Role of Activin A and Myostatin in Human Cancer Cachexia

Abstract

Context:

Cachexia is a multifactorial syndrome, characterized by the loss of skeletal muscle mass and not fully reversible by nutritional support. Recent animal observations suggest that production of Activin A (ActA) and Myostatin (Mstn) by some tumors might contribute to cancer cachexia.

Objective:

Our goal was to investigate the role of ActA and Mstn in the development of the human cancer cachexia.

Design/Setting:

The ACTICA study is a cross-sectional study, which prospectively enrolled patients from a tertiary-care center between January 2012 and March 2014.

Subjects/Outcome Measures:

One hundred fifty two patients with colorectal or lung cancer had clinical, nutritional and functional assessment. Body composition was measured by CT-scan, anthropometry, and bioimpedance. Plasma concentrations of ActA, Mstn, and Follistatin were determined.

Results:

Cachexia was associated with reduced lean and fat mass (p < .01 and p < .001), reduced physical function, lower quality of life, and increased symptoms (QLQC30; p < .001). Anorexia (SNAQ score < 14) was more common in cachectic patients (CC) than in noncachectic patients (CNC) (p < .001). ActA concentrations in CC patients were higher than in CNC patients (+40%; p < .001) and were correlated positively with weight loss (R = 0.323; p < .001) and negatively with the SNAQ score (R = −0.225; p < .01). In contrast, Mstn concentrations were decreased in CC patients compared to CNC patients (−35%; p < .001).

Conclusions:

These results demonstrate an association between circulating concentrations of ActA and the presence of the anorexia/cachexia syndrome in cancer patients. Given the known muscle atrophic effects of ActA, our study suggests that increased circulating concentrations of ActA may contribute to the development of cachexia in cancer patients.

Cachexia is a complex multifactorial syndrome associated with underlying illness and characterized by loss of skeletal muscle and fat mass. In contrast to malnutrition, cachexia cannot be reversed simply by increasing nutritional intake (1). Estimates reveal that cachexia is observed in up to 80% of patients with advanced cancer and accounts directly for 20% of cancer-related deaths (2). Cachexia is associated with functional impairment, altered quality of life (QOL) and reduced tolerance and response to anticancer therapies (1, 3). More importantly, loss of skeletal muscle mass in advanced cancer is recognized as an independent predictor of mortality (4). Thus, cachexia represents a major challenge in the care of cancer patients. Nevertheless, the precise molecular mechanisms of cancer cachexia remain poorly characterized. It is thought to result from a combination of reduced food intake and abnormal metabolism, induced by both tumor- and host-derived factors (2).

Activin A (ActA), a member of the TGFβ superfamily, is a homodimer of β-Activin chains which is produced by several cell types and has a broad spectrum of biological effects (5). ActA is present in the circulation where its concentrations are increased in acute inflammation (6, 7), acute respiratory failure (8), renal failure, and in some cancers mainly with bone metastases (9, 10). ActA exerts most of its biological actions by binding to the membrane Activin type II receptor B (ActRIIB) (5), a receptor shared with Myostatin (Mstn), another TGF-β superfamily member, which is a potent negative growth factor of muscle mass (11, 12). It is therefore considered that ActA reproduces the biological action of Mstn on skeletal muscle. Many factors regulate ActA bioactivity, including Inhibin and Follistatin (FS). This last one may be considered as the major regulator of circulating ActA and activity (5). Mstn bioactivity is mainly regulated by its propeptide and Follistatin (12).

Recent works indicate that ActA and Mstn contribute to skeletal muscle atrophy observed in several animal models of tumor-induced cachexia. Indeed, Inhibin-deficient mice, characterized by high concentrations of circulating ActA, exhibit a loss of skeletal muscle and fat mass leading to death (13). Even in the absence of underlying diseases or tumor, increasing local or circulating concentrations of ActA induces muscle atrophy, supporting the muscle atrophic effect of ActA by itself (13–15). On the other hand, muscle overexpression (16) or systemic administration of Mstn induces skeletal muscle atrophy (17). Moreover, Mstn expression is upregulated in the muscle of animals with tumor-induced cachexia (13). Interestingly, in an animal model of cancer cachexia, the blockade of ActA and Mstn by a soluble form of its receptor (sActRIIB) prevents muscle atrophy and increases survival, without affecting tumor growth (13, 18). In humans, increased ActA circulating concentrations are observed in cancer patients (9, 10). Furthermore, high circulating levels of Mstn were observed in muscle wasting conditions (19–22). Finally, several human tumor cell lines secrete ActA and Mstn in vitro (13, 23)

