If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Fasting hyperglycemia in diabetes mellitus is caused by unregulated glucagon secretion that activates gluconeogenesis (GNG) and increases the use of pyruvate, lactate, amino acids, and glycerol. Studies of GNG in hepatocytes, however, tend to test a limited number of substrates at non-physiologic concentrations. Therefore, we treated cultured primary hepatocytes with three identical substrate mixtures of pyruvate/lactate, glutamine, and glycerol at serum fasting concentrations, where a different U-13C or 2-13C labeled substrate was substituted in each mix. In the absence of glucagon stimulation, 80% of glucose produced in primary hepatocytes incorporated either one or two 13C-labeled glycerol molecules in a 1:1 ratio, reflecting the high overall activity of this pathway. In contrast, glucose produced from 13C-labeled pyruvate/lactate or glutamine rarely incorporated two labeled molecules. While glucagon increased glycerol and pyruvate/lactate contribution to glucose carbon by 1.6- and 1.8-fold, respectively, glutamine contribution to glucose carbon was increased 6.4-fold in primary hepatocytes. To account for substrate 13C carbon loss during metabolism, we also performed a metabolic flux analysis, which confirmed that the majority of glucose carbon produced by primary hepatocytes was from glycerol. In vivo studies using a PKA-activation mouse model that represents elevated glucagon activity confirmed that most circulating lactate carbons originated from glycerol, but very little glycerol was derived from lactate carbons, reflecting glycerol’s importance as a carbon donor to GNG. Given diverse entry points for GNG substrates, hepatic glucagon action is unlikely to be due to a single mechanism.
Diabetes mellitus (DM) is currently a global health epidemic that affects nearly 10% of the population worldwide (
). The secretion of glucagon is dependent on paracrine insulin signaling in the pancreatic islet, and the loss of paracrine regulation markedly increases glucagon levels in poorly controlled T1DM and in the late stage of T2DM (
). There are three ways for carbons to enter GNG: 1) glycerol can be converted to glycerol-3-phosphate and enter the pathway as dihydroxyacetone phosphate (DHAP); 2) lactate and other 3-carbon amino acids (e.g. alanine) can be converted to pyruvate and enter the pathway as oxaloacetate; and 3) other gluconeogenic amino acids (e.g. glutamine) can first enter the tricarboxylic acid (TCA) cycle before being metabolized to oxaloacetate. At the early stage of fasting, glycogen catabolism in muscle and other peripheral organs produces circulating lactate (Lac), which has long been assumed to be the dominant source of gluconeogenic carbon (
). A signaling cascade activated by glucagon ultimately leads to increased protein kinase A (PKA) activity, mediating increases in gluconeogenic enzymes such as glucose 6-phosphatase (G6PC) and phosphoenolpyruvate carboxykinase (PCK1)(
). However, a recent study suggests an alternative mechanism whereby glucagon enhances GNG by allosterically activating pyruvate carboxylase (PCX) through calcium-induced hepatic lipolysis and beta-oxidation (
) have generally not considered glycerol or have employed substrates at super-physiological concentrations to study glucagon’s effect. One in vivo study concluded that lactate has a dominant role in glucose metabolism and by extension GNG, given its very high turnover flux in mammals (
). However, given the many substrates that feed GNG, it is essential to evaluate multiple substrates to determine glucagon’s overall effect.
In this study, we utilized a novel 13C tracing method to study the effect of glucagon in mouse primary hepatocytes treated with physiological concentrations of substrates including glycerol, pyruvate, lactate, and glutamine and in mice treated with non-perturbative concentrations of these substrates. Glucagon increased glucose production from all four substrates in primary hepatocytes, indicating PCX activity alone is insufficient to explain glucagon’s effect. Importantly, in the presence of all four substrates at physiologic concentrations, 76% of GNG carbon originates from glycerol. Glucagon had three effects on substrate utilization in primary hepatocytes, which are correlated with increased expression of G6PC and PCK1: 1) absolute GNG flux from all four substrates was increased; 2) glucagon preferentially increased glutamine use, and 3) relative contribution of glycerol is decreased although glycerol still remains the dominant GNG substrate in the absence or presence of glucagon stimulation. In vivo tracer infusion in PKA-activated mice confirm these findings.
Glucagon-induced glucose production in primary hepatocytes
To recapitulate physiological fasting substrate concentration in glucagon-induced hepatic gluconeogenesis (GNG), mouse primary hepatocytes (PH) were cultured in physiological substrate concentrations found in mouse serum after a 12 h fast (
). Under these in vitro conditions, issues of in vivo hepatic zonation of GNG and changes in substrate concentrations across vascular beds are eliminated. When cultured with individual substrates, glucagon significantly increased glucose production from each substrate, but glucose production was greatest when PH were cultured in the presence of all substrates (Fig. S1A). These media concentrations were maintained at a constant level over the 8h period by supplementation (Fig. S1B).
