Restoration of juvenile baseline growth hormone secretion with preservation of the ultradian growth hormone rhythm by continuous delivery of growth hormone-releasing factor R. Vasilatos-Younken, P. H. Tsao, D. N. Foster, D. L. Smiley, H. Bryant and M. L. Heiman Department of Poultry Science, The Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A.

*Department of Animal Science, University of Minnesota, St Paul, Minnesota 55108, U.S.A. fLilly Research Laboratories, Eli Lilly and Company, Indianapolis and Greenfield, Indiana 46140, (P. H. Tsao is now at Pig Research Institute, Miaoli, Taiwan, R.O.C.) received 8 January 1992

U.S.A.

ABSTRACT

The ability of continuously delivered GH-releasing factor (GRF) to enhance GH secretion while maintaining the normal ultradian GH rhythm was investigated. Synthetic human GH-releasing factor (hGRF(1\p=n-\44)NH2) was continuously infused for 4 days by means of i.v. catheters to 11-week-old broiler chickens. At this age, overall endogenous GH secretion is low, and baseline GH is barely detectable. Six birds per treatment received vehicle (control), 0\m=.\324 mg hGRF(1\p=n-\44)NH2/kg body weight per day

(low dose) or 3\m=.\24 mg hGRF(1\p=n-\44)NH2/kg body weight per day (high dose). After 4 days of GRF conditioning, concurrent with continued GRF infusion, serial blood samples were removed via atrial catheters at 15-min intervals for 6 h and GH plasma

profiles determined. High dose GRF significantly increased GH plasma concentrations over tenfold compared with controls;

however, most of this increase reflected an increase in basal GH, which was reinstated to juvenile baseline levels. Augmentation of pulse amplitude above this increased baseline was not proportionately as high,

INTRODUCTION

The episodic release of growth hormone (GH), known to occur in man and essentially all other animal models thus far studied, results from the co¬ ordinated actions of the hypothalamic neuropeptides, GH-releasing factor (GRF) and somatostatin (SRIF) (for recent reviews see Tannenbaum, Painson, Lapointe et al. 19906; Tannenbaum, 1991). There is

and failed to reach juvenile levels. The ultradian rhythm of GH was not altered by continuous GRF administration. Both low and high dose GRF treatments resulted in significant enlargement of the anterior pituitary gland. Total pituitary GH mRNA levels, although elevated over twofold by GRF treatment, were not significantly different from controls. Measures of plasma GH magnitude (overall and baseline mean, and peak amplitude) were significantly correlated with pituitary GH mRNA for control birds, but were not correlated for GRF treatments. Feed intake was markedly depressed (33%) on the high dose GRF treatment, in conjunction with total inhibition of body weight gain over the 4-day period of administration. Longitudinal bone growth and width of the epiphyseal growth plate were also significantly reduced by high dose GRF treatment, probably reflecting the reduced level of nutrient intake, despite high circulating concentrations of GH. Journal of Endocrinology (1992) 135, 371\p=n-\382

evidence both in vitro and in vivo which suggests that repeated exposure to GRF results in desensitization of pituitary somatotrophs, such that the GH response is not sustained, and subsequent responses to previ¬ ously effective bolus doses are attenuated (Ceda & Hoffman, 1985; Gelato, Rittmaster, Pescovitz et al. 1985; Losa, Schopol, Konig et al. 1986). Although the structure of chicken GRF is unknown, acute expo¬ sure to synthetic human pancreatic GRF (hpGRF)

