about 25% of cancer cases globally are due to excess weight and sedentary lifestyle (15). Contrarily, regular physical activity has been demonstrated to protect against various cancers. In 2007, the World Cancer Research Fund systematically reviewed the more than 50 studies regarding a possible association between physical activity and cancer and concluded that physical activity reduces the average risk for developing breast and colon cancers by 25–30% (28). These epidemiological findings led the World Cancer Research Fund to propose general recommendations of 30 min of daily moderate exercise. The anti-cancerous effects of physical activity are not yet entirely understood, but the protective effects are believed to depend on the intensity, type, frequency, and duration of the activities (5, 27). The list of proposed mechanisms is long and includes changes in endogenous hormones, immune function, DNA repair, low-grade inflammation, and energy balance (8, 15, 17).

Research during the past 10 years has demonstrated that exercise induces the release of cytokines from working muscles (26). These muscle-derived cytokines are collectively named myokines, and the first myokine to be discovered was IL-6 (20, 24). However, recent research has demonstrated that skeletal muscles may produce and secrete numerous myokines belonging to distinctly different families (22). This highlights the fact that muscle tissue has an inherent ability to work as a secretory organ. The myokines have been studied extensively for their role in metabolism, e.g., as mediators of muscle-to-fat cross-talk and in controlling muscular glucose and fat metabolism (9, 14, 19, 22).

However, myokines may also play a role in cancer protection. Recently, it was shown that conditioned media from electrostimulated C₂C₁₂ muscle cells could inhibit mammary cancer cell proliferation (3), suggesting that the contracting myotubes release humoral factors that can inhibit cancer cell growth. In continuation, Barnard et al. (2) have shown that serum obtained from subjects after 2 wk of a low-fat, high-fiber diet and exercise intervention could inhibit mammary cancer cell growth in vitro. Similar interventions in men were shown to inhibit prostate cancer cell proliferation (25).

The establishment of cancer is dependent on several factors, including immune reactivity toward neoplastic cells and growth factors. In this study, we hypothesize that factors secreted from muscles during physical activity have a direct effect on the cancer cells by inhibiting their proliferation, and thus physical activity, in addition to the effects on reduced adiposity, estrogen, and glucose levels, mediates independent protective effect against cancer through the release of antiproliferative myokines from working muscles.

MATERIALS AND METHODS

Cells and reagents.

The human mammary cancer cell line MCF-7 (estrogen receptor positive) was grown at 37°C in a humidified atmosphere of 5% CO₂ in Dulbecco's modified Eagle's medium containing (DMEM) 4.5 mg/l d-glucose and supplemented with 10% fetal bovine serum (FBS; Invitrogen) or 10% postexercise serum and 1% penicillin-streptavidine (Invitrogen). For incubation studies, the cytokines mouse recombinant GDF5 and human recombinant oncostatin M (OSM) (both R & D Systems) and human recombinant IL-10 (ImmunoTools) were used at 10 ng/ml. Mouse recombinant GDF5 was used since no human recombinant protein was available.

Mouse exercise trial and PCR arrays.

All animal experiments were conducted in accordance with the recommendations of the European Convention for the Protection of Vertebrate Animals Used for Experimentation and after permission from the Danish Animal Experiments Inspectorate. Ten-week-old Naval Medical Research Institute mice (own breed) swam in 37°C warm water for 60 min and were euthanized immediately after, 2 h, or 5 h after the swimming bout. Control mice did not swim but were otherwise treated similarly to the exercising mice. For instance, food was withdrawn 1 h before euthanasia. Blood was collected by decapitation, and the tibialis cranialis, gastronemius, and soleus muscles, liver, spleen, and subcutaneous and visceral adipose tissues were quickly excised and frozen on dry ice and 100% alcohol. Sample size is 8 for serum and all tissue samples.

Total RNA was isolated from frozen tissues by tissue homogenization and RNA extraction using the TRIzol reagent (Invitrogen Life Technologies), and cDNA synthesis was performed using the RT2 PCR Array First Strand kit (SABiosciences). Custom-made PCR arrays (CAPM-0838E; SABiosciences) containing predeveloped reactions for 28 customized genes and four housekeeping genes were performed according to the manufacturer's instructions. N was 10 for each condition (pre-exercise, 0H, and 2H). Data analysis was performed using the web-based analysis tool (www.sabiosciences.com/pcrarraydataanalysis.php), which includes descriptive statistics, including fold change, volcano plots, and cluster analyses. For OSM expression, RT-PCR was performed using random p(dN)₆ primers (Applied Biosystems) and MultiScripe Reverse Transcriptase (High-Capacity cDNA Reverse Transcription Kit; Applied Biosystems). All amplifications were done using the Taqman SYBR green PCR Master Mix (Applied Biosystems) and sequence-specific primers.

