Guanine inhibits the growth of human glioma and melanoma cell lines by interacting with GPR23

Abstract

Guanine-based purines (GBPs) exert numerous biological effects at the central nervous system through putative membrane receptors, the existence of which is still elusive. To shed light on this question, we screened orphan and poorly characterized G protein-coupled receptors (GPRs), selecting those that showed a high purinoreceptor similarity and were expressed in glioma cells, where GBPs exerted a powerful antiproliferative effect. Of the GPRs chosen, only the silencing of GPR23, also known as lysophosphatidic acid (LPA) 4 receptor, counteracted GBP-induced growth inhibition in U87 cells. Guanine (GUA) was the most potent compound behind the GPR23-mediated effect, acting as the endpoint effector of GBP antiproliferative effects. Accordingly, cells stably expressing GPR23 showed increased sensitivity to GUA. Furthermore, while GPR23 expression was low in a hypoxanthine-guanine phosphoribosyl-transferase (HGPRT)-mutated melanoma cell line showing poor sensitivity to GBPs, and in HGPRT-silenced glioma cells, GPR23-induced expression in both cell types rescued GUA-mediated cell growth inhibition. Finally, binding experiments using [³H]-GUA and U87 cell membranes revealed the existence of a selective GUA binding (K_D = 29.44 ± 4.07 nM; Bmax 1.007 ± 0.035 pmol/mg prot) likely to GPR23. Overall, these data suggest GPR23 involvement in modulating responses to GUA in tumor cell lines, although further research needs to verify whether this receptor mediates other GUA effects.

Keywords: G protein-coupled receptor 23 (GPR23); antiproliferative effects; glioma cell lines; guanine (GUA); guanine-based purines (GBPs); lysophosphatidic acid (LPA); melanoma cell lines; purine nucleoside phosphorylase (PNP).

FIGURE 1

Effects of lipotransfection with different GPCR-siRNAs on sensitivity to GUO (A) and GMP (B). U87 cells, grown up to 50% confluency into 96 multiwell plates, were lipotransfected in the absence of siRNAs (mock transfected), in the presence of 50 nM ctrlRNA or in the presence of 50 nM GPR3-siRNA, GPR21-siRNA, GPR22-siRNA, LPAR6/P2Y₅-siRNA and GPR23-siRNA and then incubated in growth medium. After 24 h cells were treated with GUO or GMP (300 μM) for 72 h. Cell growth was evaluated by the MTT assay. Results are expressed as the percentage of cell growth evaluated in untreated cells (control). Each point represents the mean ± SD of 3 independent experiments. Statistical significance: *p < 0.05 **p < 0.01 vs. mock transfected cells (Student’s t test).

FIGURE 2

Effect of lipotransfection with GPR23-siRNA on sensitivity to GUA, GUO and GMP. Twenty-four hours after lipotransfection with GPR23-siRNA, U87 cells were treated with GUO or GMP (50–500 μM) or GUA (1–50 μM) for 72 h. Results are expressed as the percentage of cell growth evaluated in untreated cells (control) by the MTT assay. Each value is the mean ± SD of 10 different samples. Statistical significance of data obtained in GPR23-siRNA cells vs. control (ctrl) RNA transfected cells: p < 0.001 for data relating to GUA effect and p < 0.01 for those relating to GUO and GMP effects (two-way ANOVA test).

FIGURE 3

Effect of GPR23 overexpression on sensitivity to GUA. (A,B) Immunofluorescence microphotographs of a GPR23-overexpressing U87 cell clone (U87cl12). Expression of V5-tagged recombinant GPR23 was induced in U87cl12 cells by TAG on Demand adenoviral transduction and detected by the anti-V5-FITC conjugated antibody. (A) not-transduced U87cl12 cells stained with anti-V5-FITC antibody; (B) transduced U87cl12 cells stained with anti-V5- FITC antibody. The images are representative of six independent experiments performed with different clones overexpressing GPR23, which gave similar results. (C) Following stable overexpression of GPR23, some of these cells (U87cl8) were lipotransfected in the presence of 50 nM ctrlRNA (ctrl RNA) or GPR3-siRNA (GPR23siRNA). Twenty-four hours after this procedure, all cells were exposed to GUA (1–100 μM) for 72 h. Results, expressed as the percentage of cell growth evaluated in untreated cells (control) by the MTT assay, are the mean ± SD of 4 different experiments for each cell type. *p < 0.05 **p < 0.001 and ^# p < 0.05 ^### p < 0.001: statistical significance vs. U87 ctrl RNA and U87 cl8, respectively (Student’s t test).