stimulates pituitary release of GH in a dose-dependent both in vitro and in vivo in chickens (Leung & Taylor, 1983; Huybrechts, Scanes, Rivier & Vale, 1984; Huybrechts, Decuypere, Scanes et al. 1985). However, a refractory period during which further stimulation of GH secretion cannot be elicited by additional GRF occurs in anaesthetized birds (Harvey, Scanes & Phillips, 1985). Given that responsiveness to other GH secretagogues is main¬ tained during such refractory periods (Scanes & Harvey, 1985), pituitary GH depletion is not the limit¬ ing factor. Studies demonstrating refractoriness, how¬ ever, were of relatively short duration and positive biological effects reflective of GH action have been induced by prolonged continuous GRF administra¬ tion in other species (Moseley, Krabill, Friedman & Olsen, 1985; Vance, Kaiser, Martha et al. 1989). The ability of long-term GRF administration to enhance GH secretion, while maintaining the endogenous pulse pattern (ultradian rhythm) of GH in normal birds has not been investigated. The secretory profile of GH has been well charac¬ terized in the domestic chicken, and displays an agerelated decline characterized by diminution of peak amplitude and baseline levels with increasing age and decreasing growth rate (for review see VasilatosYounken & Scanes, 1991). Such a decline in GH secretion has been recognized in many species, and is suggested to arise from a diminished GH response to GRF (Ceda, Valenti, Butturini & Hoffman, 1986; Foltzer-Jourdainne, Harvey & Mialhe, 1988; Johnson & McMurtry, 1989; Wheeler & Styne, 1990). The

manner

ability of continuous GRF administration to reinstate

juvenile level of pituitary GH secretion after the endogenous age-related peak in GH secretion is passed has not been investigated. The objectives of the present study were to deter¬ mine if continuous sustained delivery of GRF for 4 days (1) alters the ultradian rhythm of GH in chickens of a rapidly growing commercial broiler line, and (2) a

can effectively enhance GH secretion and reinstate a juvenile GH plasma profile at an age when GH pulse amplitude is low, and basal GH secretion is barely

detectable.

MATERIALS AND METHODS

All birds used in these studies were males of the Hubbard Hubbard cross, a commercial broiler line displaying rapid growth rate. Chicks were obtained (Metz Hatchery, Belleville, PA, U.S.A.) at 1 day of age and a commercial broiler starter ration was avail¬ able ad libitum. Chicks were maintained on litter until 9 weeks of age, at which time they were transferred to individual wire cages in a temperature-, light- and

ventilation-controlled room. A cycle of 16 h light: 8 h darkness was maintained throughout these studies. Birds had free access to fresh feed and water at all times except prior to surgery and blood sampling, as noted.

Surgical preparation of birds At 10 weeks of age, birds were fasted overnight and anaesthetized with sodium pentobarbitol (31 mg/ kg body weight i.V.). Dual catheters were inserted into the right jugular vein as described previously (Cravener & Vasilatos-Younken, 1989). One catheter was threaded proximally until the intravenous end just entered the right atrial chamber and was used for blood sampling. The second catheter was threaded proximally until the intravenous end was positioned 6 cm distal to the sampling catheter. This second catheter was used for infusion. All birds were allowed a minimum 72-h postoper¬ ative recovery period before the start of infusions. Birds which had not exceeded their presurgery weight at the start of the experimental period were not used. Each catheterized bird was maintained in a harness/ spring tether/fluid swivel system connected to a micro¬ processor-controlled infusion pump (Model AS20A; Travenol Laboratories, Hookset, NH, U.S.A.), as described by Cravener & Vasilatos-Younken (1989), which allowed the birds unrestricted movement within the cage, including 360° rotation, even while infusion and blood sampling occurred. Body weights were recorded at the initiation and termination of infusions, and feed consumption from the start of infusion until the start of blood sampling (4 days) was determined for each bird. Infusion and blood

sampling protocol Synthetic human (h) GRF(1-44)NH2 (Eli Lilly and Co., Greenfield, IN, U.S.A.) was dissolved in 0-01 mol H3PO4/I and diluted in 0-9% (w/v) NaCl containing 0-1% chicken serum albumin (Sigma Chemical Co., St Louis, MO, U.S.A.; vehicle). Begin¬ ning at 75 days of age, six birds per treatment were infused with vehicle (control), 0-324 mg hGRF(l44)NH2/kg body weight (low dose GRF) or 3-24 mg hGRF(l-44)NH2/kg body weight (high dose GRF).

