If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Oncology Institute, Loyola University Medical Center, Cardinal Bernardin Cancer Center Room 301, 2160 S. First Ave, Building 112, Maywood, IL 60153-5385, USA.
Notch receptor-mediated intracellular events represent an ancient cell signaling system, and alterations in Notch expression are associated with various malignancies in which Notch may function as an oncogene or less commonly as a tumor suppressor. Notch signaling regulates cell fate decisions in the epidermis, including influencing stem cell dynamics and growth/differentiation control of cells in skin. Because of increasing evidence that the Notch signaling network is deregulated in human malignancies, Notch receptors have become attractive targets for selective killing of malignant cells. Compared with proliferating normal human melanocytes, melanoma cell lines are characterized by markedly enhanced levels of activated Notch-1 receptor. By using a small molecule gamma-secretase inhibitor (GSI) consisting of a tripeptide aldehyde, N-benzyloxycarbonyl-Leu-Leu-Nle-CHO, which can block processing and activation of all four different Notch receptors, we identified a specific apoptotic vulnerability in melanoma cells. GSI triggers apoptosis in melanoma cells, but only G2/M growth arrest in melanocytes without subsequent cell death. Moreover, GSI treatment induced a pro-apoptotic BH3-only protein, NOXA, in melanoma cells but not in normal melanocytes. The use of GSI to induce NOXA induction overcomes the apoptotic resistance of melanoma cells, which commonly express numerous cell survival proteins such as Mcl-1, Bcl-2, and survivin. Taken together, these results highlight the concept of synthetic lethality in which exposure to GSI, in combination with melanoma cells overexpressing activated Notch receptors, has lethal consequences, producing selective killing of melanoma cells, while sparing normal melanocytes. By identifying signaling pathways that contribute to the transformation of melanoma cells (e.g. Notch signaling), and anti-cancer agents that achieve tumor selectivity (e.g., GSI-induced NOXA), this experimental approach provides a useful framework for future therapeutic strategies in cutaneous oncology.
The Notch signaling pathway plays a key role in cell fate decisions during development, and also regulates cell-to-cell interactions involving stem cell behavior, cell proliferation, differentiation, and death (
). Notch was originally described by Thomas Hunt Morgan in 1917 because Drosophila-bearing mutations in Notch had a distinctive phenotype in which the structure of the wing blades was irregular in shape (
). Dr Morgan received the 1933 Nobel Prize for his discoveries concerning the role played by the chromosome in heredity. Note how, compared with the smooth edge of normal wings, hemizygous Notch deficiency imparts a serrated-like appearance to the edge of the wing Figure 1a. The actual gene responsible for producing the distinctive phenotype was cloned in 1985, and it was recognized that this sequence encoded a cell surface receptor (
). Focusing on Notch by investigative skin biologists is not a coincidence, as homozygous Notch deficiency (Notch null genotype) characteristically causes a “neurogenic phenotype” in which the number of neurons in Drosophila was increased secondary to a decrease in epidermal cells. This is due to unrestrained differentiation of neuroectodermal cells toward the neural phenotype (
). Thus, Notch signaling has continued to attract interest by cutaneous biologists since it first received notoriety in 1937, when Poulson initially described this neurogenic cell fate switch involving Drosophila's ectodermal cuticle (
Figure 1Historical perspective and overview of Notch signaling. (A) Phenotypic appearance of Drosophila wings in normal, wild type as well as in flies bearing hemizygous Notch mutation. Note the serrated, rather than smooth contour, on the wing edges in the mutant fly wings as first described by Thomas Morgan in 1917. (B) Diverse roles for Notch signaling that influence various cell fate decisions including four different Notch receptors and multiple Notch ligands. (C) Simplified overview of Notch signaling pathway, in which the activation of the Notch receptor in the plasma membrane results in regulated proteolysis involving a γ-secretase/presenilin complex to generate a truncated intracellular form (e.g., NIC) that translocates to the nucleus and functions as a transcription factor to influence gene expression including HES and other targets (see text for further explanation).
Besides its role as a highly evolutionarily conserved pathway regulating epithelial development, Notch signaling has also been implicated beyond the epidermis to include hair follicles and feathers (
). The Notch receptors are designated as Notch-1 through Notch-4, and the different ligands include Jagged-1, Jagged-2, and Delta-like (DLL1, DLL3, DLL4—see Figure 1b). The four known Notch receptors differ by the number of epidermal growth factor-like (EGF-like) repeats in the extracellular domain, as well as by the length of the intracellular domain. The cell fates regulated by Notch signaling include renewal of stem cells, and control of proliferation, differentiation, and apoptosis Figure 1b.
