Elongated Fascicle-Inspired 3D Tissues Consisting of
High-Density, Aligned, Optogenetically Excitable
Muscle Tissue Using Sacrificial Outer Molding
by
Devin Michael Neal
S.M., Mechanical Engineering
Massachusetts Institute of Technology, 2009
M
S.B., Mechanical Engineering
Massachusetts Institute of Technology, 2007
'MSACHUSETTS INSTRrtTE,
OF TECFNIOL0OCUY
AUG 1 5 2014
LIBRARIES
Submitted to the Department of Mechanical Engineering
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2014
0 Massachusetts Institute of Technology, 2014. All Rights Reserved.
Signature redacted
.... .
.........
Department of Mechanical Engineering
______
Certified by......................................................S
a
4January
z
r
23, 2014
acted..
.
Author.....................................................................................
V
Harry Asada
F *Wrofessor of Mechanical Engineering
Supervisor
-
Accepted by......................................................
Signature redacted
David Hardt
-man,Department Committee on Graduate Students
1
2
Elongated Fascicle-Inspired 3D Tissues Consisting of High-Density, Aligned,
Optogenetically Excitable Muscle Tissue Using Sacrificial Outer Molding
by
Devin Michael Neal
Submitted to the Department of Mechanical Engineering
on January 23, 2014, in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy in Mechanical Engineering
ABSTRACT
The majority of muscles, nerves, and tendons are composed of fiber-like fascicle
morphology. Each fascicle has a) elongated cells highly aligned with the length of the construct,
b) a high volumetric cell density, and c) a high length-to-width ratio with a diameter small
enough to facilitate perfusion. Fiber-like fascicles are important building blocks for forming
those tissues of various sizes and cross-sectional shapes, yet no effective technology is currently
available for producing long and thin fascicle-like constructs with aligned, high-density cells.
Here we present a method for molding cell-laden hydrogels that generate cylindrical tissue
structures that are 100 pim in diameter with an extremely high length to diameter ratio (>100:1).
Using this method we have successfully created skeletal muscle tissue with a high volumetric
density (-50%) and perfect cell alignment along the axis. A new molding technique, Sacrificial
Outer Molding, allows us to i) create long and thin cavities of desired shape in a mold that is
solid at a low temperature, ii) release gelling agents from the sacrificial mold material after cellladen hydrogel is injected into fiber cavities, iii) generate a uniform axial tension between anchor
points at both ends that promotes cell alignment and maturation, and iv) perfuse the tissue
effectively by exposing it to media after melting the sacrificial outer mold at 37'C. Effects of key
parameters and conditions, including initial cavity diameter, axial tension, and concentrations of
hydrogel and gelling agent, upon tissue compaction, volumetric cell density, and cell alignment
are presented. Furthermore, the tissue is characterized in a custom designed mechanical
characterization system. Characterization has shown that an optimal diameter exists at which
muscle constructs exhibit the greatest contraction performance, and that optical and electrical
stimulation of optogenetic muscle cells result in similar performance if the tissue is developed
sufficiently.
Thesis Supervisor: H. Harry Asada
Title: Ford Professor of Mechanical Engineering
3
To Mom
4
ACKNOWLEDGMENTS
I want to start by thanking the members of my thesis committee: Professor Harry Asada,
Professor Roger Kamm, Professor Krystyn Van Vliet, and Professor Rashid Bashir. Your
comments, encouragement both in committee meetings and out have helped make this document
something I'm truly proud of. I'd like to thank my advisor in particular. He has shown me time
and again that his true product, his true pride is his students.
Thank you to the members of the d'Arbeloff lab. Every one of you was always willing to
help me with my research. I'd like to thank a few current members in particular for their
substantial contributions to the work presented in this thesis: Dr. Min-Cheol Kim, Dr. Sharon
Ong, and Dr. Vincent Chan. I would like to thank a former member in particular: Dr. M. Selman
Sakar. Selman was a mentor in every sense of the word. Thank you to members of the Kamm
Group. The Kammsters were always willing to help and share facilities and knowledge. With
them, I was privileged to be welcome in a lab away from lab. I'd like to thank two current
members for their contributions to the work presented here: Sebastien Uzel, and Jordan Whistler.
I want to thank Dr. Frances Hill for going through this process before me and being an
invaluable resource in managing the PhD process.
I want to thank the NSF and EBICS, and the NRF of Singapore and BioSyM. I am
grateful for these programs as funding sources as well as for all the people they allowed me to
work and collaborate with.
5
TABLE OF CONTENTS
LIST OF FIGURES ........................................................................................................................
Chapter 1: Introduction.................................................................................................................
Intro 1.1 Significance of engineered skeletal muscle tissue......................................................
Intro 1.2 Literature review of engineered skeletal muscle tissue...........................................
Intro 1.3 Thesis Framing .......................................................................................................
Chapter 2: Fascicle-Inspired 3D Tissues Using Sacrificial Outer Molding ..............................
2.1 Introduction .........................................................................................................................
2.2 Results and Discussion.....................................................................................................
2.2.1 Sacrificial External M olding ....................................................................................
9
11
2.2.2 Fascicle-like tissue formation..................................................................................
22
11
13
16
18
18
19
19
2.2.3 Importance of axial stress.........................................................................................
23
2.2.4 Construct diam eter...................................................................................................
23
2.2.5 Cell density...................................................................................................................
24
2.2.6 Fibrin com ponent variation ......................................................................................
25
2.3 M aterials and Methods.....................................................................................................
25
2.3.1 Fabrication of Permanent M olds ...............................................................................
26
2.3.2 Culture of C2C12 M yoblasts....................................................................................
26
2.3.3 Casting of Sacrificial Mold and Hydrogel Constructs .............................................
26
2.3.4 Immunostaim ng.....................................................................................................
27
2.3.5 Automated Image Analysis ......................................................................................
27
2.4 Conclusions .........................................................................................................................
27
2.5 Acknowledgem ents..........................................................................................................
28
2.6 Supporting Inform ation...................................................................................................
28
2.7 Figures and Captions........................................................................................................
31
Chapter 3: Mechanical Characterization of Optically and Electrically Stimulated Fascicle-Like
Constructs .....................................................................................................................................
39
3.1 Introduction ...................................................
39
3.2 M echanical characterization system design ...................................................................
41
3.2.1 Concept: lateral displacem ent with a cantilever.....................................................
41
3.2.2 Characterization System ............................................................................................
42
3.2.3 Electrode design .......................................................................................................
43
3.3 Materials and methods .....................................................................................................
44
6
3.3.1 Tissue Constructs......................................................................................................
44
3.3.2 Mechanical Characterization....................................................................................
44
3.3.3 Im age and Video A nalysis.......................................................................................
45
3.4 Results and discussion.....................................................................................................
46
3.4.1 Optical vs. Electrical Stimulation.............................................................................
46
3.4.2 Diam eter variation.....................................................................................................
47
3.4.3 Probe stiffness variation ...........................................................................................
48
3.5 Conclusion...........................................................................................................................
49
3.6 Figures and Captions........................................................................................................
52
Chapter 4: Design of Multi-Degree of Freedom Skeletal Muscle Powered Systems........ 61
4.1 Introduction .........................................................................................................................
61
4.2 Bio-A ctuator Building Blocks.........................................................................................
63
4.3 Scaling from Tw o Unit to multi-Unit System s ...................................................................
64
4.3.1 Massively Parallel System s ......................................................................................
65
4.3.2 Highly Serial/Netw orked System .............................................................................
65
4.4 Experim ental Materials and M ethods ............................................................................
66
4.5 Experim ental Results........................................................................................................
67
4.5.1 Single Building-Block Muscle Actuator ...................................................................
67
4.5.1 Serial and Parallel Prototypes..................................................................................
67
4.5.2 Multi-U nit A ctuator System Prototypes.......................................................................
68
4.6 Conclusion...........................................................................................................................
68
4.7 Figures and Captions........................................................................................................
69
CHAPTER 5: CONCLUSION AND FUTURE DIRECTIONS ...............................................
76
5.1 Conclusion...........................................................................................................................
76
5.2 Future directions..................................................................................................................
76
5.3 A cknowledgem ents ..........................................................................................................
77
5.4 Figures and captions........................................................................................................
78
References.....................................................................................................................................
80
A ppendix A : Fascicle construct protocol..................................................................................
89
Appendix B: Optogenetic Control of Live Skeletal Muscles: Non-Invasive, Wireless, and Precise
A ctivation of M uscle Tissues...................................................................................................
97
B .1 Introduciton ........................................................................................................................
97
B.2 Myogenesis and Optogenetics.........................................................................................
99
B .3 Experim ental Evaluation ..................................................................................................
100
B Applications......................................................................................................................
102
7
B.4.1 Muscle-on-a-Chip Drug Screenings ..........................................................................
B.4.2 M any DO F Robotic Devices......................................................................................
B.5 Conclusion........................................................................................................................
B.6 Acknow ledgm ents ............................................................................................................
B.7 Figures and Captions ........................................................................................................
Appendex C: Patent Description of Patent Application PCT/US2012/027483..........................
102
103
103
104
105
111
8
LIST OF FIGURES
Fig. 2.11 Fascicle-like tissue construct production technique ..................................................
31
Fig. 2.2 1Fascicle-inspired tissue constructs.............................................................................
32
Fig. 2.3 1Geometric changes in fascicle construct structure over time. ..................................
33
Fig. 2.4 1Variable parameters may be used to control the development of fascicle-like constructs.
.......................................................................................................................................................
34
Fig. 2.S1 I Demonstration of how the presented method may be expanded.................. 35
Fig. 2.S2 ILow gelatin concentration.....................................................................................
35
Fig. 2.S3 ISuitable gelatin concentrations...............................................................................
36
Fig. 2.S4 IEffects of gelatin viscosity ......................................................................................
37
Fig. 2.S5 INonuniformities from mixing thrombin and fibrinogen ........................................
38
Fig. 3.1 1Force probe diagram ...................................................................................................
52
Fig. 3.2 1Mechanical characterization system..........................................................................
53
Fig. 3.3 1Electrode design..........................................................................................................
54
Fig. 3.4 | Compliant electrodes functioning.............................................................................
55
Fig. 3.5 1Stimulation protocol ..................................................................................................
55
Fig. 3.61 Twitch data analysis...................................................................................................
56
Fig. 3.7 1Electric vs. optical performance results.....................................................................
57
Fig. 3.8 1Stress and force over various initial pin diameters.....................................................
58
Fig. 3.91 Uniform vs. nonuniform strip force..........................................................................
59
Fig. 3.10 1Tissue performance under varying load conditions................................................
60
Fig. 4.11 Optogenetic control of skeletal muscles...................................................................
69
Fig. 4.2 1Skeletal muscle structure ...........................................................................................
69
Fig. 4.3 1Fascicle-inspired building block muscle strip actuators...........................................
70
Fig. 4.41 The combination of parallel and serial connections of muscle strips........................ 71
Fig. 4.5 1Numerous parallel muscle strips on both sides of a floating node ............................
72
Fig. 4.61 Increasing serial connections....................................................................................
73
Fig. 4.7 | Individual building block actuator contracting with 4.3% measured strain .............. 74
Fig. 4.8 1Prototype muscle actuator systems consisting of fascicle-like muscle actuators.......... 74
Fig. 4.9 | Prototype muscle parallel and series actuator system................................................
75
Fig. 5.1 1Future Directions .......................................................................................................
78
9
Fig. 5.2 Initial PEG Fascicle-Like Strips ...............................................................................
79
Fig. A.1 I Schematic of fascicle tissue device with pins.........................................................
92
Fig. A.2 Schematic of fascicle tissue device with pins removed...........................................
93
Fig. B. 1 IOptogenetic control of live cells and tissue. ...............................................................
105
Fig. B.2 Membrane-bound Channelrhodopsin 2 serves as an ion gate......................................
105
Fig. B.3 I Plasmid containing ChR2............................................................................................
105
Fig. B.4 Mouse skeletal cells (C2C12) transfected with ChR2. ................................................
106
Fig. B.5 j Myogenic process........................................................................................................
106
Fig. B.6 Experiment of optogenetic contraction control of skeletal muscles.............................
106
Fig. B.7 Experiment of spatial resolution of optogenetic control..............................................
107
Fig. B.8 I Microfabricated tissue gauge for formation of 3D skeletal muscle microtissue......... 107
Fig. B.9 Cross-striations characteristic to sarcomere formation. Stained for a-actinin............ 107
Fig. B. 10 I Impulse response test of 3D skeletal muscle.............................................................
108
Fig. B.11I 2nd order ARMAX model siumulation with measured contraction data. ................ 108
Fig. B.12 Tetanus-like response from simulation.....................................................................
109
Fig. B.13 IMuscle-on-a-Chip Drug Screening...........................................................................
109
Fig. B.14 Highly Networked Mouth-Like Muscle System.......................................................
109
TABLE B.1 I Individual and group activations ..........................................................................
110
10
CHAPTER 1: INTRODUCTION
INTRO 1.1 SIGNIFICANCE OF ENGINEERED SKELETAL MUSCLE TISSUE
Muscle tissue is the material that moves us. Contractile tissue is what makes animals
move. Muscle generated motion is what makes animals animals (except for sea sponges). This
is fundamentally, where interest in engineering muscle tissue comes from; the desire to mimic or
repair the actuators of animals.
Engineered muscle tissue can and will be used for many
practical purposes with time frames ranging from immediately implementable and useful, to
developing and becoming practical over the next few decades. The uses include studying the
biology of muscles in vitro, testing the efficacy of pharmaceuticals, implantation in patients, and
implementation as engineering actuators.
Muscle biology is currently being studied using in vivo, ex vivo, and in vitro, muscle
tissue assays to study multiple phenomenon such as how satellite cells and myoblasts
differentiate into nonproliferation myotubes, the roles of muscle-specific genes and proteins such
as MyoD, Myf5, myogenin, myosin heavy chain (MHC), and muscle creatine kinase (MCK)
[Kuang et al., 2008]. The more developed the tissue used in these experiments, the greater
applicability the information gathered has to naturally formed muscle systems.
Drug screening is another practical use of engineered muscle tissue and is already
producing results. A company called Myomics is already using 3D functional muscle tissue to
test therapeutics for muscle diseases such as Duchenne Muscular Dystrophy (DMD). Previously,
2D cell culture assays have been used to identify potential new drugs. However they typically
analyzed specific proteins or pathways (such as the myosin heavy chain gene [Cross-Doersen et
al., 2003] and utrophin protein expression [Courdir-Fruh et al., 2002]), or the growth in size of
myofibers (called hypertrophy) [Semsarian et al., 1999], rather than the actual functional
performance of the muscle tissue. A common model for DMD is the mdx murine model which is
an in vivo model producing functionally developed muscle tissue [Granchelli et al., 2000]. In
vitro screening is expensive and low throughput, while the 2D assays only provide indirect
performance assessment. For example, hypertrophy has been shown to poorly relate to muscle
strength [Hoffman et al., 2006].
For these reasons, 3D tissue with measureable contractile
performance and which can be produced in a high throughput manner is preferable for drug
11
screening [Vandenburgh et al., 2010]. As engineered muscle tissue develops to be closer to
physiological tissue, its usefulness as a drug-screening platform increases.
Engineered tissue implantation is the beautiful golden story motivates grants and
continues to show promise but has yet to be practically implemented for any complex, noncosmetic tissue. Medical implant applications are arguably the strongest argument for continued
research in tissue engineering, but we are still roughly a decade away from seeing practical
therapies commonly implemented in hospitals. Specifically for muscle, there is the potential to
replace or augment muscle tissue weakened or damaged by disease (such as DMD), age (such as
the elderly), or physical damage (such as combat veterans). Promising attempts have been made
to implant undifferentiated myoblasts along with scaffolding material [Beier et al., 2006],
however, full muscle tissue has yet to be practically implanted. Of significant interest is the
development of engineered muscle tissue containing viable neurons (along with neuromuscular
junctions (NMJs)) and vascularization with endothelial cells [Koning et al., 2009].
These
auxiliary cell types are required if fully developed muscle tissue is to be used for implantation.
Therefore, engineered muscle tissue should be developed in a manner that will facilitate scaling
up in size using multiple constructs and the integration of these auxiliary cell types.
Turning the focus away from medical applications, fully developed muscle tissue could
also be used as a practical engineering actuator. Skeletal muscle is capable of producing 350 kPa
of active stress and -20% strain. The energy per cycle it can therefore transfer is on the order of
1 kJ/m3 [Hunter et al., 1992]. It can change stiffness by many orders of magnitude, allowing it to
be back drivable or not depending on contraction and strain, granting stiffness control if an
antagonistic arrangement is used. It is approximately 35% energy efficient. Muscle is frequently
used as a benchmark for developing translational actuators because of it proven performance in
mobile robotic systems; animals. Skeletal muscle performance allows it to be used in many
different ways by animals [Dickinson et al., 2010]. No robot exists today that can exhibit the
dexterity of a 4 month old puppy, and this is in no small part due to the mechanical features of
skeletal muscle [Hunter et al., 1992]. Engineered biobots have already been produced using 2D
substrate grown muscle tissue utilizing cantilever bending. Such biobots can walk [Xi et al.,
2005], swim [Feinberg et al., 2007], and strut [Chan et al., 2012]. Muscle cells have been used
to pump fluid, bend cantilevers, and in other mechanical systems [Tanaka et al., 2007, Kim et al.,
2007, Pilarek et al., 2011, Chan et al., 2012] . No 3D engineered muscle tissue powered biobots
12
exist yet. Biobots powered by excised muscle tissue have been developed [Herr et al., 2004].
These biobots are far from providing power similar to natural muscle tissue.
More fully
developed muscle tissue is required in order to get performance close to what we find in animals.
For all these applications (scientific study, drug screening, implantation, and use as
actuators), the more fully developed the engineered muscle tissue is, the more useful it will be.
INTRO 1.2 LITERATURE REVIEW OF ENGINEERED SKELETAL MUSCLE TISSUE
Skeletal muscle tissue has many features that all contribute effectiveness and efficiency
to transmitting mechanical force and/or energy to a load.
These features include 1) 3
dimensional structure, 2) high density of force producing units, 3) highly aligned force
transmission, and 4) a highly hierarchical structure. Muscle tissue is inherently 3D. There is no
naturally occurring skeletal muscle structure composed of muscle cells exerting force to bend a
2D substrate. 3D structure allows for a truly volumetric structure that can produce stress along
the axis of force transmission. Within the 3D structure, most of the volume is occupied by stress
producing protein structures.
This high density of stress producing material increases
mechanical energy transmission in two ways; first, a greater number of stress producing units
means more stress will be generated, and second, less non-stress producing material means less
stiff material the muscle has to work against when changing length. All of these stress producing
units are aligned along a single axis of force transmission. This alignment of force production
maximizes the mechanical energy transmission along a single axis rather than dividing the
transmission along different axes. All of these features are packed into multiple hierarchical
levels. Starting from the level of contractile protein structure is the sarcomere. These structures
repeat and are serially connected along the entire length of the muscle cell. These myofibrils are
packed tightly within each muscle cell. Around each muscle cell is a layer of extra cellular
matrix basal lamina called the endomysium. These individual muscle cells are bundled together
forming muscle fascicles and each fascicle is surrounded by another layer of'matrix called the
perimysium. The fascicle is the largest subunit of muscle. Fascicles are the structural level at
which all of the features described above exist: 1) fascicles are 3D scaling up from the largely
ID single muscle cell to many, 2) fascicles are highly dense in contractile protein structures just
as an individual muscle cell is, 3) all of the muscle cells within a fascicle are aligned along the
length of the fascicle bundle, and 4) fascicles are the largest hierarchical building block of
13
muscles tissue. By bundling together multiple fascicles, skeletal muscle structures of every kind
found in nature are formed, whereas the individual fascicles found in all muscle groups are
largely the same morphologically and functionally regardless of overall muscle shape and size
[Lieber 2010].
The field of tissue engineering of skeletal muscles over the past several decades has
resulted in various tissue constructs having many of these 4 features of muscle described above,
but none have possessed all at once.
The alignment of muscle tissue on 2D substrates has been an area of focus before 3D cell
culture techniques had developed [Curtis et al., 1997].
Topographical cues of numerous
dimensions have been used to guide cells and influence differentiation.
Muscle tissue
performance has been evaluated based on the induced bending of the 2D substrate on which
muscle cells are patterned [Alford et al., 2010, Grosberg et al., 2011].
Bulk hydrogel/cell molds have been generated to create muscle tissue that 3D. Some
construct fabrication methods do not provide alignment cues, generating 3D muscle tissue, but
without alignment and fully contractile muscle cells [Okano et al., 1997]. Tissues that do provide
alignment cues typically do so by constraining the tissue between anchoring points that the liquid
hydrogel becomes solid around. Suspending cell laden hydrogels between anchors has been
shown to produce aligning stress patterns [Legant et al., 2009]. These anchoring points have
included Velcro [Bian et al., 2009a, Bian et al., 2009b, Hinds et al., 2011], metal mesh [Shansky
et al., 1997], and PDMS posts [Vendenburgh et al., 2008, Vandenburgh 2010, Boudou et al.,
2011, Sakar et al., 2012]. In some situations, the anchoring structures have had compliance
engineered into them. This known compliance has then been used to assess muscle performance
as the tissue works against the anchoring structures [Vendenburgh et al., 2008, Vandenburgh
2010, Boudou et al., 2011, Sakar et al., 2012]. In other situations in which the muscle was
grown between rigid anchors, the tissue has been removed and assessed isometrically within a
separate testing structure [Hinds et al., 2011, Bian et al., 2012].
Another method of producing 3D muscle tissue is to first culture myoblasts on a 2D
substrate until they fuse and differentiate into myotubes, and then allow the 2D tissue to detach
from the substrate and roll up around anchors resulting in a 3D mass of tissue. The method first
reported by Strohman [Strohman et al., 1990] used numerous pins as anchors. The technique
evolved to use just 2 anchors in the form of pinned sutures to form muscle strips termed myooids
14
[Dennis and Kosnik, 2000]. This method was coupled with isometric force measurements of
contractions resulting from electrical stimulation. The method was used to characterize muscle
tissue and tissue performance from multiple myoblast sources [Dennis et al., 2001]. In this
study, they found that C2C12 cells (an immortalized cell line of adult mouse myoblasts) could
not for myooids using this method without the addition of fibroblasts, and then, only formed
myooids "-50% of the time." They also found that myooids generated from mouse soleus tissue
performed approximately 4 times better than myooids generated by C2C12 and fibroblast cells
[Dennis et al. 2001]. This method was also used to show that myooids formed from adult rat
tissue performed in a superior manner to myooids formed from neonatal rat cells [Kosnik et al.,
2001]. Plating myoblasts on a fibrin gel that rolls up with the 2D sheet of cells after myotube
differentiation was shown to increase the speed of formation of myooids [Huang et al., 2005],
and patterning of the fibrin gel increases alignment and overall performance when using a fibrin
substrate [Lam et al. 2009]. This roll up method consistently produces 3D muscle tissue with
aligned muscle cells. However, the method loses significant density when scaffolding hydrogels
are used, and the method does not lend itself well to scalability.
Of the reported methods of muscle tissue engineering, none have fully satisfied the 4 key
points described above.
The control of skeletal muscle in vitro is another important issue. In order to control the
contraction of any tissue grown in vitro, the muscle must be stimulated with a control signal.
Discounting mechanical intervention, there exist 4 reported methods of stimulating muscle
tissue: chemical, electrical, neuronal, and optical.
To stimulate chemically, high levels of
potassium in the extracellular environment can induce tetanic contractions in engineered muscle
tissue [Vandenburgh et al., 1991].
Electrical stimulation offers tighter control of timing over
chemical stimulation and can be used to excite targeted muscle cells [Nagamine et al., 2011]. A
drawback of both electrical and chemical stimulation however is the lack of tight spatial control.
Chemicals will diffuse throughout a volume of liquid and electrodes produce parasitic electric
fields that may stimulate non-targeted cells. Optical control is a recently developed method that
has tight spatial and temporal control. Optogenetics involves the transfection of cells with a gene
for channelrhodopsin 2 (ChR2), a membrane ion channel that opens when stimulated optically.
Optogenetics allows for localized stimulation merely by concentrating the light signal and has
been successfully used in neurons and muscle tissue, in vitro and in vivo [Zhang et al., 2006,
15
Fenno et al., 2011, Sakar et al., 2012]. An important aside about innervation and neuromuscular
junction formation is the effect on development and performance [Wagner et al, 2003]. Myooids
formed via the roll up method that have included spinal cord tissue with dorsal root ganglia
attached have demonstrated superior performance as well as increased mature, and decreased
immature muscle protein content [Larkin et al., 2006].
INTRO 1.3 THESIS FRAMING
The objective of this thesis is to develop a robust method to produce and characterize
functional muscle tissue exhibiting 4 fundamental features of skeletal muscle: 1) three
dimensional tissue 2) high volumetric density of contractile proteins, 3) highly aligned
contractile muscle cells, and 4) be scalable as to provide the means to use this tissue as a building
block for larger scale tissue constructs. To accomplish this objective, I have sought to recreate
the smallest hierarchical structure that possesses the 4 features described above which happens to
be the largest hierarchical fundamental building block of skeletal muscle tissue; the fascicle.
The goal of creating a physiologically relevant fascicle-like tissue grown in 3D is a
necessary step in creating larger scale, muscle tissue. This goal has never been achieved based
on the depth of my literature review. Muscle tissues have been grown primarily from muscle
cells formed on 2D substrates, or within bulk hydrogel scaffolds. The goal of the substrate
grown tissues was to utilize the early observed phenomenon that myoblasts readily fuse into
myotubes in 2D cultures. In large part, effort has moved toward developing muscle in more
physiologically relevant 3D hydrogel scaffolds. However, within typical bulk hydrogel scaffolds
contractile cells only form on the shell of the scaffold leaving considerable unused volume,
volume that should be occupied by the contractile proteins of muscle tissue. This lack of density
is typically attributed to a lack of perfusion due to the thickness of the bulk hydrogel constructs.
It is almost as if the fascicle level of hierarchical structure has been skipped in favor of jumping
to produce larger full scale tissue. However, the fascicle structure must be utilized to achieve
high contractile cell density.
Additionally, it is within this structure that auxiliary cell types such as neurons (for
innervation), and endothelial cells (for perfusion) interact with the contractile muscle cells. If
these necessary cell types are going to be eventually integrated in a physiologically relevant
manner, then the hierarchical structure in which they naturally form should be adhered to.
16
The fascicle constructs detailed in this dissertation have been engineered to have the 4
features described above. The system has been designed with scaling in mind. Fascicle
constructs such as these will provide a necessary structural level out of which full scale muscle
tissue will be made.
I have engineered a multi-stage molding technique to create the large length to diameter
aspect ratio found in fascicles. I utilized passively generated stress to align the cells as they
differentiated and developed. In producing and evaluating this tissue, I have made generalizable
methods and products for general tissue engineering, and for general mechanical characterization
of engineered tissue.
This thesis is organized as follows. This first chapter provides motivation for producing
tissue with the 4 features discussed above and describes previous work that has been done to
produce this kind of muscle tissue. Chapter 2 describes how I have made tissue possessing these
4 features that are necessary for producing efficient and scalable muscle tissue. In this chapter, a
tissue engineering platform is described that may be used more generally to produce other tissues
benefiting from 3D structure, a high length to diameter aspect ratio, aligned cells, and high
volumetric density of cells. Within this chapter, the adjustable parameters of the method are
evaluated against resulting muscle tissue size and contractile protein density.
Chapter 3
discusses the development of a novel mechanical performance evaluation system. The system
described is generally applicable to other tissues, and can evaluate passive and actuated
displacements and forces under varying load conditions. In this chapter, the muscle tissue
described in chapter 2 is evaluated against multiple performance metrics. Chapter 4 presents
plans and initial results for producing larger scale systems possessing multiple tissue constructs
presented in chapter 2. In this chapter, complex systems having multiple degrees of freedom are
explored that utilize numerous individual tissues in serial and parallel arrangements.