Infusion pumps were set to deliver each treatment at rate of 0-7 pi per min (1005 ml per day) continu¬ ously for 4 days. New disposable syringes containing fresh infúsate solutions were loaded onto pumps daily. After 4 days of GRF conditioning, feed was removed from all birds at 07.00 h and, while infusions continued, blood sampling commenced at 09.00 h. Blood samples (0- 5 ml per sample) were removed at 15-min intervals for 6 h (24 total samples; 12 ml

a

removed, representing approximately 7% of estimated

total blood volume). Catheter lines were backflushed with 0-5 ml sterile 0-9% (w/v) NaCl solution after each blood withdrawal. Upon collection into heparinized tubes (35 U/tube), samples were placed on ice and centrifuged within 10 min of collection for recov¬ ery of plasma. Plasma samples were stored at —80 °C until analysed. Somatic measurements

At the completion of blood sampling, birds were killed by cervical dislocation concurrent with collec¬ tion of trunk blood into tubes containing diprotin A (25 pi of a 1 mmol/1 solution, lyophilized, per 0-5 ml blood) (Sigma Chemical Co.) for determination of immunoreactive hpGRF(l 44) concentration. Diprotin A was used to prevent possible degrada¬ tion of GRF by dipeptidylpeptidase (i.v.) activity (Umezawa, Aoyagi, Ogawa et al. 1984). Pituitary glands were harvested, weighed and rap¬ idly frozen on dry ice. Glands were stored at —80 °C for quantitation of GH mRNA. The abdominal fat pad, liver, and left and right tibiotarsi were removed and weighed. The length of each tibiotarsus was deter¬ mined using micrometer calipers. Thin sections of the right tibiotarsal proximal epiphysis were fixed in buffered 10% (v/v) formalin, and the width of the epiphyseal growth plate was determined using a stage micrometer. Bone mineral density was determined on each left tibiotarsus by means of single photon absorptiometry using a Norland 2780 X-ray densitometer (Black, Rowley, Herring & Williams, 1989; Wahner, 1989). Triplicate readings were made at the proximal aspect of the patellar groove (distal metaphysis). Bone mineral density was calculated as a function of bone mineral content and bone width at the point of measurement, and expressed as g/cm2. GH mRNA

analysis

The RNA extraction procedure of Chomczynski & Sacchi (1987) was followed exactly for individual pituitaries except that the purified total cellular RNA was suspended in 100 pi sterile diethylpyrocarbonate (DEPC)-treated distilled H20 prior to use. Triplicate RNA samples (20, 10 and 10 pi respectively) from each of the individual pituitaries were denatured in formamide (Maniatis, Fritsch & Sambrook, 1982) and immobilized on nitrocellulose paper using a slotblot manifold (S & S Inc., Keene, NH, U.S.A.). For a standard curve, known amounts (5000, 2500, 1250, 625, 313, 167, 83, 42, 21, 10, 5 and 2-5 pg) of a chicken GH (cGH) cDNA (Lamb, Galehouse & Foster, 1988) insert were denatured and blotted as above. Filters were baked, then pre-hybridized in

(Seigel & Bresnick, overnight at 42 °C using an oligonucleotide-primed (Feinberg & Vogelstein, 1983) 32P-labelled cGH insert at a specific activity of 4-5 x 107 d.p.m./ml in a fresh solution of DEPC-treated Blotto. Following hybridization, filters DEPC-treated Blotto solution 1986). Filters were hybridized

were

washed twice in 2

SSC

(1 x SSC 150 mmol =

NaCl/1,15 mmol sodium citrate/1, pH 7-0) at 42 °Cfor 30 min and finally in 0-2 SSC at 45 °C for 1 h. Filters were exposed overnight to X-ray film. For quantifying

