Serum Growth Hormone-Binding Protein, Insulin-like Growth Factor-I, and Growth Hormone in Patients With Liver Cirrhosis Naoki
Hattori,
Hiroyuki
Kurahachi,
Katsuji Ikekubo, Yasuhiko
Takashi
Ishihara,
Saiki, and Hiroo
Kunisaburo
Moridera,
Megumu
Hino,
lmura
We determined serum growth hormone-binding protein (GHBP), insulin-like growth factor-l (IGF-I), and growth hormone (GH) levels in patients with cirrhosis and in age-matched control subjects, and investigated their relationships. Serum GHBP levels in cirrhotic patients (14.6% f 3.9%) (means f SD) were significantly lower than those in normal subjects (20.4% f 4.7%). GHBP levels had positive correlations with cholinesterase (I = 56, P < .OOl) and Normotest (r = .66, P < .OOl), both of which represent liver function in cirrhotic patients. Basal GH levels in cirrhotic patients (range, 0.35 to 13.0 pg/L; median, 3.9 fig/L) were significantly higher than those in normal subjects (0.015 to 6.0 pg/L; 0.19 Kg/L). GHBP levels in cirrhotic patients correlated positively with IGF-I levels (I = .39, P < .Ol), and negatively with GH levels (r = -.33, P c .Ol). These results may indicate that the serum GHBP level reflects the number of hepatic GH receptors, and that the high basal GH level observed in cirrhotic patients is, at least in part, attributable to decreased clearance of GH by these receptors. Copyrig& 0 1992 by W. B. Saunders Company
H
IGH FASTING LEVELS of serum growth hormone (GH) and abnormal GH responses to various stimuli, such as glucose and thyrotropin-releasing hormone (TRH), have been observed in patients with cirrhosis.‘” Since the liver is a major target organ of GH and is abundant in GH receptors,4 its dysfunction may be accompanied by hormonal imbalance. Growth hormone-binding protein (GHBP) has recently been detected in human serum.5,6 The amino acid sequence identities of GHBP and the hepatic GH receptor have been shown in rabbits,’ rats,8 and mice,9 although they remain to be shown in man. In Laron-type dwarfism, a syndrome characterized by resistance to GH, due to the lack of GH receptors, serum GHBP is absent.“~” These observations suggest that serum GHBP is derived from hepatic GH receptors. However, it is not clear whether serum GHBP levels in man reflect the number of hepatic GH receptors. In the present study, we measured GHBP, insulin-like growth factor-I (IGF-I), and GH levels in patients with cirrhosis, and investigated their relationships to examine this hypothesis.
SUBJECTS AND METHODS
We studied 74 patients with cirrhosis (46 men and 28 women, aged 59.3 2 10.4 years) and 51 normal subjects (30 men and 21 women. aged 58.3 ? 11.7 years). The diagnosis of cirrhosis had been confirmed as postnecrotic in 65 patients, and as alcoholic in nine patients, by clinical, biochemical, and histological findings. All cirrhotic patients were in a compensated state and were not malnourished, and had no clinical evidence of edema or ascites. All subjects had normal renal function and no history of diabetes mellitus. Blood was taken early in the morning after an overnight fast. Biochemical tests were performed with an autoanalyzer. Although cholinesterase is a good indicator of liver function,” it is also influenced by the nutritional state. Therefore, we also performed the Normotest, which reflects clotting factors II, VII, and X, and which is a more precise indicator of liver function,” using a commercially available kit (Eisai, Ltd, Tokyo, Japan). Aliquots of serum were stored at -20°C and thawed before the performance of the GHBP assay within 1 month. Clinical features of the subjects studied are listed in Table 1. The study was approved by the local ethical committee of Kobe City General Hospital. Metabolism,
