Regression of AK7 malignant mesothelioma established in ... - Nature

Cancer Gene Therapy (2003) 10, 481–490 All rights reserved 0929-1903/03 $25.00

r 2003 Nature Publishing Group

www.nature.com/cgt

Regression of AK7 malignant mesothelioma established in immunocompetent mice following intratumoral gene transfer of interferon gamma Laurence Cordier Kellerman,1 Laurence Valeyrie,1 Nadine Fernandez,1 Paule Opolon,1 Jean-Christophe Sabourin,1 Eve Maubec,1 Pierre Le Roy,2 Agnes Kane,3 Agnes Legrand,5 Mohamed Amine Abina,1,4 Vincent Descamps,1,5 and He´di Haddada1 1

Institut Gustave Roussy, 94805 Villejuif Cedex-France; 2Transgene, Strasbourg, France; 3Department of Pathology and Laboratory Medicine, Brown University, Providence, Rhode Island 02912, USA; 4 Departement de microbiologie et d’immunologie, Universite´ de Montreal, Ont., Canada; and 5Hopital Bichat, Paris, France. Malignant mesothelioma (MM) is a lethal tumor linked with a prior exposure to asbestos in which limited progress has been made so far using conventional therapies. MM is an example of a ‘‘nonimmunogenic’’ tumor characterized by a fibrous stroma and an absence of infiltrating T lymphocytes. High levels of transforming growth factor-beta (TGF-b) produced by mesothelioma cells have been related to the immune tolerance towards the tumor. In order to evaluate the effect of local delivery of cytokines such as interferon gamma (IFN-g) by gene transfer, we characterized and used a murine model, AK7, which appeared very similar to human mesothelioma. AK7 cells expressed low levels of major histocompatibility class I and class II antigens and secreted high levels of latent TGF-b. The TGF-b pathway in AK7 cells is operative but inefficient because endogenous TGF-b is predominantly inactive. Treatment of pre-established AK7 tumors by direct intratumoral injection of an adenovirus vector expressing murine IFN-g, Ad.mIFN-g, led to significant tumor regression. Peripheral tumor infiltration by CD4+ and CD8+ T lymphocytes in the treated tumors appeared to be because of the induction of an immune response. Tumor relapse was observed, which could be due to local TGF-b secretion by remaining tumor cells. Cancer Gene Therapy (2003) 10, 481–490. doi:10.1038/sj.cgt.7700594 Keywords: malignant mesothelioma; TGF-b; Gene transfer; recombinant adenovirus; IFN-c; apoptosis; tumor regression

alignant mesothelioma (MM) is an aggressive cancer of the mesothelium most commonly occurM ring in the pleural cavity and often associated with exposure to asbestos.1–3 Although rare, this aggressive tumor is responsible for 3000 deaths per year in the US (similar to Hodgkin’s disease). As a result of the widespread use of asbestos in the past and the long latency period characteristic of this tumor, deaths due to MM are expected to increase at least until 2020.4 MM is an example of ‘‘nonimmunogenic’’ tumor characterized by a fibrous and angiogenic stroma and the lack of tumor-infiltrating T lymphocytes. Transforming growth factor-beta (TGF-b) seems to be the pivotal cytokine in the tumor progression. High levels of TGF-b are produced by mesothelioma cells which contributes to the generation of this stroma and the development of a state of tolerance towards the tumor. Other additional angiogenic cytokines such as vascular endothelium Received November 4, 2002.

Address correspondence and reprint requests to: Dr He´di Haddada, VRC/NIH, Blg 40, Bethesda, MD, 20892, USA. E-mail: mhaddada50 @yahoo.fr

growth factor (VEGF), fibroblast growth factor-1 (FGF-1) or interleukin 6 also participate in the genesis of the stroma.5 The failure of conventional therapies to combat MM progression6,7 has prompted investigation into new antitumoral strategies, including local delivery of cytokines to inhibit angiogenesis and to overcome this immune tolerance. Interferon gamma (IFN-g) represents an attractive candidate because of its immunomodulatory functions and its ability eradicate this nonimmunogenic tumor. IFN-g has opposite effects to TGF-b as recently demonstrated by the induction of Smad7 expression, which prevents the interaction of Smad3 with the TGF-b receptor.8 IFN-g therefore inhibits the TGF-b activation of Smad3 and may prevent the immunosuppressive effect of TGF-b. Furthermore, IFN-g directly stimulates host immune response by inducing the expression of MHC class I and class II molecules. Additionally, the antitumoral response seen with interleukin-12 (IL-12) may be related to IFN-g as IL-12 is the main inducer of IFN-g.9 In this study, a new mouse mesothelioma cell line, AK7, was established and used to evaluate local delivery of IFN-g in MM. This sarcomatous type of tumor is

Intratumoral interferon gamma gene therapy of marine malignant mesothelioma L Cordier et al

482

highly suggestive of the most aggressive variant of human mesothelioma, which is refractory to all forms of therapy.10 Pathological examination demonstrated that this model is similar to severe types of human MM. Moreover, AK7 cells expressed low levels of MHC class I and class II antigens and secreted high levels of IL-6 and TGF-b. We studied the effect of IFN-g gene transfer into AK7 cells in vitro and on tumor progression in vivo by intratumoral delivery in mice using an adenoviral system. Our data indicate that local delivery of IFN-g induces tumor regression and no apparent tumor was observed for 50 days. The antitumoral activity of local delivery of IFN-g may be mediated through inhibition of TGF-b functions. Thus, local cytokine delivery could provide a useful approach to target agressive tumors.

371C in a minimum volume of medium at a multiplicity of infection of 100–300 PFU (plaque forming unit) per cell. b-galactosidase activity was determined 48 hours following infection with Ad.bgal using standard histochemical methods.

