This article is mentioned in:

Introduction

Despite advances in surgical techniques, radio- and chemotherapy, the outcome of patients suffering from glioblastoma multiforme (GBM) remains poor, with a median survival of <15 months (1). In recent years, immunotherapeutic approaches to malignant glioma have advanced rapidly. This is based on findings revisiting the traditional concept of the central nervous system (CNS) as an immunoprivileged locus. The discovery of the meningeal lymphatic system (2) and an improved understanding of brain T cell trafficking into the brain via the leptomeninges (3) and the blood-brain-barrier (4) represent important communication channels between the CNS and the peripheral immune system.

Whereas research has focused on T-cells as the critical component of the specific antigen-mediated antitumor response in malignant glioma, there is increasing evidence that non-specific local immunotherapies may aid in glioma defense. Clinical observations that postoperative infections within or close to tumor sites promote a prolonged survival or even complete remission in patients with GBM (5–9) were supported by results of two retrospective single-center studies (10,11) that reported an effect of non-specific systemic immune responses on glioma growth and surveillance. This has been confirmed experimentally with a novel approach using the local administration of heat-inactivated staphylococci as potent immunomodulators in an experimental gliosarcoma model, which led to oncolysis and prolonged survival associated with a distinct peri- and intratumoral infiltration of macrophages (3).

In contrast to the effects of a topical application, it is still debated if systemic immunostimulation exerts a significant antitumor effect. This is the basis and potential key for immunotherapies. Non-specific local immunotherapies have proven to be ineffective, whereas specific local immunotherapies have suggested a response (12). Non-specific systemic immunotherapies have not been validated in studies, but numerous suggestions were made regarding effects in allergies and brain abscesses (13). Host factors, including the tumor microenvironment, considerably influence glioma growth and targeting angiogenetic and inflammatory properties recently evolved as an effective treatment strategy (14).

In 2004, spontaneous regression was observed in animal studies with experimental gliomas (15). Spheroids of C6 cells were implanted into the brain and tumor growth was measured by magnetic resonance imaging (MRI)-based volumetry. The aim of the current study was to evaluate effects of a stimulation of the innate immune response, particularly of macrophages, on the proliferation of glioblastoma. For that purpose, recombinant human granulocyte macrophage colony stimulating factor (rhGM-CSF), which has been established in the treatment of neutropenia with a known safety profile, was used in this study.

Materials and methods

Tumor cell culture and generation of 3D spheroids

Multicellular spheroids from C6 tumor cells (CLS Cell Lines Service GmbH, Eppelheim, Germany) were generated by seeding cells in 75 cm2 culture flasks filled with Dulbecco's modified Eagle's medium (DMEM). Flasks were base-coated with 1% noble agar (Difco; BD Biosciences, Franklin Lakes, NJ, USA) dissolved in medium. Incubation was performed at 37°C in the presence of 5% CO2 with 100% humidity. Spheroids were screened for signs of necrosis using inverted light microscopy. Following 6 days, spheroids with a diameter of 200–300 µm without a necrotic core were selected for implantation.

Animal preparation, operative procedure and stimulation of systemic macrophages

The animal care committee of the district authorities (Regierung von Unterfranken, Veterinärwesen, Bavaria, Germany; AZ 55.2–2531.01-65/10) approved the experimental procedures. The implantation procedure has previously been described in detail (15,16). For the current study, 29 male Sprague-Dawley rats (weight, 250–300 g) were anesthetized by an intraperitoneal (i.p.) injection of ketamine hydrochloride (Ketavet™; Pfizer, Inc., New York, NY, USA) and xylazine hydrochloride (Rompun™; Bayer AG, Leverkusen, Germany). The animals' heads were fixed in a stereotactic frame using non-perforating bars, a midline incision of the scalp was performed and a 2 mm burr hole was placed 2 mm left of the bregma. Following excision of the dura, the cortex was incised in a semicircular fashion using a microscalpel. Under microscopic view, a single spheroid was then placed 2–3 mm subcortically. Following surgery, animals were housed using a 12-h dark/light cycle with free access to food and water and were monitored for signs of discomfort or neurological abnormalities daily. Of the 29 animals implanted with C6 glioma spheroids, 20 rats received a subcutaneous injections of 10 µg/kg rhGSM-CSF every other day. Nine animals served as a control group.

