High Level of Irisin as a Marker of Malnutrition in Head and Neck Cancer Patients Subjected to Radiotherapy

Background

Head and neck cancers (HNC) are the 7th most prevalent neoplasms in the world. Among them, the majority (about 90%) are squamous cell carcinomas (SCCHN). In 50% of these patients, body weight loss and malnutrition occur even before the beginning of therapy. The main methods of SCCHN treatment are surgery, radiotherapy (RTH), chemotherapy (CTH), and combination of these methods. It is estimated that in the course of therapy the proportion of malnourished patients increases to 90%. Nutritional problems in patients with SCCHN are multifactorial. The definition of malnutrition includes progressive loss of muscle and fat tissue. Nutritional problems involve socio-psychological factors, complications of treatment (including dysphagia, xerostomia), or symptoms associated with tumor location. The symptoms associated with tumor location include gastrointestinal obstruction and odynophagia [1,2]. In patients with neoplastic malnutrition or cachexia, various degrees of body weight loss may occur in association with metabolic disturbances, lean body mass loss, and increased lipolysis [2,3]. It should be noted that not all cancer patients, including those with the tumors of the head and neck area (at the same stage of disease, subject to the same treatment method), develop malnutrition or cachexia [3]. Early detection of nutritional disturbances, especially before significant body weight loss, is associated with improved quality of life, longer time to disease progression, and better overall survival (OS) [2,4]. Identification of patients at risk of nutritional disorders may allow for the implementation of appropriate nutritional treatment. Therefore, the search for predictors of malnutrition and/or cachexia seems to be warranted.

Cancer cachexia is not always present in malnourished patients but all cachectic patients experience nutritional disturbances4[5]. Cancer cachexia is a multifactorial syndrome characterized by appetite disorders, body weight loss, metabolic changes, and intensified inflammation [5,6]. Lean and fatty body mass atrophy and skeletal muscle function loss are also observed. These changes lead to steady body weight loss [5]. Knowledge concerning the pathomechanism of malnutrition and cachexia has been expanding in the recent years. However, its complexity still leaves many aspects of these conditions unexplained. It is known that an important role is played by the development of inflammation, degradation of muscle fibers, and browning of white adipose tissue (WAT). One of the most important myokines is irisin, which is a fragment of extracellular domain of FNDC5 protein [5,7]. Irisin is regulated by proliferator-activated receptor-γ coactivator-1α (PGC1α) and the muscles can excrete it into the bloodstream. As a result, irisin can activate adipose tissue thermogenesis and cause browning of WAT [8]. The relationship between irisin and neoplasms is unclear. The level of irisin in patients with different types of cancer may be lower or higher compared to healthy people [9–11]. However, the major aim of research has not been to correlate the level of irisin with the nutritional status of these patients. Perhaps the differences in irisin levels were due to the nutritional status of these patients [11]. Therefore, because irisin is involved in hydrolysis triacylglycerols in WAT and inhibits de novo lipogenesis [12], we hypothesized that irisin may be associated with the development of malnutrition or cachexia in patients with HNC.

The aim of the study was to evaluate the level of serum irisin in HNC patients subjected to RTH and to assess its correlation with the occurrence of malnutrition.

Material and Methods

STATISTICAL ANALYSIS:

All data were collected in the Excel database and then imported into the MedCalc statistical program (MedCalc Software, Belgium). Since there were no similar studies comparing irisin levels in patients with different nutritional statuses, the sample size was calculated in the acquired data set retrospectively. We decided to use NRS status as a primary outcome. As input data, we used arithmetic means and corresponding standard deviations (SD) of irisin serum concentration. In most medical studies, P values below 0.05 are used to reject the null hypothesis, thus a type I error (alpha) of 5% was used. To achieve 80% of statistical power for type II error, we set a cut-off of beta on 0.2. Considering the calculated difference in means, which was equal to 0.89, SD for NRS<3 group equal to 0.30, SD for NRS≥3 equal to 1.21, and the ratio of sample sizes of compared groups equal to 2.125, the minimal sample size of the study group was estimated as 50. Data were expressed as a percentage (for categorized variable), median, and 95% confidence interval (CI) (since continuous variables had not normal distribution). We considered P values below 0.05 as statistically significant. Because data were not normally distributed, we used the nonparametric Mann-Whitney U test for comparison of irisin level according to selected categorical variables, as well as Spearman’s correlation test for the evaluation of the correlation between irisin and other continuous variables. Odds ratio (with 95% CI) was calculated to evaluate the risk of abnormal nutrition status (according to SGA scale) or nutritional risk (NRS scale). In addition, we used the analysis of ROC curves (with the area under curve - AUC and its 95% CI) to assess the diagnostic usefulness of irisin in the differential diagnosis of patients with normal and abnormal nutritional status (SGA) and with low and high nutritional risk (NRS).

