The first two authors contributed equally to this article.
Expression of caveolin-1 in mucoepidermoid carcinoma of the salivary glands: Correlation with vascular endothelial growth factor, microvessel density, and clinical outcome
Version of Record online: 6 MAR 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 8, pages 1523–1531, 15 April 2007
How to Cite
Shi, L., Chen, X.-M., Wang, L., Zhang, L. and Chen, Z. (2007), Expression of caveolin-1 in mucoepidermoid carcinoma of the salivary glands: Correlation with vascular endothelial growth factor, microvessel density, and clinical outcome. Cancer, 109: 1523–1531. doi: 10.1002/cncr.22573
- Issue online: 4 APR 2007
- Version of Record online: 6 MAR 2007
- Manuscript Accepted: 3 JAN 2007
- Manuscript Revised: 13 DEC 2006
- Manuscript Received: 30 OCT 2006
- vascular endothelial growth factor;
- microvessel density;
- mucoepidermoid carcinoma;
Caveolin-1, which has been proposed as a candidate tumor suppressor, plays a regulatory role in several signaling pathways. The importance of caveolin-1 in endothelial cells in angiogenesis has been confirmed. The clinicopathologic significance of caveolin-1 expression and its correlation with angiogenesis remains unknown in mucoepidermoid carcinoma (MEC) of the salivary glands.
Based on an immunohistochemical study, the expression levels of caveolin-1 and vascular endothelial growth factor (VEGF) and the intratumoral microvessel density (MVD) (labeled by CD34) in 75 patients with MEC were investigated, and correlations with clinicopathologic variables were evaluated statistically.
The expression rates of both caveolin-1 and VEGF were 54.7% (41 of 75 tumors). MVD varied from 9 to 56 (24.45 ± 10.72)/×200. Caveolin-1 expression was correlated inversely with duration of tumor, clinical stage, histologic grade, and MVD (P = .027, P = .011, P = .04, and P = .025; respectively). VEGF expression was associated positively with MVD (P = .000). Advanced clinical stage, higher grade, and tumors that originated from minor salivary glands exhibited higher MVD (P = .029, P = .002, and P = .008, respectively). The presence of clinical symptoms, male gender, advanced clinical stage, higher grade, increased MVD, and down-regulated caveolin-1 were correlated significantly with the development of recurrent disease, as indicated by a shorter disease-free interval (P < .05). Both univariate and multivariate analyses indicated that clinical stage, histologic grade, and MVD were independent prognostic factors (P < .05). The presence of clinical symptoms and the down-regulation of caveolin-1 were identified as negative prognostic predictors in the univariate analysis (P < .05) but did not achieve significance in the multivariate analysis (P > .05).
The current results suggest that caveolin-1 may function as a tumor suppressor in MEC of the salivary glands. Reduced expression of caveolin-1 and increased MVD may indicate a poor prognosis for certain patients. Cancer 2007. © 2007 American Cancer Society.
Mucoepidermoid carcinoma (MEC) is the most common malignant epithelial salivary gland tumor with a widely diverse biologic behavior. The prognosis for patients with MEC usually has been related to clinical stage and histologic grade.1 However, clinical behavior unrelated to histologic morphology also has been reported.2 Therefore, the identification of biologic markers that correlate with clinicopathologic variables or prognosis is important in understanding the characteristics of this neoplasm.
