Neuroblastoma-derived Gangliosides Inhibit Dendritic Cell Generation and Function

INTRODUCTION

Two-thirds of NB 3 patients present with advanced disease, and this tumor represents a common cause of cancer-related deaths in children. The long-term survival of patients with advanced stage disease remains poor, despite aggressive therapy such as bone marrow transplantation, intensive surgery, radiotherapy, and chemotherapy (1) .

Interactions between tumor cells and cells of the immune system appear to be critical for tumor growth and progression. Tumor-derived factors dramatically alter the function and survival of immunocompetent cells in the local tumor microenvironment as well as systemically (2 , 3) . One class of immunosuppressive molecules, which may be responsible for the rapid progression of tumor, are gangliosides (4, 5, 6, 7, 8) . Blood levels of gangliosides have been shown to positively correlate with neoplastic growth in patients with diverse tumors. Sialylated lacto-series gangliosides have also been identified as antigens in human epithelial cancers (9) . A number of tumors, including NB, melanoma, retinoblastoma, and hepatoma, are known to overexpress and shed gangliosides into the circulation. For example, analysis of serum obtained from patients with NB showed a 50–100-fold increase in the levels of G_D2 ganglioside (10) . Increased levels of G_D2 and G_M3 were also detected in patients with retinoblastoma and hepatoma, respectively (11 , 12) . G_D3 and G_M3 gangliosides were identified in the sera of patients with malignant melanoma (12) and head and neck carcinoma, respectively (13) .

Evidence suggests that exogenous or tumor-derived gangliosides are able to inhibit function of immune cells and promote tumor growth. For instance, addition of tumor-derived gangliosides to the tumor cell innoculum enhances tumor formation in mice (14) . Gangliosides, predominantly G_D2, isolated from the human NB cell line LAN5, inhibit murine cellular immune responses in vivo (15, 16, 17) . Purified or tumor-derived gangliosides inhibit mitogen-stimulated proliferation of human peripheral blood lymphocytes (18 , 19) , as well as IL-2-dependent growth of murine T cell lines (20) . Gangliosides also inhibit production of TNF-α and antigen presentation by human monocytes (18 , 19 , 21) , natural killer cell cytotoxicity (22) , and IFN-γ-induced immune responses (23) .

Gangliosides have also been shown to contribute to tumor-induced bone marrow suppression and may alter hematopoiesis (24 , 25) . For instance, patients with NB bone marrow metastases often suffer from disturbed hematopoiesis, resulting in pancytopenia (26) . Furthermore, Sietsma et al. (27) have demonstrated that murine NB neuro-2a (N2a) cell-derived gangliosides, mainly G_M3, dose dependently inhibit erythropoiesis and myelopoiesis. Gangliosides isolated from the plasma of NB patients, such as G_D2, G_M2, and G_M3, down-regulated erythroid and myeloid colony formation from human CD34+ precursors in vitro (27) . However, the role of gangliosides in the regulation of DC generation (dendropoiesis) has not been studied. DCs, the most potent professional antigen-presenting cells, play a crucial role in the induction of specific antitumor immune responses (28) , and the inhibition of their function and/or formation represents an important mechanism by which tumors escape immune recognition (29) . Thus, the primary goal of this study was to evaluate the role of NB-derived gangliosides on DC generation and function. We have demonstrated that NB-derived gangliosides, as well as purified gangliosides added to DC cultures, inhibit the generation of functionally active DCs from both human CD34+ and murine hematopoietic precursor cells.

MATERIALS AND METHODS

Mice.

Male C57BL/6 mice, 6–8 weeks of age, were obtained from Taconic (Germantown, NY) and housed in a pathogen-free facility under controlled temperature, humidity, and a 12-h light/dark cycle with food and water available ad libitum.

NB Cell Lines.

Human SK-N-BE and murine Neuro-2a NB cell lines were obtained from American Type Culture Collection and were grown in complete medium:RPMI 1640 (Life Technologies, Inc., Grand Island, NY) supplemented with 10% heat-inactivated FCS, 1 mm sodium pyruvate, 2 mm l-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin (Life Technologies, Inc.).

Preparation of Mouse and Human DCs.

