Postprandial hyperglycemia and impaired hepatic glucose uptake in type 2 diabetes

Yoshifumi Tamura, Masataka Niwa, Hiroshi Uchino, Toyoyoshi Uchida and Ryuzo Kawamori
Department of Medicine, Metabolism and Endocrinology, Juntendo University School of Medicine, Tokyo, Japan

Introduction
One of the features of type 2 diabetes is a decrease in insulin-stimulated glucose uptake following a meal. In the postprandial state, the degree of the rise in blood glucose is determined by the difference between the amount of glucose entering and the amount leaving the circulation.
In healthy subjects, glucose is rapidly absorbed after oral glucose ingestion. As soon as the blood glucose concentration starts to rise, it is met by rapid pulsatile insulin secretion. Raised hepatic sinusoidal and peripheral insulin concentrations suppress hepatic glucose production and increase glucose uptake by insulin-dependent tissues. Glucose excursions are therefore kept within a narrow range as a result of the interaction between insulin and its target organs. It has been suggested that one-third of an oral glucose load is taken up by muscle and fat, one-third by the liver, and the rest by non-insulin-dependent tissues [1]. Thus, hepatic glucose uptake appears to play an important role in determining postprandial hyperglycemia.
In this review we will describe the physiological mechanisms of hepatic glucose uptake, its measurement, and the features of impaired hepatic glucose uptake in type 2 diabetes. We will also consider possible drug interventions to improve postprandial hyperglycemia by increasing hepatic glucose uptake. Finally, we will try to elucidate the biochemical mechanisms of impaired hepatic glucose uptake in type 2 diabetes.

Regulation of hepatic glucose uptake
It has been demonstrated that in conscious dogs, hepatic glucose uptake is regulated by three main factors: (1) hepatic sinusoidal insulin concentration, (2) the signal generated by the arterial–portal glucose gradient (the portal
signal), and (3) the amount of glucose reaching the liver (the hepatic glucose load) [1].
Myers and colleagues [2] investigated the relationship between hepatic sinusoidal insulin concentration and net hepatic glucose uptake during a twofold increase in hepatic glucose load in both the presence and absence of the portal signal (Fig. 1). The portal signal was generated in one group of dogs by infusing glucose intraportally, while infusing saline by the same route in the other group. In both groups, glucose was also infused into the peripheral vein to continuously increase the hepatic glucose load twofold. Somatostatin was given to inhibit endogenous insulin and glucagon secretion. Glucagon was maintained at basal values, while insulin concentrations were adjusted to three different levels. In both situations, with or without the portal signal, hepatic sinusoidal insulin level correlated with hepatic glucose uptake (Fig. 1).

Fig. 1: Relationship between net hepatic glucose uptake and hepatic sinusoidal insulin concentration in the presence and absence of portal glucose delivery in 42-h fasted, conscious dogs. Plasma glucagon concentration was maintained at basal values and insulin was varied using the pancreatic clamp technique. The amount of glucose reaching the liver was twofold that of basal values and equal in the two protocols.

Myers and colleagues [3] also established the relationship between hepatic glucose load and net hepatic glucose uptake. In fasted, conscious dogs, they used a somatostatin pancreatic clamp to keep the glucagon level at basal values and to increase the insulin level fourfold. A small amount of glucose was infused into the portal vein to generate a portal signal in one group but not in the other. To maintain the same hepatic glucose load in both groups of animals, glucose was infused into the peripheral vein. Thus a negative arterial–portal glucose gradient (portal signal) was present in one group but absent in the other, while the hepatic glucose load and hepatic insulin level were similar in both groups. Figure 2 shows the results of the study and illustrates the relationships between net hepatic glucose uptake, hepatic glucose load and the portal signal. It is clear that hepatic glucose uptake is dependent on the amount of glucose reaching the liver and is augmented by the portal signal.

Fig. 2: Relationship between net hepatic glucose uptake and hepatic (A) or portal (B) glucose loads in 42-h fasted, conscious dogs. Plasma glucagon concentration was kept at basal levels, while plasma insulin level was increased fourfold using the pancreatic clamp technique.

Although the portal signal clearly increased hepatic glucose uptake and was generated by a negative arterial–portal glucose gradient, the sites of glucose sensing were not determined. To establish the reference site of arterial glucose concentration, Matsuhisa et al. [4] infused glucose into the portal vein in the presence or absence of glucose in the central nervous system. In these studies, hepatic portal and central nervous system glucose gradients were measured in relation to the net hepatic glucose balance (NHGB; the difference between glucose efflux and influx). Glucose infusion into a peripheral vein resulted in a low NHGB (7.4 mmol/kg per min), indicating low hepatic glucose uptake. When glucose delivery was switched to the portal vein to generate the portal signal, NHGB increased to 41.5 mmol/kg per min, indicating increased hepatic glucose uptake. Infusion of glucose into the central nervous system, in addition to the portal vein, abolished the hepatic portal–central nervous system gradient and NHGB decreased to 21.7 mmol/kg per min. These results indicate that the portal signal is generated, at least partially, by the hepatic portal–central nervous system glucose gradient. Hsieh et al. [5] have shown, however, that the head arterial glucose level is not the reference site for generation of the portal signal; the site of glucose sensing has yet to be fully established.

Impaired hepatic glucose uptake in type 2 diabetes
Since at least one-third of an oral glucose load is taken up by the liver, impaired hepatic glucose uptake may contribute to postprandial hyperglycemia in type 2 diabetes. As hepatic glucose uptake is regulated by: (1) the hepatic sinusoidal insulin concentration, (2) the portal signal, and (3) the hepatic glucose load, hepatic glucose uptake in type 2 diabetes is assumed to be impaired partly by a lack of each factor after meal ingestion. Some experiments using the double tracer method have demonstrated the contribution of impaired hepatic glucose uptake to postprandial hyperglycemia. Using this method, Mitrakou et al. [6] reported that hyperglycemia, after oral glucose intake (1 g/kg body weight), resulted from a 6.8 g greater endog-enous glucose production (13.7 ± 1.1 g vs. 6.8 ± 1.0 g) and a 3.8 g lower glucose splanchnic sequestration (31.4 ± 1.5 g vs. 27.5 ± 0.9 g) in type 2 diabetic subjects compared with age- and weight-matched non-diabetic volunteers. Using the double tracer method, Firth et al. [7] also demonstrated excess glucose in diabetic subjects compared with control subjects. Their results showed that 13 g of extra glucose was endog-enous glucose produced by the liver and 2 g of extra glucose was due to the 50 g oral glucose intake. These two studies indicated the contribution of impaired hepatic glucose uptake to excess glucose, which induced postprandial hyperglycemia by 36% and 13%, respectively, in each study.
Other double tracer experiments, however, have suggested that splanchnic glucose uptake is normal [8] or increased [9] in type 2 diabetes. Thus it is not firmly established whether impaired hepatic glucose uptake, as determined by the double tracer method, does indeed contribute to postprandial hyperglycemia in type 2 diabetes.
This discrepancy could be explained by the mechanisms which regulate hepatic glucose uptake. In the above studies [6–9], after oral ingestion of glucose or a mixed meal, peripheral glucose concentrations were much higher in diabetic patients than in normal subjects. Therefore, impaired hepatic glucose uptake, induced by a lack of rapid pulsatile insulin secretion and portal signal deficiency [10], might be countered by a greater hepatic glucose load.
‘Accurate’ hepatic glucose uptake should therefore be determined under conditions in which the sinusoidal insulin level and hepatic glucose load are controlled. To evaluate hepatic glucose uptake, Kawamori and colleagues [11] developed an innovative non-invasive method — the euglycemic-hyperinsulinemic clamp combined with an oral glucose load. The validity of this method was subsequently confirmed by Ludvik et al. [12]. These studies demonstrated that splanchnic glucose disposal was impaired during euglycemic-hyperinsulinemic conditions with an oral glucose load [11, 12], and with enteral glucose administration [13], in poorly controlled type 2 diabetic patients.

Effect of oral hypoglycemic agents on hepatic glucose uptake
In patients with type 2 diabetes, portal insulin levels tend to be lower after meal ingestion than in normal subjects due to the lack of rapid pulsatile insulin secretion. Hepatic glucose uptake reaches a peak after only 15 min exposure to the portal signal in normal dogs [14]. However, a fourfold increase in insulin took 75 min to produce a maximal increase in hepatic glucose uptake in the same dogs [14]. Restoration of rapid pulsatile insulin secretion would thus seem to be very important in suppressing postprandial hyperglycemia, both by increasing the total amount of hepatic glucose uptake and by changing the timing of maximum glucose uptake.
We have examined the effects of restoring endogenous rapid insulin secretion using the newly developed insulin secretagogue, nateglinide (120 mg), during postprandial glycemic excursions in 10 obese Japanese patients with type 2 diabetes (BMI 28.4 kg/m2) [15]. Serum insulin concentrations were higher in the obese patients at 15 and 30 min and lower at 150 and 180 min after the oral glucose load than in the control subjects (Fig. 3).

