BRAF mutations result in constitutively active BRAF kinase activity and increased extracellular signal-regulated kinase (ERK) signaling and cell proliferation. Initial studies have shown that BRAF mutations occur at a high frequency in melanocytic nevi and metastatic lesions, but recent data have revealed much lower incidence of these mutations in early-stage melanoma, implying that other factors may contribute to melanoma pathogenesis in a wild-type (WT) BRAF context. To identify such contributing factors, we used microarray gene expression profiling to screen for differences in gene expression between a panel of melanocytic and melanoma cell lines with WT BRAF and a group of melanoma cell lines with the V599E BRAF mutation. We found that SPRY2, an inhibitor homologous to SPRY4, which was previously shown to suppress Ras/ERK signaling via direct binding to Raf-1, had reduced expression in WT BRAF cells. Using small interfering RNA-mediated SPRY2 knockdown, we showed that SPRY2 acts as an inhibitor of ERK signaling in melanocytes and WT BRAF melanoma cells, but not in cell lines with the V599E mutation. We also show that SPRY2 and SPRY4 directly bind WT BRAF but not the V599E and other exon 15 BRAF mutants. These data suggest that SPRY2, an inhibitor of ERK signaling, may be bypassed in melanoma cells either by down-regulation of its expression in WT BRAF cells, or by the presence of the BRAF mutation.
BRAF mutations may play an important role in initiation and/or progression of melanoma (1, 2, 3) via sustained BRAF–mitogen-activated protein kinase kinase–extracellular signal-regulated kinase (BRAF-MEK-ERK) activation (4). As such, they may be important, long searched for therapeutic targets. In melanoma, the majority of BRAF mutations (>90%) involve a T1796A transversion in exon 15, resulting in a V599E missense mutation (2, 5). Rarely in this cancer, exon 11 mutations, such as K438Q, are also described (5). Exon 15 mutations cluster to the kinase domain of BRAF and lock the molecule in its active conformation (6). In general, BRAF mutants lead to increased ERK activation via sustained MEK phosphorylation or interaction with Raf-1 (6), but their exact role with respect to the course and stage of melanoma remains to be determined. The high incidence (up to 80%) in nevi and metastatic lesions (2, 7) but low frequency (10%) in early-stage melanomas suggest that BRAF mutations may correlate with progression rather than initiation of melanoma (3, 8). Other factors may therefore be responsible for tumor initiation in the absence of a BRAF mutation. In the present study, we aimed to identify such factors that may act in a wild-type (WT) BRAF context. Using high-throughput microarray screening, we identified the inhibitor SPRY2 being under-expressed in WT BRAF melanoma cells.
Materials and Methods
Cell Lines and Cell Culture.
Normal melanocytes and primary melanoma cell lines were obtained from M. Herlyn (Wistar Institute, Philadelphia, PA). BRAF mutation status and tumor stage were determined previously (4).3 All cell lines have WT NRAS (4).3 Human melanocytes and melanoma cells were maintained in media as described previously (4). Fetal bovine serum was omitted according to the experimental setting. HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (Sigma, St. Louis, MO) at 37°C and 10% CO2.
Microarray-Based Gene Expression Analysis.
Total RNA was isolated from cell cultures at 70% confluency by use of TRIzol reagent (Invitrogen, Carlsbad, CA), as suggested by the manufacturer. We used 5 μg of total RNA to synthesize double-stranded cDNA (Superscript Choice System for cDNA Synthesis kit; Invitrogen) and after cleanup (GeneChip Sample Cleanup Module; Qiagen/Affymetrix, Santa Clara, CA), to produce fragmented biotin-labeled cRNA (Enzo RNA Transcript Labeling Kit; Enzo, Farmingdale, NY), using the manufacturers’ protocols. Human HG-U133A chips (Affymetrix, Inc.) were hybridized with 15 μg of fragmented labeled cRNA overnight at 45°C, washed (Genechip Fluidics Station 400; Affymetrix), and scanned (GeneArray Scanner; Affymetrix) according to Affymetrix protocols. Scanned images were analyzed with the MAS 5.0 software (Affymetrix), and intensities were scaled to a value of 500. Cluster3.0 was used for visualization purposes (Fig. 1 A).4 The files of the analyzed data and the list of the 108 genes of the Ras/ERK pathway screened for the analysis can be found on the Internet.5
WT and V599E pEF MYC-hBRAF were gifts from R. Marais (Institute of Cancer Research, London, United Kingdom). Point mutations in hBRAF were introduced with Quikchange (Stratagene, La Jolla, CA), according to the manufacturer’s instructions, and verified by capillary sequencing. pCDNA3 FLAG-mSPRY2 and pCDNA3 FLAG-mSPRY4 were supplied by J. Licht and have been described previously (10).
