Endocrine and Metabolic Effects of Growth Hormone (GH) Compared ...

0013-7227/04/$15.00/0 Printed in U.S.A.

Endocrinology 145(1):205–213 Copyright © 2004 by The Endocrine Society doi: 10.1210/en.2003-1005

Endocrine and Metabolic Effects of Growth Hormone (GH) Compared with GH-Releasing Peptide, Thyrotropin-Releasing Hormone, and Insulin Infusion in a Rabbit Model of Prolonged Critical Illness FRANK WEEKERS, MARINA MICHALAKI, WILLY COOPMANS, ERIC VAN HERCK, JOHANNES D. VELDHUIS, VEERLE M. DARRAS, AND GREET VAN DEN BERGHE Department of Intensive Care Medicine, Burn Unit and Center for Experimental Surgery and Anesthesiology (F.W., M.M., G.V.d.B.), Laboratory for Experimental Medicine and Endocrinology (E.V.H., W.C.), and Laboratory for Comparative Endocrinology (V.M.D.), Catholic University of Leuven, Leuven, Belgium; and Department of Medicine, Mayo Graduate and Medical Schools (J.D.V.), Mayo Clinic, Rochester, Minnesota 55905 Treatment with recombinant human GH (rhGH) increases the mortality of critical illness. We postulated that combined GHreleasing peptide-2 (GHRP-2), TRH, and insulin infusion is a less toxic anabolic strategy through a putative inability to overstimulate the GH axis and a capacity to normalize thyroid hormone concentrations while foregoing excessive hyperglycemia. Burn-injured, parenterally fed, New Zealand White rabbits were randomized to receive 4-d treatment with saline (n ⴝ 8); 60 ␮g/kg䡠h GHRP-2 and 60 ␮g/kg䡠h TRH, iv (n ⴝ 9); or 3.5 mg/kg rhGH, sc (n ⴝ 7). In the GHRP-2ⴙTRH group, insulin was adjusted to maintain blood glucose below 180 mg/dl. Endocrine function and biochemical organ system function markers were studied. Animals were killed for assay of deiodinase activity in snap-frozen liver. Mortality, organ system function, hyperglycemia, and insulin requirement were equal in the three groups. GHRP-2ⴙTRH increased pulsatile rabbit

GH (rGH) and TSH release on d 1. After 4 d, rGH secretion and T4 and T3 concentrations were elevated, with a significant increase in hepatic activity of type 1 deiodinase and a decrease in type 3 deiodinase. Exogenous rhGH suppressed endogenous rGH secretion and increased IGF-I more than GHRP-2ⴙTRH without altering thyroid hormone levels. Unlike GHRP-2ⴙTRH, rhGH down-regulated liver type 3 deiodinase and did not affect type 1 deiodinase. We conclude that in experimentally induced critical illness, GHRP-2ⴙTRH reactivated the GH and TSH axes and altered liver deiodinase activity, driving T4 to T3 conversion. In contrast to the human model, high dose rhGH was not rapidly lethal in this rabbit model. Whether this is explained by lack of rhGH-induced insulin resistance and hyperglycemia remains unclear. (Endocrinology 145: 205–213, 2004)

D

would require very high doses of GH to overcome this resistance (3, 4). This concept was recently invalidated (5–7). There appears to be a biphasic neuroendocrine response to critical illness. During the first few hours and days after an insult, the adaptive neuroendocrine response to critical illness consists primarily of activated anterior pituitary function with peripheral tissue resistance. In the later phase of critical illness, from 7 d onward, pulsatile secretion of anterior pituitary hormones is reduced, and organ function is further impaired. Reduced availability of hypothalamic TRH and the endogenous ligand of the GH-releasing peptide (GHRP) receptor (ghrelin) are postulated factors in reduced TSH and GH secretion. Pulsatile secretion of TSH and GH in critically ill patients can be reamplified by combined infusions of the releasing factors TRH and GHRP-2, which increase circulating levels of T4, T3, IGF-I, and GH-dependent IGF-binding proteins (8). Anabolism is achieved when GHRP-2 and TRH are administered together, but not individually. Nonetheless, both GHRP and rhGH infusions in the critically ill patient exacerbate the stress-induced insulin resistance and hyperglycemia (2, 9). Indirect data suggest that adequate availability of thyroid hormone may be crucial for GHRP-2 or rhGH to evoke anabolic responses (10) and prevention of hyperglycemia may be essential to improve survival (11).

ESPITE FEEDING, prolonged critically ill patients continue to lose lean body mass during their stay in the intensive care unit. This is mainly due to an accelerated loss of protein from skeletal muscle, whereas adipose tissue is paradoxically preserved (1). Erosion of muscle mass induces weakness and thereby prolonged dependency on intensive care. In an attempt to counteract the feeding-resistant hypercatabolism of critical illness, treatment with high doses of recombinant human GH (rhGH) has been tried (2). This strategy, however, did not induce significant whole body anabolism and quite unexpectedly doubled mortality and worsened morbidity. The exact mechanisms explaining this outcome are incompletely understood. The rationale for choosing the very high doses of rhGH (⬎10-fold that used in replacement therapy in GH-deficient adults) was the assumption that critically ill patients remain GH resistant throughout their ICU stay and that an anabolic response Abbreviations: AUC, Area under the curve; D1, type 1 deiodinase; D3, type 3 deiodinase; GHRP, GH-releasing peptide; rGH, rabbit GH; rhGH, recombinant human GH; Lv, distribution volume; TT3, total T3; TT4, total T4. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

