SEARCH

SEARCH BY CITATION

Keywords:

  • galectin-1;
  • glioblastoma;
  • tumor-infiltrating myeloid cells;
  • dendritic cell immunotherapy;
  • angiogenesis

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Galectin-1 is a glycan-binding protein, which is involved in the aggressiveness of glioblastoma (GBM) in part by stimulating angiogenesis. In different cancer models, galectin-1 has also been demonstrated to play a pivotal role in tumor-mediated immune evasion especially by modulating cells of the adaptive immune system. It is yet unknown whether the absence or presence of galectin-1 within the glioma microenvironment also causes qualitative or quantitative differences in innate and/or adaptive antitumor immune responses. All experiments were performed in the orthotopic GL261 mouse high-grade glioma model. Stable galectin-1 knockdown was achieved via transduction of parental GL261 tumor cells with a lentiviral vector encoding a galectin-1-targeting miRNA. We demonstrated that the absence of tumor-derived but not of host-derived galectin-1 significantly prolonged the survival of glioma-bearing mice as such and in combination with dendritic cell (DC)-based immunotherapy. Both flow cytometric and pathological analysis revealed that the silencing of glioma-derived galectin-1 significantly decreased the amount of brain-infiltrating macrophages and myeloid-derived suppressor cells (MDSC) in tumor-bearing mice. Additionally, we revealed a pro-angiogenic role for galectin-1 within the glioma microenvironment. The data provided in this study reveal a pivotal role for glioma-derived galectin-1 in the regulation of myeloid cell accumulation within the glioma microenvironment, the most abundant immune cell population in high-grade gliomas. Furthermore, the prolonged survival observed in untreated and DC-vaccinated glioma-bearing mice upon the silencing of tumor-derived galectin-1 strongly suggest that the in vivo targeting of tumor-derived galectin-1 might offer a promising and realistic adjuvant treatment modality in patients diagnosed with GBM.

Abbreviations
DC

dendritic cell

DCm

mature dendritic cells

GBM

glioblastoma

PTI

post-tumor inoculation

MDSC

myeloid-derived suppressor cells

TAMs

tumor-associated macrophages

Glioblastoma (GBM) is the most frequent and malignant human brain tumor, accounting for ∼50% of all primary brain tumor cases in adults.[1] Despite the availability of multimodal treatments, including maximal, safe neurosurgical resection and chemoradiotherapy, the prognosis of GBM remains dismal with a median survival expectancy of ∼15 months and a mortality rate of 90% within 3 years.[2] Hence, these patients are in high need of new, efficient treatment strategies that improve their quality of life and clinical outcome. During the last decade, significant advances have been made in the development of immunotherapeutic strategies to combat primary intracranial tumors. Our research group[3, 4] and others[5] have demonstrated that DC-based immunotherapy is a promising new treatment strategy in the fight against GBM. However, the extent of regression of GBM that is induced by tumor vaccination remains limited. The absence of robust tumor regression can be partially explained by the abundance of immune modulating mechanisms used by tumor cells to dampen antitumor immune responses.[6] These findings suggest that additional interference with these local immunosuppressive mechanisms will be required to facilitate the development of more potent antiglioma immune responses.

In this respect, galectin-1 could potentially offer an interesting target. Galectin-1 is a natural immunosuppressive glycan-binding protein whose expression is upregulated in several types of cancer, including GBM. In a variety of cancer models, the presence of galectin-1 within the tumor microenvironment has been shown to contribute significantly to the establishment of local immune resistance.[7] Through interactions with β-galactoside-expressing glycoproteins on the T cell surface, galectin-1 can negatively regulate T cell survival,[8, 9] antagonize effector T cell signaling,[10] block pro-inflammatory cytokine secretion[11] and suppress transendothelial migration.[12] Furthermore, galectin-1 can blunt T cell responses by promoting the accumulation and expansion of regulatory T cells (Tregs).[13] The first link between galectin-1 expression at tumor sites and T cell-mediated tumor rejection was reported by Rubinstein et al., who demonstrated that the targeted inhibition of galectin-1 gene expression in mouse melanoma cells resulted in increased T cell-mediated tumor rejection.[14] In addition, animal studies have demonstrated that the metabolic inhibition of galectin-1-binding carbohydrates can delay tumor growth by enhancing tumor lymphocyte infiltration, increasing IFN-ɣ production while lowering IL-10 production.[15] Moreover, in several human malignancies, an inverse correlation was demonstrated between lymphocyte infiltration and galectin-1 expression.[16] Thus far, most studies that have reported an immunosuppressive role for galectin-1 in the tumor microenvironment have revealed a modulating effect of galectin-1 on cells of the adaptive immune system. It is yet unknown whether tumor-derived galectin-1 can also induce quantitative or qualitative differences in tumor-infiltrating myeloid cells.

