p75 neurotrophin receptor regulates glucose homeostasis and insulin sensitivity

Abstract

Insulin resistance is a key factor in the etiology of type 2 diabetes. Insulin-stimulated glucose uptake is mediated by the glucose transporter 4 (GLUT4), which is expressed mainly in skeletal muscle and adipose tissue. Insulin-stimulated translocation of GLUT4 from its intracellular compartment to the plasma membrane is regulated by small guanosine triphosphate hydrolases (GTPases) and is essential for the maintenance of normal glucose homeostasis. Here we show that the p75 neurotrophin receptor (p75^NTR) is a regulator of glucose uptake and insulin resistance. p75^NTR knockout mice show increased insulin sensitivity on normal chow diet, independent of changes in body weight. Euglycemic-hyperinsulinemic clamp studies demonstrate that deletion of the p75^NTR gene increases the insulin-stimulated glucose disposal rate and suppression of hepatic glucose production. Genetic depletion or shRNA knockdown of p75^NTR in adipocytes or myoblasts increases insulin-stimulated glucose uptake and GLUT4 translocation. Conversely, overexpression of p75^NTR in adipocytes decreases insulin-stimulated glucose transport. In adipocytes, p75^NTR forms a complex with the Rab5 family GTPases Rab5 and Rab31 that regulate GLUT4 trafficking. Rab5 and Rab31 directly interact with p75^NTR primarily via helix 4 of the p75^NTR death domain. Adipocytes from p75^NTR knockout mice show increased Rab5 and decreased Rab31 activities, and dominant negative Rab5 rescues the increase in glucose uptake seen in p75^NTR knockout adipocytes. Our results identify p75^NTR as a unique player in glucose metabolism and suggest that signaling from p75^NTR to Rab5 family GTPases may represent a unique therapeutic target for insulin resistance and diabetes.

Insulin resistance is a key feature of type 2 diabetes and is characterized by decreased glucose disposal, increased hepatic glucose production, with a reduced ability of insulin to maintain normal glucose homeostasis (1). Whole-body glucose metabolism is regulated by a complex communication network between different tissues, including adipose tissue, liver, muscle, and brain (1, 2). Glucose transporter 4 (GLUT4) is the principal insulin-stimulated glucose transporter expressed primarily in adipose tissue and skeletal muscle (1). Insulin stimulates glucose uptake by inducing GLUT4 translocation to the plasma membrane (3). GLUT4 trafficking from intracellular compartments to the plasma membrane is regulated by a number of small guanosine triphosphate hydrolases (GTPases), including Rab5 and its family member Rab31 (3, 4). Identification of the molecular components that regulate glucose homeostasis is essential to understand the mechanisms of insulin resistance and to discover novel therapeutic targets.

p75^NTR is a member of the tumor necrosis factor (TNF) receptor superfamily that was originally identified as a receptor for neurotrophins (5). p75^NTR is expressed in the nervous system and in adult nonneuronal tissues (6), such as white adipose tissue (WAT) (7), and muscle (8). p75^NTR has surprisingly diverse cellular functions, including modulation of cell survival, apoptosis, and differentiation (5, 9), which underlie its in vivo biologic functions, including liver and muscle regeneration (8, 10), extracellular matrix remodeling (11), sensory neuron development (12), hypoxia, and the angiogenic response (13). p75^NTR induces apoptosis in neurons either after activation by neurotrophins or in a neurotrophin-independent manner potentially due to receptor multimerization (5, 14). Whereas p75^NTR lacks a catalytic motif, its intracellular domain interacts with several adapter proteins to activate multiple signaling pathways (5). For example, p75^NTR directly binds the GTPases Rho and Ras to regulate neurite formation (15), and the intracellular domain of p75^NTR is sufficient to induce activation of RhoA (16) and Rac GTPases (17). Because p75^NTR is expressed in WAT and skeletal muscle (7) and regulates GTPase activity (16), we hypothesized that p75^NTR might participate in the regulation of glucose metabolism.

