Total RNA isolation and RT-qPCR

Total RNA was isolated using RNA mini kit (Promega, Madison, WI, USA) according to the manufacturer's instructions. Quality and yield were determined by NanoDrop 1000 (PeqLabBiotechnologie, Erlangen, Germany). cDNA was synthesized using 100 ng total RNA with SuperScript II reverse transcriptase kit (Invitrogen, Waltham, MA, USA). RT-qPCR gene amplification was performed in duplicate using SYBR Green PCR mix (Applied Biosystems, Waltham, MA, USA) and 7500 Fast Real-Time PCR System (Applied Biosystems). Primer sequences used were: glycoprotein nonmetastatic melanoma protein B (Gpnmb; murine: sense 5′-AGAAATGGAGCTTTGTCTACGTC-3′, antisense 5′-CTTCGAGATGGGAATGTATGCC-3′; human: sense 5′-TGCGGTGAACCTGATATTCCC-3′, antisense 5′-GTCCTCTGACCATGCTGTCC-3′) and TATA-box binding-protein (Tbp; murine: sense 5′-AAGGGAGAATCATGGACCAG-3′, antisense 5′-CCGTAAGGCATCATTGGACT-3′; human: sense 5′-AGCGCAAGGGTTTCTGGTTT-3′, antisense 5′-CTGAATAGGCTGTGGGGTCA -3′). The results were analyzed by 2−ΔΔCT ways normalized to tbp and were presented as fold change normalized to control group, if not labeled differently.

Protein extraction and Western Blot

Whole-cell protein extracts were prepared from briefly cultured RCAS-PDGFb (≤ 7 passages) and GL261 glioma cells or primary naïve astrocytes using RIPA lysis buffer (Sigma-Aldrich, St. Louis, MO, USA) containing EDTA-free protease inhibitor cocktail tablets (Roche Diagnostics, Mannheim, Germany). Protein concentration was determined by a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA), and 20 µg of total protein of each sample was resolved on a 10% SDS-PAGE gel, followed by wet transfer of resolved proteins onto a PVDF membrane (GE Healthcare, Chicago, IL, USA). Membranes were blocked and followed by overnight incubation at 4 °C for murine GPNMB (goat, AF2330; R&D Systems, Minneapolis, MN, USA) and for murine GAPDH (mouse, ab8245; Abcam, Cambridge, UK). Membranes were incubated with a secondary for anti-rabbit HRP antibody at 1:2000 (#7074; Cell Signaling Technology) and for anti-goat with IRDye 680RD (925-68071; Li-Cor, Lincoln, NE, USA), developed with SuperSignal West Pico Chemiluminescence substrate kit (Thermo Fisher Scientific). Signal was detected by Molecular Imager Gel Doc XR system (Bio-Rad, Hercules, CA, USA).

Immunofluorescent staining and image processing

For murine tumor slices, mouse brains were harvested and perfused with phosphate buffered salt solution (PBS) followed by 4% paraformaldehyde (PFA) solution (Sigma-Aldrich); 40 µm free-floating tumor sections were prepared as previously described [33]. Human glioma specimens were prepared following the procedure previously described [33]. Slices were washed 3 times with PBS for 5 min and blocked with 5% of donkey serum and 0.1% Triton-X (Sigma-Aldrich). Primary antibodies were added overnight 1:100 for murine and human GPNMB (goat; AF2550; R&D Systems), 1:600 for IBA1 (goat; ab5076; Abcam), 1:600 IBA1 (rabbit; ab178847; Abcam), 1:100 for Ki67 (rabbit; ab16667; Abcam), 1:200 for MHCII (rat; 14-5321-81; Invitrogen), 1:100 for CD8A (rat; 14-0081-82; Invitrogen), 1:200 for Gzmb (rabbit; ab4059; Abcam), 1:200 for Foxp3 (rabbit; mAb #12653; Cell Signaling), 1:100 for CD3 (rat; 14-0032-81; Invitrogen), 1:200 for PD-1 (rabbit; ab214421; Abcam), 1:200 for CD44 (rabbit; 14-0441-81; Invitrogen) at 4 °C. As secondary antibody, we used 1:200 anti-rabbit IgG (711-545-152; Dianova, Hamburg, Germany), anti-goat IgG (705-605-147; Dianova), and anti-rat IgG (712-545-153; Dianova), 1:200 anti-rat IgG (Cy5-labelled; 712-175-150; Jackson Immuno Research, Cambridgeshire, United Kingdom) and anti-rabbit IgG (Cy3-labelled; 711-165-152 Dianova). Nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Images were taken using a confocal microscope (LSM710 & LSM700; Carl Zeiss, Oberkochen, Germany) with 10×, 20×, 40× or 63× oil objectives. Cell counting and area of staining was performed using Imaris software (Bitplane, Zürich, Switzerland) and Fiji [34]. For human tumor and control slices, tissue was cut at 5 µm sections and floated onto charged slides for preparation of immunohistochemistry.