RNA, modifications were used of the Hollander & Fornance (1990) procedure, which measures the relative constant amount of poly A mRNA found in total cellular RNA samples using a radioactive poly probe. For a standard curve, doubling dilutions of known amounts of total cellular turkey pituitary RNA (obtained from ten 6-week-old birds of mixed sex) were used. Duplicate 10 pi samples of standards (4000, 2000, 1000, 500, 250, 125, 63 and 31 ng) were mixed with formamide as above and slotblotted. For each of the individual pituitaries, dupli¬ cate 10 pi samples were mixed with formamide as above and loaded onto the manifold. Both sets of filters with immobilized total cellular RNA were baked and prehybridized as above. A 35S-labelled poly probe was prepared using 10 pg poly A and 10 ng dT(12 18) primer (both from Sigma Chemical Co.) in the presence of dATP, dCTP, dGTP plus 35S-labelled TTP, murine reverse transcriptase (BMB, Indiana¬ polis, IN, U.S.A.) and RNAsin (ProMega Biotech, Madison, WI, U.S.A.). The 50 pi reaction was stopped, and base hydrolysed in 1 mol NaOH/1 at 65 °C for 30 min. 35S-Labelled TTP incorporated into the poly probe was separated from soluble material by column chromatography. Filters were hybridized in DEPC-treated Blotto plus the poly probe (6-8 x 106 d.p.m./ml) at 42 °C overnight. Filters were washed twice with 2 x SSC at room temperature, then in 0-2 x SSC at 42 °C for 30 min, followed by exposure to X-ray film. After autoradiography, individual slots were care¬ fully removed from nitrocellulose filter paper and counted by liquid scintillation spectroscopy. The amount of RNA hybridized to the cGH probe from individual birds was plotted against the cGH standard curve. Values reported were the average of the three samples that were slot-blotted. The amount of cGH that was determined from the standard curve was nor¬ malized to the amount of total cellular RNA slotblotted on a per pg basis. amounts of slot-blotted

GH

radioimmunoassay profile analysis

and GH

plasma

Plasma immunoreactive GH concentration was deter¬ mined by a homologous chicken radioimmunoassay

described in detail previously (Vasilatos-Younken, 1986), using a recombinant chicken GH (r-cGH) preparation as standard and for iodination, and rab¬ bit anti-r-cGH serum as primary antibody (both pro¬ vided by Amgen, Thousand Oaks, CA, U.S.A.). The as

minimum detectable concentration of this assay, defined as the least concentration of unlabelled hor¬ mone (standard) that could be distinguished from no unlabelled hormone (B0), based on confidence inter¬ vals for Bo and for standards, was 0-5 ng/ml. The concentration at 50% binding was 3-8 ng/ml. Intraassay coefficient of variation (C.V.) was 3-8%, based on quadruplicate samples of high (44 ng/ml) and low (4 ng/ml) concentration quality control plasma pools that reflected the range of the assay, and were repeated every 300 assay tubes (all samples were run in a single assay). Plasma GH concentration data were subjected to PULSAR (Merriam, 1983), a Fortran 10based computer program that implements the pulse identification algorithms described by Merriam & Wächter (1982) for quantitation of overall and base¬ line mean, peak amplitude (defined here as the abso¬ lute concentration at a given peak, inclusive of baseline), peak duration and interpeak interval. The smoothing window used was 5-75 h in width, and G(i_5) values were 3-8, 2-6, 1-9, 1-5 and 1-2 respec¬ tively. Area under the plasma GH profile curves was determined based on Simpson's rule for numerical

integration (Rektorys, 1969). GRF

radioimmunoassay

Plasma immunoreactive hGRF( 1- 44)NH2 concentra¬ tions were determined using synthetic hGRF(l44) for standards, a rabbit anti-hGRF(l-44)OH

