Cancer cachexia and skeletal muscle atrophy in clinical studies: what do we really know?

Abstract Research investigators have shown a growing interest in investigating alterations underlying skeletal muscle wasting in patients with cancer. However, skeletal muscle dysfunctions associated with cancer cachexia have mainly been studied in preclinical models. In the present review, we summarize the results of clinical studies in which skeletal muscle biopsies were collected from cachectic vs. non‐cachectic cancer patients. Most of these studies suggest the presence of significant physiological alterations in skeletal muscle from cachectic cancer patients. We suggest a hypothesis, which connects structural and metabolic parameters that may, at least in part, be responsible for the skeletal muscle atrophy characteristic of cancer cachexia. Finally, we discuss the importance of a better standardization of the diagnostic criteria for cancer cachexia, as well as the requirement for additional clinical studies to improve the robustness of these conclusions.


Introduction
Cancer cachexia is a complex multifactorial syndrome characterized by involuntary and pathological weight loss, mainly due to skeletal muscle wasting. Patients experience a deterioration of their nutritional status, which is associated with a profound weakening of the body. Formation of oedemas, loss of appetite (or anorexia), and persistent fatigue are also observed. These effects considerably reduce patients' quality of life and overall survival. Cancer cachexia is a cause of death for 20-25% of patients, 1 and it is also a co-morbidity significantly affecting their overall survival. In 2011, a consensus of international experts defined cancer cachexia as being characterized by a progressive loss of skeletal muscle mass, with or without loss of adipose mass. This cachexia cannot be fully supported by conventional nutritional support and gradually leads to functional impairment. 2 It has been described as a continuum with three stages of diagnosis: precachexia, cachexia, and refractory cachexia. The risk of progression depends on the stage and the type of tumour, the presence of systemic inflammation, the response to antitumour treatment, as well as inter-individual variations such as genetic predisposition, body composition, food intake, physical activity, and co-morbidities. 3,4 It is however important to note that the severity of cachexia does not appear to correlate with the size of the tumour. 5 The prevalence of cachexia in cancer patients is estimated to be~35%. 6 It can even reach~80-90% for gastric and pancreatic cancers, and it is particularly prevalent in advanced stages of cancer. 7,8 Because of the pathophysiological complexity and multifactorial characteristics of this clinical syndrome, there is currently no effective treatment for cancer cachexia. Current research have focused on a multimodal approach, 9 which includes adapted anticancer treatment [10][11][12] ; pharmacological treatment that aims in particular at reducing systemic inflammation, counteracting the hypercatabolic state of patients, and/or stimulating their appetite 13,14 ; nutrition care [15][16][17][18][19][20][21] ; adapted physical activity [22][23][24] ; and psychosocial care. 25,26 It is therefore essential to further improve our understanding regarding the interplay of the molecular mechanisms involved in the onset and progression of cancer cachexia. In particular, it is of crucial importance to better understand the pathophysiological basis of skeletal muscle atrophy, which represents a major clinical feature of cachexia.
To improve our knowledge of the factors regulating skeletal muscle mass loss, clinical studies are essential. The critical point for clinical studies investigating cancer cachexia is to identify cachectic vs. non-cachectic cancer patients groups. The clinical definition established by Fearon et al. distinguishes between cachectic patients and non-cachectic patients based on weight loss and body mass index. 2 This definition has subsequently been validated in an international multicentre study with 861 patients. 27 However, recent studies have demonstrated the importance of continually developing and updating these clinical criteria, which take into account the evolution of the overweight and obesity prevalence 28 or the advanced age of most cachectic patients. 29 The aim of this review was to offer an up-to-date synthesis of the data, which suggest the presence of alterations in skeletal muscle from cachectic cancer patients. In addition, we discuss the importance of a standardization for the diagnostic criteria of cancer cachexia, as well as the need for additional clinical studies to improve the robustness of the conclusions.