FIGURE 4

Involvement of PNP activity in GBPs-mediated U87 cell growth inhibition. (A) Activity of PNP present in the U87 cell culture medium. The growth medium of cells was collected at different times (i.e., 1, 6 and 12 h) and the PNP activity determined (see Materials and Methods). (B) Impact of PNP inhibition on GBPs-dependent antiproliferative effect. U87 cells were exposed to GMP, GUO or GUA for 72 h, in the presence or absence (w/t) of forodesine, a PNP inhibitor. Results are expressed as the percentage of cell growth evaluated by the MTT assay in untreated cells (control). All values are the mean ± SD of 4 independent experiments. Statistical significance: **p < 0.01 vs. cells not treated with forodesine (Student’s t test).

FIGURE 5

Involvement of HGPRT enzyme in GUA-mediated cell growth inhibition. Concentration-response curves of GUA-mediated antiproliferative effect in (A) C32 melanoma cells expressing normal (WT, wild type) or inactive HGPRT form (C32TG cells) or (B) U87 wild type cells (WT) or silenced for HGPRT (U87 HGPRTsiRNA). Furthermore, the mRNA expression of GPR23 was evaluated in C32 (C) and U87 (D) cells by qRT-PCR. Finally, (E) C32TG cells with mutated HGPRT were transfected to overexpress GPR23 (C32TGcl19 and C32TGcl10) or LacZ construct (C32TGLacZ). Likewise, (F) U87 WT or HGPRT silenced cells were transfected to overexpress GPR23. Afterwards, cells were exposed to GUA as described above. For the entire duration of experiments, C32TG, -LacZ cells and GPR23 overexpressing clones were cultured in the medium without thioguanine. Results are expressed as the percentage of cell growth evaluated by the MTT assay in untreated cells (control) in (A,B,E,F) Panels. In all panels, values represent the mean ± SD of 3 independent experiments. Statistical significance: ***p < 0.001 vs. C32 or U87 WT cells (as for Panels (A,B)) or vs. C32TGLacZ cells (only as for Panel (E) (ANOVA two-way test); **p < 0.01 and ***p < 0.001 vs. C32 cells with normal HGPRT form or U87 WT cells; ^### p < 0.001 vs. U87 si-HGRPT cells (Student’s t test as for panels (C,D,F).

FIGURE 6

Characterization of [³H]-GUA radioligand binding to U87 cells membrane extracts. In all experiments membrane extracts were incubated with the indicated drugs under standard assay conditions, i.e., 30 min incubation time, 25°C temperature, pH 7.4). (A) Displacement of 50 nM [³H]-GUA binding from U87 cell membrane (50 μg protein) by unlabeled GUA or guanosine (GUO) (both at concentrations in the range from 0.001 to 10 μM). The values are the mean ± SD of three independent experiments, each point in triplicate. Statistical significance (two-way ANOVA test): ***p < 0.001 GUO vs. GUA displacement curve. (B) Saturation binding of [³H]-GUA at 25°C using control and GPR23 silenced U87 cell membranes (50 μg protein), which were incubated with increasing concentrations (6.25–300 nM) of [³H]-GUA under standard assay conditions. Non-specific binding was defined in the presence of 500 μM GUA. Values are the means ± SD of four experiments, each point performed in triplicate. Data in the Panel (A,B) were fitted by a computerized nonlinear regression analysis and resolved with a one site model. (C) Scatchard analysis of the data shown in (B). For the panels B and C, Bs, bound specific; statistical significance (two-way ANOVA test): ***p < 0.001 vs. control.

SCHEME 1

An outline of the effects promoted by GBPs, mainly GUA, in tumor cell lines. (A) GMP and guanosine (GUO), which are converted to guanine (GUA) by the sequential activity of the enzymes ecto-5′-nucleotidase and purine nucleoside phosphorylase (PNP), reduced the proliferation of cancer cell lines used in this study. Indeed, GUA was the actual effector of the effects induced by GMP and GUO, as the inhibition of PNP activity hindered the effects of these compounds but not that of GUA. GUA inhibited cell growth by interacting with GPR23 receptor, which was selected by a bioinformatic approach among potential candidates within a list of orphan and/or poorly characterized G-protein coupled receptors with high purinoceptor similarity. In fact, the enhanced expression of GPR23 in glioma cells or hypoxanthine-guanine phosphoribosyl-transferase (HGPRT)-mutated melanoma cell lines increased the sensitivity of these cells to GUA. (B) Conversely, silencing of the GPR23 receptor, as induced in glioma cells, or its low expression, as observed in melanoma cell lines with HGPRT-mutated form, reduced GUA antiproliferative effects. These findings, together with the results from binding experiments, confirmed the involvement of GPR23 in modulating responses to GUA in cancer cell lines, although further research is needed to better investigate the relationship between the activity of HGPRT and GPR23 receptor expression as well as to verify whether this orphan receptor mediates other effects of GUA.