Employing a Combination of Chemoattractants to Trap Glioblastoma Cells in a Macroporous Hydrogel

Abstract

Background: A new paradigm for treating glioblastoma multiforme (GBM) cells was proposed. Instead of trying to eliminate cancer cells infiltrated in the brain, this new treatment is based on attracting them into a macroporous gel-based trap, where they are retained and then irradiated with a localized, higher radiation dose. The objective of this study is to identify a cytokine combination that would attract GBM cells while considering heterogeneity among GBM cell lines. Methods: The ability of different combinations of cytokines CXCL12, IL-1β, IL-6, and EGF to stimulate the migration of the GBM cell lines U87, U87 CXCR4+, F98, and U118 was assessed with a two-layer Matrigel device that simulates the extracellular environment in brain. The accumulation of GBM cells within a cancer cell trap made from a macroporous hydrogel consisting of 1% alginate, 0.75% chitosan, and 0.05% genipin was determined. This hydrogel was grafted with RGD and features fully interconnected pores with an average diameter of 300 µm. CXCL12 is the most frequently used for attracting GBM cells. The other cytokines were chosen to enhance CXCR4 expression, the receptor for CXCL12, increase matrix metalloproteinase-2 and -9 (MMP-2 and -9) production, and promote the epithelial–mesenchymal transition (EMT), a phenotype shift that facilitates cell migration. Results: IL-1β significantly enhanced CXCR4 expression in the F98 and U118 cells. The production of MMP-2 was significantly stimulated with IL-1β and IL-6 in F98 cells. The combination of the cytokines IL-1β + IL-6 + CXCL12 + EGF, on the other hand, induced a decrease in MMP-2 levels. The EMT was induced by EGF in all GBM cells tested. The results obtained using the two-layer Matrigel device showed that the combination of the cytokines IL-1β + CXCL12 + EGF was the most effective in promoting the migration of the four GBM cell lines. Regarding accumulation in the macroporous hydrogel, U118 cells showed the best response to this cytokine combination. Conclusions: A significant challenge in developing a cancer cell trap is to identify a cytokine combination to attract the heterogeneous population of GBM cells. In this study, the cytokine combination IL-1β + CXCL12 + EGF was found to be the most effective in promoting the migration of GBM cells.

1. Introduction

Glioblastoma multiforme (GBM) is the most common type of malignant brain tumor, with a median survival time of only 14.6 months after diagnosis [1]. This underscores the urgent need for developing new treatments. The limited effectiveness of current therapies is primarily due to the high resistance of GBM cells to radiotherapy and the chemotherapeutic agent Temozolomide (TMZ) [2].

One significant challenge in treating GBM is delivering these chemotherapeutic agents to the peritumoral GBM cells and those that have infiltrated the brain’s stroma, which is complicated by the blood–brain barrier [3]. Additionally, subpopulations of GBM cells exhibit highly invasive characteristics, enabling them to penetrate the brain tissue up to 20 mm deep [4,5]. This characteristic also implies that the irradiated volume encompasses the tumor area, along with a 20–25 mm margin to account for the surrounding GBM cell clusters [5]. However, the maximum radiation dose that can be safely delivered is limited by the brain parenchyma’s sensitivity [6]. For example, the prescribed radiation dose delivered by stereotactic radiosurgery in one fraction is 16 Gy for a planning target volume between 10 and 20 cm³ [7]. Consequently, recurrence is almost inevitable [8].

To overcome these limitations, the concept of trapping cancer cells in a confined area to facilitate their elimination has emerged [9]. To this end, traps were developed to capture circulating and infiltrated tumor cells [10,11]. For example, circulating tumor cells (CTCs) can be isolated from blood with a hyaluronic gel loaded with the chemoattractant CXCL12 [12]. Others have been designed to attract and trap cancer cells that have infiltrated the brain tissue. Of these, the hydrogel based on hyaluronic/collagen II called GliaTrap is notable; it incorporates liposomes loaded with the chemoattractant CXCL12. GliaTrap is injected as a liquid, then quickly forms a gel at 37 °C and conforms to the shape of the surgical cavity after the tumor is removed. An orthotopic xenograft model demonstrated that GBM cells were efficiently attracted to the surface of the GliaTrap [13].

Random nanofibrous biomaterials based on bacterial cellulose (BC) are easy to synthesize and can be modified to create traps. Disks of BC membrane with small mesh sizes can efficiently trap F98 GBM cells attracted by conditioned media of glioblastoma-associated stromal cells [14].

Injectable hydrogels can also be loaded with anticancer drugs. Hydrogels made of lipid nanocapsules loaded with the anticancer drug gemcitabine (GemC12-LNC) have demonstrated the ability to slow down the recurrences of U87 cells in a xenograft orthotopic mouse model [15] and in 9 L tumor-bearing resected rats [16].

Although promising, such gels used to release chemotherapeutic drugs are not appropriate for trapping cancer cells. Their crosslinked polymer networks have a typical porosity of ≈100 nm or smaller, which is relevant to the release of molecular compounds [17,18,19]. However, larger interconnected pores (>50 μm) are needed to facilitate the penetration of cancer cells into a hydrogel [20]. Average pore size, pore interconnectivity, and other parameters such as gel “firmness”—i.e., the mechanical properties, need to be tuned in order to optimize the distribution and accumulation of cancer cells [20].

We have already prepared different gel formulations and achieved important milestones, and significant challenges have been addressed to meet the clinical requirements of this macroporous GBM cell trap. These challenges include ensuring structural stability and a compression modulus compatible with the biological environment, achieving the right formulation of the alginate (Alg) and chitosan (Chit) polymers that make up the hydrogel, optimizing pore diameter for the accumulation, distribution, and retention of GBM cells, and grafting the cell adhesion peptide CGGRGDS (RGD) onto the inner surface of the pores to enhance GBM cell retention. This adhesion peptide binds to the αvβ3 and αvβ5 integrins that are overexpressed by the GBM cells, and they correlate with tumor grade and poor survival [21,22]. The best hydrogel formulation for trapping GBM cells that our team has developed consists of 1% alginate, 0.75% chitosan, and 0.05% genipin, grafted with RGD and with fully interconnected pores with an average diameter of 300 µm [23,24,25,26]. This combination of polysaccharides has a number of functions: (1) alginate provides mechanical strength, whereas (2) chitosan crosslinked with genipin stabilizes the hydrogels in biological media and provides amine groups for cellular adhesion [23,27]. This hydrogel formulation was used in the present study. The encapsulation of the cytokine CXCL12 in Alg/Chit nanoparticles, its release rate under static conditions, and the impact of simulated cerebral interstitial fluid flow on the release of CXCL12 from an alginate-based hydrogel were also determined [24,25,26,28,29].

