Lack of Brain Insulin Receptor Substrate-1 Causes Growth Retardation, With Decreased Expression of Growth Hormone–Releasing Hormone in the Hypothalamus

Insulin and insulin-like growth factor 1 (IGF-1) play crucial roles in a wide variety of biological processes, including somatic growth and metabolism. These actions are mediated by intracellular signaling cascades activated downstream of two highly related tyrosine kinase receptors for insulin and IGF-1, the insulin receptor (IR) and the IGF-1 receptor (IGF1R), respectively (1). Both stimulation of the IR and IGF1R are known to activate similar signaling molecules. Insulin receptor substrates 1 and 2 (Irs1 and Irs2) are the 2 major substrates for IR and IGF1R tyrosine kinases, and the phosphorylated Irs proteins interact with the SH2 domains of downstream molecules such as phosphatidylinositol 3-kinase (PI3K), to produce numerous actions (2,3). Although IR and IGF1R are expressed in various organs and tissues of the body and phosphorylate the same downstream targets, Irs1 and Irs2, the physiological functions of the signaling pathways activated by them are thought to be different. We and other investigators have revealed the different roles of IR and IGF1R signaling in the whole body/specific tissues by using systemic and tissue-specific gene-knockout (KO) mice for these signaling molecules (4–7).

Mice with genetic deletion of IR in the brain (NIRKO mice) showed normal development but exhibited obesity with increased food intake and insulin resistance (8). Systemic Irs2KO mice also showed increased food intake, body weight, and adiposity, with changes in the mRNA expression levels of appetite-related neuropeptides (9–11). Furthermore, mice with genetic deletion of Irs2 in the brain (NIrs2KO mice) have been reported to display hyperphagia, obesity, and glucose intolerance, but not growth retardation, similar to the phenotype of the NIRKO mice (12). On the other hand, mice with systemic or brain-specific heterozygous deletion of IGF1R showed normal food consumption but growth retardation (13,14). Systemic Irs1KO mice also showed growth retardation, with IGF-1 and insulin resistance (15,16). In other words, growth-related phenotypes were also observed in the IGF1R- and Irs1-deficient mice. On the other hand, obesity-related phenotypes are known to be observed in the IR- and Irs2-KO mice. These data suggest that IR signaling and IGF1R signaling in the brain appear to be functionally associated with Irs2 and Irs1, respectively.

In this study, we generated mice with brain-specific deletion of Irs1 (NIrs1KO mice) to investigate the role of Irs1 in brain. The NIrs1KO mice showed growth failure, in terms of both the weight and length, with decreased pituitary GH and hepatic IGF-1 expressions. These mice also showed increased insulin sensitivity and enhanced glucose utilization in the skeletal muscle. Hypothalamic GHRH mRNA expression levels in the NIrs1KO mice were significantly lower than those in the control mice. The expression levels of GHRH protein in the median eminence (ME), where Irs1 is expressed and GHRH neurons project to, were markedly decreased in the NIrs1KO mice. The downregulation of Irs1 attenuated the IGF-1–induced neurite elongation in the cultured N-38 hypothalamic neurons. With findings taken together, brain Irs1 appears to play important roles in the regulation of somatic growth and glucose metabolism.

NIrs1KO mice were generated by mating of Irs1^flox/+ female mice (17,18) with transgenic mice expressing Cre under control of the Nestin promoter (Nestin-Cre mice). The Irs1^flox/+:Nestin-Cre male offspring were then crossed with Irs1^flox/+ female mice to obtain wild-type (Irs1^+/+), Nestin-Cre (Irs1^+/+:Nestin-Cre), Irs1^flox/lox, and NIrs1KO (Irs1^flox/flox:Nestin-Cre) mice. Syn-Irs1KO mice were generated by mating of Irs1^flox/+ female mice with transgenic mice expressing Cre under control of the synapsin-1 promoter (Synapsin-Cre mice). The Irs1^flox/+ :Synapsin-Cre female offspring were then crossed with Irs1^flox/+ male mice to obtain wild-type (Irs1^+/+), Synapsin-Cre (Irs1^+/+:Synapsin-Cre), Irs1^flox/flox, and Syn-Irs1KO (Irs1^flox/flox:Synapsin-Cre) mice. We confirmed the expression levels of GH, IGF-1, and GHRH in these mice (Supplementary Fig. 1A–F). Generation of liver-specific Irs1-KO (LIrs1KO) mice has previously been described (17). Genotyping was performed by PCR amplification of the tail DNA from each mouse at 4 weeks of age, as previously reported (17,18). All the mouse lines were maintained on a C57BL/6 background. All experiments in this study were performed with male mice. The mice were housed under a 12-h light/dark cycle at 22.5–23.5°C and given ad libitum access to regular chow, CE-2 (CLEA Japan), consisting of 25.6% (w/w) protein, 3.8% fiber, 6.9% ash, 50.5% carbohydrates, 4% fat, and 9.2% water. The animal care and experimental procedures used in this study were in compliance with the guidelines of the Animal Care Committee of the University of Tokyo.

