Glucose provides a major energy supply for mammalian cells; however, some tissues are much more dependent than others on this source. For example, glucose is the primary energy source for the brain and renal medulla and is essentially the sole provider for red blood cells and the retina. The total daily consumption of glucose in a 70-kg person is about 160 g. The brain uses about 120 g of this. Therefore, tissues that make up about 5% of the body weight consume more than 75% of the glucose metabolized each day. Some 10 to 15 g of glucose is available in the extracellular fluid (ECF), and ∼300 g can be stored in the glycogen reservoirs. Thus, less than a 24-hour supply of glucose is directly available in circumstances when no exogenous supply is available. A constant supply of glucose must be available, as moderate hypoglycemia (<50 mg/dL) causes disabling symptoms and severe hypoglycemia (<20 mg/dL) can have devastating and irreversible effects on the central nervous system (CNS). By contrast, hyperglycemia must be avoided, as this causes its own set of deleterious consequences, albeit on a considerably longer time scale.

An intricate mechanism has evolved for maintaining ECF glucose within a relatively narrow range. In broad physiologic terms (Fig. 15.1), this involves the regulated production of glucose by the liver from glycogenolysis and gluconeogenesis and, to a lesser extent, from gluconeogenesis in the kidneys and small intestine. This is counterbalanced by the regulated peripheral clearance of glucose by tissues such as skeletal muscle, adipose tissue, and the splanchnic bed, including the liver. This chapter concerns the homeostatic role the liver plays in these processes by serving as a consumer, producer, and reservoir for glucose. Descriptions of the control processes involved in glucose homeostasis generally center on hormonal regulation (Fig. 15.1). A major thesis of this chapter is that glucose and/or glucose metabolism is a (the) central player in these events. This point will be made by first outlining the general steps in metabolic control, then by illustrating how these processes affect glucose metabolism, and vice versa.

Multicellular organisms have elaborate mechanisms for sensing changes in the external environment and for generating signals that can then be transduced into an appropriate compensatory response (Table 15.1).

Glucose homeostasis is a case in point. Acute hypoglycemia results in the immediate release of glucagon, which binds to the hepatocyte plasma membrane and thereby promotes the generation of cyclic adenosine monophosphate (cAMP). cAMP is therefore an intracellular symbol of hypoglycemia (1). Hyperglycemia, through insulin release and the subsequent activation of phosphodiesterases that hydrolyze cAMP, reduces the concentration of this nucleotide. A low-molecular-weight activator
of phosphofructokinase-1 (PFK1) was found in the livers of rats given a glucose load and was absent in the livers of rats treated with glucagon (2,3). This molecule, fructose-2,6-bisphosphate (F-2,6-P₂), is thus a symbol of hyperglycemia. As discussed below, F-2,6-P₂ activates glycolysis and inhibits gluconeogenesis, as would be expected of a symbol of hyperglycemia, for which the maintenance of homeostasis requires that glucose utilization be promoted and its production inhibited. Glucose metabolism is a major source of adenosine triphosphate (ATP), which is synthesized by successive phosphorylation steps from AMP. The intracellular levels of AMP and ATP, or the AMP/ATP ratio, are thus a reflection of energy sufficiency or depletion, as the case may be, within the cell. Citrate, an early intermediate in the citric acid cycle, is a symbol of the carbon supply necessary for a variety of biosynthetic pathways. Alanine, synthesized in a single step from pyruvate, serves a similar purpose. Although probably not a complete list, cAMP, F-2,6-P₂, AMP/ATP, citrate, and alanine play key roles, by a variety of mechanisms that will be described below, in regulating hepatic glucose metabolism, which is a major source of energy and carbon for various biosynthetic reactions.

The elucidation of metabolic pathways, along with the principles that govern the regulation of these pathways, is one of the major achievements of 20th century biological research. The deciphering of how thousands of reactions are coordinated, as well as the purification and characterization of the enzymes involved, was accomplished using techniques that often were very primitive by contemporary standards. This set the stage for more recent developments that allow for the analysis of components of these pathways by overexpression of an enzyme in a tissue of interest, or the selective removal of an enzyme, by transgenic and gene knockout technologies. These, and other newer techniques, are providing supportive evidence for the fundamental importance of certain enzymatic steps and have led to the discovery of regulatory processes heretofore unappreciated. The regulation of glucose metabolism in the liver, the topic of this chapter, has been of central interest to investigators for more than 100 years. As will be discussed below, many of the principles of metabolic regulation were derived from these studies.

