Tub Has a Key Role in Insulin and Leptin Signaling and Action In Vivo in Hypothalamic Nuclei

All experiments were approved by the Ethics Committee of the State University of Campinas. Eight-week-old male C57BL/6 mice obtained from the University of Campinas, São Paulo, Brazil, were assigned to receive a standard rodent chow or a high-fat diet (HFD) as previously described (37,38) and water ad libitum. For fasting and refeeding experiments, fasted (24 h) mice were allowed to refeed for 1 h or 2 h, and hypothalamic nuclei were dissected for protein studies, as described. All feeding tests were conducted between 8:00 a.m. and 10:00 a.m.

Anesthetized mice were stereotaxically instrumented (Ultra Precise model 963; Kopf) to implant stainless steel cannulas (26-gauge; Plastics One) in the right lateral ventricle. The coordinates used from the bregma were: anterior/posterior, −0.5 mm; lateral, −1.3 mm; and dorso/ventral, −2.2 mm. Mice were single-housed after surgery and were allowed to recover for 5–7 days. The correct implantation of cannulas was checked by 10 ng angiotensin II intracerebroventricular (ICV) injection, which elicits an intake of water (39). Animals that did not reach this criterion were excluded from the experiments.

Arc, medial hypothalamus (MH; ventromedial and dorsomedial), PVN, and LH were quickly dissected in a stainless steel matrix with razor blades as described previously (40) and frozen in liquid nitrogen for further protein studies.

To determine if insulin or leptin induces Tub tyrosine phosphorylation in vivo, overnight fasted mice fed chow or HFD received an ICV injection of insulin (human recombinant insulin; Eli Lilly, Indianapolis, IN) or recombinant leptin (Calbiochem, San Diego, CA), and hypothalamic nuclei were quickly dissected and frozen in liquid nitrogen for further protein studies. The time and dose responses to ICV insulin or leptin on Tub tyrosine phosphorylation were also investigated. To inhibit Tub (5′AGG AAC ACC TTC TTG CCA T 3′) or protein tyrosine phosphatase 1B (PTP1B) (5′CCC AGC AGC GGC TTC TGC AT 3′) expression in the hypothalamus, we developed antisense oligonucleotide (ASO) and administered them by ICV injections in mice fed chow or fed HFD, respectively. Both ICV ASO treatments were administered twice per day (8:00 a.m. and 5:00 p.m.) for 5 days. Body weight, epididymal fat mass, cumulative food intake, 4-h to 8-h food intake in response to ICV insulin and leptin, fasting blood glucose and blood glucose levels in response to pyruvate (pyruvate test) (41), and PEPCK expression in fasting livers were measured in Tub ASO-treated mice. We also determined O₂ consumption and respiratory exchange rate (38). Akt, FoxO1, JAK2, and STAT3 phosphorylation in response to ICV insulin or leptin, respectively, were evaluated as described.

A pool of five to eight of each nucleus (Arc, MH, PVN, LH) was homogenized in extraction buffer. Immunoprecipitation (IP) and immunoblotting were performed as previously described (37 –39) using the following antibodies: antiphosphotyrosine, anti-IRβ (α-IR), anti-Tub (M-19), anti-β actin, anti-JAK2, anti-Akt, anti-STAT3, anti-FoxO1, anti-PTP1B, and anti-PEPCK (H-300) antibodies from Santa Cruz Technology (Santa Cruz, CA); and antipAkt Ser473, antipSTAT3, antipJAK2, and antipFoxO1 antibodies from Cell Signaling.

The protocol described was adapted for ICV injections (42). Briefly, a low dose of insulin (0.2 μg) or leptin (0.1 ng) was injected into the lateral ventricle to trigger IR and JAK2 autophosphorylation. After 7–10 min, Arc, MH, PVN, and LH were dissected and homogenized in extraction buffer (38) for IR or JAK2 IP. In another group of animals, Arc, MH, PVN, and LH were dissected without insulin or leptin stimulation and samples underwent IP with anti-Tub antibody. Samples were collected on protein A-Sepharose and combined as followed: IP IR plus IP Tub and IP JAK2 plus IP Tub. Then, the kinase activity was performed as previously described (42).