Methodology
Articles indexed in PubMed were queried to identify clinical studies, which have analysed skeletal muscle from cachectic cancer patients. Search terms included '[(cancer) OR (carcinoma) OR (tumor) OR (malignant) OR (metastasis)] AND [(cachexia) OR (sarcopenia) OR (weight loss) OR (malnutrition)] AND [(skeletal muscle) OR (muscle mass) OR (lean body mass) OR (rectus abdominis) OR (quadriceps)] ± [(biopsy) OR (biopsies)]' ± filters: humans; adult: 19 + years; and publication dates from 01/01/1900 to 18/02/2020 were included. Among 749 publications identified through the database search, we excluded duplicates, review articles, preclinical studies, non-relevant or non-full-text clinical studies, as well as clinical studies where cachectic cancer patients were compared with healthy control patients. [30][31][32][33][34][35][36][37][38][39][40][41][42][43] Importantly, in that case, it is not possible to distinguish between the specific effects of cachexia and cancer on the various parameters examined. We also excluded a case study, which was carried out on a cachectic cancer patient without comparison with a population of non-cachectic patients. 44 The identified articles were manually searched to identify additional relevant publications.
Thirty-one clinical studies were identified by this review of the literature ( Table 1). The vast majority of these studies discuss patients with gastrointestinal cancer, including cancer of the oesophagus, stomach, liver, pancreas, colon, and anus. The main mechanisms studied cover the structure and typing of muscle fibres, pathways involved in proteolysis and protein synthesis, as well as lipid and mitochondrial metabolisms.

Skeletal muscle alterations in cancer cachexia
Structure and typing of muscle fibres Up until now, cancer cachexia has mainly been studied in preclinical models generally characterized by transplantation of cancer cells or injection of carcinogens. The most studied and best described models are colon-26 adenocarcinoma 77 and Lewis lung adenocarcinoma. 78 In these models, the cross-sectional area of muscle fibres decreases, 79 and this atrophy is more likely to affect type II fibres. 80 In clinical studies, little is known about the possible alterations in the structure and typing of muscle fibres in patients with cancer cachexia ( Table 1). Judge et al. have reported increased fibrosis and collagen content in skeletal muscle from cachectic pancreatic cancer patients. 45 Skeletal muscle ultrastructure also appears to be impaired, with an apparent disorganization and autophagosome formation in gastric cancer patients with cachexia. 46 The protein expression of myosin heavy chains, a major component of the muscular contractile system, was increased (for isoforms 1, 4, and 8), 47 decreased, 48 or unchanged 49 in cachectic patients with gastrointestinal cancer. Other muscular structural components have also been studied. The protein expression of actin and tropomyosin 1 and 2 was shown to be reduced 47,48 while that of β-dystroglycan was increased, and those of β-sarcoglycan and dystrophin remained unchanged 49 in cachectic patients with gastrointestinal cancer. Skorokhod et al. also identified genes associated (positively for most of them) with cancer cachexia. 50 These genes are involved in muscle contraction and development (e.g. actin, titin, tropomyosin, and troponin) and actin cytoskeleton rearrangement (e.g. cofilin, dystonin, and vinculin) in pancreatic cancer patients. 50 While a significant reduction in the cross-sectional area of muscle fibres was observed in gastrointestinal cancer patients with cachexia and muscle loss, 46,51 Op den Kamp et al. found no such change in cachectic patients with advanced non-small-cell lung cancer compared with pre-cachectic patients. 52 Finally, no significant alteration in fibre typing associated with cancer cachexia has been reported in clinical studies, either of gastrointestinal 51,53 or lung 54 cancer.  (P = 0.046) and phosphorylated p53 (P = 0.041) (apoptosis).

