Full length article
A new glioblastoma cell trap for implantation after surgical resection
Lila Autier
a,b,c, Anne Clavreul
a,b,⇑, Maximiliano L. Cacicedo
d, Florence Franconi
e,f, Laurence Sindji
b, Audrey Rousseau
b,g, Rodolphe Perrot
h, Claudia N. Montero-Menei
b, Guillermo R. Castro
d, Philippe Menei
a,baDépartement de Neurochirurgie, CHU, Angers, France
bCRCINA, INSERM, Université de Nantes, Université d’Angers, Angers, France
cDépartement de Neurologie, CHU, Angers, France
dNanobiomaterials Lab, CINDEFI, School of Sciences, National University of La Plata-CONICET (CCT La Plata), Buenos Aires, Argentina
ePRISM, Plate-forme de recherche en imagerie et spectroscopie multi-modales, PRISM-Icat, UNIV Angers, Angers, France
fMINT, Micro & Nanomedecines Translationnelles, UNIV Angers, INSERM U1066, CNRS UMR 6021, Angers, France
gLaboratoire Pathologie Cellulaire et Tissulaire, CHU, Angers, France
hSCIAM, Service Commun d’Imageries et d’Analyses Microscopiques, UNIV Angers, Angers, France
a r t i c l e i n f o
Article history:
Received 27 July 2018
Received in revised form 9 November 2018 Accepted 18 November 2018
Available online 19 November 2018
Keywords:
Bacterial cellulose Biomaterial Cell trap Glioblastoma
a b s t r a c t
Glioblastoma (GB) is a highly infiltrative tumor, recurring, in 90% of cases, within a few centimeters of the surgical resection cavity, even with adjuvant chemo/radiotherapy. Residual GB cells left in the margins or infiltrating the brain parenchyma shelter behind the extremely fragile and sensitive brain tissue and may favor recurrence. Tools for eliminating these cells without damaging the brain microenvironment are urgently required. We propose a strategy involving the implantation, into the tumor bed after resection, of a scaffold to concentrate and trap these cells, to facilitate their destruction by targeted therapies, such as stereotactic radiosurgery. We used bacterial cellulose (BC), an easily synthesized and modifiable random nanofibrous biomaterial, to make the trap. We showed that the structure of BC membranes was ideal for trapping tumor cells and that BC implants were biocompatible with brain parenchyma.
We also demonstrated the visibility of BC on magnetic resonance imaging, making it possible to follow its fate in clinical situations and to define the target volume for stereotactic radiosurgery more precisely.
Furthermore, BC membranes can be loaded with chemoattractants, which were released and attracted tumor cellsin vitro. This is of particular interest for trapping GB cells infiltrating tissues within a few centimeters of the resection cavity. Our data suggest that BC membranes could be a scaffold of choice for implantation after surgical resection to trap residual GB cells.
Statement of Significance
Glioblastoma is a highly infiltrative tumor, recurring, in 90% of cases, within a few centimeters of the sur- gical resection cavity, even with adjuvant chemo/radiotherapy. Residual tumor cells left in the margins or infiltrating the brain parenchyma shelter behind the extremely fragile and sensitive brain tissue and con- tribute to the risk of recurrence. Finding tools to eliminate these cells without damaging the brain microenvironment is a real challenge. We propose a strategy involving the implantation, into the walls of the surgical resection cavity, of a scaffold to concentrate and trap the residual tumor cells, to facilitate their destruction by targeted therapies, such as stereotactic radiosurgery.
Ó2018 Published by Elsevier Ltd on behalf of Acta Materialia Inc.
1. Introduction
Glioblastoma (GB) is the most aggressive and frequent primary tumor of the brain in adults. Its annual age-adjusted incidence is
about 0.59–3.69 cases/100,000 persons [1]. Surgical resection is the first step in GB treatment. Ideally, the surgical margins after tumor removal should contain no tumor cells on microscopic examination. Unfortunately, in most cases, residual tumor cells remain, some of which may have infiltrated the normal parench- yma. Furthermore, a macroscopic residual tumor may even remain after resection. A standard treatment including radiotherapy plus concomitant and adjuvant temozolomide is usually proposed after
https://doi.org/10.1016/j.actbio.2018.11.027
1742-7061/Ó2018 Published by Elsevier Ltd on behalf of Acta Materialia Inc.
⇑Corresponding author at: CHU, Département de Neurochirurgie, 4 rue Larrey, 49 933 Angers Cedex 9, Angers, France.
E-mail address:anne.clavreul@univ-angers.fr(A. Clavreul).
Contents lists available atScienceDirect
Acta Biomaterialia
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a c t a b i o m a t
surgery[2,3]. Despite this treatment, recurrences are observed in 90% of cases, after 6–7 months on average, close to the excision cavity[4–6].
Recurrence may be favored by the residual GB cells left in the surgical margins or infiltrating the normal brain parenchyma [7–9]. This infiltration may result from a centrifugal migration of tumor cells away from the tumor mass and towards the normal brain parenchyma and/or a centripetal migration of GB stem cells, located centimeters away from the tumor mass at the time of sur- gery, towards the resection site[7–11]. These centrifugal and cen- tripetal migrations may be favored by the brain tissue response to surgical injury and by GB-associated stromal cells (GASCs) present in the peritumoral brain zone, which produce extracellular matrix proteins, cytokines and chemokines [12–15]. Residual GB cells, which are sheltered by the extremely fragile and sensitive brain tissue, cannot be imaged and maximum dose of radiotherapy that can be administered given the tolerance of normal brain tissue is inadequate to destroy these cells without specific targeting, result- ing in a high probability of recurrence. Chemotherapy also fails to eliminate these cells completely, and none of the many attempts to improve drug delivery has yielded a major breakthrough. Finding tools to eliminate residual GB cells left in the surgical margins or infiltrating the brain parenchyma without damaging the brain microenvironment therefore remains a major challenge.
We developed a tool based on the breakthrough concept of can- cer cell traps, which emerged from the ecological trap strategy [16]. This strategy is designed to concentrate and localize tumor cells within a predefined area that can be targeted by treatment.
We aimed to achieve this goal by placing a scaffold for attracting and trapping the residual GB cells within the tumor bed after resection. Once the cells have been immobilized on the scaffold, they can be destroyed by a targeted therapy, such as stereotactic radiosurgery, which can deliver a higher dose of radiation than is possible with conventional radiotherapy to a specific location in one shot, thereby sparing normal adjacent brain tissue[17]. Mole- cules or particles with radio-sensitizing and/or chemotherapy characteristics could also be embedded within the trap to improve tumor cell destruction.