Caveolae are defined morphologically as spherical or flask-shaped invaginations of the plasma membrane of strikingly regular shape and size (≈70 nm average outer diameter).3, 4 Caveolae are found primarily in terminally differentiated mesenchymal cells, including adipocytes, endothelial cells, fibroblasts, type I pneumocytes of the lung, and striated and smooth muscle cells.5 Caveolin proteins (caveolin-1, caveolin-2, and caveolin-3) serve as the structural components of caveolae and also function as scaffolding proteins, which are capable of recruiting numerous signaling molecules to caveolae and regulating their activity. Studies in caveolin-deficient mice have indicated that caveolae and caveolins participate in human disease processes, including diabetes, cancer, cardiovascular disease, atherosclerosis, pulmonary fibrosis, and a variety of degenerative muscular dystrophies.6 Caveolin-1, a 22- to 24-kDa protein of 178 amino acids, initially was identified as a major substrate for tyrosine phosphorylation in Rous sarcoma virus-transformed chicken embryonic fibroblasts, suggesting that it may be a target for inactivation during oncogenesis.7 It has been demonstrated that caveolin-1 is down-regulated in sarcoma, lung carcinoma, and ovarian carcinoma.8–10 However, elevated expression of caveolin-1 has been associated with the metastasis of esophageal squamous cell carcinoma and prostate cancer and negatively correlated with patient survival.11, 12 These findings indicate that the role of caveolin-1 may vary considerably, depending on the tissue involved. Until now, to our knowledge, no literature has been published about the expression of caveolin-1 in MEC of the salivary glands.
Angiogenesis is a crucial process for tumor growth and metastasis regulated by the balance of positive and negative factors. Vascular endothelial growth factor (VEGF) is a specific mitogen for endothelial cells that has demonstrated overexpression in a variety of tumors and other inflammatory diseases.13 Previous studies have revealed that the extent of tumor vascularization correlates with prognosis in a variety of cancers.14, 15 Few reports are available on angiogenesis and its clinical relevance in salivary gland tumors.16, 17 Moreover, caveolin-1, which is highly expressed in endothelial cells, is down-regulated during the proliferative phase and up-regulated during the differentiation phase of angiogenesis.18 However, the correlation of caveolin-1 expressed in tumor cells and angiogenesis remains unknown.
Thus, the objectives of the current study were to investigate the expression levels of caveolin-1 and VEGF and the intratumoral microvessel density (MVD) (labeled by CD34) in MEC, their correlation with clinicopathologic features, and the prognosis for patients with these tumors. The correlations among immunohistochemical markers also were assessed.
MATERIALS AND METHODS
Patients and Specimens
Tumor specimens from 75 patients with previously untreated MEC of the salivary glands (diagnosed between 1970 and 2000) were collected from the files of the Department of Oral Pathology, School and Hospital of Stomatology at Wuhan University. Neither tumors that originated in bone nor tumors from seromucous glands of the sinonasal regions were included.19 It was confirmed that all samples were totally extirpated with sufficient margins, according to the pathologic diagnosis and the operation records of the Department of Oral and Maxillofacial Surgery. Patient recall was available for 53 patients. The follow-up periods were no less than 5 years unless the patient had a recurrence, or metastasis, or died before that time. The day of the primary surgery was considered as the start of follow-up. The time to first recurrence was identified as the endpoint (disease-free survival). Clinical stage was determined according to the criteria used for salivary gland carcinoma as recommended by International Union Against Cancer in 2002.20 Histologic grading was determined according to the 1991 World Health Organization grading scheme.21 All specimens were fixed in 10% formalin and embedded in paraffin wax. Representative blocks were selected, and serial 4-μm-thick sections were examined immunohistochemically.
The following primary antibodies were used: polyclonal rabbit antihuman caveolin-1 (dilution, 1:125; Cell Signaling Technology, Danvers, Mass), monoclonal mouse antihuman VEGF (dilution, 1:200; Santa Cruz Biotechnology, Santa Cruz, Calif), monoclonal mouse antihuman CD34 (dilution, 1:350; Lab Vision and Neomarkers, Fremont, Calif). Sections were dewaxed and rehydrated. Endogenous peroxidase was blocked with 3% H2O2 for 15 minutes, and the sections were subjected to microwave antigen retrieval (750 Watts) in 0.01 mol/L phosphate-buffered saline (PBS), pH 6.0, for 5 minutes. The sections were incubated with normal goat serum for 20 minutes to block nonspecific background staining and subsequently were incubated with the primary antibody at 4°C overnight. After they were washed thoroughly with PBS solution, the sections were incubated with biotinylated goat antimouse/antirabbit immunoglobulin for 30 minutes followed by horseradish peroxidase-labeled avidin-biotin complex for 30 minutes. Diaminobenzidine-hydrogen peroxidase was used as the chromogen. Sections were counterstained lightly in hematoxylin. Positive staining of smooth muscle cells or endothelium, which are abundant in caveolin-1, provided an internal positive control for caveolin-1 immunostaining. Negative controls were obtained by omitting the primary antibodies. Twenty normal salivary gland samples were used as normal controls.
