Exaggerated CRR in mice lacking LepRb from PBN LepRb^CCK neurons

To understand whether PBN LepRb neurons might enhance the CRR in low-leptin states, we sought a molecular marker to permit the manipulation of PBN LepRb neurons. Since cholecystokinin (CCK)-containing PBN neurons project to the VMH^35,36, we examined the potential expression of LepRb in PBN CCK neurons. We bred Cck^cre mice to the ROSA26-eYFP background to generate CCK^eYFP mice and examined the induction of pSTAT3-IR in CCK^eYFP cells, demonstrating that many PBN LepRb cells express CCK (LepRb^CCK neurons) (Fig. 2a, b). While some LepRb^CCK cells were also observed in other brainstem regions (including the Edinger-Westphal (EW), periaqueductal grey (PAG), and NTS) (data not shown), LepRb^CCK neurons were absent from the hypothalamus and other brain areas. As for PBN LepRb neurons, we observed increased cFos-IR in PBN CCK cells following treatment with 2DG (Supplemental Fig. 5), suggesting that PBN LepRb^CCK cells represent the subpopulation of PBN LepRb cells that are activated by glucoprivation.

Open in a separate window

Figure 2

Deletion of LepRb from PBN CCK cells blunts IIH and augments the CRR to glucoprivation

(a) Cck^cre mice were crossed onto the Rosa26-eYFP background to generate CCK^eYFP mice expressing eYFP in CCK neurons and/or were crossed onto the Lepr^flox background to generate LepRb^CckKO mice. (b, c) CCK^eYFP (b) and LepRb^CckKO^eYFP (c) mice were injected with IP leptin (5 mg/kg) 2 hours prior to perfusion and immunostaining for pSTAT3 (purple) and GFP (eYFP; green). The PBN of representative mice is shown; representative of at least 3 mice per genotype. Scale bars= 100 μM. Inset-digital zoom of the indicated area; arrowheads indicate colocalized neurons. (d) Hypoglycemic response to IP insulin administration (1 U/kg) in male LepRb^CckKO mice or control genotype littermates. N=11 (con) and 12 (ko) animals; p=0.01 (15min),0.03 (30min){F(4,21)=2.55} test; all data are plotted as mean +/− SEM. (e-h) Male Control and LepRb^CckKO mice were injected with 2DG (250 mg/kg, IP) and monitored for 180 minutes. (e) Blood glucose was monitored at the indicated times; *p=0.004 (30min),0.002 (45min),<0.001 (60min),0.003 (90min){F(7,22)=3.3}, N= 11 (con) and 13 (ko) animals. Plasma glucagon (f) *p=0.008 {t(20)=-2.641}, N=11 animals per condition, corticosterone (g) p=0.029 {t(20)=-2.01}, N=11 animals per condition, and epinephrine (h) *p=0.044 {t(17)=-1.80}, N=11 (con) and 8 (ko) animals, 90 minutes after 2DG administration. Panels d-h: n=8-12 animals per condition, representative of 6 similar experiments; data are plotted as mean +/− SEM (d-e) or Q1, Q2, and Q3 with 10^th and 90^th percentiles as whiskers (f-h)..*p<0.05 d and e were analyzed by two way repeated measures ANOVA with Fisher LSD post hoc test. Hormone levels (f-h) were analyzed by one tailed t-test, consistent with our directional hypothesis.

To understand the function of these brainstem LepRb^CCK cells in leptin action, we crossed Cck^cre onto the Lepr^flox background to generate mice lacking LepRb expression specifically in LepRb^CCK neurons (LepRb^CckKO animals) (Fig. 2a). As expected, leptin-stimulated pSTAT3-IR was absent from CCK-expressing PBN neurons in LepRb^CckKO mice, and overall leptin-stimulated PBN pSTAT3-IR was reduced by approximately 75% in the PBN of LepRb^CckKO animals compared to controls (Fig. 2b, c). Leptin-stimulated pSTAT3-IR was also significantly reduced in the EW, but not in other brain areas (Supplemental Fig. 6). Thus, Cck^cre ablates LepRb expression in the LepRb^CCK cells of the brainstem PBN and EW nuclei.

