Dr Sadamu Homma, Department of Oncology, Institute of DNA Medicine, Jikei University School of Medicine, 3-25-8, Nishi-shimbashi, Minato-ku, Tokyo 105–8461, Japan. Email: email@example.com Senior author: Sadamu Homma
When BALA/c mice with BNL hepatocellular carcinoma (HCC) were treated with dendritic cells fused with BNL cells (DC/BNL) and recombinant murine interleukin (IL)-12, tumour development was significantly suppressed, whereas treatment with either DC/BNL or IL-12 alone did not show a tumour-suppressive effect. Antitumour activity induced by DC/BNL + IL-12 was abrogated by depletion of CD4+ T cells, but not by depletion of CD8+ T cells or natural killer cells. Splenic CD4+ T cells and CD8+ T cells from DC/BNL-treated mice showed cytotoxic activity against BNL cells after 3 days of incubation with DC/BNL, although BNL cells do not express major histocompatibility complex (MHC) class II molecules even after treatment with interferon (INF)-γ. Furthermore, CD4+ T cells killed syngeneic-irrelevant CT26 cells and even allogeneic Hepa1-6 cells. This cytotoxicity was blocked by concanamycin A, but not by an anti-Fas ligand (FasL) monoclonal antibody, indicating that cytotoxic activity was mediated by perforin. Immunofluorescence microscopy demonstrated that abundant CD4+ T cells and MHC class II-positive macrophages, but not CD8+ T cells, had infiltrated tumour tissue in mice treated with DC/BNL + IL-12. Flow cytometric analysis of tumour-infiltrating cells in mice treated with DC/BNL + IL-12 showed increases in CD4+ T cells and MHC class II+ CD11b+ cells but not in CD8+ T cells or MHC class I+ CD11b+ cells. Our results suggest that, in BNL-bearing mice treated with DC/BNL + IL-12, tumour macrophages activated by INF-γ produced by IL-12-stimulated T cells might present BNL tumour antigens and activate DC/BNL-primed CD4+ cytotoxic T lymphocytes (CTLs) in a MHC class II-dependent manner, leading to perforin-mediated bystander killing of neighbouring MHC class II-negative tumour cells.
CD8+ T cells expressing specific T-cell receptors that recognize tumour-associated antigens (TAAs) on major histocompatibility complex (MHC) class I molecules on the surface of tumour cells are known as antitumour cytotoxic T lymphocytes (CTLs) and play a central role in antigen-specific antitumour immunity.1 However, most studies of tumour-specific CD4+ T cells have focused on their ‘helper’ function.2 Professional antigen-presenting cells (APCs), such as macrophages and dendritic cells (DCs), present antigens in a MHC class II-dependent manner to TAA-specific CD4+ T cells, which function as helper T cells producing cytokines to activate tumoricidal CD8+ CTLs, natural killer (NK) cells, and macrophages.3,4
Although CD4+ CTLs have been reported by several investigators,5–7 little is known about their significance in antitumour immunity. Specific CD4+ CTLs kill APCs that present specific antigenic epitopes8 and even kill specific CTLs,9 suggesting that their role in immunoregulation is to avoid excess antigen-specific immune responses. CD4+ CTLs have been found to exhibit antitumour activity in several malignancies,10–12 but they cannot directly recognize TAAs on tumour cells except for those that express MHC class II molecules.13 However, malignant tumour cells rarely express MHC class II molecules.14
Several CD4+ CTL clones have been isolated from mice and humans.6,7,9,11 These clones recognize specific antigenic peptides on MHC class II molecules expressed on tumour cells and show tumoricidal activity mediated by perforin,15 the Fas–Fas ligand (FasL) system,16 the Fas-independent pathway,6,11 or tumour necrosis factor-related apoptosis-inducing ligand (TRAIL).17 Such situations, however, are probably unusual because MHC class II molecules are rarely expressed on tumour cells, except in haematopoietic malignancies, such as chronic B-cell leukaemia.13
Tumour cells transfected with genes encoding MHC class II show features of a tumour vaccine that could induce specific antitumour immunity,18,19 suggesting that expression of MHC class II molecules might make tumour cells immunogenic. Because MHC class II molecules are rarely expressed on tumour cells under natural conditions, we have concluded that CD4+ CTLs are not closely associated with direct tumoricidal activity. However, CD4+ T cells may occasionally infiltrate tumour tissue, even in cases not treated with immune-associated therapies.20–22 The significance of these tumour-infiltrating CD4+ T cells is unclear, but they might represent an incomplete immune response to tumour cells by tumour-specific CD4+ T cells.23,24
Tumour cells possess various capabilities that enable them to suppress host immune responses and escape immunological tumour rejection.25,26 Tumour cells produce immunosuppressive cytokines and prostanoids that inhibit induction of specific antitumour CTLs. Professional APCs in tumour tissue probably cannot induce antitumour T cells.27 Furthermore, macrophages in tumour-bearing hosts contribute to suppression of antitumour immunity.28 If professional APCs in tumour sites engulf TAAs from surrounding tumour cells, become fully activated, and present antigenic peptides of TAAs, CD4+ CTLs might be activated and exhibit tumoricidal activity.
