Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate

Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate

G Model ARTICLE IN PRESS YCECA-1787; No. of Pages 12 Cell Calcium xxx (2016) xxx–xxx Contents lists available at ScienceDirect Cell Calcium journ...

4MB Sizes 1 Downloads 55 Views

G Model

ARTICLE IN PRESS

YCECA-1787; No. of Pages 12

Cell Calcium xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Cell Calcium journal homepage: www.elsevier.com/locate/ceca

Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate Stacey Ann N. D’mello a,b , Wayne R. Joseph b , Taryn N. Green a , Euphemia Y. Leung b , Matthew J. During c , Graeme J. Finlay a,b , Bruce C. Baguley b , Maggie L. Kalev-Zylinska a,d,∗ a

Department of Molecular Medicine and Pathology, University of Auckland, Private Bag 92019, Auckland, New Zealand Auckland Cancer Society Research Centre, University of Auckland, Auckland, Private Bag 92019, Auckland, New Zealand c Cancer Genetics and Neuroscience Program, Department of Molecular Virology, Immunology and Medical Genetics, The Ohio State University, Columbus, OH 43210, United States d LabPlus Haematology, Auckland City Hospital, Private Bag 92024, Auckland, New Zealand b

a r t i c l e

i n f o

Article history: Received 15 July 2016 Received in revised form 29 August 2016 Accepted 13 September 2016 Available online xxx Keywords: Calcium Cancer Ion channel Melanoma NMDA receptor Oncogene Tumor suppressor

a b s t r a c t GRIN2A mutations are frequent in melanoma tumours but their role in disease is not well understood. GRIN2A encodes a modulatory subunit of the N-methyl-d-aspartate receptor (NMDAR). We hypothesized that certain GRIN2A mutations increase NMDAR function and support melanoma growth through oncogenic effects. This hypothesis was tested using 19 low-passage melanoma cell lines, four of which carried novel missense mutations in GRIN2A that we previously reported. We examined NMDAR expression, function of a calcium ion (Ca2+ ) channel and its contribution to cell growth using pharmacological modulators; findings were correlated with the presence or absence of GRIN2A mutations. We found that NMDAR expression was low in all melanoma cell lines, independent of GRIN2A mutations. In keeping with this, NMDAR-mediated Ca2+ influx and its contribution to cell proliferation were weak in most cell lines. However, certain GRIN2A mutations and culture media with lower glutamate levels enhanced NMDAR effects on cell growth and invasion. The main finding was that G762E was associated with higher glutamate-mediated Ca2+ influx and stronger NMDAR contribution to cell proliferation, compared with wild-type GRIN2A and other GRIN2A mutations. The pro-invasive phenotype of mutated cell lines was increased in culture medium containing less glutamate, implying environmental modulation of mutation effects. In conclusion, NMDAR ion channel function is low in cultured melanoma cells but supports cell proliferation and invasion. Selected GRIN2A mutations, such as G762E, are associated with oncogenic consequences that can be modulated by extracellular glutamate. Primary cultures may be better suited to determine the role of the NMDAR in melanoma in vivo. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction The high frequency of heterogeneous GRIN2A mutations in melanoma invites research into their roles in disease [1–3]. The main aim is to determine therapeutic opportunities that typically are greater for oncogenic targets. The current view is that GRIN2A mutations in melanoma induce loss of tumor suppressor function of wild-type GRIN2A. This is based on the observation that GRIN2A knock-down increases proliferation of melanoma cells that

Abbreviations: NMDAR, N-methyl-d-aspartate receptor; NZM, New Zealand Melanoma cell lines. ∗ Corresponding author at: Department of Molecular Medicine and Pathology, University of Auckland, ACM1142, Auckland, New Zealand. E-mail address: [email protected] (M.L. Kalev-Zylinska).

carry wild-type but not mutant GRIN2A [4]. However, this loss-offunction effect has so far been shown for a relatively small number of GRIN2A mutations, most of which were non-sense. Other GRIN2A mutations may have different effects and their examination is warranted. GRIN2A encodes the GluN2A subunit of the N-methyl-daspartate receptor (NMDAR) [5,6]. Typical NMDARs combine two GluN1 subunits with another two of four possible GluN2 components (designated GluN2A to GluN2D). Glutamate is the main NMDAR ligand that binds to the GluN2 subunits; glycine is a co-ligand that binds to GluN1 [6]. In response to the binding of glutamate and glycine, NMDARs facilitate intracellular influx of predominantly calcium ions (Ca2+ ) [7]. Not much is known about NMDARs in melanoma but the presence of GluN1-GluN2A complexes and NMDAR-mediated Ca2+ fluxes have been shown in

http://dx.doi.org/10.1016/j.ceca.2016.09.003 0143-4160/© 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

2

melanoma cell lines [4,8]. The role of NMDAR as a tumor suppressor is thought to engage pro-apoptotic signaling that becomes restrained by loss-of-function mutations in melanoma [4]. However, there is also evidence that active NMDARs facilitate survival, proliferation and invasion of melanoma cells, more in keeping with the pro-oncogenic role for the NMDAR [8–10]. Such opposing NMDAR effects are intriguing but not surprising, as NMDARs are known for their dichotomous effects on cell death and survival [11]. New Zealand has the highest rate of invasive melanoma in the world [12,13]. To help determine the cause, our institution established a unique collection of cell lines grown from metastatic melanoma tumors excised from local patients. We used 19 of these cell lines before and discovered four novel missense mutations in GRIN2A [3]. Significantly, patients from whom the mutated cell lines were derived had shorter overall survival, compared with patients with wild-type GRIN2A. This important clinical association motivated us to examine functional consequences of GRIN2A mutations in our cell lines; findings are presented in this report. We hypothesized that certain GRIN2A mutations increase NMDAR function, which enhances tumor growth. This hypothesis was examined using well-established small molecule modulators of the NMDAR. Results will demonstrate that consequences of GRIN2A mutations in melanoma are heterogeneous but some contribute growth-promoting effects that can be modulated by extracellular glutamate.

CA) was used as a positive control. Concentration and integrity of RNA were tested on a NanoDrop 2000 (ThermoFisher Scientific). To synthesize cDNA, 2 ␮g RNA was reverse transcribed using a Superscript III First-Strand synthesis system with oligo(dT) and random primers (all from Life Technologies). Primer sequences to detect NMDAR transcripts and PCR conditions were essentially as before (Supplemental Table S1) [16,17].