We detected no alteration in body weight, endocrine function, glucose homeostasis or glucose tolerance in LepRb^CckKO mice compared to controls (Supplemental Fig. 7). Consistent with an enhanced CRR, however, LepRb^CckKO mice displayed blunted IIH (Fig. 2d). Indeed, like fasted mice (Supplemental Figure 3), LepRb^CckKO mice exhibited more robust glycemic excursions in response to 2DG than did controls (Fig. 2e). Furthermore, while insulin was unchanged (data not shown), circulating glucagon, corticosterone and epinephrine concentrations were increased in LepRb^CckKO mice compared to controls 90 minutes after 2DG administration (Fig. 2f-h). The increased glycemic excursion in response to glucoprivation exhibited by in LepRb^CckKO mice is similar to that observed in fasted control animals, and even prolonged fasting fails to further enhance this response in LepRb^CckKO mice, suggesting overlapping mechanisms for leptin action on LepRb^CCK neurons and the augmented CRR observed with prolonged fasting (Supplemental Fig. 8). Furthermore, these cells are at least partly specific to hypoglycemia (as opposed to all stressors), since restraint stress provoked no increase in cFos in PBN LepRb neurons, and the hyperglycemic response to restraint stress was normal in LepRb^CckKO mice (Supplemental Fig. 9).

To examine IIH directly, we subjected the LepRb^CckKO mice to hypoglycemic clamp analysis, in which high levels of insulin are continuously infused while the glucose infusion rate is varied to maintain the desired level of hypoglycemia, thereby examining counter-regulatory glucose production during IIH (Fig. 3). The glucose infusion rate required to prevent blood glucose from falling below the clamped level was decreased in LepRb^CckKO mice compared to controls, consistent with increased counter-regulatory hepatic glucose production in these animals (Fig. 3a-c). Indeed, the hepatic expression of glucose-6-phosphatase (G6Pase, which dephosphorylates glucose to permit its efflux from the cell) was increased in LepRb^CckKO mice compared to controls at the end of the clamp (Fig. 3d). Note that G6Pase expression, which is decreased by insulin, is augmented by counter-regulatory stimuli that increase cAMP, including epinephrine and glucagon³⁷, and thereby provides an indirect indication of the liver’s potential to release glucose. Thus, leptin action via LepRb^CCK neurons does not contribute to baseline glucose tolerance, but rather enhances the CRR by promoting increased counter-regulatory hormone release and hepatic glucose output.

Open in a separate window

Figure 3

Enhanced CRR during hypoglycemic clamp in LepRb^CckKO mice

Male LepRb^CckKO mice or littermates of control genotypes with arterial catheters were infused with insulin to suppress blood glucose and activate the CRR. For the subsequent 120 minutes, (a) blood glucose was monitored and glucose infusion rate (b) was adjusted to maintain blood glucose at approximately 50 mg/dl; *p=0.004 (30min), 0.002 (40min), <0.001 (50min), <0.001 (60min), 0.002 (70min) {F(12,13)=2.70}, N= 7(con) and 8 (ko) animals. (c) Area under the curve (AUC) for the data graphed in (b) *p=0.014, {t(13)=2.84} N=7(con) and 8 (ko) animals. (d) hepatic G6Pase mRNA expression (fold over control) at the end of the clamp *p=0.015, {t(14)=−2.41}, N=7 (con) and 9 (ko) animals. data in a-b are plotted as mean +/− SEM and as Q1, Q2, and Q3 with whiskers at the 10^th and 90^th percentiles in c-d. Data in b were analyzed by two way repeated measures ANOVA with Fisher LSD post hoc test; c-d were analyzed by one tailed t-test, consistent with our directional hypothesis.

