Serum irisin levels are lower in patients with breast cancer: association with disease diagnosis and tumor characteristics

Irisin is a muscle-derived factor, discovered and characterized three years ago [1]. Even though it is a relatively new molecule, a significant number of reports has already been published investigating its biology and function in several healthy and disease states. Originally, studies focused on the regulation of irisin by exercise with variable observations [1, 9‐15]. At the same time, extensive research is being carried out on the role of irisin in metabolic diseases, while new findings regarding its implication in other chronic conditions gradually accumulate [32‐42]. As far as the role of irisin in cancer is concerned, current data are scarce and only include in vitro reports [43, 44]. To our knowledge, the present study is the first aiming to investigate the potential role of the newly discovered irisin in human breast cancer. Our results demonstrate that serum irisin levels are significantly lower in breast cancer patients compared to healthy controls. Logistic regression analysis indicated that decreased irisin expression is closely associated with the presence of breast cancer. Most importantly, multivariate analysis showed that serum irisin is an independent predictor of breast cancer. It was estimated that 1 unit increase in irisin levels leads to almost 90 % reduction in the probability of women having breast cancer. According to ROC curve analysis, irisin can effectively discriminate between patients and healthy individuals with 62.7 % sensitivity and 91.1 % specificity at a cut-off point of 3.21 μg/ml, suggesting its potential value in breast cancer early detection.

Our observation of lower expression levels of irisin in breast cancer is supported by the findings of Gannon et al. evaluating the effect of various concentrations of irisin on the behavior of malignant and non-malignant breast epithelial cell lines [44]. The authors reported a significant suppressive effect of irisin on cell number, migration and viability in breast cancer cells, the induction of apoptotic cell death and the suppression of NFκB activity in malignant cell lines. Based on these data, they hypothesized that irisin may alter malignant characteristics similarly to other myokines and affect the development and aggressiveness of breast cancer [44]. It is thus reasonable to assume that reduced levels of irisin would favor breast cancer initiation, in agreement with our observations. The findings of Gannon et al. oppose previous data that reported lack of regulation of cell proliferation and malignant potential of other obesity-related cancer cell lines, including endometrial, colon, thyroid and esophageal, by irisin [43]. Nevertheless, these discrepancies have been attributed to different experimental techniques [44]. Since these are the only currently available data on the function of irisin in cancer, further studies are warranted to elucidate its molecular effects during disease development and to identify the implicated mechanisms.

Interestingly, a decreased expression of irisin has been associated with other chronic conditions, known to be associated with altered energy expenditure and metabolism. Most data arise from studies on patients with diabetes mellitus type 2 (DMT2), indicating a close relationship between the presence of disease and decreased circulating irisin levels [18, 32‐34]. These observations are further supported by reports of down-regulation of the FNDC5 gene in these patients [18, 45]. Similarly, circulating irisin levels have been found significantly decreased in patients with chronic kidney disease (CKD) [35‐37]. Since the development of DMT2 and CKD is often associated with insulin resistance, it has been suggested that irisin may play an important role in the pathology of insulin-resistance related disorders while it may provide a novel therapeutic target for the treatment of metabolic diseases [18, 32, 35]. Two recent reports provide evidence for the implication of irisin in the development of polycystic ovary syndrome (PCOS) and also suggest that it may be a marker of insulin resistance in these individuals [41, 42].

Adiponectin is another adipocyte-derived peptide hormone, able to stimulate the sensitivity of peripheral tissues to insulin, thus acting to enhance insulin sensitivity and protect against conditions of insulin-resistance [46]. Indeed, it has been documented that decreased levels of adiponectin are associated with increased insulin levels that accompany insulin resistance [47]. Low circulating levels of adiponectin have been observed in several diseases characterized by insulin resistance and hyperinsulinemia, including obesity, DMT2 and cancer [48‐56]. Evidence from studies on breast cancer cell lines and animal models indicate that adiponectin can suppress cell proliferation, inhibit tumor growth, increase apoptosis and inhibit angiogenesis through multiple pathways [29‐31]. Moreover, epidemiological data support that it exerts a protective role against breast cancer development, by displaying anti-proliferating, pro-apoptotic, anti-estrogen and anti-inflammatory properties [31]. More specifically, several studies on breast cancer patients have observed reduced adiponectin expression, particularly in postmenopausal women, suggesting a close association with breast cancer risk and providing a potential mechanistic link between obesity and breast cancer [53‐55, 57‐60]. Among the many mechanisms proposed to mediate adiponectin function in breast cancer, the mitogenic effect of hyperinsulinemia appears to be of particular importance [57]. According to our results, serum levels of both irisin and adiponectin were lower in breast cancer patients compared to controls (data not presented) but there was no significant association between them. Current knowledge on the functions of irisin and the mechanisms involved in exerting its actions is exceptionally limited and only allows for hypotheses to be made. Nevertheless, the decreased expression levels of both molecules in our cohort of patients support that irisin may play a role similar to that of adiponectin. Moreover, irisin’s downregulation in conditions characterized by insulin resistance suggest that this mechanism may be a major determinant of its function. Whether the effect of irisin on breast cancer can be attributed to obesity-related hyperinsulinemia and insulin resistance extends beyond the scope of this work, but is a particularly interesting field of research that is worthy of thorough investigation in future studies.

