Article pubs.acs.org/Biomac

Hyaluronic Acid-Based 3D Culture Model for In Vitro Testing of Electrode Biocompatibility Andrea F. Jeffery,†,‡ Matthew A. Churchward,‡,§ Vivian K. Mushahwar,‡,∥,⊥ Kathryn G. Todd,§,‡,⊥ and Anastasia L. Elias*,†,‡ †

Chemical and Materials Engineering, University of Alberta, Edmonton, AB T6G 2V4, Canada Department of Psychiatry, University of Alberta, Edmonton, AB T6G 2G3, Canada ∥ Division of Physical Medicine and Rehabilitation, University of Alberta, Edmonton, AB T6G 2E1, Canada ⊥ Centre for Neuroscience, University of Alberta, Edmonton, AB T6G 2E1, Canada ‡ Alberta Innovates-Health Solutions Interdisciplinary Team in Smart Neural Prostheses (Project SMART), University of Alberta, AB, Canada §

S Supporting Information *

ABSTRACT: This work describes the development of a robust and repeatable in vitro 3D culture model of glial scarring, which may be used to evaluate the foreign body response to electrodes and other implants in the central nervous system. The model is based on methacrylated hyaluronic acid, a hydrogel that may be photopolymerized to form an insoluble network. Hydrogel scaffolds were formed at four different macromer concentrations (0.50, 0.75, 1.00, and 1.50% (w/v)). As expected, the elastic modulus of the scaffolds increased with increasing macromer weight fraction. Adult rat brain tested under identical conditions had an elastic modulus range that spanned the elastic modulus of both the 0.50 and 0.75% (w/v) hydrogel samples. Gels formed with higher macromer weight fraction had decreased equilibrium swelling ratio and visibly thicker pore walls relative to gels formed with lower macromer weight fractions. Mixed glial cells (microglia and astrocytes) were then encapsulated in the HA scaffolds. Viability of the mixed cultures was most stable at a cell density of 1 × 107 cells/mL. Cell viability at the highest macromer weight fraction tested (1.50% (w/v)) was significantly lower than other tested gels (0.50, 0.75 and 1.00% (w/v)). The inflammatory response of microglia and astrocytes to a microelectrode inserted into the scaffold was assessed over a period of 2 weeks and closely represented that reported in vivo. Microglia responded first to the electrode (increased cell density at the electrode, and activated morphology) followed by astrocytes (appeared to line the electrode in a manner similar to glial scarring). All together, these results demonstrate the potential of the 3D in vitro model system to assess glial scarring in a robust and repeatable manner.

■

INTRODUCTION Neural prostheses are a promising approach to restore function after neural injury or disease, by interfacing external devices directly to the nervous system.1 A major limiting factor for clinical implementation of neural prostheses is electrode biocompatibility in the central nervous system (CNS).2 Recording electrodes must provide stable and consistent signals for decades before they may be reliably used in neural prostheses. Current electrode designs are incapable of mitigating the cellular and tissue response of the CNS in order to enable chronic neural recordings.3−5 Electrode insertion causes damage resulting in many physical, chemical, cellular, and physiological changes including increased cell proliferation, compromised blood−brain barrier integrity, CNS infiltration of circulating macrophages, and increased expression of proinflammatory cytokines.4 In addition, the chronic presence of the electrode causes a cellular response that extends beyond the immediate damage due to © 2014 American Chemical Society

insertion including migration and proliferation of activated glial cells at the electrode interface.6 The increased density and activation of glial cells at the electrode interface is known as the glial scar, and is a major contributor to electrode failure. The formation of the glial scar has been well-documented using in vivo animal models.6−8 Microglia respond first to the electrode within hours to days followed by astrocytes within days to weeks (Figure 1).3 The sustained activation of microglia and astrocytes at the electrode forms a dense glial sheath that persists for many months after insertion. This glial sheath may increase impedance as well as prevent the growth of neurites toward the electrode thereby resulting in loss of electrode functionality.4,5 Received: February 28, 2014 Revised: May 14, 2014 Published: May 18, 2014 2157

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

Figure 1. Schematic of the cellular response in the CNS after electrode insertion. Microglia (red) are the first to react, becoming activated and taking on an amoeboid morphology. Astrocytes (green) react later becoming hypertrophic and forming a dense glial scar around the electrode.

matrix may also be modified by varying HAMA macromer weight fraction prior to photo-cross-linking. As the macromer weight fraction of the gel increases, so does its modulus.24 In this work, we investigated the use of a hyaluronic acid based 3D culture system with a tunable modulus for the analysis of glial scarring in response to microwire insertion. Gels were formed using various macromer weight fractions of HAMA (0.50, 0.75, 1.00, 1.50% (w/v)), and uniaxial compression testing was performed to characterize their mechanical properties. Whole rat brain was tested under the same conditions to see if the modulus of CNS tissue could be matched by the gels. The gel structure and porosity were then investigated to observe changes that occur as macromer weight fraction is altered. This is particularly important because cell viability and behavior are highly dependent on the structure and pore size of the gel.24,30−32 Primary mixed glial cultures were then photoencapsulated in the HAMA matrix. Cells were seeded at densities ranging from 1 × 106 to 1 × 107 cells/mL, to determine optimal viability. Cell morphology of microglia and astrocytes was studied, and their response to microwire insertion was analyzed. Together these results were used to determine the feasibility of the hyaluronic acid based 3D culture as a model of glial scarring.

