antimetastatic vaccination against lewis lung carcinoma with autologous tumor cells modified to...

Introduction

The past several years have witnessed a renewedinterest in active immunotherapy approaches for thetreatment of solid tumors. Many cytokine genes havebeen introduced into tumor cells with varying effectson both tumorigenicity and immunogenecity [1,2].Interleukin 12 was identified several years ago, butits potential use in cancer therapy has only been recognized relatively recently. IL-12 may be pro-

duced by both hematopoietic and nonhematopoieticcell types, although the major producers appear to be‘professional’ antigen-presenting cells, such asmonocytes, dendritic cells and activated B cells [3–5].IL-12 exerts a variety of biological effects on both Tand NK cells including the strong induction of IFNproduction and the promotion of TH1-type response[6,7]. IL-12 was also shown to induce in vivo anti-angiogenic effects [8]. Administration of IL-12 totumor-bearing animals enhanced the expression ofVLA4/FLA1-dependent T-cell migration to tumorsites indicating a direct role for IL-12 in lymphocytehoming to sites of tumor growth [9]. More recently,coadministration of an IL-12 expression cassette with

111112345678910111123456789201111234567893011112345678940111123456789501111234111

Clin. Exp. Metastasis, 1998, 16, 623–632

© 1998 Kluwer Academic Publishers Clinical & Experimental Metastasis Vol 16 No 7 623

Antimetastatic vaccination against Lewis lung carcinoma with autologous tumor cells modified toexpress murine Interleukin 12

Dan Popovic, Khaled M. El-Shami, Ezra Vadai, Michael Feldman, Esther Tzehoval and Lea Eisenbach

Department of Immunology, The Weizmann Institute of Science, Rehovot 76100, Israel

(Received 17 April 1998; accepted in revised form 12 June 1998)

Interleukin 12 (IL-12) is a disulfide-linked heterodimer molecule produced predominantly by professionalantigen presenting cells. It promotes the induction of sundry biological effects with significant relevance to antitumor immunity, such as enhancing a TH1 helper response, an in vivo antiangiogenic effect, induction of adhesion molecules that assist in lymphocyte homing to sites of tumor growth, and a directstimulatory effect on both T-cells and NK cells. We tested the efficacy of an antimetastatic vaccine composedof autologous murine D122 cells transfected with both subunits of IL-12 cDNA to express biologically-activeIL-12 molecule. Expression of IL-12 by D122 cells significantly reduced their tumorigenicity and metastaticpotential in immunocompetent syngeneic hosts. Furthermore, vaccination of mice with 2 � 106 irradiatedIL-12-transfected D122 cells engendered a protective CTL response which rejected a subsequent challengewith parental D122 cells and eradicated lung micrometastasis in animals whose primary tumors have beensurgically removed. The antitumor effects of IL-12 were mediated primarily by its ability to induce IFN expression in vivo. CD8+ T-cells as well as NK cells were crucial in the execution of the antitumoreffects of IL-12. These results suggest that autologous tumor cells expressing IL-12 by gene transfer are apotent antitumor vaccine able to induce a systemic immune response against poorly immunogenic and spontaneously metastatic tumors.

Keywords: interleukin 12, gene therapy, antimetastatic immunity, 3LL, gamma interferon

Address correspondence to: L. Eisenbach Department ofImmunology, The Weizmann Institute of Science, Rehovot 76100,Israel. Tel: +972–8–934 3108. Fax: +972–8–934 4141. E-mail:[email protected]

DNA vaccine formulations for HIV-1 Ag was foundto enhance the cellular immune response coupledwith reduction of specific Ab response [10].Vaccination of syngeneic animals with fibroblastsgenetically engineered to express IL-12 admixedwith tumor cells was shown to engender an antitu-mor immune response in murine models [11].Furthermore, intratumoral inoculation of a recom-binant adenovirus expressing the mIL-12 geneincreased significantly survival time of animals withexperimental liver metastasis [12]. In the presentstudy we seek to further explore the antitumorpotential of tumor cells manipulated to secrete IL-12by gene transfer, using a model of high metastatic3LL-D122 tumor [13]. We chose this model as metastasis rather than the primary growth constitutethe major cause of mortality and morbidity in humanmalignancies. Similarly D122 cells metastasizespontaneously in syngeneic C57BL/6 mice. Thus, the effect of IL-12 production on the metastatic competence of the producer cells and on theirimmunogenic capacity against the generation ofmetastasis by parental cells can be tested. We showthat the reduced malignancy of the highly metasta-tic, low immunogenic D122 cells supertransfectedwith the two subunits of IL-12 is a function of biologically-active IL-12 secretion. Irradiated IL-12secretors have the ability to induce a protective CTLresponse in vivo against a subsequent challenge ofparental cells as well as lung micrometastasis in micewhose primary tumor has been surgically resected.The effects of IL-12 was mediated essentially viainduction of IFN expression in vivo, while the antitumor immunotherapeutic effects were primarilyexecuted by CD8+ T lymphocytes as well as NK cells.Because cellular-mediated immunity is consideredcrucial in mounting antitumor immune response, ourresults show that IL-12, which particularly promotesCMI, appears to hold significant promise in the context of autologous cytokine gene-modified tumorvaccines.

