Hepatic PKA Mediates Liver and Pancreatic α-Cell Cross Talk Free

Glucagon is a peptide hormone secreted from α-cells in the pancreas and serves as a counter-regulatory hormone to insulin. Glucagon is secreted in response to a decline in blood glucose level and activates the glucagon receptor (GCGR) in the liver to stimulate glycogenolysis and gluconeogenesis (1). Glucagon levels are often elevated in patients with type 2 diabetes (T2D), exacerbating hyperglycemia (2). To investigate therapeutic potential of glucagon action blockade to lower blood glucose in patients with diabetes, GCGR inhibitors, including small molecules, antisense oligonucleotides, and antagonist antibodies, have been developed. In clinical trials, these molecules showed glucose-lowering efficacies without hypoglycemia but were often associated with potential safety signals, including increased circulating levels of ALT, LDL cholesterol, amino acids, and/or liver fat (3–7).

One of the major safety concerns associated with GCGR inhibition is hyperaminoacidemia, because it may lead to pancreatic α-cell hyperplasia and tumor formation (8). Glucagon controls whole-body amino acid homeostasis via regulation of hepatic amino acid use. Glucagon promotes amino acid uptake by hepatocytes and their breakdown into gluconeogenesis precursors via increased expression of hepatic genes involved in amino acid catabolism and transport (9,10). Gcgr-null mice present decreased hepatic expression of amino acid catabolism genes and hyperaminoacidemia (11–13), demonstrating a tight link between liver amino acid catabolism and plasma amino acid levels. The effects of glucagon on amino acid metabolism described in mice are conserved in humans. Patients with pancreatectomy develop hypoglucagonemia and hyperaminoacidemia (14), whereas patients with glucagonoma exhibit hyperglucagonemia and hypoaminoacidemia (15). Recent reports by us and others established that hyperaminoacidemia causes α-cell hyperplasia observed in whole-body and liver-specific GCGR-knockout mice (16,17). Mouse and zebrafish studies uncovered that amino acids promote α-cell proliferation in a calcium-sensing receptor and mTOR-dependent process (10,18) and that amino acid transporter SLC38A5 relays the signal of hyperaminoacidemia to induce α-cell proliferation (19,20).

These data highlight the role of glucagon in regulating hepatic and whole-body amino acid homeostasis and indicate that a lack of glucagon signaling results in hyperaminoacidemia and α-cell hyperplasia. However, the hepatic GCGR signaling pathway that controls amino acid metabolism remains unclear. GCGR is a G-protein–coupled receptor that signals through G_sα and possibly through G_qα (21). G_sα seems to be the key downstream mediator regulating amino acid metabolism, because G_sα-knockout mice develop α-cell hyperplasia (22). Activated G_sα stimulates adenyl cyclase to produce cAMP, the rise of which activates downstream effectors. Two such downstream effectors have been described in hepatocytes: protein kinase A (PKA) and exchange protein directly activated by cAMP 2 (EPAC2), also known as Rap guanine nucleotide exchange factor 4 (RAPGEF4). PKA has been described as a critical effector of hepatic GCGR signaling controlling glucose metabolism, because it phosphorylates key metabolic enzymes and activates transcriptional programs for gluconeogenesis through its downstream transcription factor cAMP response element-binding protein (CREB). EPAC2 is another hepatic cAMP effector, reported to regulate glucose metabolism, membrane current, cell survival, and cholesterol metabolism, independently of PKA (23–26). The roles of PKA, EPAC2, or other effectors downstream of G_sα in hepatic amino acid metabolism are not fully understood. In this study, we used AAV8-shRNA to knock down liver GCGR, G_sα, PKA, and EPAC2 in diet-induced obese (DIO) and lean mice and examined hepatic expression profiles, plasma levels of amino acids and glucagon, and pancreas morphology. We report here that PKA is the major GCGR downstream effector that regulates amino acid metabolism and α-cell mass.

