Inhibition of hepatic gluconeogenesis in type 2 diabetes by metformin: complementary role of nitric oxide

Abstract

Metformin is the first-line treatment for type 2 diabetes mellitus. Type 2 diabetes mellitus is associated with decreased nitric oxide bioavailability, which has significant metabolic implications, including enhanced insulin secretion and peripheral glucose utilization. Similar to metformin, nitric oxide also inhibits hepatic glucose production, mainly by suppressing gluconeogenesis. This review explores the combined effects of metformin and nitric oxide on hepatic gluconeogenesis and proposes the potential of a hybrid metformin-nitric oxide drug for managing type 2 diabetes mellitus. Both metformin and nitric oxide inhibit gluconeogenesis through overlapping and distinct mechanisms. In hepatic gluconeogenesis, mitochondrial oxaloacetate is exported to the cytoplasm via various pathways, including the malate, direct, aspartate, and fumarate pathways. The effects of nitric oxide and metformin on the exportation of oxaloacetate are complementary; nitric oxide primarily inhibits the malate pathway, while metformin strongly inhibits the fumarate and aspartate pathways. Furthermore, metformin effectively blocks gluconeogenesis from lactate, glycerol, and glutamine, whereas nitric oxide mainly inhibits alanine-induced gluconeogenesis. Additionally, nitric oxide contributes to the adenosine monophosphate-activated protein kinase-dependent inhibition of gluconeogenesis induced by metformin. The combined use of metformin and nitric oxide offers the potential to mitigate common side effects. For example, lactic acidosis, a known side effect of metformin, is linked to nitric oxide deficiency, while the oxidative and nitrosative stress caused by nitric oxide could be counterbalanced by metformin’s enhancement of glutathione. Metformin also amplifies nitric oxide -induced activation of adenosine monophosphate-activated protein kinase. In conclusion, a metformin-nitric oxide hybrid drug can benefit patients with type 2 diabetes mellitus by enhancing the inhibition of hepatic gluconeogenesis, decreasing the required dose of metformin for maintaining optimal glycemia, and lowering the incidence of metformin-associated lactic acidosis.

Introduction

Type 2 diabetes mellitus (T2DM) is characterized by carbohydrate, lipid, and protein metabolism dysregulation due to impaired insulin secretion and insulin resistance.1 In 2021, approximately 483 million people were diagnosed with T2DM globally, representing 9.45% of the population,2 and it remains a leading cause of mortality, claiming over a million lives each year.3 Currently, metformin is the first-line standard treatment for T2DM and the most common glucose-lowering medication,4 with more than 200 million patients using it daily.4 Metformin decreases both basal and postprandial plasma glucose levels.5 The glucose-lowering effects of metformin are mediated by increasing skeletal muscle glucose uptake and utilization,6 lowering intestinal carbohydrate absorption,6 and suppressing glucose production6 mainly by inhibiting hepatic gluconeogenesis,7,8 which is increased in T2DM.1 The activities of cytoplasmic enzymes that participate in gluconeogenesis are increased in rats with diabetes.9 Increased hepatic gluconeogenesis in T2DM is attributed to insulin resistance, insulin deficiency, hyperglucagonemia, increased glucagon sensitivity, and elevated substrate supply for gluconeogenesis.1

Animal studies indicate that nitric oxide (NO) supplementation has beneficial metabolic effects in T2DM by increasing insulin secretion10,11 (via increasing insulin synthesis and exocytosis,12 decreasing pancreatic islet oxidative stress,13 and increasing pancreatic islet blood flow10) and by stimulating peripheral glucose utilization (via increasing insulin sensitivity14 and increasing glucose uptake in skeletal muscle cells14 and adipocytes15). This issue has been reviewed elsewhere.16,17,18 Additionally, similar to metformin, NO inhibits hepatic glucose production mainly by inhibiting gluconeogenesis.19,20 The mechanisms underlying this effect include the suppression of activities of cytoplasmic phosphoenolpyruvate carboxykinase (PEPCK1),21,22 glucose 6-phosphatase (G6Pase),23 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH)24 by NO. Since both metformin and NO inhibit hepatic gluconeogenesis, a metformin-NO hybrid drug may offer superior glycemic control for patients with T2DM. Supporting this hypothesis, evidence shows that combination therapy with metformin plus nitrate nearly doubles the efficacy of reducing glycated hemoglobin levels and enhancing adenosine monophosphate-activated protein kinase (AMPK) activity, indicated by the phosphorylated-AMPK/AMPK ratio, in high-fat diet-fed mice compared to monotherapy with either agent alone.25 Furthermore, in a rat model of T2DM, combination therapy with metformin and L-arginine (an NO precursor) reduces C-reactive protein levels by about 75%, which is higher than the effect observed by metformin alone (~50%) and L-arginine alone (~25%).26 This review explores how combining metformin with NO-releasing agents can improve hepatic gluconeogenesis suppression, suggesting the potential efficacy of a metformin-NO hybrid drug to manage T2DM more effectively than metformin alone.

Search Strategy

In this narrative review, we conducted a literature search using the PubMed, Google Scholar, and Scopus databases to identify relevant studies on metformin and NO-releasing agents in hepatic gluconeogenesis. The keywords used included type 2 diabetes, gluconeogenesis, metformin, nitric oxide, nitric oxide synthase, and hybrid drug. In addition, the references cited in the retrieved articles were reviewed to identify further related studies. The included studies were primarily conducted on mammals, such as rodents and humans. Only articles written in English were considered.

Hepatic Gluconeogenesis: An Overview

Hepatic gluconeogenesis accounts for about 90% of glucose production during fasting.27 Key precursors for this process include lactate,28 glycerol,28 all amino acids (except leucine and lysine, which are only degraded to acetyl-coenzyme A or acetoacetate),29 and all intermediates of the tricarboxylic acid (TCA) cycle.30 However, lactate, glycerol, alanine, and glutamine are the main precursors for hepatic gluconeogenesis.28 Besides glycerol, which contributes to gluconeogenesis in later stages, other main gluconeogenesis precursors flow through oxaloacetate, indicating the crucial role of oxaloacetate as a metabolic intermediate in gluconeogenesis.30

Figure 1 illustrates the intricate nature of gluconeogenesis, where most reactions are largely the reverse of those in glycolysis, except for three irreversible glycolytic reactions. Reactions 1 [(glucose→glucose-6-phosphate (G6P)], 3 [fructose-6-phosphate (F6P)→fructose-1,6-bisphosphate (F1,6BP)], and 10 [phosphoenolpyruvate (PEP)→pyruvate)]. These reactions require the involvement of additional enzymes/mechanisms to be reversed, allowing the progression of hepatic gluconeogenesis.29 The conversion of pyruvate to PEP, a critical step in gluconeogenesis, is particularly complex. It involves several enzymatic pathways that shuttle intermediates between the mitochondria and cytoplasm.31

Figure 1.

An overview of hepatic gluconeogenesis.

