**3.2.2 Genetic overexpression of GK in the liver**

Several liver GK gain-of-function studies, both using transgenic animals and by means of adenovirus gene transfer, have been performed in healthy animals and models of diabetes such as streptozotocin induced type I and type II induced by ingestion of high fat/high carbohydrate diet. Due to heterogeneity, these studies will be examined according to the animal model and analysis conditions.



Table 2. Hepatic GK overexpression studies in healthy fed animals. Comments: decrease, increase and no change are referred to control group. "~" means no statistically significant; "-", no determined; "CMV", cytomegalovirus; "PEPCK-C", cytosolic phosphoenolpyruvate carboxykinase; "L-PK", liver pyruvate kinase; "TAT", tyrosine aminotransferase; "ACC1", Acetyl-Coenzyme A carboxylase 1; "FAS", fatty acid synthase; and "G6Pase", glucose-6 phosphatase.

In these models, enhancing hepatic glucose uptake by GK overexpression results in a direct reduction of glycaemia. As a consequence of lower blood glucose levels, pancreatic β-cell

Liver Glucokinase and Lipid Metabolism 245

In fast state, the influence of hepatic GK overexpression on glycaemia is not clear. Hariharan et al showed a decrease in glycaemia, accompanied by a decrease in insulinemia that could explain a reduction of glycolysis in skeletal muscle, causing the observed decline in serum lactate. Low insulin levels can also explain the increment of blood fatty acids. Interestingly, 20-weeks old mice were smaller than controls and presented reduced body mass index. On the contrary, long-term analysis of transgenic mice developed by Ferre et al showed that increasing GK activity in the liver lead to hepatic steatosis, hyperglycemia, hyperinsulinemia, obesity and insulin resistance. On the other hand, adenoviral gene transfer models for hepatic GK overexpression in fasting revealed induction of lipogenesis and consequently a tendency to increase blood triglycerides, without affecting glycaemia. c. Studies on hepatic GK overexpression in the context of type 1 diabetes mellitus: This is an autoimmune disease with specific destruction of insulin-producing β-cells in the pancreas, and results in loss of insulin production. As insulin stimulates the transcription of Gck gene in the liver, type 1 diabetic subjects do not have GK protein in their livers and consequently hepatic glucose metabolism is impaired. Gene therapy has been tested to restore liver glucose uptake capacity by increasing hepatic GK protein (Ferre, 1996a; Morral, 2002, 2007). In type 1 diabetic liver, all models present a similar phenotype. When restoring glucose signaling in diabetic hepatocytes via GK, glucose catabolic pathways are induced and, on the contrary, hepatic glucose production is inhibited. Consequently there is a reduction of diabetic hyperglycemia accompanied by incremented hepatic glycogen depots and de novo lipogenesis. Decreasing blood glucose levels forces muscle and adipose tissue to use fatty acids as energetic substrates, and in consequence, serum fatty acids are decreased in type 1 diabetic mice expressing GK in the liver. Lower blood fatty acids, together with increased glucose metabolism in the liver, inhibit hepatic β-oxidation of fatty acids. Therefore, these models suggest that hepatic overexpression of GK in type 1 diabetes leads to normoglycaemia thanks to increments in hepatic glucose uptake and fatty acid oxidation in peripheral tissues. d. Finally, hepatic GK overexpression in the context of type 2 diabetes: type 2 diabetes is a complex metabolic disorder caused by two physiologic defects: insulin resistance in combination with insulin secretion deficiency. Type 2 diabetes is characterized by glucose metabolism alterations such as failure of insulin to inhibit hepatic gluconeogenesis and impaired skeletal muscle glucose uptake. However, lipid metabolism is also altered. This is often reflected by increased circulating free fatty acids and triglycerides together with increased fat accumulation in non-adipose tissues. Thus, changes in the equilibrium between glucose and fatty acid metabolism in liver and muscle could be responsible for glucose homeostasis alterations. Obesity, hyperinsulinemia, in combination with hyperglycemia, inhibits fatty acid oxidation in many tissues. As a result, lipogenesis is favored over fatty acid oxidation leading to an increase in fat accumulation and a decrease in energy expenditure. A hypothetical strategy for type 2 diabetes therapy is increasing glucokinase activity, with the aim of enhancing glucose uptake in the liver that could contribute to gluconeogenesis inhibition with consequent restoration of glycaemia. If glycaemia is restored, plasma insulin levels could be secondarily lowered and it could be able to elevate energy

However, liver GK activity is increased in mild type 2 diabetes, but diminished in morbid obese diabetic patients. Animal diabetic models linked to obesity, show that GK deficiency in the liver occurs only in the case of obesity, and in severe or long-term forms of the disease. Although hepatic GK expression is different depending on disease stage, some strategies

expenditure and reduce obesity.

secretes less insulin. Therefore, a decrease in insulinemia is a secondary effect of increasing hepatic GK activity. But, O'Doherty et al demonstrate that the influence of GK activity on blood glucose and insulin levels could be dose-dependent, as it occurs only with high doses of their transgene. In the hepatocyte, glucose-6-phosphate derived from GK activity is directed to glycogen synthesis and, consequently, hepatic glycogen levels are increased in the study by Ferre et al. However, glycogen content is not modified by GK overexpression in the study of O'Doherty et al, maybe for intrinsic limitations. In both animal models, increasing GK activity results in glucose signaling that activates transcription of glycolytic and lipogenic genes. Lipogenic proteins together with high availability of citrate and ATP (derived from augmented glucose metabolism) lead to enhanced *de novo* lipogenesis in the liver, and consequently, higher secretion of VLDL to bloodstream that could explain the observed increase in blood triglycerides. Augmented blood fatty acids might be explained by insulin levels; low levels of this hormone result in low inhibition of lipolysis in the adipose tissue, and consequently fatty acids raise in the bloodstream. Importantly, Ferre et al show that long-term GK overexpression drives to hyperinsulinemia and hepatic steatosis.

b. Studies of GK overexpression in the liver of fasted, healthy mice are listed in Table 3 (Hariharan, 1997; O'Doherty, 1999; Desai, 2001; Ferre, 2003 & Scott, 2003)


Table 3. Hepatic GK overexpression studies in healthy fasted animals. Comments: decrease, increase and no change are referred to control group. "~" means no statistically significant; "-", no determined; "CMV", cytomegalovirus; "RSV", rose sarcoma virus; "apoA1-SV40", apolipoprotein A1 enhancer and simian vacuolating virus 40 promoter; "PEPCK-C", cytosolic phosphoenolpyruvate carboxykinase; "L-PK", liver pyruvate kinase; "ACC1", Acetyl-Coenzyme A carboxylase 1; "PFK-2", 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase.

secretes less insulin. Therefore, a decrease in insulinemia is a secondary effect of increasing hepatic GK activity. But, O'Doherty et al demonstrate that the influence of GK activity on blood glucose and insulin levels could be dose-dependent, as it occurs only with high doses of their transgene. In the hepatocyte, glucose-6-phosphate derived from GK activity is directed to glycogen synthesis and, consequently, hepatic glycogen levels are increased in the study by Ferre et al. However, glycogen content is not modified by GK overexpression in the study of O'Doherty et al, maybe for intrinsic limitations. In both animal models, increasing GK activity results in glucose signaling that activates transcription of glycolytic and lipogenic genes. Lipogenic proteins together with high availability of citrate and ATP (derived from augmented glucose metabolism) lead to enhanced *de novo* lipogenesis in the liver, and consequently, higher secretion of VLDL to bloodstream that could explain the observed increase in blood triglycerides. Augmented blood fatty acids might be explained by insulin levels; low levels of this hormone result in low inhibition of lipolysis in the adipose tissue, and consequently fatty acids raise in the bloodstream. Importantly, Ferre et al show that long-term GK overexpression drives to hyperinsulinemia and hepatic

b. Studies of GK overexpression in the liver of fasted, healthy mice are listed in Table 3

**Animal model** *M. musculus M. musculus M. musculus R. norvegicus* 

injection

**Promoter** apoA1-SV40 RSV PEPCK-C CMV

**Glycaemia** Decrease no change - no change **Blood lactate** Decrease no change - ~ decrease **Blood triglycerides** no change no change increase increase **NEFA** ~ increase no change - no change **Insulin** Decrease decrease increase no change **Hepatic glucose-6-P** - - ~ increase - **Hepatic glycogen** - - no change increase **Hepatic triglycerides** - - increase -

Table 3. Hepatic GK overexpression studies in healthy fasted animals. Comments: decrease, increase and no change are referred to control group. "~" means no statistically significant; "-",

apolipoprotein A1 enhancer and simian vacuolating virus 40 promoter; "PEPCK-C", cytosolic

no determined; "CMV", cytomegalovirus; "RSV", rose sarcoma virus; "apoA1-SV40",

phosphoenolpyruvate carboxykinase; "L-PK", liver pyruvate kinase; "ACC1", Acetyl-Coenzyme A carboxylase 1; "PFK-2", 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase.

