Muscle wasting in animal models of severe illness

Summary

Muscle wasting is a serious complication of various clinical conditions that significantly worsens the prognosis of the illnesses. Clinically relevant models of muscle wasting are essential for understanding its pathogenesis and for selective preclinical testing of potential therapeutic agents. The data presented here indicate that muscle wasting has been well characterized in rat models of sepsis (endotoxaemia, and caecal ligation and puncture), in rat models of chronic renal failure (partial nephrectomy), in animal models of intensive care unit patients (corticosteroid treatment combined with peripheral denervation or with administration of neuromuscular blocking drugs) and in murine and rat models of cancer (tumour cell transplantation). There is a need to explore genetically engineered mouse models of cancer. The degree of protein degradation in skeletal muscle is not well characterized in animal models of liver cirrhosis, chronic heart failure and chronic obstructive pulmonary disease. The major difficulties with all models are standardization and high variation in disease progression and a lack of reflection of clinical reality in some of the models. The translation of the information obtained by using these models to clinical practice may be problematic.

Introduction

Muscle wasting is a characteristic feature of cachexia, which is a common complication of various clinical conditions such as cancer, sepsis, liver cirrhosis, heart failure and renal and pulmonary insufficiency, and significantly increases morbidity and suffering of patients. The presence of cachexia can be found in 20–60% of patients with chronic infections, up to 50% of patients with cancer and in 10–50% of patients with chronic heart failure, chronic obstructive pulmonary disease (COPD), liver cirrhosis and chronic kidney disease (Schols et al. 1993; Strassburg et al. 2005; Morley et al. 2006; Peng et al. 2007; Tan & Fearon 2008). A characteristic feature of the loss of skeletal muscle in cachexia, distinguishing it from the loss of muscle protein in starvation, is a poor responsiveness to the anabolic stimuli. In a number of studies, it was demonstrated that conventional nutritional strategies are not sufficient to stop the loss of muscle mass (Streate et al. 1987; Shaw & Wolfe 1989; Ng & Lowry 1991).

Much has been learnt about muscle wasting from studies using administration of various mediators of cachexia, such as cytokines and glucocorticoids. In these studies, a number of therapeutic agents and nutritional supplements (e.g. carnitine, glutamine and branched-chain amino acids) have been studied and have shown promising results. Unfortunately, their favourable effects mostly failed to be confirmed in human clinical trials, and their clinical utility remains an open problem.

The disappointing difference between the results of experimental and human studies can be explained at least partially by the lack of experiments on clinically relevant models that are essential for preclinical testing of selected agents to avoid hazardous experiments in humans. Obvious advantages of animal models are their homogeneity, usually good reproducibility, no interference with different therapies and a wide range of other unpredictable factors affecting the development of the disease in clinical conditions. It would be beneficial if animal models resembled the human cachectic disorders as closely as possible. Optimal parameters evaluating the development of muscle wasting and effectiveness of its treatment in animal models of severe illness might be changes in protein synthesis and proteolysis in muscle in short-term studies and changes in muscle weight and protein content in long-term studies. As fluid shifts related to onset or treatment for oedema or ascites are frequent in severe illness, the weight loss is not a good parameter of cachexia and muscle wasting. Rather, when using young animals, instead of the loss of body weight there is slower weight gain in animal models of severe illness.

The aim of this article is to assess the suitability of various models of severe illnesses for the investigation of pathogenesis and therapy for muscle wasting. Therefore, the article is targeted on alterations in protein balance, protein synthesis and proteolysis in skeletal muscle in various models of the main muscle wasting disorders – sepsis, cancer, liver cirrhosis, renal insufficiency, chronic heart failure, chronic obstructive lung disease – and in intensive care unit patients.

Models of muscle wasting in sepsis

Sepsis can be defined as an uncontrolled systemic inflammatory response resulting from microbial infection mediated by activated immune cells. Occurrence of sepsis is increasing during the course of time, mainly as a result of the growing number of elderly people and the widespread use of antibiotics, which encourages the growth of drug-resistant microorganisms. Muscle dysfunction occurs in 40–70% of septic patients (Bolton 2000). Studies in septic patients and experimental animals have provided evidence that the myofibrillar proteins are particularly sensitive to the effects of sepsis (Hasselgren & Fischer 1998). The ubiquitin-proteasome system has a major role in skeletal muscle proteolysis and is activated by several pathways, including proinflammatory cytokines, reactive oxygen species and cyclooxygenases (Rabuel et al. 2004; Holecek 1996).

Sepsis development in humans follows two distinct hemodynamic phases. The first, characterized by increased cardiac output and low systemic vascular resistance, is referred to as hyperdynamic. During progression of sepsis, cardiac output declines and gradually develops the second, hypodynamic phase of sepsis, resulting in septic shock. The main commonly used models of sepsis are administration of endotoxin, administration of bacteria and caecal ligation and puncture (CLP).

Endotoxin administration

Endotoxin (or lipopolysaccharide, LPS) is a major component of the cell wall of Gram-negative bacteria recognized by Toll-like receptors 4 (TLR4) of innate immune cells. Activated immune cells release various mediators of inflammatory response such as cytokines and chemokines. As endotoxaemia may be observed not only in sepsis and inflammatory states, but also in multiple trauma, during cardiac, abdominal and vascular surgery, rheumatoid arthritis, liver cirrhosis, Crohn’s disease and many other disorders as well, animal models of endotoxaemia are very useful in investigation of pathogenesis and treatment possibilities of a wide spectrum of diseases. Although a huge amount of essential findings concerning the effects of endotoxin originate from mouse experiments, most of the studies evaluating the effect of endotoxaemia on protein turnover in skeletal muscle have been performed on rats.

The response to LPS administration is dependent on the source, dose and route of its administration (Breuille et al. 1999). Continuous infusion and/or repeated administration of endotoxin leads to endotoxin tolerance (Fish & Spitzer 1984; Jepson et al. 1986; Fink & Heard 1990). Therefore, the most frequently used LPS administration is as a single i.p., i.v. or s.c dose.