The malate (blue arrows), direct (orange arrows), fumarate (green arrows), and aspartate (red arrows) pathways for exporting oxaloacetate from the mitochondrion to the cytoplasm. Transparent green and red arrows show the general direction of glycolysis and gluconeogenesis, respectively. In hepatic gluconeogenesis, most reactions are the reverse of those in glycolysis, except three irreversible reactions of glycolysis, reactions 1 (glucose→G6P), 3 (F6p→F1,6BP), and 10 (PEP→ pyruvate). Created with BioRender.com. 1,3BPG: 1,3-Bisphosphoglycerate; 2PG: 2-phosphoglycerate; 3PG: 3-phosphoglycerate; ACON: aconitase; ADP: adenosine diphosphate; AGC2: aspartate-glutamate carrier 2; ALT: alanine aminotransferase; AMP: adenosine monophosphate; ARG1: arginase; ASL: argininosuccinate lyase; ASS: argininosuccinate synthase; AST: aspartate aminotransferase; ATP: adenosine triphosphate; CIC: citrate carrier; CS: citrate synthase; CP: carbamoyl phosphate; CPS: carbamoyl phosphate synthetase; cNOS: constitutive nitric oxide synthase; DHAP: dihydroxyacetone phosphate; DIC: dicarboxylate carrier; EX: anion transporter; FAD: flavin adenine dinucleotide; FADH: flavin adenine dinucleotide (reduced form); FH: fumarase; F1,6BP: fructose 1,6-bisphosphate; F1,6BPase: fructose 1,6-bisphosphatase; F2,6B: fructose 2,6-bisphosphate; F6P: fructose 6-phosphate; G3P: glycerol 3-phosphate; G6P: glucose-6-phosphate; G6Pase: glucose 6-phosphatase; GADP: glyceraldehyde 3-phosphate; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; GCK: glucokinase; GDH: glutamate dehydrogenase; GDP: guanosine diphosphate; GK: glycerol kinase; GLAST: glutamate-aspartate antiporter; GLS: glutaminase; GPDH: glycerol 3-phosphate dehydrogenase; GTP: guanosine triphosphate; IDH: isocitrate dehydrogenase; KGDH: ketoglutarate dehydrogenase; LDH: lactate dehydrogenase; malate-α-KG antiporter: malate-α-ketoglutarate antiporter; MCT1: monocarboxylate transporter 1; MDH: malate dehydrogenase; MGT: mitochondrial glutamine transporter; MPC: mitochondrial pyruvate carrier; NAD⁺: nicotinamide adenine dinucleotide; NADH: nicotinamide adenine dinucleotide (reduced form); OGC: 2-oxoglutarate carrier; ORNT1: mitochondrial ornithine transporter; OTC: ornithine transcarbamylase; PC: pyruvate carboxylase; PEP: phosphoenolpyruvate; PEPCK: phosphoenolpyruvate carboxykinase; PFK1: phosphofructokinase 1; PGI: phosphoglucose isomerase; PGK: phosphoglycerate kinase; PGM: phosphoglycerate mutase; PK: pyruvate kinase; SCS: succinyl coenzyme-A synthetase; SDH: succinate dehydrogenase; SNAT: sodium-coupled neutral amino acid transporters; TCA: tricarboxylic acid; TPI1: triosephosphate isomerase 1.

Lactate, mainly produced in the skeletal muscle, enters the hepatocyte through monocarboxylate transporter 1.28,32 Then, it is converted to pyruvate in a reaction catalyzed by lactate dehydrogenase (LDH), which reduces nicotinamide adenine dinucleotide (NAD⁺) to nicotinamide adenine dinucleotide reduced form (NADH).28 Alanine, mainly produced in the skeletal muscle, is transported into hepatocytes through sodium-coupled neutral amino acid transporters (SNAT) 2 and 433; transamination of L-alanine by alanine aminotransferase then converts it to pyruvate.34 Pyruvate enters the mitochondria by the mitochondrial pyruvate carrier35 and is converted to oxaloacetate by pyruvate carboxylase in the mitochondrial matrix28 (Figure 1).

Glutamine, another major precursor of gluconeogenesis, mainly derived from the skeletal muscle,36 is transported into the hepatocyte via SNAT2, SNAT3, and SNAT5.33 Then, it enters the mitochondria through the mitochondrial glutamine transporter.37 Glutamine is converted to glutamate by glutaminase in mitochondria, consuming H₂O to produce ammonium (NH₄⁺). Then, it is converted to α-ketoglutarate by glutamate dehydrogenase coupled with the reduction of NAD⁺ to NADH and consuming H₂O to produce NH₄⁺. α-Ketoglutarate enters the TCA cycle to produce oxaloacetate28,30,38 (Figure 1).

The mitochondria cannot directly transport oxaloacetate into the cytoplasm.28 Thus, there are four primary pathways for its transport during gluconeogenesis31 (Figure 1): malate, direct, aspartate, and fumarate. As shown in Table 1, the contribution of each pathway is determined by (1) the fasting stage and its effect on the mitochondrial NADH/NAD⁺ ratio, (2) the gluconeogenesis substrate and the different enzymes it engages to provide the cytoplasmic NADH that GAPDH requires in the later stages of gluconeogenesis, and (3) accumulation of NH₄⁺ in the hepatocytes that forces the urea cycle.29,31,39

Table 1.

Comparison of pathways for exporting oxaloacetate from mitochondrion to cytoplasm during hepatic gluconeogenesis

Pathways	Contribution to fasting stages		Substrate dependency	NADH provision		Energy provision
	Early stage	Late stage		Mitochondrial	Cytoplasmic
Malate	High	Low	None	TCA cycle	MDH1	Independent
Direct	Low	High	Lactate	Independent	LDH	SCS (GTP)
Aspartate	Low	Low	Lactate	Independent	LDH	Independent
Fumarate	Low	Low	Glutamine/AA	GDH/TCA cycle	MDH1	Independent

AA: Amino acid; GDH: glutamate dehydrogenase; GTP: guanosine triphosphate; LDH: lactate dehydrogenase; MDH1: cytoplasmic malate dehydrogenase; NADH: nicotinamide adenine dinucleotide (reduced form); SCS: succinyl coenzyme-A synthetase; TCA: tricarboxylic acid.

Malate pathway

In the malate pathway, mitochondrial oxaloacetate is converted into malate coupled with the oxidation of NADH to NAD⁺ by the mitochondrial malate dehydrogenase (MDH2).29 The same enzyme contributes to the TCA cycle by driving the conversion of malate to oxaloacetate paired with the reduction of NAD⁺ to NADH.29 Mitochondrial malate is transported to the cytoplasm by dicarboxylate carrier, 2-oxoglutarate carrier, and citrate carrier.40 Then, it is converted to oxaloacetate by cytoplasmic malate dehydrogenase (MDH1), providing NADH that GAPDH needs in the later steps of gluconeogenesis.28,31 The malate pathway can be considered the main pathway in the earlier stages of fasting when the mitochondrial NADH/NAD⁺ ratio is still high.39 However, this pathway becomes limited during the later stages of fasting due to a decreased mitochondrial NADH/NAD⁺ ratio because of NADH consumption by both the electron transport chain (ETC) and MDH2.28,39 As a result, the malate-aspartate shuttle (MAS) becomes activated (see below) to replenish mitochondrial NADH41 (Figure 1). This shift in metabolic activity alters MDH2, making it more efficient at converting malate back to oxaloacetate,39,42 which hampers malate export. Five-day starvation decreases MDH1 activity by 16% in rat liver homogenates receiving mono, di, and TCA precursors.9 In the malate pathway, MDH1 provides the cytoplasmic NADH, rendering the pathway independent of lactate and glycerol for providing NADH via NADH-producing enzymes of cytoplasm, LDH, and cytoplasmic glycerol 3-phosphate dehydrogenase (GPDH1). Thus, the malate pathway is considered substrate-independent31 (Table 1).