Hydrogels are hydrophilic polymer networks with tissue-like properties and applications in tissue engineering and drug delivery [18,19]. Bacterial cellulose (BC) is a widely studied hydrogel[20]
and was used as the scaffold for this trap. This substance closely resembles the absorbable hemostat of oxidized regenerated cellu- lose (SurgicelÒ) widely used by neurosurgeons. BC is an extracellu- lar polysaccharide synthesized in the air/culture medium interface by Gram-negative bacteria, such asKomagataeibacter hansenii. BC is produced in nanofibrils composed ofb-(1,4) glucose units stabi- lized by inter- and intra-chain hydrogen bonds. The nanofibrillar structure of BC has a high water content (more than 90%), with considerable mechanical strength and thermal stability, and well- defined biocompatibility [21,22]. BC is highly versatile and can be obtained in the form of nanofibrils, micro- and nanoparticles (by chemical or enzymatic modification), and as a biofilm matrix with various degrees of crystallinity, depending on the microbial strain and fermentation strategy used [22]. BC is very flexible and porous, like collagen, and can be purified with simple tech- niques to satisfy the <20 endotoxin units/device threshold approved by the FDA (Food and Drug Administration, USA) for implants. All these features make BC an attractive material for tis- sue engineering applications[20,22]. It can be loaded with anti- cancer drugs, anti-inflammatory drugs, antibiotics, proteins and DNA, and is thus suitable for use in a wide range of diseases, from superficial skin infections to cancers[20,23–26].
In this study, we evaluated the suitability of BC as the scaffold of choice for tumor cell traps in the F98 rat glioma model. This model closely resembles human GB in its biological behavior[27,28]. In
particular, it has an infiltrative pattern of growth within the brain and is refractory to various modes of treatment. Based on our knowledge of GASCs, we used the conditioned medium (CM) pro- duced by these cells (GASC-CM) to evaluate the capacity of BC to load and release chemoattractants for trapping tumor cells at a dis- tance. The visibility on magnetic resonance imaging (MRI) of BC and its tolerability were also analyzed.
2. Materials and methods
2.1. Cell culture
F98 tumor cells (LGC, Molsheim, France) were maintained in high-glucose Dulbecco’s modified Eagle’s medium (DMEM-HG) (Lonza, Ozyme, St Quentin en Yvelines, France) containing 10%
fetal calf serum (FCS) (Gibco, Thermo Fisher Scientific, Illkirch, France) and 1% antibiotics (Sigma, Saint Quentin Fallavier, France), in a humidified incubator, under an atmosphere containing 5% car- bon dioxide (37°C), until 80–90% confluence was reached.
2.2. BC membranes
BC is a pureb-glucan biopolymer composed only ofb-(1,4) glu- cose units. The production and biophysical characterization of BC membranes by microscopy, gravimetry and spectroscopic tech- niques have been described elsewhere[21,22,24,25,29,30]. Briefly, BC was synthesized by Komagataeibacter hansenii (ATCC 23769, Manassas, Virginia, USA) in a medium containing 25.0 g/L manni- tol, 5.0 g/L yeast extract, and 3.0 g/L peptone, the pH of which was adjusted to 6.5 with 0.1 M NaOH before sterilization. The cul- ture was maintained in 48-well plates at 30°C for 14 days, without shaking. BC membranes were collected from the plates and washed with distilled water. BC was purified by incubating the membranes in 100 mM NaOH at 50°C for 24 h, then washing them successively several times in distilled water, and adjusting the pH to 7.0. Discs of BC membrane 1.0 or 1.5 cm in diameter and 1.0 mm in thick were cut and sterilized by autoclaving (121°C for 20 min).
2.3. Chemoattractant loading and release characteristics of BC membranes
Human serum albumin (HSA) (Vialebex, LFB, Paris, France) was selected as a model protein for investigating the loading of chemoattractants onto and their release from BC membranes. For loading, three BC discs of 1.0 cm in diameter were immersed in 1.5 mL of DPBS (pH 7.4, Lonza) supplemented with 10 mg/mL HSA for 24 h at room temperature on a KS125 orbital shaker (IKA Labortechnik, Staufen, Germany, 300 rpm). BC discs incubated in DPBS without HSA, and HSA solution without BC discs were used as controls. HSA content was quantified with a PierceTMBCA pro- tein assay kit (Thermo Fisher Scientific). The difference between the amount of HSA in the control solution without BC discs and that remaining after the loading of BC discs was used to calculate the uptake capacity of BC membranes (loaded HSA as a percentage of the amount present in the control HSA solution without BC discs, which was considered to correspond to 100%). For visualization of the distribution of HSA in BC relative to controls, loaded samples were incubated in 1 mL BCA assay reagent at 37°C for 30 min and photographed. For release experiments, BC discs were trans- ferred to 1.5 mL DPBS and incubated at 37°C. Samples (100mL) were collected at various time points (1, 2, 4, 8, 24, 30 and 48 h), and 100mL DPBS was added to each sample. The HSA released into the supernatant was measured in a BCA protein assay. The amount of protein released is expressed as a percentage of the initial amount of protein loaded.
2.4. Transwell migration assays
Transwell inserts (MillicellÒcell culture inserts, 8
l
m pore size, Merck Millipore, Molsheim, France) were used to assess the migra- tion of F98 cells in response to factors secreted by GASCs, either free or loaded onto BC membranes. GASC-CM was prepared as pre- viously described[13]. The control medium consisted of DMEM supplemented with 0.1% HSA and 0.2% antibiotics. BC membranes were loaded with control medium or GASC-CM by incubation with gentle stirring for 24 h, as described above for HSA. F98 cells (3104cells) were added to the upper chamber, and control med- ium, GASC-CM, control medium-loaded BC membrane or GASC- CM-loaded BC membrane was added to the bottom chamber. After 24 h, the cells that had migrated to the underside of the insert membrane were stained with hematoxylin and counted with a computerized image-analysis system (MetaView, Roper Scientific, Evry, France). Five fields at 200 magnification were selected at random for each insert for the determination of migrating cell counts.2.5. Tumor cell escape assays
Two tests were developed to analyze F98 escape from BC mem- branes: 1) F98 cells (1104cells/cm2) were cultured on BC discs of 4 mm in diameter placed on MillicellÒinserts (0.4
l
m pore size, Merck Millipore) in a six-well plate filled with DMEM-HG supple- mented with 10% FCS and 1% antibiotics. After 24 h, each BC disc was transferred to 2 mL of culture medium in one of the wells of a six-well plate. We assessed the ability of the cells to escape from the BC membrane and reach the plastic surface, by photographing fields at the edge of the BC membrane at times of 0, 24, 48 and 72 h. 2) BC discs were placed in the upper chamber of a Transwell migration system and F98 cells (3104 cells) were carefully placed on the surface of the BC. We then added 500l
L GASC-CMto the bottom chamber. We determined whether the cells could pass through the BC membrane, by staining the underside of the insert membrane after 24 h, as described above.
2.6. Scanning electron microscopy
BC membranes with or without F98 cells were fixed in glu- taraldehyde, desiccated in hexamethyldisiloxane and coated with a thin layer of platinum. Images were acquired with a JEOL JSM- 6301 F microscope. (SCIAM platform, Angers, France).
2.7. Organotypic brain-slice invasion assays
Brain slice cultures were prepared by the interface method, as described by Stoppini et al. (1991) [31]. Nine-day-old albino wild-type Sprague-Dawley rats from the SCAHU (Angers, France) were used and the animal procedures were approved by the animal experimentation ethics committee of Pays de la Loire (authoriza- tion no. A49-2012-16). Brains were rapidly removed. The frontal pole of the hemispheres and the cerebellum were dissected out, and the brains were then cut into 400mm-thick horizontal slices with a vibratome (Motorized Advance Vibroslice MA752, Campden Instruments, Loughborough, UK) in Gey’s balanced salt solution (Sigma) supplemented with 6.5 mg/L glucose and 1% antibiotics.
