In Vitro Cell. Dev. Biol.—Animal 42:75–82, March and April 2006
? 2006 Society for In Vitro Biology
FUNCTIONAL EVALUATION OF NERVE–SKELETAL MUSCLE CONSTRUCTS
ENGINEERED IN VITRO
LISA M. LARKIN,1JACK H. VAN DER MEULEN, ROBERT G. DENNIS, AND JEFFREY B. KENNEDY
Department of Biomedical Engineering (L. M. L., R. G. D.), Division of Geriatric Medicine (L. M. L.), Muscle Mechanics Laboratory
(L. M. L., J. H. V., J. B. K.), Department of Mechanical Engineering (R. G. D.), University of Michigan, Ann Arbor, Michigan 48109-
2007, and Biomedical Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7575 (R. G. D.)
(Received 27 September 2005; accepted 15 December 2005)
Previously, we have engineered three-dimensional (3-D) skeletal muscle constructs that generate force and display a
myosin heavy-chain (MHC) composition of fetal muscle. The purpose of this study was to evaluate the functional char-
acteristics of 3-D skeletal muscle constructs cocultured with fetal nerve explants. We hypothesized that coculture of
muscle constructs with neural cells would produce constructs with increased force and adult MHC isoforms. Following
introduction of embryonic spinal cord explants to a layer of confluent muscle cells, the neural tissue integrated with the
cultured muscle cells to form 3-D muscle constructs with extensions. Immunohistochemical labeling indicated that the
extensions were neural tissue and that the junctions between the nerve extensions and the muscle constructs contained
clusters of acetylcholine receptors. Compared to muscles cultured without nerve explants, constructs formed from nerve–
muscle coculture showed spontaneous contractions with an increase in frequency and force. Upon field stimulation, both
twitch (2-fold) and tetanus (1.7-fold) were greater in the nerve–muscle coculture system. Contractions could be elicited
by electrically stimulating the neural extensions, although smaller forces are produced than with field stimulation. Severing
the extension eliminated the response to electrical stimulation, excluding field stimulation as a contributing factor. Nerve–
muscle constructs showed a tendency to have higher contents of adult and lower contents of fetal MHC isoforms, but the
differences were not significant. In conclusion, we have successfully engineered a 3-D nerve–muscle construct that
displays functional neuromuscular junctions and can be electrically stimulated to contract via the neural extensions
projecting from the construct.
Key words: myooid; coculture; neuromuscular junction.
The production and use of engineered muscle from myogenic
cells harvested from muscle biopsies is potentially a powerful tool
for the restoration of muscle function following acute injury, surgery,
or disease. Musculoskeletal tissues are highly integrated elements
within the context of a living system, none of which are known to
achieve adult phenotype without input from other components of
the living system. Skeletal muscle has three principal tissue inter-
faces: vascular, myotendinous, and neuromotor. The inclusion of a
functional nerve–muscle interface in vitro greatly expands the po-
tential to control the phenotype of the muscle tissue in culture and
thus expands the usefulness of engineered muscle for virtually all
of its possible applications. Nerve–muscle cocultured monolayers
have been used previously to study nerve–muscle interaction and
the formation of neuromuscular junctions (Ecob, 1983; Ecob et al.,
1983; Bryers and Ecob, 1984; Ecob, 1984; Whalen et al., 1984,
1985; Dennis and Kosnik, 2000; Wagner et al., 2003). Therefore,
the goal of this experiment was to combine previously described
methods for constructing three-dimensional (3-D) muscle constructs
1To whom correspondence should be addressed at Institute of Gerontology,
NIB, Room 956, 300 North Ingalls Street, University of Michigan, Ann Arbor,
Michigan 48109-2007. E-mail: firstname.lastname@example.org
(Dennis and Kosnik, 2000; Dennis et al., 2001; Kosnik et al., 2001;
Baker et al., 2003) and the methods for nerve–muscle coculture to
construct and evaluate the contractile and structural characteristics
of 3-D skeletal muscle constructs cocultured with fetal nerve ex-
Our laboratory has developed (Dennis and Kosnik, 2000; Dennis
et al., 2001; Kosnik et al., 2001) a repeatable technique for engi-
neering scaffold-free 3-D skeletal muscle tissue for the study of the
