Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 1.
Published in final edited form as: Brain Behav Immun. 2020 May 16;88:230–241. doi: 10.1016/j.bbi.2020.05.043

Interleukin 6-independent metabolic reprogramming as a driver of cancer-related fatigue

Aaron J Grossberg a,b, Elisabeth G Vichaya a,c, Phillip S Gross a,1, Bianca G Ford a, Kiersten A Scott a, Darlene Estrada a, Daniel W Vermeer d, Paola Vermeer d, Robert Dantzer a
PMCID: PMC7415540  NIHMSID: NIHMS1596669  PMID: 32428555

Abstract

Fatigue is a common and debilitating symptom of cancer with few effective interventions. Cancer-related fatigue (CRF) is often associated with increases in inflammatory cytokines, however inflammation may not be requisite for this symptom, suggesting other biological mediators also play a role. Because tumors are highly metabolically active and can amplify their energetic toll via effects on distant organs, we sought to determine whether CRF could be explained by metabolic competition exacted by the tumor. We used a highly metabolically active murine E6/E7/hRas model of head and neck cancer for this purpose. Mice with or without tumors were submitted to metabolic constraints in the form of voluntary wheel running or acute overnight fasting and their adaptive behavioral (home cage activity and fasting-induced wheel running) and metabolic responses (blood glucose, ketones, and liver metabolic gene expression) were monitored. We found that the addition of running wheel was necessary to measure activity loss, used as a surrogate for fatigue in this study. Tumor-bearing mice engaged in wheel running showed a decrease in blood glucose levels and an increase in lactate accumulation in the skeletal muscle, consistent with inhibition of the Cori cycle. These changes were associated with gene expression changes in the livers consistent with increased glycolysis and suppressed gluconeogenesis. Fasting also decreased blood glucose in tumor-bearing mice, without impairing glucose or insulin tolerance. Fasting-induced increases in wheel running and ketogenesis were suppressed by tumors, which was again associated with a shift from gluconeogenic to glycolytic metabolism in the liver. Blockade of IL-6 signaling with a neutralizing antibody failed to recover any of the behavioral or metabolic outcomes. Taken together, these data indicate that metabolic competition between the tumor and the rest of the organism is an important component of fatigue and support the hypothesis of a central role for IL-6-independent hepatic metabolic reprogramming in the pathophysiology of CRF.

1. Introduction

Fatigue is the most common symptom associated with cancer or cancer-therapy, affecting the majority of patients at some point during their disease course (Miller et al., 2008). Cancer-related fatigue (CRF) is defined as a “distressing, persistent, subjective sense of physical, emotional, and/or cognitive tiredness or exhaustion related to cancer and/or cancer treatment that is not proportional to recent activity and interferes with usual functioning” (Mock et al., 2000). Although common, effective treatment approaches to CRF have remained elusive, and the global burden of this disruptive symptom continues to rise as survivorship increases (Bower, 2007). The etiology of CRF is multi-factorial, including social, cognitive, psychological, and biological contributions (Bower, 2014). Yet fatigue is common among patients without pre-existing risk factors and its surrogate, reduced locomotor activity, is reported in multiple mouse models of cancer, implicating a cancer-associated biological mechanism (Dantzer et al., 2014).

Inflammation is the most-commonly investigated pathway linking cancer to fatigue. Multiple cross-sectional studies report an association between fatigue and circulating inflammatory cytokines, such as interleukin-6 (IL-6), c-reactive protein (CRP), and soluble tumor necrosis factor receptors (sTNFRs) in patients with cancer (Clevenger et al., 2012; Lutgendorf et al., 2008; Meyers et al., 2005; Pertl et al., 2013; Wratten et al., 2004). Although inflammatory cytokine action within the brain is clearly sufficient to induce fatigue (Dantzer et al., 2014), several observations suggest that other mechanisms may be at work as well. Anti-cytokine therapy, effective in ameliorating fatigue in patients with psoriasis, remains investigational in CRF, with no phase 3 data demonstrating clinically meaningful efficacy in this population (Bower, 2014; Jatoi et al., 2010; Tyring et al., 2006). In a preclinical attempt to mechanistically link CRF to neuroinflammation, we found that neither genetic nor pharmacologic blockade of IL-1 signaling improved wheel running or exploratory behavior deficits in an inflammatory syngeneic head and neck cancer model (Grossberg et al., 2018). Together, these findings strongly suggest the involvement of yet uninterrogated biological pathways in CRF pathophysiology.

The link between cancer and altered metabolism was first described by Otto Warburg in 1924, when he noted that many cancers engage in fermentation of glucose into lactate, even in the presence of oxygen (Warburg and Posener, 1924). Subsequent research notes that this not only provides energy, but generates building blocks for rapid cancer expansion (Vander Heiden et al., 2009). Clinically, this increased metabolic rate is leveraged using 18F-deoxyglucose positron emission tomography imaging to localize regions that take up excess glucose. However, cancer cells do not simply function as a metabolic sink, as cancer may also reprogram host metabolism, expanding the total metabolic burden of a tumor. This concept is demonstrated by cancer-associated cachexia, in which wasting of fat and lean tissue occurs out of proportion to the caloric deficit (Baracos et al., 2018). Therefore, both the cancer and its effects on global host metabolism generate a competition for resources with the organism. The biological significance of this competition is evidenced by the observations that cancer growth is inhibited in the context of strenuous forced exercise or significant nutrient deprivation (Cheney et al., 1983; Lee et al., 2012; Pedersen et al., 2016; Rusch and Kline, 1944).

Although the mechanism of metabolic fatigue in skeletal muscle and brain is poorly understood, fatigue is commonly associated with the accumulation of metabolic by-products, such as lactate, in the muscles, and by exhaustion of metabolic fuels, including glucose (Fan et al., 2017; Fitts, 1994). Both glucose partitioning and lactate metabolism are largely engineered by the liver in response to nutrient availability and metabolic demands. During the fed or resting states, glucose uptake is increased and the liver stores energy in the form of glycogen. When food is scarce or energetic demands outstrip immediate supply, the liver maintains blood glucose by breaking down glycogen and generating glucose from amino acid and lactate precursors via the Cahill and Cori cycles, respectively (Cahill, 2006; Cori and Cori, 1946). Not only are these substrate cycles important for glucoregulation when energy demand exceeds nutrition, but, in the case of the Cori cycle, it also serves to prevent lactate acidosis in the muscle and allows the muscle to replenish NAD+ stores for further rounds of glycolysis via fermentation of pyruvate to lactate. The liver also plays an important role in the inflammatory response, producing and releasing acute phase proteins, including CRP and inflammatory cytokines—a metabolically costly process (Venteclef et al., 2011). In cancer, not only can the hepatic inflammatory program expand the metabolic footprint of the tumor, but it may also limit its ability to maintain a consistent metabolic environment.

We reasoned that metabolic reprogramming of the liver could underlie CRF by limiting the availability of metabolic fuels for prolonged activity. The present work was undertaken to determine whether hepatic glucose metabolic reprogramming underlies CRF by depletion of metabolic resources via Cori cycle inhibition. For testing our hypothesis, we measured fatigue by wheel running activity that, in addition to serving as a convenient end point to assess fatigue, increases metabolic constraints on the organism. We then used acute fasting as an orthogonal metabolic constraint that potently increases wheel running behavior in healthy mice. As tumor-induced reprogramming of hepatic metabolism has been shown to be mediated by interkeukin-6 (IL-6) (Flint et al., 2016; Wunderlich et al., 2010) and reduction of wheel running activity is associated with increased circulating levels of this cytokine (Baltgalvis et al., 2010; Grossberg et al., 2018), we tested whether cancer-associated fatigue and metabolic changes are mediated by IL-6.

2. Materials and Methods

2.1. Animals.

Male C57BL/6J mice were purchased from The Jackson Laboratory or bred in our own colonies. Experiments were conducted on 10-to 13-week-old male mice individually housed in humidity-controlled environments maintained at 22°C on 12-h light–dark cycles (700-1900h). Food and water were available ad libitum unless otherwise specified. All procedures described in this study were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committees (IACUC) of the University of Texas MD Anderson Cancer Center (Houston, TX).

