Creative Touch Permanent Makeup and Aesthetics

Insulin signalling begins at the insulin receptor (IR) – a heterotetrameric receptor tyrosine kinase expressed on hepatocytes, myocytes, adipocytes, keratinocytes, sebocytes, and vascular endothelial cells – and bifurcates immediately downstream into two structurally and functionally distinct pathways. The PI3K/AKT branch, activated primarily via IRS-1 and IRS-2 scaffold proteins, mediates the metabolic actions of insulin: GLUT4 translocation, glycogen synthesis, suppression of gluconeogenesis, and lipogenesis. The MAPK/ERK branch, activated via the Shc adaptor protein, mediates insulin’s proliferative and growth actions: gene transcription, cell survival, and differentiation. In insulin resistance, PI3K/AKT signalling is selectively impaired by DAG/PKC-mediated insulin receptor kinase phosphorylation and ceramide/PP2A-mediated AKT dephosphorylation, whilst MAPK/ERK signalling is relatively preserved. The practical consequence – particularly in vascular endothelium and skin – is that hyperinsulinaemia drives the intact proliferative branch with a chronically elevated signal, producing vasoconstriction, keratinocyte hyperproliferation, and fibroblast overstimulation in tissues where the metabolic branch is no longer adequately suppressing glucose production or enabling disposal. A self-limiting negative feedback loop – mTORC1/S6K1 → IRS-1 serine phosphorylation – modulates the PI3K/AKT branch under nutrient-replete conditions; its constitutive activation in chronic overnutrition contributes independently to the PI3K/AKT arm’s progressive failure.

Insulin signalling is not a single linear pathway from receptor to biological outcome – it is a branched network whose two primary arms are anatomically separated at the receptor, serve different biological functions, and fail differently under metabolic stress.

Understanding the architecture of the cascade is prerequisite to understanding why insulin resistance produces the specific pattern of tissue consequences it does, rather than simply reducing all insulin actions proportionally.

The Insulin Receptor: Architecture Before Signalling

The insulin receptor is a disulfide-linked heterotetrameric glycoprotein – two extracellular α-subunits (ligand binding) connected to two transmembrane β-subunits (tyrosine kinase domains) – that exists in the plasma membrane as a pre-formed dimer. Insulin binds the α-subunits at two partially overlapping sites (Site 1 and Site 2), inducing conformational change that repositions the β-subunit intracellular kinase domains into proximity and triggers trans-autophosphorylation at three tyrosine residues in the activation loop: Y1158, Y1162, and Y1163. Full phosphorylation of all three sites is required for maximal kinase activity. The PKCε-mediated phosphorylation at T1160 – positioned between Y1158 and Y1162 – that characterises hepatic lipotoxic insulin resistance (detailed in the DAG entity) directly impairs this activation loop and explains why IRK kinase activity is reduced without the receptor being structurally absent. ^[2]

IR-A and IR-B isoforms arise from alternative splicing of exon 11 of the INS gene. IR-B (exon 11 included) adds 12 amino acids to the α-subunit, reducing insulin binding affinity slightly but producing a receptor that is selective for insulin and that preferentially couples to the PI3K/AKT metabolic branch. IR-B is the dominant isoform in the canonical metabolic tissues – liver, skeletal muscle, and adipose – and is the primary isoform through which insulin’s fuel-switching and glucose-disposal functions operate. IR-A (exon 11 excluded) has higher affinity for insulin and IGF-2, and couples more readily to the MAPK/ERK proliferative branch. IR-A is expressed in foetal tissue, the central nervous system, haematopoietic cells, and, importantly, in skin – where keratinocytes co-express both isoforms alongside IGF-1R and IR/IGF-1R hybrid receptors. The hybrid receptors – formed when an IR half-receptor pairs with an IGF-1R half-receptor in the membrane – signal predominantly like IGF-1R and respond to IGF-1 at physiological concentrations. ^[1]

The practical consequence of this receptor landscape in skin is that keratinocytes have access to metabolic (IR-B/PI3K/AKT), proliferative (IR-A/MAPK/ERK), and IGF-1-like (hybrid/MAPK) signalling from the same insulin ligand, depending on receptor availability and concentration. This multi-receptor co-expression explains why hyperinsulinaemia in skin produces proliferative effects (via IR-A and hybrids) even as the metabolic PI3K/AKT arm is being progressively impaired.

The PI3K/AKT Branch: Metabolic Execution

The PI3K/AKT branch is activated primarily through IRS-1 and IRS-2 – large scaffold proteins with multiple tyrosine phosphorylation sites that serve as docking platforms for downstream effectors. The sequence: ^[6]

IRS-1 phosphorylation: The activated IRK phosphorylates IRS-1 at multiple tyrosine residues (including Y612 and Y632 in human IRS-1). Tyrosine-phosphorylated IRS-1 exposes SH2 (Src homology 2) binding domains.

