O-GlcNAc Signaling: A Metabolic Link Between Diabetes and Cancer?

O-GlcNAc: A metabolic Signaling Molecule

One of the most pressing medical issues facing industrialized countries is the rapid rise of type 2 diabetes, a metabolic disorder characterized by severely elevated blood glucose and insensitivity to insulin. In the United States alone, 7.8% of the population is diagnosed with diabetes, and up to 57 million Americans have pre-diabetes. The cost for treating type 2 diabetes was $174 billion in 2007, and consumed 32% of Medicare spending (http://www.cdc.gov/diabetes/pubs/pdf/ndfs_2007.pdf). Furthermore, diabetes is a risk factor for Alzheimer’s disease, cardiovascular disease, and cancer, but the underlying mechanisms remain poorly understood¹. Both diabetes and cancer show cellular alterations in energy metabolism, stress responsiveness, and signaling. Multiple biochemical processes are involved; however, one potential link in all of these diseases is disrupted O-GlcNAc signaling.

O-GlcNAc is a post-translational protein modification consisting of a single N-acetylglucosamine moiety attached via an O-β-glycosidic linkage to serine and threonine residues^2–4. O-GlcNAc modified proteins are generally either cytoplasmic or nuclear proteins, and unlike asparagine-linked or mucin-type O-glycosylation, O-GlcNAc is not further processed into a complex oligosaccharide^{2, 3}. In many ways, O-GlcNAc is similar to protein phosphorylation; for example the sugar can be attached or removed dynamically in response to changes in the cellular environment triggered by stress, hormones, or nutrients [5,6]. Because O-GlcNAc is attached to serine/threonine residues, the sugar is in direct competition with phosphorylation (eg. c-myc⁷). Furthermore, O-GlcNAcylated or phosphorylated residues can be in close proximity to each other and sterically impair the attachment of the other modification⁵. O-GlcNAcylation is catalyzed by a highly-conserved and unique enzyme, uridine diphospho-N-acetylglucosamine:polypeptide β-N-acetylglucosaminyltransferase (O-GlcNAc transferase, OGT). In cells, the OGT catalytic subunit dynamically forms many specific holoenzyme protein complexes that regulate its specific activity toward its myriad of target protein substrates. By contrast, phosphorylation involves hundreds of individual kinases, each with their own unique activity⁵. Similarly, whereas there exist several protein phosphatases that remove phosphorylation, there is only a single cytosolic or nuclear β-N-acetylglucosaminidase (O-GlcNAcase, OGA) that is also targeted to substrates by forming many transient holoenzyme complexes to remove the sugar moiety⁵.

ATP, the donor substrate for kinases, is present at high cellular concentrations (approximately 100 nmol/g)^{8, 9}. Although ATP is mostly used to provide energy for cellular processes, it also directly links energy metabolism to signaling. Uridine diphosphate N-acetylglucosamine (UDP-GlcNAc), the high energy donor substrate for OGT, sits at the nexus of glucose, nitrogen, fatty acid and nucleic acid metabolic pathways (Figure 1), all of which dynamically influence its cellular concentration. Just like ATP, UDP-GlcNAc is often found at relatively high cellular concentrations (approximately 40 nmol/g in tissue)⁹, and is used not only for O-GlcNAcylation, but also for the biosynthesis of complex extracellular glycans¹⁰. In this review, we highlight the underlying mechanisms behind altered O-GlcNAc signaling in diabetes and cancer, and discuss the latest research linking glucose metabolism to modulation of signaling cascades by O-GlcNAc.

