- Top of page
- Materials and Methods
- Disclosure Statement
- Supporting Information
CUB (C1r/C1s, urchin embryonic growth factor, BMP1) domain-containing protein 1 (CDCP1) has been implicated in promoting metastasis of cancer cells through several mechanisms, including the inhibition of anoikis, which is cell death triggered by the loss of extracellular matrix interactions. However, the mechanism inhibiting cell death regulated by CDCP1 remains elusive. Inhibition of CDCP1 expression using small interfering RNA (siRNA) induced the cell death of suspended cancer cells without cleaving caspase-3, a marker of apoptosis; cell death was not inhibited by a general caspase inhibitor, suggesting that the loss of CDCP1 induces caspase-independent cell death. In contrast, knockdown of CDCP1 as well as protein kinase Cδ (PKCδ), a downstream effector of CDCP1, in a suspension culture of lung cancer cells resulted in marked induction of membranous microtubule-associated protein 1 light chain 3 (LC3)-II protein, a hallmark of autophagy, and caused the formation of an autophagosome structure visualized using green fluorescent protein-tagged LC3-II. Expression and phosphorylation of exogenous CDCP1 by Fyn kinase reduced the formation of autophagosomes and inhibited phosphorylation of CDCP1 by PP2, a Src kinase inhibitor or inhibited PKCδ by rottlerin, stimulating autophagosome formation. Moreover, death of suspended lung cancer cells induced by CDCP1 siRNA or by PKCδ siRNA was reduced by the autophagy inhibitor 3-methyladenine. These results indicate that CDCP1-PKCδ signaling plays a critical role in inhibiting autophagy, which is responsible for anoikis resistance of lung cancer cells.
CUB (C1r/C1s, urchin embryonic growth factor, BMP1) domain-containing protein 1 (CDCP1), also known as SIMA135 and TRASK,[1, 2] is a type I transmembrane protein with three putative CUB domains as extracellular domains and several tyrosine residues that are phosphorylated by Src family kinases (SFK) in the cytoplasmic domain.[1-6] Expression of CDCP1 has been reported in several human malignancies, such as colon and breast cancers.[3, 7] CDCP1 expression is strongly associated with progression and poor prognosis of various cancers, including renal cell carcinoma, lung adenocarcinoma and pancreatic cancer.[8-10] CDCP1 was reported to bind directly to protein kinase Cδ (PKCδ) at a unique C2 domain in a phosphorylation-dependent manner. We recently determined the biological significance of this interaction with PKCδ. We found that tyrosine-phosphorylated CDCP1 regulates cell migration, invasion and anoikis resistance of cancer cells by physically linking between SFK and PKCδ.[8, 11]
Anoikis is a form of cell death triggered by the loss of cell survival signals based on interactions with the extracellular matrix (ECM). Anoikis is considered to be physiologically important for maintaining homeostasis and tissue architecture. However, anoikis resistance acquired during carcinogenesis has been described as a core aspect for tumor progression and metastasis of cancer cells. Because anoikis resistance is unique to metastatic cancer cells, it might be a good target for developing antimetastatic therapy, which has minimal effects on normal tissue cells. However, it is unknown how CDCP1 signaling confers resistance to anoikis in cancer cells.
In the present study, we found that the cellular response compatible with autophagy is involved in anoikis of lung cancer cells caused by the suppression of CDCP1 signaling. Autophagy is an evolutionally conserved process that is characterized by an increase in the number of autophagosomes surrounding cellular organelles such as the Golgi complex, mitochondria and endoplasmic reticulum. High levels of autophagy can lead to cell death. In contrast to apoptosis, cell death mediated by autophagy is caspase independent and does not involve classic DNA laddering. Accumulating evidence suggests that cancer cells generally show reduced autophagy during cancer progression, supporting the hypothesis that defective autophagy provides resistance to metabolic stress such as hypoxia, acidity and chemotherapeutics, promotes tumor cell survival and plays a role in tumorigenesis.
Here, we present in vitro evidence that CDCP1-PKCδ signaling plays an essential role in suppressing autophagy in anchorage-free lung cancer cells, thus protecting the cells from anoikis during the development of cancer metastasis.
- Top of page
- Materials and Methods
- Disclosure Statement
- Supporting Information
Both apoptosis and autophagy are highly regulated forms of programmed cell death and play crucial roles in physiological processes such as the development, homeostasis and elimination of abnormal cells. Our previous study showed that CDCP1 inhibits anoikis in cancer cells. Anoikis is an important mechanism for maintaining tissue homeostasis by killing cells that have lost contact with the ECM. The ability to survive in the absence of ECM interactions is considered to be a critical feature of metastasis because cancer cells moving through the vasculature or growing to secondary sites are either deprived of matrix or exposed foreign matrix components. Indeed, CDCP1 promotes distant metastasis and peritoneal dissemination of cancer cells in mouse models.[11, 24] High CDCP1 expression has been observed in various types of human cancers[1, 8, 24, 25] and has been associated with poor prognosis of patients in some cases.[8, 10] In the present study, we are the first to provide evidence that CDCP1 regulates autophagy in cancer cells and further demonstrate that suppression of autophagy by CDCP1-PKCδ signaling plays an important role in preventing anoikis of lung cancer cells.
