Skip to main content
NCBI home page
Molecular Oncology logoLink to Molecular Oncology
. 2013 Sep 18;8(1):50–58. doi: 10.1016/j.molonc.2013.09.002

Glioma‐derived macrophage migration inhibitory factor (MIF) promotes mast cell recruitment in a STAT5‐dependent manner‡

Jelena Põlajeva 1,†, Tobias Bergström 1,†, Per-Henrik Edqvist 1, Anders Lundequist 2, Anna Sjösten 1, Gunnar Nilsson 3, Anja Smits 4, Michael Bergqvist 4, Fredrik Pontén 1, Bengt Westermark 1, Gunnar Pejler 2, Karin Forsberg Nilsson 1, Elena Tchougounova 1,✉
PMCID: PMC5528513  PMID: 24091309

Abstract

Recently, glioma research has increased its focus on the diverse types of cells present in brain tumors. We observed previously that gliomas are associated with a profound accumulation of mast cells (MCs) and here we investigate the underlying mechanism.

Gliomas express a plethora of chemoattractants. First, we demonstrated pronounced migration of human MCs toward conditioned medium from cultures of glioma cell lines. Subsequent cytokine array analyses of media from cells, cultured in either serum‐containing or ‐free conditions, revealed a number of candidates which were secreted in high amounts in both cell lines. Among these, we then focused on macrophage migration inhibitory factor (MIF), which has been reported to be pro‐inflammatory and ‐tumorigenic. Infiltration of MCs was attenuated by antibodies that neutralized MIF. Moreover, a positive correlation between the number of MCs and the level of MIF in a large cohort of human glioma tissue samples was observed.

Further, both glioma‐conditioned media and purified MIF promoted differential phosphorylation of a number of signaling molecules, including signal transducer and activator of transcription 5 (STAT5), in MCs. Inhibition of pSTAT5 signaling significantly attenuated the migration of MCs toward glioma cell‐conditioned medium shown to contain MIF. In addition, analysis of tissue microarrays (TMAs) of high‐grade gliomas revealed a direct correlation between the level of pSTAT5 in MCs and the level of MIF in the medium.

In conclusion, these findings indicate the important influence of signaling cascades involving MIF and STAT5 on the recruitment of MCs to gliomas.

Keywords: Mast cell, Glioma, MIF, pSTAT5

Highlights

  • MC accumulation in glioma is malignancy grade‐dependent.

  • Neutralization of MIF produced by glioma cells lowers the extent of MC migration.

  • The extent of MC recruitment is correlated with the level of MIF.

  • MIF‐induced accumulation of MCs in vivo is associated with activation of STAT5.

Abbreviations

MC

mast cell

MIF

microphage migration inhibitory factor

STAT

signal transducer and activator of transcription

TMA

tissue microarray

GBM

glioblastoma multiforme

BBB

blood–brain barrier

IHC

immunohistochemistry

CI

cytoplasmic intensity

CF

cytoplasmic fraction

CBMC

cord blood mast cell

1. Introduction

Despite complex treatment involving surgery and radio‐chemotherapeutic intervention, the prognosis for high‐grade gliomas remains among the poorest, with a median survival time for patients with glioblastoma (GBM) just over one year (Stupp et al., 2005). The poor prognosis of high‐grade glioma is associated with pronounced invasiveness and therapeutic resistance. Development of novel therapeutic modalities is hampered by the lack of reliable prognostic markers and molecular targets.

The hallmarks of GBM include disruption of the blood brain barrier (BBB), as a consequence of abnormal neovasculature and excessive vessel leakiness (Anderson et al., 2008), and tumor invasiveness, with characteristic migration of cancer cells into surrounding brain parenchyma orchestrated by both autocrine and paracrine factors. Importantly, these multistep processes begin already in low‐grade gliomas and include alterations in cellular morphology and interactions with the extracellular matrix and secretion of proteolytic enzymes (Kwiatkowska and Symons, 2013).

The contemporary anti‐tumor strategies are shifting from tumor‐cell centric to those targeting various components of the tumor microenvironment. Their contribution to the proliferative and migratory potential of cancer cells is being increasingly recognized (Hanahan and Coussens, 2012). For instance, the effective evasiveness of glioma from immune responses is considered to involve chronic inflammation and recruitment of myeloid suppressor cells, microglia and T‐regulatory cells that effectively obstruct anti‐tumor immune response (Albesiano et al., 2010). However, the ultimate interrelation between the immune system and glioma remains unclear and is being actively investigated.

There are growing indications that MCs are key modulators of the tumor microenvironment, influencing angiogenic and immunoenvironmental processes, as well as tissue remodeling (Maltby et al., 2009). At the same time, the prognostic value of MC infiltration into tumors is controversial, since related inflammatory processes can either facilitate or hinder tumor development, depending on the setting. We were the first to report that gliomas also contain MCs (Polajeva et al., 2011). Moreover, we presented evidence suggesting that chemotactic infiltration of MCs into glioma is facilitated by the CXCL12/CXCR4 axis, where CXCL12 is a chemoattractant produced by the glioma itself.