Nutrient Flux through the Hexosamine Biosynthetic Pathway (HBP)

The HBP integrates a variety of metabolic inputs in the synthesis of UDP-GlcNAc (Figure 1). First, approximately 2–3% of total cellular glucose is funneled into the HBP^{5, 10, 11}, although the glucose flux is potentially different in various cell types and we know little about the regulation of the flux of glucose into the HBP. The HBP shares its first two steps with glycolysis; first, hexokinase phosphorylates glucose to produce glucose-6-phosphate, which is then converted into fructose-6-phospate by phosphoglucose isomerase. At this point the pathways diverge, fructose-6-phosphate is converted by the HBP rate-limiting enzyme glutamine: fructose-6-phosphate aminotransferase (GFAT1) into glucosamine-6-phosphate. GFAT1 catalyzes the irreversible transfer of the amino group from glutamine and the isomerization of fructose-6-phosphate into glucosamine-6-phosphate and glutamate¹². Not only is GFAT1 dependent on glucose flux into the cell, it also must integrate signals from amino acid metabolism.

GFAT1 is the rate-limiting step of the HBP due to feedback inhibition by both the enzymatic product glucosamine-6-phosphate and the final product UDP-GlcNAc¹². Intuitively, increased cellular glucose and flux through the HBP would only increase UDP-GlcNAc levels to a specific concentration and no higher. For example, adipocytes exposed to increasing amounts of glucose in the presence of insulin show a 30% increase in UDP-GlcNAc concentration compared to a 365% increase in glucose uptake¹³. Is this slight increase in UDP-GlcNAc concentration enough to allow O-GlcNAc to act as a nutrient sensor? Potentially, the answer is yes because of the unique ability of O-GlcNAc transferase to respond to UDP-GlcNAc concentrations. OGT purifies as a multimer with a low apparent Km for UDP-GlcNAc (545 nM) allowing the enzyme to be active in times of low UDP-GlcNAc concentrations, and allowing the enzyme to out-compete nucleotide transporters in the Golgi and endoplasmic reticulum (ER)¹⁴. However, in vitro, the rate of OGT’s transfer of O-GlcNAc to peptides is directly responsive to an extraordinary range of UDP-GlcNAc concentrations, from low nanomolar to above fifty millimolar. Furthermore, increasing UDP-GlcNAc concentrations change OGT’s apparent Km for peptide substrates, with most peptides becoming better substrates as UDP-GlcNAc increases¹⁵. Even a slight increase in UDP-GlcNAc concentration, due to nutrient excess, causes different proteins to become more extensively O-GlcNAcylated.

Regulation of GFAT1 also occurs at the translational and posttranslational level. Diabetic patients exposed to chronic nutrient excess demonstrate increased GFAT1 activity correlated with increased GFAT1 mRNA levels in lymphocytes¹⁶ and skeletal muscle¹⁷. However, it is not clear if the increase in GFAT1 activity is an adaptive response of the cells caused directly by the nutrient overload, or alternatively if the hyperglycemic environment increases GFAT1 mRNA levels through altered regulation of a signaling pathway. In skeletal and heart muscle, a GFAT1 splice variant, termed GFAT1-L, is expressed¹⁸. GFAT1-L retains full enzymatic activity and has a 54 base pair insertion compared to GFAT1; little is known about its regulation. Additionally, a highly homologous GFAT2 isoform is expressed in heart and nervous tissue¹⁹.

Both GFAT1 and GFAT2 are targets for protein phosphorylation. GFAT1 has two cyclic AMP dependent protein kinase (PKA) sites at serines 205 and 235. Whereas phosphorylation of S205 decreases cellular enzymatic activity, S235 phosphorylation does not affect activity^{20, 21}. GFAT2 has a PKA site at serine 202 homologous to S205 of GFAT1, but PKA phosphorylation of this site increases its cellular activity²². Thus, GFAT2 appears to be regulated differently than GFAT1, possibly mediated via interactions with different proteins. Owing to the different tissue expression of these isoforms, the HBP in these tissues could be regulated differentially by hormonal signals.

