Abscopal Effect in Non-injected Tumors Achieved with Cytokine-Armed Oncolytic Adenovirus

O iginal A icle Abscopal E ec in Non-injec ed Tumo s Achie ed wi h Cy okine-A med Oncoly ic Adeno i us Riikka Ha unen, 1,2 João M. San os, 1,2 Su i So sa, 1,2 Tommi Ran ape o, 3 Da e Lumen, 4 Mikko Siu ala, 1,2 Anu J. Ai aksinen, 4 Vic o Ce e a-Ca ascon, 1,2 Si i Täh inen, 1 Anna Kane a, 1,5 and Akseli Hemminki 1,2,6 1 Cance Gene The apy G oup, Facul y o Medicine, Uni e si y o Helsinki, Helsinki, Finland; 2 TILT Bio he apeu ics L d., Helsinki, Finland; 3 BioMediTech, Uni e si y o Tampe e, Tampe e, Finland; 4 Labo a o y o Radiochemis y, Depa men o Chemis y, Uni e si y o Helsinki, Helsinki, Finland; 5 Depa men o Obs e ics and Gynecology, Helsinki Uni e si y Hospi al, Helsinki, Finland; 6 Comp ehensi e Cance Cen e , Helsinki Uni e si y Hospi al, Helsinki, Finland Cance ea men wi h local adminis a ion o a med oncoly ic i uses could po en ially induce sys emic an i umo e ec s, o he abscopal e ec , as hey sel -ampli y in umo s, induce dange signaling, and p omo e umo -associa ed an igen p e- sen a ion. In his s udy, oncoly ic adeno i us coding o hu- man umo nec osis ac o alpha (TNF-a) and in e leukin-2 (IL-2) Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 (also known as [a.k.a.] TILT-123) p o oked an i umo e ficacy in umo s ha we e injec ed wi h Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 and hose ha we e le non-injec ed in he same animal. Impo an ly, he i us was able o a el o dis an umo s. To dissec he e ec s o oncolysis and cy okines, we s udied eplica ion-incompe en i uses in mice. Sys emic an i umo e ec s we e simila in bo h models, highligh ing he impo - ance o he a ming de ice. The cy okines induced posi i e changes in immune cell infil a es and induced he exp ession o se e al immune- eac ion- ela ed genes in umo s. In addi- ion, Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 was able o inc ease homing o adop i ely ans e ed umo -infil a ing lympho- cy es in o bo h injec ed and non-injec ed umo s, possibly media ed h ough chemokine exp ession. In summa y, local ea men wi h Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 esul ed in sys emic an i umo e ficacy by inducing immune cell infil- a ion and a ficking in o bo h ea ed and un ea ed umo s. Mo eo e , he oncoly ic adeno i us pla o m had supe io sys emic e ec s o e eplica ion-deficien ec o h ough sp eading in o dis an umo s. INTRODUCTION Eigh million cance dea hs occu globally each yea , and almos all o hem esul om me as a ic cance . 1 New he apeu ic app oaches a e hus u gen ly needed. Because he pa ien s in need o no el ea - men s ypically ha e me as a ic disease, sys emic he apeu ic e ficacy is equi ed. A e a cen u y o de eloping immuno he apies, he fi s p oduc s ha e ecen ly en e ed ou ine use. Many o hem, including monoclonal checkpoin blocking an ibodies and ecombinan cy o- kines, a e used sys emically, which can cause se e e ad e se e en s and e en mo ali y by a ec ing no mal issues. 2,3 In con as , he ypical embodimen o oncoly ic immuno he apy is a local injec ion in o umo s. Se e al ypes o oncoly ic i uses a e being in es iga ed in p eclinical and clinical s udies, and one p oduc , alimogene lahe - pa ep ec (also known as [a.k.a.] T-Vec o Imlygic), has al eady been app o ed. 4–6 E en hough T-Vec is no capable o sp eading o dis an umo s, local injec ion causes immunological eac ions in dis an me as ases, a phenomenon known as abscopal e ec . 