The neuropilin-like protein ESDN regulates insulin signaling and sensitivity

Endothelial and smooth muscle cell-derived neuropilin-like protein is a novel partner of insulin receptor, and regulates insulin signaling and subsequently a wide array of insulin effects in cultured cells and in vivo through modulation of insulin receptor interaction with adaptor proteins. This novel mechanism of action may be exploited as target for drug development.

Abstract

Insulin effects on cell metabolism, growth, and survival are mediated by its binding to, and activation of, insulin receptor. With increasing prevalence of insulin resistance and diabetes there is considerable interest in identifying novel regulators of insulin signal transduction. The transmembrane protein endothelial and smooth muscle cell-derived neuropilin-like protein (ESDN) is a novel regulator of vascular remodeling and angiogenesis. Here, we investigate a potential role of ESDN in insulin signaling, demonstrating that Esdn gene deletion promotes insulin-induced vascular smooth muscle cell proliferation and migration. This is associated with enhanced protein kinase B and mitogen-activated protein kinase activation as well as insulin receptor phosphorylation. Likewise, insulin signaling in the liver, muscle, and adipose tissue is enhanced in Esdn^−/− mice, and these animals exhibit improved insulin sensitivity and glucose homeostasis in vivo. The effect of ESDN on insulin signaling is traced back to its interaction with insulin receptor, which alters the receptor interaction with regulatory adaptor protein-E3 ubiquitin ligase pairs, adaptor protein with pleckstrin homology and Src homology 2 domain-c-Cbl and growth factor receptor bound protein 10-neuronal precursor cell-expressed developmentally downregulated 4. In conclusion, our findings establish ESDN as an inhibitor of insulin receptor signal transduction through a novel regulatory mechanism. Loss of ESDN potentiates insulin's metabolic and mitotic effects and provides insights into a novel therapeutic avenue.

NEW & NOTEWORTHY

Endothelial and smooth muscle cell-derived neuropilin-like protein is a novel partner of insulin receptor, and regulates insulin signaling and subsequently a wide array of insulin effects in cultured cells and in vivo through modulation of insulin receptor interaction with adaptor proteins. This novel mechanism of action may be exploited as target for drug development.

insulin is a pleiotropic hormone that, in addition to its well-recognized effects on glucose homeostasis, regulates a large number of key physiological responses, including cell proliferation and differentiation, and immunity (30, 37). Insulin binding to insulin receptor (IR) tyrosine kinase triggers receptor autophosphorylation in the β-subunits (30, 37). The activated receptor recruits and tyrosine phosphorylates adaptor proteins, insulin receptor substrate (IRS), and Src homology 2 domain-containing (Shc) transforming protein to initiate a complex highly integrated signaling network that controls metabolism and gene expression. Two canonical signaling pathways mediate insulin effects: the phosphatidylinositol 3-kinase (PI3K)-AKT/protein kinase B pathway, which is responsible for most of the metabolic actions of insulin, and the Ras-mitogen-activated protein kinase (MAPK) pathway, which in cooperation with the PI3K pathway regulates cell growth and differentiation. Both pathways also regulate gene expression at the transcriptional and translational level (39).

In addition of IRS and Shc, various signaling adaptors, including growth factor receptor bound protein 10 (Grb10), adaptor protein with pleckstrin homology (PH), and Src homology 2 (SH2) domain adaptor protein with pleckstrin homology and Src homology 2 domain (APS) or SH2B2, directly bind to IR and provide docking sites for other binding partners (8). Ubiquitin ligases neuronal precursor cell-expressed developmentally downregulated 4 (Nedd4) and c-casitas B-lineage lymphoma (Cbl) are examples of such binding partners for Grb10 and APS, respectively (8). Overall, Grb10 negatively regulates insulin signal transduction through multiple mechanisms of action (37). APS modulates insulin signaling through recruitment of the ubiquitin ligase c-Cbl and enhancing IR internalization following its multiubiquitination, a process that, while critical to signal transduction, does not affect IR degradation (8, 17, 37).