Human DCs were generated from CD34+ precursors isolated from umbilical cord blood using a MACS isolation kit (Miltenyi Biotec, Inc., Auburn, CA). After positive selection, CD34+ progenitors were cultured at 10–20 × 10⁵ cells/ml in 24-well plates in complete medium supplemented with 1000 units/ml GM-CSF, 1000 units/ml IL-4, 2.5 ng/ml TNF-α, and 10 ng/ml SCF. Fresh cytokines were added every 3 days except for SCF, which was added for only 1 week. Cultured human DCs were collected 17–18 days later and used for further analyses. Confirmation of DC phenotype was performed using DC-related markers, including FITC- or PE-labeled antibodies recognizing CD83, CD86, CD80, and HLA-DR molecules (PharMingen, San Diego, CA). Flow cytometric analyses of stained cells was performed on the Becton Dickinson FACScan and LYSIS II software package.

Murine DCs were generated as described previously (30) . Briefly, mouse hematopoietic progenitors were isolated from bone marrow and cultured in complete RPMI 1640 in six-well plates (10⁵ cells/ml) in the presence of 1000 units/ml GM-CSF and 1000 units/ml IL-4. Mouse DCs were harvested on day 7 and stained with CD11c, CD86, CD80, or MHC class II FITC- or PE-conjugated antibodies (PharMingen). Both murine and human GM-CSF and IL-4 were gifts from Schering-Plough (Kenilworth, NJ). TNF-α and SCF were obtained from R&D Systems (Minneapolis, MN).

MLR.

Functional activity of human and mouse DCs was determined in a one-way allogeneic MLR using T lymphocytes as responders. Cultured human or mouse DCs, serving as stimulators, were irradiated at 3000 rads and added to responders (1 × 10⁵ cells/well) at different ratios. Cells were mixed in 96-well round-bottomed plate and incubated for 96 h. Cell cultures were pulsed with 1 μCi of [ 3 H]thymidine/well for the last 16–18 h and harvested using a Skatron (Skatron, Lier, Norway) cell harvester. Uptake of [ 3 H]thymidine was determined on a beta scintillation counter (Beta-Plate, Wallac, Finland).

Experimental Design.

Cell culture inserts with 0.4-μm pore size (Fisher Scientific, Pittsburgh, PA) were used to separate tumor cells and DCs in 6-well or 24-well plates to study the effect of NB-derived immunosuppressive factors on DC generation and function. Human or mouse DC progenitors were cultured in the lower chamber. Murine Neuro-2a (N2a), human SK-N-BE cells, or culture medium, which served as the control, was placed in the upper chamber inserts. Human peripheral blood mononuclear cells and murine splenocytes were used as additional controls. NB cells were seeded into the upper chamber on days 1 or 5 for the mouse DC cultures or on days 1 or 7 for human DC cultures. After coincubation, DCs were harvested and analyzed by FACScan for the expression of DC-specific markers and the percentage of positive cells. The functional ability of DCs to stimulate proliferation of T cells was assessed by using an MLR assay.

Inhibition of Ganglioside Synthesis.

To inhibit ganglioside synthesis, tumor cells were cultured in the presence of 10 μm d-PDMP (Matreya, Pleasant Gap, PA) for 72 h. Expression of G_D2 on treated and nontreated NB cells was detected using primary anti-disialoganglioside (PharMingen) and secondary FITC-conjugated antibodies. To evaluate the effect of purified gangliosides on DCs, G_D2 and G_M3 were dissolved in ethanol and added to DC cultures at a final concentration 5μ m for G_D2 and 25 μm for G_M3. Ethanol was used as a control solvent.

Apoptosis Assay.

To determine whether G_D2 causes apoptosis of human CD34+ progenitor cells, CD34+ cells were treated with 5μ m of G_D2 or ethanol (0.05% vol/vol, control solvent) for 24 and 48 h. The early apoptotic alterations on cell membrane were measured by an Annexin V binding assay (PharMingen), and dead cells were identified using the propidium iodide staining of DNA (Life Technologies, Inc.).

Statistical Analysis.

For a single comparison of two groups, the Student’s t test was used after evaluation for normality. If data distribution was not normal, a Mann-Whitney rank sum test was performed. For all statistical analysis, the level of significance was set at a probability of 0.05 to be considered significant. Data are presented as the mean ± SE.

