ARMS Depletion Facilitates UV Irradiation–Induced Apoptotic Cell Death in Melanoma

Tumorigenesis is currently viewed as a consequence of abnormal developmental program, often resulting from aberrant expression of molecules involved in embryonic development. Malignant melanoma, which is associated with high morbidity and mortality, is a tumor resulting from transformed melanocytes or nevocytes that are derived from neural crest. Consequently, neuron-specific markers are frequently expressed in neoplastic melanocytes to highlight its ontology as neuroepithelium. Examples include intermediate filament protein periperin, neuron-specific enolase, neuropeptide substance P, muscarinic acetylcholine receptors, and p75^NTR (1–5).

Ankyrin repeat-rich membrane spanning (ARMS), also known as Kidins220 (kinase D–interacting substrate of 220 kDa), is a tetra-spanning transmembrane protein abundantly expressed in the developing and adult neural tissues (6, 7). ARMS functions as a downstream molecule in the neurotrophin/Trk and ephrin/Eph pathways and plays a pivotal role in prolonged mitogen-activated protein kinase (MAPK) signaling associated with cell differentiation in primary neurons and PC12 cell lines (6, 8, 9). Interestingly, both neurotrophin- and ephrin-mediated signaling pathways have been shown to be involved in melanoma tumorigenesis and progression (10–13). We thus wonder whether ARMS also participates in the carcinogenesis of malignant melanoma.

In this report, we investigated the expression of ARMS in human melanoma cell lines and in archival human surgical specimens with melanocytic lesions. Our results showed that ARMS was significantly overexpressed in most of the melanoma cell lines and human melanoma. To investigate the role of ARMS in the pathogenesis of melanoma, we exploited siRNA to generate stable ARMS-knockdown/B16F0 cells, which were used for both in vitro and in vivo tumorigenesis assays. Our results show that depletion of ARMS leads to inhibition of anchorage-independent growth, restrictive xenograft growth in severe combined immunodeficient (SCID) mice, and increased apoptosis after UVB irradiation via up-regulation of the MAPK/extracellular signal–regulated kinase (ERK) kinase (MEK)/ERK signaling activity. Importantly, most of the cases we studied are acral lentiginous melanoma, which did not harbor active BRAF mutation but instead overexpressed the BRAF upstream molecule ARMS, suggesting that overexpression of ARMS could serve as a mechanism to promote melanoma formation via its effect on inhibiting apoptotic cell death and sustaining MEK signaling.

Antibodies. We generated rabbit polyclonal anti-ARMS antibody by immunizing rabbits with a synthetic peptide that corresponds to 1,696–1,714 amino acids of mouse ARMS (Sigma Genosys). The polyclonal ARMS antibodies were purified by affinity chromatography (CNBr-activated Sepharose 4B beads, Amersham). Antibodies used in immunoblotting include anti-ARMS, anti–poly(ADP-ribose) polymerase (PARP; Cell Signaling Technology), anti–phospho-ERK1/2, anti-ERK1/2 (Promega), anti–phospho-p38, anti-p38, anti–phospho-c-jun NH₂-terminal kinase (JNK), anti-JNK (Cell Signaling Technology), anti–phospho-Akt (R&D Systems), anti-Akt (Promega), anti-Myc (Invitrogen Corp.), and anti-hemagglutinin (Sigma) antibodies. The secondary antibody used in Western blot was goat anti-rabbit IgG (Amersham) or goat anti-mouse IgG (Pierce) conjugated to horseradish peroxidase.

Cell culture. Normal human melanocytes were obtained from foreskin of infants after elective circumcision. Foreskins were cut into small pieces and incubated with 0.2% protease (Sigma) in SMEM medium (Invitrogen) overnight at 4°C. The epidermis was separated from the dermis with forceps, incubated with 0.25% trypsin for 5 min at 37°C, and cell suspension was obtained and cultured with the medium consisting of Ham's F12 supplemented with 50 μg/mL gentamicin, 20 ng/mL recombinant β-FGF (PeproTech, Inc.), 20 μg/mL 3-isobutyl-1-methylxanthine, 10 μg/mL cholera toxin (Sigma), and 20% fetal bovine serum (FBS).

Human melanoma cell lines SK-Mel-1, SK-Mel-2, SK-Mel-5, SK-Mel-28, and RPMI 7951 were purchased from the American Type Culture Collection (ATCC) and cultured in essential minimal Eagle's medium (EMEM) supplemented with 10% FBS, nonessential amino acids, and 1 mmol/L sodium pyruvate. Mouse B16F0 melanoma cell line was from the ATCC and maintained in DMEM supplemented with 10% FBS, 4 mmol/L l-glutamine, 100 μg/mL streptomycin, and 100 units/mL penicillin.

