interferon α and cpg oligodeoxynucleotides elicit additive immunostimulatory and antitumor effects

Interferon � and CPGoligodeoxynucleotides elicit additiveimmunostimulatory and antitumoreffectsLloyd Brown, MD,a Julie Roda, BS,b Catherine Terrell, BS,c Abhik Ray Chaudhury, MD,d

Tim Crespin, PhD,e William E. Carson, MD,a,b and Gregory B. Lesinski, PhD,b Columbus and Hilliard,Ohio

Background. We hypothesized that interferon � (IFN-�) and unmethylated cytosine-phosphothioate-guanine (CpG)-rich oligodexoynucleotides (CpG ODNs) would elicit potent antitumor activity due tothe ability of this treatment combination to activate complimentary signal transduction intermediates.Methods. Peripheral blood mononuclear cells treated with CpG ODNs, IFN-�, or both agents combinedwere evaluated for cytotoxicity against human melanoma cells, Jak-STAT signal transduction by flowcytometry, and ISG-15 gene expression by real-time polymerase chain reaction. The effects of CpG ODNsand IFN-� were evaluated in murine models of melanoma in wild-type, IFN-�–deficient, and STAT1-deficient mice. Negative controls in all experiments included treatment with control ODN or phosphate-buffered saline.Results. Treatment of peripheral blood mononuclear cells with a combination of CpG ODNs and IFN-� resulted in enhanced cytotoxicity, activation of natural killer cells, IFN-�–induced STAT1phosphorylation, and transcription levels of ISG-15. These immunostimulatory effects of CpG ODNswere associated with increased expression of STAT1 and STAT2 proteins. Administration of CpGODNs plus IFN-� elicited superior antitumor activity in a murine model of B16 melanoma comparedwith either agent alone. The antitumor properties of CpG ODNs were dependent on STAT1-mediatedsignal transduction within the host but independent of endogenously produced IFN-�.Conclusions. CpG ODNs represent potent immune stimulants that elicit antitumor effects throughSTAT1-mediated signal transduction. (Surgery 2006;140:297-306.)

From the Department of Surgery,a Human Cancer Genetics Program, Columbus, and the Departments ofMolecular Virology, Immunology and Medical Genetics,b Psychology,c and Pathology,d Columbus; and Primetrics
Inc.,e Hilliard
Recombinant interferon � (IFN-�) is FDA-ap-proved for the treatment of malignant melanomaand produces durable regressions of disease in a

Presented at the 67th Annual Meeting of the Society of Univer-sity Surgeons, First Annual Academic Surgical Congress, Febru-ary 7-11, 2006, San Diego, California.

Supported by National Institutes of Health grants CA93670,P01-CA95426, P30-CA16058, the Valvano Foundation for Can-cer Research Award, and the Ohio State University Departmentof Surgery Clinical Science Seed Grant.

Reprint requests: Gregory B. Lesinski, PhD, Human CancerGenetics Program, Dept of Molecular Virology, Immunologyand Medical Genetics. 424 Comprehensive Cancer Center, 410W 12th Ave, Columbus, OH 43210. E-mail: [email protected].

0039-6060/$ - see front matter

© 2006 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2006.05.005

subset of patients.1 IFN-� can elicit direct anti-proliferative, proapoptotic, and antiangiogeniceffects on melanoma cells; however, several linesof evidence indicate that the immunomodulatoryeffects of IFN-� are critical for mediating its an-titumor properties. IFN-� elicits potent immuno-stimulatory effects through its ability to activateantigen-presenting cells (APCs), promote Th1cytokine production, and contribute to theclonal expansion and formation of memoryCD8� T cells.2 Importantly, we have demon-strated previously that components of the innateimmune system, particularly natural killer (NK)cells, are critical for the antitumor actions ofIFN-� in murine models of melanoma.3 IFN-� isthe only therapy approved currently for use as anadjuvant after surgical resection of high-risk mel-
anoma lesions4; however, its impact on overall
SURGERY 297

298 Brown et al SurgeryAugust 2006

survival remains controversial.5,6 Furthermore,treatment with IFN-� is associated with frequentand dose-limiting toxicities.7 Consequently, in-vestigation of novel therapeutic agents or treat-ment combinations is needed in the setting ofmalignant melanoma.