6,7 The abscopal e ec has been p oposed as po en ially ele an in pa ien s being ea ed wi h sys emic immuno he apy, such as check- poin blocking an ibodies, and local adia ion. 8,9 The biological a io- nale is ha adia ion can cause immunogenic cell dea h, allowing he induc ion o new T cells agains he umo , while concu en check- poin inhibi o s p e en immunosupp ession om occu ing. This app oach is now being es ed in dozens o ials. 10 Oncoly ic i uses, such as T-Vec, a e able o induce an abscopal e ec wi hou he need o adia ion. 6,7 The biological a ionale is he same: oncoly ic eplica- ion can induce immunogenic cell dea h and immunological dange signaling, bo h o which can induce de no o immuni y agains he umo . Wi h ega d o oncoly ic i uses unde de elopmen , bu no ye app o ed (excep oncoly ic adeno i us Onco ine in China), oncoly ic adeno i uses a e well ole a ed in humans and excellen de ices o ansgene deli e y. 5,11,12 Fo example, oxic sys emic deli e y o IL-2, egula ly used in adop i e cell he apy p o ocols, is eplaceable wi h i us- ec o ed IL-2 gene he apy in he con ex o T cell ans e . 13 In addi ion o immune s imula ion by he ansgene, adeno i al oncol- ysis induces immunogenic cell dea h and he elease o dange signals and umo -associa ed an igens, which inc ease umo immunoge- nici y. 14–16 Impo an ly, adeno i ally deli e ed cy okines p o ide enhanced an i umo e ficacy wi h minimal o nonexis en oxici y. 13,17 Recei ed 31 Oc obe 2018; accep ed 31 Oc obe 2018; h ps://doi.o g/10.1016/j.om o.2018.10.005. Co espondence: Akseli Hemminki, Cance Gene The apy G oup, Facul y o Medicine, Uni e si y o Helsinki, Haa maninka u 3, 00290 Helsinki, Finland. E-mail: akseli.hemminki@helsinki.fi Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 ª2018 The Au ho s. 109 This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). To dec ease he oxici y and inc ease he e ficacy o T cell- ela ed im- muno he apies, such as adop i e cell he apy and checkpoin inhibi- o s, we ha e de eloped an oncoly ic adeno i us coding o human umo nec osis ac o alpha (TNF-a) and in e leukin-2 (IL-2) (Ad5/3-E2F-d24-hTNF-a-in e nal ibosome en y si e [IRES]- hIL-2, a.k.a. TILT-123). 17–19 We hypo hesized ha oncoly ic adeno- i us eplica ion accompanied by IL-2 and TNF-ap oduc ion om umo cells induces immunological e ec s ha a e powe ul no only locally bu also sys em-wide. Because we ha e seen Ad5/ 3-E2F-d24-hTNF-a-IRES-hIL-2 inducing posi i e changes locally in he umo -infil a ing immune cell milieu, as well as on a sys emic le el, 17 we s udied whe he a local ea men would be able o gene a e an abscopal e ec on dis an umo s and he mechanisms behind i . RESULTS Cy okine-A med Oncoly ic Adeno i uses Induce Sys emic An i umo Responses The sys emic e ec s o a local ea men wi h oncoly ic Ad5/3-E2F- d24-hTNF-a-IRES-hIL-2 we e s udied in Sy ian hams e s ha a e semi-pe missi e o human adeno i us eplica ion. 20 In addi ion, ce ain human cy okines, including TNF-aand IL-2, a e bioac i e in hams e s. 17,20 Because his i us was de eloped o enable T cell he apies, he expe imen al se ings included a ea men wi h u- mo -infil a ing lymphocy e (TIL) g a . We obse ed umo g ow h educ ion in bo h injec ed and non-injec ed umo s wi hou di e - ences in umo sizes be ween hese umo s (Figu es 1A, 1B, and 1D). The a ming de ices esul ed in a benefi in umo con ol o e he espec i e una med i us. Wi h ega d o injec ed umo s, he bes g oups we e Ad5/3-E2F-d24-hTNF-aand Ad5/3-E2F-d24-hTNF- a-IRES-hIL-2 (p = 0.002 and p = 0.0034 compa ed wi h TILs alone, espec i ely; p = 0.002 and p = 0.01 compa ed wi h mock). Rega ding non-injec ed umo s, all a med i uses had enhanced an i umo e ficacy a he non-injec ed si e, unlike he una med i us, when compa ed wi h mock and TILs alone (TIL e sus TIL + TNF-a: p = 0.001; TIL e sus TIL + IL-2: p = 0.000427; TIL e sus TIL + TNF-a-IL-2: p = 0.00007; mock e sus TIL + TNF-a: p = 0.011; mock e sus TIL + IL-2: p = 0.022; mock e sus TIL + TNF-a-IL-2: p = 0.006). The i uses we e p esen in injec ed umo s a high le els on day 16, 8 days a e he las in a umo al injec ion. The highes alues we e obse ed in he g oup ea ed wi h he una med i us (Figu e 1C). Vi al DNA le els we e low in non-injec ed umo s and no mal issues Figu e 1. T ea men wi h Oncoly ic Vi us Con ols he G ow h o Bo h Injec ed and Non-injec ed Tumo s Hams e s we e ea ed on