First, using two different approaches, we showed that anoikis induced by CDCP1 siRNA in A549 lung cancer cells is caspase independent. Cell death caused by CDCP1 inhibition did not produce cleaved caspase-3, one of the key molecules involved in apoptosis (Fig. 1a). Activation of caspase-3 requires proteolytic processing of its inactive zymogen into activated fragments such as cleaved caspase-3. Furthermore, a general caspase inhibitor did not affect anoikis following treatment with CDCP1 siRNA (Fig. 1b). Thus, our data demonstrate that cell death induced by CDCP1 siRNA in lung cancer cells differs from typical caspase-dependent apoptosis. We examined cell autophagy and its involvement in this type of cell death.
Autophagy is characterized by an increase in the number of autophagosomes, vesicles that surround such cellular organelles as the Golgi complex, mitochondria, polyribosomes and endoplasmic reticulum. The role of autophagy in cell death remains controversial and in some contexts autophagy might even promote cell survival through a number of mechanisms, such as generating nutrients and energy to sustain starving cells or stressed cells, degrading toxic proteins and protecting cells from oxidative stress.[27-29] Most of these protective functions of autophagy are responsive and temporary. In contrast, cancer cells generally tend to undergo a lower level of autophagy than their normal counterparts, which supports the concept that defects in autophagic cell death play a role in the process of tumorigenesis. Several autophagy regulators are downregulated in human cancers. In fact, the importance of autophagy in tumor suppression has been supported by studies of tumors in Beclin 1 ± mice.[30, 31] It was also reported that Beclin-1 is monoallelically deleted in 40–75% of human breast, prostate and ovarian tumors. Previous studies have demonstrated that autophagy induces cancer cell death.[21, 33, 34]
Multiple models have been proposed for the progression from autophagy towards cell death; one of the models is that autophagosomes merge with lysosomes and digest organelles, leading to cell death (autophagic cell death). In contrast to apoptosis, autophagic cell death is caspase independent. Our data show that a reduction of CDCP1 led to autophagy by inducing the LC3-II protein and GFP-LC3 dots in lung cancer cells cultured under suspension conditions (Fig. 2). Electron microscopy revealed that massive formation of large vacuoles is induced by the loss of CDCP1 in suspension, while typical autophagosome-like bilayer structures were not clearly observed (Fig. 2c). Because autophagic markers such as LC3-II were observed under these conditions, it might be possible that the autophagosome structure was destroyed during the process of the formation of prominent large vacuoles. Expression of phosphorylated CDCP1 reduced autophagic markers; in contrast, an SFK inhibitor induced LC3-II protein (data not shown) and GFP-LC3 dots (Fig. 3b), which supports previous reports regarding the effects of Src inhibitors on autophagy. Because CDCP1 is a major substrate of SFK in suspended cancer cells,[11, 24] CDCP1 might be a key regulator of autophagy controlled by SFK. It has been reported that induction of autophagy was observed in T lymphocytes by the selective knockdown of Fyn. In addition, the SFK inhibitor induces autophagy in cancer cells.[33, 36, 37] These reports might also intensify the potential role of CDCP1, which is a main substrate of SFK as a general regulator of autophagy.
Recent reports suggest that cleavage of CDCP1 is critical for activation of CDCP1 signaling.[38-41] Cleaved CDCP1 was observed in A549 cells (Fig. 1a) and autophagy induced by CDCP1 siRNA was rescued by NtCDCP1, which mimics the cleaved form of CDCP1 (Fig. 3c). Results of the current study support that cleaved CDCP1 might also have an essential role in suppression of autophagy. It has been revealed that tyrosine phosphorylation of CDCP1 is required for binding to the unique C2 domain of PKCδ. The association between phosphorylated CDCP1 and PKCδ causes enzymatic activation of PKCδ, resulting in multiple malignant phenotypes in cancer cells including anoikis resistance, cell migration and matrix degradation.[6, 8, 11] By examining autophagy induced by suspension using siRNA and a rottlerin, we found that PKCδ can act as a downstream effector of CDCP1 also for autophagy, which is supported by a previous report showing that PKCδ constitutively suppresses autophagy and inhibits autophagic cell death in cancer cells. As one possible mechanism, expression of TG2, which is reported to suppress autophagy, was reduced by knockdown of CDCP1 during induction of autophagy in several cancer cells (Fig. S2). Therefore, it is possible that induction of TG2 by CDCP1-PKCδ signaling might have a role in the suppression of autophagy in a wide range of cancers. The results of the current study and our previous study demonstrate that CDCP1-PKCδ signaling is essential for both suppression of autophagy and cell death in suspension cultures of cancer cells. Finally, we revealed that anoikis induced by suppression of either CDCP1 or PKCδ in cancer cells can be prevented by inhibiting autophagy with 3MA (Fig. 4), showing that autophagy causes anoikis following the loss of CDCP1. Our data show that CDCP1-PKCδ signaling might contribute to tumor metastasis by protecting matrix-detached cancer cells from autophagic cell death.
It is still under investigation whether anoikis is generally associated with autophagy. Our results show that the induction of LC3-II and aggregation of GFP-LC3 in cell death is induced in a suspension culture of anchorage-dependent cancer cells (Figs 3a,4b), indicating the possibility that autophagy is widely involved in suspension-induced cell death. Further studies are required to determine the exact mechanism through which autophagy induces anoikis in various types of cells. Our results might suggest that induction of autophagy can act as protective force for cancer metastasis through anoikis sensitization.
Our recent studies demonstrated that CDCP1 is required for various malignant aspects of cancer cells during the progression of human cancers and is a promising therapeutic target of these cancers. Detection of autophagy might be a good indicator of therapeutic effects when targeting the CDCP1-PKCδ pathway.