GFAT1 activity is directly linked to changes in cellular energy. AMP-activated protein kinase (AMPK) is regulated by the cellular ATP:AMP ratio²³; it is allosterically activated by AMP, which is normally present at low concentrations, although stress, hormones, or energy imbalance can increase AMP concentrations dramatically²³. Upon AMPK activation, GFAT1 and GFAT2 are phosphorylated at S243^{24, 25}. Much like S205 phosphorylation, GFAT1 phosphorylated at S243 in vivo is less active²⁵. Furthermore, activation of AMPK causes an increase in OGT mRNA levels and protein expression²⁶, and AMPK appears to either be O-GlcNAc modified or associated with an O-GlcNAcylated protein²⁷. When GFAT1 is bypassed by glucosamine treatment, which elevates O-GlcNAc levels²⁸, AMPK activity is higher; by contrast, hexosaminidase treatment of AMPK lowers its activity²⁷. Therefore, O-GlcNAcylation of AMPK or an interacting partner potentially increases AMPK activity leading to GFAT1 inhibition and paradoxically stimulating OGT expression. The regulation of the HBP is complex and is sensitive to both nutrient alteration and phosphorylation, ultimately impinging upon O-GlcNAc signaling (Figure 1).

O-GlcNAc Signaling in a Diabetic Background

In normal liver cells exposed to insulin, a complex signal transduction cascade is initiated resulting in altered gene expression and the translocation of the insulin responsive glucose transporter GLUT4 to the plasma membrane²⁹. Insulin binds to the insulin receptor causing the receptor to autophosphorylate, which recruits the insulin receptor substrate (IRS1 or IRS2)²⁹. IRS is phosphorylated at multiple tyrosine residues, activating the docking of the phosphatidylinositol 3-kinase (PI3K) regulatory subunit, p85, to IRS²⁹. Once docked, PI3K catalyzes the production of phosphatidylinositol-3,4,5-triphosphate (PIP3) at the plasma membrane²⁹. PIP3 recruits PIP3-dependent kinase (PDK1) and AKT (protein kinase B)²⁹. PDK1 subsequently activates AKT via phosphorylation at threonine 308. AKT then phosphorylates numerous substrates, including Rab-GTPase activating proteins, which leads to GLUT4 translocation to the membrane²⁹; Forkhead (FOXO) family transcription factors, which become excluded from the nucleus and can no longer activate gluconeogenic genes (Box 1); and glycogen synthase kinase 3β (GSK3β), which in turn signals for increased glycogen production (Figure 2, Panel A).

Not surprisingly, increased flux though the HBP causes insulin resistance. Following a switch from normal to extremely high glucose concentrations, cultured cells show a slight increase in O-GlcNAcylation with some signs of altered insulin signaling³⁰. Glucosamine treatment causes insulin resistance¹⁰. Of course, a major outcome of glucosamine treatment is elevated O-GlcNAc levels³¹. However, exogenous glucosamine impinges on many pathways. Therefore linking glucosamine-induced insulin resistance directly to O-GlcNAc signaling is difficult. One way to avoid the myriad effects of glucosamine or even high glucose is to instead treat cells with pharmacological inhibitors of O-GlcNAcase to increase cellular O-GlcNAcylation. Treating adipocytes in culture with PUGNAc, an O-GlcNAcase inhibitor, increases O-GlcNAc levels and impairs insulin signaling³¹. In the presence of PUGNAc, the activating AKT phosphorylation on Thr308 is reduced, downstream signaling through GSK3β is impaired, and glucose uptake is less efficient^{31, 32}.

