Unlike Mice, Dogs Exhibit Effective Glucoregulation During Low-Dose Portal and. Peripheral Glucose Infusion

Transcription

1 Articles in PresS. Am J Physiol Endocrinol Metab (September 30, 2003) /ajpendo Unlike Mice, Dogs Exhibit Effective Glucoregulation During Low-Dose Portal and Peripheral Glucose Infusion Mary Courtney Moore, PhD, Sylvain Cardin, PhD,* Dale S. Edgerton, PhD, Ben Farmer, Doss W. Neal, Margaret Lautz, and Alan D. Cherrington, PhD Department of Molecular Physiology & Biophysics and Diabetes Research and Training Center, Vanderbilt University School of Medicine, Nashville, TN *Current affiliation: Wellstat Therapeutics Corporation, Gaithersburg, MD Running title: Peripheral and portal glucose infusion at the rate of EGP Correspondence: Mary Courtney Moore, PhD 702 Light Hall Vanderbilt University School of Medicine Nashville, TN Telephone: FAX: genie.moore@vanderbilt.edu Copyright (c) 2003 by the American Physiological Society.

2 Abstract Portal infusion of glucose in the mouse at a rate equivalent to basal endogenous glucose production causes hypoglycemia, while peripheral infusion at the same rate causes significant hyperglycemia. We used tracer and arteriovenous difference techniques in conscious 42-h-fasted dogs to determine their response to the same treatments. The studies consisted of 3 periods: equilibration (100 min), basal (40 min), and experimental (180 min), during which glucose was infused at 13.7 µmol. kg -1. min -1 into a peripheral (PE, n=5) or the hepatic portal (PO, n=5) vein. Arterial blood glucose increased ~0.8 mmol/l in both groups. Arterial and hepatic sinusoidal insulin concentrations were not significantly different between groups. PE exhibited an increase in nonhepatic glucose uptake (non-hgu; 8.6±1.2 µmol. kg -1. min -1 ) within 30 min, whereas PO showed a slight suppression (3.7±3.1 µmol. kg -1. min -1 ). PO shifted from net hepatic glucose output (NHGO) to uptake (NHGU; 2.5±2.8 µmol. kg -1. min -1 ) within 30 min, but PE still exhibited NHGO (6.0±1.9 µmol. kg -1. min -1 ) at that time and did not initiate NHGU until after 90 min. Glucose R a and R d did not differ between groups. The response to the two infusion routes was markedly different. Peripheral infusion caused a rapid enhancement of non-hgu, while portal delivery quickly activated NHGU. As a result, both groups maintained near-euglycemia. The dog glucoregulates more rigorously than the mouse in response to both portal and peripheral glucose delivery. Key words: hypoglycemia, hyperglycemia, portal vein, glucosensor, glucoregulation

3 3 Introduction Delivery of glucose into the hepatic portal vein, as opposed to a peripheral vein, of the conscious dog enhances net hepatic glucose uptake (NHGU), even when the hepatic glucose load is kept equivalent in the two circumstances and somatostatin is infused to allow insulin and glucagon to be fixed at the same concentrations with the two routes of delivery (1, 34, 38). The enhancement of NHGU is accompanied by a concomitant reduction in nonhepatic (primarily skeletal muscle) glucose uptake (1, 18, 30). This coordinated outcome, which occurs in response to the portal signal, not only operates in the presence of a somatostatin-controlled pancreatic clamp but also in the absence of somatostatin (2, 30, 37). The portal signal apparently functions in non-canine species, including rats (8) and probably humans (14, 54). However, the portal signal has generally been studied during infusion of glucose at rates that would induce hyperglycemia, since it is considered a response to feeding. Recently it has been reported that portal glucose infusion at a rate equivalent to the basal endogenous glucose production (EGP) caused hypoglycemia in 6-h-fasted conscious mice (7). In contrast, infusion of glucose into the femoral vein at the same rate resulted in significant hyperglycemia (7). Hypoglycemia was averted when somatostatin was infused simultaneously into the portal vein, and it did not occur in GLUT-2 knockout mice (6), suggesting that a hepatoportal glucose sensor, requiring GLUT-2 for its action, was inhibited by somatostatin (6, 7). These findings led us to conduct analogous studies, i.e., infusion of glucose via the portal or peripheral circulation at a rate designed to mimic EGP in the dog in order to determine whether the findings in the mouse are applicable to other species.

4 4 The dog exhibits very precise regulation of glycemia, in part by co-ordinating glucose uptake in muscle and liver. The reduction of the proportion of the glucose extracted by nonhepatic tissues concomitantly with the enhancement of NHGU by the portal signal is an example of this co-ordination. Similarly, during peripheral glucose infusion there is marked enhancement of nonhepatic glucose uptake with very little NHGU unless sufficient glucose is infused to create unphysiologic (>12 mm) glycemic levels (reviewed in ref. 29). Because of this precise reciprocity of hepatic and nonhepatic glucose uptake, arterial blood glucose concentrations were virtually identical with the two delivery routes when glucose was infused into conscious dogs at 55.5 µmol. kg -1. min -1 via the hepatic portal or a peripheral vein. Therefore we hypothesized that glycemia would be indistinguishable in dogs receiving peripheral or intraportal glucose infusions at a rate equivalent to resting glucose production. RESEARCH DESIGN AND METHODS Animals and surgical procedures. Studies were carried out on 10 conscious 42-h-fasted mongrel dogs of either sex with a mean weight of 22 ± 1 kg. Diet, housing, and protocol approval have been previously described (38). The 42-h-fast was chosen primarily because it represents the time at which hepatic glycogen concentrations in the dog (and human) reach a stable minimum, and the liver exhibits net lactate uptake rather than release. Both the 6-h-fasted mouse and the 42-h-fasted dog are thus in the postabsorptive state, but hepatic glycogen concentrations are virtually depleted in the 6-h-fasted mouse (5), while the dog, like the human, maintains a minimal but stable glycogen reserve through prolonged fasting (19, 45). To ensure that the period of fasting did not alter our findings, an additional dog was studied after an 18-h fast, the shortest possible period of

