0021-972X/92/7406-1355$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1992 by The Endocrine Society
Insulin Binding
Regulation Protein-l
Vol. 74, No. 6 Printed in U.S.A.
of Insulin-Like Growth in Obese and Nonobese
CHERYL A. CONOVER, PHILLIP D. K. LEE, JILL JAY T. CLARKSON, AND MICHAEL D. JENSEN
Factor Humans*
A. KANALEY,
Endocrine ResearchUnit (C.A.C., J.A.K., J. T.C., M.D.J.), Mayo Clinic, Rochester,Minnesota 55905;and Department of Pediatrics (P.D.K.L.), Baylor Collegeof Medicine, Houston, Texas 77030
ABSTRACT. Insulin is the principal regulator of hepatic insulin-like growth factor binding protein-l (IGFBP-1) production, mediating the rapid decrease in plasma IGFBP-1 in response to nutritional intake. In this study, we defined IGFBP-1 regulation by insulin in upper and lower body obesity, conditions associated with insulin resistance and chronic hyperinsulinemia. Overnight postabsorptive IGFBP-1 levels in obese and nonobese women showed an inverse, nonlinear relationship with plasma insulin concentrations. Maximum suppression of IGFBP-1 was seen at 70-90 pmol/L plasma insulin. Both groups of obese women had mean fasting plasma insulin concentrations above this threshold level and, consequently, markedly suppressed IGFBP-1 levels. To assess the dynamics of insulin regulated IGFBP-1, 10 obese and 8 nonobeae women were studied during sequential saline infusion (O-90 min), hyperinsulinemia (insulin infusion; 90-210 min) and hypoinsulinemia (somatostatin + GH infusion; 210-330 min). Insulin infusion rapidly decreased plasma IGFBP-1 levels in nonobese subjects (60% decrease in 2
I
NSULIN-LIKE growth factor (IGF) I and II are GHdependent polypeptides, structurally and functionally related to insulin, with potent anabolic effects in uiuo and in vitro (1). Unlike insulin, the stability, availability, and bioactivity of the IGFs appear to be regulated by specific binding proteins which vary in molecular size, hormonal control, and functional significance (2). One member of this family of IGF binding proteins (BP) is IGFBP-1, a 25kilodalton protein produced by liver, ovarian granulosa cells, and secretory phase or decidualized endometrium (3-7). Insulin is a principal regulator of hepatic IGFBP-1 production (8-13). In normal humans, rapidly fluctuating plasma insulin concentrations in response to meals result in converse changes in plasma IGFBP-1 levels (8, 14,
h), but had little or no further suppressive effect in obese subjects. Complete insulin withdrawal resulted in a significant rise in plasma IGFBP-1 concentrations in all subjects, but the response was blunted in obese compared to nonobese groups. In contrast to nlasma IGFBP-1. IGF-I concentrations did not varv during hyper- and hypoinsulinemic infusion periods and were not significantly different between groups. Basal GH levels were significantly higher in nonobese when compared to obese women, but did not change with infusions. In conclusion, low IGFBP-1 levels in obesity are related to elevated insulin levels which are, in turn, related to body fat distribution and insulin resistance. The chronically depressed levels of IGFBP-1 may promote IGF bioactivity as well as its feedback regulation of GH secretion, thus contributing to the metabolic and mitogenic consequences of obesity. In addition, our findings imply that hepatic insulin sensitivity in terms of IGFBP-1 production is preserved despite peripheral insulin resistance in obesity. (J Clin Endocrinol Metab 74: 1355-1360,1992)
15). Although its precise biological role is still undefined, plasma IGFBP-1 is largely unsaturated and probably accounts for most of the free IGF binding activity in normal adult plasma (16-18). IGFBP-1 sequesters free IGF-I and inhibits its metabolic and mitogenic actions in most in vitro systems studied thus far and a similar in uiuo role has been postulated (19-23). Thus, insulin regulation of hepatic IGFBP-1 production may coordinate insulin and IGF action with nutritional signals. Obesity is associated with insulin resistance and chronic hyperinsulinemia; upper body obesity results in greater insulin resistance than lower body obesity (24). If hepatic IGFBP-1 production in obese individuals remained normally responsive to insulin, chronic suppression of plasma IGFBP-1 concentrations could be expected. In turn, this could potentially result in increased IGF bioavailability and activity inappropriate for nutritional status (e.g. fasting). Previous studies have indicated that fasting plasma IGFBP-1 levels are decreased in obesity (25). We conducted studies under controlled conditions to further define IGFBP-1 regulation by insulin in upper body and lower body obese individuals,
Received August 9, 1991. Address all correspondence and reprint requests to: Dr. Cheryl Conover, Endocrine Research Unit, Mayo Clinic, 5-164 W. Joseph, Rochester, Minnesota 55905. * This work was supported in part by the Mayo Foundation (C.A.C., J.A.K., M.D.J.), National Institutes of Health Grants DK-43258-01 (C.A.C.), DK-40484 (M.D.J.), RR-00585 (G.C.R.C. ofthe Mayo Clinic), and a Feasibility Grant from the American Diabetes Association (P.D.K.L.). 1355
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
CONOVER
1356
and to relate changes in plasma IGFBP-1 the insulin/GH/IGF-I axis. Subjects
to changes in
and Methods
Materials Human regular insulin and recombinant DNA-derived hGH were kindly provided by Eli Lilly and Co. (Indianapolis, IN) and Genentech, Inc. (South San Francisco, CA), respectively. Somatostatin was obtained from Bachem, Inc. (Torrence, CA). Subjects and protocol Informed written consent was obtained from 32 healthy, moderately obese women (ages 36 + 2; body mass index 30-36 kg/m’) and 17 nonobese women (ages 36 + 1; body mass index 22 f 0.4 kg/m*]) who were participating in two separate studies examining in uiuo regulation of lipolysis. Body composition was determined by body potassium counting, tritiated water space, and dual energy x-ray absorptiometry, as described previously (24, 26). Extracellular water was calculated by subtracting intracellular water from total body water. Intracellular water was estimated by assuming an intracellular potassium concentration of 140 mmol/L intracellular water; thus, intracellular water = total body potassium/l40. Fourteen of the obese women had waist:hip ratios (WHR) less than 0.76 (lower body obese, LB-Ob) and 18 had waist:hip ratios greater than 0.85 (upper body obese, UB-Ob). All women were studied during the follicular phase of their menstrual cycle to avoid the variable of secretory endometrium-derived IGFBP-1. Each subject was admitted to the Mayo Clinic General Clinical Research Center the evening before the study, given a standard meal at 1800 h, and then fasted overnight. The following morning at 0730 h, blood samples were collected from 9 nonobese, 10 LB-Ob, and 12 UB-Ob women. For studies on insulin regulation of IGFBP-1, an 18-gauge infusion catheter was placed in a forearm vein in another set of 8 nonobese, 6 LB-Ob, and 6 UB-Ob women; a separate catheter was placed for blood sampling. A saline infusion was started at 0730 h (0 min). After 90 min a primed, continuous infusion of insulin was begun [0.15 mUa kg lean body mass (LBM)-’ emin-’ in nonobese and 0.25 mUa kg LBM-‘amin-’ in obese subjects] and continued until 210 min. Infusions of somatostatin (0.14 pg. kg LBM-’ . min-‘) and GH (5 ng- kg LBM-’ . min-‘) were then administered from 210330 min. From 90-330 min, 50% dextrose was infused as needed to maintain each subject’s plasma glucose concentration at the level observed during the O-90 min saline infusion. The level for each group (mean f SEM) was 5.2 + 0.14 (nonobese), 5.5 f. 0.19 (LB-Ob), and 5.8 + 0.22 (UB-Ob). The study intervals O90 min, 90-210 min, and 210-330 min are subsequently referred to as saline, hyperinsulinemia, and hypoinsulinemia, respectively. Blood was sampled at lo-min intervals from 60-90 min, 180210 min, and 300-330 min, and assayed for insulin, insulin Cpeptide, IGF-I, and GH. IGFBP-1 was assayed in the endpoint sample of each infusion period. Due to sample mishandling, IGFBP-1 assays could not be performed for two of the obese
ET AL.