Vol41, No
4
(April),1992: pp 377-381
GH-Binding
Study
GH was iodinated by the lactoperoxidase method. Specific activity was calculated at approximately 50 pCi/pg. Serum (0.4 mL) was incubated in 0.01 mol/L sodium phosphate buffer (pH 7.0) containing 0.1 mol/L NaCl, 0.1% bovine serum albumin, and 0.1% NaN,, at 37°C for 60 minutes, with ‘zI-labeled GH (15,000 to 20,000 cpm; _ 0.3 to 0.4 ng), in a final volume of 0.5 mL. Free GH was separated from bound GH by an Ultrogel AcA44 (IBF, France) minicolumn (0.9 x 25 cm) at 4°C. Samples were eluted with the same buffer, and 0.5-mL fractions were collected. Nonspecific binding was determined by adding excess unlabeled GH (0.4 *g/IO FL phosphate buffer), and specific binding was calculated by subtracting nonspecific binding from total binding. When the endogenous GH level in a sample was greater than 8 pg/L, the value was corrected based on the saturation curve, as has been described previously.‘4 The majority of the samples did not require correction, as endogenous GH was less than 8 pg/L, which is too low to substantially affect the bound fraction. The coefficient of variation of the GHBP level in the pooled reference serum, on 35 runs, was 6.7%. When the amount of serum was increased by 0.1 to 0.4 mL, the GHBP level in serum from cirrhotic patients was increased in parallel with that from normal subjects. This finding may indicate that the decrease in GHBP level observed in cirrhosis is not due to the presence of serum factors that inhibit the binding of GH to GHBP, or to the deficiency in cirrhotic serum of some factors present in normal serum that enhance GH binding to GHBP. Assays
Serum GH concentrations were measured in duplicate by a highly sensitive enzyme immunoassay (EIA), which has previously been described.j5 In brief, serum samples or GH standards (World Health Organization [WHO] First International Reference Prepa-
From the Depamnents of Endocrinology and Nuclear Medicine, Kobe Ci@ General Hospital, Kobe, Japan; and the Second Division, Department of Medicine, Kyoto UniversityFaculty of Medicine, Kyoto, Japan. Supported in part by the Foundation of Kobe City General Hospital, Kobe, Japan. Address reprint requests to Naoki Hattori, MD, Department of Endocrinology, Kobe City General Hospital, 4-6 Minatojima Nakamachi, Chuo-ku, Kobe 650, Japan. Copyright 0 1992 by W.B. Saunders Company 0026.0495/9214104-0006$03.00/O
377
HATTORI ET AL
378
Table 1. Clinical Features of the Subjects Studied Cirrhosis
No. of subjects (M/F)
NOUlM
74 (46/28)
51 (30/21)
59.3 + 10.4
58.3 + 11.7
Age (vrj Ideal BW (%)
94.8 r 13.7
97.4 r 11.6
Albumin (g/L)
32.0 i 4.9
40.0 + 2.4
Bilirubin (nmol/L)
45.8 k 40.0
11.8 + 5.3
Cholinesterase (A pH)
0.35 + 0.15
0.91 * 0.19
Normotest (%)
56.1 r 16.8
ND
IGF-I fng/Lj
90.0 + 41.6
193.1 ‘- 44.7
NOTE. Values are expressed
as mean
+ SD. Normal value of
Normotest, according to the standard of the kit, is 100 + 30% (22 SD). Abbreviation: ND, not determined.
