Regulation of mGluR1 expression in human melanocytes and melanoma cells

Abstract

We demonstrated that ectopic expression of metabotropic glutamate receptor 1 (mGluR1/Grm1) in mouse melanocytes was sufficient to induce melanoma development in vivo with 100% penetrance. We also showed that about 60% of human melanoma biopsies and cell lines, but not benign nevi or normal human melanocytes expressed mGluR1, suggesting that GRM1 may be involved in melanomagenesis. mGluR1 is expressed primarily in neurons. In various non-neuronal cells, mGluR1 expression is regulated via binding of Neuron-Restrictive-Silencer-Factor (NRSF) to a Neuron-Restrictive-Silencer-Element (NRSE). Here, we report on the possibility that aberrant mGluR1 expression in melanoma is due to alterations in NRSF and/or NRSE. We show that in human melanocytes, binding of NRSF to NRSE in the GRM1 promoter region is necessary for the suppression of mGluR1 expression. We also show that inhibiting the expression of the transcription factor Sp1 or interference with its ability to bind DNA can result in increased mGluR1 expression perhaps via its function as a negative regulator. In addition, we also provide evidence that demethylation within the promoter region of GRM1 may also be a mechanism for the derepression of mGluR1 expression in melanocytes that progress to cell transformation and tumor formation.

Figure 1

A: Aberrant mGluR1 expression is detected in human melanoma biopsies and dysplastic nevi. HEM is included as normal melanocyte control. A melanocyte-specific marker, tyrosinase related protein 1 (TYRP1), is used as loading control. In HEM, the TYRP1 is not fully modified thus it migrates faster. The arbitrary units of the intensity of mGluR1 to TYRP1 were determined with Optiquant Software and are shown below each lane. B: NRSF expression in several melanoma cell lines and a subset of biopsies from panel A. IMR-32, a neuroblastoma cell line was included as a control. Varying levels of NRSF were detected among melanoma cell lines and biopsies; there was no correlation with aberrant mGluR1 expression. The arbitrary units of the intensity of NRSF to α-Tubulin were determined with Optiquant Software and are shown below each lane. C: Examination of REST and REST4 expression by RT/PCR on RNA isolated from human melanoma cell lines with primers that discriminate between full-length REST and its variant, REST4. IMR-32 is a neuroblastoma cell line and is used as a positive control for REST4. HEK293, a human embryonic kidney cell line, is used as a positive control for the full-length REST. mGluR1 expressing human melanoma cell lines are C8161, Skmel187, WM239A and WM3211. mGluR1 negative melanoma cell lines are C81-61, UACC930, WM8, and WM1382. NHM and HEM are normal human melanocytes. RT-PCR showed wild type REST (NRSF) and REST4 at 311 and 361bp, respectively. GAPDH was used as a control for equal amounts of template in RT/PCR. The numbers represent arbitrary units of intensity of Rest or Rest 4 to GAPDH.

Figure 2

A: NRSF represses GRM1 gene transcription through an NRSE dependent mechanism. Wild type (hGRM1NE) and mutated NRSE (hGRM1NEm) linked luciferase constructs were transfected into C8161, UACC903, HEK293 and HEM. Promoter activity of each construct is expressed as luciferase activity relative to the activity of each corresponding luciferase reporter with pGL3 basic vector, which is assigned an activity value of 1. The data shown is the average of three independent experiments ± S.D. B: Examination of Co-REST and mSin3A expression by Western immunoblots on protein lysates isolated from mGluR1-negative HEK293 and normal human primary melanocytes (HEM), mGluR1 expressing human melanoma cell line C8161, and human biopsies: nevi (mGluR1 negative) and eight other mGluR1 expressing biopsies. α-Tubulin was used as a loading control. The numbers represent arbitrary units of intensity of coREST or mSin3A to α-Tubulin. C: Examination of three other known NRSF regulated genes by RT-PCR. Only IMR-32, a medulloblastoma cell line showed expression of all three, none was detected in the other four non-neuronal cell lines. D: NRSF binds to the NRSE region on the GRM1 promoter and is detected by Chromatin Immunoprecipitation (ChIP) in C8161, UACC903 and HEM. Input is an aliquot of the DNA-protein complexes before immunoprecipitation. The remaining DNA-protein complexes are subjected to immunoprecipitation with antibodies against NRSF, HDAC1 and control rabbit IgG. Immunoprecipitated DNA was amplified with specific primers across the NRSE region of the GRM1 promoter

Figure 3

A: Putative Sp1 transcription factor binding GC box is detected adjacent to the NRSE site in the human GRM1 promoter. B: Co-immunoprecipitation of the endogenous NRSF and Sp1 in C8161 and HEM cells. Immunoprecipitation assays were performed first with nuclear extracts from C8161 cells and HEM human primary melanocytes using antibodies against Sp1, Sp2 and Sp3. The immuno-complexes were subjected to Western blot with NRSF antibody. Both C8161 and HEM showed NRSF expression; only C8161 showed binding of NRSF to Sp1. C: Western blot of C8161, UACC903, M19 and MV3 cells treated with 100nM and 500nM mithramycin A for 24 hours and probed for mGluR1 expression normalized to α-Tubulin. The numbers represent arbitrary units of intensity of mGluR1 to α-Tubulin determined with Optiquant Software. D: Western blots of C8161 and M19 cells transiently transfected with siSP1 after 72 hours, protein lysates were prepared and probed for Sp1 expression (left panel). Same lysates probed for mGluR1 expression and normalized to α-Tubulin (right panel). The numbers represent arbitrary units of intensity of mGluR1 to α-Tubulin determined with Optiquant Software.

Figure 4

A: Western immunoblots of cell lysates isolated from hTERT/CDK4^R24C/p53^DD and B: FOM 139-1 cells on day 0, 3, 5, 7, or 9 after 5-azacytidine treatments and probed with antibody to mGluR1 with mGluR1-positive melanoma SK-Mel187 cells as positive control. α-Tubulin was used to show equal loading. MAGE A4 expression was used as a positive control for demethylated DNA. C8161 human melanoma cell lysates were used as positive control for MAGEA4. α-Tubulin was used to show equal loading. C: Promoter methylation PCR assays were performed with genomic DNA isolated from hTERT/CDK4^R24C/p53^DD and FOM 139-1 according to manufacturer’s instruction. Left panel; PCR performed with primer set flanking GRM1 promoter region (bases −4846/−5267) and eluted complexes of MeCP2 and Mse1 digested hTERT/CDK4^R24C/p53^DD and FOM 139-1 genomic DNA. M, marker, Hae III digested ϕx174 DNA, Lanes 1–3 are hTERT/CDK4^R24C/p53^DD; no DNA control, Mse1 digested genomic DNA, MeCP2 complexed Mse1 digested DNA. Lanes 4–6 are FOM 139-1; no DNA control, Mse1 digested genomic DNA, MeCP2 complexed Mse1 digested DNA. Right panel, PCR performed with primer set flanking GRM1 promoter region (bases −3257/−3874) and eluted complexes of MeCP2 and Mse1 digested hTERT/CDK4^R24C/p53^DD and FOM 139-1 genomic DNA, lanes 7–9 are hTERT/CDK4^R24C/p53^DD; no DNA control, Mse1 digested genomic DNA, MeCP2 complexed Mse1 digested DNA. Lanes 10–12 are FOM 139-1; no DNA control, Mse1 digested genomic DNA, MeCP2 complexed Mse1 digested DNA.