The FASEB Journal • Research Communication

Sequenced response of extracellular matrix deadhesion

and fibrotic regulators after muscle damage is
involved in protection against future injury in human
skeletal muscle
Abigail L. Mackey,*,†,1 Simon Brandstetter,*,† Peter Schjerling,*,† Jens Bojsen-Moller,*,†
Klaus Qvortrup,‡ Mette M. Pedersen,*,† Simon Doessing,*,† Michael Kjaer,*,†
S. Peter Magnusson,*,† and Henning Langberg*,†
*Institute of Sports Medicine, Department of Orthopaedic Surgery M, Bispebjerg Hospital, †Center
for Healthy Aging, Faculty of Health Sciences, and ‡Department of Biomedical Sciences, The Panum
Institute, University of Copenhagen, Denmark

ABSTRACT The purpose of this study was to test the muscle have been employed to investigate muscle sore-
hypothesis that remodeling of skeletal muscle extracel- ness, inflammation, and regeneration (e.g., 1, 2– 4).
lular matrix (ECM) is involved in protecting human The majority of these studies have used high-force
muscle against injury. Biopsies were obtained from lengthening (eccentric) contractions. However, we re-
medial gastrocnemius muscles after a single bout of cently provided evidence for the capacity also of elec-
electrical stimulation (B) or a repeated bout (RB) 30 d trically stimulated isometric contractions to induce
later, or 30 d after a single stimulation bout (RBc). A muscle damage and initiate regenerative events in
muscle biopsy was collected from the control leg for humans, evidenced by infiltration of inflammatory
comparison with the stimulated leg. Satellite cell con- cells, the presence of desmin" and dystrophin" fibers,
tent, tenascin C, and muscle regeneration were assessed z-line disruption, and activated satellite cells (SCs;
by immunohistochemistry; real-time PCR was used to refs. 5, 6).
measure mRNA levels of collagens, laminins, heat- Lengthening contractions have also been the tradi-
shock proteins (HSPs), inflammation, and related tional model of choice in investigations into under-
growth factors. The large responses of HSPs, CCL2, standing how one exercise bout could protect against
and tenascin C detected 48 h after a single bout were muscle damage when a similar bout was repeated (the
attenuated in the RB trial, indicative of protection phenomenon known as the repeated-bout effect). The
against injury. Satellite cell content and 12 target genes, fact that a muscle can be protected from damage and
including IGF-1, were elevated 30 d after a single bout. soreness purely by a single bout of exercise performed
Among those displaying the greatest difference vs. from 1 wk up to several months (e.g., 7–9) prior to
control muscle, ECM laminin-!1 and collagen types I reexposure to the same exercise is intriguing and has
and III were elevated !6- to 9-fold (P<0.001). The implications not only for strategies in injury prevention
findings indicate that the sequenced events of load- but also potentially in understanding muscle plasticity
induced early deadhesion and later strengthening of in the context of acute and chronic muscle wasting. At
skeletal muscle ECM play a role in protecting human the cellular level, heat shock proteins (HSPs) have
muscle against future injury.—Mackey, A. L., Brand- received some attention, with support for and against a
stetter, S., Schjerling, P., Bojsen-Moller, J., Qvortrup, role in the repeated-bout effect (10 –12). Similarly,
K., Pedersen, M. M., Doessing, S. Kjaer, M., Magnus- evidence suggests that inflammatory cells might be
son, S. P., Langberg, H. Sequenced response of extra- involved (13, 14). While many candidates are likely, the
cellular matrix deadhesion and fibrotic regulators after connective tissue making up the endomysium and
muscle damage is involved in protection against future perimysium has received little attention, despite its
injury in human skeletal muscle. FASEB J. 25, 1943–1959 many roles, including accommodation of nociceptors,
(2011). www.fasebj.org transmission of force, registration and conversion of
mechanical signals, and as a niche regulator of myo-
Key Words: insulin-like growth factor-1 ! satellite cells ! colla- genic precursor cells (15–18). It has been shown clearly
gen ! tenascin C ! repeated bout effect
1
Correspondence: Institute of Sports Medicine, Bispebjerg
Hospital (building 8, 1st floor), Bispebjerg Bakke 23, DK-2400
The adaptation of skeletal muscle after injury to Copenhagen NV, Denmark. E-mail: abigail.mackey@gmail.
resist future damage is not understood. Models induc- com
ing controlled and standardized damage to skeletal doi: 10.1096/fj.10-176487

0892-6638/11/0025-1943 © FASEB 1943

that physiological mechanical loading is associated with
dynamic adaptation of collagen, the major adhesion-
promoting component of muscle extracellular matrix
(ECM; refs. 19 –24). In addition, evidence is apparent
for attenuation of contraction-induced injury with in-
creased muscle collagen content in rats (25). Tenascin
C, belonging to the matricellular class of ECM proteins
(26), has been reported to demonstrate dramatic re-
sponses in skeletal muscle subjected to unaccustomed
exercise (4, 27, 28) and indeed to be essential for
full-scale muscle injury repair, providing deadhesive
effects and potentially a coordinating role at least in
stages of the regeneration process (29). In light of this,
we hypothesized that coordinated deadhesion and
strengthening of muscle ECM constitutes a vital adap-
tation in providing protection against contraction-in-
duced injury in human skeletal muscle.
To investigate the role of ECM remodeling in pro-
tection of human skeletal muscle against damage, we Figure 1. Outline of study design and timeline indicating the
time points for muscle soreness measurements, blood sam-
subjected human volunteers to a single or repeated
ples, muscle biopsies and whether groups were subjected to a
bout of electrically stimulated contractions and mea- single bout (B) or repeated bout (RB) of electrical stimula-
sured the response of factors involved in the early and tion (ES). RBc, control group for the RB group.
late stages of muscle tissue regeneration (HSPs, ECM
components, myogenic precursor cells, inflammatory
cells, and related growth factors). We hypothesized Experimental design and setup
that, along with lower soreness and circulating creatine
kinase (CK) activity, muscle biopsies collected after a The experimental setup and ES protocol were identical to
those described previously (5). Briefly, stimulation electrodes
repeated bout of ES isometric contractions would dem-
(Stimtrode, ST32D; Axelgaard Manufacturing Co., Fallbrook,
onstrate lower mRNA levels of HSPs, myogenic factors, CA USA; electrode size !25$25 mm, interelectrode center-
matricellular proteins, inflammatory factors, and insu- center distance !35 mm) were positioned over the middle
lin-like growth factor 1 (IGF-1), together with a lesser aspect of the gastrocnemius medialis muscle and connected
extent of morphological and ultrastructural damage to a stimulator (Elpha II 3000; Danmeter; Biotin, Odense,
than after a single bout, reflecting an adaptation of the Denmark). During a short familiarization session, the muscle
supporting ECM. contraction was visually confirmed, after which intermittent
muscle stimulation was performed for 30 min (60-Hz stimu-
lation, pulse width 300 %s, duty cycle 40%, total cycle time
10 s, rise time 1 s, total stimulation time 4 s, descending time
MATERIALS AND METHODS 1 s, rest time 6 s). Subjects held the stimulator themselves and
were accordingly in control of the stimulation intensity (mA).
They were instructed to increase the current as much as they
Ethical approval could tolerate and were prompted to do this every 5 min
throughout the 30-min period. The level of current was
The study was approved by the Ethics Committees of the recorded every 5 min during the first bout, and each partic-
Municipalities of Copenhagen and Frederiksberg (ref. H-A- ipant received the same level of current during the second
2007-0037) and conformed to the standards set by the Dec- bout as was delivered during the first bout.
laration of Helsinki. All volunteers gave written informed
consent before inclusion. Soreness

Volunteers Muscle soreness was assessed prior to stimulation and on each

of the 7 d postexercise with the aid of a visual-analog linear
An overview of the participant groups is illustrated in Fig. 1. soreness scale, which ranged from 0 (normal, no pain) to 10
Twenty-two young healthy untrained men volunteered for the (extremely painful). Soreness was assessed in 3 different ways:
study and were divided into 2 groups of 11: repeated bout by self-palpation in the belly of the muscle, active muscle
(RB; mean # sd age 25#4 yrs; height 1.83#0.08 m; weight contraction (heel-raise), and passive muscle stretch, as de-
76#7 kg) and repeated bout control (RBc; mean # sd age scribed previously (5). Soreness was recorded in the RB group
22#3 yrs; height 1.83#0.05 m; weight 74#10 kg). Participants after bout 1 and bout 2, and in the RBc group after biopsy
assigned to the RB group were subjected to two bouts of alone.
electrical stimulation (ES), separated by 1 mo. RBc group
participants followed a similar protocol but were not sub- Blood sampling
jected to the second bout of ES. Single-bout (B) group
participants, who were also only subjected to a single bout of Blood samples were collected by venepuncture before and on
ES, consisted of 7 young healthy untrained men (mean # sd d 2, 4, and 7 following both bouts of ES in the RB group, and
age 22#3 yrs; height 1.84#0.05 m; weight 78#6 kg) from a at the corresponding time points in the RBc group, to
previous study (5). monitor the effects of muscle biopsy sampling alone on