All procedures were conducted in compliance with protocols approved by the Institutional Animal Care and Use Committee of Regeneron Pharmaceuticals. Male DIO mice (cat. no. 380050; The Jackson Laboratory) and lean mice (cat. no. 000664; The Jackson Laboratory) were housed with one to five mice per cage in a controlled environment (12-h light/dark cycle; 22 ± 1°C), with free access to food and water. DIO mice were fed a high-fat diet (cat. no. D12492i; Research Diet) starting at 6 weeks of age for 20–22 weeks at the start of all studies, except the study evaluating the off-tissue effect of AAV8, in which mice were fed a high-fat diet for 29 weeks. Lean mice were fed a regular chow diet (cat. no. 5053; LabDiet) and were 13 weeks of age at the start of the study. All mice were sacrificed in the fed condition. Mice were assigned to experimental groups based on baseline body weight and blood glucose levels. Experimenters were not blinded during the studies.

AAV8 expressing a control scrambled shRNA or shRNAs against Gcgr, Gnas, Prkaca/b, Epac2 (Rapgef4), and Prkar1a, under the U6 promoter, was designed and generated by Vector Biolabs. Mice were injected with AAV8 particles intravenously via the tail vein at a dose of 1 × 10¹² (DIO) or 5 × 10¹¹ (lean) genome copies per mouse on day 0. For antibody treatment, mice were injected subcutaneously with GCGR or control antibody (10 mg/kg; diluted with sterile PBS) weekly. Body weight was monitored weekly. Fed blood glucose was measured in the morning via the AlphaTrak glucose monitoring system (Zoetis). Blood samples were collected via submandibular bleeds weekly into lithium heparin microtainers (cat. no. 365985; Becton Dickinson) and mixed with DPP4 and protease inhibitors (cat. no. DPP4-010; EMD Millipore; cat. no. 11836170001; Sigma). The blood samples were centrifuged at 10,000 rpm for 10 min at 4°C, and the supernatant plasma was collected for storage at −80°C until analysis.

Total plasma amino acid concentration, except for glycine, was measured with a colorimetric kit (cat. no. MAK002; Sigma). Individual amino acid concentration was measured with high-performance liquid chromatography by the Texas A&M University Protein Chemistry Laboratory. Plasma glucagon and insulin were measured by ELISA (cat. nos. 10-1281-01 and 10-1247-01; Mercodia).

Liver glycogen content was measured with a colorimetric kit (cat. no. MAK016; Sigma) according to the manufacturer’s instructions. Glycogen was extracted from the liver by homogenizing 10–50 mg frozen liver tissue in water (20 μL/mg) using the FastPrep-24 instrument (MP Biomedicals). The homogenates were boiled for 10 min, and insoluble material was removed after centrifugation at 18,000g for 10 min. Samples were diluted 20-fold (for PKA activation) or 100-fold before assay. Glycogen content was normalized to the weight of the liver used for extraction.

Liver samples were lysed with ice-cold radioimmunoprecipitation assay lysis buffer in the presence of protease and phosphatase inhibitors (cat. nos. 78430 and 78420; Thermo Fisher Scientific) using the FastPrep-24 instrument. For GCGR Western blot, liver samples were extracted with a G-coupled protein receptor extraction reagent (cat. no. A43436; Thermo Fisher Scientific) using a tissue grinder (cat. no. 47732; VWR International). Lysates for GNAS and PRKACA blotting were diluted with sample loading buffer and incubated for 5 min at 90°C, whereas lysates for GCGR and EPAC2 blotting were incubated at room temperature for 30 min. Samples were separated by SDS-PAGE, transferred to nitrocellulose membranes, and probed with the following antibodies: GCGR (cat. no. ab75240; Abcam), GNAS (cat. no. PA5-19315; Invitrogen), PRKACA (cat. no. PA5-17626; Invitrogen), and EPAC2 (cat. no. 43239; Cell Signaling Technology).