ration [IRP] 66/217) were incubated with an anti-GH IgG-coated polystyrene ball, at 37°C for 6 hours with continuous shaking. After washing twice with 150 mmol/L saline, the balls were incubated with affinity-purified anti-GH Fab’-peroxidase conjugate, at 4”C, for 16 hours, without shaking, followed by 6 hours of continuous shaking at 20°C. The balls were then washed twice with 150 mmol/L saline, and the activity of the peroxidase bound to the balls was determined by an enzyme reaction with 3-(p-hydroxyphenyl) propionic acid (Nakarai Chemical, Osaka, Japan), for 90 minutes at 30°C. Fluorescence intensity was measured by a spectrofluorophotometer (RF-5000, Shimadzu, Kyoto, Japan). The detection limit for 20 PL serum samples \uas 0.0015 kg/L. The intraassay and interassay coefficients of variation were 6.0% and 8.6%, respectively. Serum IGF-I content was determined by radioimmunoassay (RIA) (Nichols Institute, San Juan Capistrano, CA) after acidethanol extraction. To 0.9 mL of a mixture of 90% ethanol and 10% 1N HCI, 0.1 mL serum was added and mixed for 5 minutes. After centrifugation, the supernatant was neutralized and subjected to RIA immediately. The standard used in the assay was recombinant human IGF-I (supplied from Fujisawa Pharmaceutical Company, Osaka, Japan). The intraassay and interassay coefficients of variation were 6.8% and 7.1%, respectively.
(29.0 2 8.9 kg/L). This may indicate that the decrease in GH binding in cirrhotic serum is not due to a decrease in the affinity of GHBP, but rather to a decrease in binding capacity. Basal GH levels, determined by a highly sensitive EIA, ranged widely in normal subjects, from 0.015 to 6.0 &g/L (median, 0.19 pg/L), whereas that in cirrhotic patients was distributed in a relatively narrow range, with significant elevation (0.35 to 13.0 kg/L, 3.9 *g/L; P < .Ol), as shown in Fig 3. IGF-I levels in cirrhotic patients (90.0 2 41.6 kg/L) were significantly (P < .OOl) lower than those in normal subjects (193.1 2 44.7 kg/L). The relationships of serum GHBP level to liver function tests in cirrhotic patients are shown in Fig 4. GHBP level had a positive correlation with both cholinesterase (r = .58, P < .OOl) and Normotest (r = .66, P < .OOl). There was a positive correlation between GHBP and IGF-I levels in cirrhotic patients (r = .39, P < .Ol), as shown in Fig 5, whereas there was no significant relationship in normal
j-
P
Hattori,
Hiroyuki
Kurahachi,
Katsuji Ikekubo, Yasuhiko
Takashi
Ishihara,
Saiki, and Hiroo
Kunisaburo
Moridera,
Megumu
Hino,
lmura
We determined serum growth hormone-binding protein (GHBP), insulin-like growth factor-l (IGF-I), and growth hormone (GH) levels in patients with cirrhosis and in age-matched control subjects, and investigated their relationships. Serum GHBP levels in cirrhotic patients (14.6% f 3.9%) (means f SD) were significantly lower than those in normal subjects (20.4% f 4.7%). GHBP levels had positive correlations with cholinesterase (I = 56, P < .OOl) and Normotest (r = .66, P < .OOl), both of which represent liver function in cirrhotic patients. Basal GH levels in cirrhotic patients (range, 0.35 to 13.0 pg/L; median, 3.9 fig/L) were significantly higher than those in normal subjects (0.015 to 6.0 pg/L; 0.19 Kg/L). GHBP levels in cirrhotic patients correlated positively with IGF-I levels (I = .39, P < .Ol), and negatively with GH levels (r = -.33, P c .Ol). These results may indicate that the serum GHBP level reflects the number of hepatic GH receptors, and that the high basal GH level observed in cirrhotic patients is, at least in part, attributable to decreased clearance of GH by these receptors. Copyrig& 0 1992 by W. B. Saunders Company
H
IGH FASTING LEVELS of serum growth hormone (GH) and abnormal GH responses to various stimuli, such as glucose and thyrotropin-releasing hormone (TRH), have been observed in patients with cirrhosis.‘” Since the liver is a major target organ of GH and is abundant in GH receptors,4 its dysfunction may be accompanied by hormonal imbalance. Growth hormone-binding protein (GHBP) has recently been detected in human serum.5,6 The amino acid sequence identities of GHBP and the hepatic GH receptor have been shown in rabbits,’ rats,8 and mice,9 although they remain to be shown in man. In Laron-type dwarfism, a syndrome characterized by resistance to GH, due to the lack of GH receptors, serum GHBP is absent.“~” These observations suggest that serum GHBP is derived from hepatic GH receptors. However, it is not clear whether serum GHBP levels in man reflect the number of hepatic GH receptors. In the present study, we measured GHBP, insulin-like growth factor-I (IGF-I), and GH levels in patients with cirrhosis, and investigated their relationships to examine this hypothesis.