days 1, 8, 15, 22, and 29 wi h 1 10 8 VPs in a umo ally (i. .) and wi h 5 10 7 TILs on day 1 in ape i oneally (i.p.). The g ow h o injec ed (A) and non-injec ed (B) hams e umo s (n = 5–6) was measu ed e e y 2–3 days un il day 33. Du ing he ollow-up pe iod, wo animals we e sac i iced om he mock g oup (day 24), wo animals om he g oup ecei ing TILs only (day 22), and one animal om g oups ea ed wi h Ad5/3-E2F-d24-hTNF-a(day 29) and Ad5/3-E2F-d24-hIL-2 (day 29). Small amoun s o i al DNA we e de ec able also in non-injec ed umo s on day 16 (C). The e we e no di e ences be ween he injec ed and non-injec ed umo sizes on day 33 (D). The g aphs show mean plus SEM. S a is ical di e ences we e e alua ed wi h mixed model analysis; ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05. Molecula The apy: Oncoly ics 110 Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 (Figu es 1C and S1A). The highes indi idual alues we e de ec ed in spleen, li e , and lung, bu he e we e no di e ences in biodis ibu ion be ween i uses o o gans. A e ea men s, animals we e moni o ed and sac ificed acco ding o animal egula ions ( umo size eaching 20 mm). The g oup ea ed wi h he double-a med i us had he bes su i al (p = 0.03 and p = 0.0159 compa ed wi h mock and TILs alone, espec- i ely), whe eas TILs alone had a minimal e ec on su i al (Figu e S1B). A ming wi h TNF-aand IL-2 Resul s in an Abscopal E ec E en wi hou Oncolysis Nex , we sough o dissec he e ec o he ansgenes om he e ec s o oncolysis. This could be achie ed by using eplica ion-incompe en adeno i uses wi h mu ine cy okines in immunocompe en mice bea ing B16 melanoma umo s exp essing chicken o albumin (OVA). By day eigh , an ea ly ime poin , a di e ence in umo size could be seen in bo h injec ed and non-injec ed umo s (Figu es 2A and 2B). The umo s injec ed wi h IL-2- and TNF-a-a med i uses we e abou 70% smalle han he PBS-injec ed ehicle con ol umo s (p < 0.0001). The con ol i us Ad5-Luc1, which lacks an immuno- logically ac i e ansgene, had a mino ye s a is ically significan e ec (p = 0.0004). Rega ding non-injec ed umo s, he a med i uses we e he only ones able o induce umo g ow h con ol ( ehicle e sus Ad5-CMV- mTNF-a: p = 0.0162; ehicle e sus Ad5-CMV-mIL-2: p = 0.0475; ehicle e sus Ad5-CMV-mTNF-a/mIL-2: p = 0.0212). The bes umo con ol was induced by he combina ion o he cy okines, mean umo size being hal o he size o he ehicle umo s. The addi- ion o OVA-specific OT-I T cells esul ed in simila ou comes bu did no enhance umo g ow h con ol a his ea ly ime poin (Figu es 2C and 2D). The expe imen was epea ed wi h he mos ele an ea men s: OT-I T cells wi h ehicle con ol, con ol i us Ad5-Luc1, and Ad5-CMV- mTNF-a/mIL-2. Tumo g ow h was ollowed o 6 days. The ea - men did no influence he g ow h o injec ed umo s, bu he animals ea ed wi h cy okine-a med i uses clea ly showed a delay in non- injec ed umo g ow h as compa ed wi h ehicle ea men (p = 0.04; Figu es 2E and 2F). O no e, he e ec was no due o i us sp ead because i al genomes a he non-injec ed umo s we e unde- ec able (Figu e 2G). Thus, he i al ansduc ion o dis an umo s seems o equi e oncolysis and subsequen shedding o he i us om umo s in o blood. Day eigh umo s we e collec ed and analyzed o in a umo al im- mune cell popula ions. The pe cen age o na u al kille (NK) cells Figu e 2. Replica ion-Incompe en Vi us Induces G ow h Con ol Also in he Non-injec ed Tumo a an Ea ly Time Poin The sizes o injec ed (A and C) and non-injec ed (B and D) B16-OVA umo s in mice on day 8 (n = 3–8) wi hou T cell ans e (A and B) o wi h OT-I cells (C and D) we e compa ed wi h a e age mock umo size on he same ime poin . Simila esul s we e ob ained o injec ed (E) and non-injec ed (F) umo s when he expe imen was epea ed. Vi al DNA was de ec ed only in he injec ed umo s (G). The ba s show mean plus SEM. S a is ical signi icance was e alua ed wi h 2-way ANOVA: ****p < 0.0001; ***p < 0.001; *p < 0.05. www.molecula