|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
IDENTIFICATION OF DIFFERENTIALLY EXPRESSED PROTEINS AS A RESULT
OF RAF KINASE ACTIVITY
SARAH ANN REED
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Sarah Ann Reed
This document is dedicated to my Grandpa, who always believed I could do anything.
Foremost, I would like to thank my advisor, Dr. Sally Johnson, who believed in
me enough to first employ me and then accept me as a student. Her ever present
guidance, expert advice, and friendship were crucial to the completion of my degree. I
also thank my committee members, Dr. Lesley White and Dr. William Buhi, for their
excellent suggestions and ideas as well as for reviewing my thesis.
I could not have completed this research without the aid of my former and present
lab mates: Jeanine Page, Jessica Starkey, Xu Wang, Dane Winner, Jennelle McQuown,
Ju Li, and Sara Ouellette. Dr. Alan Ealy was gracious enough to assist with the statistical
analysis and introduce me to Dr. Johnson. In addition, Dr. Stan Stevens and Scott
McClung were extremely helpful in troubleshooting and assisting the protein
My Mom, Dad, and brother have shown more love and support than I could ever
ask for, for which I thank them endlessly. Most importantly, I thank my husband Jared
and my dogs, Annie and Bella. They have kept me smiling through the hard times and
have been sounding boards for ideas and frustrations. Jared has graciously dealt with my
unending pre-occupation, supported every decision I made, and kept me focused.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
LIST OF TABLES .................. .................. .................. ........... .............. vii
LIST OF FIGU RE S ................ ............................ ............ ........... .......... viii
A B S T R A C T ............................................ ... ......... ................................... x
1 IN TR O D U C TIO N ......................................................................... .... .. ........
2 LITER A TU R E R EV IEW ............................................................... ...................... 2
Satellite C ells: D definition and H istory................................... .................................... 2
D distinguishing Features of Satellite Cells................................... ....................... 5
c -M e t ................................................................................................... ...... . 5
P ax 7 ............................................................. . 6
C D 34 ........................................................ . 8
m -cadherin ................................................................................................... ........
M yocyte N nuclear F actor ............................................................ ............... 9
V-CAM 1 and N-CAM ................................. .......................... ......... 10
Syndecan 3/4 .......................................................................................... .............11
N o tc h .............................................................................1 3
Side P population C ells .............................................. 14
Satellite Cell Activation: Progression from Go to G1 ...................................... 15
H epatocyte Growth Factor ............................................................................. 15
A g e d ifferen ces...... .................................................................................... 17
E x e rc ise ............................................................................................................... 1 9
N itric O x ide .............................................................. ......................... 22
Notch ............................. .............................24
B asic Fibroblast G row th Factor ......................................................... ........... 25
Insulin-like G row th Factor 1.............................................. .......................... 29
Transform ing G row th Factor .................................................. ....................30
3 MATERIALS AND METHODS ........................................... .......................... 33
Myoblast Culture ................... ......................... ........... 33
R af Protein Expression ............................................... ........ .. ............ 33
Western Blots............................................. 34
A cidic P-galactosidase Staining..................................................................... .. ... ..34
Grow th A rrest and R recovery ......................................................... .............. 35
N nuclear Protein Extracts ..................................... .................................................. 35
Protein D esalting and Concentration................................................. ............... 35
Two Dimensional Polyacrylamide Gel Electrophoresis (2D-PAGE) ......................36
M ass Spectrophotom etry ................................................................. .....................37
Phosphorylated Protein Isolation.......................................... ........... ............... 39
Im m unocytochem istry ........................................................................ .................. 39
4 R E S U L T S .............................................................................4 1
Raf Induced Growth Arrest is Rapid and Reversible ..............................................41
Raf-Initiated ERK1/2 Activation is Sustained at 60 Minutes.................................45
Identification of Nuclear Proteins: Protocol Definition............... ...............46
R em oval of Salts...........................................................................................50
Isoelectric Focusing (IEF) .............................. ............................................50
Activation of Raf/ERK1/2 Causes Changes in Nuclear Protein Expression Profiles 51
Raf Signaling Causes Nuclear Translocation of E2F5 and LEK 1............................56
5 DISCUSSION .............. ...... ........................................ ......... 75
LEK1 Expression Changes in Response to Raf Induction .............. ............. ....77
E2F5 is Involved in Cell Cycle Exit....................................................81
pRb Localization Changes in Response to Raf Activity .................................. 90
Phosphorylated ERK1/2 Remains Cytoplasmic in Raf Induced Cells......................91
6 FU TU R E D IR E C T IO N S ........................................ .............................................94
Do ERK1/2, E2F5, pRb, and/or LEK1 Cooperate in Raf Induced Quiescence?........ 94
Do High Levels of Raf Activity Promote a More Naive State? .............................96
Does High Raf Activity Correlate with Quiescence in vivo? ............ ..................98
A PROTEINS IDENTIFIED IN VEHICLE-ONLY CELLS ............ ... ................. 102
B PROTEINS IDENTIFIED IN RAF INDUCED CELLS ..........................................105
L IT E R A T U R E C IT E D ......................................................................... ..................... 108
BIOGRAPHICAL SKETCH ............................................................. ............... 122
LIST OF TABLES
I Primary antibodies used for immunocytochemistry .......................................40
II Isoelectric focusing param eters. ........................................ .......................... 52
III Unique proteins identified from control and Raf-induced nuclear extracts ............59
LIST OF FIGURES
1 Location of m uscle satellite cells. ........ ......... ............ ............. .........3
2 High levels of Raf activity inhibit cell number increases. ............. ................. 42
3 High levels of Raf activity inhibit BrdU incorporation ........................... ........43
4 High levels of Raf activity result in morphological changes ................................44
5 Raf induced growth arrest is reversible.................. ............................................. 47
6 Raf induced growth arrest is negated by the presence of high serum
7 Tim e course of ERK 1/2 activity........................................ ........................... 49
8 Comparison of IEF protocols............................... ........... ............. 53
9 Representative pi 6-11 and pi 4-7 two dimension gels of control and Raf
induced cells. ..........................................................................54
10 C arbam ylyte tw o dim ension gel................................................................... ......55
11 Representative two dimension gel of nuclear extracts from control cells ..............57
12 Representative two dimension gel of nuclear extracts from Raf induced cells........58
13 Localization of LEK 1 and E2F5. .................................................. .....................61
14 L ocalization of E2F5 over tim e ..................................................................... ....62
15 Localization of phospho-ERK 1/2 over time. ................................... ..................... 63
16 No phosphorylated ERK1/2 is present in control or Raf induced nuclear extracts..64
17 The presence of a MEK inhibitor blocks the activation of ERK1/2. .....................65
18 Pocket protein expression in control and Raf induced cells ..............................68
19 Differences in pRb expression in the nucleus. ................................. ............... 69
20 Pocket protein expression in nuclear extracts of control and Raf induced cells......70
21 E2F4 expression does not change in response to Raf activity. .............................71
22 Inhibition of pERK1/2 blocks the translocation of E2F5 and pRb to the nucleus...72
23 M yogenin is not expressed in Raf induced cells .............. ............ .....................73
24 Recovery from ERK1/2 stimulation results in the partial restoration of
cytoplasmic location of E2F5 and pRb. ....... ........... ........... .. ............. 74
25 Schematic drawing of the conserved structures in the LEK family of proteins
(A) and potential interaction sites (B). ......................................... ...............79
26 Schematic representation of conserved E2F4 and E2F5 domains. ..........................83
27 Comparison of pRb, p130, and p107 conserved domains..................................86
28 Potential ERK1 binding domains and phosphorylation sites.............................92
29 Proposed model. ....................................... .... .. .... .............. .. 93
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
IDENTIFICATION OF DIFFERENTIALLY EXPRESSED PROTEINS AS A RESULT
OF RAF KINASE ACTIVITY
Sarah Ann Reed
Chair: Sally E. Johnson
Major Department: Animal Sciences
Satellite cells exist in the quiescent state in healthy muscle. Upon stimuli, they
become active and subsequently enter the cell cycle. In the absence of appropriate
mitogenic signals, these cells have the ability to exit the cell cycle and return to the
quiescent state, replenishing the satellite cell pool. The events that lead to reversible cell
cycle exit are largely unknown. 23A2RafERDD myoblasts withdraw from the cell cycle
upon initiation of high levels of Raf activity. Phosphorylation of ERK1/2 occurs within
the first hour of Raf activation. High levels of Raf activity induce a growth arrest within
24 hours without causing senescence or apoptosis. This growth arrest is reversible upon
removal of stimuli, resembling the process of satellite cell self-renewal. Two
dimensional PAGE followed by mass spectrometry identified changes in nuclear protein
expression after one hour of Raf activity. E2F5 and LEK1 translocate to the nucleus in
response to high Raf activity. This response is specific to ERK1/2 phosphorylation and is
a specific response of these proteins, as E2F4 expression profiles were unaffected. In
addition, pRb, a pocket protein capable of interaction with E2F transcription factors and
LEK1, also translocated to the nucleus in response to ERK1/2 activation. Interestingly,
two additional pocket proteins associated with maintenance of quiescence, p107 and
p130, are unaffected at this early time. These data indicate that the relocation of E2F5
and LEK1 to the nucleus may be an important early step in the progression from G1 to
Satellite cells are responsible for muscle growth and regeneration. These muscle
stem cells exist in a quiescent state in normal muscle tissue. Upon appropriate stimuli,
satellite cells activate prior to proliferation and subsequent fusion into myotubes. An
important characteristic of satellite cells is the ability to self-renew. Following activation
or proliferation, one or more daughter cells are capable of returning to the quiescent state
to replenish the satellite cell population. To elucidate the potential mechanism of self-
renewal, Raf induced quiescence was examined in 23A2RafERDD myoblasts. This cell
line contains a Raf molecule fused to a mutated estrogen receptor that responds
specifically to 4-hydroxytamoxifen. High levels of Raf stimulation inhibit the myogenic
program (139). In addition, these myoblasts also express Pax7 protein, suggesting that
they are more akin to muscle satellite cells than myoblasts (Ouellette et al. unpublished
The objectives of this work are to (1) characterize the cell cycle response to high
levels of Raf activity leading to quiescence and (2) identify changes in nuclear protein
expression that may play a role in cell cycle exit.
Satellite Cells: Definition and History
In 1961, Alexander Mauro described the first noted satellite cells. In electron
micrograph studies of frog skeletal muscle, Mauro discovered mononucleated cells
existing between the plasma membrane and basal lamina (Figure 1). These cells had a
low cytoplasm to nucleus ratio, and were fewer in number than normal myonuclei.
Mauro named these cells "satellite cells" due to their position relative to the muscle fiber
and proposed that these cells are "remnants from the embryonic development of the
multinucleate muscle cell which results from the process of fusion of individual
myoblasts." He put forth the theory that satellite cells were "dormant myoblasts that
failed to fuse" and were "ready to recapitulate the embryonic development of skeletal
muscle fiber when the main multinucleate cell is damaged" (89, p. 493-494).
Seventeen years later, Schultz et al. furthered the basic knowledge of the satellite
cell population. Confirming that these cells exist in a mononucleated state under the
basal lamina, this group found that upon 3H-thymidine perfusion into adult muscle, very
few (0.07%) incorporated radiolabeled nucleotides into DNA, indicating that few cells
existed in S phase. Together with the organelle poor cytoplasm and abundance of
condensed chromatin, this group concluded that satellite cells exist in a quiescent state
until needed (117).
Bischoff investigated the in vitro activation of satellite cells in their natural
position using isolated myofibers from rats. When isolated fibers were destroyed,
Figure 1. Location of muscle satellite cells. Arrow indicates satellite cell located adjacent
to the sarcolemma and under the basal lamina (141).
virtually all satellite cells (>97%) were mitotically active after 24 and 48 hours as
determined by 3H-thymidine incorporation. Subsequent fusion of these satellite cells into
elongated syncytia confirmed their myoblast nature and ability to recapitulate the entire
myogenic program. Because muscle injury stimulated satellite cell mitosis/proliferation,
Bischoff proposed that damaged muscle fibers release growth factors and/or mitogens
that enable cell proliferation. After 15 hours of exposure to soluble protein extracts of
crushed muscles (CME), satellite cells on living fibers became proliferative. This time
period shortened considerably to five hours on dead fibers. However, in the absence of
CME, no proliferation occurred, regardless of the amount of serum present. Satellite
cells on basement membrane shells that were stimulated briefly with CME and
subsequently incubated in basal medium proliferated during the following 48 hours,
indicating that CME is necessary for the initial activation of satellite cells, but not for
continued proliferation. Interestingly, contact with a live fiber is enough to reverse or
limit the cell cycle commitment of satellite cells (16).
Mauro found satellite cells to be fewer in number than normal myofiber nuclei.
Confirmation of this occurred in 1976, when Schmalbruch and Hellhammer found that
only 3.8% of nuclei were satellite cells in varied muscle groups of ten human subjects of
both sexes, aged 7-73 (116). In young rats (17-18 days), more satellite cells incorporated
thymidine after mincing than did satellite cells in old rats (35-40 days) (128). In Moss
and Leblond's study, 12% of satellite cells incorporated thymidine in 14-17 day old rats
following injury. These cells proceeded to divide and fuse to existing myofibers in
increasing numbers. This group concluded that upon injury, satellite cells are activated
and work to replace the damaged myonuclei something Mauro had theorized ten years
Distinguishing Features of Satellite Cells
The identification of satellite cells was initially based on anatomical location. In
order to study pure populations of satellite cells, these cells had to be distinguished from
other mononucleated cells residing in muscle tissue. To this end, a number of protein
markers have been identified that are specific to satellite cells alone or in combination
with other proteins. These markers include cell membrane receptors, signaling
molecules, adhesion molecules, and transcription factors. It is important that these
proteins be expressed specifically in satellite cells and not in the adjacent myofiber or
myonuclei to best identify satellite cells.
The first protein marker identified in satellite cells was the membrane receptor c-
Met. Originally identified as a protein encoded by the oncogene met, this tyrosine kinase
receptor is the receptor for hepatocyte growth factor (HGF) and is necessary for proper
migration of hepatocytes and liver regeneration (21, 65, 94). C-Met is a transmembrane
tyrosine kinase receptor composed of a and 0 chains linked by disulfide bonds. The a
chain is highly glycosylated and entirely extracellular while the 0 chain has a large
extracellular region, a transmembrane region and an intracellular tyrosine kinase domain.