functional development of muscle that is based largely on the work
of Strohman (Strohman et al., 1990). The 3-D muscle tissues pro-
duced in this manner, termed ‘‘myooids,’’ display many important
functional similarities with skeletal muscle, including positive force
frequency, normal length–tension relationships, and a normal met-
abolic profile (Dennis and Kosnik, 2000; Dennis et al., 2001; Kos-
nik et al., 2001). We have expanded this muscle model and devel-
oped a new model for a 3-D nerve–muscle construct using the co-
culture of a confluent monolayer of myotubes and embryonic day
15 (E-15) rat spinal cord with dorsal root ganglia attached. The use
of the intact embryonic spinal cord has been previously shown to
maintain the cytoarchitecture of the dorsal root ganglion and pro-
mote myocyte innervation (Kobayashi et al., 1987). Using this model
of nerve–muscle coculture, Wagner et al. (2003) demonstrated that
mouse myotubes were innervated and that further differentiation
LARKIN ET AL.
cord explants from E-15 fetal rats with tail still attached were pinned onto the
muscle cell monolayer 12 mm apart with the tail portion oriented toward the edge
of the dish (A). E-15 neural tissue integrates with muscle monolayer (B). Ap-
proximately 1 wk later, monolayers will roll up around the tail anchors and form
a cylindrical construct. Following formation ofa three-dimensionalconstruct,some
of the neural tissue forms extensions projecting from the construct (C).
Two weeks after the initial plating of the muscle stem cells, two spinal
from myotube to mature muscle fibers occurred. Further investiga-
tion of this nerve–muscle coculture system demonstrated sensitivity
to d-tubocurarine, an adult-like metabolic enzyme profile, and or-
ganized sarcomeric structure (Dorchies et al., 2001).
In vivo, as the nervous system develops and innervation increas-
es, fetal muscles shift toward the adult phenotype. Just as motor
neurons control the expression of myosin heavy chain in adult mus-
cle (Close, 1965, 1969), the motor neurons also exert this control
during development of muscle cells (Schiaffino et al., 1986). The
control over fiber type expression is by two distinct mechanisms
(Schiaffino et al., 1999). The first is a local mechanism involving
the release of neural regulatory factors at the level of the neuro-
muscular junction and a second mechanism that is mediated by the
activation pattern generated by the nerve (Schiaffino et al., 1999).
Acetylcholine receptors are among the first proteins expressed
during myogenesis (Burden, 1998; Sanes and Lichtman, 1999). Pri-
or to innervation, acetylcholine receptors are randomly dispersed
along the surface of the developing myoblast. During the formation
of a neuromuscular junction, the chemotrophic, electrical, and me-
chanical signals between the nerve and muscle induces the aggre-
gation of nicotinic acetylcholine receptors in the plasma membrane
of the developing muscle fiber. The maintenance of the neuromus-
cular junction depends on continual cross talk between the nerve
and muscle at the site of nerve–muscle contact.
This study will investigate whether the introduction of fetal neu-
ral tissue to developing engineered muscle constructs will lead to
functional motor innervation of the muscle constructs and a shift
toward the adult phenotype. We hypothesize that the innervation of
muscle constructs will shift the phenotype of the nerve–muscle from
fetal toward adult and that nerve–muscle constructs will exhibit an
increase in force production and adult myosin heavy-chain expres-
MATERIALS AND METHODS
Animal model and animal care. Tissue engineering studies were carried
out using muscle tissue from female Fischer 344 pregnant rats obtained from
the Charles River Laboratories, Inc. (Wilmington, MA). All animals were
acclimated to our colony conditions (i.e., light cycle and temperature) for 1
wk before any procedure. Rats were housed in pairs in hanging plastic cages
(28 ? 56 cm) and kept on a 12:12 light:dark light cycle at a temperature of
20–22? C. The animals were fed Purina Rodent Chow 5001 laboratory chow
and water ad libitum. At E15 gestation day, surgical procedures were per-
formed to remove both soleus muscles and tail tendon from the pregnant
mom. E15 fetuses were cesarean derived and spinal cord explants obtained.
