Primary Glial Cell and Glioblastoma Morphology in Cocultures Depends on Scaffold Design and Hydrogel Composition

3D cell cultures better replicate the in vivo environment compared to 2D models. Glioblastoma multiforme, a malignant brain tumor, highly profits from its cellular environment. Here, the U87 glioblastoma cell line in the presence/absence of primary astrocytes is studied. Thiolated hyaluronic acid (HA‐SH) hydrogel reinforced with microfiber scaffolds is compared to Matrigel. Hyaluronic acid is a major extracellular matrix (ECM) component in the brain. Poly(ɛ‐caprolactone) (PCL) scaffolds are written by meltelectrowriting in a box and triangular shaped design with pore sizes of 200 µm. Scaffolds are composed of 10‐layers of PCL microfibers. It is found that scaffold design has an impact on cellular morphology in the absence of hydrogel. Moreover, the used hydrogels have profound influences on cellular morphology resulting in spheroid formation in HA‐SH for both the tumor‐derived cell line and astrocytes, while cell viability is high. Although cocultures of U87 and astrocytes exhibit cell–cell interactions, polynucleated spheroid formation is still present for U87 cells in HA‐SH. Locally restricted ECM production or inability to secrete ECM proteins may underlie the observed cell morphologies. Thus, the 3D reinforced PCL‐HA‐SH composite with glioma‐like cells and astrocytes constitutes a reproducible system to further investigate the impact of hydrogel modifications on cellular behavior and development.


Introduction
Glioblastoma multiforme (GBM) is the most aggressive and malignant brain tumor.Despite advances in cancer treatment, life expectancy after diagnosis is around 12-16 months. [1]High DOI: 10.1002/adbi.202300029proliferation and invasion activity as well as the heterogeneous appearance of the tumor are responsible for its malignancy. [2]The interaction with its tumor microenvironment (TME) is of particular importance for tumor proliferation, e.g., it was shown that tumor cells take advantage of the interactions with other cell types in the brain. [3,4]BM´s TME is composed of microglia, neurons, astrocytes, and vascular cells.[7] Furthermore, tumor-neuron interactions have been identified to enhance tumor proliferation and invasiveness. [8]Astrocytes from the TME produce cytokines and growth factors that maintain an immunosuppressed environment and promote cell proliferation. [9,10]Moreover, extracellular vesicles, gap junctions, and nanotubes are used to alter and manipulate the phenotype of cells in the TME. [11]Hence, glioma cell interactions with surrounding cell types represent an important feature underlying brain tumor pathology.
To further improve our current knowledge and understanding of TME interactions, 3D in vitro cell culture models represent www.advanced-bio.com a suitable tool and have many benefits compared to 2D culture techniques. [12]Advances in additive manufacturing techniques and biomaterials offer new platforms for the design and development of 3D in vitro models to mimic the native tissue microenvironment in terms of mechanical properties and composition. [13]elt electrowriting (MEW) is an additive manufacturing technique that can produce extremely porous scaffolds (>90% by volume). [14]MEW reinforced hydrogels provide numerous opportunities to simulate the tumor microenvironment in its content and mechanical properties.Using MEW, the internal architecture and design of the scaffold can be tailored to the mechanical characteristics and biological responses to the hybrid hydrogel scaffold composite. [15,16]ere, we evaluate the cellular morphology, viability, and cell interactions of glioma cells and astrocytes using different MEW fiber designs reinforced with thiolated hyaluronic acid (HA-SH) and Matrigel as hydrogels.The focus of the present study is to characterize the influence of scaffold design and hydrogel on cellular morphology and behavior.