In hepatocytes, the MAS in particular, and to a lesser extent, α-glycerophosphate maintain the redox balance between mitochondria and cytoplasm during gluconeogenesis.41 MAS maintains a balance between mitochondrial and cytoplasmic NADH/NAD⁺ ratio by modulating the flux through malate-oxaloacetate-aspartate between mitochondria and cytoplasm.42 To transport cytoplasmic NADH to mitochondria, MAS recruits MDH1 to convert oxaloacetate to malate, paired with the oxidation of NADH to NAD⁺, malate-α-ketoglutarate antiporter to transport malate into mitochondria, MDH2 to convert malate to oxaloacetate, paired with the reduction of NAD⁺ to NADH, mitochondrial aspartate aminotransferase (AST) to convert oxaloacetate to aspartate, glutamate-aspartate antiporter to transport aspartate to the cytoplasm, and cytoplasmic AST to convert aspartate back to oxaloacetate41,42 (Figure 1).

Direct pathway

In the direct pathway, oxaloacetate is directly converted to PEP by mitochondrial PEP carboxykinase (PEPCK2) at the cost of guanosine triphosphate (GTP) dephosphorylation to guanosine diphosphate.28,31 PEP is transported to the cytoplasm by an anion transporter, exchanging PEP with another metabolite specific to each transporter43 (Figure 1). The direct pathway is mainly used in the later stages of fasting when the mitochondrial NADH/NAD⁺ ratio is further decreased. It depends on mitochondrial GTP, produced by succinyl coenzyme-A synthetase (SCS) during the TCA cycle.39,44 This pathway contrasts with the malate pathway, which is more active in earlier stages of fasting when NADH levels are higher and the malate-aspartate shuttle is not as necessary.39 In the later stages of fasting, the direct pathway compensates for the malate pathway’s incapacitation in exporting oxaloacetate to the cytoplasm.39 Unlike the other three pathways for exporting oxaloacetate, the direct pathway is MAS-independent, making it feasible to export oxaloacetate in the presence of GTP. Lactate is also required as a gluconeogenic substrate to provide the cytoplasmic NADH for GAPDH by LDH. Thus, the direct pathway compensates for the limitations of the malate pathway in exporting oxaloacetate in these conditions, making it both energy- and substrate-dependent since it relies on both mitochondrial GTP and lactate as a gluconeogenic substrate. This pathway ensures continued gluconeogenesis despite changes in mitochondrial redox states during fasting28,31 (Table 1).

Aspartate pathway

In the aspartate pathway, mitochondrial oxaloacetate is converted to aspartate via mitochondrial AST.40 Aspartate-glutamate carrier 2 transports aspartate to the cytoplasm in exchange of glutamate.40 In the cytoplasm, AST converts aspartate to oxaloacetate40 (Figure 1). The aspartate pathway contributes less to gluconeogenesis than the direct and the malate pathways.9,39 In support of this notion, the activity of cytoplasmic AST was lower than MDH1 by 8-fold (23 vs. 198 μmol/min/g liver) and 4-fold (44 vs. 166 μmol/min/g liver) in the fed and fasting states, respectively.9 Enzyme activities were measured during gluconeogenesis in the soluble fraction of rat liver homogenates supplied with mono-, di-, and TCA precursors.9 The aspartate pathway is substrate-dependent and requires lactate as the substrate for gluconeogenesis to produce cytoplasmic NADH by LDH.31 Regarding the absence of production of NADH and NH₄⁺ in mitochondria through this pathway, activation of LDH is needed to provide the NADH that is required by GAPDH in later stages29 (Table 1).

Fumarate pathway

In the fumarate pathway, mitochondrial oxaloacetate is converted to aspartate by AST and transported to the cytoplasm by aspartate-glutamate carrier 2. Cytoplasmic aspartate then contributes to the urea cycle, where argininosuccinate synthetase catalyzes the formation of argininosuccinate from aspartate and citrulline, consuming two adenosine triphosphate (ATP) equivalents. In the next step, argininosuccinate lyase uses argininosuccinate to produce fumarate and arginine. Fumarate is then converted to malate by cytoplasmic fumarase. MDH1 converts malate to oxaloacetate coupled with the reduction of NAD⁺ to NADH in the cytoplasm29,31,40 (Figure 1).

The fumarate pathway is the main pathway when glutamine and other amino acids are used as substrates for gluconeogenesis.31 This pathway emphasizes the interaction of the urea cycle with gluconeogenesis and the dual role of glutamine, serving as a gluconeogenesis precursor and an NH₄⁺ carrier.31,45 Glutamate accepts free NH₄⁺ derived from protein catabolism in different organs. Then, it is converted to glutamine, which the liver utilizes to convert the toxic NH₄⁺ into urea through the urea cycle.45 In the liver, other amino acids taken by hepatocytes are degraded to the TCA cycle intermediates, losing their amine group as NH₄⁺. Their carbonic skeleton flows through the TCA cycle to attain oxaloacetate.29 Toxic accumulation of NH₄⁺ recruits the fumarate pathway, which forces aspartate to enter the urea cycle, converting NH₄⁺ to urea through the urea cycle.31 Also, glutamate dehydrogenase supplies NADH in this pathway.30 Therefore, the substrate-dependency of the fumarate pathway is related to NH₄⁺ accumulation and not cytoplasmic NADH-producing enzymes.31 Supplementation of rat liver homogenates with di and TCA precursors in the presence of aspartate as the only source of amino acid showed the following results. The activity of fumarase is almost one-eighteenth of MDH1 in the fed state (198 vs. 10.9 μmol/min/g liver for MDH1 and fumarase, respectively) and one-eighth in fasting states (166 vs. 20 μmol/min/g liver for MDH1 and AST, respectively).9 These findings indicate that except for when amino acids are the primary substrate for gluconeogenesis, the contribution of the fumarate pathway to gluconeogenesis is low (Table 1).

Except for the direct pathway, in which PEP is directly transported to the cytoplasm, within other pathways, the cytoplasmic oxaloacetate is converted to PEP by PEPCK1.30 Cytoplasmic PEP is converted to 2-phosphoglycerate by enolase; 2-phosphoglycerate is further converted to 3-phosphoglycerate by phosphoglycerate mutase.28 Phosphoglycerate kinase converts 3-phosphoglycerate into 1,3-bisphosphoglycerate, coupled with the conversion of ATP to adenosine diphosphate.28 The NADH produced in the cytoplasm by MDH1 and LDH is used by GAPDH to transform 1,3-bisphosphoglycerate into glyceraldehyde 3-phosphate (GADP).28 The reversible conversion of GADP into dihydroxyacetone phosphate (DHAP) is catalyzed by triose phosphate isomerase 128 (Figure 1).

Glycerol, produced from adipose tissue lipolysis, is transported into liver cells through aquaporin-9. Then, it is converted to glycerol 3-phosphate via glycerol kinase46 and becomes DHAP by fluxing through the α-glycerophosphate shuttle. Depending on the mitochondrial redox state, the conversion of glycerol 3-phosphate to DHAP is catalyzed either by mitochondrial glycerol 3-phosphate dehydrogenase (GPDH2) or GPDH1.30 GPDH2 is located in the inner mitochondrial membrane and pairs DHAP formation with the conversion of FAD to FADH₂ within the mitochondria. GPDH1 couples DHAP formation and the reduction of NAD⁺ to NADH in cytoplasm.30,47 Aldolase combines DHAP and GADP into F1,6BP, which becomes F6P by fructose 1,6-bisphosphatase (F1,6BPase) and is subsequently transformed into G6P by phosphoglucose isomerase. Finally, within the endoplasmic reticulum, G6P becomes glucose via G6Pase28 and exits from hepatocytes by GLUT248 (Figure 1).

