| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
NEUROENDOCRINOLOGY |
Department of Intensive Care Medicine (F.W., G.V.d.B., M.M.), Burn Unit and Center for Experimental Surgery & Anesthesiology and Laboratory for Experimental Medicine and Endocrinology (E.V.H., W.C.), Catholic University of Leuven, Leuven B-3000, Belgium; Department of Medicine (C.Y.B.), Division of Endocrinology, Tulane University Medical Center, New Orleans, Louisiana 70112-2699; and Department of Medicine (J.D.V.), Division of Endocrinology, University of Virginia Health Sciences Center, Charlottesville, Virginia 29908
Address all correspondence and requests for reprints to: Frank Weekers, M.D., Department of Intensive Care Medicine, University of Leuven, B-3000 Leuven, Belgium. E-mail: . Frank.Weekers{at}uz.kuleuven.ac.be
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Seven days after arterial and venous cannulae placement, male New Zealand White rabbits were randomly allocated to a control or a critically ill group. To induce prolonged critical illness, a template controlled 15% deep dermal burn injury was imposed under combined general and regional (paravertebral) anesthesia. Subsequently, critically ill rabbits received supplemental analgesia and were parenterally fed with glucose, insulin, amino acids, and lipids. On d 1 and d 8 after randomization, acute and chronic spontaneous hormonal profiles of GH, TSH, and PRL secretion were obtained by sampling blood every 15 min for 7 h. Furthermore, GH, TSH, and PRL responses to an iv bolus of GH-releasing peptide 2 (GHRP-2) + TRH were documented on d 0, 1, and 8. Hemodynamic status and biochemical parameters were evaluated on d 0, 1, 3, 5, and 8, after which animals were killed and relative wet weight and water content of organs was determined.
Compared with controls, critically ill animals exhibited transient metabolic acidosis on d 1 and weight loss, organ wasting, systolic hypertension, and pronounced anemia on d 8. On d 1, pulsatile GH secretion doubled in the critically ill animals compared with controls, and decreased again on d 8 in the presence of low plasma IGF-I concentrations from d 1 to d 8. GH responses to GHRP-2 + TRH were elevated on d 1 and increased further on d 8 in the critically ill animals. Mean TSH concentrations were identical in both groups on d 1 and 8, in the face of dramatically suppressed plasma T4 and T3 concentrations in the critically ill animals. PRL secretion was impaired in the critically ill animals exclusively on d 8. TSH and PRL responses to GHRP-2 and TRH were increased only on d 1.
In conclusion, this rabbit model of acute and prolonged critical illness reveals several of the clinical, biochemical, and endocrine manifestations of the human counterpart.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Further investigation of the exact mechanisms underlying the neuroendocrine disruption in, and the adverse outcome of high dose GH treatment in, prolonged critical illness requires a valid animal model, which is hitherto lacking. The aim of this study was to implement a model fulfilling the following requirements: 1) a disease state severe enough to evoke metabolic derangements observed in intensive care patients, including lean tissue wasting despite feeding; 2) a condition with a standardized severity and mortality rate and limited spontaneous recovery over time and no spontaneous healing to allow study of acute and chronic responses; and 3) a biphasic neuroendocrine response pattern to sustained, severe physical stress, as observed in the human.
Rat models were considered less suitable as it is well known that rats, in contrast to the human, respond to physical stress and starvation with an immediate suppression of pituitary function (5, 6). In contradistinction to the rodent, an increase in GH secretion in response to starvation has been reported in the rabbit, like in the human (7). Accordingly, initial studies used the rabbit as a model of critical illness (8), wherein a controlled, delimited, and consistent thermal injury in the fed state could be standardized, allowing investigations of a severe but stable condition over several days. We hypothesized that this rabbit model of acute and prolonged critical illness reveals several of the clinical, biochemical, and endocrine manifestations of the human counterpart. We performed dynamic studies in this animal model via frequent blood sampling over several hours and via documenting responses to bolus injections of GH-releasing peptide 2 (GHRP-2) and TRH, to identify the time-dependent and hormone specific alterations within the somatotropic, thyrotropic, and lactotropic axes in the acute and prolonged phases of critical illness.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study 1: GHRP-2 and TRH dose finding study in healthy animals
As it was unknown which dose of GHRP-2 (Kaken Pharmaceutical Co., Ltd., Tokyo, Japan) and TRH (200 µg/ml 0.9% NaCl; Ferring Pharmaceuticals Ltd., Kiel, Germany) effectively releases GH, TSH, and PRL in the studied species, we first performed a dose finding study in healthy rabbits.
