The novel CXCR4 antagonist,PRX177561,reduces tumor cell proliferation and accelerates cancer stem cell differentiation in glioblastoma preclinical models

Abstract

Glioblastoma is the most frequent and the most lethal primary brain tumor among adults. Standard of care is the association of radiotherapy with concomitant or adjuvant temozolomide. However, to date, recurrence is inevitable. The CXCL12/CXCR4 pathway is upregulated in the glioblastoma tumor microenvironment regulating tumor cell proliferation, local invasion, angiogenesis, and the efficacy of radio-chemotherapy. In this study, we evaluated the effects of the novel CXCR4 antagonist, PRX177561, in preclinical models of glioblastoma. CXCR4 expression and PRX177561 effects were assessed on a panel of 12 human glioblastoma cells lines and 5 patient-derived glioblastoma stem cell cultures. Next, the effect of PRX177561 was tested in vivo, using subcutaneous injection of U87MG, U251, and T98G cells as well as orthotopic intrabrain inoculation of luciferase-transfected U87MG cells. Here we found that PRX177561 impairs the proliferation of human glioblastoma cell lines, increases apoptosis, and reduces CXCR4 expression and cell migration in response to stromal cell–derived factor 1alpha in vitro. PRX177561 reduced the expression of stem cell markers and increased that of E-cadherin and glial fibrillary acidic protein in U87MG cells consistent with a reduction in cancer stem cells. In vivo, PRX177561 reduced the weight and increased the time to progression of glioblastoma subcutaneous tumors while increasing disease-free survival and overall survival of mice bearing orthotopic tumors. Our findings suggest that targeting stromal cell–derived factor 1 alpha/CXCR4 axis by PRX177561 might represent a novel therapeutic approach against glioblastoma and support further investigation of this compound in more complex preclinical settings in order to determine its therapeutic potential.

Keywords

Glioblastoma CXCR4 angiogenesis monocyte infiltration

Introduction

Glioblastoma (GBM), also known as glioblastoma multiforme, is an aggressive and highly heterogeneous brain tumor¹ characterized by high local invasiveness, extensive necrosis, and dysregulated vascularization. Despite improvements in radiation therapy and chemotherapeutics, outcomes of this neoplasia remain poor, with a median survival of 15 months when radiation therapy and adjuvant chemotherapy (standard of care) were used.² It has been demonstrated that GBM contains an elevated percentage of transformed, self-maintaining, multipotent, tumor-initiating cancer stem cells (CSCs)^3–5 that are mainly present in highly hypoxic areas with palisading necrosis. GBM are also surrounded by infiltrating lymphocytes/monocytes and glioma-associated microglia/macrophages (GAMs). The presence of these cells influences the malignancy of GBM,^6,7 supporting tumor structure, angiogenesis, growth, and invasiveness, as responses to molecular regulators (growth factors and cytokines) secreted by tumor cells.^8,9 These cytokines can instruct GAMs to downregulate the M1 pro-inflammatory (potentially anti-tumor) functions and to acquire a M2 anti-inflammatory, immunosuppressive, and pro-angiogenic phenotype.^10–14 CXCL12/stromal cell–derived factor 1 (SDF-1α) and its receptor CXCR4, a G protein–coupled receptor initially linked with leukocyte trafficking and with HIV infection,¹⁵ are widely implicated in these phenomena. High expression levels of CXCL12 and CXCR4 generally have negative prognostic significance.^16–18 CXCR4 is overexpressed in more than 50% of the primary GBM tissues¹⁹ and in the majority of GBM cell lines.²⁰ Moreover, low-grade, astrocytic tumors are rarely positive for SDF1α whereas all GBMs showed moderate to intense immunostaining with SDF1α, with particularly intense staining in the pseudopalisading cells and the proliferating microvessels.²¹ In addition, CXCR4 overexpression was reported in GBM stem cells.^22,23 It has been also reported that high SDF1α levels in the tumor may attract CXCR4-positive vascular and inflammatory cells that, once within the tumor, secrete tumor promoting cytokines as well as growth and pro-angiogenic factors.^24–26 Pseudopalisades and proliferating microvessels are two markers associated with accelerated tumor growth. Several preclinical studies have demonstrated that the inhibition of SDF1α/CXCR4 pathway reduced tumor growth and vasculogenesis as well as tumor recurrence after radiotherapy (RT). These observations highlighted the role of CXCR4 as a promising strategy for GBM therapy to be associated with RT.^27–29 In addition, it has been demonstrated that CXCR4 antagonism reduced both endothelial sprouting from nearby non-irradiated blood vessels and tumor vasculature regrowth from surviving endothelial cells in and around the tumours.^30–33 SDF1α/CXCR4 pathway also participates in the recruitment of pro-angiogenic CD11b+ monocytes and bone marrow mononucleated cells.^25,33 The first and most studied CXCR4 antagonist, AMD3100 (Plerixafor), represents a clinically advanced compound^34,35 and it has been reported to inhibit the growth of GBM³⁶ and reduce GBM stem cell survival³⁷ in preclinical studies. AMD3100 is an allosteric agonist of CXCR7, the second SDF1α receptor.³⁸ AMD31000 also shows elevated cardiotoxicity associated with some other adverse events after long-term usage of drug,³⁹ prompting the search for new safer and selective CXCR4 inhibitors suitable as anti-GBM agents.

The experiments performed in this report were designed to assess the effect of CXCR4 blockade by PRX177561,³¹ a novel, highly potent, and selective compound, which penetrates the brain. The effect of this compound on the growth of xenograft tumors (subcutaneous and intrabrain) in nude mice was investigated. Our data support the translation of this compound to the clinical setting in combination with standard therapies and to test the efficacy of this compound in association with RT and/or temozolomide (TMZ) administration.

Materials and methods

Cell lines

A total of 12 human glioma cell lines (U251, U373, U118, U138, A172, U87MG, LN19, SW1783, SNB19, LN229, T98G, and D54) were cultured at 37°C in 5% CO₂ and were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS), 4 mM glutamine, 100 IU/mL penicillin, 100 µg/mL streptomycin, and 1% non-essential amino acid (Invitrogen, Life Technologies, Rockville, MD, USA). To minimize the risk of working with misidentified and/or contaminated cell lines, we stocked the cells used in this report at very low passages and used for <20 subcultures. Periodically, to check the authenticity of each cell line used, a short tandem repeat (STR) analysis was performed by GenePrint^® 10 System (Promega Corporation, Madison, WI, USA). This method allows amplification of STR regions for the detection of 10 human loci (TH01, TPOX, vWA, amelogenin, CSF1PO, D16S539, D7S820, D13S317, D5S818, and D21S11) useful for human cell line authentication. The original STR profiles were collected from American Type Culture Collection (ATCC) and Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) cell depositors. Karyotypically distinct U251, SNB19, and U373 cell lines were defined to be of the same origin but have different drug treatment sensitivities.³⁹ Similarly, U138-MG cells show strong similarity to U118-MG cells, sharing at least six derivative marker chromosomes. Luciferase-transfected U87MG cells were kindly provided by Jari E Heikkila, Department of Biochemistry and Pharmacy, Abo Akademi University, Turku, Finland. GBM patient-derived stem cell lines are five (BT12M, BT25M, BT48EF, BT50EF, and BT53M) were kindly provided by J. Gregory Cairncross and Samuel Weiss (Hotchkiss Brain Institute, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada),⁴⁰ and CSCs-5 and CSCs-7 from Marta Izquierdo (Departamento de Biología Molecular, Universidad Autónoma de Madrid, Spain)⁴¹ were maintained as neurosphere cultures in Neurocult medium (StemCell Technologies, Vancouver, BC, Canada) supplemented with epidermal growth factor (EGF; 20 ng/mL) and fibroblast growth factor (FGF; 10 ng/mL). Non-commercial patient-derived cells were analyzed using STR profiles obtained from producers of these cells.