Endotoxin models using low doses (below 0.5 mg/kg b.w.) induce an increase in body temperature and a hypermetabolic state. However, animals become rapidly resistant to endotoxin and exhibit rapid growth recovery (Jepson et al. 1986; Fink & Heard 1990; Cross et al. 1993). Moderate doses of endotoxin (0.5–5 mg/kg b.w.) are not lethal; they induce transient hypothermia and transient increase in proinflammatory cytokines for 3 h (Schotanus et al. 1995; Holecek et al. 1995b; Holecek et al. 1998). Then activation of the hypothalamo-pituitary axis and a febrile response are induced. Growth recovery of the animals usually occurs after 48 h. Administration of a high dose of endotoxin (10 mg/kg b.w. and more) induces intestinal ischaemia, decreased sensibility, hypoglycaemia, hyperlacticacidaemia and other deteriorations typical of endotoxic shock.

The alterations in protein metabolism in skeletal muscle induced by LPS administration in different laboratories are summarized in Table 1. The data indicate that the LPS administration activates proteolysis in muscle via the ubiquitin-proteasome pathway and inhibits protein synthesis. The response can be modulated by proteasome inhibitors (Kadlcikova et al. 2004; Safranek et al. 2006a; Holecek et al. 2009). In most experiments, moderate or high, shock-inducing dose of LPS was administered. Models using moderate doses of LPS, in which hyperdynamic cardiovascular response is observed, seem to be clinically more relevant. The high dose of LPS is accepted as a model of septic shock (Fink & Heard 1990). On the basis of a series of experiments performed in our laboratory, we have chosen intraperitoneal application of 5 mg/kg b.w. of LPS isolated from of E. coli, which induces significant increase in myofibrillar proteolysis at 24 h after surgery (Kovarik et al. 2010).

Table 1.

Muscle protein metabolism in LPS model of sepsis

Model	Species	Results	References
LPS (E. coli) 10 mg/kg i.p. † at 2, 6 and 12 h	Rat	↑myofibrillar proteolysis in EDL ↑mRNA for ubiquitin in EDL (maximum after 6 h)	Chai et al. (2003)
LPS (E. coli) 12 mg/kg i.p. † at 48 or 96 h	Rat	↓diaphragm mass and protein content at 96 h ↑proteolysis at 48 and 96 h ↑caspase-3 activity in diaphragm	Supinski et al. (2009)
LPS (E. coli) 3 mg/kg, i.p. † at 6, 24, 48 and 72 h	Rat	↓muscle weight and protein content (at 24, 48 and 72 h after LPS)	Macallan et al. (1996)
LPS (E. coli) 3 mg/kg s.c. † between 6 and 72 h	Rat	↓protein synthesis, maximum at 30 h (GM) ↑protein degradation, maximum at 24 h (GM)	Jepson et al. (1986)
LPS (E. coli) infusion of 0.08 mg/kg † at 18 h	Rat	↓protein synthesis ↑protein degradation	Ash and Griffin (1989)
LPS (S. enteritidis) 2 mg/kg i.p. and i.v. infusion (0.1 mg) † at 3.5 h	Rat	↓protein synthesis in GM ↑whole-body proteolysis	Holecek et al. (2000b)
LPS (P. aeruginosa) 1 mg/kg i.p. † at 12 and 18 h	Mouse	↓weight and protein content 18 h (GM) ↑mRNA for atrogin-1/MAFbx, MuRF1 and polyubiquitin at 12 h (GM)	Jin and Li (2007)
LPS (E. coli) 25 μg/mouse † at 4 h	Mouse	↓protein synthesis in GM	Lang et al. (2010)

^†sacrifice time after LPS treatment; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius. LPS, lipopolysaccharide.

The main problem of LPS models is that they may not reproduce all the sequelae of the sepsis adequately because of very high, although transitory, increase in plasma cytokine levels, and rapid clearance of LPS from the circulation, which is not observed in humans. There is a need for a model of endotoxaemia characterized by persistent hypermetabolism and muscle protein wasting for longer periods of time.

Bacteria administration

Administration of live bacteria may reproduce a long-lasting model with many hallmarks of clinical sepsis. The main problem of this approach is that injected bacteria frequently fail to colonize and replicate (Buras et al. 2005), and the consequent bacterial lysis leads rather to a model of toxaemia than to the model of sepsis. As a limited number of bacterial strains have the characteristics allowing their replication and dissemination in the animal, the choice of a proper pathogen is important for the efficacy of the model (Cross et al. 1993). The route of bacteria administration seems to be important as well. Breuille et al. (1999) demonstrated that intravenous route of injection of E. coli resulted in more significant and much more prolonged decrease in body weight than i.p. administration.

There are a limited number of studies evaluating the effect of administration of live bacteria on muscle protein metabolism (Table 2). Voisin et al. (1996) described a model of septic state lasting for 10 days after i.v. injection of E. coli, in which they distinguished three phases: acute septic phase (day 2 postinfection), chronic septic phase (day 6) and a late phase (day 10). In this model, a significant increase in protein synthesis in acute phase and in proteolysis in acute and chronic phase was demonstrated in epitrochlearis muscle isolated from septic rats in comparison with pair-fed animals.

Table 2.

Muscle protein metabolism in models of sepsis induced by bacteria administration

Model	Species	Results	Reference
E. coli, i.v. † at 2, 6 and 10 days	Rat	↓muscle mass (SOL, EDL, TA, EPI) ↑protein synthesis in incubated EPI at day 6 ↑proteolysis in incubated EPI at days 2, 6 and 10	Voisin et al. (1996)
S. aureus, i.m. † at 24 h	Mouse	↑proteolysis in incubated GM ↑mRNA for proteasome C5 subunit (GM) ↑chymotrypsin-like enzyme activity (GM)	Khal and Tisdale (2008)

^†sacrifice time after bacteria administration; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius.

Caecal ligation and puncture (CLP)

This model provides a focus on necrotic tissue and polymicrobial contamination of the peritoneal cavity by the host’s own gut flora and resembles the situation in patients with perforated appendicitis or diverticulitis and peritoneal sepsis. The model reproduces the hemodynamic changes seen in humans with sepsis, and it is more clinically relevant than somewhat artificial models of bacteria infusion, faecal inoculation and colon ascendens stent peritonitis.