Insulin resistance is a complex process that involves the interplay of multiple pathways downstream of IR in the muscle and the liver (35). With increased prevalence of insulin resistance and type 2 diabetes, there is growing interest in identifying novel regulators of insulin signaling that potentially can serve as therapeutic targets. Endothelial and smooth muscle cell-derived neuropilin-like protein (ESDN), also called DCBLD2 or CLCP1, is a widely expressed transmembrane protein with a domain structure reminiscent of neuropilins (18, 19). Recent work by our groups has identified ESDN as a novel regulator of vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) signaling in vascular cells (13, 28). In the case of VEGF signaling, in endothelial cells ESDN interferes with VEGF receptor (VEGFR)-2 interaction with negative regulators of VEGF signaling, including protein tyrosine phosphatases protein-tyrosine phosphatase 1B (PTP1B) and T cell protein tyrosine phosphatase, and promotes downstream signaling and angiogenesis (28). The inhibitory effect of ESDN on PDGF signaling in vascular smooth muscle cells (VSMCs) is through regulation of ligand-induced PDGF receptor (PDGFR)-β ubiquitination (13). Similar to PDGF and VEGF, insulin effects are mediated by a receptor tyrosine kinase (RTK), IR. This similarity led us to investigate whether ESDN modulates insulin signal transduction and cellular effects. Here, we demonstrate that ESDN deficiency promotes mitotic and metabolic effects of insulin in vitro and in vivo. The effects of ESDN are traced back to its interaction with IR, which modulates IR interaction with its regulatory adaptor proteins APS and Grb10.

MATERIALS AND METHODS

Reagents.

The polyclonal rabbit anti-ESDN antibody (against an intracellular immunogen) used for Western blotting and immunostaining was obtained from Sigma-Aldrich (St. Louis, MO). The polyclonal goat anti-ESDN (against the extracellular domain) and goat anti-APS used for immunostaining were obtained from Santa Cruz Biotechnology (Dallas, TX), and the rabbit anti-Grb10 antibody used for immunostaining was obtained from Abcam (Cambridge, MA). The antibodies for phospho-MAPK p44/42 (Thr^202/204), MAPK p44/42, phospho-Akt (Ser⁴⁷³), Akt, phospho-IR-β (Tyr¹¹⁴⁶), IR-β (L55B10 and 4B8), c-Cbl, APS, Grb10, Nedd4, ubiquitin, and GAPDH used for Western blotting were obtained from Cell Signaling Technologies (Danvers, MA). Other reagents were purchased from Sigma-Aldrich, unless stated otherwise.

Cell culture.

VSMCs were isolated from 6- to 8-wk-old wild-type (WT) and Esdn^−/− mice. Briefly, the thoracic aorta was harvested from three pairs of mice, and the loosely adhering adventitia was removed under a dissecting microscope after 5 min exposure to collagenase A (1 mg/ml) in Hanks' balanced salt solution. The vessels were cut into small pieces, kept in culture medium overnight, and then digested with collagenase A (2 mg/ml) and elastase (0.5 mg/ml) for 20–30 min. The cells were collected (1,000 revolutions/min centrifuge for 10 min), plated on a culture flask in medium 199 containing 20% FBS, glutamine, and penicillin-streptomycin, and identified as VSMC by immunostaining for smooth muscle α-actin. The experiments were performed with cells at early confluence, unless specified otherwise.

[³H]thymidine incorporation assay.

DNA synthesis was quantified by measuring [³H]thymidine incorporation as described (28). Briefly, 10,000 WT or Esdn^−/− VSMCs were seeded on 24-well culture plates, and the medium was switched to medium 199 containing 1% FBS after 24 h. The next day, cells were treated with insulin or control buffer in triplicates for 18 h, at which point [³H]thymidine (0.074 MBq; GE Healthcare) was added to each well for 6 h. To account for baseline differences in cell proliferation in 1% FBS, the data with insulin are shown as relative to [³H]thymidine incorporation in nonstimulated cells.

Migration assay.

The modified Boyden chamber migration assay was performed as described (28). Medium 199 with 0.5% bovine serum albumin (BSA) containing insulin (70 nM) or control buffer was added in the lower well, and cell migration was quantified after 6 h in triplicates. The data represent the average number of transmigrated cells per microscope field (6–8 areas).

Immunoblotting.

Western blotting was performed on protein extracts as described with minor modification (28). Various tissues were ground into powder in liquid nitrogen. VSMCs were serum starved in FBS-free medium 199 for 16 h before stimulation with insulin (105 nM) or control buffer for the indicated time. Proteins were extracted in ice-cold RIPA buffer (150 mM NaCl, 50 mM Tris·HCl, pH 8.0, 1.0% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS) supplemented with complete proteinase inhibitor (Roche Applied Sciences, Indianapolis, IN) and phosphatase inhibitor cocktails (Sigma-Aldrich). For the phosphorylated antibodies, the membranes were stripped by restore Western blot strip buffer (Thermo Fisher Scientific, Rockford, IL) and reprobed with total protein antibodies. GAPDH was used as loading control. The blots were scanned and quantitated using ChemiDoc MP Imager systems (Bio-Rad Laboratories, Hercules, CA).

Immunoprecipitation.

Cell or tissue lysates (containing 500 μg of protein) were incubated with 1.5 mg of Dynabeads Protein G for Immunoprecipitation (Invitrogen Life Technologies, Carlsbad, CA), and immunoprecipitation was performed as described (28).

Immunofluorescence staining.

