Am J Physiol Endocrinol Metab 316: E940–E947, 2019.

First published February 19, 2019; doi:10.1152/ajpendo.00414.2018.

RESEARCH ARTICLE

IL-6 release from muscles during exercise is stimulated by lactate-dependent

protease activity
X Pernille Hojman,1 Camilla Brolin,2 Nynne Nørgaard-Christensen,1 Christine Dethlefsen,1
Britt Lauenborg,1 Cecilie Køllner Olsen,1 Mette Marie Åbom,1 Thomas Krag,3 Julie Gehl,4,5
and Bente Klarlund Pedersen1
1
Centre of Inflammation and Metabolism and Centre for Physical Activity Research, Copenhagen University Hospital,
University of Copenhagen, Copenhagen, Denmark; 2Department of Cellular and Molecular Medicine, University of
Copenhagen, Copenhagen, Denmark; 3Copenhagen Neuromuscular Center, Copenhagen University Hospital, University of
Copenhagen, Copenhagen, Denmark; 4Center for Experimental Drug and Gene Electrotransfer, Department of Oncology and
Palliative Care, Zealand University Hospital, Roskilde, Denmark; and 5Department of Clinical Medicine, Faculty of Health
and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
Submitted 27 September 2018; accepted in final form 17 February 2019

Hojman P, Brolin C, Nørgaard-Christensen N, Dethlefsen C, INTRODUCTION

Lauenborg B, Olsen CK, Åbom MM, Krag T, Gehl J, Pedersen
BK. IL-6 release from muscles during exercise is stimulated by
lactate-dependent protease activity. Am J Physiol Endocrinol Metab
316: E940 –E947, 2019. First published February 19, 2019; doi:
10.1152/ajpendo.00414.2018.—IL-6 is secreted from muscles to the
circulation during high-intensity and long-duration exercise, where
muscle-derived IL-6 works as an energy sensor to increase release of
energy substrates from liver and adipose tissues. We investigated the
mechanism involved in the exercise-mediated surge in IL-6 during
exercise. Using interval-based cycling in healthy young men, swim-
ming exercise in mice, and electrical stimulation of primary human
muscle cells, we explored the role of lactate production in muscular
IL-6 release during exercise. First, we observed a tight correlation
between lactate production and IL-6 release during both strenuous
bicycling and electrically stimulated muscle cell cultures. In mice,
intramuscular injection of lactate mimicked the exercise-dependent
release of IL-6, and pH buffering of lactate production during exercise
attenuated IL-6 secretion. Next, we used in vivo bioimaging to
demonstrate that intrinsic intramuscular proteases were activated in
mice during swimming, and that blockade of protease activity blunted
swimming-induced IL-6 release in mice. Last, intramuscular injection
of the protease hyaluronidase resulted in dramatic increases in serum
IL-6 in mice, and immunohistochemical analyses showed that intra-
muscular lactate and hyaluronidase injections led to release of IL-6-
containing intramyocellular vesicles. We identified a pool of IL-6
located within vesicles of skeletal muscle fibers, which could be
readily secreted upon protease activity. This protease-dependent re-
lease of IL-6 was initiated by lactate production, linking training
intensity and lactate production to IL-6 release during strenuous
exercise.
interleukin-6; metalloproteinase; muscle-derived factors; myokines;
physical activity
Address for reprint requests and other correspondence: B. K. Pedersen,
Centre of Inflammation and Metabolism (CIM) and Centre for Physical
Activity Research (CFAS), Copenhagen University Hospital, 7641, Univ.
of Copenhagen, Blegdamsvej 9, DK-2100 Copenhagen, Denmark (e-mail:
Bente.Klarlund.Pedersen@regionh.dk).
E940 0193-1849/19 Copyright © 2019 the American Physiological Society
Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
Release of soluble factors plays a pivotal role in cell-to-cell
and organ-to-organ communication and is a widespread phe-
nomenon, extending across virtually all organs and cell types in
the body. In line with this, skeletal muscles have been demon-
strated to act as a secretory organ, secreting factors, known as
myokines, during exercise training (15). The best-described
myokine is interleukin 6 (IL-6), which increases in an expo-
nential fashion in response to exercise and then drops to baseline
levels following exercise cessation. The magnitude of the in-
creases in plasma IL-6 is related to exercise duration, training
intensity, and amount of muscle mass involved in the exercise
(16). In particular, the duration of the performed exercise is the
single most important factor for determining the levels of IL-6
release. Training intensity also plays a major role, especially if
the performed exercise leads to depletion of intramuscular
glycogen and energy stores. Accordingly, studies have shown
that muscular glycogen depletion markedly augments IL-6
release from the working muscles (26).
In line with this intrinsic energy dependency, IL-6 release
has been suggested to correlate with lactate production in
working muscles (9). Exercise-mediated IL-6 release stimulates
lipolysis and release of free fatty acids from adipose tissue, as well
as glycogenolysis in the liver, resulting in discharge of glucose [as
extensively reviewed by Pedersen and Febbraio (14)]. Thus, IL-6
can be regarded as a muscular energy sensor, which is secreted
due to energy depletion in the muscles during exercise, leading
to systemic energy substrate mobilization in energy-storing
tissues.
Traditionally, secretion through constitutive or activated
exocytosis has been a hallmark for release of soluble factors
(20). In addition, the extracellular matrix may function as a
reservoir of growth factors (6), and shedding of matrix- or
membrane-bound molecules by activated proteases is also a po-
tential source of soluble factors (2, 3). However, it is currently not
fully understood how IL-6 is secreted from contracting muscles
during exercise. Endogenous IL-6 has been shown to be localized
in vesicle-like structures in resting muscle fibers, and muscle
contraction resulted in a reduction of these IL-6-containing
vesicles (7). In parallel, exercise has been shown to be asso-
http://www.ajpendo.org
 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES
ciated with activation of intramuscular proteases, including the
two matrix metalloproteinases MMP2 and MMP9, as evident
from spillover of MMP2 and MMP9 to the blood circulation,
increased intramuscular mRNA and protein expression levels,
and proteolytic cleavage activity following exercise (8, 18, 19).
Yet, whether these intramuscular proteases play a role in IL-6
secretion during exercise has to our knowledge not been
investigated.
In this study, we aimed to elucidate the mechanism involved
in exercise-mediated release of IL-6 from muscles. We hypoth-
