Hepatic insulin resistance is a key pathological feature of type 2 diabetes, metabolic syndrome, and metabolic dysfunction-associated steatotic liver disease (MASLD). It is characterized by a reduced responsiveness to insulin in hepatocytes and is closely associated with genetic susceptibility, chronic inflammation, and lipotoxicity/glucotoxicity1. However, the intrinsic molecular mechanisms remain poorly defined, largely due to the limited accessibility of human liver tissues. Circulating factors, including cytokines and metabolites, as well as adjoining stellate cells, endothelial cells, and circulating inflammatory cells, can regulate hepatic insulin signaling in vivo, which confounds the investigations of cell-intrinsic mechanisms. In a recent study published in The Journal of Clinical Investigation, Gattu et al.2 developed a human induced pluripotent stem cell (iPSC)-derived hepatocyte model and utilized liquid chromatography with tandem mass spectrometry-based (LC–MS/MS-based) global phosphoproteomics to systematically reveal a novel cell-intrinsic signaling network of hepatic insulin resistance in type 2 diabetes (Figure 1). Under physiological conditions, insulin signals through its receptor to recruit insulin receptor substrates (IRS) 1 and 2, thereby activating the phosphoinositide 3-kinase/protein kinase B (PI3K/AKT) pathway. Activated AKT suppresses forkhead box protein O1 (FOXO1)-mediated gluconeogenesis, inhibits glycogen synthase kinase 3β (GSK3β) to enhance glycogen synthesis, and activates mechanistic target of rapamycin (mTOR) complex 1 to stimulate de novo lipogenesis (DNL), thereby maintaining hepatic and systemic metabolic homeostasis3. In insulin-resistant states, the liver displays a distinctive "selective insulin resistance": the inhibitory effect of insulin on gluconeogenesis is impaired, while its lipogenic action is preserved or even amplified. Traditional perspectives ascribe this phenomenon to the weakening of classical PI3K/AKT signaling; however, this mechanism alone cannot fully explain the sustained activation of DNL. Current hypotheses propose several potential mechanisms: (1) differential subcellular localization and activity of IRS1 and IRS2; (2) distinct sensitivities of the hepatic glucose production and DNL pathways to AKT phosphorylation; and (3) the presence of an independent insulin signaling that specifically regulates DNL4. Due to the difficulty in obtaining primary human hepatocytes (PHHs), these hypotheses have not been validated at the level of intrinsic cellular signaling. To investigate the intrinsic mechanisms of selective insulin resistance in human hepatocytes, Gattu et al.2 enrolled 8 patients with type 2 diabetes and 8 age-matched healthy controls, and successfully generated human iPSC-derived hepatocytes (iHeps) from these individuals using classical four-stage growth factor protocol. The iHeps expressed multiple key hepatocyte markers, including ASGR1 (encoding asialoglycoprotein receptor 1) and ALB (encoding albumin). In type 2 diabetic iHeps, the insulin-mediated transcriptional repression of PCK1 (encoding phosphoenolpyruvate carboxykinase) was impaired, while the basal and insulin-stimulated expression level of FASN (encoding fatty acid synthase) was nearly doubled compared with that in control iHeps, which well reproduced the selective insulin resistance phenotype. In addition, control iHeps displayed robust insulin responses in early insulin signaling events, while these events were significantly downregulated in type 2 diabetic iHeps following insulin stimulation. Subsequently, Gattu et al.2 used LC–MS/MS-based global phosphoproteomics in control and type 2 diabetic iHeps with or without insulin stimulation to define the spectrum of insulin signaling changes. They quantified 21,863 class I phosphosites (which had a localization probability of 75% or higher) and categorized the altered sites in iHeps into the following four groups (wherein the upregulated and downregulated phosphosites were termed class 1 and 2, respectively): These sites exhibited similar insulin-induced phosphorylation change patterns in both control and type 2 diabetic iHeps, reflecting conserved signaling mechanisms. The upregulated phosphorylation sites (class 1A) were enriched in the proteins involved in classical insulin signaling, ribonucleic acid (RNA) metabolism, and multiple Rho guanine nucleotide exchange factors (GEFs) that have been well-studied in adipose and muscle tissues. By contrast, the sites with downregulated phosphorylation (class 2A) were associated with Rho guanosine triphosphatase (GTPase) effectors, vesicle transport, RNA modification, and small ubiquitin-like modifier conjugation. This group included sites in which the effect of insulin stimulation was significantly attenuated in type 2 diabetic iHeps versus control iHeps. Class 1B referred to the sites that exhibited an increased phosphorylation in control iHeps after insulin stimulation, but this increase was blunted in type 2 diabetic iHeps. By contrast, class 2B comprised the sites whose phosphorylation was decreased in control iHeps but showed little change in type 2 diabetic iHeps. Specifically, class 1B phosphosites included the proteins related to the Rho GTPase pathways, lipid metabolism (e.g., ACOX1S26 involved in fatty acid oxidation) and signal transduction regulators such as NDRG2S332, which modulates the activity of phosphatase and tensin (PTEN). Class 2B phosphosites were involved in the Rho GTPase effectors, RNA metabolism and Notch signaling/helix–loop–helix (HLH) transcription pathway, including AUP1S385, which participates in lipid droplet formation and degradation. These emergent phosphorylation sites represent the primary focus and highlight of the study. They generally exhibited little or no response to insulin stimulation in control iHeps, while their phosphorylation levels in type 2 diabetic iHeps were notably altered, with 249 sites upregulated (class 1C) and 128 sites downregulated (class 2C). Class 1C phosphosites primarily mapped to