Effects of Acute versus Recurrent Insulin-Induced Hypoglycemia on Ventromedial Hypothalamic Nucleus Metabolic-Sensory Neuron AMPK Activity: Impact of Alpha₁-Adrenergic Receptor Signaling

Abstract

Mechanisms that underlie metabolic sensor acclimation to recurring insulin-induced hypoglycemia (RIIH) are unclear. Norepinephrine (NE) regulates ventromedial hypothalamic nucleus (VMN) gluco-stimulatory nitric oxide (NO) and gluco-inhibitory γ-aminobutryic acid (GABA) neuron signaling. Current research addressed the hypothesis that during RIIH, NE suppresses 5’-AMP-activated protein kinase (AMPK) reactivity in both populations and impedes counter-regulation. The brain is postulated to utilize non-glucose substrates, e.g. amino acids glutamine (Gln), glutamate (Gln), and aspartate (Asp), to produce energy during hypoglycemia. A correlated aim investigated whether NE controls pyruvate recycling pathway marker protein (glutaminase, GLT; malic enzyme, ME-1) expression in either metabolic-sensory cell population. Male rats were injected subcutaneously with vehicle or insulin on days 1–3, then pretreated on day 4 by intracerebroventricular delivery of the alpha1-adrenergic receptor (α₁-AR) reverse-agonist prazocin (PRZ) or vehicle before final insulin therapy. PRZ prevented acute hypoglycemic augmentation of AMPK activation in each cell group. Antecedent hypoglycemic repression of sensor activity was reversed by PRZ in GABA neurons. During RIIH, nitrergic neurons exhibited α₁-AR – dependent up-regulated GLT and α₂-AR profiles, while GABA cells showed down-regulated α₁-AR. UHPLC-MS/ESI analysis documented a decline in VMN Glu, Gln, and Asp concentrations during acute hypoglycemia, and habituation of former two profiles to RIIH. PRZ attenuated glucagon and corticosterone secretion during acute hypoglycemia, but reversed decrements in output of both hormones during RIIH. Results implicate adjustments in impact of α₁-AR signaling in repressed VMN metabolic-sensory AMPK activation and counter-regulatory dysfunction during RIIH. Antecedent hypoglycemia may up-regulate NO neuron energy yield via α₁-AR – mediated up-regulated pyruvate recycling.

Introduction:

Insulin-induced hypoglycemia (IIH) is an incessant, worrisome complication of requisite strict therapeutic management of type I diabetes mellitus. IIH can cause neurological dysfunction and injury as decrements in glucose supply, the primary fuel source to the brain, impair vital high energy-demand nerve cell functions. Hypoglycemia-associated neuro-glucopenia triggers counteractive autonomic, neuroendocrine, and behavioral responses that boost systemic glucose availability. The ventromedial hypothalamic nucleus (VMN), a component of the mediobasal hypothalamus (MBH) and key element of the brain gluco-regulatory network, processes nutrient, endocrine, and neurochemical indicators of metabolic status to shape glucose counter-regulation [Watts and Donovan, 2010, 2014]. Detection of neuro-energetic sequelae of hypoglycemia within the MBH is obligatory for optimal counter-regulatory endocrine and gluconeogenic outflow [Borg et al., 1997; 2003]. MBH metabolic-sensory neurons produce a dynamic cellular energy readout by increasing (‘glucose-inhibited’; GI) or decreasing (‘glucose-excited’; GE) synaptic firing as ambient glucose levels fall [Oomura et al., 1969; Ashford et al., 1990; Silver and Erecinska, 1998]. Within the MBH, activation of the energy gauge 5’-AMP-activated protein kinase (AMPK) elicits production of the gaseous neurotransmitter nitric oxide (NO), which is critical for glucoprivic stimulation of GI neuron firing and optimum in vivo counter-regulatory hormone release [Routh et al., 2014]. Counter-regulatory outflow during hypoglycemia is also amplified by repression of the MBH gluco-inhibitory signal γ-aminobutyric acid (GABA) [Chan et al., 2006; 2008; Zhu et al., 2010].

Recurring insulin-induced hypoglycemia (RIIH) can trigger hypoglycemia-associated autonomic failure, a pathophysiological mal-adaptation that manifests as diminished hypoglycemic awareness and defective glucose counter-regulation [Cryer et al., 2003; Cryer, 2010]. Experimental models for RIIH that replicate insulin delivery route, frequency of administration, and duration of action in the clinical setting show blunted neuron genomic activation in brain gluco-regulatory loci, inferring neurological desensitization to hypoglycemia [Paranjape and Briski, 2005; Kale et al., 2006]. Sensory, integrative, and motor elements of the brain gluco-regulatory circuitry are conceivably capable of habituation to RIIH, but available studies point to liable development of malfunctioning metabolic monitoring [Smith and Amiel, 2002]. In the VMN, expression of the NO marker protein neuronal nitric oxide synthase (nNOS) is suppressed, whereas glutamate decarboxylase_65/67 (GAD_65/67) is up-regulated by antecedent hypoglycemia [Mandal et al., 2017], outcomes that align with prior reports of RIIH-associated repression of nNOS activity [Fioramonti et al., 2011] and up-regulation of GABA release within the broader MBH [Chan et al., 2008].

