0021-972x196/$03.00/0
Journal
of Clinical
Endocrinology
and Metabolism
Copyright
0 1996 by The Endocrine
Society
Vol. 81, No. 4
Printed
in U.S.A.
Recombinant
Synthesis
of Insulin-Like
Growth
FactorBinding
Protein-4
(IGFBP-4):
Development,
Validation,
and Application
of a Radioimmunoassay
for IGFBP-4 in
Human
Serum and Other Biological
Fluids*
YOKO HONDA,
EDWIN
C. LANDALE,
AND SUBBURAMAN
MOHAN
Departments
of Medicine,
Biochemistry,
Pettis Veterans Administration
Medical
DONNA
D. STRONG,
Microbiology
and Physiology, Loma Linda
Center, Loma Linda, California
92357
ABSTRACT
I
-.
I
NSULIN-LIKE growth factor I (IGF-I) and IGF-II, previously known assomatomedins, are structurally related to
insulin and are the two most abundant polypeptide growth
factors that circulate in human plasma. In plasma and other
biological fluids, they are bound to a family of structurally
related proteins, the IGF-binding proteins (IGFBPs). This
family of at least six IGFBPs, distinct from the IGF receptors,
modulates the effects of the IGFs in different tissues,including bone (l-7). IGFBP-4 was initially purified from human
bone cell-conditioned medium and adult rat serum as 25and 32- to 36-kDa forms, respectively (8, 9). Subsequently,
IGFBP-4 was also purified from conditioned medium collected from a variety of cell types, and Northern analysis of
cultured cells has also shown that IGFBP-4 is expressed in
various cell types (2,6). The primary structure of the IGFBP-4
Received
August
9, 1995. Revision
received
October
27, 1995. Accepted November
6, 1995.
Address all correspondence
and requests for reprints to: Subburaman
Mohan,
Ph.D., Research Service, Jerry L. Pettis Veterans Administration
Medical
Center, 11201 Benton Street, Loma Linda, California
92357.
* This work was supported
by funds from the NIH (AR-31062
and
AR-07543),
the V.A., the Departments
of Medicine
and Pediatrics,
Loma
Linda University
and the Japanese Foundation
for Aging and Health.
J. BAYLINK,
University
and
reacted with the IGFBP-4
antiserum,
and that recovery
of IGFBP-4
from serum samples
exceeded
90% when exogenous
IGFBP-4
was
added and was unaffected
by the addition
of IGFs or by repeated
freezing
and thawing
of the sample. We employed
this IGFBP-4
RIA
to demonstrate
an increase in IGFBP-4
in TE85 human osteosarcoma
cell-conditioned
medium
after treatment
with dibutyryl
CAMP, PTH,
and 1,25dihydroxyvitamin
Da, agents known
to increase
the IGFBP-4 messenger
ribonucleic
acid level. Application
of this RIA to the
measurement
of IGFBP-4
in human
serum revealed
that the circulating
level of IGFBP-4
in 41 individuals
in the 61-87
yr age group
(546 ? 135 pg/L) was 35% higher
than that in 24 individuals
in the
23-40 yr age group (404 t 156 pg/L). The mean circulating
level of
PTH was also 20% higher in the 61-87
yr group compared
to that in
the 23-40
yr group (P < 0.01). In addition,
serum IGFBP-4
amounts
showed a significant
positive correlation
with age (r = 0.54; P < 0.001)
and serum PTH (r = 0.26; P < 0.01). These datavalidate
this IGFBP-4
RIA and illustrate
its utility
in illuminating
the physiological
mechanisms that regulate
IGFBP-4
in uiuo and influence
its effects on the
IGFs in both normal
and abnormal
pathology
and in aging. (J Clin
Endocrinol
Metub
81: 1389-1396,
1996)
Insulin-like
growth
factor-binding
protein-4
(IGFBP-41,
like the
five other IGFBPs
present
in human
serum, acts as a transport
protein for insulin-like
growth
factor I (IGF-I)
and IGF-II
and modulates
their biological
effects. To investigate
the role of IGFBP-4
in the
physiology
of the IGF system,
we developed
a sensitive
RIA for
IGFBP-4
employing,
as antigen,
tracer,
and standard,
recombinant
human
IGFBP-4
(rhIGFBP-4)
expressed
in Escherichia
coli as a fusion protein with glutathione
S-transferase
and affinity
purified
with
dutathione-derivatized
resin. Antibodv
against
the rhIGFBP-4
fusion protein
was raised in guinea
pigs; tracer
and standard
were
provided
by the rhIGFBP-4
moiety
that had been cleaved
from the
rhIGFBP-4
fusion protein
and repurified
by reverse
phase high pressure liquid chromatography.
We report
that both IGFBP-4
purified
from PC3 human
prostate
cell-conditioned
medium
and rhIGFBP-4
bound IGF and migrated
in electrophoresis
gels in an identical
manner; that in gel permeation
chromatography,
rhIGFBP-4
coeluted
with the IGFBP-4
present in human
serum; and that both are equally
immunoreactive
with
the IGFBP-4
antiserum.
Employing
this
IGFBP-4
RIA, we determined
that no IGFBP
other than IGFBP-4
I
DAVID
protein has been deduced from clones isolated from human
placenta and osteosarcoma complementary DNA (cDNA)
libraries (10, 11). The cDNA for human IGFBP-4 encodes a
258-residue protein that is processed,by removal of the signal sequence,to a mature protein of 237 residues (25.6 kDa)
with a single asparagine-linked glycosylation site (10). Although various cell types when in culture secrete both glycosylated (28-29 kDa) and nonglycosylated (24-25 kDa)
forms of IGFBP-4, the latter is typically the most abundant
(2, 6).
IGFBP-4 inhibits IGF actions under a variety of experimental conditions. Mohan et al. (8, 12) demonstrated that
IGFBP-4 inhibited both IGF-I- and IGF-II-induced cell proliferation in embryonic chick calvaria cells and MC3T3-El
mouse osteoblasts; IGFBP-4 inhibited IGF-I- and IGF-IIstimulated DNA synthesis in a variety of cell types (6,7). In
addition to having significant biological effects on cell proliferation, in vitro studies using Western ligand blot analyses
have shown that the level of IGFBP-4 protein in the conditioned medium of a variety of cell types is affected by a
number of physiological regulators (3-6). IGFBP-4 synthesis
may be regulated by systemic hormones and local growth
factors at the transcriptional or posttranscriptional level (13).