To determine individual substrate contribution to GNG, we cultured mouse primary hepatocytes in the same four-substrate mixture, substituting one of the substrates with a U-13C labeled version (pyruvate and lactate mixture (PL) are considered one substrate because of rapid interconversion). This results in three labeling schemes (Fig. 1A): 13C3 glycerol (*Gro) with all other substrates unlabeled (*Gro/PL/Gln), 13C3 pyruvate and lactate (*PL) with all other substrates unlabeled (Gro/*PL/Gln), and 13C5 glutamine (*Gln) with all other substrates unlabeled (Gro/PL/*Gln). Since all substrates were tested at identical concentrations without or with 13C labeling, glucose production could be measured at non-perturbative concentrations and fractional labeling of glucose was maximized. Total glucose production rate in the media of all schemes was similar (Fig. 1A), indicating that the U-13C labeled substrates did not alter overall glucose production. Interestingly, however, 80% of the glucose produced after glucagon stimulation contained either one (m+3) or two (m+6) glycerol molecules. When all labeled substrates were considered, > 95% of all glucose in the media was from GNG, indicating glucose released from glycogen into the media was minimal. *Gro/PL/Gln was the only substrate mix that generated m+6 glucose (Fig. 1A), indicating the incorporation of two 13C3 glycerol molecules and increased efficiency of this pathway. In contrast, the predominant form of glucose synthesized from either *PL or *Gln was m+3, followed by m+2 and m+1, the latter two species are due to 13C loss in the TCA cycle (Fig. 1A), which was confirmed using 2-13C labeled substrates (see below). Average carbon enrichment was also lower with TCA-derived substrates due, in part, from 13C loss in the TCA cycle (note m+1 and m+2 species of *PL and *Gln, Fig. 1B).
Shift of carbon source preference in glucagon-induced gluconeogenesis
We next calculated the 13C flow from each substrate to glucose in mouse primary hepatocytes using a previously validated in vitro model (Figs. 1C and 1D, ref. 17). Under basal conditions, relative contribution to the triose-phosphate (TPhos; glyceraldehyde 3-phosphate and dihydroxyacetone phosphate) pool can be determined using the flux ratio of glycerol to TCA-derived substrates (pyruvate/lactate and glutamine). This ratio of 340/110 or 3.1-fold indicates that 76% of the carbons in glucose were derived from glycerol (Fig. 1C and 1D). As expected, overall GNG flux increased approximately 2-fold after glucagon treatment (450 to 972 nmol C/h, Fig. 1C), but the flux ratio of glycerol to TCA-derived substrates dropped to 547/425 or 1.3-fold, indicating a preferential increase in TCA-derived substrate use. Further analysis suggests that relative glutamine flux to the triose-phosphate pool was dramatically increased by glucagon (319/50 or 6.4-fold), while relative pyruvate/lactate flux to the triose-phosphate pool was increased only 1.8-fold (106/60), similar to the 1.6-fold flux increase from glycerol after glucagon (547/340). These glucagon-induced changes reduced glycerol carbon contribution to glucose from 76% to 56% (Fig. 1D). Given that glucagon is thought to increase both glutamine and pyruvate/lactate flux (
), we also determined the average 13C enrichment of gluconeogenic intermediates (Fig. S1C). After glucagon treatment, 13C enrichment of TCA intermediates was increased from glutamine but decreased from the pyruvate/lactate (Fig. S1C), consistent with previous findings (
In primary hepatocytes, U-13C pyruvate/lactate can be consumed in β-oxidation, leading to lower glucose labeling when free fatty acids are absent from the media. To test this hypothesis in primary hepatocytes, we supplemented our four-substrate culture conditions with free fatty acid (FFA) and found that FFA modestly increased total glucose production (Fig. S1D). However, FFA treatment had minimal or no effect on glucose production from pyruvate/lactate and glutamine. In fact, FFA supplementation reduced glucose production from glycerol after glucagon stimulation (Fig. S1D), likely due to higher cellular NADH levels from β-oxidation reducing glycerol flux through glycerol 3-phosphate dehydrogenase. Consistent with a recently described model (
), Fig. S2E demonstrates higher cellular acetyl-CoA levels after either glucagon or FFA treatment. Higher cellular NADH levels from β-oxidation would be predicted to reduce glycerol flux through glycerol-3-phosphate dehydrogenase and glutamine flux through the TCA cycle and might explain these results.
