Muscle atrophy in experimental cancer cachexia: Is the IGF-1 signaling pathway involved?



Skeletal muscle wasting, one of the main features of cancer cachexia, is associated with marked protein hypercatabolism, and has suggested to depend also on impaired IGF-1 signal transduction pathway. To investigate this point, the state of activation of the IGF-1 system has been evaluated both in rats bearing the AH-130 hepatoma and in mice transplanted with the C26 colon adenocarcinoma. In the skeletal muscle of tumor hosts, the levels of phosphorylated (active) Akt, one of the most relevant kinases involved in the IGF-1 signaling pathway, were comparable to controls, or even increased. Accordingly, downstream targets such as GSK3β, p70S6K and FoxO1 were hyperphosphorylated, while the levels of phosphorylated eIF2α were markedly reduced with respect to controls. In the attempt to force the metabolic balance toward anabolism, IGF-1 was hyperexpressed by gene transfer in the tibialis muscle of the C26 hosts. In healthy animals, IGF-1 overexpression markedly increased both fiber and muscle size. As a positive control, IGF-1 was also overexpressed in the muscle of aged mice. In IGF-1 hyperexpressing muscles the fiber cross-sectional area definitely increased in both young and aged animals, while, by contrast, loss of muscle mass or reduction of fiber size in mice bearing the C26 tumor were not modified. These results demonstrate that muscle wasting in tumor-bearing animals is not associated with downregulation of molecules involved in the anabolic response, and appears inconsistent, at least, with reduced activity of the IGF-1 signaling pathway.

Cancer cachexia is a syndrome characterized by loss of body weight, skeletal muscle proteins, depletion of lipid stores, weakness, and perturbations of the hormonal homeostasis.1, 2 According to a definition recently emerged, “cachexia is a complex metabolic syndrome associated with underlying illness and characterized by loss of muscle with or without loss of fat mass… …. Anorexia, inflammation, insulin resistance and increased muscle protein breakdown are frequently associated with wasting disease… .” (quoted from Ref.3). Anorexia is frequently underdiagnosed and may significantly contribute to the nutritional deterioration of cachexia.4 Finally, muscle loss is shared by several conditions such as starvation, malnutrition, ageing, bed rest and physical inactivity. In this regard, cachexia should be distinguished from other forms of muscle depletion, such as sarcopenia of ageing.

Cachectic patients show increased morbidity and mortality rates, benefit less from antineoplastic therapies, and have a poor quality of life. Since basically refractory to both nutritional and pharmacological interventions, cachexia is a major problem in clinical oncology and is responsible for about 25% of cancer deaths.5 In this regard, understanding the mechanisms involved in cachexia to design appropriate therapeutic approaches has important clinical implications.

The skeletal muscle, which accounts for approximately half of total body protein, is severely affected in cancer cachexia, and its negative protein balance mainly results from hypercatabolism.6 A complex interplay among hormones, cytokines and other humoral factors is crucially involved in the pathogenesis of muscle wasting. In this regard, catabolic states such as sepsis and cancer have been associated with reduced levels of circulating IGF-17–9 as well as with increased concentrations of proinflammatory cytokines, such as TNFα or IFNγ.10

In the last years, the role played by IGF-1 in the regulation of skeletal muscle mass has been investigated with particular attention. IGF-1 is an anabolic growth factor that stimulates muscle protein synthesis as well as proliferation and differentiation of satellite cells.11 It exerts antiapoptotic effects on muscle cells,12 suppresses proteolysis and inhibits the ubiquitin-proteasome system.13, 14 IGF-1 overexpression prevents the skeletal muscle atrophy induced in rats by disuse or glucocorticoid treatment, promotes the regenerative response in aging and in a murine model of muscle injury, and improves muscle damage in dystrophic mdx mice.1, 15–19

The effects exerted by IGF-1 on muscles mainly result from stimulation of the phosphatidylinositol-3 kinase (PI3K)/Akt pathway, leading to activation of downstream targets required for protein synthesis.20, 21 In the last years, however, several reports have proposed that signaling through the PI3K/Akt pathway also inhibits the expression of the muscle-specific ubiquitin ligases atrogin-1 and MuRF1 by inactivating the FoxO transcription factors.22, 23 Accordingly, FoxO-1 silencing has been shown to prevent protein degradation in a murine model of cancer cachexia.24

Perturbations of the IGF-1 signaling pathway have been reported in both in vitro and in vivo models of muscle atrophy.25 Indeed, the levels of active Akt are significantly reduced in C2C12 myotubes undergoing atrophy following nutrient deprivation or glucocorticoid treatment.22 Consistently, decreased activity of the PI3K/Akt pathway has been shown to occur in muscle wasting induced by denervation,26 disuse,27 aging,28 or glucocorticoid treatment,29 while no information is presently available about the state of activation of the IGF-1 signaling pathway in cancer cachexia.