Muscle proteolysis and protein synthesis
In preclinical studies of cancer cachexia, there is abundant evidence for increased proteolysis, particularly through autophagy and the ubiquitin-proteasome system (UPS), and decreased protein synthesis in skeletal muscle (see review [81][82][83] ). Activation of NF-κB by inflammatory cytokines may result in muscle wasting in mouse models. 84 NF-κB induces the transcription of UPS genes, which activation has also been observed in preclinical models. 82,83 Apoptotic processes may also, at least in part, be responsible for muscle atrophy, as suggested by the data from different preclinical studies. [85][86][87][88] However, results from clinical studies are much more controversial ( Table 1). Some studies have reported an increased expression of autophagy markers (Beclin-1, LC3B, ATG5, ATG7, and p62) 46

Myosteatosis
Myosteatosis, which is defined as a pathological fat accumulation in skeletal muscle, has been studied in a preclinical tumour-bearing model (Ward colon tumour model). 89 In that case, the authors observed increased neutral lipids and total triglyceride content within the rat gastrocnemius muscle fibres. This observation is indicative of the presence of lipid droplets, as well as increased mRNA levels of key transcription factors involved in adipocyte gene expression (i.e. C/EBPδ, C/EBPα, and PPARγ). In cachectic cancer patients, weight loss has been associated with a greater number of intramyocellular lipid droplets, as observed by electron microscopy of abdominal muscle biopsies. 68 In the clinical practice, studies of skeletal muscle radiological attenuation by computed tomography scan have shown that it is very variable in cancer patients. Its decrease may, at least in part, reflect an accumulation of intramuscular lipids 90,91 and is associated with shorter survival. 92 The specific pathophysiological mechanisms that lead to myosteatosis have not yet been well-characterized. Transcriptomic analyses have suggested that disruption in oxidative phosphorylation and lipid accumulation may contribute to myosteatosis, as observed in abdominal skeletal muscle biopsies of patients with pancreatic or periampullary cancer and with significantly low radiological attenuation of skeletal muscle on computed tomography scan. 93 Genetic and transcriptomic studies of skeletal muscle biopsies obtained from cachectic patients (mainly gastrointestinal cancers) have identified single nucleotide polymorphism 69 and miRNA 70 involved in lipid biosynthesis ( Table 1). Narasimhan et al. also found alternatively spliced genes, which are mostly up-regulated, and encode enzymes and binding proteins from the lipid biosynthesis pathways. 60

Mitochondrial metabolism
Previous studies have suggested an association between mitochondria dysfunctions and skeletal muscle atrophy. 94,95 Several preclinical and clinical studies have therefore investigated mitochondrial metabolism dysregulation in skeletal muscle in the context of cancer cachexia. Muscle atrophy associated with mitochondrial dysfunctions has been observed in cachectic rodents. Among these dysfunctions, several parameters have been described, and they include increased mitochondrial surface 96,97 ; impairment of mitochondrial dynamics such as increased fission (Fis1), decreased fusion (Mfn1 and Mfn2), or biogenesis (PGC1α) 98 ; a decrease in the activity of the respiratory chain complexes [99][100][101] ; and an increase in UCP-2 and UCP-3 gene expression, 102,103 a questionable indicator of mitochondrial energy coupling. The study by Brown et al. on preclinical models of cancer cachexia suggested that alterations in dynamics (biogenesis and fusion), mitochondrial quality, and function precede muscle atrophy and that a decrease in mitochondrial protein content as well as an increase in mitophagy appears at a later stage of cachexia. 104 Regarding clinical studies, mitochondrial metabolism has rarely been studied in the context of cancer cachexia ( Table  1). A recent clinical study by de Castro et al., including patients with gastric or colorectal cancer, revealed an increase in Fis1 mRNA expression in skeletal muscle from cachectic patients compared with non-cachectic, but there was no modification in the levels of the fusion marker (Mfn2) and mitochondrial biogenesis transcripts (TFAM and PGC1α). 55 On the other hand, they showed by electron microscopy that there was an increase in the intermyofibrillary mitochondrial area, without modification in the number of mitochondrial DNA copies. This finding could illustrate the absence of modification in mitophagy activity. The authors suggested that these last observations supported the results from the preclinical study by Brown et al. who showed that impaired mitochondrial protein content and increased mitophagy only appear at a much later stage in cancer cachexia. 104 72 This finding suggests that increased proton leak may contribute to skeletal muscle catabolism through enhancement of energy expenditure. However, increase in skeletal muscle UCP-3 mRNA is a poor index of mitochondrial energy coupling. Finally, to our knowledge, it is important to note that no clinical study has explored the mitochondrial functionality in skeletal muscle from cachectic cancer patients.

Discussion and other perspectives
The overall result of clinical studies suggests the presence of alterations in skeletal muscle from cachectic cancer patients ( Table 1). These alterations may affect the structure of muscle fibres, the different pathways involved in proteolysis and protein synthesis, lipid metabolism (myosteatosis), and mitochondrial metabolism (i.e. mitochondrial surface and dynamics and mitochondrial DNA).
