ING3 Promotes UV-induced Apoptosis via Fas/Caspase-8 Pathway in Melanoma Cells*

The novel ING tumor-suppressor family proteins (ING1-5) have been discovered during the past decade and are recognized as the regulators of transcription, cell cycle checkpoints, DNA repair, apoptosis, cellular senescence, angiogenesis, and nuclear phosphoinositide signaling. ING proteins contain a few conserved domains, including plant homeodomain motif, nuclear localization signal, and potential chromatin regulatory domain, suggesting that the ING family proteins may share common biological functions. ING3 has been shown to modulate p53-mediated transcription, cell cycle control, and apoptosis, possibly by modulating the NuA4 complex histone acetyltransferase activity. Because ING1b and ING2 have been shown to be involved in cellular stress responses such as nucleotide excision repair and apoptosis after UV irradiation, we investigated whether ING3 also mediated UV-induced apoptosis. We found that ING3 expression was rapidly induced by UV irradiation at both mRNA and protein levels. Using the stable clones of melanoma cells overexpressing ING3, we showed that overexpression of ING3 significantly promoted UV-induced apoptosis. Unlike its homologues ING1b and ING2, ING3-increased apoptosis was independent of functional p53. Furthermore, ING3 did not affect the expression of mitochondrial proteins but increased the cleavage of Bid and caspases-8, -9, and -3. Moreover, ING3-mediated apoptosis was blocked by inhibition of caspase-8 or Fas activation. In addition, ING3 up-regulated Fas expression at both mRNA and protein levels. Knock down of ING3 decreased UV-induced apoptosis remarkably. These data indicate that ING3 plays an important role in cellular response to UV irradiation by enhancing UV-induced apoptosis through the activation of Fas/caspase-8 pathway.

Cutaneous malignant melanoma is a life-threatening skin cancer, and the incidence of melanoma has doubled in the last decade (1,2). Melanoma metastasizes rapidly to other organs, and there is no effective treatment for metastatic melanoma. Even the most established adjuvant immunotherapy using interferon ␣ has encountered the resistance caused by the overexpression and activation of signal transducers and activators of transcription 5 (3)(4)(5). Patients with metastatic melanoma have a poor prognosis, with a 5-year survival of a mere 10% of patients (1,6). Epidemiological studies strongly implicate ultraviolet (UV) exposure in early life as the main environmental factor for the development of malignant melanoma (6,7). UV-induced mutation and/or aberrant expression of certain genes, including cdk6, p16 INK4a , N-ras, and BRAF, have been known to mediate the pathogenesis of melanoma (8 -10).
The novel ING tumor-suppressor family proteins (ING1-5) have been discovered during the past decade and are recognized as regulators of transcription, cell cycle checkpoints, DNA repair, and apoptosis (11). They are also reported to promote cellular senescence, inhibit angiogenesis, and function as nuclear phosphoinositide receptors (11)(12)(13). All these proteins share a highly conserved C-terminal plant homeodomain motif that mediates signal transduction, possibly by regulating chromatin remodeling, histone acetyltransferase/deacetylase activities, and phosphoinositide signaling (11,12). The ING1 proteins were frequently down-regulated but less frequently mutated in human malignancies, including neuroblastomas, colon carcinoma, head and neck squamous cell carcinomas, breast, gastric, esophageal, lymphoid, lung, and brain tumors, whereas they were increased in melanoma, papillary thyroid carcinoma, and ductal breast carcinoma, concomitant with loss of nuclear localization (11, 14 -17). The ING4 protein was also frequently deleted and reduced in human breast cancer, head and neck squamous cell carcinomas, and gliomas (18 -20). Although most studies have focused on ING1, especially its splicing isoform ING1b, the multiple amino acid sequence alignment of human ING proteins revealed a few conserved domains other than plant homeodomain motifs, including nuclear localization signals and potential chromatin regulatory domains (21). These conserved domains suggest that the ING family proteins may share common biological functions. However, the exact functions of these domains remain to be elucidated.