All tissues were collected in RNALater (cat. no. AM7021; Thermo Fisher Scientific) and stored frozen. RNA was extracted from the tissue with the GeneJET RNA purification kit (cat. no. K0732; Thermo Fisher Scientific) or RNeasy Plus Mini Kit (cat. no. 74134; Qiagen). Genomic DNA was removed using the TURBO DNA-free kit (cat. no. AM1907; Invitrogen). mRNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (cat. no. 4374966; Thermo Fisher Scientific). TaqMan analysis was performed in triplicate with TaqMan Fast Advanced Master Mix (2×; cat. no. A44360; Thermo Fisher Scientific) using the QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific). The following probes from Thermo Fisher Scientific were used: Gapdh (Mm99999915_g1), Gcgr (Mm00433546_m1), Gnas (Mm01242435_m1), Prkaca (Mm00660092_m1), Prkacb (Mm01312555_m1), Rapgef4 (Mm01327556_m1), and Prkar1a (Mm00660315_m1). Gapdh was used as the internal control gene for normalization. Expression of genes of interest was calculated relative to the Gapdh housekeeping gene and control group using the ΔΔ Ct method.

Pancreata were fixed in 10% neutral buffered formalin solution for 48 h, dehydrated in 70% ethanol, and then embedded in paraffin. Tissue sections 6 mm in thickness were prepared from paraffin-embedded blocks. Analysis for α- and β-cell mass was carried out as previously described (27).

Immunofluorescence was carried out using another set of sections from the same pancreata. Slides were stained for glucagon with human α-glucagon antibody (REGN745) and for insulin with guinea pig α-insulin antibody (cat. no. A0564; Dako) or mouse α-insulin antibody conjugated to AF488 (cat. no. 53-9769-82; Invitrogen). Slides were scanned with a Zeiss Axio Scan.Z1 slide scanner (Zeiss) with a 20× objective, and the images were analyzed with HALO (version 3.6; Indica Labs). Islets from each section were identified using the classifier function in HALO based on glucagon, insulin, and DAPI staining. The islet selections were then manually checked and adjusted. All identified islets were used for subsequent analyses. Glucagon-positive cells were counted with the CytoNuclear (version 2.0) analysis module (Indica Labs). α-Cell size was estimated by dividing the glucagon-positive cell area in the islets by the number of glucagon-positive cells in the islets. Glucagon-positive cell area was measured by a classifier in HALO trained to identify glucagon-positive cells. Mean fluorescence intensity of glucagon-positive pixels in the islets was measured using Area Quantification FL (version 1.0) analysis module (Indica Labs).

RNA extraction, library preparation, and analysis were performed as previously described (28). Differentially expressed genes were obtained using cutoffs P_adj < 0.05 and fold-change >1.5 with the DESeq2 package (29). Fold-change for an individual sample was calculated by dividing the transcripts per million (TPM) for each gene of each sample by the average TPM for each gene of the control group. The combined differentially expressed genes from the four knockdown (KD) groups from gluconeogenesis (cat. no. MM15392; Reactome), amino acid transporters (30), and amino acid catabolism (Gene Ontology identifier 0009063) are displayed in the heat map. All urea cycle genes are displayed in the heat map. The RNA sequencing (RNA-seq) data have been deposited in the Gene Expression Omnibus (GEO) under accession no. GSE268864.

Statistical and graphical data analyses were performed using Prism 9 (GraphPad Software, Inc.). Data are expressed as mean ± SEM. When two independent groups were present, mean values were compared with unpaired two-tailed t tests. When three or more independent groups were present, parameters were analyzed by one-way ANOVA when one independent variable was present or two-way ANOVA with post hoc Dunnett test when two independent variables were present; a threshold of P < 0.05 was considered statistically significant.

The RNAseq data have been deposited in GEO under accession no. GSE268864. This article does not report original code. Any additional information required to reanalyze the data reported in this article is available from the corresponding author on request.

