- Email: [email protected]
Hippocampal insulin resistance links maternal obesity with impaired neuronal plasticity in adult offspring
Psychoneuroendocrinology 89 (2018) 46–52
Contents lists available at ScienceDirect
Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen
Hippocampal insulin resistance links maternal obesity with impaired neuronal plasticity in adult offspring
T
Lisa Schmitza,1, Rebecca Kuglina,1, Inga Bae-Gartza, Ruth Janoscheka, Sarah Appela, Andrea Mesarosb, Igor Jakovcevskic,d, Christina Vohlena, Marion Handwerka, Regina Ensenauere, ⁎ Jörg Dötscha, Eva Hucklenbruch-Rothera, a
Department of Pediatrics and Adolescent Medicine, University Hospital of Cologne, Kerpener Str. 62, Cologne, 50924, Germany Max Planck Institute for Biology of Ageing, Phenotyping Core Facility, Joseph-Stelzmann-Str. 9b, Cologne, 50931, Germany c Institute for Molecular and Behavioral Neuroscience, Center for Molecular Medicine Cologne, Cologne, Germany d Experimental Neurophysiology, German Center for Neurodegenerative Diseases, Bonn, Germany e Experimental Pediatrics and Metabolism, University Children’s Hospital, Heinrich Heine University Düsseldorf, Düsseldorf, Germany b
A R T I C L E I N F O
A B S T R A C T
Keywords: Insulin resistance Hippocampus Maternal obesity Synapse
Objective: Maternal obesity and a disturbed metabolic environment during pregnancy and lactation have been shown to result in many long-term health consequences for the offspring. Among them, impairments in neurocognitive development and performance belong to the most dreaded ones. So far, very few mechanistic approaches have aimed to determine the responsible molecular events. Methods: In a mouse model of maternal diet-induced obesity and perinatal hyperinsulinemia, we assessed adult offspring’s hippocampal insulin signaling as well as concurrent effects on markers of hippocampal neurogenesis, synaptic plasticity and function using western blotting and immunohistochemistry. In search for a potential link between neuronal insulin resistance and hippocampal plasticity, we additionally quantified protein expression of key molecules of synaptic plasticity in an in vitro model of acute neuronal insulin resistance. Results: Maternal obesity and perinatal hyperinsulinemia result in adult hippocampal insulin resistance with subsequently reduced hippocampal mTor signaling and altered expression of markers of neurogenesis (doublecortin), synaptic plasticity (ampaloxO1, pSynapsin) and function (vGlut, vGAT) in the offspring. The observed effects are independent of the offspring’s adult metabolic phenotype and can be associated with multiple previously reported behavioral abnormalities. Additionally, we demonstrate that induction of insulin resistance in cultured hippocampal neurons reduces mTor signaling, doublecortin and vGAT protein expression. Conclusions: Hippocampal insulin resistance might play a key role in mediating the long-term effects of maternal obesity and perinatal hyperinsulinemia on hippocampal plasticity and the offspring’s neurocognitive outcome.
1. Introduction Obese pregnancies are on the verge of becoming the rule, not the exception in many developed countries. Only recently, neurocognitive impairments have emerged as a very unfavourable outcome in the offspring of obese pregnancies (Rivera et al., 2015). Specifically, human and animal studies have shown that maternal obesity is associated with numerous neurodevelopmental and psychiatric disorders, including intellectual deficits, anxiety, depression, attention deficit hyperactivity disorder, and autism (Edlow, 2017). Insulin has been identified as an important modulator of neuronal network development – mostly in the hypothalamus – during early phases of life (Dearden and Ozanne,
⁎
1
2015). In detail, hyperinsulinemia following maternal overnutrition has been shown to alter neuronal projections of feeding neurons located in the arcuate nucleus of the hypothalamus affecting the offspring’s pancreatic parasympathetic innervation and glucose-stimulated insulin secretion (Vogt et al., 2014). Despite the undeniable relevance of neurocognitive long-term consequences, hippocampal neurogenesis, synaptic plasticity and function have not been the major focus of mechanistic studies on maternal obesity–related offspring pathology to date (Rivera et al., 2015; Dearden and Ozanne, 2015). Previously, we have shown that maternal diet-induced obesity in the mouse results in transiently increased body weight and body fat content at the end of lactation, going along with profound
Corresponding author. E-mail address: [email protected] (E. Hucklenbruch-Rother). Authors contributed equally.
https://doi.org/10.1016/j.psyneuen.2017.12.023 Received 15 August 2017; Received in revised form 20 November 2017; Accepted 27 December 2017 0306-4530/ © 2017 Elsevier Ltd. All rights reserved.
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
software, respectively (Bruker, Belgium). Images were segmented based on tissue density for both total volume and fat volume. Total fat volume was further segmented into visceral and subcutaneous fat using the abdominal muscular wall as orientation.
hyperinsulinemia and impaired glucose tolerance at postnatal day (P) 21 (Janoschek et al., 2016; Bae-Gartz et al., 2016). After being weaned to standard chow, however, 70 day old offspring of obese mouse dams appear to be metabolically indistinguishable from control offspring (Janoschek et al., 2016; Bae-Gartz et al., 2016). In the current study, we aimed to determine the long-term effects of maternal obesity and the consecutive major changes in early metabolic environment on hippocampal insulin sensitivity and synaptic plasticity and function in the offspring at P70. Since insulin’s intracellular signaling cascade involves activation of mTor (mammalian target of rapamycin), a well characterized serine/threonine kinase which has recently gained attention due to its involvement in neurocognitive function (Graber et al., 2013; Bergeron et al., 2014) and synaptic plasticity (Sosanya et al., 2015; Dwyer et al., 2015), we additionally set out to examine whether reduced mTor activation in the hippocampus might link hippocampal insulin resistance with altered plasticity of hippocampal neuronal networks in adult offspring of obese mothers.
2.4. Cell culture experiments HT22 cells were plated with a cell count of 10^6 per 10 cm dish and maintained in their feeding medium (DMEM (Gibco, 41966-029) + 10% FBS (Biochrom, S0615) + 1% Pen/Strep (Sigma Aldrich, P4458)) for 16 hrs. For 24hrs-induction of insulin resistance (Fig. 4 a, c, Suppl. Fig. 2 a, b), cells were exposed to fresh feeding medium either with (IR) or without (CO) 20 nM insulin for 24 hrs as described before (Kim et al., 2011). Following induction of insulin resistance, cells were washed with PBS and either acutely stimulated with fresh feeding medium containing 20 nM insulin for another 15 min or not. Cells were washed with PBS, harvested, and frozen at −80 °C for protein extraction and Western blotting. For 72hrs-induction of insulin resistance (Fig. 4 a, b, Suppl. Fig. 2 c), cells were exposed to fresh feeding medium either with (IR) or without (CO) 20 nM insulin every 24 hrs for 72 hrs. After 72 hrs, cells were washed with PBS and either acutely stimulated with fresh feeding medium containing 20 nM insulin for another 15 min or not. Cells were again washed with PBS, harvested, and frozen at −80 °C for protein extraction and Western blotting.
2. Material and methods 2.1. Experimental design and animal procedures All animal procedures were conducted in compliance with protocols approved by local government authorities (Land NRW) and were in accordance with National Institutes of Health guidelines. Mice (C57BL/ 6N) were bred locally at a designated animal unit of the University Hospital of Cologne (Cologne, Germany). For detailed information on experimental design, animal housing, diet content, and mating scheme, see (Janoschek et al., 2016; Bae-Gartz et al., 2016). Briefly, control (CO) females, receiving standard chow at all times, and high fat diet (HFD)-fed females, receiving high fat diet starting upon their weaning at three weeks of age for the rest of the experiment, were mated at 12–14 weeks of age. All offspring studies were performed using male offspring. On P3, litter size was adjusted to six for each litter. At P21, in different subsets of only one animal per litter, either blood samples were taken, glucose tolerance testing (GTT) was performed, or animals underwent micro computed tomography (μCT) measurement. At P70, again, in subsets of one animal per litter, either blood samples were taken, GTT was performed, or animals were sacrificed. Hippocampus from both hemispheres was rapidly dissected and stored at −80 °C. Also, liver, skeletal muscle, and white adipose tissue were dissected, snap-frozen, and stored at −80 °C. The brain of one animal per litter was paraformaldehyde (PFA) fixated, mounted in tissue freezing medium, and stored at −80 °C for immunohistochemistry.
