microfluidic approach for the scalable formation of 3d soft ......we have developed a...
TRANSCRIPT
Microfluidic Approach for the Scalable Formation of 3D Soft Materials with Tailored Biomechanical Properties for
Tissue Engineering Applications
by
Lian Leng
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Department of Mechanical and Industrial Engineering University of Toronto
© Copyright by Lian Leng 2015
ii
Microfluidic Approach for the Scalable Formation of 3D Soft
Materials with Tailored Biomechanical Properties for Tissue
Engineering Applications
Lian Leng
Doctor of Philosophy
Mechanical and Industrial Engineering
University of Toronto
2015
Abstract
Through evolution, nature has defined materials in a highly organized and hierarchical fashion,
with a complex organization on the nano and microscale that is closely linked to the different
tissues and organs function. A variety of macroscale parameters such as mechanical strength,
elasticity, and permeability to various molecules are tissue-specific and are a direct result of their
intricate structural organization. The field of tissue engineering has seen the development of a
broad range of technologies with the common goal of enabling the generation of complex tissues
that mimic the intricate architecture and resulting properties of natural tissues. The methods
developed so far are promising but are faced with short-comings in terms of scalability,
throughput, and dynamic control over the local material composition and properties. These
strategies are often limited in size (<1cm), rely on sequential assembly processes, and the
resulting mechanical properties of these engineered tissues are often not comparable to native
tissues.
We have developed a microfluidic-based printing platform for the continuous and scalable
formation of planar soft materials with high degree of control over the material composition,
iii
cellular composition, nanoscale to macroscale assembly, and mechanical properties. The spatial
control over the different compositions can be precisely defined, with a resolution of 75µm. The
incorporation of a variety of payloads such as biomolecules, fluorescent particles and viable cells
including cardiomyocytes, keratinocytes, and fibroblasts is demonstrated. The platform was used
with a variety of matrix materials including alginate and collagen type I from rat tail. The
mechanical properties of the printed soft material sheets produced could be tailored by either a
planar assembly of various materials possessing distinct mechanical properties, or by controlling
the degree of alignment of collagen fibers. Demonstrated applications include the formation of
aligned collagen sheets with mechanical properties comparable to native tissues, and up to 6cm
wide cell-populated skin substitute with physiological properties similar to the native skin. The
approach presented promises wide applications in the field of regenerative medicine where the
ability to tailor the material and cellular composition, as well as the mechanical properties of the
replacement tissue have become a key requirement.
iv
Acknowledgments
This dissertation is dedicated to my family. To my parents, for being supportive and encouraging
through ups and downs. To my sister, for always being there for me and for being such a good
listener. To my family for teaching me the value of hard work and sharing their passion for what
they do.
To my dissertation committee, for their guidance and valuable inputs. To Dr. Axel Guenther, for
teaching me how to be a thorough and meticulous experimentalist, to be inquisitive and take any
failure with excitement for it can only mean new discoveries, and for sharing his passion for
science.
To our collaborators from Dr. Elliot Chaikof’s laboratory. Particularly Dr. Stephanie Grainger
for all the mechanical testing, and TEM-SEM imaging of the collagen samples. To our
collaborators from Dr. Marc Jeschke’s laboratory at the Sunnybrook Research Institute,
particularly Shermineh Minai for being such a great help in providing us with large cell numbers,
and Cassandra Belo for the long hours spent on in vitro and in vivo sample staining and imaging.
I would also like to give my appreciation to Dr. Saeid Amini-Nik and Cassandra Belo for
performing the surgeries for our in vivo experiments. To Dr. Milica Radisic and Boyang Zhang
for our collaborative work on fibroblasts and cardiomyocyte printing. I would like to extend my
thanks to Boyang Zhang for our extensive discussions and for being such a pleasure to work
with. To Prof. Craig Simmons for access to the tensile testing apparatus in his laboratory.
To all members of the Guenther lab, particularly Sanjesh Yasotharan, Arianna McAllister, and
Geoff Vishloff for insightful discussions and help.
v
Table of Contents
Acknowledgments .......................................................................................................................... iv
Table of Contents ............................................................................................................................ v
List of Figures ............................................................................................................................... vii
List of Tables .................................................................................................................................. x
Contributions .................................................................................................................................. xi
1 Chapter 1 - Introduction ............................................................................................................. 1
1.1 Strategies for the Formation of Organized Materials ......................................................... 2
1.1.1 Photo and Laser Polymerization ............................................................................. 2
1.1.2 Molding and Templating ......................................................................................... 3
1.1.3 Self-Assembly of Living Cells ................................................................................ 5
1.1.4 Bioprinting .............................................................................................................. 6
2 Chapter 2 - Formation of Aligned Soft Material Sheets .......................................................... 10
2.1 Fibrillar Structure of Materials in Nature ......................................................................... 10
2.2 Collagen Organization in Tissues ..................................................................................... 11
2.3 Strategies for Controlling Collagen Self-Assembly In Vitro ............................................ 13
2.4 Strategy for the Continuous Formation of Collagen Sheets with Aligned
Nanostructure .................................................................................................................... 16
2.4.1 Materials and Methods .......................................................................................... 17
2.4.2 Results ................................................................................................................... 21
3 Chapter 3 - Continuous Formation of Patterned Soft Material Sheets Populated with
Molecular and Cellular Payloads ............................................................................................. 34
3.1 Alginate in Tissue Engineering ......................................................................................... 34
3.2 Strategy for the Continuous Formation of Soft Material Sheets of Heterogeneous
Composition ...................................................................................................................... 36
3.2.1 Material and Methods ........................................................................................... 36
3.2.2 Mosaic Hydrogel Formation ................................................................................. 41
vi
3.2.3 Spatiotemporal Control and Payload Incorporation ............................................. 43
3.2.4 Information Encoding ........................................................................................... 45
3.2.5 Geometric Control over Mosaic Hydrogel Properties .......................................... 46
3.2.6 Scalable Formation of 3D Hydrogel Assemblies .................................................. 47
3.2.7 Planar Co-localization of Single and Multiple Cell Types ................................... 47
4 Chapter 4 - Skin Printer: Continuous and Scalable Organization of Layered Tissue .............. 58
4.1 Skin Structure .................................................................................................................... 58
4.2 Skin grafts and Artificial Skin Substitutes ........................................................................ 59
4.2.1 Skin Grafts ............................................................................................................ 60
4.2.2 Artificial Skin Substitutes ..................................................................................... 61
4.2.3 Microfluidic approaches ....................................................................................... 62
4.3 3D Printer Strategy for the Continuous Formation of Skin Substitute for Wound
Healing .............................................................................................................................. 63
4.3.1 Materials and Methods .......................................................................................... 63
4.3.2 Formation of Cell-Populated Skin Grafts ............................................................. 67
4.3.3 Pattern Formation – No Cells ............................................................................... 68
4.3.4 Characterization of Printed Tissue Substitutes ..................................................... 68
4.3.5 Cell-Populated Skin Grafts in Vitro ...................................................................... 71
4.3.6 Scalable Formation of Tissue Substitutes ............................................................. 72
4.3.7 Cell-Populated Skin Grafts In Vivo ....................................................................... 72
Summary and Future Work ........................................................................................................... 81
References ..................................................................................................................................... 86
Appendix ..................................................................................................................................... 107
vii
List of Figures
Figure 1. Schematic illustration of the flowable conversation of a collagne solution to a
cross-linked collagen sheet with a high degree of molecular alignment…………...
27
Figure 2. Schematic illustrations of experimental setup………………………...………….... 28
Figure 3. Detailed study of flow-focusing and strain-inducing pulling using a vertical
constriction unit……………………………………………………………………..
29
Figure 4. Measured collagen sheet width and thickness as a function of V* ranging from 0.1
to 10…………………………………………………………………………………
30
Figure 5. Characterization of collagen fibril compaction, alignment, and resulting
mechanical properties…………………………………………………………….....
32
Figure 6. One-step formation of mosaic hydrogels……………………………...………….... 50
Figure 7. Analytical model of the time-dependent concentration of free cross-linker, free
alginate, and cross-linked gel...……………………………………………………..
51
Figure 8. Continuous formation of hydrogel sheets: experimental setup and device designs..
52
Figure 9. Dynamically encoded information in planar hydrogels…………………………....
54
Figure 10. Mosaic hydrogels…………………………..………………………...………….... 56
Figure 11. Intact human skin and bioprinted skin grafts..…………………………………….
74
Figure 12. Skin printer…………………..……………………………………………………
75
Figure 13. Patterned single-layered sheets…………………………………………………....
76
Figure 14. In vitro characterization of printed skin grafts.……………………...………….... 77
Figure 15. Material optimization for keratinocyte printing…..……………………………….
78
Figure 16. Scalable formation of skin grafts………………………………….....……………
79
Figure 17. In vivo characterization of printed skin grafts…………………….....……………
80
Figure 18. Benefit of cell clustering and characterization of skin microtissues...……………
84
Figure 19. In vivo characterization of bilayered cell-populated skin grafts....………………..
85
viii
Appendix
Figure A1. Rendered device designs…………………..…………………….....……………
108
Figure A2. Characterization of velocity profiles within the constriction………...…………
110
Figure A3. Numerical simulation of the flow behavior of the focusing solution within
½ of the constriction (½ HC)...…………………………………………………..
112
Figure A4. Characterization of collagen sheet width and thickness produced at varying
V* and QF, with constant QM..…………………………………………………...
112
Figure A5. Mechanical properties of collagen sheets subjected to FIB incubation and
drying post-extrusion……………...……………………………………………..
113
Figure A6. TEM images of collagen sheets produced at various V*……………...…………
113
Figure A7. Characterization of fibril spacing and compaction by autocorrelation of TEM
and SEM images………...……………………………………………………….
114
Figure A8. Photograph of microfluidic device with constriction unit…..………...…………
115
Figure A9. Constriction manifold assembly………………………...…..………...…………
116
Figure A10. Constriction manifold – Top piece……………………...…..………...……..…
116
Figure A11. Constriction manifold – Bottom piece……………………...…..………...……
117
Figure A12. XZ constriction manifold assembly….……………………...…..………...……
117
Figure A13. XZ constriction manifold – Frame...….……………………...…..………...…...
118
Figure A14. XZ constriction manifold – Left constriction bracket..……...…..………...…...
118
Figure A15. XZ constriction manifold – Right constriction bracket……...…..………...…...
119
Figure A16. Control over soft material thickness as a function of QF………………………
121
Figure A17. Characterization of pressure in on-chip reservoirs……..………………………
122
Figure A18. Shear stress profile within a microfluidic channel………..……………………
123
Figure A19. Viability and distribution of printed cells……………...………………………
124
Figure A20. Modulus of elasticity for a homogeneous soft material composed
of 2%w.t. alginate produced in free and pulled extrusion modes…..…………
125
ix
Figure A21. Line camera intensity measurements of the UN Charter………………………
126
Figure A22. Full-thickness burns and current treatments…………………….…..…………
128
Figure A23. Characterization of spot volume as a function of valve actuation….…………
128
Figure A24. Degradation studies of skin grafts……………………………….…..…………
129
Figure A25. Mechanical properties of skin grafts as a function of culture time...………….
130
Figure A26. Stress-strain curves of skin grafts materials as a function of culture time…….
131
Figure A27. Stress-Strain curves of skin graft printed with parallel spots………………….
132
Figure A28. Stress-Strain curves of skin graft printed with parallel stripes………………...
133
Figure A29. Stress-Strain curves of skin graft printed with alternating voids..……………..
134
Figure A30. Stress-Strain curves of skin graft printed with alternating spots……………….
135
x
List of Tables
Table A1. Natural tissues and their mechanical properties.………………….....………….
109
Table A2. Mechanical and structural properties of natural and synthetic collagen gels
created using a variety of strategies……………….....…………………………..
110
Table A3. Parameters used in numerical model, treating the focusing solution as a
Newtonian fluid…………………...……………….....…………………………..
111
Table A4. Mechanical properties of alginate gels………….………………….....………….
120
Table A5. Literature data of elastic moduli of human skin measured in vivo using
various strategies..………………...……………….....…………………………..
127
xi
Contributions
Publications
Leng, L., McAllister, A., Zhang, B., Radisic, M., and Guenther, A., "Mosaic Hydrogels:
One-Step Formation of Multiscale Soft Materials", Advanced Materials, 24, July 2012,
3650-3658
Singh, R., Genov, R., Leng, L., Guenther, A., "A Hybrid CMOS-Microfluidic
Luminescence Contact Imaging Microsystem," accepted at 2009 SPIE Optics+Photonics
(Invited), Aug 2009. SPIE Paper Number 7397-39
Manuscript in progress
Leng, L., Grainger, S., Chaikof, E., Guenther, A., “In-Flow Preparation of Collagen
Sheets with Tunable Molecular Alignment and their Influence on Cell Behavior”.
Leng, L., Ba, Q., Amini-Nik, S., Jeschke, M., Guenther, A., “Skin Printer: Continuous
and Scalable Organization of Layered Tissue”.
Patent
Leng, L., Zhang, B., McAllister, A., Wollard, A., Radisic, M., Guenther, A. "Devices and
methods for producing planar polymeric materials using microfluidics, WO2013075248A1.
Others
Chapter 2: TEM/SEM images and tensile data were obtained by Dr. Stephanie Grainger.
Chapter 4: In vitro and in vivo sample staining and imaging was performed by Cassandra Belo.
Animal surgeries and histology data were performed and obtained by Dr. Saeid Amini-Nik and
Cassandra Belo.
1
1 Chapter 1 - Introduction
Natural selection and evolution through time has resulted in the formation of a wide variety of
materials in living organisms, shaping them to adapt to environmental changes and optimize their
performance. From microbes to plants and mammals, the materials shape and microstructure are
closely related, and are formed in a simultaneous process, that is through growth of the organ 1.
As these materials form at multiple length scales, from nano to macroscale, the resulting
structure of these biological materials becomes highly hierarchical 2. Through evolution, this
complex organization has been optimized to efficiently serve tissue-specific functions and define
the characteristic properties of the materials.
In this thesis, we present a novel platform for the continuous and high throughput formation of
complex soft materials with tunable composition and structure that can be adapted to specific tissues.
In the first chapter, various strategies for the assembly of organized materials are reviewed. In the
second chapter, we explore our ability to define directionally dependent microstructures with the
formation of collagen sheets of highly anisotropic fibril alignment. The anisotropic microstructures
and resulting mechanical properties are investigated. In the third chapter, we introduce an adapted
platform for the formation of materials with heterogeneous composition, possible through the
addition of on-chip wells that can be loaded with a variety of payloads ranging from biomaterials and
fluorescence particles to living cells. The resulting mechanical properties such as elastic modulus and
diffusivity, as well as the viability and attachment of living cells were investigated. In the fourth
chapter, we focus on the application of our platform to the formation of skin grafts composed of
islands of microtissues mimicking the epidermal and dermal layer of skin. In vitro and in vivo
characterization of the grafts mechanical properties, cell proliferation, and wound healing on murine
models were performed.
2
1.1 Strategies for the Formation of Organized Materials
Over the last few decades, research groups have developed a broad range of strategies to
generate synthetic organized materials in an attempt to recreate the multicellular composition and
structural organization of natural materials. The ultimate goal of these technology developments
is to facilitate the directed assembly of biologically relevant materials, with prescribed three-
dimensional hierarchical organization and physiologically relevant composition. Some of these
approaches include, but are not limited to the formation of microscale building blocks that are
either randomly or orderly assembled into higher-order structures. These building blocks may be
in the form of spherical or spheroidal bubbles and droplets, or polymer particles and disks 3. The
formation of building blocks can be achieved through a variety of approaches such as
photopolymerization 4-17, molding 18-31, self-assembly of living cells 32-40, and bioprinting 26,27,41-
70.
1.1.1 Photo and Laser Polymerization
Photoinitiated polymerization techniques have been adapted to generate such building blocks
with the ability to control the geometry and sizes of these particles within the xy-plane 10-13 or
even in three dimensions 14. These particles can be subsequently assembled either manually 13,
through surface tension at a liquid-liquid interface 11, through shape-affinity and fluid flow 15,16,
or microfluidic templating 17. Similarly, these planar assembly approaches can be scaled-up to
three-dimensional assemblies of droplets 4, polymer particles 5, and microbubbles 6. Cheung et
al. used confocal scanning as well as conventional fluorescence microscopy to polymerize three-
dimensional polymer particles with various geometries, sizes up to 1mm2, and spatial resolution
down to 3µm 7. Select regions could be sequentially photopolymerized, with each step requiring
a wash before flowing the following photocurable reagent. A variety of polymers were tested,
including PEG-DA and other acrylate-based polymers.
Larger-scale photopolymerized scaffolds of heterogeneous cellular and material compositions
were also created by through a bottom-up assembly using stereolithography 8. A reservoir filled
with PEG-DA and RGDS peptide solution containing various cell types was photopolymerized
using a laser beam. Multiple layers were built up in a sequential process where a second layer of
3
uncrosslinked cell-seeded polymer solution was coated onto the cross-linked layer, exposed, and
the process repeated for subsequent layers. Complex structures 2mm in thickness were produced
with a total of 20 layers. This approach remains sequential, results in wasteful discarding of
excess uncrosslinked cell-seeded polymer, and is limited in thickness due to the lack of a
vascular network. Cell death was noticeable in the centre of the scaffold, with an increase in cell
number at the periphery of the scaffolds after 14 days of culture.
Laser microablation techniques have also been applied to create three-dimensional scaffolds with
geometrically defined micropores 9. Multilayered poly (glycerol-sebacate) (PGS) scaffolds with
controlled pore microarchitectures were fabricated and assembled through a layer-by-layer
oxygen plasma treatment. The final assembly required storage under compression for 18hrs to
fully bond the different layers, resulting in a final construct 6mm in diameter and 500µm thick.
The scaffolds were subsequently immersed in a suspension of cardiomyocytes for seeding and
showed contractile behavior of the entire construct.
1.1.2 Molding and Templating
Hierarchical biological structures can also be produced through sequential templating, molding,
and layer-by-layer assembly approaches.
Template-assisted self-assembly of spherical colloids and hydrogel beads was achieved by
capillary forces and geometric confinement 18,19. Matsunaga et al. produced three-dimensional
tissue architectures by packing cell-populated beads into silicon molds 19. The collagen type I
beads were produced at a rate of 104 beads/min using a flow-focusing device and had diameters
ranging from 100-300µm. Cells were seeded and allowed to attach and penetrate the collagen
beads. The resulting cell-populated beads were stacked into polydimethylsiloxane (PDMS)
molds and observed to attach to one another and contract as a bulk structure from an initial
1.5mm thickness down to 1.2mm. However, the lack of a vascularized network within such thick
substrate presents a challenge in diffuse transport of nutrients throughout the scaffold, a
phenomenon also observed in the formation of large alginate molds for cartilage regeneration 20.
Cell-seeded alginate implants were injection-molded to anatomical shapes, with average molded
construct size of 25.3 × 17.4 × 6.6mm 20. Although articular chondrocytes in cartilage tissue
4
engineering require less vascularization and do not undergo necrosis in thick (>0.3mm)
scaffolds, it is commonly observed that low oxygen concentration in the chore of thick hydrogel
implants result in an enhanced extracellular matrix (ECM) formation and cell necrosis 21.
Molding and layering approaches have been used separately or in combination to generate cell-
populated multilayered biopolymers constructs. Nishiguchi et al. employed a layer-by-layer
assembly technique to create tissue models such as blood vessels, skeletal muscle, and
connective tissue 22. Less than 10nm thick fibronectin-gelatin films were prepared and deposited
on cell surfaces, promoting cell-cell adhesion between layers. Their approach resulted in the
formation of 1cm2 and 50µm thick cell-populated substrates. Although promising due to the
control over the cell type and location in the z-direction, this sequential approach is lengthy and
therefore limited to submillimeter-thick constructs. In another approach, layer-by-layer assembly
processes were combined with microfluidic geometries to create three-dimensional layered
biopolymer matrices containing cell suspensions at a concentration of 3 × 105 cells/mL 23.
Matrices consisting of collagen, collagen-chitosan, matrigel, and fibrin were sequentially flown
through microfluidic channels 350µm [W] × 300µm [δ] and cross-linked by incubation at 37 ̊C
for each layer introduced, creating multilayered composite biopolymer constructs. The thickness
was however limited by soft-lithography fabrication of the microfluidic channels.
Biomaterials layers were also molded with perfusable channels to mimic in a simplified fashion
the vascular network of natural tissues, thereby promoting the transport and diffusion of nutrients
throughout the entire scaffold 24-30. Choi et al. molded a solution of primary chondrocytes
suspended in 4%w.t. calcium alginate into an aluminum jig 24. A silicon master patterned with a
connected array of 100µm wide microchannels was placed within the jig and serve to create a
perfusable network of microchannels, ensuring cell viability throughout the 2mm thick scaffold.
Perfusable networks were also produced using sacrificial molds 30,31. A few examples of
sacrificial templating include the incorporation of phosphate glass fibers within a 50µm thin
planar collagen layer 31. The structure was subsequently built-up by rolling onto itself and the
fibers degraded, producing a perfusable tubular structure 2mm in diameter. Yoshida et al. used a
similar templating approach to generate biodegradable multi-layered construct mimicking the
bilayered structure of blood capillaries . A parallel array of microchannels within a
biodegradable hydrogel was created by the molding and later removing silica tubes 720µm in
5
diameter. The hollow microchannels created within the disulfide-crosslinked γ-PGA gels were
then sequentially coated with smooth muscle cells followed by endothelial cells. Through this
approach, they were able to construct multilayered blood capillaries possessing barrier functions
similar to the native blood capillaries. Permeability assay was performed by flowing FITC-
Dextran molecules through the tubular network and comparing the diffusion over 24hrs between
the cell-coated and uncoated constructs. They also attempted to collect these cell-populated
capillaries by decomposing the casting hydrogel by the addition of cysteine, a biocompatible
reductant. The collected capillaries were 1cm in length and maintained their hollow structure,
promising potential use as implantable artificial blood capillary.
1.1.3 Self-Assembly of Living Cells
Another strategy for the assembly of organized three-dimensional tissues relies on the template-
free self-assembly of living cells into sheets 32-37. Yang et al. successfully harvested intact cell
sheets from thermally responsive dish, preserving cell-cell junction protein and ECM that would
be degraded by conventional trypsinization for cell recovery 36. The temperature-responsive
culture dishes were created by covalent grafting of a temperature-responsive polymer poly(N-
isopropylacrylamide) (PIPAAm) to ordinary tissue culture dishes. Under normal culture
conditions at 37ºC, the dish surfaces are hydrophobic and cells are able to attach, spread, and
proliferate. Upon temperature reduction below the polymer’s lower critical solution temperature
of 32ºC, the polymer surface becomes hydrophilic and swells, forming a hydration layer between
the dish surface and the cultured cells. This layer allows the spontaneous detachment of cell
sheets without the need for enzymatic treatments such as trypsinization. By avoiding proteolytic
treatment, critical cell surface proteins such as ion channels, growth factor receptors and cell-to-
cell junction proteins remain intact, and cells can be noninvasively harvested as intact sheets
along with their deposited ECM. These sheets are subsequently able to spontaneously attach to
the region of implant due to ECM present. Relying on the cell sheets spontaneous attachment
capability, up to five cell sheets were layered to engineer transplantable corneal epithelial
sheets 38 as well as cardiac sheets 39. This technique was later on adapted to produce cell sheets
for co-culture of different cell types 35. In this approach, two thermoresponsive polymers
exhibiting different transition temperatures were sequentially patterned onto culture dishes,
6
namely hydrophilic poly(N-isopropylacrylamide) (PIPAAm) and hydrophobic monomer n-butyl
methacrylate (BMA). Sequential seeding of viable cells enabled the patterned co-culture of rat
primary hepatocytes and bovine carotid endothelial cells. Chen et al. produced in a similar
approach cell sheets by seeding onto thermoreversible methylcellulose hydrogels 32. Human
foreskin fibroblasts were seeded at a density of 4 × 104 cells/cm2 and their attachment was
improved by evenly coating the hydrogel with a layer of neutral aqueous collagen. Cell sheets
were completely detached within 20min, and multilayered sheets were produced by directly
seeding a second layer of cells onto a non-detached cell sheet.
However, due to the reduced diffusion of oxygen and nutrient through the non-vascularized
construct, the maximum number of layers is limited to five layers (~200-300µm total thickness).
In addition, cell sheet strategies require large cell numbers to produce centimeter scale sheets.
Currently, seeding of 2×104 cells/cm2 produces 5.98 ± 0.12cm2 sheets before detachment from
the thermally-reversible dish. Once detached, the cell sheet area is reduced to 1.16 ± 0.08cm2
with a thickness of 45 ± 8µm due to cytoskeletal tensile reorganization 39. The steps involved in
the preparation of cell sheets are also timely as they required approximately 2-3 weeks before the
sheets can be manipulated, about 1h for sheet detachment, and about 30min for sheet
reattachment to create multilayer constructs.
More recently, natural ECM sheets were decellularized and subsequently repopulated with viable
cells to generate scaffold sheets possessing natural structure of tissues. Kim et al. extracted
ECM from human adipose tissue and explored the potential application of this natural substrate
as a scaffold for other types of tissues 40. Decellularized ECM sheets (26mm [L] × 10 mm [W] ×
0.15mm [δ]) were seeded with a variety of human cells including dermal fibroblasts, aortic
smooth muscle cells, chondrocytes, umbilical vein endothelial cells, and adipose-derived stem
cells, and the respective cells attachment and proliferation were assessed.
1.1.4 Bioprinting
Ink-jet printing technologies has gained increasing attention over the last decade for applications
in regenerative medicine 26,27,41-51, with the first international workshop on bioprinting and
biopatterning taking place at the University of Manchester in September 2004. Bioprinting
7
technologies have emerged from the adaptation of rapid prototyping strategies used to construct
structures made generally of hard plastic for rapid casting 52, ceramics for dental prostheses 53,54,
or the low-cost production of laboratory components 55 and microwells 56 that would have
normally required dedicated infrastructures. These have been adapted to manufacture, through a
layer-by-layer bottom-up approach, scaffolds onto which cells could attach and grow until they
were mature enough for implantation into the body. However through further development, a
more appealing approach is to simultaneously print living cells and scaffolding materials. Early
brioprinters were adapted from existing desktop printers 45,57, and used to create either droplets
of biomaterials and cell suspensions that are dispensed by heating and vaporizing the bioink 58,
mechanical actuation using a piezoelectric element 51,59, or microextrusion of a filament 60.
These bioprinters can be used to create heterogeneous architectures either by controlling the
micropositioning of specific cell types within a plane 43,44,57,61 , or to generate a biological
template that recreates the cellular environment 51,60,62. Direct cell printing can be achieved by
printing cell-adhesion proteins and monoclonal antibodies onto a substrate material and allowing
the cells to specifically attach to patterned regions 57,59,61,64. Roth et al. adapted ink-jet printing
technology to pattern collagen films with a resolution of 350µm onto glass slides dip-coated in
agarose 44. Regular printer cartridges were rinsed thoroughly with ethanol and sterile water prior
to use. Smooth muscles cells and neural cells were seeded onto the protein-patterned glass slides
and their culture monitored over time, showing preferential attachment to the protein patterns.
Boland et al. extended the cell patterning approach from seeding onto a planar surface to
embedding within a gel layer, creating a three-dimensional microenvironment for the cells. Their
approach relied on the sequential deposition of collagen gel and cell suspension 41. Specifically, a
layer of collagen type I was first coated onto a hosting substrate, with thickness ranging from
200-500µm. Bovine aortal endothelial cells were suspended in culture media and subsequently
printed onto the collagen layer, followed by coating with a final layer of collagen. This process
relies on the cells to migrate into the collagen gel, form clusters and aggregate within the
construct. They found the collagen layers to be essential in promoting fusion of adjacent cell
aggregates. They attempted to reproduce this result by replacing the collagen layers with a
thermo-reversible gel and showed successful cell fusion, although not as effective as the collagen
case. One of the limitations of direct cell printing lies in the need for the cell suspensions to
remain stable and homogeneously distributed over several minutes (beyond 20minutes). Cell
8
agglomeration or sedimentation affecting the printing performance is a common occurrence with
this approach 59. In addition, studies have shown that the dispensing pressure through a
microscale nozzle during printing has more significant effect on the cell viability than the nozzle
diameter, with 38.75% reduction in cell viability when printed at 40 psi as opposed to 5 psi 63.
Complex hierarchical structures mimicking anatomical features of the body were first created by
Vacanti et al. through a molding-casting strategy utilizing alginate gels, creating various shapes
such as human ear 65, trachea 66, nasal tip 67, and nasal septum 68. Cohen et al. utilized printing
strategies to form alginate-based geometries with a surface roughness of 160 ± 20µm in the z-
direction for cartilage tissue regeneration 62. Alginate solutions at a concentration of 2%w.t.
containing chondrocytes at a concentration of 33 × 106 cells/mL were ejected through a 0.84mm
diameter tip at a rate of 0.6mL/min. Gelation of the cell-suspended alginate solution was initiated
by mixing with 10mg/mL of CaSO4 as a crosslinker prior to introducing into the printer. Due to
the time-dependent gelation, this lead to a limited time frame for optimal printing of only 15min.