Binding of the extracellular ligand results in autophosphorylation of intracellular tyrosine
residues and subsequent phosphorylation of downstream proteins (94). This receptor is
found on all quiescent satellite cells but is not expressed in myofibers. Satellite cells
identified by anatomic location were positive for c-Met transcripts (141). The c-Met (+)
cells had a narrow band of darkly stained cytoplasm located around the nuclei.
Following crush injury, mononucleated cells surrounding the necrotic fiber express c-Met
protein. C-Met colocalizes with hepatocyte growth factor (HGF) in actively proliferating
satellite cells and is responsible for conducting the HGF signal (132). Isolated cells
negative for the surface marker CD34 but positive for c-Met were capable of giving rise
to myotubes in culture (113). C-Met null mouse embryos exhibit no skeletal muscle
precursor cells in the limb buds or diaphragm, in contrast to the normal formation of the
axial muscles (19). This indicates that c-Met is necessary for proper limb and diaphragm
skeletal muscle formation.
A member of the paired box family of transcription factors, Pax7 is expressed in
proliferating primary myoblasts. Pax7 protein levels are down-regulated upon myoblast
differentiation (118, 119). Pax7 co-localizes with myostatin, c-Met, and m-cadherin in
satellite cells resting beneath the basal lamina (60, 84, 90, 111, 119). While Pax 7 is
restricted to satellite cells in the post-natal animal, not all satellite cells, identified by c-
Met expression, contain Pax7. The heterogeneity of Pax7 in daughter cells does not
correlate with differences in MyoD expression (96, 111). Satellite cells quickly down-
regulate Pax7 upon commencement of terminal differentiation (96). Myoblasts
immunostaining Pax7(+) and MyoD(-) may return to quiescence to replenish the satellite
cell pool while myoblasts that acquire MyoD proliferate and fuse to form myotubes
(147). Myofiber nuclei in chicks immunostain negative for Pax7 but positive for MyoD
and myogenin (60). Whether myoblasts express Pax7 and myogenin simultaneously is
the source of some debate. Olguin and Olwin found Pax7 and myogenin to be mutually
exclusive, while Pax7, Myf-5, and MyoD may be present in the same myoblast (96). In
contrast, Halevy et al. and Zammit et al. found few cells that were positive for both Pax7
and myogenin indicating that these may be an intermediate population that is exiting the
cell cycle and entering terminal differentiation (60, 147) .
In myogenic cell lines, quiescent, undifferentiated myoblasts appear to be uniquely
marked with high levels of Pax7. Overexpression of Pax7 in satellite cells induces cell
cycle exit, prevents incorporation of BrdU, decreases expression of MyoD, and prevents
the induction of myogenin. However, the loss of Pax7 does not induce differentiation,
indicating that other factors must be present or absent for myoblasts to commence
terminal differentiation. While myogenic conversion induced by MyoD is inhibited by
Pax7, MyoD remains nuclear. However, Pax7 was unable to inhibit the effects of a
MyoD-E47 heterodimer, indicating that Pax7 may interfere with MyoD function or
compete for proteins necessary for MyoD-dependent transcription (96).
Adult Pax7-/- mice are smaller in size and have smaller myofibers than wild type
counterparts (100). Fibers isolated from Pax 7 null mice give rise to no quiescent satellite
cells in culture systems. In addition, no quiescent satellite cells could be identified in
more than 300 sublaminar nuclei from gastrocnemius muscles in 7-10 day old knockout
mice or day 18 embryos (119). Oustanina et al. indicated that Pax7 null mice have
satellite cells, but these are very few and the numbers decrease with age (100). The
mitotically active satellite cells express MyoD and divide at rates comparable to wild
types suggesting that Pax7 may be needed for maintenance of the satellite cell population
rather than proliferation. Pax7-/- myoblasts demonstrate a reduced differentiation
capacity in vitro (100). Importantly, injury in adult Pax7 null mice indicates that
regeneration is not as efficient or complete as wild type or heterozygous individuals as
indicated by a lengthened time of healing and the presence of residual necrotic material.
This may be due to the decreasing number of satellite cells and/or inefficient
CD34 as a true marker of satellite cells remains controversial. CD34 positive cells
express no cardiac, hematopoietic, or skeletal muscle cell mRNA transcripts, indicating a
lack of lineage. By contrast, skeletal muscle progenitor cells lacking CD34 express c-
Met but no other skeletal muscle transcripts, indicative of a muscle cell lineage. After
14-21 days in culture, both CD34(+) and (-) myoblast populations fuse and exhibit
spontaneous contractile behavior, expressing MyoD, myocyte nuclear factor a (MNFa),
and desmin. Thus, both CD34 (+) and CD34 (-) progenitor cells may give rise to
myotubes in vitro (113). Single fiber isolates from mice contain mononucleated cells that
express both CD34 and m-cadherin. After five days of differentiation, some cells remain
mononucleated and express CD34 but are MyoD and Myf5 negative. After 48 hours in
culture, CD34 expression declines in satellite cells that migrate and have begun to
proliferate. This group did note that CD34, Myf-5, and m-cadherin do not mark all
satellite cells (15).
CD34 may mark a population of satellite cells that is less committed to the
myogenic lineage. Muscle-derived stem cells isolated from mice that express CD34 lack
Pax7 and m-cadherin expression. These cells are capable of contributing to regeneration
of mdx muscle at low passage numbers but lose this ability with time (39). In mouse
gastrocnemius sections, CD34 expressing cells constitute a small population of muscle-
derived cells located under the basal lamina co-expressing m-cadherin (85). Jankowski et
al. found CD34 expressing primary myoblasts to be more efficient at participating in
regeneration, although CD34(-) cells had a higher fusion index in vitro (68).
Interestingly, mononucleated cells remaining after cultures were induced to differentiate
expressed high levels of CD34. Myoblasts expressed similar levels of myogenic proteins
(MyoD, myogenin, MRF4, m-cadherin) regardless of the presence of CD34.
Several groups have identified the Ca-dependent cell adhesion molecule m-
cadherin in both quiescent and activated satellite cells (15, 31, 35, 84). Cells expressing
m-cadherin account for 5.3% of sublaminar nuclei in the six week old mouse hind limb
(31). The number of m-cadherin positive cells increases upon satellite cell activation
(35). Following cardiotoxin injury, cells staining positive and negative for m-cadherin
express MyoD, although cells lacking m-cadherin stain less intensely for MyoD. All
cells expressing Myf-5 also express m-cadherin (31). M-cadherin expressing cells also
express CD34 (15).
Myocyte Nuclear Factor
Expressed in cardiac and skeletal muscle, myocyte nuclear factor (MNF) was
identified in a screen of binding proteins for the myoglobin CCAC box. This
transcription factor has also been identified as Foxkl. Two forms of MNF were
identified as having reciprocal expression during regeneration (52, 144). MNFa is a
90kDa protein expressed in committed myoblasts and myotubes. A splice variant of
MNFa, MNFp expression peaks on day two of in vitro differentiation and then wanes
(144). Further characterization showed MNFP expression primarily in quiescent satellite
cells while MNFa occurred predominantly in proliferating cells. MNF protein is present
as early as 8.5 days post coitum (dpc) in the heart tube and rostral somites. By 10.5 dpc,
MNF is abundantly expressed in the myotome of the somite and cells migrating to the
limb bud and two days later can be found throughout the limb bud and axial musculature.
Expression in differentiated myotubes wanes around 16 dpc but persists in small
mononucleated cells in the limbs and body wall. Approximately 2-5% of sublaminar
nuclei in post-natal sections contain MNF. Electron microscopy of post natal sections
shows MNF expression in the nuclei of satellite cells as classified by location beneath the
basal lamina yet above the sarcolemma. Upon cardiotoxin injury, the number of
positively stained MNF cells expands and expression persists in the central nuclei of
regenerating fibers (53).
MNF null mice show significant growth retardation, at -60% the size of wild type
littermates. These mice also have impaired regeneration; three weeks following
cardiotoxin injury, a hypercellular myonecrotic response remained. The regeneration that
did occur was accompanied by extensive replacement of muscle tissue with fat (52).
Myoblasts isolated from MNF null animals can differentiate into myotubes but proliferate
much more slowly than heterozygous or wild type cells (52). In addition, progenitor cells
that are capable of differentiation are able to co-express MNFa and CD34 (113). When
crossed with mdx mice, the resulting MNF -/- mdx mouse is born even smaller than
MNF-/- or mdx littermates. Compared to littermates, these mice also are less active and
appear to be more fragile. Upon closer investigation, the chest wall and diaphragm have
widespread myonecrosis and fibrosis. MNF-/-mdx mice die within the first few weeks of
V-CAM1 and N-CAM
Two cell adhesion molecules, vascular cell adhesion molecule 1 (V-CAM1) and
neuronal cell adhesion molecule (N-CAM) have been identified in satellite cells (69, 88,
118). V-CAM1 is expressed in endothelial and skeletal muscle cells in adult animals, and
in skeletal muscle is limited to quiescent satellite cells and activated muscle precursor
cells in vivo. A possible interaction between cells staining positive for V-CAM1 and
infiltrating lymphocytes may increase cytokine accumulation, which is required for
satellite cell activation and muscle precursor cell proliferation and differentiation during
Little information is available regarding N-CAM expression in satellite cells.
Malm et al. used N-CAM (CD56) as a marker for activated satellite cells following
exercise in human skeletal muscle (88). Ectopic expression of human N-CAM in mouse
muscle resulted in an increased size of the neuromuscular junction and an increase in the
number of terminal sprouts in the absence of denervation (138). Ectopic N-CAM
enhanced regeneration of the neuromuscular junction following denervation by
increasing the length of the sprouts and the number of secondary sprouts.
Syndecan 3 and 4 are cell surface transmembrane heparin sulfate proteoglycans
involved in FGF signaling. Both proteins are implicated in satellite cell maintenance and
regeneration. Syndecan 3 immunostaining is present in the skeletal muscle of the mouse
forelimb throughout development and in young adult skeletal muscle. Syndecan 3
appears at embryonic day (dE) 14.5 in individual myoblasts and by dE18.5 individual
myoblasts and myotubes are outlined with syndecan 3. By neonatal day 2, syndecan 3 is
becoming localized to discrete sites at the myotube periphery and becomes even more
localized in young adult skeletal muscle (33). Syndecan 4 is present around the lumen in
developing myoblasts at dE14.5 and at the periphery of myotubes at dE18.5. While
expression becomes localized in young adult cells, it is not as restricted as syndecan 3.
Syndecan 3 and 4 are both associated with a subset of myonuclei. Co-staining
syndecan 3/4 expressing cells with c-met shows complete overlap of staining in quiescent
and proliferating cells. Laminin staining locates the cells between the basal lamina and
myotube. In young adult muscle syndecan 3 staining colocalizes with FGFR1, staining
satellite cells exclusively. These proteins are expressed prior to detectable MyoD
expression and remain after MyoD induction and proliferation. Primary mouse satellite
cells require heparan sulfate for normal proliferation. When signaling events requiring
heparan sulfate (including syndecan signaling) are blocked, activation and initiation of
myogenesis is delayed in vitro (33).
Syndecan 3 null mice have extensive fatty infiltrates between and within muscle
fibers. Quiescent satellite cells appear normal in morphology in terms of size, position,
and marker gene expression. However, the knockout mouse has approximately 6.7 times
the number of satellite cells/100 myofiber nuclei than the wild type mouse. There also is
an increase in the number of differentiated myonuclei compared to wild type but no
difference in the fiber number/muscle or the average fiber diameter. The appearance of
significantly more central nuclei in mature fibers suggests that regeneration is occurring
in the absence of experimental damage. These observations are consistent with the
defects in hind limb locomotion exhibited by these animals. Upon activation of satellite
cells, fewer satellite cells from syndecan 3 null mice expressed MyoD at 96 hours than
wild type. This lower expression is accompanied by a mislocation of MyoD to the
cytoplasm rather than the nucleus. Satellite cells from syndecan 3 null mice proliferate
extensively but form large, irregular syncytia rather than fibers. These cells do not
express MyoD or MyHC (34).
Syndecan 4 null mice also have adipose infiltrates between fibers, but unlike
syndecan 3 null mice, have none within the fibers. Satellite cells in syndecan 4 null mice
have decreased c-Met and syndecan 3 immunostaining along with a decreased thickness
and increased disorder of the basal lamina. Unlike the syndecan 3 null mouse, the
syndecan 4 null mouse has only a slight increase in central myonuclei above the wild
type mouse. There is no difference in the number of satellite cells or mature myonuclei
in syndecan 4 null mice and wild type. Satellite cells in these mice are delayed in
activation and proliferation, and after proliferation myoblasts appear in large, aberrant
clusters. Few cells express MyoD at 96 hours post stimulation and those that do express
the protein cytoplasmically. These results indicate that syndecan 4 is required for normal
activation and proliferation. Isolated colonies of satellite cells from syndecan 4 null mice
are delayed in adhesion, fail to express MyoD and MyHC, and fail to fuse. Addition of
exogenous heparan partially rescues the syndecan 3 null mouse by correcting the MyoD
localization but fails to correct syndecan 4 defects (34).
The Notch signaling pathway is an evolutionarily conserved signaling cascade
that is involved in tissue development in a multitude of organisms. The binding of an
extracellular ligand (such as Delta or Serrate) to a transmembrane receptor of the Notch
family results in the enzymatic cleavage of the intracellular domain of the receptor. This
cleavage product represents the active form of Notch and translocates to the nucleus
where it affects gene transcription through association with various transcription factors.
Muscle injury results in the increased expression of desmin and activated Notch 1 in the
injured tissue. However, desmin and Notchl are co-expressed in mononucleated cells
only at the site of injury. Therefore, muscle injury results in the activation of Notch 1
and expansion of Notchl and Desmin expressing myoblasts in vivo. Interestingly, Numb
is expressed asymmetrically in crescent shaped patterns. During cell division, Numb
localizes to one pole, thus only one daughter cell receives Numb. Myofibers were
isolated 72 hours after injury and cultured overnight. Numb and Pax3 (a premyoblast
gene) were mutually exclusive in satellite cells on the cultured myofibers. The exclusion
of Numb in Pax3 expressing satellite cells indicates that these cells are less committed to
a specific phenotype. In contrast, Numb expressing cells also expressed Pax7, suggesting
that Pax7 is a marker of a more committed cell. The expression of a constitutively active
Notchl resulted in the up regulation of Pax3 and the down regulation of Myf-5, MyoD,
and desmin as well as a modest reduction in Pax7. Over expression of Numb resulted in
the down regulation of Pax3 and the up regulation of Myf-5 and desmin. Notch 1
appears to promote a less committed myogenic phenotype while Numb may promote
progression along the myogenic lineage (29).