Many factors influence the extent of glial scarring including electrode size,9 electrode material,3 and mechanical mismatch between the tissue and the electrode.10−14 Different strategies are currently being pursued to reduce the degree of glial scarring by altering electrode design parameters and device handling. These strategies include the use of bioactive coatings to modify cell/implant interactions,15−17 development of different insertion strategies,18 reduction of electrode size,9 and integration of more flexible materials into the devices themselves.10,19,20 Extensive testing is required to understand the impact of these approaches and to compare their relative effect on glial scar formation. In vivo animal studies currently used to assess electrode biocompatibility are time-consuming and costly due to the number of animals required. In addition, comparing results obtained from in vivo analysis is challenging due to variability inherently present in these systems. Various 2D culture models of in vitro glial scarring have been developed to allow controlled comparisons of electrode coatings over time.21−23 These models offer robust and controlled investigations of the interaction between glial cells and the electrode, but remove cells from their natural 3D environment and do not model the mechanical nature of the CNS. In addition to the reductionist perspective these models offer (microglia and astrocytes cultured in isolation from other CNS cell types), it would be beneficial for the in vitro model to mimic the modulus of CNS tissue. Mimicking the modulus of the CNS enables the influence of electrode flexibility, size, and micromotion to be assessed in vitro. This cannot be achieved in 2D culture, but may be modeled in 3D culture systems. To the authors’ knowledge, one other 3D model of glial scarring exists.22 This model investigated the increase of electrode impedance over time due to glial scarring, but did not model the mechanical properties of the CNS.22 A scaffold with a tunable modulus would allow modeling of glial scarring in different regions of the CNS (e.g., brain vs spinal cord). The primary constituent of CNS extracellular matrix (ECM) is hyaluronic acid (HA), a glycosaminoglycan capable of binding to cell surface receptors necessary for cell proliferation and attachment.24−26 The hydrophilic nature of hyaluronic acid allows it to imbibe large amounts of water (forming a hydrogel) and is easily modified to form methacrylated hyaluronic acid (HAMA) to enable photo-cross-linking.27 HAMA has been used previously to encapsulate cells at physiological temperature (37 °C) via irradiation with visible light (∼523 nm) in the presence of biocompatible precursors.24,25,27,28 Typically, cell densities of 1 × 106 to 1 × 108 cells/mL29,30 are used during 3D culture of microglia and astrocytes. The modulus of an HAMA

2.0. MATERIALS AND METHODS 2.1. Materials. 1-Vinyl-2-pyrrolidinone (NVP, V3409−5G), antimouse GFAP (G 3892), eosinY (EY, 119830), hyaluronic acid (HA, 53747−10G, source: Streptococcus equi, MW: 1.5 to 1.8 × 106 Da), methacrylic anhydride (MA, 275585−100 ML), poly-L-lysine (PLL, P-6282), and triethanolamine (TEA, 90279) were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.) and used as received. Alexa Fluor 488 donkey antirabbit (A21206), Alexa Fluor 647 donkey antimouse (A31571), Syto13Green Nucleic Acid Stain (S7575), and SytoxOrange Nucleic Acid Stain (S11368) were purchased from Molecular Probes (Life Technologies, Burlington, ON, Canada). The primary antibody antirabbit Iba1 (019−19741) was purchased from Wako (Osaka, Japan). Ethanol (EtOH, E112) was purchased from Commercial Alcohols Inc. (Brampton, ON, Canada). Polydimethylsiloxane (PDMS, Sylgard 184 Silicone Elastomer Kit) was purchased from Paisley Products of Canada (Toronto, ON). Platinum/iridium microwires, 30 μm in diameter insulated with 4 μm of polyimide were ordered from California Fine Wire (Grover Beach, CA, U.S.A.). All cell culture components were purchased from Gibco (Life Technologies, Burlington, ON, Canada). 2.2. Microglia and Astrocyte Isolations. All animal protocols were approved by the Animal Care and Use Committee at the University of Alberta and conducted in accordance with animal ethics guidelines. Mixed glial cultures (astrocytes and microglia) were obtained from postnatal day one Sprague−Dawley rats as described previously.33 Rats were decapitated and the brain exposed. Whole 2158