SUBJECTS AND METHODS
We studied 74 patients with cirrhosis (46 men and 28 women, aged 59.3 2 10.4 years) and 51 normal subjects (30 men and 21 women. aged 58.3 ? 11.7 years). The diagnosis of cirrhosis had been confirmed as postnecrotic in 65 patients, and as alcoholic in nine patients, by clinical, biochemical, and histological findings. All cirrhotic patients were in a compensated state and were not malnourished, and had no clinical evidence of edema or ascites. All subjects had normal renal function and no history of diabetes mellitus. Blood was taken early in the morning after an overnight fast. Biochemical tests were performed with an autoanalyzer. Although cholinesterase is a good indicator of liver function,” it is also influenced by the nutritional state. Therefore, we also performed the Normotest, which reflects clotting factors II, VII, and X, and which is a more precise indicator of liver function,” using a commercially available kit (Eisai, Ltd, Tokyo, Japan). Aliquots of serum were stored at -20°C and thawed before the performance of the GHBP assay within 1 month. Clinical features of the subjects studied are listed in Table 1. The study was approved by the local ethical committee of Kobe City General Hospital. Metabolism,
Vol41, No
4
(April),1992: pp 377-381
GH-Binding
Study
GH was iodinated by the lactoperoxidase method. Specific activity was calculated at approximately 50 pCi/pg. Serum (0.4 mL) was incubated in 0.01 mol/L sodium phosphate buffer (pH 7.0) containing 0.1 mol/L NaCl, 0.1% bovine serum albumin, and 0.1% NaN,, at 37°C for 60 minutes, with ‘zI-labeled GH (15,000 to 20,000 cpm; _ 0.3 to 0.4 ng), in a final volume of 0.5 mL. Free GH was separated from bound GH by an Ultrogel AcA44 (IBF, France) minicolumn (0.9 x 25 cm) at 4°C. Samples were eluted with the same buffer, and 0.5-mL fractions were collected. Nonspecific binding was determined by adding excess unlabeled GH (0.4 *g/IO FL phosphate buffer), and specific binding was calculated by subtracting nonspecific binding from total binding. When the endogenous GH level in a sample was greater than 8 pg/L, the value was corrected based on the saturation curve, as has been described previously.‘4 The majority of the samples did not require correction, as endogenous GH was less than 8 pg/L, which is too low to substantially affect the bound fraction. The coefficient of variation of the GHBP level in the pooled reference serum, on 35 runs, was 6.7%. When the amount of serum was increased by 0.1 to 0.4 mL, the GHBP level in serum from cirrhotic patients was increased in parallel with that from normal subjects. This finding may indicate that the decrease in GHBP level observed in cirrhosis is not due to the presence of serum factors that inhibit the binding of GH to GHBP, or to the deficiency in cirrhotic serum of some factors present in normal serum that enhance GH binding to GHBP. Assays
Serum GH concentrations were measured in duplicate by a highly sensitive enzyme immunoassay (EIA), which has previously been described.j5 In brief, serum samples or GH standards (World Health Organization [WHO] First International Reference Prepa-