Conversely, when adipocytes are treated with a more selective O-GlcNAcase inhibitor, they show no signs of insulin resistance³³. PUGNAc inhibits other hexosaminidases, through off-target effects, potentially altering oligosaccharide structures at the plasma membrane³³. These oligosaccharide structures are critical in transducing extracellular signals. A major caveat in studies using O-GlcNAcase inhibitors to elevate O-GlcNAc is that cells rapidly respond to elevated O-GlcNAc by increasing the expression of O-GlcNAcase. Furthermore, reduction of O-GlcNAc rapidly results in the upregulation of OGT expression³⁴. Cells attempt to maintain normal O-GlcNAc function, but not necessarily the absolute stoichiometry. One way of circumventing the problem of the lack of specificity of pharmacological agents is with genetic manipulation. Transgenic mice over-expressing human OGT in their adipocytes and skeletal/cardiac muscles³⁵ appear to have normal weight, fat pad and muscle histology, as well as fasting glucose levels, but the animals suffer from hyperinsulinemia, lower glucose disposal rates, and a reduction in serum leptin levels³⁵. However, over-expression of O-GlcNAcase or knockdown of OGT did not impair insulin signaling in adipocytes³⁶. Although this study is in contradiction to several other genetic studies, no measurement of OGT protein levels was performed when OGA was over-expressed; likewise OGA levels were not examined when OGT was knocked down. Again, alterations in OGT or OGA protein levels cause a reciprocal change in the other’s expression³⁴, which might partially explain this discrepancy. Current data suggest that altered O-GlcNAcylation per se might not be sufficient to induce profound insulin resistance, and that other unknown factors are involved.

O-GlcNAc Transferase Regulates Insulin Signaling

Studies from several laboratories suggest that O-GlcNAcylation plays a direct role in insulin signaling (for reviews^1,3,10). Under normal cellular conditions, OGT is localized to the nucleus and the perinuclear region with only moderate staining observed in the cytoplasm³⁷. However, upon insulin stimulation a subset of OGT translocates to the plasma membrane^{38, 39}, mediated by a non-canonical PIP3 binding domain on the OGT carboxy-terminus ³⁸. At the membrane, OGT associates with the insulin receptor³⁹ and undergoes tyrosine phosphorylation, thus leading to an increase in activity³⁹. OGT then O-GlcNAcylates several components of the signaling pathway³⁸, including IRS1^{38, 40}. IRS1 is O-GlcNAcylated in a time-dependent manner with maximum O-GlcNAcylation occurring approximately 30 minutes after insulin stimulation. Thereafter, O-GlcNAcylation of IRS1 rapidly declines³⁸.

Mass spectrometry based site-mapping approaches have determined at least 3 O-GlcNAc sites on the IRS1 C-terminus, one of which is close to a putative Src homology domain 2 (SH2) binding domain for the PI3K p85 subunit ^{41, 42}. Increased O-GlcNAc on IRS1 potentially causes a reduction in its interaction with p85 and a dampening of the activating tyrosine phosphorylation at Y608 of IRS1⁴³. O-GlcNAcylation of IRS1 reduces Y608 phosphorylation, and concomitantly promotes IRS1 Ser632 and Ser635 phosphorylation, resulting in attenuated insulin signaling^{38, 43}. OGT trafficking to the plasma membrane requires its PIP3 binding domain and occurs upon PI3K activation³⁸. Mice injected with adenovirus expressing OGT bearing a mutated PIP3-binding domain do not show hepatic insulin resistance. Clearly, IRS1 O-GlcNAcylation plays a role in the normal attenuation of insulin signaling, and might also play a direct role in insulin resistance during nutrient excess, which promotes hyper-O-GlcNAcylation (Figure 2, Panel B).

Other components of the insulin signaling pathway are modified by O-GlcNAc. OGT in vitro O-GlcNAcylates IRS2, PI3K, and PDK⁴³, and AKT is O-GlcNAcylated in vivo after insulin stimulation (Figure 2 Panel B)^{38, 44}. O-GlcNAcylation of AKT lowers phosphorylation on Thr308 and alters downstream signaling^{31, 38}. For example, GLUT4-mediated glucose uptake would decrease; this is part of the normal insulin attenuation process in healthy individuals, but premature dampening of the insulin signal or excessive O-GlcNAcylation could keep blood glucose high and put more stress on the pancreas to produce even more insulin^{31, 38}. How then would chronic hyperglycemia elevate O-GlcNAc levels if GLUT4-mediated glucose uptake is impaired? Passive glucose uptake by the ubiquitous GLUT1 transporter might provide an explanation; however, the mechanism remains unclear.

