Creative Touch Permanent Makeup and Aesthetics

Diacylglycerol is a glycerolipid intermediate in which two fatty acid chains are esterified to a glycerol backbone, leaving one position unoccupied. Three stereo/regioisomers exist – sn-1,2-DAG, sn-2,3-DAG, and sn-1,3-DAG – but only sn-1,2-DAG at the plasma membrane holds PKC-activating signalling properties and mediates lipid-induced insulin resistance. In normal physiology, plasma membrane sn-1,2-DAG is generated transiently by phospholipase C from PIP₂ hydrolysis, activating PKC briefly as part of receptor signalling cascades; the signal is rapidly terminated by DAG kinase phosphorylation and DGAT re-esterification. In the metabolic impasse created by CPT-1 inhibition under simultaneous glucose and fat excess, DAG accumulates chronically at the plasma membrane from backed-up esterification – activating PKCε in the liver (which phosphorylates insulin receptor kinase at Thr1160, destabilising the activation loop) and PKCθ in skeletal muscle (which phosphorylates IRS-1 at Ser1101). Both impair the insulin signalling cascade upstream of PI3K, explaining the proximal nature of insulin resistance in lipid-overloaded tissue. ^[5]

Diacylglycerol occupies a rare position in biochemistry: the same molecular species serves as both a legitimate physiological signalling molecule and, when its concentration is dysregulated, a potent driver of pathological insulin resistance. The distinction between these two functions is not simply a matter of quantity – it is a matter of stereoisomeric form, subcellular location, and temporal pattern. Understanding DAG’s structure is essential before its metabolic roles can be interpreted accurately. ^[1]

Structure and Stereochemistry: Not All DAG Is Equal

Glycerol has three hydroxyl positions: sn-1 and sn-3 on the outer carbons (primary alcohols), and sn-2 at the centre (secondary alcohol). Diacylglycerol has fatty acids esterified at two of these positions; which two determines its stereo/regioisomeric form. Three distinct forms exist in mammalian cells: ^[1]

• sn-1,2-DAG: Fatty acids at positions 1 and 2; the sn-2 carbon is a chiral centre (the fatty acid at sn-2 is the secondary-alcohol ester). This is the exclusively biologically active stereoisomer – it is the only form that activates novel and conventional protein kinase C isoforms. It is generated by phospholipase C (PLC) from PIP₂ hydrolysis, by the Kennedy (G3P) pathway during de novo lipid synthesis, and by sphingomyelin synthase. It accumulates at the plasma membrane in the lipotoxic state.

• sn-2,3-DAG: The mirror-image enantiomer of sn-1,2-DAG. Generated by adipose triglyceride lipase (ATGL) with co-activator CGI-58. Does not activate novel PKC isoforms. Accumulates in lipid droplet lipolysis but does not drive insulin resistance.

• sn-1,3-DAG (rac-1,3-DAG): Generated by ATGL (without CGI-58) from TAG hydrolysis at the sn-2 position. Can be achiral or chiral depending on fatty acid species. Does not activate novel PKC isoforms.

This stereoisomeric specificity resolves an apparent paradox in the earlier lipotoxicity literature: some experimental models with elevated total cellular DAG showed no insulin resistance, whilst others with similar or lower total DAG showed profound insulin resistance. The lipid droplet-associated DAG pool is predominantly sn-2,3 and sn-1,3 – the non-signalling forms generated by lipolysis – whereas plasma membrane-associated DAG is predominantly sn-1,2. Total cellular DAG measurement conflates these pools; subcellular fractionation with stereoisomer-specific mass spectrometry is required to identify the biologically active fraction. ^[5]

Where DAG Comes From: Physiological Sources

DAG is generated by at least four distinct metabolic routes, each with different stereochemical outcomes and cellular locations: ^[1]

1. PLC-mediated PIP₂ hydrolysis (signalling pool): Phospholipase C cleaves phosphatidylinositol 4,5-bisphosphate (PIP₂) at the plasma membrane, releasing sn-1,2-DAG and inositol trisphosphate (IP₃). Since PIP₂ carries its phosphate at the sn-3 position, PLC-derived DAG is exclusively the sn-1,2 isomer. This is the canonical second-messenger DAG: produced rapidly and transiently following receptor activation (GPCR, receptor tyrosine kinases), activating PKC before being terminated by DAG kinase and DGAT re-esterification. Lifetime is typically seconds to minutes.