Macrophage migration inhibitory factor (MIF) is currently emerging as a key factor in glioma. For example, its expression correlates well with glioma recurrence and poorer prognosis and, thereby, MIF is currently being considered as a valuable and independent prognostic indicator for patients with glioma (Wang et al., 2012). MIF is also expressed by a variety of immune cells, including macrophages, lymphocytes, and eosinophils, as well as by endothelial cells and epithelial cells. Further, MIF has been shown to promote tumor progression in many types of malignancies and been implicated as a direct link between the process of inflammation and tumor growth, thus making it a potential target for anti‐cancer treatment (Conroy et al., 2010).

Here, we demonstrate that accumulation of MCs, which is dependent on the malignancy grade of the glioma, correlates with the level of MIF expression, and that MIF upregulates pSTAT5 levels. In addition, analysis of TMAs of high‐grade gliomas revealed a direct correlation between the level of pSTAT5 in MCs and the level of MIF. In conclusion, the present investigation helps clarify the mechanism through which gliomas recruit MCs, demonstrating the influence of the STAT5‐related signaling cascade downstream of MIF on the migratory properties of MCs, a finding with potential therapeutic implications.

2. Materials and methods

2.1. Patients, tissue microarrays and immunostaining

Informed consent for the use of human brain tissue and for access to medical records for research purposes was obtained, and all material obtained in compliance with the Declaration of Helsinki. The experiments involving human tissue samples were approved both by the Ethics Committee of Uppsala University (Application Dnr Ups 02‐330 and Ups 06‐084) and the Ethics Committee of Karolinska Institutet (Application Dnr Ki 02‐254) and written informed consent was solicited prior to collection of these samples.

Tissue microarrays (TMAs), immunohistochemistry (IHC) and slide scanning were performed using the strategies of The Human Protein Atlas project (www.proteinatlas.org) (Kampf et al., 2012; Uhlen et al., 2010).

TMAs involved tumor tissue from high‐grade gliomas (anaplastic gliomas and glioblastomas, n = 101) and gliomas WHO grade II (Elsir et al., 2011) (astrocytomas, oligoastrocytomas and oligodendrogliomas grade II, n = 87). For immunohistochemical analysis, primary antibodies directed against human MC tryptase, hTPS (sc‐33676, Santa Cruz Biotech, Santa Cruz, CA, USA) and macrophage migration inhibitory factor MIF (HPA003868, Atlas Antibodies, Stockholm, Sweden) were employed. Briefly, slides of tissue samples were deparaffinized in xylene, rehydrated in a series of aqueous solutions with decreasing concentrations of ethanol, boiled (125 °C, 4 min) in epitope retrieval buffer (Thermo Scientific, Waltham, USA) in a pressure boiler and subsequently cooled for 30 min. Immunochistochemistry was then performed using the Autostained 480 instrument (Lab Vision, Freemont, USA) with 3′3′‐diaminobenzidine (DAB) as a substrate. Thereafter, the slides were counterstained with hematoxylin, mounted and scanned using the ScanScope XT system (Aperio Technologies, Vista, USA).

All MIF staining was evaluated by the same neuropathologist. Cytoplasmic intensity (CI) was subdivided into 4 groups (negative, weak, moderate, strong) and the cytoplasmic fraction (CF) into 6 groups (0–1%, 2–10%, 11–25%, 26–50%, 51–75%, >75%). For graphical representation, the moderate and strong CI groups were pooled. The numbers of MCs per TMA core were also determined and grouped into three categories (0–5, 6–20, ≥21) for statistical analysis. During the evaluation of MIF staining, the data regarding the numbers of MCs were blinded and vice versa.

Immunofluorescent staining was carried out as described elsewhere (Lindberg et al., 2009; Polajeva et al., 2011) employing primary antibodies against human MC tryptase, hTPS (sc‐33676) and pSTAT5 (sc‐11671, Santa Cruz Biotech, Santa Cruz, USA) and chicken anti‐goat Alexa 488 and donkey anti‐mouse Alexa 555 (Invitrogen, Carlsbad, USA) as secondary antibodies. Identification and quantification of human MCs co‐expressing pSTAT5 was performed. The samples were dichotomized into ‘HIGH MIF’ (with moderate or strong MIF cytoplasmic intensity and high MC numbers (≥21 MC/TMA core)) and ‘LOW MIF’ (with negative or weak MIF cytoplasmic intensity and low MC numbers (0–5 MC/TMA core)).

2.2. Cell cultures

All cells were cultured at 37 °C under 5% CO2. U‐2987 MG, a human glioma cell line (Hagerstrand et al., 2006), was cultured in 10% FBS‐containing MEM supplemented with 4 mM l‐glutamine and 100 units/ml penicillin and 0.1 mg/ml streptomycin (Sigma Aldrich, St Loius, USA). U‐3054 MG cells were cultured on laminin (10 μg/ml) in serum‐free (stem cell) conditions, to enrich for stem‐like glioma cells, as described previously (Ferletta et al., 2011).

