tisol on hepatic carbohydrate metabolism. Am. J. Physiol. 264. (Endocrinol. Metab. 27): Ell9-E127, 1993.-This study was undertaken to investigate the effects of ...
Effects of chronic elevation in plasma cortisol on hepatic carbohydrate metabolism RICHARD E. GOLDSTEIN, D. BROOKS LACY, ALAN
DAVID H. WASSERMAN, OWEN P. McGUINNESS, D. CHERRINGTON, AND NAJI N. ABUMRAD
Department of Molecular Physiology and Biophysics, Department of Surgery, and Diabetes Research and Training Center, Vanderbilt University School of Medicine, Nashville, Tennessee 37232 Goldstein, Richard E., David H. Wasserman, Owen P. McGuinness, D. Brooks Lacy, Alan D. Cherrington, and Naji N. Abumrad. Effects of chronic elevation in plasma cortisol on hepatic carbohydrate metabolism. Am. J. Physiol. 264 (Endocrinol. Metab. 27): Ell9-E127, 1993.-This study was undertaken to investigate the effects of chronic physiological elevations in plasma cortisol on glycogenolysis and gluconeogenesis in conscious, overnight-fasted dogs. Experiments consisted of an 80-min tracer and dye equilibration period and a 40-min sampling period. Infusions of D-[3-3H]glucose, L-[Ul*C]alanine, and indocyanine green dye were used to assess glucose production (R,) and gluconeogenesis using tracer and arteriovenus (a-v) difference techniques. In the cortisol group, (n = lo), a continuous infusion of hydrocortisone (3.5 I-cg*wl l rein+) was begun 5 days before the experiment and continued throughout the sampling period. In the saline group (n = lo), there was no infusion of cortisol. The fivefold elevation in plasma cortisol increased plasma insulin from 12 t 2 to 19 & 2 pU/ml. Glucose R, was elevated in the cortisol group (3.5 + 0.2 vs. 2.8 t 0.2 mgekg-l emin-l) but net hepatic glucose output was markedly diminished (1.2 t 0.4 vs. 2.7 t 0.3 mg=kg-l. min-l). Gluconeogenic conversion of alanine to glucose was increased slightly by cortisol (0.60 t 0.13 to 0.99 & 0.12 pmol kg-l . min-l), but the gluconeogenic efficiency of the liver was unchanged. Cortisol increased hepatic glycogen content evident at the end of the study greater than twofold (76.4 t 7.9 vs. 30.0 t 4.7 g/liver). These results suggest that cortisol I) promotes glucose cycling through glycogen, 2) greatly inhibits nonhepatic glucose utilization, 3) increases hepatic gluconeogenesis in vivo primarily through enhanced substrate delivery to the liver, and 4) raises plasma insulin levels, which restrains intrahepatic gluconeogenesis. glycogenolysis; gluconeogenesis; tracers; hepatic glucose output; glycerol; lactate; alanine; ,&hydroxybutyrate; glucose production l
in plasma cortisol levels has been reported in patients who have sustained trauma (32) and sepsis (11). In vitro studies (14, 16) have consistently demonstrated the ability of glucocorticoids to stimulate gluconeogenesis in the isolated hepatocyte and the perfused liver and suggested that glucocorticoids may play an important role in the stress response by promoting hepatic glycogen formation (29). Later, in vivo work (22, 23) supported these findings by showing that chronic hypercortisolemia increased tracer-determined glucose production. The studies by Issekutz and associates (22, 23) also showed that hypercortisolemia increased plasma lactate and alanine levels and suggested that there was increased conversion of lactate and alanine into glucose. The relative contributions of glycogenolysis and gluconeogenesis to the increase in hepatic glucose production was not, however, discerned. Bessey et al. (6) infused cortisol alone or in combination with epinephrine plus glucagon into human volunA
FOUR-
TO
FIVEFOLD
ELEVATION
0193-1849/93
$2.00
Copyright
0 1993
