Induction of Antitumor Immunity by Transduction of CD40 Ligand Gene

D2001 Nature Publishing Group 0929-1903/01/$17.00/+0 www.nature.com/cgt

Induction of antitumor immunity by transduction of CD40 ligand gene and interferon- gene into lung cancer Masahiro Noguchi,1 Kazuyoshi Imaizumi,1 Tsutomu Kawabe,1 Hisashi Wakayama,1 Yoshitsugu Horio,1 Yoshitaka Sekido,2 Toru Hara,1 Naozumi Hashimoto,1 Masahide Takahashi,3 Kaoru Shimokata,2 and Yoshinori Hasegawa1

1First Department of Internal Medicine, Nagoya University School of Medicine, Nagoya, Japan; Departments of 2Clinical Preventive Medicine and 3Pathology, Nagoya University School of Medicine, Nagoya, Japan.

CD40±CD40 ligand (CD40L) interaction is an important costimulatory signaling pathway in the crosstalk between T cells and antigen-presenting cells. This receptor±ligand system is known to be essential in eliciting strong cellular immunity. Here we demonstrate that murine lung cancer cells (3LLSA) transduced with the CD40L gene (3LLSA-CD40L) were rejected in syngeneic C57BL/6 mice, but grew in CD40-deficient mice to the same extent as control tumor cells. Immunohistochemical study showed that inflammatory cells, including CD4+, CD8+ T cells and NK cells, infiltrated into the inoculated 3LLSA-CD40L tumor tissue. Inoculation of 3LLSA-CD40L cells into mice resulted in the induction of 3LLSA-specific cytotoxic T-cell immunity, and the growth of parental 3LLSA tumors was inhibited when 3LLSA cells were inoculated into C57BL/6 mice mixed with 3LLSA-CD40L cells or when they were rechallenged 4 weeks after 3LLSA-CD40L cells were rejected. Furthermore, co-inoculation of interferon (IFN)- ± transduced cells (3LLSA-IFN ) with 3LLSA-CD40L cells enhanced the antitumor immunity efficiently in vivo. These results indicate that the in vivo priming with CD40L- and IFN- gene±transduced lung cancer cells is a promising strategy for inducing antitumor immunity in the treatment of lung cancer. Cancer Gene Therapy (2001) 8, 421±429

Key words: Antitumor immunity; CD40 ligand; interferon- ; cytotoxic T lymphocyte; lung cancer.

espite recent increases in therapeutic options, lung surface proteins expressed on B cells, macrophages, and Dcancer remains one of the leading causes of cancer dendritic cells. It binds a CD40 ligand (CD40L/CD154), death.1 This malignancy is often resistant to chemotherapy which is a member of the TNF superfamily expressed on and frequently presents metastasis to distant organs in the activated T cells, basophils, and mast cells.7 Although the early phase of the disease. Even if the primary tumor is CD40±CD40L interaction is primarily implicated in the removed surgically, relapse at the primary or distant sites establishment of humoral immunity,8 recent reports dem- frequently occurs. To overcome these problems, new onstrated that the CD40±CD40L interaction also plays a therapeutic approaches including cancer immunogene critical role in the induction of antitumor immunity,9 therapy have recently been investigated.2 Several reports especially for the priming of tumor-specific T cells.10 We have identified tumor-specific antigens for some tumors, previously reported that alveolar macrophages were and tumor-specific cytotoxic T cells (CTLs) have been activated by stimulation through CD40±CD40L interaction successfully isolated from patients with cancer.3 However, and developed tumoricidal activity against lung cancer most malignant tumors, including lung cancer, evade host cells, and that interferon (IFN)- induced a high immune surveillance.4 It has also been suggested that expression of CD40 molecule on the surface of alveolar antigen-specific T cells may be rendered tolerant early in macrophages.11 the course of tumor progression.5 Recent studies have Considering these findings, we hypothesized that CD40L revealed that the state of activation and/or maturation of gene transduction into lung cancer cells could activate APCs antigen-presenting cells (APCs) may determine whether T surrounding the tumor tissue, and induce antitumor cells are primed or rendered tolerant.6 CD40 is a member immunity against parental lung cancer cells in vivo, although of the tumor necrosis factor (TNF) receptor family of cell lung cancer cells generally reveal low antigenicity, and it seems to be difficult to induce lung cancer±specific cellular immunity, unlike in melanoma cells. In the present study, we Received March 22, 2001. studied the effect of CD40L gene transduction combined Address correspondence and reprint requests to Yoshinori Hasegawa, with IFN- gene transduction into lung cancer cells for the MD, PhD, First Department of Internal Medicine, Nagoya University efficient induction of antitumor immunity against lung School of Medicine, Nagoya 466-8550, Japan. cancer.

