IL-10 gene transfer to intracranial 9L glioma: tumor inhibition and cooperation with IL-2

Ž .Journal of Neuroimmunology 92 1998 50–59

IL-10 gene transfer to intracranial 9L glioma: tumor inhibition andcooperation with IL-2

Adam A. Book a,b, Kevin E. Fielding b, Namita Kundu c, Mary Ann Wilson b,d,Amy M. Fulton c, John Laterra a,b,d,e,)

a Department of Neuroscience, The Johns Hopkins UniÕersity School of Medicine, Baltimore, MD, USAb The Kennedy Krieger Research Institute, 707 N. Broadway, Baltimore, MD 21205, USA

c Department of Pathology and The Greenebaum Cancer Center, UniÕersity of Maryland, 22 S. Greene Street, Baltimore, MD 21201, USAd Department of Neurology, The Johns Hopkins UniÕersity School of Medicine, Baltimore, MD, USAe Department of Oncology, The Johns Hopkins UniÕersity School of Medicine, Baltimore, MD, USA

Received 20 April 1998; accepted 30 June 1998

Abstract

Ž .This study examines the effects of interleukin-10 IL-10 and combination IL-10q IL-2 gene transfer on experimental brain tumorŽ .growth in vivo. 9L gliosarcoma cells were engineered to stably express murine IL-10 9L–IL-10 cells and implanted subcutaneously or

to the caudaterputamen of syngeneic rats. The growth of tumors expressing IL-10 was substantially reduced compared to that of controlŽ .tumors p-0.05 . Intracranial tumors expressing IL-10 and IL-2 were established by co-implanting 9L–IL-10 cells with endothelial cells

engineered to express IL-2. At 14 days post-implantation, tumors expressing IL-10q IL-2 were 99% smaller than control-transfectedŽ .tumors p-0.0001 . This extent of anti-tumor effect could not be achieved by expression of IL-10 or IL-2 alone within tumors. Neither

Ž .IL-10 nor a combination of IL-10q IL-2 gene delivery inhibited tumor growth in severe combined immunodeficient SCID-Beige miceŽ .p)0.05 . Immunohistochemical analysis revealed that IL-10q IL-2 gene delivery markedly increased T-cell infiltration within thestriatum ipsilateral to tumor cell implantation. These findings establish that IL-10 expression, particularly in combination with IL-2expression, can have significant immune-dependent anti-tumor actions within intracranial gliomas. q 1998 Elsevier Science B.V. Allrights reserved.

Keywords: Anti-tumor immunity; Gene therapy; Neuroimmunology; Cytokine

1. Introduction

The malignant gliomas are the most aggressive of pri-Ž .mary human central nervous system CNS tumors. Their

pronounced proliferative activity, invasiveness, and angio-Ž .genesis Brem et al., 1972; Burger et al., 1985 coupled

with their relative isolation by the blood-brain barrierrenders them very resistant to both surgical and systemictherapies. Median survival for the most malignant of thesetumors, glioblastoma multiforme, remains less than 1 yearŽ .Neuro-oncology, 1990 .

While distal metastasis of malignant glioma is ex-tremely rare, its tendency to diffusely invade brain empha-sizes the need for therapies capable of targeting tumor

) Corresponding author. Tel.: q1 410 5029494; fax: q1-410-5028003;e-mail: [email protected]

cells throughout brain. One such approach involves theactivation of anti-glioma immune responses. Promisingmethods include the intratumoral expression of therapeutictransgenes encoding specific immune-activating cytokinesor the expression of antisense sequences that inhibit im-

Žmunosuppressive cytokines Trojan et al., 1993; Glick et.al., 1995; Fakhrai et al., 1996 . One prototypic immune-

Ž .activating cytokine, interleukin-2 IL-2 , inhibits solid tu-mor growth by virtue of its ability to stimulate T-lympho-

Ž .cyte and natural killer NK cell cytotoxicity. IL-2 genetransfer has been shown to inhibit the growth of experi-mental tumors and to induce anti-tumor immunity in cer-

Žtain brain tumor model systems Gansbacher et al., 1990;.Roth et al., 1992; Levraud et al., 1997 . Similar anti-glioma

Žresponses have resulted from interleukin-4 Yu et al.,. Ž . Ž .1993 and interferon-gamma IFNg Tjuvajev et al., 1995

gene delivery. Therapeutic efficacy has been limited byinefficient transgene expression, inadequate host response,

0165-5728r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.Ž .PII: S0165-5728 98 00172-6

( )A.A. Book et al.rJournal of Neuroimmunology 92 1998 50–59 51

Žand cytokine-associated toxicities Fujita et al., 1991;.Hanisch et al., 1996 .

Ž .Interleukin-10 IL-10 , produced by the Th subset of2

CD cells, was originally discovered on the basis of its4

ability to suppress cytokine production by the Th1 subsetq Ž .of CD4 helper T-lymphocytes Fiorentino et al., 1989 .

IL-10 was subsequently shown to also inhibit the produc-tion of numerous pro-inflammatory cytokines by mono-

Ž .cytes Malefyt et al., 1991 . IL-10 expression has beendetected in human malignant gliomas and at higher levels

Ž .in malignant vs. low grade tumors Huettner et al., 1995 .This has led to the hypothesis that endogenous IL-10functions to suppress anti-glioma immunity within brain.However, the response of CNS tumors to transgenic IL-10has not been determined. Despite the potentially immuno-suppressive and anti-inflammatory actions of endogenousIL-10, evidence is mounting that transgenic IL-10 pro-duced at high levels by engineered tumor cells can inhibitgrowth of systemic tumors by either stimulating anti-tumorimmunity or inhibiting tumor-associated angiogenesisŽSchwarz et al., 1994; Giovarelli et al., 1995; Kundu et al.,

.1996; Huang et al., 1997 .In the present report, we examine the response of

subcutaneous and intracranial rat 9L gliomas to IL-10 genetransfer. Our findings establish for the first time that IL-10gene delivery can have significant anti-glioma actionswithin the rat CNS. Since the Th1 cytokine IL-2 has beenshown to synergize with the Th cytokine IL-10 in vitro2ŽChen and Zlotnik, 1991; Fluckiger et al., 1993; Itoh et al.,

.1994 and the combined effects of these cytokines had notbeen evaluated previously in any tumor model, we exam-ined the effect of IL-2 gene transfer in combination withIL-10 gene transfer on glioma growth. We show that IL-10and IL-2 cooperate in vivo to significantly inhibit gliomagrowth within the brain and to prolong animal survival.Finally, we show that an anti-tumor immune responseunderlies the inhibition of glioma growth following IL-10gene transfer and combination IL-10q IL-2 gene transfer.