To identify the GCGR signaling molecules mediating amino acid regulation, we used AAV8-shRNA to knock down liver Gcgr, Gnas, Prkaca, and Prkacb (two catalytic subunits of PKA) and Epac2 (Rapgef4) in DIO mice, a model of T2D, for which therapeutic utility of GCGR antagonism has been considered. Liver TaqMan analysis 4 weeks after AAV injection showed >70% mRNA reduction for all target genes, including Epac2 (Fig. 1A). Reduction in the target protein levels was confirmed by Western blotting (Supplementary Fig. 1). GCGR inactivation in DIO mice led to 30–45% reduction in ad lib glucose levels (6,20). Therefore, changes in blood glucose levels at any point were expected to reflect changes in GCGR KD efficiency of the time point. Starting 2 weeks postinjection, ad lib fed glucose levels were 30% reduced in mice with liver Gcgr KD and remained at that level until the end of study, suggesting that maximal KD by the shRNA was likely achieved for the last 2 weeks of the study (Fig. 1B). Similar glucose lowering was observed in mice with Gnas and Prkaca/b KD, but not in mice with Epac2 KD. (Fig. 1B). Body weight of mice with Prkaca/b KD was slightly lower than in the control group at week 4 (Supplementary Fig. 2A). Terminal liver glycogen content was not different between the groups (Supplementary Fig. 2B).

To examine the effect of target gene KD on the liver transcriptome, we conducted bulk RNA-seq analysis using the liver harvested at study termination. Consistent with the glucose lowering in mice with liver Gcgr, Gnas, and Prkaca/b KD, these mice showed reduced expression of key gluconeogenic genes, such as G6pc, Pck1, and Fbp1 (Fig. 1C). Genes involved in the catabolism and transport of amino acids and the urea cycle were largely downregulated in these mice, but not in mice with Epac2 KD (Fig. 1C). These results show that inactivation of GCGR, GNAS, and PKA limits hepatic amino acid use in mice, whereas EPAC2 does not affect liver amino acid metabolism.

We next tested if the amino acid metabolism gene changes observed in the liver of mice with Gcgr, Gnas, and Prkaca/b KD accompanied alterations in circulating amino acid levels and α-cell mass. Starting 2 weeks after AAV injection, mice with liver Gcgr, Gnas, and Prkaca/b KD showed elevated plasma amino acids levels, which persisted throughout the study (Fig. 2A).

Of the 19 measured amino acids, 10, seven, and six amino acids were increased significantly by liver Gcgr, Gnas, and Prkaca/b KD, respectively (Fig. 2B). α-Cell mass was increased in mice with liver Gcgr and Prkaca/b KD and had a trend toward increase in mice with Gnas KD (Fig. 2D and E). There were no differences in pancreas weight or β-cell mass between the groups (Fig. 2C and F). Circulating glucagon levels were more than 10-fold elevated in mice with liver Gcgr, Gnas, and Prkaca/b KD (Fig. 2G). Plasma insulin levels were not different between the groups (Fig. 2H). Changes in pancreas morphometry or circulating glucagon or insulin levels were not detected in mice with Epac2 KD compared with the control mice. These data demonstrate that PKA, not EPAC2, is the cAMP-responsive element downstream of GCGR-GNAS, controlling hepatic amino acid catabolism, plasma amino acid levels, and α-cell mass.

