NK cells are migrated and indispensable in the anti-tumor activity induced by CCL27 gene therapy

Abstract

Natural killer (NK) cells have been demonstrated could play an important role in the treatment of a number of tumors in mice. In the present study, chemokine CCL27, which be considered only selectively chemoattracts cutaneous lymphocyte antigen positive memory T cells and Langerhans cells, firstly demonstrated that it could induce the accumulation of NK cells into tumor by the intratumoral injection of CCL27-encoding fiber-mutant vector, AdRGD-CCL27. Experiments using spleen cell fractionation and RT-PCR showed CCL27 receptor, mCCR10, was strongly expressed in NK cells, suggesting the accumulation of NK cells in tumor was attributed to chemoattractant activity of CCL27 itself. Moreover, the combination of AdRGD-CCL27 and AdRGD-IL-12 induced the synergistic anti-tumor activity via NK-dependent manner and induced more NK cells infiltration into tumor nodule than that induced by AdRGD-CCL27 alone or AdRGD-IL-12 alone.

Introduction

To date, various studies of anti-tumor immunotherapy have focused on immunostimulatory cytokines, accessory molecules, or tumor antigen-loaded dendritic cells (DC) [1–3]. Unfortunately, most of these approaches have shown a limited anti-tumor response. Because of cancer cells are “self” cells that have bypassed normal homeostatic regulation mechanisms, as well as the specific characteristics of tumor tissue, it is known that infiltration of a large amount of effector cells into tumor is difficult [4, 5]. Insights into cellular and molecular events that lead to recruitment and activation of immune cells suggest that obstacles present at tumor site might be overcome and tumor immunity initiated by providing pro-inflammatory cytokines at sites of solid tumors [6]. There is some evidence to indicate that the presence of tumor-infiltrating lymphocytes is a favorable prognostic sign. In ovarian cancer, 5-year survival was significantly increased (38% versus 4.5%) in patients whose tumor biopsy samples contained CD3⁺ tumor-infiltrating lymphocytes compared with patients whose biopsy samples lacked these cells [7]. Therefore to induce an efficient anti-tumor response, large numbers of cells, such as T cells, natural killer (NK) cells and DC, capable of eliciting an effector function on presentation and activation by tumor antigen, must be attracted to the site of malignancy.

In this setting, a family of cytokines, known as chemokines, which are small (8–14 kDa), secreted basic proteins that regulate relevant leukocyte-migration and -invasion into tissue by interacting with their specific receptors, has been paid great attention [8–10]. Chemokines show chemotactic activity for a variety of immune cells as well as angiostatic activity of some kinds [11, 12]. In addition, some tumor cells express lower level of chemokines than that of normal cells [13]. Therefore, it is likely that chemokines act as important adjuvants for mediating the recruitment of relevant leukocyte subsets to the tumor site and potentiating effective anti-tumor immunity.

The results obtained by some research groups using chemokines demonstrated the tumor-suppressive activity after chemokine genes had been transduced in a variety of experimental tumors [10, 14–18]. Such as a CXC chemokine, CXCL12, induced the DC infiltration into tumor and reduced tumorigenicity [16]. Another chemokine, CX₃CL1, elicited anti-tumor activity through NK- and T cell-dependent manner [18]. Some CXC family chemokines, which deficiency in ELR motif, showed the potent anti-tumor effect both by the immune cell accumulation and the suppression of blood vessel growth in tumor [19–21].

In our previous study, a CC family chemokine, CCL27, firstly has been demonstrated that could induce tumor-suppressive activity and accumulate NK cells infiltration via transfection into OV-HM tumor cells through ex vivo approach [22]. Among CC family, several chemokines, such as CCL3, CCL7, CCL8, CCL19, and CCL21 show the chemoattractant activity to NK cells. By now, there are evidences suggesting the relationship between NK cells and anti-tumor effect induced by chemokines. Such as chemokines CCL2, CCL3, and CCL5 have been reported could against either NK-sensitive K562 leukemia cells or NK-resistant RAJI lymphoma cells via enhancing the killing activity of NK cells [23]. Chemokines CCL19 and CXCL10 also demonstrated their anti-tumor effects in mice through angiostatic mechanisms that required the participation of NK cells [24–27]. It has been supposed that activated NK cells also kill the cells in the surrounding tissues including endothelial cells, thereby impairing neovascularization of the tumor.

