Insulin acts on peripheral tissues including liver, muscle and adipose tissue to directly control glucose metabolism, while also acting in the brain to concordantly regulate nutrient fluxes, feeding behaviour and energy homeostasis (Varela and Horvath, 2012). The CNS effects of insulin on metabolism are mediated by different brain regions, in particular the hypothalamus, acting via autonomic circuits to influence peripheral organs, including the pancreas, liver, white adipose tissue and brown adipose tissue to modulate insulin secretion, endogenous glucose production and glucose uptake (Dodd et al., 2015; Könner et al., 2007; Obici et al., 2002b; Pocai et al., 2005; Steculorum et al., 2016; Vogt and Brüning, 2013; Vogt et al., 2014). Although our understanding of the neural circuitry controlling feeding behaviour and energy expenditure has grown considerably in the last few years, the neural processes by which insulin elicits its effects on glucose metabolism are less clear.

The insulin receptor (IR) is widely expressed throughout the brain (Havrankova et al., 1978). Mice lacking the IR in the brain exhibit whole-body insulin resistance accompanied by increased food intake and obesity (Brüning et al., 2000). The contributions of the IR in controlling feeding and energy expenditure are best understood in the arcuate nucleus (ARC) of the hypothalamus, a specialized brain region comprising of neurons proximal to the fenestrated capillaries of the median eminence. This position allows ARC neurons to respond to peripheral circulating factors such as leptin and insulin that convey information on the nutritional and metabolic status of the organism. In particular, the ARC comprises of two opposing neuronal populations, the appetite suppressing proopiomelanocortin (POMC)-expressing neurons and the orexigenic agouti-related peptide (AgRP)-expressing neurons that co-express neuropeptide Y (NPY) and γ-aminobutyric acid (GABA). The activation of POMC neurons results in POMC processing to generate α-melanocyte stimulating hormone (α-MSH), which is released and agonises melanocortin-4 receptors (MC4R) on neurons in other regions of the brain, such as the paraventricular nucleus of the hypothalamus (PVH) (Varela and Horvath, 2012). The activation of AgRP/NPY neurons promotes the synthesis and release of AgRP and GABA that antagonise α-MSH/MC4R interactions and directly inhibit POMC neurons, respectively (Atasoy et al., 2012; Cowley et al., 2001; Cowley et al., 2003; Tong et al., 2008).

Insulin binds to its tyrosine kinase receptor and signals via the phosphatidylinositol 3-kinase (PI3K)/protein kinase PKB/AKT pathway to hyperpolarise and inhibit AgRP/NPY neurons and repress Agrp expression (Könner et al., 2007; Spanswick et al., 2000; Varela and Horvath, 2012; Vogt and Brüning, 2013; Zhang et al., 2015). Leptin also hyperpolarizes and inhibits NPY/AgRP neurons and represses Npy/Agrp expression acting via several pathways, including the Janus-activated kinase (JAK)−2/signal transducer and activator of transcription (STAT)−3 pathway and the PI3K/AKT pathway (Cowley et al., 2001; Elias et al., 1999; van den Top et al., 2004; Varela and Horvath, 2012; Zhang et al., 2015). In this way leptin and insulin act to alleviate the inhibitory constraints on POMC neurons and the melanocortin response. IR signaling in AgRP neurons is important for the control of peripheral glucose metabolism (Dodd et al., 2018; Könner et al., 2007) acting through vagal efferents to the liver and α7-nicotinic acetylcholine receptors to repress hepatic glucose production (HGP) (Kimura et al., 2016). Central insulin signaling suppresses HGP by activating Kupffer cells to release interleukin- 6 (IL-6) (Obici et al., 2002a; Obici et al., 2002b; Pocai et al., 2005). IL-6 in turn acts on hepatocytes via STAT-3 to repress the expression of gluconeogenic enzymes such as glucose-6-phosphatase (encoded by G6pc) (Inoue et al., 2006; Inoue et al., 2004). IR deletion in AgRP neurons in mice results in defective inhibition of HGP accompanied by decreased Il6 and increased G6pc hepatic gene expression (Könner et al., 2007).

Leptin depolarises and activates POMC neurons and promotes Pomc expression and α-MSH secretion to repress feeding, increase energy expenditure and repress HGP (Berglund et al., 2012; Caron et al., 2018; Varela and Horvath, 2012). By contrast insulin has been traditionally viewed as an inhibitor of POMC neuronal excitation, hyperpolarising POMC neurons through the engagement of ATP-sensitive K⁺ channels (Hill et al., 2008; Könner et al., 2007; Plum et al., 2006; Spanswick et al., 2000; Williams et al., 2010). More recent studies have shown that insulin can depolarise and activate POMC neurons via the activation of transient receptor potential (TRPC)−5 channels (Qiu et al., 2018; Qiu et al., 2014), whereas several studies have shown that insulin can promote Pomc expression and the melanocortin response (Varela and Horvath, 2012). The advent of single-cell transcriptomics has yielded unprecedented insight into the heterogeneity of hypothalamic cell types and has led to the identification of different subsets of POMC- and AgRP-expressing neurons that exhibit distinct transcriptional responses to changing energy status (Campbell et al., 2017; Chen et al., 2017; Henry et al., 2015; Lam et al., 2017). Such heterogeneity may at least in part explain the seeming discordant effects of insulin on POMC neuronal excitability, but this remains to be formally examined. Irrespective, whereas insulin signaling in AgRP neurons overtly affects glucose metabolism through the control of HGP and brown adipose tissue glucose uptake (Könner et al., 2007; Shin et al., 2017), gene deletion and pharmacogenetic experiments have suggested that insulin signaling in POMC neurons may not be so important; neither IR deletion in POMC neurons (Könner et al., 2007; Shin et al., 2017) nor the acute activation of POMC neurons using DREADDs (Steculorum et al., 2016) influences glucose metabolism. Nonetheless, the combined deletion of the IR plus leptin receptor in POMC neurons profoundly affects glucose homeostasis and results in systemic insulin resistance and increased HGP (Hill et al., 2010). Thus, the contribution of insulin signaling in POMC neurons on glucose metabolism and energy expenditure may be more nuanced and dictated by the hormonal milieu.

We report that insulin can elicit discordant effects on the electrophysiological activity of POMC neurons and that this is regulated by T-cell protein tyrosine phosphatase (TCPTP) (Tiganis, 2013), an IR phosphatase whose abundance in the ARC is increased by fasting and repressed by feeding (Dodd et al., 2017). Moreover we demonstrate that the regulation of IR signaling by TCPTP dictates whether POMC neurons are activated or inhibited by insulin and that this serves to coordinate hepatic glucose metabolism with feeding and fasting. Our studies define a mechanism for linking POMC neuronal plasticity with the nutritional and metabolic state of the organism.

In this study we demonstrate that POMC neurons can be either inhibited or activated by insulin (or otherwise not respond) and that the balance of excitation versus inhibition is dictated by the phosphatase TCPTP. Moreover, we demonstrate that TCPTP may orchestrate these effects through the regulation of IR signaling to coordinate the melanocortin response and hepatic glucose metabolism in response to feeding and fasting. This provides an elegant system for integrating the melanocortin circuit with the nutritional state of the organism and coordinating feeding with the CNS control of glucose metabolism.

The IR is widely expressed in the hypothalamus and is present in both POMC and AgRP neurons (Havrankova et al., 1978; Könner et al., 2007). Although its role in AgRP neurons and HGP has been firmly established (Könner et al., 2007), its importance in POMC neurons and glucose metabolism has received less attention, as early studies showed that IR deletion in POMC neurons had no effect on HGP and insulin sensitivity (Könner et al., 2007). Our studies suggest that at least one explanation for this could be that any effect on glucose metabolism would be negated by the concomitant deletion of IR in POMC neurons that are excited and inhibited by insulin. Moreover, our studies indicate that any response of POMC neurons to insulin would be dictated by the nutritional state of the organism. C57BL/6 mice feed predominantly at night during the first four hours of the dark cycle (Dodd et al., 2017). Previous studies addressing the role of IR deletion in POMC neurons would have been conducted during the day (Könner et al., 2007; Shin et al., 2017), when mice are effectively fasted and TCPTP levels are elevated (Dodd et al., 2017) and/or in overnight fasted mice (Könner et al., 2007). Under these conditions, the TCPTP-mediated suppression of insulin signaling would render POMC neurons largely unresponsive or inhibited by insulin and antagonise the CNS insulin-mediated repression of HGP. Therefore, it is likely that the responses seen in our studies would have been missed by previous studies. Consistent with this assertion we found that ICV insulin repressed HGP in fed mice, but not in fasted mice. In addition, Steculorum et al. (Steculorum et al., 2016) have reported that the acute activation of AgRP neurons using DREADDs has no effect on HGP, but represses brown adipose tissue glucose uptake and insulin sensitivity, whereas the acute activation of POMC neurons has no discernible influence on glucose metabolism. Although this is consistent with the acute CNS regulation of glucose metabolism being independent of the melanocortin circuit, our studies demonstrate that the chronic inhibition or activation of POMC neurons, via DREADDs, or deletion of the phosphatase TCPTP, profoundly influences hepatic insulin sensitivity and glucose metabolism. The chronic activation or inhibition of POMC neurons decreased or increased hepatic gluconeogenesis and was accompanied by the promotion or repression of hepatic Il6 expression and STAT3 signaling and the repression or induction of gluconeogenic gene expression respectively. The impact of the pharmacogenetic inhibition of POMC neurons with DREADDs on hepatic glucose metabolism was substantiated in hyperinsulinemic euglycemic clamps where endogenous glucose production was markedly increased and systemic insulin sensitivity decreased. Conversely TCPTP deletion in POMC neurons resulted enhanced IR signaling and POMC neuronal excitation to repress HGP and significantly enhanced hepatic and whole-body insulin sensitivity. Moreover, the acute activation of AgRP neurons may elicit melanocortin-independent effects on brown adipose tissue glucose uptake (Steculorum et al., 2016), our recent studies indicate that TCPTP-deletion in AgRP enhances the insulin-mediated inhibition of AgRP neurons (Dodd et al., 2017) and the IR-dependent repression of HGP (Dodd et al., 2018). Therefore, the chronic activation of POMC neurons and concomitant inhibition of AgRP neurons by insulin would drive melanocortin-dependent responses to modulate HGP and glucose homeostasis. This may allow for a graded response to changing nutritional states, with acute changes in insulin acting via AgRP-dependent but melanocortin-independent pathways to affect BAT glucose uptake, and the sustained hyperglycemia and hyperinsulinemia accompanying more intense caloric intake additionally engaging the melanocortin circuit and the liver to allow for a concerted systemic response to lower postprandial blood glucose levels.