2.5. Western blotting Frozen tissue or cells were homogenized in lysis buffer as previously described (Bae-Gartz et al., 2016). Protein concentration was determined with a BCA-Protein Assay Kit (Thermo Scientific, Waltham, USA). Lysates resolved on a 10% reducing SDS-PAGE gel were transferred to a PVDF membrane. Blots were probed with the antibodies (Suppl. Table 1). 2.6. Immunohistochemistry For immunohistochemistry, 20 μm thick coronal cryosections were washed in PBS and permeabilized with 0.3% Triton-X-100 (SigmaAldrich, T8532) in PBS for 30 min. After blocking of non-specific binding components with Sea Block blocking Buffer (ThermoFisher, 37527) for 2 h, samples were incubated with primary antibodies (Suppl. Table 1) dissolved in antibody diluent (Dako, #S202230). After an overnight incubation at 4 °C and intensive washing in PBS, samples were incubated with secondary antibodies at room temperature for 2 hrs. Slices were rinsed in PBS and cover-slipped in Fluoroshield Mounting Medium containing 4′,6-Diamidin-2-phenylindol (DAPI) (Abcam. ab104139). For morphometric analysis, fluorescence images were taken by Olympus BX43F with cellSens Dimension software (DP80 dual CCD Camera, cellSens Dimension (V1.8)) and analyzed with the aid of ImageJ software supplemented with Olympus viewer plugin. All analyses are performed in a double-blind setting by at least two independent researchers.
2.2. Intraperitoneal glucose tolerance test and analytical procedures Glucose tolerance tests were performed as previously described (Janoschek et al., 2016; Bae-Gartz et al., 2016). Briefly, after overnight fasting (16 hrs), each animal received an intraperitoneal (ip) injection of 2 g glucose/kg body weight. Blood glucose levels were measured before glucose injection and after 15, 30, 60, and 120 min using an automatic glucose monitor (GlucoMen; A. Menarini Diagnostics, Berlin, Germany). Serum levels of insulin were measured by ELISA using mouse standards according to the manufacturer's guidelines (mouse insulin ELISA (EZRMI–13 K); Millipore CorpBillerica, MA).
2.7. Statistics Values are reported as mean +/− SD. All statistical analysis was performed in GraphPad Prism 6 software. After testing for normality, we performed an unpaired t-test or a Mann-Whitney t-test (for nonparametric distribution). Significance was set at p < 0.05.
2.3. Quantification of fat by μCT (micro computed tomography) Whole mice were scanned post mortem with a μCT scanner (SkyScan 1176, Bruker, Belgium) with an isotropic voxel size of 35.26 μm3. The x-ray settings for each scan were 45 kV and 475 μA using a 0.5 mm aluminum filter. All scans were performed over 360 ° with a rotation step of 0.6 ° and a frame averaging of 2. Images were reconstructed, analyzed and visualized using NRecon, CTAn and CTVox
3. Results 3.1. Offspring metabolic phenotype at P21 and P70 Offspring of obese mouse dams show increased body weight, body fat content, serum insulin levels, and significantly increased 15 min 47
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 1. Offspring metabolic phenotype at P21 and P70. (a) Representative μCT scan of a control (CO) offspring at P21 for quantification of subcutaneous (yellow) and visceral (red) fat. (b) Representative μCT scan of a maternal obesity (HFD) offspring at P21 for quantification of subcutaneous (yellow) and visceral (red) fat. (c) Summary of phenotypical characteristics of CO and HFD offspring at P21 (as previously reported in [5,6]); displayed values are means +/− SD; body weight: CO n = 87, HFD n = 58, p < 0.0001; serum insulin: CO n = 21, HFD n = 15, p < 0.0001; GTT and AUC: CO n = 12, HFD n = 9, P < 0.05 at 15 min. (d) Summary of phenotypical characteristics of CO and HFD offspring at P70 (as previously reported in [5,6]); displayed values are means +/− SD; body weight: CO n = 66, HFD n = 35, p = 0.6105; serum insulin: CO n = 10, HFD n = 10, p = 0,1607; GTT and AUC: CO n = 24, HFD n = 22.
48
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 2. Hippocampal insulin signaling at P70. (a) Western blots of the insulin receptor (InsR), phosphorylated and total protein kinase B (pAKT, AKT), forkhead box protein (FoxO) 1, phosphorylated and total mammalian target of rapamycin (pmTor, mTor), phosphorylated and total ribosomal protein s6 kinase beta-1 (pp70s6 K, p70s6 K), phosphorylated and total 4E-binging protein 1 (p4E-BP1, 4E-BP1), and their respective loading controls (glyceraldehyde 3-phosphoate dehydrogenase (GAPDH)) in CO and HFD offspring at P70. (b) Densitometrical quantification of western blots displayed in a). Relative protein expression normalized to GAPDH in HFD offspring compared to CO with CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. * p < 0.05, ** p < 0.01. (c) Simplified schematic overview of the insulin signaling cascade summarizing the effects of maternal obesity on hippocampal protein expression at P70. Arrows indicate significant downregulation of InsR, pmTor, pp70s6 K, and FoxO1 in the offspring’s hippocampus following maternal obesity.
neuronal network plasticity, we first determined protein expression of doublecortin (DCX), a specific marker for new-born neurons, in the hippocampus of offspring of lean and obese mouse dams at P70. DCX protein expression in total hippocampus tissue homogenate was reduced to roughly 30% of control levels following maternal obesity (Fig. 3a and b). Using the same method, we additionally quantified phosphorylation of Synapsin, which is known to facilitate the transport of neurotransmitter-filled vesicles to the synapse during an action potential and thereby promotes synaptic plasticity. We detected a significant 70% reduction in Synapsin phosphorylation in offspring of obese mouse dams at P70 (Fig. 3a and c). As an indicator of neuronal activation and indirect measure for synaptic function we next visualized phosphorylation of CREB (cAMP response element-binding protein) using immunohistochemistry. Despite visible reduction in immunoreactivity in the DG, statistical testing failed to yield significance (p = 0.073) in our analyses of 3 vs. 4 animals (Fig. 3d, k, for orientation see h). To characterize any qualitative differences in hippocampal synaptic input organization, we additionally visualized and quantified glutamatergic (excitatory) and GABAergic (inhibitory) nerve terminals by immunohistochemistry. For quantification of excitatory synapses, we first quantified immunoreactivity for the total and phosphorylated glutamate receptor subunit GluR1 at P70. We detected only a slight reduction in total hippocampal GluR1 immunoreactivity in offspring of obese dams (Fig. 3e, I, for orientation see h), but no significant differences for GluR1 in specific sub-regions or for phosphorylated GluR1 (Fig. 3e, i, l, for orientation see h). Additionally, we determined immunoreactivity for the vesicular glutamate transporter 2 (vGlut2) and detected significant local reduction in immunoreactivity in the Cornu
peak in the GTT at P21 (Fig. 1a–c) (Janoschek et al., 2016). AUC of the GTT at P21 shows an almost 20% increase in HFD offspring (p = 0.065). AT P70, however, all of these parameters have returned to control levels (Fig. 1d). 3.2. Hippocampal insulin signaling at P70 To determine the long-term effects of maternal obesity on hippocampal sensitivity to insulin, we first quantified protein expression levels of insulin’s specific receptor (InsR) in the hippocampus of adult offspring of lean and obese dams at P70 and found a significant 30% reduction (Fig. 2a and b). Next, we assessed intracellular AKT/mTor signaling by quantifying phosphorylation of AKT, mTor, and several of their downstream targets in the hippocampus at P70. We found a statistically significant reduction in activation of mTor and p70S6 K while phosphorylation of AKT and 4EBP1 only showed a tendency towards reduction in HFD offspring (Fig. 2a and b). Furthermore, protein expression of FoxO1, an important transcription factor expressed by developing hippocampal neurons in the granule cell layer of the dentate gyrus and responsible for intact neuronal polarity and axon outgrowth (la Torre-Ubieta et al., 2011) was found to be significantly reduced by more than 50% (Fig. 2a and b). At the same time, protein expression of InsR and phosphorylation of AKT was found to be unaltered in skeletal muscle and white adipose tissue, and even significantly increased in the liver of HFD offspring (Suppl. Fig. 1 a). 3.3. Hippocampal neurogenesis, synaptic plasticity and function at P70 To evaluate the effects of maternal obesity on hippocampal 49
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 3. Hippocampal neurogenesis, synaptic plasticity and function at P70. (a) Western blots of doublecortin (DCX) and phosphorylated and total Synapsin (pSyn, Syn), and their respective loading controls (GAPDH) in CO and HTFD offspring at P70. (b) Densitometrical quantification of the DCX western blot displayed in a). Relative protein expression normalized to GAPDH, CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. ** p < 0.01. (c) Densitometrical quantification of the pSyn/Syn western blot displayed in a). Relative protein expression of pSyn/Syn, CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. ** p < 0.01. (d) Representative immunohistochemical co-staining for the phosphorylated cAMP response element binding protein (pCreb) and the neuron marker NeuN in the dentate gyrus (DG) of CO (upper panel) and HFD (lower panel) offspring at P70. Neuron marker NeuN (green), pCreb (red), Nuclear stain Dapi (blue). (e) Representative immunohistochemical co-staining for the glutamate receptor 1 (GluR1) and the neuron marker NeuN in the dentate gyrus (DG) of CO (upper 4 panels) and HFD (lower 4 panels) offspring at P70. Each set of 4 shows the nuclear staining alone (Dapi, top left), GluR1 alone (top right), neuron marker alone (NeuN, bottom left), and a merge of all three stainings (bottom, right). Neuron marker NeuN (green), GluR1 (red), nuclear stain Dapi (blue). (f) Representative immunohistochemical staining for the vesicular GABA transporter (vGAT (red)) in the CA1 region of CO (upper panel) and HFD (lower panel) offspring at P70. (g) Representative immunohistochemical staining for the vesicular glutamate transporter 2 (vGlut2 (green)) in the CA 2 region of CO (upper panel) and HFD (lower panel) offspring at P70. (h) Overwiew of hippocampal structures in coronal brain slices. White boxes indicate the subregions displayed in d, f, and g, respectively. Scale bar = 300 μm. (i) Quantification of immunoreactivity for GluR1 (calculated as raw intensity/area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 5, HFD n = 6. * p < 0.05. (j) Quantification of immunoreactivity for vGAT (calculated as vGAT positive area/total area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 4, HFD n = 7. * p < 0.05. (k) Quantification of immunoreactivity for pCreb (calculated as pCreb positive area/NeuN positive area) in total hippocampus and the DG. Displayed values are means +/− SD; CO n = 4, HFD n = 7. (l) Quantification of immunoreactivity (calculated as raw intensity/area) for phosphorylated GluR1 (pGluR1) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 5, HFD n = 6. * p < 0.05. (m) Quantification of immunoreactivity for vGlut2 (calculated as raw intensity/area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/ − SD; CO n = 4, HFD n = 7. * p < 0.05.