Bacterial unmethylated cytosine-phosphothio-ate-guanine (CpG)-rich oligodexoynucleotides (CpGODNs) elicit immunostimulatory effects similar tothat of bacterial DNA.8 CpG ODNs bind to toll-likereceptor 9 (TLR9) and stimulate the innate im-mune system through their ability to induce cyto-kine secretion (ie, interleukin 12, IFN-�, IFN-�),and through their activities on NK cells and theirability to induce maturation of dendritic cells(DCs).9 CpG ODNs elicit antitumor activity medi-ated by both NK cells and tumor-specific T cells inexperimental models.9 In addition, CpG ODNshave shown promise as immune adjuvants to pep-tide-based vaccines in human clinical trials for mel-anoma and hepatitis B.10,11

We hypothesized that IFN-� and CpG wouldelicit superior antitumor activity than either agentalone because of the ability of this treatment com-bination to activate complimentary signal transduc-tion intermediates. In the present report, we havedemonstrated that CpG ODNs and IFN-� exhibitenhanced cytotoxic activity in vitro and antitumoractivity in a murine model of B16 melanoma. Pre-treatment of human peripheral blood mononu-clear cells (PBMCs) in vitro with CpG ODNs wasshown to enhance IFN-�-induced activation of theJak-STAT signal transduction pathway and down-stream gene expression. In addition, treatment ofmice with CpG ODNs alone and in combinationwith IFN-� was associated with increased lympho-cyte infiltration (CD45�, CD3�) into tumor depos-its. Finally, we demonstrated the antitumor effectsof CpG ODNs in vivo were dependent on STAT1-mediated signal transduction within the host butnot endogenous production of IFN-�.

MATERIAL AND METHODSOligodeoxynucleotides and cytokines. The fol-

lowing CpG-containing ODNs were used: CpG2336 (GGGGACGACGTCGTGGGGGGG, activat-ing for NK cells, and plasmacytoid DCs). A non–CpG-containing ODN (2243) was used as a control(GGGGGAGCATGCTGGGGGGG). All ODNs werepurchased from the Coley Pharmaceutical Group(Wellesley, Mass). Recombinant human IFN-�-2b(specific activity of 2 � 108 IU/mg) was purchasedfrom Schering-Plough, Inc (Nutley, NJ). UniversalIFN-� (IFN-A/D) was purchased from R & D Sys-
tems, Inc (Minneapolis, Minn).
Isolation of PBMCs and NK cells. Healthy donorPBMCs were isolated directly from fresh peripheralblood Leukopacks (American Red Cross, Colum-bus, Ohio) by density gradient centrifugation withFicoll Hypaque (Sigma-Aldrich, St Louis, Mo). NKcells (�95% CD56�) were purified by a 30-minincubation with RossetteSep mixture (Stem CellTechnologies, Vancouver, British Columbia, Can-ada) before centrifugation.

Cytotoxicity assay. To evaluate cytotoxic activityof immune cells, we plated PBMCs from healthydonors in triplicate at effector to target ratios of50:1, 25:1, 12.5: 1 in a 96-well V-bottom plate. Cellswere stimulated with IFN-�2b, CpG ODNs, or bothagents combined for 16 h. PBMCs were also stim-ulated with control ODNs or phosphate-bufferedsaline (PBS; negative controls). Chromium-labeled(51Cr) human melanoma cells (A375, obtainedfrom American Type Culture Collection, Manassas,Va; MEL 1259 obtained from Dr S. Ferrone, Ro-swell Park Cancer Institute, Buffalo, NY)12 or theNK cell–sensitive, K562 chronic myelogenous leu-kemia cell line (American Type Culture Collec-tion) were used as targets (in a standard 51Cr-release assay as described.13 Cytotoxicity againsteach target cell line was evaluated with the use ofPBMCs from n � 3 separate normal donors.

Intracellular flow cytometry. Staining for STAT1and phosphorylated STAT1 (P-STAT1) was accom-plished by using intracellular flow cytometry, asdescribed previously, after in vitro stimulation ofPBMCs with CpG ODNs, IFN-�-2b, or both agentscombined.14 Negative controls consisted of cellsstimulated with control ODNs or PBS. Amplifiedfluorescence signals were expressed as specific flu-orescence (Fsp � Ft – Fb), where Ft represents themedian value of total staining, and Fb representsthe median value of background staining (obtainedby staining with the isotype control Ab).14

Real-time PCR. Real-time polymerase chain re-action (PCR) was used to measure transcript levelsof the interferon-stimulated gene ISG-15 after invitro stimulation of PBMCs with CpG ODNs, IFN-�,or both agents combined.14,15 Negative controlsconsisted of cells stimulated with control ODNs orPBS. Total RNA was isolated from the treated PB-MCs, converted to complementary DNA by reversetranscription, and used as a template for real-timePCR assays. Commercially available, predesignedprimer/probe sets were utilized according to themanufacturer’s recommendations (Assays On De-mand; Applied Biosystems, Foster City, Calif). Pre-designed primer/probe sets for human �-actinwere used as an internal control in each reaction
well (Applied Biosystems), and the data were ana-

Surgery Brown et al 299Volume 140, Number 2

lyzed withe use of Sequence Detector software ver-sion 1.6 (Applied Biosystems).