WT VSMCs grown on optical glass cover slips (MatTek, Ashland, MA) were fixed with 4% paraformaldehyde in PBS for 20 min at room temperature. Autofluorescence was quenched by 100 mM glycine in PBS for 10 min, and cells were permeabilized with 0.1% Triton X-100 in PBS for 5 min at room temperature. After being blocked with 2% BSA in PBS for 1 h, the cells were incubated with primary antibodies overnight at 4°C. Cover slips were then incubated for 1 h in a 1:200 dilution of either Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary antibodies (Thermo Fisher Scientific, Waltham, MA) and mounted using Prolong Gold Antifade Mountant containing DAPI (Thermo Fisher Scientific). Fluorescence images were acquired using a Nikon AR1 confocal microscope.

Real time-PCR.

Total RNA was isolated and reverse transcribed using QIAGEN kits (Valencia, CA). Real time (RT)-PCR was performed in triplicates on this cDNA using Taqman gene assays (Applied Biosystems) according to the manufacturer's instructions. The results were normalized to GAPDH. Mm00439700_g1 and Mm99999915_g1 Taqman primers were used for mouse IR and GAPDH.

Animal models.

The generation of Esdn^−/− mice on C57BL/6 background was reported previously (28). To induce insulin resistance, sex-matched 10- to 12-wk-old Esdn^−/− and WT mice were fed a high-fat diet (HFD; 45% of calories from fat, mostly lard; Research Diets, New Brunswick, NJ) ad libitum for 8 wk. To evaluate insulin signal transduction in vivo, regular insulin (1 U/kg) was administered intraperitoneally. Various organs were harvested under isoflurane anesthesia after 50 min and kept in liquid nitrogen until further analysis. All animal procedures were performed in accordance with protocols approved by Yale University and Veterans Affairs Connecticut Healthcare System Institutional Animal Care and Use Committees.

Glucose and insulin tolerance tests.

Sex-matched WT and Esdn^−/− mice derived from the same parents (14–16 wk) were kept on a 12:12-h day-night light cycle and provided ad libitum access to food and water, except where indicated. Body weights were measured weekly. Glucose tolerance testing was performed by intraperitoneal injection of glucose (2 g glucose/kg body wt) after overnight (15 h) fasting, as described (43). Insulin tolerance testing was performed by intraperitoneal injection of insulin (0.75 U insulin/kg body wt) after 6 h fasting, as described (43). Effort was made to minimize stress. Briefly, matched animals acclimatized to being handled were brought in the laboratory in the morning and kept in a separate quiet space. Blood was collected with a quick and gentle stick with massaging from the tip of the tail before or after glucose/insulin administration. Blood glucose levels were measured using a Contour blood glucose monitoring system (Bayer HealthCare LLC, Mishawaka, IN) at different time points (0, 10, 20, 30, 60, 90, and 120 min for glucose tolerance test; 0, 20, 30, 60, 90, and 120 min for insulin tolerance test). Insulin secretory response was monitored in overnight (15-h)-fasted animals using blood samples collected at indicated times after intraperitoneal glucose loading (2 g/kg body wt). Plasma insulin was measured with the Rat/Mouse Insulin ELISA Kit (Millipore, Billerica, MA) using mouse standards.

Body composition analysis.

Body composition was assessed by ¹H-magnetic resonance spectroscopy on unanesthetized awake male mice (14–16 wk old on normal chow and 18–20 wk old on HFD) using a Bruker Minispec analyzer mq10 (Bruker Optics) at the Yale Mouse Metabolic Phenotyping Center. The assessment of fat, lean muscle mass, and free water content by this technique is based on differences in their nuclear magnetic resonance relaxation times determined from the time-domain signals (14). Measurements are performed at 37°C and take <3 min.

Targeted inactivation of the genes.

VSMCs were transfected with target-specific pooled short-interfering RNAs (siRNAs) for either APS (sc-40329), Grb10 (sc-40962), or a negative control (sc-37007) siRNA (Santa Cruz Biotechnology) using Lipofectamine RNAiMAX Transfection Reagent (Life Technologies, Grand Island, NY) according to the manufacturers' instructions. Later (48 h), the cells were serum starved overnight and treated with insulin (105 nM, 20 min).

Statistical analysis.

All values are expressed as means ± SD. The significance of the results was assessed by two-tailed t-test or two-way ANOVA with Bonferroni post hoc test (for >2 groups). P <0.05 was considered as significant.

RESULTS

ESDN deficiency promotes insulin-induced VSMC proliferation and migration.