The
conclusions and proposed future work are described in the final chapter, chapter 5. Included in
appendices are a) a detailed protocol for producing the fascicle like tissue constructs, b)
discussion (with results) of optical control of muscle cells, and b) the description section of a
patent submitted covering the tissue engineering method.
17
CHAPTER 2: FASCICLE-INSPIRED 3D TISSUES USING SACRIFICIAL
OUTER MOLDING
2.1 INTRODUCTION
There is a class of tissues that transmit forces, displacements, and even signals over
significant distances. These tissues include muscles, tendons, and nerves. Cells within these
tissues are significantly elongated along the axis of transmission, and tightly bundled with other
cells into fascicle structures [Williams 1995, Lieber 2002]. Multiple fascicles are bundled along
with auxiliary cell types into the full tissue. Although the size and structure of the whole tissue
varies greatly depending on specific location and function, all of these consist of fascicles that
are largely the same in size and function. These tissues exhibit a hierarchical structure, where
fascicles in vivo may be viewed as the structural unit at the highest hierarchical level, the
assembly of which is the whole tissue. They act as functional building blocks to form tissue of
diverse sizes and shapes. Fascicles have some morphological traits and features common to all
of them: 1) their significant length is much larger than the width of an individual cell, 2) they are
aligned along the axis of transmission, and 3) they have high volumetric density of cells to
concentrate transmission. Furthermore, fascicles size is limited to a particular diameter so that
centrally located cells may be perfused effectively [Paul , 2001]. The long and thin, cylindrical
shape of the fascicle would facilitate to construct tissues with diverse cross-sectional shapes to
meet the needs of specific tissue groups. Rather than directly building monolithic bulk tissue,
the formation of building-block tissues inspired by fascicles that possess these features is an
important step for constructing a full-scale tissue of this class. We present a method to produce
fascicle-inspired tissues with tightly controlled parameters by generating a scaffold that is
engineered to have 1) a truly 3D structure with controlled diameter and length, and 2) an aligned
hydrogel nanostructure generated through anisotropic stress along the length of the construct.
Formation of fascicle-like tissues having all the three morphological features is a
challenge. Prior works have successfully produced tissues focusing on one or two features.
Photomasks [Tsang et al., 2007], additive rapid prototyping [Chan et al., 2010], and gas foaming
methods [Partap et al., 2006] have been used to create 3D tissue where highly aligned cell
morphology is not required [Zorlutuna et al., 2012].
Various surface techniques have been
utilized that pattern substrates with channels of diverse sizes, or chemical patterns, or even
18
optical patterns for hydrogel polymerization on a surface. Such methods are designed to achieve
highly aligned cell morphology and have successfully been employed to study fibroblast [Yang
et al., 2009] and kidney epithelial cell orientation [Schumacher et al., 2008], muscle cell
alignment [Lam et al., 2006] and axon guidance [Jang et al., 2010]. Another substrate based
technique to generate 3D tissue is to start with 2D cell sheets and either roll the cell sheets [Lam
et al., 2006, Dennis et al., 2001] or stack them [Takahashi et al., 2013]. To generate 3D, aligned
tissue scaffolds of biocompatible materials and hydrogels, electrospinning [Chew et al., 2008,
Wang et al., 2009], and microfluidic spinning [Hwang et al., 2009, Kang et al., 2012, Onoe et al.,
2013] techniques have been used. Spun scaffold materials are typically composed of fibers with
diameters larger than the cells seeded onto them. Therefore, the volumetric density of cells is
inevitably low. Another method of generating highly aligned, 3D cell morphology is to seed
cells in bulk hydrogel molds and allow the initial cell mediated compaction to align the
nanostructure of the hydrogel around controllable post/anchoring architectures [Bian and Bursac,
2009, Hinds et al., 2011, Vandenburgh et al., 2008]. This approach produces tissue constructs
where cells tend to migrate towards the peripheral surface of the tissue, leaving a low density of
cells inside. The average cell density across the full tissue cross section is therefore low for
larger diameter constructs. The micro-post technology has successfully produced aligned 3D
skeletal muscle tissues [Vandenburgh et al., 2008, Sakar et al., 2012]. While the compliance of
the posts facilitates to measure force and displacement, it allows the tissue to relax, reducing
stress in the tissue.
Despite these advances, to our knowledge, no technology is currently available that
successfully produces fascicle-like 3D constructs, grown in a 3D scaffold environment, having a
high aspect ratio on the order of 100:1, and a high volumetric density of highly aligned cells. In
this work, we present a method of producing cell-seeded hydrogel strips with controllable
properties that produce aligned 3D fascicle-like tissue constructs with a high cell density and a
high aspect ratio.
2.2 RESULTS AND DISCUSSION
2.2.1 SACRIFICIAL EXTERNAL MOLDING
To produce an elongated, highly dense fascicle structure in a robust and consistent
manner, we form the fascicle hydrogel within a soft sacrificial mold that constrains the hydrogel
19
to generate a desired shape. Fig. 2.1A shows a permanent mold with a horizontal through-hole. A
removable spacer, or a pin, is inserted into the through-hole (Fig. 2.lAi)
.
Warm liquid
sacrificial mold material, such as gelatin, is filled into the permanent mold with the spacer in
place. A key aspect of this method is that reagents to facilitate the gelling of the scaffold
hydrogel are dissolved in the sacrificial mold material. Once cooled solid (typically at 4* C), the
spacer is removed from the sacrificial mold leaving a cavity tube (Fig. 2.lAii), and cells mixed
with a dissolved hydrogel precursor are seeded into the cavity tube (Fig. 2.lAiii). The device is
then warmed to a temperature where the sacrificial mold melts. As the mold material liquidizes,
the dissolved gelling reagents freely diffuse into the scaffold hydrogel (Fig. 2.lAiv). The
scaffold hydrogel solidifies due to the combined effects of temperature and gelling reagents.
Lastly, 370 C media is infused via a media-changing well adjacent to the seeding well to remove
the melted sacrificial mold material (Fig. 2.lAv). The entire fascicle-like construct is fully
exposed to the surrounding media, facilitating perfusion throughout the experiment.
The novelty in this method is that the outer mold becomes liquid as the inner hydrogel
becomes solid. This process is diagramed in detail in Fig. 2. 1B. At a low temperature (such as
40
C) the mold is solid and the inner cell/hydrogel solution is liquid. As the device is warmed to
37 *C, the mold melts releasing a gelling agent. At 37 *C, the inner hydrogel has become solid,
and the mold has become completely liquid. The inner hydrogel is then freely suspended across
the well, anchored to the PDMS walls.
Our method of creating suspended hydrogel strips exploits the properties that 1) the
melting temperature of a sacrificial mold material overlaps with the temperature range of the
hydrogel gelation and 2) hydrogel crosslinking agents embedded in the sacrificial mold material
are released into the hydrogel as the melting/gelation occurs. See Fig. 2.1B. These phenomena
allowed us to mold the hydrogel within a sacrificial mold and then simultaneously melt the
sacrificial mold and crosslink the hydrogel. This strategy prevents the hydrogel solution from
solidifying during the injection process, which occurs if the hydrogel is mixed with crosslinking
agents before injection. Furthermore, no hard substrate or solid surface is used for forming the
hydrogel. This eliminates any difficulty in removing the hydrogel from a mold without sticking,
facilitating highly uniform, consistent, and reliable production of very high aspect ratio
constructs.
20
The use of sacrificial molds has been reported in prior cell culture works, however, only
as a removal layer [Muraoka et al., 2010] or as internal molds (positive molds) [King et al., 2004,
Golden and Tien, 2007, Bellan et al., 2012] rather than outer molds (negative molds). For
example, microvascular-like channels have been produced in a Collagen I gel by using a
removable or sacrificial material as an internal mold. Once the internal mold is removed, a
scaffold with embedded microchannels is created, and endothelial cells were seeded into the
scaffold. We present a contrasting method as we invert the "inside" and "outside" materials. By
exchanging the removable and permanent materials, we produce a device (Fig. 2.1) with the
unique features described above. The resultant fascicle-like construct is not in contact with any
hard substrate, yet a desired shape is created. To our knowledge there is no prior work on using
1) a sacrificial material, like gelatin, as an outer mold for cell seeded hydrogels, or on utilizing 2)
the overlapping external melting/internal gelation temperatures of the materials used.
In order for this method to work, the sacrificial mold material must be solid at the
temperature at which the hydrogel solution is injected. If the mold material is not solid, no
cavity will be formed in which to inject the hydrogel solution. The mold material must be liquid
at a temperature below a level affecting cell viability, ideally below 37 *C. See Fig. 2.1B. For
practical implementation purposes, the ideal temperature for injection is room temperature (20
*C), and the material should have a viscosity low enough to inject into the seeding wells around
the spacer pins without introducing inconsistencies and/or bubbles in the well. We have found
that gelatin dissolved in culture media at a concentration of 5-20% satisfies these requirements
(See supplementary information for more details).
Using fibrin as an exemplary hydrogel, we demonstrate that we can release a gelling
agent (thrombin) from the sacrificial mold material. Fibrin is produced through interaction of
two materials. Fibrinogen is the primary precursor component of fibrin, mixed with seeding
cells. Thrombin is a serine protease that converts fibrinogen into gel-forming fibrin strands
[Weisel, 2005]. Thrombin is dissolved in the sacrificial mold material and reacts with fibrinogen
when the mold melts. Additionally, fibrin has been shown to be a good scaffold for producing
aligned contractile skeletal muscle cells [Hinds et al., 2011, Huang et al., 2005].
An alternative to dissolving thrombin in the sacrificial mold material is to mix the
thrombin and fibrinogen solution immediately before injecting. There are significant drawbacks
to using this alternative method because of the quick onset of gelation once thrombin is mixed
21
with fibrinogen. The viscosity of the cell laden hydrogel keeps changing as the property of
fibrinogen changes during injection, leading to a non-homogeneous fibrin strip. It is particularly
difficult to inject it uniformly into multiple holes to create multiple fascicle strips. These nonuniformities may also generate randomly distributed stress concentrations affecting cell
behavior. Conversely, dissolving thrombin in the sacrificial mold allows us to fully control the
gelation timing; gelation does not occur until the temperature is increased to approximately 37
'C. (See supplementary information for more details).
We have used this method to embed multiple cell types including motor neurons in
embryoid bodies, fibroblasts, tenocytes, and skeletal myoblasts in multiple hydrogels including
fibrin, collagen, and PEG hydrogels (See Supplementary Fig. 2.S1). Here, we focus on results
from seeding C2C12 mouse myoblasts in fibrin hydrogels.
2.2.2 FASCICLE-LIKE TISSUE FORMATION
Fig. 2.2A shows a developed fascicle-like tissue 14 days after seeding with C2C12
formed myotubes distributed throughout the length of the tissue. The total length of the tissue is
5 mm, while the diameter is -100 pm, resulting in an aspect ratio of 50:1. Striations of a-actinin
along the length of each myotube are clearly visible (Fig. 2.2C) indicating sarcomeric structure
capable of contraction (Supplemental Movie 1). Phase contrast images (Fig. 2.2D) show that
cells become more aligned, as the tissue develops from 2 to 14 days post seeding. By day 14,
these tissues contain a high volumetric density of packed contractile muscle cells aligned axially
along the full length of the fascicle. The cross section images in Fig. 3.2B demonstrate that
myotubes are densely populated throughout the tissue. Two distinct features of the method were
exploited to produce these fascicle-like constructs with dense, aligned cells; 1) controlled
scaffold geometry ensures that diffusion-based transport is possible for internally located cells,
and 2) cells align passively along the axis of the scaffold material due to cell mediated
compaction of the hydrogel.
In the above experiments we have observed the fascicle-like construct becomes taut,
indicating that internal tension is generated between the two ends. The whole construct is not
constrained by any substrate except for both ends where the hydrogel adheres to the permanent
mold. Therefore, the tension is uniform throughout the construct along the longitudinal axis. This
22
internal stress acts to align the fibrils of the hydrogel that provide alignment cues to the seeded
cells [Brown et al., 2009, Sander et al., 2011]. It is important to generate uniform tension
throughout gel development to align the hydrogel fibrils uniformly. This is a passive process
requiring no external effort to induce such aligning stress.
2.2.3 IMPORTANCE OF AXIAL STRESS
To investigate the role of the stress generated between two anchored ends of the
hydrogel, and its effects on cell density and alignment, we conducted a series of experiments. In
one experiment, we relieved the generated stress by disconnecting one end of each taut construct
from the permanent mold two days after seeding as diagramed in Fig. 2.2G. The resulting
constructs all compacted significantly towards the remaining anchored end (Fig. 2.21). The cross
section of such constructs reveals that cells occupy only the periphery of the compacted construct
and that the volumetric cell density is low (Fig. 2.2J). Within these single anchored constructs,
a-actinin, a contractile protein, was found in only a few short, unaligned myotubes. In contrast,
constructs maintained in suspension between two anchors produced significantly higher density
of a-actinin (Fig. 2.2H). In another experiment, we examined the potential aligning effects due to
cell fusion (myotube formation). We prohibited myoblasts from differentiating to fused
myotubes by increasing E-aminocaproic acid concentration to 10 mg ml~1 in the media [DiazRamos et al., 2012], and examined whether the individual cells still align or not. Fig. 2.2E and F
show that, although the cells were arrested in the myoblast phase, the cells densely populated the
interior of the constructs and were aligned with their nuclei elongated along the longitudinal axis.
Therefore, the cells align without fusion under stress.
2.2.4
CONSTRUCT DIAMETER
To examine the effect of scaffold geometry, constructs made with various removable pin
diameters were produced, and changes in construct diameters were monitored via phase contrast
images over the course of two weeks as the myoblasts developed into functional myotubes (Fig.
2.3A). The diameter varies along the length of tissue construct. We measured the diameter at
each point along the length and at each time point by using custom MATLAB code. Fig. 2.3B
shows changes to the average diameter computed for the entire length of each construct over 10
days.
23
In each construct, significant compaction occurred within the first day. This is consistent
with prior works on initial cell mediated compaction of cell laden fibrin [Bian and Bursac, 2009,
Sander et al., 2011, Kim et al., 2011].
After the first day, the diameter of the developing
constructs slowly decreased continuously. The ratio of the cross sectional area at a particular
time point to the initial cross sectional area (At/Ao) of a construct (the 'area ratio'), was
independent of initial diameter for the first few days (Fig. 2.3C). We observed that the initial gel
compaction is fairly independent of the gel's geometric size. However, as time progressed,
constructs with smaller diameters had substantially smaller area ratios, compared to those of
larger diameter constructs. This indicates that over time, smaller diameter constructs degrade the
fibrin scaffold more than the larger diameter constructs. In other words, the longer-term
interactions between cells and their surrounding scaffold material are more active for the thinner
constructs, where the cells are perfused properly. Increasing construct diameter limits diffusion
based processes of internally located cells.
We normalize the variations in construct diameter by dividing the standard deviation of
the construct diameter with its mean value. Fig. 2.3D shows this normalized standard deviation
for constructs with different initial diameters.
Over time, the standard deviation tends to
increase, indicating that the initial construct geometry is more uniform. Of significance is the
result that smaller diameter constructs tend to increase in variation, (i.e. decrease in uniformity),
more so than larger diameter constructs. Therefore, in designing tissue constructs, there is a
trade-off between volumetric density of contractile cells (Fig. 2.3C) and uniformity of the cross
section (Fig. 2.3D) along the construct.
2.2.5
CELL DENSITY
To quantify the volumetric cell density, a-actinin, a marker of mature myotubes, was
imaged via confocal microscopy, and its total volume was determined. The volumetric density of
a-actinin was calculated as the ratio of the total volume of a-actinin to the total volume of the
construct (Fig. 2.4C). Total volume is approximated from final diameter data generated from
phase contrast images using custom MATLAB code (Fig. 2.4A), while a-actinin volume (Fig.
2.4B) is calculated from 3D reconstruction via Imaris software generated from confocal image
stacks (Fig. 2.4D). The ratio of a-actinin volume to total volume (Fig. 2.4C) serves as a measure
24
of developed myotube density within the construct that can objectively be used to compare
constructs grown under different conditions.
The presented method has produced a high cell density of over 50% in terms of a-actinin
volume ratio, as shown in Fig. 2.4C. It allows us to vary several key parameters that affect the
fascicle formation. These parameters include initial diameter, fibrinogen concentration within the
injected hydrogel-cell suspension, and thrombin concentration dissolved in the sacrificial mold.
As expected, both the total volume and a-actinin volume are positively correlated with initial
diameter size (Fig. 2.4A,B). Interestingly though, the volumetric density of a-actinin is higher
for constructs with smaller initial diameters; over 50% volumetric density was achieved for the
diameter of 254pm (Fig. 2.4C).
Smaller initial diameter constructs end up with higher
concentrations of contractile cells, leading us to hypothesize that increasing construct diameter
limits diffusion based processes of internally located cells.
2.2.6 FIBRIN COMPONENT VARIATION
When initial fibrinogen concentration was varied, constructs with higher fibrinogen
concentration decreased in volume significantly more (Fig 2.4A). Despite this, there was no
statistically significant variation in a-actinin volume for various concentrations of fibrinogen
(Fig 2.4B). Combining these two results, the volumetric density of a-actinin is greater for higher
concentrations of fibrinogen than for lower concentrations (Fig. 2.4C).
Varying thrombin concentration produced similar results to varying fibrinogen
concentration; the total resulting volume of the constructs was negatively correlated with
thrombin concentration (Fig. 2.4A), while the volume of a-actinin was uncorrelated to thrombin
concentration (Fig. 2.4B). The resulting volume ratio of a-actinin was thus positively correlated
with thrombin concentration (Fig. 2.4C). We hypothesize that these results occur because a
more structurally dense and cross linked hydrogel may lead to cells exerting greater compaction
forces if the cell density is maintained high enough.
These results demonstrate that the these three groups of molding parameters, a) sacrificial
mold geometry, e.g. diameter, b) hydrogel concentration, and c) crosslinking agent
concentration, can be used to vary and tune key properties of fascicle-like constructs, such as cell
density and construct size, in a statistically significant manner.
2.3 MATERIALS AND METHODS
25
2.3.1 FABRICATION OF PERMANENT MOLDS
Multiple molding steps were utilized in the fabrication of the device. An initial shallow
aluminum mold with steel pins spanning the width of the mold was used to cast PDMS slabs
with numerous properly sized and spaced through holes. The aluminum mold was milled using a
3/16" end mill and a #77 or larger drill bit. The default pin diameter was nominally 356
sm.
When pin diameter was treated as the independent variable, pin diameters of 254, 305, 356, 508,
787 and 1016 pm were used (McMaster-Carr). The aluminum and a single pack of steel pins
cost less than $25. The PDMS slabs were cut into multiple PDMS chips, each with a single
through hole. 5 mm holes were punched in a line
-
3 mm apart along the through hole of each
chip. These PDMS chips were then bonded to glass forming 3 wells per device. In each device,
a steel pin was inserted on each side of the through hole such that each pin spanned one of the
two outer 5 mm wells to be used as seeding wells. The central well was to be used as a media
reservoir.
2.3.2 CULTURE OF C2C12 MYOBLASTS
C2C12 mouse myoblasts (Ameican Type Culture Collection) were cultured in growth
medium (GM) containing DMEM (American Type Culture Collection), supplemented with 10%
fetal bovine serum (FBS, Sigma-Aldrich), 1% penicillin-streptomycin 10OX (Invitrogen), and 0.1
mg ml-1 Normacin (Invivogen). Confluence was kept below 70%.
2.3.3 CASTING OF SACRIFICIAL MOLD AND HYDROGEL CONSTRUCTS
The wells of each device were rinsed with PBS (Invitrogen) and aspirated.
Gelatin
(Sigma-Aldrich) was added to GM at a concentration of 5% w/v and melted at 370 C. 1% 0.5 M
NaOH and thrombin were added to the gelatin solution which was then immediately injected into
the seeding wells of the devices, submerging the steel pin spanning the each seeding well. The
devices were then cooled at 4' C for 20 minutes to solidify the gelatin. The default concentration
for thrombin was 1 U ml-1. Thrombin concentrations of 0, 0.1 U ml-1, 1 U ml~ 1 , and 10 U ml-1
were used when thrombin concentration was treated as the independent variable.
The cell/hydrogel mixture was prepared by mixing fibrinogen (Sigma-Aldrich), Matrigel
(BD Biosciences) (20% v/v), and C2C12 myoblasts (10x10 6 cells ml-i) with GM and kept on ice.
26
The default concentration of fibrinogen was 5 mg mi1 . When fibrinogen was treated as the
independent variable, its concentration varied between 1.25, 2.5, 5, and 10 mg mI.
The devices were seeded by removing the steel pin spacers from the solid gelatin mold,
and injecting the cell/hydrogel mixture, via micropipette, into the cavity formed by the pins.
Immediately after seeding, the devices are placed into an incubator at 370 C and 5% C02. After
30 to 60 minutes, GM is added to the medium reservoir well. GM in the media reservoir is
replaced daily twice, then replaced with differentiation medium (DM) comprised of DMEM,
10% horse serum (Sigma-Aldrich), 1% penicillin-streptomycin 100X, and 0.1 mg ml'
Normacin. Experiments were terminated 14 days after initial seeding.
2.3.4 IMMUNOSTAINING
Samples were fixed with 4% paraformaldehyde at room temperature for 15 minutes,
rinsed with PBS, permeabilized with 0.5% Triton-X (Sigma-Aldrich) in PBS for 20 minutes, and
blocked with a mixture of 10% goat serum (Sigma-Aldrich) v/v, 1% w/v bovine serum albumin
(Sigma-Aldrich) in PBS for 1 hour all at room temperature.
Samples were treated with the
primary antibody for sarcomeric a-actinin (Invitrogen) at a 200:1 dilution for 1 hour at room
temperature, followed by the Alexa Fluor 488 secondary antibody (Invitrogen) at a dilution of
200:1 overnight at 4* C.
2.3.5 AUTOMATED IMAGE ANALYSIS
Custom scripts written in MATLAB (Mathworks), determined the changes in diameter
along the muscle strips. To determine the strip from the background, the differences in image
intensity in the x and y directions at each pixel in the image were combined. Larger pixel
intensities correspond to the muscle strip. Otsu thresholding was applied to segment the strip in
the image [Otsu, 1979]. We determined the first and last thresholded pixel along the x direction
of the image to be its top and bottom edges. A straight line was fitted on centerline pixels. The
angle of the strip to the horizontal (x-axis) was calculated from the slope of this line. The top
and bottom pixel coordinates were rotated at this angle so that the strip was horizontally aligned.
The diameter at each x-coordinate was calculated as the difference between the top and bottom
y-coordinates.
2.4 CONCLUSIONS
27
In summary, the sacrificial outer molding method allows us to produce fascicle-like 3D
constructs consisting of densely populated (over 50% volumetric cell density), aligned cells with
a high aspect ratio (~100:1). This was made possible by exploiting three major features of the
method: 1). The outer mold is sacrificial, and releases hydrogel crosslinking agents when
melting, resulting in a scaffold simply suspended between two anchoring ends, with no solid
substrate; 2). Cell-mediated compaction of the hydrogel creates uniform internal tension along
the entire construct, which aligns the cells and promotes maturation; and 3) The sacrificial outer
mold with a tunable diameter provides the hydrogel with an initial geometric constraint and
allows the 3D construct to be exposed to a surrounding media, facilitating perfusion across the
entire cross section of the construct. The present molding method is flexible and expandable to a
broad class of cell types, providing a unique approach to formation of fascicle-like constructs.
2.5 ACKNOWLEDGEMENTS
This material is based on work supported in part by the National Science Foundation,
under Grant No. CBET-093951 1, the Science and Technology Center for Emergent Behaviors of
Integrated Cellular Systems (EBICS), and in part by the Singapore-MIT Alliance of Research
and Technology, BioSyM IRG.
2.6 SUPPORTING INFORMATION
Sacrificial Mold Requirement
In order for this method to work, the sacrificial mold material must be solid at the
temperature at which the hydrogel solution is injected. If the mold material is not solid, no
cavity will be formed in which to inject the hydrogel solution.
For practical implementation
purposes, the ideal temperature for injection is room temperature (20 'C). Additionally, the mold
material must be liquid at a temperature below a level affecting cell viability, ideally below 37
*C. For implementation purposes, the material should have a viscosity low enough to inject into
the seeding wells around the spacer pins without introducing inconsistencies and/or bubbles in
the well. If a bubble is introduced that touches the pin, then when the pin is removed, the cavity
will include the geometry of the bubble.
Gelatin dissolved in PBS or culture media meets these requirements when dissolved at
certain concentrations. Gelatin dissolved at 0.5%, 1.5%, 5 %, 20%, and 30% w/v was used to
28
test for appropriate concentration levels to meet the following criteria: melting temperature
below 37 *C, solid at room temperature (20 *C), and implementable viscosity. Melting below 40
*C is a hard requirement for gelatin because the strength and viscosity gradually weaken upon
prolonged heating above 40 'C.
A gelatin concentration of 30% did not melt at 37 *C, and could not be injected around
the spacer pin. Concentrations of 0.5% and 1.5% were liquid at 37 'C, but do not solidify at
room temperature. At these lower concentrations, no cavity was formed upon removal of the
spacer pin.
Concentrations of 5% and 20% met the temperature requirements. After filling the
seeding wells, and removing the spacer pins, cavities remained within the gelatin.
It may be worth noting that the filled cavities of the 5% concentration wells are slightly
larger in diameter than the 20% concentration wells. This is likely due to the increase in gelatin
stiffness at higher concentrations.
The most important difference between 5% and 20% is the difference between their
viscosities and how this effects injection in the seeding well. The warm liquid 20% gelatin is
much more difficult to work with and results in more inconsistencies/bubbles when filled into the
seeding wells than the 5% concentration. The 5% gelatin does not liquefy for several minutes
and is easily injected via 200 pIl pipette tip. Conversely, the 20% gelatin liquefies readily at
room temperature, and even quicker while injecting into the seeding well leaving numerous
bubbles and requiring careful removal of the pipette tip so as not to remove the gelatin along
with the tip.
In conclusion, when gelatin dissolved in an aqueous solution is used as the sacrificial
mold material, w/v concentrations between 5% and 20% should be used. Within this range, the
melting point is between 20 *C and 37 *C allowing for a solid mold when seeding and release of
the mold when incubating. Additionally, concentrations at the lower end of this range are easier
to work with, however gels at this concentration are less stiff affecting resulting hydrogel
geometry slightly.
Gelling reagent release from sacrificial mold
A significant advantage to using the sacrificial mold comes from being able to dissolve
additional reagents within the sacrificial mold material. These reagents are released into the final
29
gel solution as the sacrificial mold melts. It is particularly useful to dissolve a gelling reagent
into the sacrificial mold material. To form a fibrin gel strip using this method, first thrombin is
dissolved into the sacrificial mold, next fibrinogen solution is injected into the cavity formed by
removing the spacer pin from the mold, and then the thrombin is released into the fibrinogen
solution as the sacrificial mold melts. This process creates a uniform strip of fibrin spanning the
seeding well.
An alternative method is to not include thrombin in the sacrificial mold and mix the
thrombin and fibrinogen solution immediately before injecting.
There are 2 significant
drawbacks to using this alternative method arising from the quick onset of gelation once
thrombin is mixed with fibrinogen; 1) there is only time enough to fill a single channel after
mixing, and 2) the fibrin strip formed will be non-homogeneous. The first issue, that there is
only time for one channel to be filled for each mixing, adds greater deviation from one strip to
the next because of the slight differences in mixing and increased time difference between
injections. This time difference is especially important for larger experiments with many strips.