In the present study, we demonstrate tumoricidal activity of CD4+ CTLs primed by DCs loaded with tumour cells. The DC-primed CD4+ CTLs are presumed to be activated at the tumour site by macrophages, whose antigen-presenting capacity is up-regulated by interferon (IFN)-γ induced by systemically administered interleukin (IL)-12. Activation of DC-primed CD4+ CTLs by macrophages presumably enables them to kill surrounding MHC class II-negative tumour cells by releasing perforin.
Materials and methods
Eight-week-old female BALB/c mice were supplied by Nihon SLC Co., Ltd. (Hamamatsu, Japan) and maintained in our facilities with unlimited water and standard laboratory chow under 12-hr light/dark cycles. All animals received human care according to the criteria outlined in the ‘Guide for the Care and Use of Laboratory Animals’ prepared by the National Academy of Sciences.29
Cell line, media, cytokines and antibodies
BNL, a murine hepatocellular carcinoma cell line with a genetic background of H-2d,30 was kindly provided by Dr Shigeki Kuriyama of Nara Medical University (Nara, Japan). RPMI-1640 medium, which was used for BNL cell culture, was purchased from Nissui Pharmaceutical Co., Ltd (Tokyo, Japan). Recombinant murine IL-12 was kindly provided by the Genetic Institute (Cambridge, MA). Recombinant murine granulocyte-macrophage colony-stimulating factor (rmGM-CSF) was kindly provided by Kirin Brewery Co., Ltd (Tokyo, Japan). Recombinant murine (rm) IL-4 was purchased from Becton-Dickinson (Bedford, MA). Fetal calf serum (FCS) and γ-globulin were products of Daiichi Fine Chemicals (Tokyo, Japan) and Cappel (Aurora, OH), respectively. A hamster monoclonal antibody (mAb) against murine FasL (MFL1) was kindly provided by Dr Hideo Yagita, Juntendo University (Tokyo, Japan). Phycoerythrin (PE)-conjugated mAbs against murine CD4, CD8, CD11b, CD11c, H-2Kd and I-Ad/I-Ed were purchased from Pharmingen BD Biosciences (San Diego, CA). Polyethylene glycol (PEG, molecular weight 1450) and concanamycin A were purchased from Sigma Chemical Co. (St. Louis, MO).
Preparation of DCs
Bone marrow-derived DCs were prepared from mice as described by Inaba et al.31 with slight modifications.32 Briefly, bone marrow cells were flushed out from the femurs and tibias of mice and depleted of red blood cells by treatment with 0·83% NH4Cl/0·01 m Tris-HCl (pH 7·5). Bone marrow cells were depleted of granulocytes and macrophages by short-term incubation of the cells on a dish coated with human γ-globulin.32 Floating cells were collected; suspended in the complete medium (RPMI-1640 supplemented with 5% FCS, 100 U/ml penicillin, 100 µg/ml streptomycin, 50 µmol/l 2-mercaptoethanol and 2 mmol/l glutamine) containing 10 ng/ml rmGM-CSF and 10 ng/ml recombinant murine IL-4; plated at 5 × 105 cells/well on 24-well culture plates; and cultured for 5 days. Non-adherent cells and loosely attached cells were collected by gentle pipetting. After overnight incubation in 100-mm-diameter Petri dishes, non-adherent cells were collected as the DC-rich fraction. The DC population showed dendritic features on phase-contrast microscopy; expressed MHC class I and II molecules, CD11c, CD80, CD86 and CD54; and elicited allostimulatory ability. A total of 1–2 × 106 DCs could be obtained from one mouse.