extrahippocampal brain regions within the same sections were also analyzed. In the cortex of HFD offspring at P70, we found a significant reduction in pCreb expression along with significantly reduced immunoreactivity for vGlut2. vGAT, GluR1, and pGluR1 immunoreactivity was not significantly altered (Suppl. Fig. 1 b). Within the amygdala, only vGAT, GluR1, and pGluR1 immunoreactivity was quanitifiable. No significant changes in immunoreactivity were
Ammonis (CA) 3 region and the Dentate Gyrus (DG) of the hippocampus following maternal obesity (Fig. 3g, m, for orientation see h). For the vesicular GABA transporter (vGAT), a marker for inhibitory nerve terminals, we detected significantly lower immunoreactivity in total hippocampus and all sub-regions except CA3 (Fig. 3f, j, for orientation see h). To evaluate specificity of the hippocampal findings, 50
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 4. In vitro effect of insulin resistance on neuronal plasticity. (a) Quantification of protein expression in insulin resistant (IR) HT22 cells relative to control (CO) cells in vitro (0 equals mean of controls and negative values represent downregulation in IR cells (e.g. −0.5 = 50% reduction compared to controls)). Relative protein expression of phosphorylated protein kinase B (pAKT), phosphorylated mammalian target of rapamycin (pmTor), doublecortin (DCX), the vesicular glutamate transporter 2 (vGlut2), and the vesicular GABA transporter (vGAT) was normalized to GAPDH. Lower panel (light grey): 24 hrs induction of insulin resistance. Upper panel (dark grey): 72 hrs induction of insulin resistance. (b) Exemplary bands from Western blots of pAKT, pmTor, DCX, vGlut2, and vGAT of IR und CO cells 72 hrs upon induction of insulin resistance (original blots displayed in Suppl. Fig. 2 c). (c) Exemplary bands from Western blots of pAKT, pmTor, DCX, vGlut2, and vGAT of IR und CO cells 24 hrs upon induction of insulin resistance (original blots displayed in Suppl. Fig. 2 b).
reported by ourselves and others (Janoschek et al., 2016; Bae-Gartz et al., 2016; White et al., 2009a), hippocampal insulin resistance as a potential underlying mechanism for impaired neurocognitive development and outcome has to our knowledge so far not been investigated. Insulin resistance is a key feature of the metabolic syndrome in mice and men. In our model, however, hippocampal insulin resistance is detected while mice lack signs of insulin resistance in liver, skeletal muscle, or white adipose tissue and do not display any increase in body weight, body fat content, or serum hormone levels. This is intriguing and implies that a constitutional difference in hippocampal hormone receptor abundance might be a consequence of transient disturbances, such as increased body weight, body fat content, or high levels of circulating hormones and cytokines during vulnerable developmental stages (Janoschek et al., 2016; Bae-Gartz et al., 2016). Independent of its origin, the clear reduction in hippocampal insulin action is inevitably linked with chronic inflammatory processes and oxidative stress – two very important contributors to cognitive decline (for review see (Verdile et al., 2015)). And indeed, multiple studies already showed that maternal obesity is associated with increased oxidative stress and inflammation in the offspring’s hippocampus resulting in impaired hippocampus-dependent behavior (Edlow, 2017; Tozuka et al., 2010; Bilbo and Tsang, 2010; White et al., 2009b). A connection between oxidative stress, inflammation and insulin resistance in the hippocampus of offspring of obese mothers is thus highly likely. Yet, the question of cause or consequence remains unanswered. With regard to neurogenesis, the interpretation of our data lies at hand: hippocampal neurogenesis, marked by DCX, is known to be clearly reduced in diabetic animals and can be restored by treatment with metformin, a wellstudied insulin sensitizer (Hwang et al., 2010). Reduced hippocampal neurogenesis, on the other hand, strongly correlates with reduced hippocampus-dependent cognition (Loxton and Canales, 2017). Thus, a clear reduction in hippocampal DCX expression in our model might at least partly result from local insulin resistance and gate way to cognitive and behavioral abnormalities. Grillo et al. provided first evidence for the role of hippocampal insulin resistance in long-term potentiation and hippocampus-dependent behavior (Grillo et al., 2015). While they reported an associated reduction in phosphorylation of the glutamate receptor, we observed reduced abundance of the total glutamate receptor without changes in its phosphorylation, with reduced hippocampal immunoreactivity of the vesicular glutamate transporter (vGlut2) in CA3 and the DG. Additionally, we see a clear reduction in vGAT immunoreactivity across most hippocampal subregions, which has interestingly previously been described as a consequence of gestational or early postnatal hypothyroidism in rats with concomitant cognitive and behavioral abnormalities (Navarro et al., 2015). Reciprocally, increased vGAT expression is
detected (Suppl. Fig. 1 c). 3.4. In vitro effects of insulin resistance on neuronal plasticity Hypothesizing a mechanistic link between insulin resistance in hippocampal neurons and neurogenesis, neuronal plasticity and function, we aimed to induce insulin resistance in an immortalized murine hippocampal cell line (HT22) and screened for changes in protein expression and activation of markers and enzymes found to be altered in our in vivo model at P70. As described in the literature, treatment of cells with insulin for 24 or 72 hrs renders them insulin resistant (IR) based on a downregulation of the insulin receptor (Suppl. Fig. 2a) with blunted AKT phosphorylation response to an acute insulin stimulus (15 min) (Kim et al., 2011) (Suppl. Fig. 2b, c). Paralleling the findings in our in vivo model, IR hippocampal neurons also showed reduced levels of phosphorylated mTor, most prominent after 72 hrs (Fig. 4a and b, Suppl. Fig. 2b, c). Furthermore, DCX protein expression appeared to be reduced in insulin resistant cells after 24 and 72 hrs, while vGlut2 expression seemed unaffected at both time points (Fig. 4a–c, Suppl. Fig. 2 b, c). A reduction was also observed for vGAT protein expression 24 hrs upon induction of insulin resistance, pointing towards a potential connection between neuronal insulin action and inhibitory synapse function in vitro. At 72 hrs, however, vGAT expression had returned to control levels. 4. Discussion Neurocognitive impairments range among the most dreaded longterm consequences for the offspring of obese mothers (Edlow, 2017). Here, we report impaired insulin signaling in the hippocampus of adult offspring of obese mouse dams paralleled with significantly altered hippocampal expression and distribution of markers of neurogenesis, synaptic plasticity and function. Additionally, we suggest a mechanistic link between neuronal insulin signaling and protein expression of synaptic transmission markers in vitro. It has long been suggested that synaptic insulin receptor signaling directly affects learning and memory (Cordner and Tamashiro, 2015). But only recently, an elegant study by Grillo et al. was able to show that site-specific hippocampal knockdown of the insulin receptor in adult rats results in changes in hippocampal plasticity and in hippocampusdependent behavior (Grillo et al., 2015). Furthermore, maternal obesity has repeatedly been linked to impaired hippocampus-dependent behavior in the offspring, including impaired hippocampal learning, decreased sociability, sex-dependent hyperactivity, and increased anxiety (for review see (Edlow, 2017)). Although impaired systemic glucose tolerance in the offspring of obese mothers has been previously 51
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
associated with memory enhancement and anxiolysis (Jansone et al., 2016), letting a functional role of vGAT in maternal obesity-induced cognitive impairment seem quite conceivable. In vitro, we demonstrate a potential connection between neuronal insulin signaling and vGAT protein expression, supporting the in vivo results. Lastly, the strong reduction in hippocampal pSynapsin in our model might directly contribute to the reduction in vGAT expression as shown before (Bogen et al., 2006), but is of course also on its own indicative of disturbed hippocampal transmitter trafficking and impaired plasticity following maternal obesity (Cesca et al., 2010).