5 5 fasting, since meal absorption continues for h (12). The glycemic and insulin responses of that animal, as well as the rate of glucose disposal, were indistinguishable from those in the 42-h-fasted animals, but its data are not included in the results because it exhibited net hepatic lactate release during the basal period. Approximately 16 d before study, each dog underwent a laparotomy, with insertion of sampling catheters in a hepatic vein, the hepatic portal vein, and a femoral artery and infusion catheters into a splenic and a jejunal vein (38, 39). Ultrasonic flow probes (Transonic Systems, Ithaca, NY) were placed around the portal vein and the hepatic artery. Dogs were used for study only if they met established criteria (38). On the morning of the study, catheters and flow probe leads were exteriorized from their subcutaneous pockets and catheter contents were aspirated. The splenic and jejunal catheters were used for intraportal infusion of glucose (when applicable). Angiocaths (Deseret Medical, Sandy, UT) were inserted into two peripheral veins. Experimental design. Each experiment consisted of a 100-min equilibration period (- 140 to 40 min), a 40-min basal period (-40 to 0 min), and a 180-min experimental period (0 to 180 min). At 140 min, a primed (36 µci), constant infusion of [3-3 H]glucose (0.35 µci/min) was initiated via peripheral vein. A constant infusion of cold glucose at 13.7 µmol. kg -1. min -1 was started at 0 min and continued until 180 min. This was equivalent to the rate of tracer-determined EGP in the basal state in 42-h-fasted dogs previously studied in our laboratory (21, 38). Five of the dogs (PO group) received the glucose intraportally, and five (PE group) received it via a peripheral vein. Femoral artery, portal vein, and hepatic vein blood samples were taken every 20 min during the basal period and every 30 min during the experimental period as previously described (39).

6 6 Processing and analysis of samples. Hematocrit; blood lactate, alanine, and glycerol; and plasma glucose, nonesterified fatty acid (NEFA), insulin, and glucagon concentrations and 3 H glucose were determined as described previously (21, 38, 48). Calculations and data analysis. The rate of substrate delivery to the liver, or hepatic substrate load, was calculated by a direct (D) method as: Load in (D) = ([S] A AF) + ([S] P PF), where [S] is the substrate concentration, A and P refer to artery and portal vein, respectively, and AF and PF refer to blood or plasma flow (as appropriate) through the hepatic artery and portal vein, respectively. Hepatic glucose load was also calculated by an indirect (I) method: Load in (I) = (G A HF) + GIR Po GUG, where G A is the arterial blood glucose concentration, HF refers to total hepatic blood flow, GIR Po is the intraportal glucose infusion rate, and GUG is the uptake of glucose by the gastrointestinal tract (34). The load of a substrate exiting the liver was calculated as: Load out = ([S] H HBF), where H represents the hepatic vein. Net hepatic balance was calculated as: NHB = Load out Load in. NHB for glucose (NHGB) was calculated with both the direct and indirect calculations. Only the direct calculation was employed for substrates other than glucose. A negative value indicates net uptake by the liver. Nonhepatic glucose uptake (non-hgu) was calculated as the glucose infusion rate minus NHGB. During the first h of glucose infusion, the non-hgu was corrected for the glucose required to fill the pool, using a pool fraction of 0.65 (11) and assuming that the volume of distribution for glucose equaled the volume of the extracellular fluid, or ~22% of the dog s weight (49). Net fractional glucose extraction by the liver (FE) was calculated by direct and indirect methods as NHGB/Load in. Net hepatic carbon retention, an index of the carbon available

7 7 for glycogen synthesis, equaled NHGU + net hepatic uptake of glycerol + (2 net hepatic uptake alanine) net hepatic lactate output; we have previously determined that doubling the net hepatic uptake of alanine provides a close approximation of net hepatic amino acid uptake (48). Glucose concentrations were converted from plasma to blood values by using correction factors (ratio of the blood to the plasma concentration) previously described (38). The rates of glucose appearance (R a ) and disappearance (R d ) were calculated with a 2-compartment model, using dog parameters (16, 27). In PE the glucose infusion rate was subtracted from total R a to yield endogenous R a (EndoR a ). EndoR a in PO = total R a [glucose infusion rate (1 FE)]. Unidirectional hepatic glucose uptake (HGU) was equal to the NHB of [ 3 H]glucose using the direct calculation and dividing by the inflowing [3-3 H]glucose specific activity (in dpm/µmol glucose). Hepatic glucose release (HGR) was NHGB - HGU. Hepatic sinusoidal insulin and glucagon concentrations were calculated as Load in (D)/HF, using plasma concentrations and plasma flow data. This provides an estimate of the hormone concentrations at the beginning of the sinusoid, where the hepatic artery and portal vein blood become confluent. Area under the curve (AUC) was calculated by the trapezoidal rule, using the change from basal in each animal. Results for the glucose infusion period are the mean for the entire period unless specified otherwise. Statistical analysis. Data are presented as means ± SE. Time course data were analyzed with repeated-measures analysis of variance, with post hoc comparisons by the Student- Newman-Keuls procedure. A t-test was used for comparisons of AUC. Statistical significance was accepted at P < 0.05.