Preparation of DCs fused with BNL cells
DCs were fused with BNL cells by treating a mixture of tumour cells and DCs with 50% PEG as described previously.33–36 BNL cells were irradiated prior to PEG treatment to avoid the growth of BNL cells not fused to DCs. Briefly, the DCs and BNL cells were mixed at a ratio of 3 : 1 and centrifuged. A volume of 0·5 ml of 50% PEG solution warmed to 37° was gently added to the pellet and left for 1 min at room temperature. The PEG was removed by addition of an excessive volume of RPMI-1640 and subsequent centrifugation. The PEG-treated mixture of DCs and BNL cells was suspended in the complete medium containing 10 ng/ml rmGM-CSF and 10 ng/ml rmIL-4, and cultured overnight at 37° in a flask. Non-adherent and loosely attached cells were then collected by gentle pipetting and used as DCs fused with BNL cells (DC/BNLs). Approximately 35% of PEG-treated cells were viable DC/BNLs.34–36
Treatment of mice
Female BALB/c mice were inoculated subcutaneously in the back with BNL cells (106/mouse) on day 0. DC/BNLs were then suspended in phosphate-buffered saline (PBS) and injected subcutaneously into the backs of mice on days 3 and 10 (DC/BNLs generated by PEG treatment from 106 DCs and 3 × 105 BNL cells per mouse). A solution of IL-12 (200 ng/mouse) was injected intraperitoneally on days 5, 7, 9, 12, 14 and 16. Tumour growth was observed thereafter, with the size of tumours being measured once a week. Control mice were untreated or treated with either subcutaneous injections of DC/BNL alone or intraperitoneal injections of IL-12 alone. In some experiments, mice treated with DC/BNL + IL-12 also received intraperitoneal injections of an antimurine CD4 or CD8 mAb (0·5 mg/mouse, obtained from ascitic effusion accumulated by inoculation of American Type Culture Collection hybridoma GK1·5 or 56·6.73) on days 4, 8, 12, 16 and 20. Depletion of CD4+ and CD8+ was achieved by treatment with these antibodies.35 Some mice were also treated with an antiasialo GM1 antibody (0·5 mg/mouse; Wako Pure Chemical Industries, Osaka, Japan) in the same way as with mAbs against CD4 or CD8.
Assay of cytotoxic activity of splenocytes against BNL cells
The cytotoxic activity of splenocytes against BNL cells was assayed as described previously.34 Briefly, splenocytes were obtained by gentle disruption of the spleen on a steel mesh and depletion of red blood cells by hypotonic treatment. Splenocytes from the mice were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FCS in the presence of DC/BNL (splenocytes:DC/BNL, 500 : 1). The BNL cells (104 cells/well) were labelled with 51Cr and incubated with splenocytes (effector cells) at the indicated effector:target ratios in a final volume of 200 µl in triplicate in a 96-well plate for 4 hr at 37°. After incubation, 100 µl of supernatant was collected, and the per cent specific 51Cr release was calculated with the following formula: per cent cytotoxicity = 100 × (c.p.m. experimental – c.p.m. spontaneous release)/(c.p.m. maximum release – c.p.m spontaneous release), where c.p.m. is counts per minute, maximum release was that obtained from target cells incubated with 0·33 N HCl, and spontaneous release was that obtained from target cells incubated without the effector cells. In some experiments, splenocytes and purified CD4+ T cells or CD8+ T cells were incubated with concanamycin A (100 nm) or an antimurine FasL mAb (25 µg/ml) for 2 hr before the cytotoxic assay was performed in the presence of each agent.
Preparation of cells by a magnetic cell-sorting system
To obtain cell fractions depleted of NK cells, CD4+ T cells, or CD8+ T cells, splenocytes were treated with mAbs against DX5, CD4 or CD8, respectively, conjugated with iron microbeads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) according to the manufacturer's instructions. Cells treated with each mAb were passed through a column in a magnetic field, and cells not trapped in the column and eluted were collected as NK cell-, CD4+ T cell- or CD8+ T cell-depleted fractions of splenocytes.