The catecholamine neurotransmitter norepinephrine (NE) links hindbrain energy sensors with the VMN and other downstream gluco-regulatory loci [Shrestha and Briski, 2014]. Glucoprivic augmentation of local NE activity promotes GABA signaling [Beverly et al., 2000; 2001]. VMN NO and GABA neurons are likely direct targets for NE action by virtue of express alpha₁-adrenergic receptor (α₁-AR), alpha2-adrenergic receptor (α2-AR), and beta1-adrenergic receptor (β1-AR) protein expression [Ibrahim et al., 2018]. VMN nNOS profiles are decreased by exogenous NE administration; taken together with evidence for noradrenergic stimulation of GABA, these findings imply that noradrenergic input singularly enhances energy stability [Ibrahim et al., 2018; Mahmood et al., 2019]. Current research utilized high spatial-resolution microdissection/high-sensitivity molecular analytical techniques along with a validated RIIH model [Paranjape and Briski, 2005] to address the hypothesis that NE regulates VMN NO and GABA neuron AMPK activity and marker protein expression during IIH by α₁-AR signaling, and that this receptor mediates accommodated metabolo-sensory neuron responses to RIIH. Here, male rats were pretreated by intracerebroventricular (icv) administration of the α₁-AR reverse-agonist prazocin (PRZ) or vehicle before exposure to a single versus final episode of hypoglycemia. VMN neurons were identified by immunocytochemical labeling as nNOS-or GAD_65/67-immunoreactive (-ir)-positive prior to laser-catapult microdissection for Western blot analysis of nNOS or GAD profiles as well as AMPK, phosphoAMPK (pAMPK), α₁-AR, α₂-AR, and β₁-AR protein expression.

There is new focus on the prospect that brain neurons adjust to hypoglycemia by utilizing alternative energy fuels, and that such reliance may, in turn, impede detection and/or signaling of sequential neuro-glucopenic episodes. Glutamine (Gln) and glutamate (Glu) are potential cerebral energy substrates during glucose deficiency as brain levels of these amino acids decline during IIH [Behar et al., 1985]. Glutaminolysis is an energy-yielding pathway that utilizes glutamine metabolism via the pyruvate recycling pathway [Cerdan, 2017], via conversion of Gln to Glu to support tricarboxylic acid (TCA) cycle function. Complete glutamine oxidation in the TCA cycle involves entrance and exit of glutamate in the form of alpha-ketoglutarate and malate, respectively, followed by processing of pyruvate to acetyl-CoA. The amino acid aspartate (Asp) is an alternative source of substrate for pyruvate recycling as aspartate transaminase catalyzes conversion of Asp and α-ketoglutarate to Glu and the TCA cycle component oxaloacetate. Glucoprivation is postulated to up-regulate nerve cell pyruvate recycling pathway activity [Amaral, 2013]. Current studies examined the premise that α₁-AR mediate differential effects of acute and chronic hypoglycemia on expression profiles of the rate-limiting pyruvate recycling pathway enzymes glutaminase (GLT) and NADP-dependent malic enzyme-1 (ME-1) in VMN nitrergic and/or GABAergic neurons. In addition, VMN tissue samples obtained by micropunch dissection were analyzed by HPLC-electrospray ionization mass spectrometry (LC/MS-ESI) to evaluate effects of acute versus recurring hypoglycemia on Gln, Glu, Asp, and GABA amino acid content.

Methods and Materials:

Experimental Design:

Adult male Sprague-Dawley rats (3–4 months of age) were maintained in groups (2–3 animals per cage) under a 14-h light/10-h dark lighting schedule (light on at 05:00h), and allowed ad-libitum access to standard laboratory rat chow (Harlan Teklad LM-485; Harlan Industries, Madison, WI, USA) and tap water. Animals were accustomed to daily handling for a minimum of one week prior to experimentation. All protocols were conducted in accordance with NIH guidelines for care and use of laboratory animals, under ULM Institutional Animal Care and Use Committee approval. On day 1, animals were anesthetized by subcutaneous (sc) injection of ketamine/xylazine (0.1 mL/100 g bw; 90 mg ketamine: 10 mg xylazine/mL; Henry Schein Inc., Melville, NY), then implanted with a left lateral ventricular (LV) PE-20 cannula [Singh and Briski, 2005] at the following coordinates: 0.0 mm posterior to bregma; 1.5 mm lateral to bregma; 5.0 mm ventral to skull surface. After surgery, rats were injected with enrofloxacin (baytril 2.27%, 10 mg/kg; intramuscular) and ketoprofen (3 mg/kg; sc), then transferred to individual cages. Animals were divided into five treatment groups (n=5/group) and injected sc at 10.00 hr on days 8–11 with sterile diluent (V; Eli Lilly and Co., Indianapolis, IN) or neutral protamine Hagedorn insulin (INS, 5.0 U/kg bw/mL; Butler Schein Animal Health, Dublin, OH) [Mandal et al., 2017], according to the treatment schedule shown in Table 1. On day 11, groups injected with one or three INS doses were pretreated at 09.30 hr by intracerebroventricular (icv) administration of the alpha1-adrenergic receptor inverse-agonist prazocin (PRZ; 1.0 μg/2.0 μL [Briski and Shakya, 2018]) or vehicle 0.9% saline (V). Animals were sacrificed at 11.30 hr for brain tissue and trunk blood collection. Dissected brains were immediately snap-frozen in liquid nitrogen-cooled isopentane and stored at −80°C. Plasma was obtained by immediate centrifugation and stored at −20°C.

Table 1:

Experimental Design

		Treatment Days
Treatment Groups		1Day 8	2Day 9	3Day 10	4Day 11	LVa pretreatment on Day4
1	n=5	V scb	V sc	V sc	V sc	Vc
2	n=5	V sc	V sc	V sc	INS scd	V
3	n=5	V sc	V sc	V sc	INS sc	PRZe
4	n=5	INS sc	INS sc	INS sc	INS sc	V
5	n=5	INS sc	INS sc	INS sc	INS sc	PRZ

^aLateral ventricle

^bSterile diluent; 100 μL/100 g bw sc

^cArtificial cerebrospinal fluid; 2.0 μL

^d12.5 U neutral protamine Hagedorn insulin/kg bw sc

^e1.0 ug/2.0 uL

Western Bot Analysis of Laser-Catapult Microdissected VMN Nitrergic and GABAergic Neurons:

Each VMN was cut into alternating series of thin (10 μm thickness; intended for laser-microdissection) or thick (200 μm thickness; intended for micropunch dissection) frozen sections along its rostro-caudal, over repeating distances of 100 μm (10 × 10 μm sections) and 400 μm (2 × 200 μm thick sections), respectively. Sections of 10 μm-thickness were mounted on polyethylene napththalate, e.g. PEN membrane slides (Carl Zeiss MicroImaging LLC, Thornwood, NY), fixed with acetone, and immunostained to identify GABAergic or nitrergic neurons. Immunolabeling was achieved by 48 hr (4°C) incubation of tissues with rabbit primary antibodies raised against GAD_65/67 (prod. no. ABN904, 1:1500; MilliporeSigma, Burlington, MA) or nNOS (prod. no. NBP1–39681, 1:500; Novus Biologicals, LLC, Littleton, CO) [Ibrahim et al., 2019]. Sections were then exposed to goat anti-rabbit biotinylated secondary antibody (prod. no. BA-1000, 1:000, Vector Laboratories, Burlingame, CA), followed by R.T.U. Vectastain Elite ABC-HRP reagent (prod. no. PK-7100, Vector Lab.). GAD_65/67 and nNOS-immunoreactive (ir)-positive neurons were visualized using Vector ImmPACT DAB peroxidase substrate kit reagents (prod. no. SK-4105; Vector Lab.). Individual immunolabeled neurons featuring a visible nucleus and complete labeling of the cytoplasmic compartment were removed from sections using a Zeiss P.A.L.M. UV-A microlaser IV, and collected into lysis buffer (2.0% sodium dodecyl sulfate (SDS; VWR Intl., Radnor, PA), 0.05 M dithiothreitol (Sigma Aldrich), 10.0% glycerol (Sigma Aldrich), 1.0 mM EDTA (Sigma Aldrich), 60 mM Tris-HCl (Sigma Aldrich), pH 7.2). Within each treatment group, triplicate pools of n=50 nNOS-or GAD_65/67-nerve cell lysates were prepared for separation of individual target proteins in BioRad TGX 10–12% stain-free gels (prod. no. 161–0183, Bio-Rad Laboratories Inc., Hercules CA) [Shakya et al., 2018]. After electrophoresis, gels were UV light-activated (1 min.) in a Bio-Rad ChemiDoc TM Touch Imaging System before overnight transblotting (30 V, 4°C; Towbin buffer) to 0.45-μm PVDF membranes (ThermoFisherScientific; Waltham, MA). After blocking with Tris-buffer saline (TBS), pH 7.4, containing 0.1% Tween-20 and 2% bovine serum albumin, membranes were incubated overnight (4°C) with primary antisera raised in rabbit against glutamate decarboxylase_65/67 (GAD_65/67; 1:10,000; prod. no. ABN904; MilliporeSigma, Burlington, MA), neuronal nitric oxide synthase (nNOS; 1:2,000; prod. no. NBP1–39681; Novus Biologicals, Littleton, CO), AMPK_α1/2 (Thr 172; 1:2,000; prod. no. 2532s, Cell Signaling Technology, Danvers, MA), phosphoAMPK_α1/2 (pAMPK, Thr 172; 1:2,000; prod. no. 2535; Cell Signaling Technol.), α₁AR/ADRA1A (prod. no. NB100–78585, 1:1,000; Novus Biol.), α₂AR/ADRA3A (prod. no. NBP2–22452, 1:1,000; Novus Biol.), GLS2 (NBP1–89766, 1:1000; Novus Biol.), or ME-1 (NBP1–32398, 1:1200; Novus Biol.), or in goat against β₁AR/ADRB1 (prod. no. NB600–978, 1:2,000; Novus Biol.). Membranes were then incubated for 1 hr with peroxidase-conjugated goat anti-rabbit (prod. no. NEF812001EA, 1:5000; PerkinElmer, Waltham, MA) or rabbit anti-goat (prod. no. AP106P, 1:5000; MilliporeSigma) secondary antibodies before exposure to Supersignal West Femto chemiluminescent substrate (prod. no. 34095; ThermoFisherScientific, Rockford, IL). Membrane buffer washes and antibody incubations were carried out by Freedom Rocker™ Blotbot® automation (Next Advance, Inc., Troy NY). Chemiluminescence band optical density (O.D.) values were normalized to total in-lane protein using Image Lab™ 6.0.0 software (Bio-Rad). Precision plus protein molecular weight dual color standards (prod. no. 161–0374, Bio-Rad) were included in each Western blot analysis.

LC-MS/ESI Analysis of VMN Amino Acid Content:

For each animal, the VMN was removed bilaterally from thick frozen sections using a calibrated 0.50 mm hollow punch tool (prod. no. 57401; Stoelting, Wood Dale, IL), pooled by collection into 100 μL ultrapure water, and stored at −80° C. After thawing to room temperature, samples were vortexed (30 sec.) and 100 μL aliquots were derivatized by addition of carbonate buffer, pH 9.0, (100μL), L-glutamic acid 5-benzyl ester (50 μL), and FMOC (100 μL), as described with modification [Rebane and Herodes, 2012]. After vortexing (30 sec.) and maintenance at 25°C (40 min.), 1-Adamantanamine hydrochloride (AD; 50 μL) was added prior to further vortexing (30 sec.) and maintenance at 25°C (5 min.). After centrifugation, a clear supernatant was transferred to 350μL inserts for analysis using a UHPLC ThermoFisherScientific Vanquish pump, Vanquish Autosampler, Vanquish UHPLC+ column compartment, ISQ EC Thermo Scientific mass spectrometer, and ThermoFisherScientific™ Dionex™ Chromeleon™ 7 Chromatography Data System software (Waltham MA). For mobile phase A, 10 mM ammonium acetate or formic acid (60μL in 900 mL of water) was used for negative or positive analysis mode, respectively; mobile phase B consisted of LC-MS grade acetonitrile. Linear gradient mobile phase flow featured an increase in acetonitrile from 50% to 80% from 0–4 min., followed by a reduction to 50% from 4 to either 8 or 15 min. for both modes. Using optimized parameters for vaporizer pressures (VT; 250°C), ion transfer tube pressure (ITT; 200°C), sheath gas pressure (SGP; 25psig), auxiliary gas pressures (AGP; 2psig), and sweep gas pressure (SWGP; 0.5psig), a C18 column (4.6 mm ID x 100 mm L, 5μm, 120Å, Acclaim TM 120; ThermoFisher Sci.), 0.25mL/min. flow rate, 1 μL injection volume, was used for negative mode analysis of Gln, GABA and creatine (Cre) at m/z 367.1, 324.1, and 352.1, while a C18+ column (Accucore Vanquish C18+, ThermoFisherSci.). 0.2 mL/min flow rate, was used for positive mode analysis of Glu and Asp at m/z 369.1, and 326.1.

