- Email: [email protected]
The role of pore size on vascularization and tissue remodeling in PEG hydrogels
Biomaterials 32 (2011) 6045e6051
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
The role of pore size on vascularization and tissue remodeling in PEG hydrogels Yu-Chieh Chiu a, Ming-Huei Cheng b, Holger Engel c, Shu-Wei Kao b, Jeffery C. Larson a, Shreya Gupta a, Eric M. Brey a, d, * a
Department of Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, USA Division of Reconstructive Microsurgery Department of Plastic and, Reconstructive Surgery, Chang Gung Memorial Hospital, Taoyuan, Taiwan c Department of Hand-, Plastic-, and Reconstructive Surgery e Burn Center -,BG, Trauma Center Ludwigshafen, University of Heidelberg, German d Research Service, Hines Veterans Administration Hospital, Hines, IL, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 March 2011 Accepted 23 April 2011 Available online 12 June 2011
Vascularization is influenced by the physical architecture of a biomaterial. The relationship between pore size and vascularization has been examined for hydrophobic polymer foams, but there has been little research on tissue response in porous hydrogels. The goal of this study was to examine the role of pore size on vessel invasion in porous poly(ethylene glycol) (PEG) hydrogels. Vascularized tissue ingrowth was examined using three-dimensional cell culture and rodent models. In culture, all porous gels supported vascular invasion with the rate increasing with pore size. Following subfascia implantation, porous gels rapidly absorbed wound fluid, which promoted tissue ingrowth even in the absence of exogenous growth factors. Pore size influenced neovascularization, within the scaffolds and also the overall tissue response. Cell and vessel invasion into gels with pores 25e50 mm in size was limited to the external surface, while gels with pores larger pores (50e100 and 100e150 mm) permitted mature vascularized tissue formation throughout the entire material volume. A thin layer of inflammatory tissue was present at all PEG-tissue interfaces, effectively reducing the area available for tissue growth. These results show that porous PEG hydrogels can support extensive vascularized tissue formation, but the nature of the response depends on the pore size. Published by Elsevier Ltd.
Keywords: Porous Poly (ethylene glycol) Vascularization Hydrogels
1. Introduction Mass transfer limitations are a critical barrier to clinical application of engineered tissues [1e3]. In general, cells need to be within 100e200 mm of a blood vessel in order to receive oxygen and nutrients at levels sufficient to support proper function. [4]. Various strategies have been investigated for improving the vascularization of biomaterials used as tissues engineering scaffolds, including the delivery of therapeutic genes or proteins, encapsulation of endothelial cells within the scaffolds and prevascularization prior to implantation. Regardless of approach, it is likely that scaffold properties will need to be optimized to facilitate rapid vascularized tissue ingrowth within the material. Tissue response to an implanted material is influenced not only by its chemical and mechanical properties, but also its physical architecture. Studies using polymer membranes suggest that in some cases physical properties may be more important than * Corresponding author. Department of Biomedical Engineering, Illinois Institute of Technology, 3255 South Dearborn Street, Wishnick Hall Room 223, Chicago IL 60616. USA. Tel.: þ1 312 567 5098; fax: þ1 312 567 5707. E-mail address: [email protected] (E.M. Brey). 0142-9612/$ e see front matter Published by Elsevier Ltd. doi:10.1016/j.biomaterials.2011.04.066
chemical composition for maximizing vascularization. [5]. Vascular density near the surface of an implanted synthetic membrane was shown to be highest when membrane pore sizes were between 0.8 and 8.0 mm regardless of the material used [5]. Studies using porous chitosan scaffolds observed neovascularization in scaffolds with pore size of w90 mm but not w30 mm [6]. One recent study investigated the effect of pore sizes from 300 to 700 mm and interconnections of vascularization in porous b-tricalcium phosphate scaffolds. The results show that above a pore size 400 mm there is no difference for vascularization. [7,8]. While the relationship between pore size and tissue response has been examined for membranes and polymer foams, there is little information about how pore size may be used to regulate vascularized tissue response in hydrogels. Hydrophilic cross-linked polymer matrices, or hydrogels, have a variety of properties that make them popular as scaffolds for tissue engineering [9e11]. They can be synthesized with mechanical properties similar to soft tissues, and can be modified with peptides and proteins to mimic chemical composition of extracellular matrices (ECM). Poly(ethylene glycol) (PEG)-based hydrogels have been investigated for a number of tissue engineering applications including bone regeneration [12], cartilage repair [13], and vascularized tissue formation [14]. PEG chemistry provides
6046
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
relatively straightforward methods for the incorporation of proteins or ECM-based peptides that can be used to enhance vascularized tissue invasion into the hydrogel [15]. However, the polymer spacing in PEG hydrogels is much smaller than cells resulting in a requirement for degradation of polymer cross-links for cell migration or tissue invasion [16]. This is true of most polymer hydrogels. The spacing available is not consistent with ECM architecture and is orders of magnitue smaller than sizes that have been shown to be optimal for tissue response into other polymeric scaffolds (Fig. 1). PEG hydrogels with larger interconnected pores can be generated via polymerization of a salt saturated precursor solution containing particulate crystals of defined size. [8,20,21]. Particulate leaching of the crystals following polymerization provides the ability to control pore size within the hydrogels. The goal of this study was to investigate how pore size in these hydrogels influences tissue response and vascularization. This information could be used to help optimize scaffold design for applications in regenerative medicine. 2. Materials and methods 2.1. Materials PEG (Mn z 8000), hexane (reagent grade, 95%), acryloyl chloride (98%), PKH26GL, methyl cellulose (viscosity, 4000 cP), triethylamine (99.5%), and 2-hydroxy-
2-methylpropiophenone (Irgacure 1173) were obtained from Sigma (St. Louis, MO). Sodium chloride (99.5%), dichloromethane (99.9%), and ethyl ether (anhydrous) were from Fisher Scientific (Pittsburgh, PA). Fluorescein isothiocyanate (FITC)-conjugated bovine serum albumin (FITC-BSA) was purchased from Invitrogen (Pittsburgh, PA). 2.2. Synthesis of PEG diacrylate The methods to synthesize and purify PEG-diacrylate (PEG-DA) have been described previously [22]. Briefly, 10 g of PEG (Mn z 8000) were lyophilized overnight to remove any water. The lyophilized PEG was then placed into a three neck round bottom flask with 60 ml of anhydrous dichloromethane. Two mole of triethylamine (TEA) per mole of PEG was added to the flask and stirred for 5 min. Four moles of acryloyl chloride per mole of PEG were added dropwise to the flask and stirred for 24 h. The resulting products were washed with 5 ml of 2 M K2CO3 and then precipitated into 2 L of ice-cold ethyl ether to remove the residual acryloyl chloride. The structure and purity of the product was determined using 1H NMR (Advance 300 Hz; Bruker). To perform 1H NMR, the product was dissolved in CDCl3 with 0.05% (v/v) tetramethylsilane as an internal standard. The acrylation efficiency of PEG-DA was quantified based on 1H NMR. 2.3. RGD conjugation The methods to conjugate peptides to PEG was performed as described previously [20]. Briefly, a solution of 50 mM NaHCO3 (pH 8.3) was prepared as a buffer. Ten milligrams of YRGDS (American Peptide, Sunnyvale, CA) was dissolved in 5 mL of 50 mM NaHCO3. Acryl-PEG-SVA (3400 Da; Laysan, Arab, AL) was dissolved in 7 mL of 50 mM NaHCO3 and then added drop-wise into the stirred YRGDS solution in the dark. The molar ratio of YRGDS to acryl-PEG-SVA was 1:1.5. The solution was stirred for 2 h in the absence of light. The final product was dialyzed (2000 Da molecular weight cut-off) in 2 L of DI water for 24 h (with replacement after 12 h). The resulting product was lyophilized and stored at -80 C until use. 2.4. Synthesis of porous PEG hydrogels A particulate leaching technique was used to generate porous PEG hydrogels as described by Chiu et al. [20]. Hydrogels were generated with particulate crystals in ranges of 25e50, 50e100, and 100e150 mm. Salt crystals of defined size were selected using sieves (Precision E-forming LL, Cortland, NY). Ten mg/ml (based on pre-polymer conditions) of YRGDS was included in each hydrogel. Pore size was assessed using a selective partitioning of fluorescent proteins within the pore structure followed by confocal microscopy [20]. Entire hydrogels were incubated in a 0.5% (w/v) solution of FITC-BSA (rs ¼ 3.5 nm) overnight. [23]. The next day a center portion of the hydrogel was dissected out and imaged using a PASCAL laser scanning microscopy system from Carl Zeiss MicroImaging, Inc.(Thornwood,NY). The hydrogel was imaged using a 488 nm laser with a 505 nm low pass filter. Images had both x and y resolution of 3.5 mm/pixel and z resolution of 20 mm/pixel. Resulting serial images were exported to Axiovision 4.5 (Carl Zeiss,
Fig. 1. Three-dimensional reconstruction of confocal images of collagen fibers in dermis. The spacing between fibers is on the level of microns. The table provides a comparison of pore sizes in model ECM and optimal pore sizes determined previously for other tissue engineering applications to the the mesh size of PEG hydrogels.