Chemoattractant agents, such as CXCL12, have been proposed to reverse the direction of GBM cell migration from the brain tissue toward the macroporous hydrogel [29,30,31]. The accumulation and retention of GBM cells within the macroporous hydrogel will enable the safe delivery of a higher radiation dose concentrated on the hydrogel. This approach aims to improve the effectiveness of radiotherapy while reducing adverse effects on the surrounding healthy brain tissue.

The heterogeneity of the cancer cell population within the tumor, as well as between patients, presents a challenge in identifying the cytokines and their concentration needed to attract GBM cells infiltrated in the brain parenchyma toward the macroporous hydrogel. Is it sufficient to use a single cytokine, or is a combination of them required? In this study, combinations of cytokines were assessed with the rodent GBM cell line F98 and the human cell lines U118, U87, and U87 CXCR4+. These latter GBM cells exhibit increased expression of CXCR4, the receptor of CXCL12 [32]. In combination with CXCL12, cytokines that promote different aspects of cancer cell migration were evaluated. IL-1β was studied because of its ability to increase CXCR4 expression and enhance cancer cell migration [33,34,35,36,37]. IL-1β also increases the level of matrix metalloproteinases MMP-2 and -9, which cleave extracellular matrix proteins, thereby creating a passage for cancer cells [34,35,36,37]. Regarding IL-6, its ability to increase MMP-2 and -9 production has been reported [38,39]. Epidermal Growth Factor (EGF) was selected for its ability to induce epithelial–mesenchymal transition (EMT), stimulate the cancer cell migration, and increase MMP-9 level [40,41].

A two-compartment migration chamber was utilized to assess the ability of these cytokines to attract the GBM cells. Their concentrations were determined based on their ability to be loaded into nanoparticles (NPs) that would be incorporated into the hydrogel trap. For Alg/Chit nanoparticles with an average size of ≈300 nm, the entrapment efficiency for CXCL12 reached ≈98%, with a mass loading varying from 0.372 to 1.490 µg CXCL12/mg NPs [29].

2. Materials and Methods

2.1. Cell Culture

The murine cell line F98 and human cell lines U87 and U118 were obtained from the American Type Culture Collection (Manassas, VA, USA), while the U87 CXCR4+ cells were provided by the NIH Reagent Program, catalog # 4036 (Germantown, MD, USA). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM glutamine, 1 mM sodium pyruvate, 100 units per mL penicillin, and 100 mM streptomycin at 37 °C and 5% CO₂.

2.2. Cell Lines Expressing Fluorescent Markers

Gene encoding for the fluorescent marker Green Fluorescent Protein (GFP) was incorporated into the U87 cells by transfection with the GFP/pLenti6/V5 construct [42]. The human 293 T cell line was transfected with the plasmids pLP1, pLP2, pLP/VSV-G, and the lentiviral plasmid GFP/pLenti6/V5. After 48 h of incubation, the supernatant containing the lentivirus was collected, filtered with a 0.45 µm membrane, and then 700 µL was added to the U87 cells in the presence of 4 µg/mL polybrene. After 48 h of incubation, the virus solution was replaced with culture medium containing blasticidin as a selection agent, and the cells were incubated for 10 days. Regarding the fluorescent marker mCherry, it was introduced into F98, U87 CXCR4+, and U118 cells by using the lentiviral vector pCDH-CMV-mCherry-T2A-Puro (Addgene Catalog#72264, Watertown, MA, USA). After transfection, cells were incubated for 10 days in complete DMEM containing the selection agent puromycin at 2.5 μg/mL. Expression of GFP or mCherry was confirmed by observations with a Leica DM-IRBE fluorescence microscope (Leica Microsystems, Wetzlar, Germany).

2.3. CXCR4 Receptor Activity

The dose–response of CXCR4 and its ligand CXCL12 was determined with the Fluo-4 NW calcium Assay Kit (Molecular Probes, Invitrogen, #F36206, Carlsbad, CA, USA), which measures the increase in the intracellular level of Ca²⁺ ions in response to CXCL12. The GBM cell lines U87, U87 CXCR4+, F98, and U118 (2 × 10⁴/well) were incubated with complete DMEM in 96-well black opaque plates with a clear bottom (Corning Incorporated Costar, Catalog#3603, Corning, NY, USA) for 24 h at 5% CO₂ and 37 °C. After removing the culture medium, 100 µL of dye loading solution supplied with Fluo-4 NW calcium assay kit was added to each well. After 30 min of incubation at 37 °C, the first fluorescence reading was taken to determine the background level using the plate reader HT Synergy (λex = 494 nm, λem = 516 nm) (Bio-Tek Instrument, Winooski, VT, USA). The second fluorescence reading was performed after adding CXCL12 at 2, 4, 6, 8, and 10 nM. In supplementary assays, the specific inhibitor of CXCR4, AMD11070 (Sigma-Aldrich, St. Louis, MO, USA), was added at 40 nM and 6 nM CXCL12.

2.4. Quantification of CXCR4 Expression by Semi-Quantitative PCR

F98, U87, U87 CXCR4+, and U118 cells were cultured in DMEM supplemented with 0.1% BSA, 5% CO₂ and 37 °C. Total RNA was extracted using the Absolutely RNA Microprep Kit (Stratagene, La Jolla, CA, USA) according to the manufacturer’s instructions. RNA integrity was evaluated using the Agilent 2100 Bioanalyzer (Agilent Technologies, Mississauga, ON, Canada). Reverse transcription was performed with 2 µg of total RNA in a 20 µL reaction volume using Transcriptor reverse transcriptase, random hexamers, dNTPs (Roche Diagnostics, Laval, QC, Canada), and 10 units of RNaseOUT (Invitrogen, Burlington, ON, Canada), following the manufacturer’s protocol. Semi-quantitative PCR (qPCR) reactions were carried out on a Realplex2 thermocycler (Eppendorf, Mississauga, ON, Canada). Each reaction contained 5 μL of 2x FastStart Universal SYBR Green Master Mix (Roche Diagnostics, Laval, QC, Canada), 10 ng (3 µL) of cDNA, and 200 nM (2 µL) of the primer pair solution. The cycling protocol included an initial denaturation step at 95 °C for 10 min, followed by 50 cycles of 15 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C. Relative expression levels were calculated using the qBASE framework with housekeeping genes MRPL19, PUM1, and YWHAZ for murine cells F98; Psmc4, Pum1, and Tbp for human U87 and U87 CXCR4+ cells; and L3MBTL2, TMEM199, and VAMP7 for U118 cells. No template controls were included for each primer pair in each qPCR run, consistently yielding negative results. The primer sequences are listed in Table S1.