The slices were placed on MillicellÒ inserts (0.4
l
m pore size, Merck Millipore), which were then transferred to the wells of six-well plates filled with 1 mL pre-warmed culture medium con- sisting of 50% MEM, 25% horse serum (Sigma), 25% Hanks’ balanced salt solution (Sigma), supplemented with 6.5 mg/L D-glucose, 1%antibiotics (Sigma) and 1 mM L-glutamine (Lonza). One day after slice preparation, the medium was aspirated and 0.8 mL pre- warmed culture medium consisting of Neurobasal medium (Gibco,
Thermo Fisher Scientific) supplemented with 6.5 mg/L D-glucose, 1 mM L-glutamine, 1x B27 supplements (Gibco, Thermo Fisher Sci- entific) and 1% antibiotics (Sigma), was added. On the same day, adherent F98 cells were stained by incubating for 45 min at 37°C in serum-free DMEM containing 5
l
M CellTracker green CMFDA (5-chloromethylfluorescein diacetate) fluorescent dye (Invitrogen, Thermo Fisher Scientific). The dye solution was then replaced with DMEM-HG supplemented with 10% FCS and 1% antibiotics and the cells were incubated at 37°C for a further hour. The cells were then washed twice with DPBS and 5103cells in a total volume of 1mL of HBSS with Ca2 +and Mg2+(Lonza) were then implanted into or close to the corpus callosum. Two days later, the medium was changed and 18 brain slices containing identified tumors were selected and assigned to three groups: 1) brain slices with control tumors (n= 6), 2) brain slices with tumors close to control medium-loaded BC membranes (n= 6), 3) brain slices with tumors close to GASC-CM-loaded BC membranes (n= 6). Tumor size dif- fered between the brain slices, and was therefore taken into account during the assignment of the slices to the groups, to ensure a similar tumor size distribution in each of the groups. After four days, brain slices were fixed in 4% PFA pH 7.4 for 3 h. The attraction of tumor cells towards BC membranes was followed with a Leica TCS SP8 laser-scanning confocal microscope (Leica Microsystems, Heidelberg, Germany) with an HC PL FLUOTAR 10(N.A. 0.3) objective and 0.75magnification. The excitation wavelength was set at 488 nm, and fluorescence was measured between 491 and 594 nm, with a hybrid detector. Multiple stacked images (10241024 pixels) were taken at 400 Hz, with pixel sizes of 1.521.52mm and 4mm between slices. Tumor surface and vol- ume were measured on reconstructed image stacks, with ImageJ 1.50a (National Institutes of Health, Bethesda, USA) and Imaris 6.4 (Bitplane AG, Zurich, Switzerland), respectively.2.8. Implantation of BC membranes in rat brain parenchyma Female syngeneic Fisher 344 rats aged 9 to 10 weeks were obtained from Charles River Laboratories (L’Arbresle, France). The protocol was approved by the French Ministry of Higher Education and Research (authorization no. APAFIS#9080). Animals were anesthetized by an intraperitoneal injection of xylazine (8 mg/kg body weight) and ketamine (80 mg/kg body weight) and posi- tioned in a Kopf stereotaxic instrument. On day 0 (D0), BC mem- branes of 3 mm in diameter were implanted with fine forceps (Dumont no. 5) in the cerebral cortex close to the corpus callosum [coordinates: 2.5 mm lateral to bregma, 1 mm anterior and 3 mm interior to the outer edge of the cranium]. In control rats, the same surgical gesture was performed with the forceps, but no membrane was implanted. Two groups of rats were then set up: one group for MRI analysis and one group for histological analysis.
2.9. MRI analyses 2.9.1. In vitro MRI
BC membrane discs were embedded in agarose for MRI analysis.
We used agarose gel because of its rheological properties, similar to those of the brain. Imaging data were acquired in a 7 T MRI sys- tem (Bruker Biospec Avance III 70/20).In vitro, T2, T1 and diffusion maps were acquired with an axial acquisition of 9 slices, and a slice thickness/gap of 0.7/0.3 mm. For the T2 map, a multispin echo sequence was used with field of view (FOV): 2626 mm, image matrix 256256, repetition time (TR) = 2 s, 25 echo times (TE) = 8, 16, 24. . .200 ms, one average. For the T1 map, a fast spin echo sequence was used (RARE) with an echo train length of 2, FOV = 2626 mm, image matrix 256192, TE = 7 ms, 6 TR = 200, 400, 800, 1500, 3000 and 5500 ms.
For the trace diffusion acquisition to measure the apparent dif- fusion coefficient (ADC), a segmented (4 segments) echo planar sequence was used with FOV = 2618 mm, a matrix of 10890, TE = 17 ms, TR = 3 s. The diffusion weighting gradients were applied in three perpendicular directions with the following parameters: duration d= 2.5 ms, separation D= 8.4 ms corre- sponding to a bmaxvalue of 1431 s/mm2completed by one acquisi- tion without the diffusion weighting gradients.
2.9.2. In vivo MRI
MRI analysis was performed on six rats (four implanted with BC and two controls) at D1, D21, M3, M6 and M12 after BC implanta- tion, with a Bruker Biospec 70/20 system operating at 7 T, under isoflurane (0.5% 1 L/min O2) anesthesia, with the monitoring of res- piratory parameters. Several scans were acquired: T1, T2 and diffusion-weighted imaging (DWI). T1-weighted images were obtained with a FLASH sequence [TR = 200 ms, TE = 3.5 ms, flip angle = 60°, 2 averages, 9 axial 0.8 mm slices (gap 0.3 mm), FOV = 3535 mm, matrix 384384]. T2-weighted images were acquired with a multispin echo sequence [FOV = 3535 mm, 7 axial 0.8 mm slices (gap = 0.3 mm), matrix 256256, TR = 2 s, 25 TE = 8, 16, 24. . .200 ms, one average]. T2 maps were calculated from these datasets by fitting the signal intensity on a pixel-by- pixel basis with a single-exponential curve. Regions of interest (ROIs) were drawn manually on T2 maps for quantification of the volume of BC membrane and edema over time. T2 and ADC values were also obtained. ADC values were obtained from a diffusion acquisition [7 slices of 0.8 mm (gap 0.3 mm), FOV = 3535 mm, matrix 128128, TE = 56 ms, TR = 3 s, 2 segments echo-planar imaging (EPI), one average]. The diffusion weighting gradients were applied in 126 directions with the following parameters:
d= 2.5 ms, D= 7.5 ms corresponding to a bmax value of 1262 s/mm2and 12 acquisitions were obtained without the diffu- sion weighting gradients. Diffusivity and fractional anisotropy (FA) and tractography maps were processed from DWI acquisitions [32,33].