The structural and functional similarity between insulin [and insulin-like growth factor (IGF)] receptors in VSMCs and metabolic tissues (29) led us to investigate the effect of ESDN on insulin signaling in VSMCs. As expected, insulin stimulated cell proliferation, mirrored in [³H]thymidine incorporation, in aortic VSMCs (Fig. 1A). Insulin-induced [³H]thymidine incorporation, expressed relative to [³H]thymidine incorporation in nonstimulated cells, was significantly higher in Esdn^−/− VSMCs compared with WT cells at all insulin concentrations studied (Fig. 1A). Similarly, ESDN deficiency significantly enhanced insulin-induced VSMC migration in a modified Boyden chamber migration assay (Fig. 1, B and C). Collectively, these data indicate that ESDN regulates the mitotic and migratory insulin responses in VSMCs.

Fig. 1.

Endothelial and smooth muscle cell-derived neuropilin-like protein (ESDN) deficiency enhances insulin signal transduction in cultured vascular smooth muscle cells (VSMCs). A: insulin-induced [³H]thymidine incorporation in wild-type (WT) and Esdn^−/− VSMCs. The data represent the fold increase in [³H]thymidine incorporation relative to unstimulated cells in response to three concentrations of insulin from 3 independent experiments. B and C: insulin (70 nM)-induced VSMC Transwell migration. B, top, representative low magnification; B, bottom, high-magnification images of crystal violet-stained filter membranes. C: no. of transmigrated cells/high-power microscopic field from 3 independent experiments. D–G: Western blot analysis of insulin signal transduction in VSMCs. Examples and quantification from 3 independent experiments of insulin-induced mitogen-activated protein kinase (MAPK) p44/42 Thr^202/204 (D and E), protein kinase B (Akt) Ser⁴⁷³ (D and F), and insulin receptor (IR)-β Tyr¹¹⁴⁶ phosphorylation (D and G) in WT and Esdn^−/− VSMCs. The WT and Esdn^−/− samples in D were run on the same gel, but the order of the lanes was changed for consistency. H: RT-PCR analysis of IR mRNA expression in WT and Esdn^−/− VSMCs. I and J: representative Western blot (I) and quantification from 3 independent experiments (J) of IR expression in WT and Esdn^−/− VSMCs using a second IR-specific antibody. *P < 0.05 and **P < 0.01, 2-way ANOVA.

ESDN deficiency enhances insulin signaling in VSMCs.

To elucidate the molecular mechanism(s) of the observed ESDN effect on insulin-induced VSMC proliferation and migration, we assessed the effect of Esdn gene deletion on canonical insulin signaling pathways. Western blot analysis of WT and Esdn^−/− VSMCs exposed to insulin demonstrated enhanced p44/42 MAPK and Akt activation in Esdn^−/− compared with WT cells (Fig. 1, D–F). To trace back the effect of Esdn gene deletion on upstream signaling events, we next examined the state of IR phosphorylation after insulin treatment in Esdn^−/− and WT VSMCs. ESDN deficiency significantly increased insulin-induced IR Tyr¹¹⁴⁶ phosphorylation without altering the total IR expression in VSMCs (Fig. 1, D and G), pointing to IR phosphorylation as the upstream mediator of the ESDN effect on insulin signaling in VSMCs. We confirmed the absence of an effect of Esdn gene deletion on IR expression by RT-PCR and Western blotting using a second anti-IR antibody (Fig. 1, H–J).

ESDN regulates insulin signaling in vivo.

ESDN expression was confirmed by Western blotting in murine liver, muscle, and abdominal adipose tissue (data not shown). To investigate the scope of the observed ESDN effect, we assessed the effect of Esdn gene deletion on insulin signaling in key metabolic organs following intraperitoneal insulin administration in vivo. Similar to cultured VSMCs, in Esdn^−/− mice, insulin-stimulated liver phosphorylation of ERK, Akt, and IR (Tyr¹¹⁴⁶) was enhanced compared with WT animals (Fig. 2, A–D). A similar effect of ESDN deficiency was observed on insulin signaling in the skeletal muscle (Fig. 2, E–H) and abdominal adipose tissue (data not shown), indicating the general nature of the ESDN effect on canonical insulin signaling that regulates cell metabolism and proliferation.

Fig. 2.

ESDN deficiency promotes insulin signal transduction in vivo. A–D: Western blot analysis of insulin signaling in the liver. Examples and quantification from 3 independent experiments of insulin-induced MAPK p44/42 Thr^202/204 (A and B), Akt Ser⁴⁷³ (A and C), and IR-β Tyr¹¹⁴⁶ phosphorylation (A and D) in the liver of WT and Esdn^−/− mice in the absence of, and 50 min after, insulin (1 U/kg) administration. E–H: Western blot analysis of insulin signaling in the muscle. Examples and quantification from 3 independent experiments of insulin-induced MAPK p44/42 Thr^202/204 (E and F), Akt Ser⁴⁷³ (E and G), and IR-β Tyr¹¹⁴⁶ phosphorylation (E and H) in muscles of WT and Esdn^−/− mice in the absence of, and 50 min after, insulin administration (1 U/kg). *P < 0.05, 2-way ANOVA.