Electrical stimulation of C₂C₁₂ muscle cells.

Murine C₂C₁₂ muscle cells were grown in triplets in six-well plates at a starting concentration of 2 × 10⁴ cells/well. When plates were confluent, differentiation was initiated by switching to differentiation medium 1 (DMEM, glutamax, low glucose, 10% FBS) followed by differentiation medium 2 (DMEM, high glucose, 2% HS). When cells were fully differentiated, electrical stimulation (2-ms duration, 40.0 V, 1 Hz) was performed on half of the wells, whereas the other half was nonstimulated controls. Media were changed before the stimulation and collected after 3 h of stimulation/nonstimulation and frozen at −20°C.

Cell culture assays.

Serum from the exercising mice was pooled within the groups, followed by heat inactivation at 56°C for 30 min and sterile filtration. Cells were grown in triplets in six-well plates at a starting concentration of 100,000 cells/well and with 2.5 ml of medium containing 10% conditioned serum instead of FBS or 10 ng/ml recombinant protein.

A Caspase/CPP32 Colorimetric Assay kit (Biovision) was used to assay the activity of caspase. After incubation (24 h for postexercise serum experiment and 96 h for recombinant proteins) the cells were counted, and 1,000,000 cells from each well were spun down. The supernatant was removed, and cells were resuspended in 50 μl of chilled cell lysis buffer (Biovision) and incubated on ice for 10 min. Following centrifugation, the cytosolic extract was transferred to a fresh tube and put on ice. Fifty micrograms of protein was diluted to 50 μl of cell lysis buffer, followed by the addition of 50 μl of 2× reaction buffer (containing 10 mM DTT) and 5 μl of 4 mM DEVD-pNA substrate (200 μM final concentration). The plate was incubated at 37°C for 1 h, and absorbance was read at a microtiter plate reader at 405 nm.

An Annexin V-Biotin Apoptosis Detection kit (Biovision) was used to determine translocated phosphatidylserine. After incubation with the cytokines, 5 × 10⁵ cells were collected from each group and resuspended in 200 μl of binding buffer (kit), and 5 μl of Annexin V-biotin and 5 μl of propidium iodid were added, followed by 5 min of incubation in the dark. Following washing, the cells were fixed in 2% formaldehyde (Bie & Berntsen) in phosphate-buffered saline (PBS; Invitrogen) and washed again in PBS. The cells were resuspended in 100 μl of PBS with 1 mg/ml BSA and 5 μl streptavidine FITC (ElectraBox). After the final wash, the cells were resuspended in 200 μl of PBS and analyzed by flow cytometry.

Caspase activity in MCF-7 cells incubated with conditioned media.

MCF-7 cells were plated in six-well plates in at concentration 60,000 cells/well. The next day, the cells were added to medium (+10% FBS) from the electrical stimulated/non-stimulated C₂C₁₂ cells, and half of the wells were added to anti-OSM antibodies (R & D Systems). Additionally, control cells were grown in normal growth medium. All stimulations were performed three times in triplets. After 24 h, caspase activity was determined by the use of a caspase assay, as described earlier.

Serum OSM and metabolite measurements.

Serum was collected from control mice (n = 8) or immediately after 1 h of swimming (n = 10) and stored at −80°C until assayed. Blood glucose was determined by Hemocue glucose 201 measurements. Enzyme-linked immunosorbent assay kit (Uscnk Life Science) was used for the quantitative measurement of mouse OSM according to the manufacturer's protocol. Meso Scale Discovery 3-plex kit was for used for quantitative analyse of serum insulin, glucagon, and glucagon-like peptide-1 (GLP-1).

Statistics.

One-way ANOVA was used to compare the groups, and ad hoc posttesting by t-test with Bonferroni correction for multiple testing was used to determine differences among the individual groups. Two-way ANOVA was used to compare groups and time points in the swimming experiment. This analysis was also followed by ad hoc posttesting with Bonferroni correction to determine differences among the individual groups. A P value of <0.05 was considered to be statistically significant.

RESULTS

Exercise-conditioned serum inhibits mammary cancer cell proliferation and induces apoptosis.

Mice were euthanized before, immediately after, or 2 h into recovery after a 1-h swimming exercise bout, and serum was collected. Incubating MCF-7 cells with media containing 10% of this exercise-conditioned serum resulted in a 52% reduction in cell proliferation, with the serum collected immediately after swimming (P < 0.005; Fig. 1, top). Serum taken 2 h into recovery had no effect on the proliferation rate. Similarly, caspase activity was analyzed in the cells incubated with the exercise-conditioned serum.Serum taken immediately after swimming increased caspase activity by 54% compared with resting levels (P < 0.05), whereas serum taken 2 h into recovery had little effect on caspase activity (Fig. 1, botttom).