Side Population Cells
A fraction of muscle-derived stem cells, side population cells are theorized to
contribute marginally to muscle regeneration. This population of cells is derived from
bone marrow. Side population cells may reside in skeletal muscle but do not assume the
typical satellite cell position in healthy muscle tissue (9, 83). However, upon gamma
irradiation, bone marrow derived side population cells may reside in the satellite cell
niche and assume a myogenic lineage, expressing c-met and Myf-5 (83). Side population
cells injected into injured muscle contribute to regeneration (32, 41, 83). Isolated bone
marrow cells that are CD45Lin-c-kit+ contribute to muscle regeneration (41).
Expression of CD34 or Scal did not distinguish cells capable of assisting regeneration.
The side population cells residing in muscle are also capable of giving rise to
hematopoietic cells, unlike satellite cells (9).
Satellite Cell Activation: Progression from Go to Gi
In normal adult muscle, satellite cells exist in the quiescent state (Go). Upon
injury or hypertrophic stimulus, these cells are activated and enter the cell cycle. Satellite
cell activation occurs during the lag phase from the first extracellular signal to cell cycle
entrance. This phase is often incorrectly grouped with the proliferative state: they are
two separate stages. In essence, the activation phase is the preparation for cell cycle
entry. It is important to note that activation does not imply or lead to proliferation. Early
inhibition may return the satellite cell to the quiescent state.
Hepatocyte Growth Factor
The first satellite cell activating factor was identified by Richard Bischoff in
1986. The soluble material obtained from crushed rat leg muscles was capable of
inducing proliferation of satellite cells on single fibers in culture without stimulating
proliferation of the surrounding fibroblasts. Crushed muscle extract (CME) also
enhanced proliferation of myogenic cells isolated from 19 day old rat embryos and
promoted differentiation in these cells. Satellite cells stimulated with CME enter S phase
after 18 hours and proliferate with a generation time of 12 hours. When injected into the
flexor digitorum brevis (FDB) muscle of 1 week old rat pups, CME caused a significant
increase in muscle growth reflected by increased DNA content (18). Satellite cells on
living fibers proliferated 15 hours after CME exposure, but only 5 hrs after exposure on
degenerating fibers. Satellite cells on killed fibers that briefly were stimulated with CME
and then incubated in basal medium proliferated over the following two days. Contact
with a live fiber was enough to reverse or limit the cell cycle commitment (16). In
primary cultures of satellite cells isolated from three week and nine month old rats, CME
increased PCNA expression, indicating cell cycle re-entry (71).
Immunoblotting of CME identified two bands of hepatocyte growth factor (HGF),
a 90 kDa uncleaved and 60 kDa cleaved alpha chain. In fibers isolated from normal adult
rat tibialis anterior (TA) muscle, c-Met positive cells were elongated and located under
the basal lamina in the typical satellite cell position. In regenerating muscle, c-Met and
HGF co-localize to the cytoplasm of satellite cells with levels of both proteins decreasing
as the cells enter terminal differentiation (132). HGF is present in the extracellular
matrix of muscle (131). The HGF message was detectable from 12-72 hr in culture of
adult rat satellite cells. At 12 hrs, HGF appears to be predominantly located on the cell
surface but is intracellular at 72hr. Serum-free conditioned medium contained HGF,
indicating that it was synthesized, secreted, and dissociated from activated cells (122). In
the presence of anti-HGF, satellite cells stimulated with CME do not re-enter the cell
cycle at rates comparable to controls, indicating that HGF is responsible for activation.
HGF injection into uninjured rat hind limbs causes satellite cell proliferation (over 50%
more than that caused by IGF/FGF2) (132). Conditioned medium from stretched fibers
stimulated BrdU incorporation into satellite cells on unstretched fibers, indicating the
release of an activating factor. Addition of anti-HGF to the conditioned medium
inhibited S-phase entry of satellite cells on unstretched fibers (133). The addition of
exogenous HGF in isolated chick satellite cells increased satellite cell DNA synthesis and
resulted in an earlier entry into S-phase. Exogenous HGF also inhibited transcription
from muscle-specific reporter genes as well as MyoD and myogenin expression.
Inhibition of differentiation through these pathways suggests that HGF may regulate
differentiation through inhibition of the myogenic regulatory factors (49). Injection of
exogenous HGF on the day of muscle injury results in significantly more myoblasts
although there is no difference in the regenerative ability of the fibers. Multiple
injections of HGF (once daily for three days) result in a large area devoid of regenerated
myofibers. Those fibers that do regenerate have a significantly smaller cross sectional
area. After washout of exogenous HGF, muscle regeneration returns to normal (92).
Thus, while HGF promotes satellite cell proliferation, it inhibits differentiation, allowing
enlargement of the pool of muscle precursor cells.
Differences in the extent and the length of time for muscle regeneration in the
elderly are a common cause of age-related myopathies such as sarcopenia and
polymyositis. Several differences in the satellite cell population are readily apparent,
such as the lag time until activation and the size of the nuclear domain. Many age-related
myopathies are related to muscle atrophy and lack of regeneration. Old rats (24 mos)
incur more sarcopenia than adult (9 mos) or young (1 mos) rats (12). The relative muscle
weight (muscle weight/body weight) decreases from adult (6-20 mos) to elderly (20-32
mos) in rats. Hind limb suspension decreases muscle weight in young and old rats,
causing a muscle mass loss due to inactivity that is similar to the loss that occurs during
aging. Intermittent reloading of the suspended muscle decreases atrophy in rats of 6 and
20 months, but not of 32 months of age (50).
While the number of satellite cells present in young adult, adult, and old animals
remains the same, the number of cells that can be isolated from these animals tends to
decrease with age, most likely due to changes in activation. Because these animals have
the same number of satellite cells prior to activation, the difference in proliferative cells
is a decline in the activation of satellite cells, not the total number present (12, 20, 27).
The percent of quiescent cells per fiber was similar in all three groups (young adult,
adult, aged) (12).
Important in maintaining proper protein production, the amount of cytoplasm a
nucleus controls is termed the nuclear domain. Nuclear domain tends to differ in aged
and young muscle. The number of myonuclei per fiber is lower in younger individuals
than in older individuals, resulting in a larger nuclear domain in younger muscle (72).
Similarly, Gallegly et al. found that despite the decline in muscle size and strength in old
rats, the number of nuclei present remains similar between groups. The smaller nuclear
domain (nuclei/mg protein) is reflected in a lower myofiber cross sectional area per
In addition to the differences in nuclear domain size, satellite cells from older
animals take longer to become mitotic following stimuli. Satellite cells isolated from
young rats re-enter the cell cycle 24-30hr post plating, while satellite cells isolated from
old rats remain quiescent for 42-48 hours (12, 70, 112, 143). Go exit is accompanied by a
higher number of c-Met positive cells upon serum stimulation in young rats as compared
to old rats and an overall higher percentage of PCNA positive cells in young rats (12, 87).
Similar to rodents, no significant difference was found in the proliferation or fusion rates
of myogenic cells isolated from 60-69 and 70-79 yr old humans (20).
Conboy et al. completed a rather elegant set of experiments with parabiotic
pairings between young and old mice (28). The young partners in each pair were either
transgenic for green fluorescent protein (GFP) or expressed a distinct CD45 allele to
confirm blood chimaerism. Mice were paired isochronically (young-young and old-old)
or heterochronically (young-old). After five weeks of parabioses, a hind limb on each
mouse was injured. Five days after injury, young mice in isochronic and heterochronic
parabioses exhibit regenerating muscles as indicated by central nuclei in embryonic
myosin heavy chain (eMHC) expressing myotubes. Old isochronic pairs exhibited poor
muscle regeneration with few proliferating cells, prominent fibrosis, and incomplete
myofiber formation. Injured muscle in old mice (19-26 months) exposed to a young
systemic environment through heterochronic parabiotic pairing with a young mouse
regenerated similarly to injured muscle in young mice. This regeneration was due almost
exclusively from the activation of aged satellite cells, not from grafts from the young
The act of exercise evokes stress upon muscle fibers through stretching, as well as
physiological changes in pH, changes in growth factor concentration, and occasionally
injury to the myofibrils or myofibers. The amount of damage caused to the cell (and
therefore the level of satellite cell activation) is dependent upon the type of exercise as
well as the intensity and duration.
Both endurance and resistance training activate satellite cells. Mice participating in
daily exercise show a significant increase in the number of proliferating satellite cells.
Daily exercise resulted in the addition of significantly more new myonuclei than one bout
or no exercise (127). The number of satellite cells increased gradually (but significantly)
over 90 days of resistance training in young men and ten weeks of strength training in
young women (73, 74). After 30 days of detraining, satellite cell number gradually
decreased to pre-training levels. mRNA for the cell cycle markers cyclin Dl and p21
increased following the first 30 days of training and returned to basal levels during de-
training. Along with changes in satellite cell number and muscle cross sectional area,
acute resistance loading (RL) increased mechano growth factor (MGF) mRNA
expression, the muscle-specific form of IGF-1. Acute RL also induced an overall
increase in levels of cyclin Dl gene expression and inhibited myostatin expression (78).
There was no significant change in the number of myonuclei following resistance training
or during the detraining period in contrast to changes that occur during endurance
exercise. The gradual increase in the CSA gained during training (significant by 90 days)
decreased during detraining. The significant increase in cross sectional area (CSA)
coupled with little increase in myonuclear number resulted in a larger myonuclear
domain during training in the vastus lateralis (which consists of type I and type II fibers),
indicating that the number of nuclei present were capable of handling the increased
myonuclear domain (73). In a previous study, this group indicated that the fibers of the
trapezius in women increased by 36% following a ten week resistance exercise protocol.
The increased CSA was concurrent with an increase in myonuclei to CSA ratio in
accordance with the myonuclear domain theory. The differences between these two
studies may be the differences in the percentage of cross sectional area gained (17% in
the 2004 study, 36% in the 2000 study) (74).
Maximal voluntary exercise stimulus in humans is not enough to cause myofiber
lesions in most individuals, however it is enough for satellite cell activation. There was a
significant increase in the number of cells positive for N-CAM following a single bout of
high intensity exercise (four and eight days post exercise) concurrent with a significant
increase in the percent of mononuclear cells expressing Fetal Antigen 1 (FA1) in the
exercised leg. FA1 is expressed in undifferentiated mononucleated cells in fetal skeletal
muscle. Following high intensity eccentric exercise, there was a significant increase in
the number of cells positive for FA1. FA1 colocalizes with Pax7 in mononuclear cells in
damaged adult muscle fibers but is not present in healthy adult fibers (36).
Irradiation attenuates satellite cell proliferation in mouse hind limbs. In healthy
muscle, eccentric contraction-induced injury causes regeneration demonstrated by
centrally located nuclei. Muscles irradiated prior to eccentric contraction-induced or
freeze-induced injury had few fibers with central nuclei. Not only does irradiation affect
regeneration, it also prevents the recovery of muscle strength following injury. Even
after 35 days of recovery, muscle that was irradiated and then injured did not recover the
same amount of strength (compared to pre-irradiation/injury) as non-irradiated injured
Spinal cord transaction causes a rapid and progressive decrease in the cross
sectional area of soleus and extensor digitorum longus (EDL) in rats. The CSA of the
soleus declines more rapidly and to a greater extent than the EDL. Exercise following
spinal cord transaction results in an increase in the CSA of soleus fibers to nearly control
values but has no effect on the EDL. While all three fiber types are affected in the soleus
muscle, only types IIx and IIa are affected in the EDL but these fibers aren't stimulated
enough to offset the atrophy caused by a lack of neural input. Satellite cell activation
occurs in the soleus and EDL muscles following spinal cord transaction regardless of
exercise regime (42).
Both muscle contraction and stretching occur during exercise, so it follows that
mechanical stretch may activate satellite cells in a similar manner to the stretch and
contraction caused by exercise. Both Tatsumi et al. and Wozniak et al. noted that stretch
activated satellite cells on isolated fibers although not all cells were activated (133, 141).
Satellite cell activation peaked at thirty minutes and two hours after stretching (141). The
differences in activation resulting from stretch may occur because of differences in the
intensity or duration of stretch.
Recent work has focused on the interaction between nitric oxide (NO) release and
satellite cell activation. Anderson was the first to show an effect of NO on satellite cells.
Using L-Arginine, a stimulus for nitric oxide synthase (NOS) and L-NAME, a NOS
inhibitor, this group showed that the presence of NO (through the addition of L-Arg)
increased the cell yield (presumably from activated satellite cells) following muscular
injury. Treatment with L-NAME both restricted and delayed the normally rapid changes
in size and position of activated satellite cells. Thin cells that were c-Met positive but
expressed no HGF immediately following injury were present in L-NAME treated mice.
By 10 minutes post injury, enlarged, HGF expressing satellite cells were present but these
were fewer in number than in control muscles. The persistence of necrotic fiber
segments in mice supplemented with L-NAME during 6 days of repair indicated
inefficient regeneration (5). Inhibition of satellite cell activation by L-NAME was
restored with the addition of HGF (5, 133). The combination of CME and L-NAME
treatments increased activation above control levels, but only to approximately half the
level of CME alone. Addition of HGF or L-Arg increased activation two fold. Stretch
and injury both released HGF to bind the c-met receptor through a NO dependent
mechanism (4). Satellite cells on unstretched fibers were activated with SNP, a NO
generating compound. NO increased as early as 1 hr after muscle fiber stretch and
continued to be present as long as 20 hr (133).
On fibers isolated from the FDB, many c-Met(+) mononucleated cells remained
attached to the fiber for 44-48 hours post plating. However, some c-Met expressing cells
lifted from the fibers and entered the cell cycle. The percentage of proliferative cells
increased to 55% at 48 hours post plating. The addition of CME to fibers increased the
number of proliferating cells 76% by 48 hours post plating. While CME increased
activation by three fold, the addition of L-NAME decreased satellite cell activation and
Mdx mice treated with L-Arg and deflazacort (a corticosteroid used to treat
Duchene's muscular dystrophy) resulted in earlier myoblast differentiation. The positive
effects of deflazacort and L-Arg are positively correlated with NOS-Ii expression in
regenerating muscles. Deflazacort treatment alleviated dystrophy, decreased the central
nucleation index, and increased fiber diameter. NOS inhibition decreased myotube
formation and decreased or blocked the c-Met and myf-5 increase that occurs with
satellite cell activation (7).