a third subset of the proteins involved in Rho GTPase signaling, RNA metabolism, and mTOR signaling-related proteins. Importantly, IRS2Y675, RPS6KB1S452, and FOXK1S468 have been reported to be insulin-responsive and contribute to the regulatory mechanisms of insulin signaling. Class 2C phosphosites were enriched in another subset of the proteins associated with membrane trafficking and retrograde Golgi transport, including ARID1AS1754, S1755 (which have been confirmed to regulate hepatic fatty acid oxidation), NR3C2S283, S387 (which influences glucose metabolism regulation in skeletal muscle), and FKBP5S13 (a regulator of AKT1 activity). Additionally, the proteins involved in chromatin modification (e.g., KMT2D) also exhibited a decreased phosphorylation in response to insulin stimulation in type 2 diabetic iHeps. This group revealed alterations in basal phosphorylation levels between control and type 2 diabetic iHeps. In the absence of insulin stimulation, type 2 diabetic iHeps displayed increased basal phosphorylation on the proteins involved in the Rho GTPase cycle, cell–cell communication, and vesicle-mediated transport. By contrast, the proteins with reduced basal phosphorylation, except for classical insulin signaling proteins such as IRS2S391, T527, T350, S1100 and FOXO1S287, S329, were mainly enriched in the pathways related to the Rho GTPase cycle and mRNA splicing. The gender of different individuals may affect the pathogenesis of type 2 diabetes and MASLD. In the study, Gattu et al.2 also identified significant sexual dimorphism in iHeps, with 2,680 phosphorylation sites (12.2%) displaying sex-specific differences. Female-dominant sites occurred in the pathways related to the cell cycle, cellular response to stress, and sphingolipid metabolism (e.g., the sphingosine kinase SPHK2T198). Male-dominant sites were enriched in the pathways involved in autophagy, signal transduction e.g., protein kinase C γ (PRKCGT635) and cell junction e.g., Sushi domain-containing 5 protein (SUSD5S105). The exact kinases responsible for many altered phosphorylation sites are unknown, and many kinase-substrate relationships remain undefined. To uncover the molecular basis of IRSs, Gattu et al.2 further identified the dysregulated kinase activities. They integrated the atlas of serine/threonine kinase-substrate specificities with computational modeling to construct a kinase-substrate regulatory network. The results showed that "impaired" insulin signaling in type 2 diabetic iHeps (class 1B/2B) was linked to a reduced activity of kinases, including AKT2, PKCθ, adenosine monophosphate-activated protein kinase catalytic subunit α2 (AMPKA2), and checkpoint kinase 2 (CHK2), while "emergent" insulin signaling (class 1C/2C) was driven by an enhanced activity of Rho-associated coiled-coil-containing protein kinase 1/2 (ROCK1/2), branched-chain α-ketoacid dehydrogenase kinase (BCKDK), and mammalian sterile 20-like kinase 4 (MST4). These observations redefine "selective insulin resistance" as not only a weakening of the classical insulin signaling pathways but also a dynamic network imbalance orchestrated by kinase reprogramming. Notably, functional experiments indicated that the selective ROCK1/2 inhibitor ripasudil could partially restore proximal insulin signaling defects in type 2 diabetic iHeps, highlighting the therapeutic potential of these newly identified kinases. The study innovatively employed iHeps to propose that insulin resistance in type 2 diabetic hepatocytes involved dual insulin signaling defects, namely both impairments in classical insulin signaling pathways and alterations in numerous emergent phosphorylation sites, particularly those associated with non-classical insulin signaling pathways, such as Rho GTPase signaling, RNA metabolism, and chromatin modification. However, the study also has some issues that need to be addressed. First, although iHeps highly express several hepatocyte markers, their phenotype and metabolic functions do not fully reproduce those of PHHs5. Therefore, the phosphorylation signaling networks identified in iHeps require further confirmation in PHHs and/or animal models to clarify their functional roles. Second, the limited samples of control individuals and patients with type 2 diabetes, and the absence of detailed clinical stratification may affect the accuracy of insulin signaling site quantification. Third, the pathophysiological functions of emergent phosphorylation sites, including ROCK1/2 and MST4, in type 2 diabetic iHeps need to be experimentally substantiated. In addition, future studies combining animal models with clinical cohorts will be essential to assess the feasibility and long-term effects of therapeutic interventions selectively targeting the phosphorylation signaling networks. In summary, based on a systematic phosphoproteomic analysis in iHeps, the study has innovatively revealed the dual signaling characteristics of hepatic insulin resistance in patients with type 2 diabetes, namely the co-occurrence of impaired classical insulin signaling and activated emergent insulin signaling (Figure 1). These findings not only extend the understanding of cell-intrinsic insulin signaling regulation but also provide a theoretical framework and potential targets for the treatment of metabolic disorders such as type 2 diabetes and MASLD. Nevertheless, further studies should focus on the functional validation of the phosphorylation signaling networks to advance the clinical translation of precision therapeutic strategies. The authors declare no conflict of interest. Tianpei Hong is an Editorial Board member of the Journal of Diabetes Investigation and a coauthor of this article. To minimize bias, he was excluded from all editorial decision-making related to the acceptance of this article for publication. Approval of the research protocol: N/A. Informed consent: N/A. Registry and the registration no. of the study/trial: N/A. Animal studies: N/A. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Zheng et al. (Thu,) studied this question.