A number of different studies have already demonstrated a pivotal role for tumor-derived galectin-1 in the aggressiveness of GBM.[17] Galectin-1 has been shown to enhance glioma cell migration, stimulate angiogenesis and to promote the development of chemotherapy and radiotherapy resistance. The importance of galectin-1 in glioma growth and progression is further underscored by the observation that intratumoral galectin-1 expression levels are positively correlated with the grade of malignancy and with worse prognosis.[18] Although the abundance of galectin-1 in malignant glioma is well-known[17] and several mechanisms of galectin-1-mediated immune modulation have been described in different types of cancer,[7] the role of galectin-1 in glioma-mediated immune evasion is still elusive.[19] The primary goal of this study was to examine whether the absence or presence of galectin-1 in the glioma microenvironment results in quantitative or qualitative differences in innate and/or adaptive antitumor immune responses. Moreover, we explored the potency of prophylactic DC-based immunotherapy to protect mice against primary tumor inoculation in the presence or absence of tumor- or host-derived galectin-1. To address these questions, we used the orthotopic GL261 mouse glioma model, which is the most abundantly used immune-competent model to study the potency of immunotherapy approaches against intracranial high-grade glioma.[20]

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Mice

Eight-to-ten week-old female C57BL/6J mice were purchased from Harlan (Horst, The Netherlands). Galectin-1-/- C57BL/6 mice were developed in the laboratory of Françoise Poirier[21] and were bred at KUL animal facilities for the experiments in this study. NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). The mice were maintained under conventional pathogen-free conditions. All of the mouse experiments were approved by the bioethics committee of the KU Leuven, which follows international guidelines.

Tumor cell lines

Methylcholanthrene-induced murine C57BL/6J syngeneic GL261 glioma cells were kindly provided by Dr. Eyüpoglu (University of Erlangen, Germany) and were cultured as previously described.[20, 22] Cell morphology was evaluated by microscopic examination three times per week.

The generation of galectin-1 knockdown GL261 clones

Parental GL261 tumor cells (GL261-WT) were transduced with an HIV-based vector encoding the galectin-1-targeting, miRNA-based, short-hairpin RNAs with a blasticidin resistance cassette (Supporting Information Fig. S1). For the design of this construct, we made use of the antimurine-galectin-1 siRNA sequence of which specificity and efficacy was demonstrated.[23] The stable polyclonal knockdown cells were selected using blasticidin (3 µg/ml; Invivogen, Toulouse, France) and 9 monoclonal galectin-1 knockdown cell lines were generated. Selection of the best clone (GL261-KD) was based on the level of galectin-1 expression (determined by real time RT-PCR and Western Blot analysis) and on the in vitro growth rate (determined by the colorimetric MTT assay). In addition, GL261-WT and GL261-KD tumor cells were analyzed for their expression levels of H-2Kb, I-A/I-E, CD80, CD86, CD40, CTLA-4, FasL and PD-L1 by means of flow cytometry. The GL261 tumor cells transduced with a HIV-based vector encoding an irrelevant, miRNA-based, short-hairpin RNA were used as a non-specific control (GL261-Mock). The detailed procedure is described in the Supporting Information Material and Methods section.

RT-PCR and Western blot analysis

The quantification of mouse GAL-1, CCL2 and VEGF by RT-PCR was performed using an ABI prism 7700 Sequence Detector (Applied Biosystems, Foster). Western blotting was performed as previously described.[23] For details, see the Supporting Information Material and Methods section.

In vitro proliferation rate

In vitro proliferation rates were measured using a colorimetric MTT assay. Briefly, cells were seeded and after 24, 48 or 72 hr, the culture medium was replaced with 100 μl of a 0.5 mg/ml 3-(4,5-dimetylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich, Haverhill, England) solution. After 2 h of incubation at room temperature (RT), the MTT solution was replaced with 100 μl of pure dimethylsulfoxide (DMSO, Merck, Darmstadt, Germany). The optical density was measured at 570 and 620 nm using an ELISA reader (Thermo Labsystems, Franklin, MA). The OD570–620nm value was used as a measurement of cell growth.

The generation and maturation of DCs

Dendritic cells (DCs) were generated from bone marrow progenitor cells as previously described.[22] Recombinant murine granulocyte/macrophage colony-stimulating factor (GM-CSF) was kindly provided by Prof. Dr. Kris Thielemans (University of Brussels, Belgium). Briefly, bone marrow progenitor cells were cultured in the presence of GM-CSF (20 ng/ml) for 7 days. On day 7, immature DCs (DCi) were harvested and co-incubated with lysates derived from GL261-WT or GL261-KD tumor cells. Maturation was initiated by adding GM-CSF and 1 μg/ml E. coli LPS (Sigma-Aldrich). After 24 h, DCm were harvested and resuspended in phosphate-buffered saline (DPBS, Lonza, Verviers, Belgium) at a concentration of 1 million DCs per 100 μl. For each vaccination, flow cytometric quality control of the DCs was performed by staining the DCs for H-2Kb, I-A/I-E, CD80, CD86, CD40 and CD11c.

Orthotopic glioma inoculation

The mice were intracranially injected with GL261-WT, GL261-Mock or GL261-KD tumor cells as previously described.[22] Briefly, 0.5 × 106 tumor cells were injected at 2 mm lateral and 2 mm posterior from the bregma at a depth of 3 mm below the dura mater by using a stereotactic frame (Kopf Instruments, Tujunga, CA). Stereotactic inoculation was performed under sterile conditions. The mice were considered long-term survivors when their survival exceeded 3-fold of the median survival of the untreated animals. In the cases of vaccination, the mice received 2 i.p. injections with 1 × 106 DCs at day 14 and 7 prior to the tumor inoculation.