Figure 1c presents an overview of Notch signaling with a simplified portrayal of intracellular biochemical events that participate in Notch signaling (
). Regulated proteolysis plays a major role in Notch signaling. Notch receptors are produced as large single-chain precursors, which are cleaved by serine proteases related to furin in the trans-Golgi to generate calcium-stabilized heterodimeric receptors consisting of an extracellular subunit and a trans-membrane subunit. Once a cell expressing a Notch receptor is stimulated by an adjacent cell via a Notch ligand on the cell surface, the extracellular subunit is trans-endocytosed in the ligand-expressing cell. The remaining transmembrane subunit undergoes two consecutive enzymatically mediated cleavages. The first activating cleavage event is mediated by a metalloprotease-dependent TNF-α converting enzyme (TACE) belonging to the disintegrin and metalloprotease family (
) to generate an intracellular truncated version of the receptor designated as NICFigure 1. Very recently, it was discovered that γ-secretase cleavage requires and is preceded by mono-ubiquitination and endocytosis (
). Thus, the rate of cleavage of Notch-1 is finely modulated by multiple post-translational modifications and cellular compartmentalization events.
The truncated or intracellular form of the Notch-1 receptor (e.g., NIC-1) can then translocate to the nucleus, where it forms a multimeric complex with a highly conserved transcription factor (CBF1), and other transcriptional co-activators that influence the intensity and duration of Notch signals (
). CBF-1 exists as a repressor in the absence of Notch-1, forming a complex with adaptor protein SKIP and one or more co-repressors, including SMRT (signaling mediator of retinoid and thyroid receptor), N-CoR and/or CIR (
), and histone deacetylase 1 (HDAC1). Notch-1 dissociates the co-repressors and HDAC1 from CBF-1, instead recruiting co-activator MAML1 (Mastermind in Drosophila), and histone acetyltransferases P/CAF, GCN5, and/or p300. The ultimate result is activation of transcription at promoters containing CBF-1-responsive elements. Since many of the genes regulated in this manner are themselves transcription factors, this process can ultimately result in induction or repression of numerous genes. Thus, the ultimate outcome for Notch signaling is to either stimulate or repress various target genes (
). Three of the most commonly studied target genes for Notch signaling include Hairy-Enhancer of split (HES), and HES-related proteins (HERP/HEY) as well as p21WAF1, although cross-talk between Notch and the Stat3 signaling pathway has been demonstrated as well (
In addition to this “classical” pathway of Notch signaling, other pathways are beginning to emerge. The role of Notch-associated protein Deltex remains poorly understood. Deltex, of which there are several mammalian isoforms, binds to the ankyrin repeats of Notch and has been proposed to act as a downstream Notch mediator involved in JNK inhibition, as a nuclear mediator of Notch signaling, and as a feedback inhibitor of Notch signaling (
). More recently, convincing evidence has been presented that Notch-1 acts as a kinase activation scaffold, forming a complex with Src-family protein tyrosine kinase Lck and phosphoinositide-3-kinase (PI3K). This process leads to activation of the PI3K-AKT pathway (
) and may be responsible for survival signals mediated by Notch activation. There is considerable evidence of cross-talk between the Notch and NF-κB pathways, but the molecular details remain unclear. Transcriptional induction of all five NF-κB subunits (
Given the multitude of cell fate-regulatory processes modulated by Notch signaling and the number of cell fate decisions and post-developmental influences that Notch signaling mediates in a diverse array of cell types, it is not surprising that Notch plays a complex role in pathophysiology, and that multiple positive and negative regulatory circuits tightly control the activity of Notch receptors. In the next section, we review the role of Notch signaling in a physiological context (i.e., epidermal keratinocyte differentiation), as well as in pathology, with special emphasis on deregulation of Notch signaling in various neoplastic conditions (
Notch Signaling as a Proto-Oncogene or Tumor Suppressor
