ABSTRACT
Glucose levels are tightly regulated at all times. Gluconeogenesis is the metabolic pathway dedicated to glucose synthesis from non-hexose precursors. Gluconeogenesis is critical for glucose homoeostasis, particularly during fasting or stress conditions. The renal contribution to systemic gluconeogenesis is increasingly recognized. During the post-absorptive phase, the kidney accounts for ∼40% of endogenous gluconeogenesis, occurring mainly in the kidney proximal tubule. The main substrate for renal gluconeogenesis is lactate and the process is regulated by insulin and cellular glucose levels, but also by acidosis and stress hormones. The kidney thus plays an important role in the maintenance of glucose and lactate homoeostasis during stress conditions. The impact of acute and chronic kidney disease and proximal tubular injury on gluconeogenesis is not well studied. Recent evidence shows that in both experimental and clinical acute kidney injury, impaired renal gluconeogenesis could significantly participate in systemic metabolic disturbance and thus alter the prognosis. This review summarizes the biochemistry of gluconeogenesis, the current knowledge of kidney gluconeogenesis, its modifications in kidney disease and the clinical relevance of this fundamental biological process in human biology.
INTRODUCTION
Glucose is one of the main body fuels and its blood levels are tightly regulated. Plasma glucose values are maintained within a narrow range throughout the day despite wide fluctuations in the delivery and removal of glucose from the circulation in order to fuel the body organs and notably the brain. Gluconeogenesis is the pathway by which glucose is synthesized from non-hexose precursors such as glycerol, lactate, pyruvate and glucogenic amino acids and is crucial in maintaining normoglycaemia during fasting and stress conditions. After an overnight fast, glucose production relies on endogenous production by both glycogenolysis and gluconeogenesis.
Glycogenolysis is the breakdown of glycogen to glucose-6-phosphate (G6P) and further hydrolysis to glucose. As fasting progresses, glycogen, stored mainly in the liver and the muscles, is depleted. After around 60 h of starvation in humans, gluconeogenesis becomes the only source of glucose production [1]. Gluconeogenesis is classically attributed to the liver. However, in 1937, the ability of the mammalian kidney to produce glucose from non-carbohydrate precursors was discovered [2] and further investigated in animal models, healthy persons and—more recently—in patients with kidney disease [3]. This review summarizes the biochemistry of gluconeogenesis, the current knowledge of renal gluconeogenesis and its potential clinical relevance.
Metabolic pathway
Gluconeogenesis is an anabolic pathway consisting of 11 reactions catalysed by enzymes (summarized in Figure 1), resulting in the production of glucose from non-hexose precursors. Most of these reactions are the reverse steps of glycolysis, except for four reactions, which are irreversible and specific to gluconeogenesis. These specific reactions are catalysed by four enzymes: pyruvate carboxylase (PC), phosphoenolpyruvate carboxykinase (PEPCK, PCK1), fructose-1,6-bisphosphatase (FBP) and G6P. Of them, PC is the only one localized in the mitochondria and converts pyruvate to oxaloacetate [4]. Pyruvate can originate from glucose through the final steps of glycolysis, from lactate oxidation via lactate dehydrogenase or from the transformation of amino acids such as alanine. Oxaloacetate cannot cross the mitochondrial membrane and is first transformed to aspartate or malate through malate dehydrogenase or aspartate aminotransferase in order to be transported to the cytosol, where it is reconverted back to oxaloacetate. In the cytosol, oxaloacetate is both decarboxylated and phosphorylated by PEPCK/PCK1. This reaction is widely accepted as a rate-limiting step for gluconeogenesis [5]. PEPCK/PCK1 activity appears well related to the amount of PCK1 messenger RNA (mRNA), although post-transcriptional regulations are described [6]. FBP catalyses the hydrolysis of FBP to frutcose-6-phosphate. The activity of FBP is regulated either at the pre- or the post-translational level [7]. Finally, G6P dephosphorylates to glucose. G6P is also regulated at the transcriptional level [8]. In addition to the classical pathway, two shunts exist in the gluconeogenesis pathway: some ketogenic and glucogenic amino acids (in particular glutamine) can join the TCA cycle and be transformed directly to oxaloacetate, whereas glycerol, released by the hydrolysis of triglycerides, can also join gluconeogenesis via its conversion to dihydroxyacetone phosphate [9]. Gluconeogenesis is an energy-requiring process. For each molecule of glucose generated, six high-energy phosphate bonds are needed.