The following linear equation was used to quantify tissue concentrations of amino acid derivatives:

$$\text{Analyte}\text{Area}\text{/}\text{IS}\text{Area}\text{=}\text{(Analyte}\text{Conc.}\text{/}\text{IS}\text{Conc.)}\text{x}\text{slope}\text{+}\text{Const}.$$

Analyte Area	=	Extracted at suitable m/z from total ion current chromatogram (TIC).
IS Area	=	Extracted at m/z IS 458.2 or 460.2 from TIC.
Const.	=	y-intercept of standard calibration equations below:
		Compound	Linear Equation	Range (μg/mL)	R2
		Gln-FMOC	y=0.3482x+0.0088	0.48–250	0.9912
		Glu-FMOC	y=2.539x+0.01511	0.48–250	0.9958
		Asp-FMOC	y=0.0073x-0.0043	250–1000	0.9768
		Cre-FMOC	y=0.0216x+0.0023	31.25–1000	0.9969
		GABA-FMOC	y=2.0909x+0.0413	0.07–37.5	0.9683
Slope	=	Value derived from standard calibration equations above.
IS Conc.	=	190 μg/mL

From the above equation:

$$\text{Analyte}\text{Conc.}=[((\text{Analyte}\text{Area}\text{/}\text{IS}\text{Area})\text{-Const}.)\text{/Slope}]\text{x}\text{IS}\text{Conc}.$$

VMN amino acid levels were expressed as ug/g tissue.

Plasma Analyte Measurements:

Plasma glucose levels were determined with an Accu-Check Aviva Plus glucometer (Roche Diagnostics, Indianapolis, IN; Kale et al., 2006). Circulating corticosterone (ADI-900–097; Enzo Life Sciences, Inc., Farmingdale, NY) and glucagon (EZGLU-30K, EMD Millipore, Billerica, MA) concentrations were determined using commercial ELISA kit reagents, as described [Mahmood et al, 2018].

Statistics:

Mean plasma glucose and counter-regulatory hormone measures, nitrergic or GABA nerve cell normalized protein O.D. values, and VMN tissue amino acid levels were evaluated by one-way analysis of variance and Student Newman Keul’s test, using IBM SPSS software. Differences of p<0.05 were considered significant.

Results:

Figure 1 depicts effects of icv pretreatment with the α₁-AR reverse-agonist PRZ on VMN NO nerve cell AMPK, pAMPK, and nNOS protein expression during a single hypoglycemic episode versus the fourth of four consecutive daily bouts of hypoglycemia. Data in Panel A show that exposure to antecedent hypoglycemia increased total AMPK levels relative to hypoglycemia-naïve animals and vehicle-injected controls [V/AH-INS versus V/INS and V/V] (F_(4,10) = 12.34; p < 0.0001), and that PRZ further augmented this response [PRZ/AH-INS versus V/AH-INS]. Nitrergic neuron pAMPK levels (Panel B) were augmented after injection of one [V/INS versus V/V], but not a fourth dose of INS (F_(4,10) = 4.26; p = 0.02). Blockade of α₁-AR signaling normalized pAMPK expression during acute hypoglycemia, but increased this protein profile during RIIH. The mean pAMPK/AMPK ratio in NO neurons (Panel 1C) as elevated or equivalent to controls during acute versus recurrent hypoglycemia (F_(4,10) = 4.59; p = 0.01). As indicated in Panel D, expression of the NO marker protein nNOS was significantly augmented during singular, but not serial induction of hypoglycemia [V/INS versus V/V] (F_(4,10) = 5.37; p = 0.008). PRZ pretreatment prevented the positive nNOS response to acute hypoglycemia [PRZ/INS versus V/INS], but did not modify nNOS expression during a fourth hypoglycemic episode.

Figure 1.

Effects of Lateral Ventricular (LV) Administration of the Alpha₁-Adrenergic Receptor (α₁–AR) Inverse Agonist Prazocin (PRZ) on Ventromedial Hypothalamic Nucleus (VMN) Nitric Oxide (NO) Neuron 5’-AMP-Activated Protein Kinase (AMPK), Phospho-AMPK (pAMPK), and Neuronal Nitric Oxide Synthase (nNOS) Protein Expression during Acute versus Chronic Insulin-Induced Hypoglycemia. Groups of male rats (n=5/group) were injected subcutaneously (sc) with vehicle (V) or neutral protamine Hagedorn insulin (INS; 5.0 U/kg bw) on days 1–3 of the experiment, then pretreated on day 4 by intraventricular administration of V or PRZ (1.0 μg/μL) before singular or final insulin therapy. VMN nNOS-immunopositive neurons were laser-catapult microdissected from 10 μm-thick frozen sections collected over the rostro-caudal length of the VMN, and pooled with treatment groups for Western blot analysis of AMPK (Panel A), pAMPK (Panel B), and nNOS (Panel D). Data in these panels depict mean normalized protein optical density (O.D) measures ± S.E.M. for the following treatment groups: V/V (solid white bars); V/INS (solid light gray bars); PRZ/INS (diagonal-striped light gray bars); AH (antecedent hypoglycemia)/V (solid dark gray bars); AH/INS (diagonal-striped dark gray bars). Panel C depicts mean ratios of pAMPK/AMPK ± S.E.M. *p<0.05; **p<0.01; ***p<0.001.