1389
HONDA
1390
The 16- to l&kDa form of IGFBP-4 is attributed to the latter
process (14-16). Studies in vitro revealed that PTH, 1,25dihydroxyvitamin
D,, IGF-I, IGF-II, transforming growth
factor-p, and osteogenicprotein-l /bone morphogenetic protein-7 are major regulators of IGFBP-4 production in human
bone cells (5,10,17-19), but little has been learned about the
serum regulation of this binding protein becausemeasuring
physiological concentrations of IGFBP-4 was problematic before the development, reported herein, of an IGFBP-4 RIA.
The sensitivity provided by this assay is essentialfor studying IGFBP-4 regulation both in vitro and in vivo and for
evaluating the contribution of IGFBP-4 to the IGF binding
capacity in serum and other biological fluids. As we were not
successful in our efforts to produce high titer antibodies
specific to IGFBP-4 using either synthetic peptides or limited
amounts of purified native IGFBP-4, we undertook studies to
express large amounts of IGFBP-4 by recombinant DNA
technology to develop and validate a specific IGFBP-4 RIA.
We also demonstrate the utility of this assay for in vitro and
in vivo studies.
Materials
and
Methods
Recombinant
human
(rh) IGFBP-1
and rhIGFBP-2
were purchased
from Upstate Biotechnologv
(Lake Placid, NY) and Austral
Biologicals
(San Ramon, CA), respectyvely.
Nonglycosylated
rhIGFBP-3
was> gift
from Dr. A. Sommer
(Santa Clara, CA). Human
IGFBP-4
and IGFBP-6
were purified
from human prostate PC3 cell-conditioned
medium using
established
procedures
(8). rhIGFBP-5
was a gift from Drs. Carola Dony
and Kurt Lang (Boehringer
Mannheim,
Mannheim,
Germany).
rhIGF-I
was a gift from Ciba-Geigy
(Basel, Switzerland);
rhIGF-II
was purchased
from Bachem Chemicals
(Torrance,
CA). Rabbit polyclonal
antiserum
against IGF-I (kindly
provided
by L. E. Underwood
and J. J. Van Wyk)
was obtained
from the National
Hormone
and Pituitary
Program
(Baltimore,
MD). Monoclonal
antibody
against rat IGF-II was purchased
from Amano
International
(Troy, VA).
Subjects
Serum samples were obtained from normal healthy men and women,
aged 23-87 yr. None of these subjects was receiving
medical treatment,
suffering
from malnutrition,
or showing
signs of acute or chronic disease. A single nonfasting
blood sample was drawn
for IGF-I, IGF-II,
IGFBP3,
and IGFBP-4
determinations.
Serum samples were collected
with patient
consent under
the ethical guidelines
established
by the
Initial Review
Board of Loma Linda University
Medical
Center.
of the IGFBP-4
expression
JCE
vector
IGFBP-4 was expressed
from a construct based on the cDNA derived
IGFBP-4 insert that had been cloned into the pTZlSRBstX1
(Invitrogen,
San Diego, CA) vector to generate clone TE89S3a14BP
(pTE89S3)
(10).
The Escherichia coli strain HBlOl was used throughout,
and recombinant
DNA and biochemical
manipulations
employed
standard
methods
(20,
21). The identity
of the DNA constructs
was confirmed
by restriction
analysis. The upstream
untranslated
region of IGFBP-4 was deleted from
plasmid
pTE89S3 by digestion
with EcoRI and ApaI. The gel-purified
fragment
was recircularized
by intramolecular
ligation
with an adapter
complimentary
to the resultant
sticky ends to bring the coding sequence
for IGFBP-4
into frame with the LacZa peptide.
To facilitate
IGFBP-4
protein
purification,
the cDNA-coding
region from glycine 17 to glutamic acid 258 was excised by digestion
with Srfr and XhoI, and the
gel-purified
fragment
was ligated
into the glutathione
S-transferase
fusion vector uGEX-5X-2
(Pharmacia
Fine Chemicals,
Uuusala. Sweden)
at SmnI and ‘XhoI. The resultant
construct,
pSB4 iFig.’ l), expressed
protein
after induction
with 1 mmol/L
isopropylthio-p-o-galactoside
& M . 1996
Vol81
l No 4
‘PstI,2677
1. Expression
plasmid
for rhIGFBP-4
fusion
protein
Restriction
enzyme
sites (in base pairs) and protein-encoding
including
that of the fusion protein
(IGFBP-41,
are shown.
FIG.
(IPTG)
(Difco
during
the last several hours of growth
in either
Laboratories,
Detroit,
MI) or 2xYT medium
(21).
Purification
Materials
Construction
ET AL.
Terrific
(pSB4).
regions,
Broth
of rhIGFBP-4
Protein
was obtained
from harvested
bacteria
that were lysed by
sonication
in phosphate-buffered
saline (PBS)-0.1
mmol/L
phenylmethylsulfonylfluoride
at 4 C. The sonicate was centrifuged,
and fusion
protein was obtained
from either the supernatant,
in which case the yield
was enhanced
by growing. the bacteria at 30 C rather than 37 C, or from
the pellet. The p&t
was solubilized
in 10 mmol/L
Tris-HCl
(pH 7.4) and
4 mol/L
euanidine-HCl
(10 mL/g)
and then dialvzed
at 4 C against PBS
to removi
the guanidine.
The fusion protein
was subsequenzy
affinity
purified
on glutathione-derivatized
Sepharose
4B resin (Pharmacia)
and
eluted with 5 mmol/L
glutathione
and 50 mmol/L
Tris-HCl
(pH 8.0)
following
the manufacturer’s
instructions.
After confirming
purity
by
SDS-PAGE,
the fusion protein was employed
for antibody
production.
The IGFBP-4
moiety was released after cleavage
of the 53-kDa
fusion
protein with 1 mg/mL
protease factor Xa (Boehringer
Mannheim)
in 0.1
mol/L
NaCl, 1 mmol/L
CaCl,, and 50 mmol/L
Tris-HCl
(pH 8.0). This
resultant
rhIGFBP-4
had an 11-amino
acid amino-terminal
extension
derived from uGEX-5X2,
GIPGIPGPSLG,
not oresent in native orocessed
IGFBP-4,
followed
by five amino acids derivld
from the IGFB’P-4 signal
sequence. The cleaved IGFBP-4
protein
was subsequently
purified
by
high pressure liquid chromatography
C, reverse phase chromatography
using a gradient
of acetonitrile,
as described
previously
(8, 12). The
fractions
containing
IGFBP-4
were dried by Speed-Vat
centrifugation
(Savant Instruments,
Farmingdale,
NY), reconstituted
in PBS, and used
as tracer and standard.