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

brains were removed using a metal spatula and placed in Hank’s balanced salt solution (HBSS) with 2% penicillin-streptomycin (PS). Meninges and surface vasculature were removed using forceps. The brain was then minced into ∼1 mm3 pieces using a scalpel. Minced whole brains were dissociated by enzymatic digestion in 0.25% Trypsin and 1 mM Ethylenediaminetetraacetic acid (Trypsin-EDTA) at 37 °C. The solution was then further dissociated into a single cell suspension via trituration. Cells were cultured in 12 well tissue culture plates precoated with PLL at 37 °C in a 5% CO2 incubator. Culture media, Dulbecco’s modified Eagle’s medium and Ham’s nutrient mixture F-12 in a 1:1 ratio (DMEM/F12) supplemented with 10% fetal bovine serum (FBS) and 2% PS, was changed twice weekly. Mixed glial cultures were maintained in culture for 2−3 weeks before encapsulation in HAMA gels. Prior to encapsulation, cellular characterization was determined via immunofluorescence.33 2.3. Macromer Synthesis. Hyaluronic acid for photo-crosslinking was prepared following a protocol adapted from the literature.22 Hyaluronic acid sodium salt (375 mg) was dissolved in distilled water (37.5 mL). A total of 20 molar equivalents of MA (∼3000 μL) were added to the solution in 200 μL increments under vigorous mixing at room temperature for 2 h. NaOH (5.0 M) was used to maintain pH between 8 and 12. The reaction was then left overnight at 4 °C. Methacrylated HA (HAMA) was then precipitated in cold EtOH for 24 to 72 h. The white fibrous precipitate was then lyophilized (−55 °C and 0.01 Torr) using a Savant SuperModulyo Freeze-Dryer (ThermoScientific, Waltham, Massachusetts, U.S.A.) and ground to form a white powder, which was stored at −20 °C. The presence of methacrylate groups on the HA backbone was confirmed via 1H NMR spectroscopy,24 as shown in Figure S1 of the Supporting Information. The degree of methacrylation was calculated from the ratio of relative peak integrations of the methacrylate protons (peaks at ∼6.1 and ∼5.6 ppm) and methyl protons on HA (∼1.9 ppm). The average degree of methacrylation was determined to be 95 ± 11%. 2.4. Gel Formation. HAMA powder was rehydrated to the desired monomer weight percent (0.5−1.5% (w/v)) in phosphate buffered saline (PBS) to form an aqueous solution. Immediately before crosslinking, precursors were added to the solution (0.1% (w/v) TEA, 0.1% (w/v) NVP, and 0.01 mM EY) and mixed vigorously. The aqueous HAMA was then applied to PDMS molds (a 1 mm thick PDMS gasket with an outer diameter of 15 mm and an inner diameter of 10 mm placed on top of a glass coverslip) and exposed to LED light (LED Supply, Cree XPE, Indus Star 3-Up Green High Power LED, Randolph, VT, U.S.A., ∼520 nm, 60 mW) for 2 min. 2.5. Mechanical Characterization. The compressive modulus of the HAMA gels at various macromer weight fractions (0.50 to 1.50% (w/v)) was determined using an Instron 5943 (Instron, Norwood, MA, U.S.A.) with a 10 N load cell. Gels were formed in a circular mold 20 mm in diameter and 3 mm thick. The compressive tests were performed at room temperature and the strain rate for each test was 1.0 mm min −1. Each test was ended when the gel had reached 50% strain. The elastic modulus (E) was then calculated from the slope of the stress−strain curve at 10% strain (in the linear region). For each macromer weight fraction, three independent gels were tested (N = 3). Three separate samples from one adult rat brain were also tested (N = 1). 2.6. Gel Swelling. To investigate the maximal amount of water that could be taken up by the gels, gels were prepared on coverslips at each macromer weight fraction (0.50, 0.75, 1.00, and 1.50% (w/v)) in a manner identical to that used for cell encapsulation and immersed in PBS for 24 h. This was done to simulate swelling as it would occur during cell encapsulation and facilitate handling of the soft hydrogels. Subsequently, gels were removed from PBS and excess liquid was carefully absorbed with Kimwipes (Kimberly-Clark, Irving, Texas, U.S.A.); the gels were weighed again to determine final gel weight (wwet). Equilibrium swelling was calculated as the ratio between wwet and the dry weight of the HAMA powder (wdry), as expressed in eq 1.34 The average weight of a coverslip was accounted for in the calculations. For each macromer weight fraction, three independent gels were tested (N = 3).

equilibrium swelling ratio(Q ) =

wwet wdry

(1)

2.7. Gel Porosity. Cryo-SEM was utilized to explore the structure of the pores in a hydrated state. A Zeiss NV40 Dual Beam (Carl Zeiss AG, Oberkochen, Germany) focused ion beam/scanning electron microscope (FIB/SEM) with a modified Leica VCT 100 attachment (Leica Camera AG, Solms, Germany) was used for imaging. Samples with macromer weight fractions between 0.50 and 1.50 % (w/v) were prepared on coverslips and left overnight in distilled water (ddH2O). The samples were rapidly frozen through immersion in liquid nitrogen to avoid formation of crystalline water and reduce expansion during the freezing process. The sample was fractured while completely immersed to provide a clean face for imaging. To preserve the cleaved face, the samples were loaded into a transfer shuttle purged with dry nitrogen gas and subsequently evacuated and pumped to high vacuum. To keep the samples frozen during transfer, the samples were loaded into the microscope through a specialized feed-through port onto a premounted, precooled stage. The stage was warmed up from −140 °C (133 K) to −60 °C (213 K) to expedite the sublimation of water from the sample surface, expose the sample’s inner structure, and allow for the imaging of wall thickness. During sublimation, chamber pressure reached 5.0 × 10−4 Pa and stabilized to 7.0 × 10 −5 Pa. Standard imaging conditions consisted of using a 5 kV accelerating voltage through 30 μm aperture. High-resolution images were captured using the secondary electron in-lens detector. 2.8. 3D Cell Encapsulation. Mixed glial cultures were photoencapsulated into the HAMA gel, as is outlined in Figure 2. After 1

Figure 2. Three dimensional culture technique. (a) Cells are mixed with the aqueous HAMA solution with photosensitive precursors and dispensed into a mold. (b) The macromer is cross-linked through exposure to ∼523 nm high intensity LED light, 60 mW for 2 min. (c) The mold is removed, leaving an insoluble cross-linked HAMA gel with encapsulated cells. (d) A total of 3−4 days post-encapsulation, a microwire is inserted into the HAMA gel and the cellular response is observed at select time points. week of culture, mixed glial cells were lifted from 12 well plates via 0.063% Trypsin and 0.25 mM EDTA in DMEM/F12. Cells were then collected, resuspended in DMEM/F12, and triturated to form a single cell suspension. Seeding densities and appropriate cell suspension volumes were calculated using Trypan blue staining. Cell suspensions were centrifuged to produce a pellet (200 g for 2 min), and this pellet was then resuspended in the precursor solution containing HAMA monomer (0.50 to 1.50% (w/v)) and photoinitiators (0.1% (w/v) TEA, 0.1% (w/v) NVP, and 0.01 mM EY) in PBS. This cell solution was then applied to PDMS molds (a 1 mm thick gasket with an outer diameter of 15 mm and an inner diameter of 10 mm) on PLL precoated coverslips. After 2 min of exposure to an LED light (∼520 nm, 60 mW), PDMS molds were removed. Gels were then placed into 12-well plates, covered with DMEM/F12 containing 10% FBS and 2% PS and incubated at 37 °C in a 5% CO2. Culture media, DMEM/F12 supplemented with 10% FBS and 2% PS, was changed twice weekly for the duration of all experiments. 2.9. Cell Viability. Cell viability was determined using a live/dead membrane permeability assay. Syto13 green fluorescent nuclear stain (5 mM) was used to label cell nuclei of all cells and SytoxOrange nuclear stain (5 μM) was used to label cells with compromised plasma membranes, as per distributors protocols. Syto13 was applied for 60 min and SytoxOrange for 30 min, followed by thorough washing with DMEM/F12. Gels were imaged with a Leica DMI6000B inverted fluorescent microscope (Leica Microsystems, Wetzlar, Germany) to 2159

