Augmented TME O-GlcNAcylation Promotes Tumor Proliferation through the Inhibition of p38 MAPK

Cancer cells rely preferentially on glycolysis as the main energy source even in the presence of oxygen. This phenomenon of preferential aerobic glycolysis is known as the Warburg effect (1). In cancer cells, glycolysis is further stimulated by tumor hypoxia (2), oncogenes, and mutant tumor suppressors (3). In spite of being an inefficient way to generate adenosine 5′-triphosphate, aerobic glycolysis is selected by cancer cells, which require a lot of energy to proliferate. This discrepancy can be explained by the fact that aerobic glycolysis produces acetyl-CoA and NADPH, which can then be used to efficiently generate all the nutrients (nucleotides, amino acids, and lipids) that are needed to produce a new cell (4). Therefore, cancer cells require a lot of glucose for proliferation and, in this respect, hyperglycemia could be considered a favorable condition for cancer. Although a possible association between diabetes and cancer risk has long been speculated, a correlation has not been established. However, accumulating epidemiologic evidence suggests an increased incidence and progression of several cancers in diabetic patients with cancer (5).

In diabetic patients, the elevated glucose increases flux through the hexosamine biosynthesis pathway (6), which often leads to increased concentrations of uridine diphospho-N-acetylglucosamine (UDP-GlcNAc). UDP-GlcNAc is the donor substrate for dynamic posttranslational modifications of nucleocytoplasmic proteins on serine (Ser) and threonine (Thr) residues with a single N-acetylglucosamine (O-GlcNAc; ref. 7). This modification is known as O-GlcNAcylation and is regulated by two enzymes, O-GlcNAc transferase (OGT; ref. 8) and O-GlcNAcase (OGA; ref. 9). Several studies showed that O-GlcNAc and phosphate often compete for the same sites (10), and the crosstalk between the two is thought to be essential for the regulation of cellular signaling (11–13). It is thought that diabetic complications, such as retinopathy (14), nephropathy (15), and cardiomyopathy (16) are partly due to increased O-GlcNAcylation.

Recent studies reported that increased O-GlcNAcylation is also a general feature of cancer cells and contributes to various cancer phenotypes, including cell proliferation, survival, invasion and metastasis, energy metabolism, and epigenetics (17–20). O-GlcNAcylation upregulates cancer cell proliferation by modification of transcription factors such as Myc, p53, NF-κB, or β-catenin (21). O-GlcNAcylation is not only involved in proliferation, but also in invasion or metastasis through the modification of the transcriptional repressor SNAIL1 (22). SNAIL1, stabilized by O-GlcNAcylation, induces the downregulation of E-cadherin expression, resulting in epithelial–mesenchymal transition.

O-GlcNAcylation in cancer cells is thought to be one of the effectors related to the Warburg effect (23). O-GlcNAcylation within breast cancer cells regulates cancer cell metabolism via regulation of HIF-1α and its downstream target, GLUT1, which is an effector of the Warburg effect (24). Because cancer cell metabolism is affected by the tumor microenvironment (TME; ref. 25), O-GlcNAcylation in the TME could also play pivotal roles in cancer cell metabolism and proliferation. Although many studies reported the effect of O-GlcNAcylation within cancer cells, the effect of increased O-GlcNAcylation in the TME has not been fully elucidated. Here, we examined the role of O-GlcNAcylation in the TME using a transplantation model system in which melanoma cells were subcutaneously transplanted into Ogt transgenic (Ogt-Tg) mice and clarified the increase in tumor growth and the mechanism by which O-GlcNAcylation in the TME promotes cancer progression.