Six healthy male New Zealand White rabbits were anesthetized with ketamine (25 mg/kg im, Imalgene 1000, Merial, Lyon, France) and medetomidine (0.15 ml/kg im, Domitor, Orion Corp., Finland). Under anesthesia, the neck and back areas were shaved. Then 4 liters of oxygen per minute were administered through a face mask to the spontaneously breathing animal. Halothane (Fluothane, AstraZeneca, Macclesfield, UK), at a dose of 0.2–0.4 vol% was added by a regular vaporizer to maintain anesthesia. The right neck area was disinfected with chlorhexidine in alcohol and then properly draped. Both the internal jugular vein and carotid artery were dissected free and secured. First, the distal carotid artery was ligated and the proximal part clamped. Through the arterial cutdown, a polyurethane catheter (5CH, 38 cm, Sherwood Medical, Tullamore, Ireland) was advanced 5–6 cm into the vessel and then secured in place. Subsequently, the jugular vein was clamped proximally and the distal part ligated. Through a venous cutdown, a double lumen catheter (4F, 20 cm, Vygon, Ecouen, France) was advanced for 5 cm and secured. Both arterial and venous catheters were tunneled sc and exteriorized to the back of the animal. Catheters were de-aired and filled with heparin (Heparine 5000 IU/ml; Rh\|[ocirc ]\|ne Poulenc Rorer, Brussels, Belgium), 0.3 ml for the internal jugular catheter and 0.4 ml for the arterial catheter. Thereafter, the animals returned to their cages for 72-h recovery. They were individually housed and fed with conventional rabbit diet. Water was freely available.
After the 3-d recovery period, and on 5 d consecutively, each animal was injected iv with saline (0.9% NaCl), 10 µg/kg GHRP-2 + 10 µg/kg TRH, 30 µg/kg GHRP-2, 30 µg/kg GHRP-2 + 30 µg/kg TRH, and 60 µg/kg GHRP-2 + 60 µg/kg TRH, in a random order. Sixty micrograms per kilogram GHRP-2 injected sc in healthy male rabbits was known to evoke a maximal GH response (Bowers, C. Y., unpublished data). Each day at 0920 h, two baseline blood samples with a 10-min interval were taken before injection of the study drug at 0930 h. Subsequently, blood samples were taken every 10 min for 1 h. An additional sample was taken 120 min after bolus injection. After clotting and centrifugation (15 min on 5000 rpm), plasma was stored at -70 C until assay.
In line with previous human studies, we chose to inject rabbits with the combination of GHRP-2 and TRH to document the GH, TSH, and PRL responses to increasing doses of these secretagogues (10). We found that the addition of 30 µg/kg TRH to 30 µg/kg GHRP-2 had no influence on the GH response (mean ± SEM peak GH concentration 26.6 ± 2.7 µg/liter after 30 µg/kg GHRP-2 and 36.7 ± 5.9 µg/liter after 30 µg/kg GHRP-2 + TRH; P = 0.13) and did not induce a significant TSH release. Thirty micrograms per kilogram GHRP-2 alone slightly suppressed PRL levels (P = 0.04 vs. saline), whereas 30 µg/kg GHRP-2 + TRH evoked a significant increase in PRL (P = 0.01 vs. saline).
Study 2: study of the spontaneous release of GH, TSH, and PRL in the acute and chronic phase of critical illness and of the responses to iv bolus injection of 60 µg/kg GHRP-2 and 60 µg/kg TRH
Rabbits were weighed and taken to the lab for instrumentation, as described above, 7 d before initiation of the thermal stressor. After the procedure, animals were fitted into a homemade jacket to secure position of the catheters and returned to their cages for a 7-d period of recovery.