2.5. Zymography

The increase in MMP-2 and MMP-9 levels in response to different concentrations of chemoattractants was determined by gelatin zymography [34,35]. F98 cells (15 × 10⁴/2 mL), U87, U87 CXCR4+, and U118 cells (5 × 10⁵/2 mL) in complete DMEM were seeded in 6-well plates. After 24 h of incubation, the medium was replaced with DMEM 0.1% BSA. Cells were incubated separately or with combinations of IL-1β (10 ng/mL), IL-6 (20 ng/mL), EGF (20 ng/mL), and CXCL12 (100 ng/mL) for 24 h. Supernatants were collected and concentrated by centrifugation in an Amicon ultra-4 filter (Millipore Sigma-Aldrich, St. Louis, MO, USA) at 4000 rpm for 10 min and then stored at −20 °C. Gel zymography was then performed as previously reported [34,35]. Bands corresponding to MMP-2 and -9 were revealed using a light plate and quantified with ImageJ (version 20250529-2217). The first well of each zymography gel was loaded with MMP-2 (0.07 ng/well) or MMP-9 (0.3 ng/well) to confirm their location and to calculate their relative quantities in the other wells.

2.6. EMT Stimulation by EGF

The stimulation of epithelial–mesenchymal transition in response to EGF was first evaluated via qPCR by determining the expression levels of VIM, CDH1, and CDH2 in cells of different GBM lines. GBM cells (10⁵ cells/well) were seeded in 12-well plates using standard DMEM. After 24 h of incubation, the culture medium was changed to DMEM 0.1% BSA and then EGF was added at a concentration of 20 ng/mL. VIM, CDH1, and CDH2 expression were determined after incubation for 24, 48, and 72 h.

The housekeeping genes that were used are MRPL19, PUM1, and YWHAZ. Primer sequences are provided in Table S2. Morphological changes associated with EMT were assessed by incubating U87 cells with or without EGF (20 ng/mL) for 48 and 72 h, followed by imaging using a Bioimager microscope (Bioimager Inc., Concord, ON, Canada).

2.7. In Vitro Migration System

A two-compartment in vitro migration system was developed to measure GBM cell attraction by a combination of cytokines through an artificial extracellular matrix called Matrigel (Growth Factor Reduced (GFR) Basement Membrane Matrix; Corning, Glendale, AZ, USA; #catalog: 356231) (Figure 1). A 2 mm layer of Matrigel/DMEM 0.1% BSA (1:1) was deposited on the porous membrane of the migration chamber (TC-Insert, Sarstedt, Catalog#83.3922.800, Nümbrecht, Germany), followed by a second 2 mm layer of Matrigel/DMEM 0.1% BSA (1:1) containing the GBM cells (50 × 10⁴ cells/mL). U87 cells were expressing the fluorescent marker GFP, while the marker mCherry was inserted in U87 CXCR4+, F98, and U118 cells. The inserts were transferred to 24-well plates containing different combinations of the cytokines IL-1β (10 ng/mL), IL-6 (20 ng/mL), EGF (20 ng/mL), and CXCL12 (100 ng/mL) in the bottom. After 2, 24, and 48 h, the migration of GBM cells from the upper layer to the lower layer was determined using the EVOS™ FL Auto Imaging System epifluorescence microscope (Life Technologies, Carlsbad, CA, USA). F98, U118, and CXCR4+ cells labeled with mCherry were visualized using the RFP spectrum (excitation: 531 ± 40 nm; emission: 593 ± 40 nm), while GFP-labeled U87 cells were visualized with the GFP light Cube spectrum (excitation: 470 ± 22 nm; emission: 525 ± 45 nm).

The images obtained were then processed and analyzed with ImageJ software to determine the number of cells present at the focal point in 5 levels, as illustrated in Figure 1. The percentage of relative migration was then calculated as follows.

$$\%\mathrm{R}\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e} \mathrm{m}\mathrm{i}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}={\sum }_{\mathrm{L}=1}^5\left(\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s} \mathrm{i}\mathrm{n} \mathrm{L}\mathrm{i}÷\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s} \mathrm{i}\mathrm{n} \mathrm{L}1 \mathrm{t}\mathrm{o} \mathrm{L}5\right)\times 100.$$

2.8. Wound Healing Assay

F98, U87, U87 CXCR4+, and U118 cells were seeded in Petri dishes at a density of 1 × 10⁵ cells/mL. Once 90–100% confluence was reached, a linear scratch was made in the center using a sterile tip. Two PBS washes were performed to remove detached cells, after which DMEM supplemented with 0.1% BSA was added with or without IL-1β (10 ng/mL), CXCL12 (100 ng/mL), and EGF (20 ng/mL). Wound closure was monitored over time (0, 24, 48, and 72 h) using the EVOS™ FL Auto Imaging System epifluorescence microscope (Life Technologies, Carlsbad, CA, USA). Images were analyzed using ImageJ software to measure the residual wound area at each time point and calculate the percentage of closure over time. The wound healing size tool plugin in ImageJ/Fiji (version 20250529-2217) was used to automatically segment the photos obtained by using the EVOS. This method relies on the application of a variance filter to distinguish the cellular area (monolayer) from the empty area (the “wound”), then binarizes the image, fills in the holes, and identifies the largest open region as the “wound area” (https://github.com/AlejandraArnedo/Wound-healing-size-tool/wiki, accessed on 6 October 2025).