Previous results obtained in our laboratory have shown that muscle wasting in rats bearing the Yoshida AH-130 hepatoma is associated with reduced IGF-1 expression, both in the liver and in the skeletal muscle, suggesting that impaired IGF-1 signaling may be also involved in the pathogenesis of muscle depletion in cancer bearers.8 In this regard, the aim of our study has been to investigate the IGF-1 signal transduction pathway in the muscle of animals transplanted with 2 different cachexia-inducing tumors, the Yoshida AH-130 hepatoma in rats and the Colon26 adenocarcinoma in mice. In addition, since previous observations have shown that exogenous IGF-1 is unable to correct muscle atrophy in AH-130 bearing rats,8 the effectiveness of local IGF-1 overexpression in tumor bearers has been tested and compared to that observed in aged animals.

Material and Methods

Animals and experimental design

Male Wistar rats weighing about 150 g and Balb-c mice weighing about 20 g were obtained from Charles River Laboratories (Calco, LC, Italy). For the senescence experiment, young adult (3 months, n = 5) and aged (22 months, n = 5) Balb-c mice were used. Animals were maintained on a regular dark–light cycle (light from 8.00 a.m. to 8.00 p.m.), with free access to food (Piccioni, Brescia, Italy) and water during the whole experimental period, including the night before sacrifice. They were cared for in compliance with the Italian Ministry of Health Guidelines (n° 86609 EEC) and the Policy on Humane Care and Use of Laboratory Animals.30 The experimental protocol has been approved by the Bioethical Committee of the University of Torino.

Tumor-bearing rats (n = 8 for each group) received an intraperitoneal inoculum of Yoshida AH-130 ascites hepatoma cells (˜108 cells/rat), whereas tumor-bearing mice (n = 8 for each group) were inoculated s.c. dorsally with 5 × 105 C26 adenocarcinoma cells. Healthy animals served as controls (n = 6 for each group, for both rats and mice). Animal weight and food intake were recorded daily. Tumor-bearing rats and mice were sacrificed under anesthesia 7 and 12 days after tumor transplantation, respectively. Several muscles were rapidly excised, weighed, frozen in isopentane cooled with liquid nitrogen, and stored at −80°C.

Plasmids and electroporation

An expression vector encoding IGF1-Ea (the muscle isoform of IGF-1 containing the Ea-peptide) was used (MGC 105288, American Type Culture Collection, Rockville, MD). Transfection efficiency has been assessed by a plasmid encoding EGFP (Clontech, Palo Alto, CA). Plasmids were purified with a NucleoBond Xtra Maxi kit (Macherey-Nagel GmbH, Duren, Germany).

The left tibialis anterior muscle was injected with 25 μl of 0.5 U/μl hyaluronidase (to improve transfection efficiency; see 28) and 2 hr later injected with 50 μg of plasmidic DNA, whereas the contralateral muscle served as control. One minute after DNA injection, transcutaneous pulses were applied by 2 stainless steel plate electrodes (gap between plates: 4 mm). Electrical contact with the leg skin was ensured by shaving each leg and applying conductive gel. Electric pulses with a standard square wave were delivered by an electroporator (ECM-830, BTX-Harvard Apparatus, Holliston, MA). Three pulses (20 msec each) of 75 V/cm were administered to the muscle with a delivery rate of 1 pulse/sec. The polarity was then reversed and a further 3 pulses were delivered to the muscle. The electroporation was performed 10 days before animal sacrifice. With the transfection procedure described, no sign of muscle damage and inflammatory infiltrate could be seen on histological analysis (see Supporting Information Figure S1). By contrast, muscle damage and consequent regeneration were evident when electroporation was performed at 150 V/cm (see Supporting Information Figure S1).

Reverse transcription-PCR

Total RNA extraction and RT-PCR were conducted as previously described.8 Briefly, RNA was obtained using the TriPure reagent (Roche, Indianapolis, IN) following the instructions provided by the manufacturer. IGF-1 mRNA levels were determined by semiquantitative reverse-transcription polymerase chain reaction using the kit “Ready-to-Go RT-PCR Beads” (GE Healthcare, Milano, Italy). Total RNA (0.5 μg) and 400 nM mixture of each couple of primers were added to a RT-PCR reaction mixture. PCR products were electrophoresed on 2% agarose gels and visualized with ethidium bromide. A 100 bp-standard DNA ladder (Fermentas, Burlington, ON, Canada) was used to estimate the length of each PCR product. Quantification was performed by densitometric analysis. The results were normalized according to 18S rRNA expression. Comparisons among groups were made in the linear phase of amplification.