2.10. Histological analysis
Histological analysis was performed on rats one day, one month, three months and one year after BC implantation. At each time point, seven animals were killed (four implanted with BC and three controls). Brains were fixed in formalin, embedded in paraffin and sections were cut and stained with hematoxylin- phloxin-saffron.
2.11. Statistical analysis
Results are expressed as means ± SEM. Statistical analyses were performed with the Mann-WhitneyUtest. Differences were con- sidered significant if thep-value was < 0.05.
3. Results
3.1. Analysis of BC membranes 3.1.1. Morphology of BC membranes
The BC membranes produced byKomagataeibacter hanseniicon- sist of a random network of nanofibrils that form an elastic and resistant hydrogel when exposed to water (Fig. 1). BC membranes have a denser flatter surface on one side and a gelatinous layer enabling cells to anchor themselves in its structure.
3.1.2. Chemoattractant loading onto and release from BC membranes HSA was selected as a model protein for the investigation of chemoattractant loading onto and release from BC membranes.
Three BC discs of 1.0 cm in diameter were incubated with 15 mg HSA with gentle stirring for 24 h. An uptake capacity of 11.9 ± 0.5% was achieved. BCA staining of BC membranes resulted in a uniform purple coloration on both surfaces and within the membrane, indicating that the distribution of HSA was homoge- neous (Fig. 2A). A time-dependent biphasic cumulative release pro- file was obtained, with an initial burst (90.7 ± 3.6%) within the first
Fig. 1.Morphology of BC membranes. (A): Photograph of a BC membrane. (B): SEM images of BC membranes. The bar indicates, from left to right, 10mm and 1mm. The random network of nanofibrils can be seen.
4 h, and equilibrium conditions, with a total release of 98.2 ± 4.2%, achieved after 24 h (Fig. 2B). BC membranes were loaded with con- trol medium or GASC-CM in the conditions used for HSA.In vitro, GASC-CM-loaded BC membranes induced the migration of larger numbers of F98 cells than control medium-loaded BC membranes (Fig. 2C, D, E). Less migration was observed with GASC-CM-loaded BC membranes than with GASC-CM, due to the limited uptake capacity of BC membranes.
3.1.3. MRI parameters of BC membranes
We defined the MRI parameters, by including BC membrane in agarose and measuring the T2, T1 and diffusion coefficient values (Fig. 3A, B). The T1 and diffusion characteristics of agarose gel were close to those of the BC, whereas the T2 value of the BC was higher than that of the agarose gel. BC is, therefore, visible on MRI.
3.2. Analysis of the trapping capacity of BC membranes
We assessed the trapping capacity of BC membranes, by analyz- ing the escape of F98 cells from these membranes by two methods.
First, after the adhesion of F98 cells on the BC membrane, we
placed the membrane on a plastic plate and checked at various time points whether the F98 cells had moved (Fig. 4). The F98 cells grew in the form of spheroids on BC membranes and remained in place (Fig. 4B). The SEM images presented inFig. 4C show a net- work of F98 spheroids with lamellipodia contacting the BC mem- brane, demonstrating cell adhesion. The second method used was Transwell migration assays. F98 cells were able to migrate towards GASC-CM. We therefore investigated whether this migration also took place when the insert contained a BC membrane. In the absence of the BC membrane, F98 cells migrated homogeneously to the underside of the insert (Fig. 4E). When the BC membrane was present inside the insert, F98 cells were located exclusively around the edge of the insert suggesting that they passed between the edge of the insert and the BC membrane but not through the membrane (Fig. 4E).
3.3. Analysis of tumor cell behavior at a distance from BC membranes, unloaded or loaded with GASC-CM
Organotypic brain slice assays were used to analyze the behav- ior of F98 tumors at a distance from BC membranes. F98 cells Fig. 2.Chemoattractant loading onto and release from BC membranes. (A) Visualization of the loading of the model protein (HSA) and its distribution on BC membranes by BCA staining. Plan view of a control BC membrane incubated in DPBS without HSA, followed by three HSA-loaded BC membranes. (B) Cumulative HSA release of BC membranes as a function of time. (C) Schematic diagram of the principle of the Transwell migration assay, with movement towards GASC-CM-loaded BC membranes (GASC- BC). The pore size of the insert (8mm) allowed the passage of F98 cells. (D) Number of F98 cells migrating towards GASC-CM either free in solution or loaded onto BC membranes, after 24 h of incubation. The control medium, loaded onto BC membranes (Control-BC) or alone was used as a control. Data are expressed as the mean ± SEM (n= 2).*P< 0.05 versus the corresponding controls. (E) Representative fields of F98 cells responding to chemoattractive medium, after 24 h of migration. The bar indicates 100mm.
labeled with CellTracker green CMFDA were implanted in rat brain slice cultures, within or close to the corpus callosum (Fig. 5A). Two days later, when the tumor was visible, control medium-loaded BC membranes or GASC-CM-loaded BC membranes were added some distance away from the tumor. Tumor volume and surface area were monitored and quantified on bright-field (data not shown) and fluorescence images four days after membrane addition (Fig. 5B, C, D). Tumor area did not differ significantly between the tumor alone and the tumor close to control medium-loaded BC membranes or GASC-CM-loaded BC membranes (tumor with- out membrane: mean = 0.93 ± 0.11 mm2, tumor close to control medium-loaded BC membrane: mean = 0.96 ± 0.10 mm2, tumor close to GASC-CM-loaded BC membrane: mean = 1.04 ± 0.16 mm2) (Fig. 5B, C). Tumor volume was also similar in each of the tumor groups (tumor without membrane: mean = 0.07 ± 0.01 mm3, tumor close to control medium-loaded BC membrane:
mean = 0.08 ± 0.01 mm3, tumor close to GASC-CM-loaded BC membrane: mean = 0.07 ± 0.01 mm3) (Fig. 5B, D). We observed no migration of F98 cells towards GASC-CM-loaded BC membranes, suggesting that the trapping of tumor cells requires the cells to be close to the membrane.
3.4. Analysis of the fate and biocompatibility of the BC membrane in brain parenchyma
3.4.1. MRI analysis of changes in the BC membrane in brain parenchyma
The BC membrane was implanted in the cerebral cortex of rats and changes in the membrane were followed by MRI at D1, D21, M3, M6 and M12 after implantation (Fig. 6). As expected from thein vitroresults, one day after implantation, the BC membrane was not visible on T1-weighted MRI (data not shown) but clearly visible on the T2 map, due to the BC T2 value being higher than the classical T2 value obtained for the brain (Fig. 6A). However, contrary to thein vitroresults, the BC membrane was clearly visi- ble on diffusivity and FA maps, as a hypersignal on the diffusivity map because it had a higher diffusivity than brain tissue, and as a hyposignal on the FA map due to its isotropic structure, whereas brain tissue is anisotropic (Fig. 6A). Tractography showed a disrup- tion in the map of neuronal pathways caused by BC insertion (Fig. 6A). Mean values of T2 and ADC for BC membranes obtained one day after BC membrane implantation are presented in Fig. 6B. The BC membrane remained clearly visible over time on T2 and diffusivity maps. Its volume decreased slightly until M3
after implantation, remaining stable thereafter until at least M12 (Fig. 6C). One day after BC implantation, a diffuse T2 hypersignal corresponding to an edema was observed around the membrane.