More recently, Hockaday et al. successfully printed heterogeneous aortic valve scaffolds a
combination of 3D-printing and photocrosslinking strategy 60. Native anatomic and axisymmetric
aortic valve geometries were printed with poly-ethylene glycol-diacrylate (PEG-DA)
supplemented with 10-15%w.t. alginate, which was found to increase sufficiently the bio-ink
viscosity for efficient extrusion. A range of PEG-DA with alginate blends were considered and
showed a wide range in elastic modulus from 5.3 ± 0.9 to 74.6 ± 1.5kPa. Heterogeneous material
printing was achieved by interchanging the bio-ink solutions, enabling a stiffer material to be
printed for the aortic root wall, and a more compliant and extensible material for the leaflets. The
print time ranged from 14 to 45min, depending on the size of the aortic valves, with improved
shape fidelity from 66.6% for small valves (12mm inner diameter) to 93.3% for larger valves
(22mm inner diameter). Porcine aortic valve interstitial cells were subsequently seeded onto the
scaffolds and cultured for up to 21 days.
Other complex structures were also created by printing cell-free and cell-loaded alginate droplets
and building-up into branching constructs 600µm wide and approximately 500µm tall, with
roughly 90µm wide channels throughout the structure 51. Alginate droplets 50µm in diameter
were deposited onto a hydrated gelatin substrate acting as a Ca2+ reservoir (50mM Ca2+
concentration). The Ca2+ diffused upwards into the printed droplets, inducing gelation and
simultaneous fusion of the alginate beads to one another. Due to the coalescence of partially-
9
gelled droplets, the printed features are often loss when droplets are deposited in a row-by-row
process. They adapted their printing sequence by splitting the planar matrix into four sequential
but alternating patterns, such that each printed droplet would be allowed to gel for tens of
seconds before a neighboring droplet was added.
Printed bio-ink droplets containing viable cells were patterned similarly using a layer-by-layer
approach onto bio-paper consisting of a layer of collagen type I from rat tail 39,55. The technique
takes advantage of the printed cells generating their own extracellular matrix and fusing to one
another into three-dimensional constructs 43,69. The bio-ink particles containing cell suspensions
are made of cell pellets that are incubated to form what they call a cellular “sausage” 69. These
were cut into cylinders which were left overnight to round into spheres 500µm in diameter. The
cell spheroids were printed one-by-one, and the collagen bio-paper was essential between each
spheroid layers to promote fusion of the individual cell beads. Since the particle fusion relies on
cell movement, too rapid cell motion resulted in fragile structures, with distortion resulting from
uneven gelation of successive collagen sheets. This resulted in reduced precision beyond a few
layers, with the need to remove the collagen sheets after printing proving to be challenging.
Norotte et al. created vascularized constructs using the same cellular spheroids printing
approach 46. Improved fusion time between the cellular building-blocks was achieved by
preparing cellular cylinders rather than spheres (2-4 days fusion versus 7 days). These were
prepared by printing cell-suspensions of a bio-ink into non-adhesive Teflon or agarose molds,
and incubation overnight. Template agarose rods were 300-500µm in diameter were prepared
and simultaneously printed with the cell-populated rods, requiring their manual removal to
produce vascularized networks. Final constructs resulted in tubes with 900µm inner diameter and
300µm wall thickness. The need for agarose rod removal post printing limits the complexity of
the vascular networks that can be printed. In addition, the additional time required to prepare a
large number of cellular spheroids and rods before printing, and a wait time of up to a week for
the fusion of spheroids makes this printing strategy timely. In addition, the large diameter of the
cell spheroids and cylinders lead to apoptotic cells throughout the vascular construct after 3 days
of fusion.
10
2 Chapter 2 - Formation of Aligned Soft Material Sheets
2.1 Fibrillar Structure of Materials in Nature
Nature possesses the unique ability to organize tissues with respect to their cellular and material
composition. In plants, animals and humans, biological tissues possess a hierarchical
organization of their extracellular matrix with characteristic length scales that often extend across
six orders of magnitude: from macromolecular length scales of tens of nanometers to tissue
length scales of tens of millimeters. In several tissues, the multiscale organization of the
extracellular matrix relies on a high degree of molecular alignment to satisfy critical functional
requirements 1. The molecular composition and mechanical properties of selective biological
tissues are summarized in Table A1. For instance, palm trees have been able to survive tornados
due to their fiber-reinforced composite structure 71-73. Many mammalian tissues are characterized
by fibrillar alignment, e.g., blood vessels 74,75, the cornea 76-78, skin79-81, and the tendon 1,82,83.
A key contributor to achieving the tensile properties associated with connective, epithelial, and
muscular tissues found in blood vessel, skin, and cardiac tissues, is associated with the multiscale
organization of collagen. Collagen accounts for 25-35% of the total protein mass in mammals
and is one of the main components of the extracellular matrix (ECM) 75,84. The collagen family
consists of 28 different proteins, with type I represents over 90% of the weight of all collagen in
humans 75,84. The mechanical properties of tissues, i.e., elastic modulus, elasticity, and strength,
are highly tissue-specific and strongly influenced by the collagen fibrils 85. Fibrils are composed
of three polypeptide strands (alpha peptides), each in the form of a left-handed helix. The three
helices self-assemble into a right-handed triple helix through a process dictated by the
distribution of polar charged and hydrophobic amino acid residues in each strands, forming the
collagen molecule that is approximately 300nm in length and 1.5nm in diameter 86,87. These
molecules self-assemble through an entropy-driven process known as fibrillogenesis, forming
fibrils with diameters ranging from 20nm to 70nm 75. The collagen fibrils in series are separated
by a gap zone of approximately 40nm, and the adjacent collagen fibrils within these fibers are
staggered from one another with a periodicity known as the D-period that is approximately 54-
67nm long 84. Collagen fibers have diameters between 10nm and 300nm and constitute of
microfibril functional units 86.
11
2.2 Collagen Organization in Tissues
The higher order organization of collagen fibrils varies between tissues, and its structure, size,
and orientation are closely linked to functional tissue characteristics. The different types of
collagen organization can be divided into two general categories: those subjected to
unidirectional tensile stress, and those subjected to multidirectional stresses75. The first group
consists mostly of large, heterogeneous fibrils that are tightly packed in parallel and subjected to
tensile stress in the direction of their axis. These include highly tensile structures such as
tendons, ligaments and bone. The second group consists of small, homogeneous fibrils arranged
in helical wavy bundles forming a three-dimensional network, enabling them to be tough yet
highly compliant to multidirectional stresses. Tissues falling within this category include the
cornea, blood vessel walls, skin and nerve sheaths.75
Collagen fibers in the tendon follow a straight and highly aligned organization along the
longitudinal axis of the tendon. These collagen bundles however possess a microscopic crimped
morphology travelling parallel to the axis 82,88. Most of the tendon structure is made up of
collagen type I (>95%) 88, which molecules are arranged and polymerized as fibrils, fibers, and
fiber bundles 83. The organization of these bundles in terms of packing density, diameter, and net
orientation is a determinant factor of the tendon function 82. In the rat tail tendon, collagen fibers
are aligned with the long axis of the tissue and grouped into bundles approximately 61.9-
123.8µm in diameter 82,88,89. These bundles form thick, straight, parallel fascicles 80-320µm
diameter, each possessing a planar-zig-zag structure with a wavelength of approximately
175µm 75,90. The crimp angle was found to be around ±20 ̊ to the longitudinal plane of the
fascicles, with the angle decreasing to ±12 ̊ as it approaches the center of the fascicles 89. At
small strains (2-13% depending on the age of the mice), a very small stress (up to 2MPa) is
sufficient to elongate the tendon by straightening the crimped architecture 90. Under larger
strains, occurring closer to the center of the fascicle, the stiffness of the tendon increases
considerably with extension 1,90,91.
Collagen fibers found in the cornea follow a highly aligned morphology 73,76,78. The cornea is
composed of an epithelium and endothelium separated by a tough collagenous stroma about
500µm in total thickness 77. This dense and organized collagen structure (mainly composed of
12
type I and V) makes up 90% of the corneal volume and 70% of its dry weight, and is arranged in
orthogonal lamellae 76,78,92. Each lamella contains a highly oriented array of collagen fibrils of
homogeneous diameter. In the bovine cornea, these fibrils are around 36nm, and composed of
approximately 4nm diameter microfibrils that are tilted by about 15 ̊ to the fibril long axis in a
right-handed helix 73. The human cornea is composed of smaller fibrils that are 15.4 ± 0.5nm in
persons younger than 65 years of age, and 16.1 ± 0.5nm in persons older than 65 93. In humans,
lamellar size varies considerably as a function of depth within the stroma. Anterior lamellae are
0.5-30µm wide and 0.2-1.2µm thick, whereas those situated in the posterior stroma are 100-
200µm wide and 1.0-2.5µm thick 92. As a result of this dense collagen fibril packing, the stromal
layer of the cornea possesses an architecture that effectively resists tensile loads and internal
swelling pressures of up to 60mm Hg 77, giving the human cornea an elastic modulus of
15.9 ± 2.0MPa and UTS of 3.3 ± 0.2MPa 94,95.
In other tissues such as skin and the intestine wall 75,86,96-98, the wavy collagen fibers are not
aligned in a preferential direction and instead form an almost random woven network. Skin
consists of an upper layer, the epidermis, and a lower layer, the dermis. In the latter, 85% of the
dermal content consists of collagen type I, which plays an essential role in maintaining the
strength and elasticity associated with intact skin 80. Unlike the tendon, the collagen fiber
organization in skin is more random with a complex network of interlaced fibrils that provides
structural support to the epidermis and gives skin its firmness 84,97. The ability of the collagen
fibers to reorient allows large extensions of the tissues 99. As a result, the tissue also becomes
progressively stiffer and its orientation of the collagen fibers more aligned with the direction of
stretching.
Collagen present in the vasculature is crucial for the determination of the tensile strength and
stiffness of blood vessels. The vascular wall is mainly composed of collagen type I and III, and
their removal was found to reduce up to 50 times the local stiffness in the aorta, as measured by
atomic force microscopy (AFM) 100. In the rat aorta, collagen fibrils were measured to be
81.7 ± 7.6nm in diameter 101. These collagen fibrils form a “helical” arrangement with a winding
angle of approximately 18 ̊ 75. In addition, these fibrils are organized into wavy lamellae where
the collagen fibers are generally aligned in the circumferential direction. This circumferential
organization is characterized by a crimped structure that, at low wall tension, is elongated and
straightened with increasing wall tension 96,98,102. At low transmural pressures, this capacity for
13
distension without a significant increase in stiffness 102,103, provides physiological levels of
compliance that serve to dampen the pulsatility of blood output from the heart74,75. At high
transmural pressures the vanishing pitch and crimping significantly increase the stiffness and
allow large blood vessels to withstand high burst pressures 103. Large veins display a drastic
increase in stiffness when subjected to pressures beyond 30-50mm Hg as opposed to pressures
beyond 200mm Hg for arteries 102.
2.3 Strategies for Controlling Collagen Self-Assembly In Vitro
Collagen fibrillogenesis depends on electrostatic and hydrophobic interactions that are strongly
dependent on temperature ionic strength, pH and temperature 104. The controlled multiscale
assembly of collagen fibrils in vitro remains a major challenge. Particularly the difficulty in
consistently promoting high degrees of fibrillar alignment and compactness result in low
mechanical properties (e.g. elastic modulus and ultimate tensile strength) compared to native
tissues.
Early protocols for collagen gel preparation favored the inclusion of viable cells. This required
the formation of these gels in culture media, at neutral pH, and cultured in an incubator, thereby
limiting the range of ionic strength, pH, and temperature. This resulted in weak collagen gels
(UTS of 0.01MPa and E of 0.1MPa) that required from 3 weeks up to 3 months of culture until
cellular remodeling increased the elastic modulus of the construct to levels that allowed
manipulation 105-107. Achilli and Mantovani investigated the effect of varying the ionic strength,
pH and temperature and optimized these parameters to 174 mM salt solution, pH 10, and 4 ̊C to
achieve improved mechanical properties of 10kPa UTS and E of 100kPa 104.
In addition to the magnitude of tensile properties, their directional dependence plays a key role in
controlling cell behavior 108,109. For instance, the presence of aligned collagen fibers in the
vascular wall acts as a signaling factor for platelet activation 75,110. Collagen alignment was also
shown to play an important role in directing cell proliferation and migration in vivo after
injury 111. In order to mimic the natural and tissue-specific structure of the ECM such that both
the mechanical properties and bioinductive aspects of the synthetic scaffold can be tailored, it is
vital to be able to manipulate the spatial organization of the cellular microenvironment.
14
Several reports have focused on achieving various degrees of anisotropic collagen fibril
alignment through: application of anisotropic mechanical forces through fluid flow-induced
shear stress 112-116 or tensional forces 112,117,118, geometric confinement 119, electric currents 120,
magnetic fields 121-126, and electrospinning 74,127-129. The structural and mechanical properties of
the resulting collagen gels produced are summarized in Table A2.
Caves et al. 112 demonstrated the continuous fluid-extrusion of collagen fibers at a rate of
60m/hr. Fibers were collected on a spinning mandrel that exerted a tensional force in the
direction of extrusion and were manually transferred to a collection frame that maintained
tension during incubation and drying. The fibers had a high degree of fibrillar alignment, a cross
section of 53 ± 14µm 21 ± 3µm, UTS up to 94 ± 19 MPa and E up to 775 ± 173 MPa. Lai et al.
combined hydrodynamic shear and translation-induced shear to generated aligned collagen slabs
that were approximately 1mm 1.2mm 30mm. Acidic solutions of collagen with
concentrations ranging from 0.2mg/ml to 0.8mg/ml were extruded from a syringe onto a
translating glass slide immersed in a salt buffer solution at pH 7.4. The hydrodynamic shear
induced alignment of the collagen monomers and fibrillogenesis was triggered from the change
in pH from acidic to neutral. As a result, the collagen formed were highly aligned and resulted in
elastic modulus of 3.6 ± 1.9MPa 116.
Cell-laden collagen gels have been subjected to anisotropic mechanical forces for the formation
of aligned collagen gels with anisotropic mechanical properties. Thomopoulous et al. constrained
fibroblast-laden collagen sheets (4cm × 4cm) uniaxially or biaxially for 72h and reported gel
compaction in the unconstrained axis, with an obvious structural and mechanical anisotropy 118.
The authors found collagen gels to develop anisotropic molecular alignment under uniaxial
tension even in the absence of cells.
Geometric confinements have also been utilized to define collagen fiber alignment during
fibrillogenesis. Lee et al. used this approach to generate three-dimensional collagen gels with
aligned fibers for the investigation of cell signaling in vitro 119. Collagen gels were cast and
geometrically confined within 1cm long, 40µm deep, and 10-100µm wide microfluidic channels.
Under static flow conditions at room temperature, the first fibers appeared 10min after casting
and fibrillogenesis was completed after 30min. These narrow microfluidic channels resulted in
15
the formation of 20µm long collagen fibers that were 20-40% aligned within 5 degrees from one
another.
Electrochemical cues have been similarly employed to create aligned and densely packed
elongated collagen bundles. Cheng et al. used parallel wire electrodes to apply electrical currents
(current density 0.55A/m2, nominal electric field strength 2.5kV/m) to collagen solutions spread
onto a glass slide. The currents created a pH gradient within the footprint of the glass slide,
causing the collagen molecules to migrate and assemble within a plane 120. This approach
enabled the formation of single collagen bundles 50-400µm in diameter and 3-7cm in length
depending on the length of the electrodes, achieving half the mechanical strength of the native
tendon (UTS of 24-88MPa and E of 277-671MPa).
Due to the small negative diamagnetic susceptibility of collagen molecule, the exposure of
collagen solution to a strong static magnetic field on the order of a few teslas (T) during gelation
aligns collagen fibrils perpendicular to the direction of the field 122,130,131. Torbet et al. subjected
collagen solutions to a 7T magnetic field for a duration of 30min in an incubator at 37°C 126. The
authors repeated the process by rotating the collagen samples along the z-axis and adding an
ungelled layer of collagen solution onto the already gelled and aligned collagen films. This
resulted in the formation of a laminated structure of varying fibril orientations that mimicked the
organization of the stroma in the cornea. Others have demonstrated the effect of interstitial fluid
flow on collagen gels, where initially entangled fibers were successfully disentangled and
aligned perpendicular to the flow field 132,133. Guo et al. combined magnetic field and fluid-flow
effects to create collagen films with aligned microfibrils 122. They incorporated 2.5µm diameter
magnetic beads within collagen gels, utilizing the movement of the beads attracted to an external
magnetic field and fluid-flow to pull the collagen fibers during gelation.
Electrospinning is a well-established technique used by many researchers to prepare nanofibrous
matrices for tissue engineering applications 74,128,129. Although the resulting fibers are generally
random in orientation, it is possible to control the degree of collagen fiber alignment by
collecting the spun fibers onto a rotating mandrel 74,128. Oryan et al. created highly aligned
collagen nanofibers by combining electrospinning with magnetic fields to reproduce the collagen
fibrils in the tendon 127. The film of electrospun collagen fibers formed was subsequently
embedded into collagen solution and polymerized at 4 ̊C for 48hr under 12 T magnetic fields to
16
induce fibril alignment. Final collagen gel composite were cut to the size of rabbit’s Achilles
(L = 2cm, H = 3.5mm, W = 3mm) before implantation.
These approaches promote collagen alignment in small dimensions between 0.05-3mm, but with
a limited degree of fibril alignment and packing density the gelled collagen has a low mechanical
strength in comparison to native tissue (E and UTS up to 775MPa and 94MPa respectively 112,
Table A2), and often require hours to complete gelation 134,135. There is therefore a currently
unaddressed need for the controlled alignment of anisotropic matrices in centimeter scale in
several tissue engineered applications.
2.4 Strategy for the Continuous Formation of Collagen Sheets with Aligned Nanostructure
Here, we present a method for the continuous formation of nanofibrous collagen sheets with
precise control over the sheet thickness, width, and the degree of fibril alignment. Our strategy is
illustrated in Figure 1, and consists of the continuous in-flow formation of collagen sheets using
a multilayer microfluidic device with a flow constriction that attached downstream. A layered
fluid consisting of a central sheath of collagen solution sheathed by streams of PEG solutions on
the top and bottom, simultaneously exited the microfluidic device. Downstream of the device
exit, the focusing and collagen solutions were hydrodynamically focused in the sheath-normal
direction, collagen fibrils were formed and aligned in the flow direction and cross-linking
progressed starting from the top and bottom boundaries where the collagen solution was in direct
contact with the focusing PEG solution. A flowable thin sheet of cross-linked collagen was
formed and an initial compaction of the collagen solution achieved. A strain was applied on the
partially cross-linked sheet when collecting it onto a drum at a location downstream of the
confinement. We report the obtained changes in the tensile properties of the collagen sheet as a
function of the degree of fibril alignment and compaction, and demonstrate the scalability of this
approach to form meter-long highly aligned collagen sheets with very large aspect ratios (sheet
width divided by thickness) of up to 400.
17
2.4.1 Materials and Methods
2.4.1.1 Experimental Setup
Figure 2a shows a schematic illustration of the experimental setup, which consists of a multilayer
microfluidic device, a constriction manifold and a rotating drum. Two syringe pumps
continuously supply the collagen solution and the focusing solution at their respective flow rates,
QM and QF. The multilayered microfluidic device distributes through microchannel networks in
separate device feature layers (Fig. 2b) a focusing solution (top), a collagen solution (center) and
a second focusing solution (bottom). These solutions meet at the device exit, form a layered or
sheath flow, and are hydrodynamically focused in the sheath-normal direction while passing
through a downstream flow constriction with a gap height of HC =1mm (Fig. 2c). Prior to the
constriction, an opening gap HG = 4mm was introduced to allow some degree of freedom in the
alignment of the device exit region with the constriction manifold. The continuously produced
collagen sheet was pulled onto a rotating drum, decreasing the sheet thickness and enabling the
degree of collagen fibril alignment and their packing density to be precisely tailored. The flow-
focusing constriction plays three roles in our system by (1) inducing through hydrodynamic
focusing the alignment of collagen in its monomeric state, by (2) preventing through
hydrodynamic focusing any unwanted deformation of the collagen liquid sheet that would be
expected in the case of an elastomeric substrate material 136, thereby maintaining thickness
uniformity of the collagen sheet, and by (3) stabilizing the co-flow of the focusing and collagen
solutions by isolating the co-flowing system from any flow recirculation within the liquid filled-
reservoir created by the rotating drum. The microfluidic device and constriction manifold are
immersed in a reservoir filled with the same liquid as the focusing solution. As the collagen and
focusing solutions meet at their common interface, the specific composition of the focusing fluid
(10% w/v PEG at pH 8) triggers the gelation of collagen, thereby fixing the structural change
imposed on the collagen sheet. The composition of the focusing fluid was chosen after numerous
optimization studies by our collaborators from Dr. Chaikof’s lab for optimal and rapid collagen
fibril formation 112,137,138. Such manipulation of the material structure results in the formation of
collagen sheets with a wide range of mechanical properties directly linked to the degree of their
fibril alignment and packing density.
18
2.4.1.2 Fabrication of Microfluidic Device and Constriction Manifold
The microfluidic device was fabricated using standard soft-lithography techniques 139 and
consists of three polydimethylsiloxane (PDMS) layers that were individually fabricated and
subsequently bonded to form the final multilayered device. The top and bottom layers distribute
a flow-focusing solution, while the middle layer distributes an acid solubilized collagen solution
(Fig. 2b). On the top and bottom sides, the collagen solution is confined by a layer of the
focusing solution as it exits the device and enters a liquid-immersed constriction unit.
Hydrodynamic focusing takes place at a location downstream of the microfluidic device within a
constriction unit that is 12mm wide, has a LG=2mm long section with gap height HG=4mm and a
LC=6mm long flow constriction. The horizontal distance between the end of the constriction unit
and the edge of the rotating drum (LP) (Fig. 2c). We machined a device-external constriction unit
in aluminum in order to retain a uniform constriction height, HC, across the 1:12 aspect ratio slit,
and thereby avoid any unwanted deformation that would be expected in the case of an
elastomeric substrate material 136. The constriction unit also prevented any flow instabilities
within the co-flowing system caused by flow recirculation induced by the rotation of the drum
within the liquid-filled reservoir. The constriction gap was horizontally aligned and tightly sealed
against the exit section of the microfluidic device. The value of HG exceeded slightly the height
of the device exit section by approximately 2.5mm to ensure fluids from all three layers are
consistently guided through the constriction.
2.4.1.3 Collagen Gelation through Molecular Crowding
In this work, we rely on a phenomenon known as molecular crowding to initiate the rapid
gelation of collagen at room temperature, enabling physical handling of the collagen sheets
immediately after they exit the microfluidic device. In past studies, the assembly and
disassembly of biomolecules of cytoskeletal filaments have been investigated 140, and revealed
that the self-association of molecules into organized bundles of actin filaments is highly
influenced by volume exclusion and confinement provided by the crowded cell environment 141-
143. Cuneo et al. utilized 6000kDA polyethylene glycol (PEG) molecules to mimic the
macromolecules of the cell and found that concentrations of PEG between 6 to 7% w/v resulted
in massive conversion of actin filaments into bundles 140. Self-assembly and organization of
19
collagen fibrils can be controlled in vitro through either geometric confinement that concentrates
pure collagen monomers, or in open spaces using fibril forming buffers. These fibril forming
buffers essentially consists of a solution containing macromolecules that occupy a significant
fraction of the volume of the medium, thereby commonly referred to as ‘crowded’ solutions.
These can be produced by adding high concentrations of a synthetic or biomolecular co-
nonsolvent such as polyethylene glycol or hyaluronic acid 143,144. Due to the large size of the
crowding molecules, little free space is left in the solution, causing volume exclusion and
confinement, thereby forcing the collagen molecules to align themselves in order to maximize
entropy.In 1994, Cavallaro et al. adapted a collagen fibril extrusion developed by Kato et al.
where collagen was extruded into a fibril forming buffer 145. Cavallaro have used polyethylene
glycol (PEG) as the fibril forming buffer and demonstrated that molecular crowding of the
collagen monomers was induced by this hypertonic environment (i.e. higher osmotic
pressure) 144. The osmotic pressure difference between the PEG and collagen solution generates
an osmotic pressure gradient, leading to the dehydration of the collagen sheath through depletion
forces, which results in compaction and coagulation of the collagen. As a result, the collagen
fibers produced were denser and could therefore be continuously extruded without rupture.
Following this observation, other groups have utilized the same approach to generate organized
collagen fibers 112,143,146-148. In our approach, in addition to PEG as a molecular crowding agent,
the focusing solution used is also pH and salt-balanced (see materials and methods section),
ensuring the simultaneous self-assembly of tropocollagen molecules (consisting of a triple helix
composed of three alpha-peptide strands) into aligned fibers through the process of
fibrillogenesis. As the collagen sheath is focused into the constriction, hydrodynamic flow
focusing and strain-induced pulling align the collagen in its monomeric state. The process of
fibrillogenesis occurs through a longer time period through the pH and salt concentration change
as they diffuse from the focusing solution into the collagen sheath. In recent works, salt and pH
induced fibrillogenesis was observed to form after 10min at room temperature 119.
2.4.1.4 Isolation and Purification of Monomeric Collagen
Acid-soluble, monomeric rat-tail tendon collagen (MRTC) was obtained from Sprague-Dawley
rat tails following Silver and Trelstad 149. Frozen rat tails (Pel-Freez Biologicals, Rogers, AK)
were thawed at room temperature and tendon was extracted with a wire stripper, immersed in 10
mM HCl (pH 2.0; 150 mL per tail) and stirred for 4 hr at room temperature. Soluble collagen
20
was separated by centrifugation at 30,000g and 48 ̊C for 30 minutes followed by sequential
filtration through P8, 0.45 µm, and 0.2 µm membranes. Addition of concentrated NaCl in 10 mM
HCl to a net salt concentration of 0.7 M, followed by 1 hr stirring and 1 hr centrifugation at
30,000g and 48 ̊C, precipitated the collagen. After overnight redissolution in 10 mM HCl the
material was dialyzed against 20 mM phosphate buffer for at least 8 hr at room temperature.
Subsequent dialysis was performed against 20 mM phosphate buffer at 48 ̊C for at least 8 hr and
against 10 mM HCl at 48 ̊C overnight. The resulting MRTC solution was stored at 48 ̊C for the
short-term or frozen and lyophilized.
2.4.1.5 Preparation of Collagen Neutralization Buffer
The rapid gelation of collagen sheets during extrusion was induced by the addition of
polyethylene glycol (PEG) in the focusing solution. PEG triggers collagen molecular crowding
and gelation150, a phenomenon only seen in a much smaller scale during collagen wet spinning
(~10s of microns), but yet demonstrated on a macro-scale 112,137. The focusing solution consisted
of a neutralization buffer, which that contained 10%w.t. PEG (MW 35kDa), 4.14mg/mL
monobasic sodium phosphate, 12.1mg/mL dibasic sodium phosphate, 6.86mg/mL TES, and
7.89mg/mL sodium chloride.
2.4.1.6 Collagen Sheet Incubation and Drying
After collagen extrusion and pulling onto the rotating drum, the sheets were collected and
immersed in flow focusing for 1 hr, after which they were washed three times with ddH2O.
Sheets were subsequently incubated in a fiber incubation buffer (FIB) (7.89 mg/mL sodium
chloride, 4.26 mg/mL dibasic sodium phosphate, 10 mM Tris, pH 7.4) at 37 ̊C for 48 hr.
Following incubation, the collagen sheets were rinsed in ddH2O for 1 hr and dried on a glass
slide under constant forced air flow.
21
2.4.2 Results
The presented approach enables the continuous formation of collagen sheets with a controlled
width, w, thickness, , and angle of fibrillar alignment, . The sheet width was determined at the
drum from measurements performed with three microfluidic that had the exit widths: w0=5mm,
10mm, and 25mm. The thickness and fibril alignment of the collagen sheets depended on the
following experimental parameters: the collagen flow rate, QM, the flow rate of the focusing
fluid, QF, and the pulling velocity, VP. In the following, we will assess the roles of hydrodynamic
focusing and stretching on the formed sheets.
2.4.2.1 Flow Confinement
We hypothesize that the initial thickness reduction is dependent on the re-organization of the
focusing and collagen streams while they pass through the confinement. In order to
experimentally characterize how the local sheet thickness locally varies at different locations
downstream of the exit section of the microfluidic device while the collagen sheath flows
through the constriction and pulling-induced strain is being applied, a separate constriction unit
was fabricated. The experimental setup consisted of a vertically arranged microfluidic device,
provided for visual access within the constriction and allowed the collagen sheet thickness
variation to be imaged in the (x, z)-plane (Fig. 3a), using an inverted microscope. Figure 3b
shows bright-field images of the exit region of a microfluidic device placed within the vertically
positioned constriction unit (top), and of a collagen sheet being formed within the constriction
(bottom). The bottom image was captured for a device with w0=10mm at conditions QM =
100µl/min, QF = 1ml/min, V* = 4.5. Here, we introduce a non-dimensionalized velocity
parameter V* obtained by relating the pulling velocity with the total velocity of the working
fluids. Specifically V* = (VP – VTotal)/VTotal, where VTotal = (QF + QM)/AConst, and the cross-
sectional area at the constriction AConst = W × HC. The flow profile of the focusing solution
within the confinement was further visualized by incorporating fluorescent microbeads (Nile red
carboxylate microbeads 1µm in diameter) at a concentration of 0.08% v/v. Long-term exposure
images (exposure time 400ms) captured the streamlines within the two regions of interests that
are indicated in Figure 3c. Specifically, streamlines within the entrance region of the sheath flow
entering the constriction (window 1) and the upper wall of the chamber before the constriction
22
(window 2) were investigated. The images shown in Figure 3c-1 were obtained at
QM = 100µl/min, QF = 1ml/min, V* = 10 and illustrate the streamlines of the focusing fluid travel
parallel to the moving collagen sheath in its proximity. In Figure 3c-2, no collagen was flown
through the microfluidic device and QF = 1ml/min. The presence of recirculating flows can be
observed in the upper wall of the open region before the constriction. The size of the
recirculation zone decreased when increasing QF from 1ml/min to 6ml/min (see Appendix Fig.