The ING3 gene, mapped to 7q31.3, consists of 12 exons and encodes a 46.8-kDa protein that modulates p53-mediated transcription, cell cycle control, and apoptosis (22). Gunduz et al. (23) showed that the expression of ING3 gene is reduced in human head and neck squamous cell carcinomas because of loss of heterozygosity. As a subunit of the NuA4 2 histone acetyltransferase multisubunit complex, ING3 protein is associated with p53 function and can reconstitute robust nucleosomal histone acetyltransferase activity in vitro by forming a recombinant trimeric complex with Tip60 and EPC1 (24 (26,27), whereas ING2 is induced by the DNA double strand break-inducing agents etoposide and neocarzinos-tatin (25). In response to UV irradiation, both ING1b and ING2 can promote UV-induced apoptosis in a p53-dependent manner (26 -28). They also enhance the p53-mediated repair of UV-damaged DNA (29,30). However, the role of ING3 in cellular response to UV irradiation is not clear. In this study, we found that ING3 was rapidly induced by UV irradiation in melanoma cells. Overexpression of exogenous ING3 in melanoma cells can dramatically promote UV-induced apoptosis. Interestingly, unlike with its homologues ING1b and ING2, ING3-mediated apoptosis was independent of p53 function. We also demonstrated that ING3 induced Fas expression and promoted UV-induced apoptosis through the Fas/caspase-8 pathway.

EXPERIMENTAL PROCEDURES
Cell Culture and UV Irradiation-The MMRU and MMAN cell lines were kind gifts from Dr. R. Byers, Boston University. The Sk-mel-5 cell line was obtained from the Tissue Bank at NCI, National Institutes of Health. The MEWO, Sk-mel-3, and Sk-mel-24 cell lines were kind gifts from Dr. A. P. Albino (Memorial Sloan-Kettering Cancer Center). MMRU, MMAN, Sk-mel-5, and Sk-mel-24 cells harbor the wild-type p53 gene, but there is no detectable p53 protein in MMAN and the p53 protein in Sk-mel-5 is inactivated (31)(32)(33)(34). The MEWO cells carry a mutated p53 gene (35), and the status of p53 in Sk-mel-3 is unknown. All melanoma cell lines were maintained in Dulbecco's modified Eagle's medium (Invitrogen), supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, 100 g/ml of streptomycin, and 25 g/ml of amphotericin B in a 5% CO 2 atmosphere at 37°C. Normal human epithelial melanocytes were purchased from Clonetics (Walkersville, MD) and fed in melanocyte growth medium (Clonetics) at 37°C in a 5% CO 2 atmosphere. The wild-type p53 and p53-null HCT116 colorectal cancer cells were obtained from American Type Culture Collection (ATCC) and Dr. B. Vogelstein, respectively, and maintained in McCoy's 5A medium (Invitrogen) supplemented with 10% fetal bovine serum. For UV irradiation, cells were rinsed with PBS and exposed to ultraviolet B (UVB) (290 -320 nm) as previously described (29).
Plasmids, Transfection, and ING3 Stable Clone Generation-The pcDNA3-ING3 and the dominant negative caspase-8 plasmids were kindly provided by Drs. O. Mamoru, and K. Arul, respectively. All the transfections were performed with Effectene reagent (Qiagen, Mississauga, ON, Canada). For the ING3 stable clone generation, MMRU cells were transfected with pcDNA3 or pcDNA3-ING3 and incubated at 37°C for 48 h followed by a 14-day selection in the culture medium supplemented with 800 g/ml of G418 (Sigma). Single clones were then picked up and maintained in the culture medium containing 100 g/ml of G418.