Additionally, from a practical standpoint, this adds time and tedium to the process. Conversely,
dissolving thrombin in the sacrificial mold allows for rapid filling of many channels. The second
issue, that the fibrin strip is non-homogenous, results from the gel forming as it is being injected.
These nonuniformities may generate randomly distributed stress concentrations affecting cell
behavior.
To demonstrate these drawbacks, four conditions were tested. For every condition, 5%
w/v gelatin in PBS was used for the sacrificial mold, and 5 mg/ml fibrinogen in PBS was
injected. In the first condition, thrombin was dissolved in sacrificial molds. For the second
condition, 120 pl of fibrinogen solution was mixed with thrombin via pipetting resulting in a
solution of 5 mg/ml fibrinogen and 4 U/ml thrombin, and this solution was injected into the mold
cavity. The third condition was the same as the second, except 20 pl was used instead of 120 pl.
The fourth condition was a control with no thrombin. All four conditions were immediately
warmed in a 37 C incubator after injection of the fibrinogen solution.
The control condition with no thrombin formed no fibrin strip as expected. The condition
incorporating thrombin in the sacrificial mold formed strips that were visibly uniform. The
fibrin strips generated by first mixing the thrombin and fibrinogen before injection were visibly
nonuniform.
30
2.7 FIGURES AND CAPTIONS
A
Top View
B Outer
Material
I)
Inner
Material
0
4
C
i
ii)
Cavit
Cavity
iii)
iv)
370 C
v)
N_
Fig. 2.1 | Fascicle-like tissue construct production technique utilizing a sacrificial outer mold,
the overlap of mold melting temperature/hydrogel gelling temperature, and cell mediated
compaction. (A) Schematic cross sectional and top view of multi-molding stage technique. (B)
Upon heating to 370 C, the gelatin melts and the hydrogel begins to solidify as gelling agents
dissolved in the gelatin are released into hydrogel upon melting of the gelatin, then cell mediated
gel compaction generates anisotropic stress.
31
hi3
H 4
IO 2
>ci
uw
Fig. 2.2 1Fascicle-inspired tissue constructs. Longitudinal stack (A), and axial slice (B)
confocal images of developed fascicle-like structure of C2C12 myotubes showing high
volumetric density of aligned cells (white arrows) and contractile protein, a-actinin. (C) Clear
striations of a-actinin within myotubes. (D) Phase contrast images of a cell seeded construct
showing progressive alignment of cells from 2 days to 14 days. (E) 2-Photon microscope
images of a construct with C2C12 cells arrested in the myoblast state showing alignment of
actin and elongated nuclei (white arrows) and (F) Imaris renderings of F-actin showing cell
populated internal regions without myotube formation. (G) Schematic sideview of a seeding
well with a construct with one end disconnected from the wall 2 days after seeding. (H)
Volumetric density of sarcomeric protein, a-actinin, with both ends anchored (control) vs.
single-anchor-removed (disconnected) constructs (n=5). Longitudinal stack (I) and axial
section (J) confocal images of fascicle like constructs having had one end disconnected 2 days
after seeding showing poor volumetric density, poor myotube formation and a large central
region void of cell (orange arrow). Scale bars 50 pm for A-E,I,J and 10 pm for F.
32
A
B 300
350 micron initial diameter after 10 days
200
100
0
1
2
D~
Cl
4
7
10
254 pm
=356 pm
30
8
508 pm
787 pm
W
6
3
Days ater seeding
790 micron initial diameter after 10 days
28
=1016 pm
10
S
2
0
0
1
2
0
3
4
7
Days ater seedig
~0L
10
1
2
3
4
7
Days alter seeding
10
Fig. 2.3 Geometric changes in fascicle construct structure over time. (A) Phase contrast
images analyzed with custom MATLAB code to measure diameter at each point along the
construct. The green, red, and blue lines indicate the top, center, and bottom of the tissue. (B)
Mean diameters of individual constructs with differing initial diameters decrease gradually
over time. (C) The compacted area ratio (the ratio of the mean compacted cross sectional area
to its initial cross sectional area) shows little initial dependence on diameter 1 day after
seeding, and decreases over time. After 10 days, the compacted area ratio is dependent on
initial diameter. (D) The standard deviation of diameter along single constructs for various
initial diameters is independent of initial diameter after 1 day, and increases over time, but
increases faster for constructs with smaller initial diameters.
33
A
co
B0
30
20
2
15-
20-
04]
-
15-
10-
0
10
5
5
0
0
4
4
8
3
3
6-
x 2.
E3
C
20
--
0--
2.
4
22
0
0.4
1.5 .
.20.35-
.
.
.
.0.8-
0113
TI0.4
S0.2-
0.~ j_
F2r
Frge
C5W 1.25
5
10
Fibinogen
Concentration [mg/ml]
Thobn1nta
0
0.1 1 10
Thrombin
Concentration [ULml]
0.2LI~
iLLL
Daee
254 305 356 508
Initial Diameter
Fig. 2.4 | Variable parameters may be used to control the development of fascicle-like
constructs. Total volume (A), volume of contractile protein a-actinin (B), and volumetric
density of contractile protein a-actinin after 14 days of development (C) in relation to initial
fibrinogen density, thrombin concentration, and initial diameter (n=5-6). Increasing
concentration of fibrinogen and thrombin decreases total volume (A) while leaving a-actinin
volume unchanged (B), resulting in high a-actinin density for higher concentrations of
fibrinogen and thrombin (C). Increasing initial diameter increases both total volume (A) and aactinin volume (B) but decreases a-actinin density (C). (D) A top view (1) and isotropic view
(2) of the volumetric density of a-actinin in a typical construct. Scale bars 50 pm.
34
Fig. 2.S1
I Demonstration of how the presented method may be expanded to a broad class of
cells and hydrogels providing a unique approach to the formation of fascicle-like constructs. Phase
contrast images of (A) 3T3 fibroblasts and (B) horse tenocytes seeded in fibrin. (B) Phase contrast
image of C2C 12 cells seeded in PEG hydrogel. (C) Florescent image of GFP expressing motor neurons
with axons extending from embriod bodies seeded in fibrin.
Fig. 2.S2
I Low gelatin concentration. 0.5% (left) and 1.5% (right) gelatin concentrations resulted
in no cavity remaining after spacer pin removal.
The cavities visible are those inside the PDMS
permanent mold.
35
Fig. 2.S3
I Suitable gelatin concentrations. 5% (top 2), and 20% (bottom 2) gelatin concentrations
in cavities remaining after spacer pin removal (1st and 3rd from top). Upon injection of another fluid,
these cavities are filled (2nd and 4th from top).
36
Fig. 2.S4 I Effects of gelatin viscosity. 5% (left) and 20% (right) gelatin concentrations seeded into
the top and bottom seeding wells of each device. The 5% gelatin wells are clear and without bubbles.
The 20% gelatin wells contain bubbles and have inconsistent texture resulting from tip removal.
37
Fig. 2.S5
I Nonuniformities from mixing thrombin and fibrinogen. Strips of fibrin formed
using different methods to incorporate thrombin.
Thrombin dissolved in the gelatin sacrificial mold
(top), thrombin mixed with fibrinogen for 120 pl of solution (mid), and thrombin mixed with fibrinogen
for 20 pl of solution (bottom).
38
CHAPTER 3: MECHANICAL CHARACTERIZATION OF OPTICALLY AND
ELECTRICALLY STIMULATED FASCICLE-LIKE CONSTRUCTS
3.1 INTRODUCTION
The mechanics of muscle tissue has been practiced for many decades. Understanding the forces
that are generated under varying conditions is important in understanding how muscle functions
when healthy, damaged, or grown in in vitro.
A common practice is to measure the isometric tension generated by skeletal muscle
tissue. This is the measurement of force generated while the tissue maintains a constant length
by holding the two ends of the tissue at a constant position relative to each other throughout the
measurement.
The length may be maintained either by coupling it to rigid supports or by
applying feedback position control of the ends. Feedback position control has been used as far
back as 1965 when live frog muscle was tested to determine the relationship between sarcomere
length and isometric tension generation [Gordon et al., 1966]. In vitro grown 3D muscle tissue
constructs have been tested using stiff couplings [Dennis and Kosnik, 2000], and stiff coupling
muscle testing tissue products are available from companies such as Aurora Scientific. One
potential issue with using stiff couplings is that a force transducer must be serially coupled to the
tissue, and force sensors are necessarily compliant; they work by measuring displacement.
However, if the force sensor is sensitive enough such that it may be made significantly stiffer
than the tissue being measured, then the compliance of the force sensor does not significantly
influence the measurement.
Isometric force measurements are useful in finding peak forces achievable by muscle
tissue, and in comparing muscle tissue under various conditions, such as testing effects of
various influences like alignment, drugs, or damage.
However, isometric force measurements
cannot provide time dependent information about muscle contraction such as power, or twitch
energy. Isometric force measurements cannot provide information on the energy that can be
imparted on a load for a single muscle twitch.
Twitch energy transmissibility is critically
important for cardiac tissue as well as for any engineered mechanical system powered muscle
tissue (biobots) [Chan et al., 2014].
With the addition of velocity control, a typical isometric muscle testing apparatus may be
used to measure power output directly by testing isovelocity shortening [Brooks and Faulkner,
39
1991]. However, this power measured is dependent on the magnitude of the velocity, and a
constant velocity is not necessarily the conditions in which muscle will be functioning.
An alternative method of characterizing skeletal muscle tissue performance is to have the
muscle tissue work against as load such as a known stiffness. With this aim muscle tissue has
been grown such that it's two terminal ends are supported by the ends of two compliant posts
[Vandenburgh et al., 2008, Sakar et al., 2012, Boudou et al., 2012]. This method allows for the
direct measurement of muscle tissue output energy transmission to a load. However, the
performance of any given tissue can only be assessed for driving a single load, the stiffness of
the cantilevered posts. Fully characterizing the tissues against various loads is not possible.
Additionally, it is impossible to remove the muscle tissue from the posts devices without
damaging or altering its performance in the process. One final concern with using polymer posts
is the nonlinearity of their stiffness. The Young's Modulus of PDMS can change by roughly a
factor of 2 simply by changing temperature from 20 *C to 37 *C, and other factors such as
moisture or pH may also change compliance. Polymers are also viscoelastic, therefor the resting
stiffness must be measured separately from the short term, activation stiffness of polymer posts.
For these reasons, using highly linear metal cantilevers with the ability to use multiple
cantilevers on the same tissue is desirable.
The primary objective of the mechanical characterization presented here is to determine
the optimal diameter at which to grow 3D, aligned muscle tissue having a high density of
contractile proteins. 2 additional objectives include 1) determining the performance disparity
between optical vs. electrical stimulation of the tissue, and 2) demonstrating that an optimal load
impedance exists for a particular muscle construct to transmit mechanical work to. In the context
of work presented here, performance is assessed in terms of maximum work transmitted to a load
due to a single stimulated twitch cycle. Force and/or displacement may also be used in twitch
performance comparison because these two values are monotonically coupled to work.
There are multiple factors that may influence measurements and are issues with all muscle
characterization methods described above. They include time dependent mechanics (such as
creep and stress relaxation), and the stress/strain of the tissue at the time the measurements are
taken (pre-stress/pre-strain). To account for these influences the time-displacement profiles of
each experimental measurement must be kept constant between samples.
In isometric
contraction experiments, the tissue is typically pre-strained to the desired value, and left to rest
40
for a specific time period (typically on the order of minutes) before contractions are stimulated
[Wang et al., 1993, Dennis and Kosnik, 2000].
Incubator like conditions are used in the
experiment environments because experimental time is increased due to waiting for time
dependent forces to attenuate. In compliant post based methods for muscle characterization,
there is no need to wait for time dependent forces to attenuate, because there is no change in prestrain of the tissue before experimental characterization.
The method presented here uses the lateral displacement of a long muscle tissue
constructs, using a copper cantilever. Lateral displacement entails displacing a tissue construct
at point between its two anchors in a direction perpendicular to the primary axis of the tissue.
Lateral displacement for mechanical characterization has been used for neurite characterization
[Dennerll et al., 1988] as well as for very simple skeletal muscle characterization [Vandenburgh
et al., 1991]. Muscle tissue made via the sacrificial outer mold method described in the previous
chapter facilitates this mechanical characterization method in two ways. First, the tissue is
anchored between the stiff well walls of the culture device, ensuring no serial compliance at the
anchor points that would affect measurements.
This is in contrast to tissue developed on
compliant posts. Second, the tissue can be tested without transferring the tissue. This is in
contrast to the roll up method of producing myooids, which require at least one anchor to be
decoupled from the culture dish on which the tissue is grown. Multiple cantilevers of different
stiffness may be used on a single tissue. This allows for the testing of tissue against different
mechanical loads. Isometric and post based methods alone are incapable of testing a single
tissue against differing loads.
By changing the load, it is possible to determine the load
impedance that matches the specific tissue to maximize the output work of the tissue. This
method has never been used to characterize muscle tissue grown under varying conditions, nor
with different cantilever stiffnesses and displacements.
3.2 MECHANICAL CHARACTERIZATION SYSTEM DESIGN
3.2.1 CONCEPT: LATERAL DISPLACEMENT WITH A CANTILEVER
The basic concept of the force probe is to laterally displace the tissue strip at its center
using the end of a cantilever of known stiffness. This is diagramed in Fig. 3.1. Combining
geometric relations, force balance, and Hooke's law, we can get expressions for the axial force
along the tissue, Fai, and the displaced length of half the tissue, L'. The displacement of center
41
of the tissue is in the direction orthogonal to the un-displaced tissue axis.
As the tip of the
cantilever displaces the center of the strip, the length of the strip elongates, and the cantilever
bends. Knowing the position of the cantilever base and tip, and the stiffness of the cantilever, we
can find the force the tip, Fob,, is exerting on the tissue using Hooke's law:
Fprobe = (Xbase
-
Xtip)kprobe,
where x,, is the lateral displacement of the cantilever tip,
xa,,e
is the lateral displacement of the
cantilever base, and kpo, is the stiffness of the cantilever probe.
The geometric relationships are:
Faxial,x_ Xtp
L'
Faxiai
xtip =
L'2 -
L2
where Faxix is the lateral component of the axial force held by the tissue, and L and L' are the
unforced and forced half-length of the tissue respectively.
And a simple force balance along the direction of displacement is:
Fprobe = 2 Faxial,x
Combining these equations yields an expression for the elongated length, L', and axial force,
L'=
Faxial = (Xbase
-
xt
2+
Xtip) k
L2
t 2
Xro
+ L2
From these equations, we can determine axial force, and tissue elongation.
3.2.2 CHARACTERIZATION SYSTEM
42
For this concept to work, 1) we must be able to finely position the base of the cantilever,
2) the positions of the cantilever base and tip, must be measured simultaneously, and 3) the
muscle tissue must be excitable. How these requirements are achieved is diagramed in Fig. 3.2.
The base of the cantilever is positioned with a 3-axis stage, coupled to the cantilever probe via a
stiff aluminum beam.
The position of the base of the cantilever is measured with a laser
micrometer, which measures the position of a flag coupled to the aluminum beam that extends
away from the sample. The position of the tip is measured via images from a microscope. Video
from the microscope computer coupled with the time history data from the laser micrometer are
combined to give time dependent data.
In order to excite the muscle tissue electrically,
electrodes that are coupled to the cantilever base extend into the sample well and terminate with
ends that are parallel to the strip. In order to excite the tissue optically, the fluorescent imaging
light source of the microscope is used.
3.2.3 ELECTRODE DESIGN
In order to provide electrical stimulation to the muscle strips, I had to build electrodes
into the system that would be capable of providing an electric field across the strip;
perpendicular to the axis of the strip. This field had to be uniform and repeatable.
The
electrodes should be parallel to the strip to generate a uniform field across its length. The
electrodes must be precisely at the depth of the tissue and independent of the displacement of the
tissue as shown in Fig. 3.3a. This means that the electrodes either had to 1) follow the probe tip
as it displaced the tissue (Fig. 3.3b) or 2) stay fixed while the probe tip moves (Fig. 3.3c). The
problem with the first strategy is that we control the probe base, not the probe tip. We would
need continuous feedback to move the electrodes with the tip. The second option was selected.
It was preferable if the electrodes did not need to be positioned independently of the
probe tip or each other. The electrodes and probe tip are required to be at the same depth; the
depth of the tissue in the well. Positioning electrodes and probe tip independently would require
additional hardware and increase setup time for each tissue being tested. Coupling the depth of
the electrodes to the depth of the probe tip simplified the system and ensured the depths of the
electrodes and the probe tip were the same.
The solution was to use compliant cantilevered electrodes rigidly coupled to the base of
the cantilever tip. The electrodes are platinum wires coupled closely to the cantilever tip, that
43
splay out away from the cantilever probe as they extend down. Unforced, the two electrodes are
separated by approximately 3 times the width of the sample well at their ends (Fig. 3.3d). When
inserted into the sample well, the electrodes must first be brought together with tweezers (Fig.
3.3e). The elastic energy stored in the compliant electrodes keeps them pinned to the walls of
the sample well even as the system translates the cantilever probe (Fig. 3.4).
3.3 MATERIALS AND METHODS
3.3.1 TISSUE CONSTRUCTS
Fascicle-like muscle tissue constructs were produced as described in the previous chapter.
For the experiments in this chapter the mold material consisted of 5% w/v gelatin, 10 U/ml
thrombin, and 1% 0.5 N NaCl in PBS, and the cell solution consisted of 5 mg/ml fibrinogen, 10 7
cells/ml, and 0.5 mg/ml aminocaproic acid. The cells used were c2c12 cells (ATCC) transfected
with channel rhodopsin 2 (ChR2) as described in Sakar et al., 2012. For all samples, growth
media was changed to differentiation media 2 days after seeding.
For the initial tissue diameter variation results, two separate experiments were run. The
first consisted of 3 devices using each of the following nominal pin diameters: 0.010" (254 Am),
0.012" (305 pm), 0.014" ( 356 Am) and 0.020" (508 im). The second consisted of 3 devices
using each of the following nominal pin diameters: 0.014" (356 Am) and 0.020" (508 Am),
0.030" (762 Am), and 0.040" (1016 gm). Note the overlap of the 356 Am and 508 Am pin
diameters in these two experiments.
For the optical vs. electrical stimulation experiment and variable probe stiffness
experiments, nominal pin diameters used were 0.020" (508 Jim).
3.3.2 MECHANICAL CHARACTERIZATION
For all experiments, mechanical stimulation testing was performed 14 days after seeding.
For all experiments, multiple twitches were generated via stimulation at the same pre-strain
conditions. In this system, pre-strain means the percent increase in tissue construct length due to
displacement via the cantilever probe tip. Each strip was displaced to approximately 2% strain.
The optical vs. electrical stimulation protocol was designed to account for the effects of muscle
fatigue. 6 stimulation cycles were performed for each strip with 60 seconds of rest between
each. Each stimulation cycle consisted of two sub cycles with 60 seconds of rest between each.
44
Each sub-cycle consisted of pre-straining the tissue, stimulating the tissue 5 times with one form
of stimuli (electric or optical) followed immediately by stimulating the tissue 5 times with the
other form of stimuli (optical or electric), then relaxing the tissue. This protocol is shown in Fig.
3.5. For the variable stiffness experiment, each strip was tested with each cantilever twice using
the following protocol: the strip was displaced by the cantilever tip to a pre-strained to 2% strain
and stimulated 5 times optically, then relaxed.
For the diameter variation experiments, the
protocol was the same as the variable stiffness experiment protocol, except only 1 cantilever
probe was used.
For electrical stimulation, an electric field of 30 V/cm was used with bipolar pulses of 1
ms each, at a frequency of 1 Hz. For optical stimulation, a 300-W mercury lamp coupled with a
GFP (473 nm wavelength) was used to generate 30 ms pulses at 1 Hz using custom scripts
written into the microscope controlling software (Metamorph).
A single cantilever and pair of electrodes were used in the diameter variation and optical
vs. electrical experiments. 4 different cantilevers with different stiffnesses were used in the
variable stiffness experiment. Cantilever stiffnesses were evaluated before and after testing to
make sure the stiffnesses were consistent and that they were not accidentally deformed during
testing. The stiffnesses were evaluated by hanging various small amounts of 40 AWG wire from
the tip against gravity. The stiffness of the cantilever used in the variable diameter and optical
vs. electrical experiments was 0.0278 N/m.
The stiffnesses used in the variable stiffness
experiment were the following: 0.0091 N/m, 0.023 N/m, 0.13 N/m, and 0.42 N/m.
3.3.3 IMAGE AND VIDEo ANALYSIS
Data from the laser micrometer (measuring cantilever probe base position) and the
microscope camera video (measuring cantilever tip position) was processed to generate axial
tissue force and displacement. The video data provides probe tip data with the use of a software
called Tracker. It is free and open source. Tracking the cantilever tip is robust and reliable
giving sub-pixel resolution. Each individual twitch was analyzed independently (Fig. 3.6). The
tip and base data are synchronized based on a step-like input to the base position. The two sets
of data run through an interpolation ensuring that each point in the interpolated data sets are
paired (Fig. 3.6c). Each twitch is segmented and combined with probe base data to determine
axial strip displacement and force (Fig. 3.6d). Finally, each twitch is creep corrected assuming
45
linear creep with a slope equal to the post twitch position minus the initial tip position divided by
the twitch time (Fig. 3.6e). This creep correction is used to ensure that the data presented
represent the force and displacements generated only by the contraction of the muscle.
3.4 RESULTS AND DISCUSSION
3.4.1 OPTICAL VS. ELECTRICAL STIMULATION
Electric stimulation performance was slightly superior to optical stimulation performance
overall. The difference in mean twitch amplitude performance was less than 10% for stronger
strips. However, for poor performing strips, optical stimulation resulted in almost 40% worse
mean twitch amplitude than electrical stimulation. Twitch data for the best performing strip,
strip A (Fig. 3.7a) and the worst performing strip, strip B (Fig. 3.7b) clearly show examples of
this. The optically stimulated twitches nearly follow the electrically generated twitches for the
high performing strip, strip A, while the optically generated twitches are generally much lower in
force than the optically generated twitches for the poor performing strip, strip C. Aggregate data
of the maximum twitch force generated for multiple strips with various overall performance (Fig.
3.7c) show that electric stimulation is mildly superior to optical stimulation. A trend can clearly
be seen when the aggregate of the ratio of optical to electrical stimulation performance for each
strip is plotted against mean maximum twitch force of the electrically stimulated twitches for
each strip (Fig. 3.7d). This clearly demonstrates that, as the overall performance of the strip
increases, the relative performance of optically generated twitches approaches the performance
of electrically generated twitches.
This is likely due to superior strips having more highly developed contractile myotubes
that have a greater number of nuclei. Having more nuclei means that more myoblasts were fused
into the myotubes and increases the chances of having ChR2 producing nuclei. Conversely, it is
likely that poorer performing strips have fused with fewer myoblasts, decreasing the chance of
containing ChR2 producing nuclei. This results in a greater proportion of functional myotubes,
that are not optically excitable.
Another theory that has not been shown explicitly is that
immature/poor performing myotubes may need a greater level of stimulation than optical
stimulation can provide. Further exploring this phenomenon may provide valuable insights on
how myotubes develop.
46
These results show that muscle constructs developed to a certain level may be stimulated
optically and electrically with similar resulting contractions. This means that optical control of
muscle tissue may be used for stimulation purposes in situations where electrodes are too
intrusive or otherwise disadvantageous.
3.4.2 DIAMETER VARIATION
Numerous twitches from multiple strips of differing diameter clearly show an optimal
diameter exists for generating maximum stress (Fig. 3.8a). The maximum peak stress occurs
using the 508 gm nominal diameter pins. The stress performance quickly drops off after this as
the diameter increases. For the constructs formed using the smaller diameters (250, 300, and 350
microns) , the mean stress generated is approximately the same and is independent of initial
diameter, however, the standard deviation in peak twitch force is much greater for the smallest
diameter pins used.
There are clear reasons for the existence of a local maximum stress generating pin
diameter. For the smaller diameter strips, the stress is lower because there exists a greater degree
of non-uniformity as described in Chapter 2 of this document and shown in Fig. 2.3d. This is
phenomenon is diagramed in Fig. 3.9.
Thinner regions of the construct act as compliant
elements in series with the rest of the construct.
These compliant "weak links" absorb
displacement with the elastic tissue, decreasing the resulting output force (Fig. 3.9a). The larger,
more uniform strips have less internal displacement due to more uniform stiff elements in series
(Fig. 3.9b).
For the larger diameter strips diffusion limiting processes discourage myotube formation
within the central region of the construct yielding fewer myotubes per cross sectional area. This
can be seen in Figure 2.4c of Chapter 2 of this document where the volume of a-actinin is used
as a surrogate of myotube volume. Additionally, if there are fewer cells near the central region
to degrade the scaffold, then more scaffold will exist in the central region, resulting in increased
passive scaffold the tissue must work against as diagramed in Fig. 3.9c.
These two phenomenon (decreasing contractile cell density as diameter increases, and
increasing uniformity as diameter increases) can be put together in the simple model linear
component model shown in Fig. 3.9 to describe these results (Fig. 3.8c,d). The simple model
assumes the cross section of the tissue has two regions, 1) an outer annulus of contractile tissue,
47
and 2) and inner disk of non-contractile scaffold. In the model, the thickness of the contractile
ring/annulus is assumed to be -100 microns. When the thickness of the construct is around this
value or lower, all of the tissue is modeled as contractile. As the thickness of the construct
becomes larger, a non-contractile region in the center begins to grow. Reported values of fibrin
stiffness (-kPa) were used to model this inactive region [Duong et al., 2009]. The contractile
region is modeled as producing -1 kPa of peak stress based on observation and previous research
[Dennis et al., 2000], and the stiffness of the cantilever load the tissue is working against is
modeled as -0.01 N/m, which is similar to the stiffness used to gather the data. This model of the
effects of cell density results in the descending regions of the full model plotted in Fig. 3.8.
The other phenomenon observed and modeled is the greater degree of non-uniformity for
smaller diameter constructs. This is modeled as compliant elements in series (Fig. 3.9). The
overall stiffness is dominated by the most compliant elements. Fig. 2.3d in Chapter 2 of this
document shows that variation of the smallest strips (250 micron initial diameter) vary in
diameter by 25%. The effect of this non-uniformity on stiffness can be modeled as two springs
in series having different stiffnesses. Assuming stiffness is proportional to diameter squared, if
the strip is simply modeled as 2 halves with one having 75% of the mean diameter, and the other
having 125% of the mean diameter, then the effective stiffness of this structure is -65% of a
uniform structure having the mean diameter. This disparity was modeled as simply going to zero
as the initial pin diameter increases. This model of non-uniformity results in the decreasing
performance as tissue diameter decreases.
While more detailed models are of course possible, this was the simplest model based on
the physics of the system that captures the general trend observed in the twitch data. This same
model fits well with both the stress and forces generated by the tissues.
Concerning the variation in twitch force between strips, there is greater variability in
smaller construct twitch performance due in large part to the greater dependence on individual
myotubes. Individual myotube performance varies from one cell to the next. The performance
of the smaller diameter constructs varies more significantly than the larger constructs which have
a greater number of myotubes to average performance over.
3.4.3 PROBE STIFFNESS VARIATION
48
The twitches of a single muscle construct working against multiple cantilever probes of
varying stiffnesses are shown in Fig. 3.10. Each data point represents a single probe. The stiffer
the cantilever, the larger the peak twitch force and lower the peak twitch displacement (Fig.
3.10a). A linear fit matches the force-displacement curve with an r2 value of 0.955. This shows
that the higher the impedance of a load, the higher the maximum force of the tissue, but the
lower the displacement. The peak energy stored in the cantilever probe resulting from a twitch is
equal to the half of the product of the peak twitch force and peak twitch displacement. For a
linear relationship between force and displacement, the energy output as a function of
displacement is a parabola with a local maximum.