Gluconeogenesis pathway.
Gluconeogenic organs and regulation of renal gluconeogenesis.
Renal gluconeogenesis in the proximal tubule cells.
Renal gluconeogenesis in the healthy kidney
The liver was generally considered as the site of gluconeogenesis. In 1937, the ability of kidney slices to produce glucose in response to pyruvate and lactate exposure suggested the gluconeogenic role of the kidney [2]. One year later, another group showed that, compared with hepatectomized rabbits, those who additionally underwent renal artery ligatures needed more glucose to maintain euglycaemia [10]. Four decades later, with the isotopic method, the kidney was finally well recognized as a major organ for gluconeogenesis [3]. In the kidney, gluconeogenesis takes place in the proximal tubule, where the key specific enzymes are exclusively expressed in segments S1–S3 [11–13], with much lower activities in the latter portion of the nephron [14, 15]. Even under maximal stimulation, distal nephron segments are unable to perform gluconeogenesis. We confirmed using transcriptomic data from single-nucleus RNA sequencing of mice kidneys that the expression of rate-limiting enzymes PEPCK/PCK1 and FBP are restricted to segments S1, S2 and S3 of the proximal tubule in mice [16]. In contrast, the distal nephron consumes glucose and expresses the key enzymes of the glycolytic pathway (hexokinases, 6-phosphofructokinases and pyruvate kinases), which are not present in the proximal tubule in normal conditions [15, 17, 18].
Due to this functional partition resulting in both consumption and production of glucose, renal gluconeogenesis has been underestimated by net organ balance measurements [19–22]. In 1995, the first isotopic method able to measure separately renal glucose uptake and release was described [23]. In the post-absorptive phase, after 14–16 h of fasting, serum glucose levels were stable, implying that systemic glucose release was equivalent to systemic glucose uptake, at a rate of ~10 µmol/kg/min. In this study, the kidney accounted for ∼30% of the systemic glucose rate of appearance in 10 healthy human volunteers after an overnight fast. Interestingly, because of the considerable renal glucose uptake observed in these patients (20% of systemic glucose disposal), the renal net balance was small (˂1 µmol/kg/min) compared with that of the liver (8 µmol/kg/min). Subsequent human studies obtained similar results, with a renal contribution to systemic gluconeogenesis equivalent to 20 ± 2% of the total body glucose production [3] after an overnight fast. Considering that gluconeogenesis accounts for ∼50% of glucose production after an overnight fast [24–29] and that all renal glucose release is produced by gluconeogenesis, as normal kidney contains only limited glycogen stocks in the medulla [30] with only low levels of G6P enzymes expression [31], one can conclude that renal gluconeogenesis is responsible for ∼40% of the systemic gluconeogenesis after an overnight fast [3]. As fasting progresses, glycogen stores are depleted and the role of gluconeogenesis increases to maintain stable glucose levels, reaching 90% of the whole-body glucose production in healthy humans after fasting >40 h [1, 25, 26]. After prolonged starvation, the kidney contributed up to 50% of the total systemic glucose production [32]. Altogether, kidney gluconeogenesis plays a critical role in glucose homoeostasis, particularly in fasting conditions (Figure 2).
Substrates of renal gluconeogenesis
Gluconeogenesis can produce glucose from various precursors such as glycerol, lactate and amino acids, mostly alanine and glutamine. In post-absorptive healthy volunteers, renal gluconeogenesis was shown to rely mainly on lactate, glutamine and glycerol that accounted for ∼50, 20 and 10% of renal glucose release, respectively, in two different studies [29, 33]. The kidney is also able to metabolize fructose, since the infusion of 2 mmol/min fructose in healthy humans after a 60-h fast induced an increase in net renal glucose and lactate production, suggesting an enhancement of both glycolytic and gluconeogenic pathways [34]. In freshly microdissected human nephron segments, proximal tubules exhibited heterogeneity in the gluconeogenic rate from lactate but not glutamine, with the S2 and S3 segments synthesizing more glucose from lactate than the S1 segment [35]. These findings thus highlighted the role of the kidney in lactate metabolism.