2. Materials and methods

2.4. Recordings of intracellular Ca2+ responses

2.1. Primary cells and cell lines

Cells were plated at a density of 50 × 103 cells mL−1 on black clear-bottom 96-well plates and incubated in supplemented ␣-MEM overnight. Cytosolic Ca2+ was visualized using Fluo-4AM (acetoxymethyl ester) and a Fluo-4-NW Calcium Assay kit (Thermo-Fisher Scientific). Prior to loading, culture medium was removed by washing cells three times with the kit buffer consisting of Hanks Balanced Salt Solution with 20 mM HEPES. Cells were loaded with 5 ␮M Fluo-4-AM in the presence of 0.06% pluronic acid and 2.5 mM probenecid over 15 min at 37 ◦ C in the dark, then washed, incubated in buffer for another 15 min at room temperature to complete de-esterification, and washed again. The imaging buffer contained 2.3 mM CaCl2 and 1% FBS. L-glutamic acid (glutamate) and NMDA (both at 50, 100 and 200 ␮M; Sigma-Aldrich) were applied to determine if NMDARs contribute Ca2+ fluxes in melanoma cells. Ionomycin (1 ␮g mL−1 ; Sigma-Aldrich) was used as a positive control and the imaging buffer as a negative control. The fluorescence signal was read from the bottom of the plate at 1 s intervals (3 wells per s) using an EnSpire 2300 Multimode Plate Reader (Perkin-Elmer, Waltham, MA) with excitation of 494 nm and emission 506 nm (25 ◦ C, 3 mm measurement height, 50 flushes). Baseline fluorescence was recorded for 10 s, after which the activator or buffer were added and imaging continued for a further 90 s. Fold changes in Ca2+ levels were calculated relative to a buffer control using average fluorescence values recorded at baseline and between 20 and 30 s after the addition of modulators, when Ca2+ responses were maximal.

All work on patient samples has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) and was approved by Northern A Health and Disability Human Ethics Committee. This study used 19 New Zealand Melanoma (NZM) cell lines previously generated in our institution from metastatic melanoma tumors excised from patients undergoing diagnostic or treatment procedures [14]. Briefly, tumor samples were disaggregated with a scalpel, finely minced and passed through a sieve using a metal probe. Disaggregated cells were collected in ␣-modified Minimal Essential Medium (␣-MEM; Sigma-Aldrich, Saint Louis, MO), centrifuged at 400 rpm for 1 min to remove debris, then seeded in culture flasks, fed weekly and passaged when confluent. Cell line numbers (NZM 1 to NZM 100) follow the order in which they were established. The number of passages was less than 30 for all cell lines. NZM cells were maintained in ␣-MEM supplemented with the following additives: 5% fetal bovine serum (FBS; Life Technologies, Carlsbad, CA); penicillin and streptomycin (both at 100 Units mL−1 ; Sigma-Aldrich); insulin and transferrin (both at 5 ␮g mL−1 ); sodium selenite (5 ng mL−1 ; all from Roche Applied Science, San Diego, CA). In selected experiments, cells were cultured in RPMI-1640 medium (Thermo-Fisher Scientific, Waltham, MA), supplemented as above. Normal human melanocytes (HEMa-LP cells) were cultured in M-254-500 medium with Human Melanocyte Growth Supplement (all from Life Technologies). Melanoma and melanocyte cultures were maintained at 37 ◦ C in a humidified incubator supplied with 5% O2 – 5% CO2 in nitrogen. Fetal rat hippocampal neurons were derived from fetuses of Wistar rats and cultured as described before [15]. 2.2. RNA isolation, cDNA synthesis and reverse transcription (RT) PCR Total RNA was extracted from cells using TRIzol (Life Technologies). Total RNA from human cerebellum (Clontech, Mountain View,

2.3. Western blotting Proteins were extracted from cells using a radioimmunoprecipitation assay buffer and quantified using a bicinchoninic acid kit (Life Technologies). Samples containing 40 ␮g proteins were separated on 5–14% Mini-PROTEAN precast polyacrylamide gels and transferred to nitrocellulose membranes using a semi-dry transfer apparatus (all from Bio-Rad, Hercules, CA). Membranes were blocked with 5% weight per volume skim milk powder in TBSTween20 and incubated with the following primary antibodies overnight: anti-NMDAR1 (D65B7), anti-Slug (C19G7; both from Cell Signalling, Beverly, MA), anti-GluN2A (Ab133265; Life Sciences, Cambridge, MA) and anti-Actin (MAB1501; Millipore, Billerica, CA). Membranes were washed and incubated with appropriate secondary antibodies conjugated to horseradish peroxidase (HRP; Santa Cruz Biotechnology, Dallas, TX). The HRP signal was developed using Clarity Western ECL (Bio-Rad) and imaged on Fujifilm LAS 3000 (Biocompare, San Francisco, CA).

2.5. Cell proliferation assay Cell proliferation was measured from the incorporation of tritiated [3 H]thymidine as before [14]. The following NMDAR antagonists were used: (+)-MK-801 hydrogen maleate (MK801), 3,5-dimethyl-1-adamantanamine hydrochloride (memantine), D(−)-2-amino-5-phosphonopentanoic acid (AP5) and 2amino-6-(trifluoromethoxy)benzothiazole (riluzole; 10–200 ␮M for all). MK-801 and memantine block open NMDAR channels and AP5 competes with glutamate for its binding sites on the GluN2 sub-