Nitric Oxide Synthesis in Hepatocytes

In the human body, NO is produced by (1) NO synthase (NOS)-dependent L-arginine pathway49 and (2) nitrate-nitrite-NO pathway.50,51 Almost 90% of the body’s NO synthesis is related to the L-arginine-NO pathway, with the nitrate-nitrite-NO pathway responsible for the other 10%.52 NO is produced from L-arginine by three isoforms of the enzyme NOS.53 Two of them, i.e., endothelial NOS (eNOS) and neural NOS (nNOS), are constitutively expressed and are further activated following intracellular calcium rise.53 The third isoform, inducible NOS (iNOS), is expressed following transcriptional activation by cytokines, such as interferon-gamma and lipopolysaccharide (LPS), interleukin-1, interleukin-2, and tumor necrosis factor-alpha.54,55

The L-arginine–NOS pathway of NO production is oxygen-dependent, whereas the nitrate-nitrite-NO pathway is further activated during hypoxia and acidosis.56 Within the nitrate-nitrite-NO pathway, nitrate reduction by bacterial nitrate reductases in the oral cavity and gastrointestinal tract produces nitrite.56 In addition, studies on germ-free mice displayed the presence of a mammalian nitrate reductase in different organs,57 registering the colon as the site of the highest activity, followed by the stomach, kidney, small intestine, liver, heart, and lung.57 NO oxidation through auto-oxidation in tissues or via ceruloplasmin in plasma also produces nitrite.56 Nitrite is reduced to NO in blood and tissues by deoxyhemoglobin/myoglobin, cytoglobin, xanthine oxidoreductase, aldehyde oxidase, polyphenols, ascorbate, and protons.56,58,59

A study on the human liver using the citrulline assay, Western immunoblots, immunohistochemistry, and in situ hybridization shows the expression of eNOS and iNOS but not nNOS in hepatocytes.60 eNOS is also expressed in the endothelium of hepatic arteries, terminal hepatic venules, sinusoids, and the biliary epithelium of the human liver.60 iNOS is expressed in Kupffer cells of normal human liver tissue from patients undergoing partial hepatectomy for liver cancer.61 In addition, studies on the livers of iNOS knockout (iNOS^−/−) mice indicated the iNOS expression in the liver Kupffer cells.62,63,64 Hepatocyte iNOS is induced in response to Kupffer cell activation, reactive oxygen species generation, endotoxin, hepatic ischemia/reperfusion, and interleukin-1β.65,66,67 iNOS-containing hepatocytes are mainly localized in the periportal zone of the liver acinus in humans.60

The expression of nNOS in the liver is controversial in animal studies; some studies have reported the expression and activation of nNOS in mice’s primary isolated hepatocytes,68,69 and others have shown no expression of nNOS in rat hepatocytes.70 Immunohistochemistry of nNOS in rat liver revealed the presence of nNOS only in the liver endothelium of normal rats. In contrast, in type 1 diabetic and LPS-treated rats, nNOS presence was detected in hepatocytes, endothelium, and Kupffer cells.71

Given the inducible nature of iNOS and the lack of conclusive studies on the presence of nNOS in the liver, it appears that eNOS is the predominant NOS isoform in the liver and the main isoform contributing to glucose metabolism in the normal state.60 In support of this notion, hyperinsulinemic-euglycemic clamp studies after a 24-hour fast during an insulin infusion in eNOS knockout (KO) and nNOS knockout mice demonstrated hepatic insulin resistance only in the eNOS knockout group.72

Effects of Metformin on Gluconeogenesis in Hepatocytes

Enterocytes uptake metformin through the apical membrane via plasma membrane monoamine transporter and organic cation transporter 3 (OCT3), then transport it through the basolateral membrane to the portal vein by OCT1.73 The liver is the primary site of action for metformin,7 and metformin uptake in hepatocytes is mainly through OCT1.74 OCT3 and multidrug and toxin extrusion protein 1 transporter also contribute to hepatocyte metformin uptake.47 Evaluation of metformin effects on patients with T2DM by a combination of isotope dilution, indirect calorimetry, bioimpedance, and tissue-balance techniques indicates that metformin decreases hepatic glucose production mainly through inhibition of gluconeogenesis.8 The inhibitory effect of metformin on gluconeogenesis in type 2 diabetic patients has also been reported in a clinical trial using stable isotope tracers.75

As shown in Figure 2, the main mechanisms underlying the inhibitory effects of metformin on gluconeogenesis include (1) increasing the ratio of glutathione reductase (GR)-to-glutathione peroxidase (GPx) activities,76 (2) inhibition of mitochondrial oxidative phosphorylation by (a) inhibition of complex I and complex IV of the ETC76 and (b) inhibition of mitochondrial GPDH2.76 These effects increase NADH/NAD⁺ and AMP/ATP ratios; (3) induction of sirtuin silent information regulator 1 (SIRT1) and general control non-depressible 5 (GCN5) transcriptional regulators.77 These mechanisms eventually inhibit liver gluconeogenesis by transcriptional, allosteric, substrate-specific, and redox-dependent mechanisms.47

Figure 2.

Mechanisms underlying the inhibitory effect of metformin on hepatic gluconeogenesis.

Metformin inhibits hepatic gluconeogenesis by transcriptional, allosteric, substrate-specific, and redox-dependent mechanisms. Created with BioRender.com. AC1: Adenylyl cyclase 1; AMP: adenosine monophosphate; AMPK: AMP-activated protein kinase; aPKC: atypical protein kinase C; ATP: adenosine triphosphate; CBP: CREB binding protein; CREB: cAMP-response element binding protein; CRTC2: CREB-regulated transcription coactivator 2; DHAP: dihydroxyacetone phosphate; ETC: electron transport chain; F1,6BPase: fructose 1,6-bisphosphatase; FAD: flavin adenine dinucleotide; FADH: flavin adenine dinucleotide (reduced form); FoxO1: forkhead box protein O1; G3P: glycerol 3-phosphate; G6PC: glucose-6-phosphatase catalytic subunit; GCN5: general control non-depressible 5; GPDH: glycerol-3-phosphate dehydrogenase; GPx: glutathione peroxidase; GR: glutathione reductase; GSH: glutathione; GSSG: glutathione disulfide; HNF4α: hepatocyte nuclear factor 4 alpha; Let-7: lethal microRNA; LDH: lactate dehydrogenase; NAD⁺: nicotinamide adenine dinucleotide; NADH: nicotinamide adenine dinucleotide hydrogen; OCT1: organic cation transporter; PEPCK1: cytoplasmic phosphoenolpyruvate carboxykinase; PGC-1α: peroxisome proliferator-activated receptor gamma (PPAR-γ) coactivator-1 alpha; PK: pyruvate kinase; PKA: protein kinase A; PP1: protein phosphatase 1; SHP: small heterodimer partner; SIRT1: sirtuin silent information regulator 1; TET3: tet methylcytosine dioxygenase 3.

Increasing the GR-to-GPx activity ratio

Metformin increases the GR/GPx activity ratio in HepG2 cells78 and thus increases the ratio of reduced (GSH) to oxidized (GSSG) glutathione (GSH/GSSG).79 Elevated GSH/GSSG ratio changes the cytoplasmic redox state. This increases the expression of lethal microRNA (let-7),80 which binds tet methylcytosine dioxygenase 3 (TET3) mRNA and causes posttranscriptional inhibition of TET3 mRNA expression.80 TET3 is a DNA demethylase81 that binds to the P2 promoter of hepatocyte nuclear factor 4 alpha (HNF4α) and promotes its transcription.82 HNF4α is a transcription factor that increases the mRNA expression of gluconeogenic enzymes PEPCK1 and G6Pase catalytic subunit (G6PC).80 In addition to HNF4α, other transcription factors also regulate the expression of gluconeogenic enzymes, of which cyclic AMP (cAMP)-response element binding protein (CREB) and Forkhead box protein O1 (FoxO1) are affected by metformin.83 Therefore, through a redox-dependent increase in let-7 microRNA, metformin decreases the expression of PEPCK1 and G6PC and inhibits hepatic gluconeogenesis by a redox-dependent transcriptional mechanism84 (Figure 2).

