GLUT5 (SLC2A5) enables fructose-mediated proliferation independent of ketohexokinase

Abstract

Background: Fructose is an abundant source of carbon and energy for cells to use for metabolism, but only certain cell types use fructose to proliferate. Tumor cells that acquire the ability to metabolize fructose have a fitness advantage over their neighboring cells, but the proteins that mediate fructose metabolism in this context are unknown. Here, we investigated the determinants of fructose-mediated cell proliferation.

Methods: Live cell imaging and crystal violet assays were used to characterize the ability of several cell lines (RKO, H508, HepG2, Huh7, HEK293T (293T), A172, U118-MG, U87, MCF-7, MDA-MB-468, PC3, DLD1 HCT116, and 22RV1) to proliferate in fructose (i.e., the fructolytic ability). Fructose metabolism gene expression was determined by RT-qPCR and western blot for each cell line. A positive selection approach was used to "train" non-fructolytic PC3 cells to utilize fructose for proliferation. RNA-seq was performed on parental and trained PC3 cells to find key transcripts associated with fructolytic ability. A CRISPR-cas9 plasmid containing KHK-specific sgRNA was transfected in 293T cells to generate KHK^-/- cells. Lentiviral transduction was used to overexpress empty vector, KHK, or GLUT5 in cells. Metabolic profiling was done with seahorse metabolic flux analysis as well as LC/MS metabolomics. Cell Titer Glo was used to determine cell sensitivity to 2-deoxyglucose in media containing either fructose or glucose.

Results: We found that neither the tissue of origin nor expression level of any single gene related to fructose catabolism determine the fructolytic ability. However, cells cultured chronically in fructose can develop fructolytic ability. SLC2A5, encoding the fructose transporter, GLUT5, was specifically upregulated in these cells. Overexpression of GLUT5 in non-fructolytic cells enabled growth in fructose-containing media across cells of different origins. GLUT5 permitted fructose to flux through glycolysis using hexokinase (HK) and not ketohexokinase (KHK).

Conclusions: We show that GLUT5 is a robust and generalizable driver of fructose-dependent cell proliferation. This indicates that fructose uptake is the limiting factor for fructose-mediated cell proliferation. We further demonstrate that cellular proliferation with fructose is independent of KHK.

Keywords: Fructose; GLUT5 (SLC2A5); Hexokinase; Ketohexokinase; Metabolism.

Fig. 1

Cellular gene expression and tissue of origin do not determine cellular proliferation in fructose. a PC3 and HepG2 were seeded into 12-well plates (20,000 cells/well) and cultured in the absence or presence of 10 mM fructose, or 10 mM glucose media for approximately 3 days. Cell density (% confluency) was monitored over time using live cell imaging (n = 2 per media condition). b Fructolytic index (fructose-mediated growth/glucose-mediated growth) of the indicated cell lines arranged in order of least to most fructolytic (n = 3). c Fructolytic index of cell lines in b grouped by tissue of origin. d Normalized expression of the indicated genes for each cell line shown as a heatmap. Cell lines ordered by fructolytic index (n = 2 per gene per cell line). *Denotes C_t > 30. e Immunoblot of the indicated proteins using lysates from the indicated cell lines, ordered from least to most fructolytic. The murine muscle, liver, and Khk knockout liver were used as controls

Fig. 2

Cells can be trained to metabolize fructose for proliferation. a Schematic for the positive selection strategy to generate fructolytic cell lines. b PC3 and PC3 passage 10 (P10) cells were seeded into 96-well plate (1500 cells/well) and cultured in media containing various amounts of sugar. Cell density (% confluency) was monitored over time using live cell imaging (n = 2 per condition). c Schematic for the competition growth assay between PC3-red (parental PC3 cells transduced with RFP reporter) and fructose-trained cell lines. d 40,000 of PC3, semi-trained PC3 passage 20 (P20), and trained-PC3 cells in 10 mM fructose or 11 mM glucose over time (n = 2 per condition). e Cells from d were grown in 10 mM fructose or 11 mM glucose for 96 h. They were then fixed and stained with crystal violet solution (n = 2 per condition). f 20,000 PC3-red and 20,000 trained PC3 cells were seeded in the same well and cultured for 96 h in 10 mM fructose or 10 mM glucose-containing media. Live fluorescent imaging was performed and the proportion of PC3-red cells to total PC3 cells is shown over time (n = 2 per condition). Supplemental Video 1 and Supplemental Video 2 are of competition assays monitored with live cell imaging