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

determine the total number of live and dead cells. A minimum of six images were analyzed per gel. For each macromer weight fraction, three independent animal preparations (N = 3) were tested in triplicate. 2.10. Immunocytochemistry for Gels. At desired time points, cells were fixed for 20 min with 10% formalin and then washed three times in PBS. Following fixation, cells were permeabilized for 60 min with 0.50% triton X-100 in PBS, supplemented with 10% normal horse serum (NHS). The cells were then washed three times in PBS and incubated with primary antibodies (rabbit Iba1 1:1000 and mouse GFAP 1:1000) and 0.1% triton X-100 overnight at 4 °C. Primary antibodies were then aspirated and cells were rinsed three times in PBS, followed by incubation for 60 min with the secondary antibodies, Alexa Fluor 488 donkey anti-rabbit (1:200) and Alexa Fluor 647 donkey anti-mouse (1:200), at room temperature. Gels were inverted onto glass coverslips and imaged under a Leica DMI6000B inverted fluorescent microscope (Leica Microsystems, Wetzlar, Germany). 2.11. Statistical Analysis. Statistical analyses were performed using GraphPad Prism (GraphPad Software Inc., La Jolla, CA, U.S.A.). One-way ANOVA was conducted with Tukey post hoc tests and twoway ANOVA was performed in combination with Bonferroni tests. The group comparisons were made using a Mann-Whitney U test. All values are indicated as mean ± standard deviation (SD) of the mean. P-values were taken as an indicator of statistical significance as follows: *p < 0.05, **p < 0.01, and ***p < 0.0001, and where significant differences were present between two means in a figure, these were indicated using a line. N was used to represent independent experiments. Figure 3. Results of mechanical compression tests. (a) Stress−strain curve of HAMA gels after photo-cross-linking as a function of macromer weight fraction. Uniaxial compression testing was performed at a strain rate of 1.0 mm min−1 and presented as mean of three independent experiments, with straight lines representing linear regression analysis. (b) The elastic portion of the curve (to 10%) was used to calculate the elastic modulus for each macromer weight fraction and are presented as mean ± SD (N = 3). Gel moduli were compared to that of adult rat brain tissue tested under identical conditions. Three independent measurements were taken from one adult rat brain (N = 1). The black line represents the mean and the dotted black lines represent the SD.

3.0. RESULTS AND DISCUSSION 3.1. Establishing the Tunable Range of HAMA Modulus in Relation to CNS Tissue. Gels were formed by photo-cross-linking HAMA solutions with macromer weight fractions between 0.50% (w/v) and 1.50% (w/v). Increasing the macromer weight fraction upon cross-linking was expected to increase the elastic modulus of the hydrogel. Elastic modulus (E) of the HAMA gels with macromer weight fractions between 0.50−1.50% (w/v) was analyzed via uniaxial compressive testing. The stress−strain curves for the gels were similar to those of low-density and open-cell foams with a linear region from 0−10%.35 Elastic moduli were calculated from the slope of the stress−strain curve in the linear region (up to 10%) and increased with higher macromer weight fractions (Figure 3a). The calculated E values were as follows: 0.50% (w/v) samples had an E of 203 ± 54 Pa, the 0.75% (w/v) samples had an E of 290 ± 28 Pa, the 1.00% (w/v) had an E of 524 ± 108 Pa, and the 1.50% (w/v) had an E of 962 ± 148 Pa (Figure 3b). HAbased hydrogels have been considered by many researchers as cell scaffolds, and the mechancial properties of these gels have been tested previously both through rheometry24,26,27 and compression testing.25,28,35 The mechanical properties of these gels may vary depending on the degree of cross-linking, the molecular weight of HA used and how the gels were processed (e.g., lyophilization post cross-linking). Gels may also be chemically cross-linked or photo-cross-linked. In the case of photo-cross-linking, the degree of methacrylation, concentration of initiator used and the intensity and duration of LED exposure may alter the mechanical properties of the scaffold. 24,25,36 Although all of these factors make it challengining to directly compare the mechanical properties of the HAMA gels described in this paper with others reported in literature, the E values do fall within the reported range for related materials tested under similar conditions. For example, Her et al. measured the properties of a chemically cross-linked

HA scaffold (1.00% (w/v)) via uniaxial compression and found the E to be 780 ± 340 Pa.35 Elastic modulus values for the gels also fell within the range of reported modulus values for CNS tissue (rodent and porcine brain tissue; 0.1 to 1 kPa).37,38 As mechanical measurements are strongly dependent on testing conditions, the E of an adult rat brain was measured under identical testing parameters for comparison (results are shown in black in Figure 3). The E of the adult rat brain was found to be 273 ± 180 Pa, which covers the range of the 0.50% (w/v) and 0.75% (w/v) HAMA samples as well as part of the range of the 1.00% (w/v) sample (range shown with black dotted lines in Figure 3b). This demonstrates the potential of the HAMA scaffold to model soft biological tissues such as adult rat brain. 3.2. Investigating HAMA Gel Porosity and Structure at Various Macromer Weight Fractions. As elastic modulus increased with increasing HAMA macromer weight fraction, we expected to observe differences in gel microstructure for different HAMA macromer weight fractions. Scanning electron microscopy was utilized to image the pore structure of samples, which were frozen in their hydrated state to preserve the structure of the pores for imaging. Once the frozen samples were inserted into the cooled imaging chamber, water was sublimated directly out of freeze fractured pores, revealing the 2160