To isolate purified CD4+ or CD8+ T cells, splenocytes were treated with a cocktail of mAbs conjugated with iron microbeads that recognize immune cells other than CD4+ or CD8+ T cells (Miltenyi Biotec) according to the manufacturer's instructions. Splenocytes treated with each mAb cocktail were passed through a column in the magnetic field; cells not trapped in the column and eluted were collected as a purified CD4+ or CD8+ T cell fraction of splenocytes. The CD4+ or CD8+ T cell fraction showed a purity of more than 90% of each cell.
The tumours of the mice were resected, then immediately minced, immersed in Tissue-Tek O.C.T. Compound (Sakura Finetechnical Co., Ltd, Tokyo, Japan), and frozen at −80°. The frozen tissue samples were fixed in acetone and treated with 10% normal goat serum. After being washed with PBS, the sections were incubated with PE-conjugated mAbs against murine CD4, CD8 and I-Ad/I-Ed for 1 hr at room temperature. The sections were washed with PBS and examined under a fluorescence microscope (LSM410; Carl Zeiss, Oberkochen, Germany).
Preparation of tumour-infiltrating cells
The tumours of the mice were resected, minced and washed well with Mg2- and Ca2+-free PBS. The fragments of tumour tissue were placed in RPMI-1640 medium containing 10% FCS and 1000 U/ml Dispase (Godo Shusei Co., Ltd, Tokyo, Japan) and agitated at 100 oscillations/min at 37° for 45 min. Dispersed cells were suspended in PBS, and those passing through a cell strainer (70 µm; BD Biosciences, Discovery Labware, Bedford, MA) were collected as the tumour-infiltrating cell fraction.
Flow cytometric analysis
To examine the expression of MHC class I or class II molecules on BNL cells, the cells were stained with fluorescein isothiocyanate (FITC) conjugated to mAbs against H-2Kd or I-Ad/I-Ed (Pharmingen). In some experiments, expression was examined after BNL cells had been incubated overnight in the presence of 100 ng/ml rmIFN-γ (Pepro Tech, Inc., Rocky Hill, NJ). Tumour-infiltrating cells were also examined for expression of CD4, CD8, CD11b, H-2Kd and I-Ad/I-Ed. For analysing the expression of MHC class I or class II molecules on tumour macrophages, CD11b+ cells in tumour-infiltrating cells were gated and expression of H-2Kd or I-Ad/I-Ed on CD11b+ cells was examined. Fluorescence profiles were generated with a FACSCalibur flow cytometer (BD Biosciences Immunocytometry Systems, San Jose, CA). Histograms and density plots were generated with the CellQuest software package (BD Biosciences Immunocytometry Systems).
Combined treatment of BNL-bearing mice with DC/BNL and IL-12 significantly suppressed tumour development
When mice received subcutaneous injections of DC/BNL 3 and 10 days after inoculation of BNL cells, tumour growth and incidence were similar to those in untreated control mice. Tumour growth and incidence were also not suppressed by treatment with IL-12 alone (200 ng/mouse, 6 times in 2 weeks) started 5 days after inoculation of BNL cells. However, tumour growth and incidence were significantly suppressed when BNL-inoculated mice were treated with both DC/BNL and IL-12 (DC/BNL + IL-12), in which each treatment was performed as described above (Fig. 1).
Suppressive effect on tumour growth of DC/BNL +IL-12 was abrogated by depletion of CD4+ T cells, but not CD8+ T cells or NK cells
When BNL-bearing mice treated with DC/BNL + IL-12 received intraperitoneal injections of mAb against murine CD4, the tumour-suppressive effect of DC/BNL + IL-12 was completely abrogated and tumour growth and incidence were similar to those in untreated control mice (Fig. 2). In contrast, tumour growth and incidence were still inhibited when mice treated with DC/BNL + IL-12 received intraperitoneal injections of mAb against murine CD8. The tumour-suppressive effect of DC/BNL + IL-12 was not modified by treatment with antiasialo GM1 antibody (Fig. 3). In an in vitro study, splenocytes from mice treated with DC/BNL + IL-12 and antiasialo GM1 antibody showed significant cytotoxic activity against BNL cells despite the loss of that against Yac-1 cells by NK cells.