PRX177561 was provided by Proximagen Ltd, (Babraham Research Campus, Cambridge, UK) and it is a highly selective CXCR4 antagonist (Ki at human and mouse CXCR4 receptor approximately 3 nM), which shows no activity at the other chemokine receptors nor at 75 other drug targets (enzymes, receptors, and ion channels) at 10 µM. In mice, the maximum plasma and brain concentrations are achieved approximately 30 min after oral dosing (50 mg/kg), and brain levels exceeded the IC₉₀ for at least the next 24 h. PRX177561 was dissolved in 0.9% NaCl (25 mg/mL) and the pH was brought to 6.5 with HCl. TMZ was purchased from Selleckchem Labs (Aurogene, Rome, Italy).

Reagents and drugs preparation

All the materials for tissue culture were purchased from HyClone (Cramlington, NE, USA). Plasticware was obtained from Nunc (Roskilde, Denmark). Antibodies for β-actin (sc-130065), Ang2 (F1) (sc-74403), Ang-1 (C-19) (sc-6320), FAS (C-20) (sc-715), FAS-L (N-20) (sc-834), CD68 (H-255) (sc-9139), CD20 (M20) (sc7735), matrix metalloproteinase 2 (MMP-2) (4D3) (sc-53630), transforming growth factor beta (TGFβ) RI (V-22) (sc-398), CXCR4 (4G10) (sc-53534), cyclin D1, p27, p21, LC-3-II, beclin 1, ATG5, ATG7, Bcl2, Bax, cytochrome C, caspases 3, caspases8, caspases-9, poly (ADP-ribose) polymerase (PARP), β-catenin, and octamer-binding transcription factor 4 (Oct4) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Tie2 (AB33) Mouse mAb (#4224) and Phospho-Tie2 (Ser1119, antibody #4226) were purchased from Cell Signaling Technology (Leiden, The Netherlands). Survivin antibody was purchased from Biorbyt (Cambridge, UK). Anti-MIB1 (Ki67) was purchased from Dako (Dako Italia SPA, Milan, Italy). Pan-caspase inhibitor z-Val-Ala-Asp(Ome)-fluoromethyl ketone and autophagy inhibitor 3-methyladenine were purchased from Sigma-Aldrich SRL (Milan, Italy).

Growth assays

Cells were seeded at a density of 2 × 10⁴ cells/mL in 24-well plates. Cells were left to attach and grow in 10% fetal calf serum (FCS) DMEM for 24 h. After 24 h of culture in complete medium, cells were serum deprived and treated with SDF1α 100 ng/mL with different concentrations of PRX177561. Cells were trypsinized and resuspended in 1.0 mL of saline, thus viable cells were counted using the NucleoCounter^™ NC-100 (Chemotec, Cydevang, Denmark). Results were represented as data from three independent experiments performed in triplicate. IC₅₀ values for PRX177561 were calculated by the GraFit method (Erithacus Software Ltd, Staines, UK).

Oncosphere formation/limiting dilution assay

GBM stem-like tumor propagating cells were plated at 8 × 10³cells/cm² in NeuroCult NS-A medium (StemCell Technologies) containing 20 ng/mL EGF and 10 ng/mL (FGF; PeproTech Inc., Rocky Hill, NJ, USA). The resulting tumor spheres were collected for every 3–5 days in vitro, mechanically dissociated to a single-cell suspension with StemPro^® Accutase^® Cell Dissociation Reagent (Thermo Fisher Scientific Italia (Rodano, Italy)) and replaced in culture medium. The total number of viable cells was assessed at each passage by trypan blue (Sigma Aldrich SRL (Milan, Italy)) exclusion. For clonogenic assays, cells derived from the dissociation of clonal single neurospheres were seeded on 48-well plates in the presence of complete stem cell medium. The number of secondary spheres generated was assessed after 7, 14, and 21 days.

Cell cycle and apoptosis analysis

Apoptosis was analyzed using Alexa Fluor^® 488 Annexin V/Dead Cell Apoptosis Kit (Life Technologies, Monza, Italy). Then, the fluorescence for all the cells were measured on Tali^® Image-Based Cytometer at 530 nm emission wavelength (e.g. FL1) and >575 nm. The results were expressed as the percentage of cell death by apoptosis in controls and in treated cultures.

Wound-healing assay

Cells were cultured until confluency in six-well plates (SonicSeal Slide; Nalge Nunc, Rochester, NY, USA) and were starved for 24 h. Each well was divided into a 2 × 3 grid. A linear wound was made in each hemisphere of the well using a 1-mL pipette tip and the medium was replaced with starvation medium. After 24 h of inhibitor treatment, images were obtained of the intersections of the linear cell wound and each grid line. Migration rate was estimated from the distance that the cells moved, determined microscopically.

Transwell chamber migration and invasion assays

Migration of GBM cells was analyzed by a Transwell chamber assay³⁶ using 8 µm-pore inserts (BD Biosciences Italia (Milan, Italy)) which stood in six-well plates (Corning Incorporated Life Sciences (Tewksbury, MA, USA)). Cells were seeded in serum-free medium (1 × 10⁶ cells/well) in the Transwell chambers either with or without PRX177561 and allowed to migrate for 20 h at 37°C. To stimulate migration, either FBS (10%), as a positive control of migration, or SDF1α was added to the medium in the well underneath the insert. Quantitative analysis and scanning electron microscopy (SEM) observations were performed as previously described.⁴² Six fields for each condition were examined.

Vasculogenic mimicry formation assay

Vasculogenic mimicry (VM) formation assay was performed using a commercial Matrigel assay kit (BD Biosciences Italia (Milan, Italy)). Extracellular matrix (ECM) Matrigel (200 µL) was placed in 48-well tissue culture plates and incubated at 37°C for 2 h. GBM cells were treated with SDF1α and/or PRX177561 and seeded onto the coated plate. After incubation for 24 h with or without inhibitors, VM formation was assessed using an inverted microscope.

Preparation of cell lysates and western blot analysis

Following treatments, cells, grown in 90 mm diameter Petri dishes, were washed with cold phosphate-buffered saline (PBS) and immediately lysed with 1 mL lysis buffer containing a proteinase and phosphatase inhibitor cocktail. Total lysates were electrophoresed in 7% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), and separated proteins were transferred to nitrocellulose membrane and probed with the appropriate antibodies using the conditions recommended by the suppliers. Total extracts were normalized using an anti β-actin antibody.

Enzyme-linked immunosorbent assay determinations

After appropriate treatments, tumor cell cultures and tissues harvested were lysed and cell lysates were analyzed for cytokine expression. Cell pellets were washed with PBS and lysed with radioimmunoprecipitation assay (RIPA) buffer. Cell lysates and conditioned media were assayed by enzyme-linked immunosorbent assay (ELISA) for the following determinations: (1) active human Caspase-3 (CBA045; Merck Chemicals Ltd, Nottingham, UK), (2) beclin-1 (E98557Hu; USCN Life Sciences, Houston, TX, USA), and (3) DNA damage assay (EpiQuik in situ DNA Damage Assay Kit (EpiGentek, East Farmingdale, NY, USA)). Tumor extracts were analyzed for the presence of (1) MMP-2 (KHC3081; Life Technologies), (2) tumor necrosis factor (TNFα; KHC3011; Life Technologies), (3) CXCR4 (Cyto Glow CXCR4 (pSer339); cell based ELISA; Assay Biotech, Sunnyvale, CA, USA), (4) Human FasL ELISA (ELH-FASL; RayBiotech (Norcross GA, USA)), (5) IL-6 (orb50052; Biorbyt), (6) survivin (orb50135; Biorbyt), (7) TGFβ1 (Human TGF-beta 1 Quantikine ELISA Kit; DB100B; R&D Systems, Minneapolis, MN, USA), and (8) SDF1α (Human CXCL12/SDF-1 alpha Quantikine ELISA Kit; DSA00; R&D Systems). All determinations were performed following the manufacturer’s protocols. Analyses were performed in triplicate and presented data represent mean ± standard error (SE) considering three replicated experiments. Cytokine secretion determined by ELISA was normalized to total protein concentration of tissue lysates.