IGFBP-4
antisera
Antisera
were raised in guinea pigs by dissolving
400 Kg purified
rhIGFBP-4
fusion
protein
in 0.1 mL PBS and mixing
with 0.1 mL
Freund’s
complete
adjuvant
before injecting
10 animals SC with 200 pg
purified
rhIGFBP-4.
For subsequent
boosts, 200 pg rhIGFBP-4
fusion
protein in 0.1 mL were mixed with 0.1 mL alum and injected im every
4 weeks. After eight boosts, two animals (no. 11 and 12) developed
high
titer antibodies
and precipitated
30-40%
of [‘*sIJIGFBP-4
at a 1:2500
dilution.
Antiserum
from the former animal was used throughout
this
study.
IGFBP-4
tracer
One microgram
of rhIGFBP-4
was iodinated
using a modified
chloramine-T
method,
as described
previously
(22). [‘251JIGFBP-4
was separated from free rz51 using a Sip-Pak
C& cartridge
(22). The speciiic
activitv
of the radiolabeled
IGFBP-4
tracer varied
between
100-200
&i/&g
protein.
Aliquots
of radiolabeled
IGFBP-4
were stored at -70
C and used for up to 4 weeks for RIAs.
RIA FOR IGFBP-4
IGFBP-4
RIA
RIA assay buffer was comprised
of 0.05 mol/L
sodium phosphate
(pH 7.5), 0.25% BSA, and 0.02% sodium
azide. Standards
(0.1 mL of
0.24-31.2
pg/L solutions)
and unknowns
were added to 0.2 mL RIA
assav buffer and ureincubated
for 1 h at room temperature
with 0.1 mL
primary
antibod;
(1:2500 dilution)
before addingb.1
mL [‘25111GFER’-4
tracer (40,000 cpm) and incubating
for 16 hat 4 C. The bound [125111GFBP4 was separated
from free tracer by adding 0.1 mL of a 1:20 dilution
of
normal
guinea pig serum, 0.4 mL of the appropriate
dilution
of goat
antiguinea
pig IgG, and 0.2 mL 8% polyethylene
glycol 8000. After
mixine.
the samoles were incubated
at room temperature
for 2 h. The
precipytates
were pelleted
by centrifugation
and-quantitated
using a
y-counter.
Serum samples
were diluted
1:20 with RIA buffer before
assay unless otherwise
specified.
IGF-I
and IGF-II
RIAs
IGF-I and IGF-II were measured
by RIAs after separation
of IGFBPs,
as previously
described
(22). As IGFBPs have been shown to produce
artifacts in IGF radioligand
assays, it is essential to completely
separate
the IGFBPs from the IGFs for the IGF determinations
to be valid. The
separation
of the IGFBPs from IGFs was achieved
by a recently
developed and validated
rapid acid gel filtration
protocol
(22).
Western
immunoblot
analysis
A 0.05-mL
sample was diluted with 0.05 mL nonreducing
SDS-dissociation
buffer 10.125 mol/L
Tris (pH 6.8),4%
SDS, and 20% glycerol],
then loaded onto a 1.5-mm discontinuous
SDS-polyacrylamide
gel and
electrophoresed
at 10 milliamperes
overnight
through
a 4% stacking gel
and lo-20%
gradient
separating
gel. After electrophoresis,
the proteins
were electroblotted
onto a 0.45-pm
BA-S nitrocellulose
membrane
(Schleicher
and Schuell, Keene, NH) at 50 volts for 3 h according
to the
method
of Towbin
et al. (23). The filters were blocked for 1 h with 5%
nonfat dry milk and incubated
for 1 h with IGFBP-4
antiserum
(1:2000
dilution).
The nitrocellulose
membrane
was subsequently
incubated
for
1 h with horseradish
peroxidase-conjugated
rabbit antiguinea
pig IgG
(1:lOOO dilution;
Zymed
Laboratories,
San Francisco,
CA). Antigenantibody
reactions
were visualized
using
ECL chemiluminescence
reagents
according
to the manufacturer’s
instructions
(Amersham
Life Sciences, Arlington
Heights,
IL).
High performance
and collection
of conditioned
28 kga-
medium
Normal
human bone cells derived
from rib, TE85, and U2
osteosarcoma
cells were cultured
as previously
described
(14).
lection of conditioned
medium,
cells were plated in 60-mm petri
in 2 mL DMEM
containing
2% calf serum and incubated
for 24
removal
of serum, fresh DMEM
containing
0.1% BSA (2 mL) was
After 24 h, fresh serum-free
medium
containing
the effecters was
Forty-eight
hours later, the medium
was collected,
centrifuged
move cellular debris, and stored at -70 C until assayed.
Statistical
rhIGFBP-4 fusion protein, which migrated as a 53-kDa protein in SDS-polyacrylamide gel under reducing conditions
(data not shown). Of the four bacterial strains used, HBlOl
provided the highest rhIGFBP-4 yield (10% of the extractable
protein) and was subsequently employed throughout the
remainder of the study. Western ligand blot analysis employing [‘251]IGF-II tracer indicated that the 53 kDa rhIGFBI’-4 fusion protein bound IGF-II. A maximal increase in
IGF-II-binding activity was found between 4-6 h after induction with IPTG. IGF-II-binding activity in the HBlOl cell
extract in the absenceof IPTG was not detectable (data not
shown). HBlOl cells grown at 30 C, to enhance the production of soluble protein, yielded about 0.5-1.0 mg soluble
rhIGFBP-4 fusion protein/L bacterial culture, and affinity
purification of the cleaved rhIGFBP-4 yielded 100-200 Fg
homogeneous preparation from 1 L bacterial culture. Comparison of rhIGFBP-4 purified from E. coli HBlOl and
IGFBP-4 purified from PC3 cell-conditioned medium revealed that the two IGFBP-4 preparations comigrated in SDSPAGE and showed similar IGF-binding activity upon Western ligand blot analysis (Fig. 2). The purified rhIGFBP-4
preparation was shown to be homogeneous by SDS-PAGE,
where it comigrated with the PC3 cell-derived 24- to 25-kDa
IGFBP-4 protein. The sequencewas confirmed by N-terminal
amino acid sequenceanalysis, and the concentration of rhIGFBI’-4 in the purified fraction was determined by microamino acid analysis (data not shown). The comparative
IGF-binding activities of rhIGFBP-4 and PC3 cell derived
IGFBP-4, determined by a polyethylene glycol precipitation
assay, were similar (data not shown).