In this study, we report an unexpected function of p75^NTR as a regulator of insulin-stimulated glucose transport and systemic insulin sensitivity. We show that genetic loss of p75^NTR increases insulin sensitivity in lean, chow-fed mice. In adipocytes, loss of p75^NTR increases the activity of Rab5, which promotes GLUT4 trafficking, while decreasing the activity of Rab31, which promotes GLUT4 retention, thus resulting in increased plasma membrane translocation of GLUT4 and increased glucose uptake. While p75^NTR interacts with both Rab5 and Rab31, it primarily regulates glucose uptake via Rab5, as shown by rescue experiments whereby dominant negative Rab5 rescues the increase in glucose uptake in p75^NTR−/− adipocytes. Taken together, these results identify an unanticipated function of p75^NTR as a unique regulator of glucose metabolism and insulin sensitivity.

Results

Loss of p75^NTR Increases Insulin Sensitivity Independently of Body Weight.

On normal chow, p75^NTR−/− mice weigh the same as wild-type (WT) mice (Fig. 1A and Fig. S1A), consume the same amount of food (Fig. S1B), and show similar body composition (Fig. S1C), suggesting that loss of p75^NTR does not regulate appetite or body weight on normal diet. To determine whether p75^NTR regulates other metabolic pathways, we tested glucose and insulin homeostasis in WT and p75^NTR−/− mice. p75^NTR−/− mice showed lowered glycemic excursions during glucose tolerance tests (GTTs) (50% decrease in area under curve, AUC) (Fig. 1B) with a 30% increase in the hypoglycemic effect of insulin during insulin tolerance tests (ITTs) (Fig. 1 C and D) compared with WT controls. These findings indicate that lean, chow-fed p75^NTR−/− mice are insulin sensitive relative to weight-matched lean controls, and this was quantitated by euglycemic-hyperinsulinemic glucose clamp studies. As seen in Fig. 1E, the glucose disposal rate (GDR) was greater in p75^NTR−/− mice, as was the insulin stimulated GDR (IS-GDR) value (Fig. 1F) compared with WT mice. Basal hepatic glucose production (HGP) was not affected by p75^NTR deficiency, but suppression of HGP by insulin was significantly improved (Fig. 1G) by 20% (Fig. 1H). Furthermore, free fatty acid (FFA) levels were lower in the p75^NTR−/− mice during the clamp studies, whereas basal FFA, glucose, and insulin levels were unchanged (Fig. S2 A–C). It is important to note that genetic manipulations that improve insulin sensitivity in lean, normal chow-fed mice are unusual, and these results demonstrate significantly enhanced insulin sensitivity in p75^NTR−/− mice independent of food intake or body weight.

Fig. 1.

p75^NTR−/− mice exhibit improved insulin sensitivity on normal diet. (A) Body weight of WT and p75^NTR−/− mice on normal chow (n = 10; NS, not significant). (B) Glucose tolerance of fasted WT and p75^NTR−/− mice on normal chow (n = 10, *P < 0.05, **P < 0.01, ***P < 0.001). (C) Insulin tolerance of fasted WT and p75^NTR−/− mice on normal chow (n = 10, **P < 0.01, ***P < 0.001). (D) Percentage of initial blood glucose in insulin tolerance test (n = 10, *P < 0.05). (E) Glucose disposal rates (GDRs) from WT and p75^NTR−/− mice on normal chow (n = 10, **P < 0.01). (F) Insulin-stimulated glucose disposal rates (IS-GDRs) of WT and p75^NTR−/− mice on normal chow (n = 10, *P < 0.05). (G) Hepatic glucose production (HGP) from WT and p75^NTR−/− mice on normal chow (n = 10, *P < 0.05). (H) Suppression of HGP in WT and p75^NTR−/− mice on normal chow (n = 10, *P < 0.05). Data are shown as the means ± SEM. Statistical comparisons between means were made with one-way ANOVA (B–D) and Student's t test (A and E–H).