Western blotting

About 50 mg of muscle were homogenized in 80 mM TRIS-HCl, pH 6.8 (containing 100 mM DTT, 70 mM SDS, and 1 mM glycerol) with freshly added protease and phosphatase inhibitor cocktails (Sigma, St. Louis, MO), kept on ice for 30 min, centrifuged at 15,000g for 10 min at 4°C, and the supernatant collected. Protein concentration was assayed by the method of Lowry using BSA as working standard. Equal amounts of protein (30 μg) were heat-denaturated in sample-loading buffer (50 mM TRIS-HCl, pH 6.8, 100 mM DTT, 2% SDS, 0.1% bromophenol blue, 10% glycerol), resolved on a SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA). The filters were then blocked with Tris-buffered saline (TBS) containing 0.05% Tween and 5% nonfat dry milk and incubated overnight with the following antibodies directed against: p-Akt (Ser473 #9271 and Thr308 #9275), p-FOXO1 (Ser256 #9461), p-GSK3β (Ser9 #9336), p-p70S6K (Thr389 #9205), and p-eIF2α (Ser51 #9721) from Cell Signaling Technology (Danvers, MA); p-IRS-1 (Tyr989 sc-17200), IRS-1 (sc-559), Akt (sc-8312), p-PTEN (Ser380 sc-31714), PP2A (sc-6110), and GFP (sc-9996) from Santa Cruz Biotechnology (Santa Cruz, CA), p-AMPKα (Thr172), AMPK-pan from Millipore (Temecula, CA) or Tubulin (T5168, Sigma, St. Louis, MO). Peroxidase-conjugated IgG (Bio-Rad, Hercules, CA) were used as secondary antibodies. The membrane-bound immune complexes were detected by an enhanced chemiluminescence system (Santa Cruz Biotechnology) on a photon-sensitive film (Hyperfilm ECL; GE Healthcare, Milano, Italy). Protein loading was normalized according to tubulin expression. Quantification of the bands was performed by densitometric analysis using a specific software (TotalLab, NonLinear Dynamics, Newcastle upon Tyne, UK).

Electrophoretic mobility shift assay

To prepare nuclear extracts, the gastrocnemius (50 mg) was homogenized in ice-cold 10 mM HEPES, pH 7.5, containing 10 mM MgCl2, 5mM KCl, 0.1 mM EDTA, pH 8.0, 0.1% Triton X-100, 0.1 mM phenylmethanesulfonyl fluoride (PMSF), 1 mM DTT, 2 μg/ml aprotinin, 2 μg/ml leupeptin. Samples were then centrifuged (5 min, 3000g), pellets resuspended in ice-cold 20 mM HEPES, pH 7.9, containing 25% glycerol, 500 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, pH 8.0, 0.2 mM PMSF, 0.5 mM DTT, 2 μg/ml aprotinin, 2 μg/ml leupeptin and incubated on ice for 30 min. Cell debris were removed by centrifugation (5 min, 3000g) and the supernatant collected and stored at −80°C. Oligonucleotide labeling and binding reactions were performed by using the reagent supplied in the Gel Shift Assay System (Promega, Milano, Italy). Binding reaction mixtures, containing nuclear proteins (10 μg) and Gel Shift Binding Buffer (10 mM TRIS-HCl, pH 7.5, containing 1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 0.05 μg/μl poly(dI-dC)·poly(dI-dC) and 4% glycerol), were incubated at room temperature for 10 min in the presence of 0.035 pmol 32P-ATP end-labeled double-stranded oligonucleotide (5′-ACTAGTCACAGTTGTTTACTTATAG-3′). At the end of the incubation, samples were electrophoresed in 0.5× Tris-borate-EDTA (TBE) buffer at 350 V for 40 min on a 4% nondenaturing acrylamide gel. The gel was dried for 45 min and exposed overnight or longer to a film for autoradiography (GE Healthcare, Milano, Italy) at –80°C with intensifying screens. Specificity of the bands was confirmed by adding an excess amount of unlabeled oligonucleotide (1.75 pmol) to a control sample. HeLa cell nuclear extract was used as positive control (Promega).


Tibialis anterior muscles were excised, weighed, mounted in OCT (VWR, Milan, Italy) and frozen in melting isopentane cooled in liquid nitrogen. Ten micrometers of transverse sections from the midbelly region were cut on a cryostat and later stained with hematoxylin and eosin (H&E). Fiber-type assessment by myosin ATPase staining31 showed that the tibialis muscle is composed almost exclusively by fast fibers (Supporting Information Figure S1). Fiber cross-sectional area (CSA) was determined on 100 individual fibers (irrespective of GFP expression) by the Image J software,32 and expressed in pixels. Differences in absolute values are due to changes in photo magnification and/or resolution. Both variables, however, were maintained fixed within each experiment.

Data analysis and presentation

All results were expressed as mean ± SD (figures are also reported in Supporting Information showing median and lower/upper quartiles). Normal distribution of data has been ascertained by the Kolmogorov–Smirnov test. Significance of the differences was evaluated by analysis of variance followed by Tukey's test.


