Previous studies in mice suggest that portal venous infusion of glucose at a low rate paradoxically causes hypoglycemia; this does not occur in dogs, rats, and humans. A possible explanation is that fasting status in the mouse studies may have altered the response. We sought to determine whether the response to portal glucose delivery in the mouse was similar to that seen in other species and whether it was dependent on fasting status. Studies were performed on chronically catheterized conscious mice. Catheters were placed into the portal and jugular veins and carotid artery 5 days before study. After a 5- or 16-h fast, glucose was infused into either the portal (PO) or the jugular vein (JU) for 6 h at 25 μg·g−1·min−1. [3-3H]glucose was infused into the JU to measure glucose turnover. In 5-h-fasted mice, PO and JU exhibited similar increases in arterial blood glucose from 155 ± 11 to 173 ± 19 and 147 ± 8 to 173 ± 10 mg/dl, respectively. Endogenous glucose production decreased and arterial insulin increased to the same extent in both PO and JU. A similar response was observed in 16-h-fasted mice; however, the proportion of hepatic glycogen synthesis occurring by the indirect pathway was increased by fasting. In summary, portal glucose delivery in the mouse did not cause hypoglycemia even when the duration of the fast was extended. The explanation of the differing response from previous reports in the mouse is unclear.
- portal vein
the liver plays a major role in blood glucose homeostasis by being able to rapidly transition from a site of glucose production in the fasting state to a site of glucose utilization in the fed state. The magnitude of hepatic glucose utilization during feeding is dependent on both the plasma insulin and glucose levels and the route of glucose delivery. In fact, significant hepatic glucose utilization is not observed if only hyperglycemia and hyperinsulinemia are present, but an intraportal or intraduodenal glucose infusion prompts substantial hepatic glucose uptake (24). Portal vein glucose delivery is known to produce a signal (“portal signal”), which enhances net hepatic glucose uptake and hepatic glycogen deposition in humans (9, 29) and rats (7, 12, 27). The portal signal also impairs peripheral glucose utilization in dogs and rats (20), thus augmenting the importance of the liver in whole body glucose disposal following a meal.
A series of studies in the mouse suggested that delivery of glucose into the portal vein at a rate of 25 μg·g−1·min−1, which is similar to the basal endogenous glucose production (EGP) paradoxically induced hypoglycemia (5, 6). Hypoglycemia was precipitated by an activation of peripheral glucose clearance. This response was inhibited by somatostatin infusion and dependent upon GLP-1 receptor activation (4). Prior work observed an inhibitory or no effect of the portal signal on glucose clearance using somatostatin with insulin and glucose replacement to allow careful matching of the insulin and insulin concentrations (7, 20, 28). However, the use of somatostatin may have blocked an augmentation of the peripheral glucose uptake seen with portal glucose delivery. Subsequent studies were done in other species where somatostatin was not infused. In nondiabetic humans, intraduodenal glucose infusion at rates bracketing normal EGP modestly increased plasma glucose and insulin concentrations (29). In conscious 42-h-fasted dogs, peripheral glucose infusion at a rate near the basal EGP caused a rapid enhancement of nonhepatic glucose uptake, whereas portal delivery at the same rate quickly activated net hepatic glucose uptake (18). Another study did not observe hypoglycemia in 12- to 16-h-fasted dogs (16). All groups maintained near euglycemia. A continuous 24-h infusion of glucose into the hepatic portal vein or a peripheral vein at three different rates (4, 8.5, and 14 μg·g−1·min−1) increased arterial glucose and insulin concentrations in the 24-h-fasted conscious, unrestrained rat (23). Thus the later studies suggest that in the absence of somatostatin infusion, portal glucose delivery in other species does not cause hypoglycemia.