To assess the contribution of hepatic PKA in mediating GCGR regulation of amino acid homeostasis and α-cell mass, we activated PKA in the liver of DIO mice administered with GCGR-blocking antibody, which promotes hyperaminoacidemia and α-cell hyperplasia (6,20). We activated PKA via AAV8-shRNA–mediated KD of Prkar1a, the major inhibitory PKA regulatory subunit (31–33). The Prkar1a mRNA level was reduced by 80% in the liver of DIO mice receiving AAV (Fig. 3A) and not in other tissues tested (Supplementary Fig. 3). Increased hepatic PKA activity by this approach was confirmed by Western blotting for phosphorylated PKA substrates. Band intensities for phosphorylated PKA substrates increased in mice with Prkar1a KD, irrespective of the antibody treatment, whereas the signal decreased in mice treated with control AAV and GCGR antibody (Fig. 3B). Terminal body weight was higher for the control group than for groups with liver PKA activation alone or GCGR inhibition alone (Supplementary Fig. 4A). Blood glucose levels were decreased by GCGR antibody treatment alone, consistent with our previous report (6). Mice with PKA activation alone showed transient hyperglycemia, reaching a peak of 330 mg/dL at day 14, before returning to baseline glucose levels at day 28 (Fig. 3C). Mice with GCGR inhibition and PKA reactivation showed glucose lowering at day 7, when the KD of Prkar1a may have been incomplete, followed by transient hyperglycemia, similar to mice with Prkar1a KD alone (Fig. 3C). Terminal liver glycogen content was significantly lower in mice with Prkar1a KD, with or without GCGR inhibition (Supplementary Fig. 4B).

Terminal liver RNA-seq showed that gluconeogenic genes (Pck1, G6pc, Fbp1) were downregulated by GCGR antibody treatment alone and upregulated by PKA activation with or without GCGR antibody administration (Fig. 3D). These data confirm previously reported effects of GCGR inhibition or PKA activation on glucose homeostasis in mice (6,31–34). Next, effects of GCGR inhibition alone, PKA activation alone, or the combination of both on mRNA expression levels of hepatic amino acid metabolism genes were compared. Expression of genes involved in amino acid catabolism, amino acid transporters, and the urea cycle was decreased in mice with GCGR antibody treatment alone and increased in mice with liver PKA activation alone (Fig. 3D). Mice with both GCGR inhibition and PKA reactivation showed similar expression profiles to mice with PKA activation alone. These data demonstrate that PKA is the major signaling molecule mediating GCGR regulation of hepatic amino acid metabolism.

We assessed effects of GCGR inhibition alone, liver PKA activation alone, or the combination of both on plasma amino acid levels and pancreatic α-cells. Plasma amino acid levels were increased twofold by GCGR antibody treatment alone and 50% reduced by liver PKA activation alone (Fig. 4A). Mice with GCGR inhibition and PKA activation showed plasma amino acid levels similar to those in mice with PKA activation alone (Fig. 4A). GCGR antibody treatment increased the levels of all amino acids except aspartic acid and glutamic acid (cysteine was not measured), whereas PKA activation, alone or in combination with GCGR antibody, significantly reduced all amino acids except arginine, aspartic acid, glutamic acid, and lysine (Fig. 4B). Pancreas weight and α-cell mass was increased by GCGR antibody treatment alone and decreased by liver PKA activation alone or in combination with GCGR antibody treatment (Fig. 4C–E). β-Cell mass was unchanged across groups (Fig. 4F). Terminal plasma glucagon levels were increased by >25-fold in mice receiving GCGR antibody alone, whereas the levels were below the detection limit in mice with liver PKA activation alone or in combination with GCGR inhibition (Fig. 4G). Terminal plasma insulin levels were decreased in mice with GCGR antibody treatment, liver PKA activation, or the combination of both (Fig. 4H). Further pancreas analysis revealed that GCGR inhibition induced α-cell mass expansion accompanied by increases in the number and size of α-cells, as well as glucagon staining intensity (Fig. 5). Reduced α-cell mass observed with PKA activation alone, or in combination with GCGR inhibition, accompanied reductions in α-cell size and glucagon staining intensity without changes in α-cell number (Fig. 5). Collectively, these data demonstrate that PKA activation and GCGR inhibition have opposing effects on plasma amino acid levels and α-cell mass and that liver PKA activation fully reverses GCGR inhibition–induced hyperaminoacidemia and α-cell hyperplasia.