Tumor growth negatively affected body weight: as expected the increase in body weight was lower than in controls in the AH-130 bearing rats (Fig. 1a), while significant weight loss occurred in the C26 bearing mice (Fig. 1b). At the time of sacrifice, the skeletal muscle was severely atrophic in tumor bearers. In particular, the reduction in size of the gastrocnemius and tibialis anterior muscles was about 20% in the AH-130 bearers (7 days), and ∼30% of control values in the C26 hosts (12 days; Figs. 1c and 1d). Previous results have shown that insulin and IGF-1 plasma levels were significantly reduced in the AH-130 hosts.8, 33, 34 Cumulative food intake markedly decreased with respect to controls in both the AH-130 (C: 127.1 ± 5.7 g; AH-130: 87.5 ± 4.8 g; p = 0.0048) and the C26 bearers (C: 40.4 ± 2.1 g; C26: 28.6 ± 2.9 g; p = 0.01).

Figure 1.

Loss of body weight (a, b) and muscle mass (c, d) in AH-130 and C26 hosts. Muscle mass is expressed as mg/100 g (AH-130) or mg/10 g (C26) of initial body weight (i.b.w.). GSN = gastrocnemius. Data are means ± SD, n = 6 (C: controls) or 8 (tumor bearers). Significance of the differences versus controls: *p < 0.05, **p < 0.01, ***p < 0.001.

IGF-1 signaling pathway

Activation of the insulin receptor substrate (IRS)-1 by Tyr989-phosphorylation is the first event following IGF-1 receptor engagement. The levels of Tyr-phosphorylated IRS-1 significantly decreased in the gastrocnemius muscle of the AH-130 hosts (Fig. 2a). However, not only the degree of phosphorylation (activation) of Akt, the kinase downstream of IRS-1, on Thr308 was comparable to controls, but its phosphorylation on Ser473 was increased. Besides PI3K, also the activity of phosphatases such as PTEN35 or PP2A36 is relevant to phosphorylated Akt levels. The expression of both phosphorylated PTEN and the active Cα subunit of PP2A were comparable in the gastrocnemius of both AH-130 hosts and controls (Fig. 2a). To investigate if the maintenance of the IGF-1-dependent signaling is a common feature in experimental cachexia, a few relevant points were also investigated on mice bearing the C26 tumor. Results in Figure 2b show that, as in the AH-130 hosts,8 IGF-1 mRNA levels were markedly reduced in the gastrocnemius of the C26 bearers, while the expression of the muscle-specific ubiquitin ligase atrogin-1, an accepted indicator of the occurrence of protein hypercatabolism,22 was significantly increased with respect to controls. The observation that in the skeletal muscle of the C26 hosts Akt phosphorylation on both Thr308 and Ser473 was comparable to controls (Fig. 2c) demonstrates that maintenance of active Akt levels is not exclusive of the AH-130 model system. The lack of downregulation of Akt phosphorylation, and in general of the IGF-1 signaling pathway, also stands for muscles different from the gastrocnemius such as the tibialis anterior (AH-130 and C26 hosts), the EDL and the soleus (AH-130 hosts only; see Supporting Information Figures S2–S3)

Figure 2.

Akt activation is not compromised in the skeletal muscle of tumor-bearing animals. (a) Representative western blot patterns and densitometric analysis for indicated signaling proteins involved in Akt activation in the gastrocnemius of AH-130 hosts; protein loading was normalized according to tubulin expression. Data (means ± SD) are expressed as percentage of controls (in arbitrary units (a.u.): p-IRS: 0.70 ± 0.15; p-Akt(Thr): 0.96 ± 0.14; p-Akt(Ser): 1.03 ± 0.27; p-PTEN: 1.29 ± 0.32; PP2A: 0.27 ± 0.03); n = 6 (C) or 8 (tumor bearers). Significance of the differences versus controls: *p < 0.05. (b) Representative PCR pattern and densitometric analysis of IGF-1 and atrogin-1 gene expression in the gastrocnemius muscle of mice bearing the C26 tumor. mRNA levels were normalized according to 18S expression. Data (means ± SD) are expressed as percentage of controls (IGF-1: 4.24 ± 1.47 a.u.; atrogin-1: 5.02 ± 0.34 a.u.); n = 6 (C) or 8 (tumor bearers). Significance of the differences versus controls: *p < 0.05; **p < 0.01. (c) Representative western blot pattern (3 different animals) and densitometric analysis of Akt (total and phosphorylated) in the gastrocnemius of C26 bearing mice. Data (means ± SD) are expressed as percentage of controls (p-Akt(Thr): 0.20 ± 0.05 a.u.; p-Akt(Ser): 0.40 ± 0.13 a.u.); n = 6 (C) or 8 (tumor bearers).

FOXO transcription factors

A number of reports have proposed that muscle protein wasting is associated with dephosphorylation (activation) of FoxO transcription factors owing to downregulation of the PI3K/Akt signaling pathway.22, 23 At variance, our study shows that FoxO1 activation was reduced in the skeletal muscle of tumor-bearing animals, since its degree of phosphorylation on Ser256 was increased in both AH-130 and C26 hosts (Figs. 3a and 3b) and its nuclear translocation and DNA binding activity was decreased in the gastrocnemius of rats bearing the AH-130 hepatoma (Fig. 3c).

