In Vitro Cytotoxicity of Unsaturated Oligo[poly(ethylene glycol) fumarate]

Biomacromolecules 2003, 4, 552-560

552

In Vitro Cytotoxicity of Unsaturated Oligo[poly(ethylene glycol) fumarate] Macromers and Their Cross-Linked Hydrogels Heungsoo Shin,† Johnna S. Temenoff,† and Antonios G. Mikos* Department of Bioengineering, Rice University, MS-142, P.O. Box 1892, Houston, Texas 77251-1892 Received November 12, 2002; Revised Manuscript Received January 22, 2003

Currently, oligo[poly(ethylene glycol) fumarate] (OPF) hydrogels are being investigated as an injectable and biodegradable system for tissue engineering applications. In this study, cytotoxicity of each component of the OPF hydrogel formulation and the resulting cross-linked network was examined. Specifically, OPF synthesized with poly(ethylene glycol) (PEG) of different molecular weights (MW), the cross-linking agent [PEG-diacrylate (PEG-DA)], and the redox initiator pair [ammonium persulfate (APS) and ascorbic acid (AA)] were evaluated for cytotoxicity at 2 and 24 h using marrow stromal cells (MSCs) as model cells. The effect of leachable byproducts of OPF hydrogels on cytotoxicity was also investigated. Upon exposure to various concentrations of OPF for 2 h, greater than 50% of the MSCs were viable, regardless of OPF molecular weight or concentration in the media. After 24 h, the MSCs maintained more than 75% viability except for OPF concentrations higher than 25% (w/v). When examining the cross-linking agent, PEG-DA of higher MW (3400) demonstrated significantly higher viability compared to PEG-DA with MW 575 at all concentrations tested. Considering initiators, when MSCs were exposed to AA and APS, as well as the combination of AA and APS, higher viability was observed at lower concentrations. Once cross-linked, the leachable products from the OPF hydrogels had minimal adverse effects on the viability of MSCs (percentage of live cells was higher than 90% regardless of hydrogel types). The results suggest that, after optimization of cross-linking parameters, OPF-based hydrogels hold promise as novel injectable scaffolds or cell carriers in tissue engineering. Introduction Synthetic hydrogels have been widely explored for tissue engineering applications as they possess several advantages. One of these is high water content, which facilitates exchange of gases and nutrients in a biological environment.1 Additionally, the cross-linked network structure of hydrogels is similar to that of the extracellular matrix (ECM) and allows for maintenance of tissue-like elastic properties.2 Synthetic hydrogels have been studied as a substrate on which cell populations can attach and migrate,3 or they also have been utilized as a delivery vehicle for cells in conjunction with a drug to activate specific cellular functions in a localized region.4 Our laboratory is currently developing a biodegradable and injectable hydrogel system for tissue engineering, which can be applied to tissue defect in a minimally invasive way, aid in new tissue development, and degrade over time. We have recently synthesized a novel oligomer, oligo[poly(ethylene glycol) fumarate] (OPF), based on fumaric acid and poly(ethylene glycol) (PEG). The presence of multiple double bonds along the linear macromolecular chain allows for chemical cross-linking of an aqueous oligomer solution in the presence of a pair of redox initiators, ammonium persulfate (APS) and ascorbic acid (AA), and a cross-linking agent, PEG-diacrylate (PEG-DA).5 The formed hydrogels * To whom correspondence should be addressed. Tel: (713) 348-5355. Fax: (713) 348-4244. E-mail: [email protected]. † These authors have equally contributed to this work.

have been studied as biomimetic scaffolds for bone tissue engineering, especially after modification with cell adhesion peptides.6 OPF hydrogels have also been investigated as injectable cell carriers to aid the regeneration of articular cartilage in osteochondral defects.7 The evaluation of biocompatibility and cytotoxicity of these novel hydrogels is a critical step in further development of this material for tissue engineering applications. In particular, to examine potential use of synthetic hydrogels in injectable applications, several tests should be performed8 according to generalized guidelines.9 The cytotoxicity of hydrogel constituents including linear macromers, crosslinking agents, and initiators should be examined as these components may be cytotoxic to surrounding cells during the cross-linking reaction. Second, cytotoxicity of any substances leached from the cross-linked hydrogels (hydrogel sol fraction) should be determined since unreacted residues and byproducts from the radical polymerization reaction released from the hydrogels may prove detrimental to cell viability. Finally, for biodegradable hydrogels, the cytotoxicity of the degradation products should also be tested. The objective of this study was to evaluate the cytotoxicity of the constituents of an OPF hydrogel formulation and their cross-linked hydrogels. Specifically, OPF macromers synthesized with PEG of various molecular weights and PEGDA of two molecular weights, as well as the redox initiator system (APS and AA), were evaluated for cytotoxicity. Marrow stromal cells (MSCs) were isolated from the tibias and femurs of rats and used as a model cell type for the

