Potential conflict of interest: Nothing to report.
Immunotherapy is often effective only for small tumor burdens and, in many cases, is restricted to subcutaneous tumors. Here, we investigated the antitumor effects of combination therapy with GM-CSF and IL-12 on orthotopic liver tumors with intermediate or large tumor volumes, or on chemically-induced multifocal liver tumors in animals. Adenoviruses encoding GM-CSF or IL-12 were injected intratumorally to animals bearing transplanted tumors, or injected via intrahepatic artery in animals with primary multifocal liver tumors induced by diethylnitrosamine. Our results demonstrated that IL-12, but not GM-CSF, monotherapy displayed significant therapeutic effects, whereas combination therapy with both cytokines displayed synergistic antitumor effects not only on transplanted tumor models with intermediate or large tumor loads, but also on carcinogen-induced multifocal liver tumors. Effector cell analyses, revealed by in vivo cell subset depletion, flow cytometry analysis, and immunohistochemical staining of tumor infiltrates, indicated that NK cells were the prominent antitumor effectors for the IL-12–mediated antitumor activity, whereas CD8+ T cells, NKT cells, and macrophages were more important than NK cells in the combination therapy–mediated antitumor effects. Both IL-12 monotherapy and combination therapy could induce various types of effectors and high levels of IFN-γ; however, the latter induced much higher levels than the former, which may explain why combination therapy is superior to IL-12 monotherapy. Conclusion: Combination therapy with GM-CSF and IL-12 represents a promising immunotherapy strategy for treating orthotopic, widespread liver tumors. (HEPATOLOGY 2007;45:746–754.)
Hepatocellular carcinoma (HCC) is one of the most important malignancies in the world. The hallmarks are that it is always identified clinically at an advanced stage and usually together with impaired liver function.1 Surgical resection has been considered the only curable approach, but only a small portion of patients are operative candidates. Most of those who can not tolerate an operation receive loco-regional therapy, such as percutaneous ethanol injection and transcatheter arterial chemoembolization.2 Unfortunately, however, a reduced hepatic reservoir resulting from underlying liver cirrhosis or repeated antitumor treatment restricts the use of these therapeutic modalities for HCC.3 As for patients who suffer from multiple tumors, distant metastasis, or cancer recurrence after initial treatment, no optimal treatment can be offered because conventional chemotherapy or radiotherapy is ineffective in countering HCC. Therefore, novel therapeutic strategies are urgently needed.
Immunotherapy has been thought to have significant advantages when applied to cancers, particularly multifocal tumor nodules or tumor metastasis. The successful immunosurveillance makes it an ideal tool for eradicating tumors systematically. Additionally, the established immunological memory following active vaccination with tumor-associated antigens or tumor cell vaccines often induces persisting tumor-specific T cells, providing a system for long-term prevention of cancer recurrence. Cytokines are often used in immunotherapy to augment antitumor immunity. Among them, granulocyte macrophage colony-stimulating factor (GM-CSF) is one of the most potent cytokines in cancer treatment.4, 5 The underlying mechanisms of GM-CSF action mainly involve the enhanced capacity of antigen presentation of dendritic cells induced by GM-CSF, which subsequently activates tumor-specific cytotoxic T lymphocytes (CTLs) and causes tumor regression.6 Interleukin-12 (IL-12) is another potent antitumor cytokine that can augment the cytotoxic activity of CTLs, natural killer (NK) cells, or natural killer T (NKT) cells against a wide variety of target cells, and induces them to secrete interferon-γ (IFN-γ).7, 8
We previously reported the effectiveness of a recombinant adenovirus carrying a GM-CSF and an endostatin gene on an early rat HCC model; however, the efficacy was quite disappointing in the treatment of large tumors.9 Similarly, the efficacy of IL-12 gene therapy, though significant in treating HCC, is also limited to small tumor burdens and, in many cases, restricted to subcutaneous tumor models. To augment the antitumor effects of immunotherapy on larger tumor burdens, especially on orthtopic liver tumors, this study performed immunotherapy by adenovirus (Ad)-mediated gene transfer of IL-12 and GM-CSF simultaneously to animals bearing large hepatic tumors or animals with chemically-induced multiple HCC nodules. Our results show that in situ tumor therapy with coadministration of Ad/GM-CSF and Ad/IL-12 synergistically reduces tumor volumes, compared to single cytokine gene therapy, and that NK cells are the main effector cells responsible for tumor regression in the IL-12–mediated gene therapy, whereas CD8+ T cells, NKT cells, and possibly macrophages as well, are the major effectors involved in the (IL-12 + GM-CSF) combination therapy.