However, the role of ActA and Mstn in the development of cancer cachexia has poorly been investigated in humans. This hypothesis seems particularly attractive to be tested, since inhibitors of ActA and Mstn are currently under clinical investigation. The aim of this study was therefore to investigate the role of ActA and Mstn in the development of the human cancer cachexia, in order to identify a new biomarker predictive of cachexia and to eventually select patients susceptible to benefit from ActA and Mstn antagonists treatment.

Patients and Methods

Study design

This prospective study was performed at the Cliniques Universitaires Saint-Luc, Brussels, Belgium. The protocol was approved by the local ethical committee of the Université Catholique de Louvain. Patients with colorectal or lung cancer, confirmed by anatomopathology, were recruited at the diagnosis or at relapse, before any therapeutic intervention, from January 2012 to March 2014. This was a cross-sectional study of cancer patients with or without cachexia. Cachexia was defined, according to the definition proposed by Fearon et al, as an involuntary weight loss > 5% over the past 6 months or weight loss > 2% and BMI < 20 kg/m² or weight loss > 2% and low muscularity (LM) (1).

Eligibility criteria

Subjects were at least 18 years old, had an expected survival of more than 3 months, and no previous history of any other cancer in the last 5 years. Written consent was given prior to entry into the study. Exclusion criteria were non-Caucasian subjects, obvious malabsorption, major depression, artificial nutrition, high doses of steroids (> 1 mg/kg hydrocortisone equivalent), hyperthyroidism, other causes of malnutrition, major walking handicap, ECOG performance status ≥ 4 and psychological, familial, social or geographic conditions that would preclude participation in the full protocol.

Study measures

Body weight loss

Current body weight and height were measured at the inclusion. Body weight change during the previous 6 months was calculated and expressed as a percentage of preillness body weight. For relapsed patients, the median of the cancer-free period was 7 (3–240) months and the weight loss was calculated by comparing prerelapse and current weights.

Nutritional and functional assessment

The nutritional intake was evaluated by the Simplified Nutritional Appetite Questionnaire (SNAQ) score. Anorexia was defined by a SNAQ score < 14 (24). The functional status was assessed by two previously validated scales, namely the Eastern Cooperative Oncology Group (ECOG) and EORTC QOL questionnaire (QLQ-C30) (25).

Skeletal muscle and fat mass measurement

Skeletal muscle and fat mass were assessed by abdomen CT-scan (CT), anthropometry, and bioelectrical impedance (BIA).

CT scans used for the analysis were performed for standard cancer care between 3 months before and 1 month after the inclusion date and before any therapeutic intervention. A transverse CT image from the third lumbar vertebrae (L3) was analyzed for each patient and tissue area estimated, using previously described Hounsfield unit (HU) thresholds and quantified by the Slice-O-Matic software, version 4.3 (Tomovision). The most common and accepted HU range for muscle tissue is −29 to +150 HU and for adipose tissue is −190 to −30 HU. Cross-sectional area for muscle and adipose tissues was normalized for stature and expressed as muscle and fat indexes (cm²/m²) (26, 27). All CT images were analyzed by a single trained observer, who was blinded to the patient's status. The intra-observer coefficient of variation was 2.2%.

The anthropometric assessment included the measurement of triceps skinfold (TSF) thickness (cm) using a Holtain Ltd. caliper (Crymych) and midarm circumference (cm) together to calculate the midarm muscle area (MAMA) (cm²) (28). Skeletal muscle mass was estimated by the MAMA ratio (MAMA/MAMA 50th percentile for age and sex) and fat mass by the TSF ratio (TSF/TSF 50th percentile for age and sex) (28).

