Abstract
Background/aim: Obesity increases the risk of breast cancer. It is established that adipocyte secretions, i.e. adipokines, may play a role in mammary carcinogenesis. We have shown that two major adipokines, leptin and adiponectin, were expressed in mammary adenocarcinoma. Patients and Methods: Here, we evaluated zinc-α2-glycoprotein (ZAG) expression in tumor (n=55) and healthy (n=6) breast tissue by immunohistochemistry and examined whether it was correlated with that of major adipokines, usual tumor biomarkers (sex steroids receptors, i.e. estrogen (ER) and progesterone; Ki-67; cErb2), or apoptosis markers (Bcl2 and Bax). Results: ZAG expression was detected in ductal carcinoma and normal epithelial adjacent tissue but not in normal tissue of healthy women. In cancer tissue, its expression was correlated positively to leptin receptor and negatively to adiponectin receptor and ER. Conclusion: These preliminary results suggest both a relationship between ZAG expression and pathways involving adipokines or estrogen and that ZAG may be a potential breast cancer biomarker.
By the year 2000, 300 million adults were obese worldwide (1). Obesity is now recognized as a real risk factor for postmenopausal breast cancer development. Nevertheless, the specific mechanisms implicating obesity in the breast cancer process are not well understood (2, 3). After menopause, aromatization processes that convert androstenedione into estrogen, take place primarily in the adipose tissue whose mass is increased in the case of obesity (4). The mammary epithelial microenvironment (adipose tissue, fibroblasts, myoepithelial cells) could also be implicated in the development of this tumoral pathology and especially adipose tissue. Indeed, in healthy or tumoral breast, epithelial cells are directly in contact with adipocytes which could also influence them.
Adipocytes secrete a large number of molecules named adipokines (5). Approximately 250 adipokines have been described to date including new adipokines such as zinc-α2-glycoprotein, apelin, visfatin, adiponutrin, resistin and hepatocyte growth factor. Some of them, such as adiponectin and leptin, are known to play a role in breast cancer development (6). We have previously established that leptin had pro-carcinogenic effects and adiponectin anticarcinogenic effects (7, 8). Indeed, in ductal tumors, we showed a strong expression of leptin and conversely a low expression of adiponectin. Moreover, normal tissue within normal breast, e.g. non-cancerous lesion holders, does not express leptin, contrary to the normal tissue surrounding a breast ductal carcinoma. Additionaly, we have established a strong expression of adiponectin in myoepithelial cells known for their tumor suppressor properties (9). In vitro, we have reported on the expression of adiponectin and leptin receptors on the breast cancer cell line MCF7 by qRT-PCR, and that adiponectin induces an anti-proliferative response to these cells in contrast to leptin. In addition, we have shown that adiponectin decreases aromatase activity and expression of the estrogen receptor in MCF7 tumor cells (8, 9). These results suggest an important local role of the adipocyte microenvironment in the mammary gland, which could be of importance in the case of obesity.
Other adipokines may also play an important role in breast cancer progression. For example, hepatocyte growth factor (HGF) has been associated with proliferation, migration and invasion of tumor cells, and several tumors, including breast cancer, produce HGF and overexpress its receptor (10). In a Korean study, the frequency of tumors with the highest histological grade was significantly increased in patients presenting a higher resistin level than the median, suggesting that high resistin levels were likely to be associated with increased breast cancer risk in Korean women (11). Another study suggested that apelin behaved as a potent activator of tumor neoangiogenesis by a paracrine effect (12). Overall, several of the adipokines mentioned seem to have pro-carcinogenic effects, and it would be wise to identify a molecule with anti-carcinogenic properties such as adiponectin. Among these adipokines, zinc-α2-glycoprotein (ZAG) seems to be of potential interest for breast cancer cases.
ZAG, also named GCDFP-44 (gross cystic disease fluid protein) (13), is a 44 kDa single chain polypeptide (14-16), identified in 1961 (17, 18). A high degree of similarity between ZAG and class I MHC molecules has been shown both at sequence and structural levels (14). ZAG is secreted in various body fluids and is present at high concentrations in human seminal plasma (14) and breast cyst fluid (19). It has been shown that ZAG can bind to the adrenoreceptor (14) and its expression can be regulated through TNF-α and the PPARγ nuclear receptor (20). Interestingly, the ZAG gene and protein has been shown to be down-regulated (−70%, p<0.05) in adipose tissue of obese compared to lean subjects (21-23). ZAG, also called lipid-mobilizing factor (14, 15, 24), stimulates lipolysis (25) and is associated with an extensive reduction of fat in the human body (14).