Computed tomography (CT) was performed to determine the fat volume (Hitachi Aloka Medical). The bone length (femur and tibia), bone volume, bone mineral content (BMC), bone mineral density (BMD), and bone thickness were also measured with the software provided by the manufacturer. For the CT imaging, the mice were anesthetized by inhalational anesthesia (isoflurane at 3%, induction, and 1%–2%, maintenance).

Plasma levels of insulin, IGF-1, growth hormone, adrenocorticotropic hormone (ACTH), and prolactin, were determined with mouse insulin (Morinaga Institute of Biological Science, Inc.), mouse/rat IGF-1 (R&D Systems), rat/mouse growth hormone (Millipore), ACTH (Phoenix Pharmaceuticals), and mouse prolactin (CUSABIO) ELISA kits, respectively. Plasma triglyceride, nonesterified fatty acid, and total cholesterol levels (Wako Pure Chemical Industries, Ltd.) were assayed with enzymatic methods. Blood samples for IGF-1, growth hormone, ACTH, and prolactin assay were collected from the mice fed ad libitum at 9:00 a.m.

GHRP-2 was administered as previously described (19), with some modifications. The mice were administered GHRP-2 by intraperitoneal injection (LKT Laboratories) at the dose of 0.4 μg/g body wt at 9:00 a.m. Ten min later, blood samples were collected and centrifuged in heparinized tubes, and the separated plasma samples were used for the measurement of the plasma growth hormone concentrations.

RNA was isolated from the cells and tissues with the QIAGEN RNeasy Kit (QIAGEN, Hilden, Germany) in accordance with the manufacturer’s instructions. RNA (1 μg) was used for generating cDNA with use of random hexamers with MultiScribe reverse transcription reagents (Applied Biosystems). TaqMan quantitative PCR (50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min) was then performed with ABI PRISM 7900 PCR (Applied Biosystems) to amplify IGF-1 (Mm00439560_m1), GH (Mm00433590_g1), GHRH (Mm00439100_m1), somatostatin (Mm00436671_m1), POMC (Mm00435874_m1), AgRP (Mm00475829_g1), NPY (Mm00445771_m1), TSH-β (Mm00437190_m1), LH-β (Mm00656868_g1), and β-actin (Mm00607939_s1) cDNA from the samples. The primers were purchased from Applied Biosystems. The expression level of each of the transcripts was normalized to the β-actin mRNA expression level.

Isolated brain was fixed overnight in Bouin solution (picric acid, acetic acid, and formaldehyde) and then dehydrated and embedded in paraffin. Isolated liver was fixed overnight in 4% paraformaldehyde, appropriately cryopreserved, embedded in optimal cutting temperature compound, and frozen in liquid nitrogen. Hypothalamic and hepatic sections were cut at 4 and 10 μm, respectively, and mounted on silanized slides. For the immunohistochemical analysis, the sections were stained with anti-rabbit GHRH antibodies (1:200) and anti-rabbit Irs1 antibodies (hypothalamus, 1:300, and liver, 1:500). Secondary antibodies for immunofluorescence staining were Alexa Fluor 568 goat anti-rabbit and Alexa Fluor 488 goat anti-rabbit (Invitrogen).