esized that lactate production during high-intensity exercise
initiates the activation of pH-dependent proteases, which may
explain the association between myokine release and muscular
metabolism during exercise.
METHODS
Human Exercise Intervention Study
The study was approved by the Ethics Committee of the Capital
Region of Denmark (H-8-2014-019) and by the local safety board. All
subjects gave written consent after oral and written information was
given. Thirteen healthy male subjects were included in the study, and
their baseline characteristics are presented in Table 1. Prior to the
experiments, the subjects were screened, which included a DEXA
scan and V̇O2max test. The exercise was performed on an electrody-
namically loaded Monark LC-4 cycle ergometer and after overnight
fasting. The subjects performed one session of 120 min of cycling,
involving a high-intensity interval-based biking, comprised of biking
for 10 min at 50% of V̇O2max, followed by 45 s at 120% of V̇O2max,
15-s break (0% V̇O2max), and 3 min at 30% of V̇O2max. Such intervals
were repeated until the subject had cycled for a total of 120 min. If the
subject could not perform the 45-s interval at the given workload,
workload was decreased to 100% of V̇O2max. Catheters were inserted
in the antecubital vein, and blood samples were drawn following 10,
30, 60, and 120 min of cycling. Rest samples were drawn 15 min
before cycling and 120 min postexercise (240 min).
Animal Studies
All animal experiments were conducted in accordance with the
ARRIVE guidelines and after permission from the Danish Animal
Inspectorate. Fully grown female NMRI mice (⫹12 wk) were bred
in-house at the Animal Facility at the University Hospital Herlev,
except for the lactate injection study (see Fig. 2I), where the NMRI
mice were purchased directly from Taconic. Breeding pairs were
obtained from Taconic, and these were replaced after four litters. Prior
to the exercise bouts, the mice were fasted overnight during the dark
Table 1. Baseline characteristics of 13 health young men
participating in acute-exercise study
Baseline characteristics (n ⫽ 13) Mean (Range)
Age, yr 26 (21–33)
Height, cm 185.6 (178–192)
Weight, kg 80.8 (71.9–93.3)
BMI, kg/m2 23.5 (19.5–27.4)
V̇O2max, l/min 4.48 (3.81–5.56)
Systolic blood pressure, mmHg 128 (115–142)
Diastolic blood pressure, mmHg 77 (65–96)
Resting heart rate, beats/min 59 (49–71)
Hemoglobin, mmol/l 9.7 (8.7–10.8)
Fat mass, kg 12.4 (4.8–18.7)
Fat mass, % 15.9 (7.1–24.5)
Lean body mass, kg 64.6 (57.7–73–5)
V̇O2max, maximal oxygen consumption.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
E941
period. In the morning, the mice performed a 1 h acute bout of
swimming in 35°C water. Blood samples were collected, and muscles
[tibialis anterior (TA)] were excised and immediately snap-frozen in
isopentane-cooled liquid nitrogen and stored at ⫺80°C. For drug
treatments, the following reagents were used: for intramuscular injec-
tion, 20 ␮l of hyaluronidase at 0.16 U/mg muscle (H3506, Type I-S, 400
U/mg, Sigma-Aldrich) (22) or 20 ␮l of lactate at 50 nmol/mg muscle
(L7022, Sigma-Aldrich); or for intraperitoneal injections, Marimastat at 5
mg/kg (M2699, Sigma-Aldrich) (24). Bicarbonate (Sigma-Aldrich) was
provided in the drinking water at a concentration of 200 mmol/l from
2 wk before the acute exercise bout.
For in vivo bioimaging, anesthetized mice were placed in a custom-
made bed, and in vivo scanning was performed using the Optix MX-2
Optical Molecular Image System (Advanced Research Technologies,
Montreal, QC, Canada), which uses time domain optical imaging.
Excitation was performed with a 635 nm (LDH-P-635) pulsing laser,
and emission was detected with a 650 long-pass filter. This scanning
was performed after the mice had been injected with 20 ␮l of
ProSense 750 FAST (40 nmol/ml, PerkinElmer, NEV11171) in the
TA muscle for detection of protease activity and subjected to 1 h of
swimming.
In Vitro Cell Culture Model
Skeletal muscle biopsies from the vastus lateralis of the quadriceps
femoris muscle were obtained under local anesthesia with 2% Lidocaine
(SAD) by the Bergstrom needle biopsy method from healthy young
premenopausal women. Isolation of satellite cells and differentiation
into mature myotubes were performed as previously described (21).
Fully differentiated myotubes were incubated with hyaluronidase
(0.5714 mg/ml, H3506, Type I-S, 400 U/mg, Sigma-Aldrich) for 48 h
or electrical pulse stimulation (EPS). Myotubes were stimulated by
EPS at 11 mV, 1 Hz, 2 ms (C-Pace EP, Ion Optix) for 6 and 24 h in
DMEM containing 1% horse serum (HS) and 1% penicillin strepto-
mycin (P/S). Nonstimulated (CON) myotubes served as controls.
Media were changed immediately before stimulation and collected
and centrifuged before being frozen at ⫺80°C.
Molecular Analyses
Determination of IL-6, MMPs, and other cytokines. Concentrations
of lactate and glucose in the plasma samples were measure by ABL
800 flex blood gas analyzer (Radiometer). IL-6 concentrations in
human plasma, murine serum, or murine muscle protein lysates were
measured by Meso Scale Discovery (MSD) ultrasensitive mouse IL-6
kit or MSD human single-plex IL-6. Serum C-X-C motif ligand 1
(CXCL1), IL-10, IL-1␤, and tumore necrosis factor-␣ (TNF␣) were
determined by MSD mouse 7-plex. MMP2 and MMP9 were deter-
mined by Quantikine ELISA for MMP2 and MMP9, respectively,
(Biotechne, RD Systems). All kits were analyzed according to the
manufacturer’s instructions, and the concentration of the analyte was
measured relative to a standard curve of known concentrations of the
analytes. All samples were analyzed in duplicate with intrasample
coefficient of variation (CV) below 20%.
Immunohistochemistry. Ten-micrometer cryosections from control
or lactate- or hyaluronidase-treated TA muscles were fixed in ice-cold
acetone for 3 min and subsequently blocked in buffer (3% BSA in
PBS) before staining. Sections were stained with antibodies against
IL-6 at 1:500 (LS-C70904, LSBio) and ␣7-integrin at 1:100 (3C12,
AM20011AF-N, Acris Antibodies). Goat anti-rabbit Alexa Fluor 594
and goat anti-mouse Alexa Fluor 488 were used as secondary anti-
bodies at 1:500 in PBS (ThermoFisher). All sections were additionally
stained with the nuclear stain 4=,6-diamidino-2-phenylindole (DAPI,
ThermoFisher). Negative control sections were stained without IL-6
primary antibody but with the two secondary antibodies. All sections
were observed at room temperature using a Nikon ⫻10 Plan Apo,
⫻20 Plan Apo VC and a ⫻60 Plan APO water immersion mounted on
a Nikon Ti-E epifluorescence microscope with motorized X/Y table
E942 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES

(Nikon Instruments). Images were acquired with a 5-Mpixel Andor
Neo camera for fluorescence imaging using NIS-Elements Advanced
Research (AR) software (Nikon Instruments) and merged in software.
Scale bar in figure micrographs is 50 ␮m.
RNA isolation and PCR analyses. Prior to RNA isolation, muscles
were homogenized at ⫺80°C using a mortar and then lysed in a
Qiagen Tissuelyser Retsch MM300 at 25 Hz suspended in TRIzol.
RNA was isolated by chloroform-isopropyl alcohol extraction.
RNA concentrations were determined using the NanoDrop 1000
spectrophotometer (Thermo Scientific). cDNA preparations were
synthesized from 250 ng of RNA, using a High Capacity cDNA
Transcription Kit (Applied Biosystems), and the resulting cDNA
was used for qPCR detecting IL-6 expression by Power Up SYBR
Green PCR Master Mix (Life Technologies), 7.5 ng cDNA, 300
nM forward primer, and 300 nM reverse primer in a total reaction
volume of 10 ␮l per well in MicroAmp Optical 384-well reaction
plates (Life Technologies).
Statistics
The statistical significance of the difference in variables over
time was determined by one-way ANOVA with repeated measure-
ments for the human exercise intervention study or without re-
peated measurements for the mouse studies. For multiple compar-
isons of changes over time or exercise and different treatments,
statistical analysis was performed using two-way ANOVA fol-
lowed by Bonferroni post hoc tests. Linear regression analyses
were used to investigate the association between lactate and IL-6
concentrations in both the human exercise intervention study and
the cell culture model. Last, Student’s t-tests were used to deter-
mined changes in MMPs after a swimming intervention. Data
analysis was performed using the GraphPad Prism version 8.0
software. Results are expressed as means ⫾ SE. The criterion for
significance is a probability of ⬍0.05.
RESULTS
Lactate Production During High-Intensity Exercise
Correlates with IL-6 Secretion in Humans
To explore the role of lactate production in IL-6 secretion,
we designed a high-intensity interval-based acute exercise
intervention, aiming to increase plasma and intramuscular
lactate levels. Thirteen young healthy men were recruited to
perform a 2-h interval-based cycling intervention, and had
blood samples collected at baseline, 30, 60 and 120 min
during the cycling, and 2 h into the recovery period (240
min). The exercise protocol involved intervals of 45 s at
120% of VO2max, 15 s breaks (0% of VO2max) and 3 min
at 30% of VO2max. The exercise was perceived as ex-
tremely strenuous on the Borg scale (Fig. 1a), and heart
rates increased markedly (Fig. 1b). Plasma lactate levels
increased 8-fold, peaking 1 h into the cycling exercise (P ⬍
0.001, Fig. 1c). Correspondingly, plasma IL-6 levels in-
creased 5.1-fold from 0.39 ⫾ 0.32 pg/ml to 2.00 ⫾ 1.37
pg/ml (P ⬍ 0.001) and peaked at exercise cessation (120
min) (Fig. 1c). To study a possible association between
lactate production and IL-6 release, correlational analyses
were performed, showing that plasma lactate concentration
at 60 min tightly correlated with plasma IL-6 concentration
at 120 min (R2 ⫽ 0.70, P ⬍ 0.001, Fig. 1d).

