| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
The Medical Research Laboratories, Clinical Institute, and Medical Department M (Diabetes and Endocrinology), Aarhus University Hospital, DK-8000 Aarhus C, Denmark
Address all correspondence and requests for reprints to: Dr. Jan Frystyk, Medical Research Laboratories, Nørrebrogade 44, Aarhus University Hospital, DK-8000 Aarhus C, Denmark. E-mail: jan{at}frystyk.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Objective: Our objective was to develop an assay for free GH in serum.
Methods: We used ultrafiltration by centrifugation. Due to the small molecular difference between GH and GHBP, the size of GHBP and GHBP-GH complexes was increased by preincubation of serum with a monoclonal GHBP antibody (MAb 263).
Results: The ultrafiltration membrane almost completely retained all GHBP (>98.5%) and allowed free passage of unbound GH (>98.4%). Addition of increasing concentrations of GHBP reduced free GH dose dependently, and measured and calculated levels of free GH changed in parallel. During an insulin-tolerance test, free and total GH changed in parallel in all individuals (n = 11) and their peak values as well as area under the curve values were positively correlated (r = 0.89; P < 0.001 and r = 0.92; P < 0.001, respectively). Of note, the relative levels of free GH (calculated as the area under the curve of free to total GH) was inversely correlated with GHBP (r = –0.94; P < 0.001).
Conclusion: It is possible to measure free GH in human serum. Free GH correlated positively with total GH and inversely with GHBP. Measurement of free GH may be a helpful future tool in the management of GH disorders and in studies of GH-GHBP interrelationships.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
75%) and considered as the prototype of pituitary GH. After secretion, GH associates with two plasma proteins: a well-characterized high-affinity GH-binding protein (GHBP) and a lesser characterized low-affinity GHBP (2, 3). The high-affinity GHBP (for simplicity hereafter referred to as GHBP) is a heavily glycosylated protein composed of 246 amino acids and with a molecular mass of 60–65 kDa (4, 5, 6). GHBP originates from proteolytic cleavage of the extracellular domain of the GH receptor (GHR), and therefore, its plasma levels are believed to reflect the tissue density of the GHR. GHBP circulates in molar excess of GH and is far from saturated, whereas about 50% of GH is bound to GHBP during basal conditions (7, 8). GH normally interacts with GHBP in a 1:1 molar ratio, and only at supraphysiological GHBP concentrations has a 2:1 stoichiometry been observed (9). In contrast, the GHRs exist in preformed dimers and therefore normally interact with GH in a 2:1 molar ratio (10, 11).
Five to 20% of plasma GH is reported to complex with low-affinity proteins. Early studies identified one or more peptides of between 124 and 174 kDa (6, 12, 13). A later study demonstrated that GH specifically bound to 
2-macroglobulin, which therefore appears to represent a high-capacity, low-affinity GHBP (14).
The biological role of GHBP remains unclear, but it appears to influence GH bioactivity as well as kinetics. In cell culture systems, GHBP inhibits GH-stimulated actions (15, 16). In contrast, in experimental animals, administration of GHBP enhanced the bioactivity of GH in two of three studies (17, 18, 19) and prolonged the half-life of GH, most likely by confining GH to the vascular compartment (20, 21). In humans, on the other hand, an inverse association between GH half-life and GHBP has been observed (22). These findings are suggested to reflect opposing effects of GHR and GHBP on circulating GH. A high GHR density may increase the cellular uptake of GH and consequently its plasma clearance. Conversely, elevated GHBP levels may serve to decrease GH clearance (6). However, this notion needs to be confirmed, and one study using pegvisomant GHR blockage failed to confirm that the receptor participated in the clearance of GH (23).