Ad, adenovirus; CTL, cytotoxic T lymphocyte; DEN, diethylnitrosamine; GM-CSF, granulocyte macrophage colony-stimulating factor; IFN, interferon; IL, interleukin; NK, natural killer; NKT, natural killer T cell.
Materials and Methods
Cell Lines and Animals.
The mouse hepatoma cell line BNL and human embryonic kidney cell line 293 were purchased from the American Type Culture Collection (Manassas, VA). Both cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) obtained from Seromed, Berlin, Germany, supplemented with 10% fetal bovine serum obtained from Biological Industries Ltd., Kibbutz Beit Haemek, Israel. Male BALB/c mice aged 7-8 weeks and male Wistar rats aged 6-7 weeks were used in these experiments. All animal experiments were performed in accordance with the guidelines of the Animal Welfare Committee of National Taiwan University College of Medicine.
Construction of Adenoviral Vector.
The adenoviral vector containing a mouse GM-CSF cDNA (Ad/GM-CSF) or a GFP gene (Ad/GFP) under the control of a CMV immediate early gene promoter was constructed using the AdEasy system10 as described.9 Ad/IL-12, kindly provided by Dr. B.L. Chiang (National Taiwan University College of Medicine), is the adenoviral vector containing a murine single-chain IL-12 gene which encodes the two IL-12 subunits (p35 and p40) linked by a polypeptide linker.11
Generation of Orthotopic Liver Tumors and In Vivo Gene Therapy.
For single HCC nodule model, 3 × 105 BNL cells were injected into the left liver lobe of mice on day 0. The needle hole was sealed with an electric coagulator (Aaron, Petersburg, FL) immediately after the withdrawal of the needle to avoid leakage of the injected substance. The incision was subsequently sutured. A single injection of 30 μl of adenoviruses, 2 × 109 Ad/GFP, 1 × 109 Ad/GM-CSF, 1 × 109 Ad/IL-12, or 1 × 109 Ad/GM-CSF + 1 × 109 Ad/IL-12 (i.e., Ad/combined) or 30 μl PBS was administered intratumorally on day 7 or day 14 (n = 5 for each group) after tumor implantation. Tumors were measured using calipers on day 28 by an investigator blinded to the treatment groups. Tumor volume was calculated using the formula: volume = width2 × length × 0.52.
For primary multifocal HCC model, Wistar rats received 0.02 ml/kg/day of diethylnitrosamine (DEN) (Sigma, St. Louis, MO) for 10 weeks by giving weekly doses of DEN in a volume corresponding to the estimated water consumption of 7 days of drinking water (100 ppm). The weights of the rats were recorded and DEN solution was freshly prepared every week. After 10 weeks, a PE10 silicon tube was inserted into the gastroduodenal artery of the animal using an operating microscope (μ20 magnification). A single injection of adenovirus (same dosages as described above) in 100 μl or 100 μl PBS (n = 10 for each group) was infused through the hepatic artery into the liver by the silicon tube, then the gastroduodenal artery was ligated. Two weeks after treatment, rats were killed and livers were harvested and weighed. Another 10 age-matched Wistar rats without DEN feeding were killed at the same time and used as healthy control group. Therapeutic effects were determined by the modified tumor burden index which was defined as the difference of the ratios of liver weight/body weight between the treated groups and the healthy controls.
Antibody-Mediated Depletion of CD4+ or CD8+ T-Cells, NK Cells or IFN-γ.
BNL cells (3 × 105) were inoculated in the liver of BALB/c mice on day 0. Adenoviruses were intratumorally injected on day 7 after tumor implantation. CD4+ and CD8+ T cells or IFN-γ were depleted by intraperitoneal injection of 0.5 mg of anti-CD4 monoclonal antibody (mAb) (GK1.5), anti-CD8 mAb (53-6.72), or anti-IFN-γ mAb (R4-6A2), respectively, on day 5 (i.e., 2 days before the administration of adenoviruses), and then with 0.25 mg of the same mAb on days 8, 10, 12, and 19 after tumor implantation. NK cells were depleted by intraperitoneal injection of 20 μl of rabbit anti-asialoGM1 antiserum (Wako, Osaka, Japan) using the same schedule. Mice intraperitoneally injected with normal rat IgG or normal rabbit serum at the same dose and schedule were used as controls. Depletion of CD4+, CD8+, or NK cells was confirmed by flow cytometry. Tumor growth in each group was monitored on day 28 after tumor implantation.