Using mixtures of uniformly 13C-labeled substrates, our data demonstrate that glycerol is most efficient at labeling glucose synthesized in mouse primary hepatocytes without any loss of 13C label (only m+3 and m+6 glucose was produced, Fig. 1A). Unlike glycerol, however, pyruvate/lactate and glutamine are likely to lose 13C label in the TCA cycle, potentially underestimating their contribution to glucose synthesis. All GNG substrates label the triose-phosphate pool, consisting of glyceraldehyde 3-phosphate (GAP) and dihydroxyacetone phosphate (DHAP) in presumed rapid equilibrium- an assumption needed to perform this analysis. To explore further the labeling pattern of glucose and relative substrate use, we employed 2-13C labeled glycerol, pyruvate, and lactate (the latter in a 1:10 ratio as above and indicated as *PL) in physiological fasting concentrations and repeated the mouse primary hepatocytes study (Figs. 1, E-H). In this study, we continued to use U13C glutamine to label glucose. Interestingly, the glucose labeling results were identical to those using U13C labeled substrates (compare Figs 1A and 1E), demonstrating in both cases more than 80% labeling of glucose by 2-13C glycerol and nearly 100% labeling of glucose when all labeled substrates were considered. We also found that physiological 2-13C glycerol predominantly labeled the triose phosphate pool (>70% of GAP labeled, Fig. 1G) versus minor labeling by either 2-13C pyruvate/lactate or U13C glutamine. Therefore, low labeling of glucose by 2-13C pyruvate/lactate or U13C glutamine in the unstimulated state reflects their intrinsically lower labeling of the triose-phosphate pool, not loss of 13C label in the TCA cycle.
Study of triose-phosphate pool labeling before and after glucagon stimulation was instructive when determining changes in substrate preference. Glucagon reduced the GAP labeled fraction from 2-13C glycerol from 71% to 53%, maintained the GAP labeled fraction from 2-13C pyruvate/lactate at ∼10%, and increased the GAP labeled fraction from U13C glutamine from 3% to 9% (Fig. 1G) These data suggest dilution of glycerol-labeled GAP by glutamine-labeled GAP. Consistent with these data, the m+1/m+2 ratio from 2-13C glycerol was also increased (i.e., more of the triose-phosphate pool is derived from glutamine and less from labeled glycerol after glucagon stimulation).
Finally, we used mass isotopomer distribution analysis (MIDA) to confirm the binomial relationship between the fractional labeled product (glucose) and fractional concentration of both labeled substrates (DHAP and GAP). Many previous studies have shown that this relationship is valid even in the setting of unlabeled substrates (
). Here we show that MIDA predicts the 13C fractional labeling pattern of glucose from fractional GAP labeling, assuming that GAP and DHAP are in rapid equilibrium and constitute 3-carbon monomers that are subsequently assembled as a 3-carbon dimer (compare Figs. 1F and 1H). Moreover, the increase in the glucose m+1/m+2 ratio after glucagon treatment reflects a dilution of fractional GAP labeling by 2-13C glycerol with unlabeled 3-carbon substrate from glutamine.
Intracellular mediators of glucagon action
Glucagon is known to stimulate GNG through cyclic adenosine monophosphate (cAMP)-mediated activation of PKA and increased transcription of GNG enzymes such as phosphoenolpyruvate carboxykinase 1 (PCK1) and glucose-6-phosphatase (G6PC) (Fig. 2A) (
) in the same labeling scheme as shown in Fig. 1. Both PKA and INSP3R1 inhibitors blocked glucagon-mediated increases in glucose production in a concentration-dependent manner (Figs. 2B-D and Figs. S2A and S2B). In contrast, ATGL and PCK1 inhibitors only partially inhibited glucagon’s effect even at very high concentrations (Figs. 2B-D and Figs. S2C and S2D). Importantly, both PKA and INSP3R1 inhibitors (H89 and 2APB, respectively) neutralized the increased gluconeogenic flux from all three groups of substrates (Figs. 2B-D). The ATGL inhibitor (Atgi) slightly suppressed flux from pyruvate/lactate but not from either glycerol or glutamine (Figs. 2B-D). On the other hand, the PCK1 inhibitor (3MPA) clearly suppressed flux from both pyruvate/lactate and glutamine but not from glycerol (Figs. 2B-D). Interestingly, however, 3MPA did alter the m+3/m+6 ratio after glycerol labeling (Fig. 1B) from 1.3 to 0.5, indicating less unlabeled TCA-derived substrates supplied the other 3-carbon molecule to glucose.