JCE & M .1992 Vol74.No6
subjects. Plasma glucose concentrations were measured at lomin intervals from 60-330 min to assist in determining the amount of dextrose to infuse in order to maintain euglycemia. Assays IGFBP-I was measured by RIA using a polyclonal rabbit antihuman IGFBP-1 antiserum. The characteristics and specificity of this assay have been previously described (8, 16). Inter- and intraassay coefficients of variation at 7.86 ng/mL were 17.3 + I.4 and 13.4 f 3.8%, respectively, with an assay range of l-200 ng/mL or 0.1-20 rig/tube. IGFBP-3, the major plasma IGF binding protein, does not cross-react in the IGFBP1 RIA (data not shown). Samples from each of the study protocols were grouped to minimize the effect of interassay variability on the data analysis. IGF-I levels were measured by RIA after sample pretreatment using a modification of the procedure described by Daughaday et al. (27) to remove interfering binding proteins. Briefly, 100 PL serum samples were mixed with 400 PL 12.5% 2 N HCl/ 87.5% absolute ethanol, pH 2-3, at room temperature for 30 min. Precipitates were removed by microcentrifugation at 10,000 rpm times 3 min, and supernatant was mixed 1:l with 50 mM Tris, pH 9.5. Samples were then diluted 1:lO in assay buffer (50 mM phosphate-buffered saline, pH 7.4,0.1% Tween20) before assay (1:lOO dilution of original sample). Final pH of the sample was 7.4, and recovery of added [1251]IGF-I was greater than 99% using this extraction procedure. For the assay, 100 ML final sample were mixed with 50 rL IGF-I antiserum UBK487 1:18,000 final titer and 300 CCLassay buffer for 1 h at room temperature. The antiserum UBK487 was provided by Drs. L. E. Underwood and J. J. Van Wyk, Division of Pediatric Endocrinology, University of North Carolina, Chapel Hill, NC and distributed by the National Hormone and Pituitary Program, University of Maryland. Fifty microliters [‘251]thr5gIGFI (Amersham, Arlington Heights, IL) were then added for 16 h at 4 C. Bound and free counts were separated using agaroseimmobilized goat and anti-rabbit immunoglobulin (Bio-Rad). Pure recombinant DNA-derived IGF-I (kindly provided by T. L. Jeatran, Lilly Research Laboratories, Indianapolis, IN) was used for standards. All reported results are from a single assay. Assay range was 0.04-5.81 ng/mL, with intraassay coefficient of variation of 7.8% at 0.48 ng/mL. Plasma insulin, insulin C-peptide, and GH were measured by RIA, as previously described (8, 24). The lower limit of detection in these respective RIAs is 20 pmol/L insulin, 0.01 nmol/L C-peptide, and 0.5 rg/L GH. Plasma glucose concentrations were determined using a glucose analyzer (Yellow Springs Instruments, Yellow Springs, OH). Statistical analyses Results are expressed as mean + SEM. Statistical comparisons between the same study periods among groups were performed using analysis of variance and subsequent nonpaired t test. Comparisons between one study interval and another within the same group were made using a two-tailed paired t test. The relationship between IGFBP-1 and insulin was evaluated using a variety of models. In addition to simple linear
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
INSULIN
REGULATION
regression, several nonlinear models were considered: tial, the sum of two exponentials, log transformation IGFBP-1 and insulin, repeated regression analyses tioned data, and B-spline transformation. “Goodness the different models was compared based on the deviation of the residuals [S,., = Jsum of squared residuals/(n
exponenof both of partiof fit” of standard
- 2)].
The general linear method, nonlinear, and transformed regression procedures of the SAS program (SAS Institute Inc., Cary, NC) were used for linear and nonlinear regressions and curve fitting. General linear method uses the method of least squares to fit general linear models, nonlinear fits nonlinear regressions with a multivariate secant method, and transformed regression extends the general linear model by providing optimal variable transformations, including B-splines, derived by the method of alternating least squares. Natural log transformation of both IGFBP-1 and insulin concentration data minimized the weighting effect of uneven variances over a broad range of concentrations. Statistical significance is defined as P less than 0.05.
Relutiomhip between fasting plasma insulin and IGFBP1 concentrations After an overnight fast, plasma glucose concentrations were similar in nonobese, UB-Ob, and LB-Ob women (5.3 f 0.2,5.7 + 0.1, and 5.6 f 0.2 mmol/L, respectively, P = NS). However, baseline plasma insulin concentrations were greater in UB-Ob (105 + 13 pmol/L) than in LB-Ob women (85 f 11 pmol/L), and both were significantly elevated when compared to nonobese women (42 f 4 pmol/L). Analysis of fasting insulin us. IGFBP-1 concentrations indicated that the best fit was a curvilinear relationship, rather than a simple linear one (Fig. l), confirming that high insulin concentrations have a very different effect on IGFBP-1 response than low concentrations with a marked change in the relationship occurring at 70-90 pmol/L. With insulin concentrations above 70 pmol/L, there was little change in IGFBP-1 levels. Thus, maximum suppression of IGFBP-1 was associated with systemic insulin concentrations of approximately 70 pmol/L. Five of 10 LB-Ob and 10 of 12 UB-Ob women had fasting insulin concentrations above 70 pmol/L, reflecting their insulin-resistant state, and had markedly suppressed IGFBP-1 levels. Data from nonobese subjects insulin
(0.4 mU insulin/kg.min
for 3 h) are
included in the figure, but not in the analysis, to demonstrate the effect of hyperinsulinemia independent of obesity on plasma IGFBP-1 concentrations. Insulin
regulation
IN OBESITY
of IGFBP-1
Eight nonobese and 10 obese subjects were studied during sequential saline (O-90 min), hyperinsulinemic
1357
200 WY)
-y .
-
-1.535
l
h(x)
+
10.08
160
60
40
0 20
Results
infused with
OF IGFBP-1
60
100
140
160
220
260
Insulin (pmol/L) FIG. 1. Relationship betweenplasmainsulinandIGFBP-1 concentra-
tions.Bloodsamples weretakenat 0730h after an overnightfastfrom 9 nonobese (O), 10 LB-Ob (O), and 12 UB-Ob (A) women. (m), Data from 5 nonobese subjects infused with 0.4 mu. kg-’ .min-’ insulin for 3 h.
(90-210 min), and hypoinsulinemic periods (210-330 min). Plasma IGFBP-1, insulin, insulin C-peptide, IGFI, and GH concentrations for this experiment are presented in Table 1; changes in plasma IGFBP-1 concentrations with hyper- and hypoinsulinemic infusions are illustrated in Fig. 2. During insulin infusion (90-210 min), plasma insulin concentrations increased (P < 0.05) in each group of subjects with little or no change in insulin C-peptide levels (Table 1). In nonobese women, the IGFBP-1 levels decreased (P < 0.05) by 60% in 120 min.’ Plasma IGFBP-1 concentrations in the obese women tended to decrease during the insulin infusion, but the difference in IGFBP-1 values between the saline and insulin infusion periods was significant only in the UB-Ob women (Table 1). During the subsequent somatostatin infusion, plasma insulin and insulin C-peptide concentrations decreased to levels at or near the limits of assay sensitivity. With this complete insulin withdrawal, IGFBP-1 levels increased (P < 0.05) in all three 1The mean IGFBP-1 during the saline infusion in the eight healthy subjects in Table 1 was lower than the mean of the corresponding group in Fig. 1 despite no difference in circulating insulin concentrations. We do not have a complete explanation for the apparent discrepancy between the two studies. However, interassay variability of 17% could account for some of the difference. All samples were -run in a single assay; therefore, relative differences in IGFBP-1 values within each experiment are valid.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
1358
CONOVER
TABLE 1. Plasma peptide concentrations means f SEM) Saline (O-90 min) Insulin (pmol/L) Non-Ob LB-Ob UB-Ob C-Peptide (nmol/L) Non-Ob LB-Ob UB-Ob IGFBP-1 (cg/L) Non-Ob LB-Ob UB-Ob IGF-I WL) Non-Ob LB-Ob UB-Ob GH (a/L) Non-Ob LB-Ob UB-Ob
45 f I 71 f 5 121 f 21’sd 0.49 f 0.12 0.38 f 0.02 0.78 f 0.10” 33 + 4 10 f 2’ 4 f 1’
ET AL.
