>GJEK)GF)L@=),));=DD)HGHMD9LAGFj) ) 'F);@9HL=J)Åg)')J=HGJLg)LG?=L@=J)OAL@)KGE=);GkOGJC=JKg)L@9L),));=DD)HGHMD9LAGFK) OAL@) >=J=FL) !"ÇÉ) AKG>GJE) HJG>AD=g) FGL9:DQ) L@=) AF;J=9K=<) >J=IM=F;Q) G>) !"ÇÉ0Fã!"ÇÉ0-ã) t@=J=9>L=J) !"ÇÉ0F0-) ,)u) AK) DAFC=<) OAL@) K=N=J9D) @=E9LGDG?A;9D)

! ÅÄ! ;9F;=JK)t)JRQOAFKC9)=L)9Dj)Ä}~Ég)Ä}~Ñuj)0=KMDLK)>JGE)L@AK)KLMGJE) G>)HM:DAK@=<)9JLA;D=)AF)L@=)BGMJF9D) #7>CA:9>8>B: j) ) )!@9HL=J)Ç)HJ=K=FL=<)L@=)%F*#,)KLMMF;LAGFK) AF) J=D9HK=<) H9LA=FLK) OAL@) FGFk&G) ;GE:AF9LAGF) LJ=9LE=FL) G>) %F~}~) t9FLAk!"Ä}) EGFG;DGF9D) 9FLA:G) L@AK) KLMGJE) G>) L@=) HM:DAK@=<) 9JLA;D=) AF) BGMJF9D)G>) -B8C>AAHBC@C)>MF;LAGF9D)AEH9AJE=FL)AF)H9LA=FLK)OAL@)N9JAGMK)H9L@GDG?A;9D) ;GF) ,)) ;=DD) HGHMD9LAGF) AF) 9) KE9DD) ;G@GJL) G>) &'4) AF>=;L=<)AFj).A=JJ=)!GMJ:=9Mj) ) 'F) ;@9HL=J) Ög) O=) HJ=K=FL=<) 9) E=L@G<) G>) ?=F=J9LAF?) 9;LN9L=<) ,)) ;=DDK) >B8 I>GEC ) >JGE) @=9DL@Q) GJE) =>>A;A=FL) F"!!) >B8 I>GEC) 9F<) >B8 I>IC ) O@=F) MK=<) AF) ;GE:AF9LGF) OAL@) K=N=J9D) ;DAFA;9Dk 9HHJGN=<)EF:Kj) ) !@9HL=J)Ük~})HJ=K=FL=<)K=EAF9D)J=K=9J;@)9JLA;D=K)GF);9F;=J);=DD)E=L9:GDAKE)L@9L) ') @9N=) K@9J=<) ;Gk9ML@GJK@AH) AF) L@=K=) KLM) L@=) .")~) AF@A:ALGJg) ) *"*0) t!"ÅÑuj) $MJL@=JEGJ=g) AF)!@9HL=J)á)O=)K@GO=<)L@=)#0)Ék+#$Ä)KA?F9DAF?)H9L@O9Q);GFLJGDK)F !)LJ9FKHGJL=JK) AF)EML9FL)HÉÅ);9F;=J);=DDK):ML)FGL)OL)HÉÅ);=DKg)AF>=J=FL)JGD=)G>)L@AK)KLJ=KKk AF) #0)Ék+#$Ä)

! ÅÅ! H9L@O9QK)AF)L@=)EALG;@GFJA9Dk;GEHD=Pk~)<=H=F=FL)9FLAGPA<9FL)J=KHGFK=)NA9)J=?MD9LAGF) G>),0$Ä)=PHJ=KKAGFj)"M=)LG)KH9;=);GFKLJ9AFKg)J=KMDLK)>JGE)L@=K=);@9HL=JK)OADD)FGL):=) >G;MK=<)AF))L@=)J=KMDL)K=;LAGFKj) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! ÅÇ! ) ) ) )

)

)

)

) ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÅÉ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

)

!'".1#/!eZ!

! +"1#/("*0!","!+#1'-"0! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÅÑ! eZd!&"*#,!KLM7QY! ! -3D;7?DC! LEE! I:MB>GML! ;>EHG@! MH! MA>! BH%L*#,! LMN=R! :G=!LB@G>=!LI>=! GM! ?HKF! ;>?HK>! ;BHEH@B<:E! L:FIE>L! =! BG! P>;LBM>! !EBGB<:E2KB:ELi@HO! PBMA! GNF;>K! ,!2{|ÅÑ}ÉÉÇi! .A:L>! ' ! P:L! ?HK! ?HEEBGML! :G=! .A:L>! ''! ?HK! ?HEEBLLBO>! q"* !*! :G=! +!*r! j<>EE! ERFIAHF:!I:MB>GMLi!~!Q!~!FE!H?!A>I:KBGBS>=!;EHH=!HK!Ä!FE!H?!;HG>!F:KKHP!:LIBK:M>!P>K>! =!:M!=:R!{i!!LM!MA>!>G=!H?!?BKLM!!HK!:M!MA>!>G=!H?!MK>:MF>GM!qLNIIE>F>GM:E!$B@i! |r!P>!I:KBGBS>=!;EHH=i! ! Cell culture

The B cell lymphoblastoid Daudi cell line was maintained in logarithmic growth in RPMI 1640 medium (Gibco® GlutaMAX™ media) with 10% fetal bovine serum (FBS) (Gibco®).

Cells were cultured at 37°C in a humidified chamber with 5% CO 2 in air, and passaged 1:10 twice a week. ! Peripheral blood mononuclear cell (PBMC) purification Bone marrow and peripheral blood samples were obtained from patients and total PBMC were isolated using Ficoll. KB>?ERf! ~jÇ! FE! H?! |h}! =BENM>=! ;EHH=! HK! |h~! =BENM>=! ;HG>! F:KKHP! L:FIE>L!BG!0.+'!P>K>!:==>=!HG!MHI!H?!Å!FE!H?!&BLMHI:JN>!q1B@F:ri!!>EEL!P>K>!<>GMKB?N@>=! :M!|Ç{{!KIF!:G=!:M!}{å!!PBMAHNM!;K>:D!?HK!~{!FBGNM>Li!+HGHGN:K!<>EEL!P>K>!=! ?KHF!MA>!BGM>KE:R>K!PABM>!KBG@i!L?M>K!P:LABG@!BG!0.+'f!<>EEL!P>K>!L>KO>=!BG!EBJNB=! GBMKH@>G! BG! F>=BNF! ! H?! $ 1! IENL! |{Ü! j@K:=>! "+1-! q!EBGB+L!1r! NGMBE! :G:ERSBG@i! ! Flow Cytometry analysis 'LHE:M>=!. +!L!P>K>!LM:BG>=!PBMA!ÉLL"!q >GMB?R!OB:;E>!<>EEL!:G=!PBMA!MA>! ?HEEHPBG@! j!"ÄÅ0-j$'2!f! j!"|Ç|j$'2!f! j!"~j.#f! j!"|Öj.#f! j!"Ç}*j.#f! j!"ÇÖj.#f! j !"~|Äq,)%}"rj.#f! j!"~j#!"f! j!"|Öj#!"f! j!"ÅÇj.#!RÉf! !"~jL.!f! j!"ÅÇjL.!f! j%S j LE>Q:$ENHKÉ{{f! j!"|ÖjLE>Q:$ENHKÉ{{f! j!"}{jL.!jLE>Q:$ENHKÉÅ{f! j!"ÄÅ0LjL.!j LE>Q:$ENHKÉÅ{f! j!"Åj.:f! j!"|Çj.:f! j!"ÅÉj.:f! j!"|Çj )KHF>-K:G@>!q >

! ÅÖ! )BjÇÉj4ÄÅ{f!&*LjL !j 4É||!q "! BHLG<>Lrf!+'!jLo j.#f!j!"ÄÅ0Lj$'2!f!j!"ÄÅ0-j.#f! j!"|ÅÖ:q,)%}Lrj.#f!j!"ÖÄj.#j4BHÉÉ{f!j!"ÄÅ0-jL.!f!j!"|Öj4BH EN>f!j!"|ÅÑ>j4BH EN>! q+BEM>GRB! BHM>L! :@:BGLM! LNK?:<>! F:KD>KLi! KB>?ERf! |! MH! |{Q|{ Ç! <>EEL! P>K>! BG=!PBMA!MA>!=B??>K>GM!:GMB;H=B>L!BG!. 1!Li! !>EEL! P>K>! MA>G! P:LA>=! PBMA! . 1! :G=! LNLI>G=>=! BG! }{{j}Å{! çE! . 1! }Ü! $ 1i! $BG:EERf! L:FIE>! :K?HKF>=! NLBG@! %:EEBHL! ?EHP! M>K! q >LL:! q "! BHLG<>Lri! L=! L:FIE>L! P>K>! E:M>K! :G:ERS>=! NLBG@! ):ENS:! LH?MP:K>! OÅi|! q >Interleukin 2 (eBiosciences) and incubated overnight. After stimulation, cell mixture was collected and stained for FACS using an antibody cocktail containing 7AAD, the anti-CD45RO-FITC, -CD69-PE, -CD19-ECD, -CD56- PECy7, -CD3-APC, -CD45RA-APCAlexaFluor750, -CD107a-HV500 and -CD16-KO antibodies (BD Biosciences, Beckman). A bivariate plot of CD56 versus CD3 was used to acquire at least 10,000 NK cells. ! Multicolor Staining for Intracellular Markers Cell permeablization and intracellular staining was performed as previous described (Krzywinska et al. 2016). Briefly,1-10 million cells were incubated with 10% normal human serum at RT for 15 min and then stained with an antibody mix for cell surface markers (anti- CD45RO-FITC, -CD19-ECD, -CD56-PC7, -CD3-APC, -CD45RA-APCAlexaFluor750 and - CD16-KO antibodies) (BD Biosciences, Beckman). After surface staining, cells were washed twice and permeabilize with CytoFix/CytoPerm (BD Biosciences) reagent according to the manufacturer protocol. After fixation and permeablization, cells were washed twice in BD Perm/Wash solution and follow FACS staining for intracellular markers Granzyme B- PE (Miltenyi Biotec) and Ki-67-V450 (BD Biosciences) at 4°C for 30 minutes in the dark. Finally,

! Ç{! cells were washed twice in BD Perm/Wash solution and resuspended in PBS 2% FBS prior to acquisition on flow cytometer Gallios (Beckman). A bivariate plot of CD56 versus CD3 was used to acquire at least 10,000 NK cells. ! Statistical analysis #QI>KBF>GM:E! ?B@NK>L! :G=! LM:MBLMB<:E! :G:ERLBL! P>K>! I>K?HKF>=! NLBG@! %K:IA.:=! .KBLF! qOÇi{ri! 1M:MBLMB<:E! LB@GB?B<:G<>! ;>MP>>G! =:R! {! :G=! MA>! ?HEEHPBG@! MBF>jIHBGML! P:L! =>M>KFBG>=!NLBG@!I:BK>=!1MN=>GM!MjM>LM!HG!MA>!L:FIE>!I:MB>GML!?HK!>:M>KFBG>! LM:MBLMB<:E! LB@GB?B<:G<>! ;>MP>>G! A>:EMAR! =HGHKL! :G=! I:MB>GMLf! HG>jP:R! L,-4L!M>LM!P:L!NL>=!MH!!;>MP>>G!A>:EMAR!=HGHKL!O>KLNL!I:MB>GML!:M!>O>KR!MBF>j IHBGMLi! LEE! LM:MBLMB<:E! O:EN>L! IK>L>GM>=! :L! uh! Iâ{i{Åg! uuh! Iâ{i{|g! uuuh! Iâ{i{{|i! LO>K:@>! O:EN>L!P>K>!>QIK>LL>=!:L!F>:G!IENL!HK!FBGNL!MA>!LM:G=:K=!=>OB:MBHG!q1"ri! ! eZe!"+*!KLM7QY! ! -3D;7?DC! L+*! I:MB>GM! ;EHH=! L:FIE>L! P>K>! =! ?KHF! MA>! !EBGBF:MHEH@R! ">I:KMF>GM! H?! !&3!+HGMI>EEB>K!NG=>K!I:MB>GMlL!GMi!.:MB>GM!;EHH=!L:FIE>L!P>K>!=!BG!#"2L! HK!&>I:KBG!MN;>L!:G=!LN;L>JN>GMER!IKH<>LL>=!?HK!. +!!BLHE:MBHG!LM>Ii! ! - *!!;C@=3D;@?! .:MB>GMLlL!. +!!P:L!IKH<>LL>=!;R!$BGM!<>GMKB?N@:MBHG!:L!=>L=!:;HO>i!L?M>K! <>GMKB?N@:MBHG! :G=! P:LABG@! LM>If! I:MB>GMlL! . +!! P:L! =! :G=! L>KO>=! NLBG@! $ 1!IENL!|{Ü!j@K:=>!"+1-!q!EBGB+L!1r!:G=!LMHK>=!:M!FBGNL!Ñ{!<>ELBNL!qå!r!NGMBE! :G:ERLBLi! ! *E=D;A3B3>7D7B!8=@G!5ID@>7DBI!3?3=IC;C! ! !KRHIK>L>KO>=! I:MB>GM! . +!L! P>K>! MA:P>=! BG! IK>jP:KF>=! 0.+'j%ENM:F:Q! q%B;KNFj?K>>! F>=BNF! :G=! P:LA! HG<>! BG! MABL! F>=BNFi! '?! <>EE! :K>=f! MA>R! P>K>! MK>:M>=!PBMA!:!",:L>!=B@>LMBHG!LM>I!q1M>F!>EEr!?HK!|ÅF!:M!02!:G=!?HEEHP>=!;R!:GHMA>K! P:LABG@!LM>Ii!$BG:EERf!{i}!j!|!Q!|{ Ç!. +!!I>K!L:FIE>!:K>!LM:BG>=!PBMA!:!FNEMBE!PBMA!."j|!p!$'2!f!!"ÅÉ!p!.>K!.j!RÅiÅf!!"Ç}*!p!.#j!$ÅÖÄf!!"|Ç!p!L$É{{f!

! Ç|! !"|Ä! p! 4ÇÅ{f! !"~~! p! 4É||f! !"~! p! 4ÉÑÇf! !"|{É:! p! 34~ÖÅf! !"|Ö! p! 34É~É! q "! BHLG<>Lrf! !"ÅÇ! p! .#j4BHÉÉ{f! ,)%}!! p! L.!f! !"ÄÅ0L! p! L.!j4BHÉÉ{f! !"ÄÅ0-! p! 4BH%K>>Gf!!"É!p!4BH EN>!q+BEM>GRB! BHM>Kr!:L!=>L=! BG! D34=7!X i!!>EEL!P>K>!LM:BG>=!?HK!~{F!:M!Äå!f!P:LA>=!PBMA!. 1!:G=!K>LNLI>G=>=!BG!$L!1! LM:BGBG@!;N??>K!;>?HK>!:M>Ki!1M:BG>=!L:FIE>L!P>K>!:=!;R!$HKM>LL:! M>K!q "r!?HK!:M!E>:LM!|{l{{{!>O>GML!HG!,)!<>EE!@:M>!=>?BG>=!;R!!"ÅÇá!"~ji!0>LNEML! P>K>!:G:ERL>=!;R!):ENS:!LH?MP:K>!Åi|!q >Kri!! !! /D3D;CD;53=!3?3=IC;C! #QI>KBF>GM:E! ?B@NK>L! :G=! LM:MBLMB<:E! :G:ERLBL! P>K>! I>K?HKF>=! NLBG@! %K:IA.:=! .KBLF! qOÇi{ri! 1M:MBLMB<:E! LB@GB?B<:G<>! ;>MP>>G! L:FIE>L! :M! m=B:@GHLMBi! m;>?HK>! MK>:MF>GMnr! :G=!m:?M>K!MK>:MF>GMn!H?!MA>!L:F>!I:MB>GM!P:L!=>M>KFBG>=!NLBG@!I:BK>=!1MN=>GM!MjM>LMi!LEE! %!O:EN>L!IK>L>GM>=!:L!uh!Iâ{i{Åg!uuh!Iâ{i{|g!uuuh!Iâ{i{{|i!LO>K:@>!O:EN>L!P>K>!>QIK>LL>=! :L!F>:G!IENL!HK!FBGNL!MA>!LM:G=:K=!=>OB:MHG!q1"ri! ! eZf!'(3!KLM7Q! ! -3D;7?DC! EHH=!L:FIE>L!?KHF!&'4á!I:MB>GML!:G=!A>:EMAR!=HGHKL!q&"r!P>K>!DBG=ER!IKHOB=>=!;R!.Ki! .B>KK>! !HK;>:N! q"XI:KM>F>GM! =l'FFNGHEH@B>f! !&3! =>! +HGMI>EEB>Kri! "NKBG@! MA>! BGBMB:E! IA:L>!LMN=Rf!;EHH=!L:FIE>L!?KHF!&'4á!I:MB>GML!q;HMA!:OBK>FBFBK>! ERL>=!MH!>EBFBG:M>!K>=!;EHH=!<>EEL!:G=!=BK>=!?HK!LN;L>JN>GM!$L!1!:G:ERLBLi! ! *E=D;A3B3>7D7B!8=@G!5ID@>7DBI!3?3=IC;C! L?M>K!K>=!;EHH=!<>EE!ERLBG@!LM>If!PAHE>!E>NDHL!P>K>!LM:BG>=!PBMA!MA>!$L!1!:GMB;H=R! I:G>E!q:L!=>LKB;>=!BG! 034=7!Y r!PBMA!!"ÄÅ0-!p!$'2!f!,)%}"!p!.#f!!"|Ö!p!.#j!RÅiÅf!!"|Ä! p! .#j!RÅiÅf! !"|Ç! p! )KHF>-K:G@>! q >Krf! !"ÅÇ! p! .#j4BHÉÉ{f! !"~! p! L.!f! !"ÄÅ0L! p! L.!j4BHÉÉ{! :G=! !"|{É:! p! 4BH EN>! q+BEM>GRB! BHM>EEL! P>K>! LM:BG>=! ?HK! ~{F!:M!Äå!!:L!BG!MA>!LM:G=:K=!IKHMHL=!;>?HK>!:G=!LN;L>JN>GMER!:G:ERL>=!NLBG@! >K! %:EEBHL! ?EHP! MKR! MH! H;M:BG! :M! E>:LM! Ål{{{! >O>GML! HG! ,)! <>EEL! q=>?BG>=!:L!!"ÅÇá!"~jri!! ! !

! Ç}! ! #$%)J## +NEMBI:K:F>M>K!?EHP!MKR!I:G>E!?HK!L+*!L:FIE>L!HG! "!$HKM>LL:! BGLMKNF>GMi!

! ! !

! Ç~! $HK! MA>! L>! H?! &'4! LMN=Rf! PAHE>! . +!! ?KHF! I:MB>GML! :G=! &"! P>K>! LM:BG>=! :<! I:G>E! =>L=! BG! 034=7! ZQ[! :G=! ?BQ>=! PBMA! }Ü! .$L! BG! . 1! ;>?HK>! :G:ERLBLi! 1M:BG>=! L:FIE>L! P>K>! :=! BG! I:K:EE>E! NLBG@! "! $HKM>LL:! :G=! >M>KLi! !

! #$%)T!!+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!?HK!&'4!L:FIE>L!HG! >GMi!

!

! ÇÄ! ! #$%)^#+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!q>QM>G=>=r!?HK!&'4!L:FIE>L!HG! "! $HKM>LL:!BGLMKNF>GMi!

! ! ! !

! ÇÅ! !

! ! #$%)&#:4#+NEMBI:K:F>M>K!?EHP!MKR!I:G>E!q>QM>G=>=r!?HK!&'4!L:FIE>L!HG! >GMi!

! !+#1(/., !DB@9@5ID@C;C!3CC3I! $HK! LMN=RBG@! ,)! <>E! MKH@HGM! . +!L! P>K>! =! PBMA! !"Ä!:G=!!!0Åj>QIK>LLBG@!M:K@>M!<>EEL!q+2Äj0Å!:G=!!#+ri!'G!LNFF:KRf!M:K@>M!<>EEL!NLBG@! BG! MA>! :LL:R! P>K>! E:;>E>=! PBMA! <>EEjMK:<>! OBHE>M! q!24r! q'GOBMKH@>Gr! :<! F:GN?:K! IKHMHG! ?NKMA>K! =! PBMA! I:MB>GM! PAHE>! . +!L! ?HK! ÄA! :M! ~Éå!! BG! |{Ü! $ 1! LNIIE>F>GM>=! 0.+'j%ENM:F:Q! q%B;K! ! M:K@>M! <>EEL! :G=! . +!! FBQ! P:L! =! :G=! LM:BG>=! ?HK! $L!1! :G:ERLBL! NLBG@! >M>K! MH! =>M>F;K:G>! MK:GL?>K! ?KHF! M:K@>M! <>EEL! MH! ,)! <>EEL! ;R! @:MBG@! HG! !"Ä! :G=oHK! !!0Å! >QIK>LLBHG! BG! I:MB>GM! ,)! <>EE! :L! =>?BG>=! ;R! !"ÅÇá!"~j!:G=!>QMLi! !

! ÇÇ! /D3D;CD;53=!3?3=IC;C! #QI>KBF>GM:E! ?B@NK>L! :G=! LM:MBLMB<:E! :G:ERLBL! P>K>! I>K?HKF>=! NLBG@! %K:IA.:=! .KBLF! qOÇi{ri!1M:MBLMB<:E!LB@GB?B<:G<>!;>MP>>G!&"!L:FIE>L!:G=!&'4á!I:MB>GML!q;HMA!:OBK>FBFBM>KFBG>=!NLBG@!HG>jP:R!L,-4Li!LEE! %!O:EN>L!IK>L>GM>=!:L!uh!Iâ{i{Åg!uuh! Iâ{i{|g!uuuh!Iâ{i{{|i!LO>K:@>!O:EN>L!P>K>!>QIK>LL>=!:L!F>:G!IENL!HK!FBGNL!MA>!LM:G=:K=! =>OB:MBHG!q1"ri! ! eZg! "#&!%$# !,)!68DD!8PH4FK?GF!84DL>Q!7GFGJ!2! ! ! 2H! LMN=R! ,)! <>EE! KBLMBEEL! P>K>! :=! :G=! >QI:G=>=! ?KHF! A>:EMAR! =HGHK! 3! ! :<L=! IK>OBHNLER! q1UGSj+:KM[G>S! >M! :Ei! }{|Åf! }{|Çri! 'G! LAHKMf! 3! ! FHGHGN:K! <>EEL! q3! +!r! ?KHF! A>:EMAR!=HGHKL!P>K>!H;M:BG>=!?KHF!!&3!+HGMI>EEB>K!&HLIBM:Ei!3! +!!P>K>!=!;R! $BGM!<>GMKB?N@:MBHG!:G=!?KHS>G!BG!F>=BNF!F>GM>=!PBMA! |{Ü! ! "+1-! q+BEM>GRB! BHM>EE! >QI:GLBHG! LM:KM! PBMA! >BMA>K! ?K>LAER! BLHE:M>=! HK! L>KO>=! 3! +!i! 'G! <:L>! H?! L>KO:MBHGf! :! IK>EBFBG:KR! ",:L>! =B@>LMBHG!LM>I!q",:L>!'!1HENMBHGf!1M>F!>EEr!<:G!;>!=HG>!B?!?KHS>G!<>EEL!P>K>!:@@K>@:M>=! =N>!MH!F>LLi! >?HK>!>QI:GLBHGf!2!<>EEL!P>K>!>EBFBG:M>=!NLBG@!!"~á!IHLBMBO>! L>E>F!>EEr! :<! F:GN?:K! IKHMHEE! =>IE>MBHG! P>K>! <:KKB>=!;R!:!MHM:E!H?!~!!INKB?B<:MBHG!LM>I!;R!F:@G>M!L>I:K:MBHG!MH!:OHB=!2!<>EE! MBMBO>! >QI:GLBHGi! $BG:EER! !"~j! <>EEL! P>K>! =! PBMA! MA>! @:FF:jBKK:=B:M>=! q|{{%Rr! :<<>LLHKR! .*&! <>EE! EBG>! q# 4á! ERFIAH;E:LMHB=! <>EE! EBG>r! :M! MA>! K:MBH! ,)hB.*&! s|hÄtf!:M!MA>!BGBMB:E!GMK:MBHG!|fÅ!Q!|{ Ç!<>EEoFE!BG!,)!<>EE!F>=BNFh!|{Ü!$ 1!!0.+'j %ENM:F:Q! KE>NDBGj}! q.>IKHM>KE>NDBGj|Å! q+BEM>GRB! BHM>=! ,)! <>EE! ! BL! K>?K>LA>=! >O>KR! ! H?! ~jÄ! =:RL! ;R! K>IE:! PBMA! ?K>LA! F>=BNF! q,)! F>=BNF! M>=! LNIIE>F>GMr! :G=! F:BGM:BGBG@!,)hB.*&!K:MBH!:M!s|h|t!NGMBE!,)!<>EE!I>K<>GM:@>!BG!MA>!!BL!AB@A>K!MA:G! Ö{Ü! =>M>KFBG>=! ;R! !"~j!"ÅÇá! IHINE:MBHGi! 2RIB<:EERf! "##("'&$ ! >QI:G=>=! ,)! <>EEL! P>K>! NL>=! ?HK! ?NKMA>K! GHMRIB<:E! :G:ERLBL! ?KHF! =:R! |Ä! NGMBE! =:R!}|!IHLM!>QI:GLBHGi!0>F:BGBG@!<>EEL!<:G!;>!D>IM!?HK!?NMNK>!NL>!;R!L>KO:MBHG!:L! IK>OBHNLER!=>L=!BG!EBJNB=!GBMKH@>Gi! ! )

! ÇÉ! ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! ÇÑ! ) ) ! ! ! ! !

!'".1#/!fZ!! ! !

(78FL?!>8E4LGDG=?64D!64F68JK! ! ! I<@087>632+/=?7183+G+3 7;932717:3J+ + + K&:+ ?67>+ >?@2D+ ?6/?+ &+ >6/=32+ 1;G/@?6;=>67+ 7:A;8A32+ 7:+ <8/::7:5+ /:2+ 3C31@?7;:+;4+?63+3C<3=793:?>E+1;8831?7:5+/:2+7>;8/?7:5++>/9<83+) '!E+ <3=4;=97:5+$D!++/:/8D>7>+/:2+"##)/==/D+79/57:5+3C<3=793:?L+ ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÇÖ! EBioMedicine 2 (2015) 1364 –1376

Contents lists available at ScienceDirect

EBioMedicine

journal homepage: www.ebiomedicine.com

Research Article Identi fication of Anti-tumor Cells Carrying Natural Killer (NK) Cell Antigens in Patients With Hematological Cancers

Ewelina Krzywinska a, Nerea Allende-Vega a, Amelie Cornillon a, Dang-Nghiem Vo a, Laure Cayrefourcq b,c, Catherine Panabieres b,c, Carlos Vilches d, Julie Déchanet-Merville e, Yosr Hicheri f, Jean-François Rossi f, Guillaume Cartron f, Martin Villalba a,g,⁎ a INSERM U1183, Université de Montpellier, UFR Médecine, Montpellier, France b Laboratory of Rare Human Circulating Cells (LCCRH), Department of Cellular and Tissular Biopathology of Tumors, University Medical Centre, Montpellier, France c EA2415 — Help for Personalized Decision: Methodological Aspects, University Institute of Clinical Research, Montpellier University, Montpellier, France d Inmunogenética-HLA, Hospital Univ. Puerta de Hierro, Manuel de Falla 1, 28220 Majadahonda, Madrid, Spain e CNRS UMR 5164, Université Bordeaux, 33076 Bordeaux, France f Département d'Hématologie Clinique, CHU Montpellier, Université Montpellier, 80 Avenue Augustin Fliche, 34295 Montpellier, France g Institut for Regenerative Medicine and Biotherapy (IRMB), CHU Montpellier, Montpellier 34295, France article info abstract

Article history: Natural killer (NK) cells, a cytotoxic lymphocyte lineage, are able to kill tumor cells in vitro and in mouse models. Received 4 February 2015 However, whether these cells display an anti-tumor activity in cancer patients has not been demonstrated. Here Received in revised form 11 August 2015 we have addressed this issue in patients with several hematological cancers. We found a population of highly Accepted 11 August 2015 activated CD56 dim CD16+ NK cells that have recently degranulated, evidence of killing activity, and it is absent Available online 13 August 2015 in healthy donors. A high percentage of these cells expressed natural killer cell p46-related protein (NKp46), Keywords: natural-killer group 2, member D (NKG2D) and killer inhibitory receptors (KIRs) and a low percentage expressed CD45 NKG2A and CD94. They are also characterized by a high metabolic activity and active proliferation. Notably, we Trogocytosis found that activated NK cells from hematological cancer patients have non-NK tumor cell antigens on their NK cell surface, evidence of trogocytosis during tumor cell killing. Finally, we found that these activated NK cells are dis- Cytotoxicity tinguished by their CD45RA +RO + phenotype, as opposed to non-activated cells in patients or in healthy donors − Hematological cancer displaying a CD45RA +RO phenotype similar to naïve T cells. In summary, we show that CD45RA +RO + cells, which resemble a unique NK population, have recognized tumor cells and degranulate in patients with hemato- logical neoplasias. © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/ ).

1. Introduction (Fujisaki et al., 2009; Sanchez-Martinez et al., 2014 ). In peripheral blood, − human NK cells are mostly CD3 CD56 dim cells with high cytotoxic activ- − The immune system recognizes and eliminates tumor cells ( Dunn ity, while CD3 CD56 brigth cells excel in cytokine production ( Bryceson et al., 2002 ), which defend themselves by different mechanisms et al., 2011 ). In vitro evidence indicates that CD56 bright NK cells are precur- (Villalba et al., 2013 ). The lymphocyte lineage natural killer (NK) cell be- sors of CD56 dim NK cells and this might also be the case in vivo ( Domaica longs to the innate immune system ( Lanier, 2008; Vivier et al., 2008 ) et al., 2012 ). In addition, combined analysis of CD56 and CD16 expression and show strong anti-leukemia activity when engrafted in allogeneic during NK cell development indicates that their pro files change as − settings in hematological cancer patients ( Velardi, 2008; Ruggeri et al., follows: CD56 brigth CD16 → CD56 brigth CD16 dim → CD56 dimCD16 dim → 2007; Anel et al., 2012 ). However, the presence of an anti-leukemia CD56 dim CD16 +. Additional markers can be used to identify speci fic sub- NK cell population in patients with hematological malignancies has sets within these NK cell populations ( Moretta, 2010; Freud et al., 2014 ). not been proven. Identi fication of antileukemic NK cells in vivo is complex. CD69 NK cells are not a homogenous population and different subsets expression increases after NK cell activation but it is not exclusive of have different physiological activities. Moreover, different stimuli NK cells encountering tumor cells ( Fogel et al., 2013; Elpek et al., (e.g., cytokines vs targets cells) give rise to different immunophenotypes 2010 ). Due to the clinical interest of NK cells, it is therefore highly rele- vant to identify more precisely the NK cell population(s) with antitumor functions. ⁎ Corresponding author at: INSERM U1183, Institute de Regenerative Medicine and fi Biothérapie, 80, Avenue Augustin Fliche, 34295 Montpellier Cedex 5, France. CD45 is a protein tyrosine phosphatase that is speci cally expressed E-mail address: [email protected] (M. Villalba). in leucocytes ( Kaplan et al., 1990 ). CD45 regulates receptor signaling by

http://dx.doi.org/10.1016/j.ebiom.2015.08.021 2352-3964/© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ). E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1365 direct interaction with components of the receptor complexes or by de- 37. All samples from cancer patients were collected at diagnosis and in- phosphorylating and activating various Src family kinases (SFK) ( Rhee cluded HD , Healthy donor (n bs = 10); MM , multiple myeloma (n bs = and Veillette, 2012 ). However, it can also hinder cytokine receptor sig- 19, n bms = 20); B-CLL , B-cell chronic lymphocytic leukemia (n bs = naling by inhibiting Janus kinases (JAK) ( Irie-Sasaki et al., 2001 ) or by 15); BCL, B-cell lymphoma (n bs = 14); AML, acute myeloid leukemia dephosphorylating Src activating residues (Rhee and Veillette, 2012). (n bs = 14); bs , blood samples; bms , bone marrow samples. CD45 activity is critical for ef ficient immune response, because its de fi- ciency results in severe combined immunode ficiency (SCID) in mice 1.1.5. Multicolor Staining of Cell Surface Markers (Kishihara et al., 1993; Byth et al., 1996; Mee et al., 1999 ) and humans PBMCs were stained with 7AAD (Beckman) to identify viable (Kung et al., 2000; Tchilian et al., 2001 ). cells and with the following anti-CD25-FITC, −CD45RO-FITC, −CD161- Total CD45 expression increases with T cell maturation ( Hermiston FITC, −CD3-PE, −CD19-PE, −CD62L-PE, −CD69-PE, −CD138-PE, et al., 2009 ), but the CD45 family comprises several isoforms derived −CD314(NKG2D)-PE, − CD3-ECD, − CD19-ECD, −CD38-ECD, from a single complex gene ( Hermiston et al., 2009 ). Naive T lympho- −CD56-PECy7, CD3-APC, − CD56-APC, − GzB-AlexaFluor700, cytes usually express the long CD45RA isoform. Activated and memory −CD19-AlexaFluor700, −CD20-APC-AlexaFluor750, −CD45-APC- T cells express CD45RO, the shortest CD45 isoform, generated by AlexaFluor750, −CD45RA-APC-AlexaFluor750, −CD5-Paci ficBlue, activation-induced alternative splicing of CD45 pre-mRNA ( Warren −CD16-Paci ficBlue, −CD57-Paci ficBlue, −CD45-KromeOrange, −CD16- and Skipsey, 1991; Roth, 1994; Lynch and Weiss, 2000; Hermiston KromeOrange (Beckman), −CD158b-FITC, −CD158a-PE, −CD107a- et al., 2009 ). It has been proposed that CD45RO expression also iden- HV500, −Ki-67-V450 (BD Biosciences), −CD45RA-FITC, −CD45RO-PE, ti fies memory NK cells ( Fu et al., 2011 ). Little is known about CD45 ex- − CD159a(NKG2A)-PE, −CD335(NKp46)-PE, − CD94-PE-Vio770, pression and function in NK cells, although it is commonly accepted that −CD335(NKp46)-PE-Vio770, −CD45RO-APC, −CD14-VioBlue, −CD19- CD45 positively regulates their activation by dephosphorylating the in- VioBlue, −CD158e-VioBlue (Miltenyi Biotec) and -CD71-APC (Immuno- hibitory site of SFKs, thus leading to cytokine and chemokine produc- Tools) antibodies against surface markers for cell phenotyping. Brie fly, tion. However, in vitro cytotoxicity is only slightly impaired in NK cells 1x10 6 cells were incubated with the different antibodies in PBS with 2% derived from CD45-de ficient mice ( Huntington et al., 2005; Hesslein FBS at 37 °C for 30 min. Cells were then washed and suspended in et al., 2006; Mason et al., 2006 ). As most of what is known was obtained 200 –250 μl PBS 2% FBS and staining was analyzed using a Gallios flow in mouse models and cannot be transposed to humans, the role of CD45 cytometer (Beckman) and the Kaluza software. in human NK cells is an open issue and could depend on the type and Viable lymphocytes were gated using FSC-SSC and 7AAD staining. B − strength of the activation. lymphocytes (CD19 +), T lymphocytes (CD3 +CD56 ) and NK cells − Here we show that, in contrast to T cells, NK cells can express both (CD56 +CD3 ) were differentiated based on CD19, CD3 or CD56 expres- CD45 isoforms: CD45RA and CD45RO. Moreover, these CD45RARO NK sion. NK cells were then separated in four distinct populations based on − cells recognize tumor cells in patients with hematological cancers and, CD45RA and CD45RO expression: CD45RA +RO (CD45RA ), CD45RA +- − subsequently, degranulate. RO + (CD45RARO ), CD45RA dim RO (CD45RA dim ), CD45RA dim RO + (CD45RAdimRO ). These different populations were then analyzed for 1.1. Experimental Procedures CD16, CD57, CD62L, CD69, CD71, CD94, CD107a, CD158a, CD158b, CD158e, CD159a (NKG2A), CD161, CD314 (NKG2D), CD335 (NKp46), 1.1.1. Cell Culture Ki-67, GzB expression and cell size and granularity (FSC and SSC). The K562 cell line (ATCC CCL 243) and the lymphoblastoid EBV cell line PLH (IHW Number: 9047) were maintained in logarithmic growth 1.1.6. In Vitro CD107a Degranulation Assay in RPMI 1640 medium (Gibco® GlutaMAX ™ media) with 10% fetal bo- After PBMC puri fication and NK cell quanti fication, 3 million cells vine serum (FBS) (Gibco®). Cells were cultured at 37 °C in a humidi fied were incubated at 37 °C for 4 h or overnight with K562 target cells at chamber with 5% CO 2 in air, and passaged 1:10 twice a week. an Effector (NK cell): Target ratio of 1:10 in a final volume of 500 μl (RPMI Glutamax with 10% FBS and 10U/ml IL2). The medium also 1.1.2. Peripheral Blood Mononuclear Cell (PBMC) Puri fication contained 1.5 μl anti-CD107a antibody (BD Biosciences, Franklin Lakes, Bone marrow and peripheral blood samples were obtained from pa- NJ) and 1 μl monensin to prevent CD107a degradation (BD Golgi-Stop tients with different hematological diseases and from healthy donors BD Biosciences). Then, cells were resuspended in 50 μl of an antibody after informed consent. Cells were puri fied by Ficoll-Hypaque (Sigma) cocktail containing the anti-CD45RO-FITC, −CD69-PE, −CD19-ECD, density-gradient centrifugation as described earlier ( Allende-Vega −7AAD, −CD56-PECy7, −CD3-APC, −CD45RA-APCAlexaFluor750, et al., 2015 ). Brie fly, 3 –6 ml of 1:2 diluted blood or 1:3 diluted bone mar- −CD107a-HV500 and −CD16-KO antibodies (BD Biosciences, row samples in RPMI were added on top of 5 ml of Histopaque. Cells Beckman). Samples were analyzed on a Beckman Coulter FACS Gallios were centrifuged at 1600 rpm and at 20 °C without break for 30 min. flow cytometer using the Kaluza software. Events were initially gated Mononuclear cells were collected from the interlayer white ring. After on forward and side scatter (SSC) to identify lymphocytes. A bivariate washing in RPMI, cells were suspended in complete RPMI medium sup- plot of CD56 versus CD3 was used to acquire at least 10,000 NK cells. plemented with 10% FBS (Invitrogen). 1.1.7. Multicolor Staining for Cell Surface and Intracellular Markers 1.1.3. In Vitro NK Cell Stimulation Protocol After PBMC puri fication, 1 million cells were pre-blocked by incuba- PBMCs, 1.10 6 cells/ml, were stimulated during 10 or 20 days with a tion with 10% normal human serum at RT for 15 min and then stained high dose of IL-2 (1000 U/ml, eBiosciences) or with the lymphoblastoid with 50 μl of the PANEL Ki-67 antibody cocktail against cell surface EBV cell line PLH together with IL-2 (100 U/ml) and IL-15 (5 ng/ml, markers (anti-CD45RO-PE, −CD19-ECD, −CD56-PC7, −CD3-APC, Miltenyi). −CD45RA-APCAlexaFluor750 and −CD16-KO antibodies) (BD Biosci- ences, Beckman). Cells were washed twice with Staining Buffer and re- 1.1.4. Selection of Patients and Healthy Donors suspended in 250 μl BD Cyto fix-Cytoperm solution at 4 °C for 20 min. Data and samples from patients with different hematological can- Cells were washed twice in BD Perm-Wash solution. Next, cells were cers were collected at the Oncology and Clinical Hematology Depart- fixed-permeabilized in 50 μl BD Perm-Wash solution containing an an- ment of the CHU Montpellier, France, after patient's informed consent tibody cocktail against intracellular markers (anti-GzB-AlexaFluor700, (Allende-Vega et al., 2015 ). Patients were enrolled in two independent −Ki-67-V450) as described in the figures at 4 °C for 30 min in the clinical programs approved by the “Comités de Protection des dark. Cells were washed twice in BD Perm-Wash solution and resus- Personnes Sud Méditerranée I ”: ref 1324 and ID-RCB: 2011-A00924- pended in Staining Buffer prior to flow cytometric analysis on a 1366 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

Beckman Coulter FACS Gallios flow cytometer using the Kaluza soft- Similar increases in the CD45RA dim and CD45RO populations were ware. Events were initially gated on forward and side scatter (SSC) to also observed in bone marrow samples from patients with acute mye- identify lymphocytes. A bivariate plot of CD56 versus CD3 was used to loid leukemia (AML) or in blood samples of patients with B-cell chronic acquire at least 10,000 NK cells. lymphocyte leukemia (B-CLL) and B-cell lymphoma (BCL) ( Fig. 1 A and supplemental Table 3). In summary, the C45RARO cell population was 1.1.8. Identi fication of Pure Single NK Cells statistically increased in all analyzed samples from patients with blood Primary CD56 + NK cells were enriched and puri fied from PBMCs of malignancies compared to healthy controls ( Fig. 1 B and supplemental B-cell lymphoma patients with the CD56 + NK cell isolation kit (Miltenyi Fig. 1). The gating strategy to identify CD45RARO cells is described in − Biotec, Auburn, CA, USA). The purity (% of CD56 +CD3 ) of CD56 + NK supplemental Fig. 1B). cells, measured by flow cytometry, was N90%. Puri fied CD56 + NK cells have been stained with anti-CD335(NKp46)-PE to formally identify NK 2.2. Phenotypic Characterization of CD45RARO Population cells together with anti-CD45RA-FITC, −CD45RO-APC and −CD19- VioBlue (detection of trogocytosis-capable NK cells). Puri fied and stained As indicated in Fig. 1 , CD45RARO cells belonged to the CD56 +CD16 + NK cells have been analyzed with the DEPArray ™ System (Silicon subset ( Fig. 2 A) and mostly express the maturation marker CD57 Biosystems, Menarini). This technology allowed us to detect, enu- (Fig. 2 B) although CD62L was coexpressed by half of them. The merate and take pictures of single cells. CD45RARO population contained higher percentage of cells that expressed KIRs, although it was statistically signi ficant only for 1.1.9. DEPArray TM Procedure CD158e ( Fig. 2 C and supplemental Fig. 2). The percentage of gran- Cell sorting experiments were performed as described in the manu- zyme B (GzmB) + cells was similar to other subsets, but the intracel- facturer's instructions and in ( Lianidou et al., 2013 ). Brie fly, DEPArray car- lular level of this cytokine was lower ( Fig. 2 C). This could be due to a tridges were manually loaded with 14 μl of sample and 800 μl of the buffer de ficient production or a recent degranulation that has emptied the solution in which puri fied and stained NK cells had to be recovered. After intracellular stores. CD45RARO cells also expressed similar levels loading the cartridge into the DEPArray system, ∼9.26 μl of sample was than CD45RA of another maturation marker the CD161-Killer cell automatically injected by the system into a microchamber of the cartridge lectin-like receptor subfamily B, member 1 (KLRB1) or the natural where the cells were spontaneously organized into a preprogrammed cytotoxicity receptor (NCR) NKP46 and slightly higher levels of the electric field consisting of 16,000 electrical cages in which individual activating NKG2D receptor ( Fig. 2 D and supplemental Fig. 3). How- cells are trapped. Image frames covering the entire surface area of the ever, they showed lower levels of the CD94 glycoprotein and, prob- microchamber for each of four fluorescent filter cubes (FITC, PE, APC ably, the inhibitory NK receptor NKG2A ( Fig. 2 D and supplemental and DAPI-Hoechst-VioBlue-Paci ficBlue) and bright field images were cap- Fig. 3). In summary, CD45RARO cells are fully mature NK cells that tured. Captured images were digitally processed and presented in a soft- mainly express NK receptors of mature cells. ware module that enables selection of cells of interest by the operator. 2.3. Expression of Different CD45 Isoforms in Vivo: Patients With 1.1.10. Statistics Cytomegalovirus (CMV)-Reactivation All the experiments shown in the figures were performed at least with samples from six patients for each malignancy and the same num- We next asked whether other conditions that lead to NK cell activa- ber of healthy donors (HD). The statistical analysis was performed using tion, such as viral infections, could give rise to a similar phenotype. We the Student t test: *p b 0.01; **p b 0.001; ***p b 0.0001. Average values thus analyzed peripheral blood mononuclear cell (PBMC) samples from were expressed as mean plus or minus the standard error (SD). patients with reactivation (CMV +) or not (CMVneg) of CMV infection following kidney transplantation. CMV reactivation induced an increase 2. Results in the total number of NK cells ( Fig. 3 A). In addition, CMV + patients showed an increase in CD56 dimCD16dim cells associated with a reduc- 2.1. Expression of Different CD45 Isoforms in Patients With Hematological tion of the CD56 brigth subsets compared to CMVneg patients ( Fig. 3 B). Malignancies The reason of these changes is not clear to us, but could be due to dif- ferent factors, such as CMV-induced NK cell maturation ( Della Chiesa In healthy donors, NK cells were mainly CD45RA cells with few et al., 2013 ), or an effect on the expression of the different NK cell CD45RA dim cells, found particularly in immature NK cell subsets. markers in CMV-infected cells, as previously described for decidual NK CD45RARO cells represented between 0 and 0.75% of all NK cells and cells ( Siewiera et al., 2013 ). These changes were accompanied by belonged exclusively to the fully mature CD56 +CD16 + subset ( Fig. 1 A minor variations in the expression pattern of CD45 isoforms ( Fig. 3 C). top panels and supplemental Table 1). NK cells derived from healthy These results indicate that the expression pattern of CD45 isoforms in donor bone marrows showed equal distribution ( Fig. 1B). Blood sam- NK cells activated by viral infection or hematological cancers is different. ples from patients with multiple myeloma (MM) contained four times Speci fically, CD45RARO cells are mainly present in samples from more CD45RA dim cells and between 1 and 20% of CD45RARO cells patients with hematological cancers, whereas they represent a minor (Fig. 1 A and supplemental Table 2). As MM is characterized by accumu- fraction in virus-infected patients. lation of tumor cells in the bone marrow, we also investigated whether bone marrow NK cells, which should be in closer contact with tumor 2.4. Metabolic Characterization of CD45RARO Population cells, were more activated than circulating NK cells. This was not the case as the percentage of CD45RA dim and CD45RARO cells was similar Activated lymphocytes generally increase their size ( Zarcone et al., in blood and bone marrow samples ( Fig. 1 A and supplemental Table 2). 1987; Skak et al., 2008 ) and become highly metabolically active cells

Fig. 1. Patients with hematological malignancies and healthy donors have different NK cell subset pro files. A) PBMCs from blood samples (bs) of a healthy donor and of a patient with multiple myeloma (MM) or from bone marrow (bms) of the patient with MM or samples of patients with other hematological diseases were stained for FACS analysis with anti-CD19 − − (B cells), −CD3 (T cells, CD3 +CD56 ) and −CD56 (NK cells, CD56 +CD3 ), to identify the different lymphocyte populations, and also with anti-CD16, to identify NK cell subsets at different stage of maturation, and with −CD45RA, and −CD45RO antibodies. Numbers in the quadrants indicate the percentage of cells. B) Percentage of different NK cell populations based on CD45RA and RO expression in healthy donors and in patients with hematological cancers. The populations correspond to the quadrants in A: upper left (CD45RA), bottom left (CD45RAdim), upper right (CD45RARO) and bottom right (CD45RA dimRO). The bars show the mean ± SD for each medical condition, Student t-test compare to healthy donor blood (left panel) or bone marrow (right panel) samples: *p b 0.01; **p b 0.001; ***p b 0.0001. HD , Healthy donor; MM , multiple myeloma; B-CLL , B-cell chronic lymphocytic leukemia; BCL , B-cell lymphoma; AML , acute myeloid leukemia; bs , blood samples; bms , bone marrow samples. E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1367 1368 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

(Sanchez-Martinez et al., 2014; Sanchez-Martinez et al., 2015 ). The Ser- 2.5. Functional Characterization of CD45RARO Population Thr kinase mammalian target of rapamycin (mTor) probably links the metabolic shift and the cytoskeletal organization after NK cell activation To identify the function of the different NK cell subsets, first we (Marcais and Walzer, 2014 ). We observed a subset of NK cells from MM assessed cell degranulation by ex vivo staining of PBMCs with anti- patients that was larger in size (FS) and with high granularity (SS) and CD107a antibodies ( Fig. 5 A). In healthy donors, around 1% NK cells corresponded to CD45RARO cells ( Fig. 4A and supplemental Fig. 5C). were CD107a +. Most of these cells were CD45RA, with a small number This suggested that they were activated cells that had a higher metabol- of CD45RO + cells. Remarkably, most CD45RARO and half of CD45RA- ic activity compared to the other NK subsets. Hence, we activated NK dim RO cells were CD107a + (Fig. 5 B). cells in vitro by incubating them with the Epstein Barr Virus (EBV) NK cells from patients with hematological cancers showed a large lymphoblastoid cell line PLH. To support NK cell survival we added increase in CD107a + cells ( Fig. 5 A and supplemental Fig. 6A), particular- low concentrations of two NK cell activating cytokines: IL-2 (100 U/ ly among the CD45RO + subsets, which are speci fically increased in ml) and IL-15 (5 ng/ml) ( Anel et al., 2012 ). We incubated NK cells for these patients ( Fig. 1 ). Reduction of CD45RA expression was not associ- up to 20 days to re flect long-term activation. In agreement with previ- ated with increased degranulation ( Fig. 5 B). Like in healthy donors, the ous reports ( Zarcone et al., 1987; Skak et al., 2008 ), 10-day activation in- CD45RARO and, to a lower extent, CD45RA dimRO fractions contained duced an increase in size (FCS) and granularity (SSC) that was more mostly cells that had degranulated ( Fig. 5 B and supplemental Fig. 6B). relevant at day 20 (supplemental Fig. 4A). The median CD107a-mean fluorescence intensity (MFI) of these two − After 3 days of in vitro activation, NK cells started losing CD45RA populations was largely increased compared to CD45RO populations (supplemental Fig. 4B). However, it was questionable if a real CD45RARO (supplemental Fig. 6 A). This was not exclusive of circulating NK cells, population appeared or cells were losing CD45RA whereas gaining because similar results were obtained also for NK cells derived from − CD45RO. 10 days after initial activation, most cells are CD45RA bone marrow samples of patients with MM and AML ( Fig. 5 B and C CD45RO +. However, at day 20 a CD45RARO population appeared in the upper panels). In contrast, CD45RARO cells showed low GzmB content culture. This was not exclusive of the presence of accessory cells because (Fig. 2 C). Our explanation is that CD45RARO cells had recently long-term activation with cytokines produced a similar pattern (supple- degranulated in vivo. mental Fig. 4B). In summary, CD45RARO NK cells also exist in vitro after CD45RARO cells continued to show the higher degranulation rate long activation. Next, we evaluated in vitro activation of patient after an in vitro analysis using K562 as target cells ( Fig. 5 C bottom CD45RARO population. Three days of cytokine-induced activation in- panels), although other populations signi ficantly increased degranula- duced a strong rearrangement on the expression of CD45 isoforms and tion. Interestingly, the different CD45 NK cell subsets did not change it was impossible to evaluate the faith of individual populations (Supple- after the 4-hour in vitro cytotoxic assay (supplemental Fig. 7). This mental Fig. 4C). These results additionally suggested that CD45RARO cells and the in vitro activation results (supplemental Fig. 4) showed that ex- could change their CD45 phenotype, at least in vitro. pression of CD45RA and CD45RO is stable at short times but can change We then assessed the expression of transferrin receptor protein 1 after long lasting activation. (TfR1 or CD71), which is required for iron delivery from transferrin to the cells. CD71 expression increases in active metabolic cells be- 2.6. CD45RARO Have Performed Trogocytosis in Vivo cause iron is a cofactor for fundamental biochemical activities, such as oxygen transport, energy metabolism and DNA synthesis To investigate if CD45RARO cells were performing antitumor activity (Wang and Pantopoulos, 2011 ). In agreement with the superior in vivo, we investigated if these cells have performed trogocytosis on metabolic activity suggested by high FS and SS of CD45RARO cells, tumor targets. Trogocytosis is a process whereby lymphocytes, i.e. NK CD71 expression was higher in CD45RO + cells in both healthy con- cells ( Suzuki et al., 2015; Nakamura et al., 2013 ), gain surface molecules trols and patients with hematological malignancies ( Fig. 4 B and sup- from interacting cells and express them on their own surface and has plemental Fig. 5A). Moreover, most of these cells also expressed the been observed in B lymphoblastic leukemia (B ALL) ex vivo ( Soma proliferation marker Ki-67 ( Fig. 4 C and supplemental Fig. 5B). In et al., 2015 ). We observed that long-time activated NK cells (see sup- summary, CD45RARO cells represent a NK subset of highly metabol- plemental Fig. 4) performed trogocytosis in two AML cell lines (supple- ic cells in proliferation. mental Fig. 8). In fact, NK cells extracted at least two proteins expressed CD69 expression increases after NK cell stimulation and is con- in AML cells, CD14 and CD33, with considerable ef ficiency. This showed sidered a bona fide marker of NK cell activation ( Elpek et al., 2010 ), in- that human NK cell ef ficiently performed trogocytosis and we investi- cluding ex vivo ( Vey et al., 2012 ). Analysis of CD69 expression in the gated if this was the case in vivo. Because in this experiment we studied different CD45 populations in patients showed that CD45RO + cells markers of other cell types, we used a double labeling to identify NK were mainly CD69 + (Fig. 4 D); but not vice versa, as most CD69 + cells cells and gated on CD56 +NKP46 + cells. In a BCL patient, we observed were not CD45RO. The very low amount of CD45RO + cells in healthy that 14% of the NK cells expressed the BCL marker CD19 in their mem- donors precluded any meaningful analysis of this population. brane ( Fig. 6 A). This value increased to 52% in the CD45RARO popula- − In healthy donors, CD45RA dim cells were mainly CD69 (Fig. 4E). tion and it was much lower in the other populations. NK cells also CD45RA dim cells were signi ficantly increased in patients and many gained at lower level expression of the myeloid marker CD14, although were also CD69 +. However, reduction of CD45RA expression was not al- the population was predominantly CD45RA dim RO. However, the NK ways associated with gain of CD69 expression. In fact, patients' samples cells that stained positive for both CD19 and CD14 were very rare. This − were enriched particularly in CD45RA dim CD69 and CD45RA CD69+ suggested that two different NK cell populations were performing and, to a lower extent, in CD45RA dim CD69 + cells. This finding suggests trogocytosis. The CD45RARO cells were doing it on tumor cells. We that loss of CD45RA and gain of CD69 expression identify two different observed the very similar results in another CD19 + disease: B-CLL (sup- physiological processes and that these two populations might have dif- plemental Fig. 9). Next, we used the puri fied NK cells (CD56 + selection) ferent functions. from whole blood of a B-CLL patient and analyzed them with the

Fig. 2. The phenotypic characterization of CD45RARO shows that they are fully mature cells. PBMCs from a representative BCL patient were stained as in Fig. 1 to identify the CD45RARO population and the maturation development was revealed by expression of CD56 CD16 (A) or CD57 CD62L (B). Numbers in the quadrant indicate the percentage of cells. C –D) PBMCs from 5 BCL patients were stained as in Fig. 1 to identify the CD45RARO population and the expression of different molecules on the different NK cell subsets was revealed by using antibodies against KIRs 158a, b and e, GzmB, the Lectin Like Transcript-1 (LLT1) receptor CD161, the NCR NKP46, the activating receptor NKG2D, the inhibitory receptor NKG2A (D) and the molecule CD94. E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1369 1370 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

Fig. 3. CMV + patients and patients with hematological cancer have different NK cell subset pro files. A) PBMCs from patients with reactivation (CMV +) or not (CMVneg) of CMV infection following kidney transplantation were puri fied as in Fig. 1 and the percentage of NK cells was calculated. B) The abundance (in percentage) of NK cells at different stages of maturation (CD56 CD16) was analyzed in the samples described in (A). C) The percentage of each NK cell subsets (CD45 isoforms) is shown. There were not any differences between the two groups of patients. In (B) and (C) bars represent the mean ± SD of at least four individuals for each medical condition.