Lactate is fully filtered by the glomerulus and reabsorbed in the proximal tubule [36] until a threshold of 10 mmol/L is reached, where urinary excretion occurs [37]. Lactate uptake from the tubular lumen occurs through a carrier-mediated process involving sodium-coupled monocarboxylate transporters (MCTs) 1 and 2 and proton-linked MCTs (MCT8), all using the sodium gradient as a driving force [38, 39]. Tubular proximal cells also expressed proton-linked MCTs (MCT1 and MCT2) at the basolateral side and may act as an efflux pathway of the reabsorbed lactate towards the blood or in taking up lactate or pyruvate from the circulation for gluconeogenesis [40].
The kidney is both a lactate producer and consumer, with marked differences between the cortex and the medulla. Due to the low oxygen tension [41], the distal nephron segments rely mainly on glycolysis to generate adenosine triphosphate (ATP) [42], resulting in net lactate generation [43]. Thus the distal nephron, including the thick ascending limb, the distal convoluted tubules and the collecting duct, all display the highest activity of glycolytic enzymes [43]. This produced lactate acts as a substrate for cortical gluconeogenesis, leading to a corticomedullary glucose lactate recycling loop [44]. In contrast, the cortex relies mainly on oxidative phosphorylation to generate ATP [45–47]. Renal lactate uptake is a non-saturable process over a wide range of lactate levels, as shown after lactate infusion in healthy sheep [48]. As a consequence, at an arterial concentration of 9 mmol/L, the liver is not the dominant organ for lactate clearance [48]. Renal lactate uptake is not directly related to oxygen consumption, suggesting that most lactate reabsorbed in the kidney is used for gluconeogenesis [49]. Finally, in rats submitted to an intravenous lactic acid load, the kidney was responsible for 30% of lactate removal, highlighting the major role of the kidney in systemic lactate clearance [50] (Figure 3).
Regulation of renal gluconeogenesis
Renal gluconeogenesis is hormonally regulated. A 3-fold increase in renal gluconeogenesis was observed after the infusion of stress hormones (combining hydrocortisone, epinephrine, norepinephrine and glucagon) in dogs, accounting at the end for 22% of the overall glucose production. In this experiment, the kidney seemed to be more sensitive to the stimulation than the liver [51]. Further studies focused on each hormone separately. Methylprednisolone infusion was shown to increase glucose net balance of isolated perfused rat kidney 2-fold [52]. This effect could be mediated by an increase in PEPCK and PC activities [53–55] related to upregulation in gene expression [56, 57] and translation [58]. In post-absorptive healthy volunteers, a continuous infusion of epinephrine at a rate of 50 ng/kg/min increased renal glucose release by 2-fold, accounting for 40% of total systemic glucose [23]. This response was sustained in the kidney, whereas it was only transient in the liver. Epinephrine may also increase gluconeogenic substrate availability and increase the efficiency of renal gluconeogenesis [29]. Renal denervation may also directly regulate renal glucose production, as unilateral renal denervation leads to a decrease in renal glucose release [59].
The effects of insulin on gluconeogenesis have been extensively studied. In healthy post-absorptive humans, two independent groups reported a decrease in renal glucose release and an increase in renal glucose uptake during a hyperinsulinaemic–euglycaemic clamp experiment. Several mechanisms link insulin to gluconeogenesis. At the transcriptional level, insulin decreases PCK1 and G6P mRNA expression both in vivo and in vitro [60] via tyrosine phosphorylation of insulin receptor substrates (IRSs) 1 and 2 and further phosphorylation of forkhead box protein O1 (FOXO1) via the canonical pathway IRS/Pi3k/Akt/FOXO1 [61], leading to its nuclear exclusion. Insulin also reduces the availability of gluconeogenic substrates, mainly glycerol and glutamine [62, 63]. The suppression of plasma-free fatty acids could itself decrease renal gluconeogenesis [64]. Finally, insulin could induce a redirection of the gluconeogenic substrates to the oxidative pathways to compensate for the decrease in free fatty acid and sustain energy production [3, 62].