Evaluation of Immunostaining
The scoring and interpretation of immunohistochemical results were performed by 2 independent pathologists without any information about clinical outcomes or other clinical data. Consensus scores were achieved by reexamining the specimens with discrepant scores. For caveolin-1 and VEGF immunostaining, semiquantitative estimates were made by using a composite score that was obtained by multiplying the values of the immunoreaction intensity and relative abundance of positive cells. The intensity was graded as 0 (negative), 1 (weakly positive), 2 (moderately positive), or 3 (strong positive). The abundance of positive cells was graded from 0 to 4 as follows: 0, ≤5% positive cells; 1, 5% to 25% positive cells; 2, 26% to 50% positive cells; 3, 51% to 75% positive cells; and 4, 76% to 100% positive cells.22 For further analysis, a mean score of 6 was used as a cut-off value to divide the samples into a high-expression group (≥6) and a low-expression group (<6). This cut-off value was adopted from a previous report.22 For MVD assessment, microvessels were counted by light microscopy in a ×200 field (Olympus Optical Company, Ltd., Tokyo, Japan) in each of the 5 most vascularized, separately located areas (hot spots). The hot spots were identified during the scanning of the entire section at low magnification (×100) as areas with the highest density of CD34-positive cells. MVD was expressed as the mean count of CD34-immunostained vessels in each sample,23 and the mean value was used as a cut-off point for survival analysis.
The correlation between caveolin-1 expression, VEGF expression, and clinicopathologic data was analyzed by using chi-square tests. Significant differences in MVD between 2 variables were determined by using the Mann-Whitney U test. Spearman correlation analysis was applied for correlations among caveolin-1 expression, VEGF expression, and MVD. Survival rates were obtained by using the Kaplan-Meier method, and the differences were compared by log-rank tests. Univariate analysis with a Cox proportional-hazards model was used to determine each identified prognostic factor, and multivariate analysis with a Cox proportional-hazards model was used to explore combined effects. A difference of P < .05 was considered significant. Statistical analysis was calculated using the SPSS version 11.5 software package (SPSS Inc., Chicago, Ill).
Clinicopathologic Features of the Patients
The study population was comprised of 31 men and 44 women between ages 8 years and 75 years (median age, 38 years). Tumors were graded histologically, and there were 49 low-grade tumors and 26 high-grade tumors. Fifty-one tumors were clinical stage I or II, and 24 tumors were clinical stage III or IV. The primary tumor sites were parotid in 33 patients, submandibular in 7 patients, sublingual in 2 patients, palate in 16 patients, retromolar area in 7 patients, pharynx in 1 patient, and other sites in 3 patients (1 each in the buccal mucosa, floor of the mouth, and tongue). Duration of the tumor varied from 6 months to 20 years (median duration, 11 months). Thirty-eight patients displayed clinical symptoms, such as pain, paresthesia, numbness, and bleeding. Tumor sizes ranged from 0.5 cm to 15 cm (median, 3 cm). Twelve patients had positive lymph nodes at diagnosis, including submandibular, superior deep cervical, jugulodigastric, parotid, and retropharyngeal lymph nodes.