Over the past two decades, many processes have been identified in which Notch signaling is operative in normal cells to maintain tissue homeostasis. Furthermore, with regard to tumorigenesis, Notch may play a role as either a tumor suppressor or as an oncogene (
). In this section, we briefly review the role of Notch signaling in the differentiation of epidermal keratinocytes Figure 2a, as well as the evidence that persistent activation of Notch signaling can produce malignancies such as human T cell acute lymphoblastic leukemias Figure 2b. In general, Notch signaling tends to delay and restrict differentiation pathways, thereby maintaining cells in an undifferentiated state (
Figure 2Opposite roles of Notch-1 receptor signaling in keratinocytes as a tumor suppressor, or in T-ALL cells as a proto-oncogene. (A) As Notch-1 receptor activation triggers growth arrest and differentiation in normal epidermal keratinocytes, it can function as a tumor suppressor. Presumably highly regulated interaction between Jagged-1-expressing keratinocytes can deliver a physiological stimulus induced by Notch-1 receptors to trigger differentiation in healthy skin. (B) Either a chromosomal translocation producing enhanced levels of activated Notch-1 receptors, or point mutations in the Notch-1 receptor that enhance γ-secretase-mediated activation of Notch-1 receptors, generate sustained Notch signaling culminating in tumorigenesis. (T cell acute lymphoblastic leukemia, T-ALL). Although a chromosomal translocation is rare, occurring in approximately 10% of patients with T-ALL, the point mutations are observed in over 50% of T-ALL patients (
An important exception to this general rule is the epidermal keratinocyte, because we and others have clearly demonstrated that activation of Notch signaling promotes epidermal keratinocyte differentiation (
). Characterizing expression patterns for Notch-1 in human tissues has revealed that in the skin, Notch-1 was expressed by hair follicles, sebaceous glands, and sweat glands (
). All four Notch receptors are expressed in the epidermis, with Notch-2 being more basal than Notch-1, and Notch-3 and -4 being more superficial than Notch-1. Notch ligand Delta-1 is expressed in “epidermal stem cells” (
showed that in murine keratinocytes, Notch-1 induces cell cycle arrest and early differentiation, whereas inhibiting late differentiation. Using a submerged epidermal equivalent model system, we demonstrated that activation of Notch signaling (very likely, all four Notch receptors in sequence) is necessary and sufficient for triggering terminal differentiation, including cornification (
Thus, in the skin, Notch-1 signaling could function in keratinocytes as a tumor suppressor because of its ability to trigger induction of differentiation Figure 2a. The precise roles of the other three Notch receptors in epidermopoiesis remain unknown and are presently under investigation. Additional and more direct evidence that Notch-1 can function as a tumor suppressor is derived from in vivo studies in which the long-term consequences of Notch-1 deficiency were characterized using a chemical carcinogenesis protocol (
). Using a tissue-specific, gene-targeting approach, Notch-1 -/- mice were found to have enhanced β-catenin signaling and increased frequency of skin tumors resembling basal cell carcinoma and squamous cell carcinoma. This is consistent with the anti-proliferative role of Notch-1 in murine keratinocytes. Whether Notch-1 is a bona-fide tumor suppressor in the epidermis, i.e., whether functional inactivation of Notch-1 is required for BCC or SCC progression in humans, is not clear. Selective loss of Notch-1 in advanced cervical cancer and cervical cancer cell lines had been reported (
Specific down-modulation of Notch1 signaling in cervical cancer cells is required for sustained HPV-E6/E7 expression and late steps of malignant transformation.
) have shown that not only is Notch-1 expressed in advanced cervical cancer cells and clinical specimens, but it is necessary for survival of cervical cancer cells through its anti-apoptotic activity.
Thus, it appears that the role of Notch-1 in keratinocytes is not so simple (
), as one group has reported that forced overexpression of activated Notch-1 can participate with HPV oncoproteins E6 and E7 to transform primary human keratinocytes (
Activated Notch1 inhibits p53-induced apoptosis and sustains transformation by human papillomavirus type 16 E6 and E7 oncogenes through a PI3K-PKB/Akt-dependent pathway.