2.9. Migration to Macroporous Hydrogels

2.9.1. Hydrogel Preparation

Macroporous hydrogels made of chitosan (CHI, 85/60/A1 grade, BioLog Heppe GmbH, GER) and sodium alginate (SA, Kimica Corporation, Santiago, Chile) crosslinked with genipin (GNP, purity ~98%, Thermo Fisher Scientific, Waltham, MA, USA) and functionalized with the RGD peptide (CGGRGDS, 98%, N-terminal acetyl, C-terminal amide, EZBiolab, Carmel, IN, USA) were prepared from melt-processed porous polylactic acid (PLA) molds. The full procedure is described in a previous publication [23]. Briefly, polymer rods composed of a co-continuous blend of PLA (Ingeo 4032D, NatureWorks, Plymouth, MN, USA) and polystyrene (PS, MC3650, Americas Styrenics, The Woodlands, TX, USA) were melt-extruded at 190 °C, then annealed under a hot press at 190 °C for 30 min to let the morphology coarsen. The annealed bars were then trimmed by CNC machining to obtain 5 mm× 4 mm cylinders. The PS phase was finally extracted with cyclohexane in a soxhlet apparatus for a week, and the resulting porous PLA molds were dried overnight in a vacuum oven at 40 °C.

Next, the PLA molds were plunged in falcons containing a solution of RGD-functionalized SA (1% w/v, 3.5 × 10⁻⁷ mol of cysteine-terminated RGD per 400 mg of SA), CHI (0.75% w/v), and genipin (0.05% w/v based on CHI content), which were deposited in a custom-built injection system applying 3 to 4 vacuum N₂ pressure cycles to fill the molds. Once filled, the molds were placed in closed Eppendorf tubes and deposited in an oven at 37 °C for 24 h to crosslink CHI with genipin. Following crosslinking, the filled molds were placed in a 0.22 µm filtered 4% w/v CaCl₂ solution at 4 °C overnight to complete SA gelation. The macroporous gels were finally obtained by dissolving the PLA molds in chloroform, followed by thorough rinsing at least twice in MilliQ water (Darmstadt, Germany) under stirring for a day on a shaker plate. The gels were then stored in a 4% w/v CaCl₂ solution at 4 °C until use. The resulting porosity of the macroporous hydrogels consisted of fully interconnected macropores with an average diameter of 300 µm and a volume fraction of 45% v/v.

2.9.2. Migration Assay

The hydrogels were deposited onto the porous membrane of a migration chamber insert (TC-Insert, Sarstedt, Catalog#83.3922.800, Nümbrecht, Germany), and then Matrigel/DMEM (1:1) was poured around the hydrogels, with a thin layer applied on top. GBM cells U87 and U118 (50 × 10⁴ cells/mL) loaded in a 2 mm layer of Matrigel/DMEM (1:1) were then added to the top of the hydrogel (Figure 2). The inserts containing the hydrogel and the GBM cells were transferred to 24-well plates with or without IL-1β (10 ng/mL), CXCL12 (100 ng/mL), and EGF (20 ng/mL) in the bottom. After 72 h of incubation, the hydrogels were gently removed, transferred to 24-well plates, and fixed with a solution of paraformaldehyde (PFA) 3% w/v for 15 min at room temperature. The hydrogels were next washed twice with DMEM for 5 min, immersed in a solution of 1% BSA (blocking agent) for 30 min, and then washed twice with DMEM for 2 min. The cells were then stained by incubating the hydrogels in a solution of Hoechst 33,342 (nuclei, blue channel) at 5 μg/mL in 0.1% w/v red phenol-free BSA/DMEM for 30 min at room temperature in the dark. Finally, they were washed twice with DMEM for 2 min. The epifluorescence microscope of the EVOS™ FL automated imaging system was used to quantify the distribution of GBM cells at the focal plane across five equidistant levels in the hydrogels. This measurement began approximately 1 mm below the top face of the hydrogels (referred to as L1) and extended to about 4.5 mm (L5) [26]. Hoechst-labeled U87 and U118 cells were visualized using the DAPI spectrum (excitation: 357 ± 44 nm; emission: 447 ± 60 nm).

3. Results

3.1. Concentration of CXCL12 That Activates the CXCR4 Receptor

CXCL12 is the central chemokine around which others will be added to optimize the migration of GBM cells towards a macroporous hydrogel trap. The Fluo-4 NW Calcium Assay kit was used to determine the concentration of CXCL12 that activates its receptor, CXCR4, on GBM cell lines U87, U87 CXCR4+, F98, and U118. This test measures the increase in cytoplasmic Ca²⁺ released from the endoplasmic reticulum following activation of CXCR4, a pathway involved in the migration and invasion of cancer cells [43,44] (Figure 3). Increasing concentrations of CXCL12 from 2 to 10 nM were used. The CXCR4 allosteric antagonist AMD11070 (AMD 40 nM) followed by 6 nM CXCL12 treatment was also used to verify the involvement of the CXCR4 receptor.

The highest and most sustained CXCR4 activation for U87 cells was obtained with 2 nM CXCL12 (p = 0.003 vs. negative control at 225 s). In the U87 CXCR4+ cell line, which expresses CXCR4 2 times more than the U87 wild-type (Figure 4), the highest activation of CXCR4 measured by the increase in intracellular Ca²⁺ was observed at 6 nM CXCL12 (p = 0.0002 vs. negative control at 225 s), reaching a plateau 2.4 times higher than measured in the U87 cells. Regarding the F98 cells, their CXCR4 expression was 2.3 times more elevated than in the U87 cells (Figure 4), resulting in 1.6 times higher activation at 6 nM CXCL12, which is similar to a previously reported value [28] (p = 0.003 for F98 vs. U87 at 675 s). For the U118 cells, 1.8-fold higher CXCR4 activation at 6 nM CXCL12 was obtained compared to the U87 cells. Activation of CXCR4 was also more sustained and stable over time for the cell lines U87 CXCR4+, F98, and U118 compared to U87 cells. In these four GBM cell lines, the level of CXCR4 activation was proportional to its expression (Figure 4).

A concentration of 6 nM (52.2 ng/mL) CXCL12 was chosen for the subsequent study involving the CXCR4 allosteric antagonist AMD11070. The role of the CXCR4 receptor was confirmed by adding AMD11070 (40 nM), followed by the addition of 6 nM CXCL12. This treatment completely inhibited CXCL12-induced CXCR4 activation in all GBM cell lines tested, supporting that this pathway is active (U87 cells, p < 0.005; U87 CXCR4+ cells, p < 0.005; F98 cells, p = 0.0023; U118 cells, p < 0.005) (Figure 3B).