Materials and methods

MiceEight- to 12-week-old C57BL/6 were bred in theWeizmann Institute (Rehovot, Israel). Animals weremaintained and treated according to NIH guidelines.

Tumor cells and cell culturesCells of the highly metastatic, poorly immunogenic,low Kb expressor D122 clone of 3LL Lewis lungcarcinoma, of C57BL/6 (H-2b) origin, were used as

parental cells for gene transfer. 39.5 is a lowmetastatic and highly immunogenic H-2Kb transfec-tant of D122 cells [14]. Cells were maintained inDMEM supplemented with 10% heat-inactivatedFCS (Biological Industries, Beth Haemek, Israel),2 mM glutamine, 1 mM nonessential amino acids,1 mM sodium pyruvate, and 0.4% combined antibi-otics. IL-12 transfectants were grown in maintenancemedium containing 600 �g/ml G418 (GIBCO).

Plasmid expression vectors and transfectionsMurine p35 and p40 subunits of IL-12 were clonedin BL-pSV expression plasmid vector. D122 cellswere cotransfected with these as well as pSV2-neoplasmid using standard calcium phosphate tech-niques [15]. To obtain stably transfected clones(D122-IL-12), transfected cells were grown in G418-containing medium for 14 days, and resistant clonespropagated separately, with subsequent determina-tion of IL-12 bioactivity in the supernatant. ControlD122 cells containing only the neomycin phospho-transferase gene (D122-neo) were prepared bytransfection with only pSV2-neo using the sameprocedures.

Screening of IL-12 transfectant by RT-PCRExpression of mRNA for both p35 and p40 subunitsin transfected clones was confirmed using PCR.mRNA was extracted from approximately 1 � 107

cells using a single step RNA extraction method withTri-reagent (Molecular Research Center, Cincinnati,OH). Complementary DNA was synthesized withMoloney murine leukemia virus reverse transcrip-tase (Promega, Madison, WI) using the mRNAsample. The sets of 5� and 3� primers for PCR were:GTGTCTTAGCCAGTCCCG and ATGGTCAC-GACGCGGGTG for p35 and GAGGTGGACTGGACTCCC and CAGGGAACACATGCCCACfor p40.

IL-12 T-cell growth factor bioassayTo determine the bioactivity of IL-12 produced bytransfectants, the proliferative response of day 3Concanavalin-activated lymphoblasts was measuredas previously described with minor modifications[14]. Results were extrapolated from the ‘King Leo’linearization method using a defined dose of recom-binant human IL-12 (Cetus Co., Emeryville, CA;specific activity : 18 � 106 IU/mg protein).

In vitro cytotoxicity assayMice were immunized i.p. three times at 7-day inter-vals with 2 � 106 irradiated (5000 rad) cells.Spleenocytes were harvested from immunized mice

111112345678910111123456789201111234567893011112345678940111123456789501111234111

D. Popovic et al.

624 Clinical & Experimental Metastasis Vol 16 No 7

10 days after the third booster and restimulated in vitro on monolayers of irradiated (5000 rad) and mitomycin C-treated (80 �g/ml/107 cells for 1hour). Restimulation was carried out for 5 days inRPMI supplemented with 10% FCS, 2 mM gluta-mine, and 2 � 10–5 M �-mercaptoethanol. Viablelymphocytes were separated by lympholyte centri-fugation (Cedarlane, Ontario, Canada) and admixedat different ratios with 5000 L-[35S] methionine-labeled target cells in U-shaped microtiter wells. The plates were incubated for 5 hours at 37°C.Cultures were terminated by centrifugation at 300 � g for 10 min. at 4°C, and 100 �l of supernatantswere assayed in a �-scintillation counter. Per-cent specific lysis was calculated as described before[15].

Injection of antibodies for depletion of immunocytesand IFN neutralizationMonoclonal Abs against CD4+ cells (GK1.5; ratmAb, IgG2b; Ref. 16), CD8+ cells (YTS-169–4; ratmAb IgG2b; Ref. 17), and NK1.1+ (PK136; mousemAbIgG2a; Ref.18) were injected to deplete subsetsof immune cells. Mice were given i.v. injections oncewith 100 �l (GK1.5, YTS-169–4) or 150–200 �l(PK136) of hybridoma ascites 5 days before inocu-lation of live transduced tumor cells or 4 days before the first vaccination in postsurgical immuno-therapy experiments. Starting 7 days after the firstantibody injection, mice were injected i.p. weekly(YTS-169–4, PK 136) or each third week (GK1.5)with 100 �l (GK1.5, YTS-169–4) or 150–200 �l(PK136) of hybridoma ascites. Mice given i.f.p. injec-tions of live transduced cells received a total of 7doses of YTS-169–4 or PK136 and 3 doses of GK1.5, whereas mice in postsurgical immunotherapyexperiments received a total of 5 doses of YTS-169–4 and PK136 and 2 doses of GK1.5. Accordingto direct immunofluorescence-activated cell sortinganalysis of splenocytes from depleted mice, specificdepletion achieved by injection of the different antibodies was at least 85% (data not shown).