teers for 4 days. The chronic cortisol infusion increased nitrogen excretion and metabolic rate and decreased whole body glucose disposal and insulin-stimulated forearm glucose uptake. The above increases were less than those that occurred with the “triple infusion,” but the decreases in glucose disposal and forearm glucose uptake were no greater with the triple infusion. Johnston et al. (24) increased cortisol levels threefold over a 3-day period with Synacthen (tetracosactrin; Depot). Increases were noted in serum insulin, blood glucose, blood lactate, and blood alanine levels. The authors suggested that the elevation in both lactate and alanine provided substrates for increased gluconeogenesis. Previous studies investigating the effects of chronic hypercortisolemia employed only arterial blood sampling. Studies that employ the triple-catheter technique (sampling blood entering and exiting the liver) can assess the hepatic balance of gluconeogenic substrates and can yield data that clarify effects of cortisol on the direct and indirect pathways of glycogen synthesis. The general hypothesis for the present study was that cortisol enhances glucose production in vivo by increasing the contribution of gluconeogenesis. The present study was undertaken to more fully examine the ability of chronic hypercortisolemia to modulate gluconeogenesis and glycogenolysis in vivo. Overnight-fasted conscious dogs were studied after a &day period of continuous hydrocortisone infusion that increased the plasma cortisol level fivefold. MATERIALS
AND
METHODS
Animals and surgical procedures. Experiments were performed on a total of 20 mongrel dogs (16-25 kg) of either sex that had been fed a standard diet of meat and chow (31% protein, 52% carbohydrate, 11% fat, and 6% fiber based on dry weight; Kal Kan meat, Kal Kan Foods, Vernon, CA, and Wayne Dog Chow, Allied Mills, Chicago, IL). Techniques used in the preparation of the animals have been previously described (1). Two weeks before each experiment, a laparotomy was performed under general endotrachial anesthesia (pentobarbital sodium, 25 mg/kg iv) with ventilation maintained with the use of a Harvard Apparatus respiratory pump (Boston, MA). Using sterile surgical technique, we placed Silastic infusion catheters (0.03 mm ID) in jejunal and splenic veins. Silastic sampling catheters (0.047 mm ID) were placed in the portal vein such that the tip lay 1 in. below the liver, in the left common hepatic vein so as to sample blood draining 60% of the liver, and in the left femoral artery. -After insertion, the catheters were filled with 200 U/ml heparinized saline, their free ends were knotted, and they were placed in a subcutaneous pocket so that complete closure of skin incisions was possible. For the protocol that specified a 5-day infusion of hydrocortisone and in three of the saline control animals, the dogs were The
American
Physiological
Society
El19
El20
CHRONIC
HYPERCORTISOLEMIA
AND
again similarly anesthetized 1 wk prior to the experiment, and a Silastic infusion catheter was directed into the right jugular vein via the hyoid vein after a small cut down was performed. The catheter was tunneled subcutaneously around the neck and brought out through the skin over the distal cervical spine and secured. The catheter was heparinized, knotted, and secured to a dog jacket placed on the animal. In three of the dogs that were to receive a &day infusion of hydrocortisone, a Silastic sampling catheter was placed in the left renal vein (20) during the initial surgery. All studies were performed on 18-h, overnight-fasted, conscious dogs. Five days before the experiment the leukocyte count and hematocrit were determined. Only animals that had a leukocyte count ~18,000, a hematocrit >35%, a good appetite, normal stools, and no evidence of infections were studied. Five days before the experiment, in those dogs with a jugular infusion catheter, the catheter was connected to an infusion pump capable of delivering fluid at a rate of 64 pi/h (Mill Hill, Harvard Apparatus, South Natick, MA). On the day of the experiment the subcutaneous ends of the catheters were exteriorized through small skin incisions made using local anesthesia (2% lidocaine, Astro, Worcester, MA). The contents of the catheters were aspirated, and the catheters were flushed with saline. The conscious dog was then allowed to stand calmly in a Pavlov harness for 20 min before the start of the experiment. Experimental procedures. Each experiment consisted of an 80-min tracer and dye equilibration period (-80 to 0 min) and a 40-min sampling period (O-40 min) during which blood was sampled every 10 min from the artery (5 samples) and every 20 min from the hepatic and portal veins (3 samples from each). At -80 min, a primed (50 &i) constant infusion of D-[~-~H]glucose (0.35 &i/min) and a constant infusion of L-[U-~~C]alanine (0.026 PCi. kg-l min-l) were started via an 18-gauge angiocath in the cephalic vein and continued throughout the experiment. The priming dose of D- [3-3H]glucose equaled the amount of tracer infused in 140 min. At -80 min, a constant infusion of indocyanine green (0.1 mg mm2 min-‘) was also started in the cephalic vein and continued throughout the entire experiment. Two protocols were employed. In the first protocol (Cortisol, n = IO), a continuous infusion of hydrocortisone (3.5 pg. kg-l min-l) was begun 5 days before the experiment and continued until time equaled 40 min. Five of these dogs were subsequently entered into an additional protocol that is not the subject of this report. The remaining five animals were killed, and open liver biopsies were taken from each of four liver lobes. The tissue was immediately frozen in liquid nitrogen and stored at -7OOC. In the second protocol (Saline, n = IO), cortisol was not chronically infused, and five dogs were killed at 40 min so that open liver biopsies could be taken. The other five dogs were subsequently entered into an additional protocol that is not the subject of this report. In 3 of the 10 saline dogs, a continuous saline infusion was begun 5 days before the day of the experiment, and the infusion was continued for 40 min. This confirmed that the infusion catheter did not induce any metabolic changes. Animals were maintained, and experiments were performed, in accordance with the guidelines of the Animal Care Committee of Vanderbilt University in an American Association for Accreditation of Laboratory Animal Care accredited facility. Processing of blood samples. The collection and processing of blood samples have been described (1, 8). Plasma glucose concentrations were determined by the glucose oxidase method in a Beckman glucose analyzer (Beckman Instrument, Fullerton, CA). Plasma glucose radioactivity (3H and 14C)was determined by liquid scintillation counting after deproteinization with barium hydroxide and zinc sulfate and reconstitution in 1 ml of distilled water and 10 ml Aqueous Counting Solution (Amerl
l
l
l
CARBOHYDRATE
METABOLISM
sham, Arlington Heights, IL). Concentrations of indocyanine green were determined spectrophotometrically (805 nm) in arterial and hepatic vein plasma samples. Labeled and unlabeled plasma alanine and lactate concentrations were determined with a short-column ion exchange chromatographic system that has been described previously (1). Whole blood lactate, alanine, glycerol, and P-hydroxybutyrate concentrations were determined with the use of the Technicon AutoAnalyzer in samples deproteinized with 4% perchloric acid. Immunoreactive glucagon concentrations were determined in plasma samples containing 50 ~1 Trasylol (FBA Pharmaceuticals) by radioimmunoassay with 30K antiserum; the interassay coefficient of variation (CV) was 22%. Phadebas insulin radioimmunoassay kits were purchased from Pharmacia (Piscataway, NJ); the interassay CV was 11%. Plasma cortisol was measured with the Clinical Assays gamma coat radioimmunoassay kit (Clinical Assays, Travenol-Genetech Diagnostics, Cambridge, MA); the interassay CV was 6%. Plasma epinephrine and norepinephrine levels were determined with the Cat-A-Kit radioenzymatic kit (Upjohn, Kalamazoo, MI). The interassay CVs were 13 and 11% for epinephrine and norepinephrine, respectively. Hepatic glycogen content and specific radioactivity were determined using an enzymatic method (7). Materials. [ 3-3H] glucose (New England Nuclear, Boston, MA) was used as the glucose tracer (500 &i/O.005 mg), and [U-14C]alanine (Amersham, Chicago, IL) was used as the labeled gluconeogenic precursor (171 mCi/mmol). Indocyanine green was purchased from Hynson, Westcott, and Dunning (Baltimore, MD). Glucagon 30K antiserum was obtained from the University of Texas Southwestern Medical School (Dallas, TX), and the standard glucagon and 1251-labeledglucagon were obtained from Novo (Copenhagen, Denmark). Catecholamine assay kits (Cat-A-Kit) and hydrocortisone were obtained from Upjohn (Kalamazoo, MI). Insulin was obtained from SquibbNovo (Princeton, NJ), and glucagon was obtained from Lilly (Indianapolis, IN). Cortisol radioimmunoassay kits were obtained from Micromedic Systems. (Horsham, PA). The [3-3H] glucose infusate contained cold glucose so that its final concentration was 1 mg/ml. The indocyanine green infusate was prepared with sterile water, and the [U-14C]alanine and [ 3-3H]glucose i&sates were prepared with normal saline. Calculations. In all studies, measurement of hepatic extraction of indocyanine green dye was used to assesstotal hepatic plasma flow (HPF). The proportion of hepatic blood flow (HBF) provided by the hepatic artery was assumed to be 28% based on a compilation of data by Greenway and Stark (18). Recent studies in this laboratory using pulsed Doppler flow probes indicate that the proportion of blood flow provided by the hepatic artery may be closer to 20%, and results were also calculated using this percentage. This change had little effect on the data and did not alter the conclusions. To be consistent with our previous publications in which indocyanine green was used to measure HBF we chose to depict the data obtained using an assumption of 28% arterial flow. Net hepatic balances (whole blood) of alanine, lactate, glycerol, and p-hydroxybutyrate were determined by the formula [H - (0.28A + 0.72P)] X HBF, where H, A, and P represent the concentrations of the given substrate in the hepatic vein, femoral artery, and portal vein, respectively. HBF was derived from the hepatic plasma flow; HBF equals HPF/(l - hematocrit). All hepatic balances are depicted as positive values but are labeled appropriately as either output (production) or uptake. Net hepatic blood gIucose output was determined using the above equation after converting the plasma glucose levels to blood glucose levels by multiplying by 0.73 (1). Hepatic fractional extraction of alanine was determined by the formula [(0.28A + 0.72P) - H]/(0.28A + 0.72P). Hepatic glucose production (R,) and glucose utiliza-
CHRONIC
HYPERCORTISOLEMIA
AND
tion (Rd) were determined by the equations for isotope dilution during a constant infusion of radioactive glucose ( [3-3H] glucose) as modified by DeBodo et al. (12). Although the equation allows for the determination of glucose kinetics during non-steady state, in the present study glucose production and utilization were determined during steady state. Therefore, the equations for R, and Rd simplify to R, =
tracer infusion rate 9 SA
R, = Rd
where SA is the arterial [3H]glucose specific activity (dpm/ mg] . The hepatic [ 14C]glucose production rate (dpm kg-l min-l) was determined using the tracer technique described by Chiasson et al. (8) l
[ 14C]glucoseproduction =
CARBOHYDRATE
El21
METABOLISM
the liver from circulating glucose was determined by multiplying the mass (mg) of glycogen deposited in the liver by the average inflowing [‘14C]glucosespecific activity (dpm/mg) [14C]glucosefrom circulating glucose
glycogen from glucose
O.iB-e d.2ia
+ 0.72+ 0.72-
The [14C]glucose derived from gluconeogenesis (dpm) was calculated by subtracting the [14C]glucose deposited from circulating glucose from the total [14C]glucose in the liver.
l
d[ 14C]glucose x pV + (Rd x SA) dt ( )
= [ 14C]glUCOSeliver
-
[ 14C]glucosefrom circulating glucose
where the rate of glycogen deposition from gluconeogenesis was determined by dividing the radioactivity deposited via gluconeogenesis by the hepatic vein [ 14C]lactate specific activity (presumed to reflect the intrahepatic pyruvate specific activity), dividing by 2 to account for &carbon precursor molecules being converted to 6-carbon glucose molecules, and converting micromoles to miligrams of glucose
where [ 14C]glucose is arterial [ 14C]glucose radioactivity [disintegrations/min (dpm)/ml], p is the pool fraction, V is the glucose distribution volume (ml), and SA is the arterial [ 14C]glucose specific activity (dpm/mg). Once the [14C]glucose production rate was determined, the gluconeogenic conversion rate ~‘4cl~1ucose,l~~o~~o~~~~~i~ of alanine to glucose was calculated by dividing the [14C]glucose glYCogengluconeogenesis = SAk;E$ x t x body wt x 5.6 pmol/mg production rate (dpm kg-l min-l) by the specific activity (dpm/pmol) of alanine entering the liver (weighting the arterial Tracer-determined hepatic glucose uptake (HGU), which and portal specific activities) should approximate the difference between glucose R, and net hepatic glucose output (HNGO) if the liver is the only source [ 14C]glucoseproduction rate of glucose production, was calculated by dividing the net heconversion = patic [3H]glucose balance by the inflowing [3H]glucose specific 0.28Alai$&,l+ 0.72Al$zk, activity. 0.28Alaz$,, + 0.72Alq$$ ( ) The methods used for calculating gluconeogenic conversion rate and efficiency yield minimal estimates due to isotopic diwhere Alahot represents the plasma [ 14C]alanine radioactivity (dpm/ml) , and Alacoldrepresents the plasma alanine concentra- lution in the oxaloacetate pool. Calculation of the absolute rate tion (pmol/ml) in the artery or portal vein. The efficiency of of gluconeogenesis from alanine would require the determinahepatic gluconeogenesis was calculated by dividing the [‘“Cl- tion of a correction factor as described by Hetenyi (21) or the glucose production rate by the rate of hepatic [14C]alanine up- use of several tracers and specific activity determinations of the glucose skeleton as discussed by Katz (25). Experiments take by Hetenyi (21) suggested that the correction factor in fact varies only slightly in normal, insulin-deprived, and steroidefficiency treated dogs. [ 14C]glucoseproduction rate Results were obtained by first finding the mean of the data = (O.28[14C]A1a,,te,ial + 0.72[ “C]Alqotil - [‘4C]Alah,,& X flOW from each individual animal’s sampling period so as to obtain 10 values for each group. These 10 values were then averaged where flow is total hepatic plasma flow (ml kg-l min-l). and are expressed as means t SE. Statistical analyses were The livers of these dogs were all net producers of labeled performed using analysis of variance with repeated measures lactate during measurement periods and therefore the net to compare values between the two groups. Statistical calculacontribution of [ 14C]lactate to [ 14C]glucose production was tions were performed using SPSS-PC+ and SAS-PC. A P value zero, and [ 14C]lactate was not considered in the gluconeogenic of 0.05 was used to define statistical significance between calculations. groups. Hepatic glycogen content was determined by multiplying glycogen concentration (mg/g of liver) by liver weight. Rate of glycogen synthesis from glucose (mg kg-l imin-l) was deter- RESULTS mined by dividing total [3H]glucose (dpm) per liver by the avArterial plasma hormone concentrations (Fig. 1). The erage inflowing [3-3H]glucose specific activity (weighted for the arterial and portal contributions), time (min), and body weight plasma cortisol levels rose approximately fivefold in Cortisol (9.5 t 0.4 pg/dl) compared with Saline (1.8 t 0.3 (kg) (20) pg/dl). The plasma insulin level was elevated 70% (19 t 2 vs. 12 t 2 flu/ml, P c 0.01) in the presence of cortisol. glvcogen = synthesis The arterial plasma glucagon levels were not significantly 0.28+ 0.72xtxbodywt different in the two protocols. On the other hand, the l
l
l
l
l
VY
v
0.28-
+ 0.72-
)
where Glchotis the plasma [3H]glucose radioactivity (dpm/ml), and Glccoldis the plasma glucose concentration (mg/ml) in the artery or portal vein. The amount of [14C]glucose deposited in
arterial plasma epinephrine and norepinephrine levels were significantly lower in the presence of cortisol (29 t 10 vs. 75 t 13 pg/ml and 66 t 12 vs. 110 t 7 pg/ml, respectively, P < 0.01).
El22
CHRONIC
HYPERCORTISOLEMIA
AND CARBOHYDRATE
METABOLISM
25
0
B
I
0 =
SALINE CORTISOL
SALINE CORTISOL
p=NS
i
i
L vo.01
p