1944 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
circulating CK. Plasma CK levels were measured at the Regeneration
Department of Clinical Biochemistry at Bispebjerg Hospital,
Copenhagen, as described in detail previously (5). Regeneration was investigated by double-staining serial sec-
tions with the following combinations of primary antibodies:
Muscle biopsies embryonic myosin (F1.652; Developmental Studies Hybrid-
oma Bank, Iowa City, IA, USA) and dystrophin; CD56
(347740; Becton Dickinson, San Jose, CA, USA) and laminin;
Muscle biopsies were obtained from the belly of the gastroc- laminin '5 (4C7, ab17107; Abcam) and laminin; myosin type
nemius medialis by percutaneous needle biopsy with suction. I (A4.951; Developmental Studies Hybridoma Bank) and
None of the participants in this study had previously had laminin; myosin type II (A4.74; Developmental Studies Hy-
biopsies taken from either gastrocnemius medialis muscle. bridoma Bank) and laminin.
Two biopsies were collected from each individual in the RB
group 48 h after the second bout of ES, one from the leg that Tenascin C
had been stimulated and one from the contralateral leg as a
control. In the RBc group, which did not undergo a second
bout of ES, biopsies were collected 30 d after the single bout Tenascin-C immunoreactivity was digitally evaluated from
of ES in order to investigate any residual regeneration at this sections incubated with the primary antibody NCL-Tenas-C
time point. On extraction, part of the sample was immersed (Novocastra; Leica Microsystems, Newcastle On Tyne, UK), a
in 2% glutaraldehyde in 0.05 M sodium phosphate buffer biotinylated secondary antibody (E0433; Dako), Vector Elite
(pH 7.2) for analysis by transmission electron microscopy. ABC kit (PK6100; Vector Laboratories, Peterborough, UK)
The remaining part of the sample was aligned, embedded in and visualized with Immpact diaminobenzidine (DAB) sub-
Tissue-Tek (Sakura Finetek Europe, Zoeterwoude, The Nether- strate (SK-4105, Vector Laboratories). The percentage of
lands) and frozen by immersion in isopentane, precooled by pixels demonstrating tenascin-C immunoreactivity was calcu-
liquid nitrogen. Serial transverse sections (10 %m) were cut at lated from 2 nonoverlapping images of dimension 1323 $
"24°C using a cryostat and stored at "80°C. Muscle biopsies 1757 %m with the aid of the threshold function in ImageJ
from a previous study (5), stored in Tissue-Tek at "80°C since 1.42q (U.S. National Institutes of Health, Bethesda, MD,
collection, were included for the mRNA analysis, and new USA).
immunohistochemical analyses were performed for the pres-
ent study. These biopsies were collected 48 h after a single Collagen types I and III
bout of ES identical to the one used in the current study and
are referred to here as the B group. We have previously Immunohistochemical staining for collagen types III (C7805,
published data documenting muscle damage and SC activa- Sigma-Aldrich Denmark A/S, Copenhagen, Denmark) and
tion in these biopsies (5, 6). The CD68 and laminin immu- type I (C2456, Sigma-Aldrich) was also performed to investi-
nohistochemical double-staining performed for one of the gate the localization of these collagen types in regenerating
previous studies (5) was reevaluated in the present study. All muscle.
other stainings of biopsies in the B group described in the
present study were performed in conjunction with staining
SCs
biopsies from the RB and RBc groups.
The number of SCs was determined from cross-sections
Immunohistochemical analysis of muscle biopsies stained with a combination of Pax7, Type I myosin, and
laminin, as described recently in detail (30). Briefly, sections
Immunohistochemical stainings were performed using the were incubated overnight with a mouse anti-Pax7 antibody
same antibodies as in a prior study (5). Briefly, visualization of (MO15020; Neuromics, Edina, MN, USA). A biotinylated
primary antibody binding was achieved with a combination of secondary antibody (E0433; Dako) was then applied, followed
two of the Molecular Probes Alexa Fluor 488/568 goat by Vector Elite ABC kit (PK6100; Vector Laboratories). DAB
anti-mouse/rabbit secondary antibodies (A11031, A11034, substrate visualized antibody binding. The sections were then
A11036, A11029; Invitrogen A/S, Taastrup, Denmark). 4&,6- incubated with primary antibodies against type I myosin
Diamidino-2-phenylindole (DAPI) in the mounting medium (A4.951; Developmental Studies Hybridoma Bank) and laminin.
(P36931; Molecular Probes ProLong Gold; Invitrogen) Alexa Fluor 488 goat anti-rabbit and 568 goat anti-mouse
stained the nuclei blue. As well as the rabbit anti-laminin secondary antibodies rendered type I myosin red and laminin
antibody (Z0097; Dako Denmark A/S, Glostrup, Denmark) green. Sites of Pax7 antigenicity were stained with DAB, visible
used as a general basement membrane marker, we also by light microscopy. See Fig. 2 for an example of this staining on
performed staining with a mouse antibody against laminin '5 one of the biopsies from this study. The number of Pax7 cells
chain (4C7, ab17107; Abcam, Cambridge, UK). To distin- associated with type I or type II fibers was determined and
guish between these two, the general laminin antibody is expressed relative to the total number of type I or type II fibers
referred to here as laminin, and the laminin '5 chain-specific included in the assessment.
antibody is referred to as laminin '5.
Macrophages and cellular activity
Myofiber damage
Macrophage staining was assessed in a similar manner to that
described above for tenascin C from 3 nonoverlapping im-
Macrophage infiltration of myofibers was evaluated from ages of dimensions 664 $ 882 %m, both to calculate the
sections double-stained with primary antibodies for CD68 percentage of pixels demonstrating CD68 immunoreactivity
(M0718; Dako) and laminin (Z0097; Dako). Intermediate and the number of CD68( cells per square millimeter of
filament and sarcolemma disruption was investigated by dou- tissue. As an indicator of general cellular activity, staining for
ble-staining with primary antibodies for desmin (18-0016; Ki67 (CP249, Biocare Medical, Concord, CA, USA) was
Zymed, San Francisco, CA, USA) and dystrophin (ab15277; performed and the number of positive cells expressed relative
Abcam), respectively. to 100 fibers.

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1945

Figure 2. Immunohistochemical detection of
Pax7 cells, type I myosin, and laminin on a single
cross-section of regenerating healthy gastrocne-
mius medialis muscle 30 d after a single bout of
ES (RBc group). In this series of images with an
unusually high density of Pax7 cells, 8 Pax7 cells
are clearly visible, 2 of which are associated with
type I fibers (red) and 6 with type II fibers
(unstained). Laminin staining (green) defines
the fiber borders. Scale bar * 100 %m.