After abdominal incision, the caecum is ligated below the ileocaecal valve and punctured. The severity of the inflammation outcome can be influenced by the needle size, number of punctures and location of the caecal ligation (Singleton & Wischmeyer 2003). In most studies, the caecum is ligated below the ileocaecal valve and punctured twice with an 18- or 20-gauge needle. In the CLP model of sepsis, hyperdynamic cardiovascular response associated with enhanced cytokine production develops within 6 h after surgery (Safranek et al. 2006a,b).

Data demonstrated in Table 3 indicate a marked increase in proteolysis in skeletal muscle in rats 16 or 18 h after surgery and the suitability of this model for studying muscle wasting by using incubated skeletal muscle. Unfortunately, because of reduced permeability of muscles isolated from rats weighing more than 60 g, most of these studies have been performed using young animals.

Table 3.

Muscle protein metabolism in model of sepsis induced by CLP

Model	Species	Results	Reference
CLP † at 16 h	Rat	↑proteolysis in incubated EDL	Fischer et al. (2000); Fareed et al. (2006); Zamir et al. (1994)
CLP † at 18 h	Rat	↑proteolysis in incubated EDL NS effect on protein synthesis in SOL and EDL	Safranek et al. (2006b)
CLP † at 18 h	Rat	↓protein synthesis in incubated SOL and EDL ↑proteolysis in incubated SOL and EDL	Kadlcikova et al. (2004)
CLP † at 16 h	Rat	↑proteolysis in incubated EDL and SOL	Hummel et al. (1988)
CLP † at 16 h	Rat	↑proteolysis in incubated EDL ↓protein synthesis in incubated EDL	Hobler et al. (1998)
CLP † at 5 days	Mouse	↓weight and protein content of GM ↑mRNA of MuRF1 and Atrogin-1	Nystrom et al. (2009)
CLP † at 7 days	Mouse	↓muscle mass (GM)	Lang et al. (2010)

^†sacrifice time after LPS or bacteria administration or surgery; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius. CLP, caecal ligation and puncture.

Other models of sepsis

A clinically relevant model of sepsis is also the faecal peritonitis model in which faecal inoculum or bacteria suspended in a fibrin clot is implanted into peritoneal cavity (Lang et al. 1983) or a stent is placed by incision of the colon ascendens allowing faecal matter to flow from the colon into the peritoneal cavity (Zantl et al. 1998). In these models, the animals develop a hypermetabolic response; the serum cytokine response is similar to that observed in septic patients with signs of sepsis that progressively worsens and culminates in a septic shock and death within several (2–5) days. Unfortunately, in these models of sepsis, there is limited information concerning alterations in muscle protein metabolism. The useful model of acute-phase response associated with muscle protein catabolism is injection of turpentine oil (Holecek et al. 2006; Muthny et al. 2008).

Models of muscle wasting in cancer

Cachexia and muscle wasting occur in up to one half of patients with cancer, and it is estimated that up to 20% of all cancer deaths are caused directly by cachexia. The greatest incidence of muscle wasting is seen among patients with gastric, pancreatic, lung and colorectal tumours (Tan & Fearon 2008). Loss of skeletal muscle in patients with cancer results from decreased protein synthesis at the level of protein translation and increased protein degradation predominantly through the activation of ubiquitin-proteasome system. The preferred substrate degraded through ubiquitin-proteasome system is myosin heavy chain (Acharyya et al. 2004).

Although anorexia is frequently present, depression of food intake alone seems not to be responsible for the muscle wasting as nutritional supplementation or pharmacological manipulation of appetite is unable to reverse the catabolic process (Tisdale 2002). It seems that the pathogenesis of muscle wasting is the outcome of a variety of interactions between the host and tumour. The tumour’s role may include secretion of proinflammatory cytokines and tumour-specific factors (e.g. proteolysis-inducing factor). The principle mediators of host response are cytokines, particularly TNF-α and interleukins IL-1 and IL-6, and altered hormone concentrations (e.g. cortisol and glucagon), resulting in insulin resistance, anorexia and altered metabolism of all three nutrients.

Cancer cachexia has been studied in animals using tumour models produced by transplanting tumour cells and by administering large amounts of potent carcinogens. A new perspective is to use the genetically engineered mouse models that have been developed recently. The metabolic response of the host to tumour differs according to the type of the tumour and the phase of its development. Usually, three phases can be distinguished – the latent phase, the phase of progressive loss of body weight and anorexia associated with the increased rates of protein turnover and energy expenditure, and the premortal phase, which is characterized by severe cachexia and decreased energy expenditure. The main characteristics of tumours commonly used to study cachexia development are the following:

Walker 256 carcinosarcoma (rat model)

This tumour is a solid tumour which grows after s.c. injection of tumour cells. Within 10 days marked changes in the plasma levels of hormones and cytokines develop, including a factor with similar metabolic effects to the proteolysis-inducing factor isolated from the cachexia-inducing MAC16 tumour and urine of cachectic patients (Yano et al. 2008), which results in a marked decline in protein and fat content in tissues (Pizato et al. 2006).

Morris hepatoma 7777 (rat model)

After s.c. implantation, this tumour grows slowly, reaching the size of about 10% of body weight within 4–5 weeks. Cachexia develops within 2 weeks, and the death of the host occurs within 5 weeks (Fields et al. 1981; Le Bricon et al. 1995). It seems that cachexia development in this model is independent of proinflammatory cytokines, and the weight loss is attributable mostly to anorexia (Ruud & Blomqvist 2007).

Yoshida ascites hepatoma 130 (AH130) (rat model)

After i.p. injection of tumour cells, rapid development of anorexia and cachexia occurs, followed by death in about 2 weeks. Cachexia develops at a tumour burden not exceeding 0.1% of the host body weight (significantly less than in most experimental models), and thus, Yoshida AH130 appears to be appropriate for the investigation of the pathogenesis of tumour-associated cachexia in humans (Tessitore et al. 1993).

Lewis lung carcinoma (murine model)

This is probably the most commonly used model of cancer cachexia in mice. There is no decrease in food intake despite marked loss of body weight within 2 weeks after i.m. inoculation (Busquets et al. 2004).

Murine adenocarcinoma 16 (MAC16)

This model shows no effect on food intake, induces significant weight loss within 2 weeks and produces a proteolysis-inducing factor and a lipid mobilizing factor (Cannon et al. 2007).

