Voluntary Wheel Running Reduces the Acute Inflammatory Response to Liver Carcinogen in a Sex-specific Manner

Inflammation plays an important role in cancer development, affecting all processes from tumor initiation, promotion, and progression to metastasis, and more than 20% of all cancers arise as a consequence of persistent infections and chronic inflammation (1, 2). Yet, the inflammatory response is a complex system, and the acute inflammatory response is needed to eradicate nascent tumors and trigger the adaptive immune response to combat cancer (3, 4). In line with this, a high level of infiltration by cytotoxic immune cells in tumors is associated with a better prognosis for most cancer diagnoses (5–7). In contrast, chronic low-grade inflammation with continuous production of proinflammatory cytokines promotes cancer and impairs cytotoxic immune cells, thus further accelerating cancer progression (8, 9). Therefore, it is of major clinical importance to obtain in-depth understanding of the regulation of the inflammatory response.

Physical activity and exercise are associated with reduced risk of cancer, as well as lower risk of disease recurrence after primary cancer treatment (10, 11). Importantly, exercise is recognized for its immunomodulatory effects. With long-term training, basal levels of low-grade inflammation are reduced, as is inflammatory signaling in major metabolic tissues (12). In continuation, acute exercise results in the release of anti-inflammatory cytokines, that is, IL6, IL10, and IL1ra, to the circulation, thereby creating an anti-inflammatory environment (13). A number of studies have investigated how exercise modulates the response to acute and chronic inflammatory response. For instance, it has been demonstrated that LPS-induced TNFα production decreased after a session of exercise (14), and mechanistic studies furthermore suggest that exercise may have regulatory effects on important inflammatory signaling pathways, such as toll-like receptors (TLR) and NFκB signaling (15). To this end, IL6 plays an important role, linking the acute proinflammatory response to factors involved in the resolution of the inflammatory response.

In this study, we explore the effect of voluntary wheel running on the acute inflammatory response to the carcinogen diethylnitrosamine (DEN). DEN is a well-established experimental model for induction of hepatocarcinoma (HCC) initiation and proliferation (16, 17), yet the model displays a strongly sex-specific response in tumor development, as male mice develop multiple HCC lesions in an IL6-dependent manner (16, 18, 19). In contrast, female mice are essentially protected from DEN-induced HCC (16). A similar sex-specific pattern exists in the incidence of HCC in patients, where men are almost twice as likely as women to develop HCC according to the NORDCAN database (20). Yet, it should be noted that differences in lifestyle such as alcohol consumption, eating habits, and frequency of hepatitis may add to the biological difference between men and women. In addition to the sex dependency, we have recently shown that voluntary wheel running markedly reduced the incidence of DEN-induced HCC, as well as decreased tumor burden in the male mice (21). In light of this, we aim to investigate the effect of voluntary wheel running and sex on the acutely DEN-induced inflammatory response in mice.

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. The studies were compliant with the ARRIVE guidelines. Mice were placed in standard housing cages and maintained in a thermostated environment under a 12-hour light/dark cycle with free access to food (normal chow, 2,844 kcal/kg, 4% crude fat, Altromin pellets, Spezialfutter-Werke) and drinking water. Breeding pairs were obtained from Harlan.

Male or female NMRI mice of 20–24 weeks of age were randomized to individual cages fitted with light running wheels (12 cm in diameter) or no wheels. Running wheels were only installed in the cages of the exercising mice, but both the control and exercise groups had enrichment in form of nesting material, biting sticks and small toys in the cages. The mice were kept in these cages for 6 weeks as training intervention, running between 2–7 km per night. After the training period, all mice had a single injection of 100 mg/kg DEN (N-Nitroso-diethylamine, #N0258, Sigma-Aldrich), or saline as control. At 2, 24, 48, and 96 hours, groups of mice (10 at each time point) were euthanized, and blood, liver, spleen, and muscles were excised and snap frozen. During this 96-hour period, food intake and body weight were monitored daily.