Transmission electron microscopy precipitated again with the help of sodium-acetate (10 %l 3M
NaAc pH 5.5) and 200 %l 96% ethanol. The obtained pellet
Following 3 rinses in 0.15 M sodium phosphate buffer (pH was washed again with 75% ethanol, air-dried, and dissolved
7.2), the specimens were postfixed in 1% OsO4 in 0.15 M in 20 %l RNase-free water. RNA concentrations were deter-
sodium phosphate buffer (pH 7.2) for 2 h. The specimens mined by spectroscopy at 260 nm. In addition, absorbance at
were dehydrated in a graded series of ethanol, transferred to 240 and 280 nm was measured to ensure the purity of the
propylene oxide, and embedded in Epon (TAAB Laborato- sample (260/280- and 260/240-nm ratio). Furthermore, RNA
ries Equipment Ltd., Aldermaston, UK) according to stan- quality was verified by RNA gel electrophoresis using a
dard procedures. Ultrathin sections were cut with a Reichert- formaldehyde agarose gel.
Jung Ultracut E microtome (Leica Microsystems), collected
on 1-hole copper grids with Formvar supporting membranes Real-time PCR
(Merck, Darmstadt, Germany) and stained with uranyl ace-
tate and lead citrate. The sections were examined using a Total RNA (450 ng)was reverse transcribed into cDNA in 20
Philips CM 100 transmission electron microscope (Philips, %l using the OmniScript reverse transcriptase (Qiagen, Va-
Amsterdam, The Netherlands) operated at an accelerating lencia, CA, USA) according to the manufacturer’s protocol.
voltage of 80 kV. Digital images were obtained with an Following dilution of cDNA (1:20), 5 %l of diluted cDNA was
Olympus Soft Imaging Solutions (OSIS) Veleta, side-mounted amplified in a 25 %l reaction batch using 1$ QuantiTect
digital slow scan 2000 $ 2000 CCD camera (Olympus, Tokyo, SYBR Green (Qiagen) Master Mix and 100 nM of each primer
Japan) and the AnalySIS ITEM software package (ITEM (Table 1). The amplification was monitored in real time using
Software, Whitely,UK). Images from the RB and RBc groups a Mx3005P real-time PCR machine (Stratagene, La Jolla, CA,
were combined with the original images for the B group (5) USA). Cycle threshold (Ct) values were related to a standard
and were assessed as described previously (5) in a masked curve made with the cloned PCR products. Specificity was
procedure by an investigator not involved in any of the biopsy ensured by melting curve analysis after the PCR run. Ct values
analyses. ranged from 15 to 40. RPLP0 mRNA was chosen as an
internal control because it was assumed to be expressed
Gene expression analysis constitutively. To validate this assumption, another unrelated
and constitutively expressed mRNA, GAPDH, was measured
Muscle biopsies from all 3 groups were processed at the same and normalized to RPLP0. GAPDH-RPLP0 ratio showed a
time for this analysis. Biopsies were sectioned in batches (with nonconstitutive expression pattern, but RPLP0 was still con-
biopsies from each of the 3 groups represented in each sidered the best choice, as described in Discussion. mRNA
batch) at "24°C using a cryostat, and the frozen sections were data are presented as fold changes relative to the mean of all
collected into a precooled tube. After adding beads (five control values for muscle. Samples were successfully analyzed
2.3-mm steel beads and 1 silicium carbide grain; Biospec from all groups: 5 pairs from the B group, 9 from the RB
Products, Bartlesville, OK, USA) and 1000 %l TriReagent group, and 11 from the RBc group.
(Molecular Research Center, Cincinnati, OH, USA), the
mixture was shaken (Fastprep24; MP Biomedicals, Solon, Statistics
OH, USA) immediately for 15 s at speed 4 m/s and cooled on
ice for !2 min. This procedure was repeated twice. 1-bromo- The level of statistical significance for all tests was P ) 0.05.
3-chloropropane (BCP; 100 %l; Molecular Research Center) Unless otherwise specified, all data are presented as means #
was then added to separate the homogenized muscle samples se. mRNA data were log transformed and analyzed using
into an aqueous and an organic phase. The aqueous phase SigmaPlot for Windows 11.0 (Systat Software Inc., San Jose,
(450 %l; containing the RNA) was used to precipitate RNA CA, USA) by 2-way repeated measures ANOVA. Where a
with isopropanol (450 %l). Following the first wash with 75% significant group $ leg interaction was detected, comparisons
ethanol, the pellet was dissolved in 100 %l RNase-free water, between control and ES legs within and across groups were

1946 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
 TABLE 1. Primer sequences for PCR

mRNA Sense Antisense

aB-crystallin GTGTTGGGAGATGTGATTGAGGTG CTGGGATCCGGTATTTCCTGTGG

CCL2 GCCCTTCTGTGCCTGCTGCT GCAGGTGACTGGGGCATTGATT
cmet AACCCGAATACTGCCCAGACCC TGATATCCGGGACACCAGTTCAG
Col1A1 GGCAACAGCCGCTTCACCTAC GCGGGAGGACTTGGTGGTTTT
Col3A1 CACGGAAACACTGGTGGACAGATT ATGCCAGCTGCACATCAAGGAC
Col4A1 TCGCTGTGGATCGGCTACTCTT CGATGAATGGCGCACTTCTAAAC
Collagen XII CCCAGGTCCTCCTGGATACTGTGA GCAGCACTGGCGACTTAGAAAATGT
CTGF TGCGAAGCTGACCTGGAAGAGA GCCGTCGGTACATACTCCACAGAA
FAK1 CCAGTCCGAGGTCCAGCGAAG CATGTGAACCAGGGTAGCCAGAA
GAPDH CCTCCTGCACCACCAACTGCTT GAGGGGCCATCCACAGTCTTCT
HGF TGAAATATGTGCTGGGGCTGAAA ACAAACAAGTGGGCCACCATAATCC
HSP27 GCTGACGGTCAAGACCAAGGATG TGAAGCACCGGGAGATGTAGCC
HSP70 GTGGCTGGACGCCAACACCTT TTACACACCTGCTCCAGCTCCTTC
IGF1a GACATGCCCAAGACCCAGAAGGA CGGTGGCATGTCACTCTTCACTC
IGF1b GCCCCCATCTACCAACAAGAACAC CAGACTTGCTTCTGTCCCCTCCTTC
IGF1c GCCCCCATCTACCAACAAGAACAC CGGTGGCATGTCACTCTTCACTC
IL-1b TCCAGGGACAGGATATGGAGCA AGGCCCAAGGCCACAGGTATTT
LAMB 1 GGACAAGAGCAATGAGGAGCTGAGA AGCAACTGCTTCAATGCTGTCCAA
LAMB 2 TGAAATTGAAACGGGCAGGAAA GATCACCCAGAGGACCGCGTAG
Myf6 GGGCTCGTGATAACGGCTAAGGA TGTCCACGATGGAAGAAAGGCA
Myogenin CTGCAGTCCAGAGTGGGGCAGT CTGTAGGGTCAGCCGTGAGCAG
p21 CTCAGGGGAGCAGGCTGAAGG AGCCGGCGTTTGGAGTGGTAG
RPLP0 GGAAACTCTGCATTCTCGCTTCCT CCAGGACTCGTTTGTACCCGTTG
Tenascin-C CAACCATCACTGCCAAGTTCACAA GGGGGTCGCCAGGTAAGGAG
TGF-1b GAGGTCACCCGCGTGCTAATG CACGGGTTCAGGTACCGCTTCT
TNFa TTCCCCAGGGACCTCTCTCTAATC GAGGGTTTGCTACAACATGGGCTAC

performed using the Student-Newman-Keuls method. mRNA 4, 45 # 4, and 48 # 6 mA, respectively, increasing
data are presented as geometric means # back-transformed continually with time. All subjects received exactly the
se. All other data were analyzed with GraphPad Prism for same current at each 5-min interval during bout 2 as
Macintosh 4.0c (GraphPad Software, San Diego, CA, USA).
Force production under stimulation during bouts 1 and 2 in during bout 1 and could tolerate the stimulation current
the RB group was compared using a 2-tailed paired t test. delivered. Force production under stimulation was similar
Soreness and CK data for bouts 1 and 2 in the RB group were for both bouts (bout 1 vs. 2: 101#41 vs. 91#44 N;
analyzed using a 2-way repeated measures ANOVA with P*0.47). The stimulation current over the 30 min period
subsequent Bonferroni post hoc tests where a main interaction for the B, RB, and RBc groups was 29 # 10, 41 # 6, and
was observed. Soreness and CK data for the RBc group (effect 28 # 7 mA, respectively. Significant main effects of time
of biopsying alone) were analyzed using a 1-way repeated
and group (P)0.0001) were detected, indicating that the
measures ANOVA with subsequent Dunnett’s multiple com-
parison tests of each post-treatment time point vs. baseline, mean stimulation current delivered to the RB group was
where a significant main effect was detected. Soreness area greater than that delivered to the B and RBc groups.
under the curve was analyzed with a 1-way ANOVA test and
Newman-Keuls multiple comparison post hoc tests. Tenascin-C,
Ki67, z-line, and macrophage (RB and RBc group) data were Soreness
analyzed by Kruskal-Wallis test, with Dunn’s multiple compar-
ison test where the main outcome was significant. Macro-
Muscle soreness was observed to increase in all individ-
phage data from the B group were compared with the
Wilcoxon signed rank test. Since the stability of the Pax7 uals in the RB trial following both bout 1 and bout 2,
antigen over time is unknown, the Pax7 analysis of the older peaking on d 2 or 3 (Fig. 3). A 2-way repeated measures
biopsies from the B group were subject to separate statistical ANOVA revealed a significant effect of time (P)0.0001),
analysis (paired t test) from the RB and RBc groups. Pax7 data but not bout or interaction, for all 3 soreness measures
from the RB and RBc groups were analyzed by 2-way repeated (palpation, contraction, and stretching). The pattern for
measures ANOVA. Stimulation current for the B, RB, and RBc the contraction and stretching data was similar to the
groups was also tested by 2-way repeated measures ANOVA.
palpation data displayed in Fig. 3.
Participants in the RBc group reported increased mus-
cle soreness after biopsy sampling alone, peaking on d 3
RESULTS to 2 AU, 1 d after the biopsy procedure (P)0.001). To
compare muscle soreness between bouts in the RB group
Stimulation current and evoked force production with the effect of biopsy sampling alone (RBc group), the
area under the curve from d 0 to 7 was calculated. Again,
Levels of stimulation current during the first bout at 5, 10, no significant differences were observed between bout 1
15, 20, 25, and 30 min were 32 # 5, 37 # 5, 40 # 4, 43 # and bout 2 for the RB group. The active contraction area

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1947

 assessed by muscle stretching followed a similar pattern,
with biopsy sampling area under the curve (5.6) being
smaller than RB bout 2 (12.9; P)0.05) and bout 1 (15.1;
P)0.05). The area under the curve for palpation of the
stimulated muscle with biopsy sampling (4.1) was also
significantly smaller than both RB group bout 2 (16.9;
P)0.01) and bout 1 (18.8; P)0.01). Accordingly, soreness
as assessed by palpation, stretch, or active contraction
following RB bout 2 was imilar to that observed following
bout 1, both bouts resulting in a greater degree of
soreness than that observed with biopsy sampling alone.