The mouse-derived hypothalamic cell line N-38 was purchased from CELLutions Biosystems (Toronto, Canada) and used for the siRNA transfection, neurite outgrowth, cell viability, cell proliferation, cell count, and Akt phosphorylation experiments. Cells were maintained at 37°C under 5% CO₂ in DMEM (Invitrogen) supplemented with 10% FBS (Gibco), 20 mmol/L glucose, 0.15% sodium bicarbonate, and 1% penicillin/streptomycin. For investigation of neurite outgrowth, gene expressions, cell viability, cell proliferation, number of cells, and Akt phosphorylation, the cells were cultured for 24 h with or without siRNA targeting Irs1. Thereafter, the cells were incubated in low serum media (0.1%–1% FBS) for 60 min and then stimulated with 100 nmol/L human recombinant IGF-1 (R&D Systems) for 60 min. The harvested samples were used for quantification of neurite outgrowth, determination of the DNA content, gene expression profiles, cell proliferation assay, cell viability assay, cell count, and Western blot analysis. The lengths of neurites originating from different neurons (12 neurites for each condition) were measured and normalized against their respective controls (basal condition). DNA was prepared by overnight incubation with proteinase K (Merck Millipore) in digestion buffer (50 mmol/L Tris-HCl, pH 8.0; 100 mmol/L EDTA, pH 8.0; 100 mmol/L NaCl; and 1% SDS) at 55°C, followed by extraction with phenol/chloroform and ethanol precipitation.

siRNA targeting Irs1 was purchased from Applied Biosystems (s129868). Nontargeting siRNA (4390843; Life Technologies) was used as the negative control. siRNA was transfected into cultured cells with use of Lipofectamine RNAiMAX Reagent (Life Technologies) in accordance with the manufacturer’s instructions.

For evaluation of the cell viability, N-38 cells were incubated with 10 μL WST-1 (Roche) solution for 1 h before the measurements. We measured the absorbance at 450 nm using a microplate reader. For measurement of the cell proliferation, 10 μL BrdU (Roche) solution was added to these cells. We measured the BrdU incorporation rate by chemiluminescence assay in accordance with the manufacturer’s instructions.

For preparation of the lysates, the tissues were homogenized in buffer A (25 mmol/L Tris-HCl, pH 7.4; 10 mmol/L sodium orthovanadate; 10 mmol/L sodium pyrophosphate; 100 mmol/L sodium fluoride; 10 mmol/L EDTA; 10 mmol/L EGTA; 1% NP-40; and 1 mmol/L phenylmethylsulfonyl fluoride). For immunoprecipitation (IP) of Irs1 and Irs2, the lysates were incubated overnight at 4°C with rabbit polyclonal antibodies against Irs1 and Irs2, respectively. Then, Protein G-Sepharose was added, followed by incubation for 1 h at 4°C. Thereafter, after washing three times with buffer A, the immunocomplexes were resolved by 7% SDS-PAGE. Protein was analyzed by immunoblotting (IB) with specific antibodies against Irs1 (1:2,000) and Irs2 (1:2,000). Phosphorylated Akt (1:2,000) and total Akt protein (1:2,000) were visualized by IB with specific antibodies after the tissue lysates had been resolved by 10% SDS-PAGE and transferred to a Hybond-P PVDF transfer membrane (Amersham Biosciences). Bound antibodies were detected with horseradish peroxidase–conjugated secondary antibodies with use of ECL detection reagents (Amersham Biosciences).

Antibodies against Irs1 (IP and IB: cat. no. 06-248) and Irs2 (IB: MABS15) were purchased from Merck Millipore. Antibodies against Irs2 (IP: 3089), Akt (9272), and phosphorylated Akt (Ser⁴⁷³) (9271) were purchased from Cell Signaling Technology. Antibody against Irs1 (immunohistochemical analysis: PAC546 MU01) was purchased from Cloud-Clone Corp. Antibodies against GHRH (PAA438 MU01, LS-C482423) were purchased from Cloud-Clone Corp. and LSBio, respectively. Mouse monoclonal antibody directed against β-actin was purchased from Sigma-Aldrich (A5441). Horseradish peroxidase–conjugated goat anti-rabbit IgG (111-035-144) and goat anti-mouse IgG (155-035-003) were purchased from Jackson Immuno Research.