A concentration-dependent increase in gluconeogenic enzyme expression by glucagon
To understand the roles of gluconeogenic enzymes in substrate utilization, we treated mouse primary hepatocytes with various concentrations of glucagon and measured gluconeogenic enzyme expression (Figs. 3A-C) and corresponding protein levels (Fig. 3D). G6pc expression correlates with the glucose production rate as both showed a significant increase at 1 nM glucagon (Figs. 3A, 3D, and 3E). The expression of Pck1 showed a similar trend except that the increase at 1 nM concentration is not statistically different from control primary hepatocytes (Fig. 3B). In contrast, pyruvate carboxylase (Pcx) expression was not changed by glucagon treatment (Fig. 3C) as also reported (
). Contrary to a previous report, these data clearly show that glucagon’s stimulatory effect on gluconeogenic enzyme expression is a plausible mechanism for increased glucose production in primary hepatocytes (Fig. 3F, ref. 8).
Forskolin recapitulates glucagon’s effect on substrate usage in gluconeogenesis
Glucagon’s effect on primary hepatocytes can be largely explained by increased Pck1 and G6pc expression secondary to cAMP, calcium signaling, or both (Fig. 2A). Given our focus on determining substrate-specific gluconeogenic flux, we wanted to determine if glucagon’s effect on GNG could be reproduced in primary hepatocytes using forskolin (a direct adenylyl cyclase activator). Using identical labeling schemes, glucagon and forskolin produced the same changes in glucose production and relative substrate utilization (Figs. 4A-D) and expression of G6pc and Pck1 (Figs. 4E and 4F). These data indicate that forskolin-mediated PKA activation recapitulates glucagon mediated regulation of substrates use in GNG.
Hepatic PKA activation increases gluconeogenesis in vivo
The findings on glucagon action presented so far were derived from in vitro studies in primary hepatocytes. Some investigators have attempted to study the global action of glucagon in vivo using either acute injections or chronic infusions of glucagon (
). Systemic treatments, however, affect both the liver and peripheral tissues, such as adipose depots and the pancreas, making localization of glucagon’s effect problematic. Given that forskolin, a direct adenylyl cyclase activator, completely recapitulated glucagon’s effect in primary hepatocytes (Fig. 4), we modeled elevated hepatic glucagon action by constitutive PKA activation using hepatic knockout of the PKA regulatory subunit in Prkar1afl/fl mice (
Littermates of Prkar1afl/fl mice were placed on a normal chow for 16 weeks and then injected with either Ad-CRE or control Ad-GFP virus to generate experimental groups: L-GFP and L-PKA. Ad-CRE significantly reduced the expression of Prkar1a proteins in liver to activate hepatic PKA without differential expression of Prkar1a proteins in other tissues such as skeletal muscle (Fig. S3A). L-GFP and L-PKA mice showed similar body weight (Fig. S3B) and fat mass (Fig. 3SC). As expected, increased PKA activity was found in the liver of L-PKA mice compared to L-GFP control (Fig. 5A), resulting in increased fasting glucose (Fig. 5B) due to enhanced GNG in the liver. Consistent with findings in primary hepatocytes, hepatic PKA activation in L-PKA mice significantly increased G6pc (Fig. 5C) and Pck1 (Fig. 5D) expression compared to L-GFP mice. As expected, PKA activation (L-PKA) significantly increased glucose production from i.p. injection of either Pyr (pyruvate tolerance test, PTT, Fig. 5E) or Gro (glycerol tolerance test, GroTT, Fig. 5E). Higher glucose levels found in L-PKA mice were positively correlated with serum insulin levels (Fig. S3D) and negatively correlated with serum glucagon levels (Fig. S3E).
PKA activation affects lactate contribution to GNG in vivo.