RNA interference and selection of stably transfected clones. The siRNA expression vector pSUPER.neo+GFP (OligoEngine) was used for the construction of ARMS-siRNA targeting vector. The specific sequence for paired RNA oligonucleotides contained nucleotides 4,271–4,289 for ARMS-RNAi-1 and nucleotides 4,840–4,858 for ARMS-siRNA-2 in mouse ARMS (accession no. AK122478). The oligonucleotides were annealed and ligated into BglII and HindIII sites of the pSUPER plasmid containing a H1-RNA promoter. B16F0 cells were transfected with ARMS-RNAi plasmid or pSUPER plasmid by LipofectAMINE 2000 (Invitrogen) and stable cell clones were selected with 700 μg/mL G418 (geneticin; Sigma) starting at 24 h after transfection. After 2 weeks of selection in G418, clones of resistant cells were isolated and allowed to proliferate in medium containing 500 μg/mL G418. Knockdown of ARMS was confirmed by reverse transcription-PCR (RT-PCR) and Western blot analysis. The RNA interference (RNAi) rescue experiment was done by transfecting the ARMS-RNAi-2/B16F0 stable clone with the ARMS-pcDNA3.1C-Myc 4849T→C expression mutant, which was constructed with the site-directed mutagenesis kit (Stratagene) according to the manufacturer's instruction.

RNA extraction and reverse transcription-PCR. Total RNA was extracted with the TRIzol reagent (Invitrogen) and reverse transcribed with the Superscript amplification kit (Invitrogen). The human ARMS primers used were 5′-CAGATCAGTCAGGCAGTAAG-3′ (sense) and 5′-GAGTTCGGTTCAGGTTGTAG-3′ (reverse) to yield a 590-bp amplification product. The amplification program consisted of initial denaturation at 94°C for 5 min, followed by 25 cycles at 94°C for 1 min, 58°C for 1 min, 72°C for 1 min, and final polymerization at 72°C for 7 min. Samples were then electrophoresed on a 1.2% agarose gel, stained with ethidium bromide, and photographed under UV.

Genomic DNA extraction and mutation detection for BRAF. Genomic DNA was extracted from paraffin-embedded tumor tissue dissected by laser microdissection. The tissue was incubated with 10 μg/mL proteinase K in a 150-μL aliquot of ATL buffer (pH 8.3) overnight at 56°C. After that, 1-mL TRIzol and 200-μL chloroform were added, incubated, and centrifuged. The supernatant was supplemented with glycogen and 100% ethanol and stored at −20°C for 24 h to precipitate the genomic DNA, which was recovered and used for subsequent PCR amplification. PCR primers for BRAF exon 15 were 5′-TCATAATGCTTGCTCTGATAGGA-3′ and 5′-GGCCAAAAATTTAATCAGTGGA-3′ to detect the hotspot of T→A missense mutation at nucleotide 1,799. The 240-bp PCR products were gel eluted and analyzed by direct automated sequencing (ABI 377 DNA Sequencer, Applied Biosystems).

Immunohistochemistry, tissue arrays, and terminal deoxyribonucleotidyl transferase–mediated dUTP nick end labeling assay. Paraffin-embedded tissues were obtained from the Department of Pathology, National Taiwan University Hospital, with the approved written informed consent by each patient, who underwent surgery at the National Taiwan University Hospital between 1997 and 2005.

Five-micrometer-thick, paraffin-embedded sections were deparaffinized and rehydrated. A microwave-based antigen retrieval method with 10 mmol/L citrate buffer (pH 6.0) for 10 min was done. The sections were washed with PBS for 5 min and immersed in 0.3% hydrogen peroxide in PBS for 15 min to block endogenous peroxidase activity. After blocking with 10% nonimmune goat serum, sections were incubated with anti-ARMS polyclonal antibody overnight at 4°C. The sections were washed and then incubated with biotin-labeled antirabbit antibody (Vector Laboratories) for 30 min at room temperature. After three washes with PBS, sections were incubated with a solution of avidin and biotin-conjugated peroxidase complex (Vector Laboratories) for 30 min at room temperature. 3-Amino-9-ethylcarbazole (DAKO) was applied for colorization and slides were counterstained for nuclei with hematoxylin. The sections were examined double blindly by three pathologists.