Fasted animals received an ICV injection of insulin or leptin or acetylcholine (Ach) (100 µmol/L; Sigma-Aldrich, St. Louis, MO). The specific phospholipase Cβ (PLCβ) inhibitor U73122 (1 μmol/L; Sigma-Aldrich) was previously injected to block Tub nuclear translocation induced by the hormones. The U73343 (1 μmol/L; Sigma-Aldrich), which weakly inhibits PLCβ, was injected as a negative control. Hypothalamic nuclei or whole hypothalamus were quickly dissected and frozen in liquid nitrogen for further processing. A pool of 20 hypothalamic nuclei or a pool of four hypothalami were used for plasma membrane and nucleus separation as described previously (43,44). The procedure was performed at 4°C. Plasma membrane and nucleus fractions were analyzed by IP and immunoblotting with anti-Tub (M-19) antibody.

Random fed or fasted (24 h) mice treated for 5 days with Tub ASO or sense received ICV insulin or leptin injection, and the hypothalami or hypothalamic nuclei were harvested after 6 h, quickly frozen in liquid nitrogen, and stored at −80°C for processing. Total RNA was obtained using RNeasy Mini Kit (Cat. 74106; Qiagen). Real-time PCR was performed using TaqMan RT-PCR Master Mix (Applied Biosystems) in an Mx3000P thermocycler (Stratagene). The Mx3000P software was used to calculate the cycle threshold for each reaction. Relative expression levels were determined using the comparative cycle threshold method with normalization of target gene expression levels to 18s. Primers and probes were purchased from Applied Biosystems and were described as TRH, Mm01182425_g1; NPY, Mm00445771_m1; AgRP, Mm00475829_g1; POMC, Mm00435874_m1; MCH, Mm01242886_g1; and Ore, Mm01964031_s1 for mouse. The polymerase chain reaction conditions were 2 min at 50°C, 10 min at 95°C, followed by 40 cycles at 95°C for 15 s and 60°C for 60 s. Real-time data were analyzed using the engine provided by Applied Biosystems.

Protein tyrosine phosphatase (PTPase) activity was measured as previously described (45,46).

Results are expressed as means ± SD. Significance was determined using two-tailed Student t test or one-way or two-way ANOVA with Bonferroni posttest, as appropriate, and differences were considered significant if P < 0.05. We used GraphPad Prism (GraphPad Software, San Diego, CA).

Previous studies have shown that Tub is expressed in the central nervous system (13). Here, we confirmed that Tub was expressed in multiple hypothalamic nuclei (Fig. 1A).

The Tub protein abundance was not different in hypothalamic nuclei after fasting and refeeding states. However, Tub-p-tyr was strongly induced after 1 h of refeeding, followed by a decrease after 2 h of refeeding in Arc, MH, PVN, and LH in control mice (Fig. 1B).

Insulin-induced Tub-p-tyr was dose-dependent. The presence of phosphorylated Tub was detectable after ICV injection of as little as 0.02 μg insulin, half-maximal stimulation occurred with 0.2 μg of the hormone (Fig. 1C), and maximal stimulation was observed with 2.0 μg insulin (n = 3). Insulin-induced Tub-p-tyr occurred after 5 min of ICV insulin injection, and the maximal phosphorylation was observed at 10 min (Fig. 1D). The presence of phosphorylated Tub was detectable after ICV leptin injection of as little as 0.1 ng leptin, and the maximal stimulation was observed with 10.0 ng leptin (Fig. 1E). Leptin-induced Tub-p-tyr occurred after 2 min of ICV leptin injection, and the maximal phosphorylation was observed at 7 min (Fig. 1F).

Next, we compared the maximal effect of insulin and leptin in Arc, MH, PVN, and LH hypothalamic nuclei, and the results showed that in all these nuclei insulin or leptin induced similar increases in Tub-p-tyr (Fig. 1G).

The IR kinase activity was increased significantly in Arc lysates after ICV insulin in vivo and there was a further increase after ATP addition in vitro in mice fed chow. Most importantly, Tub-p-tyr was increased in Arc of mice injected with a low dose of insulin, with a further increase after ATP addition in vitro, demonstrating enhanced IR kinase activity with Tub. These results were also demonstrated in MH, PVN, and LH (Fig. 2A).