Animals

Mice were housed in the animal facilities of the Max Delbrueck Center and handled according to governmental guidelines (LaGeSo G125/17). GPNMB−/− (GPNMB-KO; KO) mice were generated as described [35]. The first base after the start codon ATG of the gene GPNMB was deleted with Crispr-Cas9 technology in the C57BL/6N background strain (SI Methods). KO mice were held homozygously knockout for GPNMB alleles and were compared to the C57BL/6N (wildtype; WT) strain (Charles River, Sulzfeld, Germany) in animal experiments.

Generation of intracranial mouse gliomas

Ntv-a/Ink4a-Arf−/− mice develop pro-neural high-grade gliomas 6–8 weeks following intracranial injection of RCAS-PDGFb-producing DF-1 chicken fibroblast cells at 4.5–10 weeks of age [36]. RCAS-PDGFb tumors were dissociated as described below and intracranially re-transplanted into KO and WT mice. Injections were performed using a stereotactic frame (Stoelting, Wood Dale, IL, USA). Mice used for these experiments were 5–10-week-old (Ntv-a/Ink4a-Arf−/− mice for DF-1 RCAS-PDGFb injection), or 8–14-week-old (C57BL/6N WT or KO for RCAS-PDGFb tumor cell re-implantation). Mice were anesthetized with intraperitoneal injections of 0.1 mg/g ketamine (Pharmazeutischen Handelsgesellschaft, Garbsen, Germany) and 0.02 mg/g xylazine (Bayer, Leverkusen, Germany). Animals were also provided 0.25% Marcaine in the volume of about 0.1 ml/25 g administered right before the surgery, which provided pain relief from the sutures for 6–8 h. One microliter cell suspension (4 × 104 transfected DF-1 cells, or 5 × 104 RCAS-PDGFb tumor cells) was delivered using a 30-gauge needle attached to a Hamilton syringe (Hamilton; Reno, NV, USA). Coordinates for injections of DF-1 cells into Ntv-a/Ink4a-Arf−/− mice and RCAS-PDGFb tumor cells into WT or KO mice, respectively were bregma 1.5 mm anterior, Lat − 0.5 mm, and a depth 2.0 mm. Tumors for the GL261 mouse model were implanted as previously described [16].

Analysis of disease progression

RCAS-PDGFb tumor-bearing wild type and KO mice were monitored for signs of disease progression using a score from 1 (no symptoms) to 5 (severe disease). Score 1: No neurological deficit; when lifting by the tail, front paws remain on the ground and animal resists strong pull. Score 2: Shaggy coat with persistent grooming; when lifting by the tail, front paws remain on the ground and animal resists moderate pulling. Score 3: Weight loss 5%; unsteady walking pattern; when lifting by the tail, front paws remain on the ground and animal resists slight pull. Score 4: Increasing weight loss up to 15%; delayed reaction to touch and sounds; resistance of the front paws when lifting not present; low curvature of the back. Score 5: Eyes mostly closed; blindness; weight loss 20% of initial weight; no reaction to touch or sound; hunchback. A mouse was euthanized whenever it reached a score between 3 and 4. Monitoring of the mice occurred daily; weighing and behavioral assessment reflecting the mice’s health condition started on day 21 post surgery.

Tumor volume quantification

WT and KO mice were inoculated with RCAS-PDGFb cells and sacrificed together whenever the first mouse reached a disease score between 3 and 4. Brains were fixed after perfusion and cut into 40 μm thick slices. Slices were stained with DAPI and antibodies against IBA1 and GPNMB. The tumor volume was reconstructed through the Cavalieri method by measuring the RFP+ area in every 12th tumor slice.