To test if our DIO mouse findings were relevant to lean mice, we knocked down liver Gcgr alone, Prkaca/b alone, or Prkar1a with or without GCGR antibody treatment. TaqMan analysis 4 weeks after AAV injections confirmed that the target gene mRNAs were knocked down >70% (Supplementary Fig. 5A). Blood glucose levels were reduced by Gcgr KD, Prkaca/b KD, and GCGR antibody treatment and were increased by Prkar1a KD (Supplementary Fig. 5B). Terminal body weight was significantly reduced only by Prkaca/b KD (Supplementary Fig. 5C). Plasma levels of total amino acids and glucagon were increased by Gcgr KD, Prkaca/b KD, and GCGR antibody treatment and were reduced by PKA activation with or without GCGR antibody treatment (Supplementary Fig. 5D and E). These data demonstrate that the dominant role hepatic PKA plays in amino acid metabolism is applicable to lean mice.

Liver GCGR signaling is well recognized as a critical regulator of amino acid homeostasis; however, GCGR downstream mediators essential to convey the signal are not well characterized. Reduced expression of hepatic amino acid catabolism genes, hyperaminoacidemia, α-cell hyperplasia, and hyperglucagonemia are common features of mice with whole-body or liver-specific GCGR inhibition (6,16,17,19,20,35). We now extend the same observation to mice with hepatic PKA inactivation. PKA inactivation in the mouse liver was achieved by knocking down the catalytic subunits of PKA without affecting liver Gcgr expression. The mice showed changes in hepatic amino acid metabolism gene expression, hyperaminoacidemia, α-cell hyperplasia, and hyperglucagonemia, all to the same extent as mice with hepatic GCGR inactivation. In contrast, mice with hepatic Epac2 KD did not show changes in plasma levels of amino acids or glucagon or α-cell mass. These data indicate that PKA is the primary downstream effector of GCGR-GNAS, controlling hepatic amino acid metabolism, and that EPAC2 does not play a role in the pathway.

To further investigate the contribution of PKA in GCGR regulation of amino acids, PKA was activated via inhibition of the major inhibitory PKA regulatory subunit in mice, and simultaneously, GCGR signaling was inhibited by administration of GCGR-blocking antibody. Mice with hepatic PKA activation alone developed hypoaminoacidemia, hypoglucagonemia, and reduced α-cell mass, which are opposite phenotypes of mice with hepatic GCGR or PKA inactivation. Mice receiving GCGR-blocking antibody with hepatic PKA activation showed identical phenotypes to mice with hepatic PKA activation alone, suggesting that PKA activation fully overrides the effect of GCGR inhibition on amino acids, glucagon, and α-cells. These data highlight PKA as the dominant effector of GCGR-GNAS, controlling hepatic amino acid metabolism and α-cell mass.

Although PKA has been considered the major effector of GCGR signaling downstream of GNAS, the other effector, EPAC2, has been reported to regulate diverse liver functions in response to glucagon. A recent study found that liver EPAC2 and its downstream factor RAP1 regulate glucose homeostasis (26), and the inhibition of EPAC2 leads to increased gluconeogenesis. In our DIO mouse model, liver Epac2 KD resulted in a small yet persistent increase in fed blood glucose levels (Fig. 1B), which agrees with the previous report, although gluconeogenesis gene expressions were not significantly altered. Our results suggest that EPAC2 does not affect hepatic amino acid catabolism or plasma amino acid levels.

Several amino acids are known to play critical roles in the liver–α-cell axis. It has been reported that alanine, glutamine, and arginine stimulate α-cell proliferation (19,20,36), whereas alanine, arginine, cysteine, and proline stimulate glucagon secretion from α-cells (37). Among the five key amino acids in the liver–α-cell axis, cysteine was not measured in this study. Significant increases or a trend toward such increases in the circulating levels of alanine, glutamine, arginine, and proline were observed in mice with Gcgr, Gnas, and Prkaca/b KD (Fig. 2B). Mice receiving GCGR antibody showed further elevation in these amino acids compared with mice with the KDs, consistent with greater α-cell mass and plasma glucagon levels (Fig. 4B, E, and G). Liver PKA activation in mice with or without GCGR blockade lowered most amino acids, including alanine, glutamine, and proline, similar to reported changes in mice receiving a GCGR agonist treatment (35) (Fig. 4B). Heterozygosity in a GCGR variant that reduces cAMP signaling does not accompany changes in plasma amino acid levels in humans (38), similar to normal plasma amino acid levels detected in GCGR^+/− mice (8). In this study, changes in plasma amino acids were observed in mice with 70–90% reductions in mRNA levels of Gcgr, Gnas, and Prkaca/b (Fig. 1A), suggesting the extent of GCGR-GNAS-PKA pathway inhibition required to induce detectable changes in plasma amino acids. Overall, the present mouse study demonstrates that the GCGR-GNAS-PKA pathway controls homeostasis of amino acids, including key amino acids in the liver–α-cell axis.