JCE&M.W!Z Vol74.No6
during infusions (results are
HyperlNS
HypolNS I
HyperINS” (90-210 min) 70 It: 8 119 f 6’ 169 f 2W
HypoINS” (210-330 min) 520’ 520’ 520’
. 80
I
Non-Obrre
0
LB-Obar.
n
UB-Oberb
T R
0.30 f 0.04 0.04 f 0.01’ 0.33 f 0.01 0.04 * 0.01’ 0.67 f O.llc*d 0.09 f 0.02+ 13 f 3’ 6fl 2 + 1’9’
78 f 7’ 22 + 2’1’ 27 f 8”
171 f 25 151 f 22 144 f 23
154 f 16 140 f 26 142 & 24
156 f 13 140 f 18 149 f 19
1.0 f 1.4 2.2 f 0.4” 1.4 f 0.4’
3.4 f 1.7 0.8 f 0.1 1.0 + 0.4
2.0 f 0.3 1.0 f 0.1 1.2 f 0.2
Eight nonobese (Non-Ob), 4 LB-Ob, and 6 UB-Ob women were studied during sequential saline (O-90 min), hyperinsulinemic (HyperINS; 90-210 min), and hypoinsulinemic (HypoINS; 210-330 min) infusion periods. Insulin, C-peptide, IGF-I, and GH values for the last 30 min of each infusion period were averaged for each subject. GH concentrations for one LB-Ob subject were below the level of RIA sensitivity and were assigned a value of 0.05 pg/L. IGFBP-1 was measured at the endpoint of each infusion period. ’ HyperINS: Insulin infusion at 0.15 mu. kg LBM-‘a min-’ (nonobese) and 0.25 mU +kg LBM-’ . min-’ (obese). * HypoINS: Somatostatin + GH infusion. ' P < 0.05ct.nonobeee (Non-ob). d P < 0.05 between LB-Ob and UB-Ob. ’ P c 0.05 cf. preceding study period.
groups. In the 120 min following initiation of the hypoinsulinemic state, plasma IGFBP-1 levels in nonobese and obese subjects increased by (range values) 44-102 pg/L and by 5-45 gg/L, respectively (Table 1). From 210-330 min (hypoinsulinemic period), the net increase in plasma IGFBP-1 was 260-400% greater for nonobese compared with obese groups (P < 0.05). The absolute increase in plasma IGFBP-1 was not different between UB-Ob and LB-Ob. In contrast to plasma IGFBP-1 concentrations, IGF-I levels did not vary during the study periods, and were not different between groups (Table 1). Basal GH levels were higher (P C 0.05) in the nonobese when compared to LB-Ob and UB-Ob women, but did not change significantly during hyperinsulinemic infusions. During somatostatin-induced hypoinsulinemia, GH was infused to maintain levels of l-2 pg/L in all three groups. Discussion In this study, we defined the regulation of IGFBP-1 by insulin in obesity-related insulin resistance. The as-
90
210
330
Minutes FIG. 2. Insulin regulation of plasma IGFBP-1 concentrations. Eight nonobese (O), 4 LB-Ob (O), and 6 UB-Ob (A) women were studied during sequential saline (O-90 min), hyperinsulinemic (HyperINS; 90210 min), and hypoinsulinemic (HypoINS; 210-330 min) infusion periods. HyperINS: Insulin infusion at 0.15 mu. LBM-‘emin-’ (nonobese) and 0.25 mU=LBM-‘.min-’ (obese). HypoINS: Somatostatin + GH infusion.
sociation between plasma insulin concentrations and IGFBP-1 was examined in healthy nonobese, LB-Ob, and UB-Ob women in order to evaluate a spectrum of insulin resistance. Our data indicated a nonlinear, inverse relationship between postabsorptive plasma insulin concentrations and IGFBP-1. An apparent maximal suppression of hepatic IGFBP-1 production was obtained with systemic insulin concentrations of approximately 70-90 pmol/L, in agreement with findings from our earlier study on meal-related changes in plasma IGFBP1 and insulin (8). In the present study, 50% of LB-Ob and the majority of UB-Ob women had fasting insulin concentrations above 70 pmol/L and markedly reduced IGFBP-1 levels. Raising insulin concentrations above fasting levels (Le. insulin infusion) had little or no further suppressive effect on already low plasma IGFBP-1 levels in obese subjects but had pronounced effects on nonobese subjects, decreasing plasma IGFBP-1 60% in 2 h. These observations in normal lean subjects are similar to those reported previously showing a disappearance tliz for plasma IGFBP-1 of 60-120 min during meal-related increases in insulin and with euglycemic hyperinsulinemic clamp (8, 10, 28, 29). Our findings support the proposal
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
INSULIN
REGULATION
that insulin is the major regulator of hepatic IGFBP-1 production (8-13), and strongly suggest that the insulin resistance of obesity, best characterized with respect to insulin’s regulation of fuel (glucose, fatty acid, and amino acid) metabolism (24, 30), does not extend to its effects on IGFBP-1 production. Since IGFBP-1 is synthesized primarily by the liver (4-6,12,13), IGFBP-1 response to insulin may provide a marker of hepatic sensitivity to this hormone in ho. This sensitivity would be comparable to that of insulin regulation of adipose tissue lipolysis (31) and is 4-5 times more sensitive than insulin suppression of hepatic glucose production (32). In agreement with this notion, Brismar et al. (33) recently reported a highly significant inverse correlation between plasma IGFBP-1 and insulin secretion in healthy subjects. Thus, the low fasting IGFBP-1 levels in obesity are appropriate for the relatively high ambient insulin concentrations, as has been noted in other subject groups with insulin resistance (25, 28, 34). However, insulin response patterns were not fully investigated in these previous studies. Furthermore, our hypoinsulinemic infusion studies suggested differences between lean and obese populations in hepatic IGFBP-1 responsiveness to reduced insulin concentrations. Nonobese subjects had a brisk response to insulin withdrawal with increases in IGFBP-1 concentrations of 65 pg/L in 120 min. This agrees with our previous studies demonstrating a rate of increase in plasma IGFBP-1 concentrations of 0.5 pg/Lmin in normal subjects 3-4 h postprandially (8) and under controlled hypoinsulinemic conditions (Lee, P.D.K., et al., submitted manuscript). In comparison, IGFBP-1 concentrations increased approximately 25 pg/ L in 120 min in obese individuals following complete insulin withdrawal; the range in these subjects was 5-45 pg/L. This apparent blunting of IGFBP-1 in response to reduced insulin concentrations in obesity cannot be explained solely by differences in body mass between lean and obese women. Although the volume of distribution for IGFBP-1 is unknown, this protein is present in plasma and lymph (35), and therefore can be assumed to distribute with extracellular water. The 33 + 3% increase in extracellular water in obese individuals from these studies (data not shown) cannot totally account for the 3- to 4-fold difference in plasma IGFBP-1 response. Possible explanations include impaired hepatic IGFBP1 gene expression and/or protein synthesis induced by the chronic hyperinsulinemia or a delayed rise in IGFBP1 due to elevated portal insulin concentrations (8). In addition, a differential effect of insulin on IGFBP-1 clearance in obese subjects cannot be excluded from these data. We also found that GH levels were reduced in obese women while “GH-dependent” IGF-I levels were not
OF IGFBP-1
IN OBESITY
1359
significantly different between groups. The discordant relationship between GH and IGF-I in obesity may be explained by insulin stimulation of IGF-I production with resultant feedback suppression of GH (36). A decrease in circulating IGFBP-1 could promote available IGF-I for direct inhibition of somatotroph GH secretion. Free IGF peptides have been shown to inhibit GH expression and release in uiuo and in vitro (37, 38). Thus, chronic suppression of IGFBP-1 could contribute to “good growth without GH” in obesity (36). Although the role of insulin-mediated suppression of IGFBP-1 is uncertain, IGFBP-1 inhibits IGF-stimulated cell metabolism in vitro (19-23), and a similar action in uiuo has been proposed. One to 10% of plasma IGF-I circulates as free bioactive peptide (39,40). Thus, during normal daily fluctuations in nutritional status, insulin regulation of hepatic IGFBP-1 production could coordinate insulin and IGF-I action at the cellular level. Longterm overnutrition and hyperinsulinemia of obesity may alter this regulated growth response by persistent marked suppression of hepatic IGFBP-1 production. Given the normal and relatively stable IGF-I levels, it is interesting to speculate that chronically depressed levels of IGFBP1 accentuate the growth promoting effects of IGF-I as well as its feedback regulation of GH secretion. Additional studies will be needed to clarify the contribution of reduced plasma IGFBP-1 to the metabolic and mitogenie abberations associated with obesity. Acknowledgments The authors gratefully acknowledge the staff of the Mayo Clinic Clinical Research Center for excellent technical support, and Dr. Peter O’Brien of the Department of Biostatistics for his assistance with the statistical analyses.