The human MC line LAD2 (obtained from Prof Dean Metcalfe at NIH/NIAID, MD, USA) was cultured as described previously (Kirshenbaum et al., 2003) with addition of 10% FBS (Invitrogen, Carlsbad, USA). Culturing of MCs derived from human cord blood (CBMC) has also been described elsewhere. The cultures contained >95% of tryptase‐positive MCs as assayed by MC‐specific serine proteinase tryptase staining (Sjostrom et al., 2002).

2.3. Transwell migration assays

To evaluate the capacity of glioma cells to recruit MCs, conditioned media from confluent U‐2987 MG and U‐3054 MG cells were used as the chemoattractant with unconditioned media serving as controls. LAD2 and CBMCs were grown in the absence of serum overnight, resuspended in control medium (5 × 104 cells/ml), then allowed to migrate through transwells with 8 μm pores for 8 h, and thereafter counted. Migration of MCs toward glioma cell‐conditioned medium was set to 100%. These experiments were performed in duplicate.

Neutralizing experiments were performed in triplicates as described previously (Polajeva et al., 2011) using two MIF antibody concentrations (50 and 0.5 μg/ml; R&D Systems, Abingdon, UK) and control IgG (50 μg/ml; R&D Systems, Abingdon, UK). The selection of concentration range for MIF neutralizing antibodies was based on previous publications (Bernhagen et al., 2007; Ferro et al., 2008).

To examine the potential influence of STAT5 on MC migration, LAD2 transwell migration toward U‐2987MG media was performed for 8 and 20 h in the presence and absence of pimozide, (P1793, Sigma Aldrich, St Loius, USA), an inhibitor of STAT5 phosphorylation. The efficacy of this inhibitor was confirmed by western blotting as described elsewhere (Ferletta et al., 2011). The primary antibodies used were directed against pSTAT5 (sc‐11671) and β‐actin (A5441, Sigma Aldrich, St Loius, USA).

2.4. Human cytokine and phospho‐kinase arrays

The human cytokine array panel A allowing for detection of 36 different cytokines and chemokines (R&D Systems, Abingdon, UK) was utilized in accordance with the manufacturer's instructions. Briefly, media from confluent cultures of glioma cells (collected 96 h after seeding) and corresponding unconditioned (control) media were pre‐incubated with the antibody cocktail and incubated overnight on the membranes. Streptavidin‐HRP was used for signal detection with the Amersham ECL Plus western blotting detection system (GE Healthcare, Buckingamshire, UK).

The human phospho‐kinase array kit which allows for simultaneous detection of the changes in relative levels of phosphorylation of 46 kinase phosphorylation sites (R&D Systems; Abingdon, UK) was also used in accordance with the manufacturer's directions. In brief, LAD2 cells grown overnight in the absence of serum were placed for 5 min in medium from confluent 96‐h cultures of U‐2987 MG cells or control medium with or without MIF (R&D Systems, Abingdon, UK) at previously described concentrations (Kim et al., 2008). Lysates were prepared and applied to membranes overnight, following which signals were detected as above after appropriate application of antibody cocktails and streptavidin‐HRP solutions.

To quantify and analyze the data a custom‐made pipeline was specifically written using CellProfiler software (r11710) where pixel intensity (ranging from 0 to 616) was determined for each spot. Since capture antibodies are spotted onto the nitrocellulose membrane in duplicates, for quantification of both arrays, the pixel intensity of each spot was calculated, integrated and corrected for background and the resulting average values from duplicate spots plotted.

2.5. Statistical analysis

For group‐wise comparisons the Student's unpaired t‐test was used. The chi‐square test was applied to analyze the distribution of malignancy (GraphPad Sowtware 6.0). To measure statistical dependence, Spearman's rank correlation coefficient ρ was calculated and a p‐value of <0.05 was considered to be statistically significant.

3. Results

3.1. The numbers of MCs in human glioma is dependent on malignancy grade

Since we observed more MCs in human GBM compared to low‐grade gliomas using a limited number of samples (Polajeva et al., 2011), we extended this finding here by staining immunohistochemically for MC tryptase in tumor tissues from 188 patients. As expected, the extent of MC infiltration into low‐ and high‐grade gliomas differed significantly (Figure 1, p < 0.001). Most (95%) of the former (gliomas grade II) contained fewer than 5 MCs per TMA core and only 4 (5%) had between 6 and 20. In contrast, among high‐grade gliomas (i.e., gliomas grade III and IV), 17 (17%) exhibited 6–20 MCs per TMA core and 10 (10%) ≥21.

Figure 1.

Figure 1

Open in a new tab

The number of MCs in human gliomas is malignancy grade‐dependent. Staining of TMA samples for MC tryptase was used to determine the number of MC. *** p < 0.001.

3.2. Glioma cells recruit mast cells and express a distinct set of chemokines

As shown previously, mouse glioma cells recruit MCs by secreting chemoattractants (Polajeva et al., 2011). Here, we extended these observations to humans. As depicted in Figure 2A, conditioned media from cultures of U‐2987 MG and U‐3054 MG cells promoted the migration of human MCs (LAD2 cells), as well as of MCs derived from human cord blood (CBMCs) (data not shown).

