A Genomic Screen Identifies TYRO3 As a MITF Regulator in Melanoma

A genomic screen identifies TYRO3 as a MITF regulator in melanoma

Shoutian Zhua, Heiko Wurdaka,1, Yan Wangb,1, Anna Galkinb, Haiyan Taob, Jie Lib, Costas A. Lyssiotisa, Feng Yanb, Buu P. Tub, Loren Miragliab, John Walkerb, Fanxiang Sunb, Anthony Orthb, Peter G. Schultza,b,2, and Xu Wub,2

aThe Skaggs Institute of Chemical Biology and the Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037; and bGenomics Institute of the Novartis Research Foundation, 10675 John Jay Hopkins Drive, San Diego, CA 92121

Contributed by Peter G. Schultz, August 19, 2009 (sent for review July 1, 2009) Malignant melanoma is the most aggressive form of cutaneous including Wnt (7), and more specifically transcription factors carcinoma, accounting for 75% of all deaths caused by skin cancers. such as LEF-1, CREB, SOX10, and PAX3 have been demon- Microphthalmia-associated transcription factor (MITF) is a master strated to bind the MITF-M promoter and regulate its expression gene regulating melanocyte development and functions as a (8, 9). Recent genomic analyses have shown that the MITF locus ‘‘lineage addiction’’ oncogene in malignant melanoma. We have is amplified in melanoma cells (Ϸ10% of cutaneous tumors and identiﬁed the receptor protein tyrosine kinase TYRO3 as an up- 15–20% of metastatic tumors) and that MITF can function as a stream regulator of MITF expression by a genome-wide gain-of- lineage-survival oncogene (10, 11). Expression of MITF-M is function cDNA screen and show that TYRO3 induces MITF-M also deregulated in clear cell sarcoma (CCS), driving CCS cell expression in a SOX10-dependent manner in melanoma cells. survival and tumor growth (12). Expression of TYRO3 is signiﬁcantly elevated in human primary To further clarify the biological role of MITF-M and to identify melanoma tissue samples and melanoma cell lines and correlates additional regulators of MITF-M expression, we performed a with MITF-M mRNA levels. TYRO3 overexpression bypasses genome-wide gain-of-function cDNA screen with a MITF-M pro- BRAF(V600E)-induced senescence in primary melanocytes, induc- moter reporter assay. TYRO3 was identified as a positive regulator ing transformation of non-tumorigenic cell lines. Furthermore, of MITF-M transcription in melanoma cells. Increased expression TYRO3 knockdown represses cellular proliferation and colony of TYRO3 was detected in established melanoma cancer cell lines, formation in melanoma cells, and sensitizes them to chemothera- as well as primary tumor samples. Overexpression of TYRO3 peutic agent-induced apoptosis; TYRO3 knockdown in melanoma overcomes BRAF(V600E)-induced senescence in melanocytes, cells also inhibits tumorigenesis in vivo. Taken together, these data whereas knockdown of TYRO3 inhibits melanoma cell prolifera- indicate that TYRO3 may serve as a target for the development of tion, sensitizes them to chemotherapeutic agents in vitro, and therapeutic agents for melanoma. inhibits tumorigenesis in vivo. These data suggest that targeting TYRO3 with small molecules or antibodies may potentially be elanoma is a rare subtype of skin cancer that arises from useful for the treatment of melanoma.

melanocytes, the specialized pigment cells that synthesize CELL BIOLOGY M Results melanins and are responsible for skin color. Approximately 212,000 new cases of melanoma are diagnosed annually worldwide, resulting A Genome-Wide Gain-Of-Function Screen Identifies TYRO3 as a Posi- in 65,000 melanoma-related deaths (1). Although not as common tive Regulator of Mitf-M Transcription in Mouse Melanoma Cells. In as other subtypes of cutaneous carcinomas, melanoma is the most an effort to identify upstream factors regulating Mitf-M tran- dangerous form, accounting for 75% of skin cancer deaths (http:// scription, we performed a genome-wide gain-of-function cDNA www.aafp.org/afp/20000715/357.html, OMIM 155600). When diag- screen using a Mitf-M promoter driven luciferase reporter gene nosed early, more than 80% of cases can be successfully treated as a readout (13). The Mammalian Gene Collection (MGC) through surgical resection. However, metastatic malignant mela- consisting of approximately 16,000 cDNAs was screened by noma is resistant to currently available treatments and carries a very co-transfection of reporter plasmids and cDNA expressing plas- poor prognosis with a less than 10% 5-year survival rate (2–4). mids into B16-F0 mouse melanoma cells using a 384-well, Ras/Raf and PI3K/Akt signaling have been implicated in melanoma high-throughput screening format. The average fold activation development; however, the exact mechanism(s) of melanocyte (afa) value, calculated for each gene based on the signal strength transformation, disease progression, and development of chemore- in two independent experiments, was used for hit selection. Ͼ sistance are still not well understood. Using afa 3.0 as the cutoff point, 263 hits were selected from As a master transcription factor in melanocytes, microphthal- the primary screen. To filter out false positives that non- mia-associated transcription factor (MITF) controls the expres- specifically activate luciferase activity, a counter screen was sion of genes important for melanocyte development, function performed using a group of reporter constructs driven by and survival (5). Alpha-melanocyte stimulating hormone (MSH) alternative promoter sequences, including Mitf-A, Mitf-H, and Ͼ activates melanocortin-1 receptor (MC1R), resulting in a cAMP- cAMP-responsive element (CRE). Using an afa 3.0 for the Ͻ mediated phosphorylation/activation of the CREB/ATF family Mitf-M reporter and an afa 3.0 for all other reporters as a of transcription factors, induces MITF expression, and regulates selection criteria, 23 genes were chosen for further follow-up as pigmentation in melanocytes. The MITF locus contains several positive regulators of Mitf-M reporter expression in melanoma distinct promoters and alternative transcription initiation sites that generate multiple isoforms of MITF, which differ in their Author contributions: S.Z., H.W., P.G.S., and X.W. designed research; S.Z., H.W., Y.W., A.G., amino-termini and expression patterns (Fig. S1). Among them H.T., J.L., F.Y., and B.P.T. performed research; S.Z., H.W., Y.W., A.G., H.T., C.A.L., L.M., J.W., MITF-A and MITF-H are most abundant and widely expressed F.S., A.O., P.G.S., and X.W. analyzed data; and S.Z., H.W., C.A.L., P.G.S., and X.W. wrote the in many cell types; MITF-E is expressed abundantly in cultured paper. mast cells (6); and MITF-M is predominantly expressed in The authors declare no conﬂict of interest. melanocytes and pigmented melanoma cells. The differential 1H.W. and Y.W. contributed equally to this work. expression of MITF isoforms in a variety of tissues suggests that 2To whom correspondence may be addressed. E-mail: [email protected] or [email protected]. other mechanisms may affect their expression in addition to This article contains supporting information online at www.pnas.org/cgi/content/full/ MSH signaling. For example, a variety of signaling cascades, 0909292106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0909292106 PNAS ͉ October 6, 2009 ͉ vol. 106 ͉ no. 40 ͉ 17025–17030 Downloaded by guest on September 26, 2021 A AB * * TYRO3 SOX10 2.0 * * MITF γ−Tubulin * * * * * * * * * * * SOX10 CTL Si-1 Si-2 (Nuclear) MITF-M 1.0 8 Empty vector CREB *