Murine tumor model. Animal studies were con-ducted according to guidelines of the InstitutionalLaboratory Animal Care and Use Committee ofThe Ohio State University. An intraperitoneal (IP)model of malignant melanoma was employed todetermine whether combined treatment with CpGODNs and IFN-� was superior in antitumor activityto either agent alone.3,16,17 Female C57BL/6 mice(n � 6 mice per group; 6 weeks of age; TaconicFarms, Inc, Germantown, NY) were injected IP onday 0 with 106 B16F1 murine melanoma cells. Onday 1, mice were treated IP with PBS (vehicle),IFN-A/D alone (2 � 104 U/day), CpG, or controlODNs alone (100 �g on days 1, 3, 7, 14, 21, 28, 35).Two additional groups of mice were treated witheither IFN-A/D � CpG or IFN-A/D � controlODNs combined according to the same dosingregimens. All mice were monitored for survival.18

Studies were also performed in B6.129S7-Ifngtm1Ts

(IFN-�-/- or GKO mice; Jackson Laboratory, BarHarbor, Me) or STAT1-/- mice19 to investigate theantitumor effects of CpG ODNs as a single agent.For all studies, n � 6 mice were utilized per treat-ment group.

Immunohistochemical analysis of murine tu-mors. Intraperitoneal B16F1 melanoma tumors wereharvested from mice in each group (described abovein Murine tumor model) after 10 days of treatment.Formalin-fixed, paraffin-embedded tumor sectionswere examined after staining with hematoxylin-eosin(H&E). To identify infiltrating lymphocytes, westained tumor sections with Abs directed againstCD45 (general lymphocyte marker; BD Biosciences,San Jose, Calif) and CD3 (T cell marker; Dako,Carpintera, Calif) or appropriate isotype controls. De-tection was achieved with the Vectastain Elite ABCsystem and Novared Chromogen (Vector, Burlin-game, Calif). Slides were counterstained with hema-toxylin, and dehydrated through graded alcohols andxylene. Tissue specimens were processed in a singlebatch to ensure valid comparisons between samples.Analysis was performed in a blinded fashion by anexperienced dermatopathologist. Three or morehigh-power fields per tumor section were analyzedwith the use of a Nikon Eclipse E400 microscope(Tokyo, Japan).

Statistical analysis. A conservative, nonparamet-ric Mann-Whitney test was selected to evaluate sur-vival data in murine studies. The Mann-Whitneytest made planned comparisons between survivaltimes in the treatment groups without distribu-tional assumptions. To compare treatment groups,
we pooled and ranked survival times (the average
rank was assigned in the case of ties). The numberof instances in which a mouse from one treatmentgroup out-survived a mouse from the comparisongroup, and vice versa, was then calculated. Groupswere different if the ranks of survival times wereunlikely to be randomly mixed between them (P �.05). No adjustment for experiment-wise error wasmade because of the very limited number of pre-planned comparisons in each dataset.

RESULTSEnhanced cytotoxicity and activation of NK

Cells in response to IFN-� and CpG ODNs. Wehypothesized that IFN-� and CpG ODNs would acttogether to enhance the cytotoxicity of immuneeffector cells against human melanoma cell lines(1259 MEL, A375). Stimulation of PBMCs fromhealthy donors with media alone or control ODN(1 �mol/L) produced a minimal degree of cyto-toxicity. As expected, a moderate degree of cyto-toxicity against this target was evident in PBMCstreated with either low doses of IFN-� (500 U/mL)or CpG ODNs (1 �mol/L). Costimulation of hu-man PBMCs with IFN-� (500 U/mL) and CpGODNs (1 �mol/L; 16 hours) resulted in enhancedcytotoxicity against the human 1259 MEL cell line,compared with IFN-� alone or CPG ODNS alone(Fig 1, A). A similar pattern of cytotoxicity was alsoobserved after incubation of stimulated PBMCswith the NK cell–sensitive K562 cell line (Fig 1, B)and the A375 human melanoma cell line (Fig 1, C).Because CpG ODNs are known to elicit potentstimulatory activity on NK cells,9 surface levels ofCD69, a cellular activation marker, were measuredin purified human NK cells. CD69 levels were ele-vated after overnight stimulation with IFN-� (102-104 U/mL) plus CpG ODNs (1 �mol/L) comparedwith either agent alone, confirming enhanced ac-tivation of NK cells in response to this treatmentcombination (Fig 2, A and B).