Glucose levels play a direct role in renal gluconeogenesis regulation. Proximal tubules reabsorb glucose via sodium-dependent glucose transporters (SGLTs) 1 and 2. In a glucose-rich medium, glucose reabsorption increases, increasing the ratio between the oxidized and reduced form of nicotinamide adenine dinucleotide (NAD+/NADH), inhibiting deacetylation of peroxisome proliferator-activated receptor gamma coactivator 1α (PGC1α) in sirtuin-1SIRT1–dependent manner, leading to a downregulation of PCK1 gene expression [60]. Interestingly, the novel SGLT2 inhibitors stimulate renal gluconeogenesis in the proximal tubule [60].
Acidosis induces PEPCK mRNA expression [14] and activity [58] in the proximal tubule. This transcriptional regulation is independent of nuclear factors but requires an intact hepatocyte nuclear factor (HNF)-1 binding site in the gene promoter [65]. It results in an increase in renal glucose production [66]. Whereas extrarenal lactate clearance decreases with acidosis, renal ability to remove lactate increases with a low pH [50]. Coordinated with ammoniogenesis, renal gluconeogenesis results in glucose production and acid excretion, contributing to maintenance of systemic acid–base homoeostasis [67, 68].
Both in vitro and in vivo studies have reported the relation between gluconeogenic substrate concentrations and the rate of gluconeogenesis. For example, isotopic lactate or alanine infusion in patients with non-insulin-dependent diabetes mellitus increased the systemic gluconeogenesis rate >2-fold, accounting for ∼50% of glucose [69]. Similar results were reported for glycerol [70, 71] or alanine [72] infusion in dogs. In addition, fatty acids [73], ketone bodies [74] and glycerol [71] could inhibit the oxidation of lactate and pyruvate and simultaneously increase their conversion to glucose through a sparing effect on other precursors and action on enzymatic activity. Altogether, renal gluconeogenesis appears to be particularly enhanced by stress hormones, insulin and acidosis, suggesting a critical role in critical illness.
Gluconeogenesis and kidney disease
Renal function and hypoglycaemia risk
Hypoglycaemia has classically been related to liver failure. Considering the high rate of renal gluconeogenesis and its importance in glucose homoeostasis, impairment of renal function could also lead to hypoglycaemia. In 1986, a retrospective study reported in 94 hospitalized patients that chronic kidney disease [CKD; defined as a serum creatinine level >265 µmol/L, estimated glomerular filtration rate (eGFR)˂40 mL/min or the need for renal replacement therapy] was the factor most associated with the incidence of hypoglycaemia (defined as glucose levels ˂49 mg/dL) [75]. A possible link with renal gluconeogenesis impairment was then suggested [76]. In a woman with diabetes mellitus and chronic renal failure, endogenous glucose production was measured because of repetitive episodes of fasting hypoglycaemia. Using glucose isotope infusion, the authors described a severe impairment of systemic gluconeogenesis [77]. A report from 1978 mentioned four non-diabetic patients on dialysis developing spontaneous hypoglycaemia associated with metabolic acidosis related to hyperlactataemia [78]. Reduced insulin clearance has been speculated to explain part of this observation [79], but a recent study showed that insulin secretion, sensitivity and clearance were not associated with eGFR when adjusted for confounding factors in non-diabetic CKD patients with an eGFR ˂60 mL/min/1.73 m2 compared with healthy volunteers [80]. In 2009, a large retrospective cohort analysis including 243 222 patients reported a higher incidence of hypoglycaemia in patients with CKD. This was also true in non-diabetic patients (2.23 versus 3.46/100 patient-months, P < 0.0001), excluding the participation of anti-diabetic therapies [81]. Altogether these data demonstrate an association between renal function and the risk of hypoglycaemia.