Circulating CK activity (Fig. 4)

A significant bout $ time interaction (P*0.0096) was

Figure 3. Self-assessment of muscle soreness in the days detected for log-transformed CK values following RB
following a single bout (bout 1) or a second bout (bout 2) of group bouts 1 and 2, with significantly lower CK activity
stimulated muscle contractions, performed by the same indi- levels on d 4 and 7 following bout 2 when compared to
viduals 1 mo after bout 1 (RB). A significant main effect of the same time points following bout 1. The biopsy
time was detected, but not bout or interaction. procedure alone (RBc group) resulted in a very small
mean 1.5-fold increase in CK levels (P)0.05), peaking
under the curve for biopsy sampling alone (5.6) was on d 4 (P)0.05), 2 d after biopsy sampling.
significantly smaller than that of RB group bout 2 (mean
13.8; P)0.05) and bout 1 (14.3; P)0.05). Soreness as Immunohistochemical analysis of muscle biopsies

Myofiber damage

Out of all the biopsies collected from the control legs in

t1 the B, RB, and RBc groups, only 1 affected fiber
t2

Figure 5. Scatter plot (with median bars) displaying the propor-

tion of myofiber cross-sections that were positive for embryonic
Figure 4. Circulating CK activity was measured in the RB group myosin or laminin '5, following single bout (B group) or
following 2 bouts of ES-induced isometric contractions of the repeated bout (RB group) of ES. Both markers of regeneration
medial gastrocnemius muscle (A), and in the RBc group follow- were clearly detectable in RB group and in RBc control group
ing biopsying alone (B) on d 2. Biopsies were also collected from 30 d after a single bout. Higher values recorded for laminin '5 than
the RB group 2 d after bout 2. Data are back-transformed embryonic myosin suggest that indications of regenerative pro-
geometric means with back-transformed error bars, presented cesses are preserved in the fiber membrane for longer than in the
on a logarithmic scale. Actual mean values for each bar are myofibrillar component and signify that this marker is less sensitive
given. *P ) 0.05 vs. bout 1; #P ) 0.05 vs. d 0. to the number of fibers in the cross-section.

1948 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
(desmin", dystrophin", infiltrated with CD68( cells) fiber each. No laminin '5( fibers were detected in any
was observed. In the biopsies collected from the mus- of the control leg biopsies. One biopsy from the
cles subjected to ES, desmin" fibers were observed in 2 stimulated leg of a member of the RBc group, clearly
subjects in the RB group and 2 subjects in the RBc demonstrating ongoing regeneration, was investigated
group vs. 4 subjects in the B group (5). Fibers infil- further (Fig. 6). The areas of laminin '5( endomysium
trated with CD68( cells were present in 4 biopsies from staining were associated with other signs of regenera-
the RB group and 2 from the RBc group, compared to tion (see Fig. 5 for details).
5 of 7 biopsies in the B group (5).
Tenascin C
Regeneration

Figure 5 displays the proportion of fibers from the Tenascin-C analysis revealed a greater extent of immu-
stimulated leg that were positive for embryonic myosin noreactivity in the B group when compared with the RB
or laminin '5. Two biopsies from the control leg of the and RBc groups (Fig. 7). The mean value for control
RBc trial contained 1 embryonic myosin (F1.652)( biopsies was 0.2%.

Figure 6. Five consecutive 10-%m sections (rows

1–5) of a biopsy taken 30 d after a single bout of
ES-induced isometric contractions (RBc group),
clearly demonstrating ongoing regeneration.
Each row is a series of images of 1 double-stained
section, the 2 individual stainings in the left and
middle panels merged digitally in the right
panel. It was observed that the areas of laminin
'5( endomysial staining exhibited other signs of
regeneration, such as CD56( fibers, a high density
of CD68( cells in the ECM, embryonic myosin
(Emb Myo)( fibers, occasional small fibers (arrow-
heads), and central nuclei. A laminin( dystro-
phin( membrane (arrow), observed on these se-
rial sections to encroach into one fiber, was seen to
define a separate fiber on a section deeper into the
biopsy (row 6). Most of the affected fibers (aster-
isks) were negative for type I myosin (Myo I; row 6)
and positive for type II myosin (Myo II; row 7).
Lam, laminin. Scale bar * 100 %m.

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1949

 Macrophages and cellular activity

A significantly greater number of CD68( cells per

square millimeter of tissue (Fig. 9) was detected in the
stimulated vs. control leg of B (P)0.05). Comparison
of RB and RBc data revealed a significant overall effect
(P*0.0002) with significant differences between ES
and control in both groups (P)0.05). The fraction of
CD68( immunoreactive area demonstrated the same
pattern (data not shown). A significant overall effect
(P)0.0001) was detected for the number of Ki67( cells
per 100 fibers (Fig. 9), with significant differences
between ES and control in all 3 groups (P)0.05).

Transmission electron microscopy

A significantly greater extent of z-line disruption was

observed in the stimulated leg of the B and RB vs. RBc
trials (see Fig. 10). Out of the 26 control biopsies
evaluated in the B, RB, and RBc groups, 3 were
categorized under the grade 0.5 and all others were 0.
It is worth noting for the stimulated leg of the RBc trial
that discrete and inconsistent irregularities in z-line
integrity, graded 0.5 to 1, were observed in 9 of the 11
samples. These include occasional deviation in the
strict z-line alignment and z lines out of register with
adjacent sarcomeres, as displayed in Fig. 10.

Figure 7. Tenascin-C immunoreactivity. Top panel: scatter Gene expression

plot (with median bars) indicates significantly higher levels
of tenascin-C immunoreactivity 48 h following a single bout mRNA levels and details of statistical outcomes are
(B group) compared to a repeated bout (RB group) of ES
or the response 30 d after a single bout (RBc group). presented in Figs. 11–14. Overall, of the 25 targets
Bottom panels: images illustrate varying extents of tenas- analyzed, 5 demonstrated an attenuated response to RB
cin-C immunoreactivity from control (Con) muscle and trial when compared to B trial: chemokine ligand 2
muscle subjected to ES (Stim). Percentage of immunoreac- (CCL2), connective tissue growth factor (CTGF), HSPs;
tive area for each image was quantified; respective results for 12 demonstrated significantly elevated levels 30 d after
these representative images (a portion of each image is shown) a single bout of ES (RBc group) vs. their respective
are given. Scale bar * 200 %m. **P ) 0.01, ***P ) 0.001;
controls: CTGF, TGF-+, CCL2, IGF1-Ea, IGF1-Eb, IGF1-
Dunn’s multiple comparison test.
Ec, laminin-+1, laminin-+2, and collagen types I, III, IV,
and XII; 5 targets displayed an attenuated response in
Collagen types I and III the ES leg following RB trial when compared to RBc:
CTGF, laminin-+2, and collagen types I, III, and XII.
As can be seen in Fig. 11, HSPs demonstrated an
Collagen type I and III staining patterns are displayed attenuated response in the RB group when compared
and described in Fig. 8. From comparisons of the to the B group. At 48 h after a single bout of ES (B
control and stimulated muscle, a more intense immu- group), '+-crystallin and HSP27 mRNA levels were
noreactivity for both collagen types was apparent !3-fold significantly higher in ES vs. control treatment
around small fibers, only present in the stimulated groups, while HSP70 displayed a !10-fold elevation.
muscle and therefore likely to be regions of ongoing Matricellular tenascin-C and CTGF values were also
regeneration. elevated in the B vs. RB or RBc groups. [A tendency for
leg $ group interaction (P*0.11) for tenascin C was
SCs detected.]
Extracellular matrix components (Fig. 12) generally
demonstrated greatest mRNA levels 30 d after a single
No difference between ES and control treatment was bout (RBc group), with the greatest !6- to 9-fold
observed in the B group (Table 2). For the RB and RBc response observed in collagen types I and III and
groups, a main effect of leg (P)0.01), but not group or laminin-+1 at this time point. For collagen types IV and
interaction, was observed for SC content per fiber, and XII and laminin-+1, this was the only time point where
when the SC content associated with type I or type II a significant difference vs. control was detected. A
fibers was examined. tendency for elevated mRNA levels in the ES leg vs.

1950 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
 Figure 8. Immunohistochemical staining pat-
tern of collagen types I and III on cross-
sections of skeletal muscle 30 d after a single
bout of ES-induced isometric contractions
(RBc group). Two serial sections from the
stimulated and control legs of one individual
were double-stained with laminin and type I
(series 1 and 2) or type III (series 3 and 4)
collagen. The collagen (a) and laminin (b)
stainings are displayed separately and as com-
puter-generated merged images (c). Both colla-
gen types were observed in perimysium (ar-
rows) and endomysium (arrowheads), with
negligible staining of capillaries. Endomysium
around regenerating fibers appeared to dem-
onstrate more intense immunoreactivity for col-
lagen types I and III (circled area). Scale bars *
200 %m.