(Nuclear) 6 TYRO3 4 Relative mRNA Level,Relative mRNA PAX3 4 0.5 C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 (Nuclear) 10 x 2 γ-Tubulin Luciferase Signal Luciferase 0 CTL Si-1 Si-2 B CTL TYRO3 12 C D Flag-SOX10 9 (Nuclear) WB: TYRO3 p-Tyr Flag-SOX10

TYRO3 6 (Total) γ 3 -Tubulin

Relative mRNA Level, Level, mRNA Relative CTL WT KD 8 0 E

III IIIA IIIB IIIC IV 6 * 4 Normal Melanoma Tissue Samples 4 CTL WT KD CTL WT KD 10 x Fig. 1. A genomic screen identiﬁes TYRO3 as a regulator of Mitf-M tran- 2 0

scription in melanoma cells. (A) Quantitative RT-PCR for the mRNA of Mitf-M Signal Luciferase CTL WT KD in melanoma cells transfected with cDNA hits. Asterisks indicate signiﬁcant difference from the empty vector control (P Ͻ 0.05). Data are presented as mean Ϯ standard deviation. A complete list of the genes used can be found in F TYRO3 Table S1.(B) Real time quantitative RT-PCR for the mRNA of TYRO3 in human SOX10 primary melanoma tissues. The Roman numerals on the x axis indicate the (Nuclear) melanoma stages of the samples. The ﬁrst three samples are from normal skin SOX10 tissues. (Total) γ−Tubulin

CTL shRNA1 shRNA2 CTL shRNA1 shRNA2 cells. To confirm their function in wild-type melanoma cells, plasmids containing these cDNAs were transfected into B16-F0 A2058 Cells MDA-MB-435 Cells Mitf-M cells and mRNA levels of mouse were analyzed using Fig. 2. TYRO3 regulates SOX10 nuclear localization in melanoma cells. (A) quantitative reverse-transcription PCR (qRT-PCR, Fig. 1A; see Western blot of TYRO3, MITF-M, and ␥-tubulin for the whole cell lysate and the gene list in Table S1). Among these 23 genes, 15 genes SOX10, CREB, and PAX3 for the nuclear fraction of empty vector (CTL) and significantly increased Mitf-M transcription as compared to the TYRO3-transfected B16-F0 melanoma cells. (B) Sox10 knockdown abolishes empty vector control. A number of these genes have been Tyro3 overexpression induced Mitf-M reporter activity in B16-F0 melanoma associated with cancers previously, including retinoic acid re- cells. Si-1 and -2 are two siRNAs targeting mouse Sox10 mRNA with different ceptor, RARa [translocation mutation in acute promyelocytic sequences. CTL: cells were transfected with non-targeting siRNA. Asterisks indicate signiﬁcant difference from the empty vector control (P Ͻ 0.05). Data leukemia, (14)], and receptor tyrosine kinase, Tyro3 [overex- are presented as mean ϩ standard deviation. (C) Overexpression of wild-type pressed in mouse mammary tumors (15)]. (WT) TYRO3 in 293T cells increases tyrosine phosphorylation in the cells. Left: Analysis of the tissue and cell line specific expression of the Western blot of TYRO3 for whole cell lysate; Right: Western blot of phospho- selected 15 genes revealed elevated levels of TYRO3 in human tyrosine for whole cell lysate, CTL: empty vector transfected cells; WT: wild- melanoma cell lines [GNF SymAtlas, http://symatlas.gnf.org/ type TYRO3-transfected cells; KD: kinase-dead mutant (K550A) TYRO3- SymAtlas/ (16)] but not in other human cancer lines across the transfected cells. (D) Western blot of nuclear localized Flag-SOX10 for cells co-transfected with Flag-SOX10 and empty vector (CTL), wild-type TYRO3 NCI60 panel. Furthermore, we detected increased expression of (WT), or kinase-dead mutant TYRO3 (KD). (E) MITF-M luciferase reporter assay TYRO3 in primary human melanoma samples at different stages in B16 cells co-transfected with empty vector (CTL), wild-type TYRO3 (WT), or compared to normal skin tissues using real time quantitative kinase-dead TYRO3 (KD). Asterisks indicate signiﬁcant difference from the RT-PCR (Fig. 1B). Twenty out of 40 melanoma tissue samples empty vector control (P Ͻ 0.05). Data are presented as mean ϩ standard showed increased TYRO3 mRNA levels (over 3-fold) compared deviation. (F) Western blot for TYRO3, nuclear SOX10, total SOX10, and ␥ to normal skin tissues, although there was no obvious correlation -Tubulin in TYRO3 knockdown A2058 and MDA-MB-435 melanoma cells. between TYRO3 mRNA levels and tumor stage. TYRO3 expression correlated well with elevated expression levels of MITF-M a variety of cellular systems (20–22), and play important roles in in these samples (Fig. S2 ) as well. Neither AXL nor MER, the different types of cancer (23–25). other two members of the TYRO3/AXL/MER (TAM) receptor protein-tyrosine kinase family, is differentially expressed be- TYRO3 Regulates MITF-M Transcription Through the Control of SOX10 tween melanoma tumor samples and normal skin tissues (Fig. Nuclear Localization. MITF-M expression is controlled by the S2). These results suggest that TYRO3 may play an important CREB, PAX3, and SOX10 transcription factors (8, 9). To role in melanoma etiology. Other genes from the screen do not determine whether TYRO3 affects the activation of established show significant overexpression in melanoma cell lines, and MITF-M regulators, we analyzed the effects of TYRO3 on the therefore, we focused our further follow-up studies on TYRO3. nuclear protein levels of CREB, PAX3, and SOX10 when As a member of the TAM receptor family, TYRO3 has been overexpressed in B16-F0 cells. Concomitant with an increase in implicated in the regulation of the immune system (17), includ- TYRO3 and MITF-M expression, we observed an increase in ing natural killer cell differentiation (18), and macrophage and SOX10 nuclear localization, whereas nuclear protein levels of neutrophil functions (19). The TAM family receptor kinases are CREB or PAX3 were not affected (Fig. 2A). To verify the role also involved in cell proliferation, survival and transformation in of SOX10 in mediating MITF-M transcription downstream of