2. Materials and methods

2.1. Cell culture

The 9L cell line was originally established from anitrosourea-induced gliosarcoma in Fisher 344 ratsŽ .Schmidek et al., 1971 . 9L cells and endothelial cells wereboth grown at 378C in 5% CO r95% air in Dulbecco’s2

Žmodified Eagle’s medium DMEM; Mediatech, Washing-. Ž .ton, DC supplemented with 10% vrv fetal bovine serum

Ž .HyClone , 2 mM L-glutamine, 50 mlrml gentamycin, andŽ .300 mgrml Geneticin G418, GIBCO . Endothelial cell

cultures were supplemented with 5 ngrml basic fibroblastgrowth factor. Endothelial cells were originally isolatedfrom brains of Lewis rats and immortalized by transfectionwith adenovirus 2 EIA gene under transcriptional control

Ž .of the SV40 promoter Roux et al., 1994 . One clone,designated RBE4, was subsequently transfected with thereplication-defective MFG–NB retroviral vector contain-

Ž . Ž .ing a modified lacZ gene nls–lacZ Lal et al., 1994 .ŽThese cells, designated RBEZ kindly provided by Pierre-

.Olivier Couraud, Neurotech, Paris, France , were culturedon fibronectin-coated substrata as previously describedŽ .Lal et al., 1994 .

2.2. Cytokine gene transfections

Ž .Murine IL-10-producing 9L cells 9L–IL-10 were cre-ated by transfection with the plasmid pBMGneo. IL-10 in

Ž . Žthe presence of lipofectamine GIBCO Kundu et al.,.1996 . 9L–neo control cells were produced by transfection

under identical conditions with the plasmid pBMGneolacking the IL-10 cDNA insert. Stable transfectants were

Žselected in the presence of G418 300 mgrml; Life Tech-.nologies .

Ž .Murine IL-2-producing endothelial cells RBEZ–IL-2Žwere constructed as previously described Nam et al.,

.1996 by transfecting RBEZ cells in the presence ofŽlipofectamine with the plasmid pBCMGhygro.IL-2 kindly

. Žprovided by Dr. Drew Pardoll Karasuyama and Melch-.ers, 1988 . Control RBEZ–hygro cells were constructed by

transfecting RBEZ cells under identical conditions with thepBCMGhygro expression vector lacking the IL-2 cDNAinsert. Stable transfectants were selected in the presence of

Ž . Žhygromycin B 250 mgrml as previously described Nam.et al., 1996 .

2.3. Assays for cytokine production by transfected cells

ŽTransfected clonal cell lines 9L–neo, 9L–IL-10,.RBEZ–hygro, and RBEZ–IL-2 were grown to confluence

in 24-well tissue culture plates. Cells were subsequentlyŽ .incubated with serum-free DMEM 0.5 mlrwell at 378C

for 24 h. Conditioned media were removed, centrifuged,and supernatants were assayed by ELISA according to the

Žsupplier of capture and detection antibodies PharMingen,.San Diego, CA . Standard curves were established using

Ž .purified recombinant murine IL-10 and IL-2 PharMingen .Biological activity of IL-2 produced by RBEZ cells wasconfirmed using the IL-2 dependent CTLL cell line as

Ž .previously reported Nam et al., 1996 .

2.4. Tumor and endothelial cell implantation

Tumor cells were harvested, counted using a CoulterŽ .counter Coulter Electronics, Hialeah, FL , and resus-

pended in sterile DMEM immediately before implantationto host animals. For intracranial implantation into the

Ž .caudate–putamen of anesthetized 200–250 g adult maleŽFisher 344 rats Charles River Laboratories, Wilmington,

. Ž 5 .MA , tumor cells alone 10 cells; 9L–neo or 9L–IL-10

( )A.A. Book et al.rJournal of Neuroimmunology 92 1998 50–5952

Table 1Expression of IL-10 by transfected 9L glioma cells in vitro and in vivo

Ž .IL-10 ngrml

Conditioned media9L control 2.3"0.29L–IL-10 )106.2"4.9

Plasma9L control 0.05"0.019L–IL-10 )1.67"0.59

Ž .IL-10 within conditioned media of control-transfected 9L cells ns4Ž .and IL-10-transfected 9L cells ns3 or within plasma obtained from

ŽSCID-Beige mice bearing intracranial control-transfected 9L tumors ns. Ž .6 and IL-10-transfected 9L tumors ns6 was determined by ELISAŽ .limit of assay sensitivity G31 pgrml .Conditioned media was collected for 24 h from confluent cell monolay-ers.Plasma was obtained from mice bearing intracranial 9L-control and

Ž .9L–IL-10 tumors of comparable size see Fig. 5B at the time of sacrificeŽ .post-implantation day 14 .) p-0.001, Student’s t-test.Data represents means"S.E.M.

Ž 5.or a mixture of tumor cells 10 and endothelial cellsŽ 6 .2=10 cells; RBEZ–hygro or RBEZ–IL-2 in DMEMŽ .2–5 ml were injected stereotactically with a 26-gauge,beveled-tip Hamilton syringe according to our previously

Ž .described methods Nam et al., 1996 . Briefly, injectionswere made 3.0 mm to the right of bregma, at a depth of 4.5mm from the dural surface. Cells were injected over a2-min period and the needle was left in place for 2 minafter injection and then withdrawn slowly to limit leakage.Intracranial implantations of 9L and endothelial cells into

the caudaterputamen of severe combined immune defi-Žcient SCID-Beige mice C.B-17rIcrCR1-SCIDrBeige;

.Charles River were performed in a similar manner to ratimplantations. For mice, injections were made 2.0 mm tothe left of bregma, at a depth of 2.5 mm from the duralsurface. For subcutaneous implantation into the anteriorflanks of anesthetized Fisher 344 rats or SCID-Beige mice,

Ž 5 .tumor cells 10 ; 9L–neo or 9L–IL-10 in sterile DMEMŽ .100 ml were injected bilaterally using a 22-gauge syringe

Ž .as previously described Arosarena et al., 1994 . Subcuta-neous tumors were measured with dial calipers and vol-

Žumes were calculated using the formula: Volumes length2 . Ž=width r2, as previously described Tamargo et al.,

.1990 .