he apy.o g Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 111 (legend on nex page) Molecula The apy: Oncoly ics 112 Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 (NK1.1. and CD11b double-posi i e cells in he CD3-nega i e cell popula ion) was ele a ed in bo h injec ed and non-injec ed umo s ollowing ea men wi h cy okine-coding i uses (Figu e 3A). In addi ion, he cy okine combina ion was able o inc ease he le els o CD11c- and CD86-posi i e dend i ic cells in ea ed umo s, and he end was simila in non-injec ed umo s (Figu e 3B). Mo eo e , he e was a posi i e co ela ion be ween he p esence o NK cells and dend i ic cells (Figu es S2A and S2B). An inc ease o F4/80 and CD11b-posi i e mac ophages was obse ed in bo h umo s (Figu e 3C). In e es ingly, he po ion o immunosup- p essi e M2-like mac ophages (di e en ia ed by CD206 exp ession) was dec eased when umo s ecei ed cy okine-coding i uses (Fig- u e 3D). Again, a simila , ye no significan , end was seen in non-injec ed umo s. Ou o o he ma ke s o immunosupp ession, he ea men s did no a ec he exp ession o TGF-bo FoxP3 (Fig- u es S2C and S2D). No majo di e ences we e obse ed be ween he g oups ega ding melanoma-specific (OVA, T p2, and gp100) CD8- posi i e T cells (Figu e 3E). The p esence o any o he immune cell popula ions s udied he e did no co ela e wi h umo olumes (Fig- u es S2E–S2J). A med Adeno i us Induces he T Cell G a T a icking in o Bo h Injec ed and Non-injec ed Tumo s Because he oncoly ic i us was able o a el o he non-injec ed umo s, we wan ed o s udy whe he he i us induces TIL g a a - ficking in o bo h injec ed and non-injec ed umo s. Dis ibu ion and umo accumula ion o he adiolabeled cells ( 111 In-oxine) we e de e mined by single-pho on emission compu ed omog aphy/ compu ed omog aphy (SPECT/CT) imaging a 48, 72, and 96 h a e adminis a ion (Figu es 4A and 4B). A end o inc eased a ficking in o bo h injec ed and non-injec ed umo s was obse ed in animals ea ed wi h Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 compa ed wi h he una med i us and he ehicle con ol (Figu e 4C). In addi ion, he cells seemed o be mo e pe sis en in he umo s when he animals we e ea ed wi h he cy okine-a med i us: he e was no dec ease o e ime in he injec ed o non-injec ed umo s, whe eas in he ehicle and una med i us con ol g oups he signal dec eased o e ime. To es ima e he e ec o he cy okines on T cell g a a ficking, fluo- escen ly labeled OT-I T cells we e adminis e ed o animals ecei ing ehicle, Ad5-Luc1 con ol i us, o Ad5-CMV-mTNF-a/IL-2 in o one o he wo umo s. Fi e days la e , he umo s we e collec ed and he p esence o ans e ed T cells de ec ed wi h flow cy ome y. E en in he absence o oncolysis, ea men o one umo was able o induce OT-I cell a ficking in o bo h injec ed and non-injec ed umo s (Figu e 4D). The ans e ed T cells did no find hei way in o he umo s when he animals we e ea ed wi h ehicle con ol, and only a low signal was de ec ed in non-injec ed umo s in animals ecei ing he con ol i us. The e we e no s a is ically significan di e ences be ween he injec ed and non-injec ed umo s. Oncoly ic Adeno i us Induces he Exp ession o Immunologically Rele an Genes in Injec ed and Non-injec ed Tumo s In o de o unco e he mechanisms o ac ion unde lying he sys- emic e ec s and he a ficking o adop i ely ans e ed cell g a , we analyzed he gene exp ession p ofiles o injec ed and non-in- jec ed umo s in hams e s ecei ing TIL he apy. In he ehicle g oup, PBS injec ion induced immune eac ions by he up egula- ion o he genes ela ed o humo al immune esponses, chemo- kine and cy okine p oduc ion, inflamma o y esponses, and complemen ac i a ion, among o he s (Table S1). In addi ion, he PBS injec ion down egula ed genes ela ed o cy oskele on o ganiza ion and o he filamen - ela ed p ocesses. To allow o he e ec s o he needle punc u e and ehicle injec ion, we analyzed ea men g oup exp ession le els agains co esponding