ESDN deficiency improves insulin sensitivity in vivo.

To investigate a potential role of ESDN in regulating the metabolic effects of insulin in vivo, we monitored the body weight and fasting blood glucose level in Esdn^−/− and WT mice fed on normal chow or a HFD for 8 wk to induce insulin resistance (34, 35). As shown in Fig. 3A, there was no difference in body weights evaluated over time between Esdn^−/− and WT mice on normal chow. Similarly, fasting blood glucose was not different at 3 wk, 10 wk, or 5 mo of age between Esdn^−/− and WT mice (Fig. 3, B and C). High-fat feeding increased the body weight and fasting blood glucose concentration to a similar degree in both Esdn^−/− and WT mice (Fig. 3, D and E). In normal chow-fed mice, body composition analysis by ¹H-nuclear magnetic resonance spectroscopy revealed that Esdn^−/− mice had an increase in adipose tissue mass that was matched by a reciprocal decrease in muscle mass (Fig. 3F). This difference between Esdn^−/− and WT mice was not present in animals fed on HFD, which had a higher relative adipose tissue content (Fig. 3F). To assess the effect of ESDN on glucose homeostasis, we performed intraperitoneal glucose tolerance tests in normal chow and high-fat-fed Esdn^−/− and WT mice after an overnight fast. Feeding mice a HFD led to glucose intolerance in both Esdn^−/− and WT mice. Despite similar fasting blood glucose levels, the glycemic excursion in normal chow-fed Esdn^−/− mice was significantly lower compared with WT animals (Fig. 3G). Similarly, HFD-fed Esdn^−/− mice had improved glucose tolerance compared with WT mice (Fig. 3H). Indeed, the rise in blood glucose in response to glucose administration in HFD-fed Esdn^−/− mice was comparable to the levels observed in normal chow-fed WT animals. The improvement in glucose tolerance was not associated with increased insulin secretion, since both basal and glucose-stimulated insulin concentrations were identical between Esdn^−/− and WT mice in each diet condition (Fig. 3I). Next, we assessed insulin sensitivity in Esdn^−/− and WT mice by insulin tolerance test. Consistent with the observed effect of ESDN on insulin signaling in the liver, muscle, and adipose tissue, the blood glucose levels dropped to significantly lower levels following insulin administration in Esdn^−/− compared with WT mice under both diets (Fig. 3, J and K). As such, similar to its effect on VSMC proliferation and migration, ESDN deficiency enhanced IR-mediated metabolic responses in vivo.

Fig. 3.

ESDN deficiency enhances insulin sensitivity in vivo. A: body weight of WT and Esdn^−/− mice at different ages; n = 3 mice in each group. B: fasting (overnight) blood glucose level in WT and Esdn^−/− mice of different ages; n = 4–5 in each group. C: effect of fasting duration on blood glucose level of 10-wk-old WT and Esdn^−/− mice; n = 6 in each group. D: effect of high-fat diet (HFD) on body weight over time in 10- to 12-wk-old WT and Esdn^−/− mice; n = 8 in each group. E: blood glucose level following 6 or 15 h fasting in WT and Esdn^−/− mice fed a HFD for 8 wk; n = 6 in each group. F: magnetic resonance spectroscopic analysis of body mass composition in WT and Esdn^−/− mice on normal chow or HFD; n = 3 in each group. G: blood glucose levels in the course of glucose tolerance testing in WT and Esdn^−/− mice on normal chow; n = 6 in each group. H: blood glucose levels in the course of glucose tolerance testing in WT and Esdn^−/− mice on HFD; n = 6 in each group. I: changes in insulin level following ip glucose loading in WT and Esdn^−/− mice on normal chow or HFD; n = 3 in each group. J: blood glucose levels in the course of insulin tolerance testing in WT and Esdn^−/− mice on normal chow; n = 6 in each group. K: blood glucose levels in the course of insulin tolerance testing in WT and Esdn^−/− mice on HFD; n = 6 in each group. *P < 0.05, Esdn^−/− vs. WT, analyses performed using 2-way ANOVA, except for B, C, E, and F, which were analyzed using 2-tailed t-test.

ESDN interacts with IR and modulates receptor ubiquitination.

To address the molecular mechanisms of the ESDN effect on IR signaling, we investigated whether, similar to its interaction with VEGFR-2, ESDN forms a complex with IR. In coimmunoprecipitation studies of WT mouse liver or skeletal muscle tissue extracts, IR-β brought down ESDN, independent of insulin stimulation status (Fig. 4A). We next determined whether, similar to its effects on PDGFRβ, ESDN modulates IR ubiquitination. Contrary to the ESDN effect on PDGFRβ ubiquitination in human VSMCs (which is reduced following ESDN downregulation), IR immunoprecipitation from mouse liver or skeletal muscle tissue extracts followed by immunoblotting for ubiquitin showed markedly higher receptor ubiquitination in Esdn^−/− organs (Fig. 4A).