Mouse GBM xenograft model

Female CD1-nu/nu mice, at 6 weeks of age, were purchased from Charles River (Milan, Italy) under the guidelines established by our Institution (University of L’Aquila, Medical School and Science and Technology School Board Regulations, complying with the Italian government regulation n.116 January 27 1992 for the use of laboratory animals). All mice received subcutaneous flank injections (two each) of 1 × 10⁶ U251, U87MG, and T98G cells representing models for MGMT (O⁶-alkylguanine DNA alkyltransferase) negative and MGMT positive cells. Tumor growth was assessed bi-weekly by measuring tumor diameters with a Vernier caliper (length × width). Tumor weight (TW) was calculated according to the formula: TW (mg) = tumor volume (mm³) = d² × D/2, where d and D are the shortest and longest diameters, respectively. The effects of the treatments were examined as previously described.⁴² At about 10 days after the tumor injection, 30 mice with tumor volume of approximately 0.5–0.8 cm³ were retained and randomly divided into two groups (10 mice per group): (1) Vehicle and (2) PRX177561 (50 mg/kg/day p.o., 5 days in per week). At the end of treatments (35 days after the start of drug administration), animals were sacrificed by carbon dioxide inhalation and tumors were subsequently removed surgically. Half of the tumor was directly frozen in liquid nitrogen for protein analysis and the other half was fixed in paraformaldehyde overnight for immunohistochemical analyses. Indirect immunoperoxidase staining of tumor xenograft samples was performed on paraffin-embedded tissue sections (4 µm). Briefly, sections were incubated with primary antibodies overnight at 4°C. Next, avidin–biotin assays were performed using the Vectastain Elite kit obtained from Vector Laboratories Ltd (Peterborough, UK). Mayer’s hematoxylin was used as nuclear counterstain. Tumor microvessels were counted at 400× in five arbitrarily selected fields, and the data were presented as number of CD31+ microvessels/100× microscopic field for each group. Ki67 labeling index was determined by counting 500 cells at 100× and by determining the percentage of cells staining positively for Ki67. Apoptosis was measured as the percentage of tunnel-positive cells ± standard deviation (SD) measured on five random fields (400×). The presence of red cells in tumor tissue and in blood vessels as well as the presence of microthrombi and bleeding zones was demonstrated by Martius yellow-brilliant crystal scarlet blue technique. Tumor hemoglobin levels were quantified as described elsewhere.⁴²

Evaluation of treatment response in vivo (xenograft model)

The following parameters were used to quantify the anti-tumor effects upon different treatments as previously described:³⁴ (1) tumor volume measured during and at the end of experiments; (2) tumor weight measured at the end of experiment; (3) complete response (CR), defined as the disappearance of the target lesion with respect to baseline; (4) partial response (PR), defined as a reduction of greater than 50% of tumor volume with respect to baseline; (5) stable disease, defined as a reduction of less than 50% or an increase of less than 100% of tumor volume with respect to baseline; (6) tumor progression (TP), defined as an increase of greater than 50% of tumor volume with respect to baseline; and (7) time to progression (TTP).

Orthotopic intrabrain model

For orthotopic luciferase-transfected U87 tumor growth and therapy studies in vivo with an approved animal-use protocol, nude mice were inoculated intracerebrally as follows:²¹ Animals were anesthetized with 100 mg/kg ketamine, 15 mg/kg xylazine, and 0.05 mL atropine (IM). The surgical zone was swabbed with betadine solution, the eyes were coated with Lacri-lube. The head was fixed in a stereotactic frame (Mouse Stereotaxics Instrument; Stoelting Europe, Dublin, Ireland) and a midline scalp incision was made. A small hole was made at 1.0 mm anterior and 2 mm lateral to the exposed bregma. A sterile 5-µL Hamilton syringe with a 26-gauge needle was inserted at a depth of 3.0 mm from the skull surface and withdrawn by 0.5 mm to inject 3 × 10³ U87MG cells in a volume of 3 µL. The injection rate was set upto 1 µL/min. After the implantation of the tumor cells, the needle was left in place for 5 min to prevent reflux. The needle was then completely withdrawn from the brain over the course of 4 min (1.0 mm/min), and the skin was sutured. Just before treatment initiation (5 days after injection), animals were randomly distributed to treatment groups of 10 mice each. In vivo bioluminescence images were obtained using the UVITEC Cambridge Mini HD6 (UVItec Limited, Cambridge, UK) to identify intracranial implants similar to the method described by Kemper et al.⁴² Animals were anesthetized and luciferin (150 mg/kg) was injected intraperitoneally (IP) 15 min prior to imaging. The mice were photographed while placed on their front and the bioluminescence intensity (BLI) was measured in the region of interest (ROI). The BLI value just prior to the initiation of the treatment was used to calculate the percentage of BLI of increment for each individual animal. We deliberately inoculated a small amount of cells (3 × 10³) to simulate a chemo-radiotherapeutic treatment made after surgery in which a low number of tumor cells, remaining the operatory bed, regrows and gives arise to a recurrence. Treatments were started 5 days after cell injection when no luciferase activity was intracranially detectable. Mice were euthanized when they displayed neurological signs (e.g. altered gait, tremors/seizures, lethargy) or weight loss of 20% or greater of presurgical weight.

Statistical analysis

Continuous variables were summarized as mean and SD or as median and 95% confidence interval (CI) for the median. For continuous variables that are not normally distributed, statistical comparisons between control and treated groups were established by carrying out the Kruskal–Wallis tests. When Kruskal–Wallis tests revealed a statistical difference, pairwise comparisons were made by Dwass–Steel–Chritchlow–Fligner method and the probability of each presumed “non-difference” was indicated. For continuous variables that are normally distributed, statistical comparisons between control and treated groups were established by carrying out the analysis of variance (ANOVA) test or by Student t test for unpaired data (for two comparisons). When ANOVA test revealed a statistical difference, pairwise comparisons were made by Tukey’s HSD (honest significant difference) test and the probability of each presumed “non-difference” was indicated. Dichotomous variables were summarized by absolute and/or relative frequencies. For dichotomous variables, statistical comparisons between control and treated groups were established by carrying out the exact Fisher’s test. For multiple comparisons, the level of significance was corrected by multiplying the p value by the number of comparisons performed (n) according to Bonferroni correction. TTP was analyzed by Kaplan–Meier curves and Gehan’s generalized Wilcoxon test. When more than two survival curves were compared, the log-rank test for trend was used. This tests represents the probability of a trend in survival scores across the groups. All tests were two-sided and were determined by Monte Carlo significance. p < 0.05 were considered to be statistically significant. SPSS^® (statistical analysis software package) version 10.0 and StatsDirect Ltd (version 2.3.3., Altrincham, UK) were used for statistical analysis and graphical presentation.