2.2. Tumor model.

Adult male C57BL/6J mice were injected with 1 x 106 murine E6/E7/hRas (mEER) tumor cells suspended in 100 μl of sterile PBS or an equivalent volume of PBS vehicle into the right flank subcutaneous space, as previously described (Grossberg et al., 2018). Because the mEER cell stock was generated from tonsillar epithelium harvested from a male mouse, attempts to grow these cells in female have not yielded viable tumor growth (Mermod et al., 2018; Spanos et al., 2009). Tumor cells were grown in eMedia (Spanos et al., 2008) and maintained in exponential growth at 37°C in 95% O2/5% CO2 in a humidified incubator. Day of injection was considered “day 0” for the purposes of daily recording. Tumor volume was recorded weekly using Vernier calipers to measure tumor diameter (d) in three orthogonal dimensions [volume = (π/6) x (d1 x d2 x d3)] (Mason et al., 2005). Body weight and food intake were measured at least weekly, and daily food intake was estimated by dividing the total food consumed by the number of days since prior measurement. Mice were euthanized by CO2 approximately 4 weeks after tumor injection, once tumors had reached at least 500 mm3. Mice were monitored daily for general health and euthanized earlier if needed based on IACUC-defined endpoint criteria including tumor diameter ≥1 cm, lethargy, loss of ≥10% body weight, or other signs of sickness or distress. Mice that were euthanized early were not included in analyses. Blood was collected by percutaneous cardiac puncture immediately after euthanasia. Mice were then perfused transcardially with ice-cold PBS. Liver and brain were dissected, sectioned, and snap frozen. Tumors and bilateral gastrocnemius and soleus muscles were isolated, weighed on an analytic balance, and snap frozen. Carcass mass was calculated as terminal body mass less tumor mass. All tissues were stored at −80°C until subsequent analysis.

2.3. Animal studies

2.3.1. Wheel running intervention.

Eighteen adult male mice were counterbalanced into 4 groups (n=4-5/group) based on body weight using a 2x2 factorial design (+/− mEER, +/− wheel). Imbalance in group sizes was to allow for attrition due to rapid or absent tumor growth. Two groups were provided access to counting running wheels, while the other two groups were individually housed within the home cage activity photobeam frames (vide infra). Baseline activity was measured for 7 days prior to tumor or vehicle injection. On day 25, mice were euthanized and tissues harvested between Zeitgeber time (ZT) 14-16, during peak running activity.

2.3.2. Fasting intervention.

Thirty adult male mice were divided into 4 groups (n=7-8/group), counterbalanced by baseline wheel running, using a 2x2 factorial design (+/− mEER, +/− fast). Imbalance in group sizes was to allow for attrition due to rapid or absent tumor growth. On day 34 after tumor injection, at ZT 12, food was removed from the mice assigned to the fasting groups. Mice were euthanized and tissues harvested the morning of day 35 at ZT 2-4.

2.3.3. Glucose and insulin tolerance tests (GTT & ITT).

Twenty adult male mice were individually housed without running wheels and divided into 4 groups (glucose tolerance test +/− tumor & insulin tolerance test +/− tumor; n=5/group), counterbalanced by body weight. For the GTT groups, mice were fasted overnight each week (food removed at ZT 10) and fasting blood glucose was measured using a One Touch Ultra 2 glucose monitor (Lifescan Inc., Milipitas, CA) at ZT 2, after which food was returned. On day 29, following fasting blood glucose measure, blood glucose was measured at 0, 15, 30, 60, and 120 minutes after intraperitoneal injection of 2 g/kg dextrose (Ayala et al., 2010). For the ITT groups, food was removed at ZT 3 on day 30. Following a 4 h fast, mice were given an intraperitoneal injection of 0.5 U/kg insulin and blood glucose was measured at 0, 15, 30, and 60 minutes (Ayala et al., 2010).

2.3.4. IL-6 neutralizing antibody studies.

In the first experiment, mice were placed in 4 groups (n=5/group) counterbalanced by baseline wheel running, using a 2x2 factorial design (+/− mEER; +/− IL-6na). Starting on day 21, mice received i.p. injections of IL-6na (200 μg; BioXcell, BE0046) or isotype control (BioXcell, BE0088) for 3 consecutive days, then every other day (Flint et al., 2016). Mice were euthanized on the morning of day 30 and tissues harvested. For the second experiment, a 3-group design was used (control, IgG; mEER, IgG; mEER, IL-6na, n=6-7/group). Mice were individually housed with running wheels and groups were counterbalanced based on baseline wheel running. Starting on day 14, mice received i.p. injections of IL-6na or isotype control for 3 consecutive days, then every other day until the end of study. At ZT12 on day 28, food was removed from all cages at the time of i.p. treatment. On day 29 blood glucose was measured by glucometer at ZT 3, and mice were euthanized. Plasma was collected and snap frozen.

2.4. Behavioral measures

2.4.1. Voluntary wheel running activity.

Mice were individually housed with in-cage low-profile running wheels, wirelessly connected to an activity counter (Wheel Manager, Med Associates, Inc.). Acclimation to running conditions and activity calculations were conducted as previously described (Grossberg et al., 2018). Fasting-induced change in overnight wheel running was calculated by subtracting the previous night’s overnight wheel running total with ad libitum access to food from the wheel running counts during fasting and expressed as counts.

2.4.2. Home cage activity.

Mice were individually housed in standard shoebox cages that were positioned in 8 x 4 photobeam array frames (PAS HomeCage, San Diego Instruments). Activity was recorded as beam breaks in the x- and y-directions. Overnight home cage activity was expressed as number of beam breaks during the 12-hour dark phase. No running wheel was available for mice undergoing home cage activity measurement. Mice were recorded for 5 consecutive days each week during the study.

2.5. RNA and quantitative PCR.

RNA was extracted using the E.Z.N.A. Total RNA Kit II (Omega) according to the manufacturer’s instructions. cDNA was generated using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). PCR reactions were run on a CFX384 (BioRad) using TaqMan Mastermix with the gene expression assays from Integrated DNA Technologies or with SYBR green primer-based assays and SYBR Green Master Mix (ThermoFisher; see Supplementary Table 1 for primer and probe information). Relative expression was calculated using the ΔΔCt method and was normalized to PBS-injected control. Statistical analysis was performed on the normally distributed ΔCt values. Inflammatory qPCR targets were selected based on our prior inflammatory characterization of the mEER tumor (Grossberg et al., 2018; Vichaya et al., 2017). Metabolic qPCR targets were selected based on previously demonstrated roles in regulating rate limiting steps in glycolysis, lactate production and transport, gluconeogenesis (GNG), lipid metabolism, uncoupling of oxidative phosphorylation, or regulation of metabolic transcriptional programs. Full names and functions of targets are provided in Supplementary material.

2.6. Plasma and tissue analytes.

Blood was stored on ice in EDTA-treated tubes for 30 min and spun at 1,500 x g for 15 minutes before plasma was collected and snap frozen in liquid nitrogen. Plasma was analyzed for glucose using a One Touch Ultra 2 glucometer. β-hydroxybutyrate concentration was measured using colorimetric assay (Cayman Chemical) and read on a plate reader (Synergy HTX). Plasma concentrations of corticosterone (Enzo) and IL-6 (BioLegend) were quantified using ELISA following manufacturer’s instructions. Glucagon and insulin were quantified using multiplex magnetic bead assay (R&D Systems) per manufacturer instructions. The Luminex 200 TM system and Milliplex Analyst were used for detection and analysis. To quantify lactate in plasma or tissue, tissue was first powdered on liquid nitrogen, then deproteinized in 1M perchloric acid on ice. Sample was neutralized with KOH and lactate concentration was quantified using a colorimetric kit (BioVision), per manufacturer’s instructions. Liver glycogen was measured by preparing two aliquots of each sample: one boiled first in 2M HCI, the other boiled first with 2M NaOH. After neutralization and homogenization, glycogen content was assessed by measuring total glucose using a colorimetric assay (Sigma) and subtracting the quantity from the samples boiled first in NaOH from the corresponding samples boiled first in HCI.