2015.06.008. Dearden, L., Ozanne, S.E., 2015. Early life origins of metabolic disease: developmental programming of hypothalamic pathways controlling energy homeostasis. Front. Neuroendocrinol. 39, 3–16. http://dx.doi.org/10.1016/j.yfrne.2015.08.001. Dwyer, J.M., Maldonado-Avilés, J.G., Lepack, A.E., DiLeone, R.J., DUMAN, R.S., 2015. Ribosomal protein S6 kinase 1 signaling in prefrontal cortex controls depressive behavior. Proc. Natl. Acad. Sci. U. S. A. 112 (19), 6188–6193. http://dx.doi.org/10. 1073/pnas.1505289112. Edlow, A.G., 2017. Maternal obesity and neurodevelopmental and psychiatric disorders in offspring. Prenat. Diagn. 37 (1), 95–110. http://dx.doi.org/10.1002/pd.4932. Graber, T.E., McCamphill, P.K., Sossin, W.S., 2013. A recollection of mTOR signaling in learning and memory. Learn. Mem. (Cold Spring Harbor, N.Y.) 20 (10), 518–530. http://dx.doi.org/10.1101/lm.027664.112. Grillo, C.A., Piroli, G.G., Lawrence, R.C., Wrighten, S.A., Green, A.J., Wilson, S.P., et al., 2015. Hippocampal insulin resistance impairs spatial learning and synaptic plasticity. Diabetes 64 (11), 3927–3936. http://dx.doi.org/10.2337/db15-0596. Hwang, I.K., Kim, I.Y., Joo, E.J., Shin, J.H., Choi, J.W., Won, M.-H., et al., 2010. Metformin normalizes type 2 diabetes-induced decrease in cell proliferation and neuroblast differentiation in the rat dentate gyrus. Neurochem. Res. 35 (4), 645–650. http://dx.doi.org/10.1007/s11064-009-0115-5. Janoschek, R., Bae-Gartz, I., Vohlen, C., Alcázar, M.A.A., Dinger, K., Appel, S., et al., 2016. Dietary intervention in obese dams protects male offspring from WAT induction of TRPV4, adiposity, and hyperinsulinemia. Obesity 24 (6), 1266–1273. http:// dx.doi.org/10.1002/oby.21486. Jansone, B., Kadish, I., van Groen, T., Beitnere, U., Plotniece, A., Pajuste, K., et al., 2016. Memory-enhancing and brain protein expression-stimulating effects of novel calcium antagonist in Alzheimer's disease transgenic female mice. Pharmacol. Res. 113 (Pt. B), 781–787. http://dx.doi.org/10.1016/j.phrs.2016.06.020. Kim, B., Sullivan, K.A., Backus, C., Feldman, E.L., 2011. Cortical neurons develop insulin resistance and blunted akt signaling: a potential mechanism contributing to enhanced ischemic injury in diabetes. Antioxid. Redox Signal. 14 (10), 1829–1839. http://dx. doi.org/10.1089/ars.2010.3816. Loxton, D., Canales, J.J., 2017. Long-term cognitive, emotional and neurogenic alterations induced by alcohol and methamphetamine exposure in adolescent rats. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 74, 1–8. http://dx.doi.org/10.1016/j. pnpbp.2016.11.003. Navarro, D., Alvarado, M., Navarrete, F., Giner, M., Obregon, M.J., Manzanares, J., et al., 2015. Gestational and early postnatal hypothyroidism alters VGluT1 and VGAT bouton distribution in the neocortex and hippocampus, and behavior in rats. Front. Neuroanat. 9 (8), 9. http://dx.doi.org/10.3389/fnana.2015.00009. Rivera, H.M., Christiansen, K.J., Sullivan, E.L., 2015. The role of maternal obesity in the risk of neuropsychiatric disorders. Front. Neurosci. 9, 194. http://dx.doi.org/10. 3389/fnins.2015.00194. Sosanya, N.M., Cacheaux, L.P., Workman, E.R., Niere, F., Perrone-Bizzozero, N.I., RaabGraham, K.F., 2015. Mammalian target of rapamycin (mTOR) tagging promotes dendritic branch variability through the capture of Ca2+/Calmodulin-dependent protein kinase II α (CaMKIIα) mRNAs by the RNA-binding protein HuD. J. Biol. Chem. 290 (26), 16357–16371. http://dx.doi.org/10.1074/jbc.M114.599399. Tozuka, Y., Kumon, M., Wada, E., Onodera, M., Mochizuki, H., Wada, K., 2010. Maternal obesity impairs hippocampal BDNF production and spatial learning performance in young mouse offspring. Neurochem. Int. 57 (3), 235–347. http://dx.doi.org/10. 1016/j.neuint.2010.05.015. Verdile, G., Keane, K.N., Cruzat, V.F., Medic, S., Sabale, M., Rowles, J., et al., 2015. Inflammation and oxidative stress: the molecular connectivity between insulin resistance, obesity, and alzheimer's disease. Mediators Inflamm. 2015 (6), 105828–105917. http://dx.doi.org/10.1155/2015/105828. Vogt, M.C., Paeger, L., Hess, S., Steculorum, S.M., Awazawa, M., Hampel, B., et al., 2014. Neonatal insulin action impairs hypothalamic neurocircuit formation in response to maternal high-fat feeding. Cell 156 (3), 495–509. http://dx.doi.org/10.1016/j.cell. 2014.01.008. White, C.L., Purpera, M.N., Morrison, C.D., 2009a. Maternal obesity is necessary for programming effect of high-fat diet on offspring. AJP: Regul. Integr. Comp. Physiol. 296 (5), R1464–R1472. http://dx.doi.org/10.1152/ajpregu.91015.2008. White, C.L., Pistell, P.J., Purpera, M.N., Gupta, S., Fernandez-Kim, S.-O., Hise, T.L., et al., 2009b. Effects of high fat diet on Morris maze performance, oxidative stress, and inflammation in rats: contributions of maternal diet. Neurobiol. Dis. 35 (1), 3–13. http://dx.doi.org/10.1016/j.nbd.2009.04.002. la Torre-Ubieta, de, L., Bonni, A., 2011. Transcriptional regulation of neuronal polarity and morphogenesis in the mammalian brain. Neuron 72 (1), 22–40. http://dx.doi. org/10.1016/j.neuron.2011.09.018.
5. Conclusions To conclude, we suggest that a disturbed early metabolic environment results in long-term hippocampal insulin resistance, reduced mTor activity, and deteriorative effects on multiple mechanisms of neuronal network plasticity. Restoring hippocampal insulin signaling – possibly via targeting mTor – might thus be key to the prevention and therapy of behavioral or cognitive impairments following maternal obesity. Acknowledgements We thank Professor Axel Methner at the Johannes Gutenberg University Mainz for donating the HT22 cell line. This work was supported by the Deutsche Forschungsgemeinschaft (DGF) (RO 4109/2-1 to EHR). Each author has made an important scientific contribution to the study: LS, RK, IJ, RE, JD, and EHR participated in the study design. Animal experiments were carried out by IBG, RJ, SA, AM, and MH. Molecular analyses and cell culture experiments were performed by LS, RK, and CV. LS, RK, RJ, AM, and EHR analyzed and interpreted the data. EHR is the guarantor of this work. All authors acknowledge that no conflicts of interest exist. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.psyneuen.2017.12.023. References Bae-Gartz, I., Janoschek, R., Kloppe, C.-S., Vohlen, C., Roels, F., Oberth R, A., et al., 2016. Running exercise in obese pregnancies prevents IL-6 trans-signaling in male offspring. Med. Sci. Sports Exerc. 48 (5), 829–838. http://dx.doi.org/10.1249/MSS. 0000000000000835. Bergeron, Y., Chagniel, L., Bureau, G., Massicotte, G., Cyr, M., 2014. mTOR signaling contributes to motor skill learning in mice. Front. Mol. Neurosci. 7 (e35885), 26. http://dx.doi.org/10.3389/fnmol.2014.00026. Bilbo, S.D., Tsang, V., 2010. Enduring consequences of maternal obesity for brain inflammation and behavior of offspring. FASEB J. 24 (6), 2104–2115. http://dx.doi. org/10.1096/fj. 09-144014. Bogen, I.L., Boulland, J.-L., Mariussen, E., Wright, M.S., Fonnum, F., Kao, H.-T., et al., 2006. Absence of synapsin I and II is accompanied by decreases in vesicular transport of specific neurotransmitters. J. Neurochem. 96 (5), 1458–1466. http://dx.doi.org/ 10.1111/j.1471-4159.2005.03636.x. Cesca, F., Baldelli, P., Valtorta, F., Benfenati, F., 2010. The synapsins: key actors of synapse function and plasticity. Prog. Neurobiol. 91 (4), 313–348. http://dx.doi.org/ 10.1016/j.pneurobio.2010.04.006. Cordner, Z.A., Tamashiro, K.L.K., 2015. Effects of high-fat diet exposure on learning & memory. Physiol. Behav. 152 (Pt. B), 363–371. http://dx.doi.org/10.1016/j.physbeh.