205

206

Endocrinology, January 2004, 145(1):205–213

In view of lethal side-effects, studies with rhGH are now impossible in patients. We described an animal model marked by typical endocrine and metabolic characteristics of prolonged critical illness (7). This model allows possible investigation of the primary mechanisms of toxicity of rhGH in critical illness. We hypothesized that the toxicity of high dose rhGH treatment during critical illness is explained by the induction of insulin resistance and hyperglycemia and that impaired anabolism reflects low thyroid hormone availability. According to this postulate, infusion of GHRP-2 in combination with TRH and insulin may provide a superior anabolic strategy. Materials and Methods Study set-up and protocol (Fig. 1) All animals were treated according to the Principles of Laboratory Animal Care formulated by the U.S. National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the NIH. The study protocol was approved by the University of Leuven ethical review board for animal research. The model has been described in detail previously (7). In brief, male New Zealand White rabbits were purchased from a local rabbitery and housed individually. Animals had free access to water, were fed once daily, and were exposed to artificial light for 14 h/d. After weighing and cannulation of both carotic artery and jugular vein under general anesthesia [30 mg/kg ketamine 1000, im (Merial, Lyon, France), and 0.15 ml/kg medetomidine, im (Orion Corp., Ospoo, Finland)], animals were put into a homemade jacket to secure the positions of the catheters. To assure the patency of the catheters, all lines were filled with heparin (5000 IU/ml; Heparine, Rhoˆ ne Poulenc Rorer, Brussels, Belgium). Two days later, animals were reweighed, and blood samples for biochemical and blood gas analyses were taken via the arterial line. Then, after shaving both flanks and under deep general anesthesia supplemented with a local paravertebral block, a 15% full thickness burn injury was imposed (Fig. 1). After a 1-h recovery period with supplemented oxygen, the animals returned to their cages, and fluid resuscitation was started overnight with a continuous infusion of Ringer lactate at six drops per minute (⫾12 ml/h) through a volumetric infusion pomp. To assure free movement of the rabbits without twisting of the catheter and infusion lines, a swivel device was incorporated. In the evening, a supplemental dose of a major analgesic (Dipidolor, Janssen-Cilag, Beerse, Belgium) was given. The next morning at 0800 h ⫾ 30 min, parenteral nutrition was initiated in all rabbits at four drops per minute (⫾12 ml/h). Blood glucose levels were kept within narrow limits (⬍180 mg/dl) by frequent blood glucose monitoring and titration of insulin infusion (100 IU/ml Actrapid, Novo Nordisk, Bagsvaerd, Denmark; via an SE200B infusion pump, Vial Medical, Brezins, France) when necessary. All iv infusions were weighed before and after administration, allowing accurate determination of the infused quantity. Parenteral nutrition was prepared daily in the hospital pharmacy under laminar air-flow conditions. The infusion bags contained 103 ml 20% glucose, 37.5 ml Aminoplasmal L10 (Braun, Melsungen, Germany), 34 ml 20% Intralipid (Fresenius Kabi, Stockholm, Sweden), and 150 ml sterile water. Thus, bags with 325 ml

FIG. 1. Flow chart of study design. Animals were randomized into three interventional groups 4 d after induction of the injury. Interventions were continued for 4 d. N, Number of animals in each cohort at randomization.

Weekers et al. • rhGH vs. GHRP-2⫹TRH in Critically Ill Rabbits

solution contained 150 kcal nonprotein calories and 0.6 g nitrogen. Of all calories, 56% were delivered as carbohydrates, and 44% as fat. Protein intake equaled 1.25 g amino acids/kg䡠d. No additional vitamins or trace elements were added. Total parenteral nutrition was continuously administered. Animals had free access to water, but oral food intake was denied. Parenteral nutrition bags were changed daily. On d 1 and 3 blood was sampled for later analysis. On the morning of d 4, surviving rabbits were randomly (by sealed envelopes) allocated to one of the following treatment groups. Group 1 (saline group) received a 4-d continuous infusion of 0.9% NaCl. Goup 2 (rhGH group) was injected sc on 4 consecutive d with 3.5 mg/kg䡠d of rhGH (Genotropin, Amersham Pharmacia Biotech, Stockholm, Sweden) in two divided injections (one at 0800 h ⫾ 30 min and one at 2000 h ⫾ 30 min), and insulin was kept at the same infusion rate as on the morning of d 4 regardless of glucose levels during further treatment. Group 3 (from hereon referred to as the GHRP-2⫹TRH group) received an initial bolus of 60 ␮g/kg GHRP-2 (Kaken Pharmaceutical Co. Ltd., Tokyo, Japan) plus 60 ␮g/kg TRH (200 ␮g/ml 0.9%NaCl; Ferring, Kiel, Germany), followed by a continuous infusion of 60 ␮g/kg䡠h GHRP-2 and 60 ␮g/ kg䡠h TRH. Infusion syringes and infusion lines were protected from the light. The dose of GHRP-2 and TRH was defined in a previous study (7). In groups 1 and 3, insulin was titrated to maintain blood glucose levels below 180 mg/dl (9.9 mm/liter) throughout the study. On d 4, between 0830 h ⫾ 30 min and 1530 h ⫾ 30 min, 1 ml blood was sampled every 15 min for determination of plasma rabbit GH (rGH) and the TSH concentration-time series over 7 h. Blood was drawn from the arterial line using a Vamp system (Baxter Healthcare Corp., Irvine, CA), which allows undiluted blood sampling without undue blood loss. Blood was collected into glass tubes containing 25 U heparin and then stored at 4 C until centrifugation. All samples were centrifuged, and plasma was stored at ⫺70 C until assay. In all groups, study drugs were only injected after two baseline samples (i.e. the first two samples of the profiles) had been withdrawn. Blood for blood gas and biochemical analyses was taken before and after the profile. On d 5, 6, and 7, parenteral nutrition bags were changed daily, and additional blood samples were taken on d 5. On d 8, again during TPN infusion and continuation of the study drugs, another 7-h sampling time series (sampling every 15 min) was performed to evaluate rGH and TSH secretion. After the last blood sample, animals were weighed and euthanized with sodium-pentobarbital (60 mg/ml Nembutal, Sanofi-Winthrop, New York, NY). Organs were harvested, and heart, lungs, liver, spleen, and kidney were weighed. Subsequently, tissue samples were taken from left and right ventricles, lung, liver, spleen, kidney, ileum, muscle, intact skin, and burned skin. In each of these tissue samples, a part was used to determine dry weight and water content. Water content was determined by weighing the samples before and after 24 h of lyophilization (Alfa I-5 Chriss, Aichach, Germany). Part of the liver was snap-frozen in liquid nitrogen-chilled isopentane (Sigma-Aldrich Corp., Bornem, Belgium) and stored at ⫺70 C until further analysis of deiodinase activity. All blood gas analyses were performed immediately after sampling. Blood for biochemical analyses was sampled in lithium citrate tubes (BD Biosciences, Temse, Belgium), immediately centrifuged, and then stored at ⫺70 C until assay.


Endocrinology, January 2004, 145(1):205–213 207

Assays

Statistics

Arterial blood was analyzed on an ABL 700 analyzer (Radiometer Medical A/S, Copenhagen, Denmark) to quantitate pH, pO2, pCO2, hemoglobin, lactate, and glucose. Furthermore plasma concentrations of sodium, potassium, chloride, bicarbonate, ureum, creatinine, total calcium, phosphorus, total protein content, alkaline phosphatase, transaminases (aspartate aminotransferase and alanine aminotransferase), ␥ glutamyl transpeptidase, total bilirubin, uric acid, creatine kinase, lactate dehydrogenase, amylase, lipase, cholesterol, and triglycerides were determined using the Roche Modular with Roche reagents (Roche, Mannheim, Germany). Plasma rGH concentrations were measured with a specific RIA (reagents provided by Dr. A. Parlow, National Pituitary Agency). The detection limit was 1 ␮g/liter, and the within-assay coefficient of variation was 2.3%. All samples were analyzed in duplicate within a single assay run. All samples contained detectable plasma concentrations. Plasma IGF-I concentrations were measured in duplicate by RIA, using a slightly modified version of that described by Verhaeghe et al. (12). After acidification of the plasma samples with formic acid, binding proteins were separated by acid gel filtration on an Econo-Pac column (Bio-Rad Laboratories, Richmond, CA). Des-IGF-I was used as tracer to avoid binding to residual small IGF-binding proteins. The intraassay coefficient of variation was 4.6%. The detection limit was 20 ␮g/liter. Plasma concentrations of rabbit TSH were measured by a specific RIA (reagents provided by Dr. A. Parlow, National Pituitary Agency). The detection limit was 1.2 mU/liter. All samples were analyzed in duplicate within a single assay run. For samples with undetectable levels, a value representing half the detection limit was entered. Total plasma T4 and T3 concentrations were determined by a specific RIA (Immunotech, Marseilles, France). The sensitivity for T3 was 0.2 nmol/liter, and that for T4 was 6 nmol/liter. All samples were assayed in duplicate. Plasma insulin concentrations were determined by immunoradiometric assay (Medgenix INS-IRMA, BioSource, Fleurus, Belgium).

Comparisons of normally distributed data among the three groups were made by ANOVA. Within-group changes were analyzed by t test with Bonferroni correction for multiple comparisons when indicated. Nonnormally distributed data were analyzed by Friedman test or Kruskal-Wallis test as appropriate. All data are expressed as the mean ⫾ sd unless specified otherwise. P ⬍ 0.05 was considered significant.

Hepatic deiodinase activity For determination of hepatic type 1 (D1) and type 3 (D3) iodothyronine deiodinase activities, liver microsomal fractions were prepared and assayed as described by Darras et al. (13). Final incubations were in a total volume of 200 ␮l. For D1 activity, final incubation mixtures contained 1 ␮m rT3, 50,000 cpm/tube [3⬘,5⬘-125I]rT3 (labeled using 17 Ci/mg carrier-free 125I from NEN Life Science Products, Boston, MA), 250 ␮g/ml microsomal liver protein, 2 mm EDTA, and 5 mm dithiothreitol and were incubated for 30 min at 37 C. For D3 activity, incubations contained 1 nm T3, 200,000 cpm/tube [3⬘-125I]T3 (labeled as described above), 250 ␮g/ml liver microsomal protein, 1 ␮m rT3, 0,1 mm propylthiouracil, 2 mm EDTA, and 50 mm dithiothreitol and were incubated for 120 min at 37 C.

Analytical procedures The time series of sequential plasma rGH concentrations measured every 15 min for 7 h on d 4 and 8 was transformed into pituitary secretion profiles using multiple parameter deconvolution analysis (14). This method is designed to compute hormonal half-life and the number and mass of underlying pituitary secretory bursts and to estimate basal secretion. The following parameters were calculated: half-life, temporal position of the secretory bursts, basal rGH secretion rate (amount of rGH released from pituitary to achieve serum concentrations approximating the mean of the lowest 5% of all values observed [micrograms per liter of distribution volume (Lv) per minute]), mass [area of the resolved secretion burst (micrograms per Lv)], mean pulsatile production [product of the number of secretory bursts and the mean secretory burst mass over the time interval considered (micrograms per Lv over 7 h)], and total rGH secretion (sum of basal and pulsatile released rGH) were calculated. Deconvolution analysis of the plasma TSH concentration profiles was not feasible because too many samples had undetectable TSH concentrations. Thus, TSH concentrations are given as the mean and the area under the curve (AUC) calculated by trapezoidal rule.

Results

For this experiment, we used 44 rabbits. Twenty-four (54%) rabbits survived the instrumentation and acute burn injury phase and could be randomized into 1 of the treatment groups. In the rhGH and GHRP-2⫹TRH group, 5 rabbits survived the 8-d experimental period, whereas in the saline group 6 rabbits were alive on d 8. Body weight

The mean body weight of the rabbits at the start of the study was 3004 ⫾ 266 g in the saline group (n ⫽ 8), 3050 ⫾ 157 g in the GHRP-2⫹TRH group (n ⫽ 9), and 3136 ⫾ 211 g in the rhGH group (n ⫽ 7). All rabbits lost weight 4 d after burn injury: to 2828 ⫾ 196 g in the saline group (P ⫽ 0.04 vs. baseline), to 2891 ⫾ 134 g in the GHRP-2⫹TRH group (P ⫽ 0.01 vs. baseline), and to 2921 ⫾ 214 g in the rhGH group (P ⫽ 0.007 vs. baseline). Loss of body weight was similar in the three groups. Thereafter there was no further change in body weight toward d 8. Equal amounts of parenteral feeding were administered in the three groups. GH axis

