The present study shows that: (1) the level of caveolin-1 expression is elevated in pancreatic adenocarcinoma relative to non-neoplastic ductal epithelium, (2) caveolin-1 expression is related to tumour diameter and histopathologic grade, and (3) positive caveolin-1 expression is an independent unfavourable prognostic factor following surgical resection.
The expression of caveolin-1 is elevated in various cancer tissues, including prostate cancer, oesophageal squamous cell carcinoma, colon cancer and breast cancer (Yang et al, 1998
; Fine et al, 2001
; Kato et al, 2002
). In addition, a correlation between caveolin-1 expression and clinicopathological variables has been shown in these cases. In prostate cancer, caveolin-1 expression positively correlated with Gleason score, positive surgical margin and lymph node metastasis (Yang et al, 1999
). In oesophageal squamous cell carcinoma, expression of caveolin-1 was positively correlated with histopathologic stage, lymph node metastasis and distant metastasis (Kato et al, 2002
). In the present study, positive caveolin-1 expression was detected in 32 out of 79 tumours (40.5%) in pancreatic carcinoma, while non-neoplastic ductal epithelium showed little or no immunoreactivity. Moreover, caveolin-1 immunopositivity is positively correlated with tumour diameter and histopathologic grade in this cancer as well. These results suggest that caveolin-1 overexpression may contribute to the aggressiveness of pancreatic carcinoma.
Except for the case of prostate cancer, however, the molecular and cellular underpinnings of the relationship between caveolin-1 expression and cancer progression remain unclear. In prostate cancer, caveolin-1 both protects against androgen withdrawal-induced apoptosis in vitro
and in vivo
(Nasu et al, 1998
) and blocks c-myc
-induced apoptosis in cancer cells (Timme et al, 2000
). Moreover, caveolin-1 mediates testosterone-stimulated survival/clonal growth and promotes metastatic activity (Li et al, 2001
), while overexpression of caveolin-1 potentiates ligand-dependent androgen receptor activation (Lu et al, 2001
). It is well known that prostate cancer is androgen-dependent. Androgen receptor has also been detected in cancerous tissues of pancreatic carcinoma (Corbishley et al, 1986
), and it is thought that testosterone increases growth of this tumour (Greenway, 1998
). Thus, a similar mechanism may be at work in prostate cancer and pancreatic carcinoma.
Interestingly, it has recently been reported that a caveolin-1 mutation at codon 132 was found in human breast cancer specimens and that the mutated caveolin-1 induced cellular transformation, activated the mitogen-activated protein kinase (MAPK)-signalling pathway, and promoted invasion ability in NIH3T3 cells (Hayashi et al, 2001
). Following up this result, we searched for this caveolin-1 mutation in 11 pancreatic cancer cell lines, but found no mutation in the predicted functional domains (including the scaffolding and membrane-spanning domain) (data not shown). In light of this finding, we suggest that overexpression of wild-type caveolin-1 may be associated with cancer progression in pancreatic carcinoma.
Contrary to the above-mentioned studies, it has been reported that caveolin-1 levels were reduced in a variety of cancer cell lines and cancer specimens (including human breast cancer, lung cancer, colon cancer, ovarian cancer and sarcoma) (Lee et al, 1998
; Racine et al, 1999
; Bender et al, 2000
; Wiechen et al, 2001a
). Under some conditions, caveolin-1 has been shown to suppress growth of specific cell lines in vitro
and in vivo
(Koleske et al, 1995
; Engelman et al, 1997
; Suzuki et al, 1998
), and some have suggested that caveolin-1 functions as a tumour suppressor gene (Engelman et al, 1998b
). The reasons behind this seemingly contradictory evidence remain unclear. Lee et al (2000)
suggest that the diverse effects of caveolin-1 may be mediated by different regions of the caveolin-1 molecule, and may depend on the expression levels of other coexpressed molecules. It has been reported that the oncosuppressive effect of caveolin-1 is mediated through the caveolin-1 scaffolding domain (residues 82–101) (Okamoto et al, 1998
). c-Src, however, induces phosphorylation of caveolin-1 at residue tyrosine 14. Tyrosine 14-phosphorylated caveolin-1 confers binding to growth factor receptor-binding protein 7 (Grb7) and augments both anchorage-independent growth and epidermal growth factor (EGF)-stimulated cell migration (Lee et al, 2000
). In pancreatic carcinoma, Src kinase overexpression and activation has been reported (Lutz et al, 1998
). Thus, caveolin-1 might cooperate with other molecules, such as c-Src and Grb7, to stimulate tumour growth in pancreatic carcinoma.
This is the first study demonstrating the prognostic significance of caveolin-1 expression in pancreatic carcinoma. The 3-year survival rate following surgical resection of the caveolin-1 negative group was 33.8%, while that in the caveolin-1 positive group was only 4.8%. Furthermore, multivariate analysis demonstrated that positive caveolin-1 expression is an independent negative prognostic factor. These results can increase the accuracy of prognosis for patients with pancreatic carcinoma, following surgical resection. Caveolin-1 overexpression in resected specimens may be a useful index of adjuvant therapy for the patients with a high risk of poor prognosis.
In summary, overexpression of caveolin-1 in pancreatic carcinoma may contribute to tumour progression and be a negative prognostic predictor following surgery. For patients with a tumour overexpressing caveolin-1, closer follow-up should be performed to find recurrence, and adjuvant therapy may be beneficial. However, at present, the role of caveolin-1 in pancreatic carcinoma remains unclear, and elucidation awaits further investigation.