Body composition was also assessed by bioimpedance (BIA) using the BIA101 device (Akern) and the software provided by the manufacturer. Whole body fat-free mass and fat mass indexes (kg/m²) were obtained using the in-built equations (29).

The cutoffs for low muscularity were those proposed by Fearon et al (1).

Muscle function and quality

Skeletal muscle strength was assessed with a Jamar hand-held dynamometer. Three measures were made on the nondominant side in a time interval of 30 seconds. The highest value was retained. Muscle strength ratio was calculated as hand grip strength (kg)/hand grip strength (kg) 50th percentile for age and for sex (28).

Skeletal muscle quality was assessed by skeletal muscle density (SMD) based on the mean muscle radiation attenuation for the entire muscle area at the third lumbar vertebra level, measured by a CT scan in HU. This parameter is inversely related to muscle fat content and is linked to reduced muscle strength and performance. The generally accepted lower limit of normal mean attenuation for muscle is 30 HU, a value defined as two standard definitions below the mean attenuation value of muscles of young healthy persons (27, 30).

Biomarkers measurement

Blood samples were collected from patients at the time of recruitment, in standardized conditions. Total plasma ActA, Mstn, and FS were measured by solid-phase two-sites enzyme linked immunoassays (R&D Systems) according to the recommendations of the manufacturer. Nutritional and inflammatory markers (albumin, prealbumin, and CRP) were determined by clinical routine methods in our Clinical Chemistry Department.

Statistical analysis

Comparisons between groups were performed using the nonparametric Mann Whitney U-test for continuous parameters and by Fischer Exact test for categorical variables. Correlations were estimated using Spearman's Rank correlation coefficient. Data are expressed as median (min-max). The program SAS® version 9.3. was used for all the statistical tests. Statistical significance was set at P < .05.

Results

Patient and tumor characteristics

Among 700 patients who were screened for this study, 356 patients met the initial inclusion criteria. From the 356 patients, 193 patients had one or more exclusion criteria and 11 patients refused to participate at the study. Thus, a total of 152 patients with colorectal (n = 94) or lung cancer (n = 58) were recruited at the time of diagnosis (n = 125) or at relapse (n = 27). The patient median age was 67 (26–95) years with a predominance of male subjects (57%) (Table 1). Most patients had a high tumor stage T3 or T4 (73%). The prevalence of cachexia reached 49%. Sixty-four patients had a weight loss > 5%, 10 patients had a weight loss > 2% and a BMI < 20 kg/m², and 39 patients had a weight loss > 2% with a low muscularity. Since the same patient could fit into more than one single criteria, the total number is higher than 74, the number of cachectic patients in our study. The two groups of patients (cancer cachectic or CC vs cancer noncachectic or CNC) were comparable for age, sex ratio, and usual BMI. The presence of cachexia was unrelated to the cancer site, the tumor staging and the presence of metastases. The CNC group counted a greater proportion of relapse than the CC one (P < .01).

Table 1.

Patients and Tumor Characteristics

Characteristics	Total	CNC	CC	P value (CC vs CNC)
n	152	78	74
Age (y)	67 (26–95)	66 (26–86)	69 (31–95)	NS
Sex ratio (F/M) (%)	43/57	38/62	49/51	NS
Usual BMI (kg/m2)	26 (16–47)	27 (16–47)	26 (18–42)	NS
Current BMI (kg/m2)	25 (16–47)	27 (16–47)	23 (16–39)	<.001
Weight loss (%)	3.6 (0.0–25.0)	0.0 (0.0–4.0)	8.0 (2.0–25.0)	<.001
Tumor nature				NS
Colorectal (%)	62	65	58
Lung (%)	38	35	42
Type				<.01
Primitive (%)	82	73	92
Relapse (%)	18	27	8
Stage
T 1/2/3/4 (%)	11/16/46/27	12/21/47/20	10/12/44/34	NS
N 0/1/2/3 (%)	47/27/16/10	54/30/11/5	41/25/21/13	NS
M 0/1 (%)	64/36	73/27	56/44	NS

Usual BMI was calculated as preillness weight (kg)/height (m²) for new diagnosis patients and prerelapse weight (kg)/height (m²) for relapsed patients. One hundred twenty five patients had a staging of cancer with TNM classification, 57 patients for CNC, and 68 patients for CC. NS, not significant.