Figure 2.

Figure 2

Open in a new tab

MCs are recruited by chemokines released by in glioma cells. Neutralization of glioma‐derived MIF attenuates the migration of MCs toward conditioned medium. (A) Migration of MCs towards conditioned media from cultures of glioma cells. The experiments were performed 2 times, with duplicates in each case. The error bars represent the SD. ** p < 0.01, *** p < 0.001. (B) The cytokine profiles in the media of U‐2987 MG cell line grown in serum‐containing medium and (C) a low‐passage U‐3054 MG grown under stem cell conditions. The intensity of the dots (left panels) was quantified and the average integrated pixel intensity corrected from the control medium is plotted (right panels). The error bars depict the SD. (D) Neutralization of MIF in U2987‐conditioned medium (C.M.) with antibodies (MIF ab) attenuates the migration of MCs. Three independent experiments with duplicates were performed. The error bars represent the SD. * p < 0.05.

To examine the underlying mechanism, we performed cytokine array analyses and detected in the medium of a glioma cell line, U‐2987 MG, grown in the presence of serum high levels of MIF and PAI‐1, as well as some IL‐6 and IL‐8 (Figure 2B). This profile was partially similar to that of conditioned medium from a newly established low‐passage glioma cell line, U‐3054 MG (passage 12), grown under stem cell conditions, which contained high levels of MIF and PAI‐1 but also sICAM‐1 (Figure 2C).

3.3. Neutralization of MIF produced by glioma cells attenuates MC migration

A main goal of the present investigation was to use unbiased technology to identify novel MC chemoattractants secreted by human glioma cells. In this process we focused on macrophage migration inhibitory factor (MIF) as a candidate chemoattractant because MIF was found to be the most highly upregulated cytokine of the array. Furthermore, MIF has been implicated in the development of a variety of malignancies (Babu S 2012) and is acknowledged as a marker of glioma progression (Mittelbronn 2011). We therefore attempted to block MC chemotaxis by inhibition of MIF. As documented in Figure 2D, neutralization of the MIF in U‐2987 MG‐conditioned medium with antibodies significantly lowered the motility of LAD2 MC cells in a dose‐dependent manner. A similar, although less pronounced effect was observed with U‐3054 MG‐conditioned medium (data not shown).

3.4. The extent of MC recruitment is correlated with the level of MIF

Analysis of consecutive sections of the TMAs revealed a significant correlation between the number of infiltrating MCs and the cytoplasmic intensity of staining for MIF (CI) (Figure 3A). Thus negative staining was associated with 0–5 MCs per TMA core in all cases (n = 21). The proportion of TMA cores with low numbers of MCs was greatest (n = 20, 77%) among those with weak MIF CI, with intermediate (6–20 MC/TMA core) and high (≥21 MC/TMA core) numbers in 8% (n = 2) and 15% (n = 4) of those samples, respectively. The proportion of TMA cores exhibiting low numbers of MCs was lowest with moderate or strong CI staining for MIF. These values were 64% (n = 35) of all samples combined, 29% (n = 16) with intermediate and 7% (n = 4) with high MC numbers.

Figure 3.

Figure 3

Open in a new tab

The level of MIF is correlated with the extent of MC recruitment. The relationship between the number of MCs and (A) cytoplasmic intensity of MIF (CI). Spearman's rank correlation coefficient ρ = 0.248 and p = 0.012. (B) Representative TMA cores of negative (left panels) and positive staining (middle panels) for tryptase and MIF. Selected areas from middle panels are magnified in the right panels.

A significant positive correlation between the number of infiltrating MCs and the cytoplasmic fraction (CF) of staining for MIF was also observed (p = 0.011, data not shown). Furthermore, MIF CI and CF scores were correlated (p = 9.53 × 10−6, data not shown). Representative positive and negative staining for MCs and MIF is illustrated in Figure 3B.

3.5. Both MIF and glioma‐conditioned medium promote phosphorylation of STAT5

In order to investigate which targets downstream of MIF are affected, a phospho‐kinase array was performed. Analysis of protein lysates from MIF‐stimulated LAD2 cells by this array revealed intense augmentation of the intracellular level of phosphorylated STAT5 (pSTAT5) as compared to control medium (Figure 4A, bottom panel). Capture antibodies against different phosphorylation sites of isoforms STAT5a and STAT5b yielded similar results. In the case of STAT5b, higher concentration of MIF resulted in slightly more intense degree of phosphorylation at Y699 amino acid. This effect was less pronounced in the case of phosphorylation at Y694 on STAT5a. Significantly, at the concentrations tested MIF‐stimulated phosphorylation of both isoforms of STAT5 to almost the same degree as U‐2987 MG‐conditioned medium.

Figure 4.

Figure 4

Open in a new tab

MIF alters signaling networks in MCs. (A) Upper panels: the phosphorylation profiles of lysates of MCs cultured in control medium (top left), medium containing MIF (10 μg/ml, top right; 100 μg/ml, bottom left) or U2987‐conditioned medium (C.M., bottom right). In the lower graph quantification of the dots of interest is plotted in terms of the as integrated pixel intensities. The error bars represent the SD. (B) Inhibition of STAT5 phosphorylation by pimozide (PMZ) reduces the migration of MCs. Three independent experiments were performed with triplicates. C.M. – conditioned medium. The error bars represent the SD. ** p < 0.01, n.s. = not significant.