Flow Cytometric Analysis of Tumor-infiltrating Lymphocytes.
Tumor-Infiltrating Lymphocytes (TILs) from tumor tissues were prepared on day 4 after adenovirus injection as described.12 Briefly, resected liver tumors were cut into small pieces using a razor blade. The tissue fragments were incubated for 15 minutes at 37°C in HBSS solution (1 g/10 ml) containing collagenase type I (0.05 mg/ml), collagenase type IV (0.05 mg/ml), hyauronidase (0.025 mg/ml) and soybean trypsin inhibitor (1 mg/ml) (all from Sigma-Aldrich), and DNase I (0.01 mg/ml; Roche Applied Science). Cells were recovered by centrifugation and suspended again in a fresh aliquot of the HBSS digestion solution for 15 minutes at 37°C. Undigested material was removed on a 40-μm mesh sieve, and the liberated cells were recovered and washed with RPMI 1640 medium. They were further separated on a Ficoll-Paque gradient to remove dead cells. The cells obtained were used for cytometric analysis. The surface markers of these cells were stained with directly conjugated antibodies (all from BD Biosciences Pharmingen): isothiocyanate (FITC)-conjugated anti-CD4 mAb (GK1.5), anti-CD8 mAb (53-6.72), or phycoerythrin (PE)-conjugated anti-CD3 mAb (145-2C11). NKT cells were detected with α-GalCer-loaded DimerX I (CD1d:Ig fusion protein) (BD Biosciences Pharmingen) which was probed with PE-conjuated A85-1 mAb (anti-mouse IgG). After surface staining, cells were fixed and permeabilized according to the manufacturer protocol (BD Biosciences Pharmingen), and then stained with Alexa Fluor647-conjugated anti-IFN-γ mAb (XMG1.2) or isotype-matched control Ab. The stained cells were analyzed with a FACScan (Becton Dickinson, Mountain View, CA), and the data were processed using CELLQest Software (BD Biosciences Pharmingen).
In Vitro Activation of Tumor-specific CD8+ T Cells.
To determine BNL-specific CD8+ T cells, 1 × 105 TILs were activated by incubating with 1 × 105 irradiated BNL at 37°C for 24 hours in the presence of 20 ng/ml IL-2, 1 μg/ml anti-CD28, and 2 μM monensin. After overnight activation, the cells were stained with anti-CD8 antibody, followed by intracellular IFN-γ staining as described above.
NK Activity Assay.
NK cytolytic activity was determined by lactate dehydrogenase (LDH) assay (Promega, Madison, WI) using YAC cells as target cells at the E/T ratios indicated according to manufacturer instructions. The percentage of specific lysis was calculated by the following formula: percent cytotoxicity = [(experimental LDH release − spontaneous LDH release by effector and target)/(maximal LDH release − spontaneous LDH release)] × 100. Target cells were incubated either in culture medium alone to determine spontaneous LDH release or in a mixture of 2% Triton X-100 to define maximal LDH release. All assays were performed in triplicate.
All results are expressed as means ± SE. One-way ANOVA was used to evaluate the statistical significance of the difference in tumor volumes between different groups.
Adenoviral Delivery of GM-CSF and IL-12 Synergistically Regresses Orthotopic Liver Tumors.
To test immunotherapy strategies in a clinically relevant situation, this study used orthotopic liver tumor models representing either an intermediate or a large tumor load. BNL cells (3 × 105) were injected in the left liver lobe of BALB/c mice. Usually, a tumor nodule of 10≈20 mm3 and 60≈100 mm3 could be observed on day 7 and day 14, respectively, after tumor implantation, representing the intermediate tumor burden and the large tumor burden, respectively. A single injection of adenoviruses (Ad/GFP, Ad/GM-CSF, Ad/IL-12, or Ad/GM-CSF + Ad/IL-12) was administered intratumorally to the 7-day-old and 14-day-old tumors. Liver tumors were measured on day 28 using calipers. Animals treated with Ad/GM-CSF only had marginal effects, but those treated with Ad/IL-12 showed significant tumor reduction either in the 7-day-old tumor model (Fig. 1A, P < 0.001) or in the 14-day-old tumor model (Fig. 1B, P < 0.05), compared to the control PBS or Ad/GFP-treated group. Remarkably, animals treated with Ad/GM-CSF + Ad/IL-12 (i.e., Ad/combined) almost completely regressed tumor in the 7-day-old tumor model (P < 0.001) and synergistically reduced tumor volumes in the 14-day-old tumor model (P < 0.001) compared to either monotherapy.