DEPArray ™ System, which allowed identifying, visualizing and taking activated (high size and granularity, CD69 +CD71 +KI67 +NKG2D +) pictures of single cells. We labeled cells with NKp46 to formally identify and that have degranulated (CD107a + and low GzmB content) and per- NK cells together with CD45RA, CD45RO and CD19. Fig. 6 B showed that formed trogocytosis (CD19 + in BCL and B-CLL and CD14 + in AML). single cells expressed all markers. Thus, we showed a picture of a Moreover, they are prompted to perform trogocytosis on different human NK cell that has just performed in vivo trogocytosis in a tumor target cells. These findings suggest that they are effector cells with max- cell (see also the graphical abstract). imal cytotoxic activity against cancer cells. It seems that a population of To exclude that NK cells were not gaining CD19 in all tumors, we in- highly mature NK cells encounters its targets and respond by becoming vestigated NK cells from an AML patient ( Fig. 6B). Around 10% of NK effector cells. In addition, we observed that a population of NK cells has cells expressed the AML marker CD14 and only 3% expressed CD19, performed trogocytosis in non-tumor, myeloid, cells at least in BCL and which was expressed in all NK cell populations. In contrast, 90% of B-CLL patients. It is well known that NK cells kill dendritic cells and mac- CD45RARO expressed CD14 showing that they have massively per- rophages in several contexts, but the role here is unknown. Moreover, formed trogocytosis in a CD14 + population in AML patients. the population that has performed it is mainly CD45RA dim RO, a general- Next, we investigated if CD45RARO cells derived from a AML patient ly minor population. that had performed trogocytosis and gained CD14 expression were able The large size and granularity of CD45RARO cells could preclude to take CD19 from tumor cells of a BCL patient. Tumor CD19 + cells and their observation when standard FCS-SSC parameters for the classical NK cells from the AML patient did not express the same membrane lymphocyte populations are used. It is essential to understand that acti- markers ( Fig. 7 A). We distinguished BCL cells by CD19 and CD10 stain- vated lymphocytes increase in size and granularity, which distinguish ing and after 16-h cytotoxic assay, we observed that 4% of NK cells form them for naïve lymphocytes. This is important for future studies of the AML patient gained expression of both CD10 and CD19 ( Fig. 7 A). The CD45RARO NK cells in solid cancers, which could also induce a similar CD45RARO population was mainly stable all through the assay ( Fig. 7 B). phenotype because NK cell in filtration is associated with a good progno- The population that performed trogocytosis mainly was the CD45RARO sis in several cancers ( Senovilla et al., 2012; Mamessier et al., 2013; (Fig. 7 C). This showed that the CD45RARO population was prompted to Mamessier et al., 2012 ). However, our work does not show the irrefut- recognize and interact with allogeneic tumor cells. In summary, our able proof that CD45RARO cells are bona- fide NK cells, although all data showed that NK cells performed trogocytosis on tumor cells and results point in this direction. Alternative analyses are needed to de fin- that the CD45RARO population is mainly responsible of this. itively state the nature of these cells. Target cell availability is probably maximal for NK cells in blood 3. Discussion borne cancers, hence, we believe that these diseases will show the highest CD45RARO NK cell numbers; although these cells are unable Identi fication of human NK cell populations is important for under- to control the disease. Leukemogenesis in mouse is enhanced when standing their physiology and for improving their therapeutic use in the host immune system is impaired ( Garaude et al., 2008; Kaminski the clinic. Altogether our results indicate that CD45RARO cells are fully et al., 2012 ) and more hematological cancer patients present severe mature NK cells (CD56 dim CD16 +CD57+KIR +CD161 +), which are NK cell dysfunctions ( Baier et al., 2013 ). Others and we have shown E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1371

Fig. 4. Functional characterization of CD45RARO NK cells. A) FS and SS values of the different NK cell subsets (based on the expression of CD45 isoforms) derived from a blood sample of a patient with MM. B –C) Expression of CD71 and Ki67 in the different NK cell subsets in a representative BCL patient. D –E) Representative graphs showing the expression of CD45RO or CD45RA versus CD69 in NK cells from blood (bs) or bone marrow samples (bms) of patients with different blood-borne cancers. Numbers in the quadrant indicate the percentage of cells. the requirement of fully functional NK cells to eradicate blood-borne tu- cell transplantation (UCBT) in the clinic ( Willemze et al., 2009; Stern mors in several mouse models ( Karre et al., 1986; van den Broek et al., et al., 2008 ). Moreover, NK cells: i) are not responsible of GvH disease 1995; Pardo et al., 2002; Aguilo et al., 2009; Charni et al., 2009; Charni (GvHD); ii) can be injected as “differentiated ” cells and thus do not et al., 2010; Ramírez-Comet et al., 2014 ). The use of alloreactive NK need to survive within the patient's body for a long time; iii) protect cells may represent a new cancer treatment, speci fically for tumors of from opportunistic infections ( Willemze et al., 2009 ), probably through hematopoietic origin. Indeed, KIR –KIR ligand incompatibility in the their immunoregulatory effects on B and T cells, macrophages and, graft-versus-host (GvH) direction, which is mainly based on NK cell more importantly, polymorphonuclear cells ( Bhatnagar et al., 2010 ). alloreactivity, improves the outcome after unrelated cord blood stem However, evaluation of NK cell activation in vivo is dif ficult because 1372 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

Fig. 5. CD45RARO identi fies degranulating NK cells. PBMCs from healthy donors (HD) and patients with different hematological malignancies were puri fied as in Fig. 1 . A) Number of CD107a + cells in each NK cell subset (CD45RA RO expression described in Fig. 1 A) per million of NK cells. Bars represent the mean ± SD for each medical condition; Student t-test compare to healthy donor samples. B) Percentage of CD107a + NK cells in the four different subsets. C) Upper panels, Percentage of CD107a + cells in different NK cell subsets isolated from bone marrow samples (bms) of patients with MM (shown also the percentage in the corresponding blood sample, bs, for comparison) or AML. Bottom panels, Percentage of CD107a + cells in different NK cell subsets after exposure to target K562 tumor cells (in vitro cytotoxicity assay). PBMCs were incubated for 4 h with target K562 tumor cells at the effector:target ratio of 10:1. E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1373

Fig. 6. CD45RARO cells have performed trogocytosis on tumor cells. PBMCs from patients with BCL (A) or AML (C) were puri fied as in Fig. 1 and were stained with different antibodies. Numbers in the quadrant indicate the percentage of cells. In this experiment, the NK cell population corresponded to CD56 +NKP46 + cells. B) Puri fied NK cells (CD56 + selection) from a B-CLL patient have been were stained with NKp46, to formally identify NK cells, together with CD45RA, CD45RO and CD19. They were analyzed with the DEPArray ™ System. 1374 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

Fig. 7. CD45RARO cells are prompted to perform trogocytosis on allogeneic tumor cells. PBMCs from an AML patient were incubated with puri fied tumor cells from a BCL patient (E: (NK cell):T ratio 0.3:1), which are CD10CD19 for 16 h before staining with different antibodies. A) Top panels identi fied the BCL cells, the NK cells before and after cytotoxic assay. In the bottom panels CD10CD19 expression was analyzed in the NK cells described in the top panels. Numbers in the quadrant indicate the percentage of cells. B) CD45RA RO expression before and after cytotoxic assay was analyzed in NK cells. C) The expression of CD10CD19 was analyzed in different NK cell CD45 subsets. The numbers in graphics represent the percentages of cells in the speci fic quadrant.

we lack effective methods for their analysis. CD69 expression has rou- chemicals that can be associated with immunotherapy to boost the im- tinely been used ( Elpek et al., 2010; Vey et al., 2012 ); though, our results mune response ( Villalba et al., 2014 ) and that could improve the NK show that CD69 expression does not imply degranulation, which is be- cell-mediated response ( Catalán et al., 2015 ). lieved to be the most essential component of the NK cell anti-tumor ac- CD45 activity is regulated by dimerization and spontaneous CD45 tivity ( Bryceson et al., 2011 ), or trogocytosis. Conversely, our work homodimerization at the plasma membrane inhibits its activity ( Xu indicates that CD45RO expression identi fies degranulating NK cell sub- and Weiss, 2002 ). The size of CD45 extracellular domain is inversely sets in patients with hematological malignancies. We believe that ef fi- proportional to the extent of CD45 dimerization and thus self- cient antitumor treatments that involve also NK cell activity, such as inhibition ( Xu and Weiss, 2002 ). Larger CD45 isoforms, such as monoclonal antibodies against tumor antigens, should also increase CD45RA, dimerize less ef ficiently and, accordingly, they should bet- these NK cell populations. Other options for treatment include new ter promote TCR signaling than smaller isoforms, such as CD45RO E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376 1375

(Rhee and Veillette, 2012 ). However, CD45 activity also depends on Con flicts of Interest its plasma membrane localization and thus on its extracellular do- main ( Mustelin et al., 2005; Rhee and Veillette, 2012 ). At least in T The authors EK and MV have presented a patent application for the cells, too high CD45 activity leads to dephosphorylation of the acti- use of CD45RARO cells as a biomarker of hematological cancers (Martin vating residues in Src kinases, whereas too low CD45 activity might Villalba and Ewelina Krzywinska. Methods for Diagnosing Hematologi- leave phosphorylated the inhibitory residues. Therefore, it is impor- cal Cancers. EP14306134.9.). tant for ef ficient NK cell activation that CD45 activity remains within a speci fic window ( Hermiston et al., 2009 ) and the amount of specif- Appendix A. Supplementary Data ic CD45 isoforms will regulate the final activity. We found that CD45RARO NK cells show maximal degranulation and trogocytosis, Supplementary data to this article can be found online at http://dx. suggesting that expression of both CD45RA and CD45RO isoforms doi.org/10.1016/j.ebiom.2015.08.021 . might give to NK cells the appropriate level of CD45 activity for ef fi- cient signaling to boost cytotoxicity. CD45 is required for full NK cell References cytotoxicity in vivo in mice ( Hesslein et al., 2011 ); however, it is not required in vitro ( Mason et al., 2006; Hesslein et al., 2006; Aguilo, J.I., Garaude, J., Pardo, J., Villalba, M., Anel, A., 2009. Protein kinase C-theta is Huntington et al., 2005 ). In agreement, we observed that other NK required for NK cell activation and in vivo control of tumor progression. J. Immunol. 182, 1972 –1981. cell subsets, which express different CD45 isoforms, improved de- Allende-Vega, N., Krzywinska, E., Orecchioni, S., Lopez-Royuela, N., Reggiani, F., Talarico, granulation in vitro. This suggests that NK cells depend less of G., Rossi, J.F., Rossignol, R., Hicheri, Y., Cartron, G., Bertolini, F., Villalba, M., 2015. CD45 expression in vitro than in vivo. The presence of wild type p53 in hematological cancers improves the ef ficacy of com- binational therapy targeting metabolism. Oncotarget 6, 19228 –19245. Expression of different CD45 isoforms changes the recognition of Anel, A., Aguilo, J.I., Catalan, E., Garaude, J., Rathore, M.G., Pardo, J., Villalba, M., 2012. CD45 ligands. For example, the abundance and types of O-glycans on Protein kinase C-theta (PKC-theta) in natural killer cell function and anti-tumor the different CD45 isoforms regulate the cell sensitivity to galectin-1 immunity. Front. Immunol. 3, 187. Baier, C., Fino, A., Sanchez, C., Farnault, L., Rihet, P., Kahn-Perles, B., Costello, R.T., 2013. (Earl et al., 2010 ). Galectin-1, which is abundantly produced by tumor Natural killer cells modulation in hematological malignancies. Front. Immunol. 4, 459. cells, blocks T cell-mediated cytotoxic responses ( Ito et al., 2012 ) and in- Bhatnagar, N., Hong, H.S., Krishnaswamy, J.K., Haghikia, A., Behrens, G.M., Schmidt, R.E., duces apoptosis of thymocytes and T cells ( Earl et al., 2010 ). It is possible Jacobs, R., 2010. Cytokine-activated NK cells inhibit PMN apoptosis and preserve – that anti-tumor NK cells are selected based on their resistance to their functional capacity. Blood 116, 1308 1316. Bryceson, Y.T., Chiang, S.C., Darmanin, S., Fauriat, C., Schlums, H., Theorell, J., Wood, S.M., galectin-1 or to other ligands through expression of CD45RO. Indeed, 2011. Molecular mechanisms of natural killer cell activation. J. Innate Immun. 3, as O-glycans bind mainly to CD45 extracellular domain, cells that 216 –226. express short CD45 isoforms, like CD45RO, will have relatively fewer Byth, K.F., Conroy, L.A., Howlett, S., Smith, A.J., May, J., Alexander, D.R., Holmes, N., 1996. CD45-null transgenic mice reveal a positive regulatory role for CD45 in early thymocyte O-glycans and thus will be more resistant to galectin-1. development, in the selection of CD4+CD8+ thymocytes, and B cell maturation. J. Exp. Ex vivo we found very few NK cells that express CD45RO in periph- Med. 183, 1707 –1718. eral blood samples from healthy donors. This is surprising, especially if Catalán, E., Charni, S., Aguiló, J.-I., Enríquez, J.-A., Naval, J., Pardo, J., Villalba, M., Anel, A., fi 2015. MHC-I modulation due to metabolic changes regulates tumor sensitivity to CD45RO expression identi es memory NK cells, as it has been proposed CTL and NK cells. Oncoimmunology 4, e985924. http://dx.doi.org/10.4161/ (Fu et al., 2011 ). This finding suggests that the amount of memory NK 2162402X.2014.985924 . cells might be extremely low in blood or bone marrow samples, or Charni, S., Aguilo, J.I., Garaude, J., De Bettignies, G., Jacquet, C., Hipskind, R.A., Singer, D., Anel, A., Villalba, M., 2009. ERK5 knockdown generates mouse leukemia cells with that CD45RO may not be a marker of memory NK cells. Alternatively, low MHC class I levels that activate NK cells and block tumorigenesis. J. Immunol. CD45RO expression in NK cells could have been speci fically lost during 182, 3398 –3405. ex vivo sample handling. We think that this is unlikely because NK Charni, S., De Bettignies, G., Rathore, M.G., Aguilo, J.I., Van Den Elsen, P.J., Haouzi, D., Hipskind, R.A., Enriquez, J.A., Sanchez-Beato, M., Pardo, J., Anel, A., Villalba, M., cells express slightly higher levels of total CD45 than other lymphocyte 2010. Oxidative phosphorylation induces de novo expression of the MHC class I in types (data not shown). In fact, we found that CD45RO is mostly associ- tumor cells through the ERK5 pathway. J. Immunol. 185, 3498 –3503. ated with effector NK cells; however, differently from what observed in Della Chiesa, M., Muccio, L., Moretta, A., 2013. CMV induces rapid NK cell maturation in – most T cell populations, CD45RA down-regulation is not required for NK HSCT recipients. Immunol. Lett. 155, 11 13. Domaica, C.I., Fuertes, M.B., Uriarte, I., Girart, M.V., Sardanons, J., Comas, D.I., Di Giovanni, cell activation. This suggests that in NK cells the expression of different D., Gaillard, M.I., Bezrodnik, L., Zwirner, N.W., 2012. Human natural killer cell matura- CD45 isoforms plays a different role than in T cells. tion defect supports in vivo CD56(bright) to CD56(dim) lineage development. PLoS In summary, we show here that NK cells that recognize tumor cells One 7, e51677. Dunn, G.P., Bruce, A.T., Ikeda, H., Old, L.J., Schreiber, R.D., 2002. Cancer immunoediting: are present in all examined patients of hematological cancers. Hence, from immunosurveillance to tumor escape. Nat. Immunol. 3, 991 –998. NK cells are actively recognizing tumor cells in leukemia patients; but Earl, L.A., Bi, S., Baum, L.G., 2010. N- and O-glycans modulate galectin-1 binding, CD45 – this seems to be insuf ficient to eradicate disease. Protocols to enhance signaling, and T cell death. J. Biol. Chem. 285, 2232 2244. Elpek, K.G., Rubinstein, M.P., Bellemare-Pelletier, A., Goldrath, A.W., Turley, S.J., 2010. such population should improve patient's prognosis. Finally, the pres- Mature natural killer cells with phenotypic and functional alterations accumulate ence of this population identi fies blood-borne cancer patients. upon sustained stimulation with IL-15/IL-15Ralpha complexes. Proc. Natl. Acad. Sci. U. S. A. 107, 21647 –21652. Fogel, L.A., Sun, M.M., Geurs, T.L., Carayannopoulos, L.N., French, A.R., 2013. Markers of nonselective and speci fic NK cell activation. J. Immunol. 190, 6269 –6276. 4. Financial Support and Acknowledgements Freud, A.G., Yu, J., Caligiuri, M.A., 2014. Human natural killer cell development in second- ary lymphoid tissues. Semin. Immunol. 26, 132 –137. Fu, X., Liu, Y., Li, L., Li, Q., Qiao, D., Wang, H., Lao, S., Fan, Y., Wu, C., 2011. Human natural All our funders are public or charitable organizations. This work was killer cells expressing the memory-associated marker CD45RO from tuberculous supported by the program “Chercheur d'avenir ” from the Region pleurisy respond more strongly and rapidly than CD45RO-natural killer cells follow- Languedoc-Roussillon (09-13195) (MV), a scienti fic program from the ing stimulation with interleukin-12. Immunology 134, 41 –49. “Communauté de Travail des Pyrénées ” (CTPP5/12 to MV), the charities Fujisaki, H., Kakuda, H., Shimasaki, N., Imai, C., Ma, J., Lockey, T., Eldridge, P., Leung, W.H., Ca mpana, D., 2009. Expansion of highly cytotoxic human natural killer cells for cancer CIEL, L'Un pour l'Autre and Ensangble (09/2013) (MV), a grant from cell therapy. Cancer Res. 69, 4010 –4017. FEDER Objectif Competitivite (10-007762) (MV), a grant from the Garaude, J., Kaminski, S., Charni, S., Aguilo, J.I., Jacquet, C., Plays, M., Hernandez, J., European Community Program SUDOE (CLiNK SOE2/P1/E341 to MV Rodriguez, F., Hipskind, R.A., Anel, A., Villalba, M., 2008. Impaired anti-leukemic immune response in PKCtheta-de ficient mice. Mol. Immunol. 45, 3463 –3469. and J-FR), an AOI from the CHU Montpellier (N°221826) (GC and MV) Hermiston, M.L., Zikherman, J., Zhu, J.W., 2009. CD45, CD148, and Lyp/Pep: critical phos- and fellowships from the ARC (DOC20121206007) and La Ligue Contre phatases regulating Src family kinase signaling networks in immune cells. Immunol. – le Cancer (TDKB13362) (EK) and Ministère de l'Enseignement Supérieur Rev. 228, 288 311. Hesslein, D.G., Takaki, R., Hermiston, M.L., Weiss, A., Lanier, L.L., 2006. Dysregulation of et de la Recherche (MESR) (DNV). FACs analysis was performed at the signaling pathways in CD45-de ficient NK cells leads to differentially regulated cyto- platform Montpellier Rio Imaging (MRI). toxicity and cytokine production. Proc. Natl. Acad. Sci. U. S. A. 103, 7012 –7017. 1376 E. Krzywinska et al. / EBioMedicine 2 (2015) 1364 –1376

Hesslein, D.G., Palacios, E.H., Sun, J.C., Beilke, J.N., Watson, S.R., Weiss, A., Lanier, L.L., 2011. Ruggeri, L., Mancusi, A., Burchielli, E., Aversa, F., Martelli, M.F., Velardi, A., 2007. Natural Differential requirements for CD45 in NK-cell function reveal distinct roles for Syk- killer cell alloreactivity in allogeneic hematopoietic transplantation. Curr. Opin. family kinases. Blood 117, 3087 –3095. Oncol. 19, 142 –147. Huntington, N.D., Xu, Y., Nutt, S.L., Tarlinton, D.M., 2005. A requirement for CD45 distin- Sanchez-Martinez, D., Krzywinska, E., Rathore, M.G., Saumet, A., Cornillon, A., Lopez- guishes Ly49D-mediated cytokine and chemokine production from killing in primary Royuela, N., Martinez-Lostao, L., Ramirez-Labrada, A., Lu, Z.Y., Rossi, J.F., Fernandez- natural killer cells. J. Exp. Med. 201, 1421 –1433. Orth, D., Escorza, S., Anel, A., Lecellier, C.H., Pardo, J., Villalba, M., 2014. All-trans Irie-Sasaki, J., Sasaki, T., Matsumoto, W., Opavsky, A., Cheng, M., Welstead, G., Grif fiths, E., retinoic acid (ATRA) induces miR-23a expression, decreases CTSC expression and Krawczyk, C., Richardson, C.D., Aitken, K., Iscove, N., Koretzky, G., Johnson, P., Liu, P., granzyme B activity leading to impaired NK cell cytotoxicity. Int. J. Biochem. Cell Rothstein, D.M., Penninger, J.M., 2001. CD45 is a JAK phosphatase and negatively reg- Biol. 49, 42 –52. ulates cytokine receptor signalling. Nature 409, 349 –354. Sanchez-Martinez, D., Azaceta, G., Muntasell, A., Aguilo, N., Nunez, D., Galvez, E.M., Naval, Ito, K., Stannard, K., Gabutero, E., Clark, A.M., Neo, S.Y., Onturk, S., Blanchard, H., Ralph, S.J., J., Anel, A., Palomera, L., Vilches, C., Marzo, I., Villalba, M., Pardo, J., 2015. Human NK 2012. Galectin-1 as a potent target for cancer therapy: role in the tumor microenvi- cells activated by EBV lymphoblastoid cells overcome anti-apoptotic mechanisms of ronment. Cancer Metastasis Rev. 31, 763 –778. drug resistance in haematological cancer cells. Oncoimmunology 4, e991613. Kaminski, S., Adjali, O., Jacquet, C., Garaude, J., Keriel, A., Lassaux, A., Hipskind, R., Sitbon, Senovilla, L., Vacchelli, E., Galon, J., Adjemian, S., Eggermont, A., Fridman, W.H., Sautes- M., Taylor, N., Villalba, M., 2012. The protooncogene Vav1 regulates murine leukemia Fridman, C., Ma, Y., Tartour, E., Zitvogel, L., Kroemer, G., Galluzzi, L., 2012. Trial virus-induced T-cell leukemogenesis. Oncoimmunology 1, 600 –608. watch: prognostic and predictive value of the immune in filtrate in cancer. Kaplan, R., Morse, B., Huebner, K., Croce, C., Howk, R., Ravera, M., Ricca, G., Jaye, M., Oncoimmunology 1, 1323 –1343. Schlessinger, J., 1990. Cloning of three human tyrosine phosphatases reveals a Siewiera, J., El Costa, H., Tabiasco, J., Berrebi, A., Cartron, G., Le Bouteiller, P., Jabrane-Ferrat, multigene family of receptor-linked protein-tyrosine-phosphatases expressed in N., 2013. Human cytomegalovirus infection elicits new decidual natural killer cell ef- brain. Proc. Natl. Acad. Sci. U. S. A. 87, 7000 –7004. fector functions. PLoS Pathog. 9, e1003257. Karre, K., Ljunggren, H.G., Piontek, G., Kiessling, R., 1986. Selective rejection of H-2- Skak, K., Frederiksen, K.S., Lundsgaard, D., 2008. Interleukin-21 activates human natural de ficient lymphoma variants suggests alternative immune defence strategy. Nature killer cells and modulates their surface receptor expression. Immunology 123, 319, 675 –678. 575 –583. Kishihara, K., Penninger, J., Wallace, V.A., Kundig, T.M., Kawai, K., Wakeham, A., Timms, E., Soma, L., Wu, D., Chen, X., Edlefsen, K., Fromm, J.R., Wood, B., 2015. Apparent CD19 ex- Pfeffer, K., Ohashi, P.S., Thomas, M.L., et al., 1993. Normal B lymphocyte development pression by natural killer cells: a potential confounder for minimal residual disease but impaired T cell maturation in CD45-exon6 protein tyrosine phosphatase- detection by flow cytometry in B lymphoblastic leukemia. Cytometry B Clin. Cytom. de ficient mice. Cell 74, 143 –156. 88, 145 –147. Kung, C., Pingel, J.T., Heikinheimo, M., Klemola, T., Varkila, K., Yoo, L.I., Vuopala, K., Stern, M., Ruggeri, L., Mancusi, A., Bernardo, M.E., De Angelis, C., Bucher, C., Locatelli, F., Poyhonen, M., Uhari, M., Rogers, M., Speck, S.H., Chatila, T., Thomas, M.L., 2000. Aversa, F., Velardi, A., 2008. Survival after T cell-depleted haploidentical stem cell Mutations in the tyrosine phosphatase CD45 gene in a child with severe combined transplantation is improved using the mother as donor. Blood 112, 2990 –29 95. immunode ficiency disease. Nat. Med. 6, 343 –345. Suzuki, E., Kataoka, T.R., Hirata, M., Kawaguchi, K., Nishie, M., Haga, H., Toi, M., 2015. Lanier, L.L., 2008. Up on the tightrope: natural killer cell activation and inhibition. Nat. Trogocytosis-mediated expression of HER2 on immune cells may be associated Immunol. 9, 495 –502. with a pathological complete response to trastuzumab-based primary systemic ther- Lianidou, E.S., Mavroudis, D., Georgoulias, V., 2013. Clinical challenges in the molecular apy in HER2-overexpressing breast cancer patients. BMC Cancer 15, 39. characterization of circulating tumour cells in breast cancer. Br. J. Cancer 108, Tchilian, E.Z., Wallace, D.L., Wells, R.S., Flower, D.R., Morgan, G., Beverley, P.C., 2001. A de- 2426 –2432. letion in the gene encoding the CD45 antigen in a patient with SCID. J. Immunol. 166, Lynch, K.W., Weiss, A., 2000. A model system for activation-induced alternative splicing of 1308 –1313. CD45 pre-mRNA in T cells implicates protein kinase C and Ras. Mol. Cell. Biol. 20, Van Den Broek, M.F., Kagi, D., Zinkernagel, R.M., Hengartner, H., 1995. Perforin depen- 70 –80. dence of natural killer cell-mediated tumor control in vivo. Eur. J. Immunol. 25, Mamessier, E., Bertucci, F., Sabatier, R., Birnbaum, D., Olive, D., 2012. “Stealth ” tumors: 3514 –3516. breast cancer cells shun NK-cells anti-tumor immunity. Oncoimmunology 1, Velardi, A., 2008. Role of KIRs and KIR ligands in hematopoietic transplantation. Curr. 366 –368. Opin. Immunol. 20, 581 –587. Mamessier, E., Pradel, L.C., Thibult, M.L., Drevet, C., Zouine, A., Jacquemier, J., Vey, N., Bourhis, J.H., Boissel, N., Bordessoule, D., Prebet, T., Charbonnier, A., Etienne, A., Houvenaeghel, G., Bertucci, F., Birnbaum, D., Olive, D., 2013. Peripheral blood NK Andre, P., Romagne, F., Benson, D., Dombret, H., Olive, D., 2012. A phase 1 trial of cells from breast cancer patients are tumor-induced composite subsets. J. Immunol. the anti-inhibitory KIR mAb IPH2101 for AML in complete remission. Blood 120, 190, 2424 –2436. 4317 –4323. Marcais, A., Walzer, T., 2014. mTOR: a gate to NK cell maturation and activation. Cell Cycle Villalba, M., Rathore, M.G., Lopez-Royuela, N., Krzywinska, E., Garaude, J., Allende-Vega, N., 13, 3315 –3316. 2013. From tumor cell metabolism to tumor immune escape. Int. J. Biochem. Cell Biol. Mason, L.H., Willette-Brown, J., Taylor, L.S., Mcvicar, D.W., 2006. Regulation of Ly49D/ 45, 106 –113. DAP12 signal transduction by Src-family kinases and CD45. J. Immunol. 176, Villalba, M., Lopez-Royuela, N., Krzywinska, E., Rathore, M.G., Hipskind, R.A., Haouas, H., 6615 –6623. Allende-Vega, N., 2014. Chemical metabolic inhibitors for the treatment of blood- Mee, P.J., Turner, M., Basson, M.A., Costello, P.S., Zamoyska, R., Tybulewicz, V.L., 1999. borne cancers. Anti Cancer Agents Med. Chem. 14, 223 –232. Greatly reduced ef ficiency of both positive and negative selection of thymocytes in Vivier, E., Tomasello, E., Baratin, M., Walzer, T., Ugolini, S., 2008. Functions of natural killer CD45 tyrosine phosphatase-de ficient mice. Eur. J. Immunol. 29, 2923 –2933. cells. Nat. Immunol. 9, 503 –510. Moretta, L., 2010. Dissecting CD56dim human NK cells. Blood 116, 3689 –3691. Wang, J., Pantopoulos, K., 2011. Regulation of cellular iron metabolism. Biochem. J. 434, Mustelin, T., Vang, T., Bottini, N., 2005. Protein tyrosine phosphatases and the immune 365 –381. response. Nat. Rev. Immunol. 5, 43 –57. Warren, H.S., Skipsey, L.J., 1991. Loss of activation-induced CD45RO with maintenance of Nakamura, K., Nakayama, M., Kawano, M., Ishii, T., Harigae, H., Ogasawara, K., 2013. NK- CD45RA expression during prolonged culture of T cells and NK cells. Immunology 74, cell fratricide: dynamic crosstalk between NK and cancer cells. Oncoimmunology 2, 78 –85. e26529. Willemze, R., Rodrigues, C.A., Labopin, M., Sanz, G., Michel, G., Socie, G., Rio, B., Sirvent, A., Pardo, J., Balkow, S., Anel, A., Simon, M.M., 2002. Granzymes are essential for natural killer Renaud, M., Madero, L., Mohty, M., Ferra, C., Garnier, F., Loiseau, P., Garcia, J., Lecchi, L., cell-mediated and perf-facilitated tumor control. Eur. J. Immunol. 32, 2881 –2887. Kogler, G., Beguin, Y., Navarrete, C., Devos, T., Ionescu, I., Boudjedir, K., Herr, A.L., Ramírez-Comet, N., Aguiló, J.I., Rathoré, M.G., Catalán, E., Garaude, J., Uze, G., Naval, J., Gluckman, E., Rocha, V., 2009. KIR-ligand incompatibility in the graft-versus-host di- Pardo, J., Anel, A., Villalba, M., 2014. IFN- α signaling through PKC- θ is essential for rection improves outcomes after umbilical cord blood transplantation for acute leu- anti-tumor NK cell function. Oncoimmunology 3. http://dx.doi.org/10.4161/ kemia. Leukemia 23, 492 –500. 21624011.2014.948705 . Xu, Z., Weiss, A., 2002. Negative regulation of CD45 by differential homodimerization of Rhee, I., Veillette, A., 2012. Protein tyrosine phosphatases in lymphocyte activation and the alternatively spliced isoforms. Nat. Immunol. 3, 764 –771. autoimmunity. Nat. Immunol. 13, 439 –447. Zarcone, D., Prasthofer, E.F., Malavasi, F., Pistoia, V., Lobuglio, A.F., Grossi, C.E., 1987. Ultra- Roth, M.D., 1994. Interleukin 2 induces the expression of CD45RO and the memory phe- structural analysis of human natural killer cell activation. Blood 69, 1725 –1736. notype by CD45RA+ peripheral blood lymphocytes. J. Exp. Med. 179, 857 –864. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! Ñ~! ! ! ! ! ! ! ! ! ! ! ! ! ! !

!'".1#/!gZ!! !

,)!68DD!46L?N4L?GF!4F7!J86GN8JQ!GGE4!H4L?8FLK!4632+/=?7183+G+(:1;799@:;8;5DJ+ ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÑÄ! ONCOIMMUNOLOGY 2018, VOL. 7, NO. 4, e1409322 (11 pages) https://doi.org/10.1080/2162402X.2017.1409322

ORIGINAL RESEARCH NK cell activation and recovery of NK cell subsets in lymphoma patients after obinutuzumab and lenalidomide treatment

Dang-Nghiem Vo a,b, Catherine Alexia a, Nerea Allende-Vega a,b, Franck Morschhauser c, Roch Houot d,e, Cedric Menard e,f, Karin Tartee,f, Guillaume Cartron g,h, and Martin Villalba a,b aINSERM U1183, Universite de Montpellier 1, UFR M edecine, Montpellier, France; bInstitute for Regenerative Medicine and Biotherapy (IRMB), CHU Montpellier, Montpellier, France; cUniv. Lille, CHU Lille, EA 7365 – GRITA – Groupe de Recherche sur les formes Injectables et les Technologies Associees, Lille, France; dDepartment of Clinical Hematology, University Hospital Rennes, Rennes, France; eUMR U1236, INSERM Universite Rennes 1, Etablissement Francais¸ du Sang, Rennes, France; fSITI, P^ole de Biologie, CHU de Rennes, Rennes, France; gDepartement d’Hematologie Clinique, CHU Montpellier, Universite Montpellier I, 80 avenue Augustin Fliche, Montpellier, France; hCNRS UMR5235, Universite de Montpellier, Montpellier, France

ABSTRACT ARTICLE HISTORY Obinutuzumab (OBZ) shows stronger antibody-dependent cell cytotoxicity (ADCC) compared to Received 29 September 2017 rituximab and improved clinical activity for treating certain CD20 C neoplasia. However, the ef ficacy Revised 20 November 2017 of monoclonal antibody (mAb) as a monotherapy is limited. Natural Killer (NK) cells are mediators of Accepted 20 November 2017 ADCC. Hematological cancer patients possess antitumor NK cells that are unable to control disease, KEYWORDS possibly because they are dysfunctional. The immunomodulatory drug lenalidomide (LEN) could be diffuse large B-cell a treatment to restore exhausted NK cell cytotoxic functions. The clinical trial GALEN is a Phase Ib/II lymphoma (DLBCL); follicular study of OBZ combined with LEN for the treatment of relapsed/refractory follicular and aggressive lymphoma (FL); (DLBCL and MCL) B-cell Lymphoma. During treatment, we analyzed speci fic aspects of NK cell lenalidomide; obinutuzumab; biology. Treatment reversed the immature NK phenotype of patients and increased expression of NK NK cell activating receptors. Inhibitory receptors were either unchanged or decreased. There was a strong NK response at the end of the 1st cycle: NK number and intracellular granzyme B (GrzB) expression decreased, degranulation increased and NK responded better to allogeneic target challenge. Moreover, the interaction of NK cells with B cell targets, measured by trogocytosis, decreased during treatment. At the end of treatment, when target cells had been wiped out, the proportion of reactive NK cells (CD69 C, CD45RARO C, CD107a C, CD19 C) strongly decreased. Because all patients received LEN and OBZ, it was uncertain which drug was responsible of our observations, or even if a combination of both products was necessary for the described effects on this lymphocyte lineage.

Introduction lymphocytic leukemia (CLL). 5 This clinical bene fit has been The anti-CD20 IgG1 monoclonal antibody (mAb) rituximab observed in other B-cell malignancies. 4,6,7 OBZ is approved (RTX) has improved the treatment of B-cells lymphocytic for first-line CLL in association with chlorambucil and in leukemia (B-CLL) and B-cells non-Hodgkin lymphomas (B- combination with bendamustine for the treatment of NHL). Its success is related to its capacity to induce Fc- patients with follicular lymphoma (FL) who relapse or are (antibody-dependent cell-mediated cytotoxicity (ADCC). refractory to RTX-containing regimen. 8 One receptor for human IgG1 is Fc gRIIIa (CD16 a), which However, it is remarkable to note that the mAbs themselves is expressed on natural killer (NK) cells and macrophages. have modest clinical activity. For example, RTX or OBZ when The in fluence of Fc gRIIIa-158VF polymorphism on RTX used as monotherapy in patients with relapsed follicular lym- clinical response strongly suggests that ADCC is critical.1 phoma have demonstrated short progression-free survival Based on these results, there has been an attempt to pro- (PFS).8 These data indicate that there is a need to optimize their duce new anti-CD20 mAbs that exhibit higher af finity for use in co-therapy. In this sense, hematological cancer patients Fc gRIIIa either by Fc mutations or by glycoengeenering. 2,3 possess antitumor NK cells that are unable to control This later strategy, leading to low fucose content of the N- disease.9,10 Blood-borne cancer cells use different mechanisms glycan, is currently under clinical investigations in B-cell for immune escape, 11 ,12 e.g. inducing NK cell dysfunction. 13 ,14 malignancies with the mAb obinutuzumab (OBZ; previously In addition, NK cell differentiation may be inhibited by the GA101, Roche, Genentech), which shows stronger ADCC in presence of tumor cells e.g. acute myeloid leukemia (AML) cells vitro and in a lymphoma xenograft mouse model compared in filtrating bone-marrow.15 ,16 Therefore, the failure of mAb as to RTX 4 and improved clinical activity for treating chronic monotherapy could be related to impaired NK cell function

CONTACT Martin Villalba (MV) [email protected] INSERM U1183, Institute for Regenerative Medicine and Biotherapy. CHU Montpellier H ^opital Saint- Eloi 80, av. Augustin Fliche. 34295 Montpellier Cedex 5 France. Supplemental data for this article can be accessed on the publisher’s website . © 2018 Taylor & Francis Group, LLC e1409322-2 D.-N. VO ET AL. and hence, there is a clinical interest to reactivate patient NK showed increased expression of the major histocompatibility cells.17 complex-I (MHC-I), as has been observed in other hemato- Lenalidomide (LEN; Revlimid; Celgene) is an immune-mod- logical neoplasias 14 ; but also of the stress ligands MHC class – ulatory drug that can activate NK cells. 14 ,18 21 LEN treatment I polypeptide-related sequence A (MICA) and MICB during and after stem cell transplantation (SCT) increases (Fig. 1D ). The increased expression of MHC-I and MICA/B NK cell proliferation, enhances NKp44 expression on NK could have countervailing effects on NK cells because they cells14 and increases circulating NK-cell numbers in leukemia are recognized by KIRs, inhibitory receptors, and NKG2D, patients.22 ,23 LEN increases co-stimulatory receptor expression activating receptor, respectively. Hence the final effect on on NK cells, such as CD16 and Lymphocytes Function-associ- NK cell recognition in remaining target cells is unclear. ated Antigen (LFA)14 and stabilizes NK cell:target cell immuno- logical synapse.20 ,23 ,24 These effects lead to increased cytotoxic Treatment induces maturation of the immature NK cell activity and increased proliferation of LEN-stimulated NK population cells.14 ,19 ,20 LEN has similar effects in B-NHL patients restoring synapse formation, ADCC, and cytotoxic functions in NK We next directly investigated the physiological status of NK cells.25 ,26 Of particular clinical importance, LEN allows NK cells cells during treatment. In the peripheral blood, human NK cells to be activated by lower doses of RTX. 20 Finally, it also favors are mostly CD3 ¡CD56dim cells with high cytotoxic activity, target recognition by inducing expression of NKG2D and while CD3¡CD56bright cells excel in cytokine production 28 . In DNAM-1 ligands on malignant cells. 27 LEN mechanism of vitro evidence indicates that CD56 bright NK cells are precursors action is thus predominantly immune-mediated, making LEN of CD56 dim NK cells and this might also be the case in vivo 29 . a suitable treatment to restore exhausted NK cell cytotoxic In addition, combined analysis of CD56 and CD16 expression functions. during NK cell development indicates that their pro files With this view, the clinical trial GALEN is a Phase Ib/II changes as follows: CD56 brightCD16¡ ! CD56brightCD16dim ! study of OBZ combined with LEN for the treatment of CD56dim CD16dim ! CD56dim CD16C. Additional markers can relapsed/refractory follicular and aggressive B-cell lym- be used to identify speci fic subsets within these NK cell popula- phoma (diffuse large B-cell lymphoma (DLBCL) and mantle tions30 ,31 . As previously described 9,10 , we observed a tendency cell lymphoma (MCL) by the LYSA Lymphoma Study Asso- to a higher proportion of immature NK cells in patients com- ciation. The primary objective of the Phase IB part of the pare to HD, which correlated with a decrease in the full mature study was to determine the recommended dose (RD) of CD56dim CD16C (Fig. 2A ). At the end of treatment most LEN when administered in association with OBZ. The pri- patients lost the immature subsets and gained a NK distribu- mary objective of the Phase II part of the study was to tion similar to healthy donors ( Fig. 2A ), i.e. with less immature assess the ef ficacy of the association of the recommended cells. dose of LEN in combination with OBZ, as measured by the We next analyzed the maturation marker CD161-killer cell overall response rate (ORR) at the end of 6 cycles in these lectin-like receptor subfamily B, member 1 (KLRB1) that is 2 different populations of lymphoma patients. We devel- expressed early in NK cell development and before CD56 32 . oped a pilot exploration of some speci fic aspects of NK cell The expression of this marker did not change during treatment biology. In this respect, we monitored the following time in the CD56 C NK compartment ( Fig. 2B ). points: i) C1D1 predose; ii) C1D28 and iii) C6D28 (supple- During in vivo maturation CD56 bright cells become mental Fig. 1). CD56 dim CD62L CCD57 ¡ cells that produce perforin, while maintaining high IFN- g production in response to cyto- kines 28 ,33 . On the other hand, CD56 dim CD62L ¡CD57 C cells Results show low response to cytokines and higher cytotoxic capac- ity 28 ,34 . CD62L was slightly increased in patients and the Effect of treatment on lymphocyte populations treatment decreased the expression ( Fig. 2B ). In contrast Patients were treated with a combination of LEN (orally CD57 was lower in patients and remained unchanged by administered) and 3 doses of OBZ in the first cycle and a the treatment ( Fig. 2B ). This suggests that at the end of single dose on the first day of the following for total of six treatment the NK cells show decreased expression of an consecutive treatment cycles (see supplemental Fig. 1 for immature marker, i.e. CD62L. When NK cells reach fully treatment and sampling protocol). We did not observe dif- mature CD56 dim CD16 C status, they gain full expression of ferences in the NK cell parameters tested between the dif- killer inhibitory receptors (KIRs) receptors. KIR expression ferent lymphoma types in our pilot study, hence we in patients before and after treatment was variable and analyzed them together (both FL and DLBCL patients). We expression of the 3 KIRs taken together was similar in observed a transient decrease in hemoglobin levels, a signifi- patients and healthy donors ( Fig. 2C ). cant decrease in leucocytes and a trend towards a decrease The CD94 glycoprotein heterodimerizes with the natural- in lymphocytes ( Fig. 1A ). T cell numbers were unchanged killer group 2 (NKG2) receptors, which are type II trans- and there was a transient decrease in NK cells at the end of membrane proteins. CD94/NKG2 A is an inhibitory recep- the first cycle ( Fig. 1B ). B cells (CD19 C) decreased in num- tor that recognizes HLA-E and it is the first inhibitory bers ( Fig. 1C ). The CD20 C population, which is the main receptor expressed during NK cell maturation 32 ,35 . CD94 target of OBZ, showed a tendency to decrease ( Fig. 1C ), can also associate with the activating receptors NKG2C and similar to CD5 C cells ( Fig. 1C ). The remaining CD19 C cells E32 ,35 . The activating receptor NKG2D represents an ONCOIMMUNOLOGY e1409322-3

Figure 1. Effect of treatment on lymphocyte populations. (A) Absolute values of hemoglobin (left), total leukocyte count (middle) and total lymphocyte count (right) are reported as before the first induction (day 0), after first round of treatment (end cycle 1) and after the final round of treatment (end study) respectively (n D 16). (B) Abso- lute count of T lymphocyte (CD3 CCD56-) population and NK cell (CD3-CD56 C) populations from total PBMC (n D 16). (C) Absolute count of lymphocyte populations car- rying specific markers associated with B-cell lymphoma (n D 16). (D) Expression of HLA and the stress ligands MIC-A and MIC-B on CD19 C population in term of mean fluorescence intensity (HD: n D 4; patients: n D 10). Significance was determined by paired t-test between day 0 versus following time-points, and one-way ANOVA between HD and patients at every time-points with à p  0.05, Ãà p  0.01 and ÃÃà p  0.001. exception: it is a homodimer 32 ,35 . CD94 was lower in elevated values remained unchanged during treatment. In con- patients and increased on treatment ( Fig. 2D ). In contrast, trast, levels of the activation marker CD69, which were similar NKG2A was higher in patients and was not modi fied by to healthy donors, decreased at the end of treatment ( Fig. 3A ). treatment ( Fig. 2D ). NKG2D, which was lower in patients, The antitumor NK cell population is easily recognized by signi ficantly increased after treatment ( Fig. 2D ). In sum- expression of CD45RO (CD45RO cells) in general together mary, NK activating receptors tend to increase while inhibi- with CD45RA (CD45RARO cells). Patients show high levels tory receptors are either unchanged or decreased. of these cells, leading to a decrease in the CD45RA CRO ¡ population (CD45RA cells). 9,10 Patients in our cohort clearly showed this phenotype ( Fig. 3B ). At the end of treat- Treatment decreases the activated NK cell population ment this phenotype tended to converge versus a healthy The proliferation marker Ki-67 is increased in NK cells from donor phenotype with increase in CD45RA cells and a hematological cancer patients. 9,10 Fig. 3A showed that the decrease in CD45RO C cells. Taken together these data e1409322-4 D.-N. VO ET AL.

Figure 2. Treatment induces maturation of the immature NK cell population. (A) Analysis of NK cell subpopulations based on level of CD56 and CD16 expression that divides NK cells in 4 subpopulations: CD56 br CD16¡ (top, left panel), CD56br CD16C (top, right panel), CD56dim CD16¡ (bottom, left panel) and CD56dim CD16C (bottom, right panel). (B) Assessment of NK cell maturation status determined by expression of the maturation markers CD161, CD62 L and CD57. (C) Expression of several KIR receptors on healthy donor and patient NK cells: CD158 a (KIR2DL1), CD158b (KIR2DL2/3) and CD158e (KIR3DL1). (D) Expression of the inhibitory heterodimer complex NKG2 A/ CD94 and the activating receptor NKG2D. Patient: n D 16, healthy donor: n D 4. Signi ficance was determined by paired t-test between day 0 versus following time-points, and one-way ANOVA between HD and patients at every time-points with à p  0.05, Ãà p  0.01 and ÃÃà p  0.001. suggest that elimination of target cells decreases NK cell (Fig. 4A ). The amount of granzyme, as measured by the median activation status. fluorescence intensity (MFI) values, was higher in patients, underwent a signi ficant reduction at the end of cycle 1, and recovered to normal values at the end of treatment ( Fig. 4A ). In Treatment modulates NK cell cytotoxic activity agreement with our previous results, 9,10 we observed that more If NK cell target cells were disappearing, we also expect to find a NK cells were degranulating in patients, measured by CD107a decrease in NK degranulation because cytotoxicity is probably expression in the plasma membrane ( Fig. 4B ). At the end of the main antitumor function of these lymphocytes in hemato- treatment the proportion of degranulating cells had signi fi- logical neoplasias.9,10 The proportion of granzyme C cells was cantly decreased (Fig. 4B ). NK cells degranulated at similar lev- similar to healthy donors and slightly decreased after treatment els at the beginning and at the end of the treatment against the ONCOIMMUNOLOGY e1409322-5

Figure 3. Treatment decreases the activated NK cell population. (A) Proliferation potency of NK cell presented as intracellular staining of the nuclear marker Ki-67 and NK activation status determined by expression of the activation marker CD69. (B) Analysis of NK populations based on differential expression of CD45 isoforms: CD45RA C (left panel) and CD45RACR0 C plus CD45R0C (right panel). Patient: n D 16, healthy donor: n D 4. Signi ficance was determined by paired t-test between day 0 versus fol- lowing time-points, and one-way ANOVA between HD and patients at every time-points with à p  0.05, Ãà p  0.01 and ÃÃà p  0.001. allogeneic non-Hodgkin B lymphoma cell line Daudi ( Fig. 4C subset expressed the highest level of CD19 and treatment suc- and D). Interestingly, after the first cycle NK cells were more cessfully decreased CD19 expression (Fig. 5A ). Other subsets active against the allogeneic targets ( Fig. 4C and D). This sug- also decreased CD19 expression (Fig. 5A ). These results suggest gests that the presence of targets cells and the treatment activate that ef ficient treatments leading to elimination of NK cell target NK cells in vivo . CD56 dim cells were responsible for CD107a cells reduce NK: target cell interaction and lead to lack of target expression ex vivo , whereas CD56 bright cells lacked expression antigens on NK cell surface. of this marker ( Fig. 4E ). CD56 dim CD16¡ cells are considered We next investigated which populations were responsible immature cells and precursors of CD56 dim CD16C cells Unex- for other variations in NK cell markers and focused on three of pectedly, CD56dim CD16¡ cells expressed higher CD107a levels them that changed after treatment. CD69 expression was than CD56dim CD16C cells.30 ,31 This is probably related to higher in patients and decreased after treatment ( Fig. 3A ). CD16 downregulation after NK cell activation 36 and also CD45RO C cells showed higher CD69 levels, and treatment did explains the relative high CD56 dim CD16¡ cell numbers in not decrease it. CD45RO¡ cells showed lower levels that patients (Fig. 2A ). decreased with treatment ( Fig. 5B ). Therefore the decrease in CD69 expression is linked to both the decrease in the number of CD45RO C cells, which expressed higher CD69 levels and the Treatment decreases the trogocytosis of tumor-associated decrease of CD69 in CD45RO ¡ cells. markers by NK cells NKG2D expression was lower in patients and increased after These results suggested that NK cells were actively killing their treatment ( Fig. 2D ). NKG2D was lower in CD45RO C cells and targets during treatment and the absence of such targets at the treatment increased expression in both CD45RO ¡ and end of treatment generated resting NK cells. To test this CD45RO C (Fig. 5B ). Hence the increase in NKG2D levels is hypothesis, we investigated the proportion of NK cells that due to the decrease in CD45RO C cell numbers and the increase have killed CD19 C targets at the different time-points. We took in NKG2D expression by all NK cells. advantage of the fact that NK cells gained target cell antigens, CD94 expression was not modified in patients but increased e.g. CD19, by trogocytosis. 9,10 As expected, the percentage of during treatment ( Fig. 2D ). CD45RO C cells expressed less CD19C NK cells was higher in patients than in healthy donors CD94 and both CD45ROC and RO ¡ non-signi ficantly and decreased with treatment ( Fig. 5A ). The CD45RARO NK increased CD94 expression during treatment ( Fig. 5B ). The e1409322-6 D.-N. VO ET AL.

Figure 4. Treatment modulates NK cell cytotoxic activity . (A) Percentage of GrzB C NK cells (left panel) and the MFI of GrzB C population (right panel). (B) Ex vivo degranu- lation of NK cells determined by CD107 a staining (left panel: % of CD107 a C NK cells; right panel: MFI of CD107 a C NK cells. (C) Degranulation of NK cells upon overnight incubation with Daudi target cells at [E:T] D 1:10 determined by % of CD107 a C cell (left panel) and MFI of CD107 a C NK (right panel). (D) The increase in CD107 a in in vitro assays (DCD107 a) was calculated as the difference between the % of CD107 a C in degranulation assay versus the % of CD107 a C NK cell detected ex vivo for each data point. (E) Percentage of the different NK cell subsets that expressed CD107 a ex vivo . Patient: n D 16, healthy donor: n D 4. Signi ficance was determined by paired t-test between day 0 versus following time-points, and one-way ANOVA between HD and patients at every time-points with à p  0.05, Ãà p  0.01 and ÃÃà p  0.001. increase of CD94 on the whole NK cell population is thus stimulatory CD137 receptor and the inhibitory PD-1 receptor related to both, the decrease in CD45RO C cells and the general 17 . We used samples from a different cohort of patients in the increase of CD94. GALEN clinical study and investigated the effect of OBZ treat- ment at the following time points: 1) during LEN course, just before OBZ injection (D7 before OBZ); 2) 1 hour after the end Treatment does not exhaust NK cells of OBZ infusion (D7 after OBZ); 3) at D0 of cycle 2 before Finally, we investigated the effect of treatment on two receptors LEN (cycle 2); and 4) at assessment of clinical response after 6 regulated on NK cells by CD16-mediated activation: the cycles (end of induction). OBZ increased CD137 in both FL ONCOIMMUNOLOGY e1409322-7

Figure 5. Trogocytosis of tumor associated marker on NK cell population. (A) Trogocytosis of CD19 tumor marker on the total NK population (left panel) or on different NK cell subpopulations regarding expression of CD45 isoforms: CD45RA C, CD45RAR0 C, CD45RA dim and CD45R0C. (B) The percentage of NK cells expressing CD69, NKG2D and CD94 was analyzed in the CD45RO ¡ and CD45ROC populations. Patient: n D 16, healthy donor: n D 4. Signi ficance was determined by paired t-test between day 0 versus following time-points, and one-way ANOVA between HD and patients at every time-points with à p  0.05, Ãà p  0.01 and ÃÃà p  0.001. and DLBCL patients (Fig. 6 ). The effect was found within hours first cycle), NK cells from patients degranulated more than after OBZ treatment and increased until the end of treatment. those from healthy donors ( Fig. 4). Once target cells disappear In contrast, PD-1 expression was unchanged ( Fig. 6). This sug- (Fig. 1 ), the activated markers CD69 and CD45RO ( Fig. 3), the gests that OBZ induces NK CD137 receptor expression in the degranulation marker CD107a (Fig. 4) and the marker of trogo- presence of LEN and that continued infusion of the mAb keeps cytosis CD19 (Fig. 5) decrease on NK cell membrane. This sug- levels of this activating receptor high. In contrast, under these gests that is possible to follow disease development by studying conditions, the inhibitory PD-1 receptor was not expressed. NK cell markers; at least, when NK cells are the effectors of the therapy, e.g. some clinical mAb. At the end of first cycle, when target cells are still present in relative numbers, NK cells show Discussion increased cytotoxicity in vitro and ex vivo and low GrzB levels Although RTX-based therapy is ef ficient in a large number of (Fig. 4 ). However, the NK cell number decreases ( Fig. 1 ). We patients, there is still a need for improvement. The develop- propose the following scenario. NK cells are constitutively kill- ment of new mAbs such as OBZ aims to ful fill this demand. ing target cells ( Fig. 5 and9,10 ). Some NK cells die during this However, even the best-designed mAb could be inef ficient in immune response generating an increase in immature cells. some patients if they lack proper effector cells. The use of LEN OBZ and LEN induce improved target cell recognition and NK to activate NK cells could address this problem. Here we cell activation. NK cells degranulate in larger numbers but also observe that treatment with LEN reverses the immature pheno- die in larger numbers. At the end of treatment, most targets type of patient NK cells ( Figs. 2 and 3) and induces expression cells have disappeared and NK cells are no longer dying, so of activating ligands, i.e. NKG2D ( Fig. 2 ) and CD137 ( Fig. 6 ). there is less de novo formation of NK cells and they are becom- During treatment and in the presence of target cells (at end of ing more mature. However, NK cells continue to show high Ki- e1409322-8 D.-N. VO ET AL.

Figure 6. NK cell increased expression of CD137, but not PD1, in FL and DLBCL patients. Expression of CD137 (TNFRSF9) and the exhaustion marker PD1 on NK cells before the treatment (day 0 – cycle 1), before the first infusion of OBZ (day 8 – cycle 1), before the treatment of 2 nd cycle (end cycle 1) and finally at the end of the last cycle (end cycle 6). Statistical significance was determined by paired t-test between “day 0 – cycle 1” and the following time points; p values à à p  0.05, Ãà p  0.01 and ÃÃà p  0.001.

67 levels. Perhaps this can be explained by adaptive differentia- tumor cell apoptosis and blocks bone marrow stromal sup- – tion of NK cells and subsequent growth of a larger population port,39 but also activates immune cells, e.g. NK cells. 14 ,18 21 Our of mature “memory” NK cells, as has been suggested after results support that NK cells are important mediators of the CMV infection. 37 Almost all phenotypic changes observed in clinical benefits of LEN COBZ. NK cells disappear with the lack of target cells suggesting that It is noteworthy that PD-1 is absent on NK cells isolated treatment keeps NK activated only in the presence of target from healthy donors but it is expressed on those from MM cells. Several populations are responsible for the changes in NK patients.40 We observe that there is a large heterogeneity of cell phenotype ( Fig. 5B) although cytotoxicity ex vivo is almost PD-1 expression in our patient cohort, and only a few of them exclusively associated to CD56 dim cells (Fig. 4E). Unexpectedly, constitutively express PD-1 (Fig. 6 ). As discussed above, there we observed that CD45RO C NK cells showed low NKG2D is probably continual production of mature NK cells to replace expression (Fig. 5B). We speculate that once NK cells are those dying during the immune response. These new cells are engaged on killing, NKG2D expression is not anymore probably not exhausted and lack PD-1 expression. Hence, the required. continual renewal of NK cells might preclude PD-1 expression Our results show an increase in immature, CD56 bright, NK on NK cells in some patients. Treatment did not signi ficantly cells in lymphoma patients ( Fig. 2A ). CD56 bright cells produce affect PD-1 expression. This suggests that the role of PD-1 on high cytokine levels. 28 In the context of anti-CD20-induced treatment is minor. Perhaps LEN partially reversed the exhaus- ADCC, we believed that NK cell cytotoxic function would be tion of effector cells as previously suggested. 41 LEN is probably more relevant than cytokine production. In hematological can- the most active treatment (alone or combined with anti-PD-1/ cer patients, cytotoxicity is mainly mediated by CD56 dim NK PD-L1 antibodies or other drugs) able to restore cytotoxic func- cell subsets.9,10 This is con firmed in the current study ( Fig. 4E ). tion to exhausted NK cells. 14 Our results show the hypothesis When we planned our analysis, we decided to maximize the that LEN in combination with OBZ increases several NK cell study of cytotoxic, CD56 dim , NK cells and did not investigate biological parameters associated with maturation and activa- cytokine production because it is believed that CD56 dim cells tion. However, because we did not obtain samples in mono- produce low cytokine levels.28 However, in view of our current therapy, i.e. LEN or OBZ alone, we cannot identify the relative results it would be interesting to investigate the cytokine pro file contribution of these two drugs. of the immature NK cell populations that accumulate in lym- Cancer patients show NK cell subsets that are signi ficantly phoma patients. different of those found in healthy donors. 9,10 However, one LEN targets the E3 ligase cereblon that degrades the Ikaros question was unresolved: what is the fate of these NK cell sub- transcription factors IKZF1 and IKZF3. 38 In vivo , LEN induces sets when their target cells disappear? Here we show for the first ONCOIMMUNOLOGY e1409322-9 time that in our situation the NK cell subsets come back to a Flow cytometry analysis normal situation, i.e. similar to healthy donors, for the vast Isolated PBMCs were stained with 7AAD (Beckman) to identify majority of markers. This is probably related to the disappear- viable cells and with the following -CD45RO-FITC, -CD161- ance of target cells because both LEN and OBZ are NK cell acti- FITC, -CD3-PE, -CD19-PE, -CD62 L-PE, -CD69-PE, -CD314 vating molecules that should not promote NK cell resting (NKG2D)-PE, -CD3-ECD, -CD19-ECD, -CD56-PECy7, CD3- markers. This suggests that NK cells strongly react to effective APC, -CD56-APC, -GzB-AlexaFluor700, -CD19-AlexaFluor treatment and that NK cell monitoring could be interesting to 700, -CD20-APC-AlexaFluor750, -CD45RA-APC-AlexaFluor follow-up anti tumor treatments; mainly those involving mAb 750, -CD5-Paci ficBlue, -CD16-Paci ficBlue, -CD57-Paci ficBlue, therapy. -CD16-KromeOrange (Beckman), -CD158 a-V450, -CD158b- FITC, -CD158 a-PE, -CD107 a-HV500, -Ki-67-V450, HLA- ABC-BV711 (BD Biosciences), MIC-A/B-PE, -CD45RA-FITC, Disclosure information -CD45RO-PE, -CD159 a(NKG2 A)-PE, -CD94-PE-Vio770, fl -CD45RO-APC, -CD19-VioBlue, -CD158e-VioBlue (Miltenyi The authors declare the following con ict of interests: Roch Biotec) antibodies against surface markers. Brie fly, 1 to 10 £ 10 6 Houot: Honoraria from Celgene and Roche; Guillaume Car- cells were incubated with the different antibodies in PBS contain- tron: Honoraria and consultancy from Roche and Celgene ; ing 2% FBS at 4 C for 30 minutes. Cells were then washed with Franck Morschhauser: honoraria Celgene, Roche advisory – m fi PBS and suspended in 200 250 l PBS 2% FBS. Finally, sample boards and scienti c lectures; Karin Tarte: Celgene, Roche for acquisition was performed using Gallios flow cytometer (Beck- advisory boards and scientific lectures; Cedric Menard: Celgene fi man) or Fortessa (BD Biosciences). Acquired samples were later for scienti c lectures. analyzed using Kaluza software v5.1 (Beckman).

