Communication
Micropatterning Electrospun Scaffolds to Create Intrinsic Vascular Networks Eric M. Jeffries, Shintaro Nakamura, Kee-Won Lee, Jimmy Clampffer, Hiroyuki Ijima, Yadong Wang*
Sufficient vascularization is critical to sustaining viable tissue-engineered (TE) constructs after implantation. Despite significant progress, current approaches lack suturability, porosity, and biodegradability, which hinders rapid perfusion and remodeling in vivo. Consequently, TE vascular networks capable of direct anastomosis to host vasculature and immediate perfusion upon implantation still remain elusive. Here, a hybrid fabrication method is presented for micropatterning fibrous scaffolds that are suturable, porous, and biodegradable. Fused deposition modeling offers an inexpensive and automated approach to creating sacrificial templates with vascular-like branching. By electrospinning around these poly(vinyl alcohol) templates and dissolving them in water, microvascular patterns were transferred to fibrous scaffolds. Results indicated that these scaffolds have sufficient suture retention strength to permit direct anastomosis in future studies. Vascularization of these scaffolds is demonstrated by in vitro endothelialization and perfusion.
Vascularization is critical for maintaining cell survival by providing effective nutrient delivery and waste removal beyond the 200 mm diffusion limit.[1,2] Although angiogenic factors can induce sprouting of vessels into ischemic tissue, this slow (0.1 mm per day) ‘‘extrinsic vascularization’’ can take months to perfuse large organs, eliciting necrosis within the core.[3,4] Even organs with a pre-formed internal vasculature can take up to 8 d to perfuse by inosculation.[5] Consequently, an intrinsic network that can be sutured to host vessels is needed to provide immediate perfusion upon
Prof. Y. Wang, E. M. Jeffries, Dr. K.-W. Lee, J. Clampffer Department of Bioengineering, McGowan Institute for Regenerative Medicine, Pittsburgh, PA 15261, USA E-mail: [email protected] S. Nakamura, Prof. H. Ijima Department of Chemical Engineering, Faculty of Engineering, Graduate School, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan E-mail: [email protected]; [email protected]
ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
implantation.[4] Current sources of intrinsic vascular networks rely on harvesting vascularized pedicles from donor sites. Vascular networks within TE constructs have been developed through techniques including soft lithography,[6–17] cast hydrogels,[18–27] coculture,[28–30] and solid freeform fabrication (SFF).[31–42] Recent advances in these methods have made strides toward building 3D constructs, improving porosity, and incorporating different cell types.[7,18,43] However, shortcomings such as low mechanical strength (hydrogels),[44] limited diffusion through thick solid walls separating microvessels from parenchyma (soft lithography),[7] and nondegradability (poly(dimethylsiloxane) (PDMS)-based method) still hamper direct anastomosis and perfusion of TE constructs after implantation. We address these issues by introducing a new technique that uses fused deposition modeling (FDM) as an indirect method to pattern vasculature networks within electrospun scaffolds. The electrospun polydioxanone (PDO) scaffolds are strong and suturable, to allow direct anastomosis with host vessels. Additionally, electrospun
wileyonlinelibrary.com
1
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306
R
1. Introduction
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
fibers offer high porosity and surface area to permit diffusion and promote cell attachment, respectively.[45] Furthermore, PDO resorbs within 6 months.[46,47] Feasibility of endothelializing these vascular networks is demonstrated by in vitro culture and perfusion.
2. Experimental Section 2.1. Micropatterned Template and Scaffold Fabrication 2.1.1. Fused Deposition Modeling A FDM printer (AO-101, Lulzbot, Loveland, CO) was used to extrude capillary networks from 3 mm poly(vinyl alcohol) (PVA) filament (Ultimachine, South Pittsburg, TN) (Figure 1a–c). Printing was performed at 210 8C nozzle temperature and 120 8C bed temperature. Designs for the micropatterned template were written in
G-code and optimized to maximize template uniformity and reproducibility. The printer was run using Pronterface open-source program.
2.1.2. Electrospinning PDO (Sigma–Aldrich, St. Louis, MO) was dissolved in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP, Oakwood Products, Inc., West Columbia, SC) to prepare an 11% solution. This solution was pumped at 29 mL min 1 through a 22 gauge needle at 7 kV (þ) and collected on an aluminum plate at 7 kV () and 20 cm away. The first layer of PDO fibers was deposited to an aluminum collector by vertically electrospinning 350 mL of the solution (Figure 1d–i). The printed PVA template was placed on the PDO fibers (Figure 1d–ii) before electrospinning a second layer of 350 mL PDO (Figure 1d–iii). The scaffold was removed from the aluminum collector and washed twice in deionized water with agitation for 24 h to remove PVA (Figure 1d–iv).