Daily L-Arg injections over four weeks increased the nNOS activity and nitric
oxide levels in mdx mice, although there were no changes in the number of central nuclei.
The percent of fibers taking up Evans Blue Dye decreased from 10% to 4% in mdx mice
in the absence or presence respectively of L-Arg, indicating that L-Arg increases the
stability of the fibers. L-Arg protected mdx mice against damage caused by eccentric
contraction, had positive effects on the force frequency relationship, and increased
utrophin (a substitute for dystrophin) in mdx fibers (14).
Overload of the rat plantaris causes an increase in skeletal a-actinin and MyHC
type I mRNA as well as a higher expression of HGF, IGF-I, MGF, and VEGF mRNA.
TRIM (a specific inhibitor for iNOS and nNOS) doubled IGF-I and MGF mRNA
expression in overloaded but not control muscle and eliminated the increase of skeletal a-
actinin and MyHC type I mRNA seen in overloaded muscle. L-NAME and TRIM
increased phosphorylation of p70s6k in overloaded muscles which correlated to increased
protein synthesis (120).
In 2002, Conboy and Rando used a mouse myofiber explant system to monitor
satellite cell activation and proliferation. No bromodeoxyuridine (BrdU) incorporation
was seen before 24 hours in vitro. Proliferation gradually increased from 24-72 hr in
vitro and then increased dramatically between 72 and 96 hr in vitro. Full length Notchl
was present at 0 hours in vitro. Activated Notchl was undetectable at 0 hours, but
increased between 0 and 96hr. The Notchl ligand Delta also was undetectable at 0 hours
in vitro and increased until 96 hr in vitro. At 96 hours in vitro, Notchl, Delta, and Numb
were present in the majority of mononucleated cells associated with the myofiber.
Notchl activation and the increased expression of Delta are coincident with satellite cell
The Notchl/Numb balance also effects myoblast proliferation and differentiation.
Overexpression of Numb, an inhibitor of intracellular Notch, resulted in lower cell
proliferation, while a constitutively active Notchl enhanced proliferation of primary
mouse myoblasts. The constitutively active Notchl continued to promote proliferation
even upon serum withdrawal. In addition to increased proliferation, Notchl appears to
inhibit differentiation, significantly reducing the expression of eMHC and production of
multinucleated myotubes. The use of RNAi to inhibit Notchl signaling decreased the
rate of BrdU incorporation by two fold. These transient effects were most severe in
rapidly proliferating, recently established myoblast cultures. Cells deprived of serum
expressed more Numb and withdrew from the cell cycle (29).
Resting muscle expresses little of the Notch ligand, Deltal, but has increased
levels of Numb and baseline levels of Notchl regardless of the age of mouse from which
it was isolated. However, in response to injury Deltal expression increases in young and
adult satellite cells but fails to do so in old cells. The increase in Deltal is associated
with a decrease in Numb expression and an increase in proliferation (identified through
increased PCNA expression). When satellite cells are activated in culture, similar levels
ofNotchl are expressed, but cells from old mice have consistently lower levels of
activated Notch (27).
After injury, Delta expression in young animals is induced in satellite cells and at
the periphery of myofibers adjacent to the injury as well as in cells -300 [tm caudal to the
injury. Very little Delta up-regulation occurs in older animals and none occurs at a
distance from the injury site. Inhibition ofNotchl signaling by a soluble Jagged-Fc
fusion protein blocked satellite cell activation in young injured muscle and resulted in a
reduced number of regenerating myotubes. By contrast, activation of Notchl by an
antibody directed to its extra-cellular domain increased cell proliferation and inhibited
differentiation. Forced Notchl activation markedly improved regeneration of injured old
muscle and significantly enhanced formation of regenerated fibers (27).
Basic Fibroblast Growth Factor
Basic fibroblast growth factor (bFGF or FGF2) stimulates myoblast proliferation
while inhibiting differentiation into mature myofibers. bFGF represses differentiation
while stimulating proliferation in MM14 myoblasts and is 30 times more effective at
maintaining a proliferative state than acidic FGF. bFGF is required for the initiation of a
new cell cycle (26). MM14 and C2C12 myoblasts exhibit specific bFGF binding.
Differentiating cells lose expression of the fibroblast growth factor receptor and therefore
lose the ability to bind bFGF while serum-starved quiescent cells retain high levels of
FGFR on the cell surface (97). Unlike differentiating MM14 myoblasts, proliferating
MM14 myoblasts express bFGF mRNA. Addition of exogenous bFGF inhibits
differentiation. Transfection of a bFGF cDNA into MM14 myoblasts inhibits
differentiation and increases BrdU incorporation. Transfection also affects neighboring
cells which incorporated BrdU in the presence of bFGF expression, suggesting a
paracrine action (61). Addition of HGF and bFGF to C2C12 myoblasts has a synergistic
effect on proliferation (137). 23A2 myoblasts supplemented with TGF-3 or bFGF were
incapable of fusion. Removal of the inhibitory growth factor resulted in fusion and
troponin I expression (134).
Inhibition of FGF signaling through overexpression of a dominant negative FGF
receptor 1 (dnFGFR1) resulted in withdrawal from the cell cycle and the formation of
small myotubes containing few nuclei. A small number (7%) of cells continued to
proliferate, indicating the presence of FGF independent myoblasts. Inhibition of FGFR1
resulted in a decrease in muscle mass during embryonic development due to decreased
numbers of myoblasts. No change in fiber diameter was observed, however a 50%
decrease in packing density in dnFGFR1 limbs existed corresponding to an increased
space between fibers (46).
bFGF is a heparin binding growth factor that can signal through a variety of
intracellular signaling pathways. bFGF stimulated the mitogen activated protein kinase
kinase 1 (MEK1, a dual specificity protein kinase upstream ofERK1/2 in the
Raf/MEK/ERK pathway) but not ERK1/2 or S6 kinase in MM14 myoblasts. However,
addition of bFGF following ten hours of serum starvation resulted in the activation of
MEK, MAPK, and S6 kinase indicating that the response to this growth factor may be
dependent upon the extracellular environment (23). Syndecans are heparan sulfate
proteoglycans (HSPGs) implicated in the binding of growth factor and extracellular
matrix components. Proliferating but not differentiating MM14 myoblasts express
syndecan 3/4. Chlorate treatment reduces the low affinity binding of 125I-aFGF and 1251_
bFGF to HSPGs. This effect is reversed by the restoration of glycosaminoglycan
sulfation. Chlorate induces terminal differentiation in bFGF supplemented cells,
indicating that it blocks the bFGF driven inhibition of differentiation. Addition of
heparin to chlorate treated cells restores bFGF action (98).
In addition to affecting myoblasts, bFGF increases satellite cell proliferation,
promoting muscle regeneration (71, 86, 121, 143). Proliferating rat satellite cells express
FGFR1/2/3/4 mRNA with 1 and 4 being the most prominent. Addition ofbFGF to
isolated adult rat satellite cells elicited a greater mitogenic response than IGF-I or TGF-P.
The combination of bFGF and HGF acted in an additive manner with regard to
proliferation (121). Addition of bFGF increases proliferation and PCNA expression in
satellite cells isolated from three week old rats but not from 9 month old rats (71). High
affinity FGF binding did not occur until 42 hours post-plating in adult rat satellite cells,
occurring only 18 hours after plating in cells from young rats. The activated receptors
were capable of generating the intracellular signaling through tyrosine phosphorylation
characteristic of the bFGF signaling cascade (70).
In healthy mouse muscle, bFGF is expressed around the fiber periphery, around
myonuclei, and in non-muscle cells (6). Mononucleated cells in mdx mice are positive
for bFGF and myogenin (mgn), especially around and within damaged fibers (51).
Twelve hours after muscle injury, disrupted myofibers express bFGF, particularly in
regions of hypercontraction (6). There is a corresponding increase in the number of
mononucleated cells distal to the injury that express bFGF. This population likely
includes inflammatory cells. bFGF also is present in newly formed myotubes (51).
Injection of bFGF into the TA of male mdx mice at the time of the first round of
spontaneous necrosis results in enhanced satellite cell proliferation (86). The number of
regenerated fibers is positively correlated to the dose of bFGF. Addition of bFGF to
satellite cells isolated from rats increased the number of activated cells, possibly by
allowing an increased number of satellite cells to enter the cell cycle or shortening the
time it takes to undergo division thus, achieving more cells in the same amount of time
Faster proliferating satellite cells isolated from the pectoralis major of the turkey
expressed higher levels of bFGF and FGFR1 earlier than slower proliferating cells and
produced more HSPG. The faster proliferating satellite cells also showed a greater
mitogenic responsiveness to bFGF than slower proliferating cells. Expression of HSPG
decreased as differentiation proceeded (91).
Satellite cells isolated from the extensor digitorum longus and tibialis anterior of
rat hind limbs were most responsive to bFGF when added at a concentration of 2 ng/ml.
The addition of heparin to cultures supplemented with bFGF had no effect on the number
of PCNA (+) cells (77). As in myoblasts, bFGF stimulated a strong mitogenic signal in
isolated satellite cells and isolated myofibers by increasing proliferation and inhibiting
differentiation (3, 17, 58, 59, 146).
Chicken anterior latissimus dorsi muscle expressed decreased levels ofFGF1,
normal levels of bFGF, and increased levels of FGFR1 after 11 days of stretching when
compared to unstretched muscle. After 11 days of stretch, increases were seen
immunohistochemically in bFGF and FGF4 protein levels within the endomysium and
perimysium. Nuclear bFGF localization was identified in a specific population of
mononucleated cells at the fiber periphery. Chronic low frequency electrical stimulation
over five days resulted in a 3 fold increase in FGF1 and bFGF mRNA. In addition,
FGFR1 increased two fold, while FGFR4 only slightly increased in response to electrical
Insulin-like Growth Factor 1
Similar to bFGF, insulin like growth factor I (IGF-I) increases satellite cell
proliferation in vitro and in vivo (1, 3, 76). IGF-I stimulated proliferation of isolated
satellite cells even in the presence of TGF-P. However, TGF-P decreased the
proliferation rate in a dose dependent manner. In the absence of TGF-P, IGF-I greatly
stimulated satellite cell differentiation although high amounts of IGF-I could not induce
differentiation in the presence of TGF-P (3). The presence of IGF-I increased the
magnitude of the proliferative response elicited by bFGF as well as stimulating
differentiation alone (59). IGF-I binding protein (IGFBP) 3 and IGFBP 5 are produced
by satellite cells but not fibroblasts in vitro. IGFBP3 and 5 expression were increased by
addition of IGF-I, TGF-P, and bFGF. IGF-I was the most potent stimulus for IGFBP5,
while TGF-P greatly stimulated IGFBP3. IGF binding proteins extend the half life of
IGFs and serve as a reservoir for IGF-I in the blood (146). The addition of estrogen or
trenbolone to bovine satellite cell cultures resulted in increased amounts of IGF-I mRNA
and an increase in cell proliferation (76).
Infusion of IGF-I into rat TA muscles caused no systemic effects. IGF-I infusion
over two-three weeks resulted in a -9% increase in muscle mass as well as an increase in
muscle DNA content (1). Muscles recovering from atrophy (hind limb suspension)
regained more muscle mass than control recovering muscles when infused with IGF-I
during regeneration. Colonies from satellite cells isolated from atrophied muscles treated
with IGF-I were larger than colonies isolated from control atrophied muscles. Satellite
cells with the highest potential for replication appeared to be lost with repeated atrophy-
regeneration. The addition of IGF-I to cells after repeated atrophy-recovery resulted in
larger colonies with a greater potential for replication, indicating that IGF-I may help
restore the replicative capacity of these cells (24). In young rats, overloaded muscles
have increased IGF-I receptor mRNA. A significant increase in mechano growth factor
mRNA occurs in all ages following muscle overload (101).
While muscles injected with rAAV-IGF-I demonstrated increased muscle mass,
gamma irradiation prior to IGF-I injection suppressed the increase in size. A small
muscle mass increase did occur in muscles that were irradiated and subsequently injected
with IGF-I, indicating that not all of the hypertrophy was due to satellite cells. This
increased muscle mass may result from the ability of differentiated fibers to produce
more protein by increasing the myonuclear domain without the addition of new
Transforming Growth Factor P
One of the most potent inhibitors of myoblast proliferation and differentiation, the
transforming growth factor 0 (TGF-P) family of growth factors also elicit dramatic
effects on satellite cells. TGF-P greatly inhibits differentiation and to a lesser extent,
proliferation, of satellite cells (2, 47, 59, 107, 129). The inhibition of fusion caused by
TGF-P is dose dependent and reversible. Removal of the growth factor allows myoblasts
to fuse and express myosin heavy chain as well as creatine kinase (47, 129). While the
inhibition of fusion is not overcome by addition of IGF-I, bFGF is capable of overcoming
at least a portion of the inhibition of proliferation resulting from TGF-P activation (3).
In Sol 8 mouse myoblasts addition of anti-TGF-P was capable of blocking the
inhibition of differentiation resulting from addition of exogenous TGF-P (2).
Supplementation with bFGF or TGF-P blocked differentiation of C2C12 myoblasts.
These cells accumulated cyclin Dl protein, a cyclin active in mid-G1 of the cell cycle.
Cyclin D3 responded to bFGF or TGF-P in a biphasic manner, decreasing 2-4 hours post
stimulation and increasing 32-64 hours post stimulation. Ectopic expression of cyclin Dl
inhibited differentiation when expressed at low levels but had no effect on fusion at high
levels of expression. Activation of muscle gene transcription by myogenic basic helix-
loop-helix regulators also is prevented by the ectopic expression of cyclin D (107).
Both TGF-P and family member GDF8 (myostatin) increase IGFBP3 mRNA and
protein expression. IGFBP3 suppresses proliferation of porcine embryonic myogenic
cells. Antibody neutralization of IGFBP3 relieved 50-70% of the suppression of
proliferation caused by TGF-P or GDF8 (75).