The specific IGFBP-4 tracer binding to antiserum under the
assay conditions described in this study was 14.4 + 2.2% of
the total radioactivity (n = 15) and varied from 11.5-l&0%
gel chromatography
A 1 X 30-cm Superdex
G-75 column
(Pharmacia)
was equilibrated
with 50 mmol/L
sodium phosphate
and 0.1 mol/L
sodium chloride, pH
7.0 (elution buffer), at a flow rate of 0.5 mL/min.
Equal volumes
of fresh
normal
human
serum were mixed with elution
buffer
or 500 p.g/L
IGFBP-4, and 0.2 mL of each of these mixtures
was applied to the column
separately.
The proteins were eluted with the elution buffer at a flow rate
of 0.5 mL/min.
Two minute
fractions
were collected,
and O.l-mL
aliquots were assayed for IGFBP-4
immunoreactivity.
Cell culture
1391
human
For coldishes
h. After
added.
added.
to re-
analysis
19 kDa-
2:g
lB,ng
rhlGFBP-4
Assay results are presented
as the meant SD. Datawereanalyzed
using standard
statistical methods,
including
Student’s
t test and linear
regression
analysis (CSS, Stat Soft, Tulsa, OK).
Results
Induction with IPTG of various bacterial strains that had
been transformed with pSB4 resulted in the expression of
2ng
long
I
PC-3b-l
IGFBP-4
FIG. 2. Western
ligand blot analysis
of rhIGFBP-4
and IGFBP-4
purified from PC3 cells. Purification
procedures
for rhIGFBP-4
are described in Materials and Methods.
IGFBP-4
produced
by PC3 human
prostate
cells was purified
from serum-free
conditioned
medium
(PC3
CM) as previous&
described
(8). Two and 10 ng rhIGFBP-4
and
PC3-derived
IGFBP-4
were electrophoresed,
blotted,
and detected as
described
in Materials
and Methods.
The migration
positions
of size
markers
are indicated
at the left (in kilodaltons).
HONDA
1392
depending on the age of the tracer. The nonspecific binding
of IGFBP-4 tracer measured in the absenceof primary antibody was 1.9 ? 0.5% (n = 15). Addition of increasing concentrations of purified rhIGFBP-4 inhibited the binding of
[‘251]IGFBP-4 tracer to the antiserum in a dose-dependent
manner (Fig. 3). Half-maximal displacement (ED,,) occurred
at 0.36 2 0.05 rig/tube, and 20% (ED,,) displacement of tracer
binding to the antiserum occurred at 0.12 2 0.023 rig/tube
(n = 15). Inter- and intraassay variations were lessthan 8.1%
and 5%, respectively. IGFBP-4 purified from prostate PC3
cell-conditioned medium displaced the binding of tracer to
the antiserum in a manner identical to that of rhIGFBP-4 (Fig.
3). Normal human serum and conditioned medium collected
from TE85 human osteosarcoma cells, which produce
IGFBP-4, gave parallel dose-responsedisplacement curves
(Fig. 3). In contrast, conditioned medium collected from U2
human osteosarcomacells, which do not produce IGFBP-4,
as determined by the absence of the 2.2-kilobase IGFBP-4
messengerribonucleic acid (mRNA) transcript (lo), did not
displacethe binding of IGFBP-4 tracer to the antiserum (Fig. 3).
Competition experiments with sera from various species
revealed the capability of the human IGFBP-4 RIA to detect
IGFBP-4 from these species. Sera from sheep and mouse
inhibited the binding of [‘251]IGFBP-4 to anti-IGFBP-4 antiserum, but were much lesseffective than human serum (Fig.
4). Rabbit serum contained immunoreactive material in
amounts equal to that of human serum. Rat (Fig. 4) and chick
(data not shown) sera contained little if any material that
cross-reacted with the anti-IGFBP-4 antiserum (Fig. 4). RIA
assaysof the IGFBPs at concentrations up to 0.5 mg/L revealed that the anti-IGFBP-4 antiserum reacted only with
human IGFBP-4 and not with other human IGFBPs (Fig. 5).
The specificity of the IGFBP-4 antiserum was also examined by Western immunoblotting after electrophoresis under
nonreducing conditions of conditioned medium collected
from PC3 prostate cells, which do produce IGFBP-4, and U2
human osteosarcoma cells, which do not produce (10)
IGFBP-4 (Fig. 6A). A major 25-kDa band and two minor
bands at 28-29 and 40-45 kDa (seeDiscttssion)were detected
in the preparation of IGFBP-4 purified from PC3 prostate
1
0
*
,
0
0.8
5 0.6
rhIGFBP-4
PC3CMBP-4
TE85 CM (cmtrd)
TE35 CM @BcAMP)
m
0.4
loo
10’
102
lo3
,uL or cLg/L IGFBP4
FIG.
3. Displacement
of the IGFBP-4
tracer
from the IGFBP-4
antiserum.
Competition
between
the 11251]IGFBP-4
tracer and purified
rhIGFBP-4,
IGFBP-4
purified
from PC3 cell-conditioned
medium,
conditioned
medium
derived
from TE85 and U2 human osteosarcoma
cells, and normal
human
serum for binding
to the IGFBP-4
antiserum. All samples
were tested in duplicate.
ET AL.
JCE & M . 1996
Vol81.
No 4
1
0.8
5 0.6
!a
0.4
0.2
GFBP4
0
10-i
100
&
10’
102
Sera or ccg/L IGFBP-4
FIG. 4. IGFBP-4
RIA reactivity
with animal
sera.
[125111GFBP-4
tracer
from the IGFBP4
antiserum
standard
(IGFBP-4)
and by sera from the animals
The y-axis values (B/B0 ratio) are the counts per min
in the presence
of competitor
divided by the counts
determined
in the absence of competitor.
Displacement
of
by rhIGFBP-4
named in figure.
of tracer assayed
per min of tracer
1.2
1
0
*
W
A
0.8
5 0.6
F9
0.4
BP-1
BP-2
BP-3
BP-4
x BP-5
t BP-6
0.2
10-l
loo
10’
lo2
lo3
lo4
IGFBP (,&L)
5. IGFBP-4
RIA
reactivity
with
IGFBPs.
Displacement
of
[iz51]IGFBP-4
tracer from IGFBP-4
antiserum
by various
concentrations of purified
human
IGFBP-1,
-2, -3, -4, -5, and -6. The y-axis
values (B/B0 ratio) are the counts per min of tracer assayed
in the
presence
of competitor
divided
by the counts per min of tracer determined in the absence of competitor.
FIG.
cell-conditioned medium (Fig. 6A, lane 1); a major band at 25
kDa was detected in the purified rhIGFBP-4 preparation (Fig.
6A, lane 2), and an additional minor 16-kDa band was detected in conditioned medium collected from PC3 prostate
cells (Fig 6A, lane 3). Immunoreactive material was reduced
considerably in PC3 cell-conditioned medium after it was
subjected to IGF-II affinity chromatography (Fig. 6A, lane 4)
and was absent from conditioned medium collected from U2
human osteosarcoma cells (Fig. 6A, lane 5). After electrophoresis under reducing conditions, the immunoreactive
material in human serum (Fig. 6B, SERUM) was shown to
include a major band of equivalent electrophoretic mobility
to rhIGFBP-4 (Fig. 6B, BP-4) as well as minor 28- to 29- and
16-kDa bands. The specificity of the IGFBP-4 RIA was further
examined by size fractionation of human serum, to some
aliquots of which had been added 250 pg/L rhIGFBP-4, on
a Superdex G-75 column in a fast protein liquid chromatography system followed by RIA assayof the column fractions.
Both with and without added IGFBP-4, the major peak of
IGFBP-4 immunoreactivity eluted at an equivalent retention
RU
A
B
FOR IGFBP-4
1393
(Table 1). RIAs with exogenous IGFBP-4 added to serum
revealed that more than 90% of the added IGFBP-4 was
recovered (Table 2). The amount of IGFBP-4 detected by this
RIA was not significantly affected by repeated freezing and
thawing of the serum assayed (data not shown).
Consistent with the observation that dibutyryl CAMP,
PTH, and 1,25-dihydroxyvitamin Da increased both IGFBP-4
mRNA and protein levels, as determined by Northern blot
and Western ligand blot analysis, respectively (10, 17), the
IGFBP-4 RIA detected increased amounts of IGFBP-4 in the
conditioned medium after treatment of TE85 human osteosarcomacellswith these compounds. Treatment of TE85 cells
with dibutyryl CAMP increased the IGFBP-4 protein level in
a dose-dependent manner, with a maximal increaseof 5-fold
seen at a 0.1 mmol/L concentration (data not shown). In
addition, treatment of TE85 cells with PTH and 1,25-dihydroxyvitamin Da also increased IGFBP-4 levels in the conditioned medium (Fig. 7).
Serum IGF-I showed a negative correlation with both age
(r = -0.63; P < 0.001) and serum IGFBP-4 (r = -0.18; P <
0.05; Table 3; data not shown). Employing the IGFBP-4 RIA
to assay the sera of normal healthy men and women demonstrated a significant positive correlation with age (r = 0.50;
P < 0.001; Fig. 8 and Table 3) and serum PTH (r = 0.26; P <
0.01; Fig. 9). The mean serum PTH level was significantly
higher (mean ? SD, 212 ? 72 DS.169 ? 40 rig/L; P < 0.001)
in the 61-87 yr age group than that in the 23-40 yr age group
(Fig. 9).
12345
klla
26+
Discussion
18+
7-W
BP-4
SERUM
FIG. 6. Immunoblot analysis employing IGFBP-4 antiserum. A, Electrophoresis under nonreducing conditions of 10 ng IGFBP-4 purified
from PC3 cell-conditioned media (lane 11, 2 ng rhIGFBP-4 (lane 21,
0.05 mL lo-fold concentrated PC3 cell-conditioned medium before
IGF-II affinity chromatography (lane 3),0.05 mL IGF-II affinity column flow-through fraction of PC3 cell-conditioned medium concentrate (lane 4), and 0.05 mL lo-fold concentrated U2 cell-conditioned
medium (lane 5). B, Electrophoresis under reducing conditions of 10
ng rhIGFBP-4 (BP-41 and 0.01 mL human serum (SERUM). Samples
were electrophoresed through a lo-20% gradient SDS-polyacrylamide gel, blotted to a nitrocellulose membrane, incubated with a
1:2000 dilution of the IGFBP-4 antiserum, and visualized with the
chemiluminescent detection system described in Materials
and Methods. Migration
positions of size markers are indicated at the left (in
kilodaltons).
volume, between the 43 and 25 kDa size markers (data not
shown).
The potential for IGFs to interfere with the IGFBP-4 RIA
was evaluated by adding purified IGF-I and IGF-II to the
reaction mixtures. Addition of as much as 1 mg/L IGF-I or
IGF-II to either IGFBP-4 standard or serum had no significant
effect on IGFBP-4 recovery, demonstrating that this antiserum detects both free and IGF-bound forms of IGFBP-4
A specific and sensitive RIA for IGFBP-4 has been developed by employing antiserum raised against purified rhIGFBI’-4 fusion protein and using purified rhIGFBP-4 astracer
and standard. Employing this assaywe have shown 1) identical displacement of [1251]IGFBP-4tracer from antiserum by
purified PC3 cell-derived IGFBP-4 and rhIGFBP-4; 2) parallel
displacement by human serum and TE85 cell-derived
IGFBP-4; 3) no displacement by conditioned medium derived from U2 cells, which do not express IGFBP-4 (10); 4)
TABLE
1. Effect of preincubation
immunoreactivity
of IGFBP-4
with IGF-I or IGF-II on the
Sample
+IGF-I
+IGF-II
BP-4 + 0 pg/L IGF=
4.63b
4.69
BP-4 + 1 &L IGF
4.45
4.80
BP-4 + 10 &L IGF
4.95
4.90
BP-4 + 100 pg/L IGF
4.83
4.83
BP-4 + 1000 CLgn IGF
4.33
4.46
Serum + 0 pg/L IGF
4.70
5.20
Serum + 1 pg/L IGF
4.99
4.55
Serum + 10 yg/L IGF
5.32
5.13
Serum+ 100 ,Ugn IGF
4.74
4.47
Serum + 1000 pg/L IGF
4.84
4.52
Purified rhlGFBP-4 0.05 mL of 10 pg/L or 0.05 mL of 1:40 diluted
normal human serum samples was incubated for 1 hat 22C with 0.05
mL IGF-I, IGF-II, or vehicle. Incubated samples were assayed for
IGFBP4 by RIA as described in Materials
and Methods.
IGF-I or
IGF-II alone assayed at concentrations as great as 500 &L did not
have any effect on the binding of IGFBP-4 tracer to the antiserum.
a IGF in this column refers to IGF-I or IGF-II.
b IGFBP-4 values shown are in micrograms per L and are the mean
of duplicate determinations.
HONDA
TABLE
2. Recovery
Serum no.
1
of exogenous
Added (fig/L)
IGFBP-4
0
1.95
3.9
7.8
15.6
2
1000
in serum
Observed
(pg/L)
% Recovery
2.9
5.2
7.4
11.9
17.3
I'TH
r=0.54
106.9
118.4
97.7
95.2
0”
po
0
A@A
0
o "Q4 Aati A
%0 80
oA ",
@'O
A
400
B
-
0
P<O.OOl
600
200
A
0
0
Serum (0.05 mL of 1:20 dilution)
was mixed with various
concentrations
of IGFBP-4
(0.05 mL) or buffer before RIA. IGFBP-4
values
are in micrograms
per L and are the mean of duplicate
determinations. The amount
of IGFBP-4
obtained
without
added IGFBP-4
represents endogenous
IGFBP-4
levels in 1:lOO diluted serum for the two
separate
serum samples studied. The percent recovery
obtained
in the
presence of exogenous
IGFBP-4
was calculated
by taking
the ratio of
the observed
IGFBP-4
level and the expected
sum of the endogenous
and exogenous
IGFBP-4
levels.
250
800
108.3
108.8
111.2
93.5
5.3
7.7
10.9
12.8
19.9
0
1.95
3.9
7.8
15.6
JCE & M . 1996
Vol81.
No 4
ET AL.
20
40
60
100
80
Age (years)
FIG. 8.
normal
determine
between
(a), and
for the
Relationship
between
serum
levels of IGFBP-4
and age in
healthy
men and women.
An IGFBP-4
RIA was employed
to
the IGFBP-4
concentrations
in serum from 24 individuals
21-40 yr of age (O), 37 individuals
between
41-60 yr of age
41 individuals
between
61-87 yr of age (0). Linear regression
entire population
yielded
r = 0.54 and P < 0.001.
1000
* P<O.OOl
800
r=0.26
0
P<O.Ol
0
0
00
0
0
cl
!j
0
0
1
IO"
Age group
23-40
41-60
61-87
Values
(yr)
of IGF-I,
IGF-II,
Age (yr)
No.
IGF-I
32% 8
522 6
68 2 5
24
37
41
224 2 86
133 -c 48”
116 5 35”
are the mean
t-
(pg/L)
and IGFBP-4
IGF-II
200
300
400
500
PTH (rig/L)
lO-'
mol/L
3. Serum levels
men and women
I
100
I
NY8
FIG. 7. Effects of stimulators
of IGFBP-4
production
on the amount
of IGFBP-4
in the conditioned
medium
of TE85 human osteosarcoma
cells. Cells were treated
under serum-free
conditions
with PTH dissolved in DMEM
containing
0.1% BSA or 1,25-dihydroxyvitamin
Da
[1,25-(OH),D,]
dissolved
in ethanol
to a final ethanol
concentration
of 0.01% in the culture.
Conditioned
medium
samples
were collected
after 48 h and subjected
to the IGFBP-4
RIA to determine
IGFBP-4
concentrations.
Values
are the mean ? SD of four replicate
cultures
per treatment.
Significance
was determined
by comparisons
with the
control.
TABLE
healthy
600
(pg/L)
563 2 78
642 +- 153b
573 2 120
in normal
IGFBP-4
(pg/L)
404 ? 156
447 2 87
546 2 135”
SD.
a P < 0.001 us. 23-40 yr.
bP < 0.05 us. 23-40 yr.
"P < 0.05 vs. 41-60 yr.
comigration of immunoreactive material in serum with rhIGFBP-4; 5) coelution of rhIGFBP-4 and material from human
serum that was immunoreactive with IGFBP-4 antiserum
upon high pressure liquid chromatography gel filtration using a Superdex G-75 column (data not shown); 6) more than
90% recovery of the exogenously added recombinant
FIG. 9. Relationship
between
serum levels of IGFBP-4
and PTH in
normal
healthy
men and women.
IGFBP-4
RIA was employed
to
determine
the IGFBP-4
concentration,
and a PTH midmolecule/Cterminal
RIA kit (Nichols
Institute,
San Juan Capistrano,
CA) was
used to determine
the PTH concentration
in serum from 24 individuals between
21-40 yr of age (O), 37 individuals
between
41-60 yr of
age (a), and 41 individuals
between
61-87
yr of age (0). Linear
regression
for the entire population
yielded
r = 0.26 and P < 0.01.
IGFBP-4 to serum; 7) no reaction of the antiserum with other
purified IGFBPs when tested at concentrations ashigh as0.5
mg/L; and 8) increasesin IGFBP-4 mRNA (10,17) and protein after treatment with agents that increase IGFBP-4 production. These findings collectively demonstrate that this
IGFBP-4 RIA is specific for IGFBP-4 and applicable for
IGFBP-4 measurements in both serum and conditioned
medium.
