| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Center for Eating Disorders and Department of Endocrinology (R.K.S., D.G., K.B., K.H.), Odense University Hospital, DK-5000 Odense, Denmark; and Medical Research Laboratories (J.-W.C., A.F., J.F.), Clinical Institute, Aarhus University Hospital, DK-8000 Aarhus C, Denmark
Address all correspondence and requests for reprints to: René Klinkby Støving, Department of Endocrinology M, Odense University Hospital, DK-5000 Odense, Denmark. E-mail: rene.k.stoving{at}dadlnet.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to examine the relationship between IGF-I bioactivity and IGF-binding capacity as expressed as formation of the binary complex of IGF-binding protein-1 (IGFBP-1) and IGF-I in patients with AN at different stages and with different subtypes of the disease.
Design: This was a longitudinal study.
Setting: The study took place at a clinical research center at a university hospital.
Study Participants: We studied a total of 45 women with AN and 24 age-comparable healthy controls.
Main Outcome Measures: IGF-I bioactivity was determined using an IGF-I receptor activation assay, and IGF-I/IGFBP-1 complex formation was determined by an assay that allows direct determination of the binary complex.
Results: IGF-I bioactivity was significantly decreased in serum from patients with AN. We found significant correlations between total, ultrafiltered free, and bioactive IGF-I. Despite increased IGFBP-1 concentrations, levels of IGF-I/IGFBP-1 binary complex were not significantly increased in AN. Oral contraceptives were associated with increased levels of IGF-I, IGFBP-1, and binary complex formation. Ghrelin levels were only significantly raised in those patients who had lost more than 5% of the body weight during the last 4 wk, whereas ghrelin levels in weight-stable as well as in weight-gaining patients did not significantly differ from the controls.
Conclusions: Total IGF-I level is a suitable marker of IGF-I bioactivity in emaciated patients with AN irrespective of the clinical subtype and acute nutritional state.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
IGF-I is a hormone with widespread metabolic and mitogenic actions. Regulation of IGF-I activity appears crucial in AN in adaptation to chronic starvation as well as in regenerative processes during nutritional restoration. In the circulation, approximately 99% of IGF-I is bound to specific, high-affinity IGF-binding proteins (IGFBPs). IGFBP-1 differs from the other IGFBPs in its rapid regulation by metabolic status (4). In the circulation, most IGF-I is bound in a ternary complex with IGFBP-3 and the acid-labile subunit (ALS). We previously found that the ALS level was decreased in AN (5). Dissociation from ALS is required for IGF-I to reach target tissues, whereas the binary complexes leave the circulation rapidly (6).
Numerous studies in AN have measured total, extractable IGF-I (1), which ignores the modifying effect of the IGFBPs at the level of the IGF-I receptor (IGF-IR) (7), and, therefore, data on IGF-I bioactivity are lacking. Total IGF-I levels have been shown to be positively related to body mass index (BMI) (8). However, total IGF-I is less sensitive to short-term nutritional changes than free IGF-I (4), and, therefore, we speculated that free and/or bioactive IGF-I are better indications of the actual nutritional condition. Also, total IGFBP-1 level has been studied in numerous studies, but it remains unknown whether the elevated serum levels of IGFBP-1 are accompanied by an increased binding of IGF-I and how this relates to bioactive IGF-I.
GH is an important regulator of IGF-I levels. Because of pulsatile secretion, single measurements of serum GH concentrations are inappropriate. We reported previously changed pulsatility and entropy of GH secretion in AN (9), primarily attributable to decreased hypothalamic somatostatinergic tone (10).
Ghrelin is a peptide with GH-releasing activity, and circulating ghrelin has been shown to correlate with total IGF-I levels in healthy subjects (11). Furthermore, ghrelin seems to play an essential role in regulating human eating behavior, and it has been suggested that ghrelin plays an important role in the pathophysiology of eating disorders, especially in binge-eating behavior (12, 13)
Insulin has been shown to stimulate hepatic IGF-I mRNA in vitro and to inhibit hepatic synthesis of IGFBP-1 at the transcriptional level (14). Low insulin levels are a well recognized feature in AN (1), but studies on insulin sensitivity determined by clamp techniques are conflicting (15). There is evidence that the adipocyte-derived hormone adiponectin mediates changes in insulin sensitivity and energy metabolism (16). In AN, serum levels of adiponectin have been found to be highly increased (17), normal (18), or low (19). Hypothetically, the discrepant findings may be attributable to the existence of a lower threshold limit for body fat in extreme emaciation to maintain hyperadiponectinemia. The potential role of adiponectin in adaptation to starvation and refeeding, however, has not been elucidated.