Portacaval anastomosis (PCA) rats mimic the muscle atrophy seen in liver
cirrhosis. PCA rats exhibit decreased expression of myosin heavy chain, MyoD,
myogenin, Myf5, and PCNA protein levels. Myostatin levels are three times higher in
PCA than control mice, coinciding with increases in activinR2b, the myostatin receptor,
and CDKI p21 which mediates the effects of myostatin. Changes in mRNA levels are
consistent with the identified protein changes (38).
MATERIALS AND METHODS
Stock cultures of 23A2RafERDD embryonic mouse myoblasts (139) were
maintained on 10 cm plastic plates coated with 0.1% w/v gelatin and passage at
approximately 70-75% confluency. Cells were cultivated in basal medium eagle (BME)
supplemented with 15% v/v fetal bovine serum (FBS), 1% v/v L-glutamine, 1% v/v
penicillin/streptomycin, 0.1% v/v gentamycin reagent solution, and 10mM puromycin
and incubated in 5% CO2 at 37C. All cell culture media, supplements, and sera were
purchased from Invitrogen (Carlsbad, CA). Cells (3.5x 104) for immunofluorescence
were cultured on 35 mm glass-bottomed plates (World Precision Inst., Sarasota, FL)
coated with 10% v/v BD Matrigel Matrix HC (BD Biosciences, San Jose, CA).
Raf Protein Expression
23A2RafERDD myoblasts stably express a tamoxifen-inducible chimeric Raf
protein (139). The estrogen receptor-Raf kinase domain chimera is unstable in the
absence of the estrogen analog. Addition of 4-hydroxy-tamoxifen (4HT) binds to the
estrogen receptor and allows for a dose-dependent increase in Raf protein expression and
kinase activity. Sub-confluent 23A2RafERDD myoblasts were washed twice with
phosphate buffered saline (PBS), treated with 10lg/mL protamine sulfate (CalBioChem,
San Diego, CA) in serum-free BME for 10 minutes, and washed twice with PBS. Cells
were starved in serum free BME for one hour. For long term (>2hr) Raf induction, cells
were treated with 1 M 4-hydroxy-tamoxifen (4HT; Sigma, St. Louis, MO) in 2% FBS
BME. For short term induction (< 2 hr) cells were treated with 1 iM 4HT in serum free
BME. Control cells were maintained in the appropriate media supplemented with ethanol.
When necessary, cells were pulsed with bromodeoxyuridine (BrdU) during the last thirty
minutes of treatment and then fixed with methanol.
Following Raf activation, plates were washed twice with PBS. Cells were lysed in
4X SDS PAGE sample buffer. Lysates were briefly sonicated, heated for 5 minutes at
95C and stored frozen at -200C until further use. Equal amounts of protein were
separated through 10% polyacrylamide gels. Blots were incubated in 5% w/v non fat dry
milk (NFDM) in TRIS-buffered saline Triton X (TBS-T; 10 mM Tris, pH 8.0, 150 mM
NaC1, 0.1% v/v Tween 20) for thirty minutes at room temperature. Primary antibodies
were diluted in blocking solution and incubated with the blots overnight at 40C. After
washing three times for five minutes each in TBS-T, peroxidase conjugated anti-mouse
(or rabbit) antibodies diluted 1:5000 in blocking solution were added to the blots for sixty
minutes at room temperature. Blots were washed a further three times for five minutes
each in TBS-T. Immunoreactive complexes were visualized by enhanced chemi-
luminescence (Amersham Biosciences, Piscataway, NJ) and X-ray film.
Acidic P-galactosidase Staining
Following Raf activation, myoblasts cells were stained for acidic P-galactosidase
to visualize senescent cells. Myoblasts were fixed in 2% v/v formaldehyde, 0.2% v/v
glutaraldehyde in PBS for 4 minutes at room temperature. Plates were incubated
overnight at 370C in P-galactosidase staining solution (20% v/v citric acid sodium
phosphate solution [126mM sodium phosphate, 36.8mM citric acid, pH 6.0], 150mM
NaC1, 2mM MgCl2, 5mM potassium ferricyanide, 5mM potassium ferrocyanide, Img/ml
X-galactosidase, to 100ml with H20). Staining was viewed by light microscopy the
following day after three five minute washes with PBS.
Growth Arrest and Recovery
23A2RafERDD myoblasts were culture in 2% fetal bovine serum containing 1 [iM
4HT or vehicle only. After 48 hours, cells were washed twice with PBS and placed in
normal growth media (15% FBS BME, no 4HT) for 24 hours. Cells were pulsed with 10
iM BrdU for 30 minutes prior to fixing in methanol.
Nuclear Protein Extracts
Myoblasts were rinsed in cold TBS and scraped from the plates into hypotonic
buffer (25 mM Tris, pH7.5, 1 mM MgC12, 5 mM KC1, 0.05% v/v NP40, 5 mM
orthovanadate, 5 mM sodium fluoride, 5 mM pyrophosphate, 1 mM PMSF, 10 mg/ml
aprotinin). Cells were allowed to swell for 15 minutes on ice in polypropylene Falcon
tubes (Fisher, Pittsburgh, PA). Lysates were centrifuged at 3220 x g for 10 minutes at
4C in a swinging bucket rotor to recover nuclei. Nuclei were resuspended in high salt
buffer (20 mM Tris, pH8.0, 20% v/v glycerol, 300 mM NaC1, 1.5 mM MgC12, 200 [iM
EDTA, 1 mM DTT, 1 mM PMSF, 10 mg/ml aprotinin, 5 mM orthovanadate, 5 mM
pyrophosphate, 5 mM sodium fluoride, 1% v/v NP40, 0.01% v/v SDS) and rocked at 4C
for one hour. DNA and debris were pelleted for 15 minutes at 16,100 x g at 40C. The
supernatants containing the nuclear proteins were frozen at -800C until further use.
Protein Desalting and Concentration
Resin from D-Salt Excellulose Desalting columns (Pierce, Rockford, IL) was
added to small spin columns (Pierce, Rockford, IL). The resin was centrifuged at 1500 x
g for one minute. Isoelectric focusing (IEF) sample buffer (4% w/v CHAPS, 7M urea,
2M thiourea) was passed through the column at 1500 x g for one minute. The column
wash was repeated twice, with the final spin lasting two minutes. Nuclear proteins were
applied to the resin bed and the column was centrifuged at 1500 x g for two minutes.
Eulates were stored at -80 oC or used immediately for isoelectric focusing. Alternatively,
salts were removed from the nuclear protein extracts by dialysis (Slide-a-lyzers; Pierce,
Rockford, IL) into water followed by IEF sample buffer.
Nuclear proteins were concentrated by acetone precipitation. In brief, an equal
amount of acetone was added to the extracts and proteins were precipitated at -200C for
two hours. Proteins were collected by centrifugation at 16,100 x g for 10 minutes at 40C.
The pellet was washed in 70% ethanol, vortexed and centrifuged at 16,100 x g for 10
minutes at 40C. After a second wash, the pellet was allowed to air dry before
resuspension in IEF sample buffer at 25% of the original volume. Protein concentration
was measured using the Bradford method (Pierce, Rockford, IL).
Two Dimensional Polyacrylamide Gel Electrophoresis (2D-PAGE)
Isoelectric focusing was performed using the Ettan IPGphor II system (Amersham
Biosciences, Piscataway, NJ) Precast gel strips with a fixed pi range of 4-7 or 6-11
(Amersham Biosciences, Piscataway, NJ) were loaded with 250 [g of protein for 13 cm
gels and 100 [g of protein for 7 cm gels in IEF sample buffer plus 0.4% v/v IPG buffer
(Amersham Biosciences, Piscataway, NJ) and 0.01% bromophenol blue. IEF strips were
placed in the electrophoresis unit and proteins were separated as follows: (1)
Rehydration, 16 hours, 50 V; (2) 500V, 500Vhr; (3) 1000V, 1,000Vhr; and (4) 8000V,
16,000Vhr. Maximum amperage of 50 [A and temperature of 200C were maintained
throughout the course of IEF. Focused gel strips were equilibrated with equilibration
buffer (50 mM Tris HC1, pH8.8, 6 M Urea, 2% w/v SDS, bromphenol blue) for 15
minutes with 1% w/v dithiothreitol (Fisher, Pittsburgh, PA) and alkylated for 15 minutes
with 2.5% w/v iodoacetamide (Sigma, St. Louis, MO). The gel strips were overlayed on
to 10% SDS polyacrylamide gels and sealed in place with 1% w/v agarose in SDS-PAGE
running buffer. Proteins were separated through the polyacrylamide gels at constant
amperage (30 mA). Subsequently, gels were fixed in 20% v/v methanol, 10% v/v acetic
acid for one hour at room temperature. Proteins were detected by modified silver
staining methodology. Gels were incubated sequentially with gentle shaking as follows:
(1) 30% v/v methanol, 2% w/v sodium thiosulfate and 6.8% w/v sodium acetate for one
hour (2) washed five times for eight minutes each in ddH20 (3) 0.25% w/v silver nitrate
solution for one hour (4) washed four times for one minute in ddH20 (5) 0.025% (w/v)
sodium carbonate, 0.004% formaldehyde (37% stock) until satisfactory color is attained.
Development was stopped with ethylenediaminetetraacetic acid (EDTA; 3.65g EDTA
w/v in 250 ml water) for 45 minutes. Gels were then washed in water.
Proteins of interest were rated based on the intensity of staining as determined by
comparison with equally stained gels. Intensely stained proteins were excised and
subjected to in-gel tryptic digestion prior to analysis by MALDI-TOF or MALDI MS/MS
by the University of Florida Protein Chemistry Core. The type of analysis was
determined by the perceived staining intensity of the proteins of interest. For intensely
stained proteins, mass spectrometric analysis of the tryptic digests was accomplished by a
hybrid quadrupole time-of-flight instrument (QSTAR, Applied Biosystems, Foster
City, CA) equipped with the o-MALDI ionization source. A two-point mass calibration
was performed in MS/MS mode of operation using the known fragment ion masses of
[Glu]-Fibrinopeptide (m/z 175.119 and m/z 1056.475). Peptide mass fingerprint
data generated via the QSTAR were searched against the NCBI nr sequence database
using the Mascot (Matrix Science, Boston, MA) database search engine. Probability-
based MOWSE scores above the default significant value were considered for protein
identification in addition to validation by manual interpretation of the mass spectra.
Proteins stained with moderate intensity were analyzed by capillary rpHPLC
separation of protein digests performed on a 10 cm x 75 um i.d. PepMap C18 column
(LC Packings, San Francisco, CA) in combination with a home-built capillary HPLC
System operated at a flow rate of 200 nL/min. Inline mass spectrometric analysis of the
column eluate was accomplished by a quadrupole ion trap instrument (LCQ,
ThermoFinnigan, San Jose, CA) equipped with a nanoelectrospray source. Fragment ion
data generated by data dependent acquisition via the LCQ were searched against the
NCBI nr sequence database using the SEQUEST (ThermoFinnigan) and Mascot (Matrix
Science, Boston, MA) database search engines. In general, the score for SEQUEST
protein identification was considered significant when dCn was equal to 0.08 or greater
and the cross-correlation score (Xcorr) was greater than 2.2. MASCOT probability-based
MOWSE scores above the default significant value were considered for protein
identification in addition to validation by manual interpretation of the tandem MS data.
The least intensely stained proteins were analyzed by capillary rpHPLC separation
of protein digests performed on a 15 cm x 75 um i.d. PepMap C18 column (LC Packings,
San Francisco, CA) in combination with an Ultimate Capillary HPLC System (LC
Packings, San Francisco, CA) operated at a flow rate of 200 nL/min. Inline mass
spectrometric analysis of the column eluate was accomplished by a hybrid quadrupole
time-of-flight instrument (QSTAR, Applied Biosystems, Foster City, CA) equipped with
a nanoelectrospray source. Fragment ion data generated by Information Dependent
Acquisition (IDA) via the QSTAR were searched against the NCBI nr sequence database
using the Mascot (Matrix Science, Boston, MA) database search engine. Probability-
based MOWSE scores above the default significant value were considered for protein
identification in addition to validation by manual interpretation of the tandem MS data.
Phosphorylated Protein Isolation
Phosphorylated proteins were purified using the PhosphoProtein Purification Kit
(Qiagen, Valencia, CA) according to manufacturer's recommendations. In brief,
myoblasts were rinsed twice with ice-cold TBS and scraped into TBS. Cell pellets were
collected by centrifugation and resuspended in lysis buffer. Lysates were incubated at
4C for 30 minutes with vortexing every 10 minutes. Cell debris and insoluble material
was removed by centrifugation at 10,000 x g for 30 minutes at 40C. The supernatant was
collected and diluted (1:5) with lysis buffer before application to immobilized metal ion
resin. The column resin bed was washed with lysis buffer once before elution of the
phosphoproteins. The eluate was concentrated and exchanged into 10mM Tris, pH 7.0
using Nanosep columns (Qiagen, Valencia, CA). Protein concentration was determined
using the Bradford method. Proteins were frozen at -200C until use.
23A2RafERDD myoblasts were fixed with 4% v/v paraformaldehyde for 20
minutes at room temperature. For the detection ofLEK1, myoblasts were fixed with 70%
v/vethanol for 15 minutes at room temperature. Non specific antigen sites were blocked
by incubation in 5% v/v horse serum, 0.1% v/v Triton X-100 in PBS for 60 minutes at
room temperature. Antibody dilutions, sources, and conditions are listed in Table I.
After exhaustive washes with PBS, AlexaFluor 488 conjugated anti-mouse and anti-
mouse antibodies diluted 1:500 were added for 60 minutes at room temperature. Hoechst
dye was included as a nuclear stain. After washing with PBS representative images were
captured using a DM1200F digital camera.
Table I. Primary antibodies used for immunocvtochemistry.