Our studies indicate that the abundance of the IR phosphatase TCPTP determines whether POMC neurons are excited or inhibited by insulin and controls the intensity of IR signaling. TCPTP deletion in POMC neurons (marked by the Pomc-eGFP reporter), or inhibition with the TCPTP selective inhibitor compound 8, shifted POMC neural responses so that the majority of POMC neurons were activated by insulin. Our recent studies have shown that TCPTP abundance in AgRP and POMC neurons is subject to diurnal fluctuations linked to feeding (Dodd et al., 2017). Increases in circulating corticosterone that accompany fasting drive hypothalamic TCPTP expression (Dodd et al., 2017). By contrast, feeding inhibits TCPTP expression and promotes its rapid degradation via the proteasome (Dodd et al., 2017). We have reported that in AgRP neurons this serves to coordinate the browning and thermogenic activity of white fat with the expenditure of energy linked to feeding (Dodd et al., 2017). In other studies we have shown that such fluxes in TCPTP in AgRP neurons also influence hepatic glucose metabolism (Dodd et al., 2018), which is in keeping with the previously established role for the IR in AgRP neurons in glucose metabolism (Könner et al., 2007). Although we did not assess the influence of feeding and fasting on the activation of POMC neurons by insulin, we demonstrated that insulin administered ICV repressed hepatic gluconeogenesis in fed mice, when TCPTP is repressed (Dodd et al., 2017), but not in food restricted or fasted mice, when TCPTP expression is increased (Dodd et al., 2017). Importantly we found that deleting TCPTP in POMC neurons, which enhanced IR signaling and resulted in a recruitment of POMC neurons activated by insulin, reinstated the ICV insulin-induced repression of gluconeogenesis and hepatic glucose output in otherwise unresponsive fasted mice. In keeping with this assertion, we found that chow-fed mice lacking TCPTP in POMC neurons exhibited a marked improvement in insulin sensitivity (as assessed in hyperinsulinemic euglycemic clamps) that was accounted for by a reduction in endogenous glucose production (a measure of HGP) and decreased gluconeogenesis, as assessed in pyruvate tolerance tests and the decreased expression of hepatic gluconeogenic genes. The improvement in glucose homeostasis was corrected in POMC-TC-IR mice, where the enhanced insulin-induced PI3K/AKT signaling in mice lacking TCPTP in POMC neurons was corrected to that of control mice. This would suggest that TCPTP deletion, or decreased TCPTP in POMC neurons in response to feeding, represses HGP by enhancing IR signaling and permitting POMC neurons to be activated by insulin. A caveat of these findings is that early in development Pomc-Cre is active in roughly 25% of NPY neurons (Padilla et al., 2010; Xu et al., 2018), so that the improved glucose metabolism in POMC-TC mice might at least in part reflect the enhancement of insulin signalling in AgRP/NPY neurons. However, given that (i) TCPTP abundance in POMC neurons (marked by the Pomc-eGFP reporter) correlated with insulin-induced electrophysiological responses, (ii) TCPTP inhibition in POMC neurons (marked by the Pomc-eGFP) increased the proportion of activated neurons, (iii) α-MSH staining in the PVH was increased in POMC-TC mice, and iv) DREADD-mediated activation of POMC neurons repressed HGP as seen in POMC-TC mice, our findings are consistent with feeding-associated TCPTP fluxes in POMC neurons coordinating POMC neuronal activity and thereby hepatic glucose metabolism in response to feeding.

Precisely how TCPTP influences POMC neuronal electrophysiological responses to insulin remains to be determined. One mechanism could involve TCPTP affecting synaptic plasticity by influencing pre-synaptic excitatory and inhibitory inputs, including GABAergic inputs (Cowley et al., 2001; Dietrich and Horvath, 2013; Horvath et al., 2010; Pinto et al., 2004; Sternson et al., 2005; Vong et al., 2011; Yang et al., 2011). However, our studies indicate that there is no overt change in relative excitatory and inhibitory inputs into POMC perikarya (Horvath et al., unpublished observation). Another mechanism could involve IR-dependent pathways that directly influence the activation of TRPC channels that depolarise POMC neurons (Qiu et al., 2010), or inwardly rectifying K⁺ channels that inhibit POMC neurons (Plum et al., 2006). Yet another mechanism could involve TCPTP altering mitochondrial dynamics, as mitochondrial fusion in AgRP neurons drives ATP production and excitation (Dietrich et al., 2013; Nasrallah and Horvath, 2014), whereas mitochondrial fusion in POMC neurons drives α-MSH production and the melanocortin response (Schneeberger et al., 2013). In cardiomyocytes, insulin-induced PI3K signaling promotes mitochondrial fusion (Parra et al., 2014), but it remains to be established if insulin similarly alters mitochondrial dynamics in POMC neurons and whether TCPTP influences POMC neuron electrophysiologial responses through the IR-mediated control of mitochondrial fusion. Irrespective, our studies define an exquisite mechanism by which to link feeding and fasting and the associated changes in insulin levels and glucocorticoids that drive TCPTP expression with the control of POMC neuronal excitability and the coordination of hepatic glucose metabolism for the maintenance of euglycemia. Our studies point towards this glucoregulatory mechanism being defective in obesity. The increased hypothalamic TCPTP in obesity probably occurs as a consequence of the heightened glucocorticoid and leptin levels that we have shown previously to drive TCPTP expression in the ARC (Dodd et al., 2017; Loh et al., 2011). Irrespective, our studies point towards the heightened hypothalamic TCPTP levels in obesity resulting in the majority of POMC neurons being inhibited by insulin, to contribute to the increased HGP and fasting hyperglycemia that accompany the obese state. In humans, peripheral insulin resistance in obesity has been linked to CNS insulin resistance (Anthony et al., 2006; Hirvonen et al., 2011; Tschritter et al., 2006). Moreover the ability of intranasally administered insulin to elicit effects in the hypothalamus and suppress HGP and promote glucose uptake in lean individuals is defective in overweight individuals (Heni et al., 2017; Heni et al., 2014). Therefore, we propose that perturbations in glucose metabolism in obesity, may at least in part be due to TCPTP repressing the IR-dependent activation of POMC neurons and the overall melanocortin output response.

All data generated or analysed during this study are included in the manuscript and supporting files.

1. 1. Alexander GM
   2. Rogan SC
   3. Abbas AI
   4. Armbruster BN
   5. Pei Y
   6. Allen JA
   7. Nonneman RJ
   8. Hartmann J
   9. Moy SS
   10. Nicolelis MA
   11. McNamara JO
   12. Roth BL

   (2009) Remote control of neuronal activity in transgenic mice expressing evolved G protein-coupled receptors

   Neuron 63:27–39.

   https://doi.org/10.1016/j.neuron.2009.06.014

   • PubMed
   • Google Scholar
2. 1. Anthony K
   2. Reed LJ
   3. Dunn JT
   4. Bingham E
   5. Hopkins D
   6. Marsden PK
   7. Amiel SA

   (2006) Attenuation of insulin-evoked responses in brain networks controlling appetite and reward in insulin resistance: the cerebral basis for impaired control of food intake in metabolic syndrome?

   Diabetes 55:2986–2992.

   https://doi.org/10.2337/db06-0376

   • PubMed
   • Google Scholar
3. 1. Armbruster BN
   2. Li X
   3. Pausch MH
   4. Herlitze S
   5. Roth BL

   (2007) Evolving the lock to fit the key to create a family of G protein-coupled receptors potently activated by an inert ligand

   PNAS 104:5163–5168.

   https://doi.org/10.1073/pnas.0700293104

   • PubMed
   • Google Scholar
4. 1. Atasoy D
   2. Betley JN
   3. Su HH
   4. Sternson SM

   (2012) Deconstruction of a neural circuit for hunger

   Nature 488:172–177.

   https://doi.org/10.1038/nature11270

   • PubMed
   • Google Scholar
5. 1. Balthasar N
   2. Coppari R
   3. McMinn J
   4. Liu SM
   5. Lee CE
   6. Tang V
   7. Kenny CD
   8. McGovern RA
   9. Chua SC
   10. Elmquist JK
   11. Lowell BB

   (2004) Leptin receptor signaling in POMC neurons is required for normal body weight homeostasis

   Neuron 42:983–991.

   https://doi.org/10.1016/j.neuron.2004.06.004

   • PubMed
   • Google Scholar
6. 1. Berglund ED
   2. Vianna CR
   3. Donato J
   4. Kim MH
   5. Chuang JC
   6. Lee CE
   7. Lauzon DA
   8. Lin P
   9. Brule LJ
   10. Scott MM
   11. Coppari R
   12. Elmquist JK

   (2012) Direct leptin action on POMC neurons regulates glucose homeostasis and hepatic insulin sensitivity in mice

   Journal of Clinical Investigation 122:1000–1009.

   https://doi.org/10.1172/JCI59816

   • PubMed
   • Google Scholar
7. 1. Brüning JC
   2. Michael MD
   3. Winnay JN
   4. Hayashi T
   5. Hörsch D
   6. Accili D
   7. Goodyear LJ
   8. Kahn CR

   (1998) A muscle-specific insulin receptor knockout exhibits features of the metabolic syndrome of NIDDM without altering glucose tolerance

   Molecular Cell 2:559–569.

   https://doi.org/10.1016/S1097-2765(00)80155-0

   • PubMed
   • Google Scholar
8. 1. Brüning JC
   2. Gautam D
   3. Burks DJ
   4. Gillette J
   5. Schubert M
   6. Orban PC
   7. Klein R
   8. Krone W
   9. Müller-Wieland D
   10. Kahn CR

   (2000) Role of brain insulin receptor in control of body weight and reproduction

   Science 289:2122–2125.

   https://doi.org/10.1126/science.289.5487.2122

   • PubMed
   • Google Scholar
9. 1. Buettner C
   2. Muse ED
   3. Cheng A
   4. Chen L
   5. Scherer T
   6. Pocai A
   7. Su K
   8. Cheng B
   9. Li X
   10. Harvey-White J
   11. Schwartz GJ
   12. Kunos G
   13. Rossetti L
   14. Buettner C

   (2008) Leptin controls adipose tissue lipogenesis via central, STAT3-independent mechanisms

   Nature Medicine 14:667–675.

   https://doi.org/10.1038/nm1775

   • PubMed
   • Google Scholar
10. 1. Campbell JN
    2. Macosko EZ
    3. Fenselau H
    4. Pers TH
    5. Lyubetskaya A
    6. Tenen D
    7. Goldman M
    8. Verstegen AM
    9. Resch JM
    10. McCarroll SA
    11. Rosen ED
    12. Lowell BB
    13. Tsai LT

    (2017) A molecular census of arcuate hypothalamus and median eminence cell types

    Nature Neuroscience 20:484–496.

    https://doi.org/10.1038/nn.4495

    • PubMed
    • Google Scholar
11. 1. Caron A
    2. Dungan Lemko HM
    3. Castorena CM
    4. Fujikawa T
    5. Lee S
    6. Lord CC
    7. Ahmed N
    8. Lee CE
    9. Holland WL
    10. Liu C
    11. Elmquist JK

    (2018) POMC neurons expressing leptin receptors coordinate metabolic responses to fasting via suppression of leptin levels

    eLife 7:e33710.

    https://doi.org/10.7554/eLife.33710

    • PubMed
    • Google Scholar
12. 1. Chen R
    2. Wu X
    3. Jiang L
    4. Zhang Y

    (2017) Single-Cell RNA-Seq reveals hypothalamic cell diversity

    Cell Reports 18:3227–3241.