A2). However, the recirculating vortices do not interact with the collagen sheath, and follow a
laminar flow-profile with linear streamline parallel to the flow direction as they pass through the
constriction region, suggesting that the formation of collagen with consistent control over the
width and thickness is unaffected by their presence. A numerical model of the velocity profile
within the constriction region was developed using a multiphysics solver based on the finite
element method (COMSOL). The collagen sheet was treated as a moving wall with velocity
equal to the pulling velocity and no-slip boundary condition. The focusing solution (10% w/v
PEG) was treated as a Newtonian fluid with a density of 131mg/mL and constant dynamic
viscosity of 54.4cP confirmed from viscosity measurements under increasing shear rates using a
rheometer (DV-III Rheometer, Brookfield, Massachusetts, US) (see Appendix Fig. A3a). The
velocity profile of the focusing solution was investigated, with QF = 1-16ml/min and
VP = 2mm/s, with Reynolds number ranging from 0.02-0.32 confirming that the experimental
conditions remain within the laminar region (fig. A3b,c). In agreement with the experimental
data, the numerical data showed no backflow or recirculation within the constriction region HC.
2.4.2.2 Inducing Alignment
We experimentally investigated the increase of fibril alignment and compaction within the
collagen sheet with increasing strain being applied by the pulling drum. The vertically arranged
microfluidic device and constriction unit were used to produce collagen sheets at varying pulling
velocities V* while was measured from bright-field microscopic images. Figure 3f shows the
results obtained at different values of V* and QM = 100µl/min, QF = 0.5 and 1ml/min.
Measurements were taken at four different streamwise locations within the confinement: (A) the
microfluidic device exit, x = 0, (B) x = 0.5LG, (C) x = LG+0.5LC, (D) x=LG+LC+LP. For all
considered values of V*, as the focusing flow rate QF increases, the overall thickness of the
collagen sheet decreases, suggesting a focusing and stabilizing effect induced by the focusing
solution. For both QF at low V*= 0.1-2, the collagen sheet thickness increases slightly as it exits
23
the device into the opening region within LG (B), similar to the ‘die swell effect’ observed in
flow extrusion where the melt expands as it exits from the die 151,152. However, at higher pulling
velocities, V* > 2 (VP>5mm/s), the sheath thickness monotonically decreases. We attribute this
effect to the rapid onset of collagen fibril formation and gelation locally within the regions of the
collagen sheath that are in direct proximity to the PEG solution. We assume molecular crowding
to induce the rapid gelation of collagen within the thin outermost regions of the collagen sheath,
resulting in the thickness of the collagen sheath along its entire length to be affected by an
increased pulling velocity exerted at the drum.
2.4.2.3 Sheet Formation
The effect of the constriction unit on sheet thickness and width was investigated experimentally
by comparing results of collagen sheets formed with and without the use of the flow-focusing
unit. The experiments were conducted using three devices with w0=5mm, 10mm, and 25mm. For
the device with w0=5mm conditions QM=50µl/min, QF at 1ml/min, and VP = 1-20mm/s were
applied. In the case of the two other devices, the same range of VP was considered, and QM and
QF were adjusted proportionally with the increase in device width (i.e., the flow rates were
twofold higher in case of w0=10mm, and fivefold higher in case of w0=25mm), and the
corresponding V* were calculated accordingly The use of a constriction unit produced wider and
thinner collagen sheets. Collagen sheets formed without constriction were between
0.65±0.21mm and 3.3± 0.17mm wide. With a constriction they were at the same flow rates
between 3.3± 0.09mm and 17.3 ± 0.1mm wide (Fig. 4a). The constriction unit reduced by up to
88% the thickness of the produced sheets, from =260±8µm to 1140±10µm without and from
30±3µm to 213±15µm with constriction (Fig. 4b). The measurements of the external sheet
dimensions w and for all three devices were non-dimensionalized by w0 and HC, respectively.
The self-similarity of the results demonstrates the utility of the approach for the predictive
formation of a large aspect ratio collagen sheet with a certain target width, by selecting a
microfluidic device with an appropriate width w0. Sheet dimensions w and , were studied for
w0=10mm, V*=0.1-10, QM=100µl/min, and QF=1-6ml/min (Fig. A4). The obtained data suggest
a decrease in both width and thickness for an increasing flow rate of the PEG solution, QF, with
w/w0= 0.32-0.8, and /HC=0.025-0.3.
24
2.4.2.4 Nano and Microscale Properties
At a distance LP downsteam of the confinement unit, the formed collagen sheet is being collected
on a rotating drum. Along with the applied flow rates QF and QM, the speed of drum rotation, V,
or the corresponding dimensionless parameter V*, affect not only sheet dimensions but also the
alignment of collagen fibrils and their packing density. The cross-sectional area of the wet
collagen sheets, w , was calculated, plotted against V*, and compared to the calculated cross-
sectional area. The calculated values were obtained from QM/VP (Fig. 4c). Data were obtained
from experiments were conducted with three device widths, s w0=5mm, 10mm and 25mm; where
QM = 50µl/min (w0/5mm) and QF=1.5ml/min (w0/5mm). For values of V* below a threshold V*th,
the measured cross-sectional area exceeded the one predicted under the assumption of a
conserved volume. For V* > V*th, the opposite case was observed, suggesting compaction of the
collagen sheet. The degree of compaction was estimated by comparing the measured cross-
section of the hydrated sheets with the initial sheet cross-section, equivalent to w0HC. Values of
compaction were found to range from 3.3 % to 95.5 % (Fig. 4d). This can be explained by the
relationship between the flow rates QM and QF with the pulling velocity VP. At an initially low
VP, the average total velocity of the collagen and focusing solutions through the constriction is
larger than the pulling velocity, suggesting that the fibril alignment is solely due to
hydrodynamic focusing and no strain is being exerted by the drum rotation. However, once VP
exceeds the average velocity of the collagen sheath leaving the microfluidic device, a strain is
applied by the pulling drum, that causes the alignment of fibrils along the length of the sheet, a
reduction of the average fibril-to-fibril spacing and a contraction of the sheet.
The degree of compaction and fibril packing density was also characterized based on
transmission electron microscopic (TEM) and scanning electron microscopic (SEM) images of
dried collagen sheets. Collagen samples were produced across a range of V* were examined, and
the TEM and SEM images revealed the degree of fibril alignment and packing density with an
increase in fibril packing density and alignment observed with V* increasing from 0 to 10 (Fig.
5a, b; see Fig. A6 for TEM and SEM of samples obtained at V* = 0.6 and 4.5). In addition, D-
period banding typical of collagen fibers can be observed in both TEM and SEM images of the
highly aligned collagen sheets produced (Fig. 5a-2, b-2). The degree of compaction was
measured by image analysis of the SEM images of collagen sheets formed at V* = 0.1, 0.6, 4.5,
and 10. An autocorrelation function was calculated for the intensity distributions in SEM and
25
TEM images using the software program Matlab (Mathworks, Econometrics Toolbox, Natick,
MA, USA). Fibril spacing was measured from the resulting plots. As a sample, Fig. 5c shows an
autocorrelation function of the TEM image in Fig. 5a-2 (see Supplement Fig. A7 for all
autocorrelation plots). Fibril spacing for all V* conditions are summarized in Fig. 5d and indicate
a 95.3% decrease, from 139.3 ± 37.3nm for V* = 0.1 to 6.51 ± 1.16nm for V* = 10 (Fig. 5d
insert). It is important to note that the degree of compaction is much larger in the dried samples
as compared to the wet samples. We attribute this difference to the drying process post-extrusion
which further compacts the fibers through water evaporation. The D-period banding of collagen
fibers was calculated by applying an autocorrelation function to line intensity plot obtained in the
x-axis of the SEM image in Fig. 5e (V* = 7). A banding period of 67nm was obtained,
characteristic of collagen fibrils in the body and confirming that triple helical fibrils are formed.
In addition to the degree of compaction and the banding length, fibril alignment of the collagen
sheets was characterized by applying a Fast Fourier Transform (FFT) algorithm to the SEM
images obtained using an image processing software (ImageJ). The percentage of aligned fibrils
was plotted as frequency (%) versus the angle of alignment (Fig. 5f), confirming an increased
degree of alignment with increasing V* from 0 to 10. We have achieved up to 40% alignment of
fibrils within ± 5 ̊ from one another in collagen sheets with aspect ratio w/δ ranging from 15 to
375.
2.4.2.5 Macroscale Properties
The direct impact of fibril alignment on the overall mechanical properties of these collagen
sheets was confirmed through tensile measurements. Samples were prepared and mechanically
tested using an inverted DMTA (Dynamic Mechanical Thermal Analysis). Sample drying prior
to mechanical testing was critical to compact the fibers and prevent gel compaction in the
potential case where cells would be seeded on the surface (thickness drops to 1/10th of the initial
wet state). Sheets were placed on a glass substrate and subjected to a constant air flow (fan) to
accelerate the drying process. Figure A5 demonstrates the effect drying and incubation in fiber
incubation buffer (FIB) have on the mechanical properties of collagen sheets. The dehydration
process is shown to improve the upper tensile strength in both cases with or without incubation
in FIB post-extrusion. On the other hand, the Young’s modulus of samples without FIB
26
incubation was reduced by 67% when dried, while when incubated in FIB, the drying process
brought the Young’s modulus back to its highest value post-extrusion. Mechanical testing of our
FIB-incubated and dehydrated samples was performed by rehydration with a drop of DI water,
lift-off from the glass slide, and immersion in PBS at 37˚C for 30min. Samples were kept
immersed in PBS at 37 ̊C throughout the experiment. As expected, the structural anisotropy
resulted in mechanical anisotropy with increased, significantly higher ultimate tensile strengths
(UTS) and Young’s moduli in the direction of alignment when compared to unaligned matrices.
Specifically, V* between 0.6-10 lead to the formation of sheets with ultimate tensile strengths
between 1.25-13MPa and Young's modulus between 1.3-130MPa (Fig. 5g). These values are
well within the range of those found for native blood vessels, 1.4-11.1MPa (UTS) and
1.5±0.3MPa (E).153
We have demonstrated a high throughput approach for the continuous formation of wide
collagen sheets (width:thickness ratio up to 400) with tunable alignment and compaction of
collagen fibrils. The combination of flow-focusing and strain-induced pulling of the collagen
sheet results in sustained collagen fibril alignment in the direction of flow, with the degree of
alignment and the density of fibrils consistent throughout the entire sheet width, a result
unprecedented at the macroscale. Large aspect-ratio collagen sheets with dimensions that ranged
from 3-12mm in width and 30-250µm in thickness were continuously produced. The degree of
alignment and compaction of the collagen fibrils was controlled, with up to 40% of fibers aligned
within ± 5° of one another, and up to 95.5% of compaction. As a result, these highly aligned
collagen sheets achieved mechanical properties comparable to native native blood vessel, with
E between 1.3-130MPa, UTS between 1.25-13MPa, and strain to failure from 15-35%. In
addition, the presence of D-banding periods of ~67nm typical of collagen fibrils was consistently
observed. We believe the presented strategy to be promising in developing large collagen
substrates of biologically relevant composition and tunable mechanical properties for
applications in tissue engineering such as vascular grafts.
27
Figure 1. Schematic illustration of the flowable conversion of a collagen solution (shown in
orange) to a cross-linked collagen sheet with a high degree of molecular alignment (shown in
red). The illustrated approach involves (1) the reduction of the thickness of the collagen sheath
and initiation of fibril formation and cross-linking during hydrodynamic focusing in the sheath-
normal direction, (2) the application of axial strain between the constriction (blue arrow) and
collection on a rotating drum. The combination of these two steps, along with rapid gelation of
the collagen through molecular crowding, result in the formation of collagen sheets with high
aspect ratio of w/δ from 15 to 375, and highly aligned fibers (3).
28
Figure 2. (a) Schematic illustration of experimental setup consisting of microfluidic device for
preparation of biopolymer sheet, constriction unit, and rotating collection drum to collect and
pull the cross-linked collagen sheet (VP). The drum is located at a distance LP = 20 mm from the
constriction unit. (b) Three-layer multilayer microfluidic device. Layers 1 and 3 distribute
focusing solutions (shown in green) and layer 2 distributes collagen solution (shown in orange)
to the device exit where the three fluid layers are brought in contact and cross-linking is initiated
(symbolically illustrated by the sheet color turning red). (c) A machined confinement unit (LG =
2 mm, LC = 6 mm, HC = 1 mm, HG = 4mm) reduces the thickness of the biopolymer layer, .
Schematic of collagen sheet viewed from the xy plane (top). Schematic of the sheet formation
within the confinement (bottom). Scale bars 10 mm (b) and 2 mm (c).
29
Figure 3. Detailed study of flow-focusing and strain-inducing pulling using a vertical
constriction unit. (a) Schematic of experimental setup for imaging of collagen sheet formation
in the (xy)-plane. Sheat extruded into flow-focusing solution. (b) Bright-field images of collagen
sheet formation using vertically oriented manifold in (a). Images taken at device exit (top) and
within constriction region (bottom). QM = 100µl/min, QF = 1ml/min, V* = 4.5, w0=10mm. (c)
Schematic of regions (1) and (2) investigated using fluorescence microscopy. (1) Flow profile at
30
entrance to constriction: QM = 100µl/min, QF = 1ml/min, V* = 10. (2) Recirculation profile
within focusing solution in entrance region to constriction, QF = 1ml/min. Fluorescent
microbeads added to focusing solution at a concentration of 0.08% v/v. (d) Sheet thickness
measured at four different locations within the constriction manifold. Data obtained with QM =
100µl/min, QF = 0.5ml/min and QF = 1ml/min (*), V*= 0.1, 2, 4.5, 10 (light to dark color bars).
Scale bars 250µm (b), 1mm (c), 200µm (c, 1-2), 2mm (d).
Figure 4. Measured collagen sheet width and thickness as a function of V* ranging from 0.1
to 10. (a,b) Collagen sheet width and thickness obtained with (full line) and without (dotted line)
use of constriction manifold. Three devices of varying exit width (5, 10, 25mm) were used with
the manifold. 5mm wide device was considered in the case without manifold. For the 5mm wide
device, QM = 50µl/min, QF = 1ml/min. QM and QF were varied proportionally with device width.
(c) Comparison of experimental and calculated collagen sheet cross-sectional area as a function
of V* ranging from 0.1 to 10. Dotted lines represent calculated value (QM/VP), while full lines
31
represent experimental values obtained for 5mm (square), 10mm (triangle), and 25mm
(diamond) wide devices. (d) Degree of compaction measured as a percentage change in cross-
sectional area calculated from comparison between experimental cross-sectional areas and cross-
sectional area of the device exit section. Data plotted for all three devices (5mm, 10mm, 25mm)
as a function of V*.
32
33
Figure 5. Characterization of collagen fibril compaction, alignment, and resulting
mechanical properties. (a) TEM images of fibrillar alignment in collagen sheet obtained at V* =
0 (1), 10 (2). (b) SEM images of collagen fiber alignment obtained at V* = 0 (1), 10 (2). (c)
Autocorrelation function of a TEM image of collagen sheets produced at V* = 10 showing an
average spacing of ~6.5nm. (d) Degree of compaction quantified by autocorrelation of SEM
images of collagen fibers obtained at V* = 0.1, 0.6, 4.5, and 10. Compaction quantified as
percent change in fibril spacing in reference to the fibril spacing at V* = 0.1. Results plotted in
comparison to percent change in cross-sectional area in Fig. 4c (insert). (e) Autocorrelation
function of SEM image (insert) showing repeated banding pattern (D-period) of ~67nm. (f)
Collagen fibril alignment obtained from SEM image processing of sheets formed at V* = 0, 0.1,
0.6, and 10. Full width half max (FWHM) summarized in table insert. (g) Young’s modulus (E),
ultimate tensile strength (UTS), and strain to failure (%) of collagen sheets formed by passing
through constriction and subsequent alignment induced by different values of V* = 0.6-10
(*p<0.05). All experiments conducted at QM = 100µl/min, QF = 1ml/min. SEM and TEM images
were obtained by Dr. Grainger. Scale bars 200nm (a), 1µm (b, 1 left), 500nm (b, right), 50nm (b,
2-left), 500nm (e, insert).
34
3 Chapter 3 - Continuous Formation of Patterned Soft Material Sheets Populated with Molecular and Cellular Payloads
Soft materials with a spatially non-uniform composition that is closely linked to their function
are abundant in nature. Such materials often possess a hierarchical architecture that extends from
cell to tissue scales in several directions. In this report, we demonstrate a one-step, continuous
process for the scalable formation of soft material sheets while controlling their local and global
composition. A ten-layer microfluidic device enabled us to first dynamically define mosaic
hydrogels by incorporating within a flowing biopolymer sheet a secondary biopolymer and to
retain the microstructure in a subsequent cross-linking step. The secondary biopolymer was
either a different hydrogel or it carried a biomolecular, colloidal or cellular payload. We
continuously organized hydrogel sheets to 2D and 3D soft material assemblies with millimeter to
centimeter length scales, stored information within unsuspended hydrogel sheets, incorporated
void regions, created mosaic stiffness and diffusivity patterns and populated tessellations with
different viable primary cells. We envision mosaic hydrogels to become continuous, automatable
and physiologically meaningful formats for engineering cell instructive microenvironments and
3D tissues.
3.1 Alginate in Tissue Engineering
Alginate, a naturally derived hydrogel forming polymer, has been widely used as a biomaterial
for tissue engineering and drug delivery applications 154-156, due to its gentle gelling kinetics and
low toxicity 157-159. Alginate is derived primarily from brown algae and is a linear polysaccharide
co-polymer of (1-4)-linked β-mannuronic acid (M) and α-guluronic acid (G) monomers. Within
this polymer, the M and G monomers are sequentially assembled in repeating MM or GG blocks,
or alternating MG blocks 160,161. Factors such as seaweed species, age, and section define the
amount and distribution of these M and G monomers. An attractive property of alginate is its
ability to rapidly form a hydrogel at room temperature and conditions that are mild to living
35
cells. The gelation process is triggered by the cross-linking of alginate molecules with divalent
cations such as Ca2+. These calcium cations preferentially interact with G-block monomers in the
polymer chains to form ionic bridges between adjacent polymer chains, resulting in an “egg-box”
structure. This highly cooperative binding process requires more than 20 G-monomers 162. For
this reason, the ratio of M : G blocks is critical in defining the resulting mechanical strength of
the cross-linked alginate hydrogels. The mechanical properties of various types of alginate
(M : G ratio from 0.3-1.56) at concentrations ranging from 1.5-4% w.t. gelled with different
concentrations and types of cross-linkers have been summarized in Table A4.
In tissue engineering applications, the main function of alginate is to provide mechanical
integrity while transmitting initial mechanical signals to the cells and developing tissue. The
ability to tune the alginate mechanical properties by controlling the cross-linking time and M : G
ratio makes it an attractive material for a variety of tissues. Alginate has been widely utilized as a
scaffold material for engineered cartilage 163-165, bone 166,167, and skeletal muscles 168. It has also
been processed in the form of hydrogel beads to be used as a carrier matrix for encapsulating
molecules such as enzymes, drugs, microbial 169, or viable cells 170,171, with applications as
scaffolds for tissue engineering or the controlled release of drugs and biological molecules.
Bioprinting of alginate microdroplets 50µm in diameter was utilized to construct branched
microvasculatures 51. These were printed onto gelatin substrates soaked in 10mM CaCl2, relying
on the upward diffusion of Ca2+ cations into the printed droplets. The gelation of the alginate
beads was completed in approximately 8s, generating a fused bifurcated construct with a 90µm
diameter channel and 300µm thick walls. Due to the soft material properties, the hollow channel
collapsed slightly as the structure was built-up in height, resulting in a slightly deformed and
smaller opening of approximately 40µm in the upper region. Although alginate is a
biocompatible material non-toxic to the cells, it is limited in directing the fates of cells into
organized structures and cannot be remodelled by the cells 158. They incorporated collagen type I
into the alginate solution, relying on the biologically active property of collagen and the fast-
gelling characteristic of alginate. The rapid gelation of alginate prevents the diffusion and loss of
collagen, allowing it enough time to complete gelation at 37ºC. Once completed, the alginate
hydrogel was completely removed using a chelating agent (EDTA), leaving behind a structured
biologically relevant material that can be remodelled by the cells.
36
3.2 Strategy for the Continuous Formation of Soft Material Sheets of Heterogeneous Composition
We report an approach for the tessellation and coding of planar soft materials that is scalable and
continuous, does not involve substrate support or moving device components, and is compatible
with a range of biopolymers and different cell types. Figure 6 shows a schematic illustration. A
microfluidic device dynamically incorporates one or several secondary biopolymer solutions
within one layer of a base polymer. The secondary biopolymers are either chemically distinct
from the base biopolymer or they carry a molecular, colloidal or cellular payload. Upon exiting
the device, the spatial organization within the fluid layer is retained via diffusion-mediated ionic
cross-linking and a mosaic hydrogel is formed. The presented strategy promises a one-step
process for information to be encoded, concentration gradients of diffusing or binding molecules
to be established, directionally dependent mechanical and transport properties to be realized and
cells to be co-localized and co-cultured within the same soft material substrate. Depending on the
choice of biopolymers, payloads, tessellations and microenvironmental conditions, the mosaic
hydrogel may either display time-constant or dynamically changing characteristics. Three-
dimensional bulk structures of homogeneous or organized heterogeneous composition can be
subsequently produced using the same microfluidic platform, in a single continuous step.
3.2.1 Material and Methods
3.2.1.1 Materials
Alginate (alginic acid sodium salt) and calcium chloride were purchased from Sigma-Aldrich
(St. Louis, MO, US). The alginate sample contained 2%w.t. alginate in a solution of 60% v/v
glycerol in DI water. The pectin-alginate solution was obtained by incorporating 1%w.t. pectin
(Sigma-Aldrich) into an aqueous solution containing 1%w.t. alginate and 65% v/v glycerol. The
crosslinking solutions consisted of 50mM, 100mM, and 150mM CaCl2 in DI water containing
65%, 63%, and 61% v/v of glycerol respectively. The density of all solutions was 1.168g/mL.
Two types of fluorescence microbeads were used either for continuously projecting wide-field
fluorescence images of the formed hydrogels from an upright fluorescence microscopic setup
(Nikon Eclipse E600, Nikon, Japan) onto a line camera (LC1-USB, Thorlabs, Newton, NJ, USA)
or for off-line characterization using laser-scanning confocal microscopy (Olympus IX81
37
Inverted Microscope with FluoView FV1000, Olympus, Pennsylvania, USA). Specifically,
microspheres with mean diameter of 1μm with excitation/emission of 505/515nm and
535/575nm were purchased (F8852 and F8819, Invitrogen, Canada). Microbeads were added to
the biopolymer solutions at a ratio of 1:200, followed by 20min sonication (B5510-MT, Branson
Ultrasonics, Danbury, Connecticut, USA) to minimize aggregation.
3.2.1.2 Alginate Crosslinking Kinetics
Alginate gelation is triggered by an external crosslinking process during alginate extrusion into a
calcium chloride solution. As the two solutions come into contact at the device exit, the outer
surface of the sheet is instantly crosslinked, followed by further diffusive flux of calcium ions
from the reservoir solution into the gel to complete the gelation throughout the alginate sheet.
The maximum growth rate of the alginate gel crosslink density occurs within the first 15 minutes
of crosslinking 172. The gelation time is directly proportional to the alginate and crosslinker
concentration. Higher concentration of alginate directly translates to an increased number of
binding sites, therefore increasing the time needed for gelation. On the other hand, a higher
concentration of crosslinker results in a faster initial crosslinking rate due to the higher ionic
gradient created. Crosslinking then continues within the gel until either the ion source is depleted
or uncrosslinked positions in the gel are depleted.
Since the binding kinetics of calcium and alginate are so rapid compared to the diffusive
transport of calcium ions, the diffusion of calcium is the rate-limiting step in this gelation
process. In general, the diffusivity of calcium ions in porous gels and alginate is estimated to that
in water and has a value of DC = 1 × 10-9m2/s 173,174. This assumption can be made as no
significant diffusive resistance is present for molecules that are less than 20kDa (calcium
chloride molecular weight = 110.98Da). Mikkelsen and Elgsaeter developed a numerical model
to predict the calcium, alginate, and gel concentration over time 175. Assuming a homogeneous
system, the reaction should be consistent along the flow-direction. The analysis presented by
Mikkelsen could therefore be simplified to a one-dimensional diffusion analysis, with
concentration of the solutions dependent on time t and thickness of the alginate sheet x. For
simplicity, we assumed the diffusion coefficient of alginate to be much smaller than the diffusion
38
coefficient of Ca2+ ions, therefore negligible. This resulted in the following equations predicting
the time-dependent concentration of Ca2+, alginate, and gel respectively.
(1)
(2)
(3)
(4)
(5)
(6)
Where c is the free Ca2+ concentration, a the free alginate concentration, g the gel concentration,
NC is the average number of Ca2+ ions per alginate-alginate dimer formation (=50), k is the
reaction rate (= 2 × 10-2 M-2s-1) 172.
The initial concentrations are zero for calcium and gel, and 1 for free alginate molecules. The
boundary conditions are summarized in equations (4) and (5) where in the center of the sheet (x
= 0), fluxes are equal to 0. At the external boundary of the sheet (x = δ), it is assumed that only
the free Ca2+ penetrate the alginate sheet and the fluxes of alginate and gel are therefore equal to
zero. Time-dependent concentrations of free Ca2+, alginate, and gel are summarized in figure 7.
The data obtained suggests complete gelation of a 100µm thick alginate sample after 15 minutes
exposure to the cross-linking solution from one side. Since our alginate samples are immersed in
a reservoir filled with cross-linking solution and exposed to diffusive transport of Ca2+ from all
directions, our extruded 200µm thick alginate sheets are predicted to fully gel after 15 minutes.
39
For this reason, printed samples are kept immersed in the cross-linking solution for an additional
15 minutes before collecting them.
3.2.1.3 Neonatal Rat Heart Isolation
Neonatal Sprague–Dawley rats (1-2 day old) were euthanized according to the procedure
approved by the University of Toronto Committee on Animal Care. The cells from the heart
ventricles were isolated by treating with trypsin overnight (4○C, 6120 U/mL in Hanks’s balanced
salt solution, HBSS) followed by serial collagenase digestion (220 U/mL in HBSS)176. The
supernatant from five collagenase digests of the tissues was centrifuged at 750 rpm (RCF = 94 ×
g) for 4 min, resuspended in culture medium, and pre-plated into T75 flasks (Falcon) for 1 h
intervals to separate the adherent cells (non-myocyte) from the non-adherent cells (enriched
cardiomyocyte). Primary cardiac fibroblasts were obtained by cultivating for up to 7 days the
cells adhered to the T75 flask during the pre-plating. Culture medium for both cardiomyocyte
and fibroblast consisted of Dulbecco’s modified Eagle’s medium (DMEM) with 4.5 g/L glucose,
4 mM L-glutamine, 10% certified fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL
streptomycin and 10 mM 4-2-hydroxyethyl-1-piperazineethanesulphonic acid buffer (HEPES)
(Gibco, Invitrogen, Canada). Human umbilical vein endothelial cells (HUVEC) were purchased
from Lonza, Canada.
3.2.1.4 Cell Patterning
Cell were suspended in a 1:1 ratio of cell suspension solution and RGDS (arg-gly-asp-ser)
peptide-functionalized alginate solution. The cell suspension solution consisted of 12.3%v/v DI
water, 1.2%v/v glucose solution (0.3g/mL), 7.7%v/v 10x Medium 199 (Sigma-Aldrich, Canada),
1.1%v/v NaOH solution (1N), 2.0%v/v NaHCO3 solution (0.075g/mL), 0.8%v/v HEPES
(Invitrogen, Canada), 19.1%v/v MatrigelTM, and 55.9%v/v collagen type I from rat tail
(3.66mg/mL, BD Biosciences, Canada). The peptide-functionalized alginate solution consisted
of 1.5%w.t. RGDS-alginate and 0.08%w.t. collagen type I from rat tail. Peptide-functionalized
alginate was obtained following a previously described procedure177. Briefly, RGDS peptide
(American Peptide 44-0-14) was conjugated to alginate using carbodiimide chemistry with N-
40
hydroxysulfosuccinimide ester (sulfo-NHS) stabilizer (Pierce, Fisher 24510). The resulted
solution was purified by dialysis, dried by lyophilize, and stored at -20 °C until use.
3.2.1.5 Cell Tracking
CellTracker Red CMPTX (C34552, Molecular Probes, Invitrogen, Canada) was used for
fibroblasts and CellTracker Green for cardiomyocytes (C2925, Molecular Probes). A 10mM
concentration of CellTracker dyes in DMSO was further diluted in serum-free culture medium
(DMEM) to create a working concentration of 10 μM. The cells were incubated in 1mL of dye
solution for 30 min at 37ºC in 5% CO2. Following the incubation step, the dye-cell suspension
was centrifuged and the pellet was washed two times with DMEM.