Effects of acute versus chronic hypoglycemia on pyruvate recycling pathway marker protein expression in VMN nitrergic neurons are illustrated in Figure 2. As shown in Panel A, GLT profiles were progressively increased during single versus repetitive exposure to hypoglycemia [V/AH-INS > V/INS > V/V] (F_(4,10) = 7.72; p = 0.0002). Pharmacological suppression of α₁-AR signaling did not alter NO neuron GLT expression after one INS injection, but significantly diminished GLT levels in serially dosed animals. Nitrergic neuron ME-1 levels (Panel B) were augmented to an equivalent extent during acute versus chronic hypoglycemia (F_(4,10) : 7.25; p = 0.001); neither positive ME-1 response was affected by PRZ pretreatment.

Figure 2.

Impact of PRZ Pretreatment on VMN NO Neuron Pyruvate Recycling Pathway Marker Protein Expression during Acute versus Chronic Hypoglycemia. Groups of adult male rats (n=5/group) were pretreated by V or PRZ (1.0 μg/μL) administration to the LV ahead of single or final (fourth of four) injection of neutral protamine Hagedorn insulin (INS; 5.0 U/kg bw). Results show mean normalized VMN nitrergic neuron glutaminase (GLT; Panel A) or malic enzyme-1 (ME-1; Panel B) O.D. values ± S.E.M. for V/V, V/INS, PRZ/INS, AH/V, and AH/INS treatment groups. *p<0.05; **p<0.01; ***p<0.001.

Figure 3 illustrates the role of α₁-AR signaling on patterns of α₁-AR (Panel A), α₂-AR (Panel B), and β₁-AR (Panel C) protein expression in VMN NO neurons during acute versus chronic hypoglycemia. Results indicate that α1-AR levels in these cells were significantly increased after one or four INS injections, and that PRZ normalized expression of this protein (F_(4,10) = 8.02; p = 0.0003). Nitrergic neuron α₂-AR profiles were elevated during chronic hypoglycemia, a response that was reversed by PRZ, but were unaffected by a single hypoglycemic episode (F_(4,10) = 11.28; p < 0.0001). Expression of β₁-AR was up-regulated during acute and chronic hypoglycemia (F_(4,10) = 9.06; p = 0.0001) protein expression during acute versus chronic hypoglycemia. PRZ pretreatment further augmented β₁-AR protein in antecedent hypoglycemia-exposed, but not hypoglycemia-naive animals.

Figure 3.

Impact of α₁–AR Signaling on VMN Nitrergic Neuron Alpha₁ Adrenergic Receptor (α₁-AR), Alpha₂-AR (α₂-AR), and Beta₁-AR (β₁-AR) Protein Responses to Acute versus Chronic Hypoglycemia. Data illustrate mean normalized α₁-AR (Panel A), α₂-AR (Panel B), and β₁-AR (Panel C) protein O.D. measures ± S.E.M. for the following treatment groups: V/V (solid white bars); V/INS (solid light gray bars); PRZ/INS (diagonal-striped light gray bars); AH/V (solid dark gray bars); AH/INS (diagonal-striped dark gray bars). *p<0.05; **p<0.01; ***p<0.001.

Effects of α₁-AR blockade on VMN GABA neuron AMPK, pAMPK, and GAD_65/67 protein expression after one or the fourth of four INS injections are illustrated shown in Figure 4. As shown in Panel 4A, total AMPK protein was significantly diminished during acute hypoglycemia and PRZ pretreatment reversed this response (F_(4,10) = 7.32; p = 0.005). Yet, the same profile was refractory to chronic hypoglycemia. Conversely, GABAergic nerve cell pAMPK protein levels (Panel B) were unchanged during a single bout of hypoglycemia, but were significantly suppressed during recurring exposure (F_(4,10) = 10.17; p = 0.002). PRZ prevented RIIH inhibition of pAMPK expression in this neuron population. The mean pAMPK/AMPK ratio in GABA neurons (Panel 4C) as elevated or equivalent to controls during acute versus recurrent hypoglycemia (F_(4,10) = 12.91; p = 0.001). GAD_65/67 protein (Panel 4D) was significantly decreased during acute, but not chronic hypoglycemia; this profile was not impacted by PRZ pretreatment (F_(4,10) = 8.65; p = 0.001).

Figure 4.

Effects of PRZ on VMN γ-Aminobutyric Acid (GABA) Neuron AMPK, pAMPK, and Glutamate Decarboxylase_65/67 (GAD_65/67) Protein Expression during Acute versus Chronic Hypoglycemia. Groups of male rats (n=5/group) were injected sc with V or INS (5.0 U/kg bw) on study days 1–3, then pretreated on day 4 by intra-LV administration of V or PRZ (1.0 μg/μL) before singular or final insulin therapy. VMN GAD_65/67-immunolabeled neurons were harvested by laser-catapult microdissection and pooled with treatment groups for Western blot analysis of AMPK (Panel A), pAMPK (Panel B), and nNOS (Panel D). Data in these panels depict mean normalized protein optical density (O.D) measures ± S.E.M. for the following treatment groups: V/V (solid white bars); V/INS (solid light gray bars); PRZ/INS (diagonal-striped light gray bars); AH (antecedent hypoglycemia)/V (solid dark gray bars); AH/INS (diagonal-striped dark gray bars). Panel C depicts mean ratios of pAMPK/AMPK ± S.E.M. *p<0.05; **p<0.01; ***p<0.001.

Hypoglycemia-associated patterns of VMN GABAergic neuron GLT and ME-1 protein expression as depicted in Figure 5. Neither GLT (Panel A; F_(4,10) = 11.39; p = 0.0007) nor ME-1 (Panel B; F_(4,10) = 4.527; p = 0.02) profiles were affected by either acute or chronic hypoglycemia. Inhibition of α₁-AR signaling elevated expression of both proteins in animals exposed to one, but not four bouts of hypoglycemia.