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1951

 TABLE 2. Assessment of Pax7 SCs on cross-sections of muscle biopsies from 3 groups of volunteers
48 h after a single bout (B group) or repeated bout (RB group) of ES-induced contractions separated
by 1 mo

B group RB group RBc group

Type Control ES Control ES Control ES

Pax7/f 0.15 # 0.03 0.14 # 0.02 0.11 # 0.01 0.16 # 0.01 0.07 # 0.01 0.13 # 0.02*
Pax7/ft I 0.10 # 0.01 0.14 # 0.01 0.07 # 0.01 0.11 # 0.01*
Pax7/ft II 0.12 # 0.01 0.18 # 0.02 0.08 # 0.01 0.16 # 0.02*

Control RBc group was biopsied 30 d after a single bout. Table shows number of SCs associated with
the 2 main fiber types (ft I and ft II), or combined fiber types (f), in the ES leg and control leg. Values
are means # se. Statistical analysis of data from B group was performed separately from analysis of data
from RB and RBc groups. SC content in control and ES legs of B group was found to be similar. *P )
0.01, main effect of leg.

control following B trial was seen for collagen type I after a single bout (B group). With regard to inflam-
(P*0.06) and type III (P*0.11). matory targets (Fig. 14), CCL2 [monocyte chemoattrac-
Myogenic-related factors (Fig. 13) did not demon- tant protein 1 (MCP-1)] was the only one to demon-
strate a dramatic response at the time points examined
in this study. Myf6 (MRF4) and p21 demonstrated a
main effect of leg (control)ES), and a tendency for a
main effect of leg (P*0.054) was observed for myoge-
nin. A suppression of c-met and HGF gene expression
levels was detected in the RB vs. RBc groups.
The greatest mRNA levels of the 3 IGF-1 isoforms
(Fig. 14) were observed in the stimulated leg of the RBc
trial, with IGF1-Ec displaying the largest (!3-fold)
elevation vs. control treatment. No elevation was seen
in any of the IGF-1 isoforms with ES vs. control at 48 h

Figure 10. Top panel: scatter plot (with median bars) displaying
extent of z-line disruption (0 * none, 3 * severe disruption)
following a single bout (B group) or repeated bout (RB group)
of ES. Biopsies in RBc group were collected 30 d after a single
Figure 9. Scatter plots (with median bars) displaying the bout. Bottom panels: transmission electron micrographs of longi-
number of macrophages (CD68( cells) per square millimeter tudinal sections of human medial gastrocnemius from RBc group
of biopsy cross-section (A) and the number of Ki67( cells per biopsies. Representative images from 2 subjects (Sub 1 and 2) are
100 fibers (B) from electrically stimulated (ES) or control displayed, along with images from control (Con) legs of the same
(Con) muscle. Broken vertical line in panel A indicates that individuals. Images illustrate that, while strict z-line alignment and
biopsies in B group were subject to separate immunohisto- register have generally been restored, discrete and subtle differ-
chemical and statistical analyses from RB and RBc groups. ences are still visible at this time point. Scale bars * 1 %m. **P )
*P ) 0.05; **P ) 0.01. 0.01; ***P ) 0.001.

1952 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
 Figure 11. HSP and matricellular protein
gene expression levels in control (Con)
muscle and muscle subjected to a single
bout (B group) or a repeated bout (RB
group) of ES 1 mo after first bout. Biopsies
in RBc group were collected 30 d after a
single bout of ES. mRNA levels of GAPDH,
'+-crystallin, HSP27, HSP70, FAK1, tenas-
cin C, CTGF and TGF-+ are presented,
expressed relative to RPLP0 mRNA. Data
are back-transformed geometric means #
se, displayed on a logarithmic scale y axis.

strate a response in the B group, where a !4-fold content. The outcome of the delayed response is an
elevation vs. control was detected. attenuated early response following subsequent reexpo-
sure to a damaging stimulus.

DISCUSSION Normalization of gene expression values

The present findings indicate that, in human skeletal In an attempt to validate RPLP0 mRNA as an internal
muscle, acutely induced damage by ES results in an reference for mRNA levels throughout the experi-
initiation of deadhesive, disassembly, and disorganization ment, we first normalized another housekeeping mRNA,
responses in the contractile connective tissue, followed by GAPDH, with RPLP0. However, the GAPDH-RPLP0 ratio
a delayed anabolic response in the supporting ECM, showed an alteration in the expression pattern resulting
providing the basis for a prolonged strengthening of the in apparent lower GAPDH mRNA levels in the ES leg vs.
muscle matrix (Fig. 15). This results in a diminished control leg of the B and RB groups. (P)0.05; Fig. 11).
degradation of the skeletal muscle and its connective mRNA levels of ES and control in the RBc group were
tissue when subjected to a similar damaging exercise at a similar. GAPDH mRNA levels in muscle tissue are usually
later stage. The fact that sustained enhancement of SC !10 times higher than in other tissues of the body (31).
content was observed in damaged muscle beyond muscle Furthermore, a huge amount of invading cells was ob-
fiber regeneration suggests that crosstalk between muscle served after ES. Taken together, we argue that in muscle
stem cell activity and the matrix components of skeletal cell lysates with high GAPDH mRNA levels, invading cells
muscle exists, with the overall goal to protect the individ- with lower GAPDH mRNA levels, such as macrophages,
ual muscle fiber from damage on reexposure to extreme would lead to a dilution of total GAPDH mRNA within this
loading. We suggest a sequential series of responses to cell lysate, thus resulting in lower concentrations of GAPDH
damage in skeletal muscle, with an early response favoring mRNA, as was clearly observed after each bout of stimulation
muscle damage, inflammation, disassembly, and disorga- (B and RB groups) but not after a longer period of recovery
nization. Following this, a more anabolic matrix-oriented (RBc group). Hence, normalization of GAPDH with RPLP0
response is seen, which includes growth factors and showed a changing expression of a housekeeping gene,
collagen, in concert with a prolonged elevation of SC which we assume to be caused by a simple dilution effect of

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1953

Figure 12. Comparisons of extracellular matrix gene expression levels between control (Con) muscle and muscle subjected to
a single bout (B group) or a repeated bout (RB group) of ES 1 mo after first bout. Biopsies in RBc group were collected 30 d
after single bout of ES. mRNA levels of collagen types I (Col1A), III (Col3A1), IV (Col4A1), XII, and laminin-+1 (LAMB 1) and
-+2 (LAMB 2) are presented, expressed relative to RPLP0 mRNA. Data are back-transformed geometric means # se, displayed
on a logarithmic scale y axis.

GAPDH. We therefore conclude that, although not perfect, We observed an attenuated increase in circulating CK
the most reasonable was to normalize all other targets to following the second bout when compared to the first
RPLP0. It should be noted, though, that other muscle- bout, in line with the only other human study (32)
specific mRNA might be subject to the same dilution effect investigating repeated bouts of electrically stimulated isomet-
(seen as a 2-fold drop in B). For example, the similar drop in ric contractions. Measurement of muscle HSP gene expres-
laminin-+2 in the B group might accordingly be explained sion in the 3 sets of biopsies in the present study provides
by the dilution effect, whereas the increase in the RBc group further evidence for a repeated bout effect with a large
is likely to be a “real” effect. up-regulation following B treatment and an attenuated in-
crease in the small HSPs (HSP27 and '+-crystallin) following
Attenuated early damage response to a repeated RB treatment. Previous studies have either reported an
damaging stimulus attenuated HSP response with a repeated bout of exercise
(11, 13), or no attenuation with the second bout (10–13),
This is the first study of human muscle biopsies subjected providing an unclear picture with regard to the response of
to 2 bouts of electrically stimulated isometric contractions. HSPs to repeated bouts. While our HSP results do not

Figure 13. Comparisons of SC-related

gene expression levels between con-
trol (Con) muscle and muscle sub-
jected to a single bout (B group) or a
repeated bout (RB group) of ES 1 mo
after first bout. Biopsies in RBc group
were collected 30 d after a single bout
of ES. mRNA levels of c-met, HGF
(HGF1), myogenin, Myf6 (MRF4),
and p21 are presented, expressed relative to RPLP0 mRNA. Data are back-transformed geometric means # se,
displayed on a logarithmic scale y axis.