In vitro CD107a degranulation assay Patients and methods In vitro degranulation assay was performed to evaluate NK Patients reactivity to the B cell target Daudi by measuring CD107a All patients belong to the BioGALEN study and signed speci fic expression on the surface after cytotoxic granule release. In informed consent form before biological samples collection of summary, isolated PBMC were pre-stained with CD3/CD56 to BioGALEN. This study is recorded in website ClinicalTrials. determine NK frequency in the sample. Next, PBMC were gov with number NCT01582776. Phase Ib was for follicular incubated with Daudi cells at a 1:10 ratio NK:Daudi in the pres- lymphoma (FL) patients and Phase II for follicular and aggres- ence of 1.5 ul of anti-CD107a (BD Biosciences, Franklin Lakes, sive (DLBCL and MCL) B-cell lymphoma patients. 3 £ 3 ml of NJ) and 1 ul Golgi-stop (BD Biosciences) (containing monen- heparinized blood or 4 ml of bone marrow aspirate were col- sin) to inhibit vesicle traf ficking. Cell mixture was then resus- lected at day 0. At the end of first cycle or at the end of treat- pended in RPMI Glutamax 10% supplemented with 10 IU/ml ment (supplemental Fig. 1) 3 £ 3 ml of heparinized blood was Interleukin 2 (eBiosciences) and incubated overnight. After collected. stimulation, cell mixture was collected and stained for FACS using an antibody cocktail containing 7AAD, the anti- CD45RO-FITC, -CD69-PE, -CD19-ECD, -CD56-PECy7, -CD3-APC, -CD45RA-APCAlexaFluor750, -CD107a-HV500 Cell culture and -CD16-KO antibodies (BD Biosciences, Beckman). A The B cell lymphoblastoid Daudi cell line was maintained in bivariate plot of CD56 versus CD3 was used to acquire at least logarithmic growth in RPMI 1640 medium (Gibco Ò Gluta- 10,000 NK cells. MAXTM media) with 10% fetal bovine serum (FBS) (Gibco Ò). Cells were cultured at 37 C in a humidi fied chamber with 5% Multicolor staining for intracellular markers CO 2 in air, and passaged 1:10 twice a week. Cell permeablization and intracellular staining was performed as previous described. 9,10 Brie fly,1 –10 million cells were incu- bated with 10% normal human serum at RT for 15 min and Peripheral blood mononuclear cell (PBMC) puri fication then stained with an antibody mix for cell surface markers Bone marrow and peripheral blood samples were obtained (anti-CD45RO-FITC, -CD19-ECD, -CD56-PC7, -CD3-APC, from patients and total PBMC were isolated using Ficoll. -CD45RA-APCAlexaFluor750 and -CD16-KO antibodies) Briefly, 3 –6 ml of 1:2 diluted blood or 1:3 diluted bone marrow (BD Biosciences, Beckman). After surface staining, cells were samples in RPMI were added on top of 5 ml of Histopaque washed twice and permeabilize with CytoFix/CytoPerm (BD (Sigma). Cells were centrifuged at 1600 rpm and at 20 C with- Biosciences) reagent according to the manufacturer protocol. out break for 30 minutes. Mononuclear cells were collected After fixation and permeablization, cells were washed twice in from the white ring at the interface. After washing in RPMI, BD Perm/Wash solution and follow FACS staining for cells were cryopreserved in liquid nitrogen in medium com- intracellular markers Granzyme B- PE (Miltenyi Biotec) and prise of FBS plus 10% culture-grade DMSO (CliniMACS) until Ki-67-V450 (BD Biosciences) at 4 C for 30 minutes in the analyzing. dark. Finally, cells were washed twice in BD Perm/Wash e1409322-10 D.-N. VO ET AL. solution and resuspended in PBS 2% FBS prior to acquisition the phase II GAUGUIN study. J Clin Oncol. 2013;31:2912 –9. on flow cytometer Gallios (Beckman). A bivariate plot of doi:10.1200/JCO.2012.46.9585. PMID:23835718. CD56 versus CD3 was used to acquire at least 10,000 NK cells. 8. Cartron G, Watier H. Obinutuzumab: what is there to learn from clin- ical trials? Blood. 2017;130:581–9. doi:10.1182/blood-2017-03-771832. PMID:28584136. 9. Krzywinska E, Allende-Vega N, Cornillon A, Vo D, Cayrefourcq L, Statistics Panabieres C, Vilches C, D echanet-Merville J, Hicheri Y, Rossi JF, fi Experimental figures and statistical analysis were performed et al. Identi cation of anti tumor cells carrying natural killer (NK) cell fi antigens in patients with hematological cancers. EBioMedicine. using GraphPad Prism (v6.0). Statistical signi cance between 2015;2:1364–76. doi:10.1016/j.ebiom.2015.08.021. PMID:26629531. day 0 and the following time-points was determined using 10. Krzywinska E, Cornillon A, Allende-Vega N, Vo DN, Rene C, Lu ZY, paired Student t-test on the sample patients for each sampling Pasero C, Olive D, Fegueux N, Ceballos P, et al. CD45 Isoform Pro file point. To determine statistical signi ficance between healthy Identifies Natural Killer (NK) Subsets with Differential Activity. PLoS donors and patients, one-way ANOVA test was used to com- One. 2016;11:e0150434. doi:10.1371/journal.pone.0150434. PMID:27100180. pare between healthy donors versus patients at every time- Ã ÃÃ 11. Villalba M, Rathore MG, Lopez-Royuela N, Krzywinska E, Garaude J, points. All statistical values presented as : p <0.05; : p <0.01; Allende-Vega N. From tumor cell metabolism to tumor immune ÃÃÃ : p <0.001. Average values were expressed as mean plus or escape. Int J Biochem Cell Biol. 2013;45:106 –13. doi:10.1016/j. minus the standard error (SD). biocel.2012.04.024. PMID:22568930. 12. Villalba M, Lopez-Royuela N, Krzywinska E, Rathore MG, Hipskind RA, Haouas H, Allende-Vega N. Chemical metabolic inhibitors for Acknowledgments the treatment of blood-borne cancers. Anti-Cancer Agents Med Chem. 2014;14:223 –32. doi:10.2174/18715206113136660374. FACs analysis was performed at the platform Montpellier Rio Imaging PMID:24237221. (MRI). 13. Baier C, Fino A, Sanchez C, Farnault L, Rihet P, Kahn-Perles B, Cost- ello RT. Natural Killer Cells Modulation in Hematological Malignan- cies. Front Immunol. 2013;4:459. doi:10.3389/ fimmu.2013.00459. Funding details PMID:24391641. 14. Giuliani M, Janji B, Berchem G. Activation of NK cells and disruption This project has been conducted owing to the Bio_GALEN funding as a of PD-L1/PD-1 axis: two different ways for lenalidomide to block part of the ancillary GALEN study. DNV has received a fellowship from myeloma progression. Oncotarget. 2017;8:24031–44. PMID:28199990. the Ministere de l ’Enseignement Superieur et de la Recherche (MESR). 15. Ambrosini P, Loiacono F, Conte R, Moretta L, Vitale C, Mingari MC. IL-1beta inhibits ILC3 while favoring NK-cell maturation of umbilical cord blood CD34(C) precursors. Eur J Immunol. 2015;45:2061 –71. ORCID doi:10.1002/eji.201445326. PMID:25847448. 16. Moretta L, Pietra G, Vacca P, Pende D, Moretta F, Bertaina A, Mingari Martin Villalba http://orcid.org/0000-0002-4385-4888 MC, Locatelli F, Moretta A. Human NK cells: From surface receptors to clinical applications. Immunol lett. 2016;178:15 –9. doi:10.1016/j. imlet.2016.05.007. PMID:27185471. References 17. Muntasell A, Ochoa MC, Cordeiro L, Berraondo P, Lopez-Diaz de Cerio A, Cabo M, L opez-Botet M, Melero I. Targeting NK-cell check- 1. Cartron G, Dacheux L, Salles G, Solal-Celigny P, Bardos P, Colombat P, points for cancer immunotherapy. Curr Opin Immunol. 2017;45:73– Watier H. Therapeutic activity of humanized anti-CD20 monoclonal anti- 81. doi:10.1016/j.coi.2017.01.003. PMID:28236750. body and polymorphism in IgG Fc receptor FcgammaRIIIa gene. Blood. 18. Benson DM, Jr., Cohen AD, Jagannath S, Munshi NC, Spitzer G, Hof- 2002;99:754 –8. doi:10.1182/blood.V99.3.754. PMID:11806974. meister CC, Efebera YA, Andre P, Zerbib R, Caligiuri MA. A Phase I 2. Jefferis R. Glycosylation as a strategy to improve antibody-based ther- Trial of the Anti-KIR Antibody IPH2101 and Lenalidomide in apeutics. Nat Rev Drug Discov. 2009;8:226–34. doi:10.1038/nrd2804. Patients with Relapsed/Refractory Multiple Myeloma. Clin Cancer PMID:19247305. Res. 2015;21:4055 –61. doi:10.1158/1078-0432.CCR-15-0304. 3. Evans SS, Clemmons AB. Obinutuzumab: A Novel Anti-CD20 Mono- PMID:25999435. clonal Antibody for Chronic Lymphocytic Leukemia. J Adv Pract 19. Grossenbacher SK, Aguilar EG, Murphy WJ. Leveraging natural killer Oncol. 2015;6:370–4. PMID:26705497. cells for cancer immunotherapy. Immunotherapy. 2017;9:487–97. 4. Wang W, Erbe AK, Hank JA, Morris ZS, Sondel PM. NK Cell-Medi- doi:10.2217/imt-2017-0013. PMID:28472904. ated Antibody-Dependent Cellular Cytotoxicity in Cancer Immuno- 20. Lagrue K, Carisey A, Morgan DJ, Chopra R, Davis DM. Lenalidomide therapy. Front immunol. 2015;6:368. doi:10.3389/fimmu.2015.00368. augments actin remodeling and lowers NK-cell activation thresholds. PMID:26284063. Blood. 2015;126:50 –60. doi:10.1182/blood-2015-01-625004. 5. Goede V, Fischer K, Busch R, Engelke A, Eichhorst B, Wendtner CM, PMID:26002964. Chagorova T, de la Serna J, Dilhuydy MS, Illmer T, et al. Obinutuzu- 21. Paiva B, Mateos MV, Sanchez-Abarca LI, Puig N, Vidriales MB, mab plus chlorambucil in patients with CLL and coexisting condi- Lopez-Corral L, Corchete LA, Hernandez MT, Bargay J, de Arriba F, tions. N Engl J Med. 2014;370:1101 –10. doi:10.1056/NEJMoa1313984. et al. Immune status of high-risk smoldering multiple myeloma PMID:24401022. patients and its therapeutic modulation under LenDex: a longitudinal 6. Salles GA, Morschhauser F, Solal-Celigny P, Thieblemont C, Lamy T, analysis. Blood. 2016;127:1151 –62. doi:10.1182/blood-2015-10- Tilly H, Gyan E, Lei G, Wenger M, Wassner-Fritsch E, et al. Obinutu- 662320. PMID:26668134. zumab (GA101) in patients with relapsed/refractory indolent non- 22. Chanan-Khan AA, Chitta K, Ersing N, Paulus A, Masood A, Sher T, Hodgkin lymphoma: results from the phase II GAUGUIN study. J Swaika A, Wallace PK, Mashtare TL Jr, Wilding G, et al. Biological effects Clin Oncol. 2013;31:2920 –6. doi:10.1200/JCO.2012.46.9718. and clinical signi ficance of lenalidomide-induced tumour flare reaction in PMID:23835715. patients with chronic lymphocytic leukaemia: in vivo evidence of immune 7. Morschhauser FA, Cartron G, Thieblemont C, Solal-Celigny P, activation and antitumour response. Br J Haematol. 2011;155:457 –67. Haioun C, Bouabdallah R, Feugier P, Bouabdallah K, Asikanius E, Lei doi:10.1111/j.1365-2141.2011.08882.x. PMID:22010965. G, et al. Obinutuzumab (GA101) monotherapy in relapsed/refractory 23. Hagner PR, Chiu H, Ortiz M, Apollonio B, Wang M, Couto S, Wald- diffuse large b-cell lymphoma or mantle-cell lymphoma: results from man MF, Flynt E, Ramsay AG, Trotter M, et al. Activity of ONCOIMMUNOLOGY e1409322-11

lenalidomide in mantle cell lymphoma can be explained by NK cell- 33. Juelke K, Killig M, Luetke-Eversloh M, Parente E, Gruen J, Morandi B, mediated cytotoxicity. Br J Haematol. 2017;179 :399 –409. doi:10.1111/ Ferlazzo G, Thiel A, Schmitt-Knosalla I, Romagnani C. CD62 L bjh.14866. PMID:28771673. expression identifies a unique subset of polyfunctional CD56 dim NK 24. Shanafelt TD, Ramsay AG, Zent CS, Leis JF, Tun HW, Call TG, LaP- cells. Blood. 2010;116:1299–307. doi:10.1182/blood-2009-11-253286. lant B, Bowen D, Pettinger A, Jelinek DF, et al. Long-term repair of T- PMID:20505160. cell synapse activity in a phase II trial of chemoimmunotherapy fol- 34. Lopez-Verges S, Milush JM, Pandey S, York VA, Arakawa-Hoyt J, lowed by lenalidomide consolidation in previously untreated chronic Pircher H, Norris PJ, Nixon DF, Lanier LL. CD57 de fines a function- lymphocytic leukemia (CLL). Blood. 2013;121:4137–41. doi:10.1182/ ally distinct population of mature NK cells in the human blood-2012-12-470005. PMID:23493782. CD56dimCD16C NK-cell subset. Blood. 2010;116:3865–74. 25. Gribben JG, Fowler N, Morschhauser F. Mechanisms of Action of doi:10.1182/blood-2010-04-282301. PMID:20733159. Lenalidomide in B-Cell Non-Hodgkin Lymphoma. J Clin Oncol. 35. Pegram HJ, Andrews DM, Smyth MJ, Darcy PK, Kershaw MH. 2015;33:2803–11. doi:10.1200/JCO.2014.59.5363. PMID:26195701. Activating and inhibitory receptors of natural killer cells. 26. Kritharis A, Coyle M, Sharma J, Evens AM. Lenalidomide in non- Immunol Cell Biol. 2011;89:216 –24. doi:10.1038/icb.2010.78. Hodgkin lymphoma: biological perspectives and therapeutic opportu- PMID:20567250. nities. Blood. 2015;125:2471–6. doi:10.1182/blood-2014-11-567792. 36. Grzywacz B, Kataria N, Verneris MR. CD56(dim)CD16( C) NK cells PMID:25736312. downregulate CD16 following target cell induced activation of matrix 27. Fionda C, Abruzzese MP, Zingoni A, Cecere F, Vulpis E, Peruzzi G, metalloproteinases. Leukemia. 2007;21:356–9; author reply 9. Soriani A, Molfetta R, Paolini R, Ricciardi MR, et al. The IMiDs tar- doi:10.1038/sj.leu.2404499. PMID:17251901. gets IKZF-1/3 and IRF4 as novel negative regulators of NK cell-acti- 37. Lopez-Botet M, Vilches C, Redondo-Pachon D, Muntasell A, Pupu- vating ligands expression in multiple myeloma. Oncotarget. leku A, Yelamos J, Pascual J, Crespo M. Dual Role of Natural Killer 2015;6:23609–30. doi:10.18632/oncotarget.4603. PMID:26269456. Cells on Graft Rejection and Control of Cytomegalovirus Infection in 28. Bryceson YT, Chiang SC, Darmanin S, Fauriat C, Schlums H, Theorell Renal Transplantation. Front Immunol. 2017;8:166. doi:10.3389/ J, Wood SM. Molecular mechanisms of natural killer cell activation. J fimmu.2017.00166. PMID:28261220. Innate Immun. 2011;3:216 –26. doi:10.1159/000325265. PMID: 38. Stewart AK. Medicine. How thalidomide works against cancer. Sci- 21454962. ence. 2014;343:256–7. doi:10.1126/science.1249543. PMID:24436409. 29. Domaica CI, Fuertes MB, Uriarte I, Girart MV, Sardanons J, Comas 39. Vallet S, Palumbo A, Raje N, Boccadoro M, Anderson KC. DI, Di Giovanni D, Gaillard MI, Bezrodnik L, Zwirner NW. Human Thalidomide and lenalidomide: Mechanism-based potential drug natural killer cell maturation defect supports in vivo CD56(bright) to combinations. Leuk Lymphoma. 2008;49:1238 –45. doi:10.1080/ CD56(dim) lineage development. PLoS One. 2012;7:e51677. 10428190802005191. PMID:18452080. doi:10.1371/journal.pone.0051677. PMID:23240056. 40. Benson DM, Jr., Bakan CE, Mishra A, Hofmeister CC, Efebera Y, 30. Moretta L. Dissecting CD56 dim human NK cells. Blood. 2010;116:3689 – Becknell B, Baiocchi RA, Zhang J, Yu J, Smith MK, et al. The PD-1/ 91. doi:10.1182/blood-2010-09-303057. PMID:21071612. PD-L1 axis modulates the natural killer cell versus multiple myeloma 31. Freud AG, Yu J, Caligiuri MA. Human natural killer cell development effect: a therapeutic target for CT-011, a novel monoclonal anti-PD-1 in secondary lymphoid tissues. Semin Immunol. 2014;26:132 –7. antibody. Blood. 2010;116:2286 –94. doi:10.1182/blood-2010-02- doi:10.1016/j.smim.2014.02.008. PMID:24661538. 271874. PMID:20460501. 32. Montaldo E, Del Zotto G, Della Chiesa M, Mingari MC, Moretta A, 41. Gorgun G, Samur MK, Cowens KB, Paula S, Bianchi G, Anderson JE, De Maria A, Moretta L. Human NK cell receptors/markers: a tool to White RE, Singh A, Ohguchi H, Suzuki R, et al. Lenalidomide Enhan- analyze NK cell development, subsets and function. Cytometry Part ces Immune Checkpoint Blockade-Induced Immune Response in A: the journal of the International Society for Analytical Cytology. Multiple Myeloma. Clin Cancer Res. 2015;21:4607–18. doi:10.1158/ 2013;83:702–13. doi:10.1002/cyto.a.22302. PMID:23650273. 1078-0432.CCR-15-0200. PMID:25979485. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÖÇ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

!'".1#/!hZ!! !

0LM7Q!G8EGL>8J4HQ!LJ84LE8FL! ! I@:<@087>632+=3>@8?>J+ ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ÖÉ! 1GL4D!. +!!4F4DQK?K!GF!"+*!H4L?8FLK!7MJ?F=!6>8EGL>8J4HQ!?FL8JN8FL?GF! ! 2H! LMN=R! ,)! <>EE! IA>GHMRI>! BG! L+*! =BL>:L>f! P>! =! :G=! :G:ERL>=! ;EHH=! L:FIE>L!?KHF!É!I:MB>GML!qBG!PABGM!=:M:!P>K>!:G:ERL>=!;HMA!;>?HK>!:G=!:?M>K! MK>:MF>GM!:G=!|!I:MB>GM!PBA!m:?M>K!MK>:MF>GMn!P:L!NGj:G:ERL>=!=N>!MH!<>EE!EHLM!=NKBG@! MA>!?K>>S>oMA:P!LM>ILri!. +!!L:FIE>L!?KHF!I:MB>GML!PBMA!E>ND>FB:f!BG!I:KMB! MRIB<:EER! HO>KIHINE:M>=! ;R! BFF:MNK>! FR>EHB=! ;E:LML! MA>K>?HK>! :M>=! :! M>!?HK!GMBHG:E!@:MBG@!LMK:M>@R!?HK!,)!<>EE!qBi>i!!"~j!"ÅÇár!=N>!MH!HO>E:IIBG@! K>@BHGL! BG! :! ;Bj=BK>! :;NG=:GM! E>ND>FBEELi! 2H! HO>K! MABL! BLLN>f!P>!NEMBEBS>=!:!L>Kf!!"É!BGMH!MA>!:G:ERSBG@!IBI>EBG>i!'G! MABL! LMN=Rf! ,)! <>EE! IHINE:MBHG! P:L! =>?BG>=! :L! !"~j!"Éá!"ÅÇá! q $;9EB7! [PW ri! %>G>K:E! GML!. +!!P:L!=!:?M>K!~{!=:RL!qHK!FHK>r!IHLMjBG=NK! I>K<>GM:@>! H?! 2! <>EE! :G=! ,)! <>EE! qGHM! LB@GB?B<:GMr! :G=! =>:L>! BG! !"~~á!"|Äj! FR>EHB=!IHINE:MBHG!q $;9EB7![PX ri!2ABL!!BG!MHM:E!. +!!IKHIHKMBHG!BG!I:MB>GML!:K>! FHLM!EBD>ER!=N>!MH!E>ND>FBEEL!??>FHMA>K:IR!:@>GMLi! ! ,)!68DD!7?KHD4Q87!EGJ8!?EE4LMJ8!H>8FGLQH8!4!GHMBHG!MA:M!,)!<>EE!F:MNK:MBHG!LM:MNL!<:G!;>!>Q:FBG>=!OB:!L>O>K:E!P>EEj LMN=B>=!F:KD>KL!LNOBHNLER!K>IHKM>=!MH!=>L=!,)!<>EE! :M! =B??>K>GM! LM:@>L! H?! =>O>EHIF>GMf! MA>! GM! FH=>E! BG! ANF:G! ,)! <>E! LN@@>LML! MA:M! !"ÅÇ;K! @:BGBG@! F:MNK:MBHG! IA>GHMRI>! MH! ;>! !"ÅÇ=BF! OB:! ?HEEHPBG@! LM>ILh! !"ÅÇ ;K !"|Ç j! jä! !"ÅÇ =BF !"Ç}* á! jä!"ÅÇ =BF !"Ç}* j!"ÅÉ á! BFIERBG@! MA:M! @:BGBG@! !"ÅÉ! F:KDL!MA>!M>KFBG:E!=B??>K>GMBM:MBHGoF:MNK:MBHG!LM>I!BG!MA>!,)!<>EE!EBG>:@>!q(N>ED>!>M!:Ei! }{|{g!)KSRPBGLD:!>M!:Ei!}{|Çri! >:KBG@!MABL!FH=>Ef!,)!<>EE!BG!HNK!L+*!=!:G! BG:L>! BG! !"Ç}*! E>O>E! :G=! EHHLBG@! !"ÅÉ! >QIK>LLBHG! LBEFNEM:G>HNLER! :?M>K! MA>! MK>:MF>GM!q $;9EB7![PY!!Q"ri!2H@>MA>K!PBMA!MABLf!!"ÅÇ;K!IHINE:MHGL!q;HMA!!"ÅÇ;K!"|Çj! :G=! !"ÅÇ;K!"|Çár! BG:L>=! BG! ?K>JN>GK! MK>:MF>GM! BG=B<:MBG@! PAHE>! ,)! IHINE:MBHG! LD>P>=! MHP:K=! :! E>LL! F:MNK>! IA>GHMRI>! q $;9EB7! [P[! ri! 2ABL! I>KA:IL! LN@@>LML! :! MNKGHO>K! H?! @>G>K:E! ,)! <>EE! IHINE:MBHG! :?M>K! BGMKH=NFHMA>K:IR! :@>GMf!>BMA>K!=BK>=!DBEEBG@!:GO87!D8KK!?F!N?NG!6QLGLGP?6?LQ\4KKG6?4L87!E4JC8JK!4

! ÖÑ! ! 5>!G>QM!BGO>LMB@:M>!,)!<>EE!BG!OBOH!:E:M>=!F:KD>KL! MH!L>>!AHP!MK>:MF>GM!BGM>KO>GMBHG!F:R!:??>L>!F:D>KLi!LL!>QI>=f!MA>! )#("($ ! =>@K:GNE:MBHG! F:KD>K! !"|{É:! HG! ,)! <>EEL! P:L! LB@GB?B<:GMER! EHP>K! :?M>K! MK>:MF>GM! :L! H;L>KO>=!;>?HK>!BG!:GHMA>K!:L>! BL!:ELH!BG!:@K>>F>GM!PBMA!E>LL!,)!<>EEL!<:KKRBG@!FR>EHB=jLI>K!q>i@f!!"|Ä!:G=! !"~~rf! :G! >??>=B:M>=! ;R! MKH@H! ,)! <>EE! q $;9EB7! [PZ! !Q"ri! L?M>K! MK>:MF>GMf! ,)! <>EEL! @:BG>=! FHK>! ,)%}!! >QIK>LLBHGf! :G! :<>IMHK! MA:M! K>!&*Lj#!PABQIK>LL>=!HG!MK:GL?HKF>=!MNFHK!<>EEL!q $;9EB7! [PZ! ,ri!3G?HKMNG:M>ER!P>!=B=!GHM!DGHP!PA>MA>K!MABL!E:M>L!PBMA!O>E! HG! L+*! E>ND>FBEEL! :L! &*Lj#! :GMB;H=R! P:L! GHM! NL>=! BG! MA>! :G:ERLBG@! $L!1! I:G>E! q2:;>E! }ri! 'GM>K>LMBG@ERf! MA>! :QA:NLMBHG! F:KD>K! ."j|! HG! ,)! <>EEL! :ELH! =>:L>=! :?M>K! MK>:MF>GM! PABLML! ,)! <>EEL! :K>! BG! E>LL! :! :?M>K! =>:L>!BG!E>ND>FBK>?HK>!MABL!H;L>KO:MBHG!BGM>K@K:M>=!P>EE!PBMA!HMA>K! IA>GHMRIBL! q $;9EB7! [PY! , ri! 1NKIKBLBG@ER! :! K>=NO>Ef! :! F:KD>K! :LLH=!PBMA!:K!MK>:MF>GM!P:L!GHM!H;L>KO>=!BG!PAHE>!,)!<>EE!IHINE:MBHG! q$;9EB7! [PY! rf! MABL! E>=! NL! MH! ?NKMA>K! :G:ERL>! HG! ,)! LN;IHINE:MBHG! ;:L>=! HG! !"ÄÅ! BLH?HKFL!MH!;>MM>K!NG=>KLM:G=!AHP!>:L>!LN;L>ML!=!BG!!"ÇÖ!E>O>E!=NKBG@! MK>:MF>GMi! ! !"gh/"/-m!,)!68DD!6GJJ8D4L8K!O?L>!!"dg!EQ8DG?7!E4JC8J!4F7!!"il!46L?N4L?GF! E4JC8J! ! 5>!A:O>!IK>OBHNLER!MA>!=BLMBG!IA>GHMRI>!H?!MA>!!"ÄÅ0L0-á!LN;IHINE:MBHG! PABK>?HK>!BM!BL!H?!BGM>K>LM!MH!K>jO:EB=:M>! HNK!?BG=BG@L!BG!MABL!G>P!LMN=R!q !:3AD7B!Y ri!'G!@>G>K:Ef!MA>!,)!<>EE!!"ÄÅ!BLH?HKF!IKH?BE>L! =B=!GHM!=!=K:F:MB<:EER!:?M>K!I:MB>GMlL!MK>:MF>M!:EMAHN@A!:!LEB@AM!BG:L>!MK>G=!BG! !"ÄÅ0Lá!IHINE:MBHG!P:L!H;L>KO>=!q $;9EB7![P[!, ri!2H!?NKMA>K!>Q:FBG>!MA>!MKH@HGMB:E!MNFHKj:LLH=!F:KD>KL!H?!L+*!qBi>i!!"|Ä!:G=!!"~~f! PAB! FR>EHB=jLI>! F:=>! :EMBHG! :G:ERLBL! ;>MP>>G! Ü! H?! FR>EHB=! IHINE:MBHGL!qBGEE!BG! >:! :L! H;L>KO>=! ;>?HK>! :G=! :?M>K! MK>:MF>Mi! LL! LAHPG! BG! $;9EB7! [P\! , f! MA>! IK>L>G<>! H?! !"~~á!"|Äá! FR>EHB=! IHINE:MBHG! BL! LMKHG@ER! :M>=! PBMA! !"ÄÅ0L0-á! ,)! IHINE:MBHG! ;NM! GHM! !"~~á!"|Äj! FR>EHB=! IHINE:MBHGi! 5BMA! :! LBFBE:K! E:MBHG!

! ÖÖ! :G:ERLBL! ;>MP>>G! !"~~o!"|Ä! MKH@H=! ,)! <>EEL! PBMA! IKHIHKMBHGL! H?! !"ÄÅ0L0-á! ,)! LN;L>Mf! !"|Ä! ;NM! GHM! !"~~! BL! LMKHG@ER! K>E:M>=! PBMA! AB@A>K! !"ÄÅ0L0-á! ,)! I>K<>GM:@>!q $;9EB7![P\! ri!%BO>G!MA:M!!"|Ä!FHE>L!HGER!>QIK>LL>=!HG!A>F:MHIH>MBEEL!=!MH!MA>!FHGH!EBG>:@>f!MABL!L>>FL!MH!BG=B<:M>!!"ÄÅ0L0-á!,)! IK>?>K>GMB:EER!BGM>K:G<>f!MKH@HoF:!IHINE:MBHGLi!! ! ! 2H!;>MM>K!NG=>KLM:G=!AHP!!"ÄÅ0L0-á!,)!<>EEL!<:G!;>!NL>=!MH!IK>=BEE! @>G>K:E! IA>GHMRIBf! P>! MHHD! :=O:GM:@>! H?! MA>! ?:K! H?! I:MB>GML! qBi>! ~! H?! MHM:E! Ç! I:MB>GMLr! BG! HNK! L+*! =! :! =>:L>! BG! Ü! !"ÄÅ0L0-á! ,)! <>EEL! MH! =BOB=>! MHM:E! I:MB>GML! BGMH! }! @KHNILh! m0L0-j=HPGn! qBi>i! Ü!"ÄÅ0L0-á! ,)! =>:L>! :?M>K! MK>:MF>GMr! :G=! m0L0-jNIn! qBi>! Ü!"ÄÅ0L0-á! ,)! NG=oBG:L>=r! :G=! ?NKMA>K! BGO>LMB@:M>! MA>L>! }! @KHNIL! BG! =RG:FB! H?! IK>OBHNL! F:KD>KL! q $;9EB7! [P]! , ri! 'G! ! NG=! E>O>E! H?! !"ÇÖ! BG! MA>! MHM:E!,)!IHINE:MBHGf!!"ÇÖ!E>O>E!P:L!=>:L>=!BG!m0L0-j=HPGn!@KHNI!qGHM!LB@GB?B<:GMr! :G=!BG:L>!BG!m0L0-jNIn!@KHNI!q $;9EB7![P]!! ri!1BG<>!MA>!GNF;>K!H?!!"ÄÅ0L0-á!,)! MA:M! =! =NKBG@! MK>:MF>M! P:L! FNK! BG! L<:E>! MH! MHM:E! ,)! IHINE:MBHGf! MA>! =>:L>!H;L>KO>=!BG!!"ÇÖ!q:IIKHQBF:M>ER!}{Ü!BG!:O>K:@>r!<:GGHM!;>!:<=!?HK!MA>! GNF;>K! H?! !"ÄÅ0L0-á! ,)! <>EEL! :EHG>i! $NKMA>KFHK>f! MA>! =>:L>! BG! ."j|! E>O>E! H;L>KO>=!HG!MA>!m0L0-j=HPGn!@KHNI!P:L!FHK>!=K:F:MB!m0L0-jNIn!q $;9EB7! [P]! ri!&HP>O>Kf!MA>!IA>GHMRIBK>G<>!BG!,)!<>EEL!BG!m0L0-j=HPGn!O>KLNL!m0L0-j NIn!@KHNIL!:K>!GHM!=BLMBG@NBLA:;E>!?HK!MA>!F:CHKBMR!H?!HMA>K!F:KD>KL!q $;9EB7! [P]! "Q#f! CEAA=7>7?D3BI!$;9EB7![PWN ![PX ri!2:D>G!MH@>MA>Kf!MA>!=RG:FBL!BG!!"ÄÅ0L0-á! ,)!<>EE!<:G!;>!:!?:=BK!BG!MHM:E!,)!<>EE!IHINE:MBHG!BG! L+*!I:MB>GMLi!! ! ! ! ! ! ! ! ! !

! |{{!
  
   





   


   



! !

   



 



!

        !    OB:;E>!<>EE   ! !  
# "     ! LBG@E>M| LBG@E>M}  

           ! !               
  
   ! ! ! ! 

 



 ! 

 ! OB:;E>!<>EE 



"
# " 

!   !  

    !        ! ! !   
  !
   !"~j!"Éá !"~j!"Éj

 !
   ! ,)!<>EE !      

!


!"~ !  

!          &  &
   

   
 !   & #!"ÅÇ  !   &#  ! ' &$ !& "  $#!  ! ! $;9EB7! :46 "B:@K:F!H?!@:MBG@!LMK:M>@R!?HK!,)!<>EE!:G=!HMA>K!ERFIAHL!IHINE:MBHGL!BG! L+*i! 0>IK>L>GM:MBO>! IEHML! H?! :! I:MB>GMlL! . +!! q:?M>K! MK>:MF>GMr! LAHPBG@! MA>! @:MBG@!

LMK:M>@R! :IIEB>=! BG! MA>! LMN=Ri!
 'G! MABL! LMN=Rf! ,)! <>EE! IHINE:MBHG! =>?BG>=! :L! !"~j !"Éá!"ÅÇái #
 ! ! ! !
 ! ! ! ! ! ! ! ! ! ! !

! |{|! !

! T cell B cell NK cell ! 80 15 GL 15 GL ! u 60 ! 10 10

! 40 ! 5 5 % of total cell total of % % of total cell total of % ! 20 cell total of % ! 0 0 0

diagnostic diagnostic ! diagnostic

! 30d post-induction 30d post-induction 30d post-induction ! ! CD33+CD14- monocyte CD33+CD14+ monocyte u 40 ! 100 GL ! 80

! 60 20 ! 40 % of total cell total of % % of total cell total of % ! 20 ! 0 0 ! ! diagnostic diagnostic ! # 30d post-induction 30d post-induction

$;9EB7! [PX!!!.KHIHKMBHG!H?!ERFIAH!:G=!FHGH!IHINE:MBHGLi#.KHIHKMBHGL!H?!>:! :G=! FHGH! IHINE:MBHGL! IK>L>GM>=! BG! I>K<>GM:@>! BG! 2f! ! :G=! ,)! <>EE! GML! qNII>K! I:G>ELr! :G=! FHGHL! qEHP>K! I:G>ELr! H?! !"~~á! :G=! !"~~á!"|Äá!IA>GHRI>Li!1M:MBLMB<:E!LB@GB?B<:GM!P:L!=>M>KFBG>=!;R!I:BK>=!MjM>LM!;>MP>>G! =B:@GHLMBKLNL!:?M>K!MK>:MF>Mi!,àÉ#PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! |{}! ! ! ! !" !" ! PD1 CD69 ! 10 u 80 GL ! 8 60 ! 6 ! 40 ! 4 % of total NK total of % % of total NK total of % 20 2

! 0 0 !

diagnostic diagnostic

! 30d post-induction 30d post-induction ! ! !" !" ! CD62L CD57 60 100 ! GL u ! 80 ! 40 60 ! 40 ! 20 % of total NK total of % % of total NK total of % ! 20 ! 0 0 ! ! diagnostic diagnostic

! 30d post-induction 30d post-induction ! $;9EB7![PY!!,)!<>EE!:KLi!!Lr!.>K<>GM:@>!H?!."j|!>QIK>LBG@! ,)! <>EEi! r! !"ÇÖ! :K! HG! ,)! <>EE! IHINE:MBHG! =NKBG@! MK>:MF>GMi! !j"r! #QIK>LLBHG! H?! F:MNK:MBHG! F:KD>KL! H?! !"Ç}*! :G=! !"ÅÉ! HG! ,)! <>EE! IHIN:MBHG! =NKBG! MK>:MF>GMi! 1M:MBLMB<:E! LB@GB?B<:GM! P:L! =>M>KFBG>=! ;R! I:BK>=! MjM>LM! ;>MP>>G! =B:@GHLMBKLNL!:?M>K!MK>:MF>Mi!,àÉ#PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! ! ! ! ! ! ! ! ! ! ! ! ! !

! |{~! ! ! !" !" ! NKG2C CD107aCD (degranulation) ex vivo 40 15 ! GL u ! 30 ! 10 ! 20 5 % of total NK total of % ! 10 NK total of % ! 0 0 ! ! diagnostic diagnostic

30d post-induction 30d post-induction ! ! !" !" ! CD14-CD33+ NK CD14+CD33+ NK 6 ! 15 u GL ! ! 10 4 ! 5 2 % of total NK total of %

! NK total of % ! ! 0 0 !

! diagnostic diagnostic

! 30d post-induction 30d post-induction ! ! $;9EB7![PZ!!,)!<>EE!>Q!OBOH!=>@K:GNE:MBHG!:G=!MKH@HO>E!H?!,)%}!!:<>IMHK!HG!,)!<>EE!=NKBG@!MK>:MFGMii! r!,)!=>@K:GNE:MBHG!>Q!OBOH!B=>GMB?B>=!;R!!"|{É:i! !j"r! +>:LNK>! H?! ,)! <>EE! MKH@HEHB=! F:D>KL! ;>?HK>! :G=! :?M>K! MK>:MF>GMi! 1M:MBLMB<:E! LB@GB?B<:GM! P:L! =>M>KFBG>=! ;R! I:BK>=! MjM>LM! ;>MP>>G! =B:@GHLMBKLNL!:?M>K!MK>:MF>Mi!,àÉ#PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! ! ! ! ! ! ! ! ! ! ! ! ! !

! |{Ä! ! ! !"! ! CD45RA+ CD45RAR0+ CD45RAdim CD45RO+ 6 GL 4 GL ! 105 8 GL GL ! 100 6 3 ! 4 ! 95 4 2 2 % of total NK total of % % of total NK total of % % of total NK total of % ! 90 NK total of % 2 1

! 85 0 0 0 ! diagnostic diagnostic diagnostic diagnostic

30d post-induction 30d post-induction ! 30d post-induction 30d post-induction ! !

!"! ! CD56br CD16- CD56br CD16+ CD56dimCD16- CD56dimCD16+ ! 40 uu 25 GL 100 GL uu 60 ! 20 80 30 ! 15 40 60 ! 20 ! 10 40 % of total NK total of % % of total NK total of % % of total NK total of % 10 NK total of % 20 ! 5 20

! 0 0 0 0 !

! diagnostic diagnostic diagnostic diagnostic ! ! 30d post-induction 30d post-induction 30d post-induction 30d post-induction #

$;9EB7! [P[! ! ,)! LN;IHINE:MBHG! :G:ERLBLi! LG:ERLBL! H?! ,)! <>EE! LN;IHINE:MBHGL! ;:L>=! HG! !"ÄÅ!BLH?HKF!LN;L>ML!q!"ÄÅ0Láf!!"ÄÅ0L-áf!!"0L=BF!:G=!!"ÄÅ0-ár!qLr!:G=!E>O>E!H?! !"ÅÇ! :G=! !"|Ç! >QIK>LLBHG! MA:M! =BOB=>L! ,)! <>EEL! BG! Ä! LN;IHINE:MBHGLh! !"ÅÇ;K!"|Çjf! !"ÅÇ;K!"|Çá! f! !"ÅÇ=BF!"|Çj! ! :G=! !"ÅÇ=BF!"|Çá! q ri! 1M:MBLMB<:E! LB@GB?B<:GM! P:L! =>M>KFBG>=! ;R! I:BK>=! MjM>LM! ;>MP>>G! =B:@GHLMBKLNL! :?M>K! MK>:MF>Mi! ,àÉ! PBMA! u! I! ã! {i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! !

! ! ! ! ! ! ! !

! |{Å! ! !"! !"%&!&'&!('#!$' !!% $&%$ !%" #% ! 100 100 ! 80 80 ! !"!!#!" " j! 60 60 ! ! # 40 40 ! 20 20 % CD33+CD14-mono % % CD33+CD14-mono % 0 0 ! 0 5 10 15 0 1 2 3 4 ! % CD45RAR0+ NK % CD45RAR0+ NK

Kà!{iÑÖÖÄ 60 60 ! Iâ{i{{{|!uuuu

! 40 40 ! ""$! !#$' ! ! # 20 20

! CD14+CD33+mono % CD14+CD33+mono % 0 0 0 5 10 15 0 5 10 15 ! % CD45RAR0+ NK % CD45RAR0+ NK ! ! !"! ! !"%&!&'&!('#!$' !!% $&%$ !%" #%

! 15 8

! 6 ! 10 !"!!#!" " j! 4 ! 5 "#!"## 2 % CD33+CD14-NK % ! CD33+CD14-NK % 0 0 ! 0 5 10 15 0 1 2 3 ! % CD45RAR0+ NK % CD45RAR0+ NK ! Kà!{iÖ~~Å ! 15 1.5 Iâ{i{{{|!uuuu ! 10 1.0 !! ""$! !#$' 5 0.5 ! "#!"## % CD33+CD14+NK % ! CD33+CD14+NK % 0 0.0 ! 0 5 10 15 0 1 2 3 % CD45RAR0+ NK % CD45RAR0+ NK ! ! $;9EB7! [P\## Lr! IKHIHKMBHG! H?! !"ÄÅ0L0-á! ,)! <>EE! E:M>L! PBMA! !"~~á!"|Äá! FHGH! IHINE:MBHG! ;NM! GHM! !"~~á!"|Äj! FHGH! IHINE:MBHGi! r! IKHIHKMBHG! H?! !"ÄÅ0L0-á! ,)! <>EE! E:M>L! PBMA! !"~~á!"|Äá! MKH@H=! ,)! <>EEi! !HKK>E:MBHG! ;>MP>>G!IHINE:MBHGL!:G:ERS>=!;R!.>:KLHG!F>MAH=!PBMA!Kà!E:MBG!??BGM!:G=!I!à! LM:MBLMB<:E!LB@GB?B<:<>!H?!E:MBHGi!

! ! ! !

! |{Ç! ! !" !"Z[.,.,`!+( ! 6 ! patient 8 !"ÄÅ0L0-á!,)! patient 10 4 m=>:L>n ! patient 14 patient 2 !"ÄÅ0L0-á!,)! ! 2 patient 15 mBG:L>n % of total NK total of % patient 20 ! 0

! after treatment after treatment before treatment before treatment

! !" !" PD1 CD69 ! 10 GL 80 decrease decrease GL increase 8 increase u ! uu 60 6 ! 40 4 % of total NK total of % % of total NK total of % 20 ! 2 ! 0 0 ! after treatment after treatment after treatment after treatment before treatment before treatment before treatment before treatment !

! !" CD62L !" % CD57 NK cell 60 GL decrease 100 GL GL decrease ! increase GL 80 increase 40 ! 60 40 20 % of total NK total of % ! NK total of % 20 ! 0 0 ! after treatment after treatment after treatment after treatment before treatment before treatment ! before treatment before treatment $;9EB7! [P]! ! ! !"ÄÅ0L0-á! LMK:MB?B<:MBHGh! :G:ERLBL! H?! O:KBHNL! LNK?:<>! F:KD>KL! ;:L>=! HG! L! BG! !"ÄÅ0L0-á! ,)! <>EEL! I>K<>GM:@>! :?M>K! MK>:F>GMi! Lr! 2HM:E! I:MB>GML! =BOB=>=! BGMH! }! @KHNIL! ;:L>=! HG! !"ÄÅ0L0-á! K>LIHGL>! :?M>K! MK>:MF>GMi! j#r! !HFI:KBLHG! ?HK! O:KBHNL!F:KD>L!."j|!q rf!!"ÇÖ!q!rf!!"Ç}*!q"r!:G=!!"ÅÉ!q#r!;>MP>>G!}!@KHNIL!;:L>=!HG! !"ÄÅ0L0-á! LMK:MB?B<:MBHG! =>L=! BG! qLri! 1M:MBLMB<:E! LB@GB?B<:GM! P:L! =>M>KFBG>=! ;R! I:BK>=!MjM>LM!;>MP>>G!=B:@GHLMBKLNL!:?M>K!MK>:MF>Mi!,à~!PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|! :G=!uuu!I!ã!{i{{|i!

!

! |{É! NKG2C CD107a ex vivo ! 50 15 decrease decrease GL ! 40 GL increase increase GL 10 30 ! GL 20 5 % of total NK total of % % of total NK total of % ! 10 0 0 !

! after treatment after treatment after treatment after treatment before treatment before treatment before treatment before treatment

CD33+ NK cell CD14+CD33+ NK cell ! 20 6 GL decrease decrease increase 15 GL increase 4 {i{É~ 10 ! GL 2 % of total NK total of % ! NK total of % 5 0 0 !

after treatment after treatment after treatment after treatment ! before treatment before treatment before treatment before treatment ! "0--)&*&+/$.3# ('0.&# :46 ! !"ÄÅ0L0-á! LMK:MB?B<:MBHG! q=ri! LG:ERLBL! H?! LN?:<>! F:KD>KL!;:L>=!HG!!"ÄÅ0L0-á!LMK:MB?B<:MBHG!:L!=>L=!BG!$B@i!ÅiÉi!

! CD56brCD16- NK cell CD56brCD16+ NK cell 25 40 GL GL decrease decrease 20 increase increase 30 15 ! {i{Ç} u 20 10 % of total NK total of % ! NK total of % 10 5

! 0 0 !

after treatment after treatment after treatment after treatment ! before treatment before treatment before treatment before treatment

! CD56dimCD16- NK cell CD56dimCD16+ NK cell 120 GL decrease decrease ! 60 100 {i{Å increase GL increase ~ 80 ! 40 60

GL 40 ! 20 % of total NK total of % % of total NK total of % 20

! 0 0 !

after treatment after treatment after treatment after treatment ! before treatment before treatment before treatment before treatment "0--)&*&+/$.3# ('0.&# :47 ! !"ÄÅ0L0-á! LMK:MB?B<:MBHG! q=ri! LG:ERLBL! H?! LN?:<>! F:KD>KL!;:L>=!HG!!"ÄÅ0L0-á!LMK:MB?B<:MBHG!:L!=>L=!BG!$B@i!ÅiÉi!

! |{Ñ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! |{Ö! ! ! ! ! ! ! ! ! ! !

!'".1#/!iZ!! !

"?KL?F6L!H>8FGLQH?6!?EHJ?FLK!?F!,)!68DD!!'ME4F!(EEMFG78632+=3>@8?>J+ ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ||{! #P4E?F8!L>8!=8F8J4D!D4F7K64H8!?F!DQEH>G6QL8!HGHMD4L?GFK!?F!'(3m!H4L?8FLK! ! 2H!H;M:BG!NG=>KLM:G=BG@!:;HNM!&'4!BG?>??>EEf!P>!:G:ERL>=!;EHH=! L:FIE>L! ?KHF! &'4á! I:MB>GML! qBG! PABGML! :K>! :OBK>FBi! GH! L>KNF! OBKNL! I:KMBLf! :G=! ~! OBK>FBGML! Bi>i! OBKNL! IK>L>G<>! BG! ;EHH=! L>KNFr! :G=! ~! A>:EMAR! =HGHKL!:L!=!MA>!@:MBG@!LMK:M>@R!NL>=!BG!MABL!LMN=Ri! "N>! MH! EBFBM:MBHG! BG! ?ENHK>L<>GM! E! HIMBHGL! =! MH! HNK! =>LBK:;E>! BFFNGHIA>GHMRIBEf! !"|Ä! :G=! !"|Ö! P>K>! IHHE>=! :G:ERL>=! q 034=7! Y ri! &>G<>! BG! LN;L>JN>GM! . +!! :G:ERLBLf! ERFIAH! IHINE:MBHGL! :K>! =>?BG>=! :L! !"|Äj!"|Öj! qBi>i! PBMAHNM! ! <>EELri! .KHIHKMBHG:E! :G:ERLBL! HG! MA>! !"|Äj!"|Öj! ERFIAHL! BG=B<:M>=! :! =>:L>!BG!;HMA!?K>JN>GL!H?!,)2!:G=!,)!<>EE!IHINE:MBHGLf!LBFNEM:G>HNLER!!"~á!2!<>EE! I><>GM:@>! P:L! BG:L>! q $;9EB7! \PX r! =! MH! &"! LML! }! IHLLB;BEBM>Lh!,)!:G=!,)2!<>EE!=>:L>=!BG!<>EE!GNF;>K!BG!&'4á!I:MB>GML!E>:=BG@!MH!AB@A>K! IKHIHKMBHG!H?!!"~á!2!<>EELf!HK!,)!:G=!,)2!<>EE!GNF;>K!K>F:BG>=!LM>:=R!PABE>!2!<>EEL! IKHEB?>K:M>! FHK>! BG! BG?>=! I:MB>GMLi! 5>! IHLMNE:M>=! MA:M! MA>! ?HKF>K! L<>G:KBH! BL! FHK>! IARLBHEH@B<:E!K>E>O:GM!AHP>O>K!:;LHENM>!ERFIAH!L!:K>!G>>=>=!MH! O:EB=:M>!MABL!ARIHMA>LBLi!'G!@>G>K:Ef!MA>!=BLMKB;NMBHG!E:G=L<:I>!BG!! IHINE:MBHGL!BG!&'4jBG?>=!I:MB>GML!BL!=BLMBGER!=B??>K>GM!?KHF!A>:EMAR!BG=BOB=N:ELi! ! (EH4?J8E8FL!?F!L>8!46L?N4L?F=!J868HLGJ!,)&e"!?F!'(3m!H4L?8FL!,)!68DDK! ! 'G! MA>! QM! H?! &'4! OBK:E! BG?>:LM! HG>! OBK:E! IKHM>BGf! G:F>ER! + !# A:L! ;>>G!K>IHKM>=!MH!BGM>K?>K>!PBMA!,)%}"j,)%}"*!LB@G:EBG@!:QBL!;>MP>>G!BG?>=!<>EE!:G=! ,)!<>EEf!>FIA:LBSBG@!MA>!BFIHKM:G<>!H?!MABL!P>EE!DGHPG!mLMK>LL!BG=N<>=n!F>=B:M>=!OBK:E!:K:G<>!q!>K;HGB!>M!:Ei!}{{Éri!2H!:==K>LL!MA>!KHE>!H?!,)%}"!BG!HNK! LMN=Rf!P>!:G:ERS>=!MA>!E>O>E!H?!,)%}"!>QIK>LLBHG!HG!MA>!,)!<>EEL!?KHF!&'4á!I:MB>GML! :G=!&"!QI>=f!P>!A:O>!H;L>KO>=!;HMA!K>=NEEL! :G=!K>=NO>E!BG!,)%}"j>QIK>LLBG@!<>EEL!HG!&'4á!I:MB>GML!=!MH! A>:EMAR!K>G<>!BG!,)%}"!IKH?BE>!BG!I:MB>GM!,)!<>EE!q $;9EB7!\PY ri! 2ABL! BMA>K! F>:G! ,)! <>EEL! =HPGjFH=NE:M>=! ,)%}"! FHE>L! BG! :! EB@:G=j =>I>G=>GM! >G=HK! IHLLB;ER! =N>! MH! ,)%}"*j>QIK>LLBG@! BG?>=! <>EEL! HK! I>KA:IL! ,)%}"! E>O>E! =>:L>=! BG=BK>=B:M>=! BFFNG>! >L<:I>! F>

! |||! ! "?<<8J8FL!KM5HGHMD4L?GF!HJGMM>K! NG=>KLM:G=BG@! HG! ,)! <>EE! =RG:FBGLf! P>! =BOB=>=! MHM:E!,)!IHINE:MBHG!BGMH!>:ML!;:L>=!HG!MA>BK!IKH?BE>!H?!!"ÅÇ! O>KLNL! !"|Ç! HK! !"ÄÅ0L! O>KLNL! !"ÄÅ0-! q $;9EB7! \PZ ri! 'G! GHKF:E! I>HIE>f! F:CHKBMR! H?! I>KBIA>K:E!,)!<>EEL!LAHP!!"ÅÇ=BF!"|Çá!IA>GHMRI>!qyÖ{Ür!:G=!K>F:BGBG@!<>EEL!A:O>! !"ÅÇ;K!"|Çj! PABE:M>=! P>EE! PBMA! HNK! =:M:i! &HP>O>Kf! BG! &'4á! I:MB>GML! q;HMA! :OBK>FBFB! AB@AER! L! BG=B<:M>! :! M>G=>K! IKHIHKMBHGf! >LI>FBGML! MA>! !"ÅÇ=BF! OL! !"ÅÇ;K! >FL! MH! LD>P! MHP:K=! AB@A>K! BFF:MNK>! IA>GHMRI>! qBi>i! BG:L>! !"ÅÇ;K! I>K<>GM:@>r!q $;9EB7!\PZ!,ri!1BFBE:KERf!!"ÄÅ!IKH?BE>!HG!,)!<>EEL!?KHF!BG?>=!I:MB>GM!:K>! =BLMBGER!=B??>K>GM!?KHF!MA:M!H?!&"f!PBMA!:G!BG:L>!BG!,)!<>EE!A:OBG@!!"ÄÅ0L=BF! :G=!!"ÄÅ0-á!IA>GHMRI>!PA>K>!FHLM!H?!,)!<>EEL!?KHF!&"!:K>!!"ÄÅ0Lá!LIHG=BG@! MH!:!mK>LMBG@n!LM:M>!q $;9EB7!\PZ! Q!ri!! ! "GOF!EG7MD4L?GF!G!AB@AER!EE!LN;L>Mf!=HPGjFH=NE:MBHG!H?!!"|Ç! :?M>K!>G@:@>F>GM!MH!M:K@>M!<>EEL!A:L!;>>G!IK>OBHNLER!K>IHKM>=!q%KSRP:KG>KBL! }{{Ég! *:CHB>! >M! :Ei! }{|Äri! +HK>HO>Kf! !"|Ç! =HPGjFH=NE:MBHG! :EL! H<=! BG! :LLH=! PBMA! BG! OBMKH! K! ?HK! =>@K:GNE:MBHG! qNGIN;EBLA>=! H;L>KO:MBHGri! 5BMA! MABL! GHMBHGf! P>! :MM>FIM>=! MH! BGM>KIK>M! MA>!!BG!!"ÄÅ!IKH?BE>L!BG!I:MB>GM!,)!<>EEL!MH!EBGD!;>MP>>G!=B??>K>GM!!"ÄÅ!LN;L>ML! PBMA!!"|Ç!E>O>Ei!'GM>K>LMBG@ERf!BG!;HMA!@KHNIL!H?!I:MB>GML!P>!H;L>KO>=!:!EBGD!;>MP>>G! !"ÄÅ0L=BF!IENL!!"ÄÅ0-á!,)!qPABL!GHM!>QBLM!BG!&"r!:G=!:!@K:=N:E!=>:L>!BG! !"|Ç! E>O>E! q $;9EB7! \P[! ,Q ri! ,HMB<>:;ERf! MA>! m@K:=>n! H?! !"|Ç! =HPGjFH=NE:MBHG! BL! IKH@K>LLBO>! :L! ,)! <>EE! LN;L>ML! @:BGBG@! !"ÄÅ0-! >QIK>LLBHG! PABE>! EHHLBG@! !"ÄÅ0L! T$;9EB7! \P[! ri! 2ABL! E>=! NL! MH! IKHIHL>! :! ! L>JN>G<>! H?! ,)! <>EE! :G:KBH! "##("($ !:L!?HEEHPh!!"ÄÅ0Lá!jä!!"ÄÅ0L=BF!jä!!"ÄÅ0-á!BG!PABEEL! PBMA! AB@A! E>O>E! H?! !"|Ç! >QIK>LLBHG! LAHNE=! ;>! K>@:K=>=! :L! mG:BO>n! <>EELf! :L! MA>R! >GK! MA>! BG?>=! <>EEL! "## ("($ ! MA>L>! >O>GML! IKHFHM>=! :!H?!!"ÄÅ0Lj!"ÄÅ0-á!:G=!!"|Ç!EHLLi!&>G<>!BG!MABL!LBMN:MBHG!!"ÄÅ0L=BF!,)!

! ||}! <>EE! <:G! ;>! K>@:K=>=! :L! :G! mBGM>KF>=B:M>n! LM:M>! ;>MP>>G! mG:BO>n! :G=! m:=n! <>EELf! PAB>G!IKHIHL>=!;>?HK>!BG!I:MB>GML!PBMA!A>F:MHEH@B<:E!<:G<>KL!q)KSRPBGLD:!>M! :Ei!}{|Çri!! ! 1J4FK<8JJ?F=!G4F?KE!GOBHNL! K>LNEML! IKHO=BG@! >OB=>GML! MA:M! ,)! <>EEL! ?KHF! &'4! BG?>=! I:MB>MGL! :LLH=!FHK>!PBMA!:=!LM:M>! "##("($ f!IKH;:;ER!=N>!MH!BG?>EEL!>G@:@>F>GMi! 2ABL!E>=!NL!MH!BGO>LMB@:M>!MA>!HNM!H?!MABL!<>EEj<>EE!BGM>K:MP>>G!,)!<>EE!:G=! OBKNL!BG?>=!<>EEi!,)!<>EEL!A:O>!;>>G!LAHPG!MH!I>K?HKF!MKH@HJN>GM! H?! MABL! ,)! <>EE! <:G! <:KKR! I:ML! H?! M:K@>M! <>E! F>F;K:G>! BGF;K:G>j:G=!K><>IMHK!FHE>L!q+BG>K!>M!:Ei!}{|Åg!,:D:FNK:!>M!:Ei! }{|~ri!2H!>Q:FBG>!B?!,)!<>EE!<:G!I>K?HKF!MKH@HEE!qPABQM! F>:GL! M:K@>M! <>EErf! P>! =! IKBF:KR! BLHE:M>=! ,)! <>EEL! q?KHF! :! A>F:MHEH@B<:E!<:G<>K!I:MB>GMr!PBMA!2!<>EE!EBG>L!!#+!:G=!+2Ä!MA:M!:K>!IK>OBHNLER!E:;>E>=! PBMA!:!OBHE>M!=R>!q!24ri!LL!LAHPG!BG! $;9EB7!\P\!, f!:?M>K!EE!M:K@>MLf!OBHE>M!=R>!:LLH=!PBMA!M:K@>M!<>EEL!<:G!;>!=>M>=!?KHF!,)!<>EEL!PBMA! LHF>! IKHIHKMBHG! H?! ,)! <>EEL! @:BG>=! :L! AB@A! :FHNGM! H?! =R>! :L! IKBF:KR! LM:BG>=! M:K@>M! <>EELi! 5>! G>QM! :LD>=! B?! MABL! F>F;K:G>jMK:GL?>KKBG@! IA>GHF>G:! ;>MP>>G! ,)o2! <>EEL! E>O:GM!JN>GMi!'M!P:L!DGHPG!MA:M!&'4!OBKNL!I:KMB! >QIEHBML! MA>! 2! <>EE! F>F;K:G>! K><>IMHK! !"Ä! MH@>MA>K! PBMA! FHDBG>! K><>IMHK! !6!0Äo!!0Å!?HK!OBKNL!>GMRi!2ABL!E>=!NL!MH!ARIHMA>LBS>!B?!MA>L>!&'4jK><>IMHKL!<:G!:ELH!;>! MK:GL?>K>=!?KHF!OBKNL!BG?>=!<>EE!BGMH!,)!<>EE!OB:!MKH@HK!MABL!JN>LMBHGf! P>!K>I>:M>=!MA>! "##("'&$ !MKH@HMP>>G!IKBF:KR!BLHE:M>=!,)!<>EEL!MH!=R>j E:;>E>=! !#+o+2Ä! :G=! LBFNEM:G>HNLER! NLBG@! :GMBj!"Ä! :G=! :GMBj!!0Å! :GMB;H=B>L! MH! =>M>r!MK:GL?>KKBG@!H?!MA>L>!FHE>L!?KHF!MA>L>!M:K@>ML!<>EEL!MH!,)!<>ELi!LL! LAHPG! BG! $;9EB7! \P\! Q!f! !#+! :G=! +2Ä! ;HMA! >QIK>LL>=! !"Ä! qI:KMB:EER! ?HK! !#+r! AHP>O>K! HGER! +2Ä! ;NM! GHM! !#+! >QIK>LL! MA>! FHDBG>! K><>IMHK! !!0Åi! 'G! EEL! BLHE:M>=! ?KHF! }! =B??>K>GM! I:MB>GML! PBMA! A>F:MHEH@B<:E! =BL>:L>L! =!PBMA!E:;>E>=!!#+o+2Ä!@:BG>=!LN;LM:GM:E!E>O>E!H?!!24!=R>!:L!P>EE!:L!!"Ä!q;HMA! PBMA! !#+! :G=! +2Är! :G=! !!0Å! qHGER! PBMA! +2Äri! 2A>K>?HK>f! BM! P:L! BG=B<:M>=! MA:Mf! OB:! MKH@HEEL! <:G! @:BG! MK:GLB>GM! >QIK>LLBHG! H?! OBKNLjK><>IMHKL! ;R! BGM>K:

! ||~! PBMA!!"Äo!!0Åj>QIK>LLBG@!M:K@>M!<>EELi!2ABL!;>@L!MA>!G>QM!BGM>K>LMBG@!JN>LMBHGh!=H>L!,)! <>EE!;>!&'4!LNL<>IMB;E>!AHLM!;R!@:BGBG@!!"Äo!!0Å!OB:!MKH@HOBHNL! BG?>=!<>EELk!2H!HNK!DGHPE>=@>!BM!A:L!GHM!R>M!;>>G!=>LKB;>=!;>?HK>i! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ||Ä!
    
     

    
    !
     

     ! ! #$% $"     !    ) * +  LBG@E>M| LBG@E>M}

  ## 

% #&%

$ % " 

$1j& $ 11! 

$1!j2-$ " #        $      $     # %  #  ) * +  $1!jL $1jL         ( # ,- ,  $"  ( .#j!RÅiÅ!p! $" !"|Äo!"|Ö  

# !"|Ä j!"|Ö j
       !  ' 2!ERFIAH $"

  #   .#j!RÅiÅ!p! !"|Äo!"|Ö  ,)2 ."       ( L.!   # !"ÅÇj.#j!RÉ,- , '" $ !"~!j!L.!  ,) 

 $         ' # !"ÅÇ!p!.#j!RÉ " 

('0.&# ;46# #%:MBG@! LMK:M>@R! ?HK! ,)! <>EE! :G=! HMA>K! ERFIAHL! IHINE:MBHGL! BG! A>:EMAR! =HGHKL!:G=!&'4á!I:MB>GMLi!0>IK>L>GM:MBO>!IEHML!H?!:G!:OBK>FBGMlL!. +!!LAHPBG@!

 MA>!@:MBG@!LMK:M>@R!:IIEB>=!BG!MA>!LMN=Ri"! "! #

! ! ! ! ! ! ! ! ! ! "! ! ! ! ! ! !

! ||Å! lymphocyte

100 u GL HD 80 aviremic viremic 60 GL 40 {i{ÉÅ GL {i{ÑÄ 20

0

T cell NK cell NKT cell

! $;9EB7! \PX! ! %>G>K:E! IKHIHKMBHG! H?! ERFIAH! IHINE:MBHGL! BG! &"! :G=! &'4á! I:MB>GMLi! .>K<>GM:@>!H?!2f!,)2!:G=!,)!<>EE!IHINE:MBHGL!K>LI>ER!HG!BG?>=!I:MB>GML!q:OBK>FB:! :G=! OBK>FB:r! :G=! HG! A>:EMAR! =HGHKLi! &"h! Gà~f! LOBK>FBFBM>KFBG>=!;R!HG>jP:R!L,-4L ;>MP>>G!&" O>KLNL!I:MB>GML PBMA!u!I!ã {i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! !

! ! ! ! NKG2D NKG2D+ NK HD 110 5 aviremic u GL viremic {i{ÇÄ u 100 4 MFI 90 3 % of total NK total % of

80 2

HD HD

viremic viremic aviremic aviremic ! ! $;9EB7! \PY! ! !0>=N<>=!,)%}"!>QIK>LLBHG!BG!,)!<>EEL!BG!BG?>=!I:MB>GML!F>:LNK>=!;R! I><>GM:@>! H?! ,)%}"á! ,)! <>EEL! qE>?M! I:G>Er! :G=! >QIK>LLBHG! E>O>E! H?! ,)%}"á! ,)! <>EEL! qKB@AM! I:G>Eri! &"h! Gà~f! LOBK>FBFBM>KFBG>=!;R!HG>jP:R!L,-4L!;>MP>>G!&"!O>KLNL!I:MB>GML!PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|! :G=!uuu!I!ã!{i{{|i!

! ||Ç! !" CD56/CD16 subpopulations GL 120 GL HD 100 aviremic 80 viremic GL 60 u GL 40 GL
    
   
    

% of total NK total of %
     
          !  20

0

CD56br

CD56dimCD16+ CD56dimCD16-

!" %!)$#&('& *&"$&&

K>IK>L>GM:MBO>! K>IK>L>GM:MBO>! K>IK>L>GM:MBO>!&" :OBK>FBGM OBK>FBGM $ '( , ''( ,#  +,%
$#!+"+,%
( # $ ,1('!%&,1*  %  )  

    

      '( ,# +,%
( ,1* !"ÄÅ0L   

  

           %         )         '(!"ÄÅ0- , ' +,%
$#!+ ,1('!%

!" CD45RA/RO {i{Å 110 {i{Ö}{i{ Ö}Ñ 100 HD 90 aviremic GL 80 viremic GL 70 u 30 GL {i{Ñ  20 GL  "  % of total NK total of % 10 0

CD45RA+ CD45RO+ CD45RARO+ CD45RAdim

! $;9EB7!\PZ!!,)!LN;IHINE:MBHG!:G:ERLBL!HG!&"!:G=!&'4á!I:MB>GMLi!Lr!LG:ERLBL!H?!,)!<>EE! LN;IHINE:MBHGL!;:L>=!HG!E>O>E!H?!!"ÅÇ!:G=!!"|Ç!>QIK>LLBHG!MA:M!=BOB=>L!,)!<>EEL!BG!~! LN;IHINE:MBHGLh! !"ÅÇ;Kf! !"ÅÇ=BF!"|Çj! ! :G=! !"ÅÇ=BF!"|Çái! r! 0>IK>L>GM:MBO>! =HMj IEHL! LAHPBG@! !"ÄÅ0L! O>KLNL! !"ÄÅ0-! >QIK>LLBHG! IKH?BE>L! ;>MP>>G! &"! :G=! &'4á!

! ||É! I:MB>GMLi!!r!1NFF:KBS>=!;:Kj@K:IA!LAHPBG@!I>K<>GM:@>!H?!,)!<>EE!LN;L>ML!=BOB=>=!BGMH!Ä! IHINE:MBHGL! ;:L>=! HG! !"ÄÅ0L! O>KLNL! !"ÄÅ0-! >QIK>LLBHGi! &"h! Gà~f! LOBK>FBFBM>KFBG>=! ;R! HG>jP:R! L,-4L! ;>MP>>G! &"! O>KLNL!I:MB>GML!PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! # ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ||Ñ!
    
     

   
     !  ! !

!" !

MHM:E!,) !"ÄÅ0Lá!,)
     !"ÄÅ0L=BF! !"ÄÅ0-á!,) !
     
$ '( $ '( ' '% &' #&' & ,)  %  $ !

  !  
    
   $ #$  % "&'&!' * !       !         !  &' & '( ' !" !   !  

 
  
      ! ! !    $ #$  (& "       % #+$  '% "    #$ '% "  #+$ '% "  + * * *  '( $  &' #  

        $ #$ % "&'&!' * !         !%"$!#" ( ! "  "  "  " 

    ""# %!#

   

 
   
       !         !   !     + # $  , (&,"-   #+*$  + '%+",      * #+$  + '%+",      * #+*$  + '%+",      *      (  '  '  ' !

  !   !      !"ÅÇ     $ #$ % "&'&!' * +", +", ,"- +", $!# "!&' !         ""# %!# !    

            !                     , (!"|Ç + ' + ' + ' ! ! "! "!* ! ! !" aviremic!$!# "!' n=4

120 $ ! ! CD56br ! GL CD56dimCD16+ 90 GL ! ! CD56dimCD16- "  "  60 ! ! 30 ! "!* "!* 0 ! !

total NK CD45RA+ CD45RO+ ! CD45RAdim ! ! viremic n=3 ! 120 !!# "!& "  "  CD56br ! u CD56dimCD16+ 90 CD56dimCD16- ! GL uuu 60 uu ! ! 30 !

0 !

total NK CD45RA+ CD45RO+ ! CD45RAdim ! ! $;9EB7!\P[!! !0>=N<>=!!"|Ç!>QIK>LLBHG!BG!!"ÄÅ0L=BF!:G=!!"ÄÅ0-á!,)!LN;IHINE:MBHGL! BG! &'4á! I:MB>GMLi! Lr! K>IK>L>GM:MBO>! =HMjIEHML! HG! &"! :G=! BG?>=! I:MB>GML! LAHPBG@!

! ||Ö! !"ÅÇo!"|Ç! LN;L>ML! MP>>G! MHM:E! ,)! <>EELf! !"ÄÅ0Láf! !"ÄÅ0L=BF! :G=! !"ÄÅ0-á! ,)! <>EELi! r! /N:GMBM:MBO>! ;:Kj@K:IA! LNFF:KBS>=! !"ÅÇo!"|Ç! ! =B??>K>GM! !"ÄÅ! ,)! <>EE! LN;>ML! :L! =>FHGLMK:M>=! BG! Li! LOBK>FBFBM>KFBG>=! ;R! HG>jP:R! L,-4L! ;>MP>>G! !"ÄÅ0Lá! O>KLNL!!"ÄÅ0L=BF!:G=!!"ÄÅ0-áf!PBMA!u!I!ã!{i{Åf!uu!I!ã!{i{|!:G=!uuu!I!ã!{i{{|i! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! |}{!
     

!

!"
      ! IÅÑÇ!á!+2Ä!p @:M>=!HG!,) IÅÑÇ!á!!#+!p!@:M>=!HG!,) ! IÅÑÇ!:EHG>!p!@:M>=!HG!,) +2Ä!:EHG> ! !#+!:EHG> ! ! ! ! !

! ! <>EEMK:<>!OBHE>M

   
    
      !      !"   !
   !#+!:EHG>
   +2Ä!:EHG>!!!!!           

! !"  

  !  

    ! ))(), ) ,

   
!+ ! $  +

    !   )       (       ! IÅÇ}!q,)r!. +!!
%   $%# IÅÇ}!á!!#+ IÅÇ}!á!+2Ä <>EEMK:<>jOBHE>M :EHG>

! &  



!

! ) + + - +-

! $* '#, $,

!  

 &     ! $% # ('# '("& ! !"Ä!p!.# ! ! !"
   
   ! !#+!:EHG>      

      +2Ä!:EHG>                !  !        

!     $' !  #$ "#       &    % ! !       

  ! EMK IÅÑÇ!q,)r!. +!!                  

!
 " <>EEMK:<>jOBHE>M <>E :EHG>
 
    
     !             !"# IÅÑÇ!á!!#+

  IÅÑÇ!á!+2Ä!"            %  $ !  

! &) &) % (   !      (  $( " '

!    