). The work of Lathion et al shed some light on this apparent contradiction. These authors showed that intracellular Notch-1 expressed from a very strong (i.e., cytomegalovirus) promoter at very high levels inhibits the expression of E6 and E7 HPV oncoproteins, as reported by Talora et al. When expressed in more moderate amounts from a Rous sarcoma virus promoter, however, intracellular Notch did not inhibit E6 or E7 expression, and transformed keratinocytes in cooperation with them. These findings underscore two crucial characteristics of Notch signaling, concentration dependence and context dependence. The effects of exogenously overexpressed Notch receptors do not necessarily correspond to those of the spontaneously expressed proteins. Additionally, the co-presence of other oncogenic alterations (such as inactivation of retinoblastoma (Rb) and p53 by viral oncoproteins) affects the outcome of Notch signaling. This context dependence has relevance to the concept of synthetic lethality in that it suggests possible molecular explanations for differential sensitivity of neoplastic versus normal cells to Notch inhibition. In this case, inactivation of Rb and p53 results in the disabling of the G1 restriction point, thus limiting the effects of p21 induction by Notch-1, while synergizing with the anti-apoptotic signals mediated by Notch-1. It should be noted that with the possible exception of T-cells, intracellular Notch-1 (and Notch-2) transforms cells of several lineages in the co-presence of viral oncoproteins that inactivate Rb and p53 (
). It should be kept in mind that the end result of Notch signaling is dictated by the cellular context, and the signal strength and duration, which in turn are regulated by post-translational modifications and pathways that influence receptor recycling and membrane availability (
). In the following paragraph, we briefly review evidence linking excessive or sustained Notch receptor activation to its proto-oncogene activity in other cell types.
A recent report documented the presence of point mutations in strategic sites within Notch-1 that ultimately produced a receptor that was more readily processed by γ-secretase, leading to a sustained Notch-1 receptor signaling in approximately 50% of the cases of T cell acute lymphoblastic leukemias (
). This makes it the most frequently mutated oncogene in this disease Figure 2b. Earlier, a chromosomal translocation (t 7; 9) was identified in a rare (∼10%) subset of T-ALL patients that also resulted in elevated Notch-1 receptor signaling (
). Using a mouse model, forced overexpression of the truncated form of the human Notch-1 gene in the hematopoietic compartment also produced T cell leukemia (
). This consisted of CD4+, CD8+T-cells, and appeared after a latency period, suggesting that other mutations are required. Besides T-ALL, there are many other examples of deregulated expression of wild-type Notch receptors in human tumors including: head and neck carcinomas, renal carcinomas, endometrial carcinomas, mesothelioma, breast, lung, and pancreatic carcinomas, as well as chronic lymphoid and a subset of acute myeloid leukemias, anaplastic large cell non-Hodgkin lymphomas, and a subtype of Hodgkin lymphoma (
). As regards interactions between Notch and other oncogenic pathways of relevance to investigative skin biologists, it should be noted that cross-talk involving the Wnt signaling pathway as well as the Sonic-Hedgehog pathway has been observed (
). In the next section, a brief review of the mechanism by which Ras and Notch signaling can influence the neoplastic phenotype in melanoma will be presented.
Ras and Notch Signaling in Melanoma
Prior to the discovery of the aforementioned activating point mutations in Notch-1 found in T-ALL patients, several laboratories had focused on the interaction between Ras signaling and the Notch pathway Figure 3. It was discovered that oncogenic Ras could enhance Notch signaling by at least two distinct, but complementary, pathways (
). The first way in which oncogenic Ras can enhance Notch signaling is by increasing the level of Notch ligands such as Delta-1, and the second mechanism is by increasing the levels of the γ-secretase/presenilin 1 complex (
). It also appears that Ras-mediated transformation requires the presence of intact Notch signaling, and if such Notch receptor signaling is disrupted, the ability of Ras to transform cells is significantly reduced (
). Although the precise mechanistic details by which Notch signaling can contribute to tumorigenesis remain to be defined, several interesting insights have been gained by focusing on cell survival pathways. It appears that Notch signaling can enhance the survival of cells by interacting with the PKB/AKT (
Activated Notch1 signaling cooperates with papillomavirus oncogenes in transformation and generates resistance to apoptosis on matrix withdrawal through PKB/Akt.
Figure 3Role of cross-talk between oncogenic Ras and Notch receptors in melanoma, and portrayal of therapeutic targets derived from this signaling pathway. Note that there are at least two mechanisms by which activated Ras can enhance Notch signaling: either by inducing Notch ligand expression, or by enhancing γ-secretase activity. Transcriptional targets for Notch signaling in tumor cells include Hes-1, STAT3, NF-κB, and p21. Therapeutic approaches for interrupting Notch signaling include use of GSI (γ-secretase inhibitor), or a genetic silencing approach featuring RNAi targeting individual Notch receptors (see text for further explanation).