Identification of exercise-induced myokines.

To identify which potential myokines were expressed in the working muscles, we performed PCR array analysis on 28 selected genes. These were chosen on the basis of previous reports that had proposed them as potential myokines. We found that seven of the 28 selected genes were significantly upregulated, and none were downregulated immediately after swimming (Fig. 2). The upregulated genes included OSM, IL-6, IL-11, GDF2, GDF5, CSF2, and IL-10. Two hours into recovery, IL-6, IL-11, GDF2, and GDF5 remained significantly upregulated (Fig. 2).

Control of tumor cell growth by individual myokines.

We chose to investigate OSM, IL-10, IL-11, and GDF5 in detail. This selection was based on the availability of the candidates as recombinant proteins, and furthermore, we chose not to investigate IL-6 since numerous reports suggest that IL-6 does not possess an anti-cancer effect (16, 18). OSM and IL-10 increased immediately after swimming but returned to baseline level after 2 h, whereas IL-11 and GDF5 remained significantly elevated both immediately and 2 h into recovery.

Adding these factors to the cell media at a concentration of 10 ng/ml, we determined cell proliferation and induction of apoptosis as measured by caspase activity and annexin-streptavidin staining. Incubation of MCF-7 cells for 5 days showed growth inhibitory effects of OSM (P < 0.05) and a borderline decrease in growth by IL-10 (P = 0.07) but no effect of IL-11 and GDF5 on MCF-7 proliferation, and these two candidates were not investigated further (Fig. 3). Pooling of the cytokines also showed a significant reduction in cell number after 5 days of incubation (P < 0.001; data not shown). The apoptosis assays showed a significant upregulation (P = 0.034) in caspase activity in MCF-7 cells stimulated for 96 h by OSM compared with the control group (Fig. 4, top). Similarly, annexin-streptavidin staining showed a significant increase in the number of apoptotic cells after 96 h of incubation with OSM compared with the control group (Fig. 4, middle). No effect of IL-10 on caspase activity or annexin-streptavidin staining was observed (data not shown).

To further assess whether OSM was responsible for the induction of caspase activity in stimulated MCF-7 cells, we electrically stimulated murine C₂C₁₂ muscle cells to obtain a pure pool of muscle-derived myokines in response to an exercise stimuli. We then incubated the MCF-7 cells with the conditioned media from the muscle cells with or without anti-OSM antibodies. In line with the results from postexercise serum incubation studies, media from electrically stimulated muscle cells induced a 100% increase in caspase activity (P = 0.01), whereas the presence of anti-OSM antibodies reduced this increase to a 49% increase compared with incubation with conditioned media from nonstimluated myotubes (P = 0.1) (Fig. 4, bottom).

Is OSM a myokine?

Because OSM proved to be the most efficient cytokine, we measured the serum levels of OSM following the exercise intervention. At the control level, serum OSM was expressed very weakly and was detectable in only four of eight mice (Fig. 5). After 1 h of swimming, OSM increased significantly (P < 0.001) and was detectable in nine of nine mice. Two hours into recovery, serum OSM had returned to control levels. To verify the source of systemic OSM, we determined OSM expression in different tissues (Fig. 5). Following the swimming exercise bout, OSM expression was upregulated in all three muscles investigated, i.e., tibialis cranialis (4-fold, P < 0.001), gastronemius (2-fold, P < 0.05), and soleus (2-fold, P < 0.05). OSM expression did not change in spleen (mononuclear cells), liver, or subcutaneous and visceral adipose tissue following the exercise bout, strongly suggesting that muscle tissue is the primary source of OSM.

DISCUSSION

Our main finding is that exercise-conditioned serum is capable of inhibiting mammary cancer cell proliferation and inducing apoptosis through caspase activation. The direct use of exercise-conditioned serum in this study bypasses in vivo modulations like the immune system and hypoxia and emphasizes the role of systemic factors, which change during an exercise bout. Our results clearly suggest that muscle-derived humoral factors (myokines) can inhibit tumor cell growth and that at least some of the antiproliferative response of physical activity is due to a direct release of humoral factors to the circulation during exercise bouts.

Our approach was to incubate mammary cancer cells with conditioned serum from exercising mice. A similar approach was undertaken by Barnard et al. (2), who conducted studies with obese postmenopausal subjects and collected serum 24 h after the exercise intervention. In these studies, the exercise intervention was combined with a low-fat, high-fiber diet for 2 wk, and thus the effect of the metabolic profile in the serum caused by the diet or exercise intervention could not be differentiated. The intervention was associated with reductions in serum estradiol, insulin, and IGF-I and proved to significantly reduce MCF-7 proliferation (2). In our study, we found that acute changes in serum could inhibit mammary cancer cell proliferation. At this time point, the exercise-induced myokines are at the highest level. In addition, we observed reductions in serum insulin (Table 1), whereas serum estradiol was not investigated since these studies were performed in male mice.