However, CCL27 has been reported expressed in skin and selectively chemoattracts cutaneous lymphocyte antigen positive memory T cells and Langerhans cells [28, 29]. Therefore, it remains several interesting questions such as if CCL27 gene transfer could induce the accumulation of NK cells infiltrating into established tumor mass? What is the mechanism if it happens? Is it directly induced by the ligand–receptor interaction? Moreover, if the gene therapy with CCL27 will elicit tumor-suppressive effects in vivo like that happened ex vivo, which factor can be combined to enhance the anti-tumor activity of chemokine? In the present research, all these questions were elucidated.

Materials and methods

Animal and cells

OV-HM ovarian carcinoma cells and Meth-A fibrosarcoma cells were kindly provided by Dr Fujiwara Hiromi (School of Medicine, Osaka University, Japan). OV-HM cells were maintained in RPMI 1640 supplemented with 10% heat-inactivated FBS. Meth-A cells were maintained by intraperitoneal passage in syngeneic BALB/c mice. Human embryonic kidney (HEK) 293 cells were obtained from JCRB cell bank (Tokyo, Japan) and were cultured in DMEM supplemented with 10% FBS. Female (C57BL/6xC3H/He)F1 (B6C3F1) mice, C57BL/6 mice and BALB/c mice were purchased from SLC Inc. (Hamamatsu, Japan) and used at 6–8 week of age.

Development of adenovirus vectors

Replication-deficient adenovirus vectors with a fiber mutation used in this study were based on the adenovirus serotype 5 backbone with deletions of E1 and E3 and the expression cassette in the E1 region. The RGD sequence was inserted into the HI loop of the fiber knob using a two-step method developed by Mizuguchi et al. [30, 31]. The construction of adenovirus vector encoding chemokine CCL27, AdRGD-CCL27 was reported previously [22]. And AdRGD-IL-12, which carried the murine IL-12 gene derived from mIL-12 BIA/pBluescript II KS(−) (kindly provided by Prof. Hiroshi Yamamoto; Graduate School of Pharmaceutical Sciences, Osaka University, Suita, Japan), were constructed. The expression cassette, which was designed to be transcribed in order from IL-12 p35 cDNA to the internal ribosome entry site sequence to IL-12 p40 cDNA under the control of the cytomegalovirus promoter, was inserted into the E1-deletion region of each E1/E3-deleted Ad vector by an improved in vitro ligation method as described [32]. The AdRGD-Luc vector, serving as a control vector, is identical to the AdRGD-CCL27 and AdRGD-IL-12 vectors with the luciferase gene in the expression cassette. The adenovirus vectors were propagated in HEK 293 cells and purified by cesium chloride gradient ultracentrifugation, and their titer was determined by plaque-forming assay using 293 cells according to the approach reported previously [22].

Analysis of IL-12 gene transduction in vitro and in vivo

A total of 5 × 10⁵ OV-HM cells cultured on six-well plates were transfected with AdRGD-IL-12 for 24 h at 0, 1, 5, 10 multiplicity of infection (MOI) in 2 ml of RPMI 1640 medium complemented with 10% FBS, respectively. AdRGD-Luc was used as control. After washing three times with PBS, a 1.5-ml aliquot of culture medium was added to each well. After 24 h, the supernatants were collected and the IL-12 p70 level was measured using a murine IL-12 p70 ELISA KIT (Biosource International, Camarillo, CA, USA) following the manufacturer’s instruction. Furthermore, the IL-12 p70 production in tumor was also evaluated. A total of 2 × 10⁷ PFU of AdRGD-IL12 were intratumorally injected into OV-HM tumor after 7 days inoculation of 1 × 10⁶ cells. After 2 days, the tumor nodule was harvested and homogenated, IL-12 p70 was measured using the method described above.