3.2.1.6 Immunofluorescence Staining
Cell samples were fixed in 4% Paraformaldehyde in PBS at room temperature for 15 minutes
followed by incubation in mouse anti-vimentin (Sigma, 1:100 dilution) overnight at 4°C.
Samples were then incubated with anti-mouse Alexa 488 (Sigma, 1:100) at room temperature for
1 hour and imaged with confocal microscope (Olympus FV5-PSU confocal with IX70
microscope, Canada).
3.2.1.7 Sample Preparation for Scanning Electron Microscopy
Hydrogel samples were fixed in 2% glutaraldehyde in a 0.05M sodium cacodylate buffer at
pH 7.4 for 1hr at room temperature, followed by gradual replacement of the liquid phase with
100% ethanol. Dehydration of the samples was achieved with liquid CO2 at 10°C in a critical
point dryer. Samples were subsequently heated to 31°C with a pressure increase to 7.2MPa,
transitioning the CO2 to supercritical fluid conditions. Lowering the pressure from the
supercritical state allowed a direct transition into the gas phase without causing any unwanted
liquid-gas phase transitions. The dehydrated sample was then transferred into a vacuum and
vapour-deposited with a thin film of gold to render the outer surface of the substrate electrically
conductive.
41
3.2.1.8 Tensile Test
Samples were cut to lengths of approximately 20 mm and fixed with a cyanoacrylate adhesive
(Krazyglue Advanced Formula, Elmer's Products, Columbus, OH, USA) to cardboard strips,
which were vertically clamped between tensile grips for testing. A ramp of 0.1 mm/s was
applied using a 1000g load cell until failure.
3.2.2 Mosaic Hydrogel Formation
Mosaic hydrogels (thickness δ= 150 - 350µm, width ~3mm) were formed using a multilayer
microfluidic device along with the experimental setup shown in figure 8. Layers were
individually molded and vertically attached using a partial curing process 178, resulting in a 10-
layer-device that was able to withstand pressures of up to 600kPa (Appendix figure A1-b). The
center layer (indicated in blue color in figure 8b and as layer #6 in figure A1-b) carried to the
device exit via a set of parallel microchannels a time-varying content of biopolymer solutions.
Additional layers located above and below delivered, at the device exit, focusing streams
containing the crosslinker (indicated in green color in figure 8b). The produced biopolymer
sheet flowed into a liquid-filled reservoir which contained the same solution as the focusing
streams (Fig. 8a). To reduce the unwanted effect of flow instabilities at the device exit and to
ensure a uniform sheet thickness, δ, two co-flowing fluids were delivered from above and below
the soft biopolymer sheet in a flow-focusing configuration 179,180. The focusing fluids carried
cross-linking ions and induced gelation of the sheet. In a case study, we used a 2%w.t. alginate
solution, a biopolymer with well-known biocompatibility 159 and ionic crosslinking mechanism
181. To increase the fluid viscosity and render the produced biopolymer sheet neutrally buoyant
with respect to the focusing fluids, glycerol was added to both the biopolymer and focusing
streams, with the latter containing CaCl2 as the crosslinker. The focusing fluids were
continuously supplied by an annular gear pump (mzr-2921, HNP Mikrosysteme, Parchim,
Germany) at a rate of 8 mL/min. At a location approximately 50 mm downstream of the device
exit, the sheet was manually attached to a collection drum (21.3 mm in diameter) that rotated at a
constant tangential velocity, UP (Fig. 8a). The focusing fluid was subsequently stopped while the
hydrogel sheet continuously exited the device and was collected by the drum. Although the
shear stress exerted by the focusing fluid alone (Fig. A16) was sufficient to consistently form
hydrogel sheets, we relied on the rotating drum as this configuration allowed the continuous
42
formation, image-based inspection, and collection of mosaic hydrogels. The sheet thickness δ
was dynamically controlled by varying the flow rate of the base biopolymer, QB, using a syringe
pump (model PHD, Harvard Apparatus, Massachusetts, US) and by varying UP (Fig. 8b). The
employed base biopolymers and their pore sizes (Fig. 8e,f) along with the increased alignment of
the polymer fibers due to the axial stress imposed by the pulling drum affected the elastic moduli
of the produced sheets (Fig. A20).
The continuous collection of the biopolymer sheet onto the rotating drum provides an avenue to
spatially organize soft materials in three dimensions. For the purpose of conveniently imaging
the architecture of produced bulk material, a two-layered hydrogel sheet was produced by adding
an additional layer (blue layer in Fig. 8b) to the microfluidic device, resulting in an 11-layer-
device. The primary biopolymer layer feed a 2%w.t. alginate solution containing a payload of
red fluorescence microbeads, while the secondary layer feed a 2%w.t. alginate solution
containing green fluorescence microbeads. Figure 8g shows a confocal scan of the produced
layered hydrogel.
We now discuss how our experimental strategy allows the various secondary biopolymers to be
incorporated within a planar, unsupported hydrogel sheet, at a spatial resolution of approximately
up to 130µm (Fig. 8e). The ability to precisely control the incorporation of the secondary
hydrogel in the lateral direction and in time allows us to consistently define a variety of
tessellations in the (x, y)-plane. Until now, the ability to controllably form heterogeneous soft
materials in a one-step process was limited to microparticles 182-184 and coded fibers 185.
Recently, a stepwise approach was developed to incorporate within a 200µm thick paper
substrate a secondary biopolymer that consisted of Matrigel containing a payload of cells 186. The
paper substrate was photolithographically patterned. Subsequently, the cell-loaded Matrigel
wicked through spaces between the cellulose fibers, in areas that were not protected by the
photoresist.
Computer-controlled solenoid valves (The Lee Company, Connecticut, US) (Fig. 8h-j and Fig.
A17) initiated the outflow of secondary biopolymers from one of the seven on-chip reservoirs
during a time period tV at which the head pressure was raised from the atmospheric pressure level
43
P1 to P2. A biopolymer spot was then predictably incorporated within the hydrogel sheet and
cross-talk between different reservoirs was prevented.
3.2.3 Spatiotemporal Control and Payload Incorporation
The experimental configuration shown in Fig. 9a was used to assess the spatiotemporal control
of the process. The seven reservoirs were typically filled with a secondary biopolymer that
consisted of an alginate solution with different payloads (see Experimental section). A custom
computer interface allowed us to individually incorporate localized spots of the secondary
biopolymer on demand. In the first case we substituted the secondary biopolymer with a density-
matched aqueous solution with a composition identical to the focusing fluids (i.e., it contained
100mM CaCl2). A planar soft material sheet with an array of void areas was obtained (Fig. 9b),
at the following experimental conditions Up = 10mm/s, QB = 200µl/min, inlet pressure P =
3.5kPa, and tv = 65ms. Ultimately, the combination of void spaces with the ability to produce
millimeter thick multilayers of hydrogel sheets as illustrated in Fig. 8g, provides a strategy
towards 3D vascularized soft materials.
In a second case, we investigate the extent to which the incorporated biopolymer replaced the
base biopolymer by first considering fluorescently labelled microspheres as the payload. We
performed confocal microscopic scans and found the smallest ellipsoidal spot (lengths of semi-
principal axes: 100µm [w], 130µm [L], 130μm [δ]) that completely replaced the base hydrogel
across the entire sheet, as shown in Fig. 9c. The spot was produced with the conditions
Up = 12 mm/s, QB = 160 µl/min, inlet pressure P = 3.5kPa, and tV = 50ms.
In a third case, we selected viable cells as the payload. Consequently, the secondary biopolymer
was modified to improve cell viability and functionality, since alginate alone is insufficient to
promote cell proliferation, attachment, and migration 187,188. Briefly, a payload of neonatal rat
cardiomyocytes at a density of 10 × 106 cells/mL was suspended in a peptide-functionalized
hydrogel solution (see Methods for more detail). Confocal microscopic scans revealed a uniform
distribution of cardiomyocytes across the hydrogel sheet (Fig. 9d) in a configuration that is not
attainable in a single step using conventional top-down patterning approaches. The cell-loaded
secondary biopolymer pattern clearly extended throughout the entire cross-section (Fig. A19) of
a sheet sufficiently thin (δ~250µm) to be adequately penetrated by oxygen and nutrient
44
molecules contained in the culture solution 189. The fact that δ is at least tenfold greater than the
average size of the incorporated cells renders mosaic hydrogels a candidate format for 3D cell
culture. Although the described approach can be extended up to δ ~ 700µm (Fig. A15), sheet
thicknesses δ > 250µm were not considered for cell immobilization as they would require
internal vascularization to guarantee viability of the cellular payload throughout the cross
section. Similarly, the fibroblasts were suspended in the same secondary biopolymer as the
cardiomyocytes and were incorporated as patterned spots at a cell density of 10 × 106 cells/mL.
Confocal fluorescence images of the patterned spots were obtained on Day 5, demonstrating the
cells ability to attach onto the biopolymer (Fig. 9e,f). Patterned sheets were fixed and
immunostained following the protocol described in the Experimental section.
In a fourth case, the secondary biopolymer constituted of alginate containing fluorescently
labeled diffusible molecules (concentration 100µM, molecular weights 4kDa and 40kDa FITC-
dextran, and 10kDa rhodamine-dextran, Sigma-Aldrich, Missouri, US). Spots of the secondary
biopolymer (~4 nL) were incorporated in either 2%w.t. alginate (I) or in 1%w.t. pectin-1%w.t.
alginate (II) and the diffusive release of the fluorescent marker was followed in time-sequences
of fluorescence micrographs. Figure 9g shows two fluorescence images that were taken from
such a sequence for a spot with a 40kDa FITC-dextran payload, the first one right after gelation
and the second one 3hrs later. The diffusivity of the three dextran molecules in the base
hydrogels (I) and (II) was determined by fitting similarly measured intensity distributions to an
analytical solution of the concentration field (Fig. 9h). Specifically, the diffusion coefficient for
molecular transport of dextran (4kDa, 10kDa, and 40kDa) through two different hydrogel
matrices that were composed of either 2%w.t. alginate or 1%w.t. pectin-1%w.t. alginate were
calculated by curve fitting the time-lapsed experimental data with the analytical solution for one
dimensional diffusive transport into a semi-infinite domain190.
Dt
xerfcItxI
2),( 0 (7)
We obtained the diffusivity for a best fit using the least mean squares method (LMS). The LMS
value is defined as the sum of the residuals squared,
n
i irS1
2 , where the difference between
the experimental intensity value and the value predicted by the model is ),(exp, txIIr iii . For
45
all considered molecular payloads, a higher diffusivity was consistently obtained for hydrogel (I)
as compared to hydrogel (II), an effect that we attribute to the larger average pore size of
hydrogel (I) that was confirmed by scanning electron microscope (SEM) images (inset in Figs.
8e,f). As expected, the diffusivity in both base hydrogels decreased with increasing molecular
weight of the payload.
3.2.4 Information Encoding
The ability to incorporate isolated spots of a secondary biopolymer into unsupported soft
material sheets allows information to be encoded in a compact manner. The secondary
biopolymer alginate contained fluorescent microspheres as the payload. At a location
downstream of the device exit, the encoded information was continuously projected onto a line
camera using a fluorescence imaging configuration. The word “TORONTO” was patterned in
14s into a sheet of an alginate base polymer that was supplied at a flow rate QB = 160 μl/min, and
subsequently imaged (Fig. 9i). Valves were actuated at a pressure of P2 = 5 kPa with opening
and closing times of 75ms and 1000ms, respectively. The velocity of the drum was
UP = 12 mm/s. Each letter was represented by 7-20 individual spots and occupied an area of
approximately 6.25 mm2. Similarly, cardiomyocytes as a payload were pre-labelled (CellTracker
Green, Molecular Probes, Invitrogen, Canada) and predictably incorporated in multiple spots that
represented the letters “T” and “O” (Fig. 9j). The base biopolymer was 2%w.t. alginate and the
secondary biopolymer was a suspension of 10 × 106 cells/mL in the same peptide-functionalized
hydrogel as described previously for cardiomyocytes.
The density of the encoded information was increased 19 fold by employing the 7-bit American
Standard Code for Information Interchange (ASCII) where each of the seven solenoid valves was
assigned to one bit. The intensity values recorded from the formed hydrogel sheet were
interpreted by a custom computer program, translated back into text and validated against the
original text. In ASCII format, “TORONTO” was incorporated within a 37.5 mm long hydrogel
sheet during approximately 7.5 s at QB =160 μl/min and UP = 8.25 mm/s (Fig. 9k). To
demonstrate the ability of consistently writing and reading information, article 1, chapter 1 of the
UN Charter (165 words and 1,047 characters including spaces) was encoded in the same format
46
(see Fig. 9l and Fig. A21). We produced a 5.6 m long sheet within 18.8 min and subsequently
validated the encoded information with 100% accuracy.
3.2.5 Geometric Control over Mosaic Hydrogel Properties
The ability to dynamically control the local material composition provides an effective means of
altering local and bulk material properties, such as the permeability and the elasticity, and of
creating soft materials with directionally dependent properties. We formed and characterized,
using confocal and wide-field fluorescence microscopy, mosaic hydrogels with a variety of
tessellations including square tiles (Fig. 10b), stripe patterns of variable width (Fig. 10c), axially
inter-connected spots (Fig. 10d) and uniform-width stripe sections (Figs. 10a, e-h). In cases
where the base hydrogel and the secondary biopolymer were chemically distinct from each other
(i.e., they did not differ by the presence or absence of a payload only), we investigated how the
different tessellations affected the bulk elastic modulus. Homogeneous and mosaic alginate
sheets were formed via cross-linking with three CaCl2 concentrations, 50, 100, 150mM, and
tensile tests were conducted (Custom 840LE2 tensile tester, Test Resources Inc., Minnesota,
USA) (Fig. 10j, see Methods for sample fixation procedure). Two homogeneous hydrogel
samples with the previously described compositions (I) and (II) were prepared, along with
mosaic hydrogels with the tessellations shown in Figs. 10d and 10f. Figure 10i summarizes the
elastic moduli that were obtained for the different crosslinker concentrations. The values
obtained for mosaic hydrogels fall in between the ones corresponding to homogeneous samples.
A comparison between the two mosaic hydrogels suggests that axially aligned tessellations
(Fig. 10f) resulted in higher elastic moduli than laterally aligned ones (Fig. 10d). All samples
exhibited an increase in the bulk elastic modulus when the crosslinker concentration increased
from 50mM to 100mM. As the concentration of CaCl2 increased, the crosslinking rate increased
proportionally. As a result, a mosaic hydrogel with a locally increased stiffness in proximity of
the sheet surface was formed, limiting the diffusion of CaCl2 into the hydrogel and thereby
creating weaker internal polymer networks 174. We associate the decrease in elastic modulus that
was observed at 150mM with this effect. Elasticity only serves as one illustration for the
acquired ability of tuning a macroscale property with a tailored microscale material composition,
and can be extended to a wide range of other material properties.
47
3.2.6 Scalable Formation of 3D Hydrogel Assemblies
The obtained mosaic hydrogel sheets may be stacked-up to produce millimeter-size 3D
assemblies with a heterotypic composition. As an example of this approach, five hydrogel sheets
with the tessellations indicated in Fig. 10f were stacked in an alternating orientation (90° offset
between layers). In order to decrease absorbance of the 3D assembly during confocal imaging,
only one of the biopolymers contained a payload of fluorescence microspheres. The resulting 3D
structure had dimensions 5mm[w]×5mm[L]×1.5mm[δ] (Fig. 10j). Alternatively, the presented
platform can be employed for the tubular assembly of hydrogel sheets. As an illustration, the
rotating drum was replaced by a rotating capillary tube (22-690-943, Fisher Scientific, Canada)
which was manually translated to collect a continuously extruded hydrogel sheet with 50%
overlap in the sheet surface area. The overlap ensures the tubular architecture to be retained upon
the removal of the capillary tube. Homogeneous and heterogeneous hydrogel tubes with inner
diameters of approximately 1.5 mm and lengths of up to several centimeters were produced (Fig.
10k-n).
3.2.7 Planar Co-localization of Single and Multiple Cell Types
In a case study, we form mosaic hydrogels by locally incorporating single or multiple cell types
as a payload within the secondary biopolymer. In tissue engineering applications it is necessary
to authentically represent the physiological environmental milieu of a particular tissue or organ.
Resembling the structure and function of tissues and organs requires multiple cell types and
ECM molecules to be co-localized in two or three dimensional patterns at length scales that
exceed several millimeters. Currently available cell patterning methods allow to either
incorporate multiple cell types in microparticles and subsequently organize them in one or two
directions 191-194, or achieve co-localization along one direction within a fiber 185, but do not yet
provide dynamic control over the matrix composition and the incorporation of multiple cell types
in two or more directions.
We demonstrate continuous two-directional patterning of cardiomyocytes, endothelial cells and
fibroblasts, major components of the native heart 195,196, at a resolution of ~400µm and at length
48
scales of several millimeters. The cellular payloads suspended in the biopolymer streams are
exposed to shear levels less than 2 dyne/cm2 while passing through the microfluidic device
(Fig. A18), which is within physiological ranges 197-199 and well below shear stresses of 167-
200,000 dyne/cm2 commonly associated with direct-printing26 and ink-jet printing strategies
59,63,200. Neonatal rat fibroblasts were incorporated at a concentration of 10 million cells/mL and
the conditions Up=12mm/s, QB=160µl/min, inlet pressure P=3.5kPa, and tV=65ms (Fig. 10o, top
panel). The homogeneous distribution of cells within the hydrogel sheet was assessed by z-stack
confocal scans of five spots containing cardiomyocytes pre-labelled with CellTracker Green
(Molecular Probes, Invitrogen, Canada) and incorporated at a density of 10 × 106 cells/mL. Z-
stack scans were collected with a 30µm step size to prevent cells from being counted twice. For
each spot sample, three slices located in the middle, top, and bottom, were selected and cells
were counted from these x-y plane slices within an area of 400 × 400µm2 (Fig. A19).
Cell survival of fibroblasts within 15 days of culture and neonatal rat cardiomyocytes within 7
days of culture was investigated using a Live/Dead viability/cytotoxicity kit for mammalian cells
(L3224, Invitrogen, Canada). On Day 15, fibroblasts had a 88.7% survival, and on Day 7,
neonatal rat cardiomyocytes were observed to have 93.1% (Fig. A19). The co-localization of two
cell types (cardiomyocytes or endothelial cells with fibroblasts) within separate tessellations
within a mosaic hydrogel was illustrated by patterning parallel stripes (Fig. 10h) or islands (Fig.
10o, center and bottom figures). The ability to pattern multiple cell types in close geometrical
proximity offers the potential of systematically exploring cell-cell interactions via secreted
factors as well as the interrogation of heterotypic and homeotypic cell interactions. Figure 4p
illustrates how the incorporation of different cell types can be combined with the ability to record
the associated experimental parameters in the form of a barcode that can be tracked throughout
the duration of cell culture. We co-localized patterns consisting of cardiomyocytes (green) and
fibroblasts (red) by using four on-chip reservoirs. The remaining three reservoirs were dedicated
to a 6 bit computer-readable code where a 2%w.t. alginate with a payload of fluorescence
microspheres was used as the secondary biopolymer.
We presented a flowable format for the continuous tessellation of mosaic hydrogels. We further
explored some of the key advantages of our approach that go beyond the encoding of
information and include (a) the ability to predictably, accurately and dynamically control the
49
composition of a soft material in two dimensions, (b) the continuous and scalable soft material
formation at high throughput, and (c) the mask-free patterning approach. These capabilities
exceed previously demonstrated lower-dimensional heterogeneous soft materials that were
formed using an in-flow lithography technique 201-203 and lead to barcoded polymer particles with
a high information density or to coded fibers 185. The mosaic tessellations created using our
approach can be composed of chemically distinct hydrogels or may differ by the type or
concentration of the payload they carry: diffusing or binding molecules, microscale particles, or
cells. Spatiotemporal control over the hydrogel composition allowed the encoding of information
in the soft material at a resolution of approximately 130µm and 50ms.
The effects of tessellations on local and bulk material properties such as permeability to three
dextran molecules of varying sizes, as well as bulk elastic moduli of mosaic sheets were
investigated. In the cell patterning case studies, the ability to incorporate dense cell populations
within an unsupported soft material sheet with precise spatiotemporal control was demonstrated.
We selected neonatal rat cardiomyocytes to show the compatibility of our method with
patterning a primary cell type that is highly sensitive to adverse culture conditions, including
factors such as supra-physiological shear stress 204,205 and hypoxia.206 Additionally, as
cardiomyocytes are a terminally differentiated cell type that lacks the ability to proliferate, the
achieved spatiotemporal control and the compliance of our strategy with high cell densities is
critical in establishing a physiologically relevant functional tissue. The presented strategy may
enable a fully-automated and continuous format for culturing cells in physiologically relevant
microenvironments, the systematic investigation of cell-cell and cell-matrix interactions and,
ultimately, define 3D functional tissues.
50
Figure 6 One-step formation of mosaic hydrogels. Schematic illustration of presented
approach. Solutions of two distinct biopolymers, with the option of preloading the second with
microparticles, biomolecular or cellular payloads, are organized into a planar uncrosslinked fluid
network using a microfluidic device (outlined with dashed lines). At the device exit, a mosaic
hydrogel with a well-defined spatial composition is formed upon cross-linking (e.g. ionic
exchange). The mosaic hydrogel properties (e.g. elasticity, diffusivity of different molecular
payloads) can be tailored by controlling its microscale composition. The ability to controllably
define planar biomaterials with heterogeneous properties enables the predictable time dependent
diffusion of molecular payloads. Bottom-up stacking or continuous collection of mosaic
hydrogel sheets onto a rotating drum enables the formation of multilayered soft materials with
compositional control in three dimensions, as well as millimeter to centimeter-scale tubular
structures.
51
Figure 7. Analytical model of the time-dependent concentration of free cross-linker, free
alginate, and cross-linked gel. (a) Normalized concentration profiles of free Ca2+, free alginate,
and crosslinked gel as a function of time and position x. (b) 2D plot of the concentration profiles
as a function of position x, at five time points t = 0, 200, 400, 600, 1000s.
52
53
Figure 8 Continuous formation of hydrogels sheets: experimental setup and device designs.
(a) Experimental setup consisting of a microfluidic device with inlets for a base biopolymer
solution (supplied by a syringe pump) and focusing fluid (supplied by a gear pump). The
extruded soft material sheet is collected on a drum with diameter D. (b) Rendered image and (c)
photograph of device exit section. (d) Control over planar soft material thickness by varying
drum rotation speed UP, with base biopolymer flow rate QB = 80 μl/min (●), 120 μl/min (▲),
160 μl/min (■). (e, f) SEM images showing the pore structure of planar biopolymer of
homogeneous composition: (e) 2%w.t. alginate, (f) 1%w.t. pectin-1%w.t. alginate. (g) Confocal
fluorescence scan of multilayered biopolymer sheets obtained by collecting onto the rotating
drum a continuous sheet composed of two layers of biopolymers (green and red). Final bulk
dimensions of 5mm[w] × 3.5mm[δ] (thickness δ obtained by ~18 double layers 200µm thick)
with a total volume of 2 mL. (h) Rendered image and photograph (i) of multilayered microfluidic
device used for the formation of mosaic hydrogels. The device supplied biopolymer solutions 1-7
from on-chip reservoirs. Individual computer-controlled solenoid valves selected reservoir head
pressures between the states P1 and P2. (j) Schematic illustration of valve actuation. Scale bars
are 1 mm (b, c), 2 μm (e, f), 500 µm (g), 5 mm (h, i).
54
55
Figure 9. Dynamically encoded information in planar hydrogels. (a) Illustration of encoding
information by dynamically incorporating spots of a secondary (fluorescently labelled)
biopolymer into a base biopolymer and subsequently decoding the information contained in the
hydrogel sheet. (b) Hydrogel sheet with array of void areas as imaged by confocal fluorescence
(top) and scanning electron microscopy (bottom). (c) Confocal fluorescence image illustrating
dimensions and shape of spots created by incorporating a secondary biopolymer with a payload
of fluorescently labelled microspheres at conditions P = 3.5 kPa, QB = 160 µl/min, UP = 12 mm/s,
tv = 50 ms. Inserts represent the xy-plane (center location of sheet). d) Confocal fluorescence
image (x-z plane) of cardiomyocytes incorporated within a planar biomaterial. The planar
material of thickness δ is outlined by dashed lines. e) Confocal fluorescence image of a patterned
spot of fibroblasts incorporated at a cell density of 10 million cells/mL (40×, Day 5). f) 5x
magnification confocal fluorescence scan of fibroblasts spot shown in (e) (40×, Day 5). g)
Fluorescence image and corresponding intensity curves of 100 μM 40kDa FITC-dextran loaded
in 2%w.t. alginate and incorporated into the same base material. Images captured at times 0 and
3 hrs. h) Diffusivity of 4 kDa, 10 kDa, and 40 kDa dextran in 2%w.t. alginate ( ),
1%w.t. pectin-1%w.t. alginate ( ). i) Line camera intensity scan (top) and fluorescence image
(bottom) of encoded letters. j) Fluorescence image of pattern formed with 10 million/mL
cardiomyocytes in 1.2%w.t. alginate and in 0.08%w.t. collagen type I from rat tail (Day 0).
Approximately 25,000 cells were incorporated, operating conditions: P = 3.5 kPa,
QB = 160 µl/min, UP = 12 mm/s, valve 65 ms open. k) Fluorescence line scan of binary code
(top) and schematic of valve actuation with white sections corresponding to valve open (bottom)
(n = 7 binary characters). l) Sample fluorescence line scan of the UN charter in ASCII code (n=1,
2, ... 1047 binary characters including space). Scale bars are 500 μm (b), 150 μm (c), 200 μm
(d), 50 μm (e), 10 μm (f), 100 μm (g) and 2mm (i-k).
56
57
Figure 10 Mosaic hydrogels. Confocal and wide-field fluorescence images of mosaic hydrogels
with various tessellations. Two to three distinct material compositions are illustrated. Inserts
represent schematic of desired patterns including: two parallel stripes of distinct materials (a),
squares (b), alternating wave pattern (c), axially connected spots (d), and multiple parallel stripes
(e-h). Continuous inlet gas pressures ranging from 2-14 kPa were used, with a range of valve
opening times starting from 50 ms to infinity (for continuous stripe pattern). g) SEM image of
striped heterogeneous material (See Supplementary Information for SEM sample preparation). h)
Wide-field fluorescence image of two parallel stripes containing 10 million cells/mL of
cardiomyocytes (green) and fibroblasts (red). i) Modulus of elasticity of homogeneous and
heterogeneous planar materials with CaCl2 concentrations of 50, 100, and 150 mM: 2%w.t.
alginate ( ), 1%w.t. pectin-1%w.t. alginate ( ), 2%w.t. alginate with patterns of 1%w.t. pectin-
1%w.t. alginate ( ) as illustrated in (d), and ( ) in (f). j) Millimeter-scale 3D organization of
heterogeneous soft materials. Obtained by stacking, with alternating orientation, hydrogel sheets
with patterns as illustrated in insert (brightfield image) and (f). k-n) Centimeter length tubular
structures obtained by rolling onto a translating capillary tube. Confocal fluorescence images of
tubes with homogeneous (k) and heterogeneous (l,m) composition (contain fluorescent
microbeads as payload). n) Stereomicroscopy image of tubular structure. o) Single (top) and
multiple (middle and bottom) cell incorporation into a base planar material. Top and bottom
figures are fibroblasts (red) and endothelial cells (green) at a cell density of 10 million cells/mL.
Middle figure consists of fibroblasts (red) and cardiomyocytes (green) at a cell density of 2
million cells/mL. Images were captured on Day 0. p) Combination of multiple cell types
incorporation along with 6-bit barcoding of a planar material. Scale bars 500 μm (a-h, k-m,o, p),
1mm (i), 2mm (n).
.
58
4 Chapter 4 - Skin Printer: Continuous and Scalable Organization of Layered Tissue
We present a continuous approach for scalable and consistent incorporation of human skin
microtissues into handleable skin substitutes. Our skin printer consists of a microfabricated
cartridge and has a throughput of up to 0.5 square meters per hour (equivalent to ~ 2×108 cells
per hour). Skin equivalents with different sizes and morphologies were bioprinted without
substrate support, and consisted of a biopolymer matrix into which regular arrays of skin
microtissues were incorporated. Bioprinted skin substitutes were consistently produced with
widths between 6mm and 60mm. Tensile properties, cell viability, attachment and proliferation
of bioprinted sheets populated with patterns containing 2 × 106 fibroblasts per ml and 9 × 106
keratinocytes per ml were characterized. The presented approach promises the scalable formation
of cell-populated skin grafts with precise control over the cell types and location, enabling the
physiological properties of the native skin to be reproduced with improved accuracy. In this
approach, the unique ability to create localized arrays of microtissues rather than homogeneously
populating the entire area of the graft with cells may reduce the need for large cell number,
thereby accelerating the post-operative culture time.
4.1 Skin Structure
Many tissues of the body including skin and muscular tissue, e.g., blood vessels and heart
muscle, possess a unique hierarchical organization that is characterized by a layered architecture.
Skin is the largest organ of the body and forms a protective barrier against its environment, and
is composed of an upper epidermal and a lower dermal layer. A unique hierarchical organization
of different cells and extracellular matrix components is critically important for the structure and
biological function of intact skin. The epidermis prevents foreign organisms from entering the
body and limits evaporative water loss. Its thickness varies between 0.05mm on the eyelids to
1.5mm on the palms and soles. It is avascular and populated by 95% keratinocytes, which form
from 37 to 51 layers depending on the region of the body.207-209 The dermis of human skin varies
59
in thickness between 0.6mm, in the case of the eyelids, and 3mm, in the cases of the back, the
palms, and the soles210,211. The extracellular matrix content of the dermal layer is dominantly
comprised of collagen types I and III (> 90%) secreted by fibroblasts. Collagen is organized in
an irregular intertwined meshwork of fibers to accommodate to minor stresses during normal
activity while resisting severe stretch by aligning in rope-like structures in the direction parallel
to the tension applied207-209.