To investigate if PKA activation affects gluconeogenic flux in vivo, we infused L-GFP and L-PKA mice with 13C3 glycerol (Fig. S4A) or a mixture of 13C3 pyruvate/lactate (in a physiological ratio of 1:10) (Fig. S4B) following a 12h fast. After a 6h infusion of either tracer, all isotopologues of serum glucose reached steady-state, and no significant changes in metabolite pool size were observed after the infusion (Figs. S4C and S4D), indicating non-perturbative conditions were achieved. 13C enrichment after either glycerol or pyruvate/lactate infusion was also similar (∼15%) among the groups. The endogenous turnover rate (Fcirc) of circulating glycerol and lactate was calculated as previously described (
Both L-GFP and L-PKA mice showed a similar glycerol turnover rate (Fig. 6A), indicating similar glycerol production and consumption. Because glycerol is the only 13C source at steady-state conditions, its fractional contribution to glucose can be estimated based on a ratio of the average 13C enrichment (Fig. 6B). We previously reported that glycerol contributed the majority of the glucose carbon in normal fasted mice both as a direct hepatic substrate and by labeling lactate in the Cori cycle (
). We now show that glycerol also contributes approximately 60% to glucose carbon in both L-GFP and L-PKA (Fig. 6B). Given that primary hepatocytes do not generate m+1 or m+2 glucose labeling after treatment with 13C3 glycerol, m+1 and m+2 serum glucose found after in vivo glycerol labeling (Fig. S4A) most likely represent glucose synthesized from labeled lactate generated from a glycerol→glucose→lactate pathway that undergoes 13C carbon loss in the TCA cycle.
The lactate turnover rate was also similar in both groups (Fig. 6C), and pyruvate/lactate contributed ∼65% of glucose carbon in L-GFP and L-PKA mice (Fig. 6D). It should be noted that both glycerol and pyruvate/lactate each contributed more than 50% to glucose carbon because a significant portion of pyruvate/lactate carbon originates from glycerol at steady-state (Fig. S4E) (
). In contrast, < 10% of glycerol carbon originated from pyruvate/lactate (Fig. S4F), indicating that the net flow of carbon is from glycerol to pyruvate/lactate. We also measured the glucose turnover rate (Fig. 6E) and its contribution to circulating glycerol and lactate (Fig. 6F) by infusing mice with 13C6-glucose. In contrast to glycerol and lactate turnover, L-PKA mice showed significantly increased glucose turnover rate compared to the L-GFP control (Fig. 6E). Glucose contributed more than 85% of lactate carbon but only 10% of glycerol carbon in both groups (Fig. 6F), consistent with a predominant glycerol→glucose→lactate carbon flow.
To illustrate more clearly the in vivo carbon flows among glycerol, lactate, and glucose, we performed a metabolic flux analysis (MFA) using the enrichment and Fcirc data and previously developed model (Fig. 6G and ref 24). L-PKA mice showed a significant increase in gluconeogenic flux from lactate compared to L-GFP mice (Fig. 6I), while gluconeogenic flux from glycerol showed a similar flux (Fig. 6H).
Gluconeogenesis has been extensively studied by a number of investigators. While other studies have linked both glutamine and pyruvate/lactate metabolism to glucagon’s action in primary hepatocytes (
), these studies were performed at super-physiological fasting concentrations and failed to consider glycerol’s role as a primary substrate. They must now be interpreted with caution, given that glycerol is the primary substrate used by primary hepatocytes in the absence or presence of glucagon. In particular, glucagon stimulation of primary hepatocytes markedly increased glycerol production (
), suggesting it is necessary to separate glucose produced by pyruvate/lactate due to lipolysis (FA oxidation→acetyl-CoA→PCX activation, Fig. 2A) from glucose produced from glycerol liberated during lipolysis.
Under basal conditions, primary hepatocytes 13C enrichment data indicate that non-glycerol substrates are minor substrates compared to glycerol. We rigorously considered the hypothesis that 13C carbon loss from all non-glycerol substrates in the TCA cycle minimized their glucose labeling contributions. Ultimately, we rejected this hypothesis because both the MFA (U13C labeled substrates) and MIDA (2-13C labeled substrates) studies demonstrate that glycerol is the primary substrate labeling glucose (Figs. 1D and 1F) or the triose phosphate pool metabolites (Figs. 1G and S1C) under physiologic fasting conditions.
After glucagon stimulation, PH 13C enrichment data demonstrate increased use of all substrates, but glutamine use is preferentially increased (Figs. 1D and 1F). We also considered if 13C carbon loss from pyruvate/lactate minimizes its glucose labeling contribution, but we rejected this hypothesis for two reasons: 1) Glutamine’s but not pyruvate/lactate’s contribution to glucose carbon increased after glucagon stimulation, even though both enter the TCA cycle (Figs. 1C and D); and 2) 13C enrichment of TCA intermediates by glutamine but not pyruvate/lactate increases after glucagon (Fig. S1C). Consistent with the latter finding, Miller et al. showed that glucagon increased glutamine flux in GNG through activation of α-ketoglutarate dehydrogenase (
). We conclude that our results reflect the relative activity of these GNG pathways in mouse primary hepatocytes towards different substrates depending on their concentration and glucagon stimulation.