Cell Extraction and Immunoblotting.
Whole-cell extracts were prepared in ELB+ [250 mm NaCl, 50 mm HEPES (pH 7.0), 5 mm EDTA, 10 mm β-glycerol phosphate, 10 mm NaF, 10 mm sodium vanadate, 0.5 mm dithiothreitol, 1 mm phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 0.2% Triton X-100], diluted with Laemmli buffer, and separated by SDS-PAGE. Samples were transferred to nitrocellulose membranes (trans-Blot; BIO-RAD, Hercules, CA), blocked in Tris-buffered saline containing 0.1% (v/v) Tween 20 and 5% (w/v) milk powder, and then probed with murine M2 anti-FLAG (Sigma-Aldrich; St. Louis, MO) and murine 9E10 anti-MYC antibodies (Institute of Cancer Research, London, United Kingdom) in Tris-buffered saline containing 0.1% (v/v) Tween 20 and 5% (w/v) milk powder. Primary antibodies were detected with an Alexa Fluor680-conjugated goat antimouse secondary antibody (Molecular Probes, Eugene, OR) and an Odyssey IR imaging system (Li-Cor Biosciences, Lincoln, NE) according to the manufacturers’ guidelines. For detection of endogenous SPRY2 and BRAF, membranes were probed with rabbit anti-SPRY2 (Upstate, Chicago, IL) and rabbit anti-BRAF (Upstate) antibodies, respectively. The primary antibodies were detected with goat antirabbit horseradish peroxidase-conjugated antibody (Bio-Rad) incubated with horseradish peroxidase substrate (Bio-Rad) and detected by chemiluminescence using X-OMAT film (Kodak, Rochester, NY).
HEK293T cells were transfected with pEF MYC-hBRAF and pCDNA3 FLAG-mSPRY constructs, using Effectene (Qiagen, Chatsworth, CA) according to the manufacturer’s instructions. MYC-hBRAF complexes were immunoprecipitated from HEK293T cell extracts by mixing with rabbit A14 anti-MYC antibody (Santa Cruz Biotechnology, Santa Cruz, CA) bound to protein G-Sepharose beads for 1 h at 4°C. Samples were washed in ELB, eluted with Laemmli buffer, and subjected to SDS-PAGE. Membranes were probed for MYC-hBRAF and the associated FLAG-mSPRY, as described above. FLAG-mSPRY and MYC-hBRAF levels were determined with the Odyssey IR imaging system using Odyssey v1.0 software (Li-Cor Biosciences). Immunoprecipitations were repeated at least three times in independent experiments. The averaged results are presented as the “relative association” of FLAG-mSPRY with MYC-hBRAF (the ratio of FLAG-mSPRY2 signal, normalized for expression levels, to MYC-hBRAF signal detected with the Alexa Fluor680 secondary antibody). For immunoprecipitation of endogenous BRAF, cell extracts were passed through protein G-Sepharose columns (Sigma) bound to anti-SPRY2 antibody (Upstate). Samples were washed in ELB and eluted with Laemmli buffer. After SDS-PAGE and transfer, membranes were probed with anti-BRAF antibody (Upstate).
Small Inhibitory RNA Transfections.
Small inhibitory duplex RNAs (siRNA) were prepared from DNA templates by use of the Silencer siRNA construction kit (Ambion, Austin, TX), according to the manufacturer’s protocol. Sequences are available on request. To achieve transient knockout of SPRY2, cells were plated in 6-well plates at 30–50% confluency and transfected with 50 nm siRNA and 4 μl of oligofectamine reagent (Invitrogen Life Technologies, Inc., Gaithersburg, MD) and OptiMEM (Invitrogen), as recommended by the manufacturer. After 4 h at 37°C, the transfection medium was removed and replaced with MCDB153. Cells were harvested at post-transfection time points as indicated.