Scaffold Design Has an Impact on Cellular Morphology
Primary neurons of the central nervous system (CNS) in their physiological environment are surrounded by mechanical stimuli essential for proper maturation and development. [17]In addition, structural guidance is also necessary for migration and thus, neuronal differentiation and maturation at the final cell localization.During these processes, extracellular matrix (ECM) proteins secreted by neighboring astrocytes play important guiding roles. [18]Tumor cells in the brain, e.g., glioblastoma cells, use structural and mechanical stimuli for their progression and development but also take advantage of the close proximity and function of neurons. [19]o further understand disease pathology, 3D cell culture models are required which simulate the native conditions in vitro.MEW provides a platform for precise and reproducible scaffold design, that may serve as structural guides for in vitro 3D disease modeling. [20]However, the geometric organization of the PCL fibers in MEW printed scaffold might have an impact on cellular morphology.To address this issue, we compared healthy brain cells (astrocytes) and tumor cells of a glioblastoma cell line (U87) for their cellular morphology at different scaffold designs composed of PCL fibers spun in box-forming or triangularforming structures with pore sizes of 200 μm with 3.7 ± 0.4 and 9.7 ± 0.2 μm in fiber diameter (Figure 1A), respectively.Primary mouse astrocytes infected with a lentivirus inducing green fluorescent protein (GFP) expression under the control of the cytomegalovirus (CMV) promoter and farnesylated-tdTomato expressing U87 glioma cells [21] were monitored for their morphology at those two MEW scaffolds.
Both cell types attach and grow along the PCL fibers of the scaffolds.Interestingly, morphological differences between both cell types were obvious when grown at the box scaffold compared to the triangular scaffold (Figure 1B,C).Using box scaffolds with a 200 μm pore size, astrocytes, and U87 cells grow along the PCL fibers (Figure 1B,C).Both cell types appear as stretched cells sometimes wrapping around the PCL fibers (Figure 1B,C).
Stretching throughout the pore has only been observed for astrocytes in the vertices of the boxes (Figure 1B).Within the vertices, the astrocyte cell bodies can be localized and surrounded by branches that further divide to branchlets.In vivo, numerous end feet of astrocytes contact neurons or vascular cells and thus are important for the nutrient supply of the neurons. [22]The localization within the vertices demonstrates that most likely, the physiological morphology is not supported by box scaffolds.
The natural morphology of U87 glioma cells is pleiomorphic depending on the environment.This cell type is highly mobile to adapt to environmental changes, e.g., nutrient supply, ECM, and mechanical cues. [23]Moreover, U87 glioma cells resemble a skating snail phenotype due to durotaxis of aligned PCL fibers. [24]he term skating snail refers to quick changes in morphology from moving around toward long attached cells along the fibers and again scrolling.Even though the box design is more complex than the aligned linear PCL fibers, the morphology and behavior of U87 cells remained the same (Figure 1C; and Video S1, Supporting Information).
The triangular scaffold design was produced with a similar 200 μm pore size.In contrast to the box scaffolds, the triangular scaffolds contain numerous crossing points of the fibers generating smaller pores in the shape of triangles (Figure 1A).Although cells grow again along the PCL fibers of the triangular scaffolds, the vertices with the smaller pores serve as a platform for astrocytes and U87 glioma cells to better develop their native morphology (Figure 1B,C).Interestingly, U87 glioma cell processes extend from the cell body to various directions, contact other PCL fibers, and even cross the pores of the scaffold (Figure 1B).Such behavior has not been observed for U87 cells seeded at box scaffolds.For astrocytes, the triangular PCL fiber design provides a surrounding that allows sending branches extending through the pore of the scaffolds.Astrocytic processes contact the fibers directly but also were able to use the small interspaces between single PCL fibers spun on top of each other to grow through.The contacts the astrocytes create with the PCL fibers remind to contacts of astrocytic end feet with vasculature structures.PCL frame stiffness depends on the porosity in a range of 1-15 kPa comparable to blood vessels. [14]