17026 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0909292106 Zhu et al. Downloaded by guest on September 26, 2021 TYRO3, we knocked down Sox10 expression in B16-F0 cells

A B 6 CTL using gene-specific siRNAs (Dharmacon). As expected, Sox10 1.2 60 knockdown diminished Tyro3 overexpression induced Mitf-M shRNA1 shRNA2 reporter activation (Fig. 2B). Knockdown of Sox10 also slightly 0.8 40 decreased the reporter signal in empty vector transfected cells, 0.4 20

consistent with the fact that endogenous SOX10 contributes 0 X10 Number Cell 0 partially to Mitf-M expression in these cells. RT-PCR analysis of LevelRelative mRNA CTL shRNA1 shRNA2 0 2 4 6 8 10 Days after cell seeding Mitf-M mRNA levels in wild-type B16-F0 cells gave similar results (Fig. S3). CD* CTL 60 To determine whether human TYRO3 overexpression is suf- DMSO * shRNA1 shRNA2 ficient to drive the ligand-independent activation of downstream 40 cell signaling, we ectopically expressed wild-type (WT) TYRO3 * * and its kinase-dead (KD) mutant (lysine 550 in the conserved Cisplatin 20

ATP-binding site, which is analogous to lysine 614 of MER (26) Caspase3/7 Activity was replaced by alanine, K550A) in 293T cells. TYRO3 overex- 0 pression increased the overall level of tyrosine phosphorylation CTL shRNA1 DMSO Cisplatin TMZ Docetaxel in the whole cell lysate (Fig. 2C), indicating hyperactivation of E F BRAF TYRO3 the downstream signaling components, which is consistent with phospho-ERK Cisplatin previous findings that Src/Fyn/Yes family kinases and PI3K can γ−Tubulin be recruited and phosphorylated by TYRO3 upon activation/ Total ERK autophosphorylation (27, 28). The kinase-dead mutant form of CTL shRNA1 BRAF(V600E) - + TYRO3 - + TYRO3 failed to elevate phospho-tyrosine levels, confirming the requirement for kinase activity in cell signaling. G To further investigate the effect of TYRO3 on SOX10 localization, we co-transfected wild-type and kinase-dead TYRO3, and Flag-tagged SOX10 into 293T cells. Assessment of nuclear BRAF(V600E) - - + + - + - + Flag-SOX10 level by Western blot using an anti-Flag antibody TYRO3 indicated that the levels of nuclear SOX10 increased upon H wild-type TYRO3 overexpression, as compared to the control cells (Fig. 2D). Overexpression of the kinase-dead mutant did not enhance SOX10 nuclear localization, consistent with the BRAF(V600E) - - + + absence of increased phospho-tyrosine levels. In addition, the TYRO3 - + - + kinase-dead mutant failed to activate the MITF-M reporter gene as well (Fig. 2E), again indicating that the effects of TYRO3 are Fig. 3. TYRO3 is involved in apoptosis and survival in melanoma cells and dependent on its kinase activity. However, the kinase dead protects melanocytes from BRAF(V600E)-induced senescence. (A) Real time mutant of TYRO3 does not function as a ‘‘dominant negative’’ RT-PCR for TYRO3 knockdown in A2058 melanoma cells. (B) Cell proliferation CELL BIOLOGY by competing with the wild-type receptor and further inhibiting assay for TYRO3 knockdown in A2058 melanoma cells. (C) Immunoﬂuorescent the basal level of signaling, suggesting that the basal level of staining for cleaved caspase 3 in A2058 cells treated with cisplatin or DMSO. Red: cleaved caspase 3; blue: DAPI (nuclei). (Scale bars, 50 ␮m.) (D) Caspase 3/7 SOX10 translocation and MITF expression might be regulated by activity assay for A2058 cells treated with chemotherapeutic agents. TMZ: other mechanisms. To confirm this mechanism in human mel- temozolomide. Asterisks indicate signiﬁcant difference from DMSO treated anoma cells, we performed shRNA-mediated TYRO3 knock- control (P Ͻ 0.05). Data are presented as mean ϩ standard deviation. (E) down in A2058 and MDA-MB-435 melanoma cells. Knockdown TUNEL staining for A2058 cells treated with cisplatin. Green: TUNEL; blue: of TYRO3 decreased nuclear SOX10 levels without affecting DAPI (nuclei). Arrowheads indicate the positive nuclear TUNEL staining. (Scale overall protein levels (Fig. 2F) or affecting SOX10 transcription bars, 50 ␮m.) (F) Overexpression of BRAF(V600E) and TYRO3 using retroviral (Fig. S4). These experiments demonstrate that TYRO3 acts particles in human primary melanocytes. (G) Staining of cell colonies for upstream of SOX10 to regulate its nuclear