Inhibition of mitochondrial oxidative phosphorylation

Metformin inhibits complex I of the mitochondrial respiratory chain, thereby blocking ETC.85 Complex I accepts electrons from NADH and helps to create a proton gradient that serves as the driving force for ATP synthesis.47,86 Metformin also inhibits complex IV, which catalyzes the final step of electron transfer, and fortifies ETC blockage.76 Metformin-induced ETC block inhibits GPDH2 through decreasing electrons accepted by ubiquinone.87 Metformin also directly inhibits GPDH2 noncompetitively in primary hepatocytes isolated from rats84 (Figure 2). GPDH2 is an enzyme in the α-glycerophosphate shuttle47 (Figure 1). Thus, metformin-induced inhibition of ETC76 and GPDH284 increases NADH/NAD^+76,84 and AMP/ATP76 ratios.

In addition to providing redox-dependent inhibition of gluconeogenesis, increased NADH/NAD⁺ ratio suppresses glucose formation from those substrates that their contribution to gluconeogenesis is paired with the reduction of NAD⁺ to NADH.76 Such substrates include lactate and glycerol, which are catalyzed by LDH and GPDH1.76 Thus, metformin inhibits hepatic gluconeogenesis by a redox-dependent substrate-specific mechanism.47 Metformin-induced increased NADH/NAD⁺ ratio reverses the conversion of lactate to pyruvate catalyzed by LDH.88 Lactate accumulation in hepatocytes results in the suppression of hepatocyte lactate uptake.89 Increased lactate production and impaired hepatic lactate removal lead to lactic acidosis.88 The absence of lactic acidosis in OCT1^–/– mice treated with metformin supports the critical role of hepatocytes in metformin-associated lactic acidosis (MALA).90

Increased AMP/ATP ratio inhibits hepatic gluconeogenesis by three mechanisms: (1) allosteric activation of pyruvate kinase (PK)91 and inhibition of F1,6BPase92; (2) inhibition of adenylyl cyclase 1 (AC1)93; and (3) stimulation of AMPK.76 Metformin-induced elevated AMP levels allosterically inhibit F1,6BPase in cultured hepatocytes in mice.94 In addition, the hypoglycemic effect of metformin is abolished in an AMP-independent F1,6BPase knockin mouse model.92 The metformin-induced decrease in ATP concentration stimulates PK in the isolated rat hepatocytes perifused with dihydroxyacetone as the gluconeogenic substrate and thus inhibits hepatic gluconeogenesis.91 Inhibition of F1,6BPase47 and stimulation of PK91 by metformin-induced elevated AMP/ATP ratio is known as allosteric inhibition of gluconeogenesis47 (Figure 2).

Metformin counters glucagon-induced hepatic gluconeogenesis by inhibiting AC193 and activating AMPK.95 Glucagon binds to its receptor, a G_S-protein-coupled receptor, in the hepatocyte membrane and activates AC1, which stimulates the conversion of ATP to cAMP.93 Elevated cAMP activates protein kinase A (PKA).96 PKA mediates the phosphorylation of transcription factor, CREB as well as dephosphorylation of CREB-regulated transcription coactivator 2 (CRTC2) and CREB binding protein (CBP) to form the CREB-CBP-CRTC2 complex, which is a transcriptional and a coactivator complex.97 Phosphorylated CREB of the CREB-CBP-CRTC2 complex binds the promoter of PEPCK1 and G6PC genes and increases their expressions.95 Phosphorylated CREB of the CREB-CBP-CRTC2 complex also binds the promoter of peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α) and FoxO1 genes and induces their expressions.96 PGC-1α is a transcriptional coactivator,98 which forms complexes with transcription factors FoxO196 and HNF4α99 to bind the promoter of PEPCK1 and G6PC genes and increase their expression.97 Glucagon-induced increased cAMP also increases protein phosphatase 1 regulatory subunit 3C expression and enhances protein phosphatase 1 activity, further promoting gluconeogenesis100 (Figure 2).

AMP-dependent AC1 inhibition in mouse hepatocytes and decreased cAMP levels block glucagon-induced expression of G6PC and PEPCK1,93 an energy-dependent transcriptional inhibition of gluconeogenesis.47 However, a study on AMPK-deficient mice showed no impairment in hepatocyte cAMP production by metformin.101 In addition, in prediabetic individuals with hyperglucagonemia, the decremental effect of metformin on gluconeogenesis is decreased, debating the inhibitory effect of metformin on glucagon-stimulated gluconeogenesis in humans.102

AMPK activation induced by an increased ratio of AMP/ATP decreases protein phosphatase 1 regulatory subunit 3C expression in mouse hepatocytes,100 activates atypical protein kinase C,95 and inhibits small heterodimer partner (SHP) gene expression.103 Inhibition of protein phosphatase 1, a regulator of hepatic gluconeogenesis,104 decreases CRTC2 dephosphorylation,100 while atypical protein kinase C phosphorylates CBP.95 Downregulation of SHP, an orphan nuclear receptor, disrupts the CREB-CBP-CRTC2 complex105 and suppresses FoxO1, HNF4α, and FoxA2 transcriptional factors; all AMPK-dependent mechanisms inhibit PEPCK1 and G6PC gene promoters by an energy-dependent transcriptional mechanism103 (Figure 2).

Induction of SIRT1 and GCN5 transcriptional regulators

Metformin treatment of T2DM mice stimulates SIRT1,77 an NAD⁺-dependent cellular energy sensor that deacetylates several substrates.106 SIRT1 activation deacetylates CRTC2 and breaks the CREB-CBP-CRTC2 complex,77 inhibiting gluconeogenesis. SIRT1 also deacetylates PGC-1α and induces the PGC-1α-mediated gluconeogenic genes transcription; however, this effect is inhibited by metformin-induced activation of GCN5, which acetylates and inhibits PGC-1α.77 GCN5 (also known as lysine acetyltransferase 2A) is a histone acetyltransferase and lysine acetyltransferase.107 Stimulation of SIRT1 and GCN5 counter the glucagon-induced cAMP-dependent CRTC2-mediated PEPCK1 and G6PC gene expression77 through an acetylation-dependent transcriptional inhibition of gluconeogenesis (Figure 2).

Effect of Nitric Oxide on Gluconeogenesis in Hepatocytes

NO inhibits hepatic gluconeogenesis. eNOS-derived NO,22,108 iNOS-derived NO,20,21 and NO donors, including S-nitroso-N-acetylpenicillamine,19 3-morpholinosydnonimine,19 S-nitroso-glutathione,19 sodium nitroprusside,22 nitrate,109 and nitrite23 inhibit hepatic gluconeogenesis as demonstrated in type 2 diabetic rats,23,109 type 2 diabetic mice,108 and type 1 diabetic mice,22 streptozotocin-induced diabetic mice,22 isolated mice hepatocytes,22,108 isolated rat hepatocytes,19,21 and cultured rat hepatocytes.20 NOS inhibitors, including N^ω-nitro-L-arginine methyl ester108 and N^G-monomethyl-L-arginine,110 restored the NO-induced inhibited gluconeogenesis in isolated mouse108 and rat110 hepatocytes. Considering the dominance of eNOS in hepatocytes60,72 and the similar recovery of gluconeogenesis in the presence or absence of iNOS stimulators in isolated rat hepatocytes incubated with N^G-monomethyl-L-arginine,110 it can be speculated that eNOS-derived NO is the main NOS-derived NO contributing to gluconeogenesis inhibition.22,108

Mechanisms underlying inhibitory effects of nitric oxide on hepatic gluconeogenesis

As shown in Figure 3, NO inhibits gluconeogenesis mainly through (1) cyclic guanosine monophosphate (cGMP)-dependent mechanisms111,112,113 and (2) enzyme S-nitrosylation.114 cGMP-dependent mechanisms include (a) inhibition of CREB phosphorylation,111 (b) activation of AMPK,112 and (c) activating glucokinase (GCK).115,116 S-nitrosylation mechanisms include (a) GAPDH inhibition,110,117 (b) TCA cycle inhibition,114,118,119 and (c) ETC inhibition.120,121,122

Figure 3.