This energy maximum represents the
maximum energy that a single twitch can transmit to a linear elastic load, such as the cantilever
probe.
The performance of an engineered muscle construct and the impedance of the load it will
drive should match in order for the muscle to transmit the most energy per contraction cycle to
the load. This is clearly shown in Fig. 10b. There is a maximum energy transmission when the
cantilever with an intermediate stiffness of 0.023 N/m is used. Notice that the energy transmitted
to the stiffest cantilever results highest force, but lowest energy transmission of the four probes
used.
Thus, the system presented here has been used to find an impedance that maximized the
output energy and that it is not equal to the impedance that produces the maximum output force.
3.5 CONCLUSION
Using the mechanical characterization system described in this chapter, three key
observations about engineered muscle constructs have been made. First, optical stimulation of
myotubes formed from cells transfected with ChR2 encoding DNA is potentially as powerful as
electrical stimulation if the constructs are developed sufficiently. Second, there exists an optimal
form factor at which to grow muscle tissue constructs seeded in a 3D scaffold where the
performance metric is contraction performance per unit volume. Third, there exists an optimal
load impedance to drive with any individual muscle construct where the performance metric is
energy transmission to a load.
The key observation about optical stimulation broadens the potential use of optically
stimulated muscle tissue. Muscle tissue that can be optically stimulated has been presented, but
has not yet been methodically compared to electrical stimulation [Sakar et al., 2012]. The
49
proposed benefits of optical stimulation over electrical stimulation include having no need to use
potentially invasive electrodes, and the high spatial-temporal accuracy in stimulating specific
muscle cells that may exist near other muscle cells. However a significant question existed about
the performance of optical stimulation. Now, knowing that optical stimulation can produce
similar results to electrical stimulation, future research may be conducted with optogenetic
muscle tissue without the potential concern for vastly inferior performance.
The key observation about optimal form factor validates the hypothesis set forth in this
thesis; that the fascicle level of hierarchical structure should be engineered and developed first,
before scaling up to larger tissue constructs. This optimal form factor also address the concern
addressed in Chapter 2, that maximizing density of contractile proteins may be an insufficient
optimization parameter due to the results of non-uniformity in smaller diameter constructs.
When comparing twitch performance to other constructs made from bulk hydrogels, the directly
measured generated stress is roughly twice as much. This is even more impressive given that the
maximum stress measured in the system is not with isometric contractions, but with contractions
against a compliant load, and that an immortalized cell line is used here rather than primary
derived myocytes typically used in bulk hydrogel constructs which tend to have superior
contraction performance to cell lines [Dennis et al., 2001].
The key observation about characterizing a tissue to find an optimal impedance for
energy transmission demonstrates the utility of this unique tissue characterization platform.
Further, more complex characterization of muscle tissue performance using this system is
possible and the subject of future work. This system may be used to characterize muscle tissue
working against a wide variety of impedance loads. The cantilever probe loads presented are all
simply linear elastic impedances, but don't need to be. Loads with dynamic impedance such as
inertia and damping may also be implemented. Loads may include hydrodynamic conditions
and/or highly nonlinear materials. The probe loading conditions may be made to match the
loading conditions the tissue will be under when eventually put into use. Additionally, thanks to
the similar performance to electrical stimulation, optical stimulation may be used to quickly and
easily test any individual tissue under these arbitrary loading conditions. A valuable potential
use will be testing constructs using cardiomyocytes instead of skeletal muscle cells. Cardiac
muscle operates exclusively in a twitch-like manner and drives a hydrodynamic load as it pumps.
50
Using this system, heart muscle tissue may be more fully characterized than isometric
characterization can do alone.
The key findings of engineered muscle tissue, as well as the characterization system
presented in this chapter will be used to advance the field of muscle tissue research, muscle
related drug testing, and scaling engineered muscle constructs to produce muscle tissue systems
of much larger size.
51
3.6 FIGURES AND CAPTIONS
A
Translate cantilever base
Cantilever, k,
n
Tissue elongates as
cantilever bends
Tissue
B
Xbase
F
Probe
tip
trip
C
Faxial'
L
Xtip
FaXia
Fprobni F
L
Fa iF
Fig. 3.1
I Force
probe diagram. A) Diagram showing how when the cantilever probe base is
displaced, the tip of the cantilever displaces the tissue. B) Free body diagram of cantilever tip when a
tissue is displaced and a diagram with labeled lengths of the cantilever and tissue. C) Simplified free body
diagram and length relationships.
52
Laser
Micromete
A
XYZ stage
Cantilever
Electrodes
Electric Circuit
Microscope
Images
q
Fig. 3.2 | Mechanical characterization system.
Fluorescent
Light Source
A) Schematic of mechanical characterization
system. B) Key components of mechanical characterization system. C) A 3D printed tip with a cantilever
probe.
53
A
E field H
Peobe
Muscle Strip
B
C
E
D
Electrodes
Fig. 3.3 1 Electrode design. A) Desired electric field lines independent of probe tip displacement. B)
Electrode design in which the electrodes move with the probe tip to maintain a uniform electric field. C)
Alternative electrode design in which electrodes do not move as probe tip displaces tissue. D) Image and
diagram of compliant electrodes above tissue well. E) Image and diagram of compliant electrodes inserted
into tissue well. Well diameters are 5 mm.
54
Fig. 3.4 1 Compliant electrodes functioning. As the cantilever tip is displaced to the left (A), center
(B), and right (C) of a well, the compliant electrodes do not lose contact with the well walls. Well
diameters are 5 mm.
A
B
Electric Stimuli First
3.5
-
2.5k
1.5
1.5
-
1
~-
--
-
-
0.5
0
1
0.5
-
Base Position
we
Optical Stimuli
mm Electric Stimuli
4-
-
2T5
First
Optical Stimuli
-. -..-
r.---- -
3
-
- -- -
3
3.5
0
0
5
10
Time
C
15
20
0
[$
5
10
15
20
Tne is]
First Sub Cycle
Second Sub C
Fig. 3.5 1 Stimulation protocol. For the electric vs. optical stimulation experiment, stimulation subcycles consisted of either electric stimuli first (A) or optical stimulation first (B). These cycles were
repeated alternatively in six full stimulation cycles (C) with 60 s of rest between each sub-cycle.
55
50
A
Raw Data From Video
00
~-50
E -100
U150
-250
C
2
6
4
10
8
12
14
16
Tune [sI
Ra w Twitch
B.
D.-210
.S0
E -22
a
19 -23D
93
3
3.5
4
r
C Interpolated Twitch
-200
.S-210
S-22D
o-23D0\
-240L
4.5
-23
ine [4i
E
6
z 8
6
4
X0
CL
CL
3.5
4
4.5
Time [s]
D . Geometry Corrected
z
4
S.
Creen Corrected
4
2
01
-2
0
0.5
Tune [S]
~0
0.5
TuEM [S]
1
Fig. 3.6 1 Twitch data analysis. A) Raw probe tip data is generated from video processing software.
B) Individual twitch data. C) Raw data is interpolated for point-to-point matching with probe base data
recorded with a laser micrometer. D) Interpolated twitch data from base and tip data sets are
geometrically combined to generate data for the axial force of the tissue. E) Twitch data is adjusted for
the slight creep occurring during the twitch assuming a constant creep rate.
56
A
B
Strip A
6
strip C
-
2
trical
Opti cal
- -z
Electrical
1
\
.
- - -Elec
\
5
tical
4
0
03
LL
IL
72
~0-5
-J
0
200
400
600
8O
0!
0
1000
400
200
Tine [Imsi
C
800
600
1000
Time [ms]
1.3-
Optical
Electrical
I
D
-
12
E
1.1
-
C
K~X\
I1
11
1
*~
0
0
r
0
LL
0
*
0R
ir
0.9-
-E4+
0-80-7-
I
0
0.6
A
B
C
D
Strip Name
0.5L-
0
2
4
6
8
Max Force LNI
Fig. 3.7 1 Electric vs. optical performance results. Exemplary twitch data for multiple optical and
electrically stimulated twitches for a good performing strip (A) and poor performing strip (B).
C)
Aggregate data of maximum twitch force achieved for numerous twitches stimulated via electrical and
optical signals. D) The ratio of optically to electrically stimulated maximum twitch force plotted against
mean maximum twitch force generated electrically. Error bars represent
1 standard deviation. Asterisks
represent statistical differences with p > .95. Error bars are standard deviation values.
57
A 2000
B
1500
2
I
T
z
lili.>
1000
CO
0
L-
~5W
0
15
10
254 305 356 508 762 1016
254 I (5 356 508 762 1016
iitial Pin Dianeter smJ]
C
1=,
*
M
140D
A
120D
V
a
*
1fED
25Opmpnda
300 pm pn dda
350 pm pn dda
510 pm pin ddat
760 pin pn dda
1020 pm pi dda
Mod
.-
10
Ar
N
93
/
CL
1~T
800
600-
0D
N
15-
-
L-
20-
L-
/4
Initial Pn Diameter h'nj
4005
200
0
50
10
150
20
250
Tissue Diameter [prm]
30
0
0
50
100
150
200
250
300
Tissue Dianeter [pm]
Fig. 3.8 | Stress and force over various initial pin diameters. Peak stress (A) and force (B) from
twitches generated by constructs made using various initial nominal diameter pins. Asterisks represent
statistically significant differences with p>0.95. Peak twitch stress (C) and force (D) plotted against final
construct diameter from constructs of differing initial pin diameter and a simple model. All error bars are
standard deviations.
58
A
minit-a diameter
-5
compliant region
offtI
interria displacemnt
on
smaller stress
500 pm initialdiameter
B
uniforml stiff regions
off
on
no internal displacement
larger stress
1000 trm initial diameter
C
Offk
passive stiffness inhibits contraction
smaller stress
Fig. 3.9 | Uniform vs. nonuniform strip force. Diagram of consequences of nonuniform strip
development (a) vs. uniform strip development (b) showing that compliant regions in tissue result in
internal displacements of the tissue that decrease the muscle's transmission of mechanical energy to an
external load. (C) As the diameter increases, there is a greater volume of passive material that acts as a
stiffness to decrease contraction stress generation. Scale bars 50 pm.
59
6-
18-
A
-
5
B
0.42 N/m probe
A
*
*
*
0.13 N/m probe
0.023 N/m probe
0.0091 N/m probe
1614-
Fit (r = 0.955)
124).
10-
2C
6-
S24)
40~
2i
I
I
I
I
0
2
4
6
8
10
12
Peak Twitch Displacment [pm]
Data (
Fit
a)
0-
,
I
0
14
0
2
8
10
Tissue Displacement [pm]
4
6
121
1
14
Fig. 3.10 1 Tissue performance under varying load conditions. A) Peak twitch force and peak
twitch displacement for a single tissue stimulated while displacing 4 different cantilevers of different
stiffness with a linear regression fit. B) Peak energy stored in the probe cantilever as a function of tissue
displacement under 4 different cantilever loads and the fit from the linear regression found in A). Note:
probe stiffness values are not equal to the geometry corrected effective stiffness seen by the tissue. Error
bars represent standard deviation. Error bars are smaller than symbol where not shown.
60
CHAPTER 4: DESIGN OF MULTI-DEGREE OF FREEDOM SKELETAL
MUSCLE POWERED SYSTEMS
Note: This chapter was previously published in the proceedings of the IEEE Engineering
in Medicine and Biology Conference (EMBC) 2013, August 26-30, Osaka, Japan.
Bio-integrated robot design using live skeletal muscles as actuators is presented.
Millimeter-scale, optically excitable 3D skeletal muscle bio-actuators are created by culturing
genetically coded precursory muscle cells that are activated with light: optogenetics. These
muscle bio-actuators are networked together to create a distributed actuator system. Unlike
traditional robots where fixed axis joints are rotated with electric motors, the new networked
muscle bio-actuators can activate loads having no fixed joint. These types of loads include
shoulders, the mouth, and the jaw. The optogenetic approach offers high spatiotemporal
resolution for precise control of muscle activation, and opens up the possibility to activate
hundreds of interconnected muscles in a spatiotemporally coordinated manner. In this work, we
explore the design of robotic systems composed of multiple light-activated live muscular
actuator units. We describe and compare massively parallel and highly serial/networked
distributions of these building-block actuator units. We have built functional fundamental
prototypes and present experimental results to demonstrate the feasibility of the construction of
larger scale bio-integrated robots.
4.1 INTRODUCTION
Live cells and tissues cultured in microfabricated in vitro environments can be used as
components for building robots [Feinberg et al., 2007, Xi et al., 2005, Tanaka et al., 2007, Park
et al., 2008, Chan et al., 2012]. Skeletal muscles, for example, can be actuators for powering a
micro-robot or an artificial "animal". Muscle strips can be formed from their precursory cells,
myoblasts, by guiding them through multi-stage myogenic process [Lieber, 2010]. Muscle strips
self-assembled within a robotic structure have potential to activate a high degree of freedom
micro-mechanism, which is a feature which is difficult to attain using current actuator
technology [Neal and Asada, 2011]. Such live biological materials have the potential to be a
significant paradigm changing technology in designing robotic systems and extending their
applications to broader fields.
61
Skeletal muscles can produce over 20 % of strain and over 200 kPa of stress in vivo.
They directly convert chemical energy to mechanical energy without transducing intermediate
energy. Skeletal muscles can survive for 8 ~ 12 weeks in vitro environment. Various shapes and
types of actuators can be created by molding myoblasts and culturing them in a microfluidic
environment [Bian and Bursac, 2009].
Activation and control of skeletal muscles has been a challenging issue. Forming
functional neuromuscular junctions remains a difficult task despite progress in recent years
[Keynes et al., 2011]. Electric stimuli using electrodes attached to skeletal muscles has been a
standard technique [Nagamine et al., 2011]. However, electric stimulation is limited in
performance due to several drawbacks. First, it is hard to secure electrodes to muscles that are
contracting and are thus not stationary. Second, the spatial resolution of electrical stimulation is
coarse since in wet environment it is hard to insulate a single muscle strip from others and, third,
electrical stimulation is invasive, leading to rapid degradation of contraction.
Optogenetics is an emergent technology that genetically codes an excitable cell so that it
becomes sensitive to an optical stimulus [Zhang et al., 2006]. Recently, we have applied this
optogenetic control technology to 3D engineered skeletal muscles microtissues [Sakar et al.,
2012]. Simply projecting a light beam on a muscle strip causes contraction of the muscle. No
mechanical contact is involved; it allows for wireless control of each muscle strip. The temporal
as well as spatial resolution of optical stimuli is significantly higher than electric stimulation,
achieving millisecond-order temporal precision and 10 pm of spatial resolution. It is also less
invasive than electric stimulation.
This optogenetic control of skeletal muscles opens up the possibility to activate hundreds
of individual muscles in a manner that is spatiotemporally coordinated. This chapter aims to
explore novel actuation technology using in vitro skeletal muscles and their optogenetic control.
Particularly unique is that live skeletal muscle actuators can activate degrees of freedom having
no fixed axis. Examples of such systems include shoulder joints, the jaw, and the mouth. Fixedaxis-joints are conveniently activated by traditional actuators such as electromagnetic motors.
However, those floating-axis-joints or floating loads are particularly difficult to activate. Live
skeletal muscle actuators can conform to those floating loads and create high DOF motion or
even distributed DOF motion in a compact body. This high DOF, distributed actuation
62
technology will be useful not only for android and humanoid face expression, but also for
creating and exploring entirely new robot applications.
4.2 Bio-ACTUATOR BUILDING BLOCKS
In order to construct a complex system of numerous degrees of freedom composed of
multiple actuator units, we first established and characterized the fundamental building block
unit. From this building block unit, numerous hnits may be combined in parallel and series
producing redundant degrees of freedom and floating nodes.
The building block we have developed is similar to the fascicle structural level of natural
hierarchical muscle. Muscle is primarily composed of bundles of muscle fascicles as shown in
Fig. 4.2. A muscle fascicle is a bundle of individual muscle cells that contract together. The
form factor of a fascicle is significantly influenced by diffusion transport of chemicals.
Essentially, it is an optimized building block that is robust against individual cell failure, and can
be combined in a parallel arrangement allowing for scalability.
We have developed a muscle cell culturing technique for forming a fascicle-like muscle
strip. See Fig. 4.3a. The self-assembling muscle strips can be an arbitrary length between 1 and
10 mm. Furthermore, we can vary nominal initial diameter between 250-500 pm. Maximal
diameter is limited by chemical diffusion of nutrients and waste products into and out of the
strip. Minimal diameter is limited by the size of contractile muscle cells within the strip. Based
on data from our previous work [Sakar et al., 2012], functional properties of the in vitro muscle
tissues are estimated as passive force of 11 pN and optogenetically activated force 13 pN for a
-100 pm diameter muscle strip.
Additionally we have control over the reagents that constitute the fibrin hydrogel scaffold
used to hold the precursor muscle cells, myoblasts. These reagents include fibrinogen, thrombin,
and Matrigel.
The forces generated and stiffness of individual strips are tested using a custom-built,
micro-force displacement apparatus. It is valuable to note that upon contraction, muscle stiffness
increases and is thus overall actuator stiffness may be controllable.
Beyond producing the building block actuator unit, our next objective was to produce two
units in parallel (Fig. 4.3b) and two units in series (Fig. 4.3c). Note that the rectangular block in
63
the middle is a floating node to which two muscle strip actuators are connected. These initial
steps are vital to producing massively parallel and highly serial/networked units.
4.3 SCALING FROM Two UNIT TO MULTI-UNIT SYSTEMS
The next important milestone in producing a larger scale system is to combine parallel
and series configurations into a single system.
Figure 4.4 illustrates a simple parallel and series arrangement: two parallel muscle strips
in series with another two parallel muscle strip. This configuration combines major underlying
techniques of constructing a modular actuator and is sufficient proof that these techniques can be
combined. Thus a system of arbitrary scale can be produced where further scaling is simply a
matter of incorporating additional building block actuators in parallel or series.
Recall that each muscle strip is individually addressable with optogenetic stimulation. In
the 2D system shown in Fig. 4.4a there are 4 controllable inputs to the system. The central
floating node has 2 controlled degrees of freedom; one translational, and one rotational. When 2
parallel muscle strips are contracted, the floating node translates as shown in Fig. 4.4b. When a
pair of muscle strips located at diagonal positions is contracted, the floating node rotates as
shown in Fig. 4.4c.
Note that the degrees of freedom of the floating node within the plane are three, while the
number of individually addressable actuators is four. This redundancy yields robustness against
individual unit failure. If a single unit fails, as shown in Fig. 4.4e, the control is still maintained
over the 2 degrees of freedom of the floating node. Additionally, the nonlinear stiffness
characteristics of muscle allow the redundant actuators to potentially be used to control stiffness.
As an example of this, consider when all 4 muscle units are activated as shown in Fig. 4.4d. The
activated muscles are stiffer than when relaxed, thus increasing the stiffness of both the
translational and rotational degrees of freedom of the floating node.
The above example is the simplest case of parallel and serial arrangement. This can be
scaled-up to massively parallel arrangements as well as to massively serial/networked system. In
a massively parallel system, there are highly redundant degrees of freedom and the system is
much more robust to individual unit failure. A massively parallel system is also the method with
which to produce large net forces from the relatively small forces produced by the individual
64
muscle units. Conversely, in a massively networked system, there are numerous independent
degrees of freedom, which will allow us to create a type of distributed motion.
4.3.1 MASSIVELY PARALLEL SYSTEMS
A massively parallel system is composed of many individual units acting together to
apply force to a single point. Skeletal muscle examples from the human body include the
pectoralis major (chest muscle), or the deltoid (shoulder muscle). Both of these are composed of
a great number of individual fascicles in bundled structures that originate from different parts of
the skeletal frame, and converge on a single area. In both of these examples, a small number of
degrees of freedom are controlled, but the strength is tremendous due to their parallel
architecture.
To produce a massively parallel system, we can expand upon our initial multi-unit system
in a series of logical steps outlined in Fig. 4.5. First we can scale up the number of individual
units in parallel as shown in Fig. 4.5a. This step is primarily what increases the redundancy and
force producing capabilities.
The next step in producing large force is to direct the force
production along a concentrated axis. To do this, the individual muscle units are arranged at an
angle in what is called a bipennate configuration as shown in Fig. 4.5b. In this configuration, the
output node is connected to a load, and when the muscle strips contract, their combined force is
directed toward pulling the output node. Further increase in scale is achievable by applying this
bipennate structure recursively to produce a multi-pennate system as shown in Fig. 4.5c. This is
similar to the structure found in the deltoid (shoulder) muscle. This configuration consists of
numerous anchor points converging on a single anchor point.
With the use of optogenetic control, individual muscle strips may be addressed in an ONOFF manner. Because of the plurality of building block actuators applying force redundantly,
the total applied force may be finely discretized to facilitate a smoother, near continuous, range
of applicable force.
4.3.2 HIGHLY SERIAL/NETWORKED SYSTEM
A massively serial or networked system is composed of many individual units acting
independently to apply force to numerous floating nodes. A muscle example from the human
body is the mouth and lips. This structure contains numerous interconnected nodes that have
numerous degrees of freedom as a system. The mouth is a highly controllable orifice with
65
functions including a controllable nozzle, a fine manipulator, and effective social communication
tool. In this example a large number of degrees of freedom are controlled through utilizing
numerous floating nodes interconnected via muscle tissue.
To produce a massively serial/networked system, we can expand upon our initial multiunit system in a series of logical steps outlined in Fig. 4.6. First, we scale up the number of
floating nodes as shown in Fig. 4.6a. This step achieves the goal of having numerous serial
nodes. The next step is to produce nodes connected beyond a series of connections, but as a
network. To do this, the individual nodes must be connected in a non-collinear fashion as shown
in Fig. 4.6b.
In this configuration, an additional translational degree of freedom is made
controllable for each floating node in the network. Further increase in scale produces a highly
networked system of floating nodes. This finer level of discretization leads to a smoother, more
continuous muscle powered network as shown in Fig. 4.5c. This is similar to the structure found
in the human mouth. To control one's mouth a plurality of nodes are interconnected with each
other and to grounding points on the skull.
With the use of optogenetic control, controlled individual muscle strips grant control of
each degree of freedom. This is in contrast to the parasitic electric field generation from
electrical stimulation that may limit the specificity of control of individual units in a wet
environment.
4.4 EXPERIMENTAL MATERIALS AND METHODS
C2C12 mouse myoblasts (Ameican Type Culture Collection) are cultured in growth
medium (GM) containing DMEM (American Type Culture Collection), supplemented with 10%
fetal bovine serum (FBS, Sigma-Aldrich), 1% penicillin-streptomycin IOOX (Invitrogen), and 0.1
mg/ml Normacin (Inv'ivogen). Confluence was kept below 70% and cells were seeded into the
experimental device at the 5th passage.
Cell/gel suspension is molded such that a uniform cylinder of controlled dimensions is
suspended in medium anchored at both ends to the walls of a well. The cell/gel suspension
contains GM with 20% Matrigel (BD Biociences), with C2C12 cells at a concentration of 15e6
cells/ml, fibrinogen (Sigma-Aldrich) at a concentration of 5 mg/ml, and 6-aminocaproic acid
(Sigma-Aldrich) at a concentration of 0.5 mg/ml. The initial fluid surrounding the construct
contains thrombin (Sigma-Aldrich) at a concentration of 5 U/ml. These components become the
66
hydrogel that acts as a scaffold to hold the developing muscle cells. The wells are 1 to 6 mm in
length, with walls made of polydimethylsiloxane (PDMS, Dow Corning) plasma bonded to a
glass coverslip (VWR) which constitutes the bottom of the well. The length of the molded gels
are nominally 1-6 mm, and the initial nominal diameters of the molded gels range from 250 pm
to 500 gm.
After 24 hours, the initial fluid is changed to GM supplemented with 0.5 mg/ml 6aminocaproic acid.
This medium is replaced daily for 2 more days, then switched to
differentiation medium, DM, containing DMEM supplemented with 10% horse serum (SigmaAldrich) 1% pen-strep, 0.1 mg/ml Normacin, and 0.5 mg/ml 6-aminocaproic acid.
DM is
changed daily until the termination of the experiment.
Optical stimulation is accomplished using one of two methods depending upon the degree
of spatial specificity required. Both methods utilize an IX81 Olympus inverted microscope. For
less specific, broadcasts illumination, light from a mercury lamp is passed through a blue filter
and controlled with a mechanical shutter. For greater spatial control, an Epson PowerLite 1965
LCD projector is used as the light source granting XGA resolution to the microscope field of
view.
4.5 EXPERIMENTAL RESULTS
4.5.1 SINGLE BUILDING-BLOCK MUSCLE ACTUATOR
Within a single muscle strip, multinucleated contractile muscle cells form in parallel with
others suspended in fibrin gel. Each muscle cell extends ~1 mm in length along the axis of the
strip. These cells populate the full length of the fascicle-like muscle construct and have generated
4.3% strain.
4.5.1 SERIAL AND PARALLEL PROTOTYPES
Multiple building block actuators have been combined in series and parallel. Parallel
muscle strips behave identically to individual muscle strips. Our initial prototype demonstrates
that parallel arrangements diagramed in Fig. 4.3b are possible as shown in Fig. 4.8a.
A serial connection between multiple muscle strips as diagramed in Fig. 4.3c is
demonstrated in functional prototypes such as the one shown in Fig. 4.8b. In this figure, muscle
actuators are anchored to a PDMS block that functions as a floating series node. Fibrin gel and
67
non-functional cells occupy and anchor to the hollow cylinder in the center of the PDMS block.
Light stimulation of the left or right muscle actuators moves the floating series node left or right,
respectively.
4.5.2 MULTI-UNIT ACTUATOR SYSTEM PROTOTYPES
We have produced functional serial/parallel systems using our building block actuators as
components as shown in Fig. 4.9. The unforced system in Fig. 4.9a contains four muscle
actuators anchored two hollow cylinders in a transparent PDMS block.
The un-actuated young's modulus has been measured using our custom-built force probe
to be 210 kPa as shown in Fig. 4.9b. In the loaded configuration, the geometric structure is
bipennate further showing the potential concentration of generated forces.
4.6 CONCLUSION
We have generated functional prototypes demonstrating the first few initial steps
necessary for producing large scale muscle actuators based on optimized building block muscle
units. Our basic units produce force and displacement upon optical stimulation. These quantities
along with material stiffness have been measured using custom force sensing equipment.
These bio-actuators allow us to build unique distributed actuator system having no fixed
axis; they are floating. This type of system is not easily produced using traditional actuators and
arises because muscle is both an actuator and a structural material.
We are advancing our characterization and control techniques to keep up with the
development of our actuator systems. We are developing a feedback controlled light stimulation
system to track complex actuators and maintain tight control of their numerous degrees of
freedom.
Future muscle actuator systems may be adapted to specific applications, thus potentially
utilizing different muscle strip geometries.
Additionally, with more advanced stimulation
methods such as light stimulus delivery coming from multiple angles or embedded within the
system, we will be able to produce more complex 3D actuator systems.
68
4.7 FIGURES AND CAPTIONS
Fig. 4.1 1 Optogenetic control of skeletal muscles [1]. Illumination of a skeletal muscle strip with
blue light causes contraction of the muscle with high spatiotemporal resolution.
Fig. 4.2 1 Skeletal muscle structure. Skeletal muscle contains many levels of hierarchical structure.
The fascicle structure is what our building block muscle construct is inspired by.
69
Fig. 4.3 1 Fascicle-inspired building block muscle strip actuators (a). Basic scaling techniques
include combining strips in (b) parallel and (c) series.
70
Fig. 4.4 | The combination of parallel and serial connections of muscle strips (a).
Such
arrangement with light control enables translational (b), rotational (c), and stiffness (d) control, as well as
robustness to failure (e).