localization and primary melanocytes infected with BRAF(V600E)- and TYRO3-expressing ret- ␤ ␤ provide a link between TYRO3 and MITF-M through SOX10 rovirus. (H) Senescence-associated - galactosidase (SA- -Gal) activity staining for primary melanocytes infected with BRAF(V600E) and TYRO3 expressing activation. retrovirus. (Scale bars, 50 ␮m.) To further explore the function of TYRO3 in melanoma cells, we performed gene expression analysis in A2058 and MDA- MB-435 melanoma cells upon TYRO3 knockdown. Genes whose TYRO3 Is Involved in Melanoma Cell Proliferation and Survival. Mem- expression levels changed more than 2-fold by TYRO3 knock- bers of TAM protein kinase receptor family are known to play down were subjected to pathway analysis using Ingenuity Path- a role in cell survival and protection against apoptosis (22, 29, way Analysis software. The pathways affected by TYRO3 knock- 30). To further address the biological function of TYRO3 in down were similar for two different shRNAs and two different melanoma, we analyzed its role in proliferation and apoptosis in melanoma cell lines (Fig. S5A). Pathway analysis revealed that different melanoma cell lines using RNAi-mediated gene silenc- TYRO3 knockdown has effects on a wide range of signaling ing. Lentiviral particles carrying scrambled control or TYRO3 pathways (Fig. S5B and Tables S2 and S3), including PI3K/AKT TYRO3 signaling, Wnt/␤-Catenin signaling, ERK/MAPK signaling, targeting shRNA sequences were used to induce stable SPAK/JNK signaling, JAK/Stat signaling, TGF-␤ signaling, and knockdown in melanoma cell lines. Three out of five shRNA RAR activation, to name a few. Interestingly, RARa, PIK3c3, sequences demonstrated 60–80% knockdown efficiency at both and GPR114 were identified as hits from the genomic screen, the protein and mRNA levels, and two of those constructs were indicating potential crosstalk and involvement of these signaling used for subsequent functional studies (Figs. 2F and 3A). pathways in melanoma. Furthermore, genes whose expression is First, we analyzed the effect of TYRO3 knockdown on mel- affected by TYRO3 knockdown are involved in a wide range of anoma cell proliferation in vitro. TYRO3 knockdown signifi- cellular processes, including proliferation, survival, transforma- cantly reduced the proliferation rates of A2058 melanoma cells tion, growth, apoptosis, and senescence, suggesting possible roles (Fig. 3B); both shRNA constructs decreased melanoma cell for TYRO3 in these processes. proliferation to a similar extent (Ϸ10% of the control shRNA),

Zhu et al. PNAS ͉ October 6, 2009 ͉ vol. 106 ͉ no. 40 ͉ 17027 Downloaded by guest on September 26, 2021 indicating that TYRO3 plays an important role in the regulation A of cell proliferation. Next, we examined whether TYRO3 mediates chemoresis- tance in melanomas, similar to its downstream effector MITF (10). The effect of TYRO3 knockdown on caspase activation was determined in the presence of chemotherapeutic agents com- CTL shRNA1 shRNA2 monly used for the treatment of melanoma. TYRO3 knockdown B C alone induced low levels of caspase 3 activity in A2058 cells as 50 compared to scrambled shRNA infected cells. When TYRO3 60 40 knockdown was coupled with cisplatin (10 ␮M) or docetaxel (50 40 30 20 nM) treatment, a significant increase (5-fold for cisplatin and 20

approximately 2- to 3-fold for docetaxel, respectively) in caspase * * 10 * Number of Colonies of Number * Number of Colonies of Number 0 0 activity was observed (Fig. 3 C and D). Interestingly, no syner- CTL shRNA1 shRNA2 CTL shRNA1 shRNA2 gistic effect with temozolomide (TMZ, 50 ␮M) treatment was observed. Consistent with the above results, cells treated with D cisplatin showed increased TUNEL staining with TYRO3 knockdown compared to control shRNA (Fig. 3E). These experiments indicate that TYRO3 knockdown sensitizes melanoma cells to chemotherapeutic agents in a manner consistent with MITF’s CTL shRNA1 shRNA2

protective role against apoptosis (31, 32). They also provide 3 further support for the interaction of the TYRO3 and MITF E 600 CTL shRNA1 signaling cascades in melanoma, and suggest that TYRO3 may 400 shRNA2 be a therapeutic target for melanoma treatment in combination with other chemotherapeutic agents. 200