The dermis and epidermis are separated by a basal lamina, a 50-90nm thin layer that is composed
mainly of collagen Type IV, laminin, fibronectin and proteoglycans212, and prevents the
fibroblasts from the underlying dermis from direct contact with the epidermal keratinocytes,
while allowing immune cells to penetrate. On the epidermal side, the basal lamina nests a source
of self-renewing keratinocytes which, upon reaching the skin surface, stratify, shed their nuclei
and produce keratins that define the epidermal barrier. Instead of a flat boundary between the
epidermal and dermal layers, the basal lamina of intact skin exhibits periodic protrusions or
ridges (also known as rete ridges and papillary projections) that are 50-400µm wide and 50-
200µm deep 213. These ridges increase the area of contact between the dermis and epidermis by
approximately 4 to 7 times, promote a better adhesion between layers211 and enhance the
proliferation and differentiation of keratinocytes 212,214,215. The absence of a basal lamina results
in scaring and loss of biological function during wound healing 216.
The mechanical properties of skin have been investigated in vivo using a variety of approaches
including indentation 217, suction 218,219, tension 220, and torsion 221,222. The resulting values are
summarized in Table A5, and vary from one region of the body to another and are highly
dependent on the methods used (from 1.5kPa on the forearm using indentation 217 to 20MPa on
the leg using tension 220).
4.2 Skin grafts and Artificial Skin Substitutes
In severe burn injuries where both the epidermal and dermal layers are destroyed, effective
strategies that ensure prompt wound closure and result in favorable clinical outcomes and patient
survival rates are essential. Patient survival is inversely proportional to the time required for
wound coverage and stabilization, and the mortality rate increases by 10% for every additional
60
10% surface area burn223. However, in cases of large burns, rapid wound closure is often
hindered by the insufficient source of autologous skin for grafting. A succession of numerous
surgical interventions are required instead224,225.
An ideal skin substitutes possesses both epidermal and dermal layers, with the respective
keratinocytes and fibroblasts incorporated within each layer. The two cell types are known to
exhibit a strong co-dependence in forming functional and vascularized skin grafts226,227. Graft re-
vascularization is critical in supplying the epidermal layer with nutrients. For grafts thicknesses
larger than 400µm, angiogenesis alone does not establish a rapid enough nutrient supply,
compromising cell viability and often resulting in graft loss228,229.
4.2.1 Skin Grafts
As of today, the golden standard of burn treatment consists in the use of split-thickness autograft,
allograft, acellular or cell-populated artificial substitute, and cell spraying strategies230-237 (Fig.
A22). Limitations associated with autografts and commercially available skin substitutes include
their limited size, the temporary nature of some of the grafts, and the long preparation time of
cell-populated grafts (over two weeks culture prior to surgery).
A split-thickness autograft relies on the use of the patient’s own skin from an unwounded area.
Therefore, the procedure results in a final wound area that is larger than the original burn.
Wound size for split-thickness grafts can be reduced by the use of a meshing technique. It allows
the surface area of the graft to be increased by up 6 times, with however a downside of reduced
cosmesis230. However, grafting is only to a limited extent applicable to severe burn patients due
to the lack of sufficient skin donor sites. The absence of a full dermis limits the extent to wich
the split-thickness autograft resembles the structure and function of the patient’s normal skin230.
Allograft consists of skin from human donors (cadaveric, Alloderm®, LifeCell Corporation, The
Woodlands, TX). It is comprised of a fully formed basement membrane. The associated
morbidity is reduced as compared to split-thickness autografts235,238. However, this strategy
comes at the risk of transmitting pathogens, e.g., the human immunodeficiency virus. Even
though the donor skin becomes vascularized upon grafting, the upper epidermal layer needs to be
removed after approximately 3 weeks to prevent the immunologic rejection by the patient235.
61
Recently, micrografting techniques from the late 50s have been readapted to regenerate the
epidermis of full-thickness porcine wounds239. Due to the lack of donor sites in large burns, the
use of micrografts from autologous split-thickness skin graft is promising as it enables up to
1:100 expansion ratio, while still enabling full-thickness regeneration in porcine wounds240-242. A
total of 39 micrografts 0.8mm × 0.8mm × 0.35mm were cut using a mincing device and
manually spread over a 5cm × 5cm porcine wound bed 239. This approach was combined with
clinically available moist dressings in view of potential future clinical applications.
4.2.2 Artificial Skin Substitutes
A critical function of skin substitutes is to promote the formation of an intact epidermal barrier
that reduces transepidermal water loss, infection, and hypertrophic scarring by accelerating
wound closure and patient recovery, while enable large surface coverage without the need for
lengthy pre-operative cell culture time243. In order to best reproduce functional elements of intact
skin, biological substitutes must mimic its structure and cellular organization. Such a
requirement must rely on a technology that will enable the simultaneous control over the
combined incorporation of cells, growth factors, and biomaterials.
Boyce et al. have developed skin substitutes consisting of autologous fibroblasts and
keratinocytes cultured for two weeks on collagen sponges. Although limited in thickness
(<500µm), their approach promises qualitative outcomes comparable to meshed split-thickness
autografts, with the advantage of reduced need for donor skin harvesting244,240. Similarly, Kempf
et al. created cultured skin grafts composed of spunges made of electrospun collagen type I
fibers, onto which human fibroblasts and keratinocytes were seeded. Fibroblasts and
keratinocytes naturally separated into two distinct layers over time and resulted in a cell-laden bi-
layer structure with a typical thickness of 50-70µm, smaller than the 1-3mm typical for human
skin245.
Integra is a bilayered, commercially available skin graft consisting of type I bovine collagen and
chondroitin-6-sulfate, covered with a silicone membrane 231. However, the graft is limited by its
high cost, the two stage procedure, rejection by patients allergic to bovine products, and the
inability of cultured cells to efficiently adhere to and penetrate through the matrix 246.
62
Cell therapy is another strategy that relies on cultured epidermal cells from patients or cell lines.
Cultured epithelial autografts and skin keratinocyte sheets (Epicel247) are based on keratinocytes
only. Cultured epithelial autografts is an appealing alternative as it enables expansion ratios of up
to 1000 times248, well beyond the maximum 6 times achievable using standard meshing
techniques249,250. However, culture of these cells to desirable numbers takes up to weeks, putting
further at risk the life of patients. In addition, these cellular grafts are fragile to handle and
require a supporting scaffold229,251. A feeder layer of fibroblasts is often required230. Zaulyanov
et al. investigated the use of Apligraf in the treatment of foot and leg ulcers and found it to be
equally fragile, and its incorporated neonatal keratinocytes did not remain viable when applied to
the wound bed. They found its function to be limited to stimulating wound healing through the
delivery of growth factors232,233. Fragility in handling can be overcome by employing a cell
spraying technique252 using a large number of cells (0.16 million cells/cm2), but results in limited
attachment of keratinocytes to the underlying dermis234.
4.2.3 Microfluidic approaches
The formation of spatially organized, cell-laden biomaterials has also been accomplished through
microfluidic-based approaches such as sequential replica molding or templating24,253, or by using
continuous-flow microfluidic formats that form cell-laden fibers or sheets185,254,255.
Atac et al. proposed a micro-bioreactor for the long-term dynamic culture of skin equivalents, ex
vivo skin tissue samples, and single hair follicular units 256. Similarly, Morimoto et al.257
prepared spheroid-based skin microtissues enabling the co-culture of both fibroblasts and
keratinocytes and their study as a simplified skin tissue model. The spheroids were 150µm in
diameter, composed of collagen Type I, and encapsulated human dermal fibroblasts at a density
of 1×106cells/mL. Human epidermal keratinocytes were subsequently seeded and attached to the
collagen spheroids at a density of 0.06×106 cells/cm2. Although promising for screening
purposes, microsystems-based advances have to date lacked the ability to routinely define skin
substrates larger than 1cm2 in size that closely recapitulate the structure and function of human
skin. Cerqueira et al. have remedied to the small scale limitation by utilizing a cell-sheet
technique 250, resulting in 0.8cm2 human adipose stem cell derived sheets for the treatment of
murine full-thickness excisional wounds239. Although there is potential in producing larger
63
sheets 139, this approach is time consuming as it requires manual stacking of up to three
individual sheets, or the use of a support substrate to produce sheets that are sufficiently robust
and handleable for grafting. 258
4.3 3D Printer Strategy for the Continuous Formation of Skin Substitute for Wound Healing
Here, we present a skin printer for the scalable formation of cell-populated wound dressings that
accurately reproduce key features of human skin, and are scalable to grafts with large surface
areas (tens of cm2). Our approach is based on a microfabricated printer cartridge that enables the
scalable incorporation of skin microtissue regions within continuously produced hydrogel sheets.
The printed skin grafts consist of multiple viable cell types, have a precisely controlled graft
thickness, structure (e.g. epidermis and dermis), and biomolecular composition that are
characteristic of human skin. We demonstrate the provided skin substitutes are easy to routinely
handle and apply to the wound bed, and promise to reduce wound recovery times.
4.3.1 Materials and Methods
4.3.1.1 Materials
The following materials were purchased for the preparation of both the base graft material and
the cell printing solution: sodium alginate (Molecule-R), high G:M alginate (G:M >1.5,
Novamatrix), Hyaluronic acid (HA Grade 80, Novamatrix), collagen type I from rat tail (BD),
Matrigel (BD).
The base sheet material consisted of 1%w.t. sodium alginate (Molecule-R). Cells were
suspended in 3.5%v/v DI water, 0.6 %v/v glucose solution (0.3g/mL), 3.7%v/v 10× Medium 199
(Sigma-Aldrich, Canada), 4.5%v/v NaOH solution (1 N), 0.96%v/v NaHCO3 solution
(0.075g/mL), 0.4%v/v HEPES (Invitrogen, Canada), 18.7%v/v Matrigel, 2.5mg/mL collagen
type I from rat tail (BD Biosciences, Canada), 42%v/v culture media (DMEM and Epilife for
fibroblasts and keratinocytes respectively), and 8.34%v/v of cell printing alginate. The cell
printing alginate differed for each cell type to be printed. Fibroblast suspension material was
64
composed of 0.05%w.t. sodium alginate (Molecule-R), whereas keratinocyte suspension material
consisted of 0.2%w.t. high G:M ratio alginate (G:M>1.5) with 0.23mg/mL hyaluronic acid (HA
Grade 150).
4.3.1.2 Printer Cartridge Fabrication
The printer cartridge consisted of four to six layers, which were fabricated using standard soft
lithography techniques139. Masters with 150µm tall features were defined, from which each
layers were individually molded and vertically attached using a partial curing process178. The
cartridge design for bilayered skin grafts consisted of six layers (Fig. A1c, left). Layers 1 and 4
distribute the flow focusing solution. Layer 5 and 6 distribute the respective fibroblast and
keratinocyte-loaded solutions into layer 2 and 3 respectively. Each layer 2 and 3 are populated
with additional channels for the distribution of a base material. This configuration results in the
formation of a bilayered skin graft containing keratinocytes on the top layer and fibroblasts on
the bottom layer. Both cartridges for single layered and 6cm wide grafts consisted of four
aligned and bonded layers. The design for the formation of a single layered graft is composed of
layers 1, 2, 4, and 5 in figure A1c (left). The cartridge design for a 60mm wide graft is
represented in figure A1c (right). Layers 1 and 3 distribute the focusing solution. Layer 4
distributes the cell-populated solution into layer 2, which guides both the base material solution
and cell-populated solution in a planar fashion until they exit the device as an organized planar
graft.
4.3.1.3 Tensile Measurements
Tensile measurements of patterned sheets were measured using a custom tensile tester (840LE2
tensile tester, Test Resources Inc., Minnesota, USA). Wet samples were cut to lengths of
approximately 20 mm and sandwhiched between two cardboard strips, which were vertically
clamped between tensile grips for testing. A ramp of 0.1 mm/s was applied using a 1000g load
cell until failure. Results are expressed as mean ± standard deviations (each sample has n = 5
experiments). All data were analysed using multiple comparison ANOVA tests in SigmaStat 3.5
(Systat Software Inc.; San Jose, CA, USA) using the Tukey method.
65
4.3.1.4 Cell Sources
Fibroblasts were obtained from healthy human normal skin after surgery. Cells were cultured in
growth medium (DMEM, 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic
(Ab/Am)) until near confluency and split into further passages by trypsin-EDTA treatment
(0.05% Trypsin-EDTA). Human umbilical epidermal keratinocytes (Gibco Invitrogen, C-011-
5C) were cultured according to company instructions in EpiLife Medium with 1% HKGS and
1% Ab/Am and trypsinized using the same trypsin-EDTA solution as for fibroblasts.
4.3.1.5 Immunofluorescent Staining
Cells in sheets were fixed with 4% paraformaldehyde in HBSS for 1 hour at room temperature
then washed with HBSS. They were permeabilized with 0.5% Triton X-100 in HBSS for 30 min
at room temperature and then washed with HBSS. Cells were blocked with block buffer (1%
BSA in 0.25% Triton X-100 in HBSS) for 1 hour. Antibodies were diluted in block buffer and
incubated overnight at 4°C. Primary antibodies included fluorescein phalloidin (Life
Technologies) and cytokeratin 14 (Santa Cruz Biotechnology). If only phalloidin was used, the
mounting step was performed next. With keratinocytes, samples were washed with HBSS then
incubated with secondary Alexa Fluor antibodies (Life Technologies). Cells were washed then
slides were mounted with Vectashield mounting medium with DAPI (Vector Laboratories).
Images were taken on Apotome Axiovert fluorescent imaging system or Zeiss Observer
Z1 spinning disk confocal microscope.
4.3.1.6 Murine Model
Male nude mice (J:NU, Jackson Laboratory) at 6-8 weeks old were ordered and allowed to
acclimate to the environment for a week prior to use. Mice were anesthetized with isofluorane
and buprenorphine was given as an analgesic. Two 6mm diameter full-thickness excisional
wounds were created on the back of each mouse using a biopsy punch. Silicon rings were
sutured onto the wound to prevent contraction. Circular patches of the same diameter were
punched from the printed sheets and placed onto the wound model, followed by coverage with
66
Tegaderm. The mice were monitored and health scored twice daily for 10 days. Mice were
sacrificed using CO2 euthanasia and the wound area was taken and stored in 10% neutral-
buffered formalin for histological analysis.
4.3.1.7 Histology
Tissue specimens were fixed in 10% buffered formalin overnight at 4°C, and stored in 70%
ethanol and embedded in paraffin. Specimens were cut into 5 µm sections in the centre of the
wound.
Trichrome reagents were from EMS (Hatfield, PA) unless otherwise stated. Briefly, paraffin
embedded slides were deparaffinized with citrosol, followed by rehydration through grades of
ethanol to water. Slides were placed in Bouin’s solution 1 hour at 60°C and washed in water.
Hematoxylin (Sigma) and Biebrich scarlet-acid fuchsin solution were stained for 10 minutes
each, respectively with washes in between. Slides were differentiated in phosphomolybdic-
tungstic acid for 15 min, and transferred to aniline blue for 5 min. Slides were rinsed and
differentiated in 1% acetic acid for 2 min. Slides were dehydrated through 95% ethanol and
absolute ethanol followed by clearing in citrosol. Slides were mounted with SHUR/Mount
xylene-based liquid mounting media (Triangle Biomedical Sciences). Images were acquired
using LeicaDM 2000LED light microscope.
For immunohistochemistry staining, paraffin embedded skin tissue slides were deparaffinized
with citrosol followed by rehydration. Antigen decloaker (1X, Biocare) was added to the slides
in a preheated decloaking chamber for 4 minutes at 110°C. Samples were blocked with 3% H2O2
for 10 min, then washed with washing buffer (0.05 M Tris-HCl, 0.15 M NaCl, 0.05% Tween 20
in DI water). Primary antibody was diluted in PBS and incubated at room temperature for 1 h.
Primary antibody used was cytokeratin 14 (Santa Cruz Biotechnology). Next, slides were
incubated for 15 minutes first with goat-on-rodent probe (Biocare Medical), and secondly with
goat-on-rodent HRP-polymer. The betazoid DAB chromogen kit (Biocare Medical) was added
for 5-10 min and the reaction was terminated with running water. Nuclear staining was done with
hematoxylin for 30 sec, followed by differentiation with 3 dips in 1.5% acid alcohol and bluing
in 0.1% sodium bicarbonate for 10 sec. Sections were dehydrated through 95% and absolute
67
ethanol to citrosol and mounted with SHUR/Mount as previously described. Images were
acquired using LeicaDM 2000LED light microscope.
4.3.2 Formation of Cell-Populated Skin Grafts
The skin printer shown in figure 12a consists of a microfabricated cartridge, a receiving
reservoir, a stepper motor-controlled drum and a pneumatic control unit. The printer cartridge
distributes and organizes in a planar fashion the biopolymer matrix that provides mechanical
support for the printed grafts, along with the printing solution that contains a mixture of
biopolymers with dermal and epidermal cells. Three microfabricated cartridges were fabricated
for the formation of single layered, bilayered, and wide (60mm in width) skin substitutes
respectively. The biopolymer solution for the support structure consisted of 1%w.t. sodium
alginate (Molecule-R) and was supplied to the cartridge using syringe pumps. Up to two
different biopolymer-cell mixtures were supplied at the top side of the cartridge, from separate
on-chip wells. Pneumatic control of the individual head pressures along with the microfabricated
channel networks within the cartridge enabled us with precise spatio-temporal patterning of cells
within the continuously formed skin graft 255. Dedicated cartridges allowed single and bilayered
sheets that were populated with viable human fibroblasts and keratinocytes to be continuously
formed and resembled the dermal and epidermal layers of intact human skin. To reduce the
unwanted effect of flow instabilities at the device exit and to ensure a uniform sheet thickness, a
co-flowing focusing solution was delivered from above and below the soft biopolymer sheet in a
flow-focusing configuration.179 The focusing fluids carried cross-linking ions and induced
gelation of the sheet. The ability to spatially control the cellular composition of these grafts
allowed the formation of skin substitutes with defined epidermis and dermis, appropriate cell
types, and control over the cell density. The exit of the microfabricated printer cartridge was
immersed in a liquid-filled reservoir. The calcium-loaded liquid within the reservoir facilitated
cross-linking of the skin graft and guided the cross-linked graft at a distance 50 mm away from
the cartridge exit to a rotating drum (drum velocities: 1-5mm/s). Varying the velocity of drum
rotation allowed the sheet thickness to be precisely tailored. Skin grafts with thicknesses between
100µm and 500µm, and widths between 3mm and 60mm were prepared.
68
Cells were incorporated in a variety of patterns within single and bilayered grafts, as alternating
spots, parallel spots, parallel void regions, and parallel stripes (Fig. 12b). The skin grafts were
subsequently collected, cultured for three days in vitro, and implanted in vivo on murine wound
models (Fig. 12c). A photograph of a microfabricated cartridge within the liquid-filled reservoir
is shown in figure 12d. Figure 12e focuses on the exit region of the cartridge, where a parallel
array of a secondary material (labelled with rhodamine for visualization) is printed into the
continuously extruded sheet.
4.3.3 Pattern Formation – No Cells
Controlled 3D cell printing and incorporation of void spaces within the hydrogel sheets was
achieved through up to two computer-controlled solenoid valves. Hydrogel solutions with a
known concentration of a particular cell type were loaded into the on-chip wells. After loading,
the wells were closed and connected to individual solenoid valves that selected between two
pressure levels. Upon electrical valve actuation, the well head-pressure either suddenly increased
or decreased, thereby initiating or ending the local incorporation of a cellular pattern within a
continuously formed hydrogel sheet. Different patterns in the form of stripes and co-localized
spots were formed, defined by the selected valve actuation times and well head-pressures (Fig.
13a). Spot volumes were precisely controlled and varied between 45-450nL (Fig. A23).
As an illustration of the patterning ability and accuracy, and to distinguish between the base
material and the patterning solution, hydrogel sheets were initially patterned with an alginate
solution (0.75%w.t. sodium alginate) while the base material was loaded with 0.5%w.t. BaSO4.
Single-layer hydrogel sheets that contained regular arrays of parallel spots, stripes, alternating
spots and void regions were continuously incorporated into the continuously extruded base
material (Fig. 13b).
4.3.4 Characterization of Printed Tissue Substitutes
An important practical requirement for skin grafts is their handleability during tissue culture and
the application to the wound bed. The elastic moduli were measured for sheets printed using
69
different biopolymer solutions and patterns. Materials included homogeneous sheets of 1%w.t.
sodium alginate (Molecule-R), 4%w.t. sodium alginate (Sigma), sheets patterned with parallel
stripes, parallel spots, and alterning spots of cell printing solution, and alternating void regions.
Degradation studies over the course of 11 days were performed for all skin graft compositions
and their elastic moduli were measured during a period of four days culture in both fibroblast and
keratinocyte culture media (DMEM and Epilife, respectively) (Figs. A24, 25). Mechanical
properties of the printed skin grafts were strongly dependent on three parameters: base material,
pattern, culture media, and culture time.
For the homogeneous case, graft degradation was characterized by incubating in DMEM and
Epilife culture media over the course of 11 days. The dry weight of five samples for each time
point and condition (material type and culture media) was measured and normalized by the value
of dry weight after 2h incubation on Day 0 (Fig. A24a). Epilife culture media degraded the
materials more rapidly than in DMEM culture media, with the 1%w.t. sodium alginate
(Molecule-R) exhibiting a slower degradation than the 4%w.t. sodium alginate (Sigma). The
elastic moduli measurements shown in figure A25 confirmed our degradation results. Sheets
made of 1%w.t. sodium alginate (Molecule-R) exhibited a higher elastic modulus than the
4%w.t. sodium alginate (Sigma). Specifically, the initial elastic modulus of 4%w.t. sodium
alginate (Sigma) of 51.4 ± 4.6kPa (Day 0) was reduced to 18.8 ± 8.2kPa and 14.2 ± 7.6kPa on
Day 4 in DMEM and Epilife respectively, equivalent to a reduction by 63.4% and 72.4% of the
original value. The elastic modulus of 1%w.t. sodium alginate (Molecule-R) in DMEM however
remained constant at approximately 51kPa ± 4.4kPa during the first four days and reduced by
36.4% by day 4 during culture in Epilife. Similar trends were observed for the ultimate tensile
strength of 1%w.t. sodium alginate (Molecule-R) and 4%w.t. sodium alginate (Sigma). For the
first, the initial value of 62.5 ± 2.1kPa was reduced by 28% in DMEM and 56.8% in Epilife on
day 4. For the second, the initial ultimate tensile strength of 29.2 ± 1.9kPa dropped by 58.7% in
DMEM and 86.3% in Epiflife on day 4. We therefore selected 1%w.t. sodium alginate
(Molecule-R) as the base material of our printed skin grafts.
Although non-toxic to cells, commonly employed for cell encapsulation and 3D culture, and
FDA approved for clinical use in skin grafts, alginate alone does not provide an adequate
environment for cell attachment, proliferation, and migration. The cell printing solution was
70
therefore adapted to cater to the above requirements. Briefly, two cell printing solutions were
optimized for fibroblasts and keratinocytes respectively, with a general decrease in alginate
concentration of 80-95% as compared to the base material. We compensated for the reduced
alginate concentration by selecting an alginate with a higher elastic modulus. Alginate is a linear
co-polymer composed of blocks of mannuronic (M) and gluoronic (G) residues covalently linked
in different concentrations and sequences. Alginate gelation is triggered through an ionic
mechanism. Divalent cations (Ca2+) dissolved in the liquid-filled reservoir interacted with blocks
of G residues to form a 3D gel network. A higher G:M ratio is equivalent with an increased
number of binding sites and therefore produced mechanically stronger gels.
In the case of keratinocytes, hyaluronic acide (HA) was added to the material composition. HA is
a nonsulfated glycosaminoglycan widely found in epithelial and connective tissues. It is a major
component of the ECM, contributes significantly to cell proliferation and migration, exhibits
anti-inflammatory behavior and, reduces scar formation and graft contraction by regulating
collagen synthesis85,242,259,260.
Similar to the homogeneous samples, the degradation of the cell printing solutions was
characterized by incubating in DMEM and Epilife culture media over the course of 11 days (Fig.
A24b). Overall, Epilife culture medium degraded the materials more rapidly than in DMEM
culture medium, with the fibroblast printing solution exhibiting slower degradation than
keratinocyte printing solution for both culture media.
The cell printing solution remains considerably weaker than the base material. The composition
of cell-laden printed grafts with different patterns was therefore optimized to ensure consistent
handling without rupture. The elastic moduli of four distinct patterns were investigated and
compared (Fig. 13c). Tensile tests were performed on printed sheets in both the extrusion
direction (x) and the lateral direction (y) (Fig. 13c insert), and on Day 0 and Day 1 during culture
in DMEM. Sheets patterned with alternating spots and void regions showed comparable elastic
modulus of approximately 35kPa, and exhibited satisfactory stability after one day of culture in
DMEM, with approximately 12% decrease in elastic modulus in both x and y. Patterns of parallel
spots and stripes however, showed much weaker properties when tested in the y-direction, with a
drop of 50% in elastic modulus as compared to measurements performed in the x-direction.
Similarly, the ultimate tensile strength of sheets formed with alternating spots and void regions
71
was around 13kPa in x and 8kPa in y on day 0, with a decrease of approximately 15% on day 1.
On the other hand, patterns consisting of parallel stripes and spots exhibited up to 87% drop in
UTS in the y direction compared to x direction.
4.3.5 Cell-Populated Skin Grafts in Vitro
4.3.5.1 Single Layer Grafts
The presented platform to continuously pattern soft material sheets with spots and void regions
was utilized to controllably incorporate human fibroblasts and human epidermal keratinocytes
within single and bilayer skin grafts. Fibroblasts were incorporated at a concentration of
2×106cells/mL, while keratinocytes were incorporated at a concentration of 4×106cells/mL.
Figures 14a-f reveal fibroblasts printed within single layer sheets in the form of spots and
parallel stripes. The cell printing solution used consisted of 0.05%w.t. sodium alginate
(Molecule-R), 2.5mg/mL collagen type I, and 18.7%v/v Matrigel. After two days of culture in
DMEM, the printed fibroblasts were fully confluent within the three-dimensional environment of
the skin graft. The cell printing solution for keratinocytes was optimized to maximized cell
proliferation. Four candidates were tested, namely: (1) 0.05% w.t. sodium alginate (Molecule-R)
with 2.5mg/mL HA grade 80 (MW ~ 620-1200kDa), (2) 0.05% w.t. sodium alginate (Molecule-
R) with 2.5mg/mL HA grade 150 (MW ~ 1200-1900kDa), (3) 0.2% w.t. high G:M alginate
with 2.5mg/mL HA grade 80, and (4) 0.2% w.t. high G:M alginate with 2.5mg/mL HA grade
150, all of which also contained 2.5mg/mL collagen type I and 18.7%v/v Matrigel. Material
samples were prepared by pipetting 500µl of each solution into individual wells in a 24 well
plate. The solutions were gelled by gently pipetting in equal volume of alginate cross-linking
solution containing 50mM CaCl2 in DI water. The well plate was then placed in the incubator for
an additional 30min to further cross-link the collagen component of the solutions. Subsequently,
the excess calcium solution was removed and gels were gently rinsed twice with Epilife.
Keratinocytes were then seeded onto each gel sample at a concentration of 0.1million cells/cm2
and cultured over 13 days in Epilife culture media. Their proliferation was quantified by
measuring the percentage increase in surface area covered by the cells at different type points
throughout the culture period. Keratinocytes began to form clusters on Day 2. On Day 3 of
culture, samples were placed on 6-wellplate transwells and brought to the air-liquid interface.
72
Compiled results in figure 15 confirmed that material 3 had the optimal composition, with a
645±28% increase in cell surface area on day 13. The insert in figure 15 shows a fluorescence
microscopy image of keratinocytes stained with Dapi, Phalloidin, and Keratin14.
4.3.6 Scalable Formation of Tissue Substitutes
The presented platform is scalable to produce skin grafts up to 10 folds wider than the smaller
grafts, equivalent to up to 60mm wide sheets. These larger scale grafts are obtained using a
microfluidic device with a 75mm wide exit section. The widths and thicknesses of skin grafts as
a function of pulling velocity obtained using four cartridges of different width were measured
and summarized in figure 17. Overall, 10mm, 25mm, 60mm, and 75mm wide devices were
characterized. The matrix flow rates used for the 10mm cartridge were: 100µl/min (green),
150µl/min (red), 200µl/min (blue). For the other three devices, flow rates were scaled
proportionally by multiplying by 2.5, 6, and 7.5 respectively. Rendered designs of 10mm and
75mm wide cartridges are represented in figure 16a. For all devices, there is on average a 35%
decrease in final sheet width compared to the initial cartridge width (Fig. 16b). We hypothesize
this phenomenon to be caused by the pulling of the alginate sheet as it exits the cartridges and
enters a stagnant liquid reservoir. This width reduction must be taken into account when
designing new cartridges for the formation of different skin grafts of specific widths. Finally,
since the matrix flow rates were scaled proportionally to the changes in cartridge widths, the
thicknesses obtained within the range of pulling velocity was consistent for all devices (Fig.