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

units [6,18]. Glutamate and NMDA were applied as NMDAR agonists (25–200 ␮M). NZM cells were cultured in the presence of modulators for five days, and primary cultures of melanoma cells for 7 days. DNA synthesis was measured using 10 nM thymidine deoxyd-ribofuranosyl, 10 nM fluoro-deoxyuridine and 0.25 mCi ␮L−1 [3 H]thymidine. At the end of cultures, cells were harvested onto a membrane using a TomTec Harvester (Life Sciences, Cherstertown, ML), dried overnight and sealed in a plastic pocket containing scintillation fluid. Radioactivity of incorporated [3 H]thymidine was measured on a Wallac Microbeta 1450-021 TriLux Luminometer Liquid Scintillation Counter (LabEquip, Markham, Canada). The effect of NMDAR modulation was determined from the modulator concentrations required to inhibit cell proliferation by 50% (IC50 values). 2.6. Cell invasion assay The invasion assay was conducted using a BD BiocoatTM MatrigelTM Invasion system (BD Biosciences). The kit contains a 24-well plate with trans-well inserts with porous (8 ␮m) membranes coated with Matrigel. Cells were seeded on the apical side of membranes at 80 × 103 cells mL−1 in FBS-free ␣-MEM. To stimulate invasion, 10% FBS was added to the basal side of membranes. Cells were settled over 30 min at 37 ◦ C in an atmosphere containing 5% O2 − 5% CO2 . NMDAR modulators were added on both sides of membranes to 100 ␮M final concentration. Twenty hours later, membranes were taken out of the plate and swabbed to remove Matrigel. Cells that migrated to the basal side of membranes were fixed in methanol and stained with Wright-Giemsa staining. Cells were imaged using a TE2000E inverted light microscope equipped with a Plan Fluor 10×/0.45 NA objective lens and Digital Sight color camera (both from Nikon, Tallahassee, FL). Four images were taken per membrane capturing most its surface. Image files were separated into red, green and blue channels; cell confluency was analyzed on the green channel using ImageJ software (version 1.49). 2.7. Measurements of glutamate concentrations Cells were cultured for 9 days in 6-well plates in either ␣-MEM or RPMI-1640 supplemented with 5% FBS in an atmosphere containing 5% O2 − 5% CO2 . NZM3, NZM6, NZM7 and NZM40 cell lines were seeded at 5 × 103 cells mL−1 ; NZM11, NZM61 and NZM100 at 10 × 103 cells mL−1 . Seeding densities were determined in prior experiments to produce 90–100% confluency after 9 days in culture. Media samples were collected on days 0 and 9, and stored at −80 ◦ C until testing. Glutamate concentrations were measured using an Amplex® Red Glutamic Acid/Glutamate Oxidase Assay kit (ThermoFisher Scientific) according to manufacturer’s instructions with minor modifications. The recommended reaction volume was 100 ␮L in a black 96-well plate but we used 50 ␮L in a black 384-well plate with appropriate positive and negative controls as recommended. The samples of medium were diluted 40× before being assayed. The fluorescence signal was read from the top of the plate on an EnSpire 2300 Multimode Plate Reader (Perkin-Elmer) at excitation of 571 nm and emission 585 nm (25 ◦ C). 2.8. Statistical analysis Data are expressed as mean ± standard error of the mean (SEM). Statistical analysis was conducted using GraphPad Prism 5.0 software for Windows (San Diego, CA). Differences between groups were examined using an unpaired Student t-test (2-tailed), one-way analysis of variance (ANOVA) with Dunnett’s post-hoc for continuous variables or chi-square analysis for dichotomous variables. P values less than 0.05 were considered statistically significant.

3

Table 1 NZM cell lines with non-synonymous mutations in GRIN2A. NZM cell line

Nucleotide substitution and position number

Amino acid change

Exon

NZM61

C>T c.1046 g.291693 G>A c.2285 g.384407 G>A c.2666 g.417877 C>T c.3395 g.418606 C>T c.3397 g.418608

p.S349F

5

p.G762E

12

p.G889E

14

p.P1132L

14

p.P1133S

14

NZM61

NZM7

NZM100

NZM3

Mutations are listed in the order of location along the GRIN2A sequence (NCBI accession number NM 000833). All mutations were heterozygous and reported before [3]. Abbreviations: c = cDNA; g = genomic DNA; p = protein.

3. Results 3.1. GRIN2A mutations affect Ca2+ responses in melanoma cells Nine NZM cell lines, including four with missense mutations in GRIN2A (Table 1) were examined for expression of NMDAR transcripts, termed GRIN1 and GRIN2 (A to D; Fig. 1A; Supplemental Table S2). RNA from human cerebellum and normal melanocytes were used as controls. We found that transcripts of GRIN1, GRIN2A and GRIN2D were present in all cell lines and normal melanocytes. In comparison, signals for GRIN2B and GRIN2C were very weak and documented only in the minority of cell lines. GRIN2D transcripts were dominant in most melanoma lines, except for NZM3 and NZM40 that showed stronger expression of GRIN2A. The overall pattern of GRIN genes expression was virtually identical in melanoma cells and normal melanocytes but differed from that in the brain, where GRIN2A and GRIN2B transcripts dominated over GRIN2D. These differences suggested that melanoma NMDARs are likely to be distinct from their neuronal counterparts. However, the presence of GRIN2A mutations was not associated with any obvious change in GRIN genes expression at the transcript level. The GluN1 and GluN2A proteins were difficult to detect using traditional Western blotting. Most signals were weak and insufficient for quantitation. Nevertheless, the GluN1 protein was detected in all melanoma cell lines, except for NZM1 that carries wild-type GRIN2A (Fig. 1B). GluN1 bands were of slightly larger size in melanoma cells than in the brain but this was found to be similar to the pattern detected in HCT116 and HepG2, two other non-neuronal cancer cell lines (colorectal and hepatocellular, respectively) that were previously shown to express GluN1 [19] (Fig. 1C). As GluN1 is the obligate NMDAR component, its wide expression implied that most, if not all NZM cell lines had the ability to form functional NMDARs. Western blot signals for GluN2A were even weaker. Most distinct bands were detected for the NZM40 cell line that carries wild-type GRIN2A (Fig. 1D). Weak signals were also detected in early but not later passages of NZM1 (also with wildtype GRIN2A) and in normal melanocytes, but not in cell lines with GRIN2A mutations (Fig. 1D,E). A fluorimetric Ca2+ flux assay was used to examine if melanoma NMDARs were functional. Cells were washed and loaded with Fluo4-AM. Intracytoplasmic Ca2+ levels were monitored in response to 50 ␮M NMDA and 100 ␮M glutamate applied to cells in a buffer containing 2.3 mM CaCl2 and 1% FBS (Fig. 2). Glutamate is a

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12 4

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

Fig. 1. Expression of the NMDAR transcripts and proteins in selected NZM cell lines. (A) Conventional RT-PCR to detect GRIN1 and GRIN2 transcripts in nine NZM cell lines, human cerebellar neurons and normal human melanocytes (HEMa-LP cells). NZM cell lines have been numbered and GRIN2A mutations are indicated for positions of the predicted amino acid changes. Transcripts of ␤-actin were used as a positive control. The identity of GRIN1 and GRIN2A transcripts was confirmed by Sanger sequencing. (B–E) Representative Western blots demonstrating expression of the GluN1 (B and C) and GluN2A (D and E) proteins in melanoma cell lines and selected controls as indicated. All tests were performed independently at least twice. Abbreviations: HCT116, human colorectal cancer cell line; HEMa-LP, normal human melanocytes; HepG2, human hepatocellular cancer cell line; Neg, negative (no template) control; NZM, New Zealand Melanoma cell lines; p4 and p7 refer to the passage numbers 4 and 7, respectively; # effects of G762E were previously modeled in silico; cells also harbored S349F [3].