All surgical procedures were performed in an aseptic environment with an-
imals in a deep plane of anesthesia induced by i.p. injections of sodium
pentobarbital (65 mg/kg). Supplemental doses of pentobarbital were admin-
istered as required to maintain an adequate depth of anesthesia. All animal
care and animal surgery were in accordance with the Guide for Care and
Use of Laboratory Animals (Public Health Service, 1996, NIH Publication
No. 85-23); the experimental protocol was approved by the University Com-
mittee for the Use and Care of Animals.
Preparation of solutions and media. Unless otherwise indicated, all solu-
tions and media were prepared and stored at 4? C before isolation and culture
of muscle cells and warmed to 37? C in a heated water bath immediately
before use. The media, with slight modifications from Dennis and Kosnik
(2000; Baker et al., 2003) were as follows: growth medium (GMA) consisted
of 400 ml of HAM F-12 Nutrient Mixture (Gibco BRL Cat# 11765-054), 100
ml fetal bovine serum (Gibco BRL Cat# 10437-028), and 5 ml A9909 (Sigma
A9909). Differentiation medium (DMA) consisted of 465 ml Dulbecco mod-
ified Eagle medium (DMEM; Gibco BRL Cat# 11995-065), 35 ml 100% horse
serum albumin (Gibco BRL Cat# 16050-122), and 5 ml A9909. The tissue
was dissociated in a dispase and collagenase solution (D&C) that was pre-
ENGINEERED NERVE–MUSCLE CONSTRUCTS
dimensional construct. For field stimulation of the entire construct, platinum
wire electrodes were positioned on either side of the construct (A). Following
the direct field stimulation of the entire construct, a microelectrode was used
to electrically stimulate the neural extensions projecting from the construct
(B). Nerve–myooid with neural extension (C).
Equipment setup for measuring contractile properties of a three-
d of coculture with muscle cell monolayers. (A) When fetal spinal cord ex-
plants are placed in a dish with a myooid, neural cells first migrate away
from the site of explant (B) then aggregate to form larger neural extensions.
Light microscopy of E-15 fetal spinal cord explants following 3
pared in the amount of 20 ml per four soleus muscles and consisted of 8
units Dispase (Sigma Cat# P-3417; 0.4 units/mg) and 200 units of type 4
collagenase (Gibco BRL Cat# 17104-019; 239 units/mg) per ml DMEM.
Transport medium (TM) was prepared in the amount of 5 ml per muscle
dissected at the concentration of 2% A9909 in DPBS. Preincubation medium
(PIM) was prepared in the amount of 3 ml per plate and consisted of 2.5 ml
of 0.05% sodium azide (NaN3; Sigma Cat# S-8032) in DPBS solution, 22.5
ml DMA, and 0.25 ml A9909 per muscle dissected.
Preparation of culture dishes and anchors. Myooids were engineered in
individual 35-mm plates as described earlier (Dennis and Kosnik, 2000).
Briefly, individual plates allowed for functional evaluation of individual
myooids and facilitated removal of any contaminated plates. Each 35-mm
plate was coated with 1.5 ml of Sylgard (Dow Chemical Corporation, Midland,
MI; type 184 silicon elastomer) and allowed to cure for 3 wk before use. One
week before use, Sylgard-coated plates were then coated with laminin at 1.0
?g/cm2per plate (10 ?g of natural mouse laminin [Gibco BRL Cat# 23017-
015] and 3 ml of Dulbecco phosphate-buffered saline [DPBS] pH 7.2 [Gibco
BRL Cat# 14190-144] per plate) and left to dry for 48 h. Salt crystals were
dissolved and removed by rinsing the plates with 3 ml DPBS. The plates
were then filled with 2 ml of previously described GMA and decontaminated
with UV light (wavelength 253.7 nm) for 90 min and placed in a 37? C 5%
CO2incubator for 1 wk before plating muscle cells.