Interestingly, we found that irisin levels were higher in stage III disease compared to stage I, resulting in a positive association with tumor stage. Based on our observation of reduced levels of irisin in breast cancer compared to controls, a further downregulation with the extent or progression of the disease would possibly be expected. This, however, may be a rather simplified hypothesis. Cancer is an exceptionally complex disease, accompanied by major dysregulation of a vast number of interacting molecular mechanisms and signaling pathways. At present, it is still impossible to know how irisin may behave and interact either in physiological conditions or in disease states. One assumption for our observation would be that, after the establishment of breast cancer, the secretion and metabolism of irisin may be affected by other, yet unidentified, cancer-related pathways and factors, resulting in alterations in its expression levels. Otherwise, irisin might exert a different function trying to counteract disease progression that is reflected in its expression levels. Nevertheless, it should be noted that even at stage III disease, irisin levels remained significantly lower than in healthy controls. Since this the first approach, our findings arise from a relatively restricted patient group and their verification in larger cohorts of patients is mandatory.

According to our findings, irisin levels were not associated with serum levels of the established cancer markers CEA, CA15-3 and Her2/neu. Despite their widespread use in routine clinical practice, it is well-recognized that the traditional tumor markers are only suitable for detection of recurrent or metastatic breast cancer as well as for management and surveillance of patients with advanced disease [61‐63]. Their lack of specificity and sensitivity for in situ and primary breast cancer bears serious limitations and renders them rather inappropriate biomarkers for screening and diagnosis of low-volume disease [61, 63]. Thus, our observation of no association between irisin and cancer markers is rather expected and may further support the potential diagnostic value of irisin.

Aiming to explore a potential link between irisin and obesity in human breast malignancy, we investigated potential correlations of irisin levels with body mass index (BMI) and with levels of the adipokines leptin, adiponectin and resistin in breast cancer patients. Our results indicate that irisin was positively associated with leptin, but not with adiponectin, resistin or BMI. Previously, several studies have investigated irisin expression in relation to obesity. BMI is the most traditional and widely-used measure of obesity. It has not yet been clarified if irisin levels are associated with BMI, as some studies have found positive correlations while others negative or no correlation [9, 10, 16‐18]. It was recently proposed that irisin is positively correlated to BMI in healthy individuals but negatively correlated in subjects with metabolic diseases, suggesting a different function according to the metabolic state [64, 65].

Leptin is the most prominent and best studied adipokine, and has been suggested to be implicated in the development of hormone-dependent malignancies including breast cancer [66]. Many studies have investigated leptin as a risk factor for breast cancer mainly in postmenopausal women, with conflicting and opposing findings [67‐73]. At the same time, few reports have explored the role of leptin during disease progression and/or in relation to its clinical behavior. Although data are diverse, studies have provided evidence that increased expression of leptin in patients with breast cancer is associated with large tumor size, advanced tumor stage, high tumor grade and lymph node metastasis [59, 60, 67, 74‐76]. These observations have led to hypotheses that leptin is not only involved in breast cancer initiation but also in disease progression, and that it may be associated with more aggressive phenotypes [59, 60, 67, 74, 75]. In addition, leptin has been associated with distant metastasis and short survival, suggesting that it may have a negative prognostic value [76, 77]. Considering the potential relationship of leptin expression with the characteristics and biological behavior of breast tumors, we consider the association between irisin and leptin levels in our patients is a reflection of the positive correlations of irisin with tumor stage and marginally with tumor size and lymph node involvement. The association of irisin levels with the presence of disease and tumor characteristics but not with BMI suggests that this factor is more likely to be related to systemic disease and tumor extent rather than to adiposity, as previously reported for leptin [76].

Altogether, our findings provide preliminary data for the implication of irisin in breast cancer development and suggest that it may serve as a potential novel biomarker of the presence of disease. Cancer serum markers may have a number of clinical applications including early diagnosis, differential diagnosis, prognosis, prediction of response and resistance to therapy, patient monitoring and surveillance. In breast cancer, established tumor markers are characterized by limited specificity and sensitivity precluding their use for screening and early diagnosis [60, 63]. Mammography, in combination with clinical examination and breast self-examination, remains the primary modality for breast cancer screening and early detection of the disease, and has significant contribution to the decrease in mortality rates [78]. Nevertheless, mammography has important limitations particularly regarding the detection of suspicious lesions in women with very dense breasts and the early diagnosis of interval tumors [78]. As a result, improvements in the detection and early diagnosis of breast cancer are required and are expected to have a major impact on morbidity and mortality from the disease. In that aspect, intensive research is being carried out to exploit the use of novel serum biomarkers in screening, early diagnosis and differential diagnosis of breast cancer [79]. The identification of irisin, as well as other novel proteins, as candidate biomarkers that could assist in the early detection of the disease is an exciting perspective, and remains to be further investigated and verified by future prospective studies.