References 1. Froesch ER, Schmid C, Schwander J, Zapf JA. Actions of insulinlike growth factors. Annu Rev Physiol. 1985;47:443-67. 2. Baxter RC, Martin JL. Binding proteins for the insulin-like growth factors: structure, regulation and function. Prog Growth Factor Res. 1989;1:49-68. 3. Drop SLS, Hintz RL. Introduction. On the nomenclature of the IGF binding proteins. In: Drop SLS, Hintz RL (eds) Insulin-Like Growth Factor Binding Proteins. Amsterdam: Excerpta Medica; 1989:v-vii. 4. Lee YL, Hintz RL, James PM, Lee PDK, Shively JE, Powell DR. Insulin-like growth factor (IGF) binding protein complementary deoxvribonucleic acid from human HEP Gz henatoma cells: uredieted protein sequence suggests an IGF binding domain different from those of the IGF-I and IGF-II receptors. Mol Endocrinol. 1988,2:404-11. 5. Brinkman A, Groffen C, Kortleve DJ, Geurta van Kessel A, Drop SLS. Isolation and characterization of a cDNA encoding the low molecular weight insulin-like growth factor binding protein (IBP1). EMBO J. 196&7:2417-23. 6. Julkunen M, Koistinen R, Aalto-Setala K, Seppala M, Janne OA, Kontula K. Primary structure of human insulin-like growth factor binding protein/placental protein 12 and tissue-specific expression of its mRNA. FEBS Lett. 196&236:295-302.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
1360
CONOVER
7. Suikkari AM, Jalkanen J, Koistinen R, et al. Human granulosa cells synthesize low molecular weight insulin-like growth factorbinding protein. Endocrinology. 1989;124:1088-90. 8. Conover CA, Butler PC, Wang M, Rizza RA, Lee PDK. Lack of growth hormone effect on insulin-associated suppression of insulin-like growth factor binding protein 1 in humans. Diabetes. 1990;39:1251-6. 9. Suikkari AM. Koivisto VA. Rutanen EM. Yki-Jarvinen H. Karonen SL, Se&ala M. Insulin regulates the serum levels ‘of low molecular weight insulin-like growth factor binding proteins. J Clin Endocrinol Metab. 1988:66:266-72. 10. Brismar K, Gutniak M, Povoa G, Werner S, Hall K. Insulin regulates the 35 kDa binding protein in patients with diabetes mellitus. J Endocrinol Invest. 1988:11:599-602. 11. Holly JMP, Biddlecombe RA, Dun&r DB, et al. Circadian variation of GH-independent IGF-binding protein in diabetes mellitus and ita relationshin to insulin. A new role for insulin? Clin Endocrinol (Oxf). 1988;29’667-75. 12. Conover CA, Lee PDK. Insulin regulation of IGF binding protein production in cultured HepGn cells. J Clin Endocrinol Metab. 1990;70:1062-7. 13. Lewitt MS, Baxter RC. Regulation of growth hormone-independent insulin-like growth factor-binding protein (BP-28) in cultured human fetal liver explants. J Clin Endocrinol Metab. 1989;69:24652. 14. /Cotterill AM, Cowell CT, Baxter RC, McNeil D, Silinik M. Regulation of the growth hormone-independent growth factor binding protein in children. J Clin Endocrinol Metab. 1988;67:882-7. 15. Busby WH, Snyder DK, Clemmons DR. Radioimmunoassay of a 26,OOOdalton plasma insulin-like growth factor binding protein: control by nutritional factors. J Clin Endocrinol Metab. 1988,67:1225-30. 16. Lee PDK, Hintz RL, Sperry JB, Baxter RC, Powell DR. Serum insulin-like growth factor binding proteins in growth-retarded children with chronic renal failure. Pediatr Res. i989;36:308-15. 17. Baxter RC. Cowell CT. Diurnal rhvthm of arowth hormone independent binding protein for insulin-like growth factors in human plasma. J Clin Endocrinol Metab. 1987;65:432-40. 18. Hintz RL, Liu F, Rosenfeld RG, Kemp SF. Plasma somatomedinbinding proteins in hypopituitarism: changes during growth hormone therapy. J Clin Endocrinol Metab. 1981;53:100-4. 19. Rutanen EM, Pekonen F, Makinen T. Soluble 34K binding protein inhibits the binding of insulin-like growth factor I to its cell receptors in human secretory phase endometrium: Evidence for autocrine/paracrine regulation of growth factor action. J Clin Endocrinol Metab. 1988;66:173-80. 20. Ritvos 0, Ranta T, Jalkanen J, et al. Insulin-like growth factor (IGF) binding protein from human decidua inhibits the binding and biological action of IGF-I in cultured choriocarcinoma cells. Endocrinology. 1988;122:2150-7. 21. Conover CA, Ronk M, Lombana F, Powell DR. Structural and biological characterization of bovine insulin-like growth factor binding protein-3. Endocrinology. 1990;127:2795-803. 22. Burch WM, Correa J, Shively JE, Powell DR. The 25-kilodalton insulin-like growth factor (IGF)-binding protein inhibits both basal and IGF-I-mediated growth of chick embryo cartilage in vitro. J Clin Endocrinol Metab. 1990;70:173-80. 23. Liu L, Brinkman A, Blat C, Hare1 L. IGFBP-1, an insulin-like growth factor binding protein, is a cell growth inhibitor. Biochem Biophys R.es Commun. 1991;174:673-9.
ET AL.
JCE&M.1992 Vol74’No6
24. Jensen MD. Regulation of forearm lipolysis in different types of obesity: In viva evidence for adipocyte heterogeneity. J Clin Invest. 1991;87:187-93. 25. Weaver JU, Holly JMP, Kopelman PG, et al. Decreased sex hormone binding globulin (SHBG) and insulin-like growth factor binding protein (IGFBP-1) in extreme obesity. Clin Endocrinol. 1990;33:415-22. 26. Jensen MD, Braun JS, Vetter RJ, Marsh HM. Measurement of body potassium with a whole-body counter: Relationship between lean body mass and resting energy expenditure. Mayo Clin Proc. 1988,63:864-8. 27. Daughaday WH, Mariz IK, Blethen SL. Inhibition of access of bound somatomedin to membrane receptor and immunobinding sites: A comparison of radioreceptor and radioimmunoassay of somatomedin in native and acid-ethanol-extracted serum. J Clin Endocrinol Metab. 1980;51:781-8. 28. Suikkari AM, Koivisto VA, Koistinen R, Seppala M, Yki-Jarvinen H. Dose-response characteristics for sunnression of low molecular weight plasma insulin-like growth factor-binding protein by insulin. J Clin Endocrinol Metab. 1989:68:135-40. 29. Yeoh SI, Baxter RC. Metabolic regulation of the growth hormone independent insulin-like growth factor binding protein in human plasma. Acta Endocrinol. 1988:119:465-73. 30. Groop LC, Saloranta C, Shank M, Bonadonna RC, Ferrannini E, DeFronzo RA. The role of free fatty acid metabolism in the pathogenesis of insulin resistance in obesity and noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab. 1991;72:96107. 31. Jensen MD, Caruso M, Heiling V, Miles JM. Insulin regulation of lipolysis in nondiabetic and IDDM subjects. Diabetes. 1989;38:1595-1601. 32. Rizza RA, Mandarin0 W, Gerich JE. Dose-response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol. 1981;240:E630-9. 33. Brismar K, Grill V, Efendic S, Hall K. The insulin-lie growth factor binding protein-l in low and high insulin responders before and during dexamethasone treatment. Metabolism. 1991;40:72832. 34. Conway_ GS. Jacobs HS. Hollv JMP. Wass JAH. Effects of luteinizing hormone, insulin, ins&n-like growth factor-I and insulinlike growth factor small binding protein 1 in the polycystic ovary syndrome. Clin Endocrinol. 1990:33:593-603. 35. Binoux M, Hossenlopp P. Insulin-like growth factor (IGF) and IGF-binding proteins: comparison of human serum and lymph. J Clin Endocrinol Metab. 1988;67:509-14. 36. Phillips LS, Harp JB, Goldstein S, Klein J, Pao C-I. Regulation and action of insulin-like growth factors at the cellular level. Proc Nutr Sot. 1990,49:451-8. 37. Snencer GSG. Macdonald AA. Buttle HL. Moore LG.? Carlvle --.-- SS. --~ Intracerebral ‘administration of insulin-like growth factor I decreases circulating growth hormone levels in the fetal pig. Acta Endocrinol. 1991;124:563-8. 38. Yamasaki H, Prager D, Gebremedhin S, Moise L, Melmed S. Binding and action of insulin-like growth factor I in pituitary tumor cells. Endocrinologv. 1991:128:857-62. 39. Guler HP, Zapf J, Froe&h ER.. Short-term metabolic effects of recombinant human insulin-like growth factor I in healthy adults. N End J Med. 1987:317:137-40. 40. Mukku VR. A 96-well microtiter plate assay for free IGF-I in ulasma usine immobilized IGFBP-3. Presented at The Endocrine Society, Waihington DC, June 19-22,199l (Abstract).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
Insulin Binding
Regulation Protein-l
Vol. 74, No. 6 Printed in U.S.A.