**Desai 2001 Ferre 2003 O'Doherty 1999** 

12 months 4-5 days post-

Transgenic Adenovirus Transgenic Adenovirus

x5 x1.5 x2 x2.1 or x3


**Scott 2003** 

injection

No change: PEPCK-C, PFK-2

(Hariharan, 1997; O'Doherty, 1999; Desai, 2001; Ferre, 2003 & Scott, 2003)

**1997** 

**Age at analysis** 5 weeks 3 weeks post-

steatosis.

**GK activity over** 

**Modulation of enzymes and transcription factors** 

**control** 

**Study variables Hariharan** 

In fast state, the influence of hepatic GK overexpression on glycaemia is not clear. Hariharan et al showed a decrease in glycaemia, accompanied by a decrease in insulinemia that could explain a reduction of glycolysis in skeletal muscle, causing the observed decline in serum lactate. Low insulin levels can also explain the increment of blood fatty acids. Interestingly, 20-weeks old mice were smaller than controls and presented reduced body mass index. On the contrary, long-term analysis of transgenic mice developed by Ferre et al showed that increasing GK activity in the liver lead to hepatic steatosis, hyperglycemia, hyperinsulinemia, obesity and insulin resistance. On the other hand, adenoviral gene transfer models for hepatic GK overexpression in fasting revealed induction of lipogenesis and consequently a tendency to increase blood triglycerides, without affecting glycaemia.


However, liver GK activity is increased in mild type 2 diabetes, but diminished in morbid obese diabetic patients. Animal diabetic models linked to obesity, show that GK deficiency in the liver occurs only in the case of obesity, and in severe or long-term forms of the disease. Although hepatic GK expression is different depending on disease stage, some strategies

Liver Glucokinase and Lipid Metabolism 247

d. Transgene dose: Desai et al and O'Doherty et al described different metabolic impact of hepatic GK overexpression depending on the dose of transgene that they used. In our laboratory we aimed to re-examine the conclusions of these studies and the differentiated effects that GK activity could have on the metabolism, clearly differentiated, of periportal and perivenous hepatocytes. To evaluate the issue, we have developed a hydrodynamic gene transfer technique that served us to pursuit GK overexpression studies exclusively in perivenous liver (Liu, 1999; Zhang, 1999; Gomez-Valades, 2006; Budker, 2006 & Suda, 2007). With the injection of a plasmid for green fluorescent protein (GFP) and immunohistochemistry for PEPCK (periportal marker), we could visualize that hydrodynamic injection generate two separate populations of hepatocytes: green hepatocytes that expressed GFP and red hepatocytes showing PEPCK-C staining (Figure 4). We could conclude that in our conditions the hydrodynamic gene transfer technique delivered the transgene only in the

the periportal area of the liver during fasting, and not in fed state.

hepatocytes surrounding the central vein of the liver acinus.

a periportal marker.

**A B C**

Fig. 4. Visualization of liver transfection achieved with adenoviral and hydrodynamic gene transfer techniques. (A) Healthy mice were injected with 5.5·109 IU of an adenovirus that codified for the green fluorescent protein (GFP). Green fluorescence was observed in liver sections (200X), demonstrating a homogeneous presence of the transgene all over the liver acinus. (B) A plasmid for GFP was hydrodynamically injected to healthy mice and, as it can see appreciated in liver sections, resulted in non-homogenous green fluorescence signal. (C) Slices from hydrodynamically-injected mice were immunostained for PEPCK-C (red signal),

Our results represent the first attempt to overexpress pGK in perivenous hepatocytes. The first approach was the hydrodynamic injection of a plasmid with the Gck gene to healthy mice (Vidal-Alabró; publication pending). Forty-eight hours post-injection, increased GK in perivenous hepatocytes had clear effects on glucose homeostasis (Figure 5A). There was a reduction of glycaemia and insulinemia in the fed state, probably as a direct consequence of increased hepatic glucose uptake. Therefore perivenous GK gain of function reproduced results of periportal GK (Ferre, 1996), and liver-homogeneous GK overexpression (O'Doherty, 1999; Desai, 2001 & Scott, 2003). However, 16 hours-fasted mice did not show differences in blood glucose and insulin levels (data not shown), as Desai et al and O'Doherty et al had obtained with adenoviral GK transfer. Fifty days post-injection, perivenous GK overexpressing-mice presented blood glucose levels similar to control animals but accompanied by hyperinsulinism (Figure 5B). Long-term augmented GK activity in perivenous liver resulted in hepatic insulin resistance, since mice presented a phenotype very similar to liver-specific insulin receptor knock-out mice named LIRKO (Michael, 2000). Briefly, hyperinsulinism was probably due to reduced hepatic insulin clearance. Since peripheral tissues were still insulin-sensitive, hyperinsulinism inhibited lipolysis and induced lipogenesis in adipose tissue. Adipose tissue function together with

and insulin. Therefore, hepatic GK transgenic mice described by Ferre et al express GK at

based on increasing GK activity in the liver have been tested in some models of high fat diet induced type 2 diabetes (Desai, 2001 & Ferre, 2003) , in obesity models (Wu, 2005 & Torres, 2009) and in transgenic mice with hepatic insulin resistance (Okamoto, 2007). All these studies have in common that the increase in hepatic GK activity produces glycaemia normalization. Hepatic GK, through glycolysis and glycogenesis activation, increases blood glucose clearance while it inhibits hepatic glucose production. On the other hand, liver GK activity results in increased malonyl-CoA, a lipogenic substrate and inhibitor of β-oxidation. It is difficult to draw clear conclusions when evaluating consequences of liver GK overexpression on lipid metabolism in type 2 diabetic models. Wu et al report an expected increase in hepatic and serum triglycerides, together with higher serum fatty acids. However, Wu et al report that, although hepatic fatty acid β-oxidation was decreased, muscle increased fatty acid oxidation as a consequence of lower glycaemia and insulinemia. Conversely, Desai et al showed no changes in hepatic and serum lipid levels. Otherwise, Torres et al & Okamoto et al obtained an increase in serum triglycerides with no changes in fatty acid levels. The most striking model is presented by Ferre et al: under high fat diet, liver GK-transgenic mice became insulin resistant faster than controls and showed hepatic steatosis. It contrasts with results obtained in GK gene locus transgenic mice (Shiota, 2001). Besides exhibiting a reduction of the blood glucose concentration, mice with a greater than normal amount of GK also exhibited a dramatic resistance to the development of hyperglycemia and hyperinsulinemia normally brought on by consumption of a high fat diet.

Taken together, all these models have convincingly demonstrated that increasing GK protein in the liver leads to a direct reduction of glycaemia, but sometimes it can be accompanied with the risk of serious alterations in lipid metabolism deriving in hepatic steatosis and/or overt dyslipidemia. This aspect is essential when considering the possibility of using GK overexpression in the liver for diabetes therapy. At this point, it would be important to find out which are the causes of the different phenotypes observed in those animal models of hepatic GK overexpression previously described. There are several possible reasons:


based on increasing GK activity in the liver have been tested in some models of high fat diet induced type 2 diabetes (Desai, 2001 & Ferre, 2003) , in obesity models (Wu, 2005 & Torres, 2009) and in transgenic mice with hepatic insulin resistance (Okamoto, 2007). All these studies have in common that the increase in hepatic GK activity produces glycaemia normalization. Hepatic GK, through glycolysis and glycogenesis activation, increases blood glucose clearance while it inhibits hepatic glucose production. On the other hand, liver GK activity results in increased malonyl-CoA, a lipogenic substrate and inhibitor of β-oxidation. It is difficult to draw clear conclusions when evaluating consequences of liver GK overexpression on lipid metabolism in type 2 diabetic models. Wu et al report an expected increase in hepatic and serum triglycerides, together with higher serum fatty acids. However, Wu et al report that, although hepatic fatty acid β-oxidation was decreased, muscle increased fatty acid oxidation as a consequence of lower glycaemia and insulinemia. Conversely, Desai et al showed no changes in hepatic and serum lipid levels. Otherwise, Torres et al & Okamoto et al obtained an increase in serum triglycerides with no changes in fatty acid levels. The most striking model is presented by Ferre et al: under high fat diet, liver GK-transgenic mice became insulin resistant faster than controls and showed hepatic steatosis. It contrasts with results obtained in GK gene locus transgenic mice (Shiota, 2001). Besides exhibiting a reduction of the blood glucose concentration, mice with a greater than normal amount of GK also exhibited a dramatic resistance to the development of hyperglycemia and hyperinsulinemia normally brought on

Taken together, all these models have convincingly demonstrated that increasing GK protein in the liver leads to a direct reduction of glycaemia, but sometimes it can be accompanied with the risk of serious alterations in lipid metabolism deriving in hepatic steatosis and/or overt dyslipidemia. This aspect is essential when considering the possibility of using GK overexpression in the liver for diabetes therapy. At this point, it would be important to find out which are the causes of the different phenotypes observed in those animal models of

a. Species-specific results: one possibility is that GK overexpression in mouse liver may be more effective stimulating glucose disposal than the same degree of expression in a

b. Side-effects of gene transfer technology: when using adenoviral gene transfer, adenoviruses involve *per se* hepatic metabolic changes. When using transgenic, germline manipulated animals overexpress GK throughout life, including intrauterine life, possibly resulting in compensatory changes in insulin secretion, insulin action, or in other metabolic variables that do not occur with acute manipulation of GK via

c. Promoter that directs transgene expression can affect two important variables. On one hand, taking into account the metabolic hepatic zonation concept (Jungermann, 1995), the promoter determines which set of hepatocytes express the transgene. It is well known that physiological GK expression predominates in the perivenous area of the liver (Moorman, 1991; Jungerman, 1995 & Jungerman, 2000). However, most studies of hepatic GK gain of function did not use perivenous promoters. For instance, Ferre et al used a PEPCK promoter that directs the transgene to the periportal area of the liver, specialized in gluconeogenesis. In contrast, RSV or CMV promoters are ubiquitous promoters that transfect both perivenous and periportal hepatocytes. On the other hand, promoter directs the regulation of transgene expression by nutrients and hormones. For instance, GK under the PEPCK promoter is expressed under glucagon signaling and is inhibited by glucose

hepatic GK overexpression previously described. There are several possible reasons:

by consumption of a high fat diet.

larger animal such as rat.

adenovirus technology.

and insulin. Therefore, hepatic GK transgenic mice described by Ferre et al express GK at the periportal area of the liver during fasting, and not in fed state.

d. Transgene dose: Desai et al and O'Doherty et al described different metabolic impact of hepatic GK overexpression depending on the dose of transgene that they used.

In our laboratory we aimed to re-examine the conclusions of these studies and the differentiated effects that GK activity could have on the metabolism, clearly differentiated, of periportal and perivenous hepatocytes. To evaluate the issue, we have developed a hydrodynamic gene transfer technique that served us to pursuit GK overexpression studies exclusively in perivenous liver (Liu, 1999; Zhang, 1999; Gomez-Valades, 2006; Budker, 2006 & Suda, 2007). With the injection of a plasmid for green fluorescent protein (GFP) and immunohistochemistry for PEPCK (periportal marker), we could visualize that hydrodynamic injection generate two separate populations of hepatocytes: green hepatocytes that expressed GFP and red hepatocytes showing PEPCK-C staining (Figure 4). We could conclude that in our conditions the hydrodynamic gene transfer technique delivered the transgene only in the hepatocytes surrounding the central vein of the liver acinus.

Fig. 4. Visualization of liver transfection achieved with adenoviral and hydrodynamic gene transfer techniques. (A) Healthy mice were injected with 5.5·109 IU of an adenovirus that codified for the green fluorescent protein (GFP). Green fluorescence was observed in liver sections (200X), demonstrating a homogeneous presence of the transgene all over the liver acinus. (B) A plasmid for GFP was hydrodynamically injected to healthy mice and, as it can see appreciated in liver sections, resulted in non-homogenous green fluorescence signal. (C) Slices from hydrodynamically-injected mice were immunostained for PEPCK-C (red signal), a periportal marker.

Our results represent the first attempt to overexpress pGK in perivenous hepatocytes. The first approach was the hydrodynamic injection of a plasmid with the Gck gene to healthy mice (Vidal-Alabró; publication pending). Forty-eight hours post-injection, increased GK in perivenous hepatocytes had clear effects on glucose homeostasis (Figure 5A). There was a reduction of glycaemia and insulinemia in the fed state, probably as a direct consequence of increased hepatic glucose uptake. Therefore perivenous GK gain of function reproduced results of periportal GK (Ferre, 1996), and liver-homogeneous GK overexpression (O'Doherty, 1999; Desai, 2001 & Scott, 2003). However, 16 hours-fasted mice did not show differences in blood glucose and insulin levels (data not shown), as Desai et al and O'Doherty et al had obtained with adenoviral GK transfer. Fifty days post-injection, perivenous GK overexpressing-mice presented blood glucose levels similar to control animals but accompanied by hyperinsulinism (Figure 5B). Long-term augmented GK activity in perivenous liver resulted in hepatic insulin resistance, since mice presented a phenotype very similar to liver-specific insulin receptor knock-out mice named LIRKO (Michael, 2000). Briefly, hyperinsulinism was probably due to reduced hepatic insulin clearance. Since peripheral tissues were still insulin-sensitive, hyperinsulinism inhibited lipolysis and induced lipogenesis in adipose tissue. Adipose tissue function together with

Liver Glucokinase and Lipid Metabolism 249

In the context of type 1 diabetes induced with streptozotocin, perivenous liver GK gain of function restored hepatic glucose uptake and reduced gluconeogenesis. Therefore, typical increases in hepatic glucose depots (glycogen, triglyceride) occurred and resulted in a reduction of diabetic glycaemia, albeit small. But, perivenous GK expressing mice showed a significant increase in triglyceride and free fatty acid serum concentration, and hepatic lipids (Figure 6) (Vidal-Alabró; publication pending). Therefore our work in type 1 diabetes model reproduces those of periportal GK overexpression (Ferre, 1996a) and those of liver homogeneous GK overexpression (Morral, 2002, 2007) in terms of glycaemia. However, our results on lipid metabolism are more deleterious, probably because perivenous hepatocytes

have higher lipogenic potential than periportal hepatocytes (Jungermann, 1995).

**1.0 \***

**0**

**50**

**Triglycerides (mg / dL)**

**<sup>100</sup> \***

**0.0**

**0**

**5**

**10**

**Triglyceride (mg/g liver)**

**C D**

**15**

**0.5**

**Free fatty acids (mM)**

**0**

**<sup>20</sup> \***

**10 20**

**mRNA (relative units)**

*Fasn*

\*

**0**

**10**

**Glycogen (mg/g liver)**

**B**

*Mod1*

and \*\*\* p<0.001 vs control, determined by t-Student.

*Srebf*

*Nr1h3*

*Chrebp*

Fig. 6. Analysis of perivenous GK overexpression in type 1 diabetic mice. (A) Shows glycaemia, serum triglycerides and free fatty acids, 48 hours post-injection of the plasmid that contained the GK gene. Columns represent media ± standard error. (B) Hepatic glucose storage was evaluated by measuring glycogen and triglyceride levels. (C) Expression of lipogenic genes in the liver was analyzed by Real-Time PCR. Calculations were done

**0.0**

following ΔΔCt algorithm (Applied Biosystems), using β-microglobulin gene expression as a housekeeping gene. (D) The same for gluconeogenic and lipolysis genes. **\*** p<0.05, **\*\***p< 0.01

*Pck1*

*G6pc*

*Hnf4*

*Ppargc1*

\*\*\*

\* \* \* \*\*

*Cpt1a*

pControl pGK

*Ucp2*

**0.5**

**1.0**

**1.5 2.5**

**250**

**glycemia (mg / dL)**

**500**

**A**

reduced hepatic lipogenesis *de novo* could explain the observed decrease in circulating triglycerides and free fatty acids. Although having increased GK activity in the liver, neither glycogen synthesis nor glycolysis was stimulated in those mice. Besides, gluconeogenesis was not inhibited in fed state. Therefore, considering the bibliography, our perivenous model resembled transgenic mice that expressed GK transgene under PEPCK-C promoter at periportal hepatocytes (Ferre, 2003). However, periportal GK overexpressing model showed whole-body insulin resistance linked to obesity and hepatic steatosis. It must be considered that their analysis was in 12 months old mice. If the study was extended to 12 months, we would be able to tell if hepatic insulin resistance observed in our mice model leads to general insulin resistance or, on the contrary, confirm its resemblance to LIRKO animals.

Fig. 5. Analysis of GK-overexpressing healthy mice. (A) Shows glycaemia and insulinemia, 48 hours post-injection of the plasmid that contained the GK gene. Columns represent media ± standard error. (B) 50 days post-injection results on serum nutrients (glucose, free fatty acids and triglycerides) together with insulin levels are represented. (C) After 50 days post-injection, expression of glycolytic and gluconeogenic genes from liver were analyzed by Real-Time PCR. Calculations were done following ΔΔCt algorithm (Applied Biosystems), using β-microglobulin gene expression as a housekeeping gene. (D) The same for lipogenic and lipolysis genes. **\*** p<0.05 and **\*\***p< 0.01 vs control, determined by t-Student.

reduced hepatic lipogenesis *de novo* could explain the observed decrease in circulating triglycerides and free fatty acids. Although having increased GK activity in the liver, neither glycogen synthesis nor glycolysis was stimulated in those mice. Besides, gluconeogenesis was not inhibited in fed state. Therefore, considering the bibliography, our perivenous model resembled transgenic mice that expressed GK transgene under PEPCK-C promoter at periportal hepatocytes (Ferre, 2003). However, periportal GK overexpressing model showed whole-body insulin resistance linked to obesity and hepatic steatosis. It must be considered that their analysis was in 12 months old mice. If the study was extended to 12 months, we would be able to tell if hepatic insulin resistance observed in our mice model leads to general insulin resistance or, on the contrary, confirm its resemblance to LIRKO animals.

**0**

*Pfkfb3*

*Slc2a2*

**\***

*Myc*

**0.07**

*G6pc*

**0.0**

**2.5** \*

**0.0 0.5 1.0 1.5 2.0**

**Insulinemia (g/L)**

**0.5**

**Insulinemia (g/L)**

**1.0**

**\***

**1.5**

*Hnf4*

and lipolysis genes. **\*** p<0.05 and **\*\***p< 0.01 vs control, determined by t-Student.

*Ppargc1a*

*Foxa3*

Fig. 5. Analysis of GK-overexpressing healthy mice. (A) Shows glycaemia and insulinemia, 48 hours post-injection of the plasmid that contained the GK gene. Columns represent media ± standard error. (B) 50 days post-injection results on serum nutrients (glucose, free fatty acids and triglycerides) together with insulin levels are represented. (C) After 50 days post-injection, expression of glycolytic and gluconeogenic genes from liver were analyzed by Real-Time PCR. Calculations were done following ΔΔCt algorithm (Applied Biosystems), using β-microglobulin gene expression as a housekeeping gene. (D) The same for lipogenic

**0**

*Fasn*

*Mod1*

*Chrebp*

*Srebf1*

*Nr1h3*

*Pparg*

**\*\***

*Cpt1a*

*Hmgcs2*

**1**

**\***

**D**

**0.0**

**2**

**0.5**

**Free fatty acids (mM)**

**1.0**

**\***

**0**

**\* \*\* \***

**100**

**Triglycerides (mg/dL)**

pControl pGK

**\***

**200**

**1.5**

**1**

**2**

**mRNA (relative units)**

**4 6 8**

**0**

**Glycaemia (mg/dL)**

**C**

**B**

**50**

**100**

**Glycaemia (mg/dL)**

**150**

**A**

**\***

In the context of type 1 diabetes induced with streptozotocin, perivenous liver GK gain of function restored hepatic glucose uptake and reduced gluconeogenesis. Therefore, typical increases in hepatic glucose depots (glycogen, triglyceride) occurred and resulted in a reduction of diabetic glycaemia, albeit small. But, perivenous GK expressing mice showed a significant increase in triglyceride and free fatty acid serum concentration, and hepatic lipids (Figure 6) (Vidal-Alabró; publication pending). Therefore our work in type 1 diabetes model reproduces those of periportal GK overexpression (Ferre, 1996a) and those of liver homogeneous GK overexpression (Morral, 2002, 2007) in terms of glycaemia. However, our results on lipid metabolism are more deleterious, probably because perivenous hepatocytes have higher lipogenic potential than periportal hepatocytes (Jungermann, 1995).

Fig. 6. Analysis of perivenous GK overexpression in type 1 diabetic mice. (A) Shows glycaemia, serum triglycerides and free fatty acids, 48 hours post-injection of the plasmid that contained the GK gene. Columns represent media ± standard error. (B) Hepatic glucose storage was evaluated by measuring glycogen and triglyceride levels. (C) Expression of lipogenic genes in the liver was analyzed by Real-Time PCR. Calculations were done following ΔΔCt algorithm (Applied Biosystems), using β-microglobulin gene expression as a housekeeping gene. (D) The same for gluconeogenic and lipolysis genes. **\*** p<0.05, **\*\***p< 0.01 and \*\*\* p<0.001 vs control, determined by t-Student.

Liver Glucokinase and Lipid Metabolism 251

noteworthy loss in insulin secretion or changes in fasting blood glucose concentrations. Moreover, when challenged with a high-sucrose/high-fat diet the knock-out and normal mice gained body weight at a similar rate but the knock-out mice were hyperglycaemic and hyperinsulinemic. Importantly, no changes in plasma triglycerides and non-esterified fatty acids were observed in basal conditions as well as with a high-sucrose/high-fat diet. In summary, absence of GKRP results in decreased hepatic GK protein content, affecting

On the other hand, GKRP gain of function in the liver has also been assessed. In vitro studies with HepG2 cells simultaneously transduced with an adenoviral vector expressing GKRP and another adenoviral vector for GK had significantly elevated GK protein and activity levels compared with cells transduced with the GK adenovirus alone (Slosberg, 2001). These data suggest that GKRP serves to stabilize and protect a pool of GK protein (i.e., extend half-life), and is consistent with data obtained in GKRP knock-out studies. But in vivo studies revealed a more complicated situation. Adenoviral-mediated hepatic overproduction of GKRP in mice with high-fat diet-induced diabetes resulted in 23% decrease in GK enzymatic activity. Although reduction of GK activity is commonly associated to diabetes, hepatic GKRP-expressing mice had improved fasting and glucoseinduced glycaemia with a concomitant increase in insulin sensitivity and TAG levels, and a decrease in leptin levels. A possible explanation for discrepancies between in vivo and in vitro results on GK levels when overexpressing GKRP is that GK expression in vivo is influenced by insulin and other physiological regulators. To understand how decreased GK activity improved type 2 diabetes phenotype in this model, a possibility is that GK activity may be applied in a more efficient manner toward metabolizing blood glucose. The subcellular compartmentalization by scaffolding proteins of enzymes or signaling proteins

Coming back to genome-wide studies that associate Gckr with fast hypoglycemia and high triglycerides, Beer et al reported that P446L-GKRP has reduced regulation by physiological concentrations of fructose-6-phosphate, resulting indirectly in increased GK activity (Beer, 2009). They predicted that this increased GK activity in the liver enhanced glycolytic flux, promoting hepatic glucose metabolism and elevating concentrations of malonyl-CoA, a substrate for *de novo* lipogenesis, providing a mutational mechanism for the reported association of this variant with raised triglycerides and lower glucose levels. However, their predictions are conflictive with in vivo studies by Slosberg et al (Slosberg, 2001), since GKRP gain of function reduced hepatic GK activity and also resulted in a decrease of blood glucose levels accompanied by an increase of blood triglycerides. Therefore, any other undetermined factor/s must exist to really understand the complex physiology of the GK/GKRP system. Another possibility is that brain P446L-GKRP and β-cell P446L-GKRP (if

Another study that may bring light to this issue, relates to defects in glucokinase translocation identified in Zucker diabetic fatty (ZDF) (Fujimoto, 2004 & Shin, 2007). Although having normal GK protein content, GK was predominantly localized in the nucleus regardless of plasma glucose and insulin levels. Nevertheless, sorbitol restored GK translocation. Clearly, there must be two distinct mechanisms bringing about the dissociation of GK from GKRP. How they are related and what differentiate them are questions currently under investigation. Since this defect was discovered in early stage of diabetes, it could cause of the progression to diabetes seen in the adult ZDF rat. Consistently, a MODY-2 mutation in the Gck gene has been reported to increase the physical

glucose metabolism without disturbing lipid parameters.

into clusters is often used as a means of increasing system efficiencies.

existent) may exert determinant influences on phenotype.

All in all, our review of the literature together with our own results on the subject will convey that pGK-overexpression in the liver, independent of zonation, will result in changes in glycaemia but with the risk of non-desirable lipid alterations and insulin resistance. However, several undetermined factors influence the results obtained in GK overexpression studies, reinforcing the concept that hepatic GK is a key regulator of whole-body homeostasis, so that little changes in its activity and/or in its regulation affect glucose and lipid metabolism.