 !       %  $   !!0Å!p!.#j!RÅiÅ

#
 ! ! ('0.&# ;4;# # 2K:GL?>K! H?! &'4! K><>IMHK! !"Ä! :G=! !!0Å! ?KHF! M:K@>M! <>EE! MH! ,)! <>EE! ;R! MKH@HK?HKF>=! ;R! E>=! !"Äá!M:K@>M!<>EEL!q!#+!:G=!+2Är!MH@>MA>!PBMA!I:MB>GMlL!. +!f!,)!<>EE!MKH@HM!<>EEL!:K>!F>:LNK>=!;R!MK:GL?>KBG@!H?!!24!;R!$L!1!:G:ELRBL!:?MK!ÄA!H?!

!  " 

! |}|!

   #$ !" j!r! 'G! OBMKH! MKH@HK?HKF>=! :L! =>L=! :;HO>! PBMA! !"Ä! q ! I:G>Er! :G=! !!0Å!q!!I:G>Er!&'4!<>IMHK!LM:BGBG@!LAHPBG@!MK:GL?>KBG@!H?!MA>L>!K><>IMHKL!HGMH!,)! <>EE!LNK?:<>!NIHG!KBH=i!

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

!

!

! !

! |}}! ! ! ! ! ! ! ! ! !

!'".1#/!jZ!! !

#PH4FK?GF!G! 8<632+/=?7183+G+,63=/:;>?71>J+ + + K&:+ ?67>+ >?@2D+ ?6/?+ &+ >6/=32+ 1;G/@?6;=>67+ 7:A;8A32+ 7:+ <8/::7:5+ /:2+ 3C31@?7;:+;4+?63+3C<3=793:?>E+1;8831?7:5+/:2+7>;8/?7:5++>/9<83+) '!+ /:2+-! +2;:;=>E+/:2+<3=4;=97:5+/:/8D>7>+;4+=3>@8?>+7:18@232+7:+>@<<8393:?/=D+ 9/?3=7/8>L+ ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! |}~! Theranostics 2018, Vol. 8, Issue 14 3856

Ivyspring

International Publisher Theranostics 2018; 8(14): 3856-3869. doi: 10.7150/thno.25149 Research Paper Expansion of allogeneic NK cells with efficient antibody-dependent cell cytotoxicity against multiple tumors

Diego Sanchez-Martinez 1,#, Nerea Allende-Vega 1,2,# , Stefania Orecchioni 3, Giovanna Talarico 3, Amelie Cornillon 1, Dang-Nghiem Vo 1, Celine Rene 1, Zhao-Yang Lu 1, Ewelina Krzywinska 1, Alberto Anel 4, Eva M. Galvez 5, Julian Pardo 4, Bruno Robert 6, Pierre Martineau 6, Yosr Hicheri 7, Francesco Bertolini 3, Guillaume Cartron 7 and Martin Villalba 1,2,

1. IRMB, Univ Montpellier, INSERM, CHU Montpellier, Montpellier, France. 2. IRMB, CHU Montpellier, France. 3. Laboratory of Hematology-Oncology, European Institute of Oncology, Milan, Italy. 4. University of Zaragoza/Institute of Health Research of Aragón (IIS-Aragón), Spain. 5. Instituto de Carboquimica, Consejo Superior de Investigaciones Científicas (CSIC). 6. IRCM, INSERM U1194, Univ Montpellier, France. 7. Département d'Hématologie Clinique, CHU Montpellier, Univ Montpellier, France.

# These two authors have equally contributed to this manuscript

Corresponding author: [email protected] (MV)

© Ivyspring International Publisher. This is an open access article distributed under the terms of the Creative Commons Attribution (CC BY-NC) license (https://creativecommons.org/licenses/by-nc/4.0/). See http://ivyspring.com/terms for full terms and conditions.

Received: 2018.01.25; Accepted: 2018.04.26; Published: 2018.06.14

Abstract Monoclonal antibodies (mAbs) have significantly improved the treatment of certain cancers. However, in general mAbs alone have limited therapeutic activity. One of their main mechanisms of action is to induce antibody-dependent cell-mediated cytotoxicity (ADCC), which is mediated by natural killer (NK) cells. Unfortunately, most cancer patients have severe immune dysfunctions affecting NK activity. This can be circumvented by the injection of allogeneic, expanded NK cells, which is safe. Nevertheless, despite their strong cytolytic potential against different tumors, clinical results have been poor. Methods: We combined allogeneic NK cells and mAbs to improve cancer treatment. We generated expanded NK cells (e-NK) with strong in vitro and in vivo ADCC responses against different tumors and using different therapeutic mAbs, namely rituximab, obinutuzumab, daratumumab, cetuximab and trastuzumab. Results: Remarkably, e-NK cells can be stored frozen and, after thawing, armed with mAbs. They mediate ADCC through degranulation-dependent and -independent mechanisms. Furthermore, they overcome certain anti-apoptotic mechanisms found in leukemic cells. Conclusion: We have established a new protocol for activation/expansion of NK cells with high ADCC activity. The use of mAbs in combination with e-NK cells could potentially improve cancer treatment.

Key words: NK cells, monoclonal antibodies (mAbs), antibody-dependent cell cytotoxicity (ADCC), cancer

Introduction Recent progress in cancer treatment is primarily that are mainly expressed by the tumor cell related to the development of novel targeted therapies population and/or playing a critical role in neoplastic [1]. These require the identification of suitable targets cell growth. Therapeutic monoclonal antibodies

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3857

(mAbs) particularly illustrate this concept. Indeed, possess antitumor NK cells that are unable to control rituximab (RTX), an IgG1 mAb directed against CD20 disease [10, 11]. Notably, blood-borne tumor cells use antigen, has now become the treatment of choice for different mechanisms for immune escape [12, 13], e.g., most B-cell chronic lymphocytic leukemias (B-CLL) by inducing NK cell dysfunction [7, 14]. This and B-cell non-Hodgkin’s lymphomas (B-NHL). The mechanism has also been observed in a variety of combination of RTX with conventional chemotherapy patients of solid tumors [3]. In addition, NK cell has shown better efficacy in randomized clinical trials. differentiation may be inhibited by the presence of Similar success has been found with other cytotoxic tumor cells, e.g., acute myeloid leukemia (AML) cells mAbs, such as trastuzumab in breast cancer or infiltrating bone marrow [15, 16]. Therefore, the cetuximab in colorectal carcinoma and squamous cell failure of mAbs in monotherapy could be related to carcinoma of the head and neck [2, 3]. Nevertheless, impaired NK cell function. Hence, there is a clinical mAbs alone generally have modest clinical activity. interest to reactivate or replace patient NK cells [17]. For example, the anti-CD20 mAbs RTX and Clinical-grade production of allogeneic NK cells is obinutuzumab (OBZ; previously GA101, Roche, efficient and NK cell-mediated therapy after Genentech), when used as monotherapy in patients hematopoietic stem cell transplantation (HSCT) seems with relapsed follicular lymphoma (FL), have only led safe [16, 18, 19]. Despite the strong cytolytic potential to short progression-free survival (PFS) [4]. Thus, of expanded NK cells against different tumors, clinical there is a need to optimize their use in combination results have been very limited [16, 18, 19]. therapy. The combination of allogeneic NK cells with RTX success is related to its capacity to induce mAb could improve cancer treatment by replacing the Fc-antibody-dependent cell-mediated cytotoxicity defective effector immune cells. In addition, mAbs $'&&  2QH UHFHSWRU IRU KXPDQ ,J* LV )FDŽ5,,,D would effectively guide these effectors to their tumor (CD16a), which is expressed on natural killer (NK) targets. Several groups have tried this combination cells and macrophages. The link between with varying results that could be due to deficient )FDŽ5,,,D -158VF polymorphism and RTX clinical CD16 expression or lack of proper activation of responses strongly suggests that ADCC is critical [5]. expanded NK [20-23]. In addition, these studies did This polymorphism is located on the extra-cellular not include a systematic evaluation of the effect of GRPDLQRI)FDŽ5,,,DDQGDPLQR -acid 158 is involved in these cells in combination with several mAbs on the interaction with CH2 of human IgG1 [4]. Human different tumors, nor did they include primary tumor IgG1 has a higher affinity for VV-NK cells compared cells. to FF-NK cells [5]. Based on these observations, there The aim of this work was to generate allogeneic has been an attempt to produce new anti-CD20 mAbs NK cells with strong ADCC response against different by either Fc mutations or by glycoengeenering that tumors and mediated by different therapeutic mAbs. H[KLELWKLJKHUDIILQLW\IRU)FDŽ5,,,D [4, 6]. Lowering the In addition, NK cell production should be easily fucose content of the N-glycan is currently under scaled up and developed with good manufacturing clinical investigation in B-cell malignancies with the practices (GMP). We have produced umbilical cord mAb OBZ, which shows stronger ADCC in vitro and blood (UCB)-derived NK cells because UCB are in a lymphoma xenograft mouse model relative to rapidly available, present low risk of viral RTX. It also demonstrated improved clinical activity transmission and have less strict requirements for for treating B-CLL and other B-cell malignancies [4]. HLA matching and lower risk of graft-versus-host OBZ is approved for first-line B-CLL in association disease (GvHD) [18]. For NK cell expansion we used with chlorambucil, and in combination with Epstein–Barr virus (EBV)-transformed bendamustine for the treatment of patients with FL lymphoblastoid B cell lines as accessory cells, which who relapse or are refractory to a RTX-containing induce a unique genetic reprogramming of NK cells regimen [4]. Initial results show that lenalidomide, [24]. This generates effectors that overcome the which stimulates NK cell activity [7], activates NK anti-apoptotic mechanism of leukemic cells [25] and cells in OBZ-treated patients[8]. that are able to eliminate tumor cells from patients NK cells mediate ADCC but also possess natural with poor prognosis [26]. We show that NK cells cytotoxicity, which is mediated by engagement of obtained with our protocol are able to perform ADCC their natural cytotoxicity receptors (NCRs). These in vitro and in vivo . The ADCC response was induced play a central role in triggering NK activation. In by using different therapeutic antibodies and against humans, NKp30, NKp46, and NKp80 are multiple target cells. constitutively expressed on resting and activated NK cells [9]. The NK cell-activating receptor CD16 mediates ADCC. Hematological cancer patients

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3858

Methods adenocarcinoma), SK-OV-3 (ovarian carcinoma) and SKBR3 (breast adenocarcinoma) were cultured in 10% Ethics statement FBS DMEM medium. Experimental procedures were conducted according to the European guidelines for animal PBMC and UCBMC purification welfare (2010/63/EU). Protocols were approved by UCBs were obtained from healthy donors from the Animal Care and Use Committee “Languedoc- the CHU Montpellier. Prof. John de Vos is the Roussillon” (approval number: CEEA-LR-12163). The responsible of the “Collection du Centre de Resources use of human specimens for scientific purposes was Biologiques du CHU de Montpellier” – approved by the French National Ethics Committee. http://www.chu-montpellier.fr/fr/plateformes All methods were carried out in accordance with the (Identifiant BIOBANQUES - BB-0033-00031). PBMC approved guidelines and regulations of this and UCB mononuclear cells (UCBMC) were committee. Written informed consent was obtained respectively collected from peripheral blood samples from each patient prior to surgery. and UCB units using Histopaque -1077 (Sigma). Briefly, 13 mL Histopaque was added to 50 mL Chemicals centrifugation tubes and 30 mL of 1/2 diluted blood The D1D2 peptide has been previously described in RPMI, (Invitrogen) was slowly added at the top. [27]. IL-2 and IL-15 were obtained from Miltenyi Tubes were centrifuged at 400 rcf for 30 min at 20 °C Biotec. To produce deglycosylated cetuximab, a without brake. Mononuclear cells were collected from commercial cetuximab solution was treated overnight the interlayer white ring, washed in RPMI and with PNGAseF (Promega) at 37 °C in 50 mM sodium suspended in RPMI medium supplemented with 10% bicarbonate buffer (pH 7.8) at 125 U/mg of FBS (Invitrogen). cetuximab. Deglycosilated cetuximab was purified by Isolation and activation of human NK cells gel filtration using a Sephadex 75 column in PBS, and sterilized by filtration. The AF 647 Goat F(AB')2 anti Frozen UCBMCs were depleted of T cells by human IgG (H+L) min x (BOV, HRS, MS) was from using EasySep TM CD3 Positive Selection Kit Interchim. (STEMCELL technologies). Cells were cultured for 10 WR  GD\V ZLWK DŽ -irradiated PLH cells at 1:1 NK B-CLL patients cell:accessory cell ratio in the presence of IL-2 (100 Data and samples from patients were collected at U/mL) and IL-15 (5 ng/mL), or with IL-2 alone (1000 the Clinical Hematology Department of the CHU U/mL). PLH cells were added every four days and Montpellier, France, after patients’ written consent fresh cytokines every two days. At the end of the and following French regulations. Patients were process, NK cell purity (CD56+/CD3-) was always enrolled in two independent clinical programs higher than 90%. approved by the “Comités de Protection des FACS analysis Personnes Sud Méditerranée I”: ref 1324 and HEMODIAG_2020 (ID-RCB: 2011-A00924-37). For phenotype analysis, cells were stained with Samples were collected at diagnosis and kept by the 7AAD (Beckman) to identify viable cells and CHU Montpellier [11, 28]. For analysis, peripheral antibodies against the surface markers CD25-FITC, blood mononuclear cells (PBMCs) were obtained by CD45RO-FITC, CD69-PE, CD62L-PE, CD19-PE, ficoll gradient and stored frozen in liquid nitrogen CD3-PE, CD19-ECD, CD56-PECy7, CD56-APC, until use. The percentage of CD19 +CD5 + cells was CD3-APC, CD45-APCAlexaFluor750, CD45RA-APC- always higher than 90%. Other samples from AlexaFluor750, CD16-PacificBlue, CD57-PacificBlue, hematological cancer patients were obtained from the CD45-KromeOrange, CD16-KromeOrange (Beck- same collections. man), CD158b-FITC, CD158a-PE, CD107a-HV500 (BD Biosciences), CD158e-Vioblue (Miltenyi). 1×10 5–3×10 5 Cell lines cells were incubated for 20-30 min at 4 °C with The (EBV)-transformed lymphoblastoid B cell different antibodies in PBS containing 2.5% FBS. Cells line PLH, the hematopoietic cell lines HL60 and were then washed and suspended in 200-250 µL of the MV4-11 (acute myeloid leukemia), Daudi and Raji same media. Staining was analyzed on a Gallios flow (Burkitt's lymphoma derived, CD20 +), K562 cytometer (Beckman) using the Kaluza software. (erythroleukemia), MM.1S and U266 (multiple Alive lymphocytes were gated using FSC/SSC and myeloma), MEC1 (B-chronic lymphocytic leukemia) 7AAD staining. B lymphocytes (CD19+), T and its variants MEC-1-BCL-XL and MEC-1-MCL-1 lymphocytes (CD3+CD56-) and NK cells were cultured in 10% FBS RPMI medium. The (CD56+CD3-) were distinguished using, respectively, adherent cell lines Calu-1 (lung cancer), A549 (lung CD19, CD3 and CD56 antibodies.

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3859

NK cell-mediated cytotoxicity cytometer (Millipore) with the count and viability kit Fresh or frozen (stored in liquid nitrogen) NK (Millipore). As a control, NK cells were incubated + cells were labeled with 3 µM of CellTracker™ Violet without targets. CD107a NK cells were analyzed on a BMQC Dye (Life Technologies) and incubated Gallios flow cytometer (Beckman Coulter) using overnight with target cells at different E:T ratios. 7AAD, CD45RO-FITC, CD19-PE, CD56-PECy7, Subsequently, phosphatidylserine (PS) translocation CD3-APC, CD45RA-APCAlexaFluor750, CD16-Kro- and membrane damage were analyzed in the violet meOrange and CD107a-HV500 (BD Biosciences). fluorescence-negative target cell population by flow Results were analyzed using Kaluza software. cytometry using Annexin V-FITC (Immunostep) and In vivo experiments 7AAD (BD Biosciencies) or propidium iodide (PI) as In vivo experiments were carried out using previously described [25, 29]. We consider all cells ï -week-old male NOD scid gamma (NSG) mice. positive for annexin-V and/or PI (or 7-ADD) as dead Mice were bred and housed in pathogen-free (or dying). conditions in the animal facility of the European In ADCC experiments, we incubated target cells Institute of Oncology–Italian Foundation for Cancer with the relevant antibodies (RTX and OBZ at 10 Research (FIRC), Institute of Molecular Oncology µg/mL; daratumumab, cetixumab and trastuzumab (Milan, Italy). For engraftment of human cells, mice at 5 µg/mL) for 30 min at 37 °C. To arm NK cells, we were subcutaneously engrafted with 5×10 6 BCL-P2 or incubated them at the same concentration of 10×10 6 LNH1 primary tumor cells derived from a antibodies before washing and incubation with target B-cell lymphoma (BCL) patient (BCL P2) or a diffuse cells. EGTA was used at 1 mM to block the granular large B-cell lymphoma (DLBCL) patient (LNH1). At exocytosis pathway and MgCl 2 at 1.5 mM to maintain day 4, we engrafted 15 (BCL-P2) or 10 (LNH1) million the osmotic pressure. e-NK cells and at day 6, mice were treated i.p. with We used the tetrazolium dye MTT RTX (in saline medium) 3 mg/kg once a week for 3 (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazomiu weeks; or with a combination of both treatments e-NK m bromide) to determinate cellular viability. We and RTX. Tumor growth was monitored at least once DGGHGǍ/RI077 PJP/ WRWKHDGKHUHQWFHOOV a week using a digital caliper, and tumor volume was  Ǎ/ RI PHGLXP DIWHU  ZDVKHV ZLWK 3%6  DQG calculated according to the formula: L × W 2/2 (mm 3), LQFXEDWHGIRUKDWƒ&WKHQDGGHGǍ/RI where W represents the width and L the length of the M HCl in isopropanol to dissolve the crystals and tumor mass. quantified absorption at 550 nm using a spectrophotometer. Statistical analysis In all experiments, we calculated the basal cell Experimental figures and statistical analysis death in the absence and presence of the different were performed using GraphPad Prism (v6.0). All mAbs. These values were subtracted from those statistical values are presented as * p<0.05; ** p<0.01; obtained after NK cell or NK cell+mAb treatments to *** p<0.001 and **** p<0.0001. Mean values are generate the specific natural cytotoxicity or specific expressed as mean plus or minus the standard error of ADCC, respectively. All mAbs gave very low levels the mean (SEM). (<3%) of cytotoxicity in presence of heat-inactivated serum media. Results Evaluation of RTX-armed e-NK Costimulation with the EBV lymphoblastoid e-NK cells (2×10 5) were incubated for 1 h with 10 PLH cell line more efficiently expands UCB µg/mL RTX, washed and incubated with 1:800 NK cells for clinical use than IL-2 stimulation solution of a goat F(ab')2 anti-human IgG (H+L) for 30 Cytokines and encounter with target cells induce min at 4 °C. After incubation, NK cells were washed dissimilar gene expression on NK cells [24]. We used with PBS and RTX binding was analyzed by FACs. As umbilical cord blood (UCB) cells and compared two a control, cells were only stained with the goat F(ab')2 NK cell activation/expansion protocols: one using a anti-human IgG. high dose of the cytokine IL-2 (1000 U/mL) and the NK degranulation assay other using cell costimulation. The costimulation protocol was performed with the EBV cell line PLH Briefly, 50×10 3 target cells per well were placed together with low concentrations of IL-2 (100 U/mL) in RPMI, 10% FBS, IL-2 100 U/mL with monensin (BD and IL-15 (5 ng/mL) [9]. NK cell expansion is Biosciences) in a 96-well V-bottom plate. NK and jeopardized by T cells, therefore we depleted them target cells were incubated overnight at 37°C in 5% from UCB before expansion. NK cell cultures CO and living cells were counted using a Muse 2 underwent massive cell death at day 12 in IL-2-driven

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3860 expansion (data not shown). So, we compared Frozen/thawed NK cells keep their cytolytic different parameters that reflect NK activity at day 10: activity Proliferation. Costimulation-driven expansion For clinical purposes, it would be advantageous was considerably more efficient ( Figure 1A ). to have a bank of cryopreserved expanded NK cells Cytotoxicity. IL-2-driven expansion led to NK ready to use [34]. Compared with fresh expanded NK cells with superior natural cytotoxicity against K562 cells, frozen/thawed NK cells lost roughly 35% of + (Figure 1B ), Daudi ( Figure S1A ) and primary CD20 CD16 expression and 50% of their cytolytic activity B cell lymphoma cells (BCL P2; Figure 1C ). Moreover, (Figure 1E ). As shown in Figure 1D , 20 day-activation these NK cells also showed higher ADCC activity showed higher cytolytic activity than shorter with RTX ( Figure 1C and Figure S1A ). However, expansions. For the next experiments, we used 20-21 natural cytotoxicity gradually increased in the days costimulation-induced expansion of costimulation protocol when cells were activated for UCB-derived NK cells containing more than 90% of longer periods of time ( Figure 1D ). This correlated NK cells that were kept frozen until use. Hereafter, we with a notable large-scale expansion of cells (median ± call them e-NK. SD, IL-2 10 d (16.8 ± 22.2), costimulation 10 d (140.5 ± 235.8) and costimulation 20 d (260.9 ± 141.2), n=10). e-NK cells mediate ADCC against target cells Activation markers. Both protocols increased the with diverse CD20 levels expression of the activation marker CD69 ( Figure We observed that e-NK performed ADCC S2B ) and decreased CD45RA expression to that of similarly on Raji and Daudi cells, which express high CD45RA dim cells ( Figure 2B ). This was associated with CD20 levels ( Figure S2A ), as on primary tumor cells, an increase in the activation marker CD45RO, as which express low levels ( Figure S2B ). Even though previously published [10]. Costimulation maintained P2 cells probably express more CD20 than P148, they higher CD16 expression ( Figure S2B ). were slightly less sensitive to RTX-mediated ADCC. Exhaustion markers. We investigated the In fact, e-NK performed ADCC even if their natural expression of two markers that could suppress NK cytotoxicity against some patient cells was low or cell-mediated cytotoxicity: TIM-3 [30] and PD-1 [31]. absent ( Figure 2A and Figure S2B-E ). Hence, e-NK While both protocols did not affect their expression, show strong ADCC with RTX independently of their the mean fluorescence intensity (MFI) of positive natural cytotoxicity and with lower variability ( Figure populations tended to increase ( Figure S2B ). This 2A-B ). The glycoengineered mAb OBZ [4, 35] induced probably reflects that, after long activation, some NK higher ADCC than RTX ( Figure 2C and Figure cells become exhausted. S2B-C ). Maturation and homing markers. UCB NK cells showed a low percentage of CD62L + cells (18.1% ± e-NK cells can be “armed” with mAbs to 6.7%, n=3) that increased 10 d after IL-2 treatment facilitate treatment (68.2% ± 13.5%) and costimulation (56.7% ± 11.9%). We next used e-NK cells coated with anti-CD20 However, at day 20 post-costimulation, CD62L mAb (“armed” e-NK cells) as an alternative to expression was lost (1.6% ± 0.4). IL-2-stimulated cells antibody-coated target cells. “CD20-armed” e-NK did not survive this long; so, we could not measure showed similar results to opsonizing tumor cells with CD62L levels. CD62L is a "homing receptor" anti-CD20 ( Figure S3A ). The presence of RTX after facilitating naive lymphocytes to enter secondary e-NK “arming” was visualized by using a fluorescent lymphoid tissues. Mature NK cells express low anti-IgG ( Figure S3B ). “OBZ-armed” e-NK also CD62L, which favors their peripheral trafficking [32]. efficiently generated ADCC ( Figure S4 ). In agreement with others [33], few naïve UCB Statistical analysis of several e-NK productions NK cells expressed CD57 (1.2% ± 1.3%, n=3). IL-2 on cells from 5 CLL patients did not show any stimulation barely increased expression (7.3% ± 3.9%) differences between opsonizing targets or “arming” and costimulation did not change it (0.6% ± 0.6). e-NK ( Figure 3A). Moreover, the analysis of these 4 Longer costimulation, i.e., 20 days, also had no effect e-NK expansions on the CLL targets showed that all (1.6% ± 0.5%). The lack of CD57 expression did not productions could be armed ( Figure 3B ). Combining impair NK cell cytolytic activity (see below). all these results statistically showed that significant In summary, costimulation led to higher ADCC was mediated by e-NK ( Figure 3C ). numbers of activated and functional NK cells with Cytotoxicity requires degranulation and cell higher CD16 expression. This prompted us to only use costimulation for the next experiments. On the other interaction by ICAM hand, IL-2 induced higher and faster cytotoxicity and Primary human NK cell cytotoxicity is largely could be the best option for autologous NK cell grafts. independent of degranulation [36] and resides in

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3861 death receptor binding on tumor cells by ligands the degranulation inhibitor EGTA ( Figure 4A ), which expressed by NK cells. In agreement with this, e-NK in contrast, largely blocked ADCC. natural cytotoxicity was only partially diminished by

Figure 1. Optimization of NK cell expansion protocol. (A) Comparison of costimulation (PLH accessory cells + IL-2 100 U/mL + IL-15 5 ng/mL) and IL-2 (1000 U/mL) expansion protocols using three UCB donors (2903, 3464, 2928). (B) K562 cells were incubated overnight with costimulation- (circles) or IL-2-activated (triangles) NK cells from two different donors at different effector: target (E:T) ratios. Cell death was analyzed by 7-AAD staining. (C) BCL Patient 2 cells were incubated overnight with costimulation- or IL-2-activated NK cells from three different donors, in the presence (black symbols) or absence (white symbols) of RTX (10 µg/mL). (D) NK cells from 2 donors were expanded by costimulation for different days. At these days, NK cells were frozen. They were thawed at the same time and tested for cytotoxicity against the cell lines. (E) NK cells from 2 donors were expanded for 20 days and frozen or kept fresh before testing their cytotoxicity.

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3862

Figure 2. e-NK mediate ADCC with anti-CD20 mAbs. (A) PBMCs from CLL patient 148 were incubated for 1 h with 10 µg/mL of RTX and overnight with e-NK cells obtained from nine different donors at a 1:1 E:T ratio. Cell death was analyzed by 7AAD staining. The right panel shows the statistical analysis. (B) 7 e-NKs were tested against 13 CD20+ target samples at different E:T ratios as described in (A). (C) 5 e-NKs were tested against 6 CD20+ target samples at different 1:1 E:T ratios as described in (A). Graphics represent mean ± SEM. Significance was determined by paired t-WHVW Sģ SģDQG Sģ

The interaction of NK cell-expressed LFA-1 with A549 (data not shown). Both lines are targets of the its target cell ligand ICAM modulates NK cell anti-EGFR cetuximab, which has been proposed for cytotoxicity [37]. Blocking this interaction with the use with adoptively-transferred expanded allogeneic D1D2 peptide [27] partly reduced natural cytotoxicity NK cells in clinical trials for cervical cancer [39]. In and almost completely abolished ADCC ( Figure 4B ). fact, in vitro and clinical data suggest that cetuximab Therefore, our e-NK use similar mechanisms for mediates ADCC through NK cells [23, 34]. We eliminating target cells as primary human cells. observed a relatively large variation in the natural cytotoxicity of the different e-NK donors versus these e-NK mediate ADCC with daratumumab target cells. The decrease in cell viability, as measured Next we tested if e-NK produced ADCC with the by MTT formation, was low ( Figure S6A ). However, anti-CD38 daratumumab [38]. We used 3 target cells the increase in cell death, measured by annexin-V / that express CD38 (MM.1S, MV4-11 and BCL-P2; 7-ADD staining, was higher. This showed that e-NK Figure S5A ) and observed that three different e-NK had induced the initial steps of apoptosis (annexin-V preparations showed ADCC with daratumumab staining), but longer times were required to evaluate (Figure S5B-D ). In contrast, daratumumab failed to cell viability with MTT. Cetuximab increased early induce ADCC against U266 that are negative for CD38 apoptosis and accelerated the process of cell death, (Figure S5A, E ). Several e-NK productions efficiently decreasing viability. Several e-NK productions performed ADCC with daratumumab on MM.1S and efficiently performed ADCC with cetuximab on P2 that was statistically significant ( Figure 4C ). CALU-1 and A549 that was statistically significant e-NK mediate ADCC with cetuximab (Figure 5A ). EGTA diminished ADCC but insignificantly Next, we analyzed if e-NK cells mediate ADCC (Figure S6B-C ). This suggested that the mechanism of with other mAbs used to treat solid tumors. We used action only partly involved degranulation, indicating the cell lines Calu-1 and A549. Calu-1 cells express a possible participation of death ligand-induced more epidermal growth factor receptor (EGFR) than apoptosis.

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3863

Figure 3. Anti-CD20-armed e-NK show ADCC activity. PBMCs from CLL samples were incubated for 1 h with 10 µg/mL of RTX and overnight with donor e-NK cells at a 3:1 E:T ratio (antibody-coated target cells condition). Alternatively, e-NK cells were incubated for 1 h with 10 µg/mL of RTX before incubating them overnight with target cells (antibody-armed NK cell condition). (A) The bars represent cell death of each individual CLL sample to 4/5 e-NK preparations. (B) The bars represent cell death of each individual e-NK preparation to 4/5 CLL samples. B-CLL cell death was analyzed by 7-AAD. Graphics represent mean ± SEM. Significance was determined by paired t-WHVW Sģ  Sģ SģDQG Sģ

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3864

Figure 4. ADCC requires degranulation and LFA-1/ICAM interaction. (A-B) Daudi cells were incubated overnight with e-NK cells from two different donors and/or RTX (10 µg/mL) as described in Figure 2 . Cytotoxic assays were performed also in the presence of 1 mM EGTA (A) or 15 µg/mL D1D2 protein (B). Cell death was analyzed by 7-AAD staining. (C) e-NK produced ADCC with daratumumab. Three different e-NK cell productions were tested against the CD38 + cell line MM.1S and BCL-P2 cells that express CD38. Target cells were pre-incubated for 1 h with 5 µg/mL daratumumab before overnight incubation at different E:T ratios with e-NK. Cell death was analyzed by 7-AAD staining. Graphics represent mean ± SEM. Significance was determined by paired t-t HVW Sģ SģDQG Sģ

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3865

Figure 5. e- NK perform ADCC with cetuximab and trastuzumab against EGFR- and HER2-positive cell lines, respectively. (A-B) Tumor cells were incubated with 5 µg/mL cetuximab (A) or trastuzumab (B) for 1 h and overnight with 4 e-NK preparations at 3:1 E:T ratio. Subsequently, we measured cell viability (MTT) and cell death (annexin-V/PI). (C) 1×10 4 SK-OV-3 cells were plated and 24 h later treated with trastuzumab and cultured with 5×10 5 NK cells. After 6 days, medium was removed, and cells were fixed and stained. The right graph shows the statistical analysis comparing cells incubated with e-NK alone or with trastuzumab. Two areas of 2 different experiments were counted and the mean of cells/area is depicted in the graphic. Graphics represent mean ± SEM. Significance was determined by paired t-WHVW Sģ SģDQG Sģ 0.001.

e-NK mediate ADCC with trastuzumab decreasing viability or increasing apoptosis ( Figure S7A ). Natural cytotoxicity, as well as ADCC, heavily We next tested the anti-HER2 mAb trastuzumab depended on degranulation because EGTA largely on SK-BR-3 cells, which express high HER2, and A549 decreased both ( Figure S7A-B ). Statistical analysis of cells, which express low levels. Under this condition, several e-NK productions on SK-BR-3 and A549 e-NK performed ADCC in both cell lines by

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3866 showed that cell viability was significantly reduced surface receptors for several cytokines, including IL-2 and apoptosis tended to increase, although this was and IL-15. Therefore, the signaling pathways for these not statistically significant. This suggested that e-NK cytokines are blocked in ,OUDŽ knockout mice. They efficiently performed ADCC with trastuzumab should lack functional NK cells, which require IL15 (Figure 5B ). However, the increase in apoptosis was signaling to develop. Four days later, mice were not statistically significant. Finally, we extended this engrafted with e-NK and, 2 days later, treated with study to the ovarian cell line SK-OV-3 that was RTX; the latter decreased tumor growth ( Figure 6A ), resistant to both natural cytotoxicity and ADCC showing that RTX possesses direct, non-effector during short treatment (data not shown). A 6-day functions independently of NK cell-mediated treatment revealed that NK cells, mainly with effects. While e-NK also decreased tumor growth trastuzumab, efficiently killed these cells ( Figure 5C ). (Figure 6A ), co-treatment was more effective, protecting all mice from BCL cells and 4 out of 5 mice e-NK mediated ADCC in vivo from DLBCL cells. We next evaluated e-NK activity in vivo by engrafting primary tumor cells from a B-cell e-NK cells showed ADCC against lymphoma (BCL) patient (P2) or from a diffuse large chemoresistant cells B-cell lymphoma (DLBCL) patient (LNH1) into NSG EBV-activated NK cells overcome anti-apoptotic mice. NSG mice have a complete null mutation by mechanisms active in chemoresistant cells [25, 26]. NQRFNRXWRIWKHDŽFKDLQRIWKHLQWHUOHXNLQUHFHSWRU Overexpression of BCL-X L and MCL1 are common (Il2r DŽ), which is a common component of the cell features of several hematological cancers [40]. Jurkat

Figure 6. e-NK show ADCC in vivo and overcome mechanisms of drug resistance. (A) 5 NSG mice/group were subcutaneously engrafted with 5×10 6 BCLP2 (left) or 10×10 6 LNH1 (right) cells and treated with e-NK and/or RTX. (B) e-NK cell-induced ADCC overcome anti-apoptotic mechanisms of drug resistance. CD20 + MEC-1 cells overexpressing BCL-X L and MCL1 were incubated with RTX (10 µg/mL). After 1 h, e-NK cells from 3 different donors were added overnight. Cell death was analyzed by 7AAD/annexin-V labeling. Graphics represent mean ± SEM. Significance was determined by paired t-WHVW Sģ SģDQG Sģ

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3867 cells over-expressing BCL-X L are resistant to Second, NK are the first lymphocytes to recover after doxorubicin and to soluble TRAIL [41]. MCL1 over- HSCT including after umbilical cord blood expression protects them from ibrutinib cytotoxicity transplantation (UCBT). The speed of recovery [42]. e-NK killed the CD20 + B-CLL cell line MEC-1 correlates with the prognosis [18]. In spite of these overexpressing BCL-X L or MCL1, neither of which findings, NK cell adoptive immunotherapy has conferred protection against ADCC compared to wild provided clinical benefit. Perhaps current expansion type cells ( Figure S8 ). Several e-NK productions kill protocols fail to produce enough NK cells to support chemoresistant cell lines to statistically significant clinical success or generate cells with impaired levels ( Figure 6B ). However, BCL-X L overexpression activity [16]. An inconvenience of engraftment of significantly decreases ADCC at high E:T ratios. This allogeneic expanded NK cell is their low survival in suggests that chemoresistance could partially protect vivo [49]. The persistence of ex vivo haploidentical tumor cells from e-NK-mediated ADCC, but not from IL-2-activated and -expanded NK-DLIs reaches a e-NK natural cytotoxicity. maximum of 7 days in lymphoma patients [50]. This leaves grafted NK cells little time to eliminate their Discussion targets. The advantage is that NK will be less likely to Clinical mAbs fail to improve prognosis in a generate the clinical problems found with CAR-T large number of patients. This could be due to the cells, which produce some chronic effects related to impairment of NK cells observed in most cancer their long-term persistence ([51]; http://www patients [3, 16, 18]. Therefore, restoring this immune .medscape.com/viewarticle/876591). One of the main function should improve mAb clinical benefits. We concerns in using allogeneic immune cells is the focused on developing a protocol to obtain NK cells in incidence of GVHD. Allogeneic NK cells infusion is sufficient number and with high ADCC activity well tolerated in cancer patients [3, 18] and the together with different therapeutic mAbs. From severity of aGVHD correlates with impaired UCB-derived NK, we produced e-NK that only reconstitution of the NK cell compartment [52]. To our partially lost ADCC function after cryopreservation knowledge, engraftment of NK cells has been linked (Figure 1 ) and preserved ADCC in vivo ( Figure 6 ). to GVHD only when combined to HLA-matched, e-NK do not require relatively high Ag levels to T-cell-depleted nonmyeloablative peripheral blood perform ADCC, since they were effective with stem cell transplantation [53]. NK cells likely different cell lines expressing variable Ag levels. contributed to GVHD in this setting by augmenting The coupling of mAbs and e-NK should underlying T-cell alloreactivity [53]. synergize in several clinical contexts. First, e-NKs An interesting alternative to allogeneic NK is should bypass NK dysfunction by increasing KIR/KIRL blocking antibodies that activate mAb-induced ADCC in patients with immune endogenous NK cells [54]. This approach has the defects. Second, the clinical activity of NK cells is inconvenience that cancer patient NK cells are uncertain in solid cancers [16, 18]. Probably these hyporeactive [3, 16, 18], suggesting that they are too effectors scarcely recognize solid tumor targets in vivo inefficient to totally eliminate tumors. Moreover, and/or fail to infiltrate these tumors—e.g., NK have recent clinical data suggest that such antibodies been detected in the tumor stroma but not within the modify the endogenous NK repertoire. This would tumor lesion in some cases [43-45]. Moreover, the make KIR-expressing NK cells, which are those with adoptive transfer of autologous NK cells in patients as higher cytolytic activity, hyporeactive [55]. This raises single therapy maintained high levels of circulating concerns about the clinical use of these antibodies. NK cells but did not mediate tumor regression [3, 18, There are other ways to activate endogenous NK cells, 46]. mAbs should recruit e-NK to the selected targets such as the use of lenalidomide (LEN) [7, 8]. and can also facilitate target elimination by favoring Preliminary results from the Phase Ib/II clinical trial the recognition of opsonized cells. In fact, GALEN suggest that LEN could facilitate haploidentical NK cells combined with anti-GD2 mAb OBZ-mediated NK cell activation [8], as was observed therapy has shown promising antitumor activity in with RTX [56]. pediatric recurrent/refractory neuroblastoma [47, 48]. In collaboration with the University Hospital of Third, e-NKs overcome anti-apoptotic mechanisms Montpellier, we wish to test the clinical efficiency of active in leukemic cells ( Figure 6B and [25]), allowing e-NK in lymphoma patients resistant to standard elimination of tumor cells from patients with poor treatments, including RTX. prognosis [26]. e-NK could also have anti-tumor activity per se . Abbreviations First, high numbers of tumor-infiltrating NK ADCC: antibody-dependent cell-mediated correlates with a better prognosis in some tumors [16]. cytotoxicity; AML: acute myeloid leukemia; B-CLL:

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3868

B-cell chronic lymphocytic leukemia; B-NHL: B-cell 4. Cartron G, Watier H. Obinutuzumab: what is there to learn from clinical trials? Blood. 2017; 130: 581-9. non-Hodgkin’s lymphoma; BCL: B-cell lymphoma; 5. Cartron G, Dacheux L, Salles G, Solal-Celigny P, Bardos P, Colombat P, et al. Therapeutic activity of humanized anti-CD20 monoclonal antibody and DLBCL: Diffuse large B-cell lymphoma; EBV: polymorphism in IgG Fc receptor FcgammaRIIIa gene. Blood. 2002; 99: 754-8. Epstein–Barr virus; EGFR: epidermal growth factor 6. Jefferis R. Glycosylation as a strategy to improve antibody-based therapeutics. Nat Rev Drug Discov. 2009; 8: 226-34. receptor; FL: follicular lymphoma; GMP: good 7. Giuliani M, Janji B, Berchem G. Activation of NK cells and disruption of manufacturing practices; GvHD: graft-versus-host PD-L1/PD-1 axis: two different ways for lenalidomide to block myeloma progression. Oncotarget. 2017; 8: 24031-44. disease; HSCT: hematopoietic stem cell transplan- 8. Vo DN, Alexia C, Allende-Vega N, Morschhauser F, Houot R, Menard C, et al. tation; LEN: lenalidomide; NCRs: natural cytotoxicity NK cell activation and recovery of NK cell subsets in lymphoma patients after obinutuzumab and lenalidomide treatment. Oncoimmunology. 2018; 7: receptors; NK cells: natural killer cells; OBZ: e1409322. obinutuzumab; PFS: progression-free survival; RTX: 9. Anel A, Aguilo JI, Catalan E, Garaude J, Rathore MG, Pardo J, et al. Protein Kinase C-theta (PKC-theta) in Natural Killer Cell Function and Anti-Tumor rituximab; UCB: umbilical cord blood; UCBT: Immunity. Frontiers in immunology. 2012; 3: 187. umbilical cord blood transplantation; e-NK: expanded 10. Krzywinska E, Allende-Vega N, Cornillon A, Vo D, Cayrefourcq L, Panabieres C, et al. Identification of anti tumor cells carrying natural killer (NK) cell NK cells; mAbs: monoclonal antibodies. antigens in patients with hematological cancers. EBioMedicine. 2015; 2: 1364-76. 11. Krzywinska E, Cornillon A, Allende-Vega N, Vo DN, Rene C, Lu ZY, et al. Supplementary Material CD45 Isoform Profile Identifies Natural Killer (NK) Subsets with Differential Activity. PLoS One. 2016; 11: e0150434. Supplementary figures. 12. Villalba M, Rathore MG, Lopez-Royuela N, Krzywinska E, Garaude J, http://www.thno.org/v08p3856s1.pdf Allende-Vega N. From tumor cell metabolism to tumor immune escape. Int J Biochem Cell Biol. 2013; 45: 106-13. 13. Villalba M, Lopez-Royuela N, Krzywinska E, Rathore MG, Hipskind RA, Acknowledgements Haouas H, et al. Chemical metabolic inhibitors for the treatment of blood-borne cancers. Anti-cancer agents in medicinal chemistry. 2014; 14: The authors are in debt to Drs. O. Gonzalo and I. 223-32. 14. Baier C, Fino A, Sanchez C, Farnault L, Rihet P, Kahn-Perles B, et al. Natural Marzo, Doctoral Thesis, University of Zaragoza, 2017 Killer Cells Modulation in Hematological Malignancies. Frontiers in for giving us the BCL-X L and MCL1-overexpressing immunology. 2013; 4: 459. 15. Ambrosini P, Loiacono F, Conte R, Moretta L, Vitale C, Mingari MC. IL-1beta Jurkat cells. We acknowledge the imaging facility inhibits ILC3 while favoring NK-cell maturation of umbilical cord blood MRI, member of the national infrastructure CD34(+) precursors. Eur J Immunol. 2015; 45: 2061-71. 16. Moretta L, Pietra G, Vacca P, Pende D, Moretta F, Bertaina A, et al. Human NK France-BioImaging supported by the French National cells: From surface receptors to clinical applications. Immunology letters. 2016; Research Agency (ANR-10-INBS-04, «Investments for 178: 15-9. 17. Muntasell A, Ochoa MC, Cordeiro L, Berraondo P, Lopez-Diaz de Cerio A, the future»). The Région Languedoc Roussillon Cabo M, et al. Targeting NK-cell checkpoints for cancer immunotherapy. Curr supports the clinical data and samples (HEMODIAG_ Opin Immunol. 2017; 45: 73-81. 18. Sarvaria A, Jawdat D, Madrigal JA, Saudemont A. Umbilical Cord Blood 2020). We gratefully thank Dr. Robert A. Hipskind for Natural Killer Cells, Their Characteristics, and Potential Clinical Applications. his critical revision of this manuscript. Frontiers in immunology. 2017; 8: 329. 19. Kottaridis PD, North J, Tsirogianni M, Marden C, Samuel ER, Jide-Banwo S, et al. Two-Stage Priming of Allogeneic Natural Killer Cells for the Treatment of Funding Patients with Acute Myeloid Leukemia: A Phase I Trial. PLoS One. 2015; 10: e0123416. This work was supported by an AOI from the 20. Besser MJ, Shoham T, Harari-Steinberg O, Zabari N, Ortenberg R, Yakirevitch CHU Montpellier (N°221826;GC/MV), La Ligue A, et al. Development of allogeneic NK cell adoptive transfer therapy in metastatic melanoma patients: in vitro preclinical optimization studies. PLoS Regionale contre le Cancer (GC), Fondation de France One. 2013; 8: e57922. (0057921;MV), the PRT-K program 2018 (MV/GC/ 21. Deng X, Terunuma H, Terunuma A, Takane T, Nieda M. Ex vivo-expanded natural killer cells kill cancer cells more effectively than ex vivo-expanded BR), the “Investissements d’avenir” Grant LabEx gammadelta T cells or alphabeta T cells. International immunopharmacology. MAbImprove: ANR-10-LABX-53 (GC/PM/BR) and 2014; 22: 486-91. 22. Voskens CJ, Watanabe R, Rollins S, Campana D, Hasumi K, Mann DL. Ex-vivo the NK 001 projet financed by the "Fonds Europeen de expanded human NK cells express activating receptors that mediate Developpement Regional (FEDER-FSE-IEJ 2014/2020) cytotoxicity of allogeneic and autologous cancer cell lines by direct recognition and antibody directed cellular cytotoxicity. Journal of experimental & clinical et par la region Occitanie Pyrénées-Méditerranée, cancer research: CR. 2010; 29: 134. 23. Wang W, Erbe AK, Hank JA, Morris ZS, Sondel PM. NK Cell-Mediated SAF2014- 54763-C2-1-R (JP), SAF2017-83120-C2-1-R Antibody-Dependent Cellular Cytotoxicity in Cancer Immunotherapy. (JP) and SAF2014-54763-C2-2-R (EMG) from Spanish Frontiers in immunology. 2015; 6: 368. 24. Sanchez-Martinez D, Krzywinska E, Rathore MG, Saumet A, Cornillon A, Ministry of Economy and Competitiveness. Lopez-Royuela N, et al. All-trans retinoic acid (ATRA) induces miR-23a expression, decreases CTSC expression and granzyme B activity leading to Competing Interests impaired NK cell cytotoxicity. Int J Biochem Cell Biol. 2014; 49: 42-52. 25. Sanchez-Martinez D, Azaceta G, Muntasell A, Aguilo N, Nunez D, Galvez EM, The authors have declared that no competing et al. Human NK cells activated by EBV lymphoblastoid cells overcome anti-apoptotic mechanisms of drug resistance in haematological cancer cells. interest exists. Oncoimmunology. 2015; 4: e991613. 26. Sanchez-Martinez D, Lanuza PM, Gomez N, Muntasell A, Cisneros E, Moraru M, et al. Activated Allogeneic NK Cells Preferentially Kill Poor Prognosis References B-Cell Chronic Lymphocytic Leukemia Cells. Frontiers in immunology. 2016; 1. Neves H, Kwok HF. Recent advances in the field of anti-cancer 7: 454. immunotherapy. BBA clinical. 2015; 3: 280-8. 27. Nunez D, Domingo MP, Sánchez-Martínez D, Cebolla V, Chiou A, 2. Scott AM, Wolchok JD, Old LJ. Antibody therapy of cancer. Nat Rev Cancer. Velázquez-Campoy A, et al. Recombinant production of human ICAM-1 2012; 12: 278-87. chimeras by single step on column refolding and purification. Process in 3. Hu Z. Overcome the Impairment of NK Cells for Icon and Antibody Biochemistry. 2013; 48: 708-15. Immunotherapy of Cancer. Journal of Immune Based Therapies, Vaccines and 28. Allende-Vega N, Krzywinska E, Orecchioni S, Lopez-Royuela N, Reggiani F, Antimicrobials. 2013; 2: 1-8. Talarico G, et al. The presence of wild type p53 in hematological cancers

http://www.thno.org Theranostics 2018, Vol. 8, Issue 14 3869

improves the efficacy of combinational therapy targeting metabolism. 53. Shah NN, Baird K, Delbrook CP, Fleisher TA, Kohler ME, Rampertaap S, et al. Oncotarget. 2015; 6: 19228-45. Acute GVHD in patients receiving IL-15/4-1BBL activated NK cells following 29. Aguilo JI, Garaude J, Pardo J, Villalba M, Anel A. Protein kinase C-theta is T-cell-depleted stem cell transplantation. Blood. 2015; 125: 784-92. required for NK cell activation and in vivo control of tumor progression. J 54. Vey N, Bourhis JH, Boissel N, Bordessoule D, Prebet T, Charbonnier A, et al. A Immunol. 2009; 182: 1972-81. phase 1 trial of the anti-inhibitory KIR mAb IPH2101 for AML in complete 30. Ndhlovu LC, Lopez-Verges S, Barbour JD, Jones RB, Jha AR, Long BR, et al. remission. Blood. 2012; 120: 4317-23. Tim-3 marks human natural killer cell maturation and suppresses 55. Carlsten M, Korde N, Kotecha R, Reger R, Bor S, Kazandjian D, et al. cell-mediated cytotoxicity. Blood. 2012; 119: 3734-43. Checkpoint Inhibition of KIR2D with the Monoclonal Antibody IPH2101 31. Benson DM, Jr., Bakan CE, Mishra A, Hofmeister CC, Efebera Y, Becknell B, et Induces Contraction and Hyporesponsiveness of NK Cells in Patients with al. The PD-1/PD-L1 axis modulates the natural killer cell versus multiple Myeloma. Clin Cancer Res. 2016; 22: 5211-22. myeloma effect: a therapeutic target for CT-011, a novel monoclonal anti-PD-1 56. Lagrue K, Carisey A, Morgan DJ, Chopra R, Davis DM. Lenalidomide antibody. Blood. 2010; 116: 2286-94. augments actin remodeling and lowers NK-cell activation thresholds. Blood. 32. Luetke-Eversloh M, Killig M, Romagnani C. Signatures of human NK cell 2015; 126: 50-60. development and terminal differentiation. Frontiers in immunology. 2013; 4: 499. 33. Nielsen CM, White MJ, Goodier MR, Riley EM. Functional Significance of CD57 Expression on Human NK Cells and Relevance to Disease. Frontiers in immunology. 2013; 4: 422. 34. Trivedi S, Srivastava RM, Concha-Benavente F, Ferrone S, Garcia-Bates TM, Li J, et al. Anti-EGFR Targeted Monoclonal Antibody Isotype Influences Antitumor Cellular Immunity in Head and Neck Cancer Patients. Clin Cancer Res. 2016; 22: 5229-37. 35. Evans SS, Clemmons AB. Obinutuzumab: A Novel Anti-CD20 Monoclonal Antibody for Chronic Lymphocytic Leukemia. Journal of the advanced practitioner in oncology. 2015; 6: 370-4. 36. Zhu Y, Huang B, Shi J. Fas ligand and lytic granule differentially control cytotoxic dynamics of Natural Killer cell against cancer target. Oncotarget. 2016. 37. Barber DF, Faure M, Long EO. LFA-1 contributes an early signal for NK cell cytotoxicity. J Immunol. 2004; 173: 3653-9. 38. Shallis RM, Terry CM, Lim SH. The multi-faceted potential of CD38 antibody targeting in multiple myeloma. Cancer Immunol Immunother. 2017. 39. Veluchamy JP, Heeren AM, Spanholtz J, van Eendenburg JD, Heideman DA, Kenter GG, et al. High-efficiency lysis of cervical cancer by allogeneic NK cells derived from umbilical cord progenitors is independent of HLA status. Cancer Immunol Immunother. 2017; 66: 51-61. 40. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, et al. The landscape of somatic copy-number alteration across human cancers. Nature. 2010; 463: 899-905. 41. De Miguel D, Basanez G, Sanchez D, Malo PG, Marzo I, Larrad L, et al. Liposomes decorated with Apo2L/TRAIL overcome chemoresistance of human hematologic tumor cells. Molecular pharmaceutics. 2013; 10: 893-904. 42. Bojarczuk K, Sasi BK, Gobessi S, Innocenti I, Pozzato G, Laurenti L, et al. BCR signaling inhibitors differ in their ability to overcome Mcl-1-mediated resistance of CLL B cells to ABT-199. Blood. 2016; 127: 3192-201. 43. Balsamo M, Vermi W, Parodi M, Pietra G, Manzini C, Queirolo P, et al. Melanoma cells become resistant to NK-cell-mediated killing when exposed to NK-cell numbers compatible with NK-cell infiltration in the tumor. Eur J Immunol. 2012; 42: 1833-42. 44. Carrega P, Morandi B, Costa R, Frumento G, Forte G, Altavilla G, et al. Natural killer cells infiltrating human nonsmall-cell lung cancer are enriched in CD56 bright CD16(-) cells and display an impaired capability to kill tumor cells. Cancer. 2008; 112: 863-75. 45. Halama N, Braun M, Kahlert C, Spille A, Quack C, Rahbari N, et al. Natural killer cells are scarce in colorectal carcinoma tissue despite high levels of chemokines and cytokines. Clin Cancer Res. 2011; 17: 678-89. 46. Parkhurst MR, Riley JP, Dudley ME, Rosenberg SA. Adoptive transfer of autologous natural killer cells leads to high levels of circulating natural killer cells but does not mediate tumor regression. Clin Cancer Res. 2011; 17: 6287-97. 47. Federico SM, McCarville MB, Shulkin BL, Sondel PM, Hank JA, Hutson P, et al. A Pilot Trial of Humanized Anti-GD2 Monoclonal Antibody (hu14.18K322A) with Chemotherapy and Natural Killer Cells in Children with Recurrent/Refractory Neuroblastoma. Clin Cancer Res. 2017; 23: 6441-9. 48. Talleur AC, Triplett BM, Federico S, Mamcarz E, Janssen W, Wu J, et al. Consolidation Therapy for Newly Diagnosed Pediatric Patients with High-Risk Neuroblastoma Using Busulfan/Melphalan, Autologous Hematopoietic Cell Transplantation, Anti-GD2 Antibody, Granulocyte-Macrophage Colony-Stimulating Factor, Interleukin-2, and Haploidentical Natural Killer Cells. Biol Blood Marrow Transplant. 2017; 23: 1910-7. 49. Shevtsov M, Multhoff G. Immunological and Translational Aspects of NK Cell-Based Antitumor Immunotherapies. Frontiers in immunology. 2016; 7: 492. 50. Bachanova V, Burns LJ, McKenna DH, Curtsinger J, Panoskaltsis-Mortari A, Lindgren BR, et al. Allogeneic natural killer cells for refractory lymphoma. Cancer Immunol Immunother. 2010; 59: 1739-44. 51. Ramos CA, Savoldo B, Dotti G. CD19-CAR trials. Cancer journal. 2014; 20: 112-8. 52. Ullrich E, Salzmann-Manrique E, Bakhtiar S, Bremm M, Gerstner S, Herrmann E, et al. Relation between Acute GVHD and NK Cell Subset Reconstitution Following Allogeneic Stem Cell Transplantation. Frontiers in immunology. 2016; 7: 595.

http://www.thno.org ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! |~Ñ! ) ) ) ) ) ) )

!'".1#/!kZ!! !

1>8!.")d!(F>?5?LGJ!"?6>DGJG468L4L8!!GFLJGDK! !>GD8KL8JGD!'GE8GKL4K?K!1>JGM=>!L>8! #/)h[+#%e!.4L>O4Q! ) I<@087>632+/=?7183+H++173:?7471+*3<;=?>J+ ) ) K&:+ ?67>+ >?@2D+ ?6/?+ &+ >6/=32+ 1;G/@?6;=>67+ 7:A;8A32+ 7:+ <8/::7:5+ /:2+ 3C31@?7;:+;4+?63+3C<3=793:?>E++7:18@27:5+<=3+@>32+7:+?63+>?@2DL+ ) ) ) ) ) ) ) ) ) )

! |~Ö! ™™™Ǥƒ–—”‡Ǥ ‘Ȁ• ‹‡–‹ˆ‹ ”‡’‘”–•

 Š‡ͷ Š‹„‹–‘”‹ ŠŽ‘”‘ƒ ‡–ƒ–‡ ‘–”‘Ž•Š‘Ž‡•–‡”‘Ž ‘‡‘•–ƒ•‹• Š”‘—‰Š–Š‡ͻȀ ͸ƒ–Š™ƒ› Received: 25 April 2017 „”ƒ”Ž ƒ“Šƒ ͷǡ͸ ǡ‡”‡ƒŽŽ‡†‡Ǧ‡‰ƒ ͷǡ͸ ǡ‡Ž’Š‹‡ ‹–‡ƒ› ͷǡ͸ ǡƒ„‹‡ ‡”„ƒŽǦŠƒŽ‘‹ ͷǡ͸ ǡ Accepted: 8 August 2017 Žƒ‹”‡ ‘†‡ƒ— ͷǡ͸ǡ͹ ǡƒ‰Ǧ‰Š‹‡‘ ͷǡƒƒ‡ŽƒŠŽƒ ͷǡ–‡ˆƒ‹ƒ”‡ Š‹‘‹ ͺǡ ‹‘˜ƒƒ Published: xx xx xxxx ƒŽƒ”‹ ‘ ͺǡ ”ƒ ‡• ‘‡”–‘Ž‹‹ ͺǡ‹Ž‹ ƒ‘œ‹ ͻǡ ‘•‡ǤƒŽ†‹˜‹‡Ž•‘ ͻǡ ƒ„‹‡‡ Bejj ƒ‹ ͼǡ •ƒ„‡ŽŽ‡ ƒ”‹‡Ž ͼǡ •ƒ„‡ŽǤ‘’‡œǦ‡Œ‹ƒ ͽǡŽ—‹• ƒŒƒ• ͽǡŠƒ”Ž‡•Ǧ ‡”‹‡ ‡ŽŽ‹‡” ͼǡ ƒ˜‹‡” ‡”ƒ†‡œ ͷǡƒ”–‹‡ƒ—Œƒ– ͷǡ͸ Ƭƒ”–‹‹ŽŽƒŽ„ƒ ͷǡ͸

‘–”‘ŽŽ‹‰ Š‘Ž‡•–‡”‘ŽŽ‡˜‡Ž•‹•ƒƒŒ‘” ŠƒŽŽ‡‰‡‹Š—ƒŠ‡ƒŽ–Šǡ•‹ ‡Š›’‡” Š‘Ž‡•–‡”‘Ž‡‹ƒ ƒ Ž‡ƒ†–‘•‡”‹‘—• ƒ”†‹‘˜ƒ• —Žƒ”†‹•‡ƒ•‡Ǥ”—‰•–Šƒ––ƒ”‰‡– ƒ”„‘Š›†”ƒ–‡‡–ƒ„‘Ž‹• ƒƒŽ•‘‘†‹ˆ› Ž‹’‹†‡–ƒ„‘Ž‹•ƒ†Š‡ ‡ Š‘Ž‡•–‡”‘Ž’Žƒ•ƒŽ‡˜‡Ž•Ǥ –Š‹••‡•‡ǡ†‹ ŠŽ‘”‘ƒ ‡–ƒ–‡ȋȌǡƒ’›”—˜ƒ–‡ †‡Š›†”‘‰‡ƒ•‡‹ƒ•‡ȋȌ‹Š‹„‹–‘”ǡƒ—‰‡–•—•ƒ‰‡‘ˆ–Š‡‰Ž› ‘Ž›•‹•Ǧ’”‘†— ‡†’›”—˜ƒ–‡‹ –Š‡‹–‘ Š‘†”‹ƒ‹ ”‡ƒ•‹‰‘š‹†ƒ–‹˜‡’Š‘•’Š‘”›Žƒ–‹‘ȋ ȌǤ •‡˜‡”ƒŽƒ‹ƒŽ‘†‡Ž•ǡ †‡ ”‡ƒ•‡•’Žƒ•ƒ Š‘Ž‡•–‡”‘Žƒ†–”‹‰Ž› ‡”‹†‡•ǤŠ—•ǡ™ƒ•—•‡†‹–Š‡ͽͶ•–‘–”‡ƒ–†‹ƒ„‡–‡• ‡ŽŽ‹–—•ǡŠ›’‡”Ž‹’‘’”‘–‡‹‡‹ƒƒ†Š›’‡” Š‘Ž‡•–‡”‘Ž‡‹ƒ™‹–Š•ƒ–‹•ˆƒ –‘”›”‡•—Ž–•Ǥ ‘™‡˜‡”ǡ–Š‡ ‡ Šƒ‹•‘ˆƒ –‹‘”‡ƒ‹‡†—‘™ƒ†™‡†‡• ”‹„‡‹–Š‡”‡Ǥ‹ ”‡ƒ•‡• LDLR ƒ† ’”‘–‡‹Ž‡˜‡Ž•ƒ•™‡ŽŽƒ•‹–ƒ‡‹•‡˜‡”ƒŽ ‡ŽŽŽ‹‡•ǡ’”‹ƒ”›Š—ƒŠ‡’ƒ–‘ ›–‡•ƒ†–™‘†‹ơ‡”‡– ‘—•‡‘†‡Ž•ǤŠ‹•‡ơ‡ –‹•‡†‹ƒ–‡†„›–”ƒ• ”‹’–‹‘ƒŽƒ –‹˜ƒ–‹‘ƒ•‡˜‹†‡ ‡†„› ͹ƒ ‡–›Žƒ–‹‘ ‘Ž›•‹‡͸ͽ‘–Š‡ LDLR ’”‘‘–‡”Ǥ‹†— ‡•‡š’”‡••‹‘‘ˆ–Š‡ͻ–Šƒ––—”•‘–Š‡ –”ƒ• ”‹’–‹‘ˆƒ –‘” ͸Ǥ Š‹„‹–‹‘‘ˆ–Š‹•ͻȀ ͸’ƒ–Š™ƒ›„›‰‡‡–‹ ‘”’Šƒ”ƒ ‘Ž‘‰‹ ƒŽ‡ƒ• †‡ ”‡ƒ•‡•‡š’”‡••‹‘ƒ†‹–ƒ‡Ǥ •—ƒ”›ǡ‘—””‡•—Ž–•‹†‹ ƒ–‡–Šƒ–ǡ„›‹†— ‹‰  ǡ’”‘‘–‡•ͻȀ ͸ƒ –‹˜ƒ–‹‘Ž‡ƒ†‹‰–‘‡š’”‡••‹‘ǤŠ‡ͻȀ ͸’ƒ–Š™ƒ›‘ơ‡”• ƒ‹–‡”‡•–‹‰’Šƒ”ƒ ‘Ž‘‰‹ ƒŽ–ƒ”‰‡–ˆ‘”†”—‰†‡˜‡Ž‘’‡–Ǥ

Elevated level of low-density lipoprotein (LDL) in blood is a predominant risk factor for atherosclerosis, a large cause of mortality 1. Control of plasma cholesterol levels largely depends on low-density lipoprotein receptor (LDLR), which mediates the endocytosis of cholesterol-rich LDL 2, 3. !is process takes place mainly in the liver 2, 3. LDL is degraded in lysosomes and cholesterol made available for repression of microsomal enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase, the rate-limiting step in cholesterol synthesis 2, 3. Hence, LDLR regulates intracellular and extracellular cholesterol homeostasis and is involved in atherosclerosis due to accumulation of LDL-cholesterol in blood 4. Lipid and carbohydrate metabolic pathways are interconnected and targeting the latter may result in altered cholesterol levels. !e pyruvate dehydrogenase (PDH) kinase (PDK) inhibitor dichloroacetate (DCA) acti- vates PDH, the rate-limiting enzyme of aerobic glucose oxidation 5. PDH converts glycolysis-produced pyru- vate in acetyl-CoA that enters the mitochondria and is consumed in the process of oxidative phosphorylation (OXPHOS). !us, DCA inhibits glycolysis and lactate production and induces OXPHOS 6–10 . DCA concentration in DCA-treated patients is unclear because its half-life is less than 1 hour and it is not detectable in patients during the initial phase of treatment that can last 2 to 3 months 11 , 12 . However, DCA inhibits its own metabolism

ͷ ǡͷͷ;͹Ǣ‹˜‡”•‹–±†‡‘–’‡ŽŽ‹‡”ǡ ‡†‡ ‹‡ǡ;Ͷǡƒ˜Ǥ—‰—•–‹ Ž‹ Š‡ǡ͹ͺ͸Ϳͻǡ‘–’‡ŽŽ‹‡”‡†‡šͻǡ ”ƒ ‡Ǥ ͸ •–‹–—–†‡±†‡ ‹‡±‰±±”ƒ–”‹ ‡‡–‹‘–Š±”ƒ’‹‡ȋ Ȍǡ ‘–’‡ŽŽ‹‡”ǡ‘–’‡ŽŽ‹‡”ǡ͹ͺ͸Ϳͻǡ ”ƒ ‡Ǥ ͹±’ƒ”–‡‡–†ǯ ±’ƒ–‘Ǧ‰ƒ•–”‘‡–±”‘Ž‘‰‹‡ǡ ؒ‹–ƒŽƒ‹–Ž‘‹ǡ ǡ‘–’‡ŽŽ‹‡”ǡ ”ƒ ‡Ǥ ͺƒ„‘”ƒ–‘”›‘ˆ ‡ƒ–‘Ž‘‰›Ǧ ‘Ž‘‰›ǡ—”‘’‡ƒ •–‹–—–‡‘ˆ ‘Ž‘‰›ǡ‹Žƒǡ –ƒŽ›Ǥ ͻƒ• —Žƒ”ƒ†‡ƒŽ”ƒ•Žƒ–‹‘ƒŽ‡•‡ƒ” Š ”‘—’Ǥ •–‹–—–†‡‡ ‡” ƒ‹‘‡†‹ ƒ†‡Ž‡‹†ƒȋ  ȌǡŽ‡‹†ƒǡ’ƒ‹Ǥ ͼ ǡǡ‹˜Ǥ‘–’‡ŽŽ‹‡”ǡ ‘–’‡ŽŽ‹‡”ǡ ”ƒ ‡Ǥ ͽ‡–‡”ˆ‘” –‡‰”ƒ–‹˜‡ ‡‘‹ •ǡ‹˜‡”•‹–›‘ˆƒ—•ƒ‡ǡƒ—•ƒ‡ǡ™‹–œ‡”Žƒ†Ǥ ‘””‡•’‘†‡ ‡ƒ†”‡“—‡•–•ˆ‘”ƒ–‡”‹ƒŽ••Š‘—Ž†„‡ƒ††”‡••‡†–‘ǤǤȋ‡ƒ‹Žǣ ƒ”–‹Ǥ˜‹ŽŽƒŽ„ƒ̻‹•‡”Ǥˆ” Ȍ

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 1 www.nature.com/scientificreports/

and serum concentrations increase, eventually reaching a plateau, with plasma concentrations around 0.3 mM 11 . Michelakis et al . gave 50 mg/Kg/day of DCA to patients and obtained similar values: 0.44 ± 0.16 mM 12 . DCA treatment decreases plasma triglyceride and cholesterol in animal models and humans 13 , 14 . !is e"ect is likely indirect due to the decrease in plasma of very-low-density lipoproteins (VLDL) by stimulation of triglyc- eride oxidation 5, 14 , although the mechanism remains unknown. A#er these $ndings there was an attempt to use DCA to treat two cases of familial hypercholesterolemia (FH) 15 . DCA reduced circulating cholesterol levels in both patients by decreasing LDL cholesterol 15 . Unfortunately, DCA had to be discontinued following neuropatho- logical e"ects and this precluded its use to treat high cholesterol levels 15 . !e biological mechanism underlying DCA e"ects on cholesterol was unknown and, to the best of our knowledge, no further studies were conducted to investigate it. Nowadays, the most accepted explanation would be that the change in carbohydrate metabolism alters lipid metabolism. In this sense, DCA, by inhibiting glycolysis, activates AMPK9, which inhibits cholesterol synthesis 16 , 17 leading to LDLR expression 18 , 19 . Enhanced LDLR expression is mediated by a MAPKK because the MAPKK inhibitors PD98059 18 and U0126 20 , 21 restrain it. !is would indicate that DCA has a similar mechanism of action than the alkaloid berberine and its analogs 18 , 19 . An alternative explanation could be that DCA induces ROS production 9, 10 and the cellular oxidative state regulates LDLR expression 22 . Hence, changes in ROS could mediate DCA-induced LDLR expression. In preliminary experiments, we found by transcriptome analysis that LDLR was one of the most downregu- lated genes in hematopoietic cells expressing a small hairpin RNA for ERK5 (shERK5) and one of the most upreg- ulated a#er DCA treatment. We have previously observed that DCA increased mRNA and protein expression of the MAPK ERK5, which is essential for cell survival in OXPHOS conditions 8, 10 , 23 –26 . ERK5 activates the MEF2 family of transcription factors 27 –30 that in turn mediates most of the e"ects of ERK5 on metabolism 8, 10 , 23 –26 . In fact, DCA activates multiple genes with promoters containing MEF2 binding sites. Genomic analysis using the UCSC genome browser ( http://genome.ucsc.edu/ ) showed that LDLR is one of such genes because its promoter contains several binding sites for at least two MEF2 proteins, MEF2A and MEF2C, which have been validated in several cell lines by Chromatin Immunoprecipitation (ChIP; (http://genome.ucsc.edu/). We investigate here how DCA reduces cholesterol levels and the role of the ERK5/MEF2 pathway. ‡•—Ž–• ‡Šƒ ‡•‡š’”‡••‹‘Ǥ We $rst observed that DCA increased LDLR mRNA in hematopoietic cells (Fig. 1A , le# panel). Since liver is the main organ that takes up LDL-cholesterol from blood, we tested the e"ect of DCA in two hepatic cell lines, $nding that LDLR mRNA levels were also increased (Fig. 1A , right panel). Augmented LDLR mRNA levels correlated with an increase of LDLR protein in the plasma membrane (Fig. 1B and Supplemental Fig. 1A ). We then tested the functional consequence of this enhanced expression by incubating control or DCA-treated cells with &uorescently labeled LDL. DCA increased LDL transport into these cell lines (Fig. 1C and Supplemental Fig. 1B ).