Profound Influence of Hydrogels on Cellular Morphology Independent of Scaffold Design
Hydrogels are widely used for 3D cell culture as they attempt to mimic ECM from native tissue.The chemical composition as well as mechanical properties of the hydrogels are important for cell viability and survival in various types of matrices. [25]As the brain is one of the weakest materials in the human body, ultraweak matrices (≈100 Pa) have been used to enable functional 3D neuronal network formation. [26]13a] Here, we used ultraweak Matrigel (31 Pa elastic modulus) reinforced with the two MEWprinted scaffold design (box and triangular) to study the impact of the hydrogels on astrocyte and U87 cell morphology. [26]A disadvantage of Matrigel is its origin from a mouse sarcoma resulting in high lot to lot variability of protein content and cytokine concentrations. [27]Therefore, we compared the cellular morphol- ogy to a hyaluronic acid-based hydrogel.Hyaluronic acid is a major component of the ECM in the brain. [28]13c,29] Cryo-scanning electron microscopy (cryo-SEM) was used to reveal contacts of the fibers and its hydrogel surrounding (Figure 1D).The pores of scaffolds were filled with the hydrogels forming the MEW-scaffold/hydrogel composites.The cryo-SEM images also provided insights to the porous structure and pore size of HA-SH and Matrigel.Even though HA-SH has a greater elastic modulus (95 Pa) compared to Matrigel (31 Pa), larger pore sizes have been observed for HA-SH than for Matrigel.In a recent study, HA-SH with similar molecular weight (230 kDa) and porous structure allowed the diffusion of molecules with sizes between 40 and 500 kDa. [30]For Matrigel, molecules between 20 and 70 kDa as well as liposomes of 100 nm are able to diffuse the hydrogel. [31]As the hydrogels used here, Matrigel and HA-SH, present similar mechanical properties and porous structure, we do not expect obstructed nutrient diffusion.
Interestingly, the comparison between Matrigel and HA-SH showed clear differences in the cellular morphology and behavior of astrocytes and U87 glioma cells.These morphological differences depend on the hydrogel used and were mostly independent from scaffold designs (Figures 2 and 3).
When seeded in the HA-SH/PEGDA hydrogel, astrocytes showed up as round cells and developed as spheroids of single cells (Figure 2A).This morphology was observed for both scaffold types, the triangular and the box shape, and was different from the natural star-like morphology seen in the brain (Figure 2A,B). [32]Astrocytic interactions with the scaffolds, while grown in HA-SH were rarely seen.This observation is in clear contrast to astrocytes seeded without hydrogel, which exhibited clear and scaffold-dependent interaction with the PCL fibers (Figure 1B).Astrocytes seeded in Matrigel developed a physiological morphology (Figure 2C,D) with processes spreading from the cell soma to neighboring astrocytes and the PCL fibers of the scaffolds.Quantification of the morphological changes of astrocytes determined significant differences in terms of area, circularity, solidity, and aspect ratio of the cells.Astrocytes in Matrigel displayed significantly larger surface areas compared to HA-SH for both scaffold designs (Matrigel-Triangular: 972 ± 107 μm 2 , HA-SH-Triangular: 230 ± 34 μm 2 , p < 0.0001; Matrigel-Box: 686 ± 82 μm 2 , HA-SH-Box: 133 ± 13 μm 2 , p < 0.0001; Figure 2E).The increased cell area was not accompanied by increased circularity.Astrocytes in vivo also do not appear round rather show numerous processes within well-delineated bushy territories. [32]A significant decrease of circularity was observed for astrocytes in Matrigel compared to HA-SH (Matrigel-Triangular compared to HA-SH-Triangular p < 0.0001 and Matrigel-Box compared