Figure 1. Fabrication of microvascular (MV) templates (a–c) and electrospun MV scaffolds (d–e) by FDM and electrospinning, respectively. (a) FDM extrudes a filament through a small (0.35 mm) nozzle to fuse the material to the underlying layer. (b) Automation of FDM process allows many templates to be quickly and uniformly produced. Supporting branches shown connecting the inlet and outlet are printed as part of the continuous path and trimmed before use. (c) Optimization of the printing path and extrusion parameters permit fabrication of branched structures with small dimensions. (d) Cross-sectional view of MV scaffold fabrication using templated electrospinning. (e–f) Completed MV scaffolds with 2 (e) and 4 (f) bifurcations.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
R
2
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
Micropatterning Intrinsic Vascular Networks . . . www.mbs-journal.de
2.2. Scaffold Characterization 2.2.1. Scanning Electron Microscopy The morphology of microchannels in electrospun PDO scaffolds was examined by scanning electron microscopy (SEM). Samples were fixed with 2.5% glutaraldehyde and dehydrated with graded ethanol and hexamethyldisilazane (HMDS). Dehydrated samples were sputter-coated with 3.5 nm gold and images were taken at 3 kV acceleration voltage with a field emission SEM (JSM-6330F, Jeol, Tokyo, Japan). Average diameter of fibers was measured using ImageJ software (NIH, Bethesda, MD).
pump was used, it is likely that this slow flow rate resulted in near continuous flow resembling physiological flow through capillaries. Medium was changed every other day. Fibrous PDO scaffolds were harvested on day 10 and washed with PBS. Both inlet and outlet channels were cut from scaffolds, embedded in Tissue-Tek optimal cutting temperature compound (Sakura Finetek, Inc., Torrance, CA) and snap-frozen in liquid nitrogen. Cryosections were used for immunofluorescence staining to examine HUVEC lining and cell–cell junction. Detail procedures are reported in the Supplemental Information.
2.5. Perfusion
2.2.2. Mechanical Testing Electrospun PDO sheets were punched using a dogbone-shaped die (28.75 4.75 mm2 outer dimensions, 8.25 1.5 mm2 narrow region). The thickness of hydrated samples was measured with a dial caliper (Mitutoyo). Uniaxial tensile tests were performed (n ¼ 7) to failure at 25 mm s1 using a 50 N load cell on a MTS Insight (Eden Prairie, MN). Suture retention strength was measured by methods described previously.[48] Briefly, a 6-0 suture was inserted 2 mm from the end of square sample and pulled until rupture (n ¼ 4). Suture retention strength was calculated as load/ (suture diameter sample thickness). Suturability of the microvascular scaffolds was demonstrated by end-to-end anastomosis with 9-0 interrupted sutures.
The bioreactor chamber was detached from the perfusion setup on day 10 and culture medium was aspirated from the lumen of microvascular channels of fibrous PDO scaffolds. Fluorescein isothiocyanate (FITC)-labeled dextran (150 kDa, Sigma) solution was prepared in PBS (20 mg mL1) and injected into the lumen to visualize perfused microvascular channels through endothelial cell barrier. Fluorescent images were captured on an inverted microscope. The intensity of fluorescence in perfused microvascular channels was measured using the NIS-Elements software (Nikon). Acellular PDO scaffolds were used as a negative control.
3. Results and Discussion 2.3. Cell Attachment in Static Culture
3.1. Combining FDM with Templated Electrospinning
HUVECs (Lonza, Walkersville, MD) were cultured on fibronectincoated (1 mg cm2; Chemicon International, Temecula, CA) plates using endothelial basal medium-2 (Lonza) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 1% L-glutamine–penicillin–streptomycin (Mediatech, Manassas, VA), and endothelial growth medium-2 SingleQuot Kit (Lonza). Fibrous PDO sheets were cut to circular samples 12 mm in diameter and placed under stainless steel rings (inner diameter 7.5 mm) in 24-well plates. They were sterilized with 70% ethanol, rinsed with PBS, and incubated with culture medium at 37 8C for 24 h. HUVECs (passage 4) were seeded at 1.0 106 cells mL1 and cultured for 1 d under static conditions.