Mice were fed freely and then denied access to food for 2 h prior to the study. They were intraperitoneally challenged with regular human insulin (0.75 units/kg) in the morning at 9:00 a.m. We determined blood glucose levels in whole venous blood specimens obtained from the tail, using an automatic glucose monitor (Glutest Ace, Sanwa Chemical Industrial Co., Ltd.).

Mice were loaded with oral glucose at 1.5 mg/g body wt at 9:00 a.m., after being denied access to food for 16 h. Blood samples were taken at different time points, and the blood concentrations of glucose were measured with an automatic glucometer (Glutest Ace). Blood samples were collected and centrifuged in heparinized tubes, and the separated plasma samples were used for the measurement of the plasma insulin levels.

For investigation of insulin signaling in the liver, epididymal white adipose tissue, and skeletal muscle, insulin (6 ng/mL) (Humulin R; Lilly) was injected via the inferior vena cava. The liver, epididymal white adipose tissue, and skeletal muscle were dissected 10 min after the insulin injection and immediately frozen in liquid nitrogen. The samples were then analyzed by Western blot analysis.

Oxygen consumption of the mice was measured every 3 min for 24 h with an O₂/CO₂ metabolism measurement device (model MK-5000; Muromachikikai, Tokyo, Japan) as previously described (20). The respiratory quotient was calculated with the following equation based on the volume of O₂ consumed and the volume of CO₂ produced: respiratory quotient = VCO₂ / VO₂. Locomotor activity was measured with an activity-monitoring system (ACTIMO-100; Shinfactory, Fukuoka, Japan).

Data are expressed as means ± SEM. Statistical significance was calculated with the unpaired Student t test unless otherwise specified. For experiments involving multiple comparisons, data were analyzed with the Tukey-Kramer test. P values of <0.05 were considered indicative of statistical significance.

The data sets generated during the current study are available from the corresponding author upon reasonable request.

Nestin-Cre mice are known to show growth retardation compared with their littermates (21,22). We measured the body lengths and body weights of the wild-type, Irs1-floxed, Nestin-Cre, and NIrs1KO mice. Although the body length and body weight of the Nestin-Cre mice were significantly lower than those of the wild-type and Irs1-floxed mice, those of the NIrs1KO mice were even lower than the body length and body weight of the Nestin-Cre mice (Fig. 1A–C). We used Nestin-Cre mice as the control mice for all the experiments. Although Irs1 protein expression levels in the pancreas, heart, and kidney were equivalent between the Nestin-Cre and NIrs1KO mice, Irs1 protein expression was absent in the entire brain, including the cerebral cortex, cerebellum, hippocampus, and hypothalamus in the NIrs1KO mice (Fig. 1D and E). On the other hand, no significant differences in the Irs2 protein levels in the brain, including in the cerebral cortex, cerebellum, hippocampus, and hypothalamus, were seen between the Nestin-Cre and NIrs1KO mice (Fig. 1E). The bone length was also significantly shorter in the NIrs1KO mice compared with the Nestin-Cre mice (Fig. 1F). No significant differences were observed in the ratio of the brain, heart, liver, kidney, soleus muscle, or brown adipose tissue to body weight between the Nestin-Cre and NIrs1KO mice (Fig. 1G). CT showed that total fat weight–to–body weight ratio and the ratio of the visceral to subcutaneous fat area were not different between the Nestin-Cre and NIrs1KO mice (Fig. 1H). The bone volume, BMC, and BMD measured by CT were significantly reduced in the NIrs1KO mice compared with the Nestin-Cre mice (Fig. 1I). The bone thickness in the NIrs1KO mice tended to be smaller than that in the Nestin-Cre mice (Fig. 1I). Moreover, we investigated the growth-related phenotype in another brain-specific Irs1-KO mouse model (Syn-Irs1KO mice) to rule out the effect of Nestin promoter (Fig. 2D). Although the body length and body weight of the Synapsin-Cre mice were not different from those of the wild-type and Irs1-floxed mice, the body length and body weight of the Syn-Irs1KO mice were significantly shorter and lower, respectively, compared with those of the wild-type, Irs1-floxed, and Synapsin-Cre mice (Fig. 2A–C). The bone length was also significantly shorter in the Syn-Irs1KO mice compared with the Synapsin-Cre mice (Fig. 2E). No significant differences were observed in the ratio of the weights of the brain, heart, liver, kidney, soleus muscle, brown adipose tissue, and total fat to the body weight or in the ratio of the visceral to the subcutaneous fat area between the Synapsin-Cre and Syn-Irs1KO mice (Fig. 2F and G). The bone volume, BMC, and BMD measured by CT were significantly lower in the Syn-Irs1KO mice compared with the Synapsin-Cre mice (Fig. 2H). There were no differences in the food intake, mRNA expression levels of appetite-related neuropeptides, or the body temperature between the Nestin-Cre and NIrs1KO mice (Supplementary Fig. 2A–C). These data suggest that the lack of brain Irs1 causes growth retardation with normal proportions and impaired bone modeling.