To identify potential muscle-derived factors that could explain the reduction in cancer cell proliferation, we performed a partial expression profile screening using PCR arrays. In this, we included genes encoding 28 candidate myokines that represent known and potential myokine families, i.e., IL-6 superfamily, transforming growth factor-β family, NGFβ family, and other miscellaneous candidates, which have been suggested to play a role in muscle-to-fat cross-talk (21). In continuation of this selection, seven genes were upregulated acutely after 1 h of swimming, whereas none were downregulated. Several of the genes remained elevated 2 h after the termination of swimming, including IL-11, GDF-2, and GDF5, whereas others returned to baseline expression, i.e., OSM and IL-10.

OSM proved to be of particular interest because this cytokine significantly inhibited MCF-7 cell proliferation and induced apoptosis. Moreover, blockade of OSM signaling by addition of anti-OSM antibodies halved the induction of caspase activity. We also found that OSM increased significantly from nearly undetectable levels to significant serum concentrations immediately after the exercise bout and declined 2 h after the exercise bout. Of note, only serum obtained immediately after the exercise, but not 2 h after, had inhibitory effects on cancer cell growth. This serum increase was paralleled by an increase in the muscular expression of OSM in all three muscles investigated. This implies that muscle tissue is the primary source of OSM during exercise and that OSM can be considered as a myokine. Interestingly, the concentration of OSM goes from undetectable levels at resting conditions to measurable picogram per millilter levels immediately after exercise, suggesting that OSM works even at very low concentrations. Previous studies have shown a dose-dependent inhibition on MCF-7 cells with OSM (4, 12). Doses down to 400 pg/ml have been shown to have inhibitory effects in in vitro systems, whereas the main effect occurs above 2 ng/ml. Our serum OSM increases following exercise might not reach these levels, but we believe that OSM still plays a role in our system since addition of anti-OSM antibodies to conditioned media from electrically stimulated C₂C₁₂ cells can partly blunt the induction of caspase activity in the MCF-7 cells.

OSM is a pleiotropic cytokine belonging to the IL-6 superfamily, and studies suggest that OSM is expressed in 66% of breast tumors, with the prevalence being even higher in inflammatory breast cancers (6). The antiproliferative role of OSM has been documented in many in vitro studies, and in line with our results, OSM has been shown to inhibit proliferation of tumor cell lines derived from a variety of tissues, including ovarian, lung, human cerebral meningioma, mammary epithelial cells, and melanoma cells (7, 10, 13). However, following the initial interest in OSM for its antiproliferative effects, it has become increasingly apparent that the effects of OSM are also associated with breast cancer cell detachment and angiogenesis (11).

We also tested IL-10, IL-11, and GDF5 for their antiproliferative effects. Regarding IL-10, a large and contradicting literature exists proving that IL-10 can both inhibit and stimulate tumor cell growth. We found that IL-10 tended to reduce the proliferation of MCF-7 cells, whereas it had no effect on induction of apoptosis. Little evidence on the role of IL-11 and GDF5 on tumor proliferation exists, and we found no effect of IL-11 GDF5 on the proliferation of MCF-7.

In addition to the myokines that were induced by exercise, the swimming exercise bout also tended to decrease a number of proinflammatory cytokines and growth factors that have been shown to play a role in tumor progression, including TNFα, IGF-I, and leptin (1, 23). Although not significant, the reduction in muscular expression was seen to last at least 2 h into recovery. We measured TNFα, IL-6, and monocyte chemoattractant protein-1 in the serum samples, but none of these proinflammatory cytokines were detectable either before or after the swimming bout (data not shown). Furthermore, we measured insulin, glucagon, and GLP-1 (Table 1). We found that insulin decreased after the exercise bout and remained depressed 2 h into recovery, where the growth inhibitory effect of the exercise-conditioned media was reversed. If insulin should play a role, this does not explain why growth repression is not seen in MCF-7 cells incubated with serum taken 2 h into the recovery. Glucagon and GLP-1 did not change during the exercise or recovery period.

CONCLUSION

Physical activity is associated with significant reductions in the development of cancers. In this study, we showed that factors secreted from working muscles could inhibit cancer cell growth and induce apoptosis and that addition of anti-OSM antibodies halved this response, indicating that myokines might play a role in the protective effect of exercise on cancer.

GRANTS

The Centre of Inflammation and Metabolism is supported by a grant from the Danish National Research Foundation (no. 02-512-55). P. Hojman was supported by the Danish Medical Research Council, and the study was further supported by grants from the Augustinus Foundation, Fabrikant Einar Willumsens Foundation, and Bdr. Hartmanns Foundation.

DISCLOSURES

The authors have no competing interests to declare.