The simplest type of control is afforded by substrate availability. Flux through an enzyme step is directly related to the concentration of substrate and to the affinity of the enzyme for that substrate. For example, hexokinases catalyze the phosphorylation of glucose to form glucose-6-phosphate (G-6-P). The Michaelis-Menten constant (K_m) for the 100-kDa hexokinases (HK-I–III) is low (20–120 μM), so these enzymes are very efficient at low ECF glucose concentrations (4). Glucokinase, a 50-kDa hexokinase found in liver and pancreatic β-cells, has a K_m for glucose of 5 to 10 mM, so it is very active when glucose concentrations in the ECF are high, particularly in the portal circulation, as happens after a meal (5). These high glucose concentrations are well above the K_m of the 100-kDa hexokinases, which are relatively less influential under such circumstances. Glucokinase therefore plays an important role in hepatic glucose clearance during the postprandial period. Substrate availability provides a very rapid and simple control mechanism and is a major control mechanism in unicellular organisms. However, these organisms are relatively defenseless in the face of large changes of substrate availability, particularly when the substrate is not abundant.

Substrates and reaction products afford another type of direct regulation in which they affect their own concentration
by influencing the activity or amount of an enzyme by positive or negative mechanisms. Positive control occurs when a downstream component activates a subsequent enzyme and promotes forward flux, or induces the synthesis of an enzyme that accomplishes the same purpose. Fructose-1,6-bisphosphate (F-1,6-P₂) activates a downstream enzyme, pyruvate kinase (PK), and thus promotes glycolytic flux (6); (Fig. 15.2). This provides a mechanism for ensuring that flux continues in the forward direction, but it generally does not afford fine-tuning of substrate concentration or flux because it is unidirectional. Negative feedback control, which is much more common than positive control, occurs when the product of a reaction is capable of feedback inhibition of upstream enzyme activity or synthesis. Negative feedback control can be exerted at many upstream reaction steps in a pathway and helps define the points at which metabolic-control strength is exerted. This mechanism, analogous to the function of a thermostat in temperature control, adjusts substrate concentration in a narrow range around a preset value. When the concentration of the reaction product falls, negative feedback inhibition is relieved and the reaction (or reactions) produces more of the regulatory molecule. High concentrations shut down the system, thus providing up-and-down fine-tuning. For example, physiologic levels of G-6-P, the product of the hexokinase reaction, provide feedback inhibition of the 100-kDa hexokinases but not of glucokinase (7). This provides for appropriately regulated levels of the branch-point substrate G-6-P in most tissues. Tissues or organs wherein glucokinase is the predominant hexokinase, such as liver and pancreatic β-cells, are able to continuously phosphorylate glucose owing to the lack of feedback inhibition by G-6-P, the reaction product. The liver is thus able to clear glucose in the postprandial state, and β-cells are able to employ glucokinase as a glucose sensor for regulated secretion of insulin (8). Substrate availability and feedback/ feedforward control are prominent mechanisms in single-cell organisms.

Simple regulatory systems are advantageous in that they provide very rapid and quickly reversible control. The major disadvantages are that the unicellular organism is susceptible to major environmental changes of substrate availability and the regulation is provided by molecules that are themselves reaction
components. As noted, these mechanisms are frequently employed in glucose metabolism in multicellular organisms. However, the complexity of these organisms is such that certain environmental changes that are threats to survival are often perceived by one tissue or organ and the responses that allow for homeostasis are facilitated by one or more distant organs. Communication between the sensors of the environment and the organs or tissues that provide the appropriate response to recognized challenges is accomplished by the nervous and endocrine systems. For example, hypoglycemia is perceived by the CNS, which triggers the release of glucocorticoids, catecholamines, and glucagon and the inhibition of insulin secretion (Fig. 15.1). These hormones, the so-called first messengers, provide much of the cell-cell communication necessary for the coordinated response that restores the ECF glucose level to the normal range. A variety of intracellular molecules serve as so-called second messengers, including cAMP, cyclic guanosine monophosphate (cGMP), calcium ion, and various phosphatidylinositides. These molecules share several properties. They act as intracellular symbols of particular environmental changes, and more than one of these may be symbols of the same challenge, e.g., hypoglycemia. The concentration of these molecules is labile within the cell, as befits their regulatory potential. In contrast to the simple mechanisms of regulation described above, none of these molecules is a component (substrate or product) of a reaction pathway, nor are they structural analogues of such components. Finally, these regulatory molecules all initiate signal transduction cascades that result in the ultimate biologic response. Again, these cascades are not mutually exclusive, as different signals can activate the same, or complementary, steps in a given pathway. A wide variety of metabolic control mechanisms are affected by these signals, including allostery, posttranslational protein modification, protein translocation, protein-protein interaction, and changes in the amount of key proteins (Table 15.2).

Until recently the guiding principle in metabolic control has been the idea that there is a single rate-controlling step in any metabolic pathway. Flux through such a pathway could thus be increased or decreased by changing the activity of that single rate-determining enzyme. The advent of recombinant DNA technology enabled investigators to rather easily change the expression of a given enzyme in a cell; yet this often does not result in the change of flux predicted by the rate-limiting enzyme hypothesis. These results were not surprising to students of metabolic control analysis, a concept first proposed by Kacser some 30 years ago (30). Metabolic control is assessed by measuring all enzyme activities, their kinetic properties, the substrates and products of each reaction, and the relevant regulatory molecules in a particular pathway. An analysis of the data allows one to assign a quantitative control strength value (C) to each step in this pathway. The sum of the C values for a pathway equals 1.0. High C values indicate high control strength. When appropriately analyzed, it is apparent that control strength can be distributed among the several steps in a pathway, and the C value for a given step can be changed by different metabolic circumstances. An examination of glucose metabolism in the isolated, beating heart provides a good example of these principles. In a heart preparation perfused with a physiologic concentration of glucose, control of the early steps of glucose metabolism is distributed between glucose transport (C value of 0.4) and glucose phosphorylation (C value of 0.6) (31). The addition of insulin results in a remarkable change in the distribution of control strength. Under these conditions the C value for glucose transport is now negligible (0.0002), whereas the C value for glucose phosphorylation increases to 0.97 (31). Phosphoglucoisomerase accounts for the other 0.02. Metabolic control analysis also allows one to assign relative importance to a set of reactions within a complex pathway. For instance, in the case cited, the distribution of control provided by transport/ phosphorylation (0.58), glycolysis (0.25), and glycogen synthesis (0.17) was assessed.

At equilibrium, the supply of an end product by the series of enzymes and substrates equals the demand afforded by product utilization and degradation. Thus, the demand side can also influence metabolic control by providing a pull on the system. Two recent studies illustrate the importance of this concept. In studying tryptophan biosynthesis in yeast, investigators first showed that exogenous provision of excessive amounts of the first enzyme in the biosynthetic pathway did not increase flux to tryptophan. All of the enzymes in the pathway had to be overexpressed to increase flux (32). This illustrates the point made above about so-called rate-limiting steps. Of more interest was the observation that the biosynthetic enzymes (“supply side”) accounted for only about 25% of the control strength. The other 75% was presumably provided by the “demand side,” e.g., the use of tryptophan for protein synthesis. This is similar to the example cited above, in which, in the heart, glycolysis and glycogen synthesis exert control strength downstream from glucose transport/phosphorylation (31). That demand may also be an important control step as illustrated in another recent report. Glycolysis, as noted above, provides ATP from glucose. Yeasts engineered to have an excess of the ATP-degrading enzyme adenosine triphosphatase (ATPase) have a remarkable increase in glycolytic flux (33). This observation is pertinent to the topic of this chapter. The energy charge in these cells, as
reflected in a low AMP/ATP ratio, resulted in increased glycolytic flux. As described in detail below, a low AMP/ATP ratio promotes flux through the critical glycolytic enzymes PFK1 and PK in hepatocytes. It thus is important to recognize that supply and demand are both significant, that control is distributed, and that this distribution is subject to change according to metabolic circumstances.

Glucose enters all cells by facilitated diffusion (from a high to low concentration) through one of the members of the glucose transporter family of proteins (34). This is basically a unidirectional process in cells that cannot produce glucose, as one of the four hexokinase family members (HK-I–IV) rapidly converts glucose to G-6-P, which cannot be exported by the glucose transporter. The activity of these hexokinases keeps the intracellular free-glucose concentration low, so most cells have a net uptake of glucose due to inward, passive diffusion. The only regulation of this process is afforded by the extracellular glucose concentration (substrate control) and by G-6-P, which inhibits (by negative feedback) the 100-kDa hexokinases (isoforms I–III).

Hepatocytes, which can produce glucose from glycogenolysis and gluconeogenesis, can export and import glucose. The direction of glucose flux depends on the concentration gradient. Under certain metabolic conditions (e.g., fasting or hormonal conditions that favor glycogenolysis and/or gluconeogenesis), hepatocytes release glucose into the ECF. The ability of the hepatocyte to effect net glucose release is a function of its high K_m hepatic transporter coupled with a regulatable high K_m hexokinase (HK-IV, or glucokinase). When hormonal conditions favor glucose utilization (high insulin, low glucagon), glucokinase activity is high, intracellular glucose is rapidly phosphorylated, and glucose entry is facilitated along a downhill gradient. When conditions favor glycogenolysis or gluconeogenesis (low insulin, high glucagon), glucokinase activity is reduced, the low rate of glucose phosphorylation results in intracellular accumulation of glucose, and glucose is exported by facilitated diffusion. The hepatic GLUT2 transporter has a K_m for glucose of ∼40 mM and generally functions with pseudo first-order kinetics with respect to the intracellular or extracellular glucose concentration (34). Because the portal vein glucose concentration normally does not exceed 10 mM, transport is not limiting for either hepatic glucose import or export. Because of this feature there is no need for regulation of the hepatic glucose transporter.

Hepatic glucose production and utilization involve the movement of substrates through three major cycles (2,35) (Fig. 15.2). Hepatic glucose production, by gluconeogenesis, is a biosynthetic pathway. Glucose utilization, by glycolysis, is a degradative pathway. Degradative and biosynthetic pathways are usually catalyzed by separate sets of enzymes. This affords unique, but flexible, control of each pathway. The latter is particularly important in this instance, as the production of a single molecule of glucose from gluconeogenesis requires the expenditure of six high-energy bonds [four ATP and two guanosine-triphosphate (GTP)], whereas glucose degradation to two pyruvate molecules results in the production of two ATP molecules. Futile cycling between the cycles involved in gluconeogenesis and glycolysis could thus be very wasteful in terms of energy consumption and heat production. In general, a negative free-energy charge (AMP, or a high AMP/ATP ratio) promotes glycolysis, and a high free-energy charge (ATP, or a low AMP/ATP ratio) favors the gluconeogenic pathway.

The direction and magnitude of substrate movement, along either the glycolytic or the gluconeogenic pathway, is controlled by six enzymes that catalyze essentially irreversible reactions and thus have high metabolic control strength (2). The three enzymes that catalyze the glycolytic reactions must be bypassed by three different enzymes that direct the gluconeogenic pathway and vice versa. The activity of these six enzymes is modulated by acute (seconds to minutes) and chronic (minutes to hours) regulatory mechanisms. The acute regulation of hepatic glucose metabolism occurs through hormone-mediated changes in enzymatic activity, principally the phosphorylation or dephosphorylation of two key enzymes, PK and phospho-fructokinase-2/fructose-2,6-bisphosphatase (PFK2/FBPase2, the bifunctional enzyme), and the allosteric regulation of GK by GKRP and of PFK1 and FBPase1 by F-2,6-P₂ (2,3). Hormones also exert chronic effects on glucose metabolism by changing the rate of enzyme synthesis. These chronic effects are mediated through alterations of the rate of mRNA synthesis and, in some cases, by changes in the rate of degradation of specific mRNAs.