‡ŽŽ•’‡”ˆ‘”‹‰ ‹ ”‡ƒ•‡ƒ –‹˜‹–›Ǥ DCA induces OXPHOS in leukemic cells 8–10 , 24 , 25 , 31 . To investigate whether the metabolic switch from aerobic glycolysis to OXPHOS mediated by DCA a"ect LDLR expression, we used a glucose-free culture medium containing a $nal glutamine concentration of 4 mM and 10 mM galactose. Glutamine is used to drive mitochondria to utilize OXPHOS and galactose allows cells to syn- thesize nucleic acids through the pentose phosphate pathway 8, 24 , 32 , 33 . We called it ‘OXPHOS medium’ because it forced leukemic cells to use OXPHOS as primary ATP source 8, 9, 23 . We observed that acute myeloid leukemia (AML) OCI-AML3 cells growing in OXPHOS medium for 2 weeks, such as those treated with DCA, presented an increase of ERK5 and LDLR mRNA (Supplemental Fig. 2A ), LDLR protein and LDL intake (Fig. 1D ). !ese results indicated that, as expected, the e"ect of DCA on LDLR expression was due to a metabolic switch. DCA and OXPHOS also increased LDLR mRNA and protein as well as LDL intake in primary lymphoma cells derived from a B cell lymphoma patient (BCL-P2; Fig. 1A , Supplemental Fig. 1B and Fig. 1E ). We found similar results in the hepatic cell lines HepG2-C3A and HuH7, with OXPHOS media increasing LDLR protein and uptake (Supplemental Fig. 3). In primary human hepatocytes, DCA also induced LDLR expression at 6 and 24 h post treatment (Fig. 2A ). However, e"ects disappeared at 48 and 72 h with a net decrease (Fig. 2A ). !is is likely due to the short DCA half-life, since LDLR mRNA levels were kept high when fresh DCA was added to the medium every 24 h (Fig. 2B ).

‡ƒ –‹˜‡š›‰‡’‡ ‹‡•ȋȌ†‘‘–‡†‹ƒ–‡Ǧ‹†— ‡†‡š’”‡••‹‘Ǥ !e cellular oxidative state can regulate LDLR expression 22 and DCA induces ROS production in certain hematopoietic cell lines but not in others 9, 10 . We observed a similar phenomenon in two di"erent hepatic cell lines. In Huh7 cells there was an increase of ROS, but not in HepG2-C3A cells, a#er DCA treatment (Fig. 3A ). In contrast, in all cell lines utilized in this study, DCA increased LDLR expression, suggesting that ROS production was not essential (Figs 1A and 3B ). Next, we incubated both hepatic cell lines with the antioxidant N-acetyl-cysteine (NAC). NAC e*ciently blocked DCA-induced ROS production in HuH7 cells (Fig. 3A ); however, it did not a"ect DCA-induced LDLR mRNA (Fig. 3B ) or plasma membrane protein (Fig. 3C ). To de$nitively exclude that ROS played any role in LDLR induction, we incubated primary hepatocytes with DCA in presence or absence of NAC. We found that NAC did not inhibit and in fact increased LDLR mRNA expression (Fig. 3D ). !ese results exclude a major role for de novo ROS production in LDLR expression a#er DCA treatment.

‹†— ‡• in vivo Ǥ We next assessed whether DCA enhances LDLR expression in vivo . We engrafted human AML primary cells in non-obese diabetic/severe combined immunodeficient (NOD/ SCID)-interleukin-2 receptor γ chain null (NSG) mice, as previously described 9, 10 . Mice with established tumors (day 80 post-gra#) were treated daily with DCA (Fig. 4A ). !e treatment was not toxic and did not show any

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 2 www.nature.com/scientificreports/

Figure 1. OXPHOS induced LDLR expression and LDL uptake. ( A) !e hematopoietic cell lines Jurkat and OCI-AML3 and primary cells from a BCL patient (BCL-P2) as well as HepG2-C3A and Huh7 hepatic cell lines were treated with 10 mM DCA for 24 h and LDLR mRNA was analyzed by RT-qPCR. ( B) Cell lines were treated for 72 h with 5 mM DCA and LDLR protein in plasma membrane was analyzed by FACs. (C) Cells were treated as in ( B) and &uorescent LDL intake analyzed by FACs. ( D) OCI-AML3 cells were grown in OXPHOS medium for 2 weeks and LDLR expression (le#) and LDL intake (right) were analyzed by FACs. ( E) BCL-P2 cells were treated with 5 mM DCA for 1 week (le#) or were grown in OXPHOS medium for 2 weeks (center) and LDLR protein in plasma membrane analyzed by FACs. LDL intake (right) was analyzed in cells growing in OXPHOS. !e bar graphs represent means ± SD of 3 independent experiments performed in triplicate; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test compared to control cells.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 3 www.nature.com/scientificreports/

Figure 2. DCA induced LDLR expression in primary human hepatocytes. ( A) Primary hepatocytes were treated with the indicated concentrations of DCA for the indicated times. ( B) Cells were treated at time 0 with DCA and some were treated every 24 h before harvesting as indicated. LDLR mRNA was analyzed by RT-qPCR. !e bar graphs represent means ± SD of 3 independent donors performed in triplicate; *p < 0.05, ** p < 0.01, ***p < 0.005 one-way ANOVA with post-hoc Tukey test.

notable e"ect on mice survival 9. Human tumor AML cells gather in murine spleen and bone marrow, hence we isolated mRNA from these organs. We used human-speci$c primers and observed that DCA signi$cantly increased expression of LDLR mRNA (Fig. 4A ). We also found augmented mouse LDLR mRNA levels in normal liver and spleen from wt mice that were treated daily, for 1 and up to 3 days, with DCA (Fig. 4B ). !e e"ect was $rst observed in hematopoietic cells

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 4 www.nature.com/scientificreports/

Figure 3. Increase in ROS levels was not required for LDLR expression. ( A) Both hepatic cell lines were treated with 2 mM NAC 1 h before adding DCA (10 mM) for 24 h. Cells were labeled with CH-H2DCFDA and analyzed by FACs for ROS production. LDLR mRNA ( B) or protein ( C) from these cells were analyzed as described in Fig. 1. Results represent the means ± SD of these donors with experiments performed in triplicate. !e data represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test compared to cells non treated with DCA. D) Primary hepatocytes from 2 independent donors were treated for 6 and 24 h as in ( A) but with two di"erent DCA concentrations before analyzing LDLR mRNA expression. !e data represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 one-way ANOVA with post-hoc Tukey test.

gathering in spleen and, later, in liver. !us, DCA induced LDLR expression in multiple cell populations in vivo . !is could, at least partially, explain the reduction in plasma cholesterol levels a#er DCA treatment in several species including humans 5, 13 –15 , 34 .

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 5 www.nature.com/scientificreports/

Figure 4. DCA induced LDLR expression in vivo . ( A) NSG mice were engra#ed with primary human AML cells. At day 80 post-gra#, they were treated with DCA (n = 4) or leave untreated (n = 4). At day 140, mRNA from bone marrow or spleen was isolated and human LDLR mRNA expression was quanti$ed by qPCR. !e data represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test compared to non treated mice. (B) B6 wt mice (n = 4/5 per group) were treated with a dose of DCA (50 mg/kg) everyday intraperitoneally and mouse LDLR mRNA was analyzed in spleen and liver at indicated time points. !e data represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 one-way ANOVA with post-hoc Tukey test.

ͻ”‡‰—Žƒ–‡•‡š’”‡••‹‘Ǥ We further investigated the underlying mechanism of DCA-induced LDLR expression and the role of the ERK5/MEF2 pathway, which is activated by DCA 8, 10 , 23 , 24 . To this end, we targeted ERK5 utilizing a small hairpin RNA (shERK5). Reducing ERK5 expression resulted in decreased LDLR mRNA levels in non-treated hematopoietic cells (Fig. 5A ). We could not investigate DCA e"ects on cells express- ing shERK5 because ERK5 is essential to perform OXPHOS and hence DCA is highly toxic in cells with reduced ERK5 levels 8, 10 , 23 , 24 . Conversely, overexpression of ERK5 increased LDLR mRNA levels (Fig. 5A ). As shown in Supplemental Fig. 4A , transfection with shERK5 or ERK5 expressing vectors e*ciently decreased and increased ERK5 protein levels respectively. Decreasing ERK5 levels with small interference RNA for ERK5 (siERK5) also impaired LDLR expression in primary hepatocytes (Fig. 5B ) or in HuH7 hepatic cells (Supplemental Fig. 5A), in which we observed a reduction of 40% on ERK5 protein levels (Supplemental Fig. 4B ). Overexpression of ERK5 in Jurkat cells augmented LDLR protein and enhanced LDL uptake (Fig. 5C ). !e MAPKK MEK5 activates ERK5 in di"erent physiological contexts 35 . !us, we next used the MEK5 inhibitor BIX02189, which inhibits its catalytic function, and showed that it decreased LDLR protein and LDL uptake in Jurkat and OCI-AML3 cells (Fig. 5D ). We validated these $ndings in primary tumor cells derived from a BCL patient (Supplemental Fig. 6). Treatment of these cells with an extremely selective ERK5 inhibitor, XMD8–92, resulted in the reduction of LDLR protein (Supplemental Fig. 6). Taken together, these results indicate that ERK5 is essential for LDLR expression and function in multiple cell lines.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 6 www.nature.com/scientificreports/

Figure 5. ERK5 controled LDLR expression and LDL uptake. ( A) 10 7 Jurkat-TAg cells were transfected with 5 µg of the empty pSUPER Neo vector (control) or with this vector containing a small hairpin RNA for ERK5 (shERK5) or with a pcDNA vector expressing ERK5 (ERK5). Forty-eight hours later mRNA expression of the whole population was analyzed by qPCR and represented as the % of mRNA compared to cells transfected with the empty vector. ( B) Primary hepatocytes were transfected with control siRNA (control) or with siRNA against ERK5 (siERK5). 96 h later mRNA was collected and ERK5 and LDLR mRNA expression was analyzed by qPCR. (C) Jurkat cells were transfected with ERK5 as described in ( A) and LDLR plasma membrane protein (le#) and LDL intake (right) were analyzed by FACs. D) Jurkat (le# and center) and OCI-AML3 (right) cells were treated with 5 µM of the MEK5 inhibitor BIX02189 for 24 h and LDLR protein (le#) or LDL intake (center and right) were analyzed by FACs. Bar graphs represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test compared to empty vector transfected cells (control).

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 7 www.nature.com/scientificreports/

†‘‡•‘–‡†‹ƒ–‡Ǧ‹†— ‡†‡š’”‡••‹‘Ǥ DCA induces AMPK activation, the main cellular metabolic sensor 36 . AMPK, by blocking de novo cholesterol production 16 , 17 , could mediate LDLR expres- sion as it has been observed with berberine or its analogs 18 , 19 . !e e"ect of berberine on LDLR expression was inhibited by PD98059 18 and U0126 20 , 21 , In fact, a PD98059 18 and U0126 20 , 21 sensible pathway mediated ber- berine e"ect. Although these MAPKK inhibitors were initially described as speci$c MEK1 inhibitors, they also inhibit the ERK5 upstream kinase MEK5 37 . !is indicates that DCA may have a similar mechanism of action than berberine. To test this hypothesis, we used metformin, which stimulates AMPK in Jurkat and OCI-AML cells 36 . Surprisingly, metformin did not increase, but rather decreased, LDLR mRNA levels (Fig. 6A ), protein levels (Fig. 6B ) and LDL intake (Fig. 6C ) in two hematopoietic cell lines. Metformin also decreased LDL uptake in HEPG2-C3A cells (Fig. 6D ). Moreover, blocking expression of the catalytic subunit of AMPK, AMPK α, with two di"erent siRNA that e"ectively decrease AMPK α levels 38 , did not statistically decrease LDLR mRNA. In summary, these data indicated that the AMPK pathway did not modulate DCA-induced LDLR expression and suggested that we uncovered a totally new pathway that controlled LDLR expression.

‡š’”‡••‹‘”‡“—‹”‡• ͸Ǥ ERK5 mediates part of its metabolic functions through the MEF2 fam- ily of transcription factors 8, 10 , 23 , 24 . Interestingly, LDLR promoter contains predicted binding sites for MEF2A and C that have been validated in several cell lines ( http://genome.ucsc.edu/ ). !erefore, we performed a knockdown of MEF2A and C in OCI-AML3 cells by siRNA. !ese siRNA halved mRNA expression of both transcription factors (Fig. 7A ) and reduced between 25 and 50% protein levels (Supplemental Fig. 4C ). !e level of knockdown was su*cient to signi$cantly decrease LDLR mRNA levels (Fig. 7A ). Finally, to further investigate whether DCA activated LDLR promoter, we checked the Histone H3 lysine 27 acetylation (H3K27Ac), a modi$cation associated with active gene expression 39 . We observed that H3K27Ac increased a#er DCA treatment indicating an increase of LDLR gene transcription (Fig. 7B ).

Š‡ͻȀ ͸’ƒ–Š™ƒ›ƒŽ•‘ ‘–”‘Ž•‡š’”‡••‹‘‘ˆ–Š‡”‡ ‡’–‘”Ǧƒ†ƒ’–‡”’”‘–‡‹ͷ ȋͷȌǤ !e LDL receptor-adapter protein 1 (LDLRAP1) is a cytosolic protein that interacts with the cytoplasmic tail of LDLR 40 . LDLRAP1 promoter contains MEF2 binding sites ( http://genome.ucsc.edu/ ), sug- gesting that the same ERK5/MEF2 pathway could regulate this gene. Consequently with this hypothesis, DCA enhanced LDLRAP1 expression in hepatic cell lines and primary hepatocytes (Fig. 8A ). OCI-AML3 and primary tumor B cells also increased LDLRAP1 mRNA a#er DCA treatment or a#er incubation in OXPHOS medium (Fig. 8B ). In non-stimulated cells, siERK5 reduced LDLRAP1 mRNA in primary hepatocytes, Huh7 cells and primary tumor cells (Fig. 8C (le#), Supplementary Fig. 5C ). Similarly, OCI-AML3 cells transfected with siMEF2 repressed the expression of LDLRAP1 mRNA (Fig. 8C (right). In summary, cells performing OXPHOS increased the expression of an additional protein involved in LDLR activity. ‹• —••‹‘ Carbohydrate and lipid metabolism are intrinsically bound and their dysfunction play a major role in cardiovas- cular disease. Diabetes is typically associated to dyslipidemia, but vice versa , lipid changes also disturb glucose metabolism 41 . DCA, by stimulating PDH activity, decreases glucose catabolism and stimulates OXPHOS 6, 7. To fuel it, cells could rely on fatty acid oxidation (FAO), suggesting that DCA could increase lipid catabolism. LDL particles transport cholesterol and triglycerides; hence an increase in LDLR should allow cells to increase fat avail- ability. We propose that the avidity for fatty acids induces DCA-treated cells to increase LDLR expression and, subsequently, cholesterol uptake. !is is supported by our observations showing that OXPHOS, which mediates FAO, reproduces DCA e"ects on LDLR expression in vitro (Fig. 1 and Supplemental Figs 2 and 3). !e mecha- nism requires ERK5, which directs the choice of catabolic substrates 8, 10 , 23 –26 and, hence, is a good candidate to control increase on fat avidity. Inhibiting ERK5 function by pharmacological or genetic means decrease LDL intake (Fig. 5). !is reduces the consumption of exogenous cholesterol but should also a"ect the consumption of fatty acids. In contrast, in vivo ERK5 stimulation, e.x; by DCA treatment, induces LDLR mRNA (Fig. 4) and would induce triglyceride oxidation as previously observed 5, 14 . !e ERK5 target MEF2 also regulates this pathway, likely because LDLR promoter contains MEF2 binding sites (see Introduction). ERK5 induces activation of several members of the MEF2 family of transcription factors by several mechanisms. It induces direct phosphorylation at di"erent serines and threonines on MEF2A, C and D 27 , 28 . It can also activate MEF2A and D by direct interaction because ERK5 serves as a MEF2 coactivator through its signal-dependent direct association with the MEF2 MADS domain. Although, at least MEF2A-dependent tran- scription requires ERK5 kinase activity 29 , 30 . MEF2 mediates several ERK5 e"ects on metabolism 10 . However, the exact mechanism via which DCA stimulates LDLR expression could be multifactorial, although it is a general phenomenon that we have con$rmed in several hematopoietic, colon and hepatic cells (Fig. 1), including primary human hepatocytes (Fig. 2), which are the main regulators of cholesterol levels 2, 3. We cannot exclude that other mechanisms play a role in DCA-induced LDLR expression. For example, DCA inhibits HMG CoA reductase activity in liver and leukocytes 6. !is could lead to an even higher demand on exogenous cholesterol and subsequently to an increase in LDLR levels. In addition, we have also recently shown that the MAPK ERK5 induces Sirt1 expression 24 that also stabilizes LDLR protein 42 . !erefore, ERK5 could target LDLR function in multiple ways and some of them independently of MEF2 family. In contrast, we have mainly excluded that ROS levels or AMPK activation play a major role in this process (Figs 3 and 6). Other drugs could modulate the ERK5/MEF2 pathway to regulate the expression of LDLR. Berberine induces LDLR expression by a MAPK pathway sensitive to PD98059 18 and U0126 20 , 21 , two inhibitors that block the ERK5 pathway 37 . !en, ERK5 could also partly mediate berberine e"ects, although independently of AMPK (Fig. 6). DCA decreases cholesterol plasma levels in several animal models and humans 5, 13 –15 . In this study, we have mainly used high (10 mM) DCA concentrations for acute responses and “physiological” concentrations (1 to

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 8 www.nature.com/scientificreports/

Figure 6. AMPK did not regulate DCA-induced LDLR expression and LDL uptake. Two di"erent hematopoietic cell lines were treated with 5 mM metformin for 24 h and LDLR mRNA ( A), protein ( B) and LDL uptake ( C) were analyzed. ( D) HepG2-C3A cells were treated as in ( A) and LDL uptake was measured. ( E) HCT116 cells were transfected with 2 small interference RNA (siRNA) for AMPK α or with control siRNA and treated with 20 mM DCA for 6 h before mRNA analysis.

5 mM) for chronic treatments. !ese latter values are in the range of those used in DCA-treated patients 11 , 12 . Michelakis et al . gave 50 mg/Kg/day of DCA to patients 12 . On average, this amount of DCA should give a blood concentration of 4.6 mM, i.e. by considering 70 Kg/patient and a total of 5 L of blood. However, DCA plasma con- centration was around 0.4 mM 11 , 12 and the ultimate destination of the DCA that was not in blood was unknown.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 9 www.nature.com/scientificreports/

Figure 7. DCA required the transcription factor MEF2 to target LDLR promoter. ( A) OCI-AML3 cells were transfected with 40 nM siRNA control or with 20 nM siRNA for each MEF2A and MEF2C (siMEF2). Twenty- four hours later cells were incubated for 24 h with 10 mM DCA. mRNA expression was analyzed by qPCR and represented as the % of mRNA compared to cells transfected with the empty vector. ( B) OCI-AML3 cells were incubated for 72 h with 10 mM DCA. Cells were prepared for ChIP analysis using an antibody against H3 acetylation on lysine 27. Acetylation was revealed at di"erent points of the LDLR promoter by using speci$c oligonucleotides. Bar graphs represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test compared to empty vector transfected cells (control).

Perhaps the liver processes DCA very fast as suggested by our results using primary hepatocytes (Fig. 2). !ere was an attempt to use DCA for treating hypercholesterolemia using similar DCA doses to ours in vivo 15 . DCA reduced circulating cholesterol levels in two patients through a mechanism involving a reduction in LDL cho- lesterol 15 , although both patients initially showed low LDLR surface activity. However, DCA was halted due to its neuropathological e"ects 15 and this precluded its use to treat high cholesterol. !ese pathological e"ects have been observed in other clinical contexts, e.g. lactic acidosis and stroke-like episodes (MELAS) 43 . At the beginning of their DCA treatment, patients quickly eliminate DCA 11 , 12 . In this manuscript, we have observed that in pri- mary hepatocytes fresh DCA should be daily added to media to keep physiological e"ects (Fig. 2). As expected, hepatocytes probably metabolize DCA faster than other cell types 6. A#er several weeks of treatment, DCA plasma concentrations reach a plateau around 0.4 mM 11 , 12 , and during this phase neuropathology appears 11 . From a phar- maceutical point of view, it would be desirable to use drugs with the LDLR-stimulating e"ect of DCA, but without its neuropathological e"ects. !e uncovering of ERK5/MEF2 pathway as a regulator of LDLR expression opens an interesting pharmacological possibility.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 10 www.nature.com/scientificreports/

Figure 8. DCA induced LDLRAP1 expression. ( A) Two hepatic cell lines or primary hepatocytes were treated with DCA as in Figs 1 and 2 and LDLRAP1 mRNA was analyzed. ( B) OCI-AML3 cells (le#) and primary cells from a BCL patient (BCL-P2; right) were treated with 5 mM DCA or grown in OXPHOS medium for 2 weeks and LDLRAP1 mRNA was measured. ( C) Primary hepatocytes were transfected as in Fig. 5 and expression of LDLRAP1 was analyzed by q-PCR. !e bar graphs represent means ± SD of 3 independent experiments performed in triplicate; *p < 0.05, ** p < 0.01, ***p < 0.005 student t-test or one-way ANOVA with post-hoc Tukey test.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 11 www.nature.com/scientificreports/

‡–Š‘†• –Š‹ ƒŽ•–ƒ–‡‡–Ǥ Experimental procedures were conducted according to the European guidelines for animal welfare (2010/63/EU). Protocols were approved by the Animal Care and Use Committee “Languedoc- Roussillon” (approval number: CEEA-LR-12163). The use of human specimens for scientific purposes was approved by the French National Ethics Committee. All methods were carried out in accordance with the approved guidelines and regulations of this committee. Written informed consent was obtained from each patient prior to surgery.

In vivo ‘—•‡‡š’‡”‹‡–•Ǥ In vivo experiments were carried out using 6 to 8 weeks/old male NSG mice. Mice were bred and housed in pathogen-free conditions in the animal facility of the European Institute of Oncology–Italian Foundation for Cancer Research (FIRC), Institute of Molecular Oncology (Milan, Italy). For engra#ment of human cells, 1 million AML cells were injected intravenously (i.v.) through the lateral tail vein in non-irradiated mice. NSG mice with established human AML tumors (day 80 post-gra#) were treated with DCA (50 mg/kg, 1 dose/day by gavage, starting at day 1 for 16 consecutive days). Human tumor AML cells gather in mouse spleen and bone marrow, hence we isolated mRNA from these organs. We used human-speci$c primers to visualize expression of human LDLR mRNA. In a di"erent experiment B6 wt mice were treated with a daily single dose of DCA (50 mg/kg/day) intraperitoneally and mouse LDLR mRNA was analyzed in spleen and liver at di"erent time points.

‡ŽŽŽ‹‡•ƒ† —Ž–—”‡ ‘†‹–‹‘•Ǥ !e leukemic human cell lines T Jurkat TAg and OCI-AML3 were grown in RPMI 1640–Glutamax (GIBCO) supplemented with 5% (Jurkat) or 10% (OCI-AML3) FBS 10 , 24 . Primary cells from a lymphoma B cell patient (BCL-P2) were grown in the same medium with 10% FBS. In certain exper- iments cells were grown in RPMI 1640 without glucose (GIBCO 11879) with the addition of 2 mM glutamine and 10 mM galactose (OXPHOS medium). !e Jurkat TAg cells carry the SV40 large T Ag to facilitate cell trans- fection. HepG2-C3A and HuH7 cells were grown in MEM and DMEM respectively supplemented with FBS, sodium pyruvate, glutamine, penicillin and streptomycin. !e HCT116 human colon cancer cells were cultured in low glucose (5 mM) DMEM medium supplemented with 10% FBS. Cellular con&uence during experiments was between 80–85%.

—ƒŽ‹˜‡”•ƒ’Ž‡•ƒ†’”‡’ƒ”ƒ–‹‘‘ˆ • —Ž–—”‡•Ǥ Liver samples were obtained from liver resections performed in adult patients for medical reasons. Human hepatocytes isolation and culture were per- formed as described previously 44 . Brie&y, a#er liver perfusion, hepatocytes were counted and cell viability was assessed by trypan blue exclusion test. A suspension of 1 × 10 6 cells/mL per well was added in 12-well plates pre-coated with type I collagen (Beckton Dickinson) and cells were allowed to attach for 12 h. !en, the super- natant containing dead cells and debris was carefully removed and replaced with 1 mL of serum-free long-term culture medium (Lanford medium, LNF). !e number of con&uent attached cells was estimated at ~1.5 × 10 5 cells/cm 2.

‡ƒ‰‡–•ƒ†ƒ–‹„‘†‹‡•Ǥ DCA was from Santa Cruz Technologies. Galactose and glutamine were from GIBCO. Human anti-LDLR-PE and IgG were from BD Biosciences and 7AAD from Beckman. !e MEK5 inhib- itor BIX02189 and the ERK5 inhibitor XMD8–92 were from Selleck. RIPA bu"er to prepare protein extracts was from Euromedex. !e complete protease inhibitor cocktail (Complete EDTA-free) and the phosphatase inhibitor cocktail (PhosSTOP) were from Roche. ERK5 and MEF2A/C antibodies were from Cell Signaling Technology and Abcam respectively. !e antibody against β-Actin and HRP-labeled secondary antibodies were from Sigma.

”ƒ•‹‡––”ƒ•ˆ‡ –‹‘Ǥ Jurkat cells in logarithmic growth phase were transfected with the indicated amounts of plasmid by electroporation 45 , 46 . In each experiment, cells were transfected with the same total amount of DNA by supplementing with empty vector. Cells were incubated for 10 min at RT with the DNA mix and elec- troporated using the Gene Pulser Xcell ™ Electroporation system (Bio-Rad) at 260 mV, 960 mF in 400 µl of RPMI 1640. Expression of the di"erent proteins was con$rmed by western blot. !e transfection e*ciency in Jurkat TAg cells is between 60 and 80%. OC-AML-3 cells were transfected using Amaxa TM D-Nucleofector TM Lonza Kit according to manufactured protocol. In HuH7 and HCT116 cells, transfection of 30–50 nM siRNAs was carried out using Lipofectamine RNAiMAX (Invitrogen) in Opti-MEM (Invitrogen), according to the manufacturer’s instructions. Primary hepatocytes were transfected twice at days 1 and 3 post-seeding. Cells were harvested 48 to 96 h post-transfection.

Žƒ•‹†•Ǥ The expression vectors for ERK5, the pSUPER expression vector for GFP alone or GFP plus shERK5 and the pSiren-retroQ-puro (BD Biosciences) retroviral vectors for shERK5 and control have been pre- viously described 45 . Control, MEF2A and C and ERK5 siRNA were ON-TARGETplus SMARTpools (mixture of 4 siRNA) were from Dharmacon.

‘—–‹‰ƒ††‡–‡”‹ƒ–‹‘‘ˆ ‡ŽŽ˜‹ƒ„‹Ž‹–›Ǥ Cell number, viability and cell death was analyzed with the Muse Cell Analyzer (Millipore) by incubating cells with Muse Count & Viability and Annexin V and Dead Cell kits respectively, following manufacturer’s instructions.

‡ƒ•—”‡‡–Ǥ Cells lines were plated at 300,000 cells/ml and treated with DCA for the indicated times, harvested and counted to perform further analysis. To evaluate ROS levels, we labeled cells with CellROX ® Deep Red Reagent or with CH-H2DCFDA (Life Technologies) for 30 minutes and analyzed them by FACs fol- lowing manufacturer’s instructions.

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 12 www.nature.com/scientificreports/

Ǧƒ†•‡“—‡ ‹‰Ǥ Total RNA was extracted using NucleoSpin RNA isolation col- umns (Macherey-Nagel), reverse transcription was carried out using iScript ™ cDNA Synthesis Kit (Biorad). Quantitative PCR was performed with KAPA SYBR Green qPCR SuperMix (Cliniscience) and a CFX Connect ™ Real-Time qPCR machine (Biorad) with LDLR, LDLRAP1, ERK5 and actin primers (Supplemented Fig. 7). All samples were normalized to β-actin mRNA levels. Results are expressed relative to control values arbitrarily set at 100.

—‘„Ž‘––‹‰Ǥ Protein analysis by immunoblotting was performed essentially as previously described 45 . Brie&y, samples were collected, washed out with PBS and lysed with RIPA bu"er. Protein concentration was deter- mined by BCA assay (Pierce) before electrophoresis in 4–15% TGX gels (BioRad) and equal amount of protein was loaded in each well. Protein transfer was performed in TransTurbo system (BioRad) in PVDF membranes. A#er blocking for 1 h with 5% non-fat milk, membranes were incubated overnight at 4 °C in agitation with pri- mary antibodies, washed three times with PBS-Tween 0,1% and incubated with the appropriate HRP-labeled secondary antibody for 1 h. Membranes were washed out three times with PBS-Tween 0,1% and developed with Substrat HRP Immobilon Western (Millipore). Band quanti$cation was performed using the “ImageLab” so#- ware from BioRad and represented as the ratio between the protein of interest and a control protein i.e. actin. !e value of 1 is arbitrarily given to control cells. One blot representative of several experiments is shown.

 –ƒ‡Ǥ A#er treatment cells were incubated with BODIPY FL LDL (Invitrogen) in PBS with 2% FBS and incubated at 37 °C for 30 min. Cells were then washed and suspended in 200–250 µl PBS 2% FBS and analyzed using a Gallios &ow cytometer (Beckman) and the Kaluza so#ware.

Ž‘™›–‘‡–”›Ǥ Brie&y, 1 × 10 6 cells were stained with antibody in PBS with 2% FBS and incubated at 37 °C for 30 min. Cells were then washed and suspended in 200–250 µl PBS 2% FBS and staining was analyzed using a Gallios &ow cytometer (Beckman) and the Kaluza so#ware.

Š ƒƒŽ›•‹•Ǥ OCI-AML3 cells were treated with 10 mM DCA for 72 h. Ten million cells were centrifuged (5 min; 1200 rpm) and the pellet was washed two times in 1X phosphate-bu"ered saline (PBS) at room tempera- ture and suspended in 10 mL of 1X PBS. Cells were $xed in 1% paraformaldehyde (Electron Microscopy Sciences) at room temperature for 5 min. Fixation wa lysed in 1 ml of cell lysis bu"er (5 mM PIPES, 85 mM KCl, 0.5% NP40, Na Butyrate 10 mM + 2X protease inhibitor cocktail (Halt ™ Protease Inhibitor Cocktail, EDTA-Free (100X), !ermo$scher)) at 0 °C for 10 min. Nuclei were recovered by centrifugation (10 min, 5000 rpm) at 4 °C and lysed in 250 µl nuclei lysis bu"er (50 mM Tris-HCl pH 7.5, 1% SDS, 10 mM EDTA, Na Butyrate 10 mM + Halt ™ Protease Inhibitor Cocktail (3X)) at 4 °C for at least 2 hours. 250 µl of each sample were then sonicated 2 times for 5 min (30 s on/o") at 4 ◦C using a Bioruptor (Diagenode). A#er sonication, absorbances at 280 nm (A280) of 1/100 diluted samples were measured and A280nm and was adjusted to 0.133 with nuclei lysis bu"er. One hun- dred microliter were used for ChIP experiments in a $nal volume of 1 ml. Samples were incubated under gentle agitation at 4 °C overnight in the presence of 3 µg of either a speci$c antibody or a negative control. Antibodies (anti-K27Ac Ab4729 (Abcam) and negative control IgG (Diagenode)) were previously bound to DYNA Beads Protein G Novex (Life Technology) according to the supplier’s recommendations. Dynabeads-bound immu- noprecipitates were sequentially washed once with a low salt bu"er (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% triton, 0.1% SDS, 1 mM EDTA, 1 mM Na Butyrate + Halt ™ Protease Inhibitor Cocktail (1X)), a high-salt bu"er (50 mM Tris-HCl pH 7.5, 500 mM NaCl, 1% triton, 0.1% SDS, 1 mM EDTA, 1 mM Na Butyrate) and a LiCl-containing bu"er (20 mM Tris-HCl pH 7.5, 250 mM LiCl, 1% NP40, 1% Na deoxycholate, 1 mM EDTA, 1 mM Na Butyrate) and, then, twice with a TE bu"er (10 mM Tris-HCl pH 7.5, 1 mM EDTA, Tween 20 0.02%). Samples were then eluted in 250 µL of elution bu"er (100 mM NaHCO3, 1% SDS), and DNA-protein complexes were incubated at 65 °C for 5 hours to reverse crosslinks. Samples were then treated with 100 mg/ml proteinase K and 100 mg/ml RNAse A at 45 °C for 2 hours to digest proteins and contaminating RNA. DNA was puri$ed with an extraction kit (NucleoSpin Gel and PCR clean-up, Macherey-Nagel) according to the manufacturer’s recom- mendations and qPCR analysis was performed using the Roche LightCycler 480 real-time PCR system. !e data were normalized with inputs taken from samples before the immunoprecipitation and treated under the same conditions. !e primers used to amplify various regions of LDLR gene promoter.

–ƒ–‹•–‹ ƒŽƒƒŽ›•‹•Ǥ !e statistical analysis of the di"erence between means of paired samples was per- formed using the paired t test. Analysis of multiple comparisons with single control was performed using one-way ANOVA with post-hoc tukey test. !e results are given as the con$dence interval ( *p < 0.05, ** p < 0.01, *** p < 0.005). All the experiments described in the $gures with a quantitative analysis have been performed at least three times in duplicate. Other experiments were performed three times with similar results.

ƒ–ƒƒ˜ƒ‹Žƒ„‹Ž‹–›Ǥ All data are available upon request. ‡ˆ‡”‡ ‡• 1. Moza"arian, D. et al . Executive Summary: Heart Disease and Stro@e Statistics-2016 Update: A Zeport From the American Heart Association. Circulation 133, 447–454 (2016). 2. Goede@e, L. et al . MicroZNA-148a regulates LDL receptor and ABCA1 expression to control circulating lipoprotein levels. Nat Med 21 , 1280–1289 (2015). 3. Goldstein, J. L. & Brown, M. S. !e LDL receptor. Arteriosclerosis, thrombosis, and vascular biology 29 , 431–438 (2009). 4. Go, G. W. & Mani, A. Low-density lipoprotein receptor (LDLZ) family orchestrates cholesterol homeostasis. !e Yale journal of biology and medicine 85 , 19–28 (2012). 5. Stacpoole, P. W. & Greene, Y. J. Dichloroacetate. Diabetes care 15 , 785–791 (1992).

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 13 www.nature.com/scientificreports/

6. Stacpoole, P. W. !e pharmacology of dichloroacetate. Metabolism: clinical and experimental 38 , 1124–1144 (1989). 7. Michela@is, E. D., Webster, L. & Mac@ey, J. Z. Dichloroacetate (DCA) as a potential metabolic-targeting therapy for cancer. Br J Cancer 99 , 989–994 (2008). 8. Charni, S. et al . Oxidative phosphorylation induces de novo expression of the MHC class I in tumor cells through the EZ[5 pathway. J Immunol 185, 3498–3503 (2010). 9. Allende-Vega, N. et al . !e presence of wild type p53 in hematological cancers improves the e*cacy of combinational therapy targeting metabolism. Oncotarget 6, 19228–19245 (2015). 10. [han, A. U. et al . Human Leu@emic Cells performing Oxidative Phosphorylation (OXPHOS) Generate an Antioxidant Zesponse Independently of Zeactive Oxygen species (ZOS) Production. EBioMedicine 3, 43–53 (2016). 11. Stacpoole, P. W., [urtz, T. L., Han, Z. & Langaee, T. Zole of dichloroacetate in the treatment of genetic mitochondrial diseases. Adv Drug Deliv "ev 60 , 1478–1487 (2008). 12. Michela@is, E. D. et al . Metabolic modulation of glioblastoma with dichloroacetate. Sci Transl Med 2, 31ra34 (2010). 13. Zibes, G., Valette, G. & Loubatieres-Mariani, M. M. Metabolic e"ects of sodium dichloroacetate in normal and diabetic dogs. Diabetes 28 , 852–857 (1979). 14. Stacpoole, P. W., Moore, G. W. & [ornhauser, D. M. Metabolic e"ects of dichloroacetate in patients with diabetes mellitus and hyperlipoproteinemia. N Engl J Med 298, 526–530 (1978). 15. Moore, G. W. et al . Zeduction of serum cholesterol in two patients with homozygous familial hypercholesterolemia by dichloroacetate. Atherosclerosis 33 , 285–293 (1979). 16. Li, Y. et al . AMP[ phosphorylates and inhibits SZEBP activity to attenuate hepatic steatosis and atherosclerosis in diet-induced insulin-resistant mice. Cell Metab 13 , 376–388 (2011). 17. Liu, S. et al . AICAZ-Induced Activation of AMP[ Inhibits TSH/SZEBP-2/HMGCZ Pathway in Liver. PLoS One 10 , e0124951 (2015). 18. Brusq, J. M. et al . Inhibition of lipid synthesis through activation of AMP @inase: an additional mechanism for the hypolipidemic e"ects of berberine. Journal of lipid research 47 , 1281–1288 (2006). 19. Wang, Y. X. et al . Synthesis and structure-activity relationship of berberine analogues in LDLZ up-regulation and AMP[ activation. Bioorganic & medicinal chemistry 20 , 6552–6558 (2012). 20. [ong, W. et al . Berberine is a novel cholesterol-lowering drug wor@ing through a unique mechanism distinct from statins. Nat Med 10 , 1344–1351 (2004). 21. Abidi, P., Zhou, Y., Jiang, J. D. & Liu, J. Extracellular signal-regulated @inase-dependent stabilization of hepatic low-density lipoprotein receptor mZNA by herbal medicine berberine. Arteriosclerosis, thrombosis, and vascular biology 25 , 2170–2176 (2005). 22. Zheng, X. et al . !e e"ects of chylomicron remnants enriched in n-3 or n-6 polyunsaturated fatty acids on the transcription of genes regulating their upta@e and metabolism by the liver: in&uence of cellular oxidative state. Free "adic Biol Med 32 , 1123–1131 (2002). 23. Zathore, M. G. et al . !e NF-@appaB member p65 controls glutamine metabolism through miZ-23a. Int J Biochem Cell Biol 44 , 1448–1456 (2012). 24. Lopez-Zoyuela, N. et al . Extracellular-signal-regulated @inase 5 modulates the antioxidant response by transcriptionally controlling Sirtuin 1 expression in leu@emic cells. Int J Biochem Cell Biol 53 , 253–261 (2014). 25. Villalba, M. et al . Chemical metabolic inhibitors for the treatment of blood-borne cancers. Anti-cancer agents in medicinal chemistry 14 , 223–232 (2014). 26. Villalba, M. et al . From tumor cell metabolism to tumor immune escape. Int J Biochem Cell Biol 45 , 106–113 (2013). 27. [ato, Y. et al . BM[1/EZ[5 regulates serum-induced early gene expression through transcription factor MEF2C. EMBO Journal 16 , 7054–7066 (1997). 28. [ato, Y. et al . Big mitogen-activated @inase regulates multiple members of the MEF2 protein family. Journal of Biological Chemistry 275, 18534–18540 (2000). 29. [asler, H. G., Victoria, J., Duramad, O. & Winoto, A. EZ[5 is a novel type of mitogen-activated protein @inase containing a transcriptional activation domain. Mol Cell Biol 20 , 8382–8389 (2000). 30. Yang, C. C., Ornats@y, O. I., McDermott, J. C., Cruz, T. F. & Prody, C. A. Interaction of myocyte enhancer factor 2 (MEF2) with a mitogen-activated protein @inase, EZ[5/BM[1. Nucleic Acids "es 26 , 4771–4777 (1998). 31. Catalán, E. et al . MHC-I modulation due to metabolic changes regulates tumor sensitivity to CTL and N[ cells. Oncoimmunology 4, e985924 (2015). 32. Zeitzer, L. J., Wice, B. M. & [ennell, D. Evidence that glutamine, not sugar, is the major energy source for cultured HeLa cells. J Biol Chem 254, 2669–2676 (1979). 33. Zossignol, Z. et al . Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. Cancer "es 64 , 985–993 (2004). 34. Stacpoole, P. W., Harwood, H. J. Jr. & Varnado, C. E. Zegulation of rat liver hydroxymethylglutaryl coenzyme A reductase by a new class of noncompetitive inhibitors. Effects of dichloroacetate and related carboxylic acids on enzyme activity. J Clin Invest 72 , 1575–1585 (1983). 35. Drew, B. A., Burow, M. E. & Bec@man, B. S. ME[5/EZ[5 pathway: the $rst $#een years. Biochim Biophys Acta 1825, 37–48 (2012). 36. Allende-Vega, N. et al . p53 is activated in response to disruption of the pre-mZNA splicing machinery. Oncogene 32 , 1–14 (2013). 37. [ama@ura, S., Moriguchi, T. & Nishida, E. Activation of the protein @inase EZ[5/BM[1 by receptor tyrosine @inases. Identi$cation and characterization of a signaling pathway to the nucleus. Journal of Biological Chemistry 274, 26563–26571 (1999). 38. Allende-Vega, N., Dias, S., Milne, D. & Mee@, D. Phosphorylation of the acidic domain of Mdm2 by protein @inase C[2. Mol Cell Biochem 274, 85–90 (2005). 39. Wang, Z. et al . Combinatorial patterns of histone acetylations and methylations in the human genome. Nat Genet 40 , 897–903 (2008). 40. Fellin, Z., Arca, M., Zuliani, G., Calandra, S. & Bertolini, S. !e history of Autosomal Zecessive Hypercholesterolemia (AZH). From clinical observations to gene identi$cation. Gene 555, 23–32 (2015). 41. Parhofer, [. G. Interaction between Glucose and Lipid Metabolism: More than Diabetic Dyslipidemia. Diabetes & metabolism journal 39 , 353–362 (2015). 42. Miranda, M. X. et al . !e Sirt1 activator SZT3025 provides atheroprotection in Apoe −/− mice by reducing hepatic Pcs@9 secretion and enhancing Ldlr expression. European heart journal 36 , 51–59 (2015). 43. [aufmann, P. et al . Dichloroacetate causes toxic neuropathy in MELAS: a randomized, controlled clinical trial. Neurology 66 , 324–330 (2006). 44. Pichard, L. et al . Human hepatocyte culture. Methods Mol Biol 320, 283–293 (2006). 45. Garaude, J. et al . EZ[5 activates NF-@appaB in leu@emic T cells and is essential for their growth in vivo . J Immunol 177, 7607–7617 (2006). 46. Garaude, J. et al . SUMOylation regulates the transcriptional activity of JunB in T lymphocytes. J Immunol 180, 5983–5990 (2008).  ‘™Ž‡†‰‡‡–• FACs analysis was performed at the platform Montpellier Rio Imaging (MRI). !e collection of clincial data and samples (HEMODIAG_2020) at the CHRU Montpellier was supported by funding from Région Languedoc

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 14 www.nature.com/scientificreports/

Roussillon. All our funders are public or charitable organizations. !is work was supported by a grant from Fondation de France (0057921), fellowship from the Higher Education Commission, Pakistan (AK) and fellowships from the Ministère de l’Enseignement Supérieur et de la Recherche (MESR) (DNV). —–Š‘”‘–”‹„—–‹‘• A.K., N.A.-V., D.G., S.G.-C., C.G., D.-N.V., S.B., S.O., G.T., M.B., F.B. and I.L.-M. perform experiments. F.B., J.-M.V., I.J., L.F., C.-H.L., J.H., M.D.-C. and M.V. wrote the main manuscript text. All authors reviewed the manuscript. ††‹–‹‘ƒŽ ˆ‘”ƒ–‹‘ Supplementary information accompanies this paper at doi: 10.1038/s41598-017-10339-5 Competing Interests: !e authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional a*liations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. !e images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

© !e Author(s) 2017

SCIENTIFIC REPORTS ȁͽǣ 10654 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷͽǦͷͶ͹͹ͿǦͻ 15 ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! |ÅÅ! ) ) ) ) ) )

!'".1#/!lZ!! !

!>4F=8K!?F!E8L45GD?KE!4<<86L!8PHJ8KK?GF!GJGM=>!#/)h!4F7! 78H8F7?F=!GF!Hhf!KL4LMK! ) I<@087>632+/=?7183+H+(:1;?/=53?J+ ) ) K&:+ ?67>+ >?@2D+ ?6/?+ &+ >6/=32+ 1;G/@?6;=>67+ 7:A;8A32+ 7:+ <8/::7:5+ /:2+ 3C31@?7;:+;4+?63+3C<3=793:?>E++7:18@27:5+<=3+@>32+7:+?63+>?@2DE+<=3+<3=4;=97:5+(.)%(+L+ ) ) ) ) ) ) ) ) ) ) )

! |ÅÇ! www.impactjournals.com/oncotarget/ Oncotarget, 2018, Vol. 9, (No. 1), pp: 1114-1129

Research Paper Changes in metabolism affect expression of ABC transporters through ERK5 and depending on p53 status

Sana Belkahla 1, Abrar Ul Haq Khan 1,2 , Delphine Gitenay 1,2 , Catherine Alexia 1,2 , Claire Gondeau 1,2,3 , Dang-Nghiem Vo 1, Stefania Orecchioni 4, Giovanna Talarico 4, Francesco Bertolini 4, Guillaume Cartron 5, Javier Hernandez 1, Martine Daujat- Chavanieu 1,2 , Nerea Allende-Vega 1,2,* and Martin Villalba Gonzalez 1,2,* 1Department of Lymphocyte Differentiation, Tolerance and Metabolism: Basis for Immunotherapy, Institut De Médecine Régénératrice Et Biothérapie (IRMB), INSERM, Univ De Montpellier, Montpellier, France 2Department of Lymphocyte Differentiation, Tolerance and Metabolism: Basis for Immunotherapy, Institut De Médecine Régénératrice Et Biothérapie (IRMB), CHU Montpellier, Montpellier, France 3Département d'Hépato-gastroentérologie A, Hôpital Saint Eloi, CHU Montpellier, Montpellier, France 4Department of Oncology and Hemato-Oncology, European Institute of Oncology, Milan, Italy 5Département d'Hématologie Clinique, CHU Montpellier, Université Montpellier I, Montpellier, France *These two authors share senior authorship Correspondence to: Martin Villalba Gonzalez, email: [email protected] Nerea Allende-Vega, email: [email protected] Keywords: ABC transporter; p53; ERK5; oxidative phosphorylation (OXPHOS); dichloroacetate (DCA) Received: June 10, 2017 Accepted: December 05, 2017 Published: December 14, 2017 Copyright: Belkahla et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

ABSTRACT

Changes in metabolism require the efflux and influx of a diverse variety of metabolites. The ABC superfamily of transporters regulates the exchange of hundreds of substrates through the impermeable cell membrane. We show here that a metabolic switch to oxidative phosphorylation (OXPHOS), either by treating cells with dichloroacetate (DCA) or by changing the available substrates, reduced expression of ABCB1, ABCC1, ABCC5 and ABCG2 in wild-type p53-expressing cells. This metabolic change reduced histone changes associated to active promoters. Notably, DCA also inhibited expression of these genes in two animal models in vivo . In contrast, OXPHOS increased the expression of the same transporters in mutated (mut) or null p53- expressing cells. ABC transporters control the export of drugs from cancer cells and render tumors resistant to chemotherapy, playing an important role in multiple drug resistance (MDR). Wtp53 cells forced to perform OXPHOS showed impaired drug clearance. In contrast mutp53 cells increased drug clearance when performing OXPHOS. ABC transporter promoters contain binding sites for the transcription factors MEF2, NRF1 and NRF2 that are targets of the MAPK ERK5. OXPHOS induced expression of the MAPK ERK5. Decreasing ERK5 levels in wtp53 cells increased ABC expression whereas it inhibited expression in mutp53 cells. Our results showed that the ERK5/ MEF2 pathway controlled ABC expression depending on p53 status.