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

by a gel and is a function of the chemical potential (driving force for water uptake) and elastic retractive forces of gel crosslinks (restricting water adsorption). For cell scaffolds, a high Q value is desired as it encourages free diffusion of molecules, thereby enabling cells access to nutrients and signaling factors.39 A value of Q > 10 generally indicates a high degree of swelling and therefore a large amount of water and diffusion inside the gel.34 Equilibrium swelling ratio is often used to characterize gels and is related to many gel properties including mesh pore diameter,34,39,40 modulus,41 and porosity.42−45 An equilibrium gel swelling assay was performed to assess how Q changes with increasing macromer weight fraction. As HAMA macromer weight fraction was increased, Q decreased (Figure 5). This decreased swelling ratio indicates greater retractive

delicate structure and porosity of HAMA gels at macromer weight fractions between 0.50−1.50% (w/v). Lower magnification images demonstrate the irregular shape of pores formed during photo-cross-linking (Figure 4). This gives rise to a

Figure 5. Equilibrium swelling ratio (Q) after 24 h at macromer weight fractions between 0.50 to 1.50% (w/v). N = 3, values are mean ± SD.

forces in the 1.5% (w/v) gel, and correlates with the higher elastic modulus values obtained during mechanical testing. Since cryo-SEM results suggested that pore size does not vary significantly between samples, we hypothesize that pores with thicker walls are stiffer and imbibe less water at equilibrium (have greater retractive forces). Nonetheless, HAMA gels at all macromer weight fractions had Q values above 10 and were determined to be suitable for cell encapsulation in vitro. 3.4. Altered Cell Viability in Response to Cell Density. One variable known to affect cell viability in culture is cell density.46 Therefore, it was important to optimize this parameter for photoencapsulation of mixed glial cultures. Cells interact through chemical cell−cell signaling, which have been known to play a role in cell survival.46−49 Cell density also influences the supply of oxygen and nutrients in culture, which may also mitigate cell survival.46,49 In this study, seeding density was optimized to increase viability after photoencapsulation using a Syto13/SytoxOrange membrane permeability assay. Cells were encapsulated in a 0.50% (w/v) HAMA matrix at cell densities between 1 × 106 to 1 × 107 cells/mL. Distinct nuclei were observed for both live and dead cells. Cells that were out of plane appeared out of focus as the images were taken with an inverted fluorescent microscope and not a confocal microscope: however this did not interfere with cell counting. At day one post-encapsulation, cell viability for the lowest cell density (1 × 106 cells/mL) was 42.60 ± 7.67%, while the highest cell density (1 × 107 cells/mL) had a significantly higher viability of 80.20 ± 4.82% (Figure 6; p < 0.0001, 2-way ANOVA). Although some recovery of cell viability was seen at day 7 with 1 × 106 cells/mL cell density (42.60 ± 7.67% on day

Figure 4. HAMA gel structure investigated using cryo-SEM after photo-cross-linking as a function of macromer weight fraction (0.50 to 1.50% (w/v)). Lower magnification images were taken to compare pore size qualitatively and higher magnification images were used to assess changes in wall thickness between gels prepared with various macromer weight fractions.

variety of pore shapes and sizes, as has been seen in previous studies of hydrogel structure after cross-linking.31 Quantitative analysis of pore size performed using ImageJ (results not shown) found no significant difference in pore size with increasing macromer weight fraction. However, at higher magnifications, it became apparent that wall thickness of the HAMA pore walls increased with increasing macromer weight fraction. Walls of the 0.50% (w/v) samples were almost electron transparent, while walls of the 1.50% (w/v) samples appeared thicker. The increasing wall thickness observed with higher macromer weight fraction may contribute to the increasing modulus of elasticity observed as a function of macromer weight fraction during mechanical testing (Figure 3). 3.3. Equilibrium Swelling of HAMA Gels as a function of macromer weight fraction. Equilibrium swelling ratio (Q) is a measure of the amount of water that may be imbibed 2161

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

assess cells based on morphology due to the large number of cells in close proximity, and therefore higher densities were not used. All subsequent experiments were carried out with a cell density of 1 × 107 cells/mL, allowing for consistent cell viability analysis while still enabling cell morphology analysis. 3.5. Mixed Glial Cell Characterization in HAMA Gel. Next, the morphology of microglia and astrocytes within the HAMA gel was assessed after encapsulation. Microglia and astrocytes were labeled using immunofluorescence 7 days post encapsulation in a 0.50% (w/v) HAMA gel with a cell density of 1 × 107 cells/mL. Since morphology is commonly used to determine the state of cell activation, it was important to evaluate any changes cell encapsulation may have on glial cell morphology. Astrocytes (GFAP, red) exhibited many processes (Figure 7a) and were sometimes clustered in the gel (Figure

Figure 7. (Left) Differential interface contrast images; (Right) Astrocytes (red, GFAP) and microglia (green, Iba1). Cells cultured for 1 week in 0.50% (w/v) HAMA gel. Scale bar 20 μm. Figure 6. Live/dead (Syto13/SytoxOrange) membrane permeability cell viability assay of cell densities between 1 × 106 cells/mL and 1 × 107 cells/mL at 1 and 7 days post encapsulation. Differential interface contrast micrographs showing presence of processes are also shown. Quantification of cell viability was performed by counting. Representative images of N = 3 independent experiments 2-way ANOVA (***p < 0.0001, *p < 0.05). Values are mean ± SD and bars are used to show significance between groups. Scale bar 20 μm.