2.5. Histology

At the indicated post-implantation times, rats were anes-thetized and transcardially perfused with 100 ml of phos-

Ž .phate-buffered saline PBS followed by 350 ml of fixativeŽ .4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4 .Mice were perfused with 75 ml of PBS followed by 150ml of fixative. Brains were removed, placed in fixative for

Ž .4 h, and cryoprotected in 15% and then 30% wrvŽ .sucrose in PBS. Coronal sections 30 mm through the

level of the tumor were cut using a freezing microtomeŽ .Leitz . A series of sections from each brain was stained

Ž .with hematoxylin and eosin H and E , and maximal tumorcross-sectional areas were determined by computer-as-

Ž .Fig. 1. Inhibition of subcutaneous and intracranial 9L glioma growth by IL-10 gene transfer. A Systemic IL-10 and control tumors were generated by5 Ž .subcutaneously injecting 10 9L–IL-10 or 9L–neo control cells, respectively, into the flanks of Fisher 344 rats. Tumor growth was monitored biweekly

using tissue calipers. Tumor volumes were calculated as described in Section 2. Subcutaneous IL-10 tumors grew at a significantly slower rate compared toŽ . Ž . 5control tumors Fisher’s PLSD, ) p-0.05, )) p-0.01, qp-0.001 . B Intracranial IL-10 and control tumors were generated by implanting 10

Ž .9L–IL-10 or 9L–neo control cells, respectively, to the right caudaterputamen of adult Fisher 344 rats. Animals were sacrificed 11 days later and tumorvolumes were calculated as described in Section 2. Intracranial IL-10 tumors were substantially smaller in volume than control tumors at 11 days

Ž .post-implantation p-0.05, unpaired Student’s t-test . Data represents mean"S.E.M.


sisted image analysis using the Microcomputer ImagingŽ . ŽDevice MCID M4 software package Imaging Research,.Canada . Intracranial tumor volumes were estimated by the

Žformula: Volumes square root of maximal tumor cross-.3 Ž .sectional area , as previously described Nam et al., 1996 .

For the immunohistochemical detection of T-cells, brainsections were placed for 1 h at room temperature in asolution containing 3% normal horse serum, 0.2% TritonX-100, 0.2% porcine gelatin, and 0.02% sodium azide inPBS. Sections were then incubated for 3 days at 48C inMRC OX-52 at 1:1500, diluted in the above diluent. MRC

Ž .OX-52 Serotec recognizes a cell surface antigen which islargely restricted to cells of the T-lymphocyte lineageŽ .Robinson et al., 1986 . Antibody staining was visualizedby the avidin–biotin–peroxidase complex method, using a

Ž . ŽVectastain Elite ABC kit Vector Laboratories Hsu et al.,.1981 . Following a 9 min incubation in diaminobenzidine

and three rinses, sections were mounted on slides, dehy-drated through an alcohol gradient, cleared in Hemo–De,and coverslipped.

3. Results

3.1. Inhibition of 9L tumor growth by IL-10 gene transfer

The in vitro production of IL-10 by stably transfectedclonal 9L cell lines was confirmed by an enzyme-linked

Ž . Ž .immunoadsorbent ELISA assay Richter et al., 1993 . Aclonal cell line that produced IL-10 at levels approximately50-fold higher than control cells was selected for subse-

Ž .quent in vivo studies Table 1 . Production of IL-10 invivo was determined by comparison of plasma samplesobtained from animals bearing intracranial 9L–IL-10 or

Ž .9L–neo control tumors. The approximate 30-fold in-crease in plasma IL-10 levels found in animals implantedwith IL-10-secreting tumor cells confirmed the secretion of

Ž .transgenic IL-10 in vivo Table 1 .To determine whether transgenic IL-10 production

Ž .within subcutaneously s.c. implanted tumors was growthinhibitory, tumor growth was examined following the s.c.

Ž .implantation of 9L–IL-10 or 9L–neo control tumor cells

Fig. 2. Histological evaluation of the effects of single cytokine and combined IL-10q IL-2 gene delivery on intracranial 9L tumor growth. Control, IL-2,5 Ž . 6IL-10, and IL-10q IL-2 tumors were generated by implanting to caudaterputamen of Fisher 344 rats mixtures of 10 glioma 9L cellsq2=10

Ž . Ž . Ž .endothelial RBEZ cells as follows: Control: control-transfected glioma cells 9L–neo qcontrol-transfected RBEZ cells RBEZ–hygro ; IL-2: 9L–neocellsqRBEZ–IL-2 cells; IL-10: 9L–IL-10 cellsqRBEZ–hygro cells; IL-10q IL-2: 9L–IL-10 cellsqRBEZ–IL-2 cells. Animals were sacrificed 6 and

Ž .14 days post-implantation ns6 for each tumor type at each time point , brains were sectioned, and sections were stained with hematoxylin and eosin.Representative photomicrographs depict maximum tumor cross-sections. At both time points, IL-10q IL-2 tumors were smaller than any of the othergenetically modified tumors. Bars1 mm.


into syngeneic Fisher 344 rats. Tumors developed in allanimals, regardless of whether they received IL-10-produc-ing cells or control-transfected cells. 9L–IL-10 tumorsgrew at a significantly slower rate than control tumorsŽ . Ž .p-0.05 Fig. 1A . By 34 days post-implantation, thevolumes of 9L–IL-10 tumors averaged approximately 25%of control tumors.

To determine the effect of transgenic IL-10 on intracra-Ž . 5 Žnial i.c. tumor growth, 10 9L–IL-10 or 9L–neo con-

.trol cells were implanted into the caudaterputamen ofFisher 344 rats. Animals were sacrificed on post-implanta-tion day 11 and the sizes of intracranial tumors werequantified. Similar to our results with s.c. tumors, the i.c.9L–IL-10 tumors were found to be significantly smaller

Ž . Ž .than control tumors p-0.05 Fig. 1B . These anti-tumoreffects of IL-10 gene transfer required an intact host

Ž .immune system see below and were, therefore, not sec-ondary to autocrine effects of gene transfer on the inherentmalignant characteristics of the tumor cell lines.