Loss of p75^NTR Increases Glucose Uptake.

We next explored the mechanisms by which p75^NTR might regulate insulin sensitivity in cell autonomous in vitro systems. Because p75^NTR−/− mice displayed increased in vivo glucose disposal (Fig. 1E), we examined the effects of p75^NTR on insulin-stimulated glucose uptake in adipocytes and myocytes. Adipocytes derived from p75^NTR−/− mouse embryonic fibroblasts (MEFs) showed a threefold increase in insulin-stimulated glucose uptake compared with WT (Fig. 2A), whereas basal glucose uptake was identical between WT and p75^NTR−/− MEF-derived adipocytes (Fig. 2A). Similar differences in insulin-stimulated glucose transport were observed in primary adipocytes isolated from WT and p75^NTR−/− mice (Fig. 2B). Likewise, shRNA knockdown of p75^NTR increased insulin-stimulated glucose uptake in 3T3L1 adipocytes (Fig. 2C). To analyze the potential role of p75^NTR in skeletal muscle, we examined the influence of p75^NTR on insulin-stimulated glucose uptake in L6 myocytes. Similar to adipocytes (Fig. 2C), L6 cells transfected with p75^NTR siRNA showed no change in basal glucose transport but displayed a significant 20% increase in insulin-stimulated glucose uptake compared with cells transfected with a control siRNA (Fig. 2D). Efficiency of the p75^NTR knockdown in 3T3L1 adipocytes and L6 myoblasts (Fig. 2 C and D), and expression of p75^NTR in MEF-derived adipocytes, WAT and skeletal muscle (Fig. S2D) is shown by immunoblot.

Fig. 2.

p75^NTR regulates glucose uptake in adipocytes and skeletal muscle cells. (A) Basal and insulin-stimulated glucose uptake in WT and p75^NTR−/− MEF-derived adipocytes (NS, not significant; *P < 0.05). (B) Basal and insulin-stimulated glucose uptake in WT and p75^NTR−/− primary isolated adipocytes (NS, not significant; *P < 0.05). (C) p75^NTR expression in 3T3L1 adipocytes infected with p75^NTR shRNA or scrambled shRNA lentivirus. Gapdh was used as a loading control (Upper). Basal and insulin-stimulated glucose uptake in 3T3L1 cells differentiated to adipocytes and infected with lentivirus containing p75^NTR shRNA or scrambled shRNA control (*P < 0.05; **P < 0.01) (Lower). (D) p75^NTR expression in L6 myocytes tranfected with p75^NTR siRNA or siRNA control. Gapdh was used as a loading control (Upper). Basal and insulin-stimulated glucose uptake in L6 myocytes cells transfected with p75^NTR siRNA or siRNA control (**P < 0.01; ***P < 0.001) (Lower). (E) Basal and insulin-stimulated glucose uptake in WT and p75^NTR−/− MEF-derived adipocytes electroporated with p75FL, p75ICD, or GFP control expression vectors. After 72 h, cells were serum starved for 3 h and then stimulated with 100 nM insulin (*P < 0.05). Data are shown as the means ± SEM of a minimum of three different experiments performed in triplicate. Statistical comparisons between means were made with one-way ANOVA (A, B, and E) and Student's t test (C and D).