MCG 101 (murine model)

This is a sarcoma, originally induced by methylcholanthrene, which grows s.c. in mice. The tumour reaches 15% of body weight after 14 days. It induces rapid development of cachexia and anorexia, the tumour produces PGE₂, and animals die within 15 days (Svaninger et al. 1983, Cahlin et al. 2000). Lundholm et al. (1978) reported changes in protein metabolism in skeletal muscle, heart and liver in mice bearing MCG 101 tumour that are similar to those observed in patients with cancer.

C26 colorectal adenocarcinoma (murine model)

The tumour reaches 10% of body weight within 4 weeks after s.c. injection into the flank region of the mouse (Zhou et al. 2010).

Human melanoma (murine model)

There are various cell lines (e.g. SEKI, G36, A375, MEWO) that can be used in nude mice that cannot mount an immune reaction and thus are permissible for the tumour growth. Severe body weight loss and cachexia-like appearance develop within 2–3 weeks after s.c. transplantation of tumour blocks of SEKI and G361 lines, but not after transplantation of A375 and MEWO lines (Mori et al. 1991; Hanada et al. 2004). There is a lack of information about alterations in protein metabolism in skeletal muscle.

Genetic models of cancer

The models are based on mutations leading to gain of functions of proto-oncogenes (e.g. Kras) or to loss of functions of tumour-suppressor genes (e.g. APC and p53). Widely used are Apc^Min/+ mouse and KPC mouse.

The Apc^Min/+ mouse is a model of intestinal polyposis and cancer. APC (adenomatous polyposis coli) gene is a tumour-suppressor gene mutated in a large percentage of colon cancers in humans. The loss or mutation of APC results in dysregulation of the Wnt signalling pathway and alterations in cell adhesion and chromosome stability (McCart et al. 2008). A wide spectrum of Apc mutant mice has been developed. The Apc^Min/+ (multiple intestinal neoplasia) mice are heterozygotes for a mutation in APC gene and spontaneously develop intestinal and colon adenomas, muscle wasting, fatigue and enhanced Il-6 levels. Animals die by 4 months of age from bowel obstruction (Baltgalvis et al. 2010; White et al. 2011).

The KPC mouse (LSL-Kras^G12D/+; LSL-Trp53^R172H/+; Pdx-1-Cre mouse) is a model of pancreatic ductal adenocarcinoma. KPC mice develop lesions called pancreatic intraepithelial neoplasia that progress to invasive metastatic carcinoma resistant to chemotherapy, cachexia, jaundice and ascites. The median survival is nearly 5 months (Hingorani et al. 2005; Olive et al. 2009; Ijichi 2011).

The main alterations observed in food intake and protein turnover in skeletal muscle of tumour-bearing animals are summarized in Table 4. The findings clearly indicate that most of these models could be a suitable tool in the investigation of muscle wasting in cancer cachexia.

Table 4.

Changes in protein metabolism in skeletal muscle in tumour-bearing animals

Model	Species	Results	Reference
Walker 256 † at 20 days	Rat	↑proteolysis (incubated GM) ↓protein synthesis (incubated GM) ↓myosin content (GM)	Ventrucci et al. (2004)
Yoshida AH 130 † at 8 days	Rat	↓muscle weight (EPI, SOL, EDL) ↑proteolysis (incubated EPI) ↓protein synthesis (incubated EPI)	Strelkov et al. (1989)
Yoshida AH 130 † at 2, 4 and 7 days	Rat	↓muscle weight at days 4 and 7 (EDL, SOL, TA) ↑atrogin-1 and MuRF-1 mRNA levels in GM	Costelli et al. (2006)
Yoshida AH 130 † at 4 days	Rat	↓muscle weight (GM)	Muscaritoli et al. (2003)
Yoshida AH 130 † at 4 days	Rat	↑proteolysis (EPI)	Baracos et al. (1995)
Yoshida AH 130 † at 7 days	Rat	↓muscle weight (GM, TA, SOL, EDL) ↑expression of proteasome system	Busquets et al. (2004)
Yoshida AH 130 † at 2, 4, 6, 8 and 10 days	Rat	↓muscle weight (GM) ↑proteolysis (GM)	Tessitore et al. (1987)
Morris hepatoma 7777 † at 21 days	Rat	↓muscle weight (EPI, SOL, EDL) ↑proteolysis (incubated EPI) ↓protein synthesis (incubated EPI)	Strelkov et al. (1989)
MCG 101 sarcoma † at 14 days	Mouse C-57	↓protein synthesis in muscle (hindleg) ↑cathepsin-D activity (hindleg)	Lundholm et al. (1978)
MCG 101 sarcoma † at 5 days	Mouse C57BL/6J	NS changes in proteolysis ↓protein synthesis (incubated EDL)	Svaninger et al. (1983)
Lewis lung carcinoma † at 15 days	Rat	↓muscle weight (GM, TA, SOL) ↑expression of ubiquitin-dependent, calcium-dependent and lysosomal system	Busquets et al. (2004)
C26 adenocarcinoma † at 25 days	Mouse CDF1	↓muscle weight (EDL, SOL, TA, calf) ↑ubiquitin, atrogin-1 and MuRF-1 in GM ↑production of proinflammatory cytokines	Zhou et al. (2010)
MAC 16 † at 1825 days	Athymic mouse	↓muscle weight (hind leg) ↑production of proinflammatory cytokine	Cannon et al. (2007)
ApcMin/+ mice † at 26 weeks	Mouse	↓muscle mass (GM) ↑apoptotic signalling (GM) ↓muscle fibre cross-sectional area (GM)	Baltgalvis et al. (2010)
ApcMin/+ mice † at 14–20 weeks	Mouse	↓muscle mass (GM) ↑proteolysis (↑ atrogin-1 and MuRf-1 in GM) ↓protein synthesis (GM)	White et al. (2011)

^†sacrifice time after inoculation of tumour cells; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius.