Fig. 3: Glycemic responses (A) and serum insulin concentrations (B) during an OGTT with and without nateglinide in obese type 2 diabetic patients. *p < 0.001, †p < 0.01 and ‡p < 0.05 compared with controls. Data shown are mean ± SEM.

Thus, during the first 60 min, insulin secretion was augmented and glucose excursion curves were maintained within the near-normal range.
Two main mechanisms may reduce postprandial hyperglycemia via rapid pulsatile insulin secretion. One is the rapid inhibition of hepatic glucose production by pulsatile insulin secretion. In the early prandial phase, endogenous insulin secretion increases the portal insulin level, which indirectly suppresses hepatic glucose production by reducing lipolysis in visceral fat and regulating the glucagon response. At the same time, increased hepatic sinusoidal insulin directly suppresses hepatic glucose production [1]. The importance of this mechanism has been reported by several investigators using the double tracer method [6–8].
The second mechanism may be an alteration to the timing of maximum glucose uptake by rapid pulsatile insulin secretion. Interestingly, it was demonstrated that glucose disappearance was not significantly different in type 2 diabetic subjects and non-diabetic controls, although the insulin level was lower in the former [6, 7]. Thus, rapid pulsatile insulin secretion could decrease postprandial hyperglycemia by altering the timing of maximum glucose uptake. Pulsatile insulin secretion is clearly an important factor in the timing of hepatic glucose uptake, as insulin required a long time (75 min) to produce a maximal increase in hepatic glucose uptake [14].
We have also examined the effect of pioglitazone in type 2 diabetes in a double-blind, controlled trial [16]. Using the euglycemic-hyper-insulinemic clamp technique, the mean glucose infusion rate — an index of muscle glucose uptake — was significantly increased in patients taking pioglitazone: from 8.2 mg/kg per min during the run-in period to 9.2 mg/kg per min at the end of 12 weeks of treatment. Splanchnic glucose uptake, determined by the euglycemic-hyperinsulinemic clamp combined with an oral glucose load, significantly increased from 28.5% to 59.4% (Fig. 4); the difference between the pioglitazone and placebo treatments was significant.

Fig. 4: Changes in rates of splanchnic glucose uptake in type 2 diabetic patients before and after 12 weeks of treatment with pioglitazone or placebo. Data shown are mean values ± SD. *p < 0.05 vs. run-in period; †p < 0.05 pioglitazone vs. placebo.

This study demonstrated the effectiveness of pioglitazone at increasing both hepatic glucose uptake following meal ingestion and insulin-induced muscle glucose uptake.

Glucokinase (GK) translocation and hepatic glucose uptake
Unlike the physiological mechanisms of hepatic glucose uptake, the biochemical mechanism has yet to be fully elucidated. Some reports, how-ever, support an association between hepatic glucose uptake and glucokinase (GK) translocation in hepatocytes. Hepatic GK is bound to GK regulatory protein (GKRP) during basal glucose concentrations (~5.5 mM) and exists in the nucleus. However, when hepatocytes are exposed to high glucose (~10–30 mM) or fructose (~50 mM–1 mM) concentrations, GK is released from GKRP and moves to the cytosol, which is augmented by insulin [17].
Recent work has shown that there is a good correlation between the increase in cytoplasmic GK and the phosphorylation induced by fructose in cultured hepatocytes [18]. Thus GK, which is translocated from the nucleus to the cytosol in the presence of high glucose or fructose concentrations, catalyzes the phosphorylation of glucose and promotes glucose entry into the hepatocyte [1].
Under experimental conditions, intraportal infusion of a small amount of fructose markedly increased net hepatic glucose uptake in 42-h fasted, conscious dogs [19], and GK translocation from the nucleus to the cytosol was observed in the hepatocytes of these animals [19]. This research suggests that GK translocation is the principal factor regulating hepatic glucose uptake.