The purpose of this study was to examine the relationship between IGF-I bioactivity and IGF-binding capacity as expressed as formation of the binary complex of IGF-I and IGFBP-1 in serum from patients with AN at different stages and in different subtypes of the disease. IGF-I bioactivity was determined by a highly specific and sensitive IGF-IR kinase receptor activation assay (KIRA) based on cells transfected with the human IGF-IR gene (20). Furthermore, we aimed to investigate whether IGF-I bioactivity and IGF-I/IGFBP-1 complex formation was related to ghrelin, adiponectin, and insulin sensitivity estimated by the homeostatic model assessment (HOMA).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We studied a total of 45 women with AN: 20 with restricting subtype, 7 with binge-purging subtype, 8 AN patients taking oral contraception (OCP), and 10 weight-recovered AN patients, as well as 24 age-comparable healthy controls. The patients who used OCP (n = 8) consisted of seven with the restricting subtype and one with the binge-purging subtype. The AN diagnosis was based on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (2). Recovery was defined according to the Morgan-Russell criteria, which means that the patient had achieved at least 85% of the ideal body weight for age and height and obtained a regular menstruation during the last 6 months (21). None of the studied subjects suffered from medical diseases. The clinical and biochemical characteristics of the subjects are summarized in Table 1
. To study the significance of acute nutritional changes, the subjects were grouped according to the relative weight change during the last 4 wk (Table 2). Information about weight change during the last 4 wk were obtained from medical records. Furthermore, 15 patients (10 from the restricting group and five from the binge-purging group) were restudied 2–6 months after the first blood sample. The restudied patients had a minor weight gain, resulting in an increase in BMI from 14.8 ± 1.0 to 15.8 ± 1.2 kg/m2 (mean ± SD). All restudied women were outpatients and had gained less than 5% body weight during the last 4 wk. The controls did not use OCP or any other medications. They had regular menstrual cycles, and blood sampling took place in the follicular stage (d 2–8). All blood samples were obtained in the supine position between 0800 and 0930 h after an overnight fast. The study was approved by the local ethical committee, and all participants signed written informed consent.
|
|
All measurements were performed in duplicate within the same assay. With the exception of ultrafiltered free IGF-I and bioactive IGF-I, all intraassay and interassay coefficients of variation (CVs) were less than 5 and 13%, respectively. Serum-free IGF-I was determined using ultrafiltration by centrifugation as described previously (22). For free IGF-I the intraassay CVs, including ultrafiltration and immunoassay, averaged 19%. The detection limit of free IGF-I in the ultrafiltrates was 0.050 µg/liter. Serum IGF-I bioactivity was determined by an in-house KIRA based on human embryonic renal cells (EBNA 293 from Invitrogen, Karlsruhe, Germany) transfected with the human IGF-IR gene (20). In this assay, cultured cells are stimulated with either IGF-I standards or unknown serum samples. After stimulation of the cells, samples are aspirated and the cells were lysed. Thereafter, crude cell lysates were transferred to an assay that detects the concentration of phosphorylated (i.e. activated) IGF-IRs. This assay uses a monoclonal antibody against the extracellular IGF-IR for coating and an europium-labeled monoclonal anti-phosphotyrosine antibody (PY20) as tracer. The assay is sensitive (detection limit <0.08 µg/liter), specific (IGF-II cross-reactivity is 12%, and proinsulin, insulin, and insulin analogs have a cross-reactivity <1%), and precise (mean within and between assay CVs were <7 and 15%) (20). The serum level of bioactive IGF-I is most likely composed of the sum of two moieties: free, unbound IGF-I plus IGF-I being dissociated from the IGFBPs during incubation of serum with the IGF-IR-transfected cells. The latter, often referred to as "readily dissociable IGF-I," has been suggested to be biologically active (23), and, accordingly, the level of bioactive IGF-I measured by the KIRA has been higher than that of free IGF-I but lower than total IGF-I (20). Serum total IGF-I levels were determined after acid ethanol extraction using noncompetitive time-resolved monoclonal immunofluorometric assays (TR-IFMAs) as described previously (24). IGFBP-1 was determined by an in-house RIA with modifications as described previously (25). The RIA is based on a monoclonal IGFBP-1 antibody that recognizes all human phospho-forms of IGFBP-1 in serum (MAB 6303; Medix Biochemica, Kauniainen, Finland). The dimeric complex of IGF-I and IGFBP-1 was determined by a specific in-house TR-IFMA as described previously (26). In brief, the dimeric complex was captured by an IGFBP-1 antibody (MAB 6303) and detected by a europium-labeled monoclonal IGF-I antibody (Diagnostic Systems Laboratories, Webster, TX). The assay is highly specific for the dimeric complex of IGFBP-1 and IGF-I because the monoclonal IGFBP-1 antibody shows no cross-reactivity with the other IGFBPs (information supplied by the manufacturer; Medix Biochemica). This information is in accordance with our investigations that showed that neither IGFBP-2, IGFBP-3, nor IGFBP-4 cross-reacted. Furthermore, IGF-II did not cross-react, and, importantly, no signal was obtained unless both IGF-I and IGFBP-1 were present.