ehicle g oup umo s. O e 2- old up egula ion was obse ed o 445 genes in injec ed u- mo s and o 165 genes in non-injec ed umo s (Figu es S3 and 5). The numbe o down egula ed genes was 45 and 70 in injec ed and non-injec ed umo s, espec i ely (Figu es S3 and 6). TIL ea men wi hou i uses induced he exp ession o chemokines (Ccl4,Ccl7, Ccl11), bu also immune checkpoin molecules Pdl1 (o Cd274)in he PBS-injec ed umo s and Lag3 in he non-injec ed umo s. In addi ion, we saw up egula ion o a a ie y o lymphocy e- ela ed genes (Sash3,Fgl2,Txk,Gzmk,Bs 1,Pik3ap1,Pik3cd,P kcb,Nkg7, Spn), T cell ac i a o s (P kcq,Tagap) and inhibi o s (P p c, P pn22), dend i ic cell ma ke I gax (o Cd11c), genes p omo ing cy okine p oduc ion (Themis2,Fg ,T em1), and Rn 144b, whose p oduc unc ions in majo his ocompa ibili y complex (MHC) class I an igen p ocessing and p esen a ion (Figu es 5 and S5). Vi us Injec ion Induces he Up egula ion o Genes Rela ed o Bo h Inna e and Adap i e Immuni y Ou o immunologically ele an genes, injec ion wi h ei he i us induced he exp ession o mac ophage ma ke Sca a5, chemokine gene Cxcl12, lymphocy e di e en ia ion ma ke N 5e, and wo genes, Re nla and Re nlb, ela ed o T cell and dend i ic cell ec ui men , espec i ely. In addi ion, we saw up egula ion o T cell inhibi o Ido1 in non-injec ed umo s in he g oup ea ed wi h he una med i us and TILs. In addi ion, as seen in mice, i us injec ion was Figu e 3. T ea men wi h A med Vi uses Induces Posi i e Changes in Immune Cells in he Tumo Mic oen i onmen Tumo samples we e collec ed on day 8, and he p esence o na u al kille cells (A), ma u e dend i ic cells (B), mac ophages (C), and M2-like mac ophages (D) in umo s was de ec ed wi h low cy ome y. Melanoma-speci ic T cells we e de ec ed wi h pen ame s and he esul s wi h OVA + , T p2 + , and gp100 + cells pooled in o one g aph (E). S a is ical di e ences we e analyzed agains co esponding mock umo . The ba s show mean plus SEM. Black ba indica es mock con ol, whi e ba s una med i us o OT-I con ols, and g ay ba s ea men g oups wi h a med i uses. Unpai ed es was pe o med o analyze s a is ical di e ences; ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05. www.molecula he apy.o g Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 113 able o up egula e NK cell ac i a ion ma ke s and dend i ic cell di e - en ia ion (Figu e S5). The una med i us induced he up egula ion o he highes numbe o genes. Among he genes up egula ed in he una med g oup, he e we e se e al T cell ma ke s (I m2a,P kcq,Dpp4,Apln ), B cell- ela ed genes (Jchain,Eb 1,Fgd2,Mzb1), chemokines and hei ecep o s (Ccl21,Ack 1,Cc 5,Cmkl 1,Cxcl14), complemen sys em- ela ed genes (C1s,C1 ,C3a 1 C4,C6,C7), and an igen p esen a ion- ela ed genes (B2m,Tap1,Tap2,Tapbp;Figu e S4). In e es ingly, Ad5/3-E2F- d24 also up egula ed genes ha a e equi ed o lymphocy e mig a- ion and in asion, such as Vcam1,Pecam1, and Esam, in addi ion o CD34 and Sell, coding o he binding pa ne o CD34 on T cells, L-selec in. Ano he in e es ing de ail unique o he una med i us was he up egula ion o genes ela ed o i al p ocesses, o example, an i i al Rsad2 (Figu e S5). In addi ion, we saw down egu- la ion o inflamma ion- ela ed genes such as S100a9,S100a8,O m1, L b (o Tn c), and Gc (Figu es 6 and S6). A med Adeno i us Induces Chemokine Exp ession in Bo h Injec ed and Non-injec ed Tumo s Tumo s injec ed wi h a med i us showed unique up egula ion o cy okine genes Cs 3 and Il1b. In addi ion, we saw an up egula- Figu e 4. A med Vi us Induces TIL and TCR G a T a icking o and Pe sis ence in Bo h Tumo s TILs we e labeled wi h adioac i e indium, and he a icking in o (A) injec ed and (B) non-injec ed umo s was ollowed wi h SPECT/CT imaging o e ime. On day 0, he hams e s ecei ed he labeled TILs in ape i- oneally and i uses o PBS con ol in a umo ally in o one o he wo umo s (n = 2/g oup). The di e ence be ween he g oups was mos p ominen 96 h a e he