Fig. 1. Lactate production is associated with IL-6 release. Thirteen healthy young men performed a 2-h high-intensity interval-based cycling intervention (gray
area), during which rate of perceived exhaustion (Borg scale, n ⫽ 13; A), heart rate (n ⫽ 13; B), and plasma lactate and IL-6 (C) were determined (n ⫽ 13).
D: plasma lactate at 60 min was correlated with plasma IL-6 levels at 120 min. Primary human muscle cells (n ⫽ 10 individual donors) were stimulated with
electrical pulse stimulation (EPS), and IL-6 fold increase in the media in response to 24-h stimulation (n ⫽ 10 for each time point; E), lactate concentration (F),
and glucose concentration (G) were determined (n ⫽ 10 for each time point. H: media lactate concentration was correlated with media IL-6 levels after 24 h
of EPS. Statistical significance was determined by 1-way ANOVA (A–C, F and G); linear regression analysis was performed to determine the correlation between
lactate and IL-6 levels (D and H). Underlined stars indicate differences of the ANOVA analyses. *P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org

Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES
Lactate Production Correlates with IL-6 Release in
Electrically Stimulated Muscle Cells
In muscle cell culture studies, electrical pulse stimulation
(EPS) is frequently used as exercise mimic, thus we ex-
plored the effect of this stimulation on lactate production
and IL-6 secretion in vitro. EPS stimulation for 6 and 24 h
increased IL-6 concentration in the supernatant by 1.9-fold
and 2.1-fold, respectively (P ⬍ 0.05, Fig. 1e), and as
expected, EPS augmented lactate production and glucose
consumption (Fig. 1f, g). The EPS-stimulated IL-6 release
correlated with the accumulations of lactate in the media
(R2 ⫽ 0.41, P ⬍ 0.001, Fig. 1h).
E943
Lactate Stimulation Is Associated with IL-6 Secretion
To further investigate a possible mechanistic role of lactate,
we injected lactate directly into the quadriceps, gastrocnemius,
and TA muscles of both legs in resting mice and found that
serum IL-6 increased from 2.2 ⫾ 2.4 to 59.6 ⫾ 35.1 pg/ml
across 105 min from lactate injection (P ⬍ 0.01; Fig. 2A). In
addition to the release of IL-6 to the circulation, lactate injec-
tion was followed by a 5.4-fold increase in mRNA expression
of IL-6 within the injected muscles (P ⬍ 0.01; Fig. 2B). During
exercise, muscles produce lactate, and we aimed to prevent the
lactate-associated increase in intramuscular pH during exercise
by pretreating the mice for 2 wk with bicarbonate. In this study,
swimming exercise increased serum IL-6 only from 27.9 ⫾

Fig. 2. Lactate and exercise stimulates protease activity and IL-6 release. Female NMRI mice were injected with lactate (50 nmol/mg muscle im), and blood and
muscles were collected at baseline (0), and 5, 30, 60, and 105 min after injection for determination of serum IL-6 (n ⫽ 5 per time point; A) and muscular IL-6
mRNA expression (n ⫽ 5 per time point; B). C: IL-6 release in mice subjected to 1 h of swimming after 2 wk of bicarbonate treatment (n ⫽ 4 –7). D: IL-6 protein
content in protein lysates from tibialis anterior (TA) or soleus muscles (n ⫽ 5– 6). E: representative pictures of in vivo bioimaging after intramuscular injection
of the ProSense probe and 1 h of swimming. F: peak fluorescent intensity of ProSense fluorescence of pictures in E (n ⫽ 4 – 6). Serum concentrations of matrix
metalloproteinse-2 (MMP2; G) and MMP9 (H) in mice randomized to no intervention (Con, n ⫽ 12) or 1 h of swimming (SWIM, n ⫽ 12). I: serum IL-6 levels
in mice after 1 h of swimming and randomized to Marimastat or saline pretreatment (n ⫽ 7– 8 per group). Statistical significance was determined by 1-way
ANOVA (A–C), 2-way ANOVA with Bonferroni post hoc tests (D, I, J), and Student’s t-test (G and H). Underlined stars indicate differences of the ANOVA
analyses. *P ⬍ 0.05, **P ⬍ 0.01.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org

Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
E944 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES
10.8 to 36.0 ⫾ 11.2 (P ⫽ 0.16), whereas bicarbonate pretreat-
ment attenuated the minor swimming-induced increase in serum
IL-6 (Fig. 2C). In addition, we compared IL-6 protein content
within untreated TA muscles, which are primarily comprised of
type II fibers, and soleus muscles, which are primarily comprised
of type I fibers. Untreated TA muscles contained 3.0 ⫾ 0.6
pg/mg IL-6 compared with 7.7 ⫾ 2.4 pg/mg IL-6 in soleus
muscles (P ⬍ 0.01, n ⫽ 5– 6).
Exercise-Dependent IL-6 Release Is Associated with
Activation of MMP2 and -9
Proteases are sensitive to physical and chemical stimuli,
including pH levels, providing a possible link between exercise
intensity, lactate production, and IL-6 secretion. Therefore, we
investigated whether endogenous proteases were involved in
the exercise response. First, we evaluated whether the intra-
muscular proteases were activated by exercise. To this end, we
injected the fluorophore ProSense, which become fluorescent
upon cleavage by proteases, into the TA muscle and subjected
the mice to 1 h of swimming (Fig. 2E). One hour of swimming
changed fluorescent peak intensity from 666.5 ⫾ 69.1 (mean ⫾
SD, n ⫽ 4) to 726.9 ⫾ 58.2 (n ⫽ 6, P ⫽ 0.17; Fig. 2F). MMP2
and MMP9 are metalloproteinases that are highly expressed in
muscles, and we next investigated whether these two proteases
might spill over to the circulation during exercise (Fig. 2G).
One hour of swimming increased the release of MMP9 to the
circulation (P ⬍ 0.01; Fig. 2G), whereas no change in systemic
MMP2 concentration was observed at this early time point
(Fig. 2G). Next, we administered the MMP2/9-specific blocker
Marimastat during 1 h of swimming. This treatment com-
pletely blunted IL-6 release to the circulation during swimming
(Fig. 2H), further highlighting the importance of MMP activa-
tion in IL-6 release during exercise. Finally, we returned to the
in vitro model of EPS-stimulated release of IL-6 in muscle
cells. Here, we added the protease inhibitor pepstatin during
EPS to explore whether IL-6 release in vitro also depended on
protease activity. In line with the results in Fig. 1E, EPS
increased IL-6 secretion 2.0-fold (P ⬍ 0.05), whereas addition
of pepstatin attenuated this induction by 28% (Fig. 2J).
Intramuscular Location of IL-6
At the immunohistochemical level, we found that IL-6 was
expressed in individual fibers in clusters near the sarcolemma
in untreated muscles (Fig. 3A). Costaining with ␣7-integrin, a
marker for the sarcolemma, showed that these IL-6 clusters
were mainly located within the muscle fibers but close to
external capillaries (Fig. 3A). After hyaluronidase, IL-6 stain-
ing was mainly observed in the capillaries and extracellular
space, whereas the clusters observed within the muscle fibers
had disappeared (Fig. 3B). Of note, in a limited area of the
hyaluronidase-injected muscles (Fig. 3B, bottom right part of
the IL-6 staining), there was a complete absence of IL-6. This
area corresponded to the injection area. In lactate-injected
muscles, we also observed an increase in IL-6 staining in the
capillaries and extracellular space, but most noticeably we
observed an overall more intense staining of IL-6, which is in
line with our finding that intramuscular lactate injection en-
hanced IL-6 expression levels and thus IL-6 production.
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
Hyaluronidase Injection in Muscles Stimulate Release of
IL-6 and Other Myokines
Last, we explored whether intramuscular injection of the
protease hyaluronidase, which cleaves the extracellular matrix,
could mimic the exercise-mediated secretion of IL-6. At the
systemic level, hyaluronidase injection into the quadriceps,
gastrocnemius, and TA muscles of both legs in resting mice
resulted in dramatic increases in the systemic concentrations of
IL-6 (73-fold, P ⬍ 0.001), CXCL1 (5-fold, P ⬍ 0.001), and
IL-10 (5-fold, P ⬍ 0.001), all of which are recognized as
exercise-induced myokines (Fig. 4, A–C). In contrast, we
observed no changes in serum levels of TNF␣ and IL-1␤ (Fig.
4, D and E). Time course studies of hyaluronidase demon-
strated that the increase in serum IL-6 was time dependent,
peaking 2 h after injection after hyaluronidase injection (Fig. 4F).
Hyaluronidase injection also increased muscular IL-6 mRNA
expression, showing the same kinetics as IL-6 secretion, with
expression levels peaking 2 h after hyaluronidase injection
(Fig. 4G). In continuation, we found that the IL-6 increase in
the serum was dose dependent according to the number of
muscles injected with hyaluronidase (Fig. 4H). Finally, the
IL-6 release was independent of the age of the mice (Fig. 4I).
DISCUSSION
The idea of a circulating exercise factor, which can coordi-
nate energy homeostasis during exercise, has enthralled re-
searchers for more than half a century (17). In 2000, skeletal
muscles were formally demonstrated to secrete such exercise
factors, and these muscle-derived contraction-induced factors
were named myokines (25). Myokines have since provided a
conceptual platform for understanding how muscles commu-
nicate with other organs during exercise. IL-6 is the most
widely studied myokine, with numerous reports characterizing
its secretion from contracting muscles during exercise, as
reviewed in, e.g., Refs. 1 and 13). Whereas large efforts have
been put into identifying additional myokines (27), little atten-
tion has been given to the mechanisms by which myokines are
actually released from contracting muscles. On the basis of the
findings from this study, we propose a model of IL-6 release,
wherein exercise-mediated lactate production initiates pro-
tease-dependent release of IL-6 from the muscle. This model
links exercise intensity, lactate production, and IL-6 release
during strenuous exercise and can explain why IL-6 secretion
is augmented during high-intensity training or glycogen deple-
tion (25).
Il-6 has been shown to have numerous beneficial effects on
whole body metabolism and energy homeostasis (14) as well as
muscle-specific adaptation, including muscle hypertrophy and
regeneration and glucose metabolism (23). The exercise-medi-
ated IL-6 release is tightly correlated with exercise intensity
and duration; accordingly, IL-6 is reckoned an energy sensor
that is released from contracting muscles during strenuous
exercise concurrently with energy and glycogen depletion. In
line with a previous study (5), we here show that IL-6 release
during strenuous exercise is closely linked to local intramus-
cular lactate production; yet so far, no causal link has been
made between these two phenomena. Moreover, we have
demonstrated that lactate can directly stimulate IL-6 release
from murine muscles and that IL-6 release is blunted by
buffering the lactate-associated decrease in pH during exercise.
 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES
These data indicate that the exponential increase in IL-6 release
results from lactate production during the energy-exhausting
stage of strenuous exercise. Of note, with our design we cannot
determine whether it is the direct effect of lactate or the
acidifying role of lactate that is driving the observed response.
Exercise is known to be associated with intramuscular acti-
vation of MMP2 and MMP9 (4, 8, 18, 19). During an acute
bout of exercise, both the activity and the transcription of
MMP2 and MMP9 have been shown to increase, and activation
of matrix metalloproteinases have been measured as a spillover
into the bloodstream. Matrix metalloproteinases are, together
with other proteases, known to mediate ectodomain shedding
of cytokines, chemokines, growth factors, and their receptors,
which, for instance, occur as part of the regulation of the innate
and adaptive immune system (12). Our study was not designed
to explore the role of each individual protease in the exercise-
mediated IL-6 release. In contrast, our results indicate that
several or a cascade of proteases might be involved in this
response. First, intramuscular hyaluronidase injection leads to
a dramatic release of IL-6 to the circulation. This response to
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
E945
Fig. 3. Distribution of IL-6 after intramus-
cular hyaluronidase injection. A: cross-sec-
tional sections of tibialis anterior (TA) mus-
cles of 12-wk-old female NMRI mice were
stained with antibodies against ␣7-integrin
(green), IL-6 (red), and DAPI (blue) and
presented as individual pictures or in overlay
(merge). B: cross-sectional scan of TA mus-
cles, which were excised 2 h after intramus-
cular injection of hyaluronidase and stained
for (␣7-integrin (green), IL-6 (red), or DAPI
(blue) and overlay (merge). Darker area in
IL-6 staining corresponds to injection area of
hyaluronidase with no IL-6 present. C:
cross-sectional scan of TA muscles, which
were excised 105 min after intramuscular
injection of lactate and stained for (␣7-integ-
rin (green), IL-6 (red), or DAPI (blue) and
overlay (merge). Far right: IL-6 staining and
overlay at high magnitude. Bar, 50 ␮m.
an unspecific protease implies that any alteration of the extracel-
lular matrix can initiate the membrane conformational changes
needed to release the IL-6-containing intrafibrillar vesicles. Next,
and in line with previous studies, we observed that MMP9 was
activated and released from muscles during acute exercise. We
did not observe any increase in serum MMP2, which might be
explained by the fact that this matrix metalloproteinase is
activated later in the process, i.e., 3–12 days after muscle
injury, and thus much later than the 1- to 2-h time point, when
we collected the serum samples. Blocking MMP2 and MMP9
activity by Marimastat during exercise blunted the exercise-
mediated IL-6 release, suggesting that these matrix metallo-
proteinases are involved in the membrane conformational
changes, which are needed to release the IL-6-containing
intrafibrillar vesicles.
All matrix metalloproteinases are synthesized as latent zy-
mogenic propeptides and require cleavage by proteases in the
extracellular space for their activation. This step may be governed
by pH-sensitive proteases residing in the extracellular matrix of
muscles, providing a link between lactate production during
E946 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES

Fig. 4. IL-6 response to hyaluronidase injec-

tion and swimming. Female NMRI mice
were randomized to saline injection (CON,
n ⫽ 8) or hyaluronidase injection (HYA,
n ⫽ 8) and/or 1 h of swimming (SWIM, n ⫽
8). Immediately after the hour of swimming,
blood samples were collected, and serum
IL-6 (A), serum C-X-C motif ligand 1
(CXCL-1; B), serum IL-10 (C), serum IL-1␤
(D), and serum tumor necrosis factor-␣
(TNF␣; E) were determined. Female NMRI
mice were injected with hyaluronidase, and
blood and muscles were collected at baseline
(0), and 5, 30, 60, 120, and 240 min after
injection for determination of serum IL-6 (F)
and muscular IL-6 expression (G) (n ⫽ 5 per
time point). H: female NMRI mice were
injected with hyaluronidase in one tibialis
anterior muscle (1TA), both tibialis anterior
muscles (2TA), or both TA and both quad-
rips muscles (2Ta⫹2QUAD), and serum
was collected at 120 min for determination
of IL-6 levels (n ⫽ 5). I: 3-mo-old (Young)
and 8-mo-old (Old) mice were randomized
to saline injection (CON), 1 h of swimming
(SWIM), or hyaluronidase injection (HYA),
and blood was collected immediately after
swimming for determination of serum IL-6
(n ⫽ 6 –7). Statistical significance was de-
termined by 2-way ANOVA with Bonfer-
roni post hoc tests (A–E, I), or 1-way
ANOVA (F–H). Underlined stars indicate
differences between CON and HYA; indi-
vidual stars indicate differences between
REST and SWIM. **P ⬍ 0.01, ***P ⬍
0.001.