GM-CSF and IL-12 Combination Therapy Significantly Regresses Multifocal HCC Induced by DEN in Rats.
We further pursued the antitumor effects of cytokine immunotherapy on a multifocal liver tumor model. Primary liver tumors were induced in Wistar rats with DEN as described.13 Normally, multifocal tumors were generated within a 6-week to 8-week period. Adenoviruses or PBS were injected via hepatic artery. The liver weights and body weights of the rats were measured 2 weeks after treatment. Tumor burden was expressed as a modified tumor burden index (MTBI), which indicates the difference of the ratio of liver weight/body weight between tumor-bearing rats and normal healthy rats. Usually, normal healthy animals have a constant liver weight/body weight ratio; whereas animals bearing liver tumors have higher ratios than healthy ones.14 In our model, the mean ratio of liver weight/body weight of healthy rats was around 0.0405 ± 0.004 (n = 10), whereas that of tumor-bearing animals treated with PBS increased to 0.0697 ± 0.01 (n = 9). Thus, the MTBI of the PBS control group is 0.0292 ± 0.0123 (Fig. 1C). In contrast, the MTBI of the Ad/combined-treated animals was very close to that of healthy animals, and was synergistically reduced (92% reduction) compared to Ad/GM-CSF (26% reduction) or Ad/IL-12 (55% reduction) monotherapy (Fig. 1C). Representative photographs of the livers from the adenovirus-treated animals are shown in Supplementary Fig. 1. These results demonstrate that combination therapy with GM-CSF and IL-12 also has enormous effects on the multifocal HCC model.
GM-CSF and IL-12 Combination Therapy Induces Significantly High Levels of IFN-γ.
IFN-γ is a main cytokine induced by IL-12 and is critically involved in the development of cell-mediated immune responses, so we analyzed IFN-γ production in the animals treated with adenoviruses. The serum levels of IFN-γ in Ad/IL-12-treated BLAB/c mice were high (≈1,000 pg/ml), which lasted for about 12 to 20 days (Fig. 2). Notably, the serum levels of IFN-γ in Ad/combined-treated group were about 4-fold of those of Ad/IL-12-treated group (≈4,000 pg/ml). These results thus demonstrate that combined administration of Ad/GM-CSF and Ad/IL-12 greatly enhances IFN-γ production.
Ad/IL-12 and Ad/Combined Treatments Employ Different Effector Cells for Tumor Regression.
To further assess the effectors involved in IL-12-mediated or combination therapy-mediated antitumor immunity, BALB/c mice were depleted of CD4+ or CD8+ T cells, or NK cells by anti-CD4 or anti-CD8 mAb or anti-asialoGM1 antiserum, respectively. Mice treated with an irrelevant rat monoclonal IgG2a or normal rabbit serum at the same dose and schedule were included as controls. The depletion efficiency of each cell subset by respective antibody is shown in Fig. 3A, which shows a 97.9%, 96.2%, and 93.1% depletion for CD4+, CD8+ T cells and NK cells, respectively; whereas depletion of IFN-γ was nearly 99% as measured by ELISA (data not shown). In the animals treated with Ad/IL-12, depletion of NK cells significantly impaired the antitumor effects of Ad/IL-12 (Fig. 3B, P < 0.005). Depletion of CD4+ or CD8+ T cells also had some, but minor, effects on antitumor activity (P < 0.05). On the contrary, in animals treated with Ad/combined depletion of NK cells had no effect at all on antitumor activity (Fig. 3C), whereas tumor growth was greatly restored in the CD4+ T cell–depleted mice (P < 0.001), and was partially restored in the CD8+ T-cell–depleted mice (P < 0.05). In both treatments, neutralization of IFN-γ greatly impaired the antitumor effects (Fig. 3B,C), establishing that IFN-γ is critical to the antitumor effects of Ad/IL-12 and Ad/combined treatments.