    https://doi.org/10.1016/j.celrep.2017.03.004

    • PubMed
    • Google Scholar
13. 1. Claret M
    2. Smith MA
    3. Batterham RL
    4. Selman C
    5. Choudhury AI
    6. Fryer LG
    7. Clements M
    8. Al-Qassab H
    9. Heffron H
    10. Xu AW
    11. Speakman JR
    12. Barsh GS
    13. Viollet B
    14. Vaulont S
    15. Ashford ML
    16. Carling D
    17. Withers DJ

    (2007) AMPK is essential for energy homeostasis regulation and glucose sensing by POMC and AgRP neurons

    Journal of Clinical Investigation 117:2325–2336.

    https://doi.org/10.1172/JCI31516

    • PubMed
    • Google Scholar
14. 1. Cowley MA
    2. Smart JL
    3. Rubinstein M
    4. Cerdán MG
    5. Diano S
    6. Horvath TL
    7. Cone RD
    8. Low MJ

    (2001) Leptin activates anorexigenic POMC neurons through a neural network in the arcuate nucleus

    Nature 411:480–484.

    https://doi.org/10.1038/35078085

    • PubMed
    • Google Scholar
15. 1. Cowley MA
    2. Smith RG
    3. Diano S
    4. Tschöp M
    5. Pronchuk N
    6. Grove KL
    7. Strasburger CJ
    8. Bidlingmaier M
    9. Esterman M
    10. Heiman ML
    11. Garcia-Segura LM
    12. Nillni EA
    13. Mendez P
    14. Low MJ
    15. Sotonyi P
    16. Friedman JM
    17. Liu H
    18. Pinto S
    19. Colmers WF
    20. Cone RD
    21. Horvath TL

    (2003) The distribution and mechanism of action of ghrelin in the CNS demonstrates a novel hypothalamic circuit regulating energy homeostasis

    Neuron 37:649–661.

    https://doi.org/10.1016/S0896-6273(03)00063-1

    • PubMed
    • Google Scholar
16. 1. Dietrich MO
    2. Horvath TL

    (2013) Hypothalamic control of energy balance: insights into the role of synaptic plasticity

    Trends in Neurosciences 36:65–73.

    https://doi.org/10.1016/j.tins.2012.12.005

    • PubMed
    • Google Scholar
17. 1. Dietrich MO
    2. Liu ZW
    3. Horvath TL

    (2013) Mitochondrial dynamics controlled by mitofusins regulate Agrp neuronal activity and diet-induced obesity

    Cell 155:188–199.

    https://doi.org/10.1016/j.cell.2013.09.004

    • PubMed
    • Google Scholar
18. 1. Dodd GT
    2. Decherf S
    3. Loh K
    4. Simonds SE
    5. Wiede F
    6. Balland E
    7. Merry TL
    8. Münzberg H
    9. Zhang ZY
    10. Kahn BB
    11. Neel BG
    12. Bence KK
    13. Andrews ZB
    14. Cowley MA
    15. Tiganis T

    (2015) Leptin and insulin act on POMC neurons to promote the browning of white fat

    Cell 160:88–104.

    https://doi.org/10.1016/j.cell.2014.12.022

    • PubMed
    • Google Scholar
19. 1. Dodd GT
    2. Andrews ZB
    3. Simonds SE
    4. Michael NJ
    5. DeVeer M
    6. Brüning JC
    7. Spanswick D
    8. Cowley MA
    9. Tiganis T

    (2017) A hypothalamic phosphatase switch coordinates energy expenditure with feeding

    Cell Metabolism 26:577.

    https://doi.org/10.1016/j.cmet.2017.08.001

    • PubMed
    • Google Scholar
20. 1. Dodd GT
    2. Lee-Young RS
    3. Brüning JC
    4. Tiganis T

    (2018) TCPTP regulates insulin signaling in AgRP neurons to coordinate glucose metabolism with feeding

    Diabetes 67:1246–1257.

    https://doi.org/10.2337/db17-1485

    • PubMed
    • Google Scholar
21. 1. Elias CF
    2. Aschkenasi C
    3. Lee C
    4. Kelly J
    5. Ahima RS
    6. Bjorbaek C
    7. Flier JS
    8. Saper CB
    9. Elmquist JK

    (1999) Leptin differentially regulates NPY and POMC neurons projecting to the lateral hypothalamic area

    Neuron 23:775–786.

    https://doi.org/10.1016/S0896-6273(01)80035-0

    • PubMed
    • Google Scholar
22. 1. Ferguson SM
    2. Eskenazi D
    3. Ishikawa M
    4. Wanat MJ
    5. Phillips PE
    6. Dong Y
    7. Roth BL
    8. Neumaier JF

    (2011) Transient neuronal inhibition reveals opposing roles of indirect and direct pathways in sensitization

    Nature Neuroscience 14:22–24.

    https://doi.org/10.1038/nn.2703

    • PubMed
    • Google Scholar
23. 1. Fueger PT
    2. Lee-Young RS
    3. Shearer J
    4. Bracy DP
    5. Heikkinen S
    6. Laakso M
    7. Rottman JN
    8. Wasserman DH

    (2007) Phosphorylation barriers to skeletal and cardiac muscle glucose uptakes in high-fat fed mice: studies in mice with a 50% reduction of hexokinase II

    Diabetes 56:2476–2484.

    https://doi.org/10.2337/db07-0532

    • PubMed
    • Google Scholar
24. 1. Gelling RW
    2. Morton GJ
    3. Morrison CD
    4. Niswender KD
    5. Myers MG
    6. Rhodes CJ
    7. Schwartz MW

    (2006) Insulin action in the brain contributes to glucose lowering during insulin treatment of diabetes

    Cell Metabolism 3:67–73.

    https://doi.org/10.1016/j.cmet.2005.11.013

    • PubMed
    • Google Scholar
25. 1. Havrankova J
    2. Roth J
    3. Brownstein M

    (1978) Insulin receptors are widely distributed in the central nervous system of the rat

    Nature 272:827–829.

    https://doi.org/10.1038/272827a0

    • PubMed
    • Google Scholar
26. 1. Heni M
    2. Wagner R
    3. Kullmann S
    4. Veit R
    5. Mat Husin H
    6. Linder K
    7. Benkendorff C
    8. Peter A
    9. Stefan N
    10. Häring HU
    11. Preissl H
    12. Fritsche A

    (2014) Central insulin administration improves whole-body insulin sensitivity via hypothalamus and parasympathetic outputs in men

    Diabetes 63:4083–4088.

    https://doi.org/10.2337/db14-0477

    • PubMed
    • Google Scholar
27. 1. Heni M
    2. Wagner R
    3. Kullmann S
    4. Gancheva S
    5. Roden M
    6. Peter A
    7. Stefan N
    8. Preissl H
    9. Häring HU
    10. Fritsche A

    (2017) Hypothalamic and striatal insulin action suppresses endogenous glucose production and may stimulate glucose uptake during hyperinsulinemia in lean but not in overweight men

    Diabetes 66:1797–1806.

    https://doi.org/10.2337/db16-1380

    • PubMed
    • Google Scholar
28. 1. Henry FE
    2. Sugino K
    3. Tozer A
    4. Branco T
    5. Sternson SM

    (2015) Cell type-specific transcriptomics of hypothalamic energy-sensing neuron responses to weight-loss

    eLife 4:e09800.

    https://doi.org/10.7554/eLife.09800

    • Google Scholar
29. 1. Hill JW
    2. Williams KW
    3. Ye C
    4. Luo J
    5. Balthasar N
    6. Coppari R
    7. Cowley MA
    8. Cantley LC
    9. Lowell BB
    10. Elmquist JK

    (2008) Acute effects of leptin require PI3K signaling in hypothalamic proopiomelanocortin neurons in mice

    Journal of Clinical Investigation 118:1796–1805.

    https://doi.org/10.1172/JCI32964

    • PubMed
    • Google Scholar
30. 1. Hill JW
    2. Elias CF
    3. Fukuda M
    4. Williams KW
    5. Berglund ED
    6. Holland WL
    7. Cho YR
    8. Chuang JC
    9. Xu Y
    10. Choi M
    11. Lauzon D
    12. Lee CE
    13. Coppari R
    14. Richardson JA
    15. Zigman JM
    16. Chua S
    17. Scherer PE
    18. Lowell BB
    19. Brüning JC
    20. Elmquist JK

    (2010) Direct insulin and leptin action on pro-opiomelanocortin neurons is required for normal glucose homeostasis and fertility

    Cell Metabolism 11:286–297.

    https://doi.org/10.1016/j.cmet.2010.03.002

    • PubMed
    • Google Scholar
31. 1. Hirvonen J
    2. Virtanen KA
    3. Nummenmaa L
    4. Hannukainen JC
    5. Honka MJ
    6. Bucci M
    7. Nesterov SV
    8. Parkkola R
    9. Rinne J
    10. Iozzo P
    11. Nuutila P

    (2011) Effects of insulin on brain glucose metabolism in impaired glucose tolerance

    Diabetes 60:443–447.

    https://doi.org/10.2337/db10-0940

    • PubMed
    • Google Scholar
32. 1. Horvath TL
    2. Sarman B
    3. García-Cáceres C
    4. Enriori PJ
    5. Sotonyi P
    6. Shanabrough M
    7. Borok E
    8. Argente J
    9. Chowen JA
    10. Perez-Tilve D
    11. Pfluger PT
    12. Brönneke HS
    13. Levin BE
    14. Diano S
    15. Cowley MA
    16. Tschöp MH

    (2010) Synaptic input organization of the melanocortin system predicts diet-induced hypothalamic reactive gliosis and obesity

    PNAS 107:14875–14880.

    https://doi.org/10.1073/pnas.1004282107

    • PubMed
    • Google Scholar
33. 1. Inoue H
    2. Ogawa W
    3. Ozaki M
    4. Haga S
    5. Matsumoto M
    6. Furukawa K
    7. Hashimoto N
    8. Kido Y
    9. Mori T
    10. Sakaue H
    11. Teshigawara K
    12. Jin S
    13. Iguchi H
    14. Hiramatsu R
    15. LeRoith D
    16. Takeda K
    17. Akira S
    18. Kasuga M

    (2004) Role of STAT-3 in regulation of hepatic gluconeogenic genes and carbohydrate metabolism in vivo

    Nature Medicine 10:168–174.

    https://doi.org/10.1038/nm980

    • PubMed
    • Google Scholar
34. 1. Inoue H
    2. Ogawa W
    3. Asakawa A
    4. Okamoto Y
    5. Nishizawa A
    6. Matsumoto M
    7. Teshigawara K
    8. Matsuki Y
    9. Watanabe E
    10. Hiramatsu R
    11. Notohara K
    12. Katayose K
    13. Okamura H
    14. Kahn CR
    15. Noda T
    16. Takeda K
    17. Akira S
    18. Inui A
    19. Kasuga M

    (2006) Role of hepatic STAT3 in brain-insulin action on hepatic glucose production

    Cell Metabolism 3:267–275.

    https://doi.org/10.1016/j.cmet.2006.02.009

    • PubMed
    • Google Scholar
35. 1. Kimura K
    2. Tanida M
    3. Nagata N
    4. Inaba Y
    5. Watanabe H
    6. Nagashimada M
    7. Ota T
    8. Asahara S
    9. Kido Y
    10. Matsumoto M
    11. Toshinai K
    12. Nakazato M
    13. Shibamoto T
    14. Kaneko S
    15. Kasuga M
    16. Inoue H