Splenic T cells from mice inoculated with DC/BNL showed cytotoxic activity against BNL cells
Cytotoxic activity against BNL cells was determined in splenocytes from untreated mice and from mice treated with DC/BNL or IL-12 or both (Fig. 4a). The cytotoxic activity of splenocytes from untreated mice or mice treated with IL-12 alone was extremely low. Splenocytes from mice treated with DC/BNL or DC/BNL + IL-12 showed significant cytotoxic activity against BNL cells, but the activity of splenocytes from mice treated with DC/BNL alone was higher than that of splenocytes from mice treated with DC/BNL + IL-12. Splenocytes from mice treated with DC/BNL + IL-12 readily became apoptotic on incubation with DC/BNL, probably because they had already been strongly activated in vivo and their incubation with DC/BNL provided excess antigen stimulation, resulting in activation-induced cell death. Accordingly, splenocytes from mice treated with DC/BNL were examined for cytotoxic activity thereafter.
Effector cells among splenocytes were identified by stepwise depletion of immune cell types from DC/BNL-activated splenocytes using a magnetic sorting system (Fig. 4b). Splenocytes depleted of NK cells showed cytotoxic activity similar to that of whole splenocytes. In contrast, depletion of CD8+ or CD4+ T cells from NK cell-depleted splenocytes significantly decreased cytotoxic activity, suggesting that T cells, both CD8+ and CD4+, are associated with the cytotoxic activity.
For the recognition of CD4+ and CD8+ T cells, MHC class I and class II molecules, respectively, must be expressed on target cells. Expression of H-2Kd and I-Ad/I-Ed molecules on BNL cells was examined with flow cytometric analysis (Fig. 4c). H-2Kd molecules were consistently expressed on BNL cells, and expression was significantly enhanced by treatment with IFN-γ. In contrast, no significant expression of I-Ad/I-Ed molecules was observed on BNL cells, even after treatment with IFN-γ.
CD4+ and CD8+ CTLs showing non-specific cytotoxic activity were induced in DC/BNL-activated splenocytes
After splenocytes from DC/BNL-inoculated mice had been incubated with DC/BNL for 3 days, CD4+ and CD8+ CTLs were isolated using a magnetic cell-sorting system. Both CD4+ and CD8+ CTLs showed significant cytotoxic activity against BNL cells (Fig. 5). Moreover, both CD4+ and CD8+ T cells killed irrelevant syngeneic CT26 cells and allogeneic Hepal-6 cells. In contrast, neither CD4+ nor CD8+ T cells from untreated mice showed cytotoxic activity after incubation with DC/BNL.
Cytotoxic activity of CD4+ CTLs was mediated by perforin, but not by FasL
We attempted to identify the mediator molecule of the cytotoxic activity of CD4+ cells in DC/BNL-activated splenocytes. Cytotoxic activity of CD4+ CTLs against syngeneic BNL cells or allogeneic Hepa1-6 cells was significantly decreased when the cytotoxic assay was performed in the presence of concanamycin A, but not when it was performed in the presence of an anti-FasL mAb (Fig. 6). These results indicate that the cytotoxic activity of activated CD4+ CTLs is mediated by perforin, but not by the Fas–FasL system. The cytotoxic activity of CD8+ T cells was also mediated by perforin (data not shown).
Abundant CD4+ T cells and MHC class II-positive macrophages infiltrated tumour tissue of mice treated with DC/BNL + IL-12
Small tumours developed in some mice treated with DC/BNL + IL-12. Tumours of DC/BNL + IL-12-treated mice were resected 17 days after BNL cell inoculation and examined with immunofluorescence microscopy. Many CD4+ T cells were observed in tumour tissue from mice treated with DC/BNL + IL-12 (Fig. 7a), but CD8+ T cells (Fig. 7b) and DX5+ cells (NK cells; data not shown) were rarely observed. Abundant MHC class II+ cells were seen in tumour tissue of mice treated with IL-12 or DC/BNL + IL-12, and many more MHC class II+ cells were observed in mice treated with DC/BNL + IL-12 than in mice treated with IL-12 alone (Fig. 7c). Many MHC class II+ cells were also positive for CD11b, a macrophage marker (data not shown). The tumour tissue of mice treated with DC/BNL alone showed little immune cell infiltration, as in tumour tissue of untreated control mice.