Ad/Combined Treatment Elicits Much Higher Levels of Diverse Effectors in the Tumor Regions Than Ad/IL-12 Treatment.
The markedly different results of subset depletion experiments between Ad/IL-12 and Ad/combined treatments prompted us to compare further the propensities of the effectors induced by these 2 treatments. Tumor infiltrating cells were isolated on day 4 after adenovirus injection and analyzed by flow cytometry. Significantly higher levels of IFN-γ–secreting CD4+ T cells, CD8+ T cells, and NKT cells were detected in the tumor regions of the animals treated with Ad/combined than in the tumor regions of the animals treated with Ad/IL-12 (Fig. 4A). In contrast, Ad/IL-12 treatment induced higher levels of NK infiltrates than Ad/combined treatment. As for Ad/GM-CSF treatment, it only induced moderately higher levels of CD4+ and CD8+ cells than Ad/GFP or PBS control treatment.
Activation of NKT cells requires the expression of CD1d molecule on the surface of antigen-presenting cells, e.g., dendritic cells (DCs). So, we measured the levels of CD1d+ DCs in tumor infiltrating cells. Data illustrated in Fig. 4B show that Ad/combined treatment induced markedly higher levels of CD1d+CD11c+ DCs than the other treatments. By activity assays, we further confirmed that Ad/combined treatment induced higher levels of tumor-specific CTLs than did Ad/IL-12 treatment (Fig. 4C), whereas Ad/IL-12 treatment induced higher cytolytic NK activity than Ad/combined treatment (Fig. 4D). All these results are in good agreement with the data presented in Fig. 4A.
We further characterized the CD4+ population in tumor infiltrating lymphocytes by triply staining them with anti-IFN-γ, anti-CD4, and α-galactosylceramide (α-GalCer)-loaded CD1d:Ig DimerX I which stains NKT cells. It was found that in mice treated with Ad/IL-12, 59.6% of the CD4+ IFN-γ+ cells were NKT cells; whereas Ad/combined treatment further increased the NKT population to 65.2% (Fig. 5A). Moreover, we also found that in the activated NKT cells, the CD4− population, probably representing the CD4/CD8 double negative NKT cells, was significantly increased in Ad/combined treatment group compared to Ad/IL-12 treatment group (30.3% versus 13.6%) (Fig. 5B). It was recently reported that the CD4/CD8 double negative NKT subset might exert higher antitumor activity than the CD4+ NKT subset.15
Collectively, these data indicate that GM-CSF and IL-12 combination therapy seems able to induce much higher levels of all types of effectors, except NK cells, than IL-12 monotherapy, which may explain why combination therapy has better antitumor effects than IL-12 monotherapy.
Combination Therapy Induces Extraordinarily High Levels of Activated Macrophages in Tumor Regions.
IFN-γ is also known able to activate macrophages and induce them to produce NO by iNOS expression, which exhibits their tumoricidal activity.16 So, we performed immunohistochemical staining with anti-Mac-3 or anti-iNOS antibody to detect any macrophages activated in tumor tissues of the animals treated with adenoviruses. Intense infiltration of activated macrophages was observed in tumor beds of the animals treated with Ad/combined, but much less in those with other treatments (Fig. 6A). The infiltration was in response to IFN-γ secretion, as macrophage infiltration was greatly reduced when animals were depleted of IFN-γ upon Ad/combined treatment. (Fig. 6B). Thus, tumoricidal macrophages were recruited and activated by the local high levels of IFN-γ.
Large tumor burden or multifocal tumor nodules are often the most challenging problems that have been encountered when treating cancer patients. Unfortunately, some preclinical studies suggest that immunotherapy would be effective only when provided in the minimal tumor setting.5, 17 However, we present data from this study demonstrating that combination therapy with GM-CSF and IL-12 significantly regressed orthotopic liver tumors not only in inoculated tumor models with intermediate or large tumor burdens, but also in DEN-induced multifocal HCCs. We show that this combination mode could be particularly promising for treating orthotopic liver tumors.