Figure 3.

FoxO1 activity is impaired in the muscle of tumor bearing animals. Representative western blots (3 different animals) and densitometric analysis of phosphorylated FoxO1 in the gastrocnemius of AH-130 (a) and C26 hosts (b). Data (means ± SD) are expressed as percentage of controls [(a) 0.42 ± 0.06 a.u.; (b) 0.52 ± 0.19 a.u.]; n = 6 (C) or 8 (tumor bearers). Significance of the differences versus controls: **p < 0.01. (c) Representative pattern (3 different animals) and densitometric analysis of FoxO DNA-binding activity in the gastrocnemius of AH-130 bearing rats. Data (means ± SD) are expressed as percentage of controls (0.24 ± 0.06 a.u.); n = 6 (C) or 8 (tumor bearers); *positive control (HeLa cells), §control muscle incubated in the presence of an excess amount of cold oligonucleotide (specific competition). Significance of the differences versus controls: *p < 0.05.

Anabolic pathways downstream of Akt

Several molecules downstream of the IGF-1/PI3K/Akt pathway are involved in the regulation of protein synthesis.37 Their state of activation in the skeletal muscle of tumor-bearing animals was evaluated in our study. In spite of the marked muscle atrophy, the inactivating phosphorylation of GSK3β was increased in both the AH-130 and C26 hosts (Fig. 4). The levels of phosphorylated p70S6K in the C26 and in the AH-130 bearers were unaffected or even increased, respectively, while the phosphorylation of eIF2α was markedly decreased.

Figure 4.

Muscle wasting in experimental cancer cachexia is associated with upregulation of anabolic signals. Representative western blots (3 different animals) and densitometric analysis of phosphorylated GSK3β, p70S6K and eIF2α in the gastrocnemius of AH-130 (a) and C26 hosts (b). Data (means ± SD) are expressed as percentage of controls [(a) p-GSK3: 0.21 ± 0.03 a.u.; p-p70: 0.05 ± 0.01 a.u.; p-EIF2α: 0.50 ± 0.09 a.u.; (b) p-GSK3: 1.03 ± 0.60 a.u.; p-p70: 0.77 ± 0.17 a.u.; p-EIF2α: 0.89 ± 0.26 a.u.]; n = 6 (C) or 8 (tumor bearers). Significance of the differences versus controls: *p < 0.05, **p < 0.01, ***p < 0.001.

IGF-1 overexpression

Since the impairment of IGF-1 signaling pathway was considered an early event in the induction of muscle atrophy,22 IGF-1 was hyperexpressed in the tibialis muscle of the C26 hosts 2 days after tumor injection, in the attempt to force further the metabolic balance toward anabolism. As expected, IGF-1 overexpression in control mice induced fiber and muscle hypertrophy (Figs. 5a and 5b) associated with increased levels of phosphorylated Akt (Figs. 5a5c). The pattern of GFP-positive fibers, broadly indicating transfection efficiency, is shown in Figure 5c.38, 39 Of interest, IGF-1 electroporation could not modify either the loss of muscle mass (data not shown) or the reduction of fiber CSA in mice bearing the C26 tumor (Fig. 5d). Since previous reports demonstrated that IGF-1 transgenic mice were protected against aging-related sarcopenia,17 we compared the effect of IGF-1 overexpression in tumor bearers and aged animals. In the latter, muscle wasting was associated with reduced phosphorylation of both GSK3β and p70S6K, without changes in the levels of phosphorylated Akt (Fig. 6b). Differently from tumor-bearing animals, IGF-1 overexpression by gene transfer was able to increase fiber CSA in the tibialis of senescent mice (Fig. 6c).

Figure 5.

IGF-1 induces muscle hypertrophy in controls but not in cachectic mice. (a) H&E staining of tibialis transverse sections of control mice overexpressing IGF-1 or expressing green fluorescent protein (GFP) as transfection control; scale bar = 50 μm. Representative western blot pattern of GFP and Akt (phosphorylated and total) in the tibialis of control mice overexpressing IGF-1 or expressing GFP. (b) Tibialis mass and myofiber mean cross-sectional area (CSA) in mice transfected with either IGF-1 or GFP (expressed as percentage of GFP-transfected; mass: 19.0 ± 0.85; CSA: 5206 ± 952, see “Material and Methods” section for details). (c) Densitometric analysis of phosphorylated Akt expression in healthy mice transfected with GFP (GFP-T) or IGF-1 (IGF-1-T). Data are expressed as percentage of GFP-T (0.26 ± 0.21 a.u.). Significance of the differences versus GFP-T: *p < 0.05 (2 independent experiments, n = 4 for each group/experiment). Inset shows GFP expression by immunofluorescence microscopy (scale bar = 100 μm). (d) Myofiber mean CSA (expressed as percentage of controls; NT: 1700 ± 343; IGF-1: 2340 ± 463) in C26 bearing mice either transfected with IGF-1 (IGF-1) or not (NT). Significance of the differences versus NT controls: *p < 0.05, **p < 0.01, ***p < 0.001; n = 6 (C) or 8 (C26).