8 8 RESULTS Blood glucose concentrations, hepatic blood flow, and hepatic glucose load. In response to glucose infusion, the arterial blood glucose levels increased from 4.5 ± 0.1 to 5.3 ± 0.1 mmol/l in PE and from 4.6 ± 0.1 to 5.4 ± 0.1 mmol/l in PO (Fig. 1). In PE the arterial-portal glucose gradient remained positive (0.1 ± 0.0 mmol/l), while in PO the portal glucose concentration increased to a greater extent than the arterial glucose level, creating an arterial-portal glucose gradient of 0.4 ± 0.1 mmol/l. Hepatic blood flow did not change significantly in either group (32 ± 4 and 28 ± 3 ml. kg -1. min -1 in PE and 27 ± 3 and 26 ± 2 ml. kg -1. min -1 in PO during the basal and experimental periods, respectively; data not shown). Hepatic blood flow was not different between the groups at any time. The hepatic glucose load increased significantly in both groups during glucose infusion, although the change was modest (6% and 22% in PE and PO, respectively; P < 0.05 between groups). Even though the increment in the hepatic glucose load was greater in PO than in PE, the hepatic glucose load itself did not differ significantly between groups (Fig. 1). Net hepatic glucose balance, net hepatic fractional glucose extraction, nonhepatic glucose uptake, and glucose turnover. The livers in PE and PO exhibited net hepatic glucose output (NHGO) at 8.2 ± 1.9 and 10.6 ± 2.1 µmol. kg -1. min -1, respectively, during the basal period (Fig. 2). In PE, there was a slight ( 2.2 ± 1.4 µmol. kg -1. min -1 ) decrease in NHGO after 30 min of glucose infusion. NHGO was suppressed to ~0 µmol. kg -1. min -1 by 60 min, and eventually a low rate of NHGU was evident (averaging 2.5 ± 1.5 µmol. kg - 1. min -1 during the last h). In PO, there was a shift to NHGU at a rate of 2.5 µmol. kg -1. min - 1 by 30 min, creating a change in NHGB of 13.1 ± 2.8 µmol. kg -1. min -1. By the last h of

9 9 the experiment, NHGU had increased to 4.1 ± 2.2 µmol. kg -1. min -1. The AUC of the change in net hepatic glucose balance from basal was 1.16 ± 0.21 mmol/kg 3 h in PE vs 1.94 ± 0.52 (direct calculation) and 2.30 ± 0.49 (indirect calculation) mmol/kg 3 h in PO (NS between calculations for PO; P < 0.05 for PE vs PO with both the direct and indirect calculations). Since there were no significant differences between the results of the direct and the indirect calculation for any parameter, and the indirect calculation minimizes potential errors introduced by incomplete mixing of the glucose infusate with the blood in the portal vein, the results of the indirect calculation are used in the figures and the remainder of the text. Non-HGU increased significantly during peripheral glucose infusion, from a basal rate of 8.2 ± 1.9 to a maximum of 16.0 ± 1.9 µmol. kg -1. min -1, and then declined so that the mean during the last h was 10.9 ± 1.3 µmol. kg -1. min -1 (Fig. 2). Non-HGU tended to decrease ( -3.7 ± 3.1 µmol. kg -1. min -1 ) initially after the portal glucose infusion began but then returned to basal values for the duration of the infusion (P < 0.05 between groups). Net hepatic carbon retention during the first 90 min averaged 0.5 ± 1.0 and 5.6 ± 2.8 µmol. kg -1. min -1 in PE and PO, respectively (P < 0.05). During the last 90 min it was very similar between groups, at 6.0 ± 1.2 and 8.2 ± 2.8 µmol. kg -1. min -1 in PE and PO, respectively. Glucose EndoR a decreased similarly in both groups during glucose infusion (16.5 ± 0.7 to 6.9 ± 0.5 µmol. kg -1. min -1 in PE and 15.8 ± 0.8 to 7.5 ± 0.7 µmol. kg -1. min -1 in PO)(Fig. 3). HGU did not differ between groups during the basal or glucose infusion periods (Fig. 4). HGR tended to be suppressed less in the PE than in the PO group ( from basal 5.7 ± 3.2 vs 11.1 ± 2.0 µmol. kg -1. min -1 ; P = 0.07). Glucose R d increased

10 10 similarly in the two groups ( 4.4 ± 1.4 and 4.7 ± 1.0 µmol. kg -1. min -1 in PE and PO, respectively; Fig. 3). Plasma hormone concentrations. Glucose infusion induced a significant rise in the insulin concentrations in both groups. The arterial plasma insulin concentration in PE increased from 41 ± 6 to a peak of 94 ± 34 pmol/l, with the AUC of the change from basal being 4706 ± 2391 pmol/l 3 h (Fig. 5). In PO, the arterial concentration increased from 55 ± 14 to a peak of 120 ± 41 pmol/l, with the AUC of the change from basal being 7502 ± 2471 pmol/l 3 h (P = 0.2 between groups). The hepatic sinusoidal insulin concentrations in PE increased from 113 ± 31 to a peak of 198 ± 70 pmol/l, with the AUC of the change totaling ± 5656 pmol/l 3 h. The hepatic sinusoidal insulin concentrations in PO increased from 214 ± 82 to a peak of 427 ±151 pmol/l (AUC of the change: ± pmol/l 3 h; P = 0.4 between groups). Arterial and hepatic sinusoidal glucagon concentrations declined 10-20% in both groups and were indistinguishable between the two groups throughout the studies (Table 1). Arterial concentrations and net hepatic balances of non-glucose substrates. Arterial blood concentrations of lactate increased 30-40% in both groups during glucose infusion (Table 1). The groups exhibited net hepatic lactate uptake during the basal period, then shifted to a low rate of net hepatic lactate output in response to glucose infusion. Net hepatic lactate balance was near 0 µmol. kg -1. min -1 during the last h of study in both groups. Neither the arterial concentrations nor the net hepatic balances of lactate differed between groups at any time.

11 11 Arterial alanine concentrations and net hepatic alanine uptake remained stable in both groups throughout the studies (Table 1). Arterial blood glycerol concentrations and net hepatic glycerol uptakes fell by ~30% in both groups, and arterial plasma NEFA concentrations and net hepatic NEFA uptakes declined by ~55-65%. DISCUSSION During either peripheral or portal glucose infusion at a rate approximating the basal rate of EGP (13.7 µmol. kg -1. min -1 ), arterial blood glucose concentrations in the dog increased ~20%. The tendency toward higher insulin concentrations during portal vs peripheral glucose delivery was anticipated, because hyperglycemia during portal glucose delivery has previously been shown to stimulate pancreatic insulin secretion more than equivalent hyperglycemia achieved with peripheral glucose infusion (17). However, it is not clear whether the trend toward enhancement of insulin release might have occurred solely because of the route of glucose delivery in the present study or was also the product of a higher basal secretion rate in that group. The increase in glycemia in the PE group was virtually identical to that exhibited by the pig during peripheral infusion of glucose at 11.1 µmol. kg -1. min -1 (32). The results in the PO group, on the other hand, were very similar to those in normal overnight-fasted humans receiving a continuous 7-h intraduodenal infusion of glucose at either 8.6 or 17.2 µmol. kg -1. min -1 (rates that bracketed the basal EGP, which averaged ~13 µmol. kg -1. min -1 in these subjects) (54). Those individuals exhibited an increase in arterialized venous plasma glucose levels of ~16 and 19%, respectively, during glucose infusion at 8.6 and at 17.2 µmol. kg -1. min -1. Peak plasma insulin concentrations in the humans were 2- and 3.5-fold basal, respectively, with the two glucose infusion rates (54). The ability of the human and the

12 12 dog to avert hypoglycemia during intraduodenal/intraportal infusion of glucose, despite robust increases in circulating insulin concentrations, highlights the impressive glucoregulatory capacity of these species. Appropriate titration of insulin secretion to the glycemic level represents an important difference between these two species vs the mouse. In the mouse, insulin concentrations continued to rise inappropriately during portal glucose infusion, even after hypoglycemia had developed (7). Portal glucose infusion resulted in a change in NHGB equivalent to 96% of the glucose infusion rate within 30 min. This change resulted from stimulation of HGU (5.2 ± 5.1 µmol. kg -1. min -1 ) and suppression of HGR (7.8 ± 4.2 µmol. kg -1. min -1 ). Non-HGU had been suppressed by 3.7 ± 3.1 µmol. kg -1. min -1 by 30 min, however. Consequently the arterial glucose concentration increased in spite of the near-complete compensation by the liver for the exogenous glucose. It is well known that under hyperglycemic conditions dogs demonstrate a rapid enhancement of NHGU and concomitant suppression of non- HGU in response to portally-delivered glucose (18, 38). In this study, we demonstrate for the first time that both the liver and the nonhepatic tissues respond to the portal signal in the same way in the presence of only very mild hyperglycemia. It is worth noting that, while HGU was near-maximal at 30 min, HGR continued to decrease (nadir 0.3 ± 3.2 µmol. kg -1. min -1 at 150 min). In contrast to the effect with portal glucose delivery, peripheral glucose infusion resulted in a much smaller decline in NHGO (equivalent to 16% of the glucose infusion rate) by 30 min. At that time there was no stimulation of HGU (in fact, the rate tended to decline, with a change of -1.3 ± 4.1 µmol. kg -1. min -1 from basal), and a decline of only 3.4 ± 4.8 µmol. kg -1. min -1 in HGR. However, non-hgu had increased by 8.9 ± 1.2 µmol. kg -

13 13 1. min -1, an amount equal to ~65% of the glucose infusion rate. As a consequence, the arterial glucose concentration in PE increased to almost the same extent as in PO. Whether the enhancement of NHGU in PO vs PE at 30 min resulted from the slightly larger increment in hepatic glucose load (despite the similarity between the absolute rate of glucose delivery to the liver during the PO and PE glucose infusion), modestly higher insulin levels, and/or the presence of the portal signal is not clear. A striking difference in non-hgu was evident between the two groups for 90 min. Despite similar arterial glucose levels and a tendency toward higher arterial insulin concentrations in the presence of portal glucose delivery, the PO group displayed blunted non-hgu, compared to that observed in PE. Relative peripheral insulin resistance persisted throughout the glucose infusion period in the PO group, with no stimulation of non-hgu evident in spite of continuously elevated insulin and glucose levels. The rapid enhancement of non-hgu in PE paralleled the increase in plasma insulin, and the gradual fall in non-hgu in that group followed the decline in the arterial insulin concentration. If the insulin concentrations had remained high in PE, as in PO, the difference in non-hgu between the groups evident over the first 90 min would likely have been sustained. The mechanism for the peripheral insulin resistance caused by portal glucose delivery is not clear but may be reciprocally linked to the increase in NHGU or a change in CNS sympathetic outflow as discussed below. In contrast to the findings in the dog, pig, and human, infusion of glucose at a rate equivalent to basal EGP in the mouse resulted in significant hyperglycemia (as high as 7.7 mmol/l) during peripheral glucose infusion and hypoglycemia (as low as 2.3 mmol/l) during portal glucose infusion (7). Similar findings were previously reported in

14 14 anesthetized cats (44). In 6 out of the 8 cats examined, slow intraportal injection (1 min) of small amounts (5.5, 27.8, or 55.5 µmol/kg) of an isotonic glucose solution induced a reduction (mean ~1.1 mmol/l) in arterial blood glucose within 5 min that lasted >30 min. Conversely, injection of 5.5 and 27.8 µmol/kg of glucose into the radial vein had a mild hyperglycemic effect. Only at the 55.5 µmol/kg dose did peripheral glucose delivery produce a hypoglycemic effect, and even then the decrease in blood glucose was much smaller (~0.09 vs 0.28 mmol/l) and more transient (<15 vs >30 min) than the corresponding response to portal glucose injection. It is difficult to put the data from cats into context because no pancreatic hormone data were obtained and the presence of anesthesia complicates interpretation of the data. However, the discrepant findings in the response to low-rate peripheral and portal glucose infusions in dogs, pigs, and humans vs mice and cats suggest species differences. Species-specific responses to portal and peripheral glucose infusion might be explained by differences in the predominant tissues involved in glucose disposal among species. In the mouse, both peripheral and portal glucose infusion stimulated glucose utilization in the heart, brown adipose tissue (BAT), diaphragm, and soleus muscle, but not in the liver (7). Moreover, in the mouse, unlike the dog (Fig. 3), whole body glucose turnover was significantly greater with portal vs peripheral glucose infusion. The heart and the BAT exhibited the greatest enhancement in glucose clearance during portal vs peripheral glucose infusion in the mouse (7), suggesting that glucose uptake by these tissues was a major contributor to the development of hypoglycemia. Burcelin et al (6, 7) proposed that the stimulation of peripheral glucose uptake was activated by a hepatoportal glucose sensor, and peripheral glucose delivery did not trigger the response.

15 15 In contrast to the findings in the mouse, in the dog the liver has significant involvement in disposal of a glucose load whichever route of delivery is used, although it plays a much larger role in response to portal glucose. Myocardial glucose extraction is relatively low in both humans and dogs except under ischemic conditions (15, 33, 53), and adult humans and dogs have little or no BAT (9, 22, 23). Instead, skeletal muscle is the predominant nonhepatic tissue responsible for insulin-stimulated glucose disposal in the dog (18). Portal glucose delivery is associated with a decrease in the firing rate in the hepatic branch of the splanchnic nerve and in the adrenal nerve (both primarily sympathetic) and an increase in efferent firing the pancreatic branch of the vagus (primarily parasympathetic) (36). Presumably these effects remove an inhibition to hepatic glucose uptake (31) and stimulate insulin secretion (17). In contrast to the decrease in sympathetic signaling to the liver, sympathetic nervous system (SNS) input to peripheral tissues, as evidenced by norepinephrine (NE) levels and turnover, is increased after mixed meals (40, 50) and oral carbohydrate intake (41, 47, 51, 52). Similarly, circulating serotonin (5-hydroxytryptamine, or 5-HT) concentrations rise following high-carbohydrate meals or glucose ingestion (3). NE is generally regarded as having an inhibitory effect on glucose disposal in vivo (10, 24), but this effect has not always been observed (46). The disparate findings may relate to dosages utilized in the different studies since the dose-response of NE on glucose and oxygen uptake in vitro is bell-shaped (42) and/or the differing study conditions, e.g., insulinemia and glycemia. Nevertheless, studies on isolated tissues indicate that sympathetic neurotransmitters can have different effects on different tissues. NE (25) and

16 16 the 3-adrenoreceptor-specific agonist CL-316,243 (13, 26) stimulate adipose tissue glucose uptake in the rat, with BAT having 100-fold greater capacity for NE stimulation of glucose uptake than white adipose tissue (25). On the other hand, NE reduces insulinmediated glucose uptake in perfused rat hindlimb (42). In regard to a potential effect of 5- HT on extrahepatic glucose uptake, it has been demonstrated that dexfenfluramine, a stimulator of 5-HT release and inhibitor of 5-HT reuptake, can increase glucose uptake in perfused rat white adipose tissue (4). However, 5-HT produced acute insulin resistance in the perfused rat hindlimb (35, 43). When individual muscles were examined, it was determined that those rich in slow twitch oxidative fibers actually increased their glucose uptake during 5-HT perfusion, while the muscles with predominantly fast-twitch glycolytic fibers showed suppressed glucose uptake (43). Since the latter fiber type predominated in the hindlimb, the net effect was a reduction in glucose uptake (43). Thus, a likely explanation of the discrepant findings of Burcelin et al (7) and the current report is that SNS activity is responsible for the extrahepatic effects in both cases, and the results of the increase in SNS activity are dependent on the predominant tissues involved in glucose disposal. SNS activity markedly stimulates BAT glucose uptake where that tissue is present but can suppress insulin-mediated glucose uptake in skeletal muscle. Interestingly, the only skeletal muscle that Burcelin et al (7) observed to exhibit a substantial enhancement of glucose uptake following portal glucose infusion was soleus (composed primarily of oxidative fibers), suggesting that 5-HT is involved in the extrahepatic effects of the portal signal. In conclusion, during both portal and peripheral glucose delivery at a rate approximating the basal rate of endogenous glucose release, the dog was able to maintain

17 17 blood glucose at a near-euglycemic level. There was, however, a pronounced difference in the adaptive mechanisms employed in the PE and PO groups, particularly during the early portion of the infusion period. During peripheral glucose infusion there was a marked increase in glucose uptake by the nonhepatic tissues with only a small suppression of NHGO occurring as a consequence of a fall in HGR. Portal glucose infusion, on the other hand, prompted a rapid conversion to NHGU stemming from both enhancement of hepatic glucose uptake and suppression of HGR, but there was no stimulation (and in fact slight suppression) of nonhepatic glucose uptake. This difference resulted in a marked alteration in the site of glucose storage, with an augmentation of the liver's role during portal glucose delivery, such that net hepatic carbon retention was equivalent to ~41% of the infused glucose during the first 90 min in PO vs 4% in PE. Since overall glucose R d was similar in the two protocols, it is clear that muscle and liver played reciprocal roles. There is general agreement (6, 20, 28) that a hepatoportal glucose sensor is involved in the regulation of glucose metabolism, with both canine and murine data indicating that this sensor is involved in regulation of the disposition of portally delivered glucose in nonhepatic tissues. Nevertheless, it would appear that the sensor may operate differently in the mouse (where it directs glucose into the extrahepatic tissues even at the expense of hypoglycemia) than in the dog (where it partitions glucose between the liver and extrahepatic tissues in a manner that sustains near- euglycemia). While further study of the hepatoportal glucose sensor mechanism in the mouse will undoubtedly yield useful insights, glucoregulation in the mouse does not appear to be as rigorous as it is in the dog (current data) or the human (54). Therefore, data regarding

18 18 glucose sensing in the mouse must be viewed in context with data from other models before it can be extrapolated to human physiology.

19 19 ACKNOWLEDGEMENTS The authors appreciate the assistance of Wanda Snead and Angelina Penaloza of the Vanderbilt Diabetes Research and Training Center Hormone Core Lab and of Jon Hastings. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R-01-DK and Diabetes Research and Training Center Grant SP-60-AM

20 20 REFERENCES 1. Adkins BA, Myers SR, Hendrick GK, Stevenson RW, Williams PE, and Cherrington AD. Importance of the route of intravenous glucose delivery to hepatic glucose balance in the conscious dog. J Clin Invest 79: , Bergman RN, Bier JR, and Hourigan PM. Intraportal glucose infusion matched to oral glucose absorption. Lack of evidence for "gut factor" involvement in hepatic glucose storage. Diabetes 31: 27-35, Blum I, Vered Y, Graff E, Grosskopf Y, Don R, Harsat A, and Raz O. The influence of meal composition on plasma serotonin and norepinephrine concentrations. Metabolism 41: , Boschmann M, Adams F, and Klaus S. In situ metabolic and hemodynamic response to dexfenfluramine in white adipose tissue of rats. Ann Nutr Metab 45: 24-29, Burcelin R, del Carmen Munoz M, Guillam MT, and Thorens B. Liver hyperplasia and paradoxical regulation of glycogen metabolism and glucose-sensitive gene expression in GLUT2-null hepatocytes. Further evidence for the existence of a membrane-based glucose release pathway. J Biol Chem 275: , Burcelin R, Dolci W, and Thorens B. Glucose sensing by the hepatoportal sensor is GLUT2-dependent: in vivo analysis in GLUT2-null mice. Diabetes 49: , Burcelin R, Dolci W, and Thorens B. Portal glucose infusion in the mouse induces hypoglycemia: evidence that the hepatoportal glucose sensor stimulates glucose utilization. Diabetes 49: , 2000.

21 21 8. Cardin S, Emshwiller M, Jackson PA, Snead WL, Hastings J, Edgerton DS, and Cherrington AD. Portal glucose infusion increases hepatic glycogen deposition in conscious unrestrained rats. J Appl Physiol 87: , Champigny O, Ricquier D, Blondel O, Mayers RM, Briscoe MG, and Holloway BR. Beta 3-adrenergic receptor stimulation restores message and expression of brown-fat mitochondrial uncoupling protein in adult dogs. Proc Natl Acad Sci U S A 88: , Connolly CC, Steiner KE, Stevenson RW, Neal DW, Williams PE, Alberti KG, and Cherrington AD. Regulation of glucose metabolism by norepinephrine in conscious dogs. Am J Physiol 261: E , Cowan JS and Hetenyi G, Jr. Glucoregulatory responses in normal and diabetic dogs recorded by a new tracer method. Metabolism 20: , Davis MA, Williams PE, and Cherrington AD. Effect of a mixed meal on hepatic lactate and gluconeogenic precursor metabolism in dogs. Am J Physiol 247: E , de Souza CJ, Hirshman MF, and Horton ES. CL-316,243, a beta3-specific adrenoceptor agonist, enhances insulin-stimulated glucose disposal in nonobese rats. Diabetes 46: , DeFronzo RA, Ferrannini E, Hendler R, Wahren J, and Felig P. Influence of hyperinsulinemia, hyperglycemia, and the route of glucose administration on splanchnic glucose exchange. Proc Natl Acad Sci USA 75: , Depre C, Vanoverschelde JL, Gerber B, Borgers M, Melin JA, and Dion R. Correlation of functional recovery with myocardial blood flow, glucose uptake, and

22 22 morphologic features in patients with chronic left ventricular ischemic dysfunction undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 113: , Dobbins RL, Davis SN, Neal DW, Cobelli C, Jaspan J, and Cherrington AD. Compartmental modeling of glucagon kinetics in the conscious dog. Metabolism 44: , Dunning BE, Moore MC, Ikeda T, Neal DW, Scott MF, and Cherrington AD. Portal glucose infusion exerts an incretin effect associated with changes in pancreatic neural activity in conscious dogs. Metabolism 51: , Galassetti P, Shiota M, Zinker BA, Wasserman DH, and Cherrington AD. A negative arterial-portal vein glucose gradient decreases skeletal muscle glucose uptake in the conscious dog. Am J Physiol 275: E101-E111, Hendrick GK, Frizzell RT, Williams PE, and Cherrington AD. Effect of hyperglucagonemia on hepatic glycogenolysis and gluconeogenesis after a prolonged fast. Am J Physiol 258: E841-E849, Hevener AL, Bergman RN, and Donovan CM. Hypoglycemic detection does not occur in the hepatic artery or liver: findings consistent with a portal vein glucosensor locus. Diabetes 50: , Holste LC, Connolly CC, Moore MC, Neal DW, and Cherrington AD. Physiological changes in circulating glucagon alter hepatic glucose disposition during portal glucose delivery. Am J Physiol 273: E488-E496, Lean ME, James WP, Jennings G, and Trayhurn P. Brown adipose tissue in patients with phaeochromocytoma. Int J Obes10: , 1986.