Cell fractionation

Spleen cells of C57BL/6 mice were mechanically dispersed and magnetic cell sorting using MACS (Miltenyi Biotech, Germany) was conducted to purify the cells with indicated surface markers. In brief, CD4⁺ cells, CD8⁺ cells, and B220⁺ cell (Miltanyi) were labeled with anti-CD4 microbeads, anti-CD8 microbeads, and anti-B220 microbeads, respectively, positively selected on a VS separation column (Miltenyi). For separation of macrophages, plastic adherent cells were sequentially resuspended with EDTA treatment and pipet vigorously, stained with FITC-conjugated F4/80 (Pharmingen), and positively selected on a VS column after labeling with anti-FITC microbeads (Miltenyi). For separation of NK cells, spleen cells were double-stained with biotinylated anti-CD3 and FITC-conjugated anti-NK1.1. CD3⁺ cells were negatively isolated on a BS depletion column (Miltanyi) after reacting with streptavidin microbeads. Subsequently, NK1.1⁺ cells were positively selected on a VS column after reacting with anti-FITC microbeads. The purity of each cell fraction was measured by flow cytometer after staining with FTIC-labeled antibodies to respective cell surface markers.

RT-PCR of fractionated spleen cells

Trizol (Gibco-BRL) and SuperScript αreverse transcriptase (Gibco-BRL) were used for total RNA extraction and cDNA reverse transcription, respectively. cDNA was amplified by PCR in the presence of primers. Primer sequences for mouse CCR10 were 5′-AGAGCTCTGTTACAAGGCTGATGTC-3′ (sense) and 5′-CAGGTGGTACTTCCTAGATTCCAGC-3′ (antisense), and for glyceraldehydes 3-phosphate dehydrogenase (GAPDH) were 5′-GCCAAGGTCATCCATGACAACTTTGG-3′ (sense) and 5′-GCCTGCTTCACCACCTTCTTGATGTC-3′ (antisense). PCR cycling condition were 94°C 30 s, 55°C 30s, 72°C 30s, 30 cycles. PCR products were run on 2% agarose gel and stained in ethidium bromide.

In vivo gene therapy of established OV-HM tumor with adenovirus vectors encoding chemokine and cytokine

B6C3F1 mice were inoculated intradermally with OV-HM cells (1 × 10⁶ cells/100 μl/mouse) and injected intratumorally with recombinant adenovirus vectors (AdRGD-IL-12, AdRGD-CCL27, or AdRGD-IL-12 + AdRGD-CCL27 in a ratio of 9:1) at 2 × 10⁷ PFU in a volume of 50 μl diluted in PBS after the diameter of tumor reached about 7–8 mm. AdRGD-Luc was used as control vector and same dose was also injected intratumorally. Tumor growth was monitored by measurement of the tumor size twice a week by two perpendicular tumor diameters using a precision caliper. Animals showing severe distress or with tumors that exceeded 20 mm in one diameter were sacrificed for ethical reasons according to institutional guidelines.

Depletion of lymphocytes and tumor growth

Depletion experiment was carried out as intraperitoneal injection of asialo GM1 antiserum (40 μl/dose; Wako, Japan) six times on days −2, −1, 0, 5, 10, and 15. OV-HM tumor was intratumorally injected with indicated adenovirus vectors on day 0. The depletion of NK cells was confirmed by FACScan flow cytometer. Tumor growth was examined twice a week.

Immunohistochemical staining

Immunohistochemical analysis was utilized to determine NK cells infiltrated into tumors. When the length of OV-HM tumor or Meth-A tumor reached about 7–8 mm, intratumoral administrations of indicated Ad were carried out. Tumor-bearing mice were sacrificed in 2 days after the intratumoral administration of AdRGD-Luc, AdRGD-CCL27, AdRGD-IL-12, or AdRGD-IL-12 + AdRGD-CCL27. The tumor nodules were harvested, embedded in the compound OCT (Sakura, Torrance, USA), and stored at −80°C. Frozen thin sections (6 μm in thickness) of the nodules were fixed in 4% PFA washed with PBS, and then were incubated with 0.3% hydrogen peroxide in methanol for 30 min at room temperature to block endogenous peroxidase activity. The sections were preincubated with 5% BSA in PBS for 10 min and sequentially incubated with optimal dilution of primary antibody, rabbit anti-asialo GM1 (WAKO, Japan), or normal rabbit IgG (Santa Cruz Biotechnology, USA). Each primary antibody bound was detected with biotinylated goat anti-rabbit immunogloblins (DakoCytomation, Kyoto, Japan), and streptavidin-HRP (Vector Laboratories, USA). Each of the incubation lasted for 30 min and was followed by a 15-min wash in Tris-buffered saline (pH 7.6). The sections were stained with diaminobenzidine solution (WAKO, Japan), and finally counterstained with hematoxylin. The number of immunostained cells was counted under a light microscope with ×400 magnification. For counting the positive cell number infiltrated into tumor tissue, six fields were randomly selected.