Cytokine assays AK7 cells were infected with Ad.mIFN-g. or Ad.bgal. After 48 hours, the medium was removed, centrifuged, and diluted. Cytokine secretion by uninfected and Ad.mIFN-g-infected AK7 cells was assayed by ELISA using the following kits: IFN-g (R & D Systems, England), and murine IL-6 (Endogen, France). The active and total amount of TGF-b1 was determined by ELISA according to the manufacturer’s instructions (Promega, USA and R&D Systems, England).

Materials and methods

Fluorescence-activated cell sorting (FACS) analysis

Mice and cell lines

At 72 hours after infection with Ad.mIFN-g (MOI 75), AK7 cells were harvested with 2 mM EDTA in phosphate-buffered saline (PBS) and washed in PBS containing 1% bovine serum albumin (BSA) and 0.1% sodium azide. For detection of MHC class I and class II molecules, the following antibodies were used: FITCconjugated anti-mouse H-2Kb (CTKb, mouse IgG2a, Caltag, San Francisco, CA), FITC-conjugated anti-mouse I-Ab (25-5-16S, mouse IgGM, Caltag, San Francisco, CA), and FITC-conjugated isotype-matched. Cells infected with Ad.bgal were used as controls.

Female C57BL/6 and C57BL/6 nu/nu, purchased from the ‘‘Centre d’e´levage Janvier’’ (le Genest-St-Ile, France), were kept in the animal facilities of the Institut Gustave Roussy. Mice were 8 to 9-week old at the time of tumor inoculation. A murine mesothelioma cell line (AK7) was established following peritoneal lavages of a C57BL/6 mouse after 41 weekly intraperitoneal injections of crocidolite asbestos fibers. The cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 1 mM MEM sodium pyruvate solution, penicillin (75 U/ml), and streptomycin (10 mg/ml) at 371C in a humidified atmosphere of 5% CO2. HEK 293 cells were used to propagate the recombinant adenovirus vectors.11,12

Recombinant adenovirus vectors A replication-deficient adenovirus vector containing the b-galactosidase gene (Ad.bgal). was used as a control to verify in vitro and in vivo gene transfer into AK7 cells. Two replication-deficient recombinant adenoviral vectors Ad.mIFN-g were constructed by the insertion of murine IFN-g encoding cDNA under the control of the RSV.LTR into the pRSVbgal or pCMVbgal expression plasmid as previously described.13 For the Ad.TGF-b vector, the TGF-b1 cDNA was placed under the control of the CMV promoter.14 An adenovirus expressing wildtype p53 (Ad.p53) was used as a control for caspase 3 cleavage analysis. Viral stocks were prepared by infection of the HEK 293 cell line. After 36–40 hours, viral particules were harvested, concentrated by double cesium chloride gradient purifications,11 dialyzed, and stored in 10% glycerol at 701C. The titers of the stocks were determined by plaque assay using HEK 293 cells.12

Gene transfer Murine mesothelioma cells (AK7) were cultured for 48 hours, then infected with rAd vectors for 2 hours at

Cancer Gene Therapy

Intravenous delivery of recombinant adenovirus vectors Five C57BL/6 were injected in the orbital vein with 109 PFU of either Ad.bgal or Ad.mIFN-g in a volume of 200 ml. At specified times, blood was drawn from the retro-orbital sinus of mice and sera were stored at 201C until the quantification of the mIFN-g level by ELISA was performed.

Tumor studies Murine AK7 cells were harvested by trypsinization, washed twice in PBS, and viable cells were counted using the trypan blue dye exclusion test. Cells were inoculated subcutaneously in C57BL/6 and nude mice in a 0.1 ml volume using 1-ml syringe. The minimal tumorigenic dose (MiTD) of tumor cells had been previously determined (2 106 cells). The mice were scored for tumor growth, twice a week, and tumor size was documented by measuring two perpendicular diameters in millimeters using a caliper. The tumor volume was calculated from the average diameter assuming that a spheroid had formed. To determine AK7 immunogenicity, 10 MiTD irradiated AK7 cells were injected three times into C57BL/6 mice, once a week. After 15 days, mice were challenged with 1 MiTD live tumor cells. Ad.bgal or Ad.mIFN-g viral stocks were diluted in PBS. 109 PFU in 100 ml of PBS were injected into day 15 pre-established AK7 tumors. Tumor size was measured twice a week as described above.


RNA extraction and cDNA synthesis Total RNA was isolated according to the method described by Gerwin et al.15 cDNA was synthetized from 10 mg of RNA, using a (dT)17 primer, AMV reverse transcriptase (Boerhinger Mannheim, Mannheim, Germany) and its provided buffer, with addition of RNAsin (Promega, Madison, WI).

Oligonucleotides PCR. primer specific for the sequence of the adenovirus RVS-LTR promoter was: RSV2-LTR-50 , 0 0 5 CATTCACCACATTGGTGTGC3 ; PCR primer specific for the sequence of mIFN-g was: INF-g 2-30 , 50 CCACATCTATGCCACTTGAG30 .

PCR PCR analysis was performed using a 9600 GeneAmpCyclers (Perkin-Elmer, Foster City, CA) with the following mixture: 1 U of Taq polymerase (Eurogentec, Seraing, Belgium), with the provided buffer; 2.5 mM MgCl2; 0.2 mM dNTP; 0.5 mM of each primer; and template (the equivalent of 10–100 ng of reverse-transcribed RNA) in a final volume of 30 ml. After an initial step of 2 minutes at 941C, reaction consisted of 40 cycles of 25 seconds at 941C, 25 seconds at 601C, and 30 seconds at 721C, followed by a final step of 4 minutes at 721C.

Histopathology Ad.bgal- or Ad.mIFN-g-treated tumors were dissected, fixed in Bouin’s reagent, embedded in paraffin, sectioned, and stained with hematoxylin–eosin–saffron. The sections were examined independently by two pathologists who confirmed the histopathologic diagnosis.