Results

No statistically significant differences in the level of irisin were observed in relation to the presence of the analysed demographic and clinical factors (Table 2). However, higher irisin values (1.57 vs 1.18 [ng/ml], P=0.0004) were observed in patients with higher nutritional risk (≥3) evaluated according to NRS scale. Similarly, in patients assessed as B or C on the SGA scale, higher values of irisin concentration (1.38 vs 1.07 [ng/ml], P=0.0139) were noted. Moreover, higher level of irisin (1.56 vs 1.27 [ng/ml], P=0.0416) was observed in patients in whom parenteral nutrition was implemented. A trend was noticed toward significantly higher concentrations of irisin in underweight patients (BMI<18.5 vs >18.5: 2.53 vs 1.33, P=0.0585). In the remaining studied factors, no statistically significant differences were observed. The influence of selected factors associated with the evaluation of the nutritional status on the level of irisin is presented in Table 3. It was also observed that the level of irisin before treatment was correlated negatively (rho=−0.30, p=0.0350) with FM% and positively (rho=0.30, p=0.0340) with FFM% in the case of BIA measurements performed after the 7th cycle of RTH. A similar correlation was not found in with absolute values of these parameters or with the remaining studied variables (Table 4). Among the studied variables, only the performance status (PS) and the level of irisin significantly influenced the risk of occurrence of malnutrition assessed according to the SGA scale. A 20-fold lower risk of occurrence of severe malnutrition (SGA C) was noticed in patients with better performance status (PS<2; OR=0.05, 95% CI: 0.01–0.98, P=0.0487), whereas low level of irisin was associated with a 10-fold lower risk of moderate or severe malnutrition (SGA B or C; OR=0.11, 95% CI: 0.01–0.95, P=0.0449). Similarly, in the case of nutritional risk assessed using NRS scale, the only factor that affected it was the level of irisin. In patients with low level of irisin, a 5-flod lower nutritional risk was observed (NRS≥3; OR=0.21, 95% CI: 0.05–0.78, P=0.0197). The data on the influence of a series of variables on the risk of malnutrition assessed according to SGA or NRS scale are presented in Tables 5 and 6. On the basis of analysis of ROCs, it was observed that the level of irisin (with the cut-off point >1.24) significantly differentiated patients with normal nutritional status assessed according to SGA scale (A) from the ones with moderate or severe risk of malnutrition (B or C) (sensitivity: 64.3%, specificity: 87.5%; AUC=0.777 [0.637–0.882]; P=0.0029; Figure 1). This marker was similarly useful (with the cut-off point >1.43) in the differentiation of patients assessed using SGA scale as normally nourished or moderately undernourished (A or B) from these with severe malnutrition (C) (sensitivity: 55.6%, specificity: 75%; AUC=0.660 [0.512–0.788]; P=0.0487; Figure 2). The NRS scale also significantly differentiated the level of irisin (with the cut-off point >1.23) between patients with high and low nutritional risk (NRS>3 vs <3) (sensitivity: 93.7%, specificity: 58.8%; AUC=0.811 [0.675–0.908]; P<0.0001; Figure 3).

