Quercetin Impairs the Growth of Uveal Melanoma Cells by Interfering with Glucose Uptake and Metabolism

Abstract

Monosomy 3 in uveal melanoma (UM) increases the risk of lethal metastases, mainly in the liver, which serves as the major site for the storage of excessive glucose and the metabolization of the dietary flavonoid quercetin. Although primary UMs with monosomy 3 exhibit a higher potential for basal glucose uptake, it remains unknown as to whether glycolytic capacity is altered in such tumors. Herein, we initially analyzed the expression of n = 151 genes involved in glycolysis and its interconnected branch, the "pentose phosphate pathway (PPP)", in the UM cohort of The Cancer Genome Atlas Study and validated the differentially expressed genes in two independent cohorts. We also evaluated the effects of quercetin on the growth, survival, and glucose metabolism of the UM cell line 92.1. The rate-limiting glycolytic enzyme PFKP was overexpressed whereas the ZBTB20 gene (locus: 3q13.31) was downregulated in the patients with metastases in all cohorts. Quercetin was able to impair proliferation, viability, glucose uptake, glycolysis, ATP synthesis, and PPP rate-limiting enzyme activity while increasing oxidative stress. UMs with monosomy 3 display a stronger potential to utilize glucose for the generation of energy and biomass. Quercetin can prevent the growth of UM cells by interfering with glucose metabolism.

Keywords: glucose uptake; glycolysis; oxidative stress; pentose phosphate pathway; proliferation; quercetin; uveal melanoma.

Figure 1

Differentially expressed genes in the monosomy 3 and metastatic tumors of the TCGA-UM cohort (n = 80 patients). (A) The tumor samples were aligned according to the copy numbers of chromosome 3p and 3q (dark blue: normal, red: loss) as well as the metastatic status in the uppermost three rows. The expression heatmap was constructed using the z-scores, with red and blue indicating mRNA levels that were up to three standard deviations above or below the mean (black), respectively. The gene symbols and loci are stated on the left. All of the genes had an adjusted p-value < 0.05 with regard to the copy number of chromosome 3 and metastases. n/a: Not available. (B) Fold changes (FC) of median gene expression in the monosomy 3 tumors and patients with metastases. The up- and downregulated genes are highlighted with a red or blue background, respectively. The red lines indicate an FC of |2|. Chr3: Chromosome 3. (C) Gene set enrichment analysis demonstrating the biological pathways and phenotypes that were over-represented among the differentially expressed genes. All of the p-values and false discovery rates were <0.05. For simplicity, the five most enriched processes or pathways were demonstrated for the Gene Ontology (GO), KEGG, and Reactome databases.

Figure 2

Validation of gene expression in two independent cohorts and the correlation of validated genes with prognostic factors. (A) The expression of PFKP and ZBTB20 with regard to the metastases in the GSE22138 and GSE44295 cohorts is presented as boxplots. The mean values are connected by the sloped lines. For the GSE22138 cohort, the average expression of all of the available isoforms is presented (n = 2 isoforms for PFKP; n = 12 isoforms for ZBTB20). FC: fold change of median gene expression in the patients with metastases. The raw p-values were determined using the Mann–Whitney U test. Gene loci were indicated underneath the gene symbols. (B) The probability of overall survival with regard to the expression of PFKP and ZBTB20 in the primary UMs of the TCGA cohort is demonstrated by the Kaplan–Meier curves. The median gene expression was taken as the cut-off value. p-values were determined via the log-rank test. (C) Expression of PFKP and ZBTB20 with regard to the clinical and histopathological factors in the UM cohort of the TCGA study is presented as bar charts. The fold change of median gene expression in the subgroup with the unfavorable condition was calculated based on the classification of the prognostic factors according to the categories indicated in parentheses. The log₂(Fold Change) values above or below 0 were depicted in pink or blue, respectively. TILS: tumor-infiltrating lymphocytes, TAMS: tumor-activated macrophages. The raw p-values were determined using the Mann–Whitney U test. * p < 0.05, ** p < 0.005, and *** p < 0.001.