23 Lean ME, James WP, Jennings G, and Trayhurn P. Brown adipose tissue uncoupling protein content in human infants, children and adults. Clin Sci (Lond) 71: , Lembo G, Capaldo B, Rendina V, Iaccarino G, Napoli R, Guida R, Trimarco B, and Sacca L. Acute noradrenergic activation induces insulin resistance in human skeletal muscle. Am J Physiol 266: E , Liu X, Perusse F, and Bukowiecki LJ. Chronic norepinephrine infusion stimulates glucose uptake in white and brown adipose tissues. Am J Physiol 266: R914- R920, Liu X, Perusse F, and Bukowiecki LJ. Mechanisms of the antidiabetic effects of the beta 3-adrenergic agonist CL in obese Zucker-ZDF rats. Am J Physiol 274: R , Mari A. Estimation of the rate of appearance in the non-steady state with a twocompartment model. Am J Physiol 263: E , Moore MC and Cherrington AD. Regulation of net hepatic glucose uptake: interaction of neural and pancreatic mechanisms. Reprod Nutr Dev 36: , Moore MC, Cherrington AD, and Wasserman DH. Regulation of hepatic and peripheral glucose disposal. Best Pract Res Clin Endocrinol Metab 17: , Moore MC, Hsieh PS, Neal DW, and Cherrington AD. Nonhepatic response to portal glucose delivery in conscious dogs. Am J Physiol Endocrinol Metab 279: E , 2000.

24 Moore MC, Rossetti L, Pagliassotti MJ, Monohan M, Venable C, Neal D, and Cherrington AD. Neural and pancreatic influences on net hepatic glucose uptake and glycogen synthesis. Am J Physiol 271: E215-E222, Muller MJ, Moring J, and Seitz HJ. Regulation of hepatic glucose output by glucose in vivo. Metabolism 37: 55-60, Myears DW, Sobel BE, and Bergmann SR. Substrate use in ischemic and reperfused canine myocardium: quantitative considerations. Am J Physiol 253: H , Myers SR, Biggers DW, Neal DW, and Cherrington AD. Intraportal glucose delivery enhances the effects of hepatic glucose load on net hepatic glucose uptake in vivo. J Clin Invest 88: , Newman JM, Rattigan S, and Clark MG. Nutritive blood flow improves interstitial glucose and lactate exchange in perfused rat hindlimb. Am J Physiol Heart Circ Physiol 283: H , Niijima A. Reflex control of the autonomic nervous system activity from the glucose sensors in the liver in normal and midpontine-transected animals. J Auton Nerv Syst 10: , Ogihara N, Ebihara S, Kawamura W, Okamoto M, Sakai T, Takiguchi K, Morita T, Uchida R, Matsuyama Y, Hayashi Y, Arakawa Y, and Kikuchi M. Characterization of the portal signal in a nonsteady hyperglycemic state in conscious dogs. Am J Physiol Endocrinol Metab 284: E , 2003.

25 Pagliassotti MJ, Holste LC, Moore MC, Neal DW, and Cherrington AD. Comparison of the time courses of insulin and the portal signal on hepatic glucose and glycogen metabolism in the dog. J Clin Invest 97: 81-91, Pagliassotti MJ, Myers SR, Moore MC, Neal DW, and Cherrington AD. Magnitude of negative arterial-portal glucose gradient alters net hepatic glucose balance in conscious dogs. Diabetes 40: , Patel J, Coppack SW, Rawesh A, Miles JM, and Eisenhofer G. Regional and total body changes in sympathoadrenal activity in response to eating. Int J Obes Relat Metab Disord 24 Suppl 2: S , Rappaport EB, Young JB, and Landsberg L. Initiation, duration and dissipation of diet-induced changes in sympathetic nervous system activity in the rat. Metabolism 31: , Rattigan S, Dora KA, Colquhoun EQ, and Clark MG. Inhibition of insulinmediated glucose uptake in rat hindlimb by an alpha-adrenergic vascular effect. Am J Physiol 268: E , Rattigan S, Dora KA, Colquhoun EQ, and Clark MG. Serotonin-mediated acute insulin resistance in the perfused rat hindlimb but not in incubated muscle: a role for the vascular system. Life Sci 53: , Rodriquez-Zendejas AM, Vega C, Soto-Mora LM, and Russek M. Some effects of intraperitoneal glucose and of intraportal glucose and adrenaline. Physiol Behav 3: , 1968.