major physiological NMDAR agonist, although not NMDAR-specific. NMDA is a synthetic and weaker but NMDAR-specific agonist [6]; we used both to corroborate findings. Compared with robust Ca2+ responses in cultured fetal neurons, recordings from melanoma cells were weak (Fig. 2A–F). Nevertheless, NMDAR-mediated Ca2+ influx was detected in cells carrying wild-type GRIN2A, consistent with the presence of functional NMDARs (Fig. 2B; Supplemental Table S3). GRIN2A mutations were associated with variable Ca2+ responses. In the presence of G889E, Ca2+ influx was not reliably detected (Fig. 2C), which was in keeping with previous data that G889E inhibits NMDAR assembly [4]. Unexpectedly, other mutations did not abolish Ca2+ responsiveness in melanoma cells. In the

presence of P1133S, peak Ca2+ responses were similar to controls (Fig. 2D), or higher for G762E and P1132L (Fig. 2E,F), arguing against the loss-of-function effect for the latter mutations. 3.2. Cells carrying G762E are more sensitive to inhibition by MK-801 Encouraged by the evidence of NMDAR functionality in the presence of both wild-type and mutated GRIN2A, we proceeded to examine functional NMDAR effects in all 19 NZM cell lines. Cells were cultured for five days in the presence of well-established small molecule NMDAR antagonists (MK-801, memantine, AP5; all

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

5

Fig. 2. Peak Ca2+ responses in melanoma cells to the NMDAR agonists. Cytosolic Ca2+ levels (indicated by fluorescence units) recorded from rat fetal neurons (A) and melanoma cell lines as indicated (B–F) in response to 50 ␮M NMDA, 200 ␮M glutamate, 1 ␮g/mL ionomycin (positive control) and a negative buffer control containing 2.3 mM extracellular Ca2+ and 1% FBS. Line graphs show mean relative levels of Ca2+ over 100 s calculated from triplicate wells. Peak Ca2+ responses shown were detected from at least three independent experiments for each NZM cell line (fetal neurons were tested twice). Corresponding bar graphs show a fold change in Ca2+ levels relative to the buffer control (mean ± SEM) calculated from average fluorescence values measured between 20 s to 30 s time points. Statistical significance is indicated (one-way ANOVA with Dunnett post-hoc; *p < 0.05, **p < 0.01, ***p < 0.001). Abbreviations: glut, glutamate; ns, non-significant; NZM, New Zealand Melanoma cell lines.

at 10–200 ␮M) and effects on cell proliferation were examined from [3 H]thymidine incorporation. Both MK-801 and memantine inhibited cell proliferation (Fig. 3A–F) but AP5 did not (data not shown). Cells that harboured GRIN2A mutations, in particular G762E were more sensitive to inhibition by MK-801 than cells with wild-type GRIN2A (Fig. 3A,B; p = 0.03). In addition, the presence of G762E was

associated with higher sensitivity to MK-801 than G889E or P1133S (Fig. 3A,E), which was in keeping with the pattern of Ca2+ fluxes recorded from these cells (Fig. 2). Memantine inhibited melanoma cell proliferation more potently than MK-801 but with no dependency on GRIN2A mutations (Fig. 3C,D). Nevertheless, cells carrying G762E were more

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12 6

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

Fig. 3. Anti-proliferative effects of MK-801 and memantine in melanoma cells. IC50 values (mean ± SEM) for MK-801 (A, B) and memantine (C, D) that inhibited proliferation of melanoma cell lines. A and B show data for individual cell lines; box plots in B and D show cumulative data for GRIN2A mutated and wild-type cell lines. Each cell line was tested independently three times. Statistical significance is shown (Student’s t-test). (E, F) Representative patterns of cell proliferation in the presence of MK-801 and memantine, respectively. Each point represents mean ± SEM of an experiment performed in triplicate. Response curves for cells carrying G762E and P1133S are highlighted for comparison; cells carrying G889E were not tested at this time. (G) Mean IC50 values of memantine that inhibited proliferation of primary melanoma cells from one experiment performed in duplicate.

sensitive to memantine than cells with P1133S, which was similar to MK-801 (Fig. 3E,F). Intriguingly, two of five melanoma cells in primary cultures responded better to memantine than any of the 19 NZM cell lines, suggesting that culture conditions select for lower NMDAR activity (Fig. 3G).

3.3. Cells carrying G762E are more sensitive to stimulation by glutamate To examine contribution of glutamate to melanoma cell growth, we first used riluzole that inhibits glutamate release [20,21]; its effects on cell proliferation were tested as above. Riluzole reduced

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

7

Fig. 4. Effects of riluzole, NMDA and glutamate on the proliferation of melanoma cells. (A, B) IC50 values (mean ± SEM) for riluzole that inhibited proliferation of melanoma cell lines. A shows data for individual cell lines; box plots in B compare cumulative results for GRIN2A mutated and wild-type cell lines. Each cell line was tested independently three times. Statistical significance is shown (Student’s t-test). (C, D, E) Representative patterns of cell proliferation in the presence of riluzole, NMDA and glutamate, respectively. Each point represents mean ± SEM of an experiment performed in triplicate. Response curves for cell lines carrying G762E and P1133S are highlighted for comparison; cells carrying G889E were not tested at this time.

proliferation of melanoma cell lines with no obvious dependency on GRIN2A mutations (Fig. 4A,B; p = 0.14). Nevertheless, cells that harboured G762E were harder to suppress by riluzole than a number of other cell lines, including those with other GRIN2A mutations (Fig. 4A,C). When NMDA or glutamate were added to medium, proliferation of cells increased, implying NMDAR reactivity in culture (Fig. 4D,E). Cells carrying G762E were more responsive to NMDA and glutamate than cells with P1133S, consistent with their higher sensitivity to glutamate. To examine if melanoma cells release glutamate, cells were cultured in media that varied in glutamate content, either RPMI1640 containing 136 ␮M glutamate or ␣-MEM containing 510 ␮M glutamate (manufacturers’ specifications were confirmed; Fig. 5A,