Immunohistochemistry The following monoclonal antibodies were obtained from PharMingen: rat anti-mouse CD4 (L3T4/IgG2a), rat antimouse CD8a (Ly-2/IgG2a) all biotinylated, and a purified rat anti-mouse Mac-1 (IgA). For the biotinylated primary antibodies, a streptavidin–horseradish peroxidase (Vector Laboratories) was used; for Mac-1, the secondary antibody was a rabbit anti-mouse Ig coupled with alkaline phosphatase (DAKO). Serial cryosections (6 mm) were collected on Super Frost/plus slides (Menzel-Glaser), fixed in 0.02% H2O2 in acetone, and immunolabeled in a double-layer protocol as described above. All incubations were done at room temperature: 1 hour with the primary antibody and 30 minutes with each of the conjugates followed by 5 minutes washes in Tris-buffered saline, then Tris-buffered saline plus 0.3% BSA. Slides were revealed by incubation in 3-amino-9-ethylcarbazol (AEC, Sigma Immunochemicals) for 6 minutes at room temperature, counterstained with Mayer’s hematoxylin, and mounted in Aquatex (Merck). For Mac-1 staining, the revelation included mouse APAAP (1/25) (DAKO) and fast red (DAKO) supplemented with 0.2 mM levamisole (Levamisole, DAKO).

Slides were examined independently by two pathologists and the number of immunostained aggregates was counted taking into account 10 fields per section.

Western blot analysis Subconfluent cells were transduced with the adenoviral vectors at a MOI of 100 PFU/cell. At 27 hours after transduction, the cells were washed with ice-cold PBS, then harvested by trypsinization. After washing two times, lysis was performed at 41C for 10 minutes in a buffer containing Tris PH 8 1 M, NaCl 120 mM, NP40 0.5%. Western blot analysis was performed according to standard procedures using a primary antibody directed against either phosphorylated Smad2 or Smad7 (kindly provided by Professor CH Heldin, Ludwig Institute for Cancer Research, Uppsala, Sweden) , caspase 3 or b-actin (Sigma).

Statistical analysis Mean tumor volumes between treated groups were compared utilizing the Student’s t-test. Results

Characterization of the AK7 mouse mesothelioma cell line The AK7 mouse mesothelioma cell line has biological characteristics similar to human aggressive nonimmunogenic mesothelioma. AK7 cells expressed low levels of class I (H-2Kb) and class II (H-2I-Ab) MHC molecules, while the B7.1 costimulatory molecule was not expressed constitutively (data not shown). Moreover, this cell line released high amounts of IL-6 (850 pg/106 cells/48 h) and TGF-b1 (1076 pg/106 cells/48 h) in vitro. Active TGF-b1 was estimated to be 4% of the total TGF-b1. Since AK7 cells mainly produced inactive latent TGFb1, we tested the growth response of AK7 cells to active TGF-b1, which is known to be a potent inhibitor of tumor growth. This experiment was performed under specific in vitro conditions plating the cells at a low cell density (102 cells/well). Inhibition of the growth of AK7 cells by exogenous active TGF-b1 was only observed under these conditions. This result may suggest the absence of anomaly of the TGF-b1 transducing pathway in these tumor cells. At a higher density of cells (104 cells/ well), exogenous TGF-b1 had no effect on AK7 proliferation. The supernatant of AK7 cells as well as recombinant TGF-b1 inhibited lymphocytes stimulation by PHA (data not shown). At 10 days after subcutaneous injection of 1 MiTD containing 2 106AK7 cells, animals developed homogeneous tumors that extended to the adjacent tissues. Tumors were hard, white in color with a fibrous stroma. Histologically, the tumors exhibited a sarcomatoid pattern, similar to the most aggressive forms of human mesothelioma. Injection of irradiated AK7 cells at a total dose of 10 MiTD did not provide any protection against subsequent challenge.

Cancer Gene Therapy

483


484

Efficient adenovirus-mediated gene transfer into AK7 mesothelioma tumor cells Efficiency of mesothelioma cell infection with adenovirus vectors was evaluated by in vitro Ad.bgal infection of AK7 cells and revealed by nuclear staining. The results indicate that 70 and 95% of cells expressed the transgene upon infection with 75 and 150 PFU/cell, respectively (data not shown). The viability of the infected cells was around 90% as compared to uninfected cells even when higher doses of virus (up to 200 PFU/cell) were used. The level of IFN-g secretion by Ad.mIFN-g-infected AK7 cells was determined by ELISA and was dependent on the dose of the virus used. As can be seen in Figure 1a, an amount of 8.7 ng of IFN-g was detected during a period of 6 hours following 48 hours postinfection with 100 PFU and IFN-g rose to 13 ng following infection with 200 PFU. No IFN-g was detected in the supernatants of uninfected or Ad.bgal-infected cells (Fig 1a).

Evidence of in vivo mIFN-g secretion after Ad.mIFN-g injection C57BL/6 mice were injected in the orbital vein with 2 109 PFU of Ad.mIFN-g; then, sera were collected at different times after injection and analyzed by ELISA. The level of IFN-g was on average 1.8 ng/ml 3 days postinjection, reached 3.5 ng/ml at day 6, then decreased to 1.5 ng/ml at day 14 (Fig 1b). Two out of three animals intravenously injected with Ad.mIFN-g died at day 7, while control Ad.bgal-injected animals showed no decrease in viability. However, intratumoral injection of Ad.mIFN-g induced neither mortality nor detectable IFN-g in the sera of the animals. Using a nonquantitative PCR assay, IFN-g gene expression in Ad.mIFN-g-treated tumors was evaluated by measuring IFN-g mRNA levels. The data confirmed that transgenic murine IFN-g mRNA was detectable in tumors treated with Ad.mIFN-g, but not in those treated with PBS or Ad.bgal (Fig 1c).