Figure 5.

Impact of PRZ Pretreatment on VMN GABA Neuron Pyruvate Recycling Pathway Marker Protein Expression during Acute versus Chronic Hypoglycemia. Results show mean normalized VMN GABA neuron GLT (Panel A) and ME-1 (Panel B) O.D. values ± S.E.M. for V/V, V/INS, PRZ/INS, AH/V, and AH/INS treatment groups. *p<0.05; **p<0.01; ***p<0.001.

Figure 6 shows effects of single versus serial INS dosing on α₁-AR (Panel A), α₂-AR (Panel B), and β₁-AR (Panel C) protein expression in VMN GABA neurons. Results reveal that α₁-AR protein profiles were up-regulated by α₁-AR-dependent mechanisms after one, but not four INS injections (F_(4,10) = 6.60; p = 0.007). Expression of α₂-AR was unaffected by either acute or chronic hypoglycemia despite an inhibitory tone imposed by α₁-AR signaling (F_(4,10) = 10.57; p = 0.001). Profiles for β₁-AR were diminished during single or repetitive exposure to hypoglycemia; these decrements in protein expression were exacerbated by PRZ pretreatment (F_(4,10) = 15.96; p < 0.0001).

Figure 6.

Impact of α₁–AR Blockade on VMN GABAergic Neuron α₁-AR, α₂-AR, and β₁-AR Protein Profiles during Acute versus Chronic Hypoglycemia. Data illustrate mean normalized α₁-AR (Panel A), α₂-AR (Panel B), and β₁-AR (Panel C) protein O.D. measures ± S.E.M. for the following treatment groups: V/V (solid white bars); V/INS (solid light gray bars); PRZ/INS (diagonal-striped light gray bars); AH/V (solid dark gray bars); AH/INS (diagonal-striped dark gray bars). *p<0.05; **p<0.01; ***p<0.001.

Effects of acute versus chronic hypoglycemia on VMN tissue amino acid levels are depicted in Figure 7. Data show that Gln concentrations (Panel A) were significantly decreased during acute, but not chronic hypoglycemia (F_(4,10) = 7.32; p = 0.005). VMN Glu levels (Panel B; F_(4,10) = 38.61; p < 0.0001) and Asp (Panel D; F_(4,10) = 38.57; p < 0.0001) content were uniformly diminished irrespective of number of hypoglycemic episodes. PRZ amplified acute hypoglycemic reductions in VMN Asp, but blunted partially reversed Asp decline chronic hypoglycemia. As shown in Panel C, tissue GABA concentrations were reduced during acute, but not chronic hypoglycemia (F_(4,10) = 2.92; p = 0.03). VMN Cre levels were unaffected by hypoglycemia, but were significantly reduced in PRZ plus INS-treated rats (F_(4,10) = 7.18; p = 0.0002).

Figure 7.

Impact of PRZ on VMN Tissue Glutamine (Glu), Glutamate (Glu), GABA, Aspartate (Asp), and Creatine (Cre) Content during Acute versus Chronic Hypoglycemia. VMN tissue was micropunch-dissected from 200 μm-thick frozen sections and analyzed by LC-MS/ESI for amino acid content. Graphs depict mean normalized VMN Gln (Panel A), Glu (Panel B), GABA (Panel C), Asp (Panel D), and creatine (Panel E) measures ± S.E.M. for V/V, V/INS, PRZ/INS, AH/V, and AH/INS treatment groups. *p<0.05; **p<0.01; ***p<0.001.

Figure 8 illustrates effects of PRZ pretreatment on acute versus chronic hypoglycemia and counter-regulatory hormone secretion. Data in Panel A show that the magnitude of glucose decline was equivalent between single versus serial INS-dosed rats (F_(4,10) = 66.38; p < 0.0001), and that PRZ caused a small, but significant reversal of acute hypoglycemia. Plasma glucagon concentrations (Panel B) were significantly elevated following acute INS administration, a response that was prevented by PRZ pretreatment (F_(4,10) = 14.26; p < 0.0001). RIIH suppressed circulating glucagon levels below baseline, a negative response that was averted by PRZ. Corticosterone secretion (Panel C) was increased during acute, but not chronic hypoglycemia (F_(4,10) = 10.15; p = 0.002). PRZ prevented acute hypoglycemic hypercorticosteronemia, yet increased hormone release during chronic hypoglycemia.

Figure 8.

Effects of PRZ on Glycemic, Glucagon, and Corticosterone Responses to Singular or Serial INS Injection to Male Rats. Data show circulating plasma glucose (Panel A), glucagon (Panel B), and corticosterone (Panel C) levels ± S.E.M. for V/V, V/INS, PRZ/INS, AH/V, and AH/INS treatment groups (n=5/group). *p<0.05; **p<0.01; ***p<0.001.

Discussion:

Maintenance of vital, high energy-demand neuron functions, e.g. preservation of trans-membrane electrical potential and neurotransmitter uptake/recycling, requires uninterrupted glucose delivery to the brain. Mal-adaptive repression of glucose counter-regulation during RIIH involves habituated metabolic monitoring in the brain. Acute hypoglycemia up-regulates gluco-stimulatory NO signaling and represses gluco-inhibitory GABA release within the MBH to drive optimum counter-regulatory hormone outflow; yet, these responses are attenuated during RIIH. Noradrenergic activity within the neural glucostatic network, including the VMN, is broadly affected by hypoglycemia, [Shrestha et al., 2014; Mandal et al., 2018]. Exogenous NE governs VMN nNOS and GAD_65/67 marker protein expression [Ibrahim et al., 2019]. Here, combinatory immunocytochemistry/laser-catapult microdissection/Western blot techniques were used to determine if VMN nitrergic and/or GABAergic neurons express the metabolic-sensory molecular marker AMPK, and whether α₁-AR signaling governs AMPK reactivity to acute hypoglycemia and mediates sensor acclimation to recurring episodes. Immunoblotting was also employed to evaluate metabolic-sensory neuron pyruvate recycling pathway activity, alongside LC-MS/ESI analysis of VMN content of amino acids processed via glutaminolysis to yield energy.

