| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Departments of Protein Engineering (Y.D., G.N., H.B.L.), Pharmacokinetics and Metabolism (D.L.M., P.J.F.), Bioanalytical Technology (A.I., M.D.S.), Endocrinology (D.A.H., E.F.), and Recovery Sciences (P.R., P.L.), Genentech, Inc., South San Francisco, California 94080
Address all correspondence and requests for reprints to: Dr. Henry B. Lowman, Department of Protein Engineering, 1 DNA Way, Genentech, Inc., South San Francisco, California 94080. E-mail: hbl{at}gene.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The IGFBP family includes six well characterized proteins (IGFBP-1 to -6) with molecular masses ranging from 23–31 kDa and a more distantly related group of proteins (7). The structure of the classical IGFBPs consists of two conserved domains joined by a variable linker region. They are often glycosylated and contain a characteristic signature of 16–20 cysteines. The variable linker region harbors cleavage sites for IGFBP-specific proteases, which include kallikreins, cathepsins, and matrix metalloproteinases (6, 8). Proteolytic cleavage generally inactivates the binding proteins by substantially reducing their affinity for IGF-I. This controlled IGFBP degradation adds an additional regulatory loop to the complex biological IGF-I system.
The biodistribution of IGF-I critically depends on 1) the formation of long-lived high molecular mass complexes, and 2) the absolute IGFBP concentrations. IGFBP-3 and -5 are apparently unique in their ability to form a ternary complex with an 85-kDa glycoprotein called acid-labile subunit (ALS). ALS association occurs only in the presence of IGF-I, and a basic motif in the carboxyl-terminal domains of IGFBP-3 and -5 seems to mediate this interaction (9, 10, 11). These approximately 150-kDa ternary complexes are unable to permeate the vasculature, and therefore act as a circulating reservoir of IGF-I. As a consequence, the serum half-life of IGF-I in ternary complexes is reported to be 12–15 h, as opposed to 30 min in binary complexes or 10 min in the free form (11). The second determinant of IGF-I biodistribution is the total concentration of binding proteins. IGFBP-3 is the most abundant binding protein, and IGFBP-2 is the second most abundant, followed by IGFBP-1; the serum concentrations of IGFBP-4, -5, and -6 are quite low (12). IGFBP-3, therefore, represents the main IGF-I carrier in the blood. In contrast, a substantial portion of IGFBP-1 and -2 in the blood is unoccupied. Hence, they appear to be the major modulators of free IGF-I levels (12).
The levels of IGFBP-1 are negatively regulated by insulin and can fluctuate more than 10-fold in response to changes in serum glucose concentrations (6). Experimental evidence showed that IGFBP-1 participates in glucose metabolism by causing insulin release (13). Furthermore, certain pathophysiological conditions, such as diabetes or renal failure, are accompanied by highly elevated IGFBP-1 levels (1, 6, 14). It is believed that under these conditions IGF receptor signaling is inhibited due to sequestration of free ligand by IGFBP-1. Lack of the IGF-I survival signal may result in a catabolic state of the cells, eventually leading to organ failure.