To test this indication that signaling by STAT5 downstream of MIF promotes MC migration, migration in the presence of pimozide, an inhibitor of STAT5 phosphorylation (Nelson et al., 2011), was examined. The inhibitory effect was verified by western blotting (data not shown). Indeed, at both concentrations tested, pimozide significantly attenuated the migration of MCs toward U‐2987 MG‐conditioned medium at 20 h (Figure 4B). The trend was similar, although less pronounced already at 8 h (data not shown).

3.6. MIF‐induced accumulation of MCs in vivo is associated with activation of STAT5

To further evaluate the biological relevance of the observations above on cell lines, TMA samples were co‐stained for pSTAT5 and MC tryptase (Figure 5A). On average, 88% of the MCs in the ‘HIGH MIF’ samples (n = 4, as defined in the Materials and methods section) also stained positive for pSTAT5; whereas the corresponding value for the ‘LOW MIF’ samples (n = 5) was only 58%, p < 0.05 (Figure 5B).

Figure 5.

Figure 5

Open in a new tab

High MIF expression in human glioma correlates with high STAT5 activation. (A) Representative images of TMA samples with high levels of MIF and ≥21 MC/TMA core (HIGH MIF, top panels), and those with low levels of MIF and 0–5 MC/TMA core (LOW MIF, bottom panels) stained for pSTAT5 and MC tryptase (hTPS). Scale bar = 20 μM. (B) Quantification of positive staining for pSTAT5 in HIGH MIF (n = 4) and LOW MIF (n = 5) samples. The error bars represent the SEM. * p < 0.05.

4. Discussion

It remains uncertain whether inflammation is a prerequisite for or a consequence of tumor development and growth. Despite much effort, the interplay between inflammatory cells and tumor cells is still not understood. The prognosis for most patients with glioma is poor and for those with glioblastoma mean survival is only 16 months.

Our understanding of the involvement of inflammation in gliomagenesis is gradually improving. However, despite numerous investigations, the underlying concepts implicating various immune cells, including helper T cells, microglia and NK cells, in the development of glioma, remain unclear. In general, these immune cells have been proposed to be predominantly pro‐tumorigenic, with numbers rising in more malignant tumor types (Ghosh and Chaudhuri, 2010; Tran Thang et al., 2010).

We demonstrated recently that gliomas recruit MCs (Polajeva et al., 2011), another immunomodulatory cell that plays a key role in inflammation. This recruitment is more pronounced as malignancy grade increases, suggesting that the role played by MCs is detrimental. Here, we performed immunohistochemistry on TMAs of both low‐grade and high‐grade gliomas and discovered a significant positive correlation between the numbers of MCs and the malignancy grade of the tumor.

This correlation suggests that gliomas secrete factors that recruit MCs and that MCs may promote glioma development, introducing interference with MC migration as a potential therapeutic strategy. Therefore, in an attempt to identify the chemoattractants involved, we observed pronounced migration of human MCs toward media from cultures of the glioma cell line U‐2987 MG, as well as low‐passage glioma cell line U‐3054 MG grown in neural stem cell medium. Arrays analysis revealed several candidates, of which two, PAI‐1 and MIF, were expressed at high levels by both cell lines, while more moderate expression of IL‐6, IL‐8 and sICAM was observed. These differences may be due to intratumoral heterogeneity, as well as differences in culturing conditions.

Subsequently, we focused on MIF, which is secreted by GBM cells in response to hypoxic and hypoglycemic stress, and is associated with recurrence of the disease (Bacher et al., 2003; Wang et al., 2012). Neutralization of MIF attenuated the migration of MCs. However, residual chemoattractant activity toward MC was present even after adding the blocking antibody against MIF, suggesting that additional, as yet unidentified MC chemoattractants are secreted by glioma cells. MIF binds to the CXCR4 receptor (Bernhagen et al., 2007), through which it may exert its pro‐migratory action. Indeed, we previously showed that CXCR4 plays a role in the recruitment of MCs (Polajeva et al., 2011). Regulation of MC recruitment to gliomas by MIF is supported further by our TMA data demonstrating a positive correlation between the intensity and fraction of cytoplasmic staining for MIF and the number of intratumoral MCs.

In line with a functional influence of MIF on MCs, we found here that signaling networks within these cells are affected by MIF, which promotes differential phosphorylation of a number of intracellular signaling molecules, including STAT5. The receptor linking the action of MIF to the phosphorylation of STAT5 may be CXCR4, as reports have been published demonstrating that ligand binding of this receptor leads to phosphorylation of STAT5 (Honczarenko et al., 2006; Mowafi et al., 2008). Interestingly, glioma‐conditioned medium exerted effects on the phosphorylation profile of MC proteins, including elevated phosphorylation of STAT5a at amino acid Y694, similar to those exposed to MIF. Accordingly, it seems likely that MIF secreted by gliomas is responsible for much of STAT5a phosphorylation in MCs.