Fig. 2. Porous hydrogel implanted in a pocket between the fascia and muscle created using dissection scissors.
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
6047
Fig. 3. Volume rendered confocal images of porous PEG hydrogels showing the interconnected porous structure. The hydrogels shown were generated with salt sizes ranging from (A) 150e100, (B) 100e50, and (C) 50e25 mm. The pores can be imaged based on the selective partitioning of FITC-labeled protein into the pores. The images were inverted so the hydrogels structure is white and pores black.
Göttingen, Germany) and volume rendered. For each sample, pore diameters were measured manually using Axiovision. Approximately 20 pores were examined per image. Pore size was defined as the longest axis of the pore. 2.5. Neovascularization in vitro Vessel invasion into the porous hydrogels was evaluated in vitro using a modification of an endothelial cell/smooth muscle cell co-culture model described previously [24]. Five thousand PKH26 stained human umbilical vein endothelial cells (HUVECs Lonza, Walkersville, MD) and 5,000 PKH26 stained human umbilical artery smooth muscle cells (HUASMCs, Lonza) were assembled into co-culture spheroids that were placed in the hydrogels. Briefly, 1 mL of 500,000 labeled HUVEC and 1 mL of 500,000 labeled HUASMC were placed in sterile 15 mL centrifuge tubes and 10 mL of endothelial cell growth medium added (EGM, Lonza). Three mL of 20% methyl cellulose in endothelial basal medium (EBM, Lonza) was added and mixed. One hundred fifty mL of the resulting cell solution was added to each well of 96-well, non-adherent, round bottom plates (Greiner bio-one, Monroe, NC), and incubated at 37 C overnight. The resultant cell spheroids were seeded in the porous gels and imaged at 0 days, 1 week, 2 weeks and 3 weeks. Both the HUVEC and the HUASMC used in this study were between passages 5 and 10. For each sample, the invasion area was measured manually using the PASCAL laser scanning microscopy system. The spheroid was imaged using a 488 nm laser with a 505 nm low pass filter. Images had both x and y resolution of 3.5 mm/pixel. Resulting images were exported to Axiovision 4.5 for quantification.
disk shapes with a height of 2 mm and a diameter of 0.5 cm. Sprague Dawley rats (weight ¼ 1958g, n ¼ 5 per time point) were anesthetized, their backs shaved and skin scrubbed with isopropyl alcohol followed by a povidone-iodine antiseptic solution. A longitudinal incision was made alone the spine and the skin separated using blunt dissection. The panniculus carnosus was carefully removed using forceps and blunt dissection, exposing the thin fascia covering the muscle body. The hydrogel was implanted in a pocket between the fascia and muscle created using dissection scissors (Fig. 2). The incision was closed by nylon suture. The rats were housed in individual cases and the implants harvested at 1, 2, and 3 weeks. Harvested tissues, which included the entire hydrogel and the surrounding muscle, were placed in 10% formaldehyde, paraffin embedded and serially sectioned (5 mm thickness) for histological staining. Sections were cut through the center of the sample containing cross sections of the gel through the underlying muscle [25]. 2.7. Histological staining and image analyses Specimens were stained for hematoxylin and eosin (H&E) and Masson’s trichome. Tissue sections were digitally imaged (0.17 mm/pixel, 20objective) using an Axiovert 200 inverted microscope. The percent collagen area was quantified from Masson’s trichrome stained sections by manually selecting the collagen (blue) area and dividing the collagen area by the total tissue area in the material. H&E stained
2.6. Animal model All animal procedures were approved by the institutional animal care and use committee (IACUC) at Chang Gung Memorial Hospital. Porous PEG-DA hydrogels of varying pore size with 10 mg/ml YRGDS were synthesized as described for the in vitro studies. Non-porous hydrogels with the same YRGDS concentration and PEG concentration were used as controls. In each condition, the hydrogels were cut into
Fig. 4. Invasion area of cocultured aggregates into porous PEG hydrogels generated with salt crystals ranging in size between 50e25, 100e50, and 150e100 mm (* indicates statistical difference between groups (p < 0.001)).
Fig. 5. Masson’s trichome stain of the tissue interface with a non-porous gel 3 weeks post-implantation. A thin layer of inflammatory tissue can be seen between the hydrogel (PEG) and underlying skeletal muscle (muscle) suggesting that these materials induce a low foreign body response (arrow). The non-porous PEG hydrogels are lost during processing due to the lack of tissue invasion.
6048
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
Fig. 6. Masson’s trichome stained of porous gels at 1 (A, D, G), 2 (B, E, H), and 3 (C, F, I) weeks post-implantation. Images are presented of gels with pores ranging from 50 to 25 mm (A, B, C), 100-50 mm (D, E, F) and 150-100 mm (G, H, I). Areas of vascularized (arrows) collagen (blue) can be observed in large (G, H, I) and medium (D, E, F) pore size gels at all time points.
tissue sections were also used to assess pore size of the PEG gels. The sections were digitally imaged (0.17 mm/pixel, 20objective) using an Axiovert 200 inverted microscope. Four images were taken per slide, and two slides were imaged per condition. All pores in a given image were quantified. Pore size was defined as the longest axis in the pore cross-section. 2.8. Statistical analysis Data are presented as means standard deviations. Significant differences between groups of data were determined by analysis of variance with HolmeSidak post-test. In all cases, p < 0.05 was considered statistically significant.