Flow cytometric analysis showed an increase in CD4+ T cells and MHC class II+ macrophages in tumour-infiltrating cells of mice treated with DC/BNL + IL-12
Tumour-infiltrating cells obtained through enzymatic treatment of tumour tissue were submitted to flow cytometric analysis. The frequency of CD4+ T cells was significantly higher in tumour tissue from mice treated with DC/BNL + IL-12 (Table 1) than in tissue from untreated control mice; the frequency of CD8+ T cells was also higher than in untreated control mice but was much less than the number of CD4+ T cells.
Table 1. Frequency of CD4+ and CD8+ T cells in tumour-infiltrating cells
Treatment of mice
Tumour-infiltrating cells (%)
BNL-bearing mice were untreated or treated with dendritic cells fused with BNL cells (DC/BNL) and/or interleukin (IL)-12. Tumours were removed on day 17. Tumour tissue was digested by dyspase treatment and passed through a nylon mesh. Tumour-infiltrating cells thus obtained were examined for the expression of CD4 and CD8 by flow cytometric analysis.
DC/BNL + IL-12
Expression of H-2Kd and I-Ad/I-Ed molecules on CD11b+ intratumoral macrophages was examined. In mice treated with DC/BNL + IL-12, the expression of H-2Kd on macrophages was slightly higher than in untreated control mice, but the frequency of H-2Kd-positive macrophages in mice treated with DC/BNL + IL-12 was not significantly higher than in controls (Fig. 8), whereas the frequency of I-Ad/I-Ed-positive intratumoural macrophages was significantly higher than in untreated control mice (Fig. 8).
Inoculation with DC/BNL and administration of IL-12 had a significant therapeutic effect against established BNL hepatocellular carcinoma (HCC) tumours in BALB/c mice. As we reported previously,34 splenocytes from mice inoculated with DC/BNL elicited cytotoxic activity against BNL cells in vitro, suggesting that BNL-specific cytotoxic T cells were induced. The tumour-suppressive effect elicited by treatment with DC/BNL + IL-12 was abolished by treatment with an anti-CD4 mAb, but not with an anti-CD8 mAb, indicating that CD4+ T cells are associated with the antitumour activity. Because IL-12 receptor molecules could be induced on CD4+ T cells by stimulation through T-cell receptors and costimulatory molecules of DC/BNL,37 it is likely that secretion of IFN-γ by CD4+ T cells induced by IL-12 would be greatly enhanced, leading to induction of non-specific antitumour activity mediated by NK cells. However, depletion of NK cells by treating mice with antiasialo GM1 antibody did not modify the tumour-suppressive effects of DC/NML + IL-12. This finding indicates that NK cells are not associated with the tumour-suppressive effect.
Next, we tried to identify immune cells exhibiting cytotoxic activity against BNL cells in DC/BNL + IL-12-induced antitumour activity in vitro. Splenocytes from DC/BNL + IL-12-treated mice readily became apoptotic when incubated with DC/BNL, suggesting that T cells from DC/BNL + IL-12-treated mice had already been activated in vivo strongly enough to elicit antitumour activity, and that in vitro re-stimulation by DC/BNL might cause the T cells to undergo activation-induced cell death by excess antigen stimulation. It is likely that splenocytes from mice inoculated with DC/BNL without administration of IL-12 could be functionally activated and show cytotoxic activity against BNL cells when stimulated again by antigen-presenting cells in vitro. Splenocytes from naïve mice did not show cytotoxic activity against BNL cells, even after cocultivation with DC/BNL, indicating that in vivo priming by DC/BNL is necessary for induction of cytotoxic activity. Accordingly, for analyses of cytotoxic activity, splenocytes obtained from DC/BNL-inoculated mice and restimulated with DC/BNL in vitro, termed DC/BNL-activated splenocytes, were used.