To the best of our knowledge, only few studies reported a link between ZAG and cancer process. In breast fluids, the presence of ZAG was demonstrated in about 40% of breast carcinomas (26). Moreover, in breast tumor cytosols, the concentration of ZAG ranged from 0 to 23.5 μg/mg of total soluble protein, with an average value of 2.4 μg/mg (19). In in situ ductal carcinoma, ZAG was expressed in 40% of the studied cases (13). Moreover, ZAG was also described as a potential serum marker of prostate cancer that may be early elevated in tumor growth (14) and seem to be expressed in cancer cachexia (14, 15, 20, 24, 27, 28).
Cancer is characterized by a dysfunction of different cell mechanisms including apoptosis, cell cycle, angiogenesis, proliferation and migration. Dysregulation of normal programmed cell death pathways plays an important role in the pathogenesis and progression of breast cancer but no relationship between ZAG and apoptosis has been described to date. Among many apoptosis-inducing factors, Bax and Bcl2 have been identified functionally antagonistic proteins which control this mechanism. Bax is a 21 kDa proapoptotic protein with extensive amino acid homology with Bcl2 and is suggested as a good prognostic marker (29). Bax has been shown to form heterodimers with Bcl2 and the ratio of Bcl2 to Bax determines the survival or cell death following an apoptotic stimulus such as removal of a growth factor (29). Bcl2 gene encodes a 26 kDa protein, which appears to play a key role in cell regulation by inhibiting apoptosis (29). These proteins are expressed differentially depending on cell types and stage of differentiation; and their biological effects depend on their level, selective expression and dimerisation status (30).
In this study, in an attempt to elucidate the relationship between ZAG and breast carcinogenesis, we investigated ZAG expression in different types of breast tumor, in normal tissue adjacent to breast tumor and in tissue of healthy women by immunohistochemistry. We also evaluated the relationship between ZAG expression and the expression of two major adipokines, leptin and adiponectin, with apoptosis and clinical classical markers.
Patients and Methods
All chemicals were purchased from Sigma (Saint-Quentin-Fallavier, France) except for the anti-ZAG primary antibody (Santa Cruz Biotechnology, Santa Cruz, United States), the biotinylated anti-goat IgG (Vector Laboratories, Burlingame, United States), the avidin/biotin blocking kit, the Vectastain ABC kit, the diaminobenzidine (DAB) substrate and the Vectamount mounting medium (Vector Laboratories, Abcys, Paris, France).
Patients. Primary tumors from 39 women, aged 30-80 years, not treated previously by radiotherapy or chemotherapy, were surgically resected in the Department of Surgery, Centre Jean-Perrin, France. Control tissue samples were collected from 6 healthy women aged 33-63 years with mammary hypertrophy by the Tissue and Cell bank, Hôpital Edouard-Herriot, France. Tissue samples were immediately paraffin-embedded and cut into 3 μm wide sections.
Tissue classification. The diagnosis was made on formalin-fixed paraffin-embedded tissue sections after haematoxylin-eosin-saffron staining. Routinely, the expression of estrogen receptors (ER), progesterone receptors (PR) and Ki-67 were evaluated and scored as previously described (31). Tissues were classified according to histological subtypes as malignant lesions corresponding to in situ ductal carcinoma (n=16) and invasive ductal carcinoma of different grades [grade 1 (n=11), 2 (n=14) and 3 (n=14)]. Invasive tumors were evaluated according to the SBR grade classification modified by Elston and Ellis (32).
Immunohistochemistry. The expression of ZAG was investigated by immunohistochemical staining using affinity-purified goat polyclonal biotinylated antibodies against ZAG. Sections were deparaffined in toluene through graded concentrations of ethanol to distilled water. Sections to be stained with antibodies were pretreated by boiling them for 40 minutes in citrate buffer. Nonspecific binding sites were blocked using the avidin/biotin kit for 30 min. Slides were then incubated overnight at 4°C in a humid chamber with the primary anti-ZAG antibody (0.5 μg/ml) and 4 hours with the secondary anti-ZAG biotinylated antibody. Endogenous peroxidase activity was inhibited with 0.3% hydrogen peroxide for 5 min. Visualisation was carried out using a Vectastain ABC peroxidase-conjugated streptavidine kit for 30 min. The sections were then treated with DAB substrate for 10 min. Finally, slides were contrasted using haematoxylin, dehydrated and mounted using the Vectastain mounting medium. For each assay, control samples without the anti-ZAG antibody were used to establish the specificity of the immunohistochemical analysis.