Upregulation of cell surface MHC molecules following infection of AK7 cells with Ad.mIFN-g As AK7 cells express low levels of MHC class I and class II antigens, we assessed whether IFN-g gene transfer increased the expression of these molecules on the cell surface. Therefore, MHC antigen expression following treatment with Ad.mIFN-g or Ad.bgal (as a control) were analyzed 72 hours postinfection. FACS analysis (Fig 2a) showed that following Ad.mIFN-g infection, the level of class I antigens expression was increased as well as the percentage of positive cells (44 versus 14% for control Ad.mIFN-g-injected tumors. Established (around day 15) AK7 tumors were treated with either PBS or, 109 PFU of Ad.bgal or Ad.mIFN-g. At 2 and 4 days after treatment, mice were killed, tumors excised, total RNA extracted, and cDNA synthesized. PCRs were then performed to test the expression of murine IFN-g mRNA (RSV-50 and mIFN-g 30 primers). Bands of specific size (220 pb), visualized in an ethidium bromide-stained agarose gel, are shown.

Cancer Gene Therapy

Figure 1 Efficient adenoviral IFN-g gene transfer and expression (a) in vitro after infection of AK7 cells, (b) in vivo following intravenous injection and (c) in established tumors after intratumoral administration of Ad.mIFN-g. (a) Level of IFN-g secreted by infected AK7 cells: 48 hours postinfection of 1 106 AK7 cells with Ad.mIFN-g at an MOI of 100, 200, and 300, the cells were washed and incubated in fresh medium for a novel period of 6 hours. Then, the supernatants were collected and the secreted IFN-g was quantified by ELISA. Supernatants from Ad.bgal-infected AK7 cells were used as controls. (b) Concentration of IFN-g in the serum of Ad.mIFN-g-injected C57BL/6: Three mice per group were inoculated in the orbital vein with either 109 PFU of Ad.mIFN-g or, as controls, 109 PFU of Ad.bgal. Sera were collected at day 3, 6, 9, and 14 postinjection and, presence of IFN-g was quantified by ELISA. (c) Expression of transgenic mIFN-g in


cells). However, MHC class II expression was only slightly enhanced (Fig 2b).

Ad.mIFN-g induced downregulation of Smad2 phosphorylation in AK7 cells In order to determine whether IFN-g could modulate TGFb signaling pathway, we studied the state of activation of Smad2 in Ad.mIFN-g-infected cells by Western blot analysis using a monoclonal antibody specific for the phosphorylated form of Smad2. As shown on Figure 3a, the expression of phosphorylated Smad2 was dramatically decreased in Ad.mIFN-ginfected AK7 cells as compared to that of uninfected cells or cells infected with Ad.TGFb which encodes the active form of TGFb. This result is in agreement with the Ad.mIFN-g inhibitory effect on TGFb activation in AK7/fibroblasts coculture (Maubec et al, manuscript in preparation). Smad7 was constitutively expressed in AK7 cells and was not affected after treatment (Fig 3b).

Ad.mIFN g-induced apoptosis of AK7 cells In order to determine whether Ad.mIFN-g had a direct effect on the viability of AK7 cells, we evaluated the cell cytotoxicity following infection, and found that 60% of the cells were dead 48 hours after infection with 100 PFU/cell, versus 10% with the control vector. Western blot analysis of cellular extracts from infected AK7 cells showed the induction of caspase 3 cleavage (17 kd) confirming apoptosis in the cells (Fig 4).

Intratumoral injection of Ad.mIFN-g-induced AK7 tumor regression in syngeneic immunocompetent mice C57BL/6 mice were injected subcutaneously with 2 106 AK7 cells and when the tumor volume reached 30–50 mm3, the animals were injected intratumorally with PBS (eight animals) or 109 PFU of either Ad.mIFN-g (12 animals), or Ad.bgal (eight animals) as a control. The data illustrated in Figure 5 showed important differences in tumor growth between control animals and

Figure 2 Membrane expression of MHC class I (a) and class II (b) on Ad.mIFN-g-infected cells. AK7 cells were infected with Ad.mIFN-g at an MOI of 75 and analyzed by flow cytometric 72 hours postinfection. Cells were stained with mAb to H-2Kb (CTKb) and I-Ab (25-5-16S). As a control, AK7 cells were infected with Ad.bgal.

Cancer Gene Therapy

485


486

Figure 5 Treatment of AK7 mesothelioma tumor-bearing C57BL/6 mice by intratumoral IFNg gene transfer. AK7 tumors (30–50 mm3) were injected with 109 PFU of either (a) Ad.mIFN-g or (b) of Ad.bgal as control. Tumor size was recorded twice a week.

Transfer of active TGFb1 gene instead of IFN-g does not inhibit the tumorigenicity of AK7 cells in mice

Figure 3 Western blot analysis of phosphorylated-Smad2 (a) and Smad7 (b) expression in AK7 cells following in vitro TGF-b or IFN-g gene transfer. At 24–27 hours post-adenoviral gene transfer of TGFb or INF-g, cellular extracts were prepared and equivalent amounts of proteins (30 mg) were loaded. Expression of phosphorylated Smad2 (P-Smad2) and Smad7 was analyzed using specific antibodies.

The paradoxical pivotal role of TGF-b1 in this tumor was evaluated by subcutaneous injection in C57BL/6 mice of AK7 cells transduced in vitro with Ad.TGF-b1 (100 PFU/ cell). Mesothelioma tumor was detected at day 6 in 80% of C57BL/6 mice (8/10) injected with Ad.TGF-b1transduced AK7 cells, whereas tumors were only present in 50% of the control mice (5/10) injected with parental or Ad.bgal-transduced AK7 cells (P ¼ .3). At day 18, 90– 100% mice in both groups possessed tumor.