We now report the characterization of two IGF-I variants with greatly reduced IGFBP-1 affinity and with low nanomolar binding to IGFBP-3. Both variants display wild-type (wt) potency in cellular kinase receptor activation assays, stimulate proteoglycan synthesis in human articular cartilage, and are still capable of forming ternary complexes with IGFBP-3 and ALS. Hence, the half-lives of these variants are still predominantly determined by IGFBP-3, but their activities are no longer regulated by IGFBP-1. The ability of these molecules to activate the IGF receptor in the presence of elevated inhibitory IGFBP-1 levels is of potential therapeutic interest.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
IGFBPs. IGFBP-1 was expressed in mammalian Chinese hamster ovary cells and purified from conditioned medium as previously described (13). The gene encoding mature IGFBP-3 (17) was fused C-terminally to the gp67 signal sequence of the baculovirus transfer vector pAcGP67C (PharMingen, San Diego, CA), yielding plasmid pIGFBP3bac. Primary infection of SF9 insect cells using the BaculoGold system was performed according to the supplier’s manual (PharMingen), and viral stock amplification was also performed. For protein expression, typically 1 liter protein-free expression medium ESF 921 (Expression Systems LLC, Woodland, CA) was incubated with 2–5 x 105 Hi5 cells/ml in a 4-liter Spinner flask. Ten milliliters of viral stock solution were added, and cells were grown for 2.5 days at 27 C in the dark. Secreted protein was harvested by pelleting cells at 9000 rpm for 20 min at 4 C (centrifuge stopped without braking). The decanted supernatant was supplemented with one protease inhibitor mix tablet per liter (Complete, Roche Molecular Biochemicals, Mannheim, Germany) and stored at 4 C. An IGF-I affinity column was generated by coupling 180 mg IGF-I to 4 g controlled pore glass beads (18) (Sigma, St. Louis, MO). The volume of the column was approximately 15 ml. After equilibration at room temperature with PBS, the IGF-I column was loaded at 15 ml/min with 2 liters insect cell supernatant containing IGFBP-3. After loading, the column was washed with 5 vol buffer A (50 mM HEPES, pH 7.2). Contaminants and proteolyzed IGFBP-3 fragments were eluted with 100 ml 50% buffer B (2 M KSCN in buffer A). Intact IGFBP-3 was eluted with 100% buffer B in 10-ml fractions, monitoring the absorption at 280 nm. The protein- containing fractions were pooled, diluted 1:1 with 20 mM 2[N-morpholino]ethanesulfonic acid (MES) (pH 6.0; containing protease inhibitors), and concentrated to 1 ml in Centriplus 30/Centricon 30 microconcentrators (Amicon, Beverly, MA). Part (0.5 ml) of the concentrated protein mixture was loaded at 0.35 ml/min onto two tandem Superdex 75 gel filtration columns (Pharmacia Biotech, Uppsala, Sweden). The running buffer was 20 mM MES (pH 6.0), 150 ml NaCl, and protease inhibitors, and the fraction size was 1 ml. The fractions judged to be 95% or more pure by SDS-PAGE were pooled and stored in aliquots at -70 C. The identity of IGFBP-3 was verified by N-terminal sequencing and mass spectrometry. The yield of purified IGFBP-3 was approximately 1 mg/liter insect cell supernatant.
ALS purification. Recombinant human (rhu) ALS (19, 20) was expressed in Chinese hamster ovary cells and purified using anion exchange and hydrophobic interaction chromatography. The purified ALS was comparable to material purified using rhuIGF-I:rhuIGFBP-3 affinity chromatography methods similar to those previously described (19). Purified ALS was evaluated using quantitative amino acid analysis, SDS-PAGE, and an ALS ELISA (21). In addition, IGF-I:rhuIGFBP-3:ALS complex formation was observed by analytical size-exclusion chromatography. This purification yields an ALS preparation that is essentially free of IGFBP-3 and IGF-I (Lester, P., manuscript in preparation).
Surface plasmon resonance experiments
All experiments were performed on a BIAcore 2000 or 3000
instrument (Piscataway, NJ) at 25 C.
Competitive binding measurements. Approximately 200
RU (response units) of wt IGF-I were immobilized on the surface
of a B1 chip (BIAcore, Inc.) by conventional amine-coupling. IGFBP-1
samples at 5 nM were incubated with a 3-fold dilution
series of wt IGF-I, F49A, or E3A/F49A, as indicated in Fig. 1
. The incubation and running buffer was
PBS, 0.05% Tween-20, and 0.01% sodium azide. Protein mixtures were
injected at 20 µl/min for 5 min, with a dissociation time of 2 min.
The amount of free IGFBP-1 in solution was monitored by subtracting
nonspecific binding from the total relative response after 90 sec of
dissociation. Regeneration of the chip surface was achieved by a 60-sec
injection of 20 mM HCl, followed by a 3-min baseline
stabilization. The data were fit to a four-parameter curve to calculate
the IC50 of IGFBP-1 inhibition
(22).
|
Circular dichroism analysis
CD spectra were recorded at 25 C on an AVIV spectrometer (model
202, Lakewood, NJ) using a 1-cm path length cuvette. The protein
concentration was 1 µM in PBS. The spectra were corrected
for signal of the buffer alone.