Mechanisms underlying inhibitory effect of NO on hepatic gluconeogenesis.

NO inhibits hepatic gluconeogenesis via cGMP-dependent pathways and S-nitrosylation. Created with BioRender.com. AMPK: AMP-activated protein kinase; cGMP: cyclic guanosine monophosphate; CREB: cAMP-response element binding protein; ETC: electron transport chain; G6PC: glucose-6-phosphatase catalytic subunit; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; GCK: glucokinase; GTP: guanosine triphosphate; MDH2: mitochondrial malate dehydrogenase; NO: nitric oxide; PEPCK: phosphoenolpyruvate carboxykinase; PKG: cGMP-dependent protein kinase; sGC: soluble guanylate cyclase; TCA: tricarboxylic acid; VASP: vasodilator-stimulated phosphoprotein.

cGMP-dependent mechanisms

NO stimulates soluble guanylate cyclase, a heterodimer composed of two subunits, each containing catalytic and heme-binding domains. Both subunits are essential for enzyme activity. NO activates soluble guanylate cyclase through interaction with its heme component, leading to the conversion of GTP into cGMP in hepatocytes.123,124 Subsequently, cGMP binds to allosteric sites in the regulatory domain of cGMP-dependent protein kinase (PKG),125 enhancing its activity up to 10-fold.126 The NO/cGMP/PKG signaling pathway initiates a cascade of reactions,126 including the inhibition of phosphorylated CREB protein expression in hepatocytes111 (Figure 3). As discussed earlier, phosphorylated CREB is a transcriptional factor and coactivator involved in glucagon-induced expression of gluconeogenic enzymes G6Pase and PEPCK197 (Figure 2). It has been shown that exogenous cGMP can inhibit basal, glucagon-stimulated127 and epinephrine-stimulated128 gluconeogenesis in rat liver.

cGMP also activates AMPK as indicated in nitrite-treated human liver cell line HepG2 cells and causes AMPK-dependent inhibition of gluconeogenic enzyme expressions.112 cGMP-induced AMPK activation in hepatocytes is mediated by vasodilator-stimulated phosphoprotein (VASP), a protein associated with filamentous actin formation and plays a role in cell adhesion and motility.129 In support, in hepatocytes of VASP^−/− mice, AMPK phosphorylation was lower compared to wild-type mice.129 AMPK mainly inhibits glucagon-induced gluconeogenesis.130 According to experiments conducted in eNOS knockout mice,22,108 it has been shown that eNOS-derived NO plays a critical role in AMPK-dependent inhibition of hepatic gluconeogenesis. cGMP also increases GCK activity, which converts glucose to G6P in the glycolysis pathway in isolated rat hepatocytes, thereby inhibiting gluconeogenesis115,116 (Figure 3). Adding cGMP to the culture medium of rat hepatocytes increased GCK activity and GCK mRNA.115 In rat hepatocytes, N^ω-nitro-L-arginine methyl ester decreases cGMP levels, GCK activity, and G6P concentrations but increases glucose levels in rat hepatocytes; however, L-arginine, the precursor of NO, has opposite effects on these parameters.113

Enzyme S-nitrosylation

S-nitrosylation, a reversible post-translational regulation of enzymatic activity,131 results from NO binding to a transition metal132 or cysteine thiols of proteins.114 S-nitrosylation mediates the post-translational regulation of nearly 1000 proteins.133 In hepatocytes, eNOS-derived NO, which is the primary source of hepatic NO, stimulates protein S-nitrosylation134 and protects against apoptosis135 and ischemic complications,136 and mediates the regulation of glucose metabolism.119 Decreased eNOS expression due to a high-fat diet decreases protein S-nitrosylation in mice hepatocytes.134 NO S-nitrosylates and inhibits GAPDH in isolated hepatocytes from rats injected with Cryptosporidium parvum to induce iNOS, thereby decreasing gluconeogenesis.24 Exogenously delivered NO via S-nitroso-N-acetylpenicillamine and sodium nitroprusside inhibits GAPDH in isolated rat hepatocytes, as does endogenously synthesized NO through induction of NOS by a combination of cytokines and LPS.110

eNOS-derived NO also S-nitrosylates other enzymes involved in gluconeogenesis in mice liver,114 including triose phosphate isomerase 1, aldolase, phosphoglycerate mutase, Phosphoglycerate kinase, and pyruvate carboxylase, as well as TCA enzymes (i.e., aconitase, isocitrate dehydrogenase, succinate dehydrogenase, and MDH2)114 (Figure 1). Limited data on the regulatory consequence of S-nitrosylation of these enzymes are available,114 and the overall effect of their S-nitrosylation on gluconeogenesis is controversial.137 It is only assumed that NO inhibits rat liver mitochondrial isocitrate dehydrogenase via S-nitrosylation118 and results in TCA cycle suppression.137 However, the inhibitory effect of S-nitrosylation on other TCA cycle enzymes is speculated to be related to the redox-sensitive cysteine thiol groups of these proteins at their catalytic or regulatory sites.138 NO inhibits mitochondrial aconitase in isolated rat hepatocytes exposed to NO gas or treated with a combination of cytokines and LPS to induce endogenous NO synthesis.120 Although it has not been shown that NO S-nitrosylates aconitase in hepatocytes, the inhibitory effect of NO on pig heart aconitase via S-nitrosylation has been reported.139 In addition, it has been shown that S-nitrosylated ketoglutarate dehydrogenase and SCS have lower activity in human pluripotent stem cell-derived neurons, leading to suppression of the TCA cycle.140 Mitochondrial ketoglutarate dehydrogenase, isocitrate dehydrogenase, and SCS inhibition and the overall suppression of the TCA cycle decrease mitochondrial NADH119 and GTP synthesis,44 resulting in the inhibition of MDH2119 and PEPCK2 within gluconeogenesis flux44 (Figure 3).

NO inhibits mitochondrial NADH-ubiquinone oxidoreductase and succinate-ubiquinone oxidoreductase (complexes I and II of the mitochondrial ETC) in isolated rat hepatocytes exposed to NO gas and or treated with a combination of cytokines and LPS to induce endogenous NO synthesis.120 Among different mechanisms suggested for the inhibition of complex I by NO, S-nitrosylation of critical thiols in the enzyme complex141 is the most supported.142 S-nitrosylation of bovine heart mitochondrial membrane proteins by S-nitroso-N-acetylpenicillamine121 and MitoSNO1 (mitochondria-targeted NO donor and S-nitrosating agent)143 showed lowered activities of complex I, and, to a lesser degree, complex II of the ETC.