adminis a ion o he cells (C). TCR-modi ied OT-I cells we e labeled wi h luo escen -labeled nanopa icles, and hei p esence in injec ed and non-injec ed umo s was in es iga ed a e 5 days (n = 6) (D). Tumo s injec ed wi h i uses coding o TNF-aand IL-2 had signi ican ly highe numbe s o in used OT-I cells compa ed wi h ehicle- ea ed umo s (K uskal-Wallis es , *p < 0.05). ion o mac ophage- ela ed gene Ma co, and Ack 3 coding o an a ypical chemokine ecep- o . In he same g oup, he non-injec ed u- mo s showed up egula ion o immunoglobulin lambda- and kappa-like genes. When expanding he obse a ions o all significan ly di e en ially exp essed genes p esen bo h in injec ed and non-injec ed umo s uniquely in his g oup, we obse ed up egula ion o Cxcl5 coding o TNF-a-inducible chemokine and Rnase2 ha a ac s dend i ic cells (Table S2). Fu he mo e, we saw up egula ion o he ollowing genes ela ed o immune eac ions: Ie 3 ha has unc ions in TNF-a-s imula ed apop osis and T cell apop osis inhibi ion; L mp, whose p o ein p oduc deli e s pep ides o MHC class I molecules; B cell egula o Rgs13; and Sul 1e1 ha is in ol ed in inflamma o y- esponse egula ion. By con as , im- mune- eac ion- ela ed genes ha we e down egula ed in his g oup in bo h umo s we e Jak3 (media es IL-2R signaling in T cells and NK cells) and h ee genes coding o pa s o MHC class II complex (Table S2). Because he double-a med i us had mo e p ominen an i umo ab- scopal e ec s on he non-injec ed umo s in compa ison o he same i us wi hou a ming de ices, i was o in e es o s udy hose genes ha changed only in non-injec ed umo s, and only when he dou- ble-a med i us was used o injec ion o he o he umo . In heo y, his could allow dissec ion o he abscopal e ec s o oncolysis om he e ec s o he ansgenes. Howe e , only a ew such genes we e iden ified (up egula ion o LOC101839749 and LOC101842437, anno a ed as immunoglobulin lambda-1 ligh chain-like and immu- noglobulin kappa a iable 4-1-like, espec i ely, and down egula ion o Ca3,Mylp ,Ckm, and Hspb1), bu hey a e cu en ly o unknown significance in he con ex o immuno he apy. Fu he cell biology esea ch could e en ually help unde s and he ole o hese ansc ip s in ou findings. Molecula The apy: Oncoly ics 114 Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 (legend on nex page) www.molecula he apy.o g Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 115 DISCUSSION Cy okines, such as TNF-aand IL-2, a e po en induce s o an i umo immuni y. TNF-ahas bo h di ec and indi ec e ec s on cance cells by inducing nec osis and apop osis, bu also inducing immunologic eac ions ia acu e inflamma ion. 21 Recombinan TNF-ais ou inely used in isola ed limb pe usion o , o example, sa coma and mela- noma, bu i is oo oxic o be used sys emically. 22,23 Wi h adeno i al deli e y, high local concen a ion o TNF-aand IL-2 is achie able wi hou significan sys emic exposu e. 17,18 IL-2 unc ions as a T cell p opaga o and ac i a o , and i is used as a ea men o melanoma and enal cell ca cinoma. IL-2 is also included in many adop i e cell he apy ials, especially in TIL he apy and in many solid umo ials wi h chime ic an igen ecep o T cells (CART) o wi h ecep o - modified T cells (TCR). 24–26 Again, ec o ed deli e y can achie e he beneficial e ec s o IL-2 wi hou sys emic oxici y. 13 He e, we in- se ed TNF-aand IL-2 in o an oncoly ic adeno i us and s udied he sys emic an i umo e ec s o Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 wi h adop i e T cell ans e . Oncoly ic adeno i uses whose capsid is a chime a be ween se o ype 5 and se o ype 3 appea use ul wi h ega d o sys emic deli e y. In line wi h he esul s ob ained in his s udy, we ha e p e iously shown in bo h labo a o y animals and humans ha i us adminis e ed in a u- mo ally o in a enously can ansduce dis an me as ases. 11,27 I has e en been possible o g ow ou he ea men i us om non-injec ed b ain me as ases o a cance pa ien . 