At termination, blood glucose was measured using Precision X system (Abbott Diabetes Care), while the rest of the blood was processed to serum. Serum aspartate aminotransferase (ASAT) and alanine aminotransferase (ALAT) were determined by standard clinical biochemical measurement at the Department of Clinical Biochemistry, Herlev Hospital (Herlev, Denmark). Serum IL6, IL1β, IFNγ, TNFα, and IL10 were determined by 10-plex Mouse Pro-inflammatory panel ELISA (Meso Scale Discovery) according to the manufacturer's instructions. Detection limits were 0.90 pg/mL for IL6, 0.37 pg/mL for IL1β, 0.20 pg/mL for IFNγ, 0.13 pg/mL for TNFα, and 0.70 pg/mL for IL10. Interassay CV was 5% and intra assay CV were <15% for all five factors.

RNA was isolated from 10–20 mg of snap frozen liver tissue after homogenizing in the Qiagen Tissuelyser Retsch, and RNA extraction performed according to the TRIzol-chloroform protocol. RNA concentrations were determined using the NanoDrop 1000 spectrophotometer (Thermo Scientific). Preparations of cDNA were synthesized from 250 ng RNA using the High Capacity cDNA Transcription Kit (Applied Biosystems) in a total reaction volume of 20 μL. The reactions were submitted to the following PCR program: 25°C for 10 minutes, 37°C for 120 minutes, and 85°C for 5 minutes followed by cooling to 4°C. The resulting cDNA was then used for real-time PCR detecting the genes encoding the following immune cell–specific markers or cytokines: CD335/NKp46 (NK cells), CD68 (macrophages), CD209 (antigen-presenting dendritic cells), iNOS, IL6, IL1β, IL10, TNFα, and IFNγ. Primer sequences were designed using the Primer-BLAST tool from NCBI and ordered from TAG Copenhagen A/S, and the primer sets were optimized according to efficacy determined by a 2-fold dilution standard and no byproduct when performing postanalysis melting curves. Primer sequences and primer optimization are shown in Supplementary Table S1. All samples were run in triplicates using Power Up SYBR Green PCR Master Mix (Life Technologies), 7.5 ng cDNA, 300 nmol/L forward primer, and 300 nmol/L reverse primer in a total reaction volume of 10 μL per well in MicroAmp Optical 384-well reaction plates (Life Technologies). qPCR was performed using the ViiA 7 system (Thermo Scientific): 50°C for 2 minutes, 95°C for 2 minutes, and 42 cycles of 95°C for 15 seconds followed by 60°C for 1 minute. Gene expression levels are normalized to GAPDH expression, as this household gene showed no regulation by DEN and exercise interventions (data not shown).

Protein was isolated from approximately 10 mg of hepatic tissue by homogenizing in the Qiagen Tissuelyser Retsch in a lysis buffer [0.05 mol/L Tris-HCL, 0.15 mol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 0.05 mol/L NaF, 5 mmol/L NaP, 2 mmol/L NaV, 1 mmol/L DTT, 1xPP3, 1xPP2, 0.2% Ipegal-CA-630, cOmplete mini Protease Inhibitor Cocktail (Roche)]. Concentrations were determined using the NanoDrop 1000 spectrophotometer (Thermo Scientific). The samples were loaded on Criterion TGX 4%–20% gels and submitted to SDS-PAGE in TGS buffer at 125 V, 300 A, and 300 W for 60 minutes. Markers Precision Plus Protein Dual Color Standards and Precision Plus Protein All Blue Standards (Bio-Rad) were also loaded. The samples were blotted onto Trans-Blot Turbo Midi PVDF membranes using the Trans-Blot Turbo Transfer System. The membranes were then blocked in TBST buffer with 5% milk for 1 hour at room temperature, incubated with primary antibody overnight at 4°C, and subsequently incubated with secondary antibody for 2 hours at room temperature, before staining with Femto Stable Peroxide Solution (Thermo Fisher Scientific) mixed 1:1 with Femto Luminol Enhancer Solution (Thermo Fisher Scientific) and detection of the bands using the ChemiDoc XRS system from Bio-Rad. All washing steps and antibody dilutions were made in TBST. The following antibodies from Cell Signaling Technology were used for detection: NFκB p65 (C22B4) rabbit mAb #4764 1:1,000 in 5% BSA and phospho-NFκB p65 (Ser536; 93H1) rabbit mAb #3033 1:1,000 in 5% BSA. Secondary antibody was purchased from DAKO: anti-rabbit-HRP #P0448 1:5000 in 5% milk. All detected bands were normalized to a total protein staining of the membrane using Reactive Brown (Sigma), and image analyses were performed in Image Lab 5.2.1.