    (2016) Central insulin action activates kupffer cells by suppressing hepatic vagal activation via the nicotinic alpha 7 acetylcholine receptor

    Cell Reports 14:2362–2374.

    https://doi.org/10.1016/j.celrep.2016.02.032

    • PubMed
    • Google Scholar
36. 1. Könner AC
    2. Janoschek R
    3. Plum L
    4. Jordan SD
    5. Rother E
    6. Ma X
    7. Xu C
    8. Enriori P
    9. Hampel B
    10. Barsh GS
    11. Kahn CR
    12. Cowley MA
    13. Ashcroft FM
    14. Brüning JC

    (2007) Insulin action in AgRP-expressing neurons is required for suppression of hepatic glucose production

    Cell Metabolism 5:438–449.

    https://doi.org/10.1016/j.cmet.2007.05.004

    • PubMed
    • Google Scholar
37. 1. Krashes MJ
    2. Koda S
    3. Ye C
    4. Rogan SC
    5. Adams AC
    6. Cusher DS
    7. Maratos-Flier E
    8. Roth BL
    9. Lowell BB

    (2011) Rapid, reversible activation of AgRP neurons drives feeding behavior in mice

    Journal of Clinical Investigation 121:1424–1428.

    https://doi.org/10.1172/JCI46229

    • PubMed
    • Google Scholar
38. 1. Lam TK
    2. Pocai A
    3. Gutierrez-Juarez R
    4. Obici S
    5. Bryan J
    6. Aguilar-Bryan L
    7. Schwartz GJ
    8. Rossetti L

    (2005) Hypothalamic sensing of circulating fatty acids is required for glucose homeostasis

    Nature Medicine 11:320–327.

    https://doi.org/10.1038/nm1201

    • PubMed
    • Google Scholar
39. 1. Lam BYH
    2. Cimino I
    3. Polex-Wolf J
    4. Nicole Kohnke S
    5. Rimmington D
    6. Iyemere V
    7. Heeley N
    8. Cossetti C
    9. Schulte R
    10. Saraiva LR
    11. Logan DW
    12. Blouet C
    13. O'Rahilly S
    14. Coll AP
    15. Yeo GSH

    (2017) Heterogeneity of hypothalamic pro-opiomelanocortin-expressing neurons revealed by single-cell RNA sequencing

    Molecular Metabolism 6:383–392.

    https://doi.org/10.1016/j.molmet.2017.02.007

    • PubMed
    • Google Scholar
40. 1. Lamont BJ
    2. Visinoni S
    3. Fam BC
    4. Kebede M
    5. Weinrich B
    6. Papapostolou S
    7. Massinet H
    8. Proietto J
    9. Favaloro J
    10. Andrikopoulos S

    (2006) Expression of human fructose-1,6-bisphosphatase in the liver of transgenic mice results in increased glycerol gluconeogenesis

    Endocrinology 147:2764–2772.

    https://doi.org/10.1210/en.2005-1498

    • PubMed
    • Google Scholar
41. 1. Loh K
    2. Deng H
    3. Fukushima A
    4. Cai X
    5. Boivin B
    6. Galic S
    7. Bruce C
    8. Shields BJ
    9. Skiba B
    10. Ooms LM
    11. Stepto N
    12. Wu B
    13. Mitchell CA
    14. Tonks NK
    15. Watt MJ
    16. Febbraio MA
    17. Crack PJ
    18. Andrikopoulos S
    19. Tiganis T

    (2009) Reactive oxygen species enhance insulin sensitivity

    Cell Metabolism 10:260–272.

    https://doi.org/10.1016/j.cmet.2009.08.009

    • PubMed
    • Google Scholar
42. 1. Loh K
    2. Fukushima A
    3. Zhang X
    4. Galic S
    5. Briggs D
    6. Enriori PJ
    7. Simonds S
    8. Wiede F
    9. Reichenbach A
    10. Hauser C
    11. Sims NA
    12. Bence KK
    13. Zhang S
    14. Zhang ZY
    15. Kahn BB
    16. Neel BG
    17. Andrews ZB
    18. Cowley MA
    19. Tiganis T

    (2011) Elevated hypothalamic TCPTP in obesity contributes to cellular leptin resistance

    Cell Metabolism 14:684–699.

    https://doi.org/10.1016/j.cmet.2011.09.011

    • PubMed
    • Google Scholar
43. 1. Nasrallah CM
    2. Horvath TL

    (2014) Mitochondrial dynamics in the central regulation of metabolism

    Nature Reviews Endocrinology 10:650–658.

    https://doi.org/10.1038/nrendo.2014.160

    • PubMed
    • Google Scholar
44. 1. Novak A
    2. Guo C
    3. Yang W
    4. Nagy A
    5. Lobe CG

    (2000) Z/EG, a double reporter mouse line that expresses enhanced green fluorescent protein upon Cre-mediated excision

    Genesis 28:147–155.

    https://doi.org/10.1002/1526-968X(200011/12)28:3/4<147::AID-GENE90>3.0.CO;2-G

    • PubMed
    • Google Scholar
45. 1. Obici S
    2. Feng Z
    3. Karkanias G
    4. Baskin DG
    5. Rossetti L

    (2002a) Decreasing hypothalamic insulin receptors causes hyperphagia and insulin resistance in rats

    Nature Neuroscience 5:566–572.

    https://doi.org/10.1038/nn0602-861

    • PubMed
    • Google Scholar
46. 1. Obici S
    2. Zhang BB
    3. Karkanias G
    4. Rossetti L

    (2002b) Hypothalamic insulin signaling is required for inhibition of glucose production

    Nature Medicine 8:1376–1382.

    https://doi.org/10.1038/nm1202-798

    • PubMed
    • Google Scholar
47. 1. Padilla SL
    2. Carmody JS
    3. Zeltser LM

    (2010) Pomc-expressing progenitors give rise to antagonistic neuronal populations in hypothalamic feeding circuits

    Nature Medicine 16:403–405.

    https://doi.org/10.1038/nm.2126

    • PubMed
    • Google Scholar
48. 1. Parra V
    2. Verdejo HE
    3. Iglewski M
    4. Del Campo A
    5. Troncoso R
    6. Jones D
    7. Zhu Y
    8. Kuzmicic J
    9. Pennanen C
    10. Lopez-Crisosto C
    11. Jaña F
    12. Ferreira J
    13. Noguera E
    14. Chiong M
    15. Bernlohr DA
    16. Klip A
    17. Hill JA
    18. Rothermel BA
    19. Abel ED
    20. Zorzano A
    21. Lavandero S

    (2014) Insulin stimulates mitochondrial fusion and function in cardiomyocytes via the Akt-mTOR-NFκB-Opa-1 signaling pathway

    Diabetes 63:75–88.

    https://doi.org/10.2337/db13-0340

    • PubMed
    • Google Scholar
49. 1. Parton LE
    2. Ye CP
    3. Coppari R
    4. Enriori PJ
    5. Choi B
    6. Zhang CY
    7. Xu C
    8. Vianna CR
    9. Balthasar N
    10. Lee CE
    11. Elmquist JK
    12. Cowley MA
    13. Lowell BB

    (2007) Glucose sensing by POMC neurons regulates glucose homeostasis and is impaired in obesity

    Nature 449:228–232.

    https://doi.org/10.1038/nature06098

    • PubMed
    • Google Scholar
50. 1. Pinto S
    2. Roseberry AG
    3. Liu H
    4. Diano S
    5. Shanabrough M
    6. Cai X
    7. Friedman JM
    8. Horvath TL

    (2004) Rapid rewiring of arcuate nucleus feeding circuits by leptin

    Science 304:110–115.

    https://doi.org/10.1126/science.1089459

    • PubMed
    • Google Scholar
51. 1. Plum L
    2. Ma X
    3. Hampel B
    4. Balthasar N
    5. Coppari R
    6. Münzberg H
    7. Shanabrough M
    8. Burdakov D
    9. Rother E
    10. Janoschek R
    11. Alber J
    12. Belgardt BF
    13. Koch L
    14. Seibler J
    15. Schwenk F
    16. Fekete C
    17. Suzuki A
    18. Mak TW
    19. Krone W
    20. Horvath TL
    21. Ashcroft FM
    22. Brüning JC

    (2006) Enhanced PIP3 signaling in POMC neurons causes KATP channel activation and leads to diet-sensitive obesity

    Journal of Clinical Investigation 116:1886–1901.

    https://doi.org/10.1172/JCI27123

    • PubMed
    • Google Scholar
52. 1. Pocai A
    2. Lam TK
    3. Gutierrez-Juarez R
    4. Obici S
    5. Schwartz GJ
    6. Bryan J
    7. Aguilar-Bryan L
    8. Rossetti L

    (2005) Hypothalamic K(ATP) channels control hepatic glucose production

    Nature 434:1026–1031.

    https://doi.org/10.1038/nature03439

    • PubMed
    • Google Scholar
53. 1. Qiu J
    2. Fang Y
    3. Rønnekleiv OK
    4. Kelly MJ

    (2010) Leptin excites proopiomelanocortin neurons via activation of TRPC channels

    Journal of Neuroscience 30:1560–1565.

    https://doi.org/10.1523/JNEUROSCI.4816-09.2010

    • PubMed
    • Google Scholar
54. 1. Qiu J
    2. Zhang C
    3. Borgquist A
    4. Nestor CC
    5. Smith AW
    6. Bosch MA
    7. Ku S
    8. Wagner EJ
    9. Rønnekleiv OK
    10. Kelly MJ

    (2014) Insulin excites anorexigenic proopiomelanocortin neurons via activation of canonical transient receptor potential channels

    Cell Metabolism 19:682–693.

    https://doi.org/10.1016/j.cmet.2014.03.004

    • PubMed
    • Google Scholar
55. 1. Qiu J
    2. Wagner EJ
    3. Rønnekleiv OK
    4. Kelly MJ

    (2018) Insulin and leptin excite anorexigenic pro-opiomelanocortin neurones via activation of TRPC5 channels

    Journal of Neuroendocrinology 30:e12501.

    https://doi.org/10.1111/jne.12501

    • PubMed
    • Google Scholar
56. 1. Schneeberger M
    2. Dietrich MO
    3. Sebastián D
    4. Imbernón M
    5. Castaño C
    6. Garcia A
    7. Esteban Y
    8. Gonzalez-Franquesa A
    9. Rodríguez IC
    10. Bortolozzi A
    11. Garcia-Roves PM
    12. Gomis R
    13. Nogueiras R
    14. Horvath TL
    15. Zorzano A
    16. Claret M

    (2013) Mitofusin 2 in POMC neurons connects ER stress with leptin resistance and energy imbalance

    Cell 155:172–187.

    https://doi.org/10.1016/j.cell.2013.09.003

    • PubMed
    • Google Scholar
57. 1. Shin AC
    2. Filatova N
    3. Lindtner C
    4. Chi T
    5. Degann S
    6. Oberlin D
    7. Buettner C

    (2017) Insulin receptor signaling in POMC, but not AgRP, neurons controls adipose tissue insulin action