71
Fig. 4.5 1 Numerous parallel muscle strips on both sides of a floating node (a).
Force
generation is directed and concentrated by orienting parallel strips in bipennate (b) and multi-pennate (c)
arrangements.
72
(c) Highly Networked Mouth-ike Muscle System
Fig. 4.6
I Increasing serial
connections (a). Demonstration of networked non-fixed axis nodes (b).
A mouth-inspired highly networked system of floating loads controlled by individually addressable
muscle strips (c).
73
Fig. 4.7 1 Individual building block actuator contracting with 4.3% measured strain.
Vertical lines are constant horizaontal position across both images. Ovals are identifiable marks within
the tissue, that translate upon contraction. Scale bar is 50 pm.
(a) Prototype parallel muscle actuators
Fig. 4.8
Prototype series muscle actuators
I Prototype muscle actuator systems consisting of fascicle-like muscle actuators in
parallel (a), and series (b). Scale bar is 350 pm.
74
(b) Loaded bipennate system
Fig. 4.9 1 Prototype muscle parallel and series actuator system consisting of four fascicle-like
muscle actuators and a PDMS node. Force displaces the central node downwards generating bipennate
geometry (b). Scale bar is 350 pm.
75
CHAPTER 5: CONCLUSION AND FUTURE DIRECTIONS
5.1 CONCLUSION
The sacrificial outer molding method I have developed allows us to produce fascicle-like
3D constructs consisting of densely populated, aligned cells with a length to diameter ratio
(~100:1). This was made possible by exploiting three major features of the method: 1). The outer
mold is sacrificial, and releases hydrogel crosslinking agents when melting, resulting in a
scaffold simply suspended between two anchoring ends, with no solid substrate; 2). Cellmediated compaction of the hydrogel creates uniform internal tension along the entire construct,
which aligns the cells and promotes maturation; and 3) The sacrificial outer mold with a tunable
diameter provides the hydrogel with an initial geometric constraint and allows the 3D construct
to be exposed to a surrounding media, facilitating perfusion across the entire cross section of the
construct. The method is flexible and expandable to a broad class of cell types, providing a
unique approach to formation of fascicle-like constructs.
The mechanical characterization system and method has allowed for the performance
assessment of developed muscle tissue. Valuable features of the system include 1) strain
generation of up to 40%, 2) optical and electrical stimulation of varying light intensity and
electric field strength, and 3) quick and easy assessment of work performed on various loads by
changing probe cantilevers. Using this system I have shown that there exists an optimal diameter
at which to grow fascicle-like muscle tissue constructs in order to maximize energy transfer to an
external load, and that optical and electrical stimulation results in similar performance.
5.2 FUTURE DIRECTIONS
The tissue engineering platform will continue to be developed both for skeletal muscle
tissue, as well as for novel hydrogel scaffolds and other cell types. Early work has already begun
on expanding applications for bundling multiple strips to do mechanical work (as described in
Chapter 3).
New directions currently being pursued include the following:
* Bundling of muscle tissue strips (Fig. 5.la)
& The formation of cardiomyocyte containing contractile strips (Fig. 5.1b)
76
* Studying the effects of prolonged, and intermittent stress on skeletal muscle
development (Fig. 5.lc)
" Using synthetic hydrogels such as polyethylene glycol (PEG) as the scaffolding
material for muscle strips (Fig. 5.2)
This tissue engineering platform is currently used to study deeper scientific phenomenon
pertaining to the differentiation and development of muscle tissue. It has potential to be used as
a test bed for pharmaceuticals due to simple performance assessment. With success of initial
bundling attempts, larger scale tissue bundles show potential for implantation into patients in
need of muscle tissue. Finally, further development of the tissue as an actuator may eventually
lead to in vitro grown muscle's use as the mechanical power source in mechanical systems.
5.3 ACKNOWLEDGEMENTS
I'd like to thank the contributions to this chapter by Hyeon-Yu Kim, Dr. Vincent Chan,
and Apoorva Kalasuramath.
77
5.4 FIGURES AND CAPTIONS
A
Bundling
B
.................
...................... . . . . .....
Fig. 5.1 I Future Directions. (A) Fascicle-like strips grown in parallel and then bundled together. (B)
Cardiac cells grown in fascicle-like strips. (C) Strain control during development of muscle strips.
78
Fig. 5.2 | Initial PEG Fascicle-Like Strips. Initial work using PEG hydrogels consisted of using
concentrations of PEG and fibrin as scaffolding material such as 100% PEG (A), and 80% PEG and 20%
5 mg/ml fibrin (B).
79
REFERENCES
Alford PW, AW Feinberg, SP Sheehy, and KK Parker, "Biohybrid thin films for
measuring contractility in engineered cardiovascular muscle," Biomaterials,vol. 31, pp. 36133621, 2010.
Beier JP, Stern-Straeter J, Foerster VT, Kneser U, Stark GB, Bach AD, "Tissue
engineering of injectable muscle: three-dimensional myoblast-fibrin injection in the syngeneic
rat animal model," PlastReconstr Surg., vol. 118(5), pp. 1113-21, 2006.
Bellan LM, M Pearsall, DM Cropek, and R Langer, "A 3D Interconnected Microchannel
Network Formed in Gelatin by Sacrificial Shellac Microfibers," Advanced Materials, vol. 24, pp.
5187-5191, 2012.
Bian W, and N Bursac, "Engineered skeletal muscle tissue networks with controllable
architecture", Biomaterials,vol. 30, pp. 1401-1412, 2009a.
Bian W, B Liau, N Badie and N Bursac, "Mesoscopic hydrogel molding to control the 3D
geometry of bioartificial muscle tissue," Nat Protocols, vol. 4, pp. 1522-1534, 2009b.
Bian W, M Juhas, TW Pfeiler ,and N Bursac, "Local tissue geometry determines
contractile force generation of engineered muscle networks," Tissue Eng: PartA, vol. 18, pp.
957-967, 2012.
Boudou T, Legant WR, Mu A, Borochin MA, Thavandiran N, Radisic M, et al., "A
microfabricated platform to measure and manipulate the mechanics of engineered cardiac
microtissues," Tissue EngineeringPartA, vol. 18, pp. 910-919, 2012.
Brooks SV, JA Faulkner, "Forces and Powers of Slow and Fast Skeletal Muscles in Mice
During Repeated Contractions," Journal of Physiology, vol. 436, pp. 701-710, 1991.
80
Brown AEX, RI Litvinov, DE Discher, PK Purohit, and JW Weisel, "Multiscale
mechanics of fibrin polymer: gel stretching with protein unfolding and loss of water," Science,
vol. 325, pp. 741-744, 2009.
Chan V, P Zorlutuna, JH Jeong, H Kong and R Bashir, "Three-dimensional
photopatterning of hydrogels using stereolithography for long-term cell encapsulation," Lab on a
Chip, vol. 10, no. 16, pp. 2062-2070, 2010.
Chan V, Park K, Collens M, Kong H, Saif TA, Bashir R, "Development of Miniaturized
Walking Biological Machines," Scientific Reports, vol. 2 : 857, 2012.
Chan V, Jeong JH, Bajaj P, Collens M, Saif T, Kong H, et al., "Multi-material biofabrication of hydrogel cantilevers and actuators with stereolithography," Lab Chip, vol. 12, pp.
88-98, 2012.
Chan V, Asasda HH, Bashir R, "Utilization and control of bioactuators across multiple
length scales," Lab on a Chip, vol. 14, pp. 653-670, 2014.
Chew SY, R Mi, A Hoke, and KW Leong, "The effect of the alignment of electrospun
fibrous scaffolds on Schwann cell maturation," Biomaterials, vol. 29, pp. 653-661, 2008.
Courdier-Fruh I, Barman L, Briguet A, Meier T, "Glucocorticoid-mediated regulation of
utrophin levels in human muscle fibers." Neuromuscul Disorders, vol. 12 (Suppl 1), pp. S95104,2002.
Cross-Doersen D, Isfort RJ, "A novel cell-based system for evaluating skeletal muscle
cell hypertrophy-inducing agents," In Vitro Cell Dev Biol Anim, vol. 39, pp. 407- 412,2003.
Curtis A, Wilkinson C, "Topographical control of cells," Biomaterials, vol. 18, pp. 15731583, 1997.
81
Mammalian Skeletal Muscle Constructs Engineered In Vitro," In Vitro Cellular
&
Dennis RG, and PE Kosnik, "Excitability and Isometric Contractile Properties of
Developmental Biology, vol. 36, pp 327-335, 2000.
Dennis RG, Kosnik PE II, Gilbert ME, Faulkner JA, "Excitability and contractility of
skeletal muscle engineered from primary cultures and cell lines," Am. J. Physiol. Cell Physiol.,
vol. 280, pp. C288-C295, 2001.
Dennerll TJ, HC Joshi, VL Steel, RE Buxbaum, and SR Heidemann, "Tension and
Compression in the Cytoskeleton of PC-12 Neurites II: Quantitative Measurements," Journalof
Cell Biology, vol. 107, pp. 665-74, 1988.
Diaz-Ramos
A,
A Roig-Borrellas, A Garcia-Melero, A Llorens, and R L6pez-Alemany,
"Requirement of Plasminogen Binding to Its Cell-Surface Receptor a-Enolase for Efficient
Regeneration of Normal and Dystrophic Skeletal Muscle," PLos ONE, vol. 7, no. 12, 2012.
Dickinson MH, Farley CT, Full RJ, Koehl MAR, Kram R, Lehman S, "How animals
move: an integrative view," Science, vol. 288, pp. 100-106, 2010.
Duong H, Wu B, Tawil B, "Modulation of 3D Fibrin Matrix Stiffness by Intrinsic
Fibrinogen-Thrombin Compositions and by extrinsic Cellular Activity," Tissue Eng PartA, vol.
15, pp. 1865-1875, 2009.
Feinberg AW, Feigel A, Shevkopyas SS, Sheehy S, Whitesides GM, Parker KK,
"Muscular Thin Films for Building Actuators and Powering Devices," Science, vol. 317, pp.
1366-1370, 2007.
Fenno L, 0 Yizhar and K Deisseroth, "The development and application of
optogenetics," Annu Rev Neurosci, vol. 34, pp. 389-412, 2011.
82
Golden AP, and J Tien, "Fabrication of microfluidic hydrogels using molded gelatin as a
sacrificial element," Lab on a Chip, vol. 7, pp. 720-725, 2007.
Gordon AM, AF Huxley, and FJ Julian, "The Variation in Isometric Tension With
Sarcomere Length in Vertebrate Muscle Fibres," The Journal of Physiology, vol. 184, pp. 170192, 1966.
Granchelli JA, Pollina C, Hudecki MS, "Pre-clinical screening of drugs using the mdx
mouse," Neuromuscul Disorders,vol. 10, pp 235-239, 2000.
Grosberg A, PW Alford, ML McCain, and KK Parker, "Ensembles of engineered cardiac
tissues for physiological and pharmacological study: Heart on a chip," Lab Chip, vol. 11, pp.
4165-4173, 2011.
Herr H, Dennis RG, "A swimming robot actuated by living muscle tissue," Journal of
NeuroEngineeringandRehabilitation,vol 1:6, 2004.
Hinds S, W Bian, RG Dennis, and N Bursac, "The role of extracellular matrix
composition in structure and function of bioengineered skeletal muscle," Biomaterials, vol. 32,
pp. 3575-3583, 2011.
Hoffman EP, Escolar D, "Translating mighty mice into neuromuscular therapeutics: is
bigger muscle better?" Am JPathol,vol. 168, pp. 1775-1778, 2006.
Huang YC, RG Dennis, L Larkin, and K Baar, "Rapid formation of functional muscle in
vitro using fibrin gels," JournalofApplied Physiology, vol. 98, pp. 706-713, 2005.
Hunter IW, Lafontaine S, "A comparison of muscle with artificial actuators," Solid-State
Sensor and Actuator Workshop, pp. 178-185, 1992.
83
Hwang CM, Y Park, JY Park, K Lee, K Sun, A Khademhosseini, and SH Lee,
"Controlled cellular orientation on PLGA microfibers with defined diameters," Biomedical
microdevices, vol. 11, pp. 739-746, 2009.
Jang KJ, MS Kim, D Feltrin, NL Jeon, K-Y Suh, and 0 Pertz, "Two distinct filopodia
populations at the growth cone allow to sense nanotopographical extracellular matrix cues to
guide neurite outgrowth," PLoS One, vol. 5, p. e15966, 2010.
Kang E, YY Choi, SK Chae, JH Moon, JY Chang, and SH Lee, "Microfluidic Spinning
of Flat Alginate Fibers with Grooves for Cell-Aligning Scaffolds," Advanced Materials, vol. 24,
pp. 4271-4277, 2012.
Keynes R, D Aidley, and C Huang, Nerve and Muscle, Cambrige University Press, NY,
2011.
Kosnik PE, Faulkner JA, Dennis RG, "Functional development of engineered skeletal
muscle from adult and neonatal rats," Tissue Eng. vol. 7, pp. 573-5 84, 2001.
Kim J, Park J, Yang S.,Baek J, Kim B, Lee SH, et al., "Establishment of a fabrication
method for a long-term actuated hybrid cell robot," Lab Chip, vol. 7, pp. 1504-1508, 2007.
Kim E, OV Kim, KR Machlus, X Liu, T Kupaev, J Lioi, AS Wolberg, DZ Chen, ED
Rosen, Z Xu, and M Alber, "Correlation between fibrin network structure and mechanical
properties: an experimental and computational analysis," Soft Matter, vol. 7, pp. 4983-4992,
2011.
King KR, CCJ Wang, MR Kaazempur-Mofrad, JP Vacanti, and JT Borenstein,
"Biodegradable microfluidics," Advanced Materials,vol. 16, pp. 2007-2012,2004.
Kuang S, Gillespie M, Rudnicki M, "Niche Regulation of Muscle Satallite Cell SelfRenewal and Differentiation," Cell Stem Cell, vol. 2, pp 22-31,2008.
84
Lam MT, Sim S, Zhu X, Takayama S, "The effects of continuous wavy micropatterns on
silicone substrates on the alignment of skeletal muscle myoblasts and myotubes," Biomaterials,
vol. 27, pp. 4340-4347, 2006.
Lam MT, Huang YC, Birla RK, Takayama S, "Microfeature guided skeletal muscle
tissue engineering for highly organized 3-dimensional free-standing constructs," Biomaterials,
vol. 30, pp. 1150-1155, 2009
Larkin LM, Ven Der Meulen JH, Dennis RG, Kennedy JB, "Functional evaluation of
nerve-skeletal muscle constructs engineered in vitro," In Vitro Cell. Dev. Biol. - Animal, vol. 42,
pp. 75-82, 2006.
Legant WR, A Pathak, MT Yang, VS Deshpande, RM McMeeking, and CS Chen,
"Microfabricated tissue gauges to measure and manipulate forces from 3D microtissues," Proc
NatlAcadSci USA, vol. 106, pp. 10097-10102,2009.
Lieber RL, "Skeletal Skeletal Muscle Structure, Function, and Plasticity: The
Physiological Basis of Rehabilitation, Third Edition," Lippincott Williams & Williams, 2010.
Muraoka M, T Shimizu, K Itoga, H Takahashi, and T Okano, "Control of the formation
of vascular networks in 3D tissue engineered constructs," Biomaterials, vol. 34, no. 3, pp. 696703,2010.
Nagamine K, Kawashima T, Sekine S, Ido Y, Kanzaki M, Nishizawa M,
"Spatiotemporally controlled contraction of micropatterned skeletal muscle cells on a hydrogel
sheet," Lab Chip, vol. 11, pp. 513-517, 2011.
Neal D, HH Asada, "Co-fabrication of live skeletal muscles as actuators in a millimeter
scale mechanical system," Proceedings of IEEE International Conference on Robotics and
Automation, pp. 3251-6, 2011
85
Okano T, S Satoh, T Oka, and T Matsuda, "Tissue engineering of skeletal muscle. Highly
dense, highly oriented hybrid muscular tissues biomimicking native tissues," Asaio J, vol. 43, pp.
M749-53, 1997.
Onoe H, T Okitsu, A Itou, M Kato-Negishi, R Gojo, D Kiriya, K Sato, S Miura, S
Iwanaga, K Kuribayashi-Shigetomi, YT Matsunaga, Y Shimoyama, and S Takeuchi, "Metrelong cell-laden microfibres exhibit tissue morphologies and functions," Nature Materials, vol.
12, pp. 584-590, 2013.
Otsu, N., "A Threshold Selection Method from Gray-Level Histograms," IEEE
Transactionson Systems, Man, and Cybernetics, vol. 9, pp. 62-66, 1979.
Paul AC, "Muscle length affects the architecture and pattern of innervation differently in
leg muscles of mouse, guinea pig, and rabbit compared to those of human and monkey muscles,"
The Anatomical Record, vol. 262, no. 3, pp. 301-309, 2001.
Park J, et al., "Biohybrid microsystems actuated by cardiomyocytes: microcantilever,
microrobot, and micropump." Proceedings of IEEE International Confernce on Robotics and
Automation, pp. 880-885, 2008.
Partap S, I Rehman, JR Jones, and JA Darr, ""Supercritical Carbon Dioxide in Water"
Emulsion-Templated Synthesis of Porous Calcium Alginate Hydrogels," Advanced Materials,
vol. 18, pp. 501-504, 2006.
Pilarek M, Neubauer P, Marx U, "Biological cardio-micro-pumps for microbioreactors
and analytical micro-systems," Sens. Actuators B, vol. 156, pp. 517-526, 2011.
Sakar MS, D Neal, T Boudou, MA Borochin, Y Li, R Weiss, RD Kamm, CS Chen and
HH Asada, "Formation and optogenetic control of engineered 3D skeletal muscle bioactuators,"
Lab on a Chip, vol. 12, no. 23, pp. 4976-4985, 2012.
86
Sander EA, VH Barocas, and RT Tranquillo, "Initial fiber alignment pattern alters
extracellular matrix synthesis in fibroblast-populated fibrin gel cruciforms and correlates with
predicted tension," Annals of biomedical engineering, vol. 39, pp. 714-729, 2011.
Schumacher KM, SC Phua, and JY Ying, "Controlled formation of biological tubule
systems in extracellular matrix gels in vitro," Kidney international,vol. 73, pp. 1187-1192, 2008.
Semsarian C, Wu MJ, Ju YK, Marciniec T, Yeoh T, Allen DG, et al., "Skeletal muscle
hypertrophy is mediated by a Ca2+-dependent calcineurin signalling pathway," Nature, vol. 400,
pp. 576-581, 1999.
Shansky J, Vandenburgh H, "A simplified method for tissue engineering of skeletal
muscle organoids in vitro," In Vitro Cell. Dev. Biol., vol. 33, pp. 659-661, 1997.
Strohman RC, Bayne E, Spectra D, Obinata T, Micou-Eastwood J, Maniotis A,
"Myogenesis and histogenesis of skeletal muscle on flexible membranes in vitro," In Vitro Cell.
Dev. Biol, vol.26A, pp. 201-208, 1990.
Tanaka Y, Sato L,Shimizu T, Yamato M, Okano T, Kitamori T, "A microspherical heart
pump powered by cultured cardiomyocytes," Lab Chip, vol. 7, pp. 207-212, 2007.
Takahashi H, T Shimizu, M Nakayama, M Yamato, T Okano, "The use of anisotropic
cell sheets to control orientation during the self-organization of 3D muscle tissue," Biomaterials,
vol. 34, pp. 7372-7380, 2013.
Tsang VL, Chen AA, Cho LM, Jadin KD, Sah RL, DeLong S, West JL, and Bhatia SN,
"Fabrication of 3D hepatic tissues by additive photopatterning of cellular hydrogels," The
FASEB Journal,vol. 21 pp. 790-801, 2007.
87
Vandenburgh HH, Swasdison S, Karlisch P, "Coputer-aided mechanogenesis of skeletal
muscle organs from single cells in vitro," FASEB J., vol. 13, pp. 2860-2867, 1991.
Vandenburgh H, J. Shansky, F Benesch-Lee, V Barbata, J Reid, and L Thorrez, et al.,
"Drug-screening platform based on the contractility of tissue-engineered muscle," Muscle Nerve,
vol. 37, pp. 438-447, 2008.
Vandenburgh H, "High-content drug screening with engineered musculoskeletal tissues,"
Tissue Eng: PartB, vol. 16, pp. 55-64, 2010.
Wagner S, Dorchies OM, Stoeckel H, Wafter JM, Poindron P, Takeda K, "Functional
maturation of nicotinic acetylcholine receptors as an indicator of murine muscular differentiation
in a new nerve-muscle co-culture system." PflugersArch., vol. 447, pp. 14-22, 2003.
Wang K, McCarter R, Wright J, Beverly J, Ramirez-Mitchell R, Viscoelasticity of the
sarcomere matrix of skeletal muscles. The titin-myosin composite filament is a dual-stage
molecular spring." Biophys J., vol. 64, pp. 1161-1177, 1993.
Wang HB, ME Mullins, JM Cregg, A Hurtado, M Oudega, MT, and Trombley and RJ
Gilbert, "Creation of highly aligned electrospun poly-L-lactic acid fibers for nerve regeneration
applications," Journalof neuralengineering, vol. 6, p. 016001, 2009.
Weisel JW, "Fibrinogen and fibrin," Adv. Protein Chem., vol. 70, pp. 247-299,2005.
Williams PL, Gray'sAnatomy, New York: Churchill Livingstone, 1995.
Xi J, Schmidt JJ, Montemagno CD, "Self assembled microdevices driven by muscle,"
Nature Materials,vol. 4, pp. 180-184, 2005.
88
Yang J, FR Rose, N Gadegaard, and MR Alexander, "A High-Throughput Assay of CellSurface Interactions using Topographical and Chemical Gradients," Advanced Materials, vol. 21,
pp. 300-304, 2009.
Zhang F, L-P Wang, ES Boyden, and K Deisseroth, "Channelrhodopsin-2 and optical
control of excitable cells," Nat Methods, vol. 3, pp. 785-792, 2006.
Zorlutuna P, N Annabi, G Camci-Unal, M Nikkhah, JM Cha, JW Nichol, A Manbachi, H
Bae, S Chen, and A Khademhosseini, "Microfabricated biomaterials for engineering 3D tissues,"
Advanced Materials, vol. 24, pp. 1782-1804, 2012.
APPENDIX A: FASCICLE CONSTRUCT PROTOCOL
Materials
Reagents
PDMS Mold fabrication
" PDMS monomer
" PDMS curing agent
Hydrogel and cell culture
"
"
"
"
*
*
"
"
"
NaOH
Thrombin
Fibrinogen
DMEM
Fetal Bovine Serum (FBS)
Gelatin
Penicillin-streptomycin
Aminocaproic acid
trypsin
Equipment
PDMS mold fabrication
" Aluminum or 3D printed ,old
* Steel wire (0.014", 0.02", or 0.031")
89
Hydrogel and cell culture
0
9
a
0
9
*
1.5 ml tubes
Water bath
BSC
1 ml pipette and tips
Tweezers
Refrigerator
Reagent Setup
* Growth Medium Plus (GM+): DMEM from ATCC, with 10% fetal bovine serum, 1 mg
ml 1 aminocaproic acid, 100 U ml' penicillin, 100 pg ml1 streptomycin.
Procedure
Aluminum mold manufacturing
e Not included
Casting PDMS molds
* Not included
Preparation of PDMS molds
1. Start with an aluminum mold with a central pocket -3 mm deep with a smooth sanded
bottom
2. Thread wire through holes in the aluminum pocket walls such that they span the pocket
a. The holes are 1.5 mm above the bottom of the pocket and 1.5 mm below the
surface of the pocket
3. Mix and de-gas PDMS (1:10 monomer:curing agent ratio)
4. Pour PDMS
a. For better imaging PDMS may be poured until the surface is just noticeably
covering the spanning wires, and using this top surface to bond to glass. (Must be
less than 0.5 mm for working distance with our 20X 0.75 NA microscope
objective)
5. Cure in 60-80* C oven for at least 4 hours
6. Remove wires leaving cavities in the PDMS slab
7. Remove PDMS slab from the aluminum mold
8. Cut PDMS slab into numerous devices
9. Use biopsy punch to punch holes (1 mm to 6 mm diameter)
a. Punch from top of device to bottom of device
b. Punch through the perpendicular channel(s) created by with wires
c. Punch three holes per device (central hole is for a medium reservoir)
90
10. Bond PDMS slab to glass via plasma bonding
11. Cut wires into two pieces
12. Re-insert both pieces of wire into the ends of the PDMS cavities. Two wires now span
all three punched holes and meet at the central hole
91
Punched Holes
Wires
Glass
Medium Reservoir PDMS
Hole
Top View
Seeding
Hole
Fig. A.1 Schematic of fascicle tissue device with pins
Casting sacrificial gelatin hydrogel
Reagent Checklist (for 18 devices, and T150 flask of cells)
*
*
*
*
*
*
*
*
Gelatin (3 40-65 mg tubes)
Thrombin solution 100 U/ml (36 pl)
NaOH solution 0.5 N (18 pl)
GM+ (GM + 1 mg/ml aminocaproic acid) (1800 pl for reservoir + 210 pl for gelatin
1000 pl for fibrinogen < 4 ml)
GM (18 ml for spinning)
Trypsin (6 ml)
PBS (24 ml for trypsinizing)
Fibrinogen (5 mg)
Cells (T150 less than 70% confluent)
Ice
Cell Counting Chips, trypan blue
Tip Jar
+
"
"
"
*
1. Add at least 10 mg Gelatin powder per device to 1.5 mL tubes (65 mg/tube maximum)
a. Weigh 1.5 mL centrifuge tube and record
b. Add gelatin in sterile BSC
c. Weigh again. Mark difference between measurements on tube.
2. Add GM+ to make 0.05 mg/pl gelatin in GM+ (X mg gelatin * 20 pl/mg GM+= 20 X
GM+)
a. Tap tube on side to scatter gelatin along the side of the tube to ensure the gelatin
powder is not compressed on the bottom
b. Add GM+ and quickly vortex until gelatin particles are well distributed (-5
seconds)
92
)
3. Place tube(s) in 37*C water bath until gelatin particles are completely dissolved (-5
minutes). Vortex and tap up-side down if gelatin is settling and solidifying on the bottom
of the tube.
4. Add 1% vol/vol NaOH solution to neutralize PH ((X GM+) (0.01)).
5. Add 1% thrombin solution ((X GM+) (0.01)
6. Vortex for 5 seconds and return to 37*C water bath until immediately before adding to
devices
7. Prepare the BSC for adding gelatin solution to the devices. The gelatin solution solidifies
very quickly at room temperature.
a. Position devices, maximum of 6 per gelatin solution tube.
b. Set lml pipette to at least 100 pl per device (150 p11 per device is safer) (900 P1 for
6 devices).
c. Position tweezers and pipette tips near the devices.
8. Add gelatin solution to experiment well
a. Try NOT to get gelatin solution foam on pipette tip when drawing gelatin solution
b. Draw -150 pl gelatin solution per device
c. Hold the device with tweezers to prevent it from moving while filling.
d. Place the tip of the pipette at the bottom of the experiment well, just underneath
steel pin to ensure no air pockets are left. (Important: do not allow air pockets
between pin and glass)
9. Place devices in 4*C refrigerator for at least 30 minutes before seeding cell suspension,
but not more than 4 hours or the gelatin solution will dry up.
Cavity left
by wire
Insert gelatin with
wire in place
Inject cell
suspension here
1
V
Glass
Seeding
Wells
Medium Reservoir
Hole
Top View
Fig. A.2 Schematic of fascicle tissue device with pins removed
Casting fibrin hydrogel
matrix (replace steps
1-7 above
with the following)
Optimal Fibrinogen Concentration: 5 mg/ml
10. Add at least 1 mg fibrinogen powder per device to 1.5 mL tubes (6.5 mg/tube maximum)
a. Weigh 1.5 mL centrifuge tube and record
b. Add Fibrinogen in sterile BSC
c. Weigh again. Mark difference between measurements on tube.