3.2. Role of IL-1β in Increasing Expression of CXCR4

IL-1β can enhance CXCR4 gene expression, which, in turn, increases cell migration in response to CXCL12 [33]. To establish a link between IL-1β and the enhanced migration of GBM cells in response to CXCL12, CXCR4 expression levels across the four GBM cell lines were assessed by semi-quantitative PCR, normalized to the expression of housekeeping genes (Figure 4). In U87 cells, which have lower CXCR4 expression, no significant change was observed in the presence of IL-1β (p = 0.45). Similarly, U87 CXCR4+ cells showed no substantial difference (p = 0.25). However, treatment with 10 ng/mL IL-1β resulted in a significant 1.5-fold increase in CXCR4 expression in F98 cells compared with untreated F98 cells (p = 0.006). In U118 cells, a significant, although more modest, increase of 1.3 times induced by IL-1β treatment was measured (p = 0.038). While our analysis was limited to four cell lines, these results indicate that IL-1β may not be a universal sensitizer for CXCL12-based attraction strategies.

3.3. Increase in Matrix Metalloproteinases MMP-2 and -9 Levels

MMP-2 and -9 play an important role in cancer cell invasion by cleaving proteins of the extracellular matrix, thereby opening a passage for cancer cells [34,35,36,37]. The ability of the GBM cell lines to produce these MMPs in the presence of IL-1β (10 ng/mL) [36], IL-6 (20 ng/mL) [45], EGF (20 ng/mL) [46], CXCL12 (100 ng/mL) [47], and their combinations were determined by gel zymography (Figure 5). In U87 cells, a modest, non-statistically significant increase in MMP-2 was measured with CXCL12 and the combination of all cytokines. In U87 CXCR4+ cells, MMP-2 was lower in control without cytokines than in U87 cells. While IL-1β and EGF induced a slight, though not statistically significant, increase in MMP-2 levels, the combination of four cytokines (IL-1β, IL-6, EGF, and CXCL12) resulted in a reduction in MMP-2. In F98 cells, MMP-2 was significantly increased with both IL-1β and IL-6 (p = 0.04 and p = 0.007, respectively). No increase in MMP-2 was measured in U118 cells. Regarding MMP-9, it was detectable only in F98 cells when incubated with IL-1β, CXCL12, and the combination of cytokines.

3.4. EMT Induction by EGF

Morphological changes associated with EMT were evaluated after incubation with 20 ng/mL EGF. After 48 h of incubation, the morphology of U87 cells changed from their initial squamous shape to an elongated, spindle-like morphology, characterized by specialized membrane structures known as lamellipodia. This spindle-like appearance became more pronounced at 72 h, with many cells exhibiting multiple cytoplasmic extensions (Figure 6A, red arrows). Quantification of lamellipodes for the cell lines U87, U87 CXCR4+ CXCR4+, F98, and U118 is reported in

Table 1. Spindle-like elongation induced by EGF.

		Control		EGF
Cells Time (h)		% Cells with Lamellipodes	Average Number of Lamellipodes/Cell	% Cells with Lamellipodes	Average Number of Lamellipodes/Cell
U87	24 h	26.1 ± 3.3	1.97 ± 0.04	34.5 ± 3.02	2.38 ± 0.38
	72 h	31.9 ± 4.0	1.34 ± 0.11	41.4 ± 1.2	1.62 ± 0.04
U87 CXCR4+	24 h	28.5 ± 8.2	0.91 ± 0.11	38.9 ± 1.8	1.32 ± 0.14
	72 h	36.3 ± 1.6	0.74 ± 0.07	39.7 ± 1.6	1.02 ± 0.06
F98	24 h	29.1 ± 4.7	2.32 ± 0.12	40.2 ± 4.4	1.82 ± 0.25
	72 h	33.4 ± 1.2	1.73 ± 0.07	40.4 ± 0.2	1.34 ± 0.11
U118	24 h	39 ± 14.6	2.40 ± 0.46	46.2 ± 4.4	2.26 ± 0.91
	72 h	42.8 ± 1.6	1.04 ± 0.05	48.9 ± 0.2	1.70 ± 0.08

. This analysis demonstrates that EGF increased the average percentage of cells displaying lamellipodes by 31% after 24 h of incubation compared to the control that did not receive this cytokine. The most significant increases were observed in U87 cells (p = 0.045), F98 cells (p = 0.008) and U118 cells (p = 0.021). After 72 h, the formation of lamellipodia continued; however, the effect of EGF was less pronounced. These results indicate a cell transition from an exploratory morphology, characterized by several thin and unstable extensions, to a polarized and elongated morphology, favoring directional migration. This phenomenon reflects a shift from amoeboid/exploratory migration to polarized mesenchymal migration [48].

Changes in two markers related to EMT, E-cadherin and N-cadherin, were measured by qPCR for U87 cells (Figure 6B). Gene expression of the epithelial marker E-cadherin was significantly reduced at both 48 and 72 h in response to EGF, with a particularly notable decrease at 72 h (p = 0.007). On the other hand, the expression of N-cadherin, which is associated with the mesenchymal phenotype, increased 1.7-fold in U87 cells incubated with EGF at 72 h (p = 0.023) (Figure 6B).

3.5. Attraction of GBM Cells by Cytokines Through Matrigel Layers

The cytokine concentrations (IL-1β, 10 ng/mL [36]; IL-6, 20 ng/mL [45]; EGF, 20 ng/mL [46]; and CXCL12, 100 ng/mL [47]) were chosen for their ability to create a gradient that increases the migration rate of cancer cells, while allowing for their encapsulation within nanoparticles that can be incorporated into a macroporous hydrogel designed to trap cancer cells.

We evaluated the effectiveness of various combinations of these cytokines in attracting GBM cells by mixing them with a layer of Matrigel, which simulates the extracellular environment in brain. Their migration toward a second layer of Matrigel enabled us to determine how effectively each cytokine combination promoted cell movement.

The migration of GBM cells loaded into an upper Matrigel layer toward a lower layer was measured using an EVOS™ FL Auto Imaging System epifluorescence microscope after 2, 24, and 48 h of incubation. The EVOS microscope was used because it allows for imaging at different depths within the hydrogels. Since the cells are distributed in a three-dimensional arrangement in the Matrigel layers, some cells appear blurry when images are captured at a specific depth. In this study, only those that were clearly in focus at specific levels of the hydrogel were included in the count.