2. Kennedy (G3P) pathway (de novo synthesis intermediate): During de novo TAG and phospholipid biosynthesis from glycerol-3-phosphate, phosphatidic acid (PA) is dephosphorylated by lipins (PA phosphatase) to generate sn-1,2-DAG – the direct precursor for both TAG synthesis (via DGAT1/2) and phospholipid synthesis (via CPT). This DAG resides primarily in the ER membrane.

3. Sphingomyelin synthase (SMS pool): SMS1 and SMS2 transfer a phosphocholine headgroup from phosphatidylcholine onto ceramide to produce sphingomyelin and sn-1,2-DAG. SMS1 operates at the trans-Golgi; SMS2 at the plasma membrane. Golgi-derived DAG from SMS1 activates protein kinase D (PKD) to drive secretory vesicle formation. Plasma membrane SMS2-derived DAG may regulate local sphingomyelin replenishment and serves as a site-specific PKC activator.

4. TAG lipolysis (lipid droplet pool): ATGL generates predominantly rac-1,3-DAG from TAG hydrolysis in adipocyte lipid droplets; CGI-58 co-activation allows generation of some sn-2,3-DAG. HSL then hydrolyses the sn-3 position of DAG to produce MAG and free fatty acid. Neither ATGL- nor HSL-derived DAG from lipid droplet lipolysis efficiently activates novel PKC isoforms – explaining why elevated lipid droplet DAG does not consistently track with insulin resistance. ^[1]

The Metabolic Impasse: How Lipotoxic DAG Accumulates

When glucose and fat are simultaneously elevated and CPT-1 is inhibited by malonyl-CoA, incoming dietary fatty acids cannot enter the mitochondria for beta-oxidation and are rerouted to esterification pathways in the cytosol and ER. This esterification process generates sn-1,2-DAG as a mandatory intermediate in the Kennedy pathway – GPAT and AGPAT convert glycerol-3-phosphate and fatty acyl- CoA to lysophosphatidic acid and then phosphatidic acid; lipin dephosphorylates PA to sn-1,2-DAG; DGAT1 or DGAT2 then acylate DAG to TAG for storage. ^[1]

When the rate of fatty acid input exceeds the capacity of DGAT to complete TAG synthesis, sn-1,2-DAG accumulates as a backed-up intermediate. It is rapidly trafficked from the ER to the plasma membrane – confirmed directly by the Lyu et al. (2020, Cell Metabolism) subcellular fractionation experiments, which showed that acute DGAT2 knockdown in rat liver caused ~50% higher sn-1,2-DAG in the ER and ~80% higher sn-1,2-DAG at the plasma membrane. The plasma membrane is where PKCε resides in its inactive, membrane-unbound state and where it is available for DAG-driven translocation and activation. ^[5]

The critical distinction from a physiological standpoint: the same metabolic impasse that keeps malonyl-CoA elevated (CPT-1 inhibited, fatty acids cannot be oxidised) simultaneously generates the substrate for accumulating plasma membrane sn-1,2-DAG via the esterification overflow. Both effects – CPT-1 inhibition and DAG accumulation – stem from the same metabolic miscommunication described in the Randle Cycle page: cells receiving more fuel than they can route, in a state where the metabolic gates are directing one substrate while blocking the other.

Liver: DAG → PKCε → Insulin Receptor Impairment

In hepatic tissue, the primary PKC isoform activated by plasma membrane sn-1,2-DAG accumulation is PKCε – a novel ( calcium-independent, DAG-dependent) PKC isoform. ^[5]

The mechanistic sequence, established by the Shulman group at Yale (Lyu et al., 2020, Cell Metabolism) with subcellular resolution and confirmed in human liver biopsies: ^[5]

1. Plasma membrane sn-1,2-DAG accumulates (from esterification overflow or hepatic de novo lipogenesis under insulin/glucose excess) → 2. PKCε translocates from cytosol to plasma membrane, where DAG activates it → 3. Active PKCε phosphorylates the insulin receptor kinase (IRK) at Threonine 1160 in the activation loop → 4. IRK-T1160 phosphorylation destabilises the activation loop, positioned between the two canonical activation tyrosines Y1158 and Y1162 → 5. IRK-Y1162 autophosphorylation (the kinase’s own activation step) is reduced by >50% → 6. Downstream insulin signalling (Akt-S473, GSK3β-S9, FOXO1-S256) is correspondingly impaired → 7. Hepatic glycogen synthesis falls; endogenous glucose production (EGP) fails to be suppressed by insulin.