For Western blotting, anti-OGT rabbit polyclonal antibody (pAb; sc-32921) and anti-NF-κB p65 rabbit pAb (sc-372) were obtained from Santa Cruz Biotechnology. Mouse mAbs against β-actin (clone AC-15) and O-GlcNAc (clone RL2) were obtained from Sigma-Aldrich Japan and Thermo Fisher Scientific, respectively. Antibodies against ERK1/2 (#4696), p-ERK1/2 (Thr 202/Tyr 204; #9101), MEK1/2 (#9122), p-MEK1/2 (Ser 217/221; #9154), p-B-RAF (Ser 445; #2696), C-RAF (#9422), p-C-RAF (Ser 259; #9421), AKT (#9272), p-AKT (Ser 473; #9271), p38 MAPK (#9212), p-p38 MAPK (Thr 180/Tyr 182; #4511), SRC (#2109), p-SRC (Tyr 416; #2101), IκBα (#4814), p-IκBα (Ser 32/36; #9246), and Histone H3 (#4499) were purchased from Cell Signaling Technology Japan. Anti-B-RAF rabbit (ab33899) and anti-p-NF-κB p65 (Ser 536; ab86299) antibodies were from Abcam. Anti-GAPDH antibody (010-25521) was from Wako Pure Chemical Industries. Secondary horseradish peroxidase (HRP)-conjugated antibodies were from The Jackson Laboratory. For immunofluorescence, anti-CD3 (550275), anti-B220 (553085), and anti-CD31 (clone MEC13.3) antibodies were purchased from BD Pharmingen. Anti-F4/80 rat monoclonal (clone Cl:A3-1), FITC-conjugated anti-α-SMA mouse monoclonal (clone 1A4), Alexa 488–conjugated anti-O-GlcNAc mouse monoclonal (clone RL2), NF-κB p65 rabbit polyclonal (sc-372), iNOS rabbit polyclonal (NB300-605), and arginase-1 rabbit monoclonal (#93668) antibodies were obtained from AbD Serotec, Sigma-Aldrich Japan, Abcam, Santa Cruz Biotechnology, and Novus Biologicals, respectively. Secondary Alexa 488-, Alexa 546–conjugated antibodies, and 4',6-diamidino-2-phenylindole (DAPI) were purchased from Invitrogen Japan. Cy3-, Cy5, and DyLight 488–conjugated antibodies were from Jackson Laboratory. For cell culture experiments, a p38 MAPK–specific inhibitor, SB203580, and a MEK1/2-specific inhibitor, PD98059, were purchased from Cayman Chemical Company and Calbiochem, respectively.

B16 mouse melanoma were kindly provided by Professor M. Miyasaka (The University of Osaka, Japan) and maintained in DMEM (Life Technologies Japan) supplemented with 10% FBS (Invitrogen Japan) and penicillin/streptomycin (100 IU/50 μg/mL; Invitrogen Japan) at 37°C in a humidified atmosphere containing 5% CO₂.

For the generation of Ogt-Tg mice, the coding sequence of the mouse Ogt (GenBank accession number NM_139144.4) cDNA was cloned into pCAGGS vector (26) in the multiple cloning sites under the control of the CAG promoter. Ogt-Tg mice were generated by standard pronuclear injection into C57BL/6J mice. Ogt-Tg mice were genotyped by PCR using a forward primer (5′-GTATCATTTTCTCACCTGTGGCTCCT-3′) specific for the inserted mouse Ogt cDNA sequence and the reverse primer (5′-GCAGAGGGAAAAAGATCTCAGTGGTA-3′) specific for the pCAGGS vector region. All mice were maintained in a laminar air-flow cabinet in a barrier facility. All studies on animal models were approved by the Ethical Committee of the Osaka Medical College and performed according to its guidelines.

B16 mouse melanoma cells were dissociated with Trypsin (0.05%)/EDTA (0.02%) and single-cell suspensions in sterile PBS were prepared. The viability of the cells was checked before injection (>95%). Under 2,2,2-tribromoethanol (Sigma-Aldrich Japan) anesthesia, 10- to 14-week-old WT and Ogt-Tg/+ female mice (C57BL/6J) were injected subcutaneously with 25,000 cells/40 μL B16 melanoma cells using a 27-gauge needle. Ten days after transplantation, the growing tumors were excised and tumor weights were measured.