Ghrelin was determined by a validated in-house RIA as described previously (27) based on a polyclonal rabbit antibody, with acetylated recombinant human ghrelin serving as standard (both generously provided by Novo Nordisk A/S, Bagsværd, Denmark). Iodinated acetylated recombinant human ghrelin was obtained from Peninsula Laboratories (a division of Bachem, San Carlos, CA). The assay detects both acetylated and nonacetylated ghrelin, and it has been calibrated against the commercial ghrelin RIA from Phoenix Pharmaceutical (St. Joseph, MO). However, the RIA is unable to differentiate between acetylated and nonacetylated ghrelin, and therefore we consider the obtained levels as total ghrelin. Adiponectin was determined by a validated TR-IFMA as described previously (28). Serum estradiol was determined by RIA (Orion Diagnostica, Espoo, Finland), serum LH and FSH by TR-IFMAs (Delfia; Wallac Oy, Turku, Finland), insulin by a double-antibody RIA (Kabi Pharmacia Diagnostics, Uppsala, Sweden), and plasma glucose by the glucose dehydrogenase method (Merck, Darmstadt, Germany).
Dual-energy x-ray absorptiometry (DXA)
Total body fat, body fat percentage, and lean body mass were measured by DXA using a Hologic (Waltham, MA) QDR 2000 densitometer. All scans were performed in single-beam mode. The machine has been fully operable, cross-calibrated with other Danish DXA scanners, and serviced on a regular basis (at least once yearly), and quality control measurements have been performed daily with fully acceptable results. The software was updated by the manufacturer during the entire period.
Statistics
The statistical analyses were performed using SPSS version 13.0 (SPSS, Chicago, IL). The results are expressed as mean ± SDs, and comparisons between groups were performed using the Mann-Whitney test, with P < 0.05 indicating statistical significance. Bivariate correlations were estimated using Spearman’s coefficients (Table 3).
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
.
The mean age was not significantly different between patients and controls. Total, ultrafiltered free, and bioactive IGF-I serum levels in normal weight controls and in different subgroups of AN are illustrated in Fig. 1A. IGFBP-1, IGFBP-1 complexed IGF-I, and IGFBP-1 saturation in controls and in different subgroups of AN are illustrated in Fig. 1B. None of the anorectic groups differed significantly from the controls in ghrelin levels (Table 1). However, when stratified according to relative weight changes during the last 4 wk, the ghrelin levels in the acute weight-losing groups were significantly higher than that in controls (Table 2).
|
Refeeding study
During refeeding, BMI increased from 14.8 ± 1.7 to 15.8 ± 1.2 kg/m2 (mean ± SD) during 2–6 months in 15 patients, leading to significant increases in total, free, and bioactive IGF-I as illustrated in Fig. 2. IGFBP-1, IGF-I/IGFBP-1 binary complex, and ghrelin did not change significantly. HOMA increased significantly from 6.8 ± 5.4 to 10.3 ± 6.2, whereas adiponectin levels did not change significantly after the weight gain.
|
). Compared with the restricting and the binge-purging AN subgroups, the use of OCP was associated with significantly increased total, free, and bioactive IGF-I, IGFBP-1, and IGF-I/IGFBP-1 binary complex level, whereas the IGFBP-1 saturation was unchanged (Fig. 1, A and B). The OCP group did not differ from healthy regular menstruating women in total, free, or bioactive IGF-I (Fig. 1A). |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). These observations indicate that total IGF-I can be used as a suitable marker of the circulating IGF-I bioactivity in severely malnourished patients with AN at different stages and subtypes. We estimated the IGF-I bioactivity of serum by measurement of its ability to phosphorylate the IGF-I receptor in vitro, using cultured cells transfected with the human IGF-IR. We know from previous experiments that the cells respond not only to IGF-I but also to IGF-II, which has a cross-reactivity of 12%, when comparing pure IGF-peptide solutions (20). Thus, the observed receptor phosphorylation is likely to reflect the bioactivity of both peptides. However, activation of the IGF-II receptor does not lead to any biological signal, and therefore IGF-II is assumed to exert its biological effects by interaction with the IGF-IR. Because we measured the overall ability of serum to stimulate/activate the IGF-IR in vitro, our present data do not allow us to discriminate between the relative contribution of IGF-I and IGF-II. However, from a biological point of view, it is hardly important whether IGF-IR activation is caused mostly by IGF-I or IGF-II.