After depleting tissue glycogen, carbon reservoirs for GNG include gluconeogenic amino acids from proteolysis or glycerol from lipolysis. Among all gluconeogenic substrates, lactate, pyruvate, alanine, glutamine, and glycerol together contribute more than 97% of all carbons to glucose (
). Our data extend this finding to show that glycerol is the ultimate carbon source for more than 60% of circulating lactate and glucose in mice with or without activation of the PKA signaling pathway by glucagon. Thus, glycerol can enter GNG directly or after conversion to lactate via glycerol→glucose→lactate metabolism in the Cori cycle. Fasting hypoglycemia, hyperketonemia, and developmental delay in patients with defects in glycerol metabolism also suggest that glycerol is essential for GNG in humans (
Age-matched C57BL/6J albino males between 3-4 months were used for PH culture isolation. All mice were anesthetized using a ketamine/xylazine mixture (Henry Schein) before isolation. The hepatic portal vein was cannulated and perfused with Kreb’s Ringer Bicarbonate Buffer (Sigma) containing EGTA for 10 min at 37 °C. After the first wash, a second Kreb’s Ringer wash containing CaCl2 and Liberase™ (Roche) was applied for 10 minutes at 37 °C. Hepatocytes were filtered through a gauze mesh and resuspended in plating media: William’s Media E (Sigma) containing 10 % FBS (Sigma), 200 nM dexamethasone (Sigma), 1× penicillin/streptomycin (Fisher) and 2 mM L-glutamine (Fisher). Cells were plated at a density of 3×105/ml on a six-well collagen (Sigma) coated plate. Hepatocytes were allowed to recover over-night and all experiments were started 24 h post isolation, unless otherwise indicated.
In vitro isotope labeling experiments
After recovery at 37 °C and 5% CO2 from isolation and viral treatment if applicable, PH were serum-starved for 3 h in a basal media (no-glucose DMEM supplemented with 3.52 mg/ml HEPES, 1× penicillin/streptomycin and 2 mM L-Gln). For in vitro glucose production, the cells were washed once with 1× PBS and treated with glucose production medium (no-glucose DMEM with 3.52 mg/ml HEPES and 1×penicillin/streptomycin) and substrates with or without 20 nM glucagon. For the physiological-concentration substrate scheme, 0.5 mM L-glutamine, 0.33 mM glycerol, 0.25 mM sodium pyruvate and/or 2.5 mM sodium lactate were used. For the high-concentration substrate scheme, 5mM glycerol and 50 mM sodium lactate were used. Culture medium were collected every 2h for the measurement of glucose production and labeling patterns. To compensate the loss of substrates and maintain the physiological concentration, 0.22 μmol glycerol, 0.33 μmol pyruvate and 0.05 μmol L-glutamine per ml culture were added every 2 h after the sampling of media.
For the inhibition of PKA or PCK1, 25 μM H-89 dichloroacetate hydrate (Sigma) or 500 μM 3-Mercaptopropionic acid (3-MPA, Sigma) was directly dissolved in the glucose production medium. For the inhibition of INSP3R1 or ATGL, 50 μM 2APB (R&D Systems) or 50 μM atglistatin (Sigma) was first dissolved in the vehicle of 0.5% DMSO or 0.1% ethanol, respectively and then added to the glucose production medium. The total amount of glucose in the media was measured with Glucose Assay Kit (Abcam Cat# ab65333).
For 13C labeled substrate experiments, the same conditions were used but substrates were substituted one at a time with 13C3-sodium pyruvate/lactate, 13C5-L-glutamine or 13C3-glycerol (Cambridge Isotope Lab) as indicated. The medium metabolites were directed extracted in 100 volumes of ice-cold 40:40:20 methanol:acetonitrile:water solution with 0.1% formic acid. The cells were washed once with prewarmed 1× PBS and extracted with ice-cold 40:40:20 methanol:acetonitrile:water solution with 0.1% formic acid (2 ml per million cells). Both media and cell extracts were incubated on ice for 5 min and neutralized with 0.7% NH4HCO3, followed by centrifugation for 10 min at 16,000 × g. The supernatant was then transferred to another clean tube for mass spectrometry (LC-MS) analysis (
). To avoid the variation caused by estrus cycles, only male mice were used in this study. All mice were maintained on a C57BL/6J-albino background (Jackson Labs; B6(Cg)-Tyrc-2J/J). Littermates of Prkar1afl/fl mice (Jackson Labs; Prkaa1tm1.1Sjm/J) were placed on a normal chow for 11 weeks after weaning and then injected with either Ad5-CMV-CRE or Ad5-CMV-GFP (3 x 109 pfu/mouse; University of Iowa Viral Vector Core) via the tail vein, resulting in L-GFP and L-PKA mice. All experiments were performed between 2-5 weeks after the injection. All mice described were housed in a pathogen-free barrier facility with a 12-hour light/dark cycle. All animal protocols were approved by the Institutional Animal Care and Use Committee of Rutgers University.