2.7. Statistical Analyses.

Statistical significance was calculated using Prism (GraphPad Software, version 8). Wheel running data were analyzed using 2- or 3-way repeated measures ANOVA with multiple comparisons testing performed using Bonferroni-corrected t-test. Cross-sectional data were analyzed using Student’s t test (2 groups, normally distributed data), one-way ANOVA with Bonferroni-corrected t test (3 or more groups, normally distributed data), or 2-way ANOVA with multiple comparisons testing performed using Bonferroni-corrected t-test (2x2 factorial studies), as appropriate. Differences between groups were considered significant when p<0.05. Statistically non-significant comparisons are not reported in the description of the results.

3. Results

3.1. Tumor-associated fatigue is elicited by running wheel challenge

Previously, we observed that tumor-bearing mice with continuous access to a running wheel exhibited a gradual decline in both peak amplitude and duration of wheel running, without any changes in the time of running initiation or cessation (Grossberg et al., 2018). Based on this finding, we questioned whether tumor-bearing mice would exhibit a similar decrease in home cage activity in the absence of a running wheel. We employed a 2x2 factorial design to identify the interaction between the tumor and wheel running on overnight activity (Figure 1A). All mice were acclimated to non-counting running wheels prior to tumor injection to ensure equivalent conditioning. In the absence of a running wheel, tumor-bearing mice showed no decrement in overnight home cage activity throughout the course of tumor growth [FmEER x time (9,63) = 0.65, p=.75, Figure 1B]. In contrast, mice with running wheels showed a progressive decrease in overnight wheel running beginning 16 days after tumor injection [FmEER x time (19,133) = 52.23, p<.01, Figure 1C]. As previously observed, there was no alteration in the time of running onset or termination, but the mice demonstrated a decrease in both peak and duration of wheel running (Figure 1D). We therefore decided to terminate the mice during the period of peak running activity, to interrogate the metabolic status of the mice at maximal metabolic demand (shown in orange hash in Figure 1D). Access to a running wheel did not significantly impact tumor growth [FmEER x time (4,32) = 0.58, p=0.68, Figure 1E & Supplementary Figure 1A], body mass [FmEER x wheel x time (5,70) = 0.42, p=.84, Figure 1F], or food intake (Supplementary Figure 1B). Because decreased wheel running could be secondary to cancer-associated muscle wasting, we looked for evidence of cancer cachexia in these mice. We found no effect of wheel running or tumor on carcass mass (Supplementary Figure 1C), mass of the cachexia-resistant type I myofiber-containing soleus muscle (Supplementary Figure 1D), mass of the cachexia-sensitive type II myofiber-dominant gastrocnemius muscle (Supplementary Figure 1E), or the expression of the catabolic E3 ubiquitin ligase genes Fbxo32 or Trim63 in the gastrocnemius muscle (Supplementary Figures 1F, G) (Braun et al., 2011). We observed a very close correlation between tumor volume and overnight wheel running (Supplementary Figure 1H), but not homecage activity (Supplementary Figure 11). Thus, the tumor has little effect on the activity of an unstimulated mouse with freely available food and water, but in the presence of a running wheel, fatigue becomes apparent in the form of reduced wheel running, without evidence of tumor-related immobility. We propose that this is because the metabolic requirements of running can no longer be met in an organism in which there is a competition for resources between the tumor and skeletal muscles.

Figure 1.

Figure 1.

Open in a new tab

Tumor-induced fatigue elicited by metabolic challenge. (A) Study design for Figures 1 & 2 (n = 5/group for mice injected with mEER tumor cells with or without running wheel; n = 4/group for mice injected with PBS with or without running wheel). (B) Overnight home cage LMA in mEER and CTL mice housed without running wheels. (C) Overnight wheel running activity in mEER and control mice. (D) Hourly wheel running activity 1 d and 24 d after tumor implantation. (E) Tumor growth in mice with and without running wheels. (F) Body mass. Data represented as mean ± SEM. Orange hashed area in (D) represents time of mouse termination. Statistics calculated using 2- or 3-way ANOVA and Bonferroni-corrected t-test. *, p<0.05; **, p<0.01.

We then analyzed the levels of macronutrients in the plasma and tissues of mice terminated during their activity peak. Plasma glucose remained normal in unchallenged tumor-bearing mice, whereas running mEER mice exhibited lower plasma glucose levels [FmEER x wheel (1,14) = 4.85, p<.05, Figure 2A]. The plasma levels of the ketone β-hydroxybutyrate, which is typically elevated following prolonged fasting and glycogen depletion when lipid oxidation becomes the primary fuel, were not significantly elevated by wheel running. However, an interaction between tumor and wheel running resulted in modestly decreased plasma levels of this metabolite in tumor-bearing, running mice [FmEER x wheel (1,14) = 4.85, p<.05, Figure 2B]. Lactate, the principal metabolic waste product from anaerobic glycolysis was decreased in wheel running mice, suggesting that the intensity of voluntary wheel running is inadequate to engage anaerobic metabolism in conditioned mice. An interaction between tumor and wheel running appeared to modestly suppress circulating lactate levels, although this may reflect the decreased running in these animals [FmEER x wheel (1,14) = 5.10, p<.05; Fwheel (1,14) = 5.02, p<.05, Figure 2C]. Circulating lactate is taken up by the liver, where it provides a major substrate for gluconeogenesis in the hepatic arm of the Cori Cycle, thereby both renewing the glycolytic substrate, glucose, while allowing the muscle to regenerate NAD+ to continue glycolytic metabolism. Neither wheel running nor the presence of a tumor significantly impacted lactate concentration in the liver, and wheel running decreased brain lactate concentration with no impact of tumor (Supplementary Figure 2A, B). Despite elevated rates of glycolytic metabolism, tumor lactate concentrations were not significantly different in running mice, indicating no major contribution to total lactate generation (Supplementary Figure 2C). Together, these data suggest that the production of lactate by wheel running did not exceed the rate of global uptake and metabolism (Bergman et al., 1999). Lactate concentration in the gastrocnemius muscle, however, was elevated nearly two-fold in running tumor-bearing mice as compared to the CTL group as well as both groups without running wheels [FmEER x wheel (1,14) = 5.69, p<.05, Figure 2D]. Skeletal muscle lactate buildup is associated with muscle fatigue, supporting a metabolic explanation for the observed decrease in wheel running.

Figure 2.

Figure 2.

Open in a new tab

Wheel running elicits alterations in circulating metabolites and hepatic gene expression. Plasma glucose (A), β-hydroxybutyrate (B), and lactate (C) concentrations in CTL and tumor-bearing mice with or without access to running wheels. (D) Lactate quantity in the gastrocnemius muscle. (E) Ratio of Pkm1 to Pkm2 expression in the liver. (F) Plasma IL-6. (G) Heat map showing differential expression of metabolic and inflammatory genes in the liver. Data represented as mean ± SEM (A-F) or log2-transformed relative expression (G). Statistical significance calculated using 2-way ANOVA with Bonferroni-corrected t-test. *, p<.05; **, p<.01; ***, p<.001. Abbreviations: GNG, gluconeogenesis; TF, transcription factors; m, mEER; w, wheel; m x w, mEER x wheel.

