Hyperglycemia exacerbates colon cancer malignancy through hexosamine biosynthetic pathway

Abstract

Hyperglycemia is a common feature of diabetes mellitus, considered as a risk factor for cancer. However, its direct effects in cancer cell behavior are relatively unexplored. Herein we show that high glucose concentration induces aberrant glycosylation, increased cell proliferation, invasion and tumor progression of colon cancer. By modulating the activity of the rate-limiting enzyme, glutamine-fructose-6-phosphate amidotransferase (GFAT), we demonstrate that hexosamine biosynthetic pathway (HBP) is involved in those processes. Biopsies from patients with colon carcinoma show increased levels of GFAT and consequently aberrant glycans' expression suggesting an increase of HBP flow in human colon cancer. All together, our results open the possibility that HBP links hyperglycemia, aberrant glycosylation and tumor malignancy, and suggest this pathway as a potential therapeutic target for colorectal cancer.

Figure 1

Hyperglycemia increases proliferation and invasion in MC38 cells. (a) Population doubling time (PDT) of MC38 cells cultured in high (HG) or low glucose (LG) concentration. n=(3); error bars indicate mean±se.m. (b) Photomicrography of lower chamber of transwell Matrigel-coated membrane of (LG and HG) MC38-GFP cells. (c) All cells that invaded to the lower chamber were counted after 24 h of incubation. n=3; ***P<0.0001 t test; scale bar, 100 μm.

Figure 2

Hyperglycemia exacerbates tumor progression in vivo. (a) Tumor area and (b) weight of MC38-LG or MC38-HG injected in the flank of euglycemic (EuG) and hyperglycemic (HyG) mice. n=9–18 per group; error bars indicate mean±s.e.m.; *P<0.05, **P<0.001, ***P<0.0001, ^#P<0.05, ^###P<0.0001; (*) represents comparison between MC38-HG and MC38-HG (HyG) with MC38-LG; (#) is relative to comparison between MC38-HG and MC38-HG (HyG). Anova two way, Bonferonni post test. (c) Magnetic resonance whole body of euglycemic (EuG) and hyperglycemic (HyG) mice, 5 days after subcutaneous MC38-HG cells injection. (d) Fluorescence levels of lung of euglycemic mice, 3 days after injection of MC38-GFP-LG or MC38-GFP-HG cells into lateral tail vein. n=3 per group. (e) Representative fluorescent images of lung 21 days after MC38-GFP cell infusion. The picture shows five distinct groups: Non-injected mice; euglycemic mice (EuG) injected with MC38-LG or MC38-HG cells; hyperglycemic mice (HyG) injected with MC38-HG or MC38-LG. After the acquisition of images, the nodules were counted (f) and lung was homogenized and diluted for a fluorescence read-out (g); n=5 per group, *P<0.05; **P<0.001;±s.e.m.; one-way ANOVA; Tukey's multiple comparison test.

Figure 3

Hyperglycemia induces aberrant glycosylation. (a) Scheme representing binding specificities of the lectins and RL2 antibody used in this study. (b) Flow cytometry histograms show representative binding profile of different lectins in MC38 cells cultured in high (HG, orange) or low glucose (LG, blue) concentration. Red empty histogram refers to cells stained with the FITC-conjugated streptavidin, and bar graph shows the differences between the fluorescence intensity for each lectin. O-GlcNAc labeling was accessed by immunolabeling with RL2 antibody, and its red histogram represents cells stained with Alexa Fluor 488-conjugated antibody. The results represent four experimental replicates. (n=4); unpaired t-test. (c) Left, chromatograms of polar metabolites of cell extracts from MC38-LG (gray line) and MC38-HG cells (black dashed line), showing regions corresponding to UDP-HexNAc and pNP retention times. Right, UDP-HexNAc quantification of MC38-LG (gray) and MC38-HG cells (black). Quantitative analyses are shown as mean±s.d. of two independent experiments. (d) Comparison of m/z localization in subcutaneous tumor from euglycemic and hyperglycemic mice. MALDI-MSI analysis showing the distribution of m/z 606.073 (UDP-hexosamine), m/z 202.10 (acetylcarnitine) and m/z 743.074 (NADPH) and right panels are H&E-stained. (e) Lectin binding to tumor tissue of hyperglycemic versus euglycemic mice. Photomicrografies are representative of subcutaneous tumor tissue from hyperglycemic (HyG) versus euglycemic (EuG) mice analyzed 28 days after MC38 implantation. Scale bar, 100 μm. (n=3).