Results

Intratumoral injection of AdRGD-CCL27 induced the migration of NK cells into tumor but could not elicit the anti-tumor activity

AdRGD-CCL27 was constructed and the chemotactic activity for cells expressing specific receptor induced by AdRGD-CCL27 has been confirmed according to the approach reported previously [27]. As shown in Fig. 1a, b, there were NK cells infiltrated into tumor tissue with the administration of AdRGD-CCL27, whereas no NK cells accumulation with the control vector, AdRGD-Luc. However, there was no notable anti-tumor activity appeared by the injection of AdRGD-CCL27 (Fig. 1c).

Fig. 1.

Intratumoral injection of AdRGD-CCL27 induced the migration of NK cells into tumor nodule but did not arise the anti-tumor activity. 1 × 10⁶ of OV-HM cells were inoculated and CCL27 encoded AdRGD was intratumorally injected after 1 week. AdRGD-Luc was used as control vector. Part of tumor-bearing mice (six mice per group) were sacrificed in 2 days after the intratumoral administration of a AdRGD-Luc; b AdRGD-CCL27 in total of 2 × 10⁷ PFU for determining NK cells infiltrated into tumors by immunohistochemical staining. Another part of mice (seven mice per group) were maintained and the tumor diameter was measured and the size was calculated in the indicated days after treatment c

RT-PCR analysis of mCCR10 expression in fractionated spleen cells

Expression of mCCR10, the specific receptor of CCL27, in various mouse immune cells was carried out by RT-PCR. The spleen cells of C57BL/6 mice were fractionated using various surface markers. As shown in Fig. 2, mCCR10 expression was detected in CD4⁺ T cells, CD8⁺ T cells, B cells, and NK cells, but no signal was detected in macrophage.

Fig. 2.

RT-PCR analysis of mCCR10 expression in fractionated spleen cells. Total RNA was isolated from the indicated cell fractions, reverse transcribed and subjected to PCR amplification using mCCR10 specific primers. Simultaneously, amplification of GAPDH was performed as an internal control. Amplification products were run on 2% agarose gel and stained with ethidium bromide. Representative results from two separate experiments are shown. The purity of each cell fraction used was as follows: CD4⁺ 95.1%, CD8⁺ 94.5%, B220⁺ 97.1%, macrophage 81.6%, NK1.1⁺ CD3⁻ cells 88%

IL-12 p70 can be detected in the supernatant of OV-HM cells transfected and in tumor injected intratumorally with AdRGD-IL-12

The mouse IL-12-encoding fiber-mutant Ad was constructed. Here, the production of IL-12 p70 induced by the OV-HM cells transfected with AdRGD-IL-12 was confirmed by ELISA (Fig. 3). IL-12 p70 production appears dose-dependent with the transfection of viral MOI. In contrast, there is no detectable IL-12 p70 existed in the supernatant of AdRGD-Luc-transfected cells. The results demonstrated the secretion ability of developed IL-12-encoding vector. In tumor, IL-12 p70 also could be detected and the results demonstrated the expression of transferred gene in vivo.

Fig. 3.