Antigen Antibody Type Source Dilution
Mouse monoclonal IgGi
Mouse polyclonal IgGi
Mouse monoclonal IgG
Mouse monoclonal IgGI
Raf Induced Growth Arrest is Rapid and Reversible
Previous work from this lab has demonstrated that high levels of Raf activity
inhibit mouse myogenesis (139). To clarify the mechanisms behind this effect,
23A2RafERDD myoblasts were treated with 1 pM 4HT in low serum media for 48 hours
to induce high levels of Raf protein expression and kinase function. After 48 hours,
control and treated myoblasts were fixed and cell numbers measured. Total cell number
in plates treated with 4HT remained similar to t=0 hour plates, while the cell number in
vehicle-only plates increased two-fold (Figure 2, p<0.01). Total cell number in 4HT-
treated plates never decreased below the number of cells at t=0 indicating Raf signaling
during the treatment period was not lethal. To determine how soon the negative effect on
mitosis occurs, myoblasts were treated with 1 pM 4HT for 24 hours and pulsed with 10
IM bromodeoxyuridine (BrdU) during the final thirty minutes. Myoblasts were fixed
and immunostained for BrdU. 23A2RafERDD myoblasts containing elevated amounts of
Raf incorporate significantly less BrdU than control myoblasts, indicating fewer cells in
S phase (Figure 3, p<0.02). Coincident with the abrupt cessation of proliferation are
subtle changes in cell morphology. Control and 4HT treated 23A2RafERDD myoblasts
were fixed and stained with phalloidin to visualize the actin cytoskeleton (Figure 4).
Myoblasts that synthesized abundant Raf kinase display an ordered, distinct localization
of actin filaments. Control myoblasts posses a more diffuse acting staining pattern
surrounding the periphery of the cell.
20 "I (-) 4HT
1 (+) 4HT
0 hr 48 hr
Figure 2. High levels of Raf activity inhibit cell number increases. 23A2RafERDD
myoblasts treated with 1 [iM 4HT or vehicle-only for 48 hours in low serum
media do not proliferate to the same extent as untreated cells. Asterisk
indicates significant difference, p<0.01.
1 0 (-) 4HT
E (+) 4HT
0 hr L-12 hr -J-- 24 hr -
Figure 3. High levels of Raf activity inhibit BrdU incorporation. 23A2RafERDD
myoblasts were treated for 12 or 24 hours with 1 iM 4HT or vehicle-only in
low serum media and pulsed with BrdU for the last thirty minutes of the
incubation. Letters indicate significant difference, p<0.02.
Figure 4. High levels of Raf activity result in morphological changes. 23A2RafERDD
myoblasts treated with 1 [iM 4HT or vehicle-only for 48 hours in low serum
were subsequently immunostained for phalloidin, highlighting the
Elevated Raf signaling in human fibroblasts causes senescence typified by
irreversible growth arrest. Myoblasts senescence was measured in 23A2RafERDD cells
treated with 4HT or vehicle only. After 48 hours in low serum media in the presence of
4HT, the culture media was replaced with fresh growth media containing 15% FBS.
Cultures were fixed and myoblast cell number measured at 24 hour intervals. After 48
hours in 4HT, cell number was not different from t=0 (Figure 5). Control myoblasts
doubled over the 48 hour time-frame. Replacement of media with fresh growth media
lacking 4HT caused a two-fold increase in cell number (p<0.01). In data not shown,
23A2RafERDD myoblasts treated for 48 hours with 1 pM 4HT failed to express detectable
amounts of acidic galactosidase, a hallmark of cellular senescence. Satellite cells exiting
the cell cycle enter the quiescent state, from which cell cycle re-entry is possible.
Alternatively, myoblasts exiting the cell cycle often enter the senescent state, in which
proliferation is no longer an option. Interestingly, high serum medium prevents Raf-
induced growth arrest. Myoblasts were treated with 4HT in 15% FBS for 48 hours.
Cultures were fixed at 24 hour intervals and cell number measured (Figure 6). No
difference in cell number was apparent between control and 4HT treated myoblasts,
indicating that high serum concentration is sufficient to prevent the growth arrest caused
by elevated levels of Raf activity.
Raf-Initiated ERK1/2 Activation is Sustained at 60 Minutes
The rapid reversible growth arrest imposed by Raf is a similar feature of
myogenic stem cells. Skeletal muscle stem cells or satellite cells exist in mature muscle
as quiescent precursors that reactivate the myogenic program during periods of growth
and regeneration. Thus, 23A2RafERDD myoblasts may be a useful model to examine
early events during entry into Go or self-renewal. To clarify the early events associated
with Raf induced quiescence, the temporal activation of ERK1/2 was examined.
23A2RafERDD myoblasts were treated with 1 [tM 4HT and proteins from three replicates
of parallel plates were isolated at 15 minute intervals. Total cell lysates prepared from
equal numbers of cells were electrophoretically separated through 10% polyacrylamide
denaturing gels. Proteins were transferred to nitrocellulose and analyzed for total and
phospho-ERK1/2 expression by Western blot. No phosphorylation of ERK1/2 was
observed in 23A2RafERDD cells treated with vehicle-only (Figure 7, representative blot).
Phosphorylation of ERK1 occurs as early as 15 minutes in 23A2RafERDD cells treated
with 4HT and continues to be phosphorylated over the next two hours. ERK2
phosphorylation occurs at 30 minutes and is strongly phosphorylated at one hour.
ERK1/2 phosphorylation does not occur in 23A2 myoblasts treated with 4HT (139).
Sustained activation of both ERK1 and ERK2 was detected at 60 minutes post Raf
induction. Based upon the strong sustained phosphorylation of ERK1/2 at 60 minutes,
this time point was selected for the identification of proteins uniquely expressed in the
nucleus during entry into Go. The preferred method of protein identification involves 2D-
PAGE separation and MALDI-TOF or MALDI MS/MS identification. Several
parameters were refined to accomplish our goals.
Identification of Nuclear Proteins: Protocol Definition
Because the use of two-dimensional polyacrylamide gel electrophoresis was a new
technique for this laboratory, it was necessary for the protocol to be refined to suit our
needs. Protocol definition included sample preparation, isoelectric focusing and
subsequent vertical PAGE, and silver staining techniques.
(-) 4HT O
Figure 5. Raf induced growth arrest is reversible. 23A2RafERDD myoblasts were treated
with 1 iM 4HT or vehicle-only in low serum media for 48 hours. Treatment
media was replaced with normal growth media lacking tamoxifen. Arrow
indicates replacement of low serum treatment media with high serum growth
media. Different letters indicate significant difference, p<0.05.
O (-) 4HT
S6 (+) 4HT *
0 hr -24 hr- L-48 hr-
Figure 6. Raf induced growth arrest is negated by the presence of high serum
concentrations. 23A2RafERDD myoblasts were cultured in high serum media
with 1 lM 4HT or vehicle-only. Cell numbers have been normalized to one
based on the t=0 cells. Asterisk indicates significant difference, p<0.01.
0 15 30 60 120
0 15 30 60 120
|. i~-4 |.,*w ..
I 4trW 4
I eP ft u** a-Tubulim
Figure 7. Time course ofERK1/2 activity. 23A2RafERDD myoblasts were treated with
1 iLM 4HT (Raf-Induced) or vehicle-only (Growth Arrested) over the times
shown. Cell lysates were probed for phosphorylated and total ERK1/2 and
tubulin (loading control).
Removal of Salts
Proteins from control and Raf induced nuclei were extracted using 3M NaC1, a
solute incompatible with isoelectric focusing and 2D-PAGE. Three methods of salt
removal were tested. First, nuclear proteins were precipitated with an equal volume of
ice-cold acetone. Recovery of proteins using this method was approximately 20% of
original concentration. Second, proteins were dialyzed against IEF sample buffer. This
method efficiently removed salt and retained protein. However, the final protein
concentration was too dilute and required an additional concentration step. The third, and
preferred, method of salt removal involves the use of D-Salt Excellulose resin. Recovery
of protein was similar to dialysis with regard to final amount of protein recovered. This
method was chosen due to the ease of use and time savings.
Isoelectric Focusing (IEF)
Separation of proteins based upon their isoelectric point was accomplished using
the Ettan IPGphor system. Two protocols were tested to determine optimum focusing
times and voltages using 3-10 non-linear pi gradients. The first protocol used a fixed
time period protocol and the second protocol used fixed voltages (Table II). Identical
samples (250 pg) were electroporetically separated using each protocol. Proteins were
visualized by silver stain. The fixed time protocol occasionally produced gel strips that
failed to reach maximum voltage and resulted in poor protein focusing and separation
(Figure 8). The fixed voltage protocol resulted in sharper protein spots and less vertical
smearing. Therefore all experimental isolations and 2D-PAGE were performed at fixed
The overwhelming number of protein spots on each gel posed a limitation to
excision of single proteins for further identification. Therefore, pi 4-7 and pi 6-11
Immobiline dry strips were tested for focusing sharpness as well as protein patterns. pi 6-
11 gel strips were difficult to electrophoresis, often failing to reach maximum voltage or
taking an extraordinary amount of time to resolve. The proteins on pi 6-11 gels resolved
poorly and two vertical streaks of proteins were evident at -pI 8 and pi 10 (Figure 9).
Separation of proteins with a pi 4-7 gradient resulted in a sharper resolution of protein. A
vertical streak was present on the extreme basic side of the gel, representing the
accumulation of proteins with pIs outside of the covered range.
To estimate the molecular weight and isoelectric points of proteins of interest,
carbamylytes were run in parallel with nuclear protein samples. Calibration standards
included carbonic anhydrase (CA; MW 30kDa, pi range 4.8-6.7), creatine phosphokinase
(CP; MW 40kDa, pi range 4.9-7.1), and glyceraldehyde-3-phosphate dehydrogenase
(GAPDH; MW 36kDa, pi range 4.7-8.3). Carbamylytes are trains of the same protein
with different isoelectric points. The positively charged amino groups of each protein
interact with urea upon heating, forming protein derivatives with a lower pi. These three
carbamylytes resulted in protein trains of 20-30 spots each, providing a good estimate of
pi values (Figure 10). In summary, the final protocol for purification, electrophoresis and
visualization of nuclear proteins uses desalting resin, pi 4-7 IEF gel strips, electrophoresis
at a constant voltage (Table II), and silver staining.
Activation of Raf/ERK1/2 Causes Changes in Nuclear Protein Expression Profiles
To determine specific changes in protein expression, nuclear proteins isolated
from control and Raf-induced cells were separated by 2D PAGE and visualized by silver
staining. Gel images were analyzed by Melanie (GeneBio), a software program that
allows comparisons of 2D PAGE pattern profiles. Protein landmarks were used to align
gels, orient molecular weight, and calculate pi. A number of similar proteins were
present in high levels in both the control and treated gels, making them ideal as
landmarks. Unique proteins were excised from the control and treated gels (Figures 11
and 12). A total of 57 protein spots were excised from the treatment groups, 24 from the
Table II. Isoelectric focusing parameters.
Rehydration 50V, 16 hours
Step One 500V, 1 hour
Step Two 1000V, 1 hour
Step Three 8000V, 2 hours
50V, 16 hours
Fixed Time Period
Figure 8. Comparison of IEF protocols. First dimension separation protocols were tested
for optimum focusing conditions. The fixed time period protocol was based
on maximum amperage for a pre-determined amount of time. The fixed
voltage protocol used a set amount of volt hours and maximum voltage.
CTL pill p16
Figure 9. Representative pi 6-11 and pi 4-7 two dimension gels of control and Raf
induced cells. 23A2RafERDD myoblasts were treated with 1 giM 4HT or
vehicle-only for one hour. Nuclear proteins were isolated and separated over
pl6-11 or p14-7 first dimension gels and subsequently separated down a 10%
Figure 10. Carbamylyte two dimension gel. Creatine phosphokinase (CP),
glyceraldehydes-3-phosphate dehydrogenase (GAPDH), and carbonic
anhydrase (CA) carbamylytes were separated over a 10% SDS gel and silver
-- ~ 4--
Raf-induced proteins and 33 from the control proteins. A minimum of four gels were
analyzed for each group (control, 4HT). Proteins were digested in gel with trypsin and
peptides were identified by MALDI-TOF and MS/MS at the University of Florida Protein
Chemistry Core facility. Identification was verified by the presence of the protein in at
least three gels in the correct location relative to pi and molecular weight. Twelve unique
proteins were identified in the Raf-induced cells and ten unique proteins were identified
in the control cells (Table III).
Raf Signaling Causes Nuclear Translocation of E2F5 and LEK1
LEK1 and E2F5, two proteins of interest identified in extracts from the Raf
expressing myoblasts, were chosen for further analysis (Figure 12). E2F5 and LEK1 are
pocket protein binding proteins that are involved in cell cycle control (44, 104, 115, 145).
E2F5 and LEK1 proteins are located in the nucleus of Raf-induced myoblasts (Figure
13). On closer inspection, LEK1 was localized throughout the cytoplasm and nucleus in
control serum starved myoblasts. Treatment with 1 pM 4HT treatment caused a marked
redistribution of LEK1 to the nucleus. Residual cytoplasmic LEK1 immunoreactivity
was apparent in Raf induced myoblasts. E2F5 expression was similar to LEK1. In
control myoblasts, E2F5 appeared primarily cytoplasmic, entering the nucleus upon Raf
induction. Nuclear accumulation of E2F5 begins as early as 15 minutes post Raf
induction, but does not occur at anytime during the one hour treatment in serum starved
controls (Figure 14).
The translocation of E2F5 and LEK1 prompted the examination of ERK1/2
localization in response to Raf activity. Surprisingly, phosphorylated ERK1/2 remained
cytoplasmic throughout the first hour of Raf induction (Figure 15). This was confirmed
through Western blotting of nuclear extracts for phospho-ERK1/2 (Figure 16). To verify
o i I
I t:;~ bi(( ;"-a
Figure 11. Representative two dimension gel of nuclear extracts from control cells.
Nuclear extracts from vehicle-only treated cell were separated over a p14-7 gel
strip and subsequently over a 10% SDS gel and silver stained. Circled
proteins are those excised for identification.
Figure 12. Representative two dimension gel of nuclear extracts from Raf induced cells.
Nuclear extracts from Raf induced cells were separated over a p14-7 gel strip,
a 1-% SDS gel and subsequently silver stained. Circled proteins are those
excised for identification. LEK1 and E2F5 were chosen for further analysis.
Table III. Unique proteins identified from control and Raf-induced nuclear extracts.