Consistent with the results of p75^NTR depletion, overexpression of full-length p75^NTR (p75FL) rescued the increase in glucose uptake in p75^NTR−/− adipocytes (Fig. 2E). We also overexpressed the intracellular domain (ICD) of p75^NTR, which lacks the extracellular domain that binds neurotrophins. Similar to the p75FL, the p75ICD also rescued the increase in glucose uptake in p75^NTR−/− adipocytes (Fig. 2E), whereas p75FL and p75ICD decreased glucose uptake in WT adipocytes (Fig. 2E and Fig. S3A). These results indicate that the ability of p75^NTR to regulate glucose uptake is neurotrophin independent. Consistent with this, addition of the p75^NTR neurotrophin ligands nerve growth factor (NGF) or brain-derived neurotrophin factor (BDNF) or neutralization of neurotrophins either by antibodies to NGF, BDNF, or BDNF scavenger Fc-TrkB, pan-neurotrophin scavenger Fc-p75^NTR, or inhibition of Trk signaling by K252a did not affect glucose uptake (Fig. S3 B and C). Together, these results indicate that p75^NTR regulates glucose uptake in adipocytes and myoblasts and are consistent with the improved in vivo insulin sensitivity and glucose disposal in p75^NTR−/− mice (Fig. 1).

p75^NTR Regulates GLUT4 Retention.

Because regulation of insulin-stimulated glucose transport is a unique cellular function for p75^NTR, we sought to determine the molecular mechanism downstream of p75^NTR that regulates glucose uptake. Insulin-stimulated glucose uptake is largely due to translocation of GLUT4 from its basal intracellular location to the plasma membrane (3, 18). To determine whether p75^NTR regulates GLUT4 trafficking, we examined the intracellular localization of GLUT4-GFP in p75^NTR−/− adipocytes. In response to insulin, p75^NTR−/− adipocytes exhibited increased GLUT4 translocation to the plasma membrane compared with WT adipocytes (Fig. 3A). We also performed plasma membrane fractionation from WAT derived from WT and p75^NTR−/− mice injected with either saline or glucose to assess translocation of GLUT4 in response to endogenous insulin. As shown in Fig. 3B, plasma membrane GLUT4 content was increased in p75^NTR−/− WAT. p75^NTR deletion did not affect total protein levels of GLUT1 or GLUT4 in 3T3L1 adipocytes, MEF-derived adipocytes, or primary adipocytes (Fig. S4 A–C). Protein levels of p75^NTR were shown by immunoblot (Fig. S4 A–C). p75^NTR also did not affect insulin receptor signaling, as loss of p75^NTR did not increase insulin-stimulated Akt phosphorylation (Fig. S4D).

Fig. 3.

p75^NTR regulates GLUT4 trafficking in adipocytes. (A) Immunocytochemistry of WT and p75^NTR−/− MEF-derived adipocytes transfected with GFP-tagged GLUT4 and stimulated with 100 nM of insulin for 30 min. (Scale bar, 75 μm.) Representative images of two independent experiments are shown (Left). GLUT4 plasma membrane (PM) translocation was quantified by determining the percentage of cells with GFP rim staining. At least 50 cells per condition were counted (*P < 0.05 by Student's t test). (B) Western blot for GLUT4 in the plasma membrane fraction of epididymal WAT from WT and p75^NTR−/− mice on normal chow injected with (1 g/kg) of dextrose after 6 h fasting. Ponceau red was used as a loading control.

p75^NTR Regulates Rab5 and Rab31 Activities.

Stimulation of GLUT4 translocation by insulin results from trafficking of GLUT4 vesicles from intracellular storage sites and their subsequent fusion with the plasma membrane (3). Activation of Rab5 GTPase functions in the initial endocytosis of GLUT4, thus regulating the intracellular pools of GLUT4 (19). The intracellular retention of GLUT4 is sustained by the activity of Rab31, a Rab5 subfamily GTPase implicated in trans-Golgi network-to-endosome trafficking (4). Insulin increases the activity of plasma membrane-associated Rab5 (20) and decreases the activity of Rab31 in adipocytes, and Rab31 knockdown potentiates insulin-stimulated movement of GLUT4 to the cell surface with increased glucose uptake (4). Because p75^NTR is known to regulate other GTPases, such as Rac (17), Ras (15), and RhoA (16), we sought to determine whether p75^NTR is required for activation of Rab5 and Rab31. Early endosome antigen-1 (EEA1) binds to Rab5 and Rab31 in a GTP-dependent manner (4). A pulldown assay with EEA1 showed increased activity of Rab5 and decreased activity of Rab31 in p75^NTR−/− adipocytes (Fig. 4A), suggesting that loss of p75^NTR increases GLUT4 endocytosis, while decreasing GLUT4 intracellular retention. Overexpression of dominant negative Rab5, but not dominant active Rab31, rescued the increased insulin-induced glucose uptake in the p75^NTR−/− adipocytes (Fig. 4B). Therefore, loss of p75^NTR differentially regulates Rab5 and Rab31 to increase the flux of GLUT4 through the recycling pathway, while decreasing GLUT4 intracellular retention, thus resulting in increased plasma membrane GLUT4 translocation and glucose uptake.