We next investigated glucose and lipid metabolism in the NIrs1KO mice. A significant decrease of the blood glucose levels was observed in the NIrs1KO mice at 40 and 60 min after insulin injection (Fig. 3A). Although the plasma insulin levels did not differ between the Nestin-Cre and NIrs1KO mice, a significant decrease of the blood glucose levels was observed in the NIrs1KO mice at 15 min after oral glucose administration (Fig. 3B). Insulin-induced Akt phosphorylation in the skeletal muscle was increased in the NIrs1KO mice compared with the Nestin-Cre mice, whereas insulin-induced Akt phosphorylation in the liver and white adipose tissue did not differ between the Nestin-Cre and NIrs1KO mice (Fig. 3C). The respiratory exchange ratio in the dark cycle was significantly higher in the NIrs1KO mice compared with the Nestin-Cre mice, suggesting that the glucose utilization rate is increased in the skeletal muscle of the NIrs1KO mice (Fig. 3D). No differences were observed in the locomotor activities in the dark and light cycles between the Nestin-Cre and NIrs1KO mice (Fig. 3E). No significant differences in the plasma total cholesterol, triglyceride, or nonesterified fatty acid levels were observed between the Nestin-Cre and NIrs1KO mice (Fig. 3F). These data suggest that deletion of Irs1 in the brain results in increased insulin sensitivity and glucose utilization in the skeletal muscle.

To investigate the mechanism of growth retardation and increased insulin sensitivity observed in the NIrs1KO mice, we measured the plasma and mRNA levels of GH and IGF-1. Plasma and pituitary expression levels of GH were significantly reduced in the NIrs1KO mice (Fig. 4A and B and Supplementary Fig. 1A). Plasma and hepatic expression levels of IGF-1 were also significantly decreased in the NIrs1KO mice (Fig. 4C and D and Supplementary Fig. 1B). Plasma and mRNA levels of GH and IGF-1 were also significantly decreased in the Syn-Irs1KO mice, as in the NIrs1KO mice (Supplementary Figs. 1D and E and 3A and B). On the other hand, no significant differences were observed in the plasma levels of adrenocorticotropic hormone (ACTH) and prolactin (PRL) or in the mRNA expression level of thyroid-stimulating hormone (TSH) and luteinizing hormone (LH) between the Nestin-Cre and NIrs1KO mice (Fig. 4E–H). To investigate GH secretion from the pituitary, we administered growth hormone–releasing peptide-2 (GHRP-2), which is known as a strong secretagogue of GH, by intraperitoneal injection to the mice. Plasma GH levels after GHRP-2 injection did not differ between the Nestin-Cre and NIrs1KO mice (Fig. 4I), suggesting that GH secretion from the pituitary is not impaired in the NIrs1KO mice. Consistent with these data, no decrease in the Irs1 protein level in the pituitary was observed in the NIrs1KO mice (Fig. 4J). We next measured the hypothalamic GHRH and somatostatin expression levels. GHRH mRNA expression levels in the hypothalamic arcuate nucleus were significantly reduced in the NIrs1KO mice compared with the Nestin-Cre mice (Fig. 4K and Supplementary Fig. 1C). GHRH mRNA expression levels in the hypothalamic arcuate nucleus were also significantly decreased in the Syn-Irs1KO mice, as in the NIrs1KO mice (Supplementary Fig. 1F). On the other hand, hypothalamic somatostatin expression levels did not differ between the Nestin-Cre and NIrs1KO mice (Fig. 4L). Consistent with the pattern of Irs1 staining, immunohistochemistry for GHRH (primary antibody: cat. no. PAA438 MU01; Cloud-Clone Corp.) in the hypothalamic ME demonstrated strong staining in the wild-type mice (Fig. 4M and Supplementary Fig. 4A), although a clear merged image was not observed in the serial sections by immunofluorescence staining (Supplementary Fig. 4A). We carried out the immunohistochemical staining using another antibody for GHRH (LS-C482423; LSBio) in the wild-type mice. The latter antibody (LS-C482423) stained the GHRH neurons in almost the same area as the former antibody (PAA438) (Supplementary Fig. 4B). No nonspecific binding with the secondary antibody used for GHRH or Irs1 staining was observed in the ME (Supplementary Fig. 4C). Positive staining for GHRH was also found in the ME of the Nestin-Cre mice, but only weak staining was found in the ME of the NIrs1KO mice (Fig. 4N). These data suggest that lack of Irs1 in the hypothalamus causes a reduction of the hypothalamic GHRH expression levels, resulting in a decrease of the pituitary GH and hepatic IGF-1 expression levels. We confirmed positive staining with Irs1 antibody in the liver of the wild-type, but not liver-specific Irs1 KO, mice (17) (Supplementary Fig. 4D).