1954 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
Figure 14. IGF-1 and inflammatory-related gene expression levels in control (Con) muscle and muscle subjected to a single bout
(B group) or a repeated bout (RB group) of ES 1 mo after first bout. Biopsies in RBc group were collected 30 d after a single
bout of ES. mRNA levels of IGF1-Ea (IGF1a), IGF1-Eb (IGF1b), IGF1-Ec (IGF1c; MGF), CCL2 (MCP-1), interleukin (Il)-1+, and
tumor necrosis factor ' (TNF-') are presented, expressed relative to RPLP0 mRNA. Data are back-transformed geometric
means # se, displayed on a logarithmic scale y axis.

provide support for a protective role of HSPs against muscle previously. The immunohistochemistry data generally
damage with electrically stimulated isometric contractions, mirror the gene expression findings and not only
they do, together with the data on CK, CCL2, and the suggest a greater tenascin-C response following a single
qualitative assessment of macrophage-infiltrated and bout than a repeated bout but also reveal sustained
desmin" fibers, provide strong support for the occurrence tenascin-C up-regulation 30 d into muscle regenera-
of less muscle damage after the second bout in this model. tion. The presence of tenascin-C protein, as well as
The observation of a similar extent of z-line disruption after during embryogenesis (33), has been observed in re-
the single and repeated bouts, though, suggests more intri- sponse to mechanical stress, such as previously reported
cate adaptations in providing protection against ES-induced in loaded human (4, 24, 28) and chicken muscle (27),
damage at the level of the sarcomere. or at mRNA level (28). In contrast to the adhesion-
Tenascin C rapidly creates a deadhesive ECM environment promoting ECM proteins, the role of matricellular
tenascin C in skeletal muscle is in deadhesion, the
The response of tenascin C at protein and mRNA levels process of disassembly of focal adhesion complexes,
in loaded human muscle has not been investigated which is believed to promote a more favorable environ-

Figure 15. Schematic illustration of human

skeletal muscle responses to damaging ES (red
arrows). In response to the first bout, marked
muscle damage and associated early responses
are observed (orange line), followed by a de-
layed anabolic response of the muscle ECM
(black line). This late response raises the resis-
tance toward muscle damage and disorganiza-
tion after a second repeated bout in the same
muscle. Solid lines represent response to the
first bout; broken lines illustrate how a re-
peated bout may alter this response. At top left,
a muscle fiber (mf) with myonuclei (dark gray
circles) and SCs (green circles) is depicted
during the early response, illustrating the dis-
organized ECM and damage to the myofibrils
and sarcolemma (sl). Myofiber at top right
represents the late phase, characterized by a
higher SC content, a repaired sarcolemma, and
a strengthened ECM, factors that are likely to
be involved in protecting the muscle from
damage on exposure to subsequent injuring
stimuli.

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1955

ment for cell motility, survival, and repair (26, 29). In that the degree of muscle damage and associated
line with these data, we also detected up-regulated gene inflammation in the present experiment resulted in
expression of CTGF (CCN2), a matricellular protein signaling that suppresses the usual immediate IGF-1
facilitating migration and chemotaxis in injured envi- response and does not allow any increased expression
ronments (34). It is possible that altered activity of until later stages of regeneration when accelerated
matricellular proteins, central modulators of the ECM, formation of new matrix tissue is needed. While it has
and its resident or migrating cells, plays an important been reported that IGF-1Ec (MGF) in particular re-
role in repair and protection of human skeletal muscle sponds within hours of the stimulus, IGF-1Ea (and
on re-exposure to a muscle-damaging stimulus. perhaps also IGF-1-Eb) is typically more prominent 1 to
several days later (37, 39 – 43). The elevated levels of
ECM remodeling is a late event in regenerating IGF-1Ea mRNA observed in the present study at such a
human muscle late stage after damage are perhaps not surprising in
light of the extent of ongoing muscle remodeling and
Our analysis of adhesion-promoting matrix compo- the suggested role of this IGF-1 isoform, not only in
nents revealed a general pattern for a stronger up- differentiation (41, 43) but perhaps more importantly
regulation at 30 d than at 2 d postexercise. Accordingly, in moderating inflammatory and fibrotic activity during
it appears that ECM remodeling is not prioritized
the later stages of muscle regeneration (44). Alterna-
immediately after exercise. When compared to the B
tively, it has been suggested that macrophages are the
group, the RBc group displayed a stronger up-regula-
source of IGF-1 in regenerating muscle (45), in line
tion in gene expression of collagen types I and III and
laminin-+2, suggesting that while ECM remodeling, with our heightened numbers of macrophages at this
rather than a marked synthesis, might be initiated in time. Surprisingly though, mRNA levels of MGF were
the early phase of recovery, it is not until the later stages also elevated at 30 d compared to the control leg, either
that matrix structure synthesis is augmented. Further- implying that proliferation of some cell types, poten-
more, the lower levels of ECM mRNA in the RB group tially SCs, is still required at this stage, or suggesting
compared to the RBc group suggest a partial suppres- that MGF plays a role in the regeneration of muscle
sion after the second bout of exercise, supporting a ECM. In support of the latter, increased gene expres-
complex reordering at the level of gene expression. sion of MGF has been reported in loaded rat tendon,
This finding is in line with a matrix synthesis suppress- where a role for this IGF-1 isoform in collagen adapta-
ing response in the early stage after damage, and thus tion was suggested (46) and also more recently in
the response is diminished after the second bout of mouse skeletal muscle (47). Taken together, the data
exercise, where the structural changes induced by the presented here provide new insight into the magni-
first bout have already taken place. While we cannot tude, time course, and potential role of IGF-1 isoform
rule out that changes in type IV collagen and laminin responses in regenerating human skeletal muscle.
might be related to capillary remodeling, the preva-
lence of collagen types I and III in endomysium and
perimysium (confirmed by immunohistochemistry in A persistent elevation of SC content at 30 d
the present study), together with the apparent intensi-
fied staining around regenerating fibers, suggests that SC content of the regenerating muscle at 30 d was
the increased collagen gene expression reflects remod- significantly greater than the corresponding control
eling of connective tissue surrounding individual fibers,
muscle, which is a novel observation on the time course
as reported earlier (21), as well as fiber fascicles.
of stem cell activity in human skeletal muscle. It is not
Further support for this finding was provided by rarely
known how long this enhancement of the SC pool
reported staining of laminin-'5 around some fibers.
This staining pattern has been described previously in would be retained and indeed if the unneeded cells
the muscle of power lifters (35) and likely indicates would be permitted to remain as SCs, or be forced to
processes reminiscent of development (36), which, differentiate or undergo apoptosis as the muscle rees-
together with the ECM gene expression data, provides tablishes its predamage environment. It is possible that
strong new evidence of ongoing endomysial remodel- the elevated levels of tenascin C, CTGF, and the 3 IGF-1
ing 30 d after a muscle injuring stimulus. isoforms contribute to the survival of the new SCs. In
addition, contact between SCs and a mature myofiber
A delayed IGF-1 response mirrors the ECM time course has been reported to be a potent factor signaling SC
quiescence (15) and the analysis of regenerating mark-
The response of IGF-1 to repeated bouts of exercise in ers in the present study indicates that many myofibers
humans and at a relatively late stage of muscle regen- had not yet reached full maturity and thus potentially
eration has not been documented previously. Interest- lacked the quiescence-signaling factor. While further
ingly, the mRNA response for IGF-1 in this study was investigation is required to clarify the reason behind
delayed following the damaging exercise compared to the elevated SC content, it is interesting, in light of the
the more rapid response observed with more physio- delayed ECM remodeling, to consider that crosstalk
logical nondamaging exercise (37–39). Thus, it seems between the muscle ECM and the SC exists.