Analysis of IL-12 gene transduction in vitro and in vivo. a OV-HM cells were transfected with AdRGD-IL-12 for 24 h at indicated MOI. AdRGD-Luc was used as control. After washing three times with PBS, a 1.5 ml aliquot of culture medium was added to each well. After 24 h, the supernatants were collected and the IL-12 p70 level was measured using a murine IL-12 p70 ELISA KIT. b One week after 1 × 10⁶ OV-HM cells were intradermally inoculated, 2 × 10⁷ PFU of AdRGD-IL-12, AdRGD-Luc, or PBS were intratumorally injected and tumors were harvested after 2 days. Tumor tissues were homogenated and the level of IL-12p70 was measured

Synergistic anti-tumor effect induced by the combination of intratumoral injection of AdRGD-CCL27 and AdRGD-IL-12

In vivo therapeutic experiment was carried out to explore the anti-tumor activity of chemokine + cytokine. OV-HM cells were inoculated and indicated chemokine and/or cytokine encoded Ad vectors were intratumorally injected after one week. The results showed that there was no notable anti-tumor effect induced by the administration of AdRGD-CCL27 or a control vector, AdRGD-Luc, at a dose of 2 × 10⁷ PFU (Fig. 4b, c). To the contrary, topic injection of AdRGD-IL-12 at the same dose elicited tumor regress in about 50% of OV-HM tumor-bearing mice (Fig. 4d), while the co-injection of both AdRGD-IL-12 and AdRGD-CCL27 (in ratio of 9:1), also at an in total dose, 2 × 10⁷ PFU, induced markedly enhanced anti-tumor activity, in which group all the tumors were regressed and all mice survive 3 months after treatment (Fig. 4e).

Fig. 4.

In vivo anti-tumor activity of chemokine + cytokine. 1 × 10⁶ of OV-HM cells were inoculated and indicated chemokine or cytokine encoded AdRGDs were intratumorally injected after 1 week. a PBS, b AdRGD-Luc, c AdRGD-CCL27, d AdRGD-IL-12, or e co-injection of both IL-12 and CCL27 (9:1 in ratio) at the same dose of in total 2 × 10⁷ PFU. Tumor diameter was measured and the size was calculated in the indicated days after treatment

Combination of chemokine and cytokine induced more NK cells infiltration into tumor tissue

Immunohistochemical staining was used to determine the lymphocyte infiltrating OV-HM tumors after the treatment with all vectors. To determine the relative number of NK cells in the tumors, six high-power fields were randomly selected, per section and positively staining cells were counted and averaged. Sections from three individual mice were used assess the relative number of stained cells per group. As shown in Fig. 5a–d, the results demonstrated that the combination of cytokine and chemokine induced more NK cells infiltrated into OV-HM tumor tissues than that of AdRGD-IL-12 alone or AdRGD-CCL27 alone. The results were confirmed by another tumor, Meth-A fibrosarcoma, in which IL-12 + CCL27 also significantly enhanced NK cells infiltration compared to that induced by AdRGD-IL-12 alone or AdRGD-CCL27 alone (Fig. 6).

Fig. 5.

Co-injection of AdRGD-IL-12 and AdRGD-CCL27 induced more NK cells accumulation in OV-HM tumor and its anti-tumor activity is NK-dependent. Immunohistochemical analysis was utilized to determine lymphocytes infiltrated into tumors. When the length of tumor reached about 7–8 mm, intratumoral administrations of indicated Ad were carried out. Tumor-bearing mice were sacrificed in 2 days after the intratumoral administration of a AdRGD-CCL27, b AdRGD-IL-12, or c AdRGD-IL-12 + AdRGD-CCL27 in total of 2 × 10⁷ PFU. The tumor nodules were harvested, embedded in the compound OCT, and stored at −80°C. Frozen thin sections of the nodules were fixed and stained for NK cells. d The number of immunostained cells was counted under a light microscope with ×400 magnification. For counting the positive cell number infiltrated into tumor tissue, six fields were randomly selected. *P < 0.05 using Student’s t-test to compare difference between two groups (IL-12 versus IL-12 + CCL27 and CCL27 versus IL-12 + CCL27). e 1 × 10⁶ OV-HM cells were intradermally inoculated at −7 day. For NK-cell depletion, mice were treated with asialo GM1 antiserum by intraperitoneal injection six times on days −2, −1, 0, 5, 10, and 15. OV-HM tumor was intratumorally injected with AdRGD-IL-12 + AdRGD-CCL27 in a ratio of 9:1 (in total of 2 × 10⁷ PFU) on day 0. The depletion of NK cells was confirmed by FACScan flow cytometer. Tumor growth was examined twice a week

Fig. 6.