To investigate the role of Irs1 in neurite outgrowth and proliferation of GHRH neurons, we induced downregulation of Irs1 expression in a GHRH-producing cultured N-38 hypothalamic cell line by the addition of an siRNA (Fig. 5A). IGF-1 stimulation is reported to promote neurite elongation of GHRH neurons via inducing PI3K/Akt signaling (23,24). Downregulation of Irs1 was associated with significant impairment of IGF-1–induced phosphorylation of Akt and neurite outgrowth in these cells (Fig. 5B and C). Consistent with these data, the GHRH mRNA expression levels after IGF-1 treatment were significantly decreased in the cells with downregulated Irs1 (Fig. 5D). Similarly, enhanced cell viability after IGF-1 stimulation was significantly suppressed by downregulation of Irs1 in these cells (Fig. 5E). There were no differences in the BrdU incorporation rate or number of cells after IGF-1 administration between the control and Irs1-downregulated neurons (Fig. 5F and G). These data suggest that Irs1 plays a role in neurite outgrowth but not in the proliferation of GHRH-producing cultured N-38 hypothalamic neurons.

In this study, we demonstrated that the NIrs1KO mice showed growth retardation, with decreased GHRH expression in the hypothalamus. Protein levels of GHRH in the ME were also significantly reduced in the NIrs1KO mice. While the pituitary Irs1 expression levels were maintained, the pituitary GH and hepatic IGF-1 expression levels were reduced in the NIrs1KO mice. Knockdown of Irs1 inhibited IGF-1–induced neurite elongation of GHRH-producing cultured N-38 hypothalamic neurons. Although the neuronal number or neural proliferative activity remained unchanged, IGF-1–induced enhancement of cell viability was significantly suppressed in the Irs1-knockdown neurons. These data suggest that brain Irs1 plays important roles in somatic growth via inducing elongation of the GHRH neurons in the hypothalamus (Fig. 6).

The NIrs1KO mice were insulin sensitive, just like the GHRHKO mice and GHRKO mice (25,26). It has been reported that GH specifically increases the monomeric p85, which competitively inhibits the binding of the p85-p110 heterodimer to the Irs protein, resulting in insulin resistance (27,28). GH levels have been shown to be negatively correlated with insulin-induced glucose uptake in the skeletal muscle in GH-deficient and GH-overexpressing mice (29). Insulin sensitivity was increased in muscle-specific GH receptor–KO mice (30). On the other hand, no change in insulin sensitivity was noted in adipose tissue–specific GH receptor–KO mice and muscle-specific IGF1R-KO mice (5,31). These data suggest that the increased insulin sensitivity and glucose utilization observed in the NIrs1KO mice might be induced by the decreased effect of GH on the skeletal muscle (Fig. 6).