3. Result 3.1. Generation of porous PEG hydrogels The structure of the porous PEG hydrogels under fully swollen conditions can be visualized by exploiting the selective partitioning
of fluorescently-labeled protein into the bulk pores but not the cross-linked polymer network structure. [20]. The images reveal an interconnected porous structure within the hydrogels with the mean pore size varying with the range of salt crystals used (Fig. 3) Porous hydrogels generated with salts ranging from 150-100, 10050, and 50-25 mm sieved had mean pore sizes of134 28, 82 6, and 41 0.1 mm, respectively. 3.2. In vitro vessel invasion Cell invasion into the hydrogels was first examined using a 3D co-culture assay. There was no cellular invasion into non-porous hydrogels, but all pore conditions allowed extensive invasion into the hydrogels. At 1 week there were no statistical differences in invasion area between the three pore conditions (Fig. 4).However, the largest pore size (150e100 mm) gel allowed greater invasion area
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
6049
Fig. 7. H&E stained, porous gels at 1 (A, D, G), 2 (B, E, H), and 3 (C, F, I) weeks pot implantation. Images are presented of gels with pores ranging from 50 to 25 mm (A, B,C), 100-50 mm (D, E, F) and 150-100 mm (G, H, I).
(p < 0.001) than other sizes after two weeks. By week three, there was no difference between the two larger pore sizes, but the largest pore size gels (150e100 mm) had significantly greater invasion area in comparison to the smallest size (50-25 mm, p < 0.001). 3.3. In vivo response Surgically, both porous and non-porous hydrogels were easy to handle. Visually, the porous hydrogels could be observed rapidly absorbing the surrounding wound fluid after implantation within the pocket between the fascia and muscle. The non-porous hydrogels did not appear to absorb the wound fluid. All hydrogels (non-porous and porous) could be easily identified within tissues upon harvest at 1, 2, and 3 weeks following implantation. No macroscopic evidence of degradation or loss of polymer volume was observed. Histological sections of the hydrogels were stained with H&E and Masson’s trichrome to assess tissue response
and invasion. There was no histological indications of degradation at any time point for all gel conditions, and the pore size of the gels remained constant with time. There was no cell or tissue invasion into the non-porous gels. Instead, a thin layer of inflammatory tissue less than 15 mm in thickness was observed, consistent with a mild foreign body response at all time points (Fig. 5). No foreign body giant cells were observed in any of the groups. There was little cell invasion and no vessels observed in gels made with the smallest pore size at weeks 1 and 2 (Fig. 6(A),(B) and 7(A),(B)). By week 3 there was some remodeled collagen and vessels present in the pores adjacent to the surrounding tissue (Fig. 6(C)), but not deep within the hydrogel. Gels made with the two larger pore sizes, show significant cellular invasion with vessels present in the pores at week 1 (Fig. 7(D)e(I)). By week 2, vessels were observed in pores throughout the entire volume of gels with both the two larger pore sizes (Fig. 6). The vessels were contained within mature, remodeled collagen. Many of the vessels
6050
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
were filled with red blood cells. The density of new blood vessels in the pores appeared to be highest in the largest pore size. Interestingly, neither the newly remodeled collagen nor the blood vessels appear to ever contact the hydrogel surface (Fig. 8). Instead a thin layer of tissue appears adjacent to the walls of the pores. This region was highly cellular with some cells appearing spread on the hydrogel surface. The histological appearance was similar to the inflammatory tissue seen surrounding the non-porous gels. The extent of collagen remodeling within the porous PEG hydrogels was further examined by quantifying the percent of tissue area within the pores that contained collagen-rich tissue determined from the Masson’s trichrome stains (Fig.9). There were no statistical differences in percent collagen area at week 1 for all three pore sizes (Fig.9). In the second and third weeks, percent collagen area was significantly greater in the hydrogels with the largest pore size (150e100 mm) relative to the smallest (p < 0.001) There were no differences between the small (50e25 mm) and medium (100e50 mm) pore sizes. Throughout the time examined the percent of vascularized collagen formed within the porous gels was never greater than 30% of the total tissue area. The remaining tissue area appeared to be fibrin at 1 week, but at 2 and 3 weeks inflammatory tissue at the hydrogel interface made up the balance of the tissue area. 4. Discussion Hydrogels have received a significant amount of attention as tissue engineering scaffolds. The importance of pore size to vascularized tissue response has been studied in other material classes, but the importance of pores in hydrogels has not been extensively studied. PEG-based hydrogels provide an excellent model system for studying hydrogels because degradation of chemical crosslinks is required for cell and tissue invasion. Hydrogels formed from PEG allow controlled examination of the influence of pore size in the absence of degradation. Previously, we demonstrated a technique to generate porous PEG hydrogels [20]. In this study we examined the role of pore size on cell invasion and vessel assembly in both cell culture and animal models. A 3D cell culture model of sprouting angiogenesis was used to first examine invasion into the hydrogels in vitro. Models similar to this are often used as first-pass methods for evaluation of
Fig. 8. Representative Masson’s trichome stain of a porous gel generated with crystals ranging from 150 to 100 mm two weeks post-implantation. The mature collagen (blue, A) containing blood vessels (B) can be seen within the pores. A thin layer of inflammatory tissue (C) can be seen between the mature collagen and hydrogel (D). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. The fraction of tisssue within the pores of hydrogels that was mature collagen determined from Masson’s trichrome images. Collagen percentages were quantified as collagen area relative to total tissue area within the pores (* signifies statistical difference (p < 0.001)).