Fig. 4.

Adaptor protein with pleckstrin homology and Src homology 2 domain (APS)-IR interaction mediates the ESDN effect on insulin signaling. A: IR-β immunoprecipitation from the liver and muscle tissue extracts of WT and Esdn^−/− mice in the absence of, and 50 min after, insulin (1 U/kg) administration, followed by immunoblotting for IR-β, ESDN, and ubiquitin. Data are representative of 3 independent experiments. B: IR-β immunoprecipitation from the liver and muscle tissue extracts of WT and Esdn^−/− mice followed by immunoblotting for IR-β, APS, c-casitas B-lineage lymphoma (Cbl), growth factor receptor bound protein 10 (Grb10), and neuronal precursor cell-expressed developmentally downregulated 4 (Nedd4). Data from two representative pairs of animals are displayed. C: Western blot analysis of APS, c-Cbl, Grb10, and Nedd4 expression in the liver and muscle of WT and Esdn^−/− mice, demonstrating a similar expression level. D: representative coimmunofluorescence staining of ESDN (in green, left) and APS or Grb10 (in red, middle) and fused images (right) demonstrating ESDN-APS and ESDN-Grb10 colocalization. Insets show lower-magnification whole cell images in each panel and fused control staining on right, where nuclei are stained blue with DAPI. Arrows in the insets point to the magnified cell surface area. Scale bar: 6 μm (12 μm for insets). E–H: representative Western blots (E and G) and quantification (F and H) of the effect of siRNA-mediated APS (E and F) or Grb10 (G and H) downregulation on insulin (105 nM, 20 min)-induced IR-β phosphorylation in WT and Esdn^−/− VSMCS; n = 3. *P < 0.05, 2-way ANOVA.

Adaptor protein APS mediates the ESDN effect on insulin signaling.

Nedd4 and c-Cbl are two key IR E3 ubiquitin ligases that interact with the receptor via specific binding partners/adaptors, Grb10 and APS, respectively, and regulate insulin signaling (2, 3, 17, 20, 25, 26, 40). To address whether any of these two ubiquitin ligases and their binding partners are involved in the observed effect of ESDN on insulin signaling, we investigated the interaction between IR and these proteins in liver and muscle tissue extracts from WT and Esdn^−/− mice (Fig. 4B). Both Nedd4 and c-Cbl and their adaptor proteins, Grb10 and APS, coimmunoprecipitated with IR (Fig. 4B). However, in the absence of ESDN, and despite comparable total protein levels (Fig. 4C), more APS and c-Cbl and less Grb10 and Nedd4 associated with IR (Fig. 4B), indicating that ESDN alters the balance between these two regulatory mechanisms of IR. Immunofluorescence staining of WT VSMCs showed that both adaptor proteins, Grb10 and APS, colocalize with ESDN at the cell surface and intracellular compartments (Fig. 4D). Of note, the anti-ESDN antibodies used for APS and Grb10 costaining, which were generated using different immunogens (intracellular cytoplasmic and extracellular domains, respectively) in different hosts, both stained the cell membrane, but yielded distinct intracellular staining patterns. The basis for this difference remains to be studied.

To establish whether either APS or Grb10 mediates the observed effect of ESDN on insulin signaling, these adaptor proteins were downregulated by siRNA in Esdn^−/− and WT VSMCs. APS downregulation significantly attenuated insulin-induced IR phosphorylation in both WT and Esdn^−/− VSMCs. However, this reduction was more prominent in Esdn^−/− cells, where receptor phosphorylation declined to the same level as in WT cells (Fig. 4, E and F), indicating that APS mediates the modulatory effect of ESDN on insulin signaling. siRNA-mediated downregulation of Grb10 in WT VSMCs significantly enhanced insulin-induced IR phosphorylation. On the other hand, in Esdn^−/− cells, Grb10 downregulation abrogated insulin-induced IR phosphorylation (Fig. 4, G and H), highlighting the importance of the specific cellular context on Grb10 function (31).