Data document that VMN NO and GABA neurons express hypoglycemia-sensitive AMPK, and that acute hypoglycemia-associated augmentation of AMPK activity is mediated, in part, by α₁-AR up-regulation of pAMPK profiles in nitrergic neurons or down-regulation of total AMPK in GABAergic nerve cells. Results also show that antecedent hypoglycemia repressed sensor activation in both cell groups during renewed hypoglycemia. In nitrergic cells, this acclimation in sensor function may involve loss of α₁-AR stimulation of AMPK activity documented during acute hypoglycemia, as PRZ did not modify sensor activation from control-like levels during RIIH. Antecedent hypoglycemia may trigger, in the absence of α₁-AR input to these cells, α₁-AR-independent mechanisms that protect cellular energy balance, and thus stabilize AMPK activity, during ensuing hypoglycemia. Alternatively, α₁-AR-reliant and-non-reliant mechanisms may converge on AMPK, such that the latter remains operational during pharmacological blockade of α₁-AR. Indeed, AMPK is governed by several endocrine and neurochemical cues in addition to nutrient availability. In contrast, interestingly, GABAergic neurons exhibit a switch from α1-AR stimulation to inhibition of AMPK activity during acute versus chronic hypoglycemia; further effort is required to characterize the mechanisms that mediate this change in direction of α₁-AR influence on AMPK activity state. As data here demonstrate α₁-AR expression by NO and GABA neurons, it is likely that PRZ effects involve inhibition of direct NE input. However, as the current experimental paradigm involved PRZ delivery into the LV, blockade of upstream α₁-AR that control afferent input may also contribute to nerve cell reactivity to drug treatment. Our presumption is that the change in α₁-AR impact on GABA AMPK activity from positive to negative involves adjustments in downstream receptor and/or signal transduction pathways. However, it is possible that antecedent hypoglycemia may alter increase α₁-AR – mediated inhibitory neurochemical cues to VMN GABA AMPK.

PRZ pretreatment reduced NO nerve cell nNOS profiles during acute hypoglycemia, but did not modify this protein profile during RIIH. It should be noted measurable changes in nNOS protein levels do not constitute definitive evidence for commensurate adjustments in enzyme activity and NO production; as treatment effects on phosphorylated nNOS were not evaluated here, it remains speculative that nNOS activity habituates to RIIH. These outcomes align with observed differences in drug effects on AMPK activity in nitrergic neurons during acute versus chronic hypoglycemia. Present data suggest that this sensor exerts a major influence on hypoglycemic patterns of VMN NO neurotransmission. Conversely, α₁-AR blockade did not modify GABA neuron GAD_65/67 protein profiles during acute or chronic hypoglycemia, despite corresponding stimulatory or inhibitory drug effects on AMPK activation under those conditions. These results imply that mechanisms unrelated to α₁-AR-dependent AMPK status may act alongside AMPK to both diminish gluco-inhibitory GABA neurotransmission during acute hypoglycemia and up-regulate this signal during recurring hypoglycemia.

VMN NO and GABA neurons were characterized by both common and dissimilar patterns of AR protein expression during hypoglycemia as well as differential AR habituation to recurring hypoglycemia. Both neuron types exhibited α₁-AR – driven increases in α₁-AR protein expression during acute hypoglycemia, but this up-regulated response occurred only in nitrergic cells during RIIH. Expression of α₂-AR protein was refractory to acute hypoglycemia in both cell groups, but NO neurons showed a α₁-AR – dependent increase in this profile during chronic hypoglycemia. Interestingly NO (increased) and GABA (decreased) nerve cell β₁-AR expression diverged during both acute and recurring hypoglycemia. Taken together, these data indicate that antecedent hypoglycemia may promote augmented noradrenergic input to nitrergic neurons as a consequence of augmented α₂-AR expression. On the other hand, GABA nerve cell habituation to RIIH may involve diminished α₁-AR signaling. Nonetheless, the possibility that downstream signal transduction pathways may acclimate to recurring hypoglycemia independent of changes in receptor expression cannot be overlooked.

VMN nitrergic neurons exhibited elevated GLT protein expression during acute hypoglycemia, a response that was amplified by antecedent hypoglycemia exposure. ME-1 profiles in these cells were likewise augmented during acute and chronic hypoglycemia. α₁-AR signaling a statistically non-significant decline in GLT protein during acute hypoglycemia, but mediated the massive increase in this profile during chronic hypoglycemia. On the other hand, VMN GABAergic neurons showed no change in expression of either pyruvate recycling pathway marker protein during acute or chronic hypoglycemia. These data suggest that NO, but not GABA neurons may habituate to RIIH by α₁-AR – mediated up-regulation of this non-glucose energy pathway. LC-MS/ESI analyses showed that VMN Gln, Glu, and Asp concentrations were each diminished during acute hypoglycemia, and that tissue Gln and Asp content was corresponding normalized or further decreased during chronic hypoglycemia. Current results, obtained by application of methodological advances permitting quantification of these amino acids in small-volume neural tissue samples, affirm previous observations of hypoglycemia-associated reductions in whole-brain Gln and Glu [Behar et al., 1985]. Further adjustments in analytical sensitivity will be required to assess effects of hypoglycemia on concentrations of these energy-yielding amino acids in pure NO or GABA neuron samples. It is important to note that while current data support the notion of increased utilization of these non-essential amino acids versus production, they do not shed light on how hypoglycemia may influence relative rates of their synthesis versus metabolism in the VMN, nor do outcomes identify usages other than energy production. Thus, observed differences in Gln levels between acute versus chronic hypoglycaemia may reflect, in part, adjustments in that ratio such that yield and use are each increased, alongside expansion of the available amino acid pool. As the current study did not measure baseline amino acid levels after antecedent hypoglycemia (e.g. prior to the final INS injection), it is not known if some or all pools were increased as an adaptation. PRZ pretreatment did not modify effects of hypoglycemia on VMN Gln or Glu content. Whole-VMN measurements performed in conjunction with pharmacological receptor blockade might have obscured possible differences in extent of noradrenergic regulation of pyruvate recycling pathway activity between individual resident neuron populations. Evidence that decrements in VMN Asp levels are intensified by α₁-AR – dependent mechanisms during RIIH infer that conversion of this amino acid to Glu may be correspondingly increased; however, additional experimentation will be required, including analysis of aspartate transaminase expression and determination of rates of synthesis versus metabolism at the cell population level, to resolve this issue. Present findings that VMN Cre levels are unaffected by hypoglycemia suggest that the total amount of phosphate that can be stored for future transfer to ADP to replenish intracellular ATP is likely stable. Ongoing research seeks to develop suitable methods for quantification of phospho-Cre / Cre ratios in distinctive VMN cell groups in hypoglycemia-naïve and antecedent hypoglycemia-exposed animals.