The expressions of leptin, leptin receptor (Ob-R), adiponectin and adiponectin receptors (adipoR1 and R2) were determined by immunohistochemistry as previously reported (8). By addition, the expressions of Bax and Bcl2 were evaluated with the same technique.
Microscopic examination. Assessment of immunostaining was performed by a anatomopathologist blinded to the clinical data. The expression of ZAG in cancerous and normal adjacent tissues was classified as negative (<5% labeled cells) or positive (≥5% labeled cells). The total percentage of cases with ZAG expression was determined.
Statistical analysis. The statistical analysis was performed using StatView statistical software (SAS Institute, Carry, NC, USA). Relation-ships between ZAG and others adipokines, histological features and usual biomarkers were analysed using the Spearman rank correlation. In this study, we combined ductal, in situ and invasive, carcinomas for Spearman rank correlation. Differences with p<0.05 were considered to be statistically significant.
Results
ZAG expression in breast cancer tissue, in normal tissue adjacent to tumor and in normal tissue from healthy women. In ductal carcinoma, positive ZAG expression was noted in 81% of in situ (DCIS) cases studied (Figure 1A) and 79% of invasive (IDC) cases (Table I, Figure 1B). Concerning IDC cases, ZAG expression was detected in 82%, 71% and 86% of grade I, II and III, respectively. In normal tissue adjacent to breast cancer, positive ZAG immunostaining was observed in 94% of studied cases (Figure 1C). Nevertheless, ZAG expression was undetectable in normal cells from healthy breast (Figure 1D).
Correlations between expression of ZAG and other adipokines. As previously reported, in breast cancer, leptin (79%) and Ob-R (81%) were strongly expressed whereas adiponectin (7%) and adipoR1 (15%) were weakly expressed (9). In ductal carcinoma (IDC and DCIS cases), ZAG expression was correlated positively with the expression of leptin receptor (p=0.015, r=0.341) and negatively with the expression of adiponectin receptor R2 (p=0.024, r=-0.315). The expression of ZAG showed no statistically significant correlation with expression of leptin, adiponectin and adiponectin receptor R1 (Table II).
Correlations between expression of ZAG and apoptosis biomarkers, Bax and Bcl2. All of the DCIS cases were Bax negative while 87.5% of cases were Bcl2 positive. Concerning IDC, 25% of cases were Bax positive and 60% of cases expressed Bcl2. Normal tissue adjacent to breast cancer were Bax negative whereas 94% of cases were Bcl2 positive. In ductal carcinoma, Bax and Bcl2 were correlated negatively (p=0.025, r=−0.430). Bax and Bcl2 were not significantly correlated with ZAG expression (Table III).
Relationships between expression of ZAG and usual breast tumor biomarkers. In ductal carcinoma, ZAG expression was correlated negatively with ER (p=0.046, r=−0.279). The expression of ZAG was not significantly correlated with the expression of PR, c-Erb2 or Ki-67 (Table IV).
Relationships between expression of ZAG and tumor size. ZAG expression was correlated negatively with tumor diameter (p=0.008, r=−0.377) in ductal carcinoma.
Discussion
Obesity is now recognized as a risk factor for breast cancer development. We previously reported that adipokines, such as leptin and adiponectin, were expressed in breast cancer tissue. ZAG, a well known lipid mobilizing factor, which is down-regulated in obesity, is actually considered as an adipokine. In human breast cancer tissue, the implication of ZAG in mammary carcinogenesis has been explored in only a few studies. The aim of the present study was therefore to research the expression of this adipokine in breast cancer tissue and to determine its relationship with other adipokines and common biomarkers, in an attempt to better understand the impact of the tumor adiposity microenvironment on breast cancer development.