11 Injec ed and non-injec ed u- mo s we e ansduced o he same deg ee in pa ien s. Thus, he 5/3 chime ic pla o m is appealing o achie ing sys emic e ec s also h ough i al ansduc ion, in addi ion o immunological e ec s. Consis en wi h p e ious esul s, he oncoly ic i us s udied he e could be de ec ed in bo h injec ed and non-injec ed umo s. The i us sp ead was linked o eplica ion, because he eplica ion-incompe en i uses we e no ound in non-injec ed umo s. Oncoly ic i us le els we e highe in injec ed umo s han in non-injec ed umo s, bu he p es- ence in non-injec ed umo s in immune-compe en hos s cons i u es an impo an p oo -o -concep . Mo eo e , he le el o una med i us in injec ed umo s was highe han ha o a med i uses, sugges ing mo e ex ensi e immunogenici y o he a med i uses. Because he i- uses we e ound in non-injec ed umo s, i was no su p ising o find hem also in no mal issues. The selec i i y o he i us, howe e , is no a he le el o en y, bu on he le el o eplica ion. 28 Because he i us does no eplica e in no mal cells, hese cells a e no damaged and no ad e se e en s a e seen. 17 In ac , highe DNA copy numbe can some- imes be seen in no mal human issues han in umo s. 11 In con as wi h oncoly ic adeno i us, he fi s oncoly ic i us app o ed by au ho i ies in he Uni ed S a es and Eu opean Union (EU), T-Vec, has no been de ec ed om non-injec ed umo s in humans o in animals. 29 Ne e heless, T-Vec induces immunological esponses in dis an umo s, e en hough a a lowe le el. 6,7 In he phase 3 OPTIM ial, T-Vec esul ed in a 26% esponse a e in in- jec ed umo s and 15% in non-injec ed isce al me as ases. 6 Because ou app oach induced simila e ficacy in injec ed and in non-injec ed umo s, one can specula e ha o e all clinical benefi s could be mo e p onounced i he i us is capable o ansducing non-injec ed u- mo s. Thus, he e is a majo di e ence be ween oncoly ic he pes ype 1 and oncoly ic 5/3 chime ic adeno i us, and his di e ence is po en ially e y impo an clinically. The o me can p oduce sys- emic e ec s h ough he immune sys em, whe eas he la e can achie e body-wide e ec s hough wo mechanisms: (1) i al dissem- ina ion h ough he ascula u e, ansduc ion o dis an umo s ol- lowed by ansgene exp ession, and oncolysis; and (2) sys emic immunological e ec s. Conside ing he e a e se e al oncoly ic i uses in de elopmen , i is clea ha hey a e no all alike wi h ega d o he mechanisms o ac ion and sys emic e ficacy. As a esul o i us sp ead in o non-injec ed umo s, he e we e no significan di e ences be ween he sizes o injec ed and non-injec ed umo s in his s udy. Mo eo e , cy okine-a med i uses influenced ci cula ing T cell g a seen as inc eased a ficking in o dis an umo s, bo h wi h oncoly ic and eplica ion-incompe en i uses. The expe imen wi h eplica ion-incompe en i uses wi hou T cell he apy highligh ed he impo ance o he ansgenes in inducing he sys emic an i umo e ec s. E en in he absence o oncolysis, he eplica ion-incompe en i uses we e able o inhibi he g ow h o he non-injec ed umo s when a med wi h TNF-aand IL-2. Mo e- o e , wi h oncoly ic i uses, he una med i us was no able o induce as s ong an i umo e ec s as he a med i uses in he non-injec ed umo s despi e highe copy numbe . The esul s a e in line wi h ou p e ious esul s wi h he same cons uc s: TNF-aand IL-2 a e necessa y o enabling cu a i e ea men wi h TIL he apy and inducing immunological memo y agains umo echallenge. 