Discussion

Many factors influence the development of malnutrition and cachexia. The initial mechanism leading to hypercatabolism is a systemic inflammation characterized by, among other features, increased secretion of pro-inflammatory cytokines (interferon-γ, TNF-α, NFκβ, IL), catabolic mediators and reactive forms of oxygen, both by the cells of the host and cancer cells [16,17]. Chronic inflammation can cause changes in the metabolism of carbohydrates, fats, and proteins. The main changes in the metabolism of carbohydrates result from the lack of tolerance to glucose, resistance to insulin, and increase of gluconeogenesis. In patients, we can observe an increase of factors, which directly cause lipolysis (including the lipid-mobilizing factor (LMF) and zinc-alpha-2glycoprotein) and raise sensitivity to lipolytic factors, thus contributing to the loss of fat tissue [7,16]. However, one of the main reasons for the decrease of FFM (mainly muscular) is the decreased synthesis of new proteins with simultaneous increase of degradation of the existing ones [7,16]. The observed decrease of body weight in malnourished and cachectic patients is mainly associated with lipolysis (mostly WAT) in response to negative energetic balance and anorexia related to the development of cancer [18]. The influence of lipolysis on the development of cancer cachexia was demonstrated in a study in which the inhibition of lipid mobilization improved the patients’ condition [19]. Lipolysis of WAT causes loss of adipose tissue and contributes to decreased mass of skeletal muscles. In turn, WAT browning increases the expression of uncoupling protein 1 (UCP1), which leads to increased energy expenditure. Likewise, the activated BAT increases the expression of UCP1 and also increases energy expenditure. Both these processes cause negative energy balance, which contributes to the development and progression of malnutrition and cancer cachexia [19]. PGC1α Participates in many pathways regulating metabolic transformations and energy expenditure. PGC1α stimulates the expression of FNDC5 and synthesis of transmembrane protein FNDC5, whose sequence contains hydrophobic transmembrane domain, signal peptide, carboxy-terminal domain located in the cytoplasm, and fibronectin III domain. Irisin is released after proteolytic cleavage and glycosylation of FNDC5 and, probably, there is also dimerization of FNDC5 before irisin release. In humans, FNDC5 expression is observed in skeletal muscles and muscular organs (including heart and tongue) and in adipose tissue [20], and is released into the blood, milk, urine follicular fluid, and saliva [21]. However, it is estimated that FNDC5 expression is 100–200 timeshigher in muscular tissue than in adipose tissue, which may suggest that skeletal muscles are the main source of irisin, and low expression of FNDC5 is observed in the liver and pancreas, among other organs. Irisin, which is secreted from a skeletal muscle, stimulates the expression of UCP1 in adipocytes leading to WAT browning by means of the extracellular signal-regulated kinase (ERK) and mitogen-activated protein kinase p38 (MAPK) [20]. Studies were also performed to establish a correlation between irisin, cachexia, and cancer in mouse models with experimentally triggered stomach cancer. They demonstrated a correlation between the development of cancer and the increase of FNDC5 expression in subcutaneous adipose tissue, while it was not observed in cancer tissue and skeletal muscles. Significantly higher FNDC5 expression and, in consequence, higher concentration of irisin in BAT was observed in pre-cancer and cancer groups compared to controls. During these studies it was also noticed that the concentration of circulating irisin in the cancer group subjects was higher than in the control group [22]. Studies conducted by Petruzzelli et al, who concentrated of the influence of WAT browning on the development of cancer cachexia, demonstrated that this process leads to increased demand for energy expenditure and lipid mobilization, which leads to progression of cancer cachexia. On the basis of studies in rats, it was noticed that the main factors leading to progression of cachexia are Il-6 and UPC1. Data from the performed studies showed that Il-6 induced WAT browning, which leads to increased UPC1 expression and activation of the process of thermogenesis [12]. Irisin and myostatin are recognized markers of muscle strength/mass described in healthy people or in older people [23]. However, its role in neoplastic diseases has also been recently investigated [9–11]. Panagiotou et al showed a higher level of irisin in both patients with benign breast diseases and malignant tumor of the breast compared to controls [10]. However, Provatopoulou et al found a lower concentration of irisin in patients with breast cancer compared to healthy volunteers (2.47 vs 3.24 μg/ml; P<0.001) [9]. Shahidi et al tested irisin levels in gastric cancer patients and healthy controls, finding significantly higher irisin levels in the gastric cancer group compared to healthy controls (0.41 vs 0.35 ng/ml; P=0.032) [11]. Altay et al found higher levels of irisin in renal cancer patients compared to healthy controls (208 vs 110 pg/ml, respectively; P=0.0001) [24].

The development of malnutrition and cachexia is associated with the decrease of fat-free and fat body mass as well as a change from WAT to BAT. An active inflammatory process, which is strongly induced at the time of cachexia progression, contributes to the degradation of muscle fibers [24]. The studies that have been conducted so far prove that irisin has a strong impact on the transformation of WAT into BAT. The present study has some limitations. To assess the nutritional status, we used methods such as the NRS and SGA scales, the results of laboratory tests, and BIA to assess body composition. We did not use any subjective anthropometric methods to evaluate muscle mass. We did not evaluate white and brown fat tissue in the patients. We used BIA method to evaluate the fat mass, percent of fat mass, fat-free mass, and percent of fat-free mass.

Conclusions

Patients with cachexia tend to have higher irisin values compared to normally nourished patients. Low level of irisin is associated with low risk of malnutrition in patients with HNC subjected to RTH. High level of irisin may be a useful marker of malnutrition in patients with HNC subjected to RTH.