The proximal level of impairment is mechanistically significant. Ceramides impair insulin signalling at the level of AKT (distal). DAG/PKCε impairs signalling at the insulin receptor kinase itself (proximal). These are parallel, non-redundant mechanisms – explaining why some clinical scenarios show dissociation between ceramide content and hepatic insulin resistance. The Lyu et al. paper found no consistent ceramide-HIR association in their five-compartment subcellular fractionation across multiple compartments, whilst plasma membrane sn-1,2-DAG tracked tightly with hepatic insulin resistance in both rats and a human NAFLD cohort. ^[5]

The causal relationship was experimentally established through three independent interventions:

• Acute DGAT2 knockdown (raises DAG, induces HIR) – DAG accumulation → HIR
• Liver-specific PKCε knockdown (prevents PKCε activation) – protects against HFD-induced HIR
• Liver-specific constitutively active PKCε overexpression (activates PKCε without DAG) – sufficient to induce HIR independently

PKCε is therefore both necessary and sufficient for hepatic insulin resistance in this model. ^[5]

Mechanistic precision note Previous iterations of the lipotoxicity literature attributed liver insulin resistance to DAG → PKCε → IRS-1 serine phosphorylation, consistent with the muscle mechanism described below. The Lyu et al. (2020) data refines this: in liver, the primary PKCε target is the insulin receptor kinase itself (IRK-T1160), not IRS-1. IRS-1 is substrate of the insulin receptor and cannot be tyrosine-phosphorylated if the receptor is already impaired upstream. The distinction is clinically relevant because it establishes that hepatic lipotoxic insulin resistance is a receptor-level failure, not a post-receptor adapter protein failure – earlier and more fundamental in the signalling cascade.

Skeletal Muscle: DAG → PKCθ → IRS-1 Impairment

In skeletal muscle, the primary PKC isoform implicated in lipid-induced insulin resistance is PKCθ – also a novel, calcium-independent PKC, but distinct from PKCε and expressed predominantly in skeletal and cardiac muscle. ^[7]

The sequence: plasma membrane sn-1,2-DAG accumulation → PKCθ translocation to membrane and activation → PKCθ phosphorylates IRS-1 at Serine 1101 (and to a lesser extent Ser307) → serine phosphorylation of IRS-1 prevents tyrosine phosphorylation at the IRS-1 activation sites normally achieved by the insulin receptor kinase → PI3K cannot be recruited to activated IRS-1 → AKT2 activation (the primary isoform mediating GLUT4 translocation to the plasma membrane) is reduced → skeletal muscle glucose uptake is impaired under insulin stimulation. ^[4]

Human experimental evidence: Itani et al. (Diabetes, 2002) – the first human study demonstrating that acute lipid infusion with heparin (to raise plasma free fatty acids and drive intramyocellular DAG accumulation) increased membrane-associated PKCθ in vastus lateralis muscle biopsies within four hours, coinciding with reduced IRS-1 tyrosine phosphorylation and impaired PI3K activation during a hyperinsulinaemic clamp. The temporal relationship – DAG rise → PKCθ activation → IRS-1 serine phosphorylation → impaired insulin signalling – was observed in real time. ^[2]

DAG and Ceramide: Parallel Lipotoxic Mechanisms

DAG and ceramide both accumulate in the metabolic impasse and both impair insulin signalling – but through distinct molecular targets and at different positions in the signalling cascade. They should not be conflated: ^[3]

DAG acts proximal to the cascade, at the insulin receptor kinase (liver) or IRS-1 adapter (muscle). Its effects are rapid and potentially reversible if the lipid accumulation resolves – PKCε/θ activation is concentration-dependent and terminated when DAG is cleared by DGAT re-esterification or DAG kinase phosphorylation.

