вход по аккаунту


Local disruption of the insulin-like growth factor system in the arthritic joint.

код для вставкиСкачать
Vol. 39, No. 9, September 1996, pp 1556-1565
8 1996, American College of Rheumatology
O&ective. To identw differences in levels of insulinlike growth factor (IGF) and IGF binding proteins (IGFBPs) between 30 patients with arthritis (14 with rheumatoid arthritis [RA], 16 with osteoarthritis [OA]) and 11
normal control subjects. IGF and IGFBP levels were
correlated to the disease activity marker C-reactiveprotein
(CRP) to determine whether they were disease related. We
also examined the degree of proteolytic modification of the
Methods. Radioimmunoassays were used for measuring IGF and IGFBP-3 levels; CRP was measured by
enzyme-linked immunosorbent assay. Sodium dodecyl
sulfate-polyacrylamide gel electrophoresis followed by
Western blotting, chemiluminescence, and autoradiography were used for visualizing binding proteins.
Results. There was a significant increase in synovial fluid levels of both IGF-1 and IGFBP-3 in both RA
and OA. This resulted in an elevated IGFBP-3 to IGF
molar ratio of 1.49 in the OA group and 1.47 in the RA
group, compared with 0.86 in the normal control group
(P = 0.0002 for both). A significantly lower degree of
IGFBP-3 proteolysis was also seen in the synovial fluids
from the patients compared with the controls. There
were significant correlations between the CRP level and
levels of IGF-1, IGF-2, and IGFBP-3 in the RA patients
(r = 0.62-0.898, P = 0.04-0.0007).
Conclusion. There was significant local disruption of the IGF system in patients with arthritis. This
may result in a lower amount of IGF that is able to bind
to IGF receptors in the arthritic joint. Levels of IGF-1,
IGF-2, and IGFBP-3 all correlated with the CRP level in
Supported by the Throckmorton Scholarship, University of
J. K. Fernihough, BSc, M. E. J. Billingham, PhD, S. CwyfanHughes, PhD, J. M. P. Holly, PhD: University of Bristol, Bristol, Avon,
Address reprint requests to J. K. Fernihough, BSc, University
of Bristol Department of Medicine, Rheumatology Unit, Bristol Royal
Infirmary, Lower Maudlin Street, Bristol, Avon, BS2 8HW, UK.
Submitted for publication February 5, 1996; accepted in
revised form April 24, 1996.
patients with RA, which indicates the possibility that the
IGF system is involved in the disease process.
Arthritic diseases, among other characteristics,
result in an eventual loss of articular cartilage. Cartilage
damage has been partly explained as the final result of
the catabolic mechanisms in the joint space exceeding
the anabolic pathways in osteoarthritis (OA): in rheumatoid arthritis (RA) this occurs by mechanisms akin to
tumor invasion (for review, see ref. 1).The predominating catabolic processes are thought to be cytokinedriven, up-regulating proteases responsible for the
degradation of collagens and proteoglycans.Since insulinlike growth factor 1 (IGF-1) is partly able to reverse the
effect of such catabolic stimuli as interleukin-1 (IL-1)
and tumor necrosis factor (TNF) in vitro (2,3), it is
important to assess the regulation of IGF in vivo. IGF-1
has a major role in promoting the anabolic metabolism
of chondrocytes, originally demonstrated by the stimulation of 35~-sulfateincorporation into cartilage (4).
Addition of IGF-1 can result in similar levels of stimulation as that gained by 20% fetal calf serum (5,6),
indicating that IGF-1 is the major factor in serum
responsible for its growth-promoting properties. This
response depends on the differentiated state of the
chondrocytes (7) and the age of the cartilage (8,9); it is
not dependent on growth hormone stimulation (7),
being therefore a direct function of IGF. This anabolic
effect is partially able to reverse the inhibitory action of
IL-1 (2,3), confirming its important role in maintaining a
balance between anabolic and catabolic processes within
the joint space. Owing to the lack of human tissue, much
of this work has been based on studies of animal tissues.