vascularization of tissue engineering scaffolds in vitro as they are thought to recapitulate many of the steps of neovascularization that occur in vitro [26]. The speed of sprout invasion in this model depended on pore size. While all sizes supported extensive vascular ingrowth, invasion increased with pore size. Interestingly, the in vivo vascularization results did not correlate with the cell culture model. After subfascia implantation, gels with 150e100 mm and 100e50 mm pores allow complete scaffold vascularization by two weeks, but the 50e25 mm pore size supported little or no vascularization at any time point. For the smallest pores, the in vitro model was not a good predictor of in vivo response. One primary difference between the cell culture and animal models was the cells actually interacting with the material. In the in vitro model the vascular cells come into direct contact with the material, while in vivo the vessels were never observed interacting with the material (Fig. 6). The cells interacting with the surface appeared to be inflammatory cells (Fig. 8). This suggests that peptides covalently incorporated within PEG may not directly interact with ECs following implantation, meaning they may not directly influence EC behavior. Instead, they may indirectly influence vascularization by mediating the behavior of inflammatory cells. These results are consistent with other PEG studies where peptide sequences alone have been found to be inadequate to stimulate vascularization at levels that are significantly higher than control levels [27,28]. In these studies, soluble factors that diffuse away from the material increased vascularization. Significant vascularization was observed in the porous gels in the absence of soluble factors. This appears to result from a rapid uptake of wound fluid into the hydrogels after implantation. A fibrin clot is present within the scaffolds at one week post-implantation that already had begun to vascularize. Hydrogels with pore sizes of 150e100 mm and 100e50 mm supported the formation of collagen rich, vascularized tissue within the pores. As described in the previous section, these vessels and the collagen rich regions do not appear to come into contact with the hydrogel surface. Instead there is a thin layer of inflammatory tissue adjacent to the pore surface. A recent study found that non-porous PEG hydrogels containing RGD elicited a moderate inflammatory response, with a 20e40 mm thick layer of inflammatory cells near the surfaces [29]. This fibrous capsule was consistent with a normal foreign body reaction (FBR) [29]. Our results with non-porous gels agreed with these findings. In addition, we found that a similar layer existed at all PEG-tissue
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
interfaces even within the porous hydrogels. This results in an “effective pore size” available for mature vascularized tissue formation that was less than the actual pore size. A previous study’s results also suggested that PEG hydrogels degraded in vivo possible due to oxidative degradation. The porous and non-porous PEG hydrogels were very stable in vivo. While degradation may occur in some highly oxidative environments, the moderate inflammatory response in our case did not appear sufficient to degrade the gels over the time frame we examined. We did not observe any signs of material degradation histologically and were able to recover all hydrogels implanted. The ideal tissue engineered scaffold will need to be biodegradable. In future studies we will examine the effects of controlled degradation of porous PEG-based hydrogels on tissue response and vessel persistence. 5. Conclusions In this study, we examined the role of pore size on neovascularization and tissue invasion in PEG hydrogels using both 3D cell culture and small animal models. The result from both in vitro and in vivo studies show that pore size mediates neovascularization. Interestingly, the results from the in vitro model were not consistent with in vivo, possibly due to different cells mediating interactions with the material. This study provides important insight into the role of physical structure on tissue response to porous biological scaffolds. Acknowledgements This research was supported by funding from Veterans Administration, the National Science Foundation (Grants 0852048,0731201) and the Taiwan National Science Council (982314-B-182-015-MY3). We thank Sarah Lopez for assistance with analysis of collagen content in the hydrogels. References [1] McIntire LV. Vascular assembly in engineered and natural tissues. Ann N Y Acad Sci 2002;961:246e8. [2] Chiu YC, Cheng MH, Uriel S, Brey EM. Materials for engineering vascularized adipose tissue. J Tissue Viability 2011;20:37e48. [3] Papavasiliou G, Cheng MH, Brey EM. Strategies for vascularization of polymer scaffolds. J Investig Med 2010;58:838e44. [4] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407:249e57. [5] Brauker JH, Carr-Brendel VE, Martinson LA, Crudele J, Johnston WD, Johnson RC. Neovascularization of synthetic membranes directed by membrane microarchitecture. J Biomed Mater Res 1995;29:1517e24. [6] Lim TC, Bang CP, Chian KS, Leong KF. Development of cryogenic prototyping for tissue engineering. Virtual Phys Prototyping 2008;3:25e31. [7] Bai F, Wang Z, Lu JX, Liu JA, Chen GY, Lv R, et al. The correlation between the internal structure and vascularization of controllable porous bioceramic materials In vivo: a quantitative study. Tissue Eng Part A 2010;16:3791e803.
6051
[8] Chen CW, Betz MW, Fisher JP, Paek A, Chen Y. Macroporous hydrogel scaffolds and their characterization by optical coherence tomography. Tissue Eng Part C Methods 2010;17:101e12. [9] Van de Wetering P, Metters AT, Schoenmakers RG, Hubbell JA. Poly(ethylene glycol) hydrogels formed by conjugate addition with controllable swelling, degradation, and release of pharmaceutically active proteins. J Control Release 2005;102:619e27. [10] Aimetti AA, Machen AJ, Anseth KS. Poly(ethylene glycol) hydrogels formed by thiol-ene photopolymerization for enzyme-responsive protein delivery. Biomaterials 2009;30:6048e54. [11] Hahn MS, Taite LJ, Moon JJ, Rowland MC, Ruffino KA, West JL. Photolithographic patterning of polyethylene glycol hydrogels. Biomaterials 2006;27: 2519e24. [12] Kim J, Lee KW, Hefferan TE, Currier BL, Yaszemski MJ, Lu L. Synthesis and evaluation of novel biodegradable hydrogels based on poly(ethylene glycol) and sebacic acid as tissue engineering scaffolds. Biomacromolecules 2008;9: 149e57. [13] Sontjens SH, Nettles DL, Carnahan MA, Setton LA, Grinstaff MW. Biodendrimer-based hydrogel scaffolds for cartilage tissue repair. Biomacromolecules 2006;7:310e6. [14] Moon JJ, Lee SH, West JL. Synthetic biomimetic hydrogels incorporated with ephrin-A1 for therapeutic angiogenesis. Biomacromolecules 2007;8: 42e9. [15] Lin CC, Anseth KS. Controlling affinity binding with peptide-functionalized poly(ethylene glycol) hydrogels. Adv Funct Mater 2009;19:2325. [16] Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Celldemanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth. FASEB J 2003;17:2260e2. [17] Francis ME, Uriel S, Brey EM. Endothelial cell-matrix interactions in neovascularization. Tissue Eng Part B Rev 2008;14:19e32. [18] Yannas IV, Lee E, Orgill DP, Skrabut EM, Murphy GF. Synthesis and characterization of a model extracellular matrix that induces partial regeneration of adult mammalian skin. Proc Natl Acad Sci U S A 1989;86:933e7. [19] Canal T, Peppas NA. Correlation between mesh size and equilibrium degree of swelling of polymeric networks. J Biomed Mater Res 1989;23: 1183e93. [20] Chiu YC, Larson JC, Isom A, Brey EM. Generation of porous poly(ethylene glycol) hydrogels by salt leaching. Tissue Eng Part C-Methods 2010;16: 905e12. [21] Brey EM, Appel A, Chiu YC, Zhong Z, Cheng MH, Engel H, et al. X-Ray i maging of poly(ethylene glycol) hydrogels without contrast agents. Tissue Eng Part CMethods 2010;16:1597e600. [22] Chiu YC, Larson JC, Perez-Luna VH, Brey EA. Formation of microchannels in poly(ethylene glycol) hydrogels by selective degradation of patterned microstructures. Chem Mater 2009;21:1677e82. [23] White JA, Deen WM. Agarose-dextran gels as synthetic analogs of glomerular basement membrane: water permeability. Biophysical J 2002;82:2081e9. [24] Moya ML, Cheng MH, Huang JJ, Francis-Sedlak ME, Kao SW, Opara EC, et al. The effect of FGF-1 loaded alginate microbeads on neovascularization and adipogenesis in a vascular pedicle model of adipose tissue engineering. Biomaterials 2010;31:2816e26. [25] Brey EM, King TW, Johnston C, McIntire LV, Reece GP, Patrick CW. A technique for quantitative three-dimensional analysis of microvascular structure. Microvasc Res 2002;63:279e94. [26] Brey EM, Uriel S, Greisler HP, McIntire LV. Therapeutic neovascularization: contributions from bioengineering. Tissue Eng 2005;11:567e84. [27] Seliktar D, Zisch AH, Lutolf MP, Wrana JL, Hubbell JA. MMP-2 sensitive, VEGFbearing bioactive hydrogels for promotion of vascular healing. J Biomed Mater Res Part A 2004;68A:704e16. [28] Phelps EA, Landazuri N, Thule PM, Taylor WR, Garcia AJ. Bioartificial matrices for therapeutic vascularization. Proc Natl Acad Sci U S A 2010;107:3323e8. [29] Lynn AD, Kyriakides TR, Bryant SJ. Characterization of the in vitro macrophage response and in vivo host response to poly(ethylene glycol)-based hydrogels. J Biomed Mater Res Part A 2010;93:941e53.