When we next characterized the functional consequences of STAT5 phosphorylation in MCs, we found that blockage of STAT5 signaling inhibited the migration of MCs toward glioma‐conditioned medium, an effect similar to that caused by neutralization of MIF. Together, these findings indicate that gliomas recruit MCs by secreting MIF, which induces the migration of these cells through mechanisms involving signaling by STAT5. This conclusion is in agreement with observations that STAT5 enhances migration by other immune cells, e.g., B‐cells (Pierau et al., 2012).

The association between the level of MIF and the number of pSTAT5‐positive MCs was also examined by immunoflourescent analysis of the TMAs. Samples of malignant gliomas with high levels of MIF were found to accumulate large numbers of pSTAT5‐positive MCs and vice versa.

The role of MCs in cancer is not elucidated and indicates the existence of versatile interactive mechanisms between glioma cells and MCs may exist. We hypothesize that in low‐grade gliomas low MIF expression by cancer cells and modest numbers of MCs are associated with activated antitumorigenic machinery. First of all, this affects the function of MCs, their migratory properties and the profile of released mediators but also effector molecules expressed by glioma cells. The situation would be opposite in high‐grade gliomas, where glioma expressing effector molecules profile is altered and, subsequently, the role of MCs is switched into pro‐tumorigenic.

Here we suggest that the increased expression of glioma‐derived MIF and abundant level of MCs, with subsequently activated STAT5, represent a phenomenon specifically occurring in high‐grade gliomas. Our findings depict the pro‐tumorigenic features of MCs associating with migratory properties and activation of STAT5‐related intracellular signaling pathways. A number of our present observations have therapeutic implications. For example, since mediators released by MCs promote the invasiveness of cancer cells (Strouch et al., 2010), inhibition of MC degranulation by agents currently used to treat other MC‐associated disorders (Jensen et al., 2008) could be of therapeutic value. In addition, this investigation emphasizes inhibition of MIF as a potential strategy for treatment of glioma (Wang et al., 2012). Finally, on the basis of our results, inhibition of STAT5 signaling may prevent recruitment of MCs to gliomas, thereby preventing their detrimental effects in these tumors.

Acknowledgements

This work was financed by grants from the Swedish Research Council (Vetenskapsrådet), the Gunvor and Josef Anérs Foundation, the Åke Wibergs Foundation, the Knut and Alice Wallenberg Foundation, the Swedish Cancer Society and the Department of Immunology, Genetics and Pathology and Medical Faculty of Uppsala University.

The authors wish to thank Annika Hermansson for providing the media from culture of U‐3054 MG cells and Marianne Kastemar, and Jeremy Adler for technical assistance. We are also grateful to Martin Simonsson at the Center for Image Analysis and Science for Life BioVis platform in Uppsala for assistance in analyzing cytokine images and phospho‐arrays.

Põlajeva Jelena, Bergström Tobias, Edqvist Per-Henrik, Lundequist Anders, Sjösten Anna, Nilsson Gunnar, Smits Anja, Bergqvist Michael, Pontén Fredrik, Westermark Bengt, Pejler Gunnar, Forsberg Nilsson Karin and Tchougounova Elena, (2014), Glioma‐derived macrophage migration inhibitory factor (MIF) promotes mast cell recruitment in a STAT5‐dependent manner, Molecular Oncology, 8, doi: 10.1016/j.molonc.2013.09.002.

‡

This work was financed by grants from the Swedish Research Council (Vetenskapsrådet), the Gunvor and Josef Anérs Foundation, the Åke Wibergs Foundation, the Knut and Alice Wallenberg Foundation, the Swedish Cancer Society and the Department of Immunology, Genetics and Pathology and Medical Faculty of Uppsala University.

Contributor Information

Jelena Põlajeva, Email: jelena.polajeva@igp.uu.se.

Tobias Bergström, Email: tobias.bergstrom@igp.uu.se.

Per-Henrik Edqvist, Email: per-henrik.edqvist@igp.uu.se.

Anders Lundequist, Email: lundequist@gmail.com.

Gunnar Nilsson, Email: gunnar.p.nilsson@ki.se.

Anja Smits, Email: anja.smits@neuro.uu.se.

Michael Bergqvist, Email: michael.bergqvist@onkologi.uu.se.

Fredrik Pontén, Email: fredrik.ponten@igp.uu.se.

Bengt Westermark, Email: bengt.westermark@igp.uu.se.

Gunnar Pejler, Email: gunnar.pejler@slu.se.

Karin Forsberg Nilsson, Email: karin.nilsson@igp.uu.se.

Elena Tchougounova, Email: elena.chugunova@igp.uu.se.