For multiple comparisons of the effect of exercise and time, two-way ANOVA followed by post hoc tests with Bonferroni correction were performed. The effects of DEN administration in male and female mice were evaluated by investigating statistical alterations in concentrations/expression levels at the different time points after DEN injection in relation to the levels at baseline, that is, in the control mice receiving saline. Linear regression was used for determination of correlation. Data analysis was performed using GraphPad Prism version 6.0 software. Results are depicted as means ± SEM. The criterion for significance was set at a probability of less than 0.05.

To evaluate the effect of sex and voluntary running on DEN-induced inflammatory response, 20–24 weeks old male and female NMRI mice were randomized to standard cages or cages containing running wheels for 6 weeks prior to intraperitoneal injection of DEN or saline. Liver damage was determined after 2, 24, 48, and 96 hours by measuring serum levels of the liver enzymes ASAT and ALAT (Fig. 1). At 48 and 96 hours post-DEN injection, female mice showed significant release of both ASAT (80% and 73%, respectively, P < 0.001) and ALAT (4.0-fold and 4.1-fold, P < 0.001). In contrast, DEN injection in male mice showed very little effect on release of ASAT and only a significant release of ALAT after 24 hours (1.04-fold increase, P < 0.001).

In the female mice, 6 weeks of wheel running significantly reduced the DEN-induced liver damage as evident by the 26% attenuation of ASAT (P < 0.05) and a 43% attenuation of ALAT (P < 0.01) 96 hours post-DEN injection (Fig. 1A). As the mice were housed in pairs to avoid isolation stress, we evaluated training adaptation as an increased expression of the transcriptional coactivator PGC-1α in the tibialis cranialis muscle (22), and we found that increased PGC-1α expression correlated inversely with serum ALAT (P < 0.01) and serum ASAT (P = 0.08) at 96 hours (Fig. 1B). In the male mice on the other hand, voluntary wheel running did not affect the low response to DEN (Fig. 1C). Of note, when comparing muscle tissue from the exercising female mice to the control mice in the 96-hour groups, we found significantly increased PGC-1α expression in the exercise group (Fig. 1D), and furthermore, these NMRI mice showed no significant difference in the running distance between females and males (Fig. 1E). Taken together, these results clearly indicate that female and male mice respond differently when challenged with DEN, with the female mice showing a much greater systemic inflammatory response to the same stimuli, which voluntary wheel running was able to attenuate.

In response to DEN injection, both male and female mice lost body weight in the first days. While the female mice regained all lost weight within 72 hours, the male mice did not restore the original body weight within 96 hours (Fig. 2A). After DEN injection, both female and male mice displayed a significant drop in food intake (P < 0.001, Fig. 2B), and while the male mice restored their food intake within 48 hours, the female mice ate less for up to 72 hours (Fig. 2B). Exercise increased food intake in the SAL groups (males: P < 0.01, females: P < 0.05), but did not prevent DEN-induced anorexia. Of note, the first 24 hours after DEN injection, the mice did not run despite access to running wheels (data not shown). To evaluate whether the drop in food intake was regulating the inflammatory response, we tested the response to DEN injection under fasting (no food), restricted diet (1 g per day equivalent to the observed food intake in Fig. 2B), or free access to food (Fig. 2C) in female mice. As expected, the mice on fasting or restricted diet lost body weight compared with the control at 24 hours post-DEN administration (−19%, P < 0.001; Fig. 2C). While the mice on ad libitum diet mounted a significant IL6 response to DEN injection already after 2 hours (151%, P < 0.05, Fig. 2D), the fasting mice and the mice on restricted diet also showed significant increases in IL6 expression at 24 hours after DEN injection (8.9-fold, P < 0.05 and 33.9-fold, P < 0.001, respectively, Fig. 2D). At this time point, the IL6 responses to DEN were not significantly different between the diet groups.