Ad.mIFN-g transfer induces stabilization of preestablished mesothelioma AK7 tumors in nu/nu mice

Figure 4 Detection of caspase 3 cleavage by Western blot analysis following in vitro, IFN-g or p53 (as a positive control) gene transfer in AK7 cells. At 24–27 hours postinfection of AK7 cells with recombinant adenovirus (rAd), IFN gamma or p53 cellular extracts were prepared and equivalent amounts (30 mg) were loaded in each lane, probed with anti-caspase 3 and visualized with an ECL detection system. Cellular extract of uninfected AK7 cells was used as a negative control.

Ad.mIFN-g-treated animals During the first 10 days after the injection of Ad.mIFN-g, the tumors regressed while Ad.bgal-treated tumors stabilized before continuing to grow. The size of the tumor remained stable for at least 1 month post-Ad.mIFN-g treatment and three out of 12 mice had no palpable tumor (Student’s t-test, Po.001). However, 2 months after Ad.mIFN-g injection, a relapse in tumor-free mice was observed.

Cancer Gene Therapy

IFN-g has been shown to act by a number of mechanisms, including direct cytotoxicity, inhibition of angiogenesis, NK, and macrophage activation, as well as induction of T-cell-mediated specific immune response.16,17To determine which of these mechanisms were the most operative in our model, we tested the efficacy of Ad.mIFN-g in athymic nude mice. Therefore, C57BL/6 nu/nu mice were injected subcutaneously with AK7 cells. When the tumor volume was around 50 mm3, the animals were intratumorally injected with either 109 PFU of Ad.mIFN-g or, as a control, Ad.bgal or PBS. This representative experiment involved six animals per group. Until day 20 after tumor treatment, the volume of Ad.mIFN-g-treated tumors stabilized, while PBS- and Ad.bgal-treated tumors continued growing. This stabilization was effective but transient as the Ad.mIFN-g-treated tumors started to grow again after this delay (Fig 6). These data may indicate the implication of the T-cell immune response in the tumor regression induced by Ad.mIFN-g treatment in immunocompetent mice.

Histological and immunohistological examinations of Ad.mIFN-g-treated tumors Untreated AK7 tumors showed the following histological and macroscopic patterns: polylobulated nodules enclos-

Q5


487

Figure 6 Treatment of AK7 mesothelioma tumor-bearing nu/nu mice by intratumoral gene transfer of IFN-g. AK7 tumors (40–50 mm3) were injected with 109 PFU of Ad.mIFN-g. As controls, AK7 tumors were treated with PBS. Tumor size was recorded twice a week (one representative experiment out of two is illustrated in this figure).

ing large tumor cells with ill-defined cytoplasm, hyperchromatic and prominent nucleoli, and numerous mitotic features. Tumor cells were spindle shaped with a pseudosarcomatous pattern (Fig 7a) and exhibited densely packed nuclei with very small amounts of cytoplasm.10,14,18 No interstitial edematous areas were seen. This pattern is suggestive of a mechanical barrier. Even large nodules (3 cm diameter) rarely had necrotic centers. No calcifications were seen. The nodules were unencapsulated and tumor cells extended between striated muscle fibers and into other surrounding tissues. Inflammatory cells were absent or rarely observed within these tumors. Following Ad.mIFN-g treatment, many tumor cells looked dystrophic with irregular dense nuclei, occasionally associated with necrosis. An extensive fibrotic reaction was observed. Some of these aspects are observed within human tumors following chemotherapy or radiotherapy treatment. A moderate inflammatory infiltrate was distributed throughout the tumor and its fibrous stroma (Fig 7b), while at the periphery, a marked lymphocytic infiltrate was observed (Fig 7c). At 5 days post-Ad.mIFN-g injection, there was a diffuse increase of CD3+ cells at the periphery of the tumor compared to the Ad.bgal-treated tumors. Immunostaining for CD4+ (Figure 7f) and CD8+ cells (Fig 7d–h) also showed an increase in this lymphocytic subset at the periphery of the Ad.mIFN-g-treated tumors, whereas the bulk of the tumor was devoid of immune cells.

Figure 7 Histological and immunohistological analysis of Ad.mIFNg-treated tumors. (a–c) hematoxylin–eosin–saffron. Paraffin sections. (a) Untreated mesothelioma AK7 tumor. Numerous spindlecells disclosing mitoses. No necrosis, and little inflammation is present ( 200). (b, c) AK7 tumors 5 days postinjection of Ad.mIFNg. (b) Foci of necrosis and fibrosis with inflammatory cells and persistent tumoral cells ( 200). (c) Lymphocytic infiltration at the periphery of the tumor ( 100). (d–h) HPR. Immunohistochemistry on frozen sections of AK7 tumors 5 days postintratumoral injection of recombinant adenoviruses. (e–g) Ad.bgal-treated AK7 tumors ( 630). (d) Ad.mIFN-g-treated tumors ( 100). (f–h) Ad.mIFN-gtreated tumors ( 630). (e–f) CD4+ immunostaining. (d,g,h) CD8+ immunostaining. Ad.bgal-treated AK7 tumors show few T cells compared with Ad.mIFN-g-infected AK7 tumors, which contain numerous aggregates of CD4+ and CD8+ cells localized at the periphery of the tumor.

Mac-1 cells (macrophage) were present in control tumors treated 5 days before with either PBS or Ad.bgal. Nevertheless, their number increased in Ad.bgal-treated tumors examined 11 days after injection. In Ad.mIFN-gtreated tumors, the expression of Mac-1 was already highly visible at day 5.