Cancer Gene Therapy, Vol 8, No 6, 2001: pp 421±429 421 422 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L

MATERIALS AND METHODS designated as pcDNA-IFN . 3LLSA cells were transfected with pcDNA-IFN using the lipofection method described Mice above. Stable transfectants were obtained by selection in 600 C57BL/6 mice (H-2b) were purchased from SLC (Shi- g/mL Zeocin (Invitrogen). These selected clones were zuoka, Japan). CD40-deficient mice were generated by a seeded into a 24-well culture dish (105 per well) and gene-targeting technique previously reported.8 A CD40+ / cultured for 24 hours, then supernatants from each well were À mouse was produced by backcrossing the originally collected to measure IFN- concentration using a murine described CD40À / À mouse to a C57BL/6 mouse. The IFN- enzyme-linked immunosorbent assay kit (R&D heterozygous littermates were intercrossed to generate the Systems, Minneapolis, MN) following the manufacturer's CD40À / À mice used in this study. These mice were instructions (data not shown). The clone that produced the genotyped by a polymerase chain reaction (PCR) of highest amount of IFN- was selected (designated as genomic DNA obtained from a tail biopsy using primers to 3LLSA-IFN ) and used in the following experiments. identify the rearranged CD40 locus, as described previ- 3LLSA-MOCK, which was established by transfecting the ously.8 All these mice were maintained at the Institute for BCMGS-neo plasmid without a cDNA insert, was used in Laboratory Animal Research of the Nagoya University this study as a control. All these transfectants were cultured School of Medicine. in RPMI 1640/10% FCS.

Tumor cells Tumor growth A murine lung carcinoma cell line (Lewis lung cancer cells; Parental 3LLSA or transfectants (3LLSA-CD40L, 3LLSA- 3LLSA), which was originally established from the lung of a IFN or 3LLSA-MOCK) (106 cells/mouse) were inocu- C57BL mouse bearing a tumor, was obtained from the lated subcutaneously (s.c.) into the right shoulder of Japanese Cancer Research Resources Bank. A murine C57BL/6 mice or CD40À / À mice (8±10 weeks old). melanoma cell line, B164A5, which also originated from a In some experiments, 106 cells of 3LLSA were inoculated C57BL/6 mouse,12,13 was obtained from Riken Cell Bank with 106 cells of one of the transfectants (3LLSA-CD40L, (Tsukuba, Japan). 3LLSA was maintained in RPMI 1640 3LLSA-IFN or 30 Gy±irradiated 3LLSA-MOCK) simul- supplemented with 1% L -glutamine, 1% penicillin±strepto- taneously into C57BL/6 mice at the same or distant sites. In mycin, and 10% fetal calf serum (RPMI/10% FCS). other experiments, 3LLSA (106 cells) were injected s.c. B164A5 was maintained in high glucose±Dulbecco's simultaneously with a mixture of transfectants [3LLSA- modified Eagle's medium (DMEM) (Gibco BRL, Grand CD40L (106 cells)+3LLSA-IFN (106 cells) or 3LLSA- Island, NY) containing 10% fetal calf serum (DMEM/10% CD40L (106 cells)+30 Gy±irradiated 3LLSA-MOCK FCS). (106 cells)] at the same or distant sites. Similarly, B164A5 (106 cells) were inoculated into C57BL/6 mice 6 cDNA and transfection with 3LLSA transfectant [3LLSA-CD40L (10 cells), 3LLSA-IFN (106 cells), or 30 Gy±irradiated 3LLSA- The expression vector containing murine CD40L cDNA MOCK (106 cells)] or a mixture of the transfectants encoding the entire coding region (BCMGS-neo-CD40L) [3LLSA-CD40L (106 cells)+3LLSA-IFN (106 cells) or was a kind gift from Dr. Hideo Yagita (Department of 3LLSA-CD40L (106 cells)+30 Gy±irradiated 3LLSA- Immunology, Juntendo University School of Medicine, MOCK (106 cells)] as described above. The mice were Tokyo, Japan). BCMGS-neo-CD40L was transfected into monitored for tumor growth weekly, and tumor size was 11 3LLSA cells as previously reported. In brief, transfections determined by measuring two perpendicular diameters with were performed on 100-mm plates using 10 g of plasmid a caliper. DNA per plate by the lipofection method with Lipofectace reagent (Gibco BRL, Gaithersburg, MD), according to the Tumor rechallenges manufacturer's instructions. After 24 hours of exposure, cells were washed three times with medium and cultured in 10 mL Either 3LLSA transfectant (3LLSA-CD40L, 3LLSA- of normal medium. Twenty-four hours after medium IFN , or 30 Gy±irradiated 3LLSA-MOCK) or a mixture exchange, cells were selected in medium containing 800 of the transfectants (3LLSA-CD40L+3LLSA-IFN or g/mL G418 (Gibco, Grand Island, NY). After 2 weeks of 3LLSA-CD40L+30 Gy±irradiated 3LLSA-MOCK) was injected s.c. into the left shoulder of C57BL/6 mice. Four selection, G418-resistant clones were selected randomly 6 from the surviving colonies. Expression of the cell surface weeks later, these mice were rechallenged with 10 cells of CD40L of these selected clones was analyzed by flow the parental 3LLSA cells at the right shoulder. Tumor size cytometry with an FITC-conjugated anti-murine CD40L at the right shoulder was measured weekly as described mAb (PharMingen, San Diego, CA). The clone that showed above. the highest expression (designated as 3LLSA-CD40L) was selected and used in the following experiments. Murine Histological evaluation and immunohistochemistry IFN- cDNA subcloned into pBluescript KS+ (pBluescript 3LLSA-CD40L or 3LLSA-MOCK was inoculated into KS+mIFN- RDB No. 1482) was obtained from Riken C57BL/6 mice. Seven days after inoculation, tumors were Gene Bank (Tsukuba, Japan). After EcoRI and NotI resected, fixed in 10% formalin, blocked in paraffin, digestion, the IFN- cDNA was inserted into pcDNA3.1 sectioned at 2 m, and stained with hematoxylin and eosin. Zeo expression vector (Invitrogen, San Diego, CA), For immunohistochemical analysis, resected sections were