Anti-IFN mAb, R4–6A2 was administered i.v. ata dose of 0.5 mg/mouse on the first day post-tumorchallenge i.f.p., followed by i.p. inoculation of thesame dose at 7-day intervals.

Tumor growth and spontaneous metastasisMice, 8–12 in each experimental group, were inoc-ulated i.f.p. with 2 � 105 cells/mouse. Local tumorgrowth was determined by measuring the footpaddiameters with calipers. To measure lung metastasis,the tumor-bearing leg was amputated below theknee when the tumor diameter reached 8–8.5 mm.

D122-injected mice were monitored daily and sacri-ficed when moribund. Accordingly, mice in othergroups were sacrificed. Metastatic loads wereassayed by weighing the lungs.

Experimental metastasisMice, 8–12 in each experimental group, were inoc-ulated i.v. with 5 � 105 tumor cells. D122-injectedmice were monitored daily and sacrificed when mori-bund. Accordingly, mice in the other groups weresacrificed. Metastatic loads were assayed as above.

ImmunotherapyProtection protocol. Mice were immunized i.p.weekly for three weeks with 2 � 106 irradiated(5000 rad) tumor cells. The mice were challengedwith 2 � 105 live D122 cells i.f.p. on day 10 after thethird booster. Tumor growth was monitored bymeasuring the diameter of the footpad with calipersand tumors were surgically removed as describedabove upon reaching a diameter of 8–8.5 mm.Development of lung metastasis was assayed asdescribed above.

Postsurgical protocol. The tumor-bearing leg wasamputated when the footpad of challenged micereached 6–6.5 mm. Zero to two days after amputa-tion, mice were immunized i.p. weekly, 4–5 timeswith 2 � 106 irradiated (5000 rad) tumor cells. Micewere sacrificed according to death from lung metas-tasis monitored for the nonimmunized controlgroup. In immunotherapy experiment coupled withdepletion of immunocytes subpopulations, the firstadministration of antibodies was done 4 days beforethe amputation of tumors and continued weeklyduring immunotherapy as described in ‘Injection ofantibodies for depletion of immunocytes’. Metastaticloads were assayed as described above.

Histological analysisTumor cells (2 � 105 cells/mouse) were inoculatedwith or without anti-IFN antibody as describedabove, and tumors were harvested 25 days after chal-lenge, fixed in Bouin’s solution, and embedded inparaffin. Serial 5-�m cuts were made from thesesamples and underwent hematoxylin and eosinstaining.

Statistical analysisStatistical analysis of the results of the animal exper-iments was calculated using a Student t-test. Thedifferences were considered statistically significantwhen p was < 0.05.

111112345678910111123456789201111234567893011112345678940111123456789501111234111

Cancer gene therapy using Interleukin 12

Clinical & Experimental Metastasis Vol 16 No 7 625

Results

Tumorigenicity and metastatic phenotype of IL-12secreting D122 cellsD122 cells were transfected with BL-pSV35 and BL-pSV40 plasmids which encode the light, p35 and heavy, p40 chains of IL-12 respectively. RT-PCRanalysis of the transfected clones designated BD9,BD55, and BD56 showed expression of both mes-sages of the IL-12 heterodimer. Wild-type D122 aswell as the Kb-transfected clone of D122 (39.5)showed low levels of p35 expression whereas neitherclone expressed any discernible levels of p40 (Figure1). Secretion of biologically-active IL-12 was calcu-lated using the ‘King Leo’ linearization methodagainst the standard curve of proliferation of mouseCon A blasts in response to mrIL-12. Table 1 showslevels of IL-12 secretion by different clones. SinceD122 cells express low amounts of MHC class-I Kb

molecules which correlates with their high malig-nancy, we examined the effect of IL-12 secretion onthe expression of class I. IL-12-transfected clonesexpressed almost similar amounts of class I on theirsurface as compared to their wild-type counterparts(Table 1).

We then examined the growth and metastaticcompetence of parental and modified cells in

syngeneic C57BL/6 mice. When 2 � 105 live cellswere injected i.f.p., the D122-IL-12 secretors BD9,BD55, and BD56 hardly grew in the footpads(Figure 2a), whereas wild-type D122 as well as thenon-secreting clone BD10 grew progressively.

Metastatic potential of parental and modified cellswas evaluated 30 days after i.v. inoculation of 5 �105 live cells into C57BL/6 mice. Figure 2b shows anabrogated metastatic potential of BD9, BD55, andBD56 IL-12-expressing clones as assayed by lungweight of inoculated mice which was in accordancewith the tumorigenicity assay of the same clones.

Immune mechanisms involved in the response toirradiated IL-12-expressing D122 cellsIn order to test whether cytotoxic T lymphocytes areinduced by irradiated D122-modified cells, weperformed in vitro CTL assays. By using parentaland gene-modified cells as targets, we could alsomonitor the changes in antigenicity resulting fromexpression of class I or cytokine. Immunization byparental D122 cells did not generate significantlevels of effector cells (Figure 3). D122 neo mock-transfected cells were similar to parental D122 cellsin induction of and sensitivity to CTL (data notshown). On the other hand, immunization by thehigh IL-12 secreting clone BD9 generated effectorcells that specifically killed the unmodified D122cells as well as other clones derived from it by genetransfection.