Tumor Volume, mm Volume, Tumor 0 TYRO3 Overexpression Bypasses BRAF(V600E)-Induced Senescence in 1 17 20 22 24 27 31 Melanocytes. Melanoma arises from melanocytes under the in- Days post implantation fluence of transforming stimuli, such as activating mutations of BRAF and repression of p16INK4a expression (33). The Fig. 4. TYRO3 knockdown represses tumor formation of A2058 melanoma BRAF(V600E) mutation has been widely found in melanoma. cells in vivo. (A) Adhesion-dependent colony formation assay for TYRO3 However, it is insufficient to transform primary melanocytes by knockdown in A2058 melanoma cells. (B) Adhesion-dependent colony formation of A2058 cells upon TYRO3 knockdown. Asterisks indicate signiﬁcant itself, and rather induces senescence in the absence of additional difference from the non-targeting shRNA control (P Ͻ 0.05). Data are pre- oncogenic stimuli. Synergistic oncogenic mutations or silencing sented as mean ϩ standard deviation. (C) Adhesion-independent (soft-agar) of tumor suppressors, such as upregulation of MITF (10) or colony formation of A2058 cells upon TYRO3 knockdown. Asterisks indicate downregulation of IGFBP7 (34), are required to facilitate me- signiﬁcant difference from the non-targeting shRNA control (P Ͻ 0.05). Data lanocyte transformation and melanoma development. are presented as mean ϩ standard deviation. (D) Tumor formation of A2058 Tyro3 overexpression is able to transform 3T3 and Rat-2 fibro- melanoma cells at day 34 post-implantation. One million cells were injected blasts in a ligand-independent manner in vitro (27, 35). To address into nude mouse intradermally with a volume of 50 ␮L. (E) Tumor growth whether TYRO3 is involved in melanocyte transformation, we curves for (D). No tumor formed from either shRNA1 or shRNA2 expressing overexpressed TYRO3 together with BRAF(V600E) in early pas- A2058 melanoma cells till the end of the study (day 34). sage of human primary melanocytes using retroviral particles (Fig. 3F). Expression of TYRO3 alone did not have clear effect com- cells (Fig. 4 A and B). This decrease correlated with the pared to control virus infected melanocytes. Expression of knockdown efficiency of the shRNA sequences. In addition to BRAF(V600E) induced senescence in melanocytes and led to an inhibiting adhesion-dependent colony formation, TYRO3 increase in the percentage of senescent cells with the characteristic knockdown significantly decreased the number of colonies enlarged cell body and positive staining of senescence-associated formed under adhesion-independent soft-agar conditions (Fig. ␤-galactosidase (SA-␤-Gal) activity (Fig. 3 G and H). However, 4C). when TYRO3 was co-expressed with the BRAF(V600E) mutant, primary melanocytes formed colonies in which non-senescent cells To address the role of TYRO3 in tumorigenesis of melanoma dominate the population, as indicated by a strong reduction in the cells in vivo, we injected the TYRO3 knockdown A2058 mela- percentage of SA-␤-Gal positive cells. These results indicate that noma cells into nude mice intradermally and monitored tumor TYRO3 overexpression can bypass the BRAF(V600E) mutant initiation and growth. As expected, cells infected with non- induced senescence in melanocytes, and thus may play a role in targeting shRNA control formed detectable tumors after 12–15 melanoma development. days post implantation, and the tumors grew at a fast pace and started bleeding at day 30 (Fig. 4 D and E). In contrast, the cells TYRO3 Knockdown Reduces the Tumorigenic Potential of Melanoma with TYRO3 knockdown did not form any detectable tumor by Ϸ Cells In Vivo. Melanoma cells are capable of forming xenograft day 30. At day 52, two out of five mice had detectable tumors ( 3 tumors when injected into mice. To test the effect of TYRO3 mm3) from melanoma cells infected with shRNA1 targeting on tumor formation by melanoma cells, we performed tumor- TYRO3, and one out of five mice had a detectable tumor from igenesis assays with cells in which TYRO3 levels were knocked melanoma cells infected with shRNA2 targeting TYRO3. These down with shRNAs. First, we examined the colony forming experiments demonstrate that TYRO3 regulates tumor initia- ability of melanoma cells upon TYRO3 knockdown under tion and growth in vivo. low-density or non-adherent conditions in vitro. When plated at low density (200 cells per well of 6-well plates), melanoma Discussion cells form colonies in 2–3 weeks. We observed a substantial Melanoma is an extremely aggressive disease with high meta- decrease in the number of colonies (10–15% of the control static potential and strong resistance to conventional chemo- shRNA expressing cells) formed by A2058 cells stably express- therapeutics. The transcription factor MITF has been identified ing the TYRO3 shRNAs relative to scrambled shRNA infected as a ‘‘master regulator’’ of melanocytes and as a ‘‘lineage