DREADD-mediated inhibition of PBN LepRb neurons blunts the CRR

To evaluate the function of PBN LepRb neurons, specifically, we utilized the stereotaxic injection of AAVs that mediate the cre-dependent expression of designer receptors exclusively activated by designer drugs (DREADDs; expressed as DREADD-mCherry fusion proteins). DREADDs are genetically-engineered muscarinic receptor variants that are insensitive to endogenous ligands, but which are activated by the otherwise biologically inert clozapine-N-oxide (CNO)^38,39; CNO activates neurons containing the G_q-coupled hM3D_q and inhibits neurons that contain the G_i-coupled hM4D_i^38,39.

To determine the role for PBN LepRb neurons in the endogenous CRR, we injected the inhibitory AAV-hM4D_i bilaterally into the PBN of Lepr^cre animals, allowed the animals to recover for several weeks, and examined the response to 2DG-mediated glucoprivation in the presence or absence of CNO (Fig. 4a-c). We found that CNO significantly blunted the increase in blood glucose concentrations following 2DG treatment in mice that expressed hM4D_i in PBN LepRb neurons. Thus, inhibition of PBN LepRb cells impairs the CRR, demonstrating that PBN LepRb neurons play an integral role in the response to glucoprivation.

Open in a separate window

Figure 4

Inhibition of PBN LepRb neurons blunts the hyperglycemic response to glucoprivation

(a) Bilateral injection of the cre-inducible AAV-hM4Di into the PBN of male Lepr^cre mice induces the expression of the inhibitory DREADD hM4Di in PBN LepRb neurons. Animals were injected with Vehicle or CNO (Injx: 0.3 mg/kg, IP) and then with 2DG (250 mg/kg, IP) and blood glucose was monitored for 240 minutes. (b) Blood glucose was monitored at the indicated times; *p=0.005 (15min), 0.017 (30min), 0.021 (60min), <0.001 (90min), <0.001 (120min), 0.004 (180min), <0.001 (240min) {F(9,38)=1.91} N=20 (veh) and 20 (cno) animals. (c) Area under the curve (AUC) for the data graphed in (b); p<0.001 {t(38)=4.22} N=20 (veh) and 20 (cno) animals. all data are plotted as mean +/− SEM.The data in b were analyzed by two way repeated measures ANOVA with Fisher LSD post hoc test; c was analyzed by two tailed t-test. Data in a is plotted as mean +/− SEM and as Q1, Q2, and Q3 with whiskers at the 10^th and 90^th percentiles in b.

DREADD-mediated activation of PBN LepRb neurons mimics the CRR

Our findings predict that activation of PBN LepRb neurons should mimic the CRR. To test this, we injected the activating AAV-hM3D_q bilaterally into the PBN of Lepr^cre animals and, following a several-week recovery period, treated them with CNO (Fig. 5). CNO treatment increased cFos-IR in the PBN and the VMH of these animals compared to vehicle (Supplemental Fig. 10), as would be expected upon activation of the PBN LepRb→VMH circuit. CNO also increased blood glucose concentrations (Fig. 5b, c), along with glucagon and corticosterone, while tending to decrease insulin compared to vehicle (Fig. 5d-f). Since the effect of increased glucose on pancreatic islets would cause glucagon to fall and insulin to rise, the opposite finding (that glucagon rose while insulin tended to fall) implies increased SNS outflow to the pancreatic islets upon activation of PBN LepRb cells. Activation of PBN LepRb neurons also increased hepatic G6Pase expression (Fig. 5g). Thus, activation of PBN LepRb neurons increases circulating concentrations of counter-regulatory hormones and hepatic glucose production to raise blood glucose, similar to the native CRR.