To increase the knowledge on the relationship between circulating GH and GHBP, we developed a clinically applicable assay for determination of free, unbound GH. This paper details the assay procedure, its validation, and the relationship between free GH, total GH, and GHBP.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Free GH was isolated using ultrafiltration by centrifugation, which enables processing of undiluted serum at 37 C (24, 25). However, the relatively small difference in the molecular size of free GH and GHBP made it unlikely that an ultrafiltration membrane was able to isolate free from bound GH with a high recovery, and indeed this notion was confirmed by pilot studies using membranes with a 50-kDa cutoff (data not shown). On the other hand, ultrafiltration membranes with 100-kDa cutoff values were unable to fully retain GHBP (data not shown), and therefore it was decided to increase the size of GHBP and GH-GHBP complexes by incubating serum with a GHBP antibody before ultrafiltration. The commercially available GHBP monoclonal antibody MAb 263 (Agen Biomedical Ltd., Acacia Ridge, Queensland, Australia) was well suited for that purpose: first, binding of MAb 263 to GHBP does not occupy the GH-binding site (26); second, MAb 263 recognizes GHBP as well as GH-GHBP complexes (27); and third, concentrations up to 10 mg/liter (60 nM) showed no interference in the commercial GH immunoassay used in the present study [the GH Delfia from PerkinElmer LifeSciences, Turku, Finland, originally developed by Albertsson-Wikland et al. (28)] either in the absence or presence of GH (data not shown).
Ultrafiltration procedure
To retain all GHBPs above the membrane during ultrafiltration, MAb 263 must be added to serum in molar excess of GHBP. In our in-house GHBP assay, serum levels seldom exceed 5 nM (27, 29), and pilot studies showed that addition of MAb 263 in a final concentration of 2.5–5 mg/liter yielded comparable levels of free GH (data not shown). Hence, a final concentration of about 3 mg/liter (20 nM) was chosen. MAb 263 was dissolved in fetal calf serum (FCS), because this medium cross-reacts in neither our GH nor in our GHBP assay (27).
After addition of MAb 263, serum was incubated overnight at 5 C in sealed Eppendorf tubes. The next day, serum was transferred to Microcon YM-100 Centrifugal Filter devices (Millipore Corp., Bedford, MA) and incubated at 37 C for 1 h, before samples were centrifuged (1500 x g at 37 C). After centrifugation, the ultrafiltrates (25 µl) were immediately assayed in the GH Delfia, which according to the manufacturer has a sensitivity of less than 0.033 µg/liter and intra- and interassay coefficients of variation (CV) less than 5%.
Assay validation
To study the recovery of GH in ultrafiltrates, serial dilutions of 22-kDa recombinant human (rh)GH (0, 0.1, 1, 10, and 100 µg/liter; generously provided by Novo Nordisk A/S, Bagsværd, Denmark) dissolved in FCS were incubated overnight at 5 C, and then samples were equilibrated for 1 h at 37 C before they were centrifuged in triplicate for either 20 or 30 min. Ultrafiltrates were analyzed for free GH according to instructions by the manufacturer. Serum total GH was measured by the same assay but after an overnight incubation of the samples to minimize the effect of GHBP (30).
To study the effect of sample volume, different volumes (100, 200, 300, or 400 µl) of serial dilutions of GH (0, 0.1, 1, 10, and 100 µg/liter) dissolved in FCS were added to the ultrafiltration devices, incubated for 1 h at 37 C, and finally centrifuged for 30 min, also at 37 C. Afterward, all ultrafiltrates were assayed for GH.
Previous studies of serum free IGF-I have shown that it is possible to dilute serum in Krebs-Ringer buffer (KRB) up to 20-fold before ultrafiltration without any major changes in free IGF-I levels (24). Therefore, the effect of dilution of serum on free GH levels was investigated, and for that purpose, samples obtained during an insulin tolerance test (ITT) were ultrafiltered either without dilution or after a dilution 1:5 in KRB containing 5% human serum albumin.
The ability of the ultrafiltration membrane to retain GHBP was studied for nonglycosylated rhGHBP and glycosylated endogenous GHBP, which have similar binding affinities for GH (5). First, rhGHBP was dissolved in FCS containing 4.5 µg/liter GH and incubated at 5 C with or without MAb 263, after which samples were ultrafiltered. Second, to study endogenous GHBP, two serum samples were added a serial dilution of GH (0, 5, 10, 50, and 100 µg/liter) plus MAb 263 and processed as described.
Impact of exogenous GHBP on free GH in vitro
To study the effect of GHBP on levels of free GH, FCS spiked with about 16 µg/liter rhGH was incubated overnight at 5 C with a serial dilution from 0.08–10 nM of rhGHBP (generously provided by Novo Nordisk). The next day, samples were equilibrated and processed as described. The obtained concentrations of ultrafiltered free GH were compared with the estimated levels of free GH, calculated according to the formula by Barsano and Baumann (31) (Fig. 1
), using a GHBP binding affinity (Ka) of 0.9 x 109 liter/moles.
|
To study the effect of exogenous GH on levels of serum free GH, six normal sera were spiked with a serial dilution of GH up to 122 µg/liter dissolved in FCS and incubated overnight at 5 C. The next day, all samples were equilibrated for 1 h at 37 C and centrifuged. Measured and calculated levels of free GH were compared as described. Serum GHBP was measured with our in-house assay (27).