Control Raf induced
Protein MW pI Ascension # Protein MW pI Ascension #
Aldehyde 57kD 7.53 P47738 Dihydrolipoamide 59kD 5.71 AAL02400
AnnexinAl 39kD 7.15 P10107 E2F5 37kD 5.17 061502
Annexin A2 39kD 7.53 P07356 gprinl 96kD 6.79 AAH57044
Calumenin 37kD 4.49 035887 LEK1 284kD 4.94 090Z84
CaseinKinase 41kD 9.5 08BK63 Tumor rejection 92kD 4.74 P08113
al antigen gp96
ChaDeronin B 58kD 5.97 AAH26918 Txndc7 48kD 5.05 BC006865
LiDocortin 1 39kD 6.57 NP 034860 YL2 31kD 4.77 08R5L1
Reticulocalbin 37kD 4.7 005186 U8 51kD 8.74 BAA79193
SNEV 56kD 6.14 099KP6 Isocitrate 40kD 6.46 09D6R2
TNFa induced 37kD 8 NP 891995
Msx2 interacting 389k 8. AF156529
that the phosphorylation and cytoplasmic location of ERK1/2 was a result of Raf
activation, myoblasts were treated with 50 atM PD98059, a MEK1/2 inhibitor. In the
presence of both 4HT and the MEK inhibitor, no phosphorylated ERK1/2 was present in
the cytoplasm (Figure 17).
E2F5 and LEK1 both interact with members of the pocket protein family of
proteins. To determine whether these proteins are involved in the early events leading to
quiescence, Raf induced cells were immunostained for pRb, p130, and p107 expression
(Figure 18). No apparent changes in p130 or p107 localization or expression level occur
within the first hour of Raf induction. This was surprising, as p130 and E2F5 are
commonly found localized in the nucleus of quiescent cells (55). This suggests that
while p130 may be required for the maintenance of quiescence, it is not immediately
necessary for the induction of the quiescent state. However, pRb appears to localize to
the nucleus following one hour of 4HT stimulation, although some cytoplasmic staining
remains. The number of cells expressing nuclear pRb is increased significantly in treated
cells (Figure 19, p<0.01). Immunostaining was verified with nuclear extracts, which
appeared to show increased levels of pRb in the nucleus of Raf induced myoblasts but no
apparent changes in p130 or p107 concentration in response to Raf activity (Figure 20A).
However, upon quantification with densitometry, no significant changes in intensity were
apparent, although pRb tended to increase with 4HT treatment (Figure 20B). The
western blots may be more sensitive than the immunocytochemistry, as p130 and p107
could be detected in the nuclear extracts and more nuclear pRb was detected in control
myoblasts. Further work should be completed to define the effects of Raf activity on pRb
localization. E2F5 is often examined in conjunction with E2F4, the other major
Growth Arrested Raf-Arrested
Figure 13. Localization of LEK1 and E2F5. 23A2RafERDD myoblasts were treated with
1 [iM 4HT or vehicle only for one hour prior to immunostaining with LEK1 or
0 15 30 60 minutes
15 30 60 minutes
Figure 14. Localization of E2F5 over time. 23A2RafERDD myoblasts were treated with
1 M 4HT or vehicle-only for times stated and immunostained for E2F5.
Figure 15. Localization of phospho-ERK1/2 over time. 23A2RafERDD myoblasts were
treated for the times shown with 1 iM 4HT or vehicle-only (not shown) and
immunostained for phosphoERKl/2 and Hoechst.
c (-) (+)
Figure 16. No phosphorylated ERK1/2 is present in control or Raf induced nuclear
extracts. Nuclear extracts were isolated after 1 hour of treatment with 1 M
4HT or vehicle-only and separated over a 10% SDS gel. Western blots were
probed with phosphorylated and total ERK1/2. C control; (-)- vehicle only;
(+)- 1 iM 4HT
I (-) PD98059 (+) PD98059 "
(-)4HT (+) 4HT I (-) 4HT (+) 4HT |
Figure 17. The presence of a MEK inhibitor blocks the activation of ERK1/2.
23A2RafERDD myoblasts were treated with 1 gM 4HT or vehicle-only in the
presence of absence ofPD98059, a MEK1 inhibitor, prior to immunostaining
for phosphorylated ERK1/2.
transcriptional repressor of the E2F family. Immunostaining verified the specificity of
localization changes for E2F5, as E2F4 expression does not appear to change as a
function of Raf activity (Figure 21). Interestingly, a punctate ytoplasmic
immunostaining pattern exists in both control and treated myoblasts that appears to co-
localize with regions of rough endoplasmic reticulum. The pattern suggests that E2F4
gene expression and protein synthesis may be a product of general growth arrest.
Because the localization changes of E2F5 and pRb were found in response to high
levels of ERK1/2 activity, we sought to determine whether these effects were caused by
ERK1/2 phosphorylation or other cellular changes. To this end, the MEK inhibitor
PD98059 was used to block ERK1/2 activation. Nuclear localization of pRb and E2F5
was blocked in the presence of PD98059 indicating that these changes are a result of Raf
activation of the downstream MEK/ERK signaling axis (Figure 22).
To verify that the cells were entering a quiescent state distinct from terminal
differentiation, Raf induced cells were immunostained for myogenin (mgn), an early
marker of terminal differentiation. Nuclei were counterstained with Hoescht dye and a
labeling index was calculated at mgn(+)/total nuclei. Twenty three percent of the control
cells expressed myogenin; few (-2.5%) Raf-induced myoblasts stained positive (Figure
23A). All cells expressed MyoD (Figure 23B) indicating the effect was specific for
To further verify that the cells were entering a state that was reversible, myoblasts
were stimulated with 4HT for one hour, followed by culture in serum free media. As
before, the one hour stimulation resulted in nuclear translocation of E2F5 and pRb.
Removal of the stimulus partially restored the cytoplasmic location of both proteins
(Figure 24). The restoration of the cytoplasmic location of E2F5 and pRb is
accompanied by the presence of less cytoplasmic protein. These proteins may
potentially be targeted for degradation upon re-entrance to the cell cycle.
Figure 18. Pocket protein expression in control and Raf induced cells. 23A2RafERDD
myoblasts were treated for one hour with 1 giM 4HT or vehicle-only prior to
immunostaining for p107, p130, or pRb.
100 (-) 4HT *
0 (+) 4HT
Figure 19. Differences in pRb expression in the nucleus. 23A2RafERDD myoblasts were
treated with 1 iM 4HT or vehicle-only for one hour and immunostained for
pRb. The percent of cells positive for nuclear pRb was calculated from three
independent tests. Asterisk indicates a significant difference p<0.01.
1 (+) 4HT
I- p107 p130 -. I- pRb -I
Figure 20. Pocket protein expression in nuclear extracts of control and Raf induced cells.
Nuclear extracts from Raf induced (+) and control (-) myoblasts were probed
for pRb, p107, and p130 (A). Densitometry performed on two independent
tests reveals no significant differences in protein expression between control
((-) 4HT) and Raf induced ((+) 4HT) myoblasts, n=2 (B).
l hr CTT. 1 hr CTT
Figure 21. E2F4 expression does not change in response to Raf activity. 23A2RafERDD
myoblasts were treated for one hour with 1 tM 4HT or vehicle-only (CTL)
and immunostained for E2F4 and Hoescht.
I (-) (+) I
I (-) (+) 1 4HT
Figure 22. Inhibition of pERK1/2 blocks the translocation of E2F5 and pRb to the
nucleus. 23A2RafERDD myoblasts were treated with 1 M 4HT or vehicle-
only for one hour in the presence or absence of PD98059 prior to
immunostaining for E2F5 or pRb and Hoescht.
^ 20 -
c: 10 -
0 (+) 4HT
Figure 23. Myogenin is not expressed in Raf induced cells. Raf induced myoblasts were
immunostained for myogenin and MyoD expression (A). The percent of
myogenin positive cells was determined from three independent tests.
Asterisk indicates significant difference (B).
I ME Olo
Stimulation I Recovery
(-) 4HT (+) 4HT I (-)4HT (+) 4HT
Figure 24. Recovery from ERK1/2 stimulation results in the partial restoration of
cytoplasmic location of E2F5 and pRb. 23A2RafERDD myoblasts were
stimulated with 1 iM 4HT or vehicle-only for one hour (stimulation), allowed
to recover in serum free media for three hours (recovery), and immunostained
for E2F5 or pRb and Hoescht.
Embryonic myoblasts and proliferative satellite cells behave in similar manners.
Proliferating satellite cells have two possible fates terminal differentiation or return to
quiescence. It is controversial whether embryonic myoblasts contribute to the satellite
cell population, but it is well established that they are both capable of fusing to form
myofibers. If satellite cells arise from embryonic myoblasts exiting the cell cycle to a
quiescent state, determination of the causal signals would be extremely important.
The 23A2RafERDD cell line was derived from an embryonic mouse myoblast
population. The fusion of the Raf molecule to a tamoxifen-specific estrogen receptor
allows a dose-dependent response to 4HT. High levels of Raf kinase activity over 48
hours result in cell cycle inhibition, as evidenced by a lack of proliferation in treated cells
while control myoblasts proliferate. The stability of cell number in treated cells coupled
with the decreased BrdU incorporation and lack of apoptosis or senescence, suggests that
early changes are affecting S-phase entry. Consistent with the results of the presented
work, expression of an inducible Raf molecule was not capable of inducing S-phase entry
in serum starved quiescent 3T3 cells (114). In addition, a highly active Raf mutant also
failed to stimulate the cell cycle entry of quiescent cells but induced the expression of
cyclin D and p21Cipl, which binds cyclin E/cdk2 complexes to inhibit cell cycle
The removal of stimuli and return to a normal growth environment stimulated cell
cycle re-entry. This supports the protective nature of high Raf activity against apoptosis
and suggests that the cells are becoming quiescent rather than senescent, a state from
which the cell cycle cannot be re-entered. In addition, P-galactosidase staining was
negative in accordance with the work of DeChant et al., a group that showed the
protective effects of Raf activity against apoptosis in myoblasts (40). Furthermore, the
presence of high serum media abrogated the effects of Raf induction. It is interesting to
note that in serum starved 3T3 cells with high levels of Raf activity the addition of FBS
allows proliferation, while the addition of other growth factors results in a retained
quiescent state (114). This suggests that this pathway may be primarily active during
times of non-growth. Satellite cells proliferate following activation due to growth factors
and other stimuli. The presence of growth factors causes myoblasts to remain cyclic,
much like the presence of growth factors in muscle activates satellite cells and causes
subsequent proliferation. The abrogation of Raf induced changes by high levels of serum
may be reflective of the in vivo state of satellite cell proliferation.
To further clarify the immediate time course of Raf pathway activation, cells were
treated over time and Western blots probed for phosphorylated ERK1/2. Results indicate
that a slight activation of ERK2 was occurring at thirty minutes but both kinases were
strongly phosphorylated after one hour of Raf induction. The treated cells were stripped
of all extracellular growth factors and starved for one hour prior to Raf induction,
removing any potential mitogenic signal. The early activation of the Raf kinase suggests
that this pathway plays an early role in the morphological and cell cycle related changes
over time. The decrease in active ERK1/2 at 120 minutes is most likely reflective of
unequal loading or detection. Interestingly, there appears to be no active ERK1/2 in the
absence of stimulation suggesting a lack of ERK1/2 gene transcription and/or translation.
However, upon stimulation of upstream kinases, inactive ERK1/2 may be required to
immediately upregulate its own expression. Active ERK1/2 may be required to maintain
a quiescent state. In one study, the presence of ERK1/2 (regardless of phosphorylation
status) was used to mark individual satellite cells on isolated rat muscle fibers. The
presence of ERK1/2 may indicate that these quiescent cells require active ERK1/2 to
remain unproliferative. Interestingly, the presence of phosphorylated ERK1/2 decreases
from 0 minutes (the start of culture) to 24 hours of culture in satellite cells on isolated
myofibers. This coincides with the time satellite cell activation following plating (143).
LEK1 Expression Changes in Response to Raf Induction
The CENP-F/mitosin/CMF 1 family of proteins includes LEK 1, named for the large
amount of leucine (L), glutamic acid (E), and lysine (K) residues present in the protein
(57, 102). Human CENP-F/mitosin, mouse LEK1, and chicken CMF1 proteins share a
highly conserved C-terminus (57). CMF1 protein is 65% similar to LEK1 and CENP-F
(102). The family members share an atypical Rb binding domain, an ATP/GTP binding
site, a nuclear localization signal, and a predicted HLH dimerization domain distributed
in a collinear fashion in the C-terminal (10, 57, 102, 110). LEK1 appears to be widely
distributed amongst species with protein orthologs present in canine, mouse, and mink
derived cell lines. The core region of these proteins necessary for centromere binding is
conserved between species, including chickens and humans (145). LEK1 also expresses
several similarities to the E2F proteins, including a Myc-type HLH motif, a similar Rb-
binding site, and several leucine zippers possibly responsible for DNA binding (10)
LEK1 is expressed as early as five days after the formation of the embroid body,
but not in the post natal murine heart. Knockdown of LEK1 using siRNA resulted in a
severe compromise in the beating ability of embroid bodies, tightly following the time
course produced in Rb null embroid bodies. Fewer embroid body cardiomyocytes are
present compared to wild type and those present showed incomplete sarcomere formation
(104). These results suggest that LEK1 is required for terminal differentiation.
The expression of the LEK family of proteins differs throughout the cell cycle,
indicating varying functions although they belong to the same family. LEK1 is targeted
to the nucleus in differentiating embroid bodies (104) and is nuclear during interphase in
murine cells, showing patterns indicative of centromere localization in mitotic cells
(145). In stable fibroblast lines, CMF1 is predominantly localized to nucleus (110).
Similarly, LEK1 is present in the nuclei of actively dividing cells and newly
differentiated myotubes, but is absent in older myotubes (57). CENP-F is present in low
levels in the cytoplasm of G1 cells but increases sharply in S phase as a nuclear protein
(145). CMF1 is detected in the cytoplasm of developing cardiomyocytes and is
maintained in differentiated myocytes, differing significantly from other LEK family
In the present study, Raf induced entry into quiescence results in the majority of
LEK1 protein translocating to the nucleus, while differentiating cells retain LEK1
cytoplasmically. This contrasts with previous reports of LEK1 nuclear localization in
differentiated embryoid bodies and mitotic cells. These differences may be due to unique
cell lines as it is well known that members of the LEK1 family have different functions
according to their tissue of origin. In addition, the time point examined in the present
study is a very short time after the initial stimulation. Other work has examined later
stages of the cell cycle (104, 145). Ashe et al. report post-translational modification of
1IIII 1 I
I liili 1111 1 111________
E Leucine zipper
N Atypical Rb binding domain
Myc-type bHLH heterodimerization domain
o Spectrin repeat
E bHLH domain
* Rb binding site
Q Leucine zippers
0 Nuclear localization signal
Figure 25. Schematic drawing of the conserved structures in the LEK family of proteins
(A) and potential interaction sites (B). NLS, nuclear localization signal.