Serologic studies in humans have yielded contradictory results regarding the level of IGF-1 as it relates
to disease status (10,ll) and clinical symptoms (12), and
have provided no conclusive evidence of a systemic
change in IGF metabolism in arthritis. In human cartilage, however, there appears to be an up-regulation of
IGF-1 messenger RNA and protein in the lesions of OA
cartilage obtained from the femoral head (13), as compared with the low levels in macroscopically normal
areas. This appears to be in contradiction of the fact that
in experimental arthritis, chondrocytes show an unresponsiveness to IGF-1 (14) that does not relate to a
defect in IGF-1 receptors or a reduction in their number
(15). This discrepancy has been partly explained by the
presence of 6 binding proteins for IGF that can prevent
the binding of IGF to its receptor (15). There are no
intracellular stores of IGF; however, IGF binding protein 3 (IGFBP-3), the predominant binding protein
present in serum, binds more than 90% of circulating
IGF and prolongs its half-life from a few minutes to 18
hours (16). Release of IGF is believed to be achieved by
limited proteolysis of the binding protein, thereby reducing the affinity for IGF. T h e high levels of IGF in the
circulation are therefore latent; but proteases present at
the tissue level enable the site-specific action of IGF-l(l7).
The complex interactions between the 6 IGFBPs
and IGF-1 and IGF-2 have been reviewed by Jones and
Clemmons (18). It is known that chondrocytes and bone
cells provide a local source of binding proteins (19-21),
and that the highly negatively charged cartilage matrix
environment may be capable of stripping IGF from its
binding proteins (22). Metalloproteinases present in the
joint space may also be responsible for degrading
IGFBPs, affecting the way in which IGF acts in the
arthritic joint (23). T h e role of IGFBPs in the response
of a tissue to IGF-1 in an environment containing other
cytokines has been reported in studies outside the field
of arthritis. For example, we have previously shown that
the sensitivity of human dermal fibroblasts to stimulation by IGFs can be increased or decreased by cytokinemediated changes in IGFBP production (24). Despite
the discovery of IGF-1 in 1957 as an anabolic stimulus
for chondrocytes, there is very limited knowledge about
how the action of proteases for IGFBPs and the IGFBPs
themselves affect the function of IGF-1 in arthritic joint
tissues. In this study, we sought to determine the profile
of IGFs and IGFBPs in the arthritic joint space compared with the nonarthritic joint, and t o determine any
links with disease activity. The proteolytic potential of
the two types of synovial fluid was also evaluated.
Patients and controls. Samples of serum and synovial
fluid were obtained from 14 RA patients, who ranged in age
from 49 to 82 (mean age 63; ma1e:female ratio 1:6), and from
16 OA patients, who ranged in age from 50 to 93 (mean age 69;
male:female ratio 9:7). In both groups, disease duration varied
from <2 years to >10 years since first presentation. All
patients met the appropriate American College of Rheumatology (formerly, the American Rheumatism Association) criteria for the disease (25,26). Samples of serum and synovial
fluid were also obtained from 11volunteer donors who did not
have arthritis. The age of the controls ranged from 22 to 52
(mean age 41), and the ma1e:female ratio was 8:3.
Synovial fluid was aspirated from the suprapatellar
pouch and was stored at -70°C until analyzed. Serum was
removed from clotted venous blood, and samples were also
stored at -70°C until used.
IGF-1, IGF-2, and IGFBP-3 determinations. Levels of
IGFs 1 and 2 and IGFBP-3 were determined by radioimmunoassay (RIA). All tracers were iodinated using the chloramine-T method. For measurement of IGF-1 and IGF-2, removal of endogenous binding proteins was performed using an
extraction procedure according to the method described by
Bowsher et a1 (27). The IGF-1 used for the standard curve and
tracer was obtained from Kabi Pharmacia (Stockholm, Sweden). The monoclonal antibody to IGF-1 was obtained from
Blood Products (Elstree, Hertfordshire, England). The lower
limit for the detection of IGF-1 was 0.1 @liter.
For the IGF-2 assay, an excess of IGF-1 was added to
each sample to saturate any residual binding proteins not
removed by the extraction procedure. The IGF-2 used for the
standard curve and tracer was also obtained from Kabi Pharmacia; the monoclonal antibody to IGF-2 was a gift from Dr.