INTRODUCTION molecules (anabolism). Metabolism requires the intake of substrates and the outtake of metabolites through the The main purpose of cell metabolism is highly impermeable cell plasma membrane and multiple the conversion of metabolic substrates to energy transporters with differences in specificity have been (catabolism) or to build blocks for generating new described. One of the biggest and better-conserved www.impactjournals.com/oncotarget 1114 Oncotarget families of transporters is the ATP-binding cassette (ABC ABCB11, ABCC1, ABCC4, ABCC12 and ABCG1, have transporters [1, 2]). The ABC superfamily has 48 members MEF2A binding sites that have been validated by ChIP that have been classified into 7 groups (subfamilies A to (http://genome.ucsc.edu/). Most of these transporters G). One of their main biological functions is the transport also have NRF1 or NRF2 binding sites, i.e. ABCB8, of lipids [3], essential molecules for cell metabolism. ABCB9, ABCB10, ABCB11, ABCC1, ABCC3, ABCC5, Although it is conceivable that changes in metabolism ABCC4, ABCC10, ABCC12, ABCC13, ABCG1, which regulate ABC expression, this remains uninvestigated. have also been validated by ChIP (http://genome.ucsc. ABC transporters are well known for their role in edu/). This information is based on ChIP-seq experiments multiple drug resistance (MDR) because many of them performed by the ENCODE consortium[25]. Hence, we export anticancer drugs and render tumors resistant to investigated here if metabolism, through ERK5, controls chemotherapy [4]. The tumor suppressor gene p53 is ABC expression and the role of p53 status. an important regulator of metabolic homeostasis by promoting OXPHOS and inhibiting glycolysis [5]. It RESULTS induces the expression of different metabolic genes, such as cytochrome c oxidase 2 ( SCO2), glutaminase 2 ( GLS2 ), OXPHOS regulated ABC transporters in AML p53 up-regulated modulator of apoptosis ( PUMA ), glucose cell lines transporter 1 and 4 ( GLUT1 and GLUT4 ) and TP53- induced glycolysis and apoptosis regulator ( TIGAR ) [6, We induced OXPHOS either by inhibiting pyruvate 7]. More than half of all human tumors harbor mutations dehydrogenase kinase (PDK) with dichloroacetate (DCA; in the p53 gene. Most of these mutations abrogate its [7, 14, 16, 26–28]) or by incubating cells in the absence DNA binding and transactivation activity [8], but others of glucose. When glucose is no longer available, cells use can provide mutant p53 (mutp53) with gain-of-function alternative energy substrates, such as glutamine (Gln). (GOF) activity that is dependent on its de novo ability Gln oxidation, or glutaminolysis, generates ATP through to regulate directly or indirectly gene expression [9, 10]. OXPHOS [29, 30] and this pathway is functional in Furthermore, certain p53 mutants enhance drug resistance leukemic cells [15, 31]. We called OXPHOS medium a in liver cancer cells and B-CLL cells [11, 12]. glucose-free medium supplemented with 10 mM galactose Several ABC promoters contain consensus p53 and 4 mM Gln [14, 15, 31]. OXPHOS induces expression binding sequences [4]. The direct regulation of ABC of both wild type (wt) and mutp53 depending on the promoters by p53 is controversial and depends on the cellular context [7]. We used DCA concentrations on the architecture of the promoter, the nature of p53 (mutant range of that found in plasma of DCA-treated patients or wild type), the presence of other p53 family members [32, 33] and tested the effect on the expression of 4 ABC and variations cell- and tissue-specific [4, 13]. We have transporters, ABCB1, ABCC1, ABCC5 and ABCG2, observed that inducing OXPHOS in leukemic cells involved in MDR, in 3 AML cell lines with different p53 sensitizes wtp53-expressing cells to genotoxic drugs status: OCI-AML3 cells express wt p53, HL60 are p53 such as doxorubicin and vincristine [7]. In contrast, cells null and NB4 are mutp53 [7]. DCA decreased by half the carrying mutp53 were more resistant to these anticancer mRNA of the different ABC transporters on OCI-AML3 drugs. cells and this correlated with protein reduction in the 2 OXPHOS promotes the expression of the MAPK examined transporters (Figure 1A). In contrast, in HL- extracellular signal-regulated kinase-5 (ERK5), 60 and NB4 cells, DCA increased mRNA and protein which regulates the choice of metabolic substrates in expression (Figure 1B and 1C). hematopoietic cells [14–19]. ERK5 induces activation Consequently with our observations in cell lines, of several members of the MEF2 family of transcription we observed that DCA decreased ABC mRNA and factors by different mechanisms. It mediates direct serine protein expression in primary cells derived from a B-cell and threonine phosphorylation of MEF2A, C and D [20, lymphoma (BCL) patient with wtp53 (Figure 1D). On 21]. It activates MEF2A and D by direct interaction the contrary DCA increased expression of ABCC5 and because ERK5 serves as a MEF2 coactivator through ABCG2 in two mutp53 patients and ABCC1 in one patient. its signal-dependent direct association with the MEF2 In contrast, ABCB1 expression was unchanged (Figure MADS domain [22, 23]. The effect of activated ERK5 1E). Due to lack of enough patient samples, we could not on MEF2A-dependent transcription requires ERK5 analyze protein expression. kinase activity. MEF2 mediates several ERK5 effects on Moreover, OCI-AML3 cells growing in OXPHOS metabolism, including NRF2 activation [15, 16]. This medium displayed reduced mRNA expression of transcription factor and the related NRF1 are master the above-mentioned ABC transporters (Figure 2A). regulators of metabolism that are coordinated with MEF2 Conversely, mutant p53-expressing NB4 cells increased family to induce expression of metabolic genes [24]. The ABC expression in the same medium (Figure 2A). In promoters, or very proximal regions, of several ABC contrast to DCA, OXPHOS medium did not increase genes, i.e. ABCB1, ABCB4, ABCB8, ABCB9, ABCB10, ABC expression in null p53 HL60 cells (Figure 2A). www.impactjournals.com/oncotarget 1115 Oncotarget Figure 1: DCA-induced ABC transporters expression depended on p53 status in leukemic cells. Different hematopoietic cell lines OCI-AML3 wtp53 (A) , HL-60 nullp53 (B), NB4 mutp53 (C) and primary cells from BCL-P2 wtp53 patient (D) or B-CLL mutp53 patients were treated with 1 and 5 mM DCA for 1 week and RNA and plasma protein levels were analyzed by qPCR or FACS respectively. Data represent the % of mRNA compared to control cells. The bar graphs represent means ± SD of 3 independent experiments performed in triplicate; *p<0.05, ** p<0.01, ***p<0.005 student t-test compare to control cells.

www.impactjournals.com/oncotarget 1116 Oncotarget This showed that although DCA and OXPHOS medium To verify that OXPHOS was affecting transcription could have similar effects, they could also induce different of these genes, we assessed histone modifications that outcomes depending on p53 status. This could be related have been linked to ABCC1 promoter activation [11, to the different stimulation periods needed to optimize 12] in OCI-AML3 cells. We found that DCA inhibited changes in cell metabolism. acetylation on H3 K27 on the ABCC1 promoter (Figure

Figure 2: OXPHOS inhibited or stimulated ABC expression in wtp53 or mutp53 cells, respectively. (A) OCI-AML3, NB4 and HL60 cells were grown in OXPHOS medium for 2 weeks and expression of ABC transporters was analyzed as in Figure 1. The bar graphs represent means ± SD of 3 independent experiments performed in triplicate; *p<0.05, ** p<0.01, ***p<0.005 student t-test compare to control cells. (B) Cells were treated with 5 mM DCA for 3 days and processed for ChIP analysis. The enrichment of K27 acetylation on Histone 3 in the ABCC1 promoter was analyzed as described in material and methods by using different primers around the transcription start site (TSS). www.impactjournals.com/oncotarget 1117 Oncotarget 2B). We conclude that DCA inhibited transcription of ABC tumor cells with the genotoxic drug daunorubicin, which transporters on wtp53 cells. is a fluorescent compound. This allows monitoring its intracellular accumulation by FACs. In addition, it is used OXPHOS regulated ABC transporters in hepatic in the clinic to treat leukemia and lymphoma, which are cells at the origin of some cell lines that we have used in our study. Finally, like other anthracyclines, it is exported from The main detoxifying organ is liver. We tested DCA cells by a broad range of transporters [34]. We used OCI- effect in 2 hepatic cell lines expressing wtp53 (HepG2- AML3 cells treated for one week with 1 mM and 5mM C3A) or mutp53 (HuH7). When wtp53 was present, an DCA or growing in OXPHOS medium. Then, cells were acute dose of DCA decreased ABC transporters mRNA further incubated with daunorubicin and drug clearance and protein (Figure 3A and 3C). In contrary, and similar was followed by FACs. We found that OXPHOS OCI- to mutp53 leukemia cells, DCA increased ABC mRNA AML3 cells were impaired on decreasing drug levels expression in HuH7 cells and also ABCC1 protein (Figure (Figure 5A and 5C). The effect was mainly observed at 3B and 3C). Surprisingly, ABCB1 protein decreased short time points, but it was still noticeable as much as after treatment suggesting alternative posttranscriptional 24 h after (Figure 5A). Primary wtp53 tumor cells from regulation. Finally, we examined the effect of OXPHOS a BCL patient also showed decreased daunorubicin on ABC transporters expression in non-transformed cells. export (BCL-P2; Figure 5D). As expected, the outcome Primary hepatocytes from non-cancer patients, which were was opposite in cells with alterations in p53. In tumor obviously wtp53, showed the same pattern than cell lines cells not expressing p53 (HL-60) or expressing mutp53 expressing wtp53 (Figure 3D). (NB4) DCA induced a faster drug clearance (Figure 5E) in agreement with an increase on ABC transporters. DCA decreased mRNA of ABC transporters in Furthermore, blocking expression of wtp53 in OCI- vivo AML3 cells by treating them with a small interference RNA against p53 (sip53), which effectively decreased p53 We had observed that DCA increased doxorubicin expression (Supplementary Figure 1), prevented DCA- efficacy in vivo on tumor cells expressing wtp53 [7]. induced decreased daunorubicin clearing (Figure 5B), We engrafted human wtp53 AML primary cells in non- whereas a control siRNA did not show any effect (data REHVHGLDEHWLFVHYHUHFRPELQHGLPPXQRGH¿FLHQW 12' not shown). Therefore, changes in mRNA and/or protein 6&,' LQWHUOHXNLQUHFHSWRUȖFKDLQQXOO 16* PLFH expression of ABC transporters due to metabolic changes as previously described [7, 16]. Mice with established lead to alterations in drug clearance, which depended in tumors (day 80 post-graft) were treated daily with p53 status. DCA. The treatment was not toxic and did not show any To further investigate the role of p53 in ABC notable effect on mice survival [7]. Human tumor AML transporter expression after metabolic changes, we used cells gather in mouse spleen and bone marrow, hence two isogenic colon cancer cell lines (p53 +/+ and p53 -/- we isolated mRNA from these organs and used human- HCT116 cells) that differently respond to DCA treatment specific primers. We observed that DCA significantly [7]. DCA decreased the ABCB1 and ABCC1 mRNAs reduced expression of ABC transporters mRNA (Figure without changing those of ABCC5 and ABCG2 (Figure 4A). 6A). p53 -/- HCT116 cells did not downregulate any ABC Additionally, we assessed the effect of DCA in mRNA after DCA treatment (Figure 6B). These results normal tissues of non-tumor bearing wt mice. A daily from HCT116 +/+ cells suggest that the role of p53 in ABC injection of DCA, for 1 and up to 3 days, also reduced expression is cell type dependent and depends in multiple mouse ABC transporters mRNA in liver and spleen (Figure factors [4, 13]. The absence of DCA effect on HCT116 -/- 4B). The effect largely increased with the number of and of OXPHOS medium in null p53 HL60 cells, suggest doses received. Hence, DCA reduced ABC expression in that p53 is essential to mediate the effect of metabolism multiple wtp53 cell populations in vivo . This could, at least on ABC expression. partially, explain its effect on genotoxic drug treatment in wtp53 tumor cells in vivo [7]. ERK5 was essential for DCA-induced effects

OXPHOS modulated drug clearance from tumor We further investigated the underlying mechanism cells depending on its p53 status of OXPHOS-induced ABC transporters expression. We focused on the role of the ERK5/MEF2 pathway, which A direct functional consequence of modulating is activated in cells performing OXPHOS [14–16, 31]. ABC transporters expression in tumor cells could be However, we could not reduce ERK5 levels in cells changes in their responsiveness to chemotherapeutic performing OXPHOS because this kinase is essential drugs. To investigate if the changes in the expression of for cell survival in this metabolic status [14–16, 31]. We ABC transporters affect drug clearance, we incubated speculated that decreasing ERK5 levels in non-treated www.impactjournals.com/oncotarget 1118 Oncotarget Figure 3: DCA-induced ABC transporters expression depended on p53 status in hepatic cells. (A-B) Hepatic cell lines HepG2-C3A wtp53 and HuH7 mutp53 were treated with 20mM DCA for 24h before analyzing mRNA level as described in Figure 1. (C) Cells were treated with the indicated concentration of DCA for 1 week and protein level was analyzed by FACS. (D) Primary hepatocytes were treated with indicated amount of DCA for 72 hours. The data represent means ± SD (n=3); *p < 0.05, ** p < 0.01, ***p < 0.001 Student's t-test compared to control cells or as depicted in the graphic.

www.impactjournals.com/oncotarget 1119 Oncotarget cells would have the opposite effect than DCA treatment. (Supplementary Figure 2C), decreased mRNA of multiple Reducing expression of ERK5 with a small interference ABC transporters (Figure 7C). In agreement, the ERK5 RNA for ERK5 (siERK5), which reduced ERK5 inhibitor XMD8-92 or the MEK5 inhibitor BIX02189 also levels (Supplementary Figure 2A and 2B), decreased decreased basal ABCB1 protein expression (Figure 7D). and increased ABC transporters mRNA in HuH7 and Finally, overexpression of ERK5 (Supplementary Figure primary hepatocytes, respectively (Figure 7A and 7B). 2C) increased ABC mRNA expression (Figure 7C). This suggested that ERK5 effects, like those of DCA, depended on p53 status. To test this hypothesis, we used DISCUSSION Jurkat cells that express very low levels of mutp53 [35]. DCA or OXPHOS medium induced a small increase in ABC transporters have attracted great attention in ABCB1 protein (Supplementary Figure 3). Consequently recent years due to their role in MDR [4]. But, it should a shERK5, which encodes for a different sequence that the be kept in mind that their main physiological role in previous described siRNA and decreased ERK5 expression eukaryotes is possibly the transport of other substances

Figure 4: DCA induced ABC transcription in vivo . (A) NSG mice were engrafted with primary human AML cells. At day 80 post- graft, they were treated with DCA (n=4) or leave untreated (n=4). At day 140, mRNA from spleen or bone marrow was isolated and human ABCB1, ABCC1, ABCC5 and ABCG2 mRNA expression was quantified by qPCR. (B) B6 wt mice (n=4/5 per group) were treated with a dose of DCA (50 mg/kg) everyday intraperitoneally and mice mRNA in spleen and liver at different times was obtained to analyze ABC transporters. The data represent means ± SD; *p<0.05, ** p<0.01, ***p<0.001 student t-test compare to non treated cells or mice.

www.impactjournals.com/oncotarget 1120 Oncotarget Figure 5: Metabolism-controlled drug outtake depended on p53 status. (A, B, D and E) Different cell lines were treated with 5 mM DCA for 7 days or incubated in OXPHOS medium for 14 days (C). In (B), OCI-AML3 cells were transfected with a siRNA for p53 (sip53) 48 h before treatment. Cells were then incubated with 5 uM daunorubicin and clearance was measured at various time points (10, 30, 60, 120 minutes). Some cells were incubated with 1uM daunorubicin for 24 h. Experiments were repeated at least three times. www.impactjournals.com/oncotarget 1121 Oncotarget out of the cell, e.g. lipid transport [3]. It is somehow be exploited in the clinic to prevent MDR. Since more intuitive that changes in metabolism regulate the transport glycolytic tumors show higher MDR, inducing a metabolic of a variety of molecules. We show here that wtp53 cells, switch should decrease MDR, although as we prove here which comprise the vast majority of non-transformed the p53 status should be taken into account. cells, while performing OXPHOS decrease expression High expression of ABC transporters is often of ABC transporters. During OXPHOS, cells use most observed in drug-naive tumors compared with the tissue of the carbon of metabolic substrates to produce energy of origin and, as previously described, this correlates and CO2. Then, they should avoid releasing metabolites with a more glycolytic metabolism. Hence, constitutive to external media. In contrast, during glycolysis, cells MDR1 or MRP over expression is likely regulated in some generate a number of intermediate metabolites that cells by pathways involved in malignant transformation, can not accumulate and would need to be exported. e.g. metabolic changes [13]. Metabolic drugs such as We hypothesize that this is the physiological interest to DCA could revert tumor metabolism and decrease ABC increase expression of the exporter ABC transporters transporters expression. However, DCA accumulates during high glycolysis. However, tumor cells take both wt and mutp53 protein [7]. If the last induces ABC advantage of this fact to acquire MDR. Transformed expression as suggested in several cell lines by our results cells present a metabolism more orientated to anaerobic and in previous works [4, 13], DCA would finally increase glycolysis than their non-transformed counterparts, which MDR. HSP90 inhibitors, e.g. 17-AAG, degrade mutp53 mainly use respiration/OXPHOS. This should favor tumor by favoring its MDM2-mediated ubiquitination [40], and MDR by increasing ABC transporters activity in wtp53 we have shown that DCA cooperates with 17-AAG to kill expressing cells. In fact, MDR1-dependent chemoresistant mutp53 tumor cells [7]. This could be a possibility to treat cells show an even more glycolytic metabolism that the mutp53 tumors. In summary, it is essential to know p53 sensitive counterparts [36, 37]. Interestingly, hypoxia or status before treatment with metabolic drugs as we have changes in glucose levels increase MDR1 expression by previously shown [7]. hypoxia-inducible factor-1 (HIF-1; [38, 39]). This could ERK5 could use different ways to transcriptionally partly explain the observation that hypoxic tumor cells activate ABC SURPRWHUVHJS1)ț%15)+,) are more chemoresistant [13]. These observations could and MEF2 [13]. ERK5 inhibition induces p53 activation

Figure 6: DCA-induced ABC transporters expression required wtp53. (A and B) Two isogenic colon cancer cell lines (p53 +/+ and p53 -/- HCT116 cells) were treated with 1 and 5 mM DCA for 1 week and RNA was analyzed by qPCR. Data represent the % of mRNA compared to control cells. The bar graphs represent means ± SD of 3 independent experiments performed in triplicate; *p<0.05, ** p<0.01, ***p<0.005 student t-test compare to control cells. www.impactjournals.com/oncotarget 1122 Oncotarget and increases sensitivity to 5-fluorouracil in colon cancer independently of de novo ROS generation through NRF2 cells but the role of ABC was not investigated [41]. We activation [16]. Hence, by activating NRF2, ERK5 could propose that ERK5 activity by regulating metabolism modulate ABC expression. affects p53 effect in ABC promoters. HIF regulates ABC expression [38, 39] and ERK5 Oxidative stress induces expression of ABC regulates in normoxia an array of genes regulated by HIF1 transporters, which is likely mediated by NRF2 [42, 43]. in hypoxia [44]. In fact, ERK5 by modulating metabolism ERK5, through MEF2, generates an antioxidant response regulates genes that are regulated by HIF. This could be

Figure 7: The ERK5 pathway regulated ABC transcription. (A) HuH7 cells were transfected with control siRNA or siERK5. Thirty-six hours later mRNA expression was analyzed by qPCR. (B) Primary human hepatocytes were transfected with control siRNA or siERK5 at day 1 and 3 post seeding. 96 h later mRNA levels were assessed. (C) Jurkat cells were transfected with a small hairpin RNA for ERK5 (shERK5) or ERK5 expression vector. 36h later mRNA expression was analyzed by qPCR. (D) -XUNDWFHOOVZHUHWUHDWHGZLWKȝ0 ;0'RUȝ0%,;IRUKEHIRUHDQDO\]LQJH[SUHVVLRQRI$%&%E\)$&V7KHGDWDUHSUHVHQWPHDQV“6'Q  *p < 0.05, ** p < 0.01, ***p < 0.001 Student's t-test compared to empty vector transfected cells (control).

www.impactjournals.com/oncotarget 1123 Oncotarget the case of ABC  JHQHV ,QGHHG 1)ț% PHGLDWHV$%& cancerous proliferative conditions. For example, in expression by direct binding to ABC promoters [13, 45] coronary artery stent restenosis and pulmonary artery or by regulating intermediate mediators. In particular hypertension (PAH) there is an unwanted proliferation of GXULQJFKDQJHVLQJOXFRVHOHYHOV1)ț%DFWLYDWHV+,) endothelial cells. Notably, inhibition of ABCC4 prevents that targets ABC promoters [39]. Given the fact that ERK5 and reverses pulmonary hypertension in mice [52] and DFWLYDWHV 1)ț% >@ WKLV FRXOG PHGLDWH WKH HIIHFW RI ABCA3 deficiency is related to PAH of the newborn [53]. metabolism on ABC expression. Moreover, ABCG2 clears hypoxia-induced intracellular As previously described, the promoters of several metabolites and protects the heart from pressure overload- ABC genes studied here, i.e. ABCB1 and ABCC1, have induced ventricular dysfunction [54]. Hypoxia-induced MEF2A binding sites. MEF2 mediates a large part of aerobic glycolysis could obviously affect the effectiveness the observed ERK5 effects on metabolism [15, 16, 47]. of drug treatment in these situations. This could be also the Hence, this could also be the case of ABC transporters. case in other situations requesting high proliferation such The existence of multiple ERK5 targets mediating ABC as inflammation or wound healing. expression makes it difficult to elucidate their individual In summary, the effects of metabolism on the relevance. The contribution of each factor could depend, transcription of ABC transporters clearly depend on as discussed for p53 [4, 13], on the architecture of the p53 status. However, sensitization to chemotherapy promoter, the level of expression of each target, the by metabolic drugs does not depend only on ABC presence of other targets and variations cell- and tissue- transcription because pro-apoptotic pathways are also specific. activated that could lead to synergism between metabolic ERK5 inhibition in mutp53 cells decreased ABC drugs and standard chemotherapy. expression (our results) and should decrease MDR and favors chemotherapy as it has been shown [41], although MATERIALS AND METHODS the mechanism was not described. Our results show that metabolism targets multiple ABC promoters, depending on Ethical statement p53 status, which will have a clear effect on drug export. In this context, DCA overcomes resistance to paclitaxel Experimental procedures were conducted according in A549/taxol cells [48]; however, p53 status is unknown to the European guidelines for animal welfare (2010/63/ in this subline although the parental cell line, i.e. A549, EU). Protocols were approved by the Animal Care and is wtp53. Then DCA could decrease expression of ABC Use Committee “Languedoc-Roussillon” (approval transporters in this study as we have observed here. number: CEEA-LR-12163). However, Zhou et al proposed that ABCB1 expression did not change and expression of other ABC members Reagents and antibodies was not analyzed [48]. Authors propose that citric acid accumulation induces apoptosis, but DCA and paclitaxel DCA was from Santa Cruz Technologies. Galactose can also activate alternative pathways leading to cell death. and glutamine were from GIBCO. Human anti-ABCB1, In our case, we have shown that inducing OXPHOS in ABCC1 and control IgG1 were from Miltenyi Biotec and wtp53 cells cooperates with genotoxic drugs to eliminate 7AAD from Beckman. The ERK5 inhibitor XMD8-92 and tumor cells in vitro and in vivo [7]. the MEK5 inhibitor BIX02189 were from Selleck. An interesting clinical approach would be to treat patients whose tumors express wtp53 with DCA Plasmids before standard chemotherapy. We have used here DCA concentrations similar to that used in patients [7, 32, The expression vectors for ERK5, the pSUPER 33]. These concentrations were effective to change expression vector for GFP alone or GFP plus shERK5 have ABC expression and then should sensitize tumor cells to been previously described [46]. Control, ERK5 and p53 chemotherapy as we have observed in mice [7]. In fact siRNA were ON-TARGETplus SMARTpools (mixture of a combinatorial therapy involving DCA in preclinical 4 siRNA) from Dharmacon and has been previously used settings have been proposed with several compounds [7, 47]. ERK5 shRNA (AGCTGCCCTGCTCAAGTCT) such as metformin [49], Nutlin-3 [50] or paclitaxel [51]. was transfected in Jurkat cells as previously described [46]. But until our knowledge there is a current lack of clinical results. In vivo mouse experiments The control of ABC transporters by cell metabolism is a physiologic process required to optimize substrate In vivo experiments were carried out using 6 to 8 use. Therefore wtp53 cells in proliferative states, which weeks/old male NSG mice. Mice were bred and housed enter aerobic glycolysis, would also increase ABC in pathogen-free conditions in the animal facility of the expression and become more resistant to drug treatment. European Institute of Oncology–Italian Foundation for This could have implications for the treatment of non- Cancer Research (FIRC), Institute of Molecular Oncology (Milan, Italy). For engraftment of human cells, 1 million www.impactjournals.com/oncotarget 1124 Oncotarget AML cells were injected intravenously (i.v.) through the free long-term culture medium (Lanford medium, LNF). lateral tail vein in non-irradiated mice. NSG mice with The number of confluent attached cells was estimated at established human AML tumors (day 80 post-graft) were ~1.5x105 cells/cm 2. treated with DCA (50 mg/kg, 1 dose/day by gavage, starting at day 1 for 16 consecutive days). Human tumor Transient transfection AML cells gather in mouse spleen and bone marrow, hence we isolated mRNA from these organs. We used Jurkat cells in logarithmic growth phase were human-specific primers to visualize expression of human transfected with the indicated amounts of plasmid by ABC mRNA. In a different experiment B6 wt mice were electroporation [46, 56]. In each experiment, cells were treated in pathogen-free conditions at the INM animal transfected with the same total amount of DNA by facility at Montpellier, France, with a daily single dose of supplementing with empty vector. Cells were incubated for DCA (50 mg/kg/day) intraperitoneally and mouse LDLR 10 min at RT with the DNA mix and electroporated using mRNA was analyzed in spleen and liver after different the Gene Pulser Xcell™ Electroporation system (Bio-Rad) times. DWP9P)LQȝORI530,([SUHVVLRQ of the different proteins was confirmed by western blot. Cell lines and culture conditions The transfection efficiency in Jurkat TAg cells is between 60 and 80%. In HuH7 cells, transfection of 30–50 nM The leukemic human cell lines, HL-60, NB4, siRNAs was carried out using Lipofectamine RNAiMAX Jurkat TAg and OCI-AML3 were grown in RPMI 1640– (Invitrogen) in Opti-MEM (Invitrogen), according to the Glutamax (GIBCO) supplemented with 5% (Jurkat) or manufacturer’s instructions. Primary hepatocytes were 10% (OCI) FBS [15, 16]. Primary cells from a lymphoma transfected twice, at day first and third post-seeding. Cells B cell patient (BCL-P2) were grown in the same medium were harvested 48 to 96 h post-transfection. OCI-AML3 with 10% FBS. In certain experiments cells were grown cells (2 x 10 6LQȝO ZHUHWUDQVIHFWHGZLWKQ0 in RPMI 1640 without glucose (GIBCO 11879) with p53 siRNA, or control siRNA by electroporation using the addition of 2 mM glutamine and 10 mM galactose Nucleofactor Electroporation system (Lonza). Cells were (OXPHOS medium). The Jurkat TAg cells carry the SV40 harvested 24 to 72 h post-transfection. large T Ag to facilitate cell transfection. The hepatic cell lines HepG2-C3A and HuH7 were grown in MEM and Counting and determination of cell viability DMEM respectively supplemented with 10% FBS, 1 mM sodium pyruvate, 2 mM glutamine, penicillin and Cell viability and cell numbers were determined streptomycin. The HCT116 human colon cancer cells using the Muse ® Cell Analyzer (Millipore) Number were cultured in low glucose (5 mM) DMEM medium of cells means number of live cells and viability is the supplemented with 10% FBS. Cellular confluence during number of live cells counted divided by total cell counted experiments was between 80-85% for adherent cells. (alive and dead cells).

Human liver samples and preparation of PHHs RT-PCR cultures Total RNA was extracted using NucleoSpin Liver samples were obtained from liver resections RNA isolation columns (Macherey-Nagel). Reverse performed in adult patients for medical reasons (CRB- transcription was carried out using iScript™ cDNA CHUM - Biological Resource Center of the Montpellier Synthesis Kit (Biorad). Quantitative PCR was performed University Hospital, Dr. Jeanne Ramos and Pr. Sylvain with KAPA SYBR Green qPCR SuperMix (Cliniscience) Lehmann). The use of human specimens for scientific and a CFX Connect™ Real-Time qPCR machine (Biorad) purposes was approved by the French National Ethics with ABCB1, ABCC1, ABCC5, ABCG2 and actin Committee. All methods were carried out in accordance SULPHUV$OOVDPSOHVZHUHQRUPDOL]HGWRȕDFWLQP51$ with the approved guidelines and regulations of this levels. Results are expressed relative to control values committee. Written informed consent was obtained from arbitrarily set at 100. The primers used were: each patient prior to surgery. Human hepatocytes isolation Mouse primers and culture were performed as described previously [55]. $%&% )RUZDUG ƍ77&7&777*7&&*&** Briefly, after liver perfusion, hepatocytes were counted $*7&ƍ  5HYHUVH ƍ *$$7*&77&&$**&$7$$ and cell viability was assessed by trypan blue exclusion *&*ƍ $%&&)RUZDUG ƍ&7$*&7**7&*777&$ test. A suspension of 1x10 6 cells/mL per well was added &**7 ƍ  5HYHUVH ƍ&&7&&7*&$*&&&$&7 in 12-well plates pre-coated with type I collagen (Beckton $7$&ƍ  $%&& )RUZDUG ƍ&$$$*&&** Dickinson) and cells were allowed to attach for 12h. 7**$$$$7***ƍ  5HYHUVH ƍ *7***$ Then, the supernatant containing dead cells and debris $*$&*$*77*&7*$ƍ  $%&* )RUZDUG ƍ was carefully removed and replaced with 1 mL of serum- &7&&77*&&$*$7$$*$****ƍ  5HYHUVH ƍ www.impactjournals.com/oncotarget 1125 Oncotarget &&7&$*77$$777&$**$&*$&$*ƍ  $FWLQ 7.5, 150 mM NaCl, 1% triton, 0.1% SDS, 1 mM EDTA, )RUZDUG ƍ*&**$&7*77$*7*$*&7*&*ƍ  1mM Na Butyrate + Halt™ Protease Inhibitor Cocktail 5HYHUVH ƍ7*777*&7&&$$&&$$&7*&7*7&ƍ  (1X)), a high-salt buffer (50 mM Tris-HCl pH 7.5, 500 Human primers mM NaCl, 1% triton, 0.1% SDS, 1 mM EDTA, 1mM Na Butyrate) and a LiCl-containing buffer (20 mM Tris-HCl $%&% )RUZDUG ƍ**$**&&$$&$7$&$ pH 7.5, 250 mM LiCl, 1% NP40, 1% Na deoxycholate, 7*&&7ƍ  5HYHUVH ƍ $**&7*7&7$$&$$*** 1 mM EDTA, 1mM Na Butyrate) and, then, twice with &$&ƍ  $%&& )RUZDUG ƍ &77* a TE buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 7777*&7*&$***&7&ƍ  5HYHUVH ƍ&$&$7& 7ZHHQ 6DPSOHVZHUHWKHQHOXWHGLQȝ/RI $*$$77&&7*&*&& ƍ  $%&& )RUZDUG elution buffer (100 mM NaHCO3, 1% SDS), and DNA- ƍ&&&*&7&7***$&7**$$ƍ  5HYHUVH ƍ protein complexes were incubated at 65°C for 5 hours to *7$*$$****$$$&$**&&& ƍ  $%&* reverse crosslinks. Samples were then treated with 100mg/ )RUZDUG ƍ 7*7*777$7*$7**7&7*77**7& ml proteinase K and 100 mg/ml RNAse A at 45°C for 2 ƍ  5HYHUVH ƍ 7*77*&$77*$*7&&7***& ƍ  hours to digest proteins and contaminating RNA. DNA $FWLQ)RUZDUG ƍ*$***$$$7&*7*&*7*$&$ƍ  was purified with an extraction kit (NucleoSpin Gel 5HYHUVH ƍ$$7$*7*$7*$&&7**&&*7ƍ  and PCR clean-up, Macherey-Nagel) according to the manufacturer’s recommendations and qPCR analysis was Flow cytometry performed using the Roche LightCycler 480 real-time PCR system. The data were normalized with inputs taken 6 Briefly, 1x10 cells were stained with antibody in from samples before the immunoprecipitation and treated PBS with 2% FBS and incubated at 37°C for 30 min. under the same conditions. The primers used to amplify &HOOVZHUHWKHQZDVKHGDQGVXVSHQGHGLQ±ȝO3%6 various regions of LDLR gene promoter. Supplementary 2% FBS and staining was analyzed using a Gallios flow Table 1 shows the primers used in this experiment. cytometer (Beckman) and the Kaluza software. Daunorubicin accumulation / drug uptake ChIP experiment Studies OCI-AML3 cells were treated with 10mM DCA The intracellular daunorubicin accumulation for 72 h. Ten million cells were centrifuged (5min; in cells was examined by flow cytometer [57]. The 1200rpm) and the pellet was washed two times in 1X logarithmically growing cells were cultured in 48-well phosphate-buffered saline (PBS) at room temperature and SODWHVDQGLQFXEDWHGZLWKRUZLWKRXWȝ0GDXQRUXELFLQ suspended in 10mL of 1X PBS. Cells were fixed in 1% IRUGLIIHUHQWWLPHV     PLQXWHV DQGȝ0 paraformaldehyde (Electron Microscopy Sciences) at room for 24 h). After incubation, cells were placed on ice, temperature for 5 min and lysed in 1 ml of cell lysis buffer washed twice with PBS and analyzed by flow cytometry (5 mM PIPES, 85 mM KCl, 0.5% NP40, Na Butyrate (Beckman-coulter, Elite), excitation 488 nm (argon laser) 10mM + 2X protease inhibitor cocktail (Halt™ Protease for the mean fluorescence intensity (MFI) of intracellular Inhibitor Cocktail, EDTA-Free (100X), Thermofischer)) at daunorubicin. A minimum of 10000 events was analyzed 0°C for 10 min. Nuclei were recovered by centrifugation for each histogram. PLQUSP DWƒ&DQGO\VHGLQȝOQXFOHLO\VLV buffer (50 mM Tris-HCl pH 7.5, 1% SDS, 10 mM EDTA, Na Butyrate 10mM + Halt™ Protease Inhibitor Cocktail Statistical analysis ; DWƒ&IRUDWOHDVWKRXUVȝORIHDFKVDPSOH The statistical analysis of the difference between were then sonicated 2 times for 5 min (30 s on/off) at means of paired samples was performed using the paired 4°C using a Bioruptor (Diagenode). After sonication, t test. The results are given as the confidence interval absorbances at 280 nm (A280) of 1/100 diluted samples (*: p<0.05, ** : p<0.01, ***: p<0.005). All the experiments were measured and A280nm and was adjusted to 0.133 described in the figures with a quantitative analysis have with nuclei lysis buffer. One hundred microliter were used been performed at least three times in duplicate. Other for ChIP experiments in a final volume of 1 ml. Samples experiments were performed three times with similar were incubated under gentle agitation at 4°C overnight results. LQWKHSUHVHQFHRIȝJRIHLWKHUDVSHFLILFDQWLERG\RU a negative control. Antibodies (anti-K27Ac Ab4729 Author contributions (Abcam) and negative control IgG (Diagenode)) were previously bound to DYNA Beads Protein G Novex (Life S.B. A.K, N.A-V, D.G., C.A., S.G, C.G, D-N.V, S.O, Technology) according to the supplier’s recommendations. G. T, M. B and F. B perform experiments. F. B, G.C., J. Dynabeads-bound immunoprecipitates were sequentially H, M. D-C and M.V wrote the main manuscript text. All washed once with a low salt buffer (50 mM Tris-HCl pH authors reviewed the manuscript.

www.impactjournals.com/oncotarget 1126 Oncotarget ACKNOWLEDGMENTS AND FUNDING 10. Muller PA, Vousden KH. Mutant p53 in cancer: new functions and therapeutic opportunities. Cancer Cell. 2014; FACS analysis was performed at the platform 25:304-317. Montpellier Rio Imaging (MRI). We are indebted to 11. Chan KT, Lung ML. Mutant p53 expression enhances drug the staff of the INM animal facility (RAM) for animal resistance in a hepatocellular carcinoma cell line. Cancer care. The collection of clinical data and samples Chemother Pharmacol. 2004; 53:519-526. (HEMODIAG_2020) at the CHRU Montpellier was 12. Sturm I, Bosanquet AG, Hermann S, Guner D, Dorken supported by funding from Région Languedoc Roussillon. B, Daniel PT. Mutation of p53 and consecutive selective The authors thank Dimitris Xirodimas for providing cells drug resistance in B-CLL occurs as a consequence of prior and p53 siRNA constructs. DNA-damaging chemotherapy. Cell Death Differ. 2003; All our funders are public or charitable 10:477-484. organizations. This work was supported by an AOI from 13. Scotto KW. Transcriptional regulation of ABC drug the CHU Montpellier (N°221826) (GC and MV), a grant transporters. Oncogene. 2003; 22:7496-7511. from La Ligue Regionale contre le Cancer Comitte 14. Charni S, de Bettignies G, Rathore MG, Aguilo JI, van Languedoc-Rousillon (GC), a grant from Fondation den Elsen PJ, Haouzi D, Hipskind RA, Enriquez JA, de France (0057921) to MV and a fellowship from the Sanchez-Beato M, Pardo J, Anel A, Villalba M. Oxidative Ministère de l'Enseignement Supérieur et de la Recherche phosphorylation induces de novo expression of the MHC (DNV). class I in tumor cells through the ERK5 pathway. J Immunol. 2010; 185:3498-3503. CONFLICTS OF INTEREST 15. Lopez-Royuela N, Rathore MG, Allende-Vega N, Annicotte Authors declare no competing financial interests. JS, Fajas L, Ramachandran B, Gulick T, Villalba M. Extracellular-signal-regulated kinase 5 modulates the REFERENCES antioxidant response by transcriptionally controlling Sirtuin 1 expression in leukemic cells. Int J Biochem Cell Biol. 1. Jones PM, George AM. The ABC transporter structure and 2014; 53:253-261. mechanism: perspectives on recent research. Cell Mol Life 16. Khan AU, Rathore MG, Allende-Vega N, Vo DN, Sci. 2004; 61:682-699. Belkhala S, Orecchioni S, Talarico G, Bertolini F, Cartron 2. Ponte-Sucre A. Availability and applications of ATP-binding G, Lecellier CH, Villalba M. Human leukemic cells cassette (ABC) transporter blockers. Appl Microbiol performing oxidative phosphorylation (OXPHOS) generate Biotechnol. 2007; 76:279-286. an antioxidant response independently of reactive oxygen species (ROS) production. EBioMedicine. 2016; 3:43-53. 3. Neumann J, Rose-Sperling D, Hellmich UA. Diverse relations between ABC transporters and lipids: an overview. 17. Villalba M, Rathore MG, Lopez-Royuela N, Krzywinska E, Biochim Biophys Acta. 2017; 1859:605-618. Garaude J, Allende-Vega N. From tumor cell metabolism to tumor immune escape. Int J Biochem Cell Biol. 2013; 4. Bush JA, Li G. Cancer chemoresistance: the relationship 45:106-113. between p53 and multidrug transporters. Int J Cancer. 2002; 98:323-330. 18. Villalba M, Lopez-Royuela N, Krzywinska E, Rathore MG, Hipskind RA, Haouas H, Allende-Vega N. Chemical 5. Gottlieb E, Vousden KH. p53 regulation of metabolic metabolic inhibitors for the treatment of blood-borne pathways. Cold Spring Harb Perspect Biol. 2010; cancers. Anticancer Agents Med Chem. 2014; 14:223-232. 2:a001040. 19. Catalán E, Charni S, Aguiló JI, Enríquez JA, Naval J, Pardo 6. Puzio-Kuter AM. The role of p53 in metabolic J, Anel A, Villalba M. MHC-I modulation due to metabolic regulation. Genes Cancer. 2011; 2:385-391. https://doi. changes regulates tumor sensitivity to CTL and NK cells.. org/10.1177/1947601911409738. Oncoimmunology. 2015; 4:e985924. 7. Allende-Vega N, Krzywinska E, Orecchioni S, Lopez- 20. Kato Y, Kravchenko VV, Tapping RI, Han J, Ulevitch RJ, Royuela N, Reggiani F, Talarico G, Rossi JF, Rossignol R, Lee JD. BMK1/ERK5 regulates serum-induced early gene Hicheri Y, Cartron G, Bertolini F, Villalba M. The presence expression through transcription factor MEF2C. EMBO J. of wild type p53 in hematological cancers improves the 1997; 16:7054-7066. efficacy of combinational therapy targeting metabolism. Oncotarget. 2015; 6:19228-19245. https://doi.org/10.18632/ 21. Kato Y, Zhao M, Morikawa A, Sugiyama T, Chakravortty oncotarget.4653. D, Koide N, Yoshida T, Tapping RI, Yang Y, Yokochi T, Lee JD. Big mitogen-activated kinase regulates multiple 8. Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53 members of the MEF2 protein family. J Biol Chem. 2000; mutations in human cancers. Science. 1991; 253:49-53. 275:18534-18540. 9. Muller PA, Vousden KH. p53 mutations in cancer. Nat Cell 22. Kasler HG, Victoria J, Duramad O, Winoto A. ERK5 is a Biol. 2013; 15:2-8. novel type of mitogen-activated protein kinase containing www.impactjournals.com/oncotarget 1127 Oncotarget a transcriptional activation domain. Mol Cell Biol. 2000; switch hallmarked by reduced dependence on exogenous 20:8382-8389. glutamine. J Biol Chem. 2015; 290:8348-8359. 23. Yang CC, Ornatsky OI, McDermott JC, Cruz TF, Prody CA. 38. Chen J, Ding Z, Peng Y, Pan F, Li J, Zou L, Zhang Y, Liang Interaction of myocyte enhancer factor 2 (MEF2) with a H. HIF-1alpha inhibition reverses multidrug resistance mitogen-activated protein kinase, ERK5/BMK1. Nucleic in colon cancer cells via downregulation of MDR1/P- Acids Res. 1998; 26:4771-4777. glycoprotein. PLoS One. 2014; 9:e98882. 24. Ramachandran B, Yu G, Gulick T. Nuclear respiratory 39. Seebacher NA, Richardson DR, Jansson PJ. Glucose factor 1 controls myocyte enhancer factor 2A transcription modulation induces reactive oxygen species and increases to provide a mechanism for coordinate expression P-glycoprotein-mediated multidrug resistance to of respiratory chain subunits. J Biol Chem. 2008; chemotherapeutics. Br J Pharmacol. 2015; 172:2557-2572. 283:11935-11946. 40. Li D, Marchenko ND, Schulz R, Fischer V, Velasco- 25. Consortium EP. An integrated encyclopedia of DNA Hernandez T, Talos F, Moll UM. Functional inactivation of elements in the human genome. Nature. 2012; 489:57-74. endogenous MDM2 and CHIP by HSP90 causes aberrant 26. Stacpoole PW, Greene YJ. Dichloroacetate. Diabetes Care. stabilization of mutant p53 in human cancer cells. Mol 1992; 15:785-791. Cancer Res. 2011; 9:577-588. 27. Stacpoole PW. The pharmacology of dichloroacetate. 41. Pereira DM, Simoes AE, Gomes SE, Castro RE, Carvalho Metabolism: clinical and experimental. 1989; 38:1124-1144. T, Rodrigues CM, Borralho PM. MEK5/ERK5 signaling 28. Michelakis ED, Webster L, Mackey JR. Dichloroacetate inhibition increases colon cancer cell sensitivity to (DCA) as a potential metabolic-targeting therapy for cancer. 5-fluorouracil through a p53-dependent mechanism. Br J Cancer. 2008; 99:989-994. Oncotarget. 2016; 7:34322-34340. https://doi.org/10.18632/ oncotarget.9107. 29. Reitzer LJ, Wice BM, Kennell D. Evidence that glutamine, not sugar, is the major energy source for cultured HeLa 42. Ji L, Li H, Gao P, Shang G, Zhang DD, Zhang N, Jiang cells. J Biol Chem. 1979; 254:2669-2676. T. Nrf2 pathway regulates multidrug-resistance-associated protein 1 in small cell lung cancer. PLoS One. 2013; 30. Rossignol R, Gilkerson R, Aggeler R, Yamagata K, 8:e63404. Remington SJ, Capaldi RA. Energy substrate modulates mitochondrial structure and oxidative capacity in cancer 43. Wang X, Campos CR, Peart JC, Smith LK, Boni JL, Cannon cells. Cancer Res. 2004; 64:985-993. RE, Miller DS. Nrf2 upregulates ATP binding cassette transporter expression and activity at the blood-brain and 31. Rathore MG, Saumet A, Rossi JF, de Bettignies C, Tempe blood-spinal cord barriers. J Neurosci. 2014; 34:8585-8593. D, Lecellier CH, Villalba M. The NF-kappaB member p65 controls glutamine metabolism through miR-23a. Int J 44. Schweppe RE, Cheung TH, Ahn NG. Global gene Biochem Cell Biol. 2012; 44:1448-1456. expression analysis of ERK5 and ERK1/2 signaling reveals a role for HIF-1 in ERK5-mediated responses. J Biol Chem. 32. Stacpoole PW, Kurtz TL, Han Z, Langaee T. Role of 2006; 281:20993-21003. dichloroacetate in the treatment of genetic mitochondrial diseases. Adv Drug Deliv Rev. 2008; 60:1478-1487. 45. Breier A, Gibalova L, Seres M, Barancik M, Sulova Z. New insight into p-glycoprotein as a drug target. Anticancer 33. Michelakis ED, Sutendra G, Dromparis P, Webster L, Agents Med Chem. 2013; 13:159-170. Haromy A, Niven E, Maguire C, Gammer TL, Mackey JR, Fulton D, Abdulkarim B, McMurtry MS, Petruk KC. 46. Garaude J, Cherni S, Kaminski S, Delepine E, Chable- Metabolic modulation of glioblastoma with dichloroacetate. Bessia C, Benkirane M, Borges J, Pandiella A, Iniguez Sci Transl Med. 2010; 2:31ra34. MA, Fresno M, Hipskind RA, Villalba M. ERK5 activates NF-kappaB in leukemic T cells and is essential for their 34. Zhou Y, Wang R, Chen B, Sun D, Hu Y, Xu P. Daunorubicin growth in vivo . J Immunol. 2006; 177:7607-7617. and gambogic acid coloaded cysteamine-CdTe quantum dots minimizing the multidrug resistance of lymphoma in 47. Khan AU, Allende-Vega N, Gitenay D, Gerbal-Chaloin S, vitro and in vivo . Int J Nanomed. 2016; 11:5429-5442. Gondeau C, Vo DN, Belkahla S, Orecchioni S, Talarico G, Bertolini F, Bozic M, Valdivielso JM, Bejjani F, et al. 35. Cheng J, Haas M. Frequent mutations in the p53 tumor The PDK1 inhibitor dichloroacetate controls cholesterol suppressor gene in human leukemia T-cell lines. Mol Cell homeostasis through the ERK5/MEF2 pathway. Sci Rep. Biol. 1990; 10:5502-5509. 2017; 7:10654. 36. Broxterman HJ, Pinedo HM, Kuiper CM, Schuurhuis GJ, 48. Zhou X, Chen R, Yu Z, Li R, Li J, Zhao X, Song S, Liu Lankelma J. Glycolysis in P-glycoprotein-overexpressing J, Huang G. Dichloroacetate restores drug sensitivity human tumor cell lines. Effects of resistance-modifying in paclitaxel-resistant cells by inducing citric acid agents. FEBS Lett. 1989; 247:405-410. accumulation. Mol Cancer. 2015; 14:63. 37. Staubert C, Bhuiyan H, Lindahl A, Broom OJ, Zhu Y, 49. Voltan R, Rimondi E, Melloni E, Gilli P, Bertolasi V, Islam S, Linnarsson S, Lehtio J, Nordstrom A. Rewired Casciano F, Rigolin GM, Zauli G, Secchiero P. Metformin metabolism in drug-resistant leukemia cells: a metabolic

www.impactjournals.com/oncotarget 1128 Oncotarget combined with sodium dichloroacetate promotes B 53. Kunig AM, Parker TA, Nogee LM, Abman SH, Kinsella leukemic cell death by suppressing anti-apoptotic protein JP. ABCA3 deficiency presenting as persistent pulmonary Mcl-1. Oncotarget. 2016; 7:18965-18977. https://doi. hypertension of the newborn. J Pediatr. 2007; 151:322-324. org/10.18632/oncotarget.7879. 54. Nagy BM, Nagaraj C, Egemnazarov B, Kwapiszewska G, 50. Agnoletto C, Melloni E, Casciano F, Rigolin GM, Rimondi Stauber RE, Avian A, Olschewski H, Olschewski A. Lack of E, Celeghini C, Brunelli L, Cuneo A, Secchiero P, Zauli G. ABCG2 leads to biventricular dysfunction and remodeling Sodium dichloroacetate exhibits anti-leukemic activity in in response to hypoxia. Front Physiol. 2017; 8:98. B-chronic lymphocytic leukemia (B-CLL) and synergizes 55. Pichard L, Raulet E, Fabre G, Ferrini JB, Ourlin JC, Maurel with the p53 activator Nutlin-3. Oncotarget. 2014; 5:4347- P. Human hepatocyte culture. Methods Mol Biol. 2006; 4360. https://doi.org/10.18632/oncotarget.2018. 320:283-293. 51. Wang M, Liao C, Hu Y, Qinwen P, Jiang J. Sensitization of 56. Garaude J, Farras R, Bossis G, Charni S, Piechaczyk breast cancer cells to paclitaxel by dichloroacetate through M, Hipskind RA, Villalba M. SUMOylation regulates inhibiting autophagy. Biochem Biophys Res Commun. the transcriptional activity of JunB in T lymphocytes. J 2017; 489:103-108. Immunol. 2008; 180:5983-5990. 52. Belleville-Rolland T, Sassi Y, Decouture B, Dreano E, Hulot 57. Venne A, Li S, Mandeville R, Kabanov A, Alakhov V. JS, Gaussem P, Bachelot-Loza C. MRP4 (ABCC4) as a Hypersensitizing effect of pluronic L61 on cytotoxic potential pharmacologic target for cardiovascular disease. activity, transport, and subcellular distribution of Pharmacol Res. 2016; 107:381-389. doxorubicin in multiple drug-resistant cells. Cancer Res. 1996; 56:3626-3629.

www.impactjournals.com/oncotarget 1129 Oncotarget ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! |É~! ) ) ) ) ) ) )

!'".1#/!dcZ!! !

+?LG6>GF7J?4D!!GEHD8P!(!46L?N?LQ!K?=F4DK! 4FL?GP?74FL!J8KHGFK8!L>JGM=>!#/)h! ) I<@087>632+/=?7183+H++173:?7471+*3<;=?>J+ ) ) K&:+ ?67>+ >?@2D+ ?6/?+ &+ >6/=32+ 1;G/@?6;=>67+ 7:A;8A32+ 7:+ <8/::7:5+ /:2+ 3C31@?7;:+;4+?63+3C<3=793:?>E++7:18@27:5+<=3+@>32+7:+?63+>?@2DE+<=3+<3=4;=97:5+(.)%(+L+ ) ) ) ) ) ) ) ) ) ) !

! |ÉÄ! ™™™Ǥƒ–—”‡Ǥ ‘Ȁ• ‹‡–‹ˆ‹ ”‡’‘”–•

OPEN Mitochondrial Complex I activity signals antioxidant response –Š”‘—‰Šͻ Received: 17 November 2017 Abrar Ul Haq Khanͷ, Nerea Allende-Vega ͷǡ͸ , Delphine Gitenayͷ, Johan Garaude ͷ, Dang- Accepted: 21 March 2018 Nghiem Vo ͷ, Sana Belkhalaͷ, Sabine Gerbal-Chaloinͷ, Claire Gondeauͷǡ͹ , Martine Daujat- Published: xx xx xxxx Chavanieuͷ, Cécile Delettreͺǡͷ͹ , Stefania Orecchioniͻ, Giovanna Talarico ͻ, Francesco Bertoliniͻ, Alberto Anelͼ, José M. Cuezvaͽ, Jose A. Enriquez ;ǡͿ , Guillaume CartronͷͶ , Charles-Henri Lecellierͷͷǡͷ͸ , Javier Hernandezͷ & Martin Villalba ͷǡ͸

Oxidative phosphorylation (OXPHOS) generates ROS as a byproduct of mitochondrial complex I activity. ROS-detoxifying enzymes are made available through the activation of their antioxidant ”‡•’‘•‡‡Ž‡‡–•ȋȌ‹–Š‡‹”‰‡‡’”‘‘–‡”•Ǥ ͸„‹†•–‘•ƒ†‹†— ‡•–Š‹•ƒ–‹Ǧ‘š‹†ƒ– ”‡•’‘•‡Ǥ‡•Š‘™–Šƒ– ‡ŽŽ•ˆ”‘—Ž–‹’Ž‡‘”‹‰‹•’‡”ˆ‘”‹‰ ‹†— ‡† ͸‡š’”‡••‹‘ƒ† its transcriptional activity. The  ͸ ’”‘‘–‡” ‘–ƒ‹• ͸„‹†‹‰•‹–‡•ƒ†–Š‡ͻ‹†— ‡†  ͸Ǧ†‡’‡†‡– ͸‡š’”‡••‹‘ǤŽ‘ ‹‰ ‹ƒ‘—•‡‘†‡Ž†‡ ”‡ƒ•‡† ”ͻ and ”ˆ͸ ‡š’”‡••‹‘Ǥ —”–Š‡”‘”‡ǡƤ„”‘„Žƒ•–•†‡”‹˜‡†ˆ”‘’ƒ–‹‡–•™‹–Š‹–‘ Š‘†”‹ƒŽ†‹•‘”†‡”•ƒŽ•‘•Š‘™‡† low expression of ͻ and  ͸ mRNAs. Notably, in cells lacking functional mitochondrial complex ƒ –‹˜‹–› †‹†‘–‹†— ‡ͻ‡š’”‡••‹‘ƒ†ˆƒ‹Ž‡†–‘‰‡‡”ƒ–‡–Š‹•ƒ–‹Ǧ‘š‹†ƒ–”‡•’‘•‡Ǥ ‘’Ž‡š ƒ –‹˜‹–›‹†— ‡•ͻ‡š’”‡••‹‘–Š”‘—‰Šˆ—ƒ”ƒ–‡ƒ ——Žƒ–‹‘Ǥ—ƒ”›‘–‹  ‡ŽŽ•Šƒ˜‡ evolved a genetic program to prevent oxidative stress directly linked to OXPHOS and not requiring ROS.

Energy consumption in organisms should be !nely regulated to spare resources. "e vast majority of eukaryotic cells perform oxidative phosphorylation (OXPHOS), which uses the energy generated by mitochondrial oxida- tion to produce adenosine triphosphate (ATP). "is metabolic pathway is highly e#cient in releasing energy but it produces reactive oxygen species (ROS) as a byproduct. ROS are involved in normal cell signaling and homeo- stasis. However, under stress conditions levels may rapidly increase resulting in cell damage, a process known as oxidative stress. Hence, cells using mitochondria as !rst energy source must regulate ROS levels. Logically, ROS and mitochondria are functionally linked in several ways. First, ROS in the short-term regulate mitochondrial morphology and function via non-transcriptional pathways 1. Second, ROS lead to Kelch-like ECH-associated protein 1 (KEAP-1) degradation, thereby activating nuclear factor (erythroid-derived 2)-like 2 (NFE2L2 or NRF2) 2,3, which regulates expression of mitochondrial genes4. In addition, NRF2 controls ROS production by mitochondria 5 and mitochondrial function 6,7. NRF2 arguably mediates the strongest anti-oxidant cellular response by binding to anti-oxidant response elements (ARE) in gene promoters and, consequently, regulates oxi- dative stress 2,3. On the other hand, mitochondrial activity induced by acute exercise promotes Ref1/Nrf2 signaling and increases mitochondrial antioxidant activity and capacity in myocardial and skeletal muscle 8,9. Remarkably, restraining OXPHOS in vivo in the liver strongly decreases Nrf2 levels 10 . Moreover, tumor cells forced to perform

ͷIRMB, INSERM, Univ Montpellier, Montpellier, France. ͸Institut de Regenerative Medicine et Biothérapie (IRMB),  ‘–’‡ŽŽ‹‡”ǡ‘–’‡ŽŽ‹‡”ǡ͹ͺ͸Ϳͻǡ ”ƒ ‡Ǥ ͹Département d’Hépato-gastroentérologie A, Hôpital Saint Eloi, CHU Montpellier, France. ͺ ͷͶͻͷǡ •–‹–—–‡‘ˆ‡—”‘• ‹‡ ‡•‘ˆ‘–’‡ŽŽ‹‡”ǡ‘–’‡ŽŽ‹‡”ǡ ”ƒ ‡Ǥ ͻƒ„‘”ƒ–‘”›‘ˆ ‡ƒ–‘Ž‘‰›Ǧ ‘Ž‘‰›ǡ—”‘’‡ƒ •–‹–—–‡‘ˆ ‘Ž‘‰›ǡ‹Žƒǡ –ƒŽ›Ǥ ͼ‡’ƒ”–‡–‘ˆ‹‘ Š‡‹•–”›ƒ†‘Ž‡ —Žƒ” ƒ†‡ŽŽ—Žƒ”‹‘Ž‘‰›ǡ”ƒ‰× ‡ƒŽ–Š‡•‡ƒ” Š •–‹–—–‡ȋ ”ƒ‰×Ȍǡ‹˜‡”•‹–›‘ˆƒ”ƒ‰‘œƒǡƒ”ƒ‰‘œƒǡ’ƒ‹Ǥ ͽDepartamento de Biología Molecular, Centro de Biología Molecular Severo Ochoa, CSIC-UAM, CIBERER, Universidad ƒ—–א‘ƒ†‡ƒ†”‹†ǡ͸;ͶͺͿǡƒ†”‹†ǡ’ƒ‹Ǥ ;Centro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC) ‡Ž Š‘” ‡”ƒ†‡œŽƒŽ‰‘ǡ͹͸;͸ͶͿǡƒ†”‹†ǡ’ƒ‹ǤͿ  Ǥ‡Ž Š‘” ‡”ƒ†‡œŽƒ‰”‘ǡ͹͸;͸ͶͿǡƒ†”‹†ǡ’ƒ‹Ǥ ͷͶ ±’ƒ”–‡‡–†ǯ ±ƒ–‘Ž‘‰‹‡Ž‹‹“—‡ǡ ‘–’‡ŽŽ‹‡”ǡ‹˜‡”•‹–±‘–’‡ŽŽ‹‡” ǡ;Ͷƒ˜‡—‡—‰—•–‹ Ž‹ Š‡ǡ͹ͺ͸Ϳͻǡ Montpellier, France. ͷͷ IGMM, CNRS, Univ. Montpellier, Montpellier, France. ͷ͸ Institut de Biologie Computationnelle, Montpellier, France. ͷ͹ ‡’ƒ”–‡–‘ˆ‹‘Ž‘‰›ƒ† ‡ƒŽ–Š ‹‡ ‡•ǡ‹˜‡”•‹–›‘ˆ‘–’‡ŽŽ‹‡”ǡ‘–’‡ŽŽ‹‡”ǡ ”ƒ ‡Ǥ ‘””‡•’‘†‡ ‡ƒ†”‡“—‡•–•ˆ‘”ƒ–‡”‹ƒŽ••Š‘—Ž†„‡ƒ††”‡••‡†–‘ǤǤȋ‡ƒ‹Žǣ ƒ”–‹Ǥ˜‹ŽŽƒŽ„ƒ̻‹•‡”Ǥˆ” )

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 1 www.nature.com/scientificreports/

OXPHOS generate a NRF2-mediated anti-ROS response 11 . However, how mitochondria transcriptionally signal the genetic program to block the ROS they produce remains unknown. NRF2 activation depends on its dissociation from the repressor protein KEAP1 and its subsequent transloca- tion into the nucleus 2. In hematopoietic cells, the MAPK extracellular signal-regulated kinase-5 (ERK5), through the transcription factor MEF2, induces expression of miR-23 that inhibits KEAP-1 mRNA leading to NRF2 acti- vation 11 . Several types of oxidative stress activate ERK5 12 , notably in leukemic cells 11 ,13 ,14 . In fact, ERK5 is con- sidered a redox MAPK 15 . In endothelial cells, steady laminar blood $ow (s-$ow) activates ERK5 that induces up-regulation of NRF2-dependent gene expression, although the mechanism is not fully elucidated 16 ,17 . Growing evidence indicates that there are alternative pathways leading to de novo production of NRF2 3. In this context, KEAP-1 inhibition only partially accounts for OXPHOS-induced antioxidant response 11 . Chip-seq exper- iments performed by the ENCODE consortium have shown that the NRF2 promoter contains MEF2 binding sites 18 . Moreover, predicted networks of transcription factor interactions in skeletal muscle unveil direct regula- tion of NRF2 by MEF2A 19 and MEF2D binds and activates the Nrf2 promoter 20 . Hence, ERK5 could transcrip- tionally induce NRF2 expression through MEF2, a transcription factor that mediates some of the metabolic e%ects of ERK5 11 ,13 ,14 ,21 –24 . In fact, ERK5 regulates the choice of catabolic substrates in hematopoietic cells 11 ,13 ,14 ,21 –23 , suggesting that is a good candidate to mediate the link between OXPHOS and the antioxidant response. We hypothesize that mitochondrial activity triggers the ERK5 pathway that, through MEF2, induces NRF2 expres- sion and NRF2-mediated antioxidant response. We validate this by showing that mitochondrial complex I activ- ity and fumarate accumulation induce the transcriptional expression of ERK5 . ERK5 through MEF2 induces NRF2 de novo expression. "erefore, mitochondrial activity is directly linked to the most important antioxidant response in the absence of de novo increase in ROS levels. "is implies that eukaryotic cells have evolved a genetic program to prevent oxidative stress directly linked to OXPHOS and not requiring ROS. Results  Ǧ‹†— ‡††‡‘˜‘‡š’”‡••‹‘‘ˆ ͸Ǥ We have previously described that leukemic cells per- forming OXPHOS generated an anti-oxidant response independently of ROS 11 . "is response was partially medi- ated by an ERK5-induced increase in miR-23 that impairs expression of KEAP-1 11 . In parallel experiments, we found that NRF2 mRNA was also increased in three hematopoietic cell lines and in primary cells obtained from a B-cell lymphoma (BCL) patient growing in OXPHOS medium (Fig. 1A ). "is glucose-free culture medium has !nal concentrations of 4 mM glutamine and 10 mM galactose. Glutamine is used to drive mitochondria to utilize OXPHOS and galactose allows cells to synthesize nucleic acids through the pentose phosphate pathway 13 ,14 ,25 ,26 . We called it ‘OXPHOS medium’, because it forced leukemic cells to use OXPHOS as primary ATP producer 13 ,24 ,27 . "e PDK1 inhibitor dichloroacetate (DCA), which stimulates OXPHOS in all tested leukemic cells 11 ,13 ,14 ,22 ,27 ,28 , also increased NRF2 mRNA (Fig. 1A ). Both ways to stimulate OXPHOS also induced NRF2 protein (Fig. 1B ). "e e%ect of DCA on NRF2 mRNA and protein is reproduced in two hepatic cell lines (Supplemental Fig. 1A) and in a group of primary leukemic cells from 4 patients (Supplemental Fig. 1B). Of relevance, we observed that in primary human hepatocytes DCA also increased ERK5 and NRF2 mRNA as well as that of the NRF2 targets HO-1 and NQO-1 (Fig. 1C ). In summary OXPHOS induced expression of NRF2 in multiple cell contexts.

 ‹†— ‡† ͸–”ƒ•Ž‘ ƒ–‹‘–‘–Š‡— Ž‡—•Ǥ NRF2 must translocate to the nucleus to acti- vate its target genes and generate the antioxidant response. HuH7 hepatic cells treated with DCA showed NRF2 accumulation in the nucleus (Fig. 2A ). "ese results were reproduced in non-adherent Jurkat cells by western blotting (Fig. 2B ) and in the hepatic cell line HepG2C3A (Supplemental Fig. 2A). We observed a total increase in NRF2 that was more predominant in the nuclear fraction. OXPHOS medium also induced NRF2 translocation to the nucleus in Jurkat cells (Supplemental Fig. 2B).

OXPHOS induced de novo expression of  ͸‹˜‹˜‘ . To test if enhanced OXPHOS could exert a similar e%ect on NRF2 expression in vivo , we engra*ed AML primary cells in non-obese diabetic/severe com- bined immunode!cient (NOD/SCID)-interleukin-2 receptor γ null (NSG) mice, as previously described 27 . Mice with established tumors (day 80 post-gra*) were treated with DCA. "e treatment was not toxic and did not show any notable e%ect on mouse survival 27 . Human tumor AML cells gather in mouse spleen and bone marrow, hence we isolated mRNA from these organs. We used human-speci!c primers to analyze the expression of the selected mRNAs and found an increase in NRF2 mRNA (Fig. 3A ). "is increase paralleled that of ER K5 and NQO-1 under similar conditions 11 . DCA also induced mouse Erk5 , Nrf2 and Nqo1 mRNA in liver and spleen in a separate experiment in which C57BL/6 wild type mice were treated for di%erent periods of time, 1 to 3 days, with DCA (Fig. 3B ). "e e%ect was !rst observed in spleen and later in liver tissue. Nrf2 was likely active because we observed an increase in its target gene Nqo-1 (Fig. 3B ). Hence DCA induced NRF2 expression in multiple cell populations in vitro and in vivo .

 ”‡ƒ•‡‹Ž‡˜‡Ž•‹•‘–‡••‡–‹ƒŽˆ‘” ͸‡š’”‡••‹‘Ǥ "e cellular oxidative state can regu- late NRF2 expression 2. "erefore, we investigated whether NRF2 expression is regulated by ROS in our setting. DCA induces ROS production in some hematopoietic cell lines, e.g. OCI-AML3, but not all, e.g. Jurkat 11 ,27 ,29 . In contrast, both cell lines increased NRF2 expression suggesting that ROS production was not essential for this induction (Fig. 4). Next, we incubated both cell lines with the antioxidant N-acetyl-cysteine (NAC), which failed to consistently reduce DCA-induced ERK5 , NRF2 or NQO-1 mRNA (Fig. 4), although e#ciently blocked DCA-induced ROS increase 11 ,29 . We observed similar results in primary leukemic cells from a BCL patient (BCL-P2). DCA does not increase ROS in the hepatic cell line HepG2C3A, but it did in Huh7 29 . However, DCA signi!cantly increased ERK5 , NRF2 or NQO-1 mRNA in both cell lines and in the presence of NAC (Supplemental Fig. 3). "ese results excluded a major role of ROS in NRF2 expression a*er DCA treatment. Normally ROS

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 2 www.nature.com/scientificreports/

Figure 1. Cells performing OXPHOS upregulated NRF2 expression. ( A) Di%erent hematopoietic cells were incubated in OXPHOS medium or treated with 5 mM DCA for 2 weeks. NRF2 mRNA was quanti!ed by qPCR and values normalized to β-actin mRNA. Results were represented as the % of mRNA compared to cells growing only in glucose medium. Bars show average ± SD of 3 independent experiments performed in triplicate. ( B) NRF2 protein expression was analyzed in several cell lines by western blotting (upper panel) or by $ow cytometry in OCI-AML3 cells. ( C) Hepatocytes from 4 donors were treated with the indicated concentration of DCA for 24 h and ERK5, NRF2, NQO-1 and HO-1 mRNA were analyzed. Bars show average ± SD of the four donors performed in duplicate.

activate NRF2. Unexpectedly in AML cells, there is no relationship between high ROS levels and high nuclear NRF230 . Furthermore, the use of NAC, which successfully sequesters endogenous ROS in AML, has no e%ect on nuclear NRF2 levels 30 . Taken together, this excludes ROS as causing nuclear accumulation of NRF2 in resting human AML cells 30 .

ͻȀ ͸ ‘–”‘Ž• ͸‡š’”‡••‹‘Ǥ Next, we investigated the mechanism responsible for mito- chondria activity-induced NRF2 expression. Reducing expression of ERK5 with a small hairpin RNA (shERK5)

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 3 www.nature.com/scientificreports/

Figure 2. OXPHOS induced NRF2 translocation into the nucleus. ( A) Huh7 cells were treated with 10 mM DCA for 48 h and nuclear translocation was revealed by immuno$uorescence. ( B) Jurkat cells were treated with 10 mM DCA for 48 h and NRF2 nuclear translocation was revealed by subcellular fractionation and western blotting.

diminished NRF2 mRNA expression in hematopoietic cells under resting conditions (Fig. 5A ). NRF2 protein levels were also reduced a*er shERK5 transfection (Fig. 5B ). We could not treat shERK5-expressing cells with DCA because they die due to lack of appropriate mitochondrial functions and antioxidant response 11 ,13 ,14 ,21 –23 . Conversely, overexpression of ERK5 increased NRF2 mRNA (Fig. 5A ). Reducing expression of ERK5 in primary human hepatocytes (Fig. 5C ) and hepatic cell lines, HuH7 and HepG2C3A (Supplemental Fig. 1A,B), with small interference RNA for ERK5 (siERK5) also impaired expression of NRF2 and its target genes NQO-1 and HO-1 . To further study the role of the ERK5/MEF2 pathway in NRF2 expression, we overexpressed several pro- teins of this pathway. Strong activation of the ERK5 pathway by co-overexpression of a constitutively active mutant of MEK5 (MEK5D), the upstream kinase of ERK5, and ERK5 induced a greater increase in NRF2 mRNA (Supplemental Fig. 4C). In those experiments, only 30–60% of the cells are e%ectively transfected. To overcome this issue, we use a luciferase reporter plasmid driven by a DNA fragment of 1.5 kb of the human NRF2 pro- moter 30 . In this context, cells expressing the reporter plasmid also contain the overexpressed proteins. ERK5

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 4 www.nature.com/scientificreports/

Figure 3. Cells performing OXPHOS induce NRF2 expression in vivo . ( A) NSG mice were engra*ed with primary human AML cells. At day 80 post-gra*, they were treated with DCA (n = 4) or leave untreated (n = 4). At day 140, mRNA from bone marrow or spleen was isolated and the expression of di%erent human mRNA was quanti!ed by qPCR. ( B) B6 wt mice (n = 4/5 per group) were treated with a dose of DCA (50 mg/kg) everyday intraperitoneally and mouse Erk5 , Nrf2 and Nqo-1 mRNA was analyzed in spleen and liver at di%erent times. "e data represent means ± SD; statistics were performed using student t-test ( A) or One-way ANOVA with post-hoc Tukey test ( B); *p < 0.05, ** p < 0.01, ***p < 0.001. Di%erent times posttreatment were compared to non-treated mice (control) if not speci!ed in the graph.

signi!cantly activated the reporter and MEK5D increased this e%ect (Fig. 5D ). Expression of a dominant neg- ative form of MEF2C (MEF2C-DN) decreased the e%ect of ERK5 and MEK5D. "is DN construct also dimin- ished basal or DCA-stimulated reporter expression. In contrast, MEF2C overexpression increased both basal

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 5 www.nature.com/scientificreports/

Figure 4. Increase in ROS levels is not required for NRF2 expression. OCI-AML and HuH7 cell lines and primary leukemic cells from a BCL patient were treated with 2 mM NAC 1 h before adding DCA (10 mM) for 24 h. mRNA was analyzed as described in Fig. 1. Experiments were done in triplicate and data represent means ± SD; statistics were performed using One-way ANOVA with post-hoc Tukey test; *p < 0.05, ** p < 0.01, ***p < 0.001. Treatments were compared to non-treated cells (control) if not speci!ed in the graph.

and DCA-induced activity (Fig. 5D ). DCA, which induced strong activation, did not show a synergistic, but rather an additive, e%ect with the activating proteins. "ese results suggested that ERK5 controls NRF2 expression through MEF2. To test this, we transfected a small interference RNA for MEF2 (siMEF2) in the hepatic cell line HepG2C3A (Supplemental Fig. 4D) and the AML cell line OCI-AML3 (Fig. 5E ). "is e#ciently decreased MEF2 mRNA and protein 29 and also decreased both basal and DCA-induced NRF2 mRNA (Fig. 5E ).