to HA-SH-Box p = 0.0002) (Figure 2F).The stretched appearance with numerous processes of astrocytes in Matrigel is accompanied by a significantly increased aspect ratio compared to astrocytes in HA-SH in both scaffold types (Triangular: p = 0.0251; Box: p = 0.0431; Figure 2G).In line with those morphological observations, the solidity of astrocytes was significantly decreased in Matrigel compared to HA-SH and independent on scaffold design (Triangular: Matrigel versus HA-SH, p = 0.0002; Box: Matrigel versus HA-SH, p = 0.0097; Figure 2H).These morphological differences are most likely due to the lack of ECM proteins or growth factors present in Matrigel but not in pure HA-SH.Moreover, astrocytes show round morphology which might be due to their inability to secrete sufficient ECM proteins for their own proliferation.When astrocytes are activated, they are responsible for the production of brain ECM.It has been exhibited that astrocytes are activated when cultured in matrices with stiffnesses of around 1 kPa.In contrast, if astrocytes are grown in weak hydrogels of less than 150 Pa (ultraweak matrices), they are not activated, appear as spheroids and do not produce ECM proteins. [33]Furthermore, a significant increase in cell area was also present when comparing triangular with box frame composites in Matrigel (p = 0.0397) (Figure 2E) most likely due to the offer of smaller angles within the triangular scaffold promoting cellular attachment.Even though it was not significant, similar surface area differences were present in triangular compared to box composites in HA-SH.In addition, scaffold design had no significant impact on cellular circularity, solidity, and aspect ratio.These previous observations are in line with our data on astrocytes grown either at triangular PCL scaffolds without matrix compared to the same composite, including the hydrogel.
U87 glioma cells seeded in Matrigel developed a native morphology and differ significantly from growth in HA-SH (Figure 3).In Matrigel, U87 cells spread out throughout the gel after only 3 days in culture, independent of frame design (Figure 2C,D).The development of cell processes increased in both triangular and box frames.In contrast, U87 glioma cells formed spheroids in HA-SH.A key difference to astrocytic spheroids in HA-SH is that the spheroids formed by U87 cells were polynucleated (Figure 3A,B).The frame design did not affect U87 cell growth as spheroids in HA-SH.The area of U87 glioma cells significantly increased in Matrigel compared to HA-SH (Matrigel-Triangular 303 ± 34 μm compared to HA-SH-Triangular 124 ± 14 μm 2 , p < 0.0001, Matrigel-Box 267 ± 28 μm 2 compared to HA-SH-Box 138 ± 14 μm 2 , p = 0.0017; Figure 3E).Like astrocytes, a significant increase of circularity is present in HA-SH compared to Matrigel (for both scaffold types p < 0.0001; Figure 3F).In line with this, a low aspect ratio of U87 was observed in HA-SH with both triangular-and box-shaped scaffolds.The use of Matrigel exhibited a larger impact on the aspect ratio of the U87 cells than the scaffold design (Box: Matrigel vs HA-SH p = 0.0012; Figure 3G).Cellular solidity was significantly decreased in triangular-shaped compared to box-shaped scaffold in Matrigel (p = 0.0104) again arguing for a promoting effect of the smaller angles in the triangular scaffolds serving as cellular attachment points.A tendency to decreased solidity was also observed in HA-SH when using triangular-shaped compared to box-shaped scaffolds.However, besides scaffold designs the solidity was mainly affected by the hydrogels used with significantly decreased values when using Matrigel compared to HA-SH (Triangular: p < 0.0001, Box p < 0.0001; Figure 3H).The observation that both cell types grew with similar morphologies in HA-SH strengthens the idea that either lack of ECM components in HA-SH, the inability of the cells investigated to produce their own ECM, or the absence of a locally restricted ECM distribution is underlying these phenotypes.Hence both, U87 cells and astrocytes by themselves were unable to secrete sufficient ECM components which would allow to obtain their proper morphological development.