Fig. 4.

p75^NTR regulates Rab5 and Rab31 GTPase activity in adipocytes. (A) WT and p75^NTR−/− MEF-derived adipocytes were stimulated with insulin for 5 and 20 min. The activation state of Rab31 and Rab5 was determined using GST-EEA1/NT. Rab31-GTP and Rab5-GTP levels normalized to Rab31 and Rab5, respectively, were quantified by densitometry (Δ represents mean value). Representative blot of three independent experiments is shown. (B) Basal and insulin-stimulated glucose uptake in WT and p75^NTR−/− MEF-derived adipocytes electroporated with Rab31Q69L, Rab5S34N, or GFP control expression vector. After 72 h, cells were serum starved for 3 h and then stimulated with insulin (*P < 0.05 vs. WT basal; ^#P < 0.05 vs. p75^NTR−/− basal). Data are shown as the means ± SEM of a minimum of three different experiments performed in triplicate. Statistical comparisons between means were made with one-way ANOVA.

Rab5 Family GTPases Rab5 and Rab31 Directly Associate with the Death Domain of p75^NTR.

The ICD of p75^NTR interacts with several binding partners to regulate cellular functions (21). Coimmunoprecipitation showed interaction of p75^NTR with Rab31 (Fig. 5A) and Rab 5 (Fig. 5B). p75^NTR also interacted with Gapex5 (Fig. S5A), a guanine nucleotide exchange factor (GEF) for Rab5 and Rab31 (4). Endogenous coimmunoprecipitation showed that p75^NTR interacts with Rab31 and Rab5 in 3T3L1 adipocytes (Fig. 5C). p75^NTR colocalized with Rab31 and Rab5 in 3T3L1 adipocytes (Fig. S5B). To evaluate the subcellular distribution of p75^NTR and Rab5, we analyzed subcellular fractions of adipocyte. The majority of p75^NTR protein in adipocytes was localized together with Rab5 in low-density microsomes (LDMs) (Fig. S5C). This is in accordance with prior studies demonstrating that a majority of Rab5 resides in LDM (20). To verify the specificity of the association of p75^NTR with Rab5 family GTPases, mapping studies were conducted using deletion mutants (Fig. S5D). Rab5 and Rab31 interact with p75FL, but not p75Δ83, a deletion missing the distal 83 amino acids (Fig. 5 D and E), suggesting that the interaction between p75^NTR and Rab5 family GTPases occurs within the death domain (DD) of p75^NTR. As expected, the deletion mutant p75Δ151, which lacks the entire p75ICD, also did not interact with Rab5 and Rab31 (Fig. 5 D and E). Overexpression of p75FL suppressed insulin-stimulated glucose uptake, but not p75Δ83 (Fig. 5F), further supporting the involvement of the DD of p75^NTR in the regulation of glucose uptake by p75^NTR.

Fig. 5.