A serious problem of clinical relevance to cachexia development in tumour-bearing animals may be a large tumour burden. In some studies, the tumour formed up to 30% of body weight (Walker 256, MCG 101). The capability of such tumours to capture glucose and amino acids, thus depriving the host of needed substrates, may be enormous and unusual in human cancer. Therefore, animals bearing a large tumour burden may be a poor model of human cancer and cancer cachexia. Tumours in which cachexia manifests at a small tumour size and which are therefore more suitable to study muscle wasting, are, among others, the Yoshida AH130 in rat and Lewis lung carcinoma in murine models.

The major advantages of genetic models of cancer are that the tumours develop spontaneously in appropriate tissue, there are no requirements for an immunocompromised host and usually there is a lower ratio of tumour to body mass. Genetic models more closely mimic the dynamic interaction between host and tumour and more closely resemble human disease. Unfortunately, the studies exploring these models in the investigation of muscle wasting in cancer cachexia are surprisingly rare.

Models of muscle wasting in liver cirrhosis

The reports of the prevalence of hepatic cachexia and muscle wasting are not consistent and vary from 20% to 50% in patients with liver cirrhosis (Peng et al. 2007). Low level of muscle mass is associated with higher mortality and does not correlate with the degree of liver dysfunction (Montano-Loza et al. 2012).

It seems that two metabolically different situations may develop in chronic liver disease. In liver disease without signs of inflammation (e.g. stabilized liver cirrhosis), muscle wasting is probably more due to decreased protein synthesis than increased proteolysis (Merli et al. 1990; Holecek et al. 1996b). The decrease in protein turnover in muscle may be related to the inhibitory effect of ammonia on protein synthesis (Schott et al. 1984), decreased intake of food, maldigestion, malabsorption (Holecek et al. 1995a,b), decreased concentrations of BCAA and increased concentrations of glutamine because of hyperammonaemia (Holecek et al. 2000a,b; Holecek et al. 2011 and Holecek 2002). Bacterial infections and endotoxaemia, which result mostly from compromised gut barrier function, may induce a systemic response associated with enhanced production of cytokines, followed by hypermetabolism and increased protein turnover as observed in a number of clinical studies (O’Keefe et al. 1980; Swart et al. 1988; Blonde-Cynober et al. 1994).

The most common method for experimentally induced liver cirrhosis is to use multiple doses of carbon tetrachloride (CCl₄) in rats. Micronodular cirrhosis develops usually within 2 or 3 months and reproduces most of the features of cirrhosis in humans including portal hypertension, ascites, collateral venous channels and low-branched chain/aromatic amino acids plasma ratio (Holecek et al. 1996a,b). CCl₄ may be given s.c., i.m. or i.p. or introduced through a gastric tube. We have had good experience with intragastric administration of CCl₄ in olive oil, in dose of 2 ml/kg three times per week, for 2–3 months (Holecek et al. 1995a, 1996a,b), while i.p. or s.c. methods demonstrated significant shortcomings (inflammatory response in intraperitoneal space and poor cirrhosis development, respectively). Less frequently than CCl₄, other hepatotoxins, for example thioacetamide, dimethylnitrosamine or galactosamine, are used. A surgical alternative model to study the development of liver cirrhosis is based upon the ligation of the common bile duct (Kountouras et al. 1984). Cirrhosis develops in most animals after 3–5 weeks. Although alcohol is the main cause of hepatic cirrhosis in the Western world, alcohol alone is not capable of producing cirrhosis in animal models.

Unfortunately, animal studies evaluating the effect of chronic liver disease on muscle wasting are rare (Table 5). There are no studies examining the effect of possible therapeutic strategies to prevent or treat hepatic cachexia.

Table 5.

Changes in protein metabolism in animal models of chronic liver disease

Model	Species	Results	Reference
CCl4 i.g. + phenobarbital † at 10, 20 and 60 days	Rat	↑proteolysis (incubated m. vastus lateralis) ↓cross-sectional area of muscle fibres (diaphragm)	Weber et al. (1992)
CCl4 i.g. † at 8 weeks	Rat	↓gain of body weight ↓proteolysis in muscle (GM) ↓protein synthesis (GM)	Holecek et al. (1995a)
Bile duct ligation † at 5 weeks	Rat	↓muscle weight (diaphragm, m. scalenus, m. parasternalis, GM, m. plantaris, EDL, SOL) ↓cross-sectional area of type IIx/b muscle fibre (diaphragm, GM)	Gayan-Ramirez et al. (1998)
Bile duct ligation † at 5 weeks	Rat	↓muscle weight (EPI, GM, SOL, EDL) ↑proteolysis (incubated EPI) ↑expression of ubiquitin-dependent system (GM)	Lin et al. (2005)

^†sacrifice time after treatment; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius; SOL, m. soleus.

Models of muscle wasting in renal failure

The prevalence of muscle wasting and protein-energy malnutrition in patients with end-stage renal disease undergoing maintenance dialysis is 16–54% (Kopple 1997). Several muscle biopsy studies performed in patients with renal failure demonstrated significant atrophy across all fibre types in both locomotor and non-locomotor muscles (Kouidi et al. 1998; Sakkas et al. 2003).

Aetiology of myopathy in patients with renal failure has not been established definitively. Undoubtedly important factors are decreased physical activity and depletion of amino acids resulting from low-protein diets and losses during dialysis (Teplan et al. 2000). An important feature of chronic renal failure is elevated resting energy expenditure (Utaka et al. 2005). This indicates that other factors, including systemic inflammation, corticosteroids, high levels of angiotensin II and metabolic acidosis, play important roles in the pathogenesis of the muscle wasting. Several studies showed that metabolic acidosis may affect protein and amino acid metabolism directly in a number of tissues (Ballmer et al. 1995; Holecek et al. 2003; Safranek et al. 2003) and promote muscle protein catabolism through upregulation of the ubiquitin-proteasome system (Mitch et al. 1994; Caso & Garlick 2005).