The purified rhIGFBP-4 fusion protein was employed to
raise antibodies in guinea pigs. The antiserum raised in
guinea pigs reacted with the intact glycosylated 28- to 29-kDa
form, the intact nonglycosylated 25-kDa form, and the proteolytically cleaved 16-kDa fragment of IGFBP-4. The possibility that the band migrating at 40-45 kDa that is also
recognized by the IGFBP-4 antiserum is a dimer of IGFBP-4
that has not been completely disassociated by the electrophoresis buffer remains to be investigated. The IGFBP-4 RIA
data showing equivalent displacement of the tracer by human and rabbit sera are consistent with our previous lack of
RIA FOR IGFBP-4
success in raising rabbit antibodies against synthetic IGFBP-4
peptides. Surprisingly, although the rat and human IGFBP-4
amino acid sequences align with 90% amino acid identity, the
rat serum was ineffective in displacing the tracer, suggesting
that the conformations of human and rat IGFBP-4s may differ
in the region of the epitope with which the antiserum reacts.
The failure of added IGF to affect the recovery of IGFBP-4 is
a desirable feature of the IGFBP-4 RIA because it suggests
that serum samples can be directly assayed without further
extraction. This is consistent with an epitope distant from the
IGF-binding site. On the contrary, antiserum from the second
guinea pig (no. 12) that produced high titer antibodies to the
rhIGFBP-4 fusion protein exhibited reduced recovery of
IGFBP-4 when IGFs were added (data not shown). The
IGFBP-6 RIA recently developed by Baxter and Sanders (24)
is similar to the latter, as both IGF-I and IGF-II markedly
inhibited IGFBP-6 tracer binding, suggesting that the binding
site for this antibody was in close proximity to the IGFbinding site in IGFBP-6. These observations emphasize that
determining the interference of the IGFs in an IGFBP RIA,
which will depend on the antiserum used, is an important
facet of the validation of an IGFBP RIA.
Although it was previously known that adult plasma contains 50-kDa proteins comprised of multiple IGFBPs not fully
saturated with IGFs (25) and that IGFBP-4 circulates in the
serum, the amount of IGFBP-4 in serum had not been previously quantitated. Comparing the amount of IGFBP-4 detected by this IGFBP-4 RlA in human serum with reported
values for other IGFBPs reveals that the mean level of
IGFBP-4 in adult human serum is higher than those of
IGFBP-1 (26), IGFBP-2 (27), and IGFBP-6 (24); is similar to
that of IGFBPS (28); and is less than 10% of the IGFBP3 level
(29,30). The trend of serum IGFBP-4 to increase with age, as
revealed by this IGFBP-4 RIA, is similar to that reported for
IGFBP-1 and IGFBP-2 (31,32), but dissimilar to reports that
IGFBP3 (29,30) and IGFBP5 (28) decline with age, suggesting that the human serum levels of the different IGFBPs are
differentially regulated with advancing age. Also, in contrast
to IGFBPS, which showed a significant positive correlation
with IGF-I or IGF-II concentration
(281, serum IGFBP-4
showed a weak negative correlation with IGF-I concentration, suggesting that different mechanisms may regulate the
amounts of IGFBP-4 and IGFBP5 in human serum.
Although the exact functional role for serum IGFBP-4 is
not clear at this time, in vitro studies have shown IGFBP-4 to
be a potent inhibitor of IGF actions in bone cells and other
cell types under a variety of culture conditions (8, 12). It is
believed that IGFBP-4 binds to IGFs and thereby prevents the
IGFs from binding to their receptors (12). Western ligand blot
analysis revealing increased serum IGFBP-4 in hip fracture
patients with elevated serum PTH levels, serum PTH levels
showing a positive correlation with serum IGFBP-4 levels
(33), with PTH increasing IGFBP-4 protein and mRNA levels
in human bone cells, and the increase in serum IGFBP-4 with
age correlating to the PTH concentration are all consistent
with the interpretation
that PTH may be a regulator of
IGFBP-4 production in vitro and in vim. We suggest that
secondary hyperparathyroidism,
which occurs as a consequence of age, could induce the inhibition of osteoblast proliferation by stimulating the production of IGFBP-4 in the
locale of bone-remodeling sites. Although much remains to
be learned concerning the functional role and physiological
regulation of IGFBP-4 in human serum, this IGFBP-4 RIA will
facilitate future studies on the regulation of IGFBP-4 production in DizIo, which may lead to our understanding of the
potential role of this protein in modulating the inhibitory
effects of PTH and other hormones on bone formation.
Acknowledgments
The authors acknowledge
Jacquelyn
Douglas,
Joe Rung-Aroon,
and
Sylvia Morales
for their excellent
technical
support;
Jamie Lopez for
secretarial
assistance; and the Jerry L. Pettis Veteran’s
Hospital
Medical
Media Service for illustrations.
References
1. Spencer EM. 1991 Modern concepts of insulin-like growth factors. New York:
Elsevier.
2. Baxter RC, Martin JL. 1989 Binding proteins for the insulin-like growth factors:
structure, regulation and function. Frog in Growth Factor Res. 1:49-68.
3. Shimasaki S, Ling N. 1991 Identification
and molecular characterization
of
insulin-like growth factor binding proteins (IGFBP-1, -2, -3, -4, -5 and -6). Prog
Growth Factor Res. 1:243-266.
4. Rosenfeld
RG, Lamson
G, Pham H, et al. 1990 The insulin-like growth factor
binding proteins. Recent Frog Harm Res. 4699-163.
5. Mohan
S. 1993 Insulin-like
growth factor binding proteins in bone cell regulation. Growth Regul. 3:65-68.
6. Rechler
MM. 1993 Insulin-like
growth factor binding proteins. Vitam Harm.
471-114.
7. Jones JI, Clemmons DR. 1995 Insulin-like growth factors and their binding
proteins: biological actions. Endocr Rev. 16:3-34.
8. Mohan
S, Bautista
C, Wergedal
J, Baylink
DJ. 1989 Isolation of an inhibitory
insulin-like growth factor (IGF) binding protein from bone cell conditioned
medium: a potential local regulator of IGF action. Proc Nat1 Acad Sci USA.
86:8338-8342.
9. Shimonaka
10.
11.
12.
13.
14.
M, Schroeder
R, Shimasaki
S, Ling N. 1989 Identification
of a
novel binding protein for insulin-like growth factors in adult rat serum. Biothem Biophys Res Commun. 165:1890195.
La Tour D, Mohan
S, Linkhart
TA, Baylink
DJ, Strong DD. 1990 Inhibitory
insulin-like growth factor binding protein: cloning, complete sequence, and
physiologic regulation. Mol Endocrinol. 4:1806-1814.