Although HA-SH Lacks Extracellular Matrix Protein, Cell Viability is High
Astrocytes as well as U87 glioma cells exhibit a non-native morphology in HA-SH hydrogel.Interestingly, cell viability of both cell types, astrocytes, and U87 glioma cells, was high in both matrices used (Figure 4A-D).Astrocyte cultures were analyzed at day 7 and U87 glioma cells at day 3 in vitro due to their differences in cell division times.No significant decreases in the survival rate of both cell types were observed between the hydrogels, HA-SH and Matrigel (Figure 4B,D).For astrocytes, the percentage of living cells was between 73.37% ± 2.5 to 79.52% ± 2.7 (Figure 4A,B).For U87 glioma cells, a range of living cells between 75.18% ± 2.6 to 80.38% ± 2.4 was estimated (Figure 4C,D   and Table 1).These results show that even though HA-SH lacks the appropriate ECM components to ensure the development of a proper cell morphology, it does not have a detrimental effect on the survival of astrocytes and U87 glioma cells.

Cocultures of Astrocytes and U87 Glioblastoma Cells Promote Cell-Cell Interactions
In 3D monocultures of astrocytes and U87 glioma cells using HA-SH as a hydrogel, both cell types were unable to produce sufficient ECM components or other growth factors necessary for their proper morphological development.Various coculture methods have been utilized to study glioma-astrocytes interaction in 2D and 3D. [34,35]Thus, a coculture of these two cell types simulating their close proximity in the brain was designed to overcome the lack of cytokines or ECM components.Similarly, it has been shown that 3D cocultures of astrocytes and neurons enhance the development and growth of both cell types.13c,36] Our experimental setup included seeding of astrocytes to the scaffold 2 days before adding the U87 glioma cells together with the corresponding hydrogel.The reason for this segmental seeding was the faster growth of U87 cells in comparison to astrocytes.Using both scaffold designs, triangular and box, the hydrogels Matrigel and HA-SH were compared.When no hydrogel was used (Figure 5A,B), U87 cells and astrocytes attached to the scaffolds and were growing along the fibers, thus resembling the same behavior as in monoculture (Figure 1).At day 7, the cell viability and cellular morphology were analyzed.The presence of both cell types in HA-SH was not sufficient to overcome the roundish cell phenotype (Figure 6A,B).Cocultures of astrocytes and U87 glioma cells in HA-SH using the triangular scaffold and box scaffold differ in their morphologies compared to monocultures of both cell types.U87 cells grow mainly in the gel as polynucleated spheroids.On the other hand, astrocytes grow mainly attached to the scaffolds.They were thinner in diameter and spread less when surrounded by HA-SH compared to Matrigel (Figure 6C,D) and when cells were seeded at scaffolds alone (Figure 5).The morphological differences may however depend on the segmental seeding approach with seeding astrocytes together with scaffolds 2 days before the tumor cells were added together with the hydrogel.
Matrigel cocultures, as the monocultures, represent a suitable hydrogel for astrocytes and U87 cells to grow and proliferate.No significant distinction can be seen from cocultures in triangular scaffolds to box scaffolds.Astrocytes and U87 cells are able to interact in both hydrogels.This can be seen by the proximity of both cell types, especially evident in the 3D reconstructions of Matrigel cocultures (Figure 6C,D).
The viability of the cocultures of astrocytes and U87 glioma cells in both hydrogels, Matrigel and HA-SH was again high (Fig- ure 7A,B).A high percentage of living cells was present in all experimental conditions, with no significant differences between each other (Table 1).Our studies demonstrate that our established 3D in vitro scaffold-hydrogel-composite allows investigations of cell-cell interactions between astrocytes and glioma cells.