Figure 3

Quercetin suppresses the viability and proliferation of UM cells by increasing oxidative stress. (A) Representative light microscopy images of the UM cells after 3-day incubation with the substances indicated above followed by the MTT dye. A group of cells were incubated in culture medium with 0.5% fetal bovine serum (FBS) as the positive control for growth factor deprivation. A separate group of cells were incubated with the solvent of quercetin (dimethylsulfoxide; DMSO) at the same volume required for the administration of quercetin. (B) Quantification of the MTT assay (mean ± standard deviation (SD) of n = 4–5 independent experiments). The exponential trendlines were fitted and indicated in a dashed pattern in dark gray or green for DMSO or quercetin, respectively. Quercetin exhibited an IC50 of approximately 44.05 µM that was calculated from its respective trendline. * p < 0.05 for the pairwise comparison of the quercetin treatment to the corresponding solvent control, Mann–Whitney U test. (C) Representative images of immunofluorescence staining for the proliferation marker Ki67 (red) after 3 days. Nuclei were counterstained with DAPI (blue). Scale bar = 25 µm. (D) BrdU-assay demonstrating the dose-dependent anti-proliferative effect of quercetin after 3 days (mean ± SD of n = 3 independent experiments). p-values were determined through the use of the Mann–Whitney U or Kruskal–Wallis tests for pairwise or collective comparisons, respectively. (E) Representative images of the live/dead assay after 3 days. Scale bar = 50 µm. (F) Ratio of the live/dead intensity (mean ± SD of n = 3 wells from a representative experiment). p-values were assessed using the two-sided t-test or one-way analysis of variance for pairwise or collective comparisons, respectively. (G) Representative images demonstrating the accumulation of reactive oxygen species (ROS) in the cells treated with 50 µM quercetin for 12–13 h. Scale bar = 50 µm. (H) Quantification of the ROS intensity (mean ± SD of n = 3 independent experiments, Mann–Whitney U test).

Figure 4

Quercetin interferes with glucose uptake and metabolism in the UM cells. (A) Representative images demonstrating the uptake of the fluorescent glucose analog 6-NBDG in the cells incubated with the normal medium alone (control) and with the supplementation of 50 µM quercetin or its solvent DMSO. The overlay images of fluorescence and phase-contrast microscopy are presented to demonstrate the cell boundaries. Scale bar = 100 µm. (B) Quantification of 6-NBDG uptake (mean ± SD of n = 3–4 independent experiments, a.u.: arbitrary units, Q50: 50 µM quercetin,). (C) Quantification of the fluorometric glycolysis assay, demonstrating the significant reduction in the glycolytic rate in response to quercetin after 3 h (mean ± SD of n = 3 independent experiments). (D) The luminescent ATP-assay, demonstrating the reduction in ATP levels in the cells incubated with 50 µM quercetin for 2 days (mean ± SD of n = 3 independent experiments). (E) Activity of the glucose-6-phosphate dehydrogenase (G6PDH), which functions as the rate-limiting enzyme of the PPP [17,20,21], as determined by a colorimetric assay after exposing the cells for 11–13 h to the indicated treatments (mean ± SD of n = 3 independent experiments). G6PDH activity was normalized to the amount of total protein extracted from each group. The insignificant p-value above 0.05 is highlighted with a gray background. (F) Quantification of the total glutathione (GSH) levels after 7–10 h (mean ± SD of n = 3 independent experiments). In panels (B–E), all of the pairwise comparisons were evaluated via the Mann–Whitney U test whereas the collective comparisons of the three subgroups were performed with the Kruskal–Wallis test (B,E,F) or one-way analysis of variance (C,D).

Figure 5

Expression of the PFKP and ZBTB20 proteins in response to quercetin after 1 day. (A) Representative images of the fluorescent immunostainings for PFKP and ZBTB20 (red). The actin filaments were visualized via Alexa 488-phalloidin staining (green). Nuclei were counterstained with DAPI (blue). (B) Immunoblotting for PFKP and ZBTB20. Quercetin was administered at a concentration of 50 µM. Membranes were probed for beta-actin as a loading control. The total amount of protein in each well is also depicted for a more comprehensive evaluation of sample loading. kDa: Kilodalton. (C) Quantification of the PFKP and ZBTB20 levels in the immunoblots, which were normalized to the total protein loadings in each well. The PFKP levels are stated as the percentage of the 0.5% FBS group. a.u.: arbitrary units. Data represent the mean ± standard deviation of n = 3 independent experiments. p-values were evaluated with the Kruskal–Wallis test for the simultaneous comparison of all subgroups.