Sample stability
Repetitive freezing and thawing of serum samples were used to test the stability of free GH (and indirectly the GHBP-GH complex). Peak samples from six subjects undergoing an ITT were used. All samples were thawed in ice-cold water and left for 1 h at room temperature before they were refrozen at –20 C for at least 24 h. Free GH levels were measured after one, five, and 10 cycles.
Clinical validation of free GH
To validate the ultrafiltration assay for free GH in a clinical setting, we compared serum levels of free and total GH during an ITT. Eleven subjects (eight females and three males; age 23 ± 4 yr; range 13–60 yr) from the Outpatient Clinic of Medical Department M, Aarhus University Hospital, were studied. These subjects were all suspected of GH deficiency, and therefore they underwent a routine ITT. However, all subjects had a peak total GH above 3 µg/liter, and therefore the possibility of GH deficiency was waived. The ITT was performed according to standard guidelines. At time 0, an iv bolus injection of insulin (0.1 U/kg Actrapid from Novo Nordisk) was administered, allowing plasma glucose levels to decrease to less than 2.5 mM. Blood samples were drawn at –30, –15, and 0 min and then for every 10 min during the next 2 h. Serum free and total GH was determined in all samples as described and GHBP at time –30, 0, 30, 60, 90, and 120 min.
All serum samples used in the present study were obtained in accordance with the Helsinki declaration with amendments, after informed consent.
Statistical analyses
Data were compared by Student’s unpaired or paired t test as appropriate. During the ITT, the different time points were compared by one-way repeated-measures ANOVA, which if significant was followed by Bonferroni’s t test, which adjusts the P value for multiple comparisons. Alternatively, for nonparametric data, we used the Friedman repeated-measures ANOVA on ranks followed by Dunn’s test. Linear regression analyses were used to determine the predictors of free GH. All correlations passed the normality test and the constant variance test. Data are given as mean ± SEM. P values < 0.05 were considered statistically significant. All statistics were made by use of SigmaStat for Windows, version 3.5 (Systat Software Inc., San Jose, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After 30 min of centrifugation, the recovery of free GH in the ultrafiltrates corresponding to FCS containing to 0.1, 1, 10, or 100 µg/liter GH averaged 95.1 ± 0.8, 102.7 ± 4.3, 99.4 ± 4.4, and 97.1 ± 3.5%, respectively (grand mean = 98.6 ± 1.6%). This was similar to the recovery observed after 20 min (grand mean = 98.4 ± 2.5%). However, the CV were slightly lower after 30 min of centrifugation (5.4 ± 1.4 vs. 6.7 ± 1.6%). Therefore, for additional experiments, 30 min of centrifugation was chosen.
The recovery of ultrafiltered free GH was independent of sample volume, averaging 97.9 ± 5.5, 100.0 ± 3.7, 99.0 ± 7.1, and 97.0 ± 7.1% after addition of 100, 200, 300, or 400 µl serum. Hence, for additional experiments, for each ultrafiltration device, 150 µl serum was added.
Dilution of serum 1:5 in KRB reduced levels of free GH during the ITT. Thus, peak levels of free GH averaged 6.32 ± 0.62 µg/liter in undiluted serum but only 2.06 ± 0.25 µg/liter in serum diluted 1:5 before ultrafiltration. Therefore, it was decided to proceed with undiluted serum.
In practice, the following procedure was used. To Eppendorf tubes with 340 µl serum was added 10 µl MAb 263 (100 mg/liter) dissolved in FCS; this approach resulted in a very small dilution of serum (3%). After mixing, samples were incubated overnight at 5 C. The next day, each sample was divided into two aliquots of 150 µl, which were added to the ultrafiltration devices, incubated for 1 h at 37 C, and finally centrifuged at 37 C for 30 min. Ultrafiltrates were analyzed immediately for GH. Generally, the volume of the ultrafiltrates equaled about 40 µl, which sufficed for a single determination. Thus, the presented data on free GH represent the mean of either two or three ultrafiltrates.