3.2. IL-10 and IL-2 cooperate in ÕiÕo to produce signifi-cantly greater tumor inhibition

We have previously shown that nontumorigenic ratbrain endothelial cells engineered to express the murine

Ž .IL-2 gene RBEZ–IL-2 cells can elicit a moderate hostanti-tumor effect when co-inoculated with 9L glioma cellsŽ .Nam et al., 1996 . Considering recent reports showing

Žcertain synergistic actions of IL-2 and IL-10 in vitro Chen.and Zlotnik, 1991; Fluckiger et al., 1993; Itoh et al., 1994 ,

we asked if endothelial cell-based IL-2 gene delivery couldaugment IL-10-mediated tumor inhibition. In these experi-ments, the number of IL-2-secreting endothelial cells was

Žreduced from that of our prior report 10-fold reduction.relative to the number of 9L glioma cells to minimize the

anti-glioma effects of IL-2 and thereby increase the likeli-hood of identifying either additive or synergistic actions ofIL-2 combined with IL-10 gene transfer.

IL-10q IL-2 tumors were generated by implanting tocaudaterputamen a mixture of glioma cells that produce

Ž .IL-10 9L–IL-10 and nontumorigenic endothelial cellsŽ .that produce IL-2 RBEZ–IL-2 . Control tumors were

generated by using a mixture of control-transfected gliomaŽ .cells 9L–neo and control-transfected RBEZ cells

Ž .RBEZ–hygro . We also generated tumors expressing IL-2Ž .alone RBEZ–IL-2 cellsqcontrol 9L–neo cells and tu-

Žmors expressing IL-10 alone 9L–IL-10 cellsqcontrol. 5RBEZ–hygro cells . In each case, we implanted 10 glioma

cells and 2=106 endothelial cells. Animals were sacri-ficed 6 or 14 days post-implantation and brains wereevaluated for tumor growth. Histologic examination of

Ž .hematoxylin and eosin H and E -stained brain sectionsrevealed that, at both time points, IL-10q IL-2 tumorswere substantially smaller in volume than control tumorsand were considerably smaller in volume than tumors

Ž .expressing IL-10 or IL-2 alone Fig. 2 .

Quantitation of intracranial tumor volumes showed thatIL-10q IL-2 tumors were approximately 88% smaller than

Ž .control tumors at 6 days post-implantation p-0.001Ž .and were over 99% smaller than controls p-0.001 at

Ž .14 days post-implantation Fig. 3 . In comparison, IL-2tumors were approximately 20% smaller than control tu-mors at 6 days and 30% smaller than controls at 14 daysŽ .Fig. 3 . IL-10 tumors were about 50% smaller than con-trol tumors at 6 days and 70% smaller than controls at 14

Ž .days Fig. 3 . Interestingly, IL-2 tumors and IL-10 tumorsŽ .grew significantly in size from 6 days to 14 days p-0.01

while the IL-10q IL-2 tumors appeared to decrease inŽ . Ž .volume during that period ps0.07 Fig. 3 .

The cooperative action of IL-10 and IL-2 was alsoreflected in the survival of animals bearing intracranialtumors. Fisher 344 rats were implanted intrastriatally withthe same combinations of glioma and endothelial cells asdescribed above. Although delivery to tumors of IL-10 or

Ž .IL-2 alone resulted in inhibited tumor size Fig. 3 , thesurvival of animals bearing tumors engineered to secreteeither cytokine alone was not increased in comparison to

Ž . Ž .animals bearing control tumors p)0.05 Fig. 4 . Incontrast, simultaneous intratumoral expression of IL-10

Ž .and IL-2 significantly prolonged animal survival p-0.01Ž .Fig. 4 . Regardless of the treatment group, however,post-mortem evaluation revealed that all animals ulti-mately died with a substantial intracranial tumor burden.

Fig. 3. Quantitative analysis of the effects of IL-10qIL-2 gene deliveryon intracranial 9L tumor growth. Control, IL-2, IL-10, and IL-10qIL-2tumors were generated by implanting to caudaterputamen of Fisher 344

5 Ž . 6 Ž .rats mixtures of 10 glioma 9L cellsq2=10 endothelial RBEZ cellsŽ .as follows: Control: control-transfected glioma cells 9L–neo qcontrol-

Ž .transfected RBEZ cells RBEZ–hygro ; IL-2: 9L–neo cellsqRBEZ–IL-2cells; IL-10: 9L–IL-10 cellsqRBEZ–hygro cells; IL-10qIL-2: 9L–IL-10cellsqRBEZ–IL-2 cells. Animals were sacrificed 6 or 14 days laterŽ .ns6 for each tumor type at each time point and tumor volumes werecalculated as described in Section 2. At both time points, all tumorsexpressing cytokines were reduced in size compared to control tumors.IL-10qIL-2 tumors were reduced the greatest and averaged less than 1%

Žof the size of control tumors at 14 days post-implantation ) p-0.001compared with control tumors at same post-implantation time, Bonfer-

.onirDunn . Data represents meansqS.E.M.


Fig. 4. Effects of IL-10 and IL-2 gene transfer on the survival of animalsbearing intracranial 9L tumors. Control, IL-2, IL-10, and IL-10qIL-2tumors were generated by implanting to caudaterputamen of Fisher 344

Ž . 5 Ž .rats ns8 for each tumor type mixtures of 10 glioma 9L cellsq2=6 Ž .10 endothelial RBEZ cells as follows: Control: control-transfected

Ž . Ž .glioma cells 9L–neo qcontrol-transfected RBEZ cells RBEZ–hygro ;IL-2: 9L–neo cellsqRBEZ–IL-2 cells; IL-10: 9L–IL-10 cellsqRBEZ–hygro cells; IL-10qIL-2: 9L–IL-10 cellsqRBEZ–IL-2 cells. The time

Žof death of each animal was recorded. Single cytokine delivery IL-10 or.IL-2 had no significant effect on animal survival, whereas the combina-

tion of IL-10qIL-2 gene transfer significantly prolonged animal survivalŽ .p-0.01, Mann–Whitney U test .