17028 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0909292106 Zhu et al. Downloaded by guest on September 26, 2021 addiction’’ oncogene in malignant melanoma (36). Conse- activates murine TYRO3 in vitro (53), bovine Protein S activates quently, the identification of upstream regulators of MITF-M human TYRO3 (54, 55), but human Protein S is not able to expression may reveal potential therapeutic targets for meta- activate human TYRO3 (56). Neither GAS6 nor Protein S has static chemorefractory melanoma. In this study, we identified the yet been shown to be connected to melanoma. Moreover, TAM receptor tyrosine kinase TYRO3 as a positive regulator of family receptors are capable of autophosphorylation and ligand- MITF-M expression in melanoma cells by a genome-wide gain- independent self-activation with elevated expression, which may of-function cDNA screen. The expression levels of TYRO3 are also be involved in tumorigenesis (37, 57). Clearly, further elevated in both melanoma cells and primary melanoma samples investigation of the roles of GAS6 and Protein S in melanoma compared to other cell lines and normal tissues. Overexpression relative to TYRO3 activation is merited. of TYRO3 induces its autophosphorylation and activation of Finally, we have shown that TYRO3 knockdown in aggressive downstream signaling (28, 37). We also demonstrate that TY- melanoma cells significantly reduces the in vivo tumorigenic RO3 regulates MITF-M expression by controlling SOX10 nu- potential in both intradermal and s.c. (Fig. S6) xenograft models clear localization, and such effects are dependent on its kinase in nude mice, which is consistent with the effects of TYRO3 activity. SOX10 is a transcription factor that plays a critical role knockdown on melanoma cells in vitro. In summary, these data in neural crest development (38) and is highly expressed in indicate that TYRO3 overexpression is a driving oncogenic event primary melanocytes and melanoma cells. We are currently in melanoma, and support efforts to pursue development of investigating whether SOX10 is phosphorylated by TYRO3 or small molecule or antibody kinase inhibitors as a therapeutic downstream signaling events, and whether such events enhance approach for malignant melanoma. SOX10 nuclear localization. TYRO3 and other members of the TAM tyrosine kinase Methods receptor family regulate a variety of signaling cascades in- Cell Culture. Melanoma cell line MDA-MB-435 was propagated in DMEM/F12 volved in cellular proliferation and transformation (35, 39, 40), (Invitrogen) medium supplemented with non-essential amino acids (Invitro- as well as anti-apoptotic and survival mechanisms in cancers gen), and 10% FBS (Invitrogen). A2058 and B16-F0 cells were purchased from (24, 25, 41). TYRO3 knockdown inhibits melanoma colony ATCC and cultured under the recommended conditions. 293T (ATCC) and GP2 formation under adherent and non-adherent conditions, sug- cells were grown in DMEM (Invitrogen) supplemented with 10% FBS. Human gesting that TYRO3 may play a role in metastatic melanoma primary melanocytes were purchased from Invitrogen, and grown in 254 survival and contribute to its dissemination. In addition, medium supplemented with melanocyte growth supplements (Invitrogen). TYRO3 knockdown sensitized melanoma cells to cisplatin and High-Throughput Genomic Screening. cDNA library in pCMV-sport6 expression docetaxel chemotherapy-induced apoptosis, suggesting that plasmids were co-transfected with reporter plasmids using Lipofectamine TYRO3 signaling is important in the development of che- transfection reagent. Cells were grown for 48 h in a 37 °C, 5% CO2 incubator, moresistance and has anti-apoptotic effects in cancer cells. It firefly luciferase activity was detected using the Bright-Glo luciferase sub- remains to be determined if this chemosensitization is due to strate (Promega). the inactivation of downstream anti-apoptotic signaling pathways such as PI3K/Akt, repression of anti-apoptotic factors Cell Proliferation Assay. Cells were seeded at a density of 1 million cells per such as MITF-M and BCL2, or the re-activation of the BRAF 10-cm dish. Cells were trypsinized and cell number and viability were assessed (V600E)-driven senescence pathway. by trypan blue exclusion staining at day 2, 4, 6, 8, and 10 post cell seeding using CELL BIOLOGY The valine to glutamic acid substitution in BRAF(V600E) is ViCell system (Beckman Coulter). found in 50% to 70% of melanomas (42, 43). However, this BRAF mutation alone is not sufficient to transform melano- Apoptosis Assay. Cells were treated with cisplatin, temozolomide, or docetaxel cytes, but rather induces senescence, despite hyperactivated for 24 h, caspase 3/7 activity was determined using Caspase 3/7 Glo Assay (Promega). Activated Caspase 3 was immunostained with anti-cleaved ERK signaling and activation of cellular proliferation pathways. Caspase 3 antibody (Cell Signaling), as well. TUNEL staining was performed Acquisition of additional ‘‘driving’’ oncogenic events is required 48 h post-treatment using In Situ Cell Death Detection Kit (Roche) following to evade BRAF(V600E)-induced senescence and mediate trans- the manufacturer’s instructions. formation. For example, downregulation of IGFBP7 (34) and c-Myc overexpression (44) cooperate with activated BRAF to Colony Formation Assay. Colony formation assays were carried out as previ- transform melanocytes. Hyperactivated ERK signaling has been ously described (59). Briefly, 200 cells were plated into each well of 6-well shown to target MITF for degradation (45, 46), suggesting that plate, and cells were grown for 2–3 weeks until colonies became visible. The reduced levels of the melanocyte ‘‘master regulator’’ may play a colonies were visualized by Crystal Violet staining. role in BRAF-induced senescence. Mechanisms that restore MITF expression may therefore contribute to melanocyte trans- Senescent Cell Staining. Primary melanocytes were infected with V600E BRAF formation. Here we show that TYRO3 is a regulator of MITF and TYRO3 and stained for senescence-associated ␤-galactosidase (SA-␤-Gal) expression and that elevated levels of TYRO3 in primary activity on day 14 post infection with the senescent cell staining kit (BioVision) following the manufacturer’s protocol. melanocytes can bypass oncogenic BRAF(V600E) mutant- induced senescence. Tyro3 overexpression is capable of trans- In Vivo Tumorigenic Assay. Cells were trypsinized and resuspended in serum forming 3T3 fibroblasts in part through the activation of the free DMEM. One million cells in a volume of 50 ␮L were injected into each nude PI3K pathway (27). The combinatorial effect of TYRO3 medi- mouse intradermally. Tumor sizes were measured once every 3 days, the ated induction of MITF expression and PI3K/Akt pathway volume was calculated as length ϫ width ϫ height. activation may provide the necessary survival signals to over- come BRAF(V600E)-induced senescence and drive melanoma Statistical Analysis. Statistical analyses were performed using Microsoft Excel. development. Student’s t test (two-tailed) was performed to determine significance when Growth arrest-specific gene 6 (GAS6), a vitamin K-dependent comparing data from different treatment groups. P Ͻ 0.05 was considered to protein, is a well-studied ligand of the TYRO3/AXL/MER represent significant difference. family receptors, capable of binding to and activating TYRO3 to regulate cell cycle, survival, and apoptosis (30, 47–51). Concom- SI. Detailed experimental procedures can be found in SI Test . itant expression of GAS6 and TAM receptors have been dis- ACKNOWLEDGMENTS. We thank Nathan Englund, Sergio Briones, James covered in certain cancers, including breast cancer (52) and Watson, and Jie Zhang for technical support; and Drs. Jun Liu, Jianwei Che, glioma (24, 39). The role of Protein S, a close relative of GAS6, Venkateshwar Reddy, and Charles Y. Cho for helpful discussions. This work is as a TYRO3 ligand is less clear: human Protein S binds to and supported by the Novartis Research Foundation (P.G.S and X.W.).