Incorporation assays were performed using a colorimetric bromodeoxyuridine (BrdUrd) assay (Roche) according to the manufacturer’s instructions. Cells were cultured in 96-well plates at 0.5 × 104 cells/well and transfected with either control or SPRY2-specific siRNA, as described above. U0126 (Promega, Madison, WI) was applied at 10 μm for 24 h at the 48-h post-transfection time point. Cells were incubated for 5 h with 10 μm BrdUrd 48 h after siRNA transfection, fixed with FixDenat, incubated with anti-BrdUrd peroxidase antibody, washed, and incubated with 3,3′,5,5′-tetramethylbenzidine solution provided by the manufacturer. Absorbance was measured at 370 nm (reference wavelength, 492 nm). Maximum sensitivity was achieved at an absorbance range of 0.5–2.5, with sensitivity comparable to the [3H]thymidine incorporation assay. Within this range, 2-fold increases in absorbance correspond to a 5–10-fold increase in the number of proliferating cells. Wells without cells (to control for nonspecific binding to the plate) and background controls (cells incubated with anti-BrdUrd antibody) were included to control for nonspecific labeling.
Results and Discussion
We used oligonucleotide-based expression profiling (Affymetrix, Inc.) to identify gene expression changes between melanocytes and melanoma cell lines with WT BRAF and melanoma cells with mutant BRAF. We focused our initial analysis on changes in expression of components of the RAS/RAF/MEK/mitogen-activated protein kinase (MAPK) signaling pathway. Specifically, 108 transcripts belonging to this pathway were screened for changes in gene expression (see “Materials and Methods”). This analysis revealed that only SPRY2 transcripts were down-regulated in melanocytes and WT BRAF melanoma cell lines compared with mutant BRAF cultures (Fig. 1), with a statistically significant change in average intensity on the order of 2-fold (P < 0.05). Quantitative reverse transcription-PCR for SPRY2 in selected cell lines confirmed the microarray results (data not shown). SPRY2 has been identified as inhibitor of MAPK signaling in epithelial and fibroblast cell lines (9, 10). SPRY was originally identified as encoding a fibroblast growth factor inhibitor that regulates tracheal branching in Drosophila (11). There are four human and three murine SPRY genes with distinct tissue expression patterns (12, 13). Present evidence suggests that SPRY proteins may have a role in the regulation of angiogenesis as well as kidney, lung, and limb development (12, 13, 14). SPRY gene expression is induced as a result of increased levels of phosphorylated (activated) ERK, creating a negative feedback loop (9, 15). The low levels of SPRY2 in the melanocytes used in this study are consistent with the lower phospho-ERK levels that these cells show compared with melanoma cell lines (4). Melanoma cells that harbor the V599E mutation have constitutively activated ERK (4), which may account for the high levels of SPRY2 found. Under the culture conditions used (plus serum), most of the WT BRAF melanoma cells used in this study exhibited phospho-ERK levels comparable to those in the mutant cells (4). In this case, low SPRY2 levels may represent a disrupted responsiveness of SPRY2 to phospho-ERK in these cells. This notion is further supported by the fact that a decrease in phospho-ERK after incubation with a MEK inhibitor (U0126) down-regulated SPRY2 expression in mutant BRAF melanoma cells to levels similar to the ones found in WT BRAF cells (data not shown).
SPRY proteins may act as competitive inhibitors of MAPK signaling by at least two potential mechanisms. The COOH termini of SPRY2 and SPRY4 can bind to RAF1, inhibiting kinase activity (16). In addition, SPRY proteins uncouple receptor tyrosine kinase signaling from the activation of RAS, possibly by sequestration of the GRB2, an adaptor protein required for RAS activation (9, 10). As the inhibitory function of SPRY2 in MAPK signaling is thought to be cell-context-dependent (9, 17), we asked whether SPRY2 has an inhibitory effect in melanocytes and/or melanoma cell lines. We therefore designed siRNAs complementary to SPRY2 and tested their ability to reduce SPRY2 transcript and protein levels, regulate ERK phosphorylation, and alter cell growth characteristics. To this end, SPRY2-specific siRNA (or scrambled-sequence siRNA for control) was transfected into normal melanocytes and melanoma cell lines by use of Oligofectamine (Invitrogen). Whole-cell lysates were prepared at 24-h intervals (0–96 h). Maximum reduction in SPRY2 protein levels was achieved at the 48- and 72-h time points (data not shown). At 48 h post-transfection, normal melanocytes assumed a morphology previously described as characteristic of increased levels of phospho-ERK (Ref. 18; Fig. 2 A), whereas the phenotype of the melanoma cell lines did not change (data not shown).
To determine whether the siRNA-mediated SPRY2 knockout had an effect on ERK activation, we prepared whole-cell extracts from SPRY2 siRNA-transfected and control sequence siRNA-transfected melanocytes 48 h post-transfection. These experiments demonstrated that the observed morphological changes were correlated with ERK phosphorylation (Fig. 2,B). Both the phenotypic change (data not shown) and the increase in ERK phosphorylation (Fig. 2 B) seen after transfection with SPRY2 siRNA were reversible after exposure to the MEK inhibitor U0126, indicating the dependence of the phenotype on ERK activation. These data demonstrate that normal melanocytes respond to SPRY2 down-regulation by increasing ERK phosphorylation and that SPRY2 functions as an inhibitor of MAPK signaling in normal melanocytes.