Immunohistochemical Findings and Clinicopathologic Variables
Table 1 summarizes the correlations between immunohistochemical findings and clinicopathologic features. Immunohistochemical analysis revealed a strong, diffuse, cytoplasmic staining pattern in of caveolin-1 expression in the ductal epithelial cells in all 20 normal salivary gland samples (Fig. 1A). Forty-one samples (54.7%) of MEC showed a fine, granular pattern at the surface membrane and the cytoplasm of cancer cells (Fig. 1B-D). Reduced expression of caveolin-1 (mean value, <6) was noted in 34 samples (45.3%). VEGF was expressed in the cytoplasm of both normal ductal epithelial cells and cancer cells, as evidenced by the presence of stained, granular, immunoreaction products (Fig. 2A-C). Of the 75 specimens of MEC, 41 specimens were immunoreactive for VEGF. Only 6 normal samples were positive for VEGF expression. Immunostaining of caveolin-1 and VEGF were obtained in epidermoid and intermediate cells; mucous cells showed no staining. MVD varied from 9 to 56 CD34-immunostained vessels per ×200 field (mean ± standard deviation: 24.45 ± 10.72 CD34-immunostained vessels per ×200 field) (Fig. 3A,B).
|I and II||51||18||33||.011§||27||24||.054||22.18± 8.71||.029‡|
|III and IV||24||16||8||7||17||29.29± 12.99|
Decreased caveolin-1 expression rates were observed in tumors with shorter duration (44%), in stage III and IV tumors (37.5%), and in high-grade tumors (38.5%; P = .027, P = .011, and P = .04, respectively). MVD was 28.48 ± 10.96 in tumors of minor salivary glands and 21.29 ± 9.5 in tumors of major salivary glands, with a significant difference (P = .002). Tumors at stage III and IV exhibited higher MVD than tumors at stage I and II (29.29 ± 12.99 vs 22.18 ± 8.71; P = .029). MVD also increased with histologic grade, with an MVD of 21.98 ± 9.61 in low-grade tumors and 29.12 ± 11.33 in high-grade tumors. This correlation was significant (P = .008). No significant association was observed between other clinicopathologic variables and the immunohistochemical markers that were studied.
Correlation Between Caveolin-1 Expression, VEGF Expression, and MVD
Spearman correlation analysis revealed that MVD was correlated inversely with caveolin-1 expression (coefficient [rs] = −0.259; P = .025) and was associated positively with VEGF expression (rs = .444; P = .000). MVD was significantly higher in the tumors with low caveolin-1 expression and in the tumors with high VEGF expression (P = .027 and P = .000, respectively) (data are shown in Table 2). No significant correlation was identified between caveolin-1 expression and VEGF expression (rs = .111 and P = .344, respectively).
|Variable||No. of patients||MVD, mean± SD||P|
Follow-up ranged from 3 months to 252 months (median, 61 months). None of the 4 patients who received postoperative radiation or chemotherapy developed recurrences. Twenty patient experienced local recurrences, and 1 of those patients had 2 recurrences. Three patients died of their disease. The presence of clinical symptoms (P = .0293), clinical stage (P = .0000), histologic grade (P = .0000), MVD (P = .0008), and down-regulated caveolin-1 (P = .0317) were correlated significantly with the development of recurrent disease, as indicated by a shorter disease-free interval (Fig. 4A-E). Male gender showed a weak correlation with prognosis (P = .0446) (Fig. 4F). Other variables (age, tumor sites, tumor size, duration of tumor, and VEGF expression) did not achieve statistical significance. The results of univariate and multivariate analyses with Cox proportional-hazards models are shown in Table 3. The presence of clinical symptom, clinical stage, histologic grade, MVD, and down-regulated caveolin-1 were correlated with a worse prognosis. The multivariate analysis showed that clinical stage, histologic grade, and MVD had an independent prognostic effect on disease-free survival.