Results

CXCR4 and SDF1α expression levels in GBM and patient-derived GBM stem cells

Western blotting and ELISA were used to evaluate basal expression of SDF1α and CXCR4 in a cohort of 12 GBM and 6 patient-derived GBM stem cells as described in the Materials and methods section. We used western blotting to measure total CXCR4, and ELISA determinations, performed on permeabilized cells, for active pSer399 CXCR4 plasma membrane associated levels. In Figure 1(a), we show representative western blots of five GBM cell lines and three patient-derived GBM stem cells. In Figure 1(b) and (c), we show the comparison of total (Figure 1(b)) and pSer339 (Figure 1(c)) CXCR4 expression in differentiated and stem GBM cell lines. Overall, patient-derived GBM stem cell lines expressed higher levels than those observed in differentiated GBM cell lines (3.87 ± 0.25 vs 2.14 ± 0.22, p = 0.0395). The amount of pSer399 was higher in GBM stem cells compared to GBM cell lines (1.95 ± 0.21 vs 0.80 ± 0.32, p < 0.0054). The SDF1α ELISA determinations (Figure 1(d)) revealed that although patient-derived GBM stem cells produced higher SDF1α compared to those secreted by differentiated GBM cells, statistical analysis showed no significant differences (59.17 ± 18.59 vs 16.50 ± 6.32 pg/mL/1 × 10⁶ cells, p = 0.0979, NS). However, when U87MG or U251 cells were induced to grow in stem cell medium, the formation of spheroids was associated with significantly increased CXCR4 expression (Figure 1(e)) and SDF1α secretion (Figure 1(f)). In A172 cells, CXCR4 production was higher in hypoxic conditions and in the presence of exogenous SDF1α (Figure 1(g)).

Figure 1.

CXCR4 and SDF1α expression levels. (a) Representative western blots performed on cell extracts derived from U87MG, T98G, U251, A172, and LN229 GBM cell lines and BT12M, BT48EF, and CSCs-5 patient-derived GBM stem cells (GBMscs). Each lane was loaded with 100 µg of protein from cell extracts obtained from control and treated cells. (b) Box plot graphical analysis of total CXCR4 expressed as adjusted densitometric units (normalized vs GAPDH) in 12 GBM cell lines and 5 patient-derived GBM stem cells. (c) Box plot graphical analysis of pSer339 CXCR4 expressed in our cell cohorts. (d) Box plot graphical analysis of SDF1α production in our cell cohorts. (e) CXCR4 levels in U87MG and U251 grown as adherent culture or sphere aggregates. (f) SDF1α secretion by U87MG and U251 cells grown as above. (g) Immunoperoxidase CXCR4 expression in the A172 cells cultured in normoxia and hypoxia with or without exogenous SDF1α (100 ng/mL). Pictures were captured at 400× magnification.

SDF1α modulates cell proliferation of GBM cells

Next, we evaluated the effects of exogenous SDF1α (0–100 ng/mL) in three GBM cell lines (U87MG, U251, and A172) and in GBM stem (BT12M) cells. We found that SDF1α induced the proliferation of GBM cells (Figure 2(a)) and sustained the sphere formation in BT12M (Figure 2(b)) even in the absence of complete stem cell medium containing EGF and basic FGF (bFGF). These effects were similar in all patient-derived GBM stem cells. In Figure 2(c), we show the appearance of CSCs-5 grown as sphere treated with different concentrations of SDF1α. To investigate the possible functional effects induced by CXCR4 downmodulation and antagonism, we evaluated cell proliferation after exposure of GBM cells to different concentrations of PRX177561. We evaluated both the basal and SDF1α-mediated cell proliferation. Cells were seeded 24 h before adding PRX177561, and cell proliferation was determined 24, 48, and 72 h after treatments. The effects of PRX177561 on the different cell populations were compared through the IC₅₀ values calculated for each culture. In Figure 3, we show representative IC₅₀ growth curves generated at 72 h for U87MG (Figure 3(a)), U251 (Figure 3(b)), as well as for BT12M (Figure 3(c)) and CSCs-5 (Figure 3(d) and (e)) cells. IC₅₀ values ranged between 0.8 and 4.6 µM for differentiated GBM cell lines and between 0.5 and 1.1 µM for patient-derived GBM stem cell cultures, respectively. Overall, IC₅₀ values (Figure 3(f)) were significantly lower in GBM patient–derived stem cells compared to the differentiated GBM cell lines (0.77 ± 0.09 µM vs 2.32 ± 0.51 µM, p = 0.0234). Next, we compared the linear regressions between CXCR4 expression and IC₅₀ (Figure 3(g)) or p-ser399 CXCR4 expression and IC₅₀ values (Figure 3(h)), and we found inverse correlation coefficients for both the analyses with R = −0.37 (p = 0.09, NS, total CXCR4) and R = −0.4740 (p = 0.049, p-ser399 CXCR4), suggesting that the higher the expression of active CXCR4, the lower the concentration of PRX177561 required to inhibit proliferation. The regression coefficient for the analysis of SDF1α secretion and PRX177561 efficacy (Figure 3(i)) showed R = −0.45 (p = 0.063, NS).

Figure 2.

SDF1α modulates cell proliferation of GBM cells. Effects of exogenous SDF1α (0–100 ng/mL) on representative (a) GBM (U87MG, U251, and A172) and (b) GBM stem (BT12M and CSCs-5) cells. SDF1α induced cell proliferation of GBM cells and sustains the sphere formation in BT12M even in the absence of complete stem cell medium containing EGF and bFGF. (c) Microscopic pictures of sphere formation in CSCs-5 with increasing concentrations of SDF1α. Single results are representative of three different experiments performed in triplicate.

Figure 3.

PRX177561 reduces cell proliferation of GBM and patient-derived GBM stem cells in a dose-dependent manner. IC₅₀ values calculated for (a) U87MG, (b) U251, (c) CSCs-5, and BT12M (d) cells. (e) Microscopic representation of the effects of different doses of PRX177561 on BT12M cells grown in cancer stem cell medium with SDF1α. (f) Box plot graphical analyses for IC₅₀ values calculated in GBM and patient-derived GBM stem cell cultures. (g) Linear regression of CXCR4 expression and PRX177561 efficacy (IC₅₀). (h) Linear regression of active CXCR4 expression and PRX177561 efficacy (IC₅₀). (i) Linear regression of SDF1α expression and PRX177561 efficacy (IC₅₀).

The evaluation of apoptosis revealed no appreciable apoptotic or necrotic effects with low concentrations of PRX177561 neither in differentiated GBM nor in stem cell populations. In Figure 4(a), we show DNA fragmentation (ladder assay) in U87MG and BT12M cells treated with relatively low (<IC₅₀) and high PRX177561 concentrations (>IC₅₀) for 72 h.

Figure 4.

Effects of PRX177561 on apoptosis. (a) DNA fragmentation (ladder assay) in U87MG and BT12M cells treated with relatively low (<IC₅₀) and high (>IC₅₀) PRX177561 concentrations. (b) Colorimetric Apoptosis Detection analyses in U87MG cells and CSCs-5 treated with PRX177561. (c) Low doses of PRX177561 (0.1–1.0 µM) was, however, enough to arrest U87MG or BT12M tumor cell cycle phase with increased expression of P21 and P27 and downmodulation of cyclin D1. Western blotting analyses performed on U87MG cells treated with low doses of PRX177561 and showing increased expression of P21, P27 and reduced amount of cyclin D1. (d) Differentiation effects of 0.5 µM PRX177561 on BT12M cells able to reduce nestin and significantly increase glial fibrillary acidic protein (GFAP) expression. Reduction of stem cell markers including Sox2, Twist, and Nanog was also observed.

