Caveolin-1 P132L Mutation in Human Cancers

Abstract

This Commentary discusses the role of caveolin-1 in human cancer.

Caveolae are 50–100 nm flask-shaped invaginations of the plasma membrane that function as membrane organizing centers and play important roles in intracellular trafficking of cellular components (eg, lipid and cholesterol) and signal transduction.¹ The CAV1 gene maps to 7q31.1 and encodes the ubiquitously expressed caveolin-1, which is an essential component of caveolae. Caveolin-1 is believed to regulate the activity of a plethora of signaling molecules highly represented in caveolae, such as H-ras, c-src, and src-like kinases; endothelial nitric oxide synthase; G-proteins and membrane-associated receptors coupled to G-proteins; or tyrosine kinase receptors including epidermal growth factor receptor.¹ In normal breast tissue, caveolin-1 is expressed at high levels in adipocytes, endothelial cells, fibroblasts, and myoepithelial cells of the breast.¹

Initial studies have suggested that caveolin-1 would have properties consistent with those of a tumor suppressor protein.²,³,⁴ A dominant negative mutation of the CAV1 gene was reported to be found in up to 20% of estrogen receptor (ER)-positive breast cancers.²,⁴ Surprisingly, however, CAV1 knock-out mice failed to develop any tumor.¹,³ Furthermore, the presence of CAV1 gene mutations could not be independently validated.⁵,⁶,⁷,⁸,⁹ Concurrent with the controversies about the role of CAV1 as a tumor suppressor gene, evidence of caveolin-1 as a potential oncogene has been mounting.⁷,⁸,⁹,¹⁰,¹¹,¹²,¹³,¹⁴,¹⁵ In fact, multiple independent groups have reported overexpression of caveolin-1 in aggressive subtypes of cancer,¹⁰,¹¹,¹³,¹⁴,¹⁵ and even CAV1 gene amplification has been documented,¹⁴,¹⁶ although this has been shown to be a rare phenomenon. Furthermore, expression of caveolin-1 in stromal cells of breast cancer has also been shown to be associated with the outcome of breast cancer patients.¹⁷,¹⁸

Despite the interest in the tumor suppressive and oncogenic functions of caveolin-1, numerous issues remain unanswered. For instance, caveolin-1 expression in normal luminal epithelial cells of normal breast lobules and ducts remains a matter of controversy, as does the existence and prevalence of the CAV1 P132L dominant-negative mutation. In this issue of The Journal of Molecular Diagnostics, Koike et al¹⁹ have addressed the question of the prevalence of the CAV1 P132L mutation in human cancer. Using three different methods for mutation detection (ie, reverse-transcriptase direct sequencing, DNA-based direct sequencing, and cycleave PCR assay), the authors failed to detect the CAV1 P132L mutation in a large series of 409 human tumors, including 139 breast cancers [114 estrogen receptor (ER)-positive and 25 ER-negative tumors] and 270 cancers from other anatomical sites (ie, gastrointestinal tract, lung, ovary, and pancreas). Taken together, the findings reported by Koike et al¹⁹ and those by others⁵,⁶,⁷,⁸,⁹ call into question the presence of the CAV1 P132L mutation in breast cancers and other cancer types and heat up the debate about the roles of caveolin-1 in human cancer.

CAV1 as a Tumor Suppressor Gene

Since its first description in 1989, CAV1 has been the subject of extensive research; however, its role in cancer still remains a matter of controversy. Several lines of evidence³,⁴,²⁰ suggest that caveolin-1 may have tumor suppressor properties and that its down-regulation plays a role in breast cancer tumorigenesis. This assumption was also based on the high frequency of deletions of 7q31 (a fragile site known as FRA7G next to the CAV1 locus⁸) in human cancer,¹ a high prevalence of an inactivating CAV1 mutation in breast cancer,²,⁴ and down-regulation of caveolin-1 in cancer cell lines, possibly due to CAV1 gene promoter hypermethylation.⁵,⁹