We then evaluated the expression of metabolic and inflammatory genes in the gastrocnemius muscle, brain, and liver to better understand glucose metabolism in these three tissues. Wheel running, but not tumor, modified the expression of multiple glycolytic genes in the gastrocnemius without any significant induction of Il6 or Fbxo32 or Trim63, the E3 ubiquitin ligases associated with catabolism and cachexia (Supplementary Figure 2D and Supplementary Table 3) (Bodine et al., 2001; Gomes et al., 2001). In the brain, wheel running increased the expression of the lactate importer, Slc16a1, and the G-protein coupled receptor activated by lactate, Hcar1; the mitochondrial oxidative phosphorylation uncouplers Ucp2 and Ucp3; and inflammatory cytokines Il1b and Tnf. Tumor also increased brain expression of both Ucp2 and Ucp3 as well as Il1b, but no interaction between wheel running and tumor was observed (Supplementary Figure 2E and Supplementary Table 4). Tumor exhibited a more profound effect on liver metabolic genes, as indicated by increased expression of key factors involved in glycolysis (Slc2a1, Hk1, Hk2, Pdk4, Pkm2), lactate metabolism (Ldha, Slc16a1, Slc16a3, Hcar1), lipid metabolism (Cpt1, Srebp, Ppara), and OXPHOS uncoupling (Ucp2, Ucp3). Conversely, expression of the rate-limiting enzymes involved in gluconeogenesis, G6pc and Pck1, was decreased in tumor-bearing mice (Figure 2G and Supplementary Table 2). A decrease in the ratio of Pkm1/Pkm2 expression, associated with a shift to aerobic glycolysis (Yang et al., 2012), was also observed in tumor-bearing mice [FmEER (1,14) = 13.67, p<.01; Fwheel (1,14) = 8.81, p<.05, Figure 2E]. The expression of inflammatory cytokines Il1b, Il6, and Tnf was increased in tumor bearing mice (Figure 2G), as was the concentration of IL-6 protein in the plasma [FmEER x wheel (1,14) = 5.91, p<.05, Figure 2F]. Taken together, the tissue gene expression data show that the metabolic program of the liver, but much less so the skeletal muscle or brain, are significantly altered by the presence of a tumor. The hepatic gene expression changes are consistent with increased glycolytic and decreased gluconeogenic activity, posing altered carbohydrate metabolism in the liver as a key driver of the decreased plasma glucose levels observed in tumor-bearing mice challenged by wheel running. Given the connection between plasma glucose, hepatic gluconeogenesis, and muscle lactate in the Cori cycle, a hepatic metabolic program favoring glycolysis and blocking gluconeogenesis could explain both the decrease in circulating glucose and lactate accumulation in the skeletal muscle.

3.2. mEER tumor induces fasting hypoglycemia but does not impact glucose or insulin tolerance

As wheel running revealed differential glucose regulation, we next tested glucose homeostasis in the fasting condition. Mice injected with tumor cells underwent weekly overnight fasts. Tumor growth (Figure 3A) was similar to prior studies (Figure 1E). As early as one week after tumor implantation, when tumors were just becoming palpable, tumor-bearing mice demonstrated a decrease in fasting blood glucose compared to CTL, which persisted throughout tumor growth [FmEER (1,8) = 17.68, p<.01, Figure 3B]. Following their final fast, mice were subjected to a glucose tolerance test, which showed no impairment in glucose tolerance, but rather increased clearance from the bloodstream in tumor-bearing mice [FmEER (1,8) = 12.92, p<.01, Figure 3C]. Insulin tolerance was assessed in a second cohort of mice injected with tumors concomitantly with the cohort used for GTT. No difference was seen in ITT between mEER or CTL mice [FmEER (1,8) = 0.03, p=.88, Figure 3D]. Thus, in both the fasting and running conditions, glucose production in tumor bearing mice is outstripped by uptake. Because the liver is a major contributor to both uptake (glycolysis and glycogen synthesis) and production (gluconeogenesis and glycogenolysis), a shift in the hepatic metabolic program favoring glycolysis, as seen in running tumor-bearing mice, could influence both sides of the glucoregulatory balance and account for the decrement in circulating glucose.

Figure 3.

Figure 3.

Open in a new tab

mEER tumor induces hypoglycemia without evidence of glucose or insulin intolerance. Non-running mice were injected with mEER cells or PBS (n=5/group). Separate groups used for GTT and ITT. (A) Tumor growth. (B) Blood glucose following 16 h fast throughout tumor growth. (C) Glucose tolerance test (GTT) 29 d after tumor implantation. (D) Insulin tolerance test (ITT) 30 d after tumor implantation. Data represented as mean ± SEM. Statistics calculated using 2-way ANOVA with Bonferroni-corrected t-test. *, p<0.05; **, p<0.01.

3.3. mEER tumor suppresses adaptive behavioral and physiologic responses to fasting

If the observed wheel running decrease is secondary to metabolic limitations, we would expect these mice to be unable to increase their wheel running upon stimulation. Mice respond to fasting by acutely increasing locomotor activity, presumably in service of foraging for food, providing an optimal model to test the metabolic constraint hypothesis of fatigue. Fasting exercising mice triggers a wide variety of behavioral and metabolic adjustments affecting glucose and lipid metabolism (Acosta-Rodriguez et al., 2017). This can serve to amplify the negative consequences of metabolic reprogramming induced by tumor growth.

In this study all mice had access to running wheels, to be used as an endpoint rather than an intervention. After the tumors had exceeded 500 mm3 in volume, mice were subjected to an overnight 16 h fast, and behavioral, metabolic, and liver gene expression endpoints were evaluated (Figure 4N). Tumors grew more slowly in this study, and a decrease in wheel running was not observed until 22 days after injection [FmEER x time (33,858) = 4.74, p<.0001, Figure 4A]. On the evening of day 34 post-injection, tumor-bearing and CTL mice were either allowed ad libitum access to food or had all food removed overnight; prior to fasting there was no difference in food intake among the four groups (Supplementary Figure 3A). CTL mice showed the expected fasting-associated increase in wheel running, whereas fasted mEER mice exhibited no such increase in activity [FmEER x fast x time (2,49) = 5.16, p<.01, Figure 4B; FmEER x time (1,25) = 7.73, p=.01, Figure 4D]. The increase in activity in fasted CTL mice was manifested as an increase in maximal running duration, but not amplitude, whereas tumor-bearing mice showed no increase in running duration (Figure 4C). Both tumor and fasting independently decreased plasma glucose, with the greatest effect seen in fasted, tumor-bearing mice [FmEER (1,25) = 7.25, p<.05; Ffast (1,25) = 15.34, p<.001, Figure 4E]. The fasting-induced increase in ketone body production seen in CTL mice was absent from tumor-bearing mice [FmEER x Fast (1,25) = 5.75, p<.05, Figure 4F]. Ad libitum fed tumor-bearing mice had significantly lower plasma insulin levels than CTL mice, and although fasting decreased insulin in CTL mice, no decrease in insulin was observed in tumor-bearing mice [FmEER x Fast (1,25) = 13.53, p<.01, Figure 4G]. Corticosterone was increased by tumor and fasting, yet there appeared to be a ceiling effect induced by fasting, as no further increase was elicited by the mEER tumor [FmEER x Fast (1,25) = 4.61, p<.05, Figure 4H]. Plasma glucagon concentrations were elevated in tumor-bearing mice with no effect of fasting [FmEER (1,25) = 17.71, p<.001, Figure 4I]. Liver glycogen levels showed a slight decrease in fasted CTL mice, compared to ad libitum fed. However, in tumor bearing mice, fasting was paradoxically associated with an increase in glycogen content [FmEER x Fast (1,25) = 5.41, p<.05, Figure 4J]. In the liver, expression of the glucose transporter Slc2a1, two key glycolytic enzymes (Hk1 and Hk2) as well as genes associated with the conversion of pyruvate to lactate (Ldha) and subsequent lactate export (Slc16a3) were increased in tumor-bearing mice. Liver gene expression again favored glycolysis over gluconeogenesis (Figure 4M and Supplementary Table 5). The expression of the rate-limiting gluconeogenic enzymes G6pc and Pck1 were elevated by fasting in CTL mice, whereas their expression was suppressed in tumor-bearing mice. Similarly, the effect of fasting on the expression of the transcription factors associated with fatty acid synthesis (Srebp) and β-oxidation (Ppara) was suppressed in mice with tumors. The expression of mitochondrial uncoupling proteins Ucp2 and Ucp3 were increased in tumor-bearing mice (Figure 4M). Fasting had no effect on tumor-associated increases in circulating IL-6 [FmEER (1,25) = 19.16, p<.001, Figure 4K] or hepatic Il6 expression [FmEER (1,25) = 40.1, p<.001, Figure 4L]. Because tumors were allowed to grow for an additional week in this study, we repeated the experiment, terminating the mice one week earlier, when tumor size was similar to other experiments (Supplementary Figures 3B, 3C). We found similar effects on plasma glucose and the expression of a subset of genes involved in glycolysis, gluconeogenesis, lactate metabolism, and lipid metabolism (Supplementary Figures 3D–J). Thus, in tumor bearing mice, we see impairment in the adaptive response to fasting, leading to decreases in circulating glucose and ketones, impaired glycogen mobilization, and lack of increased wheel running. That the liver gene expression data coincide with these phenotypic changes in two models of metabolic constraint (fasting and wheel running) suggests that the behavioral manifestations may be a direct consequence of the metabolic limitations imposed by the tumor.