There are various animal models exhibiting features of renal failure. The most widely used approach to induce renal failure is reduction in renal mass. The reduction is usually achieved by a two-stage surgical procedure to avoid high early mortality. The first stage is a partial nephrectomy using several methods – ligation and resection of the two poles, ligation of branches of the renal artery, coagulation and/or freezing of the cortical area. The second stage performed a few days later is a total nephrectomy of the decapsulated contralateral kidney. In this 5/6 nephrectomy model, four stages of renal failure can be identified – the acute one (the first 48 h after the surgery), the period of improvement resulting from functional hypertrophy (1–6 weeks), the period of ‘stable’ chronic renal failure (GFR is about 20–40%) and the period of progression to the end-stage renal failure (Boudet et al. 1978). Model of 11/12 nephrectomy is less frequently used, that is, the ligation of branches of renal artery to infarct 5/6 of the kidney or resection of the two poles and part of the remaining cortex in the first stage of the procedure. Chronic renal failure develops within 2 weeks, but mortality is much higher than in the 5/6 nephrectomy model (personal experience). The major difficulties of this model are its standardization and high variability in progression of renal failure (Holecek et al. 2001).

Most findings using these models demonstrate that protein wasting in muscle in chronic renal failure is related to decreased protein synthesis and enhanced proteolysis (Table 6), that correction of acidosis may suppress accelerated proteolysis in muscle and that important mediators of protein wasting in chronic renal failure are glucocorticoids (May et al. 1987, 1996 and Bailey et al. 1996). Nevertheless, our findings in a model of acute renal failure induced by bilateral nephrectomy indicated that the cause of rapid depletion of body proteins was the significant decrease in whole-body protein synthesis rather than the increased proteolysis (Holecek et al. 2000c).

Table 6.

Changes in protein metabolism in animal models of acute and chronic renal failure

Model	Species	Results	Reference
Bilateral nephrectomy † at 24 h	Rat	↓protein synthesis (GM) ↓whole-body protein turnover	Holecek et al. (2000c)
5/6 nephrectomy † at 28 weeks	Rat	Significant uraemia only in four of 12 animals NS effect on gain of body weight ↑whole-body protein turnover NS effect on protein synthesis in muscle (GM)	Holecek et al. (2001)
11/12 nephrectomy † at 2 weeks	Rat	↓protein synthesis (perfused hindquarters) ↑proteolysis (perfused hindquarters)	May et al. (1987)
7/8 nephrectomy † at 15 days	Rat	↑proteolysis (incubated EPI) ↑expression of ubiquitin and proteasome subunit C9	Bailey et al. (1996)

^†sacrifice time after surgery; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius; SOL, m. soleus.

An important contribution to the understanding of pathogenesis of muscle wasting in renal failure is provided by studies using animal models of acidosis. The models of chronic acidosis are induced by administration of acid or ammonium chloride in food and/or in drinking water, and acute acidosis is achieved frequently by continuous acid infusion. In the ammonium chloride model, control animals receive a diet or water mixed with ammonium acetate. The diet should be given by gavage, and the controls should consume equivalent amounts of diet and water. The pH of blood plasma may drop to 7.1 (May et al. 1986). A serious shortcoming of ammonium chloride models of acidosis is an interference with the effect of ammonium. Several studies have demonstrated that hyperammonaemia affects the protein and amino acid metabolism (Holecek et al. 2000a, 2011). A significant decrease in pH in acute models of acidosis can be induced in rats by infusion of 0.2 M HCl at a rate 1.6 ml/kg/h (Safranek et al. 2003).

Models of muscle wasting in chronic heart failure

Prevalence of cachexia and muscle wasting in chronic heart failure is 10–16% (Strassburg et al. 2005). The loss of lean tissue, muscle weakness and decreased muscle strength are associated with increased resting metabolic rate, muscle fibre atrophy, reduced capillary density and lower electromyographic activity (Poehlmann et al. 1994; Schulze et al. 2004). The most important aetiological factors of cardiac cachexia are impaired muscle blood flow, enhanced production of proinflammatory cytokines, physical inactivity, malabsorption because of oedematous bowel wall, and neurohormonal abnormalities, especially insulin resistance and increased tone of adrenergic and renin–angiotensin–aldosterone systems. The main models of cardiac cachexia are based on surgical techniques performed to cause myocardial infarction, on restriction of cardiac output by aortic banding, and on monocrotaline injection.

The myocardial infarction model

This involves ligation of the left anterior descending coronary artery. Control animals undergo sham surgery without coronary artery ligation. The problem of this model is variability of response and high mortality. Palus et al. (2009) demonstrated in this model a more pronounced loss of body weight and both lean and fat mass in adult male than in adult female rats. The cause of this gender difference, indicating that in this model of cardiac cachexia male rats are a better model than female rats, is not known.

Aortic banding model

This involves constriction of the ascending aorta using titanium clips. The animals develop congestive heart failure within 2–4 weeks. The rats display malnutrition and alterations in hepatic and renal functions (Héliès-Toussaint et al. 2005).

Monocrotaline

This is an alkaloid that produces pulmonary hypertension within 3 weeks and subsequent right ventricular failure.

Following all these procedures, symptoms of cachexia, such as drop in food intake, decrease in body weight gain and loss of fat, are apparent within a couple of weeks. Activated protein breakdown in skeletal muscle with increased activity of the ubiquitin-proteasome proteolytic pathway has been demonstrated in the monocrotaline model. However, alterations in protein synthesis and proteolysis have not been characterized in myocardial infarction and aortic banding models of cardiac cachexia (Table 7).

Table 7.

Changes in protein metabolism in animal models of chronic heart failure

Model	Species	Results	Reference
Myocardial infarction † at 7 weeks	Rat	↓body weight and lean mass in adult males at first week after surgery	Palus et al. (2009)
Myocardial infarction † at 8 weeks	Rat	NS effect on lean mass	Akashi et al. (2009)
Myocardial infarction † at 7 weeks	Rat	↓muscle weight and protein content (GM) ↓protein content (EDL, SOL, GM)	Nagaya et al. (2001)
Aortic banding † at 8 weeks	Rat	↓body weight ↓muscle weight and protein content (EDL, SOL, GM)	Héliès-Toussaint et al. (2005)
Monocrotaline 30 mg/kg, i.p. † at 3 weeks	Rat	↓body weight ↓muscle weight (EDL) ↑mRNA levels for MaFbx, ubiquitin and 20S and 19S proteasome subunits (EDL)	Steffen et al. (2008)
Monocrotaline 30 mg/kg, i.p. † at 4 weeks	Rat	↓single-fibre cross-sectional area in muscle (TA) ↑caspase-3 activity (TA) ↑apoptosis of skeletal muscle nuclei (TA)	Vescovo et al. (2002)

^†sacrifice time after treatment; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius; SOL, m. soleus.