Tissue microarrays were constructed from archival formalin-fixed, paraffin-embedded cancer specimens. Sampling of tumor regions was based on visual alignment with the corresponding H&E-stained sections. Tissue cores were placed into a recipient paraffin block with a tissue microarrayer (Manual Tissue Arrayer 1, Beecher Instruments, Inc.) and 5-μm sections of the resulting microarray block were used for immunohistochemistry analysis.

Terminal deoxyribonucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining of tissue samples was done using the In Situ Cell Death Detection Kit (Roche) according to the manufacturer's instruction.

Western blot analysis. Protein was extracted from the cultured cells lysed with the sampling buffer [150 mmol/L NaCl, 50 mmol/L Tris-HCl (pH 7.4), 1% NP40, 0.25% Na-deoxycholate, 1 mmol/L EDTA, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L NaF, 1 mmol/L Na₃VO₄ and complete protease inhibitor]. Thirty micrograms of total proteins were loaded per lane and subjected to SDS-PAGE and then transferred onto polyvinylidene difluoride membrane (Millipore). After blocking in 5% nonfat milk in TBS, the membrane was subjected to immunoblotting with antibodies. An enhanced chemiluminescence reaction (Amersham) was applied for signal detection. Immunoblotting with antihuman α-tubulin antibody (Oncogene) was used as an internal loading control.

In vitro cell proliferation and soft agar clonogenic assay. Cells were seeded (3 × 10⁴ per well) in six-well plates in 2-mL growth medium on day 0. Cells were cultured, trypsinized, and the numbers were counted with a hemacytometer at 24-h interval till day 5. Triplicate experiments were done for each cell line.

A clonogenic assay was done in six-well 35-mm-diameter tissue culture dishes containing a bottom layer of 0.7% (w/v) agarose in complete culture medium and a top layer of 0.3% agarose solution mixed with 2.5 × 10³ cells. The layers were then covered with 2-mL culture medium. After 2 weeks of incubation, colonies >0.5 mm in diameter were scored. The experiments were done in triplicate.

Tumorigenicity assay in SCID mice. Animal handling followed the approved procedures issued by the Animal Studies Committee of National Taiwan University. Cells were trypsinized, washed with PBS, and adjusted to a concentration of 4 × 10⁵ in 100-μL PBS. The cell suspension was then mixed with an equal volume of Matrigel (Collaborative Biomedical Products) before s.c. injection into nonobese diabetic SCID male mice (5–6 weeks old). After 10 days, the mice were killed and the tumors were removed, sized, and weighed. Tumor tissue was either frozen sectioned or formalin fixed for paraffin sections, which were subjected to H&E staining, TUNEL assay, and immunohistochemistry for anti-HMB45 (DAKO), anti–Ki-67 (Abcam), and anti-ARMS antibodies.

Flow cytometry. Adherent and floating cells, treated or not treated for apoptosis induction, were collected, washed with ice-cold PBS, and fixed in 70% ethanol at 4°C for at least 1 h. After fixation, the cells were washed twice with PBS and then incubated with 1 mg/mL RNase A (Sigma) and 50 μg/mL propidium iodide (Sigma) in PBS at room temperature for 30 min. The cells were then analyzed by flow cytometry using CELLQuest software (Becton Dickinson). Besides, chemicals used in this assay were applied as follows: cells were pretreated with 10 μmol/L PD98059 (Calbiochem) or 50 μmol/L bocaspartyl-(OMe)-fluoromethyl-ketone (BAF; Calbiochem) for 1 h before UVB irradiation. Phosphatidylserine externalization in cells was assessed by staining with Annexin V-AlexaFluor 594 (Molecular Probes, Invitrogen). For cell cycle progression, we analyzed the DNA content of the cells with propidium iodide.

Mitochondrial membrane potential was measured with tetramethylrhodamine ethyl ester (TMRE; Molecular Probes, Invitrogen) using flow cytometry. Cells were collected with trypsinization, washed twice with PBS, and stained with 150 nmol/L TMRE in PBS in the dark at 37°C for 30 min. Samples were washed and resuspended with PBS and then analyzed with FACScan using CellQuest software.

Statistical analysis. The correlation between ARMS expression and the melanocytic lesions in immunohistochemical study was evaluated by Fisher's exact test. Other data are expressed as means ± SD, and the significance was assessed by Student's t test. Differences with P < 0.05 were considered significant.