Renal function and lactate clearance
Because the kidney mostly depends on lactate as a gluconeogenic substrate, it also plays a major role in the systemic clearance of this metabolite. Renal disease could thus be associated with decreased lactate clearance and hyperlactataemia. The clinical relevance of this metabolic function could be major since lactate clearance during acute disease is associated with mortality [82]. However, only one observational cohort study was performed in end-stage renal disease patients and reported a high prevalence of hyperlactataemia, but without the presence of a control group [83].
Renal gluconeogenesis in acute kidney injury (AKI)
Renal disease is associated with large modifications of renal metabolism [84–86]. In vitro, the medullary thick ascending limb and the cortical and outer medullar collecting duct dramatically increase their glycolysis rate in response to anoxia, leading to an increase in lactate release [42]. Several models of renal injury have confirmed this pattern, including mercuric acid-induced acute tubular necrosis [87] and septic [88] and ischaemia–reperfusion injury (IRI)-related AKI [89–91]. These changes were shown to be persistent in atrophic tubules [92]. Taken together, these results suggest a switch from oxidative phosphorylation to glycolysis during renal injury.
In contrast, little is known about gluconeogenesis regulation during AKI. Considering the facts that a competition exists between renal sodium transport and glucose production [52, 93–96] and lactate consumption [44] when ATP disposal is limited and that glycolysis and gluconeogenesis are reciprocally regulated [97], a decrease in renal gluconeogenesis during AKI seemed plausible. We recently demonstrated using human renal arteriovenous catheterization that cardiac surgery–related AKI was associated with decreased renal glucose release and lactate uptake [16]. This was in line with a recent study where renal arteriovenous catheterization was performed in kidney allograft recipients, showing that patients who experienced delayed graft function exhibited a positive net renal lactate balance and a lower glucose renal tissue content [98]. Consistent with this hypothesis, we could show that the expression of gluconeogenic enzymes dramatically decreased during the reperfusion phase of a kidney allograft compared with a more stable phase. To confirm the defect in gluconeogenesis, we further performed several in vivo experiments to demonstrate that systemic lactate clearance was altered in AKI and that renal glucose release was reduced after ischaemic AKI in rats, in agreement with the expression, translation and activities of key gluconeogenic enzymes such as PCK1 and FBP1. The cause of these changes has yet to be determined. In particular, further investigations are required to elucidate whether changes in renal gluconeogenesis represent a passive effect of impaired cell energy metabolism or a coordinated regulated biologic process. In fact, several elements suggest that the fundamental changes in metabolism occurring in the kidney responding to injury may represent a coordinated process. Notably, several recent studies have highlighted that the oxidative phosphorylation defect occurring during AKI and driving the transition to CKD [84–86, 99, 100] was related to PGC1α downregulation [85, 99, 101, 102]. PGC1α is a transcriptional coactivator acting as a global metabolism enhancer, including for stimulation of oxidative metabolism and mitochondria biogenesis [103] and gluconeogenesis through the HNF4α [104] and FOXO1 (Forkhead Box Protein O1) [105]. Overexpression of PGC1α in hepatic cells enhances PCK1, G6P and FBP1 transcription and increases glucose production [106]. NAD is required for gluconeogenesis, both directly for reduction of diphosphoglycerate to glyceraldehyde phosphate [107, 108] and indirectly for sirtuin-1 enhancement. Sirtuin-1 is a histone deacetylase shown to be protective in IRI AKI [109, 110] Sirtuin-1 deacetylates PCK1, restoring its gluconeogenic capacity [111]. Kidney NAD content decreases in different experimental models of AKI [100]. Using single nucleus RNA sequencing, we confirmed that expression of gluconeogenic rate-limiting enzymes was strongly reduced after IRI-induced AKI as well as classical transcriptional regulators such as peroxisome proliferator-activated receptor α and HNF4. Thus alteration of gluconeogenesis is likely to be an important component of a global metabolic switch occurring in the kidney tubular cells in response to injury.