52
Contents lists available at ScienceDirect
Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen
Hippocampal insulin resistance links maternal obesity with impaired neuronal plasticity in adult offspring
T
Lisa Schmitza,1, Rebecca Kuglina,1, Inga Bae-Gartza, Ruth Janoscheka, Sarah Appela, Andrea Mesarosb, Igor Jakovcevskic,d, Christina Vohlena, Marion Handwerka, Regina Ensenauere, ⁎ Jörg Dötscha, Eva Hucklenbruch-Rothera, a
Department of Pediatrics and Adolescent Medicine, University Hospital of Cologne, Kerpener Str. 62, Cologne, 50924, Germany Max Planck Institute for Biology of Ageing, Phenotyping Core Facility, Joseph-Stelzmann-Str. 9b, Cologne, 50931, Germany c Institute for Molecular and Behavioral Neuroscience, Center for Molecular Medicine Cologne, Cologne, Germany d Experimental Neurophysiology, German Center for Neurodegenerative Diseases, Bonn, Germany e Experimental Pediatrics and Metabolism, University Children’s Hospital, Heinrich Heine University Düsseldorf, Düsseldorf, Germany b
A R T I C L E I N F O
A B S T R A C T
Keywords: Insulin resistance Hippocampus Maternal obesity Synapse
Objective: Maternal obesity and a disturbed metabolic environment during pregnancy and lactation have been shown to result in many long-term health consequences for the offspring. Among them, impairments in neurocognitive development and performance belong to the most dreaded ones. So far, very few mechanistic approaches have aimed to determine the responsible molecular events. Methods: In a mouse model of maternal diet-induced obesity and perinatal hyperinsulinemia, we assessed adult offspring’s hippocampal insulin signaling as well as concurrent effects on markers of hippocampal neurogenesis, synaptic plasticity and function using western blotting and immunohistochemistry. In search for a potential link between neuronal insulin resistance and hippocampal plasticity, we additionally quantified protein expression of key molecules of synaptic plasticity in an in vitro model of acute neuronal insulin resistance. Results: Maternal obesity and perinatal hyperinsulinemia result in adult hippocampal insulin resistance with subsequently reduced hippocampal mTor signaling and altered expression of markers of neurogenesis (doublecortin), synaptic plasticity (ampaloxO1, pSynapsin) and function (vGlut, vGAT) in the offspring. The observed effects are independent of the offspring’s adult metabolic phenotype and can be associated with multiple previously reported behavioral abnormalities. Additionally, we demonstrate that induction of insulin resistance in cultured hippocampal neurons reduces mTor signaling, doublecortin and vGAT protein expression. Conclusions: Hippocampal insulin resistance might play a key role in mediating the long-term effects of maternal obesity and perinatal hyperinsulinemia on hippocampal plasticity and the offspring’s neurocognitive outcome.
1. Introduction Obese pregnancies are on the verge of becoming the rule, not the exception in many developed countries. Only recently, neurocognitive impairments have emerged as a very unfavourable outcome in the offspring of obese pregnancies (Rivera et al., 2015). Specifically, human and animal studies have shown that maternal obesity is associated with numerous neurodevelopmental and psychiatric disorders, including intellectual deficits, anxiety, depression, attention deficit hyperactivity disorder, and autism (Edlow, 2017). Insulin has been identified as an important modulator of neuronal network development – mostly in the hypothalamus – during early phases of life (Dearden and Ozanne,
⁎
1
2015). In detail, hyperinsulinemia following maternal overnutrition has been shown to alter neuronal projections of feeding neurons located in the arcuate nucleus of the hypothalamus affecting the offspring’s pancreatic parasympathetic innervation and glucose-stimulated insulin secretion (Vogt et al., 2014). Despite the undeniable relevance of neurocognitive long-term consequences, hippocampal neurogenesis, synaptic plasticity and function have not been the major focus of mechanistic studies on maternal obesity–related offspring pathology to date (Rivera et al., 2015; Dearden and Ozanne, 2015). Previously, we have shown that maternal diet-induced obesity in the mouse results in transiently increased body weight and body fat content at the end of lactation, going along with profound
Corresponding author. E-mail address: [email protected] (E. Hucklenbruch-Rother). Authors contributed equally.
https://doi.org/10.1016/j.psyneuen.2017.12.023 Received 15 August 2017; Received in revised form 20 November 2017; Accepted 27 December 2017 0306-4530/ © 2017 Elsevier Ltd. All rights reserved.
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
software, respectively (Bruker, Belgium). Images were segmented based on tissue density for both total volume and fat volume. Total fat volume was further segmented into visceral and subcutaneous fat using the abdominal muscular wall as orientation.
hyperinsulinemia and impaired glucose tolerance at postnatal day (P) 21 (Janoschek et al., 2016; Bae-Gartz et al., 2016). After being weaned to standard chow, however, 70 day old offspring of obese mouse dams appear to be metabolically indistinguishable from control offspring (Janoschek et al., 2016; Bae-Gartz et al., 2016). In the current study, we aimed to determine the long-term effects of maternal obesity and the consecutive major changes in early metabolic environment on hippocampal insulin sensitivity and synaptic plasticity and function in the offspring at P70. Since insulin’s intracellular signaling cascade involves activation of mTor (mammalian target of rapamycin), a well characterized serine/threonine kinase which has recently gained attention due to its involvement in neurocognitive function (Graber et al., 2013; Bergeron et al., 2014) and synaptic plasticity (Sosanya et al., 2015; Dwyer et al., 2015), we additionally set out to examine whether reduced mTor activation in the hippocampus might link hippocampal insulin resistance with altered plasticity of hippocampal neuronal networks in adult offspring of obese mothers.
2.4. Cell culture experiments HT22 cells were plated with a cell count of 10^6 per 10 cm dish and maintained in their feeding medium (DMEM (Gibco, 41966-029) + 10% FBS (Biochrom, S0615) + 1% Pen/Strep (Sigma Aldrich, P4458)) for 16 hrs. For 24hrs-induction of insulin resistance (Fig. 4 a, c, Suppl. Fig. 2 a, b), cells were exposed to fresh feeding medium either with (IR) or without (CO) 20 nM insulin for 24 hrs as described before (Kim et al., 2011). Following induction of insulin resistance, cells were washed with PBS and either acutely stimulated with fresh feeding medium containing 20 nM insulin for another 15 min or not. Cells were washed with PBS, harvested, and frozen at −80 °C for protein extraction and Western blotting. For 72hrs-induction of insulin resistance (Fig. 4 a, b, Suppl. Fig. 2 c), cells were exposed to fresh feeding medium either with (IR) or without (CO) 20 nM insulin every 24 hrs for 72 hrs. After 72 hrs, cells were washed with PBS and either acutely stimulated with fresh feeding medium containing 20 nM insulin for another 15 min or not. Cells were again washed with PBS, harvested, and frozen at −80 °C for protein extraction and Western blotting.
2. Material and methods 2.1. Experimental design and animal procedures All animal procedures were conducted in compliance with protocols approved by local government authorities (Land NRW) and were in accordance with National Institutes of Health guidelines. Mice (C57BL/ 6N) were bred locally at a designated animal unit of the University Hospital of Cologne (Cologne, Germany). For detailed information on experimental design, animal housing, diet content, and mating scheme, see (Janoschek et al., 2016; Bae-Gartz et al., 2016). Briefly, control (CO) females, receiving standard chow at all times, and high fat diet (HFD)-fed females, receiving high fat diet starting upon their weaning at three weeks of age for the rest of the experiment, were mated at 12–14 weeks of age. All offspring studies were performed using male offspring. On P3, litter size was adjusted to six for each litter. At P21, in different subsets of only one animal per litter, either blood samples were taken, glucose tolerance testing (GTT) was performed, or animals underwent micro computed tomography (μCT) measurement. At P70, again, in subsets of one animal per litter, either blood samples were taken, GTT was performed, or animals were sacrificed. Hippocampus from both hemispheres was rapidly dissected and stored at −80 °C. Also, liver, skeletal muscle, and white adipose tissue were dissected, snap-frozen, and stored at −80 °C. The brain of one animal per litter was paraformaldehyde (PFA) fixated, mounted in tissue freezing medium, and stored at −80 °C for immunohistochemistry.
2.5. Western blotting Frozen tissue or cells were homogenized in lysis buffer as previously described (Bae-Gartz et al., 2016). Protein concentration was determined with a BCA-Protein Assay Kit (Thermo Scientific, Waltham, USA). Lysates resolved on a 10% reducing SDS-PAGE gel were transferred to a PVDF membrane. Blots were probed with the antibodies (Suppl. Table 1). 2.6. Immunohistochemistry For immunohistochemistry, 20 μm thick coronal cryosections were washed in PBS and permeabilized with 0.3% Triton-X-100 (SigmaAldrich, T8532) in PBS for 30 min. After blocking of non-specific binding components with Sea Block blocking Buffer (ThermoFisher, 37527) for 2 h, samples were incubated with primary antibodies (Suppl. Table 1) dissolved in antibody diluent (Dako, #S202230). After an overnight incubation at 4 °C and intensive washing in PBS, samples were incubated with secondary antibodies at room temperature for 2 hrs. Slices were rinsed in PBS and cover-slipped in Fluoroshield Mounting Medium containing 4′,6-Diamidin-2-phenylindol (DAPI) (Abcam. ab104139). For morphometric analysis, fluorescence images were taken by Olympus BX43F with cellSens Dimension software (DP80 dual CCD Camera, cellSens Dimension (V1.8)) and analyzed with the aid of ImageJ software supplemented with Olympus viewer plugin. All analyses are performed in a double-blind setting by at least two independent researchers.