Ann White (Manchester University, Manchester, England).
The lower limit of detection of IGF-2 was 0.5 pg/liter. Although the IGF assay was originally designed for serum
samples, synovial fluid samples diluted in parallel with the
standard curve and were therefore treated in the same way. To
validate this method further for synovial fluid, we performed a
test to ensure that 100% recovery of exogenously added IGF
was possible.
The IGFBP-3 method used was the same as that
described by Cheetham et a1 (28). Serum samples for the
IGFBP-3 assay were diluted 1:50 before assaying; synovial fluid
samples were diluted 1:8 (RA or OA) or 1:4 (control). The
antibody used was an in-house polyclonal antibody raised
against recombinant nonglycosylated IGFBP-3 and was used at
a dilution of 1:8,000. The standard glycosylated IGFBP-3 was
a generous gift from Dr. C. Maack (Celltrix, Santa Clara, CA).
Western ligand blotting. We used Western ligand
blotting to show the distribution of all IGFBPs, according to a
modification of the method described by Coulson et a1 (29).
This method relies on the IGFBPs on the nitrocellulose
membrane retaining their binding capacity for IGF; therefore,
fragments of IGFBPs with proposed lower affinities for IGF
are not visualized. Briefly, the binding proteins in samples were
separated on a 12.5% sodium dodecyl sulfate (SDS) gel, and
then transferred to a nitrocellulose membrane. Probing with
radiolabeled IGF-1 and IGF-2 enabled the binding proteins to
be visualized by autoradiography. The radiolabeled IGF-1 and
IGF-2 used were the same as those used in the RIAs added to
the probing buffer to give -5,000 counts per minute each per
100 p1 of buffer. For serum samples, 2.5 p1was loaded per well;
for synovial fluids, the volume was increased to 4 pl per well.
Immunoblotting. Immunoblotting to detect in vivo
modification of IGFBPs was performed on the same membranes as those used in the Western ligand blotting. Immunoblotting was performed according to the method of Cwyfan-
Hughes et a1 (30). The membranes were washed in buffer
containing antibodies to either IGFBP-2 (mouse monoclonal
antibody; Sandoz Pharma, Basel, Switzerland) or IGFBP-3
(same antibody as that used for the IGFBP-3 RIA). These
antibodies bound to both whole and fragmented forms of the
protein. Detection was made possible using a secondary antibody coupled with enhanced chemiluminesence, according to
the manufacturer’s instructions (Amersham, Buckinghamshire, England), and exposure to x-ray film at room temperature. The secondary antibodies used were goat anti-mouse IgG
for the IGFBP-2 immunoblotting and goat anti-rabbit IgG for
the IGFBP-3 immunoblotting (Sigma, Poole, Dorset, England); both were coupled to horseradish peroxidase. This
method is more sensitive than a radioligand blot for detecting
IGFBPs because it does not rely on the retention of functional
binding capacity by the IGFBPs.
Proteolytic activity determination. Proteolytic activity
in synovial fluids was assessed by incubating 25 p1 of each
sample for 5 hours with a fixed amount of radiolabeled
nonglycosylated IGFBP-3, based on the method described by
Lamson et a1 (31). Inhibition of activity was tested by the
addition of EDTA. Normal human serum served as a negative/
low control, and serum from a woman in her third trimester of
pregnancy, when proteases for IGFBP-3 are up-regulated.
served as a positive control. An aliquot of this mixturewas then
applied to a 12.5% SDS gel, and the fragments of IGFBP-3
were separated from the whole protein. The gel was then dried
for 1.5 hours at 75°C and again exposed to x-ray film at -70°C.
C-reactive protein (CRP) level determination. Levels
of CRP were used as an indication of disease activity, and were
measured by 2 different methods. For samples with a CRP
value >10 mg/liter an immunoturbidimetric assay was used
according to the manufacturer’sinstructions (Randox, Cumlin,
County Antrim, Northern Ireland). The principle on which this
assay is based is that the sample is treated with a specific
antiserum to form a precipitate, which is measured turbidimetrically at 340 rim. This assay is routinely run in the Chemical
Pathology Department of the Bristol Royal Infirmary.