We showed that ZAG was expressed both in primary ductal breast cancer and in epithelial cells of normal tissue in the vinicity of a ductal breast lesion. The results presented here are consistent with the data of Selim et al. (13) who reported ZAG expression in DCIS. However, a significant association between the expression of ZAG and the degree of differentiation of CCIS could not be established, contrary to this author who demonstrated that in the cytosol of breast cancer cells, there was a positive significant association between ZAG concentration and the histological grade of the tumors. Indeed, it has been reported that ZAG levels are higher in well-differentiated tumors (mean 4.6 μg/mg) than in moderately (1.8 μg/mg) or poorly (0.9 μg/mg) differentiated tumors (19). Interestingly, in the present study, ZAG was undetectable in normal cells from healthy breast in all the cases studied. We report here for the first time that ZAG was expressed by all malignant cells, as well as in normal tissue in the vicinity of a ductal breast lesion but not in normal tissue of healthy breast. Given this observation, we may postulate that ZAG uptake may be increased by tumor cells and/or that ZAG may be directly synthesized by these malignant cells. Moreover, Sanchez et al. showed that ZAG was found in all the breast fluids obtained from most healthy women (93%) as well as in most patients diagnosed with duct ectasia, fibroadenoma, and gross cystic breast disease (88%) than in fluids from women diagnosed with breast carcinoma (43%) (26). Consequently, these data suggest that ZAG expression was enhanced in breast cancer compared to normal tissue in healthy women and that ZAG expression was induced in normal tissue adjacent to breast cancer during carcinogenesis.
Marrades et al. showed statistically significant positive correlations between ZAG gene expression and serum adiponectin (r=0.89, p<0.01) and a negative correlation with the plasma level of leptin (r=−0.82, p<0.05) in obese subjects (21). Previously, we evaluated the protein expression of these two major adipokines, leptin and adiponectin, and we reported that leptin and its receptor were strongly expressed whereas adiponectin and its receptor adipoR1 were weakly expressed. In our study, we showed that ZAG was positively correlated with leptin receptor (Ob-R) and negatively with adipoR2. These data suggest that there may be a link in the regulation of various adipokines as previously reported by Jarde et al. (8).
In the development of cancer, it is well known that apoptosis plays a key role, particularly through secretion of factors such as Bax and Bcl2. We reported that there was a low expression of Bax and a high expression of Bcl2 in ductal breast cancer and in normal tissue adjacent to lesions. These data are consistent with the results of Veronese et al. showing positive Bax immunostaining in 45% (grade 1 and 2) and 55% (grade 3) and Bcl2 in 67% of cases (30), and with those of Baccouche et al. demonstrating an overexpression of Bax (12%) and Bcl2 (41%) in tumors (33). Rehman et al. showed Bax expression in 66% of DCIS and 56% of IDC and Bcl2 expression in 40% of cases (29). In the present study, Bax was signi-ficantly negatively correlated with Bcl2, which is in contrast with the results of Rehman et al. showing a positive correlation between Bax and Bcl2 (29). Moreover, ZAG was not associated with Bax and Bcl2 apoptosis biomarkers, but was negatively correlated with estrogen receptor (ER). ZAG and estrogen seemed not to be associated even though a link between leptin and estrogen receptor has previously been reported (8). The data presented here suggest that the adipokines could interact with the estrogen pathway. ZAG was also negatively correlated with tumor size. Consequently, it could be suggested that ZAG may have anticarcinogenic effects. Nevertheless, these data are not consistent with the results of Diez-Itza et al. which demonstrated that there was no significant correlation between ZAG and tumor size or ER (19).
In conclusion, our study demonstrated that ZAG was expressed in breast cancer cases and in normal tissue adjacent to lesions but not in normal tissue of healthy women. This data supports a role for ZAG as a potential biomarker in breast cancer. ZAG correlated positively with Ob-R and negatively with adipoR2, ER and tumor size. Consequently, we hypothesize that ZAG may be linked with other adipokines and the estrogen pathway and that it could have anticarcinogenic effects. The presented data confirm our interest to study the tumoral microenvironment, particularly adipose tissue, in relation to breast cancer development. Nevertheless, further studies are necessary in order to better understand the mechanisms involved in carcinogenesis.
Acknowledgements
This study received financial support from the ‘Ligue Contre le Cancer’, France. Virginie Dubois is supported by fellowship from the Ministry of Research and Technology.
- Received March 16, 2010.
- Revision received June 8, 2010.
- Accepted June 14, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved