Gene-altered Mice and Metabolic Flux Control

Abstract

The regulation of metabolic flux is essential for life, yet our understanding of it remains incomplete. The ability to determine the role of specific enzymes in regulating metabolic flux rapidly becomes a nearly overwhelming challenge due to the various interacting pathways, multiple enzyme isoforms, and many different substrates that are involved. This complexity is further compounded by the fact that the enzymes themselves are regulated by multiple mechanisms. Moreover, the amount of an enzyme in a particular cell may be modulated by a variety of transcriptional and post-transcriptional mechanisms. Thus, it has long been apparent that an understanding of how metabolic flux through specific pathways is regulated requires detailed measurements within the context of model systems. In mammals, the opposing processes of glycolysis and gluconeogenesis are both essential for glucose homeostasis. As previously reviewed by Granner and Pilkis (1Granner D. Pilkis S. J. Biol. Chem. 1990; 265: 10173-10176Abstract Full Text PDF PubMed Google Scholar), the regulation of three substrate cycles (glucokinase (GK)/glucose-6-phosphatase, pyruvate kinase/pyruvate carboxylase/phosphoenolpyruvate carboxykinase (PEPCK), 1The abbreviations used are: PEPCK, phosphoenolpyruvate carboxykinase; GK, glucokinase; KO, knock-out; CREB, cAMP response element-binding protein.1The abbreviations used are: PEPCK, phosphoenolpyruvate carboxykinase; GK, glucokinase; KO, knock-out; CREB, cAMP response element-binding protein. and phosphofructokinase/fructose-1,6-bisphosphatase) play an important role in hepatic glucose metabolism. Of these, two have garnered most of the attention: GK and the cytosolic isoform of PEPCK. Both are widely thought to play key roles in determining flux through the glycolytic and gluconeogenic pathways, respectively, in the liver. Since a variety of genetically altered mouse models have been generated and characterized to determine the role of these two enzymes in regulating metabolic flux they now serve as excellent examples for highlighting the indispensable role that genetically manipulated mice have played in assessing metabolic flux control within the intact animal. Moreover, the findings obtained illustrate that surprises are still possible despite decades of previous studies and that additional studies are still required. In this Minireview we compare and contrast the findings obtained for both GK and PEPCK, because, by doing so, some important points about the concept of metabolic control are reinforced. GK and PEPCK are both expressed and regulated in a manner that reflects their vastly different functions. The two enzymes can be generally thought of as being expressed in a largely non-overlapping set of cell types, as is illustrated in Fig. 1. While they are co-expressed in the liver, periportal and perivenous hepatocytes possess different amounts of these and other glycolytic and gluconeogenic enzymes, and consequently different metabolic capacities, due to gradients in oxygen, substrate, hormone, and mediator levels, as well as different cell-to-cell interactions (2Jungermann K. Diabetes Metab. 1992; 18: 81-86PubMed Google Scholar, 3Jungermann K. Kietzmann T. Annu. Rev. Nutr. 1996; 16: 179-203Crossref PubMed Scopus (407) Google Scholar). Two kinetically similar but structurally distinct isoforms of GK are made due to alternate promoters in the GK gene and the alternate RNA processing (4Jetton T.L. Liang Y. Pettepher C.C. Zimmerman E.C. Cox F.G. Horvath K. Matschinsky F.M. Magnuson M.A. J. Biol. Chem. 1994; 269: 3641-3654Abstract Full Text PDF PubMed Google Scholar, 5Schuit F.C. Huypens P. Heimberg H. Pipeleers D.G. Diabetes. 2001; 50: 1-11Crossref PubMed Scopus (311) Google Scholar, 6Corbett J.A. Trends Endocrinol. Metab. 2001; 12: 140-142Abstract Full Text Full Text PDF PubMed Scopus (12) Google Scholar). One isoform is made in the liver, whereas the other is found in the pancreatic islet, enteroendocrine cells of the gut, and in certain parts of the brain. Studies of the role of GK in determining glycolytic flux have been limited largely to the liver and pancreatic β-cell, where the enzyme is regulated in a vastly different manner. Besides the use of alternate cell type-specific promoters, the liver expresses a regulatory protein that binds GK and sequesters it in the nucleus under certain metabolic conditions (7de la Iglesia N. Mukhtar M. Seoane J. Guinovart J.J. Agius L. J. Biol. Chem. 2000; 275: 10597-10603Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar, 8Grimsby J. Coffey J.W. Dvorozniak M.T. Magram J. Li G. Matschinsky F.M. Shiota C. Kaur S. Magnuson M.A. Grippo J.F. J. Biol. Chem. 2000; 275: 7826-7831Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). Due to an S0.5 for glucose of about 8 mm and the lack of significant end product inhibition by glucose 6-phosphate, GK functions to regulate intracellular glycolytic flux in a manner that parallels changes in the blood glucose concentration (9Matschinsky F.M. Diabetes. 2002; 51: S394-S404Crossref PubMed Google Scholar). In the case of PEPCK, there are two different genes that encode either a mitochondrial or cytoplasmic isoform of the enzyme (10Hanson R.W. Reshef L. Annu. Rev. Biochem. 1997; 66: 581-611Crossref PubMed Scopus (621) Google Scholar). The cytoplasmic PEPCK isoform is found in the proximal tubule of the kidney, adipose, and intestine, in addition to hepatocytes (10Hanson R.W. Reshef L. Annu. Rev. Biochem. 1997; 66: 581-611Crossref PubMed Scopus (621) Google Scholar). In both the liver and kidney this enzyme has long been thought to play a role in determining gluconeogenic flux, although the gene is regulated differently in these two sites (10Hanson R.W. Reshef L. Annu. Rev. Biochem. 1997; 66: 581-611Crossref PubMed Scopus (621) Google Scholar, 11Drewnowsk K.D. Craig M.R. Digiovanni S.R. McCarty J.M. Moorman A.F. Lamers W.H. Schoolwerth A.C. J. Physiol. Pharmacol. 2002; 53: 3-20PubMed Google Scholar). Intestinal PEPCK may also contribute to the endogenous glucose production during diabetes (12Rajas F. Croset M. Zitoun C. Montano S. Mithieux G. Diabetes. 2000; 49: 1165-1168Crossref PubMed Scopus (88) Google Scholar). In contrast, in adipose tissues, PEPCK expression is thought to play an important role in glyceroneogenesis (10Hanson R.W. Reshef L. Annu. Rev. Biochem. 1997; 66: 581-611Crossref PubMed Scopus (621) Google Scholar, 13Beale E.G. Hammer R.E. Antoine B. Forest C. FASEB J. 2002; 16: 1695-1696Crossref PubMed Scopus (53) Google Scholar). Indeed, transgene-directed overexpression of PEPCK in adipose tissue causes increased glyceroneogenesis, free fatty acid re-esterification, and adipose mass, without any changes in insulin sensitivity, findings that provide direct support for this notion (14Franckhauser S. Munoz S. Pujol A. Casellas A. Riu E. Otaegui P. Su B. Bosch F. Diabetes. 2002; 51: 624-630Crossref PubMed Scopus (173) Google Scholar). The physiological regulation of glucose homeostasis involves multiple tissue-tissue interactions, mediated by a variety of hormones, the nervous system, as well as changes in the concentrations of specific nutrients. For instance, glycolytic flux in the liver is stimulated by insulin and inhibited by glucagon, two hormones secreted by the pancreatic islet. The secretion of both of these hormones is modulated by sympathetic, parasympathetic, and sensory innervation of the pancreas (15Ahren B. Diabetologia. 2000; 43: 393-410Crossref PubMed Scopus (678) Google Scholar). Centers regulating the neuronal input into hormone secretion are thought largely to reside in the hypothalamus and brain stem, two other sites where GK is found. Moreover, incretins, such as glucagon-like peptide-1 or glucose-dependent insulinotropic polypeptide that affect insulin secretion are secreted by enteroendocrine cells of the gut, cells that are also known to express GK (16Doyle M.E. Egan J.M. Recent Prog. Horm. Res. 2001; 56: 377-399Crossref PubMed Scopus (89) Google Scholar, 17Drucker D.J. Endocrinology. 2001; 142: 521-527Crossref PubMed Scopus (310) Google Scholar, 18Pamir N. Lynn F.C. Buchan A.M. Ehses J. Hinke S.A. Pospisilik J.A. Miyawaki K. Yamada Y. Seino Y. McIntosh C.H. Pederson R.A. Am. J. Physiol. 2003; 284: E931-E939Crossref PubMed Scopus (98) Google Scholar). Thus, there exists a highly dispersed yet integrated system of GK-containing cells that interact with each other in a complex manner to maintain glucose homeostasis. Similarly, the regulation of gluconeogenic flux, which varies under different metabolic conditions, also depends on a variety of different tissue-tissue interactions. Although the liver, kidney, and intestine all contain the enzymatic machinery necessary to make glucose, changes in the hormonal milieu have a significant impact on where glucose is made, as well as the substrates that are used as precursors. For instance, the liver primarily utilizes lactate and alanine, and to a smaller extent glycerol, for gluconeogenesis. However, alanine is less important as a substrate in the kidney, with both glycerol and glutamine being more important for glucose production from this site (19Gerich J.E. Meyer C. Woerle H.J. Stumvoll M. Diabetes Care. 2001; 24: 382-391Crossref PubMed Scopus (425) Google Scholar, 20Cersosimo E. Garlick P. Ferretti J. Diabetes. 2000; 49: 1186-1193Crossref PubMed Scopus (78) Google Scholar). Moreover, the small intestine is beginning to be viewed as playing a substantial role during certain metabolic states. During the postabsorptive state systemic glucose release occurs primarily from the liver and kidney, with no glucose release from the small intestine (12Rajas F. Croset M. Zitoun C. Montano S. Mithieux G. Diabetes. 2000; 49: 1165-1168Crossref PubMed Scopus (88) Google Scholar, 19Gerich J.E. Meyer C. Woerle H.J. Stumvoll M. Diabetes Care. 2001; 24: 382-391Crossref PubMed Scopus (425) Google Scholar). However, glucose production by the small intestine rises to 25% of that made in the whole body during insulinopenic states, such as during a 48-h fast or in streptozotocin-induced diabetes (21Croset M. Rajas F. Zitoun C. Hurot J.M. Montano S. Mithieux G. Diabetes. 2001; 50: 740-746Crossref PubMed Scopus (156) Google Scholar). The relative contributions of the kidney and liver may also change during an extended fast, with renal gluconeogenesis becoming somewhat more important (22Ekberg K. Landau B.R. Wajngot A. Chandramouli V. Efendic S. Brunengraber H. Wahren J. Diabetes. 1999; 48: 292-298Crossref PubMed Scopus (190) Google Scholar). For many years a predominant view pertaining to the regulation of metabolic flux was the presence of a rate-limiting step at or near the start of a pathway. However, this view began to give way in the late 1970s to the concept of variable metabolic control strength. Fundamental to this concept was the idea that control over a metabolic pathway might be distributed over a number of enzymes in a pathway, and the flux control strength of each enzyme could be quantified by metabolic control analysis in an isolated system (23Kacser H. Burns J.A. Biochem. Soc. Trans. 1979; 7: 1149-1160Crossref PubMed Scopus (320) Google Scholar, 24Groen A.K. van Roermund C.W. Vervoorn R.C. Tager J.M. Biochem. J. 1986; 237: 379-389Crossref PubMed Scopus (145) Google Scholar). This concept has had significant impact on how the regulation of metabolic flux is both viewed and studied. However, as this concept evolved, it also became clear that the control exerted by any particular enzyme might also vary under different states and thus in some cases might be virtually impossible to determine without a detailed empirical analysis (25Rognstad R. J. Biol. Chem. 1979; 254: 1875-1878Abstract Full Text PDF PubMed Google Scholar). It is now well recognized that, only by being able to precisely alter the expression of an enzyme in a graded manner, can the role of specific enzymes in regulating metabolic flux in a pathway be accurately assessed. Even so, it remains a frequent practice to attempt to infer the function of an enzyme using knowledge of its kinetics, location, or regulation. However, there is no guarantee that such inferences are correct. Fortunately it is no longer necessary to simply guess given the increasing ease and precision for performing genetic manipulations in the mouse, especially given the ability both to overexpress enzymes in specific tissues and to eliminate their expression in selected sites via the Cre/loxP system (26Le Y. Sauer B. Methods Mol. Biol. 2000; 136: 477-485PubMed Google Scholar). Both overexpression and conditional gene knock-out (KO) strategies proved to be valuable for determining the effects changing the expression of GK on an important metabolic marker, the blood glucose concentration (27Niswender K.D. Postic C. Jetton T.L. Bennett B.D. Piston D.W. Efrat S. Magnuson M.A. J. Biol. Chem. 1997; 272: 22564-22569Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar, 28Niswender K.D. Shiota M. Postic C. Cherrington A.D. Magnuson M.A. J. Biol. Chem. 1997; 272: 22570-22575Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar, 29Postic C. Shiota M. Niswender K.D. Jetton T.L. Chen Y. Moates J.M. Shelton K.D. Lindner J. Cherrington A.D. Magnuson M.A. J. Biol. Chem. 1999; 274: 305-315Abstract Full Text Full Text PDF PubMed Scopus (1003) Google Scholar). As shown in Fig. 2A, a 50% reduction in GK gene expression leads to pronounced hyperglycemia. The total elimination of GK is incompatible with life because pups die within a week of birth from the effects of uncontrolled hyperglycemia. In contrast, a 50% increase in the amount of GK causes hypoglycemia, and doubling the amount of GK (via introduction of two extra copies of the entire GK gene locus) lowers the blood glucose concentration even more. When plotted, as shown in Fig. 2A, a curvilinear relationship between GK gene expression and the blood glucose concentration becomes readily apparent. These data clearly illustrate that small changes in the expression of GK are sufficient to have a major impact on the plasma glucose concentration, although they do not reveal which tissues are involved or precisely how they each contribute to this relationship. To determine the role of GK in particular tissues and to further explore the functional relationship between tissues that express this enzyme, both liver- and pancreatic β-cell-specific GK-KO mice have been generated and characterized (29Postic C. Shiota M. Niswender K.D. Jetton T.L. Chen Y. Moates J.M. Shelton K.D. Lindner J. Cherrington A.D. Magnuson M.A. J. Biol. Chem. 1999; 274: 305-315Abstract Full Text Full Text PDF PubMed Scopus (1003) Google Scholar). Liver-specific GK-KO mice exhibit nearly a 40% increase in blood glucose levels without a simultaneous increase in the plasma insulin levels. Hyperglycemic clamp experiments revealed that, although the basal glucose turnover rates are nearly normal, the rate of new hepatic glycogen synthesis and glucose turnover in response to hyperglycemia is only 10 and 40% that of animals that make normal amounts of hepatic GK, respectively. The effect of glucose on several glucose-responsive genes was also abnormal, further demonstrating a key role for GK in carbohydrate-induced responses by the liver (30Towle H.C. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13476-13478Crossref PubMed Scopus (37) Google Scholar). Although these functional impairments might have been predicted using in vitro systems, one observation simply could not have been anticipated. Namely, the liver-specific GK-KO mice exhibited profoundly impaired insulin secretion. Mice lacking hepatic GK did not exhibit any increase in their plasma insulin levels under basal condition, despite the 40% increase in blood glucose levels, and the increase in insulin secretion in response to hyperglycemia was only 30% of normal. These results suggest a linkage between rates of hepatic glycolysis and insulin secretion that is currently not understood. The mild hyperglycemia that occurs in these mice might, itself, be sufficient to impair insulin secretion because it has also been shown that small but chronic changes in the blood glucose concentration have marked effects on insulin secretion (31Weir G.C. Laybutt D.R. Kaneto H. Bonner-Weir S. Sharma A. Diabetes. 2001; 50: S154-S159Crossref PubMed Google Scholar, 32Laybutt D.R. Sharma A. Sgroi D.C. Gaudet J. Bonner-Weir S. Weir G.C. J. Biol. Chem. 2002; 277: 10912-10921Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar). By performing a pancreatic β-cell-specific KO of GK different information was gained. First, it was observed that mice without any GK in β-cells die shortly after birth of severe diabetes. Thus, this experiment clearly indicates that GK plays a vital role in regulating metabolic flux in β-cells. Although this was long thought to be the case, this experiment, and a related study in isolated islets, leaves almost no room for doubt (29Postic C. Shiota M. Niswender K.D. Jetton T.L. Chen Y. Moates J.M. Shelton K.D. Lindner J. Cherrington A.D. Magnuson M.A. J. Biol. Chem. 1999; 274: 305-315Abstract Full Text Full Text PDF PubMed Scopus (1003) Google Scholar, 33Piston D.W. Knobel S.M. Postic C. Shelton K.D. Magnuson M.A. J. Biol. Chem. 1999; 274: 1000-1004Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Mice that lack one functional allele of GK only in the β-cell exhibit a metabolic phenotype similar to that of the liver-specific GK-KO mice, namely mild hyperglycemia and normal basal glucose turnover rates, but more predictably, they exhibit impaired glucose turnover and insulin secretion during hyperglycemia due to the firmly established link between GK in β–cells and glucose-stimulated insulin secretion (9Matschinsky F.M. Diabetes. 2002; 51: S394-S404Crossref PubMed Google Scholar, 34Matschinsky F.M. Diabetes. 1996; 45: 223-241Crossref PubMed Scopus (0) Google Scholar, 35Matschinsky F.M. Glaser B. Magnuson M.A. Diabetes. 1998; 47: 307-315Crossref PubMed Scopus (288) Google Scholar). Mice that make an increased amount of GK are resistant to the development of both hyperglycemia and hyperinsulinemia associated with the feeding of a high fat diet (36Shiota M. Postic C. Fujimoto Y. Jetton T.L. Dixon K. Pan D. Grimsby J. Grippo J.F. Magnuson M.A. Cherrington A.D. Diabetes. 2001; 50: 622-629Crossref PubMed Scopus (49) Google Scholar). This resistance to development of obesity is thought to be due mainly to increased hepatic GK activity, which causes increased hepatic glucose disposal via increased hepatic glucose utilization and increased glycogen synthesis, particularly in response to hyperglycemia (28Niswender K.D. Shiota M. Postic C. Cherrington A.D. Magnuson M.A. J. Biol. Chem. 1997; 272: 22570-22575Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar). Ferre et al. (37Ferre T. Riu E. Bosch F. Valera A. FASEB J. 1996; 10: 1213-1218Crossref PubMed Scopus (164) Google Scholar) have also reported that the streptozotocin-induced increase in the serum concentrations of glucose, 3-hydroxybutyrate, triglyceride, and free fatty acids is smaller in GK overexpressing mice. A linkage between GK activity and glycolysis is clearly evident as a result of the genetic studies that have led to the identification of over 150 naturally occurring mutations of GK and serve as a natural experiment for assessing the control strength of GK. Mutations that impair the function of GK cause a phenotype in humans known as maturity onset diabetes of the young, type 2 (MODY-2) (38Kesavan P. Wang L. Davis E. Cuesta A. Sweet I. Niswender K. Magnuson M.A. Matschinsky F.M. Biochem. J. 1997; 322: 57-63Crossref PubMed Scopus (44) Google Scholar, 39Davis E.A. Cuesta-Munoz A. Raoul M. Buettger C. Sweet I. Moates M. Magnuson M.A. Matschinsky F.M. Diabetologia. 1999; 42: 1175-1186Crossref PubMed Scopus (116) Google Scholar). Interestingly, several mutations have been identified that increase the activity of the enzyme and that are associated with the syndrome of persistent hyperinsulinemic hypoglycemia of infancy (PHHI) (40Christesen H.B. Jacobsen B.B. Odili S. Buettger C. Cuesta-Munoz A. Hansen T. Brusgaard K. Massa O. Magnuson M.A. Shiota C. Matschinsky F.M. Barbetti F. Diabetes. 2002; 51: 1240-1246Crossref PubMed Scopus (133) Google Scholar, 41Glaser B. Kesavan P. Heyman M. Davis E. Cuesta A. Buchs A. Stanley C.A. Thornton P.S. Permutt M.A. Matschinsky F.M. Herold K.C. N. Engl. J. Med. 1998; 338: 226-230Crossref PubMed Scopus (508) Google Scholar). Thus, genetic studies in humans, coupled with biochemical analysis of the mutants, reinforce the experiments in mice (or vice versa), with both lines of study clearly indicating that the changes in the net activity of GK cause changes in the blood glucose concentration, an effect due solely to the high metabolic control strength that GK exerts on glycolysis in cells that express the enzyme. The metabolic effects of altering the amount of PEPCK in mice have also been intensely investigated. As for GK, the results of several groups have been important in being able to assess the impact of broad changes in the expression of the gene and its role in determining metabolic flux and in altering the blood glucose concentration after fasting. First, Valera et al. (42Valera A. Pujol A. Pelegrin M. Bosch F. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9151-9154Crossref PubMed Scopus (252) Google Scholar) generated transgenic mice that overexpress a PEPCK via a truncated fragment of PEPCK promoter that leads to increased expression in all normal sites of expression except adipose tissue. In these animals, PEPCK gene expression in the liver was 7-fold higher than wild-type animals, and the blood glucose concentration during fasting was 50% higher than normal. Second, Sun et al. (43Sun Y. Liu S. Ferguson S. Wang L. Klepcyk P. Yun J.S. Friedman J.E. J. Biol. Chem. 2002; 277: 23301-23307Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar) made a transgenic mouse, also under control of a truncated promoter fragment, in which hepatic PEPCK gene expression in the liver was about 2-fold higher than that of littermates. The blood glucose concentration after a 6-h fast in these animals was unaffected. Third, She et al. (44She P. Shiota M. Shelton K.D. Chalkley R. Postic C. Magnuson M.A. Mol. Cell Biol. 2000; 20: 6508-6517Crossref PubMed Scopus (188) Google Scholar) described three functionally distinct pck alleles that were made by gene targeting using a allelogenic Cre/loxP strategy. One allele was null, another was functionally normal but could be converted to null via Cre recombinase, whereas the third allele was functionally attenuated. By intercrossing mice containing these different alleles, animals with variations in PEPCK gene expression ranging from 0 to 100% of normal were produced. Similar to the analysis performed for GK, the blood glucose concentration observed in mice with variable amounts of PEPCK gene expression (0- to 7-fold increase) was plotted and is shown in Fig. 2B. This plot is based on data from adult animals except for the zero PEPCK point, which was plotted at 30% because this reflects the blood glucose concentration in 1-day-old pups immediately prior to death. She et al. (44She P. Shiota M. Shelton K.D. Chalkley R. Postic C. Magnuson M.A. Mol. Cell Biol. 2000; 20: 6508-6517Crossref PubMed Scopus (188) Google Scholar) found that mice with a 50, 90, and 95% global reduction in PEPCK gene expression were viable and maintained a normal blood glucose concentration after a 24-h fast. The points at 200 and 700% PEPCK mRNA are based on the results of Sun et al. (43Sun Y. Liu S. Ferguson S. Wang L. Klepcyk P. Yun J.S. Friedman J.E. J. Biol. Chem. 2002; 277: 23301-23307Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar) and Velera et al. (42Valera A. Pujol A. Pelegrin M. Bosch F. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9151-9154Crossref PubMed Scopus (252) Google Scholar). When the results of all three different groups are considered, it is clear that the relationship between PEPCK gene expression and the fasting blood glucose concentration is markedly different from the curvilinear relationship observed for the expression of GK and the blood glucose concentration. Whereas there is a negative slope through the point of 100% GK mRNA, the slope through the point reflecting 100% PEPCK mRNA is zero. These results clearly indicate that PEPCK is essential for life but also show that only a small amount of PEPCK is necessary for maintaining a normal blood glucose concentration in an unstressed animal. Indeed, the fasting blood glucose concentration is essentially unaffected by a wide range of changes in PEPCK gene expression. Even mice with a 95% reduction in PEPCK mRNA are not only viable but they have nearly normal blood glucose concentrations after a 24-h fast. Although these findings seem to imply that PEPCK exerts very little control over the fasting blood glucose concentration, except at the far extremes of expression (0 or 7 times the normal amount), they must be interpreted with caution. Changes in PEPCK mRNA may not accurately predict changes in PEPCK activity, and multiple compensatory mechanisms may be invoked when there is an inadequate amount of PEPCK that serve to protect the animal from the lethal consequences of hypoglycemia (44She P. Shiota M. Shelton K.D. Chalkley R. Postic C. Magnuson M.A. Mol. Cell Biol. 2000; 20: 6508-6517Crossref PubMed Scopus (188) Google Scholar, 45She P. Burgess S.C. Shiota M. Flakoll P. Donahue E.P. Malloy C.R. Sherry A.D. Magnuson M.A. Diabetes. 2003; 52: 1649-1654Crossref PubMed Scopus (88) Google Scholar). Thus, the plasma glucose concentration may be a poor reflection of metabolic alterations that occur in the face of large alterations in PEPCK gene expression. For instance, large changes in the expression of other genes in the liver and the rate of hepatic glucose production were observed in mice with only a 2-fold increase in hepatic PEPCK gene expression (43Sun Y. Liu S. Ferguson S. Wang L. Klepcyk P. Yun J.S. Friedman J.E. J. Biol. Chem. 2002; 277: 23301-23307Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar). These mice had a 2.3-fold increase in both their fasting plasma insulin concentration and impaired glucose tolerance, consistent with mild insulin resistance (43Sun Y. Liu S. Ferguson S. Wang L. Klepcyk P. Yun J.S. Friedman J.E. J. Biol. Chem. 2002; 277: 23301-23307Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar). Further metabolic characterization of these animals revealed increased rates of hepatic glucose production but normal whole-body glucose disposal during hyperinsulinemic-euglycemic clamp experiments. Thus, a modest increase in hepatic PEPCK activity may be enough to impair insulin signaling. Consistent with this, Sun et al. (43Sun Y. Liu S. Ferguson S. Wang L. Klepcyk P. Yun J.S. Friedman J.E. J. Biol. Chem. 2002; 277: 23301-23307Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar) showed there was a decrease in the amount of both insulin receptor substrate-2 and phosphatidylinositol 3-kinase activity in the liver. Mice that lack PEPCK only in the liver are also able to maintain fasting euglycemia, although at the expense of marked hepatic steatosis, a finding that also challenges both the notion of the exclusive role of the liver in gluconeogenesis and the role of PEPCK in this process. In a follow-up study of the liver-specific PEPCK-KO mice, She et al. (45She P. Burgess S.C. Shiota M. Flakoll P. Donahue E.P. Malloy C.R. Sherry A.D. Magnuson M.A. Diabetes. 2003; 52: 1649-1654Crossref PubMed Scopus (88) Google Scholar) have demonstrated that perfused livers from animals that lack hepatic PEPCK do not produce glucose from either lactate or pyruvate. This finding suggests that the mitochondrial isoform of PEPCK, which is still present in these animals, does not play any appreciable role in determining gluconeogenetic flux, which would not be surprising given that this isoform accounts for only ∼2% of activity in the liver of mice. By using both NMR spectroscopy and metabolic tracers, She et al. (45She P. Burgess S.C. Shiota M. Flakoll P. Donahue E.P. Malloy C.R. Sherry A.D. Magnuson M.A. Diabetes. 2003; 52: 1649-1654Crossref PubMed Scopus (88) Google Scholar) have also shown that mice lacking PEPCK in the liver have impaired whole-body gluconeogenesis from phosphoenolpyruvate, but not from glycerol, and that there is a global increase in tricarboxylic acid cycle activity, even though pyruvate cycling and anaplerosis is decreased. Interestingly, NMR experiments using 2H2O and [13C]propionate indicate that total body gluconeogenic flux from phosphoenolpyruvate is still 66% of that in a normal mouse, despite the absence of hepatic PEPCK. These results suggest that extra-hepatic tissues may be capable of fulfilling the void in gluconeogenesis by synthesizing more than the usual amount of glucose. As mentioned before, both the kidney and intestine, two other sites of PEPCK gene expression, may contribute significantly to systemic glucose production during certain metabolic states (12Rajas F. Croset M. Zitoun C. Montano S. Mithieux G. Diabetes. 2000; 49: 1165-1168Crossref PubMed Scopus (88) Google Scholar, 19Gerich J.E. Meyer C. Woerle H.J. Stumvoll M. Diabetes Care. 2001; 24: 382-391Crossref PubMed Scopus (425) Google Scholar). Indeed, renal gluconeogenesis itself has long been known to be sufficient to maintain nearly normal blood glucose levels in eviscerated rats (46Smith O.K. Long C.N. Proc. Natl. Acad. Sci. U. S. A. 1971; 68: 1618-1622Crossref PubMed Scopus (9) Google Scholar). More recently, it has been shown that, in humans, endogenous glucose production falls only by about 50% upon removal of the liver (47Lauritsen T.L. Grunnet N. Rasmussen A. Secher N.H. Quistorff B. J. Hepatol. (Baltimore). 2002; 36: 99-104Abstract Full Text Full Text PDF PubMed Scopus (26) Google Scholar, 48Joseph S.E. Heaton N. Potter D. Pernet A. Umpleby M.A. Amiel S.A. Diabetes. 2000; 49: 450-456Crossref PubMed Scopus (60) Google Scholar). Thus, when glucose cannot be made by the liver, compensatory adaptations in either the kidney and/or intestine appear to be invoked that allow whole body glucose production to continue. It is interesting to note that the plasma glutamine concentration is increased 2-fold in mice without PEPCK in the liver, suggesting that the kidney and/or intestine may be capable of using glutamine as a substrate to produce glucose (20Cersosimo E. Garlick P. Ferretti J. Diabetes. 2000; 49: 1186-1193Crossref PubMed Scopus (78) Google Scholar, 49Stumvoll M. Meyer C. Perriello G. Kreider M. Welle S. Gerich J. Am. J. Physiol. 1998; 274: E817-E826PubMed Google Scholar, 50Stumvoll M. Perriello G. Meyer C. Gerich J. Kidney Int. 1999; 55: 778-792Abstract Full Text Full Text PDF PubMed Scopus (157) Google Scholar). The gene expression levels versus blood glucose plots in Fig. 2 illustrate that there are striking differences in how alterations in either GK or PEPCK gene expression affect the blood glucose concentration. These markedly different glucose response profiles reflect differences in the control that GK and PEPCK exert over glycolytic and gluconeogenic flux, respectively. Whereas GK exhibits near total control over glycolysis in tissues that express the enzyme, the situation is far different for PEPCK, because large changes in PEPCK gene expression within mice do not cause fasting hypoglycemia. However, as compelling as these data may seem, it remains important to view it with some caution because a systematic analysis of gluconeogenic flux in animals with different amounts of PEPCK has not yet been performed. Nonetheless, the results obtained to date are striking enough on their own to suggest that PEPCK does not, itself, determine gluconeogenic flux. Rather, it is likely that control is distributed among other enzymes in this pathway. Recently it has been shown that PGC-1, a nuclear co-regulator that affects gene expression by binding to transcription factors, is induced in the liver by fasting and by insulin deficiency. Forced overexpression of PGC-1 in rats, via use of a PGC-1 expressing recombinant adenovirus, causes increased hepatic glucose production (51Yoon J.C. Puigserver P. Chen G. Donovan J. Wu Z. Rhee J. Adelmant G. Stafford J. Kahn C.R. Granner D.K. Newgard C.B. Spiegelman B.M. Nature. 2001; 413: 131-138Crossref PubMed Scopus (1483) Google Scholar). PGC-1, besides enhancing PEPCK gene expression, also leads to an increase in glucose-6-phosphatase and fructose-1,6-biphosphatase mRNAs, two other key enzymes in the gluconeogenic pathway. Moreover, mice that carry a targeted disruption of the cAMP response element-binding (CREB) protein gene or that express a dominant negative CREB inhibitor, exhibited both fasting hypoglycemia and a decrease in glucose-6-phosphatase, fructose-1,6-bisphosphatase, pyruvate carboxylase, and PEPCK mRNAs (52Herzig S. Long F. Jhala U.S. Hedrick S. Quinn R. Bauer A. Rudolph D. Schutz G. Yoon C. Puigserver P. Spiegelman B. Montminy M. Nature. 2001; 413: 179-183Crossref PubMed Scopus (1098) Google Scholar). These studies provide both additional support and a mechanism whereby the control of gluconeogenesis may involve the synchronous regulation of a set of genes. Thus, the control of gluconeogenic flux should be considered a more global process, consistent with the findings of others who have studied this issue in the past (53Nordlie R.C. Foster J.D. Lange A.J. Annu. Rev. Nutr. 1999; 19: 379-406Crossref PubMed Scopus (418) Google Scholar, 54Pilkis S.J. Claus T.H. Annu. Rev. Nutr. 1991; 11: 465-515Crossref PubMed Scopus (145) Google Scholar, 55Chen X. Iqbal N. Boden G. J. Clin. Invest. 1999; 103: 365-372Crossref PubMed Scopus (232) Google Scholar, 56Owen M.R. Halestrap A.P. Biochim. Biophys. Acta. 1993; 1142: 11-22Crossref PubMed Scopus (35) Google Scholar). This conclusion is also fully consistent with metabolic control theory that provides a conceptual basis whereby the regulation of metabolic flux can be distributed over multiple enzymatic steps. It is truly fortunate that the mechanisms governing glucose homeostasis in the mouse are strikingly similar to those in the human and that precise genetic alterations can now be readily engineered that allow the mouse to serve as a model system for assessing the role of specific enzymes in determining metabolic flux. Future use of conditional gene targeting strategies that make use of Cre recombinase, as well as other site-specific recombinases, will undoubtedly prove useful. In addition, the overexpression of enzymes via the generation of transgenic mice by pronuclear DNA microinjection promises to remain useful because, as demonstrated by the studies we have discussed, it is useful to be able to increase the expression of specific enzymes so as to complement results from KO mice.