The TiterTACS^™ Colorimetric Apoptosis Detection analyses revealed a strong apoptosis at higher concentrations of PRX177561 (Figure 4(b)) in both the cell models (effects >1.0 µM in BT12M and 5 µM in U87MG cells). Treatment with low doses of PRX177561 (0.1–1.0 µM) was, however, sufficient to arrest U87MG or BT12M tumor cell cycle phase with increased expression of P21 and P27 and downmodulation of cyclin D1, suggesting G0/G1 cell cycle arrest (Figure 4(c)). In contrast, treatment with low concentrations of PRX177561 did not affect D54 and A172 tumor cell proliferation (data not shown).

In addition, we observed that PRX177561 modified the expression of certain markers related to aggressiveness and epithelial–mesenchymal transition (EMT) in patient-derived GBM stem cells in a time-dependent manner. In particular, in Figure 4(d), we show that 0.5 µM PRX177561 reduced nestin expression in BT12M while a significant increase in glial fibrillary acidic protein (GFAP) expression was observed. A reduction in the expression of the stem cell markers—Sox2, Twist, and Nanog—was also observed. These data suggest that PRX177561 reduced cancer stem cell stemnes, perhaps by inducing their differentiation

Effects of PRX177561 on GBM cell migration

Since it is widely considered that CXCR4 activity modulates cell motility of different tumor cells, we tested the effects of SDF1α and PRX177561 on glioma cell migration and vasculomimicry. U87MG and U251 cells were used as models. These cells, in serum-free medium, were seeded in transwell chambers and allowed to migrate toward the chemokine SDF1α (added to the lower chamber medium) in the absence or presence of PRX177561. FBS was also used as a positive control for migration. In Figure 5, we show that SDF1α induced chemotaxis of U87MG (Figure 5(a)) and U251 (Figure 5(b)) cells. In the checkerboard analyses (Figure 5(c) and (d)), we show that changes in the chemotactic SDF1α gradient after addition of the cytokine to the upper chamber demonstrate the specificity of SDF1α-mediated cell migration. Further demonstration was obtained after the treatment of cells with increasing concentrations of PRX177561, which reduced both the basal and SDF1α (100 ng/mL)-mediated cell migration (Figure 5(e) and (f)). Figure 5(g) shows a representative experiment of wound healing in U251 cells treated with SDF1, with or without 1.0 µM PRX177561, in which an inhibition of cell migration by the antagonist can be clearly seen.

Figure 5.

Effects of PRX177561 on GBM cell migration. SDF1α induces (a) U87MG and (b) U251 cell migration in a modified Boyden Chamber assays and in a dose-dependent manner. Checkerboard analyses show that changes in the chemotactic SDF1α gradient after addition of the cytokine to the upper chamber of (c) U87MG and (d) U251 cells demonstrate the specificity of SDF1α-mediated cell migration. PRX177561 reduced SDF1α-mediated cell migration of (e) U87MG and (f) U251 cells in a dose-dependent manner. (g) Representative experiment of wound healing in U251 cells treated with SDF1α with or without 1.0 µM PRX177561.

Solid tumor growth is dependent on the development of an adequate blood supply. For years, sprouting angiogenesis has been considered an exclusive mechanism of tumor vascularization. However, over the last years, several other mechanisms have been identified, including vessel-co-option, intussusception, recruitment of endothelial precursor cells (EPCs), and even mechanisms that do not involve endothelial cells, a process called VM. We used U87MG as a model for the analyses of VM. U87MG cells grow with a spindle cell morphology and a tendency to have numerous cell contacts between cell colonies as shown in Figure 6(a). PRX177561, dose-dependently, reduced cell spindle morphology inducing cell rounding and partial detachment. When cultured in Matrigel, U87MG shows a migratory capacity and formation of tubule-like branching structures as shown in Figure 6(b). This suggests that CXCR4 antagonism by PRX177561 affects the invasive capacity of GBM cells, inhibiting migration and thus perhaps local invasion from GBM lesions. Therefore, we analyzed the expression changes of EMT markers induced by PRX177561 in U87MG cells. In Figure 6(c), we show that PRX177561 increased the expression of E-cadherin and GFAP. In parallel, PRX177561 downregulated the expression of Twist, N-cadherin, MMP-2, and MMP-9 suggesting a reduced EMT phenotype.

Figure 6.

PRX177561 influences vasculogenic mimicry of GBM cells. (a) U87MG cells grow with a spindle cell morphology and a tendency to have numerous cell contact between cell colonies. PRX177561, dose-dependently, reduced cell spindle morphology inducing cell rounding and partial detachment in absence of serum. (b) U87MG shows in Matrigel a migratory capacity and formation of tubule-like branching structures. (c) PRX177561 seems to reverse the EMT phenotype in U87MG cells with increased expression of E-cadherin and GFAP and reduction of Twist, N-cadherin, MMP-2, and MMP-9.

In vivo study: PRX177561 treatment delayed GBM tumor growth in vivo in subcutaneous xenograft models

Once we had established the in vitro anti-proliferative effects of PRX177561, we sought to further confirm these observations in vivo. So, we examined the effects of oral administration of PRX177561 (50 mg/kg p.o. q.d.) on U87MG, U251, and T98G GBM cell lines subcutaneously engrafted in female cd1 nu/nu mice. We observed that PRX177561 administration significantly reduced tumor weights by about 20% in U87MG (1.00 g ± 0.16 vs 0.81 g ± 0.13, p = 0.0076), 43% in U251 (0.79 g ± 0.12 vs 0.45 g ± 0.05, p < 0.0001), and 27% in T98G (0.88 g ± 0.11 vs 0.64 g ± 0.08, p = 0.0007) as shown in Figure 7(a), (d), and (g), respectively. The comparison analyses of TTP also show that this parameter was significantly increased after PRX177561 administration (Figure 7(b), (e), and (h)). Immunohistochemical analyses revealed that the high CXCR4 expression observed in the tumor cells was significantly reduced by PRX177561 treatment, suggesting that the oral administration of this drug could reach the tumor site for its functional activity (data not shown).

Figure 7.

In vivo study: PRX177561 treatment delayed GBM tumor growth in vivo in subcutaneous xenograft models. (a) Growth curves of U87MG tumor–bearing mice treated with Vehicle (CTRL), RT (3 Gy in single dose), PRX177561 (50 mg/kg p.o. q.d.), and temozolomide (16 mg/kg p.o. 1–5 days). (b) Time to progression (TTP) calculated was on U87MG. (c) Kaplan–Meier curves generated for U87MG xenografts showing an HR = 2.21 (RT, p = 0.0173), an HR = 2.44 (PRX177561, p = 0.0077), and an HR = 4.16 (TMZ, p < 0.0001), respectively. PRX177561 and RT show similar delay of tumor growth (HR = 1.10, p = 0.774, NS). TMZ showed better effects than both the RT (HR = 2.87, p = 0.0013) and PRX177561 (HR = 3.15, p = 0.0005). (d) Growth curves of U251 tumor–bearing mice treated as mentioned above. (e) TTP on U251. (f) Kaplan–Meier curves generated for U251 xenografts showing an HR = 2.21 (RT, p = 0.0127), an HR = 3.31 (PRX177561, p = 0.0003), and an HR = 4.37 (TMZ, p < 0.0001), respectively. Percentage of tumor progression in PRX177561 was significantly higher when compared to RT (HR = 2.05, p = 0.0244). TMZ showed better effects than both the RT (HR = 3.61, p = 0.0001) and PRX177561 (HR = 2.19, p = 0.0148). (g) Growth curves of T98G tumor–bearing mice treated as above mentioned. (h) TTP on T98G. (i) Kaplan–Meier curves generated for T98G xenografts showing an HR = 2.71 (RT, p = 0.0030), an HR = 3.92 (PRX177561, p < 0.0001), and an HR = 1.87 (TMZ, p = 0.0538, NS), respectively. Percentage of tumor progression in PRX177561 was significantly higher when compared to RT (HR = 2.85, p = 0.0054). TMZ showed lower effects than both the RT (HR = 1.68, p = 0.1074, NS) and PRX177561 (HR = 3.21, p = 0.0004).