We also measured phospho-ERK levels in melanoma cell lines, with and without a BRAF mutation (V599E), transfected with SPRY2 siRNA (Fig. 2,C). These experiments demonstrated that knock-down of SPRY2 protein is associated with increased phospho-ERK in melanoma cell lines with WT BRAF (Fig. 2,C) compared with control transfections with scrambled SPRY2 siRNA. In contrast, phospho-ERK levels remained unchanged after SPRY2 siRNA transfection in melanoma cell lines with mutant BRAF (V599E; Fig. 2 C).
To determine whether SPRY-induced modulation of MAPK signaling plays a role in regulation of cell growth, we used BrdUrd incorporation as a surrogate for cellular proliferation. As predicted by alterations in phospho-ERK levels, transfection with SPRY2 siRNA was associated with significantly increased BrdUrd incorporation (P < 0.0015), whereas incorporation was unaffected by SPRY2 siRNA transfection in melanoma cell lines with mutant BRAF (Fig. 2,D). Treatment with the MEK inhibitor U0126 showed that a functional RAF/MEK/ERK signaling cascade is required for cell proliferation in these cells, as captured by BrdUrd incorporation (Fig. 2 D).
Previous studies showed that SPRY4 binding to the COOH terminus of RAF1 is required for inhibition of ERK phosphorylation after growth factor stimulation (16). In addition, SPRY2 and SPRY4 may act upstream of Ras, interfering with Grb2-Sos complex formation (9). Because Ras is constitutively activated in melanoma cells (4), we hypothesized that SPRY2 may inhibit the ERK pathway downstream of Ras (at the level of Raf), rather than upstream. Thus, we asked whether SPRY2 and/or SPRY4 could bind to BRAF, and whether, if so, this interaction is altered by the presence of a BRAF mutation. Endogenous BRAF co-immunoprecipitated with SPRY2 from whole-cell lysates of FOM74 melanocytes that expressed WT BRAF, but not from lysates of 1205Lu melanoma cells that expressed mutant V599E BRAF (Fig. 3,A). To determine whether various tumor-derived BRAF mutations affected binding to SPRY2 and SPRY4, we coexpressed FLAG-tagged SPRY2 or FLAG-tagged SPRY4 and MYC-tagged mutant or WT BRAF in HEK293T cells. These experiments showed that exon 15 BRAF mutants (V599E, V599D, L596V, and K600E) did not bind SPRY2 or SPRY4, whereas WT BRAF and nonactivating exon 11 mutants (K438Q, T439P) efficiently bound both SPRYs (Fig. 3, B–D).
These results suggest that in the case of melanoma cell lines with WT BRAF, SPRY2-dependent inhibition may be mediated by direct interaction of SPRY2 with BRAF. In V599E mutant melanoma cells, SPRY2 does not show any inhibitory effect. This observation can be explained either by the mutation-mediated disruption of SPRY2-BRAF association, or it can simply be a result of the overwhelming increase in kinase activity of BRAF exon 15 mutants.
In summary, we have shown that SPRY2 is underexpressed in melanocytes and WT BRAF melanoma cell lines compared with V599E melanoma cells. Knocking down SPRY2 in these cells has a positive impact on ERK signaling only in a WT BRAF context. In addition, SPRY2 is capable of binding BRAF, but the interaction is disrupted by exon 15 mutations in BRAF. These findings suggest that loss of SPRY expression may enhance levels of active ERK, enabling the development and growth of melanoma cells of a WT BRAF profile. Further studies should provide insight into the role of the SPRY family in these processes, with the potential benefit of optimizing BRAF as a therapeutic target in melanoma.
Grant support: Abramson Family Cancer Research Institute, University of Pennsylvania Abramson Cancer Center, Pennsylvania Department of Health, and NIH (Grant CA59998 to J. Licht and Grant CA030721-01A1 to M. Olson).
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.
Requests for reprints: Dimitra Tsavachidou, Abramson Family Cancer Research Institute, University of Pennsylvania, Philadelphia, PA 19104. E-mail: firstname.lastname@example.org
B. L. Weber and M. Herlyn, unpublished data.
We thank Dr. Meenhard Herlyn (Wistar Institute, Philadelphia, PA) for the melanoma cell lines and helpful discussions.