Models of muscle wasting in chronic obstructive pulmonary disease (COPD)

The prevalence of cachexia in patients with COPD is up to 45% (Schols et al. 1993). Aetiology of muscle wasting in COPD is multifactorial, and the most important factors are hypoxia, reduced physical activity, anabolic/catabolic hormone imbalance, anorexia, corticosteroid treatment and systemic inflammatory response. The abnormalities are localized to the lower limbs, particularly to the quadriceps, while the strength of diaphragm and abdominal muscles is preserved. Quadriceps weakness associated with reduced fibre cross-sectional area and impaired oxidative enzyme capacity predicts mortality in patients with COPD (Swallow et al. 2007; Man et al. 2009). The main models of COPD include intratracheal instillation of elastase and cigarette smoke exposure.

Elastase-induced emphysema

Among various proteinases instilled intratracheally into the lungs, the most consistent and impressive airspace enlargement is produced by pancreatic elastase (Shapiro 2000).

Exposure to cigarette smoke

The animals are exposed to cigarette smoke using a smoking apparatus. There are species and strain differences in susceptibility and in abnormalities to smoke exposure. The most susceptible species developing symptoms of COPD within a few months of cigarette smoke exposure are guinea pigs (Wright & Churg 1990), while rats appear most resistant. Sensitive strains of mice are C57B1/6J and AKR/J (Martorana et al. 2006).

The number of papers examining muscle protein metabolism in models of COPD is extremely small (Table 8). A serious shortcoming of all these studies is the lack of information about food consumption and the absence of pair-fed controls although it is well known that hypoxia and chronic inflammation alter appetite and food intake.

Table 8.

Muscle protein metabolism in models of COPD

Model	Species	Results	Reference
Cigarette smoke † at 24 weeks	Mouse	↓body weight gain N.S. effect on muscle weight (EDL, TIB, GM, PL)	Gosker et al. (2009)
Cigarette smoke † at 18 weeks	Mouse	↓body weight gain ↓muscle weight (GM, m. biceps femoris) ↓cross-sectional area of types IIA fibres (GM)	De Paepe et al. (2008)
Porcine elastase (250 IU/kg intratracheally) † at 4 and 8 months	Hamster	↓muscle weight and fibre cross-sectional area after 8 months (EDL) NS effect on muscle force and relative muscle weight (EDL)	Mattson et al. (2008)
Porcine elastase (250 IU/kg intratracheally) † at 6 months	Hamster	↓cross-sectional area of fast-twitch types IIA, IIX and IIB fibres (m. biceps femoris, TA, GM, m. plantaris)	Mattson et al. (2004)

^†sacrifice time after treatment; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius; SOL, m. soleus.

Animal models of muscle wasting in intensive care unit (ICU) patients

Muscle wasting is found in up to 30% of ICU patients. This condition with negative effects on the recovery from primary illness, treatment costs and mortality has been called acute quadriplegic myopathy, critical illness myopathy, myopathy of intensive care, thick filament myosin myopathy, etc. (Larson 2007). Muscle wasting is associated with decreased muscle membrane excitability, muscle fibre atrophy and a decline in myofibrillar protein content (Friedrich et al. 2004; Ochala & Larsson 2008). Although the exact pathogenesis remains to be elucidated, five central risk factors can be identified including multiple organ failure, muscle inactivity, hyperglycaemia and use of corticosteroids and neuromuscular blockers (De Jonghe et al. 2009). As models of muscle wasting mimicking the conditions in ICU various models have been introduced including disuse atrophy, long-term exposure to neuromuscular blocking agents (NBA), and corticosteroid treatment and their combinations (Larson 2007).

Models of muscle disuse atrophy

The main models include denervation, joint immobilization and tail suspension. The most reliable model of muscle disuse atrophy appears to be denervation. In most denervation studies hindlimb muscles are denervated by crushing the sciatic nerve. An obvious advantage is the possibility of comparing denervated and innervated (contralateral) hindlimb muscles. It seems that tail suspension is more suited to studies on microgravity than on disuse (Ohira et al. 2002). Joint immobilization by cast or by a stick fixed with a bandage causes systemic stress response and alterations, which may interfere with changes induced by leg immobilization.

Neuromuscular blocking agents (NBA)

These are occasionally used in ICU with sedatives and/or analgesics for the management of patients who require the breathing machine. In this experimental model, the animal paralysed by NBA is mechanically ventilated and monitored. Unilateral mechanical loading of hindlimb muscles offers a unique possibility to study the effect of physical load on skeletal muscle (Larson 2007).

Corticosteroid treatment

Steroid myopathy is well documented in patients with Cushing’s syndrome; patients treated with high-dose corticosteroids develop a decrease in respiratory muscle strength and muscle weakness (Weiner et al. 1993; Decramer et al. 1994). The main synthetic glucocorticoid used in animal studies is dexamethasone administered in doses between 0.1 and 2.0 mg/kg for one or 2 weeks. Dexamethasone administration alters food intake, and pair-fed animals should be included in the experimental protocol.

There are a number of studies demonstrating that muscle wasting can be induced by each of these methods (please see Table 9). Loughna and Morgan (1999) demonstrated that passive stretch can modulate myosin heavy-chain gene expression in denervated muscles. Zhao et al. (2008) demonstrated that the glucocorticoid-induced muscle atrophy can be reversed by testosterone. Effective and clinically relevant models are corticosteroid treatment combined with peripheral denervation or with administration of NBA.

Table 9.