Analysis with a magnetic cell-sorting system showed that both CD4+ and CD8+ T cells in DC/BNL-activated splenocytes (termed DC/BNL-activated CD4+ or CD8+ T cells) exhibited cytotoxic activity against BNL cells. BNL cells constitutively express H-2Kd molecules (MHC class I) but do not express I-Ad/I-Ed molecules (MHC class II) even after treatment with IFN-γ. DC/BNL-activated CD8+ CTLs could kill BNL cells in a MHC class I-dependent manner, but it is difficult to understand how DC/BNL-activated CD4+ CTLs could kill BNL cells without MHC class II expression. Although CD4+ CTL clones can kill target cells expressing MHC class II molecules,6,7,9,11 MHC class II-negative tumour cells are reportedly susceptible to activated CD4+ CTLs.38,39 Surprisingly, DC/BNL-activated CD4+ CTLs killed irrelevant syngeneic and allogeneic tumour cells in vitro. The cytotoxic activity of DC/BNL-activated CD8+ CTLs was similar. These results suggest that DC/BNL-activated CD4+ and CD8+ T cells might kill target tumour cells without recognizing specific antigenic epitopes on tumour cells.
We found that CD4+ and CD8+ CTLs exhibited perforin-mediated tumoricidal activity without involvement of the Fas–FasL system. Perforin is an important mediator molecule of immune tumoricidal activity in combination with granzyme-B, inducing apoptosis of tumour cells.40 When ‘sensitized’ CTLs, which have already been primed to a TAA, recognize the same TAA on tumour cells, they are activated and kill the tumour cells using exocytosis of perforin-granzyme A or expression of a death receptor ligand, such as FasL, or both.1 However, antigen presentation by tumour cells should be much less efficient than that by professional APCs. If TAA-primed CTLs were restimulated with potent APCs presenting TAAs, CTLs would be more strongly activated and exhibit stronger tumoricidal activity than if they were stimulated by tumour cells. CD4+ or CD8+ CTLs that had been primed in vivo by DC/BNL would probably be activated in vitro by DC/BNL in a MHC class I- and II-dependent manner. Thus activated, the CTLs, either CD4+ or CD8+, would release perforin, which would kill neighbouring tumour cells without recognition of TAAs on the tumour cells. This would explain why DC/BNL-activated CD4+ CTLs can kill syngeneic BNL cells as well as allogeneic Hepa1-6 cells.
However, clarifying the role of IL-12 in in vivo tumour-suppressive effects induced by DC/BNL + IL-12 is important. Splenocytes from mice treated with DC/BNL + IL-12 became apoptotic when incubated with DC/BNL. This result suggests that CTLs primed by vaccination with DC/BNL were activated by stimulation with professional APCs presenting TAAs of BNL cells in DC/BNL + IL-12-treated mice. Immunohistochemical studies have demonstrated abundant I-Ad/I-Ed-positive cells in BNL tumour tissue after treatment with IL-12 alone or DC/BNL + IL-12. These cells also expressed CD11b, indicating that they are macrophages. Simultaneously, many CD4+ T cells, but not CD8+ T cells, infiltrated the tumour tissue in mice treated with DC/BNL + IL-12, but not in mice treated with either DC/BNL or IL-12 alone. Flow cytometric analysis demonstrated that CD11b+ macrophages in the tumour tissue of DC/BNL + IL-12-treated mice expressed high levels of I-Ad/I-Ed molecules but low levels of H-2Kd molecules. These results strongly suggest that tumour-site macrophages activated by IFN-γ induced by IL-12 would stimulate DC/BNL-primed CD4+ CTLs in a MHC class II-dependent manner. IFN-γ would have been produced by IL-12-stimulated T cells (predominantly CD4+ T cells received antigen presentation from DC/BNL and expressed IL-12 receptor), but not NK cells, because depletion of NK cells did not modify the antitumour effect of DC/BNL + IL-12. DC/BNL-primed CD4+ T cells might have been attracted into tumour tissue because activated macrophages in the tumour site presented BNL tumour antigen. The CD4+ CTLs thus activated in tumour sites would probably be able to kill neighbouring tumour cells by means of exocytosis of perforin-granzyme A; this tumoricidal activity would not require any recognition of TAAs on tumour cells, because CD4+ CTLs are activated for specific antigens by macrophages. This conclusion is supported by the result of an in vitro study in which DC/BNL-activated CD4+ T cells demonstrated tumoricidal activity in an MHC-non-restrictive but perforin-dependent manner. It would thus be essential in vivo that professional APCs in tumour sites activate TAA-primed CTLs. Cohen et al.41 found that APCs in tumour tissue activated adaptively transferred CD4+ T cells, leading to cytokine production for cell-mediated antitumour responses. This finding explains why depletion of CD4+ T cells in the mice abrogated the tumour-suppressive effects of treatment with DC/BNL + IL-12. It is conceivable that antigen of BNL cells presented by tumour macrophages would be tumour specific and poorly expressed in other normal organs. This would explain why an autoimmune response was not induced in this animal model.