Discussion

TGF-b1 regulates tumoral cell proliferation in a cellspecific manner. Whereas in normal cells, TGF-b1 acts as a tumor suppressor by inhibiting proliferation and by

Cancer Gene Therapy


488

inducing differentiation and apoptosis, in some tumors TGF-b1 stimulates proliferation and invasiveness. We characterized a new murine model of malignant mesothelioma similar to aggressive human mesothelioma, which illustrate the paradoxical response of some tumor cells to TGF-b1. Endogenous TGF-b1 does not directly play a role in tumor cell proliferation, but probably acts on the tumor environment. TGF-b1 is known to stimulate the production of the extracellular matrix and to increase the angiogenesis and invasiveness of the tumor cells. Moreover, TGF-b1 induces a state of tolerance toward the tumor. This peculiar link between endogenous TGF-b1 and tumor cells was already described in other tumor cell types such as prostatic and breast cells.19 It thus appears that these cells have devised a way to use the production of endogenous TGF-b to their advantage and to circumvent the growth inhibitory effect of this cytokine without becoming refractory to TGF-b1. The in vitro and in vivo TGF-b paradoxical effects on tumor cells have already been described in prostatic tumors.19 Murine AK7 mesothelioma cells are nonimmunogenic and unresponsive to immunotherapy. Contrary to other tumor models,13,20 IL-2 and costimulatory B7 gene transfer into pre-established AK7 tumors was ineffective in inducing tumor regression nor immune response stimulation. Moreover, the supernatant of AK7 cells inhibited the proliferation of lymphocytes treated with PHA or IL-2, similar to treatment with TGF-b1 (data not shown). Regarding the IFN-g multiple effects including immunomodulation, inhibition of tumor cell proliferation, tumor cells apoptosis, and TGF-b1 downregulation,8 we investigated the effects of IFN-g gene expression in this model. We showed that direct Ad.mIFN-g administration into pre-established AK7 tumors in C57BL/6 mice was associated with very significant tumor regression. The intratumoral injection of Ad.mIFN-g led to significant tumor regression in all treated mice (Fig 6) and three out of 12 mice had no palpable tumor. The tumor regression may be at least in part caused by a direct in vivo cytolytic effect of mIFN-g on AK7 cells. In vitro, IFN-g has growth inhibitory effects associated with a direct cytotoxicity on AK7 cells by induction of apoptosis as demonstrated by caspase 3 cleavage. Growth inhibition of human mesothelioma cell lines in the presence of recombinant human IFN-g has been previously reported. The initial stabilization of tumors observed in nude mice following Ad.IFN-g treatment could be the result of a direct cytotoxic effect on AK7 tumor cells and/or the activation of natural killer cells and macrophages. However, the inhibitory effect of IFN-g was more efficient in immunocompetent mice. An immune response may be implicated, as IFN-g is a potent inducer of MHC class I and class II expression.20–23 Indeed, our data showed that in vitro infection of AK7 cells with Ad.mIFN-g enhanced the expression of MHC antigens, more significantly class I antigens that are expressed at low levels by uninfected AK7 cells (Fig 2). Abundant CD4+ and CD8+ aggre-

Cancer Gene Therapy

gates, which were not detected in PBS- and Ad.bgaltreated tumors, were present in Ad.mIFN-g-treated tumors of immunocompetent mice (Fig 7f–h). These lymphocytes may be implicated in the regression observed in C57BL/6 mice bearing Ad.mIFN-g-treated AK7 tumors, as only stabilization without regression was noted in genetically athymic (nu/nu) mice bearing Ad.mIFN-g-treated AK7 tumors (Fig 6). However, contrary to recently published data using an adenovirus expressing IFNb to treat distinct murine mesothelioma models,24 the lymphocytic reaction observed in the present study was not sufficient to induce an efficient memory response since tumor relapse was observed. The localization of the immunocompetent cells to the periphery of the tumor probably affected their biological activities, which may be explained by the high cell density, thus constituting a physical barrier for lymphocyte dissemination. Alternatively, secretion of TGF-b by AK7 cells, which is in agreement with data obtained by others for human25 and murine26 cell lines, may also be implicated. TGF-b may favor tumorigenesis through a number of mechanisms such as blocking activation of macrophages via inhibition of TNF-a production, or by decreasing expression of MHC class II27 or suppressing IFN-g-stimulated NO production via nitric oxide synthase.28 Fitzpatrick et al.26 demonstrated that the suppression of TGF-b secretion using antisense constructs led to the enhancement of tumor infiltration by T lymphocytes. The IFN-g secreted by the transduced tumor cells may act on the TGF-b efficacy on the tumor environment. IFN-g was shown by others to inhibit TGF-b signaling by inducing the antagonist Smad7 and may thus inhibit TGF-b-induced angiogenesis.8 This indirect effect on stroma cells may contribute to tumor regression. In the present study, we found that Smad2 expression was downregulated in Ad-IFNg-infected cells (Fig 3a), but we were unable to detect Smad7 upregulation in AK7 infected cells. It is worth noting that high levels of Smad7 proteins were constitutively expressed in AK7 cells (Fig 3b). Similar to results obtained by other investigators using different mesothelioma cell lines15,29–33we also detected high amounts of TGF-b, VEGF, and IL-6 in the supernatant of AK7 cells, which may account for the inability to mobilize functional immune cells. In vitro infection of AK7 cells with the Ad-IFN-g vector down regulated VEGF expression and TGF-b activation (Maubec et al, manuscript in preparation). Other growth factors and cytokines such as TGFa,34 PDGF A and B,24,35 produced by several human and murine MM cell lines, may interfere with effector mechanisms and prevent complete recovery following treatment. In vitro, AK7 mesothelioma cells are efficiently transduced in a dose-dependant manner by recombinant adenovirus vectors. The successful transduction of mesothelioma cells may be because of the presence of the vitronectin receptor avb5 on cell membranes, which represents one of the two known adenovirus-cell receptors.36 This receptor has also been demonstrated to be