Bioinformatic analysis (TCGA, CGGA)

GlioVis (http://gliovis.bioinfo.cnio.es) was used to access tumor gene expression data of GBM patients for GPNMB, macrophage and glioma/astrocyte marker genes (Study: Glioblastoma, mRNA Expression z-scores HG-UG133A. The dataset from Klemm at el. [18] was made available by the authors through their open access platform: https://joycelab.shinyapps.io/braintime/. For correlationship evaluation between TAMs infiltration level and gene expression, TCGA GBM RNA-seq datasets were processed via TIMER platform (http://timer.comp-genomics.org/) [37].

Human material

All patients were operated at the Department of Neurosurgery, University Medical Center Schleswig–Holstein, Campus Kiel. The study was approved by the Ethics Committee of the University of Kiel (approval #D477/18) and was in accordance with the Helsinki Declaration of 1964 and its later amendments. Informed consent was obtained from all individual patients. For non-tumor samples, tissue resected from trauma patients undergoing brain surgery were used. Freshly resected tumor and non-tumor tissue was stored in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA, USA) at 4 °C for < 24 h until further experimental workup. Samples were either fixed in 4% PFA for 24 h and mounted in tissue tec at − 20 °C or prepared for Cell sorting via MACS isolation (described below).

Cell isolation

In tumor-bearing mice brains, only the visible tumor area around the injection site was used. Human tissue was dissociated with the Brain Tumor Tissue Dissociation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. Subsequently, the cell pellet was resuspended in sorting buffer for subsequent magnetic-activated cell sorting (MACS).

Murine tissue was mechanically dissociated as described previously [38]. To remove myelin we followed a protocol published by Guneykaya et al. [39] In brief, the brain cell suspension was mixed with a total of 25 ml of a 22% Percoll (Th.Geyer, Renningen, Germany) solution and a layer of 5 ml cold PBS (Gibco-Invitrogen) was added on top. Centrifugation at 950 g with slow acceleration and without breaks created a gradient that separated the cell pellet on the bottom of the tube from the myelin which was carefully aspirated. For the isolation of TAMs from RCAS-PDGFb tumors-bearing WT mice a 30%/70% Percoll gradient was used. After 25 min of centrifugation at 800 g TAMs and naïve microglia were enriched at the 30%/70% interphase. Cells were collected, washed once with PBS, and subsequently centrifuged again at 300 g for 10 min. Subsequently, the cell pellet was resuspended in sorting buffer for subsequent flow-cytometry.

Spleens were processed through a 70 μm cell strainer with a syringe plunger and the mesh rinsed with 10 ml of PBS per spleen. The cells were centrifuged and the pellet subjected to erythrocyte lysis by adding 5 ml of 1 × RBC lysis buffer (Biolegend, San Diego, CA, USA). The lysis was carried out by shaking the tube mildly for 5 min at RT and subsequently stopped with 20 ml of PBS. The pellet was washed once with PBS and resuspended in PBS, containing 0.5% FCS and 2 mM EDTA (FACS buffer) for subsequent fluorescence activated cell sorting (FACS) isolation.

MACS sorting

The CD11b+ samples for the microarray were generated using MACS. Following Percoll gradient centrifugation, tumor and control cell pellets were resuspended in PBS, containing 0.5% fetal calf serum (FCS) and 2 mM EDTA and labeled with anti-CD11b microbeads (Miltenyi Biotec). The MACS isolation was carried out according to the manufacturer’s instructions and cells were subsequently used for RNA isolation.

FACS analysis

FACS was employed using CD11b, CD45, Ly6c, Ly6G and GPNMB to determine the GPNMB+ population of non-immune cells (CD45−CD11b−), microglia or TAMs (CD45+CD11b+Ly6G−Ly6c−), monocytes with low Ly6c expression (CD45+CD11b+Ly6G−Ly6clow), monocytes with high Ly6c expression (CD45+CD11b+Ly6G−Ly6chigh), neutrophils (CD45+CD11b+Ly6G+Ly6c+) and lymphocytes (CD45+CD11b−). Following Percoll gradient centrifugation, cell pellets were resuspended in FACS buffer (containing 2% FCS) and stained with 2 μl of dye-coupled antibodies per 1 × 107 cells. The staining was performed with CD45-eFluor450 (48-0451-82), CD11b-Pe-Cyanine7 (25-0112-82), Ly6c-PE (12-5932-80), Ly6G-FITC (11-5931-82), and GPNMB-eFluor660 (50-5708-82; all eBioscience, San Diego, CA, USA) for 30 min at 4 °C. Thereafter, the cells were washed and resuspended in 500 μl FACS buffer per 5 × 106 cells for sorting at a Fortessa (BD Bioscience). Compensation was calculated with single-stained beads (552844; BD Bioscience, Franklin Lakes, NJ, USA) and unstained cells.