Figure 6.

IGF-1 prevents skeletal muscle atrophy in senescent mice. (a) Muscle mass of young (3 months; Mo) and aged (22 Mo) animals (expressed as mg/10 g initial body weight). Significance of the differences versus 3 Mo: ***p < 0.001, n = 5. (b) Representative western blots (3 different animals) and densitometric analysis of phosphorylated Akt, GSK3β and p70S6K in the gastrocnemius of young and aged mice. Data (means ± SD) are expressed as percentage of young mice (p-Akt: 0.83 ± 0.21; p-GSK3: 0.33 ± 0.04; p-p70: 0.63 ± 0.21); n = 5. Significance of the differences versus 3 Mo: *p < 0.05. (c) Tibialis myofiber mean CSA (expressed as percentage of untransfected young mice; 3 Mo: 4296 ± 213; 22 Mo: 4921 ± 985) in young and aged mice either transfected with IGF-1 (IGF-1) or not (NT). Significance of the differences: *p < 0.05 versus NT young mice; $p < 0.05 versus NT aged mice; n = 5 for each group.

Phosphorylation of 5′-adenosine monophosphate-activated protein kinase

To test other pathways possibly involved in the pathogenesis of muscle atrophy in tumor-bearing animals, the levels of phosphorylated adenosine monophosphate-activated protein kinase (AMPK) were analyzed. The data reported in Figure 7 show that in the skeletal muscle of both the AH-130 bearers and the C26 hosts p-AMPK levels tended to increase with respect to controls.40

Figure 7.

Levels of phosphorylated AMPK tend to increase in the skeletal muscle of tumor-bearing animals. Representative western blots (3 different animals) and densitometric analysis of phosphorylated AMPK in the gastrocnemius of AH-130 (a) and C26 hosts (b). Data (means ± SD) are expressed as percentage of controls [(a) 0.21 ± 0.10 a.u.; (b) 0.35 ± 0.20 a.u.]; n = 6 (C) or 8 (tumor bearers).


Several reports suggest that, in association with enhanced protein degradation, downregulation of anabolic pathways, such as those regulated by IGF-1, plays a major role in the pathogenesis of skeletal muscle atrophy in different catabolic states,41 including cancer cachexia. In this regard, insulin resistance has been recently involved in the pathogenesis of cancer cachexia.42 The results shown in this article, however, demonstrate that, in spite of the marked muscle wasting, at least some key factors involved in the regulation of protein synthesis are poised in an anabolic state in tumor-bearing animals.

In the IGF-1 signaling pathway, Akt plays a central role in the integration of both anabolic and catabolic responses by modulating the phosphorylation of its numerous substrates.43 The present observation that the phosphorylated state of Akt is maintained in the skeletal muscle of tumor-bearing animals does not conform to the usual pattern observed in other experimental models of muscle atrophy. The difference is particularly striking with the results reported in patients with pancreatic cancer44 or in tumor-bearing mice.42 The discrepancies between the latter and our study may be due to the markedly different experimental model (C26 tumor implanted in CD2F1 mice) and the working conditions used. Currently, we have no simple explanation for the maintenance of p-Akt levels observed in tumor bearers, particularly considering the reduced IGF-1 and insulin levels in these animals.8 In addition, previous results have demonstrated that muscle wasting can be prevented by administration of insulin, though not by IGF-1.8, 34 The phosphorylation status of Akt depends on PI3K activity as well as on the action of phosphatases such as PTEN and PP2A. In this regard, Hu et al.45 suggested that decreased levels of active PTEN can be a compensatory mechanism to prevent Akt downregulation and, consequently, muscle protein loss in fasted mice. However, no changes in the expression of both phosphatases could be observed in the skeletal muscle of tumor-bearing animals, suggesting, although quite indirectly, that their activity remains close to control levels.