Preparation of muscle and isolation of satellite cells. Both soleus muscles
were surgically removed under aseptic conditions, weighed, sterilized in 70%
ETOH, and incubated for 5 min in TM. A single-edged razor blade and a
pair of number 5 forceps (Fine Science Tools) were then used to slice the
soleus muscle longitudinally into three strips. Next, 35-mm Sylgard-treated
plates were sterilized in 70% ETOH, and muscle slices were pinned to
length, two muscle strips per plate. Then 3 ml PIM were added to each plate,
and the plates were UV treated for 90 min. The plates were then placed in
a 37? C 5% CO2incubator for 50 h.
LARKIN ET AL.
coculture. Images were taken at ?400. (A) Light microscope, (B) fluorescent
image of neural cells stained with CY3-labeled neurofilament. When placed
on a monolayer of muscle cells, neural cells migrate off the fetal spinal cord
explant and fuse with the existing myotubes.
Light microscopy of myotubes and neural cells following 1 wk of
formed between rat fetal spinal cord explants and three-dimensional muscle
constructs. Immunohistochemistry indicates that the muscle constructs con-
tain the Ach receptors necessary for innervation and that these receptors are
clustered in a structure resembling a neuromuscular junction.
Fluorescent microscopy images of neuromuscular junctions
Following the 50-h incubation period, muscle slices were inspected for
contamination, and any infected plates were discarded. The remaining muscle
strips were then removed from the plates and incubated in D&C (two soleus
muscles per 20 ml) in a 37? C shaking water bath for 4 h. The dissociation
was aided by occasionally shaking each vial slightly by hand. Once the
muscle was fully digested 4 h, the dissociated cells were filtered through a
100-micron filter and centrifuged at 700 ? g for 10 min at 25? C. Finally,
the supernatant was aspirated from the vials, and the pellet was resuspended
in GMA to obtain a concentration of 10 mg of dissociated muscle per 2 ml
Cell culture and myooid formation. The previously prepared laminin-coated
plates were examined for contamination after storage in an incubator for 1
wk, and the GMA was aspirated from the plates. Two milliliters of the cell
suspension were plated in each culture dish and placed in a 37? C 5% CO2
incubator for 5 d. Culture plates were not disturbed for at least 72 h to allow
cell adherence to the plates. After 5 d, the cells were fed with GMA every
48 h until the cells became confluent (approximately 7 d). Once the cells
achieved confluence, they were fed with DMA every 48 h until the myocytes
fused to form multinucleated myotubes that began to contract spontaneously.
Two weeks after the initial plating of the muscle stem cells, two spinal cord
explants from E-15 fetal rats with tail still attached were pinned onto the
muscle cell monolayer 12 mm apart with the tail portion oriented toward the
edge of the dish (Fig. 1). Approximately 1 wk later, monolayers will roll up
around the tail anchors and form a cylindrical construct. Sixteen to 18 d post
construct formation, the nerve–muscle constructs were tested for contractile
function. At the conclusion of functional tests, each specimen was snap frozen
between dry ice and stored for subsequent analysis of myosin heavy-chain
profile. Muscle constructs consisting of primary muscle cells and adult rat
tail tendon served as control for the nerve–muscle constructs.
Functional testing of constructs. Contractile properties were initially mea-
sured 16–18 d after the formation of a three-dimensional construct. The
protocol for measuring contractility of engineered muscle constructs was
adapted from Dennis and colleagues (Dennis and Kosnik, 2000; Dennis et
al., 2001; Kosnik et al., 2001) and Irintchev et al., (1998). Briefly, the pin
on one end of the construct was freed from the Sylgard and attached to a
force transducer with canning wax. For field stimulation of the entire con-
struct, platinum wire electrodes were positioned on either side of the con-
struct (Fig. 2). The temperature of the construct was maintained at 37 ? 1?