For samples with a level of CRP < l o mg/liter, a
high-sensitivity enzyme-linked immunosorbent assay was performed manually. Goat anti-human CRP antibody was used as
the coating antibody, and the conjugate used was a peroxidaseconjugated rabbit anti-human CRP antiserum (Dako, High
Wycombe, Buckinghamshire, England). The chromogenic substrate was 1,2-phenylenediaminedihydrochloride (Dako). This
assay is run routinely in the Chemical Pathology Department
of the Bristol Royal Infirmary.
Statistical analysis. All populations were assumed to
be normally distributed (by kurtosis and skewness analysis;
skewness range -0.69-1.4). F tests showed that homogeneity
of variances could not be assumed; therefore, unpooled t-tests
were performed on some groups of data.
IGF-1, IGF-2, and IGFBP-3 levels in serum and
synovial fluid. There was no significant difference in the
level of IGF-1 in serum from OA patients (mean 146
pg/liter, SD 60) and normal control subjects (mean 167
& l i t e r , SD 72). T h e mean level of IGF-1 in serum from
RA patients was 119 pg/liter (SD 43), which was 48
pg/liter lower than that in the normal control subjects.
This difference approached significance (P = 0.0529).
The 95% confidence interval for the difference between
the two mean levels ranged from -26 to 121 pg/liter,
indicating that the difference might have physiologic
consequences. There was no signifcant difference in
serum IGF-1 levels between RA and OA patients.
Serum IGFBP-3 levels were lower in RA patients
(mean 4,384 pg/liter, SD 1,287; P = 0.005) and in OA
patients (mean 4,904 &liter, SD 1,604; P = 0.049)
compared with normal levels (mean 6,261 pg/liter, SD
1,773). This was also the case for the serum 1GF-2
values. RA patients had a mean of 536 pgAiter (SD 148;
P = 0.0035) and OA patients had a mean of 599 pg/liter
(SD 161;P = 0.031), which was significantly lower than the
level in the normal controls (mean 753 &liter, SD 184).
The most striking differences, however, were seen
in the synovial fluid, where the levels of significance were
much higher (Figures 1A-C). T h e mean and SD values
for IGF-1 were 26 and 6 pg/liter for the control subjects,
78 and 26 pg/liter for the OA patients, and 72 and 26
pg/liter for the RA patients. For IGF-2, the values were
111 and 21 &liter for the control subjects, 134 and 42
pg/liter for the OA patients, and 217 and 109 pg/liter for
the RA patients. For IGFBP-3, the mean and SD levels
were 635 and 243 pg/liter for the control subjects, 1,656
and 594 pg/liter for the OA patients, and 2,244 and 1,004
& l i t e r for the RA patients. These differences resulted
in a molar ratio of IGFBP-3 to total IGF that was >1 in
the arthritic joint space (Table 1).
These differences also resulted in a change in the
amount of IGFs distributed between the serum and
synovial fluid. Table 2 shows the increases in synovial
fluid levels of IGF-1, IGF-2, and IGFBP-3, expressed as
a percentage of the levels in serum, for RA and OA
patients. In RA patients, there was a correlation between the serum and synovial fluid levels of IGF-1 (r =
0.711, P = 0.006). Such a correlation was not observed in
the OA patients or the normal controls. There were also
correlations between IGF-1 and IGF-2 levels in both the
serum and synovial fluid of RA patients. The correlation
coefficients for these were 0.63 (P = 0.017) for serum
and 0.64 (P = 0.015) for synovial fluid. These correlations were not found in OA patients; however, in the
normal control subjects, a correlation between IGF-1
and IGF-2 levels was seen in the synovial fluid (r =
0.723, P = 0.021).