    Diabetes 66:1560–1571.

    https://doi.org/10.2337/db16-1238

    • PubMed
    • Google Scholar
58. 1. Spanswick D
    2. Smith MA
    3. Mirshamsi S
    4. Routh VH
    5. Ashford ML

    (2000) Insulin activates ATP-sensitive K+ channels in hypothalamic neurons of lean, but not obese rats

    Nature Neuroscience 3:757–758.

    https://doi.org/10.1038/77660

    • PubMed
    • Google Scholar
59. 1. Steculorum SM
    2. Ruud J
    3. Karakasilioti I
    4. Backes H
    5. Engström Ruud L
    6. Timper K
    7. Hess ME
    8. Tsaousidou E
    9. Mauer J
    10. Vogt MC
    11. Paeger L
    12. Bremser S
    13. Klein AC
    14. Morgan DA
    15. Frommolt P
    16. Brinkkötter PT
    17. Hammerschmidt P
    18. Benzing T
    19. Rahmouni K
    20. Wunderlich FT
    21. Kloppenburg P
    22. Brüning JC

    (2016) AgRP neurons control systemic insulin sensitivity via myostatin expression in Brown adipose tissue

    Cell 165:125–138.

    https://doi.org/10.1016/j.cell.2016.02.044

    • PubMed
    • Google Scholar
60. 1. Sternson SM
    2. Shepherd GM
    3. Friedman JM

    (2005) Topographic mapping of VMH --> arcuate nucleus microcircuits and their reorganization by fasting

    Nature Neuroscience 8:1356–1363.

    https://doi.org/10.1038/nn1550

    • PubMed
    • Google Scholar
61. 1. Tiganis T

    (2013) PTP1B and TCPTP--nonredundant phosphatases in insulin signaling and glucose homeostasis

    FEBS Journal 280:445–458.

    https://doi.org/10.1111/j.1742-4658.2012.08563.x

    • PubMed
    • Google Scholar
62. 1. Tong Q
    2. Ye CP
    3. Jones JE
    4. Elmquist JK
    5. Lowell BB

    (2008) Synaptic release of GABA by AgRP neurons is required for normal regulation of energy balance

    Nature Neuroscience 11:998–1000.

    https://doi.org/10.1038/nn.2167

    • PubMed
    • Google Scholar
63. 1. Tschritter O
    2. Preissl H
    3. Hennige AM
    4. Stumvoll M
    5. Porubska K
    6. Frost R
    7. Marx H
    8. Klösel B
    9. Lutzenberger W
    10. Birbaumer N
    11. Häring HU
    12. Fritsche A

    (2006) The cerebrocortical response to hyperinsulinemia is reduced in overweight humans: a magnetoencephalographic study

    PNAS 103:12103–12108.

    https://doi.org/10.1073/pnas.0604404103

    • PubMed
    • Google Scholar
64. 1. van den Top M
    2. Lee K
    3. Whyment AD
    4. Blanks AM
    5. Spanswick D

    (2004) Orexigen-sensitive NPY/AgRP pacemaker neurons in the hypothalamic arcuate nucleus

    Nature Neuroscience 7:493–494.

    https://doi.org/10.1038/nn1226

    • PubMed
    • Google Scholar
65. 1. Varela L
    2. Horvath TL

    (2012) Leptin and insulin pathways in POMC and AgRP neurons that modulate energy balance and glucose homeostasis

    EMBO reports 13:1079–1086.

    https://doi.org/10.1038/embor.2012.174

    • PubMed
    • Google Scholar
66. 1. Vogt MC
    2. Brüning JC

    (2013) CNS insulin signaling in the control of energy homeostasis and glucose metabolism - from embryo to old age

    Trends in Endocrinology & Metabolism 24:76–84.

    https://doi.org/10.1016/j.tem.2012.11.004

    • PubMed
    • Google Scholar
67. 1. Vogt MC
    2. Paeger L
    3. Hess S
    4. Steculorum SM
    5. Awazawa M
    6. Hampel B
    7. Neupert S
    8. Nicholls HT
    9. Mauer J
    10. Hausen AC
    11. Predel R
    12. Kloppenburg P
    13. Horvath TL
    14. Brüning JC

    (2014) Neonatal insulin action impairs hypothalamic neurocircuit formation in response to maternal high-fat feeding

    Cell 156:495–509.

    https://doi.org/10.1016/j.cell.2014.01.008

    • PubMed
    • Google Scholar
68. 1. Vong L
    2. Ye C
    3. Yang Z
    4. Choi B
    5. Chua S
    6. Lowell BB

    (2011) Leptin action on GABAergic neurons prevents obesity and reduces inhibitory tone to POMC neurons

    Neuron 71:142–154.

    https://doi.org/10.1016/j.neuron.2011.05.028

    • PubMed
    • Google Scholar
69. 1. Williams KW
    2. Margatho LO
    3. Lee CE
    4. Choi M
    5. Lee S
    6. Scott MM
    7. Elias CF
    8. Elmquist JK

    (2010) Segregation of acute leptin and insulin effects in distinct populations of arcuate proopiomelanocortin neurons

    Journal of Neuroscience 30:2472–2479.

    https://doi.org/10.1523/JNEUROSCI.3118-09.2010

    • PubMed
    • Google Scholar
70. 1. Xu J
    2. Bartolome CL
    3. Low CS
    4. Yi X
    5. Chien CH
    6. Wang P
    7. Kong D

    (2018) Genetic identification of leptin neural circuits in energy and glucose homeostases

    Nature 556:505–509.

    https://doi.org/10.1038/s41586-018-0049-7

    • PubMed
    • Google Scholar
71. 1. Yang Y
    2. Atasoy D
    3. Su HH
    4. Sternson SM

    (2011) Hunger states switch a flip-flop memory circuit via a synaptic AMPK-dependent positive feedback loop

    Cell 146:992–1003.

    https://doi.org/10.1016/j.cell.2011.07.039

    • PubMed
    • Google Scholar
72. 1. Zhang S
    2. Chen L
    3. Luo Y
    4. Gunawan A
    5. Lawrence DS
    6. Zhang ZY

    (2009) Acquisition of a potent and selective TC-PTP inhibitor via a stepwise fluorophore-tagged combinatorial synthesis and screening strategy

    Journal of the American Chemical Society 131:13072–13079.

    https://doi.org/10.1021/ja903733z

    • PubMed
    • Google Scholar
73. 1. Zhang ZY
    2. Dodd GT
    3. Tiganis T

    (2015) Protein tyrosine phosphatases in hypothalamic insulin and leptin signaling

    Trends in Pharmacological Sciences 36:661–674.

    https://doi.org/10.1016/j.tips.2015.07.003

    • PubMed
    • Google Scholar

Thank you for submitting your article "Insulin regulates POMC neuronal plasticity to control glucose metabolism" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Mark McCarthy as the Senior Editor. The following individual involved in review of your submission has agreed to reveal his identity: Dong Kong (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

This interesting manuscript presents a large amount of information regarding the role of a specific protein tyrosine phosphatase, TCPTP, in the control of insulin responsiveness in POMC neurons in relation to control of systemic insulin sensitivity and hepatic glucose metabolism. The data strongly suggest both a key role for TCPTP in the responsiveness of POMC cells to insulin and that enhanced insulin signaling depolarizes and increases the firing of in these neurons, while the opposite effect is observed in POMC neurons with high levels of TCPTP. The data also offer a welcome new perspective on the role of POMC neurons in the control of HGP and a potential resolution to controversy surrounding this issue stemming from earlier work, and the impact of changes in nutritional state on POMC cell insulin responsiveness is also novel and of interest.

The reviewers of this paper found the results to be of great interest and suitable for publication in eLife after some revisions. It was not possible to assemble the comments into a few key points for consideration (eLife normal procedure) because the three reviewers had non-overlapping concerns. Thus, all of the comments are included below. The reviewers had a number of concerns that can, in general, be addressed without further experiments. However, explanations of the many points raised are necessary and caveats to the data interpretation should be included. Additional experiments that address concerns are, of course, welcome.

One issue raised by reviewers is that the use of Pomc-Cre mice is problematic because Cre recombinase is activated in neurons that go to develop into AgRP (and other) neurons. Thus, some of the data are probably derived from a mixed population of arcuate neurons.

The other issue is that the mitochondria data are correlative and do not provide compelling evidence linking mitochondrial dynamics and glucose regulation. Some reviewers, including the editor, suggest either providing a causal link between mitochondrial dynamics and glucose regulation, or removing the mitochondrial experiments from this story.

The authors should define TCPTP on first usage and note the mouse gene names do not include hyphens, e.g. Il6, not Il-6.

Reviewer #1:

In this study, Dodd et al. investigate the role of TCPTP in POMC neurons upon insulin signaling and glucose metabolism control. The authors report a correlation between TCPTP expression in POMC neurons and the level of neuron responsiveness to insulin. High expression of TCPTP in obesity or fasting hampers insulin-induced signaling, POMC neuron activation and repression of HGP. In contrast, genetic inactivation of TCPTP in POMC neurons causes enhanced insulin-stimulated signaling, activation of POMC neurons and inhibition of HGP. The authors also report changes in mitochondrial dynamics and mitochondria-ER contacts in POMC neurons upon TCPTP deletion. The authors propose that physiological (fed/fast) and pathophysiological (HFD) changes in TCPTP expression in POMC neurons may influence neuronal activity and HGP via insulin signaling-mediated modulation of mitochondrial fusion and contacts with the ER.

The study uses a nice combination of mouse genetics, electrophysiology, pharmacology as well as physiological and molecular biology techniques. Generally speaking, the conclusions are supported by the data although some statements in specific parts of the manuscript should be softened. Overall, the study fulfils the highest standards in the field and seems appropriate for publication in eLife.

I do not have major concerns that require additional experiments but I would suggest further analysis of the electrophysiology results. In particular, the authors compare the number (population size) of POMC neurons that are activated, inhibited or unresponsive to insulin amongst a number of experimental settings. Table 1 provides statistical comparisons for some of these studies, but not all (only for KO's and double Ko's). However, the authors claim that pharmacological TCPTP inhibition "significantly altered POMC responses to insulin" but statistical evidence is not provided. The same for the HFD study. The authors should provide a complete table of comparisons amongst the diverse experimental settings.

Reviewer #2:

Insulin's effect on POMC neurons, their neuronal heterogeneity, and the related blood glucose regulation have been interesting topics in the field of brain-controlled metabolism for quite a while. However, the underlying mechanisms are still unclear. Whether POMC neurons regulate glucose homeostasis is also under debate. In the current study, Dodd et al. focused on these questions, performed a series of studies, and made multiple important and exciting findings. First, they documented heterogenous responses of POMC neurons to insulin and further demonstrated a dependence of such responses on the expression of TCPTP phosphatase in these neurons. Then, the author performed glucose clamp studies by combining with chemogenetic tools and assessed the peripheral mechanisms. They found that TCPTP in POMC neurons regulated blood glucose by mainly inhibiting hepatic glucose production. Finally, they also used multiple mouse models and discovered that feeding/fasting and HFD-feeding induced TCPTP expression in POMC neurons regulated insulin receptor signaling and mitochondrial dynamics to manipulate responses to insulin and peripheral glucose homeostasis.