7b). Microglia often appeared rounded with a circular morphology as seen in Figure 7a, but were also seen with extended processes (Figure 7b), indicating a more ramified microglial morphology. Microglia also appeared to wrap processes around other cells in the gel, perhaps a sign of phagocytosis or cell−cell interaction.52 Processes of both astrocytes and microglia extended through multiple z planes of the gel indicating that the cells were interacting with the 3D environment of the HAMA scaffold as well as each other in a manner characteristic of in vivo cell behavior. 3.6. Effects of Increasing Macromer Weight Fraction on Cell Viability. Varying macromer weight percent was shown to modify the modulus of the matrix as well as alter gel structure. To evaluate how increasing macromer weight fraction may affect cell encapsulation, cell viability and processes extension was monitored over 7 days at macromer weight fractions between 0.50 and 1.50% (w/v). Cell viability was consistent for macromer concentrations between 0.50 and 1.00% (w/v) over 7 days (Figure 8), and cell processes were visible in each of these three gels by day 7. In contrast, cell processes were not visible in the 1.50% (w/v) gel. Cell viability was significantly lower in the 1.50% (w/v) gel on day 1 compared to the gels with lower macromer weight fractions (p < 0.0001). Some cell viability in the 1.50% (w/v) gel was recovered by day 7, but cell viability at this weight fraction did

1 to 48.40 ± 14.67% on day 7), cell viability did not recover to that of the 1 × 107 cells/mL cell density at day 7 (85.50 ± 5.05%). Cell viability at 5 × 106 cells/mL (82.17 ± 4.36%) was not significantly different from that of 1 × 107 cells/mL cell density at day 7 (p > 0.05), but was significantly lower at day 1 (69.60 ± 6.73% on day 1 vs 80.20 ± 4.82% on day 7, p < 0.05, 2-way ANOVA). The most consistent cell viability obtained over 1 week of culture was using a 1 × 107 cell/mL seeding density. Increasing cell viability at higher cell densities is likely due to the role of cell−cell interactions in cell survival.46−49 Other in vitro 3D culture studies of microglia and astrocytes reported similar cell densities.29,30 Glial cell density also varies with brain region, gender and age.50 The density of microglia and astrocytes reported in vivo higher than used in this study (∼3.4 × 108 cells/mL)51 and may be adjusted for future experiments to more accurately model the in vivo environment. Higher cell densities (1 × 108 cells/mL) made it difficult to 2162

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

tissue; therefore, 0.50% (w/v) was chosen to assess the utility of the HAMA gels as a model of glial scarring. In the future, the gel modulus may be tuned by varying the macromer weight fraction to model various regions of interest in the CNS. Platinum/iridium (Pt/Ir) polyimide-coated microwires with diameters of 30 μm were inserted into HAMA gels 3−4 days post-cell encapsulation. Great care was taken to insert wires parallel to glass coverslips and into the center of the gel without touching the coverslip to produce a floating microwire surrounded only by the scaffold. At 3, 6, 11, and 14 days gels were removed from culture plates and assessed with immunofluorescence. The HAMA gels supported suspended electrodes and no electrode shifting was observed during culture. The first cells to appear at the electrode interface were microglia (Figure 9). Microglia were first observed at day 3 after wire insertion and had a circular morphology characteristic of an activated state. Astrocytes were first observed lining the electrode in a manner similar to glial scarring at day 6 (Figure 9). By day 11 there was obvious astrocyte encapsulation of the microwire and this response was sustained over the 14 days of testing (Figure 9). These results follow closely the temporal

Figure 8. (Top) Live/dead (Syto13/SytoxOrange) membrane permeability and differential interface contrast images of cell viability assay for macromer concentrations between 0.5 and 1.5% (w/v) at 1, 3, and 7 days post encapsulation. Scale bar 20 μm. (Bottom) Quantification of cell viability at days 1 and 7 for seeding densities between 1 × 106 and 1 × 107 cells/mL. A two-way ANOVA, N = 3, ***p < 0.0001, values are mean ± SD, and bars are used to show significance between groups.

not reach the level of the other macromer weight fractions tested and remained significantly lower (p < 0.0001). Although HAMA gels became stiffer with increasing macromer weight percent, physical cues from increasing stiffness are an unlikely cause of increased cell death. In 2D cultures, cells are grown on substantially stiffer polystyrene culture plates without affecting cell viability. Changes in scaffold stiffness are correlated with process extension, cell differentiation, and varying cell expression, but not cell death.25,27,49,53−56 Instead, increased cell death may be due to the increased wall thickness of the gel observed during cryoSEM. This may limit the ability of glia to modify the gel structure and migrate through the gel as well as limit the transport of nutrients through the gel. 3.7. Assessing Cellular Response over 14 Days in HAMA Culture Model. The temporal response of microglia and astrocytes was monitored over 14 days in a 0.50% (w/v) HAMA gel. The modulus of elasticity of the 0.50 and 0.75% (w/v) hydrogels was well within the range measured for brain

Figure 9. Analysis of cellular reponse using 3D HAMA gel−cell constructs over 14 days. (Left) Differential interface contrast images showing electrode location in gels. Electrode is visible as a thick black line through the center of each image. (Right) Microglia (Iba1, green) and astrocytes (GFAP, red) were visualized using fluorescent microscopy and may be seen with relation to the electrode. Scale bar 100 μm. 2163

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

technology for performing CryoSEM work. The authors would also like to thank Rong Long for his helpful discussion regarding the compression testing of soft gels as well as Dr. Mirko Betti and Zied Khiari for assistance with mechanical characterization.

response of microglia and astrocytes in vivo after electrode insertion (microglia arriving first at the electrode, followed by astrocytes).3

4.0. CONCLUSION The overall goal of this study was to develop a 3D culture model of glial scarring capable of mimicking the elastic modulus of soft CNS tissue (such as brain tissue). In our work, methacrylated hyaluronic acid (HAMA) was investigated as a scaffold for mixed glial cultures and as the basis for an in vitro glial scarring model. We showed that the mechanical properties of this material could be varied by cross-linking the polymer in varying concentrations in PBS, and that these properties could be matched to those of rat brain. We determined an optimal density at which glial cells could be seeded in the gels, and monitored cell viability over 7 days. Finally, we used the model to observe the cellular response to a microwire inserted into the scaffold over 14 days. Changes in cell density and morphology at the electrode were similar to the progression of glial scarring observed in vivo. Due to the tunable mechanical properties of HAMA, electrode stability, and cellular response, this model shows great potential as a high-throughput system for preliminary electrode biocompatibility testing. Our current model does not contain other CNS cell types (neurons, oligodendrocytes or ependymal cells), nor does it model the breach of the blood−brain barrier during electrode insertion or the influence of micromotion. In the future, desired cell types (neurons), additional factors (blood components to simulate a breach of the blood brain barrier), and micromotion will be added to the model to investigate how these components influence the cellular response of the glial scarring model. The cellular response to various electrodes properties will also be compared (in isolation and in combination) to determine which significantly decrease glial scarring in a highly controlled and inexpensive manner prior to in vivo testing. This 3D model could therefore support and accelerate the development of biocompatible electrodes.