1956 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
Slow resolution of local inflammation in regenerating area of the triceps surae muscles (51), resulting in a
muscle minor contribution to the total force produced by the
triceps surae muscle group during a voluntary contrac-
Levels of MCP-1 (CCL2) were higher in the stimulated tion. While stimulation current is generally not consid-
leg of RBc vs. control, indicating sustained chemotactic ered a good indication of contraction intensity due to
signaling of the muscle at this time point. The immu- such variables as the thickness of subcutaneous fat, skin
nohistochemical analysis of macrophages confirmed impedance, and the positioning of the electrodes with
this activity and, in line with the SC data, indicated respect to the motor point, we cannot rule out the
continuing enhanced cellular activity in the regenerat- possibility that the RB group was subjected to greater
ing muscle. Over 30 yr ago it was reported that collagen contraction intensity than the B and RBc groups.
serves as a chemotactic stimulus for monocytes (48), However, the analyses performed on these biopsies do
contributing to evidence for an active role of muscle not indicate a greater extent of damage in this group
ECM, in concert with inflammatory cells, in adaptation and, given that the individuals in the RB group were
to loading. An additional finding of the present study stimulated equally during bouts 1 and 2, this potential
was the strong indication of incomplete repair 30 d limitation does not detract from the main findings of
after a damaging stimulus to the muscle, providing this study, summarized below. Lastly, collecting a mus-
novel documentation of the time course of regenera- cle biopsy from the control leg at the time of sampling
tion of human skeletal muscle. from the stimulated leg, instead of repeatedly biopsying
the stimulated leg, not only removes the potential for
Strengths and limitations of the model signal contamination from a previous biopsy (52) but
also allows us to interpret differences between the legs
as arising from the ES and not from diurnal or dietary
In general terms, given that ES does not respect the
variation, or altered levels of daily physical activity (52).
order of motor unit recruitment known for voluntary
contractions, further work is required to confirm to
what extent the current findings can be transferred to
a model of voluntary muscle contraction. However, it is CONCLUSIONS
likely that the more indiscriminate recruitment during
ES when compared to voluntary contractions would This study confirms the occurrence of a protective
result in initiation of a stronger, but not necessarily effect of an initial muscle-damaging stimulus against
dissimilar, repair response (4). The reason for using later injury in human skeletal muscle by an attenuated
our ES model was to minimize the interindividual response in circulating and biopsy markers following
variation in muscle damage observed with voluntary the second bout. We observed at 30 d after 1 bout of ES
contraction models. A further strength of the model is previously undocumented dramatic and subtle changes
that, due to its relatively small size, the medial gastroc- in the structure and activity of muscle and its surround-
nemius muscle can be stimulated wholly and uniformly ing connective tissue and identified new potential
(49), thus eliminating concerns about biopsy sampling players in this protection, which sheds light on the time
from an unaffected part of the muscle, as is often a course of muscle regeneration and its resistance toward
concern with biopsying larger human muscles. With future damage. Together, these findings demonstrate
regard to subjecting the same muscle to repeated bouts an ordered response to a damaging stimulus (Fig. 15).
of ES, direct evidence for the occurrence of a lesser An initial muscle fiber damaging, deadhesive, and
extent of damage after the second bout was provided by disassembly response with early elevated CK, soreness,
the attenuation in response of circulating CK, while z-line disruption, HSPs, tenascin, and inflammation
indirect evidence was apparent in the comparison of markers occurs, which favors a disorganization of the
biopsy markers in the B and RB groups. Despite this, no muscle-matrix tissue. This was time-wise followed by a
attenuation in muscle soreness after the second bout delayed elevation in expression of anabolic matrix
was observed in this study—a finding that was unex- growth factors and collagen, favoring ECM strengthen-
pected and is in contrast to a previous ES study (32). ing, as well as a continued elevation of SC content.
While this finding can perhaps be explained by the When the muscle was reexposed to the same damaging
differences between the two studies in the number of stimulus 1 mo later, the disorganizing responses were
contractions, the muscle group stimulated, and the markedly reduced compared to the initial response,
length of time between the two bouts, further work is which suggests a protective role of the intramuscular
required to explain why muscle soreness was not atten- connective tissue against future pronounced muscle
uated with our study design. Examination of the re- injury. Furthermore, the partial suppression of gene
sponse of pain-related substances such as bradykinin expression of CTGF and 4 components of the ECM
and nerve growth factor in our model may provide following the repeated bout suggests a complex and
some insight (50). In relation to comparing the 3 continual reordering of remodeling events of skeletal
groups, force measurements would have provided a muscle and its connective tissue, depending on expo-
solid base for comparison. However, a limitation of our sure to periods of loading or recovery. Finally, the
model is that the medialis gastrocnemius muscle only delayed SC response coupled to the increased anabolic
comprises 20% of the total physiological cross-sectional ECM response raises the possibility that SCs are in-

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1957

volved in crosstalk between myofibrillar damage and 13. Hubal, M. J., Chen, T. C., Thompson, P. D., and Clarkson, P. M.
(2008) Inflammatory gene changes associated with the repeat-
matrix adaptation in human skeletal muscle. ed-bout effect. Am. J. Physiol. 294, R1628 –1637
14. Lapointe, B. M., Fremont, P., and Cote, C. H. (2002) Adaptation
Casper Mortensen and Christine Bodilsen are gratefully to lengthening contractions is independent of voluntary muscle
acknowledged for help with recruiting subjects and collecting recruitment but relies on inflammation. Am. J. Physiol. 282,
data. The excellent technical assistance involved in taking the R323–R329
blood samples and preparing the muscle biopsies (Camilla 15. Bischoff, R. (1990) Interaction between satellite cells and skel-
Hansen) and the preparation of samples for electron micros- etal muscle fibers. Development 109, 943–952
copy (Mary-Ann Gleie) is warmly acknowledged. The F1.652, 16. Sarasa-Renedo, A., and Chiquet, M. (2005) Mechanical signals
regulating extracellular matrix gene expression in fibroblasts.
A4.951, and A4.74 antibodies developed by Helen M. Blau Scand. J. Med. Sci. Sports 15, 223–230
were obtained from the Developmental Studies Hybridoma 17. Kuang, S., Gillespie, M. A., and Rudnicki, M. A. (2008) Niche
Bank (University of Iowa, Iowa City, IA, USA), developed regulation of muscle satellite cell self-renewal and differentia-
under the auspices of the U.S. National Institute of Child tion. Cell Stem Cell 2, 22–31
Health and Human Development and maintained by the 18. Alexakis, C., Partridge, T., and Bou-Gharios, G. (2007) Implica-
Department of Biological Sciences, University of Iowa. Fund- tion of the satellite cell in dystrophic muscle fibrosis: a self-
ing is gratefully acknowledged from the Nordea Foundation perpetuating mechanism of collagen overproduction. Am. J.
(Healthy Aging grant), the Danish Agency for Science Tech- Physiol. Cell Physiol. 293, C661–C669
nology and Innovation (Medical Research Council), the 19. Miller, B. F., Olesen, J. L., Hansen, M., Dossing, S., Crameri,
R. M., Welling, R. J., Langberg, H., Flyvbjerg, A., Kjaer, M.,
Danish Rheumatism Association (R74-A1088), and the Lund- Babraj, J. A., Smith, K., and Rennie, M. J. (2005) Coordinated
beck Foundation. collagen and muscle protein synthesis in human patella tendon
and quadriceps muscle after exercise. J. Physiol. 567, 1021–1033
20. Suominen, H., and Heikkinen, E. (1975) Enzyme activities in
muscle and connective tissue of M. Vastus lateralis in habitually
training and sedentary 33 to 70-year-old men. Eur. J. Appl.
REFERENCES Physiol. Occup. Physiol. 34, 249 –254
21. Mackey, A. L., Donnelly, A. E., Turpeenniemi-Hujanen, T., and
1. Round, J. M., Jones, D. A., and Cambridge, G. (1987) Cellular Roper, H. P. (2004) Skeletal muscle collagen content in humans
infiltrates in human skeletal muscle: exercise induced damage after high-force eccentric contractions. J. Appl. Physiol. 97,
as a model for inflammatory muscle disease? J. Neurol. Sci. 82, 197–203
1–11 22. Kjaer, M. (2004) Role of extracellular matrix in adaptation of
2. Hikida, R. S., Staron, R. S., Hagerman, F. C., Sherman, W. M., tendon and skeletal muscle to mechanical loading. Physiol. Rev.
and Costill, D. L. (1983) Muscle fiber necrosis associated with 84, 649 – 698
human marathon runners. J. Neurol. Sci. 59, 185–203 23. Mackey, A. L., Donnelly, A. E., and Roper, H. P. (2005) Muscle
3. Jones, D. A., Newham, D. J., Round, J. M., and Tolfree, S. E. connective tissue content of endurance-trained and inactive
(1986) Experimental human muscle damage: morphological individuals. Scand. J. Med. Sci. Sports 15, 402– 408
changes in relation to other indices of damage. J. Physiol. 375, 24. Crameri, R. M., Langberg, H., Teisner, B., Magnusson, P.,
435– 448 Schroder, H. D., Olesen, J. L., Jensen, C. H., Koskinen, S.,
4. Crameri, R. M., Aagaard, P., Qvortrup, K., Langberg, H., Suetta, C., and Kjaer, M. (2004) Enhanced procollagen process-
Olesen, J., and Kjaer, M. (2007) Myofibre damage in human ing in skeletal muscle after a single bout of eccentric loading in
skeletal muscle: effects of electrical stimulation vs voluntary humans. Matrix Biol. 23, 259 –264
contraction. J. Physiol. (Lond.) 583, 365–380 25. Lapier, T. K., Burton, H. W., Almon, R., and Cerny, F. (1995)
5. Mackey, A. L., Bojsen-Moller, J., Qvortrup, K., Langberg, H., Alterations in intramuscular connective tissue after limb casting
Suetta, C., Kalliokoski, K. K., Kjaer, M., and Magnusson, S. P. affect contraction-induced muscle injury. J. Appl. Physiol. 78,
(2008) Evidence of skeletal muscle damage following electri- 1065–1069
cally stimulated isometric muscle contractions in humans. 26. Murphy-Ullrich, J. E. (2001) The de-adhesive activity of matri-
J. Appl. Physiol. 105, 1620 –1627 cellular proteins: is intermediate cell adhesion an adaptive
6. Mackey, A. L., Kjaer, M., Charifi, N., Henriksson, J., Bojsen- state? J. Clin. Invest. 107, 785–790
Moller, J., Holm, L., and Kadi, F. (2009) Assessment of satellite 27. Fluck, M., Tunc-Civelek, V., and Chiquet, M. (2000) Rapid and
cell number and activity status in human skeletal muscle biop- reciprocal regulation of tenascin-C and tenascin-Y expression by
sies. Muscle Nerve 40, 455– 465 loading of skeletal muscle. J. Cell Sci. 113, 3583–3591
7. Nosaka, K., Sakamoto, K., Newton, M., and Sacco, P. (2001) 28. Mikkelsen, U. R., Schjerling, P., Helmark, I. C., Reitelseder, S.,
How long does the protective effect on eccentric exercise- Holm, L., Skovgaard, D., Langberg, H., Kjaer, M., and Heine-
induced muscle damage last? Med. Sci. Sports Exerc. 33, 1490 – meier, K. M. (2010) Local NSAID infusion does not affect
1495 protein synthesis and gene expression in human muscle after
8. McHugh, M. P. (2003) Recent advances in the understanding of eccentric exercise. [E-pub ahead of print] Scand. J. Med. Sci.
the repeated bout effect: the protective effect against muscle Sports doi: 10.1111/j.1600-0838.2010.01170.x
damage from a single bout of eccentric exercise. Scand. J. Med. 29. Fluck, M., Mund, S. I., Schittny, J. C., Klossner, S., Durieux,
Sci. Sports. 13, 88 –97 A. C., and Giraud, M. N. (2008) Mechano-regulated tenascin-C
9. Clarkson, P. M., Byrnes, W. C., Gillisson, E., and Harper, E. orchestrates muscle repair. Proc. Natl. Acad. Sci. U. S. A. 105,
(1987) Adaptation to exercise-induced muscle damage. Clin. Sci. 13662–13667
(Lond.) 73, 383–386 30. Mackey, A. L., Andersen, L. L., Frandsen, U., Suetta, C., and
10. Thompson, H. S., Clarkson, P. M., and Scordilis, S. P. (2002) Sjogaard, G. (2010) Distribution of myogenic progenitor cells
The repeated bout effect and heat shock proteins: intramuscu- and myonuclei is altered in women with vs. those without
lar HSP27 and HSP70 expression following two bouts of eccen- chronically painful trapezius muscle. J. Appl. Physiol. 109, 1920 –
tric exercise in humans. Acta Physiol. Scand. 174, 47–56 1929
11. Vissing, K., Bayer, M. L., Overgaard, K., Schjerling, P., and 31. Barber, R. D., Harmer, D. W., Coleman, R. A., and Clark, B. J.
Raastad, T. (2009) Heat shock protein translocation and expres- (2005) GAPDH as a housekeeping gene: analysis of GAPDH
sion response is attenuated in response to repeated eccentric mRNA expression in a panel of 72 human tissues. Physiol.
exercise. Acta Physiol. (Oxf.) 196, 283–293 Genomics 21, 389 –395
12. Paulsen, G., Lauritzen, F., Bayer, M. L., Kalhovde, J. M., Ugel- 32. Aldayel, A., Jubeau, M., McGuigan, M. R., and Nosaka, K. (2010)
stad, I., Owe, S. G., Hallen, J., Bergersen, L. H., and Raastad, T. Less indication of muscle damage in the second than initial
(2009) Subcellular movement and expression of HSP27, al- electrical muscle stimulation bout consisting of isometric con-
phaB-crystallin, and HSP70 after two bouts of eccentric exercise tractions of the knee extensors. Eur. J. Appl. Physiol. 108,
in humans. J. Appl. Physiol. 107, 570 –582 709 –717