Day 0
After the 7-d recovery period, further referred to as d 0, at 1500 h ± 1 h, animals were weighed and a hemodynamic evaluation was performed by invasive measurement of blood pressure and heart rate (Propac, Protocol Systems Inc., Beaverton, OR). Animals were randomly allocated to either a control group or a critically ill group. Subsequently, all animals received an iv bolus injection of 60 µg/kg GHRP-2 + 60 µg/kg TRH, preceded by two blood samples taken with a 10-min interval and followed by blood sampling every 10 min for 1 h. Samples were handled as described above.
Thereafter, control animals stayed in their cages and had free access to water and were fed daily with regular chow. Animals randomized into the critically ill group were anesthetized with ketamine 30 mg/kg and piritramide 5 mg, both im. Both flanks were shaved. Oxygen and Halothane 1–1.5 vol% were administered until an adequate level of anesthesia was reached. Meanwhile, a paravertebral block with Xylocaine 1% was performed (Astra Pharmaceuticals, Brussels, Belgium). A contact thermal injury was imposed on both flanks. This was accomplished by application of an aluminum template, sized 100–50-20 mm, which had been heated in water of 100 C for at least 30 min. The template was applied to the shaved skin for 30 sec leaving a full thickness, third-degree dermal burn injury. Templates were applied to obtain a 400–500-cm2-sized injury equaling 15–20% total body surface area burned. The wound was then covered with a thin layer of sterile gauze. This type of third-degree burn injury is painless during the days after the injury as cutaneous nerve ends are destroyed (11). Furthermore, this type of injury does not heal without grafting. Hence, this experiment evaluates the combined stress of anesthesia and recovery, and a stable, nonhealing thermal injury.
At the end of the procedure, animals returned to their cages and an overnight iv infusion with Ringers Lactate was started at 6 drops per minute (±18 ml/h). The infusion was delivered by a volumetric pump (IVAC 531 infusion pump) through an ordinary administration set (Codan, Lensahn, Germany). To allow free movement of the animals without twisting of the tubing, the tubing ran over a pulley, and a homemade swivel device was incorporated into the perfusion system. In the evening, a supplemental dose of a major analgesic (piritramide 0.5 mg/kg) was given sc.
Day 1
At the beginning of d 1, parenteral nutrition was started in the critically ill animals at 4 drops/min (±12 ml/h), and blood glucose levels were kept within narrow limits (<8.4 mmol/liter) by frequent blood glucose monitoring and titration of insulin infusion (Actrapid 100 IU/ml, Novo Nordisk, Bagvaerd, Denmark) 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 airflow conditions. The infusion bags contained 103 ml of Glucose 20%, 37.5 ml Aminoplasmal L 10 (Braun, Melsungen, Germany), and 34 ml Intralipid 20% (Fresenius Kabi, Stockholm, Sweden). One hundred fifty millilters of sterile water were added. Thus, bags with 325 ml solution contained 150 kcal and 0.6 g nitrogen. Of all calories, 56% were delivered as carbohydrates and 44% as lipids. Protein intake equaled 1.25 g amino acids/kg·d. There were no additional vitamins or trace elements added. Parenteral nutrition was continuously administered.
On d 1, between 0830 h and 1530 h, blood was sampled from control and critically ill animals every 15 min for determination of GH, TSH, and PRL time series. This was followed by a bolus injection of 60 µg/kg GHRP-2 and 60 µg/kg TRH and sampling every 10 min for 1 h for hormone determination. Parenteral nutrition was continued during GHRP-2 and TRH injections which were given through the second lumen of the double-lumen central venous catheter. In both time series, blood samples of 1.2 ml were 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. The red blood cells were resuspended in NaCl 0.9%, centrifuged once again, and the red blood cell mass was reinfused both in the control and the critically ill animals.
Hemodynamic variables were determined at the beginning and end of d 1. At the end of d 1, an additional blood sample was taken for blood gas and biochemical analyses.
Days 2–8
The control animals had free access to water and were fed daily with regular chow. The critically ill animals received water ad libitum and continuous parenteral nutrition, with the parenteral nutrition infusion bag being replaced every 24 h. On the mornings of d 3, 5, and 8, hemodynamic parameters were recorded, and blood samples were taken for determination of blood gas values and biochemical analysis.