FACS Analysis-Cells were collected by trypsinization and pelleted by centrifugation at 500 ϫ g for 5 min. Cell pellets were then resuspended in 1 ml of hypotonic fluorochrome buffer (0.1% Triton X-100, 0.1% sodium citrate) containing 25 g/ml of RNase A and 50 g/ml of propidium iodide (PI) (Sigma). After being incubated at 4°C in the dark for 1 h, samples were analyzed by EPICS XL-MCL flow cytometer (Beckman Coulter, Miami, FL) to determine the percentage of subdiploid DNA. Cells in sub-G 1 phase were regarded as apoptotic cells. To quantify the cells in apoptosis from necrosis, cells were fixed and stained in Annexin-V-FLUOS staining kit (Roche Applied Science) according to the manufacturer's protocol. Cells stained by both Annexin-V-FLUOS and PI were regarded as apoptotic.
Hoechst Staining-Cells were rinsed with PBS, fixed in 2% formaldehyde at room temperature for 20 min, and stained with 2.5 g/ml of Hoechst 33342 (Sigma) in PBS for 5 min. The cells were then washed with PBS, mounted on the slide with glycerol-PBS (1:1), and visualized under a fluorescent microscope (Zeiss, Jena, Germany).
Cell Survival Assay-Cells in 24-well plates at 80% confluency were irradiated with UVB. Twenty-four hours after UV irradiation, cell survival was determined with the sulforhodamine B (Sigma) assay as described previously (36). Briefly, after the medium was removed, cells were fixed with 500 l of 10% trichloroacetic acid for 1 h at 4°C. The cells were then washed with tap water, air dried, and stained with 500 l of 0.4% sulforhodamine B (dissolved in 1% acetic acid) for 30 min at room temperature. The cells were destained with 1% acetic acid and air dried. For quantification, the cells were incubated with 500 l of 10 mM Tris (pH 10.5) on a shaker for 20 min to dissolve the bound dye followed by colorimetric determination at 550 nm for 100-l aliquots.
Measurement of Caspase-8 Activity-Cells in 100-mm dishes were treated with or without 600 J/m 2 UVB and cultured for 5 h before being harvested. Caspase-8 activity was determined using ApoAlert caspase colorimetric assay kits (Clontech, Palo Alto, CA). The protease activity was determined by comparing the absorbance of p-nitroaniline (p-NA) in treated cells with that in untreated cells.
Statistical Analysis-The data were presented as the mean Ϯ S.D. Statistical analyses were performed using Student's t-test, and p value Ͻ0.05 was considered significant.

RESULTS
ING3 Is DNA Damage Inducible-To test our hypothesis that ING3 may play an important role in the cellular response to UV irradiation, we analyzed ING3 expression level after UVB irradiation in melanoma cells. We found that the ING3 protein expression was rapidly induced in MMRU melanoma cells as early as 0.5 h after UVB irradiation, reached its peak at 6 h (4.6-fold compared with un-irradiated MMRU cells), and returned to the basal level 24 h later (Fig. 1A). Using RT-PCR, we also found that this induction was at mRNA level ( Fig. 1B). As expected, p53 levels were increased after UV irradiation. However, the p53 accumulation was obvious 2 h following UV irradiation, and its level gradually increased within 24 h (Fig. 1A). To assess whether UV-mediated induction of ING3 is a common event rather than a cell line-specific observation, we further examined ING3 expression after UV irradiation in other melanoma cell lines and HCT116 colon carcinoma cell lines. Despite the differences in basal expression level of ING3 among melanoma cell lines, we detected the induction of ING3 in all other five melanoma cell lines (MEWO, MMAN, Sk-mel-3, Sk-mel-5, and Skmel-24) (Fig. 1C). We also detected the induction of ING3 in both wildtype p53 and p53-null HCT116 cells, indicating that UV-induced ING3 expression is p53 independent (Fig. 1D). To evaluate whether ING3 can be induced by other DNA damage stimulus, we treated MMRU cells with genotoxic chemicals, including cisplatin, etoposide, doxorubicin, and camptothecin. Results showed that ING3 was induced by all the treatments except cisplatin (Fig. 1E).