3.3. Anti-tumor effects of IL-10 and combination IL-10qIL-2 gene transfer are immune-dependent

To determine if the anti-tumor responses to IL-10 andcombination IL-10 and IL-2 are immune-dependent, we

Ž .compared the growth of 9L–IL-10 and 9L–neo controlcells after s.c. and i.c. implantation in T-cell-, B-cell-, and

ŽNK cell-deficient SCID-Beige mice C.B-17rIcrCr1-.SCIDrBeige . As observed in rat hosts, tumors developed

in all animals that received inoculations of either 9L–IL-10or control-transfected tumor cells. Subcutaneous 9L–IL-10and control tumors grew at essentially identical rates in

Ž . Ž .these animals p)0.05 Fig. 5A . In accordance with thes.c. tumors, i.c. 9L–IL-10 and control tumors did not

Žsignificantly differ in size at the time of sacrifice 14 days. Ž . Ž .post-implantation p)0.05 Fig. 5B .

To determine if the anti-tumor activity of combinedIL-10q IL-2 gene delivery is also dependent on the hostimmune response, we compared the i.c. growth of IL-10q

Ž .IL-2 tumors and control 9L–neoqRBEZ–hygro tumorsin SCID-Beige mice. At post-implantation day 14, animalswere sacrificed and tumor sizes quantified. No significantdifferences in tumor volumes were observed between IL-

Ž . Ž .10q IL-2 tumors and control tumors p)0.05 Fig. 5C .The host T-cell response to control and cytokine-pro-

ducing intracranial tumors were also examined. Brain sec-tions from animals sacrificed at post-implantation days 6and 14 were immunocytochemically stained with the anti-body MRC OX-52, which is directed against a pan T-cell

Ž .surface antigen Robinson et al., 1986 . This revealed arelatively low density of T-cells within control tumors and

Ž .few stained cells in the peritumoral striatum Fig. 6 . Incontrast, T-cells were markedly increased in density withintumors and their surrounding striatum in animals im-

Ž .planted with IL-10q IL-2 tumors Fig. 6 . The increase ininfiltrating intratumoral and peritumoral T-cells seen in

Ž .Fig. 5. The anti-tumor actions of IL-10 and combined IL-10q IL-2 gene delivery are immune-dependent. A Systemic IL-10 and control tumors were5 Ž . Ž .generated by subcutaneously injecting 10 9L–IL-10 or 9L–neo control cells, respectively, into the flanks of SCID-Beige mice ns9 for each group .

Ž . Ž .Tumor growth was monitored biweekly using tissue calipers. IL-10 tumors grew at similar rates to control tumors p)0.05, Fisher’s PLSD . B5 Ž .Intracranial IL-10 and control tumors were generated by implanting 10 9L–IL-10 or 9L–neo control cells, respectively, to the caudaterputamen of

Ž .SCID-Beige mice ns6 for each group . Animals were sacrificed 14 days after tumor cell implantation and tumor volumes were calculated as describedŽ . Ž .in Section 2. IL-10 tumors were not significantly different in size than control tumors p)0.05, Student’s t-test . C Intracranial IL-10q IL-2 and

5 Ž . 6 Ž .control tumors were generated by implanting to caudaterputamen of SCID-Beige mice mixtures of 10 glioma 9L cellsq2=10 endothelial RBEZŽ .cells as follows: IL-10q IL-2: 9L–IL-10 cellsqRBEZ–IL-2 cells; control: control-transfected glioma cells 9L–neo qcontrol-transfected RBEZ cells

Ž . Ž .RBEZ–hygro . Animals were sacrificed 14 days later and tumor volumes were calculated as described in Section 2. IL-10q IL-2 tumors ns4 did notŽ . Ž .differ in size from control tumors ns5 p)0.05, Student’s t-test . Data represents mean tumor volume"S.E.M.


Fig. 6. T-cell infiltration is enhanced by combined IL-10q IL-2 gene transfer. Control, IL-2, IL-10, and IL-10q IL-2 tumors were generated by implantingŽ . 5 Ž . 6 Ž .to caudaterputamen of Fisher 344 rats ns6 for each tumor type mixtures of 10 glioma 9L cellsq2=10 endothelial RBEZ cells as follows:Ž . Ž .Control: control-transfected glioma cells 9L–neo qcontrol-transfected RBEZ cells RBEZ–hygro ; IL-2: 9L–neo cellsqRBEZ–IL-2 cells; IL-10:

9L–IL-10 cellsqRBEZ–hygro cells; IL-10q IL-2: 9L-IL-10 cellsqRBEZ–IL-2 cells. Animals were sacrificed 6 days post-implantation and brainŽ .sections were stained immunocytochemically to detect infiltrating T-cells using primary antibody MRC OX-52 as described in Section 2. In each case,

tumors are indicated by asterisks and a representative T-cell is identified with an arrow. Infiltrating T-cells were substantially increased within the striatumsurrounding IL-10q IL-2 tumors compared to control, IL-2, and IL-10 tumors. Bars500 mm.

brains bearing IL-10q IL-2 tumors was also substantiallymore pronounced than that seen in animals bearing either

Ž .IL-10 tumors or IL-2 tumors Fig. 6 . There was noevidence for T-cell infiltration in hemispheres contralateralto tumor cell implantation in any of the control or IL-10qIL-2 tumor-bearing animals.

4. Discussion

IL-10 has recently been found to stimulate host antitu-mor activity when secreted at high levels by gene-en-gineered tumor cells in a number of systemic tumor mod-els. To date, IL-10 gene transfer to tumor cells has beenshown to inhibit the growth in vivo of solitary epithelial

Žand mesenchymal tumors Richter et al., 1993; Giovarelliet al., 1995; Kundu et al., 1996; Zheng et al., 1996;

.Berman et al., 1996; Huang et al., 1997 . In addition,gene-delivered IL-10 inhibits the metastasis of mouse

Žmammary adenocarcinoma cells to lung Kundu et al.,.1996 and the metastasis of mouse and human melanoma

Ž .Zheng et al., 1996; Huang et al., 1997 . T-cell-, NK cell-,andror neutrophil-dependent effector cell functions werestrongly implicated in these studies. The effects of IL-10expression on tumor growth and anti-tumor immune re-sponses within the CNS had not been previously explored.We show in this study that transgenic IL-10 production by9L gliosarcoma cells inhibits systemic and orthotopic in-tracranial 9L tumor growth through immune-mediatedmechanisms.