PRX177561 inhibits intracranial growth of primary brain tumors

The efficacy of PRX177561 was investigated in mice with experimental brain tumors. We injected luciferase-transfected U87MG cells in the brain as described in Materials and methods section. We deliberately inoculated a small number of cells (3 × 10³) to simulate a chemo-radiotherapic treatment that is carried out after surgery in which a low number of tumor cells, remaining after resection and chemoradiotherapy, are able to regrow and develop into a recurrent lesion. Treatments were started 5 days after cell injection when no luciferase activity was intracranially detectable. Therefore, treating the animals for 35 days with a maximum 180 days of follow-up without drug administration, we were able to record tumor growth delay (recurrence time equivalent of disease-free survival (DFS), Figure 8(a) and (b)) as the time in which a luciferase activity was intracranially detectable. BLI photon counts and tumor volumes (calculated by magnetic resonance imaging (MRI)) were correlated. We also recorded the survival time (overall survival (OS); Figure 8(d) and (e)) considering the day in which mice, displaying neurological signs (e.g. altered gait, tremors/seizures, lethargy) or weight loss of 20% or greater of presurgical weight, were euthanized, as indicated in the Materials and methods section. We compared PRX177561 administration with RT, TMZ, or the combination of both RT and TMZ. Control mice developed a bioluminescent lesion between 10 and 25 days with a mean of 16.5 ± 1.7 (SE) days. RT was able to increase the DFS and slowed the recurrence upto 23.0 ± 2.1 days (p = 0.0002) with a hazard ratio (HR) = 2.23 (p = 0.017) compared to CTRLs. TMZ slowed the regrowth of the recurrent tumors to 52.0 ± 19.0 days, significantly more than RT (p = 0.0134) with an HR = 3.52 (p = 0.0003) versus CTRL and an HR = 2.2 (p = 0.0300) versus RT. PRX177561 increased significantly the time of recurrence with respect to control upto 40.5 ± 5.7 days (p = 0.0003) with an HR = 3.7 (p = 0.0001) versus CTRLs, an HR = 2.84 (p = 0.0024) versus RT, and an HR = 1.01 (p = 0.9770, NS) versus TMZ. Next, we analyzed the OS rate in the same groups. We observed that the PRX177561 treatment significantly increased the OS (77 ± 6.8 days) of animals injected with U87MG cells compared to untreated animals (43.1 ± 2.8 days) and RT-treated animals (58.0 ± 3.7 days) but not TMZ-treated animals (OS 98.0 ± 14.7 days) with an HR = 4.2 (p < 0.0001), an HR = 2.5 (p = 0.0102), and an HR = 0.64 (p = 0.2849, NS) in the comparisons with CTRLs, RT, and TMZ, respectively. Statistical data for Kaplan–Meier analyses are summarized in Figure 8(c) and (f).

Figure 8.

PRX177561 inhibits intracranial growth of primary brain tumors. After 5 days of cell injection, when no luciferase activity was intracranially detectable, treatments were performed for 35 days. After this time, we considered a follow-up of a maximum 180 days. (a) Recurrence time, equivalent of DFS, is the time in which a luciferase activity was intracranially detectable. (b) Percentage of positive mice in the time (Kaplan–Meier analysis). (c) Statistical analysis from Kaplan–Meier curves. Control mice developed a bioluminescent lesion between 10 and 25 days with a mean of 16.5 ± 1.7 (standard error) days. RT was able to increase the disease-free survival and slowed the recurrence upto 23.0 ± 2.1 days (p = 0.0002) with an HR = 2.2302 (p = 0.0172) when compared to CTRLs. TMZ slowed the insurgence of recurrent tumors upto 52.0 ± 19.0 days, resulting in a highly significant versus RT (p = 0.0134) with an HR = 3.5146 (p = 0.0003) versus CTRL and an HR = 2.2 (p = 0.0300) versus RT. In this group, 1/10 animals did not develop any tumor and were alive at the end of this experiment. PRX177561 increased significantly the time of recurrence with respect to control upto 40.5 ± 5.7 days (p = 0.0003) with an HR = 3.7 (p = 0.0001) versus CTRLs, an HR = 2.8410 (p = 0.0024) versus RT, and an HR = 1.0122 (p = 0.9770, NS) versus temozolomide. (d) Survival time (overall survival (OS)) was calculated considering the day in which mice, displaying neurological signs (e.g. altered gait, tremors/seizures, lethargy) or weight loss of 20% or greater of presurgical weight, were euthanized, as indicated in Materials and methods section. (e) Percentage of live mice in the time (Kaplan–Meier analysis). (f) Statistical analysis from Kaplan–Meier curves. We observed that the PRX177561 treatment increased the OS (77 ± 6.8 days) of animals injected with U87MG cells resulting statistically higher significance when compared to untreated (43.1 ± 2.8 days) and RT-treated animals (58.0 ± 3.7 days) but lower to TMZ-treated animals (98.0 ± 14.7) with an HR = 4.2 (p < 0.0001), an HR = 2.5 (p = 0.0102), and an HR = 0.6414 (p = 0.2849, NS) in the comparisons with CTRLs, RT, and temozolomide.

Discussion

It is widely appreciated that there is an urgent need to identify novel therapeutics that will improve survival in human GBM patients. The CXCL12-dependent signaling has emerged among the most relevant molecular pathways that can be targeted to successfully interfere with tumor cell proliferation, survival, migration, and radio-resistance.⁴³ Less clear, however, is whether altering or inhibiting CXCR4 signaling would lead to cell cycle arrest or cell death. The sole Food and Drug Administration–approved anti-CXCR4 agent, AMD3100, was recently shown to suppress the proliferation of a subpopulation of drug-resistant lung cancer A549 cells.⁴⁴ Here, we show the efficacy of the newly synthesized CXCR4 antagonist, PRX177561, in modulating the intrinsic properties of GBM cells and their microenvironment in an experimental model of GBM. Our work has contributed to the field by demonstrating for the first time the in vitro and in vivo inhibitory effects of PRX177561 against a wide range of GBM cell lines, representing diverse pathologic subtypes. We found that PRX177561 stably reduced the expression of CXCR4 in human U87MG cells and impaired their metabolic activity, cell proliferation, and migration in response to SDF1α in vitro.

EMT has been widely studied because of its effects on early tumor development and malignant cancer metastasis; the EMT process transforms a differentiated epithelial cell into a mesenchymal cell that shows stem cell–like properties and is characterized by a loss of cell polarity and increased motility.^45,46 However, the molecular mechanisms of EMT in human GBM are not well reported. In this report, SDF1α stimulation activated the protein expression of phosphorylated extracellular signal–regulated kinase (p-ERK), p-AKT, Twist, and N-cadherin in a concentration-dependent fashion and downregulated the protein expression of the epithelial marker, E-cadherin. These findings indicated that SDF1α/CXCR4 could promote the EMT in human GBM with the E- to N-cadherin switch leading to the deregulation of intercellular adhesion.⁴⁷

In addition, we demonstrated that the addition of PRX17756 inhibited SDF1α induced colony formation of CSCs and the proliferation of GBM cells in a dose-dependent manner. This led us to examine whether PRX177561 would delay spontaneous GBM tumor cell proliferation in the absence of SDF1α. In line with this prediction, we found that PRX177561 had both cytostatic and cytotoxic effects on all tested cell lines. The U87MG cell line was most sensitive to PRX177561 and the D54 and A172 cell lines were the least. Although all tested cell lines expressed CXCR4, they showed different responses to PRX177561. This may be simply because the different cell lines show differing dependencies on CXCR4 for proliferation or perhaps because CXCR7 mediates some of the effects of SDF1 in these cells.