Supplemental Table S4). After 9 days in culture, glutamate concentrations increased for all cell lines, except for NZM3 and NZM6, consistent with glutamate release (Fig. 5A). The released glutamate amplified glutamate concentrations in RPMI-1640 (Fig. 5B). However, the presence of GRIN2A mutations had no impact on the amount of glutamate released into media (p = 0.81). 3.4. Effects of GRIN2A mutations are environmentally modulated Eight melanoma cell lines were tested for their ability to invade in the trans-well system: four with GRIN2A mutations and four without (NZM6, NZM9, NZM11, NZM40; these were selected randomly). Unexpectedly, only the NZM40 cell line (with wild-type

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12 8

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

Fig. 5. Evidence for glutamate release by melanoma cells. (A) Concentrations of glutamate on day 9 of cultures for cells grown in two types of media that differed in glutamate concentrations (high in ␣-MEM and low in RPMI-1640). Each bar represents mean ± SEM from three separate experiments for which each sample was tested in triplicate. Horizontal lines have been drawn to indicate reference levels for media only. (B) Box plots demonstrate relative amounts of glutamate released into media with high (␣-MEM) and low (RPMI-1640) concentrations of glutamate. Statistical significance is shown (Student’s t-test in B).

Fig. 6. Effects of NMDAR inhibitors on the invasiveness of melanoma cell lines. (A) The invasiveness of NZM40 cells determined from the percent of area occupied by cells in the presence and absence of the NMDAR inhibitors. Each bar represents mean ± SEM from three separate experiments for which samples were tested in triplicate. Statistical significance is shown (one-way ANOVA with Dunnett post-hoc; *p < 0.05 **p < 0.01). (B) Representative images of the basal side of membranes showing NZM40 cells that invaded through Matrigel.

GRIN2A) was found to be invasive in the trans-well assay. MK-801 and AP5 inhibited invasion of NZM40 cells, indicating that the invasive phenotype was mediated at least in part through the NMDAR function (Fig. 6; p < 0.05). However, none of the mutated cell lines invaded through Matrigel (data not shown). Because all NZM cell lines were derived from metastatic melanoma, the trans-well system did

not adequately capture their invasive potential, reflecting limitation of this assay, also noted by others [22]. We therefore examined if Slug expression could be used as a biomarker of cell invasiveness. Reassuringly, Slug was expressed in all cell lines, consistent with their metastatic potential (Fig. 7A). We proceeded to test the hypothesis that environmental glutamate modulates pro-invasive potential of GRIN2A mutations. Cells were

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

9

Fig. 7. Impact of media with different glutamate content on the expression of Slug, a biomarker of cell invasiveness. (A) Western blots and (B) quantification of Slug expression relative to ␤-actin in melanoma cell lines cultured in media with high (H; ␣-MEM) and low (L; RPMI-1640) glutamate content. Statistical significance is shown; *p < 0.05 **p < 0.01 (Student t-test) and chi-square analysis for mutated versus non-mutated cell lines. (C) Box plots comparing relative levels of Slug expression between cell lines with mutated and wild-type GRIN2A cultured in RPMI-1640. Statistical significance is shown (Student t-test).

cultured in media with different glutamate content and Slug expression was re-measured. We found that Slug expression increased in medium with less glutamate (RPMI-1640) but only for cells with mutated GRIN2A (Fig. 7B,C; p = 0.047), raising the possibility that glutamate levels modulate effects of GRIN2A mutations.

4. Discussion This study demonstrates that GRIN2A mutations in melanoma cause heterogenous effects but some may be oncogenic. NMDAR functionality in cultured cells was low but sufficient to support cell proliferation and invasion. In keeping with previous data [4], we found that the G889E mutation was associated with reduced NMDAR Ca2+ channel function. However, other GRIN2A mutations, in particular G762E were associated with maintained NMDAR activity and stronger NMDAR contribution to cell proliferation, compared with wild-type GRIN2A and other GRIN2A mutations, indicating a potential dependency on the G762E mutation for cell growth. The pro-invasive phenotype of mutated cell lines increased in media containing less glutamate, suggesting environmental modulation of mutation effects. Our results demonstrate for the first time, to our knowledge, that certain GRIN2A mutations con-

tribute oncogenic effects in melanoma and can be environmentally regulated. Previous work demonstrated loss-of-function consequences for mostly non-sense (truncating) GRIN2A mutations (W372X, Q891X, R920K, W1271X) studied in transfected 31T and SK-Mel2 melanoma cell lines [4]. Our approach differed in that we did not use transfections but examined biological associations of selected endogenous and missense GRIN2A mutations (S349F, G762E, G889E, P1132L, P1133S). We found that G889E was associated with reduced Ca2+ responses to NMDA and glutamate, consistent with previous findings that G889E inhibits NMDAR assembly [4]. However, other GRIN2A mutations, in particular G762E, maintained or increased NMDAR-mediated Ca2+ responses and enhanced cell proliferation and invasiveness. Our previous modelling of G762E effects in silico predicted that this mutation disrupts NMDAR functionality, but the type of effect was unclear [3]. The experimental evidence presented in this paper suggests that G762E increases NMDAR activity and causes oncogenic effects. The NMDAR channel blockers, MK-801 and memantine, but not a competitive NMDAR antagonist, AP5, inhibited proliferation of melanoma cells, which was similar to previous studies [8–10]. The lack of effects by AP5 may be because it needed to compete with exceeding concentrations of glutamate in medium. The anti-