). Figure 4 provides a representative phase-contrast microscopic appearance of normal proliferating human melanocytes, as well as a panel of ten different melanoma cell lines. This collection of melanoma cells includes early-passage (<50 passages; RJ002L, BO-003) as well as many late-passage cells (>50 passages; C8161, MUM2B, MUM2C, C81-61, OCM1A, SK-Mel 5, SK-Mel 100, and SK-Mel 147). Note that normal melanocytes have elongated cytoplasmic processes resembling neuronal dendrites, whereas the melanoma cell lines appear either epithelioid or spindle shaped and display a tendency to pile up as they proliferate. As can be appreciated by the differences in the appearance of the cultured cells, morphology does not predict Notch expression as detailed in Figure 4 and Figure 5. CEM cells grow in suspension and are also portrayed in Figure 4 as they represent a cell line with markedly enhanced Notch-1 levels as mentioned earlier for T-ALL cells (
Figure 4Phase-contrast microscopic appearance of proliferating human melanocytes derived from normal skin, and a panel of melanoma cell lines derived from patients with metastatic melanoma. (Magnification × 85) (A) Normal human melanocytes. (B–K) Melanoma cell lines. (L) T-ALL cell line: CEM cells. Note that normal human melanocytes were derived from neonatal foreskins and grown in standard medium (
Figure 5Status of Notch-1 receptor levels and activation in normal human melanocytes and melanocytes and melanoma cell lines. Western blot analysis of ten different normal human melanocyte cultures and ten different human melanoma cell lines examined for activated Notch-1 receptors (acute T cell lymphoblastic leukemia line; CEM cells containing activating point mutations in Notch-1 as a positive control). Designation of normal melanocytes represents ten different neonatal foreskins, as follows: lane 1, MC-07; lane 2, MC-08; lane 3, MC-06; lane 4, MC-04; lane 5, MC-05; lane 6, MC-010; lane 7, MC-011; lane 8, MC-012, lane 9, MC-013; and lane 10, MC-014. Lane numbers designated for the melanoma cell lines are as follows: lane 1, C8161; lane 2, MUM2B; lane 3, RJ-002L; lane 4, BO-003; lane 5, MUM2C; lane 6, C81-61; lane 7, OCM1A; lane 8, SK-Mel 5, lane 9, SK-Mel 100; and lane 10, SK-Mel 147. Whole-cell extracts (50 μg protein per lane) were prepared as previously described (
Abnormalities involving several important proteins have been uncovered in the pathogenesis of melanoma including: inactivation of p16INK4a, cyclin-dependent kinase inhibitor (CDK4/6) with suppression of Rb protein, silencing of the pro-apoptotic protein APAF-1, activating mutations in B-Raf and N-Ras, and constitutive activation of mitogenic receptor signaling pathways including basic fibroblast growth factor receptor (
Loss of expression of the p16/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in melanocytic lesions correlates with invasive stage of tumor progression.
). Two of the most common molecular lesions in melanomas, namely, activation of Ras signaling and inactivation of the G1 checkpoint, can potentially result in enhanced Notch signaling (Ras) and co-operate with the transforming activity of Notch-1. As discussed earlier, whereas there are multiple Notch receptors and ligands, due to space constraints, this review will focus predominantly on Notch-1 receptor activation in melanoma as a paradigm for understanding the role of Notch signaling in carcinogenesis.
Since melanocytes arise from the neural crest, and because Wnt as well as Notch developmental signaling cascades regulate central nervous system stem cell dynamics, it was possible that Notch signaling may be activated in melanoma cells, as well as in various neural crest-derived brain tumors (
). In melanocytes and melanoma cells, one of the first steps in determining whether Notch signaling may have been relevant in the pathogenesis of melanoma is to assess the relative levels and state of activation for Notch receptors in both melanocytes as well as melanoma cells Figure 5. In two earlier reports, we initially used an immunohistochemical staining approach, and although we could not detect any of the four different Notch receptors to be expressed by melanocytes in normal or diseased skin, enhanced immunoreactivity was detected for all four different Notch receptors in melanoma lesions (
). Interest in Notch signaling in melanoma derived from the observation that some aggressive melanoma cell lines displayed a stem cell-like plasticity, and the question as to a role for Notch signaling as an underlying molecular mechanism for the trans-differentiation between melanoma cells and endothelial cells arose from such studies (
), the enhanced levels of Notch receptors in melanoma were consistent with this developmental-related theme for melanoma pathogenesis. To extend our immunostaining results, a polyclonal rabbit-derived antibody detecting the activated state of Notch-1 (Abcam, Cambridge, Massachusetts) was used to probe whole-cell extracts. No activated Notch-1 was detected by western blot analysis in any of the proliferating normal human melanocytes derived from ten different foreskins Figure 5. Using ten different melanoma cell lines, however, activated Notch-1 was detected in all of the cell lines, albeit at variable levels of overexpression and irrespective of their morphology (spindle shaped or epithelioid) as shown in Figure 4. The T-ALL cell line, CEM, was used as a positive control as it contains abundant activated Notch-1. As regards the overexpression of Notch receptors that we have detected in human melanomas, another group using a microarray approach has also independently verified enhanced Notch mRNA levels in melanoma (
), and our recently completed real-time quantitative polymerase chain reaction analysis of these ten melanoma cell lines confirmed elevated levels for all four Notch receptor mRNAs compared with normal proliferating melanocytes.