The in vivo bioactivity of IGF-I is usually estimated by measurement of the immunoreactivity of IGF-I in serum after extraction of the IGFBPs. Although this measurement has been shown to be valid in many cases, there are conditions in which the use of serum-free and/or bioactive IGF-I appears to more closely reflect the endogenous IGF-I bioactivity. This is true for chronic renal failure, obesity, and during fasting, in which in particular IGFBP-1 undergoes large changes (4). Based on these considerations, we compared the three different measurements of circulating IGF-I, namely total (extractable) IGF-I, (ultrafiltered) free IGF-I, and bioactive IGF-I in AN. Our data showed that, in AN patients, the three IGF-I measurements were tightly correlated, indicating that measurement of total (extractable) IGF-I may serve as a marker of the circulating IGF-I bioactivity. However, in controls, we failed to observe a close interrelationship between the three IGF-I measurements; in fact, only free and total IGF-I were significantly associated. We previously observed tight associations between free and bioactive IGF-I in other study cohorts (29) and speculate that the lack of relationship between the three IGF-I measurements is related to the homogeneity of the control group (i.e. a type 2 error).
Binge-eating and -purging behavior predict a poor outcome in AN (30). From clinical observations, we are convinced that binge-purging AN patients display more fluctuating weight than restricting AN patients, but this has received little attention in the literature. Compared with restricting AN patients, we found that the BMI comparable binge-purging subgroup showed significantly lower body fat percentage (Table 1), whereas the binge-purging subtype did not differ from the restricting subtype in IGF-I and IGFBP-1 pattern.
Previous studies have revealed that 40 h (31) as well as 72 h (26) fasting in normal-weight subjects increased the complex formation between IGF-I and IGFBP-1. As shown previously (7), IGFBP-1 levels were highly increased in all of the AN groups. Considering the hypoglycemic actions of IGF-I, the increase in IGFBP-1 during nutritional deprivation has been suggested as protection against hypoglycemia. Although a rare complication, severe hypoglycemia in AN certainly has been described in case reports (32).
Previous reports on IGF-I/IGFBP-1 binary complex formation in AN have been based on indirect correlation studies. In contrast, we used an assay that allows direct determination of the binary complex. Counterintuitively, no significant alterations were found in the levels of the binary complex in AN patients, irrespective of the subtype (Fig. 1B) or the acute nutritional state (Table 2). However, levels of the binary complex were increased in weight-recovered patients and in patients who used OCP (Fig. 1B). These results indicate that severely emaciated patients with chronic low IGF-I bioactivity differ from well-fed subjects in the way they adapt to short-term nutritional changes. Hypothetically, this could reflect a more inert or refractory adaptation in chronic emaciated subjects. This should be clarified by intervention studies, because it could have therapeutic implications for refeeding strategies.
Hypothalamic amenorrhea is included in the diagnostic criterion in AN (2). Eight patients used OCP. This group did not differ significantly from the other patients with regard to BMI or age. OCP was associated with increased total, free, and bioactive IGF-I levels (Fig. 1A), increased serum IGFBP-1, and binary complex levels but no significant rise in IGFBP-1 saturation (Fig. 1B). In well-fed aged subjects, oral hormone replacement markedly decreased total and free IGF-I levels, whereas IGFBP-1 and binary complex levels increased (33). In several studies in healthy women, OCP has been shown to decrease total IGF-I and increase IGFBP-1 levels, but, to our knowledge, the effect of OCP on IGF-I has neither been studied in AN nor in other types of hypothalamic amenorrhea or other catabolic conditions with GH resistance. The observed association does not allow any causal conclusion; however, it may indicate that the effects of estradiol on the IGF system are crucially modulated by the current nutritional status.
We found significantly increased adiponectin levels and increased insulin sensitivity as judged by HOMA in AN patients compared with control subjects (Table 1). This was observed in both subtypes of AN but not in weight-recovered patients or in OCP-using patients (Table 1). After a minor weight gain, HOMA increased significantly without significant changes in adiponectin levels. This refeeding response does not support the hypothesis that adiponectin is an essential regulator of insulin sensitivity during refeeding.