■

■

ASSOCIATED CONTENT

S Supporting Information *

A representative 1H NMR spectra of HAMA used to determine successful methacrylation and degree of modification. These spectra were also used to calculate the degree of methacrylation from the ratio of relative peak integrations of the methacrylate protons (peaks at ∼6.1 and ∼5.6 ppm) and methyl protons on HA (∼1.9 ppm). This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Jackson, A.; Zimmermann, J. B. Nat. Rev. Neurol. 2012, 8, 690− 699. (2) Bin, H.; Baird, R.; Butera, R.; Datta, A.; George, S.; Hecht, B.; Hero, A.; Lazzi, G.; Lee, R. C.; Liang, Jie; Neuman, M.; Peng, G. C. Y.; Perreault, E. J.; Ramasubramanian, M.; Wang, M. D.; Wikswo, J.; Guang-Zhong, Y.; Yuan-Ting, Z. IEEE Trans. Biomed. Eng. 2013, 60, 589−598. (3) He, W.; Bellamkonda, R. V. In Indwelling Neural Implants: Strategies for Contending with the In Vivo Environment; Reichert, W. M., Ed.; CRC Press: Boca Raton, FL, 2008. (4) Polikov, V. S.; Tresco, P. A.; Reichert, W. M. J. Neurosci. Methods 2005, 148, 1−18. (5) Moxon, K. A.; Hallman, S.; Sundarakrishnan, A.; Wheatley, M.; Nissanov, J.; Barbee, K. A. Materials 2009, 2, 1762−1794. (6) Szarowski, D. H.; Andersen, M. D.; Retterer, S.; Spence, A. J.; Isaacson, M.; Craighead, H. G.; Turner, J. N.; Shain, W. Brain Res. 2003, 983, 23−35. (7) Biran, R.; Martin, D. C.; Tresco, P. A. Exp. Neurol. 2005, 195, 115−126. (8) Griffith, R. W.; Humphrey, D. R. Neurosci. Lett. 2006, 406, 81− 86. (9) Thelin, J.; Jörntell, H.; Psouni, E.; Garwicz, M.; Schouenborg, J.; Danielsen, N.; Linsmeier, C. E. PloS One 2011, 6, e16267. (10) Harris, J. P.; Capadona, J. R.; Miller, R. H.; Healy, B. C.; Shanmuganathan, K.; Rowan, S. J.; Weder, C.; Tyler, D. J. J. Neural Eng. 2011, 8, 066011. (11) Hassler, C.; Guy, J.; Nietzschmann, M.; Staiger, J. F.; Stieglitz, T. IEEE 2011, 644−647. (12) Mercanzini, A.; Cheung, K.; Buhl, D.; Boers, M.; Maillard, A.; Colin, P.; Bensadoun, J.-C.; Bertsch, A.; Carleton, A.; Renaud, P. IEEE 2007, 573−576. (13) Biran, R.; Martin, D. C.; Tresco, P. A. J. Biomed. Mater. Res., Part A 2007, 82, 169−178. (14) Brain machine interfaces: implications for science, clinical practice and society (Progress in brain research); Schouenborg, J., Garwicz, M., Danielsen, N., Eds.; Elsevier: Amsterdam; Boston, 2011. (15) Kolarcik, C. L.; Bourbeau, D.; Azemi, E.; Rost, E.; Zhang, L.; Lagenaur, C. F.; Weber, D. J.; Cui, X. T. Acta Biomater. 2012, 8, 3561− 3575. (16) Zhong, Y.; Bellamkonda, R. V. Brain Res. 2007, 1148, 15−27. (17) Leung, B. K.; Biran, R.; Underwood, C. J.; Tresco, P. A. Biomaterials 2008, 29, 3289−3297. (18) Welkenhuysen, M.; Andrei, A.; Ameye, L.; Eberle, W.; Nuttin, B. IEEE Trans. Biomed. Eng. 2011, 58, 3250−3259. (19) Khaled, I.; Elmallah, S.; Cheng, C.; Moussa, W. A.; Mushahwar, V. K.; Elias, A. L. IEEE Trans. Biomed. Eng. 2013, 60, 2904−2913. (20) Kim, T.; Branner, A.; Gulati, T.; Giszter, S. F. J. Neural Eng. 2013, 10, 045001. (21) Polikov, V. S.; Block, M. L.; Fellous, J.-M.; Hong, J.-S.; Reichert, W. M. Biomaterials 2006, 27, 5368−5376. (22) Frampton, J. P.; Hynd, M. R.; Shuler, M. L.; Shain, W. Ann. Biomed. Eng. 2010, 38, 1031−1047. (23) Ereifej, E. S.; Cheng, M. M.-C.; Mao, G.; VandeVord, P. J. J. Neurosci. Methods 2013, 217, 17−25. (24) Hachet, E.; Van Den Berghe, H.; Bayma, E.; Block, M. R.; Auzély-Velty, R. Biomacromolecules 2012, 13, 1818−1827. (25) Seidlits, S. K.; Khaing, Z. Z.; Petersen, R. R.; Nickels, J. D.; Vanscoy, J. E.; Shear, J. B.; Schmidt, C. E. Biomaterials 2010, 31, 3930−3940. (26) Ananthanarayanan, B.; Kim, Y.; Kumar, S. Biomaterials 2011, 32, 7913−7923.