involved in the recognition and internalization of serumcoated crocidolite asbestos by rabbit mesothelial cells in vitro.37 Experiments using the intravenous injection of Ad.mIFN-g in C57BL/6 mice demonstrated secretion of IFNg in the serum. However, this cytokine synthesis was associated, just after the sixth day postintravenous injection, with the death of two out of three Ad.mIFNg-treated animals (corresponding with the highest amount of secreted IFN-g. However, intratumoral injection of Ad.mIFN-g did not affect the viability of the animals, whereas expression of mIFNg within the tumor lasted for at least 4 days (Fig 1c), while no IFN-g was detected in the serum. Thus, intratumoral injection of Ad.mIFN-g avoids unnecessary high concentrations of IFN-g within the serum, circumventing the cytotoxic effects observed with intravenous injection of Ad.mIFN-g. Adenovirus vectors harboring ‘‘suicide genes’’ were tested in other models of MM. Smythe et al.38 showed that transfer of the Herpes simplex thymidine kinase (HSVtk) gene renders human MM cells sensitive to ganciclovir in vitro and in vivo in SCID mice.36,39 However, in a syngeneic Fisher rat mesothelioma, treatment of small tumors by intrapleural injection of Ad.HSVtk led to a marked reduction in tumor weight without prolonging the survival of the tumor-bearing animals.40,41 The difficulties in obtaining a complete remission, regardless of the transgene encoded by the adenoviral vector, suggest that the use of combined procedures to treat mesothelioma may be more effective. To optimize immunotherapy, cytokine gene transfer may be associated with a cytotoxic strategy, thus allowing the destruction of cytokine-resistant cells or alternatively with administration of antagonists of immunosuppressive factors. The combination of gene therapy with conventional therapies such as surgery and/or chemotherapy should also be considered for treatment of MM.

Conclusion

The data presented in this report indicate that local IFN-g displays effective antitumoral activity and induces regression of a highly agressive malignant mesothelioma that resembles human mesothelioma cancers associated with poor clinical prognosis. The data also suggest that IFN-g may be operating through inhibition of TGF-b signaling as was recently observed following IL-12 treatment.

Acknowledgments

This work was supported by the CNRS, the Association pour la Recherche sur le Cancer (ARC), the Ligue Nationale Contre le Cancer, the Institut Gustave Roussy, the Assistance pubique-hopitaux de Paris (AP-HP), the Fondation pour la recherche me´dicale (FRM) (LV), and by a research grant (RO1 ES 03189) from the National

Institute of Environmental Health Sciences (AK). MAA had a MRC fellowship. We thank Patrice Ardouin and all the staff of the laboratory of animal experimentation (IGR), Jean-Pierre Levraud for providing oligonucleotides and his help for PCR analysis and Professor CH Heldin for providing the antibodies specific to Smad7 and phosphorylated Smad2. We also thank Dr J Dando and Professor Zaghouani for corrections of the manuscript. We are grateful to the Vector Core of the University Hospital of Nantes supported by the Association Franc¸aise contre les Myopathies (AFM) for providing the Ad.TGFb vector. LCK present address: cell genesys, Inc, South San Francisco, CA 94080, USA.

References 1. Wagner JC, Sleggs CA, Marchand P. Diffuse pleural mesothelioma and asbestos exposure in North Western Cape Province. Br J Ind Med. 1960;17:260–271. 2. Craighead JE, Mossman BT. The pathogenesis of asbestosassociated diseases. N Eng J Med. 1982;306:1446–1455. 3. Sterman DH, Kaiser LR, Albeida SM. Gene therapy for malignant pleural mesothelioma. Hematol Oncol Clin North Am. 1998;12:553–568. 4. Ramael M, Van Meerbeeck JV, Van Marck EV. Mesothelioma, currents insights. Cancer J. 1994;7:174–180. 5. Kumar-Singh S, Weyler J, Martin MJH, et al. Angiogenic cytokines in mesothelioma: a study of VEGF, FGF-1 and 2, and TGFb expression. J Pathol. 1999;189:72–78. 6. Ruffie PA. Pleural mesothelioma. Curr Opin Onc. 1991;3:328–334. 7. Peto J, Hodgson JT, Matthews FE, Jones JR. Continuing increase in mesothelioma mortality in britain. Lancet. 1995;345:535–539. 8. Ulloa L, Doody J, Massague J. Inhibition of transforming growth factor-b/SMAD signalling by the interferon-gamma/ STAT pathway. Nature. 1999;397:710–713. 9. Caminschi I, Venetsanakos E, Leong CC, et al. Cytokine gene therapy of mesothelioma. Immune and antitumoral effects of transfected interleukin-12. Am J Respir Cell Mol Biol. 1999;21:347–356. 10. Law MR, Hodson ME, Heard BE. Malignant mesothelioma of the pleura: relation between histological type and clinical behaviour. Thorax. 1982;37:810–815. 11. Graham FL, Van Der Eb EJ. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology. 1973;52:456–467. 12. Graham FL, Smiley J, Russel WC, Nairu R. Characteristics of a human cell line transformed by the DNA from human adenovirus 5. J Gen Virol. 1977;36:59–72. 13. Cordier L, Duffour MT, Sabourin JC, et al. Complete recovery of mice from a pre-established tumor by direct intratumoral delivery of an adenovirus vector harboring the murine IL-2 gene. Gene Ther. 1995;2:16–21. 14. Josien R, Douillard P, Guillot C, et al. A critical role for transforming growth factor-b in donor transfusion-induced allograft tolerance. J Clin Invest. 1998;102:1920–1926. 15. Monti G, Jaurand MC, Monnet I, et al. Intrapleural production of interleukin 6 during mesothelioma and its modulation by g-interferon treatment. Cancer Res. 1994;54:4419–4423.