Mice with plasma amino acid changes in either direction showed broad hepatic expression changes in amino acid catabolism, transport, and the urea cycle genes. Certain plasma amino acid changes seemed to be a direct result of changes in genes that metabolize them; for example, mice with GCGR blockade and PKA activation/inactivation showed changes in plasma glutamine levels and expression levels of its catabolic enzyme Gls2 (Figs. 1C, 2,B, 3,D, and 4,B). In contrast, changes in the plasma alanine and proline levels of these mice did not accompany expression changes in their catabolic enzymes Gpt1/2, Prodh1/2, and Aldh4a1 (Figs. 2B and 4,B), suggesting additional mechanisms may influence their plasma levels. Posttranscriptional regulation of catabolic enzymes represents a potential mechanism, because acute and posttranscriptional regulation of amino acid metabolism by glucagon have been reported. Glucagon increases plasma amino acid clearance and ureagenesis within minutes in mice (12) and stimulates two urea cycle enzymes, CPS1 and OTC, by promoting their deacetylation (39). In another study, CPS1 activity was increased with GCGR activation and vice versa without changes in CPS1 mRNA levels (35). Whether PKA regulates amino acid metabolism posttranscriptionally and/or acutely remains to be tested.

Decreases in pancreatic α-cell mass and proglucagon transcription have been reported in mice receiving GCGR agonist and rats and mice with glucagonoma implants (35,40,41). Our data suggest that the observed α-cell phenotypes in these rodents were likely a result of PKA-mediated increases in amino acid catabolism and hypoaminoacidemia. α-Cell atrophy and reduced glucagon production in mice with liver PKA activation (Fig. 5A, C, and D) are consistent with prior observations in glucagonoma rodent models (40,41). A decrease in α-cell number observed in glucagonoma rodent models (40,41) was not observed in mice with liver PKA activation (Fig. 5B). The reduced pancreas mass observed in mice with PKA activation is reminiscent of the reported data in rats with glucagonoma implantation (40). In both studies, reductions in α-cell mass did not fully account for the extent of pancreas mass loss, indicative of diminished exocrine pancreas mass. Because cell contraction and increased apoptosis were observed in the exocrine pancreas of rats with glucagonoma implantation (40), similar abnormalities might have developed in the pancreas of mice with hepatic PKA activation.

Studies suggest that a liver–α-cell axis also exists in human. Hyperaminoacidemia, notably elevated glutamine, alanine, arginine, lysine, ornithine, threonine, and serine, has been documented in an individual with a GCGR defect resulting from biallelic mutations in Gcgr (42). Hyperglucagonemia and α-cell hyperplasia have been reported in multiple individuals with Gcgr-inactivating mutations (43–46). In contrast, patients with glucagonoma develop hypoaminoacidemia (15). Glucagon infusion in healthy individuals reduces plasma amino acid levels, and this effect is blunted in obese individuals with hepatic steatosis, likely because of glucagon resistance (47–50). GCGR agonism in combination with GLP-1R agonism has been considered a potential therapeutic option for obesity, T2D, and metabolic dysfunction–associated steatotic liver disease (51). Clinical studies evaluating GCGR/GLP-1R coagonists have documented decreases in key amino acids in the liver–α-cell axis, such as alanine, glutamine, arginine, and proline, in patients receiving such coagonists (52,53). Whether GCGR agonism affects the human pancreas in a similar fashion to hepatic PKA activation in mice remains to be seen.