The isolation of brain-infiltrating immune cells

Brain-infiltrating immune cells were isolated from tumor-inoculated mice using protocols established previously in our laboratory.[22] The brain-infiltrating immune cells were analyzed for CD11b, Ly6C, Ly6G, CD3, CD4 and CD8 expression using flow cytometry. For each staining, the appropriate isotypes were used. The intracellular detection of FoxP3 was performed using a FoxP3 staining kit (eBioscience, San Diego, CA) according to the manufacturer's protocols. For intracellular IFN-ɣ staining, cells were stimulated for 4 h with 100 ng/ml phorbol myristate acetate, 1 μg/ml ionomycin and 0.7 μg/ml monensin. The cells were fixed, permeabilized and stained with anti-CD4-PE, anti-CD45-PerCP and and anti-IFN-ɣ-FITC. Analysis was performed using the Cellquest software on a FACSort flow cytometer (BD Pharmingen, Erembodegem, Belgium). For analysis of the mRNA expression levels in CD11b+ cells, cells were separated based on their expression of CD11b using CD11b Microbeads according to manufacturer's protocols (Miltenyi Biotec, Bergisch Glaadbach, Germany). For analysis of the mRNA expression levels in CD45high cells, brain-infiltrating immune cells were sorted with the FACSAria (Becton Dickinson, San Jose, CA) by gating on CD45high expression.

Determination of VEGF production by ELISA

GL261-WT or GL261-KD tumor cells were seeded in a 24-well plate and supernatants was collected after 24 h. VEGF production was measured by means of ELISA according to the manufacturer's guidelines (R&D systems, Oxon, United Kingdom).

In vivo depletion of CD8a+ T cells

The in vivo depletion of the CD8a+ cells was performed by 2 IP injections of the YTS169 anti-CD8 mAb (200 μg at day -1 and 100 μg at day 1). Polyclonal rat IgG (Rockland, Gilbertsville) was used as a control. The depletion efficacy exceeded 90% as monitored in the cervical lymph nodes at day 7 post-tumor-inoculation (PTI) using flow cytometry.

Quantification of blood vessels

The animals were sacrificed at day 14 PTI. The brains were prelevated, fixed in buffered formalin and embedded in paraffin. Serial coronal sections (10-μm-thick) were prepared and mounted onto glass slides. Brain sections were rehydrated, incubated in trypsin buffer (0.05 mg/ml) for 15 min at 37°C, permeabilized (methanol, 3% H2O2) and blocked in TNB buffer (0.1M Tris pH 7.4; NaCl 150 mM; 0.5% blocking reagent Perkin Elmer, Boston) for 30 min at RT. Tissues were incubated with an anti-mouse-CD31 (BD Pharmingen) diluted in TNB overnight at 4°C, washed in TNT (0.1M Tris pH 7.4; NaCl 150mM; 0.2% Triton X-100) and incubated with a biotin coupled anti-rat secondary antibody (Vector Labs, Burlingame) for 2 h at RT. TSA enhancer kit was used according to manufacturer procedure (Perkin Elmer). Staining was revealed with DAB (Dako, Carpinteria). Sections were counterstained with hematoxylin-eosin. Pictures were taken at 63× magnification. Semi-automatic quantification of the vascular density was performed with ImageJ Image Calculator plugin (ImageJ).

Immunohistochemistry for galectin-1 and F4/80

End-stage mice inoculated with GL261-WT or GL261-KD tumor cells were sacrificed, brains were prelevated, fixed in buffered formalin and embedded in paraffin. Galectin-1 was stained by using the commercially available polyclonal rabbit-anti-human galectin-1 antibody (1/750 high pretreatment; 500-P210, Peprotech, London, UK). Macrophages/microglia were stained with the rat-anti-mouse F4/80 antibody (ab6640, Abcam, Cambridge, UK). Pictures were taken at 100× magnification using a Leica DFC290 camera. For the semi-quantitative analysis of the number of F4/80+ macrophages and the galectin-1 intensity, 10 pictures were taken for each tumor at 40× magnification. Semi-automatic quantification of F4/80 positive macrophages was realized with ImageJ Analyse Particules plugin and semi-automatic quantification of the galectin-1 staining intensity was performed with ImageJ Image Calculator plugin (ImageJ).

Statistics

All of the data were analyzed using GraphPad Prism 5 (San Diego, CA) and are represented as the mean value ± SD. The survival analysis was performed using the log-rank test. For comparing multiple groups, one-way analysis of variance (ANOVA) was used. For comparison of two groups, Student's t-test was performed.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Silencing of glioma-derived galectin-1 delays in vivo tumor progression

To validate whether glioma-derived galectin-1 contributes to the aggressiveness of glioma in the murine GL261 malignant glioma model, we investigated the impact of its downregulation on in vivo glioma progression. Among the different galectin-1 knockdown clones that were generated, we selected one monoclonal cell line—subsequently named GL261-KD—in which a strong decrease in galectin-1 protein (Fig. 1a) and mRNA (Fig. 1b, p < 0.001) expression levels was observed. All of the mice injected with GL261-WT or GL261-mock tumor cells developed progressively growing tumors (Fig. 1c). Inoculation of the mice with GL261-KD tumor cells significantly improved tumor-bearing mice survival compared to inoculation with GL261-WT (p < 0.0001) or GL261-mock tumor cells (p < 0.0001). To exclude the possibility that the stable transduction of the cells with a galectin-1-targeting miRNA directly affects glioma cell proliferation, we compared in vitro proliferation rates. All three of the cell lines exhibited a similar in vitro growth rate, indicating that the GL261-WT or GL261-mock tumor cells did not have any growth advantage over the GL261-KD tumor cells (Fig. 1d). In addition, inoculation of the mice with either the polyclonal galectin-1 knockdown tumor cell line or another monoclonal galectin-1 knockdown tumor cell line also resulted in a significant prolonged survival compared to inoculation with GL261-WT tumor cells (Supporting Information Fig. S2). These findings suggest that the prolonged survival of GL261-KD tumor-inoculated mice cannot be explained by clonal variation.