PI3K recruitment: The p85 regulatory subunit of class IA PI3K binds tyrosine-phosphorylated IRS-1 via its SH2 domain, recruiting the p110 catalytic subunit to the inner leaflet of the plasma membrane. Activated p110 phosphorylates phosphatidylinositol 4,5-bisphosphate (PIP2) at the 3-position of the inositol ring, generating phosphatidylinositol 3,4,5-trisphosphate (PIP3) at the membrane.

PTEN – the PI3K brake: Phosphatase and tensin homologue (PTEN) dephosphorylates PIP3 back to PIP2, directly antagonising PI3K and acting as the primary off-switch for this branch. PTEN activity determines the amplitude and duration of PI3K signalling. Loss-of-function mutations in PTEN are among the most common events in human cancer – constitutively active PI3K/AKT signalling drives cell survival, proliferation, and metabolic reprogramming independent of insulin. In skin specifically, keratinocyte-specific PTEN deletion in mouse models produces epidermal hyperplasia and increased susceptibility to squamous cell carcinoma – confirming that the PI3K/AKT branch’s growth-restraining function (via normal PTEN activity) has direct skin structural consequences. This is mechanistic context for the insulin signalling/skin cancer relationship, not a claim that hyperinsulinaemia directly causes squamous cell carcinoma.

PDK1 → AKT activation: PIP3 serves as a docking site for pleckstrin homology (PH) domain-containing kinases. PDK1 (phosphoinositide-dependent kinase 1) translocates to the membrane, where it phosphorylates AKT at Thr308 (in the activation loop). mTORC2 (mechanistic target of rapamycin complex 2) phosphorylates AKT at Ser473 (in the hydrophobic motif) – both phosphorylations are required for full AKT activation.

AKT isoform specificity: Three AKT isoforms exist in mammals, each with distinct tissue expression and subcellular targeting that determines substrate access: ^[11]

AKT2 is the dominant isoform in liver, skeletal muscle, and adipose, and its subcellular localisation following insulin stimulation – translocation to the plasma membrane – positions it to access GLUT4/TBC1D4 and FOXO1 substrates. AKT1 remains more cytosolic and nuclear, accessing growth and survival substrates including BAD (apoptosis) and MDM2 (p53 regulation). The spatial segregation of AKT isoforms by subcellular localisation, rather than purely by expression level, is part of how the cell achieves specificity from a shared upstream PI3K/PIP3 signal. ^[4]

Principal AKT2 downstream substrates:

• TBC1D4 (AS160): Phosphorylation at Thr642 inactivates this Rab-GTPase activating protein → GLUT4 storage vesicles remain in Rab-GTP state → vesicle fusion with plasma membrane → glucose uptake in muscle and adipose
• FOXO1: Phosphorylation at Ser256 → nuclear export → FOXO1 cannot activate gluconeogenic gene transcription (PEPCK, G6Pase) → hepatic glucose production suppressed
• GSK3β: Phosphorylation at Ser9 → inactivation → glycogen synthase de-repressed → glycogen synthesis
• mTORC1 (via TSC1/2): AKT phosphorylates TSC2 → releases Rheb → mTORC1 active → protein synthesis, cell growth. Note: this also initiates the S6K negative feedback described below
• eNOS: AKT phosphorylates eNOS at Ser1177 → enhanced NO production → vasodilation in vascular endothelium
• BAD: Phosphorylation at Ser136 → dissociates from Bcl-2 → pro-survival signal

The MAPK/ERK Branch: Proliferative Signalling

The MAPK/ERK branch is activated through a parallel but structurally distinct adaptor – the Shc family of proteins – rather than through IRS-1. This anatomical separation at the receptor level is what allows the two branches to be differentially regulated in insulin resistance. ^[3]

The Shc/Grb2/SOS/Ras cascade: The activated IRK phosphorylates Shc (Src homology and collagen domain protein) at Tyr317. Phospho-Shc recruits Grb2 (growth factor receptor-bound protein 2) via its SH2 domain. Grb2 is constitutively bound to SOS (Son of Sevenless), a guanine nucleotide exchange factor (GEF) for the small GTPase Ras. The Grb2-SOS complex is brought to the plasma membrane by Shc, where it activates Ras by catalysing exchange of GDP for GTP (Ras-GDP → Ras-GTP). Active Ras-GTP recruits and activates Raf (a MAP kinase kinase kinase) → Raf phosphorylates MEK1/2 (MAP2K) → MEK phosphorylates and activates ERK1/2 at Thr202/Tyr204. Active ERK1/2 translocates to the nucleus, phosphorylating transcription factors including Elk-1, c- Fos, and RSK, driving gene programmes for proliferation, differentiation, and survival. ^[5]