Overall, this manuscript is very well prepared, the studies were properly designed, performed, and analyzed, and the conclusions were well reached. The current study has thus established an elegant system to understand the central control of glucose metabolism. I believe the manuscript will be of great interest to the readers of eLife and I have the following comments for the authors to address.

1) The authors performed a large number of current-clamp recordings on POMC neurons to assess their responses to insulin. However, the results of membrane potential seem problematic in the current study. For example, in the first paragraph of the subsection “TCPTP defines POMC neurons that are activated or inhibited by insulin”, POMC neurons after insulin stimulation exhibited an averaged membrane potential of -40.1+/-2.5 mV; in the next paragraph of the aforementioned subsection, excited neurons had a membrane potential as depolarized as -37.9+/-1.7 mV. How could POMC neurons be so depolarized?

2) As shown previously by Padilla, Carmody and Zeltser, et al. (Nat Med 2010, and in a more convincing way recently by Xu et al., 2018, POMC-Cre exhibited ectopic Cre activity during development, particularly in AgRP neuron. Breeding Pomc-Cre with floxed mice could yield a manipulation of 25% AgRP neurons as well (Xu et al.). In Figure 1G-H, since the authors used Z/EG reporter instead of Pomc-eGFP transgenic mice to indicate POMC neurons, AgRP neurons were expected to be patched on as well. Since AgRP neurons are known to be inhibited by insulin, the ratio for the insulin-inhibited POMC neurons (12/33, 36.4%) may need to be adjusted.

3) Subsection “The activation or inhibition of POMC neurons regulates hepatic glucose metabolism”, the authors utilized inhibitory DREADDs to inhibit POMC neurons and included convincing expression data of hM4Di in Figure 2—figure supplement 1. However, a direct test of CNO's activity on POMC neurons with current-clamp recording shall also be included since hM4Di works through the activation of GIRK channels and not all neurons express them. In addition, CNO was recently reported to activate other endogenous receptors as well (Gomez et al., Science 2017). CNO's effects on non-DREADD-expressing animals shall also be included as a control.

4) In the third paragraph of the Discussion, the authors discussed about the potential mechanisms with regards to synaptic regulation. Deletion of TCPTP in POMC neurons clearly suggested a post-synaptic mechanism but the authors believed in an existence of pre-synaptic regulation. This is confusing. Do the authors think of the involvement of a retrograde mechanism or different action sites of TCPTP from POMC neurons?

Reviewer #3:

This interesting manuscript presents a large amount of information regarding the role of a specific protein tyrosine phosphatase, TCPTP, in the control of insulin responsiveness in POMC neurons in relation to control of systemic insulin sensitivity and hepatic glucose metabolism. The data strongly suggest both a key role for TCPTP in the responsiveness of POMC cells to insulin and that enhanced insulin signaling depolarizes and increases the firing of in these neurons, while the opposite effect is observed in POMC neurons with high levels of TCPTP. The data also offer a welcome new perspective on the role of POMC neurons in the control of HGP and a potential resolution to controversy surrounding this issue stemming from earlier work, and the impact of changes in nutritional state on POMC cell insulin responsiveness is also novel and of interest. These considerable strengths are offset to some extent by the following concerns.

1) A key unanswered question is how increased insulin receptor signal transduction associated with reduced TCPTP is tied to a reversal of the effect of insulin on POMC neuron membrane potential and firing rate (from inhibitory to stimulatory). The authors state that this "phenotype switch" in the insulin response reflects direct (rather than an indirect) effects of insulin, but how increased IR-PI3K-Akt signal transduction in these cells might achieve this effect is unresolved. In an effort to shed light on this key question, the authors investigate whether changes in POMC cell mitochondrial dynamics underlie this phenotype switch. Although changes of insulin responsiveness appear to be coupled to morphological and genomic alterations affecting mitochondria, the data fail to establish a connection between these mitochondrial responses and changes of membrane potential and firing rate. Consequently, the data on mitochondrial dynamics seem somewhat disconnected from the rest of the paper, and might be more appropriate for a separate publication.

2) The authors finding of a robust effect of POMC-specific TCPTP deletion to enhance insulin suppression of HGP begs the question as to what other responses to insulin might also be affected. Given that ICV insulin has previously been reported to reduce food intake, body weight and body adiposity, and that brain-specific IR deletion promotes positive energy balance and weight gain, one wonders whether the effect of ICV insulin on food intake, body weight and body adiposity are also increased. Given the marked improvement of whole-body insulin sensitivity attributed to POMC-specific TCPTP deletion, one also wonders why the basal glucose and insulin levels were not reduced in these animals.

3) A related question is whether reduced body adiposity results from POMC-specific TCPTP deletion, and if so, did this contribute to enhanced insulin-induced suppression of HGP reported by the authors? Conversely, if POMC-specific TCPTP deletion has no effect on energy balance or body adiposity, this seems surprising given the expected increase of POMC neuron firing.

4) The question of how pronounced changes of POMC expression of TCPTP are associated with feeding, fasting and DIO was not addressed. Doing so would considerably strengthen the manuscript and seems of much greater relevance than the large amount of mitochondrial data that were included.

5) During embryogenesis, the POMC promoter is expressed in multiple arcuate nucleus neuronal subtypes including AgRP neurons. Since POMC-Cre was used to delete TCPTP, one expects the gene to be deleted across multiple arcuate nucleus neuronal cell types. Some effort to address this concern is warranted.

6) The authors state that diet-induced obesity reduces POMC neuron insulin responsiveness via a mechanism involving induction of TCPTP. Is this effect of DIO causally linked to associated responses such as inflammation, reactive gliosis, ER stress, etc.? Discussion of this question seems warranted.

7) The authors report (Discussion, last paragraph) that TCPTP deletion has no detectable effect on glucose homeostasis in DIO mice, which they attribute to the fact that other mechanisms can account for impaired POMC insulin responsiveness in this setting. But they stop short of investigating whether TCPTP deletion enhances insulin responsiveness to POMC neurons in this setting. This question should be addressed since, if POMC neuron insulin responsiveness is in fact restored and yet there is no whole-body phenotype, the physiological relevance of POMC neuron insulin responsiveness would be called into question.

8) What precedent exists for a change in hormone responsiveness reversing its effect on the firing of a distinct neuronal subset? If such a response has been reported previously, was it referred to as "neural plasticity"? Certainly, this is not the type of response that comes to mind in association with that term. In this context, the term "neural plasticity" seems potentially misleading and for this reason, it is recommended that it be removed from the manuscript title.

9) The effect of IP insulin to induce c-Fos in the PVH (Figure 1M) could have resulted from hypoglycemia, which was presumably well established by 90 min following the dose of insulin that was given. The effect of neuroglucopenia to activate PVH neurons is well established (e.g., Briski, Neuroreport, 1998; Briski, Brain Res Bulletin, 2000; Evans, AJP, 2001), and it is conceivable that changes of POMC cell insulin responsiveness affected the degree of insulin-induced hypoglycemia.

10) When insulin is given ICV and HGP or other highly insulin-responsive endpoints are measured, it's important to verify that none of the centrally-administered insulin leaked into the periphery to have direct effects on the liver.

11) Relevant to this point is the use of a study design in which ICV insulin (0.1 mU) or vehicle was given as 5 separate injections over 5 hours. It is difficult to imagine that 5 ICV injections over 5 hours would be well-tolerated, which in turn raises the possibility that the authors were inadvertently measuring the impact of insulin on a stress response, which might explain the basal hyperglycemia reported in some studies. Beyond this, insulin in CSF has a relatively long half-life, so it's not clear why a repeated injection protocol over 5 hours was required. Were the animals anesthetized during this period? I could not find the information in the Materials and methods section.

12) The basal glucose level (time=0 min) of both groups in the pyruvate tolerance test shown in Figure 2C is ~12 mM, which is in the diabetic range. By comparison, basal glucose levels of 7 mM are reported for the same test in Figure 2J, and are uniformly <10 mM in all other such studies (Figures 4, 5 and 6). This discrepancy is all the more notable in light of the remarkably small variances (SEMs) around the glucose data, which is inconsistent with known biological variability inherent in the response.

13) The Discussion is longer than it need be, and tends to reiterate points made in the Results section. It could be cut substantially.

[…] The reviewers of this paper found the results to be of great interest and suitable for publication in eLife after some revisions. It was not possible to assemble the comments into a few key points for consideration (eLife normal procedure) because the three reviewers had non-overlapping concerns. Thus, all of the comments are included below. The reviewers had a number of concerns that can, in general, be addressed without further experiments. However, explanations of the many points raised are necessary and caveats to the data interpretation should be included. Additional experiments that address concerns are, of course, welcome.

One issue raised by reviewers is that the use of Pomc-Cre mice is problematic because Cre recombinase is activated in neurons that go to develop into AgRP (and other) neurons. Thus, some of the data are probably derived from a mixed population of arcuate neurons.

We have now discussed this caveat in relation to the glucose metabolism studies in our POMC TCPTP knockout mice in the revised Discussion. Importantly, we make the specific point in the results that Pomc-eGFP faithfully marks POMC neurons (Padilla et al., 2010) and that Pomc-Cre in adult mice deletes exclusively in POMC neurons (Xu et al., 2018). Accordingly our studies defining the existence of POMC neurons that are activated or inhibited by insulin, the increased proportion of activated POMC neurons after TCPTP inhibition, the decreased proportion of POMC neurons in obesity that is corrected by TCPTP inhibition and the capacity of activating and inhibitory DREADDs to alter glucose metabolism are not compromised by the off target effects of the Pomc-Cre transgene. Collectively, our studies are entirely consistent with alterations in insulin-induced POMC neuronal excitability regulating hepatic glucose metabolism, but we have acknowledged that some of the effects on glucose metabolism in POMC-TC mice might be due to deletion in AgRP/NPY neurons.

The other issue is that the mitochondria data are correlative and do not provide compelling evidence linking mitochondrial dynamics and glucose regulation. Some reviewers, including the editor, suggest either providing a causal link between mitochondrial dynamics and glucose regulation, or removing the mitochondrial experiments from this story.

We agree with the editor and reviewers. Our goal was to define the mechanisms by which TCPTP could functionally modulate POMC neuronal excitability but we concede that our findings in this regard although exciting are correlative. We have revised the manuscript omitting the mitochondrial dynamics data. However, we have discussed the possibility that mitochondrial dynamics may mechanistically coordinate POMC neuronal excitability in the Discussion.

The authors should define TCPTP on first usage and note the mouse gene names do not include hyphens, e.g. Il6, not Il-6.

We have changed all gene names not to include hyphens and have defined TCPTP on fits usage in the Introduction.

Reviewer #1:

[…] I do not have major concerns that require additional experiments but I would suggest further analysis of the electrophysiology results. In particular, the authors compare the number (population size) of POMC neurons that are activated, inhibited or unresponsive to insulin amongst a number of experimental settings. Table 1 provides statistical comparisons for some of these studies, but not all (only for KO's and double Ko's). However, the authors claim that pharmacological TCPTP inhibition "significantly altered POMC responses to insulin" but statistical evidence is not provided. The same for the HFD study. The authors should provide a complete table of comparisons amongst the diverse experimental settings.

We have now included an additional table (Figure 1—figure supplement 2) detailing the statistical differences between POMC neurons that are excited, inhibited or non-responsive in response to insulin following either treatment with TCPTP inhibitor (Figure 1l) or high fat feeding (Figure 1K).

Reviewer #2:

[…] 1) The authors performed a large number of current-clamp recordings on POMC neurons to assess their responses to insulin. However, the results of membrane potential seem problematic in the current study. For example, in the first paragraph of the subsection “TCPTP defines POMC neurons that are activated or inhibited by insulin”, POMC neurons after insulin stimulation exhibited an averaged membrane potential of -40.1+/-2.5 mV; in the next paragraph of the aforementioned subsection, excited neurons had a membrane potential as depolarized as -37.9+/-1.7 mV. How could POMC neurons be so depolarized?

The data are presented as the absolute read-out at the time of experiments and do not account for the liquid junction potential, which for our recording conditions amounts to 10/11mV. Thus, the values are depolarised. To acknowledge this, we have now incorporated the following sentence into the electrophysiological recordings section in the supplemental experimental procedures section, “Resting membrane potential is expressed as the read-out from the amplifier and was not corrected for the liquid junction potential offset (11 mV).”

2) As shown previously by Padilla, Carmody and Zeltser, et al. (Nat Med 2010, and in a more convincing way recently by Xu et al., 2018, POMC-Cre exhibited ectopic Cre activity during development, particularly in AgRP neuron. Breeding Pomc-Cre with floxed mice could yield a manipulation of 25% AgRP neurons as well (Xu et al.). In Figure 1G-H, since the authors used Z/EG reporter instead of Pomc-eGFP transgenic mice to indicate POMC neurons, AgRP neurons were expected to be patched on as well. Since AgRP neurons are known to be inhibited by insulin, the ratio for the insulin-inhibited POMC neurons (12/33, 36.4%) may need to be adjusted.

See response to Editors. The reviewer is correct. Although we have discussed the caveat associated with the off-target effects of the Pomc-Cre transgene in POMC-TC mice on glucose metabolism in the Discussion and included a specific statement in the results highlighting the fact that we cannot exclude a proportion of inhibited neurons being AgRP neurons, we are reluctant to retrospectively adjust the number of neurons inhibited by insulin in POMC-TC;ZE/G mice without staining for NPY/AgRP neurons. Irrespective this would only under represent the number of POMC neurons being activated by insulin. Moreover to complement our findings we have also assessed POMC neuronal excitation/inhibition in Pomc-eGFP mice treated with the TCPTP inhibitor where we similarly show that TCPTP inhibition is accompanied by an increase in POMC neurons being activated by insulin.

3) Subsection “The activation or inhibition of POMC neurons regulates hepatic glucose metabolism”, the authors utilized inhibitory DREADDs to inhibit POMC neurons and included convincing expression data of hM4Di in Figure 2—figure supplement 1. However, a direct test of CNO's activity on POMC neurons with current-clamp recording shall also be included since hM4Di works through the activation of GIRK channels and not all neurons express them. In addition, CNO was recently reported to activate other endogenous receptors as well (Gomez et al., Science 2017). CNO's effects on non-DREADD-expressing animals shall also be included as a control.

hM4Di has been previously validated in POMC neurons (Koch et al., 2015, and Atasoy et al., 2012). These previous studies showed conclusively that in response to CNO administration, hM4Di attenuates c-Fos expression (a surrogate marker of neuronal activity) in ARC POMC neurons and reduces POMC neuronal firing rate as determined by current-clamp recordings. We have also conducted further electrophysiological experiments showing POMC neurons that do not express hM4Di, fail to respond to CNO. This indicates that CNO is not having any off-target effects and is not activating non-specific endogenous receptors on POMC neurons (Author response image 1).

Author response image 1

Download asset Open asset

To address the reviewer’s concern that the CNO metabolite, clozapine, may be having off target effects on glucose metabolism, we conducted GTTs, ITTs and PTTs in non-DREADD expressing Pomc-Cre mice following administration of vehicle or CNO (0.3mg/kg, i.p., 2 injections over 6h, Figure 2—figure supplement 1C-E). We found that CNO had no significant effect on whole body glucose metabolism (Figure 2—figure supplement 1C-E).

These findings are consistent with recent a recent study [Rossi M et al., (2018) J Clin Invest] that showed that a dose of 10 mg/kg (i.p.) CNO has no effect on fed/fasted blood glucose levels or glucose metabolism. The dose of CNO used in our study (0.3mg/kg injected twice over 6h) is ~33x lower than that described by Rossi et al. Furthermore, our results define a biphasic modulation of glucose metabolism by activating or inhibitory DREADDs in POMC neurons. It is highly unlikely that CNO alone elicits a biphasic regulation of glucose metabolism.

4) In the third paragraph of the Discussion, the authors discussed about the potential mechanisms with regards to synaptic regulation. Deletion of TCPTP in POMC neurons clearly suggested a post-synaptic mechanism but the authors believed in an existence of pre-synaptic regulation. This is confusing. Do the authors think of the involvement of a retrograde mechanism or different action sites of TCPTP from POMC neurons?

This was merely speculation as to mechanisms involved. Our EM analyses indicated that there was no overt change in relative excitatory and inhibitory inputs into POMC perikarya. However, beyond this, we did not specifically investigate presynaptic mechanisms but instead detailed postsynaptic changes in excitability. Exploring the specific mechanisms involved extends beyond the scope of this study.

Reviewer #3:

[…] 1) A key unanswered question is how increased insulin receptor signal transduction associated with reduced TCPTP is tied to a reversal of the effect of insulin on POMC neuron membrane potential and firing rate (from inhibitory to stimulatory). The authors state that this "phenotype switch" in the insulin response reflects direct (rather than an indirect) effects of insulin, but how increased IR-PI3K-Akt signal transduction in these cells might achieve this effect is unresolved. In an effort to shed light on this key question, the authors investigate whether changes in POMC cell mitochondrial dynamics underlie this phenotype switch. Although changes of insulin responsiveness appear to be coupled to morphological and genomic alterations affecting mitochondria, the data fail to establish a connection between these mitochondrial responses and changes of membrane potential and firing rate. Consequently, the data on mitochondrial dynamics seem somewhat disconnected from the rest of the paper, and might be more appropriate for a separate publication.

We agree with the reviewer and have omitted the mitochondrial dynamics data.

2) The authors finding of a robust effect of POMC-specific TCPTP deletion to enhance insulin suppression of HGP begs the question as to what other responses to insulin might also be affected. Given that ICV insulin has previously been reported to reduce food intake, body weight and body adiposity, and that brain-specific IR deletion promotes positive energy balance and weight gain, one wonders whether the effect of ICV insulin on food intake, body weight and body adiposity are also increased. Given the marked improvement of whole-body insulin sensitivity attributed to POMC-specific TCPTP deletion, one also wonders why the basal glucose and insulin levels were not reduced in these animals.

We have previously published that TCPTP deletion in POMC neurons has no effect on overnight food intake, adiposity or body weight [Dodd et al., 2015]. Nonetheless to address the reviewer’s concern, we have assessed the effects of ICV insulin (0.1mU/animal, 5 injections over 5 h as used in Figure 3, Figure 5G-H, Figure 6G-J, Figure 7A-D) on food intake, body weight and adiposity in 8-week-old male Ptpn2^fl/fl, POMC-TC and POMC-TC-IR mice (Figure 6—figure supplement 3A-D). Although the administration of insulin ICV had marked effects on glucose metabolism in POMC-TC mice, this was not accompanied by a reduction in body weight, food intake or whole-body adiposity at this dose and timeframe (Figure 6—figure supplement 3A-D).

As described in the manuscript, we noted a decrease in fasted plasma glucose and plasma insulin levels in POMC-TC mice (Figure 4A-B). We now also show that fasted blood glucose and plasma insulin levels are also reduced in high fat fed obese mice (Figure 7).

3) A related question is whether reduced body adiposity results from POMC-specific TCPTP deletion, and if so, did this contribute to enhanced insulin-induced suppression of HGP reported by the authors? Conversely, if POMC-specific TCPTP deletion has no effect on energy balance or body adiposity, this seems surprising given the expected increase of POMC neuron firing.

We have previously published that TCPTP deletion in POMC neurons has no effects on adiposity or bodyweight [Dodd et al., 2017) Cell]. TCPTP deletion in POMC neurons is only accompanied by the promotion of energy expenditure and decreased adiposity when accompanied by the concomitant enhancement of leptin signalling. Exploring the basis for this extends beyond the scope of this study. No differences in body weight or adiposity were evident in POMC-TC mice used in this study, precluding any changes in endogenous glucose production being related to this.

4) The question of how pronounced changes of POMC expression of TCPTP are associated with feeding, fasting and DIO was not addressed. Doing so would considerably strengthen the manuscript and seems of much greater relevance than the large amount of mitochondrial data that were included.

We have previously published that changes in POMC TCPTP expression are associated with feeding, fasting and diet-induced obesity [Dodd et al., 2017]; we have referenced this in the manuscript. We have shown that heightened glucocorticoid (corticosterone) levels in the fasted state drive the expression of TCPTP in hypothalamic neurons to suppress insulin signalling. By contrast feeding is associated with decreased glucocorticoid levels and concomitant Ptpn2 gene expression and the degradation of TCPTP protein. In the obese state we have shown that the feeding induced repression of TCPTP is abrogated so that TCPTP levels are always high emulating the fasted state; the precise reason for this is unclear, but probably occurs as a consequence of the heightened glucocorticoid and leptin levels [Loh et al., 2011] in obesity, which we have also shown drive hypothalamic TCPTP expression. We have discussed this in the revised manuscript

5) During embryogenesis, the POMC promoter is expressed in multiple arcuate nucleus neuronal subtypes including AgRP neurons. Since POMC-Cre was used to delete TCPTP, one expects the gene to be deleted across multiple arcuate nucleus neuronal cell types. Some effort to address this concern is warranted.

See responses to editors and reviewer 2 (point 2).

6) The authors state that diet-induced obesity reduces POMC neuron insulin responsiveness via a mechanism involving induction of TCPTP. Is this effect of DIO causally linked to associated responses such as inflammation, reactive gliosis, ER stress, etc.? Discussion of this question seems warranted.

See point 4. We have discussed this in the revised manuscript.

7) The authors report (Discussion, last paragraph) that TCPTP deletion has no detectable effect on glucose homeostasis in DIO mice, which they attribute to the fact that other mechanisms can account for impaired POMC insulin responsiveness in this setting. But they stop short of investigating whether TCPTP deletion enhances insulin responsiveness to POMC neurons in this setting. This question should be addressed since, if POMC neuron insulin responsiveness is in fact restored and yet there is no whole-body phenotype, the physiological relevance of POMC neuron insulin responsiveness would be called into question.