exercise and activation of MMP2 and MMP9. We have several
results that support this proposed line of events. For instance,
we found that intramuscular lactate injection increased MMP9
release into the circulation. Moreover, preventing the intramus-
cular acidification caused by lactate production during exercise
with bicarbonate pretreatment blunted IL-6 release during 1 h
of swimming. These results indicate that a cascade of proteases
is involved in cleavage of the extracellular matrix and thereby
remodeling of the sarcolemma, which leads to the release of
IL-6 from the intracellular stores.
Our histological analyses of muscle cross-sections demon-
strated that preformed vesicles containing IL-6 were located in
clusters residing close to capillaries. Very short-term exercise
as exemplified by 6-min “all-out” ergometer rowing has been
shown to increase plasma IL-6 threefold (11), supporting our
finding that a preformed depot of IL-6 in muscle cells can be
released into the circulation in relation to exercise of high
intensity without prior IL-6 transcription. In accord, Lauritzen
et al. (7) imaged individual muscle fibers and detected signif-
icant amounts of intracellular IL-6-positive vesicles under
resting conditions at both the sarcolemma and T-tubule re-
gions. Moreover, in this study we compared IL-6 content in the
predominantly fiber type II-positive TA muscles with the
predominantly fiber type I-positive soleus muscles and found