ING3 Promotes UV-induced Apoptosis-To investigate the role of ING3 in cellular stress response to UV irradiation in melanoma cells, we generated stable clones overexpressing ING3 in MMRU cells. ING3 protein and mRNA levels in the stable clones were analyzed by Western blot or RT-PCR, respectively. The results showed that the ING3 stable clones (F5, F6, and F9) had 3-to 5-fold higher ING3 expression compared with parental MMRU cells (Fig. 2, A and B). We noticed that the cells stably expressing ING3 are slightly bigger and less dendritic compared with the parental MMRU cells (Fig. 2C).
There were increased basal apoptotic cells in ING3 stable clones (Fig.  3, A and B), but it was statistically obvious only when being determined by Annexin-V-FLUOS and PI double staining (Fig. 3C). However, ING3 stable clones were significantly more sensitive to UV-induced apoptosis by FACS analysis or Hoechst staining (Fig. 3, A and B). Compared with parental cells, ING3 stable clones enhanced UV-induced apoptosis by 35-100% after 300 or 600 J/m 2 UVB irradiation. The ING3 stable clone F5, which has the highest expression of ING3 among all the stable clones ( Fig. 2A), had a 2-fold increase in UV-induced apoptosis compared with the parental MMRU cells (p Ͻ0.001). As expected, we observed more cleaved caspase-9, caspase-3, and PARP in ING3 stable clone F5 cells after UV irradiation (Fig. 3D). The results of sulforhodamine B cell survival assay showed that ING3 stable clone F5 has fewer survived cells compared with parental MMRU cells after 300 or 600 J/m 2 UVB irradiation (Fig. 3E).

ING3 Mediates UV-induced Apoptosis
Independently of p53-Because both ING1b and ING2, the homologues of ING3, can promote UV-induced apoptosis in a p53-dependent manner (26,27), we then investigated whether the increased apoptosis by ING3 also required the functional p53 protein. The p53 siRNA was used to knock down p53 expression in ING3 stable clone F5, which harbors the wild-type p53. Western blot analysis showed that the protein levels of p53 and its downstream target Bax were dramatically reduced in both MMRU and F5 cells after siRNA treatment (Fig. 4A, bottom panel). However, knock down of p53 in F5 cells did not cause a reduction of apoptosis after UVB irradiation (Fig. 4A, top panel), suggesting that ING3-mediated apoptosis is p53 independent. Similarly, ING3 overexpression did not have an effect on the expression of p53-targeting mitochondrial proteins Bax, Bcl-2, Bad, and Noxa (Fig. 4B), indicating that ING3 did not directly activate the mitochondrial apoptosis pathway. When HCT116 cells were employed, we found that ING3 promoted both basal and UV-FIGURE 2. ING3 was overexpressed in stable clones. MMRU cells were transfected with pcDNA3-ING3 and incubated at 37°C for 48 h followed by a 14-day selection in the culture medium containing 800 g/ml of G418. Single clones were then picked and maintained in the culture medium containing 100 g/ml of G418 for ING3 expression by Western blot (A) or RT-PCR (B). The fold of ING3 overexpression in the stable clones (below the blot) was quantitated with Quantity One software, comparing with the parental MMRU cells. Glyceraldehyde-3-phosphate dehydrogenase and actin were used as loading control for PCR and Western blot analysis, respectively. C, microscopic images were taken for MMRU and stable clones. The magnification is ϫ200.

FIGURE 1. ING3 was induced by DNA damage.
MMRU cells were irradiated with 600 J/m 2 UVB and harvested at indicated time points for protein or RNA extraction. Western blot (A) and RT-PCR (B) were used to analyze the protein and mRNA levels of ING3, respectively. The expression level was quantitated with Quantity One software; the fold of induction is listed below the blots. Several melanoma cell lines (C) and HCT116 colon cancer cells (D) were irradiated with 600 J/m 2 UVB and harvested for Western blot analysis 6 h after irradiation. E, MMRU cells were treated with 50 M cisplatin, 50 M etoposide, 2 M doxorubicin, and 2 M camptothecin for 6 h before being harvested for Western blot analysis. Glyceraldehyde-3-phosphate dehydrogenase and actin were used as loading control for PCR and Western blot analysis, respectively.