16c).
4.3.7 Cell-Populated Skin Grafts In Vivo
In vivo validation of cell-populated bilayered skin grafts was performed on 6 mm diameter
incision wounds in immune-incompetent mice. Samples were prepared and cultured in Epilife
culture media containing 1:100 BrdU for three days at 37 ̊C with 5% O2, following which they
were implanted onto murine models (fig. 17a). A representation of a bilayered skin graft was
obtained by co-localizing green fluorescent microbeads on the top layer (keratinocytes within the
epidermis) and red microbeads on the bottom layer (fibroblasts within the dermis) (fig. 17b). The
73
main sheet material was loaded with blue fluorescent microbeads. All microbeads were 0.1µm in
diameter and incorporated into the solutions at a density of 1%v/v.
Control consisted of 1%w.t. sodium alginate sheets (Molecule-R) patterned with cell-free cell
printing solution, whereas cell-populated sheets were patterned with fibroblasts at a
concentration of 2×106cells/mL. Our preliminary data suggests that our printed biopolymer skin
grafts led to an improved skin regeneration compared to our control (fig. 17c-f). Specifically,
trichrome staining showed better integration and collagen formation in the cell-populated graft as
opposed to control. Similarly, Keratin 14 staining revealed keratinization post-wounding on the
cell-populated graft compared to none in the control, where keratinocytes are rather migrating
underneath the alginate graft.
We have demonstrated a scalable approach for the formation of human skin grafts that were up
to 6cm in width and meters in length. The high-throughput of this platform enabled skin grafts to
be printed at a rate of up to 0.5m2/hr, equivalent to ~2×108 cells/hr. These grafts composed of
human skin microtissue islands (2 × 106 fibroblasts per ml and 9 × 106 keratinocytes per ml)
were printed in a variety of patterns that affected the overall mechanical properties of the grafts
and determined their handleability for application onto wound beds of in vivo murine models.
Based on the tensile properties of the printed grafts, an optimal pattern design was chosen for our
in vivo experiments and consisted of microtissue islands printed in an array of alternating spots,
which resulted in the highest mechanical properties (i.e. E = 30 ± 2.1kPa and UTS = 12 ± 1.5kPa
after one day of culture in DMEM). Preliminary in vivo data obtained over 10 days post grafting
suggests improved wound healing of wound grafted with cell-populated bilayer skin substitutes
in comparison to fibroblasts-only or cell-free substitutes. This approach offers the scalable
formation of handleable skin grafts with control over the cell type and concentration, and
promises the formation of meter scale grafts with reduced requirement for cell number.
74
Figure 11. Intact human skin and bioprinted skin grafts. (a) Multi-scale organization of skin,
where the epidermis is populated with keratinocytes, and the dermal layer with fibroblasts and
collagen type I being the predominant component of the ECM forming the dermis. The
epidermal and dermal layers are separated by a basement membrane. (b) Scalable, one-step
incorporation of microtissue arrays into bioprinted skin graft.
75
Figure 12 Skin Printer. (a) Schematic illustration of skin printer operation for continuous
formation of cell-populated skin graft by means of a microfabricated printer cartridge. The
microfluidic device distributes base materials, with the ability to locally and controllably pattern
both keratinocytes and fibroblasts. A streaming solution is distributed on the top and bottom
layers of the printer cartridge and focuses the skin graft as it is continuously extruded into a
liquid-filled reservoir and collected onto a rotating drum. (b) A variety of patterns in the form of
(1) alternating spots, (2) parallel spots, (3) void regions, (4) and parallel stripes can be formed in
both single and bilayer skin grafts (5). (c) The skin grafts are subsequently cultured in vitro, and
76
applied on murine wound models in vivo. (d) Photograph of the extrusion reservoir, with the
ability to form up to 6cm wide skin grafts. Scale bars 1cm (a), 4cm (d), 2cm (e).
77
Figure 13. Patterned Single-Layered Sheets. (a) Patterns formed with cell printing solution or
void regions are printed by pressure-driven flow of various solutions loaded into on-chip
cartridges. (b) The patterns generated include: (1) parallel spots, (2) alternating spots, (3) parallel
stripes, and (4) alternating void regions. Actuation pressure constant at 3.5psi, with duty cycles
(1) 300ms open-300ms close, (2) 600ms open-600ms close, (3) fully open, (4) 600ms open-
600ms close. Scale bars 500µm. (c) Mechanical properties of skin grafts (E and UTS, *p<0.001
Day 0 and **p<0.001 Day1, n = 5). Sheets with patterns in (b) were subjected to both x and y
strain (insert, scale bar 5mm). Skin graft patterns consist of (I) base matrix without pattern, (II)
parallel spot (b-1), (III) parallel stripe (b-3), (IV) alternating void (b-4), (V) alternating spot (b-
2). Day 1 corresponds to samples cultured in DMEM during one day.
Figure 14. In vitro characterization of printed single layer skin grafts. Bright field and
fluorescence images of fibroblasts printed as spots (a-c) and stripes (d-f). (a, b, d, e) Day 3, (c, f)
Day 5. Fibroblasts were printed at a concentration of 2×106cells/mL. Actuation pressure of
3.5psi, with duty cycle 300ms open-300ms close for the spotted pattern, and fully open for the
stripe pattern. Cell staining and fluorescence imaging obtained by Cassandra Belo. Scale bars
250µm (a, d), 100µm (b, c, e, f).
78
Figure 15. Material optimization for keratinocyte printing. Human keratinocytes were seeded
at a concentration of 0.1million cells/cm2 on four different material compositions, and their
proliferation assessed as a measure of increase in surface area covered by cells. Materials:M1%
w.t. sodium alginate (Molecule-R) with 2.5mg/mL HA grade 80, M1% w.t. sodium alginate
(Molecule-R) with 2.5mg/mL HA grade 150, 2.15% w.t. high G:M alginate with 2.5mg/mL
HA grade 80, and M 2.15% w.t. high G:M alginate with 2.5mg/mL HA grade 150 (statistical
comparison p<0.001 with * Day 2, ** Day 6, *** Day 9, **** Day 13; n = 5). Samples were
lifted at the air-liquid interface on Day 3. Insert: fluorescence image of keratinocytes stained
with Dapi, Phalloidin, and Keratin14. Cell staining and fluorescence imaging obtained by
Cassandra Belo. Scale bar 10µm.
79
Figure 16. Scalable formation of skin grafts. (a) Rendered images of a 10mm wide (1) and
75mm wide (2) cartridge. (b, c) Measured graft widths and thicknesses versus pulling velocity
produced using four different cartridges: 10mm (●), 25mm (▲), 60mm (♦), and 75mm (■). Flow
rates 100µl/min (green), 150µl/min (red), 200µl/min (blue) for 10mm device. Flow rates for
other cartridges were scaled proportionally by multiplying by 2.5, 6, and 7.5 respectively. Scale
bars 20mm (a-1), 25mm (a-2).
80
Figure 17 In vivo characterization of printed skin grafts. (a) Schematic of animal wound bed
with implanted skin graft. (b) Confocal scan of bilayered graft. Support material and cell printing
solution were loaded with fluorescent microbeads to distinguish between the different regions.
(c, d) Trichrome staining of wounded skin 10 days post-wounding. (e, f) Keratin14 staining
showing keratinization post-wounding. (c, e) Control, no cells. (d, f) 4%w.t. alginate patterned
with human fibroblasts. Histology stains obtained by Cassandra Belo. Scale bars 500µm (b, e, f),
1mm (c, d).
81
Summary and Future Work
This thesis presented a novel microfluidic approach for the continuous and scalable formation of
biologically relevant soft materials with dynamically tunable composition. Directionally varying
compositions lead to directionally-dependent properties such as elastic modulus, ultimate tensile
strength, and diffusivity to various molecules. Viable cells were printed within biomaterial
sheets, promising the generation of heterogeneous soft materials with physiologically relevant
composition. The capability of the presented platform was investigated, namely: the control over
the printed material size, shape, and patterns generated, the incorporation of a variety of
materials possessing distinct mechanical properties and their impact on the macroscale properties
of the printed sheets, and viability and proliferation of various cells post-printing. The potential
application of this technology was demonstrated in two groups of work.
In a first demonstration, we have generated high aspect ratio collagen sheets with widths ranging
from 3-17mm and thicknesses ranging from 30-250µm. The formation of highly aligned collagen
fibers within these high aspect ratio sheets was possible through a combination of flow-induced
shear through a shallow geometric constriction, and strain-induced stretching in the longitudinal
direction through a collection drum. The resulting anisotropic alignment led to high mechanical
strength (E and UTS) comparable to native blood vessels. Cellular alignment and morphological
changes of smooth muscle cells and endothelial cells seeded and cultured on these aligned
collagen sheets was observed over a period of 3 days, demonstrating the direct impact of
collagen fibril alignment on cellular behavior (data obtained by our collaborator Dr. Stephanie
Grainger but not shown in this thesis). Future work will involve the application of these cell-
populated collagen sheets as tubular constructs for replacement arteries in mice and rat models.
In a second demonstration, biomaterial sheets with structural organization and cellular
composition mimicking the physiological composition of native skin were printed. Dermal
fibroblasts were incorporated at a concentration of 2-6×106cells/mL in arrays of spots and
parallel stripes and their viability and attachment was investigated. Epidermal keratinocytes were
incorporated in similar patterns at a higher concentration of 9×106 cells/mL. Material
compositions for cell printing were optimized for each cell types to promote their viability,
82
attachment, and proliferation. A variety of patterns in the form of parallel stripes, parallel spots,
and alternating spots were investigated to select the optimal mechanical strength of the printed
skin grafts for improved handling. In our preliminary in vivo experiments, we have investigated
wound healing on immunodeficient murine models treated with a variety of skin grafts
consisting of single layers without cells or fibroblasts-populated spots. Histological data were
promising as wounds treated with cell-populated grafts regenerated with a full epidermal layer
populated with keratinocytes that had migrated from the host tissue.
As the next steps, we are currently investigating the difference between printed skin grafts
composed of spotted arrays of concentrated cells versus homogeneously cell-populated grafts.
The effects of cell concentration and cell clustering are investigated by comparing single layer
sheets containing: (1) fibroblasts homogeneously incorporated at a concentration of
2×106cells/mL, (2) fibroblasts incorporated at a concentration of 2×106cells/mL within spots
occupying 40% of the sheet surface area, and (3) fibroblasts incorporated at a concentration of
5×106cells/mL within spots occupying 40% of the sheet surface area (Fig. 18). We investigate
the effect of localizing cells within spots versus homogeneously distributing them within an
entire skin graft, while maintaining the same cell density (fig. 18i). Cases 1 and 3 will allow us
to study the effect of increased local cell density within spots, with the same total cell number in
the spotted candidate versus homogeneously populated candidate (Fig. 18ii). The resulting cell
viability and proliferation for all samples will be characterized through live/dead staining, where
cell concentration will be determined as a measure of cell-to-cell distance.
Finally, the co-culture of bilayered sheets containing both keratinocytes and fibroblasts
(epidermal and dermal layers respectively) will be investigated in vitro over a period of 14 to 28
days with the goal of obtaining a basement membrane between the two respective layers,
essential in the formation of functional skin substitutes. Two systems will be compared: bilayer
grafts containing fibroblasts in the bottom layer and keratinocytes in the top layer, and single
layer grafts containing a mixed solution of both keratinocytes and fibroblasts. In the first case,
fibroblasts will be incorporated at a concentration of 2×106cells/mL and keratinocytes at a
concentration of 9×106cells. In the second case, cell-populated spots will be printed with a
solution containing a mixture of 2×106cells/mL fibroblasts and 7×106cells/mL keratinocytes. The
total cell concentration to be printed must not exceed 9×106cells/mL, as higher cell concentration
would result in device clogging. These skin grafts printed with both cell types will be immersed
83
in culture media and cultured over a period of 3 days, following which the grafts will be raised to
the air-liquid interface and further cultured up to 28 days.
These samples will also be grafted onto immunodeficient murine models for wound healing
studies. As a preliminary set of experiments (animal surgeries and histology data obtained by Dr.
Amini-Nik and Cassandra Belo), we have grafted bilayered skin grafts patterned with: cell-free
spots, fibroblasts only, keratinocytes only, and keratinocytes with fibroblasts. In difference to our
in vivo data in figure 17, a plastic dome designed by Dr. Amini-Nik was placed onto 8mm
diameter wound, sealing the wound boundary and preventing in the process both wound
contraction and migration of host cells. As a result, we can postulate that wound healing is
strictly due to our printed skin grafts, as healing from wound contraction and host cell invasion
was inhibited. In our preliminary data, figure 19 shows histology samples obtained by Cassandra
Belo where keratinization was observed after 15 days post-surgery. Samples were stained with
trichrome (a), Keratin 14 (b), and Keratin 10 (c).
84
Figure 18. Benefit of cell clustering and characterization of skin microtissues. (i)
Comparison of homogeneously populated skin graft (1) versus locally printed cells (2). Local
cell concentration are equal between the two samples. (ii) Comparison of homogeneously
populated skin graft (1) versus patterned graft with a 2.5× higher cell density (3). Increased local
cell density in (3) vs (1), with an equal total cell number between the two candidates. Base
material (blue), cell printing solution (red), human dermal fibroblasts (yellow).
85
Figure 19. In vivo characterization of bilayered cell-populated skin grafts. An 8mm diameter
dome with a ring bracket was used to segregate the wound region from the host tissue,
preventing wound contraction and migration of host cells into the wound area. Histology samples
were stained with (a) Trichrome, (b) Keratin 14, and (c) Keratin 10. Preliminary data shows the
onset of keratinocyte layer formation after 15 days post-surgery. Scale bars 500µm.
86
References
1 Fratzl, P. & Weinkamer, R. Nature's hierarchical materials. Progress in Materials
Science 52, 1263-1334, doi:10.1016/j.pmatsci.2007.06.001 (2007).
2 Lakes, R. Materials with structural hierarchy. Nature 361, 511-515,
doi:10.1038/361511a0 (1993).
3 Xu, S. Q. et al. Generation of monodisperse particles by using microfluidics: Control
over size, shape, and composition. Angewandte Chemie-International Edition 44, 724-
728, doi:10.1002/anie.200462226 (2005).
4 Shui, L. L., Kooij, E. S., Wijnperle, D., van den Berg, A. & Eijkel, J. C. T. Liquid
crystallography: 3D microdroplet arrangements using microfluidics. Soft Matter 5, 2708-
2712 (2009).
5 Wu, H. K., Thalladi, V. R., Whitesides, S. H. & Whitesides, G. M. Using hierarchical
self-assembly to form three-dimensional lattices of spheres. Journal of the American
Chemical Society 124, 14495-14502 (2002).
6 van der Net, A., Delaney, G. W., Drenckhan, W., Weaire, D. & Hutzler, S. Crystalline
arrangements of microbubbles in monodisperse foams. Colloids and Surfaces 309, 117-
124 (2006).
7 Cheung, Y. K., Gillette, B. M., Zhong, M., Ramcharan, S. & Sia, S. K. Direct patterning
of composite biocompatible microstructures using microfluidics. Lab on a Chip 7, 574-
579, doi:10.1039/b700869d (2007).
8 Chan, V., Zorlutuna, P., Jeong, J. H., Kong, H. & Bashir, R. Three-dimensional
photopatterning of hydrogels using stereolithography for long-term cell encapsulation.
Lab on a Chip 10, 2062-2070, doi:10.1039/c004285d (2010).
9 Park, H. et al. The significance of pore microarchitecture in a multi-layered elastomeric
scaffold for contractile cardiac muscle constructs. Biomaterials 32, 1856-1864,
doi:10.1016/j.biomaterials.2010.11.032 (2011).
10 Dendukuri, D. & Doyle, P. S. The Synthesis and Assembly of Polymeric Microparticles
Using Microfluidics. Advanced Materials 21, 4071-4086, doi:10.1002/adma.200803386
(2009).
11 Dendukuri, D., Hatton, T. A. & Doyle, P. S. Synthesis and self-assembly of amphiphilic
polymeric microparticles. Langmuir 23, 4669-4674, doi:10.1021/la062512i (2007).
87
12 Dendukuri, D., Pregibon, D. C., Collins, J., Hatton, T. A. & Doyle, P. S. Continuous-flow
lithography for high-throughput microparticle synthesis. Nature Materials 5, 365-369
(2006).
13 Panda, P. et al. Stop-flow lithography to generate cell-laden microgel particles. Lab on a
Chip 8, 1056-1061 (2008).
14 Lee, S. A., Chung, S. E., Park, W., Lee, S. H. & Kwon, S. Three-dimensional fabrication
of heterogeneous microstructures using soft membrane deformation and optofluidic
maskless lithography. Lab on a Chip 9, 1670-1675, doi:10.1039/b819999j (2009).
15 Tolley, M. T., Krishnan, M., Erickson, D. & Lipson, H. Dynamically programmable
fluidic assembly. Applied Physics Letters 93, doi:10.1063/1.3048562 (2008).
16 Krishnan, M., Tolley, M. T., Lipson, H. & Erickson, D. Hydrodynamically Tunable
Affinities for Fluidic Assembly. Langmuir 25, 3769-3774, doi:10.1021/la803517f (2009).
17 Chung, S. E., Park, W., Shin, S., Lee, S. A. & Kwon, S. Guided and fluidic self-assembly
of microstructures using railed microfluidic channels. Nature Materials 7, 581-587
(2008).
18 Yin, Y., Lu, Y., Gates, B. & Xia, Y. Template-assisted self-assembly: a practical route to
complex aggregates of monodispersed colloids with well-defined sizes, shapes, and
structures. Journal of the American Chemical Society 123, 8718-8729 (2001).
19 Matsunaga, Y. T., Morimoto, Y. & Takeuchi, S. Molding Cell Beads for Rapid
Construction of Macroscopic 3D Tissue Architecture. Advanced Materials 23, H90-H94,
doi:10.1002/adma.201004375 (2011).
20 Ballyns, J. J. et al. Image-guided tissue engineering of anatomically shaped implants via
MRI and micro-CT using injection molding. Tissue Engineering Part A 14, 1195-1202,
doi:10.1089/ten.tea.2007.0186 (2008).
21 Malda, J., Martens, D. E., Tramper, J., van Blitterswijk, C. A. & Riesle, J. Cartilage
tissue engineering: Controversy in the effect of oxygen. Critical Reviews in
Biotechnology 23, 175-194, doi:10.1080/714037688 (2003).
22 Nishiguchi, A., Yoshida, H., Matsusaki, M. & Akashi, M. Rapid Construction of Three-
Dimensional Multilayered Tissues with Endothelial Tube Networks by the Cell-
Accumulation Technique. Advanced Materials 23, 3506-+, doi:10.1002/adma.201101787
(2011).
23 Tan, W. & Desai, T. A. Layer-by-layer microfluidics for biomimetic three-dimensional
structures. Biomaterials 25, 1355-1364, doi:10.1016/j.biomaterials.2003.08.021 (2004).
24 Choi, N. W. et al. Microfluidic scaffolds for tissue engineering. Nature Materials 6, 908-
915, doi:10.1038/nmat2022 (2007).
88
25 Ling, Y. et al. A cell-laden microfluidic hydrogel. Lab on a Chip 7, 756-762,
doi:10.1039/b615486g (2007).
26 Vozzi, G., Flaim, C., Ahluwalia, A. & Bhatia, S. Fabrication of PLGA scaffolds using
soft lithography and microsyringe deposition. Biomaterials 24, 2533-2540,
doi:10.1016/s0142-9612(03)00052-8 (2003).
27 Bettinger, C. J. et al. Three-dimensional microfluidic tissue-engineering scaffolds using a
flexible biodegradable polymer. Advanced Materials 18, 165-169 (2006).
28 Golden, A. P. & Tien, J. Fabrication of microfluidic hydrogels using molded gelatin as a
sacrificial element. Lab on a Chip 7, 720-725, doi:10.1039/b618409j (2007).
29 Tang, M. D., Golden, A. P. & Tien, J. Fabrication of collagen gels that contain patterned,
micrometer-scale cavities. Advanced Materials 16, 1345-+, doi:10.1002/adma.200400766
(2004).
30 Yoshida, H., Matsusaki, M. & Akashi, M. Multilayered Blood Capillary Analogs in
Biodegradable Hydrogels for In Vitro Drug Permeability Assays. Advanced Functional
Materials 23, 1736-1742, doi:10.1002/adfm.201201905 (2013).
31 Nazhat, S. N. et al. Controlled microchannelling in dense collagen scaffolds by soluble
phosphate glass fibers. Biomacromolecules 8, 543-551, doi:10.1021/bm060715f (2007).
32 Chen, C. H. et al. Novel living cell sheet harvest system composed of thermoreversible
methylcellulose hydrogels. Biomacromolecules 7, 736-743, doi:10.1021/bm0506400
(2006).
33 Haraguchi, Y. et al. Fabrication of functional three-dimensional tissues by stacking cell
sheets in vitro. Nature Protocols 7, 850-858, doi:10.1038/nprot.2012.027 (2012).
34 Masumoto, H. et al. Pluripotent Stem Cell-Engineered Cell Sheets Reassembled with
Defined Cardiovascular Populations Ameliorate Reduction in Infarct Heart Function
Through Cardiomyocyte-Mediated Neovascularization. Stem Cells 30, 1196-1205,
doi:10.1002/stem.1089 (2012).
35 Tsuda, Y. et al. The use of patterned dual thermoresponsive surfaces for the collective
recovery as co-cultured cell sheets. Biomaterials 26, 1885-1893,
doi:10.1016/j.biomaterials.2004.06.005 (2005).
36 Yang, J. et al. Cell sheet engineering: Recreating tissues without biodegradable scaffolds.
Biomaterials 26, 6415-6422, doi:10.1016/j.biomaterials.2005.04.061 (2005).
37 Yang, J. et al. Reconstruction of functional tissues with cell sheet engineering.
Biomaterials 28, 5033-5043, doi:10.1016/j.biomaterials.2007.07.052 (2007).
38 Nishida, K. et al. Functional bioengineered corneal epithelial sheet grafts from corneal
stem cells expanded ex vivo on a temperature-responsieve cell culture surface.
Transplantation 77, 379-385, doi:10.1097/01.tp.0000110320.45678.30 (2004).
89
39 Shimizu, T. et al. Fabrication of pulsatile cardiac tissue grafts using a novel 3-
dimensional cell sheet manipulation technique and temperature-responsive cell culture
surfaces. Circ.Res. 90, E40-E48, doi:10.1161/hh0302.105722 (2002).
40 Kim, B. S., Choi, J. S., Kim, J. D., Choi, Y. C. & Cho, Y. W. Recellularization of
decellularized human adipose-tissue-derived extracellular matrix sheets with other human
cell types. Cell and Tissue Research 348, 559-567, doi:10.1007/s00441-012-1391-y
(2012).
41 Boland, T., Mironov, V., Gutowska, A., Roth, E. A. & Markwald, R. R. Cell and organ
printing 2: Fusion of cell aggregates in three-dimensional gels. Anatomical Record Part
a-Discoveries in Molecular Cellular and Evolutionary Biology 272A, 497-502,
doi:10.1002/ar.a.10059 (2003).
42 Boland, T., Xu, T., Damon, B. & Cui, X. Application of inkjet printing to tissue
engineering. Biotechnology Journal 1, 910-917, doi:10.1002/biot.200600081 (2006).
43 Mironov, V., Boland, T., Trusk, T., Forgacs, G. & Markwald, R. R. Organ printing:
computer-aided jet-based 3D tissue engineering. Trends in Biotechnology 21, 157-161,
doi:10.1016/s0167-7799(03)00033-7 (2003).
44 Roth, E. A. et al. Inkjet printing for high-throughput cell patterning. Biomaterials 25,
3707-3715, doi:10.1016/j.biomaterials.2003.10.052 (2004).
45 Wilson, W. C. & Boland, T. Cell and organ printing 1: Protein and cell printers.
Anatomical Record Part a-Discoveries in Molecular Cellular and Evolutionary Biology
272A, 491-496, doi:10.1002/ar.a.10057 (2003).
46 Norotte, C., Marga, F. S., Niklason, L. E. & Forgacs, G. Scaffold-free vascular tissue
engineering using bioprinting. Biomaterials 30, 5910-5917,
doi:10.1016/j.biomaterials.2009.06.034 (2009).
47 Mironov, V., Reis, N. & Derby, B. Bioprinting: A beginning. Tissue Engineering 12,
631-634, doi:10.1089/ten.2006.12.631 (2006).
48 Campbell, P. G. & Weiss, L. E. Tissue engineering with the aid of inkjet printers. Expert
Opinion on Biological Therapy 7, 1123-1127, doi:10.1517/14712598.7.8.1123 (2007).
49 Jakab, K. et al. Tissue engineering by self-assembly and bio-printing of living cells.
Biofabrication 2, doi:10.1088/1758-5082/2/2/022001 (2010).
50 Derby, B. Printing and Prototyping of Tissues and Scaffolds. Science 338, 921-926,
doi:10.1126/science.1226340 (2012).
51 Pataky, K. et al. Microdrop Printing of Hydrogel Bioinks into 3D Tissue-Like
Geometries. Advanced Materials 24, 391-+, doi:10.1002/adma.201102800 (2012).
90
52 Bassoli, E., Gatto, A., Iuliano, L. & Violante, M. G. 3D printing technique applied to
rapid casting. Rapid Prototyping J. 13, 148-155, doi:10.1108/13552540710750898
(2007).
53 Ebert, J. et al. Direct Inkjet Printing of dental Prostheses Made of Zirconia. Journal of
Dental Research 88, 673-676, doi:10.1177/0022034509339988 (2009).
54 Mott, M., Song, J. H. & Evans, J. R. G. Microengineering of ceramics by direct ink-Jet
printing. Journal of the American Ceramic Society 82, 1653-1658 (1999).
55 Symes, M. D. et al. Integrated 3D-printed reactionware for chemical synthesis and
analysis. Nat. Chem. 4, 349-354, doi:10.1038/nchem.1313 (2012).
56 Di Biase, M., Saunders, R. E., Tirelli, N. & Derby, B. Inkjet printing and cell seeding
thermoreversible photocurable gel structures. Soft Matter 7, 2639-2646,
doi:10.1039/c0sm00996b (2011).
57 Klebe, R. J. Cytoscribing - A method for micropositioning cells and the construction of
two-dimensional and 3-dimensional synthetic tissues. Experimental Cell Research 179,
362-373, doi:10.1016/0014-4827(88)90275-3 (1988).
58 Wang, Y., Bokor, J. & Lee, A. in Emerging Lithographic Technologies Viii Vol. 5374
Proceedings of the Society of Photo-Optical Instrumentation Engineers (Spie) (ed R. S.
Mackay) 628-636 (2004).
59 Saunders, R. E., Gough, J. E. & Derby, B. Delivery of human fibroblast cells by
piezoelectric drop-on-demand inkjet printing. Biomaterials 29, 193-203 (2008).
60 Hockaday, L. A. et al. Rapid 3D printing of anatomically accurate and mechanically
heterogeneous aortic valve hydrogel scaffolds. Biofabrication 4, 1-22 (2012).
61 Yang, S. F., Leong, K. F., Du, Z. H. & Chua, C. K. The design of scaffolds for use in
tissue engineering. Part II. Rapid prototyping techniques. Tissue Engineering 8, 1-11,
doi:10.1089/107632702753503009 (2002).
62 Cohen, D. L., Malone, E., Lipson, H. & Bonassar, L. J. Direct freeform fabrication of
seeded hydrogels in arbitrary geometries. Tissue Engineering 12, 1325-1335 (2006).
63 Nair, K. et al. Characterization of cell viability during bioprinting processes.
Biotechnology Journal 4, 1168-1177, doi:10.1002/biot.200900004 (2009).
64 Phillippi, J. A. et al. Microenvironments engineered by inkjet bioprinting spatially direct
adult stem cells toward muscle- and bone-like subpopulations. Stem Cells 26, 127-134,
doi:10.1634/stemcells.2007-0520 (2008).
65 Cao, Y. L., Vacanti, J. P., Paige, K. T., Upton, J. & Vacanti, C. A. Transplantation of
chondrocytes utilizing a polymer-cell construct to produce tissue-engineered cartilage in
the shape of a human ear. Plastic and Reconstructive Surgery 100, 297-302 (1997).
91
66 Kojima, K. et al. Comparison of tracheal and nasal chondrocytes for tissue engineering of
trachea. The Annals of Thoracic Surgery 76, 1884 (2003).
67 Kamil, S. H. et al. In vitro tissue engineering to generate a human-sized auricle and nasal
tip. Laryngoscope 113, 90 (2003).
68 Puelacher, W. C. et al. Design of nasoseptal cartilage replacements synthesized from
biodegradable polymers and chondrocytes. Biomaterials 15, 774 (1994).
69 Jakab, K. et al. Tissue engineering by self-assembly of cells printed into topologically
defined structures. Tissue Engineering Part A 14, 413-421, doi:10.1089/tea.2007.0173
(2008).
70 Mironov, V. Printing technology to produce living tissue. Expert Opinion on Biological
Therapy 3, 701-704, doi:10.1517/14712598.3.5.701 (2003).
71 Wang, N., Liu, W., Huang, J. & Ma, K. The structure-mechanical relationship of palm
vascular tissue. Journal of the mechanical behavior of biomedical materials 36, 1-11,
doi:10.1016/j.jmbbm.2014.04.001 (2014).