Day 8
On d 8, another 7-h sampling time series (sampling every 15 min) was done to evaluate GH, TSH, and PRL secretion (further referred to as the chronic profile), followed by another iv bolus injection of 60 µg/kg GHRP-2 and of 60 µg/kg TRH (and sampling every 10 min for 1 h). The procedure was as described on d 1. After the last sampling, animals were weighed and euthanized with sodium-pentobarbital (Nembutal, 60 mg/ml, Sanofi Pharmaceuticals, Inc., Winthrop, NY). Organs were harvested and heart, lungs, liver, spleen, and kidney weighed. Then tissue samples were taken from left and right ventricle, lung, liver, spleen, kidney, ileum, quadriceps muscle, intact skin, and burned skin. Of 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 lyophyllization (Alfa I-5 Chriss, Aichach, Germany).
Assays
Biochemistry.
Arterial whole blood was analyzed on an ABL-analyzer (Radiometer, Copenhagen, Denmark) to quantitate blood pH, pO2, pCO2, and levels of hemoglobin (Hb), lactate, glucose, calcium, sodium, potassium, and bicarbonate.
GH
All samples were analyzed in duplicate within a single assay run. A specific RIA was used for determination of rabbit plasma GH concentrations (reagents kindly provided by Dr. A. Parlow, National Pituitary Agency). The detection limit was 1 µg/liter and the within-assay CV was 2.3%. All samples contained detectable plasma GH concentrations.
IGF-I
Plasma IGF-I was analyzed 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, Inc., Richmond, CA). Des-IGF-I was used as tracer to avoid binding to the small IGF-binding proteins. The intraassay CV was 4.6%. The detection limit was 20 µg/liter.
TSH and PRL
All samples were analyzed in duplicate within a single assay run. Specific RIAs for determination of rabbit TSH and PRL levels were used (reagents kindly provided by Dr. A. Parlow, National Pituitary Agency, Los Angeles, CA). The detection limits were 1.2 mU/liter and 0.5 µg/liter, respectively. For samples with undetectable levels, a value representing one half the detection limit was entered.
T4 and T3
Total plasma T4 and T3 concentrations were determined by a specific RIA (Immunotech, Marseille, France). The sensitivity for T3 was 0.2 nmol/liter and for T4 was 6 nmol/liter. All samples were assayed in duplicate.
Statistical analysis
The time series of sequential plasma GH concentrations measured over the 7-h profiles on d 1 and 8 were transformed into pituitary secretion profiles by adjusting for endogenous half-life using multiple parameter deconvolution analysis (13). This method is designed to compute hormonal half-life and the number, amplitude, and mass of underlying pituitary secretory bursts. GH half-life in the rabbit was estimated as (mean ± SD) 7.7 ± 1.5 min by deconvolving the GH responses to the bolus injection of GHRP-2 (60 µg/kg iv) in the healthy rabbits. This GH half-life was applied to the sequential time series to quantify pulsatile secretion. Basal GH secretion was assumed to be zero. The following parameters were calculated for each hormonal profile: GH burst amplitude [maximal secretory rate (µg per liter of distribution volume (Lv) released per min)] and temporal position of all secretory bursts; the mass of hormone secreted per burst [area of the resolved secretion burst (µg per Lv)]; and the mean pulsatile production [product of the number of secretory bursts and the mean secretory burst mass over the time interval considered (µg per Lv over 7 h)].
Independently, approximate entropy (ApEn) was determined to monitor regularity of the time series (14). Larger ApEn values correspond to a greater process randomness (15).
As several samples had undetectable values for PRL and TSH, deconvolution analysis of the PRL and TSH profiles was impossible. Therefore, PRL and TSH secretion was estimated by mean plasma concentrations with half the detection limit assigned to those samples in which no detectable amount was measured.
Responses to bolus injections of GHRP-2 and TRH were compared between groups using ANOVA for repeated measures or factorial ANOVA and Fisher’s protected least significant difference posthoc testing for multiple comparisons, as appropriate. Within-group changes in control and critically ill rabbits were analyzed by ANOVA for repeated measures, or Wilcoxon signed rank test (and Bonferroni correction for multiple comparisons). Between-group differences on the different study days were analyzed by repeated measures ANOVA or Mann Whitney U test. A P value equal or less than 0.05 was construed significant. Data are expressed as mean ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Compared with saline, only 60 µg/kg GHRP-2 + 60 µg/kg TRH induced a significant increase of GH, TSH, and PRL plasma concentrations in healthy rabbits. This dose was thus used for further study. Deconvolution analysis of the GH response after 60 µg/kg GHRP-2 + 60 µg/kg TRH injection revealed a GH half-life of 7.7 ± 0.5 min.
|
Spontaneous endocrine changes
GH.
Mean and integrated (7-h) plasma GH concentrations and calculated spontaneous GH secretion increased on d 1 in critically ill (n = 8) compared with control animals (n = 9) (Table 1, Fig. 2). No difference between the two groups was found on d 8. The doubling of GH secretion on d 1 was exclusively due to a larger size of the GH pulses (Table 1). The number of bursts was similar in the two groups and no change was observed between d 1 and d 8. GH pulse half-duration was significantly longer in both groups on d 1 compared with d 8 but equal between the groups.
|
|
). On d 1, plasma IGF-I was lower in critically ill animals compared with controls. Plasma IGF-I remained normal in controls and low in critically ill animals on d 8. Plasma IGF-I levels in critically ill animals were identical on d 1 and 8.
|
) in critically ill rabbits were indistinguishable from controls, both on d 1 and d 8.
T4, T3.
On d 0, plasma T4 and T3 concentrations were identical in control and critically ill animals (Fig. 3). On d 1, T4 and T3 levels were undetectable in critically ill animals while they remained stable in controls. On d 8, T4 and T3 levels remained low in critically ill animals compared with controls.
PRL.
On d 1, mean PRL levels in critically ill rabbits were indistinguishable from controls (Fig. 2). However, on d 8, mean PRL levels were lower in critically ill animals. Lower values in the critically ill were mainly due to a lack of an initial rise in PRL in response to manipulation at the start of the sampling session.
Changes in body weight and hemodynamics
Critically ill rabbits received 41 kcal/kg·d. Mean ± SEM body weight at baseline (d -7, before instrumentation) was 2987 ± 109 g in the control group and 3061 ± 56 g in the critically ill group. Over the 7-d recovery after instrumentation, the change in body weight was identical in the control group and the critically ill group (+3.3 ± 1.8% and +2.6 ± 1.9%, respectively; P = NS). From the day of randomization (d 0) to the end of the experiment (d 8) the critically ill animals lost weight (-10.8 ± 3.1%), whereas body weight remained unaltered in the control animals (+0.1 ± 1.9%; P = 0.03 vs. critically ill animals). An eschar was identified at the end of the experiment.
Systolic blood pressure before the injury on d 0 was 99 ± 8 mm Hg in the critically ill group and 95 ± 5 mm Hg in the control group (P = NS). On d 1, systolic blood pressure was higher in the critically ill animals (102 ± 6 mm Hg) than in controls (78 ± 5 mm Hg; P = 0.02), as it was on d 8 (104 ± 5 mm Hg vs. 83 ± 4 mm Hg, respectively; P = 0.02). Diastolic and mean blood pressure and heart rate were not detectably different between the two study groups on any of the studied time points.
Biochemistry (Fig. 4)
Arterial pO2 and pCO2, and levels of ionized calcium, sodium and potassium were not significantly different between the two study groups at any of the studied time points. At baseline, arterial pH, bicarbonate (20.1± 1.1 mmol/liter in the control group and 19.4 ± 0.5 mmol/liter in the critically ill group; P = NS) and lactate levels were identical in both study groups. On d 1, pH slightly decreased in the critically ill group (P = 0.0002) and was associated with a lower bicarbonate level (17.6 ± 0.8 mmol/liter in the critically ill group and 20.2 ± 0.8 mmol/liter in the control group; P = 0.02) and a higher lactate concentration (P = 0.006) in the critically ill group, but this effect was transient. Blood glucose levels remained identical in both groups throughout the study, as a consequence of exogenous insulin administration in all critically ill animals. On d 0, Hb level in the critically ill group was identical to that in the control group. On d 1, Hb level decreased to the same extent in critically ill and control animals and this remained stable until d 3. On d 8, Hb was significantly lower in critically ill animals than in controls.
|
. Compared with control rabbits, the proportional weight of lung, heart, and liver relative to body weight was significantly higher in critically ill animals. The water content of different tissues was comparable in critically ill and control animals except for the normal skin, which lost water in the critically ill animals, and the left ventricle, which retained water compared with controls (Table 3).
|
|
) in the acute and chronic phase of critical illness
|
PRL.