Nutritional and functional characteristics

The SNAQ score was significantly lower in CC than in CNC patients (P < .001) and anorexia was more common in the CC compared to the CNC group (P < .001) (Table 2). Cachexia was associated with reduced physical function (ECOG and QLQC30; P < .001), reduced QOL (QLQC30, P < .001), and increased symptoms (QLQC30; P < .001). The SNAQ score was positively correlated with nutritional status parameters, albumin, and prealbumin (respectively, R = 0.324 and R = 0.405; P < .001), physical function (QLQC30, R = 0.387; P < .001) and QOL (QLQC30, R = 0.429; P < .001), and negatively correlated with weight loss (R = −0.520; P < .001) and symptoms (QLQC30, R = −0.525; P < .001).

Table 2.

Nutritional and Functional Characteristics

Characteristics	Total	CNC	CC	P value (CC vs CNC)
N	152	78	74
SNAQ score	15 (6–20)	16 (10–20)	14 (6–19)	<.001
Anorexia (%)	25	10	41	<.001
ECOG				<.001
0 (%)	59	74	43
1 (%)	32	23	42
2 (%)	6	3	8
3 (%)	3	0	7
QLQ-C30
Function	77 (18–100)	84 (33–100)	70 (18–97)	<.001
Symptoms	17 (0–69)	13 (0–59)	27 (0–69)	<.001
Quality	66 (0–100)	66 (16–100)	58 (0–100)	<.001

Anorexia was defined by a SNAQ score < 14. ECOG was scored from 0 to 3, with 0 representing the best function. QLQ-C30 was scored from 0 to 100, with 100 representing best function and higher quality of life and 0 representing fewer symptoms.

Skeletal muscle and fat mass assessment

As expected, the skeletal muscle mass was lower in CC patients compared with CNC patients independently of the technique used to assess body composition (−6% by CT; P < .01, −9% by anthropometry; P < .01 and −11% by BIA; P < .001) (Figure 1). The fat mass was also lower in CC patients compared with CNC patients (−36% by CT; P < .001, −21% by anthropometry; P < .01 and −21% by BIA; P < .001).

Figure 1.

Skeletal muscle (A) and fat (B) mass, assessed by CT-scan (n = 139), muscle strength (C) and skeletal muscle density (D) (n = 139) in cachectic patients (CC) compared to noncachectic patients (CNC).

**, P < .01; ***, P < .001.

Skeletal muscle function

CC patients had less muscle strength, as evidenced by a lower hand grip strength ratio, compared with CNC patients (1.01 vs 1.15; P < .001) (Figure 1). The SMD tended to be lower in CC than in CNC patients (29.4 [6.2–54.8] vs 35.2 [0.2–62.2] HU; P = .05), reflecting a greater fatty infiltration of skeletal muscle in CC patients. Muscle strength was correlated with muscle index (R = 0.637; P < .0001) but also with SMD (R = 0.427; P < .0001).

Biomarkers

Interestingly, plasma ActA concentrations were higher in CC patients compared with CNC patients (558 pg/mL [228–17 660] vs 397 [165–2731]; P < .001), whereas concentrations of FS, the main physiological inhibitor of ActA, were not different between the two groups (2320 pg/mL [778–7534] vs 2094 [1134–3844]; NS) (Figure 2) (Table 3). Moreover, ActA concentrations above the normal range (115–665 pg/mL) were observed three times more often in CC patients than in CNC patients (38% vs 13%; P < .001). ActA concentrations were also higher in anorectic patients than in nonanorectic patients (583 pg/mL [194–9402] vs 420 [165–17 660]; P < .01). Although patients with elevated ActA had more frequently metastases (68% vs 26%; P < .001), we did not see any difference between patients with elevated and normal ActA concentrations regarding cancer type and staging of tumor (score T and N). In contrast, plasma Mstn concentrations were decreased in CC patients compared to CNC patients (1371 pg/mL [167–4989] vs 2109 [715–4907]; P < .001). In addition, Mstn concentrations below the normal range (1220–7300 pg/mL) were observed more often in CC patients than in CNC patients (36% vs 14%; P < .01). Patients who exhibited a lowered Mstn concentration had more frequently an invasive tumor (score T) (P < .01), but we did not see any difference between patients with lowered and normal Mstn concentrations regarding the nature of tumor, the score N of the tumor, and the presence of metastases. As expected, CC patients had a worse nutritional status and a more severe degree of inflammation than CNC patients, as reflected by lower albumin and prealbumin and higher CRP concentrations.