IRS-1 itself can also couple to Ras through a PI3K-independent mechanism – adding a second route from IRS-1 to MAPK/ERK beyond the Shc-primary route. However, the Shc pathway is the dominant MAPK/ERK input from the insulin receptor and, critically, is the one that is NOT impaired by IRS-1 serine phosphorylation or PI3K disruption in insulin resistance. When DAG/PKCε impairs IRK at T1160, the reduction in kinase activity preferentially affects IRS-1 tyrosine phosphorylation (PI3K/AKT) more than Shc tyrosine phosphorylation – establishing the molecular basis of differential branch sensitivity. ^[3]

In skin, MAPK/ERK drives: keratinocyte and fibroblast proliferation; keratinocyte migration during wound re-epithelialisation; sebocyte gene expression; hair follicle cycling. The MAPK/ERK pathway in keratinocytes converges with EGF receptor (EGFR) signalling at the Ras/ERK node – insulin, IGF-1, and EGF all activate ERK in keratinocytes, and the proliferative consequences of each are not easily separated at the downstream level.

Selective Insulin Resistance: When One Branch Fails and the Other Does Not

The most clinically consequential feature of insulin signalling architecture is the differential susceptibility of its two branches to lipotoxic impairment. This is not a theoretical inference – it was directly demonstrated in human skeletal muscle biopsies from individuals with type 2 diabetes in a landmark 2000 study published in the Journal of Clinical Investigation (Cusi et al., PMC377440). ^[3]

The study compared insulin-stimulated signalling responses in vastus lateralis biopsies from lean controls, obese non-diabetic individuals, and type 2 diabetic patients during a hyperinsulinaemic euglycaemic clamp. The findings were unambiguous:

• PI3K/AKT branch (IRS-1 tyrosine phosphorylation, PI3K activity, AKT activation): Significantly reduced in obese non-diabetic individuals; further reduced in type 2 diabetes. The metabolic branch was impaired in proportion to the degree of insulin resistance.
• MAPK/ERK branch (Shc tyrosine phosphorylation, Ras activation, ERK1/2 phosphorylation): Fully preserved in obese non-diabetic individuals and in type 2 diabetes. The proliferative branch was unaffected by the same lipotoxic signalling environment that was progressively disabling the metabolic branch.

The mechanistic explanation: IRK-T1160 phosphorylation by PKCε reduces IRK kinase activity, but not uniformly across all substrates. IRS-1 tyrosine phosphorylation requires higher kinase activity (or more sustained kinase engagement) than Shc phosphorylation, making the PI3K/AKT branch disproportionately sensitive to partial IRK impairment. The MAPK/ERK branch, activated through the lower-threshold Shc route, remains functional even when IRK is only partially active.

The vascular endothelium consequence is particularly well-characterised (PMC7642854 and PMC2613319): in vascular endothelium with insulin resistance, the PI3K/AKT → eNOS → NO → vasodilation arm is impaired, whilst the MAPK/ERK → endothelin-1 (ET-1) → vasoconstriction arm is preserved. At hyperinsulinaemic concentrations, the intact MAPK/ERK branch drives ET-1 production whilst NO production falls – producing net vasoconstriction, impaired vasodilatory responses, and endothelial dysfunction. This is the molecular mechanism of the hypertension and microvascular disease that characterise metabolic syndrome: it is not glycaemia driving the vascular pathology, but the branch-selective failure of insulin signalling under chronic hyperinsulinaemia. ^[8]

The skin keratinocyte consequence: In insulin-resistant skin with hyperinsulinaemia, the PI3K/AKT branch (which supports normal IRS-1-mediated keratinocyte differentiation) is progressively impaired, while the MAPK/ERK branch (which drives keratinocyte proliferation) remains active and is driven harder by elevated insulin concentrations. The result is a shift in keratinocyte fate: reduced differentiation signal (PI3K/AKT impaired), elevated proliferation signal (MAPK/ERK intact). This is the molecular-level mechanism underlying acanthosis nigricans’ keratinocyte hyperproliferation, operating through the insulin receptor’s own branched signalling architecture rather than exclusively through IGF-1R cross-activation – although both mechanisms operate simultaneously at hyperinsulinaemic concentrations.

Negative Feedback: mTORC1/S6K1 → IRS-1

The PI3K/AKT branch carries an intrinsic negative feedback mechanism that limits its own activity under nutrient-replete conditions – and that contributes independently to the PI3K/AKT arm’s failure in chronic overnutrition. This mechanism is separate from the DAG/PKCε and ceramide routes of IRS-1 impairment and operates through a distinct serine phosphorylation site. ^[10]

The pathway: Insulin → PI3K/AKT → mTORC1 activation (via AKT phosphorylation of TSC2) → S6K1 (p70 ribosomal S6 kinase 1) activation → S6K1 phosphorylates IRS-1 at Ser1101 (and also Ser636/639) → serine-phosphorylated IRS-1 has reduced capacity for IRK-driven tyrosine phosphorylation and undergoes accelerated proteasomal degradation → PI3K/AKT activation in the next insulin signal is attenuated.