This signal had decreased by D21 and was absent at M3 (data not shown).
3.4.2. Histological analysis of the biocompatibility of the BC membrane in brain parenchyma
One day after BC membrane implantation in brain parenchyma, we observed the presence of an edema and an infiltration of neu- trophils around the BC, indicating an acute inflammatory response (Fig. 7). The same reaction was observed in animals in which the surgical gesture was performed without membrane implantation (data not shown). One month after BC implantation, we observed a mild chronic inflammatory response characterized by the pres- ence of altered neutrophils, multinucleated giant cells, fibroblasts with collagen fibers, and neovascularization, forming a crown around the BC membrane (Fig. 7). This chronic inflammatory response was not observed in control rats (data not shown). Three months after BC implantation, scar remodeling was observed, with a thin fibrous capsule around the membrane, a small amount of lymphocytic infiltrate, some multinucleated giant cells and cal- cium deposition (Fig. 7). A similar reaction was present 12 months after BC membrane implantation (Fig. 7). With the exception of the inflammatory response, the brain microenvironment was not affected after BC membrane implantation in the brain parenchyma.
The deposition of a number of brain cells on the BC membrane was noted 12 months after BC implantation demonstrating the non- cytotoxicity of BC to normal cells (Fig. 7).
3.4.3. Clinical tolerance of BC film
All animals were monitored daily and weighed once weekly. All animals survived until the date on which they were scheduled to be killed. Animals displayed no changes in behavior or appearance and grew normally (data not shown).
4. Discussion
Despite maximal initial resection and standard treatment, GB recurs within 2 cm of the original tumor site in 90% of cases [4–6]. Residual GB cells left in the surgical margins or infiltrating normal brain parenchyma may contribute to the risk of recurrence.
Many current therapeutic approaches based on the ‘‘search and destroy” strategy have been less successful than hoped. We devel- Fig. 3.MRI parameters of BC membrane. (A) T1, T2 and diffusivity maps for BC membrane (the border of the BC membrane is indicated in orange). (B) T1, T2 and ADC mean values for the BC membrane relative to agarose gel.
oped a tool based on the ‘‘attract to kill” strategy described by van der Sanden et al. (2013)[16], involving the implantation, into the tumor bed after resection, of a scaffold to concentrate and trap residual tumor cells so as to improve their elimination by targeted treatments, such as stereotactic radiosurgery.
We used a BC scaffold to make the trap, because cellulose-based materials are already used in clinical practice, as exemplified by SurgicelÒ, which is widely applied in neurosurgery due to its
hemostatic effects and good tissue compatibility. Furthermore, this biomaterial has many useful properties, including high mechanical strength and thermal stability, high flexibility and an ultra-fine network structure of very pure nanofibrils[22]. The considerable flexibility of BC makes it easy to introduce into the surgical resec- tion cavity of GB, and the gelatinous layer on one side of the mem- brane enables it to mold itself to the walls and floor of the cavity.
The thermal stability of BC makes it possible to sterilize this mate- Fig. 4.Analysis of the trapping capacity of BC membranes. (A) Schematic representation of the first method used to analyze the escape of F98 cells from the BC membrane. (B) Phase-contrast microscopy of one field of F98 cells on the edge of the BC membrane taken at different times (T0, T24, T48 and T72 h). The white arrow shows the formation of a spheroid, which remained in place over time. The bar indicates 100mm. (C) SEM images of F98 cells cultured on BC membranes. The bar indicates, from left to right, 100mm, 10mm and 1mm. (D) Schematic representation of the second method used to analyze the escape of F98 cells from the BC membrane. (E) Representative fields of F98 cells after 24 h of migration in response to GASC-CM, with or without a BC membrane in the insert. The bar indicates 100mm.
rial by autoclaving, an advantage over other hydrogels, which must be sterilized by gamma irradiation.
We found that the structure of BC membranes, with their ran- dom assembly of nanofibrils, was ideal for the trapping of tumor cells. Indeed, unlike aligned nanofiber scaffolds that mimic the structure of the white matter tracts and blood vessels, thereby favoring GB cell invasion[34–37], the BC membrane trapped F98 tumor cells, which, once attached to the surface of the membrane, were unable to move on, go through the membrane or escape, even if an attractive medium was present in nearby. Surface cell adhe-
sion to BC membranes has also been reported in other studies with human smooth muscle cells and fibroblasts[30,38]. The trapping capacity of BC membrane renders this membrane a potentially interesting alternative to the use of integrin inhibitors for decreas- ing GB cell invasion [39,40]. F98 cells trapped on BC remained viable and retained the ability to grow, adopting a spheroid pattern of growth.
Our results demonstrate the biocompatibility of the BC trap with brain parenchyma. BC is generally considered to be biocom- patible, although, to our knowledge the fate of BC implants in brain Fig. 5.Analysis of the migration of F98 cells towards BC membranes loaded with GASC factors on organotypic brain slices. (A) Schematic diagram of the principle of the assay of migration towards control medium-loaded BC membranes (control-BC) and GASC-CM-loaded BC membranes (GASC-BC). (B) Migration pattern of F98 cells four days after BC membrane deposition. F98 cells, which were labeled with the CellTracker green CMFDA fluorescent dye, were detected under a laser scanning confocal microscope. The yellow dotted line delimits the outer contour of the membrane. The bar indicates 300mm. (C) Graph showing the area of F98 tumors four days after BC membrane deposition.
(D) Graph showing the volume of F98 tumors four days after BC membrane deposition. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
parenchyma has never been reported. Brain BC implants caused a mild benign inflammatory reaction that decreased over time with- out eliciting a foreign body reaction. A thin fibrous capsule also formed. This type of reaction, which has also been described for subcutaneous BC implants[41], cannot be due to residual bacterial biomolecules because BC is a pureb-glucan biopolymer composed only ofb-(1,4) glucose units and this biomaterial was sterilized by autoclaving before use. Biomaterials, like any foreign body, are known to trigger inflammatory responses on implantation. How- ever, little is known about the detection of such biomaterials by the body and how this detection might trigger an inflammatory response[42]. Phagocytes are thought to interact with the sponta- neously adsorbed fibrinogen rather than with the material itself [43]. The inflammatory response induced by BC is unlikely to cause additional complications, because GB surgery is inherently lesional, regardless of the care taken to spare brain function[44].
Furthermore, based on the post-surgery reverse migration field theory[10], this response could promote the migration of GB stem cells towards the BC trap. The formation of the fibrous capsule around the BC at a late time point, about three months after implantation, leaves a sufficiently large time window for tumor cell trapping.
We also showed that the BC trap was visible on MRI, a non- invasive imaging method. BC was visible on T2 and diffusion images, unsurprisingly given the high water content of the BC and the high visibility of water signals on T2 and diffusion sequences. The trapped water molecules reside within the pores of the BC matrix, where hydrogen bonding interactions take place [45]. MRI made it possible to monitor the fate of BC implants over time in the brain parenchyma in rats. In particular, the volume of the BC implants remained stable. The BC implants were implanted fully wet and the number of pores in the BC matrix was limited, thus reducing the water-holding capacity of the BC implants and preventing them from increasing in volume [45]. A dynamic exchange of water molecules between the inside and outside of the polymer network probably kept the volume of the BC implants stable. The BC scaffolds were still visible on MRI one year after their implantation, and displayed no major signs of degradation.