If too much O-GlcNAc and OGT disrupts insulin signaling, then would the converse of reduced O-GlcNAcylation and/or more O-GlcNAcase promote proper insulin signaling? At least under diabetic conditions, reduced O-GlcNAcylation is beneficial. O-GlcNAcase over-expression relieves hepatic insulin resistance⁴⁵, and rescues the function of diabetic cardiomyocytes⁴⁶. Mexican Americans with a single nucleotide polymorphism within the O-GlcNAcase gene (MGEA5) are at higher risk for diabetes⁴⁷.

Cancer’s Influence on the Hexosamine Biosynthetic Pathway

A physiological hallmark of tumors is the use of aerobic glycolysis (also known as the Warburg effect) instead of oxidative phosphorylation to produce ATP⁴⁸. Aerobic glycolysis is a normal function of rapidly proliferating cells which provides both bioenergetic and biosynthetic needs^{49, 50}. Herein, we will discuss aerobic glycolysis and how it influences the hexosamine biosynthetic pathway, and how disruption of the HBP alters O-GlcNAcylation. Aerobic glycolysis is characterized by the conversion of glucose into pyruvate, which is further metabolized into lactate, generating 2 net molecules of ATP, lactate, which is excreted, and NAD⁺, which feeds back into glycolysis⁵¹. Although aerobic glycolysis is an inefficient process, in a nutrient rich environment, efficient energy production is less important than biosynthesis⁴⁸. Cancerous cells are programmed to rapidly proliferate; therefore, the cell must consume carbon and nitrogen rich nutrients to biosynthesize metabolites needed for cell proliferation. Aerobic glycolysis generates pyruvate, which in turn is used for energy or is fed into the citric acid cycle where the metabolites are used in cataplerotic reactions (those that replenish intermediates of the citric acid cycle)⁴⁸.

Diabetes provides at least some additional mortality risk for all cancers due to the hyperinsulinemic and hyperglycemic state feeding a tumor by promoting insulin and insulin-like growth factor signaling and by providing excessive amounts of glucose¹. This excess glucose can be shunted into the HBP as well as into excess metabolites, such as acetyl-CoA and ribonucleotides (Figure 1). The other major metabolite funneling into the HBP is glutamine. Cancer cells are addicted to glutamine; the rate of glutamine consumption in tumors is 10 fold higher then normal cells^{52, 53}. Glutamine is the main energy source in many cancer cells: first by its conversion to glutamate by glutaminase then into α-ketoglutarate in the mitochondria⁵³; the excess ammonia generated by the reaction is excreted⁵⁴. RNAi-mediated knockdown of glutaminase decreases GFAT1 activity, and the cells experience a slight reduction in O-GlcNAc levels and in OGT O-GlcNAcylation⁵⁵. The loss of glutaminase function dramatically reduces the flux of glutamine to α-ketoglutarate, severely depleting energy and reducing flow into the HBP.

It is unclear if UDP-GlcNAc concentrations are higher in cancer cells. However, N-linked glycosylation, which uses UDP-GlcNAc as a donor substrate, generates larger, more highly branched oligosaccharides in tumor cells, suggesting an increased need for the metabolite⁵⁶. The amount of O-GlcNAc modified protein is lower and O-GlcNAcase activity is higher in solid tumors from breast cancer patients, especially in more aggressive tumors; however, the data set is relatively small, so statistically it is unclear if elevated O-GlcNAcase activity is a marker for metastatic potential⁵⁷. By contrast, tumor cell lines that mimic metastatic tumors show an increase in OGT protein expression and O-GlcNAc suggesting that an increase in O-GlcNAcylation might be beneficial to cancer cells⁵⁸.