10.1021/bm020121m CCC: $25.00 © 2003 American Chemical Society Published on Web 02/27/2003

Cytotoxicity of OPF Macromers and Hydrogels

study. MSCs can be differentiated into bone- and cartilageforming cells10-12 and be used for orthopedic tissue engineering.13,14 We also investigated the effect of leachable byproducts from the cross-linked hydrogels on the cytotoxicity of MSCs by exposing cells to conditioned media containing hydrogel extracts. This study asked three questions: (1) Do the macromolecular structures of OPF and PEG-DA affect cytotoxicity in vitro? (2) Does the concentration of each component of an injectable formulation affect MSC viability in vitro? (3) Does the sol fraction of the cross-linked hydrogels affect cell viability? Experimental Section Materials. Poly(ethylene glycol) (PEG) (nominal molecular weights 0.6K, 1.0K, 3.3K, 8.0K, and 10.0K), PEGdiacrylate (PEG-DA) (nominal molecular weight 575), ammonium persulfate (APS), and triethylamine (TEA) were purchased from Aldrich (Pittsburgh, PA). Fumaryl chloride was obtained from Acros (Milwaukee, WI) and distilled prior to use. Ascorbic acid (AA) and bovine serum albumin (BSA) were provided by Sigma (St. Louis, MO). PEG-DA (nominal molecular weight 3.4K) was purchased from Shearwater Polymers (Huntsville, AL). The LIVE/DEAD viability/ cytotoxicity assay kit was obtained from Molecular Probes (Eugene, OR). Dulbecco’s modified Eagle’s medium (DMEM) and trypsin-EDTA were purchased from Gibco Life (Grand Island, NY), and fetal bovine serum (FBS) was obtained from Gemini Bio-Products (Calabasas, CA). All solvents used in the study were of reagent grade. All reagents were used without further purification unless specified. Synthesis of Oligo(PEG fumarate) (OPF). Four formulations of OPF were synthesized from PEG nominal molecular weights of 1.0K, 3.3K, 8.0K, and 10.0K following a established procedure.5 Briefly, PEG was dehydrated by azeotropic distillation and dissolved in anhydrous tetrahydrofuran (or methylene chloride for the synthesis of OPF 10.0K). TEA and distilled fumaryl chloride were dropped concurrently into the dehydrated PEG solution that was placed in an ice bath. The reagents were mixed vigorously at room temperature overnight. Following solvent evaporation, the resulting product was recrystallized in ethyl acetate and precipitated in anhydrous ethyl ether. Remaining solvents were removed by drying at 0.1 Torr at least for 5 h. The number-average molecular weight of OPF was determined by gel permeation chromatography (GPC) based on a calibration curve obtained with monodispersed PEG standards. The purified OPF was stored at 0 °C prior to use. Marrow Stromal Cell (MSC) Culture. MSCs were isolated from 6-week-old male Wistar rats as previously described.13 Briefly, femurs and tibias were excised under aseptic conditions, and the marrow was flushed with primary media [DMEM supplemented with 10% (v/v) FBS, 1% (v/ v) antibiotics containing penicillin and streptomycin]. Following centrifugation for 8 min at 1250 rpm, the cell pellet was resuspended in 3 mL of primary media and seeded on T-75 culture flasks. Cell cultures were incubated at 37 °C, 95% relative humidity, and 5% CO2. Nonadherent cells were removed after 3 days, and fresh media were added. MSCs were cultured for 6 days prior to use in the study.