AUTHOR INFORMATION

Corresponding Author

*Tel.: +1-780-248-1589. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was funded by Alberta Innovates − Health Solutions (AIHS). A.F.J. was supported in part by an Alberta Innovates Technology Futures (AITF) scholarship. V.K.M. is an Alberta Heritage Foundation for Medical Research (AHFMR) Senior Scholar. Funding from the Canada Foundation for Innovation (CFI) is gratefully acknowledged. Gratitude is extended to Martin Kupsta and the NRC National Institute for Nano2164

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165

Biomacromolecules

Article

(27) Smeds, K. A.; Pfister-Serres, A.; Miki, D.; Dastgheib, K.; Inoue, M.; Hatchell, D. L.; Grinstaff, M. W. J. Biomed. Mater. Res. 2001, 54, 115−121. (28) Camci-Unal, G.; Cuttica, D.; Annabi, N.; Demarchi, D.; Khademhosseini, A. Biomacromolecules 2013, 14, 1085−1092. (29) East, E.; Golding, J. P.; Phillips, J. B. Tissue Eng., Part C 2012, 18, 526−536. (30) Frampton, J. P.; Hynd, M. R.; Shuler, M. L.; Shain, W. Biomed. Mater. 2011, 6, 015002. (31) Haycock, J. W. In 3D Cell Culture; Haycock, J. W., Ed.; Humana Press: Totowa, NJ, 2011; Vol. 695, pp 1−15. (32) Erickson, I. E.; Huang, A. H.; Sengupta, S.; Kestle, S.; Burdick, J. A.; Mauck, R. L. Osteoarthritis Cartilage 2009, 17, 1639−1648. (33) Lai, A. Y.; Todd, K. G. Glia 2006, 53, 809−816. (34) Bryant, S. J.; Anseth, K. S. In Scaffolding in Tissue Engineering; Ma, P. X., Elisseeff, J. H., Eds.; Taylor & Francis: Boca Raton, FL, 2005. (35) Her, G. J.; Wu, H.-C.; Chen, M.-H.; Chen, M.-Y.; Chang, S.-C.; Wang, T.-W. Acta Biomater. 2013, 9, 5170−5180. (36) Bencherif, S. A.; Srinivasan, A.; Horkay, F.; Hollinger, J. O.; Matyjaszewski, K.; Washburn, N. R. Biomaterials 2008, 29, 1739− 1749. (37) Pettikiriarachchi, J. T. S.; Parish, C. L.; Shoichet, M. S.; Forsythe, J. S.; Nisbet, D. R. Aust. J. Chem. 2010, 63, 1143. (38) Discher, D. E.; Mooney, D. J.; Zandstra, P. W. Science 2009, 324, 1673−1677. (39) Okay, O. In Hydrogel Sensors and Actuators; Gerlach, G., Arndt, K.-F., Eds.; Springer: Berlin, Heidelberg, 2009; Vol. 6, pp 1−14. (40) Chan, A. W.; Neufeld, R. J. Biomaterials 2009, 30, 6119−6129. (41) Bahney, C. S.; Lujan, T. J.; Hsu, C. W.; Bottlang, M.; West, J. L.; Johnstone, B. Eur. Cell. Mater. 2011, 22, 43−55 discussion 55.. (42) Da Silva, R.; Ganzarolli de Oliveira, M. Polymer 2007, 48, 4114− 4122. (43) Zhang, F.; He, C.; Cao, L.; Feng, W.; Wang, H.; Mo, X.; Wang, J. Int. J. Biol. Macromol. 2011, 48, 474−481. (44) LaNasa, S. M.; Hoffecker, I. T.; Bryant, S. J. J. Biomed. Mater. Res. B 2011, 96B, 294−302. (45) Smeds, K. A.; Pfister-Serres, A.; Miki, D.; Dastgheib, K.; Inoue, M.; Hatchell, D. L.; Grinstaff, M. W. J. Biomed. Mater. Res. 2001, 54, 115−121. (46) Yavin, E.; Billia, D. M. J. Neurosci. Res. 1997, 47, 471−478. (47) Van den Bos, C.; Silverstetter, S.; Murphy, M.; Connolly, T. Cell Tissue Res. 1998, 293, 463−470. (48) Stephan, S.; Johnson, W. E.; Roberts, S. Eur. Cell. Mater. 2011, 22, 97−108. (49) Morrison, B.; Cullen, D. K.; LaPlaca, M. In Neural Tissue Biomechanics; Bilston, L. E., Ed.; Springer: Berlin, Heidelberg, 2011; Vol. 3, pp 247−285. (50) Pelvig, D. P.; Pakkenberg, H.; Stark, A. K.; Pakkenberg, B. Neurobiol. Aging 2008, 29, 1754−1762. (51) Bjornsson, C. S.; Lin, G.; Al-Kofahi, Y.; Narayanaswamy, A.; Smith, K. L.; Shain, W.; Roysam, B. J. Neurosci. Methods 2008, 170, 165−178. (52) Cerbai, F.; Lana, D.; Nosi, D.; Petkova-Kirova, P.; Zecchi, S.; Brothers, H. M.; Wenk, G. L.; Giovannini, M. G. PLoS One 2012, 7, e45250. (53) De Guzman, R. C.; Ereifej, E. S.; Broadrick, K. M.; Rogers, R. A.; VandeVord, P. J. J. Microencapsulation 2008, 25, 487−498. (54) Mironi-Harpaz, I.; Wang, D. Y.; Venkatraman, S.; Seliktar, D. Acta Biomater. 2012, 8, 1838−1848. (55) Wang, X.; He, J.; Wang, Y.; Cui, F.-Z. Interface Focus 2012, 2, 278−291. (56) Engler, A. J.; Sen, S.; Sweeney, H. L.; Discher, D. E. Cell 2006, 126, 677−689.

2165

dx.doi.org/10.1021/bm500318d | Biomacromolecules 2014, 15, 2157−2165