Primary macrophages were harvested from WT mice and Otg-Tg/+ mice by intraperitoneal lavage in RPMI1640 medium (Life Technologies Japan) with 1% FBS, 72 hours after injection of 1 mL of 3% thioglycollate medium (BD Difco). The isolated peritoneal cells were filtered with a 70-μm mesh nylon gauze filter to remove some large particles. Red blood cells were removed by brief hypotonic shock. After washing, the isolated cells were cultured in 1% FBS-containing RPMI1640 medium for 2 hours to allow peritoneal macrophages to attach to the culture dish, and nonadherent cells were washed out. Attached macrophages were subsequently cultured for 2 hours in the presence of lipopolysaccharide (LPS, 1 μg/mL; Sigma-Aldrich Japan) and IFNγ (20 ng/mL; PeproTech). After being washed with PBS three times, these macrophages were lysed for protein or RNA extraction, and used for Western blot analysis and real-time PCR analysis, respectively.

Isolated normal tissues and tumor tissues were homogenized using a Teflon-stick homogenizer in ice-cold lysis buffer (50 mmol/L Tris-HCl, pH 7.4, 150 mmol/L NaCl, 10% glycerol, 1% NP-40, 1 mmol/L PMSF, 20 mmol/L NaF, 10 mmol/L Na₄P₂O₇, 2 mmol/L Na₃VO₄) supplemented with EDTA-free protease inhibitor cocktail (Calbiochem) and lysed on ice for 30 minutes. To separate nucleus from cytosol, the lysates were centrifuged at 12,000 × g for 20 minutes, the supernatants were considered to be the cytosolic fraction which was free of nuclear markers. The pellets were resuspended in the SDS-PAGE sample buffer and used for nuclear protein–enriched fraction. B16 melanoma cells and isolated peritoneal macrophages cultured in vitro were harvested at subconfluency and lysed in SDS-PAGE sample buffer. These lysates were aliquoted and stored at −80°C until used. Protein concentrations were determined using the Pierce BCA Protein Assay (Thermo Fisher Scientific) or by Coomassie brilliant blue-staining protein spot of each specimens with a standard concentration of protein. Equivalent amounts (1 or 5 μg) of protein were loaded for each condition and resolved by SDS-PAGE, electrophoretically transferred to PVDF membrane (Merck Millipore), blocked with 5% skim milk, probed with antibodies, and developed using a Luminata Western HRP substrates (Merck Millipore). Signals were detected and documented with the densitometry system LAS3000 (Fujifilm) or Fusion FX7 (Vilber Lourmat).

B16 melanoma cells were cultured in the presence or absence of SB203580 (5 μmol/L) for 24 hours. After the pretreatment, cells were trypsinized and seeded on 96-well culture plates at a density of 2,500 cells per well. The cells were cultured in the presence of DMSO (Nacalai Tesque), SB203580 (5 μmol/L), and/or PD98059 (20 μg/mL) for 72 hours. The proliferation assay was performed using a disulfonated tetrazolium salt, WST-8 (Dojindo), following the manufacturer's instructions. Briefly, at the indicated time points, WST-8 was added to each well and the cells were incubated for 3 hours at 37°C. The absorbance values for each well were measured at 450 nm on the iMark Microplate Reader (Bio-Rad Laboratories). Representative results from three independent experiments are shown as mean ± SEM of triplicate wells.

Tumor tissues were harvested, embedded in OCT compound (Tissue-Tek, Sakura Finetek), and snap frozen in liquid nitrogen. Tissue sections (10-μm thickness) of tumor tissues were prepared using a cryostat (Leica CM3050S, Leica Microsystems Japan) and the sections were fixed with 4% paraformaldehyde for 30 minutes. After blocking with 3% BSA in PBS, the sections were stained with primary antibodies overnight at 4°C, washed in PBS three times, and incubated with fluorescence-conjugated secondary antibodies for 1 hour at room temperature. DAPI counterstaining was performed. Stained tumor sections were examined using a fluorescence microscope (BZ-X710, Keyence).