Mean baseline plasma rGH concentrations were similar in the three study groups (4.0 ⫾ 2.8 ␮g/liter in the saline group, 4.7 ⫾ 4.2 ␮g/liter in the GHRP-2⫹TRH group, and 3.16 ⫾ 2.03 ␮g/liter in the rhGH group; P ⫽ NS). Within 1 d of treatment, exogenous high doses of rhGH suppressed and GHRP-2⫹TRH infusion increased pulsatile rGH secretion compared with saline (Figs. 2 and 3 and Table 1). Between d 4 and 8, pulsatile rGH secretion decreased further in the saline and rhGH groups, whereas it remained elevated in the GHRP-2⫹TRH group. The rhGH-induced suppression was mainly caused by reduced rGH burst mass. The initial GHRP-2⫹TRH-induced amplification of rGH secretion was caused by an increase in rGH burst mass, whereas elevated basal secretion accounted for most of the increase on d 8 (Table 1). At baseline (and before injury), plasma IGF-I concentrations were equal in the three study groups (122 ⫾ 53 ␮g/liter in the saline group, 119 ⫾ 35 ␮g/liter in the GHRP-2⫹TRH group, and 122 ⫾ 32 ␮g/liter in the rhGH group; P ⫽ NS). The plasma IGF-I concentration decreased progressively during the first 3 d postinjury (Fig. 4) in all three groups. The plasma IGF-I concentration remained stable between d 3 and 8 in the saline group, whereas rhGH administration significantly increased and GHRP-2⫹TRH tended to increase IGF-I by d 8. Serum IGF-I levels on d 8 were significantly higher in the rhGH group compared with the GHRP-2⫹TRH group (Fig. 4).

208



Thyroid axis

Mean baseline TSH concentrations were identical in the three groups (2.77 ⫾ 0.66 mU/liter in the saline group, 2.98 ⫾ 0.56 mU/liter in the GHRP-2⫹TRH group, and 2.71 ⫾ 0.51 mU/liter in the rhGH group). Plasma TSH concentrations (AUC) on d 4 were elevated in the GHRP-2⫹TRH group compared with the saline and rhGH groups (P ⫽ 0.007 vs. rhGH group and P ⫽ 0.003 vs. saline group), whereas on d 8 TSH levels became indistinguishable in the three groups (Figs. 5 and 6). At baseline, plasma total T4 and T3 (TT4 and TT3) concentrations were equal in the three study groups. TT4 and TT3 significantly decreased from baseline to d 3 postinjury (Fig. 7) in all three groups. In the saline and rhGH groups, both TT4 and TT3 remained low between d 3 and 8, whereas in the

FIG. 2. Representative individual rGH concentration-time series (based on blood sampling every 15 min for 7 h) on d 4 (first day of intervention) and d 8 (fourth day) in groups assigned to saline (upper panel), GHRP-2 ⫹ TRH (middle panel), and rhGH (lower panel).

GHRP-2⫹TRH group, there was a progressive increase in TT4 and TT3. On d 8, hepatic D1 activity was significantly higher in the GHRP-2⫹TRH group than in the saline and rhGH-treated groups. Both GHRP-2⫹TRH and rhGH significantly suppressed hepatic D3 activity compared with saline. Glucose control

At baseline, the mean blood glucose concentration was 156 ⫾ 38 mg/dl (Fig. 8, upper left panel). On d 3, the insulin titration protocol to keep blood glucose levels below 180 mg/dl resulted in a mean blood glucose concentration of 130 ⫾ 19 mg/dl (P ⫽ 0.02 vs. baseline and no difference

FIG. 3. The AUC of rGH concentrations on d 4 (A) and d 8 (B) in animals given saline (f), GHRP-2⫹TRH (o), and rhGH (䡺). Day 4 is the first day, and d 8 is the fourth day of intervention. Data are the mean ⫾ SEM.

TABLE 1. Deconvolution-derived characteristics of rGH secretory patterns in saline-, GHRP-2⫹TRH-, and rhGH-treated male New Zealand White rabbits on d 4 (first day of treatment) and 8 (fourth day of treatment) d4

d8

Parameter

Saline (n ⫽ 8)

GHRP-2⫹TRH (n ⫽ 9)

rhGH (n ⫽ 6)

Saline (n ⫽ 6)

GHRP-2⫹TRH (n ⫽ 5)

rhGH (n ⫽ 5)

GH half-life (min) GH pulse frequency (no. per 7 h) GH inter-pulse interval (min) GH burst mass (␮g/Lv) Pulsatile GH secretion (␮g/Lv/7 h) Basal GH secretion (␮g/Lv/7 h) Total secretion (␮g/Lv/7 h)

14.6 ⫾ 3.3 4.8 ⫾ 0.9 63.8 ⫾ 25 15.3 ⫾ 9.2 68 ⫾ 30 238 ⫾ 243 306 ⫾ 231

14.1 ⫾ 4.2 5.4 ⫾ 1.8 51.7 ⫾ 18.3 28.4 ⫾ 24.1 138 ⫾ 80c 271 ⫾ 221 410 ⫾ 250

11.8 ⫾ 0.7 5.1 ⫾ 0.7 61.9 ⫾ 11.9 5.1 ⫾ 2.5c 26 ⫾ 13c 155 ⫾ 33 181 ⫾ 29

18.4 ⫾ 2.2a 5.5 ⫾ 1.2 67.7 ⫾ 14.8 8.5 ⫾ 4.2a 45 ⫾ 20a 65.4 ⫾ 41 111 ⫾ 51a

10.9 ⫾ 1.6b 4.4 ⫾ 1.3 76 ⫾ 16.7 26.6 ⫾ 16.4c 130 ⫾ 97 353 ⫾ 76.5b 484 ⫾ 146b

16 ⫾ 3a 5.6 ⫾ 1.6 66.4 ⫾ 34.9 1.6 ⫾ 0.5a,c 9 ⫾ 4a,c 94 ⫾ 49 103 ⫾ 52a

Samples were withdrawn every 15 min for 7 consecutive hours. Values represent mean ⫾ a P ⬍ 0.05 between d 4 and d 8. b P ⬍ 0.01 vs. saline group. c P ⬍ 0.05.

SD.



FIG. 4. Serum IGF-I concentrations on d 0 (preburn), 3, 4, and 8 of study. u, Pooled IGF-I concentrations before randomization; f, saline; o, GHRP2⫹TRH; 䡺, rhGH groups. *, P ⫽ 0.04; †, P ⫽ 0.07 (vs. d 3).