RESULTS

Inhibition of Dendropoiesis by NB-derived Factors.

Human CD34+ progenitors cultured in the presence of human GM-CSF, SCF, IL-4, and TNF-α for 17–18 days differentiated into DCs, as was assessed by high levels of expression of CD83 (up to 85%), MHC class II (up to 85%), CD86 (up to 95%), and CD80 (up to 70%) molecules. Mouse bone marrow precursors stimulated in cultures with murine GM-CSF and IL-4 were analyzed for the percentage and number of DCs on day 7. Control DC cultures contained 61.1% CD11c+ CD86+ cells and 60.2% CD11c+ MHC class II+ double-positive cells.

Addition of NB cells to DC cultures demonstrated that soluble NB-derived factors inhibited DC generation in both human and mouse models (Figs. 1 ⇓ and 2) ⇓ . For instance, in the presence of N2a cells, the percentage of CD11c+ class II+ cells in mouse bone marrow-derived DC cultures decreased from 60.2 ± 3.9% in control to 24.1 ± 2.2% in NB-treated cultures (P < 0.01). In addition, the percentage of CD11c+ CD86+ cells decreased in these cultures from 61.1 ± 4.9% to 29.3 ± 3.1%, respectively (P < 0.01; Fig. 2 ⇓ ). Inhibition of expression of DC-specific markers was more significant when NB cells were added to DC cultures on day 1 as compared with day 5. Addition of N2a tumor cells to mouse DC cultures on day 5 resulted in a decrease in CD11c+ class II+ and CD11c+ CD86+ cells from 60.2 ± 5.7% to 37.1 ± 3.4% (P < 0.05) and 61.1 ± 4.6% to 45.3 ± 3.4% (P < 0.05), respectively (Fig. 2) ⇓ . However, the inhibition was significantly less than that observed in DC cultures coincubated with NB cells starting on day 1.

Similar data were obtained when human DC cultures were coincubated with human NB cells (Fig. 1) ⇓ . For example, inhibition of DC production from CD34+ precursors was also significantly greater when SK-N-BE cells were added to hematopoietic precursors on day 1 as compared with those obtained on day 5 (data not shown; P < 0.05). As shown in Fig. 1 ⇓ , SK-N-BE-derived factors caused significant decreases in the appearance of CD83+, CD86+, CD80+ cells in these cultures (P < 0.01). However, expression of MHC class II molecules was not significantly changed in the presence of NB cells (data not shown). Interestingly, NB-derived factors caused alterations of DC differentiation but did not change the number of cells in cultures. Together, these results demonstrate that murine NB cells (N2a) and human NB cells (SK-N-BE) significantly inhibited murine and human DC generation in cultures.

Inhibition of APC Function in DC Cultures by Coincubation with NB.

To determine whether NB-derived factors affect the functional activity of DCs to induce T-cell proliferation, we evaluated the stimulatory capacity of cells harvested from control or NB-treated DC cultures for their ability to induce proliferation of allogeneic T cells in an MLR assay. As shown on Fig. 3 ⇓ , coincubation of DC cultures with either SK-N-BE or N2a cells on day 1 significantly decreased their ability to induce T-cell proliferation. In human studies, the maximum inhibition of T-cell proliferation was observed at APC:T cell ratios 0.9:1 and 0.3:1:33,508 ± 1,613 cpm for control versus 17,428 ± 152 cpm for SK-N-BE treated cultures (P < 0.01) and 32,714 ± 5,785 versus 15,967 ± 1,888 cpm (P < 0.01), respectively (Fig. 3A) ⇓ . Similarly, the level of suppression of murine T-cell proliferation induced by pretreatment of DC cultures with N2a tumor cells reached 30% (P < 0.01; Fig. 3B ⇓ ). It is important to note that when the coincubation of DC cultures with NB cells was initiated at day 5 or later, a decrease in DC numbers and their capacity to stimulate allogeneic T cells was also detected but to a significantly lesser extent in both murine and human systems (data not shown). This suggests that NB-derived factor(s) affect the development of early hemopoietic progenitors. These results confirm that human and murine NB produce factors that inhibit the generation of DCs in vitro, as assessed by a significant decrease in the APC activity of cultured cells. The level of inhibition was time dependent, with the most profound inhibition observed when tumor cells were added at an early time to the culture of hemopoietic progenitors.