17 He e, he wo di e en animal models, hams e s and mice, bo h sug- ges ha he ansgenes ha e an impo ance in inducing sys emic an i umo e ec s. Due o limi ed a ailabili y o hams e -specific o c oss- eac i e e- agen s, a mouse model p o ided us a means o s udy immune cell compa men s in umo s, ha ing he ocus on inna e immuni y. In e es ingly, ea ing jus one umo induced immune cell infil a ion also in o he non-injec ed umo . Mo eo e , he p esence o immune cells in umo s did no co ela e wi h umo sizes, sugges ing ha he ea men influences umo mic oen i onmen ega dless o he ol- ume. The clea es di e ence was seen wi h NK cells ha a e known o a ack cells wi h low MHC class I exp ession, such as B16-OVA. 30,31 The p esence o any i us induced NK le els in he injec ed umo s, bu an a ming de ice was equi ed o NK cell induc ion in he Figu e 5. Lis o Genes Up egula ed o e 2-Fold Compa ed wi h Co esponding Mock G oup Genes in clus e 1 a e up egula ed only in non-injec ed umo s, genes in clus e 2 a e up egula ed in bo h umo s, genes in clus e 3 a e up egula ed wi h Ad5/3-E2F-d24 injec ion, genes in clus e 4 a e up egula ed wi h ei he i us, genes in clus e 5 a e up egula ed wi h Ad5/3-E2F-d24-hTNF-a-IRES-hIL-2 only, and genes in clus e 6 wi h any injec ion. Black box indica es o e 2- old up egula ion. Molecula The apy: Oncoly ics 116 Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 non-injec ed umo . In e es ingly, also Balasa e al. 32 epo ed he impo ance o he combina ion o IL-2 and TNF-awi h ega d o NK ac i a ion. In addi ion o NK cells, IL-2 s imula ed he p esence o dend i ic cells exp essing he ma u a ion ma ke CD86. Mo eo e , he p esence o ma u e dend i ic cells posi i ely co ela ed wi h highe numbe s o NK cells in he umo . C oss alk be ween NK cells and dend i ic cells has an in e es ing ole in bo h inna e and adap i e immuni y. NK cells p omo e umo an igen p esen a ion by dend i ic cells, bu a he same ime, CD11c-posi i e dend i ic cells a e equi ed o NK cell p iming. 33,34 Mo eo e , NK cells induce dend i ic cell ma u a ion bo h ia cell-cell con ac s and ia sec e ion o TNF-aand in e e on gamma. 35 NK cells addi ionally sec e e chemokines, such as MIP-1a, MIP-1b, and RANTES, which a ac dend i ic cells. 36 In addi ion o inducing dend i ic cell ma u a ion, TNF-ais capable o supp essing he M2 mac ophage pheno ype. 37 Immunosupp essi e M2-like mac ophages a e known o associa e wi h poo su i al in cance pa ien s, and cance cells appea o a emp o d i e mac o- phage di e en ia ion owa d his pheno ype. 38,39 We saw, howe e , a ela i e dec ease in his sub ype o mac ophages when umo s we e injec ed wi h a med i uses e en hough he o e all mac ophage pe cen age was inc eased in bo h umo s. The e ec was mos no ice- able wi h TNF-a-coding i us. A simila end was seen in gene exp ession p ofiles in hams e umo s whe e 12 ou o 19 genes specific o M1 mac ophages acco ding o K a och ill e al. 37 we e up egula ed in umo s ea ed wi h Ad5/3-E2F-d24-hTNF-a-IRES- hIL-2. Likewise, he exp ession o 13 ou o 20 genes ha a e linked o M2 pheno ype was absen in he umo s. When in e p e ing gene exp ession da a, i is impo an o ex ac he e ec s de i ed om he deli e y o he d ug. We ha e p e iously obse ed ha injec ion wi h saline causes immunological eac ions in umo s. 40 This phenomenon was confi med in ou obse a ions wi h he up egula ion o immune- eac ion- ela ed genes caused by injec ion o PBS. To con ol he e ec o PBS injec ion, we compa ed he gene exp ession p ofiles o he i us-injec ed umo s wi h he saline-injec ed umo s and he non-injec ed umo s wi h he non-injec ed umo s in he ehicle g oup. Mos changes in gene exp ession p ofile occu ed in he g oup ea ed wi h he una med i us. Because his i us causes ewe immune esponses owa d umo cells ha allow eplica ion, oncolysis is mo e p omi- nen , and he e o e he majo i y o he exp ession changes seen a e due o his phenomenon. Fo example, he i us up egula ed many o he complemen sys em componen s, sugges ing ha he un- a med i us migh ha e induced mo e an i i al immune esponses han he a med i us. Figu e 6. Lis o Genes Down egula ed o e 2-Fold Compa ed wi h Co esponding Mock G oup Genes in clus e 1 a e down egula ed in TIL g oup, genes in clus e 2 a e down- egula ed wi h a i us injec ion, and genes in clus e 3 a e down egula ed only in non-injec ed umo s. Black box indica es o e 2- old down egula ion. www.molecula he apy.o g Molecula The apy: Oncoly ics Vol. 11 Decembe 2018 117