Two differences between the studies in mice and those in other species are the route of blood sampling and the duration of the fast. In the mouse, blood samples were obtained from a tail vein. In other species, samples were obtained from an artery by use of a previously implanted arterial catheter (18, 23) or from a heated hand vein to obtain arterialized venous blood (29). Although the tail vein sampling method will raise the baseline glucose concentration in the mouse, it will not alter insulin action as assessed with the euglycemic hyperinsulinemic clamp (1). The fasting state also differed. The studies in the dog, rat, and human used fasting states that did not cause marked hepatic glycogen depletion (15, 19, 25). In contrast, studies in the mouse were done after a 6-h morning fast, but the animals were on a reverse light-dark cycle. Because mice are primarily nocturnal eaters, the equivalent duration of the fast was much longer than if they had been on a normal light-dark cycle (6)
We sought to determine whether the response to portal glucose delivery in the conscious mouse was similar to that seen in other species and whether this was influenced by the duration of the fast. We extended the chronically catheterized conscious mouse model (13, 22) in which catheters were placed into the carotid artery and jugular vein to include catheterization of the portal vein. With simultaneous access to both the jugular and portal veins for glucose infusion and the ability to take blood samples from the carotid artery, we could determine in the conscious mouse whether the mouse has a unique response to portal glucose delivery and whether the response is dependent on the fasting state.
RESEARCH DESIGN AND METHODS
Animals and surgical procedure.
Adult (8-wk-old) male C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and were studied between 12 and 16 wk of age. Mice were fed standard chow ad libitum in an environmentally controlled room with a 12:12-h light-dark cycle (lights on 0600–1800) that met Association for Assessment and Accreditation of Laboratory Animal Care International guidelines. The surgical procedures utilized were similar to those described previously (13, 22). Mice (28.6 ± 0.4 g) were anesthetized with sodium pentobarbital (70 mg/kg body wt). The left common carotid artery and portal vein were catheterized with a two-part catheter consisting of a small piece of PE-10 (inserted into the artery or portal vein) and attached Silastic (0.025 in. OD) tubing. The right jugular vein was catheterized for infusions with a Silastic catheter (0.025 in. OD). After a midline laparotomy from the xyphoid process to the pubic bone, the portal vein was punctured caudal to the entry of the splenic vein and anchored using a cellulose patch kit (Data Sciences, St. Paul, MN). The abdominal incison was closed in two layers. The free ends of catheters were tunneled under the skin to the back of the neck where they were attached via stainless steel connectors to lines made of Micro-Renathane (0.033 in. OD), which were exteriorized and sealed with stainless steel plugs. Lines were flushed daily with 10–50 μl of saline containing 200 U/ml heparin and 5 mg/ml ampicillin (Sigma, St Louis, MO). Animals were individually housed after surgery, and body weight was recorded daily. The success rate of the surgery and catheterization procedures was ∼70%. Mice were allowed to recover from surgery for ≥5 days and were studied only when body weight was restored to within 10% of presurgery body weight. All procedures performed were approved by the Vanderbilt University Animal Care and Use Committee.
Four groups of animals were studied. The animals were fasted either for 5 h (0700 to 1200) or overnight (16 h, 2000 to 1200) prior to the study. After the fast, glucose (t = 0, 1200) was infused into either the portal (PO) or the jugular vein (JU) for 6 h at 25 μg·g−1·min−1. Each study consisted of a 2-h tracer equilibration period (−120 to 0 min) and a 6-h experimental period (0 to 360 min). [3-3H]glucose (New England Nuclear, Wilmington, DE) was given at 0.2 μCi/min through the JU (−120 to 360 min) to measure glucose turnover. [U-14C]glucose (New England Nuclear) was given at 0.1 μCi/min through the PO (0 to 180 min), and then it was switched to the JU (180 to 360 min). This allowed us to calculate the first-pass hepatic glucose extraction (data will not be presented). Prior to the infusion of glucose, an arterial blood sample was taken (t = 0). After initiation of the glucose infusion, arterial blood samples were taken every 30 min for the duration of the study. Blood glucose levels were monitored using a HemoCue glucose analyzer (HemoCue, Mission Viejo, CA). During the experimental period, a continuous infusion of donor blood was used to replace blood removed. After the last blood sample had been, mice were anesthetized with an infusion of pentobarbital sodium. The liver was excised, freeze-clamped immediately in liquid nitrogen, and stored at −70°C for future tissue analysis.