72 Hutmacher, D. W. Scaffold design and fabrication technologies for engineering tissues -
state of the art and future perspectives. J. Biomater. Sci.-Polym. Ed. 12, 107-124,
doi:10.1163/156856201744489 (2001).
73 Holmes, D. F. et al. Corneal collagen fibril structure in three dimensions: Structural
insights into fibril assembly, mechanical properties, and tissue organization. Proceedings
of the National Academy of Sciences of the United States of America 98, 7307-7312,
doi:10.1073/pnas.111150598 (2001).
74 Boland, E. D. et al. Electrospinning collagen and elastin: Preliminary vascular tissue
engineering. Frontiers in Bioscience 9, 1422-1432, doi:10.2741/1313 (2004).
75 Ottani, V., Raspanti, M. & Ruggeri, A. Collagen structure and functional implications.
Micron 32, 251-260, doi:10.1016/s0968-4328(00)00042-1 (2001).
76 Muller, L. J., Marfurt, C. F., Kruse, F. & Tervo, T. M. T. Corneal nerves: structure,
contents and function. Exp. Eye Res. 76, 521-542, doi:10.1016/s0014-4835(03)00050-2
(2003).
77 Ruberti, J. W., Roy, A. S. & Roberts, C. J. in Annual Review of Biomedical Engineering,
Vol 13 Vol. 13 Annual Review of Biomedical Engineering (eds M. L. Yarmush, J. S.
Duncan, & M. L. Gray) 269-295 (2011).
78 Muller, L. J., Pels, E. & Vrensen, G. The specific architecture of the anterior stroma
accounts for maintenance of corneal curvature. British Journal of Ophthalmology 85,
437-443, doi:10.1136/bjo.85.4.437 (2001).
92
79 Mackenzie, I. C. & Hill, M. W. Connective-tissue influences on patterns of epithelial
architecture and keratinization in skin and oral-mucosa of the adult-mouse. Cell and
Tissue Research 235, 551-559 (1984).
80 Verhaegen, P. et al. Differences in collagen architecture between keloid, hypertrophic
scar, normotrophic scar, and normal skin: An objective histopathological analysis. Wound
Repair and Regeneration 17, 649-656, doi:10.1111/j.1524-475X.2009.00533.x (2009).
81 Chen, X., Nadiarynkh, O., Plotnikov, S. & Campagnola, P. J. Second harmonic
generation microscopy for quantitative analysis of collagen fibrillar structure. Nature
Protocols 7, 654-669, doi:10.1038/nprot.2012.009 (2012).
82 Birk, D. E. & Trelstad, R. L. Extracellular compartments in tendon morphogenesis -
collagen fibril, bundle, and macroaggregate formation. Journal of Cell Biology 103, 231-
240, doi:10.1083/jcb.103.1.231 (1986).
83 Sharma, P. & Maffulli, N. Current concepts review tendon injury and tendinopathy:
Healing and repair. Journal of Bone and Joint Surgery-American Volume 87A, 187-202,
doi:10.2106/jbjs.d.01850 (2005).
84 Gelse, K., Poschl, E. & Aigner, T. Collagens - structure, function, and biosynthesis.
Advanced Drug Delivery Reviews 55, 1531-1546, doi:10.1016/j.addr.2003.08.002 (2003).
85 Zhao, J.-Y. et al. Influence of hyaluronic acid on wound healing using composite porcine
acellular dermal matrix grafts and autologous skin in rabbits. International Wound
Journal 10, 562-572, doi:10.1111/j.1742-481X.2012.01023.x (2013).
86 Brodsky, B., Eikenberry, E. F. & Cassidy, K. Unusual collagen periodicity in skin.
Biochimica Et Biophysica Acta 621, 162-166, doi:10.1016/0005-2795(80)90072-0
(1980).
87 Hofmann, H., Fietzek, P. P. & Kuhn, K. Role of polar and hydrophobic interactions for
molecular packing of type-I collagen - 3-dimensional evaluation of amino-acid sequence.
Journal of Molecular Biology 125, 137-165, doi:10.1016/0022-2836(78)90342-x (1978).
88 Amiel, D., Frank, C., Harwood, F., Fronek, J. & Akeson, W. Tendons and ligaments: A
morphological and biochemical comparison. Journal of Orthopaedic Research 1, 257-
265 (1984).
89 Kastelic, J., Galeski, A. & Baer, E. Multicomposite structure of tendon. Connective
Tissue Research 6, 11-23, doi:10.3109/03008207809152283 (1978).
90 Diamant, J., Arridge, R. G. C., Baer, E., Litt, M. & Keller, A. Collagen-ultrastructure and
its relation to mechanical properties as a function of aging. Proceedings of the Royal
Society Series B-Biological Sciences 180, 293-+, doi:10.1098/rspb.1972.0019 (1972).
91 Misof, K., Rapp, G. & Fratzl, P. A new molecular model for collagen elasticity based on
synchrotron x-ray scattering evidence. Biophysical Journal 72, 1376-1381 (1997).
93
92 Komai, Y. & Ushiki, T. The 3-dimensional organization of collagen fibrils in the human
cornea and sclera. Investigative Ophthalmology & Visual Science 32, 2244-2258 (1991).
93 Daxer, A., Misof, K., Grabner, B., Ettl, A. & Fratzl, P. Collagen fibrils in the human
corneal stroma: Structure and aging. Investigative Ophthalmology & Visual Science 39,
644-648 (1998).
94 Rafat, M. et al. PEG-stabilized carbodiimide crosslinked collagen-chitosan hydrogels for
corneal tissue engineering. Biomaterials 29, 3960-3972,
doi:10.1016/j.biomaterials.2008.06.017 (2008).
95 Wollensak, G., Spoerl, E. & Seiler, T. Stress-strain measurements of human and porcine
corneas after riboflavin-ultraviolet-A-induced cross-linking. Journal of Cataract and
Refractive Surgery 29, 1780-1785, doi:10.1016/s0886-3350(03)00407-3 (2003).
96 Bigi, A., Ripamonti, A. & Roveri, N. X-ray investigation of the orientation of collagen
fibers in aortic media layer under distending pressure. International Journal of Biological
Macromolecules 3, 287-291, doi:10.1016/0141-8130(81)90044-1 (1981).
97 van Zuijlen, P. P. et al. Collagen morphologyin human skin and scar tissue: no
adaptations in response to mechanical loading at joints. Burns 29, 423-431 (2003).
98 Purslow, P. P., Bigi, A., Ripamonti, A. & Roveri, N. Collagen fiber reorientation around
a crack in biaxially stretched aortic media. International Journal of Biological
Macromolecules 6, 21-25, doi:10.1016/0141-8130(84)90042-4 (1984).
99 Jarrett, A. The structure of collagen and elastic tissues in unprocessed skin. Br. J.
Dermatol. 70, 343-347, doi:10.1111/j.1365-2133.1958.tb13794.x (1958).
100 Beenakker, J. W. M., Ashcroft, B. A., Lindeman, J. H. N. & Oosterkamp, T. H.
Mechanical properties of the extracellular matrix of the aorta studied by enzymatic
treatments. Biophysical Journal 8, 1731-1737 (2012).
101 Berillis, P. The role of collagen in the aorta's structure. The Open Circulation and
Vascular Journal 6, 1-8 (2013).
102 Canham, P. B., Finlay, H. M. & Boughner, D. R. Contrasting structure of the saphenous
vein and internal mammary artery used as coronary bypass vessels. Cardiovascular
Research 34, 557-567, doi:10.1016/s0008-6363(97)00056-4 (1997).
103 Shadwick, R. E. Mechanical design in arteries. Journal of Experimental Biology 202,
3305-3313 (1999).
104 Achilli, M. & Mantovani, D. Tailoring mechanical properties of collagen-based scaffolds
for vascular tissue engineering: the effects of pH, temperature and ionic strength on
gelation. Polymers 2, 664-680, doi:10.3390/polym2040664 (2010).
105 Weinberg, C. B. & Bell, E. A blood vessel model constructed from collagen and cultured
vascular cells. Science 231, 397-400 (1986).
94
106 L'Heureux, N., Pâquet, S., Labbé, R., Germain, L. & Auger, F. A. A completely
biological tissue-engineered human blood vessel. FASEB J. 12, 47-56 (1998).
107 Berglund, J. D., Mohseni, M. M., Nerem, R. M. & Sambanis, A. A biological hybrid
model for collagen-based tissue engineered vascular constructs. Biomaterials 24, 1241-
1254 (2003).
108 Cukierman, E., Pankov, R., Stevens, D. R. & Yamada, K. M. Taking cell-matrix
adhesions to the third dimension. Science 294, 1708-1712, doi:10.1126/science.1064829
(2001).
109 O'Shaughnessy, T. J., Lin, H. J. & Ma, W. Functional synapse formation among rat
cortical neurons grown on three-dimensional collagen gels. Neuroscience Letters 340,
169-172, doi:10.1016/s0304-3940(03)00083-1 (2003).
110 Ruggeri, Z. M. Platelets in atherothrombosis. Nature Medicine 8, 1227-1234 (2002).
111 Matsumoto, N. et al. Effect of alignment of the transplanted graft extracellular matrix on
cellular repopulation and newly synthesized collagen. Archives of Orthopaedic and
Trauma Surgery 117, 215-221, doi:10.1007/s004020050232 (1998).
112 Caves, J. M. et al. Fibrillogenesis in Continuously Spun Synthetic Collagen Fiber.
Journal of Biomedical Materials Research Part B-Applied Biomaterials 93B, 24-38,
doi:10.1002/jbm.b.31555 (2010).
113 Koester, S., Evans, H. M., Wong, J. Y. & Pfohl, T. An in situ study of collagen self-
assembly processes. Biomacromolecules 9, 199-207, doi:10.1021/bm700973t (2008).
114 Hakansson, K. M. O. et al. Hydrodynamic alignment and assembly of nanofibrils
resulting in strong cellulose filaments. Nature Communications 5,
doi:10.1038/ncomms5018 (2014).
115 Lanfer, B. et al. Aligned fibrillar collagen matrices obtained by shear flow deposition.
Biomaterials 29, 3888-3895, doi:10.1016/j.biomaterials.2008.06.016 (2008).
116 Lai, E. S., Huang, N. F., Cooke, J. P. & Fuller, G. G. Aligned nanofibrillar collagen
regulates endothelial organization and migration. Regenerative Medicine 7, 649-661,
doi:10.2217/rme.12.48 (2012).
117 Eastwood, M., Porter, R., Khan, U., McGrouther, G. & Brown, R. Quantitative analysis
of collagen gel contractile forces generated by dermal fibroblasts and the relationship to
cell morphology. Journal of Cellular Physiology 166, 33-42, doi:10.1002/(sici)1097-
4652(199601)166:1<33::aid-jcp4>3.0.co;2-h (1996).
118 Thomopoulos, S., Fomovsky, G. M. & Holmes, J. W. The development of structural and
mechanical anisotropy in fibroblast populated collagen gels. Journal of Biomechanical
Engineering-Transactions of the Asme 127, 742-750, doi:10.1115/1.1992525 (2005).
95
119 Lee, P., Lin, R., Moon, J. & Lee, L. P. Microfluidic alignment of collagen fibers for in
vitro cell culture. Biomedical Microdevices 8, 35-41, doi:10.1007/s10544-006-6380-z
(2006).
120 Cheng, X. et al. An electrochemical fabrication process for the assembly of
anisotropically oriented collagen bundles. Biomaterials 29, 3278-3288,
doi:10.1016/j.biomaterials.2008.04.028 (2008).
121 Xu, B., Chow, M.-J. & Zhang, Y. Experimental and modeling study of collagen scaffolds
with the effects of crosslinking and fiber alignment. International journal of biomaterials
2011, 172389-172389, doi:10.1155/2011/172389 (2011).
122 Guo, C. & Kaufman, L. J. Flow and magnetic field induced collagen alignment.
Biomaterials 28, 1105-1114, doi:10.1016/j.biomaterials.2006.10.010 (2007).
123 Novak, T., Shannon, G., Mousoulis, C., Voytik-Harbin, S. L. & Neu, C. P. Controlled
fibrillogenesis for improved magnetic alignment of collagen. J. Tissue Eng. Regen. Med.
8, 268-269 (2014).
124 Torbet, J. & Ronziere, M. C. Magnetic alignment of collagen during self-assembly.
Biochem. J. 219, 1057-1059 (1984).
125 Barocas, V. H., Girton, T. S. & Tranquillo, R. T. Engineered alignment in media
equivalents: Magnetic prealignment and Mandrel compaction. Journal of Biomechanical
Engineering-Transactions of the Asme 120, 660-666, doi:10.1115/1.2834759 (1998).
126 Torbet, J. et al. Orthogonal scaffold of magnetically aligned collagen lamellae for corneal
stroma reconstruction. Biomaterials 28, 4268-4276,
doi:10.1016/j.biomaterials.2007.05.024 (2007).
127 Oryan, A., Moshiri, A. & Meimandi-Parizi, A. In vitro characterization of a novel tissue
engineered based hybridized nano and micro structured collagen implant and its in vivo
role on tenoinduction, tenoconduction, tenogenesis and tenointegration. Journal of
Materials Science-Materials in Medicine 25, 873-897, doi:10.1007/s10856-013-5110-3
(2014).
128 Zhong, S. P. et al. An aligned nanofibrous collagen scaffold by electrospinning and its
effects on in vitro fibroblast culture. J. Biomed. Mater. Res. Part A 79A, 456-463,
doi:10.1002/jbm.a.30870 (2006).
129 Xie, J. W., Li, X. R. & Xia, Y. N. Putting Electrospun Nanofibers to Work for
Biomedical Research. Macromol. Rapid Commun. 29, 1775-1792,
doi:10.1002/marc.200800381 (2008).
130 Builles, N. et al. Use of magnetically oriented orthogonal collagen scaffolds for hemi-
corneal reconstruction and regeneration. Biomaterials 31, 8313-8322,
doi:10.1016/j.biomaterials.2010.07.066 (2010).
96
131 Namani, R., Wood, M. D., Sakiyama-Elbert, S. E. & Bayly, P. V. Anisotropic mechanical
properties of magnetically aligned fibrin gels measured by magnetic resonance
elastography. J. Biomech. 42, 2047-2053, doi:10.1016/j.jbiomech.2009.06.007 (2009).
132 Ng, C. P. & Swartz, M. A. Fibroblast alignment under interstitial fluid flow using a novel
3-D tissue culture model. American Journal of Physiology-Heart and Circulatory
Physiology 284, H1771-H1777, doi:10.1152/ajpheart.01008.2002 (2003).
133 Ng, C. P. & Swartz, M. A. Mechanisms of interstitial flow-induced remodeling of
fibroblast-collagen cultures. Annals of Biomedical Engineering 34, 446-454,
doi:10.1007/s10439-005-9067-3 (2006).
134 Dahl, S. L. M., Vaughn, M. E. & Niklason, L. E. An ultrastructural analysis of collagen
in tissue engineered arteries. Annals of Biomedical Engineering 35, 1749-1755,
doi:10.1007/s10439-007-9340-8 (2007).
135 Dahl, S. L. M., Rhim, C., Song, Y. C. & Niklason, L. E. Mechanical properties and
compositions of tissue engineered and native arteries. Annals of Biomedical Engineering
35, 348-355, doi:10.1007/s10439-006-9226-1 (2007).
136 Gervais, T., El-Ali, J., Gunther, A. & Jensen, K. F. Flow-induced deformation of shallow
microfluidic channels. Lab on a Chip 6, 500-507, doi:10.1039/b513524a (2006).
137 Caves, J. M. et al. The use of microfiber composites of elastin-like protein matrix
reinforced with synthetic collagen in the design of vascular grafts. Biomaterials 31, 7175-
7182, doi:10.1016/j.biomaterials.2010.05.014 (2010).
138 Caves, J. M. et al. Elastin-like protein matrix reinforced with collagen microfibers for
soft tissue repair. Biomaterials 32, 5371-5379, doi:10.1016/j.biomaterials.2011.04.009
(2011).
139 McDonald, J. C. et al. Fabrication of microfluidic systems in poly(dimethylsiloxane).
Electrophoresis 21, 27-40, doi:10.1002/(sici)1522-2683(20000101)21:1<27::aid-
elps27>3.3.co;2-3 (2000).
140 Cuneo, P., Magri, E., Verzola, A. & Grazi, E. Macromolecular crowding is a primary
factor in the organization of the cytoskeleton. Biochem. J. 281, 507-512 (1992).
141 Zhou, H. X., Rivas, G. & Minton, A. P. Macromolecular crowding and confinement:
biochemical, biophysical, and potential physiological consequences. Annual Review
Biophysics 37, 375-397 (2008).
142 Minton, A. P. The influence of macromolecular crowding and macromolecular
confinement on biochemical reactions in physiological media. Journal of Biological
Chemistry 276, 10577-10580 (2001).
143 Saeidi, N. et al. Molecular crowding of collagen: A pathway to produce highly-organized
collagenous structures. Biomaterials 33, 7366-7374,
doi:10.1016/j.biomaterials.2012.06.041 (2012).
97
144 Cavallaro, J. F., Kemp, P. D. & Kraus, K. H. Collagen fabrics as biomaterials.
Biotechnology and Bioengineering 44, 146 (1994).
145 Kato, Y. P. & Silver, F. H. Formation of continuous collagen fibres: evaluation of
biocompatibility and mechanical properties. Biomaterials 11, 169-175 (1990).
146 Paten, J. A. et al. Utility of an optically-based, micromechanical system for printing
collagen fibers. Biomaterials 34, 2577-2587, doi:10.1016/j.biomaterials.2012.12.028
(2013).
147 Kemp, P. D., Cavallaro, J. F. & Hastings, D. N. Effects of carbodiimide crosslinking and
load environment on the remodeling of collagen scaffolds. Tissue Engineering 1, 71-79
(1995).
148 Zeugolis, D. I., Paul, R. G. & Attenburrow, G. Extruded collagen fibres for tissue-
engineering applications: influence of collagen concentration and NaCl amount. J.
Biomater. Sci.-Polym. Ed. 20, 219-234 (2009).
149 Silver, F. H. & Trelstad, R. L. Type-I collagen in solution - structure and properties of
fibril fragments. Journal of Biological Chemistry 255, 9427-9433 (1980).
150 Saeidi, N., Sander, E. A. & Ruberti, J. W. Dynamic shear-influenced collagen self-
assembly. Biomaterials 30, 6581-6592, doi:10.1016/j.biomaterials.2009.07.070 (2009).
151 Goublomme, A., Draily, B. & Crochet, M. J. Numerical prediction of extrudate swell of a
high-density polyethylene. J. Non-Newton. Fluid Mech. 44, 171-195, doi:10.1016/0377-
0257(92)80050-8 (1992).
152 Mitsoulis, E., Abdali, S. S. & Markatos, N. C. Flow simulation of herschel-bulkley fluids
through extrusion dies. Can. J. Chem. Eng. 71, 147-160 (1993).
153 Kumar, V. A. C., J. M.; Haller, C. A.; Dai, E.; Liu, L.; Grainger, S.; Chaikof, E. L.
Acellular vascular grafts generated from collagen and elastin analogs. Acta biomaterialia
9, 8067-8074 (2013).
154 Tonnesen, H. H. & Karlsen, J. Alginate in drug delivery systems. Drug Development and
Industrial Pharmacy 28, 621-630 (2002).
155 Kikuchi, A. & Okano, T. Pulsatile drug release control using hydrogels. Advanced Drug
Delivery Reviews 54, 53-77 (2002).
156 Joki, T. et al. Continuous release of endostatin from microencapsulated engineered cells
for tumor therapy. Nature biotechnology 19, 35-39 (2001).
157 Drury, J. L., Boontheeku, T. & Mooney, D. J. Cellular cross-linking of peptide modified
hydrogels. Journal of Biomechanical Engineering-Transactions of the Asme 127, 220-
228, doi:10.1118/1.1865194 (2005).
98
158 Rowley, J. A., Madlambayan, G. & Mooney, D. J. Alginate hydrogels as synthetic
extracellular matrix materials. Biomaterials 20, 45-53, doi:10.1016/s0142-
9612(98)00107-0 (1999).
159 Augst, A. D., Kong, H. J. & Mooney, D. J. Alginate hydrogels as biomaterials.
Macromolecular Bioscience 6, 623-633 (2006).
160 Smidsrod, O. & Skjakbraek, G. Alginate as immobilization matrix for cells. Trends in
Biotechnology 8, 71-78, doi:10.1016/0167-7799(90)90139-o (1990).
161 Johnson, F. A., Craig, D. Q. M. & Mercer, A. D. Characterization of the block structure
and molecular weight of sodium alginates. Journal of Pharmaceutical and Pharmacology
49, 639-643 (1997).
162 Smidrod, O. & Skjakbraek, G. Alginate as immobilization matrix for cells. TIBTECH 8,
71-78 (1990).
163 Paige, K. T., Cima, L. G., Yaremchuk, M. J., Vacanti, J. P. & Vacanti, C. A. Injectable
cartilage. Plastic and Reconstructive Surgery 96, 1390-1398 (1995).
164 Chang, S. C. N. et al. Injection molding of chondrocyte/alginate constructs in the shape
of facial implants. Journal of Biomedical Materials Research 55, 503-511 (2001).
165 Stevens, M. M., Qanadilo, H. F., Langer, R. & Shastri, V. P. A rapid-curing alginate gel
system: utility in periosteum-derived cartilage tissue engineering. Biomaterials 25, 887-
894, doi:10.1016/j.biomaterials.2003.07.002 (2004).
166 Lee, K. Y., Alsberg, E. & Mooney, D. J. Degradable and injectable poly(aldehyde
guluronate) hydrogels for bone tissue engineering. J Biomed Mater Res 56, 228-233
(2001).
167 Alsberg, E., Anderson, K. W., Albeiruti, A., Franceschi, R. T. & Mooney, D. J. Cell-
interactive alginate hydrogels for bone tissue engineering. Journal of Dental Research
80, 2025-2029 (2001).
168 Okano, T. & Matsuda, T. Tissue engineered skeletal muscle: preparation of highly dense,
highly oriented hybrid skeletal muscular tissues. Cell Transplantation 7, 71-82 (1998).
169 Prabakaran, G. & Hoti, S. L. Immobilization of alginate-encapsulated Bacillus
thuringiensis var. israelensis containing different multivalent counterions for mosquito
control. Current Microbiology 57, 111-114, doi:10.1007/s00284-008-9159-z (2008).
170 Kachouie, N. N. et al. Directed assembly of cell-laden hydrogels for engineering
functional tissues. Organogenesis 6, 234-244, doi:10.4161/org6.4.12650 (2010).
171 Li, X. et al. Culture of neural stem cells in calcium alginate beads. Biotechnology
Progress 22, 1683-1689, doi:10.1021/bp060185z (2006).
99
172 Skjakbraek, G., Grasdalen, H. & Smidrod, O. Inhomogeneous polysaccharide ionic gels.
Carbohydrate Polymers 10, 31-54 (1989).
173 Tanaka, H., Matsumura, M. & Veliky, I. A. Diffusion characteristics of substrates in Ca-
alginate gel beads. Biotechnology and Bioengineering 26, 53-58 (1984).
174 Kuo, C. K. & Ma, P. X. Ionically crosslinked alginate hydrogels as scaffolds for tissue
engineering: Part 1. Structure, gelation rate and mechanical properties. Biomaterials 22,
511-521, doi:10.1016/s0142-9612(00)00201-5 (2001).
175 A., M. & Elgsaeter, A. Density distribution of calcium-induced alginate gels.
Biopolymers 36, 17-41 (1995).
176 Iyer, R. K., Chui, J. & Radisic, M. Spatiotemporal tracking of cells in tissue-engineered
cardiac organoids. J. Tissue Eng. Regen. Med. 3, 196-207 (2009).
177 Plouffe, B. D., Brown, M. A., Iyer, R. K., Radisic, M. & Murthy, S. K. Controlled
capture and release of cardiac fibroblasts using peptide-functionalized alginate gels in
microfluidic channels. Lab on a Chip 9, 1507-1510 (2009).
178 Unger, M. A., Chou, H. P., Thorsen, T., Scherer, A. & Quake, S. R. Monolithic
microfabricated valves and pumps by multilayer soft lithography. Science 288, 113-116
(2000).
179 Knight, J. B., Vishwanath, A., Brody, J. P. & Austin, R. H. Hydrodynamic focusing on a
silicon chip: Mixing nanoliters in microseconds. Physical Review Letters 80, 3863-3866
(1998).
180 Anna, S. L., Bontoux, N. & Stone, H. A. Formation of dispersions using "flow focusing"
in microchannels. Applied Physics Letters 82, 364-366 (2003).
181 Fang, Y. et al. Binding behavior of calcium to polyuronates: Comparison of pectin with
alginate. Carbohydrate Polymers 72, 334-341 (2008).
182 Chen, C. H., Shah, R. K., Abate, A. R. & Weitz, D. A. Janus particles templated from
double emulsion droplets generated using microfluidics. Langmuir 25, 4320-4323 (2009).
183 Bong, K. W., Bong, K. T., Pregibon, D. C. & Doyle, P. S. Hydrodynamic focusing
lithography. Angewandte Chemie-International Edition 49, 87-90 (2010).
184 Nie, Z. H., Li, W., Seo, M., Xu, S. Q. & Kumacheva, E. Janus and ternary particles
generated by microfluidic synthesis: Design, synthesis, and self-assembly. Journal of the
American Chemical Society 128, 9408-9412 (2006).
185 Kang, E. et al. Digitally tunable physicochemical coding of material composition and
topography in continuous microfibres. Nature Materials 10, 877-883 (2011).
186 Derda, R. et al. Multizone paper platform for 3D cell cultures. Plos One 6 (2011).
100
187 Nagakura, T. et al. Effect of viscous injectable pure alginate sol on cultured fibroblasts.
Plastic and Reconstructive Surgery 116, 831-838 (2005).
188 Pokrywczynska, M., Drewa, T., Jundzill, A. & Lysik, J. Alginate is not a good material
for growth of rapidly proliferating cells. Transplantation Proceedings 40, 1664-1667
(2008).
189 Radisic, M. et al. Oxygen gradients correlate with cell density and cell viability in
engineered cardiac tissue. Biotechnol Bioeng 93, 332-343 (2006).
190 Bird, R. B., Stewart, W. E. & Lightfoot, E. N. Transport Phenomena. Second edn, 18
(Wiley, 2007).
191 Du, Y., Lo, E., Ali, S. & Khademhosseini, A. Directed assembly of cell-laden microgels
for fabrication of 3D tissue constructs. Proc Natl Acad Sci U S A 105, 9522-9527 (2008).
192 Bruzewicz, D. A., McGuigan, A. P. & Whitesides, G. M. Fabrication of a modular tissue
construct in a microfluidic chip. Lab Chip 8, 663-671 (2008).
193 Qi, H. et al. Patterned differentiation of individual embryoid bodies in spatially organized
3D hybrid microgels. Advanced Materials 22, 5276-5281 (2010).
194 Fernandez, J. G. & Khademhosseini, A. Micro-masonry: construction of 3D structures by
microscale self-assembly. Advanced Materials 22, 2538-2541 (2010).
195 Naito, H. et al. Optimizing engineered heart tissue for therapeutic applications as
surrogate heart muscle. Circulation 114, I72-78 (2006).
196 Brown, M. A., Iyer, R. K. & Radisic, M. Pulsatile perfusion bioreactor for cardiac tissue
engineering. Biotechnol Prog 24, 907-920 (2008).
197 Davies, P. F. Flow-mediated endothelial mechanotransduction. Physiological Reviews 75,
519-560 (1995).
198 Dewey, C. F., Bussolari, S. R., Gimbrone, M. A. & Davies, P. F. The dynamic response
of vascular endothelial cells to fluid shear stress. J. Biomech. Eng.-Trans. ASME 103,
177-185 (1981).
199 Born, C., Zhang, Z., Alrubeai, M. & Thomas, C. R. Estimation of disruption of animal-
cells by laminar shear-stress. Biotechnology and Bioengineering 40, 1004-1010 (1992).
200 Lee, W. et al. On-demand three-dimensional freeform fabrication of multi-layered
hydrogel scaffold with fluidic channels. Biotechnology and Bioengineering 105, 1178-
1186 (2010).
201 Pregibon, D. C., Toner, M. & Doyle, P. S. Multifunctional encoded particles for high-
throughput biomolecule analysis. Science 315, 1393-1396 (2007).
101
202 Chapin, S. C., Pregibon, D. C. & Doyle, P. S. High-throughput flow alignment of
barcoded hydrogel microparticles. Lab on a Chip 9, 3100-3109 (2009).
203 Lee, H., Kim, J., Kim, H. & Kwon, S. Colour-barcoded magnetic microparticles for
multiplexed bioassays. Nature Materials 9, 745-749 (2010).
204 Carrier, R. L. et al. Control of the structure and metabolism of engineered cardiac muscle
by direct perfusion of culture medium. Abstracts of Papers American Chemical Society
219, 76 (2000).
205 Carrier, R. L. et al. Perfusion improves tissue architecture of engineered cardiac muscle.
Tissue Engineering 8, 175-188 (2002).
206 Casey, T. M. & Arthur, P. G. Hibernation in noncontracting mammalian cardiomyocytes.
Circulation 102, 3124-3129 (2000).
207 Ribeiro, J. F., Martins dos Anjos, E. H., Mello, M. L. S. & Vidal, B. d. C. Skin Collagen
Fiber Molecular Order: A Pattern of Distributional Fiber Orientation as Assessed by
Optical Anisotropy and Image Analysis. Plos One 8, doi:10.1371/journal.pone.0054724
(2013).