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12 10

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

proliferative potency of MK-801 but not memantine correlated with the presence of GRIN2A mutations, which may be due to the modulator kinetics. MK-801 has high affinity for the NMDAR channel and blocks it irreversibly [6]; in contrast, memantine maintains residual NMDAR activity, hence effects of hyper-active NMDARs could be preserved [23]. The impact of extracellular glutamate on the growth of cancer cells is difficult to dissect in culture. RPMI-1640 contains 136 ␮M glutamate, which aligns closer with its concentrations in plasma ranging from ∼30 to 300 ␮M [24–26]. In comparison, ␣-MEM contains 510 ␮M glutamate, which causes toxicity in both neuronal and non-neuronal cells that express functional NMDARs [27,28]. Our NZM cell lines were established in ␣-MEM, therefore cells were selected for optimal growth in glutamate-rich environment. We hypothesize that cell adaptation included down-regulation of NMDAR expression, as high glutamate levels cause NMDAR overactivation, intracellular Ca2+ overload and cell death [29]. In support, NMDAR expression and function were low in all cell lines but may be higher in vivo as memantine was more effective in primary cultures. Western blotting did not detect expression of GluN2A in cell lines with mutated GRIN2A. Technical challenges may have influenced these results [30] but it is also possible that GRIN2A mutations led to a further reduction in the GluN2A protein abundance and the number or subunit composition of mutated NMDARs. Our data implies that cultured cells contain low numbers of functional NMDARs; however, this does not exclude their relevance in vivo. Many ion channel proteins are present at low density in non-neuronal cells, including less than 30 NMDARs in a human red cell, but they still exert important effects on cell function [31,32]. To gain better insights into the NMDAR role in melanoma in vivo, we need to test patient cells in primary culture or xenograft models using methods such as radiolabeled [3 H]MK-801 binding, electrophysiological recordings and further characterization of Ca2+ fluxes with a range of NMDAR agonists and antagonists. We found that cells carrying GRIN2A mutations up-regulated expression of Slug in medium with less glutamate, suggesting environmental regulation of mutation effects. We cannot exclude that other differences between media influenced Slug expression, which limits our conclusions, but the phenomenon highlights that standard culture conditions affect the glutamate-NMDAR axis. The use of cell lines with endogenous GRIN2A mutations is a strength of this study but there are limitations. Non-specific targets of pharmacological inhibitors cannot be excluded, although the overall pattern of responses supports NMDAR functionality in cultured cells. NMDAR subunits other than GluN2A and their mutations may have influenced glutamate effects we observed; to look into this possibility we are now conducting high-throughput genomic analysis of the NZM cell lines. Our observations are relevant to the biology of melanoma growth in vivo. Glutamate in the skin is secreted by keratinocytes [33] and nerve endings [34]. The exact concentrations are not known but these are likely to be lower than in plasma in the basal layer of the epidermis where keratinocytes express glutamate transporters that remove and recycle glutamate, and higher in more superficial layers of the skin that lack glutamate transporters [33,35,36]. Melanocytes are located in the basal layer of the epidermis but their projections span its entire width [37], hence melanocytes may be placed to respond to varied levels of glutamate through a range of glutamate receptors they express [38–40]. Normal melanocytes do not secrete glutamate but melanoma cells do and glutamate is known to supports melanoma growth not only through the NMDAR but also through metabotropic glutamate receptors [40,41]. Glutamate levels are likely to be highest in advanced melanoma tumors, contributed by seeping plasma, activated platelets [42] and infiltrating white cells [43], which is akin to the wound fluid, where glutamate levels range from 300 to

Fig. 8. Model of the role of GRIN2A mutations in melanoma growth. The role of GRIN2A mutations in melanoma growth depends on glutamate concentrations in the extracellular environment. Under lower glutamate levels found in early tumors, oncogenic GRIN2A mutations facilitate cell survival, proliferation and migration. Under higher glutamate levels found in advanced tumors, NMDARs may have tumor suppressive functions. To escape glutamate-mediated toxicity, cells down-regulate NMDAR expression and accumulate further mutations to reduce NMDAR activity.

>1000 ␮M [44]. Half maximal effective concentrations (EC50 ) for glutamate binding to recombinant NMDARs range from ∼0.5 to 3 ␮M [6]. During synaptic transmission, glutamate concentrations rise above 1 mM but quickly return to <1 ␮M baseline, which avoids receptor desensitization [24,45]. Melanoma NMDARs may also operate under glutamate concentrations from <1 ␮M to >1000 ␮M, but high glutamate levels are more likely to be sustained, which is also true in culture. We don’t know how melanoma NMDARs remain reactive under conditions of chronic glutamate elevation. The following possibilities have been suggested for other nonneuronal cells: increased expression of glutamate transporters, altered receptor composition, intracellular sequestration and the use of alternative agonists [46]. Glutamate EC50 values for NMDARs in rat red cells have been reported to be 88.2 ␮M [47]. Future studies will be required to examine this in melanoma. 5. Conclusions Based on previous and our data, we propose the following model to describe the role of the glutamate-NMDAR axis in melanoma (Fig. 8 and the graphical abstract). Under low glutamate levels, relevant in early melanoma tumors, oncogenic GRIN2A mutations increase NMDAR activity, which facilitates melanoma cell survival, proliferation and invasion. However, with tumor progression, glutamate levels rise, causing toxic effects. To escape NMDARmediated excitotoxicity, melanoma cells down-regulate NMDAR expression or acquire further mutations that reduce NMDAR function, more akin to the tumor suppressor effect. The high frequency of mutations in the GRIN genes in metastatic melanoma may therefore represent a selection of cells that proliferate optimally in a glutamate-rich microenvironment. We believe, a similar phenomenon may apply in cell culture and are reviewing our practices to derive cell lines. In summary, NMDARs can provide both oncogenic and tumor suppressor effects in melanoma, depending on the type of GRIN2A mutations and glutamate levels. Our data argues that glutamate in medium affects melanoma cell biology; therefore, alternative approaches are required to gain better insights into the role of the glutamate-NMDAR axis in melanoma in vivo. Disclosure of conflict of interests The authors state that they have no conflict of interests. All authors have approved the final article that has been submitted.