O-GlcNAcylation of Oncogenes and Tumor Suppressors

c-Myc is a transcription factor whose expression increases in proliferative cells and regulates genes involved in glycolysis, purine/pryimidine, and lipid metabolism^{59, 60}, as well as genes involved in glutamine metabolism, mitochondria biosynthesis, cell cycle control, and HBP genes ^{60, 61}. C-Myc is O-GlcNAcylated⁶² at threonine 58, which is also a GSK3β phosphorylation site⁷. GSK3β requires a priming phosphorylation site before it can actively phosphorylate most substrates; in the case of c-Myc the priming site is at serine 62 (a proline-directed phosphorylation site often targeted by mitogen-activated protein kinase; MAPK)⁶³. Inhibition of GSK3β by lithium chloride or a substitution of the priming Ser62 to alanine elevates O-GlcNAcylation on Thr58. By contrast, serum stimulation increases phosphorylation at Thr58⁶⁴, an event which promotes c-Myc degradation⁶³. Notably, in Burkett lymphoma the most common mutations in MYC affect Thr58 ⁶³, suggesting that phosphorylation of this site reduces protein stability and leads to c-Myc degradation and a reduction in c-Myc target gene expression⁶³. By contrast, an O-GlcNAc residue at Thr58 would block phosphorylation and potentially stabilize the protein. This mechanism is seen in other transcription factors: the C/EBPβ transcription factor is also O-GlcNAcyalted adjacent to a MAPK priming phosphorylation site. O-GlcNAcylation severely reduces MAPK ability to phosphorylate C/EBPβ and in turn GSK3β’s ability to phosphorylate C/EBPβ⁶⁵. By blocking the phosphorylation on C/EBPβ, transactivation and DNA binding are also reduced causing delays in adipocyte differentiation⁶⁵. O-GlcNAc acts antagonistically to phosphorylation, changing the activity and function of the protein. Potentially, c-Myc O-GlcNAcylation might alter the activity of c-Myc in two ways, either by blocking the Thr58 phosphorylation or by interfering with the priming phosphorylation at Ser62. Clearly, a sophisticated interplay of phosphorylation and O-GlcNAcylation occurs at serine/threonine residues on proteins, but teasing out this relationship is challenging. For example, simply making alanine substitutions at these sites in order to determine phosphate or O-GlcNAc function is not sufficient, because this mutation will directly prevent both modifications. The availability of phosphate mimetics helps to elucidate phosphate functions; unfortunately, there are no amino acid substitutions that can mimic the actions of O-GlcNAc.

This type of interplay occurs with another transcription factor-tumor suppressor protein, p53⁶⁶. The stability of p53 is tightly regulated by phosphorylation. Normally, p53 is kept at low levels in the cells by phosphorylation at Thr155 targeting the protein for proteasomal degradation. However, when Ser149 of p53 is O-GlcNAcylated, the protein is stabilized, while Thr155 phosphorylation is decreased⁶⁶.

Several studies indicate that O-GlcNAcylation regulates the cell cycle. Because Ogt is an essential gene, mouse knockouts are lethal⁶⁷; however, conditional Ogt knockout mouse embryonic fibroblast (MEF) cells show cell cycle defects leading to an increase cyclin-dependent kinase inhibitor p27^Kip1 and subsequent senescence⁶⁸. Additionally, OGT siRNA-mediated knockdown in breast cancer cells recapitulates the data seen in the knockout MEFs⁵⁸. In HeLa (cervical carcinoma) cells synchronized at the G₁/S transition then released into serum, over-expression of either OGT or OGA causes mitotic exit defects³⁴. Interestingly, decreases in O-GlcNAc either from OGA over-expression or inhibition of GFAT1 accelerated S-phase transitions in these cells whereas OGA inhibitor treatment slowed S-phase completion³⁴. These results suggest that O-GlcNAc is required for cell cycle progression, whereas high levels or the complete lack of O-GlcNAcylation activates stress response pathways that arrest the cell cycle checkpoint (Box 3). By example, one clinical manifestation of diabetes is nephropathy. Kidney glomerular mesangial cells respond to chronic hyperglycemia by entering the cell cycle⁶⁹. The cells proliferate quickly through a few cell divisions and then become arrested at G₁/S, an event characterized by hypertrophy and increased expression p27^{Kip1 69}. Eventually the cells lose function and undergo cell death. This same phenotype can be replicated with the use of glucosamine, suggesting that HSP flux leading to O-GlcNAcylation is involved in regulating transitions throughout the cell cycle⁶⁹.

