caveolae are flask-shaped invaginations on cholesterol- and sphingolipid-rich domains of the plasma membrane that have been proposed to play a role in the localization of a wide range of signaling systems and metabolic pathways in a variety of cells. Localization of these processes to caveolae is mediated by protein-protein interactions between pathway components and caveolins, the caveolar coat proteins. Three members of the caveolin family have been described (caveolins-1, -2, and -3). Of these, caveolin-1 (Cav-1; 22 kDa) has been shown to be essential for formation of caveolae in non-muscle cells. Many proteins that interact with Cav-1 contain a caveolin-binding motif enriched in aromatic amino acids (4). Recently, an additional family of caveolae-associated proteins, the cavins (cavins-1, -2, -3, and -4), has also been identified. Although the biology of cavins is just beginning to be explored, it has been suggested that cavins may regulate such diverse functions as sequestration of caveolins within caveolae, caveolar membrane curvature, intracellular movement of caveolae, and coupling of caveolar signaling with nuclear signaling (9).
It has proven very difficult to unravel the role of caveolae in health and disease, because of the complex spatial and temporal associations of macromolecules/cell functions with caveolae (10). A key (and controversial) example is the role of caveolae in cancer cell growth, drug resistance, and metastatic potential. One notable characteristic of cancer cells is their high rate of glycolysis and lactate production (the Warburg effect), and caveolae have been shown to localize the glycolytic pathway to the plasma membrane (14, 24, 25). Hypoxic tumors, which have especially high lactate production rates and are able to survive using glycolysis, are more drug resistant and metastatic (1, 5). Therefore, the role of caveolae in cancer cell drug resistance and metastatic potential is of considerable interest. In this issue, Rungtabnapa et al. (19) begin to dissect out a novel role for caveolae in enabling cancer cells to survive detachment and gain metastatic potential.
Caveolin-1: Tumor Suppressor or Oncogene?
Evidence from in vitro studies investigating the role of Cav-1 in cancer is contradictory, with some studies supporting the conclusion that Cav-1 functions as a tumor suppressor gene, while others indicate that it is an oncogene. Early studies suggested that Cav-1 functions as a tumor suppressor gene. Cav-1 levels were shown to be reduced in oncogenically transformed fibroblasts, and the ability of the transformed cells to form colonies in soft agar was inversely correlated with Cav-1 levels; furthermore, induction of Cav-1 expression inhibited colony growth and/or induced apoptosis in transformed cells and breast cancer cells (6, 7, 12). Further studies showed that Cav-1 levels were inversely correlated with metastatic potential in breast adenocarcinoma cells (20). Simultaneously, however, evidence for an oncogenic role of Cav-1 also began to accumulate. Such evidence included the demonstration that Cav-1 levels are increased in drug-resistant variants of human lung carcinoma, ovarian carcinoma, colon adenocarcinoma, and breast adenocarcinoma cell lines (13, 29). In addition, lung adenocarcinoma cells selected for a highly invasive phenotype had elevated Cav-1 levels (11). Interestingly, prostate cancer cells were found to secrete Cav-1; the secreted Cav-1 can be taken up by tumor cells and endothelial cells and may promote tumor angiogenesis (22, 23). These in vitro studies implicated elevated Cav-1 expression in the development of drug resistance and tumor metastasis.
Contradictory results have also been obtained from in vivo studies in both animal models and human patients. Genetic deletion of Cav-1 in mice with breast cancer induced by oncogenic transformation resulted in increased tumorigenesis and lung metastasis, again supporting a role for Cav-1 as a tumor suppressor (28). However, lung metastases of mice with prostate cancer were found to have increased Cav-1 expression compared with the primary tumor; likewise, Cav-1 levels in lymph node metastases of human prostate and breast cancers were shown to have higher Cav-1 expression than normal epithelia from these tissues (30). Thus, these in vivo studies supported an oncogenic, prometastatic function for Cav-1. Indeed, increased Cav-1 levels are associated with decreased survival in human patients with several types of cancer (for a review, see Ref. 8).
A model that attempts to reconcile at least some of these contradictory findings has emerged. This model proposes that Cav-1 levels vary during the course of tumor progression, with downregulation of Cav-1 in early stages facilitating oncogenic transformation, but with reexpression of Cav-1 at later stages (in some tumors) possibly contributing to the development of drug resistance and metastatic potential (18, 27). Thus, Cav-1 may function as both a tumor suppressor and an oncogene, depending on the stage of oncogenic transformation and extent of tumor progression (Fig. 1). It also seems clear that the role of Cav-1 is highly dependent on tumor type (for a review, see Ref. 27).