This result was not surprising, because thein vivodegradability of cellulose-based materials is thought to be limited, given the absence of hydrolases capable of attacking theb-(1,4) linkage in animal and human tissues[46,47]. In the cancer cell trap system, the visibility of BC on MRI will make it possible to delimit more precisely the target volume for stereotactic radiosurgery.
Fig. 6.MRI analysis of the change in the BC membrane in rodent brain parenchyma. (A) T2, diffusivity, FA and tractography maps of the BC membrane one day after its implantation (the border of the BC membrane is indicated in orange). (B) Comparison of T2 and ADC mean values of the BC membrane with those of the brain and edema one day after BC membrane implantation. (C) Graph showing changes in BC membrane volume over time. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
We also investigated the ability of BC membranes to trap tumor cells at a distance through Transwell migration and organotypic brain slice invasion assays. The brain slice model system, which has the advantage of maintaining the cytoarchitecture and microenvironment of the brain tissue, constitutes a unique system for evaluating the invasive properties of GB cells in more physio- logically relevant conditions[48]. These assays showed that BC membranes alone could not attract F98 cells. Chemokines must, therefore, be introduced into BC membranes, to induce chemical guidance for tumor cells. BC membranes can be loaded with pro- teins during the formation of the BC material (in situloading) or after its preparation (postloading), usually by sorption techniques in submerged conditions with gently shaking for 24–48 h[25,49].
In this study, we used the postloading technique, because this technique provides mild conditions, and prevents deleterious effects on proteins during BC material formation and purification.
It has been suggested that the incorporation of proteins into BC
is related to diffusion and capillary forces [49]. Using HSA as a model protein, we measured an uptake capacity of 11.9 ± 0.5%
within 24 h of the loading of BC membranes with gentle stirring, a result similar to that reported by Muller et al. (2013)[49]. Release experiments demonstrated an initial burst of release within the first 4 h, followed by a slow-release phase until 24 h. Most studies using postloading techniques to incorporate active components into BC have reported similar profiles, with differences only in the time frames of the two phases[26]. There is growing evidence to suggest that brain extracellular components and their partners and modulators play a crucial role in GB cell invasion in the brain [15,50]. In previous studies, we identified and characterized stro- mal cells called GASCs in the peritumoral brain zone of GB[12–
14]. GASC-CM, which contains chemoattractants, such as THBS1, CXCL12, HGF and fibronectin, increased human A172 GB cell inva- sion[13]. In this study, we observed that GASC-CM induced the migration of F98 cells with a chemoattractive potency greater than Fig. 7.Histological section of rat brain one day (D1), one month (M1), three months (M3) and twelve months (M12) after BC membrane implementation. The bar indicates 200mm.
that observed with purified CXCL12 or THBS1 (data not shown).
Based on these results, we used GASC-CM as a source of chemoat- tractants for loading BC membranes. We showed that GASC-CM- loaded BC membranes attracted F98 cells in Transwell migration assays but not in organotypic brain slice invasion assays. There may be several reasons for this. For example, organotypic brain slices being inherently lesional (like resection margins), the growth factors/chemokines produced locally by the organotypic slice might overcome any directional chemoattractive gradient emanat- ing from the BC matrix. Furthermore, insufficient amounts of the factors loaded onto BC membranes and/or a rapid release of GASC factors into the incubating medium could prevent establishment of the chemoattractant gradient required for cell migration. These issues have already been highlighted for other therapeutic agents incorporated into BC[26]. Various strategies have been explored for improving the incorporation of drugs into BC and controlling their release. These strategies include chemical modifications of BC, covalent binding of the drug to BC fibers or the incorporation of an additional diffusional barrier or release system[22,26]. For example, the trapping of nanoparticles, nanosheets, self- assembled particles or micelles, cyclodextrins or gelled structures in BC is considered to slow drug release[26]. One of these methods could be used to improve the loading and release of chemoattrac- tants by BC. The chosen method of BC modification will also need to ensure the diffusion of the chemoattractants released over sig- nificant distances within the peritumoral zone of the brain. Studies of drug delivery by diffusion from polymer implants have revealed that drugs often diffuse over very short distances, typically 1–5 mm. The penetration of chemoattractants will depend on their metabolism within the brain tissue and their elimination from the interstitial spaceviapermeation into capillaries[51]. In addition, tissue composition around the implant may interfere with the pen- etration of chemoattractants. In the case of the BC membrane, the transient edema, resulting from its implantation may enhance the penetration of chemoattractantsviathe convection of interstitial fluid, and the late formation of a thin fibrous capsule around the BC may not impede this process.
5. Conclusion
These data indicate that BC membrane is a biocompatible scaf- fold that can trap tumor cells. Its high flexibility makes it easy to introduce into the tumor bed after resection and its visibility on MRI may facilitate stereotactic radiosurgery. Furthermore, we were able to load this scaffold with chemoattractants, which attracted tumor cells on their releasein vitro. Advanced methods should be developed to transform this trap into a chemotaxis device for the diffusion of chemoattractants over large distances and at high enough concentrations to establish a concentration gradient extending into the surrounding environment, for the trapping of GB cells infiltrating tissues several centimeters away from the resection cavity.
Acknowledgements
We thank the SCAHU (Service Commune d’Animalerie Hospital- Universitaire), particularly Pierre Legras and Jérôme Roux, for ani- mal care, and the SCIAM (Service Commun d’Imageries et d’Analyses Microscopiques) of Angers for SEM analysis. We also thank Jean-Luc Grandpierre (Laboratoire Pathologie Cellulaire et Tissulaire, CHU, Angers) for technical support, Prof. Franck Boury and Dr. Emma- nuel Garcion (Team 17, GLIAD, CRCINA, INSERM 1232, Angers) for allowing us to use their facilities and Alex Edelman and Associ- ates for correcting the manuscript.
Disclosures
The authors have no conflict of interest to declare.
Funding
This research was supported by French (Association en Avant la Vieand CHU Angers), and Argentinian [the National Council for Science and Technology (CONICET, PIP 0498) and the National Agency of Scientific and Technological Promotion (ANPCyT, PICT 2016-4597) to GRC] grants.
References
[1]Q.T. Ostrom, L. Bauchet, F.G. Davis, I. Deltour, J.L. Fisher, C.E. Langer, M.
Pekmezci, J.A. Schwartzbaum, M.C. Turner, K.M. Walsh, M.R. Wrensch, J.S.
Barnholtz-Sloan, The epidemiology of glioma in adults: a ‘‘state of the science”
review, Neuro-Oncol. 16 (2014) 896–913.
[2]R. Stupp, W.P. Mason, M.J. van den Bent, M. Weller, B. Fisher, M.J.B. Taphoorn, K. Belanger, A.A. Brandes, C. Marosi, U. Bogdahn, J. Curschmann, R.C. Janzer, S.K.