B. J. Nickoloff, A. Weeraratna, unpublished observations, 2004.
Concept of Synthetic Lethality
As one considers the biology of oncogenic events, it is important to note that, in general, most cancer-associated growth-promoting signals are directly coupled to increased cell death (
). Thus, initial oncogenic events can only have a net tumor growth effect when they are complemented by, or cooperate with, anti-apoptotic pathways in tumor cells. Identification of this specific apoptotic vulnerability in melanoma can then be targeted to achieve selective killing of malignant cells while sparing non-transformed (i.e., normal) cells (
). This approach has been termed synthetic lethality because it emphasizes combining a drug that does not affect normal cells that contain wild-type alleles, with induction of death in malignant cells containing a specific genetic alteration that renders the tumor cell susceptible to induction of apoptosis (
). Figure 6 provides an overview of these concepts including the pharmacogenomic strategy as applied to synthetic lethality in melanoma. The term “pharmacogenomic” is used here not to indicate the study of inter-individual genetic variability in drug susceptibility, but to indicate the genetic basis for differential drug susceptibility between normal cells and cancer cells.
Figure 6Overview of pharmacogenomic strategy for synthetic lethality including use of gamma-secretase inhibitor (GSI) to achieve selective killing of melanoma cells, but not normal melanocytes. The initial proto-oncogenic activation (portrayed by c-myc) within a normal cell (green highlights) induces both enhanced proliferation that is offset by enhanced apoptosis (portrayed by increased death receptor-DR5 levels) yielding no net tumor formation (yellow highlights). When additional genetic alterations are present (portrayed by increased Bcl-x, NF-κB, or AKT), the balance between proliferation and apoptosis is tipped toward enhanced survival pathways leading to tumor formation (red highlights). An ideal selective chemotherapeutic agent such as GSI (N-benzyloxycarboxy Leu-Leu-Nle-CHO) can trigger growth arrest, but not apoptosis in normal melanocytes, whereas GSI induces prominent killing of melanoma cell lines (
). An important molecular determinant for selective killing of melanoma cells versus melanocytes is the ability of GSI to induce the proapoptotic BH3-only protein NOXA in melanoma cells, but not in normal melanocytes. One potentially important characteristic of melanoma cells is their activation of Notch receptors relative to melanocytes, although the specific link between Notch signaling and GSI-mediated induction of NOXA remains to be defined. This figure is derived from several publications including
As portrayed in Figure 6, a normal cell is depicted with a properly balanced set of proteins regulating proliferation and apoptosis, but following proto-oncogene activation (e.g., c-myc activation), there is enhancement in both proliferation and apoptosis such that no net tumor is formed. Following an additional genetic alteration such as increased Bcl-xL, NF-κB, or AKT, the enhanced survival pathway overcomes the pro-apoptotic tendency to shift the balance in favor of tumor formation. There are several examples in which an oncogenic genetic alteration can sensitize the tumor cells to induction of apoptosis including: activation and deregulation of MYC that leads to either sustained E2F-1 levels that trigger caspase activation and apoptosis, or enhanced death receptor (DR5)-mediated susceptibility (
). By using agents that can exploit the hypersensitivity of malignant cells, a pharmacogenomics approach that is analogous to the concept of genetic synthetic lethality can be envisioned for a molecularly targeted therapeutic application. In this review, data are presented indicating that Notch receptor activation is abnormal in melanoma cells, and in the next section focus is directed toward the use of an agent initially selected to block Notch signaling. This line of inquiry led to the discovery of a novel apoptotic pathway featuring induction of a BH3-only pro-apoptotic protein known as NOXA.