Purified Gangliosides Inhibit Dendropoiesis.

Human SK-N-BE cells predominantly express and shed G_D2 ganglioside, whereas murine N2a cells express and shed high amounts of G_M3 ganglioside (27) . G_D2 and G_M3 gangliosides cause a dose-dependent inhibition (up to 70%) of erythroid and myeloid colony formation up to 70% (27) . We hypothesized that the inhibition of dendropoiesis by NB cells was mediated by gangliosides.

To test this hypothesis, we examined whether purified G_D2 (5 μm) or G_M3 (25 μm) affect differentiation of human CD34+ or mouse bone marrow progenitors into mature DCs. The doses of gangliosides were chosen according to earlier published data (27) . G_D2 and G_M3 were added to DC cultures on day 1 and then every third day. The number of DCs differentiated from CD34+ precursors in the G_D2-treated groups was 3-fold less when compared with DC number in control culture. As we have mentioned above, SK-N-BE cells express G_D2 and altered DC generation but not the number of DCs in cultures. The absence of the effect of SK-N-BE cells on DC proliferation is likely attributable to the influence of other SK-N-BE-derived factors. Expression of CD83 molecules on cultured cells decreased from 51.8 ± 6.1% in control groups to 12.9 ± 2.7% in cultures treated with G_D2 (P < 0.05). Similarly, the percentage of CD86+ cells in the same cultures decreased from 32.4 ± 2.5% to 7.4 ± 1.8% after addition of G_D2 (P < 0.01; Fig. 4 ⇓ ). The percentages of double-positive CD83+ CD80+ and CD83+ HLA-DR+ cells were three times less in G_D2-treated cultures in comparison with nontreated cultures (P < 0.05). Because we have observed that G_D2 decreases the number of DCs differentiated from CD34+ progenitors, we hypothesized that this effect might be mediated by the induction of apoptosis. To check this possibility, CD34+ precursor cells were treated with G_D2 (5 μm) or control solvent (0.05% ethanol) for 24 and 48 h, collected and analyzed by Annexin V-FITC binding. The levels of Annexin V-positive, propidium iodide-negative cells in the treated and untreated cultures were 15.9 ± 3.1% and 10.9 ± 2.7%, respectively (P > 0.05). These data suggest that G_D2-induced inhibition of dendropoiesis was not mediated by apoptotic death of early hematopoietic progenitors.

We have demonstrated that treatment of bone marrow cultures with G_M3 did not significantly change the total number of cells at the end of culture, although G_D2 markedly inhibited the yield of cultured DCs. The differences between the effect of G_M3 and G_D2 on DC proliferation may be explained by the different structure of carbohydrate tails of these gangliosides. The carbohydrate structure of gangliosides is known to be involved in the regulation of cell growth and differentiation (9 , 31) . However, G_M3 induced down-regulation of CD11c expression by 15%. For instance, 67.8 ± 3.8% of cells expressed CD11c molecules in control DC cultures, whereas 57.5 ± 1.5% of cells were CD11c+ in cultures treated with G_M3. The expression of CD86 and CD80 was also inhibited up to 53 and 17%, respectively, in DC cultures generated in the presence of G_M3 (Fig. 5A) ⇓ . Furthermore, APC function of DC cultures treated with G_M3 was also inhibited. The level of suppression of murine T-cell proliferation induced by pretreatment of DC cultures with G_M3 reached 38% (P < 0.05) at an APC:T cell ratio of 1:1 (Fig. 5B) ⇓ . Taken together, these results demonstrate that purified gangliosides G_D2 and G_M3 inhibit the generation of human and murine DCs in vitro.

NB-derived Gangliosides Inhibit DC Generation.