However, CTLs generated by immunization withthe highly immunogenic clone 39.5 (a Kb-transfectedD122 clone), used as a positive control for CTLinduction, manifested higher activity than thosegenerated with IL-12 secreting D122 clones. Theunrelated target cells YAC-1 as well as RMA-Slymphomas were not lysed by splenocytes from anyof the immunized groups (data not shown).

111112345678910111123456789201111234567893011112345678940111123456789501111234111

D. Popovic et al.


Figure 1. PCR amplification of p35 and p40 cDNA in3LL clones and IL-12 transfected 3LL clones. Both 5� and3� primers used for amplification of p35 and p40 subunitsare shown in ‘Materials and Methods’. The size of theamplified fragments are shown in comparison to the 1 kbladder marker.

Table 1. IL-12 detection assay was done as described in‘Materials and methods’. The results given are taken fromrepresentative assays. Expression of MHC class I mole-cules, detected by FACS staining for the MHC class Iallele Kb using the anti-Kb antibody (20–8–4), was compa-rable for both transfected and parental cells irrespectiveof their levels of IL-12 secretion

Clones IL-12 secretion MHC expression(IU) (% of positive cells)

D122 0. 00 37.4BD 9 46.40 37.6BD 55 11.48 39.8BD 56 9. 54 36.8BD 10 0. 00 37.1

Previous studies have shown that the immuneeffects of IL-12 are, at least partially, mediatedthrough induction of IFN expression in vivo . Inorder to check the importance of IFN in mediatingthe in vivo immune effects of IL-12 in our model,we administered the neutralizing mAb againstmIFN, R4–6A2, during an in vivo tumorigenicityassay starting from day 1 post-tumor inoculationi.f.p. in a weekly dose of 0.5 mg/mouse. Figure 4ashows that only after administration of anti-IFNAbs did the rate of tumor growth following inocu-

lation i.f.p. of BD9 cells become comparable to those of the parental D122 cells, whereas in theabsence of antibody or upon administration of anisotype-matched control antibody, D2.4, rejectionwas observed (Figure 4a). Similarly, the generationof specific CTLs from splenocytes taken 23 daysafter i.f.p. inoculation from mice which rejected the primary inoculum of BD9 cells was significantlyabrogated in the presence of neutralizing anti-IFNmAbs (Figure 4b). Histological evaluation of the site of tumor growth either by i.f.p. inoculation ofparental D122 or BD9 cells showed a histologicalpattern of tumor necrosis, macrophage accumulationand encapsulation by fibroblasts in the presence of local IL-12 production (Figure 5), whereas administration of R4–6A2 anti-IFN mAb led to agrowth pattern comparable to that of the wild-type tumor confirming the presence of a trendtoward reduction of the effectiveness of IL-12 in vivoupon administration of neutralizing anti-IFN anti-body.

Since IL-12 and IFN induce NK activity, affectmaturation of CTLs, and triggers the differentiationof TH0 into TH1 helper cells, we tested whichsubgroup of immunocytes are crucial for rejectionof BD9 cells. We injected BD9 and D122 cells i.f.p.into mice previously depleted with anti-CD4+

(GK1.5), anti-CD8+ (YTS-169–4), or anti-NK1.1(PK136) antibodies. D122 cells grew progressivelyand at similar rates in both depleted and undepleted

111112345678910111123456789201111234567893011112345678940111123456789501111234111



Figure 2. (a) C57BL mice (8 per group) were injectedi.f.p. with 2 � 105 cells of the high metastatic 3LL clone(D122) and with IL-12 transfected clones of D122.Twenty-four days later palpable tumors started to growand their diameters were measured. The figure shows themean tumor diameter on day 40 post-tumor inoculation.Statistical analysis of tumor size (Student t-test) shows p < 0.0001 for BD9, BD55, and BD56 relative to D122while p < 0.05 for BD10. Normal footpad diameter is2 mm. (b) C57BL mice (as above) were injected i.v. intothe tail vain with 5 � 105 cells of the high metastatic 3LLclone (D122) and with IL-12 transfected clones of D122(experimental metastasis model-see ‘Materials andmethods’). Thirty-five days later, when half of the controlgroup (D122 injected) died, mice in all groups were sacri-ficed and their lungs weighed. Statistical analysis with lungweights (Student t-test) shows p < 0.0001 for BD9, BD55,and BD56 relative to D122 while p > 0.05 for BD10.

Figure 3. In vitro lytic activity of CTLs elicited by D122,39.5, and IL-12 D122 transfectant (BD9). The data showpercent specific lysis obtained with E:T ratios of 100:1 and50:1. Target cells labeled with L-[35S]methionine, andmixed with the above ratios of effector cells in a 5 h assay.

mice. In nondepleted or CD4+ depleted animals,BD9 cells were rejected by 6 out of 8 mice, whereasin mice depleted of NK or CD8+ cells, 4 out of 8 and 3 out of 8 rejected the tumors respectively.Therefore, in our model, NK and CD8+ cellsbut not CD4+ cells affect the rejection of IL-12 secretors.