Open in a separate window

Figure 5

Activation of PBN LepRb neurons mimics the CRR to glucoprivation

(a) Bilateral injection of the cre-inducible AAV-hM3Dq into the PBN of male Lepr^cre mice induces the expression of the activating DREADD hM3Dq in PBN LepRb neurons. Animals were injected with Vehicle or CNO (0.3 mg/kg, IP) and monitored for 120 minutes. (b) Blood glucose was monitored at the indicated times *p<0.001 (15min), <0.001 (30min), <0.001 (45min), <0.001 (60min), <0.001 (90min), 0.002 (120min) {F(6,30)=8.23}, N=16(veh) and 16(cno) animals. (c) Area under the curve (AUC) for the data graphed in (b); *p<0.001 {t(42)=−4.38} N=22 (veh) and 22 (cno) animals. Plasma glucagon (d) *p=0.002 {t(35)=−3.03} N=18(veh) and 19(cno) animals, insulin (e) *p=0.049 {t(40)=1.69} N=21(veh) and 21(cno) animals, corticosterone (f) 30 minutes after CNO administration *p=0.034 {t(40)=−1.87}, N=21(veh) and 21 (cno) animals. (g) hepatic G6Pase mRNA expression (fold over control) at the end of the experiment *p<0.001 {t(17)=−6.64}, N=10 (veh) and 9 (cno) animals. all data are plotted as mean +/− SEM. *p<0.05 The data in b were analyzed by two way repeated measures ANOVA with Fisher LSD post hoc test. The data in c were analyzed by two tailed t-test; d-g were analyzed by one tailed t-test, due to our directional hypothesis. Data in b is plotted as mean +/− SEM and as Q1, Q2, and Q3 with whiskers at the 10^th and 90^th percentiles in c-g.

Electrophysiologic studies

3-5 week old LepRb reporter mice of either sex were euthanized by decapitation. Horizontal slices (250 μM) were prepared in oxygenated ice-cold sucrose solution containing (in mM): 220 sucrose, 2.5 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 5 glucose, 6 MgCl₂, 1 CaCl₂ using a VT 1200S vibratome (Leica Microsystems Inc., Buffalo Grove, IL). Slices were allowed to recover for at least 1 hr before electrophysiological analysis in a holding chamber containing oxygenated artificial cerebrospinal fluid (aCSF) solution containing (in mM): 125 NaCl, 3 KCl, 1.25 NaH₂PO₄, NaHCO₃, 2 glucose, 1 MgCl₂, 2.5 CaCl2, pH 7.4. Individual fluorescently labeled lPBN LeprRb neurons were visualized using an Olympus BX51WI upright microscope equipped with IR-DIC optics (Olympus, Tokyo, Japan). Patch electrodes made from borosilicate glass capillaries (Warner Instruments, Hamden, CT) were pulled to a tip resistance of 3-7 MΩ using a Brown/Flaming P-97 micropipette puller (Sutter Instr. Co., Novato, CA) and filled with a solution containing (in mM): 130 K gluconate, 10 KCl, 1 EGTA, 10 HEPES, 0.6 NaGTP, 2 MgATP, 8 phosphocreatine, pH 7.2. Slices were perfused with aCSF solution warmed to 32-33° C using a TC-344 temperature controller and preheater (Warner Instruments, Hamden, CT) and continuously bubbled with 5% CO₂ and 95% O₂. Low glucose aCSF (0.5 mM, osmolarity adjusted with sucrose) and leptin were applied via bath perfusion. Membrane potentials were measured using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA) in tight-seal whole-cell current clamp mode. Currents were filtered at 5 kHz, digitized at 10 kHz and analyzed offline. Data were not corrected for a junction potential 14.7 mV. Neurons selected for analysis had stable series resistances (Rs) < 25 MΩ that were not compensated.

Phenotypic Studies

Starting at 21 days of age, animals were singly housed. Blood glucose and body weight were monitored weekly in these animals starting at 28 days. Beginning at 10 weeks of age, glucose homeostasis was examined by glucose tolerance test (GTT; 2 g/kg, IP), insulin tolerance test (ITT; 1.2 units/kg Humulin (Eli Lilly), IP) and 2DG (Sigma) challenge (250 or 500 mg/kg, IP). The animals were allowed at least 10 days of recovery prior to each challenge. Mice were fasted for four hours prior to each challenge: Food was removed following lights on and testing followed four hours later. Tail vein blood was collected for the measurement of glucose (one touch ultra 2 glucocometer (Johnson and Johnson)). Plasma or serum was prepared from larger volume samples and stored at −20°C for later assay. Samples for the determination of epinephrine concentrations were obtained from animals with an arterial catheter (placed by the UM Animal Phenotyping Core (APC)). Glucagon (Millipore) and Glucocorticoid (MP Biomedicals) levels were determined via radioimmunoassay, while insulin and leptin were determined via multiplex assay (Millipore).