208 Gibson, T., Kenedi, R. M. & Craik, J. E. Mobile micro-architecture of dermal collagen -
A bio-engineering study. British Journal of Surgery 52, 764-&,
doi:10.1002/bjs.1800521017 (1965).
209 Sakai, S., Yamanari, M., Lim, Y., Nakagawa, N. & Yasuno, Y. In vivo evaluation of
human skin anisotropy by polarization-sensitive optical coherence tomography.
Biomedical Optics Express 2, 2623-2631 (2011).
210 Ross, M. H. P., W. in Integumentary System 442-475 (2003).
211 Kanitakis, J. Anatomy, histology and immunohistochemistry of normal human skin.
European Journal of Dermatology 12, 390-400 (2002).
212 Pins, G. D., Toner, M. & Morgan, J. R. Microfabrication of an analog of the basal
lamina: biocompatible membranes with complex topographies. Faseb Journal 14, 593-
602 (2000).
213 Odland, G. F. The morphology of the attachment between the dermis and the epidermis.
Anatomical Record 108, 399-413, doi:10.1002/ar.1091080305 (1950).
214 Downing, B. R., Cornwell, K., Toner, M. & Pins, G. D. The influence of microtextured
basal lamina analog topography on keratinocyte function and epidermal organization.
Journal of Biomedical Materials Research Part A 72A, 47-56, doi:10.1002/jbm.a.30210
(2005).
215 Bush, K. A. & Pins, G. D. Development of Microfabricated Dermal Epidermal
Regenerative Matrices to Evaluate the Role of Cellular Microenvironments on Epidermal
102
Morphogenesis. Tissue Engineering Part A 18, 2343-2353,
doi:10.1089/ten.tea.2011.0479 (2012).
216 Vracko, R. Basal lamina scaffold-anatomy and significance for maintenance of orderly
tissue structure. American Journal of Pathology 77, 314-346 (1974).
217 Bader, D. L. & Bowker, P. Mechanical characteristics of skin and underlying tissues in
vivo. Biomaterials 4, 305-308, doi:10.1016/0142-9612(83)90033-9 (1983).
218 Hendriks, F. M. et al. A numerical-experimental method to characterize the non-linear
mechanical behaviour of human skin. Skin Research and Technology 9, 274-283,
doi:10.1034/j.1600-0846.2003.00019.x (2003).
219 Diridollou, S. et al. In vivo model of the mechanical properties of the human skin under
suction. Skin Research and Technology 6, 214-221, doi:10.1034/j.1600-
0846.2000.006004214.x (2000).
220 Manschot, J. F. M. & Brakkee, A. J. M. The measurement and modeling of the
mechanical properties of human skin in vivo. 1. The measurement. J. Biomech. 19, 511-
515, doi:10.1016/0021-9290(86)90124-7 (1986).
221 Escoffier, C. et al. Age-related mechanical-properties of human-skin - An in vivo study.
Journal of Investigative Dermatology 93, 353-357, doi:10.1111/1523-1747.ep12280259
(1989).
222 Agache, P. G., Monneur, C., Leveque, J. L. & Derigal, J. Mechanical properties and
Young's modulus of human skin in vivo. Archives of Dermatological Research 269, 221-
232, doi:10.1007/bf00406415 (1980).
223 Miller, S. F. et al. National burn repository 2007 report: A synopsis of the 2007 call for
data. Journal of Burn Care & Research 29, 862-870,
doi:10.1097/BCR.0b013e31818cb046 (2008).
224 Tanner, J. C., Jr., Vandeput, J. & Olley, J. F. The mesh skin graft. Plastic and
Reconstructive Surgery 34, 287-292 (1964).
225 Warden, G. D., Saffle, J. R. & Kravitz, M. A 2-stage technique for excision and grafting
of burn wounds. Journal of Trauma-Injury Infection and Critical Care 22, 98-103,
doi:10.1097/00005373-198202000-00004 (1982).
226 Delvoye, P. et al. Fibroblasts induce the assembly of the macromolecules of the
basement-membrane. Journal of Investigative Dermatology 90, 276-282,
doi:10.1111/1523-1747.ep12456042 (1988).
227 Demarchez, M., Hartmann, D. J., Regnier, M. & Asselineau, D. The role of fibroblasts in
dermal vascularization and remodeling of reconstructed human skin after transplantation
onto the nude-mouse. Transplantation 54, 317-326, doi:10.1097/00007890-199208000-
00023 (1992).
103
228 Sahota, P. S. et al. Development of a reconstructed human skin model for angiogenesis.
Wound Repair and Regeneration 11, 275-284, doi:10.1046/j.1524-475X.2003.11407.x
(2003).
229 MacNeil, S. Progress and opportunities for tissue-engineered skin. Nature 445, 874-880,
doi:10.1038/nature05664 (2007).
230 Sheridan, R. Closure of the excised burn wound: autografts, semipermanent skin
substitutes, and permanent skin substitutes. Clinics in Plastic Surgery 36, 643-+,
doi:10.1016/j.cps.2009.05.010 (2009).
231 Burke, J. F., Yannas, I. V., Quinby, W. C., Bondoc, C. C. & Jung, W. K. Successful use
of a physiologically acceptable skin in the treatment of extensive burn injury. Annals of
Surgery 194, 413-428, doi:10.1097/00000658-198110000-00005 (1981).
232 Streit, M. & Braathen, L. R. Apligraf - a living human skin equivalent for the treatment
of chronic wounds. International Journal of Artificial Organs 23, 831-833 (2000).
233 Zaulyanov, L. & Kirsner, R. S. A review of a bi-layered living cell treatment (Apligraf
(R)) in the treatment of venous leg ulcers and diabetic foot ulcers. Clinical Interventions
in Aging 2, 93-98, doi:10.2147/ciia.2007.2.1.93 (2007).
234 Wood, F. M., Stoner, M. L., Fowler, B. V. & Fear, M. W. The use of a non-cultured
autologous cell suspension and Integra (R) dermal regeneration template to repair full-
thickness skin wounds in a porcine model: A one-step process. Burns 33, 693-700,
doi:10.1016/j.burns.2006.10.388 (2007).
235 Cuono, C., Langdon, R. & McGuire, J. Use of cultured epidermal autografts and dermal
allografts as skin replacement after burn injury. Lancet 1, 1123-1124 (1986).
236 Jones, I., Currie, L. & Martin, R. A guide to biological skin substitutes. British Journal of
Plastic Surgery 55, 185-193, doi:10.1054/bjps.2002.3800 (2002).
237 Boyce, S. T. et al. Surface electrical capacitance as a noninvasive index of epidermal
barrier in cultured skin substitutes in athymic mice. Journal of Investigative Dermatology
107, 82-87, doi:10.1111/1523-1747.ep12298286 (1996).
238 Boyce, S. T., Supp, A. P., Swope, V. B. & Warden, G. D. Vitamin C regulates
keratinocyte viability, epidermal barrier, and basement membrane in vitro, and reduces
wound contraction after grafting of cultured skin substitutes. Journal of Investigative
Dermatology 118, 565-572, doi:10.1046/j.1523-1747.2002.01717.x (2002).
239 Hackl, F. et al. Moist dressing coverage supports proliferation and migration of
transplanted skin micrografts in full-thickness porcine wounds. Burns 40, 274-280,
doi:10.1016/j.burns.2013.06.002 (2014).
240 Hackl, F. et al. Epidermal Regeneration by Micrograft Transplantation with Immediate
100-Fold Expansion. Plastic and Reconstructive Surgery 129, 443E-452E,
doi:10.1097/PRS.0b013e318241289c (2012).
104
241 Breuing, K., Eriksson, E., Liu, P. & Miller, D. R. Healing of partial thickness porcine
skin wounds in a liquid environment. J. Surg. Res. 52, 50-58, doi:10.1016/0022-
4804(92)90278-8 (1992).
242 Svensjo, T., Pomahac, B., Yao, F., Slama, J. & Eriksson, E. Accelerated healing of full-
thickness skin wounds in a wet environment. Plastic and Reconstructive Surgery 106,
602-612, doi:10.1097/00006534-200009030-00012 (2000).
243 Klingenberg, J. M. et al. Engineered human skin substitutes undergo large-scale genomic
reprogramming and normal skin-like maturation after transplantation to athymic mice.
Journal of Investigative Dermatology 130, 587-601, doi:10.1038/jid.2009.295 (2010).
244 Boyce, S. T. et al. Cultured skin substitutes reduce requirements for harvesting of skin
autograft for closure of excised, full-thickness burns. Journal of Trauma-Injury Infection
and Critical Care 60, 821-829, doi:10.1097/01.ta.0000196802.91829.cc (2006).
245 Kempf, M. et al. A denatured collagen microfiber scaffold seeded with human fibroblasts
and keratinocytes for skin grafting. Biomaterials 32, 4782-4792,
doi:10.1016/j.biomaterials.2011.03.023 (2011).
246 Boyce, S. T. et al. Comparative-assessment of cultured skin substitutes and native skin
autograft for treatment of full-thickness burns. Annals of Surgery 222, 743-752,
doi:10.1097/00000658-199512000-00008 (1995).
247 Wright, K. A. et al. Alternative delivery of keratinocytes using a polyurethane membrane
and the implications for its use in the treatment of full-thickness burn injury. Burns 24, 7-
17, doi:10.1016/s0305-4179(97)00075-2 (1998).
248 Atiyeh, B. S. & Costagliola, M. Cultured epithelial autograft (CEA) in burn treatment:
Three decades later. Burns 33, 405-413, doi:10.1016/j.burns.2006.11.002 (2007).
249 Wood, F. M., Kolybaba, M. L. & Allen, P. The use of cultured epithelial autograft in the
treatment of major burn injuries: A critical review of the literature. Burns 32, 395-401,
doi:10.1016/j.burns.2006.01.008 (2006).
250 Lumenta, D. B., Kamolz, L.-P. & Frey, M. Adult Burn Patients With More Than 60%
TBSA Involved-Meek and Other Techniques to Overcome Restricted Skin Harvest
Availability-The Viennese Concept. Journal of Burn Care & Research 30, 231-242,
doi:10.1097/BCR.0b013e318198a2d6 (2009).
251 Yannas, I. V., Burke, J. F., Orgill, D. P. & Skrabut, E. M. Wound tissue can utilize a
polymeric template to synthesize a functional extension of skin. Science 215, 174-176,
doi:10.1126/science.7031899 (1982).
252 Navarro, F. A. et al. Sprayed keratinocyte suspensions accelerate epidermal coverage in a
porcine microwound model. Journal of Burn Care & Rehabilitation 21, 513-518,
doi:10.1067/mbc.2000.110291 (2000).
105
253 Baranski, J. D. et al. Geometric control of vascular networks to enhance engineered
tissue integration and function. Proceedings of the National Academy of Sciences of the
United States of America 110, 7586-7591, doi:10.1073/pnas.1217796110 (2013).
254 Onoe, H. et al. Metre-long cell-laden microfibres exhibit tissue morphologies and
functions. Nature Materials 12, 584-590, doi:10.1038/nmat3606 (2013).
255 Boyce, S. T. et al. Skin wound closure in athymic mice with cultured human-cells,
biopolymers, and growth-factors. Surgery 110, 866-876 (1991).
256 Atac, B. et al. Skin and hair on-a-chip: in vitro skin models versus ex vivo tissue
maintenance with dynamic perfusion. Lab on a Chip 13, 3555-3561,
doi:10.1039/c3lc50227a (2013).
257 Morimoto, Y., Tanaka, R. & Takeuchi, S. Construction of 3D, layered skin, microsized
tissues by using cell beads for cellular function analysis. Advanced Healthcare Materials
2, 261-265, doi:10.1002/adhm.201200189 (2013).
258 Pereira Gomes, J. A. et al. Corneal Reconstruction with Tissue-Engineered Cell Sheets
Composed of Human Immature Dental Pulp Stem Cells. Investigative Ophthalmology &
Visual Science 51, 1408-1414, doi:10.1167/iovs.09-4029 (2010).
259 Galassi, G. et al. In vitro reconstructed dermis implanted in human wounds: degradation
studies of the HA-based supporting scaffold. Biomaterials 21, 2183-2191,
doi:10.1016/s0142-9612(00)00147-2 (2000).
260 Zacchi, V. et al. In vitro engineering of human skin-like tissue. Journal of Biomedical
Materials Research 40, 187-194, doi:10.1002/(sici)1097-4636(199805)40:2<187::aid-
jbm3>3.0.co;2-h (1998).
261 Cicchi, R. et al. From molecular structure to tissue architecture: collagen organization
probed by SHG microscopy. Journal of Biophotonics 6, 129-142,
doi:10.1002/jbio.201200092 (2013).
262 Chan, E.-S. et al. Effect of formulation of alginate beads on their mechanical behavior
and stiffness. Particuology 9, 228-234, doi:10.1016/j.partic.2010.12.002 (2011).
263 Dewey, C. F., Bussolari, S. R., Gimbrone, M. A. & Davies, P. F. The dynamic-response
of vascular endothelial-cells to fluid shear-stress. Journal of Biomechanical Engineering-
Transactions of the Asme 103, 177-185 (1981).
264 Young, E. W. K., Wheeler, A. R. & Simmons, C. A. Matrix-dependent adhesion of
vascular and valvular endothelial cells in microfluidic channels. Lab on a Chip 7, 1759-
1766, doi:10.1039/b712486d (2007).
265 Hendriks, F. M., Brokken, D., Oomens, C. W. J., Bader, D. L. & Baaijens, F. P. T. The
relative contributions of different skin layers to the mechanical behavior of human skin in
vivo using suction experiments. Medical Engineering & Physics 28, 259-266,
doi:10.1016/j.medengphy.2005.07.001 (2006).
106
107
Appendix
Microfluidic Device Fabrication
The multi-layered microfluidic devices were produced from vertically stacked and bonded
PDMS layers that were individually obtained by moulding from different masters. Figure A1
represents rendered views of the microfluidic device designs presented.
Masters with 150µm tall features were defined by spincoating negative photoresist SU8-2050
(MicroChem Corp, Newton, MA, USA) onto clean glass substrate. The final feature height was
achieved by two spincoating steps at 1900rpm (30s with a 5s linear ramp to 1900rpm), producing
a 75µm thick resist layer in each step. This two-step procedure ensured thickness uniformity
across the entire master. After the first spincoating step, the substrate was postbaked for 6min at
a temperature of 65˚C, followed by 15min at 95˚C. Following the second spincoating step, the
substrate was baked for 10min at 65˚C and 35 min at 95˚C. Features with minimum width of
230 µm at the device exit section were patterned by soft lithography with 24mW/cm2 UV
intensity and 9s exposure time (total energy of 220 mJ). The exposed substrate was baked for
30 s at 65˚C and 20 min at 95˚C, left to cool to room temperature, and developed under constant
shaking for 12 min with SU8 developer (MicroChem Corp).
Individual layers of the microfluidic device were defined by spincoating178
poly(dimethylsiloxane) (1:10 ratio of curing agent to monomer) (PDMS, Sylgard 184 Silicone
Elastomer Kit, Dow Corning, Midland, MI, USA). Spincoating PDMS at 450 rpm for 30 s
resulted in layers with uniform thickness of 400 ±7 µm. The multilayer device was obtained by
sequentially aligning and bonding individual layers that were previously partially cured for 8 min
at 80ºC, producing a final multilayer device composed of up to 10 layers. On-chip reservoirs
were obtained from 3ml BD syringe barrels cut in half, resulting in a total fluid storage volume
of 1.5ml. The section of the barrel containing the female luer lock connector was used for easy
connection to the computer-controlled solenoid valves using male luer lock connectors
(Upchurch Scientific, Oak Harbor, Washington, USA). These reservoirs were implemented onto
the microfluidic device by first fixing with epoxy and subsequently pouring a 1cm thick uncured
PDMS layer over the final device, preventing the reservoirs from delaminating. The completed
multilayer microfluidic device was further cured for 8 hrs at 80ºC. Devices consistently
withstood inlet pressures up to 600 kPa without any delamination.
108
109
Figure A1. Rendered device designs. (a) Designs for the formation of collagen sheets. Focusing
solution layer (left) and collagen layer (right). (b) Designs for the formation of patterned single
layered sheet with the ability to incorporate up to seven distinct payloads or material solutions.
(c) Designs for the formation of bilayered skin grafts 15mm wide (left) and single layered skin
grafts 60mm wide (right). On the left, the layer numbers correspond to: (1, 4) focusing solution,
(2) bottom matrix layer, (3), top matrix layer, (5) distribution of botton cell patterning solution,
(6) distribution of top cell patterning solution. On the right, the layer numbers correspond to: (1,
3) focusing solution, (2) matrix layer, (4) distribution of cell patterning solution. Scale bars 5mm
(a), 10mm (b), 20mm (c).
CHAPTER 2 – Appendix
Tissues Fiber Molecule Fibril Diameter
[nm] Fiber Diameter [µm]
UTS
[MPa]
E
[MPa]
Palm Tree73,261 -- -- 1.1-19 89-222 440-1090
Spruce Wood1,73 Cellulose -- 34.5 ± 5.2 -- 14000 ± 2000
Rat Tail Skin 86 Collagen -- -- 1-20 --
Human
Cornea 92-95 Collagen 25-35 (stroma) -- 3.3 ± 0.2 15.9 ± 2.0
Rat Tail
Tendon 1,86 Collagen 50-500 -- 100 1000-2000
Native Porcine
Arteries 134,135 Collagen 16-500 -- 6.58 ± 0.97 45.1 ± 16.8
Table A1. Natural tissues and their mechanical properties.
110
Collagen Formation Degree of
Alignment
Dimensions of
Collagen Construct
[W]×[L] [mm]
Collagen Fibril
Diameter [nm]
UTS
[MPa]
E
[MPa]
Ionic Strength, pH,
Temperature 104 Random 0.01 0.1
Microfluidic – Flow 119 20-40% aligned
Flow/Extrusion 112 aligned 53 ± 14 × 21 ± 3 94 ± 19 775 ± 173
Flow/Extrusion 116 aligned 1 × 1.2 × 30 45 - 3.6 ± 1.9
Electrochemical 120 aligned 0.050-0.4 × 30-70 24-88 277-671
Electrospun and
embedded 127
71.27 ± 12.81%
aligned 3 × 20 272 ± 183 2.69 ± 0.47 0.044 ± 0.0042
Table A2. Mechanical and structural properties of natural and synthetic collagen gels created
using a variety of strategies.
Figure A2. Characterization of velocity profiles within the constriction. Velocity streamlines
of focusing solution loaded with 0.08% v/v fluorescent microbeads were imaged using long-
exposure fluorescence microscopy. (a) Entrance region to the constriction. QM = 100µl/min,
QF = 1ml/min, V* = 10. (b) No collagen flow and no pulling. QF = 3ml/min (1), 6ml/min (2).
Scale bars 250µm.
111
Numerical simulation of the velocity profile within the constriction.
A multiphysics solver based on the finite element method (Comsol v3.5 fluid dynamics module)
was used to model the velocity profile of the focusing solution within the constriction, under
steady flow conditions. The system was treated as a two-dimensional problem, with the focusing
solution consisting of 10% w/v PEG (dynamic viscosity = 54.4cP) treated as a Newtonian fluid.
The collagen sheet was assumed a gelled sheet and modelled as a moving wall with a velocity
equal to the pulling velocity VP = 2mm/s. The focusing solution flow rate QF was varied from 1-
16ml/min and input as a constant mean velocity, which was calculated by dividing the flow rate
by the cross-sectional area of the microfluidic device (10mm [W] × 0.12mm [H]). The outlet was
set at 0Pa gauge pressure. The parameters used in the numerical study are listed in table A3, and
a schematic of the model and the resulting velocity profiles are summarized in figure A3.
Focusing Flow Rate QF [mL/min] 1-16
Pulling Velocity VP [mm/s] 2
Dynamic Viscosity Focusing Solution [Pa•s] 0.0544
Density Focusing Solution [mg/mL] 100
½ Constriction Height [mm] 0.5
Table A3. Parameters used in numerical model, treating the focusing solution as a Newtonian
fluid.
112
Figure A3. Numerical simulation of the flow behavior of the focusing solution within ½ of
the constriction (½ Hc). The cross-linked sheet is considered a moving wall with velocity equal
to the pulling velocity VP = 2mm/s. (a) Schematic of the region investigated. (b) Velocity profiles
of the focusing solution within half the constriction height HC, with QF = 1, 3, 6, 9, 12ml/min.
Focusing solution viscosity = 54.4cP.
Figure A4. Characterization of collagen sheet width and thickness produced at varying V*
and QF, with constant QM. QM = 100µl/min, QF = 1ml/min (red), QF = 3ml/min (blue), QF =
6ml/min (green). Data obtained for 10mm wide device with flow-focusing manifold, HC= 1mm.
113
Figure A5. Mechanical properties of collagen sheets following various treatments: with/without
48hr incubation in fibril incubation buffer (FIB) post-extrusion, and with/without air-drying step
post-extrusion. Young’s modulus and upper tensile strength are reported for all conditions.
Figure A6. (a) TEM images of collagen sheets produced at V* = 0.6 (1) and 4.5 (2). (b) SEM
images of collagen sheets produced at V* = 0.6 (1) and 4.5 (2). Magnification: 20× (left), 100×
(right). Scale bars 200nm (a), 1µm (b, left), 500nm (b, right).
114
Figure A7. Characterization of fibril spacing and compaction by autocorrelation of TEM
(a) and SEM (b) images. Experimental conditions are QM = 100µl/min, QF = 1ml/min, V* = 0.1,
0.6, 7, 10 (1 to 4 respectively). (c) SEM image of collagen sheet produced with QM = 100µl/min,
QF = 1ml/min, and V* = 7, showing the repeated banding pattern (D-period). Scale bar 500nm.
115
Figure A8. Photograph of microfluidic device with constriction unit. Scale bar 10mm.
116
Design Drawings of the Constriction Manifold
Figure A9. Constriction manifold assembly.
Figure A10. Constriction manifold – Top piece.
117
Figure A11. Constriction manifold – Bottom piece.
Design Drawings of the XZ Constriction Manifold
Figure A12. XZ constriction manifold assembly.
118
Figure A13. XZ constriction manifold - Frame.
Figure A14. XZ constriction manifold – Left constriction bracket.
119
Figure A15. XZ constriction manifold – Right constriction bracket.
120
CHAPTER 3 – Appendix
Referenc
e
Alginate
Concentrati
on
[% w.t.]
Alginate
M:G
Ratio
Cross-
Linker
Cross-Linker
Concentration
[mM]
Cross-
Linking
Time
[min]
Elastic
Modulus
[kPa]
UTS
[kPa]
Drury et
al.157
2 0.29 CaSO4 1540 20 52 ± 12 32 ± 3
2 0.85 CaSO4 1540 20 14 ± 5 5 ± 2
Chan et
al.262
2 0.59 CaCl2 135.2 30 300 ± 45 --
2 1.56 CaCl2 135.2 30 250 ± 10 --
4.5 0.59 CaCl2 135.2 30 570 ± 15 --
4.5 1.56 CaCl2 135.2 30 400 ± 30 --
1.5 0.59 BaCl2 50 30 592 --
1.5 0.59 CuCl2 50 30 802 --
Stevens et
al. 165
2 0.33-0.54 CaCl2 75 -- 172 ± 7 --
2 0.33-0.54 CaCl2 300 -- 155 ± 3 --
4 0.33-0.54 CaCl2 75 -- 489 ± 21 --
4 0.33-0.54 CaCl2 300 -- 471 ± 19 --
Table A4 Mechanical properties of alginate gels.
121
Control of Soft Material Thickness by Fluid Focusing
Planar soft material samples with well-defined thicknesses were also produced by employing the
outlined strategy while relying exclusively on the shear imposed by the focusing fluid, i.e.,
without additional pulling via the collection drum. Thickness of samples produced at QB =
120µl/min and QF = 2-10ml/min was characterized. Thicknesses ranging from 170-700µm were
obtained and measured by optical microscopy of the cross section.
Figure A16. Control over soft material thickness as a function of the flow rate of the focusing
stream QF, and for QB = 120µl/min
122
Dynamics of Valve Actuation for Soft Material Coding
Characterization of the dynamic behaviour of the computer-actuated solenoid valves (model
LHLA0521111H, The Lee Company, Westbrook, CT, USA) was achieved using piezoresistive
pressure transducers (pressure range: 0-30psi, time resolution: 1ms, model HSCDIP030PGAA5,
Honeywell, Morristown, New Jersey, USA). On-chip measurements obtained in the reservoirs
during valve actuation in terms of a voltage were converted to a pressure reading using a
calibration curve. Two actuation cycles were considered: 0.15 s open and 2 s closed, and 0.25 s
open and 2 s closed. The measured pressures and valve actuation times were found to be in good
agreement with the programmed input parameters.
Figure A17. Characterization of pressure in on-chip reservoirs. Valve actuation pattern: (left)
0.15 ms open – 2 s close, (right) 0.25 ms open – 2s close. Input pressure 7kPa. Inserts represent
magnified view of pressure evolution during valve actuation.
123
Shear Stress During Cell Patterning
Shear stresses during the flow of cell suspension into the microfluidic device were calculated to
ensure that the shear stress experienced by the cells did not exceed physiological levels. Given
the employed microfluidic channel geometry and experimental conditions, the inlet pressures
(wells) of 2-4 kPa and a viscosity of the (uncrosslinked) biopolymers of approximately 0.05 Pas,
the shear stress is linearly distributed between the location of the channel center (zero) and its
maximum value of 13-26 dyne/cm2 at the wall (Poiseuille flow). The cell suspended within the
biopolymer are therefore subjected to shear stresses less than 2 dyne/cm2, a level well within
physiological conditions. Endothelial cells, e.g., experience 15-20 dyne/cm2 in undisturbed
regions of the vascular system, can be transiently exposed to 40-50 dyne/cm2 in areas of
disturbed flow 197,263 and exhibit reduced adhesion above 100 dyne/cm2 264. The calcium
chloride concentrations of 50-100mM that we used for cross-linking of hydrogel sheets are
consistent with conditions previously employed for cell encapsulation and are not detrimental to
cells 171,177,185.
Figure A18. Shear stress profile within a microfluidic channel.
124
Figure A19. Viability and distribution of printed cells. (a) Statistics of cell distribution within
a single pattern (n=5). (b) Confocal fluorescence image of fibroblasts incorporated in 2%w.t.
alginate (Day 5). (c) Assessment of patterned cell survival (n=5). Neonatal rat fibroblast (left)
and cardiomyocyte (right). Scale bars 200µm (a), 50µm (b).
125
Figure A20. Modulus of elasticity for a homogeneous soft material composed of 2%w.t.
alginate produced in the free-extrusion and pulled-extrusion modes.
126
Figure A21. Line camera intensity measurements of the UN Charter, Chapter 1, Article1, “The
purposes of the United Nations”.
127
CHAPTER 4 – Appendix
Method Test Region E [kPa]
Indentation [217]
Male thigh 1.99 ± 0.43
Male forearm 1.51 ± 0.32
Female forearm 1.09 ± 0.54
Suction [219,265] Forearm 56-217
Tensile [220]
Leg (parallel to muscle) 20000
Leg (perpendicular to muscle) 4600
Torsion [222, 221]
Dorsal forearm (<30yr old) 420
Dorsal forearm (>30yr old) 850
Ventral forearm 1120
Table A5. Literature data of elastic moduli of human skin measured in vivo using various
strategies.
128
Figure A22. Full-thickness burns and current treatments. (a) Pathophysiology of skin and
comparison to full-thickness burns, (b) Clinical approaches for treatment of full-thickness burns.
Figure A23. Characterization of spot volume as a function of valve actuation.
129
Figure A24. Degradation studies of skin grafts. (a) Grafts composed of base material solutions
consisting of 1%w.t. sodium alginate (Molecule-R) (circle) and 4%w.t. sodium alginate (Sigma)
(triangle). (b) Grafts composed of cell printing solutions consisting of 0.05%w.t. sodium alginate
(Molecule-R) (circle), and 0.2%w.t. high G:M alginate with 0.23mg/mL Hyaluronic acid
(triangle). Samples were incubated at 37 ̊C in DMEM (full symbols) and Epilife (hollow
symbols), and their dry weight measured over a period of 11 days.
130
Figure A25. Mechanical properties of skin grafts as a function of culture time. Elastic
modulus and UTS of: 1%w.t. sodium alginate (Molecule-R) and 4%w.t. sodium alginate (Sigma)
as base materials. Samples were cultured in DMEM (*) and Epilife (**) over four days.
131
Sample Stress-Strain Curves of Printed Skin Grafts
Figure A26. Stress-Strain curves of skin graft materials as a function of culture time. (a)
1%w.t. sodium alginate (Molecule-R) and (b) 4%w.t. sodium alginate (Sigma). Samples were
cultured in DMEM and Epilife over four days.
132
Figure A27. Stress-Strain curves of skin graft printed with parallel spots. Tensile tests
conducted in both x and y direction on day 0 and day 1 of culture in DMEM.
133
Figure A28. Stress-Strain curves of skin graft printed with parallel stripes. Tensile tests
conducted in both x and y direction on day 0 and day 1 of culture in DMEM.
134
Figure A29. Stress-Strain curves of skin graft printed with alternating voids. Tensile tests
conducted in both x and y direction on day 0 and day 1 of culture in DMEM.
135
Figure A30. Stress-Strain curves of skin graft printed with alternating spots. Tensile tests
conducted in both x and y direction on day 0 and day 1 of culture in DMEM.