U87 cells expressed the fluorescent marker GFP, while the marker mCherry was inserted in U87 CXCR4+, F98, and U118 cells. The number of cells was determined at five equidistant levels (L1 to L5), with the third level positioned at the interface between the upper and lower Matrigel layers, as described in the Materials and Methods Section.

We have shown that IL-1β enhances the expression of CXCR4 in U118 cells, which is the receptor for CXCL12. The impact of increasing CXCR4 expression on U118 cell migration in the Matrigel layer was assessed by adding IL-1β and CXCL12, alone or in combination, to the bottom compartment of the migration setup (Figure 7).

After 2 h of incubation, U118 cells did not migrate significantly in the absence of these cytokines, as their distribution remained similar to the initial measurement (T = 0, light gray bars), primarily located in the L1 area of the upper Matrigel layer. On the other hand, the ability of IL-1β and CXCL12, alone or in combination, to promote the migration of U118 cells had already been observed. A significant number of U118 cells migrated toward L4 in the lower Matrigel layer (Figure 7). At 24 h, a significant increase in U118 cell migration to L4 was measured only with CXCL12 alone and combined with IL-1β. The U118 cell control also reached the L4 stage, which may be linked to their inherent ability to express MMP-2 and migrate [49]. At 48 h, no further increase in U118 cell accumulation in L4 was observed with CXCL12 alone and combined with IL-1β, and none were detected in the deepest L5 layer. This loss of migration-stimulating cytokines may result from the gradual distribution of cytokines, which becomes uniform throughout the Matrigel layer, thereby destroying the cytokine gradient.

These observations show that the combination of IL-1β and CXCL12 significantly stimulates U118 cell migration. To simplify the potential cytokine combinations, CXCL12 was paired with IL-1β, to which IL-6 and EGF were added, and tested in four GBM cell lines.

Figure 8 illustrates F98 mCherry cells that have reached the L4 level in the lower Matrigel layer, as determined with the EVOS™ FL Auto Imaging System epifluorescence microscope.

Significant migration of F98 cells toward L3, the deepest layer in the upper Matrigel layer, was already observed after 2 h of incubation with a combination of IL-1β + CXCL12 + EGF, compared to their initial distribution (T = 0, light gray bars) (Figure 9). After 24 h of incubation, notable migration of F98 cells toward L4 in the lower Matrigel layer was recorded with the cytokine combinations IL-1β + CXCL12 + EGF and IL-1β + CXCL12 + EGF + IL-6. Contrary to U118 cells, this enhancement was not observed in F98 cells with the combination of IL-1β + CXCL12 compared to control cells (Figure 8 and Figure 9). At 48 h of incubation, a significant enhancement in migration continued to occur with the combination of IL-1β + CXCL12 + EGF, followed by the combination of IL-1β + CXCL12 + EGF + IL-6, resulting in the majority of F98 cells reaching the L4 layer (p < 0.01 and p < 0.05, respectively).

To simplify the determination of the most effective cytokine combinations for promoting migration of the cell lines U87, U87 CXCR4+, and U118 into the lower Matrigel layer (L4 and L5), only the results obtained after 48 h of incubation are presented (Figure 10).

For U87 cells, the combination of IL-1β + CXCL12 did not improve their migration in the L4 level of the lower Matrigel layer, compared to the control without these cytokines. With both cytokine combinations of IL-1β + CXCL12 + EGF and IL-1β + CXCL12 + EGF + IL-6, nearly 100% of U87 cells were located in L4 after 48 h of incubation (Figure 10).

Contrary to expectations, the overexpression of CXCR4 did not increase the migration of U87 CXCR4+ cells compared to wild-type U87 cells when both were incubated with IL-1β + CXCL12. Although fewer U87 CXCR4+ cells than U87 cells remained in the upper layer (L1), there was no significant increase observed in the deeper layers (L3 and L4) compared to the no-cytokine control. The addition of EGF significantly enhanced their migration toward L4 (p = 0.003 compared to IL-1β + CXCL12), whereas IL-6 did not lead to further improvement. A similar trend was observed in both U87 and U87 CXCR4+ cells; however, the response was weaker for U87 CXCR4+ cells. Approximately 70% of U87 CXCR4+ cells reached layer L4, compared to nearly 100% of U87 cells, and none moved further to layer L5.

U118 cells demonstrated the highest migration capacity, both in the presence and absence of cytokines. In the no-cytokine control, approximately 75% of the cells accumulated in layer L3 of the upper Matrigel layer. However, to significantly migrate into the lower Matrigel layer, the addition of cytokines was necessary (Figure 10). The three combinations of cytokines tested similarly promoted migration to layer L4. The introduction of EGF enabled the cells to reach layer L5, and the presence of IL-6 further enhanced their migration. These findings indicate that U118 cells have the highest sensitivity to IL-6 among the cell lines studied.

3.6. Cell Migration Assessed by Wound Healing Assay

A wound healing assay was performed to confirm the ability of the best cytokine combination, i.e., IL-1β, CXCL12, and EGF, to promote the migration of GBM cells.

The enhancement of cell migration was associated with the intrinsic migratory capacity of the GBM cell lines. Among the cell lines tested, F98 cells exhibited the slowest scratch closure, achieving only 37.4% closure after 48 h in the absence of cytokines. In contrast, the other GBM cell lines achieved an average closure of 71.5%, the fastest being U118 cells. Interestingly, the most significant improvement in migration with the addition of cytokines was observed for the slowest cells, F98, which reached 74.6% of scratch closure after 48 h of incubation, an improvement of 37.2% (p = 0.0019 vs. negative control). In comparison, the other cell lines only improved by 21.3% (Figure 11). Nevertheless, the addition of the cytokine combination remained beneficial for the U87, U87 CXCR4+, and U118 cell lines, allowing more than 92.8% of the scratches to close.