Conclusion
We have systemically investigated the role of scaffold design in the absence and presence of hydrogels for cellular morphology of astrocytes (healthy) and glioblastoma (tumor) cells to establish a 3D cell culture model suitable to study interactions of neural cells with tumor cell models.Using MEW, box-shaped PCL frames and triangular-shaped frames were printed which by themselves influenced cellular shapes.Both, astrocytes and glioblastoma cells preferred growth along the PCL fibers with the formation of triangular physiological cell shape in the smaller degree angles of the triangular scaffolds only.When using hydrogels reinforced by both scaffold designs, cellular morphology differed in Matrigel, rich in growth factors and ECM proteins, compared to HA-SH, most probably due to lack of sufficient ECM production by both cell types in the HA-SH environment.While exhibiting a spread-out cell shape with processes in Matrigel, both cell types formed spheroids in HA-SH.In cocultures of astrocytes and tumor cells in HA-SH, the formation of polynucleated spheroids by the tumor cells is less likely, arguing for some subsidiary function of the astrocytes.Hence, our current 3D model suggests that the direct neighborhood of both cell types initiates cell-cell interactions to promote cellular viability and more natural morphology.Thus, our 3D model allows to systematically include other neural and non-neural cell populations present in the brain to build up a construction kit for studying brain tumor cells in a defined, reproducible environment.Such a study design would allow to objectively characterize signal transmission between disease-related cell types with other brain cells.

Experimental Section
MEW Scaffolds Printing: A custom-built MEW printer was used to fabricate scaffolds as previously described. [37]MEW was performed using medical-grade PCL (PURASORB PC 12, Lot#1 712 002 224, 05/2018 Corbion Inc, Amsterdam, Netherlands) at 21 ± 3 °C and humidity of 35 ± 12%.The following parameters were used: 77 °C; 2 bar of air pressure; 30 gauge nozzle; 4 kV voltage applied across a 1.4 mm collector distance for the triangular scaffolds.Parameters for 200 μm box scaffolds were: 80 °C; 3 bar of air pressure; 25G nozzle; 6 kV voltage applied across a 4 mm collector distance. [37]ynthesis of Thiolated Hyaluronic Acid (HA-SH) and Polyethylene Glycol-Diacrylate (PEGDA): The synthesis of HA-SH was performed with a number of modifications from Stichler et al. and has been previously described. [29]PEGDA was synthesized according to Cruise et al. [30,38] HA-SH was obtained as white foam after freeze drying.The molecular weight of HA-SH was M w 264 kDa, DS 44 %.
Scaffolds and Hydrogels Preparation: MEW scaffolds were washed once with 70% ethanol, three times with ddH 2 O, once with PBS.Washed scaffolds, HA-SH and O-rings were placed 15-30 min under UV light.The scaffolds were then placed in 3 cm dishes with four 93 mm 2 inner rings (627170, Greiner, Greiner Bio-One, Kremsmünster, Austria).Scaffolds were incubated for 30 min at 37 °C with 5% CO 2 , with 50 μL of medium.
HA-SH was cross-linked with PEGDA.1.5% w/v HA-SH was dissolved with HEPES buffer (154 mm, pH 7.6) and mixed with 1.5% PEGDA in PBS.The solution was incubated for 25 min at 37 °C before used for cell seeding.
Cell Seeding and Culture: For monocultures, each cell type was detached, centrifuged, counted, and cell pellets were mixed with the appropriate medium and the respective hydrogel solution (see Scaffold and Hydrogels preparation).70-100 μL containing a mixture of cells, media, or hydrogel were pipetted on top of each scaffold (box or triangle).An O-ring was used to fix the scaffolds.Matrigel, HA-SH/PEGDA and scaffolds with no hydrogel were incubated for 30 min at 37 °C with 5% CO 2 .Afterward 3 mL of supplemented DMEM medium was added.5000 U87 cells were seeded and 150 000 astrocytes.Cells were grown under standard growth conditions.
For cocultures, astrocytes were dissociated and seeded on top of MEW scaffolds 2 days before seeding of U87 cells.150 000 astrocytes were seeded in each scaffold.At day 3, 5000 U87 cells were added.After a 30 min incubation at 37 °C with 5% CO 2 , 3 mL of DMEM medium were added.Cells were grown under standard growth conditions.