The death domain of p75^NTR directly associates with Rab5 and Rab31 GTPases. (A) Immunoprecipitation of HA-FLp75^NTR with myc-Rab31 transfected in HEK293T cells. (B) Immunoprecipitation of HA-FLp75^NTR with GFP-Rab5 transfected in HEK293T cells. Lysates were immunoprecipitated with anti-HA antibody, and Western blots were developed with anti-GFP and anti-HA antibodies. (C) Immunoprecipitation of p75^NTR with Rab31 and Rab5 in 3T3L1 adipocytes stimulated with insulin. Lysates were immunoprecipitated with anti-p75^NTR antibody, and Western blots were developed with anti-Rab31 and anti-Rab5 antibodies. (D) Mapping of the p75^NTR sites required for interaction with Rab31. Immunoprecipitation of myc-Rab31 with truncated forms of HA-p75^NTR transfected in HEK293T cells. Lysates were immunoprecipitated with anti-myc antibody, and Western blots were developed with anti-HA and anti-myc antibodies. (E) Mapping of the p75^NTR sites required for interaction with Rab5. Immunoprecipitation of truncated forms of HA-p75^NTR with GFP-Rab5 in HEK293T cells. Lysates were immunoprecipitated with anti-HA antibody, and Western blots were developed with anti-GFP and anti-HA antibodies. (F) Basal and insulin-stimulated glucose uptake of WT MEF-derived adipocytes electroporated with p75FL, p75Δ83, or GFP control expression vectors. After 72 h, cells were serum starved for 3 h and then stimulated with 100 nM insulin (**P < 0.01; ***P < 0.001; NS, not significant). Data are shown as the means ± SEM of two different experiments performed in triplicate. Statistical comparisons between means were made with one-way ANOVA. (G) Peptide array mapping of the p75^NTR ICD sites required for the interaction with Rab5 and Rab31. Schematic diagram of the p75^NTR ICD shows the domain organization. The six helixes within the DD are highlighted in yellow. Peptide array screened with recombinant GST-Rab5 and GST-Rab31 revealed helix 4 of p75^NTR ICD as the strongest region that interacts with Rab5 and Rab31. Peptide location, length, and sequences are shown. Peptide library was also screened with recombinant GST as a control. (H) Schematic representation of the binding of Rab5 family GTPases and RhoGDI to helixes 4 and 5 of the death domain of p75^NTR, respectively. Coimmunoprecipitation of p75ICD and activated GST-Rab5 recombinant proteins. Tat-Pep5 or control peptide (Tat-ctrl) were added as indicated. GST-Rab5 was immunoprecipitated with anti-GST antibody, and Western blots were developed with anti-p75^NTR and anti-GST antibodies. (I) Coimmunoprecipitation of p75ICD with activated GST-Rab5 or GST-RhoGDI recombinant proteins. Pep5 was added as indicated. GST-Rab5 and GST-RhoGDI were immunoprecipitated with anti-GST antibody, and Western blots were developed with anti-p75^NTR and anti-GST antibodies. Western blots were performed in triplicates and in all panels, representative blots are shown.

To further characterize the interaction, we used peptide array technology, which has defined sites of direct interaction for other p75^NTR partners (11, 13) and many other proteins (22). Using Rab5-GST or Rab31-GST, we screened a peptide array library of overlapping 25-mer peptides that spanned the sequence of p75ICD. The strongest interactions were within the DD and in particular within helix 4 and the region between helixes 4 and 5 (peptide 7, amino acids 386–400) (Fig. 5G). This interaction was distinct from that previously described for binding of RhoGDI within helix 5 of p75ICD (Fig. 5H) (23). Accordingly, Tat-Pep5, an inhibitory peptide that blocks the interaction of p75^NTR with RhoGDI (23), did not block the interaction of Rab5 with p75ICD (Fig. 5H). Pep5 inhibited the interaction of p75^NTR with RhoGDI without affecting p75^NTR binding to Rab5 (Fig. 5I), further suggesting that different epitopes within the DD of p75^NTR contribute to the binding of Rho and the Rab5 family GTPases. These results suggest a direct interaction between p75^NTR and the Rab5 family GTPases that primarily requires helix 4 within the death domain of p75^NTR.