Blood samples were centrifuged at 13,000 rpm for 2 min. For the glucagon assay, 25 μl of plasma were added to 2 μl of Trasylol (500 kallikrein inhibitor units; Miles, Kankakee, IL). Plasma samples (10 μl) were deproteinized with Ba(OH)2 and ZnSO4, exposed to anion and cation resin to remove charged intermediates (17), dried, and counted (LS 3801; Beckman, Fullerton, CA) to assess plasma [3H]- and [14C]glucose specific activity (SA).
Immunoreactive insulin and glucagon were assayed with a double-antibody method (21). For glucokinase (GK), hexokinase (HK), and glucose-6-phosphatase (G-6-Pase) activity measurements, freeze-clamped liver was homogenized in 50 mmol/l HEPES, 100 mmol/l KCl, 1 mmol/l EDTA, 5 mM MgCl2, and 2.5 mmol/l dithioerythritol (2). The homogenate was centrifuged (100,000 g) at 4°C for 45 min, and GK and G-6-Pase activities were measured in the supernatant and sedimentary fractions, respectively. GK activity was calculated as the difference between activities at 100 and 0.5 mM glucose. HK was calculated as the activity in the presence of 0.5 mM glucose. G-6-Pase was measured at glucose 6-phosphate (G-6-P) concentrations between 0 and 10 mM. Protein content was assessed with the Biuret method.
Glucose turnover rate (GTR) was calculated as the ratio of the rate of infusion of [3-3H]glucose (dpm/min) and the steady-state plasma [3H]glucose SA (dpm/μmol) according to the method of De Bodo et al. (8). Fasting decreases body weight, thus the glucose infusion rate and glucose turnover data were calculated based upon the body weight the evening prior to fasting.
Net deposition of glycogen in liver was calculated using both [3H]- and [14C]glucose by dividing hepatic tracer glycogen accumulation (dpm/g liver) by the average inflowing tracer glucose SA. In the case of portal glucose delivery inflowing glucose SA will be less than the arterial SA. This was corrected by multiplying the arterial glucose SA by the ratio of (Pa·LBF) and 25 + (Pa·LBF) where Pa is the average arterial glucose concentration (mg/ml), LBF is liver blood flow [100 ml·kg−1·min−1 (26)], and 25 is the glucose infusion rate in the portal vein (mg·kg−1·min−1). Indirect glycogen synthesis was calculated as the difference between 14C determined glycogen synthesis and 3H determined glycogen synthesis.
Data are expressed as means ± SE. The significance of differences between groups of time course data was analyzed by two-way repeated-measures analysis of variance (ANOVA) and post hoc analysis with F-test (SYSTAT, Evanston, IL). The significance of differences between groups was analyzed by one-way ANOVA or Student's t-test. Differences were considered significant when P < 0.05.
The evening before the study, the body weights in the 5-h-PO, 5-h-JU, 16-h-PO, and 16-h-JU groups (28 ± 1, 28 ± 1, 26 ± 1, and 26 ± 1 g, respectively) were similar and were within 10% of their body weight prior to surgical implantation of the catheters (29 ± 1, 29 ± 1, 28 ± 1, and 28 ± 1 g). The body weight of overnight-fasted (16-h) mice decreased ∼12% (23 ± 1 g in 16-h-PO and 23 ± 1 g in 16-h-JU) overnight.
During the 360-min glucose infusion (25 μg·g−1·min−1), arterial blood glucose levels of 5-h-fasted mice increased within 30 min (155 ± 11 to 258 ± 32 mg/dl in 5-h-PO and 147 ± 8 to 228 ± 24 mg/dl in 5-h-JU) and then slowly decreased to 198 ± 11 mg/dl in 5-h-PO and 199 ± 13 mg/dl in 5-h-JU after 180 min of the infusion period (t = 180 min; Fig. 1A). The elevated arterial blood glucose levels were maintained until the end of the infusion period (173 ± 19 mg/dl in 5-h-PO and 173 ± 10 mg/dl in 5-h-JU; 180 to 360 min). The blood glucose levels were lower after a 16-h fast and increased to a lesser extent during the first 30 min (125 ± 2 to 192 ± 19 mg/dl in 16-h-PO and 113 ± 5 to 187 ± 17 mg/dl in 16-h-JU groups mice) but steady-state levels were similar to those in the 5-h-fasted group (Fig. 1B). The glucose response was not dependent on the route of glucose infusion in either the 5-h- or 16-h-fasted mice.
The plasma insulin concentrations (Fig. 1C) in 5-h-PO and -JU increased (from 0.7 ± 0.1 to 1.6 ± 0.1 ng/ml and 0.6 ± 0.1 to 1.7 ± 0.3 ng/ml, respectively; P < 0.05 vs. basal period), over the first 120 to 180 min of glucose infusion. They then gradually decreased toward the plasma insulin concentrations observed prior to glucose infusion (0.7 ± 0.1 ng/ml in 5-h-PO and 1.2 ± 0.2 ng/ml in 5-h-JU; 180- to 360-min interval). In addition, no differences in plasma insulin concentrations were apparent between the two groups at any time. In 16-h-fasted mice (Fig. 1D), arterial plasma insulin concentrations changed minimally in PO and increased from 0.5 ± 0.1 to 1.0 ± 0.2 ng/ml in JU at 120 min and remained stable until the end of the glucose infusion period. There was a significant difference between 16-h-PO and 16-h-JU at 120 min (P < 0.05).
Arterial plasma glucagon concentrations were measured at the end of the glucose infusion period (t = 360 min). The arterial glucagon concentrations were very low (≤20 pg/ml) in both 5-h-PO and 5-h-JU. In 16-h-PO and 16-h-JU, glucagon concentrations were also very low: 22 ± 2 and 25 ± 2 pg/ml, respectively.
The glucose infusion increased the tracer-determined glucose appearance rate (Ra; 18.3 ± 2.6 to 34.1 ± 3.0 μg·g−1·min−1 in 5-h-PO, 18.1 ± 1.6 to 32.8 ± 1.4 μg·g−1·min−1 in 5-h-JU; 17.8 ± 2.0 to 38.3 to 4.1 μg·g−1·min−1 in 16-h-PO, 16.23 ± 1.6 to 31.0 ± 2.1 μg·g−1·min−1 in 16-h-JU at 120 min) and glucose clearance (12.2 ± 2.1 to 16.5 ± 1.3 μl·g−1·min−1 in 5-h-PO, 12.5 ± 1.3 to 16.7 ± 2.3 μl·g−1·min−1 in 5-h-JU; 14.3 ± 2.6 to 21.0 ± 2.7 μl·g−1·min−1 in 16-h-PO and 14.8 ± 2.1 to 17.8 ± 2.2 μl·g−1·min−1 in 16-h-JU at 120 min; Fig. 2, A–D). The endogenous glucose Ra (Endo Ra) decreased in the four groups during glucose infusion (Fig. 2, E and F). Glucose Ra, clearance, and Endo Ra were not different among the groups at any time point, but the areas under the curve (AUC) of Endo Ra were significantly different between 16-h-PO and 16-h-JU (5,891 ± 967 and 3,452 ± 670 μg/g; P < 0.05).
The rates of liver glycogen synthesis via the direct and indirect pathways during the glucose infusion were not different between 5-h-PO (3.7 ± 2.3 and 3.0 ± 2.8 μg glucose·g liver−1·min−1, respectively) and 5-h-JU (2.8 ± 1.0 and 1.9 ± 1.0 μg glucose·g liver−1·min−1, respectively) or between 16-h-PO (4.5 ± 1.5 and 5.3 ± 2.0 μg glucose·g liver−1·min−1, respectively) and 16-h-JU (5.2 ± 2.9 and 7.2 ± 4.5 μg glucose·g liver−1·min−1, respectively). However, the percentage of glycogen synthesis via the direct pathway was higher in the 5-h-fasted groups (Table 1).
The activities of HK and GK were 2.9 ± 0.3 and 5.9 ± 1.4 mU/mg protein in 5-h-PO, 3.1 ± 0.1 and 9.0 ± 1.0 mU/mg protein in 5-h-JU, 4.5 ± 0.6 and 7.0 ± 0.7 mU/mg protein in 16-h-PO, and 4.1 ± 0.8 and 5.4 ± 0.8 mU/mg protein in 16-h-JU, respectively. There was a significant difference between the GK activity of 5- and 16-h-JU.
Activity of G-6-Pase was evaluated at 10 mM G-6-P. The calculated Km and Vmax of G-6-Pase were both increased in 16-h-fasted mice (18 ± 2 mM and 1,472 ± 163 mU/mg protein in PO, 19 ± 2 mM and 1,102 ± 149 mU/mg protein in JU, respectively) compared with 5-h-fasted mice (6 ± 1 mM and 419 ± 50 mU/mg protein in PO, 8 ± 1 mM and 488 ± 95 mU/mg protein in JU, respectively).
In previous reports, mice developed hypoglycemia in response to portal glucose delivery. This is not the case for rats, dogs, and humans. We hypothesized that the difference was not due to species differences but to differences in experimental procedures and fasting duration. In the present study, we observed that portal glucose infusion in the chronically catheterized, conscious mouse does not cause hypoglycemia and that the duration of fasting does not alter the overall response.
The metabolic response to intraportal glucose delivery in mice was reported by Burcelin et. al. (6). Infusion of glucose in the portal vein of mice at a rate of 25 μg·g−1·min−1 markedly increased peripheral glucose clearance (6) and caused a paradoxical fall in the plasma glucose concentration to hypoglycemic levels (5, 6). We had a similar result to that of Burcelin et. al. (6) both peripheral and portal glucose delivery (25 μg·g−1·min−1) enhanced insulin release and glucose clearance and decreased Endo Ra. However, we observed that the rise in glucose clearance with portal glucose delivery was similar to that seen with peripheral glucose delivery; thus hypoglycemia did not develop. The present results are very similar to the previous reports in other species: portal glucose delivery does not cause paradoxical hypoglycemia (18, 23, 29).
One difference between the previous reports in mice and those in the other species is that, in mice, blood samples were obtained from a tail vein rather than from an artery. In addition, in the mice studies a segment of the left sciatic nerve was removed during the catheter implantation. The latter will alter the left leg motor activity but has been reported not to change feeding behavior (6). We considered that both the left sciatic nerve removal and tail cutting (6) likely altered the baseline stress and, when combined with the reverse light-dark cycle, may have altered the metabolic response of the mice. Interestingly, baseline glucose levels were rather low in the prior studies [∼90 mg/dl (6)] considering the duration of the reported fast (6 h). It is unlikely that tail vein sampling per se can explain the fasting hypoglycemia. Our recent data suggest that tail vein sampling increases fasting glucose concentration compared with arterial sampling (1). Yet, despite this, insulin action is not altered, as glucose requirements during a euglycemic hyperinsulinemic clamp are comparable when blood is sampled from a tail vein or a previously implanted arterial catheter (1). Because the response to peripheral glucose infusion in the studies of Burcelin et al. (6) is similar to the present results where blood was obtained from an arterial catheter, it is unlikely that the differing sampling site altered the overall metabolic response to glucose infusion.
Increasing the duration of fasting did not alter the route-dependent response to glucose infusion; however, it may have augmented insulin action. In response to peripheral glucose infusion, arterial glucose and insulin concentrations were ∼10 and 70% lower after the 16-h fast than with the 5-h fast. Because whole body glucose disposal was similar, it suggests that insulin responsiveness was enhanced. This is consistent with recent studies in the mouse in which overnight fasting augmented insulin action as assessed by euglycemic hyperinsulinemic clamps (1, 14). In contrast, extended fasting (60 and 96 h) induces insulin resistance in humans (3, 11).