Isolated tumor tissues, peritoneal macrophages, and cultured B16 melanoma cells were homogenized using MagNA Lyser Instrument (Roche) in RNA lysis buffer. Isolation of total RNA was performed using NucleoSpin RNA (Takara Bio Inc.) according to the manufacturer's protocol. Total RNA was reverse transcribed into cDNA using PrimeScript RT reagent Kit (Takara Bio Inc.) following the manufacturer's protocol. Quantitative real-time PCR was performed using SYBR premix Ex Taq II (Takara Bio Inc.) and Thermal Cycler Dice Real-Time System Single TP870 (Takara Bio Inc.) using the following conditions: 40 cycles of two-step PCR (95°C for 5 seconds, 60°C for 30 seconds). Primers listed in Supplementary Table S1 were designed using Primer3. The relative gene expression was calculated using the 2^−ΔΔC_t method with β-actin as the reference gene. Results are expressed as the fold change relative to the gene expression measured in controls.

Values are presented as mean ± SEM. Differences between two groups were analyzed by the unpaired two-tailed Student t test. P values are indicated as follows (***, P < 0.001; **, P < 0.01; and *, P < 0.05).

O-GlcNAcylation is elevated in patients with diabetes mellitus and upregulated in many solid tumors. To investigate the effect of O-GlcNAcylation in the TME on tumor progression, we generated Ogt-Tg mice exhibiting systemically elevated O-GlcNAcylation. Ogt-Tg/+ mice were born at the expected Mendelian ratio (Supplementary Table S2) and developed normally to adulthood with no difference in body weight compared with that of WT C57BL/6J mice (Supplementary Table S3). First, we confirmed that Ogt-Tg/+ mice showed increased O-GlcNAcylation in various tissues, resulting from OGT overexpression by Western blot analysis (Supplementary Fig. S1). We used Ogt-Tg/+ mice exhibiting systemic elevation of O-GlcNAcylation in the following experiments.

To determine the effect of O-GlcNAcylation in the TME on tumor growth, we performed subcutaneous transplantation experiments using Ogt-Tg/+ and WT C57BL/6J mice. Ogt-Tg/+ mice showed elevated O-GlcNAcylation in back skin area in which tumor cells should be transplantated (Supplementary Fig. S1B). B16 melanoma cells were subcutaneously injected and tumor growth was analyzed 10 days after transplantation. Strikingly, tumors from Ogt-Tg/+ mice were larger than those from WT mice (Fig. 1A and B). Consistent with tumor size, the weight of tumors from Ogt-Tg/+ mice (66.0 ± 10.8 mg) was significantly higher than that of tumors from WT mice (25.7 ± 3.8 mg; Fig. 1C). Western blot analysis of these tumor tissues showed that OGT expression and O-GlcNAcylation level in the tumors from Ogt-Tg/+ mice were also significantly higher than controls (Fig. 1D). These data indicate that elevated O-GlcNAcylation in the TME can promote the growth of transplanted melanoma cells.

To identify the signaling pathways involved in cancer cell proliferation that are affected by high O-GlcNAcylation in the microenvironment, we isolated B16 tumor tissues from WT and Ogt-Tg/+ mice at 10 days after subcutaneous transplantation of B16 melanoma cells and performed Western blot analyses using the indicated antibodies (Fig. 2). As shown in Fig. 2A, the expression of phosphorylated MEK1/2, ERK1/2, and AKT was significantly higher in tumor tissues from Ogt-Tg/+ mice than in those from WT mice. In addition, B-RAF, C-RAF, and SRC tyrosine kinases, upstream regulators, were also more phosphorylated in Ogt-Tg/+ mice (Fig. 2B). These data show that the RAF/MEK/ERK and AKT pathways are upregulated in B16 melanoma cells transplanted in Ogt-Tg/+ mice.