Discussion

Our studies reveal an unanticipated mechanism for the regulation of glucose metabolism by p75^NTR. The results indicate a unique model, whereby p75^NTR differentially regulates Rab5 and Rab31 leading to decreased GLUT4 plasma membrane translocation that results in decreased insulin-stimulated glucose uptake. p75^NTR exerts corresponding in vivo effects on glucose metabolism, because lean p75^NTR−/− mice display increased insulin sensitivity on a normal chow diet, independent of body weight. p75^NTR is expressed in myocytes and adipocytes, and knockdown of this protein markedly increased insulin-stimulated glucose uptake, whereas overexpression had the opposite effect. These cell autonomous actions of p75^NTR on glucose transport indicate a direct role of this protein in the regulation of glucose homeostasis. Because skeletal muscle accounts for the great majority of glucose disposal, it is likely that the effects of p75^NTR on skeletal muscle glucose transport play a major role in the in vivo phenotype of the p75^NTR−/− mice. There are numerous examples of interventions that lead to insulin resistance in normal mice or improve insulin sensitivity in obese, insulin-resistant animals. However, outside of caloric restriction, there are very few manipulations that can enhance insulin sensitivity in normal lean mice, because these mice are already highly insulin sensitive. Therefore, the current results are notable, showing that p75^NTR deletion leads to a cell autonomous improvement in insulin action in adipocytes and skeletal muscle cells, accompanied by an overall increase in systemic insulin sensitivity. The mechanisms for this improved insulin action involve an enhanced ability of insulin to mediate translocation of GLUT4 to the cell surface.

The knockout animals display systemic insulin sensitivity in all three major insulin target tissues: fat, muscle, and liver. In these situations, it is often difficult to dissect out primary vs. secondary effects. However, in this case, we demonstrate clear-cut and substantial cellular effects of p75^NTR to regulate glucose uptake in myocytes and adipocytes. Because these in vitro events would not be influenced by potential secondary effects that might be manifested in vivo, we conclude that the in vivo insulin sensitivity in the p75^NTR−/− mice can be traced to the novel effect of p75^NTR to attenuate insulin-stimulated glucose transport. With respect to hepatic insulin sensitivity, p75^NTR is not expressed in hepatocytes (10), and liver cells clearly do not have the GLUT4 translocation machinery. Therefore, it is likely that the increased in vivo hepatic insulin sensitivity is secondary to the overall improvement in glycemia and glucose homeostasis.

Previous reports have demonstrated that Rab31 activation promotes intracellular retention of GLUT4, thereby attenuating insulin stimulated translocation of GLUT4 to the cell surface (4). The activation state of this G protein is maintained by the activity of Gapex-5, a Rab31 guanyl nucleotide exchange factor (GEF) that translocates to the plasma membrane in response to insulin, thus reducing the activity state of Rab31 (4). Moreover, Rab5 undergoes activation in response to insulin (20), and this increase in the activity of the GTPase is associated with increased GLUT4 endocytosis (19). Although the mechanism of this effect is uncertain, it may involve increased phosphatidylinositol 3-phosphate production, Akt activation, or increased endocytosis to stimulate GLUT4 recycling (24). We found that loss of p75^NTR in adipocytes led to increased Rab5 and decreased Rab31 activity, suggesting p75^NTR might act to reduce endocytosis and increase retention of GLUT4 in intracellular vesicles, thus decreasing glucose uptake. Because dominant negative Rab5, but not dominant active Rab31, rescued the effects of p75^NTR−/− on glucose uptake, it is possible that increased endocytosis of GLUT4 is a critical step in p75^NTR-mediated GLUT4 retention that is facilitated by Rab31 activation. Future experiments to directly measure the rate of the internalization and exocytosis of GLUT4 in p75^NTR−/− adipocytes will shed light onto the specific trafficking steps regulated by p75^NTR.