Similarly to insulin, JAK2 kinase activity was increased significantly in Arc lysates in response to leptin injection, as demonstrated by an increase in JAK2 autophosphorylation. The addition of Tub to the reaction also demonstrated that after leptin injection, Tub was a substrate of JAK2 tyrosine kinase in this nucleus. These results were also demonstrated in MH, PVN, and LH (Fig. 2B). We reprobed the membranes and confirmed that the bands blotted with antiphosphotyrosine antibody were IR, JAK2, and Tub (Fig. 2A and B).

Previous studies suggest that Tub acts as a transcription factor that can translocate from the plasma membrane to the nucleus after G-protein activation (47). Here, we observed that ICV insulin induced the nuclear translocation of Tub after 15 min in Arc, MH, and LH, and this translocation was maximal after 30 min in these nuclei. In PVN, insulin-induced Tub translocation was delayed compared with other nuclei and was evident at 90 min after the hormone injection. Leptin also was able to induce Tub nuclear translocation in Arc, MH, LH, and in PVN (Fig. 3A–D). In Arc, MH, and LH, leptin-induced Tub nuclear translocation was observed at 7 min, but in PVN the maximal translocation was observed 20 min after leptin. We blotted 30 μg of protein from the plasma membrane and nucleus with anti-IR and anti-TATA-box antibodies, respectively, as controls for these preparations (data not shown).

Previous data showed that Ach stimulates translocation of Tub in Neuro-2A cells, and this involves PLCβ, which is blocked by a specific PLCβ inhibitor (U73122). We investigated whether Ach is able to induce Tub translocation in vivo in the hypothalamus of mice fed chow. We injected Ach (100 µmol/L) alone or U73122 before Ach ICV injection, or U73343, which weakly inhibits PLCβ, as a negative control. We observed that Ach induced Tub nuclear translocation, an effect that was blocked by U73122, but not by U73343 (Fig. 3E–G). To investigate whether insulin- or leptin-induced Tub translocation also involved PLCβ, we injected insulin or leptin in mice previously treated with ICV U73122. The results showed that U73122 blocked insulin-induced and leptin-induced Tub translocation in vivo (Fig. 3H and I). It is interesting that ICV Ach altered TRH expression in the hypothalamus. Further, insulin or leptin induced an increase in POMC and TRH and a reduction in orexin or MCH expression, respectively. The ICV U73122 injection before Ach or insulin or leptin blunted these effects on neuropeptides (Supplementary Fig. 1A–C).

Control mice treated for 5 days with Tub ASO had a decrease of 80–90% in Tub expression in the hypothalamus compared with sense-treated mice (Fig. 4A). Mice treated with ASO had higher body weight and epididymal fat mass associated with an increase in food ingestion compared with sense-treated mice (Fig. 4B–D). Additionally, there was a decrease in O₂ consumption without changes in the respiratory exchange rate (Fig. 4E and F). There was also an increase in fasting blood glucose levels in ASO-treated mice (Fig. 4G). Furthermore, we observed higher blood glucose levels after pyruvate intraperitoneal injection at 15 and 30 min and enhanced PEPCK expression in fasting livers of mice treated with ASO (Fig. 4H and I). These differences persisted in pair-fed mice (Supplementary Fig. 2A–D).

The anorexigenic effect of ICV insulin or leptin was measured by recording 4-h and 8-h food intake. Food intake was reduced in response to ICV insulin or leptin in mice treated with sense, but not in mice treated with ASO (Fig. 5A and B).

Next, we investigated whether ASO treatment for 5 days affects the influence of insulin or leptin on neuropeptide expression in the hypothalamus. As expected, fasting increased the expression of NPY, AgRP, MCH, and orexin, and insulin or leptin was able to markedly reduce expression of these neuropeptides in sense-treated mice (Fig. 5C, D, F, and G). However, in mice treated with ASO, leptin-induced reduction of MCH mRNA was blunted (Fig. 5F) and insulin-induced reduction of orexin was also blunted (Fig. 5G). In contrast, fasting decreased the expression of POMC and TRH, and insulin or leptin increased the expression of these neuropeptides in mice treated with sense; however, in mice treated with ASO, leptin-induced or insulin-induced increases were blunted for POMC and TRH (Fig. 5E and H).