A potential genetic basis for caveolin-1 inactivation in a subgroup of breast cancer stemmed from the study by Hayashi et al,² who described a somatic point mutation of CAV1 gene at codon 132 (P132L). This mutation was reported to cause abrogation of caveolin-1 function due to its dominant negative properties²,³ and to lead to activation of mitogen-activated protein kinase signaling pathway. This mutation was found in 16% of 92 samples of human breast cancers, mainly in ‘scirrhous’ carcinomas (ie, the equivalent of a subtype of invasive ductal carcinoma of no special type in the Japanese classification of breast cancer).² Five years later, the CAV1 P132L mutation was found in one study to be present in 9% (n = 5/55) of breast cancers.⁴ Interestingly, however, six novel CAV1 mutations were described and together with CAV1 P132L mutation were shown to be restricted to ER-positive cancers, suggesting that the CAV1 gene is mutated in up to 35% of ER-positive breast carcinomas.⁴ Further characterization of the functional properties of the P132L mutation was performed using in vitro models.²,³ Surprisingly, however, several groups have now failed to independently confirm the presence of these mutations of the CAV1 gene in human cancer samples and cancer cell lines.⁵,⁶,⁷,⁸,⁹,¹⁹

Supporting their assumption of CAV1 being a tumor suppressor gene, the same group has characterized an in vivo model of CAV1 (−/−) null mice, which developed “wide-spread mammary epithelial cell hyperplasia (a pre-malignant mammary lesion).”³ It should be noted, however, that CAV1 null mice failed to develop mammary tumors, even at up to 9 months of age. Furthermore, the degree of epithelial cell hyperplasia illustrated in that study (Figures 7–9)³ would perhaps be considered ‘mild’ if human breast pathology criteria were adopted. Furthermore, the other alterations observed in those mice,³ such as fibrosis and pronounced lobular development, have not been shown to be associated with cancer development. Importantly, depending on the antibody used,⁸,¹⁰,¹³,¹⁴ caveolin-1 protein is not expressed at detectable levels in luminal epithelial cells of the normal breast; this fact may not be trivially reconciled with the potential tumor suppressor function of caveolin-1 in luminal breast cancer.

CAV1 as an Oncogene

There has been increasingly more coherent data to suggest that caveolin-1 may also have oncogenic properties in human cancer. It has been demonstrated that breast cancer cell lines, in particular those displaying more overt invasive properties as defined by invasion assays (eg, MDA-MB-231, BT549, Hs578T), harbor high levels of caveolin-1 expression at the mRNA and protein levels.¹¹,²¹ Several studies provided mechanistic evidence of the potential oncogenic properties of caveolin-1.¹¹,¹⁵,²¹ For instance, the interaction between caveolin-1 and membrane type 1 matrix metalloproteinase, which colocalize at the subcellular compartment, seems to be an essential regulator of matrix degradation and cancer cell invasion.²¹ Moreover, two independent groups described activation of Rho/ROCK signaling by caveolin-1.¹¹,¹⁵

Accordingly, numerous studies have shown caveolin-1 overexpression in human cancer samples, including breast,⁸,¹⁰,¹¹,¹³,¹⁴,¹⁵ colorectal,¹¹ lung,⁷ and other anatomical sites, as well as have demonstrated that caveolin-1 overexpression is associated with poor clinical outcome and subsets of aggressive tumors, such as inflammatory (41.4%)¹⁵ and basal-like (20 to 70%)¹⁰,¹³,¹⁴ breast carcinomas. In fact, the association of caveolin-1 up-regulation and basal-like cancers should not come as a surprise given that caveolin-1 is highly expressed at the transcriptomic and immunohistochemical levels in basal/myoepithelial cells of the normal breast,⁸,¹⁰,¹³,¹⁴ and basal-like tumors are known to be characterized by the expression of basal/myoepithelial markers. The potential mechanisms underpinning caveolin-1 overexpression include gene promoter hypomethylation¹⁵ and CAV1 gene amplification, which are found in a minority of breast cancers displaying strong caveolin-1 immunohistochemical expression¹⁴ and in gastric cancer.¹⁶ Notably, when amplified, CAV1 appears to be consistently overexpressed, a hallmark feature of amplicon ‘drivers’.