Expression of ARMS in human melanocytic lesions and melanoma cell lines. We have generated polyclonal antibodies against a peptide corresponding to amino acid residues 1,697–1,714 of mouse ARMS. The generated antibody recognized a band of ∼220 kDa, which comigrated with Myc-tagged ARMS proteins overexpressed in PC12 cells. However, immunoblotting with preimmune serum did not reveal this specific band (Fig. 1A). ARMS immunostain was thus applied in sets of tissue arrays covering a variety of human tumors to investigate whether ARMS was abnormally expressed in any specific tumors. In our screening, we found that ARMS was specifically expressed in tumors of neuroepithelium origin, including pheochromocytoma, paraganglioma, central neurocytoma (data not shown), and malignant melanoma.

ARMS expression in cutaneous melanoma was further investigated by immunohistochemical study of ARMS in 3 cases of normal skin and 79 surgical specimens of different melanocytic lesions, which include benign melanocytic lesions (n = 15), dysplastic nevus (n = 2), primary in situ or invasive melanoma (n = 42), and metastatic melanoma (n = 20). The intensity of ARMS immunostain was examined under a light microscope and scored as negative, weakly positive, moderately positive, or strongly positive. When we compared the intensity of ARMS staining between benign nevi and melanomas, positive ARMS immunoreactivity was observed in most of the cases of primary malignant melanoma (92.9%) and metastatic melanoma (60.0%), whereas weak or absent ARMS immunoreactivity was seen in normal skin and benign nevocellular nevi (Table 1; Fig. 1B and C). Patterns of cytoplasmic punctate, diffuse cytosol, or perimembranous staining could be observed in melanoma cells. Statistical analysis with Fisher's exact test showed significantly increased ARMS expression in primary cutaneous and metastatic melanoma relative to that in benign nevi (Table 1). The specific expression of ARMS in malignant transformed melanocytes was further highlighted by negative immunostain of ARMS in nonmelanocytic skin cancers including basal cell carcinoma, squamous cell carcinoma, and extramammary Paget's disease (data not shown).

ARMS expression was also examined in normal human melanocytes and a panel of human melanoma cell lines including SK-Mel-1, SK-Mel-2, SK-Mel-5, SK-Mel-28, and RPMI 7951. Semiquantitative RT-PCR analysis showed that all tested melanoma cell lines expressed increased level of ARMS compared with that in normal melanocytes (Fig. 1D,, left). Western blotting analysis also revealed that all melanoma cell lines, except RPMI 7951, overexpressed ARMS protein compared with normal melanocytes (Fig. 1D , right). The results indicated that increased ARMS expression, both at the transcriptional and translational levels, occurred in most malignant melanoma cell lines. Taken together, the results suggest that ARMS, as a neuron-predominant molecule, is overexpressed specifically in malignant transformed melanocytes and is thus can be used to differentiate melanoma from nonmelanoma skin cancers.

Acral lentiginous melanoma overexpresses ARMS without activating BRAF^T1799A mutation. A hallmark for melanoma formation arising in intermittent sun-exposed sites is a high frequency of activating BRAF^T1799A mutation, leading to constitutive activation of the MEK/ERK pathway and cell survival. However, melanomas arising from skin with minimal sun exposure such as palms and soles typically do not acquire BRAF mutation (14). Because ARMS is known as an upstream activator of BRAF in the prolonged MAPK pathway (15), we thus wonder whether BRAF is already constitutively activated in these melanocytic lesions we studied. Representative paraffin-embedded samples of 19 melanomas (including 12 acral lentiginous melanomas and 7 superficial spreading melanomas) with overexpressed ARMS were microdissected by laser capture. Genomic DNA was extracted from the dissected tumor cells. Primer sets flanking exon 15 of BRAF were used for PCR amplification and the resulting PCR product was sequence analyzed. A missense 1799T to A transversion was identified in 4 of 7 (57.1%) cases with superficial spreading melanoma, but none of the 12 cases of acral lentiginous melanoma showed such mutation (Fig. 2A,, top). We also found that SK-Mel-1, SK-Mel-5, and SK-Mel-28 cells do harbor BRAF^T1799A as previously reported (ref. 16; Fig. 2A , bottom). Notably, most of the melanoma cases we studied were acral lentiginous melanoma (33 of 42; 78.6%) that was not exposed to sun. Thus, the wild-type status of BRAF in our cases of acral lentiginous melanoma is consistent with another report (17). In addition, it further suggests that, without BRAF^T1799A mutation, overexpressing ARMS in melanoma per se could contribute to melanoma formation, especially for the tumorigenesis of acral lentiginous melanoma.