DEN induces an acute damage to the hepatocytes, which causes the release of IL1β (23). IL1β activates the hepatic Kupffer cells to produce IL6, which in turn further damages the hepatocytes and induces additional IL1β production. If this cycle continues in the hepatic tissue, compensatory proliferation and hepatocarcinogenesis will occur. IL1β was significantly elevated after 96 hours in the female mice compared with the level after 2 hours (1.78-fold, P < 0.01; Supplementary Fig. S1A), while the male mice tended to demonstrate an increase at 48 hours. Wheel running blunted the IL1β response in the male mice (P < 0.05), while wheel running augmented the serum IL1β levels after 96 hours in the female mice (49%, P < 0.05). For IL6, we observed marked inductions at 2 and 24 hours (17.4-fold and 20.1-fold, P < 0.01 and P < 0.001, respectively) in the female mice in response to DEN (Supplementary Fig. S1B), and this induction was attenuated by voluntary wheel running (−74%, P < 0.01, Supplementary Fig. S1B). For the male mice, only a small and nonsignificant increase IL6 was observed at 2 and 24 hours. No significant differences were found in serum TNFα (Supplementary Fig. S1C).

Next, we went on to evaluate the hepatic expression of DEN-catabolizing and detoxification enzymes, inflammatory cytokines, and downstream signaling.

Both female and male mice expressed the DEN-catabolizing enzyme cytochrome P450 (CYP2E1) to a similar extent (Fig. 3A), suggesting that it is not the underlying cause of the observed sex difference. In contrast, the female mice tended to have lower baseline expression levels of the detoxification enzyme glutathione S-transferase (GST; 71% lower, P < 0.001), indicating that the initial clearing of DEN was lower in female mice. Following DEN challenge, GST expression levels were markedly induced, in particular in the females, peaking at 96 hours with a 3.6-fold induction (P < 0.01), while the males only showed a significant 77% increase after 48 hours (P < 0.01, Fig. 3B). No difference was observed between the control and exercising groups (Fig. 3). The detoxification enzyme glucuronide transferase did not differ between female and male mice, or with exercise (Supplementary Fig. S2).

At baseline, the exercising female mice had lower IL1β expression compared with the controls (−62%, P < 0.001, Supplementary Fig. S3A), and the same tendency was seen in the male mice (−54%, not significant, Fig. 5A), suggesting that exercise reduced inflammation in livers at baseline. Following DEN injection, IL1β expression actually decreased (males: P < 0.05 at 24 hours, females: P < 0.01 at 2–48 hours; Supplementary Fig. S3A) with no difference between the control groups and the exercise groups in any of the sexes.

Even though female mice start out with a higher baseline expression of hepatic IL6 expression compared with the males, induction of expression following DEN administration was completely absent (Fig. 4A), in contrast to the elevations in serum levels. This phenomenon has previously been described (16), and may be the underlying mechanistic explanation of why female mice are essentially protected from DEN-induced hepatocarcinoma. To this end, the estrogen receptor alpha, ERα, has been suggested to provide this protective role by downregulating IL6 release from the Kupffer cells. As expected, our female mice expressed much higher levels of ERα than the male mice (51% more, P < 0.01; Fig. 4B), while prior exercise training had only marginally effect on ERα expression (8% more, P = 0.21). The estrogen receptor beta (ERβ) was expressed at much lower levels in the liver of both female and male mice, and accurate measures of this receptor could not be obtained.