DISCUSSION

Here we demonstrated that ESDN, a hitherto unrecognized partner of IR, regulates insulin signaling and subsequently a wide array of insulin effects in cultured cells and in vivo through modulation of IR interaction with APS and Grb10. Dysregulation of insulin signaling is the hallmark of metabolic syndrome and type 2 diabetes, two increasingly prevalent disorders associated with considerable morbidity and mortality. Insulin regulates blood glucose levels by modulating intestinal glucose absorption, enhancing glucose uptake in skeletal muscle and adipose tissue, and inhibiting hepatic glucose production while promoting glycogen synthesis (33). In addition, insulin has a wide array of other cellular functions (e.g., on cell growth and differentiation) that may contribute to insulin resistance-related pathologies. The human and economic costs of insulin resistance have led to much effort aimed at identifying novel regulators of insulin signaling as potential therapeutic targets.

The biological effects of insulin are elicited via its binding to insulin (and IGF) receptor. IR belongs to the RTK family of surface receptors, but, unlike typical members of this family, at the basal (non-insulin-binding) state it is composed of two extracellular α-subunits and two transmembrane β-subunits linked by disulfide bonds (16, 37). Upon binding to insulin, the receptor undergoes tyrosine phosphorylation (starting with Tyr¹¹⁴⁶ in the mouse) and activation of the kinase domain. The activated receptor subsequently phosphorylates a number of substrates, including IRS and Shc (37). Tyrosine phosphorylation of these substrates, in turn, activates two major signaling pathways, RAS/ERK (p44/p42 MAPK) and PI3K/Akt, that mediate much of insulin's biological actions. These two canonical signaling pathways are often differentially affected in insulin resistance states. As such, while signal transduction through the PI3K/Akt pathway, which mediates much of the metabolic effects of insulin, is perturbed in insulin resistance states, activation of the RAS/MAPK pathway, which mediates the mitogenic effects of insulin, may be less or not at all affected in this situation (7). In this context, our finding that ESDN deficiency affects both canonical insulin signaling pathways similarly is noteworthy.

ESDN deficiency promotes insulin-induced IR activation. IR activity is regulated by receptor content (controlled at the level of transcription, translation, and protein degradation), adaptors, and regulatory proteins that modulate its autophosphorylation and kinase function (37, 42). One example of such regulatory proteins is ectonucleotide pyrophosphatase phosphodiesterase 1, which through direct interaction with IR blocks autophosphorylation (12). Protein tyrosine phosphatases leukocyte common antigen-related phosphatases and PTP1B directly interact with IR and deactivate its kinase activity (6, 42). Adaptor proteins can directly interfere with receptor kinase activity or serve as a docking point for other regulatory proteins. Suppressor of cytokine signaling proteins-1 and -3, which are upregulated by cytokines, bind to specific phosphorylated tyrosines in IR and inhibit IRS interaction with IR, without affecting receptor phosphorylation (37, 42). Grb10 and APS modulate IR activity in part by serving as docking points for E3 ubiquitin ligases, Nedd4 and c-Cbl (2, 3, 17, 20, 25, 26, 40).

Our data show that ESDN downregulation enhances IR ubiquitination. RTK ubiquitination is a key step in receptor activation and trafficking (1, 11), which plays an important role in sustaining downstream signaling through MAPK and Akt pathways (11). Several E3 ubiquitin ligases, including c-Cbl, Nedd4, and MG53, mediate IR ubiquitination and regulate insulin signaling. Of these, MG53 is expressed only in the heart and skeletal muscle (38) and, therefore, is not a good candidate for mediating the observed multitissue effect of ESDN on IR ubiquitination. c-Cbl and Nedd4 interact with IR through adaptor proteins, APS (SH2B2) and Grb10, and have a broader expression pattern (5, 36). The exact role of c-Cbl- and Nedd4-mediated receptor ubiquitination in RTK signaling remains to be fully defined (11). However, while receptor ubiquitination can lead to lysosomal degradation, APS-mediated ubiquitination of IR does not induce its degradation (17). One could speculate that mono- or multiubiquitinated receptors are “primed” to respond to insulin and recycle very quickly. RTK signaling occurs intracellularly, and it is possible that recycling receptors are primed for further signaling while those on the surface, which are not ubiquitinated, are not. The increase in basal IR ubiquitination in Esdn^−/− tissues may suggest that IR is maintained in an activation-prone conformation in the absence of ESDN (10).