Effect of ARMS depletion on cell proliferation, cell cycle progression, and cell transformation. Overexpressing ARMS in dysplastic melanocytes and malignant melanoma prompted us to explore whether ARMS plays a role in tumorigenesis of cutaneous melanoma. Mouse B16F0 melanoma cell line, which overexpresses ARMS and harbors no activating mutation of BRAF (18), was used to generate stable cell lines with depleted ARMS by RNAi. Two targeting constructs, pSUPER-ARMS-RNAi-1 (designed for targeting toward ARMS nucleotides 4,271–4,289; accession no. AK122478) and pSUPER-ARMS-RNAi-2 (for targeting toward ARMS nucleotides 4,840–4,858), were exploited. More than 20 viable clones were picked and expanded and were checked for ARMS expression by RT-PCR and Western blotting. Two stable clones, one labeled as ARMS-RNAi-1 (from targeting construct 1, clone 13) and the other labeled as ARMS-RNAi-2 (from targeting construct 2, clone 15), were used in the subsequent experiments. Both clones have reduced ARMS expression to the level ∼10% of that in control-transfected cells (Fig. 2B).

The growth property of the ARMS-RNAi cells was first assessed by trypan blue exclusive assay. Figure 2C showed that the proliferation rate was not significantly changed in ARMS-RNAi cells compared with the control-transfected or untransfected B16F0 cells. Flow cytometry analysis revealed that the progression of cell cycle was unaffected by ARMS silencing in B16F0 cells (distribution ratio for G₁/S/G₂M: control, 54.31%:11.29%:21.98%; ARMS-RNAi-1, 53.57%:12.75%:21.88%; ARMS-RNAi-2, 53.65%:11.54%:25.70%). In addition, the apoptotic death assessed by sub-G₁ ratio in flow cytometry without any treatment of cells is not significantly changed in ARMS-RNAi cells compared with the control-transfected or untransfected B16F0 cells (sub-G1: control, 4.13 ± 2.01%; ARMS-RNAi-1, 3.95 ± 3.42%; ARMS-RNAi-2, 2.08 ± 2.23%).

Transformation ability was assessed by anchorage-independent growth in soft agar. A total of 2.5 × 10³ cells for each were seeded into soft agar plate and numbers of colony size >0.5 mm were counted after 14 days of incubation. Our results showed that control-transfected cells had 78.6 ± 11.7 colonies per 35-mm dish, whereas ARMS-RNAi-1 and ARMS-RNAi-2 cells had 37.0 ± 12.0 and 42.0 ± 4.3 colonies per 35-mm dish, respectively (P < 0.01; Fig. 2D). We concluded that without much influence on the basal status about cell cycle progression, apoptosis, and proliferation, ARMS-RNAi cells were somehow compromised by the disability of anchorage-independent growth in vitro.

Restricted tumor growth of ARMS-depleted melanoma cells in vivo. To investigate whether ARMS expression affects melanoma tumor formation in vivo, we injected 4 × 10⁵ ARMS-RNAi cells s.c. into one side of the flank of SCID mice with simultaneous inoculation of the control-transfected B16F0 cells into the other side of the same mouse for comparison. Ten days after inoculation, mice were sacrificed and examined (n = 9). Tumors were found only at the inoculation site, and there was no evidence for distant metastasis as examined both grossly and microscopically. When tumors at the inoculation site were examined, the tumor from ARMS-RNAi cells was smaller and lighter than that from the control [mean tumor weight, 0.12 ± 0.05 g in ARMS-RNAi cells versus 0.32 ± 0.07 g in control (P = 0.005); mean tumor volume, 74.9 ± 17.5 mm³ in ARMS-RNAi cells versus 292.3 ± 21.03 mm³ in control (P = 0.005, Student's t test); Fig. 3A and B].

We confirmed that tumors from ARMS-RNAi cells showed no detectable ARMS immunoreactivity, whereas tumors from the control B16F0 cells showed positive ARMS staining (Fig. 3C). All the tumors showed positive HMB45 immunostaining, confirming their melanocyte origin (Fig. 3C), which was also evident by the presence of melanosomes in tumor cells examined by electron microscopy (data not shown). Interestingly, broader area of apoptotic cell death was found in tumors derived from ARMS-RNAi cells as already evident by H&E staining. TUNEL study further confirmed the presence of larger area of apoptotic cells in tumors of ARMS-RNAi cells. By contrast, tumors from the control showed only scattered small areas of apoptosis (Fig. 3C, and D, bottom). When cellular proliferative status was checked by Ki-67 immunostain, we found no difference between tumors of ARMS-RNAi and those of the control (Fig. 3C  and D, top). Taken altogether, the results suggest that the restricted growth of melanoma cells caused by ARMS depletion may result from enhanced cell death.