Shimasaki
S, Uchiyama
F, Shimonaka
M, Ling N. 1990 Molecular cloning of
the cDNAs encoding a novel insulin-like growth factor binding protein from
rat and human. Mol Endocrinol. 4:1451-1458.
Mohan
S, Nakao Y, Honda Y, et al. 1995 Studies on the molecular mechanisms
by which insulin-like
growth factor (IGF) binding protein-4 (IGFBP-4) and
IGFBP-5 modulate IGF actions in bone cells. J Biol Chem. 270:20424-20431.
Mohan
S, Strong DD, Linkhart
TA, Baylink
DJ. 1994 Regulation and actions
of insulin-like growth factor binding protein (IGFBPl-4 and IGFBP-5 in bone:
physiological and clinical implications. In: Baxter RC, Gluckman PD, Rosenfeld
RG, eds. The insulin-like growth factors and their regulatory
proteins. New
York: Excerpta Medica; 205-215.
Kanzaki
S, Hilliker
S, Baylink
DJ, Mohan
S. 1994 Evidence that human bone
cells in culture produce insulin-like growth factor binding protein-4 and -5
proteases. Endocrinology.
134383392.
15. Conover
CA,
Durham
SK, Zapf
J, Masiarz
FR, Kiefer
MC.
1995
Clevage
analysis of insulin-like growth factor (IGF)-dependent
IGF-binding
protein-4
proteolysis and expression of protease-dependent
IGF-binding
protein-4 mutants J Biol Chem. 270:4395-4400.
16. Chemausek
SD, Smith
CE, Duffin
KL, Busby
WH,
Wright
G, Clemmons
DR.
1995 Proteolytic
cleavage of insulin-like
growth factor binding protein-4
(IGFBP-4). Localization of cleavage site to non-homologous
region of native
IGFBP-4. J Biol Chem. 270:11377-11382.
17. Scharla
SH, Strong DD, Rosen
C, et al. 1993 1,25-Dihydroxyvitamin
D,
increases expression of insulin-like growth factor binding protein-4 (IGFBP-4)
in human osteoblast-like cells in vitro and elevates IGFBP-4 serum levels in viva
J Clin Endocrinol Metab. 77:1190-1197.
18. Knutsen
R, Honda
Y, Strong
DD,
Sampath
K, Baylink
DJ, Mohan
S. 1995
Regulation of insulin-like
growth factor system components by osteogenic
protein-l in human bone cells. Endocrinology.
136857-865.
19. Durham
SK, Riggs L, Conover C. 1994 The insulin-like
growth factor-binding
protein-4 (IGFBP-4).IGFBP-4
protease system in normal human osteoblast-like
cells: regulation by transforming
growth factor-p. J Clin Endocrinol Metab.
791752-1758.
20. Ausubel
FM, Brent R, Kingston
RE, et al. 1989 Current protocols in molecular
biology. New York: Wiley and Sons.
HONDA
21. Sambrook
J, Fritsch
EF, Maniatis
T. 1989 Molecular
cloning-a
laboratory
manual,
2nd ed. Cold Spring Harbor:
Cold Spring
Harbor
Laboratory.
22. Mohan
S, Baylink
DJ. 1995 Development
of a simple,
valid method
for the
complete
removal
of insulin-like
growth
factor binding
proteins
from IGFs in
human
serum and other biological
fluids:
comparison
with acid-ethanol
extraction
and C-18 Sep-Pak
separation.
J Clin Endocrinol
Metab.
80637647.
23. Towbin
H, Staehlin
T, Gordon
J. 1979 Electrophoretic
transfer
of proteins
from acrylamide
gels to nitrocellulose
sheets: procedure
and some applications. Proc Nat1 Acad Sci USA. 764350-4354.
24. Baxter RC, Sanders
H. 1992 Radioimmunoassay
of insulin-like
growth
factor
binding
protein-6
in human
serum and other biological
fluids.
J Endocrinol,
134133-139.
25. Hintz
RL, Liu F, Rosenfeld
R, Kemp SF. 1981 Plasma somatomedin-binding
proteins
in hypopituitarism:
changes during
growth
hormone
therapy.
J Clin
Endocrinol
Metab. 53:100-104.
26. Busby WH, Snyder DK, Clemmons
DR. 1988 Radioimmunoassay
of a 26,000dalton
plasma
insulin-like
growth
factor-binding
protein:
control
by nutritional variables.
J Clin Endocrinol
Metab. 67122551230.
27. Clemmons
DR, Snyder
DK, Busby
WH. 1991 Variables
controlling
the
secretion
of insulin-like
growth
factor binding
protein-2
in normal
human
subjects.
J Clin Endocrinol
Metab.
73727-733.
ET AL.
JCE & M . 1996
Volt31 . No 4
28. Mohan
S, Libanati
C, Dony
C, Lang K, Srinivasan
N, Baylink
DJ. 1995
Development,
validation,
and application
of a radioimmunoassay
for insulinlike growth
factor binding
protein-5
in human
serum and other biological
fluids.
J Clin Endocrinol
Metab.
8012638-2645.
29. Baxter
RC, Martin
JL. 1986 Radioimmunoassay
of growth
hormone-dependent insulin-like
growth
factor binding
protein
in human
plasma. J Clin Endocrinol
Metab. 78x1504-1512.
30. Blum WF, Ranke MB, Kietzmann
K, Gauggel
F, Zeisel HJ, Bierich
JR. 1990
A specific
radioimmunoassay
for the growth
hormone
(GH)-dependent
somatomedin-binding
protein:
its use for diagnosis
of GH deficiency.
J Clin
Endocrinol
Metab.
70:1292-1298.
31. Hall K, Lundin
G, Povoa G. 1988 Serum levels of the low molecular
weight
form of insulin-like
growth
factor binding
protein
in healthy
subjects
and
patients
with growth
hormone
deficiency,
acromegaly
and anorexia
nervosa.
Acta Endocrinol
(Copenh).
118:321-326.
32. Blum WF, Horn N, Kratzsch
J, et al. 1993 Clinical
studies
of IGFBP-2
by
radioimmunoasssay.
Growth
Regul. 3100-104.
33. Rosen C, Donahue
LR, Hunter
S, et al. 1992 The 24/25 kD serum insulin-like
growth
factor binding
protein
is increased
in elderly
osteoporotic
women.
J Clin Endocrinol
Metab. 74:24-27.