Figure 4.

Figure 4.

Open in a new tab

mEER tumor suppresses adaptive responses to fasting. (A) Voluntary wheel running in mEER tumor-bearing and CTL mice. (B) Voluntary overnight wheel running at baseline, on last day of ad libitum feeding (d 34) and during fast. (C) Hourly voluntary wheel running corresponding to final night of ad libitum feeding and during fasting. (D) Change in overnight wheel running between final night of ad libitum feeding and during fasting. Plasma glucose (E), β-hydroxybutyrate (F), insulin (G), corticosterone (H), and glucagon (I) at time of termination. (J) Liver glycogen stores at time of euthanasia. (K) Plasma IL-6 concentrations. (L) Liver Il6 expression. (M) Heat map showing differential expression of metabolic genes in the livers of CTL and mEER mice fed ad libitum or fasted. (N) Study design for A-M (n=7-8/group). One mEER mouse was dropped from plasma analysis due to insufficient blood collection; data from this mouse were included in wheel running and anthropometric analyses. Data represented as mean ± SEM (A-L) or log2-transformed relative expression (M). Statistical significance calculated using 2- or 3-way ANOVA with Bonferroni-corrected t-test. *, p<.05; **, p<.01; ***, p<.001. Abbreviations: GNG, gluconeogenesis; TF, transcription factors; m, mEER; f, fast; m x f, mEER x fast.

3.4. mEER-associated metabolic reprogramming is independent of IL-6 signaling

Given the consistent association of IL-6 with CRF in cancer patients and preclinical models, recent work showing IL-6 mediates metabolic changes in the liver that impact muscle wasting, immune response, and metastasis (Flint et al., 2016; Goncalves et al., 2018; Lee et al., 2019), and evidence that IL-6 is elevated in mEER tumor-bearing mice that exhibit decreased wheel running, we next tested the role of IL-6 signaling on fatigue and the metabolic changes associated with the mEER tumor. In the first study, CTL and tumor-bearing mice were administered IL-6na starting after a persistent wheel running decrement was observed, using a 2x2 factorial design (Figure 5A). mEER tumor was associated with decreased wheel running, which was not improved by IL-6na [FmEER x fast x time (30,480) = 1.60, p<.05, Figure 5B]. Although IL-6na appeared to reduce wheel running slightly in this study, a repeat examination of IL-6na against isotype control in CTL mice revealed no differences in wheel running or body mass associated with IL-6na treatment (Supplementary Figure 4A, 4B). IL-6na did not alter body mass [FmEER x IL6na x time (8,128) = 0.29, p=.97, Figure 5C] or tumor volume [t(8) = 0.68, p=.52, Figure 5E]. IL-6na had no impact on mEER-associated decreases in plasma glucose [FmEER (1,16) = 21.45, p<.001, Figure 5F]. Plasma β-hydroxybutyrate levels were unchanged by tumor or IL-6na administration [FmEER x IL6na (1,16) = 0.81, p=.38, Figure 5G]. Plasma corticosterone was elevated in tumor-bearing mice, irrespective of IL-6 signaling [FmEER (1,16) = 24.1, p<.001,Figure 5H]. Hepatic gene expression again suggested an inflammatory, glycolytic phenotype associated with mEER tumor, with decreased expression of the gluconeogenic enzymes G6pc and Pck1 (Figure 5D and Supplementary Table 6). IL-6na had no effect on any of the genes evaluated and no interaction with mEER tumor.

Figure 5.

Figure 5.

Open in a new tab

mEER-associated metabolic reprogramming is independent of IL-6 signaling. (A) Study design for B-H (n=5/group). (B) Voluntary wheel running in mEER tumor-bearing and CTL mice treated with IL-6na or IgG control. (C) Body mass throughout study course. (D) Heat map showing differential expression of metabolic genes in the livers of CTL and mEER mice treated with IL-6na or IgG control. (E) Terminal tumor volume in mice treated with IL-6na or IgG control. Non-fasting terminal plasma glucose (F), β-hydroxybutyrate (G), and corticosterone (H). (I) Study design for J-P (n=6-7/group). (J) Voluntary wheel running in second study evaluating the effect of IL-6na on fasting-associated adaptations. (K) Change in overnight wheel running between final night of ad libitum feeding and during fasting. (L) Tumor growth. Body mass (M) and daily food intake (N) during study. (O) Fasting blood glucose. (P) Plasma IL-6 concentrations. Vertical arrows represent treatment with IL-6na or IgG. Data represented as mean ± SEM (B-C, E-H, J-P) or log2-transformed relative expression (D). Statistical significance calculated using 3-way, 2-way, or 1-way ANOVA with Bonferroni-corrected t-test. *, p<.05; **, p<.01; ***, p<.001. Abbreviations: GNG, gluconeogenesis; TF, transcription factors; m, mEER; Ab, IL-6na; m x Ab, mEERx IL-6na.

We then performed a companion study to determine the effect of IL-6na on tumor-associated impairment of fasting adaptations, with the IL-6na intervention started as soon as activity decrease was observed (Figure 5I). Again, wheel running was decreased in tumor-bearing mice, with no effect of IL-6na [Fgroup x time (56,501) = 4.01, p<.001, Figure 5J]. Suppression of fasting-associated increases in wheel running by the mEER tumor was not impacted by IL-6 blockade [F (2,17) = 8.06, p<.01, Figure 5K]. Tumor growth was unchanged by IL-6na [FIL-6na x time (5,45) = 0.02, p=1.0, Figure 5L], and body mass [Fgroup x time(28,266) = 0.66, p=.91, Figure 5M] and food intake were not altered by IL-6na or mEER tumor [Fgroup x time (24,225) = 1.23, p=.21, Figure 5N]. IL-6na had no effect on the decrease in fasting blood glucose induced by mEER tumor [FmEER (2,17) = 5.71, p<.05, Figure 5O]. Plasma IL-6 concentration was significantly lower in IL-6na-treated tumor-bearing animals than either isotype control-treated CTL or mEER mice [FmEER (2,17) = 20.28, p<.001, Figure 5P]. Based on these studies, we conclude that the tumor-associated fatigue and metabolic phenotypes are independent of the activity of circulating IL-6.

4. Discussion

The neuroinflammatory hypothesis of CRF is based on evidence from multiple clinical and preclinical studies linking fatigue to increases in circulating cytokines and a well-established role for central cytokine signaling in the pathogenesis of fatigue-like behavior (Dantzer et al., 2014). This is supported by preclinical work showing an improvement in wheel running with administration of ibuprofen to mice bearing the CT26 colon tumor, prior to the onset of muscle dysfunction, implicating a neuroinflammatory etiology for fatigue (Norden et al., 2015a; Norden et al., 2015b). However, we found that despite evidence of increased brain Il1b production in mice bearing mEER tumors, genetic ablation of IL-1 signaling had no impact on fatigue behaviors, suggesting non-neuroinflammatory processes were involved (Grossberg et al., 2018). Because cancer and the inflammatory response exact a significant metabolic toll on the host, we sought to investigate whether the metabolic landscape of cancer contributes to this common and pervasive morbidity. Our data suggest that CRF can be explained by a competition for resources between the skeletal muscles and tumor. This could help rationalize the penetrance and pervasiveness of fatigue among cancer patients, for whom normal daily activity likely places substantial metabolic drain on an already burdened physiologic state. Fatigue may then be an adaptive response in an organism having limited metabolic resources. We reasoned that this competition should become more visible in conditions of increased metabolic expenses due to physical activity and to acute fasting.