Correlations of IGF-1 with CRP levels. Figure 2
shows that there was a strong correlation between CRP
levels and serum IGF-1 levels in patients with RA (r =
I-test OA va. RA: NS
I-test RA vs. normal: P = 0.00001
I-test O A vs. normal: P = O.OOOOO1
Normal controls
RA patients
OA patients
RA patienb
lml ;
Normal controls
OA patients
I-t-t OA VS. RA: P=O.O15
I-test OA vs. normal: NS
I-test RA vs. normal: P = 0.003
I-test OA va. RA: NS
I-teat I
U vs. normd: P = 0.000037
I - t a t OA va. norm& P= 0.00O004
Figure 1. Levels of A, insulin-like growth factor type 1 (IGF-1), B,
IGF-2, and C, IGF binding protein type 3 (IGFBP-3) in synovial fluid
obtained from osteoarthritis (OA) patients, rheumatoid arthritis (RA)
patients, and normal control subiects. Each point represents the mean
of 3 replicates determined by radioimmunoassay. NS = not significant.
OA patients
RA patients
Normal controls
0.898, P = 0.0007). Significant correlations were also
seen with serum and synovial fluid IGF-2 and IGFBP-3,
as well as synovial fluid IGF-1, levels (Table 3). No such
Table 1. Mean molar ratios of IGFBP-3 to total IGF in the study
Osteoarthritis patients
Rheumatoid arthritis patients
Normal control subjects
Synovial fluid
* There was no significant difference in the synovial fluid and serum
values between the osteoarthritis and rheumatoid arthritis patients.
For insulin-like growth factor (IGF), 7.5 kd was used; for IGF binding
protein (IGFBP), 40 kd was used.
t P = 0.0002 versus each patient group.
correlations were seen in serum or synovial fluid from
OA patients.
Findings of Western ligand blotting. As shown in
Figure 3A, there was a large increase in the level of
Table 2. Levels of IGF and IGFBP-3 in synovial fluid, expressed as
a percentage of the levels in serum*
Osteoarthritis patients
Rheumatoid arthritis patients
Normal control subjects
* IGF = insulin-like growth factor; IGFBP = IGF binding protein.
t P < 1 x 1 0 - ~versus controls.
$ P < 0.01 versus controls.
9 P < 0.01 versus rheumatoid arthritis patients.
120 - -
100 - -
r = 0.898
80 --
g 60 -.
20 --
IGF-1 )igniter
Figure 2. Correlation of serum insulin-like growth factor type 1
(IGF-1) with levels of C-reactive protein (CRP) in patients with
rheumatoid arthritis.
virtually all of the IGFBPs in the synovial fluid samples
from RA and OA patients; the IGFBP-3 level was the
most marked. In some synovial fluids from the patients, additional bands were seen, migrating just above
IGFBP-1 (Figure 3A, lanes 7 and 9). These were not
seen in all patients, and may represent IGFBP-5 or
IGFBP-6 or glycosylated IGFBP-4. In the RA samples,
it is also important to note the striking increase in
IGFBP-4 above that in the normal samples.
Findings of immunoblotting for IGFBP-3 and
IGFBP-2. Figure 3B shows that there was a marked
difference in the ratio of intact to proteolysed IGFBP-3
in the synovial fluids from the normal subjects and the
arthritis patients. The IGFBP-3 fragment moved to the
same position as that caused by the proteolytic activity in
serum taken from a woman in her third trimester of
pregnancy. The reverse situation was observed with
IGFBP-2 (Figure 3C). There, the intact binding protein
predominated in the synovial fluid from the normal
subjects. Both the intact and fragmented forms of these
2 binding proteins were observed in the synovial fluid
samples from the arthritis patients. Thus, the extent of
IGFBP-3 and IGFBP-2 proteolysis was similar in serum
and synovial fluid from the arthritis patients. However,
in normal subjects, there was a considerable difference
between the 2 compartments.
Protease activity in synovial fluids. Figure 4
shows the ability of synovial fluids from both groups
of patients to degrade radiolabeled nonglycosylated
IGFBP-3 (lanes 5-6,9-10, and 13-14). Scanning of the
individual lanes by densitometry showed that the mean
degradation of added IGFBP-3 was 51% (range 2487%) for OA, 34% (range 4-67%) for RA, and 65%
(range 34-91%) for control synovial fluid samples; these
results are shown graphically in Figure 5. There were
significant differences between all the groups, the most
notable being that between the RA and control samples
( P = 0.00017 ). Figure 4 also shows that the IGFBP-3
protease activity in synovial fluid was not related to high
activity in the corresponding paired serum. Indeed, the
serum protease activity appeared to be similar in all
groups. The protease activity in synovial fluid could be
partially inhibited by EDTA (Figure 4, lanes 7, 11, and
15); inhibition ranged from 50% to loo%, as assessed by
densitometry scanning. Although equal volumes of synovial fluid were loaded onto the gel after correcting for
protein content, the same results were obtained (Figure 4).