CAV1 Mutations and Expression: Technical Considerations

There exists evidence in support of the role of CAV1 as either a tumor suppressor gene or an oncogene. The question that is germane, however, is how can these contradictory lines of evidence be reconciled? It would be rather trivial to claim that distinct case selection criteria, different in vitro models and different technical approaches would account for the discrepant results.

When it comes to the discrepancies between the results of Koike et al¹⁹ and previous publications²,⁴ on the existence and prevalence of the CAV1 P132L mutation, the source of material and the technologies used for mutation detection are of utmost importance. Sequencing of formalin-fixed paraffin-embedded samples is fraught with difficulties,²² owing to the high prevalence of artifacts (ie, one artifactual mutation per 500 bases²²). Importantly, though, in Hayashi et al,² the source of material is unclear, whereas in Li et al,⁴ formalin-fixed paraffin-embedded samples were used. In fact, the authors themselves stated: “recently, concern has been raised regarding the use of genomic DNA derived from formalin-fixed paraffin-embedded samples for the detection of mutations, due to the putative DNA damage-induced effects of formalin.”⁴ In both studies, tumor DNA was subjected to direct sequencing. In Koike et al¹⁹ cytological preps and paraffin-embedded tissues were used, and both cDNA and tumor DNA were sequenced with different technologies (direct sequencing and Cycleave PCR) and no mutations were found. However, only 40 samples were subjected to the three mutation detection methods. Nevertheless, owing to the putative high prevalence of the CAV1 P132L mutation in ER-positive tumors, a completely negative result in their set of ER-positive cancers¹⁹ would be unlikely. It is noteworthy that none of the breast cancer cell lines subjected to CAV1 gene sequencing to date have been shown to harbor a dominant negative CAV1 mutation.⁸

It should be highlighted that analysis of caveolin-1 expression both by gene expression or immunohistochemistry may be confounded by biases due to either the distribution of caveolin-1 expression in normal tissues or technical issues. At the mRNA level, qRT-PCR and expression array-based methods should be interpreted with caution when samples are not microdissected, as caveolin-1 is highly expressed in adipocytes, fibroblasts, and endothelial cells.¹⁰,¹⁴ This is particularly exemplified by the high level of caveolin-1 expression in the so-called normal breast-like subtype of breast cancer, which is nowadays considered as an artifact of low tumor cell content.²³

At the protein level, the choice of primary antibody, protocols, and scoring systems have a crucial effect on the results of immunohistochemistry analysis. Several anti-caveolin-1 antibody clones are available, and the most widely used are the polyclonal antiserum N-20 (Santa Cruz Biotechnology)³,⁴,²⁰ and the monoclonal 2297 (BD Transduction Laboratories).⁷,¹⁰,¹³,¹⁴ This could explain in part the diametrically opposite results related to the prevalence of caveolin-1 expression in breast cancer, which ranges from as low as 4.2%¹⁰,¹³,¹⁴ in studies performed with the monoclonal antibody to up to 82% when the polyclonal antibody was used.²⁰ Notably, a substantial part of the functional studies focusing on the CAV1 P132L mutation and leading to the conclusion that this mutation has an impact on the subcellular localization of the protein and a dominant-negative effect on cell biology was performed with the polyclonal antibody. Of note, in the same studies, different antibodies were used for Western blot and immunohistochemistry analyses,³,⁴ although the monoclonal antibody has been proven to be adequate for both techniques.¹⁴,¹⁵

CAV1: A Janus of Human Cancer?

There are several lines of evidence to suggest that protein functions may vary according to cell type and network state. This is particularly true in neoplastic cells where numerous genetic and epigenetic alterations co-exist and may modify networks and signaling pathway interactions. In addition, a tumor cell population is constantly under microenvironment pressure and requires different adaptive cellular responses. This ultimately leads to clonal selection during cancer development and progression. Following this line of evidence, it has been hypothesized that during the early phases of tumorigenesis down-regulation of caveolin-1 could promote cancer cell proliferation, whereas in later stages increased caveolin-1 expression could help tumor cells to survive in a more competitive environment.¹ This is supported by the fact that caveolin-1 expression seems to be increased in metastatic variants of tumors.⁷,¹¹ Moreover, tumor cells must also adapt to the exposure to chemotherapeutic agents. In this context, a link between multidrug resistance and caveolin-1 has been described. Multidrug resistance is defined as a cellular resistance to distinct chemotherapeutic agents, which limits their efficacy in treating cancers. Several cell lines displaying a multidrug resistance phenotype, such as the adriamycin-resistant breast cancer cell line MCF-7/AdrR or the colon cancer cell line HT29-MDR, harbor high levels of caveolin-1 and an increased number of caveolae at the membrane surface.²⁴