In contrast to the female mice, male mice showed upregulations of IL6 mRNA expression at 48 and 96 hours (4.7-fold and 5.8-fold; Fig. 4A), although not significant. This was unaffected by wheel running. TNFα expression was significantly increased in the female control mice after 96 hours (1.98-fold; P < 0.01), while male controls showed no elevations. Exercise did not have any effect on the expression level of TNFα in neither males nor females (Supplementary Fig. S3B). In line with the IL6 expression, male mice tended to show an induction of NFκβ, while this was not observed in the female mice (Supplementary Fig. S4).

The TLR pathway has been suggested to play a prominent role in DEN-induced acute inflammation and later development of carcinogenesis through the recognition of DEN (24, 25). Thus, we investigated differences in the acute immune response between male and female mice in this pathway (Fig. 5). TLR9 is a member of the TLR family, which is responsible for activating the innate immune system (26), and this receptor showed significantly increased expression in the female mice, while the expression was blunted in male mice after DEN injection. TLR9 peaked at 96 hours in the female controls (2.59-fold, P < 0.001), and it also displayed a marked induction at 2 hours (2.48 fold, P < 0.001). These inductions were attenuated by wheel running (50%, P < 0.01 and 48%, P < 0.01, respectively; Fig. 5A). TLR4 is another TLR also activating the innate immune system, and the expression of this receptor also showed a peak at 96 hours in the female controls (TLR4: 1.09-fold, P < 0.001) with attenuation by wheel running (28%, P < 0.01), but at 48 hours, exercise actually increased the induction (45%, P < 0.01; data not shown). MyD88 is a key downstream effector molecule of the TLRs, and the regulation of MyD88 expression also proved to be sex-specific with the female control mice having significantly higher expression after 2 (58%, P < 0.05) and 24 hours (1.75-fold, P < 0.001) with wheel running appearing to postpone the induction at 24 hours, displaying a peak at 48 hours (P < 0.001; Fig. 5B). In contrast, MyD88 expression was not induced in the male mice.

Further downstream effectors of the TLR signaling pathway are the interleukin receptor associated kinases (IRAK), and the mRNA expression levels of IRAK1 showed significant increases in the female mice at 24 hours (88.5%, P < 0.001), with wheel running shifting the peak in expression to 48 hours (1.44-fold, P < 0.001), while the control females had the highest expression at the 24-hour time point (19%, P < 0.05; Fig. 5C). As for MyD88, neither DEN injection nor wheel running had any effect on IRAK1 expression in male mice.

The TLR pathway is highly expressed in innate immune cells, thus we finally evaluated the expression of markers for monocytes/macrophages (CD68; Fig. 5E), dendritic cells (CD209; Fig. 5F), and NK cells (NKp46; Fig. 5D). All three markers displayed similar patterns with expression levels peaking at 96 hours in female control mice (CD68: 2.13-fold P < 0.0001; CD209: 3.33-fold P < 0.0001; NKp46: 1.61-fold P < 0.001), and exercise reduced the induction of all three markers at 96 hours post-DEN administration (CD68: −51%, P < 0.001; CD209: −41%, P < 0.001; NKp46: −49%, P < 0.01). The male mice, on the other hand, only showed induction of CD68 at 24 hours, while CD209 or NKp46 was not regulated at any time point. Here wheel running was not able to attenuate the CD68 expression.

It is well documented that male mice, but not female mice, develop liver tumors after injection of DEN, and voluntary wheel running was recently shown to reduce DEN-induced incidence and tumor burden. We have therefore comparatively investigated the effects of exercise and sex on the acute liver damage and inflammatory response in this particular model. We found that administration of DEN caused a marked acute inflammatory response in female mice as evident by release of liver enzymes and IL6, along with weight loss, reduced food intake, hepatic expression of cytokines, components of the TLR signaling pathway, and markers of immune cells. In contrast, DEN injection in the males had less effect on the systemic and hepatic inflammation, even though the male mice did display marked physiologic effects including weight loss and reduced food intake. These differences might be explained by the findings that female mice had lower expression levels of the detoxifying enzyme, GST, suggesting a longer and stronger presence of the inflammatory inducer, yet the female mice have higher ERα expression, which ameliorates IL6 expression and thus offers long-term protection from carcinogenesis despite a large acute inflammatory response. In addition to these effects, voluntary wheel running could attenuate systemic and hepatic inflammation in the female mice, and shift the peak time of the inflammatory response, which might be mediated through a training-dependent priming of the innate immune cells and TLR signaling.