Protein Extraction and Western Blotting
Protein preparations were conducted with both hepatocytes that were lysed in Laemmli Buffer with β-mercaptoethanol and tissue samples. For tissues protein preparations, the tissues had been snap-frozen in liquid nitrogen following the sacrifice of the animal and were mechanically homogenized in ice-cold RIPA buffer (Sigma) with 1× protease inhibitor (Roche) and 1× phosphatase inhibitor (Thermo Scientific) using Bullet Blender (Next Advance, Inc., NY). The homogenate was then centrifuged at 16,000 x g at 4 °C and the supernatant was collected. Protein concentrations were measured using the Pierce BCA assay (Thermo Scientific). Western blotting was performed using standard procedures utilizing the standard V3 Western Workflow (Bio-Rad). All standard western blotting reagents were obtained from Bio-Rad. Antibodies were obtained from Cell Signaling Technology (CYPB (1:1000; Cat# 43603S), Phospho-PKA substrate (1:1000; Cat# 9624S), Anti-rabbit IgG, HRP-linked (1:2000; Cat# 7074S), Anti-mouse IgG, HRP-linked (1:2000, Cat# 58802)), Abcam (PCK1 (1:1000; Cat# ab70358), PCX (1:1000; Cat# ab126707)) or Atlas (G6PC (1:250; Cat# HPA052324)).
For the pyruvate and glycerol tolerance tests, mice were fasted overnight (12h) and then injected i.p. with sodium pyruvate (9 mmol/kg; Sigma) or glycerol (9 mmol/kg; Sigma). Blood glucose measurements were obtained via a small nick in the lateral tail vein using a glucometer (Bayer Contour).
Quantitative RT-PCR analysis
Total RNA was isolated from primary hepatocytes and mouse livers using TriZol method. cDNA was obtained using iScript (Biorad) and then subjected to qRT-PCR analysis using SYBR Green (BioRad) according to manufacturer’s protocol. The primers used for the analysis were the following: G6pc Forward 5’-CAGCAAGGTAGATCCGGGA-3’ Reverse 5’-AAAAAGCCAACGT ATGGATTCCG-3’; Pck1 Forward 5’-AGCATTCAACGCCAGGTTC-3’ Reverse 5’- CGAGTCTGTCAGTTCAATACCAA-3’; Pcx Forward 5’-TGGGTTCCTCTCAGAGCGAG-3’ Reverse 5’-GTCTCCCATCTTGCGGACC-3’; Gk Forward 5’-ATCCGCTGGCTAAGAGACAACC-3’ Reverse 5’-TGCACTGGGCTCCCAATAAGG-3’ Hnf4α Forward 5’-CTTCCTTCTTCATGCCAG-3’ Reverse 5’-ACACGTCCCCATCTGAAG-3’ Actb Forward 5’-CCAGTTGGTAACAATGCCATG-3’ Reverse 5’-GGCTGTATTCCCCTCCATCG-3’.
Insulin, Glycerol and Glucagon Measurements
Serum insulin levels were measured using the Ultra-Sensitive Mouse Insulin ELISA Kit (Crystal Chem). Serum glycerol levels were measured using the Glycerol Assay Kit (Sigma). Serum glucagon levels were measured using Glucagon ELISA Kit (Crystal Chem). Liver glycerol levels were calculated by subtracting the endogenous glycerol-3-phosphate level (measured using the Sigma Glycerol Assay Kit without the addition of ATP) from the sum of glycerol-3-phosphate and liver glycerol (measured using the Sigma Glycerol Assay Kit with the addition of ATP).
In vivo Isotope Labeling Studies
For continuous infusion experiments, five mice from each group (L-GFP and L-PKA) were catheterized on the right jugular vein (
) and recovered more than 7 days. Catheterized mice were fasted for 12h and then transferred to new cages without food and infused to circulating isotope steady-state (5-6 h). For the mouse infusion, a tether and swivel system were used to allow mice free movement in the cage (Instech Laboratories). Water-soluble isotope-labeled metabolite tracers (Cambridge Isotope Laboratories) were prepared as solutions in sterile normal saline and infused via the catheter at a constant rate (0.1 ul/g body weight/min). 200 mM U-13C glucose, 150 mM U-13C glycerol or 40 mM U-13C sodium pyruvate with 360 mM U-13C lactate were infused. About 30 μl blood were collected by tail vein bleeding at each time point, left at room temperature in the absence of anticoagulant for 30 minutes and centrifuged at 4 °C to prepare serum. Serum and tissue samples were kept at −80 °C until further extraction.