‹–‘ Š‘†”‹ƒŽ ‘’Ž‡š ƒ –‹˜‹–›•‹‰ƒŽ•ͻ‡š’”‡••‹‘Ǥ "e previous experiments had shown that OXPHOS generates a signal that induces ERK5 expression, which contributes to the NRF2-mediated antioxidant response. We con!rmed this in vivo by using a transgenic Tet-O% mouse that express a mutant active form of the ATPase Inhibitory Factor 1 (IF1) in hepatocytes to restrain OXPHOS in the liver 10 . Interestingly Santacatterina et al . describe in the Fig. 8C of their MS that liver Nrf2 levels are lower in mice expressing IF1. We con!rmed it by analyzing expression of Nrf2 mRNA (Fig. 6A ). "is correlated with lower expression of Erk5 mRNA as compared to wild type mice (Fig. 6A ). "is shows that OXPHOS also induces Erk5 mRNA expression in vivo . We con!rmed the essential role of mitochondrial function on ERK5 expression by using !broblasts derived from patients with strong mitochondrial disorders (Supplemental Table 1). ERK5 and NRF2 mRNA were signi!- cantly reduced in these patients (Fig. 6B ). We next focused on the molecular mechanisms underlying our observations. When we inhibited the mito- chondrial complex I with metformin, we observed a decrease on ERK5 , NRF2 and NQO-1 mRNA and protein expression (Fig. 6C ). Both metformin and DCA induce AMPK activation 27 , however, they blocked and induced ERK5 expression, respectively. "is suggested that AMPK and its associated metabolic changes were not involved in ERK5 expression. We con!rmed this by reducing the expression of the catalytic subunit of AMPK, AMPK α, with 2 di%erent siRNA that e%ectively blocked several AMPK-mediated metabolic changes 27 . "is did not a%ect expression of ERK 5 or NRF2 mRNA (Supplemental Fig. 5). "erefore, AMPK activation was not responsible for generating the antioxidant response in cells performing OXPHOS.

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 6 www.nature.com/scientificreports/

Figure 5. ERK5 controls NRF2 expression. ( A) 10 7 Jurkat-TAg cells were transfected with 5 µg of the empty pSUPER Neo vector or with this vector containing a small hairpin RNA for ERK5 (shERK5) or with a pcDNA vector expressing ERK5. Forty-eight hours later mRNA expression was analyzed by qPCR and represented as the % of mRNA compared to cells transfected with the control vector. (B) Cell transfected with control (Neo) or shERK5 were analyzed for protein expression by western blotting at 24 and 48 h post-transfection. Graphic bars show the NRF2/actin ratio of the depicted experiment. ( C) Primary human hepatocytes were double transfected with control siRNA or with siRNA against ERK5 (siERK5). 96 h later mRNA was collected and mRNA expression was analyzed by qPCR. ( D) 10 7 Jurkat-TAg cells were co-transfected with 5 µg of the following vectors ERK5 wild type, a constitutively active MEK5 mutant (MEK5D, M5), MEF2C and MEF2C with dominant negative function (MEF2DN) together with 2 µg of a luciferase reporter plasmid driven by the NRF2 promoter along with 1 µg of β-galactosidase expression vector. Cells were incubated in regular glucose media (gray bars) or containing 10 mM DCA (black bars) 24 h a*er transfection and analyzed 2 days later for luciferase

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 7 www.nature.com/scientificreports/

and β-galactosidase activities. "e graphic represents the relative luciferase units (RLU). ( E) OCI-AML3 cells were transfected with siRNA for MEF2A and C and 24 h later treated with 10 mM DCA for 36 h. NRF2 mRNA and NRF2 protein were analyzed as in in ( A) and ( B) respectively. Experiments were done in triplicate. "e data represent means ± SD; statistics were performed using student t-test ( C) or One-way ANOVA with post- hoc Tukey test ( A,D and E); *p < 0.05, ** p < 0.01, ***p < 0.001. Treatments were compared to empty vector transfected cells (control) if not speci!ed in the graph.

In Fig. 6C we showed that complex I inhibition decreased ERK5 mRNA. "e electron transport chain complex III removes electrons from ubiquinol (QH 2) and sequentially transfer them to cytochrome c. "e reduction of ubiquinone (Q) to QH 2 could either be due to mitochondrial complex I, which removed electrons from NADH, or mitochondrial complex II, which removed them from succinate and transferred through FAD. "en, we inves- tigated the e%ect of the complex II inhibitor thenoyltri$uoroacetone (TTFA). "is drug strongly induced ERK5 mRNA expression (Fig. 6D ). DCA did not increase TTFA e%ects suggesting that both shared the same target. TTFA was slightly toxic (Supplemental Fig. 6A) and, like metformin, could have o%-target e%ects. "erefore, we used an array of cell lines with impaired activity of the di%erent mitochondrial complexes (Supplemental table 2). ρ0 cells that lack mitochondrial DNA and thus a functional ETC, did not induce ERK5 expression a*er DCA treatment (Fig. 6E , right lower panel), in agreement with our previous results showing that mitochondrial activity induced ERK5 expression (Fig. 6A,B13 . We next used 3 di%erent cell lines in 2 di%erent mitochondrial backgrounds with defects in mitochondrial complex I and observed that DCA treatment did not induce ERK5 expression (Fig. 6E , le* panels). In agreement with Fig. 6D , mutation in mitochondrial complex II did not inhibit DCA-induced ERK5 expression (Fig. 6E , top right panel). Cells with mutations in complex III and V, but not in complex IV, increased ERK5 expression a*er DCA treatment (Fig. 6E , right panels). However, complex V mutant show lower basal ERK5 mRNA levels in agreement with in vivo experiments (Fig. 6A ). Mutation in the mitochondrial tRNA Ile in the L929 cell line (mB77), which produces more ROS 31 , did not increase ERK5 mRNA (Fig. 6E , upper right panel). "is sup- ported our results in Fig. 4 showing that de novo ROS production was not involved in ERK5 expression. Mitochondria adapt the organization of the di%erent complexes and supercomplexes to optimize the use of the available substrates, mainly regulating the proportion of respiratory complex III superassembled with com- plex I for electron transport. "is is needed to avoid competition between FADH 2- and NADH-derived elec- trons 31 . DCA, by inhibiting PDK1, activates PDH and the formation of acetyl-CoA from pyruvate. "is generates 3 NADH per 1 FADH 2 (through succinate) molecules in the TCA cycle. Other substrates however, generate a di%erent proportion of NADH/FADH 2 electrons and therefore a di%erent demand of CI/CII dependent oxidation. "erefore, while both complexes are always delivering electrons to the ETC simultaneously, the requirement of complex I seems relatively favored by DCA. Complex II, or succinate dehydrogenase (SDH), is also part of the Krebs cycle and catalyzes the conversion of succinate to fumarate. Hence, if complex II is outcompeted by com- plex I activity, succinate accumulation and fumarate reduction may be induced. Both phenomena are well known in intracellular signaling. Fumarate and succinate, in their acid form, acidify culture media. Hence, we used mon- omethylsuccinate (MMS) and dimethylfumarate (DMF) to investigate the impact of fumarate and succinate accu- mulation on ERK5 expression. MMS decreased ERK5 levels (Fig. 7A ). In contrast, DMF increased them (Fig. 7A ). Next, we used metformin to inhibit complex I, forcing the use of complex II, and added MMS to increase complex II activity. When used together, they decreased even further ERK5 expression (Fig. 7B ). "is suggested that com- plex II activity reduced ERK5 expression, probably by inhibiting complex I activity. In summary, whereas succi- nate probably does not play any role per se on ERK5 expression, fumarate induces its expression. "is suggested that DCA, by accumulating fumarate, induces ERK5 mRNA. Complex I receive electrons from NADH and deliver them to CoQ. Complex II employs FADH as co-factor to deliver electrons from succinate to CoQ.FADH 2. "erefore, the ratio of NADH/FADH 2 electrons changes with di%erent substrates and the requirement of complex I for NADH oxidation vary according to the NADH/FADH 2 ratio. Oxidative metabolism of one molecule of glucose generates ten NADH and two FADH2, a NADH:FADH2 electron ratio of 5. Fatty acids (FA), e.g. palmitate, generate a ratio of 2 32 . Etomoxir inhibits FA transport into the mitochondria and blocks fatty acid oxidation (FAO), resulting in an increase of the ratio NADH/FADH 2. Interestingly, etomoxir decreased basal and DCA-induced increase ERK5 mRNA (Fig. 7C ). "erefore, we found no correlation between ERK5 expression and the expected changes in NADH/FADH 2 ratio. Etomoxir and DCA were not toxic to OCI-AML3 cells, although they decreased cell proliferation (Supplemental Fig. 6B). However when combined they induced cell death suggesting that DCA treatment requires FAO for cell survival as sug- gested by our previous results 29 . In summary complex I activity, through accumulating fumarate, induces ERK5 expression leading to NRF2-mediated antioxidant response. Discussion ROS generation is inherent to the activity of the electron transport chain, with Complex I being considered one of the main sites at which premature electron leakage to oxygen occurs and give rise to superoxide anion 33 . We show here that complex I activity initiates an antioxidant response mediated by ERK5-induced NRF2 expression. It is interesting to note that the main generator of ROS is at the same time responsible of triggering the mechanism to eliminate them. Of relevance, ROS de novo production is not required for this response. "e cell “anticipates” ROS formation and activates the pathway to avoid their uncontrolled increase. Once produced, ROS quickly originate biochemical reactions that generate damage to cell structures. Hence, it is on the cell’s own bene!t to create the anti- oxidant response when ROS production is going to occur. However, new data are challenging the “only” deleterious view of mitochondrial ROS. For example ROS increase with aging, but increasing mitochondrial ROS production speci!cally through the respiratory complex I reverse electron transport (RET) extends Drosophila lifespan 34 .

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 8 www.nature.com/scientificreports/

Figure 6. Inhibition of mitochondrial complex I and II signals ERK5 expression. ( A) Erk5 and Nrf2 mRNA expression in the liver of wild-type and T/H (Tet-O%-H49K (h-IF1) mice. mRNA from 3 mice of each genotype was quanti!ed by qPCR and represented as the % of mRNA compared to wild-type mice. ( B) ERK5 and NRF2 mRNA expression in !broblasts derived from a group of 8 healthy donors or 8 patients su%ering from mitochondrial defects (Supplemental Table 1). ( C) Di%erent hematopoietic cell lines were incubated for 24 h with 5 mM metformin. mRNA expression was quanti!ed by qPCR and represented as the % of mRNA compared non-treated cells. ERK5 and NRF2 protein expression was analyzed in these cell lines by western blotting (lower panel). ( D) Jurkat and OCI-AML3 cells were treated with 10 mM DCA and 300 µM TTFA for 24 h. NRF2 mRNA expression was quanti!ed by qPCR and represented as the % of mRNA compared to control cells. ( E) Di%erent cell lines described in Supplemental Table 2 were treated with 20 mM DCA during 24 hours. ERK5 mRNA was quanti!ed by qPCR and represented as the % of mRNA compared to non-mutant control cells. Experiments were done in triplicate and data represent means ± SD; statistics were performed using student t-test ( A–C) or One-way ANOVA with post-hoc Tukey test ( D and E); *p < 0.05, ** p < 0.01, *** p < 0.001; §p < 0.05, §§ p < 0.01 compare to the respective control cell lines. Treatments were compared to non-treated cells (control) if not speci!ed in the graph.

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 9 www.nature.com/scientificreports/

Figure 7. Fumarate/succinate regulate ERK5 expression. OCI-AML3 were treated with di%erent drug combinations and the expression of ERK5 mRNA was analyzed by qPCR. ( A) OCI-AML3 cells were treated with 10 mM DCA, 5 mM MMS and/or 300 µM DMF for 24 h. (B) OCI-AML3 cells were treated with 5 mM metformin and/or 5 mM MMS for 24 h. (C) OCI-AML3 cells were treated with 5 mM DCA and/or 100 µM Etomoxir for 48 h. "e data represent means ± SD; *p < 0.05, ** p < 0.01, ***p < 0.005 ANOVA with post-hoc Tukey test. Treatments were compared to non-treated cells (control) if not speci!ed in the graph.

We show that mitochondrial complex I activity is required for ERK5-induced NRF2 expression. We have examined several possibilities that could account for our observation. DCA inhibits glycolysis and increases FAO as suggested by the high toxicity of the combined etomoxir plus DCA treatment ((Supplemental Fig. 6) and 29 ). "en, this switch would change NADH:FADH2 electron ratio. When electron $ux from FAD overwhelms the oxi- dation capacity of CoQFAD, CI is degraded, releasing CIII from CI-containing complexes to receive FAD-derived electrons 35 . "e increased electron $ux through FAD could saturate the oxidation capacity of the dedicated coen- zyme Q (CoQ) pool and result in the generation of ROS 36 . However, in our experiments ROS do not mediate OXPHOS-induced ERK5 expression (Figs 4 and 6). Moreover, etomoxir decreases basal ERK5 expression but it does not block DCA-induced increase. Finally, if FAO is inducing ERK5 expression, complex II inhibition should decrease it, but we found an increase with TTFA and basically no e%ect by genetic approaches. "is suggests that another mechanism is responsible for triggering ERK5 expression. In non-transformed cells inhibition of OXPHOS by IF1 induces AMPK activation 10 , which could lead to ERK5 activation 37 . DCA also induces AMPK 27 . However, our pharmacological (metformin, Fig. 6C ) and genetic (siRNA, Supplemental Fig. 5) approaches suggest that AMPK is not involved on ERK5 expression during OXPHOS. Strong mitochondrial complex I activity could decrease electron transport through complex II and the subsequent accumulation of succinate or reduced fumarate be responsible for ERK5 expression. Our results using TTFA or genetically-modi!ed cells support this conclusion (Fig. 6). However, MMS fails to induce ERK5

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 10 www.nature.com/scientificreports/

even if complex I was blocked by metformin (Fig. 7). In contrast, we found that DMF induced ERK5 (Fig. 7). Interestingly, DMF induces Nrf2 expression through a PD98059-sensitive pathway 38 . Although this MAPKK inhibitor was initially described as a speci!c MEK1 inhibitor, it also inhibits the ERK5 upstream kinase MEK5 39 . "erefore, fumarate indeed mediates ERK5 expression. Accumulated fumarate can covalently modify cysteine residues of proteins, in an uncatalyzed process termed succination, modifying cellular signaling 40 . Succination occurs on KEAP1 41 and results in constitutive NRF2 activation and increased expression of its target genes 41 . "erefore, fumarate induces the NRF2-mediated antioxidant response by directly a%ecting KEAP-1 41 and by inducing de novo NRF2 expression (our results). An alternative is that succinate promotes CII activity and induces RET thereby decreasing mitochondrial membrane potential. In reverse, fumarate blocks CII thereby increasing complex I activity and this could trigger ERK5 expression. Although it is well-established that ROS induce NRF2 activation 2, recent data support that alternative path- ways independent of ROS are also operative. For example, OXPHOS decreases KEAP-1 expression independently of ROS 11 and NRF2 expression in AML depends on NF- κB but not on ROS 30 . Interestingly, ERK5 activates NF- κB in leukemic cells 42 . Hence, ERK5 could handle the NRF2-mediated antioxidant response by at least 3 mechanisms independently of de novo ROS generation: i) direct transcription through MEF2 (the results presented here); ii) direct transcription through NF- κB30 ,42 ; iii) upregulation of miR-23 and downregulation of KEAP1 mRNA 11 . "is emphasizes the central role of ERK5 in the antioxidant response 11 –17 . Transcriptome analysis shows that the ERK5 pathway regulates in normoxia several genes involved in meta- bolic remodeling, including some controlled by hypoxia inducible factor-1 α (HIF-1 α under hypoxia 13 ,43 . Also, like HIF-1 α, ERK5 is degraded by a process depending on the tumor suppressor von Hippel-Lindau (VHL), through a prolyl hydroxylation-dependent mechanism 44 . Hence, mitochondrial complex I activity through fuma- rate accumulation could also protect ERK5 from VHL-induced degradation. "is is based on the fact that suc- cinate and fumarate (and succinate) outcompete α-ketoglutarate, an essential co-factor of prolyl hydroxylase domain enzymes 45 . How ERK5 induces NRF2 mRNA expression is not totally elucidated. ERK5 directly phosphorylates MEF2A, C and D at di%erent serines and threonines 46 ,47 . It activates MEF2A and D by direct interaction because ERK5 serves as a MEF2 coactivator through its signal-dependent direct association with the MEF2 MADS domain; although, at least, MEF2A-dependent transcription requires ERK5 kinase activity 48 ,49 . Finally, forcing cells to produce energy through OXPHOS also a%ects cell viability and proliferation independently of ROS. "is is rather related to energy depletion. In this sense OXPHOS requires mitochondrial function and DCA induces cell death in ρ0 cells, while in other cells it just inhibits growth 11 ,23 ,50 . In summary forcing OXPHOS in vitro is cytostatic in “normal” tumor cells and cytotoxic in cells with major mitochondrial dysfunctions. Experimental Procedures Ethical statement. Experimental procedures were conducted according to the European guidelines for animal welfare (2010/63/EU). Protocols were approved by the Animal Care and Use Committee “Languedoc- Roussillon” (approval number: CEEA-LR-12163). The use of human specimens for scientific purposes was approved by the French National Ethics Committee. All methods were carried out in accordance with the approved guidelines and regulations of this committee. Written informed consent was obtained from each patient prior to surgery.

Reagents and antibodies. DCA was from Santa Cruz Technologies. Galactose and glutamine were from GIBCO. RIPA bu%er to prepare protein extracts was from Euromedex. "e complete protease inhibitor cocktail (Complete EDTA-free) and the phosphatase inhibitor cocktail (PhosSTOP) were from Roche. H 2O2, DMF and MMS were from Sigma. ERK5 and NRF2 antibodies were from Cell Signaling Technology and Santa Cruz respec- tively. "e antibody against β-Actin and HRP-labeled secondary antibodies were from Sigma.

˜‹˜‘ mouse experiments. In vivo experiments were carried out using 6 to 8 weeks/old male NSG mice. Mice were bred and housed in pathogen-free conditions in the animal facility of the European Institute of Oncology– Italian Foundation for Cancer Research (FIRC), Institute of Molecular Oncology (Milan, Italy). For engra*ment of human cells, 1 million AML cells were injected intravenously (i.v.) through the lateral tail vein in non-irradiated mice. NSG mice with established human AML tumors (day 80 post-gra*) were treated with DCA (50 mg/kg, 1 dose/ day by gavage, starting at day 1 for 16 consecutive days). Human tumor AML cells gather in mouse spleen and bone marrow, hence we isolated mRNA from these organs. We used human-speci!c primers to visualize expression of human mRNA. In a di%erent experiment B6 wild type mice were treated with a daily single dose of DCA (50 mg/kg/ day) intraperitoneally and mouse mRNA was analyzed in spleen and liver a*er di%erent times.

Š ͷ”ƒ•‰‡‹ ‹ ‡Ǥ "e samples from transgenic mice containing the mutant H49K version of hIF1 have been described 10 . mRNA was analyzed in liver of these mice.

Cell lines and culture conditions. "e leukemic human cell lines T Jurkat Tag, NB4 and OCI-AML3 were grown in RPMI 1640–Glutamax (GIBCO) supplemented with 5% (Jurkat) or 10% (OCI and NB4) FBS. Primary cells from a lymphoma B cell patient (BCL-P2) were grown in the same medium with 10% FBS. In certain experiments cells were grown in RPMI 1640 without glucose (GIBCO 11879) with the addition of 2 mM glutamine and 10 mM galactose (OXPHOS medium). "e Jurkat TAg cells carry the SV40 large T Ag to facilitate cell transfection. HepG2C3A and HuH7 cells were grown in MEM and DMEM respectively supplemented with 10% FBS, sodium pyruvate, glutamine, penicillin and streptomycin. "e HCT116 human colon cancer cells were cultured in low glucose (5 mM) DMEM medium supplemented with 10% FBS. Cellular con$uence during experiments was between 80–85%.

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 11 www.nature.com/scientificreports/

Primary Leukemic Cells. Data and samples from patients with di%erent hematological cancers were col- lected at the Oncology and Clinical Hematology Department of the CHU Montpellier, France, a*er patient’s informed consent. Patients were enrolled in two independent clinical programs approved by the “Comite ́s de Protection des Personnes Sud Me diterrané é I (ref 1324)” and ID-RCB: 2011-A00924-37. All samples from cancer patients were collected at diagnosis.

Human liver samples and preparation of primary human hepatocytes (PHHs) cultures. PHHs were isolated as described previously 51 from donor organs unsuitable for transplantation or from liver resec- tions performed in adult patients for medical reasons unrelated to our research program. Liver samples were obtained from the Biological Resource Center of Montpellier University Hospital (CRB-CHUM; http://www. chu-montpellier.fr ; Biobank ID: BB-0033-00031) and this study bene!tted from the expertise of Dr Jeanne Ramos (hepatogastroenterology sample collection) and Prof Sylvain Lehmann (CRB-CHUM manager). "e procedure was approved by the French Ethics Committee and written or oral consent was obtained from the patients or their families. Human hepatocytes isolation and culture were performed as described previously 51 . Brie$y, a*er liver per- fusion, hepatocytes were counted and cell viability was assessed by trypan blue exclusion test. A suspension of 1 × 10 6 cells/mL per well was added in 12-well plates pre-coated with type I collagen (Beckton Dickinson) and cells were allowed to attach for 12 h. "en, the supernatant containing dead cells and debris was removed and replaced with 1 mL of serum-free long-term culture medium (Lanford medium, LNF). "e number of con$uent attached cells was estimated at ~1.5 × 10 5 cells/cm 2.

Plasmids. "e luciferase reported plasmid driven by a DNA fragment of 1.5 kb of the human NRF2 promoter was a kind gi* from Stuart Rushworth 30 . "e expression vectors for ERK5, the pSUPER expression vector for GFP alone or GFP plus shERK5 and the pSiren-retroQ-puro (BD Biosciences) retroviral vectors for shERK5 and control have been previously described 52 . Control, MEF2A and C and ERK5 siRNA were ON-TARGETplus SMARTpools (mixture of 4 siRNA) from Dharmacon.

Transient transfection. Jurkat cells in logarithmic growth phase were transfected with the indicated amounts of plasmid by electroporation 42 ,53 . In each experiment, cells were transfected with the same total amount of DNA by supplementing with empty vector. Cells were incubated for 10 min at RT with the DNA mix and elec- troporated using the Gene Pulser Xcell ™ Electroporation system (Bio-Rad) at 260 mV, 960 mF in 400 µl of RPMI 1640. Expression of the di%erent proteins was con!rmed by western blot. "e transfection e#ciency in Jurkat TAg cells is between 60 and 80%. OC-AML-3 cells were transfected using Amaxa TM D-Nucleofector TM Lonza Kit according to manufactured protocol. In HuH7 and HCT116 cells, transfection of 30–50 nM siRNAs was carried out using Lipofectamine RNAiMAX (Invitrogen) in Opti-MEM (Invitrogen), according to the manufacturer’s instructions. Adherent primary hepatocytes were transfected twice at day !rst and third post-seeding with 20 nM siRNA. Cells were harvested 48 to 96 h post-transfection.

Reporter assay. In all experiments, Jurkat cells were transfected with β-galactosidase reporter plasmid 42 ,53 . "e transfected cells were harvested a*er 2 days and centrifuged at 1000 g for 5 min. "e cell pellet was suspended in 1 ml cold PBS and transferred to 1.5 ml Eppendorf tube for washing. Cells were lysed with 100 µl luciferase lysis bu%er (Promega) and incubated at room temperature for 10 min. "e lysates were centrifuged and luciferase assays (40 µl) performed according to the manufacturer’s instructions (Promega, Charbonnières, France) using a Berthold lumi- nometer. For β-Galactosidase assays, 40 µl of lysates were added to 200 µl of β-Galactosidase assay bu%er (50 mM phosphate bu%er pH 7.4; ONPG 200 µg; 1 mM MgCl2; 50 mM β-Mercaptoethanol) and the absorbance measured at 405 nm. "e results were expressed as luciferase units normalized to the corresponding β-galactosidase activity. "e expression level of the transfected proteins was routinely control by immunoblot analysis.

Subcellular fractionation. For preparation of nuclear extracts, Jurkat cells were grown in indicated medium. Ten million cells were taken and washed twice in cold PBS 54 . Nuclear and cytoplasmic proteins were extracted using according to manufactured instruction of Bio Basic Inc ®. Extracted soluble proteins were ana- lyzed by immunoblotting.

Counting and determination of cell viability. Cell number, viability and cell death was analyzed with the Muse Cell Analyzer (Millipore) by incubating cells with Muse Count & Viability and Annexin V and Dead Cell kits respectively, following manufacturer’s instructions 27 .

—‘ƪ—‘”‡• ‡–ƒ••ƒ›Ǥ Control or treated cells were washed with cold bu%er and !xed with paraform- aldehyde (3.2% in PBS) for 20 minutes. Cells were washed 3 times with PBS and stored at 4 °C until labelling. Cells were permeabilized with Triton (0.1% in TBS) for 5 minutes and washed with TBS-T (TBS + Tween 0.05%). Cells were labelled with primary antibody (for one hour at room temperature (dilution in TBS + 2% SVF) and washed with TBS-T. Cells were labeled with secondary antibody and Hoechst or DAPI (1/1000 dilution in TBS + 2% SVF) for 30 minutes. Cells were washed with TBS-T and !nally washed with H 2O before montage. Immuno$uorescent labeling was examined under a $uorescent microscope (Leica Microsystem, Rueil-Malmaison, France) and images were analyzed using Metamorph so*ware (Universal Imaging Corporation, Downington, PA).

RT-PCR. Total RNA was extracted using NucleoSpin RNA isolation columns (Macherey-Nagel), reverse tran- scription was carried out using iScript ™ cDNA Synthesis Kit (Biorad). Quantitative PCR was performed with KAPA

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 12 www.nature.com/scientificreports/

SYBR Green qPCR SuperMix (Cliniscience) and a CFX Connect ™ Real-Time qPCR machine (Biorad) with ERK5, NRF2, NQO1, HO-1 and actin primers. Supplemental Table 3 shows all primers used in this study. All samples were normalized to β-actin mRNA levels. Results are expressed relative to control values arbitrarily set at 100 27 . Immunoblotting. Protein analysis by immunoblotting was performed essentially as previously described 27 . Brie$y, samples were collected, washed out with PBS and lysed with RIPA bu%er. Protein concentration was deter- mined by BCA assay (Pierce) before electrophoresis in 4–15% TGX gels (BioRad) and equal amount of protein was loaded in each well. Protein transfer was performed in TransTurbo system (BioRad) in PVDF membranes. A*er blocking for 1 h with 5% non-fat milk, membranes were incubated overnight at 4 °C in agitation with pri- mary antibodies, washed three times with PBS-Tween 0,1% and incubated with the appropriate HRP-labeled secondary antibody for 1 h. Membranes were washed out three times with PBS-Tween 0,1% and developed with Substrat HRP Immobilon Western (Millipore). Band quanti!cation was performed using the “ImageLab” so*- ware from BioRad and represented as the ratio between the protein of interest and a control protein i.e. actin. "e value of 1 is arbitrarily given to control cells. One blot representative of several experiments is shown. Statistical analysis. "e statistical analysis of the di%erence between means of paired samples was per- formed using the paired t test. Multiple comparisons were performed using One-way ANOVA with post-hoc Tukey HSD test. "e results are given as the con!dence interval ( *p < 0.05, ** p < 0.01, *** p < 0.005). All the experiments described in the !gures with a quantitative analysis have been performed at least three times in duplicate. Other experiments were performed three times with similar results. We used actin as a loading control and the histograms represent the ratio (value of protein of interest)/(value of actin). References 1. Willems, P. H., Zossignol, Z., Dieteren, C. E., Murphy, M. P. & [oopman, W. J. Zedox Homeostasis and Mitochondrial Dynamics. Cell Metab 22 , 207–218 (2015). 2. [ensler, T. W. & Wa\abayashi, N. Nrf2: friend or foe for chemoprevention? Carcinogenesis 31 , 90–99 (2010). 3. [raj\a-[uznia\, V., Paluszcza\, J. & Baer-Dubows\a, W. "e Nrf2-AZE signaling pathway: An update on its regulation and possible role in cancer prevention and treatment. Pharmacological reports: P! 69 , 393–402 (2017). 4. Scarpulla, Z. C. Nuclear control of respiratory gene expression in mammalian cells. J Cell Biochem 97 , 673–683 (2006). 5. [ovac, S. et al . Nrf2 regulates ZOS production by mitochondria and NADPH oxidase. Biochim Biophys Acta 1850, 794–801 (2015). 6. Din\ova-[ostova, A. T. & Abramov, A. Y. "e emerging role of Nrf2 in mitochondrial function. Free !adic Biol Med 88 , 179–188 (2015). 7. Strom, J., Xu, B., Tian, X. & Chen, Q. M. Nrf2 protects mitochondrial decay by oxidative stress. FASEB J 30 , 66–80 (2016). 8. Muthusamy, V. Z. et al . Acute exercise stress activates Nrf2/AZE signaling and promotes antioxidant mechanisms in the myocardium. Free !adic Biol Med 52 , 366–376 (2012). 9. Wang, P., Li, C. G., Qi, Z., Cui, D. & Ding, S. Acute exercise stress promotes Zef1/Nrf2 signalling and increases mitochondrial antioxidant activity in s\eletal muscle. Experimental physiology 101, 410–420 (2016). 10. Santacatterina, F. et al . Down-regulation of oxidative phosphorylation in the liver by expression of the ATPase inhibitory factor 1 induces a tumor-promoter metabolic state. Oncotarget 7, 490–508 (2016). 11. [han, A. U. et al . Human Leu\emic Cells performing Oxidative Phosphorylation (OXPHOS) Generate an Antioxidant Zesponse Independently of Zeactive Oxygen species (ZOS) Production. EBioMedicine 3, 43–53 (2016). 12. Zhao, J. et al . Big mitogen-activated protein \inase 1 protects cultured rat aortic smooth muscle cells from oxidative damage. Journal of pharmacological sciences 116, 173–180 (2011). 13. Charni, S. et al . Oxidative phosphorylation induces de novo expression of the MHC class I in tumor cells through the EZ[5 pathway. J Immunol 185, 3498–3503 (2010). 14. Lopez-Zoyuela, N. et al . Extracellular-signal-regulated \inase 5 modulates the antioxidant response by transcriptionally controlling Sirtuin 1 expression in leu\emic cells. Int J Biochem Cell Biol 53 , 253–261 (2014). 15. Abe, J., [usuhara, M., Ulevitch, Z. J., Ber\, B. C. & Lee, J. D. Big mitogen-activated protein \inase 1 (BM[1) is a redox-sensitive \inase. J Biol Chem 271, 16586–16590 (1996). 16. [im, M. et al . Laminar $ow activation of EZ[5 protein in vascular endothelium leads to atheroprotective e%ect via NF-E2-related factor 2 (Nrf2) activation. J Biol Chem 287, 40722–40731 (2012). 17. Nigro, P., Abe, J. & Ber\, B. C. Flow shear stress and atherosclerosis: a matter of site speci!city. Antioxid !edox Signal 15 , 1405–1414 (2011). 18. Consortium, E. P. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012). 19. Mysic\ova, A. & Vingron, M. Detection of interacting transcription factors in human tissues using predicted DNA binding a#nity. BMC genomics 13 (Suppl 1), S2 (2012). 20. Nagar, S. et al . MEF2D haploinsu#ciency downregulates the NZF2 pathway and renders photoreceptors susceptible to light-induced oxidative stress. Proc Natl Acad Sci USA 114, E4048–E4056 (2017). 21. Villalba, M. et al . From tumor cell metabolism to tumor immune escape. Int J Biochem Cell Biol 45 , 106–113 (2013). 22. Villalba, M. et al . Chemical metabolic inhibitors for the treatment of blood-borne cancers. Anti-cancer agents in medicinal chemistry 14 , 223–232 (2014). 23. Catalán E, et al . MHC-I modulation due to metabolic changes regulates tumor sensitivity to CTL and N[ cells. Oncoimmunology 4, (2015). 24. Zathore, M. G. et al . "e NF-\appaB member p65 controls glutamine metabolism through miZ-23a. Int J Biochem Cell Biol 44 , 1448–1456 (2012). 25. Zeitzer, L. J., Wice, B. M. & [ennell, D. Evidence that glutamine, not sugar, is the major energy source for cultured HeLa cells. J Biol Chem 254, 2669–2676 (1979). 26. Zossignol, Z. et al . Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. Cancer !es 64 , 985–993 (2004). 27. Allende-Vega, N. et al . "e presence of wild type p53 in hematological cancers improves the e#cacy of combinational therapy targeting metabolism. Oncotarget 6, 19228–19245 (2015). 28. Catalán, E. et al . MHC-I modulation due to metabolic changes regulates tumor sensitivity to CTL and N[ cells. Oncoimmunology 4, e985924 (2015). 29. [han, A. U. H. et al . "e PD[1 Inhibitor Dichloroacetate Controls Cholesterol Homeostasis "rough the EZ[5/MEF2 Pathway. Scienti"c reports 7, 10654 (2017). 30. Zushworth, S. A. et al . "e high Nrf2 expression in human acute myeloid leu\emia is driven by NF-\appaB and underlies its chemo- resistance. Blood 120, 5188–5198 (2012).

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 13 www.nature.com/scientificreports/

31. Moreno-Loshuertos, Z. et al . Evolution meets disease: penetrance and functional epistasis of mitochondrial tZNA mutations. PLoS genetics 7, e1001379 (2011). 32. Speijer, D. Oxygen radicals shaping evolution: why fatty acid catabolism leads to peroxisomes while neurons do without it: FADH(2)/NADH $ux ratios determining mitochondrial radical formation were crucial for the eu\aryotic invention of peroxisomes and catabolic tissue di%erentiation. Bioessays 33 , 88–94 (2011). 33. Lenaz, G. et al . Mitochondrial Complex I: structural and functional aspects. Biochim Biophys Acta 1757, 1406–1420 (2006). 34. Scialo, F. et al . Mitochondrial ZOS Produced via Zeverse Electron Transport Extend Animal Lifespan. Cell Metab 23 , 725–734 (2016). 35. Lapuente-Brun, E. et al . Supercomplex assembly determines electron $ux in the mitochondrial electron transport chain. Science 340 , 1567–1570 (2013). 36. Guaras, A. et al . "e CoQH2/CoQ Zatio Serves as a Sensor of Zespiratory Chain E#ciency. Cell reports 15 , 197–209 (2016). 37. Young, A. et al . Flow activation of AMP-activated protein \inase in vascular endothelium leads to [ruppel-li\e factor 2 expression. Arteriosclerosis, thrombosis, and vascular biology 29 , 1902–1908 (2009). 38. Wang, Q. et al . Dimethyl Fumarate Protects Neural Stem/Progenitor Cells and Neurons from Oxidative Damage through Nrf2- EZ[1/2 MAP[ Pathway. Int J Mol Sci 16 , 13885–13907 (2015). 39. [ama\ura, S., Moriguchi, T. & Nishida, E. Activation of the protein \inase EZ[5/BM[1 by receptor tyrosine \inases. Identi!cation and characterization of a signaling pathway to the nucleus. Journal of Biological Chemistry 274, 26563–26571 (1999). 40. Ternette, N. et al . Inhibition of mitochondrial aconitase by succination in fumarate hydratase de!ciency. Cell reports 3, 689–700 (2013). 41. Ooi, A. et al . An antioxidant response phenotype shared between hereditary and sporadic type 2 papillary renal cell carcinoma. Cancer Cell 20 , 511–523 (2011). 42. Garaude, J. et al . EZ[5 activates NF-\appaB in leu\emic T cells and is essential for their growth in vivo . J Immunol 177 , 7607–7617 (2006). 43. Schweppe, Z. E., Cheung, T. H. & Ahn, N. G. Global gene expression analysis of EZ[5 and EZ[1/2 signaling reveals a role for HIF- 1 in EZ[5-mediated responses. J Biol Chem 281, 20993–21003 (2006). 44. Arias-Gonzalez, L. et al . EZ[5/BM[1 is a novel target of the tumor suppressor VHL: implication in clear cell renal carcinoma. Neoplasia 15 , 649–659 (2013). 45. Smol\ova, [. et al . Waves of gene regulation suppress and then restore oxidative phosphorylation in cancer cells. Int J Biochem Cell Biol 43 , 950–968 (2011). 46. [ato, Y. et al . BM[1/EZ[5 regulates serum-induced early gene expression through transcription factor MEF2C. EMBO Journal 16 , 7054–7066 (1997). 47. [ato, Y. et al . Big mitogen-activated \inase regulates multiple members of the MEF2 protein family. Journal of Biological Chemistry 275, 18534–18540 (2000). 48. [asler, H. G., Victoria, J., Duramad, O. & Winoto, A. EZ[5 is a novel type of mitogen-activated protein \inase containing a transcriptional activation domain. Mol Cell Biol 20 , 8382–8389 (2000). 49. Yang, C. C., Ornats\y, O. I., McDermott, J. C., Cruz, T. F. & Prody, C. A. Interaction of myocyte enhancer factor 2 (MEF2) with a mitogen-activated protein \inase, EZ[5/BM[1. Nucleic Acids !es 26 , 4771–4777 (1998). 50. Stoc\win, L. H. et al . Sodium dichloroacetate selectively targets cells with defects in the mitochondrial ETC. Int J Cancer 127 , 2510–2519 (2010). 51. Pichard, L. et al . Human hepatocyte culture. Methods Mol Biol 320, 283–293 (2006). 52. Charni, S. et al . EZ[5 [noc\down generates mouse leu\emia cells with low MHC class i levels that activate N[ cells and bloc\ tumorigenesis. J Immunol 182, 3398–3405 (2009). 53. Garaude, J. et al . SUMOylation regulates the transcriptional activity of JunB in T lymphocytes. J Immunol 180, 5983–5990 (2008). 54. Garaude, J., [amins\i, S., Cherni, S., Hips\ind, Z. A. & Villalba, M. "e Zole of EZ[5 in T-Cell Signalling. Scand J Immunol 62 , 515–520 (2005). Acknowledgements We thank Dr. Robert A. Hipskind for his helpful comments regarding this manuscript. FACs analysis and microscopy were performed at the platform Montpellier Rio Imaging (MRI). "e collection of clincial data and samples (HEMODIAG_2020) at the CHRU Montpellier was supported by funding from Région Languedoc Roussillon. All our funders are public or charitable organizations. "is work was supported by a fellowship from the Higher Education Commission, Pakistan (AK) and fellowships from the Ministère de l’Enseignement Supérieur et de la Recherche (MESR) (DNV). Author Contributions A.K., N.A.-V., D.G., S.G., C.G., D.-N.V., S.O., G.T., perform experiments. J.G., M.D., C.D., F.B., A.A., J.-M.C., J.-A.E., G.C., C.-H.L., J.H. and M.V. wrote the main manuscript text. All authors reviewed the manuscript. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-23884-4 . Competing Interests: "e authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional a#liations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. "e images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

© "e Author(s) 2018

SCIENTIFIC REPORTS | (2018) 8:7420 ȁ ǣͷͶǤͷͶ͹;Ȁ•ͺͷͻͿ;ǦͶͷ;Ǧ͸͹;;ͺǦͺ 14 ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

! |ÑÖ! ) ) ) ) ) ) ) ) ) !

!'".1#/!ddZ!

!

&#,#/"*!"(0!200(-,!

!

!

! ! ! ! ! ! ! ! ! ! ! ! ! )

! |Ö{! 2@=),));=DD)AK)9F)AFL=?J9D)H9JL)G>)L@=)@ME9F)AEEMF=)KQKL=Eg)HD9QAF?)H9JLA;MD9J) AEHGJL9FL)JGD=K)AF);GFLJGD)G>)LMEGJ)<=N=DGHE=FL)9F<)NAJMK)AF>=;LAGFj)'F<==) AEEMF=) =>>=;LGJKg) AF;DM) ;9F;=J) t&9F9@9F) 9F<) 5=AF:=J?)Ä}~~h)4ADD9D:9)=L)9Dj)Ä}~Åu)L@9L)AK)K@9J=<):Q)NAJMKj)'F)L@=K=)AEEMFGkH9L@GDG?Q) ;GF) ,)) ;=DDK) LG) L9J?=L) ;=DDK) t=AL@=J) LMEGJ) GJ) NAJMK) AF>=;L=<) ;=DDKu)J=KMDL)AF),));=DD)HGHMD9LAGFK)) EMDLAHD=) AEEMF=) =N9KAGF) KLJ9L=?A=K)9F<)EMDLAHD=)EG<=K)G>)AFL=J9;LAGF):=LO==F),));=DD)9F<)L@=AJ)L9J?=L);=DDKg)L@=) H9L@GDG?A;9DkJ=D9L=<) E9JC=JK) GF) ,)) ;=DD) HGHMD9LAGF) 9J=) @9J<) LG) :=) ?=F=J9DAR=AJKL)H9JL)G>)EQ)L@=KAK)9LL=EHL=<)LG)AFN=KLA?9L=)L@AK):Q)<=L9AD=<)AEEMFGkH@=FGLQHAF?) 9HHJG9;@=K) GF) ,)) ;=DD) HGHMD9LAGFKj) 5=) KLM>=J=FL) !"ÇÉ) AKG>GJEK) AF) N9JAGMK) @=E9LGDG?A;9D) E9DA?F9F;A=K) t 15/;>3<# JDLu) 9F<) AF) NAJMK) AF>=;LAGF) t&'4) s) 15/;>3<# M uj) 2@=)K=;GF<)H9JL)G>)L@=)L@=KAK)HJGNA<=<)9F)=>>A;A=FL)E=L@G<)G>)?=F=J9LAF?)9DDG?=F=A;),)) ;=DD) OAL@) @A?@DQ) 9;LAN9L=<) HGL=FLA9D) >GJ) 9FLAk;9F;=J) AEEMFGL@=J9HQj) 2@=K=) =PH9F<=<) ,)) ;=DDK) ;9F) :=) MK=<) 9K) 9) KAF?D=k9?=FL) EGFGL@=J9HQ) GJ) AF) ;GE:AF9LAGF) OAL@) =AL@=J) ;DAFA;9Dk9HHJGN=<) EGFG;DGF9D) 9FLA:G)LMEGJ);=DDK)LG)=F@9F;=)L@=),)k<=H=F<=FL)9FLAkLMEGJ)=>>=;Lj)2@=) GJ)=9;@)K=;LAGF)G>)L@=)J=KMDLK)9J=)HJ=K=FL=<)9;;GJ)L@AK) >J9E=OGJCj) ! kZd!&"*#,!KLM7Q! ! 2@AK) DGF?ALM) 9) H@9K=) ':r'') ;DAFA;9D) LJA9D) OAL@) J=D9HK=J9;LGJQ) k,&*) H9LA=FLKg) O@A;@) 9J=) >GDDGO=<) OAL@AF) L@=) LAE=>J9E=) G>) Ñ) ;Q;D=K)tAF)O@A;@)=9;@);Q;D=)AK)9HHJGPAE9L=DQ)~)EGFL@u)OAL@);GE:AF9LAGF)LJ=9LE=FL)G>) *#,)9F<)- 8)tK==) ,?;;73839>/3<# K g)>GJ)LJ=9LE=FL)9F<)K9EHDAF?) HJGLG;GDuj) *#,) :=DGF?K) LG) L@=) ;D9KK) G>) AEEMFGEG>=;LK) G>) *#,) AF;DM<=)HJGEGLAF?)AEEMF=)=>>=;LGJ)9;LAN9LAGFg)=KH=;A9DDQ)GF)AEEMF=)HGHMD9LAGFK)G>)2) 9F<) ,)) ;=DDK) t%9A<9JGN9) =L) 9Dj) Ä}}Üh) *9?JM=) =L) 9Dj) Ä}~Éh) 8@9F?) =L) 9Dj) Ä}}áuj) ,GL9:DQg) ;GE:AF9LAGF)G>)*#,)9F<)JALMPAE9:)t026u)LJ=9LE=FL)@9K):==F)K@GOF)LG)AEHJGN=)F"!!)

! |Ö|! >B8 I>GEC )E=A;A9D) KQF=J?AKLA;) =>>=;L) G>) L@=K=) LOG) 9FLAkLMEGJ) 9?=FLK)t5M)=L)9Dj)Ä}}Üuj)'F)L@AK)KLMA=<):Q)<=>M;GKQD9LGF)>GJ)@A?@=J):AF>AFALQ)LG) $;Q0'''9g)O@A;@)AK)=PHJ=KK=<)E9AFDQ)GF),));=DDK)9F<)E9;JGH@9?=Kj),));=DD)<=HD=LAGF) >GDDGO=<) :Q) - 8) AFGJ) !**) H9LA=FLK) t%9J;Z9k+M^GR) =L) 9Dj) Ä}~Üuj) 'F) GMJ) ;G@GJLg) O=) G:K=JN=<) 9) LJ9FKA=FL) ,)) <=HD=LAGF)GFDQ)9L)L@=)=F<)G>)L@=)>AJKL);Q;D=)9>L=J)Å);GFK=;MLAN=)- 8)AF>MKAGFK)t $64#H #D# !5/;>3<# K uj),=N=JL@=D=KKg)*#,)@9K):==F)J=HGJL=<)LG)9M?E=FL),));=DD)HJGDA>=J9LAGF)AF) K=N=J9D) HJ=NAGMK) KLM>=;L)G>)LJ=9LE=FL)GF),));=DD)HGHMD9LAGF)AK)DAC=DQ)9)F=L)GML;GE=)G>) :GL@)KAF?D=)9?=FLnK)=>>=;Lj)+GJ=GN=Jg)H9LA=FL),));=DDK)<=;J=9K=<)AF)AEE9LMJ=)!"ÉÑ:J) HGHMD9LAGF) O@AD=) AF;J=9K=<) AF) !"ÉÑ3<#K uj)2@AK)=>>=;L)@9K):==F)KAEAD9JDQ)J=HGJL=<):Q) 9) HJ=NAGMK) KLM) L@=) 9;LAN9LAGF) E9JC=J) !"Ñá) 9F<) L@=) <=?J9FMD9LAGF) E9JC=J) !"~}Ö9) GF) LGL9D) ,)) ;=DD) HGHMD9LAGF) 9L) L@=) =F<) G>) >AF9D) ;Q;D=) MHGF) L@=) <=;J=9K=) AF) L9J?=L) ;=DD) FME:=JK)tAF3<#K uj)2@AK)J=KMDL) =EH@9KAR=<)L@=);DGK=)J=D9LAGFK@AH):=LO==F),));=DD)9;LANALQ)AF)NANG)9F<)L@=)HJ=K=F;=)G>) L9J?=L) ;=DDKj) +GJ=) AEHGJL9FLDQg) ,)) ;=DD) AF) N9JAGMK) @=E9LGDG?A;9D) GJE) HJG>AD=i) @A?@=J) !"ÇÉ0F0-ã) 9F<) !"ÇÉ0-ãg) DGO=J) oF9\N=p) !"ÇÉ0Fã) H=J;=FL9?=) ;GEH9J=<) LG) &") t !5/;>3<# J uj) 5=) ;GF>AJE=<) GMJ) HJ=NAGMK) J=KMDLK) AF) L@AK) KLML=J) L9J?=L) ;=DD) =DAEAF9LAGF) 9L) L@=) =F<) G>) LJ=9LE=FL) t $64# J # E# !5/;>3<# K uj) $AF9DDQg) O=) K@GO=<) L@9L) !"~á)=PHJ=KKAGF)GF),));=DD)HGHMD9LAGF)AK)EGKL)HJ=3<# J uj) 'F) L@AK) KLM)=9;@)

! |Ö}! kZe!"+*!KLM7Q! ! .9LA=FLK) OAL@) 9;ML=) EQ=DGA<) D=MC=EA9) tF+*u) 9J=) LQHA;9DDQ) ;@9J9;L=JAR=<) OAL@) 9;;MEMD9LAGF) 9F<) J9HA<) HJGDA>=J9LAGF) G>) AEE9LMJ=) @=E9LGHGA=LA;) EQ=DGA<) HJ=;MJKGJK) CFGOF) 9K) EQ=DG:D9KLKj) 2@=) KL9F<9J<) LJ=9LE=FL) >GJ) L@=K=) E9DA?F9F;A=K) AK) LQHA;9DDQ) ;@=EGL@=J9HQ) OAL@) 9FL@J9;Q;DAF=K) 9F<) ;QL9J9:AF=) 9K) >AJKLkDAF=) LJ=9LE=FLKg) G>L=F) >GDDGO=<) :Q) @=E9LGHGA=LA;) KL=E) ;=DD) LJ9FKHD9FL9LAGF) t&1!2u) AF) H9LA=FLK) OAL@) AFL=JE==9KA:D=j) ,)) ;=DD) >JGE) F+*) H9LA=FLK) MKM9DDQ) MF;LAGF9DALA=K) t!GKL=DDG)=L)9Dj)Ä}}Çh)$9J9?)9F<)!9DA?AMJA)Ä}}Ñuj)2G)?9AF):=LL=J)MF<=JKL9F)L@=),)) ;=DD)H@=FGLQH=)AF)F+*)H9LA=FLK)9>L=J);@=EGL@=J9HQ)LJ=9LE=FL)t!2ug)O=)9F9DQK=<),)) ;=DD) H@=FGLQH=K) AF) Ö) H9LA=FLK) 9L) Ä) LAE=HGAFLKi) 9L) L=J) !2) t!5/;>3<#L uj)2G)>9;ADAL9L=)L@=)?9LAF?)HJG;=KK)9F<)9NGA<)L@=)HGL=FLA9D)EAKkAFL=HJ=L9LAGF) G>)!"ÉÑã)EQ=DGAGJ),));=DDK)9K) L@AK)E9JC=J)AK)J=KLJA;L=<)LG)L@=)2r,)kGJA?AF)DAF=9?=)t $64#LCH#E#!5/;>3<#L uj)Å})<9QK)9>L=J) !2g)L@=)D=MCG;QL=)H9LL=JF)AK)H9JLDQ)J=KLGJ=3<# L uj)'FL=J=KLAF?DQg)H9LA=FLnK) ,));=DD)>JGE)GMJ)KLM)!"ÑÄ*g)!"ÉÑ:J),))HGHMD9LAGF) O@AD=)J=L=J)!2) t$64#LCJ#!D"B#$64#LCL# #E#!5/;>3<#L uj)2@=K=)J=KMDLK)9J=)KGE=O@9L)KAEAD9J)LG)9)HJ=NAGMK) J=HGJL) O@=J=) 9) KA?FA>A;9FL) @A?@=J) !"ÉÑ:J) ,)) ;=DD) HJGHGJLAGF) 9>L=J) !2) @9K) :==F) G:K=JN=<)t"9M?M=L)=L)9Dj)Ä}~~uj)!@J=LA=F)9F<);GDD=9?M=K)9DKG)<=K;JA:=<)9F)9;;MEMD9LAGF) G>)L@=)D9L=)E9LMJ=),))H@=FGLQH=);GJJ=KHGFIC )<=?J9FMD9LAGF)9F<)DGO=J)."k~)D=N=D)AF)L@=)O@GD=),));=DD) ;GEH9JLE=FLg)KM??=KLAF?)9)D=KK)9;LAN9L=<)H@=FGLQH=)9>L=J)<=;J=9K=)AF)LMEGJ);=DD)DG9<) t$64# LCJDB# $64# LCK # E# !5/;>3<# L uj) $AF9DDQg) O=) ;GF>AJE=<) L@9L) !"ÇÉ0F0-ã) ,)) ;=DDK) ;GJJ=D9L=)OAL@)LMEGJ)DG9<)9K)AF)GMJ)HJ=NAGMK)KLM3<# KB# $64# LCN# D# !5/;>3<# Luj) $AF9DDQg)!"~Çg):ML)FGL)!"ÅÅg)9KKG;A9L=K)KLJGF?DQ)OAL@)L@=)9EGMFL)G>)!"ÇÉ0F0-ã),)) ;=DDK) t $64# LCM# E# !5/;>3<# Luj) %AN=F) L@9L) !"~Ç) AK) 9) E9JC=J) KH=;A>A;) >GJ) EGFG;QL=rE9;JGH@9?=) HGHMD9LAGFKg) @GO) L@=) !"ÇÉ0F0-ã) ,)) ;=DDK) HJ=>=J=FLA9DDQ) AFL=J9;LK) OAL@) L@=K=) HGHMD9LAGFK) :ML) FGL) !"ÅÅk=PHJ=KKAF?) EQ=DGA<) HJ=;MJKGJK)

! |Ö~! J=E9AF=<) LG) :=) AFN=KLA?9L=) H9LA=FL) FME:=JK) 9F<) D9;C)G>)&");GFLJGDKg)L@AK)KLMMJL@=J)AFN=KLA?9LAGFKj) ) kZf!'(3!KLM7Q! ! 2@=)JGD=)G>),));=DD)AF)&'4)AF>=;LAGF)@9K):==F)O=DD)=KL9:DAK@=) ,)) ;=DD) AF) F"!!k<=H=F<=FL) ;QLGLGPA;ALQ) :Q) 9FLAk&'4) 9FLA:G3<# HD# HCMg)FDL=J)=L)9Dj) Ä}}Öh) !G@=F) =L) 9Dj) ~áááuj) #E=J?AF?) KLM=;LAGF) ;9MK=K) <9EH=F=<),));=DD)@GE=GKL9KAK)9F<)9:=JJ9FL),));=DD)J=;=HLGJ)J=H=JLGAJ=)t+ACMD9C)=L) 9Dj)Ä}~Öh)1;MDDQ)9F<)FDL=J)Ä}~Ñuj),));=DDK)>JGE)H9LA=FLK)O@G)@9<)DGF?)L=JE)=PHGKMJ=)LG) &'4)NAJ=EA9)MF;LAGF9D) !"ÉÑF=?) ,)) ;=DDK) 9J=) ;GEHJGEAK=<) AF) ;QLGLGPA;ALQ) ;GEH9J=) LG) ALK) !"ÉÑ=;LAN=)AF);QLGCAF=)HJG) &'4) AF>=;LAGF) GF) ,)) ;=DD) HGHMD9LAGF) AF) ?=F=J9D) 9F<) AF) L@=) =PHJ=KKAGF) G>) !"ÇÉ) AKG>GJEK) AF) H9JLA;MD9Jg) O=) 9F9DQK=<) :DGG<) K9EHD=K) >JGE) H9LA=FLK) OAL@) :GL@) &'4) NAJ=EA9) 9F<) FGFkNAJ=EA9) tAj=g) 9NAJ=EA9u)tJ=KMDLK)HJ=K=FL=<)AF) !5/;>3<# M uj)'F)L@=K=)H9LA=FLKg)O=)G:K=JN=<)9)):GL@),))9F<),)2);GEH9JLE=FLK);GEH9J=<)LG)&"g)O@A;@)KM??=KL=<)&'4) AF>=;LAGF) E9Q) @9N=) 9F) AEH9;L) GF) ,)) ;=DD) @GE=GKL9KAK) t $64# MCI# E# !5/;>3<# M uj) 'FL=J=KLAF?DQg),));=DD)HGHMD9LAGF)>JGE)GMJ);G@GJL)G>)H9LA=FLK)K@GO=<)E9JCK)G>),)%Ä") J=3<# M uj) 1AF;=) ,)%Ä")=;LAGF) J=KMDL=<)AF)@A?@=J)D=N=DK)G>),)%Ä"*)AF)H9LA=FLK);GEH9J=<)LG)&"j),)%Ä"*K)HJ=K=FL)AF) Ä)>GJEKi)E=E:J9F=):GMF<)tMKM9DDQ)GF)KLJ=KK=<);=DDKg)Aj=)NAJMK)AF>=;L=<);=DDKu)GJ)KGDM:D=j) +'!Fg)9),)%Ä"*g)AF)ALK)KGDM:D=)>GJE)@9K):==F)J=HGJL=<)9L)@A?@)D=N=D)>JGE)HD9KE9)G>) &'4)H9LA=FLKj)2@AK);GJJ=D9L=K)OAL@)<=>=;LK)AF),)%Ä"kE=JGE) &'4kAF>=;L=<) H9LA=FLK) t+9NADAG) =L) 9Dj) Ä}}Éh) ,GDLAF?) =L) 9Dj) Ä}~}uj) 2@=) !"ÇÉ) AKG>GJE)

! |ÖÄ! HJG>AD=) G>) ,)) ;=DDK) <=JAN=<) >JGE) GMJ) H9LA=FLK) J=N=9D=<) 9) AD=) ;GEH9J=<) LG) &"i)@A?@=J)HJGHGJLAGFK)G>)!"ÇÉ0FJGE) &") 3<# M uj) 'EHGJL9FLDQg) L@=K=) &'4kKH=;A>A;) !"ÇÉ0F)!"~Ñ)AF)O@A;@)!"ÇÉ0-ã),));=DD)>JGE):GL@)9NAJ=EA;)9F<) NAJ=EA;)H9LA=FLK)9J=)EGKLDQ)!"ÉÑ3<# M uj)1@=<)!"~Ñ) GF)!"ÉÑ),))MF<=JO=FL)HJGDGF?=<)KLAEMD9LAGF)t0GE==)=L) 9Dj) Ä}~Åuj) 2@=J=>GJ=) AL) AK) L=EHLAF?) LG) KH=;MD9L=) L@9L) L@=) =F@9F;=<) !"ÇÉ0FJGE) ,));=DD)9F<)NAJMK)AF>=;L=<);=DD)AFL=J9;LAGFK) >B8I>IC j)&GO=N=J)AL)AK)KLADD)HGKKA:D=)L@9L)!"~Ñ) MF;LAGF9D) HJG;=KKj) F) >MJL@=J) AEEMFGH@=FGLQHAF?)GF),));=DDK)AK)MF<=J?GAF?j) ) ,)) ;=DDK) 9J=) 9:D=) LG) ;9HLMJ=) L9J?=L) ;=DDK) 9FLA?=FK) MHGF) L@=AJ) AFL=J9;LAGF) NA9) LJG?G;QLGKAKj) 5=) <=EGFKLJ9L=<) >B8 I>GEC ) L@9L) 9DDG?=FA;) HJAE9JQ) ,)) ;=DDK) 9J=) 9:D=) LG) H=J>GJE=<) LJG?G;QLGKAK) GF) !"Çã) 2) ;=DD) DAF=K) t $64# MCM# E# !5/;>3<# M uj) FK) 9) J=KMDLK) G>) LJG?G;QLGKAKg) ,)) ;=DDK) ;9JJA=<) &'4k=FLJQ) J=;=HLGJKg) Aj=j) !"Ç) 9F<) !!0Ég) O@A;@) ;GMD<) K=FKALAR=)L@=E)LG)KM:K=IM=FL)&'4)AF>=;LAGFj)&'4);9F)AF>=;L)FGL)GFDQ)!"Çã)2);=DDK):ML) 9DKG) E9;JGH@9?=K) t&GF=Q;MLL) =L) 9Dj) Ä}~Ñuj) -MJ) J=KMDLK) KM??=KL) L@9L) ,)) ;=DD) ;GMD<) :=;GE=)9)&'4kL9J?=L)9>L=J)LJG?G;QLGKAK)G>)2);=DD)9FLA?=FK)>JGE)&'4kAF>=;L=<);=DDKj)2@AK) AK)KLADD)MF<=J?GAF?)AFN=KLA?9LAGFj)) ) kZg! "#&!%$# !,)!68DD!8PH4FK?GF!4F7!?EH46L!G8J4HQ! ! !9F;=J);=DDK)HJ=>=J)H=J>GJE)?DQ;GDQKAK)AFKL=9<)G>)-6.&-1)tAj=)59J:MJ?)=>>=;Luj) 2@AK)HJGEGL=K)AF=>>A;A=FL):ML)J9HA<)F2.)?=F=J9LAGF)9F<):J=9C)E9;JGEGD=;MD=K) O@A;@)K=JN=)9K)AFL=JE=GJ)LMEGJ)E9KK)9F<)MDLAE9L=DQ)9DDGOAF?)L@=E)LG)E9AFL9AF) J9HA<) 9) HJGDA>=J9LAGF) J9L=) t*GH=Rk0GQM=D9) =L) 9Dj) Ä}~Çh) 4ADD9D:9) =L) 9Dj) Ä}~Åuj) ') @9N=) H9JLA;AH9L=<) GF) K=N=J9D) AFN=KLA?9LAGFK) D=9) N9JAGMK) 9KH=;LK)G>);9F;=J);=DD)E=L9:GDAKEj)FEGF?)L@=Eg)O=)K@GO)L@9L)