The antitumor actions of transgenic tumor cell-derivedIL-10 are interesting in light of the known immune-sup-pressive actions of this cytokine in other settings. Forexample, IL-10 has been shown to inhibit production ofthe pro-inflammatory cytokines IFNg and TNFa in a

Žvariety of systems Chen and Zlotnik, 1991; Vieira et al.,.1991; Hishii et al., 1995 and such effects would be

expected to inhibit anti-tumor immune responses. IL-10gene expression has been found within human gliomas


Ž .using reverse transcriptase PCR Huettner et al., 1995leading to a hypothesis that endogenous IL-10 functions tosuppress anti-tumor immunity within brain tumor patients.However, while levels of IL-10 observed in unmanipulatedwild-type tumors are typically low, the glioma cells engi-neered to express IL-10 in our study secrete transgenicIL-10 at rates 50-fold higher than control glioma cells andIL-10 immunoreactivity in plasma of animals bearing IL-10-secreting tumor cells was elevated 30-fold in compari-son to animals bearing control tumors. As is well-recog-nized with other cytokine-based tumor therapies, the highlevels of transgenic IL-10 expression achieved within tu-

Žmors following gene transfer relative to endogenous lev-.els are likely to be critical to the anti-tumor effects seen in

our experiments and those of other investigators usingŽ .systemic tumor models. Recently, Hagenbaugh et al. 1997

found more aggressive growth of subcutaneous Lewis lungcarcinomas in IL-10 transgenic mice constructed using amurine IL-10 transgene under the transcriptional control ofa human IL-2 promoter. The absence of an anti-tumorresponse in these transgenic hosts may be due to therelatively lower levels of intratumoral IL-10 achieved inthis model or to the cellular source of transgenic IL-10

Žproduction host lymphocytes in the transgenic model vs..tumor cells in our model .

IL-10 has immunostimulatory effects in certain contextsŽ .Blazar et al., 1995 and has been shown to have coopera-tive or synergistic effects with IL-2 in vitro in a number ofrecent reports. IL-10 and IL-2 synergize to induce theproliferation and differentiation of chronic B-cell lympho-

Ž .cytic leukemias Fluckiger et al., 1993 and to enhance thedifferentiation of Cowan I strain Staphylococcus aureus-

Žactivated B-cells into immunoglobulin-secreting cells Itoh.et al., 1994 . Culture of CD8q T-cells in the presence of

IL-10 plus IL-2 stimulates significantly higher prolifera-tion and cytolytic activity compared to cells cultured in

Ž .IL-2 alone Chen and Zlotnik, 1991; Groux et al., 1998 .In addition, both IL-2 and IL-10 stimulate NK cell activityŽSchwarz et al., 1994; Carson et al., 1995; Kundu and

.Fulton, 1997 and IL-10 alone rescues T-cells from apop-Ž .totic cell death Taga et al., 1993 . However, the capacity

of these cytokines to cooperate in anti-tumor responses invivo has not been previously explored. While combina-tions of other cytokines have been evaluated in tumor

Ž .models Heaton and Grimm, 1993 , this is the first reportevaluating the efficacy of this specific cytokine pair in anytumor model. Our findings show that combining IL-10 andIL-2 gene transfer results in a greater than additive andpossibly synergistic anti-tumor effect, by the criteria oftumor growth inhibition, animal survival, and tumorrperi-tumoral T-cell infiltration. This is the first evidence ofcooperative anti-tumor actions of these cytokines in vivoand within the CNS and is consistent with the data in vitroshowing enhanced proliferation and activation of cytolytic

ŽT-cells in the presence of both cytokines Chen and Zlot-.nik, 1991; Groux et al., 1998 .

The mechanism of action of IL-10 and combinationIL-10q IL-2 against intracranial 9L tumors was found tobe immune-mediated since the anti-tumor effect was lostin SCID-Beige mice that lack T, B, and NK cell functions.This conclusion is also supported by the enhanced infiltra-tion of T-cells within tumors secreting both IL-10 andIL-2. Our experiments do not distinguish between directimmune activation by IL-10 or an indirect effect secondaryto IL-10-mediated alterations in other immune-activating

Ž .or inhibitory factors Giovarelli et al., 1995 . While HuangŽ .et al. 1997 have suggested that IL-10 can function di-

rectly as an anti-angiogenic agent, the loss of all signifi-cant IL-10-mediated anti-glioma activity in immune-defi-cient hosts makes a direct anti-angiogenic mechanism un-likely in either the intracranial or subcutaneous 9L tumormodels. We can also conclude from the immune-deficientanimal experiments that tumor inhibition mediated by IL-10is not due to potential autocrine effects on the inherentmalignant characteristics of the tumor lines from the ma-nipulations required for gene transfer. Finally, the allo-geneic endothelial cells used for IL-2 delivery may play arole in the cooperative anti-tumor actions of combinationIL-10 and IL-2 in our experiments. However, the presenceof allogeneic endothelial cells in all experimental groupsincluding controls indicates that any important role foralloantigens must be IL-10 andror IL-2 dependent.

In conclusion, we show in this report that IL-10 genetransfer can inhibit glioma growth within brain throughimmune-mediated mechanisms. Importantly, we also showfor the first time that IL-10 and IL-2 can cooperate toenhance anti-tumor actions within brain. While the poten-tial toxic effects of IL-10 administration within the CNSremain relatively unknown, systemic administration in ratsand humans reveal little toxicity suggesting that IL-10 may

Žbe well-tolerated by patients with brain tumors Powrie et.al., 1993; Chernoff et al., 1995; Huhn et al., 1996 . Our

findings in this report justify additional studies to definethe mechanisms of IL-10 action against brain tumors andthe potential of IL-10 to augment other local or systemicbrain tumor therapies.

Acknowledgements

We thank Ms. Angela T. Williams for assistance inpreparing this manuscript. The immortalized brain en-dothelial cell line, RBEZ, was kindly provided by Pierre-

Ž .Olivier Couraud Neurotech, Paris . RBEZ cells and theirderivative lines are the property of Neurotech, Paris,France. J.L. owns stock in Neurotech, an arrangement thathas been reviewed and approved by the Johns HopkinsUniversity Office on Conflict of Interest. This work was

Ž .supported by NIH grants NS 33728 JL and CA 67173Ž .AMF .