Cancer Gene Therapy, Vol 8, No 6, 2001 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L 423 embedded in OCT compound (Tissue-Tek, Sakura Fine- Statistical analysis technical, Tokyo, Japan), snap-frozen in liquid nitrogen, and mounted for cryostat sectioning. Five-micrometer-thick Statistical analysis was performed using the Wilcoxon sections were air-dried and fixed with acetone and rank sum test and a P value below .05 was considered immunostained with biotinylated anti-mouse monoclonal significant. antibody against CD4 and CD8 (GK1.5 and 53-6.7, respectively, Pharmingen) and anti-asialo GM1 rabbit antibody (Wako Pure Chemical Industries, Osaka, Japan). For staining asialo GM1, the sections were blocked with RESULTS diluted normal goat serum for 20 minutes, washed in PBS for 5 minutes, and then incubated with diluted anti-asialo GM1 rabbit antibody for 30 minutes. After washing twice, the CD40L-transduced lung carcinoma cells are rejected in sections were incubated with a biotinylated anti-rabbit IgG syngeneic mice by a CD40±CD40L±dependent immune for 30 minutes in a humidified boxat room temperature. For response staining CD4 and CD8 , the sections were incubated with A murine lung carcinoma cell line expressing CD40L PBS diluted monoclonal antibody for 1 hour in a humidified (3LLSA-CD40L) was established by transducing murine boxat room temperature. After these procedures, all sections CD40L cDNA into parental 3LLSA cells as described in were washed in PBS for 5 minutes, incubated for 30 minutes Materials and Methods. Inoculation of 3LLSA-CD40L in 0.3% H2O2 in methanol, and then washed again in PBS for cells into syngeneic C57BL/6 mice resulted in complete 5 minutes, following incubation with Vectastatin Elite ABC rejection, though a small tumor mass was observed on day reagent (Vector Laboratories, Burlingame, CA) for 30 7 (Fig 1A). In contrast, 3LLSA-MOCK cells established minutes. After washing, the sections were incubated with by transducing the control empty vector grew as well as AEC substrate reagent (AEC Staining Kit, Sigma, St. Louis, the parental 3LLSA cells (Fig 1A). The growth rates of MO) for 10 minutes, followed by washing and counter- these cell lines in vitro were not significantly different staining with 1% methyl green stain solution (Muto Pure (data not shown). In CD40-deficient mice, no difference Chemicals, Tokyo, Japan). in the growth rate was observed between 3LLSA-CD40L and 3LLSA-MOCK cells (Fig 1B). IFN- ±transduced CTL assays 3LLSA (3LLSA-IFN ) cells, however, were rejected both in wild-type and CD40-deficient mice (Fig 1A and B). Spleens were removed from C57BL/6 mice that had been These results indicate that the rejection of CD40L- inoculated with the 3LLSA transfectant (3LLSA-CD40L, transduced lung cancer cells in syngeneic wild-type mice 3LLSA-IFN , or 30 Gy±irradiated 3LLSA-MOCK) or was mediated by the CD40±CD40L interaction. mixture of the transfectants (3LLSA-CD40L+3LLSA- We analyzed histologic sections of tumors 7 days after IFN or 3LLSA-CD40L+30 Gy±irradiated 3LLSA- 3LLSA-CD40L or 3LLSA-MOCK cells were inoculated MOCK). After preparing a single-cell suspension, 2Â106 into C57BL/6 mice. We observed the infiltration of spleen cells were cocultivated with 60 Gy±irradiated wild- inflammatory cells and some necrotic areas in 3LLSA- type 3LLSA cells (105 ) in 2 mL DMEM supplemented with CD40L tumor tissue (Fig 2A), whereas 3LLSA-MOCK 5% FCS, 50 M 2-mercaptoethanol (Sigma), 50 U/mL tumor tissue contained few inflammatory cells and no penicillin, 50 g/mL streptomycin, and 15 mM HEPES necrosis (Fig 2B). Immunohistochemical study showed buffer (Life Technologies, Grand Island, NY) in a six-well infiltration of CD4+ cells (Fig 2C), CD8+ cells (Fig 2D) tissue culture plate (Nunc, Roskilde, Denmark). Four days and many NK cells (Fig 2E) in the 3LLSA-CD40L tumor. later, nonadherent cells were harvested, counted, and used as These findings indicate that the transduction of the CD40L effector cells. Wild-type 3LLSA cells (5Â105 cells) were gene into cancer cells produced a local immune response 51 labeled with 100 Ci of Na2[ Cr]O4 in 0.5 mL RPMI/10% against tumor cells. FCS medium for 2 hours at 378C and used as target cells.14 Then, labeled target cells (5Â103) were cocultivated for 6 hours with effector cells at different effector to target ratios Elicitation of local antitumor immunity by 3LLSA-CD40L (E:T ratio) in wells of round-bottom 96-well microculture To investigate the effect of 3LLSA-CD40L inoculation on plates. The radioactivity released during the incubation was local antitumor response in vivo, we transplanted parental determined with 100 L of culture supernatant from each 3LLSA cells mixed with 3LLSA-CD40L cells into the well by a gamma counter. Experimental release was C57BL/6 mice. As shown in Figure 3A, the inoculation of determined from the amount of [ 51Cr] released by the target 3LLSA-CD40L cells inhibited the growth of parental cells when they were incubated with effector cells. Sponta- 3LLSA cells, whereas this inhibitory effect was not neous [ 51Cr] release was determined from culture containing observed when parental 3LLSA cells were inoculated with target cells alone. Total release was determined from 1% irradiated 3LLSA-MOCK cells (P<.02) (Fig 3A). We next IGEPAL CA-630 (Sigma) lysate of the target cells. The investigated whether co-inoculation of 3LLSA-IFN with percentage of specific [ 51Cr] release was calculated as 3LLSA-CD40L cells could augment the antitumor response follows: [(experimental releaseÀspontaneous release)/ against wild-type 3LLSA cells. When parental 3LLSA cells (total releaseÀspontaneous release)]Â100. All tests were mixed with 3LLSA-CD40L and 3LLSA-IFN cells were performed in triplicate and mean values were calculated. inoculated into C57BL/6 mice, the growth of the tumor was