From the Depamnents of Endocrinology and Nuclear Medicine, Kobe Ci@ General Hospital, Kobe, Japan; and the Second Division, Department of Medicine, Kyoto UniversityFaculty of Medicine, Kyoto, Japan. Supported in part by the Foundation of Kobe City General Hospital, Kobe, Japan. Address reprint requests to Naoki Hattori, MD, Department of Endocrinology, Kobe City General Hospital, 4-6 Minatojima Nakamachi, Chuo-ku, Kobe 650, Japan. Copyright 0 1992 by W.B. Saunders Company 0026.0495/9214104-0006$03.00/O
377
HATTORI ET AL
378
Table 1. Clinical Features of the Subjects Studied Cirrhosis
No. of subjects (M/F)
NOUlM
74 (46/28)
51 (30/21)
59.3 + 10.4
58.3 + 11.7
Age (vrj Ideal BW (%)
94.8 r 13.7
97.4 r 11.6
Albumin (g/L)
32.0 i 4.9
40.0 + 2.4
Bilirubin (nmol/L)
45.8 k 40.0
11.8 + 5.3
Cholinesterase (A pH)
0.35 + 0.15
0.91 * 0.19
Normotest (%)
56.1 r 16.8
ND
IGF-I fng/Lj
90.0 + 41.6
193.1 ‘- 44.7
NOTE. Values are expressed
as mean
+ SD. Normal value of
Normotest, according to the standard of the kit, is 100 + 30% (22 SD). Abbreviation: ND, not determined.
ration [IRP] 66/217) were incubated with an anti-GH IgG-coated polystyrene ball, at 37°C for 6 hours with continuous shaking. After washing twice with 150 mmol/L saline, the balls were incubated with affinity-purified anti-GH Fab’-peroxidase conjugate, at 4”C, for 16 hours, without shaking, followed by 6 hours of continuous shaking at 20°C. The balls were then washed twice with 150 mmol/L saline, and the activity of the peroxidase bound to the balls was determined by an enzyme reaction with 3-(p-hydroxyphenyl) propionic acid (Nakarai Chemical, Osaka, Japan), for 90 minutes at 30°C. Fluorescence intensity was measured by a spectrofluorophotometer (RF-5000, Shimadzu, Kyoto, Japan). The detection limit for 20 PL serum samples \uas 0.0015 kg/L. The intraassay and interassay coefficients of variation were 6.0% and 8.6%, respectively. Serum IGF-I content was determined by radioimmunoassay (RIA) (Nichols Institute, San Juan Capistrano, CA) after acidethanol extraction. To 0.9 mL of a mixture of 90% ethanol and 10% 1N HCI, 0.1 mL serum was added and mixed for 5 minutes. After centrifugation, the supernatant was neutralized and subjected to RIA immediately. The standard used in the assay was recombinant human IGF-I (supplied from Fujisawa Pharmaceutical Company, Osaka, Japan). The intraassay and interassay coefficients of variation were 6.8% and 7.1%, respectively.
(29.0 2 8.9 kg/L). This may indicate that the decrease in GH binding in cirrhotic serum is not due to a decrease in the affinity of GHBP, but rather to a decrease in binding capacity. Basal GH levels, determined by a highly sensitive EIA, ranged widely in normal subjects, from 0.015 to 6.0 &g/L (median, 0.19 pg/L), whereas that in cirrhotic patients was distributed in a relatively narrow range, with significant elevation (0.35 to 13.0 kg/L, 3.9 *g/L; P < .Ol), as shown in Fig 3. IGF-I levels in cirrhotic patients (90.0 2 41.6 kg/L) were significantly (P < .OOl) lower than those in normal subjects (193.1 2 44.7 kg/L). The relationships of serum GHBP level to liver function tests in cirrhotic patients are shown in Fig 4. GHBP level had a positive correlation with both cholinesterase (r = .58, P < .OOl) and Normotest (r = .66, P < .OOl). There was a positive correlation between GHBP and IGF-I levels in cirrhotic patients (r = .39, P < .Ol), as shown in Fig 5, whereas there was no significant relationship in normal
j-
P