Effects of CpG ODNs and IFN-� on the Jak-STAT signal transduction pathway. The effects ofCpG ODNs and IFN-� on Jak-STAT signal transduc-tion were evaluated in PBMCs. As a marker ofIFN-�–induced signal transduction, flow cytometricanalysis of Tyr701-phosphorylated STAT1 (P-STAT1) was conducted in PBMCs after 1-hour stim-ulation with IFN-� (104 U/mL), CpG ODNs (1�mol/L), or both agents combined. Controls inthis assay included cells stimulated with PBS, con-trol ODNs (1 �mol/L), or control ODNs in com-bination with IFN-�. As expected, PBMCs treatedwith IFN-� displayed a robust activation of STAT1;however, treatment with CpG ODNs alone did not
induce directly the phosphorylation of STAT1 at

sphat


the 1-hour time point (Fig 3, A). Furthermore,P-STAT1 levels were not enhanced after a 1-hourtreatment with CpG ODNs and IFN-� combined,compared with IFN-� alone. In contrast, overnightpretreatment with CpG ODNs (16 hours) sensi-tized PBMCs to the effects of IFN-� and led toenhanced STAT1 phosphorylation, compared withcontrol-treated cells (Fig 3, B). Flow cytometricanalysis of PBMCs after overnight stimulation withCpG ODNs further revealed increased levels oftotal STAT1 and STAT2 proteins, compared withPBMCs treated with PBS or control ODNs (Fig 3, Cand D). These data suggest that CpG ODNs canenhance IFN-� responsiveness through upregula-

Fig 1. CpG ODNs and IFN-� enhance cytotoxicity of PBcytotoxic activity against (A) the 1259 MEL human melaa standard 4-hour 51Cr-release assay after treatment for 1observed against the A375 human melanoma cell line. Efrom quadruplicate wells. Data are representative ofCytosine-phosphothioate-guanine; Ctrl, control; PBS, pho

tion of Jak-STAT signal transduction intermediates.

Enhanced expression of interferon-stimulatedgene-15 in response to IFN-� and CpG ODNs. Todetermine whether CpG ODNs could augmentIFN-�-induced gene expression in human PBMCs,we evaluated cells for the expression of interferon-stimulated gene-15 (ISG-15; an IFN-�–responsivegene) by real-time PCR. Consistent with our signaltransduction data, coadministration of CpG ODNsin combination with IFN-� overnight did not en-hance the transcription of ISG-15, compared withCpG ODNs or IFN-� alone (Fig 4, A). However,pretreatment with CpG ODNs (1 �mol/L, 8 hours)rendered human PBMCs more responsive to IFN-�(104 U/mL, 4 hours) at the level of gene transcrip-

against human melanoma cells. PBMCs were tested forcell line and (B) the NK cell–sensitive K562 cell line ins as indicated. C, A similar effect on cytotoxicity was alsors represent standard deviations of % lysis of target cellsiments from PBMCs of n � 3 normal donors. CpG,e-buffered saline.

MCsnoma2 hourrror baexper

tion (Fig 4).


Treatment with IFN-� and CpG ODNs enhancesthe survival of mice in a model of B16 melanoma.The effects of combined treatment with CpG ODNsand IFN-� were studied in a murine model of B16malignant melanoma (n � 6 mice per group).Administration of IFN-� alone or in combinationwith IFN-� with control ODNs led to an improve-ment in survival (median survival, 28 and 30 days,respectively), compared with mice treated with PBSalone or control ODNs alone (Fig 5; median sur-vival, 16 days and 16 days, respectively; P ’s � .004).Mice receiving CpG ODNs as a single agent also

Fig 2. Enhanced NK-cell activation by CpG ODNs andIFN-�. Expression of CD69 (NK-cell activation marker)was evaluated by flow cytometry after an overnight (16hours) stimulation of enriched human NK cells withCpG ODNs (1 �mol/L) in combination with IFN-� (104

U/mL). PBMCs stimulated with CpG ODNs alone orcontrol ODNs combined with IFN-� served as controls inthis assay. A, Data are presented as a representative his-togram, and fluorescence data are presented as specificfluorescence intensity of CD69-PE staining (Fsp). Appro-priate isotype control Abs were used to set markers (M1)in the histogram. B, Quantitative analysis of CD69 Fsp(y-axis) shown in (A). CpG, Cytosine-phosphothioate-guanine; Ctrl, control; PBS, phosphate-buffered saline.

displayed improvement in survival (median sur-

vival, 24 days), compared with negative controlmice (P � .004); however, the survival of micetreated with CpG ODNs combined with IFN-� (me-dian survival, 38 days) was enhanced when com-pared with either CpG ODNs alone (U � 1, P �.006) or IFN-� ODN alone (U � 5, P � .037).