Renal gluconeogenesis and metabolic syndrome
In patients with morbid obesity, renal and hepatic gluconeogenesis are almost similar after a prolonged fast [32]. In proteinuric obese patients, mammalian target of rapamycin complex 1 is upregulated in the proximal tubule cells [112] and the uptake of fatty acid is increased [113], leading to lipid droplet accumulation as triacylglycerol in the cortex [114] and, finally, to enhancement of renal gluconeogenesis via an increase in acetyl coenzyme A concentrations [115]. Intracellular fat accumulation has also been shown to drive insulin resistance [116]. Considering the renal effect of insulin on gluconeogenesis, one could hypothesize an increase in renal glucose production in obesity through insulin resistance. Mice with proximal tubule deletion of the insulin receptor displayed higher fasting glucose levels associated with elevated activity of G6P in renal cortex homogenates, which could be related to enhanced gluconeogenesis [117]. In insulin-resistant diabetic Zucker rats, a disinhibition of renal gluconeogenesis due to increased mRNA expression and activity of gluconeogenic enzymes was observed [118]. Nevertheless, a recent study conducted by Sasaki et al. [60] reported that, in streptozotocin-treated animals, gluconeogenesis was downregulated, likely via a different pathway than insulin resistance . Thus the net effect of metabolic syndrome on renal gluconeogenesis is still to be determined.
Clinical relevance of renal gluconeogenesis alterations
Is renal gluconeogenesis important and is the regulation observed during AKI of clinical interest? To answer these questions we retrospectively analysed glucose and lactate levels in intensive care unit (ICU) patients and observed, as already suggested by Freire Jorge et al. [119], that AKI correlates with a specific systemic metabolic pattern associating high lactate and low glucose levels. Interestingly, the impact of liver failure was less important than renal failure. Very importantly, and in agreement with the fact that both hypoglycaemia and decreased lactate clearance [82, 119] are associated with an adverse outcome, AKI-associated mortality was restricted to patients displaying higher lactate and lower glucose levels. This suggests that the metabolic function of tubule cells could have a systemic relevance beyond the kidney and may influence the systemic aspect of metabolism and finally mortality, at least in the specific context of critical illness. The enhanced glycolysis observed during renal disease may also play an important role in the phenotype [92, 120].
Whether gluconeogenesis modulation during AKI only reflects the injury severity or plays a role in disease progression should be investigated. For instance, whereas FBP1 has been shown to be depleted in clear cell renal cell carcinoma, in vitro overexpression inhibited glycolysis and restrained cell proliferation by inhibiting nuclear hypoxia-inducible factor function [121]. In hepatocarcinoma, dexamethasone, an active form of synthesized glucocorticoids, is able to restore gluconeogenesis, leading to therapeutic efficacy against hepatocarcinoma [122]. As previously described, renal gluconeogenesis may be modulated by several stimuli. Among these, SGLT2 inhibitors appear to be associated with the enhancement of renal gluconeogenesis [123] and improvement of the renal prognosis [124]. Steroids are known enhancers of renal and liver gluconeogenesis (see above). Recent meta-analyses including all major randomized controlled trials reported that steroids confer no or minimal mortality benefit in septic patients, but the very different complex of this class precludes any conclusion on a specific effect on renal gluconeogenesis [125–127]. Thiamine is a critical cofactor for several reactions occurring in the mitochondria [128]. In a retrospective study using propensity score for thiamine supplementation matching, we observed an improvement in the systemic metabolic pattern and a decrease in ICU mortality in patients receiving thiamine. This was in line with previous studies showing improvement in both lactate clearance and 28-day mortality in patients receiving thiamine [129]. If confirmed in prospective interventional trials, these data would promote renal gluconeogenesis as one of the most important functions of the kidney in ICU patients and as a new pharmacological target to improve overall prognosis in AKI.
CONCLUSION
Gluconeogenesis is a neglected metabolic function of the kidney. Alterations of renal gluconeogenesis may be an important part of the global metabolic changes observed during AKI and its systemic repercussions may be of major importance in critically ill patients. Whether modulation of this pathway can rescue the poor prognosis of AKI deserves further studies.
FUNDING
D.L. is the recipient of a STARTER grant (RS03-25) from the the foundation of the Geneva University Hospitals and the University of Geneva’s Faculty of Medicine. S.d.S. is supported by the Swiss National Science Foundation (grant PP00P3 157454).
REFERENCES
Author notes
None declared.
Comments