In controls, IGFBP-1 levels were significantly inversely correlated with insulin levels and HOMA. However, these correlations were disrupted in AN, suggesting that insulin is not a major regulator of IGFBP-1 synthesis in states of severe malnutrition. GH and cortisol levels are both highly increased in AN (9) and may interfere with the effect of IGF-I and IGFBP-1 on glucose metabolism. Furthermore, it should be stressed that, in AN patients, glucose-clamp studies have yielded conflicting results (15), and the HOMA has never been validated in AN.
There are conflicting reports on ghrelin levels in AN. In most studies, plasma levels were found to be increased (34), but, in some patients, ghrelin levels remained normal (35) and in other studies ghrelin levels normalized after a minor weight gain (13). Furthermore, in several studies, binge-eating behavior was found to be associated with increased (12) as well as decreased (36) ghrelin levels. The divergent findings may partly be explained by methodological dissimilarities, because the available assays detect different ghrelin fragments and isoforms. Using immunocomplex transfer-enzyme immunoassay, Hotta et al. (37) found that intact ghrelin levels were decreased in AN. Our data suggest that the acute nutritional condition may be another important reason for discrepant reports in the literature. In the majority of publications, it is not specified whether the anorectic subjects were in a weight-stable condition. Using an in-house RIA, we found that ghrelin levels were only significantly raised in those patients who had lost more than 5% of the body weight during the last 4 wk, whereas ghrelin levels in weight-stable as well as in weight-gaining patients did not differ significantly from the controls (Table 2).
A clear definition of full remission and recovery in AN is lacking (38). In accordance with others (39), we found that weight-recovered patients still displayed a markedly decreased body fat percentage compared with control subjects (Table 1). In the literature, a widely applied outcome criterion in AN is the Morgan-Russell Outcome Scale, but this does not take into account body composition and premorbid weight. In the present study, the recovered patients fulfilled the Morgan-Russell Outcome Scale for good outcome, and, aside from a low body fat percentage, they displayed significantly increased IGF-I bioactivity as well as an increased IGFBP-1 and binary complex levels (Fig. 1). Despite the low body fat percentage, the weight-recovered AN subjects did not differ from the controls in insulin sensitivity as assessed by the HOMA (Table 1). These observations indicate an abnormal metabolic state in weight-recovered AN patients. The rate of relapse in AN after refeeding is extremely high. Whether anthropometric parameters or alterations in the IGF-I system can predict relapses remain to be clarified.
In conclusion, total and ultrafiltered free IGF-I levels in serum from patients with AN mirrored in vitro IGF-I bioactivity, irrespective of the clinical subtype and the acute nutritional state.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Information: R.K.S., J.-W.C., D.G., and K.H. have nothing to declare. K.B. has received lecture fees and grant support from Merck Sharpe and Dohme since 2002. A.F. has received consulting fees from Hoffmann-La Roche and Taisho and lecture fees from GlaxoSmithKline and Novo Nordic. J.F. has received consulting fees from Hoffmann-La Roche.
First Published Online March 27, 2007
Abbreviations: ALS, Acid-labile subunit; AN, anorexia nervosa; BMI, body mass index; CV, coefficient of variation; DXA, dual-energy x-ray absorptiometry; HOMA, homeostatic model assessment; IGFBP, IGF-binding protein; IGF-IR, IGF-I receptor; KIRA, kinase receptor activation assay; OCP, oral contraceptive; TR-IFMA, time-resolved monoclonal immunofluorometric assay.
Received September 1, 2006.
Accepted March 21, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Rickenlund, M. Thoren, A. Nybacka, J. Frystyk, and A. L. Hirschberg Effects of oral contraceptives on diurnal profiles of insulin, insulin-like growth factor binding protein-1, growth hormone and cortisol in endurance athletes with menstrual disturbance Hum. Reprod., January 1, 2010; 25(1): 85 - 93. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Giustina, G. Mazziotti, and E. Canalis Growth Hormone, Insulin-Like Growth Factors, and the Skeleton Endocr. Rev., August 1, 2008; 29(5): 535 - 559. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Brugts, A. W. van den Beld, L. J. Hofland, K. van der Wansem, P. M. van Koetsveld, J. Frystyk, S. W. J. Lamberts, and J. A. M. J. L. Janssen Low Circulating Insulin-Like Growth Factor I Bioactivity in Elderly Men Is Associated with Increased Mortality J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2515 - 2522. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, P < 0.05 vs. restricting and binge-purging AN. Data are mean and SD.