Ludwin, T. Gorlia, A. Allgeier, D. Lacombe, J.G. Cairncross, E. Eisenhauer, R.O.
Mirimanoff, European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups, National Cancer Institute of Canada Clinical Trials Group, Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma, N. Engl. J. Med. 352 (2005) 987–996.
[3]R. Stupp, M.E. Hegi, W.P. Mason, M.J. van den Bent, M.J.B. Taphoorn, R.C. Janzer, S.K. Ludwin, A. Allgeier, B. Fisher, K. Belanger, P. Hau, A.A. Brandes, J.
Gijtenbeek, C. Marosi, C.J. Vecht, K. Mokhtari, P. Wesseling, S. Villa, E.
Eisenhauer, T. Gorlia, M. Weller, D. Lacombe, J.G. Cairncross, R.-O.
Mirimanoff, European Organisation for Research and Treatment of Cancer Brain Tumour and Radiation Oncology Groups, National Cancer Institute of Canada Clinical Trials Group, Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial, Lancet Oncol. 10 (2009) 459–466.
[4]J.-M. Lemée, A. Clavreul, P. Menei, Intratumoral heterogeneity in glioblastoma:
don’t forget the peritumoral brain zone, Neuro-Oncol. 17 (2015) 1322–1332.
[5]M. Niyazi, M. Brada, A.J. Chalmers, S.E. Combs, S.C. Erridge, A. Fiorentino, A.L.
Grosu, F.J. Lagerwaard, G. Minniti, R.-O. Mirimanoff, U. Ricardi, S.C. Short, D.C.
Weber, C. Belka, ESTRO-ACROP guideline ‘‘target delineation of glioblastomas”, Radiother. Oncol. 118 (2016) 35–42.
[6]K. Petrecca, M.-C. Guiot, V. Panet-Raymond, L. Souhami, Failure pattern following complete resection plus radiotherapy and temozolomide is at the resection margin in patients with glioblastoma, J. Neurooncol. 111 (2013) 19–23.
[7]C. Angelucci, A. D’Alessio, G. Lama, E. Binda, A. Mangiola, A.L. Vescovi, G.
Proietti, L. Masuelli, R. Bei, B. Fazi, S.A. Ciafrè, G. Sica, Cancer stem cells from peritumoral tissue of glioblastoma multiforme: the possible missing link between tumor development and progression, Oncotarget 9 (2018) 28116–
28130.
[8]M. Glas, B.H. Rath, M. Simon, R. Reinartz, A. Schramme, D. Trageser, R.
Eisenreich, A. Leinhaas, M. Keller, H.-U. Schildhaus, S. Garbe, B. Steinfarz, T.
Pietsch, D.A. Steindler, J. Schramm, U. Herrlinger, O. Brüstle, B. Scheffler, Residual tumor cells are unique cellular targets in glioblastoma, Ann. Neurol.
68 (2010) 264–269.
[9]S.G.M. Piccirillo, S. Dietz, B. Madhu, J. Griffiths, S.J. Price, V.P. Collins, C. Watts, Fluorescence-guided surgical sampling of glioblastoma identifies phenotypically distinct tumour-initiating cell populations in the tumour mass and margin, Br. J. Cancer. 107 (2012) 462–468.
[10]W. Jabbour, D. Wion, Extent of Resection and Survival in Glioblastoma Multiforme, JAMA Oncol. 2 (2016) 1509.
[11]J.-M. Lemée, A. Clavreul, M. Aubry, E. Com, M. de Tayrac, P.-A. Eliat, C. Henry, A.
Rousseau, J. Mosser, P. Menei, Characterizing the peritumoral brain zone in glioblastoma: a multidisciplinary analysis, J. Neurooncol. 122 (2015) 53–61.
[12]A. Clavreul, A. Etcheverry, A. Chassevent, V. Quillien, T. Avril, M.-L. Jourdan, S.
Michalak, P. François, J.-L. Carré, J. Mosser, Grand Ouest Glioma Project Network, P. Menei, Isolation of a new cell population in the glioblastoma microenvironment, J. Neurooncol. 106 (2012) 493–504.
[13]A. Clavreul, C. Guette, R. Faguer, C. Tétaud, A. Boissard, L. Lemaire, A. Rousseau, T. Avril, C. Henry, O. Coqueret, P. Menei, Glioblastoma-associated stromal cells (GASCs) from histologically normal surgical margins have a myofibroblast phenotype and angiogenic properties, J. Pathol. 233 (2014) 74–88.
[14]A. Clavreul, A. Etcheverry, C. Tétaud, A. Rousseau, T. Avril, C. Henry, J. Mosser, P.
Menei, Identification of two glioblastoma-associated stromal cell subtypes with different carcinogenic properties in histologically normal surgical margins, J. Neurooncol. 122 (2015) 1–10.
[15]L. Hamard, D. Ratel, L. Selek, F. Berger, B. van der Sanden, D. Wion, The brain tissue response to surgical injury and its possible contribution to glioma recurrence, J. Neurooncol. 128 (2016) 1–8.
[16]B. van der Sanden, F. Appaix, F. Berger, L. Selek, J.-P. Issartel, D. Wion, Translation of the ecological trap concept to glioma therapy: the cancer cell trap concept, Future Oncol. 9 (2013) 817–824.
[17]J.L. Shah, G. Li, J.L. Shaffer, M.I. Azoulay, I.C. Gibbs, S. Nagpal, S.G. Soltys, Stereotactic Radiosurgery and Hypofractionated Radiotherapy for Glioblastoma, Neurosurgery. 82 (2018) 24–34.
[18]Q. Chai, Y. Jiao, X. Yu, Hydrogels for Biomedical Applications: Their Characteristics and the Mechanisms behind Them, Gels. 3 (2017) 6.
[19]S.M.F. Kabir, P.P. Sikdar, B. Haque, M.A.R. Bhuiyan, A. Ali, M.N. Islam, Cellulose- based hydrogel materials: chemistry, properties and their prospective applications, Prog. Biomater. 7 (2018) 153–174.
[20]G.F. Picheth, C.L. Pirich, M.R. Sierakowski, M.A. Woehl, C.N. Sakakibara, C.F. de Souza, A.A. Martin, R. da Silva, R.A. de Freitas, Bacterial cellulose in biomedical applications: a review, Int. J. Biol. Macromol. 104 (2017) 97–106.
[21]M.L. Cacicedo, I.E. León, J.S. Gonzalez, L.M. Porto, V.A. Alvarez, G.R. Castro, Modified bacterial cellulose scaffolds for localized doxorubicin release in human colorectal HT-29 cells, Colloids Surf. B Biointerfaces. 140 (2016) 421–
429.
[22]M.L. Cacicedo, M.C. Castro, I. Servetas, L. Bosnea, K. Boura, P. Tsafrakidou, A.
Dima, A. Terpou, A. Koutinas, G.R. Castro, Progress in bacterial cellulose matrices for biotechnological applications, Bioresour. Technol. 213 (2016) 172–180.