C using a heated aluminum platform. The diameter of the construct was
determined and used to calculated cross-sectional area, assuming a circular
cross section. Passive baseline force was measured as the average baseline
passive force preceding the onset of stimulation. Twitches were elicited using
a single 1.2-ms pulse at 2.5, 5, 10, and 20 V, whereas maximum tetanic
force was determined using a 1-s train of 1.2-ms pulses at 10 V and 10, 20,
40, 60, and 80 Hz. Data files for each peak twitch force and peak tetanic
force trace were recorded at 1000 samples/s and stored for subsequent anal-
ysis using LabVIEW data acquisition software. Peak tetanic force was nor-
malized for cross-sectional area to determine maximum specific force. Fol-
lowing the direct field stimulation of the entire construct, a microelectrode
was used to electrically stimulate the neural extensions projecting from the
construct using the same stimulation parameters as described previously for
the field stimulation (Fig. 2).
Myosin heavy-chain Western blotting. Western analysis of myosin heavy
chains was conducted as previously described (Talmadge and Roy, 1993).
Briefly, muscle-only and nerve–muscle constructs (n ? 5) were homogenized
in phosphate buffer in a 1:100 (wt/vol) dilution. Twenty microliters of ho-
mogenate was used to determine proteins via Bradford protein assay (Biorad,
Richmond, CA). Western analysis was performed by using a vertical SDS
page gel electrophoresis in the presence of sodium dodecyl sulfate and ali-
quots of muscle membranes (2 ?g protein per lane). Several internal control
ENGINEERED NERVE–MUSCLE CONSTRUCTS
anus (C) in response to field stimulation. The field stimulation was across the
nerve–muscle construct and not the neural extensions.
Representative force traces for baseline (A), twitch (B), and tet-
during the direct stimulation of an attached neural extension (A) before and
(B) following the severing of the neural extension.
Representative tetanus recordings of nerve–muscle constructs
samples of adult rat soleus, rat E-15 fetal limb buds, and rat 3-d-old neonatal
limb muscle were run on all gels, and duplicate samples from each myooid
were separated on an 8% polyacrylamide-glycerol resolving gel and then
eletrophoretically transferred onto Immobilon polyvinyldifluoride membrane
(Millipore, Milford, MA). Immunoblotting was performed by using antibodies
against rat myosin heavy-chain (MHC) slow (Cat #-NCL-MHCs), MHC fast
(Cat #-NCL-MHCf), MHC neonatal (Cat #-NCL-MHCn), and MHC develop-
mental (Cat #-NCL-MHCd) obtained from (Novocastra Laboratories Ltd, New-
castle, United Kingdom), followed by chemiluminescent labeled IgG (Jackson
Laboratories, Bar Harbor, ME), and the chemiluminescent signal was moni-
tored, digitized, and analyzed using autoradiography. Myooid signals were
corrected for background and myosin heavy-chain signals were compared to
the adult soleus, fetal, and neonatal muscle signals.
Statistics. Values are presented as means ? SE. Statistical analysis was
performed by using Jump In 5.1 (SAS Institute Inc., Cary, NC). A one-way
analysis of variance was conducted to compare the differences between
myooids and nerve–muscle constructs. Differences were considered signifi-
cant at P ? 0.05.
Morphology of nerve–muscle construct. Approximately 1 wk after
the introduction of the neural explant, neuronal cells from the fetal
explants migrate away from the explant and move across the mono-
layers of myotubes (Fig. 3). Many of the neural cells fuse with the
myotubes (Fig. 4), and others form neuromuscular-like junctions
with clustering of acetylcholine receptors surrounded by neurofila-
ment-stained neural extensions (Fig. 5).
Contractile properties. A representative tracing from a nerve–mus-
cle constructs is shown in Figure 6. The nerve–muscle constructs
show the same response to field stimulation as previously described
for muscle-only constructs. The nerve–muscle constructs exhibit
spontaneous baseline activity and can elicit both a twitch and a
tetanus response to a field stimulation (Fig. 6). Following the field
stimulation of the nerve–muscle construct, microelectrodes were
used to electrically stimulate the neural extensions radiating from
the construct. The maximum tetanus generated from the neural ex-
tension was approximately 25% of the total tension elicited with
field stimulation, suggesting that the neural extension was recruiting
LARKIN ET AL.
muscle constructs. Values are means ? SE.