Cell Viability: Viability was assessed at day 3 of culture for scaffolds seeded with U87 MG cells and day 7 for astrocytes and cocultured cells.Cells were incubated at 21 °C for 30 min with Calcein-AM (1:500, green/living cells; Thermo Fisher Scientific, Waltham, MA) and Ethidium Homodimer (1:1000, red/dead cells; Sigma-Aldrich, St. Louis, MO) diluted in PBS.Using the Spots function of the Imaris Software 7.7.2(Oxford Instrumentals, Abingdon, UK) the live/dead ratio was assessed.For each experimental condition, 5 image stacks per scaffold were analyzed (n = 3).
Immunocytochemistry: Immunocytochemical stainings on MEW scaffolds and MEW reinforced hydrogels were performed using cultures of GFP-expressing astrocytes and U87 dTom cells.Composites were washed with PBS and fixed for 10-15 min with 2% paraformaldehyde solution.As both cell types have fluorescent properties, composites were directly mounted on glass slides without blocking or further incubation periods.Mounting was done with ProLong Glass Antifade Mountant containing Hoechst 33 342 (Thermo Fisher Scientific, Waltham, MA).
Image Acquisition and Confocal Microscopy: Images obtained from cell viability experiments or from immunocytochemical stainings were acquired using an Olympus IX81 microscope equipped with a FV1000 confocal laser scanning system, a FVD10 SPD spectral detector, and diode lasers of 405, 495, 550, and 635 nm (Olympus).Olympus UPLSAPO 20x (air, numerical aperture 0.75) objective and UPLFLN 40x (oil, numerical aperture 1.3) objectives were used.3D reconstructions were done with the Imaris Software (Oxford Instrumentals, Abingdon, UK).Dynamic range adjustments and projections were done with ImageJ/Fiji Software. [41]orphology Analysis: The FIJI software was used to quantify cellular morphological features, e.g., cell area, circularity, solidity, and aspect ra- tio.Image analysis was performed with maximal intensity z-projections.At least 6 cells were analyzed for each scaffold type and hydrogel.
Scanning Electron Microscopy (SEM): To prepare the samples for SEM imaging, the cells were, after an incubation period of 7 days (astrocytes) and 3 days (U87 glioma cells), fixed with 4% PFA for 15 min on ice.After two PBS incubations on ice, the samples were dehydrated using a graded ethanol series (from 50% to 100%).Samples were frozen 1 h at −80 °C and later dried at 37 °C.To ensure sufficient surface conductivity for SEM imaging, the samples were sputter-coated with a 3.5 nm platinum layer (Leica EM ACE600, Wetzlar, Germany).The SEM images of the MEW scaffolds and the cells were taken utilizing an acceleration voltage of 2 kV and by detecting secondary electrons (Zeiss Crossbeam 340, Oberkochen, Germany).
Cryo-Scanning Electron Microscopy: The samples were placed in a small notch (d = 2 mm) in between two aluminum holders (d = 3 mm) and frozen in slushed nitrogen at −210 °C.The frozen samples were then transferred with a Leica EM VCT100 cryo-shuttle into a Leica EM ACE600 Sputter Coater.Here, under vacuum (< 1 × 10 −3 mbar), the top half of the samples were knocked off, freeze etched for 15 min at −85 °C and coated with 2.5 nm of platinum at −140 °C.Finally, samples were transferred into a Zeiss Crossbeam340 Field-Emission Scanning Electron Microscope and secondary electron images of the fractured sample surface were taken at −140 °C and an electron acceleration voltage of 8 kV.
Statistical Analysis: Descriptive statistics (means, standard deviation, errors) and significance were calculated with GraphPad Prism 9.0.0 (GraphPad Software, San Diego, CA).To determine statistical significance in the live/dead assays, an unpaired t-test with Welch´s correction was performed.For morphological analysis a two-way-ANOVA test was used.p-value level are as follows *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 1 .