image

Figure 1. Silencing of galectin-1 in GL261 tumor cells prolongs the survival of glioma-bearing mice. (a) Western blot analysis of the galectin-1 expression levels in GL261-WT, GL261-Mock and GL261-KD tumor cells. (b) Galectin-1 mRNA expression levels in GL261-WT and GL261-KD tumor cells as measured by RT-PCR. The data are expressed as the mean values ± SD. ***, p < 0.001 (c) Kaplan-Meier analysis of the mice grafted with 0.5 × 106 GL261-WT (black dots), GL261-Mock (black triangles) or GL261-KD (open dots) tumor cells (n = 10 for each group). ***, p < 0.0001 (d) The in vitro proliferation rate of GL261-WT, GL261-Mock and GL261-KD tumor cells over time as measured by the colorimetric MTT assay. The data are represented as the mean optical density (OD) from one experiment performed in quintuplicate.

Download figure to PowerPoint

Silencing of glioma-derived galectin-1 decreases the accumulation of myeloid cells in the brain of tumor-bearing mice

Next, we questioned whether the silencing of glioma-derived galectin-1 induced proportional changes in the brain-infiltrating myeloid cell population of tumor-inoculated mice. To address this question, brain-infiltrating immune cells of GL261-WT or GL261-KD tumor-inoculated mice were isolated at day 14 PTI and were analyzed for the expression of macrophage or MDSC markers by means of flow cytometry (Fig. 2). Analysis of the percentage of brain-infiltrating myeloid cells (CD11bhigh) amongst the total population of brain-infiltrating CD45high immune cells revealed a significantly reduced influx of myeloid cells in the absence of glioma-derived galectin-1 (p < 0.001). A more detailed characterization of these brain-infiltrating myeloid cells revealed a significant decrease in the percentage of tumor-associated macrophages (TAMs; CD11bhigh F4/80+) in GL261-KD tumor-inoculated mice (p < 0.01). Moreover, in GL261-WT tumor-bearing mice, a significantly increased percentage of brain-infiltrating mononuclear (CD11bhigh Ly6Chigh; p < 0.001) and granulocytic (CD11bhigh Ly6Ghigh; p < 0.001) MDSC was observed. Interestingly, the percentage of brain-infiltrating myeloid cells amongst the total population of brain-infiltrating CD45high immune cells remained low during tumor progression in GL261-KD tumor-bearing mice. These findings suggest that the significant reduced percentage of myeloid cell infiltration in GL261-KD tumor-inoculated mice observed at day 14 PTI cannot be explained by the differences in tumor load at that time point (data not shown).

image

Figure 2. Silencing of glioma-derived galectin-1 reduces the accumulation of myeloid cells in the brain of tumor-inoculated mice. Flow cytometric analysis of the brain-infiltrating myeloid cells that were isolated from mice grafted with GL261-WT of GL261-KD tumor cells. Each point represents stained brain-infiltrating immune cells isolated from one mouse at day 14 PTI. All of the cells were gated on CD45high expression, thereby excluding CD45intCD11b+ microglia. ***, p < 0.001; **, p < 0.01.

Download figure to PowerPoint

In addition, we semi-quantified the number of F4/80+ macrophages/microglia on histological sections of end-stage GL261-WT or GL261-KD tumor-bearing mice. Semi-quantitative analysis of the intratumoral galectin-1 expression levels revealed that the expression level of galectin-1 was significantly lower in tumors of GL261-KD tumor-bearing mice as compared with tumors of GL261-WT tumor-bearing mice (Figs. 3a and 3b; p < 0.001). In addition, semi-quantitative analysis of the number of tumor-infiltrating macrophages/microglia revealed that tumors of GL261-WT tumor-bearing mice contained significant higher numbers of macrophages/microglia as compared to tumors of GL261-KD tumor-bearing mice (Figs. 3a and 3c; p < 0.001). Interestingly, we observed a similar link between the intratumoral galectin-1 expression levels and macrophage/microglia infiltration in biopsies of patients with newly diagnosed GBM (Supporting Information Fig. S3).

image

Figure 3. Galectin-1 expression and macrophage/microglia (F4/80) infiltration in tumors of end-stage GL261-WT or GL261-KD tumor-inoculated mice. (a) Immunohistochemical analysis of the galectin-1 protein expression level and macrophage/microglia infiltration in tumors of end-stage GL261-WT (top) or GL261-KD (bottom) tumor-bearing mice. At least two mice/group were analyzed. Representative areas are shown (100× magnification). (b,c) Semiquantitative analysis of the galectin-1 intensity and macrophage/microglial cell number. Data are expressed as the mean values ± SD. ***, p < 0.001.