(A) Diameter, (B) twitch, and (C) tetanus of muscle and nerve–
approximately 25% of the myotubes in the construct during the
contraction (Fig. 7). Following the severing of the extension without
the removal of the microelectrode, the nerve–muscle construct re-
turned to spontaneous baseline contractile activity (Fig. 7). The con-
struct would no longer respond to the stimulation of the microelec-
The average diameters of the myooid and nerve–muscle con-
structs were not significantly different (Fig. 8). The average maxi-
mum twitch and tetanus generated by field stimulation was signif-
icantly greater in the nerve–muscle versus myooid (P ? 0.005 and
P ? 0.005, respectively).
Myosin heavy-chain profile. Western analysis of the myosin heavy-
chain content in the nerve–muscle versus myooid constructs is
shown in Figure 9. The sample from 3-d-old neonatal limb muscle
showed the greatest expression of myosin per gram of protein than
any other sample with a greater signal with the developmental an-
tibody and gave a moderate signal with the neonatal antibody. The
fetal E-15 limb bud sample only gave a signal with the neonatal
antibody. The expression of developmental myosin heavy chains
tended to be greater, while the neonatal myosin heavy chains tended
to be less in the nerve–muscle versus myooid samples, suggesting
that the myosin content per total gram of protein is greater in the
nerve–muscle constructs and that the myosin heavy chains present
are more neonatal-like than fetal.
We have constructed three-dimensional nerve–muscle constructs
from cocultures of myogenic cells from adult soleus muscle with E-
15 fetal spinal cord explants of rats. We have extended our muscle-
only ‘‘myooid’’ model to a new model with functional neural mus-
cular junctions and neural muscular projections that respond to
electrical stimulation by contraction. The introduction of the neural
cells to the muscle-only construct resulted in nerve–muscle con-
structs with functional neuromuscular junctions that generate more
force per gram of tissue than muscle-only constructs. In addition,
we have altered the myosin heavy-chain profile from that of a fetal
to a more neonatal phenotype as indicated by the increase in the
neonatal form of myosin heavy chain compared to muscle-only con-
The introduction of the neural cells to a monolayer of myotubes
resulted in the formation of neuromuscular-like junctions with clus-
tering of acetylcholine receptors surrounded by neurofilament
stained neural extensions. The costaining of neuronal derived tis-
sues with neurofilament and acetylcholine receptors with alpha-bun-
garo toxin allowed us to identify neuromuscular junctions. During
development, acetylcholine receptors are dispersed randomly along
the plasma membrane of myotubes (Burden, 1998; Sanes and Licht-
man, 1999). Introduction of the E-15 fetal rat spinal cord resulted
in the clustering of the acetylcholine receptors. In addition, neu-
ronal clefts and projections form around and away from the cluster
of acetylcholine receptors. Morphologically, the introduction of the
E-15 spinal cord resulted in what resembles architecturally a neu-
Western analysis of the myosin heavy-chain content in the nerve–
muscle versus myooid constructs indicates that the expression of
both developmental myosin heavy chains tended to be greater, while
the neonatal myosin heavy chains tended to be less in the nerve–
muscle versus myooid samples, and that there was a greater content
ENGINEERED NERVE–MUSCLE CONSTRUCTS
for (A) neonatal and (B) developmental myosin heavy-chain proteins. ‘‘N’’ is
a sample obtained from neonatal rat muscle, and ‘‘F’’ is a sample obtained
from limb buds of E-15 fetal rats. Average expression of developmental (B)
and neonatal (C) myosin heavy chain expressed as a percentage of the myooid
expression. Values are means ? SE.
(A) Western blot of muscle and nerve–muscle constructs probed
of neonatal versus fetal myosin heavy-chain isoform in the nerve–
muscle versus the muscle-only constructs.