1958 Vol. 25 June 2011 The FASEB Journal ! www.fasebj.org MACKEY ET AL.
33. Jones, F. S., and Jones, P. L. (2000) The tenascin family of ECM aro, M., and Musaro, A. (2007) Local expression of IGF-1
glycoproteins: structure, function, and regulation during embry- accelerates muscle regeneration by rapidly modulating inflam-
onic development and tissue remodeling. Dev. Dyn. 218, 235–259 matory cytokines and chemokines. FASEB J. 21, 1393–1402
34. Chen, C. C., and Lau, L. F. (2009) Functions and mechanisms of 45. Lu, H., Huang, D., Saederup, N., Charo, I. F., Ransohoff, R. M.,
action of CCN matricellular proteins. Int. J. Biochem. Cell Biol. 41, and Zhou, L. (2011) Macrophages recruited via CCR2 produce
771–783 insulin-like growth factor-1 to repair acute skeletal muscle
35. Eriksson, A., Lindström, M., Carlsson, L., and Thornell, L. E. (2006) injury. FASEB J. 25, 358 –369
Hypertrophic muscle fibers with fissures in power-lifters; fiber splitting 46. Heinemeier, K. M., Olesen, J. L., Schjerling, P., Haddad, F.,
or defect regeneration? Histochem. Cell Biol. 126, 409–417 Langberg, H., Baldwin, K. M., and Kjaer, M. (2007) Short-term
36. Gullberg, D., Tiger, C. F., and Velling, T. (1999) Laminins strength training and the expression of myostatin and IGF-I
during muscle development and in muscular dystrophies. Cell. isoforms in rat muscle and tendon: differential effects of specific
Mol. Life Sci. 56, 442– 460 contraction types. J. Appl. Physiol. 102, 573–581
37. Hameed, M., Orrell, R. W., Cobbold, M., Goldspink, G., and 47. Barton, E. R., DeMeo, J., and Lei, H. (2010) The insulin-like
Harridge, S. D. R. (2003) Expression of IGF-I splice variants in growth factor (IGF)-I E-peptides are required for isoform-
young and old human skeletal muscle after high resistance specific gene expression and muscle hypertrophy after local
exercise. J. Physiol. (Lond.) 547, 247–254 IGF-I production. J. Appl. Physiol. 108, 1069 –1076
38. Bamman, M. M., Petrella, J. K., Kim, J. S., Mayhew, D. L., and 48. Postlethwaite, A. E., and Kang, A. H. (1976) Collagen-and
Cross, J. M. (2007) Cluster analysis tests the importance of collagen peptide-induced chemotaxis of human blood mono-
myogenic gene expression during myofiber hypertrophy in cytes. J. Exp. Med. 143, 1299 –1307
humans. J. Appl. Physiol. 102, 2232–2239 49. Kalliokoski, K. K., Bojsen-Moller, J., Seppanen, M., Johansson, J.,
39. McKay, B. R., O’Reilly, C. E., Phillips, S. M., Tarnopolsky, M. A., and Kjaer, M., Teras, M., and Magnusson, S. P. (2007) Contraction-
Parise, G. (2008) Co-expression of IGF-1 family members with myo- induced [18F]-fluoro-deoxy-glucose uptake can be measured in
genic regulatory factors following acute damaging muscle lengthen- human calf muscle using high-resolution PET. Clin. Physiol.
ing contractions in humans. J. Physiol. 586, 5549–5560 Funct. Imaging 27, 239 –241
40. Haddad, F., and Adams, G. R. (2002) Exercise effects on muscle 50. Murase, S., Terazawa, E., Queme, F., Ota, H., Matsuda, T.,
insulin signaling and action: selected contribution: acute cellu- Hirate, K., Kozaki, Y., Katanosaka, K., Taguchi, T., Urai, H., and
lar and molecular responses to resistance exercise. J. Appl. Mizumura, K. (2010) Bradykinin and nerve growth factor play
Physiol. 93, 394 – 403 pivotal roles in muscular mechanical hyperalgesia after exercise
41. Yang, S. Y., and Goldspink, G. (2002) Different roles of the IGF-I (delayed-onset muscle soreness). J. Neurosci. 30, 3752–3761
Ec peptide (MGF) and mature IGF-I in myoblast proliferation 51. Fukunaga, T., Roy, R. R., Shellock, F. G., Hodgson, J. A., Day,
and differentiation. FEBS Lett. 522, 156 –160 M. K., Lee, P. L., Kwong-Fu, H., and Edgerton, V. R. (1992)
42. Yang, S., Alnaqeeb, M., Simpson, H., and Goldspink, G. (1996) Physiological cross-sectional area of human leg muscles based
Cloning and characterization of an IGF-1 isoform expressed in skeletal on magnetic resonance imaging. J. Orthop. Res. 10, 928 –934
muscle subjected to stretch. J. Muscle Res. Cell Motil. 17, 487–495 52. Vissing, K., Andersen, J. L., and Schjerling, P. (2005) Are
43. Hill, M., Wernig, A., and Goldspink, G. (2003) Muscle satellite exercise-induced genes induced by exercise? FASEB J. 19,
(stem) cell activation during local tissue injury and repair. J. 94 –96
Anat. 203, 89 –99
44. Pelosi, L., Giacinti, C., Nardis, C., Borsellino, G., Rizzuto, E., Received for publication November 22, 2010.
Nicoletti, C., Wannenes, F., Battistini, L., Rosenthal, N., Molin- Accepted for publication February 17, 2011.

REMODELED ECM PROTECTS MUSCLE AGAINST INJURY 1959