ESDN downregulation enhances APS and diminishes Grb10 interaction with IR while promoting insulin-induced IR phosphorylation and downstream signaling (Fig. 5). As such, ESDN-IR interaction alters the balance of regulatory molecules that interact with IR through these adaptors. Grb10 binds to activated receptor and inhibits its catalytic activity (8, 16), and as such its reduced association with the receptor could contribute to the effect of ESDN deficiency on insulin signaling. Yet, the loss of the stimulatory effect of ESDN gene deletion on IR signal transduction by APS siRNA points to APS as the key mediator of ESDN effect on insulin responses. APS (SH2B2) is a member of the SH2B family of adaptor proteins. The members of this family, which is composed of APS, SH2B1, and SH2B3 (Lnk), share a common structure with an NH₂-terminal proline-rich stretch, a central PH domain, and an SH2 domain and regulate the activity of several RTKs (including IR, IGF receptor, and PDGFRβ) (15, 32, 41). Both SH2B2 (APS) and SH2B1 interact with IR and regulate physiological insulin responses (22, 27). However, SH2B1 function in vivo appears to be predominantly JAK2 mediated (22), and, in our hands, SH2B1 downregulation did not affect the ESDN effects on insulin signaling (data not shown). There are conflicting reports regarding APS (and c-Cbl) effects on insulin signaling (37). Whereas some reports indicate that SH2B2 (APS) promotes insulin signaling, others support a negative regulatory function (3, 8, 21). While reversing the ESDN effect, in our hands, siRNA-mediated APS downregulation modestly reduced IR phosphorylation without affecting receptor level in VSMCs. These findings are consistent with reports that indicate: 1) APS-mediated IR ubiquitination promotes its internalization but has no effect on receptor degradation (17), and 2) APS overexpression delays IR tyrosine dephosphorylation and promotes insulin signaling (2). Interestingly, it is reported that, in contrast to the effects of APS on insulin signaling in cultured cells, APS (and c-Cbl) gene deletion modestly promotes insulin sensitivity in the mouse (23, 24). The molecular mechanisms of these observations and whether they reflect chronic compensatory changes remain to be determined.

Fig. 5.

Schematic presentation of the effect of ESDN on IR interaction with key regulatory molecules, APS, c-Cbl, Grb10, and Nedd4.

ESDN is expressed in major metabolic organs, including liver, skeletal muscle, and adipose tissue. Consistent with the effects of APS on insulin signaling in cultured cells, the enhanced APS association with IR in Esdn^−/− tissues is associated with enhanced insulin signal transduction in these organs. In line with this multitissue effect of ESDN, insulin sensitivity is enhanced in vivo in Esdn^−/− mice. Esdn^−/− mice have a similar body weight and similar insulin levels as WT animals. The minor difference in body mass composition observed between Esdn^−/− and WT mice on normal chow (but not in animals fed on HFD) should have led to more insulin resistance and thus cannot explain the higher insulin sensitivity detected in Esdn^−/− mice. The liver, adipose tissue, and skeletal muscle contribute to varying degrees to insulin resistance under specific pathological conditions (35). Based on the multitissue effect of ESDN on IR signal transduction, it is likely that Esdn gene deletion will have a similar insulin-sensitizing effect in different animal models of insulin resistance and diabetes.

The transmembrane protein ESDN (DCBLD2), along with its homolog DCBLD1, are members of a novel family of neuropilin-like proteins. The extracellular domain of ESDN consists of complement C1r/C1s, Uegf, Bmp1 (CUB); LCCL; and discoidin (FV/VIII) domains. Unlike neuropilins, which are composed of two CUB, two discoidin, a MAM, and a short intracellular domain, ESDN has a relatively long intracellular domain that can function as a scaffold for CT10 regulator of kinase like (4). It is conceivable that ESDN intracellular domain can similarly be involved in its interactions with IR and adaptor proteins, APS and Grb10. Previous work has implicated ESDN in regulation of VEGF, PDGF, and epidermal growth factor (EGF) signal transduction (9, 13, 28). Our new findings on ESDN regulation of insulin signaling suggest that ESDN might regulate the biological effects of other growth factors. Of note, the ESDN effects appear to be cell and growth factor dependent, and distinct molecular mechanisms are implicated in its regulatory effects on VEGF, PDGF, EGF, and insulin signaling. These may be explained by variances in experimental settings or, more importantly, reflect interspecies differences or idiosyncratic differences in the structure and interacting partners of various RTKs (16).

In conclusion, ESDN is a novel regulator of mitotic and metabolic effects of insulin and modulates insulin signal transduction through regulation of IR interaction with its adaptor proteins, APS and Grb10. The multimetabolic tissue effects of ESDN suggest that this novel mechanism of action may be exploited as target for drug development in type 2 diabetes and insulin resistance. Ongoing studies seek to establish the feasibility of this approach and its comparative effectiveness compared with existing therapeutic interventions.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R01-HL-12992 and R01-HL-114703 and Department of Veterans Affairs Merit Award I0-BX001750.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

X.L., J.-J.J., L.N., M.R., and J.Z. performed experiments; X.L., J.-J.J., L.N., M.R., J.Z., and M.M.S. analyzed data; X.L., L.N., and M.M.S. drafted manuscript; X.L., J.-J.J., L.N., M.R., J.Z., V.T.S., and M.M.S. approved final version of manuscript; V.T.S. and M.M.S. interpreted results of experiments; V.T.S. and M.M.S. edited and revised manuscript; M.M.S. conception and design of research.