In addition to the NIrs1KO mice, mice with heterozygous deletion of IGF1R in the brain (bIGF1RKO^+/− mice), but not the NIRKO mice, also showed growth retardation (8,14). The phenotypes of NIRKO mice, which exhibit obesity and increased food intake and no growth impairment, were not observed in the NIrs1KO and bIGF1RKO^+/− mice. These data suggest that Irs1 mainly transmits signals from IGF1R, and not IR, in the brain. In the vascular smooth muscle cells, activation of IGF1R promoted Irs1-associated PI3K activation and cell proliferation (32). In Irs1-deficient 3T3 cells, the activation of IGF1R failed to promote cell growth, whereas reexpression of Irs1 in the deficient cells restored the cell cycle progression (33). In IR-deficient hepatocytes, activation of IGF1R induced Irs1 phosphorylation and binding to PI3K, resulting in an increase in cell growth, without modulating other metabolic activities, such as the activities of glucokinase and glycogen synthase (34). These data suggest that activation of IGF1R preferentially recruits Irs1 and results in growth-promoting effects in many organs or tissues, including the brain. An important question that arises from these data is why Irs1 preferentially transmits signals from IGF1R rather than from IR. The structure of IGF1R is highly homologous to that of IR; both receptors are tetramers of two α and two β subunits linked by disulfide bonds. The intracellular domains of the β subunits contain the juxtamembrane region after the transmembrane helix, followed by the tyrosine kinase domain and the C-terminal region. The structures of the tyrosine kinase domains of both the receptors are almost identical and are important for receptor autophosphorylation and tyrosine phosphorylation of the substrates. On the other hand, the homology of their juxtamembrane and C-terminal regions, which provide docking sites for many substrates and adopter proteins, is relatively low. Although the NPEY (Asn-Pro-Glu-Tyr) motif in the juxtamembrane region is crucial for the recruitment of Irs protein, amino acid residues around the NPEY motif differ between IGF1R and IR (35,36). Replacement of an amino acid residue in this region changed the affinity of Irs1 to the receptor in cultured preadipocytes (36). The differences in the amino acid sequence adjacent to the NPEY motif are among the candidate determinants of the substrate preferences of IGF1R and IR.

Although Irs1 is known to be expressed throughout the brain, the phenotype of the NIrs1KO mice was somatotroph specific, and other neuroendocrine pathways were unaffected. Despite the ubiquitous distribution of IGF1R in the brain, the bIGF1RKO^+/− mice showed the growth retardation phenotype, with decreased GHRH, but not any of the other neuroendocrine phenotypes—like the NIrs1KO mice (14). IGF-1 has been reported to promote neurite elongation of the GHRH neurons but not of the AgRP or gonadotropin-releasing hormone (GnRH) neurons (23,37). Brain size (brain-to-body ratio) was preserved in the NIrs1KO mice, as well as in the bIGF1RKO^+/− and systemic Irs1KO mice (14,38). These data suggest that impairment of neurite outgrowth by the lack of Irs1 or IGF1R may be GHRH neuron–specific and not be observed in all neurons or glia. Why does lack of Irs1 or IGF1R in the brain preferentially impair neurite elongation in the GHRH neurons, even though both Irs1 and IGF1R are normally expressed throughout the brain? The ME contains many tanycytes, which are ependymal cells, that absorb IGF-1 from the cerebrospinal fluid and transport it to the brain (39). Thus, GHRH neurons that exist in the arcuate nucleus and project to the ME are presumed to be sensitive to IGF-1 exposure. That may be one reason why growth retardation with decreased hypothalamic GHRH expression is observed in the NIrs1KO and bIGF1RKO^+/− mice. We cannot completely exclude the possibility that Irs1 deletion in other brain lesions affects the GHRH neurons at a distance, and further investigations are needed to clarify the mechanisms of neurite outgrowth in various subpopulations of neurons.