Co-injection of AdRGD-IL-12 and AdRGD-CCL27 induced more NK cells accumulation in Meth-A fibrosarcoma. Immunohistochemical analysis was utilized to determine lymphocytes infiltrated into tumors. When the length of tumor reached about 7–8 mm, intratumoral administrations of indicated Ad were carried out. Tumor-bearing mice were sacrificed in 2 days after the intratumoral administration of AdRGD-CCL27; AdRGD-IL-12, or AdRGD-IL-12 + AdRGD-CCL27 in total of 2 × 10⁷ PFU. The tumor nodules were harvested, embedded in the compound OCT, and stored at −80°C. Frozen thin sections of the nodules were fixed and stained for NK cells. The number of immunostained cells was counted under a light microscope with ×400 magnification. For counting the positive cell number infiltrated into tumor tissue, six fields were randomly selected. *P < 0.05 using Student’s t-test to compare difference between two groups (IL-12 versus IL-12 + CCL27 and CCL27 versus IL-12 + CCL27)

NK cells are involved in the anti-tumor efficacy of combined intratumoral administration of AdRGD-IL-12 and AdRGD-CCL27

In the setting of OV-HM-derived intradermal tumor nodules treated with both AdRGD-IL-12 and AdRGD-CCL27 given in a single dose, the requirement of NK cells for the therapeutic effect was evaluated. The depletion of NK cells were carried out and the results also demonstrated that the depletion of NK cells elicited the disappearance of anti-tumor activity in three of seven mice (Fig. 5e). The results indicated that NK cells are necessary in the anti-tumor responses induced by the combination of chemokine CCL27 and cytokine IL-12.

Discussion

CCL27, also called CTACK (cutaneous T cell attracting chemokine) or ILC (interleukin-11 receptor α-locus chemokine), has been reported expressed in skin and selectively chemoattracts cutaneous lymphocyte antigen positive memory T cells and Langerhans cells [28, 29]. It can induce the expression of the T cell-specific cytokine IL-2, CCR10, and the α-chain of the lymphocyte function-associated antigen-1α (LFA-1α) in a dose-dependent manner [33]. However, in our previous study, we firstly observed that CCL27 induced the NK cells migration in vivo while transfected into OV-HM tumor cells in vitro through fiber-mutant adenovirus vector [27].

Therefore, in the present study, the intratumoral injection of AdRGD-CCL27 in established tumor was conducted and the results showed that it also induced the NK infiltration into tumor nodule (Fig. 1). Subsequently, RT-PCR analysis of mCCR10 expression, which is the specific receptor of CCL27, in fractionated spleen cells was carried out and the results demonstrated that mCCR10 was strongly expressed in NK cells (Fig. 2). It is the first time indicated that the chemoattractant activity of CCL27 to NK is directly by ligand-receptor interaction.

However, compared to ex vivo experiments using CCL27 gene transferred OV-HM cells, in which about 70% tumor were regressed completely [27], in vivo anti-tumor activity was not satisfied although the intratumoral injection of AdRGD-CCL27 do induced a large number of NK cells accumulation (Fig. 1). Similarly, though the findings of recent studies provide experimental evidence that introduction of chemokines into the tumor environment results in the recruitment of relevant leukocyte subsets in vivo and decreases tumorigenicity of malignant cells. Most of studies used ex vivo method and few results show transfection with chemokine alone can induce completely regress of tumors, especially in established tumor mass [34–36]. Studies showed that chemokine genes such as the one encoding XCL1 and CCL3 have been transfected into tumor cells demonstrating that although they attracted T lymphocytes to the malignant tissue, they failed to induce rejection [26, 37]. The results suggested that only the accumulation of immune cells into tumor could not induce the notable tumor regress. On the other hand, it also has been reported that intraperitoneal administration of murine recombinant IL-12 protein failed to induce the growth inhibition of Meth-A fibrosarcoma even with 500 ng/time injected once day for 3 days [38]. Further studies showed that the failure of therapy was due to the inability of immune cells migration into tumor cells [39].