Determination of plasma epinephrine concentrations

Five μL of plasma was spiked with 1.25 μL of D6-epinephrine as an internal standard. Proteins were removed by addition of 25 μL of ice cold acetonitrile, followed by centrifugation for 10 min at 13,400 rpm. Twenty μL of the supernatant was removed and benzoylated⁴³ by sequential addition of 10 μL of 100 mM sodium tetraborate, 10 μL of benzoyl chloride (2% in acetonitrile, v/v), and 10 μL of sulfuric acid (1% in DMSO, v/v). Standard solutions of epinephrine were prepared in aCSF, which is similar in composition to plasma without protein⁴⁴, to create a calibration range of 0.1 – 20 nM. Standards were spiked with internal standard, diluted with acetonitrile, and derivatized as above. Calibration curves were prepared based on the peak area ratio of the standard to the internal standard by linear regression. All samples and standards were analyzed in triplicate using an Acquity HSS T3 C18 chromatography column (1 mm × 100 mm, 1.8 μm, 100 Å pore size) in a nanoAquity UPLC (Waters) interfaced to an Agilent 6410 triple quadrupole mass spectrometer. Mobile phase A was 10 mM ammonium formate with 0.15% (v/v) formic acid in water. Mobile phase B was acetonitrile. The gradient used was: initial, 0% B; 0.01 min, 23% B; 2.51 min, 23% B, 3 min, 50% B; 5.3 min, 60% B; 6.46 min, 65% B; 6.47 min, 100% B; 7.49 min, 100% B; 7.5 min, 0% B; 8.5 min, 0% B at 100 μL/min. Benzoylated epinephrine eluted at 6.85 min. The sample injection volume was 5 μL in partial loop injection mode. The autosampler was kept at ambient temperature and the column was held at 27 °C. Electrospray ionization was used in positive mode at 4 kV. The gas temperature was 350 °C, gas flow was 11 L/min, and the nebulizer was at 15 psi. Ions were detected in MS-MS mode with precursor ions of 496 and 502 m/z for benzyolated epinephrine and D6-epinepherine, respectively, and product ion 105 m/z for both. The following voltage settings were used: fragmentor = 120 V, collision energy = 15 V, and cell accelerator = 4 V. Automated peak integration was performed using Agilent MassHunter Workstation Quantitative Analysis for QQQ, version B.05.00. All peaks were visually inspected to ensure proper integration.

Hyperinsulinemic-hypoglycemic Clamp

Following a 5-hour fast, the UM APC infused conscious, unstressed, catheterized mice with 10mU/kg/min insulin and glucose was clamped at ~50mg/dl for 120 minutes by the infusion of glucose ^45,46. Following hyperinsulinemic-hypoglycemic clamp, animals were euthanized for the collection of liver and brain tissue.