Download figure to PowerPoint

Silencing of glioma-derived galectin-1 alters CCL2 and VEGF mRNA expression levels in cultured glioma cells and in brain-infiltrating immune cells isolated from tumor-inoculated mice

In gliomas, CC chemokine ligand 2 (CCL2) and vascular endothelial growth factor (VEGF) have been described as important chemoattractants for both macrophages and MDSC towards the tumor microenvironment.[24] Therefore, we performed RT-PCR on in vitro cultured GL261-WT and GL261-KD tumor cells to analyze the possible differences in CCL2 and VEGF expression levels (Figs. 4a and 4b). We observed a significant decrease in CCL2 (p < 0.01) and VEGF (p < 0.01) mRNA expression levels upon intratumoral silencing of galectin-1. Moreover, we compared the levels of CCL2 and VEGF secretion between in vitro cultivated GL261-WT and GL261-KD tumor cells. GL261-KD cells secreted significantly lower amounts of VEGF as compared to GL261-WT tumor cells (Fig. 4c). CCL2 could not be detected in the supernatants of both tumor cell lines (data not shown). As tumor cells are not the only source of VEGF and CCL2 within the tumor microenvironment, we analyzed whether differences in VEGF and CCL2 mRNA expression levels could also be found within the CD45high immune cells that were isolated from the brain of GL261-WT or GL261-KD tumor-inoculated mice (Figs. 4d and 4e). We observed a strong decrease in both CCL2 and VEGF expression at day 14 PTI in the CD45high immune cells isolated from the brain of the GL261-KD tumor-inoculated mice. A possible explanation for these decreased CCL2 and VEGF mRNA expression levels in isolated CD45high cells could be the significantly lower proportion of CD11b+ cells present within the brain-infiltrating CD45high population of GL261-KD tumor-inoculated mice. Therefore, we also analyzed CCL2 and VEGF mRNA expression levels in CD11b+ cells isolated from the brain of GL261-WT or GL261-KD tumor-inoculated mice (Figs. 4d and 4e). Analysis of the CCL2 mRNA expression levels in brain-infiltrating CD11b+ cells revealed a strong decrease in the mRNA expression levels of CCL2 in CD11b+ cells that were isolated from GL261-KD tumor-bearing mice. In contrast, VEGF mRNA expression levels were increased in CD11b+ cells isolated from GL261-KD tumor-bearing mice.

image

Figure 4. Silencing of glioma-derived galectin-1 alters both CCL2 and VEGF mRNA expression levels in glioma cells and in brain-infiltrating immune cells isolated from tumor-inoculated mice. (a,b) The expression of CCL2 and VEGF mRNA in in vitro cultured GL261-WT and GL261-KD tumor cells as measured by RT-PCR. The data are expressed as the mean values ± SD. **, p < 0.01. (C) Analysis of VEGF protein levels in supernatants of GL261-WT or GL261-KD tumor cells as measured by ELISA. Data are expressed as the mean values ± SD from one experiment performed in six duplicate. ***, p < 0.001. (d,e) The mRNA expression levels of CCL2 and VEGF in CD45high cells or CD11b+ cells isolated from the brains of at least 5 GL261-WT or GL261-KD tumor-inoculated mice as measured by RT-PCR. The brain-infiltrating cells were isolated as described and pooled/group.

Download figure to PowerPoint

Silencing of glioma-derived galectin-1 boosts IFN-ɣ production in brain-infiltrating CD8+ T cells of tumor-bearing mice

Next, we investigated whether the silencing of glioma-derived galectin-1 induced proportional changes in brain-infiltrating CD3+ T cells or in CD4+ Foxp3+ Tregs in tumor-inoculated mice. Although several reports have demonstrated an important role for tumor-derived galectin-1 in the induction of T cell apoptosis,[8, 9] flow cytometric analysis of the brain-infiltrating immune cell population did not reveal a significant change in the percentage of tumor-infiltrating CD3+ T cells in the absence of tumor-derived galectin-1 (Fig. 5a). In addition, no effect on the percentage of brain-infiltrating CD4+ Foxp3+ Tregs was observed upon intratumoral silencing of the galectin-1 expression. However, the downregulation of galectin-1 resulted in a weak but significant increase in the IFN-ɣ production amongst brain-infiltrating CD8+ T cells isolated from GL261-KD tumor-inoculated mice (p < 0.05; Fig. 5a). This increase was only observed amongst the CD8+ T cell population. Of note, silencing of intratumoral galectin-1 did not cause changes in the cell surface expression levels of several costimulatory or coinhibitory molecules that could boost or suppress adaptive antitumor immune responses (Supporting Information Fig. S4).

image

Figure 5. The prolonged survival of GL261-KD tumor-inoculated mice requires the presence of an intact adaptive antitumor immune response. (a) Flow cytometric analysis of the brain-infiltrating immune cells isolated from GL261-WT or GL261-KD tumor-inoculated mice. With the exception of the CD3 staining, in which each point represents cells that were isolated from one mouse, each point represents pooled brain-infiltrating immune cells isolated from three tumor-bearing mice. (b) The data are represented as Kaplan-Meier graphs of three pooled experiments. The C57/BL6 mice or NSG mice (dashed lines) were inoculated with GL261-WT (black dots, n = 14 for C57/BL6 mice and n = 8 for NSG mice) or GL261-KD tumor cells (open dots, n = 13 for C57/BL6 mice and n = 9 for NSG mice). **, p < 0.01; ***, p < 0.001. (c) Kaplan-Meier survival curves of two pooled experiments in which mice were grafted with GL261-WT (black dots) or GL261-KD cells (open dots). Mice were depleted of CD8a+ T cells (dashed lines) or were left untreated. **, p < 0.01; ***, p < 0.001.

