The connection between sugar biology and type 2 diabetes (T2D) has been studied for decades, yet actionable mechanistic details at the level of individual glycan modifications on insulin signaling proteins remain surprisingly incomplete. That gap leads to a major biomedical problem: without knowing precisely which carbohydrate-linked proteins are modified, on which residues, and under what metabolic conditions, intervention strategies are less effective and prone to side effects.
This post covers what chemical biology approaches reveal about glycan modifications and insulin resistance, with a particular focus on O-GlcNAcylation as a regulatory mechanism in insulin signaling pathways. After all, O-GlcNAc is the sugar sensor of the cell! It stands out as a key modulator of insulin sensitivity vs. resistance.
Why Glycan Modifications Matter in Insulin Signaling
Glycosylation is not a passive decoration. Post-translational modifications by monosaccharides — especially O-linked N-acetylglucosamine (O-GlcNAc) — actively regulate protein function. As a key mechanism, O-GlcNAc on serine and threonine sites can compete directly with phosphorylation, the known driver of insulin signal transduction. Several core insulin signaling proteins, including IRS-1, AKT, and PDK1, carry O-GlcNAc modifications at serine and threonine residues that overlap with phosphorylation sites, making these hotspots for how aberrantly high levels of O-GlcNAc reduce insulin sensitivity and activity!
When O-GlcNAcylation occupies those sites, downstream signaling through the PI3K-AKT axis is attenuated. The result is reduced GLUT4 translocation, impaired glucose uptake in skeletal muscle and adipose tissue, and the metabolic phenotype we recognize as insulin resistance.
Hyperglycemia itself increases flux through the hexosamine biosynthetic pathway (HBP), producing more UDP-GlcNAc — the donor substrate for O-GlcNAc transferase (OGT). The result is a feed-forward loop: high glucose increases O-GlcNAcylation, which impairs insulin signaling, which worsens glycemic control.
O-GlcNAcylation as a Metabolic Sensor
OGT and its counterpart OGA (O-GlcNAcase, the single human enzyme that removes these modifications) together control the dynamic cycling of O-GlcNAc on hundreds of nuclear and cytoplasmic proteins. In type 2 diabetes, re-adjusting the balance between their activities is a meaningful medical target.
For decades, research has shown that genetic elevation of OGT activity in mouse models produces insulin resistance. Pharmacological OGT inhibition partially restores insulin sensitivity under hyperglycemic conditions. OGA inhibition, which raises O-GlcNAc levels, can do the opposite…it can exacerbates insulin signaling defects. It should be noted that there is some controversy there, depending on which OGA inhibitor is used – some are safe!
What has been harder to establish is the site-specific picture: which O-GlcNAc modifications on which proteins are functionally consequential, and which are simply bystanders elevated because UDP-GlcNAc is abundant? That is where our chemical biology tools and advanced metabolic disease studies we do in the Fehl Lab become essential.
Chemical Tools Defining the Intervention Points
Metabolic Labeling and Bioorthogonal Chemistry
Metabolic labeling with GlcNAc analogs bearing bioorthogonal handles — such as GalNAz, Ac4GlcNAz, and our PhotoSugar tools — allow us to tag O-GlcNAcylated proteins in living cells, then enrich, identify, and quantify them using click chemistry. We apply these tools in insulin-responsive cell lines like muscle cells or adipocytes under hyperglycemic conditions, to map the O-GlcNAc proteome in a metabolic context.
The limitation is selectivity. Metabolic labeling itself elevates the entire global O-GlcNAc proteome, not the subset of modifications that causally drive insulin resistance. Distinguishing cause from correlation at the proteome level requires additional chemical strategies. We are always seeking to improve our approaches to catch these nuanced details, as an active research area in our lab.
Site-Specific Mutagenesis and Synthetic Glycopeptides
Replacing individual serine or threonine residues on IRS-1 or AKT with alanine, or introducing non-hydrolyzable O-GlcNAc mimetics at defined positions, allows direct testing of whether a specific modification impairs signaling. Synthetic glycopeptides carrying defined O-GlcNAc modifications serve as substrates for kinase assays and antibody development.