2.2. Intraperitoneal glucose tolerance test and analytical procedures Glucose tolerance tests were performed as previously described (Janoschek et al., 2016; Bae-Gartz et al., 2016). Briefly, after overnight fasting (16 hrs), each animal received an intraperitoneal (ip) injection of 2 g glucose/kg body weight. Blood glucose levels were measured before glucose injection and after 15, 30, 60, and 120 min using an automatic glucose monitor (GlucoMen; A. Menarini Diagnostics, Berlin, Germany). Serum levels of insulin were measured by ELISA using mouse standards according to the manufacturer's guidelines (mouse insulin ELISA (EZRMI–13 K); Millipore CorpBillerica, MA).
2.7. Statistics Values are reported as mean +/− SD. All statistical analysis was performed in GraphPad Prism 6 software. After testing for normality, we performed an unpaired t-test or a Mann-Whitney t-test (for nonparametric distribution). Significance was set at p < 0.05.
2.3. Quantification of fat by μCT (micro computed tomography) Whole mice were scanned post mortem with a μCT scanner (SkyScan 1176, Bruker, Belgium) with an isotropic voxel size of 35.26 μm3. The x-ray settings for each scan were 45 kV and 475 μA using a 0.5 mm aluminum filter. All scans were performed over 360 ° with a rotation step of 0.6 ° and a frame averaging of 2. Images were reconstructed, analyzed and visualized using NRecon, CTAn and CTVox
3. Results 3.1. Offspring metabolic phenotype at P21 and P70 Offspring of obese mouse dams show increased body weight, body fat content, serum insulin levels, and significantly increased 15 min 47
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 1. Offspring metabolic phenotype at P21 and P70. (a) Representative μCT scan of a control (CO) offspring at P21 for quantification of subcutaneous (yellow) and visceral (red) fat. (b) Representative μCT scan of a maternal obesity (HFD) offspring at P21 for quantification of subcutaneous (yellow) and visceral (red) fat. (c) Summary of phenotypical characteristics of CO and HFD offspring at P21 (as previously reported in [5,6]); displayed values are means +/− SD; body weight: CO n = 87, HFD n = 58, p < 0.0001; serum insulin: CO n = 21, HFD n = 15, p < 0.0001; GTT and AUC: CO n = 12, HFD n = 9, P < 0.05 at 15 min. (d) Summary of phenotypical characteristics of CO and HFD offspring at P70 (as previously reported in [5,6]); displayed values are means +/− SD; body weight: CO n = 66, HFD n = 35, p = 0.6105; serum insulin: CO n = 10, HFD n = 10, p = 0,1607; GTT and AUC: CO n = 24, HFD n = 22.
48
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 2. Hippocampal insulin signaling at P70. (a) Western blots of the insulin receptor (InsR), phosphorylated and total protein kinase B (pAKT, AKT), forkhead box protein (FoxO) 1, phosphorylated and total mammalian target of rapamycin (pmTor, mTor), phosphorylated and total ribosomal protein s6 kinase beta-1 (pp70s6 K, p70s6 K), phosphorylated and total 4E-binging protein 1 (p4E-BP1, 4E-BP1), and their respective loading controls (glyceraldehyde 3-phosphoate dehydrogenase (GAPDH)) in CO and HFD offspring at P70. (b) Densitometrical quantification of western blots displayed in a). Relative protein expression normalized to GAPDH in HFD offspring compared to CO with CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. * p < 0.05, ** p < 0.01. (c) Simplified schematic overview of the insulin signaling cascade summarizing the effects of maternal obesity on hippocampal protein expression at P70. Arrows indicate significant downregulation of InsR, pmTor, pp70s6 K, and FoxO1 in the offspring’s hippocampus following maternal obesity.
neuronal network plasticity, we first determined protein expression of doublecortin (DCX), a specific marker for new-born neurons, in the hippocampus of offspring of lean and obese mouse dams at P70. DCX protein expression in total hippocampus tissue homogenate was reduced to roughly 30% of control levels following maternal obesity (Fig. 3a and b). Using the same method, we additionally quantified phosphorylation of Synapsin, which is known to facilitate the transport of neurotransmitter-filled vesicles to the synapse during an action potential and thereby promotes synaptic plasticity. We detected a significant 70% reduction in Synapsin phosphorylation in offspring of obese mouse dams at P70 (Fig. 3a and c). As an indicator of neuronal activation and indirect measure for synaptic function we next visualized phosphorylation of CREB (cAMP response element-binding protein) using immunohistochemistry. Despite visible reduction in immunoreactivity in the DG, statistical testing failed to yield significance (p = 0.073) in our analyses of 3 vs. 4 animals (Fig. 3d, k, for orientation see h). To characterize any qualitative differences in hippocampal synaptic input organization, we additionally visualized and quantified glutamatergic (excitatory) and GABAergic (inhibitory) nerve terminals by immunohistochemistry. For quantification of excitatory synapses, we first quantified immunoreactivity for the total and phosphorylated glutamate receptor subunit GluR1 at P70. We detected only a slight reduction in total hippocampal GluR1 immunoreactivity in offspring of obese dams (Fig. 3e, I, for orientation see h), but no significant differences for GluR1 in specific sub-regions or for phosphorylated GluR1 (Fig. 3e, i, l, for orientation see h). Additionally, we determined immunoreactivity for the vesicular glutamate transporter 2 (vGlut2) and detected significant local reduction in immunoreactivity in the Cornu
peak in the GTT at P21 (Fig. 1a–c) (Janoschek et al., 2016). AUC of the GTT at P21 shows an almost 20% increase in HFD offspring (p = 0.065). AT P70, however, all of these parameters have returned to control levels (Fig. 1d). 3.2. Hippocampal insulin signaling at P70 To determine the long-term effects of maternal obesity on hippocampal sensitivity to insulin, we first quantified protein expression levels of insulin’s specific receptor (InsR) in the hippocampus of adult offspring of lean and obese dams at P70 and found a significant 30% reduction (Fig. 2a and b). Next, we assessed intracellular AKT/mTor signaling by quantifying phosphorylation of AKT, mTor, and several of their downstream targets in the hippocampus at P70. We found a statistically significant reduction in activation of mTor and p70S6 K while phosphorylation of AKT and 4EBP1 only showed a tendency towards reduction in HFD offspring (Fig. 2a and b). Furthermore, protein expression of FoxO1, an important transcription factor expressed by developing hippocampal neurons in the granule cell layer of the dentate gyrus and responsible for intact neuronal polarity and axon outgrowth (la Torre-Ubieta et al., 2011) was found to be significantly reduced by more than 50% (Fig. 2a and b). At the same time, protein expression of InsR and phosphorylation of AKT was found to be unaltered in skeletal muscle and white adipose tissue, and even significantly increased in the liver of HFD offspring (Suppl. Fig. 1 a). 3.3. Hippocampal neurogenesis, synaptic plasticity and function at P70 To evaluate the effects of maternal obesity on hippocampal 49
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 3. Hippocampal neurogenesis, synaptic plasticity and function at P70. (a) Western blots of doublecortin (DCX) and phosphorylated and total Synapsin (pSyn, Syn), and their respective loading controls (GAPDH) in CO and HTFD offspring at P70. (b) Densitometrical quantification of the DCX western blot displayed in a). Relative protein expression normalized to GAPDH, CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. ** p < 0.01. (c) Densitometrical quantification of the pSyn/Syn western blot displayed in a). Relative protein expression of pSyn/Syn, CO set to 1. Displayed values are means +/− SD; CO = control, n = 5, HFD = maternal obesity, n = 5. ** p < 0.01. (d) Representative immunohistochemical co-staining for the phosphorylated cAMP response element binding protein (pCreb) and the neuron marker NeuN in the dentate gyrus (DG) of CO (upper panel) and HFD (lower panel) offspring at P70. Neuron marker NeuN (green), pCreb (red), Nuclear stain Dapi (blue). (e) Representative immunohistochemical co-staining for the glutamate receptor 1 (GluR1) and the neuron marker NeuN in the dentate gyrus (DG) of CO (upper 4 panels) and HFD (lower 4 panels) offspring at P70. Each set of 4 shows the nuclear staining alone (Dapi, top left), GluR1 alone (top right), neuron marker alone (NeuN, bottom left), and a merge of all three stainings (bottom, right). Neuron marker NeuN (green), GluR1 (red), nuclear stain Dapi (blue). (f) Representative immunohistochemical staining for the vesicular GABA transporter (vGAT (red)) in the CA1 region of CO (upper panel) and HFD (lower panel) offspring at P70. (g) Representative immunohistochemical staining for the vesicular glutamate transporter 2 (vGlut2 (green)) in the CA 2 region of CO (upper panel) and HFD (lower panel) offspring at P70. (h) Overwiew of hippocampal structures in coronal brain slices. White boxes indicate the subregions displayed in d, f, and g, respectively. Scale bar = 300 μm. (i) Quantification of immunoreactivity for GluR1 (calculated as raw intensity/area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 5, HFD n = 6. * p < 0.05. (j) Quantification of immunoreactivity for vGAT (calculated as vGAT positive area/total area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 4, HFD n = 7. * p < 0.05. (k) Quantification of immunoreactivity for pCreb (calculated as pCreb positive area/NeuN positive area) in total hippocampus and the DG. Displayed values are means +/− SD; CO n = 4, HFD n = 7. (l) Quantification of immunoreactivity (calculated as raw intensity/area) for phosphorylated GluR1 (pGluR1) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/− SD; CO n = 5, HFD n = 6. * p < 0.05. (m) Quantification of immunoreactivity for vGlut2 (calculated as raw intensity/area) in total hippocampus and four subregions (CA1, CA2,CA3, and DG). Displayed values are means +/ − SD; CO n = 4, HFD n = 7. * p < 0.05.