Fig. 3

GLUT5 overexpression rescues cellular proliferation in fructose. a Normalized expression of genes that are differentially expressed (q = 0.4, >1.1 log₂ fold change) between PC3 and semi-trained PC3 cells (passage 20) presented in heatmap form. b Relative expression of SLC2A5 transcript in semi-trained PC3 and trained PC3 cells as compared to the parental PC3 line. Two primer sets were used (n = 2 per condition). Two-way ANOVA with Fisher’s LDS test. *P < 0.05, ****P < 0.0001. c Immunoblot of the indicated proteins using lysates from PC3, semi-trained PC3, and trained PC3 cells. The murine liver and muscle used as controls. d GLUT5 or an empty vector (EV) were overexpressed in the indicated cells lines. The cells were plated at 20,000-30,000 cells/well and then grown in the presence of no sugar, 10 mM fructose, or 10 mM glucose. After 3 days, the cells were fixed and stained with crystal violet solution (n = 2 per condition)

Fig. 4

Fructose fluxes through HK, not KHK, in order to sustain cellular proliferation. a Percent of heavy isotope (¹³C) incorporation into fructose, fructose 1-phosphate (F1P), and lactate as detected by LC/MS from polar extracts of PC3, semi-trained PC3, and trained PC3 cells (n = 2-3). The isotopic labelling is indicated by M+# designation indicated in the legend where the # represents the amount of [¹²C] replaced by [¹³C]. Two-tailed unpaired t tests were used between parental and trained cells (M+3 for lactate, M+6 for fructose/F1P). *P < 0.05, **P< 0.01, and ****P < 0.0001. b Total abundance of fructose, F1P, and lactate as detected by LC/MS from polar extracts of PC3, semi-trained PC3, and trained PC3 cells (n = 2–3). Two-tailed unpaired t tests were used between parental and trained cells. *P < 0.05, **P< 0.01, and ****P < 0.0001. c Extracellular acidification rate (ECAR) over time of PC3, semi-trained PC3, and trained PC3 cells under basal conditions and following the addition of glucose, oligomycin (Oligo), and 2-deoxyglucose (2-DG) at the indicated times. Data are the mean and SEM from 6 replicates. d ECAR over time of PC3, semi-trained PC3, and trained PC3 cells under basal conditions and following the addition of fructose, Oligo, and 2-DG at the indicated times. Data are the mean and SEM from 6 replicates. e GLUT5 or an empty vector (EV) were overexpressed in 293T or 293T KHK^-/- cells. The cells were plated at 20,000 cells/well and then grown in the presence of no sugar, 10 mM fructose, or 10 mM glucose. After 7 days, the cells were fixed, stained with crystal violet solution (n = 2 per condition). f Fold change in cell viability as assessed by ATP concentration (Cell Titer Glo) of the indicated fructolytic cell lines grown in either 10 mM glucose or 10 mM fructose containing the specified concentrations of 2-DG after 72 h (n = 3 per concentration). The half maximal inhibitory concentration (IC₅₀) is displayed on the graph for each curve. g Fold change in cell viability as assessed by ATP concentration (Cell Titer Glo) of the trained PC3 grown in the specified sugar conditions containing the specified concentrations of 2-DG after 96 h. (n = 2 per concentration). The half maximal inhibitory concentration (IC₅₀) is displayed on the graph for each curve