We did not detect an enhancement of pancreatic insulin secretion in response to portal glucose delivery. Prior work suggested that portal glucose delivery can amplify pancreatic insulin secretion in the dog (10). This is not observed when low doses of glucose are used in the dog and rat (16, 18, 23). The studies of Burcelin et. al. (6) demonstrated in the mouse that insulin secretion persisted during low-dose portal glucose infusion, which contributed to the observed hypoglycemia. In our studies, the insulin response was not amplified and tended to be lower in the 16-h-fasted group when modest glucose infusion rates were used. An explanation for the differing pancreatic insulin response is unclear. However, it is possible that inappropriate insulin secretion during portal glucose delivery is required to allow the peripheral effect of portal glucose delivery to be manifested.
Prior fasting amplified hepatic glycogen synthesis primarily via activation of the indirect pathway, and glycogen synthesis was not altered by the route of glucose delivery. Glycogen in liver is synthesized from G-6-P derived either by the direct phosphorylation of glucose (direct pathway) or via gluconeogenesis from 3-carbon intermediates (indirect pathway). We found that the percentage of glycogen synthesis via the indirect pathway was lower in 5-h-fasted groups than in overnight-fasted ones. Thus prolonging the duration of a fast (overnight) increases the percentage of glycogen synthesis from the indirect pathway irrespective of the route of glucose delivery in mice. The rate of and the relative contribution of the indirect pathway to glycogen synthesis will be underestimated to the extent that there is significant hepatic futile cycling. Futile cycling may have been even higher in the 16-h-fasted group, as G-6-Pase activity was higher. Because prolonged fasting augmented indirect glycogen synthesis and possibly had a higher rate of futile cycling during glucose infusion, the fasting-induced increase in indirect glycogen synthesis was likely an underestimate. Consistent with the lack of a change in the direct pathway, total GK was not altered. Although portal glucose delivery is expected to specifically amplify total glycogen synthesis and the relative contribution of the direct pathway, this did not occur in this study. This is consistent with studies in the rat where duodenal glucose delivery augmented glycogen synthesis but not the relative contribution of the direct pathway (27). In the dog, direct glycogen synthesis is augmented by portal glucose delivery (24). The lack of a detectable response to portal glucose delivery cannot be explained by an inadequate arterial portal glucose gradient. The predicted arterial portal vein glucose gradient is 40 mg/dl, which is more than adequate in other species to activate the portal signal (24). In the rat, hepatic glycogen deposition is dependent on the route of glucose delivery in an acute (2 h) setting when the glucose (∼220 mg/dl) and insulin (∼25 ng/ml) concentrations are elevated and glucose requirements (36 mg·kg−1·min−1; ∼6× basal glucose turnover) are high (23, 27). Although lower infusion rates of glucose (14 mg·kg−1·min−1) can augment glycogen synthesis in a route-dependent manner, the effect is not seen until after 8 h (23). Thus it is possible that the relatively low glucose infusion rates in mice for only 6 h may have not been high enough to detect a route-dependent effect on net glycogen synthesis. Consistent with that, the tracer-determined rates of liver glycogen synthesis are substantially lower than that seen in the rat when higher glucose and insulin concentrations were present (27).
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R21-DK-064877 (PI O. P. McGuinness), Mouse Metabolic Phenotyping Center Grant U24-DK-59637; Clinical Nutrition Research Unit Grant P30-DK-26657, Digestive Disease Research Center Grant P30-DK-058404, and Vanderbilt DRTC Grant DK-20593.
We are grateful for the expert technical assistance of Wanda Snead and Angie Penaloza in the Vanderbilt Diabetes Research and Training Center (DRTC) hormone core laboratory.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by American Physiological Society