NO derived from hepatic stellate cells of cytoglobin-deficient mice, lacking NO dioxygenase activity, which converts NO to nitrate, reversibly inhibits hepatocyte cytochrome c oxidase (complex IV of the ETC),122 an effect that was restored by NO synthase inhibitor (N^ω-nitro-L-arginine methyl ester) or NO scavenger (oxyhemoglobin), which displaces the binding of NO to cytochrome c oxidase.122 Although NO-induced inhibition of complex IV is only described as direct binding of NO to cytochrome c oxidase in hepatocytes,122 it is associated with S-nitrosylation of cysteine residues located on subunit II of the complex IV in porcine pulmonary artery endothelial cells.144

NO-induced inhibition of complex I, II,120 and IV122 suppresses the ETC, decreasing mitochondrial ATP.119 Increased AMP/ATP ratio inhibits gluconeogenesis by countering glucagon-induced gluconeogenesis via AMPK activation145 and AC1 inhibition93 and by allosteric stimulation of PK 91 and inhibition of F1,6BPase92 (Figure 3).

Comparison Between the Effects of Metformin and Nitric Oxide on Hepatic Gluconeogenesis

As mentioned above, mitochondria cannot transport oxaloacetate into the cytoplasm,28 and this necessary step of gluconeogenesis is done through malate, direct, aspartate, and fumarate pathways.31 Malate and direct pathways contribute the most to normal gluconeogenesis during the early and late stages of fasting, respectively.39 Metformin increases the mitochondrial NADH/NAD⁺ ratio in later stages of fasting (in favor of converting oxaloacetate to malate), in which the mitochondrial NADH/NAD⁺ ratio is decreased.39,76 In addition, metformin blocks the α-glycerophosphate shuttle (GPDH1 and GPDH2) and LDH, remaining MAS as the only means for providing cytoplasmic NADH, which is required for converting 1,3-bisphosphoglycerate to GADP by GAPDH during gluconeogenesis42,76 (Figure 1). Both of these effects attenuate metformin’s inhibitory effect on gluconeogenesis via the malate pathway. In support of this notion, metformin increases hepatic malate in type 2 diabetic mice and circulating malate in patients with T2DM.146

NO decreases the NADH/NAD⁺ ratio, probably due to TCA cycle inhibition,119 inhibits MDH2,119 and forces the MAS so that malate is converted to oxaloacetate in the mitochondrion.42 Collectively, these effects strongly inhibit gluconeogenesis via the malate pathway. Regarding the direct pathway, metformin inhibits the direct pathway of gluconeogenesis by increasing the mitochondrial NADH/NAD⁺ ratio,76 changing the activity of MDH2 in favor of converting oxaloacetate to malate and thereby decreasing the availability of oxaloacetate to the direct pathway of gluconeogenesis.39 Furthermore, NO inhibits the TCA cycle and consequently decreases GTP production by SCS, which is required for the activity of PEPCK244 (Figure 1). The increase of the mitochondrial NADH/NAD⁺ ratio by metformin can inhibit the fumarate and aspartate pathways.47,119 Metformin increases the NADH/NAD⁺ ratio and forces the direction of MAS to enter aspartate into the mitochondrion, thereby causing a stronger inhibitory effect on gluconeogenesis via fumarate and aspartate pathways than NO28,29,31,42,76 (Table 2).

Table 2.

Comparison between metformin and NO in gluconeogenesis

	Metformin	NO
Oxaloacetate export
Malate pathway	↑*	↓
Direct pathway	**	↓
Aspartate pathway	↓	↔
Fumarate pathway	↓	↔
Substrate-related gluconeogenesis
Lactate	↓	↔
Alanine	↔	↓
Glutamine	↓	↓
Glycerol	↓	↔
Gluconeogenesis-related enzymes
PEPCK1	↓	↓
GAPDH	↔	↓
G6Pase	↓	↓
F1,6BPase	↓	↓
PK	↑	↑
GCK	↔	↑

*Metformin increases the conversion of mitochondrial oxaloacetate to malate and thus oxaloacetate export; however, its overall effect on hepatic gluconeogenesis through the malate pathway is inhibitory. ** Preventing the compensatory increase. ↑: Increase; ↓: decrease; ↔: no change. F1,6BPase: Fructose 1,6-bisphosphatase; G6Pase: glucose 6-phosphatase; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; GCK: glucokinase; NO: nitric oxide; PEPCK1: cytoplasmic phosphoenolpyruvate carboxykinase; PK: pyruvate kinase.

Regarding gluconeogenic substrates, metformin increases the NADH/NAD⁺ ratio, which affects LDH and GPDH activities and prevents lactate and glycerol utilization for gluconeogenesis.47,76 In addition, glutamine contribution to gluconeogenesis can be suppressed by the metformin-induced increase of mitochondrial NADH/NAD⁺ ratio.28,31,76 Furthermore, NO decreases the NADH/NAD⁺ ratio, which affects MDH2 activity119 and inhibits alanine-induced gluconeogenesis.28 The inhibition of TCA cycle key enzymes by NO can inhibit the glutamine contribution to gluconeogenesis119 (Figure 1). To sum up, substrate-related effects of metformin and NO on inhibition of gluconeogenesis seem complementary, as metformin mainly inhibits lactate, glycerol, and glutamine-induced gluconeogenesis, whereas NO mainly inhibits alanine-induced gluconeogenesis.

Comparing the effects of metformin and NO on gluconeogenesis-related enzymes, both inhibit PEPCK1, G6Pase, and F1,6BPase and stimulate PK.47,111,120 Unlike metformin, NO inhibits GAPDH110,117 and activates GCK115,116 (Table 2).

These findings suggest that the combination of metformin and NO could provide a more potent inhibition of gluconeogenesis compared to metformin alone. This enhanced effect is likely due to their complementary actions, including preventing oxaloacetate export, suppressing gluconeogenic enzymes, and inhibiting the utilization of gluconeogenic substrates (Table 2). However, this hypothesis requires empirical validation. Two studies have investigated the combined effects of metformin and NO donors on metabolic regulation. Cordero-Herrera et al.25 reported that combination therapy with metformin (150 mg/kg/d) and nitrate (85 mg/kg/d) for 7 weeks in high-fat diet-fed mice reduced glycated hemoglobin by approximately 33%. This reduction was greater than that observed with the metformin effect alone (15%) or nitrate effect alone (12%), suggesting a synergic effect, as the combined effect exceeded the sum of the individual effects. Furthermore, the study demonstrated a synergic increase in AMPK phosphorylation and activation, with 140% increase in mice treated with both metformin and nitrate, compared to 66% and 60% increases with metformin and nitrate monotherapy, respectively. Similarly, Adeoye et al.26 recently reported that combined treatment with metformin and L-arginine (100 mg/kg) for 6 weeks reduced serum C-reactive protein levels by approximately 75%. This effect was greater than the reductions achieved with metformin alone (~50%) or L-arginine alone (~25%), indicating an additive effect.