To determine the mechanism underlying the upregulation of the MEK/ERK and AKT pathways, we analyzed the activity of p38 MAPK in the tumor tissues transplanted in Ogt-Tg/+ mice. p38 MAPK has been shown to control gene expression, cell growth, and apoptosis in response to a wide range of stimuli such as environmental stresses, inflammatory cytokines, growth factors, and extracellular matrix (27). As shown in Fig. 3A, the expression of phosphorylated p38 MAPK was significantly lower in the tumor tissues transplanted in Ogt-Tg/+ mice than in those transplanted in WT mice. Consistently, immunofluorescence analysis indicated that the expression of phosphorylated p38 MAPK was significantly reduced in tumor tissues from Ogt-Tg/+ mice (Fig. 3B).

As previous reports showed that p38 MAPK inactivates MEK1/2 and ERK1/2 and reduces the proliferation in various cancer cell lines (28), we examined whether the inhibition of p38 MAPK activity can upregulate the proliferation of B16 melanoma cells through the activation of MEK1/2 and ERK1/2 (Fig. 4). In cultured B16 melanoma cells, treatment with SB203580, a p38 MAPK-specific inhibitor, increased the phosphorylation of MEK1/2 and ERK1/2 (Fig. 4A), indicating that the activation of MEK/ERK pathway in B16 tumor tissues in Ogt-Tg/+ mice might be partly due to the inactivation of p38 MAPK. In contrast, the phosphorylation of AKT, B-RAF, and C-RAF were not affected by SB203580 treatment. As the RAF/MEK/ERK and AKT pathways were upregulated in B16 melanoma cells transplanted in Ogt-Tg/+ mice (Fig. 2), the upstream signal should also be involved in the activation of the pathways in the tumor tissues.

Next, we examined the effect of the reduction of p38 MAPK activity on the proliferation of cultured B16 melanoma cells (Fig. 4B). According to the in vitro proliferation assay, SB203580 treatment significantly promoted the proliferation of B16 melanoma cells, and the increase was inhibited by a MEK1/2-specific inhibitor, PD98059. These data indicate that reduction of p38 MAPK activation in B16 melanoma cells promotes its cellular proliferation through activation of the MEK/ERK pathway. p38 MAPK is also involved in apoptotic signaling in several cancer cells. To examine the effect of p38 MAPK downregulation on apoptotic signaling, the expression levels of apoptosis-related genes, Bax and Bcl-2, were determined in B16 melanoma cells cultured in vitro by real-time PCR analyses. The apoptotic signals were not changed (Fig. 4C). In addition, we demonstrated that the Bax/Bcl-2 ratio was not changed (Supplementary Fig. S2A) and cleaved caspase-3 was not detected in tumor tissues excised from both WT and Ogt-Tg/+ mice (Supplementary Fig. S2B). Taken together, these data indicate that the promotion of tumor growth in Ogt-Tg/+ mice may mainly depend on an increase of signaling pathways underlying cell growth rather than on the reduction of proapoptotic signaling.

Accumulating evidence revealed that the crosstalk between p38 MAPK and NF-κB transcription factor also plays crucial roles in inflammation and cancer progression (29). In addition, previous reports showed that O-GlcNAcylation of NF-κB p65 on Ser 536 increases NF-κB binding to IκB, resulting in the reduction of NF-κB function (30). Therefore, we examined the activation of NF-κB in B16 tumor tissues by Western blot analysis. In tumor tissues from Ogt-Tg/+ mice, the expression of phosphorylated NF-κB p65 (Ser 536) and the nuclear distribution of NF-κB p65 were significantly lower than that in WT mice (Fig. 5A), indicating that NF-κB was inactivated in tumor tissues from Ogt-Tg/+ mice.