|Variable||Univariate analysis||Multivariate analysis|
|HR (95% CI)||P||HR (95% CI)||P|
|Sex (men/women)||2.421 (0.982–5.968)||.055|
|Age (≤50 y/>50 y)||1.717 (0.673–4.382)||.258|
|Sites (major/minor)||2.078 (0.856–5.046)||.106|
|Duration of tumor (≤11 mo/>11 mo)||0.659 (0.265–1.64)||.37|
|Symptom (negative/positive)||0.373 (0.146–0.951)||.039*||0.409 (0.129–1.296)||.129|
|Stage (I and II/III and IV)||9.293 (3.418–25.264)||.000*||4.821 (1.562–14.878)||.006*|
|Histologic grade (low/high)||16.003 (5.082–50.387)||.000*||6.428 (1.664–24.83)||.007*|
|Caveolin-1 (<6/≥6)†||0.367 (0.14–0.962)||.041*||1.063 (0.342–3.304)||.916|
|VEGF (<6/≥6)†||1.667 (0.632–4.398)||.302|
|MVD (≤24/>24) ‡||4.414 (1.684–11.568)||.003*||3.427 (1.061–11.075)||.04*|
In this study, caveolin-1 expression in samples of normal salivary glands and in samples of MEC of the salivary glands were evaluated immunohistochemically. Immunolocalization revealed strong expression of caveolin-1 in ductal epithelial cells of normal salivary glands. Studies in rat and mice kidneys have implicated the function of caveolin-1 related to water and calcium reabsorption.24–26 Transient receptor potential canonical 1 (TRPC1) is a transmembrane protein that is expressed in salivary gland cells. The wide-ranging biologic roles of TRPC1 include salivary gland secretion27 and the store-operated Ca(2+) influx mechanism in salivary gland cells.28 TRPC1 immunoprecipitates and localizes with caveolin-1, and its function depends on the presence of cholesterol, or at least on the integrity of the cholesterol-rich caveolae.27 Therefore, the physiologic functions of caveolae and caveolin-1 in normal salivary glands deserve further research.
The down-regulated expression of caveolin-1 observed in the current study was similar to that reported in studies of sarcoma, lung carcinoma, and ovarian carcinoma.8–10 It has been suggested that transcriptional silencing through hypermethylation of the caveolin-1 gene promoter may abrogate caveolin-1 expression in human cancers.6 In the current study, down-regulated caveolin-1 expression was associated with advanced clinical stage, high-grade malignancy of MEC, and development of recurrent disease. The univariate analysis with Cox proportional-hazards model also confirmed that lower expression of caveolin-1 was correlated with a worse prognosis, suggesting that the caveolin-1 gene is likely to function as a tumor suppressor gene in human MEC of the salivary glands. However, according to our multivariate analysis, caveolin-1 expression failed to be an independent prognostic factor. This may be explained by the finding that tumor prognosis is affected by many interactive factors. Thus, further research is needed that considers other factors in more samples. A noteworthy finding in the current study was that the frequency of caveolin-1 expression was higher in the tumors that had longer duration (68%) compared with tumors that had shorter duration (44%), indicating that tumor suppressor characteristics of caveolin-1 retarded the progression of tumors and prolonged their duration. Shorter duration was identified as a clinical feature that suggested aggressive behavior in MEC.19 Contrary to the studies described above, elevated expression of caveolin-1 has been associated with the aggressiveness of esophageal squamous cell carcinoma and prostate cancer.11, 12 Lee and colleagues suggested that the diverse effects of caveolin-1 might be simply mediated by the different regions of caveolin-1 molecule and might be dependent on the levels of other molecules that are coexpressed with caveolin-1, such as c-Src and growth factor receptor-binding protein 7 (Grb7). The caveolin-scaffolding domain can function as a negative regulator of a variety of mitogenic signaling molecules, whereas tyrosine phosphorylation of caveolin-1 at residue 14 may confer binding to SH2 domain-containing proteins (such as Grb7) and subsequent growth-stimulatory or oncogenic activity. In this way, caveolin-1 would be able to function as both a negative and positive regulator of signaling and cell transformation.29 Furthermore, it has been demonstrated that the scaffolding domain serves a dual role, acting both as an anchor that holds various proteins within caveolae and as a regulatory element capable of either inhibiting or enhancing a given protein's signaling activity.6