Clinical Applications of Nitric Oxide-Based Interventions

NO-based therapies have a long history in medicine, with glyceryl trinitrate (nitroglycerin) as a well-known example, used to manage angina pectoris (oral and intravenous forms),147 anal fissure (ointment),148 and pulmonary hypertension (nebulized).149,150 Other notable NO-based therapies include isosorbide mononitrate and isosorbide dinitrate for chronic angina and heart failure,147 nitroprusside for hypertensive crises,147 sildenafil (Viagra), which inhibits phosphodiesterase V and amplifies NO signaling, for erectile dysfunction,151 and inhaled NO for persistent pulmonary hypertension in infants.152

In line with our suggestion for the development of a metformin-NO hybrid drug, several efforts have been made to develop NO-donating and NO hybrid drugs for better managing of angina pectoris (e.g., nicorandil147,153), inflammation (NO-aspirin,154 NO-naproxen,155 NO-pravastatin,156 NO-diclofenac,157 and NO-paracetamol158), and diabetic neuropathy (NO-pregabalin159). In addition, in 2017, the Food and Drug Administration approved Latanoprostene Bunod, a NO-donating prostaglandin analog, for lowering intraocular pressure in patients with open-angle glaucoma or ocular hypertension.160 In addition to NO, hybrid drugs incorporating other gasotransmitters, such as hydrogen sulfide, are being developed. A notable example is hydrogen sulfide-NO hybrid compounds (NOSH), such as NOSH-aspirin, which has shown promise in cancer therapy,161 with reported anti-cancer effects in colon,162 pancreas,163 and breast cancers.164 Moreover, NOSH-sulindac,165 NOSH-naproxen,166 and NOSH-aspirin167,168 have been evaluated for managing inflammatory pain. Other areas of research for NO-based treatments include applications in cardiovascular diseases,17,169,170 wound healing171 and metabolic disorders.16,18,170

Developing a metformin-NO hybrid drug offers potential advantages for managing patients with T2DM. First, interactions between the effects of metformin and NO on hepatocytes may enhance their combined ability to inhibit gluconeogenesis. For instance, eNOS-derived NO contributes to AMPK-dependent inhibition of hepatic gluconeogenesis by metformin.22,108 Additionally, metformin enhances the activation of AMPK by nitrate-derived NO in mouse hepatocytes.25 Second, metformin is associated with lactic acidosis in approximately 10 cases per 100,000 patient-years, with mortality rates of 30–50%.88 The therapeutic circulating concentration of metformin ranges from 0.5 to 3 mg/L,172 but levels exceeding 4 mg/L increase the risk of MALA.173 MALA is characterized by elevated plasma metformin (> 5 mg/L), lactate levels (> 5 mM), and blood pH < 7.35.174 NO deficiency also plays a role in lactic acidosis.175 Thus, a metformin-NO hybrid may mitigate MALA by reducing the required metformin dose and preventing NO deficiency-associated lactic acidosis. Notably, metformin is contraindicated in patients with severe renal dysfunction due to the risk of lactic acidosis.7 The potential effect of NO in preventing MALA175 could broaden the applicability of a metformin-NO hybrid in T2DM management, warranting further investigation in animal and human studies. Finally, while NO may exert oxidative and nitrosative stress,176 metformin can elevate GSH,76 potentially counteracting these effects. Elevated GSH also enhances NO-induced S-nitrosylation,177 amplifying its inhibitory effect on gluconeogenesis.

Despite having potential advantages, several challenges must be addressed in developing a metformin-NO hybrid drug. First, NO is a short-lived free-radical gas, making compound stability and bioavailability critical considerations.178 Second, combining the pharmacokinetic properties of metformin and NO requires precise formulation to avoid potential toxicity and ensure efficacy. Third, the effects of NO are dose-dependent. Physiological NO concentrations range from 100 pM to 5 nM,179 with varying biological effects. For example, at concentrations between 1–30 nM, NO promotes cGMP-mediated processes; at 30–100 nM, it activates protein kinase B, promoting cell survival; and at 100–300 nM, NO stabilizes hypoxia-inducible factor 1, protecting against tissue injury. NO concentrations > 400 nM activate tumor suppressor protein P53, which mediates cell cycle arrest and apoptosis, and at concentrations > 500 nM and 1 µM, NO induces oxidative and nitrosative stress, respectively.180 For developing a metformin-NO hybrid drug, key parameters such as the initial concentration of the NO donor, rate of release, decomposition and autoxidation rates, and half-life must be optimized.181,182 Finally, metformin and NO have contrasting effects on glucose metabolism. For example, while metformin upregulates TCA cycle enzymes in hepatocytes,146 NO inhibits these enzymes,114,118,119 as well as pyruvate dehydrogenase.119 This paradox is also observed in combined metformin and sodium-glucose-cotransporter-2 inhibitor treatments for T2DM.146 Careful consideration of these factors will be crucial for developing a safe and effective metformin-NO hybrid drug.

Conclusions

Metformin is the most successful medication in treating T2DM4 and exerts its antidiabetic effects primarily by inhibiting hepatic gluconeogenesis.7,8 T2DM is associated with decreased eNOS-derived NO bioavailability183 because of impaired NO synthesis from L-arginine,184 downregulation of eNOS expression,185,186 and posttranscriptional inhibition of eNOS.187 T2DM decreases hepatic NO by decreasing eNOS phosphorylation188 and increasing eNOS uncoupling.189 NO is involved in the pathophysiology of T2DM,190 and NO-based supplementations have been proposed to have beneficial metabolic effects in T2DM.16,17,18 Similar to metformin, NO inhibits hepatic gluconeogenesis.19,20 However, metformin and NO use shared, as well as unique mechanisms for inhibiting hepatic gluconeogenesis, suggesting that a metformin-NO hybrid drug may be more effective in inhibiting hepatic gluconeogenesis in patients with T2DM. First, the inhibitory effect of NO on hepatic gluconeogenesis via malate pathway is more effective than metformin. On the other hand, metformin exerts a stronger inhibitory effect on fumarate and aspartate pathways of gluconeogenesis than NO.28,29,31,42,76 Second, substrate-related effects of metformin and NO on inhibition of gluconeogenesis seem complementary, as metformin mainly inhibits lactate, glycerol, and glutamine-induced gluconeogenesis, whereas NO mainly inhibits alanine-induced gluconeogenesis. Finally, both metformin and NO inhibit PEPCK1, G6Pase, and F1,6BPase and stimulate PK.47,111,120 Unlike metformin, NO inhibits GAPDH110,117 and activates GCK115,116 (Figure 4).

Figure 4.

Effect of Met and NO on hepatic gluconeogenesis, indicating the site of combined actions.

Created with BioRender.com. DHAP: Dihydroxyacetone phosphate; G6P: glucose-6-phosphate; GADP: glyceraldehyde 3-phosphate; Met: metformin; NO: nitric oxide; PEP: phosphoenolpyruvate.

Some limitations of this review need to be acknowledged to better direct future research and development efforts. First, there is a lack of extensive studies addressing the combined effects of metformin and NO in T2DM. In particular, randomized clinical trials are needed to assess the safety and efficacy of such interventions in humans. Second, much of the evidence supporting the development of a metformin-NO hybrid drug comes from animal models. These studies utilize different types and doses of NO-releasing agents for different durations, which complicates the direct translation of findings to clinical settings. Finally, while this review primarily focuses on the inhibition of gluconeogenesis as the basis for developing a metformin-NO hybrid drug, this compound may also influence other pathways of glucose metabolism, such as glycolysis,91,94,119 the TCA,118,119,140,146 and the ETC.76,85,120,122 Both metformin and NO stimulate glycolysis: metformin achieves this through allosteric activation of phosphofructokinase 194 and PK,91 while NO enhances glycolysis by increasing the AMP/ATP ratio,119 leading to the allosteric activation of PK and GCK.115,116 In addition, both metformin and NO inhibit complex I85,120 and IV76,122 of the ETC, while NO also inhibits complex II.121,143 Further mechanistic studies are needed to determine how NO affects glucose metabolism.

In summary, a metformin-NO hybrid drug could offer more comprehensive glucose control for patients with T2DM, addressing both elevated gluconeogenesis and the risks associated with metformin therapy.

Funding Statement

Funding: This work was supported by Shahid Beheshti University of Medical Sciences, No. 43011564-1; and in part by the National Institutes of Health, USA, Nos. R01GM123508 and 2U54MD017979-01A1 (both to KK).

Data availability statement:

No additional data are available.