extrahippocampal brain regions within the same sections were also analyzed. In the cortex of HFD offspring at P70, we found a significant reduction in pCreb expression along with significantly reduced immunoreactivity for vGlut2. vGAT, GluR1, and pGluR1 immunoreactivity was not significantly altered (Suppl. Fig. 1 b). Within the amygdala, only vGAT, GluR1, and pGluR1 immunoreactivity was quanitifiable. No significant changes in immunoreactivity were
Ammonis (CA) 3 region and the Dentate Gyrus (DG) of the hippocampus following maternal obesity (Fig. 3g, m, for orientation see h). For the vesicular GABA transporter (vGAT), a marker for inhibitory nerve terminals, we detected significantly lower immunoreactivity in total hippocampus and all sub-regions except CA3 (Fig. 3f, j, for orientation see h). To evaluate specificity of the hippocampal findings, 50
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
Fig. 4. In vitro effect of insulin resistance on neuronal plasticity. (a) Quantification of protein expression in insulin resistant (IR) HT22 cells relative to control (CO) cells in vitro (0 equals mean of controls and negative values represent downregulation in IR cells (e.g. −0.5 = 50% reduction compared to controls)). Relative protein expression of phosphorylated protein kinase B (pAKT), phosphorylated mammalian target of rapamycin (pmTor), doublecortin (DCX), the vesicular glutamate transporter 2 (vGlut2), and the vesicular GABA transporter (vGAT) was normalized to GAPDH. Lower panel (light grey): 24 hrs induction of insulin resistance. Upper panel (dark grey): 72 hrs induction of insulin resistance. (b) Exemplary bands from Western blots of pAKT, pmTor, DCX, vGlut2, and vGAT of IR und CO cells 72 hrs upon induction of insulin resistance (original blots displayed in Suppl. Fig. 2 c). (c) Exemplary bands from Western blots of pAKT, pmTor, DCX, vGlut2, and vGAT of IR und CO cells 24 hrs upon induction of insulin resistance (original blots displayed in Suppl. Fig. 2 b).
reported by ourselves and others (Janoschek et al., 2016; Bae-Gartz et al., 2016; White et al., 2009a), hippocampal insulin resistance as a potential underlying mechanism for impaired neurocognitive development and outcome has to our knowledge so far not been investigated. Insulin resistance is a key feature of the metabolic syndrome in mice and men. In our model, however, hippocampal insulin resistance is detected while mice lack signs of insulin resistance in liver, skeletal muscle, or white adipose tissue and do not display any increase in body weight, body fat content, or serum hormone levels. This is intriguing and implies that a constitutional difference in hippocampal hormone receptor abundance might be a consequence of transient disturbances, such as increased body weight, body fat content, or high levels of circulating hormones and cytokines during vulnerable developmental stages (Janoschek et al., 2016; Bae-Gartz et al., 2016). Independent of its origin, the clear reduction in hippocampal insulin action is inevitably linked with chronic inflammatory processes and oxidative stress – two very important contributors to cognitive decline (for review see (Verdile et al., 2015)). And indeed, multiple studies already showed that maternal obesity is associated with increased oxidative stress and inflammation in the offspring’s hippocampus resulting in impaired hippocampus-dependent behavior (Edlow, 2017; Tozuka et al., 2010; Bilbo and Tsang, 2010; White et al., 2009b). A connection between oxidative stress, inflammation and insulin resistance in the hippocampus of offspring of obese mothers is thus highly likely. Yet, the question of cause or consequence remains unanswered. With regard to neurogenesis, the interpretation of our data lies at hand: hippocampal neurogenesis, marked by DCX, is known to be clearly reduced in diabetic animals and can be restored by treatment with metformin, a wellstudied insulin sensitizer (Hwang et al., 2010). Reduced hippocampal neurogenesis, on the other hand, strongly correlates with reduced hippocampus-dependent cognition (Loxton and Canales, 2017). Thus, a clear reduction in hippocampal DCX expression in our model might at least partly result from local insulin resistance and gate way to cognitive and behavioral abnormalities. Grillo et al. provided first evidence for the role of hippocampal insulin resistance in long-term potentiation and hippocampus-dependent behavior (Grillo et al., 2015). While they reported an associated reduction in phosphorylation of the glutamate receptor, we observed reduced abundance of the total glutamate receptor without changes in its phosphorylation, with reduced hippocampal immunoreactivity of the vesicular glutamate transporter (vGlut2) in CA3 and the DG. Additionally, we see a clear reduction in vGAT immunoreactivity across most hippocampal subregions, which has interestingly previously been described as a consequence of gestational or early postnatal hypothyroidism in rats with concomitant cognitive and behavioral abnormalities (Navarro et al., 2015). Reciprocally, increased vGAT expression is
detected (Suppl. Fig. 1 c). 3.4. In vitro effects of insulin resistance on neuronal plasticity Hypothesizing a mechanistic link between insulin resistance in hippocampal neurons and neurogenesis, neuronal plasticity and function, we aimed to induce insulin resistance in an immortalized murine hippocampal cell line (HT22) and screened for changes in protein expression and activation of markers and enzymes found to be altered in our in vivo model at P70. As described in the literature, treatment of cells with insulin for 24 or 72 hrs renders them insulin resistant (IR) based on a downregulation of the insulin receptor (Suppl. Fig. 2a) with blunted AKT phosphorylation response to an acute insulin stimulus (15 min) (Kim et al., 2011) (Suppl. Fig. 2b, c). Paralleling the findings in our in vivo model, IR hippocampal neurons also showed reduced levels of phosphorylated mTor, most prominent after 72 hrs (Fig. 4a and b, Suppl. Fig. 2b, c). Furthermore, DCX protein expression appeared to be reduced in insulin resistant cells after 24 and 72 hrs, while vGlut2 expression seemed unaffected at both time points (Fig. 4a–c, Suppl. Fig. 2 b, c). A reduction was also observed for vGAT protein expression 24 hrs upon induction of insulin resistance, pointing towards a potential connection between neuronal insulin action and inhibitory synapse function in vitro. At 72 hrs, however, vGAT expression had returned to control levels. 4. Discussion Neurocognitive impairments range among the most dreaded longterm consequences for the offspring of obese mothers (Edlow, 2017). Here, we report impaired insulin signaling in the hippocampus of adult offspring of obese mouse dams paralleled with significantly altered hippocampal expression and distribution of markers of neurogenesis, synaptic plasticity and function. Additionally, we suggest a mechanistic link between neuronal insulin signaling and protein expression of synaptic transmission markers in vitro. It has long been suggested that synaptic insulin receptor signaling directly affects learning and memory (Cordner and Tamashiro, 2015). But only recently, an elegant study by Grillo et al. was able to show that site-specific hippocampal knockdown of the insulin receptor in adult rats results in changes in hippocampal plasticity and in hippocampusdependent behavior (Grillo et al., 2015). Furthermore, maternal obesity has repeatedly been linked to impaired hippocampus-dependent behavior in the offspring, including impaired hippocampal learning, decreased sociability, sex-dependent hyperactivity, and increased anxiety (for review see (Edlow, 2017)). Although impaired systemic glucose tolerance in the offspring of obese mothers has been previously 51
Psychoneuroendocrinology 89 (2018) 46–52
L. Schmitz et al.
associated with memory enhancement and anxiolysis (Jansone et al., 2016), letting a functional role of vGAT in maternal obesity-induced cognitive impairment seem quite conceivable. In vitro, we demonstrate a potential connection between neuronal insulin signaling and vGAT protein expression, supporting the in vivo results. Lastly, the strong reduction in hippocampal pSynapsin in our model might directly contribute to the reduction in vGAT expression as shown before (Bogen et al., 2006), but is of course also on its own indicative of disturbed hippocampal transmitter trafficking and impaired plasticity following maternal obesity (Cesca et al., 2010).