26 Rothman DL, Magnusson I, Katz LD, Shulman RG, and Shulman GI. Quantitation of hepatic glycogenolysis and gluconeogenesis in fasting humans with 13C NMR. Science 254: , Sacca L, Morrone G, Cicala M, Corso G, and Ungaro B. Influence of epinephrine, norepinephrine, and isoproterenol on glucose homeostasis in normal man. J Clin Endocrinol Metab 50: , Schwartz JH, Young JB, and Landsberg L. Effect of dietary fat on sympathetic nervous system activity in the rat. J Clin Invest 72: , Sindelar DK, Chu CA, Venson P, Donahue EP, Neal DW, and Cherrington AD. Basal hepatic glucose production is regulated by the portal vein insulin concentration. Diabetes 47: , Steele R, Wall JS, DeBodo RC, and Altszuler N. Measurement of size and turnover rate of body glucose pool by the isotope dilution method. Am J Physiol 187: 15-24, Vaz M, Turner A, Kingwell B, Chin J, Koff E, Cox H, Jennings G, and Esler M. Postprandial sympatho-adrenal activity: its relation to metabolic and cardiovascular events and to changes in meal frequency. Clin Sci (Lond) 89: , Young JB and Landsberg L. Stimulation of the sympathetic nervous system during sucrose feeding. Nature 269: , Young JB, Rowe JW, Pallotta JA, Sparrow D, and Landsberg L. Enhanced plasma norepinephrine response to upright posture and oral glucose administration in elderly human subjects. Metabolism 29: , 1980.

27 Young LH, Renfu Y, Russell R, Hu X, Caplan M, Ren J, Shulman GI, and Sinusas AJ. Low-flow ischemia leads to translocation of canine heart GLUT-4 and GLUT-1 glucose transporters to the sarcolemma in vivo. Circulation 95: , Zangeneh F, Basu R, Shah P, Arora P, Camilleri M, and Rizza RA. Enteral infusion of glucose at rates approximating EGP enhances glucose disposal but does not cause hypoglycemia. Am J Physiol Endocrinol Metab 285: E , 2003.

28 28 Table 1. Arterial and hepatic sinusoidal plasma glucagon concentrations and arterial concentrations and net hepatic balance of substrates Parameter and group Basal Period Glucose Infusion Period Arterial plasma glucagon (ng/l) PE 39 ± 3 31 ± 1 PO 32 ± 4 28 ± 6 Hepatic sinusoidal plasma glucagon (ng/l) PE 42 ± 7 38 ± 11 PO 39 ± 4 32 ± 6 Arterial blood lactate (mmol/l) PE 564 ± ± 129 PO 441 ± ± 97 Net hepatic lactate balance (µmol. kg -1. min -1 ) PE -8.0 ± ± 1.4 PO -6.7 ± ± 0.7 Arterial blood alanine (mmol/l) PE 0.31 ± ± 0.02 PO 0.32 ± ± 0.01 Net hepatic alanine uptake (µmol. kg -1. min -1 ) PE 3.0 ± ± 0.3 PO 2.9 ± ± 0.4 Arterial blood glycerol (mmol/l) PE 0.10 ± ± 0.01

29 29 PO 0.10 ± ± 0.01 Net hepatic glycerol uptake (µmol. kg -1. min -1 ) PE 2.4 ± ± 0.4 PO 2.2 ± ± 0.4 Arterial plasma NEFA (mmol/l) PE 1.1 ± ± 0.1 PO 0.9 ± ± 0.1 Net hepatic NEFA uptake (µmol. kg -1. min -1 ) PE 3.6 ± ± 0.5 PO 3.3 ± ± 0.9 Data are means ± SE. Basal values are the means of the 3 sampling points during the basal period; glucose infusion period values are the mean of the 6 sampling points between 30 and 180 min. Negative values reflect net uptake. n = 5/group. There were no significant differences between groups.

30 30 FIGURE LEGENDS Fig. 1 Arterial and portal blood glucose concentrations and hepatic glucose loads in conscious 42-h-fasted dogs in the basal state and during infusion of glucose at 13.7 µmol. kg -1. min -1 via a peripheral (PE) or the hepatic portal (PO) vein. n = 5/group. *P < 0.05 between groups Fig. 2 Net hepatic glucose uptake (NHGU), net hepatic fractional extraction of glucose, and nonhepatic glucose uptake. See legend to Figure 1. The insets show the AUCs of the change from basal. * P < 0.05 between groups. Fig. 3 Endogenous glucose R a and glucose R d. See legend to Figure 1. There were no significant differences between groups. Fig. 4 Hepatic glucose uptake and release. See legend to Figure 1. There were no significant differences between groups. The change in HGR between the basal and infusion periods tended to be greater in PO than PE (P = 0.07). Fig. 5 Arterial and hepatic sinusoidal plasma insulin concentrations. See legend to Figure 1. There were no significant differences between groups.

31 31 Fig. 1 Arterial Blood Glucose (mmol/l) Portal Blood Glucose (mmol/l) Glucose Infusion Portal Peripheral * * * * * * Hepatic Glucose Load (µmol. kg -1. min -1 ) Time (min)

32 32 Fig. 2 Net Hepatic Glucose Balance (µmol. kg -1. min -1 ) Net Hepatic Fractional Extraction of Glucose Nonhepatic Glucose Uptake (µmol. kg -1. min -1 ) Output Uptake Glucose Infusion AUC from basal mmol/kg. 3 h * Portal Peripheral * AUC from basal mmol/kg. 3 h * * Time (min) * * Pe * Pe Po Po

33 33 Fig. 3 EndoR a (µmol. kg -1. min -1 ) Glucose Infusion R d (µmol. kg -1. min -1 ) 30 Portal 25 Peripheral Time (min)

34 34 Fig. 4 Hepatic Glucose Uptake (µmol. kg -1. min -1 ) Glucose Infusion Hepatic Glucose Release (µmol. kg -1. min -1 ) 20 Portal 15 Peripheral Time (min)

35 35 Fig. 5 Arterial Plasma Insulin (pmol/l) Hepatic Sinusoidal Plasma Insulin (pmol/l) Glucose Infusion Peripheral Portal Time (min)