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
The role of pore size on vascularization and tissue remodeling in PEG hydrogels Yu-Chieh Chiu a, Ming-Huei Cheng b, Holger Engel c, Shu-Wei Kao b, Jeffery C. Larson a, Shreya Gupta a, Eric M. Brey a, d, * a
Department of Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, USA Division of Reconstructive Microsurgery Department of Plastic and, Reconstructive Surgery, Chang Gung Memorial Hospital, Taoyuan, Taiwan c Department of Hand-, Plastic-, and Reconstructive Surgery e Burn Center -,BG, Trauma Center Ludwigshafen, University of Heidelberg, German d Research Service, Hines Veterans Administration Hospital, Hines, IL, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 March 2011 Accepted 23 April 2011 Available online 12 June 2011
Vascularization is influenced by the physical architecture of a biomaterial. The relationship between pore size and vascularization has been examined for hydrophobic polymer foams, but there has been little research on tissue response in porous hydrogels. The goal of this study was to examine the role of pore size on vessel invasion in porous poly(ethylene glycol) (PEG) hydrogels. Vascularized tissue ingrowth was examined using three-dimensional cell culture and rodent models. In culture, all porous gels supported vascular invasion with the rate increasing with pore size. Following subfascia implantation, porous gels rapidly absorbed wound fluid, which promoted tissue ingrowth even in the absence of exogenous growth factors. Pore size influenced neovascularization, within the scaffolds and also the overall tissue response. Cell and vessel invasion into gels with pores 25e50 mm in size was limited to the external surface, while gels with pores larger pores (50e100 and 100e150 mm) permitted mature vascularized tissue formation throughout the entire material volume. A thin layer of inflammatory tissue was present at all PEG-tissue interfaces, effectively reducing the area available for tissue growth. These results show that porous PEG hydrogels can support extensive vascularized tissue formation, but the nature of the response depends on the pore size. Published by Elsevier Ltd.
Keywords: Porous Poly (ethylene glycol) Vascularization Hydrogels
1. Introduction Mass transfer limitations are a critical barrier to clinical application of engineered tissues [1e3]. In general, cells need to be within 100e200 mm of a blood vessel in order to receive oxygen and nutrients at levels sufficient to support proper function. [4]. Various strategies have been investigated for improving the vascularization of biomaterials used as tissues engineering scaffolds, including the delivery of therapeutic genes or proteins, encapsulation of endothelial cells within the scaffolds and prevascularization prior to implantation. Regardless of approach, it is likely that scaffold properties will need to be optimized to facilitate rapid vascularized tissue ingrowth within the material. Tissue response to an implanted material is influenced not only by its chemical and mechanical properties, but also its physical architecture. Studies using polymer membranes suggest that in some cases physical properties may be more important than * Corresponding author. Department of Biomedical Engineering, Illinois Institute of Technology, 3255 South Dearborn Street, Wishnick Hall Room 223, Chicago IL 60616. USA. Tel.: þ1 312 567 5098; fax: þ1 312 567 5707. E-mail address: [email protected] (E.M. Brey). 0142-9612/$ e see front matter Published by Elsevier Ltd. doi:10.1016/j.biomaterials.2011.04.066
chemical composition for maximizing vascularization. [5]. Vascular density near the surface of an implanted synthetic membrane was shown to be highest when membrane pore sizes were between 0.8 and 8.0 mm regardless of the material used [5]. Studies using porous chitosan scaffolds observed neovascularization in scaffolds with pore size of w90 mm but not w30 mm [6]. One recent study investigated the effect of pore sizes from 300 to 700 mm and interconnections of vascularization in porous b-tricalcium phosphate scaffolds. The results show that above a pore size 400 mm there is no difference for vascularization. [7,8]. While the relationship between pore size and tissue response has been examined for membranes and polymer foams, there is little information about how pore size may be used to regulate vascularized tissue response in hydrogels. Hydrophilic cross-linked polymer matrices, or hydrogels, have a variety of properties that make them popular as scaffolds for tissue engineering [9e11]. They can be synthesized with mechanical properties similar to soft tissues, and can be modified with peptides and proteins to mimic chemical composition of extracellular matrices (ECM). Poly(ethylene glycol) (PEG)-based hydrogels have been investigated for a number of tissue engineering applications including bone regeneration [12], cartilage repair [13], and vascularized tissue formation [14]. PEG chemistry provides
6046
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
relatively straightforward methods for the incorporation of proteins or ECM-based peptides that can be used to enhance vascularized tissue invasion into the hydrogel [15]. However, the polymer spacing in PEG hydrogels is much smaller than cells resulting in a requirement for degradation of polymer cross-links for cell migration or tissue invasion [16]. This is true of most polymer hydrogels. The spacing available is not consistent with ECM architecture and is orders of magnitue smaller than sizes that have been shown to be optimal for tissue response into other polymeric scaffolds (Fig. 1). PEG hydrogels with larger interconnected pores can be generated via polymerization of a salt saturated precursor solution containing particulate crystals of defined size. [8,20,21]. Particulate leaching of the crystals following polymerization provides the ability to control pore size within the hydrogels. The goal of this study was to investigate how pore size in these hydrogels influences tissue response and vascularization. This information could be used to help optimize scaffold design for applications in regenerative medicine. 2. Materials and methods 2.1. Materials PEG (Mn z 8000), hexane (reagent grade, 95%), acryloyl chloride (98%), PKH26GL, methyl cellulose (viscosity, 4000 cP), triethylamine (99.5%), and 2-hydroxy-
2-methylpropiophenone (Irgacure 1173) were obtained from Sigma (St. Louis, MO). Sodium chloride (99.5%), dichloromethane (99.9%), and ethyl ether (anhydrous) were from Fisher Scientific (Pittsburgh, PA). Fluorescein isothiocyanate (FITC)-conjugated bovine serum albumin (FITC-BSA) was purchased from Invitrogen (Pittsburgh, PA). 2.2. Synthesis of PEG diacrylate The methods to synthesize and purify PEG-diacrylate (PEG-DA) have been described previously [22]. Briefly, 10 g of PEG (Mn z 8000) were lyophilized overnight to remove any water. The lyophilized PEG was then placed into a three neck round bottom flask with 60 ml of anhydrous dichloromethane. Two mole of triethylamine (TEA) per mole of PEG was added to the flask and stirred for 5 min. Four moles of acryloyl chloride per mole of PEG were added dropwise to the flask and stirred for 24 h. The resulting products were washed with 5 ml of 2 M K2CO3 and then precipitated into 2 L of ice-cold ethyl ether to remove the residual acryloyl chloride. The structure and purity of the product was determined using 1H NMR (Advance 300 Hz; Bruker). To perform 1H NMR, the product was dissolved in CDCl3 with 0.05% (v/v) tetramethylsilane as an internal standard. The acrylation efficiency of PEG-DA was quantified based on 1H NMR. 2.3. RGD conjugation The methods to conjugate peptides to PEG was performed as described previously [20]. Briefly, a solution of 50 mM NaHCO3 (pH 8.3) was prepared as a buffer. Ten milligrams of YRGDS (American Peptide, Sunnyvale, CA) was dissolved in 5 mL of 50 mM NaHCO3. Acryl-PEG-SVA (3400 Da; Laysan, Arab, AL) was dissolved in 7 mL of 50 mM NaHCO3 and then added drop-wise into the stirred YRGDS solution in the dark. The molar ratio of YRGDS to acryl-PEG-SVA was 1:1.5. The solution was stirred for 2 h in the absence of light. The final product was dialyzed (2000 Da molecular weight cut-off) in 2 L of DI water for 24 h (with replacement after 12 h). The resulting product was lyophilized and stored at -80 C until use. 2.4. Synthesis of porous PEG hydrogels A particulate leaching technique was used to generate porous PEG hydrogels as described by Chiu et al. [20]. Hydrogels were generated with particulate crystals in ranges of 25e50, 50e100, and 100e150 mm. Salt crystals of defined size were selected using sieves (Precision E-forming LL, Cortland, NY). Ten mg/ml (based on pre-polymer conditions) of YRGDS was included in each hydrogel. Pore size was assessed using a selective partitioning of fluorescent proteins within the pore structure followed by confocal microscopy [20]. Entire hydrogels were incubated in a 0.5% (w/v) solution of FITC-BSA (rs ¼ 3.5 nm) overnight. [23]. The next day a center portion of the hydrogel was dissected out and imaged using a PASCAL laser scanning microscopy system from Carl Zeiss MicroImaging, Inc.(Thornwood,NY). The hydrogel was imaged using a 488 nm laser with a 505 nm low pass filter. Images had both x and y resolution of 3.5 mm/pixel and z resolution of 20 mm/pixel. Resulting serial images were exported to Axiovision 4.5 (Carl Zeiss,
Fig. 1. Three-dimensional reconstruction of confocal images of collagen fibers in dermis. The spacing between fibers is on the level of microns. The table provides a comparison of pore sizes in model ECM and optimal pore sizes determined previously for other tissue engineering applications to the the mesh size of PEG hydrogels.