Cancer Gene Therapy, Vol 8, No 6, 2001 424 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L

Figure 1. 3LLSA-CD40L tumor cells are rejected in syngeneic C57BL/6 mice but not in CD40-deficient mice. Wild-type C57BL/6 mice (A) and CD40-deficient mice (B)were inoculated subcutaneously with 106 indicated tumor cells. Tumor sizes were assessed weekly by measuring perpendicular diameter with a caliper. The results are expressed as mean diameters (in millimeters) of tumors in five to eight mice in each group. Error bars represent the standard deviation of the mean. The data are representative of at least two independent experiments. significantly inhibited compared with the tumor growth growth of B164A5 (P<.05) and inoculation of B164A5 when the parental cells were mixed with either 3LLSA- cells with 3LLSA-CD40L and 3LLSA-IFN cells resulted CD40L or 3LLSA-IFN cells (P<.01) (Fig 3A). These in almost complete rejection of the tumor (P<.002) (Fig results indicate that the transduction of CD40L gene into 3B). Considering that NK cells infiltrated the 3LLSA- tumor cells could exert an antitumor immunity against CD40L tumor tissue (Fig 2E), we speculated that the parental tumor cells, and the inoculation of CD40L gene± enhancement of nonspecific antitumor immunity might be transduced tumor cells combined with IFN- gene± involved in part of the 3LLSA-CD40L±induced local transduced tumor cells could enhance the antitumor antitumor response. response. To determine whether the antitumor immunity induced by 3LLSA-CD40L was 3LLSA tumor±specific immunity Elicitation of tumor-specific and long-lasting antitumor or not, we inoculated B164A5 melanoma with the same immunity by 3LLSA-CD40L number of 3LLSA-CD40L cells into C57BL/6 mice. It It is important to determine whether transduction of the was found that 3LLSA-CD40L significantly inhibited the CD40L gene into lung cancer cells could induce systemic

Figure 2. Histopathological and immunohistochemical analyses of tumors induced by 3LLSA-CD40L cells. C57BL/6 mice were injected with 106 3LLSA- CD40L (A,C,D,E)or 3LLSA-MOCK cells (B). Tumors were resected 7days after injection and prepared for analyses. A, B: Histopathological findings of tumors induced by 3LLSA-CD40L cells (A) and 3LLSA -MOCK cells (B) by hematoxylin/eosin staining. Many inflammatory cells are infiltrated into 3LLSA-CD40L tumor but not into 3LLSA- MOCK tumor (A: 100Âmagnification, B: 200Âmagnification). C, D, and E: Immu- nohistochemical findings of 3LLSA- CD40L tumors. Cryosections were prepared from resected 3LLSA-CD40L tumor and immunostained with monoclonal anti-mouse CD4 (C) or CD8 (D) antibody. Samples were also immunostained with anti-asialo GM1 antibody for detecting NK cells (E). CD4+ (C), CD8+ (D) and asialo GM1+ cells are observed (200Âmagnification).