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

Author contributions SAND performed experiments, analyzed data and drafted the manuscript. WRJ and TNG provided technical support. EYL advised on research. MJD provided mentorship and advice. GJF and BCB contributed to experimental design, data interpretation and manuscript writing. MLK-Z designed the study, supervised research and wrote the manuscript. Acknowledgments This work was funded by Auckland Medical Research Foundation (UOA3700909) and Cancer Society of New Zealand (UOA3702269). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ceca.2016.09.003. References [1] X. Wei, V. Walia, J.C. Lin, et al., Exome sequencing identifies GRIN2A as frequently mutated in melanoma, Nat. Genet. 43 (2011) 442–446, http://dx. doi.org/10.1038/ng.810. [2] M.S. Stark, S.L. Woods, M.G. Gartside, et al., Frequent somatic mutations in MAP3K5 and MAP3K9 in metastatic melanoma identified by exome sequencing, Nat. Genet. 44 (2012) 165–169, http://dx.doi.org/10.1038/ng. 1041. [3] A. D’Mello S, J.U. Flanagan, T.N. Green, et al., Evidence that GRIN2A mutations in melanoma correlate with decreased survival, Front. Oncol. 3 (333) (2014), http://dx.doi.org/10.3389/fonc.2013.00333. [4] T.D. Prickett, B.J. Zerlanko, V.K. Hill, et al., Somatic mutation of GRIN2A in malignant melanoma results in loss of tumor suppressor activity via aberrant NMDAR complex formation, J. Invest. Dermatol. (2014), http://dx.doi.org/10. 1038/jid.2014.190. [5] G.L. Collingridge, R.W. Olsen, J. Peters, M. Spedding, A nomenclature for ligand-gated ion channels, Neuropharmacology 56 (2009) 2–5, http://dx.doi. org/10.1016/j.neuropharm.2008.06.063. [6] S.F. Traynelis, L.P. Wollmuth, C.J. McBain, et al., Glutamate receptor ion channels: structure regulation, and function, Pharmacol. Rev. 62 (2010) 405–496, http://dx.doi.org/10.1124/pr.109.002451. [7] P. Paoletti, C. Bellone, Q. Zhou, NMDA receptor subunit diversity: impact on receptor properties synaptic plasticity and disease, Nat. Rev. Neurosci. 14 (2013) 383–400, http://dx.doi.org/10.1038/nrn3504. [8] Z. Song, C.-D. He, J. Liu, et al., Blocking glutamate-mediated signalling inhibits human melanoma growth and migration, Exp. Dermatol. 21 (2012) 926–931, http://dx.doi.org/10.1111/exd.12048. [9] E.P. Seidlitz, M.K. Sharma, Z. Saikali, M. Ghert, G. Singh, Cancer cell lines release glutamate into the extracellular environment, Clin. Exp. Metastasis 26 (2009) 781–787, http://dx.doi.org/10.1007/s10585-009-9277-4. [10] M.P. Ribeiro, I. Nunes-Correia, A.E. Santos, J.B. Custodio, The combination of glutamate receptor antagonist MK-801 with tamoxifen and its active metabolites potentiates their antiproliferative activity in mouse melanoma K1735-M2 cells, Exp. Cell Res. 321 (2014) 288–296, http://dx.doi.org/10.1016/ j.yexcr.2013.11.002. [11] G.E. Hardingham, Coupling of the NMDA receptor to neuroprotective and neurodestructive events, Biochem. Soc. Trans. 37 (2009) 1147–1160, http:// dx.doi.org/10.1042/BST0371147. [12] D.C. Whiteman, A.C. Green, C.M. Olsen, The growing burden of invasive melanoma: projections of incidence rates and numbers of new cases in six susceptible populations through 2031, J. Invest. Dermatol. 136 (2016) 1161–1171, http://dx.doi.org/10.1016/j.jid.2016.01.035. [13] W.O. Jones, C.R. Harman, A.K. Ng, J.H. Shaw, Incidence of malignant melanoma in Auckland, New Zealand: highest rates in the world, World J. Surg. 23 (1999) 732–735. [14] E.S. Marshall, J.H. Matthews, J.H. Shaw, et al., Radiosensitivity of new and established human melanoma cell lines: comparison of [3H]thymidine incorporation and soft agar clonogenic assays, Eur. J. Cancer 30A (1994) 1370–1376, http://dx.doi.org/10.1016/0959-8049(94)90188-0. [15] H. Monyer, N. Burnashev, D.J. Laurie, B. Sakmann, P.H. Seeburg, Developmental and regional expression in the rat brain and functional properties of four NMDA receptors, Neuron 12 (1994) 529–540, http://dx.doi. org/10.1016/0896-6273(94)90210-0. [16] T. Kamal, T.N. Green, M.C. Morel-Kopp, et al., Inhibition of glutamate regulated calcium entry into leukemic megakaryoblasts reduces cell proliferation and supports differentiation, Cell Signal. 27 (2015) 1860–1872, http://dx.doi.org/10.1016/j.cellsig.2015.05.004.