References

Neuro-oncology, 1990. Primary malignant brain tumours. In: Thomas,Ž .D.G.T. Ed. , Neuro-oncology. Primary Malignant Brain Tumours,

Johns Hopkins Univ. Press, Baltimore, MD.Arosarena, O., Guerin, C., Brem, H., Laterra, J., 1994. Endothelial

differentiation in intracerebral and subcutaneous experimental gliomas.Brain Res. 640, 98–104.

Berman, R.M., Suzuki, T., Tahara, H., Robbins, P.D., Narula, S.K.,Lotze, M.T., 1996. Systemic administration of cellular IL-10 inducesan effective, specific and long-lived immune response against estab-lished tumors in mice. J. Immunol. 157, 231–238.

Blazar, B., Taylor, P.A., Smith, S., Vallera, D.A., 1995. Interleukin-10administration decreases survival in murine recipients of major histo-compatibility complex disparate donor bone marrow grafts. Blood 85,842.

Brem, S., Cotran, R., Folkman, J., 1972. Tumor angiogenesis: a quantita-tive method for histologic grading. J. Natl. Cancer Inst. 48, 347–356.

Burger, P.C., Vogel, F.S., Green, S.B., Strike, T.A., 1985. Glioblastomamultiforme and anaplastic astrocytoma. Cancer 56, 1106–1111.

Carson, W.E., Lindemann, M.J., Baiocchi, R., Linett, M., Tan, J.C.,Chou, C.-C., Narula, S., Caligiuri, M.A., 1995. The functional charac-terization of interleukin-10 receptor expression on human naturalkiller cells. Blood 85, 3577–3585.

Chen, W.-F., Zlotnik, 1991. IL-10: a novel cytotoxic T cell differentiationfactor. J. Immunol. 147, 528–534.

Chernoff, A.E., Granowitz, E.V., Shapiro, L., Vannier, E., Lonnemann,G., Angel, J.B., Kennedy, J.S., Rabson, A.R., Wolff, S.M., Dinarello,C.A., 1995. A randomized, controlled trial of IL-10 in humans. J.Immunol. 154, 5492–5499.

Fakhrai, H., Dorigo, O., Shawler, D.L., Lin, H., Mercola, D., Black,K.L., Royston, I., Sobol, R.E., 1996. Eradication of establishedintracranial rat gliomas by transforming growth factor b antisensegene therapy. Proc. Natl. Acad. Sci. U.S.A. 93, 2909–2914.

Fiorentino, D.F., Bond, M.W., Mosmann, T.R., 1989. Two types ofmouse T helper cell IV. Th2 clones secrete a factor that inhibitscytokine production by Th1 clones. J. Exp. Med. 170, 2081–2095.

Fluckiger, A.-C., Garrone, P., Durand, I., Galizzi, J.-P., Branchereau, J.,Ž .1993. Interleukin 10 IL-10 upregulates functional high affinity IL-2

receptors on normal and leukemic B lymphocytes. J. Exp. Med. 178,1473–1481.

Fujita, S., Puri, R.K., Yu, Z.-X., Travis, W.D., Ferrans, V.J., 1991. Anultrastructural study of in vivo interactions between lymphocytes andendothelial cells in the pathogenesis of the vascular leak syndromeinduced by interleukin-2. Cancer 68, 2169–2174.

Gansbacher, B., Zier, K., Daniels, B., Cronin, K., Bannerji, R., Gilboa,E., 1990. Interleukin 2 gene transfer into tumor cells abrogatestumorigenicity and induces protective immunity. J. Exp. Med. 172,1217–1224.

Giovarelli, M., Musiani, P., Modesti, A., Dellabona, P., Casorati, G.,Allione, A., Consalvo, M., Cavallo, F., di Pierro, F., De Giovanni, C.,Musso, T., Forni, G., 1995. Local release of IL-10 by transfectedmouse mammary adenocarcinoma cells does not suppress but en-hances antitumor reaction and elicits a strong cytotoxic lymphocyteand antibody-dependent immune memory. J. Immunol. 155, 3112–3123.

Glick, R.P., Lichtor, T., Kim, T.S., Ilangovan, S., Cohen, E.P., 1995.Fibroblasts genetically engineered to secrete cytokines suppress tumorgrowth and induce antitumor immunity to a murine glioma in vivo.Neurosurgery 36, 548–555.

Groux, H., Bigler, M.d.V.J.E., Roncarolo, M.-G., 1998. Inhibitory andstimulatory effects of IL-10 on human CD8q T cells. J. Immunol.160, 3188–3193.

Hagenbaugh, A., Sharma, S., Dubinett, S.M., Wei, S.H.Y., Aranda, R.,Cheroutre, H., Fowell, D.J., Binder, S., Tsao, B., Lockszley, R.M.,Moore, K.W., Kronenberg, M., 1997. Altered immune responses ininterleukin 10 transgenic mice. J. Exp. Med. 185, 2101–2110.

Hanisch, U.-K., Neuhaus, J., Quirion, R., Kettenmann, H., 1996. Neuro-toxicity induced by interleukin-2: involvement of infiltrating immunecells. Synapse 24, 104–114.

Heaton, K.M., Grimm, E.A., 1993. Cytokine combinations in im-munotherapy for solid tumors: a review. Cancer Immunol. Im-munother. 37, 213–219.

Hishii, M., Nitta, T., Ishida, H., Ebato, M., Kurosu, A., Yagita, H., Sato,K., Okumura, K., 1995. Human glioma-derived interleukin-10 inhibitsantitumor immune responses in vitro. Neurosurgery 37, 1160–1166.

Hsu, S.M., Raine, L., Fanger, H., 1981. The use of avidin–biotin–per-Ž .oxidase complex ABC in immunoperoxidase techniques: a compari-

Ž .son between ABC and unlabeled antibody PAP procedures. J.Histochem. Cytochem. 29, 577–580.

Huang, S., Zie, K., Bucana, C.D., Ulrich, S.E., Bar-Eli, M., 1997.Interleukin 10 suppresses tumor growth and metastasis of humanmelanoma cells: potential inhibition of angiogenesis. Clin. CancerRes. 2, 1969–1979.

Huettner, C., Paulus, W., Roggendoorf, W., 1995. Messenger RNAexpression of the immunosuppressive cytokine IL-10 in humangliomas. Am. J. Pathol. 146, 317–322.