MRI and tumor volumetry

MRI exams were performed on postoperative days (POD) 7, 14, 21, 28, 32 and 42 with a 3 T clinical scanner (Magnetom Trio®; Siemens Healthineers, Erlangen, Germany) using a round surface coil. The following sequences were performed: T1 TSE cor (0.9 mm), cor T2 TSE rs (0.7 mm) and cor T2 CISS 3D (0.3 mm). Animals were anesthetized using Ketavet® as described and were then administered 0.1 ml contrast agent i.p. 10 min prior to MRI examination. Tumor volumes were calculated using the T2 3D CISS sequences by using MRI Convert® and MIPAV® software.

Tissue preparation

Following each MRI exam, two randomly selected animals were sacrificed for histological examination. Brains and spleens were removed and immediately fixed in paraformaldehyde solution for 24 h and stored in cold PBS (pH 7.4; 4°C) for one week prior to paraffin embedding. For histological studies, spleen sections and coronal brain sections cut into 4-µm slices.

Histological and immunohistochemical analyses

Hematoxylin and eosin (H&E) staining was performed for an estimation of the gross morphology of tumors and spleen sections.

For immunohistochemistry, sections were stained with Ki67 and CD8 antibodies targeting CD8+ lymphocytes (AK CD8α; eBioscience; Thermo Fisher Scientific, Inc., Waltham, MA, USA; cat. no. 14-0084; AK CD68, Zymed; Thermo Fisher Scientific, Inc.; cat. no. 603–2210; AK Ki67; Zymed; Thermo Fisher Scientific, Inc.; cat. no. RMPD 004). Omission of primary antibodies in the control experiments resulted in the expected absence of any cellular labeling. The extent of infiltration of different immune-cell subsets was quantified by cell counting of five representative high-power fields (HPF) in each section, including the tumor margins.

In order to estimate the putative effects of rhGM-CSF stimulation, macrophages and CD8+ lymphocytes in tumor tissues were counted and the corresponding spleen tissue functioned as positive control. Sections were counted at a magnification of ×100 using a microscope. In each case, five contiguous fields of view were counted and the mean was determined. For brain sections, five visual fields were counted starting from the tumor margin. As ED1 stains macrophages and microglia, only cells with distinct phagocyte morphology were considered. In CD8+ cells, only cells with clear lymphocyte morphology were included.

Data analysis

Survival time is presented using box plots. Tumor volumes and cell counts are presented as the median ± standard deviation. Due of the small sample size, all analyses were of explorative nature. The results (survival times, tumor growth, cell counts) were evaluated graphically. Animals surviving 42 days without MRI-based evidence of tumor growth were excluded from the analysis. Overall survival was assessed by Kaplan-Meier analysis and differences between survival curves were calculated by using Graph Pad Prism® (GraphPad Software, Inc, La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference. Student's t-tests were used in pairwise comparisons.

Results

MRI studies

rhGM-CSF group

On POD 7, 2/20 (10%) animals treated with rhGM-CSF developed a visible tumor. Two animals were sacrificed on POD 7 for analysis and further two rats were excluded from the trial due to anesthesia-associated complications, leaving 16 animals for evaluation on POD 14. At that time, solid tumors were observed in 11/16 (65%) animals and on POD 21, tumors were visible in 14/15 animals. A further animal exhibited a visible tumor on POD 28. Solid tumors developed in 15/20 (75%) animals. Tumor regression was observed in six (30%) animals. As a result of severe symptoms associated with tumor growth, seven (35%) animals of the rhGM-CSF group were sacrificed during the trial.

An exemplary course of tumor growth is presented in Fig. 1. On POD 7, the first extracranial fluid accumulation caused by the surgical trauma was detected. On POD 14, minimal tumor growth with contrast uptake at the edges of the surgical cavity was observed. An increase in size caused peritumoral edema. On POD 28, a midline shift to the opposite side developed due to a mass effect. In addition, the central sparing of contrast enhancement representing necrotic changes was accompanied by a decrease of tumor volume. The peritumoral edema disappeared and midline shift was no longer visible. A small defect with a fading contrast enhancement remained at POD 32 and 42.

Figure 1. Magnetic resonance imaging performed with a 3 T clinical scanner. T1 with 0.1 ml contrast medium intraperitoneal: POD 7–42 results are presented. POD 7, extracranial fluid accumulation; POD 14, minimal tumor growth and contrast medium uptake; POD 28, midline shift due to increasing tumor mass accompanied by necrotic changes; POD 32 and 42, decreasing tumor volume, disappearing peritumoral edema and midline shift. Small defects with a remaining fading contrast enhancement. POD, postoperative day.