-- "''-'-' '
Ii I I I
LEK1 in the cytoplasm, resulting in the cleavage and nuclear translocation of the C-
terminus (10). Unfortunately, the exact cleavage site and mechanism of modification has
yet to be determined. Immunocytochemistry indicates the presence of LEK1 in both the
nucleus and cytoplasm in Raf induced quiescent cells. This is likely due to the ability of
the antibody to recognize both the full length and cleaved protein as the antibody is
targeted to the C-terminus of the protein.
The full function of LEK1 has yet to be determined. Despite the presence of
several leucine zippers with DNA binding potential, LEK1 alone cannot directly bind
DNA (10). However, LEK1 recognizes sequences in all three pRb binding pocket
subdomains and complexes with active pRb, p107, and p130 in a cellular environment
(10, 110). Similarly to E2F proteins, LEK1 binds the long pocket, the A/B pocket, and
the C pocket regions of Rb, p107, and p130. However, LEK1 is capable of interaction
with an E2F-binding incompetent Rb, which has a point mutation at amino acid 706. Pull
down assays indicate that the pRb binding domain located in the C-terminus of LEK1 is
essential for pocket protein binding (10). In the present work, LEK1 is accompanied by
E2F5 and pRb in translocation to the nucleus. This suggests that although pRb and E2F5
may not bind preferentially, these three proteins may interact to elicit entry into
The Rb family of proteins significantly affects the cell cycle. Knockdown of LEK1
by an antisense morpholino directed at LEK1 results in a reduced cell number and an
accumulation of cells in the G1 phase of the cell cycle, indicative of G1/S phase arrest.
Absence of LEK1 prevents cells from entering G2 when released from serum starvation
and subsequently induces apoptosis (10). LEK1 may potentially act as a pocket protein
suppressor by disrupting their association with other proteins. dnLEK1 enhanced the
differentation of C2C12 myoblasts indicating that this protein may play a role in
achieving or maintaining quiescence (57). In the present study, LEK1 may be
sequestering pRb in the cytoplasm until appropriately stimulated and subsequently
assisting or allowing nuclear translocation of this or other proteins to aid in the entrance
E2F5 is Involved in Cell Cycle Exit
The E2F proteins are a set of transcription factors that assist in cell cycle control.
The eight proteins are divided in to three subgroups based on function. E2F 1, E2F2, and
E2F3 are involved in positive cell cycle control and S phase entry of quiescent cells. The
second subgroup consists of E2F4 and E2F5. These proteins negatively affect the cell
cycle progression. E2F6, E2F7, and E2F8 make up the third group. Although they are
generally classified as repressors, the structure and function of these proteins is
dramatically different from the former two groups, as none bind pocket proteins.
Closely related, 69% of the amino acids in E2F4 and E2F5 are identical and 80%
are similar. The 1239 bp E2F4 cDNA open reading frame encodes a 413 amino acid
protein that is approximately 44kD (115). E2F5 is a 345 amino acid protein of
approximately 37.5kD encoded by a 1035bp open reading frame (63, 115). Both are
constitutively synthesized (115). These two factors lack a cyclin A binding domain,
resulting in shorter N-terminals than present on the other E2F factors (115). E2F5 lacks
the serine repeat region found in E2F4, however both proteins contain a C-terminal
transactivation domain (63)(Figure 26).
D'Souza et al. found abundant levels of E2F5 in the murine brain, heart, lung,
liver, and kidney with low levels expressed in the dermis and epidermis (37). Unlike
other E2F transcription factors, E2F5 levels are detectable in adult rat cardiac myocytes
(136). Murine E2F5 mRNA is transcribed and protein can be detected in the cytoplasm
from pre-implantation through 1 Id.p.c. embryos and in gametes of both sexes (103).
Upregulation of E2F5 is associated with completed neuronal circuitry in the adult murine
brain (82). Actively proliferating murine keratinocytes express lower levels of E2F5,
which increase upon differentiation (37).
Some controversy exists regarding the expression profile of E2F5 during the cell
cycle. Several labs have found E2F5 complexes present in quiescent cells. One study
found E2F5 transcript levels peaking in mid to late G1 and returning to Go levels as
human fibroblasts enter S phase (115). Unlike E2F 1/2, little to no change in E2F4/5
mRNA expression occurred throughout the cell cycle of fibroblasts. Fugita et al.
determined that in serum starved quiescent rat smooth muscle vascular cells; E2F5 was
down-regulated and increased upon serum stimulation in a time dependent manner (48).
Expression of E2F 1-4 in serum-starved quiescent mouse and rat cardiomyocytes induces
cell cycle entry, however only E2F 1/2 expression can complete the cell cycle by
stimulating mitotic division (44). Interestingly, a 24 hour treatment of keratinocytes with
TGF-P to achieve a reversible growth arrest caused a decrease in E2F5. Forced ectopic
expression of E2F5 for 24 hours in cycling keratinocytes reduced DNA synthesis (37). In
another study, a variety of cell types were used to show that the E2F5-pl30 complex was
present in Go but not throughout the remainder of the cell cycle. Hypertrophic stimulus
of rat cardiomyocytes promotes E2F 1/3/4 and DP expression while down-regulating
E2F5, consistent with a role for E2F5 in maintenance of growth arrest or quiescence
(136). In differentiated L6 myotubes, E2F5 exists primarily in the cytoplasm, while pRb,
SNuclear Localization Signal
DNA binding domain
Nuclear Export Signal
O DP dimerization domain
El Transactivation domain
E Pocket protein binding
Figure 26. Schematic representation of conserved E2F4 and E2F5 domains.
p130, and p107 are present in the myonuclei (55). E2F5 may serve different purposes in
individual cell types. In the present study, serum-starved myoblasts express E2F5
cytoplasmically before entering terminal differentiation. In response to high levels of Raf
activity, E2F5 translocates to the nucleus. The presence of nuclear E2F5 in this study is
consistent with the induction and maintenance of a quiescent state. The Raf-induced cells
are likely to be entering a true quiescent state as they are capable of resuming
proliferation. It is tempting to speculate that endogenous cytoplasmic E2F5 translocates
to the nucleus to up regulate its own transcription in preparation for maintenance of
It is widely accepted that E2F5 contains a nuclear export sequence (NES);
however, the mechanism for nuclear entry is more controversial. Localization is
mediated through the balance of import and export, as E2F5 has an atypical nuclear
export sequence in a region that is poorly conserved between E2F5 and E2F4 (AA 130-
154) as well as an atypical nuclear localization sequence (NLS) in AA 1-56 (8). The
NES contains a high number of hydrophobic residues rather than a Rev-type Leucine-rich
NES. Nuclear export is mediated through the CRM1 pathway (8). Although some
groups have claimed that E2F5 translocates to the nucleus via association with pocket or
DP proteins (54, 55), Apostolova et al. contend that association with pocket proteins or
DP proteins is not necessary for nuclear entry (8). E2F5 enters the nucleus through
nuclear pores in an energy-dependent manner. In the present study, E2F5 does not
appear to require pocket proteins for nuclear entry, as p130 and p107 patterns are
unchanged. pRb does translocate to the nucleus in response to high levels of Raf activity,
but as E2F5 and pRb have little to no binding affinity it is unlikely that pRb is
transporting E2F5. Fujita et al. found E2F5 complexes with pRb upon serum stimulation
which has not been reported by other labs (48). However, in the present work, E2F5 and
pRb translocate to the nucleus in the absence of serum, further suggesting that these two
proteins don't interact to elicit translocation of each other.
Serum stimulation affects the localization of E2F proteins. In media without
serum supplementation, E2F1 and 4 as well as DP-1, pRb, and p130 are cytoplasmic.
However, in quiescent cells supplemented with as little as 0.2% serum, both E2F 1 and 4
are nuclear. This may explain the differences seen in the Raf induced cells in the
presence and absence of high levels of serum in the present study. High serum
concentrations that inhibit nuclear translocation of E2F5 or pocket proteins may provide
mitogenic signals to the myoblast, creating inappropriate signals for cell cycle exit. In
contrast, low serum conditions with high levels of Raf activity may mimic the in vivo
environment promoting satellite cell self renewal. This may explain why Raf induced
cells in high concentrations of serum continue to proliferate while Raf induced cells in
low or no serum media enter quiescence. The presence of mitogens in the serum may
over ride the forced Raf expression.
E2F proteins exert their major effects in two ways: (1) binding and sequestering
pocket proteins and (2) binding DNA to induce or repress transcription. Members of the
pocket protein family p130 and p107, but not pRb, associate with E2F4/5, although E2F5
indicates a preference for p130 binding (63, 115)(Figure 27). E2F5 stimulated
transcription is inhibited by cotransfection with expression vectors containing p130,
p107, and pRb (63). E2F4 can induce pocket protein independent repression during Go in
glioblastoma cells, suggesting E2F5 may also have this capability (11)
1 BI I
I 1/- ____
SCdk inhibitory domain
SE2F1 binding domain
Cyclin/cdk binding domain
Figure 27. Comparison of pRb, p130, and p107 conserved domains.
Although E2F5 is capable of binding DNA as a homodimer, heterodimerization
with one of the DP family of proteins creates a stronger bond (63, 115). DP1/2 have E2F
and DNA binding regions. Heterodimers of E2F transcription factors and DP1 or DP2
cooperate in DNA transactivation (56, 62). The E2F4/DP1 heterodimer is analogous to
other winged helix proteins with the exception of presenting a continuous protein surface
and lacking a C-terminal wing region. The a3 helices bind the major DNA groove while
the N-terminal al helices and portions of 0 sheets contact the DNA backbone. The
E2F/DP residues that contact the DNA bases and backbone are identical within the E2F
and DP families. This suggests that other E2F/DP dimers bind DNA in a similar manner.
E2F5/DP-1 over expression increases E2F DNA binding in culture (105). Over
expression of E2F5 can contribute to cell transformation in a Myc-type oncogene manner
in the presence of DP-1 in primary rat BRK cells (106). How E2F5 is acting in Raf
induced cells is not clear. While it is clear that E2F5 negatively affects cell cycle
progression, further work needs to be completed to determine whether Raf induced E2F5
translocation affects gene transcription in a positive or negative manner.
The tumor suppressor protein p53 interacts with E2F5 to inhibit E2F5-activated
transcription in a dose-dependent manner (135). p53 induces expression of p21, which
then inhibits the cdks and prevents the release of free E2Fs from pRb/E2F complexes. A
transactivation defective p53 is capable of inhibiting E2F5 mediated transcription,
indicating that p21 is not involved in the inhibition. p53 inhibits the activity but not the
expression of E2F5, having no effect on E2F5-DP1 complex formation or DNA binding,
although DP1 was still required for strong binding. p53 inhibits E2F5 transcriptional
activation, but does not physically interact with the protein, indicating an indirect effect.
Regulation of E2F responsive genes in cycling cells involves recruitment of co-
repressors outside of the pocket protein family. HDAC1, a histone deacetylase associated
with transcriptional repression stemming from histone deacetylation, complexes with
E2F5 and p130 in differentiated keratinocytes (37). The reduction in DNA synthesis
resulting from E2F5 over expression is abrogated in the presence of an HDAC1 inhibitor,
indicating that HDAC1 is required for E2F5 mediated growth arrest (37). HDAC1 is
recruited to promoters during quiescence but is absent by mid-G1 (109). Recruitment of
HDAC1 to the promoter requires the E2F4-pl30 complex. In addition, histones H3 and
H4 are deacetylated on certain promoters in quiescent T98G human glioblastoma cells.
In contrast, these histones are extensively acetylated at E2F responsive promoters in S,
G2, and M phases. In G1, histones H3 and H4 are deacetylated when p107 and p130 are
present at the promoters.
mSin3B, a co-repressor that associates with chromatin modifying factors,
occupies promoters in quiescent murine embryonic fibroblasts (VMEFs). p107 or p130 is
required for the recruitment of mSin3b to some but not all E2F regulated promoters
(109). Although HDAC1 and mSin3b both associate with promoters during quiescence,
both are equally present in quiescent and S phase nuclei. In quiescent cells, mSin3B is
targeted to E2F4, E2F4/pl30, and E2F4/p107 promoters. However, in early G1 mSin3B
is not present on E2F4 targets promoters. Pocket protein independent repression does not
require the presence of mSin3B or the deacetylation of histones H3 and H4 in early G1
Somatic stem cells in the drosophila ovary require hedgehog and wingless
proteins for self renewal. Upon division, one stem cell remains anchored to the
germarium niche via cadherin-mediated cell adhesion and maintains stem cell identity.
The other cell loses stem cell identity and differentiates. Somatic stem cells require
domino, which encodes a transcriptional repressor, for the maintenance of the stem cell
population (142). Domino effects drosophila E2F independently ofRb (Lu et al.,
manuscript in preparation). The mammalian homolog of Domino, Snf2 Related CBP
activator protein (SRCAP), is capable of partially rescuing the lethality of domino
mutants and can fully rescue the female infertility associated with domino mutations (45).
SRCAP may potentially be involved in the E2F5 mediated cell cycle repression in
skeletal myoblasts seen in this work. In addition, SRCAP can functionally replace
Domino in drosophila Notch signaling, indicating that it is likely involved in mammalian
Notch signaling as well. Indeed, SRCAP is a co-activator of Notch dependent gene
expression in drosophila (45). In addition, another target of the Notch pathway, the HES
related repressor protein (HERP), can repress gene expression through association with
mSin3 and HDAC1, both of which are present in E2F5 repressor complexes (11, 37, 66,
67, 109). E2F5 repressor complexes containing SRCAP or HDAC1/mSin3/HERP may
inhibit Notch transcription, allowing cell cycle exit. Considering that Numb, the Notch
inhibitor, is present in satellite cells returning to the quiescent state, this gene repression
may be an additional "off switch" for the Notch signaling pathway (28).
Methylation of certain E2F gene elements (dhfr, E2F], and cdc2) blocks the
binding of all E2F factors while other elements (c-myb and c-myc promoters) can bind all
factors except E2F 1 regardless of methylation status. Association with pRb family
members doesn't affect the ability of E2F factors to bind methylated elements (22).
Unlike pRb, p130, and E2Fs 1-3, growth inhibitory TGF-3 signaling results in p107 and