Beyond direct competition for resources, the biochemical and gene expression data also indicate that the tumor reprograms the liver, impairing its ability to meet energetic demands during metabolic challenges, such as wheel running or fasting. This is evident not only in the behavioral alterations induced by the tumor, but also in the mice’s inability to maintain blood glucose levels and diminished fasting ketone production, two key roles of the liver during acute nutritional challenge. We do not believe that these metabolic challenges are required for hepatic reprogramming—indeed, tumor had the largest impact on liver gene expression regardless of challenge—only that these challenges better reveal the impacts of this metabolic limitation on behavior and physiology. How metabolic exhaustion at the level of the muscle is communicated to the brain and how that influences top-down behavioral control is largely unknown, as peripheral metabolite effectors of exhaustion, such as lactate, do not appear to cause fatigue like behavior when delivered to the brain (Matsui et al., 2017; Quistorff et al., 2008). Hepatic metabolic reprogramming could impact activity by inhibiting the Cori cycle, thereby limiting gluconeogenesis and leading to a proximal buildup of lactate (Figure 6). We did not observe any increase in circulating lactate, but prior studies have demonstrated that such signal is best assessed at the level of the skeletal muscle venous drainage bed (Brooks et al., 1992). Nevertheless, the muscle showed substantial elevation of lactate levels; accumulation of this metabolite in muscles is the biochemical marker most commonly associated with muscle exhaustion (Brooks, 1991). This finding may reflect muscle-intrinsic metabolic limitations; however, we observed no clear change in any glycolytic or lactate transport gene expression, nor was there any evidence of muscle catabolism. Meanwhile, in the liver we noted decreased expression of Pck1 and G6pc, the rate-limiting enzymes of gluconeogenesis, in concert with increased expression of the glucose transporter, Slc2a1, hexokinase genes (Hk1 and Hk2), and lactate dehydrogenase (Ldha) suggesting a glucose-consumptive phenotype that could account for both the decrease in circulating glucose and the buildup of lactate in muscle. It remains unclear whether this buildup of lactate in the muscle is a consequence of hepatic reprogramming or local mechanisms. The consistent decrease in circulating glucose levels could also impact muscle performance, as running-induced hypoglycemia is associated with skeletal muscle metabolic exhaustion (Fan et al., 2017). Our GTT data establish that increased glucose clearance is a component of decreased glucose levels in tumor-bearing mice. Although some of this may be explained by tumor uptake, the shifting of hepatic metabolism away from gluconeogenesis and toward glycolysis suggests that the liver may be the primary contributor to this process. Thus, fatigue may be mediated by internal competition between tumor and muscle for limited glucose resources (Figure 6).

Figure 6.

Figure 6.

Open in a new tab

Proposed model for cancer-associated fatigue. In response to energy challenge (fasting, exercise), the liver maintains blood glucose levels by suppressing glucose import and glycolysis and favoring gluconeogenesis. Lactate, produced by exercising muscles, is converted to glucose via the Cori cycle, thereby providing fuel for further skeletal muscle utilization (left). In tumor-bearing mice (right), this adaptive response is suppressed, leading to decreased glucose availability and accumulation of lactate in the skeletal muscles, limiting further exertion.

In an attempt to understand how cancer reprograms hepatic metabolism, we targeted IL-6, the circulating factor most commonly associated with CRF (Wang et al., 2012). In mouse models of lung and pancreatic cancer, IL-6 acts as a long-distance signal from the tumor to the liver, leading to changes in hepatic metabolism that are important for the immune response to tumor, metastasis, and circadian rhythmicity (Flint et al., 2016; Lee et al., 2019; Masri et al., 2016). Although neither fasting nor wheel running impacted IL-6 concentration or Il6 gene expression, we reasoned that tumor-associated elevations in IL-6 could still drive the underlying susceptibility to metabolic challenge. Inhibition of IL-6 signaling using a neutralizing antibody had no impact on fatigue, circulating glucose, or hepatic gene expression in these mice. Similarly, Borniger and colleagues found no impact of IL-6 neutralization on the metabolic phenotype observed in mice bearing mammary tumors (Borniger et al., 2018). However, in contrast to the mEER tumor, the 67NR mammary tumors used in their studies resulted in mild hyperglycemia, which was reversible by either orexin antagonist administration or sympathectomy. Given that the mEER tumor yields an opposing phenotype, we feel it is unlikely that this mechanism is at play in these mice. Despite IL-6 blockade, the livers of mEER tumor-bearing mice continued to express inflammatory cytokine genes at approximately the same level as isotype control-treated mice, indicating that the role of IL-6 in mEER tumor-associated physiology occurs downstream of the liver. The heterogeneity of phenotypes and mechanisms underlying tumor-to-liver communication mirror that of cancer biology and offers some insight into why reversing cancer-associated morbidities remains so challenging.

4.1. Limitations

Although our study provides provocative data linking CRF to metabolic constraints placed on the organism by the tumor and its effects on adaptive physiology, our experiments cannot confirm a causative relationship. Therefore, restricted to association, these data must be considered hypothesis-generating, with liver-specific approaches necessary for confirmation. It is furthermore unclear how generalizable our findings are to humans and to other tumor types. As described above, the influence of cancers on peripheral metabolites and inflammatory mediators is highly heterogeneous, with tumor type, clonality, and location of growth all playing a role (Borniger et al., 2018; Kir et al., 2014; Mayers et al., 2016). In our studies, we were limited to heterotopic growth, as the orthotopically grown tumors rapidly occlude the aerodigestive tract, and male mice, as our tumors are immunogenically cleared by female mice (data not shown). Indeed, the mEER-like tumors derived from female epithelium do not exhibit the same growth dynamics, demonstrating the inherent limitations of single-model studies. Our findings are based on small numbers of mice per group (n=4-8), which both increases the risk of false-negatives and makes studies more susceptible to outliers. Because similar measures were repeated across multiple studies with concordant results, we believe the reliability of the reported positive findings is high. Finally, wheel running is an imperfect surrogate for fatigue in patients, as mice clearly derive secondary gain from the running activity (Walker and Mason, 2018). Yet, our prior studies show no evidence of anhedonia in the mEER model, so we believe its use as both an intervention and an outcome are reasonable given the purposes of the present study (Vichaya et al., 2017).

5. Conclusions

CRF can be elicited by increasing the metabolic constraints on the host. Localized cancer can reprogram liver metabolism, impairing the adaptive responses to energy depletion in an IL-6-independent manner. These metabolic changes both decrease glucose availability and increase lactate accumulation in skeletal muscles, which may accelerate local metabolic exhaustion, offering a non-inflammatory mechanism for fatigue. This unaddressed metabolic contribution to CRF may explain the limited successes of fatigue-directed interventions.

Supplementary Material

1

Highlights.

  • Increased activity demand, such as that provided by access to a voluntary running wheel, elicits a cancer-related fatigue phenotype

  • Tumors reprogram hepatic metabolism to favor glycolysis over gluconeogenesis leading to hypoglycemia and lactate accumulation in the muscle

  • Cancer impairs adaptive behavioral and metabolic responses to fasting

  • Tumor-induced hepatic reprogramming is independent of IL-6 signaling

Acknowledgements

Funding: This work was supported by the National Cancer Institute of the National Institutes of Health [R01 CA193522, to R. Dantzer]. Additional support came from the University of Texas MD Anderson Cancer Center and the NIH MD Anderson Cancer Center Support Grant [P30 CA016672].

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of interest

Robert Dantzer has received honoraria from Danone Nutricia Research France and Pfizer USA unrelated to the present research.