Immunotherapeutic potential of IL-12-modifiedD122 cellsBecause different D122 modified cells in irradiatedform have been shown to stimulate various hostimmune responses, we next tested the efficiency ofIL-12 gene-modified D122 cells to serve as

antimetastatic vaccine against parental D122 cells.C57BL/6 mice were immunized with i.p. inoculationof 2 � 106 irradiated D122 cells, BD9 cells, 39.5 cells, or a combination of BD9 and 39.5 cells.Animals were immunized three times at 7-day inter-vals and were challenged on day 10 after the lastboost with a tumorigenic dose of parental D122 cells. As shown in Figure 6a, the date of emergenceof palpable tumors in the group immunized withBD9 and combination of BD9 and 39.5 cells wassignificantly delayed compared to the naive mice ormice immunized with irradiated wild-type D122cells. Moreover, 7 out of 12 mice and 6 out of 12mice in 39.5+BD9- and BD9- immunized groupsrespectively showed total rejection of the primarytumors. In mice which did not reject the primarytumors, the tumor-bearing leg was amputated uponreaching a diameter of 8 mm. Twenty days post-amputation when half the mice in the control group(naive mice) succumbed to lung metastasis, mice inall groups were sacrificed and metastatic coloniza-tion of the lungs was assayed by lung weight. Figure6b shows significant reduction in lung deposits inmice immunized with BD9, 39.5, and combinationof BD9 and 39.5 (p < 0.0019, p < 0.0136, p < 0.0364respectively).

111112345678910111123456789201111234567893011112345678940111123456789501111234111

D. Popovic et al.


Figure 4. (a) Growth curves of D122 and BD9 in thepresence of anti-IFN mAb (R4–6A2) and the nonspe-cific isotype-matched mAb (D2.4). C57BL mice (8 pergroup) were injected i.f.p. with 2 � 105 D122 cells andgrowth rate was measured as described in ‘Materials andmethods’. Mice were injected i.p. with mAb (0.5 mg permouse) at weekly intervals. Tumor size is shown as a func-tion of the diameters of the tumor-bearing feet. (b) Invitro lytic activity of CTLs elicited by D122, D122 + R4 –6A2, BD9, BD9 + R4 – 6A2, and BD9 + D2.4. C57BLmice were inoculated i.f.p. with 2 � 105 cells and mAb asabove and spleen cells were removed and restimulated invitro on a monolayer of autologous, irradiated, and mito-mycin-C treated tumor cells for 5 days. Cytolytic activitywas determined against L-[35S]methionine labeled D122,39.5, BD9, and RMA-S target cells.

Figure 5. Histological analysis (hematoxylin-eosinstaining) of footpads from the mice injected with (A)naive; (B) D122; (C) D122 + R4 – 6A2; (D) BD9; (E) BD9 + R4–6A2; (F) BD9+D2.4 (magnification is �400).The histological pattern of tumor necrosis and fibroblastencapsulation of BD9 cells is abolished in the presence ofR4 – 6A2 mAb.

We then further studied whether immunizationwith these clones can cure micrometastasis and thus prevent the establishment of macrometastasisin tumor-bearing mice employing the postsurgicalimmunotherapy protocol. Naive mice were chal-lenged i.f.p. with a tumorigenic dose of D122 cells,and the tumor-bearing leg was amputated uponreaching a diameter of 6–6.5 mm. Mice were thenimmunized by i.p. inoculation of 2 � 106 irradiatedcells five times at 7-day intervals. Mice were sacri-ficed when half the animals in the control group(naive mice) succumbed to lung metastasis. Asdepicted in Figure 7a, BD9 and 39.5 clones werecomparably effective in protecting against lungmetastasis by D122 tumor cells (p < 0.048 and p < 0.05 respectively) whereas a combination of BD9and 39.5 was an efficient antimetastatic vaccine (p < 0.017) compared to vaccination with irradiatedwild-type D122 cells. However, this doubly modifiedtumor cell vaccine did not show a statistically significant advantage over the singly modified tumorcells (Figure 7b). When immunocytes subpopula-tions (CD4+, CD8+, and NK) were depleted at thetime of vaccination, the average lung weights ofgroups depleted of CD8+ or NK cells were higherthan those in the nondepleted group or the groupdepleted of CD4+ cells, confirming the role of CD8+

and NK cells in therapy of lung micrometastasis(data not shown).

Discussion

IL-12 is a cytokine with potent antitumor activityagainst various tumors models [16–20]. Renal adeno-carcinoma (RENCA), melanoma (B16), reticulumcells sarcoma (M5076), sarcoma (MCA-105 andMCA-207), Lewis lung carcinoma (3LL), and coloncarcinoma (MC-38 and CT-26) respond to systemicrecombinant IL-12 administration, leading to regres-sion of established tumors and to long-term animalsurvival in some cases. However, systemic toxicitiesassociated with immunotherapy remains to be over-come. Preliminary results in one clinical trialemploying recombinant IL-12 showed severe toxi-city, with 2 patients out of 17 dying and the restexperiencing adverse side effects in the form ofgastrointestinal bleeding, asthenia, and hepatotoxi-city [21]. In this study we sought to modify tumorcells genetically with IL-12 cDNA so that the tumorcell could supply the cytokine of interest in aparacrine fashion to the antitumor responder cellspresent with, or in the vicinity of, the tumor. Thisparacrine physiology much more closely mimics thenatural biology of cytokine action than does thesystemic (e.g. intravenous) administration of recom-binant cytokines [22]. In a mammary adenocarci-noma model, it was shown that autologous tumorcells engineered to express IL-12 exhibited compa-rable efficiency to the systemic administration of