[23]M.M. Abeer, M.C.I. Mohd Amin, C. Martin, A review of bacterial cellulose-based drug delivery systems: their biochemistry, current approaches and future prospects, J. Pharm. Pharmacol. 66 (2014) 1047–1061.
[24]M.L. Cacicedo, G.A. Islan, I.E. León, V.A. Álvarez, I. Chourpa, E. Allard-Vannier, N.
García-Aranda, Z.V. Díaz-Riascos, Y. Fernández, S. Schwartz, I. Abasolo, G.R.
Castro, Bacterial cellulose hydrogel loaded with lipid nanoparticles for localized cancer treatment, Colloids Surf. B Biointerfaces. 170 (2018) 596–608.
[25]M.L. Cacicedo, G.A. Islan, M.F. Drachemberg, V.A. Alvarez, L.C. Bartel, A.D.
Bolzán, G.R. Castro, Hybrid bacterial cellulose–pectin films for delivery of bioactive molecules, New J. Chem. 42 (2018) 7457–7467.
[26]Y. Pötzinger, D. Kralisch, D. Fischer, Bacterial nanocellulose: the future of controlled drug delivery?, Ther Deliv. 8 (2017) 753–761.
[27]R.F. Barth, B. Kaur, Rat brain tumor models in experimental neuro-oncology:
the C6, 9L, T9, RG2, F98, BT4C, RT-2 and CNS-1 gliomas, J. Neurooncol. 94 (2009) 299–312.
[28]A. Clavreul, M. Delhaye, E. Jadaud, P. Menei, Effects of syngeneic cellular vaccinations alone or in combination with GM-CSF on the weakly immunogenic F98 glioma model, J. Neurooncol. 79 (2006) 9–17.
[29]M.L. Cacicedo, K. Cesca, V.E. Bosio, L.M. Porto, G.R. Castro, Self-assembly of carrageenin–CaCO3 hybrid microparticles on bacterial cellulose films for doxorubicin sustained delivery, J. Appl. Biomed. 13 (2015) 239–248.
[30]E. Goldschmidt, M. Cacicedo, S. Kornfeld, M. Valinoti, M. Ielpi, P.M. Ajler, C.
Yampolsky, J. Rasmussen, G.R. Castro, P. Argibay, Construction and in vitro testing of a cellulose dura mater graft, Neurol. Res. 38 (2016) 25–31.
[31]L. Stoppini, P.A. Buchs, D. Muller, A simple method for organotypic cultures of nervous tissue, J. Neurosci. Methods. 37 (1991) 173–182.
[32]P.J. Basser, D.K. Jones, Diffusion-tensor MRI: theory, experimental design and data analysis – a technical review, NMR Biomed. 15 (2002) 456–467.
[33]G.P. Winston, The physical and biological basis of quantitative parameters derived from diffusion MRI, Quant. Imaging Med. Surg. 2 (2012) 254–265.
[34]A. Beliveau, G. Thomas, J. Gong, Q. Wen, A. Jain, Aligned nanotopography promotes a migratory state in glioblastoma multiforme tumor cells, Sci. Rep. 6 (2016) 26143.
[35]A. Jain, M. Betancur, G.D. Patel, C.M. Valmikinathan, V.J. Mukhatyar, A.
Vakharia, S.B. Pai, B. Brahma, T.J. MacDonald, R.V. Bellamkonda, Guiding intracortical brain tumour cells to an extracortical cytotoxic hydrogel using aligned polymeric nanofibres, Nat. Mater. 13 (2014) 308–316.
[36]J. Johnson, M.O. Nowicki, C.H. Lee, E.A. Chiocca, M.S. Viapiano, S.E. Lawler, J.J.
Lannutti, Quantitative analysis of complex glioma cell migration on electrospun polycaprolactone using time-lapse microscopy, Tissue Eng. Part C Methods. 15 (2009) 531–540.
[37]F.M. Kievit, A. Cooper, S. Jana, M.C. Leung, K. Wang, D. Edmondson, D. Wood, J.
S.H. Lee, R.G. Ellenbogen, M. Zhang, Aligned chitosan-polycaprolactone polyblend nanofibers promote the migration of glioblastoma cells, Adv.
Healthc. Mater. 2 (2013) 1651–1659.
[38]H. Bäckdahl, G. Helenius, A. Bodin, U. Nannmark, B.R. Johansson, B. Risberg, P.
Gatenholm, Mechanical properties of bacterial cellulose and interactions with smooth muscle cells, Biomaterials. 27 (2006) 2141–2149.
[39]M. Paolillo, M. Serra, S. Schinelli, Integrins in glioblastoma: Still an attractive target?, Pharmacol Res. 113 (2016) 55–61.
[40] G. Tabatabai, J.-C. Tonn, R. Stupp, M. Weller, The role of integrins in glioma biology and anti-glioma therapies, Curr. Pharm. Des. 17 (2011) 2402–2410.
[41]R.A.N. Pértile, S. Moreira, R.M. Gil da Costa, A. Correia, L. Guãrdao, F. Gartner, M. Vilanova, M. Gama, Bacterial cellulose: long-term biocompatibility studies, J. Biomater. Sci. Polym. Ed. 23 (2012) 1339–1354.
[42]L. Tang, J.W. Eaton, Inflammatory responses to biomaterials, Am. J. Clin. Pathol.
103 (1995) 466–471.
[43]W.J. Hu, J.W. Eaton, T.P. Ugarova, L. Tang, Molecular basis of biomaterial- mediated foreign body reactions, Blood. 98 (2001) 1231–1238.
[44]V. Jadhav, J.H. Zhang, Surgical brain injury: prevention is better than cure, Front. Biosci. 13 (2008) 3793–3797.
[45]M. Ul-Islam, T. Khan, J.K. Park, Water holding and release properties of bacterial cellulose obtained by in situ and ex situ modification, Carbohydr.
Polym. 88 (2012) 596–603.
[46]G. Helenius, H. Bäckdahl, A. Bodin, U. Nannmark, P. Gatenholm, B. Risberg, In vivo biocompatibility of bacterial cellulose, J. Biomed. Mater. Res. A. 76 (2006) 431–438.
[47]F.G. Torres, S. Commeaux, O.P. Troncoso, Biocompatibility of bacterial cellulose based biomaterials, J. Funct. Biomater. 3 (2012) 864–878.
[48]A. Valster, N.L. Tran, M. Nakada, M.E. Berens, A.Y. Chan, M. Symons, Cell migration and invasion assays, Methods San Diego Calif. 37 (2005) 208–215.
[49]A. Müller, Z. Ni, N. Hessler, F. Wesarg, F.A. Müller, D. Kralisch, D. Fischer, The biopolymer bacterial nanocellulose as drug delivery system: investigation of drug loading and release using the model protein albumin, J. Pharm. Sci. 102 (2013) 579–592.
[50] V.P. Ferrer, V. Moura Neto, R. Mentlein, Glioma infiltration and extracellular matrix: key players and modulators, Glia. (2018).
[51]J.B. Wolinsky, Y.L. Colson, M.W. Grinstaff, Local drug delivery strategies for cancer treatment: gels, nanoparticles, polymeric films, rods, and wafers, J.
Control. Release 159 (2012) 14–26.