It is important to note that for most of the parameters studied, which could play a role in muscle wasting, the results of clinical studies diverge ( Table 1). A possible explanation for these differences may be the lack of well-established diagnostic criteria. An improvement and better standardization of the diagnostic criteria used for cancer cachexia identification is utterly necessary. These criteria should include key cachexia symptoms such as weight loss, taking body mass index into account, 28 muscle wasting, 105,106 appetite loss, performance status, and blood chemistry. On the basis of these key components, Zhou et al. have proposed a clinically applicable score for the classification of the various cachexia stages observed in cancer patients. 107 De Castro et al. have also used combined diagnostic criteria to discriminate cachectic and non-cachectic cancer patients. 55 To better improve patient care, it would be preferable to diagnose patients as soon as possible, possibly at a pre-cachectic state. Early diagnosis can reduce the risk of transition to a cachectic state with an appropriate choice of therapeutic and supportive regimens. But current diagnostic criteria are insufficient and inadequate to allow a good distinction between pre-cachectic and non-cachectic patients. 27 New clinical trials testing muscle bioptic and biochemical/metabolic parameters are necessary to develop validated criteria.
Besides diagnostic criteria, more obvious factors may explain why the results of clinical studies on skeletal muscle alterations in cancer cachexia diverge ( Table 1): (i) The heterogeneity of the cohorts, sometimes with several types of cancer and a disproportionate distribution of these types between the cachectic and non-cachectic groups. At this time, we cannot exclude that each cancer can differently regulate cancer cachexia. Additionally, there is a disproportionate distribution of men and women between groups. Sexual dimorphism exists for several skeletal muscle parameters, and hormonal differences may also have a significant impact on cancer cachexia pathophysiology. [108][109][110] (ii) Muscular localization of biopsies. Most samples were collected from the rectus abdominis, but some clinical studies have used biopsies from the quadriceps 47,52,54,61,64,73 or diaphragm muscle. 64 (iii) The number of patients included is often very small and cannot provide sufficient statistical power. Thirteen of 31 clinical studies had a total number of cancer patients ≤20, and there were <30 patients in 20 studies, with a prevalence of cachexia of around 50% most of the time. (iv) Methods of analysis. Some studies have relied on mRNA expression, while others have compared protein expression or enzyme activity. These differences may explain the varying results in clinical studies on muscle proteolysis.
Several blood components have been proposed to aid in the diagnosis of malnutrition/cancer cachexia in patients. Numerous recent clinical studies have observed an association between the components listed hereafter and the cachectic state of patients ( Table 2): (i) markers of inflammation: C-reactive protein; interleukins IL-1, IL-6, IL-8, IL-10; IFN-γ; and TNF-α; (ii) members of the TGF-β family: myostatin, activin A, and GDF-15; (iii) factors derived from the tumour: ZAG (or LMF), VEGF, and Midkine; (iv) lipolysis markers of the adipose tissue: leptin, adiponectin, resistin, free fatty acids, and glycerol; and (v) others: ghrelin, IGF-1, albumin, and angiotensin II. Other factors have also been studied, and they include the proteolysis-inducing factor 121 and markers of muscle degradation: 3-methylhistidine, titin fragments, collagen fragments, 122 testosterone, 123 and the parathyroid hormone-related protein. 124 Nevertheless, there is currently not enough clinical evidence to conclude on their validity. To date, there is no validated biomarker for cancer cachexia.
In addition to variation in the levels of factors discussed before, heterogeneity in blood analysis data may also be due to the fact that concentrations of markers, such as IL6, TNFα, and leptin, may vary between day and intraday. 125 These values may be partly associated to the patient's circadian rhythm and/or metabolic status.
It is interesting to note that C-reactive protein does not appear to be modulated by circadian rhythm or metabolic status. 126,127 Despite this observation, levels of C-reactive protein, which is a well-accepted marker of systemic inflammation, are not consistently increased in populations of patients with cancer cachexia 49,50,57,74,111,112 (Table 2). This observation challenges the use of C-reactive protein in the clinic as a complementary biomarker for cancer cachexia diagnosis.
