Portal Vein Glucose Entry Triggers a Coordinated Cellular Response That Potentiates Hepatic Glucose Uptake and Storage in Normal but Not High-Fat/High-Fructose–Fed Dogs

The protocol was approved by the Vanderbilt University Institutional Animal Care and Use Committee, and the animals were housed and cared for according to Association for Assessment and Accreditation of Laboratory Animal Care International guidelines. Adult male mongrel dogs were fed a standard meat (Kal Kan, Franklin, TN) and laboratory chow diet (control [CTR], n = 15) or a high-fat, high-fructose diet (HFFD, n = 16; PMI Nutrition TestDiet, St. Louis, MO) for 4 weeks, as described elsewhere (22). After 2 weeks of feeding, a subset of dogs (CTR, n = 10; HFFD, n = 11) underwent a laparotomy to insert sampling catheters into the hepatic portal vein, the left common hepatic vein, and the femoral artery and to insert infusion catheters into a jejunal and splenic vein (4). Ultrasonic blood flow cuffs (Transonic Systems, Ithaca, NY) were secured around the hepatic artery and hepatic portal vein for the measurement of hepatic blood flow, as described previously (4). After 4 weeks of feeding, another subset of dogs (CTR, n = 5; HFFD, n = 5) was killed following an 18-h fast for the acquisition of liver biopsies under basal conditions. All dogs studied were healthy, as indicated by: 1) hematocrit >35%; 2) leukocyte count <18,000/mm³; 3) good appetite; and 4) normal stools.

After 4 weeks of feeding, HIHG clamp experiments with (+) or without (−) PoG infusion were conducted on four groups (CTR−PoG [n = 5], HFFD−PoG [n = 5], CTR+PoG [n = 5], and HFFD+PoG HFFD [n = 6]) of 18-h–fasted dogs that had been fed one can of meat just prior to fasting (Fig. 1). Experiments consisted of a 100-min equilibration period (−120 to −20 min), a 20-min control period (−20 to 0 min), and a 180-min HIHG period divided into two subperiods (P1, 0–90 min; P2, 90–180 min). At −120 min, a priming dose of [3-³H] glucose (38 μCi) was given, followed by a constant infusion of [3-³H] glucose (0.38 μCi/min). At time 0, a constant infusion of somatostatin (0.8 µg/kg/min; Bachem, Torrance, CA) was started in a leg vein, and insulin and glucagon were then replaced intraportally at threefold basal (1.2 mU/kg/min; Lilly) and basal (0.55 ng/kg/min; Novo Nordisk) rates, respectively. A variable infusion of 50% dextrose was started in a leg vein in order to double the hepatic glucose load (HGL). In P2, normal saline (−PoG) or 20% dextrose (+PoG; 4.0 mg/kg/min) was infused intraportally. In the +PoG groups, the peripheral glucose infusion rate was adjusted as needed to clamp the HGL to that in P1. At the end of the study, animals were anesthetized with sodium pentobarbital, and liver sections were freeze-clamped and stored at −80°C.

RNA isolation, cDNA synthesis, quantitative PCR primers and analysis, and Western blotting procedures were performed as described previously (26,27). The GK antibody was a gift from Dr. Masakazu Shiota (Vanderbilt University School of Medicine, Molecular Physiology and Biophysics Department). All other antibodies were purchased from Santa Cruz Biotechnology (actin and GK regulatory protein [GKRP]) or Cell Signaling Technology (GS). Test protein bands were quantified using ImageJ software (National Institutes of Health; http://rsb.info.nih.gov/ij/).

GK activity was assessed as described previously (28). GS activity ratio was calculated as activity measured at low glucose-6-phosphate (160 µmol) versus high glucose-6-phosphate (6.7 mmol/L) concentrations, with low uridine diphosphate glucose (134 µmol), as described previously (29). GP activity was assessed by measuring the incorporation of [14C]-glucose from [14C]-glucose-1-phosphate into glycogen in the absence or presence of 2 mmol/L AMP (30).