We present a micropatterning technique that combines FDM with the method of templated electrospinning (Figure 1).[49] FDM is an inexpensive SFF platform that offers high versatility for the manufacture of templates, allowing us to print branched patterns that mimic vascular networks. While SFF and electrospinning have been used in various combinations for other applications,[50–57] the final composites contain material from both sources. In contrast, we used FDM only as an indirect fabrication method to create sacrificial templates used to pattern the electrospun meshes. Since the FDM templates are nonlinear patterns, they cannot be removed manually as previously described.[49] Instead, we fabricated templates out of PVA removed safely by dissolving in water. PVA is nontoxic, noncarcinogenic, generally regarded as safe by the USFDA, and has been used in tissue adhesion barriers, nerve guides, and cartilage replacement.[58] FDM also offers a high-throughput (Figure 1b) alternative to the previous labor-intensive and timeconsuming templating methods.[49] With FDM, template designs can be easily modified or scaled within the software. The production of intricate single-layer templates differs significantly from the large 3D structures FDM is intended to produce. This presents several challenges to using FDM for our application. Those familiar with FDM recognize that the first layer is most likely to cause errors with adhesion to
2.4. Bioreactor Culture Exel PTFE safelet I.V. catheters (Fisher Scientific, Pittsburgh, PA) were inserted into the inlet and outlet channels and secured by PTFE sealant tape to provide connections for the bioreactor tubing. The fibrous PDO scaffolds (n ¼ 3) were transferred into a bioreactor chamber made with polycarbonate sheets (thickness 2.4 mm) and a silicone rubber (thickness 3.1 mm). HUVECs were seeded in the lumen of the scaffold at 1.0 106 cells mL1 and allowed to attach uniformly to the scaffold lumen by rotating the chamber at 2 rpm for 4 h. After rotation, the bioreactor chamber was attached to the perfusion setup and the scaffolds were perfused with fresh medium by using a peristaltic pump at a rate of 0.6 ml min1 for 10 d. The slowest pump setting was used to flow media through the lumen to approximate capillary flow. Although a peristaltic
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3
R
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
the print bed and layer thickness. By using a continuous path and optimized printing conditions (e.g., nozzle temperature, bed temperature, level bed, and dry filament), we demonstrate uniform dimensions and reproducible deposition (Figure 1c). Additionally, the print bed was coated with a single layer of blue painter tape (3M ScotchBlue 2090). Online 3D printing forums have reported improved PVA adhesion on blue painter tape compared to glass surfaces, possibly due to changes in texture or wettability. Another drawback of FDM is the high heat (210 8C) needed to extrude PVA. Since this heat also melts common degradable polyesters including PDO, the templates were transferred from the print bed rather than directly printed on the electrospun fibers. 3.2. Benefits of Microvascular Electrospun Scaffolds Both the fibrous structure and the material properties of PDO contribute to the porous, suturable, and biodegradable nature of these scaffolds. Electrospun PDO produced welldefined microfibers with 1.5 mm average diameter
(Figure 2f) with highly interconnected interstices to permit rapid diffusion of small molecules. Electrospun fibers also provide mechanical benefits compared to porous scaffolds produced by other methods such as solvent casting particulate leaching.[59] Mechanical testing demonstrates that fibrous PDO scaffolds are strong (9.6 MPa) with suture retention strength of 142 N mm2 (Figure 3). This is close to commonly sutured electrospun poly(caprolactone) (PCL) (164 N mm2) and several times greater than poly(ether urethane urea) (PEUU) blends (35–59 N mm2) which were sutured for in vivo fixation.[48,60] PDO is a FDA-approved material used for sutures that resorb within 6 months or faster for fibers with high surface area.[46,47,61] In vitro culture of HUVECs on electrospun PDO fibers also demonstrate good attachment as indicated by sprawled morphology (Figure 2c and f). 3.3. Microvascular Endothelialization Electrospun fibers are typically packed too tightly to permit rapid cell infiltration.[62] To facilitate endothelialization
Figure 2. Scanning electron micrographs of electrospun PDO scaffolds with PVA template in place (a/d) and after template is dissolved (b/e). HUVECs attached to fibrous PDO with a sprawled morphology (c/f). Scale bar: a–c (100 mm), d–f (20 mm).