The functionality of the nerve–muscle constructs, that is, the
ability to convey an action potential from the nerve across the junc-
tion to elicit a muscular contraction, was verified by stimulation
with specially designed microelectrodes. Using field stimulation,
which elicits an electrical stimulation along the entire length of the
muscle construct, we observed the same mechanical response as
previously described for muscle-only constructs. The nerve–muscle
constructs exhibited spontaneous baseline activity and both a twitch
and a tetanus response to field stimulation. The contractility of the
nerve–muscle constructs was greater than that of muscle-only con-
structs. We can therefore conclude either that the ratio of muscle
tissue to connective tissue is greater in the nerve–muscle constructs
or that the contractile machinery present is capable of producing
more specific force. The cumulative signal from the myosin heavy-
chain Western suggests that there is more muscle tissue per gram
of total tissue in the nerve–muscle constructs than the muscle-only
constructs. The diameters of the nerve–muscle versus muscle-only
constructs were not different, and the addition of the neural tissue
to the monolayer of muscle would suggest a decrease in muscle per
diameter of the construct. This would imply that the muscle present
in the nerve–muscle construct has a more adult-like phenotype be-
cause it is capable of producing more force per gram of muscle
tissue. These data are in agreement with a study by Wagner et al.
(2003) that studied the electrical current patterns generated by ace-
tylcholine receptors in a nerve–muscle coculture system. Following
3 wk of nerve–muscle coculture, Wagner showed a clustering of
acetylcholine receptors in the plasma membranes of the myotubes
and a shift to a more developed pattern of electrical currents. Using
patch clamp techniques for monitoring fetal versus adult acetylcho-
line activated currents in noninnervated and innervated myotubes
from nerve–muscle cocultures, Wagner showed a shift from a fetal
current in the noninnervated myotubes to an adult current in the
innervated myotubes (Wagner et al., 2003).
Following the field stimulation of the nerve–muscle construct,
microelectrodes were used to electrically stimulate the neural ex-
tensions radiating from the construct. The maximum tetanus gen-
erated from the neural extension was approximately 25% of the total
tension elicited with field stimulation, suggesting that the neural
extension was recruiting approximately 25% of the myotubes in the
construct during the contraction. We then mechanically severed the
neural extension without altering the position of the microelectrode,
and the nerve–muscle construct returned to spontaneous baseline
contractile activity. The muscle construct no longer responded to
stimulation of the nerve via the microelectrode, indicating that the
stimulation via the microelectrode was not due to the proximity of
the electrode to the construct but actually elicited the stimulation
along the neural projection. The increase in both twitch and tetanus
force elicited by field stimulation in the nerve–muscle versus the
muscle-only constructs further suggests that the sarcomeres present
in the nerve–muscle constructs are capable of producing more force
compared to muscle-only constructs. The inclusion of a functional
nerve–muscle interface in vitro greatly expands the potential to con-
trol the phenotype of the muscle tissue in culture and thus expands
the usefulness of engineered muscle for virtually all its possible
This research was supported by DARPA Biomolecular Motors Program,
USAF, AFOSR Award FA9550-05-1-0015.
LARKIN ET AL.
Baker, E. L.; Dennis, R. G.; Larkin, L. M. Glucose transporter content and
glucose uptake in skeletal muscle constructs engineered in vitro. In
Vitro Cell. Dev. Biol. 39A(10):434–439; 2003.
Bryers, P. S.; Ecob, M. S. Differentiation in dystrophic muscle cultures from
mice of different ages. Muscle Nerve 7(4):332–336; 1984.
Burden, S. J. The formation of neuromuscular synapses. Genes Dev.
Close, R. Effects of cross-union of motor nerves to fast and slow skeletal
muscles. Nature 206(986):831–832; 1965.
Close, R. Dynamic properties of fast and slow skeletal muscles of the rat
after nerve cross-union. J. Physiol. 204(2):331–346; 1969.