In mice treated with ASO, insulin-induced Akt and Foxo1 phosphorylation and leptin-induced JAK2/STAT3 tyrosine phosphorylation were not altered, indicating that the blunted insulin and leptin effects on food intake in mice treated with ASO are independent from these pathways (Fig. 5I–L).

We next investigated the modulation of insulin-induced or leptin-induced Tub-p-tyr in obese mice. Acute treatment with ICV insulin increased Tub-p-tyr in Arc, MH, PVN, and LH from control mice, and this effect was reduced in mice fed a HFD (Fig. 6A). Similarly, ICV injection of leptin increased Tub-p-tyr in Arc, MH, PVN, and LH in control mice, and this effect was blunted in mice fed a HFD (Fig. 6B). Insulin-induced or leptin-induced Tub nuclear translocation was blunted in the hypothalamus of DIO mice (Fig. 6C).

Because PTP1B dephosphorylates proteins that bind the IR and JAK2, we next investigated if insulin and leptin can induce an increase in Tub/PTP1B association in hypothalamic nuclei using coimmunoprecipitation. Our data showed that after these hormones were injected, there was an increase in Tub/PTP1B interaction in all hypothalamic nuclei investigated (Fig. 7A–D). In addition, we investigated Tub/PTP1B association in mice fed a HFD, after leptin insulin infusion, and the results showed that in obese mice there was a marked increase in Tub/PTP1B association in response to insulin and leptin in Arc, MH, PVN, and LH nuclei (Fig. 7A–D). Then, we treated mice with ICV PTP1B ASO for 5 days, which decreased PTP1B protein level in the hypothalamus (Fig. 7E), and we measured PTPase activity associated with PTP1B (Fig. 7F) and with Tub (Fig. 7G). We observed an increase in PTP1B and Tub associated with PTPase activity in the hypothalamus of mice fed a HFD, and this result was blunted by PTP1B-ASO treatment, suggesting that PTP1B may have phosphatase activity with Tub (Fig. 7F and G). Furthermore, we investigated the modulation of insulin-induced or leptin-induced Tub-p-tyr levels in mice fed a HFD. The results showed that the reduction in PTP1B protein level in hypothalamus was accompanied by an increase in Tub-p-tyr levels after leptin or insulin ICV treatment in DIO mice, suggesting that PTP1B may have a role in the dephosphorylation of Tub (Fig. 7H).

The results of the current study demonstrate that Tub is a direct substrate of insulin and leptin in vivo. Because the insulin receptor has tyrosine kinase activity, it induced Tub tyrosine phosphorylation directly; however, for leptin, similar to other substrates, it uses JAK2 as a kinase to induce Tub phosphorylation in Arc, MH, PVN, and LH. Our data also show that Tub-p-tyr is decreased during fasting and increased after refeeding, indicating that Tub is regulated by nutritional status. The combination of an increase in cumulative food intake, lower anorexigenic effect of insulin and leptin, and reduced O₂ consumption may contribute to explain enhanced adiposity in mice treated with Tub ASO.

Previous data showed that Tub overexpression, restricted to the Arc by a herpes simplex viral system, had no effect on food intake in rats fed chow, suggesting that Tub expression beyond normal may not be functional (48). Here, we demonstrate that the inhibition of Tub expression by ASO in all hypothalamic nuclei enhanced food intake and adiposity. It is possible that the manipulation of the amount of Tub only in neurons from one nucleus induced such a discrepant result. The regulation of food intake and adiposity is made by an interaction of different kinds of neurons in multiple brain regions (1–5). The response is variable and depends on which neuron is targeted (1–5). The effect of Tub inhibition on the control of metabolism seen in our experiments with ASO may be a reflection of the circuit of leptin-responsive and insulin-responsive neurons in multiple hypothalamic nuclei. Further, although Tub acts as a transcription factor moving from the plasma membrane to the nucleus (47), it was not shown in this previous study (48) whether the elevated amount of Tub altered the translocation of Tub from the plasma membrane to the nucleus.