The results shown in our study indicate that tumor-induced muscle atrophy does not conform to the view that links univocally the increased expression of the ubiquitin ligase atrogin-1 to downregulation of the PI3K/Akt pathway.22, 23 Besides, our data show that, at least in the 2 experimental models of cancer cachexia taken into consideration, increased expression of the muscle-specific ubiquitin ligases8, 46, 47 is not associated with downregulation of the IGF-1 signaling pathway, suggesting the lack of a causal relation among low IGF-1, enhanced phosphorylated FoxO, increased ubiquitin ligases and muscle wasting. In this regard, our observations are not unique, however. Indeed, the reduction of protein degradation afforded by GSK-3β inhibitors in muscles isolated from burned rats has been reported to occur without changes in atrogin-1 mRNA levels.48 Moreover, a recent report shows that muscle atrophy in transgenic mice hyperexpressing Tpr-Met is associated with upregulation of both ubiquitin ligase expression and Akt phosphorylation.49 Finally, recent observations have proposed that high levels of atrogin-1 can be detected in hypertrophic cardiomyocytes, irrespective of Akt phosphorylation levels.50 Akt inactivation and FoxO nuclear translocation appear to be early events in the induction of muscle atrophy.22 However, in the AH-130 hosts phosphorylated Akt remained close to control values at 2, 4 (data not shown), and 7 days after tumor implantation. These observations not only suggest that muscle wasting in cancer cachexia depends just marginally on Akt downregulation, but also render unlikely that the normal levels of phosphorylated Akt in day 7 tumor hosts result from activation of unknown negative feedback loops. Since Akt is regulated by several feedback loops,51 the state of activation of downstream molecules, such as FoxO, may also change in the absence of detectable modulations of phosphorylated Akt. In this regard, is worth noting that, in 22 months old mice, p70 and GSK-3β phosphorylation is significantly reduced despite phosphorylated Akt levels remain comparable to those in 3 months animals. Finally, the possibility that FoxO phosphorylation may also depend on activation/inactivation of signaling pathways other than the one regulated by PI3K/Akt cannot be discarded: increased Akt phosphorylation in the presence of unchanged FoxO phosphorylation was recently reported in C2C12 myotube cultures exposed to IL-1.52

Besides the PI3K/Akt/FoxO pathway, other factors could trigger the transcription of muscle-specific ubiquitin ligases. For example, the increased expression of atrogin-1 and MuRF-1 induced by TNF-α in C2C12 myotubes is prevented by p38 inhibitors.53 Similarly, atrogin-1 upregulation and muscle mass depletion induced by lipopolysaccharide (LPS) in mice can be prevented by curcumin administration. These effects depend on inhibition of the LPS-stimulated p38 activation, leaving intact LPS ability to modulate both NF-κB and Akt activity.54 Stimuli leading to NF-κB activation have been demonstrated to induce MuRF-1 hyperexpression and muscle atrophy.55 Previous data, however, showed that the NF-κB transcription factor is not activated above control levels in the skeletal muscle of rats bearing the AH-130 tumor,8 suggesting that the NF-κB pathway is only marginally involved in causing muscle wasting in this model system. Expression of ubiquitin and muscle-specific E3 ligases can be induced by glucocorticoids.56 In this regard, corticosterone plasma levels are indeed increased in rats bearing the AH-130 tumor.33 However, previous data have shown that treatment of tumor hosts with the glucocorticoid receptor antagonist RU38486 is unable to prevent muscle wasting as well as the increased ubiquitin mRNA levels.57 Finally, several reports emphasize a relationship between ubiquitin ligases overexpression and myostatin.58

Muscle atrophy in tumor-bearing animals is not only characterized by a lack of correlation between atrogin-1 expression and the state of activation of the PI3K/Akt pathway, but it is also associated with positive (proanabolic) modulations of protein synthesis regulators. Indeed, in both the AH-130 and the C26 hosts the levels of phosphorylated GSK-3β and p70S6K increased, while those of eIF2α were markedly decreased.

These observations, in particular the reduced levels of phosphorylated eIF2α, differ from what reported on a different experimental model of cancer cachexia, the MAC16 tumor. In the latter, autophosphorylation of the RNA-dependent protein kinase (PKR) was shown to result in eIF2α phosphorylation, leading to reduced protein synthesis.59 In this regard, the results obtained in the present experimental models are not in contrast with those reported on the MAC16 tumor, since protein synthesis rates in the AH-130 hosts are not reduced with respect to controls,60 which is compatible with the eIF2α dephosphorylation.

Consistent with previous reports, our observation is that anabolic signals are upregulated in cancer-related muscle wasting. In particular, increased p70S6K phosphorylation and reduced eIF2α phosphorylation have been reported in the EDL muscle 7 days after denervation, in agreement with an accretion in protein synthesis rates.26 Interestingly, this anabolic setting does not prevent muscle wasting, likely because degradative stimuli overcome any promotion of protein synthesis.

In addition to insulin and Akt, also the intracellular concentration of free amino acids may activate anabolism by acting directly on the mTOR pathway.40, 61 In this regard, the increased protein degradation that occurs in the muscle of tumor-bearing animals could raise the concentrations of intracellular free amino acids. The branched chain leucine, isoleucine and valine, in particular, may result in activation of the mTOR pathway.62 Consistently with this hypothesis, the levels of phosphorylated mTOR are increased in the skeletal muscle of rats bearing the AH-130 hepatoma (Muscaritoli et al., unpublished data).