Figure 3. Mechanical properties of electrospun PDO sheets. (a) Uniaxial tensile strength (UTS). (b) Suture retention strength (SRS). (c) Suturability was demonstrated by end-to-end anastomosis of two microvascular networks using interrupted sutures.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
R
4
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
Micropatterning Intrinsic Vascular Networks . . . www.mbs-journal.de
of our scaffolds, we introduced microvascular networks by templated electrospinning. Fibrous PDO layers deposit tightly around the PVA template (Figure 2 a/d) and retain the channel shape even after PVA is dissolved (Figure 2 b/e). This process transfers the microvascular pattern to the fibrous scaffold (Figure 1e). Figure 4 demonstrates that these channels permit cells to be seeded into predetermined regions of the scaffold. The fibrous nature permits fluid extravasation while seeding cells through a filter-like process: fluid from the injected cell suspension exited through the fibers while trapping the cells against the channel walls. After culturing for 10 d in a bioreactor, HUVECs attached and proliferated on the PDO fibers and formed a near-confluent layer. The cells formed adherens junctions along the channel as demonstrated by VE-
cadherin and CD31 staining (top-view) (Figure 4a–c). The presence of HUVECs in the scaffold was further revealed by von Willibrand factor (vWF) staining of the channel cross-sections (Figure 4d–f). HUVEC staining within microvascular scaffolds revealed that despite being porous, the fibers prevent cells from immediate escape from the microchannels. Maturation of the HUVECs and confluence of EC layer were further evaluated by visualizing flow of FITC-labeled dextran solution perfused into the lumen (Figure 4g–i). The cultured microvessels show a drastic reduction in leakiness compared to the unseeded scaffold, representing barrier function of EC layer. Since some leaking still occurs, longer culture time may be necessary to allow HUVEC coverage of areas with low cell density (Figure 4b asterisk). Additionally, we expect to stabilize
Figure 4. In vitro endothelialization of fibrous microvascular scaffolds after 10 d culture. (a/b) CD31/VE-cadherin staining, respectively, of microvessel after 10 d culture (top-view). Asterisk indicates area of low cell density (Scale: 250 mm) (c) Merged CD31/VE-cadherin/DAPI image of outlined region from a/b. (Scale: 50 mm) (d-f) DAPI/vWF staining of cross-sections of inlet/outlet sections after 10 d of culture (Scale: d: 100 mm, e/f: 50 mm) (g/h) Perfusion of acellular and 10-d cultured scaffolds with FITC-labeled dextran. Channels after 10 d of culture demonstrate reduced permeability. The presence of FITC-labeled dextran beyond the channel walls indicates that channels are not yet water-tight. (Scale: 1 mm) (i) Pixel intensity along horizontal line is shown in graph.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5
R
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
HUVECs to form water-tight vessels by coculturing with mural cells such as pericytes. [63] The objective of this work was to develop a vascular network with vessels that range from large arteries and veins for anastamosis down to small capillaries for gas exchange. The 0.5–1 mm wide channels that we reported are of similar scale to small arteries and venules and should be capable of providing an intrinsic vascular supply to most of the scaffold. [64] Although FDM can achieve minimum feature sizes of 250 mm, finer resolution that is necessary for capillary formation is difficult to achieve by FDM. This is because the required extrusion pressure increases rapidly as orifice diameter decreases as described by the Hagen-Poiseuille equation.[65,66] Thus, production of future templates will explore other technologies such as pressure assisted microsyringe (PAM) and electrohydrodynamic (EHD) jet (direct melt electrospinning) which are capable of 10–30 mm feature size.[65–68] Contrarily, as template height results in electrospun fibers that encapsulate the template less tightly and create sloped sides. [50] To minimize this effect, we used templates that were short and wide. However, since PDO is a stiff material, the channels maintained this non-circular shape (Figure 4d). We plan to produce circular lumens in future iterations by fabricating scaffolds with an elastomeric material (e.g., poly(glycerol sebacate) (PGS)) that can expand once exposed to blood pressure.