References

  1. Acosta-Rodriguez VA, de Groot MHM, Rijo-Ferreira F, Green CB, Takahashi JS, 2017. Mice under Caloric Restriction Self-Impose a Temporal Restriction of Food Intake as Revealed by an Automated Feeder System. Cell Metab 26, 267–277.e262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ayala JE, Samuel VT, Morton GJ, Obici S, Croniger CM, Shulman GI, Wasserman DH, McGuinness OP, 2010. Standard operating procedures for describing and performing metabolic tests of glucose homeostasis in mice. Dis Model Mech 3, 525–534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baltgalvis KA, Berger FG, Pena MM, Mark Davis J, White JP, Carson JA, 2010. Activity level, apoptosis, and development of cachexia in Apc(Min/+) mice. J Appl Physiol (1985) 109, 1155–1161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baracos VE, Martin L, Korc M, Guttridge DC, Fearon KCH, 2018. Cancer-associated cachexia. Nat Rev Dis Primers 4, 17105. [DOI] [PubMed] [Google Scholar]
  5. Bergman BC, Wolfel EE, Butterfield GE, Lopaschuk GD, Casazza GA, Horning MA, Brooks GA, 1999. Active muscle and whole body lactate kinetics after endurance training in men. J Appl Physiol (1985) 87, 1684–1696. [DOI] [PubMed] [Google Scholar]
  6. Bodine SC, Latres E, Baumhueter S, Lai VK, Nunez L, Clarke BA, Poueymirou WT, Panaro FJ, Na E, Dharmarajan K, Pan ZQ, Valenzuela DM, DeChiara TM, Stitt TN, Yancopoulos GD, Glass DJ, 2001. Identification of ubiquitin ligases required for skeletal muscle atrophy. Science 294, 1704–1708. [DOI] [PubMed] [Google Scholar]
  7. Borniger JC, Walker Ii WH, Surbhi Emmer, K.M., Zhang N, Zalenski AA, Muscarella SL, Fitzgerald JA, Smith AN, Braam CJ, TinKai T, Magalang UJ, Lustberg MB, Nelson RJ, DeVries AC, 2018. A Role for Hypocretin/Orexin in Metabolic and Sleep Abnormalities in a Mouse Model of Non-metastatic Breast Cancer. Cell Metab 28, 118–129.e115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bower JE, 2007. Cancer-related fatigue: links with inflammation in cancer patients and survivors. Brain Behav Immun 21, 863–871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bower JE, 2014. Cancer-related fatigue--mechanisms, risk factors, and treatments. Nat Rev Clin Oncol 11, 597–609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Braun TP, Zhu X, Szumowski M, Scott GD, Grossberg AJ, Levasseur PR, Graham K, Khan S, Damaraju S, Colmers WF, Baracos VE, Marks DL, 2011. Central nervous system inflammation induces muscle atrophy via activation of the hypothalamic-pituitary-adrenal axis. J Exp Med 208, 2449–2463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Brooks GA, 1991. Current concepts in lactate exchange. Med Sci Sports Exerc 23, 895–906. [PubMed] [Google Scholar]
  12. Brooks GA, Wolfel EE, Groves BM, Bender PR, Butterfield GE, Cymerman A, Mazzeo RS, Sutton JR, Wolfe RR, Reeves JT, 1992. Muscle accounts for glucose disposal but not blood lactate appearance during exercise after acclimatization to 4,300 m. J Appl Physiol (1985) 72, 2435–2445. [DOI] [PubMed] [Google Scholar]
  13. Cahill GF Jr., 2006. Fuel metabolism in starvation. Annu Rev Nutr 26, 1–22. [DOI] [PubMed] [Google Scholar]
  14. Cheney KE, Liu RK, Smith GS, Meredith PJ, Mickey MR, Walford RL, 1983. The effect of dietary restriction of varying duration on survival, tumor patterns, immune function, and body temperature in B10C3F1 female mice. J Gerontol 38, 420–430. [DOI] [PubMed] [Google Scholar]
  15. Clevenger L, Schrepf A, Christensen D, DeGeest K, Bender D, Ahmed A, Goodheart MJ, Penedo F, Lubaroff DM, Sood AK, Lutgendorf SK, 2012. Sleep disturbance, cytokines, and fatigue in women with ovarian cancer. Brain Behav Immun 26, 1037–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cori CF, Cori GT, 1946. Carbohydrate metabolism. Annu Rev Biochem 15, 193–218. [DOI] [PubMed] [Google Scholar]
  17. Dantzer R, Heijnen CJ, Kavelaars A, Laye S, Capuron L, 2014. The neuroimmune basis of fatigue. Trends Neurosci 37, 39–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fan W, Waizenegger W, Lin CS, Sorrentino V, He MX, Wall CE, Li H, Liddle C, Yu RT, Atkins AR, Auwerx J, Downes M, Evans RM, 2017. PPARdelta Promotes Running Endurance by Preserving Glucose. Cell Metab 25, 1186–1193.e1184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fitts RH, 1994. Cellular mechanisms of muscle fatigue. Physiol Rev 74, 49–94. [DOI] [PubMed] [Google Scholar]
  20. Flint TR, Janowitz T, Connell CM, Roberts EW, Denton AE, Coll AP, Jodrell DI, Fearon DT, 2016. Tumor-Induced IL-6 Reprograms Host Metabolism to Suppress Anti-tumor Immunity. Cell Metab 24, 672–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gomes MD, Lecker SH, Jagoe RT, Navon A, Goldberg AL, 2001. Atrogin-1, a muscle-specific F-box protein highly expressed during muscle atrophy. Proc Natl Acad Sci U S A 98, 14440–14445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Goncalves MD, Hwang SK, Pauli C, Murphy CJ, Cheng Z, Hopkins BD, Wu D, Loughran RM, Emerling BM, Zhang G, Fearon DT, Cantley LC, 2018. Fenofibrate prevents skeletal muscle loss in mice with lung cancer. Proc Natl Acad Sci U S A 115, E743–e752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Grossberg AJ, Vichaya EG, Christian DL, Molkentine JM, Vermeer DW, Gross PS, Vermeer PD, Lee JH, Dantzer R, 2018. Tumor-Associated Fatigue in Cancer Patients Develops Independently of IL1 Signaling. Cancer Res 78, 695–705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jatoi A, Ritter HL, Dueck A, Nguyen PL, Nikcevich DA, Luyun RF, Mattar BI, Loprinzi CL, 2010. A placebo-controlled, double-blind trial of infliximab for cancer-associated weight loss in elderly and/or poor performance non-small cell lung cancer patients (N01C9). Lung Cancer 68, 234–239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kir S, White JP, Kleiner S, Kazak L, Cohen P, Baracos VE, Spiegelman BM, 2014. Tumour-derived PTH-related protein triggers adipose tissue browning and cancer cachexia. Nature 513, 100–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lee C, Raffaghello L, Brandhorst S, Safdie FM, Bianchi G, Martin-Montalvo A, Pistoia V, Wei M, Hwang S, Merlino A, Emionite L, de Cabo R, Longo VD, 2012. Fasting cycles retard growth of tumors and sensitize a range of cancer cell types to chemotherapy. Sci Transl Med 4, 124ra127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lee JW, Stone ML, Porrett PM, Thomas SK, Komar CA, Li JH, Delman D, Graham K, Gladney WL, Hua X, Black TA, Chien AL, Majmundar KS, Thompson JC, Yee SS, O’Hara MH, Aggarwal C, Xin D, Shaked A, Gao M, Liu D, Borad MJ, Ramanathan RK, Carpenter EL, Ji A, de Beer MC, de Beer FC, Webb NR, Beatty GL, 2019. Hepatocytes direct the formation of a pro-metastatic niche in the liver. Nature 567, 249–252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lutgendorf SK, Weinrib AZ, Penedo F, Russell D, DeGeest K, Costanzo ES, Henderson PJ, Sephton SE, Rohleder N, Lucci JA 3rd, Cole S, Sood AK, Lubaroff DM, 2008. Interleukin-6, cortisol, and depressive symptoms in ovarian cancer patients. J Clin Oncol 26, 4820–4827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mason KA, Ariga H, Neal R, Valdecanas D, Hunter N, Krieg AM, Whisnant JK, Milas L, 2005. Targeting toll-like receptor 9 with CpG oligodeoxynucleotides enhances tumor response to fractionated radiotherapy. Clin Cancer Res 11, 361–369. [PubMed] [Google Scholar]