Download figure to PowerPoint

Deficiency in adaptive immunity abrogates the survival benefit observed in GL261-KD tumor-inoculated mice

To further explore the involvement of the adaptive immune response in the control of tumor progression in GL261-WT and GL261-KD tumor-inoculated mice, GL261-WT or GL261-KD tumor cells were injected either in immunecompetent mice or in immunecompromised NSG mice which lack mature T cells, B cells and functional NK cells. Immunecompromised NSG mice inoculated with GL261-KD tumor cells exhibited more rapid and significant death compared with inoculation of the immunecompetent mice with GL261-KD tumor cells (p < 0.0001; Fig. 5b). These findings suggest that the prolonged survival of GL261-KD tumor-inoculated mice requires the presence of an intact adaptive antitumor immune response. Likewise, inoculation of the immunecompromised mice with GL261-WT tumor cells resulted in rapid death compared with the same inoculation of the immunecompetent mice (p = 0.0002).

Interestingly, we also noted a significantly improved survival rate in the immune-compromised mice upon inoculation with the GL261-KD tumor cells as compared to inoculation with the GL261-WT tumor cells (p < 0.01; Fig. 5b). We postulated that quantitative differences in the innate antitumor immune response could be involved in the observed differences in the survival rate of GL261-WT and GL261-KD-inoculated NSG mice (Supporting Information Fig. S5). Upon flow-cytometric analysis of the brain-infiltrating myeloid cell population, we indeed observed a significant reduction in the percentage of brain-infiltrating granulocytic MDSC in NSG mice bearing GL261-KD tumors as compared to GL261-WT tumor-bearing NSG mice (p < 0.01; Supporting Information Fig. S5). We found no significant difference in the number of brain-infiltrating macrophages or monocytic MDSC (Supporting Information Fig. S5).

Consistent with the observations in immunecompromised NSG mice, depletion of the CD8+ T cells in GL261-KD tumor-inoculated C57BL/6 mice resulted in a significantly reduced median survival rate as compared with immune-competent GL261-KD tumor-inoculated mice (p < 0.001; Fig. 5c). These findings confirm that the prolonged survival of GL261-KD tumor-bearing mice requires the presence of intact CD8+ T cell responses.

Silencing of glioma-derived galectin-1 impairs angiogenesis

In a next step, we explored whether glioma-derived galectin-1 also contributes to the in vivo aggressiveness of gliomas by modulating angiogenesis within the tumor microenvironment. Several reports have highlighted an important role for tumor-derived galectin-1 in the extent of angiogenesis observed in glioma-bearing mice.[25, 26] For this purpose, the brains of tumor-bearing mice were prelevated and stained with the endothelial cell marker CD31 (Fig. 6a). Comparison of the vascular density in tumors of GL261-WT and GL261-KD tumor-inoculated mice indeed revealed a pro-angiogenic role for tumor-derived galectin-1 in the GL261 murine glioma model (Fig. 6a; p < 0.001). A similar significant reduction in the vascular density was observed in tumors of GL261-KD tumor-bearing mice as compared to controls by counting the number vessels on conventional hematoxylin-eosin stained brain sections (Supporting Information Fig. S6).

image

Figure 6. Silencing of glioma-derived galectin-1 reduces vascular density and improves the outcome of vaccinated tumor-bearing mice. (a) Tumors of GL261-WT or GL261-KD tumor-bearing mice were stained with the endothelial cell marker CD31 (left). Semiquantitative analysis of the CD31 protein expression level (right). At least two mice/group were analyzed. Data are expressed as mean CD31 positive pixels ± SD. ***, p < 0.001. (b) The data are represented as a Kaplan-Meier graph of three pooled experiments: C57/BL6 mice were grafted with GL261-WT (black dots) or GL261-KD tumer cells (open dots). The mice were administered two vaccinations with DCm-WT (dashed line, n = 27) or DCm-KD (dashed line, n = 26) or were left untreated (n = 14). The DCm were loaded with the lysate from GL261-WT cells (DCm-WT) or GL261-KD (DCm-KD) tumor cells. **, p < 0.01; ***, p < 0.001. (c) The graph represents the Kaplan-Meier analysis of a single experiment in which C57/BL6 mice (black dots) or galectin-1−/− mice (open dots) were grafted with GL261-WT tumor cells. The mice were vaccinated similarly as described previously with DCm-WT (dashed lines, n = 10 for C57/BL6 mice and n = 6 for vaccinated galectin-1−/− mice) or were left untreated (n = 4). NS, not significant.