A positive association between the expression of VEGF and the increment of vessel density was identified in the current study, supporting the role of VEGF as a mitogenic growth factor for vascular endothelial cells also in MEC. The microvessel count, which gives a measure of tumor angiogenesis, may provide useful information on the biologic pathway of tumors. Bettencourt and colleagues showed that there was a significant association between the microvessel count and nuclear grade and pathologic stage in prostate cancer.30 In addition, MVD has been correlated significantly with clinical stage, tumor size, vascular invasion, recurrence, and metastasis in patients with adenoid cystic carcinoma.17 In our study, MVD in mucoepidermoid carcinoma was associated with clinical stage, histologic grade, and tumor recurrence. Both univariate and multivariate analyses revealed that MVD had an independent prognostic effect on disease-free survival. MVD in MEC of the minor salivary glands was significantly higher than in tumors of the major salivary glands, and a previous study showed that MEC in the former site occasionally may behave in a highly aggressive manner.19
Numerous studies have been performed to investigate the function of caveolin-1 expressed in endothelial cells in angiogenesis; however, few have studied the relation between caveolin-1 expressed in tumor cells and tumor-associated angiogenesis. Joo et al reported a positive association between caveolin-1 expression and MVD in clear cell renal cell carcinoma (RCC), suggesting that angiogenesis may be affected by caveolin-1 during the progression of RCC.31 Conversely, reduced expression of caveolin-1 (mean value, <6) was correlated with increased MVD in the current study. Based on previous studies, Hoffman has hypothesized that, in the quiescent vasculature, many factors that regulate angiogenesis normally are held together as part of an inactive modular unit (termed as angosome); and, when angiogenesis is stimulated, the angosome dissociates, thus enabling angiogenic regulators to become active. It is proposed that the angosome is present in the caveolae of capillary endothelial cells. The presence of caveolin-1 can inhibit proangiogenic factors: it may act as a master-switch, coordinating events during angiogenesis.32 Liu and colleagues reported that angiogenesis activators, such as VEGF, basic fibroblast growth factor, and hepatocyte growth factor, down-regulate caveolin-1 in human endothelial cells, and the down-regulation of caveolin-1 may be an important step along the pathway toward endothelial cell proliferation.33 Here, we have posed the question whether the correlation between caveolin-1 and the angiogenic regulators found in endothelial cells also exist in tumor cells. Unfortunately, no correlation was observed between caveolin-1 expression and VEGF in the current study. However, because angiogenesis is a process that involves an array of angiogenic regulators, their correlation with caveolin-1 in MEC of the salivary glands remains to be tested. Although the association between caveolin-1 and angiogenesis identified in this study and in RCC31 showed no consistency, these pilot studies provide a basis for further investigation of the role of caveolin-1 in tumor cells and angiogenesis.
In summary, the experimental data presented here showed that caveolin-1 may server as a tumor suppressor in MEC of the salivary glands. Reduction of caveolin-1 expression was correlated inversely with MVD and disease-free survival. Clinical features that suggested aggressive behavior included the presence of clinical symptom, male gender, and advanced clinical stage and histologic grade, in agreement with the features identified in previous studies.18, 34 Increased MVD and advanced clinical stage and histologic grade had independent effects on prognosis.
We thank Mr. Shi-Chun Xiong and Mrs. Yuan Li for their technical assistance.
- 17Expressions of nuclear factor kappaB, inducible nitric oxide synthase, and vascular endothelial growth factor in adenoid cystic carcinoma of salivary glands: correlations with the angiogenesis and clinical outcome. Clin Cancer Res. 2005; 11: 7334–7343., , .
- 20SobinLH, WittekindC, eds. TNM Classification of Malignant Tumors, 6th ed. New York: John Wiley & Sons, 2002.
- 21WHO: Histological Typing of Salivary Gland Tumors. Heidelberg: Springler-Verlag; 1991., , , et al.
- 33Angiogenesis activators and inhibitors differentially regulate caveolin-1 expression and caveolae formation in vascular endothelial cells. Angiogenesis inhibitors block vascular endothelial growth factor-induced down-regulation of caveolin-1. J Biol Chem. 1999; 274: 15781–15785., , , , , .