Biomacromolecules, Vol. 4, No. 3, 2003 553

Cytotoxicity of Oligomers and Cross-Linking Agents. The conditioned media were prepared by dissolving various concentrations of OPF, PEG, and PEG-DA in primary media. The concentration of OPF 1.0K conditioned media ranged from 40% (w/v) to 0.4% (w/v), and the concentration of the other types of OPF conditioned media varied from 25% (w/ v) to 0.25% (w/v). For PEG and PEG-DA conditioned media, the concentration ranged from 10% (w/v) to 0.1% (w/v). The prepared solutions were filtered using a cellulose acetate membrane filter (0.2 µm pore diameter) for sterilization. MSCs cultured on a T-75 culture flask were enzymatically lifted by an exposure to trypsin-EDTA solution (2 mL/ flask), replated on 96-well tissue culture plates in primary media at 40000 cells/cm2 (100 µL cell suspension/well), and allowed to attach for 24 h at 37 °C, 95% relative humidity, and 5% CO2. After 24 h, the media were aspirated, and the attached MSCs were exposed to 100 µL of the prepared conditioned media and cultured for either 2 or 24 h. The LIVE/DEAD reagent [combination of 4 µM ethidium homodimer-1 (EthD-1) and 2 µM calcein AM] was prepared following the protocol provided by the manufacturer. At each time point, the conditioned media were aspirated using a syringe, and the wells were rinsed with phosphate-buffered saline (PBS) solution twice to remove any remaining conditioned media. LIVE/DEAD reagent (100 µL) was added to each sample well and incubated in the dark for 30 min at room temperature. The resulting fluorescence of each well was measured using a fluorescent microplate reader (FLx800; BIO-TEK Instrument, Winooski, VT) equipped with 485/ 528 (excitation/emission) filter sets for calcein AM (detection of live cells) and 528/620 (excitation/emission) for EthD-1 (detection of dead cells). The wells in which MSCs were cultured without exposure to conditioned media served as a positive (live) control. For a negative (dead) control, MSCs were exposed to 70% (v/v) methanol solution for 30 min. The fluorescence of each sample well was normalized by that obtained from the positive and negative control groups, which was defined as the fractions of live and dead cells, respectively. The live/dead cell populations on the plates were visualized by fluorescence microscopy (Axiovert 135; Carl Zeiss, Thornwood, NY) equipped with a 35 mm camera (Nikon). Initiator Cytotoxicity. Solutions of AA and APS as well as combinations of AA/APS were prepared in primary media at 500 mM. The solutions were filtered using a cellulose acetate membrane filter (0.2 µm pore diameter) for sterilization. They were then diluted with primary media under sterile conditions to prepare 100 and 10 mM concentrations. For AA/APS combination solutions, a similar procedure was used to make sterile 1 M, 200 mM, and 20 mM solutions of each reagent (either AA or APS). A volume of 50 µL of each reagent was mixed in each well to obtain final concentrations of 500, 100, and 10 mM, respectively. Cell viability upon the exposure to initiator solutions was examined as described in the previous section, but in this case viability was determined only at 2 h. pH Change of Initiator Solutions. The initial pH for each solution (500, 100, and 10 mM) and the corresponding pH after the solutions were maintained for 2 h at 37 °C in an

554

Biomacromolecules, Vol. 4, No. 3, 2003

incubator were recorded using a electronic pH meter (AP5; Fisher Scientific). Primary media without any initiators maintained concurrently in the incubator served as a control for the study. The pH change for each solution was calculated using the equation:

Shin et al. Table 1. Number-Average Molecular Weight of Synthesized Oligomersa

∆pH ) pH(each sample solution) - pH(control) Cytotoxicity of Leachable Components from OPF Hydrogels. OPF hydrogels were prepared with PEG-DA (MW 575) in the presence of AA and APS initiators as previously described.6 OPF 1.0K, 3.3K, and 8.0K were used for the synthesis of hydrogels. Briefly, OPF and PEG-DA were dissolved in distilled deionized water (DDW) at 50% (w/v). The amount of added PEG-DA was determined by the ratio of the number of double bonds in PEG-DA to OPF (DBR) (1:1). For example, 0.041 mL of PEG-DA was added to 0.2 g of OPF 1.0K in 0.492 mL of DDW for the preparation of the OPF 1.0K hydrogel. The number of double bonds in OPF was calculated on the basis of the numberaverage molecular weight of synthesized OPF.6 The redox initiators, APS and AA, were added to this mixture at the equal concentration of 0.025 M. Subsequently, the OPF solution was cast between two glass plates (0.4 mm in thickness) and heated to 45 °C for 30 min to expedite crosslinking. The cross-linked films were lifted from the glass plate using a razor blade and punched into circular shapes with a cork borer (15 mm in diameter). These circular hydrogel films were subsequently dried overnight at room temperature and then sterilized by exposure to ultraviolet light for 3 h. After sterilization, each film was soaked in primary media and placed in an incubator at 37 °C, 95% relative humidity, and 5% CO2 for 24 h. The volume of media used to extract leachable products from OPF hydrogels was determined according to ISO/EN 10935 guidelines (e.g., 1 mL of media/3 cm2 specimen).15 Concurrently, MSCs were seeded in 96-well culture plates at 40000 cells/cm2 and cultured for 24 h as previously described. After being leached for 24 h, the OPF extract media were diluted by10 and 100 times, and attached MSCs were exposed to all three extract concentrations. The LIVE/DEAD cytotoxicity assay was performed after 2 and 24 h of exposure as previously described. Statistical Analysis. All experiments were conducted in triplicate. Single factor analysis of variance (ANOVA) and Tukey’s HSD (honestly significantly different) multiple comparison tests were used to determine statistical significance of results with a 95% confidence interval (R ) 0.05). The results are reported as means ( standard deviation. Results OPF Synthesis. The number-average molecular weight of PEG and the corresponding OPF is presented in Table 1. Cytotoxicity of Oligomers and Cross-Linking Agents. Upon exposure to OPF-containing media for 2 h, the fraction of live cells was higher than 50% regardless of OPF molecular weight (Figure 1). When MSCs were exposed to OPF 1.0K conditioned media, the cell viability increased from 0.62 ( 0.03 to 0.82 ( 0.05 for 40% (w/v) and 0.4%