Treatment with IFN-� and CpG ODNs increasestumor-infiltrating immune cells. Histologic analysiswas conducted on B16 melanoma tumors after 10days of treatment. H&E staining revealed denseimmune cell infiltrates at multiple foci in viabletumor cells obtained from mice treated with CpGODNs alone or CpG ODNs in combination withIFN-� (Fig 6, A). Only small, focal sporadic infil-trates of lymphocytes were observed in tumors frommice receiving IFN-� alone or control ODNs plusIFN-�. In these mice, a majority of infiltrate wasobserved in the peritumoral regions rather than inviable tumor. Tumors derived from PBS or controlODN–treated mice exhibited minimal infiltrationand striking necrosis (Fig 6, A). Although melaninwas visible in most tumor sections (Fig 6, B), infil-trating lymphocytes displayed distinct, membra-nous staining and were confirmed as CD45-positivestaining (Fig 6, B). Further immunohistochemicalanalysis also revealed that a large portion of theinfiltrating immune cells in tumors derived frommice treated with CpG ODNs alone or CpG ODNscombined with IFN-� were CD3-positive, suggestingthat T cells played a role in the effects of CpGODNs (Fig 6, B).

The antitumor effects of CpG ODNs are inde-pendent of endogenous IFN-�. In vivo administra-tion of CpG ODNs induces secretion of multiplecytokines including tumor necrosis factor �(TNF-�) and both type I and II IFNs (IFN-�),among others.20 To determine whether the antitu-mor actions of CpG ODNs as a single agent weredependent on host IFN-� production, we employedGKO mice (n � 6 mice per treatment group).Administration of CpG ODNs prolonged survival inGKO mice, compared with control ODNs (mediansurvival, 23 and 15 days, respectively; U � 4.0, P �.018) or PBS-treated mice, (median survival, 15days; P � .032; Fig 7). These data suggested that theantitumor properties of CpG ODNs were indepen-dent of endogenously produced IFN-�.

The antitumor effects of CpG ODNs are depen-dent on STAT1 signal transduction within the host.We investigated whether the antitumor activity ofCpG ODNs was dependent on STAT1 signal trans-duction in host cells. As expected, CpG ODNs pro-longed survival of wild-type (WT) mice, comparedwith control ODN-treated mice (median survival,
22 and 15 days, respectively; P � .007; n � 6 mice

contro


per group; Fig 8). However, no survival benefit wasobserved in STAT1 KO mice treated with CpGODNs, compared with control ODN–treated mice(median survival, 14 and 14 days, respectively; P �.336; n � 6 mice per group; Fig 8).

DISCUSSIONIn the present report, we demonstrated that CpG

ODNs and IFN-� elicit superior antitumor activity ina murine model of melanoma, compared with either

Fig 3. Effect of CpG ODNs and IFN-� on STAT1 activPBMCs after cotreatment with CpG ODNs (1 �mol/L)stimulated with PBS, control ODNs, CpG ODNs alone, oassay. B, P-STAT1 was evaluated in PBMCs after overnigh�mol/L), or PBS followed by a 15-minute stimulation windividual experiments with PBMCs from n � 3 normal dP-STAT1 staining (y-axis, Fsp). Total levels of intracellulacytometry in PBMCs from normal donors after an overni�mol/L), or PBS (negative control). Data are presentedor STAT2 staining are presented on the y-axes. Approprihistogram. CpG, Cytosine-phosphothioate-guanine; Ctrl,

agent alone. In vitro, this treatment combination en-

hanced the cytotoxicity of PBMCs against humanmelanoma cell lines and led to increased NK cellactivation. Further in vitro studies utilizing PBMCsfrom normal donors demonstrated that pretreatmentwith CpG ODNs led to increased IFN-�–induced sig-nal transduction and gene expression. This effect wasdue, in part, to increased protein expression of Jak-STAT–signaling intermediates (STAT1 and STAT2)after stimulation with CpG ODNs. Further experi-ments revealed that the in vivo antitumor properties

A, Phosphorylated STAT1 (P-STAT1) was evaluated inN-� (104 U/mL) for 1 hour by flow cytometry. PBMCsol ODNs combined with IFN-� served as controls in thisreatment (16 hours) with CpG ODNs, control ODNs (1BS or IFN-� (104 U/mL) by flow cytometry. Data fromare presented as mean-specific fluorescence intensity ofSTAT1 and (D) STAT2 proteins were evaluated by flow

imulation (16 hours) with CpG ODNs, control ODNs (1resentative histograms and fluorescence data for STAT1type control Abs were used to set markers (M1) in eachl; PBS, phosphate-buffered saline.