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org

Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
 EXERCISE-INDUCED IL-6 RELEASE DEPENDS ON LACTATE AND PROTEASES
that the protein level of IL-6 was 2.6-fold higher in the soleus
muscles. Our histological analyses were performed in female
mice, and although IL-6 has been shown to have differential
effects on whole body metabolism in mice (10), we do not know
of any evidence that the exercise-induced myokine response of
IL-6 should be affected by sex.
In conclusion, we propose a model of IL-6 release that involves
a lactate-dependent induction of proteases facilitating release of
intramyocellular IL-6-containing vesicles. This model links exer-
cise intensity, lactate production, and IL-6 release during strenu-
ous exercise and can explain why IL-6 secretion is augmented
during high-intensity training or glycogen depletion.
ACKNOWLEDGMENTS
Anne Boye and Lone Christensen are acknowledged for their technical
assistance.
GRANTS
This work was supported by grants from the Danish Cancer Society, the
Lundbeck Foundation and Copenhagen University Research funds. The Centre
for Physical Activity Research (CFAS) is supported by a grant from TrygFonden.
During the study period, the Centre of Inflammation and Metabolism (CIM) was
supported by a grant from the Danish National Research Foundation (DNRF55).
CIM/CFAS is a member of DD2, the Danish Center for Strategic Research in Type
2 Diabetes (the Danish Council for Strategic Research, grant no. 09-067009
and 09-075724).
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
P.H. and B.K.P. conceived and designed research; P.H., C.B., N.N.-C.,
C.D., B.L., C.K.O., and M.M.Å. performed experiments; P.H., C.B., N.N.-C.,
C.D., B.L., C.K.O., and M.M.Å. analyzed data; P.H., C.D., T.O.K., J.G., and
B.K.P. interpreted results of experiments; P.H. and C.B. prepared figures; P.H.
drafted manuscript; P.H., T.O.K., J.G., and B.K.P. edited and revised manu-
script; P.H., C.B., N.N.-C., C.D., B.L., C.K.O., M.M.Å., T.O.K., J.G., and
B.K.P. approved final version of manuscript.
REFERENCES
1. Benatti FB, Pedersen BK. Exercise as an anti-inflammatory therapy for
rheumatic diseases-myokine regulation. Nat Rev Rheumatol 11: 86 –97,
2015. doi:10.1038/nrrheum.2014.193.
2. Buzas EI, György B, Nagy G, Falus A, Gay S. Emerging role of
extracellular vesicles in inflammatory diseases. Nat Rev Rheumatol 10:
356 –364, 2014. doi:10.1038/nrrheum.2014.19.
3. Chen MK, Hung MC. Proteolytic cleavage, trafficking, and functions of
nuclear receptor tyrosine kinases. FEBS J 282: 3693–3721, 2015. doi:10.
1111/febs.13342.
4. Deus AP, Bassi D, Simões RP, Oliveira CR, Baldissera V, de Cássia
Marqueti R, Araujo HSS, Arena R, Borghi-Silva A. MMP(-2) expres-
sion in skeletal muscle after strength training. Int J Sports Med 33:
137–141, 2012. doi:10.1055/s-0031-1291224.
5. Helge JW, Klein DK, Andersen TM, van Hall G, Calbet J, Boushel R,
Saltin B. Interleukin-6 release is higher across arm than leg muscles
during whole-body exercise. Exp Physiol 96: 590 –598, 2011. doi:10.1113/
expphysiol.2010.056424.
6. Hynes RO. The extracellular matrix: not just pretty fibrils. Science 326:
1216 –1219, 2009. doi:10.1126/science.1176009.
7. Lauritzen HPMM, Brandauer J, Schjerling P, Koh H-J, Treebak JT,
Hirshman MF, Galbo H, Goodyear LJ. Contraction and AICAR stim-
ulate IL-6 vesicle depletion from skeletal muscle fibers in vivo. Diabetes
62: 3081–3092, 2013. doi:10.2337/db12-1261.
8. Mackey AL, Donnelly AE, Turpeenniemi-Hujanen T, Roper HP. Skeletal
muscle collagen content in humans after high-force eccentric contractions. J
AJP-Endocrinol Metab • doi:10.1152/ajpendo.00414.2018 • www.ajpendo.org
Downloaded from journals.physiology.org/journal/ajpendo (200.083.037.253) on July 15, 2020.
E947
Appl Physiol (1985) 97: 197–203, 2004. doi:10.1152/japplphysiol.01174.
2003.
9. Minetto MA, Rainoldi A, Gazzoni M, Ganzit GP, Saba L, Paccotti P.
Interleukin-6 response to isokinetic exercise in elite athletes: relationships
to adrenocortical function and to mechanical and myoelectric fatigue. Eur
J Appl Physiol 98: 373–382, 2006. doi:10.1007/s00421-006-0285-7.
10. Molinero A, Fernandez-Perez A, Mogas A, Giralt M, Comes G,
Fernandez-Gayol O, Vallejo M, Hidalgo J. Role of muscle IL-6 in
gender-specific metabolism in mice. PLoS One 12: e0173675, 2017.
doi:10.1371/journal.pone.0173675.
11. Nielsen HB, Secher NH, Christensen NJ, Pedersen BK. Lymphocytes
and NK cell activity during repeated bouts of maximal exercise. Am J
Physiol Regul Integr Comp Physiol 271: R222–R227, 1996. doi:10.1152/
ajpregu.1996.271.1.R222.
12. Parks WC, Wilson CL, López-Boado YS. Matrix metalloproteinases as
modulators of inflammation and innate immunity. Nat Rev Immunol 4:
617–629, 2004. doi:10.1038/nri1418.
13. Pedersen BK. Muscle as a secretory organ. Compr Physiol 3: 1337–1362,
2013. doi:10.1002/cphy.c120033.
14. Pedersen BK, Febbraio MA. Muscle as an endocrine organ: focus on
muscle-derived interleukin-6. Physiol Rev 88: 1379 –1406, 2008. doi:10.
1152/physrev.90100.2007.
15. Pedersen BK, Febbraio MA. Muscles, exercise and obesity: skeletal
muscle as a secretory organ. Nat Rev Endocrinol 8: 457–465, 2012. doi:10.
1038/nrendo.2012.49.
16. Pedersen BK, Fischer CP. Physiological roles of muscle-derived inter-
leukin-6 in response to exercise. Curr Opin Clin Nutr Metab Care 10:
265–271, 2007. doi:10.1097/MCO.0b013e3280ebb5b3.
17. Pedersen BK, Steensberg A, Fischer C, Keller C, Keller P, Plomgaard
P, Febbraio M, Saltin B. Searching for the exercise factor: is IL-6 a
candidate? J Muscle Res Cell Motil 24: 113–119, 2003. doi:10.1023/
A:1026070911202.
18. Rullman E, Norrbom J, Strömberg A, Wågsäter D, Rundqvist H,
Haas T, Gustafsson T. Endurance exercise activates matrix metallopro-
teinases in human skeletal muscle. J Appl Physiol (1985) 106: 804 –812,
2009. doi:10.1152/japplphysiol.90872.2008.
19. Rullman E, Rundqvist H, Wågsäter D, Fischer H, Eriksson P, Sund-
berg CJ, Jansson E, Gustafsson T. A single bout of exercise activates
matrix metalloproteinase in human skeletal muscle. J Appl Physiol (1985)
102: 2346 –2351, 2007. doi:10.1152/japplphysiol.00822.2006.
20. de Saint Basile G, Ménasché G, Fischer A. Molecular mechanisms of
biogenesis and exocytosis of cytotoxic granules. Nat Rev Immunol 10:
568 –579, 2010. doi:10.1038/nri2803.
21. Scheele C, Nielsen S, Kelly M, Broholm C, Nielsen AR, Taudorf S,
Pedersen M, Fischer CP, Pedersen BK. Satellite cells derived from
obese humans with type 2 diabetes and differentiated into myocytes in
vitro exhibit abnormal response to IL-6. PLoS One 7: e39657, 2012.
doi:10.1371/journal.pone.0039657.
22. Schertzer JD, Plant DR, Lynch GS. Optimizing plasmid-based gene
transfer for investigating skeletal muscle structure and function. Mol Ther
13: 795–803, 2006. doi:10.1016/j.ymthe.2005.09.019.
23. Serrano AL, Baeza-Raja B, Perdiguero E, Jardí M, Muñoz-Cánoves
P. Interleukin-6 is an essential regulator of satellite cell-mediated skeletal
muscle hypertrophy. Cell Metab 7: 33–44, 2008. doi:10.1016/j.cmet.2007.
11.011.
24. Skipper JB, McNally LR, Rosenthal EL, Wang W, Buchsbaum DJ. In
vivo efficacy of marimastat and chemoradiation in head and neck cancer
xenografts. ORL J Otorhinolaryngol Relat Spec 71: 1–5, 2009. doi:10.
1159/000163217.
25. Steensberg A, van Hall G, Osada T, Sacchetti M, Saltin B, Pedersen
BK. Production of interleukin-6 in contracting human skeletal muscles can
account for the exercise-induced increase in plasma interleukin-6. J
Physiol 529: 237–242, 2000. doi:10.1111/j.1469-7793.2000.00237.x.
26. Steensberg A, Febbraio MA, Osada T, Schjerling P, van Hall G, Saltin
B, Pedersen BK. Interleukin-6 production in contracting human skeletal
muscle is influenced by pre-exercise muscle glycogen content. J Physiol
537: 633–639, 2001. doi:10.1111/j.1469-7793.2001.00633.x.
27. Whitham M, Febbraio MA. The ever-expanding myokinome: discovery
challenges and therapeutic implications. Nat Rev Drug Discov 15: 719 –
729, 2016. doi:10.1038/nrd.2016.153.