Effects of rhGM-CSF on tumor growth and survival are documented in Figs. 2 and 3. Until POD 20, no significant differences were determined. The mean survival time was 35 days in the rhGM-CSF.

Figure 2. Tumor growth in the rhGM-CSF and control groups. On postoperative day 28 only one measurement of the control group was evaluated and no statistical comparisons were performed. rhGM-CSF, recombinant human granulocyte macrophage colony stimulating factor.

Figure 3. Effects of rhGM-CSF on overall survival. Treatment prolonged survival of animals in the rhGM-CSF group significantly compared with the control (35 vs. 24 days; P=0.0343). rhGM-CSF, recombinant human granulocyte macrophage colony stimulating factor.

Control group

On POD 7, small tumors were visible in 3/9 (30%) animals. On POD 14, tumor growth was observed in all control animals. A total of 6/9 (66%) animals were sacrificed due to severe symptoms of tumor growth. A total of 3 animals were used for histological examination. On POD 28, one animal remained, which was then sacrificed due to tumor mass and occurring symptoms. In contrast to the rhGM-CSF-group, none of the control animals exhibited spontaneous regression (Fig. 2). On POD 28, only one measurement was obtained for the control group and no statistical comparison was performed.

Comparison of treatment and control groups

Treatment with rhGM-CSF significantly prolonged survival in the rhGM-CSF group compared with the control group (35 vs. 24 days; P=0.0343). Tumor volume measurements suggested a delayed growth onset in the rhGM-CSF group compared with the control and a reduced median volume (134 vs. 262 mm3).

Histology

Spheroids were orthotopically implanted and in 15/20 animals tumor growth was observed. In addition, some animals exhibited extracranial tumor components. H&E staining revealed typical growth characteristics of GBM-type neovascularity, necrosis and palisade-type arrangement of the tumor cells (Fig. 4).

Figure 4. Hematoxylin and eosin staining. Typical growth characteristics of glioblastoma multiforme; neovascularity necrosis and palisade type arrangement of tumor cells are presented. Scale bar, 200 µm.

CD8+ cells, including cytotoxic T-cells, were monitored to reveal potential effects on tumor growth. There was no significant difference in CD8+ positive cells in the two groups (Fig. 5).

Figure 5. Effects of CD8+ cells on tumor growth in the rhGM-CSF and control groups. No significant difference in CD8+ cells was observed (P>0.05). rhGM-CSF, recombinant human granulocyte macrophage colony stimulating factor. CD, cluster of differentiation; rhGM-CSF, recombinant human granulocyte macrophage colony stimulating factor.

CD68 staining revealed brown-colored macrophages that were located in necrotic areas and in solid tumor tissues (Fig. 6). In addition, numerous macrophages were grouped around tumor vessels (Fig. 7).

Figure 6. CD68 staining of solid tumor tissues. Macrophages were presented with brown coloring. The arrows indicate macrophages in the solid tumor tissues. Scale bars, 200 µm. CD, cluster of differentiation.

Figure 7. CD68 staining of macrophages around tumor vessels. Macrophages were presented with brown coloring. The arrow indicates macrophages that were grouped around tumor vessels. Scale bar, 50 µm. CD, cluster of differentiation.

Analyzing the time course of macrophage invasion, the maximum number was observed when regression started (Fig. 8). Decreasing tumor size led to decreasing numbers of macrophages. Compared with the control group, rhGM-CSF animals presented a significantly higher numbers of macrophages in brain slices (P=0.0275).

Figure 8. Course of macrophage invasion. A maximum was reached at tumor regression initiation. Decreasing tumor size led to decreasing numbers of macrophages. Compared with the control, rhGM-CSF-treated animals presented a significantly higher number of macrophages in brain slices (P=0.0275). rhGM-CSF, recombinant human granulocyte macrophage colony stimulating factor.

Discussion

In the current study, tumor tissues of rhGM-CSF animals exhibited significantly higher numbers of macrophages in five-fields-of-view compared with the controls. In addition, animals of the experimental group survived significantly longer compared with animals of the control group. Based on the knowledge that macrophage invasion is correlated with tumor growth in glioblastoma and experimental models, the current data describing reduced tumor growth associated with increased macrophage invasion suggested that the pharmacological induction of macrophages may attenuate or inhibit tumor growth.