References

  1. Albesiano, E. , Han, J.E. , Lim, M. , 2010. Mechanisms of local immunoresistance in glioma. Neurosurg. Clin. N. Am.. 21, 17–29. [DOI] [PubMed] [Google Scholar]
  2. Anderson, J.C. , McFarland, B.C. , Gladson, C.L. , 2008. New molecular targets in angiogenic vessels of glioblastoma tumours. Expert. Rev. Mol. Med.. 10, e23 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bacher, M. , Schrader, J. , Thompson, N. , Kuschela, K. , Gemsa, D. , Waeber, G. , Schlegel, J. , 2003. Up-regulation of macrophage migration inhibitory factor gene and protein expression in glial tumor cells during hypoxic and hypoglycemic stress indicates a critical role for angiogenesis in glioblastoma multiforme. Am. J. Pathol.. 162, 11–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bernhagen, J. , Krohn, R. , Lue, H. , Gregory, J.L. , Zernecke, A. , Koenen, R.R. , Dewor, M. , Georgiev, I. , Schober, A. , Leng, L. , Kooistra, T. , Fingerle-Rowson, G. , Ghezzi, P. , Kleemann, R. , McColl, S.R. , Bucala, R. , Hickey, M.J. , Weber, C. , 2007. MIF is a noncognate ligand of CXC chemokine receptors in inflammatory and atherogenic cell recruitment. Nat. Med.. 13, 587–596. [DOI] [PubMed] [Google Scholar]
  5. Conroy, H. , Mawhinney, L. , Donnelly, S.C. , 2010. Inflammation and cancer: macrophage migration inhibitory factor (MIF) – the potential missing link. QJM. 103, 831–836. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Elsir, T. , Qu, M. , Berntsson, S.G. , Orrego, A. , Olofsson, T. , Lindstrom, M.S. , Nister, M. , von Deimling, A. , Hartmann, C. , Ribom, D. , Smits, A. , 2011. PROX1 is a predictor of survival for gliomas WHO grade II. Br. J. Cancer. 104, 1747–1754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Ferletta, M. , Caglayan, D. , Mokvist, L. , Jiang, Y. , Kastemar, M. , Uhrbom, L. , Westermark, B. , 2011. Forced expression of Sox21 inhibits Sox2 and induces apoptosis in human glioma cells. Int. J. Cancer. 129, 45–60. [DOI] [PubMed] [Google Scholar]
  8. Ferro, E.A. , Mineo, J.R. , Ietta, F. , Bechi, N. , Romagnoli, R. , Silva, D.A. , Sorda, G. , Bevilacqua, E. , Paulesu, L.R. , 2008. Macrophage migration inhibitory factor is up-regulated in human first-trimester placenta stimulated by soluble antigen of Toxoplasma gondii, resulting in increased monocyte adhesion on villous explants. Am. J. Pathol.. 172, 50–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ghosh, A. , Chaudhuri, S. , 2010. Microglial action in glioma: a boon turns bane. Immunol. Lett.. 131, 3–9. [DOI] [PubMed] [Google Scholar]
  10. Hagerstrand, D. , Hesselager, G. , Achterberg, S. , Wickenberg Bolin, U. , Kowanetz, M. , Kastemar, M. , Heldin, C.H. , Isaksson, A. , Nister, M. , Ostman, A. , 2006. Characterization of an imatinib-sensitive subset of high-grade human glioma cultures. Oncogene. 25, 4913–4922. [DOI] [PubMed] [Google Scholar]
  11. Hanahan, D. , Coussens, L.M. , 2012. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell. 21, 309–322. [DOI] [PubMed] [Google Scholar]
  12. Honczarenko, M. , Le, Y. , Swierkowski, M. , Ghiran, I. , Glodek, A.M. , Silberstein, L.E. , 2006. Human bone marrow stromal cells express a distinct set of biologically functional chemokine receptors. Stem Cells. 24, 1030–1041. [DOI] [PubMed] [Google Scholar]
  13. Jensen, B.M. , Akin, C. , Gilfillan, A.M. , 2008. Pharmacological targeting of the KIT growth factor receptor: a therapeutic consideration for mast cell disorders. Br. J. Pharmacol.. 154, 1572–1582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kampf, C. , Olsson, I. , Ryberg, U. , Sjostedt, E. , Ponten, F. , 2012. Production of tissue microarrays, immunohistochemistry staining and digitalization within the human protein atlas. J. Vis. Exp.. 63 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kim, J.Y. , Kwok, S.K. , Hur, K.H. , Kim, H.J. , Kim, N.S. , Yoo, S.A. , Kim, W.U. , Cho, C.S. , 2008. Up-regulated macrophage migration inhibitory factor protects apoptosis of dermal fibroblasts in patients with systemic sclerosis. Clin. Exp. Immunol.. 152, 328–335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kirshenbaum, A.S. , Akin, C. , Wu, Y. , Rottem, M. , Goff, J.P. , Beaven, M.A. , Rao, V.K. , Metcalfe, D.D. , 2003. Characterization of novel stem cell factor responsive human mast cell lines LAD 1 and 2 established from a patient with mast cell sarcoma/leukemia; activation following aggregation of FcepsilonRI or FcgammaRI. Leuk. Res.. 27, 677–682. [DOI] [PubMed] [Google Scholar]