Cancer Gene Therapy, Vol 8, No 6, 2001 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L 425

Figure 3. Induction of local antitumor immunity by 3LLSA-CD40L cells. 106 3LLSA cells (A)or 106 B164A5 cells (B) were inoculated s.c. into C57BL/6 mice simultaneously either with PBS (control), 3LLSA-CD40L (106 ), 30 Gy±irradiated 3LLSA-MOCK (106 ), 3LLSA-IFN cells (106 ), or mixture of 3LLSA-CD40L and 3LLSA-IFN (106 each) at the same site. Tumor growth was monitored weekly by measuring perpendicular tumor diameter with a caliper. The results are expressed as mean diameters (in millimeters) of tumors in six to eight mice in each group. Error bars represent the standard deviation of the mean. The data are representative of at least two independent experiments. *P<.01; **P<.02; #P<.002; ##P<.05. P values were calculated with the Wilcoxon rank sum test. tumor-specific immunity. Therefore, we investigated inhibitory effect on the growth of B164A5 melanoma cells whether inoculation of 3LLSA-CD40L could inhibit the when they were inoculated in the contralateral shoulder of parental 3LLSA growth at a distant site. Either 3LLSA- the mouse (Fig 4B). CD40L, 3LLSA-MOCK (30 Gy±irradiated) or 3LLSA- We next investigated whether primary inoculation of CD40L mixed with 3LLSA-IFN was implanted into the 3LLSA-CD40L could elicit protective immunity, the so- left shoulder of C57BL/6 mice, immediately followed by called vaccine effect, against parental 3LLSA cells chal- inoculation of parental 3LLSA into the right shoulder. lenged after the rejection of primarily inoculated 3LLSA- Inoculation of 3LLSA-CD40L with or without 3LLSA- CD40L. Four weeks after the primary inoculation of either IFN tended to inhibit the growth of the parental tumor 3LLSA-CD40L, 3LLSA-MOCK (30 Gy±irradiated), or (P=.074) (Fig 4A). However, 3LLSA-CD40L had no 3LLSA-CD40L mixed with 3LLSA-IFN , C57BL/6 mice

Figure 4. Effects of 3LLSA-CD40L cell inoculation on the growth of tumors at a distant site. C57BL/6 mice were inoculated s.c in their left shoulder with PBS (control), 3LLSA-CD40L (106 ), 30 Gy± irradiated 3LLSA-MOCK (106 ), 3LLSA- IFN cells (106 ), or mixture of 3LLSA- CD40L and 3LLSA -IFN (106 each). Immediately thereafter, mice were inoculated on the right shoulder with 106 3LLSA cells (A),or106 B164A5 cells (B). Tumor growth on the right shoulder was monitored weekly by measuring perpendicular tumor diameter with a caliper. The results are expressed as mean diameters (in millimeters) of tumors in four to nine mice in each group. Error bars represent the standard deviation of the mean. The data are representative of at least two independent experiments. *P=.074. P values were calculated with the Wilcoxon rank sum test.

Cancer Gene Therapy, Vol 8, No 6, 2001 426 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L

Figure 5. Inoculation of 3LLSA-CD40L cells induce long-lasting tumor-specific immunity against later challenge of 3LLSA cells. C57BL/6 mice were injected subcutaneously in their left shoulder with PBS (control), 3LLSA-CD40L (106 ), 30 Gy±irradiated 3LLSA- MOCK (106 ), 3LLSA-IFN cells (106 ), or a mixture of 3LLSA -CD40L and 3LLSA -IFN (106 each) at day 0. At day 30, mice were challenged on the right shoulder with 106 3LLSA cells (A),or106 B164A5 cells (B). Tumor growth on the right shoulder was monitored weekly by measuring perpendicular tumor diameter with a caliper. The results are expressed as mean diameters (in millimeters) of tumors in four mice in each group. Error bars represent the standard deviation of the mean. The data are representative of at least two independent experiments. *P<.05; **P<.02. P values were calculated with the Wilcoxon rank sum test.

were challenged with parental 3LLSA cells at a different site. To confirm the induction of tumor-specific CTLs against As shown in Figure 5A, the inoculation of 3LLSA-CD40L 3LLSA after the inoculation of 3LLSA-CD40L cells, resulted in the inhibited growth of parental 3LLSA cells spleens were removed from mice injected with either challenged after the 3LLSA-CD40L rejection (P<.02). In 3LLSA-CD40L, irradiated 3LLSA-MOCK, or 3LLSA- addition, the implantation of 3LLSA-CD40L mixed with IFN . The spleen cells were then subjected to secondary 3LLSA-IFN significantly enhanced the anti-3LLSA anti- stimulation with parental 3LLSA cells to generate specific tumor immunity (P<.05) (Fig 5A). However, primary CTLs in vitro. Spleen cells from mice inoculated with inoculation of 3LLSA-IFN without 3LLSA-CD40L did 3LLSA-CD40L showed a greater induction of CTL activity not inhibit the growth of parental 3LLSA cells challenged than those from mice inoculated with 3LLSA-MOCK or after the rejection of 3LLSA-IFN (data not shown). 3LLSA-IFN (Fig 6). Furthermore, the CTL activity in Furthermore, the growth of B164A5 melanoma cells was not vitro was significantly enhanced when the 3LLSA-CD40L affected after the rejection of 3LLSA-CD40L (Fig 5B). cells were inoculated with 3LLSA-IFN cells into C57BL/6

Figure 6. CTLs are generated in mice inoculated with 3LLSA-CD40L. C57BL/6 mice were inoculated subcutaneously with 3LLSA-CD40L (106 ), 30 Gy±irradiated 3LLSA-MOCK (106 ), 3LLSA-IFN cells (106 ), or a mixture of 3LLSA- CD40L and 3LLSA-IFN (106 each ). Four weeks later, spleens were removed from mice and splenocytes were cultured with irradiated 3LLSA cells. Subse- quently, splenocytes were assessed for their cytotoxic activity against 51Cr-labeled 3LLSA cells. The results are expressed as the percentage 51Cr release at three different E:T ratios. Error bars represent the standard deviation of the mean. These experiments were repeated three times with similar results. *P<.05 compared with spleen cells from mice inoculated with 3LLSA-CD40L. #P<.05 compared with spleen cells from mice inoculated with 3LLSA-MOCK or 3LLSA-IFN . P values were calculated with the Wilcoxon rank sum test.