The ability of the ultrafiltration membranes to retain nonglycosylated rhGHBP averaged 12.8 ± 4.2% in the absence of MAb 263 and 97.8 ± 0.4% in the presence of MAb 263. The retention of serum GHBP, which due to its glycosylation has a larger molecular mass than rhGHBP (5), was higher, averaging 32.3 ± 4.5% in the absence of MAb 263 and 98.5 ± 0.2% in the presence of MAb 263. In addition, the retention of serum GHBP was independent of the concentration of GH (data not shown).
Impact of exogenous GHBP on free GH in vitro
The relationship between GH and GHBP was studied in two ways. First, increasing concentrations of rhGHBP were added to FCS spiked with rhGH, after which the measured levels of free GH (mean CV < 2%) were compared with those obtained after calculation (Fig. 2). As can be seen, the curves for measured and calculated free GH were parallel.
|
Second, we added increasing concentrations of GH to six serum samples from healthy subjects and determined free GH. The level of GHBP in the six serum samples averaged 2.56 ± 0.40 nM (range, 1.16–3.73 nM), and the baseline free GH level was 0.16 ± 0.09 µg/liter. For comparison, the calculated levels of free GH in the presence of 1.0 and 4.0 nM GHBP are also depicted (Fig. 3). As can be seen, the vast majority of the measured free GH values were within the theoretically calculated range.
|
Freezing and thawing serum for one, five, and 10 times did not affect levels of free GH, which averaged 8.10 ± 4.02 vs. 7.48 ± 3.53 vs. 8.09 ± 4.16 µg/liter (one vs. five vs. 10 cycles; n = 6; P = 0.4), a finding that indicates that the GHBP-GH complex is stable.
Clinical validation of free GH
We used an ITT to compare the relationship between free and total GH (Fig. 4). All samples from an individual were analyzed for free GH in the same run with an intraassay CV averaging less than 7%. The interassay CV was estimated by repetitive measurement of a control sample and averaged less than 11%.
|
The area under the curve (AUC) of free and total GH were positively correlated (Fig. 5A), as were peak levels (Fig. 5B). GHBP tended to correlate inversely with the AUC of free GH (r = –0.55; P = 0.08) as well as the peak value of free GH (r = –0.53; P = 0.09). To further study the impact of GHBP on free GH levels, we calculated the ratio between the AUC of free and total GH, whereby it was possible to study the impact of GHBP on free GH independently of the absolute levels of GH (which were 100% for all subjects), and in this setting, the fraction of free GH was highly inversely correlated with GHBP (Fig. 5C). The impact of GHBP on free GH is further illustrated in Fig. 6, which shows the 11 individual relationships between peak total GH, peak free GH, and GHBP. Two pairs of subjects are shown in bold lines to emphasize that even though they had similar levels of total GH peak values, their free GH peak values differed oppositely of their GHBP levels. Thus, for the same peak total GH value, the peak value of free GH depends on GHBP.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ultrafiltration by centrifugation is well recognized as a suitable method for isolation of unbound peptide hormones, because it allows undiluted serum to be processed at 37 C (24, 25). But the methodology requires a relatively large molecular difference between the investigated peptides to make an ideal separation. As it turned out, the weight difference between GH and GHBP was too small to allow a satisfactory separation, and therefore the monoclonal GHBP antibody MAb 263 was introduced. MAb 263 has previously been used to immunoprecipitate GHBP-complexed GH (32), and it neither occupies the GH binding domain of GHBP nor interferes in the GH Delfia (26, 27). Seeing that dilution affected the equilibrium between free and bound GH, it was decided to add a small volume of highly concentrated MAb 263 so that the antibody was present in molar excess of the endogenous GHBP and at the same time diluted the samples only marginally (i.e. 3%).
It is acknowledged that the ultrafiltration method does not allow us to provide solid evidence on the impact of the low-affinity, high-molecular mass GHBP on levels of free GH. However, in the absence of MAb 263, only about two thirds of the endogenous GHBP (60–65 kDa) were able to pass the ultrafiltration membranes. Thus, it seems unlikely that a substantial fraction of molecules with a reported size of 125–175 kDa should be able to pass the 100-kDa membranes. Moreover, the contribution of the low-affinity GHBPs on the overall circulating GH binding capacity is considered to be low in comparison with that exerted by the high-affinity GHBP (7).