Cell culture

Cells of the murine glioma cell line GL261 (National Cancer Institute, MD, USA) were grown in DMEM with 10% FCS, 200 mM glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin (all from Invitrogen). DF-1 cells were purchased from ATCC (Manassas, VA, USA). Cells were grown at 39 °C according to ATCC instructions. Transfections with RCAS-PDGFb were performed using Fugene 6 transfection kit (no. 11814443001; Roche, Mannheim, Germany) according to manufacturer’s protocol. Primary microglia were prepared from neonatal and adult C57BL/6 as previously described [12]. Briefly, cortical tissue of neonatal mice was freed of blood vessel and meninges in Hank’s balanced salt solution (HBSS), and digested in 1% trypsin and 0.05% deoxyribonuclease for 5 min at room temperature. Cells were then plated in 75-cm2 flasks. After 7 days, cells were treated with L929 conditioned medium. Microglia were then shaken off and replated. All cells were maintained in a 37 °C incubator with a 5% CO2 humidified atmosphere.

Tumorsphere culture

RCAS-PDGFb tumors were excised from tumor brains using a scalpel, minced, and incubated with Accutase (eBioscience, San Diego, CA, USA) for 15 min at 37 °C. Tissue pieces were mechanically dissociated using a 1 ml pipette and washed in DMEM. Cells were passed through a 70 μm cell strainer and seeded into a T25 cell culture flask. Cells were grown in GIC medium containing DMEM-F12 GlutaMAX (GIBCO-Invitrogen, Carlsbad, CA, USA), 1% penicillin G/streptomycin sulfate (Sigma-Aldrich), 1:50 B-27 without vitamin A (GIBCO-Invitrogen), 0.2 mM HEPES (Sigma-Aldrich), 20 ng/ml insulin (Sigma-Aldrich), supplemented with 20 ng/ml fibroblast growth factor 2 (FGF2; Cell Systems, Kirkland, WA, USA) and 20 ng/ml epidermal growth factor (EGF; Cell Systems).

Organotypic brain slice (OBS) model and tumor inoculation

OBSs were prepared as described previously [15]. Briefly, 14-day-old WT or KO mice were decapitated, and brains were cut in coronal plane into 250 µm sections with a vibratome (Leica Microsystems, VT1000S, Wetzlar, Germany). Brain slices were collected with a sterile plastic pipette (7 mm diameter) and transferred onto cell culture inserts with 0.4 µm pores (Becton Dickinson, Franklin Lakes, NJ, USA), which were fitted into wells of a 6-well plate; 1 ml of culture medium containing DMEM supplemented with 10% heat-inactivated FCS, 0.2 mM glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin was added into each well; after overnight incubation, medium was changed with cultivation medium containing 25% heat-inactivated FCS, 50 mM sodium bicarbonate, 2% glutamine, 25% Hanks balanced salt solution, 1 mg/ml insulin (Invitrogen), 2.46 mg/ml glucose (Braun, Melsungen, Germany), 0.8 mg/ml vitamin C (Sigma-Aldrich), 100 U/ml penicillin, 100 mg/ml streptomycin, and 5 mM Tris in DMEM (all from Invitrogen). 5000 mCherry-GL261 or RFP-labelled RCAS-PDGFb cells were slowly injected into the caudate putamen region of the slice in 150 µm depth of both hemispheres using a Hamilton syringe. Careful control of the injection procedure ensured that no cells spilled onto the surface of the slice, which could migrate over the surface rather than invade through the tissue. After 4 days, mCherry-GL261 or, after 6 days, RFP-labelled RCAS-PDGFb slices were washed and fixed with 4% PFA. Tumor volumes were measured by confocal microscopy (LSM710; Carl Zeiss) with z-stack scanning and were reconstructed by Imaris into 3D model for exact volume evaluation.

Softwares and statistics

Graphs were created using GraphPad Prism 7 & 10 and were analyzed using an unpaired parametric 2-tailed t-test, assuming equal standard deviations. One-way ANOVA was used in experiments having more than one group to compare with controls. Linear regression was used to compare curves of disease progression. Test details are included in appropriate figure legends. Pearson correlation was calculated using GraphPad Prism 7. Imaris Version 9.3.1 was used.