Knockdown of ARMS enhances UVB-induced apoptosis in melanoma cells. The discrepancy of cell death caused by ARMS-RNAi observed in in vitro and in vivo assays merits one possibility that ARMS itself might function in the regulation of cell survival or death only when cells are challenged with certain stress. Otherwise, it contributes little to the basal state of cell survival or death in vitro. To test the hypothesis, we challenged cells with various cellular stress and apoptotic stimuli, including UVB irradiation (25 mJ/cm²), H₂O₂ (30 μmol/L), etoposide (25 μmol/L), or staurosporine (1 μmol/L) treatment. As shown in Fig. 4A, only oxidative stress (H₂O₂ treatment) and UVB irradiation triggered significant differential response of cell death with regard to the expression level of ARMS. Depletion of ARMS obviously caused melanoma cells to be more apt to apoptotic death elicited by UVB treatment, as evident by significantly increased percentage of sub-G₁ population in ARMS-RNAi cells (20.37 ± 5.02%) relative to the control (10.34 ± 4.46; P < 0.01; Fig. 4A,, left). Flow cytometry analysis of cells with Annexin V immunostaining also confirmed the increased apoptosis caused by ARMS-RNAi in melanoma cells treated with UVB irradiation (26.7 ± 1.7% versus 14.4 ± 1.7% in control; P < 0.01; Fig. 4A,, right). Reintroduction of RNAi-resistant full-length ARMS into ARMS-RNAi melanoma cells significantly rescued cells from ARMS depletion–induced apoptosis under UVB treatment (Fig. 4B). Because the apoptosis rate was rescued by expressing ARMS, this effect is specifically due to loss of ARMS and not the off-target silencing effect of RNAi.

To investigate whether or not the caspase-dependent mitochondrial pathway is involved in ARMS-mediated apoptotic death elicited by UVB irradiation in melanoma cells, we pretreated cells with the pan-caspase inhibitor BAF before UVB exposure. As shown in Fig. 4C, BAF treatment reduced UVB-induced apoptosis in both control- and ARMS-RNAi–transfected B16F0 cells, suggesting that the conventional caspase-dependent apoptotic pathway underlies UVB-induced and ARMS-mediated apoptosis. In line with this, when caspase-3 activity was measured by the effective cleavage of its substrate, poly(ADP-ribose) polymerase (PARP), increased PARP cleavage was found in ARMS-RNAi cells compared with the control cells (Fig. 4D). An early event in the mitochondria-mediated apoptotic cascade is the opening of mitochondrial membrane permeability transition pores and loss of Δψ_m in the inner membrane (19). We thus examined the integrity of mitochondrial membrane by measuring the uptake of TMRE, a dye used to monitor the inner transmembrane potential of mitochondria (Δψ_m; ref. 20). When Δψ_m decreased, TMRE uptake by cells was reduced. As clearly shown in Fig. 4E, UVB treatment induced dramatic loss of membrane potential in ARMS-RNAi cells (from 5.55% to 77.18%), whereas control-transfected cells showed less degree of loss of membrane potential (from 9.79% to 31.61%). Altogether these results suggest that when melanoma cells are challenged with UVB, the expression level of ARMS might modulate the cellular response toward apoptosis through the mitochondrial apoptosis pathway.

Depletion of ARMS diminishes UVB-induced activation of MEK/ERK and sensitizes UVB-induced apoptosis in melanoma cells. UV irradiation of skin-derived cell lines (including keratinocytes and melanocytes) could activate MAPKs and Akt, which are key regulators for cell survival (21–24). We thus examined the temporal activation status of MAPK family members (including ERK, p38, and JNK) and Akt at sequential time points after treatment with 25 mJ/cm² UVB in cells. Because B16F0 cells do not have constitutively activating T1799A mutation in BRAF and thus no increased phospho-ERK (18), we could use this cell line to examine the effect of ARMS on the MAPK signaling pathway without confounding effect of BRAF activity. As clearly shown in Fig. 5A, the expression of phospho-ERK1/2 in the control cells peaked at 30 min after UVB treatment, reaching an 8-fold increase relative to its basal level at 0 min, and gradually declined to the basal level at 2 h after UVB treatment (Fig. 5A,, top left). In contrast to the control, the phospho-ERK1/2 in ARMS-RNAi cells also peaked at 30 min, but its intensity was much diminished (only 2-fold increase relative to the basal level at 0 min; Fig. 5A,, top right). Such difference in phospho-ERK was not due to differential synthesis of ERK proteins because the total ERK protein levels remained constant in the time course examined (Fig. 5A). On the other hand, the activation of p38 and JNK induced by UVB exposure was not changed between cells with or without ARMS down-regulation (Fig. 5A,, central, and data not shown). We have also noticed that constitutive activation of Akt occurred in both cell lines and that the amount of phospho-Akt was slightly increased after UVB treatment in both cell lines (Fig. 5A , bottom).