Zhu et al. PNAS ͉ October 6, 2009 ͉ vol. 106 ͉ no. 40 ͉ 17029 Downloaded by guest on September 26, 2021 1. Lucas R, McMichael T, Smith W, Armstrong B (2006) Solar ultraviolet radiation: Global 31. McGill GG, et al. (2002) Bcl2 regulation by the melanocyte master regulator Mitf burden of disease from solar ultraviolet radiation. Environmental Burden of Disease modulates lineage survival and melanoma cell viability. Cell 109:707–718. Series, World Health Organization. 32. Dynek JN, et al. (2008) Microphthalmia-associated transcription factor is a critical 2. Bajetta E, et al. (2002) Metastatic melanoma: Chemotherapy. Semin Oncol 29:427–445. transcriptional regulator of melanoma inhibitor of apoptosis in melanomas. Cancer 3. Bastiaannet E, Beukema JC, Hoekstra HJ (2005) Radiation therapy following lymph Res 68:3124–3132. node dissection in melanoma patients: Treatment, outcome and complications. Cancer 33. Bennett DC (2008) How to make a melanoma: What do we know of the primary clonal Treat Rev 31:18–26. events? Pigment Cell Melanoma Res 21:27–38. 4. Balch CM, et al. (2001) Final version of the American Joint Committee on Cancer staging 34. Wajapeyee N, Serra RW, Zhu X, Mahalingam M, Green MR (2008) Oncogenic BRAF system for cutaneous melanoma. J Clin Oncol 19:3635–3648. induces senescence and apoptosis through pathways mediated by the secreted protein 5. Lin JY, Fisher DE (2007) Melanocyte biology and skin pigmentation. Nature 445:843– IGFBP7. Cell 132:363–374. 850. 35. Lai C, Gore M, Lemke G (1994) Structure, expression, and activity of Tyro 3, a neural 6. Steingrimsson E, Copeland NG, Jenkins NA (2004) Melanocytes and the microphthalmia transcription factor network. Annu Rev Genet 38:365–411. adhesion-related receptor tyrosine kinase. Oncogene 9:2567–2578. 7. Widlund HR, et al. (2002) Beta-catenin-induced melanoma growth requires the down- 36. Goding CR (2000) Mitf from neural crest to melanoma: Signal transduction and stream target microphthalmia-associated transcription factor. J Cell Biol 158:1079– transcription in the melanocyte lineage. Genes Dev 14:1712–1728. 1087. 37. Taylor IC, Roy S, Varmus HE (1995) Overexpression of the Sky receptor tyrosine kinase 8. Huber WE, et al. (2003) A tissue-restricted cAMP transcriptional response: SOX10 at the cell surface or in the cytoplasm results in ligand-independent activation. Onco- modulates alpha-melanocyte-stimulating hormone-triggered expression of microph- gene 11:2619–2626. thalmia-associated transcription factor in melanocytes. J Biol Chem 278:45224–45230. 38. Mollaaghababa R, Pavan WJ (2003) The importance of having your SOX on: Role of 9. Potterf SB, Furumura M, Dunn KJ, Arnheiter H, Pavan WJ (2000) Transcription factor SOX10 in the development of neural crest-derived melanocytes and glia. Oncogene hierarchy in Waardenburg syndrome: Regulation of MITF expression by SOX10 and 22:3024–3034. PAX3. Hum Genet 107:1–6. 39. Vajkoczy P, et al. (2006) Dominant-negative inhibition of the Axl receptor tyrosine 10. Garraway LA, et al. (2005) Integrative genomic analyses identify MITF as a lineage kinase suppresses brain tumor cell growth and invasion and prolongs survival. Proc Natl survival oncogene amplified in malignant melanoma. Nature 436:117–122. Acad Sci USA 103:5799–5804. 11. King R, et al. (1999) Microphthalmia transcription factor. A sensitive and specific 40. Chen J, Carey K, Godowski PJ (1997) Identification of Gas6 as a ligand for Mer, a neural melanocyte marker for melanoma diagnosis. Am J Pathol 155:731–738. cell adhesion molecule related receptor tyrosine kinase implicated in cellular trans- 12. Davis IJ, et al. (2006) Oncogenic MITF dysregulation in clear cell sarcoma: Defining the formation. Oncogene 14:2033–2039. MiT family of human cancers. Cancer Cell 9:473–484. 41. Wimmel A, et al. (1999) Synthesis and secretion of the anticoagulant protein S and 13. Fuse N, Yasumoto K, Suzuki H, Takahashi K, Shibahara S (1996) Identification of a coexpression of the Tyro3 receptor in human lung carcinoma cells. Cancer 86:43–49. melanocyte-type promoter of the microphthalmia-associated transcription factor gene. Biochem Biophys Res Commun 219:702–707. 42. Dhomen N, Marais R (2007) New insight into BRAF mutations in cancer. Curr Opin 14. Wan TS, Chim CS, So CK, Chan LC, Ma SK (1999) Complex variant 15;17 translocations Genet Dev 17:31–39. in acute promyelocytic leukemia. A case report and review of three-way translocations. 43. Thomas NE (2006) BRAF somatic mutations in malignant melanoma and melanocytic Cancer Genet Cytogenet 111:139–143. naevi. Melanoma Res 16:97–103. 15. Taylor IC, Roy S, Yaswen P, Stampfer MR, Varmus HE (1995) Mouse mammary tumors 44. Zhuang D, et al. (2008) C-MYC overexpression is required for continuous suppression express elevated levels of RNA encoding the murine homology of SKY, a putative of oncogene-induced senescence in melanoma cells. Oncogene 27:6623–6634. receptor tyrosine kinase. J Biol Chem 270:6872–6880. 