Figure 2.

Circulating Activin A (A) and Myostatin (B) concentrations in cachectic patients (CC) compared to noncachectic patients (CNC).

Dotted gray lines represent normal ranges, according to the manufacturer (R&D Systems). ***, P < .001.

Table 3.

Biomarkers

Characteristics	Total	CNC	CC	P value (CC vs CNC)
N	152	78	74
Activin A (pg/mL)	455 (165–17 660)	397 (165–2731)	558 (228–17 660)	<.001
Follistatin (pg/mL)	2174 (778–7534)	2094 (1134–3844)	2320 (778–7534)	NS
Myostatin (pg/mL)	1785 (167–4989)	2109 (715–4907)	1371 (167–4989)	<.001
CRP (mg/dL)	0.4 (0.0–25.7)	0.3 (0.0–10.3)	1.3 (0.1–25.7)	<.001
Albumin (g/dL)	4.4 (2.8–5.1)	4.5 (3.0–5.1)	4.2 (2.8–5.0)	<.001
Pre-albumin (mg/dL)	20.8 (2.2–49.5)	23.7 (3.5–40.2)	16.8 (2.2–49.5)	<.001

NS, not significant.

ActA concentrations were positively correlated with weight loss, QLQ-C30 symptoms (R = 0.326; P < .001) and with CRP concentrations and were negatively correlated with the SNAQ score, QLQ-C30 QOL and function (respectively, R = −0.287; P < .001 and R = −0.260; P < .01), albumin and prealbumin concentrations (Table 4). ActA concentrations were not correlated with muscle index but were correlated with parameters, which reflect skeletal muscle function, including muscle strength and muscle density. In contrast with ActA, Mstn concentrations were positively correlated with muscle index, muscle density, muscle strength, albumin, prealbumin, and negatively correlated with weight loss and CRP (Table 4). CRP was positively correlated with weight loss (R = 0.350; P < .001).

Table 4.

Correlations Estimate With Circulating Activin A and Myostatin Concentrations

Dependent Variable	Age (y)	Independent Variables
		WL (%)	SNAQ	M. index (cm2/m2)	SMD (HU)	Strength (kg)	CRP (mg/dL)	Albumin (g/dL)	Prealbumin (mg/dL)
Activin A (pg/mL)
Spearman's Rank correlation coefficient	0.361	0.323	−0.225	−0.026	−0.225	−0.207	0.481	−0.565	−0.566
P value	<0.001	<0.001	<0.01	NS	<0.01	<0.05	<0.001	<0.001	<0.001
Myostatin (pg/mL)
Spearman's Rank correlation coefficient	−0.097	−0.354	0.401	0.404	0.353	0.476	−0.429	0.337	0.321
P value	NS	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001

Abbreviations: WL, weight loss was calculated by comparing pre-illness or prerelapse weights and current weights; M. index, muscle index representing muscle mass, assessed by a CT scan.

Discussion

Our study demonstrates an association between the anorexia/cachexia syndrome and ActA circulating concentrations in cancer patients. Given the anorectic and cachectic effects of ActA, recently established in animal models (13), our study suggests that increased circulating concentrations of ActA in cancer cachectic patients may contribute to the development of this syndrome.

Increased circulating concentrations of ActA in cancer patients, mainly with bone metastases, have previously been reported (9, 10). Furthermore, high circulating levels of ActA have been associated with a poor prognosis in patients with myeloma (31). However, the link between ActA and cachexia has never been investigated in these studies. Therefore, to the best of our knowledge, our study is the first to assess the relationship between ActA circulating concentrations and cachexia in a population of cancer patients.