FIG. 5. Serum TSH concentration profiles during infusion of saline (F), GHRP-2⫹TRH (〫), and rhGH (E) after 1 (d 4) or 4 (d 8) d of intervention. Data are the mean ⫾ SEM. P ⬍ 0.007 denotes the difference between the effects of GHRP-2⫹TRH compared with those of saline and rhGH.

among the three groups). At the start of treatment with either rhGH or GHRP-2⫹TRH, blood glucose levels transiently increased on d 5 compared with the saline group, with equal blood glucose levels on d 8. The total amount of insulin infused in the three groups was equal (Fig. 8, upper right panel). Plasma insulin concentrations increased immediately after trauma equally in the three study groups (Fig. 8, lower panel), with a trend for a further increase on d 5 (P ⫽ 0.06). Biochemical markers of organ system function

As described previously (7), our model is characterized by anemia, moderate hyponatremia, and transient acidosis with hyperventilation (data not shown). Baseline plasma concentrations of potassium (3.9 ⫾ 0.6 mEq/liter), chloride (107 ⫾ 4.6 mmol/liter), lactate (2.4 ⫾ 2.2 mmol/liter), and bicarbonate (19.5 ⫾ 3.1 mmol/liter) remained constant throughout the experiment. Plasma urea (37 ⫾ 9 mg/dl) and creatinine (1.0 ⫾ 0.2 mg/dl) did not detectably change during the course of the experiment. Baseline plasma levels of aspartate aminotransferase (19 ⫾ 10 U/liter) and alanine aminotransferase (32 ⫾ 8 U/liter) transiently increased after burn injury on d 3 equally in the three groups and returned to baseline on d 8. Plasma bilirubin was below the detection limit of the assay in all animals (⬍0.1 mg/dl). Baseline plasma levels of alkaline phosphatase (124 ⫾ 42 U/liter) tended to progressively decline toward d 8 in the saline (68 ⫾ 79 U/liter) and GHRP-2⫹TRH (46 ⫾ 16 U/liter) groups (both P ⫽ 0.06 vs. d 0), but not in the rhGH group. The plasma amylase concentration was 277 ⫾ 56 U/liter at baseline and rose to 398 ⫾ 148 U/liter on d 3 (P ⫽

0.004). On d 8, the plasma amylase concentration remained higher in the GHRP-2⫹TRH group (316 ⫾ 29 U/liter; P ⫽ 0.01 vs. saline), whereas it normalized in the two other groups. Levels of ␥ glutamyl transpeptidase, lipase, and lactate dehydrogenase remained unaltered. Baseline plasma creatine kinase (2248 ⫾ 3133 IU/liter) levels were similar in the three groups and transiently increased on d 3 postburn (3990 ⫾ 3047 IU/liter; P ⫽ 0.0029 vs. d 0), returning to baseline on d 5. The mean plasma triglyceride concentration was 65 ⫾ 29 mg/dl at baseline. After injury and under parenteral nutrition, all animals developed hypertriglyceridemia (135 ⫾ 76 mg/dl in the saline group, 363 ⫾ 507 mg/dl in the GHRP2⫹TRH group, and 190 ⫾ 57 mg/dl in the rhGH group), which persisted throughout the experiment. There was no difference among the three groups. Organ weight

There was no significant difference among the three groups in wet weights of heart, lungs, and spleen. The liver tended to be heavier in rhGH-treated animals compared with the saline group (P ⫽ 0.07). Water contents of left and right ventricle, lungs, liver, spleen, small intestine, kidney, muscle, and burned and normal skin were identical in the three groups. Discussion

The present experimental model of prolonged critical illness shows that combined infusion of GHRP-2 and TRH

210


reverses down-regulation of hepatic D1 activity and upregulation of D3 activity and stimulates pulsatile GH and total TSH secretion. Unexpectedly, activation of the endogenous GH and TSH axes did not enhance markers of organ function or reduce lethality compared with fixed high dose rhGH or saline. Hyperglycemia and insulin requirements were also comparable in the three interventions. We hypothesized that the infusion of GHRP-2 and TRH combined with insulin-titrated maintenance of normoglycemia would be less toxic and more favorable to survival than

FIG. 6. Integrated (AUC) serum TSH concentrations after thermal injury and interventions with saline (f), GHRP-2⫹TRH (o), and rhGH (䡺). Upper panel, Data on d 4; lower panel, data on d 8.

FIG. 7. Time course of serum total T4 (upper left panel) and total T3 (lower left panel) concentrations in the three groups at baseline (d 0), d 3 postburn (before intervention on d 4), d 5, and d 8 (after 4 d of intervention). The right panel depicts hepatic D1 activity (picomoles per milligram per minute) and D3 activity (femtomoles per milligram per minute) on d 8. u, Pooled data obtained on d 0 and 3; f, response to saline; o, response to GHRP-2⫹TRH; 䡺, response to rhGH. Data are the mean ⫾ SEM. *, P ⫽ 0.02 vs. d 3; **, P ⫽ 0.001 vs. d 3; †, P ⬍ 0.02 vs. d 0.


administration of high doses of rhGH in an animal model of prolonged critical illness. Indeed, exacerbation of insulin resistance and hyperglycemia by high doses of rhGH and reduced systemic availability of thyroid hormones in critical illness are metabolic alterations that may have contributed to increased morbidity and mortality in a clinical study (2). The high dose rhGH and GH secretion driven by GHRP-2 and TRH infusion did not substantially accentuate burn injury-induced insulin resistance and hyperglycemia. rhGH elevated circulating IGF-I concentrations and suppressed endogenous pulsatile rGH secretion, demonstrating efficacy in the rabbit model. It seems, therefore, that rabbits have reduced sensitivity to insulin antagonism by rhGH. This conjecture is in line with the results of a study in transgenic rabbits overexpressing the bovine GH gene (15). Whether the absence of an adverse rhGH effect on stress-induced insulin resistance and hyperglycemia explains the apparent absence of increased mortality is speculative. However, strict glycemic control with insulin in critically ill patients limits excessive inflammation, sepsis, organ failure, and mortality (11), whereas high dose rhGH treatment predicts insulin resistance, hyperglycemia, inflammation, sepsis, multiple organ failure, and doubling of mortality (2). The present outcomes in the rabbit contrast with the findings of Angelin et al. in rats (16). Continuous infusion of rhGH, followed by endotoxin injection in the latter model, heightened the risk of cardiovascular shock, hypertriglyceridemia, and biochemical signs of multiple organ failure. Possible distinctions of relevance in the two studies include interventional sequence (17, 18), mode of rhGH administration (19 –21), type of injury imposed, feeding status, and species. Although of major importance, the precise effect of rhGH on immune function is complex. In parenterally fed trauma patients, rhGH (0.15 mg/kg䡠d) for 7 d increased TNF-␣ and IL-6 (22), whereas in patients undergoing abdominal aneurysm surgery, preoperative injection of rhGH for 6 d did not alter IL-1a receptor antagonist, IL-6, or acute phase protein concentrations (23). In vitro, pharmacological concentrations of rhGH affected the release of cytokines (24). Finally, endotoxin evokes profound, but short-lasting, cyto-



FIG. 8. Time course of whole blood glucose concentrations, insulin requirements expressed per day survived, and plasma insulin concentrations in the three study groups. F and f, Saline group; 〫 and o, GHRP-2⫹TRH group; E and 䡺, rhGH group. Data are the mean ⫾ SEM. *, P ⬍ 0.01 vs. saline.

kine release, unlike sustained cytokine production in illness in humans or evolving burn injury (25). The current study outcome is consistent with human (26) and other animal (7) studies in which prolonged critical illness depresses hypothalamic-pituitary secretion. In the control intervention, rabbits with major burns exhibited a fall in pulsatile rGH secretion and serum IGF-I concentrations between d 4 and d 8 of the experiment. As expected, sc injection of rhGH suppressed endogenous rGH secretion on d 4 and even more so on d 8. Mechanistically, negative feedback was attributable to a reduction in the amount of rGH released per burst. These data indicate that feedback inhibition is detectable during severe illness in the lagomorph, akin to inferences in the human (9, 27). Infusion of GHRP-2 and TRH stimulated pulsatile GH secretion from d 4 to 8 by the converse mechanism of augmenting GH pulse size and elevating nonpulsatile rGH release. The foregoing dynamics also operate in critically ill patients (27, 28). Reactivation of pulsatile rGH secretion by constant infusion of GHRP-2 and TRH extends the idea that blunted GH secretion in prolonged illness is not due to impaired GH synthesis by somatotropes but, rather, to deficient stimulation by hypothalamic and/or peripheral releasing factors. The significant IGF-I response to rhGH injection and the analogous trend after infusion of GHRP-2 and TRH indicates that target tissues such as the liver regain responsiveness to GH drive in prolonged critical illness, as distinguished from acute critical illness (9, 29). Pulsatile rGH secretion remained elevated after 4 d of GHRP-2⫹TRH infusion. The latter outcome may be explained by the limited initial IGF-I response with correspondingly lesser feedback inhibition (30, 31). Alternatively, continued hypothalamo-pituitary responsiveness may be relevant as inferred under low dose continuous GHRP stimulation (32). The constellation of suppressed pulsatile TSH secretion

with low serum T4 and T3 and elevated rT3 concentrations characterizes the low T3 syndrome of prolonged critical illness (5). Suppressed TSH secretion contributes to low T4 production and altered peripheral thyroid hormone metabolism depletes T3 availability. In humans, peripheral thyroid hormone metabolism is mediated by three iodothyronine deiodinases: D1, D2, and D3 (33). D1 is present in liver, kidney, and thyroid and plays a key role in the conversion of T4 to T3 and in the catabolism of rT3. D2 is present in brain, pituitary, thyroid, and skeletal muscle and also converts T4 by outer ring deiodination to T3. D3 is present in brain, skin, placenta, pregnant uterus, and fetal tissues and inactivates T4 and T3 by inner ring deiodination to rT3 and 3,3⬘-diiodothyronine, respectively. We previously showed that the reduced T3 and elevated rT3 concentrations in prolonged critically ill patients are in part due to decreased D1 activity and, hence, suppressed peripheral conversion of T4 to T3 (34). In contrast, D3 activity was higher in critically ill patients, thereby suppressing T3 and augmenting rT3 availability. Some alterations in deiodinase activity may be cytokine induced (35, 36). In the rabbit model of critical illness with signs of the low T3 syndrome, similar changes in deiodinase activity could be inferred, although we did not include healthy controls in this study. Coinfusion of GHRP-2 and TRH activated the thyroid axis by stimulating TSH secretion, elevating serum total T4 and total T3 concentrations. TSH-dependent secretion of T4 and T3 may have contributed to the latter responses. On the other hand, combined GHRP-2 and TRH stimulation, respectively, augmented and depressed the catalytic activity of hepatic D1 and D3. The observed changes in deiodinase activities may have contributed to enhanced peripheral T4 to T3 conversion in the sick rabbits. Suppression of D3, but not stimulation of D1, also occurred in response to administration of rhGH. This may point to a role for GH, either endo-

212


genously released or exogenously administered, in mediating the effect on D3 (37). Such an effect of GH on D3 activity, also shown in healthy chickens (38), would explain the increase in T3 and decrease in rT3 previously observed after treatment with 4 mg/d rhGH for 2 wk in healthy humans (39). Reactivation of hepatic D1 activity, on the contrary, was absent with exogenous GH and thus could be due to GHRP-2, TRH, or the rise in T4 and T3 concentrations (33, 40). The present data do not distinguish among these options. Although an effect of insulin on liver deiodinase activity has been suggested (41), our observation cannot be explained by such an effect, as blood glucose and insulin concentrations were identical in all intervention groups. In summary, in a rabbit model of critical illness, combined infusion of GHRP-2 and TRH stimulates GH and TSH secretion and elevates T4 and T3 concentrations inferentially by increasing T4 secretion and favoring the hepatic interconversion of T4 to T3. In contradistinction to clinical data, a high dose of rhGH in the lagomorph did not increase short-term mortality after a major burn. Whether the absence of rhGHinduced insulin resistance and hyperglycemia in this species explains the latter outcome remains to be elucidated.


10.

11. 12.

13.

14.

15. 16. 17.