This is physiologically rational: after a sufficient anabolic response to insulin, the cell limits further PI3K/AKT signalling by reducing the availability of IRS-1 as an effective substrate. The feedback couples anabolic sufficiency (via mTORC1 sensing amino acid and energy status) to the insulin signal itself. Under normal conditions it is a proportionate, transient brake.

A 2007 PNAS paper (Tremblay et al.) established that S6K1 phosphorylates IRS-1 at Ser1101 specifically in response to both insulin stimulation and nutrient excess in skeletal muscle, and that this phosphorylation is elevated in obese rodents and associates with impaired insulin action. Notably, Ser1101 is the same residue phosphorylated by PKCθ in the DAG-mediated lipotoxic pathway described in the DAG entity. Two mechanistically distinct pathways – nutrient-sensing (mTORC1/S6K1) and lipotoxic (DAG/PKCθ) – converge on the same IRS-1 serine residue, producing additive impairment of PI3K/AKT signalling in the metabolically overloaded state. ^[10]

In chronic overnutrition: mTORC1 is constitutively active (amino acid excess → mTORC1; glucose excess → AMPK suppressed → mTORC1 de-repressed) → S6K1 constitutively active → IRS-1 Ser1101 chronically phosphorylated → PI3K/AKT arm chronically attenuated → insulin resistance deepened through the feedback mechanism that was designed to prevent excessive signalling, now locked in the permanent “attenuated” state by a nutrient environment that never signals “sufficient.” ^[10]

The mTOR/S6K feedback also explains why protein overconsumption – independent of carbohydrate intake – can contribute to insulin resistance. Branched-chain amino acids (BCAAs, particularly leucine) are potent mTORC1 activators via the Ragulator/GATOR2 sensing complex. Chronically elevated BCAA intake → constitutive mTORC1/S6K1 activation → chronic IRS-1 Ser1101 phosphorylation → PI3K/AKT impairment. This is a mechanistically distinct route from the glucose/insulin/ malonyl-CoA/DAG pathway but arrives at the same IRS-1 impairment endpoint. ^[10]

Insulin Signalling in Skin: The Dual-Branch Context

The skin’s co-expression of IR-A, IR-B, IGF-1R, and hybrid receptors – described in the Insulin entity – means that both PI3K/AKT and MAPK/ERK branches are active in epidermal and dermal cells, each serving distinct biological purposes that are disrupted in different ways by the metabolic and hyperinsulinaemic states. ^[9]

PI3K/AKT in keratinocytes: Normal IRS-1 → PI3K/AKT signalling in keratinocytes supports differentiation, barrier gene expression, and cell survival. The Sadagurski et al. (2007) work established that IRS-1 has a specific and non-redundant role in the keratinocyte differentiation programme – IRS-1 deficiency impairs the organised progression from basal to spinous to granular layers. Chronic IRS-1 serine phosphorylation (from any combination of DAG/PKCθ, S6K1/mTOR, or inflammation/IKKβ) progressively reduces IRS-1’s capacity to support this differentiation signal, tilting keratinocyte biology toward the proliferative phenotype driven by the intact MAPK/ERK branch.

MAPK/ERK in keratinocytes: ERK1/2 signalling drives keratinocyte proliferation and is essential for normal wound re-epithelialisation – keratinocytes at the wound edge activate ERK in response to EGF, HB-EGF, and insulin/IGF-1 signals, providing the proliferative drive for resurfacing the denuded wound bed. The convergence of insulin, IGF-1, and EGF signalling at ERK means that wound healing’s proliferative phase is not dependent on any single growth factor – a design that provides redundancy. However, it also means that the growth-stimulating consequences of hyperinsulinaemia in skin operate through the same ERK node that normal wound healing uses, with no easy way to disentangle the two in chronically hyperinsulinaemic individuals.

The selective resistance consequence in skin: When PI3K/AKT is impaired and MAPK/ERK is intact in insulin-resistant skin – and insulin concentrations are elevated compensatorily – the net signalling output is: reduced differentiation signal, elevated proliferation signal. This manifests as the keratinocyte hyperproliferation and epidermal thickening of acanthosis nigricans, the sebocyte overstimulation contributing to acne, and the fibroblast proliferation seen in skin tag formation. The dermal vasculature’s eNOS/ET-1 imbalance (reduced vasodilatory NO, preserved vasoconstrictive ET-1) impairs skin perfusion and contributes to the microvascular context for poor wound healing. These consequences are mechanistically unified by the selective insulin resistance phenomenon rather than representing separate, unconnected skin findings. ^[8]