LC-MS/ESI analyses show that VMN GABA concentrations are diminished during acute hypoglycemia, but are restored to control range during RIIH. These data concur with previous reports that antecedent hypoglycemia minimizes hypoglycemia-associated changes in MBH GABA content [Chan et al., 2008]. Correlation of these measures with similar effects of acute versus chronic hypoglycemia on patterns of GAD_65/67 expression supports the utility of this marker protein expression as an indirect indicator of GABA transmission within the VMN.

A key outcome of the present study is the implication of α₁-AR involvement in counter-regulatory hormone acclimation to RIIH. In the model used here, plasma glucagon and corticosterone concentrations were significantly elevated above baseline after a single INS injection, but were suppressed below or at control levels, respectively, after serial INS treatment. In the absence of time-course analysis of glucagon secretion after the fourth of four daily insulin injections, we only speculate that plasma hormone levels may have been elevated relative to baseline at some point over the 1 and ½ hours between time zero and blood sampling in V/AH-INS animals. Nonetheless, data showing a reduction in glucagon levels in that treatment group compared to A/INS animals at that singular time point are suggestive of a truncated hormone response. Nonetheless, area-under-the-curve analyses would be required for verification of that supposition. PRZ was observed to prevent acute hypoglycemic hyperglucagonemia and hypercorticosteronemia, but also to potently increase hormone release during recurring hypoglycemia. Outcomes here provide novel evidence that antecedent hypoglycemia changes the direction of α₁-AR regulation of counter-regulatory secretion from stimulatory to repressive, and that noradrenergic signaling via this receptor is a principal inhibitor of counter-regulatory hormone outflow during RIIH. There remains an urgent need to identify cellular substrates for this suppressive α₁-AR action. As discussed above, the inefficacy of PRZ to alter NO or GAD_65/67 protein expression during RIIH does not preclude the likelihood that redundant, α₁-AR – independent cues may up-regulation NO and down-regulation GABA, such that pharmacological alleviation of α₁-AR input will not perturb acclimated patterns of protein expression. Moreover, NE may act on other distinctive α₁-AR – expressing nerve cells within or outside the VMN that may/may not be metabolic-sensory in function and may/may not innervate nitrergic or GABAergic nerve cells, to repress glucagon and corticosterone secretion during RIIH. As the threshold for induction of glucagon secretion was not directly evaluated here, we cannot overlook the possibility that PRZ-associated increases in glucose profiles after acute insulin treatment may have had a direct impact on counter-regulatory hormone output. Yet, evidence for significant PRZ effects on glucagon and corticosterone secretion in V-versus PRZ-injected AH-INS animals, despite equivalence of circulating glucose levels in those two treatment groups supports the notion that at least during RIIH, PRZ effects on these hormones is mediated by neural mechanisms.

In summary, current research documents noradrenergic involvement in VMN nitrergic and GABAergic nerve cell AMPK accommodation to RIIH via adjusted impact of α₁-AR signaling. Results support the possibility that VMN NO neurons may exhibit up-regulated energy yield via α₁-AR – dependent up-regulation of pyruvate recycling pathway activity. Application of high-resolution, high-sensitivity LC-MS/ESI methodology for discriminative analysis of amino acid energy substrate levels in the VMN supports the possibility that hypoglycemia-associated patterns of Gln and Asp utilization may vary between acute versus chronic hypoglycemia.

Highlights:

• Norepinephrine controls ventromedial hypothalamic nucleus (VMN) glucoregulatory signaling.

• The α₁-AR reverse agonist prazocin (PRZ) was administered intracerebroventricularly prior to hypoglycemia.

• PRZ prevented AMPK activation in VMN nitrergic and GABAergic neurons.

• Habituation of GABA nerve cell AMPK to antecedent hypoglycemia was reversed by PRZ.

• PRZ inhibited acute hypoglycemic counterregulatory hormone profiles, but prevented hormone acclimation.

Abbreviations:

α₁-AR

alpha1 adrenergic receptor

α₂-AR

alpha2 adrenergic recepto

AMPK

5’-AMP-activated protein kinase

asp

aspartate

β₁-AR

beta1 adrenergic receptor

cre

creatine

GABA

γ-aminobutyric acid

GAD_65/67

glutamate decarboxylase65/67

GLT

glutaminase

glu

glutamate

gln

glutamine

icv

intracerebroventricular

IIH

insulin-induced hypoglycemia

INS

insulin

LV

lateral cerebral ventricle

ME-1

malic enzyme

NE

norepinephrine

nNOS

neuronal nitric oxide synthase

pAMPK

phosphoAMPK

PRZ

prazocin

RIIH

recurrent insulin-induced hypoglycemia

VMN

ventromedial hypothalamic nucleus