On d 0, there was no PRL increase in response to 60 + 60 µg/kg GHRP-2 + TRH iv injection in either study group. On d 1, there was a significant PRL release in both groups (P = 0.02 for d 1 vs. d 0 in the critically ill group; P = 0.04 for d 1 vs. d 0 in the control group), and the response was higher in the critical ill group (P = 0.05 vs. controls). On d 8, however, the PRL release was higher in control than in critically ill animals (P = 0.04).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The dose-response study in healthy rabbits revealed a high dose per kg body weight of GHRP-2 and TRH needed to consistently evoke release of GH, TSH, and PRL in this species (17). In contrast to the PRL release observed in response to GHRPs in the healthy human (18) and the rat (19), a low dose of GHRP-2 alone appeared to suppress PRL levels in the healthy rabbit, an effect that was overcome by the addition of TRH. In the rat, release of PRL by GHRP appears dependent on the GH status and the gonadal steroid milieu (20). The mechanisms explaining these interspecies differences are unclear.
The high mortality (40%), the presence of wasting despite feeding, transient metabolic acidosis, and systolic hypertension in the critically ill animals surviving the 8 study days confirmed severe and sustained stress in this catabolic model. Indeed, critically ill animals lost about 11% of their body weight over the 8 study days despite parenteral nutrition and exogenous insulin treatment. Weight loss was not due to resorption of burned tissue as an eschar was identified at the end of the experiment. Lean tissue wasting, which presumptively occurred mainly in skeletal muscle (21), was also present in the spleen and kidneys, as these vital organs decreased in weight proportionately to the drop in total body weight without a detectable change in water content. The liver, heart, and lungs did not lose weight relative to the rest of the body. In the heart, this can at least in part be explained by fluid retention, as indicated by the higher water content in the left ventricle. Other possible mechanisms include infiltration by immune cells and fat accumulation. Furthermore, critically ill animals were unable to compensate for a drop in hemoglobin that occurred because of the frequent blood sampling, in contrast to their healthy counterparts. This interesting observation can be explained by impaired erythropoiesis in the critically ill animals (22), or less likely by fluid retention (23) and increased hemolysis (24).
For the first time in an animal model, we showed the presence of a biphasic neuroendocrine response to prolonged stress thereby confirming the human data. As in the human (3, 25, 26) the acute stress response within the GH axis in critically ill rabbits comprised increased GH secretion and decreased plasma concentrations of IGF-I, suggesting peripheral GH resistance. Specifically, the size of the GH bursts was significantly increased in acute illness, without a detectable change in pulse frequency. The latter indicates that the frequency of somatostatin withdrawal, presumably the main pacer for generating GH pulses (27), was comparable in critically ill and control animals. In the human condition of acute illness, a relatively high pulse frequency has been reported (3). This possible difference between the human and the rabbit model might be due to technical sampling differences and/or moderate degree of stress in the control rabbits, which may have obscured a rise in GH pulse frequency in the critically ill animals. The increase in GH pulse mass on d 1 in the critically ill animals can be explained either by reduced hypothalamic somatostatin tone, or an increased release of GHRH and/or of the endogenous ligand of the GH-releasing peptide receptor, e.g. ghrelin (28). In human models, continuous infusion of these two releasing factors indeed selectively amplifies the size of the GH pulses without altering GH pulse frequency (3, 29).
Bolus injection of the secretagogues evoked increased release of GH in both groups, although the increase was substantially higher in critically ill animals than in controls. The higher GH response in control animals on d 1 may indicate a moderate degree of stress (and associated lower somatostatin tone) in these animals, as inferred above. The pronounced GH response to GHRP-2 in critically ill animals together with spontaneously increased pulsatile GH secretion on the first day after injury are compatible with reduced somatostatin tone in response to acute stress. Theoretically, less dopaminergic inhibition of pituitary hormone release could also contribute but the lower unstimulated PRL levels on d 1, however, make this unlikely.