The free GH assay was validated by in vitro and in vivo experiments. In vitro, the relationship between free GH and GHBP was tested by spiking serum with physiological concentrations of GH, GHBP, or both peptides. Because we have no previous data to compare with, we used calculated levels of free GH as reference. As seen (Fig. 2
), the two curves were parallel but not superimposable. This finding indicates that it was inaccuracies in the added concentrations of peptides and/or in the applied association constant rather than methodological errors that accounted for the observed difference. Similarly, addition of increasing concentrations of GH to normal sera with GHBP levels ranging from 1.16–3.73 nM resulted in levels of free GH that for the vast majority of observations were within the calculated range of free GH corresponding to a theoretical level of GHBP of 1.0 and 4.0 nM, respectively (Fig. 3). Only at high GH levels did the ultrafiltration method apparently overestimate serum free GH. However, this observation most likely reflects that the GH-binding capacity of the serum samples had been exceeded.
During the ITT, levels of free and total GH changed in parallel. The time course of total GH after insulin-induced hypoglycemia was in accordance with previous observations (33), whereas there are no previous clinical data on free GH to compare with, but computer simulations data are available (7). After simulation of a GH burst reaching its half-maximal amplitude within 30 min, the time to peak was about 3 min longer for total than free GH (7). This finding is not contradictory with the present data, which were based on samples collected with 10-min intervals. However, whereas the computer simulation predicted that the ratio of free to bound GH rises after a GH burst, we were unable to detect any increase in this fraction during the ITT. There might be at least two possible explanations for this discrepancy. First, for low (i.e. baseline) levels of GH, the estimate of the ratio is less accurate, because both signals are likely to vary more for low than high readings. Second, the average levels of total GH remained less than 10 µg/liter during the ITT, and according to the formula by Barsano and Baumann (31), this results in only minor changes in the fraction of free GH; for instance, when total GH increases from 1 to 10 µg/liter, the fraction of free GH rises from 53 to 58% when GHBP equals 1 nM, and from 22 to 23% when GHBP equals 4 nM. Still, our data confirm that GHBP has a damping effect on the secretory oscillations of GH (7).
Serum levels of free and total GH were positively associated. However, for the same total GH, levels of free GH were strongly inversely correlated to GHBP. This supports the notion that GHBP controls in vivo levels of free GH. Of notice, the saturation of GHBP differed significantly between individuals, ranging from 40–80%. As recently reviewed, the production of GHBP is influenced by numerous factors, and accordingly, levels may differ up to 10-fold between various patient groups (6). As shown in the present study, this appears to have important implications for levels of free GH, which do not simply mirror total GH.
In conclusion, we have described the development and validation of a novel, ultrafiltration assay for free GH, which allows determination of free GH in undiluted serum at 37 C. The obtained data showed that free GH correlated positively with total GH and inversely with GHBP. It is our belief that the assay will enable further insights into the interrelationship between GH and GHBP. Furthermore, we speculate that free GH measurements may serve as a diagnostic tool in the management of GH disorders, particularly in those subjects with a borderline level of total GH, but this remains to be investigated.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure statement: The authors have nothing to disclose.
First Published Online May 20, 2008
Abbreviations: AUC, Area under the curve; CV, coefficients of variation; FCS, fetal calf serum; GHBP, GH-binding protein; GHR, GH receptor; ITT, insulin tolerance test; KRB, Krebs-Ringer buffer; rh, recombinant human.
Received February 15, 2008.
Accepted May 9, 2008.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-macroglobulin as a growth hormone-binding protein in human blood. J Clin Endocrinol Metab 80:585–590[Abstract]This article has been cited by other articles:
 |
|
 |
|
  N. Moller and J. O. L. Jorgensen Effects of Growth Hormone on Glucose, Lipid, and Protein Metabolism in Human Subjects Endocr. Rev., April 1, 2009; 30(2): 152 - 177. [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 |
). A serial dilution (range, 0.08–10 nM) of nonglycosylated rhGHBP was added to FCS spiked with rhGH (
), free GH (
), and GHBP (
) during an ITT in 11 subjects suspected for GH deficiency (mean ± SEM). All subjects had a peak total GH value of more than 3 µg/liter. Significant changes from baseline are indicated (
). GHBP levels at –30 min differed significantly from those at 0 and 30 min (P < 0.001). However, for clarity, these changes have not been indicated.
and 
). For both pairs, the subject with the highest level of free GH had the lowest level of GHBP. Thus, for the same peak value in total GH value, the peak value of free GH depended on GHBP. Notice the log scale of the y-axis.