In spite of the observation that several molecules involved in the regulation of protein synthesis are in an active state in the skeletal muscle of tumor-bearing animals, the rates of protein synthesis are just maintained at control levels.60 Whether this results from the lack of specific aminoacids or from activation/inactivation of other unknown mechanisms is not clear. As an example, recent data demonstrate that the initiation factor eIF3-f, a scaffold protein that coordinates both mTOR- and p70S6K-mediated translation, is one of the main targets of atrogin-1.63 Given the increased levels of this ubiquitin ligase in tumor-bearing animals, eIF3-f could be actively degraded, interfering with p70S6K activity and blocking protein synthesis. Moreover, MuRF-1, the other muscle-specific ubiquitin ligase overexpressed in tumor-bearing animals, has been proposed to act as a negative regulator of protein synthesis, independently from the PI3K/Akt pathway.64 Consistent with this report is the observation that MuRF-1 ablation is sufficient to prevent the degradation of myosin heavy chain in dexamethasone-treated mice.65

According to the literature, forcing anabolism, for example by overexpressing growth factors such as IGF-1 in the skeletal muscle, would counteract hypercatabolism.66 A number of studies question such view, however,67–70 and rather suggest that enhancing growth factor availability is not always sufficient to restore the normal metabolic balance unless the muscle environment is “permissive” for anabolism. The results shown in our study demonstrate that muscle wasting in cancer cachexia cannot be corrected by overexpressing IGF-1 and suggest that the stimuli leading to increased protein breakdown do not merely result from an altered equilibrium between anabolism and catabolism. This view is further supported by the observation that hypertrophy or correction of muscle atrophy can be promoted by IGF-1 gene transfer in healthy controls or aged mice, respectively. In this regard, healthy mice have a physiological muscle metabolism, while sarcopenia of aging is mainly related to impaired anabolism. The discrepancy between aged animals and tumor hosts in terms of response to IGF-1 overexpression could rely on the reduced food intake reported above for the latter. This hypothesis, however, could be discarded in force of a couple of observations: (i) although food intake in aged mice was not measured in our study, previous reports show that energy intake is significantly reduced in older individuals, reaching a condition of anorexia comparable to that observed in cancer patients71; (ii) IGF-1 overexpression has been demonstrated to effectively prevent fasting-induced wasting,72 rendering unlikely that modulations of food intake may significantly impinge on the response to increased IGF-1 availability. In this regard, pair-fed animals have not been included in our study, despite the significant reduction of food intake in tumor hosts. Previous observations showed that muscle depletion in pair-fed animals is more less marked than in tumor bearers (92% and 64% of controls values, respectively, at day 4 after tumor transplantation; see Ref.73). In addition, while muscle depletion in tumor hosts derives from protein hypercatabolism,6 it mainly depends on impaired synthesis in pair-fed animals.73 Indeed, these observations lead to the conclusion that pair-feeding does not provide reliable controls. In this regard, the effects of malnutrition in tumor hosts are likely overwhelmed and subverted in the frame of the tumor-host interplay, resulting in a distinctively peculiar syndrome. This view is also shared by other studies.12

The resistance to IGF-1 displayed by tumor-bearing animals may depend on a number of factors such as inflammatory mediators or stress signaling molecules. On the basis of preliminary experimental data, some hypothesis can be put forward. The 5′-AMPK, a negative regulator of cell size, is activated in response to nutrient deficiency or stress to support cellular metabolism and prevent a fall in ATP cellular level.74 Recently, AMPK has been proposed to induce expression of muscle-specific ubiquitin ligases both in vitro and in vivo.75 Our preliminary data show that AMPK phosphorylation tends to increase in the muscle of both C26 hosts and AH-130 bearers with respect to controls. On the other side, previous observations have shown that the mRNA expression of the peroxisome proliferator-activated receptor γ (PPARγ) coactivator 1 (PGC-1)α, a factor that promotes fiber-type switching from glycolytic toward more oxidative fibers, is markedly reduced in different models of skeletal muscle atrophy.76 Preliminary results show that PGC-1α nuclear levels are significantly reduced in the skeletal muscle of mice bearing the C26 tumor (Supporting Information Figure S4). On the basis of the mechanisms proposed by Sandri et al.,76 the loss of PGC-1α in C26 hosts could be responsible for an increased availability of the atrogin-1 promoter and a shift to fast-twitch glycolytic fibers, more susceptible to atrophy.

In conclusion, the results reported in our study show that muscle atrophy in tumor-bearing animals, at least in the 2 models herewith investigated, does not involve downregulation of the anabolic response, and thus appears inconsistent with a model based on downregulation of the IGF-1 signaling pathway. The present results suggest that the mechanisms underlying muscle wasting in cancer cachexia are, partially at least, different from those activated in other situations characterized by muscle atrophy.71 In this regard, Guasconi and Puri41 recently proposed that decline in anabolic signals (passive atrophy) and activation of catabolic pathways (active atrophy) may contribute differently to the pathogenesis of muscle atrophy in distinct diseases or unfavorable conditions. Further support to this hypothesis comes from reports showing that skeletal muscle atrophy as occurring in tumor hosts and in diabetic rats is attributable to different mechanisms.46, 77


The authors thank Prof. F. Cavallo and her group for the help in electroporation protocol setup.