4. Conclusion In summary, we present a new technique that uses FDM to produce water-soluble PVA templates as an indirect fabrication method for patterning microvascular networks within fibrous 2D structures. We anticipate combining this approach with rolling [69] or layering [54] of electrospun meshes will be able to create 3D organ scaffolds with an intrinsic vascular supply. This approach overcomes several significant challenges of engineered microvasculature including limited perfusion capacity, low strength associated with hydrogel approaches, and nondegradability for PDMS-based approaches. We expect that the thin (
Micropatterning Electrospun Scaffolds to Create Intrinsic Vascular Networks Eric M. Jeffries, Shintaro Nakamura, Kee-Won Lee, Jimmy Clampffer, Hiroyuki Ijima, Yadong Wang*
Sufficient vascularization is critical to sustaining viable tissue-engineered (TE) constructs after implantation. Despite significant progress, current approaches lack suturability, porosity, and biodegradability, which hinders rapid perfusion and remodeling in vivo. Consequently, TE vascular networks capable of direct anastomosis to host vasculature and immediate perfusion upon implantation still remain elusive. Here, a hybrid fabrication method is presented for micropatterning fibrous scaffolds that are suturable, porous, and biodegradable. Fused deposition modeling offers an inexpensive and automated approach to creating sacrificial templates with vascular-like branching. By electrospinning around these poly(vinyl alcohol) templates and dissolving them in water, microvascular patterns were transferred to fibrous scaffolds. Results indicated that these scaffolds have sufficient suture retention strength to permit direct anastomosis in future studies. Vascularization of these scaffolds is demonstrated by in vitro endothelialization and perfusion.
Vascularization is critical for maintaining cell survival by providing effective nutrient delivery and waste removal beyond the 200 mm diffusion limit.[1,2] Although angiogenic factors can induce sprouting of vessels into ischemic tissue, this slow (0.1 mm per day) ‘‘extrinsic vascularization’’ can take months to perfuse large organs, eliciting necrosis within the core.[3,4] Even organs with a pre-formed internal vasculature can take up to 8 d to perfuse by inosculation.[5] Consequently, an intrinsic network that can be sutured to host vessels is needed to provide immediate perfusion upon
Prof. Y. Wang, E. M. Jeffries, Dr. K.-W. Lee, J. Clampffer Department of Bioengineering, McGowan Institute for Regenerative Medicine, Pittsburgh, PA 15261, USA E-mail: [email protected] S. Nakamura, Prof. H. Ijima Department of Chemical Engineering, Faculty of Engineering, Graduate School, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan E-mail: [email protected]; [email protected]
ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
implantation.[4] Current sources of intrinsic vascular networks rely on harvesting vascularized pedicles from donor sites. Vascular networks within TE constructs have been developed through techniques including soft lithography,[6–17] cast hydrogels,[18–27] coculture,[28–30] and solid freeform fabrication (SFF).[31–42] Recent advances in these methods have made strides toward building 3D constructs, improving porosity, and incorporating different cell types.[7,18,43] However, shortcomings such as low mechanical strength (hydrogels),[44] limited diffusion through thick solid walls separating microvessels from parenchyma (soft lithography),[7] and nondegradability (poly(dimethylsiloxane) (PDMS)-based method) still hamper direct anastomosis and perfusion of TE constructs after implantation. We address these issues by introducing a new technique that uses fused deposition modeling (FDM) as an indirect method to pattern vasculature networks within electrospun scaffolds. The electrospun polydioxanone (PDO) scaffolds are strong and suturable, to allow direct anastomosis with host vessels. Additionally, electrospun
wileyonlinelibrary.com
1
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306
R
1. Introduction
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
fibers offer high porosity and surface area to permit diffusion and promote cell attachment, respectively.[45] Furthermore, PDO resorbs within 6 months.[46,47] Feasibility of endothelializing these vascular networks is demonstrated by in vitro culture and perfusion.
2. Experimental Section 2.1. Micropatterned Template and Scaffold Fabrication 2.1.1. Fused Deposition Modeling A FDM printer (AO-101, Lulzbot, Loveland, CO) was used to extrude capillary networks from 3 mm poly(vinyl alcohol) (PVA) filament (Ultimachine, South Pittsburg, TN) (Figure 1a–c). Printing was performed at 210 8C nozzle temperature and 120 8C bed temperature. Designs for the micropatterned template were written in
G-code and optimized to maximize template uniformity and reproducibility. The printer was run using Pronterface open-source program.
2.1.2. Electrospinning PDO (Sigma–Aldrich, St. Louis, MO) was dissolved in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP, Oakwood Products, Inc., West Columbia, SC) to prepare an 11% solution. This solution was pumped at 29 mL min 1 through a 22 gauge needle at 7 kV (þ) and collected on an aluminum plate at 7 kV () and 20 cm away. The first layer of PDO fibers was deposited to an aluminum collector by vertically electrospinning 350 mL of the solution (Figure 1d–i). The printed PVA template was placed on the PDO fibers (Figure 1d–ii) before electrospinning a second layer of 350 mL PDO (Figure 1d–iii). The scaffold was removed from the aluminum collector and washed twice in deionized water with agitation for 24 h to remove PVA (Figure 1d–iv).