11

[17] M.L. Kalev-Zylinska, T.N. Green, M.C. Morel-Kopp, et al., N-methyl-d-aspartate receptors amplify activation and aggregation of human platelets, Thromb. Res. 133 (2014) 837–847, http://dx.doi.org/10.1016/j.thromres.2014.02.011. [18] H.S. Chen, S.A. Lipton, The chemical biology of clinically tolerated NMDA receptor antagonists, J. Neurochem. 97 (2006) 1611–1626, http://dx.doi.org/ 10.1111/j.1471-4159.2006.03991.x. [19] F. Yamaguchi, Y. Hirata, H. Akram, et al., FOXO/TXNIP pathway is involved in the suppression of hepatocellular carcinoma growth by glutamate antagonist MK-801, BMC Cancer 13 (468) (2013), http://dx.doi.org/10.1186/1471-240713-468. [20] A. Doble, The pharmacology and mechanism of action of riluzole, Neurology 47 (1996) S233–241. [21] N. Lamanauskas, A. Nistri, Riluzole blocks persistent Na+ and Ca2+ currents and modulates release of glutamate via presynaptic NMDA receptors on neonatal rat hypoglossal motoneurons in vitro, Eur. J. Neurosci. 27 (2008) 2501–2514, http://dx.doi.org/10.1111/j.1460-9568.2008.06211.x. [22] Y.C. Du, B.C. Lewis, D. Hanahan, H. Varmus, Assessing tumor progression factors by somatic gene transfer into a mouse model: Bcl-xL promotes islet tumor cell invasion, PLoS Biol. 5 (2007) e276, http://dx.doi.org/10.1371/ journal.pbio.0050276. [23] J.W. Johnson, S.E. Kotermanski, Mechanism of action of memantine, Curr. Opin. Pharmacol. 6 (2006) 61–67, http://dx.doi.org/10.1016/j.coph.2005.09. 007. [24] M. Nedergaard, T. Takano, A.J. Hansen, Beyond the role of glutamate as a neurotransmitter, Nat. Rev. Neurosci. 3 (2002) 748–755, http://dx.doi.org/10. 1038/nrn916. [25] Y. Zhou, L. Qiu, Q. Xiao, et al., Obesity and diabetes related plasma amino acid alterations, Clin. Biochem. 46 (2013) 1447–1452, http://dx.doi.org/10.1016/j. clinbiochem.2013.05.045. [26] K. Kiessling, N. Roberts, J.S. Gibson, J.C. Ellory, A comparison in normal individuals and sickle cell patients of reduced glutathione precursors and their transport between plasma and red cells, Hematol. J. 1 (2000) 243–249, http://dx.doi.org/10.1038/sj/thj/6200033. [27] G.J. Brewer, Age-related toxicity to lactate, glutamate, and beta-amyloid in cultured adult neurons, Neurobiol. Aging 19 (1998) 561–568. [28] N.J. Anegawa, D.R. Lynch, T.A. Verdoorn, D.B. Pritchett, Transfection of N-methyl-d-aspartate receptors in a nonneuronal cell line leads to cell death, J. Neurochem. 64 (1995) 2004–2012. [29] K. Szydlowska, M. Tymianski, Calcium ischemia and excitotoxicity, Cell Calcium 47 (2010) 122–129, http://dx.doi.org/10.1016/j.ceca.2010.01.003. [30] A.E. Speers, C.C. Wu, Proteomics of integral membrane proteins-theory and application, Chem. Rev. 107 (2007) 3687–3714, http://dx.doi.org/10.1021/ cr068286z. [31] A. Makhro, P. Hanggi, J.S. Goede, et al., N-methyl-d-aspartate receptors in human erythroid precursor cells and in circulating red blood cells contribute to the intracellular calcium regulation, Am. J. Physiol. Cell Physiol. 305 (2013) C1123–1138, http://dx.doi.org/10.1152/ajpcell.00031.2013. [32] J.R. Wright, S. Amisten, A.H. Goodall, M.P. Mahaut-Smith, Transcriptomic analysis of the ion channelome of human platelets and megakaryocytic cell lines, Thromb. Haemost. 116 (2016) 272–284, http://dx.doi.org/10.1160/ TH15-11-0891. [33] M. Fischer, D. Glanz, M. Urbatzka, T. Brzoska, C. Abels, Keratinocytes: a source of the transmitter l-glutamate in the epidermis, Exp. Dermatol. 18 (2009) 1064–1066, http://dx.doi.org/10.1111/j.1600-0625.2009.00886.x. [34] E.M. Davidson, R.E. Coggeshall, S.M. Carlton, Peripheral NMDA and non-NMDA glutamate receptors contribute to nociceptive behaviors in the rat formalin test, Neuroreport 8 (1997) 941–946. [35] P.G. Genever, S.J. Maxfield, G.D. Kennovin, et al., Evidence for a novel glutamate-mediated signaling pathway in keratinocytes, J. Invest. Dermatol. 112 (1999) 337–342, http://dx.doi.org/10.1046/j.1523-1747.1999.00509.x. [36] W.K. Nahm, B.D. Philpot, M.M. Adams, et al., Significance of N-methyl-d-aspartate (NMDA) receptor-mediated signaling in human keratinocytes, J. Cell Physiol. 200 (2004) 309–317, http://dx.doi.org/10.1002/ jcp.20010. [37] J.Y. Lin, D.E. Fisher, Melanocyte biology and skin pigmentation, Nature 445 (2007) 843–850, http://dx.doi.org/10.1038/nature05660. [38] P.M. Pollock, K. Cohen-Solal, R. Sood, et al., Melanoma mouse model implicates metabotropic glutamate signaling in melanocytic neoplasia, Nat. Genet. 34 (2003) 108–112, http://dx.doi.org/10.1038/ng1148. [39] K.Y. Choi, K. Chang, J.M. Pickel, J.D. Badger 2nd, K.W. Roche, Expression of the metabotropic glutamate receptor 5 (mGluR5) induces melanoma in transgenic mice, Proc. Natl. Acad. Sci. U. S. A. 108 (2011) 15219–15224, http:// dx.doi.org/10.1073/pnas.1107304108. [40] T.D. Prickett, Y. Samuels, Molecular pathways: dysregulated glutamatergic signaling pathways in cancer, Clin. Cancer Res. 18 (2012) 4240–4246, http:// dx.doi.org/10.1158/1078-0432.CCR-11-1217. [41] V. Nygaard, L. Prasmickaite, K. Vasiliauskaite, T. Clancy, E. Hovig, Melanoma brain colonization involves the emergence of a brain-adaptive phenotype, Oncoscience 1 (2014) 82–94, http://dx.doi.org/10.18632/oncoscience.11. [42] L. Tremolizzo, J.C. DiFrancesco, V. Rodriguez-Menendez, et al., Human platelets express the synaptic markers VGLUT1 and 2 and release glutamate following aggregation, Neurosci. Lett. 404 (2006) 262–265, http://dx.doi.org/ 10.1016/j.neulet.2006.06.015. [43] I. Yawata, H. Takeuchi, Y. Doi, J. Liang, T. Mizuno, A. Suzumura, Macrophage-induced neurotoxicity is mediated by glutamate and attenuated

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003

G Model YCECA-1787; No. of Pages 12 12

ARTICLE IN PRESS S.A.N. D’mello et al. / Cell Calcium xxx (2016) xxx–xxx

by glutaminase inhibitors and gap junction inhibitors, Life Sci. 82 (2008) 1111–1116, http://dx.doi.org/10.1016/j.lfs.2008.03.010. [44] J.E. Albina, J.A. Abate, B. Mastrofrancesco, Role of ornithine as a proline precursor in healing wounds, J. Surg. Res. 55 (1993) 97–102, http://dx.doi.org/ 10.1006/jsre.1993.1114. [45] J.A. Dzubay, C.E. Jahr, The concentration of synaptically released glutamate outside of the climbing fiber-Purkinje cell synaptic cleft, J. Neurosci. 19 (1999) 5265–5274.

[46] A.D. Hogan-Cann, C.M. Anderson, Physiological roles of non-neuronal NMDA receptors, Trends Pharmacol. Sci. (2016), http://dx.doi.org/10.1016/j.tips. 2016.05.012. [47] A. Makhro, J. Wang, J. Vogel, et al., Functional NMDA receptors in rat erythrocytes, Am. J. Physiol. Cell Physiol. 298 (2010) C1315–C1325, http://dx. doi.org/10.1152/ajpcell.00407.2009.

Please cite this article in press as: S.A.N. D’mello, et al., Selected GRIN2A mutations in melanoma cause oncogenic effects that can be modulated by extracellular glutamate, Cell Calcium (2016), http://dx.doi.org/10.1016/j.ceca.2016.09.003