Download figure to PowerPoint

Silencing of tumor-derived galectin-1 improves the outcome of DC-vaccinated tumor-bearing mice

Next, we compared glioma-bearing mice survival upon intracranial inoculation with GL261-WT or GL261-KD tumor cells in DC-vaccinated mice (Fig. 6b). Hence, we implemented a DC vaccination approach in the orthotopic GL261 mouse glioma model by vaccinating the mice with lysate-loaded mature DCs (DCm) at day 14 and 7 prior to tumor inoculation. Vaccination with GL261-WT-loaded DCs significantly improved the median survival of GL261-WT tumor-bearing mice compared with that of the untreated controls (p < 0.0001), confirming the data previously published by our group.[22] Moreover, prophylactic DC vaccination induced long-term protection against primary tumor inoculation in a small subset of mice (14% respectively, Fig. 6b). In the absence of prophylactic DC immunization, the downregulation of intratumoral galectin-1 significantly delayed in vivo tumor progression (p < 0.0001; Fig. 6b). Similar as for vaccinated GL261-WT tumor-inoculated mice, prophylactic DC vaccination of GL261-KD tumor-inoculated mice significantly improved the median survival of GL261-KD tumor-bearing mice as compared to the untreated controls and induced long-term protection against primary tumor inoculation in a subset of mice (34%, respectively, Fig. 6b). Of note, all of the mice that survived the primary tumor inoculation were also protected against intracranial rechallenge using GL261-WT or GL261-KD tumor cells, suggesting that treatment with lysate-loaded DCm was able to induce immunological memory against the tumor (data not shown). Interestingly, log-rank comparison of the survival rate of vaccinated GL261-WT tumor-inoculated mice and vaccinated GL261-KD tumor-inoculated mice revealed a significant survival benefit for vaccinated mice inoculated with GL261-KD tumor cells (Fig. 6b, p < 0.01). This significant improvement in tumor-bearing mice survival is likely the result of an additive antitumor effect of intratumoral galectin-1 silencing and prophylactic DC immunization.

Host-derived galectin-1 is not involved in the aggressiveness of in vivo glioma progression

Finally, we assessed the impact of host-derived galectin-1 on glioma progression and on the outcome of vaccinated tumor-bearing mice. To address this question, GL261-WT tumor cells were grafted into galectin-1-competent or galectin-1-knockout mice in the absence or presence of prophylactic DC vaccination (Fig. 6c). The absence of host-derived galectin-1 elicited no impact on glioma progression, suggesting that the aggressiveness of the model relates mainly to tumor-derived but not to host-derived galectin-1. In addition, the absence of host-derived galectin-1 did not improve the outcome of vaccinated GL261-WT tumor-inoculated mice (Fig. 6c).

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In this study, we explored whether the presence of galectin-1 in the glioma microenvironment exerts a modulating role on antiglioma immune responses. Accumulating research in different cancer models has recognized an essential role for galectin-1 in tumor-mediated immune evasion, suggesting that specific interference with the local galectin-1 expression might contribute to overcome tumor immune resistance and boost immunotherapeutic strategies.[7, 14, 15]

In this study, we have provided evidence that the local immune responses in glioma-bearing mice are different in the presence or absence of tumor-derived galectin-1. The most remarkable difference was observed in the accumulation of brain-infiltrating macrophages of tumor-bearing mice. Both flow cytometric and pathological analysis revealed that the silencing of glioma-derived galectin-1 significantly decreased the percentage of brain-infiltrating macrophages in tumor-inoculated mice. To our knowledge, this is the first report that provides in vivo evidence for a role of tumor-derived galectin-1 in myeloid cell recruitment towards the tumor microenvironment. Although myeloid cells, such as macrophages, can exert cytotoxicity towards tumor cells and stimulate adaptive antitumor immune responses, the accumulation of these cells within the tumor microenvironment is associated with poor prognosis in many types of cancer.[27] These TAMs have been shown to promote the aggressiveness of tumors by stimulating tumor cell survival, growth and migration. Especially in high-grade gliomas, it has been shown that myeloid-derived cells in the tumor microenvironment far outweigh the presence of T cells, already revealing the underestimated importance of this innate arm of immunity in this disease. Within GBM, the number of tumor-infiltrating macrophages is increased compared to low-grade glioma.[28] Conflicting data have been reported regarding the role of tumor-infiltrating macrophages in the murine GL261 high-grade glioma model. Galarneau et al. reported enhanced tumor progression and faster mortality in GL261 tumor-inoculated mice that were depleted of macrophages.[29] In contrast, Zhu et al. noted that the systemic neutralization of CCL2 production prolonged the survival of C57BL/6 mice bearing intracranial GL261 gliomas.[30] This observation was concomitant with a strong decrease in TAM and MDSC within the tumor microenvironment, suggesting a protumoral role exerted by these cells. Of note, unpublished data from our research group revealed that prophylactic vaccination of immune-competent mice with lysate-loaded DC improved the median survival rate, which was concomitant with a reduction of MDSC within the tumor microenvironment.

To further examine the link between galectin-1 and myeloid cell recruitment, we evaluated the mRNA expression levels of CCL2 in in vitro cultured GL261-WT and GL261-KD tumor cells. This experiment revealed that the downregulation of the intratumoral galectin-1 expression significantly reduced the mRNA expression levels of CCL2 in cultivated GL261-KD tumor cells. CCL2, a chemokine that recruits both macrophages and MDSCs, was originally identified in and purified from human gliomas. Increased CCL2 expression levels have been identified in biopsies of high-grade gliomas compared with low-grade gliomas, suggesting a role in glioma aggressiveness.[31] Although GL261 cells have been shown to produce low levels of CCL2 in vitro, we were not able to detect CCL2 in nonconcentrated supernatants of in vitro cultured GL261-WT or GL261-KD tumor cells.[30] It is, however, probable that the hypoxic in vivo tumor microenvironment upregulates CCL2 in GL261 tumor cells.[30] In addition to CCL2, we also observed a downregulation of VEGF mRNA and protein expression levels in glioma tumor cells upon silencing of intratumoral galectin-1. VEGF has been shown to be an important chemoattractant for myeloid cells.[32] Moreover, VEGF also stimulates the accumulation of MDSCs within the tumor environment.[24] To present, extensive research has been dedicated to the development of drugs capable of depleting myeloid cells from the tumor microenvironment.[33] In this regard, galectin-1 might represent an important therapeutic target. Likewise, VEGF blockers could theoretically result in therapeutic advantage in combination with immunotherapeutic strategies.