In the chronically ill rabbits, spontaneous GH secretion decreased again by d 8, in the face of stable low plasma IGF-I concentrations. The late decline of GH secretion was exclusively due to loss of GH pulse mass, as GH pulse frequency remained unaltered. Impairment of GH synthesis due to disturbances within the somatotropes may theoretically explain this finding, in keeping with the general notion of multiple organ failure in critical illness and its concomitants (e.g. parenteral nutrition, anesthesia, analgesia, immobilization, etc.). However, GH hyperresponsiveness to injection of GHRP-2 on d 8 cannot be reconciled with depletion of pituitary somatotrophs and thus rather points to a hypothalamic cause of the reduced spontaneous GH release. Altered somatostatin tone alone also cannot reconcile the relatively suppressed spontaneous GH release, the stable GH pulse frequency, and the GH hyperresponsiveness to the secretagogue. Hence, a reduced availability of GHRH and/or endogenous GHRP-like ligands, such as ghrelin (23), could be responsible. Circulating ghrelin originates from the stomach and peripheral injection of ghrelin evokes GH release and increases food intake, which has been interpreted as pointing to a role for ghrelin in mediating the peripheral signal to amplify GH release in fasting conditions (30). Ghrelin, as well as the synthetic GHRP-like ligands, is dependent on sufficient endogenous GHRH availability for their normal GH releasing capacity (31, 32). Hence, reduced availability of endogenous ghrelin in the chronic phase of illness may be inferred to up-regulate GHRP-receptors and thus explain hyperresponsiveness of somatotrophes to an injection of GHRP-2 in the face of impaired spontaneous GH release.
Stable low plasma IGF-I concentrations but waning off of pulsatile GH hypersecretion in chronic stress would be consistent with an inability of GH secretion to compensate for the low IGF-I levels by an expected feedback-dependent rise (33). Improvement of the GH/IGF-I ratio may also suggest a certain degree of recovery of GH responsiveness. These disparities in GH/IGF-I regulation also typify prolonged critical illness in the human (34). Our findings in the rabbit are in contrast with observations in the rat during starvation, systemic acidosis, and other metabolic stress (35). Indeed, changes in GH secretion in response to various modes of stress in the rat have been documented to follow the inverse pattern over time, with an immediate suppression of GH secretion followed by an elevation at a later stage (36). The cause of this interspecies difference is unknown.
TSH secretion was not detectably different between the critically ill and healthy animals throughout the study episode. In contrast, T4 and T3 levels were dramatically suppressed in the critically ill rabbits from the first day onward and through to d 8. T3 changes mirror the early response to illness of the thyroid axis, which is characterized by reduced plasma concentrations of total T3, elevated reverse T3, and relatively normal total T4 levels, labeled euthyroid sick syndrome (37, 38). A transient phase of normal T4 may have been missed in our setting. Indeed, acute stress in rabbits also decreases activity of type-I deiodinase resulting in hampered peripheral conversion of T4 to T3, which leads to low plasma T3 concentrations (6). This initial effect on T4 to T3 conversion is likely induced by cytokines (39, 40). Our experimental setting, with the first evaluation taking place as late as 16 h after onset of critical illness, may have missed this very early phase of selectively low plasma T3 concentrations. Alternatively, the immediate drop in T4 may be related to the severity of critical illness in this model (41, 42). The dramatic drop in T4 from d 1 through d 8 in the face of unaltered TSH release indicates central, either pituitary or hypothalamic, neuroregulatory failure. Theoretically, this constellation may reflect altered setpoint for thyroid hormone-dependent feedback inhibition, impaired pituitary TSH synthesis, elevated somatostatin input, and/or inadequate stimulation of TSH by endogenous TRH. The TSH hyperresponsiveness to a TRH bolus on d 1, together with inadequate TSH secretion for the extremely low circulating levels of T4 and T3, could be explained by reduced TRH availability. Direct and indirect evidence for this mechanism has been obtained in the human situation of protracted critical illness (17, 30, 43). In the chronic phase of illness in our rabbit model, however, the TSH hyperresponsiveness to TRH was not maintained. Results from human studies suggest that sufficient GHRP-like ligand availability may be necessary for full TRH action (33).
Although we did observe a PRL response to 60 µg/kg GHRP-2 + TRH in the initial dose finding study, this was absent on d 0 of the second study, both in controls and in critically ill animals. This difference could be explained by a higher degree of stress (44) in the second study, as also suggested by almost twice as high baseline PRL levels compared with those in the dose-response study.