3.7. Attraction and Accumulation in a Macroporous Hydrogel

The final stage of this study aimed to validate the capacity of the cytokine combination IL-1β + CXCL12 + EGF to attract and promote the accumulation of GBM cells within the macroporous hydrogel made with 1% alginate, 0.75% chitosan, and 0.05% genipin, grafted with adhesion peptide RGD and presenting an average pore diameter of 300 μm [23]. The migration assays were performed with the GBM cells exhibiting the best response to the cytokine combination, the U87 and U118 cell lines. Their distribution into the hydrogel could not be determined using the fluorescent markers GFP or mCherry and the EVOS™ FL Auto Imaging System epifluorescence microscope because of the presence of genipin in the hydrogel.

The GBM cells were loaded onto a layer of Matrigel that was deposited on the top surface of the hydrogels, while the cytokine combination was added to the lower compartment of the migration chamber (Figure 12). After 72 h of incubation, the hydrogels containing the cells were fixed and stained with Hoefscht, allowing them to be detected using the EVOS microscope (Figure S1). The number of cells at the focal point for each level (L1 to L5) was measured to calculate the percentage of relative cell migration through the hydrogel.

In the absence of cytokines, the migration of U87 cells through the hydrogel was restricted to the upper levels, primarily in L1 (p = 0.0007 vs. the combination of cytokines), and none were observed in the deeper levels, L4 and L5 (Figure 13). In contrast, the addition of the cytokine combination promoted their migration to deeper levels, particularly at L2, L3, and L4, while only a few reached L5. This increase was highly significant, for example, at L3 (p = 0.001 vs. negative control) (Figure 13).

U118 cells exhibited more efficient migration in the absence of cytokines compared to U87 cells (Figure 13). U118 cells were more uniformly distributed at the L1 to L3 levels, with many reaching the deeper L4 and L5 levels after 72 h of incubation. Adding the cytokine combination further enhanced their migration capacity, allowing U118 cells to reach the deepest levels of the hydrogels, L4 and L5, in greater numbers. Approximately 26% of U118 cells migrated down to level L5 (p = 0.01 vs. negative control), with a maximal cell count measured at L4 (p = 0.005 vs. negative control).

4. Discussion

The overall objective is to identify a cytokine combination that would attract brain-infiltrating GBM cells into a macroporous hydrogel used as a cancer cell trap. Identifying these cytokines can be complex, given the intratumoral and interindividual heterogeneity of GBM cell populations.

The cytokine CXCL12 has been frequently chosen as a chemoattractant to promote the migration of GBM cells because its receptor, CXCR4, is overexpressed in 57% of primary GBM tumors [50]. For example, a porous 3D scaffold made of Bombyx mori silk was implanted into the rotator cuff of mice. CXCL12 was released from this scaffold using a mini pump, which allowed for the capture of circulating tumor cells originating from a tumor implanted in the mice [51]. Portella et al. also successfully captured circulating tumor cells from blood samples of patients with metastases using a hyaluronic acid-based hydrogel that released CXCL12 [12].

In the present study, we determined the ability of different combinations of the cytokines CXCL12, IL-1β, IL-6, and EGF to stimulate the migration of four GBM cell lines through a Matrigel layer into a macroporous hydrogel. These cytokines are involved in three pathways of cancer cell invasion: stimulating CXCR4 expression, increasing MMP-2 and MMP-9 levels, and inducing EMT. The individual and combined effects of these cytokines, which vary by cell type and extracellular environment, complicate the determination of their contributions to the mechanisms underlying cancer cell invasion [52]. In addition, a dose–response study is limited by the amount of these cytokines that can be encapsulated in the nanoparticles that will subsequently be inserted into the macroporous hydrogel [29]. Therefore, our results have not been analyzed to assess how these cytokines, at varying concentrations, affect migration markers in each GBM cell line. Instead, the objective was to identify a cytokine combination that would efficiently attract GBM cells into a macroporous hydrogel.

Since CXCL12 stimulates GBM cell migration via CXCR4 [53], the effect of IL-1β on CXCR4 expression was evaluated. IL-1β significantly increases CXCR4 expression in the F98 and U118 cell lines, but not in the U87 WT and U87 CXCR4+ cell lines. These results highlight that the effect of IL-1β varies across GBM cell lines. This variability may be due to differences in sensitivity to the IL-1R/NF-κB pathway. This pathway involves IL-1β, which activates NF-κB transcription, resulting in increased CXCR4 expression at both the gene and protein levels, thereby promoting cell migration in response to CXCL12 [54].

Similar observations of CXCR4 expression not being stimulated by IL-1β were noted with the GBM cell lines LN827, LN992, and LN427 [55]. These results indicate that other cytokines, such as TNF-α, which can increase CXCR4 expression [56], could be assessed for possible inclusion in a cytokine combination to attract GBM cells.

The cleavage of extracellular matrix proteins contributes to the infiltration of GBM cells into the brain. CXCL12, IL-1β, IL-6, and EGF were assessed alone and in combination. These cytokines have already shown their ability to increase the levels of MMP-2 and -9 in different cancer cell lines [38,39,57,58]. However, under the experimental conditions used in our study, a significant elevation in MMP-2 production was observed only in F98 cells treated with both IL-1β and IL-6. Interestingly, the cytokine combination IL-1β + IL-6 + CXCL12 + EGF led to the opposite result, a reduction in MMP-2 levels. Regarding MMP-9, higher levels were observed only in F98 cells incubated with IL-1β, CXCL12, and the cytokine combination IL-1β + IL-6 + CXCL12 + EGF. These results further support heterogeneous responses among the four GBM cell lines studied and indicate that some cytokine combinations may result in lower MMP-2 production compared to the cytokines used individually. This suggests that some cytokines may compete with one another, or that a combination of cytokines could activate a negative feedback loop or inhibitory pathway that reduces MMP-2 levels.

The ability of cytokines to promote GBM cell migration was first assessed using a two-layer Matrigel device that simulates their migration through the extracellular environment in the brain. GBM cells were loaded into an upper Matrigel layer and their migration through a lower Matrigel layer was determined. The assays began with IL-1β and CXCL12. U118 cells were used because we have shown that IL-1β increases the expression of the CXCL12 receptor, CXCR4, in these cells. The ability of IL-1β and CXCL12, alone or in combination, to promote the migration of U118 cells was already observed after 2 h of incubation. The best stimulation of U118 cell migration was observed with the combination of IL-1β and CXCL12. No U118 cells reached the deepest level in the lower layer of Matrigel, which may be due to a loss of the cytokine gradient caused by their gradual distribution throughout the Matrigel layer. To simplify the potential number of cytokine combinations, CXCL12 was paired with IL-1β, to which IL-6 and EGF were added, and tested in four GBM cell lines.