Hydrogen Peroxide, Cav-1, and Anoikis
The acquisition of an ability to grow without receiving signals from the extracellular matrix is a key feature of metastatic tumor cells. When deprived of contact with the extracellular matrix, anchorage-dependent cells undergo anoikis, a form of apoptotic cell death. Cell surface integrins have a major role in transmitting survival signals from extracellular matrix to the cell interior and preventing the activation of the anoikis program (reviewed in Ref. 3). Cav-1 has been shown to link integrins to tyrosine kinase signaling and to be necessary for integrin α5β1-mediated cell cycle progression (26).
Although several in vitro and in vivo studies have implicated Cav-1 in anchorage-independent cell growth and tumor metastasis, little is known about the signaling mechanisms that link extracellular matrix-derived signals, such as cell detachment, with Cav-1 and cell survival. In this issue, Rungtabnapa and colleagues (19) build on their recent studies on the role of Cav-1 in anchorage-independent cell survival (2, 15) to identify hydrogen peroxide (H2O2) as an upstream mediator of Cav-1-dependent resistance to anoikis in human lung carcinoma cells. These studies utilized NCI-H460 non-small cell lung cancer cells as a model. In a comparative study of >40 lung cancer cell lines, Sunaga et al. (21) make the striking observation that Cav-1 expression was retained in 76% of non-small cell lung cancer cell lines tested, whereas Cav-1 expression was reduced or absent in 95% of small cell lung cancer cell lines (21). Thus, the NCI-H460 line is a model for tumors that express Cav-1.
Detachment of NCI-H460 cells from the culture surface and growth under conditions in which the cells are unable to reattach causes the majority of the cells to undergo anoikis. Interestingly, the decline in cell viability is paralleled by a decrease in Cav-1 gene and protein expression, and transfection with Cav-1 reduces anoikis (2, 15, 19). Thus, Cav-1 appears to exert an inhibitory effect on anoikis, which would be consistent with a prometastatic action of Cav-1 in vivo.
Reactive oxygen species (ROS) include compounds such as superoxide (O2−), H2O2, and the hydroxyl radical (OH−). ROS have been identified as key players in various aspects of metastasis, including resistance of cells to anoikis (16). Rungtabnapa et al. (19) thus examined the role of ROS in Cav-1 expression and anoikis following cell detachment. They report that detachment of NCI-H460 cells induces a time-dependent increase in ROS. Intracellular ROS levels were examined with fluorescence staining and flow cytometry in cells treated with scavengers of the three major ROS, allowing H2O2 and OH− to be identified as the ROS that are primarily increased following cell detachment. In further experiments, the authors show that treatment of detached cells with the peroxide scavenger catalase reduces cell viability, while scavengers of O2− and OH− have no effect. Thus, peroxide is identified as a factor that limits anoikis in detached cells. To determine whether the effects of peroxide are mediated by a Cav-1-dependent mechanism, Cav-1 protein levels were measured in detached cells treated with and without a ROS scavenger. Again, peroxide was found to be the key ROS involved in this signaling pathway; scavenging H2O2 decreased Cav-1 protein levels via increased ubiquitination and proteasomal degradation of Cav-1, whereas exogenous H2O2 had the opposite effect. Finally, the ability of NCI-H460 cells to form colonies in soft agar was increased by exogenous H2O2 and blocked by catalase. Cav-1 overexpression similarly enhanced colony formation, whereas knockdown of Cav-1 with a small hairpin RNA resulted in increased cell death.
The results of Rungtabnapa and coauthors mesh nicely with findings from other recent studies of mechanisms mediating anoikis. One mechanism by which cell detachment is thought to induce anoikis is via inhibition of phosphatidylinositol 3-kinase (PI3K)/Akt signaling; in contrast, integrin-linked H2O2 production activates PI3K/Akt in adherent cells to prevent anoikis (16). The authors of the present paper and others have shown that increased Cav-1 levels cause constitutive activation of PI3K/Akt signaling (2, 17). Thus, the work presented in this issue provides an important link between these previous studies. Another recent report from this group describes a role for ROS in Cav-1-mediated migration and invasion of NCI-H460 cells (15). Along with that work, the present article provides new insights into how ROS may contribute to metastatic potential of tumor cells via altered regulation of Cav-1. Further characterization of this signaling pathway could eventually lead to new mechanism-based antimetastasis therapies for patients with Cav-1-expressing tumors.
No conflicts of interest, financial or otherwise, are declared by the author(s).
- Copyright © 2011 the American Physiological Society