Ceramide acts distal to the cascade, primarily at the AKT level (activating PP2A to dephosphorylate AKT, or activating PKCζ to phosphorylate AKT at a non-activation site). Ceramide does not explain proximal defects at the insulin receptor level – confirmed by the Lyu et al. finding that ceramide content did not consistently track with hepatic insulin resistance in their human biopsy cohort, whilst plasma membrane sn-1,2-DAG did. ^[5]

In practice, both accumulate simultaneously in the metabolic impasse, and both contribute to the overall insulin resistance phenotype – operating through mechanistically parallel but non-identical pathways. The ceramide route is covered in the Ceramides entity.

Normal DAG Signalling: Physiological PKC Activation

Before the lipotoxic context makes DAG appear inherently pathological, it is important to establish its legitimate physiological signalling role – because the lipotoxic state represents a hijacking of normal signalling machinery by a chronically elevated version of a normal messenger. ^[1]

In normal receptor signalling: ligand binds → GPCR or RTK activation → PLCβ or PLCγ activated → PIP₂ cleaved → transient sn-1,2-DAG burst at plasma membrane + IP₃ released → IP₃ opens ER calcium channels → cytosolic Ca²⁺ rises → conventional PKC isoforms (requiring Ca²⁺ and DAG: PKCα, β, γ) activated → novel PKC isoforms (requiring DAG only, not Ca²⁺: PKCδ, ε, η, θ) also activated → kinase cascade proceeds → signal terminated within seconds to minutes by DAG kinase (phosphorylating DAG to phosphatidic acid) and DGAT re-esterification. The signal is inherently self-limiting.

Lipotoxic DAG at the plasma membrane replicates the same molecular signal – sn-1,2-DAG at the plasma membrane activating novel PKC isoforms – but without the self-limiting kinetics. The signal is sustained at a chronic elevated level, producing a constitutive PKC activation that would not occur from any normal receptor-ligand interaction. The cell’s insulin-signalling machinery is being chronically impaired by a lipid intermediate whose molecular identity is identical to a normal second messenger but whose context, concentration, and duration are entirely pathological.

DAG in Skin: Keratinocyte Differentiation and the PKCα Axis

In skin, DAG-mediated PKC signalling is a central regulatory mechanism of keratinocyte differentiation, and the isoform most studied in this context is PKCα. ^[6]

A 2017 study (Palazzo et al., Cell Death & Differentiation) identified a DLX3–PKCα regulatory axis in keratinocytes in which PKCα activity is required for DLX3 (a homeobox transcription factor) expression, and DLX3 in turn regulates the PKC-dependent gene programme for terminal differentiation. Topical 12-O-tetradecanoylphorbol-13-acetate (TPA) – a DAG analogue that constitutively activates PKC by mimicking DAG’s lipid binding – strongly upregulated DLX3 and keratinocyte terminal differentiation markers, whilst PKC inhibition reversed these effects. The DLX3cKO skin phenotype – characterised by epidermal differentiation defects, inflammatory cell infiltration, and impaired barrier gene expression – was partially normalised by PKC activity restoration, confirming the DLX3–PKCα axis as the downstream executor of DAG-mediated differentiation signalling in skin. ^[6]

This is physiologically normal DAG signalling: calcium gradients rising across the suprabasal layers of the epidermis trigger PLC-mediated PIP₂ hydrolysis → sn-1,2-DAG transient bursts → PKCα activation → differentiation gene programme proceeds. The same signalling molecule that accumulates pathologically in systemic metabolic impasse is used by keratinocytes as a legitimate developmental signal during their normal differentiation journey. Context determines outcome.

DGAT1 in skin: DGAT1, the enzyme that channels DAG to TAG and thereby clears the lipotoxic pool, is highly expressed in skin and sebaceous glands. DGAT1 deficiency in mice causes sebaceous gland atrophy and a dry, scaly skin phenotype consistent with impaired sebum production – confirming that DAG → TAG esterification via DGAT1 is specifically required for sebum lipid assembly. DGAT1’s role as a DAG clearance enzyme in skin therefore has structural significance beyond metabolic protection. ^[1]