A rat brain tumor implantation model was established to investigate macrophage infiltration and tumor growth in rhGM-CSF-treated animals. Compared with mouse models, the larger size of the rat brain facilitates a more precise intracerebral implantation of tumor cells and better in vivo imaging and volumetry of the growing mass (15). Among the various rat brain tumor models, the C6 glioma model has been extensively used and is the best characterized experimental model to investigate a wide array of biological properties of glial tumors, including, their mechanisms of invasion and angiogenesis, intratumoral signaling pathways and efficacy of novel therapeutic modalities (15,17). The C6 glioma model was originally developed by Benda et al (18) by repetitive administration of methylnitrosourea to Wistar rats and was further successfully established in Long-Evans and Sprague-Dawley rats (19,20). C6 cells share certain morphologic features of human glioblastoma-type pleomorphic cells, including nuclear polymorphism, a high mitotic index, areas of necrosis and invasion into the surrounding brain tissue (21). A limitation of the model is its immunogenicity. Several rat strains challenged with C6 cells develop a vigorous immune response. Hence, studies on immunotherapies using the C6 model require careful interpretation (22).

In brain tumors, high numbers of macrophages are detected, increasing with the degree of malignancy (23,24). A total of 5–30% are typified as microglia/macrophages and account for the majority of infiltrating immune cells in gliomas (25). The majority of studies does not distinguish between microglia and systemic macrophages (CD68 and CD11b/c positive) (26–28). Badie et al (23) have reported that a differentiation of the CD11b/c positive cells is possible through quantification due to low expression of CD45 in microglia and increased expression in macrophages (1,23,29). Initially, it was assumed that high numbers of microglial cells indicate a strong antitumor response in gliomas; however, certain studies postulated a positive effect on tumor growth and a support of the immunosuppressive peritumoral region (23,30). There are various synergistic mechanisms between glioma cells and microglia/macrophages that ensure an immunosuppressant milieu. On the one hand, microglia express low levels of major histocompatibility complex class II (MHC II) in the vicinity of the glioma. The production of transforming growth factor (TGF) β by glioma cells causes downregulation of MHC class II expression (31). These molecules are essential in the interaction of antigen-presenting cells and T-lymphocytes (32–34). On the other hand, glioma cells produce substances with immunosuppressive effects. These substances include interleukin (IL) 4, IL 6, TGF β [21] and prostaglandin (PG) E2 (32).

Further studies reported a positive effect of high numbers of tumor-macrophages (35,36). Galarneau et al (37) analyzed the influence of macrophages on the growth of glioma cells using a transgenic mouse model. Depletion of the macrophages leads to an increase of tumor volume by ≤33%. It is a well-known fact that macrophages are able to recruit T cells by release of tumor necrosis factor (TNF) (38). However, it remains to be elucidated whether effects on tumor growth are driven by macrophages alone or through TNF-mediated activation of T cells. In addition, a lower vessel density of ~12% in macrophage-depleted animals was observed. Thus, it is unlikely that increased tumor growth is associated with increased vascular supply (27). Furthermore, Villeneuve et al (38) struggled to determine whether an altered vascular supply in the tumor is caused by the depletion of macrophages.

To determine effects of rhGM-CSF treatment, the M1 and M2 status of macrophages has to be considered. Macrophages present in two different forms, the M1 and M2 status (32,33,39). The M1 status describes classically-activated macrophages associated with inflammation (22,29). In the presence of various cytokines, including rhGM-CSF and interferon γ, monocytes develop to macrophages with M1 status (33). Expression of signal transducer and activator of transcription (STAT) 1, M1 macrophages exhibit antimicrobial, immunostimulatory and antitumor functions (40). TNF is another factor that contributes to the activation and recruitment of microglia/macrophages. Production of this proinflammatory cytokine contributes to recruitment (33). The M2 status arises from alternative activation and describes macrophages under normal conditions. Among other factors, this status is responsible for preventing excessive immune reactions and explains the more immunosuppressive character of M2 (38). The conversion from M1 to M2 status is induced by IL 4, IL 6, IL 10 and M-CSF. Expression of STAT3 causes M2 macrophage activation, including tissue repair and support of angiogenesis, and favors a tumor progression. This is mediated by the release of various compounds, including IL 10 and TGF (33,41).