Our studies indicated that TCPTP deficiency in POMC neurons enhanced the ICV insulin induced repression of HGP as assessed in PTTs and measuring hepatic gluconeogenic gene expression. We also found that although fasted blood glucose levels (and now in revised manuscript fasted plasma insulin levels) were reduced, ITT and GTT responses were not altered. ITTs and GTTs are crude measure of glucose homeostasis and primarily measure glucose clearance mediated by skeletal muscle. Given this and the reviewer’s concern re physiological relevance we subjected 12-week high fat fed obese mice to hyperinsulinemic euglycemic clamps. We found that glucose infusion rates were significantly increased in POMC-TC mice (Figure 7G, Figure 7—figure supplement 1E), consistent with enhanced systemic insulin sensitivity. This was attributable exclusively to the suppression of HGP and reduced gluconeogenic (Pck1 and G6pc) gene expression (Figure 7J). Taken together our findings are consistent with elevated TCPTP in obesity perturbing POMC insulin responses to drive HGP and fasting hyperglycemia in obesity.

8) What precedent exists for a change in hormone responsiveness reversing its effect on the firing of a distinct neuronal subset? If such a response has been reported previously, was it referred to as "neural plasticity"? Certainly, this is not the type of response that comes to mind in association with that term. In this context, the term "neural plasticity" seems potentially misleading and for this reason, it is recommended that it be removed from the manuscript title.

The term neural plasticity is generally attributed to changes in synaptic efficacy, for example long-term potentiation or depression (LTP and LTD, respectively). However, the term plasticity itself by definition refers to changes in neuronal function, be it learning and memory, development whatever. This is clearly as we have seen in our study in relation to hormone responsiveness and therefore we consider the term entirely appropriate. Clearly these neurons do show plasticity in their responsiveness to hormones, which leads to changes in physiological function. Whilst we understand the referee’s point of view in that tradition has restricted the use of the term to phenomena such as LTP and LTD etc. we believe our use of the term to relate to functional plasticity of hormone responsiveness under a range of physiological conditions entirely appropriate. With regards to precedents for this, hormone responsiveness (noradrenaline) has previously been shown to reverse (excitation to inhibition) in orexin neurons following a short period of sleep deprivation (Grivel et al., 2005). In a metabolic context, glucose sensing responses in ARC neurons change depending on metabolic status in an energy-status-dependent manner (van den Top et al., 2017). These studies show a functional plasticity in these neural populations and set a precedence for such changes in response to hormones or nutrients in specific neuronal populations. We also believe that we are leading the way in driving awareness of the functional, energy-status-dependent plasticity of energy-sensing neural networks. We therefore respectfully consider our use of this term is entirely appropriate.

9) The effect of IP insulin to induce c-Fos in the PVH (Figure 1M) could have resulted from hypoglycemia, which was presumably well established by 90 min following the dose of insulin that was given. The effect of neuroglucopenia to activate PVH neurons is well established (e.g., Briski, Neuroreport, 1998; Briski, Brain Res Bulletin, 2000; Evans, AJP, 2001), and it is conceivable that changes of POMC cell insulin responsiveness affected the degree of insulin-induced hypoglycemia.

To address the reviewers concerns we administered a dose of 0.85 mU/g insulin (as described in Figure 1M) and measured blood glucose levels 90 min post injection in Ptpn2^fl/fl and POMC-TC mice (Figure 1—figure supplement 4A). We found that the insulin-induced hypoglycemia was similar in Ptpn2^fl/fl and POMC-TC mice. Therefore differences in hypoglycemia could not account for the enhanced PVH c-Fos immunoreactivity seen in POMC-TC mice (Figure 1M).

10) When insulin is given ICV and HGP or other highly insulin-responsive endpoints are measured, it's important to verify that none of the centrally-administered insulin leaked into the periphery to have direct effects on the liver.

We thank the reviewer for raising this important point. To address this concern, we fasted 8-week-old male C57BL/6J mice and administered ICV saline or insulin (0.1 mU/animal, 5 injections over 5 h, as indicated in Figure 3—figure supplement 1A-B as in Figure 3A-D, Figure 5A-H, Figure 7G-J, Figure 7B and Figure 3—figure supplement 1C) and plasma insulin levels were determined at 0, 15, 30, 60 and 120 min post-injection. Plasma insulin levels were determined using an insulin ELISA (Monash Antibody Technologies Facility) capable of detecting both mouse and human insulin.

We found that no significant difference in plasma insulin level between vehicle and ICV insulin treated mice (Figure 3—figure supplement 1A-B), indicated that the effects on liver glucose metabolism that we describe inFigure 3A-D, Figure 5A-H, Figure 7G-J, Figure 7B and Figure 3—figure supplement 1Care unlikely to result from insulin leaking into the periphery.

11) Relevant to this point is the use of a study design in which ICV insulin (0.1 mU) or vehicle was given as 5 separate injections over 5 hours. It is difficult to imagine that 5 ICV injections over 5 hours would be well-tolerated, which in turn raises the possibility that the authors were inadvertently measuring the impact of insulin on a stress response, which might explain the basal hyperglycemia reported in some studies. Beyond this, insulin in CSF has a relatively long half-life, so it's not clear why a repeated injection protocol over 5 hours was required. Were the animals anesthetized during this period? I could not find the information in the Materials and methods section.

Whilst it is likely that some level of stress will be caused by 5x separate ICV injections over 5 hours, we minimized handling stress by handling the mice 2 weeks prior to the experiment and conducting dummy injections 2 days prior to the experiment. All treatment groups received repeated ICV injection of either vehicle or insulin and were thus subjected to the same amount of handling stress. We conclude that handling stress is not responsible for the effects of TCPTP deletion in POMC neurons on glucose metabolism described in Figure 3A-D, Figure 5A-H, Figure 7G-J, Figure 7B and Figure 3—figure supplement 1C. Furthermore, the enhanced insulin-POMC-mediated repression of hepatic glucose production (as assessed by the ability of ICV insulin to repress pyruvate-induced glucose excursions and hepatic gluconeogenic gene expression) was corrected in POMC-TC-IR mice (Figure 6G-I) reaffirming the importance of insulin signalling.

Mice were not anaesthetized during injections as anesthesia has confounding actions on glucose homeostasis and 5 repeated injections over 5 h of a low dose of insulin was preferred to a single high dose of insulin which could potentially result in neuroglucopenia or result in a confounding amount of insulin “spill over’ into the peripheral circulation.

12) The basal glucose level (time=0 min) of both groups in the pyruvate tolerance test shown in Figure 2C is ~12 mM, which is in the diabetic range. By comparison, basal glucose levels of 7 mM are reported for the same test in Figure 2J, and are uniformly <10 mM in all other such studies (Figures 4, 5 and 6). This discrepancy is all the more notable in light of the remarkably small variances (SEMs) around the glucose data, which is inconsistent with known biological variability inherent in the response.

We would like to make clear that the discrepancy between the basal glucose levels shown in Figure 2C and Figure 2J is due to the fact that the pyruvate tolerance tests (PTT) depicted in Figure 2C were conducted in the fed state (where blood glucose levels would be elevated), whereas the PTTs depicted in Figure 2J are were conducted in the fasted state (where glucose levels would be comparatively lower).

Figure 2 is looking at the biphasic actions of the pharmacogenetic modulation of POMC neurons on hepatic glucose production (HGP) using a PTT as a surrogate measure of gluconeogenesis. We reasoned that POMC neuronal inhibition would promote hepatic glucose production whereas POMC excitation would attenuate HGP. In order to tease apart these differences, mice were placed in opposing nutritional states, fed verse fasted, (Figures2A, B, I) which is essential to overcome any confounding aspects of endogenous HGP.

13) The Discussion is longer than it need be, and tends to reiterate points made in the Results section. It could be cut substantially.

The Discussion has been shortened.

Author details

1. Garron T Dodd

   1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
   2. Department of Biochemistry and Molecular Biology, Monash University, Victoria, Australia

   Contribution

   Conceptualization, Formal analysis, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing

   Contributed equally with

   Natalie J Michael

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7554-4876

2. Natalie J Michael

   1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
   2. Department of Physiology, Monash University, Victoria, Australia

   Contribution

   Investigation, Methodology, Writing—review and editing

   Contributed equally with

   Garron T Dodd

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9032-0862

3. Robert S Lee-Young

   1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
   2. Department of Biochemistry and Molecular Biology, Monash University, Victoria, Australia
   3. Monash Metabolic Phenotyping Facility, Monash University, Victoria, Australia

   Contribution

   Investigation, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

4. Salvatore P Mangiafico

   Department of Medicine (Austin Hospital), The University of Melbourne, Melbourne, Australia

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

5. Jack T Pryor

   1. Department of Physiology, Monash University, Victoria, Australia
   2. Warwick Medical School, University of Warwick, Coventry, United Kingdom

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

6. Astrid C Munder

   1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
   2. Department of Physiology, Monash University, Victoria, Australia

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

7. Stephanie E Simonds

   1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
   2. Department of Physiology, Monash University, Victoria, Australia

   Contribution

   Resources, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

8. Jens Claus Brüning

   1. Department of Neuronal Control of Metabolism, Max Planck Institute for Metabolism Research, Cologne, Germany
   2. Center for Endocrinology, Diabetes, and Preventive Medicine, University Hospital Cologne, Cologne, Germany
   3. Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases, University of Cologne, Cologne, Germany
   4. Center for Molecular Medicine Cologne, University of Cologne, Cologne, Germany
   5. National Center for Diabetes Research, Neuherberg, Germany

   Contribution

   Conceptualization, Resources, Writing—review and editing

   Competing interests

   No competing interests declared

9. Zhong-Yin Zhang

   Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, United States

   Contribution

   Resources, Funding acquisition, Writing—review and editing

   Competing interests

   No competing interests declared

10. Michael A Cowley

    1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
    2. Department of Physiology, Monash University, Victoria, Australia

    Contribution

    Conceptualization, Resources, Funding acquisition, Writing—review and editing

    Competing interests

    No competing interests declared

11. Sofianos Andrikopoulos

    Department of Medicine (Austin Hospital), The University of Melbourne, Melbourne, Australia

    Contribution

    Conceptualization, Resources, Funding acquisition, Writing—review and editing

    Competing interests

    No competing interests declared

12. Tamas L Horvath

    1. Program in Integrative Cell Signaling and Neurobiology of Metabolism, Department of Comparative Medicine, Yale University School of Medicine, New Haven, United States
    2. Department of Anatomy and Histology, University of Veterinary Medicine, Hungary, Europe

    Contribution

    Conceptualization, Resources, Formal analysis, Funding acquisition, Investigation, Methodology, Writing—review and editing

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-7522-4602

13. David Spanswick

    1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
    2. Department of Physiology, Monash University, Victoria, Australia
    3. Warwick Medical School, University of Warwick, Coventry, United Kingdom

    Contribution

    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing—review and editing

    For correspondence

    David.Spanswick@monash.edu

    Competing interests

    No competing interests declared

14. Tony Tiganis

    1. Metabolism, Diabetes and Obesity Program, Monash Biomedicine Discovery Institute, Monash University, Melbourne, Australia
    2. Department of Biochemistry and Molecular Biology, Monash University, Victoria, Australia
    3. Monash Metabolic Phenotyping Facility, Monash University, Victoria, Australia

    Contribution

    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Validation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    Tony.Tiganis@monash.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-8065-9942

• 3,725

  Page views

• 708

  Downloads

• 71

  Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.