Figure 1. Fabrication of microvascular (MV) templates (a–c) and electrospun MV scaffolds (d–e) by FDM and electrospinning, respectively. (a) FDM extrudes a filament through a small (0.35 mm) nozzle to fuse the material to the underlying layer. (b) Automation of FDM process allows many templates to be quickly and uniformly produced. Supporting branches shown connecting the inlet and outlet are printed as part of the continuous path and trimmed before use. (c) Optimization of the printing path and extrusion parameters permit fabrication of branched structures with small dimensions. (d) Cross-sectional view of MV scaffold fabrication using templated electrospinning. (e–f) Completed MV scaffolds with 2 (e) and 4 (f) bifurcations.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
R
2
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
Micropatterning Intrinsic Vascular Networks . . . www.mbs-journal.de
2.2. Scaffold Characterization 2.2.1. Scanning Electron Microscopy The morphology of microchannels in electrospun PDO scaffolds was examined by scanning electron microscopy (SEM). Samples were fixed with 2.5% glutaraldehyde and dehydrated with graded ethanol and hexamethyldisilazane (HMDS). Dehydrated samples were sputter-coated with 3.5 nm gold and images were taken at 3 kV acceleration voltage with a field emission SEM (JSM-6330F, Jeol, Tokyo, Japan). Average diameter of fibers was measured using ImageJ software (NIH, Bethesda, MD).
pump was used, it is likely that this slow flow rate resulted in near continuous flow resembling physiological flow through capillaries. Medium was changed every other day. Fibrous PDO scaffolds were harvested on day 10 and washed with PBS. Both inlet and outlet channels were cut from scaffolds, embedded in Tissue-Tek optimal cutting temperature compound (Sakura Finetek, Inc., Torrance, CA) and snap-frozen in liquid nitrogen. Cryosections were used for immunofluorescence staining to examine HUVEC lining and cell–cell junction. Detail procedures are reported in the Supplemental Information.
2.5. Perfusion
2.2.2. Mechanical Testing Electrospun PDO sheets were punched using a dogbone-shaped die (28.75 4.75 mm2 outer dimensions, 8.25 1.5 mm2 narrow region). The thickness of hydrated samples was measured with a dial caliper (Mitutoyo). Uniaxial tensile tests were performed (n ¼ 7) to failure at 25 mm s1 using a 50 N load cell on a MTS Insight (Eden Prairie, MN). Suture retention strength was measured by methods described previously.[48] Briefly, a 6-0 suture was inserted 2 mm from the end of square sample and pulled until rupture (n ¼ 4). Suture retention strength was calculated as load/ (suture diameter sample thickness). Suturability of the microvascular scaffolds was demonstrated by end-to-end anastomosis with 9-0 interrupted sutures.
The bioreactor chamber was detached from the perfusion setup on day 10 and culture medium was aspirated from the lumen of microvascular channels of fibrous PDO scaffolds. Fluorescein isothiocyanate (FITC)-labeled dextran (150 kDa, Sigma) solution was prepared in PBS (20 mg mL1) and injected into the lumen to visualize perfused microvascular channels through endothelial cell barrier. Fluorescent images were captured on an inverted microscope. The intensity of fluorescence in perfused microvascular channels was measured using the NIS-Elements software (Nikon). Acellular PDO scaffolds were used as a negative control.
3. Results and Discussion 2.3. Cell Attachment in Static Culture
3.1. Combining FDM with Templated Electrospinning
HUVECs (Lonza, Walkersville, MD) were cultured on fibronectincoated (1 mg cm2; Chemicon International, Temecula, CA) plates using endothelial basal medium-2 (Lonza) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 1% L-glutamine–penicillin–streptomycin (Mediatech, Manassas, VA), and endothelial growth medium-2 SingleQuot Kit (Lonza). Fibrous PDO sheets were cut to circular samples 12 mm in diameter and placed under stainless steel rings (inner diameter 7.5 mm) in 24-well plates. They were sterilized with 70% ethanol, rinsed with PBS, and incubated with culture medium at 37 8C for 24 h. HUVECs (passage 4) were seeded at 1.0 106 cells mL1 and cultured for 1 d under static conditions.