Our observations reveal that ActA circulating concentrations are increased in cachectic patients, whereas the concentrations of FS, its main physiological inhibitor, are similar in CC and CNC subjects. These data suggest that cachexia is associated with an increase of biologically active ActA. This hypothesis is further supported by the demonstration of increased phosphorylation (32) or levels (33) of Smad2/3 in the skeletal muscle of cancer patients. Indeed, ActA has been demonstrated to promote muscle wasting in rodents via the Smad2/3 signaling pathway (13, 14). Therefore, high biologically active ActA concentrations might contribute to muscle atrophy observed in cachectic patients. Although the mechanisms of the muscle atrophic effect of ActA have not yet been defined, both inhibition of protein synthesis and increased protein breakdown seem to be involved (13, 14). These observations are consistent with the model that increased ActA signaling plays a causative role in the development of muscle atrophy and cachexia in humans.

Skeletal muscle strength was reduced in our cachectic patients and negatively correlated with ActA concentrations, suggesting a role for increased ActA in the decreased muscle function in CC patients. This interpretation is supported by works performed with animal tumor-induced cachexia models, where a treatment with ActA antagonists restores muscle mass and performance (18). Even without any disease or tumor, ActA may cause reduced muscle performance as transfection of ActA gene into muscle is able to decrease muscle mass and strength (14). Muscle function in cachexia can be reduced independently of mass due to increased lipid deposition in skeletal muscle (34). Excess fat infiltration of muscle can translate in a decreased skeletal muscle density (SMD) at the CT scan measurement (27, 30). In our population, we indeed observed a positive correlation between muscle strength and SMD and CC patients tended to exhibit a lower SMD compared to CNC patients, suggesting a greater fat infiltration of the skeletal muscle. Although the role of ActA in the fat infiltration of muscle is not yet established, ActA concentrations were inversely correlated to SMD. These data may suggest that ActA impairs muscle quality, independently of its effect on mass. The clinical relevance of this observation is illustrated by recent results indicating that a low SMD is predictive of poor prognosis in cancer patients (35).

ActA concentrations were higher in anorectic compared to nonanorectic patients and were negatively correlated to the SNAQ score, suggesting a role of ActA in the development of anorexia in cancer cachexia. Supporting this hypothesis, increasing circulating concentrations of ActA by implantation of ActA-producing CHO cells or by Inhibin-α gene knock-out in rodents induced weight loss but also anorexia with reduced food intake, which were prevented by sActRIIB treatment (13). Although the mechanism of the anorectic effect of ActA remains to be established, the decreased fat mass observed in CC patients may result in part from negative energy balance caused by reduced food intake, as illustrated by the lower SNAQ score and the higher prevalence of anorexia in these patients.

Systemic inflammation is thought to be a major mediator of cancer cachexia. Based on animal models, inflammation seems most likely to contribute to the increase of ActA concentrations (36). Many human observations fit with this hypothesis as ActA concentrations are increased in the circulation of patients with inflammatory conditions (6, 7). This view is also supported by the correlation between ActA and inflammatory markers (CRP, albumin, prealbumin) observed in our study. Several research groups have identified some other proinflammatory cytokines (IL-6, TNF-α, TWEAK, …) which may contribute to cachexia in animal models (2). IL-6 in particular has been shown to correlate with weight loss and with reduced survival in cancer patients (37, 38) and also with markers of inflammation, such as CRP (39). Although we did not assess IL-6 in this study, we showed an increase of CRP levels in cachectic compared to noncachectic patients and a correlation between CRP levels and weight loss. Taken together, these data suggest indirectly that IL-6 levels are probably increased in our cachectic patients and are correlated with weight loss. Nevertheless, treatments with anti-TNF-α or anti-IL-6 antibodies provide not much benefit on muscle mass in patients with cancer cachexia (40, 41). In contrast, in murine models of cancer cachexia, treatment with sActRIIB allowed the reversal of cachexia without affecting circulating levels of TNF-α and IL-6 (13). These data suggest a dominant role of ActA over the classic proinflammatory cytokines in the development of cancer cachexia.