To examine whether NB-derived gangliosides are involved in the inhibition of DC generation, N2a and SK-N-BE cells were treated with 10μ m d-PDMP, a ganglioside synthesis inhibitor. The addition of d-PDMP to NB cell cultures for 3 days caused a significant reduction of ganglioside expression on tumor cells by up to 50% as compared with d-PDMP-untreated NB cells (Fig. 6) ⇓ , which is in agreement with others (16 , 27) . d-PDMP-treated and nontreated cells were added to DC cultures. Control DC cultures and DC cultures generated in the presence of d-PDMP-treated (PDMP+) or nontreated (PDMP−) NB cells were evaluated for their capacity to induce proliferation of allogeneic T cells in an MLR assay. Fig. 7 ⇓ represents the results of these experiments. Cultured APCs generated in the presence of PDMP-/SK-N-BE cells had a significantly decreased ability to induce T-cell proliferation compared with APCs generated with PDMP+/SK-N-BE (P < 0.05; Fig. 7 ⇓ ), suggesting that block of ganglioside synthesis in NB cells decreases their ability to inhibit generation of functionally active DCs. For instance, at an APC:T cell ratio of 0.3:1, APCs exposed to PDMP−/SK-N-BE cells had only 41 ± 4% of stimulatory activity of control APCs, whereas APCs exposed to PDMP-treated SK-N-BE cells demonstrated 69 ± 5% of control activity (P < 0.05). These data demonstrate that PDMP treatment decreases NB-mediated inhibition of APC activity up to 65%. The partial recovery of APC function after PDMP treatment of NB cells is in a good correlation with a partial inhibition of ganglioside levels observed in PDMP-treated tumor cells. The similar results were observed when murine NB cells were treated with d-PDMP. For example, N2a treatment of murine DC cultures caused inhibition of T-cell proliferation (1:1 APC:T cell ratio) from 48,320 ± 3,751 cpm for control APCs to 11,406 ± 983 cpm for APCs obtained from N2a-treated cultures (P < 0.05). Inhibition of ganglioside synthesis with PDMP treatment resulted in the partial (up to 23%) reversal of NB-induced inhibition of DC stimulatory activity.

Taken together, our data suggest that gangliosides derived from N2a or SK-N-BE cells are responsible, at least in part, for the NB-induced decrease in the number of DCs generated in cultures. Blockage of ganglioside synthesis with d-PDMP partly prevents the NB-induced inhibition of dendropoiesis in vitro.

DISCUSSION

Human SK-N-BE and mouse N2a NB cells produce soluble factors that inhibit DC generation from human CD34+ and murine bone marrow progenitor cells. As assessed by flow cytometry, expression of CD83 and CD86 on cultured cells was completely abrogated, and the expression of CD80 was strongly inhibited in the presence of SK-N-BE cells. Similarly, murine N2a tumor cells inhibited the production of mouse DCs, as was determined by a significant down-regulation of CD11c, CD86, and MHC class II expression on cells in the treated cultures. The ability of APCs, presumably DCs, harvested from NB-treated cultures to trigger proliferation of naive allogeneic T cells was also significantly decreased. The inability of DCs generated in the presence of NB cells to cause T-cell proliferation in an MLR assay could be explained by at least two different mechanisms: (a) direct inhibition of functional activity of APC; or (b) decrease in APC number. It is unlikely that NB-derived factors had a strong inhibitory effect on the APC function of DCs because we demonstrated here that both phenotypic and functional alteration in APCs induced by NB-derived factors were more profound when tumor cells were added to DC cultures at earlier stages. This suggests that NB-released factors primarily affect functional maturation of early hemopoietic precursors, thereby inhibiting DC generation rather than DC function. This does not rule out the possibility of partial inhibition of DC function mediated by NB-derived factors. Evaluation of this notion is in progress.

A common feature in patients with NB and bone marrow metastases is the occurrence of disturbed hematopoiesis. (28 , 32) . One possible group of molecules potentially responsible for NB-induced bone marrow suppression are gangliosides (32) . It has been reported recently that both gangliosides shed by NB cells and exogenous gangliosides inhibited myelopoiesis and erythropoiesis in vitro (28) . Here we demonstrated that gangliosides inhibit human and mouse DC generation. The major ganglioside expressed by human SK-N-BE cells and murine N2a cells is G_D2 and G_M3, respectively. To determine whether inhibitory effects of NB-derived factors on dendropoiesis is mediated by gangliosides, the biosynthesis of gangliosides in tumor cells was inhibited by d-PDMP (16 , 17) . d-PDMP treatment of SK-N-BE cells for 72 h caused inhibition of G_D2 expression up to 50% as compared with d-PDMP-untreated NB cells. Blockage of ganglioside synthesis partly reversed the NB-induced inhibition of DC generation. Moreover, purified G_D2 and C_M3 suppressed differentiation of CD34+ cells and bone marrow precursor cells into functionally active DCs, respectively. Thus, purified gangliosides and NB-derived gangliosides are potent inhibitors of DC generation.