We present a micropatterning technique that combines FDM with the method of templated electrospinning (Figure 1).[49] FDM is an inexpensive SFF platform that offers high versatility for the manufacture of templates, allowing us to print branched patterns that mimic vascular networks. While SFF and electrospinning have been used in various combinations for other applications,[50–57] the final composites contain material from both sources. In contrast, we used FDM only as an indirect fabrication method to create sacrificial templates used to pattern the electrospun meshes. Since the FDM templates are nonlinear patterns, they cannot be removed manually as previously described.[49] Instead, we fabricated templates out of PVA removed safely by dissolving in water. PVA is nontoxic, noncarcinogenic, generally regarded as safe by the USFDA, and has been used in tissue adhesion barriers, nerve guides, and cartilage replacement.[58] FDM also offers a high-throughput (Figure 1b) alternative to the previous labor-intensive and timeconsuming templating methods.[49] With FDM, template designs can be easily modified or scaled within the software. The production of intricate single-layer templates differs significantly from the large 3D structures FDM is intended to produce. This presents several challenges to using FDM for our application. Those familiar with FDM recognize that the first layer is most likely to cause errors with adhesion to
2.4. Bioreactor Culture Exel PTFE safelet I.V. catheters (Fisher Scientific, Pittsburgh, PA) were inserted into the inlet and outlet channels and secured by PTFE sealant tape to provide connections for the bioreactor tubing. The fibrous PDO scaffolds (n ¼ 3) were transferred into a bioreactor chamber made with polycarbonate sheets (thickness 2.4 mm) and a silicone rubber (thickness 3.1 mm). HUVECs were seeded in the lumen of the scaffold at 1.0 106 cells mL1 and allowed to attach uniformly to the scaffold lumen by rotating the chamber at 2 rpm for 4 h. After rotation, the bioreactor chamber was attached to the perfusion setup and the scaffolds were perfused with fresh medium by using a peristaltic pump at a rate of 0.6 ml min1 for 10 d. The slowest pump setting was used to flow media through the lumen to approximate capillary flow. Although a peristaltic
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3
R
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
the print bed and layer thickness. By using a continuous path and optimized printing conditions (e.g., nozzle temperature, bed temperature, level bed, and dry filament), we demonstrate uniform dimensions and reproducible deposition (Figure 1c). Additionally, the print bed was coated with a single layer of blue painter tape (3M ScotchBlue 2090). Online 3D printing forums have reported improved PVA adhesion on blue painter tape compared to glass surfaces, possibly due to changes in texture or wettability. Another drawback of FDM is the high heat (210 8C) needed to extrude PVA. Since this heat also melts common degradable polyesters including PDO, the templates were transferred from the print bed rather than directly printed on the electrospun fibers. 3.2. Benefits of Microvascular Electrospun Scaffolds Both the fibrous structure and the material properties of PDO contribute to the porous, suturable, and biodegradable nature of these scaffolds. Electrospun PDO produced welldefined microfibers with 1.5 mm average diameter
(Figure 2f) with highly interconnected interstices to permit rapid diffusion of small molecules. Electrospun fibers also provide mechanical benefits compared to porous scaffolds produced by other methods such as solvent casting particulate leaching.[59] Mechanical testing demonstrates that fibrous PDO scaffolds are strong (9.6 MPa) with suture retention strength of 142 N mm2 (Figure 3). This is close to commonly sutured electrospun poly(caprolactone) (PCL) (164 N mm2) and several times greater than poly(ether urethane urea) (PEUU) blends (35–59 N mm2) which were sutured for in vivo fixation.[48,60] PDO is a FDA-approved material used for sutures that resorb within 6 months or faster for fibers with high surface area.[46,47,61] In vitro culture of HUVECs on electrospun PDO fibers also demonstrate good attachment as indicated by sprawled morphology (Figure 2c and f). 3.3. Microvascular Endothelialization Electrospun fibers are typically packed too tightly to permit rapid cell infiltration.[62] To facilitate endothelialization
Figure 2. Scanning electron micrographs of electrospun PDO scaffolds with PVA template in place (a/d) and after template is dissolved (b/e). HUVECs attached to fibrous PDO with a sprawled morphology (c/f). Scale bar: a–c (100 mm), d–f (20 mm).