of Insulin-Like Growth in Obese and Nonobese
CHERYL A. CONOVER, PHILLIP D. K. LEE, JILL JAY T. CLARKSON, AND MICHAEL D. JENSEN
Factor Humans*
A. KANALEY,
Endocrine ResearchUnit (C.A.C., J.A.K., J. T.C., M.D.J.), Mayo Clinic, Rochester,Minnesota 55905;and Department of Pediatrics (P.D.K.L.), Baylor Collegeof Medicine, Houston, Texas 77030
ABSTRACT. Insulin is the principal regulator of hepatic insulin-like growth factor binding protein-l (IGFBP-1) production, mediating the rapid decrease in plasma IGFBP-1 in response to nutritional intake. In this study, we defined IGFBP-1 regulation by insulin in upper and lower body obesity, conditions associated with insulin resistance and chronic hyperinsulinemia. Overnight postabsorptive IGFBP-1 levels in obese and nonobese women showed an inverse, nonlinear relationship with plasma insulin concentrations. Maximum suppression of IGFBP-1 was seen at 70-90 pmol/L plasma insulin. Both groups of obese women had mean fasting plasma insulin concentrations above this threshold level and, consequently, markedly suppressed IGFBP-1 levels. To assess the dynamics of insulin regulated IGFBP-1, 10 obese and 8 nonobeae women were studied during sequential saline infusion (O-90 min), hyperinsulinemia (insulin infusion; 90-210 min) and hypoinsulinemia (somatostatin + GH infusion; 210-330 min). Insulin infusion rapidly decreased plasma IGFBP-1 levels in nonobese subjects (60% decrease in 2
I
NSULIN-LIKE growth factor (IGF) I and II are GHdependent polypeptides, structurally and functionally related to insulin, with potent anabolic effects in uiuo and in vitro (1). Unlike insulin, the stability, availability, and bioactivity of the IGFs appear to be regulated by specific binding proteins which vary in molecular size, hormonal control, and functional significance (2). One member of this family of IGF binding proteins (BP) is IGFBP-1, a 25kilodalton protein produced by liver, ovarian granulosa cells, and secretory phase or decidualized endometrium (3-7). Insulin is a principal regulator of hepatic IGFBP-1 production (8-13). In normal humans, rapidly fluctuating plasma insulin concentrations in response to meals result in converse changes in plasma IGFBP-1 levels (8, 14,
h), but had little or no further suppressive effect in obese subjects. Complete insulin withdrawal resulted in a significant rise in plasma IGFBP-1 concentrations in all subjects, but the response was blunted in obese compared to nonobese groups. In contrast to nlasma IGFBP-1. IGF-I concentrations did not varv during hyper- and hypoinsulinemic infusion periods and were not significantly different between groups. Basal GH levels were significantly higher in nonobese when compared to obese women, but did not change with infusions. In conclusion, low IGFBP-1 levels in obesity are related to elevated insulin levels which are, in turn, related to body fat distribution and insulin resistance. The chronically depressed levels of IGFBP-1 may promote IGF bioactivity as well as its feedback regulation of GH secretion, thus contributing to the metabolic and mitogenic consequences of obesity. In addition, our findings imply that hepatic insulin sensitivity in terms of IGFBP-1 production is preserved despite peripheral insulin resistance in obesity. (J Clin Endocrinol Metab 74: 1355-1360,1992)
15). Although its precise biological role is still undefined, plasma IGFBP-1 is largely unsaturated and probably accounts for most of the free IGF binding activity in normal adult plasma (16-18). IGFBP-1 sequesters free IGF-I and inhibits its metabolic and mitogenic actions in most in vitro systems studied thus far and a similar in uiuo role has been postulated (19-23). Thus, insulin regulation of hepatic IGFBP-1 production may coordinate insulin and IGF action with nutritional signals. Obesity is associated with insulin resistance and chronic hyperinsulinemia; upper body obesity results in greater insulin resistance than lower body obesity (24). If hepatic IGFBP-1 production in obese individuals remained normally responsive to insulin, chronic suppression of plasma IGFBP-1 concentrations could be expected. In turn, this could potentially result in increased IGF bioavailability and activity inappropriate for nutritional status (e.g. fasting). Previous studies have indicated that fasting plasma IGFBP-1 levels are decreased in obesity (25). We conducted studies under controlled conditions to further define IGFBP-1 regulation by insulin in upper body and lower body obese individuals,
Received August 9, 1991. Address all correspondence and reprint requests to: Dr. Cheryl Conover, Endocrine Research Unit, Mayo Clinic, 5-164 W. Joseph, Rochester, Minnesota 55905. * This work was supported in part by the Mayo Foundation (C.A.C., J.A.K., M.D.J.), National Institutes of Health Grants DK-43258-01 (C.A.C.), DK-40484 (M.D.J.), RR-00585 (G.C.R.C. ofthe Mayo Clinic), and a Feasibility Grant from the American Diabetes Association (P.D.K.L.). 1355
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
CONOVER
1356
and to relate changes in plasma IGFBP-1 the insulin/GH/IGF-I axis. Subjects
to changes in
and Methods
Materials Human regular insulin and recombinant DNA-derived hGH were kindly provided by Eli Lilly and Co. (Indianapolis, IN) and Genentech, Inc. (South San Francisco, CA), respectively. Somatostatin was obtained from Bachem, Inc. (Torrence, CA). Subjects and protocol Informed written consent was obtained from 32 healthy, moderately obese women (ages 36 + 2; body mass index 30-36 kg/m’) and 17 nonobese women (ages 36 + 1; body mass index 22 f 0.4 kg/m*]) who were participating in two separate studies examining in uiuo regulation of lipolysis. Body composition was determined by body potassium counting, tritiated water space, and dual energy x-ray absorptiometry, as described previously (24, 26). Extracellular water was calculated by subtracting intracellular water from total body water. Intracellular water was estimated by assuming an intracellular potassium concentration of 140 mmol/L intracellular water; thus, intracellular water = total body potassium/l40. Fourteen of the obese women had waist:hip ratios (WHR) less than 0.76 (lower body obese, LB-Ob) and 18 had waist:hip ratios greater than 0.85 (upper body obese, UB-Ob). All women were studied during the follicular phase of their menstrual cycle to avoid the variable of secretory endometrium-derived IGFBP-1. Each subject was admitted to the Mayo Clinic General Clinical Research Center the evening before the study, given a standard meal at 1800 h, and then fasted overnight. The following morning at 0730 h, blood samples were collected from 9 nonobese, 10 LB-Ob, and 12 UB-Ob women. For studies on insulin regulation of IGFBP-1, an 18-gauge infusion catheter was placed in a forearm vein in another set of 8 nonobese, 6 LB-Ob, and 6 UB-Ob women; a separate catheter was placed for blood sampling. A saline infusion was started at 0730 h (0 min). After 90 min a primed, continuous infusion of insulin was begun [0.15 mUa kg lean body mass (LBM)-’ emin-’ in nonobese and 0.25 mUa kg LBM-‘amin-’ in obese subjects] and continued until 210 min. Infusions of somatostatin (0.14 pg. kg LBM-’ . min-‘) and GH (5 ng- kg LBM-’ . min-‘) were then administered from 210330 min. From 90-330 min, 50% dextrose was infused as needed to maintain each subject’s plasma glucose concentration at the level observed during the O-90 min saline infusion. The level for each group (mean f SEM) was 5.2 + 0.14 (nonobese), 5.5 f. 0.19 (LB-Ob), and 5.8 + 0.22 (UB-Ob). The study intervals O90 min, 90-210 min, and 210-330 min are subsequently referred to as saline, hyperinsulinemia, and hypoinsulinemia, respectively. Blood was sampled at lo-min intervals from 60-90 min, 180210 min, and 300-330 min, and assayed for insulin, insulin Cpeptide, IGF-I, and GH. IGFBP-1 was assayed in the endpoint sample of each infusion period. Due to sample mishandling, IGFBP-1 assays could not be performed for two of the obese
ET AL.