Cell-based kinase receptor activation assay
IGF type I receptor kinase phosphorylation assays were carried
out as previously described (23, 24). Briefly, MCF-7 cells
that endogenously express IGF-I receptor, were cultured overnight in
96-well plates. The cells were then exposed to serial dilutions of wt
IGF-I, F49A, or E3A/F49A for 15 min. After ligand stimulation, cells
were lysed, IGF-I receptors were selectively immunoprecipitated, and
the extent of tyrosine phosphorylation was determined with an
antiphosphotyrosine antibody.
IGF-I activity on human articular cartilage explants
Human explants were obtained from a 79-yr-old female patient
undergoing joint replacement surgery. Articular cartilage was
dissected, minced, washed, and cultured in bulk for 24 h in a
humidified atmosphere of 95% air and 5% CO2 in
serum-free low glucose DMEM/F-12 medium with 0.1% BSA, 100 U/ml
penicillin/streptomycin (Life Technologies, Inc.,
Gaithersburg, MD), 2 mM L-glutamine, 0.1
mM MEM sodium pyruvate (Life Technologies, Inc.), 20 µg/ml gentamicin (Life Technologies, Inc.), and 1.25 mg/liter amphotericin B (Sigma).
Articular cartilage was aliquoted into Micronics (Costar, Inc.,
Cambridge, MA) tubes (
40–45 mg/tube). Twenty-four hours later,
explants were treated with 40 ng/ml IGF-I, F49A, or E3A/F49A for 5
days. Fresh medium was added on days 0, 1, 2, and 3.
To measure matrix (proteoglycan) synthesis, explants were cultured with
[35S]sulfate (10 µCi/ml) for the last 17
h of treatment. Explants were then washed twice with explant medium and
digested overnight at 50 C in a 900-µl reaction volume of 10
mM EDTA, 0.1 M sodium phosphate, and 1 mg/ml
proteinase K (Life Technologies, Inc.). Six hundred
microliters of the digestion reaction were mixed (2:1) with 10%
(wt/vol) cetylpyridinium chloride (Sigma) and centrifuged
at 1000 x g for 15 min. The supernatant was removed,
and formic acid (500 µl; Sigma) was added to dissolve
the pellets. The samples were then transferred to scintillation vials
containing 10 ml scintillation fluid (ICN Biomedicals, Inc., Costa Mesa, CA) and read in a scintillation counter. Shown
in Fig. 5 is the average value and SE of five
samples per treatment condition.
|
Iodination of IGF-I-variants, trichloroacetic acid (TCA) precipitation, and NaI administration. rhIGF-I, E3A/F49A, and F49A were iodinated by the indirect iodination method (25). The specific activity of the radiolabeled material ranged from approximately 93.1–156 µCi/µg protein. Each rat received approximately 15 µCi or 0.103–0.135 µg protein/0.2 ml dose, equivalent to 0.4–0.5 µg/kg.
NaI administration. To inhibit any nonspecific uptake of the 125I-radiolabeled material by the thyroid gland, all animals were injected with 10% NaI (0.5 ml), ip, 24 and 1 h before administration of test material.
Pharmacokinetics study. Three rats per group received an iv bolus of 15 µCi (0.2 ml) [125I]IGF-I, [125I]F49A, or [125I]E3A/F49A via a femoral vein catheter. Whole blood (0.3 ml) was collected via a jugular vein catheter before treatment and then at 1, 5, 15, and 30 min and at 1, 2, 4, 8, and 24 h after treatment. Blood volume was replaced with 0.3 ml saline after each sample collection, and catheters were locked with 0.1 ml heparin (10 U/ml) between time points. After collection, whole blood was placed into anticoagulant tubes containing EDTA and then aliquoted (0.1 ml) and placed on ice until analysis.