The results of the current study showed that after insulin and leptin stimulation, there was a clear translocation of Tub from the plasma membrane to the nucleus in the Arc, MH, PVN, and LH being slower in PVN. The delay in insulin- and leptin-induced Tub translocation in the PVN cannot be related to a lack of these hormones penetration into the area because PVN, as well as Arc, is periventricular. The IR expression in PVN is higher than in LH, but PVN has the lowest LEPR expression of all other nuclei, which may contribute to explaining the delay in leptin-induced Tub translocation in this nucleus. In addition, PVN receives neuronal projections from Arc; thus, we can speculate that modulation of neural activity in Arc induced by insulin or by leptin might delay the effect of these hormones on Tub translocation in PVN.

Previously, it was demonstrated that activation of G-protein αq by Ach releases Tub from the plasma membrane through the action of PLCβ, triggering translocation of Tub to the nucleus (47). In this regard, our data suggest that insulin and leptin also use PLCβ to induce Tub translocation, because an inhibitor of PLCβ was able to block the effects of insulin and leptin on Tub translocation. Here, we demonstrated that after reduction in Tub protein expression, insulin or leptin was unable to increase the mRNA of POMC in Arc or the mRNA of TRH in PVN. Furthermore, insulin and leptin were unable to decrease the mRNA of orexin and MCH, respectively, in LH, suggesting that Tub may have a role modulating neuropeptides expression.

The inhibition of Tub protein expression in mice treated with ASO induced an increase in fasting blood glucose, hepatic glucose production, and the PEPCK expression of the liver in mice fed a chow diet. This phenomenon seems to be independent of changes in body weight or food intake, because it persists in pair-fed mice. These data suggest that in parallel to control energy balance, Tub in the hypothalamus may play a role in the regulation of hepatic glucose metabolism.

Resistance to leptin and insulin signaling is likely to play a central role in obesity (39,49,50). Our data showing a reduction in Tub tyrosine phosphorylation and nuclear translocation in response to insulin or leptin may contribute to hypothalamic insulin and leptin resistance in mice fed a HFD.

In the current study, we observed that Tub associates with PTP1B, and this interaction is increased in mice fed a HFD. Further, we observed that there was an increase in Tub associated with PTPase activity in the hypothalamus of mice fed a HFD, and this result was blunted by PTP1B-ASO treatment, suggesting that Tub may be a PTP1B substrate. This event could be one of the molecular mechanisms by which leptin-induced or insulin-induced Tub tyrosine phosphorylation is blunted in mice fed a HFD, suggesting that in addition to dephosphorylating IR, IRS-1, and JAK2, PTP1B also may have phosphatase activity with Tub.

In summary, our data provide evidence that Tub is a substrate of IRTK and LEPR-associated JAK2 in vivo in hypothalamic nuclei. After leptin or insulin injection, Tub translocates to the nucleus (Fig. 8A), and this is associated with altered POMC, TRH, MCH, and orexin mRNA levels. The reduction in Tub protein expression in the hypothalamus by Tub ASO treatment increases food intake, adiposity, fasting blood glucose, and hepatic glucose production, and decreases O₂ consumption, suggesting that Tub has a key role in the control of insulin and leptin effects on energy balance and glucose metabolism. In the hypothalamus of mice fed a HFD, there is a marked reduction in leptin- and insulin-induced Tub-p-tyr and translocation to the nucleus, which was accompanied by an increase in Tub/PTP1B interaction (Fig. 8B). These results demonstrate that Tub is part of the hypothalamic insulin and leptin signaling in vivo that acts independently of Akt/Foxo1 and STAT3 pathways, and its regulation has marked effects on metabolism (Fig. 8A and B).

This work was supported by FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo; 2008/55674-8), Sao Paulo, Brazil, CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) Universal (480131/2009-0), and by INCT (Instituto Nacional Ciência e Tecnologia de Obesidade e Diabetes; 573856/2008-7).

No potential conflicts of interest relevant to this article were reported.

P.O.P. performed experiments and wrote the manuscript. P.G.F.Q., A.M.C., A.C.S., D.G., J.M., and L.W. performed experiments. E.R.R., J.B.C.C., and L.A.V. contributed to discussion and reviewed the manuscript. M.J.A.S. also wrote the manuscript. M.J.A.S. and P.O.P. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.

The authors acknowledge the technical assistance of Luis Janieri, Jósimo Pinheiro, Lígia Parreira Muniz, and Ramon Henrick Zorzeto dos Santos from the Department of Internal Medicine, State University of Campinas.