Acknowledgments We acknowledge the generosity of Dr. C. Y. Bowers in providing GHRP-2 for these experiments. We appreciate the technical support of Mrs. Veerle Leunens and Mr. A. Berghen (Center for Experimental Surgery and Anesthesiology), Mrs. Lut Noterdaeme, Mr. Willy Van Ham, and Mrs. Francine Voets (Laboratory of Comparative Endocrinology), Mr. J. Hellers, (Baxter, Lessines, Belgium), Mrs. F. Decraecker (Pharmacia, Brussels, Belgium), Mr. Deville´ (Hospitera, Brussels, Belgium), and Kandace Bradford (Mayo Clinic, Rochester, MN). Received August 5, 2003. Accepted September 30, 2003. Address all correspondence and requests for reprints to: Frank Weekers, M.D., Department Intensive Care Medicine, University of Leuven, B-3000 Leuven, Belgium. E-mail: [email protected]. This work was supported by Belgian Fund for Scientific Research (Grants G.3C05.95N, .0144.00, and .0278.03), Research Fund of the University of Leuven (Grants OT/99/32 and OT/03/56), a research grant from Pharmacia & Upjohn (Uppsala, Sweden), Futura Award (VKG) 1999, and National Center for Research Resources (Rockville, MD; Grant RR-00585).

References 1. Hart D, Wolf S, Herndon D, Chinkes D, Lal S, Obeng M, Beauford R, Mlcak R 2002 Energy expenditure and caloric balance after burn: increased feeding leads to fat rather than lean mass accretion. Ann Surg 235:152–161 2. Takala J, Ruokonen E, Webster N, Nielsen MS, Zandstra DF, Vondelinckx G, Hinds CJ 1999 Increased mortality associated with growth hormone treatment in critically ill adults. N Eng J Med 341:785–792 3. Dahn M, Lange M, Jacobs L 1988 Insulin-like growth factor I production is inhibited in human sepsis. Arch Surg 123:1409 –1414 4. Ross RJ 1993 The role of insulin, growth hormone and IGF-I as anabolic agents in the critically ill. Intensive Care Med 19:554 –557 5. Van den Berghe G 2002 Dynamic neuroendocrine responses to critical illness. Front Neuroendocrinol 23:370 –391 6. Van den Berghe G 2000 Novel insights into the neuroendocrinology of critical illness. Eur J Endocrinol 143:1–13 7. Weekers F, Van Herck E, Coopmans W, Michalaki M, Bowers CY, Veldhuis JD, Van den Berghe G 2002 A novel in vivo rabbit model of hypercatabolic critical illness reveals a bi-phasic neuroendocrine stress response. Endocrinology 143:764 –774 8. Van den Berghe G, de Zegher F, Bowers CY, Wouters P, Muller P, Soetens F, Vlasselaers D, Schetz M, Verwaest C., Lauwers P, Bouillon R 1996 Pituitary responsiveness to GH-releasing hormone, GH-releasing peptide-2 and thyreotropin-releasing hormone in critical illness. Clin Endocrinol (Oxf) 45:341– 351 9. Van den Berghe G, Wouters P, Weekers F, Mohan S, Baxter R, Veldhuis J,

18. 19. 20.

21. 22.

23.

24. 25. 26. 27.

28.

29.

30. 31.

Bowers CY, Bouillon R 1999 Reactivation of pituitary hormone release and metabolic improvement by infusion of growth hormone-releasing peptide and thyrotropin-releasing hormone in patients with protracted critical illness. J Clin Endocrinol Metab 84:1311–1323 Van den Berghe G, Baxter R, Weekers F, Wouters P, Bowers C, Iranmanesh A, Veldhuis J, Bouillon R 2002 The combined administration of GH-releasing peptide-2 (GHRP-2), TRH and GnRH to men with prolonged critical illness evokes superior endocrine and metabolic effects compared to treatment with GHRP-2 alone. Clin Endocrinol (Oxf) 56:655–569 Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D, Ferdinande P, Lauwers P, Bouillon R 2001 Intensive insulin therapy in critically ill patients. N Engl J Med 345:1359 –1367 Verhaeghe J, Suiker AM, Visser WJ, Van-Herck E, Van-Bree R, Bouillon R 1992 The effects of systemic insulin, insulin-like growth factor-I and growth hormone on bone growth and turnover in spontaneously diabetic BB rats. J Endocrinol 134:485– 492 Darras V, Visser T, Berghman L, Kuhn E 1992 Ontogeny of type I and III deiodinase activities in embryonic and posthatch chicks: relationship with changes in plasma triiodothyronine and growth hormone levels. Comp Biochem Physiol 103A:131–136 Veldhuis JD, Covilson ML, Johnson ML 1987 The pituitary gland secretes in bursts : appraising the nature of glandular secretary impulses by simultanuous multiple parameter deconvolution of plasma hormone concentrations. Proc Natl Acad Sci USA 84:7686 –7690 Costa C, Solanes G, Visa J, Bosch F 1998 Transgenic rabbits overexpressing growth hormone develop acromegaly and diabetes mellitus. FASEB J 12:1455– 1460 Liao W, Rudling M, Angelin B 1995 Growth hormone potentiates the in vivo biological activities of endotoxin in the rat. Eur J Clin Invest 26:254 –258 Warwick-Davies J, Lowrie D, Cole P 1995 Growth hormone activation of human monocytes for superoxide production but not tumor necrosis factor production, cell adherence, or action against Mycobacterium tuberculosis. Infect Immun 63:4312– 4316 Edwards C, Arkins S, Yunger L, Blum A, Dantzer R, Kelley K 1992 The macrophage activating properties of growth hormone. Cell Mol Neurobiol 12:499 –510 Jaffe C, Turgeon D, Lown K, Demott-Friberg R, Watkins P 2002 Growth hormone secretion pattern is an independent regulator of growth hormone actions in humans. Am J Physiol 283:E1008 –E1015 Roelfsema V, Thomas G, Lin H, Breier B, Maxwell L, Oliver M, Heineman E, Clark R, Gluckman P 2001 the metabolic effects of endotoxin are differentially affected by the pattern of growth hormone administration in the rat. J Endocrinol 171:173–181 Shapiro B, Agrawal A, Pampori N 1995 gender differences in drug metabolism regulated by growth hormone. Int J Biochem Cell Biol 27:9 –20 Jeevanandam M, Begay C, Shahbazian L, Petersen S 2000 Altered plasma cytokines and total glutathione levels in parenterally fed critically ill trauma patients with adjuvant recombinant human growth hormone. Crit Care Med 28:324 –329 Mealy K, Barry M, O’Mahony L, Sheehan S, Burke P, McCormack C, Whitehead A, Bouchier D 1998 Effects of human reconbinant growth hormone on inflammatory responses in patients undergoing abdominal aortic aneurysm repair. Intensive Care Med 24:128 –131 Kappel M, Hansen M, Diamant M, Pedersen B 1994 In vitro effects of human growth hormone on the proliferative responses and cytokine production of blood mononuclear cells. Horm Metab Res 26:612– 614 Demling R, Lalonde C 1990 Systemic lipid peroxidation and inflammation induced by thermal injury persists into the post-resuscitation period. J Trauma 30:69 –74 Van den Berghe G, de Zegher F, Bouillon F 1998 Acute and prolonged critical illness as different neuroendocrine paradigms. J Clin Endocrinol Metab 83: 1827–1834 Van den Berghe G, de Zegher F, Baxter R, Veldhuis J, Wouters P, Schetz M, Verwaest C, Van der Vorst E, Lauwers P, Bouillon R, Bowers C 1998 Neuroendocrinology of prolonged critical illness: effects of exogenous thyrotropinreleasing hormone and its combination with growth hormone secretagogues. J Clin Endocrinol Metab 83:309 –319 Van den Berghe G, de Zegher F, Veldhuis J, Wouters P, Awouters M, Verbruggen W, Schetz M, Verwaest C, Lauwers P, Bouillon R, Bowers C 1997 The somatotropic axis in critical illness: effect of continuous growth hormone (GH) releasing hormone and GH-releasing peptide-2 infusion. J Clin Endocrinol Metab 82:590 –599 Van den Berghe G, de Zegher F, Veldhuis J, Wouters P, Awouters M, Verbruggen W, Schetz M, Verwaest C, Lauwers P, Bouillon R, Bowers C 1997 The somatotropic axis in critical illness: effect of continuous growth hormone (GH) releasing hormone and GH-releasing peptide-2 infusion. J Clin Endocrinol Metab 82:590 –599 Laron Z, Frenkel J, Degenghi R, Anin S, Klinger B, Silbergeld A 1995 Intranasal administration of the GHRP hexarelin accelerates growth in short children. Clin Endocrinol (Oxf) 43:631– 635 Copinschi G, Van Onderbergen A, L’Hermite-BaleriauxM, Mendel C, Caufriez A, Leproult R, Bolognese J, De Smet M, Thorner M, Van Cauter E 1996