Because NF-κB plays an essential role in inflammation and cancer progression (29), the reduction of NF-κB activation in B16 tumor tissues from Ogt-Tg/+ mice suggests the possibility that the inflammatory response may be different between WT and Ogt-Tg/+ mice. To examine the inflammatory response in our transplantation model system, the expression levels of proinflammatory cytokines in the tumor tissues were determined using real-time PCR analysis. As expected, the expression of IL1β and TNFα was significantly downregulated, but not that of VEGF-A in tumor tissues from Ogt-Tg/+ mice (Fig. 5B). On the other hand, the serum levels of these cytokines were extremely low or under detection limit in both of WT and Ogt-Tg/+ mice according to ELISA analysis (data not shown). We speculated that the cellular source of proinflammatory cytokines in the tumor tissues was B16 melanoma cells themselves or the tumor surrounding tissues. In B16 melanoma cells cultured in vitro, SB203580 treatment had no effect on the phosphorylation of NF-κB p65 (Ser 536) and IκB (Ser 32/36; Fig. 5C) and IL1β and TNFα gene expression was undetectable (Fig. 5D), indicating that B16 melanoma cells were not the major cellular source of proinflammatory cytokines in the tumor tissues.

The infiltrated macrophage surrounding tumor is thought to be the main source of proinflammatory IL1β and TNFα (31), and the production of these cytokines requires NF-κB activation and its nuclear localization. To clarify the NF-κB activity in the macrophages in Ogt-Tg/+ mice, we examined the phosphorylation level of NF-κB and proinflammatory cytokine production in the thioglycollate-elicited peritoneal macrophages by Western blot and real-time PCR analyses, respectively. As shown in Fig. 5E, the expression of phosphorylated NF-κB p65 (Ser 536) in the macrophages stimulated with thioglycollate from Ogt-Tg/+ mice was significantly lower than that from controls. Certainly, O-GlcNAcylation level in the macrophages from Ogt-Tg/+ mice was significantly higher than controls. Consistent with the NF-κB activity, the proinflammatory cytokine production of the macrophages from Ogt-Tg/+ mice was significantly lower than that from controls (Fig. 5F). Moreover, we estimated the cellular localization of NF-κB p65 in the infiltrated macrophages in the B16 tumor tissues by immunofluorescence (Fig. 6). As a result, it was observed in a lot of cells including macrophages in WT mice, whereas the nuclear localization was reduced in the macrophages which were highly O-GlcNAcylated in Ogt-Tg/+ mice. These data indicate that NF-κB activation in the macrophages was downregulated in B16 tumors grown in Ogt-Tg/+ mice.

Next, we examined the number of infiltrated macrophages in the tumor tissues by immunofluorescence (Fig. 7A). F4/80 staining showed that the distribution of F4/80⁺ macrophages was significantly decreased in tumor tissues from Ogt-Tg/+ mice compared with WT mice. On the other hand, the distribution of CD3⁺ T lymphocytes and B220⁺ B lymphocytes in the tumor tissues was apparently not changed between WT and Ogt-Tg/+ mice (Fig. 7B). α-SMA⁺CD31⁻ cells, which were thought to be cancer associated fibroblasts, were merely detected in the tumor tissues from both of WT and Ogt-Tg/+ mice in our transplantation model system (Fig. 7C). CD31 staining showed that there were no significant differences on angiogenesis in the tumor tissues between WT and Ogt-Tg/+ mice (Fig. 7D).

In general, macrophages are divided broadly into two categories: classical M1 (antitumor functions) and alternative M2 (protumorigenic functions) macrophages, based on their function. To determine the cell type of the macrophages infiltrated into the B16 tumor tissues in our experimental model, we examined the expression of the macrophage marker by immunofluorescence using antibodies for iNOS (a M1 marker) and arginase-1 (a M2 marker). As shown in Fig. 7E and F, almost all F4/80⁺ cells were positive for iNOS in the tumor from both of WT and Ogt-Tg/+ mice, whereas, arginase-1⁺ cells were undetectable in F4/80⁺ cells in both tumors. These data suggest that the downregulation of NF-κB activation in the M1-polarized macrophages and the reduction in M1-polarized macrophage infiltration in tumor tissue may cause the decreased proinflammatory cytokines in tumor tissues grown in Ogt-Tg/+ mice.