Figure 4

Pharmacological inhibition of GFAT decreases tumor growth and invasion. (a) Proliferation analysis of MC38-HG cells cultured in the absence (Ctrl) or presence of 1 μM DON (DON). (n=3); error bars indicate mean±s.e.m.; t-test. (b) Tumor area and (c) weight of MC38-HG cells untreated, or treated, with 1 μM DON (MC38-HG+DON) and injected in the flank of euglycemic mice. MC38-HG (n=12); MC38-HG+DON (n=7). (d) Lower panel. Western blot analysis of lysate from MC38-LG cells treated with Mannitol (osmotic control), GlcNAc, DON and DON+GlcNAc analyzing O-GlcNAc levels and β-actin expression. Upper panel. Invasion of MC38-LG cells treated with Mannitol (osmotic control, 40 mM), GlcNAc (40 mM), DON (1 μM) and DON+GlcNAc by transwell analysis of 24 h of incubation; n=3,±s.e.m.; one-way ANOVA, Bonferroni's multiple comparison test *P<0.05; **P<0.001; ***P<0.001

Figure 5

Effect of GFAT deletion on tumor progression. (a, b) Protein expression of GFAT in shGFAT-MC38 cells. (n=4). (c) The population doubling time (PDT) analysis of Scrambled MC38-HG cells related to shGFAT MC38-HG cells. (d) The Matrigel-coated chamber transwell invasion assay of shGFAT MC38-HG cells compared to that of Scrambled MC38-HG cells. (e) Tumor area, (f) representative images and (g) weight of euglycemic (EuG) mice injected subcutaneously with shGFAT MC38-HG cells or Scrambled MC38-HG cells (n=6–8). (h) Representative images and (i) quantification of lung metastatic nodules in hyperglycemic (HyG) mice injected with shGFAT MC38-HG cells or scrambled MC38-HG cells (n=5). Results are expressed as mean±s.e.m.; two-tailed, unpaired t-test and Anova two way; Bonferonni post test. *P<0.05, ***P<0.001.

Figure 6

Reduction of aberrant glycosylation in GFAT silenced MC38 cells. Flow cytometry histograms of lectin binding and O-GlcNAc immunolabeling. The histograms show representative binding profile of different lectins and O-GlcNAc immunolabeling in Scrambled MC38-HG cells (Scr, green) or shGFAT MC38-HG cells (red). Yellow empty histogram refers to cells stained with the FITC-conjugated streptavidin and Alexa Fluor 488-conjgated antibody. Bar graph shows the differences between the fluorescence intensity for each marker. n=4 per group; error bars indicate mean±s.e.m. Unpaired t-test.

Figure 7

Aberrant levels of GFAT, SNA and O-GlcNAc in human colon tissue. (a–d) Western blot analysis of human colon adenocarcinoma (T) and adjacent normal tissue (N) to GFAT1, GFAT2, O-GlcNAc and actin. (e, f) GFAT1 and GFAT2 mRNA levels in tumor (T) versus adjacent normal tissue (N). (g, h) Dot blot analysis in tumor (T) and adjacent normal tissues (N), measuring the SNA levels. The images are representative from four patients (P22, P24, P28 and P36). All graphs represent means±s.e.m.; one-tailed unpaired t-test; total number of seven patients n=7; *P<0.05.