The sex-dependent differences to DEN injection was first described by Naugler and colleagues (16), demonstrating that male mice developed HCC in an IL6-dependent manner. The authors showed that female mice were protected through a lack of hepatic IL6 induction in response to DEN. This correlates with our results, as we also find that DEN injection does not cause hepatic IL6 mRNA expression in female mice. However, in contrast to the Naugler study (16), which found that acute DEN injection caused an 800-fold increase in serum IL6 in male mice, we hardly saw any increase in serum IL6 in response to DEN injection in the male mice. This difference might be explained by the age difference between the mice. In the Naugler study (16), the mice were 6–8 weeks, while our mice were 20–24 weeks of age. Studies have shown that male mice experience less liver damage in response to DEN, the older they get (17, 27). In line with this, DEN injection on its own only causes HCC if injected into very young mice (2 weeks old); however, if DEN is administered at later ages, additional substances causing liver damage (e.g., phenobarbitals, carbon tetrachloride, or partial hepatectomy) must be coadministered to initiate HCC (27). Although DEN injection might be most efficient in inducing HCC and inflammation in young mice, we chose to conduct our study in older mice. This was chosen as patients developing HCC are elderly, and as age has been shown to play an important role for immune function and inflammatory response (28), we wanted to validate the protective effect of exercise in mice that were not juvenile.

Differences in immune function and inflammation between men and women have been reported (29). Part of the difference may be related to the effect of female sex hormones. For instance, estrogen reduces influenza A virus replication in primary human nasal epithelial cells isolated from women (30), but despite lower viral replication, the disease symptoms are experienced as more severe in premenopausal women compared with men and postmenopausal women due to a stronger inflammatory response (31). Furthermore, premenopausal women with NASH (nonalcoholic steatohepatitis) show decreased risk of developing liver fibrosis compared with men and postmenopausal women (32). Despite this, premenopausal women had an increased risk of developing the hepatic inflammation itself, with lobular inflammation and hepatocyte ballooning, and this risk is further increased by use of oral contraceptives (33).

The finding that premenopausal women have stronger hepatic inflammatory responses, but are less likely to develop fibrosis supports the findings of this study. We found that the female mice had greater hepatic inflammation in response to DEN administration compared with the males, but in spite of this, male mice develop HCC in response to DEN, while the females are essentially protected (21). HCC is thought to derive from cellular hyperproliferation in response to recurrent liver damage. Here, the injection of DEN is thought to provide the initial mutations, which in the long run transforms hepatocytes into cancer cells. As our male mice hardly generate any acute inflammatory response or display signs of liver damage, that is, release of ASAT and ALAT, it implies that the hepatocytes carrying DEN-induced mutations may persist in the liver after the initial injection, and thus at later insults can flourish to hyperproliferative transformed cancer cells. For the female mice, on the other hand, it is widely acknowledged that these mice are protected from DEN-induced HCC through an estrogen-dependent suppression of IL6 signaling (34, 35). The suppression is mediated through estrogen signaling by ERα. In our mice, ERα was highly expressed in the female mice, but not in the male mice. In continuation, we observed an absence of IL6 response in the livers of female mice, regardless of marked increases in the serum levels of liver enzymes, ASAT and ALAT, induction of TNFα signaling and increased infiltration of innate immune cells. This acute response implies that a number of hepatocytes are terminally damaged in response to the DEN injection, yet the estrogen-mediated regulation of IL6 expression and chronic inflammatory ensures that even though the mice have a large acute inflammatory response, long-term carcinogenesis is prevented.