Cancer Gene Therapy

489


490 16. Fernandez N, Levraud JP, Haddada H, et al. High frequency of specific CD8+ T cells in the tumor and blood is associated with efficient local Il-12 therapy of cancer. J Immunol. 1999;162:609–617. 17. Baratin M, Ziol M, Romieu R, et al. Regression of primary hepatocarcinoma in cancer-prone transgenic mice by local interferon-gamma delivery is associated with macrophages recruitment and nitric oxide production. Cancer Gene Ther. 2001;8:193–202. 18. Johansson L, Linden CJ. Aspects of histopathologic subtype as a prognostic factor in 785 pleural mesotheliomas. Chest. 1996;109:109–114. 19. Barrack ER. TGF beta in prostate cancer: a growth inhibitor that can enhance tumorigenicity. Prostate. 1997;31:61–70. 20. Addison CL, Braciak T, Raltson R, et al. Intratumoral injection of an adenovirus expressing interleukin 2 induces regression and immunity in a murine breast cancer model. Proc Natl Acad Sci USA. 1995;92:8522–8526. 21. Wallach D, Fellous M, Revel M. Preferential effect of interferon-gamma on the synthesis of HLA antigens and their mRNAs in human cells. Nature. 1982;299:833–836. 22. Rosa F, Hatat D, Abadie A, et al. Differential regulation of HLA-DR mRNAs and cell surface antigens by interferon. EMBO J. 1983;2:1585–1589. 23. Collins T, Korman AJ, Wake CT, et al. Immune interferon activates multiple class II major histocompatibility complex genes and the associated invariant chain gene in human endothelial cells and dermal fibroblasts. Proc Natl Acad Sci USA. 1984;81:4917–4921. 24. Odaka M, Sterman DH, Wiewrodt R, et al. Eradication of peritoneal and distant tumor by adenovirus-mediated interferon-b gene therapy is attributable to induction of systemic immunity. Cancer Res. 2001;61:6201–6212. 25. Gerwin BI, Lechner JF, Reddel RR, et al. Comparison of production of transforming growth factor-B and plateletderived growth factor by normal human mesothelioma cells and mesothelioma cell lines. Cancer Res. 1987;47:6180–6184. 26. Fitzpatrick DR, Bielefeld-Ohmann H, Himbeck RP, et al. Transforming growth factor-beta: antisense RNA-mediated inhibition affects anchorage-independent growth, tumorigenicity and tumor-infiltrating T-cells in malignant mesothelioma. Growth Factors. 1994;11:29–44. 27. Torre-Amione G, Beauchamp RD, Koeppen H, et al. A highly immunogenic tumor transfected with a murine transforming growth factor type b1 cDNA escapes immune surveillance. Proc Natl Acad Sci USA. 1990;87: 1486–1490. 28. Vodovotz Y, Bogdan C, Paik J, et al. Mechanisms of suppression of macrophage nitric oxide release by transforming growth factor B. J Exp Med. 1993;178:605–613. 29. Bielefeldt-Ohmann H, Fitzpatrick DR, Marzo AL, et al. Patho- and immunobiology of malignant mesothelioma: characterisation of tumor infiltrating leucocytes and cyto-

Cancer Gene Therapy

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

kine production in a murine model. Cancer Immunol Immunother. 1994;39:347–359. Serve H, Steinhauser G, Oberberg D, et al. Studies on the interaction between interleukin 6 and human malignant nonhematopoietic cell lines. Cancer Res. 1991;51:3862–3866. Higashihara M, Sunaga S, Tange T, et al. Increased secretion of interleukin-6 in malignant mesothelioma cells from a patient with marked thrombocytosis. Cancer. 1992;70:2105–2108. Schmitter D, Lauber B, Fagg B, et al. Hematopoietic growth factors secreted by seven human pleural mesothelioma cell lines: interleukin-6 production as a common feature. Int J Cancer. 1992;51:296–301. Bielefeldt-Ohmann H, Marzo AL, Himbeck RP, et al. Interleukin-6 involvement in mesothelioma pathobiology: inhibition by interferon g immunotherapy. Cancer Immunol Immunother. 1995;40:241–250. Walker C, Everitt J, Ferriola PC, et al. Autocrine growth stimulation by transforming growth factor a in asbestostransformed rat mesothelioma cells. Cancer Res. 1995;55:530–536. Versnel MA, Claesson-Welsh L, Hammacher A, et al. Human malignant mesothelioma cell lines express PDGF Breceptors whereas cultured normal mesothelioma cells express predominantly PDGF a-receptors. Oncogene. 1991;6:2005–2011. Smythe WR, Kaiser LR, Hwang HC, et al. Successful adenovirus-mediated gene transfer in an in vivo model of human malignant mesothelioma. Ann Thorac Surg. 1994;57:1395–401. Boylan AM, Sanan DA, Sheppard D, et al. Vitronectin enhances internalization of crocidolite asbestos by rabbit pleural mesothelial cells via the integrin alpha v beta 5. J Clin Invest. 1995;96:1987–2001. Smythe WR, Hwang HC, Elshami AA, et al. Treatment of experimental human mesothelioma using adenovirus transfer of the herpes simplex thymidine kinase gene. Ann Surg. 1995;222:78–86. Hwang HC, Smythe WR, Elshami AA, et al. Gene therapy using adenovirus carrying the herpes simplex-thymidine kinase gene to treat in vivo models of human malignant mesothelioma and lung cancer. Am J Resp Cell Mol Biol. 1995;13:7–16. Kucharczuck JC, Elshami AA, Zhang HB, et al. Pleuralbased mesothelioma in immune competent rats: a model to study adenoviral gene transfer. Ann Thorac Surg. 1995;60:593–598. Elshami AA, Kucharczu JC, Zhang HB, et al. Treatment of pleural mesothelioma in an immunocompetent rat model utilizing adenoviral transfer of the herpes simplex virus thymidine kinase gene. Hum Gene Ther. 1996;7: 141–148. Boulanger P. Cell receptors for human adenoviruses. J Soc Biol. 1999;193:77–84.