MW of PEG

oligomer type

nominal

GPC

MW of oligomerb

OPF 1.0K OPF 3.3K OPF 8.0K OPF 10.0K

1000 3350 8000 10000

930 ( 10 2860 ( 30 6090 ( 90 9460 ( 30

4470 ( 60 11620 ( 500 14430 ( 380 14150 ( 360

a Data are presented as means ( standard deivation for n ) 3. Number-average molecular weight of oligomers was determined by gel permeation chromatography (GPC). b

(w/v) OPF 1.0K, respectively. The fraction of live cells for OPF 3.3K increased from 0.79 ( 0.11 for 25% (w/v) to 0.98 ( 0.0 for 0.25% (w/v). The cell viability to OPF 8.0K and 10.0K increased from 0.54 ( 0.04 for 25% (w/v) to 0.88 ( 0.02 for 0.25% (w/v) and from 0.67 ( 0.03 for 25% (w/v) to 0.84 ( 0.13 for 0.25% (w/v), respectively. While not significant between every concentration, there was a general trend of increasing cell viability with decreasing concentration of OPF in media for all OPF molecular weights. In addition, the fraction of live cells at the lowest concentration of OPF was significantly higher than that at the highest concentration of OPF, regardless of OPF type examined. No significant difference in the fraction of live cells was observed with differing the molecular weights of OPF at the same concentrations. MSCs that were exposed to the conditioned media for 24 h retained viability higher than 75% with the exception of the highest concentration of each OPF type. The fraction of live cells decreased from 0.87 ( 0.08 for 0.4% (w/v) to 0.11 ( 0.02 for 40% (w/v) for OPF 1.0K and from 0.75 ( 0.06 for 0.25% (w/v) to 0.16 ( 0.04 for 25% (w/v) for OPF 3.3K. The cell viability for OPF 8.0K and 10.0K was 0.84 ( 0.03 and 1.04 ( 0.07 for 0.25% (w/v) and decreased to 0.08 ( 0.01 and 0.14 ( 0.01 for 25% (w/v), respectively. The relative viability of MSCs was not significantly different between varying OPF concentrations except for the highest concentration of OPF, in which viability was significantly decreased for all OPF molecular weights (Figure 1c). The fluorescence uptake by nonviable cells was also measured at both time points (Figure 1b,d). The simple addition of the fraction of live and dead cells was approximately equal to 1 through all sample wells except the highest concentration of all OPF types at 24 h. After the viability assay, the appearance of MSCs on tissue culture well plates was visualized using fluorescence microscopy and presented in Figure 2. For the positive control, highly elongated morphology was observed (Figure 2a). MSCs treated with a methanol solution as the negative control remained attached to tissue culture plates and exhibited morphology similar to that as observed in the positive control (Figure 2b). The representative images of MSCs exposed to OPF are shown in Figure 2c,d. The viable cells exposed to OPF 10.0K for 24 h appeared well spread (Figure 2c), but nonviable cells had contracted and more round morphology (Figure 2d). The cytotoxicity of the cross-linking agent PEG-DA of two molecular weights, approximately 575 and 3400, with concentrations in media ranging from 0.1% to 10% (w/v)



Figure 1. Fraction of live (a, c) and dead (b, d) MSCs after exposure to various concentrations of OPF with different molecular weights for 2 h (a, b) and 24 h (c, d), normalized to positive control. Error bars stand for means ( standard deviation for n ) 3.

was also examined (Figure 3). The higher molecular weight of PEG-DA demonstrated significantly higher MSC viability upon exposure for either 2 or 24 h. When MSCs were exposed to PEG-DA-containing media for 2 h, the fraction of live cells for PEG-DA (MW 575) ranged from 0.18 ( 0.01 for 10% (w/v) to 0.51 ( 0.04 for 0.1% (w/v) while almost 100% cells were alive with PEG-DA (MW 3400) at all concentrations. For MSCs treated with PEG (MW 600) for 2 h, the fraction of live cells ranged from 0.79 ( 0.02 for 0.1% (w/v) to 0.84 ( 0.02 for 10% (w/v). For the higher molecular weight PEG (MW 3350), the fraction of live cells was close to 1.0 for all concentrations tested. At 24 h, the cell viability for PEG-DA (MW 575) showed reduced cell viability at all concentrations. However, cell viability for PEG-DA (MW 3400) was higher than 0.6 at all concentrations except at 10% (w/v). When MSCs were exposed to 10% (w/v) conditioned media, cell viability was 0.02 ( 0.01, 0.14 ( 0.01, 0.57 ( 0.13, and 1.04 ( 0.03 for PEG-DA (MW 575), PEG-DA (MW 3400), PEG (MW 600), and PEG (MW 3350), respectively. Fluorescence microscopy images supported the quantitative results. Extensive detachment and lysis of MSCs were observed in the presence of PEG-DA (MW 575) even at the lowest concentration (Figure 2e) while the morphology of MSCs treated with PEG (MW 600) was comparable to that from the control wells (Figure

2f) at each time point. MSCs exposed to higher molecular weight PEG-DA and PEG maintained elongated and polygonal morphology, which was similar to that in the positive control wells. Initiator Cytotoxicity. Cell viability data in the presence of AA and APS as well as combinations of AA/APS are shown in Table 2. The AA, APS, and AA/APS exhibited significantly higher viability at 10 mM than at the other concentrations. At 500 and 100 mM concentrations of each solution, less than 10% of the cells were viable after 2 h. Moreover, as we observed on the dead cell fraction for PEGDA, the fraction of dead cells to APS and AA/APS was low. For example, the sums of fractions of live cells and dead cells were only 0.01 ( 0.01 and 0.04 ( 0.01 for APS and AA/APS, respectively. The morphology of MSCs on each well plate was compared using fluorescence microscopy. A thin film covering MSCs was found on the bottom of the well plate at 100 mM and higher concentrations APS alone or AA/APS (pictures not shown). At 10 mM concentration of each reagent, APS (1.12 ( 0.08) and AA/APS (1.15 ( 0.17) showed cell viability comparable to the control group, while AA-exposed wells demonstrated significantly lower viability (0.47 ( 0.03). pH Change of Initiator Solutions. The pH of initiator solutions in media including AA, APS, and the AA/APS

556


Shin et al.

Figure 2. Representative morphologies of MSCs on 96-well tissue culture plates. Images represent (a) the positive control after 24 h, (b) the negative control after 24 h, (c) MSCs exposed to 10% (w/v) OPF 10.0K for 24 h (calcein AM filter), (d) MSCs exposed to 10% (w/v) OPF 10.0K for 24 h (EthD-1 filter), (e) MSCs exposed to PEG-DA (MW 575) for 2 h (calcein AM filter), and (f) MSCs exposed to PEG (MW 600) for 2 h (calcein AM filter). Scale bars represent 200 µm, and all of the figures are of the same magnification.

combination was recorded either immediately after preparation of samples or after 2 h in an incubator and compared to untreated media. In general, APS or AA alone showed a smaller pH change than the combination of APS/AA for each time point. In addition, a smaller change in pH was observed with decreasing concentrations of each solution. Cytotoxicity of Leachable Components from OPF Hydrogels. The fractions of live and dead MSCs after exposure to conditioned media containing hydrogel extracts at either 2 or 24 h are shown in Figure 4. The leachable products from the prepared hydrogels had minimal adverse effects on the viability of MSCs up to 24 h. At the time points tested, it was observed that the cell viability for the extract of OPF hydrogels without dilution was higher than 90% regardless of hydrogel type. The fraction of live cells was 1.00 ( 0.05, 0.97 ( 0.07, and 1.03 ( 0.08 when MSCs were treated with an extract of OPF 1.0K, 3.3K, and 8.0K hydrogels, respectively. Similar results were seen with varying dilutions of the extracts. Discussion The objective of this study was to examine the cytotoxicity of synthetic fumaric acid-based oligomers (OPF) and their

cross-linked hydrogels. Specifically, the effects of concentration and molecular weight of OPF hydrogel constituents on cytotoxicity of MSCs were evaluated. This study also sought to investigate the cytotoxicity of the sol fraction of crosslinked OPF hydrogels. The cytotoxicity of the OPF system was assessed using the LIVE/DEAD assay after modification of the protocol provided by the manufacturer. The evaluation of cytotoxicity is important in developing new biomaterials for both injectable and prefabricated tissue engineering applications. The modified cytotoxicity assay using a combination of fluorescent dyes allowed the quantitative examination of cellular behavior after contact with the components of the OPF hydrogel system. For most components tested, the fractions of live and dead cells could reproducibly compared since the simple addition of the two fractions was approximately 1. Also with these dyes, the morphology of viable and nonviable cells was clearly visible via microscopy. Overall, this assay provides an excellent method to study the effects of exposure time and concentration of viability for anchorage-dependent cells. Cytotoxicity of Oligomers and Cross-Linking Agents. Synthesized OPF consists of two major building blocks, PEG and fumaric acid. Fumaric acid is a naturally occurring substance in the Kreb’s cycle. PEG has been extensively



Figure 3. Fraction of live (a, c) and dead (b, d) MSCs after exposure to various concentrations of PEG-DA and PEG for 2 h (a, b) and 24 h (c, d). Error bars stand for means ( standard deviation for n ) 3. Table 2. pH Change and Corresponding Fraction of Live MSCs at 2 h for the Initiator Components APS, AA, and the Combination of APS and AA initiator ammonium persulfate (APS)

ascorbic acid (AA)

APS/AAc

concn (mM)

fraction of live cellsa

pH changeb

500 100 10 500 100 10 500 100 10

-0.08 ( 0.02 0.04 ( 0.01 1.12 ( 0.08 -0.08 ( 0.00 0.01 ( 0.01 0.47 ( 0.03 -0.13 ( 0.00 -0.11 ( 0.00 1.15 ( 0.17

-6.5 ( 0.1 -4.1 ( 1.3 -0.4 ( 0.1 -4.8 ( 0.0 -4.1 ( 0.1 -0.7 ( 0.0 -7.1 ( 0.1 -6.6 ( 0.1 -1.5 ( 0.2

a The fraction of live cells was calculated using a fluorescent microplate reader by relative fluorescence of each sample well (calcein AM stain) normalized by that of control wells prepared without initiators. b The pH change was calculated by subtracting the pH of the sample from the pH of the control media with no initiators added. c For AA/APS combination solutions, 1 M, 200 mM, and 20 mM solutions were prepared for each reagent (either AA or APS), and a volume of 50 µL of each reagent was mixed in each well to obtain final concentrations of 500, 100, and 10 mM, respectively.

explored for biomedical applications due to its relative biocompatibility and nonimmunogenicity.16,17 Although its constituents are considered biocompatible, it is also essential to assess cytotoxicity of the resulting oligomers. In the

cytotoxicity assay, approximately 60% of the cells remained alive after 2 h exposure to OPF for all molecular weights and concentrations tested. There was a general trend of increasing cell viability with decreasing concentration of OPF in media. The fraction of live cells treated with the lowest concentration of each oligomer was significantly larger than when exposed to the highest concentration of OPF. It should be noted that the concentration of OPF 1.0K containing media ranged from 40% (w/v) to 0.4% (w/v), while the conditioned media containing the other molecular weights of OPF were prepared at 25% (w/v) to 0.25% (w/v). The prepared OPF solution was sterilized by filtration using a cellulose membrane filter. However, because of increased solution viscosity, only the OPF 1.0K could be sterilized in this manner up to 40%. The relationship between cell viability and concentration of OPF as a function of culture time is particularly valuable for cell encapsulation because cross-linking monomers as well as initiators have a direct contact with MSCs for a certain period while the cross-linking reaction occurs. Our results indicate that, at the highest concentration of OPF, 60% of the MSCs remain viable after the cross-linking reaction. Furthermore, while the initial concentrations of linear OPF are similar to the highest concentrations examined

558


Shin et al.

Figure 4. Fraction of live (a, c) and dead (b, d) MSCs after exposure to various concentrations of hydrogel extracts for 2 h (a, b) and 24 h (c, d). Error bars stand for means ( standard deviation for n ) 3.

in this study, the concentration of OPF monomer will decrease within the order of minutes upon initiation of the cross-linking reaction. We have investigated physical properties of OPF hydrogels, and gelation onset in the presence of 0.025 M initiator occurred within 30 min for all molecular weights of OPF (data not shown). The cytotoxicity assay was conducted after 2 and 24 h following exposure of MSCs to OPF-containing media. Generally, the cytotoxicity of biomaterials has been reported after at least 24 h of exposure to cells. In this study, a 2 h time point was included to represent viability of MSCs present during the cross-linking reaction rather than seeded after network formation is complete. After 24 h, more than 80% of the cells were viable except for the highest concentration of each molecular weight OPF. The significant decrease in cell viability at the highest concentration of OPF after 24 h may not necessarily be due to intrinsic cytotoxicity of the macromer. Rather, the low cell viability may be a result of the high viscosity of the media, which could prevent effective exchange of nutrients and gases and, therefore, reduce cell viability at prolonged culture. Although not fully understood, there have been reports demonstrating that the toxicity of PEG may be dependent on the molecular weight.17,18 Since these studies examined

the toxicity of intravenously injected PEG, direct comparison with our study may not be possible. In our experiment, the variation in molecular weight of OPF did not affect viability of the plated MSCs at either time point, indicating that the molecular weight of PEG chains is not likely a critical factor for cell viability among synthesized macromers. To investigate the effect of molecular structure of the cross-linking agent, two molecular weights of PEG-DA and PEG were used for the study. The molecular weights of PEGDA were approximately 575 and 3400, and PEG of molecular weights 600 and 3350 were tested concurrently to decouple the effects of the presence of acrylate end groups vs PEG chain length on MSC viability. The results in Figure 5 show that there were significantly more viable cells for PEG-DA (MW 3400) than for PEG-DA (MW 575) with the same number of acrylate groups. From fluorescent microscopy images, the MSCs treated with PEG-DA (MW 575) appeared contracted and lacked cytoplasmic spaces; the considerable empty area in each field indicated that extensive detachment and lysis of the MSCs had occurred (Figure 2e). The fraction of dead cells supported these results (Figure 3). For PEGDA (MW 575), addition of the fractions of live and dead cells was much lower than 1, which suggests that, unlike the methanol-treated negative control, a large number of cells were detached from the culture plate and subsequently


Figure 5. Fraction of live MSCs after exposure to various concentrations of PEG-DA of different molecular weights for 2 and 24 h as a function of the concentration of acrylate groups in PEG-DA. Error bars stand for means ( standard deviation for n ) 3.

washed out during the rinsing or staining steps. Acrylic materials have been widely used in dentistry as restorative materials, liners, adhesives, and oral prosthetic devices.19,20 Early work has investigated the structure-cytotoxicity relationship of these materials in vitro as well as in vivo.21 Notably, it was demonstrated that acute cytotoxicity correlated with the hydrophilic/hydrophobic balance of acrylic and methacrylic compounds. Yoshii reported that the cytotoxicity of dimethacrylate with various lengths of ethylene oxide was significantly decreased when the molecular weight was higher than 1000.22 Although further experiments are needed to elucidate the exact correlation between PEG chain length and cytotoxicity of PEG-DA compounds, our results indicate that the longer chain PEG-DA molecule is less cytotoxic to MSCs than the shorter chain molecule. Initiator Cytotoxicity. The transformation of a liquid mixture of the macromer into a solid gel is initiated by the water-soluble redox initiator pairs, APS and AA. Therefore, solutions of each initiator component, as well as combinations of APS and AA, were prepared in order to investigate the effects of radicals and pH on cell viability. pH and cytotoxicity data for initiators at different concentrations showed that initiator solutions with low pH, regardless of the formation of radicals, resulted in reduced viability of MSCs. Therefore, there was a smaller change in pH and a corresponding higher viability at lower concentrations for all initiator components. For cell encapsulation, cells may react to each component alone as well as to radicals that are formed when they are combined. During the cross-linking reaction, the radicals may attack the cellular membrane.23 In addition, there is evidence that, when oxidized, AA produces a radical anion that decreases the pH of the solution during the reaction.15 Results from the dead cell assay (data not shown) presented the same trends in viability as observed in the live cell assay. Microscopic images of MSCs exposed to initiator solutions at high concentrations showed significant aggregation of cells, possibly due to denatured proteins at low pH. In addition, the combinations of APS and AA showed a pH drop at all concentrations tested, thus supporting previous observations of radical anion formation.15 Final consideration for designing OPF hydrogels is the gelation time, which is


dependent on the amount of initiators. Therefore, the 10 mM solutions that showed acceptable cell viability may delay the cross-linking reaction of OPF to times longer than those useful in a clinical setting. Thus, the relationship between the cross-linking time of OPF hydrogels and initiator concentrations should be further studied when using this system for cell encapsulation. Cytotoxicity of Leachable Components from OPF Hydrogels. Injectable materials are advantageous in that they can be applied to tissue defects with irregular shapes and form tight interfaces with surrounding tissues. However, it is hard to remove unreacted residues after injection and curing. Therefore, cytotoxicity of leachable products after cross-linking should be examined. Possible types of products in the extracts may include un-cross-linked OPF, PEG-DA, initiators, radicals, and short-chain macromers that were not incorporated into the cross-linked network. In these studies, the APS/AA was added at 0.025 M as this represents a minimal initiator concentration at which OPF underwent cross-linking at room temperature within 30 min or less (data not shown). Since the initiators and cross-linking agents showed a certain level of cytotoxicity to MSCs, it was anticipated that the residual leachable byproducts in the media could also have significant effects on cell viability. However, more than 80% of the MSCs remained alive after exposure to the media in the presence of extracts. These results indicate that the cross-linked OPF hydrogels and their sol fraction are relatively noncytotoxic although each constituent demonstrated various levels of cytotoxicity to MSCs. This can be explained since the acrylate double bond has higher reactivity relative to the fumarate double bond.24 Most of the PEG-DA may be incorporated into the cross-linked network. From the results in the present study, it is believed that free PEG-DA exists in the extract media at a minimal concentration so as to cause no major cytotoxic effects. In support of this explanation, our laboratory has recently characterized the cross-linked structure of OPF hydrogels using high-performance liquid chromatography (HPLC) and showed that the conversion of the acrylate double bond was 99%.25 Other remaining initiators or un-cross-linked molecules may be stably entrapped within the network or released slowly, which would significantly reduce their cytotoxicity. Conclusions In this study, we evaluated a chemically cross-linked OPF hydrogel system to determine the cytotoxic effects of each component on MSCs. Cell viability was examined after MSCs were exposed to various concentrations of OPF, the cross-linker PEG-DA, initiators, and hydrogel extracts. OPF had a minimal effect on cytotoxicity to MSCs whereas pH change and/or radical formation from the combination of initiators caused significant cell death. We observed an increase in cell viability for higher molecular weight PEGDA and a lower concentration of initiators (10 mM). After cross-linking, the leachable products from OPF hydrogels showed minimal cytotoxicity. These results show that through optimization of the initiator system and cross-linking

560


agents OPF-based hydrogels hold great promise for use as injectable biomimetic scaffolds and cell carriers. Acknowledgment. The authors gratefully acknowledge funding from the National Institutes of Health (Grant R01 DE13031) to A.G.M. and a Whitaker Foundation graduate fellowship to J.S.T. References and Notes (1) Nuttelman, C. R.; Mortisen, D. J.; Henry, S. M.; Anseth, K. S. J. Biomed. Mater. Res. 2001, 57, 217-223. (2) Lee, K. Y.; Mooney, D. J. Chem. ReV. 2001, 101, 1869-1879. (3) Hubbell, J. A. Bio/Technology 1995, 13, 565-576. (4) Murphy, W. L.; Mooney, D. J. J. Periodontal Res. 1999, 34, 413419. (5) Jo, S.; Shin, H.; Fisher, J. P.; Shung, A. K.; Mikos, A. G. Macromolecules 2001, 34, 2839-2844. (6) Shin, H.; Jo, S.; Mikos, A. G. J. Biomed. Mater. Res. 2002, 61, 169179. (7) Temenoff, J. S.; Athanasiou, K. A.; LeBaron, R. G.; Mikos, A. G. J. Biomed. Mater. Res. 2002, 59, 429-437. (8) Suggs, L. J.; Shive, M. S.; Garcia, C. A.; Anderson, J. M.; Mikos, A. G. J. Biomed. Mater. Res. 1998, 46, 22-32. (9) ISO/EN 10993-5. Biological evaluation of medical devices-part 5 tests for cytotoxicity, in vitro methods. (10) Pittenger, M. F.; Mackay, A. M.; Beck, S. C.; Jaiswal, R. K.; Douglas, R.; Mosca, J. D.; Moorman, M. A.; Simonetti, D. W.; Craig, S.; Marshak, D. R. Science 1999, 284, 143-147.

Shin et al. (11) Peter, S. J.; Liang, C. R.; Kim, D. J.; Widmer, M. S.; Mikos, A. G. J. Cell. Biochem. 1998, 71, 55-62. (12) Johnstone, B.; Hering, T. M.; Caplan, A. I.; Goldberg, V. M.; Yoo, J. U. Exp. Cell Res. 1999, 238, 265-272. (13) Peter, S. J.; Lu, L.; Kim, D. J.; Mikos, A. G. Biomaterials 2000, 21, 1207-1213. (14) Goldstein, A. S.; Juarez, T. M.; Helmke, C. D.; Gustin, M. C.; Mikos, A. G. Biomaterials 2001, 22, 1279-1288. (15) Masiakowski, J. T.; Lund, A.; Lindgren, M. J. Chem. Soc., Faraday Trans. 1987, 83, 893-903. (16) Lee, K. Y.; Alsberg, E.; Mooney, D. J. J. Biomed. Mater. Res. 2001, 56, 228-233. (17) Harris, J. M. Poly(ethylene glycol) chemistry; Plenum Press: New York, 1992; pp 1-14. (18) Herold, D. A.; Keil, K.; Bruns, D. E. Biochem. Pharmacol. 1989, 38, 73-76. (19) Theodore, E. J. Int. J. Prosthodontics 1989, 2, 163-172. (20) Doori, A.; Hugget, D. R.; Bates, J. F.; Brooks, S. C. Dent. Mater. 1991, 6, 25-32. (21) Lawrence, W. H.; Bass, G. E.; Purcell, W. P.; Autian, J. J. Dent. Res. 1972, 51, 526-535. (22) Yoshii, E. J. Biomed. Mater. Res. 1997, 37, 517-524. (23) Bryant, S. J.; Nuttelman, C. R.; Anseth, K. S. J. Biomater. Sci., Polym. Ed. 2000, 11, 439-457. (24) Otsu, T.; Matsumoto, A.; Shiraishi, K.; Amaya, N.; Koinuma, Y. J. Polym. Sci., Part A: Polym. Chem. 1992, 30, 1559-1565. (25) Timmer, M. D.; Jo, S.; Wang, C.; Ambrose, C. G.; Mikos, A. G. Macromolecules 2002, 35, 4373-4379.

BM020121M

In Vitro Cytotoxicity of Unsaturated Oligo[poly(ethylene glycol) fumarate]

Recommend Documents