93
11. Add GM+ to make 5 mg/mi gelatin in GM+ (X mg fibrinogen * 200 pl/mg GM+ =20 X
GM+)
a. Tap tube on side to scatter fibrinogen along the side of the tube to ensure the
powder is not compressed on the bottom
b. Add warm GM+ and quickly just enough to mix vortex until gelatin particles are
well distributed (-1 seconds). Be careful not to over vortex
12. Place tube(s) in 37*C water bath until the fibrinogen is completely dissolved (-1
minutes). Vortex gently and tap up-side down if fibrinogen is settling and solidifying on
the bottom of the tube. Be careful to not heat in water bath for too long, or the contents
wil gel
13. Prepare the BSC for adding the fibrinogen solution to the devices. The solution solidifies
quickly at room temperature.
a. Position devices, maximum of 6 per fibrinogen solution tube.
b. Set lml pipette to at least 100 pl per device (150 pl per device is safer) (900 pl for
6 devices).
c. Position tweezers and pipette tips near the devices.
14. Add fibrinogen solution to experiment well
a. Try not to get solution foam on pipette tip when drawing solution
b. Draw full volume of tube into pipette tip
c. Hold the device with tweezers to prevent it from moving while filling.
d. Place the tip of the pipette at the bottom of the experiment well, just underneath
steel pin to ensure no air pockets are left.
15. Place devices in 37*C incubator for at least 30 minutes before seeding cell suspension,
but not more than 4 hours or the solution will dry up.
Seeding hydrogel cell suspension
Optimal Cell Concentration: 10 million cells/ml
1. Trypsinize cells, count, and spin
a. Equation for cell counting chips: # / squares x 2 x 10,000 x volume of CS
2. Add fibrinogen to 1.5 ml tubes
a. Weigh 1.5 ml centrifuge tube and record
b. Add fibrinogen in sterile BSC
c. Weigh again and mark difference between measurements on tube
3. Add medium to fibrinogen to make fibrinogen solution
a. Tap tube on side to scatter gelatin along the side of the tube to ensure the
fibrinogen powder is not compressed on the bottom
b. Depending on desired fibrinogen concentration (2-10 mg/ml), add appropriate
amount of GM+ to tube ((mg of F) / (X mg/ml) x (1000 pl/ml) pIl GM+)
c. Tap tube to ensure no air bubbles are in the tube
d. Place in 37 C water bath until fibrinogen is dissolved in fluid (not more than 2
minutes or contents will gel)
e. Put tube on ice. Solution will gel in about 30 minutes
4. Add appropriate amount of fibrinogen solution to pellet to get desired cell concentration
and mix with pipette until cell suspension is homogeneous ((desired e6 cells/ml) / (e6 cell
in pellet) x (1000 pl/ml )= pl fibrinogen solution)
94
5. Seed cell suspension into device (This is a two handed multistep operation)
a. Use 100 or 200 pI pipette and draw up 20 pl of cell suspension with pipette hand
b. Pick up device with device hand between index finger and thumb
c. Remove needle with pipette hand with single smooth motion
d. Position pipette tip at conduit entrance and pipetting cell suspension
e. Stop when you see fluid entering medium reservoir (be careful not to suck fluid
back out of conduit.
f. Repeat a-e for each device
6. Place seeded devices in incubator for 30 minutes, then fill medium reservoir with GM+
Remove sacrificial gelatin hydrogel and Continued Culture:
1. Change medium daily
a. Aspirate device from central media reservoir well
b. Add -200-250 pl to central well forming a small puddle spanning all 3 wells
c. Ensure the fascicle is constantly wet, and not strained by surface tension
2. After 1-3 days, switch from using GM+ to DM+, where the + denotes 1 mg/ml
aminocaproic acid.
95
96
APPENDIX B: OPTOGENETIC CONTROL OF LIVE SKELETAL MUSCLES:
NON-INVASIVE, WIRELESS, AND PRECISE ACTIVATION OF MUSCLE
TISSUES
Note: This chapter was previously published in the proceedings of the American Control
Conference (ACC) 2013, June 17-19, Washington, DC.
Optogenetics is an emerging new technology for controlling live cell function with light.
Skeletal muscles are genetically coded to express light-sensitive proteins so that the cell's
behavior may be altered by illuminating a targeted portion of the cell. Optogenetic control
provides a non-invasive, wireless, and fast control method with high spatiotemporal resolution.
This chapter presents the technology, experimental test, and potential applications of optogenetic
control of skeletal muscles. The authors' research team has recently succeeded in controlling the
contraction of skeletal muscle tissues with light by using the light-sensitive protein
Channelrhodopsin 2 (ChR2). Precursors of skeletal cells, myoblasts, are transfected with ChR2,
creating ion channels on the cell membrane that conduct cations when exposed to blue light.
Experiments show that targeted skeletal muscles are activated rapidly and precisely with high
spatiotemporal resolution. Dynamics of optogenetically controlled skeletal muscles are
characterized based on system identification. Two potential applications are addressed. One is
muscle-on-a-chip drug screening for evaluating potential hazards of a drug on muscular function,
and the other is multi-DOF robotic devices powered by bio-artificial muscles controlled with
targeted light illumination.
B. 1 INTRODUCITON
Bringing "control" to biological systems has been a challenge. Effective intervention and
manipulation of live cells and tissues must be developed for guiding them toward a desired state
or behaviors. Such interventions or controls must not deteriorate otherwise healthy, intact cells
and tissues, yet they must be effective in altering the biological functionality. Cells have their
own regulatory systems, which tend to attenuate the effect of exogenous input and intervention.
It is a challenge to develop effective technology to intervene in the existing regulatory systems
without breaking the necessary regulatory mechanisms. In the past, various techniques ranging
from biochemical cues to mechanical, electrical, and magnetic means have been developed.
97
Recent progress in microfluidic technology allows for precise delivery of chemical, mechanical,
and electrical stimuli [1,2]. However, the responses of the cells and tissues to such cues are often
highly variable and slow [3]. The lack of effective control means is a major bottleneck for
applying control theory and techniques.
Optogenetic control is an emerging new technology to alter cell behaviors by illuminating
a targeted portion of the cell [4,5]. See Fig. B.1. Membrane-bound light-sensitive proteins, such
as Channelrhodopsin 2, open or close ion channels so that surrounding ions, such as sodium and
potassium ions, can enter or exit the cell rapidly and precisely. Compared to standard
microfluidic technology where various input cues are transported via media and bound to the
cells and tissues through diffusion, optogenetic control allows for an order-of-magnitude faster
response, and spatially accurate control that alters cell behaviors more directly. Furthermore,
optogenetic control is known to be less invasive in specific application areas, including neural
tissue stimulation [6] and cardiac cell stimulation [7]. Optogenetic control opens up new
possibilities for bringing "control" to life sciences at the cell and tissue level.
This chapter aims to introduce optogenetic control to the control community, focusing on
the optical activation of skeletal muscles that the authors' research group has recently
accomplished [8]. Bio-artificial muscles cultured in an in vitro environment have the potential to
be effective actuators that convert chemical energy directly to mechanical energy. Skeletal
muscle can produce over 20% strain with over 200 kPa of stress; no actuator technology
currently exists to create such a large strain and generated force in such a compact body.
Unlike cardiac muscles, skeletal muscles need exogenous stimuli for activating
contraction. Traditional electrical stimulation needs to attach electrodes near or directly to
muscle tissue, which causes a number of problems. First, mechanical attachment to moving soft
tissues is difficult; a second, electrical stimulus generates parasitic electric fields causing poor
spatial resolution; and third, it is invasive and tends to deteriorate the contractile function. In
contrast, optogenetic control allows us to wirelessly communicate with the muscle, stimulate
with high spatiotemporal resolution, and the stimulation is reversible and non-invasive.
In the following, the basic principle of optogenetics using Channelrhodopsin 2 is
described, followed by brief explanation of myogenesis, the process of forming a muscle tissue.
Experimental results of optogenetic control of skeletal cells are then presented. Dynamics and
key properties of muscle contraction in response to light illumination are analyzed by using a
98
system identification technique. Finally potential applications of the optogenetic control of
muscles and their impact on life sciences and medicine as well as engineering systems will be
addressed.
B.2 MYOGENESIS AND OPTOGENETICS
Optogenetics consist of three major components [9]:
Development of light-sensitive proteins that make conformational changes when
exposed to a certain wavelength of light;
Techniques to deliver the genes to targeted cells so that the light-sensitive proteins are
-
expressed within the cell; and
Optical instrumentation for illuminating a targeted cell or a specific portion of the cell
-
with a light beam.
Several kinds of light-sensitive proteins have been discovered and applied to various fields of life
sciences.
Among others,
Channelrhodopsin
2
(ChR2),
discovered
in
a microbial
Chlamydomonas reinhardtii, provided neuroscientists with a powerful tool for stimulating
neuron cells at unprecedented spatiotemporal resolution [10]. Channelrhodopsins conduct cations
and depolarize neurons upon illumination. ChR2 was a breakthrough since it can perform the
two tasks of light sensation and ion flux induction in a single component encoded by a single
gene. ChR2 is membrane-bound, and serves as a cation flux gate activated by blue light of 470
nm wavelength. See Fig. B.2.
Delivering ChR2 into target cells is a critical process in order to make the cell lightsensitive. We have made skeletal cells light-sensitive using a non-viral strategy. We used a
plasmid containing ChR2 and other key components including the genetic promoter CAG, and
puromycin resistance, shown in Fig. B.3, and had skeletal muscle cells transfected with the
plasmid. Although the transfection rate was fairly low, we selected transfected cells using an
antibiotic, puromycin. We kept selecting transfected cells during multiple passages. Fig. B.4
shows transfected cells, where the green area indicates that the cell is transfected with ChR2.
After five iterations, the cells stably express ChR2.
A skeletal muscle is formed through a multi-stage process, referred to as myogenesis.
First, precursors of skeletal muscle cells, called myoblasts, are aligned. See Fig. B.5 (a) and (b).
They proliferate in a growth media. After the media is changed to a differentiation media,
99
aligned myoblasts begin to adhere to each other, and then remove their interfacing partition; by
resolving the membranes separating adjacent cells, they become a fused, multinucleated tube.
See Fig. B.5(c). The fused muscle strip, or myotube, further differentiates into a mature
functional muscle, as it creates sarcomeric striations. Fig. B.9 illustrates our experimental result,
where clear sarcomeric striations are observed. Tension is created in the longitudinal direction,
which promote the myotube to become a mature, functional cell. The muscle cell line we chose
is C2C12 mouse cells, which have been applied broadly to bio-artificial muscle development.
B.3 EXPERIMENTAL EVALUATION
Two types of light-sensitive skeletal muscle assays were developed. One is a 2dimensional muscle sheet, and the other is a 3-dimensional muscle tissue. Fig. B.6 shows a 2D
muscle sheet contracting in response to a blue light beam of 470 nm wavelength and 10
mW/mm 2 of intensity. A significant amount of contraction was observed of the myotube, as
indicated by the displacement of markers on the muscle tube in Fig. B.6.
One of the unique features of optical stimulation is high spatial resolution. Densely
arrayed muscle strips can be activated individually by projecting a light beam at a targeted strip.
Fig. B.7 and Table B.1 show experimental results demonstrating high spatial resolution. Three
muscle strips, Ml, M2, and M3, were activated individually, and as a group. When a small light
beam of 20 pm in diameter, Rl, was projected on Ml, only Ml contracted and the other two did
not. When a large beam covering all the three strips, R2, was projected, all of them contracted, as
shown in the Table B. 1. As long as each muscle strip is isolated and the light beam is narrow
enough to illuminate only the target strip, the targeted strip alone is selectively activated. This
verifies the high spatial resolution of the optogenetic control on the order of 10 pm. Traditional
electric stimulation is unable to activate densely arrayed muscle strips one by one.
It is interesting to note that even when a small portion of the muscle strip was
illuminated, the whole strip contracted. This implies that cations entering the cell propagated
along the muscle strip to cause the whole strip to contract.
Although the 2D assay demonstrated a successful result of light-activated contraction, the
myotubes were not connected to any load and their force characteristics could not be evaluated.
Furthermore, the 2D assay significantly differs from the natural skeletal muscle, which is 3D. In
an attempt to create in vitro muscle strips that are closer to in vivo muscle tissues we have
100
developed a 3D muscle microtissue using MEMS soft lithography: Microfabricated Tissue
Gauges (pTUGs), as illustrated in Fig. B.8 [11]. Between two micro-posts made of PDMS,
optogenetically coded C2C12 myoblasts mixed with collagen gel were seeded. In the micro-well
culture they proliferated, differentiated, and became mature, forming a functional muscle tissue
as evidenced by sarcomeric striation observed in the fused multinucleated myotubes. See Fig.
B.9. The 3D skeletal microtissue formed between the two micro-posts measured 800 Pm x 100
pm x 200 pm in average.
Impulse response tests were conducted for the optogenetically controlled 3D muscle
microtissues. Fig. B.10 shows one trial of experimental results of muscle force generation in
response to a blue light impulses of 20 ms duration. The contraction and relaxation times were
consistent among many muscle constructs, 190.9
11.5 ms and 279.4
16.8 ins, respectively.
We have further quantified the response curve by using a system identification technique. An
Auto-Regressive Moving-Average with Exogenous inputs (ARMAX) model could approximate
the impulse response with a reasonable accuracy using a total of 5 parameters:
y(t)
= -ay(t
- 1) - a2 y(t -2)
+bu(t -1)+b2u(t -2)
+e(t)+ cie(t -1)
(B.1)
This method is validated by using data from a single impulse response as a training set,
and comparing it to a test set of data from multiple impulse responses as shown in Fig. B.11.
Despite the dynamics of optically stimulated contraction not being LTI, the ARMAX
model approximates well the impulse response of the system. There are two main sources of
nonlinearity: 1) the dynamics of sarcomeric contraction, and 2) the dynamics of ion
transportation. The nonlinearities arising from the sarcomeric dynamics arise from changes in
tissue stiffness and damping as a function of displacement and velocity. Fortunately, these
nonlinearities are insignificant enough to justify using the ARMAX model to simulate impulse
response contraction dynamics as seen in Fig. B. 11. The nonlinearities arising from optical
stimulation reflect the dynamics of conformational changes to the membrane-bound ChR2,
induction of cations, e.g. Na+, into the cell, and release of calcium ions Ca++ from the
sarcoplasmic reticulum, as illustrated in Fig. B.2. This is a threshold response of the cell; once a
sufficient number of ions have crossed the cell membrane, a positive feedback response releases
more ions within the cell. For this reason, the dynamic response of the cell may only be
101
appropriately modeled as a series of impulse responses. Furthermore, this part of the system is
nonlinear in that it does not scale with the magnitude of the optical stimulation; a characteristic
impulse response is triggered upon optical stimulation of a minimum threshold intensity.
Many details are unknown of the mechanisms of light-activated muscle dynamics.
Depending on the assay, the contraction magnitude varies. However, the consistent impulse
response wave form, including contraction rise time and duration as well as the relaxation time
constant and duration, indicates that the contraction dynamics are governed by fundamental
biochemical properties of ChR2 conformation and ion dynamics inside the cell.
With this model we can simulate more complex time sequence inputs to generate
arbitrary output responses such as tetanus-like contractions. Stimulation frequency is limited by
the needed duration of optical pulse to initiate an impulse response and the time required for the
cell to recover. The model begins to break down as the stimulation pulse duration approaches
the non-stimulated recovery time. Because the stimulation pulse needs to be -20 ms, the model
is only sufficient at frequencies less than -20 Hz. Shown in Fig. B.12, 10 Hz is sufficient to
generate sustained force profiles similar to those achievable by native muscle. Native muscle
tissue produces holding forces when receiving relatively high frequency control pulses from
motor neurons. We have simulated similar behavior as shown in Fig. B.12. Using twitch and
tetanus contraction profiles generated by our model derived from data, we can design tissue
constructs for particular applications and utilize this model for feedback control.
B.4 APPLICATIONS
B.4.1 MUSCLE-ON-A-CHIP DRUG SCREENINGS
Drug screening is an increasingly important issue in the pharmaceutical industry, as the
cost for new drug development soars and more stringent tests are required for all new drugs. Use
of animals for drug screening must be reduced as well. Creating an in vitro testbed that mimics in
vivo tissue environment is an effective solution to the drug screening problem [12,13].
The proposed skeletal muscle tissue
sTUG
can serve as a Muscle-on-a-Chip drug
screening testbed for two objectives. One is to test effectiveness of drugs for muscle diseases,
such as muscle dystrophies, sarcopenia of the elderly, and muscle disuse. The other is drug
screening for off-target toxicity: testing for adverse effects of new drugs upon muscular
102
functions. Both can be performed by dispensing a drug to the media of the muscle microtissue
model and observing effects on muscular functions.
Optogenetic control allows for high throughput, effective testing of muscle contractile
functions. It should be noted that dynamic contractile tests can reveal important muscular
functions and provide much more copious and reliable test data than merely observing static
properties. The impulse response test using the muscle tissue pTUG provides not only the static
stress and strain of the microtissue but also dynamic changes to the stress and strain in response
to the ion flux induced by the optical impulse. Optical stimulation also facilitates to perform high
throughput tests, since no electric probes must be inserted, unlike electric innervation.
Integrating an optical control device into a microscope, we can build an all-optical testing system
that can perform dynamic response tests wirelessly and without physical contacts. Hundreds of
micro tissues can be built in a single experimental device, and optical impulse response tests can
be performed one after another at a high throughput.
As shown in the previous section, each impulse response can be characterized with just a
few parameters. The second order ARMAX model consisting of five parameters can
approximate each response fairly accurately. When drugs intervene the muscle contractile
function, these parameters of the impulse response model are expected to shift, indicating
adverse effects of the drug upon the muscular function.
B.4.2 MANY DOF ROBOTIC DEVICES
In addition to utilization as a characterization platform, this actuation method may be
used to produce robotic configurations not practically achievable with other methods. Muscle
tissue is a type of "smart" material in that it provides mechanical structure, as well as actuator
force. With the use of optical stimulation, individual muscle actuators may be individually
controlled, and thus, numerous muscle actuators may be used in complex parallel and serial
arrangements. For example, a mouth-like structure having numerous floating loads as shown in
Fig. B.14, is possible. With our actuation model, we can begin to design dynamic systems to be
precisely controlled.
B.5 CONCLUSION
103
Optogenetic control of skeletal muscles and its applications are presented in this chapter.
Skeletal muscles can be activated with light by using membrane-bound Channelrhodopsin 2.
Optogenetic control allows for non-contact, non-invasive, and wireless control of skeletal
muscles with high spatiotemporal resolution. Impulse response of 3D muscle microtissue to
optical stimulation was recorded and characterized as a low-order ARMAX model. Two
potential applications that exploit the unique features of optogenetically controlled skeletal
muscle tissues are proposed. One is an all-optical, high throughput Muscle-on-a-Chip platform
for drug screening, and the other is multi DOF micro robotic systems using optogenetically
controlled skeletal muscles as actuators. Both applications show great promise, opening up new
possibilities of control of live skeletal muscles.
B.6 ACKNOWLEDGMENTS
I'd like to thank Thomas Boudou, Michael Borochin, and Professor Christopher S. Chen
of University of Pennsylvania, and Yinqing Li, Professor Ron Weiss, and Professor Roger D.
Kamm of MIT for their valuable suggestions and technical supports.
This material is based upon work supported by the National Science Foundation under
the Science and Technology Center Emergent Behaviors of Integrated Cellular Systems (EBICS)
Grant No. CBET-0939511.
104
B.7 FIGURES AND CAPTIONS
Fig. B.1 I Optogenetic control of live cells and tissue.
Na+
Na+
+
blueNa
a+acetylcholine
Na+I
ic
*0~4
DHP
recepto
Ca 2+ pump
Fig. B.2 I Membrane-bound Channelrhodopsin 2 serves as an ion gate. It inducing cations, e.g.
Na+, when illuminated with blue light. The sodium ion mediates release of Ca2+ from a sarcoplasmic
reticulum that leads to contraction of the skeletal muscle.
UPRE
2A
pAAV-CAG-ChR2-GFP-2A-Puro
hapR
GP
CMG
Fig. B.3
I Plasmid containing ChR2. ChR2 along with Green Fluorescent Protein (GFP), Puromycin,
and CAG.
105
Fig. B.4
ChR2.
Myoblasts
I Mouse skeletal cells (C2C12) transfected with ChR2. The green areas indicate GFP and
*
Alignment
Fusion
(a)
Fig. B.5
(b)
(c)
I Myogenic process. Precursors of muscle cells called Myoblasts (a), alignment of myoblasts
(b), and cell fusion to create a multi-nucleated myotubes (c).
Fig. B.6 I Experiment of optogenetic contraction control of skeletal muscles. The top panel with
light off and the bottom panel with light on. The markers indicate that the muscle strip contracted
approximately 10 % with the light. on.
106
Fig. B.7 I Experiment of spatial resolution of optogenetic control. The three muscle strips, Ml,
M2, and M3 can be activated individually with light beams RI and R3, or as a group with light beam R2
covering all the three muscle strips.
Fig. B.8 I Microfabricated tissue gauge for formation of 3D skeletal muscle microtissue.
Fig. B.9 I Cross-striations characteristic to sarcomere formation. Stained for a-actinin.
107
14-
Dynamic Tension
12
10
Static Tension
400
800
1600
1200
2000
Time (ms)
Fig. B.10
Impulse response
test of 3D skeletal muscle. Force generation measured using the
'"''12
microfabricated tissue gauge.
0
14
2
-
C1 -ji
1A2
10
--
Measured
--
Simulated
90
z
400
13
800
--
1200
1600
2000
z
Time (ms)
Fig. B.11 |12nd order ARMAX model siumulation with measured contraction data.
4
--- 2.5 Hz
z
0
....
10 H
Fig. B.12
I Tetanus-like
response from simulation. ARMAX model at multiple input pulse
frequencies.
80 wells of JTUG
skeletal muscle
Blue light projector
Fig. B.13 I Muscle-on-a-Chip Drug Screening.
Fig. B.14 I Highly Networked Mouth-Like Muscle System.
109
TABLE B.1 |Individual and group activations
Ri is exposed
M2
R3 is exposed
no
101"
contraction
activity
activity
R2 is exposed
no
contraction
M3
no
activity
contraction
no
contraction
110
APPENDEX
C: PATENT
DESCRIPTION OF PATENT APPLICATION
PCT/US2012/027483
APPARATUS
AND
METHOD FOR ORGANIZING THREE-DIMENSIONAL
CELL
STRUCTURES USING STIFFNESS GRADIENTS AND SACRIFICIAL GELS
RELATED APPLICATIONS
[0001] The present application claims priority under 35 U.S.C.
119(e) to U.S. provisional
application, U.S.S.N. 61/448,944, filed March 3, 2011, the entire contents of which are
incorporated herein by reference.
FIELD OF INVENTION
[0002] The invention relates to systems, devices, apparatuses and methods for organizing cell
structure.
BACKGROUND
[0003] Cellular differentiation is the process by which a progenitor cell is transformed into a cell
type having one or more specialized functions. The process of cell differentiation is influenced
by a variety of extracellular inputs including growth factors, cytokines, and other molecules that
modulate different signaling pathways in cells. Cell-cell interactions and cell-matrix interactions
also influence cell signaling and modulate the differentiation process. As an example,
myogenesis is the process by which progenitor cells differentiate to produce muscle cells and
fibers. Myogenesis involves a proliferative expansion of myogenic precursor cells (e.g. , satellite
cells, myoblasts), synthesis of a fibronectin-rich matrix, migration and alignment of cells, and
fusion of aligned cells into multi-nucleated fibers, called myotubes. Myogenesis involves an
interplay of various signaling molecules, cell-cell interactions, and cell matrix interactions. New
devices are needed for examining and influencing cell function, including differentiation, e.g.,
111
myogenesis, and for identifying factors (e.g. , test agents) that influence cell function, e.g.,
differentiation.
SUMMARY OF THE INVENTION
[0004] The present invention provides cell culture systems for organizing cells in threedimensions. In some embodiments, the invention provides cell culture devices with matrix
stiffness gradients for organizing cells in three-dimensions. In certain embodiments, the cell
culture devices include a primary matrix and one or more regions within the primary matrix that
comprise a secondary matrix, in which the stiffness of the primary matrix is different than the
stiffness of the secondary matrix. In particular embodiments, the secondary matrix comprises
cells. The cells may organize and arrange themselves within the secondary matrix as a result of
the difference in stiffness between the secondary matrix and the primary matrix. Accordingly,
various characteristics of the primary matrix and secondary matrix (including, e.g. , structural
properties (e.g., stiffness, permeability, etc.), size, and shape) may be tuned to control the
organization and arrangement of the cells within the device. In addition, various types of cells
may be cultured in the device. In some embodiments, muscle cells (e.g. , myogenic cells) are
cultured in the device. In some embodiments, combinations of different cells are cultured in the
device, either within the same region or within different regions. For example, non-muscle cells
(e.g. , endothelial cells) may be cultured within the same region or within different regions as
muscle cells.
[0005] In other aspects, the invention provides methods for culturing cells based on stiffness
gradients. The methods typically involve the use of a cell culture device that includes a primary
matrix and one or more regions within the primary matrix that comprise cells within a secondary
matrix. The cell culture device is typically maintained under conditions that support viability of
the cells (e.g. , temperature, 02 concentration, pH, etc.). For example, the device may be
maintained in vitro or may be implanted in a subject and maintained in vivo to support the
viability of the cells in the device. Any of the cell culture devices disclosed herein may be used
in the inventive methods.
112
[0006] The methods may involve influencing cells to modulate one or more phenotypic
characteristics within the device. Cells may be influenced to modulate their shape, size, or threedimensional organization within the matrix. Cells may be induced to enter into the cell cycle or
exit from the cell cycle (e.g., arrest in GO). In some embodiments, the methods involve inducing
the cells to differentiate within the secondary matrix. In some
embodiments, the methods involve inducing the cells to form a tissue or organ-like structure
within the secondary matrix. The cells may be induced to differentiate by, for example,
contacting the cells with a growth factor, cytokine, or other agent that induces differentiation of
the cells. In some cases, various characteristics of the secondary matrix (including, e.g. , its
chemical properties (e.g. , substituent molecules (e.g. , synthetic polymers, natural polymers,
biopolymers, etc.), degree of cross-linking, pH, ionic strength, hydrophilicity, polarity, etc.), its
structural properties (e.g. , stiffness), its size, and its shape) can be tuned to induce
differentiation.
[0007] The invention, in some aspect, relates to methods for producing myotubes. For example,
methods are provided that involve (a) culturing myogenic cells in a device comprising a primary
matrix and one or more regions within the primary matrix that comprise a secondary matrix, in
which the myogenic cells reside in the secondary matrix, and (b) maintaining the device under
conditions that induce differentiation of the myogenic cells. In particular embodiments, the
differentiated cells fuse to form multinucleated myotubes within the secondary matrix. Typically,
the myotubes are arranged according to the shape and size of the secondary matrix. Moreover, by
having a plurality of regions that comprise myogenic cells within a secondary matrix, multiple
regions of myotubes may be formed using a single device.