ation.and IFr contrt pretith Ponorsr (C)ght stas repate iso

of CpG ODNs were dependent on STAT1-mediated


signal transduction within the host but independentof endogenously produced IFN-�. Immunohisto-chemical analyses indicated that the antitumor effectsof CpG ODNs alone and CpG ODNs combined withIFN-� were associated with an enhanced lymphocytic

Fig 4. Effect of CpG ODNs and IFN-� on IFN-� stimu-lated gene expression. Levels of ISG-15 (an IFN-�–re-sponsive gene) were evaluated via real-time PCR inPBMCs from normal donors after (A) an overnight co-treatment (16 hours) or (B) pretreatment (8 hours) withCpG ODNs, control ODNs (1 �mol/L), or PBS followedby 4-hour stimulation with PBS or IFN-� (104 U/mL).Data were expressed as the mean fold increase relative tothe baseline levels (PBS pretreatment; 4-hour PBS stim-ulation), and standard deviations are derived from trip-licate wells. All real-time PCR data were normalized tothe level of �-actin (housekeeping gene) messengerRNA. CpG, Cytosine-phosphothioate-guanine; Ctrl, con-trol; IFN, interferon; PBS, phosphate-buffered saline.

infiltrate. These data represent the first report of CpG

ODNs administered in combination with IFN-� andsuggest a role for this drug combination in the treat-ment of advanced melanoma.

Previous in vivo studies have demonstrated thatCpG ODNs elicit antitumor activity against highlyaggressive murine tumors (AG104A, IE7 fibrosar-coma, B16 melanoma, 3LL lung carcinoma) whenadministered as a single agent.9,18 In these studies,tumor regression was dependent on NK-cell activ-ity, maturation and migration of DCs, and resultingability of DCs to stimulate tumor antigen-specific Tcells.9 Previous reports from our group have dem-onstrated that STAT1-mediated signal transductionin NK cells was critical for the antitumor actions ofIFN-� in a murine model of malignant melanoma.3

Therefore, we hypothesized that administration ofCpG ODNs combined with IFN-� would result inunique immunostimulatory effects on NK andother cell compartments, and lead to enhancedantitumor activity. Data from the present study arein agreement with these earlier reports, but furtherdefine a role for STAT1-mediated signal transduc-tion in mediating the antitumor effect of CpG

Fig 5. Antitumor activity of CpG ODNs and IFN-� invivo. C57BL/6 mice (n � 6 mice per group) were in-jected IP with 106 B16 murine melanoma cells on day 0.Beginning on day 1, mice were treated with PBS, IFN-�,control ODNs, CpG ODNs, or ODNs combined withIFN-� as described in MATERIAL AND METHODS, andmonitored for survival. CpG, Cytosine-phosphothioate-guanine; Ctrl, control; IFN, interferon; PBS, phosphate-buffered saline.

ODNs.


CpG ODNs induce the expression of several im-munostimulatory cytokines and chemokines, mostnotably IFN-� from plasmacytoid DCs (PDCs).21 Infact, previous studies demonstrated that CpGODN–induced IFN-� production involves p38MAPK-dependent STAT1 phosphorylation inPDCs.22 Therefore, it is possible that the antitumoreffects of this treatment combination are a result ofprolonged production of endogenous IFN-� orother type I IFNs. The fact that STAT1 was criticalfor the antitumor activity of both IFN-�3 and CpGODNs (Fig 8) supports this interpretation. Alterna-tively, synergistic or additive immune stimulationcould also result from the release of other cytokines

Fig 6. Immunohistochemical analysis of murine B16 menecrosis in tumors derived from PBS (original magnifioriginal magnification: �100). Lymphocytic infiltrates wderived from mice treated with IFN-� and control ODN�100), rather than in viable tumor tissue. Significant focamice treated with CpG ODNs or CpG ODNs in combinatImmunohistochemical staining confirmed that lymphocODNs or CpG ODNs in combination with IFN-� containdicate positive staining). No evidence of lymphocytic inmice (PBS- or control ODN–treated). Melanin depositsdifferentiated from CD3� and CD45� cells.CpG, Cytooxynucleotides; PBS, phosphate-buffered saline.

such as interleukin 12 or TNF-�.20 The increased

levels of STAT1 and STAT2 proteins in response topretreatment with CpG ODNs further support thenotion that additional cytokines may play a role inthe enhanced activity of this treatment combina-tion by modulating the levels of signal transductionintermediates.