A marked increase was found in the levels of
growth factors and binding proteins in arthritic knee
joints compared with those in normal knee joints. The
lack of difference between OA and normal serum levels
of IGF-1 confirms the results of previous work by
McAlindon et a1 (32). The increase in synovial fluid
levels of IGFBP-4 is particularly interesting, since this
protein is known to be produced by normal human bone
cells (33), which additionally synthesize an IGF-2dependent protease for IGFBP-4 (34). This suggests an
involvement of the underlying bone in the arthritic
process that increases the amount of intact IGFBP-4 in
the joint space. IGFBP-4 has been shown to be inhibitory to basal and IGF-mediated cartilage growth ( 3 9 ,
and in spontaneous OA of the guinea pig knee, bone loss
has been shown to precede cartilage damage (ref. 36 and
Table 3. Correlation coefficients and t-test P values for CRP and
IGF/IGFBP levels in rheumatoid arthritis patients*
Synovial fluid
Synovial fluid
Synovial fluid
Correlation coefficient
* CRP = C-reactive protein; IGF = insulin-like growth factor; IGFBP
IGF binding protein.
8 9
6 7 8 9 1 0 1 1 1 2 1 3
8 9 1 0 1 1
Figure 3. Immunoblots showing the distribution of A, insulin-like growth factor binding proteins
(IGFBPs) 1-4, B, IGFBP-3, and C, IGFBP-2 in sera and synovial fluids. Lane 1, Pooled normal human
serum used as a control; lanes 2-3 and 4-5, paired serum and synovial fluid, respectively, from 2 normal
control subjects (non-arthritic); lanes 6-7 and 8-9, paired serum and synovial fluid, respectively, from 2
rheumatoid arthritis patients; lanes 10-11 and 12-13, paired serum and synovial fluid, respectively, from
2 osteoarthritis patients.
unpublished results). Owing to the short half-life of free
IGF (16), it can be assumed that the level of IGF is
increased as a direct result of an increase in binding
proteins. This may be the result of either an increase in
the production or a decrease in the clearance of these
proteins. The results of the present study suggest that in
the synovial fluid, there is an up-regulation of the
production of IGFBP-3 and a decrease in the extent of
its proteolysis. It is known that IL-1 and, to a lesser
extent, TNF increase the level of IGFBP-3 production
by articular chondrocytes (37). IL-1 is up-regulated in
arthritis (38), and the level of TNF receptors is also
increased in arthritic cartilage (39). These cytokines may
therefore be responsible for this increase in IGFBP-3.
Intact IGFBP-3
9 1 0 1 1 1 2
Figure 4. Autoradiograph of gels showing protease activity in sera and synovial fluids. Lane 1,
Nonglycosylated insulin-like growth factor binding protein type 3 (IGFBP-3) alone; lanes 2-15, nonglycosylated IGFBP-3 plus the following: pooled normal human serum used as a low/negative control (lane
2); serum from a woman in her thud trimester of pregnancy used as a positive control (lane 3); paired
serum and synovial fluid, respectively, from a normal control subject (non-arthritic; lanes 4-5); synovial
fluid from a normal control subject (lane 6); synovial fluid from the same normal control subject as in lane
6 plus 62.5 mM EDTA (lane 7); paired serum and synovial fluid, respectively from a rheumatoid arthritis
(RA) patient (lanes 8-9); synovial fluid from an RA patient without (lane 10) and with (lane 11)62.5 mM
EDTA, paired serum and synovial fluid, respectively from an osteoarthritis (OA) patient (lanes 12-13);
synovial fluid from an OA patient without (lane 14) and with (lane 15) 62.5 mM EDTA.