The uniqueness of this study includes the use of orthotopic liver tumor models to investigate liver cancer immunotherapy. It has been shown that the local environment in different organs may affect the properties of tumors growing at those sites.18 In addition, the effectors induced may be altered by different microenvironments.19 Indeed, although we identified CD8+ T cells and NKT cells as important antitumor effectors for regressing orthotopic liver tumors, other studies, albeit using the same combination strategy, demonstrated that CD8+ and/or NK cells were the major effectors for eradicating subcutaneous tumors.20, 21 The discrepancy is likely due to the high levels of NKT cells present in the liver but not in the subcutis. Thus, to be more clinically relevant, an orthotopic tumor model is preferred than a subcutaneous model to study liver cancer immunotherapy.
GM-CSF has several roles that may synergize with IL-12 to stimulate antitumor immunity. First, in the priming phase GM-CSF can enhance the recruitment of professional APCs which prime de novo CD4+ and CD8+ T cell responses and elicit tumor-specific adaptive responses.6 Second, local secretion of GM-CSF also amplifies the effector phase by enhancing the processing and presentation of tumor antigens to memory CD4+ T cells. Through local release of IFN-γ, tumoricidal macrophages and eosinophils may be recruited and activated.22 Third, GM-CSF can induce the expansion of CD1d-restricted NKT cells through enhancing the CD1d levels on DCs.23, 24 As a result, combination of GM-CSF with IL-12 can enhance the recruitment and coordination of a wide variety of effectors, such as macrophages, NKT cells, and lymphocytes, thus leading to synergistic antitumor effects. As so, its antitumor activity is no longer NK cell-dependent (Fig. 3C).
Our results demonstrated that combination therapy with GM-CSF and IL-12 leads to significantly high levels of IFN-γ secretion (Fig. 2). The NKT cells were believed to be one of the major sources of IFN-γ production due to the following observations: (1) CD4+ T cells were important for the antitumor immunity of combination therapy (Fig. 3C), and (2) 65% of the activated CD4+ T cells in the Ad/combined-treated animals were actually NKT cells (Fig. 5A). There are 2 types of NKT cells.25 Type I NKT cells, also called invariant NKT cells, express an invariant T cell receptor (TCR) α chain (Vα14-Jα281 in mice and Vα24-JαQ in human) and recognize glycolipid ligands in the context of CD1d. Activation of invariant NKT cells is associated with rapid secretion of IFN-γ and IL-4, and is markedly enhanced by DC-produced IL-12.26 Type II NKTs are also CD1d-reactive but express diverse non-Vα14 TCRs. Although they are restricted by CD1d, they do not recognize α-GalCer. Type I NKT cells have been shown to mediate both protective and regulatory immune functions, including tumor rejection, pathogen clearance, and maintenance of transplant tolerance;27 whereas type II NKT cells are responsible for inhibiting tumor immunosurveillance, involving IL-13, myeloid cells, and transforming growth factor-β.28 In this study, we used α-GalCer-loaded CD1d dimer to probe NKT cell activity, thus representing type I invariant NKT cells. In the mouse liver, invariant NKT cells, the dominant NKT cells, constitute up to 50% of intrahepatic lymphocytes.29 Among them, around 80% are CD4+ and the remaining are CD4/CD8 double negative.30 Thus, we hypothesized that in the combination therapy, activation of liver NKT cells was greatly enhanced by IL-12,26 and further expanded by GM-CSF.23, 24 The large amounts of IFN-γ secreted by activated NKT cells were critical for later activation of diverse effector cells, leading to superior antitumor effects.
Recently, Crowe et al. demonstrated that CD4/CD8 double negative liver NKT cells displayed higher antitumor activity than CD4+ liver NKT cells.15 Notably, our study showed that combination therapy activated much higher levels of CD4/CD8 double negative NKT cells than did IL-12 monotherapy (Fig. 5B), further substantiating the higher antitumor activity of combination therapy. However, the CD4+ NKT cells, though with lower antitumor activity, may still contribute profoundly to the antitumor effects since they constitute a large part of liver NKT cells. As a result, they may primarily respond to GM-CSF and IL-12 activation and secrete high levels of IFN-γ, thus accounting for the dependence of CD4+ T cells in the combination therapy-mediated antitumor effects (Fig. 3C).
In conclusion, this study demonstrates that combination therapy with IL-12 and GM-CSF represents a promising strategy for treating orthotopic liver tumors. Remarkably, our preliminary results demonstrate that this strategy also exhibits significant antitumor effects on a woodchuck model with chronic hepatitis which develops multifocal HCCs spontaneously (unpublished results). Thus, this combination mode may be considered as a potential therapeutic option to treat patients with widespread liver tumors.