(lot no. E08-402-64-4J) as primary antibody (both produced at Eli Lilly and Co.), and 3-125Ilabelled iodotyrosyl10-hGRF(l-44)NH2 (IM.180; Amersham, Arlington Heights, IL, U.S.A.) as tracer. Competing hormones were obtained from Bachem, Inc. (Torrance, CA, U.S.A.; vasoactive intestinal polypeptide (VIP), secretin, glucagon, peptide histidserum

ine isoleucine (PHI), gastric inhibitory polypeptide (GIP) and thyrotrophin-releasing hormone (TRH)), Peninsula Labs (Belmont, CA, U.S.A.; rat GRF(144)OH), and Eli Lilly and Co. (bovine and porcine GRF(l-44)OH). Synthetic hGRF(l-44)NH2 stand¬ ard was dissolved in 0-1 mol acetic acid/1 to a stock concentration of 1 mg/ml. Standards were prepared by diluting this stock solution in buffer (10% (v/v) Dulbecco's phosphate-buffered saline (D-PBS), 10 (Gibco BRL, Grand Island, NY, U.S.A.); 0-03 mol EDTA/1; 0-01% (w/v) merthiolate; 1% bovine serum albumin (BSA); 2% (v/v) normal rabbit serum (NRS) (Gibco BRL); 0-05% Tween 20 (Sigma Chemical Co.) ; pH 7-6). All standards and unknowns

in triplicate. One hundred microlitres of standard or unknown were mixed with 300 pi buffer and 100 pi of a 1:640 dilution (in buffer B) of a stock solution of rabbit anti-hGRF(l-44)OH serum (1:100 in PBS/BSA (buffer without NRS or Tween 20)) in 12 75 mm polypropylene tubes. Tubes were vortexed and incubated at 4 °C overnight. The next day, 100 pi 125I-labelled hGRF(l-44)NH2 (100 000c.p.m./ml in buffer B) were added to all tubes, which were again vortexed and incubated at 4 °C overnight. On the third day, 500 pi of a 1:5 (v/v) mixture of sheep anti-rabbit gamma globulin (Antibodies Inc., Davis, CA, U.S.A. ; first diluted 50% (v/v) in buffer A (same as buffer but without NRS)) and 10% polyethylene glycol (w/v in buffer A) were added to assay tubes. Tubes were vortexed and allowed to incubate at room temperature for 15 min after which they were centrifuged at 1500 g for 15 min, the supernatants decanted and pellets counted for radioactivity. The minimum detectable concentration of the assay was 0-05 ng/ml, concentration at 50% binding was 0-36 ng/ml, and intraassay C.V. was 16-8%. were run

Statistical

analyses

Chickens were blocked by body weight across treat¬ ments, so that the starting mean body weight for each treatment group was equal. Data were analysed according to a randomized complete block design with terms in the model for analysis of variance (AOV) being: Yij U- + B¡ + + eij (i= 1-7, j= 13), where Y¡j plasma GH profile characteristics, organ and skeletal measurements, etc., U popu¬ lation mean, B,= block effect (initial body weight), Tj treatment effect (vehicle control ; low dose GRF ; high dose GRF) and e¡j residual error. All data are expressed as least squares means (LSM) ± standard error of the LSM, except for data on hGRF(l-44) immunoreactivity as described below. Means were separated and differences considered significant based on Bonfferoni's multiple comparison test assuming two pair-wise comparisons (low or high dose GRF versus control). The General Linear Models Proce¬ dure of the Statistical Analysis System (SAS Institute, Inc., 1982) was used to determine all AOV and means separation components, and to derive Pearson Pro¬ duct Moment correlations between GH plasma profile characteristics and pituitary GH mRNA measure¬ ments. Data for serum immunoreactive hGRF(l-44) concentrations displayed heterogeneity of variance and were log transformed prior to AOV, and signifi¬ cant differences (P values) were based on log trans¬ formed data analysis. =

·

=

·

=

=

·

·

=

pg/ml (data not shown). Mean (and range of) serum hGRF(l-44) concentrations for control, low dose and high dose GRF treatments (and probability of a difference versus control) were 3-9 (2-4-5-6)±1-29; 552-6 (145-4-1631-2)±615-93 (P