Fig. 2. Porous hydrogel implanted in a pocket between the fascia and muscle created using dissection scissors.
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
6047
Fig. 3. Volume rendered confocal images of porous PEG hydrogels showing the interconnected porous structure. The hydrogels shown were generated with salt sizes ranging from (A) 150e100, (B) 100e50, and (C) 50e25 mm. The pores can be imaged based on the selective partitioning of FITC-labeled protein into the pores. The images were inverted so the hydrogels structure is white and pores black.
Göttingen, Germany) and volume rendered. For each sample, pore diameters were measured manually using Axiovision. Approximately 20 pores were examined per image. Pore size was defined as the longest axis of the pore. 2.5. Neovascularization in vitro Vessel invasion into the porous hydrogels was evaluated in vitro using a modification of an endothelial cell/smooth muscle cell co-culture model described previously [24]. Five thousand PKH26 stained human umbilical vein endothelial cells (HUVECs Lonza, Walkersville, MD) and 5,000 PKH26 stained human umbilical artery smooth muscle cells (HUASMCs, Lonza) were assembled into co-culture spheroids that were placed in the hydrogels. Briefly, 1 mL of 500,000 labeled HUVEC and 1 mL of 500,000 labeled HUASMC were placed in sterile 15 mL centrifuge tubes and 10 mL of endothelial cell growth medium added (EGM, Lonza). Three mL of 20% methyl cellulose in endothelial basal medium (EBM, Lonza) was added and mixed. One hundred fifty mL of the resulting cell solution was added to each well of 96-well, non-adherent, round bottom plates (Greiner bio-one, Monroe, NC), and incubated at 37 C overnight. The resultant cell spheroids were seeded in the porous gels and imaged at 0 days, 1 week, 2 weeks and 3 weeks. Both the HUVEC and the HUASMC used in this study were between passages 5 and 10. For each sample, the invasion area was measured manually using the PASCAL laser scanning microscopy system. The spheroid was imaged using a 488 nm laser with a 505 nm low pass filter. Images had both x and y resolution of 3.5 mm/pixel. Resulting images were exported to Axiovision 4.5 for quantification.
disk shapes with a height of 2 mm and a diameter of 0.5 cm. Sprague Dawley rats (weight ¼ 1958g, n ¼ 5 per time point) were anesthetized, their backs shaved and skin scrubbed with isopropyl alcohol followed by a povidone-iodine antiseptic solution. A longitudinal incision was made alone the spine and the skin separated using blunt dissection. The panniculus carnosus was carefully removed using forceps and blunt dissection, exposing the thin fascia covering the muscle body. The hydrogel was implanted in a pocket between the fascia and muscle created using dissection scissors (Fig. 2). The incision was closed by nylon suture. The rats were housed in individual cases and the implants harvested at 1, 2, and 3 weeks. Harvested tissues, which included the entire hydrogel and the surrounding muscle, were placed in 10% formaldehyde, paraffin embedded and serially sectioned (5 mm thickness) for histological staining. Sections were cut through the center of the sample containing cross sections of the gel through the underlying muscle [25]. 2.7. Histological staining and image analyses Specimens were stained for hematoxylin and eosin (H&E) and Masson’s trichome. Tissue sections were digitally imaged (0.17 mm/pixel, 20objective) using an Axiovert 200 inverted microscope. The percent collagen area was quantified from Masson’s trichrome stained sections by manually selecting the collagen (blue) area and dividing the collagen area by the total tissue area in the material. H&E stained
2.6. Animal model All animal procedures were approved by the institutional animal care and use committee (IACUC) at Chang Gung Memorial Hospital. Porous PEG-DA hydrogels of varying pore size with 10 mg/ml YRGDS were synthesized as described for the in vitro studies. Non-porous hydrogels with the same YRGDS concentration and PEG concentration were used as controls. In each condition, the hydrogels were cut into
Fig. 4. Invasion area of cocultured aggregates into porous PEG hydrogels generated with salt crystals ranging in size between 50e25, 100e50, and 150e100 mm (* indicates statistical difference between groups (p < 0.001)).
Fig. 5. Masson’s trichome stain of the tissue interface with a non-porous gel 3 weeks post-implantation. A thin layer of inflammatory tissue can be seen between the hydrogel (PEG) and underlying skeletal muscle (muscle) suggesting that these materials induce a low foreign body response (arrow). The non-porous PEG hydrogels are lost during processing due to the lack of tissue invasion.
6048
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
Fig. 6. Masson’s trichome stained of porous gels at 1 (A, D, G), 2 (B, E, H), and 3 (C, F, I) weeks post-implantation. Images are presented of gels with pores ranging from 50 to 25 mm (A, B, C), 100-50 mm (D, E, F) and 150-100 mm (G, H, I). Areas of vascularized (arrows) collagen (blue) can be observed in large (G, H, I) and medium (D, E, F) pore size gels at all time points.
tissue sections were also used to assess pore size of the PEG gels. The sections were digitally imaged (0.17 mm/pixel, 20objective) using an Axiovert 200 inverted microscope. Four images were taken per slide, and two slides were imaged per condition. All pores in a given image were quantified. Pore size was defined as the longest axis in the pore cross-section. 2.8. Statistical analysis Data are presented as means standard deviations. Significant differences between groups of data were determined by analysis of variance with HolmeSidak post-test. In all cases, p < 0.05 was considered statistically significant.