! |ÖÅ! 9) .")~) AF@A:ALGJg) HJGEGL=K) EALG;@GF9LLQ) 9;A<) E=L9:GDAKE) NA9) 9F) #0)Ék+#$Äk<=H=F<=FL) MHJ=?MD9LAGF) G>) L@=) ;@GD=KL=JGD) J=;=HLGJ)!"ÅÑ)t !5/;>3<# O uj)+=L9:GDAKE)9F<)AEEMFG?=FA;ALQ)AF);9F;=J);=DDK)9J=)LOG) >=J=FL) 9KH=;LK) L@9L) @9K) :==F) AF;J=9KAF?DQ) J=;G?FAR=<) LG) :=) EMLM9DDQ) AFL=?J9L=<) t!@9JFA)=L)9Dj)Ä}~}h)*GH=Rk0GQM=D9)=L)9Dj)Ä}~Çh)4ADD9D:9)=L)9Dj)Ä}~Åuj)+G);9F;=J) ;=DD)E=L9:GDAKE)NA9)E=L>GJEAFg)9)EALG;@GF)E=L>GJEAF) 9F<)"!F)OAL@),));=DD)=PH9F<=<) >B8I>GEC )>GJ)>MLMJ=)9FLAk;9F;=J)LJ=9LE=FLK)tE9FMK;JAHL) AF)HJ=H9J9LAGFuj) ) "M=) LMEGJkE=D9EE9LAGF) MKM9DDQ) 9KKG;A9L=<) OAL@) L@=) LMEGJ) EA;JG=FNAJGFE=FLg) =F>=;LGJKg) AF;DMMF;LAGF) t3<# HB# HCNuj) 'L) AK) =KK=FLA9D) LG) J=N=JK=) L@AK) :9D9F;=) LG) >9NGJ) 9) KLJGF?=J) 9FLAkLMEGJ) AEEMF=)J=KHGFK=j) 'B8I>GEC )?=F=J9LAGF)G>)@A?@DQ)9;LAN9L=<),));=DDK)>JGE)3! )MKAF?),)) ;=DD) HJGEGLAF?) ;QLGCAF=K) AF) ;GE:AF9LAGF) OAL@) L@=) # 4k*!*) ;=DD) DAF=K) KM;;=KK>MDDQ) GN=J;GE=K) L@=) 9FLAk9HGHLGLA;) =>>=;L) G>) ;@=EGkJ=KAKL9FL) @=E9LGDG?A;9D) ;9F;=J) ;=DDK) t1TF;@=Rk+9JLZF=R)=L)9Dj)Ä}~Ég)Ä}~Ñuj)'F)L@AK);MJJ=FL)KLMMJL@=J)HJGNA<=<):9KAK) >GJ) L@=) 9HHDA;9LAGF) G>) L@AK) ,)) ;=DD) =PH9FKAGF) HJGLG;GD) AF) ;GE:AF9LAGF) OAL@) ;DAFA;9Dk 9HHJGN=<)EF:)LGO9J)LMEGJ)9FLA?=FKg)AF;DM3<# N uj) 29C=F) LG?=L@=Jg) ,)) ;=DD) ?=F=J9L=<)>B8 I>GEC ) MKAF?) L@=) E=L@G<) <=K;JA:=<) AF) !@9HL=J) Ö) HJGNA<=<) 9) HJGKH=;L) >GJ) MKAF?) ,)) ;=DDK) AF) AEEMFGL@=J9HQ) :GL@) 9K) 9) KAF?D=) 9?=FL) GJ) OAL@)EF:)LG)=F@9F;=)F"!!j) ) ) ) ) ) )

! |ÖÇ! kZh!0MEE4JQ!`!JGE) k,&*) H9LA=FLK) 9>L=J) *#,) 9F<) - 8) LJ=9LE=FLK) J=IC ) 9;LAN9LAGF)E9JC=JKg)=j?j)!"Ñág)!"~}Ö9)9F<)DGO=J)!"ÇÉ0-ã)KM:HGHMD9LAGFKj) • $MJL@=JEGJ=g)%F*#,)LJ=9LE=FL)JGE) L@=K=) H9LA=FLK) H=J>GJE=<) @A?@) LJG?G;QLGKAK) 9;LANALQ) 9L) ) !"~á) 9F<) L@AK) E9JC=J) <=;J=9K=<) AF) ,)) ;=DD) ;GEH9JLE=FL) O@=F) L9J?=L) ;=DDK) t!"~áã) ) ;=DDKu) 9J=) =DAEAF9L=<) 9L) L@=) =F<) G>) LJ=9LE=FLj) ) F+*)KLMJGE) F+*) H9LA=FLK) MF<=J?GF=) !2) L=J)!2) LJ=9LE=FLj) • 'FL=J=KLAF?DQg) L@=) !"ÇÉ0F0-ã) KM:K=L) KLJGF?DQ) ;GJJ=D9L=K) OAL@) !"~Ç) LJG?G;QLGKAK) E9JC=J) 9L) MF;LAGF9D)@9DDE9JCj) • +GJ=GN=Jg) ,)) ;=DDK) >JGE) L@=K=) H9LA=FLK) =P;DMKAN=DQ) AF;DM<=) !"ÇÉ0F)&'4k=FLJQ)J=;=HLGJK)tAj=j)!"Ç)9F<)!!0Éu)>JGE) >B8I>GEC );Gk AF;M:9LAGF)=PH=JAE=FLK)KM??=KLAF?),));=DD);9F):=;GE=)9)FGN=D)&'4)L9J?=L)9K)9) ;GFK=IM=F;=)G>)AFL=J9;LAGF)OAL@)&'4ã)!"ÇkL9J?=L);=DDKj) ) 'F) KMEE9JQg) L@=) OGJC) GF) 9F9DQRAF?) !"ÇÉ) AKG>GJEK) GF) ,)) ;=DD) KM:HGHMD9LAGF)

! |ÖÉ! >JGE)N9JAGMK)@=E9LGDG?A;9D);9F;=JK)KM;@)9K) k,&*)t!@9HL=J)Çu)9F<)F+*)t!@9HL=J)Éu) J=N=9D=<) =NA<=F;=K) >GJ) ,)) ;=DD) MF;LAGF) 9L) GJE) HJG>AD=K) GF) ,)) ;=DDK) >JGE) L@=K=) H9L@GDG?A=K) >=J=FLA9L=) L@=E) >JGE) @=9DL@Q) H=GHD=j) &=F;=g)!"ÇÉ0-ã),));=DDK)9J=)H9L@GDG?A;9DDQ)J=D=N9FL)9F<)9KKG;A9L=<)LG);D=9J9F;=)G>) LMEGJ) ;=DDK) tAj=j) *#,) HDMK) - 8g) !@9HL=J) Çuj) ,)) ;=DDK) >JGE) &'4kAF>=;L=<) H9LA=FLK) ;GFL9AF) KM:HGHMD9LAGFK) ;@9J9;L=JAR=<) :Q) !"ÇÉ0FB8 I>IC j) 2@=K=) ,)) ;=DDK)>JGE)&'4ã)H9LA=FLK)9DKG)K@GO)DGO=J),)%Ä")9;LAN9LAF?)J=;=HLGJg)O@A;@);GMD<):=)9) NAJMKkAF) >MF;LAGF9D) AEH9AJE=FLj) &=J=AFg) OAL@AF) MABL! MA>LBL! PHKD! '! :==BMBHG:EER!=>L=!:!G>P!F>i!MKH@HEE!PAB!BG!L>O>K:E!IARLBHEH@B<:E!IKH<>LL>Li!2ABL!PHNE=!;KBG@BG@!GHO>E!?NG! ,)! <>EE! =>I>G=BG@! HG! MA>! LI>QM! PA>K>! BM! M:D>! IE:<>Lh! Bi>i! <:IMNKBG@!H?!MNFHKj:LLH=!:GMB@>GL!GMB:EER!F:KDL!MA>!MKH@H=!,)!<>EE! MH!?K:MKB!PABE>:L!BGM>K:=!<>EEL!PHNE=!?:!MK:GL?>KKBG@!H?! OBKNLj>GMKR!K><>IMHKL!MH!MA>!MKH@H=!,)!<>EEL!PABER!!MA>F!:L! :! GHO>E! OBKNL! M:K@>Mi! 2A>L>! K>LNEML! <:G! ;>! QIEHBM>=! ?HK! F:DBG@! I:MB>GMzL! IKH@GHLMBKA:IL!:ELH!MH!B=>GMB?R!,)!<>EE!?:M>!PA>G!MA>R!NIM:D>!M:K@>MzL!:GMB@>GLi ) ) $AF9DDQg) ,)) ;=DD) =PH9FKAGF) >JGE) 3! ) @9K) :==F) HJGN=F) =>>A;A=FL) AF) ;GE:AF9LAGF) OAL@)K=N=J9D)EGFG;DGF9D)9FLA:GGJ)LMEGJ)9FLA?=FKj)2@AK)J=KMDLK)AF)KLJGF?)F"!!) >B8 I>GEC )9F<) >B8I>IC j)2@AK)OGJC)KMHHGJLK)L@=)MK=)G>)=PH9F<=<),));=DDK)AF)AEEMFGL@=J9HQj) ! kZi!%MLMJ8!H8JKH86L?N8K! ! .@=FGLQHA;) HJG>ADAF?) G>) ,)) ;=DD) KM:K=LKg) =j?j) :9K=<) GF) =PHJ=KKAGF) G>) !"ÇÉ) AKG>GJEKg) ;GMD<) <=L=;L) AEEMFGH9L@GDG?Q) ;GF=;LAGFKj)2@AK)K@GMD<):=)N9DA<9L=<)AF)>MJL@=J)KLM>A;A=FL) E=L@G<) G>) ?=F=J9LAF?) @A?@DQ) 9;LAN9L=<) 9DDG?=F=A;) ,)) ;=DD) >B8 I>GEC ) >JGE) 3! ) ;GMD<) HGL=FLA9DDQ) GN=J;GE=) L@=) =FJGE);9F;=J)H9LA=FLK)9F<)J=?9AF)=>>A;A=FL),));=DDk<=H=F<=FL)9FLAkLMEGJ) =>>=;Lj)#EHAJA;9D)J=KMDLK)>JGE)!@9HL=J)Ö)KMHHGJL)L@=)MK=)G>)L@=K=) >B8I>GEC )9;LAN9L=<),))

! |ÖÑ! ;=DDK)9K)=AL@=J)EGFGL@=J9HQ)GJ)AF);GE:AF9LAGF)OAL@);DAFA;9D)EF:K)>GJ);GF>=JJAF?)KLJGF?) F"!!);QLGLGPA;ALQj)2@AK)K@GMD<)OA<=DQ)KLMMLMJ=)EG<=DKj) !

!

!

!

!

!

!

!

!

!

!

!

!

! !

! |ÖÖ! /8<8J8F68! ) ) ) ) FDL=Jg)%9DAL)=L)9Dj)Ä}}Öj)o"A>>=J=FLA9D),9LMJ9D))ADD=J)!=DDk+=)&'4k~) 0=HDA;9LAGF) 9K=<)GF)"AKLAF;L))'0r&*F)1M:LQH=Kjp) 1=:8(CHEB6@8C;8#KD:E>A:BG6@8 +:9>8>B: )Ä}Çt~ÄuiÅ}ÄÖsÅÑj)) ) FDL=Jg)%9DAL)=L)9Dj)Ä}}Éj)o1=IM=FLA9D)"=J=?MD9LAGF)G>),))!=DD)1M:K=L)"AKLJA:MLAGF)9F<) $MF;LAGF)1L9JLAF?)AF)F;ML=)&'4k~)'F>=;LAGFjp) @CC9 )~}Ñt~}uiÅÅÑÑsÑáj)) ) FF>GKKAg),A;GD9K)=L)9Dj)Ä}}Ñj)o&ME9F),))!=DD)#GJ) +&!)!D9KK)'jp) 'AAHB>GL )ÄÉtÄuiÅÅ~sÇÄj) ) FJ9K=g)&jg)#)!QLGE=?9DGNAJMK):Q)F;LAN9LAF?)9F<)'F@A:ALGJQ),)) !=DD)0=;=HLGJKjp) 08>:B8: )ÄáÑtÉÉÖ~ui~ÅÄÅsÄÑj)) ) FJFGFg)29D)'jg)%9D)+9JC=Dg)9F<)->=J)+9F<=D:GAEj)Ä}}Ñj)o2MEGJ)9F<)4AJ9D)0=;G?FALAGF) :Q),9LMJ9D))ADD=J)!=DDK)0=;=HLGJKjp) 0:A>B6EF8>B8!6B8:E8 >C@CGJ) ,))!=DD)!QLGLGPA;ALQjp) 1=:8(CHEB6@8C;8'AAHBC@C),))!=DDK)9F<)2)!=DDK):Q),)%Ä"g)9)0=;=HLGJ)>GJ)1LJ=KKk 'F:B8: )ÄÜÉtÉÇÄÜuiÖÄÖsÄáj)) ) =J?g)+9JA9)=L)9Dj)Ä}}áj)o!DAFA;9Dk%J9<=)#P)4ANGk#PH9F<=<)&ME9F),9LMJ9D))ADD=J)!=DDK) MHk0=?MD9L=)F;LAN9LAF?)0=;=HLGJK)9F<)"=9L@)0=;=HLGJ)*A?9F=;LAGFK)AF)9F) F8>B: ) ÅÄ}tÄÑui~ÖÅ~sÅÉj)) ) BaJCKLJaEg),ACD9K))jg)&9FKk%MKL9>)*BMF??J=Fg)9F<)(9CG:)+A;@9YDKKGFj)Ä}~Ñj) o#E=J?AF?)'FKA?@LK)AFLG),9LMJ9D))ADD=J)!=DDK)AF)&ME9F).=JAH@=J9D)2AKKM=Kjp) ,6GHE:8 /:I>:JF8'AAHBC@C)#9JDQ),9LMJ9D))ADD=J)!=DD)"=N=DGHE=FLjp) 'AAHBC@C<>88/:F:6E8= ) Ç}tÅui~áÅsÄ}Öj)) ) aLL;@=Jg)(9F).j)=L)9Dj)Ä}~Üj)o,))!=DDK)1LAEMD9L=)0=;JMALE=FL)G>);"!~)AFLG)L@=)2MEGJ) +A;JG=FNAJGFE=FL).JGEGLAF?)!9F;=J)'EEMF=)!GFLJGDjp) !:@@ )~}ÄÄsÅÖj) )

! }{{! GLLAFGg)!j)=L)9Dj)Ä}}}j)oFF9DQKAK)G>)L@=)+GD=;MD9J)+=;@9FAKE)'FNGDN=<)AF)Ä Çk +=GJ)F;LAN9LAGF)G>)2)!=DDKjp) #HECD:6B8 (CHEB6@8C;8'AAHBC@CA;9LAGF)G>).40)t!"~ÉÉu)9F<),=;LAFkÄ)t!"~~Äu)9K) !=DD)1MJ>9;=)*A?9FGJ)L@=)&ME9F)",F+k~)t!"ÄÄÑu)F;LAN9LAF?)+GD=;MD=jp) 1=:8 (CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )~áÜtÇuiÉÉÖsÑÖj)) ) GN=FK;@=Fg),A=DK)9F<)(jFD9AF))MEE=Jj)Ä}~}j)o-JH@9F)%J9FRQE=K)$AF<)9)&GE=jp) 'AAHBC@C<>86@8/:I>:JF )ÄÅÉt~ui~~ÖsÄÖj)) ) J9FGJ)L@=)F;LAN9LAF?),9LMJ9D))ADD=J)!=DD)0=;=HLGJ),)HÅ})AF)&ME9FKjp) 1=:8(CHEB6@8C;8 #KD:E>A:BG6@8+:9>8>B: )Ä}ÑtÖui~ÇáÉs~É}Åj)) ) J9M)'F@A:ALGJQ)'FHML))'F)*BMF??J=Fg)9F<)#JA;)-j)*GF?j)Ä}}Ñj) o1QF=J?Q)9EGF?)0=;=HLGJK)GF)0=KLAF?),))!=DDK)>GJ)L@=)F;LAN9LAGF)G>),9LMJ9D) !QLGLGPA;ALQ)9F<)!QLGCAF=)1=;J=LAGFjp) @CC9 )~}Öt~ui~ÉásÑÑj)) ) JQ;=KGFg)7=F9F)2jg)+A;@9=D)#j)+9J;@g)"GEAF?G)$j) 9J:=Jg)&9FKk%MKL9>)*BMF??J=Fg) 9F<)#JA;)-j)*GF?j)Ä}}Éj)o!QLGDQLA;)%J9FMD=).GD9JAR9LAGF)9F<)"=?J9FMD9LAGF) !GFLJGDD=<):Q)"A>>=J=FL)0=;=HLGJK)AF)0=KLAF?),))!=DDKjp) 1=:8(CHEB6@8C;8 #KD:E>A:BG6@8+:9>8>B: )Ä}ÄtÖui~}}~s~Äj)) ) JQ;=KGFg)7=F9F)2jg)+A;@9=D)#j)+9J;@g)&9FKk%MKL9>)*BMF??J=Fg)9F<)#JA;)-j)*GF?j)Ä}}Ñj) oF;LAN9LAGFg)!G9;LAN9LAGFg)9F<)!GKLAEMD9LAGF)G>)0=KLAF?)&ME9F),9LMJ9D))ADD=J) !=DDKjp) 'AAHBC@C<>86@8/:I>:JF )Ä~Çt~uiÖÅsá~j)) ) MJF=Lg)$j)+j)~áÖ}j)o2@=)!GF;=HL)G>)'EEMFGDG?A;9D)1MJN=ADD9F;=jp) .ECB8 #KD:E>A:BG6@81HACE8/:F:6E8= )~Åi~sÄÖj)) ) !9JKGFg)5j)#j)=L)9Dj)~ááÖj)oF).GL=FLA9D)0GD=)>GJ)'FL=JD=MCAFk~É)AF)L@=)0=?MD9LAGF)G>) &ME9F),9LMJ9D))ADD=J)!=DD)1MJNAN9Djp) (CHEB6@8C;8!@>B>86@8'BI:FG><6G>CB )áátÉuiáÅÖsÇÅj)) ) !=J:GFAg)!JAKLAF9)=L)9Dj)Ä}}Öj)o&ME9F)'EEMFG<=>A;A=F;Q)4AJMK)~),=>).JGL=AF) "GOFEG)L@=)F;LAN9LAF?)0=;=HLGJ),)%Ä")9F<)'F@A:ALK) ,9LMJ9D))ADD=J)!=DDk+=EC@C

! }{|! !@9DMHFQg),j(9Fg)FFFA=)0=AFk5=KLGFg)1L=H@9FA=)"GK;@g)9F<)"9NA<)!GKE9Fj)Ä}}Ñj) o"GOFk0=?MD9LAGF)G>)L@=),)%Ä")*A?9F<)+'!F):Q)L@=)&ME9F)!QLGE=?9DGNAJMK) %DQ;GHJGL=AF)3*~ÇÄjp) >C8=:A>86@86B98 >CD=LF>86@8/:F:6E8=8!CAAHB>86G>CBF ) ÅÇÑt~ui~ÖÉsÜ~j)) ) !@9DMHFQg),j(9Fg)!D9AJ=)*j)1ML@=JD9FGJ)&ME9F),)%Ä"jp) >C8=:A>86@8 6B98 >CD=LF>86@8/:F:6E8=8!CAAHB>86G>CBF )Å}Ét~ui~ÄásÅÉj)) ) !@9JFAg)1j)=L)9Dj)Ä}~}j)o-PA<9LAN=).@GKH@GJQD9LAGF)'F)L@=) +&!)!D9KK)')AF)2MEGJ)!=DDK)L@JGM?@)L@=)#0)É).9L@O9Qjp) 1=:8(CHEB6@8C;8 'AAHBC@C)L@=)1*F+)0=;=HLGJjp) +C@:8H@6E86B98!:@@H@6E8 >C@C:JF8"EH<8">F8CI:EL )~ÇtÖuiÇÜÖsáÜj)) ) !@J=LA=Fg)FFF=k1GH@A=)=L)9Dj)Ä}~Éj)o'F;J=9K=<),))!=DD)+9LMJ9LAGF)AF).9LA=FLK)OAL@) F;ML=)+Q=DGA<)*=MC=EA9jp) $ECBG>:EF8>B8'AAHBC@C)L@=)&*Fk!Äk1H=;A>A;)F;LAN9LAF?))ADD=Jk!=DD)'?k DAC=)0=;=HLGJ))'0Ä"1~)GF),))9F<)2)!=DDKjp) !@>B>86@8'AAHBC@C)!D9KK)')+9BGJ) &AKLG;GEH9LA:ADALQ)!GEHD=P).JGL=AFK):Q)&'4k~).JGL=;LK)&'4k'F>=;L=<)!=DDK)>JGE) ,))!=DDKjp) 'AAHB>GL )~}tÑuiÑÑ~sÖ~j)) ) !GDM;;Ag)$J9F;=K;Gg)+A;@9=D)Fj)!9DA?AMJAg)9F<)(9E=K).j)"A)19FLGj)Ä}}Åj)o5@9L)"G=K)'L) 29C=)LG)+9C=)9),9LMJ9D))ADD=Jlp) ,6GHE:8/:I>:JFP8'AAHBC@C)!"ÉÑã) !QLGLGPA;)!=DDK)>JGE)&ME9F)3E:ADA;9D)!GJ<) DGGA:BG6@8&:A6GC@CGJ)9)"=H=F<=F;=)GF)'FL=JD=MCAF)~É)>GJ) 1MJNAN9D)G>),9LMJ9D))ADD=J)!=DDKjp) @CC9 )~}}t~}uiÅÑÅÅsÅÜj)) ) !GGH=Jg)+=?9F)Fjg)2G<<)Fj)$=@FA?=Jg)9F<)+A;@9=D)Fj)!9DA?AMJAj)Ä}}~j)o2@=) AGDG?Q)G>) &ME9F),9LMJ9D))ADD=Jk!=DD)1M:K=LKjp) 1E:B9F8>B8'AAHBC@C

! }{}! 0=;=HLGJjp) 'AAHB>GL )~ÇtÄui~ÄÅsÅÅj) ) !GKL=DDGg)0W?AK)2jg)!QJAD)$9MJA9Lg)(WJ`E=)0=Qg)(=9FkFD:=JL)%9KL9MLg)9F<)"9FA=D)-DAN=j) Ä}}Çj)o'EEMFG:AGDG?Q)G>)&9=E9LGDG?A;9D)+9DA?F9FL)"AKGJ<=JKi)2@=) 9KAK)>GJ) ,GN=D)'EEMFGL@=J9HQ).JGLG;GDKjp) 1=:8*6B8:GP8-B8C@C) GF=)+9JJGO)%J9>L)0=B=;LAGFj)'j)"=L=JEAF9FLk 1H=;A>A;)"A>>=J=F;=)G>)0=9;LANALQ)AF)2OG).9AJK)G>)'F:J=<)+GMK=)1LJ9AFKjp) 1=:8 (CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )~ÅÇt~uiÄÜ~sáÅj)) ) !MDD=Fg)1=9F).j)9F<)1=9EMK)(j)+9JLAFj)Ä}~Éj)o$9K)9F<)20F'*)m"=9L@)0=;=HLGJKn)9K) 'FALA9LGJK)G>)'F>D9EE9LAGFi)'EHDA;9LAGFK)>GJ)!9F;=Jjp) 0:A>B6EF8>B8!:@@86B98 ":I:@CDA:BG6@8 >C@C)'EE9LMJ=),9LMJ9D))ADD=J) !=DDK)AF)F;ML=)+Q=DGA<)*=MC=EA9).9LA=FLK)AF)$AJKL)!GEHD=L=)0=EAKKAGFjp) @A:E>86B8 (CHEB6@8C;8&:A6GC@C:B8:8X,:J85CE?N8 ,P5PY )ÅÇÜtÑÄÅ}ui~ÅÑsÅáj)) ) "=FE9Fg)!=;=D=)(j)=L)9Dj)Ä}~Äj)o+=E:J9F=k GMF<)'*kÄ~).JGEGL=K)1MKL9AF=<)#P)4ANG) .JGDA>=J9LAGF)G>)&ME9F),9LMJ9D))ADD=J)!=DDKjp) .*C08-,# )Öt~uj) ) "=K9Ag)+9<@9Ng))9L=),=O:=JJQg)8j)8@AK@MGg)+A;@9=D)59F?g)9F<)*A9F?)8@9F?j)Ä}~Çj) o*=F9DAJ9;LGJQ)+9FLD=)!=DD)*QEH@GE9i)-N=JNA=O)9F<) .=JKH=;LAN=jp) 1=:E6D:HG>88@9I6B8:F8>B8&:A6GC@C=F:9;@g)FF),)%Ä"jp) ,6GHE:8 'AAHBC@C)'F>=;L=<)"=FGJ),))!=DD)F;LAN9LAGF)AF)L@=)&ME9F)0=KHGFK=)LG)'F>DM=FR9) 'F>=;LAGFjp) (CHEB6@8C;8'AAHBC@CACE:N8+9P O8[_]ZY )~ÖÜtÉuiÄÑÜÜsáÜj)) ) #AKK=FKg)"A9F9),j)=L)9Dj)Ä}~Äj)o"=>AFAF?)#9JDQ)&ME9F),))!=DD)"=N=DGHE=FL9D)1L9?=K)AF) .JAE9JQ)9F<)1=;GF<9JQ)*QEH@GA<)2AKKM=Kjp)=)1j)9F<)+A;@9=D)Fj)!9DA?AMJAj)Ä}}Ñj)o'EEMFGDG?A;)FHHJG9;@=K)LG)F;ML=) *=MC=EA9)AF)L@=)#D<=JDQjp) 0:A>B6EF8>B8&:A6GC@C

! }{~! $9MJA9Lg)!QJADg)+9JLAF)9)'N9JKKGFg)&9FKk?MKL9>)*BMF??J=Fg))9JDkBG@9F)+9DE:=J?g)9F<) (9CG:)+A;@9=j)Ä}~}j)o#)&ME9F),9LMJ9D))ADD=J)!=DDK):Q)F;LAN9LAF?))ADD=J) !=DD)'EEMFG?DG:MDAFkDAC=)0=;=HLGJKjp) #9H86G>CB )~~ÉtÑui~~ÑÑsÖÇj) ) $=JFTF<=Rk+=KKAF9g)*GD9)=L)9Dj)Ä}~}j)o"A>>=J=FLA9D)+=;@9FAKEK)G>)1@=<)L@=) %DQ;GKQDH@GKH@9LAC@C<>86@8!=:A>FGEL )ÄÜÉt~ÄuiÜÉÇÅsÉ~j) ) $=JF9F<=Rg),9),9LMJ9D))ADD=J)!=DDK)F;@A=N=K)1=D>k 2GD=J9F;=)OAL@GML)#PHJ=KKAF?)'F@A:ALGJQ)0=;=HLGJK)1H=;A>A;)>GJ)1=D>k+&!) +GD=;MD=Kjp) @CC9 )~}Ét~~uiÇÇ~ÑsÄÅj) ) $DGJ=Kk4ADD9FM=N9g).j)-j)=L)9Dj)Ä}}~j)o!GFLJGD)G>)&'4k~)4AJ=EA9)9F<).JGL=;LAGF)>JGE) F'"1)FJ=)FKKG;A9L=<)OAL@)&*Fk OÇ)&GEGRQ?GKALQjp) .EC8::9>BCB6@8 @869:AL8C;808>:B8:F )áÜtáuiÉ~Ç}sÇÉj)) ) $J=M>=J=FLA9L=K)AFLG)!"ÉÑ:JA?@L),9LMJ9D))ADD=J)!=DDKjp) 'AAHB>GL )ÄÄtÅuiÄáÉsÅ}Çj)) ) $J=MGJ)"AK;J=L=)1L9?=K)G>)&ME9F),9LMJ9D))ADD=J) !=DD)"A>>=J=FLA9LAGF)AF)4ANGjp) 1=:8(CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )Ä}ÅtÇui~}ÅÅsÇÅj)) ) $J=M)&ME9F),9LMJ9D))ADD=J)!=DD)"AN=JKALQjp) 'AAHB>GL ) ÇÖtÉuiÜÄ}sÅÅj)) ) $Mg)(MFDA9F?)=L)9Dj)Ä}}Öj)o'F;J=9K=<)0=?MD9LGJQ)2)!=DDK)!GJJ=D9L=)OAL@)!"Ü)2k!=DD) 'EH9AJE=FL)9F<).GGJ)1MJNAN9D)AF)&=H9LG;=DDMD9J)!9J;AFGE9).9LA=FLKjp) %6FGEC:BG:EC@C>=;L)G>)w{)2) !=DDK)9?9AFKL)+9FLD=)!=DD)*QEH@GE9jp) @CC9 )~~Ät~~uj)) ) %9J;Z9k+M^GRg)0A;9J),))!=DDK)AF) .9LA=FLK)OAL@)!@JGFA;)*QEH@G;QLA;)*=MC=EA9jp) 'AAHBCG=:E6DL )~}tÑuiÇá~sááj)) ) %9KK=Jg)1L=H@9Fg)19F)L@=),)%Ä") 0=;=HLGJjp) ,6GHE: )ÇÅÑtÖ}ÉÇui~~ÜÑsá}j) ) %9RALg)0GA)=L)9Dj)Ä}}Ñj)o*=L@9D)'F>DM=FR9)'F>=;LAGF)AF)L@=)F:K=F;=)G>)L@=),9LMJ9D))ADD=J) !=DD)0=;=HLGJ)%=F=),;J~jp) ,6GHE:8'AAHBC@CADD9Fg)1MK9Fg)#EADQ)*j)&Gg)+9JAF9)!=DD9g)59QF=)+j)7GCGQ9E9g)9F<)+9J;G)!GDGFF9j) Ä}}Äj)o,)%Ä")0=;JMALK)2OG)"AKLAF;L)F<9HL=JK)LG)2JA??=J),))!=DD)F;LAN9LAGF)9F<) !GKLAEMD9LAGFjp) ,6GHE:8'AAHBC@C

! }{Ä! %JA::=Fg)(G@F)%jg),9L@9F)$GOD=Jg)9F<)$J9F;C)+GJK;@@9MK=Jj)Ä}~Éj)o+=;@9FAKEK)G>) F;LAGF)G>)*=F9DAB>86@8 -B8C@C)+9LJAP) +=L9DDGHJGL=AF9K=Kjp) *:H?:A>6 )Ä~tÄuiÅÉÑsÉáj)) ) %M=JJ9g),9A;A=FL)+A;=)FJ=)"=>=;LAN=)AF)2MEGJ)1MJN=ADD9F;=) AF)+G<=DK)G>)1HGFL9F=GMK)+9DA?F9F;Qjp) 'AAHB>GL )ÄÜtÇuiÉÖ~sÜ}j)) ) %MADD=J=Qg)!9EADD=g),A;@GD9K)"j)&MFLAF?LGFg)9F<)+9JC)(j)1EQL@j)Ä}~Ñj)o29J?=LAF?) ,9LMJ9D))ADD=J)!=DDK)AF)!9F;=J)'EEMFGL@=J9HQjp) ,6GHE:8'AAHBC@C)!"áÇr,)%Ä!ã),))!=DDK)AF)0=KHGFK=)LG) &ME9F)!QLGE=?9DGNAJMKk'F>=;L=<)$A:JG:D9KLKjp) @CC9 )~}ÖtáuiÅÑÄÇsÅ~j)) ) %METg)+_FA;9)=L)9Dj)Ä}}Çj)o'EHJAFL)G>)&ME9F)!QLGE=?9DGNAJMK)'F>=;LAGF)GF)L@=),)) !=DD)0=;=HLGJ)0=H=JLGAJ=jp) @CC9 )~}Çt~ÄuiÅÑÑÇsÖ~j) ) &9F9@9Fg)"GM?D9K)9F<)0G:=JL)Fj)5=AF:=J?j)Ä}~~j)o&9DDE9JCK)G>)!9F;=Ji)2@=)F=PL) %=F=J9LAGFjp) !:@@ )~ÇÇtÉuiÑÇÑsÖÇj)) ) &9QF=Kg) 9JLGF)$j)=L)9Dj)Ä}~Äj)o'EEMF=k!GJJ=D9L=K)FF9DQKAK)G>)9F)&'4k~)49;;AF=)#>>A;9;Q) 2JA9Djp) ,:J8#B<@6B98(CHEB6@8C;8+:9>8>B: )ÅÑÑt~Çui~ÄÖÉsÜÑj)) ) &=g)7MC=)9F<)8@A?9F?)2A9Fj)Ä}~Öj)o,))!=DD)#)+GMK=)*QEH@GA<)!=DDK)9?9AFKL)1QF?=F=A;)9F<)FDDG?=F=A;)2MEGJKj)''j) !@9J9;L=JAR9LAGF)G>)#>>=;LGJ)!=DDKjp) 'BG:EB6G>CB6@8(CHEB6@8C;8!6B8:E )~ÑtÄuiÄÅ}sÅáj)) ) &=JF9F<=Rk'DAR9DALMJJAg)$j)(jg),AK@AL@9)0=<A;A=FL)+GMK=) *QEH@GE9)+G<=Djp) !@>B>86@8!6B8:E8/:F:6E8= )~~t~ÑuiÉáÜÇsáÄj)) ) &=KKD=AFg)"9NA<)%j)2j)=L)9Dj)Ä}~~j)o"A>>=J=FLA9D)0=IMAJ=E=FLK)>GJ)!"ÇÉ)AF),)k!=DD) $MF;LAGF)0=N=9D)"AKLAF;L)0GD=K)>GJ)1QCk$9EADQ))AF9K=K)"A>>=J=FLA9D)0=IMAJ=E=FLK) >GJ)!"ÇÉ)AF),)k!=DD)$MF;LAGF)0=N=9D)"AKLAF;L)0GD=K)>GJ)1QCk$9EADQ))AF9K=Kjp) &:A6GC@C)1A?F9DAF?).9L@O9QK)AF)!"ÇÉk"=>A;A=FL),))!=DDK) *=9>=J=FLA9DDQ)0=?MD9L=<)!QLGLGPA;ALQ)9F<)!QLGCAF=).JGBCB6@8@869:AL8C;808>:B8:F8C;8G=:82B>G:980G6G:F8C;8@A:E>86 )

! }{Å! ~}Åt~ÜuiÖ}~Äs~Öj)) ) &a?DMF),9LMJ9D))ADD=J)!=DD) #:JF8'AAHBC@C)2)!=DDKjp) (CHEB6@8C;8!@>B>86@8'BI:FG><6G>CB )~ÄÑtÇui~ÅÉÅsÑÑj) ) &KMg)&KA9F?)2AF?)=L)9Dj)Ä}~Ñj)o,))!=DDK)!GFN=J?=)*QLA;)%J9FMD=K)LG).JGEGL=) !QLGLGPA;ALQ)9F<).J=N=FL) QKL9F<=J))ADDAF?jp) (CHEB6@8C;8!:@@8 >C@C),9LMJ9D))ADD=J) !=DDK)'K)"=L=JEAF=<):Q)'FL=J9;LAGFK)9EGF?) AEg),GP9)9F<)+;Dk~jp) ,6GHE:8 'AAHBC@C:EF8>B8'AAHBC@C)L@=),)%Ä")'EEMFGJ=;=HLGJ)AF) 'EEMF=)!=DD)F;LAN9LAGF)9F<),9LMJ9D))ADDAF?jp) 'AAHB>GL )~Öt~ui~ásÄáj) ) (G@9FKKGFg)1G>A9)=L)9Dj)Ä}}Éj)o,9LMJ9D))ADD=J)!=DD)#A:BG6@8+:9>8>B: )Ä}~tÖui~~ÇÉsÉÉj)) ) (M=DC=g))=JKLAF)=L)9Dj)Ä}~}j)o!"ÑÄ*)#PHJ=KKAGF)'<=FLA>A=K)9)3FAIM=)1M:K=L)G>) .GDQ>MF;LAGF9D)!"ÉÑ=;LAGFjp) 08>:B8: )Å}ÉtÉÑÜÉuiÜÖÄsÖÇj)) ) )A=KKDAF?g)0j)=L)9Dj)~áÖÖj)o#NA<=F;=)>GJ)9)1AEAD9J)GJ)!GEEGF)+=;@9FAKE)>GJ),9LMJ9D) )ADD=J)!=DD)F;LANALQ)9F<)0=KAKL9F;=)LG)&=EGHGA=LA;)%J9>LKjp) #HECD:6B8(CHEB6@8C;8 'AAHBC@CA;ALQ)>GJ)+GMK=)+GDGF=Q)*=MC=EA9)!=DDKj) !@9J9;L=JAKLA;K)G>)L@=))ADD=J)!=DDjp) #HECD:6B8(CHEB6@8C;8'AAHBC@CGJ),9LMJ9D)!QLGLGPA;ALQ)FJ=)0=IMAJ=<)LG)-N=J;GE=) 'F@A:ALAGF):Q);k!:D)3:AIMALAF)*A?9K=jp) 'AAHB>GL )ÅÄtÄui~ÖÉsÜÑj)) ) )AEg)1MF?BAF)=L)9Dj)Ä}}Éj)o*A;=FKAF?)G>),9LMJ9D))ADD=J)!=DDK):Q)&GKL)+9BGJ) &AKLG;GEH9LA:ADALQ)!GEHD=P)!D9KK)')+GD=;MD=Kjp) ,6GHE: )ÇÅÑtÖ}É~uiÖ}ás~Åj)) ) )GRA=J9)

! }{Ç! L@=).JGL=AF)2QJGKAF=).@GKH@9L9K=)!"ÇÉjp) (CHEB6@8C;8'AAHBC@CACE:N8+9P O8 [_]ZY )~ÉÜtÖuiÅ~Å}sÅáj)) ) )JRQOAFKC9g)#O=DAF9)=L)9Dj)Ä}~Ñj)o!"ÇÉ)'KG>GJE).JG>AD=)'<=FLA>A=K),9LMJ9D))ADD=J)t,)u) 1M:K=LK)OAL@)"A>>=J=FLA9D)F;LANALQjp) .*C08-,# )~~tÇui~s~Üj) ) *9?JM=g))9L@JQFg)FD=P)!9JAK=Qg)"9NA<)(j)+GJ?9Fg)09B=K@)!@GHJ9g)9F<)"9FA=D)+j)"9NAKj) Ä}~Éj)o*=F9DA)$;)0'''F)GF)&ME9F),))!=DDKi) '<=FLA>A;9LAGF)G>)L@=)!D=9N9?=)1AL=)9F<)0=D9LAGFK@AH)OAL@)F;LAN9LAGFjp) 1=:8(CHEB6@8 C;8'AAHBC@C)&ME9F),9LMJ9D))ADD=J)!=DDK)"=>AF=<):Q)#PHJ=KKAGF)G>)L@=)*=MkÖ) t&,)k~u)9F<)*=Mk~~)t,)k~Éu)FFLA?=FKjp) (CHEB6@8C;8'AAHBC@CACE:N8+9P O8 [_]ZY )~Å~tÇui~ÖÜásáÑj)) ) *9FA=Jg)*j)*jg)0j)2=KLAg)(j) AF)*=Mk~á)t!"ÉÑu) *=MCG;QL=)"A>>=J=FLA9LAGF)FFLA?=F)9F<),=MJ9D)!=DD)F<@=KAGF)+GD=;MD=jp) 1=:8 (CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )~ÑátÑuiÄÄÅÅsÅÜj) ) *9FA=Jg)*=OAK)*j)Ä}}Åj)o,9LMJ9D))ADD=J)!=DD)0=;=HLGJ)1A?F9DAF?jp) !HEE:BG8-D>B>CB8>B8 'AAHBC@C)'K)'FNGDN=<)AF) F;LAN9LAF?),))!=DDKjp) ,6GHE: )Åá~tÑÑÑÜuiÖ}ÅsÖj)) ) *9FA=Jg)*=OAK)*j)9F<)(GK=H@)!j)1MFj)Ä}}áj)o"G)L@=)2=JEK)'FF9L=)9F<)F<9HLAN=) 'EEMFALQ)!J=9L=)!GF;=HLM9D) 9JJA=JKlp) ,6GHE:8/:I>:JF8'AAHBC@C>J=Q)5j)=L)9Dj)Ä}~Çj)o.J=9;LAN9LAGF)OAL@)'*k~Äg)'*k~Ég)9F<)'*k~Ü)'F>AFALQ)'DkÄ)0=;=HLGJ)GF)&ME9F)!QLGCAF=k'FC@CCB ) Ä}tÇuiÇÑÅsÖÅj)) ) *BMF??J=Fg)&j)%j)9F<))j))9JJ=j)~áá}j)o'F)1=9J;@)G>)L@=)+AKKAF?)K=D>i+&!)+GD=;MD=K)9F<) ,))!=DD)0=;G?FALAGFjp) 'AAHBC@P1C96L )~~tÖuiÄÅÖsÇÇj) ) *BMF??J=Fg)&9FK)%MKL9>)9F<))9JD)(G@9F)+9DE:=J?j)Ä}}Öj)o.JGKH=;LK)>GJ)L@=)3K=)G>),)) !=DDK)AF)'EEMFGL@=J9HQ)G>)&ME9F)!9F;=Jjp) ,6GHE:8/:I>:JF8'AAHBC@C

! }{É! ) *GH=Rk0GQM=D9g),MJA9)=L)9Dj)Ä}~Çj)o#PLJ9;=DDMD9Jk1A?F9Dk0=?MD9L=<))AF9K=)É)+GCB6@8(CHEB6@8C;8 >C8=:A>FGEL86B98!:@@8 >C@C)9)3FAIM=)!"ÉÖã,)%Ä!@A),9LMJ9D))ADD=J)!=DD) 1M:K=L)=;LAGFjp) .EC8::9>BCB6@8@869:AL8C;808>:B8:F )~}ÜtÅÑui~ÇÖÄÉsÅÄj)) ) *M=LC=k#N=JKDG@g)+=JDAFg)+GFA;9))ADDA?g)9F<)!@A9J9)0GE9?F9FAj)Ä}~Åj)o1A?F9LMJ=K)G>) &ME9F),))!=DD)"=N=DGHE=FL)9F<)2=JEAF9D)"A>>=J=FLA9LAGFjp) $ECBG>:EF8>B8 'AAHBC@C)L@=) "=N=DGHE=FL9D)1QF9HK=jp) ,6GHE:8!CAAHB>86G>CBF )Öi~s~Åj) ) +9;=g)#EADQ)+j)9F<)(GJ<9F)1j)-J9F?=j)Ä}~Ñj)o%=F=LA;)!9MK=K)G>)&ME9F),))!=DD) "=>A;A=F;Q)9F<)2@=AJ)#>>=;L)GF),))!=DD)1M:K=LKjp) $ECBG>:EF8>B8'AAHBC@C) !"ÇÉ)!9MK=K)*QEH@GHJGDA>=J9LAGF)9F<)FMLGAEEMFALQjp) !:@@ )~}ÅtÖui~}ÉásÖ}j)) ) +9JLAFg)+9MJ==F).j)=L)9Dj)Ä}}Äj)o#HAKL9LA;)'FL=J9;LAGF):=LO==F))'0Å"1~)9F<)&*Fk ) "=D9QK)L@=).JG?J=KKAGF)LG)F'"1jp) ,6GHE:8%:B:G>8F )Å~tÇuiÇÄásÅÇj)) ) +9JLAF=Lg)*M) ,))!=DD)+9LMJ9LAGFjp) !:@@8/:DCEGF )~~t~uiÜÉsáÖj) ) +9JLZF=Rk*GKL9Gg)*MAKg)FD:=JLG)FF=Dg)9F<)(MDATF).9JB>86@8!6B8:E8/:F:6E8= )Ä~tÄÄuiÉ}ÇÖsÉÑj) ) +9NADAGg)"GE=FA;G)=L)9Dj)Ä}}Éj)o!@9J9;L=JAR9LAGF)G>)!"ÉÑkr!"~Ñã),9LMJ9D))ADD=J)t,)u) !=DDKi)F)&A?@DQ)"QK>MF;LAGF9D),))1M:K=L)#PH9F<=<)AF)&'4k'F>=;L=<)4AJ=EA;) 'FBCB6@8@869:AL8C;808>:B8:F8C;8G=:82B>G:980G6G:F8 C;8@A:E>86 )~}ÄtÜuiÄÜÜÑsá~j)) ) +=@L9g)0G@L=K@)1jg)#DAR9:=L@)(j)1@H9DDg)9F<))9L9QGMF)0=RN9FAj)Ä}~Ñj)o!GJ<) DGG<)9K)9) 1GMJ;=)G>),9LMJ9D))ADD=J)!=DDKjp) $ECBG>:EF8>B8+:9>8>B: )Ät(9FM9JQui~s~}j)) ) +=DK=Fg)(9FAF=)#jg)%=JLB9F)*M?L@9JLg)FJB9F)!j)*9FC=KL=Jg)9F<)+9J;G)5j)1;@AD@9Ej)Ä}~Ñj) o&ME9F)!AJ;MD9LAF?)9F<)2AKKM=k0=KA<=FL)!"ÉÑt:JA?@Lu),9LMJ9D))ADD=J)!=DD) .GHMD9LAGFKjp) $ECBG>:EF8>B8'AAHBC@CGJ)#>>=;LGJ)9F<)0=?MD9LGJQ)!"Çã)2)!=DD) 1M:K=LKjp) 1=:8(CHEB6@8C;8'AAHBC@C

! }{Ñ! ) +ACMD9Cg)(G9FF9g)$=J=;LAGF)9F<)2@=J9HQjp) @>9F )Å~t~ÖuiÄÅ~ÖsÅ}j) ) +ADD=Jg)(j)1jg))j)Fj)FDD=Qg)9F<).j)+;%D9N=j)~ááÇj)o"A>>=J=FLA9LAGF)G>),9LMJ9D))ADD=J)t,)u) !=DDK)>JGE)&ME9F).JAEALAN=)+9JJGO).JG?=FALGJK)AF)9)1LJGE9k 9K=<)*GF?k2=JE) !MDLMJ=)1QKL=Ei)'<=FLA>A;9LAGF)G>)9)!"ÅÇãÖã),)).JG?=FALGJjp) @CC9 )ÜÅtáuiÄÉáÇs ÄÑ}~j)) ) +AF=Jg)!9L@JAF=)Fjg)2MK9J))j)%AJAg)!D9AJ=)#j)+=Q=Jg)+9JC)1@9:KGNA;@g)9F<)19F<==H))j) 2JAH9L@Qj)Ä}~Éj)oF;IMAKALAGF)G>)F;LAN9LAGF)0=;=HLGJ)*A?9F<):Q)2JG?G;QLGKAK) 0=F<=JK),))!=DDK)&QHGJ=KHGFKAN=jp) 1=:8(CHEB6@8C;8'AAHBC@C:JF8!6B8:E )~Ñt~uiÖs~áj)) ) +J_R=Cg)#jg).j)FF<=JKGFg)9F<)+j)Fj)!9DA?AMJAj)~ááÑj)o0GD=)G>)'FL=JD=MCAFk~É)AF)L@=) "=N=DGHE=FL)G>)&ME9F)!"ÉÑã),9LMJ9D))ADD=J)!=DDK)>JGE)!"ÅÇã)&=E9LGHGA=LA;) .JG?=FALGJ)!=DDKjp) @CC9 )ÜÖtÖuiÄÑÅÄsÇ}j)) ) ,9C9EMJ9g))QG@=A)=L)9Dj)Ä}~Åj)o$J9LJA;A<=)G>),9LMJ9D))ADD=J)!=DDK)"J=KK=<)OAL@)2MEGJk "=JAN=<),)%Ä")*A?9FBCB6@8@869:AL8C;808>:B8:F ) ~~}tÄÅuiáÇÄ~sÄÑj)) ) ,GDLAF?g)FFF=)=L)9Dj)Ä}~}j)o+&!)!D9KK)')!@9AFk0=D9L=<).JGL=AF)F)1@=<=;LAGF)'K)FKKG;A9L=<)OAL@).JG>GMF<),))!=DD)"QK>MF;LAGFjp) 3>EC@CA;A=F;Qjp) (CHEB6@8C;8@@@:EB>86@8 'AAHBC@C)0=?MD9LGJQ)2)!=DDK)AF) .=JAH@=J9D) DGG<)G>).9LA=FLK)OAL@)&=H9LG;=DDMD9J)!9J;AFGE9jp) !6B8:E8/:F:6E8= ) ÑÉtÑuiÄÇÉÖsÑÇj)) ) -JJg)+9JC)2j)9F<)*=OAK)*j)*9FA=Jj)Ä}~}j)o,9LMJ9D))ADD=J)!=DD)#=;LAGFjp) ,6GHE:8'AAHBC@C)9) 2QH=)')!QLGCAF=)0=;=HLGJ)+GKL)0=D9L=<)LG)L@=)'*kÄ)0=;=HLGJ) =L9)!@9AFjp) .EC8::9>BCB6@8@869:AL8C;808>:B8:F )áÖtÄ~ui~~ÇÅásÇÇj)) ) .9J)!QLGLGPA;)!=DD)%J9FMD=)#PG;QLGKAK).9L@O9Qi)

! }{Ö! %J9FRQE=K)&9N=)#NGDN=<)LG)'FD9EE9LAGFjp) +>8EC7:F86B98 'B;:8G>CB )~~tÇuiÇÉÄsÉáj)) ) .9J@9Eg).=L=Jj)Ä}}Éj)o+&!)!D9KK)')+GD=;MD=K)9F<))'01)AF)&ME9F)&AKLGJQg)&=9DL@)9F<) 1MJNAN9Djp) ,6GHE:8/:I>:JF8'AAHBC@CGJ)9F)-D<)F;IM9AFL9F;=jp) ,6GHE:8'AAHBC@C=J=FLA9D).JGDA>=J9LAGF)G>),9LMJ9D))ADD=J)!=DDK)9EGF?) .=JAH@=J9D) DGG<)+GFGFM;D=9J)!=DDK)!G;MDLMJ=<)OAL@) )*QEH@G:D9KLGA<)!=DD) *AF=Kjp) ,6GHE6@8'AAHB>GL86B98!:@@8%ECJG=8/:CB )ÑtÇui~Ö~sÜÜj)) ) .=L=JKGFg)+9JQ)#j)9F<)#JA;)-j)*GF?j)Ä}}Üj)o'F@A:ALGJQ)0=;=HLGJ)1A?F9DAF?)NA9)2QJGKAF=) .@GKH@GJQD9LAGF)G>)L@=)F<9HLGJ)!JCjp) 'AAHB>GL )ÄátÇuiÉÖÜsÜÜj)) ) .@ADDAHKg)(j)&jg) j)2j)%=EDGg)5j)5j)+Q=JKg)Fj)Fj)09QF=Jg)9F<)*j)*j)*9FA=Jj)~áÜÖj)o'F)4ANG) 9F<)AF)4ALJG)F;LAN9LAGF)G>),9LMJ9D))ADD=J)!=DDK)AF)FB>86@8-B8C@C8>6@8(CHEB6@8C;8G=:8@A:E>86B80C8>:GL8C;8!@>B>86@8-B8C@CD9EE9LGJQ)!QLGCAF=K)'*kÄg)'*k~É)9F<)'*kÄ~) +G)+9LMJ=)&ME9F),9LMJ9D))ADD=J)!=DD)0=;=HLGJKjp) @EG=E>G>F8/:F:6E8=8M81=:E6DL )átÑui0~ÄÉj)) ) 0GE==g)0ARO9F)=L)9Dj)Ä}~Åj)o,))!=DD)!"~Ñ)1MJ>9;=)#PHJ=KKAGF)9F<)$MF;LAGF)'K) 0=?MD9L=<):Q)9)"AKAFL=?JAF)9F<)+=L9DDGHJGL=9K=k~Ö)tF"F+~Öujp) @CC9 )~Ä~t~ÜuiÅÉáás ÅÑ}Üj)) ) 0GK=F:=J?g)1j)Fj)=L)9Dj)~áÜÉj)o-:K=JN9LAGFK)GF)L@=)1QKL=EA;)F) FMLGDG?GMK)*QEH@GCAF=kF;LAN9L=<))ADD=J)!=DDK)9F<)0=;GE:AF9FL)'FL=JD=MCAFkÄ)LG) .9LA=FLK)OAL@)+=L9KL9LA;)!9F;=Jjp) 1=:8,:J8#B<@6B98(CHEB6@8C;8+:9>8>B: ) Å~ÅtÄÅui~ÇÜÉsáÄj)) ) 19DA@g)&=DEML)0jg)&9FKk%=GJ?)09EE=FK==g)9F<)FD=P9F<=J)1L=AFD=j)Ä}}Äj)o!MLLAF?)#)+'!F)GF)&ME9F)2MEGJK):Q).JGL=GDQLA;)1@=<ACE:N8+9P O8[_]ZY )~ÑátÜuiÇ}áÜsÇ~}Äj)) ) 1TF;@=Rk+9JLZF=Rg)"A=?G)=L)9Dj)Ä}~Ñj)oF;LAN9L=<)FDDG?=F=A;),))!=DDK).J=>=J=FLA9DDQ))ADD) .GGJ).JG?FGKAK) k!=DD)!@JGFA;)*QEH@G;QLA;)*=MC=EA9)!=DDKjp) $ECBG>:EF8>B8 'AAHBC@C)"JM?)0=KAKL9F;=) AF)&9=E9LGDG?A;9D)!9F;=J)!=DDKjp) -B8C'AAHBC@C

! }|{! 19JN9JA9g)FFMK@JMLAg)"MFA9)(9O<9Lg)(jFD=B9F:EF8>B8'AAHBC@C)+j).=FFAF?=Jj)Ä}}~j)o,=O)'FKA?@LK)AFLG) L@=)2J9FKE=E:J9F=).JGL=AF)2QJGKAF=).@GKH@9L9K=)!"ÇÉjp) 1=:8'BG:EB6G>CB6@8 (CHEB6@8C;8 >C8=:A>FGEL8M8!:@@8 >C@C>g))D9MKg)"9NA<=)$=JJ9JAg)+9J=C)*GKg)1=:9KLA9F)5=KK=D:GJ?g)9F<)+9J;MK) #j).=L=Jj)~ááÜj)oFHGHLGKAK)1A?F9DAF?):Q)"=9L@)0=;=HLGJKjp) #HECD:6B8(CHEB6@8C;8 >C8=:A>FGEL )ÄÉÇtÅuiÇÅásÉáj)) ) 1;GNADD=g)1L=N=F)"j)=L)9Dj)Ä}~Ñj)oF).JG?=FALGJ)!=DD)#PHJ=KKAF?)2J9FK;JAHLAGF)$9;LGJ) 0-0zL)%=F=J9L=K)FDD)&ME9F)'FF9L=)*QEH@GA<)!=DD)1M:K=LKjp) 'AAHB>GL ) ÇÇtÉui~~Ç}sÉ}j)) ) 1;GNADD=g)1L=N=F)"jg)F@9JGF)%j)$J=M)'FF9L=)*QEH@GA<)!=DDKjp) $ECBG>:EF8>B8 'AAHBC@C)2)*QEH@G;QL=Kjp) 'AAHB>GL )ÇtÑuiÉÖÅsÜ~j)) ) 1@A:MQ9g))9RMCG)=L)9Dj)Ä}}Åj)o!"ÄÄÑ)t",F+k~u)'K)'FNGDN=<)AF)*QEH@G;QL=)$MF;LAGFk FKKG;A9L=<)FFLA?=F)~)!GKLAEMD9LGJQ)1A?F9D)>GJ),9AN=)2)!=DD)"A>>=J=FLA9LAGF)9F<) .JGDA>=J9LAGFjp) 1=:8(CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )~áÜt~Äui~ÜÄásÅáj)) ) 1AFA;JGH=g)$J9FC)Fj)=L)9Dj)Ä}}áj)o'FLJ9=HAL@=DA9D)#>>=;LGJ)t!"ÅãurJ=?MD9LGJQ)t$GP.Åãu)2k !=DD)09LAG).J=)&ME9F)!GDGF)!9J;AFGE9jp) %6FGEC:BG:EC@C)9)4AJMKk#F;G<=<)*A?9F<):Q)9),9LMJ9D) )ADD=J)!=DD)F;LAN9LAGF)0=;=HLGJjp) .EC8::9>BCB6@8@869:AL8C;808>:B8:F8 C;8G=:82B>G:980G6G:F8C;8@A:E>86 )áát~ÅuiÜÜÄÑsÅ~j)) ) 1L=::AFKg)!@JAKLGH@=J)!j)=L)9Dj)Ä}}Åj)o49N~)"=H@GKH@GJQD9LAGF):Q)L@=)2QJGKAF=) .@GKH@9L9K=)1&.k~)9K)9)+=;@9FAKE)>GJ)'F@A:ALAGF)G>)!=DDMD9J)!QLGLGPA;ALQjp) +C@:8H@6E86B98!:@@H@6E8 >C@C:B8: )Å~ÖtÉÜÅÑuiÅÖÑsÜ~j)) ) 1MFg)(GK=H@)!jg)(GK@M9),j) =ADC=g)9F<)*=OAK)*j)*9FA=Jj)Ä}}áj)oF<9HLAN=)'EEMF=) $=9LMJ=K)G>),9LMJ9D))ADD=J)!=DDKjp) ,6GHE: )ÇÉÖtÖÄÄáuiÉÉÖsÑ~j))

! }||! ) 2;@ADA9Fg)#DE9)8j)9F<).=L=J)!j)*j) =N=JD=Qj)Ä}}Ñj)oFDL=J=<)!"ÇÉ)#PHJ=KKAGF)9F<) "AK=9K=jp) 1E:B9F8>B8'AAHBC@C=J=FLA9DDQ)FKKG;A9L=K) OAL@)1&'.k~jp) 'BG:EB6G>CB6@8'AAHBC@C>=J=FLA9D)0GD=K)AF)!9F;=J)9F<)4AJ9D)!GFLJGDjp) $ECBG>:EF8 >B8'AAHBC@C)2OG)&ME9F)"GFGJKjp) 'AAHB>GL )ÖtÑuiÖÅás É~j)) ) 49KMg)1MEAL@AJ9)=L)9Dj)Ä}~Éj)oF),GN=D)+=L@G<)LG)#PH9F<)*9J?=),ME:=JK)G>)!"ÉÑ ã) ,9LMJ9D))ADD=J)!=DDK)>JGE)9)+AFML=)$J9;LAGF)G>)1=D=;LAN=DQ)F;;=KK=<)!JQGHJ=K=JN=<) !GJ<) DGG<)>GJ)'EEMFGL@=J9HQ)9>L=J)2J9FKHD9FL9LAGFjp) !LGCG=:E6DL )~Öt~~ui~ÉÜÄsáÅj)) ) 4=ADD=LL=g)FF86@8/:I>:JF )Ä~Çt~uiÄÄsÅÇj)) ) 4ADD9D:9g)+9JLAF)=L)9Dj)Ä}~Åj)o$JGE)2MEGJ)!=DD)+=L9:GDAKE)LG)2MEGJ)'EEMF=)#K;9H=jp) 'BG:EB6G>CB6@8(CHEB6@8C;8 >C8=:A>FGEL86B98!:@@8 >C@C9;=)+GD=;MD=)1H=;A>A;9DDQ) #PHJ=KK=<):Q)F;LAN9L=<),9LMJ9D))ADD=J)!=DDKg)'K)'FNGDN=<)AF),GFs+9BGJ) &AKLG;GEH9LA:ADALQ)!GEHD=PsJ=KLJA;L=<)2MEGJ)!=DD)*QKAKjp) 1=:8(CHEB6@8C;8 #KD:E>A:BG6@8+:9>8>B: )~ÜÖt~ÄuiÄ}ÑÉsÖÄj)) ) 4GKCG:GAFACg)'DA9g)(9E=K)!j)5@AKKLG;Cg)9F<)(GK=H@)Fj)2J9H9FAj)Ä}~Éj)o.=J>GJAF)9F<) %J9FRQE=Ki)$MF;LAGFg)"QK>MF;LAGF)9F<)&ME9F).9L@GDG?Qjp) ,6GHE:8/:I>:JF8 'AAHBC@CB>CB8 >B8'AAHBC@CC@C)L@=)F;LAN9LAGF)0=;=HLGJ)Ä Ç)t!"ÄÇÇujp) (CHEB6@8C;8 'AAHBC@CACE:N8+9P O8[_]ZY )~ÑÉtÖuiÅÉÇÉsÇÜj)) ) 59LRDg)!9JKL=F)9F<)#JA;)-j)*GF?j)Ä}~}j)o1A?F9D)2J9FK),9LMJ9D))ADD=J)!=DDKjp) !HEE:BG8.ECGC8C@F8>B8'AAHBC@C

! }|}! á}ui~s~Öj) ) 5Mg)(j)=L)9Dj)~áááj)oFF)F;LAN9LAF?)'EEMFGJ=;=HLGJ)!GEHD=P)$GJE=<):Q),)%Ä")9F<) "F.~}jp) 08>:B8:8X,:J85CE?N8,P5PY )ÄÜÉtÉÇÄÜuiÖÅ}sÅÄj)) ) 5Mg)(=FFA>=J)"j)=L)9Dj)Ä}}Çj)o.J=N9D=FL)#PHJ=KKAGF)G>)L@=)'EEMFGKLAEMD9LGJQ)+&!) !D9KK)')!@9AFk0=D9L=<)+GD=;MD=)'K)!GMFL=J9;L=<):Q)1@=<B>86@8'BI:FG><6G>CB )~~ÇtÇuiÉÑ}sÑÜj)) ) 5Mg)*j)=L)9Dj)Ä}}Üj)o*=F9DA)0ALMPAE9:k2J=9L=<)!"Ä}ã) 2MEGJ)!=DDKjp) !@>B>86@8!6B8:E8/:F:6E8= )~Çt~ÇuiÇÑÉ}sÉÖj)) ) 7Mg)(A9F@M9g)F@9JGF)%j)$J=M),9LMJ9D))ADD=J)!=DD)"=N=DGHE=FLjp) 1E:B9F8>B8'AAHBC@C>=J=FLA9D)3K=)G>)20F'*)9F<)$9K)*A?9F<):Q)'EE9LMJ=)9F<)+9LMJ=).JAE9JQ) &ME9F),))!=DDKjp)~ÜÜt~ÄuiÄÅÖÉsÜ}j) ) 8@9F?g)&M9)=L)9Dj)Ä}~~j)oF;LAN9LAF?)1A?F9DK)"GEAF9L=)'F@A:ALGJQ)1A?F9DK)AF)!"~ÅÖ*r'*k~É) F;LAN9L=<),9LMJ9D))ADD=J)!=DDKjp) (CHEB6@8C;8'AAHBCG=:E6DL8X&6<:EFGCJBN8+9P O8[__^Y ) ÅÇtÄui~ÜÖsáÉj)) ) 8@9F?g)*A9F?)=L)9Dj)Ä}}áj)o1QF=J?AKLA;)FFLALMEGJ)#>>=;LK)G>)*=F9DA86B8(CHEB6@8C;8 &:A6GC@Cjp) 1=:8(CHEB6@8C;8#KD:E>A:BG6@8+:9>8>B: )Ä~Ät~ÄuiÄ~ÑÉsÜÄj)) ) 8A=?D=Jg)19:JAF9)=L)9Dj)Ä}~Öj)o!"ÉÑ)'K)9).9L@G?=F)0=;G?FALAGF)0=;=HLGJ)GF)&ME9F) ,9LMJ9D))ADD=J)!=DDKjp) 08>:BG>;>88/:DCEGF )Öt~uiÑ~ÅÜj)) )

! }|~!