To investigate whether MEK/ERK activation is involved in modulating the sensitivity of UVB-induced apoptosis in melanoma cells, we pretreated cells with PD98059, a specific inhibitor of MEK, just 1 h before UVB treatment. Addition of PD98059 indeed efficiently reduced UVB-induced ERK1/2 phosphorylation in parental B16F0 cells (Fig. 5B). When apoptosis was reevaluated between ARMS-RNAi cells and the control in the presence of PD98059, the extent of apoptosis triggered by UVB irradiation was augmented both in the ARMS-RNAi cells and the control-transfected cells. However, the augmented apoptosis was much more drastic for ARMS-RNAi cells than for control-transfected cells (44.22 ± 5.53% versus 22.7 ± 0.61%; Fig. 5C). This result suggests that activated MEK/ERK signaling could suppress UVB- and/or ARMS-mediated apoptosis in melanoma cells. Consequently and alternatively, we overexpressed constitutively active MEK1 in ARMS-RNAi cells to see whether it would rescue cells from UVB-induced apoptosis. Indeed, overexpression of constitutively active MEK1 in ARMS-RNAi cells significantly activated ERK phosphorylation (Fig. 5D) and concordantly rescued ARMS-RNAi cells from enhanced apoptosis triggered by UVB treatment (Fig. 5E). Combined together, it suggests a protective role of MEK/ERK activation in suppressing apoptosis of melanoma cells induced by UVB exposure and that ARMS might mediate apoptotic death of melanoma cells through regulation of the activity of ERK signaling.

Overexpression of growth factor receptors or its downstream adaptor proteins to activate survival signal pathway is suggested to be one of the mechanisms for cell transformation (25). Melanoma is well perceived as a tumor caused by malignant transformation of melanocytes, in which a functional autocrine or paracrine loop of growth factors and their receptors could work together during tumorigenesis (10, 25). ARMS, a neuron-predominant molecule, is a key transmembrane adaptor associated with two major groups of tyrosine kinases, Trk and Eph receptors, and it serves as a unique platform in sustained MAP kinase signaling induced by neurotrophin in neurons (15, 26). Because both Trk and Eph pathways are involved in tumorigenesis and tumor progression in melanoma, ARMS seems a logical candidate to be explored about its role in the carcinogenesis of melanoma. In this study, we showed that overexpressed ARMS per se could contribute to melanoma formation. Melanoma cells, whether derived from cell lines or human surgical specimens, overexpress ARMS. Specifically, ARMS is restrictively overexpressed in melanoma, and those nonmelanoma skin cancers thus examined, including basal cell carcinoma, squamous cell carcinoma, and extramammary Paget's disease, do not express ARMS. In this aspect, ARMS could be applied as a special marker to differentiate melanoma from nonmelanoma skin cancers. By studying the ARMS-RNAi/B16F0 stable cell lines we have generated via RNAi, we further showed that depletion of ARMS inhibited anchorage-independent growth in a soft agar colony formation assay and suppressed melanoma growth in SCID mice. These results imply a tumor-promoting role of ARMS in cutaneous malignant melanoma.

Advantageous growth in tumor could be either due to increased proliferation or decreased apoptosis of tumors. We have compared the proliferative capability of tumor cells with or without ARMS, showing that there is no difference in cell cycle progression and mitosis rate. The mechanism underlying ARMS-facilitated melanoma growth is further dissected to be mediated through regulation of the caspase-dependent, mitochondrial apoptosis pathway. When melanoma cells were challenged with UVB irradiation or H₂O₂ oxidative stress, RNAi-mediated ARMS silencing enhanced stress-induced apoptotic death in melanoma cells. Addition of pan-caspase inhibitor BAF could block UVB-induced apoptosis in melanoma cells with ARMS silencing. Alternatively, reintroduction of ARMS into ARMS-RNAi cells rescued cells from UVB-induced, ARMS depletion–augmented apoptosis, suggesting that ARMS itself specifically participates in the regulation of apoptosis via a reversible way. Our data strongly imply that overexpressed ARMS may prevent transformed melanocytes from stress-induced apoptosis tunneled through the conventional caspase-dependent mitochondrial pathway, thus facilitating advantageous growth of malignant melanoma.