Tissue distribution study. Three rats per group per time point received an iv bolus of [125I]IGF-I, [125I]F49A, or [125I]E3A/F49A via a jugular vein catheter. Five, 15, and 30 min later, animals were anesthetized with phenobarbital (25–40 mg/kg) iv. Terminal cardiac blood samples as well as spleen, heart, liver, kidney, retroperitoneal fat, and pancreas were collected. Whole blood was processed as previously described. Tissue was sectioned into 1- to 2-g pieces, rinsed in saline, blotted dry, weighed, and then placed on dry ice for measurement of 125I counts per minute.
TCA precipitations. TCA (Sigma) precipitations were performed on plasma samples used for pharmacokinetic analysis. Plasma (0.05 ml) was added to 20% TCA (0.5 ml), incubated on ice, and centrifuged. Samples were then analyzed for 125I counts per min, the supernatant was removed so that only TCA precipitate remained, and the pellet was analyzed again. Pharmacokinetic calculations were derived from TCA-precipitable counts per min.
Sample analysis. Whole blood aliquots, TCA-precipitated
plasma, and preweighed tissue samples were counted using a
Wallac, Inc., 1470 Wizard automatic 
-counter (Turku,
Finland) for 1 min. Pharmacokinetic analyses (Table 2) were based upon
a three-compartment model that best described the steep distribution
phases resulting from protein binding and tissue distribution observed
in this experiment. The three-compartment model assumes three
conceptual volumes of distribution, such as a central compartment, a
peripheral (tissue) compartment, and a deep tissue compartment.
However, these do not necessarily correspond to anatomical regions. The
goodness of fit criteria included visual inspection of the observed
data and model-predicted values, examination of residuals, and the
criteria values generated by WinNonlin (Pharsight Corp.,
Mountain View, CA). Statistical comparisons were made by one-way ANOVA
followed by Fischer’s protected least significant difference test and
were considered significant where P < 0.05. Means and
SD are shown for Fig. 6, A and B.
|
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To measure the weak apparent affinities of these mutants for IGFBP-1, a
competition binding assay was performed using the BIAcore instrument.
IGFBP-1 at a constant concentration of 5 nM was incubated
with a dilution series of the respective IGF-I variants. These protein
mixtures were injected over a biosensor chip containing immobilized wt
IGF-I, thereby capturing unbound IGFBP-1 from the solution. In Fig. 1
the amount of bound IGFBP-1 (expressed as a percentage of the maximal
response) is plotted against the respective IGF-I variant
concentration. The half-maximal inhibitory concentration
(IC50) was 870 ± 90 nM for F49A
and was estimated to be more than 100 µM for E3A/F49A.
Self-association of E3A/F49A with the immobilized IGF-I on the chip was
observed at concentrations above 100 µM. The
IC50 for wt IGF-I in this type of assay was
1.2 ± 0.1 nM. These apparent affinities correspond to
700-fold and more than 80,000-fold decreases in wt affinity,
respectively.
The corresponding affinities for IGFBP-3 were determined in kinetic
BIAcore experiments using on- and off-rates to calculate dissociation
constants (Kd). The IGF-I variants were
immobilized on the biosensor chip, and serial dilutions of IGFBP-3 were
injected. As previously determined, F49A bound to IGFBP-3 with a
Kd of 6.3 ± 1.7 nM
(16). The Kd of E3A/F49A was
22.2 ± 10.3 nM (Table 1). The selectivity factors, defined as
the apparent affinity (IC50) for IGFBP-1 divided
by the apparent affinity (Kd) for IGFBP-3, were 1
for wt IGF-I, 140 for F49A, and more than 4500 for E3A/F49A. This means
that F49A had an approximately 140-fold higher binding selectivity for
IGFBP-3 over IGFBP-1 compared with wt IGF-I. The combined mutations in
E3A/F49A increased the wt selectivity factor more than 4000-fold. This
variant still bound IGFBP-3 with a Kd of 22
nM, corresponding only to a 15-fold reduction in wt
affinity for IGFBP-3.
|
, both F49A and E3A/F49A have virtually
identical spectra to that of wt IGF-I. The spectra reveal elements of
-helix and random coil, as expected from nuclear magnectic resonance
spectroscopy of IGF-I (27). The thermal stability of
neither wt nor mutant IGF-I could be determined accurately by circular
dichroism (data not shown), presumably due to the relatively high
random coil content already present at room temperature
(28) (30%). The fact that the circular dichroism
spectra of both IGF-I variants showed no significant deviation from
that of the wild type suggests that the introduced mutations do not
globally alter the structure of IGF-I.
|
), F49A (Fig. 3B), or E3A/F49A (Fig. 3C) in the running buffer at a concentration of
1 µM. As shown in Fig. 3, the ALS binding curves were
essentially identical for all three IGF-I variants. The dissociation
rates were 6.3 ± 1.7 x
10-4
sec-1 for wt IGF-I,
4.2 ± 0.9 x
10-4
sec-1 for F49A, and
4.0 ± 2.2 x
10-4
sec-1 for E3A/F49A. In the
absence of any IGF-I in the running buffer, ALS did not associate with
IGFBP-3 on the chip (data not shown). This experiment suggests that the
novel IGF-I variants retain the ability to form ternary complexes, and
that in vivo their terminal serum half-lives therefore
should not be drastically shorter than that of wt IGF-I.
|
) and E3A/F49A (Fig. 4B) were
indistinguishable from the wt IGF-I standard curves. The half-maximal
effective concentration (EC50) was 2.6 ±
0.2 nM for F49A and 2.6 ± 0.1 nM for
E3A/F49A. The wt IGF-I EC50 values were 2.5
± 0.1 nM (Fig. 4A) and 2.7 ± 0.1 nM
(Fig. 4B).
|
, both F49A and E3A/F49A stimulated matrix synthesis as effectively as wt
IGF-I in human articular cartilage. These results strongly suggest that
both IGF-I mutants can activate the IGF type I receptor and thus retain
full biological potency and activity.
Clearance and distribution of F49A and E3A/F49A in rats
Human wt IGF-I, F49A, and E3A/F49A were radiolabeled and
administered iv to rats to assess their pharmacokinetic properties.
Figure 6a shows plasma concentration
vs. time profiles for these proteins. Both IGF-I variants
cleared significantly faster compared with wt human IGF-I (Table 2). E3A/F49A cleared at the fastest rate
(1107 ± 45 ml/h·kg), followed by F49A (427 ± 125
ml/h·kg) and wt IGF-I (151 ± 24.7 ml/h·kg). A similar trend
was observed for the intermediate half-life,
t1/2ß; E3A/F49A had the
shortest, F49A an intermediate, and wt IGF-I the longest
t1/2ß (Table 2; see also
Discussion). These findings correlate well with the
corresponding in vitro affinities for the major binding
protein in the serum, IGFBP-3 (Table 1). Interestingly, the steady
state distribution volumes were much greater for the IGF-I variants,
indicating that they are less confined to the plasma compartment than
is wt IGF-I (Table 2). Figure 6B shows the tissue to blood ratio for
the IGF-I variants in different organs. Tissue to blood ratios greater
than 1 generally indicate accumulation within that tissue. The greatest
tissue to blood ratios for the IGF-I molecules were detected in the
kidney, whereas ratios in the liver, spleen, heart, and pancreas were
much lower. It is evident from this experiment that F49A and E3A/F49A
accumulate at statistically significant higher ratios in the kidney
compared with wt IGF-I.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Another interesting structural aspect is the flexibility of IGF-I. The molecule seems to be held together primarily by three disulfide bonds that are crucial for activity (34). In solution, the entire C domain was found to be disordered by nuclear magnetic resonance spectroscopy (27). Circular dichroism analysis of IGF-I showed a high content of random coil at 25 C (28), and thermal unfolding measured by the same method revealed a complete lack of cooperativity during the transition phase (data not shown). Similar observations were made by Miller et al. (34) in denaturation studies with guanidine hydrochloride. It can be speculated that this structural flexibility is an important feature allowing for IGFBP binding, as mature insulin, an IGF-I homolog that lacks the C peptide region and appears to be more structured by CD analysis (35), has no apparent affinity for the IGFBPs (36). However, we found no evidence for direct interactions between the C region and IGFBP-1 or IGFBP-3 (16). Alternatively, insulin might be locked in an unfavorable conformation for binding to the IGFBPs.