In neurons, p75^NTR is detected at Rab5⁺ vesicles (25), and Rab5 activity is required for retrograde axonal transport (26). It is therefore possible that p75^NTR might also regulate Rab5 activity in neurons, thus affecting endocytic recycling and transport of neurotrophins and their receptors to different neuronal compartments. In neurons, p75^NTR undergoes Rab5-dependent internalization and recycling independent of neurotophins, similar to ligand-independent endocytosis described for transferrin and Notch receptors (25). The effects of p75^NTR on glucose uptake appear to be independent of neurotrophins, because the p75ICD was sufficient to rescue the increased insulin-stimulated glucose uptake in p75^NTR−/− adipocytes and neurotrophins had no effects on glucose uptake in 3T3L1 adipocytes.

Because p75^NTR does not contain a GEF domain, it is unlikely to affect Rab5 or Rab31 by direct nucleotide exchange. However, like other G proteins, Rab activity also depends on GDP dissociation inhibitors (GDIs), which release Rab proteins from their inhibitors (27). Because p75^NTR activates RhoA by displacing it from its GDI (23, 28), it might function as a displacement factor to regulate Rab31 activity. p75^NTR coimmunoprecipitates with both Rab5 and Rab31, and the region between helixes 4 and 5 within the DD is necessary for the interaction. The p75^NTR DD is similarly involved with the interaction of p75^NTR and Rho (23) or Ras (15) through helix 5 within the DD domain, which interacts with and activates small GTPases of the Rho family (23, 29). Our study suggests that different epitopes within the DD of p75^NTR contribute to the binding of Rho and the Rab5 family GTPases. As a result, Tat-Pep5, which inhibits the interaction of RhoGDI with helix 5 of p75^NTR, did not inhibit the interaction of p75^NTR with Rab5. Future experiments focusing on the design of peptide inhibitors to specifically block the interaction of the Rab5 family GTPases to helix 4 of p75^NTR and thus increase glucose uptake might reveal novel therapeutic approaches for insulin resistance.

These results indicate potential opposite biological functions for p75^NTR and BDNF in metabolism. BDNF and its tyrosine kinase receptor TrkB act in the hypothalamus to regulate body weight by suppressing appetite (30). On normal chow, BDNF^+/− or mice conditionally depleted of neuronal BDNF overeat and become obese (30, 31). In contrast, on normal chow p75^NTR−/− mice maintain normal body weight and consume the same amount of food as WT mice, suggesting that p75^NTR is not involved in appetite control. Pharmacologic studies administering high doses of recombinant BDNF in vivo have shown that BDNF-mediated TrkB activation might also enhance glucose utilization in peripheral tissues (32, 33). Because p75^NTR decreases glucose uptake, p75^NTR and BDNF might differentially regulate glucose metabolism. There are several other examples indicating that p75^NTR and BDNF/TrkB receptor signaling pathways exert opposite biological functions (34). For example, p75^NTR inhibits neurite outgrowth, whereas BDNF via TrkB enhances neurite outgrowth (35, 36). Therefore, the final metabolic outcome of neurotrophin receptor signaling in vivo may depend on the balance between central regulation of food intake by BDNF/TrkB signaling and peripheral regulation of glucose uptake by p75^NTR. In any event, the increased insulin sensitivity of the p75^NTR−/− mice suggests that p75^NTR may represent a unique therapeutic target for insulin resistance and diabetes.

Materials and Methods

Mice fasted for 6 h were injected i.p. with dextrose (1g/kg; Hospira) for GTT or insulin (0.4 units/kg, Novolin R (Novo-Nordisk). Blood samples were drawn at 0, 10, 30, 60, and 120 min after dextrose or insulin injection for glucose measurement (Onetouch; Lifescan).

Mice hyperinsulinemic-euglycemic clamp studies, glucose uptake assay, cell fractionation, Western blotting, coimmunoprecipitation, GST activity assay, immunocytochemistry, microscopy, and statistical analysis are provided in SI Materials and Methods.