Promoting Aneuploidy Through O-GlcNAc

A common phenotype in cancer cells is aneuploidy (abnormal number of chromosomes), which promotes tumorigenesis⁷⁰. Whereas partial loss of mitotic checkpoint control causes low levels of aneuploidy, complete loss of checkpoint control causes massive aneuploidy and cell death⁷⁰. Over-expression of OGT by only two-fold promotes aneuplody through disruptions in mitotic progression³⁴. Normally, during M phase, a subset of OGT localizes to the mitotic spindle, which then moves through the cleavage furrow into the midbody (Figure 3)^{34, 71}. OGT probably performs multiple functions at the spindle: i) OGT O-GlcNAcylates numerous proteins associated with spindles and midbodies regulating their function (Figure 3); and ii) OGT, probably via its tetratricopeptide repeat (TPR) domains, provides a scaffold to facilitate protein–protein interactions^{34, 72}.

Using a novel mass spectrometric approach to identify O-GlcNAc modified proteins⁷³, over 150 O-GlcNAc sites were found on numerous spindle-midbody proteins from a biochemical preparation of spindle-midbodies. This method, based on affinity enrichment of O-GlcNAc with a photocleavable tag and ETD (Electron Transfer Disassociation) mass spectrometry, mitigates many of the problems found with collision induced dissociation (CID)-based mass spectrometry site mapping. O-GlcNAc sites are difficult to map by CID due to ion suppression by unmodified peptides, preferential loss of O-GlcNAc upon ionization, and low mass to charge ratio⁷³. Upon OGT over-expression, most all O-GlcNAcylation sites identified were increased.

Cyclin-dependent protein kinase 1 (CDK1) is the master regulatory kinase for M phase progression. At M phase, CDK1 binds cyclin B and is dephosphorylated, and therefore activated, at Thr14 and Tyr15 by the duel-specificity phosphatase Cdc25C⁷⁴. CDK1 promotes M phase progression by phosphorylating numerous target substrates⁷⁴. OGT over-expression significantly elevates phosphorylation at Thr14 and Tyr15 and greatly reduces CDK1 activity as judged by downstream phosphorylation events⁷². Moreover, other mitotic kinases interact with OGT. The early M phase kinase polo-like kinase 1 (PLK1) and the late M phase kinase Aurora kinase B interact with OGT; potentially, OGT targets these complexes to specific M phase substrates^{72, 75}. PLK1 phosphorylates and inactivates MYT1, the kinase responsible for the inhibitory CDK1 phosphorylations. OGT over-expression reduces MYT1 phosphorylation, suggesting that OGT regulates PLK1 activity ⁷². These and other studies lead us to conclude that O-GlcNAc is a key regulator of multiple pathways involved in the proper function of M phase progression and that disruption of this pathway might lead to aberrant phenotypes associated with cancer.

Concluding remarks and future perspectives

Obviously, diseases such as diabetes and cancer are increasingly becoming major health risks to industrialized countries. Both diseases present major alterations in metabolism, which will impinge upon and alter O-GlcNAcylation. These alterations to O-GlcNAcylation disrupt cellular signaling cascades and potentially promote the disease state. Understanding the molecular mechanisms in common between these two diseases will provide opportunities for improved treatment. Because aberrant O-GlcNAc signaling is an underlying phenotype shared between these two diseases, a better understanding of O-GlcNAc signaling is critically important. Many questions still remain such as how OGT and OGA target their substrates. Recent work suggests that many proteins can target OGT to specific substrates under different signaling conditions^{26, 76}. How OGT and OGA are targeted to their substrates and what proteins are involved in targeting during diabetes and cancer will provide insights into the molecular mechanism of disease progression and potentially new therapeutic targets.