The ability of F49A and E3A/F49A IGF-I to form ternary complexes with the acid-labile subunit and IGFBP-3 has important consequences for the serum half-lives of these variants. Details of how ALS interacts with the IGF-I/IGFBP-3 complex are unclear due to a lack of structural information on IGFBP-3 and ALS. A recent model of ALS postulates a donut-shaped structure whose internal cavity is lined with negative charges; these regions could potentially interact with positive charges on IGFBP-3 and IGFBP-5 (37). The affinity of ALS for a cross-linked IGF-I/IGFBP-3 complex is on the order of 0.1–0.3 nM depending on the carbohydrate content of ALS (37). In our biosensor measurements, which analyzed ALS binding to a noncovalent IGF-I/IGFBP-3 complex (by including saturating amounts of IGF-I in the running buffer), the dissociation rates were consistently similar, between 4.0 x 10-4 to 6.3 x 10-4 sec-1. Interestingly, the apparent association rates generally decreased as a function of increasing analyte (ALS) concentrations, as illustrated here, for example, for wt IGF-I (1.83 x 105 M-1sec-1 (98 nM ALS), 1.32 x 105 M-1sec-1 (148 nM ALS), and 0.66 x 105 M-1sec-1 (333 nM ALS). The association constants for F49A and E3A/F49A followed a similar trend. Such an analyte-dependent phenomenon could be explained by preassociation of ALS oligomers. We are pursuing other methods to address the appropriate binding model for interpreting these data. Nevertheless, the experiment confirms that all three IGF-I variants tested form ternary complexes with ALS.
The pharmacokinetic characteristics of the IGF-I variants in rats
differed from those of wt IGF-I. Upon closer inspection of Fig. 6A, it
appears that the majority of differences are observed in the first
60–100 min of the experiment. In our three-compartment pharmacokinetic
model, this phase is characterized by the initial and intermediate
half-lives, t1/2 and
t1/2ß, which probably describe
the distribution of free and total radiolabeled molecules to
extravascular tissues and their clearance. After this initial phase,
all three IGF-I molecules are cleared at comparable rates, as suggested
by their similar terminal half-life values,
t1/2, which probably reflect
the elimination of 125I from the rat. The trends
in t1/2 and
t1/2ß observed across the
three molecules are generally consistent with a more rapid rate of
distribution and clearance for the IGF-I variants that are less bound
to IGFBPs. Furthermore, the calculated steady state distribution
volumes and clearance rates also support the interpretation that
IGFBP association plays a major role in the distribution and
clearance of the IGF-I variants. We therefore suggest that the initial
differences in the pharmacokinetic profile reflect the IGF-I/IGFBP-3
binding equilibrium, as they correlate well with the corresponding
affinities of the IGF-I variants for IGFBP-3 in vitro.
Because IGFBP-3 in the blood seems to be saturated under normal
conditions (1), it is reasonable to assume that the
injected IGF-I variants have to compete with endogenous IGF-I for
IGFBP-3 binding. The IGF-I molecules that are unable to associate with
IGFBP-3 could diffuse out of the vasculature and either associate with
other binding proteins or be cleared at a rapid rate
(t1/2 for free IGF-I, 10–15 min)
(11). Those molecules that achieve ternary complex
formation with IGFBP-3 and ALS, however, experience a more prolonged
serum half-life and become trapped in the vasculature. Ternary complex
formation of the IGF-I variants occurs with the same efficiency as wt
IGF-I in vitro, and this property appears to be reflected in
the later phase (>100 min) of the pharmacokinetic profile. The fact
that all three human IGF-I variants display terminal half-lives on the
order of 6–9 h strongly suggests that complex formation with rat
IGFBP-3 and ALS occurs in vivo. This is important, because
all in vitro experiments were performed exclusively with
human IGFBP-3 and ALS, whereas the in vivo experiments
rely on preservation of the engineered specificities for the homologous
rat binding proteins. An example where such specificity was not
preserved across species is peptide p1–01, which tightly bound human
IGFBP-1, but showed no measurable affinity for the rat homolog
(33).