Huhn, R.D., Radwanski, E., O’Connell, S.M., Sturgill, M.G., Clarke, L.,Cody, R.P., Affrime, M.B., Cutler, D.L., 1996. Pharmacokinetics andimmunomodulatory properties of intravenously administered recombi-nant human interleukin-10 in healthy volunteers. Blood 87, 699–705.

Itoh, K., Inoue, T., Ito, K., Hirohata, S., 1994. The interplay of inter-Ž . Ž .leukin-10 IL-10 and interleukin-2 IL-2 in humoral immune re-

sponses: IL-10 synergizes with IL-2 to enhance responses of human Blymphocytes in a mechanism which is different from upregulation ofCD25 expression. Cellular Immunology 157, 478–488.

Karasuyama, H., Melchers, F., 1988. Establishment of mouse cell lineswhich constitutively secrete large quantities of interleukin 2, 3, 4 or 5,using modified cDNA expression vectors. Eur. J. Immunol. 18,97–104.

Kundu, N., Fulton, A.M., 1997. Interleukin 10 inhibits tumor metastasis,downregulates MHC class I and enhances NK lysis. Cell Immunol.180, 55–61.

Kundu, N., Beaty, T.L., Hackson, M.J., Fulton, A.M., 1996. An-timetastatic and antitumor activities of interleukin 10 in a murinemodel of breast cancer. J. Natl. Cancer Inst. 88, 536–541.

Lal, B., Indurti, R.R., Couraud, P.-O., Goldstein, G.W., Laterra, J., 1994.Endothelial cell implantation and survival within experimentalgliomas. Proc. Natl. Acad Sci. U.S.A. 91, 9695–9699.

Levraud, J.-P., Duffour, M.-T., Cordier, L., Perricaudet, M., Haddada, H.,Kourilsky, P., 1997. IL-2 gene delivery within an established murinetumor causes its regression without proliferation of preexisting antitu-mor-specific CTL. J. Immunol. 158, 3335–3343.

De Waal-Malefyt, R., Abrams, J., Bennett, B., Figdor, C.G., de Vries,Ž .J.E., 1991. Interleukin 10 IL-10 inhibits cytokine synthesis by

human monocytes: an autoregulatory role of IL-10 produced bymonocytes. J. Exp. Med. 174, 1209–1220.

Nam, M., Johnston, P., Lal, B., Indurti, R., Wilson, M.A., Laterra, J.,1996. Endothelial cell-based anti-tumor gene delivery to brain. BrainRes. 731, 161–170.

Powrie, F., Menon, S., Coffman, R.L., 1993. Interleukin-4 and inter-leukin-10 synergize to inhibit cell-mediated immunity in vivo. Eur. J.Immunol. 23, 3043–3049.

Richter, G., Kruger-Krasagakes, S., Hein, G., Huls, C., Schmitt, E.,Diamantstein, T., Blankenstein, T., 1993. Interleukin 10 transfectedinto chinese hamster ovary cells prevents tumor growth andmacrophage infiltration. Cancer Res. 53, 4134–4137.

Robinson, A.P., Puklavec, M., Mason, D.W., 1986. MRC OX-52: a ratT-cell antigen. Immunology 57, 527–531.

Roth, C., Mir, L.M., Cressent, M., Quintin-Colonna, F., Ley, V., Fradelizi,D., Kourilsky, P., 1992. Inhibition of tumor growth by histoincompat-ible cells expressing interleukin-2. Intl. Immunol. 4, 1429–1436.

Roux, F., Durieu-Trautmann, O., Chaverot, N., Claire, M., Mailly, P.,Bourre, J.M., Strosberg, A.D., Couraud, P.-O., 1994. Regulation of


g-glutamyl transpeptidase and alkaline phosphatase activities in im-mortalized rat brain microvessel endothelial cells. J. Cell Physiol.159, 101–113.

Schmidek, H., Nielsen, S., Schiller, A., Messer, J., 1971. Morphologicalstudies of rat brain tumors induced by N-nitrosomethylurea. J. Neuro-surg. 34, 335–340.

Schwarz, M.A., Hamilton, L.D., Tardelli, L., Narula, S.K., Sullivan,L.M., 1994. Stimulation of cytolytic activity by interleukin-10. J.Immunother. 16, 95–104.

Taga, K., Cherney, B., Tosato, G., 1993. IL-10 inhibits apoptotic celldeath in human T cell starved of IL-2. Intl. Immunol. 5, 1599.

Tamargo, R., Leong, K., Brem, H., 1990. Growth Inhibition of the 9Lglioma using polymers to release heparin and cortisone acetate. J.Neuro-Oncol. 9, 131–138.

Tjuvajev, J., Gansbacher, B., Desai, R., Beattie, B., Kaplitt, M., Matei,C., Koutcher, J., Gilboa, E., Blasberg, R., 1995. RG-2 glioma growthattenuation and severe brain edema caused by local production ofinterleukin-2 and interferon-g. Cancer Res. 55, 1902–1910.

Trojan, J., Johnson, T.R., Rudin, S.D., Ilan, J., Tykocinski, M.L., Ilan, J.,1993. Treatment and prevention of rat glioblastoma by immunogenicC6 cells expressing antisense insulin-like growth factor I RNA.Science 259, 94–97.

Vieira, P., De Waal-Malefyt, R., Dang, M.-N., Johnson, K.E., Kastelein,R., Fiorentino, D.F., DeVries, J.E., Roncarolo, M.-G., Mosmann,T.R., Moore, K.W., 1991. Isolation and expression of human cytokinesynthesis inhibitory factor cDNA clones: homology to Epstein–Barrvirus open reading frameBCRFI. Proc. Natl. Acad Sci. U.S.A. 88,1172–1176.

Yu, J.S., Wei, M.X., Chiocca, E.A., Martuza, R.L., Tepper, R.I., 1993.Treatment of glioma by engineered interleukin 4-secreting cells.Cancer Res. 53, 3125–3128.

Zheng, L.M., Ojcius, D.M., Garaud, F., Roth, C., Maxwell, E., Li, Z.,Rong, H., Chen, J., Wang, X.Y., Catino, J.J., King, I., 1996. Inter-leukin-10 inhibits tumor metastasis through an NK cell-dependentmechanism. J. Exp. Med. 184, 579–584.

IL-10 gene transfer to intracranial 9L glioma: tumor inhibition and cooperation with IL-2

Documents