Whether glioma-derived galectin-1 also regulates macrophage function remains an open question. Additional ex vivo functional studies are required to address this issue. Glioma-derived galectin-1 might modulate macrophage phenotype and activity, as several studies have reported a regulatory effect of galectin-1 on myeloid cells.[34-36] Interestingly, in none of these reports galectin-1 binding had an effect on macrophage viability, suggesting that the differences in macrophage numbers observed in this study cannot be explained by galectin-1-mediated modulation of macrophage survival.[37, 38] Moreover, all of the data reported on galectin-1 and macrophage function have been reported in the context of the exposure of macrophages or monocytes to the monomeric form of galectin-1 and not to dimeric galectin-1, which is implicated in the regulation of T cell viability and function.

Whereas compelling evidence has accumulated regarding the modulatory effects of tumor-derived galectin-1 on T cell viability, we found no correlation between the amounts of tumor-secreted galectin-1 and the apoptosis of brain-infiltrating lymphocytes of tumor-inoculated mice. Accumulating data have demonstrated that the outcome of galectin-1-glycan interactions is complex and depends on the biochemical structure of galectin-1 and the nature of the target cell.[9] Galectin-1-triggered T cell death has been shown to require its dimerization.[39] The dimeric form of galectin-1 exhibits a rapid dissociation constant, which limits wide spread T cell-modulatory effects. In addition, the binding of galectin-1 to extracellular ligands preserves its dimeric conformation. The rapid dissociation of homodimeric galectin-1 within the glioma microenvironment might represent a possible explanation for the absence of galectin-1-mediated T cell apoptosis. We did, however, observe a difference in the extent of IFN-ɣ production amongst brain-infiltrating CD8+ T cells isolated from tumor-inoculated mice. In the absence of glioma-derived galectin-1, brain-infiltrating CD8+ T cells produced significant higher levels of IFN-ɣ suggesting that also the adaptive immune system is different in the presence or absence of tumor-derived galectin-1.

Finally, we explored the potency of prophylactic DC-mediated immunotherapy to protect mice against primary tumor inoculation in the presence or absence of galectin-1. Absence of host-derived galectin-1 did not change glioma-bearing mice survival neither in the presence nor in the absence of prophylactic DC vaccination. Interestingly, absence of tumor-derived galectin-1 prolonged the survival of glioma-bearing mice even in the absence of prophylactic DC vaccination. By analyzing the extent of angiogenesis in tumor-inoculated mice, we observed a remarkable and significant reduction in the amount of blood vessels in mice inoculated with galectin-1-depleted glioma cells. These findings further underscore the multimodal role of galectin-1 in the glioma microenvironment. Furthermore, we observed that the silencing of tumor-derived galectin-1 significantly improved the survival of DC-vaccinated tumor-bearing mice. These data are consistent with a recently published study by Stannard et al., who provided the first and thus far, the only evidence showing that combining galectin-blocking carbohydrates with immunotherapy can decrease tumor progression and improve the outcome of tumor-bearing mice.[40] However, in contrast to that study in which thiodigalactoside, a nonmetabolizable dissacharide, was used to none-specifically target the in vivo interaction between galectin-1 and β-galactoside-expressing macromolecules, we developed a galectin-1-targeting miRNA construct to permanently suppress the expression of glioma-derived galectin-1. In this way, we have avoided the non-specific silencing of other possible expressed galectins. Moreover, the use of sequence-specific, post-transcriptional gene silencing has allowed for the interference of both intracellular and extracellular functions of galectin-1. This is interesting because an intracellular role for galectin-1 has been implicated in Ras oncogene activation,[41] in temozolomide resistance (42;42;42) and also in angiogenesis.[25, 26] However, the in vitro transduction of glioma cells with a galectin-1-targeting miRNA construct remains an artificial system that is not easily translatable into a clinical setting. We therefore are currently assessing in vivo delivery systems for galectin-1-targeting siRNAs, the feasibility of which has already been demonstrated in a preliminary study.[26] Note that the presented experiment did not include a group of mice that was treated with mock-loaded DCm (DCm-mock). Previous studies from our research group have demonstrated that prophylactic vaccination with DCm-mock can induce a small, significant improvement in glioma-bearing mice survival, but was not capable of inducing long-term survival.[22]

Overall, these findings suggest that the targeting of glioma-derived galectin-1 could be a promising and realistic new treatment modality to improve the clinical outcome of high-grade patients both in the absence or presence of DC-based immunotherapy.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Ellen Dilissen for excellent technical assistance with RT-PCR, Anaïs Van Hoylandt for help with the in vivo tumor inoculation experiments, Lien Vandenberk for her assistance with the isolation of the brain-infiltrating cells and Sonja Bobic for her help with the ELISA for VEGF.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article

FilenameFormatSizeDescription
ijc28426-sup-0001-suppinfo.doc5149KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.