O-GlcNAcylation is increased in various cancers and the augmented O-GlcNAcylation in cancer cells accelerates tumor growth, progression, and metastasis (18, 19, 21). A recent report indicates that O-GlcNAcylation in cancer cells is also involved in tumorigenesis (32). O-GlcNAcylation is elevated in patients with diabetes mellitus and may aggravate tumor progression. In diabetic conditions, O-GlcNAcylation is elevated not only in cancer cells, but also in the TME. However, the role of O-GlcNAcylation in the TME is not fully understood. In this study, we examined the role of O-GlcNAcylation in the TME on cancer progression using a transplantation model system in which melanoma cells were transplanted into Ogt-Tg mice exhibiting systemically elevated O-GlcNAcylation and clarified accelerated proliferation of the transplanted tumor cells and the mechanism by which O-GlcNAcylation in the TME promotes cancer progression.

In our transplantation model system, p38 MAPK was downregulated in the tumor tissues from Ogt-Tg/+ mice (Fig. 3), although the other MAPK and AKT pathways were upregulated (Fig. 2). Given that p38 MAPK not only exhibits proapoptotic function, but also negatively regulates cell proliferation (33, 34), we postulated that the inactivation of p38 MAPK may be the key phenomenon in melanoma cells transplanted in Ogt-Tg/+ mice. In various cancer cell lines, modulation of the ERK/p38 MAPK activity ratio (ERK/p38 ratio) by multiple pharmacologic and genetic interventions showed that high ERK/p38 ratio favors tumor growth in vivo and in vitro (28). Consistent with previous reports, we revealed that MEK1/2 and ERK1/2 were upregulated by p38 MAPK inhibition in cultured B16 melanoma cells (Fig. 4). Indeed, our data suggest that the growth of the transplanted melanoma cells was accelerated by ERK1/2 activation due to the downregulation of p38 MAPK, which exerts a negative feedback on ERK1/2, in Ogt-Tg/+ mice. BRAF or NRAS mutation occurs in about 59% and 28% of human melanomas, respectively (35, 36), and human melanoma cell lines with mutations in BRAF or NRAS lose their negative feedback function (37). Because the B16 melanoma cells used in our experiments are B-RAF and N-RAS wild-type, this negative feedback system might be working.

We further examined how p38 MAPK in the transplanted melanoma cells was inactivated in Ogt-Tg/+ mice. p38 MAPK is a class of MAPK that is responsive to stress stimuli such as cytokines, ultraviolet irradiation, heat shock, and osmotic shock. Assuming that the responsible stress stimuli were inflammatory cytokines, major cytokines in the tumor tissue were evaluated. As speculated, their expression was significantly reduced (Fig. 5B), suggesting that the low activity of p38 MAPK might be due to the reduction of the expression of inflammatory cytokines. Thus, these data suggest that p38 MAPK is inactivated in B16 tumors grown in Ogt-Tg/+ mice through downregulation of inflammatory cytokine from the tumor surrounding tissues, resulting in upregulation of B16 cell proliferation partly via ERK1/2 activation. Although p38 MAPK is known to function as a proapoptotic factor, we observed no effect on apoptosis in B16 tumor tissues from Ogt-Tg/+ mice (Supplementary Fig. S2). Taken together, the upregulation of B16 cell proliferation by p38 MAPK inactivation in the tumor tissues may be caused by the ERK1/2 activation rather than by the deceleration of tumor cell apoptosis.

The role of O-GlcNAcylation on inflammation remains contradictory (38). O-GlcNAcylation has been shown to increase NF-κB transcriptional activity through some mechanisms. However, in some conditions such as acute vascular injury, O-GlcNAcylation also exerts anti-inflammatory effects via inhibition of the NF-κB pathway, suggesting a complex regulation of inflammation by O-GlcNAcylation. On the basis of these previous reports, the downregulation of NF-κB pathway in the tumors grown in Ogt-Tg/+ mice (Fig. 5A) suggested that the reduction in the expression of inflammatory cytokines might to be due to inhibition of the NF-κB pathway by O-GlcNAcylation. It could downregulate p38 MAPK, resulting in ERK1/2 activation, a potent growth engine, in the tumors grown in Ogt-Tg/+ mice.