2015.06.008. Dearden, L., Ozanne, S.E., 2015. Early life origins of metabolic disease: developmental programming of hypothalamic pathways controlling energy homeostasis. Front. Neuroendocrinol. 39, 3–16. http://dx.doi.org/10.1016/j.yfrne.2015.08.001. Dwyer, J.M., Maldonado-Avilés, J.G., Lepack, A.E., DiLeone, R.J., DUMAN, R.S., 2015. Ribosomal protein S6 kinase 1 signaling in prefrontal cortex controls depressive behavior. Proc. Natl. Acad. Sci. U. S. A. 112 (19), 6188–6193. http://dx.doi.org/10. 1073/pnas.1505289112. Edlow, A.G., 2017. Maternal obesity and neurodevelopmental and psychiatric disorders in offspring. Prenat. Diagn. 37 (1), 95–110. http://dx.doi.org/10.1002/pd.4932. Graber, T.E., McCamphill, P.K., Sossin, W.S., 2013. A recollection of mTOR signaling in learning and memory. Learn. Mem. (Cold Spring Harbor, N.Y.) 20 (10), 518–530. http://dx.doi.org/10.1101/lm.027664.112. Grillo, C.A., Piroli, G.G., Lawrence, R.C., Wrighten, S.A., Green, A.J., Wilson, S.P., et al., 2015. Hippocampal insulin resistance impairs spatial learning and synaptic plasticity. Diabetes 64 (11), 3927–3936. http://dx.doi.org/10.2337/db15-0596. Hwang, I.K., Kim, I.Y., Joo, E.J., Shin, J.H., Choi, J.W., Won, M.-H., et al., 2010. Metformin normalizes type 2 diabetes-induced decrease in cell proliferation and neuroblast differentiation in the rat dentate gyrus. Neurochem. Res. 35 (4), 645–650. http://dx.doi.org/10.1007/s11064-009-0115-5. Janoschek, R., Bae-Gartz, I., Vohlen, C., Alcázar, M.A.A., Dinger, K., Appel, S., et al., 2016. Dietary intervention in obese dams protects male offspring from WAT induction of TRPV4, adiposity, and hyperinsulinemia. Obesity 24 (6), 1266–1273. http:// dx.doi.org/10.1002/oby.21486. Jansone, B., Kadish, I., van Groen, T., Beitnere, U., Plotniece, A., Pajuste, K., et al., 2016. Memory-enhancing and brain protein expression-stimulating effects of novel calcium antagonist in Alzheimer's disease transgenic female mice. Pharmacol. Res. 113 (Pt. B), 781–787. http://dx.doi.org/10.1016/j.phrs.2016.06.020. Kim, B., Sullivan, K.A., Backus, C., Feldman, E.L., 2011. Cortical neurons develop insulin resistance and blunted akt signaling: a potential mechanism contributing to enhanced ischemic injury in diabetes. Antioxid. Redox Signal. 14 (10), 1829–1839. http://dx. doi.org/10.1089/ars.2010.3816. Loxton, D., Canales, J.J., 2017. Long-term cognitive, emotional and neurogenic alterations induced by alcohol and methamphetamine exposure in adolescent rats. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 74, 1–8. http://dx.doi.org/10.1016/j. pnpbp.2016.11.003. Navarro, D., Alvarado, M., Navarrete, F., Giner, M., Obregon, M.J., Manzanares, J., et al., 2015. Gestational and early postnatal hypothyroidism alters VGluT1 and VGAT bouton distribution in the neocortex and hippocampus, and behavior in rats. Front. Neuroanat. 9 (8), 9. http://dx.doi.org/10.3389/fnana.2015.00009. Rivera, H.M., Christiansen, K.J., Sullivan, E.L., 2015. The role of maternal obesity in the risk of neuropsychiatric disorders. Front. Neurosci. 9, 194. http://dx.doi.org/10. 3389/fnins.2015.00194. Sosanya, N.M., Cacheaux, L.P., Workman, E.R., Niere, F., Perrone-Bizzozero, N.I., RaabGraham, K.F., 2015. Mammalian target of rapamycin (mTOR) tagging promotes dendritic branch variability through the capture of Ca2+/Calmodulin-dependent protein kinase II α (CaMKIIα) mRNAs by the RNA-binding protein HuD. J. Biol. Chem. 290 (26), 16357–16371. http://dx.doi.org/10.1074/jbc.M114.599399. Tozuka, Y., Kumon, M., Wada, E., Onodera, M., Mochizuki, H., Wada, K., 2010. Maternal obesity impairs hippocampal BDNF production and spatial learning performance in young mouse offspring. Neurochem. Int. 57 (3), 235–347. http://dx.doi.org/10. 1016/j.neuint.2010.05.015. Verdile, G., Keane, K.N., Cruzat, V.F., Medic, S., Sabale, M., Rowles, J., et al., 2015. Inflammation and oxidative stress: the molecular connectivity between insulin resistance, obesity, and alzheimer's disease. Mediators Inflamm. 2015 (6), 105828–105917. http://dx.doi.org/10.1155/2015/105828. Vogt, M.C., Paeger, L., Hess, S., Steculorum, S.M., Awazawa, M., Hampel, B., et al., 2014. Neonatal insulin action impairs hypothalamic neurocircuit formation in response to maternal high-fat feeding. Cell 156 (3), 495–509. http://dx.doi.org/10.1016/j.cell. 2014.01.008. White, C.L., Purpera, M.N., Morrison, C.D., 2009a. Maternal obesity is necessary for programming effect of high-fat diet on offspring. AJP: Regul. Integr. Comp. Physiol. 296 (5), R1464–R1472. http://dx.doi.org/10.1152/ajpregu.91015.2008. White, C.L., Pistell, P.J., Purpera, M.N., Gupta, S., Fernandez-Kim, S.-O., Hise, T.L., et al., 2009b. Effects of high fat diet on Morris maze performance, oxidative stress, and inflammation in rats: contributions of maternal diet. Neurobiol. Dis. 35 (1), 3–13. http://dx.doi.org/10.1016/j.nbd.2009.04.002. la Torre-Ubieta, de, L., Bonni, A., 2011. Transcriptional regulation of neuronal polarity and morphogenesis in the mammalian brain. Neuron 72 (1), 22–40. http://dx.doi. org/10.1016/j.neuron.2011.09.018.
5. Conclusions To conclude, we suggest that a disturbed early metabolic environment results in long-term hippocampal insulin resistance, reduced mTor activity, and deteriorative effects on multiple mechanisms of neuronal network plasticity. Restoring hippocampal insulin signaling – possibly via targeting mTor – might thus be key to the prevention and therapy of behavioral or cognitive impairments following maternal obesity. Acknowledgements We thank Professor Axel Methner at the Johannes Gutenberg University Mainz for donating the HT22 cell line. This work was supported by the Deutsche Forschungsgemeinschaft (DGF) (RO 4109/2-1 to EHR). Each author has made an important scientific contribution to the study: LS, RK, IJ, RE, JD, and EHR participated in the study design. Animal experiments were carried out by IBG, RJ, SA, AM, and MH. Molecular analyses and cell culture experiments were performed by LS, RK, and CV. LS, RK, RJ, AM, and EHR analyzed and interpreted the data. EHR is the guarantor of this work. All authors acknowledge that no conflicts of interest exist. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.psyneuen.2017.12.023. References Bae-Gartz, I., Janoschek, R., Kloppe, C.-S., Vohlen, C., Roels, F., Oberth R, A., et al., 2016. Running exercise in obese pregnancies prevents IL-6 trans-signaling in male offspring. Med. Sci. Sports Exerc. 48 (5), 829–838. http://dx.doi.org/10.1249/MSS. 0000000000000835. Bergeron, Y., Chagniel, L., Bureau, G., Massicotte, G., Cyr, M., 2014. mTOR signaling contributes to motor skill learning in mice. Front. Mol. Neurosci. 7 (e35885), 26. http://dx.doi.org/10.3389/fnmol.2014.00026. Bilbo, S.D., Tsang, V., 2010. Enduring consequences of maternal obesity for brain inflammation and behavior of offspring. FASEB J. 24 (6), 2104–2115. http://dx.doi. org/10.1096/fj. 09-144014. Bogen, I.L., Boulland, J.-L., Mariussen, E., Wright, M.S., Fonnum, F., Kao, H.-T., et al., 2006. Absence of synapsin I and II is accompanied by decreases in vesicular transport of specific neurotransmitters. J. Neurochem. 96 (5), 1458–1466. http://dx.doi.org/ 10.1111/j.1471-4159.2005.03636.x. Cesca, F., Baldelli, P., Valtorta, F., Benfenati, F., 2010. The synapsins: key actors of synapse function and plasticity. Prog. Neurobiol. 91 (4), 313–348. http://dx.doi.org/ 10.1016/j.pneurobio.2010.04.006. Cordner, Z.A., Tamashiro, K.L.K., 2015. Effects of high-fat diet exposure on learning & memory. Physiol. Behav. 152 (Pt. B), 363–371. http://dx.doi.org/10.1016/j.physbeh.
52