During the early stages of glioma development, macrophages are arrested in the M1 tumor suppressive status. The percentage of microglia/macrophages arresting in M2 correlates with the histological grade of malignancy in glioma (35,37,39). Certain mediators produced by the tumor induce the conversion from M1 to M2; these include IL 4, IL 10, TGF β and M-CSF (35,39,40,42). M-CSF produced by tumor cells increases M2 microglia and macrophages and promotes tumor growth and proliferation (40). Reduced tumor growth is detected when inhibiting the M2 status (42,43). Macrophage activation by rhGM-CSF may be a novel therapeutic approach in preserving the M1 status for prolonged periods, preventing a transition into the M2 status (43). Microglial cells treated with rhGM-CSF exhibit upregulated growth and form a heterogeneous cell population in vitro, similar to macrophages. Their function is strongly associated with local environmental factors, such as rhGM-CSF and M-CSF (30,44).

Expression of the Fas-ligand enable macrophage/microglia to support the immunosuppressive environment of gliomas (41,45). There is evidence that cancer stem cells represent the major cause for reprogramming microglia/macrophages to adapt the immunosuppressive M2 status. Tumor stem cells produce soluble (s) CSF, macrophage cytokine 1 (MIC1) and TGF β1. These substances polarize microglia/macrophages in the M2 status, block phagocytosis and the production of immunosuppressive cytokines is induced, including IL 10 and TGF β1. In addition, inhibition of T-cell proliferation is observed (41,45,46). Macrophages exhibit tumor suppressive functions at later stages in the development of gliomas. Changes may be influenced pharmacologically during this period (41,47). It has been postulated that the therapeutic approach using tumor-associated macrophages is more successful with a staggered induction (43). Considering that different means of activation trigger various responses explains the comparatively late tumor growth observed in a number of rats of the GM-CSF group in the current study. A novel therapeutic approach to treating GBM via targeting macrophages at different states of activity may be considered in the future (32,39,40). Additionally, it has to be clarified, which type of macrophages is promoted by administration of rhGM-CSF that further stimulates tumor progression (35,48). Depletion of rhGM-CSF led to a reduced number of tumor-promoting macrophages and interfered with the development of proinvasive macrophages. In animal studies, rhGM-CSF-depleted animals exhibit increased overall survival (48). In contrast, Grabstein et al (49) described enhanced tumor suppressing properties of macrophages by stimulating with rhGM-CSF in vitro and noted that macrophages exhibit improved antitumor activities.

Effects of rhGM-CSF on an immune system with tumor control are associated with the dose (50). High systemic concentrations lead to recruitment of regulatory T cells, which contribute to the immunosuppressive environment and interfere with the response against the tumor exerted by the immune system itself. Low doses in the context of vaccinations with tumor antigens lead to an immune stimulation and enhance antitumor activity (50,51). To reduce inter-individual differences, investigations were performed using genetically identical mice. It is postulated that rhGM-CSF at high doses reduces the expression of M-CSF receptor, in contrast to a promotion of receptor expression that is observed at low doses (52,53). In humans, treatment with defined amounts of GM-CSF lead to varying serum levels of M-CSF receptor due to individual variance (51).

Data presented in the current study supported the hypothesis of an antitumor effect of rhGM-CSF. The number of macrophages counted in tumor tissues of the rhGM-CSF group was increased compared with the control group. An analysis of macrophage counts and tumor volume over time produced supporting results; initially an increase was observed that towards the end of the trial almost reached baseline levels. It was observed that rhGM-CSF increased macrophage accumulation and exhibited a positive effect on tumor suppression.

In conclusion, a systemic stimulation of macrophages by rhGM-CSF led to significantly reduced and delayed tumor growth in a rodent C6 glioma model. Model being aware that studies on immunotherapies using the C6 rat model have to be interpreted carefully because of its immunogenicity.

The host control of experimental gliomas by macrophages may be combined with other promising immune-based approaches, including chimeric antigen receptor T-cell technology or PD-1/PD-L1 checkpoint inhibitors. The role macrophages serve in host tumor control may have been underestimated in the past.

Acknowledgements

Not applicable.

Funding

No funding was received.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

TL and AJ performed the experiments to collect the data. TL drafted the manuscript. CMM performed the histological examinations. AFK, TW made substantial contributions to conception and design and analyzing the data. ML made substantial contributions to conception and design, was involved in drafting the manuscript, revised it critically for important intellectual content and gave the final approval of the version to be published. GHV and RIE made substantial contributions to conception and design, interpretation of the data, were involved in drafting the manuscript and revised it critically for important intellectual content. GH performed the neuroradiological examinations. ML was involved in drafting the manuscript and gave final approval for publication. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The experimental procedures were approved by the Animal Care Committee of the district authorities (Regierung von Unterfranken, Veterinärwesen, Bavaria, Germany; AZ 55.2–2531.01-65/10).

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References