Serum (10 ul each) was mixed with extraction solution (ice cold 40:40:20 methanol:acetonitrile:water solution with 0.1% formic acid), followed by vortexing for 10 sec, incubation at 4 °C for 10 min, and centrifugation at 4 °C and 16,000g for 10 min. The volume of the extraction solution (in μl) was 25× the volume of serum. The supernatant was transferred to a clean tube and neutralized with 0.7% NH4HCO3 solution. The mixture was centrifuged again at 4 °C at 16,000g for 10 min. The supernatant was then transferred to another clean tube for mass spectrometry (LC-MS) analysis.
Glycerol Derivatization for LC-MS Analysis
Due to poor ionization of glycerol, a derivatization reaction is required to detect glycerol in LC-MS. Samples containing glycerol were added into 10× volume of reaction buffer containing 25 mM Tris (pH 8.0), 10 mM Mg2+, 50 mM NaCl, 5 mM ATP and 2 U/ml glycerol kinase and incubated for 10 minutes at room temperature. The Gro ion counts equal the difference of glycerol-3-phosphate ion counts before and after the reaction.
LC conditions were optimized on an HPLC-ESIMS system fitted with a Dionex UltiMate 3000 HPLC and a Thermo Q Exactive Plus MS. The HPLC was fitted with a Waters XBridge BEH Amide column (2.1 mm × 150 mm, 2.5 μm particle size, 130 Å pore size) coupled with a Waters XBridge BEH XP VanGuard cartridge (2.1 mm x 5 mm, 2.5 μm particle size, 130 Å pore size) guard column. The column over temperature was set to 25 °C. The solvent A consisted of water/acetonitrile (95:5, V/V) with 20 mM NH4Ac and 20m M NH4OH at pH 9. The solvent B consisted of acetonitrile/water (80:20, V/V) with 20 mM NH4Ac and 20 mM NH4OH at pH 9 in the following solvent B percentages over time: 0 min, 100%; 3 min, 100%; 3.2 min, 90%; 6.2 min, 90%; 6.5 min, 80%; 10.5 min, 80%; 10.7 min, 70%; 13.5 min, 70%; 13.7 min, 45%; 16 min, 45%; 16.5 min, 100%. The flow rate was set to 300 μl/min with an injection volume 5 μl. The column temperature was set at 25 °C. MS scans were obtained in negative ion mode with a resolution of 70,000 at m/z 200, in addition to an automatic gain control target of 3 x 106 and m/z scan range of 72 to 1000. Metabolite data was obtained using the MAVEN software package50 with each labeled isotope fraction (mass accuracy window: 5 ppm). The isotope natural abundance and tracer isotopic impurity were corrected using AccuCor (
). This model reproducibly predicts flux for glycerol, pyruvate/lactate, and glutamine. For in vivo experiments, the turnover rates of circulating glycerol, lactate and glucose were calculated using the following equation (
All analysis and graphs were done on GraphPad Prism 7.0 software. Statistics were performed using either student’s t-test or one-way ANOVA where appropriate.
The data are available within the article and supplemental information.
Conflict of Interests
The authors declare no competing interests.
We thank Dr. Kirschner et al. for making the Prkar1afl/fl mice, Metabolomics Core Facility at CINJ and Princeton for LC-MS analysis, and members of the Wondisford/Radovick laboratory. We thank Drs. Moshmi Bhattacharya, Abdelfattah El Ouaamari and Christoph Buettner for scientific discussions. This work was supported by NIH grants R01 DK063349 (to F.E.W) and the Metabolomics Shared Resource of Rutgers Cancer Institute of New Jersey (P30CA072720).
Conceptualization, Y.W., H.X. and F.E.W.; Methodology, Y.W., H.K. and X.S; Software, Y.W and X.S.; Validation, Y.W., H.X., H.K. and X.S.; Formal Analysis, Y.W. and H.X.; Investigation, Y.W., H.X., K.K. and H.K., Resources, X.S. and F.E.W.; Data Curation, Y.W., H.X., H.K. and X.S.; Writing – Original Draft, Y.W., H.X. and F.E.W.; Writing – Review & Editing, K.K., H.K. and X.S.; Visualization, Y.W. and H.X.; Supervision, H.K. and F.E.W.; Project Administration, Y.W., H.X. and F.E.W.; Funding Acquisition, X.S. and F.E.W.