Use of gamma-secretase inhibitor (GSI) to Induce NOXA in Melanoma Cells
Since all four different Notch receptors appear to be activated in melanoma lesions (
). The chemical structure of GSI is depicted in Figure 6. Since the β-amyloid peptide in senile plaques requires a cleavage step similar to Notch receptor processing, many such inhibitors are already available, and some are in clinical trials that will provide valuable patient tolerability and safety data for future drug development in melanoma patients (
), when proliferating melanocytes or melanoma cells were exposed to GSI, the melanocytes underwent a G2/M growth arrest, but the melanoma cells underwent rapid and prominent apoptosis, as summarized in the bottom portion of Figure 6.
Next, studies were designed to define the mechanism of action for the selective apoptotic response of GSI in melanoma cells (
). To identify the apoptotic pathway, both melanocytes and melanoma cells were studied, and the kinetics of induction of apoptosis following addition of GSI to melanoma cells was determined. Using a panel of five different melanoma cell lines, including an early-passage metastatic melanoma cell line (RJ002L) or four different late-passage melanoma cell lines (C8161, MUM2B, SK-Mel-28, SK-Mel-100), the induction of apoptosis became detectable by Annexin V-FITC staining and flow cytometry approximately 6–8 h after exposure to GSI. To determine whether this apoptosis required new protein synthesis of pro-apoptotic factor(s), or could be attributed to the loss of survival factor(s), melanoma cells were pre-treated with cycloheximide (1 μg per mL; 1 h) to block protein synthesis, and then exposed to GSI. The inhibition of new protein synthesis almost completely blocked the GSI-mediated induction of apoptosis. This result strongly pointed toward induction of a pro-apoptotic signal following GSI, so whole-cell extracts were prepared before and at various time intervals (1, 3, 6, 18, 24 h) following GSI exposure. Western blots were probed for numerous survival and apoptotic proteins, and it was discovered that the most rapid (approximately 3 h) and prominent pro-apoptotic protein induced by GSI was the BH3-only protein, NOXA. Another BH3-only protein, Bim, was also induced but at later time points (18, 24 h). Western blots for the melanoma cells revealed constitutive levels of the important apoptotic proteins Bax and Bak, as well as high levels for survival proteins Bcl-2, Mcl-1, and survivin.
One reason why it took such a long time to discover the basis for induction of NOXA was previous reports implying that p53 was necessary for NOXA induction (
). Since several of the melanoma cell lines contained mutations in both p53 alleles, the initial survey of candidate proteins did not include p53-inducible proteins such as NOXA. But, once the rapid and prominent induction of NOXA was detected using two different antibodies that correctly detected a single protein of the correct molecular mass, we next asked four important questions. First, was GSI-induced NOXA p53 dependent? Second, did GSI induce NOXA in normal melanocytes? Third, did other agents used to treat melanoma or other cancers induce NOXA? Fourth, did blocking NOXA induction reduce the apoptotic response in melanoma cells? To determine whether p53 was required for NOXA induction, we initially reduced p53 levels using an siRNA approach and while inhibiting GSI-induced levels of two p53-dependent proteins (GADD45, MDM2), no reduction in GSI-induced NOXA was observed. Next, we used two other cell lines known to be p53 null (e.g., PC-3 prostate cancer cells, SAOS-2 osteogenic sarcoma cells), as well as verified that two of our melanoma cell lines (MUM2B, SK-Mel-28) were p53 null, and observed that whereas no p53 protein was detected, GSI could still induce NOXA and trigger cell death. When normal proliferating melanocytes were treated with GSI, no NOXA was induced, although the cells underwent a prominent G2/M growth arrest. Treatment of melanoma cells with genotoxic agents such as adriamycin (1 μg per mL), etoposide (10 μg per mL), or UV light (30 mJ per cm2), induced a slight apoptotic response, but no expression of NOXA. Finally, when an antisense oligonucleotide targeting NOXA was used to prevent NOXA induction, the apoptotic response to GSI was reduced by approximately 30%–50% depending on the cell line utilized. Taken together, these results strongly supported the conclusion that GSI was inducing apoptosis in melanoma cell lines by engaging a previously uncharacterized apoptotic pathway that was, in part, dependent on the BH3-only protein, NOXA. Importantly, GSI caused striking cell death and NOXA induction in melanoma xenografts as well, without causing observable toxicity to nude mice. This suggests that it may be possible to identify a therapeutic window for use of GSI in melanoma.