In the present in vitro study, splenocytes from DC/BNL-inoculated mice were restimulated with DC/BNL. In this situation, because DCs could present antigens in both MHC class I- and class II-dependent manners,42 DC/BNL-primed CD8+ and CD4+ T cells could be stimulated and CD8+ and CD4+ CTLs were generated. To obtain significant antitumour activity, DC/BNL-primed CD4+ CTLs are required to be activated by professional APCs in the tumour site, which express MHC class II molecules and present tumour antigen, not by tumour cells, which exhibit low antigen-presenting capacity and no expression of MHC class II. IL-12 would be required for activation of tumour macrophages which present BNL antigen to DC/BNL-primed CD4+ CTLs. Accordingly, in vitro restimulation should be performed using macrophages presenting BNL antigen. However, it was very hard to obtain intact macrophages from the tumour tissue of the mice (because ingestion of BNL antigen by macrophages occurs during tumour tissue digestion, even in control mice without IL-12 treatment). We therefore used DC/BNL for in vitro restimulation to DC/BNL-primed CD4+ CTLs as a substitute for tumour macrophages presenting BNL antigen.
In the present experiment, however, macroscopically identifiable tumours were not formed subcutaneously in DC/BNL-inoculated mice when IL-12 was administered intraperitoneally. Accordingly, as shown in Fig. 9, subcutaneous macrophages ingesting inoculated BNL cells might have been activated by IFN-γ induced by IL-12. Such macrophages would have potently activated CD4+ T cells, which had been primed by DC/BNL, leading BNL cells neighbouring the activated macrophages to be eradicated through the effects of perforin secreted by activated CD4+ T cells.
Induction of CD4+ CTLs by vaccination with DC/BNL might be associated with the method by which DC/BNL cells were generated. Treatment of a mixture of DCs and BNL cells with 50% PEG, which promotes the fusion of the two cell types, would enable DCs to ingest tumour cell components that could not be ingested via the natural endocytotic pathway. We have already reported that a DC vaccine generated by treatment with PEG induces antibody-dependent suppression of gastrointestinal tumours in adenomatous polyposis coli gene knock-out mice35 and induces macrophage-mediated inhibition of spontaneously developing HCC in old male C3H/HeNCrj mice.36 In both these animal models, CD4+ T cells, primed by DCs loaded with tumour cells produced by treatment with PEG, would play a key role in the induction of antitumour activity, as the DCs promote secretion of IL-4 from CD4+ helper T cells for antibody production35 or secretion of IFN-γ for macrophage activation.36 Indeed, Tatsumi et al.43 have reported that the suppressive effect on BNL-tumour development induced by vaccination with DCs pulsed with a lysate of BNL cells and coadministration of IL-12 is mediated by CD8+ CTLs, but not by CD4+ CTLs. These results suggest that treatment with PEG would promote the ingestion of tumour cell antigens by DCs, which cannot be promoted solely by pulsing with a tumour cell lysate. Some of these antigenic components might contain peptides to fix onto MHC class II molecules on APCs and to be presented to specific CD4+ T cells eliciting tumoricidal activity.
We would like to thank Mrs Hisako Arai for excellent technical help with the fluorescence microscopy. This study was supported (in part) by grants-in-aid for ‘High-Tech Research Centers’ (for the Institute of DNA Medicine) and for ‘Bio-Venture Research Fund Project Aid’ from the Ministry of Education, Culture, Sports, Science and Technology of Japan.