We have previously shown that voluntary wheel running markedly reduces the incidence of DEN-induced HCC, as well as the tumor burden in affected livers of male mice (21). Moreover, we have seen that female mice were essentially protected from the DEN-induced HCC, regardless of access to running wheels (data not shown). Both long-term training and acute exercise can regulate inflammatory responses primarily in an anti-inflammatory direction (12, 36). One way is through effects on the TLR signaling pathway. While long-term training can lower circulating levels of TLR ligands, that is, saturated free fatty acids and oxidized LDLs (12), the regulation of the acute inflammatory response may involve an exercise-dependent decrease in the expression of TLR4 on monocytes and other responsive cells (37, 38). In our study, 6 weeks of wheel running had no effect on the expression of CD68 or CD209 as markers of innate immune cells within the livers at baseline. However, we did observe small and nonsignificant reductions in the hepatic expression of TLR4 and its adaptor molecules, MyD88, in both sexes at baseline. Our data allowed us to follow the acute inflammatory response after DEN injection, and in the female mice, wheel running shifted the time course of the inflammatory response. Infiltration of innate immune cells (CD68, CD209 and NKp46) peaked already at 48 hours in the exercising mice, while the control mice showed higher expression at 96 hours. On the basis of our data, we cannot exclude that the inflammatory response in the control mice might last for even longer. Similarly, expression of the TLR signaling pathway (MyD88, IRAK1, and IRAK4) also showed a shifted response when comparing the control and exercising mice.

TLR signaling in immune cells results in increased expression of IL1β, IL6, and TNFα (39). After 6 weeks of wheel running and prior to DEN injection, both sexes showed lower hepatic expression of IL1β and IL6, while no detectable changes could be determined in the serum levels. In response to DEN injection, the serum levels of IL6 increased dramatically, and in the female mice this increase was markedly attenuated in the exercised mice. During exercise, the contracting muscles release several factors, known as myokines (13, 40). A prominent factor is IL6, which is produced by contracting skeletal muscle cells in a TNFα-independent fashion and released into the blood, where the level of IL6 increases in an exponential fashion during exercise. In this study, it appears that the exercise-induced IL6, which has been demonstrated to have anti-inflammatory effects, is overridden by the TNFα-driven IL6, which is released from the liver as a consequence of liver damage.

Finally, excess hepatic lipid accumulation is a strong inducer of hepatic inflammation (41, 42). We have data showing that voluntary wheel running does not affect hepatic lipid content as determined by MR imaging (data not shown). In contrast, wheel running increases hepatic glycogen content, as evident by larger livers in our exercising mice prior to DEN injection (data not shown). To evaluate whether the hepatic glycogen content played a role in the DEN-induced acute inflammatory response, we fasted or restricted food intake in female mice prior to DEN injection. Despite reduced food intake and a marked drop in body weight, we did not observe any differences in the acute inflammatory response, as determined by serum IL6 levels, underlining that energy status does not affect the acute inflammatory response in this model.

In conclusion, we observed marked induction of the acute inflammatory response in 20–24 weeks old female mice, and this inflammatory response was by exercise attenuated and shifted timewise, leading to a faster and more effective clearance of damaged liver cells. In contrast, 20–24 weeks old male mice were markedly physiologically affected with weight loss and reduced food intake, yet these mice showed little systemic and hepatic inflammation in response to DEN, not allowing for any protective effect of exercise on the acute inflammatory response. These results may hold some of the explanation for how exercise contributes to protection against DEN-induced liver cancer, and elaborate on the gender discrimination in this cancer model.

No potential conflicts of interest were disclosed.

Conception and design: M.L. Bay, P. Hojman

Development of methodology: M.L. Bay, P. Hojman

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.L. Bay

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.L. Bay, P. Hojman

Writing, review, and/or revision of the manuscript: M.L. Bay, J. Gehl, B.K. Pedersen, P. Hojman

Study supervision: P. Hojman

Anne Boye and Lone Christensen are acknowledged for their technical assistance.

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). This work was supported by grants from the Danish Cancer Society, the Novo Nordic Foundation, and Copenhagen University Research funds. The Centre for Physical Activity Research (CFAS) is supported by a grant from TrygFonden.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.