Figure 1.Scaffold design has an impact on the cell morphology of astrocytes and U87 glioma cells.A) MEW scaffold designs composed of PCL fibers with a 200 μm pore, upper image: box design, lower image: triangular design.B) Astrocytes seeded on box and triangular MEW scaffold designs.C) U87 glioma cells seeded on box and triangular scaffolds.D) Cryo-SEM images of HA-SH/PEGDA and Matrigel hydrogels.Note, both hydrogels attach to MEW fibers and fill the scaffolds pores.B,C) Left images represent confocal images of fluorescently labeled cells.Right images have been taken by SEM microscopy.Astrocytes and U87 glioma cells were color coded after image acquisition.Scale bar: confocal images 50 μm, SEM images 20 μm, Cryo-SEM images 2 μm.

Figure 2 .
Figure 2. Morphological differences of astrocytes exist when seeded in HA-SH compared to Matrigel.A) HA-SH triangular, B) HA-SH box 3D monocultures of GFP-expressing astrocytes (green).Cells grow as spheroids.C) Matrigel triangular monocultures of GFP-expressing astrocytes (green).Note, astrocytes acquire morphological states closer to their physiological state and continue to interact with MEW scaffolds.D) Matrigel box monocultures.Likewise, in Matrigel box cultures, cells grow with a physiological morphology.Scale bar: confocal/3D reconstructions/side views 50 μm, inlets of 3D reconstructions: scale bar refers to 20 μm.E) Quantification of the cell area, F) analysis of the cell circularity, G) aspect ratio and H) solidity of astrocytes in in Matrigel and HA-SH with triangular-shaped and box-shaped PCL frames.Significance values *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 4 .
Figure 4. High cell viability of astrocytes and U87 glioma cells in 3D monocultures using MEW reinforced hydrogels.Cells in both Matrigel and HA-SH/PEGDA hydrogels show high viability, regardless of the scaffold design.A) Confocal images of astrocytes from the viability assay (live/green, dead/red cells).B) Quantification of the percentage of living cells that ranges from 73.37% to 79.52%.C) Confocal images of the viability assay for U87 glioma cells in monocultures.D) Quantification of the percentage of living cells ranging from 75.18% to 80.38%.Scale bar: 50 μm.
a) SEM = Standard error of the mean;b) A = astrocytes; c) U = U87 cells.

Figure 5 .
Figure 5. Astrocytes and U87 glioma form cell-cell interactions when grown in cocultures on different scaffolds.Cocultures of astrocytes and U87 glioma cells in triangular A) and box B) MEW scaffolds without hydrogel.Both cell types attach and use the scaffolds as guidance for their own growth.Scale bars: confocal/3D reconstructions/side views 50 μm, detailed 3D reconstructions 20 μm.

Figure 6 .
Figure 6.Astrocytes and U87 glioma form cell-cell interactions in both hydrogels HA-SH and Matrigel.A) and C) Cocultures of astrocytes and U87 glioma cells in HA-SH and Matrigel in triangular scaffolds (astrocytes GFP/green, U87dTom/magenta).U87 glioma cells grew mainly as polynucleated spheroids, and when attached to the scaffolds U87 cells do not spread as much as in Matrigel.B,D) Cocultured astrocytes and U87 cells in HA-SH or Matrigel using box scaffolds.Note, in Matrigel cocultures of astrocytes and U87 glioma cells grow and spread within the hydrogel and along the PCL fibers.Detailed images show cell-cell contacts (filled white arrow head).Scale bars: confocal/3D reconstructions/side views 50 μm, detailed 3D reconstructions 20 μm.

Figure 7 .
Figure 7. Cocultures of astrocytes and U87 cells exhibit high viability.A) Confocal images obtained from the cell viability assay of cocultures between astrocytes and U87 glioma cells (live/green, dead/red).B) Quantification of cell viability of co-cultures shown in percentages of living cells (range between 77.77% and 82.63%).Scale bar 50 μm.

Table 1 .
Viability of astrocytes, glioblastoma cells and cocultures of both in different matrices and scaffolds..