32.

33.

34.

35.

36.


Effect of a 7 day treatment with a novel, orally active growth hormone secretagogue, MK-677, on 24 hour GH profiles, insulin-like growth factor-I and adrenal cortical function in normal young men. J Clin Endocrinol Metab 81:2776 –2782 Malmlof K, Bauer M, Johansen P, Ankersen M, Veldhuis J 2002 Daily lowdose GH secretagogue administration stimulates pulsatile GH secretion and elevates plasma IGF-I levels in pigs. Endocrine 16:195–199 Bianco A, Salvatore D, Gereben B, Berry M, Larsen P 2002 Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr Rev 23:38 – 89 Peeters R, Wouters P, Kaptein E, van Toor H, Visser T, Van den Berghe G 2003 Reduced activation and increased inactivation of thyroid hormone in tissues of critically ill patients. J Clin Endocrinol Metab 88:3202–3211 Baur A, Baur K, Jarry H, Ko¨ hrle J 2000 Effects of proinflammatory cytokines on anterior pituitary 5⬘-deiodinase type I and type II. J Endocrinol 167:505–515 Boelen A, Maas M, Lowik C, Platvoet M, Wiersinga W 1996 Induced illness

37. 38.

39. 40. 41.

in interleukin-6 knock-out mice: a causal role of IL-6 in the development of the low 3,5,3⬘-triiodothyronine syndrome. Endocrinology 137:5250 –5254 Darras VM, Berghman LR, Vanderpoorten A, Kuhn ER 1992 Growth hormone acutely decreases type III iodothyronine deiodinase in chicken liver. FEBS Lett 310:5– 8 Vasilatos Y, Zhou Y, Wang X, McMurtry J, Rosebrough R, Decuypere E, Buys N, Darras V, Van der Geyten S, Tomas F 2000 Altered chicken thyroid hormone metabolism with chronic GH enhancement in vivo: consequences for skeletal muscle growth. J Endocrinol 166:609 – 620 Moller J, Jorgensen J, Moller N, Christiansen J, Weeke J 1992 Effects of growth hormone administration on fuel oxidation and thyroid function in normal man. Metabolism 41:728 –731 Berry M, Kates A, Larsen P 1990 Thyroid hormone regulates type I deiodinase messenger RNA in rat liver. Mol Endocrinol 45:743–748 Tabata S, Nishikawa M, Toyoda N, Yonemoto T, Ogawa Y, Inada M 1999 Effect of triiodothyronine administration on reduced expression of type 1 iodothyronine deiodinase messenger ribonucleic acid in streptozotocin induced diabetic rats. Endocr J 46:367–374

Erratum In the full-length manuscript entitled “Phosphorylation of Insulin-Like Growth Factor (IGF) Binding Protein-3 by Breast Cancer Cell Membranes Enhances IGF-I Binding” by Suresh Mishra and Liam J. Murphy (Endocrinology 144:4042– 4050, 2003) the authors request that the following correction be noted. In Fig. 1A, an error may have occurred in the preparation of the figure. It now appears that this component of Fig. 1 may have been conducted with T47D cells enriched for IGFBP-3 membrane binding sites by panning rather than unenriched T47D cells, which were used for the rest of the experiments described in the manuscript using T47D cells. Because the only relevant difference between the enriched and unenriched T47D cells would be the abundance of the binding sites, this does not change significantly the interpretation of the data presented in the figure, the text of the manuscript, or our conclusions. In the interest of accuracy, the corrected Fig. 1 is presented below.

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.