Macrophages play central roles in acute and chronic inflammatory processes and are thought to be the major source of inflammatory cytokines (31). O-GlcNAcylation is elevated in patients with diabetes which is associated with a low-grade chronic inflammation (38) and may effect on macrophage functions (39). However, the inflammatory response of macrophages in patients with diabetes also remains contradictory. In vivo hyperglycemia may have complex effects on macrophage functions, depending on their tissue of origin and on the duration of diabetes. Increased production of proinflammatory factors was observed in peritoneal macrophages from mice two weeks after diabetes induction with alloxan or streptozotocin (40). On the other hand, TNF-α and IL-6 production in response to lipopolysaccharide stimulation was downregulated in peritoneal macrophages from mice 4 months after diabetes induction with streptozotocin (41). Moreover, the inflammatory response to multiple toll-like receptor ligands was impaired in alveolar macrophages from mice 2 weeks after diabetes induction with streptozotocin (42). Because O-GlcNAcylation is augmented in diabetic patients, we tried to reveal whether O-GlcNAcylation in macrophages may affect to the inflammatory response. As shown in Fig. 6, the activation of NF-κB was downregulated in the highly O-GlcNAcylated macrophages in the tumor tissues grown in Ogt-Tg/+ mice. Because NF-κB was inhibited through the O-GlcNAcylation of NF-κB p65 on its Ser 536 (30), O-GlcNAcylation might mediate NF-κB inactivation in macrophages resulting in impaired inflammation around the tumor tissues and increased proliferation of tumor cells in Ogt-Tg/+ mice. A lot of previous reports show that tumor development requires chronic inflammation and cancer progression also require chronic inflammation. On the other hand, it is also known that in certain contexts, proinflammatory cytokines, such as TNF-α secreted from macrophages, suppress cancer cell proliferation and progression (43–45). To elucidate how O-GlcNAcylation regulates tumor progression, further investigation should be performed about the effect of O-GlcNAcylation on inflammatory responses of respective tumor-associated cells including macrophages, lymphocytes, and fibroblasts during various stages of tumor formation.

We showed the reduction of infiltrated F4/80⁺ macrophages into the tumor from Ogt-Tg/+ mice (Fig. 7A). Macrophage recruitment into tumor tissues is regulated by the interactions between macrophages and multiple microenvironmental cues such as cytokines, chemokines, and extracellular matrix components (46). NF-κB is partially involved in the signaling of the interactions to regulate the macrophage infiltration into tumor tissues through regulating the cytokines/chemokines production and the cell movement. The downregulation of NF-κB activity was observed in the tumor tissues and macrophages from Ogt-Tg/+ mice (Fig. 5A and E). It seemed to be a cause of the reduction of infiltrated F4/80⁺ macrophages into the tumor from Ogt-Tg/+ mice. To clarify the molecular mechanisms, further detail investigation is needed about the effect of O-GlcNAcylation on macrophage infiltration into tumor tissues.

Given that the morbidity and mortality rates for various cancers are significantly higher in patients with diabetes than in patients without diabetes (47–49), it is likely that O-GlcNAcylation, not only within cancer cells, but also in the TME, may affect cancer development or progression in diabetic patients. The strategy to reduce O-GlcNAcylation in the TME as well as in cancer cells might become a substitution therapy in addition to anticancer drugs, especially for cancers where the morbidity and mortality rate is said to be higher in patients with diabetes.

No potential conflicts of interest were disclosed.

Conception and design: K. Moriwaki, M. Asahi

Development of methodology: K. Moriwaki

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Moriwaki

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Moriwaki, M. Asahi

Writing, review, and/or revision of the manuscript: K. Moriwaki, M. Asahi

Study supervision: M. Asahi

The authors thank Ms Yoshiko Takeshige, Naoko Segawa, and Nozomi Tokuhara for their excellent experimental support.

This study was supported, in part, by a Grant-in-Aid for Scientific Research (C) no.17590249 (to M. Asahi) from the Japan Society for the Promotion of Science by the Ministry of Education, Science, Culture, Sports and Technology, Japan.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.