Cancer Gene Therapy, Vol 8, No 6, 2001 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L 427 mice (Fig 6). However, the CTLs did not kill the B164A5 through CD40 shows the antiproliferative effect or increases melanoma target cells (data not shown). These results apoptosis sensitivity.29,30 In contrast, CD40 triggering favors indicate that CD40L gene transduction into lung cancer cells the growth of low-grade B-cell malignancies 29 or Kaposi's could induce the tumor-specific cell-mediated immunity in sarcoma.31 Therefore, we examined the expression of CD40 vivo, and that the secretion of IFN- in tumor tissue on the cell surface of 3LLSA cells, and we also examined the enhanced the antitumor immunity induced by CD40± direct suppression of wild-type tumor cell growth (3LLSA CD40L interaction. and B164A5) by CD40L gene±transduced clones in vitro. FACS analysis revealed no CD40 expression on wild-type tumor cells with or without IFN- , nor could we observe any DISCUSSION inhibition of wild-type tumor cell growth by stimulation with the irradiated CD40L gene±transduced tumor cells Low immunogenic tumors, such as lung cancer, may be able (data not shown). Considering these findings, we concluded to manipulate their environment and escape from the that the inhibitory effect for the growth of parental wild-type growth-restricting mechanisms of the host.4 Tumor cells cells by CD40L transfectant in vivo was due to the activation themselves or tumor-infiltrating immunomodulatory cells of host immunity induced by CD40 ligation on APCs. sometimes produce immunosuppressive cytokines and We showed that the simultaneous inoculation of CD40L 15 16 17 mediators such as TGF- , IL-10, and PGE2. Cancer transfected lung cancer cells (3LLSA-CD40L) mixed with cells also inactivate NK cells or T cells directly by expressed melanoma cells of B164A5 inhibited the subsequent cell on their surface Fas ligand,18 mucins (DF3/MUC1),19 or growth of B164A5 in vivo. Two possible mechanisms may RCAS1.20 A recent report showed that tumor-associated explain the ``nonspecific'' inhibition of tumor cell growth at macrophages suppress the expression of the CD3 chain of the site of tumor rejection. One is that macrophages or T-cell receptors.21 It was also demonstrated that T cells, dendritic cells at the tumor site are activated through CD40± which infiltrate in an early stage of the tumor lesion, suppress CD40L interaction, and they release cytokines and mediators the tumoricidal activity of NK cells.22 To overcome these such as TNF- or nitric oxide to kill tumor cells.27 They also problems and elicit effective immunity against cancer, release IL-12 to recruit NK cells those are important for efficient activation of immunomodulatory cells involved in eliciting local antitumor response.32 Our imunohistochem- innate and adaptive immunity is required.23 ical study showing the enhanced infiltration of NK cells into In this study, we demonstrated that transduction of the 3LLSA-CD40L tumor may support this scenario. Another CD40L gene into lung cancer cells resulted in the establish- possible mechanism is that B164A5 cells inoculated with ment of antitumor immunity in vivo. CD40±CD40L 3LLSA-CD40L were destroyed, phagocytosed, and pro- interaction has been shown to play a central role in the cessed by APCs activated via the CD40±CD40L interaction, induction of effective APC-T cell communication.24 APCs and tumor-associated antigens of B164A5 cells were including dendritic cells and macrophages can be activated presented to T cells, resulting in the efficient generation of via the CD40±CD40L interaction and produce IL-12, which CTLs in a short time. induces the NK cell activation and Th1 responses, resulting Previous reports have demonstrated the successful in the generation of antitumor immunity.25 Immunohisto- elicitation of in vivo protective immunity against tumors chemistry in this study revealed that NK cells and T cells such as mastocytoma, leukemia, or neuroblastoma by using infiltrated into the CD40L-transduced tumors (3LLSA- CD40L transfectant.32,33 We used lung cancer cell line CD40L) but not into the control tumor tissue (3LLSA- 3LLSA, which is known to be a very low immunogenic MOCK), suggesting that 3LLSA-CD40L cells induce the tumor.34 We did not observe the expression of H-2K (MHC activation of APCs surrounding the tumor. The activation of class I) on 3LLSA cells by FACS analysis (data not shown). APCs through CD40 molecules enhances the surface It would be difficult to establish T cell±mediated antitumor expression of costimulatory molecules and adhesion mole- immunity against low immunogenic tumors such as 3LLSA cules such as LFA-3, ICAM-1, CD80, or CD86 that are cell tumors. However, our study suggests that the enhance- important for the efficient APC±T cell interaction.26 In ment of CD40±CD40L signaling could generate tumor- addition, CD40 ligation on macrophages or monocytes specific CTLs and efficient systemic antitumor immunity enhances their TNF- production, and induces tumoricidal even in low immunogenic tumors. activity.27 We have previously demonstrated that the With regard to the treatment of lung cancer, we have 3LLSA-CD40L used in this study could activate murine reported that the alveolar macrophages did not express CD40 alveolar macrophages and enhance their tumoricidal activity molecules on the cell surface, and IFN- induced a high in vitro.11 It is also known that CD40 ligation inhibits expression of CD40 on the alveolar macrophages. Further, spontaneous or tumor-induced apoptosis of dendritic cells.28 IFN- has been known to cooperate with CD40 engagement These facts apparently indicate that stimulation of APC in modulating DC activity including high-level IL-12 through CD40 by CD40L-transduced tumor cells can production.35 Therefore, we established the IFN- gene± efficiently enhance both innate and adaptive antitumor transduced clone, and transplanted it with the CD40L gene± immunity. transduced lung cancer clone. Thus, the IFN- gene± Some malignant cells express CD40 on their surface, and transduced clone provides IFN- locally at the site of tumor the triggering via CD40 has been shown to affect the inoculation, resulting in the increase of the expression of proliferation or apoptosis of the tumor.29 In high-grade B- CD40 molecules on the macrophage or APC, and their cell lymphoma cells and breast carcinoma cells, the signaling efficient activation via CD40L on the tumor cell surface.