This work has confirmed that O-GlcNAcylation at Thr-308 of AKT — a canonical activating phosphorylation site — directly reduces kinase activity. The mechanistic link at this residue is now well-established. Analogous studies on IRS-1 have identified multiple competing glycosylation-phosphorylation sites, though the relative contribution of each in human tissue remains an active question.
OGT and OGA Inhibitor Toolkits
Small molecule inhibitors of OGT and OGA have matured considerably. OSMI-4 and related compounds offer selective OGT inhibition with improved cell permeability over earlier generations, though OGT tools still struggle for activity in live animal models and preclinicla studies. For OGA, Thiamet-G remains widely used, and next generation analogs have been developed to further reduce off-target effects in human clinical trials!
These tools are not therapeutics in their current form (none have been FDA approved yet), but they are essential for establishing causality in cell and animal models. A key methodological point: inhibitor studies work best when paired with proteomics to confirm on-target engagement, and when dose-response relationships are characterized carefully enough to avoid interpreting non-specific toxicity as a glycobiology phenotype.
Machine Learning and Pathway-Level Analysis
One of the persistent challenges in this field is that O-GlcNAcylation affects hundreds of proteins simultaneously. Understanding how those changes collectively alter insulin signaling requires moving beyond single-protein studies to pathway-level analysis.
Machine learning approaches are beginning to address this. Predictive models trained on O-GlcNAc site databases can identify sequence features that make specific residues likely OGT substrates, helping prioritize which modifications to validate experimentally. Network-level analysis of O-GlcNAc proteomics data can identify which modified proteins cluster within the insulin signaling network versus other pathways.
At the Fehl Lab, we integrate machine learning with chemical biology strategies precisely to address this kind of complexity. The goal is not to generate more data — it is to define which glycobiology pathways are actionable and to design chemical tools targeted to those specific nodes. In the context of insulin resistance, that means moving from a global picture of O-GlcNAcylation to a mechanistically grounded map of the modifications that matter most.
Open Questions in 2026
Despite real progress, several questions remain genuinely open.
Tissue specificity. O-GlcNAcylation profiles differ across skeletal muscle, liver, and adipose tissue — all of which contribute to whole-body insulin sensitivity. Most mechanistic studies use cell lines. Understanding which modifications are relevant in primary human tissue, and whether they differ by disease stage, is a priority for the field, and our lab.
Crosstalk with other protein modifications. O-GlcNAcylation is not the only glycan modification relevant to insulin signaling. N-glycosylation of the insulin receptor affects receptor trafficking and ligand binding. Sialylation patterns on cell-surface glycoproteins shift in obesity. How these modifications interact with O-GlcNAcylation at the systems level is largely uncharacterized. We are also actively looking at phospho-GlcNAc crosstalk under insulin resistance conditions via our latest NIH NIDDK-funded grant!
Temporal dynamics. Insulin signaling operates on a timescale of minutes. O-GlcNAc cycling is slower. Understanding how the timing of glycan modifications intersects with the kinetics of insulin signaling is a methodological challenge that current tools only partially address. Our Lab’s PhotoSugars and GlycoID tools were designed with this exact challenge and application in mind, so we are poised to address this challenge.
Intervention selectivity. Even if OGT is a valid target in insulin resistance, systemic inhibition carries real risks — O-GlcNAcylation is essential for normal cellular function across many tissues. Selectively targeting OGT activity toward specific substrate proteins, rather than inhibiting it globally, would require a new generation of chemical tools. This is another key space for the Fehl Lab!
Join our group!
The mechanistic picture of how glycan modifications drive insulin resistance is clearer in 2026 than it was five years ago, but the most important questions are still open. Answering them will require more precise chemical tools and better strategies for interpreting complex glycoproteomics data at the pathway level.
If your research intersects with glycobiology and metabolic disease, contact Charlie Fehl to collaborate at fehl-lab.com.