Considering the conflicting results obtained using muscle biopsies from cachectic patients, no specific biomarker of cancer cachexia has been identified at muscle levels. Myokines, such as myostatin, irisin, activin A, IL-15, FGF21, and GDF-15, were also considered as possible biomarkers and therapeutic targets in cancer cachexia (see review 128 ) ( Table 2). The advantage of dosing muscle cytokines is that it is less invasive than collecting biopsy to evaluate muscle wasting. But more clinical data are needed to confirm this suggestion.
Data extracted from preclinical and clinical studies are also ambivalent. While most preclinical studies observed increased proteolysis and mitochondrial metabolism dysregulations in cachectic rodents, such alterations were not reported in numerous clinical studies on cancer cachexia. Petruzzelli and Wagner hypothesized that these differences were due to the severity of cachexia. 129 Rodent studies are generally carried out during or after rapid and drastic weight loss, while muscle wasting is much more progressive in cachectic patients. The mismatch between the clinical reality and animal models could at least partly contribute to the poor translation observed in the field (see review 130 ).
Even if there are discrepancies in the results of clinical studies, this review suggests the presence of defects in the structure of muscle fibres, increase in proteolysis, decrease in protein synthesis, alterations in lipid metabolism (myosteatosis), and dysfunctions in mitochondrial structure and metabolism (mitochondrial surface and dynamics and mitochondrial DNA) ( Table 1). Taken together, these observations suggest associations between structural and metabolic alterations that may, at least in part, be responsible for the skeletal muscle atrophy characteristics observed in cancer cachexia.
In addition to these parameters, it is interesting to note that endoplasmic reticulum (ER) stress has recently been studied in the context of cancer cachexia (see review 131 ). An increase in the transcript levels (IRE1α, XPB-1, ATF6, and DR5) and in protein levels (CHOP and phospho-eIF2α) of certain ER stress markers has been reported in Lewis lung adenocarcinoma mice. These observations demonstrate that ER stress-induced compensatory unfolding protein response (UPR) is increased in skeletal muscle from cachectic mice. 132 Furthermore, this preclinical study suggests a close relationship between ER stress-induced UPR and muscle proteolysis 132 (see reviews 133,134 ). Furthermore, an increase in gene expression of ER stress-induced UPR markers has recently been reported in the vastus lateralis muscle in cachectic patients with lung cancer. 135 However, cancer cachectic patients were compared with healthy controls in this study. How can the parameters cited previously be associated with muscle wasting? (Figure 1) Myosteatosis is defined by lipid droplets accumulation in muscle. An accumulation of toxic lipid intermediates, such as ceramides and diacylglycerol, in lipid droplets may lead to increased proteolysis and decreased protein synthesis through inhibition of the Akt pathway 136 (Figure 1). These toxic lipid intermediates may also induce ER stress. 136 To restore ER homeostasis, several triggers of the UPR may be activated. They include PERK, which may induce a decrease in protein synthesis through the ATF4 pathway. 137 Intramuscular lipid accumulation may also cause dysfunctions in mitochondrial metabolism, through an inhibition of the electron transport chain and fatty acid β-oxidation. 136 Conversely, alterations in mitochondrial metabolism could limit fatty acid β-oxidation, accentuate the accumulation of lipid droplets, and thereby create a vicious circle. Deterioration of mitochondrial metabolism may also actively participate in muscle wasting via the production of reactive oxygen species. Mitochondria and ER can interact via specific contact sites termed mitochondria-associated ER membranes, which are necessary for several key processes, such as the formation of autophagosomes, the regulation of mitochondrial morphology, dynamics, quality control, calcium transport from the ER to the mitochondria, lipid synthesis and transport, proteolysis, as well as cell death. 138,139 Impairment of mitochondria-associated membrane function could worsen dysfunctions in mitochondrial and lipid metabolisms and participate in the increase proteolysis, thereby inducing atrophy of skeletal muscle fibres and a decrease in skeletal muscle mass.