JCE & M .1992 Vol74.No6
subjects. Plasma glucose concentrations were measured at lomin intervals from 60-330 min to assist in determining the amount of dextrose to infuse in order to maintain euglycemia. Assays IGFBP-I was measured by RIA using a polyclonal rabbit antihuman IGFBP-1 antiserum. The characteristics and specificity of this assay have been previously described (8, 16). Inter- and intraassay coefficients of variation at 7.86 ng/mL were 17.3 + I.4 and 13.4 f 3.8%, respectively, with an assay range of l-200 ng/mL or 0.1-20 rig/tube. IGFBP-3, the major plasma IGF binding protein, does not cross-react in the IGFBP1 RIA (data not shown). Samples from each of the study protocols were grouped to minimize the effect of interassay variability on the data analysis. IGF-I levels were measured by RIA after sample pretreatment using a modification of the procedure described by Daughaday et al. (27) to remove interfering binding proteins. Briefly, 100 PL serum samples were mixed with 400 PL 12.5% 2 N HCl/ 87.5% absolute ethanol, pH 2-3, at room temperature for 30 min. Precipitates were removed by microcentrifugation at 10,000 rpm times 3 min, and supernatant was mixed 1:l with 50 mM Tris, pH 9.5. Samples were then diluted 1:lO in assay buffer (50 mM phosphate-buffered saline, pH 7.4,0.1% Tween20) before assay (1:lOO dilution of original sample). Final pH of the sample was 7.4, and recovery of added [1251]IGF-I was greater than 99% using this extraction procedure. For the assay, 100 ML final sample were mixed with 50 rL IGF-I antiserum UBK487 1:18,000 final titer and 300 CCLassay buffer for 1 h at room temperature. The antiserum UBK487 was provided by Drs. L. E. Underwood and J. J. Van Wyk, Division of Pediatric Endocrinology, University of North Carolina, Chapel Hill, NC and distributed by the National Hormone and Pituitary Program, University of Maryland. Fifty microliters [‘251]thr5gIGFI (Amersham, Arlington Heights, IL) were then added for 16 h at 4 C. Bound and free counts were separated using agaroseimmobilized goat and anti-rabbit immunoglobulin (Bio-Rad). Pure recombinant DNA-derived IGF-I (kindly provided by T. L. Jeatran, Lilly Research Laboratories, Indianapolis, IN) was used for standards. All reported results are from a single assay. Assay range was 0.04-5.81 ng/mL, with intraassay coefficient of variation of 7.8% at 0.48 ng/mL. Plasma insulin, insulin C-peptide, and GH were measured by RIA, as previously described (8, 24). The lower limit of detection in these respective RIAs is 20 pmol/L insulin, 0.01 nmol/L C-peptide, and 0.5 rg/L GH. Plasma glucose concentrations were determined using a glucose analyzer (Yellow Springs Instruments, Yellow Springs, OH). Statistical analyses Results are expressed as mean + SEM. Statistical comparisons between the same study periods among groups were performed using analysis of variance and subsequent nonpaired t test. Comparisons between one study interval and another within the same group were made using a two-tailed paired t test. The relationship between IGFBP-1 and insulin was evaluated using a variety of models. In addition to simple linear
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
INSULIN
REGULATION
regression, several nonlinear models were considered: tial, the sum of two exponentials, log transformation IGFBP-1 and insulin, repeated regression analyses tioned data, and B-spline transformation. “Goodness the different models was compared based on the deviation of the residuals [S,., = Jsum of squared residuals/(n
exponenof both of partiof fit” of standard
- 2)].
The general linear method, nonlinear, and transformed regression procedures of the SAS program (SAS Institute Inc., Cary, NC) were used for linear and nonlinear regressions and curve fitting. General linear method uses the method of least squares to fit general linear models, nonlinear fits nonlinear regressions with a multivariate secant method, and transformed regression extends the general linear model by providing optimal variable transformations, including B-splines, derived by the method of alternating least squares. Natural log transformation of both IGFBP-1 and insulin concentration data minimized the weighting effect of uneven variances over a broad range of concentrations. Statistical significance is defined as P less than 0.05.
Relutiomhip between fasting plasma insulin and IGFBP1 concentrations After an overnight fast, plasma glucose concentrations were similar in nonobese, UB-Ob, and LB-Ob women (5.3 f 0.2,5.7 + 0.1, and 5.6 f 0.2 mmol/L, respectively, P = NS). However, baseline plasma insulin concentrations were greater in UB-Ob (105 + 13 pmol/L) than in LB-Ob women (85 f 11 pmol/L), and both were significantly elevated when compared to nonobese women (42 f 4 pmol/L). Analysis of fasting insulin us. IGFBP-1 concentrations indicated that the best fit was a curvilinear relationship, rather than a simple linear one (Fig. l), confirming that high insulin concentrations have a very different effect on IGFBP-1 response than low concentrations with a marked change in the relationship occurring at 70-90 pmol/L. With insulin concentrations above 70 pmol/L, there was little change in IGFBP-1 levels. Thus, maximum suppression of IGFBP-1 was associated with systemic insulin concentrations of approximately 70 pmol/L. Five of 10 LB-Ob and 10 of 12 UB-Ob women had fasting insulin concentrations above 70 pmol/L, reflecting their insulin-resistant state, and had markedly suppressed IGFBP-1 levels. Data from nonobese subjects insulin
(0.4 mU insulin/kg.min
for 3 h) are
included in the figure, but not in the analysis, to demonstrate the effect of hyperinsulinemia independent of obesity on plasma IGFBP-1 concentrations. Insulin
regulation
IN OBESITY
of IGFBP-1
Eight nonobese and 10 obese subjects were studied during sequential saline (O-90 min), hyperinsulinemic
1357
200 WY)
-y .
-
-1.535
l
h(x)
+
10.08
160
60
40
0 20
Results
infused with
OF IGFBP-1
60
100
140
160
220
260
Insulin (pmol/L) FIG. 1. Relationship betweenplasmainsulinandIGFBP-1 concentra-
tions.Bloodsamples weretakenat 0730h after an overnightfastfrom 9 nonobese (O), 10 LB-Ob (O), and 12 UB-Ob (A) women. (m), Data from 5 nonobese subjects infused with 0.4 mu. kg-’ .min-’ insulin for 3 h.
(90-210 min), and hypoinsulinemic periods (210-330 min). Plasma IGFBP-1, insulin, insulin C-peptide, IGFI, and GH concentrations for this experiment are presented in Table 1; changes in plasma IGFBP-1 concentrations with hyper- and hypoinsulinemic infusions are illustrated in Fig. 2. During insulin infusion (90-210 min), plasma insulin concentrations increased (P < 0.05) in each group of subjects with little or no change in insulin C-peptide levels (Table 1). In nonobese women, the IGFBP-1 levels decreased (P < 0.05) by 60% in 120 min.’ Plasma IGFBP-1 concentrations in the obese women tended to decrease during the insulin infusion, but the difference in IGFBP-1 values between the saline and insulin infusion periods was significant only in the UB-Ob women (Table 1). During the subsequent somatostatin infusion, plasma insulin and insulin C-peptide concentrations decreased to levels at or near the limits of assay sensitivity. With this complete insulin withdrawal, IGFBP-1 levels increased (P < 0.05) in all three 1The mean IGFBP-1 during the saline infusion in the eight healthy subjects in Table 1 was lower than the mean of the corresponding group in Fig. 1 despite no difference in circulating insulin concentrations. We do not have a complete explanation for the apparent discrepancy between the two studies. However, interassay variability of 17% could account for some of the difference. All samples were -run in a single assay; therefore, relative differences in IGFBP-1 values within each experiment are valid.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
1358
CONOVER
TABLE 1. Plasma peptide concentrations means f SEM) Saline (O-90 min) Insulin (pmol/L) Non-Ob LB-Ob UB-Ob C-Peptide (nmol/L) Non-Ob LB-Ob UB-Ob IGFBP-1 (cg/L) Non-Ob LB-Ob UB-Ob IGF-I WL) Non-Ob LB-Ob UB-Ob GH (a/L) Non-Ob LB-Ob UB-Ob
45 f I 71 f 5 121 f 21’sd 0.49 f 0.12 0.38 f 0.02 0.78 f 0.10” 33 + 4 10 f 2’ 4 f 1’
ET AL.