In this study, reductions in blood glucose levels and hepatic expression of gluconeogenic genes were detected in mice with hepatic GCGR or PKA inactivation, whereas transient hyperglycemia and increased hepatic expression of gluconeogenic genes were observed in mice with hepatic PKA activation (Figs. 1B and C and 3,C and D). Because GCGR signaling promotes use of amino acids for hepatic glucose production, glycemic modulations in mice with GCGR-GNAS-PKA inactivation or activation may be at least partially explained by reduced/increased amino acid catabolism in the liver. Indeed, increased liver alanine catabolism has been shown to promote hyperglycemia in T2D mouse models (54). Hyperglycemia, accompanying hypoaminoacidemia, and increased hepatic amino acid catabolism have been reported in mice with hepatic liver kinase B1 (LKB1) deletion (55). LKB1 suppression of amino acid catabolism is posttranscriptional and associated with protein posttranslational modifications. Whether LKB1 lies downstream of GCGR-GNAS-PKA and serves as a brake on amino acid catabolism promoted by GCGR-GNAS-PKA signaling is of interest for future research.

It is unclear if known downstream regulators of GCGR-GNAS-PKA on glucose and/or lipid homeostasis have roles in the amino acid homeostasis control by hepatic PKA. CREB is a transcriptional factor activated by PKA phosphorylation and an important regulator of glucose and lipid metabolism genes in the liver (56). Inhibition of either PKA or CREB in primary mouse hepatocytes blunts glucagon-induced expression of amino acid catabolism genes, indicating the potential role CREB plays in hepatic amino acid metabolism control (9). Whether hepatic CREB inhibition or activation affects amino acid homeostasis and α-cell mass in an animal model remains to be tested. CREB-regulated transcription coactivator 2 (CRTC2) is another PKA downstream factor implicated in the regulation of hepatic amino acid metabolism (56). Rats with ASO-mediated silencing of hepatic CRTC2 show reduced expression of several hepatic amino acid catabolism genes and increased circulating levels of glucagon, but no changes in circulating levels of amino acids (57). Identification of hepatic of GCGR-GNAS-PKA downstream factors in the control of amino acid and α-cell homeostasis in vivo will require additional studies.

One limitation of the study is that we did not prove causality between hypoaminoacidemia and the reduction in α-cell mass in mice with hepatic PKA activation. Elevating circulating amino acids in mice with liver PKA activation in a future study may clarify this point. Another limitation is that only male mice were used in the study.

In conclusion, we demonstrate that GCGR signals through PKA to regulate amino acid catabolism in the liver and circulating amino acid levels, which is a potent growth factor on pancreatic α-cells. Hepatic PKA plays a dominant role in liver–α-cell cross talk, mediated by circulating glucagon and amino acids.

Acknowledgments. The authors thank Megan Binn and Lawrence Miloscio for help with tissue collection; Biin Sung, Jinrang Kim, Judith Altarejos, Megan Binn, Neha Shrestha, and Nhu Truong for critical evaluation of the manuscript (all from Regeneron Pharmaceuticals). The authors also thank Regeneron DNA Core/Molecular Profiling for assistance with RNA-seq.

Duality of Interest. This study was funded by Regeneron Pharmaceuticals. All authors are employees and shareholders of Regeneron Pharmaceuticals. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. K.B., J.B., and E.N. performed investigations. K.B., J.B., E.N., and Q.S. performed formal analyses. K.B. and H.O. were responsible for conceptualization and wrote the original article draft. G.H., M.S., and H.O. supervised the study. All authors reviewed and edited the article. H.O. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Some of the data were presented as a poster at the Federation of American Societies for Experimental Biology Molecular Metabolism: Cells and Systems Science Research Conference, 14–18 July 2024.