[0008] In other aspects, the invention relates to methods for evaluating the effect of a test agent
on cell differentiation. In some embodiments, the methods involve culturing cells in a device
comprising a primary matrix and one or more regions within the primary matrix that comprise a
secondary matrix, in which the cells reside in the secondary matrix, and in which the stiffness of
the secondary matrix is different than (greater than or less than) the stiffness of the primary
matrix; contacting the cells with the test agent; and determining whether the cells differentiate or
113
change (e.g., undergo a phenotypic change) within the secondary matrix in the presence of the
test agent or after having been contacted with the test agent. In certain embodiments, the
invention relates to methods for evaluating the effects of a test agent on myo genesis. For
example, the methods may include culturing myogenic cells in a cell culture device of the
invention; contacting the myogenic cells with the test agent; and determining whether the
myogenic cells form myotubes within the device (e.g., within the secondary matrix of the
device).
[0009] In other aspects, the invention relates to methods for evaluating the role of a gene product
on cell differentiation. For example, the methods may involve culturing cells in a device
comprising a primary matrix and one or more regions within the primary matrix that comprise a
secondary matrix, in which the cells reside in the secondary matrix, and in which the stiffness of
the secondary matrix is different than (greater than or less than) the stiffness of the primary
matrix; contacting the cells with a test agent (e.g. , siRNA, antibody, etc.) that inhibits the gene
product; and determining whether the cells differentiate within the secondary matrix. In certain
embodiments, the invention relates to methods for evaluating the role of a gene product on
myotube formation. For example, the methods may involve culturing myogenic cells in a cell
culture device of the invention; contacting the myogenic cells with the test agent that inhibits the
gene product; and determining whether the myogenic cells form myo tubes within the device
(i.e., within the secondary matrix of the device).
[0010] The invention, in some aspects, relates to methods for evaluating the ability of a test cell
to undergo myogenic differentiation. In certain embodiments, the methods involve culturing test
cells (e.g., stem cells) in a device comprising a primary matrix and one or more regions within
the primary matrix that comprise a secondary matrix, in which the test cells reside in the
secondary matrix, and in which the stiffness of the secondary matrix is greater than the stiffness
of the primary matrix; maintaining the cell culture device under conditions that induce the
myogenic cells to form myotubes within the secondary matrix; and determining whether the test
cells form myotubes within the secondary matrix.
114
[0011] In still further aspects, the invention relates to methods for producing a cell culture
device. In some embodiments, the methods involve (a) producing a primary matrix that at least
partially encompasses one or more solid objects; (b) removing one or more of the solid objects to
produce one or more cavities; and (c) filling one or more of the cavities with a secondary matrix,
in which the stiffness of the primary matrix produced is different than the stiffness of the
secondary matrix. In certain embodiments, the secondary matrix includes cells (e.g., stem cells,
pluripotent cells, myogenic cells). The secondary matrix may include other components such as
extracellular matrix proteins (e.g. , collagen, fibronectin, etc.), growth factors, cytokines, drugs,
or other components. In some aspects, the invention relates to a method for producing one or
more matrices that have a defined shape. In some embodiments, the method involves producing a
primary matrix that at least partially encompasses one or more solid objects; removing one or
more of the solid objects to produce one or more cavities; and filling one or more of the cavities
with a secondary matrix, in which the primary matrix has a melting temperature that is less than
the secondary matrix. In some embodiments, the method also involves incubating the primary
matrix having one or more cavities filled with the secondary matrix at a temperature sufficient to
melt the primary matrix but not the secondary matrix, thereby releasing one or more secondary
matrices having a shape defined by the geometry of the one or more cavities. The invention in
some aspects, provides a cell culture device that comprises a primary matrix and one or more
regions within the primary matrix that comprise a secondary matrix, in which the primary matrix
has a melting temperature that is less than the secondary matrix. In one embodiment, the
secondary matrix comprises cells.
[0012] The invention also provides kits comprising a container or package housing any of the
cell culture devices disclosed herein or one or more components for making the cell culture
devices disclosed herein.
DEFINITIONS
[0013] As used herein, the term "appropriate standard" refers to a quantity indicative of a known
outcome, status or result (e.g. , a known differentiation state).
115
[0014] As used herein, the term "agents" or "test agents" refers to peptides, polypeptides,
proteins, small molecules, organic and/or inorganic compounds, polysaccharides, lipids, nucleic
acids, particles, antibodies, ligands, or combinations thereof.
[0015] As used herein, the term "antibody" refers to an immunoglobulin, whether natural or
wholly or partially synthetically produced. All derivatives thereof which maintain specific
binding ability are also included in the term. The term also covers any protein having a binding
domain which is homologous or largely homologous to an immunoglobulin binding domain.
These proteins may be derived from natural sources, or partly or wholly
synthetically produced. An antibody may be monoclonal or polyclonal. The antibody may be a
member of any immunoglobulin class, including any of the human classes: IgG, IgM, IgA, IgD,
and IgE. Derivatives of the IgG class, however, are preferred in the present invention.
[0016] As used herein, the term "antibody fragment" refers to any derivative of an antibody
which is less than full-length. Preferably, the antibody fragment retains at least a significant
portion of the full-length antibody' s specific binding ability. Examples of antibody fragments
include, but are not limited to, Fab, Fab', F(ab')2, scFv, Fv, dsFv diabody, and Fd fragments. The
antibody fragment may be produced by any means. For instance, the antibody fragment may be
enzymatically or chemically produced by fragmentation of an intact antibody, or it may be
recombinantly produced from a gene encoding the partial antibody sequence. Alternatively, the
antibody fragment may be wholly or partially synthetically produced. The antibody fragment
may optionally be a single chain antibody fragment. Alternatively, the fragment may comprise
multiple chains which are linked together, for instance, by disulfide linkages or other more stable
linkages. The fragment may also optionally be a multimolecular complex.
[0017] As used herein, the term "approximately" or "about" in reference to a number are
generally taken to include numbers that fall within a range of 5%, 10%, 15%, or 20% in
116
either direction (greater than or less than) of the number unless otherwise stated or otherwise
evident from the context (except where such number would be less than 0% or exceed 100% of a
possible value).
[0018] As used herein, the term "density" is used with its common technical meaning (e.g., mass
per unit volume, weight per unit volume, etc.). In some cases, density may refer to a specific
region within a matrix (e.g., density of polymers within a primary matrix, density of polymers
within a secondary matrix, etc.). The density may be measured, for example, by taking the mass
or weight divided by the geometric volume described by a shape.
[0019] As used herein, the term "in vitro" refers to events that occur in an artificial environment,
e.g., in a test tube or reaction vessel, in cell culture, etc., rather than within an organism.
[0020] As used herein, the term "in vivo" refers to events that occur within an organism.
[0021] As used herein, the term "matrix" refers to a polymeric network (e.g., a cross-linked
polymeric network). Polymers of the network may be natural, synthetic or a combination thereof.
Any suitable chemical bonding may provide cross-links for a polymeric network including, for
example, covalent bonds, ionic bonds, Van der Waals interactions, hydrogen bonds, hydrophobic
interactions, etc. Cross-links may be formed by chemical reactions that are initiated by
temperature changes, pressure changes, ionic changes, pH changes, or radiation, for example. A
matrix may a porous solid or porous solid-like material. A matrix may be a gel, such as, for
example, a hydrogel, organogel or xerogel.
[0022] As used herein, the term "nucleic acid" refers to a polymer of covalently linked
nucleotide bases. A nucleic acid can be of biologic and/or synthetic origin. The nucleic acid may
be in single-stranded or double- stranded form. Also included within the definition are nucleic
acids having modified nucleotides. Other modifications may involve, for example, modifications
of the backbone. The term nucleic acid embraces DNA, RNA, or PNA
(peptide nucleic acid), or a combination thereof.
117
[0023] As used herein, the term "peptide, " "polypeptide," or "protein" comprises a polymer of
amino acid residues linked together by peptide (amide) bonds. The amino acid residue may be
natural, unnatural, or a derivative thereof. The term(s), as used herein, refers to proteins,
polypeptides, and peptide of any size, structure, or function.
[0024] As used herein, the term "primary matrix" refers to a polymeric network (e.g., a crosslinked polymeric network) that at least partially encompasses one or more regions that have a
secondary matrix. A primary matrix may encompass multiple regions that have the same
secondary matrix. A primary matrix may encompass regions that have different secondary
matrices. The primary matrix may be more stiff or less stiff than the secondary matrix, in certain
embodiments. A primary matrix may comprise one or more matrix-free channels. The primary
matrix may have a permeability that permits the diffusion of growth factors, cytokines, carbon
sources, nitrogen sources, vitamins, and other agents that influence cell function from one region
within the matrix to another (e.g., from a region comprising a secondary matrix to another region
comprising a secondary matrix, from a matrix-free channel to a region comprising a secondary
matrix, etc.). The primary matrix may have a permeability that permits the removal (e.g., by
diffusion) of cellular waste products from a secondary matrix. The primary matrix may have a
structure that accommodates the migration of cells within it. The primary matrix may be
biodegradable.
[0025] As used herein, the term "secondary matrix" refers to a polymeric network (e.g. , a crosslinked polymeric network) that is at least partially encompassed by a primary matrix (at least
during its formation). In some embodiments, the secondary matrix has a different stiffness than
the primary matrix. Typically, the secondary matrix provides a substrate for cell attachment. The
secondary matrix may comprise growth factors, cytokines, carbon sources, nitrogen sources,
vitamins, and other agents that influence cell function (e.g., growth, proliferation, differentiation,
etc.). The secondary matrix may comprise cell culture media or a component thereof. The
secondary matrix may be biodegradable.
118
[0026] As used herein, the term "short-interfering nucleic acid" refers to a small nucleic acid
molecule (e.g. , 15 to 30 nucleotide, 19 to 23 nucleotides, a hairpin RNA, etc.) that inhibits the
expression of a non-coding RNA or an mRNA. The small interfering nucleic acid may be a
microRNA, siRNA, shRNA, antisense RNA, etc. The small interfering nucleic acid may inhibit
transcription, translation and/or may result in degradation of a target nucleic acid (e.g. , of a
target mRNA).
[0027] As used herein, the term "small molecule" is used to refer to molecules, whether
naturally- occurring or artificially created (e.g., via chemical synthesis) that have a relatively low
molecular weight. Typically, a small molecule is an organic compound (i.e., it contains carbon).
The small molecule may contain multiple carbon-carbon bonds, stereocenters, and other
functional groups (e.g. , amines, hydroxyl, carbonyls, heterocyclic rings, etc.). In some
embodiments, small molecules are monomeric and have a molecular weight of less than about
1500 g/mol. In certain embodiments, the molecular weight of the small molecule is
less than about 1000 g/mol or less than about 500 g/mol. Preferred small molecules are
biologically active in that they produce a biological effect in animals, preferably mammals, more
preferably humans. Small molecules include, but are not limited to, radionuclides and imaging
agents. In certain embodiments, the small molecule is a drug. Preferably, though not necessarily,
the drug is one that has already been deemed safe and effective for use in humans or animals by
the appropriate governmental agency or regulatory body. For example, drugs approved for
human use are listed by the FDA under 21 C.F.R.
330.5, 331 through 361, and 440 through
460, incorporated herein by reference; drugs for veterinary use are listed by the FDA under 21
C.F.R.
500 through 589, incorporated herein by reference. All listed drugs are considered
acceptable for use in accordance with the present invention.
[0028] As used herein, the term "stiffness" refers to a resistance to deformation.
Stiffness may refer to the resistance of an elastic, pseudoelastic or viscoelastic object to
deformation. Stiffness may be isotropic or anisotropic. Stiffness may or may not depend on the
shape, size or boundary conditions of the object. In some embodiments, stiffness is measured as
119
a ratio of force and displacement. In some embodiments, stiffness is measured as a ratio of an
applied moment and a rotation. In some embodiments, stiffness is measured as a ratio of applied
shear force and shear deformation. In some embodiments, stiffness is measured as a ratio of
applied torsion moment and angle of twist. In some embodiments, stiffness is measured as a ratio
of stress and strain (i.e., Elastic modulus, Young's modulus or apparent Young' s modulus).
Other appropriate measures of stiffness will be apparent to the skilled artisan.
[0029] As used herein, the term "subject" refers to a mammal, including but not limited to a dog,
cat, horse, cow, pig, sheep, goat, chicken, rodent, or primate. Subjects can be house pets (e.g.,
dogs, cats), agricultural stock animals (e.g., cows, horses, pigs, chickens, etc.), laboratory
animals (e.g., mice, rats, rabbits, etc.), zoo animals (e.g., lions, giraffes, etc.), but are not so
limited. Preferred subjects are human subjects. The human subject may be a pediatric, adult or a
geriatric subject. The human subject may be of either sex.
BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 depicts a non-limiting example of a cross- section of primary matrix poured into
the mold.
[0031] Figure 2 depicts a non-limiting example of primary matrix having a matrix-free cavity.
[0032] Figure 3 depicts a non-limiting example of primary matrix having a region comprising a
secondary matrix with cells.
[0033] Figure 4 depicts a non-limiting example of primary matrix having multiple regions
comprising secondary matrix and cells.
[0034] Figure 5 depicts a non-limiting example of primary matrix having two parallel cylindrical
regions, one comprising cells and secondary matrix, and the other being matrix-free.
120
[0035] Figure 6 depicts a non-limiting example of primary matrix having two parallel cylindrical
regions, one comprising endothelial cells and a secondary matrix, and the other comprising
muscle cells and a secondary matrix.
[0036] Figure 7 depicts a non-limiting example of bundling multiple regions of differentiated
muscle cells.
[0037] Figure 8 depicts a non-limiting example of fibrin gel contraction following cell seeding of
three dimensional tissue constructs.
[0038] Figure 9A and 9B depict a non-limiting example of myotubes produced in a three
dimensional tissue construct. Fibrous structures are F-actin. Nuclei are also stained.
[0039] Figure lOA and 10B depict non-limiting examples of fascicle-like structures produced in
a three-dimensional tissue construct. Solid bodies were calculated from a stack of 2D
microscopic images of F-actin staining and were displayed using Imaris, which is an imaging
software.
[0040] Figure 11 depicts a non-limiting example of a procedure for producing gel casting molds.
[0041] Figure 12A depicts a non-limiting example of a gel casting mold mounted to a glass
substrate with a wire passing through cell entry ports.
[0042] Figure 12B depicts a non-limiting example of a gel casting mold mounted to a glass
substrate.
[0043] Figure 12C depicts non-limiting examples of gel casting molds for single seeding and
dual seeding wells, and for creating multiple tissue constructs per well.
[0044] Figure 13 depicts a non-limiting example of a method for producing three-dimensional
tissue constructs using sacrificial gels.
121
DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS OF THE INVENTION
[0045] Cell culture devices are provided herein that are useful for evaluating and characterizing
cell growth and differentiation and preparing 3-D constructs. The devices are particularly useful
for creating tissue-like structures, including components of muscle tissue (e.g., myotubes). The
devices may be used for testing cell-cell interactions, cell-matrix interactions, or autocrine or
paracrine signaling. The devices may be used for growing artificial organs (e.g., muscles),
growing tissues, growing tissues with associated vasculature, etc. The devices provide a basis for
evaluating the effects of test agents (e.g., therapeutic candidates) on cell function (e.g., growth
and/or differentiation). The devices provide a three-dimensional context for characterizing the
differentiation capacity of progenitor cells. The devices also provide a basis for identifying and
characterizing genes that modulate cell function (e.g., genes that are involved in cell growth
and/or differentiation). Thus, the devices provide a platform for biomarker discovery and
therapeutic target identification.
Cell Culture Devices
[0046] The cell culture devices provided herein are composites that typically comprise a primary
matrix and one or more regions comprising a secondary matrix. The secondary matrix, is
typically different than the primary matrix in terms of its chemical and/or structural
characteristics (e.g., stiffness, permeability) and provides a substrate or moieties for cell
attachment. The cells organize and arrange themselves within the secondary matrix as a result of
the relative stiffness between the secondary matrix and primary matrix.
Accordingly, characteristics of the primary matrix and secondary matrix may be tuned to control
the organization and arrangement of the cells within the device.
[0047] The difference in stiffness. between the secondary matrix of each region and the primary
matrix may be designed and constructed to direct growth and/or alignment of the cells in threedimensions and/or to promote cell differentiation or change. Accordingly, the cell culture devices
122
have stiffness gradients within a three-dimensional composite matrix. The gradient may be
continuous, a step-change in stiffness, or a combination of both. For example, the boundary
between a region of the primary matrix and a region of the secondary matrix may be a stepchange in stiffness between the primary matrix and the secondary matrix. Alternatively, the
boundary between a region of the primary matrix and a region of the secondary matrix may be a
continuous gradient of change in stiffness between the primary matrix and the secondary matrix.
Alternatively, the boundary between a region of
the primary matrix and a region of the secondary matrix may define a combination of a stepchange and a continuous gradient of change in stiffness between the primary matrix and the
secondary matrix.
[0048] The stiffness of the primary matrix may be less than the stiffness of the secondary matrix
or greater than the stiffness of the secondary matrix, depending on the cell type and process
under evaluation. The primary matrix may have an elastic modulus in a range of approximately
0.1 kPa to approximately 1000 kPa, approximately 0.5 kPa to approximately 500 kPa,
approximately 0.5 kPa to approximately 250 kPa, approximately 0.5 kPa to approximately 200
kPa, approximately 0.5 kPa to approximately 100 kPa, approximately 0.5 kPa to approximately
50 kPa, approximately 0.5 kPa to approximately 10 kPa, or
approximately 0.5 kPa to approximately 6 kPa. The primary matrix may have an elastic modulus
of approximately 0.1 kPa, 0.5 kPa, 1 kPa, 2 kPa, 5 kPa, 10 kPa, 20 kPa, 50 kPa, 100 kPa, 200
kPa, 500 kPa, 1000 kPa, or more. The secondary matrix may have an elastic modulus in a range
of 0.1 kPa to 1000 kPa, 0.5 kPa to 500 kPa, 0.5 kPa to 250 kPa, 0.5 kPa to 200 kPa, 0.5 kPa to
100 kPa, 0.5 kPa to 50 kPa, 0.5 kPa to 10 kPa, or 0.5 kPa to 6 kPa. The secondary matrix may
have an elastic modulus of approximately 0.1 kPa, 0.5 kPa, 1 kPa, 2 kPa, 5 kPa, 10 kPa, 20 kPa,
50 kPa, 100 kPa, 200 kPa, 500 kPa, 1000 kPa, or more.
[0049] In some cases where the stiffness of the primary matrix is less than the stiffness of the
secondary matrix, the primary matrix may have an elastic modulus of up to approximately 0.5
123
kPa, and the secondary matrix may have an elastic modulus of greater than 0.5 kPa (optionally
up to 300 kPa, or optionally up to 1 MPa).
[0050] In some cases, the difference in stiffness between the primary matrix and secondary
matrix is a difference in elastic modulus of above 0.1 kPa, above 0.5 kPa, above 1 kPa, above 5
kPa, above 10 kPa, above 20 kPa, above 50 kPa, above 100 kPa, above 200 kPa, or above 500
kPa.
[0051] The primary matrix and/or the secondary matrix may comprise a fibrin gel, an alginate
gel, a collagen gel, an agarose gel, matrigel, gelatin, a biopolymer, biodegradable synthetic
polymers such as polyglycolic acid (PGA), poly-L-lactic acid, poly(lactic acid) (PLA),
polyhydroxyalkanoate,
poly-4-hydroxybutyrate,
polycaprolactone-co-polylactic
acid,
polyethylene glycol (PEG), poly(glycerol sebacate) (PGS), or any combination thereof. The
primary matrix and/or secondary matrix may be biodegradable. Matrices are typically
biocompatible. Where a primary matrix and secondary matrix consist of the same polymeric
constituents, different types and/or concentrations of cross-linking agents may be used, in some
cases, to achieve different structural characteristics (e.g., permeability, stiffness, etc.) in the
primary matrix compared with the secondary matrix. For example, where a primary matrix and
secondary matrix are both fibrin gels, different concentrations of fibrinogen and/or thrombin may
be used in the two matrices to achieve different structural
characteristics. A primary matrix may also comprise different polymeric constituents than a
secondary matrix to achieve different structural characteristics (e.g., permeability, stiffness, etc.)
than the secondary matrix. For example, the primary matrix may be an agarose gel, and the
secondary matrix may be a fibrin gel. The stiffness of the agarose gel may be tuned by
controlling the concentration of agarose. The fibrin gel may be tuned by controlling the
concentrations of fibrinogen and thrombin. Further examples of primary and secondary matrix
compositions will be apparent to the skilled artisan and are disclosed elsewhere herein.
[0052] The primary matrix and/or secondary matrix may comprise any one or more of a variety
of different polymers, including synthetic or natural polymers. The primary matrix and/or
124
secondary matrix may comprise polyesters, polyethers, polyamides, polycarbonates, polyureas,
polystyrenes, polypeptides, polysaccharides, polyacrylates, polyacrylamides, etc.
[0053] Typically, the primary matrix and secondary matrix comprise fibrinogen and thrombin,
and the concentrations of fibrinogen and thrombin are tuned to control the stiffness of the gel. A
non-limiting example of parameters for tuning the stiffness of a fibrin gel are disclosed in Duong
et ah, Modulation of 3D Fibrin Matrix Stiffness by Intrinsic Fibrinogen-Thrombin Compositions
and by Extrinsic Cellular Activity, Tissue Engineering; Part A Volume 15, Number 7, 2009,
incorporated herein by reference.
[0054] The primary matrix and/or secondary matrix may comprise fibrinogen at a concentration
of approximately 0.1 mg/ml, 0.5 mg/ml, 1 mg/ml, 2 mg/ml, 3 mg/ml, 4 mg/ml, 5 mg/ml, 10
mg/ml, 15 mg/ml, 20 mg/ml, 30 mg/ml, 40 mg/ml, 50 mg/ml, or more. The primary matrix
and/or secondary matrix may comprise fibrinogen at a concentration in a range of approximately
0.1 mg/ml to 5 mg/ml, 0.5 mg/ml to 10 mg/ml, 1 mg/ml to 15 mg/ml, 1 mg/ml to 20 mg/ml, 5
mg/ml to 30 mg/ml, 10 mg/ml to 50 mg/ml or 0.1 mg/ml to 50 mg/ml. The primary matrix and/or
secondary matrix may comprise thrombin at a
concentration in a range of approximately 0.1 NIH Units /ml to 100 NIH Units / ml, 0.1 NIH
Units /ml to 50 NIH Units / ml, 0.1 NIH Units /ml to 20 NIH Units / ml, 0.1 NIH Units /ml to 10
NIH Units / ml, 0.1 NIH Units /ml to 5 NIH Units / ml. The primary matrix and/or secondary
matrix may comprise thrombin at a concentration of 0.1 NIH Units /ml, 0.5 NIH Units / ml, 1
NIH Units / ml, 2 NIH Units / ml, 5 NIH Units /ml, 10 NIH Units / ml, 15 NIH
Units / ml, 20 NIH Units / ml, 50 NIH Units / ml, 100 NIH Units / ml, or more. The primary
matrix and/or secondary matrix may contain trace amounts of thrombin. The primary matrix
and/or secondary matrix may be free of thrombin. The primary matrix and/or secondary matrix
may comprise up to 1 NIH Unit/ml of thrombin.
[0055] In other embodiments, when the primary matrix and secondary matrix are both fibrin
gels, the primary matrix and secondary matrix comprise thrombin and fibrinogen at different
125
concentrations to achieve different stiffness properties. For example, the primary matrix may
comprise 1 mg/ml to 10 mg/ml of fibrinogen, and 0.1 NIH Units /ml to 5 NIH Units / ml of
thrombin, while the secondary matrix comprises above 10 mg/ml to 50 mg/ml of fibrinogen and
above 5 NIH Units /ml to 100 NIH Units / ml of thrombin. The primary matrix may comprise
about 10 mg/ml of fibrinogen and about 1 NIH Units / ml of thrombin, while the secondary
matrix comprises about 10 mg/ml of fibrinogen and no thrombin. The primary matrix may
comprise about 1% to 5% agarose (weight to volume). The primary matrix may comprise about
0.1% to 10% agarose (weight to volume). The primary matrix may comprise about 1% to 5%
alginate (weight to volume). The primary matrix may comprise about 0.1% to 10% alginate
(weight to volume). The primary matrix may comprise about 1 mg/ml to 5 mg/mi collagen TypeI. The primary matrix may comprise about 0.1 mg/ml to 10 mg/ml collagen Type-I. The
secondary matrix may comprise about 1 mg/mI to 5 mg/mi collagen Type-I. The secondary
matrix may comprise about 0.1 mg/mi to 10 mg/mi collagen Type-I.
[0056] In some embodiments, the primary matrix is produced such that it solidifies at a relatively
low temperature (e.g. , at or below room temperature) and melts at a relatively high temperature
(e.g. , at a temperature suitable for cell growth, e.g., about 37 *C). In one embodiment, the
primary matrix is maintained at a relatively low temperature (e.g., 4*C to 26 *C) while the
secondary matrix is introduced (typically together with a cell suspension) into a cavity within the
primary matrix. The secondary matrix solidifies at the relatively low temperature, forming a
solid region within the cavity of the primary matrix that conforms to the dimensions of the
cavity, and does not melt at the temperature at which the primary matrix melts (e.g. , the
secondary matrix melts at a temperature above that at which the primary matrix melts). In one
embodiment, the device is exposed (e.g. , by placing the device in an incubator) to a relatively
high temperature (e.g., 37 *C) to melt the primary matrix and release the secondary matrix. This
methodology may be used to produce a matrix having cells distributed within it and having a
defined geometry. Thus, in some
embodiments, the primary matrix functions as a sacrificial mold for making one or more
matrices that have defined geometries. In some embodiments, this methodology is used to
126
produce one or more matrices that have a defined shape (e.g., a muscle fascicle-like shape) and
within which cells (e.g., myogenic cells) are distributed.
[0057] In some embodiments, the primary matrix comprises gelatin and the secondary matrix
comprises a fibrin gel within which cells (e.g. , myogenic cells, e.g. , myoblasts) are distributed.
In some embodiments, the primary matrix comprises gelatin at a concentration of about 10 w/v
and the secondary matrix comprises a fibrin gel with a fibrinogen
concentration of about 10 mg/mi and thrombin concentration of about 5 NIH Units/ml within
which cells (e.g. , myogenic cells, e.g. , myoblasts) are distributed. In one embodiment, the
gelatin that makes up the primary matrix solidifies at a relatively low temperature (e.g. , at or
below room temperature, e.g., 4*C to 26 *C) and melts at a relatively high temperature (e.g. , a
temperature suitable for cell growth, e.g., 37 *C). Thus, in some embodiments, a primary matrix
comprising gelatin is used as a sacrificial mold. The primary gelatin matrix may be prepared
using methods well known in the art. For example, the primary matrix may be prepared by
dissolving gelatin in water at a relatively high temperature, and subsequently exposing the
dissolved gelatin to a relatively low temperature to solidify the gelatin. The primary matrix may
be produced by dissolving gelatin in water at a concentration of about 1% weight by volume
(w/v), about 5% w/v, about 10% w/v, about 15% w/v, about 20% w/v, about 25% w/v, about
30% w/v, about 35% w/v, about 40% w/v, about 45% w/v, about 50% w/v, about 55% w/v,
about 60 %w/v or more. The primary matrix may be produced by dissolving gelatin in water at a
concentration in a range of 1 %w/v to 5 %w/v, 1 % w/v to 10 %w/v, 5 % w/v to 25 % w/v, 10
% w/v to 25 % w/v, 15 % w/v to 40 %w/v, 20 % w/v to 50 % w/v, or 40 % w/v to 60 % w/v.
Other suitable concentrations of gelatin will be apparent to the skilled artisan.