Muscle protein metabolism in models of muscle wasting in ICU patients

Model	Species	Effect	Reference
Tail suspension † at 7 days	Mouse	↓muscle weight and force (SOL) ↓fibre cross-sectional area (SOL)	Labeit et al. (2010)
UHI (stick fixed with a bandage) † at 7 days	Mouse	↓fibre cross-sectional area (EDL, TA) ↑atrogin-1 and MuRF1(EDL, TA)	Madaro et al. (2008)
UHI (sciatic denervation) † at 7 days	Rat	↓muscle weight (SOL, EDL) ↑ubiquitin mRNA expression (SOL)	Beehler et al. (2006)
Dexamethasone 0.1 mg/kg/day † at 7 days	Rat	↓muscle weight ↓fibre cross-sectional area (TA)	Schakman et al. (2005)
Dexamethasone 0.5, 1 or 2 mg/kg/day † at 14 days	Rat	↓diameter of muscle fibres (GM)	Konno (2005)
Dexamethasone 1 or 5 mg/kg/day † at 4 and 10 days	Rat	↓muscle weight (TA, GM) ↓fibre cross-sectional area (TA, GM) ↑chymotrypsin-like activity of proteasome (TA, GM)	Gilson et al. (2007)
Dexamethasone 0.7 mg/kg/day † at 1 and 7 days	Rat	↓muscle weight (GM) ↑proteolysis (GM) ↑MAFbx mRNA (GM)	Zhao et al. (2008)
Dexamethasone 0.54 mg/kg/day † at 6 days	Rat	↓muscle weight (EPI) ↑ubiquitin-specific proteases mRNA (EPI)	Combaret et al. (2005)
Corticosterone 100 mg/kg/day † at 2, 4, or 8 days	Rat	↑myofibrillar proteolysis ↓protein synthesis (perfused hindquarter)	Kayali et al. (1987)
NBA (α-cobratoxin) + MV hydrocortisone (16 mg/kg/day) † at 7–13 days	Rat	↓fibre cross-sectional area (EDL, SOL) ↓myosin mRNA expression (EDL, SOL, diaphragm, intercostal muscles)	Norman et al. (2006a)
UHI (sciatic denervation) dexamethasone (5 mg/kg/day) † at 7 days	Rat	↓muscle weight ↑atrogin-1 (plantaris muscle)	Mozaffar et al. (2007)
NBA (pancuronium) + MV bethamethasone (0.2 mg/kg/day) sepsis (E. coli infusion) † at 10 h–5 days	Pig	↓myosin heavy-chain and α-actin mRNA expression (m. biceps femoris a m. masseter)	Norman et al. (2006b)

^†sacrifice time; NBA, neuromuscular blocking agent; MV, mechanical ventilation; UHI, unilateral hindlimb immobilization; SOL, m. soleus; TA, m. tibialis anterior; EPI, m. epitrochlearis; EDL, m. extensor digitorum longus; GM, m. gastrocnemius.

Discussion and conclusions

Most models of muscle wasting disorders have been developed on rodents, particularly on mice and rats. Each species has its own strengths and weaknesses, and investigators should select the most appropriate model to test the hypothesis. The obvious advantage of mice is their small size (good for dosing of expensive substances), cheap housing and rapid breeding. Rats are better for surgical models and allow serial sampling of tissue. Unfortunately, most models of muscle wasting disorders have serious shortcomings.

Animal models of sepsis are heterogeneous with regard to the type of insult. At present, the best model of muscle wasting in sepsis is probably CLP in which a marked increase in protein breakdown in skeletal muscle is observed within 18 h after surgery. However, this model is not suitable for long-term studies, and there is a need for a model of severe chronic sepsis characterized by persistent hypermetabolism. Endotoxin models of sepsis may not replicate clinical sepsis and should be considered rather as models of systemic inflammatory response, which plays a central role in the genesis of muscle wasting in a number of diseases, such as cancer, congestive heart failure, trauma, burns and haemorrhage. There is limited experience with other models of sepsis although the data presented by Voisin et al. (1996) indicate that a clinically relevant model of muscle wasting in sepsis may be induced by intravenous administration of bacteria.

The studies using tumour-bearing animals provided important information about both pathogenesis of cachexia and the efficacy of a number of therapeutic strategies. Different tumours produce different alterations in food intake, and plasma hormone and cytokine concentrations. This variability indicates that the therapy for cancer cachexia should be tailored according to the type of tumour and that we have to learn which models of tumour-bearing animals are relevant to human conditions. This indicates also the difficulties in extrapolating the results from animals to humans and that the experiments evaluating the efficacy of specific therapeutic interventions for cancer cachexia should be verified on different tumour types. There is a need to explore genetically engineered mouse models of cancer that have no requirements for immunocompromised host and more closely mimic the spectrum of tumour–host interactions.

Animal studies evaluating the effect of chronic liver disease on muscle wasting are rare. The pathogenesis of the loss of skeletal muscle in chronic liver disorders seems to be markedly affected by the presence/absence of inflammatory response, and there is a need of animal models of chronic liver diseases with and without signs of inflammation.

The pathogenesis and treatment possibilities of muscle wasting in renal failure have been studied mostly using the model of 5/6 nephrectomy. The serious problem of this model is a high variability in progression of renal failure. Valuable information about muscle wasting in uraemia has been obtained using various models of acidosis.

The main models of chronic heart failure are myocardial infarction, aortic banding and monocrotaline injection. Activated protein breakdown in skeletal muscle has been demonstrated in monocrotaline model. Alterations in protein turnover in skeletal muscle in models of myocardial infarction and aortic banding are not well characterized.

The main models of COPD include intratracheal instillation of elastase and cigarette smoke exposure. There is a lack of well-performed studies examining muscle protein metabolism in these models.

The models of muscle wasting in ICU patients provide a valuable information about the effects of immobilization and interventions used in anaesthesiology and intensive care (administration of corticosteroid hormone and NBA). Findings of experimental studies are in good agreement with observations in ICU patients (Larson 2007) and demonstrate that passive stretch can modulate myosin heavy-chain expression in a muscle-specific manner (Loughna & Morgan 1999) and that the glucocorticoid-induced muscle atrophy can be reversed by testosterone (Zhao et al. 2008).

In conclusion, the data presented in this manuscript indicate that muscle wasting has been well characterized in animal models of sepsis, cancer and chronic renal failure and in ICU patients. There is a lack of good models of muscle wasting in liver cirrhosis, chronic heart failure and COPD. Most of the commonly used models of severe illnesses are not standardized, the degree of protein degradation in muscle in these models is not well characterized, and some models do not adequately reflect the clinical reality. The translation of the information obtained using these models to clinical practice may be problematic.

Author Contributions

The author meets ICMJE authorship criteria, and nobody who qualifies for authorship has been excluded.