Cancer Gene Therapy, Vol 8, No 6, 2001 428 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L

Previous reports showed that the injection of recombinant 15. de Visser KE, Kast WM. Effects of TGF- on the immune CD40L with recombinant human fms±like tyrosine kinase 3 system: implications for cancer immunotherapy. Leukemia. ligand (Flt3L) or cotransduction of GM-CSF and CD40L 1999;13:1188±1199. gene into tumor cells enhanced the antitumor immunity 16. Salazar-Onfray F. Interleukin-10: a cytokine used by tumors to induced by CD40±CD40L interaction.36,37 In the present escape immunosurveillance. Med Oncol. 1999;16:86±94. 17. Huang M, Stolina M, Sharma S, et al. Non±small cell lung study, we showed that IFN- ±transduced cells and CD40L- cancer cyclooxygenase-2±dependent regulation of cytokine transduced cells cooperated efficiently in the generation of balance in lymphocytes and macrophages: up-regulation of immune responses against parental tumor cells. We speculate interleukin 10 and down-regulation of interleukin 12 produc- that the in vivo priming with CD40L- and IFN- gene± tion. Cancer Res. 1998;58:1208±1216. transduced lung cancer cells is a promising strategy for 18. Walker PR, Saas P, Dietrich P-Y. Tumor expression of Fas antitumor immunogene therapy for lung cancer. ligand (CD95L) and the consequences. Curr Opin Immunol. 1998;10:564±572. 19. Gimmi CD, Morrison BW, Mainprice BA, et al. Breast cancer± associated antigen, DF3/MUC1, induces apoptosis of activated ACKNOWLEDGMENTS human T cells. Nat Med. 1996;2:1367±1370. 20. Nakashima M, Sonoda K, Watanabe T. Inhibition of cell The authors thank Keiko Shimamoto for her continuous growth and induction of apoptotic cell death by the human technical assistance throughout our work. This work was tumor-associated antigen RCAS1. Nat Med. 1999;5:938± 942. supported by a grant-in-aid for COE Research from the 21. Otsuji M, Kimura Y, Aoe T, Okamoto Y, Saito T. Oxidative Ministry of Education, Science, Sports and Culture of Japan stress by tumor-derived macrophages suppresses the expres- and the 25th Aichi Cancer Research Foundation. sion of CD3 chain of T-cell receptor complexand antigen- specific T-cell responses. Proc Natl Acad Sci USA. 1996;93: 13119±13124. 22. Seo N, Tokura Y, Takigawa M, Egawa K. Depletion of IL-10± REFERENCES and TGF- ±producing regulatory T cells by administering a daunomycin-conjugated specific monoclonal antibody in 1. Lee CT, Chen HL, Carbone DP. Gene therapy for lung cancer. early tumor lesions augments the activity of CTLs and NK Ann Oncol. 1995;6:S61±63. cells. J Immunol. 1999;163:242±249. 2. Rosenberg SA. A new era for cancer immunotherapy based on 23. Ginaldi L, De Martinis M, D'Ostilio A, Marini L, Loreto MF, the genes that encode cancer antigens. Immunity. 1999;10: Quaglino D. The immune system in the elderly: III. Innate 281±287. immunity. Immunol Res. 1999;20:117±126. 3. Wang RF, Rosenberg SA. Human tumor antigens for cancer 24. van Kooten C, Banchereau J. Functions of CD40 on B cells, vaccine development. Immunol Rev. 1999;170:85±100. dendritic cells and other cells. Curr Opin Immunol. 1997;9: 4. Villunger A, Strasser A. The great escape: is immune evasion 330±337. required for tumor progression? Nat Med. 1999;5:874±875. 25. Bianchi R, Grohmann U, Vacca C, Belladonna ML, Fioretti 5. Staveley OCK, Sotomayor E, Montgomery J, et al. Induction of MC, Puccetti P. Autocrine IL-12 is involved in dendritic antigen-specific T cell anergy: an early event in the course of cell modulation via CD40 ligation. J Immunol. 1999;163: tumor progression. Proc Natl Acad Sci USA. 1998;95:1178± 2517±2521. 1183. 