ING3 Mediates UV-induced Apoptosis via Death Receptor Pathway-
Western blot analysis showed that more caspase-8 was cleaved in F5 cells upon UV irradiation, which activated caspase cascades through the cleavage of Bid (Fig. 5A). The colorimetric assay also suggested higher caspase-8 activity in F5 cells after UV irradiation (Fig. 5B). These data suggested that ING3 may mediate UV-induced apoptosis through the death receptor pathway. To confirm that caspase-8 activation was crucial for ING3-mediated apoptosis under UV stress, we transfected the MMRU and F5 cells with a dominant negative caspase-8 expression vector. Results showed that inhibition of caspase-8 completely abolished the ING3-mediated apoptosis (Fig. 5C). Furthermore, when the activation of caspase-8 was blocked, we did not observe a significant difference in caspase-9 activation between F5 and its parental MMRU cells in response to UV irradiation (Fig. 5D), confirming that caspase-9 was activated by caspase-8 via Bid cleavage in F5 cells.
The Fas death receptor is activated by binding to Fas ligands. Fas death receptors then recruit Fas-associated death domain protein to the plasma membrane, which in turn activates pro-caspase-8, leading to the activation of a cascade of caspases (37,38). Studies have shown that Fas death receptor pathway is activated by UV radiation and contributes to UV-induced apoptosis (39,40). To confirm that Fas death receptor pathway is crucial for ING3-mediated apoptosis under UV stress, MMRU and F5 cells were preincubated at 4°C for 30 min, which was shown to block the Fas aggregation (41). Our results showed that there was no significant difference in apoptosis rate between MMRU and F5 cells in response to UV irradiation (Fig. 5E). Meanwhile, when Fas expression was knocked down by Fas siRNA, we observed a significant decrease in the apoptosis of F5 cells after UV irradiation (Fig. 5F).
ING3 Induces Fas Expression-To investigate how ING3 activated Fas death receptor pathway, we examined whether ING3 could modulate Fas expression. Our data indicated that ING3 induced Fas protein expression at non-stress condition and after UVB irradiation (Fig. 6A). The highest Fas induction by ING3 was observed at 8 h after 600 J/m 2 UVB irradiation. Moreover, the induction of Fas by ING3 was at the transcriptional level (Fig. 6B). ING3 did not have effect on Fas ligand expression (Fig. 6A).  (Fig.  7A). FACS analysis showed that UV-induced apoptosis in MMRU cells was significantly reduced after ING3 knock down (Fig. 7B).

DISCUSSION
Like with its homologues ING1b and ING2, ING3 expression was rapidly induced after DNA damage. This induction by UV irradiation occurred much earlier than the p53 protein accumulation in melanoma cells (Fig. 1A), suggesting that this induction was independent of functional p53. The p53 independency of UV-induced ING3 expression was confirmed in melanoma cell lines with different p53 mutational status and in wild-type p53 and p53-deficient HCT116 cells (Fig. 1, C and D). Although UV can induce G 1 or G 2 cell cycle arrest (42,43) and ING3 may be periodically expressed during the cell cycle, this induction is not likely due to the augmented cell cycle population, because the induction of ING3 was rapid and transient and the ING3 expression returned to normal level 24 h after UV irradiation.