We demonstrated that a single daily dose of PRX177561 was, however, sufficient to induce delayed propagation of U87MG-, U251MG-, or T98G-derived tumors in vivo. Overall, these data have established for the first time the anti-proliferative effects of PRX177561 against human GBM cell lines and tumor xenografts. It has been also demonstrated that the post-irradiation recruitment of bone marrow–derived (BMD) cells can be prevented using a CXCR4 antagonist such as AMD3100 (plerixafor) and that this delays tumor recurrence following both single and fractionated doses of irradiation.⁴⁸ Accordingly, in the orthotopic U87MG mouse model, PRX177561 reduced tumor cellularity, abrogated dissemination of GBM cells at distant cerebral sites, promoted M1 features in GAMs recruited to the tumor area, and impaired intratumor vasculature as previously shown.³¹

Although the links among the various functions of macrophages, TP, and therapy responses are still poorly defined, the M1/M2 balance has emerged as a crucial factor able to control glioma biology and radio-resistance.^10–14 Interestingly, CXCR4 activation can lead to an increased mammalian target of rapamycin (mTOR) activity, which is a crucial modulator of inflammatory pathways in microglia and macrophages. Lisi et al.⁴⁹ reported that mTOR kinase inhibitors polarize glioma-activated microglia toward the M1 phenotype. Interestingly, it has been demonstrated that CXCR4 inhibition induced strong monocyte immunoreactivity in GBM for the M1 marker, inducible nitric oxide synthase (iNOS), by potentially increasing the anti-tumor killing functions of resident or recruited macrophages.⁵⁰ We observed similar results with PRX177561 (manuscript in preparation) associated with enhanced expression of iNOS indicating that PRX177561 interferes with the phenotype switching of CD68+ macrophages toward M2 pro-tumorigenic functions induced by the crosstalk with tumor cells.

Anti-angiogenic therapies in GBM patients have shown some promise,⁵¹ so, it is interesting that PRX177561 reduced the vasculomimicry of U87 cells and has also been shown to reduce the number of CD31/vascular endothelial growth factor (VEGF) positive cells and Arg-1-expressing vessel-like structures in tumor cores. These data suggest that this CXCR4 antagonist reduces aberrant intratumor vascularization, consistent with the role of SDF1 in angiogenesis and vasculogenesis. In addition, modulating the M1/M2 polarization of monocyte/macrophages may also contribute to this control of tumor angiogenesis.^10–14 Finally, taking into consideration the importance of the CXCR4/SDF1α axis in mediating tumor microenvironment interactions, future research will benefit from focusing on the mechanisms by which CXCR4 antagonists regulate tumor stroma and tumor immune cell crosstalk. Overall, our data suggest that the potential of PRX177561 as anti-cancer agent can be exploited in anti-GBM therapy. PRX177561 could enhance the efficacy of standard treatments such as radiation, chemotherapy, and anti-VEGF therapies. The upregulation of CXCL12 and CXCR4 is part of the escape program that favors tumor recurrence and dissemination, with vascular restoration and macrophage recruitment.^29,30,52 The recruitment of circulating cells that can recolonize and/or stabilize the tumor vasculature after irradiation has been suggested as a mechanism of GBM resistance to irradiation.

Conclusion

Our findings have indicated that targeting the CXCL12/CXCR4 axis with PRX177561 significantly attenuates GBM tumor growth and might augment the effects of anti-tumor chemotherapy and RT. In addition, PRX177561 increases the DFS and OS when administered to mice bearing orthotopic intrabrain xenografts. These promising results suggest that future research will benefit from delineating the downstream mechanisms of PRX177561 action and better defining PRX177561 cell susceptibility markers. We believe that by following these paths, scientists and physicians will advance the introduction of CXCR4-based therapeutics for GBM, in particular, and solid malignancies, in general.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Ricard

Idbaih

Ducray

. Primary brain tumours in adults. Lancet 2012; 379: 1984–1996.

Trent

Kong

Short

. Temozolomide as second-line chemotherapy for relapsed gliomas. J Neurooncol 2002; 57: 247–251.

Galli

Binda

Orfanelli

. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 2004; 64: 7011–7021.

Guadagno

Borrelli

Califano

. Immunohistochemical expression of stem cell markers CD44 and nestin in glioblastomas: evaluation of their prognostic significance. Pathol Res Pract 2016; 212: 825–832.

Munthe

Sørensen

Thomassen

. Migrating glioma cells express stem cell markers and give rise to new tumors upon xenografting. J Neurooncol 2016; 130(1): 53–62.

Gabrusiewicz

Rodriguez

Wei

. Glioblastoma-infiltrated innate immune cells resemble M0 macrophage phenotype. JCI Insight 2016; 1(2): e85841.

Steffen

Clark

. CSF1 overexpression promotes high-grade glioma formation without impacting the polarization status of glioma-associated microglia and macrophages. Cancer Res 2016; 76: 2552–2560.

Mangani

Weller

Seyed Sadr

. Limited role for transforming growth factor-β pathway activation-mediated escape from VEGF inhibition in murine glioma models. Neuro Oncol 2016; 18(12): 1610–1621.

Szulzewsky

Arora

de Witte

. Human glioblastoma-associated microglia/monocytes express a distinct RNA profile compared to human control and murine samples. Glia 2016; 64: 1416–1436.

10.

Kloepper

Riedemann

Amoozgar

. Ang-2/VEGF bispecific antibody reprograms macrophages and resident microglia to anti-tumor phenotype and prolongs glioblastoma survival. Proc Natl Acad Sci U S A 2016; 113: 4476–4481.

11.

Quail

Bowman

Akkari

. The tumor microenvironment underlies acquired resistance to CSF-1R inhibition in gliomas. Science 2016; 352: aad3018.

12.

Chang

Miska

Wainwright

. CCL2 produced by the glioma microenvironment is essential for the recruitment of regulatory T cells and myeloid-derived suppressor cells. Cancer Res 2016; 76: 5671–5682.

13.

Zhang

Hutter

Kahn

. Anti-CD47 treatment stimulates phagocytosis of glioblastoma by M1 and M2 polarized macrophages and promotes M1 polarized macrophages in vivo. PLoS One 2016; 11: e0153550.

14.

Hambardzumyan

Bergers

. Glioblastoma: defining tumor niches. Trends Cancer 2015; 1: 252–265.

15.

Bleul

Farzan

Choe

. The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry. Nature 1996; 382: 829–833.

16.

Gagliardi

Narayanan

Reni

. The role of CXCR4 in highly malignant human gliomas biology: current knowledge and future directions. Glia 2014; 62: 1015–1023.

17.

Liu

Pham

Luo

. Expression and functional heterogeneity of chemokine receptors CXCR4 and CXCR7 in primary patient-derived glioblastoma cells. PLoS One 2013; 8: e59750.

18.

Tabouret

Tchoghandjian

Denicolai

. Recurrence of glioblastoma after radio-chemotherapy is associated with an angiogenic switch to the CXCL12-CXCR4 pathway. Oncotarget 2015; 6: 11664–11675.