Skeletal muscle metabolism degradation during cancer cachexia has been poorly investigated in the context of obesity. Cancer-induced skeletal muscle wasting and obesity present common skeletal muscle pathological mechanisms, such as insulin resistance, inflammatory state, and oxidative stress. Indeed, obesity predisposes to a pro-inflammatory state via increased inflammatory mediators, such as TNF-α and IL-6, which stimulate the liver to synthesize and secrete C-reactive protein. It is also associated with reduced levels of adiponectin, which has anti-inflammatory function (see While some scientists discuss the paradoxically protective effect of obesity, 141 due to higher fat reserves, which would preserve from a deleterious weight loss, others highlight the danger of unseen muscle wasting. 142 It has been suggested that the alterations of adipose tissue (when it occurs) precede muscle wasting in the development of cancer cachexia. 143,144 Adipose tissue secretes adipokines, especially in the context of obesity (see review 145 ), which may have a direct effect on skeletal muscle metabolism: leptin, adiponectin, resistin, and visfatin are known mediators of tissue inflammation and insulin sensitivity (see reviews 146,147 ). Skeletal muscle may also release myokines, such as IL-6 and IL-15, which may increase white adipose tissue lipolysis/lipogenesis ratio. 128,148,149 A vicious circle may exist between the increased adiposity and skeletal muscle wasting, especially because of the reciprocal influence exerted by these two tissues (see review 150 ). For instance, adipose tissue lipolysis may induce myosteatosis. In obese patients, basal lipolysis has been reported to be increased in subcutaneous fat cells, because of their larger surface area associated with the secretion of inflammatory cytokines, such as TNF-α. 151 Excessive lipolysis leads to elevated fatty acid availability for skeletal muscle cells. Overloaded muscle fibres may present increased lipid droplet content (size and/or number), which leads to lipotoxic intermediate accumulation. In this context, muscle proteolysis, mitochondrial dysfunctions, and mitochondrial production of reactive oxygen species should be increased, which aggravates skeletal muscle wasting ( Figure 1). Potes et al. demonstrated that excess weight is associated with increased ER stress marker levels in human vastus lateralis. 152 But further studies need to be performed in patients suffering from cancer cachexia to verify the presence or absence of ER stress and its interplay with the other molecular mechanisms involved in the onset and progression of cancer cachexia. Similarly to this adipose tissue-muscle crosstalk, other tissues might affect the skeletal muscle: (i) liver acute-phase proteins, such as serum amyloid A, could participate in activating muscle wasting by enhancing proteolysis (see Figure 1 Summary of the hypothetical associations between metabolic and structural dysfunctions that may, at least in part, be responsible for skeletal muscle atrophy in cachectic cancer patients and for the aggravating role of obesity. In red are represented our hypotheses on the effects of cancer cachexia and in green, our hypotheses on the effects of obesity. An accumulation of lipid droplets, mitochondrial dysfunctions, and endoplasmic reticulum (ER) stress could lead to increased proteolysis, via the ubiquitin-proteasome system (UPS) and autophagy, and decreased protein synthesis, both of which induce atrophy of skeletal muscle fibres and loss of muscle mass. The production of reactive oxygen species (ROS) following mitochondrial alterations may also participate in muscle wasting, by activating apoptotic pathways. In the adipose compartment, lipolysis, which is increased in obese patients, could lead to a greater accumulation of lipid droplets in muscle fibres, mitochondrial dysfunctions, and disturbances in the integrity of mitochondria-associated ER membranes (MAMs), which may worsen skeletal muscle atrophy. review 153 ); (ii) in the context of cancer with bone metastases, osteolysis induces the secretion of activin and TGF-β, which are mediators of muscle proteolysis 154 ; (iii) gut barrier dysfunction leads to increasing intestinal permeability of pathologic bacteria and endotoxemia and perpetuates systemic inflammation known to drive muscle wasting 155 ; and (iv) interactions between systemic and brain inflammation 156 alter the activity of the hypothalamus, responsible for the regulation of anorexia 157 and for the secretion of glucocorticoids, which stimulate proteolysis in skeletal muscle. 158

Conclusion
The overall result of clinical studies suggests the presence of alterations in skeletal muscle from cachectic cancer patients, and these alterations could affect myofibers structure, proteolysis, and protein synthesis pathways, as well as lipid and mitochondrial metabolisms. However, evidences are still weak because clinical studies have often obtained contradictory results. Additional clinical studies are essential to further the exploring and understanding of alterations underlying skeletal muscle wasting characteristic of cancer cachexia. Moreover, an improvement and better standardization of the diagnostic criteria for cancer cachexia is crucial to ameliorate the robustness of the conclusions.