It is well known that there are three criteria are required for the immunologic destruction of established tumors: (A) sufficient numbers of immune cells with highly avid recognition of tumor antigens must be generated in vivo, (B) these cells must traffic to and infiltrate the tumor stroma, and (C) the immune cells must be activated at the tumor site to manifest appropriate effector mechanisms such as direct lysis or cytokine secretion capable of causing tumor destruction [40]. Therefore, several studies using combination of chemokines with other cytokines or co-stimulatory molecules such as IL-2, CD80, or IL-12 and resulted in marked anti-tumor effect known as “attraction–expansion” [26, 37, 41]. The data are in agreement with previous reports in which intratumor injection of fibroblasts retrovirally modified with XCL1 lack anti-tumor activity unless co-transfected with the immunostimulatory factor IL-2 [42].

For enhancing the anti-tumor activity of CCL27 in established malignant disease, which more close to the clinical state, combined with cytokine was considered. Among proinflammatory cytokines, IL-2 and IL-12, both show strong activating effect on T cells and NK cells, are widely used in clinical trials [43–45]. Here, IL-12, which is a heterodimeric protein composed of two disulfide-linked subunits (p40 and p35) and secreted by macrophages and DC, was selected and expected this inflammatory cytokine will expand the anti-tumor effect of CCL27. It is known that IL-12 promotes a potent cellular immune response in which tumor-specific cytotoxic T lymphocytes and T helper (Th) cells are induced to secrete Th1 cytokines. Also, IL-12 appears to act directly on CD8⁺ T cells to enhance homeostatic expansion. Moreover, NK and NKT cells also contribute to the anti-tumor activity of IL-12 [46].

After the construction of IL-12-encoding fiber-mutant Ad vector (AdRGD-IL-12) and confirmation of the expression of IL-12p70 in OV-HM cells both in vitro and in vivo (Fig. 3), the therapeutic study using chemokine- and/or cytokine-encoding Ad vectors was carried out. It is interesting that the combination of AdRGD-CCL27 and AdRGD-IL-12 via intratumorally injected in the same nodule of tumor induced the synergistic anti-tumor effect on OV-HM tumor, which induced completely tumor regress in all mice while total regression observed for IL-12 was low (Fig. 4). It is hypothesized that the local production of IL-12 was only sufficient to expand or activate effector populations depending on the CD4⁺, CD8⁺, or NK cell intratumoral proportion and the chemotactic activity of CCL27 for immune cells will overcome the problem by recruiting more immune cells to the tumor site.

Therefore, immunohistochemical staining was used and the results demonstrated the AdRGD-CCL27 + AdRGD-IL-12-treated group accumulated more NK cells infiltrated into both OV-HM tumor and Meth-A tumor than that of AdRGD-IL-12 alone, and CCL27 demonstrated it is a strong recruiter for NK cells. Furthermore, the experiments using NK-depleted mice demonstrated that NK cells are indispensable in the anti-tumor activity of CCL27 + IL-12 because tumors grew in two of seven mice and maintained in one mouse with NK-depletion. In our previous report, the combination of cytokine and chemokine induced more activated T cells, both CD4⁺ and CD8⁺ T cells infiltrated into tumor, whereas using CCL27 alone could not induce the activation of immune cells even though a lot of T cells have been recruited [47]. Anti-tumor activity of IL-12 + CCL27 most likely reflect the complexity of an interconnected network of immune and nonimmune mechanisms difficult to dissect and possibly redundant in achieving the outcome of tumor regress as discussed by Narvaiza [48].

Taken together, we firstly reported the accumulation of NK cells into tumor induced by the intratumoral injection of AdRGD-CCL27. Experiments using spleen cell fractionation and RT-PCR showed CCL27 receptor, mCCR10, was strongly expressed in NK cells, suggesting the accumulation of NK cells in tumor was attributed to chemoattractant activity of CCL27 itself. Moreover, the combination of CCL27 and IL-12 induced the synergistic anti-tumor activity via NK-dependent manner and induced more NK cells infiltration than that induced by CCL27 alone or IL-12 alone.

Abbreviations

Ad

Adenovirus vector

AdRGD

RGD fiber-mutant Ad

CTL

Cytotoxic T lymphocyte

DC

Dendritic cell

FBS

Fetal bovine serum

NK

Natural killer

PBS

Phosphate-buffered saline

PFU

Plaque forming unit

MOI

Multiplicity of infection

IL-12

Interleukin 12