AUTHOR'SVIEW OncoImmunology1:5,739–740;August2012;G2012LandesBioscience Cytomegalovirus infection in brain tumors A potential new target for therapy? Cecilia Söderberg-Nauclér1,* and John Inge Johnsen2 1DepartmentofMedicine,Solna;CenterforMolecularMedicineandChildhood;Stockholm,Sweden;2ChildhoodCancerResearchUnit;DepartmentofWomen’sandChildren’sHealth; KarolinskaInstitutet;Stockholm,Sweden Keywords: cytomegalovirus, immunotherapy, cancer, COX-2, brain tumor, anti-viral treatment . e t Emerging evidence demonstrate a high prevalence of cytomegalovirus (CMV) proteins and nucleic acids in different u tumors.CMVisconfinedtotumorcellsandnon-cancercellsincloseproximitytotumorsareconsistentlyvirusnegative.CMV b confersbothoncogenicandoncomodulatorymechanisms,andmaythereforeprovideanoveltargetincancertreatment. i r t s i d Emergingevidencedemonstrateafrequent virus truly plays a role in tumorigenesis exhibiting oncogenic and oncomodulara- t presence of human cytomegalovirus and tumor progression. Glioblastoma tory functions may act to aggravate tumor o (CMV)proteinsandnucleicacidsinbrain patients have a very dismal prognosis with growth and disease progression. n tumors in both adults (glioblastoma) a mean survival of 12–14mon. We Earlier studies have demonstrated that o and children (medulloblastoma).1,2 Other recently found that a low grade CMV CMV induces the expression of cyclooxy- D morecommoncancerformssuchasbreast infection in glioblastomas was associated genase-2 (COX-2). COX-2 inhibitors are cancer,colonandprostatecanceraswellas withlongertimetotumorprogressionand efficient anti-CMV drugs, as virus replica- . e salivary gland mucoepidermoid carcino- improved survival.7 These observations tion appear to depend on the synthesis of c masarealsofrequentlyviruspositive,with imply that CMV may be involved in prostaglandin E 2 (PGE ).6 Several cancer 2 n a prevalence approaching 90–100%.3-5 tumorprogressionratherthanrepresenting types including those that are frequently e CMVproteinsareexpressedonlyintumor an epiphenomenon in these tumors. CMV positive often demonstrate high i c cells, while non-tumor cells surrounding However, regardless of its role in the levels of COX-2; in some of these tumors s the tumor are CMV negative. CMV development of a tumor, the presence of highlevelsofCOX-2expressioncorrelates o proteins confer both oncogenic and onco- CMV in tumor cells but not in normal with poor patient outcome.6 COX-2 i B modulatory mechanisms; they control cell cells surrounding the tumor, makes it a inhibitors are under evaluation as addi- cycle progression by interacting with p53, potentialnewandnoveltherapeutictarget. tional treatment options for several cancer s Rbandcyclins,activatesoncogenicsignaling We recently demonstrated that 92% of forms,andarecentstudydemonstratesup e pathways(PI3K/Akt,Erk,WntandNFkB), primary medulloblastoma tumors are pos- to 70% reducedincidence of colon cancer d inhibit cellular differentiation, induce itive for CMV proteins; viral DNA and in individuals receiving long-term aspirin n chromosomal damage, affect epigenetic RNAweredetectedinprimarytumorsand treatment.8 It is possible that CMV a mechanisms, induce DNA damage and inmedulloblastomacelllines.2Theexpres- contributes to induced COX-2 levels in L inhibit DNA repair mechanisms, induce sion of CMV proteins varied over time in certain tumors and that COX-2inhibitors 2 oncogene expression and telomerase medulloblastomacelllines,andwashighly interferewithviraleffectsinCMVpositive 1 activity, induce inflammation and at the induced by xenografting in nude mice. tumors. Interestingly, we found that only 0 same time avoid recognition by the Established human medulloblastoma CMV positive medulloblastoma cells in 2 immune system.4,6 Furthermore, CMV xenograftswerepositiveforCMVimmedi- culture expressed COX-2, and CMV pro- © encoded proteins inhibit apoptosis through ately early (IE) and late proteins; two viral teins and COX-2 were also co-expressed interactions with key proteins in the proteinsthatareexpressedduringdifferent in medulloblastoma tumors in patients.2 extrinsic and intrinsic apoptotic signaling phases of CMV replication, but infectious Nude mice carrying human medulloblas- cascade, and can induce drug resistance to virus were not obtained from primary toma xenografts treated with either chemotherapeuticagents,whichmayimpair tumors or xenografts.2 These observations celecoxib or the anti-CMV drug valganci- theefficiencyofcancertherapy4(Fig.1). imply that CMV behaves differently in clovir demonstrated approximately 40% Although CMV proteins may both tumor cells compared with an acute infec- inhibition of tumor growth in vivo trigger