In this study, we further provide evidence that ARMS modulates stress-induced apoptosis of melanoma cells via regulation of the MEK/ERK signaling pathway. We show that UVB-induced ERK phosphorylation in melanoma cells is dependent on the presence of ARMS because depletion of ARMS inhibits ERK phosphorylation. Treatment with MEK inhibitor PD98059 before UVB irradiation further compromised the survival of melanoma cells with ARMS silencing. However, reintroduction of constitutively active MEK or RNAi-resistant ARMS could rescue melanoma cells from apoptosis conferred by ARMS depletion. Such finding is actually not so unexpected. It is well documented that activation of the MEK/ERK pathway is critical for cell survival in melanomas whether it is at the early, intermediate, or late stage (27). Dysregulated signaling in the MEK/ERK pathway confers resistance to melanoma cells against a variety of chemotherapeutic drugs (28, 29). Recently, it is reported that ARMS contributes to sustained activation of the BRAF/MEK/ERK pathway whereas it is tyrosine phosphorylated after neurotrophin stimulation in primary neurons and PC12 cells (15). In line with these observations, it is plausible that ARMS may mediate apoptosis of melanoma cells through regulation of the MEK/ERK signaling pathway.

The MEK/ERK pathway plays an essential role in apoptosis in many kinds of tumors in addition to melanoma. For example, ERK activation has been shown to prevent oxidant-induced apoptosis in primary alveolar epithelial cells (30). Several mechanisms have been proposed to explain tumor formation mediated by activated ERK pathway. One prevailing hypothesis describes the regulation of the caspase-dependent mitochondrial apoptotic pathway via the ERK signaling pathway. Suppression of ERK could activate procaspase-3, whereas activated ERK inhibits mitochondrial depolarization to suppress tumor necrosis factor–related apoptosis-inducing ligand–induced T-cell apoptosis (31). In agreement with this hypothesis, we reveal that overexpressing ARMS in melanoma cells suppresses ERK phosphorylation, activates procaspase-3, and leads to mitochondrial depolarization, which in turn results in apoptosis. Other possible mechanisms proposed for ERK in protecting tumor cells from apoptosis include activation of antiapoptotic Bcl-2 proteins, inhibition of proapoptotic factors such as Bim_EL or Bad, or prevention of cytosolic Bax from translocation to mitochondria (32–35). Our microarray analysis between ARMS-RNAi cells and control-transfected cells reveals no differential expression of transcripts of Bcl-2 superfamily or other apoptosis-related molecules (data not shown). How ARMS-mediated MEK/ERK signaling links to the caspase-dependent mitochondrial apoptosis pathway remains to be determined.

Distinct genetic or signaling pathways are involved in different subtypes of cutaneous melanoma. Previous studies have shown that constitutive activation of MAPK pathway resulting from activating BRAF or RAS mutation is important in melanoma arising from skin with intermittent sun exposure, which accounts for 60% to 80% of all cases (14). In contrast, most melanomas arising from sun-protected skin, such as acral lentiginous or mucosal malignant melanoma, do not acquire BRAF or RAS mutation. It is thus proposed that other signaling pathways might mediate tumor formation in acral lentiginous melanoma (17), which is the most common melanoma subtype in Asians. We have shown that most of the cases we studied are acral lentiginous malignant melanoma (accounting for 78.6%) and all of the cases of acral lentiginous malignant melanoma screened do not have activating BRAF (T1799A) mutation. We further point out that activation of the ERK/MEK pathway by overexpression of ARMS instead of constitutively active BRAF may account for tumorigenesis of acral lentiginous melanoma. It strongly implies that ARMS, as an upstream signaling molecule of RAS and BRAF in the MEK/ERK pathway, could promote advantageous growth of melanoma by overexpressing itself to activate the MEK/ERK pathway under cellular stress.

Grant support: National Science Council grant NSC-94-2314-B-002-262 (Y-H. Liao) and National Taiwan University Hospital grant NTUH-96A-03-3 (P-H. Huang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Chung-Wu Lin for helping us in the generation of tissue arrays and laser-captured microscopic dissection of tumor tissues; Prof. Zee-Fen Chang (Institute of Biochemistry and Molecular Biology, College of Medicine, National Taiwan University, Taipei, Taiwan) for the gift of the constitutively active MEK construct; Yi-Ling Huang, Wei-Ling Tsai, and Yi-Jei Huang for technical assistance; Rui-Ying Shieh for tissue sections; and all members in Huang's lab for discussion.