In the chronic phase of critical illness, PRL levels were lower in the critically ill animals compared with controls. Furthermore, the initial rise in PRL in response to handling at the start of the blood sampling time series, present in the control animals, was absent in the chronically ill rabbits and the initial PRL hyperresponsiveness to GHRP-2 + TRH turned into hyporesponsiveness. Together, these findings point to relatively impaired PRL secretion in the chronic phase of stress, mimicking the human condition (4). It is unclear which mechanism in responsible for this late suppression, but increased endogenous dopaminergic tone (45) or lack of ghrelin (46) could be involved.
In conclusion, our rabbit model of acute and prolonged critical illness reveals several of the clinical, biochemical and endocrine manifestations of the human counterpart.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: ApEn, Approximate entropy; GHRH, GH-releasing hormone; GHRP, GH-releasing peptide 2; Hb, hemoglobin.
Received August 27, 2001.
Accepted for publication October 31, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Y. Debaveye, B. Ellger, L. Mebis, T. J. Visser, V. M. Darras, and G. Van den Berghe Effects of Substitution and High-Dose Thyroid Hormone Therapy on Deiodination, Sulfoconjugation, and Tissue Thyroid Hormone Levels in Prolonged Critically Ill Rabbits Endocrinology, August 1, 2008; 149(8): 4218 - 4228. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Depke, G. Fusch, G. Domanska, R. Geffers, U. Volker, C. Schuett, and C. Kiank Hypermetabolic Syndrome as a Consequence of Repeated Psychological Stress in Mice Endocrinology, June 1, 2008; 149(6): 2714 - 2723. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Collier, L. A. Dossett, A. K. May, and J. J. Diaz Glucose Control and the Inflammatory Response Nutr Clin Pract, February 1, 2008; 23(1): 3 - 15. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Luque, S. Park, and R. D. Kineman Severity of the Catabolic Condition Differentially Modulates Hypothalamic Expression of Growth Hormone-Releasing Hormone in the Fasted Mouse: Potential Role of Neuropeptide Y and Corticotropin-Releasing Hormone Endocrinology, January 1, 2007; 148(1): 300 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Debaveye, B. Ellger, L. Mebis, E. Van Herck, W. Coopmans, V. Darras, and G. Van den Berghe Tissue Deiodinase Activity during Prolonged Critical Illness: Effects of Exogenous Thyrotropin-Releasing Hormone and Its Combination with Growth Hormone-Releasing Peptide-2 Endocrinology, December 1, 2005; 146(12): 5604 - 5611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Andersen, J. Gjedsted, C. Christiansen, and E. Tonnesen The roles of insulin and hyperglycemia in sepsis pathogenesis J. Leukoc. Biol., March 1, 2004; 75(3): 413 - 421. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Weekers, M. Michalaki, W. Coopmans, E. Van Herck, J. D. Veldhuis, V. M. Darras, and G. Van den Berghe 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 Endocrinology, January 1, 2004; 145(1): 205 - 213. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Weekers, A.-P. Giulietti, M. Michalaki, W. Coopmans, E. Van Herck, C. Mathieu, and G. Van den Berghe Metabolic, Endocrine, and Immune Effects of Stress Hyperglycemia in a Rabbit Model of Prolonged Critical Illness Endocrinology, December 1, 2003; 144(12): 5329 - 5338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Van den Berghe, D. Van Roosbroeck, P. Vanhove, P. J. Wouters, L. De Pourcq, and R. Bouillon Bone Turnover in Prolonged Critical Illness: Effect of Vitamin D J. Clin. Endocrinol. Metab., October 1, 2003; 88(10): 4623 - 4632. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||
) (P = 0.0005 vs. saline) and also significantly increased TSH (P = 0.005 vs. saline) and PRL (P = 0.006 vs. saline). Open squares depict the saline group, filled triangles the 10 µg/kg GHRP-2 + 10 µg/kg TRH group, filled circles the 30 µg/kg GHRP-2 + 30 µg/kg TRH group, and open diamonds the 60 µg/kg GHRP-2 + 60 µg/kg TRH group. Results are depicted as mean ± SEM changes from baseline. Significance testing was done by ANOVA followed by Fisher’s protected least significant difference for multiple comparisons.