  17. Kwiatkowska, A. , Symons, M. , 2013. Signaling determinants of glioma cell invasion. Adv. Exp. Med. Biol.. 986, 121–141. [DOI] [PubMed] [Google Scholar]
  18. Lindberg, N. , Kastemar, M. , Olofsson, T. , Smits, A. , Uhrbom, L. , 2009. Oligodendrocyte progenitor cells can act as cell of origin for experimental glioma. Oncogene. 28, 2266–2275. [DOI] [PubMed] [Google Scholar]
  19. Maltby, S. , Khazaie, K. , McNagny, K.M. , 2009. Mast cells in tumor growth: angiogenesis, tissue remodelling and immune-modulation. Biochim. Biophys. Acta. 1796, 19–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mowafi, F. , Cagigi, A. , Matskova, L. , Bjork, O. , Chiodi, F. , Nilsson, A. , 2008. Chemokine CXCL12 enhances proliferation in pre-B-ALL via STAT5 activation. Pediatr. Blood Cancer. 50, 812–817. [DOI] [PubMed] [Google Scholar]
  21. Nelson, E.A. , Walker, S.R. , Weisberg, E. , Bar-Natan, M. , Barrett, R. , Gashin, L.B. , Terrell, S. , Klitgaard, J.L. , Santo, L. , Addorio, M.R. , Ebert, B.L. , Griffin, J.D. , Frank, D.A. , 2011. The STAT5 inhibitor pimozide decreases survival of chronic myelogenous leukemia cells resistant to kinase inhibitors. Blood. 117, 3421–3429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Pierau, M. , Na, S.Y. , Simma, N. , Lowinus, T. , Marx, A. , Schraven, B. , Bommhardt, U.H. , 2012. Constitutive Akt1 signals attenuate B-cell receptor signaling and proliferation, but enhance B-cell migration and effector function. Eur. J. Immunol.. 42, 3381–3393. [DOI] [PubMed] [Google Scholar]
  23. Polajeva, J. , Sjosten, A.M. , Lager, N. , Kastemar, M. , Waern, I. , Alafuzoff, I. , Smits, A. , Westermark, B. , Pejler, G. , Uhrbom, L. , Tchougounova, E. , 2011. Mast cell accumulation in glioblastoma with a potential role for stem cell factor and chemokine CXCL12. PLoS One. 6, e25222 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Sjostrom, M. , Jakobsson, P.J. , Juremalm, M. , Ahmed, A. , Nilsson, G. , Macchia, L. , Haeggstrom, J.Z. , 2002. Human mast cells express two leukotriene C(4) synthase isoenzymes and the CysLT(1) receptor. Biochim. Biophys. Acta. 1583, 53–62. [DOI] [PubMed] [Google Scholar]
  25. Strouch, M.J. , Cheon, E.C. , Salabat, M.R. , Krantz, S.B. , Gounaris, E. , Melstrom, L.G. , Dangi-Garimella, S. , Wang, E. , Munshi, H.G. , Khazaie, K. , Bentrem, D.J. , 2010. Crosstalk between mast cells and pancreatic cancer cells contributes to pancreatic tumor progression. Clin. Cancer Res.. 16, 2257–2265. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Stupp, R. , Mason, W.P. , van den Bent, M.J. , Weller, M. , Fisher, B. , Taphoorn, M.J. , Belanger, K. , Brandes, A.A. , Marosi, C. , Bogdahn, U. , Curschmann, J. , Janzer, R.C. , Ludwin, S.K. , Gorlia, T. , Allgeier, A. , Lacombe, D. , Cairncross, J.G. , Eisenhauer, E. , Mirimanoff, R.O. , 2005. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N. Engl. J. Med.. 352, 987–996. [DOI] [PubMed] [Google Scholar]
  27. Tran Thang, N.N. , Derouazi, M. , Philippin, G. , Arcidiaco, S. , Di Berardino-Besson, W. , Masson, F. , Hoepner, S. , Riccadonna, C. , Burkhardt, K. , Guha, A. , Dietrich, P.Y. , Walker, P.R. , 2010. Immune infiltration of spontaneous mouse astrocytomas is dominated by immunosuppressive cells from early stages of tumor development. Cancer Res.. 70, 4829–4839. [DOI] [PubMed] [Google Scholar]
  28. Uhlen, M. , Oksvold, P. , Fagerberg, L. , Lundberg, E. , Jonasson, K. , Forsberg, M. , Zwahlen, M. , Kampf, C. , Wester, K. , Hober, S. , Wernerus, H. , Bjorling, L. , Ponten, F. , 2010. Towards a knowledge-based human protein atlas. Nat. Biotechnol.. 28, 1248–1250. [DOI] [PubMed] [Google Scholar]
  29. Wang, X.B. , Tian, X.Y. , Li, Y. , Li, B. , Li, Z. , 2012. Elevated expression of macrophage migration inhibitory factor correlates with tumor recurrence and poor prognosis of patients with gliomas. J. Neurooncol.. 106, 43–51. [DOI] [PubMed] [Google Scholar]

Articles from Molecular Oncology are provided here courtesy of Wiley

RESOURCES