Figure 3. Mechanical properties of electrospun PDO sheets. (a) Uniaxial tensile strength (UTS). (b) Suture retention strength (SRS). (c) Suturability was demonstrated by end-to-end anastomosis of two microvascular networks using interrupted sutures.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
R
4
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
Micropatterning Intrinsic Vascular Networks . . . www.mbs-journal.de
of our scaffolds, we introduced microvascular networks by templated electrospinning. Fibrous PDO layers deposit tightly around the PVA template (Figure 2 a/d) and retain the channel shape even after PVA is dissolved (Figure 2 b/e). This process transfers the microvascular pattern to the fibrous scaffold (Figure 1e). Figure 4 demonstrates that these channels permit cells to be seeded into predetermined regions of the scaffold. The fibrous nature permits fluid extravasation while seeding cells through a filter-like process: fluid from the injected cell suspension exited through the fibers while trapping the cells against the channel walls. After culturing for 10 d in a bioreactor, HUVECs attached and proliferated on the PDO fibers and formed a near-confluent layer. The cells formed adherens junctions along the channel as demonstrated by VE-
cadherin and CD31 staining (top-view) (Figure 4a–c). The presence of HUVECs in the scaffold was further revealed by von Willibrand factor (vWF) staining of the channel cross-sections (Figure 4d–f). HUVEC staining within microvascular scaffolds revealed that despite being porous, the fibers prevent cells from immediate escape from the microchannels. Maturation of the HUVECs and confluence of EC layer were further evaluated by visualizing flow of FITC-labeled dextran solution perfused into the lumen (Figure 4g–i). The cultured microvessels show a drastic reduction in leakiness compared to the unseeded scaffold, representing barrier function of EC layer. Since some leaking still occurs, longer culture time may be necessary to allow HUVEC coverage of areas with low cell density (Figure 4b asterisk). Additionally, we expect to stabilize
Figure 4. In vitro endothelialization of fibrous microvascular scaffolds after 10 d culture. (a/b) CD31/VE-cadherin staining, respectively, of microvessel after 10 d culture (top-view). Asterisk indicates area of low cell density (Scale: 250 mm) (c) Merged CD31/VE-cadherin/DAPI image of outlined region from a/b. (Scale: 50 mm) (d-f) DAPI/vWF staining of cross-sections of inlet/outlet sections after 10 d of culture (Scale: d: 100 mm, e/f: 50 mm) (g/h) Perfusion of acellular and 10-d cultured scaffolds with FITC-labeled dextran. Channels after 10 d of culture demonstrate reduced permeability. The presence of FITC-labeled dextran beyond the channel walls indicates that channels are not yet water-tight. (Scale: 1 mm) (i) Pixel intensity along horizontal line is shown in graph.
Macromol. Biosci. 2014, DOI: 10.1002/mabi.201400306 ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5
R
www.MaterialsViews.com
Early View Publication; these are NOT the final page numbers, use DOI for citation !!
E. M. Jeffries, S. Nakamura, K-W. Lee, J. Clampffer, H. Ijima, Y. Wang www.mbs-journal.de
HUVECs to form water-tight vessels by coculturing with mural cells such as pericytes. [63] The objective of this work was to develop a vascular network with vessels that range from large arteries and veins for anastamosis down to small capillaries for gas exchange. The 0.5–1 mm wide channels that we reported are of similar scale to small arteries and venules and should be capable of providing an intrinsic vascular supply to most of the scaffold. [64] Although FDM can achieve minimum feature sizes of 250 mm, finer resolution that is necessary for capillary formation is difficult to achieve by FDM. This is because the required extrusion pressure increases rapidly as orifice diameter decreases as described by the Hagen-Poiseuille equation.[65,66] Thus, production of future templates will explore other technologies such as pressure assisted microsyringe (PAM) and electrohydrodynamic (EHD) jet (direct melt electrospinning) which are capable of 10–30 mm feature size.[65–68] Contrarily, as template height results in electrospun fibers that encapsulate the template less tightly and create sloped sides. [50] To minimize this effect, we used templates that were short and wide. However, since PDO is a stiff material, the channels maintained this non-circular shape (Figure 4d). We plan to produce circular lumens in future iterations by fabricating scaffolds with an elastomeric material (e.g., poly(glycerol sebacate) (PGS)) that can expand once exposed to blood pressure.
4. Conclusion In summary, we present a new technique that uses FDM to produce water-soluble PVA templates as an indirect fabrication method for patterning microvascular networks within fibrous 2D structures. We anticipate combining this approach with rolling [69] or layering [54] of electrospun meshes will be able to create 3D organ scaffolds with an intrinsic vascular supply. This approach overcomes several significant challenges of engineered microvasculature including limited perfusion capacity, low strength associated with hydrogel approaches, and nondegradability for PDMS-based approaches. We expect that the thin (