111112345678910111123456789201111234567893011112345678940111123456789501111234111



Figure 6. (a) Growth of D122 tumors and C57BL mice (12 per group) preimmunized with D122 IL-12 transfectants.Mice were immunized i.p. (see ‘Methods’) three times at one week intervals with 2 � 106 irradiated cells and challengedwith 2 � 105 live D122 cells i.f.p. 10 days after the last immunization. Tumor growth rate was measured as described in‘Methods’. (b) Lung weights of C57BL mice in a protection experiment. Mice were immunized 3 times i.p. at weeklyintervals with 2 � 106 irradiated D122, 39,5, BD9, and 39.5+BD9 and then challenged with 2 � 105 live D122 cells i.f.p.10 days after the last immunization (see ‘Methods’). Tumor-bearing leg was amputated when the footpad reached adiameter of 8 – 8.5 mm (approximately 25 days post-tumor injection). After 20 days, when half of the control nonim-munized group died, mice in all groups were sacrificed and the metastatic load was determined by weighing the lungs.Statistical analysis of lung weights (Student t-test) shows p = 0.0019 for BD9; p = 0.0163 for 39.5; p = 0.0364 for 39.5 +BD9 relative to D122. Each dot in the graph represents the lung weight of a single mouse.

nontoxic doses of recombinant IL-12 in inducing therejection of small primary tumors but exhibitedslightly less efficacious response than the systemicadministration against larger tumors [23]. Anothergoal of this study is to evaluate the efficacy of IL-12 gene-modified cellular vaccines and vaccinecombination in a context that depicts more accu-rately the clinical situation. Of great interest is thepostsurgical immunotherapy protocol, made feasibleby the spontaneous metastatic potential of the D122model, in which vaccinations begin after amputationof the primary tumor. Thus the potential of thevaccine to cure established lung micrometastasescould be more accurately assessed. In a recent exper-iment , autologous tumor cells modified to expressIL-12 had a significant antimetastatic effect in anexperimental lung metastasis model using aBALB/c-derived colon carcinoma [24]. Our modelenabled us to explore the immunotherapeutic poten-tial of IL-12-modified tumor vaccines against a spon-taneously developing lung metastasis in a contextsimilar to the clinical situation of postsurgicalminimal residual disease.

We analyzed the malignant phenotype, immuno-genecity, and the immune mechanisms induced byD122 tumor cells manipulated by gene transfer toexpress murine IL-12 cytokine. Two cDNAs of the

p35 and p40 were transfected into D122 cells andRT-PCR-mediated characterization of positiveclones at the level of transcription was done. Wild-type D122 as well as its Kb-transfectant clone 39.5showed expression of p35 mRNA. This findingextended previous observations that detected p35mRNA in both murine lymphoid and non-lymphoidtissues as well as in some human tumors [25, 26].Expression of biologically-active IL-12 by D122tumor cells almost completely abrogated their abilityto grow as tumors in the footpads of syngeneicC57BL/6 mice. The metastatic potential of the D122-IL-12 clones were also abolished when injectedintravenously.

Zitvogel et al. [27] used NIH 3T3 fibroblasts trans-duced with a retroviral vector expressing IL-12admixed with tumor cells as a IL-12-derived anti-tumor vaccine. An antitumor response was observedwhether the fibroblasts were administered at the siteof the tumor or at a distant site, pointing to asystemic effect consequent upon the systemicdelivery of the cytokine. However, the antitumoreffect developed more rapidly, albeit with compa-rable efficacy, when the fibroblasts were adminis-tered at the tumor site compared to the effectobtained when the fibroblasts were inoculated at adistant site from the tumor growth, which indicates

111112345678910111123456789201111234567893011112345678940111123456789501111234111

D. Popovic et al.


Figure 7. (a) Postsurgical immunotherapy of mice bearingD122 micrometastases. C57BL mice (10 per group) wereinjected with 2 � 105 live D122 i.f.p. When the tumorreached the diameter of 6–6.5 mm, legs were amputated(see ‘Methods’). Three days post-amputation, all groupsstarted receiving immunizations with 2 � 106 irradiatedcells i.p. at 7-day intervals for 5 weeks. After 29 days,when half the mice in the control nonimmunized groupdied, animals in all groups were sacrificed and metastaticload determined by weighing the lungs. Statistical analysisof lung weight (Student t-test) shows p < 0.05 for 39.5; p = 0.048 for BD9; and p = 0.017 for 39.5 + BD9 relativeto D122. Each dot in the graph represents a single lung.(b) Staining in Bouin’s solution of lungs excised from thesame groups of animals in the postsurgical immunotherapyexperiment.