Another possible explanation for the controversial role of caveolin-1 in human cancer is that the function of caveolin-1 as a tumor suppressor or promoter might be context-dependent. A good example was illustrated by Sunaga et al⁹ in lung cancer. The authors demonstrated that caveolin-1 is down-regulated due to hypermethylation of CAV1 gene promoter in 95% of small cell lung cancer cell lines, while caveolin-1 expression is retained in 76% of non-small cell lung cancer cell lines, which do not display CAV1 gene promoter hypermethylation. In the last decade, it has become clear that breast cancer is heterogeneous and constitutes at least two distinct diseases, which are primarily defined by the expression of ER and ER-responsive genes.²³ Therefore, it is plausible that caveolin-1 may play different roles in the distinct molecular subtypes of breast cancer, mirroring the variety of functions that caveolin-1 is believed to have in breast cancer pathogenesis. In ER-positive tumors, caveolin-1 down-regulation may provide a growth advantage, whereas in ER-negative disease, and in particular in the so-called basal-like cancers, caveolin-1 overexpression might allow tumor progression and/or resistance to treatment.

CAV1: Dead-End Road or a Hit to be Pursued?

The identification of the CAV1 P132L mutation, putatively highly prevalent in ER-positive breast cancers, played a major role for the hypothesis that CAV1 may act as a tumor suppressor gene.²,³,⁴ Subsequent studies, however, failed to confirm the presence of CAV1 P132L mutation in tumor specimens from breast and other anatomical sites.⁵,⁶ Moreover, all functional studies on this mutation had to resort to artificial in vitro models as, to date, no cancer cell line has proven to harbor the CAV1 P132L mutation.⁷,⁸,⁹ In this setting, Koike et al¹⁹ describe the largest published series of tumors (n = 409) subjected to sequencing analysis, demonstrating the absence of CAV1 P132L mutation in human cancers. Taken together, it is reasonable to conclude that CAV1 P132L mutation is either substantially rarer than previously reported,²,⁴ or perhaps, nonexistent in human cancer. Clearly, sequencing of optimally accrued samples (ie, fresh/frozen) using state-of-the-art sequencing methods of a large cohort of breast cancer is required to settle this issue once and for all.

Nevertheless, the study of caveolin-1 remains highly relevant to cancer research. Caveolin-1 has the potential to be a prognostic and/or predictive marker, and in this context it is of paramount importance that studies aiming to address this issue take into consideration the heterogeneity of tumors, as a significant prognostic and/or predictive value may not be detected when distinct subtypes of tumors are analyzed together (eg, small and non-small lung cancers as described above⁹ or breast cancers when considered as a single disease). In the perspective of a possible clinical use of caveolin-1 as a biomarker, the standardization of immunohistochemistry protocols and validation of the sensitivity and specificity of antibodies for immunohistochemistry on formalin-fixed paraffin-embedded samples is crucial. The results of caveolin-1 expression in stromal cells and its impact on the outcome of breast cancer patients and progression from ductal carcinoma in situ to invasive breast cancer are tantalizing and certainly deserve to be thoroughly investigated.¹⁷,¹⁸ In addition, there is evidence to consider caveolin-1 as a putative therapeutic target, given that its overexpression is associated with tumor progression,⁷,¹¹ chemotherapy-resistance,²⁴ and aggressive subtypes of cancer,¹⁰,¹³,¹⁴ and, in a small minority of cancers, the CAV1 gene is overexpressed when amplified.¹⁴,¹⁶ Therefore, despite the controversies and potential technical issues, the role of caveolin-1 as a tumor suppressor, a biomarker, and a potential therapeutic target ought to be further investigated.