DNA damage caused by UV and other genotoxic stress activates checkpoint proteins ATM or ATR, which then modulate the cell fate in cell cycle arrest, DNA repair, or apoptosis through activating the downstream mediators, transducers, and effectors like p53, BRCA1, and Chk1/2 (44,45). Considering that ING3 is a key subunit of the NuA4 histone acetyltransferase complex (24), the inducible ING3 may play an important role in chromatin remodeling in cellular response to UV irradiation. In this study, we for the first time showed that stable overexpression of ING3 can remarkably promote UV-induced apoptosis in melanoma cells. This increased apoptosis was independent of p53 function, although its homologues ING1b and ING2 both enhanced UV-induced apoptosis in p53-dependent manner (28,29). Because it was reported that ING3 can mediate p53-dependent growth arrest and apoptosis in RKO cells (22), it is likely that ING3 can modulate apoptosis in both a p53-dependent and -independent manner, which may rely on the cell type.  (11, 25, and 46). Although ING1b also induced apoptosis in a p53independent manner, it was possibly through the homologues of p53, p63␣ and p73␣ (47). In this study, we did not detect any significant change of p53-responsive proteins, including Bax, Bcl-2, and Noxa, after overexpressing ING3, which further supports the notion that ING3mediated apoptosis after UV irradiation is p53 independent. This discrepancy of ING3 function from other family members is probably because ING3 is evolutionarily distinct from other family members (13). Both ING1 and ING2 contain a small C-terminal conserved proteininteracting motif that binds a defined subset of peptides and, together with plant homeodomain, binds to phosphatidylinositol monophosphates (13, 48, and 49). ING1b also contains a unique N-terminal PCNA-interacting protein (PIP) domain, through which ING1b promoted UV-induced apoptosis (27). Moreover, there are two distinct insertions of 102 and 54 amino acids located between the potential chromatin regulatory and nuclear localization signals regions of ING3, respectively (13). Therefore, it will be interesting to clarify whether the structural differences between ING3 and ING1b or ING2 are responsible for the different pathways in UV-induced apoptosis.
UV-induced apoptosis can be triggered through both p53-mediated  mitochondrial pathway and death receptor pathways (50). In this study, both Western blot analysis and caspase-8 activity assay indicated that caspase-8 was more activated in ING3 stable clone F5 cells than in parental MMRU cells in response to UV irradiation (Fig. 5, A and B). Moreover, inhibition of caspase-8 activation blocked the cleavage of caspase-9 (Fig. 5D), confirming that ING3 does not directly affect the mitochondrial apoptosis pathway but operates through the cross-talk linked by cleaved Bid. The ING3-mediated apoptosis was abrogated when cells were transfected with dominant negative caspase-8 (Fig. 5C) or preincubated at 4°C (Fig. 5E), supporting that death receptor pathways mediate ING3-triggered apoptosis. Because Fas siRNA treatment significantly reduced ING3-mediated apoptosis (Fig. 5F) and ING3 sta-ble clone F5 had higher Fas mRNA and protein levels in both basal and UV-irradiated conditions (Fig. 6, A and B   receptor pathways may also be involved in ING3-mediated apoptosis in response to UV irradiation. Like its homologue ING2 (28), ING3 up-regulated Fas expression, leading to the activation of the death receptor pathway. Fas protein is a death receptor on the cell membrane, whereas ING3 protein is mainly located in the nucleus. Similar to transcriptional factors, including NF-B and Egr-1 (52), the rapid induction and recovery of ING3 following UV irradiation suggested that ING3 may function as a transcription factor in response to DNA damage. Although ING3 does not modulate p53 function, it is likely that ING3 can act as a transcription factor or cofactor to mediate apoptosis by modulating chromatin remodeling through the NuA4 histone acetyltransferase multisubunit complex (24), which may help explain why loss of functional Tip60, the core histone acetyltransferase of NuA4 complex, failed to signal the existence of DNA damage to the apoptotic machinery (53). The exact molecular mechanism of ING3-mediated chromatin-related transcriptional regulation remains to be revealed. In summary, we have demonstrated that ING3 regulates UV-induced apoptosis through modulating Fas expression, thereby resulting in the activation of Fas/caspase8 pathway and the activation of caspase cascades via Bid cleavage.