In this study, a lack of brain Irs1 decreased GHRH expression in the hypothalamus, just like deletion of IGF1R in the brain. However, stimulation or inhibition of the IGF1R-Irs1 pathway did not change the proliferative activity or number of GHRH-producing cultured N-38 hypothalamic neurons. Consistent with these data, it has been reported that while the number of GHRH neurons was not reduced, the axonal growth of GHRH neurons was impaired in underfed mice with decreased GHRH levels (23). It has been demonstrated that IGF-1 promotes neurite outgrowth by incorporating plasmalemmal precursor vesicles into the neural cell surface via PI3K/Akt signaling at the nerve growth cone (24,40). These data suggest that the IGF1R-Irs1 pathway may play an important role in the neural outgrowth, but not in the proliferation, of growing GHRH neurons. Neurogenesis or neural outgrowth occurs not only during the prenatal and early postnatal periods but also in pathophysiological conditions such as cerebral ischemia, brain injury, and neurodegenerative disease (41–44). Although phosphorylation of Akt was decreased in the postischemic rat brain, intracerebroventricular administration of IGF-1 enhanced phosphorylated Akt expression, reducing the infarct volume in the ischemia-induced brain (45). Intranasal administration of IGF-1 restored Akt phosphorylation and promoted cell growth in the injured brain of a rat model of cerebral hypoxia-ischemia, resulting in pathological and functional neuroprotection (46). Exposure to IGF-1 of the motoneurons of a chick embryo or denervated gluteus muscle in the adult rat or mouse promoted neurite outgrowth and intramuscular nerve sprouting (47). These data suggest that IGF-1 signaling plays an important role in neural growth or neuroprotection in pathophysiological conditions. Increased phosphorylation levels of Irs1 at serine residues were observed in postmortem human brains in cases of Alzheimer disease, Pick disease, corticobasal degeneration, and progressive supranuclear palsy, as well as in those from diet-induced type 2 diabetic mice with cognitive impairment and mouse models of Alzheimer disease (48–50). Although the role of brain Irs1 in pathophysiological conditions remains unknown, IGF-1–Irs1 signaling may contribute to neural growth or provide a neuroprotective effect, and impairment of this signaling pathway may be involved in the pathogenesis or aggravation of the neurological diseases mentioned above.

Nestin-Cre mice are known to show growth retardation, along with decreased GHRH levels, compared with their littermates (21,22). In the Nestin-Cre mice, the human growth hormone (hGH) minigene is inserted downstream of Cre recombinase to ensure efficient expression of the transgene, resulting in activation of signaling molecules downstream of hGH in the hypothalamus (22). Since this activation of signaling molecules downstream of hGH may reduce GHRH expression and cause growth retardation (22), it may be better to confirm, as we did in the current study, the growth-related phenotypes using another brain-specific cre mouse model (such as Synapsin-Cre mice).

Taking together the findings in the current study, we demonstrated that a lack of brain Irs1 decreases GHRH expression in the hypothalamus, resulting in growth retardation and increased insulin sensitivity/enhanced glucose utilization in the skeletal muscle. Our findings provide insight into novel roles of Irs1 in the brain in somatic growth and glucose metabolism.

Acknowledgments. The authors thank Ayami Gouda, Yurika Shiraishi, both at The University of Tokyo, Masatsugu Takayasu, Manami Takagi, Tamao Iwakami, all at National Institutes of Biomedical Innovation, Health and Nutrition, Tomoko Asano, The University of Tokyo, Eriko Nozaki, National Institutes of Biomedical Innovation, Health and Nutrition, and Kousuke Yokota, The University of Tokyo, for excellent technical assistance and assistance with the animal care.

Funding. This work was supported by a grant for TSBMI (Translational Systems Biology and Medicine Initiative) from the Ministry of Education, Culture, Sports, Science and Technology of Japan; Grants-in-Aid for Scientific Research (18H02860 and 15H04847) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to N.K.); and a Grant-in-Aid for Young Scientists (15K19505) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to T.H.).

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

Author Contributions. T.H., T.Ku., N.K., and TKa. designed this study and wrote the manuscript. T.H., TKu., and N.K. conducted the experimental research and analyzed the data. N.W. conducted the histological and immunohistochemical analysis. T.M. contributed to mice generation. M.I., I.T., and Y.S. contributed to data discussion. All the authors gave final approval of the manuscript version submitted for publication. N.K. and T.Ka. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.