The pharmacokinetic data reported here appear to be consistent with the range of values that has been observed in the literature for total IGF-I, free IGF-I, and variants with lower affinity for IGFBPs (38, 39, 40). It has been observed, for example, that naturally occurring IGF-I variants that exhibit a reduced affinity for IGFBPs have altered pharmacokinetic profiles similar to those observed in our study. Des(1, 2, 3)IGF-I, a truncated form of wt IGF-I with a reduced affinity for IGFBP-1 and IGFBP-3, had a nearly 4-fold increase in clearance and a 2-fold greater volume of distribution than IGF-I (39). Similarly, Bastian et al. (40) observed a faster clearance of radiolabeled LR3-IGF-I, an analog with reduced binding to IGFBPs, in rats compared with IGF-I. These observations are consistent with our findings comparing wt IGF-I to F49A, a molecule with reduced affinity for both IGFBP-1 and IGFBP-3. Consistent with the observation that increased plasma clearance correlates with reduced affinity for IGFBPs, E3A/F49A, whose affinity for both IGFBP-1 and IGFBP-3 is less than that of E3A, cleared much faster and had a greater volume of distribution. These findings provide further evidence for the importance of IGFBPs in regulating the plasma levels and bioavailability of IGF-I.
Specific IGFBPs are found to be up-regulated under certain pathophysiological conditions. For example, normal levels of IGFBP-1 to -4 are found in the synovial fluid of healthy subjects (29), whereas highly elevated amounts of IGFBP-3 and IGFBP-4 were detected in that of arthritic patients (41, 42). Similarly, in serum of chronic renal failure patients the levels of IGFBP-1 and IGFBP-2 are up-regulated in relation to the degree of renal dysfunction, thereby directly inhibiting IGF-I activity (14). A potential treatment strategy to restore IGF-I activity tries to release endogenous hormone from the IGFBPs by means of a displacer molecule (32, 33). Administration of a bioactive IGF-I variant with selectively reduced affinity for the particular inhibitory IGFBP species represents an alternative strategy. In our experiments, a statistically significant increase in kidney accumulation was observed for F49A and especially E3A/F49A compared with wt IGF-I. Based on this apparent accumulation in the kidney and the fact that they do not bind IGFBP-1, we propose that these variants may have unique properties in models of renal failure (14). However, more detailed experiments are needed to characterize the activity of these molecules in the kidney cells. Histological analysis should clarify whether the IGF-I variants occupy cell surface receptors or are rapidly cleared via the kidneys. Furthermore, the ability of F49A and E3A/F49A to stimulate cartilage matrix synthesis combined with the findings that IGFBPs may regulate the biological availability of IGF-I in articular cartilage (43) suggest that specifically engineered IGF-I variants may also be useful for the treatment of cartilage disorders such as arthritis.
In contrast to the reported inhibitory mechanism, IGFBP-1 has also been shown to potentiate IGF-I activity in cell culture under certain conditions (1). Presumably, IGFBP-1 was concentrated on the cell surface through interaction with a specific binding protein receptor. This mechanism would indirectly recruit IGF-I in close proximity to the IGF receptors, resulting in increased activity. No suitable reagents have been available to discriminate clearly between these opposing modes of actions. F49A and E3A/F49A appear to be useful molecules to address the question of inhibitory vs. potentiating effects of IGFBP-1 and to better elucidate its general physiological role. In summary, the binding protein-3-selective IGF-I variants described here represent a novel tool for the elucidation of the complicated IGFBP biology and may have a potential therapeutic benefit in conditions such as renal failure, cartilage disorders, or diabetes.
|
Acknowledgments |
|---|
Received June 5, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. M. Firth, F. MCDougall, A. J. MCLachlan, and R. C. Baxter Impaired Blockade of Insulin-Like Growth Factor I (IGF-I)-Induced Hypoglycemia by IGF Binding Protein-3 Analog with Reduced Ternary Complex-Forming Ability Endocrinology, May 1, 2002; 143(5): 1669 - 1676. [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 |