(b)

(a)

an advantage in having IL-12 expression restrictedat the tumor microenvironment. In our model, morethan half of mice in groups preimmunized with theD122-IL-12 clones rejected a subsequent lethal chal-lenge of parental D122 cells. Furthermore, immu-nized mice which developed tumors did so at a muchlower rate than mice vaccinated with irradiated wild-type D122. This rejection was mediated by a CTLresponse engendered upon vaccination with the IL-12-secreting D122 clones whereas animals vacci-nated with irradiated parental D122 cells did notdevelop significant specific CTL response. This CTLwas able to reduce lung micrometastasis in a post-surgical immunotherapy experiment. As previouswork on combining MHC class I and cytokine-trans-fected autologous tumor cells showed higher efficacythan vaccination with either modified clone alone[28], we sought to test the existence of synergismbetween IL-12 expressing clones and MHC class Itransfected clones. In our model, combined vacci-nation with IL-12 D122 expressors in addition to thehigh Kb expressor D122 clone 39.5 exhibited some-what higher, albeit statistically insignificant, efficacyin eradicating pulmonary micrometastasis. Thephenotypes of the IL-12-secreting D122 cells wasdependent on CD8+ as well as NK cells whereasdepletion of CD4+ cells did not have any effect onthe primary rejection of IL-12-modified D122 cells.The immunocytes participating in the rejection oflive modified tumor cells overlapped with mecha-nisms induced by i.p. inoculation of irradiated modi-fied cells which were used as a cellular vaccine, i.e.the antimetastatic effect was dependent on CD8+

and NK cells while depletion of CD4+ did not havean effect on the vaccine efficacy. The fact that deple-tion of CD4+ did not adversely influence the IL-12-mediated antitumor response might be accountedfor by the ability of IL-12 to induce IFN expres-sion in vivo , thus bypassing the need for TH1 assis-tance in the induction of cellular immunity. Thisinduction of IFN expression by activated NK andT cells is central to IL-12-mediated-immune effectsin vitro and in vivo [29–31]. Expression of IFNmRNA is remarkably enhanced following systemicadministration of IL-12 [32] and IL-12-mediatedtumor rejection was found to be, at least partially,dependent on its unique ability to induce IFNexpression [33, 34]. In our model, we observedalmost complete abrogation to the antitumorimmune mediated effects, both in vivo and in vitroof IL-12-expressing D122 clones upon administra-tion of neutralizing anti-IFN antibodies.

These results confirm and extend previous obser-vations supporting the potent antitumor effects of

IL-12 in the context of cytokine gene modificationof autologous tumor cells and provide a rationalefor the use of this vaccination modality in postsur-gical treatment of cancer patients.

Acknowledgements

This work was supported by the Bundesministeriumfür Forschung und Technologie (BMFT) and theIsraeli Ministry of Science and Technology (MOST),Joint German–Israeli Research Program.

References1. Dranoff G, Jaffee E, Lazenby A, et al., 1993,

Vaccination with irradiated tumor cells engineered tosecrete murine granulocyte-macrophage colony stim-ulating factor stimulates potent, specific, and long-lasting anti-tumor immunity. Proc Natl Acad Sci USA,90, 3539–43.

2. Colombo MP and Rodolfo M, 1995, Tumor cellsenginnered to produce cytokines or cofactors ascellular vaccines: do animal studies really support clin-ical trials ? Cancer Immunol Immunother, 41, 265–70.

3. Schoenhaut DS, Chua AO, Wolitzky AG, et al., 1992,Cloning and expression of murine Il-12. J Immunol,148, 3433–40.

4. Banks, RE, Patel, PM and Selby PJ, 1995, Interleukin12: a new clinical player in cytokine therapy. Br JCancer, 71, 655–9.

5. Tahara H Lotze MT, 1995, Antitumor effects of inter-leukin 12 (IL-12): applications in the immunotherapyand gene therapy of cancer. Gene Therapy, 2, 96–106.

6. Romagnani S, 1995, Biology of human TH1 and TH2cells. Clinical Immunol, 15, 121–9.

7. Trinchieri G, 1993, Interleukin 12 and its role in thegeneration of TH1 cells. Immunol Today, 14, 335–338.

8. Voest EE, Kenyon BM, O’Reilly MS, et al., 1995,Inhibition of angiogenesis in vivo by interleukin 12. J Natl Cancer Inst, 87, 581–6.

9. Ogawa M, Tsutsui T, Zou J, et al., 1997, Enhancedinduction of very late antigen 4/lymphocyte function-associated antigen 1-dependent T-cell migration totumor sites following administration of interleukin 12.Cancer Res, 57, 2216–22.

10. Kim JJ, Ayyavoo V, Bagarazzi ML, et al., 1997,In vivo engineering of a cellular immune response bycoadministration of IL-12 expression vector with aDNA immunogen. J Immunol, 158, 816–26.

11. Tahara H, Zeh III HJ, Storkus WJ, et al., 1994,Fibroblasts genetically engineered to secrete inter-leukin 12 can suppress tumor growth and induce anti-tumor immunity to a murine melanoma in vivo. CancerRes, 54, 182–9.

12. Caruso M, Pham-Nguyen K, Kwong Y, et al., 1996,Adenovirus-mediated interleukin-12 gene therapy formetastatic colon carcinoma. Proc Natl Acad Sci USA,93, 11302–11306.

13. Porgador A, Bannerji R, Watanabe Y, et al., 1993,Antimetastatic vaccination of tumor-bearing mivewith two types of IFN- gene-inserted tumor cells. J Immunol, 150, 1458–70.