Gangliosides are also associated with inhibition of the activity of several other immune cells including helper T cells (33 , 34) , natural killer cells (35) , and antigen- and mitogen-stimulated T and B cells (36, 37, 38) . Ladisch et al. (19) first demonstrated that gangliosides may inhibit the accessory function of monocytes in 1984. Brain-derived gangliosides regulate the cytokine production and proliferation of activated T cells. Gangliosides suppress TNF-α production in human monocytes by acting on early stages of activation (> 21 ). They block antigen presentation by human monocytes (39) and cause down-regulation of constitutive and IFN-γ-inducible expression of MHC class I and II molecules on astrocytes (40) . The inhibitory effects of tumor-derived gangliosides on cells of the immune system have also been well demonstrated in vivo. For instance, gangliosides, isolated from the plasma of cancer patients are highly immunosuppressive and regulate a variety of immune functions in vivo (15 , 16 , 41 , 42) .

Immunosuppression caused by gangliosides may be mediated by several mechanisms. These include direct binding to cytokines such as IL-2, preventing its interaction with IL-2 receptor (20 , 23 , 43) . By a similar mechanism, the preincubation of CSFs with bovine brain mixed gangliosides resulted in decreased numbers of colonies of bone marrow-derived granulocyte/macrophage progenitor cells (44) . Direct binding of gangliosides to CSFs may also be one of the means by which NB cells inhibit dendropoiesis. In fact, competitive binding of gangliosides with growth factors may explain why transduction of C1300 murine NB cells with genes encoding IL-2 or GM-CSF results in abrogation of their tumorigenicity and decreased growth of liver metastasis (45) .

In addition, gangliosides may physiologically function to coordinate the activation of multiple receptors. They modify the binding activity of individual receptors as well as receptor-specific signal transduction pathways (46) . For instance, programmed cell death might be mediated in part by ganglioside derivatives because ceramide is a constitutive component of their structure and may be released during the interaction of tumor and immune cells. Recently, it has been demonstrated that mouse bone marrow cells treated in vitro with gangliosides, derived from T-cell lymphoma, undergo apoptotic cell death. The major ganglioside produced by this tumor is G_D3. In fact, the anti-G_D3 antibody protected bone marrow cells from tumor-induced apoptosis (47) . However, we have shown here that G_D2-induced inhibition of dendropoiesis in cultures is not mediated by apoptosis of hematopoietic precursors. It is known that the ceramide portion of gangliosides contains a fatty acid chain. The length of this chain determines the killing ability of gangliosides (48) . Thus, G_D2 does not cause the apoptosis of hematopoietic precursor cells but probably has an effect on DC generation by changing the binding of growth factors with their receptors. Ceramide itself, as an important constituent of gangliosides, blocks phosphorylation of the retinoblastoma gene product, which is an important regulator of cell cycle. Inhibition of phosphorylation of retinoblastoma results in cell cycle arrest (49) . In contrary, purified ceramide, but not ganglioside, has been shown to induce apoptosis in a model system (50) . Gangliosides have also been shown to block the activation of nuclear factor-κB in T cells, a transcriptional factor involved in the expression of several cytokine genes (23 , 51) . We have also observed NB-induced inhibition of NF-κB expression in cultured DCs. 4 This suggests that NB-derived gangliosides may not only inhibit binding of growth factors with receptors on the surface of hemopoietic precursors but also have a direct effect on their survival. We have demonstrated recently that tumor-derived factors cause inhibition of DC survival and stimulate their apoptotic death (52 , 53) . However, it is unclear whether gangliosides may induce apoptosis in hemopoietic progenitors. Further studies are necessary to clarify this mechanism.

Tumor-associated gangliosides can also be considered as antigens specific for several types of tumor including NB, melanoma, and lung cancer. Gangliosides thus could be used as a target in different immunotherapeutic approaches (31 , 54 , 55) . Further understanding and elucidating of the mechanisms of immunosuppression induced by gangliosides may lead to the discovery of a new therapeutic agent that will be directed to decrease tumorigenicity and increase immunogenicity of NB cells.