JCE&M.W!Z Vol74.No6
during infusions (results are
HyperlNS
HypolNS I
HyperINS” (90-210 min) 70 It: 8 119 f 6’ 169 f 2W
HypoINS” (210-330 min) 520’ 520’ 520’
. 80
I
Non-Obrre
0
LB-Obar.
n
UB-Oberb
T R
0.30 f 0.04 0.04 f 0.01’ 0.33 f 0.01 0.04 * 0.01’ 0.67 f O.llc*d 0.09 f 0.02+ 13 f 3’ 6fl 2 + 1’9’
78 f 7’ 22 + 2’1’ 27 f 8”
171 f 25 151 f 22 144 f 23
154 f 16 140 f 26 142 & 24
156 f 13 140 f 18 149 f 19
1.0 f 1.4 2.2 f 0.4” 1.4 f 0.4’
3.4 f 1.7 0.8 f 0.1 1.0 + 0.4
2.0 f 0.3 1.0 f 0.1 1.2 f 0.2
Eight nonobese (Non-Ob), 4 LB-Ob, and 6 UB-Ob women were studied during sequential saline (O-90 min), hyperinsulinemic (HyperINS; 90-210 min), and hypoinsulinemic (HypoINS; 210-330 min) infusion periods. Insulin, C-peptide, IGF-I, and GH values for the last 30 min of each infusion period were averaged for each subject. GH concentrations for one LB-Ob subject were below the level of RIA sensitivity and were assigned a value of 0.05 pg/L. IGFBP-1 was measured at the endpoint of each infusion period. ’ HyperINS: Insulin infusion at 0.15 mu. kg LBM-‘a min-’ (nonobese) and 0.25 mU +kg LBM-’ . min-’ (obese). * HypoINS: Somatostatin + GH infusion. ' P < 0.05ct.nonobeee (Non-ob). d P < 0.05 between LB-Ob and UB-Ob. ’ P c 0.05 cf. preceding study period.
groups. In the 120 min following initiation of the hypoinsulinemic state, plasma IGFBP-1 levels in nonobese and obese subjects increased by (range values) 44-102 pg/L and by 5-45 gg/L, respectively (Table 1). From 210-330 min (hypoinsulinemic period), the net increase in plasma IGFBP-1 was 260-400% greater for nonobese compared with obese groups (P < 0.05). The absolute increase in plasma IGFBP-1 was not different between UB-Ob and LB-Ob. In contrast to plasma IGFBP-1 concentrations, IGF-I levels did not vary during the study periods, and were not different between groups (Table 1). Basal GH levels were higher (P C 0.05) in the nonobese when compared to LB-Ob and UB-Ob women, but did not change significantly during hyperinsulinemic infusions. During somatostatin-induced hypoinsulinemia, GH was infused to maintain levels of l-2 pg/L in all three groups. Discussion In this study, we defined the regulation of IGFBP-1 by insulin in obesity-related insulin resistance. The as-
90
210
330
Minutes FIG. 2. Insulin regulation of plasma IGFBP-1 concentrations. Eight nonobese (O), 4 LB-Ob (O), and 6 UB-Ob (A) women were studied during sequential saline (O-90 min), hyperinsulinemic (HyperINS; 90210 min), and hypoinsulinemic (HypoINS; 210-330 min) infusion periods. HyperINS: Insulin infusion at 0.15 mu. LBM-‘emin-’ (nonobese) and 0.25 mU=LBM-‘.min-’ (obese). HypoINS: Somatostatin + GH infusion.
sociation between plasma insulin concentrations and IGFBP-1 was examined in healthy nonobese, LB-Ob, and UB-Ob women in order to evaluate a spectrum of insulin resistance. Our data indicated a nonlinear, inverse relationship between postabsorptive plasma insulin concentrations and IGFBP-1. An apparent maximal suppression of hepatic IGFBP-1 production was obtained with systemic insulin concentrations of approximately 70-90 pmol/L, in agreement with findings from our earlier study on meal-related changes in plasma IGFBP1 and insulin (8). In the present study, 50% of LB-Ob and the majority of UB-Ob women had fasting insulin concentrations above 70 pmol/L and markedly reduced IGFBP-1 levels. Raising insulin concentrations above fasting levels (Le. insulin infusion) had little or no further suppressive effect on already low plasma IGFBP-1 levels in obese subjects but had pronounced effects on nonobese subjects, decreasing plasma IGFBP-1 60% in 2 h. These observations in normal lean subjects are similar to those reported previously showing a disappearance tliz for plasma IGFBP-1 of 60-120 min during meal-related increases in insulin and with euglycemic hyperinsulinemic clamp (8, 10, 28, 29). Our findings support the proposal
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
INSULIN
REGULATION
that insulin is the major regulator of hepatic IGFBP-1 production (8-13), and strongly suggest that the insulin resistance of obesity, best characterized with respect to insulin’s regulation of fuel (glucose, fatty acid, and amino acid) metabolism (24, 30), does not extend to its effects on IGFBP-1 production. Since IGFBP-1 is synthesized primarily by the liver (4-6,12,13), IGFBP-1 response to insulin may provide a marker of hepatic sensitivity to this hormone in ho. This sensitivity would be comparable to that of insulin regulation of adipose tissue lipolysis (31) and is 4-5 times more sensitive than insulin suppression of hepatic glucose production (32). In agreement with this notion, Brismar et al. (33) recently reported a highly significant inverse correlation between plasma IGFBP-1 and insulin secretion in healthy subjects. Thus, the low fasting IGFBP-1 levels in obesity are appropriate for the relatively high ambient insulin concentrations, as has been noted in other subject groups with insulin resistance (25, 28, 34). However, insulin response patterns were not fully investigated in these previous studies. Furthermore, our hypoinsulinemic infusion studies suggested differences between lean and obese populations in hepatic IGFBP-1 responsiveness to reduced insulin concentrations. Nonobese subjects had a brisk response to insulin withdrawal with increases in IGFBP-1 concentrations of 65 pg/L in 120 min. This agrees with our previous studies demonstrating a rate of increase in plasma IGFBP-1 concentrations of 0.5 pg/Lmin in normal subjects 3-4 h postprandially (8) and under controlled hypoinsulinemic conditions (Lee, P.D.K., et al., submitted manuscript). In comparison, IGFBP-1 concentrations increased approximately 25 pg/ L in 120 min in obese individuals following complete insulin withdrawal; the range in these subjects was 5-45 pg/L. This apparent blunting of IGFBP-1 in response to reduced insulin concentrations in obesity cannot be explained solely by differences in body mass between lean and obese women. Although the volume of distribution for IGFBP-1 is unknown, this protein is present in plasma and lymph (35), and therefore can be assumed to distribute with extracellular water. The 33 + 3% increase in extracellular water in obese individuals from these studies (data not shown) cannot totally account for the 3- to 4-fold difference in plasma IGFBP-1 response. Possible explanations include impaired hepatic IGFBP1 gene expression and/or protein synthesis induced by the chronic hyperinsulinemia or a delayed rise in IGFBP1 due to elevated portal insulin concentrations (8). In addition, a differential effect of insulin on IGFBP-1 clearance in obese subjects cannot be excluded from these data. We also found that GH levels were reduced in obese women while “GH-dependent” IGF-I levels were not
OF IGFBP-1
IN OBESITY
1359
significantly different between groups. The discordant relationship between GH and IGF-I in obesity may be explained by insulin stimulation of IGF-I production with resultant feedback suppression of GH (36). A decrease in circulating IGFBP-1 could promote available IGF-I for direct inhibition of somatotroph GH secretion. Free IGF peptides have been shown to inhibit GH expression and release in uiuo and in vitro (37, 38). Thus, chronic suppression of IGFBP-1 could contribute to “good growth without GH” in obesity (36). Although the role of insulin-mediated suppression of IGFBP-1 is uncertain, IGFBP-1 inhibits IGF-stimulated cell metabolism in vitro (19-23), and a similar action in uiuo has been proposed. One to 10% of plasma IGF-I circulates as free bioactive peptide (39,40). Thus, during normal daily fluctuations in nutritional status, insulin regulation of hepatic IGFBP-1 production could coordinate insulin and IGF-I action at the cellular level. Longterm overnutrition and hyperinsulinemia of obesity may alter this regulated growth response by persistent marked suppression of hepatic IGFBP-1 production. Given the normal and relatively stable IGF-I levels, it is interesting to speculate that chronically depressed levels of IGFBP1 accentuate the growth promoting effects of IGF-I as well as its feedback regulation of GH secretion. Additional studies will be needed to clarify the contribution of reduced plasma IGFBP-1 to the metabolic and mitogenie abberations associated with obesity. Acknowledgments The authors gratefully acknowledge the staff of the Mayo Clinic Clinical Research Center for excellent technical support, and Dr. Peter O’Brien of the Department of Biostatistics for his assistance with the statistical analyses.