111112345678910111123456789201111234567893011112345678940111123456789501111234111



14. Gately, MK and Chizzonite, R, 1992, Measurement of human and mouse interleukin-12. In: Coligan, JE,Kruisbeek, AM, Margulies, DH, Shevach, EM andStrober, W., eds. Current Protocols in Immuno-logy,. New York: Green Publishing and Wiley-Interscience, pp. 6.

15. Porgador A, Tzehoval E, Katz A, et al., 1992, IL-6gene transfection into 3LL tumor cells suppresses themalignant phenotype and confers immunotherapeuticcompetence against parental metastatic cells. CancerRes, 52, 3679–86.

16. Brunda MJ, Luistro L, Warrier RR, et al., 1993,Antitumor and antimetastatic activity of interleukin 12against murine tumors. J Exp Med, 178, 1223–30.

17. Takeda K, Seki S, Ogasawara K, et al., 1996, LiverNK1.1+ CD4+ �� T cells activated by IL-12 as a majoreffect in inhibition of experimental tumor metastasis.J Immunol, 156, 3366–73.

18. Hashimoto W, Takeda K, et al., 1995, Cytotoxic NK1.1 Ag+ �� T cells with intermediate TCR induced inthe liver of mice by IL-12. J Immunol, 154, 4333–40.

19. Nastala CL, Edington HD, McKinney TC, et al., 1994,Recombinant IL-12 administration induces tumorregression in association with IFN- production.J Immunol, 153, 1697–706.

20. Martinotti A, Stoppacciaro A, Vagliani M, et al., 1995,CD4 T cells inhibit in vivo the CD8-mediated immuneresponse against murine colon carcinoma cells trans-duced with interleukin 12 genes. Eur J Immunol, 25,137–46.

21. Lamont AG, Adorini L, 1996, IL-12: A key cytokinein immune regulation. Immunol. Today, 17, 214–7.

22. Pardoll DM, 1995, Paracrine cytokine adjuvants incancer immunotherapy. Annu Rev Immunol, 13,399–415.

23. Cavallo F, Signorelli P, Giovarelli M, et al., 1997,Antitumor efficacy of adenocarcinoma cells engi-neered to produce interleukin 12 (IL-12) or othercytokines compared with exogenous IL-12. J NatlCancer Ins, 89, 1049–58.

24. Rodolfo M, Zilocchi C, Melani C, et al., 1996,Immunotherapy of experimental metastases by vacci-nation with interleukin gene-transduced adenocarci-noma cells sharing tumor-associated antigens:

Comparison between IL-12 and IL-2 gene-transducedtumor cell vaccines. J Immunol, 157, 5536–42.

25. Schoenhaut DA, Chua A, Wolitzky A, et al., 1992,Cloning and expression of murine IL-12. J Immunol,148, 3433–40.

26. D’Andrea A, Rengaraju M, Valliante N, et al., 1992,Production of NKSF (IL-12) by peripheral bloodmononuclear cells. J Exp Med, 180, 211–22.

27. Zitvogel L, Tahara H, Robbins BD, et al., 1995, Cancerimmunotherapy of established tumors with IL-12:effective delivery by genetically-engineered fibro-blasts. J Immunol, 155, 1393–403.

28. Porgador A, Tzehoval E, Vadai E, et al., 1995,Combined vaccination with major histocompatibilityclass I and interleukin 2 gene-transdused melanomacells synergises the cure of postsurgical establishedlung metastases. Cancer Res, 55, 4941–9.

29. Gately MK, Warrier RR, Honasoge S, et al., 1994,Administration of recombinant IL-12 to normal miceenhances cytolytic lymphocytes activity and inducesproduction of IFN- in vivo. Int Immunol, 6, 157–67.

30. Manetti R, Gerosa F, Giudizi MG, et al., 1994,Interleukin 12 induces stable priming for interferon (IFN-) production during differentiation of human T helper (Th) cells and transient IFN- production inestablished Th2 cell clones, J Exp Med, 179, 1273–83.

31. Seder RA, Gazzinelli R, Sher A and Paul WE, 1993,IL-12 acts directly on CD4+ T cells to enhance primingfor IFN- production and diminishes IL-4 inhibitionof such priming. Proc Natl Acad Sci USA, 90,10188–93.

32. Tannebaum CS, Wicker N, Armstrong D, et al., 1996,Cytokine and chemokine expression in tumors of micerecieving systemic therapy with IL-12. J Immunol, 156,693–9.

33. Zou J-P, Yamamoto N, Fuji T, et al., 1995, Systemicadministration of rIL-12 induces complete tumorregression and protective immunity: Response iscorrelated with a striking reversal of suppressed IFN- production by anti-tumor T cells. Int Immunol, 7,1135–45.

34. Fallarino F, Uyttenhove C, Boon T and Gajewski TF,1996, Endogenous IL-12 is necessary for rejection ofP815 tumor variants in vivo. J Immunol, 156, 1095–100.

111112345678910111123456789201111234567893011112345678940111123456789501111234111

D. Popovic et al.


antimetastatic vaccination against lewis lung carcinoma with autologous tumor cells modified to...

Documents