Plasma glucose, [³H]-glucose, glucagon, insulin, and nonesterified fatty acid (NEFA) levels and blood lactate and glycerol concentrations were measured using standard methods as described previously (31,32). Liver glycogen levels were determined using the amyloglucosidase method described by Keppler and Decker (33).

Unidirectional HGU was calculated by multiplying the hepatic fractional extraction of [³H]-glucose (HFrExG*) by the HGL (mg/kg/min). HFrExG* (unitless) was determined by dividing the hepatic [³H]-glucose balance by the hepatic [³H]-glucose load according to the following equation:

where G*_A, G*_P, and G*_H represent [³H]-glucose values (dpm/mL) in the artery, portal, and hepatic veins, respectively, and BF_A, BF_P, and BF_H represent blood flow (ml/kg/min) in the hepatic artery, portal vein, and liver, respectively. We converted plasma [³H]-glucose values to blood [³H]-glucose using previously established conversion factors (7). HGL was calculated according to the following equation:

where G_A represents the unlabeled blood glucose concentration in the artery (mg/mL), BF_H represents total hepatic blood flow, PoGinf represents the portal glucose infusion rate (mg/kg/min), and GUG represents the uptake of glucose by the gut (mg/kg/min). GUG was calculated as follows:

where ([G*_A − G*_P] / G*_A) represents the fractional extraction of [³H]-glucose across the gut, G_A represents the unlabeled blood glucose concentration in the artery, and BF_P represents blood flow in the portal vein, respectively. It should be stressed that this approach yields HGU rather than net HGU.

Net hepatic substrate balances were calculated with the arteriovenous difference method as described previously (22). Positive net hepatic substrate balance values indicate net production, and negative values indicate net uptake. In the +PoG groups, a correction factor was applied to net hepatic glucose balance (NHGB) to account for the percent recovery of the PoG infusate in the hepatic vein. HGP was calculated as the difference between NHGB and HGU. Plasma insulin and glucagon levels entering the hepatic sinusoids were calculated as described elsewhere (27). The incorporation of glucose into glycogen through the direct pathway was calculated by dividing hepatic [³H]-labeled glycogen by the average inflowing plasma [³H]-glucose specific activity, as described previously (34).

All data are presented as means ± SEM. Two-way ANOVA with or without repeated measures design was used (SigmaStat; Systat, Richmond, CA), and post hoc analysis was performed using the Student-Newman-Keuls multiple comparisons test. Significance was established when P was <0.05.

During transition from the control to experimental period, arterial and hepatic sinusoidal insulin concentrations (µU/mL; CTR−PoG: 22 ± 1 and 68 ± 6, HFFD−PoG: 23 ± 1 and 67 ± 3, CTR+PoG: 23 ± 4 and 78 ± 10, HFFD+PoG: 27 ± 1 and 81 ± 11; P < 0.05 vs. basal period) were increased to similar levels in all four groups, whereas arterial and hepatic sinusoidal glucagon concentrations were kept basal throughout the study (Fig. 2A–D).

During P1, arterial blood glucose concentrations were similar in all four groups (mg/dL; CTR−PoG: 159 ± 3, HFFD−PoG: 161 ± 3, CTR+PoG: 161 ± 4, HFFD+PoG: 166 ± 2; P < 0.05 vs. basal period) in order to double the HGL (mg/kg/min; CTR-PoG: 37 ± 3, HFFD-PoG: 36 ± 3, CTR+PoG: 37 ± 2, HFFD+PoG: 37 ± 4; P < 0.05 vs. basal period) (Fig. 3A and B). During P2, arterial glucose concentrations were clamped at a slightly reduced level in +PoG groups (mg/dL; CTR: 147 ± 3, HFFD: 152 ± 2) to maintain a doubling of the HGL in the presence of PoG infusion (Fig. 3A and B).

HGU and HGP were similar among all four groups during the control period (Fig. 3C and D and Table 1). In response to hyperinsulinemia and hyperglycemia, HGU increased to a similar rate in both CTR groups during P1, reaching a peak of 1.5 ± 0.3 and 1.8 ± 0.3 mg/kg/min in the −PoG and +PoG groups, respectively (P < 0.05 vs. basal period). When coupled with near complete suppression of HGP, a robust switch from net hepatic glucose output (NHGO) to uptake (NHGU) occurred in both CTR groups (Table 1). In the CTR−PoG group, mean rates of HGU and NHGU (1.6 ± 0.1 and 1.5 ± 0.2 mg/kg/min, respectively) during P2 were similar to those in P1 (Fig. 3C and Table 1). In contrast, PoG delivery rapidly augmented HGU in the CTR+PoG group, with a significant increase occurring 15 min after the start of PoG infusion (HGU, mg/kg/min; CTR−PoG: 1.4 ± 0.2 vs. CTR+PoG: 2.2 ± 0.3, P < 0.05), and an eventual peak of 3.0 ± 0.3 mg/kg/min (Fig. 3D). Given that HGP was already suppressed, the significant increase in NHGU seen in response to PoG delivery was accounted for by the increase in HGU (Table 1).

In contrast to chow-fed animals, 4 weeks of HFFD feeding rendered the liver resistant to the stimulatory effects of hyperinsulinemia, hyperglycemia, and PoG delivery on HGU. As a result, mean rates of HGU in HFFD-fed animals during P1 and P2 (mg/kg/min; 0.4 ± 0.1 and 0.6 ± 0.2 in HFFD−PoG and 0.5 ± 0.1 and 0.6 ± 0.2 in HFFD+PoG, respectively) changed minimally from those observed during the control period and were significantly diminished relative to rates observed in the corresponding CTR groups (Fig. 3C and D and Table 1). Furthermore, HGP was incompletely suppressed in both HFFD groups during P1, resulting in sustained NHGO despite the presence of hyperinsulinemia and hyperglycemia (Table 1). In the absence of PoG delivery during P2, there was no further suppression of HGP in the HFFD−PoG group. Thus, mean rates of HGU and HGP remained similar to those in P1, and the liver failed to take up glucose in net sense (Fig. 3C and D and Table 1). In contrast, PoG delivery suppressed HGP further in the HFFD+PoG group (P < 0.05, P2 vs. P1), although it did not augment HGU (Fig. 3D and Table 1). As a result, NHGB fell to a value not significantly different from zero (Table 1).

All groups exhibited net hepatic lactate uptake during the control period (Table 2). Coincident with the increase in HGU during P1, there was a significant increase in arterial blood lactate concentrations in both CTR groups that reflected a switch from net hepatic lactate uptake to output. Net hepatic lactate output waned during P2 in the absence of PoG infusion, whereas in its presence, it was sustained at an elevated rate (Table 2). In contrast, both HFFD groups exhibited net hepatic lactate uptake for the duration of the study, although it was somewhat reduced relative to the control period during both P1 and P2 (Table 2).

During the control period, arterial blood glycerol and plasma NEFA concentrations and their net hepatic balances were similar among the four groups. During P1 and P2, their levels and hepatic uptakes declined in response to hyperinsulinemia, with the steady-state values being slightly lower in the CTR groups (Table 2).

GK mRNA levels were similar between CTR and HFFD groups under basal fasted conditions (Fig. 4A). The combination of hyperinsulinemia and hyperglycemia stimulated six- and sevenfold increases in GK expression in CTR and HFFD groups, respectively (P < 0.05 vs. basal CTR), but this did not manifest as an increase in GK protein or activity (Fig. 4B and C). Strikingly, PoG delivery stimulated a significantly greater increase in GK expression in both diet groups (26- and 24-fold increase in CTR and HFFD groups, respectively; P < 0.05 vs. basal CTR). In addition, PoG infusion was associated with a significant increase in GK protein and enzymatic activity in chow-fed animals (Fig. 4B). In HFFD-fed dogs, in contrast, GK protein levels and activity were markedly reduced (P < 0.05) regardless of the experimental condition. Intriguingly, GKRP levels were also significantly lower in HFFD-fed animals (Fig. 4D).

Under basal fasted conditions, the phosphorylation of GS (Ser641) and the activity ratios of GS and GP were not different between groups. In the CTR−PoG group, hyperinsulinemia and hyperglycemia produced a significant decrease (35%) in GS phosphorylation (Ser641) (Fig. 5A). Accordingly, the activity ratio of GS increased approximately fivefold (P < 0.05 vs. basal), whereas that of GP was reduced by 40% (Fig. 5B and C). Thus, when GS and GP were themselves expressed as a ratio (GS/GP), there was an eightfold increase from basal (P < 0.05) concomitant with an increase in liver glycogen levels (Fig. 5D and E). Although portal glucose delivery did not produce a further decrement in GS phosphorylation (Ser641), it did stimulate a further increase in the GS activity ratio (∼40%; P < 0.05 vs. CTR−PoG) and an additional small decrease in the GP activity ratio (Fig. 5B and C). As a result, the GS/GP activity ratio was elevated 12-fold in the presence of PoG infusion (P < 0.05 vs. CTR−PoG), and liver glycogen levels were increased to a significantly greater extent than in its absence (Fig. 5D and E).

In the HFFD−PoG group, in contrast, hyperinsulinemia and hyperglycemia did not produce a decrease in GS phosphorylation (Ser641). Although the activity ratio of GS increased approximately threefold (P < 0.05 vs. basal), the rise was significantly reduced relative to the increase seen in the corresponding CTR group. The activity ratio of GP changed minimally from basal (Fig. 5B and C). Consequently, the GS/GP activity ratio and the change in liver glycogen were significantly lower in the HFFD−PoG group than in the CTR−PoG group (Fig. 5D and E). Furthermore, PoG delivery produced no further change in GS phosphorylation (Ser641) or in the GS, GP, or GS/GP activity ratios. Although terminal liver glycogen levels were slightly increased in the HFFD+PoG group, the change in glycogen induced by PoG delivery was markedly less in HFFD-fed dogs than in CTR-fed dogs (Fig. 5E). Notably, HFFD feeding was associated with significantly less (∼70%; P < 0.05) incorporation of glucose into glycogen through the direct pathway, in agreement with diminished HGU and GK activity and impaired activation of GS (Fig. 5F).

There was no effect of diet or experimental condition on total liver triglyceride levels (µg/mg liver; basal CTR: 2.5 ± 0.4, basal HFFD: 1.6 ± 0.1, HIHG−PoG CTR: 1.3 ± 0.2, HIHG−PoG HFFD: 1.8 ± 0.4, HIHG+PoG CTR: 1.4 ± 0.2, HIHG+PoG HFFD: 1.1 ± 0.2).

This study revealed that activation of liver GK and GS occurs in response to hepatic PoG delivery in the presence of a physiologic rise in plasma glucose and insulin. Conversely, HFFD feeding was associated with a marked reduction in liver GK content and activity and an impairment in the activation of GS in response to hyperinsulinemia, hyperglycemia, and PoG infusion. Our findings, although not definitively proving causality, thus provide compelling biochemical evidence implicating these changes in the increased postprandial hyperglycemia and diminished hepatic glucose disposal associated with chronic consumption of a diet high in fat and fructose (22,25).

Consistent with previous observations (1,4,7,11), a physiologic rise in glucose and insulin in the absence of PoG delivery resulted in a dynamic switch from NHGO to NHGU. This was due to near complete suppression of HGP concomitant with significant stimulation of HGU. Although a sixfold increase in hepatic GK expression accompanied hyperinsulinemia in the CTR−PoG group, this did not result in an increase in the functional amount of GK protein within 3 h. Previously, Iynedjian et al. (35) demonstrated in rats, and Ramnanan et al. (26) more recently in dogs, that a lag time exists between the transcriptional induction of GK by insulin and subsequent accumulation of GK protein, thus probably explaining this asynchrony. Nevertheless, hyperglycemia and hyperinsulinemia also stimulated the dephosphorylation and activation of GS while triggering the dephosphorylation and inactivation of GP, which culminated in an increase in the deposition of carbon as glycogen within the liver. These data are concordant with previous results from Pagliassotti et al. (7) in dogs and Petersen et al. (36) in humans.

PoG infusion triggered a significantly greater increase in net HGU, as reported previously by our laboratory (4,7,10,11). This was solely attributable to further augmentation of HGU, given that HGP was completely suppressed prior to PoG delivery. Intriguingly, PoG infusion increased GK mRNA 24-fold above basal, which did result in a significant increase in GK protein by 3 h. This occurred despite the fact that the level and duration of hyperinsulinemia, the canonical hormonal inducer of liver GK (35), was equivalent among the groups. Thus, activation of glucose signaling in the portal vein under HIHG conditions triggered a significantly greater increase in GK mRNA such that after 90 min of this signaling, GK protein was increased. This finding is at odds with a previous report by Iynedjian et al. (37) in which hyperglycemia had no effect on the induction of GK expression by insulin in cultured liver cells. Notably, total hepatic denervation ablates the response of the liver to PoG delivery, suggesting that the PoG signal is neurally mediated (38). Thus, it is likely that the induction of liver GK by PoG delivery is also under neural control, such that its effect would only be detected in the intact organism. Recently, Ramnanan et al. (39) demonstrated that a selective rise in brain insulin levels induced a threefold increase in hepatic GK mRNA that was ablated with intracerebroventricular infusion of a phosphatidylinositol-3 kinase inhibitor. These data raise the possibility that efferent signals sent to the liver might contribute to the transcriptional regulation of hepatic GK in response to a meal, although the precise nature of the neural or humoral factor mediating this response is currently unknown. In addition, our findings do not shed light on whether PoG delivery stimulated the translocation of GK in vivo, as we have suggested previously (8,40), but they clearly demonstrate that GK expression and activity are both augmented by PoG delivery. Likewise, PoG infusion triggered a significantly greater increase in GS activity, although this was not associated with further diminution of GS phosphorylation on Ser641. Given that GS contains several other phosphorylation sites that regulate its activity (41,42), the effect of the PoG signal on the phosphorylation state of GS could have been mediated by dephosphorylation of sites distinct from that of Ser641. Nevertheless, these data suggest that PoG signaling produces a significantly greater increase in HGU and GSYN by augmenting the activity of GK and GS under HIHG conditions.

In stark contrast to the response observed in chow-fed dogs, HFFD-fed dogs exhibited impaired suppression of HGP and virtually no stimulation of HGU in response to hyperinsulinemia and hyperglycemia. Despite a sevenfold increase in GK mRNA levels in the HFFD−PoG group, GK protein content and activity were significantly reduced relative to those in chow-fed dogs. In agreement with these findings, livers of dogs in the HFFD−PoG group continued net hepatic lactate uptake for the duration of the study. Because phosphorylation of glucose by GK is thought to be rate-limiting for HGU, this is likely to explain the impaired HGU caused by the HFFD (43,44). Although the possibility cannot be excluded that a substrate-mediated increase in GK flux and hepatic glucose cycling occurred in the HFFD−PoG group, it did not translate into an increase in committed glucose uptake and carbon retention by the liver. In addition, the phosphorylation of GS (Ser641) was not reduced from basal, and the activity ratio of GS in the presence of hyperinsulinemia and hyperglycemia was significantly lower in HFFD-fed dogs than in the corresponding CTR group. Consequently, the change in liver glycogen levels was minimal. Although our data do not prove causality, they clearly suggest that impaired regulation of hepatic GK, GS, and GP explains the deficit in HGU and storage following consumption of a HFFD.

PoG delivery in the presence of hyperinsulinemia and hyperglycemia triggered suppression of residual HGP in the HFFD+PoG group, but did not stimulate HGU. As a result, NHGB decreased to a rate not significantly different from zero. Likewise, PoG delivery had no impact on net hepatic lactate balance, on the GS or GP activity ratios, or on the incorporation of glucose into glycogen through the direct pathway, consistent with lack of stimulation of HGU and GSYN in the HFFD+PoG group. Although PoG infusion induced the expression of hepatic GK in the HFFD+PoG group, the increase in GK protein content and activity seen in response to PoG delivery in chow-fed animals was abolished in HFFD-fed animals, as their GK protein levels were decreased by >50%. Likewise, the ability of PoG delivery to activate GS was abolished in the HFFD group. Thus, 4 weeks of HFFD feeding abrogated the stimulatory effects of hyperglycemia, hyperinsulinemia, and PoG delivery on hepatic GK and GS activity. Collectively, these data are in agreement with those of Basu et al. (20), who demonstrated that impaired SGU in type 2 diabetic individuals is not dependent on the route of glucose delivery. Furthermore, Basu et al. (20,21) provided convincing metabolic data to suggest that a defect in hepatic GK was linked to impaired SGU in subjects with type 2 diabetes, given that the contribution of extracellular glucose to the uridine diphosphate glucose pool was markedly reduced in type 2 diabetic subjects despite the presence of hyperinsulinemia and hyperglycemia. Although our diet-induced model of glucose intolerance is not directly comparable to individuals with type 2 diabetes, our GK findings provide new cellular evidence in support of the claim that defective regulation of liver GK is associated with postprandial hyperglycemia and diminished hepatic glucose disposition.

Normal GK mRNA expression concomitant with diminished GK protein content suggests that the decrease in GK protein occurred posttranscriptionally. Interestingly, liver GKRP protein levels were significantly reduced in our HFFD-fed dog model as well. Studies conducted in GKRP-deficient mice (45,46) demonstrated that they have decreased GK protein despite normal basal and insulin-stimulated GK mRNA, suggesting that GKRP exerts a permissive effect on the level of GK protein through a posttranscriptional mechanism. Although we cannot directly compare our findings in the HFFD-fed dog to those of a genetically modified mouse model, it is tempting to speculate that reduced levels of GKRP in HFFD-fed animals might have contributed to the posttranscriptional decline in GK protein.

Altogether, our findings reveal new insight into the molecular physiology of the portal glucose signaling mechanism under normal conditions and the pathophysiology of aberrant postprandial hepatic glucose disposition evident in response to diet-induced glucose intolerance. These data support the possibility that nutritional modulation of hepatic GK might be a common pathogenic factor underlying the manifestation of disturbances in postprandial hepatic glucose flux and GSYN.

This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grants R01-DK-18243 (to A.D.C.) and R37-DK-27221 (to P.J.R.). Hormone and liver triglyceride analysis was performed by the Hormone Assay and Analytical Services Core, Vanderbilt Diabetes Research and Training Center, supported by NIDDK Grant DK-20593. A.D.C. is the recipient of the Jacquelyn A. Turner and Dr. Dorothy J. Turner Chair in Diabetes Research.

No potential conflicts of interest relevant to this article were reported.

K.C.C. assisted in experimental design, collected and analyzed all data, and wrote the article. G.K. assisted with all data collection. J.M.I. and P.J.R. performed GS and phosphorylase activity assays and assisted with data analysis. M.S.S. assisted with the GK activity and glycogen assays, as well as the collection of Western and mRNA data. B.F. and D.W.N. performed the surgeries and assisted with data collection. M.S. assisted with the GK activity assay and Western blot. K.C.C., G.K., J.M.I., M.S.S., B.F., D.W.N., P.J.R., M.S., and A.D.C. provided input during the writing of this manuscript. A.D.C. is the guarantor of this work and was involved in all intellectual and financial aspects of the study.

Parts of this study were presented at the 70th and 72nd Scientific Sessions of the American Diabetes Association, Orlando, Florida, 25–29 June 2010, and Philadelphia, Pennsylvania, 8–12 June 2012, respectively.

The authors thank Mary C. Moore, Margaret Lautz, Jon Hastings, Josh Roop, Wanda Snead, Patsy Raymer (Vanderbilt University), and the Vanderbilt Diabetes Research and Training Center Hormone Core for technical support in these studies, as well as Phil Williams (Department of Surgery, Vanderbilt University) for technical and surgical assistance.