The increased immune infiltrate after CpG treat-ment observed in the present report was consistentwith other studies in murine tumors.9 Because theinfiltrating cells were confirmed to consist of CD3�

lymphocytes, these data are in agreement with find-ings from other groups that demonstrate that CpGcan induce indirectly the activation of tumor-spe-cific T lymphocytes.23 We did not, however, observe

a from treated mice. A, H&E staining revealed striking: �200) and control ODN–treated mice (black arrow;bserved primarily in the peritumoral spaces of tumorsbined with IFN-� (green arrows; original magnification:hocytic infiltrates were observed in tumors derived fromth IFN-� (green arrows; original magnification: �100). B,filtrates in tumors derived from mice treated with CpGcells with a CD45� and CD3� phenotype (black arrowse was observed in tumors derived from negative controlalso observed in tumors (red arrows) however was easilyosphothioate-guanine; IFN, interferon; ODN, oligode-

lanomcationere os coml lympion wiytic ininedfiltratwere

sine-ph

any evidence for direct activity of CpG ODNs on

wild type.


T-cell proliferation (data not shown). Together,these data suggest that the antitumor properties ofCpG may be the result of the immunostimulatoryactions on multiple cellular compartments. A flowchart summarizing the data from the present studyis illustrated in Figure 9.

CpG ODNs represent a group of novel immunestimulants that are being evaluated currently inhuman clinical trials. Published reports to dateusing CpG ODNs as an immune adjuvant haveindicated that these agents are well-tolerated inpatients and effective at enhancing the immuneresponse to tumor and viral peptide vaccines.10,11

Although the present study has evaluated only asingle CpG ODNs sequence, a variety of CpG ODNssequences are currently under investigation in clin-ical and preclinical studies. In fact, recent advanceshave demonstrated that it is possible to optimizethe design of synthetic CpG ODNs to activate spe-cific cell populations (ie, PDCs, B cells).11 Furtherstudies with modified CpG ODN sequences wouldtherefore be of clinical interest for targeted immu-notherapeutic regimens.

In conclusion, the present data suggest that CpGODNs represent potent immune stimulants that

Fig 7. The role of endogenous IFN-� in mediating theantitumor activity of CpG ODNs. B6.129S7-Ifngtm1Ts

(IFN-�-/- or GKO mice; n � 6 mice per group) wereinjected IP with 106 B16 murine melanoma cells on day0. Beginning on day 1, mice were treated with PBS,control ODNs, or CpG ODNs, as described in MATE-RIAL AND METHODS, and monitored for survival. CpG,Cytosine-phosphothioate-guanine; PBS, phosphate-buff-ered saline.

elicit their antitumor effects through STAT1-medi-

Fig 8. The role of STAT1-mediated signal transductionin mediating the antitumor activity of CpG ODNs.STAT1 KO mice were injected IP with 106 B16 murinemelanoma cells on day 0. Beginning on day 1, mice weretreated with PBS, control ODNs, or CpG ODNs, as de-scribed in MATERIAL AND METHODS, and monitoredfor survival. CpG-WT, Cytosine-phosphothioate-guanine–

Fig 9. Flow chart summarizing the proposed mechanismof action of CpG ODNs and IFN-� based on data fromthe present study. CpG, Cytosine-phosphothioate-gua-nine; IFN, interferon; IL, interleukin; ODN, oligode-
oxynucleotides; TNF, tumor necrosis factor.


ated signal transduction. CpG ODNs may be usefulin combination with immune-enhancing cytokines.The ability of CpG ODNs to augment expression ofJak-STAT signal transduction intermediates sug-gests that these agents may be useful for modulat-ing the host response to cytokine immunotherapy.

The authors thank the Ohio State University HistologyCore Facility for assistance in optimizing immunohisto-chemical staining.

REFERENCES1. Belardelli F, Ferrantini M, Proietti E, Kirkwood JM. Inter-

feron-alpha in tumor immunity and immunotherapy. Cyto-kine Growth Factor Rev 2002;13(2):119-34.

2. Kolumam GA, Thomas S, Thompson LJ, Sprent J, Murali-Krishna K. Type I interferons act directly on CD8 T cells toallow clonal expansion and memory formation in responseto viral infection. J Exp Med 2005;202(5):637-50.

3. Lesinski GB, Anghelina M, Zimmerer J, et al. The anti-tumor effects of interferon-alpha are abrogated in a STAT1-deficient mouse. J Clin Invest 2003;112(2):170-80.

4. Agarwala S. Improving survival in patients with high-risk andmetastatic melanoma: immunotherapy leads the way. Am JClin Dermatol 2003;4(5):333-46.

5. Kirkwood JM, Manola J, Ibrahim J, Sondak V, Ernstoff MS,Rao U. A pooled analysis of eastern cooperative oncologygroup and intergroup trials of adjuvant high-dose inter-feron for melanoma. Clin Cancer Res 2004;10(5):1670-7.