  30. Masri S, Papagiannakopoulos T, Kinouchi K, Liu Y, Cervantes M, Baldi P, Jacks T, Sassone-Corsi P, 2016. Lung Adenocarcinoma Distally Rewires Hepatic Circadian Homeostasis. Cell 165, 896–909. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Matsui T, Omuro H, Liu YF, Soya M, Shima T, McEwen BS, Soya H, 2017. Astrocytic glycogen-derived lactate fuels the brain during exhaustive exercise to maintain endurance capacity. Proc Natl Acad Sci U S A 114, 6358–6363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mayers JR, Torrence ME, Danai LV, Papagiannakopoulos T, Davidson SM, Bauer MR, Lau AN, Ji BW, Dixit PD, Hosios AM, Muir A, Chin CR, Freinkman E, Jacks T, Wolpin BM, Vitkup D, Vander Heiden MG, 2016. Tissue of origin dictates branched-chain amino acid metabolism in mutant Kras-driven cancers. Science 353, 1161–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mermod M, Hiou-Feige A, Bovay E, Roh V, Sponarova J, Bongiovanni M, Vermeer DW, Lee JH, Petrova TV, Rivals J-P, Monnier Y, Tolstonog GV, Simon C, 2018. Mouse model of postsurgical primary tumor recurrence and regional lymph node metastasis progression in HPV-related head and neck cancer. International Journal of Cancer 142, 2518–2528. [DOI] [PubMed] [Google Scholar]
  34. Meyers CA, Albitar M, Estey E, 2005. Cognitive impairment, fatigue, and cytokine levels in patients with acute myelogenous leukemia or myelodysplastic syndrome. Cancer 104, 788–793. [DOI] [PubMed] [Google Scholar]
  35. Miller AH, Ancoli-Israel S, Bower JE, Capuron L, Irwin MR, 2008. Neuroendocrine-immune mechanisms of behavioral comorbidities in patients with cancer. J Clin Oncol 26, 971–982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mock V, Atkinson A, Barsevick A, Cella D, Cimprich B, Cleeland C, Donnelly J, Eisenberger MA, Escalante C, Hinds P, Jacobsen PB, Kaldor P, Knight SJ, Peterman A, Piper BF, Rugo H, Sabbatini P, Stahl C, 2000. NCCN Practice Guidelines for Cancer-Related Fatigue. Oncology (Williston Park) 14, 151–161. [PubMed] [Google Scholar]
  37. Norden DM, Bicer S, Clark Y, Jing R, Henry CJ, Wold LE, Reiser PJ, Godbout JP, McCarthy DO, 2015a. Tumor growth increases neuroinflammation, fatigue and depressive-like behavior prior to alterations in muscle function. Brain Behav Immun 43, 76–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Norden DM, McCarthy DO, Bicer S, Devine RD, Reiser PJ, Godbout JP, Wold LE, 2015b. Ibuprofen ameliorates fatigue- and depressive-like behavior in tumor-bearing mice. Life Sci 143, 65–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Pedersen L, Idorn M, Olofsson GH, Lauenborg B, Nookaew I, Hansen RH, Johannesen HH, Becker JC, Pedersen KS, Dethlefsen C, Nielsen J, Gehl J, Pedersen BK, Thor Straten P, Hojman P, 2016. Voluntary Running Suppresses Tumor Growth through Epinephrine- and IL-6-Dependent NK Cell Mobilization and Redistribution. Cell Metab 23, 554–562. [DOI] [PubMed] [Google Scholar]
  40. Pertl MM, Hevey D, Boyle NT, Hughes MM, Collier S, O’Dwyer AM, Harkin A, Kennedy MJ, Connor TJ, 2013. C-reactive protein predicts fatigue independently of depression in breast cancer patients prior to chemotherapy. Brain Behav Immun 34, 108–119. [DOI] [PubMed] [Google Scholar]
  41. Quistorff B, Secher NH, Van Lieshout JJ, 2008. Lactate fuels the human brain during exercise. Faseb j 22, 3443–3449. [DOI] [PubMed] [Google Scholar]
  42. Rusch HP, Kline BE, 1944. The effect of exercise on the growth of a mouse tumor. Cancer Res 4, 116–118. [Google Scholar]
  43. Spanos WC, Hoover A, Harris GF, Wu S, Strand GL, Anderson ME, Klingelhutz AJ, Hendriks W, Bossler AD, Lee JH, 2008. The PDZ binding motif of human papillomavirus type 16 E6 induces PTPN13 loss, which allows anchorage-independent growth and synergizes with ras for invasive growth. J Virol 82, 2493–2500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Spanos WC, Nowicki P, Lee DW, Hoover A, Hostager B, Gupta A, Anderson ME, Lee JH, 2009. Immune response during therapy with cisplatin or radiation for human papillomavirus-related head and neck cancer. Arch Otolaryngol Head Neck Surg 135, 1137–1146. [DOI] [PubMed] [Google Scholar]
  45. Tyring S, Gottlieb A, Papp K, Gordon K, Leonardi C, Wang A, Lalla D, Woolley M, Jahreis A, Zitnik R, Cella D, Krishnan R, 2006. Etanercept and clinical outcomes, fatigue, and depression in psoriasis: double-blind placebo-controlled randomised phase III trial. Lancet 367, 29–35. [DOI] [PubMed] [Google Scholar]
  46. Vander Heiden MG, Cantley LC, Thompson CB, 2009. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Venteclef N, Jakobsson T, Steffensen KR, Treuter E, 2011. Metabolic nuclear receptor signaling and the inflammatory acute phase response. Trends Endocrinol Metab 22, 333–343. [DOI] [PubMed] [Google Scholar]
  48. Vichaya EG, Vermeer DW, Christian DL, Molkentine JM, Mason KA, Lee JH, Dantzer R, 2017. Neuroimmune mechanisms of behavioral alterations in a syngeneic murine model of human papilloma virus-related head and neck cancer. Psychoneuroendocrinology 79, 59–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Walker M, Mason G, 2018. A comparison of two types of running wheel in terms of mouse preference, health, and welfare. Physiol Behav 191, 82–90. [DOI] [PubMed] [Google Scholar]
  50. Wang XS, Williams LA, Krishnan S, Liao Z, Liu P, Mao L, Shi Q, Mobley GM, Woodruff JF, Cleeland CS, 2012. Serum sTNF-R1, IL-6, and the development of fatigue in patients with gastrointestinal cancer undergoing chemoradiation therapy. Brain Behav Immun 26, 699–705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Warburg O, Posener E, 1924. Negelein: Ueber den Stoffwechsel der Tumoren. Biochemische Zeitschrift 152, 319–344. [Google Scholar]
  52. Wratten C, Kilmurray J, Nash S, Seldon M, Hamilton CS, O’Brien PC, Denham JW, 2004. Fatigue during breast radiotherapy and its relationship to biological factors. Int J Radiat Oncol Biol Phys 59, 160–167. [DOI] [PubMed] [Google Scholar]
  53. Wunderlich FT, Strohle P, Konner AC, Gruber S, Tovar S, Bronneke HS, Juntti-Berggren L, Li LS, van Rooijen N, Libert C, Berggren PO, Bruning JC, 2010. Interleukin-6 signaling in liver-parenchymal cells suppresses hepatic inflammation and improves systemic insulin action. Cell Metab 12, 237–249. [DOI] [PubMed] [Google Scholar]
  54. Yang W, Zheng Y, Xia Y, Ji H, Chen X, Guo F, Lyssiotis CA, Aldape K, Cantley LC, Lu Z, 2012. ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect. Nat Cell Biol 14, 1295–1304. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1596669-supplement-1.docx (898.4KB, docx)

RESOURCES