[0058] The shape of each region may be designed and constructed to direct alignment of the cells
in three-dimensions and/or to promote cell differentiation. Accordingly, the regions within the
primary matrix may have any of a variety of shapes. The regions may have a polyhedron-like
shape, a cylindrical shape, a torus-like shape , spherical shape, or a ellipsoidal shape. A plurality
of differently shaped regions may be present in the primary matrix. Typically, the regions have
an elongated shape. The elongated shape may have a length in a range of approximately 1 mm to
127
approximately 1500 mm, 1 mm to 200 mm, 1 mm to 100 mm, 5 mm to 50 mm, or 5 mm to 10
mm. The elongated shape may have a length
of up to 1 mm, 2 mm, 2.5 mm, 5 mm, 10 mm, 15 mm, 20 mm, 25 mm, 50 mm, 100 mm, 150
mm, 200 mm, 250 mm, 500 mm, 1000 mm, 1500 mm, or more. The elongated shape may have
an average width (e.g. , diameter) in a range of 10 pm to 5000 pm, 10 pxt to 2500
sm, 10 jnn to
2000 pm, 10 pm to 1500 pm^ 10 pm to 1000 pm^ 10 gm to 500 gmn, 10 pm to 250
snm, 10 gm to
200 pint, 10 pm to 150 pm, or 100 pm to 1000 pm. The elongated shape may have an average
width (e.g. , diameter) of up to 10 pm, 20 pm, 25
sm, 50 pm,
100 pm, 200 pm, 250 pm, 500 pm,
1000 pm, or more. The elongated shape may have an average width (e.g. , diameter) of less than
1600 pm. One or more of the secondary matrix regions may extend through the primary matrix.
[0059] The cell culture device often comprises a support structure (e.g. , mold) that interfaces
with the primary matrix. The support structure may, among other things, serve to immobilize the
primary matrix, provide a mold for casting the primary matrix and/or secondary matrix, and
provide one or more conduits for circulating a fluid (e.g. , a cell culture media), test agents and
other components to and/or from the device (e.g., circulating a fluid through matrix-free channels
of the primary matrix). The support structure comprises a polymer, metal, ceramic, glass, Velcro
or a combination thereof. The support structure may comprise, for example, a polymeric
organosilicon compound or an acrylic compound. A non-limiting example of a polymeric
organosilicon compound is polydimethylsiloxane (PDMS). A non-limiting example of an acrylic
compound is polymethylmethacrylate (PMMA). A non-limiting example of a metal is aluminum.
The support structure may comprise a coating (e.g. , a conformal coating, Teflon coating, etc.) to
prevent sticking of the matrix.
[0060] The primary matrix may comprise one or more matrix-free channels through which a
fluid, such as, for example, a cell culture medium may be circulated to provide nutrients and
other factors to the cells in the device and to remove waste from the device. The matrix-free
channels, like the one or more regions, may have an elongated shaped that passes through the
primary matrix. For example, the matrix-free channels may have a length in a range of
approximately 1 mm to 1500 mm, 1 mm to 200 mm, 1 mm to 100 mm, 5 mm to 50 mm, or 5 mm
128
to 10 mm. The matrix-free channels may have a length of up to 1 mm, 2 mm, 2.5 mm, 5 mm, 10
mm, 15 mm, 20 mm, 25 mm, 50 mm, 100 mm, 150 mm, 200 mm, 250 mm, 500 mm, 1000 mm,
1500 mm, or more. The matrix-free channels may have an average width (e.g. , diameter) in a
range of 10 gm to 5000 pm, 10 pm to 2500 gat, 10 pnt to 2000 pm, 10 gat to 1500 Int^ 10 pnt to
1000 pz^ 10 jm to 500 pan, 10 jun to 250 pnt, 10
sn to
200 pm, 10 jnp
to 150 pm, or 100 jpn
to 1000 jm. The elongated shape may have an
average width (e.g., diameter) of up to 10 pu, 20 pur, 25 pil, 50 guj, 100 pil, 200 gill, 250 Pm,
500 pig, 1000 pm, or more. Distances between matrix free channels and secondary matrix
regions may be in a range of 100 gu to 5000 jul, for example.
[0061] The matrix-free channels are typically designed and constructed to accommodate flow of
a fluid. The support structure may comprise one or more fluid inlets that are fluidically
connected with one or more of the matrix-free channels. The support structure may also
comprises one or more fluid outlets that are fluidically connected with one or more of the matrix-
free channels. A reservoir may be fluidically connected with one or more of the matrix-free
channels, e.g. , via the fluid inlets of the support structure. A pump may be installed to perfuse
fluid from a reservoir through one or more of the matrix-free channels. A suitable controller may
be connected to the pump to control the fluid flow rate through the device. The fluid flow rate
may be controlled to ensure a proper circulation of nutrients and removal of waste to ensure cell
homeostasis in the device.
[0062] Fluid circulating through the device may comprise one or more nutrients, (e.g., carbon
source, 02, nitrogen source, minerals, vitamins), growth factors, cytokines, or other molecules
that diffuse through the primary matrix into one or more of the regions that comprise the
secondary matrix. These growth factors, cytokines, or other molecules may, for example,
influence the growth, migration and/or differentiation of the cells in the secondary matrix. Fluid
circulating through the device may also, or alternatively, comprises a test agent or other agent
that diffuses through the primary matrix and contacts cells in the secondary matrix.
129
[0063] The matrix-free channels may contain cells that may interact with cells of the same or
different cell type in adjacent secondary matrix regions. Cells in the matrix-free channels may
include endothelial cells, neurons, fibroblasts, stem cells, muscle cells, and cancerous cells. Flow
properties through the matrix-free channels may be used to influence the cells in the matrix-free
channels, such as inducing shear stress on endothelial cells. Cells may be introduced to the
matrix-free channels before or after the secondary matrix regions have been filled with
secondary gel with or without cells. As a non-limiting example, endothelial cells may be seeded
into a matrix-free channel, and induced to form a monolayer along the channel-primary matrix
interface, followed by seeding of muscle precursor cells into an adjacent secondary matrix
region.
[0064] The cell culture device may also be configured such that a microscope may be arranged to
permit observation within one or more of the regions. Microscopic observation
provides a basis for evaluating the phenotype of cells present or growing within the device. Any
of a variety of microscopic techniques may be used to evaluate the cells including, for example,
phase contrast, fluorescence, or confocal microscopy.
[0065] The cell culture device may also comprise a heat transfer element, which may be used to
maintain cells at a predetermined temperature. The predetermined temperature may be in a range
of approximately 30 *C to approximately 45 *C. The predetermined temperature is typically
about 37 'C. The heat transfer device may also be used to melt the primary matrix and release
cells (e.g., differentiated cells, e.g., myotubes) from the device.
[0066] The cell culture device may also comprise a force-transducer configured and arranged to
measure the force of contraction produced by cells in the one or more regions, and may comprise
a strain gauge configured and arranged to measure the extent of contraction produced by cells in
the one or more regions.
Methods for Culturing Cells
130
[0067] The invention also provides methods for culturing cells using in any of the cell culture
devices disclosed herein. Accordingly, the cell culture methods typically involve the use of a cell
culture device that includes a primary matrix and one or more regions within the primary matrix
that comprise cells within a secondary matrix. The cell culture device is typically maintained
under conditions that support viability of the cells. For example, the device may be maintained in
vitro or may be implanted in a subject and maintained in vivo to support viability of the cells.
[0068] The secondary matrix of one or more of the regions typically comprises cells. The cells
may be added after the matrix has been cast or polymerized. Alternatively, the cells may be
added before casting or polymerization of the gel. The cells may be any mammalian cells. The
cells may be any human cells. The cells may be of mesenchymal, ectodermal, and endodermal
origin. The cells may be selected from the group consisting of cord-blood cells, stem cells,
embryonic stem cells, adult stem cells, progenitor cells, induced progenitor cells, autologous
cells, heterologous cells, isograft cells, allograft cells, xenograft cells, and genetically engineered
cells.
[0069] The cells may be selected from the group consisting of lymphocytes, B cells, T cells,
cytotoxic T cells, natural killer T cells, regulatory T cells, T helper cells, myeloid cells,
granulocytes,
basophil granulocytes,
eosinophil
granulocytes,
neutrophil granulocytes,
hypersegmented neutrophils, monocytes, macrophages, reticulocytes, platelets, mast cells,
thrombocytes, megakaryocytes,
dendritic cells, thyroid cells, -thyroid epithelial
cells,
parafollicular cells, parathyroid cells, parathyroid chief cells, oxyphil cells, adrenal cells,
chromaffin cells, pineal cells, pinealocytes, glial cells, glioblasts, astrocytes,
oligodendrocytes, microglial cells, magnocellular neurosecretory cells, stellate cells, boettcher
cells; pituitary cells, gonadotropes, corticotropes, thyrotropes, somatotrope, lactotrophs,
pneumocyte, type I pneumocytes, type II pneumocytes, Clara cells, goblet cells, alveolar
macrophages, myocardiocytes, pericytes, gastric cells, gastric chief cells, parietal cells, goblet
cells, paneth cells, G cells, D cells, ECL cells, I cells, K cells, S cells, enteroendocrine cells,
enterochromaffin cells, APUD cell, liver cells, hepatocytes, Kupffer cells, bone cells, osteoblasts,
131
osteocytes, osteoclast, odontoblasts, cementoblasts, ameloblasts, cartilage cells, chondroblasts,
chondrocytes, skin cells, hair cells, trichocytes, keratinocytes, melanocytes, nevus cells, muscle
cells, myocytes, myoblasts, myotubes,
adipocyte,
fibroblasts, tendon cells, podocytes,
juxtaglomerular cells, intraglomerular mesangial cells, extraglomerular mesangial cells, kidney
cells, kidney cells, macula densa cells, spermatozoa, Sertoli cells, Leydig cells, oocytes, and
mixtures thereof.
[0070] In some cases, one or more regions in a primary matrix comprise cells of a first-type and
one or more of regions in the primary matrix comprise cells of a second-type. The one or more
regions comprising cells of the first-type may be different than the one or more regions
comprising cells of the second-type; alternatively, the one or more regions
comprising cells of the first-type may be the same as the one or more regions comprising cells of
the second- type (i.e., cells of the first- type and second- type may be present in different regions
or in the same region). The cells of the first- type may secrete a growth factor or cytokine that
modulates growth and/or differentiation of the cells of the second-type. The growth factor or
cytokine secreted by the cells of the first-type may diffuse through the primary matrix into one or
more regions comprising cells of the second-type. Thus, the proximity of regions within the
primary matrix comprising different cells types may be controlled to influence the exchange of
small molecules, growth factors, and cytokines between the regions. Accordingly systems of
paracrine and autocrine signaling may be designed into the device. In some cases, cells of the
first-type migrate into and through the primary matrix, and may enter into a region comprising
secondary matrix that supports growth of cells of the second type. The methods are not limited to
only two types of cells, any number of different cell types may be used.
[0071] Moreover, any combination of cells may be used in the device. The cells of a first-type
for example may be endothelial cells, embryonic stem cells (ESC) and neurons (e.g.,
motorneurons). Cells of a second type may be myogenic cells, such as, for example, skeletal
muscle cells, smooth muscle cells, and cardiomyocytes, or cancer cells. Cells of the first-type
may be engineered cells that produce a growth factor or cytokine that modulates growth and/or
differentiation of the cells of the second-type. For example, the growth factor may be selected
132
from the group consisting of an Insulin-like Growth Factor-I (IGF-I), Insulin-like Growth FactorII (IGF- II), Serum response factor (SRF), Hepatocyte Growth Factor, , Fibroblast Growth
Factor- 1, Fibroblast Growth Factor-2, Fibroblast Growth Factor-6, Basic Fibroblast Growth
Factor, Wnt3a, Transforming Growth Factor beta (TGF-p), Angiopoietin, and Vascular
Endothelial Growth Factor (VEGF). The cytokine may be Retinoic Acid (RA), interleukin-6 or
Sonic Hedgehog Homolog (SHH).
[0072] The methods may involve influencing cells to modulate one or more phenotypic
characteristics. The methods may involve, for example, inducing the cells to differentiate within
the secondary matrix. The cells may be induced to differentiate by, for example, contacting the
cells with a growth factor, cytokine or other agent that induces differentiation of the cells. In
some cases, various parameters of the secondary matrix are tuned such that presence of the cells
within the secondary matrix induces differentiation. For example, the chemical properties of the
matrix can be selected to promote differentiation along one or another lineage. Extracellular
matrix factors (natural or synthetic) that are known to induced or promote differentiation may be
included in the matrix. The degree of cross-linking of the matrix can be controlled to achieve
certain matrix stiffness properties or permeability.
Certain cells differentiate when attached to a relatively stiff substrate; others differentiate when
attached to a relatively compliant substrate. Thus, the stiffness of the matrix may be controlled
accordingly.
[0073] In some cases, the delivery of nutrients, growth factors, cytokines, or other molecules is
important for inducing differentiation of the cells within the secondary matrix. These factors, as
described above, can be delivered via diffusion through the primary matrix from a matrix-free
perfusion channel. Additionally, a gradient of these factors may be generated and controlled
within the primary matrix and secondary matrix regions by utilizing a plurality of matrix-free
channels with different independent factor concentrations flowing through them with
independently controlled flow rates. Depending on the magnitude of pressure within matrix-free
channels, and porosity of primary and secondary matrix gels,
133
these chemicals may travel due to hydraulic force as well as diffusion. The primary matrix may
also, or alternatively, be incubated in bath containing a cell culture medium with the growth
factors, cytokines and other molecules, such that the molecules diffuse through the primary
matrix to contact cells in the secondary matrix. In some cases, where the secondary matrix
extends through the primary matrix, the secondary matrix may be in direct contract with the bath
or perfusion channel. In any event, the porosity of the primary matrix and/or secondary matrix
may be tuned to ensure that the growth factors, cytokines and other molecules can adequately
travel via diffusion or pressure driven flow through the matrices and contact the cells. Other
parameters of the matrices may also be controlled including for example the pH of the matrix,
the ionic strength in the matrix, the hydrophilicity or hydrophobicity of the matrix, the polarity of
the matrix, etc.
[0074] In some cases, the methods for producing myotubes are provided. For example, methods
are provided that involve (a) culturing myogenic cells in a device comprising a primary matrix
and one or more regions within the primary matrix that comprise a secondary matrix, in which
the myogenic cells reside in the secondary matrix, and (b) maintaining the device under
conditions that induce differentiation of the myogenic cells. The myogenic cells may be
mesenchymal stem cells, myogenic stem cells, satellite cells, myoblasts or fibroblasts. The
myogenic cells may be identifiable by any of the following biomarkers: Pax 7+, CD34+, CD45-,
and Seal-.
[0075] Typically, the differentiated myogenic cells fuse to form multinucleated myotubes within
the secondary matrix. Typically, the myotubes are arranged according to the shape and size of
the secondary matrix. The shape of each region and/or the difference in stiffness between the
secondary matrix of each region and the primary matrix may be designed and constructed to
promote differentiation of the myogenic cells. The shape of each region may be designed and
constructed to direct alignment of the myogenic cells, or myotubes formed therefrom. The
difference in stiffness between the secondary matrix of each region and the primary matrix may
be designed and constructed to direct alignment of the myogenic cells, or myotubes formed
therefrom.
134
[0076] Moreover, by having a plurality of regions that comprise myogenic cells within a
secondary matrix, multiple myotubes may be formed using a single device. The multiple
myotubes can be isolated from the primary matrix and bundled together in some cases. The
bundled myotubes can be used in vitro, e.g., to study myotube function {e.g., contractility
mechanisms) or implanted in vivo for therapeutic purposes.
Methods for Evaluating Test Agents, Gene Function and Test Cells
[0077] The invention also provides methods for evaluating the effects of a test agent on cell
function (e.g., growth, differentiation, etc.). The methods often involve culturing cells in a device
comprising a primary matrix and one or more regions within the primary matrix that comprise a
secondary matrix, in which the cells reside in the secondary matrix, and in which the stiffness of
the secondary matrix is different (greater than or less than) the stiffness of the primary matrix.
The cells residing within the secondary matrix are contacted with the test agent and the effects of
the test agent on the cells are evaluated. For example, the effect that the test agent has on
differentiation of the cells within the secondary matrix may be determined.
[0078] In some cases, if the cells do not differentiate within the secondary matrix, the test agent
is identified as inhibiting differentiation of the cells. For example, if the cells are treated with a
growth factor, cytokine or other molecule that is known to induce
differentiation and, in the presence of the test agent, do not differentiate within the secondary
matrix, then the test agent is identified as inhibiting differentiation of the cells. In other cases, if
the cells do differentiate within the secondary matrix, the test agent is identified as inducing
differentiation of the cells. For example, if the cells are maintained under conditions that are
known to not induce differentiation and, in the presence of the test agent, the cells do
differentiate, then the test agent is identified as inducing differentiation of the cells. For example,
if the cells are treated with a growth factor, cytokine or other molecule that is known to inhibit
differentiation and, in the presence of the test agent, the cells do differentiate within the
secondary matrix, then the test agent is identified as inducing differentiation of the cells.
135
[0079] Methods for evaluating the effect of a test agent on myogenesis are also provided. For
example, the methods may include culturing myogenic cells in a cell culture device of the
invention; contacting the myogenic cells with the test agent; and determining whether the
myogenic cells form myotubes within device (e.g., within the secondary matrix of the device).
[0080] In some cases, if the cells do not undergo myogenesis, the test agent is identified as
inhibiting myogenesis of the cells. For example, if the cells are treated with a growth factor,
cytokine or other molecule (e.g., IGF-I) that is known to induce myogenesis and, in the presence
of the test agent, do not undergo myogenesis, then the test agent is identified as inhibiting
myogenesis of the cells. In other cases, if the cells do undergo myogenesis the test agent is
identified as inducing myogenesis of the cells. For example, if the cells are maintained under
conditions that are known to not induce myogenesis and, in the presence of the test agent, the
cells do undergo myogenesis, then the test agent is identified as inducing myogenesis of the
cells. For example, if the cells are treated with a growth factor, cytokine or other molecule that is
known to inhibit myogenesis (e.g., TGF-pI) and, in the presence of the test agent, the cells do
undergo myogenesis within the secondary matrix, then the test agent is identified as inducing
myogenesis of the cells.
[0081] The invention also relates to methods for evaluating the role of a gene product on cell
differentiation. For example, the methods may involve culturing cells in a device comprising a
primary matrix and one or more regions within the primary matrix that comprise a secondary
matrix, in which the cells reside in the secondary matrix, and in which the stiffness of the
secondary matrix is different (greater than or less than) the stiffness of the primary matrix. The
cells are typically contacted a test agent (e.g. , siRNA, antibody, etc.) that inhibits the gene
product; and a determination is made as to whether the cells
differentiate within the secondary matrix. When the gene product is a protein, the test agent may
be a short-interfering nucleic acid that specifically targets the mRNA encoding the protein. When
the gene product is a cell-surface protein (e.g., a receptor), the test agent may be antibody or
antigen-binding antibody fragment that specifically targets and inactivates (or activates) the
136
protein. When the gene product is a non-coding RNA, the test agent may be a short-interfering
nucleic acid that specifically targets the non-coding RNA.
[0082] In some cases, if the cells do not undergo myogenesis, the gene product is identified as
inducing myogenesis of the cells. For example, if the cells are treated with a growth factor,
cytokine or other molecule (e.g. , IGF-I) that is known to induce myogenesis and, in the presence
of an agent that inhibits the gene product, do not undergo myogenesis, then the gene product is
identified as inducing myogenesis of the cells. In other cases, if the cells do undergo myogenesis
the gene product is identified as inhibiting myogenesis of the cells. For example, if the cells are
maintained under conditions that are known to not induce myogenesis and, in the presence of the
test agent that inhibits the gene product, the cells do undergo myogenesis, then the gene product
is identified as inhibiting myogenesis of the cells.
[0083] The invention also provides methods for evaluating the ability of a test cell to undergo
myogenic differentiation. The methods typically involve culturing test cells (e.g. , stem cells) in a
device comprising a primary matrix and one or more regions within the primary matrix that
comprise a secondary matrix, in which the test cells reside in the
secondary matrix, and in which the stiffness of the secondary matrix is greater than the stiffness
of the primary matrix. The cells are typically maintained in the cell culture device under
conditions that induce the myogenic cells to form myotubes within the secondary matrix; and a
determination is made as to whether the test cells form myotubes within the secondary matrix.
[0084] The effectiveness of a test agent, the role of the gene product or the extent of
differentiation may be established by comparing assay results with an appropriate standard. An
appropriate standard may be experimentally determined or may be pre-existing (e.g., a historical
quantity, etc.) For example, an appropriate standard for a positive effect on differentiation can be
established by evaluating the effects of an agent known to induce differentiation. An appropriate
standard for a negative effect on differentiation can be established by evaluating the effects of an
agent known to inhibit differentiation. In any case, the appropriate standard may be a value
indicative of differentiation status. For example, in the case of myogenic differentiation, the
137
appropriate standard may be a number of nuclei per myotube, length of myotubes, fraction of
non-differentiated to differentiated cells in a microscopic image field or within a region of a
primary matrix. An appropriate standard may also be an image or images, or descriptive and/or
quantitative features thereof, that are characteristic of a differentiated state or non-differentiated
state.
Methods for Producing Cell Culture Devices
[0085] The invention also relates to methods for producing cell culture devices. The methods
typically involve the production of a primary matrix that at least partially
encompasses one or more solid objects. The primary matrix is typically formed by adding
monomers, polymers, cross-linking agents and/or combinations thereof into a support structure
(e.g., casting mold) under conditions that result into formation of a cross-linked polymeric
network. For example, a primary matrix comprising a fibrin gel can be created by combining
fibrinogen and thrombin in a mold under appropriate conditions (e.g., appropriate temperature
conditions) for forming a fibrin matrix.
[0086] The solid objects are used to create cavities in the primary matrix. The cavities may be
filled with secondary matrix or left vacant to create matrix-free channels. To produce one or
more cavities (or regions) having a particular shape (e.g., elongated shape) within the primary
matrix, one or more of the solid objects having the desired shape are used. The cavities are
produced by removing the solid objects. However, cavities may also be created by casting a
primary matrix without solid objects and boring a hole into or through the solidified primary
matrix. The cavities can then be filled with a secondary matrix that has the desired set of
chemical and structural properties. Typically, the stiffness of the primary matrix produced is
different than the stiffness of the secondary matrix produced within the cavities.
[0087] Matrix-free channels may be created by producing a primary matrix that partially
encompasses one or more solid objects that extend through the primary matrix. The objects are
then removed without being filled with secondary matrix, leaving a matrix free channel. Matrix138
free channels may also be created by casting a primary matrix without solid objects and boring a
channel (hole) through the solidified matrix.
[0088] In some cases, a stiffening agent is introduced into the cavities either before or after they
are filled with the secondary matrix. The stiffening agent diffuses into the primary matrix and
produces in the primary matrix a gradient of increasing stiffness directed toward the one or more
cavities. When the stiffening agent is introduced after the secondary matrix is added a continuous
stiffness gradient or a combination of continuous gradient and step-change in stiffness may be
generated, depending on the composition of the primary and secondary matrices. When the
stiffening agent is introduced after the secondary matrix is added a step-change in stiffness or a
combination of continuous gradient and step-change in stiffness may be generated, depending on
the composition of the primary and secondary matrices.
[0089] Often each cavity is filled with the secondary matrix as the solid object is being removed.
This can serve in part to prevent collapse of the cavities. In cases where the solid objects extend
through the primary matrix, the solid object can be extracted from one end, while a secondary
matrix is delivered to the cavity at the other end. According to this method, the secondary matrix
is drawn in to the cavity by a negative pressure created by extraction of the solid object.
[0090] As described above the primary matrix and/or the secondary matrix may comprise a fibrin
gel, an alginate gel, a collagen gel, an agarose gel, matrigel, gelatin, a biopolymer, or any
combination thereof. When the secondary matrix comprises a fibrin gel, filling the cavities may
involve mixing a first solution comprising fibrinogen and a second solution comprising thrombin
and injecting the mixture into the one or more of the cavities. The mixing and injecting may be
accomplished using a device having a first solution comprising fibrinogen in a first syringe and a
second solution comprising thrombin in a second syringe, in which the first syringe and the
second syringe have tips forming a common port, and wherein mixing occurs as the first solution
and second solution are ejected through the common port. Cells may be present as a suspension
in the first solution and/or the second solution, and thus, may be delivered to the device
concomitantly with the secondary matrix.
139
Kits
[0091] The cell culture devices described herein may, in some embodiments, be assembled into
kits to facilitate their use in assays, diagnostics, biomarker development, research or other
applications. Components for producing the cell culture devices may also be assembled into kits
for the same purpose. A kit may include one or more containers housing the components of the
invention and instructions for use. Specifically, such kits may include one or more cell culture
devices described herein, along with instructions describing the intended application and the
proper use of these devices. Kits are also provided that include one or more components for
producing cell culture devices along with appropriate instructions. Kits may include, for
example, matrix constituents, pre-cast matrices, molds for casting matrices, solid regions for
producing cavities within primary matrices, cells, etc.
[0092] As used herein, "instructions" can define a component of instruction and/or promotion,
and typically involve written instructions on or associated with packaging of the invention.
Instructions also can include any oral or electronic instructions provided in any manner such that
a user will clearly recognize that the instructions are to be associated with the kit, for example,
audiovisual {e.g., videotape, DVD, etc.), Internet, and/or web-based communications, etc. The
written instructions may be in a form prescribed by a
governmental agency regulating the manufacture, use or sale of biological products.
[0093] The kit may contain any one or more of the components described herein in one or more
containers. As an example, in one embodiment, the kit may include instructions for mixing one
or more components of the kit, e.g., to produce a primary matrix or a secondary matrix. The kit
may include a container housing components described herein. The components may be in the
form of a liquid, gel or solid (powder). The components may be prepared sterilely, packaged and
shipped refrigerated or frozen. Alternatively the
components may be housed in a vial or other container for storage.
140
[0094] The kit may have a variety of forms, such as a blister pouch, a shrink wrapped pouch, a
vacuum sealable pouch, a sealable thermoformed tray, or a similar pouch or tray form, with the
accessories loosely packed within the pouch, one or more tubes, containers, a box or a bag. The
kit may be sterilized after the accessories are added, thereby allowing the individual accessories
in the container to be otherwise unwrapped. The kits can be sterilized using any appropriate
sterilization techniques, such as radiation sterilization, heat
sterilization, or other sterilization methods known in the art. The kit may also include other
components, depending on the specific application, for example, containers, cell media, salts,
buffers, reagents, syringes, needles, a fabric, such as gauze, for applying or removing a
disinfecting agent, disposable gloves, a support for the agents prior to administration etc.
[0095] While several embodiments of the present invention have been described and illustrated
herein, those of ordinary skill in the art will readily envision a variety of other means and/or
structures for performing the functions and/or obtaining the results and/or one or more of the
advantages described herein, and each of such variations and/or modifications is deemed to be
within the scope of the present invention. More generally, those skilled in the art will readily
appreciate that all parameters, dimensions, materials, and configurations described herein are
meant to be exemplary and that the actual parameters, dimensions, materials, and/or
configurations will depend upon the specific application or applications for which the teachings
of the present invention is/are used. Those skilled in the art will recognize, or be able to ascertain
using no more than routine experimentation, many equivalents to the specific embodiments of
the invention described herein. It is, therefore, to be understood that the foregoing embodiments
are presented by way of example only and that, within the scope of the appended claims and
equivalents thereto, the invention may be practiced otherwise than as specifically described and
claimed. The present invention is directed to each individual feature, system, article, material,
kit, and/or method described herein. In addition, any combination of two or more such features,
systems, articles, materials, kits, and/or methods, if such features, systems, articles, materials,
kits, and/or methods are not mutually inconsistent, is included within the scope of the present
invention.
141
[0096] Use of ordinal terms such as "first," "second," "third," etc., in the claims to modify a
claim element does not by itself connote any priority, precedence, or order of one claim element
over another or the temporal order in which acts of a method are performed, but are used merely
as labels to distinguish one claim element having a certain name from another element having a
same name (but for use of the ordinal term) to distinguish the claim elements.
142