3. Result 3.1. Generation of porous PEG hydrogels The structure of the porous PEG hydrogels under fully swollen conditions can be visualized by exploiting the selective partitioning
of fluorescently-labeled protein into the bulk pores but not the cross-linked polymer network structure. [20]. The images reveal an interconnected porous structure within the hydrogels with the mean pore size varying with the range of salt crystals used (Fig. 3) Porous hydrogels generated with salts ranging from 150-100, 10050, and 50-25 mm sieved had mean pore sizes of134 28, 82 6, and 41 0.1 mm, respectively. 3.2. In vitro vessel invasion Cell invasion into the hydrogels was first examined using a 3D co-culture assay. There was no cellular invasion into non-porous hydrogels, but all pore conditions allowed extensive invasion into the hydrogels. At 1 week there were no statistical differences in invasion area between the three pore conditions (Fig. 4).However, the largest pore size (150e100 mm) gel allowed greater invasion area
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
6049
Fig. 7. H&E stained, porous gels at 1 (A, D, G), 2 (B, E, H), and 3 (C, F, I) weeks pot implantation. Images are presented of gels with pores ranging from 50 to 25 mm (A, B,C), 100-50 mm (D, E, F) and 150-100 mm (G, H, I).
(p < 0.001) than other sizes after two weeks. By week three, there was no difference between the two larger pore sizes, but the largest pore size gels (150e100 mm) had significantly greater invasion area in comparison to the smallest size (50-25 mm, p < 0.001). 3.3. In vivo response Surgically, both porous and non-porous hydrogels were easy to handle. Visually, the porous hydrogels could be observed rapidly absorbing the surrounding wound fluid after implantation within the pocket between the fascia and muscle. The non-porous hydrogels did not appear to absorb the wound fluid. All hydrogels (non-porous and porous) could be easily identified within tissues upon harvest at 1, 2, and 3 weeks following implantation. No macroscopic evidence of degradation or loss of polymer volume was observed. Histological sections of the hydrogels were stained with H&E and Masson’s trichrome to assess tissue response
and invasion. There was no histological indications of degradation at any time point for all gel conditions, and the pore size of the gels remained constant with time. There was no cell or tissue invasion into the non-porous gels. Instead, a thin layer of inflammatory tissue less than 15 mm in thickness was observed, consistent with a mild foreign body response at all time points (Fig. 5). No foreign body giant cells were observed in any of the groups. There was little cell invasion and no vessels observed in gels made with the smallest pore size at weeks 1 and 2 (Fig. 6(A),(B) and 7(A),(B)). By week 3 there was some remodeled collagen and vessels present in the pores adjacent to the surrounding tissue (Fig. 6(C)), but not deep within the hydrogel. Gels made with the two larger pore sizes, show significant cellular invasion with vessels present in the pores at week 1 (Fig. 7(D)e(I)). By week 2, vessels were observed in pores throughout the entire volume of gels with both the two larger pore sizes (Fig. 6). The vessels were contained within mature, remodeled collagen. Many of the vessels
6050
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
were filled with red blood cells. The density of new blood vessels in the pores appeared to be highest in the largest pore size. Interestingly, neither the newly remodeled collagen nor the blood vessels appear to ever contact the hydrogel surface (Fig. 8). Instead a thin layer of tissue appears adjacent to the walls of the pores. This region was highly cellular with some cells appearing spread on the hydrogel surface. The histological appearance was similar to the inflammatory tissue seen surrounding the non-porous gels. The extent of collagen remodeling within the porous PEG hydrogels was further examined by quantifying the percent of tissue area within the pores that contained collagen-rich tissue determined from the Masson’s trichrome stains (Fig.9). There were no statistical differences in percent collagen area at week 1 for all three pore sizes (Fig.9). In the second and third weeks, percent collagen area was significantly greater in the hydrogels with the largest pore size (150e100 mm) relative to the smallest (p < 0.001) There were no differences between the small (50e25 mm) and medium (100e50 mm) pore sizes. Throughout the time examined the percent of vascularized collagen formed within the porous gels was never greater than 30% of the total tissue area. The remaining tissue area appeared to be fibrin at 1 week, but at 2 and 3 weeks inflammatory tissue at the hydrogel interface made up the balance of the tissue area. 4. Discussion Hydrogels have received a significant amount of attention as tissue engineering scaffolds. The importance of pore size to vascularized tissue response has been studied in other material classes, but the importance of pores in hydrogels has not been extensively studied. PEG-based hydrogels provide an excellent model system for studying hydrogels because degradation of chemical crosslinks is required for cell and tissue invasion. Hydrogels formed from PEG allow controlled examination of the influence of pore size in the absence of degradation. Previously, we demonstrated a technique to generate porous PEG hydrogels [20]. In this study we examined the role of pore size on cell invasion and vessel assembly in both cell culture and animal models. A 3D cell culture model of sprouting angiogenesis was used to first examine invasion into the hydrogels in vitro. Models similar to this are often used as first-pass methods for evaluation of
Fig. 8. Representative Masson’s trichome stain of a porous gel generated with crystals ranging from 150 to 100 mm two weeks post-implantation. The mature collagen (blue, A) containing blood vessels (B) can be seen within the pores. A thin layer of inflammatory tissue (C) can be seen between the mature collagen and hydrogel (D). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. The fraction of tisssue within the pores of hydrogels that was mature collagen determined from Masson’s trichrome images. Collagen percentages were quantified as collagen area relative to total tissue area within the pores (* signifies statistical difference (p < 0.001)).