Dennis, R. G.; Kosnik, P. E., II. Excitability and isometric contractile prop-
erties of mammalian skeletal muscle constructs engineered in vitro.
In Vitro Cell. Dev. Biol. 36A(5):327–35; 2000.
Dennis, R. G.; Kosnik, P. E., II; Gilbert, M. E.; Faulkner, J. A. Excitability
and contractility of skeletal muscle engineered from primary cultures
and cell lines. Am. J. Physiol. Cell Physiol. 280(2):C288–C295;2001.
Dorchies, O. M.; Laporte, J.; Wagner, S.; Hindelang, C.; Warter, J. M.; Man-
del, J. L.; Poindron, P. Normal innervation and differentiation of X-
linked myotubular myopathy muscle cells in a nerve-muscle cocul-
ture system. Neuromuscul. Disord. 11(8):736–746; 2001.
Ecob, M. S. The application of organotypic nerve cultures to problems in
neurology with special reference to their potential use in research
into neuromuscular diseases. J. Neurol. Sci. 58(1):1–15; 1983.
Ecob, M. The location of neuromuscular junctions on regenerating adult
mouse muscle in culture. J. Neurol. Sci. 64(2):175–182; 1984.
Ecob, M. S.; Butler-Browne, G. S.; Whalen, R. G. The adult fast isozyme of
myosin is present in a nerve-muscle tissue culture system. Differ-
entiation 25(1):84–87; 1983.
Irintchev, A.; Rosenblatt, J. D.; Cullen, M. J.; Zweyer, M.; Wernig, A. Ectopic
skeletal muscles derived from myoblasts implanted under the skin.
J. Cell Sci. 111(Pt. 22):3287–3297; 1998.
Kobayashi, T.; Askanas, V.; Engel, W. K. Human muscle cultured in mono-
layer and cocultured with fetal rat spinal cord: importance of dorsal
root ganglia for achieving successful functional innervation. J. Neu-
rosci. 7(10):3131–3141; 1987.
Kosnik, P. E.; Faulkner, J. A.; Dennis, R. G. Functional development of
engineered skeletal muscle from adult and neonatal rats. Tissue Eng.
Sanes, J. R.; Lichtman, J. W. Development of the vertebrate neuromuscular
junction. Annu. Rev. Neurosci. 22:389–442; 1999.
Schiaffino, S.; Gorza, L.; Sartore, S.; Saggin, L.; Carli, M. Embryonic myosin
heavy chain as a differentiation marker of developing human skeletal
muscle and rhabdomyosarcoma: a monoclonal antibody study. Exp.
Cell Res. 163(1):211–220; 1986.
Schiaffino, S.; Murgia, M.; Serrano, A. L.; Calabria, E.; Pallafacchina, G.
How is muscle phenotype controlled by nerve activity? Ital. J. Neurol.
Sci. 20(6):409–412; 1999.
Strohman R. C.; Bayne, E.; Spector, D.; Obinata, T.; Micou-Eastwood, J.;
Maniotis, A. Myogenesis and histogenesis of skeletal muscle on flex-
ible membranes in vitro. In Vitro Cell. Dev. Biol. 26A(2):201–208;
Talmadge, R. J.; Roy, R. R. Electrophoretic separation of rat skeletal muscle
myosin heavy-chain isoforms. J. Appl. Physiol. 75(5):2337–2340;
Wagner, S.; Dorchies, O. M.; Stoeckel, H.; Warter, J. M.; 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. Pflugers Arch. 447(1):14–22; 2003.
Whalen, R. G.; Bugaisky, L. B.; Butler-Browne, G. S.; Ecob, M. S.; Pinset,
C. Synthesis and accumulation of myosin isozymes in tissue culture.
Adv. Exp. Med. Biol. 182:193–199; 1985.
Whalen, R. G.; Johnstone, D.; Bryers, P. S.; Butler-Browne, G. S.; Ecob, M.
S.; Jaros, E. A developmentally regulated disappearance of slow my-
osin in fast-type muscles of the mouse. FEBS Lett. 177(1):51–56;