The net flux through the glucose/G-6-P, F-6-P/F-1,6-P₂, and phosphoenolpyruvate/pyruvate cycles depends on the relative activity of the enzymes illustrated in Figure 15.2. Conditions favoring gluconeogenesis (the combination of high plasma glucagon, catecholamine, and glucocorticoid levels with a low plasma insulin level, as occurs when animals are starved or fed a low-carbohydrate diet) result in increased activity of PEPCK, FBPase1, and glucose-6-phosphatase (G-6-Pase) and a coordinate decrease in the activity of GK, PFK1, and PK. Reciprocal changes in the activities of these enzymes occur when animals are fed a carbohydrate diet, particularly after a prolonged fast (2,3). In this situation, the plasma insulin concentration increases, levels of the counterregulatory hormones decrease, and glycolysis predominates. The F-6-P/F-1,6-P₂ substrate cycle is also regulated by a subcycle in which the amount of the regulatory molecule F-2,6-P₂ is controlled by the bifunctional enzyme (2,3). The relative activity of these seven enzymes, established by the interaction of several hormones, determines whether the hepatocyte is a consumer or producer of glucose. The activity of the enzymes involved in the cycles is controlled by all of the control mechanisms described above. Many other enzymes are also involved in these processes; however, they catalyze equilibrium reactions that are not rate determining.

Upon entering the hepatocyte through the GLUT2 transporter, glucose is converted to G-6-P by the enzyme glucokinase (Fig. 15.3). Glucokinase is an unusual member of the hexokinase family. Its K_m for glucose is greater than 5 mM, whereas the K_m for HK-I–III (glucokinase is HK-IV) varies between 20 and 120 μM (4,5). This difference in K_m values gives the brain, muscle, and red blood cells, which have isoforms I–III, preference when the glucose supply is limited. Portal vein and liver sinusoidal glucose concentrations are high after a meal. Hepatocytes efficiently convert glucose to G-6-P because glucokinase (HK-IV) works most efficiently at this elevated glucose concentration (7). Also, unlike HK-I–III, glucokinase is not subject to feedback inhibition by physiologic concentrations of the reaction product, G-6-P (7). Because GLUT2 operates by facilitated diffusion,
the activity of glucokinase, which keeps intracellular glucose concentrations very low, determines glucose clearance by hepatocytes.

This cycle is important because G-6-P is a branch-point substrate for, or a product of, the glycolytic, gluconeogenic, glycogen formation/degradation, hexosamine, and hexose monophosphate shunt pathways (Fig. 15.3). The balance of glucose phosphorylation and G-6-P hydrolysis, and thus the production or metabolism of glucose, as well as the avoidance of futile substrate cycling, requires the interplay of a number of control mechanisms.

Glucokinase activity is not altered by covalent modification, but its activity is regulated by an allosteric mechanism and by changes in the amount of the protein (2,19). Glucose phos-phorylation, by glucokinase, results in increased intracellular levels of G-6-P. Although physiologic concentrations of G-6-P do not inhibit glucokinase, G-6-P is in near equilibrium with F-6-P through the action of the reversible enzyme phosphoglucoisomerase. F-6-P increases the affinity of binding of the nuclear GKRP to glucokinase (19). When bound to GKRP, glucokinase is found primarily in an inactive form in the nucleus of the hepatocyte (36,37,38). This complex is subject to allosteric regulation, as the inhibition of glucokinase by F-6-P is reversed by F-1-P, which competes with F-6-P for binding to the GKRP (19). F-1-P binding causes a conformational (allosteric) change of GKRP. This results in the dissociation of glucokinase, which can then translocate to the cytoplasm as an active enzyme (38). Thus fructose, which is converted to F-1-P by fructokinase, promotes hepatic glucose clearance by its indirect activation of glucokinase (38) (Fig. 15.3).

Glucokinase gene transcription, glucokinase mRNA, and glucokinase activity are decreased when plasma glucagon is high and plasma insulin is low (e.g., fasting or diabetes) (7,37,39,40,41). Refeeding a diet high in carbohydrate results in an increase of these parameters. A 20- to 30-fold increase in trans-cription occurs within 30 to 60 minutes after the injection of insulin into a diabetic rat or after the addition of insulin to primary cultures of hepatocytes (40,41,42). Glucokinase mRNA increases accordingly (40,41), and a significant increase of glucokinase activity and attendant flux from glucose to G-6-P ensues (43). The inhibitory effect of glucagon (or cAMP, its intracellular messenger) on glucokinase gene transcription is dominant over the stimulatory effect of insulin, because cAMP blocks the effects of insulin at all concentrations of insulin in cultured cell systems (44). These hormone effects are not dependent on the presence of glucose in the medium (41,44).