The lower level of IGFBP-3 protease activity in
arthritic synovial fluid is remarkable in the context of
current literature showing high levels of metalloprotein-
I-teat OA VI. RA: P = 0.026
I-test RA vs. normal: P = 0.00017
I-tat OA va. norma1:
P = 0.044
RA patients
Normal controls
OA patients
Figure 5. Percentage degradation of radiolabeled nonglycosylated
insulin-like growth factor binding protein type 3 (IGFBP-3). Each point
represents the percentage loss of the intact nonglycosylated IGFBP-3
band relative to the control band (nonglycosylated IGFBP-3 incubated
with buffer alone, as run on each gel), as determined by densitometry
scanning. OA = osteoarthritis; RA = rheumatoid arthritis.
ases in the arthritic joint and indicating that metalloproteinases are able to degrade IGFBP-3 (40). There is an
up-regulation of tissue inhibitor of metalloproteases in
OA cartilage as well as a larger increase in metalloproteinases (23), and this may result in an overall lower
ratio of IGFBP-3 protease to inhibitors in the arthritic
joint space compared with the normal joint. While the
protease responsible for the cleavage of IGFBP-3 is
probably not specific for IGFBP-3 alone, there may be
an inhibitor of this protease that is specific for IGFBP-3
in the arthritic joint, although this has yet to be elucidated. It is not thought that the source of an IGFBP-3
protease inhibitor would be the circulation, since there is
evidence to show that outside the circulation in interstitial fluid, serum is unable to inhibit IGFBP-3 protease
activity (41). Since it is known that IGFBP-3 contains a
potential glycosaminoglycan-binding motif (42), the
abundant proteoglycans in arthritic synovial fluid may
play a role in this inhibition by binding to IGFBP-3.
Fragmentation of IGFBP-2, which occurred in
arthritic synovial fluids, was observed at a very low or
negligible level in the normal control samples. In the
very few normal synovial fluids that showed a trace of
intact IGFBP-3, there was a corresponding trace of
fragmented IGFBP-2. This suggests that there may be a
link between the processing of these 2 IGFBPs. Recent
evidence indicates that IGFBP-3 is able to inhibit the
proteolysis of IGFBP-4 (43), indicating an interaction
between individual IGFBPs and their respective IGFBP
proteases. A similar interaction between IGFBP-3 and
IGFBP-2 and IGFBP proteases may be occurring in the
synovial fluid samples described in this report.
The increase in IGFBPs in the joint causes a
marked change in the amount of IGF in the synovial
fluid compared with the serum. In a healthy adult, the
percentage of growth factor found in interstitial fluid or
lymph (i.e., outside the circulation) is 10-30% of that in
serum (44,45). This was indeed the case in the group of
normal volunteers described here, in whom the synovial
fluid IGF-1 and IGF-2 levels were 15% of the serum
However, the mean synovial fluid IGF-1 level in
RA patients was 60% of the serum level, and in OA
patients, it was 53% ( P < 0.00001 for both groups versus
normal controls). There were also significant increases
above normal in the corresponding values for IGF-2 and
IGFBP-3. This suggests that there is some regulation
occurring at the level of the joint that does not affect the
absolute circulating level of growth factor. In an inflammatory joint disease, the synovial membrane becomes
inflamed and its permeability is increased. This may also
account for the increase in IGF in the synovial fluid. The
significant correlation between IGF-1 in the synovial
fluid and that in the serum in RA patients supports this
idea. However, there was no similar association in OA
patients with OA, which is not a predominantly inflammatory disease, to account for the increased synovial
fluid level of IGF. The increased permeability of the
synovium in RA may account for the increased value
above that in OA, but it obviously does not account for
all of the increase in IGF-1 in the synovial fluid. In
serum, IGFBP-3 is bound to an acid-labile subunit,
yielding a combined molecular weight of 150 kd. This
complex is unable to cross the capillary wall (45), and
even in an inflamed synovium, is unlikely to diffuse
through passively. Indeed, there was no correlation
between serum and synovial fluid levels of IGFBP-3 in
either disease group.