Although Mstn can cause muscle atrophy, plasma Mstn concentrations were reduced in our CC patients compared to CNC patients. Most likely, this data suggests that circulating Mstn does not play a major role in the cancer cachexia development. Our observation is in the line of a recent report, who found in cancer patients a decrease of circulating Mstn propeptide levels, reflecting a decrease in Mstn production (22). This decrease could be interpreted as a protective down-regulation in response to muscle wasting. If Mstn acts as a chalone produced by skeletal muscle to control its mass (42), its production is expected to decline in response to insults causing decrease of muscle size. This is consistent with our data that show a positive correlation between plasma Mstn concentrations and muscle index. But, other surmises could nuance our interpretation. In one hand, we measured total Mstn concentrations and not only the biologically active Mstn portion. On the other hand, Mstn, which is mainly produced by skeletal muscle, has perhaps rather an autocrine and paracrine than an endocrine action on muscle, as suggested by increased expression of muscle Mstn in different animal models of cancer cachexia (13, 43). Thus, the circulating concentrations of Mstn might not reflect its real activity at the muscle level.

Now that circulating ActA and not Mstn levels have been demonstrated correlated which cachexia, future longitudinal studies will have to delineate the predictive value of circulating concentrations of ActA in terms of development of muscle atrophy, weight loss and survival, in comparison to other biomarkers such as IL-6 and TNFα. The independent correlations that we observed between ActA concentrations and different well-established prognostic factors (Table 4) provide hopeful arguments in this direction.

Our study has several limitations. It was an observational, cross-sectional study and for this reason it was not possible to document active muscle loss, whereas proportion of skeletal muscle vary widely in the population. On the other hand, we assessed only two types of cancer, albeit the most prevalent in the Caucasian population, whereas it is known that mechanisms of cachexia can vary according to tumor type, site, and mass (2). It would be interesting to extend these analyses in other types of cancer (pancreas, …). However, it is worth noting that the presence of cachexia in our work was unrelated to the cancer site, the tumor staging, and the presence of metastases. Finally, given our inclusion criteria, which comprised an expected survival more than 3 months, we excluded patients with refractory cachexia (1), who might have had high levels of circulating ActA.

In conclusion, the anorexia/cachexia syndrome in cancer patients is associated with increased circulating ActA concentrations. Given the anorectic and cachectic effects of ActA, increased circulating concentrations of ActA may contribute to the development of cachexia in cancer patients. The clinical perspectives of our study are double. On the one hand, future longitudinal studies will have to delineate the predictive value of ActA as a marker of cachexia and survival. On the other hand, interventional studies will have to investigate whether blocking this pathway preserves muscle mass and hence improves QOL and survival in cachectic cancer patients.

Acknowledgments

We especially thank Mrs Anne Moxhon, the general coordinator for clinical research at the Institute, for her valuable advice during the study. We thank also Professor Dominique Maiter for revising the manuscript critically and for continuous support. In addition, we thank research collaborators M. Hamoir, J.P. Machiels, Y. Humblet, M. Van den Eynde, A. Kartheuser, D. Leonard, C. Remue, A. Poncelet, V. Lacroix, T. Pieters, and P. Collard. Finally, this work would not have been possible without the participation and the cooperation of patients, whom we thank cordially.

This work was supported by the Cancer Plan, Belgian Ministry of Public Health (FPS Health), the Belgian Foundation against Cancer, the Saint-Luc Foundation, and the Fonds de la Recherche Scientifique Médicale (FNRS-FRS), Belgium. This work was performed in the Endocrinology and Nutrition Department with the collaboration of the King Albert II Cancer Institute of the Cliniques Universitaires Saint-Luc, Brussels, Belgium.

Disclosure Summary: The authors have nothing to disclose.

ClinicalTrials.gov number: NCT01604642

Abbreviations

 
• ActA

  Activin A

 
• ECOG

  Eastern Cooperative Oncology Group

 
• FS

  Follistatin

 
• MAMA

  midarm muscle area

 
• QOL

  quality of life

 
• SMD

  skeletal muscle density

 
• SNAQ

  Simplified Nutritional Appetite Questionnaire

 
• TSF

  triceps skinfold.