oncogenesis and confer oncomo- tion that often results in virus-induced whereas combined celecoxib and valganci- dulatoryfunctions,itisunderdebateifthe lysisofinfectedcells.Insteadviralproteins clovir treatment resulted in 72% reduced *Correspondenceto:CeciliaSöderberg-Nauclér;Email:[email protected] Submitted:01/17/12;Accepted:01/21/12 http://dx.doi.org/10.4161/onci.19441 www.landesbioscience.com OncoImmunology 739 also attenuated CMV induced patho- genesis.9 These observations imply that interfering with CMV in CMV positive tumors may provide a new therapeutic option for patients carrying such tumors. Further studies need to evaluate which chemotherapy that is most suitable to combine with celecoxib and valganciclovir in patients. Importantly, we propose that thistherapeuticstrategyisfurtherevaluated as an additional treatment option also . for other CMV positive tumors than e medulloblastoma. ut Recently, we have evaluated the b effect of valganciclovir in glioblastoma i r patients(unpublished data). The drug was well t s tolerated in patients receiving combined i chemotherapy and radiotherapy and indi- d cates an unexpectedly high survival in t patients undergoing radical surgery and o receivinglong-termtreatmentwithvalgan- n ciclovirincombinationwithchemotherapy o and radiation. However, the study was D small including only 42patients, and therefore well-powered studies have to . e furtherevaluatetheefficacyofvalganciclo- c vir in glioblastoma patients. Several small n immunotherapy trials are also ongoing to e evaluate whether enhanced CMV specific i c immunityimprovethesurvivalofglioblas- s Figure1.CMVencodesproteinsthathaveimportantfunctionsontumorcellgrowthandthetumor toma patients. The described studies o microenvironment. shouldencourageforfurtherinvestigations i onthistopic;theroleofCMVneedstobe B tumor growth without the use of chemo- of CMV late proteins was reduced by further evaluated in tumor biology and s therapy.2Importantly,nosignificanteffects about 80% in CMV positive xenografts. medical and immunotherapeutic strategies e wereobservedonthegrowthofCMVnega- In a recent study of a salivary gland tumor targetingCMVshouldbefurtherevaluated d tive tumor cells or xenografts treated with model, small molecule inhibitors of the for potentially improved outcome of n ganciclovir/valganciclovir.2 The expression COX/Amphiregulin/EGFR/Erk pathways patients carrying CMV positivetumors. a L References 5. Melnick M, Sedghizadeh PP, Allen CM, 8. Rothwell PM, Fowkes FG, Belch JF, Ogawa H, Jaskoll T. Human cytomegalovirus and muco- Warlow CP, Meade TW. Effect of daily aspirin on 1. Cobbs CS, Harkins L, Samanta M, Gillespie GY, epidermoid carcinoma of salivary glands: Cell- long-term risk of death due to cancer: analysis of 2 Bharara S, King PH, et al. Human cytomegalovirus specific localization of active viral and oncogenic individualpatientdatafromrandomisedtrials.Lancet 1 infectionandexpressioninhumanmalignantglioma. signaling proteins is confirmatory of a causal 2011;377:31-41;PMID:21144578;http://dx.doi.org/ 0 CancerRes2002;62:3347-50;PMID:12067971 relationship. Exp Mol Pathol 2012; 92:118-25; 10.1016/S0140-6736(10)62110-1 2. BaryawnoN,RahbarA,Wolmer-SolbergN,TaherC, PMID:22101257; http://dx.doi.org/10.1016/j.yexmp. 9. Melnick M, Abichaker G, Htet K, Sedghizadeh P, 2 Odeberg J, Darabi A, et al. Detection of human 2011.10.011 Jaskoll T. Small molecule inhibitors of the host cell © cytomegalovirusinmedulloblastomasrevealsapotential 6. Johnsen JI, Baryawno N, Söderberg-Nauclér C. Is COX/AREG/EGFR/ERK pathway attenuate cyto- therapeutic target. J Clin Invest 2011; 121:4043-55; human cytomegalovirus a target in cancer therapy? megalovirus-induced pathogenesis. Exp Mol Pathol PMID:21946257;http://dx.doi.org/10.1172/JCI57147 Oncotarget2011;2:1329-38;PMID:22246171 2011;91:400-10;PMID:21565184;http://dx.doi.org/ 3. HarkinsLE,MatlafLA,SoroceanuL,KlemmK,Britt 7. RahbarA,OrregoA,StragliottoG,PeredoI,Dzabic 10.1016/j.yexmp.2011.04.014 WJ,WangW,etal.Detectionofhumancytomegalo- M,Wolmer-SolbergN,etal.Humancytomegalovirus 10. Effects of Valganciclovir as an add-on therapy in virus in normal and neoplastic breast epithelium. infectionlevelsinglioblastomamultiformeareofhigh patientswithcytomegalovirus-positiveglioblastoma:A Herpesviridae2010;1:8;PMID:21429243;http://dx. prognostic value fortime totumourprogression and randomised,double-blind,proof-of-conceptstudy. doi.org/10.1186/2042-4280-1-8 survival.Herpesviridae2011;3:3;http://dx.doi.org/10. 4. SoroceanuL,CobbsCS.IsHCMVatumorpromoter? 1186/2042-4280-3-3 Virus Res 2011; 157:193-203; PMID:21036194; http://dx.doi.org/10.1016/j.virusres.2010.10.026 740 OncoImmunology Volume1Issue5