Perfusion and Immunohistochemistry

Where indicated, mice were treated with leptin (5mg/kg; the generous gift of AstraZeneca Pharmaceuticals) 2 hours prior to perfusion to label leptin-responsive neurons via the induction of pSTAT3-IR. The mice were anesthetized with sodium pentobarbital and a lobe of liver removed prior to transcardial perfusion and collection of the brain. The brains were sectioned coronally at 30μm on a freezing microtome and distributed into 4 series for analysis. Series were pre-treated with 1% Hydrogen Peroxide followed by individual washes of PBS with 0.3% Glycine and 0.03% SDS. The sections were incubated overnight at room temperature in the rabbit anti-cFos (Santa Cruz, sc-52; 1:1000), chicken anti-GFP (Abcam, ab12970; 1:1000), rabbit anti-dsred (living colors, 632496; 1:1000), and/or rabbit anti-pSTAT3 (Cell Signaling Technology,91455; 1:500) antibody and exposed the next day with either biotinylated (1:200 followed by ABC amplification and DAB reaction) or fluorescent secondary antibody (Molecular Probes, 1:200) in order to visualize proteins. Each of these antibodies is included on the Journal of Comparative Neurology Antibody database (http://onlinelibrary.wiley.com/journal/10.1002/%28ISSN%291096-9861/homepage/jcn_antibody_database.htm). The sections were mounted on glass slides and coverslipped with vectashield mounting media (Vector labs) for later image capture. Images were collected on an Olympus BX-51 microscope, and immunoreactive cells/co-labeled cells were counted manually using Adobe Photoshop. Total eYFP-IR cells were counted within the area, as well as colocalized eYFP-IR cells with nuclear marker (cfos-IR or pstat3). Cells were counted bilaterally and expressed as raw counts from each 1:4 series.

mRNA analyses

mRNA was extracted from liver samples using trizol (Invitrogen); 1μg was converted to complementary DNA (cDNA) using superscript reverse transcriptase kit (Invitrogen). Using the cDNA, GAPDH and target genes were analyzed via Taqman kits (Applied Biosystems) on an Applied Biosystems 7500 Real Time PCR system. Relative mRNA expression was calculated using the 2^−ΔΔCt method.

Stereotaxic injection of viral constructs

Following the induction of isoflurane anesthesia and placement in a stereotaxic frame, the skulls of adult Lepr^Cre mice were exposed. After Bregma and Lambda were leveled, a guide cannula with injector was lowered into the approximate PBN coordinates from Bregma (anteroposterior: −3.3, mediolateral: 1.2, dorsoventral: −4.85). 100 nl of either Ad-iN Syn-mCherry²⁹, AAV-hM4D_i, or AAV-hM3D_q (University of North Carolina Vector Core) were injected using a 500 nl Hamiltion Syringe at a rate of 20nl/minute. After 5 minutes of time post-injection for adequate dispersal and absorption of the virus, the injector and cannula were removed from the animal; the incision site was closed and sutured. The mice received pre-surgical and post-surgical analgesics. The mice injected with Ad-iN Syn-mCherry were allowed one week to recover before being sacrificed, while AAV-hM3D_q/hM4D_i were allowed at least two weeks to recover from surgery before experimentation.

Proglumide Treatment

Lepr^Cre mice were pretreated with AAV-hM3D_q bilaterally injected into the PBN (as described above) and at least two weeks later, were food deprived from 9:00 am to 1:00 pm, at which time the study commenced. Baseline blood glucose was determined and mice were pre-treated with 20 mg/kg of the pan-specific CCK-receptor antagonist proglumide sodium salt (Tocris Biosciences) or 0.9% sterile saline. After 30 min blood glucose was assessed and mice were administered 0.3 mg/kg CNO or 0.9% sterile saline. Blood glucose was monitored for a further 120 min. The study was repeated in C57BL/6 mice pretreated with 100 mg/kg proglumide followed by treatment with 500 mg/kg 2DG or 0.9% sterile saline 30 min later. Blood glucose was monitored for a further 180 min.

We thank AstraZeneca Pharmaceuticals (San Diego, CA) for the generous gift of leptin, Dr. Bryan Roth (University of North Carolina, Chapel Hill, MC) for CNO, and members of the Myers and Heisler labs for helpful discussions. Research support was provided by the Animal Phenotyping Core of the Michigan Diabetes Research Center (NIH Grant P30 DK020572) and the Michigan Nutrition and Obesity Research Center (P30 DK089503), the American Diabetes Association, the American Heart Association and the Marilyn H. Vincent Foundation and the NIH (DK098853) to MGM, NIH DK055267 & DK020595 to CJR, NIH DK046060 (RTK), NIH DK046409 (LS), and the Wellcome Trust (098012) and BBSRC (BB/K001418/1) to LKH.