References 1. Froesch ER, Schmid C, Schwander J, Zapf JA. Actions of insulinlike growth factors. Annu Rev Physiol. 1985;47:443-67. 2. Baxter RC, Martin JL. Binding proteins for the insulin-like growth factors: structure, regulation and function. Prog Growth Factor Res. 1989;1:49-68. 3. Drop SLS, Hintz RL. Introduction. On the nomenclature of the IGF binding proteins. In: Drop SLS, Hintz RL (eds) Insulin-Like Growth Factor Binding Proteins. Amsterdam: Excerpta Medica; 1989:v-vii. 4. Lee YL, Hintz RL, James PM, Lee PDK, Shively JE, Powell DR. Insulin-like growth factor (IGF) binding protein complementary deoxvribonucleic acid from human HEP Gz henatoma cells: uredieted protein sequence suggests an IGF binding domain different from those of the IGF-I and IGF-II receptors. Mol Endocrinol. 1988,2:404-11. 5. Brinkman A, Groffen C, Kortleve DJ, Geurta van Kessel A, Drop SLS. Isolation and characterization of a cDNA encoding the low molecular weight insulin-like growth factor binding protein (IBP1). EMBO J. 196&7:2417-23. 6. Julkunen M, Koistinen R, Aalto-Setala K, Seppala M, Janne OA, Kontula K. Primary structure of human insulin-like growth factor binding protein/placental protein 12 and tissue-specific expression of its mRNA. FEBS Lett. 196&236:295-302.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
1360
CONOVER
7. Suikkari AM, Jalkanen J, Koistinen R, et al. Human granulosa cells synthesize low molecular weight insulin-like growth factorbinding protein. Endocrinology. 1989;124:1088-90. 8. Conover CA, Butler PC, Wang M, Rizza RA, Lee PDK. Lack of growth hormone effect on insulin-associated suppression of insulin-like growth factor binding protein 1 in humans. Diabetes. 1990;39:1251-6. 9. Suikkari AM. Koivisto VA. Rutanen EM. Yki-Jarvinen H. Karonen SL, Se&ala M. Insulin regulates the serum levels ‘of low molecular weight insulin-like growth factor binding proteins. J Clin Endocrinol Metab. 1988:66:266-72. 10. Brismar K, Gutniak M, Povoa G, Werner S, Hall K. Insulin regulates the 35 kDa binding protein in patients with diabetes mellitus. J Endocrinol Invest. 1988:11:599-602. 11. Holly JMP, Biddlecombe RA, Dun&r DB, et al. Circadian variation of GH-independent IGF-binding protein in diabetes mellitus and ita relationshin to insulin. A new role for insulin? Clin Endocrinol (Oxf). 1988;29’667-75. 12. Conover CA, Lee PDK. Insulin regulation of IGF binding protein production in cultured HepGn cells. J Clin Endocrinol Metab. 1990;70:1062-7. 13. Lewitt MS, Baxter RC. Regulation of growth hormone-independent insulin-like growth factor-binding protein (BP-28) in cultured human fetal liver explants. J Clin Endocrinol Metab. 1989;69:24652. 14. /Cotterill AM, Cowell CT, Baxter RC, McNeil D, Silinik M. Regulation of the growth hormone-independent growth factor binding protein in children. J Clin Endocrinol Metab. 1988;67:882-7. 15. Busby WH, Snyder DK, Clemmons DR. Radioimmunoassay of a 26,OOOdalton plasma insulin-like growth factor binding protein: control by nutritional factors. J Clin Endocrinol Metab. 1988,67:1225-30. 16. Lee PDK, Hintz RL, Sperry JB, Baxter RC, Powell DR. Serum insulin-like growth factor binding proteins in growth-retarded children with chronic renal failure. Pediatr Res. i989;36:308-15. 17. Baxter RC. Cowell CT. Diurnal rhvthm of arowth hormone independent binding protein for insulin-like growth factors in human plasma. J Clin Endocrinol Metab. 1987;65:432-40. 18. Hintz RL, Liu F, Rosenfeld RG, Kemp SF. Plasma somatomedinbinding proteins in hypopituitarism: changes during growth hormone therapy. J Clin Endocrinol Metab. 1981;53:100-4. 19. Rutanen EM, Pekonen F, Makinen T. Soluble 34K binding protein inhibits the binding of insulin-like growth factor I to its cell receptors in human secretory phase endometrium: Evidence for autocrine/paracrine regulation of growth factor action. J Clin Endocrinol Metab. 1988;66:173-80. 20. Ritvos 0, Ranta T, Jalkanen J, et al. Insulin-like growth factor (IGF) binding protein from human decidua inhibits the binding and biological action of IGF-I in cultured choriocarcinoma cells. Endocrinology. 1988;122:2150-7. 21. Conover CA, Ronk M, Lombana F, Powell DR. Structural and biological characterization of bovine insulin-like growth factor binding protein-3. Endocrinology. 1990;127:2795-803. 22. Burch WM, Correa J, Shively JE, Powell DR. The 25-kilodalton insulin-like growth factor (IGF)-binding protein inhibits both basal and IGF-I-mediated growth of chick embryo cartilage in vitro. J Clin Endocrinol Metab. 1990;70:173-80. 23. Liu L, Brinkman A, Blat C, Hare1 L. IGFBP-1, an insulin-like growth factor binding protein, is a cell growth inhibitor. Biochem Biophys R.es Commun. 1991;174:673-9.
ET AL.
JCE&M.1992 Vol74’No6
24. Jensen MD. Regulation of forearm lipolysis in different types of obesity: In viva evidence for adipocyte heterogeneity. J Clin Invest. 1991;87:187-93. 25. Weaver JU, Holly JMP, Kopelman PG, et al. Decreased sex hormone binding globulin (SHBG) and insulin-like growth factor binding protein (IGFBP-1) in extreme obesity. Clin Endocrinol. 1990;33:415-22. 26. Jensen MD, Braun JS, Vetter RJ, Marsh HM. Measurement of body potassium with a whole-body counter: Relationship between lean body mass and resting energy expenditure. Mayo Clin Proc. 1988,63:864-8. 27. Daughaday WH, Mariz IK, Blethen SL. Inhibition of access of bound somatomedin to membrane receptor and immunobinding sites: A comparison of radioreceptor and radioimmunoassay of somatomedin in native and acid-ethanol-extracted serum. J Clin Endocrinol Metab. 1980;51:781-8. 28. Suikkari AM, Koivisto VA, Koistinen R, Seppala M, Yki-Jarvinen H. Dose-response characteristics for sunnression of low molecular weight plasma insulin-like growth factor-binding protein by insulin. J Clin Endocrinol Metab. 1989:68:135-40. 29. Yeoh SI, Baxter RC. Metabolic regulation of the growth hormone independent insulin-like growth factor binding protein in human plasma. Acta Endocrinol. 1988:119:465-73. 30. Groop LC, Saloranta C, Shank M, Bonadonna RC, Ferrannini E, DeFronzo RA. The role of free fatty acid metabolism in the pathogenesis of insulin resistance in obesity and noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab. 1991;72:96107. 31. Jensen MD, Caruso M, Heiling V, Miles JM. Insulin regulation of lipolysis in nondiabetic and IDDM subjects. Diabetes. 1989;38:1595-1601. 32. Rizza RA, Mandarin0 W, Gerich JE. Dose-response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol. 1981;240:E630-9. 33. Brismar K, Grill V, Efendic S, Hall K. The insulin-lie growth factor binding protein-l in low and high insulin responders before and during dexamethasone treatment. Metabolism. 1991;40:72832. 34. Conway_ GS. Jacobs HS. Hollv JMP. Wass JAH. Effects of luteinizing hormone, insulin, ins&n-like growth factor-I and insulinlike growth factor small binding protein 1 in the polycystic ovary syndrome. Clin Endocrinol. 1990:33:593-603. 35. Binoux M, Hossenlopp P. Insulin-like growth factor (IGF) and IGF-binding proteins: comparison of human serum and lymph. J Clin Endocrinol Metab. 1988;67:509-14. 36. Phillips LS, Harp JB, Goldstein S, Klein J, Pao C-I. Regulation and action of insulin-like growth factors at the cellular level. Proc Nutr Sot. 1990,49:451-8. 37. Snencer GSG. Macdonald AA. Buttle HL. Moore LG.? Carlvle --.-- SS. --~ Intracerebral ‘administration of insulin-like growth factor I decreases circulating growth hormone levels in the fetal pig. Acta Endocrinol. 1991;124:563-8. 38. Yamasaki H, Prager D, Gebremedhin S, Moise L, Melmed S. Binding and action of insulin-like growth factor I in pituitary tumor cells. Endocrinologv. 1991:128:857-62. 39. Guler HP, Zapf J, Froe&h ER.. Short-term metabolic effects of recombinant human insulin-like growth factor I in healthy adults. N End J Med. 1987:317:137-40. 40. Mukku VR. A 96-well microtiter plate assay for free IGF-I in ulasma usine immobilized IGFBP-3. Presented at The Endocrine Society, Waihington DC, June 19-22,199l (Abstract).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 05:58 For personal use only. No other uses without permission. . All rights reserved.