For the four GBM cell lines, the cytokine combination IL-1β + CXCL12 + EGF was the most effective in promoting their migration through the lower Matrigel layer. The addition of IL-6 to this combination did not result in a significant increase in their migration to the deepest level. It is noteworthy that overexpression of CXCR4, the receptor of CXCL12, did not result in a better response of U87 CXCR4+ cells compared to U87 cells when incubated with the cytokine combination IL-1β + CXCL12 + EGF. Among the cell lines tested, the U118 cell line showed the best response, being the only one to reach the deepest level in the lower Matrigel layer. In contrast, the wound healing assay results indicate that F98 cells exhibited the slowest migration in the absence of cytokines but the most significant improvement in migration speed when treated with the cytokine combination of IL-1β+ CXCL12 + EGF. These findings underscore the need to identify a combination of cytokines to address the heterogeneity of GBM cells. Further investigation in an animal model is necessary to evaluate how different cytokine combinations affect the migration of GBM cells infiltrated in the brain toward the macroporous hydrogel and to determine whether different subgroups of GBM cells respond differently. A key feature of the cancer cell trap is the macroporous hydrogel. The development of macroporous hydrogels, also known as matrices or sponges, has driven a new paradigm for treating GBM tumors [9,14,29,31,59]. Instead of trying to eliminate cancer cells infiltrated in the brain, this new treatment is based on their attraction into a gel-based trap where they will be retained and then irradiated with a localized and higher radiation dose. Classic gels used to release drugs and implanted in the brain are not appropriate to trap cancer cells. These gels are composed of a crosslinked polymer network with a typical porosity of ≈100 nm or smaller, allowing for the release of molecular compounds [17,18,19]. However, in order for cells to penetrate a hydrogel, larger interconnected pores (>50 μm) are needed [20].

The primary properties of these hydrogels must be their biocompatibility, stability in aqueous solutions, and ability to accumulate and retain cancer cells. The pore diameter of our formulation of hydrogels can be controlled within a range of 10 to 1000 µm, and full interconnectivity of pores can be achieved through a preparation process using co-continuous polymer blends as templates for porous hydrogel preparation [60,61]. The best hydrogel formulation that our team has developed consists of 1% alginate, 0.75% chitosan, and 0.05% genipin, with fully interconnected pores with an average diameter of 300 µm and grafted with cell adhesion peptide RGD (CGGRGDS), which improves the accumulation and retention of GBM cells [24,25,26].

The complementary electrostatic interactions between the negatively charged carboxyl groups of sodium alginate interact with the positively charged amino groups of chitosan and stabilize the hydrogel [62]. In contrast, pure alginate hydrogels formed in the presence of divalent cations such as Ca²⁺ are unstable in aqueous environments due to the passive diffusion of Ca²⁺ out of the hydrogel [25]. The hydrogel structure was further stabilized by chemically crosslinking chitosan polymers with genipin, a natural crosslinker, resulting in a compression modulus of 10 kPa, which is close to the value of brain tissue [23].

The feasibility of attracting GBM cells implanted in the brain of a nude rat towards a macroporous gel has been convincingly demonstrated by Molina-Peña et al. [59]. Their SF-HA-Hep sponges, with 69 µm pores, were implanted 1 mm away from a U87 CXCR4+ tumor. These GBM cells were attracted by the CXCL12 released from the sponge. However, their migration stopped in the intermediate zone located between the edge of the tumor and the edge of the sponge facing it. The inability to reach the hydrogel may be due to the difficulty of maintaining a gradient in the dynamic fluid environment of the peritumoral parenchyma [28].

Previous proof-of-concept studies on stimulating cancer cell migration and their entrapment in hydrogels have evaluated only a single cytokine used as chemoattractant [12,59]. These studies have overlooked tumor cell heterogeneity, which may respond differently to cytokines.

In our study, we have determined the capacity of the cytokine combination IL-1β + CXCL12 + EGF to attract and promote the accumulation of GBM cells within the macroporous hydrogel made with 1% alginate, 0.75% chitosan, and 0.05% genipin, grafted with adhesion peptide RGD and presenting an average pore diameter of 300 μm [23]. The assays were performed with the U87 and U118 cell lines, which showed the best response to this cytokine combination. The GBM cells were loaded onto a layer of Matrigel that was deposited on the top surface of the hydrogels, while the cytokine combination was added to the lower compartment of the migration chamber.

The addition of the cytokine combination was needed to significantly promote the migration of U87 cells to deeper levels of the hydrogel, particularly at L2, L3, and L4, while only a few reached the deepest level, L5.

U118 cells exhibited more efficient migration and distribution into the hydrogels in the absence and presence of cytokines. These results correlate with those obtained with the two-layer Matrigel device, where U118 cells were the best responders, being the only ones to reach the deepest level in the lower Matrigel layer. Adding the cytokine combination promoted their migration into the hydrogels, as they were able to reach the deepest levels of the hydrogels, L4 and L5, in greater numbers.

The next step will be to determine the maximum concentrations of the cytokine combination IL-1β + CXCL12 + EGF that can be encapsulated into alginate/chitosane nanoparticles [29], and how many of them need to be inserted into a hydrogel to attract all GBM cells loaded into an adjacent Matrigel layer. These results will then have to be validated in an animal model.

5. Statistical Analysis

The results are expressed as the mean ± standard deviation of the minimum. Two experiments were performed in triplicate. Statistical analyses were performed using two-way analysis of variance (ANOVA). p < 0.05 was considered statistically significant. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

6. Conclusions

The main challenge in developing a cancer cell trap is now attracting all GBM cells to macroporous hydrogels. In this study, the cytokine combination IL-1β + CXCL12 + EGF stimulated the migration of the GBM cell lines tested. Nevertheless, activating the major pathways involved in GBM cell migration using appropriate cytokine combinations and ensuring that the cytokine gradient is strong enough to allow for the accumulation of GBM cells into a macroporous hydrogel trap remain challenging. It will also be essential to maintain this cytokine gradient while accounting for the brain interstitial fluid flow [14]. In subsequent studies using an orthotopic GBM murine model, the ability of nanoparticles to release a continuous and strong cytokine gradient that attracts GBM cells infiltrating the brain stroma must be assessed.