vascularization of tissue engineering scaffolds in vitro as they are thought to recapitulate many of the steps of neovascularization that occur in vitro [26]. The speed of sprout invasion in this model depended on pore size. While all sizes supported extensive vascular ingrowth, invasion increased with pore size. Interestingly, the in vivo vascularization results did not correlate with the cell culture model. After subfascia implantation, gels with 150e100 mm and 100e50 mm pores allow complete scaffold vascularization by two weeks, but the 50e25 mm pore size supported little or no vascularization at any time point. For the smallest pores, the in vitro model was not a good predictor of in vivo response. One primary difference between the cell culture and animal models was the cells actually interacting with the material. In the in vitro model the vascular cells come into direct contact with the material, while in vivo the vessels were never observed interacting with the material (Fig. 6). The cells interacting with the surface appeared to be inflammatory cells (Fig. 8). This suggests that peptides covalently incorporated within PEG may not directly interact with ECs following implantation, meaning they may not directly influence EC behavior. Instead, they may indirectly influence vascularization by mediating the behavior of inflammatory cells. These results are consistent with other PEG studies where peptide sequences alone have been found to be inadequate to stimulate vascularization at levels that are significantly higher than control levels [27,28]. In these studies, soluble factors that diffuse away from the material increased vascularization. Significant vascularization was observed in the porous gels in the absence of soluble factors. This appears to result from a rapid uptake of wound fluid into the hydrogels after implantation. A fibrin clot is present within the scaffolds at one week post-implantation that already had begun to vascularize. Hydrogels with pore sizes of 150e100 mm and 100e50 mm supported the formation of collagen rich, vascularized tissue within the pores. As described in the previous section, these vessels and the collagen rich regions do not appear to come into contact with the hydrogel surface. Instead there is a thin layer of inflammatory tissue adjacent to the pore surface. A recent study found that non-porous PEG hydrogels containing RGD elicited a moderate inflammatory response, with a 20e40 mm thick layer of inflammatory cells near the surfaces [29]. This fibrous capsule was consistent with a normal foreign body reaction (FBR) [29]. Our results with non-porous gels agreed with these findings. In addition, we found that a similar layer existed at all PEG-tissue
Y.-C. Chiu et al. / Biomaterials 32 (2011) 6045e6051
interfaces even within the porous hydrogels. This results in an “effective pore size” available for mature vascularized tissue formation that was less than the actual pore size. A previous study’s results also suggested that PEG hydrogels degraded in vivo possible due to oxidative degradation. The porous and non-porous PEG hydrogels were very stable in vivo. While degradation may occur in some highly oxidative environments, the moderate inflammatory response in our case did not appear sufficient to degrade the gels over the time frame we examined. We did not observe any signs of material degradation histologically and were able to recover all hydrogels implanted. The ideal tissue engineered scaffold will need to be biodegradable. In future studies we will examine the effects of controlled degradation of porous PEG-based hydrogels on tissue response and vessel persistence. 5. Conclusions In this study, we examined the role of pore size on neovascularization and tissue invasion in PEG hydrogels using both 3D cell culture and small animal models. The result from both in vitro and in vivo studies show that pore size mediates neovascularization. Interestingly, the results from the in vitro model were not consistent with in vivo, possibly due to different cells mediating interactions with the material. This study provides important insight into the role of physical structure on tissue response to porous biological scaffolds. Acknowledgements This research was supported by funding from Veterans Administration, the National Science Foundation (Grants 0852048,0731201) and the Taiwan National Science Council (982314-B-182-015-MY3). We thank Sarah Lopez for assistance with analysis of collagen content in the hydrogels. References [1] McIntire LV. Vascular assembly in engineered and natural tissues. Ann N Y Acad Sci 2002;961:246e8. [2] Chiu YC, Cheng MH, Uriel S, Brey EM. Materials for engineering vascularized adipose tissue. J Tissue Viability 2011;20:37e48. [3] Papavasiliou G, Cheng MH, Brey EM. Strategies for vascularization of polymer scaffolds. J Investig Med 2010;58:838e44. [4] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407:249e57. [5] Brauker JH, Carr-Brendel VE, Martinson LA, Crudele J, Johnston WD, Johnson RC. Neovascularization of synthetic membranes directed by membrane microarchitecture. J Biomed Mater Res 1995;29:1517e24. [6] Lim TC, Bang CP, Chian KS, Leong KF. Development of cryogenic prototyping for tissue engineering. Virtual Phys Prototyping 2008;3:25e31. [7] Bai F, Wang Z, Lu JX, Liu JA, Chen GY, Lv R, et al. The correlation between the internal structure and vascularization of controllable porous bioceramic materials In vivo: a quantitative study. Tissue Eng Part A 2010;16:3791e803.
6051
[8] Chen CW, Betz MW, Fisher JP, Paek A, Chen Y. Macroporous hydrogel scaffolds and their characterization by optical coherence tomography. Tissue Eng Part C Methods 2010;17:101e12. [9] Van de Wetering P, Metters AT, Schoenmakers RG, Hubbell JA. Poly(ethylene glycol) hydrogels formed by conjugate addition with controllable swelling, degradation, and release of pharmaceutically active proteins. J Control Release 2005;102:619e27. [10] Aimetti AA, Machen AJ, Anseth KS. Poly(ethylene glycol) hydrogels formed by thiol-ene photopolymerization for enzyme-responsive protein delivery. Biomaterials 2009;30:6048e54. [11] Hahn MS, Taite LJ, Moon JJ, Rowland MC, Ruffino KA, West JL. Photolithographic patterning of polyethylene glycol hydrogels. Biomaterials 2006;27: 2519e24. [12] Kim J, Lee KW, Hefferan TE, Currier BL, Yaszemski MJ, Lu L. Synthesis and evaluation of novel biodegradable hydrogels based on poly(ethylene glycol) and sebacic acid as tissue engineering scaffolds. Biomacromolecules 2008;9: 149e57. [13] Sontjens SH, Nettles DL, Carnahan MA, Setton LA, Grinstaff MW. Biodendrimer-based hydrogel scaffolds for cartilage tissue repair. Biomacromolecules 2006;7:310e6. [14] Moon JJ, Lee SH, West JL. Synthetic biomimetic hydrogels incorporated with ephrin-A1 for therapeutic angiogenesis. Biomacromolecules 2007;8: 42e9. [15] Lin CC, Anseth KS. Controlling affinity binding with peptide-functionalized poly(ethylene glycol) hydrogels. Adv Funct Mater 2009;19:2325. [16] Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Celldemanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth. FASEB J 2003;17:2260e2. [17] Francis ME, Uriel S, Brey EM. Endothelial cell-matrix interactions in neovascularization. Tissue Eng Part B Rev 2008;14:19e32. [18] Yannas IV, Lee E, Orgill DP, Skrabut EM, Murphy GF. Synthesis and characterization of a model extracellular matrix that induces partial regeneration of adult mammalian skin. Proc Natl Acad Sci U S A 1989;86:933e7. [19] Canal T, Peppas NA. Correlation between mesh size and equilibrium degree of swelling of polymeric networks. J Biomed Mater Res 1989;23: 1183e93. [20] Chiu YC, Larson JC, Isom A, Brey EM. Generation of porous poly(ethylene glycol) hydrogels by salt leaching. Tissue Eng Part C-Methods 2010;16: 905e12. [21] Brey EM, Appel A, Chiu YC, Zhong Z, Cheng MH, Engel H, et al. X-Ray i maging of poly(ethylene glycol) hydrogels without contrast agents. Tissue Eng Part CMethods 2010;16:1597e600. [22] Chiu YC, Larson JC, Perez-Luna VH, Brey EA. Formation of microchannels in poly(ethylene glycol) hydrogels by selective degradation of patterned microstructures. Chem Mater 2009;21:1677e82. [23] White JA, Deen WM. Agarose-dextran gels as synthetic analogs of glomerular basement membrane: water permeability. Biophysical J 2002;82:2081e9. [24] Moya ML, Cheng MH, Huang JJ, Francis-Sedlak ME, Kao SW, Opara EC, et al. The effect of FGF-1 loaded alginate microbeads on neovascularization and adipogenesis in a vascular pedicle model of adipose tissue engineering. Biomaterials 2010;31:2816e26. [25] Brey EM, King TW, Johnston C, McIntire LV, Reece GP, Patrick CW. A technique for quantitative three-dimensional analysis of microvascular structure. Microvasc Res 2002;63:279e94. [26] Brey EM, Uriel S, Greisler HP, McIntire LV. Therapeutic neovascularization: contributions from bioengineering. Tissue Eng 2005;11:567e84. [27] Seliktar D, Zisch AH, Lutolf MP, Wrana JL, Hubbell JA. MMP-2 sensitive, VEGFbearing bioactive hydrogels for promotion of vascular healing. J Biomed Mater Res Part A 2004;68A:704e16. [28] Phelps EA, Landazuri N, Thule PM, Taylor WR, Garcia AJ. Bioartificial matrices for therapeutic vascularization. Proc Natl Acad Sci U S A 2010;107:3323e8. [29] Lynn AD, Kyriakides TR, Bryant SJ. Characterization of the in vitro macrophage response and in vivo host response to poly(ethylene glycol)-based hydrogels. J Biomed Mater Res Part A 2010;93:941e53.