6. Eggermont AM, Suciu S, MacKie R, et al. Post-surgery adju-vant therapy with intermediate doses of interferon alfa 2bversus observation in patients with stage IIb/III melanoma(EORTC 18952): randomised controlled trial. Lancet2005;366(9492):1189-96.

7. Kirkwood JM, Bender C, Agarwala S, et al. Mechanisms andmanagement of toxicities associated with high-dose inter-feron alfa-2b therapy. J Clin Oncol 2002;20(17):3703-18.

8. Krieg AM. CpG motifs: the active ingredient in bacterialextracts? Nat Med 2003;9(7):831-5.

9. Kawarada Y, Ganss R, Garbi N, Sacher T, Arnold B, Ham-merling GJ. NK- and CD8(�) T cell-mediated eradication ofestablished tumors by peritumoral injection of CpG-contain-ing oligodeoxynucleotides. J Immunol 2001;167(9):5247-53.

10. Halperin SA, Van Nest G, Smith B, Abtahi S, Whiley H,Eiden JJ. A phase I study of the safety and immunogenicityof recombinant hepatitis B surface antigen co-administeredwith an immunostimulatory phosphorothioate oligonucleo-tide adjuvant. Vaccine 2003;21(19-20):2461-7.

11. Speiser DE, Lienard D, Rufer N, et al. Rapid and strong

human CD8� T cell responses to vaccination with peptide,IFA, and CpG oligodeoxynucleotide 7909. J Clin Invest2005;115(3):739-46.

12. Restifo NP, Marincola FM, Kawakami Y, Taubenberger J,Yannelli JR, Rosenberg SA. Loss of functional beta 2-micro-globulin in metastatic melanomas from five patients receiv-ing immunotherapy. J Natl Cancer Inst 1996;88(2):100-8.

13. Fehniger TA, Carson WE, Mrozek E, Caligiuri MA. Stem cellfactor enhances interleukin-2-mediated expansion of mu-rine natural killer cells in vivo. Blood 1997;90(9):3647-53.

14. Lesinski GB, Kondadasula SV, Crespin T, et al. Multipara-metric flow cytometric analysis of inter-patient variation inSTAT1 phosphorylation following interferon alfa immuno-therapy. J Natl Cancer Inst 2004;96:1331-42.

15. Lesinski GB, Badgwell B, Zimmerer J, et al. IL-12 pretreat-ments enhance IFN-alpha-induced Janus kinase-STAT sig-naling and potentiate the antitumor effects of IFN-alpha ina murine model of malignant melanoma. J Immunol2004;172(12):7368-76.

16. Fleischmann CM, Stanton GJ, Fleischmann WR Jr. En-hanced in vivo sensitivity of in vitro interferon-treated B16melanoma cells to CD8 cells and activated macrophages.J Interferon Cytokine Res 1996;16(10):805-12.

17. Berkelhammer J, Oxenhandler RW, Hook RR Jr., HennessyJM. Development of a new melanoma model in C57BL/6mice. Cancer Res 1982;42(8):3157-63.

18. Ballas ZK, Krieg AM, Warren T, et al. Divergent therapeuticand immunologic effects of oligodeoxynucleotides with dis-tinct CpG motifs. J Immunol 2001;167(9):4878-86.

19. Durbin JE, Hackenmiller R, Simon MC, Levy DE. Targeteddisruption of the mouse Stat1 gene results in compromisedinnate immunity to viral disease. Cell 1996;84(3):443-50.

20. Roda JM, Parihar R, Carson WE 3rd. CpG-containing oli-godeoxynucleotides act through TLR9 to enhance the NKcell cytokine response to antibody-coated tumor cells. J Im-munol 2005;175(3):1619-27.

21. Kadowaki N, Ho S, Antonenko S, et al. Subsets of humandendritic cell precursors express different toll-like receptorsand respond to different microbial antigens. J Exp Med2001;194(6):863-9.

22. Takauji R, Iho S, Takatsuka H, et al. CpG-DNA-inducedIFN-alpha production involves p38 MAPK-dependentSTAT1 phosphorylation in human plasmacytoid dendriticcell precursors. J Leukoc Biol 2002;72(5):1011-9.

23. Sfondrini L, Besusso D, Bronte V, Et al. CpG-Oligode-oxynucleotides activate tyrosinase-related protein 2-specificT lymphocytes but do not lead to a protective tumor-specificmemory response. Cancer Immunol Immunother 2004;
53(8):697-704.

interferon α and cpg oligodeoxynucleotides elicit additive immunostimulatory and antitumor effects

Documents