19.

Wang

Zhang

. CXCL12-induced upregulation of FOXM1 expression promotes human glioblastoma cell invasion. Biochem Biophys Res Commun 2014; 447: 1–6.

20.

Bianco

Uno

Oba-Shinjo

. CXCR7 and CXCR4 expressions in infiltrative astrocytomas and their interactions with HIF1α expression and IDH1 mutation. Pathol Oncol Res 2015; 21: 229–240.

21.

Komatani

Sugita

Arakawa

. Expression of CXCL12 on pseudopalisading cells and proliferating microvessels in glioblastomas: an accelerated growth factor in glioblastomas. Int J Oncol 2009; 34: 665–672.

22.

Calinescu

Yadav

Carballo

. Survival and proliferation of neural progenitor derived glioblastomas under hypoxic stress is controlled by a CXCL12/CXCR4 autocrine positive feedback mechanism. Clin Cancer Res. Epub ahead of print 19 August 2016. DOI: 10.1158/1078-0432.CCR-15-2888.

23.

Richardson

. CXCR4 and glioblastoma. Anticancer Agents Med Chem 2016; 16: 59–74.

24.

Tseng

Vasquez-Medrano

Brown

. Targeting SDF-1/CXCR4 to inhibit tumour vasculature for treatment of glioblastomas. Br J Cancer 2011; 104: 1805–1809.

25.

Petritsch

. HIF1alpha induces the recruitment of bone marrow-derived vascular modulatory cells to regulate tumor angiogenesis and invasion. Cancer Cell 2008; 13: 206–220.

26.

Christoffersson

Vågesjö

Vandooren

. VEGF-A recruits a proangiogenic MMP-9-delivering neutrophil subset that induces angiogenesis in transplanted hypoxic tissue. Blood 2012; 120: 4653–4662.

27.

Rios

Hsu

Blanco

. Durable response of glioblastoma to adjuvant therapy consisting of temozolomide and a weekly dose of AMD3100 (plerixafor), a CXCR4 inhibitor, together with lapatinib, metformin and niacinamide. Oncoscience 2016; 3: 156–163.

28.

Liao

Shi

Jiang

. SDF-1/CXCR4 axis regulates cell cycle progression and epithelial-mesenchymal transition via up-regulation of survivin in glioblastoma. Mol Neurobiol 2016; 53: 210–215.

29.

Gatti

Pattarozzi

Bajetto

. Inhibition of CXCL12/CXCR4 autocrine/paracrine loop reduces viability of human glioblastoma stem-like cells affecting self-renewal activity. Toxicology 2013; 314: 209–220.

30.

Kozin

Kamoun

Huang

. Recruitment of myeloid but not endothelial precursor cells facilitates tumor regrowth after local irradiation. Cancer Res 2010; 70: 5679–5685.

31.

Gravina

Mancini

Marampon

. The brain-penetrating CXCR4 antagonist, PRX177561, increases the antitumor effects of bevacizumab and sunitinib in preclinical models of human glioblastoma. J Hematol Oncol 2017; 10(1): 5.

32.

Matsusaka

Mishima

Suenaga

. Circulating endothelial progenitors and CXCR4-positive circulating endothelial cells are predictive markers for bevacizumab. Cancer 2011; 117: 4026–4032.

33.

Duda

di Tomaso

. Direct evidence that bevacizumab, an anti-VEGF antibody, up-regulates SDF1alpha, CXCR4, CXCL6, and neuropilin 1 in tumors from patients with rectal cancer. Cancer Res 2009; 69: 7905–7910.

34.

Cheng

Huang

Zhou

. A critical role of Src family kinase in SDF-1/CXCR4-mediated bone-marrow progenitor cell recruitment to the ischemic heart. J Mol Cell Cardiol 2015; 81: 49–53.

35.

Rosenkilde

Gerlach

Jakobsen

. Molecular mechanism of AMD3100 antagonism in the CXCR4 receptor: transfer of binding site to the CXCR3 receptor. J Biol Chem 2004; 279: 3033–3041.

36.

Gravina

Mancini

Muzi

. CXCR4 pharmacogical inhibition reduces bone and soft tissue metastatic burden by affecting tumor growth and tumorigenic potential in prostate cancer preclinical models. Prostate 2015; 75: 1227–1246.

37.

Terasaki

Sugita

Arakawa

. CXCL12/CXCR4 signaling in malignant brain tumors: a potential pharmacological therapeutic target. Brain Tumor Pathol 2011; 28: 89–97.

38.

Goffart

Kroonen

Di Valentin

. Adult mouse subventricular zones stimulate glioblastoma stem cells specific invasion through CXCL12/CXCR4 signaling. Neuro Oncol 2015; 17: 81–94.

39.

Kalatskaya

Berchiche

Gravel

. AMD3100 is a CXCR7 ligand with allosteric agonist properties. Mol Pharmacol 2009; 75: 1240–1247.

40.

Cusulin

Chesnelong

Bose

. Precursor states of brain tumor initiating cell lines are predictive of survival in xenografts and associated with glioblastoma subtypes. Stem Cell Rep 2015; 5(1): 1–9.

41.

Gil-Ranedo

Mendiburu-Eliçabe

García-Villanueva

. An off-target nucleostemin RNAi inhibits growth in human glioblastoma-derived cancer stem cells. PLoS One 2011; 6: e28753.

42.

Kemper

Leenders

Kusters

. Development of luciferase tagged brain tumour models in mice for chemotherapy intervention studies. Eur J Cancer 2006; 42: 3294–3303.

43.

de Boüard

Herlin

Christensen

. Antiangiogenic and anti-invasive effects of sunitinib on experimental human glioblastoma. Neuro Oncol 2007; 9: 412–423.

44.

Jung

Rho

Kim

. Upregulation of CXCR4 is functionally crucial for maintenance of stemness in drug-resistant non-small cell lung cancer cells. Oncogene 2013; 32(2): 209–221.

45.

Lee

Joo

Lee

. Targeting the epithelial to mesenchymal transition in glioblastoma: the emerging role of MET signaling. Onco Targets Ther 2014; 7: 1933–1944.

46.

Wang

Zhang

Siu

. Glioblastoma multiforme formation and EMT: role of FoxM1 transcription factor. Curr Pharm Des 2015; 21(10): 1268–1271.

47.

Zhou

Zhuang

Wang

. Tumor necrosis factor α induces myofibroblast differentiation in human tongue cancer and promotes invasiveness and angiogenesis via secretion of stromal cell-derived factor-1. Oral Oncol 2015; 51(12): 1095–1102.

48.

Kioi

Vogel

Schultz

. Inhibition of vasculogenesis, but not angiogenesis, prevents the recurrence of glioblastoma after irradiation in mice. J Clin Invest 2010; 120(3): 694–705.

49.

Lisi

Laudati

Navarra

. The mTOR kinase inhibitors polarize glioma-activated microglia to express a M1 phenotype. J Neuroinflamm 2014; 11: 125.

50.

Mignogna

Signorelli

Vismara

. A reappraisal of macrophage polarization in glioblastoma: histopathological and immunohistochemical findings and review of the literature. Pathol Res Pract 2016; 212(6): 491–499.

51.

Zhang

. Recent advance in molecular angiogenesis in glioblastoma: the challenge and hope for anti-angiogenic therapy. Brain Tumor Pathol 2015; 32(4): 229–236.

52.

Lin

Gnatovskiy

. Macrophages regulate the angiogenic switch in a mouse model of breast cancer. Cancer Res 2006; 66(23): 11238–11246.