26. Cella M, Scheidegger D, Palmer-Lehmann K, Lane P, 6. Sotomayor EM, Borrello I, Tubb E, et al. Conversion of Lanzavecchia A, Alber G. Ligation of CD40 on dendritic cells tumor-specific CD4+ T-cell tolerance to T-cell priming triggers production of high levels of interleukin-12 and through in vivo ligation of CD40. Nat Med. 1999;5:780±787. enhances T cell stimulatory capacity: T±T help via APC 7. Banchereau J, Bazan F, Blanchard D, et al. The CD40 antigen activation. J Exp Med. 1996;184:747±752. and its ligand. Annu Rev Immunol. 1994;12:881±922. 27. Alderson MR, Armitage RJ, Tough TW, Strockbine L, 8. Kawabe T, Naka T, Yoshida K, et al. The immune responses in Fanslow WC, Spriggs MK. CD40 expression by human CD40-deficient mice: impaired immunoglobulin class switch- monocytes: regulation by cytokines and activation of ing and germinal center formation. Immunity. 1994;1:167± monocytes by the ligand for CD40. J Exp Med. 1993;178: 178. 669±674. 9. Mackey MF, Gunn JR, Ting PP, et al. Protective immunity 28. Esche C, Gambotto A, Satoh Y, et al. CD154 inhibits tumor- induced by tumor vaccines requires interaction between CD40 induced apoptosis in dendritic cells and tumor growth. Eur J and its ligand, CD154. Cancer Res. 1997;57:2569±2574. Immunol. 1999;29:2148±2155. 10. Lanzavecchia A. Licence to kill. Nature. 1998;393:413±414. 29. Costello RT, Gastaut J-A, Olive D. What is the real role of 11. Imaizumi K, Kawabe T, Ichiyama S, et al. Enhancement of CD40 in cancer immunotherapy? Immunol Today. 1999;20: tumoricidal activity of alveolar macrophages via CD40±CD40 488±493. ligand interaction. Am J Physiol. 1999;277:L49±L57. 30. Hirano A, Longo DL, Taub DD, et al. Inhibition of human 12. Riley V. Enzymatic determination of transmissible replicating breast carcinoma growth by a soluble recombinant human factors associated with mouse tumors. Ann NY Acad Sci. CD40 ligand. Blood. 1999;93:2999±3007. 1963;100:762±790. 31. Biancone L, Cantaluppi V, Boccellino M, et al. Activation of 13. Silagi S. Control of pigment production in mouse melanoma CD40 favors the growth and vascularization of Kaposi's cells in vitro. Evocation and maintenance. JCellBiol. sarcoma. J Immunol. 1999;163:6201±6208. 1969;43:263±274. 32. Nakajima A, Kodama T, Morimoto S, et al. Antitumor effect 14. Hasegawa Y, Bonavida B. Calcium-independent pathway of of CD40 ligand: elicitation of local and systemic antitumor tumor necrosis factor±mediated lysis of target cells. J responses by IL-12 and B7. J Immunol. 1998;161:1901± Immunol. 1989;142:2670±2676. 1907.

Cancer Gene Therapy, Vol 8, No 6, 2001 NOGUCHI, IMAIZUMI, KAWABE, ET AL: INDUCTION OF ANTITUMOR IMMUNITY BY CD40L 429

33. Grossmann ME, Brown MP, Brenner MK. Antitumor 36. Borges L, Miller RE, Jones J, et al. Synergistic action of fms- responses induced by transgenic expression of CD40 ligand. like tyrosine kinase 3 ligand and CD40 ligand in the induction Hum Gene Ther. 1997;8:1935±1943. of dendritic cells and generation of antitumor immunity in vivo. 34. Porgador A, Bannerji R, Watanabe Y, Feldman M, Gilboa E, J Immunol. 1999;163:1289±1297. Eisenbach L. Antimetastatic vaccination of tumor-bearing 37. Chiodoni C, Paglia P, Stoppacciaro A, Rodolfo M, Parenza M, mice with two types of IFN- gene±inserted tumor cells. J Colombo MP. Dendritic cells infiltrating tumors cotransduced Immunol. 1993;150:1458±1470. with granulocyte/macrophage colony-stimulating factor (GM- 35. Snijders A, Kalinski P, Hilkens CMU, Kapsenberg ML. High- CSF) and CD40 ligand genes take up and present endogenous level IL-12 production by human dendritic cells requires two tumor-associated antigens, and prime naive mice for a cytotoxic signals. Int Immunol. 1998;10:1593±1598. T lymphocyte response. J Exp Med. 1999;190:125±133.

Cancer Gene Therapy, Vol 8, No 6, 2001