The final aspect of this study was to further delineate the sequence of molecular events occurring in the melanoma cells once NOXA was induced. As NOXA is a BH3-only protein, it can bind the pro-survival proteins Mcl-1, Bcl-2, and Bcl-x, thus facilitating interaction between the pro-apoptotic proteins Bax and Bak located in the mitochondrial membrane Figure 7. Evidence for this interaction leading to mitochondrial dysfunction included release of cytochrome C from the mitochondria into the cytoplasm. Once cytochrome C is released, it can facilitate creation of the apoptotic complex in which APAF-1 is recruited, leading to caspase 9 activation, followed by activation of caspase 3, cleavage of PARP and other “end-stage” substrates, and culminating in DNA degradation and cell death. Interestingly, whereas APAF-1 levels were barely detectable in several melanoma cell lines, and the cells had high levels of anti-apoptotic factors (e.g. Bcl-2, Mcl-1, survivin), the induction of NOXA was sufficient to overcome all of these obstacles to trigger a prominent apoptotic response in all of the melanoma cell lines irrespective of their p53 status.
Figure 7Proposed apoptotic mechanism in which melanoma cells are treated with gamma-secretase inhibitor (GSI). Note that the primary effect is p53 independent, resulting in the requirement for new protein synthesis of pro-apoptotic proteins such as Bim and NOXA. The ability of NOXA to bind and displace Mcl-1, Bcl-2, and Bcl-xL is highlighted. Once these survival proteins are displaced, the oligomerization of Bax and Bak occurs in the outer mitochondrial membrane, triggering the release of various pro-apoptotic factors such as cytochrome C, Apaf-1, SMAC, and AIF, all of which can participate in the final aspects of the intrinsic apoptotic pathway in which DNA is degraded. A secondary, amplification loop, is also illustrated highlighting the participation of caspase 8 and tBid.
Further studies are required to more fully determine whether other stimuli (besides GSI) can trigger NOXA induction in a p53-independent fashion in melanoma cells, and to elucidate the mechanism responsible for the inability of GSI to induce NOXA in normal melanocytes. It will also be necessary to determine whether GSI induction of NOXA results from inhibition of Notch signaling alone, given that γ-secretase has numerous substrates, including: ErbB4, syndecan, CD44, and others (
Taken together, these results indicate that Notch receptor activation is a characteristic of melanoma cells, which distinguishes the malignant cells from normal melanocytes. Moreover, by using an agent that was initially chosen to target the activation of all Notch receptors, selective apoptosis of melanoma cells could be achieved while sparing normal melanocytes. The selective killing of melanoma cells was attributed, in part, to induction of NOXA, as GSI triggered the pro-apoptotic BH3-only protein NOXA exclusively in melanoma cells and not melanocytes. Moreover, the use of an antisense oligonucleotide targeting NOXA significantly reduced the GSI-induced apoptosis in melanoma cells.
Future studies are indicated to determine the precise mechanism whereby Notch receptor-mediated signaling pathways provide a survival advantage to melanoma cells. It will be important to elucidate the relationship between reduction in Notch signaling, and engagement of the apoptotic machinery in melanoma cells. Finally, devising additional reagents that can selectively target NOXA induction in melanoma cells, while sparing normal melanocytes, will provide new therapeutic options for patients with melanoma.
ACKNOWLEDGMENTS
This work was supported by NIH grants CA 59702, CA 80318 (M. J. C. H.), PO1 CA 27502 (P. M. P., J. M. T., B. J. N.), CA 84065 (L. M.), and PO1 CA 59327 (BJN). Larry Stennett and Barb Bodner performed the tissue culture and western blot analysis. Lynn Walter and Brian Bonish prepared the text and figures.
References
Alsina J.
Gorsk D.H.
Germino F.J.
Detection of mutations in the mitogen-activated protein kinase pathway in human melanoma.
Activated Notch1 inhibits p53-induced apoptosis and sustains transformation by human papillomavirus type 16 E6 and E7 oncogenes through a PI3K-PKB/Akt-dependent pathway.
Activated Notch1 signaling cooperates with papillomavirus oncogenes in transformation and generates resistance to apoptosis on matrix withdrawal through PKB/Akt.
Loss of expression of the p16/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in melanocytic lesions correlates with invasive stage of tumor progression.
Specific down-modulation of Notch1 signaling in cervical cancer cells is required for sustained HPV-E6/E7 expression and late steps of malignant transformation.
52570-0&id=gr1.jpg){kind=link}
52570-0&id=gr2.jpg){kind=link}
52570-0&id=gr3.jpg){kind=link}
52570-0&id=gr4.jpg){kind=link}
52570-0&id=gr5.jpg){kind=link}
52570-0&id=gr6.jpg){kind=link}
52570-0&id=gr7.jpg){kind=link}