45. Wellbrock C, et al. (2008) Oncogenic BRAF regulates melanoma proliferation through 16. Su AI, et al. (2002) Large-scale analysis of the human and mouse transcriptomes. Proc the lineage specific factor MITF. PLoS ONE 3:e2734. Natl Acad Sci USA 99:4465–4470. 46. Wu M, et al. (2000) c-Kit triggers dual phosphorylations, which couple activation and 17. Lemke G, Rothlin CV (2008) Immunobiology of the TAM receptors. Nat Rev Immunol degradation of the essential melanocyte factor Mi. Genes Dev 14:301–312. 8:327–336. 47. Mark MR, Chen J, Hammonds RG, Sadick M, Godowsk PJ (1996) Characterization of 18. Caraux A, et al. (2006) Natural killer cell differentiation driven by Tyro3 receptor Gas6, a member of the superfamily of G domain-containing proteins, as a ligand for Rse tyrosine kinases. Nat Immunol 7:747–754. and Axl. J Biol Chem 271:9785–9789. 19. Kiefer F, et al. (1998) The Syk protein tyrosine kinase is essential for Fcgamma receptor 48. Nagata K, et al. (1996) Identification of the product of growth arrest-specific gene 6 as signaling in macrophages and neutrophils. Mol Cell Biol 18:4209–4220. a common ligand for Axl, Sky, and Mer receptor tyrosine kinases. J Biol Chem 20. Lu Q, et al. (1999) Tyro-3 family receptors are essential regulators of mammalian 271:30022–30027. spermatogenesis. Nature 398:723–728. 49. Nakamura YS, et al. (1998) Tyro 3 receptor tyrosine kinase and its ligand, Gas6, 21. D’Arcangelo D, Gaetano C, Capogrossi MC (2002) Acidification prevents endothelial cell apoptosis by Axl activation. Circ Res 91:e4–12. stimulate the function of osteoclasts. Stem Cells 16:229–238. 22. Guttridge KL, et al. (2002) Mer receptor tyrosine kinase signaling: Prevention of 50. Ohashi K, et al. (1995) Stimulation of sky receptor tyrosine kinase by the product of apoptosis and alteration of cytoskeletal architecture without stimulation or prolifer- growth arrest-specific gene 6. J Biol Chem 270:22681–22684. ation. J Biol Chem 277:24057–24066. 51. Xiong W, et al. (2008) Gas6 and the Tyro 3 receptor tyrosine kinase subfamily regulate 23. Linger RM, Keating AK, Earp HS, Graham DK (2008) TAM receptor tyrosine kinases: the phagocytic function of Sertoli cells. Reproduction 135:77–87. Biologic functions, signaling, and potential therapeutic targeting in human cancer. 52. McCormack O, et al. (2008) Growth arrest-specific gene 6 expression in human breast Adv Cancer Res 100:35–83. cancer. Br J Cancer 98:1141–1146. 24. Hutterer M, et al. (2008) Axl and growth arrest-specific gene 6 are frequently overex- 53. Stitt TN, et al. (1995) The anticoagulation factor protein S and its relative, Gas6, are pressed in human gliomas and predict poor prognosis in patients with glioblastoma ligands for the Tyro 3/Axl family of receptor tyrosine kinases. Cell 80:661–670. multiforme. Clin Cancer Res 14:130–138. 54. Evenas P, Dahlback B, Garcia de Frutos P (2000) The first laminin G-type domain in the 25. Zhang YX, et al. (2008) AXL is a potential target for therapeutic intervention in breast SHBG-like region of protein S contains residues essential for activation of the receptor cancer progression. Cancer Res 68:1905–1915. tyrosine kinase sky. Biol Chem 381:199–209. 26. Georgescu MM, Kirsch KH, Shishido T, Zong C, Hanafusa H (1999) Biological effects of 55. Nyberg P, He X, Hardig Y, Dahlback B, Garcia de Frutos P (1997) Stimulation of Sky c-Mer receptor tyrosine kinase in hematopoietic cells depend on the Grb2 binding site tyrosine phosphorylation by bovine protein S domains involved in the receptor-ligand in the receptor and activation of NF-kappaB. Mol Cell Biol 19:1171–1181. interaction. Eur J Biochem 246:147–154. 27. Lan Z, et al. (2000) Transforming activity of receptor tyrosine kinase tyro3 is mediated, 56. Godowski PJ, et al. (1995) Reevaluation of the roles of protein S and Gas6 as ligands for at least in part, by the PI3 kinase-signaling pathway. Blood 95:633–638. the receptor tyrosine kinase Rse/Tyro 3. Cell 82:355–358. 28. Toshima J, Ohashi K, Iwashita S, Mizuno K (1995) Autophosphorylation activity and 57. Hafizi S, Dahlback B (2006) Gas6 and protein S. Vitamin K-dependent ligands for the Axl association with Src family kinase of Sky receptor tyrosine kinase. Biochem Biophys Res Commun 209:656–663. receptor tyrosine kinase subfamily. FEBS J 273:5231–5244. 29. Healy AM, et al. (2001) Gas 6 promotes Axl-mediated survival in pulmonary endothelial 58. Udono T, et al. (2000) Structural organization of the human microphthalmia- cells. Am J Physiol Lung Cell Mol Physiol 280:L1273–1281. associated transcription factor gene containing four alternative promoters. Biochim 30. Matsubara N, et al. (1996) A receptor tyrosine kinase, Sky, and its ligand Gas 6 are Biophys Acta 1491:205–219. expressed in gonads and support primordial germ cell growth or survival in culture. Dev 59. Franken NA, Rodermond HM, Stap J, Haveman J, van Bree C (2006) Clonogenic assay of Biol 180:499–510. cells in vitro. Nat Protoc 1:2315–2319.

17030 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0909292106 Zhu et al. Downloaded by guest on September 26, 2021