It is interesting to note that in RA patients, there
was a strong correlation of the synovial fluid and serum
levels of IGFBP-3, IGF-1, and IGF-2 with the serum
levels of CRP. This correlation does not, of course,
attest to a causal relationship, but it does indicate that
there is considerable disruption of the IGF system in RA
that is related to the intrinsic disease activity. This
relationship does not hold during a flare or shortly after
such local intervention as a steroid injection (results not
shown). There was no association between the CRP
level and any of the IGF system components in OA
patients. Since OA is a joint-associated, rather than a
systemic, disease, a correlation between IGF and CRP
would not be expected, even if IGF is involved in the
disease process. These results therefore confirm a role of
the IGF system in RA as well as in OA. Although there
were no significant differences in levels of serum IGF-1
in any of the groups in this study, it might be that in RA
patients, the serum IGF-1 level is partly controlled by
the disease status. With a larger sample number, the
difference between the IGF-1 level in normal and RA
sera may achieve significance with a P value below 0.05.
Although there is considerable literature suggesting that there are many proteases present in the arthritic
joint space, this study shows that proteases capable of
degrading IGFBP-3 are also present in the normal joint
space. The IGFBP-3 present in this compartment is
virtually entirely fragmented and as such, is thought to
have a reduced affinity for IGF-1 (46). This and the
molar excess of IGF to IGFBP-3 indicate an anabolic
environment. In the arthritic joint space, there is not
only a molar excess of IGFBP-3, but it is in the form that
is thought to have a higher affinity for IGF-1 (46), both
of which are conditions that could lead to decreased IGF
availability, reducing the ability of IGFs to counteract
the catabolic influence of cytokines. The extensive data
shown here clearly indicate considerable local disruption
of the regulation of IGF at the joint level in arthritis.
Findings similar to this have been presented by Matsumoto et a1 (47) and Schneiderman et a1 (48). Both
groups of investigators found low levels of IGF in
normal synovial ,fluid compared with arthritic synovial
fluids, with an accompanying low level of intact
IGFBP-3. Further work will involve identification of
potential inhibitors of IGFBP-3 proteases in the arthritic
synovial fluid, together with an analysis of the regulation
of local IGFBP protease production.
The authors would like to thank June Morgan, Chemical Pathology Department, Bristol Royal Infirmary, for her
help in performing the assays for CRP levels.
1. Fassbender HG: Inflammatory reactions in arthritis. In, The
Handbook of Immunopharmacology: Immunopharmacology of
Joints and Connective Tissue. Edited by ME Davies, JT Dingle.
London, Academic Press, 1994
2. Tyler J A Insulin-like growth factor-I can decrease degradation
and promote synthesis of proteoglycan in cartilage exposed to
cytokines. Biochem J 260543-548, 1989
3. Hardingham E, Bayliss MT, Rayan V, Noble DP: Effects of growth
factors and cytokines on proteoglycan turnover in articular cartilage. Br J Rheumatol 31 (suppl):l-6, 1992
4. Osborn KD, Trippel SB, Mankin HJ: Growth factor stimulation of
adult articular cartilage. J Orthop Res 7:35-42, 1989
5. Luyten FP, Hascall VC, Nissley SP, Morales TL, Reddi AH:
Insulin-like growth factors maintain steady-state metabolism of
proteoglycans in bovine articular cartilage explants. Arch Biochem
Biophys 267416-425, 1988
6. McQuillan DJ, Handley CJ, Campbell MA, Bolis S, Milway VE,
Herington A C Stimulation of proteoglycan biosynthesis by serum
and insulin-like growth factor-I in cultured bovine articular cartilage. Biochem J 240423-430, 1986
7. Trippel SB, Corvol MT, Dumontier MF, Rappaport R, Hung HH,
Mankin HJ: Effect of somatomedin-C/insulin-like growth factor I
and growth hormone on cultured growth plate and articular
chondrocytes. Pediatr Res 25:76-82, 1989
8. Hascall VC, Handley CI, McQuillan DJ, Hascall GK, Robinson
HC, b w t h e r D A The effect of serum on biosynthesis of proteoglycans by bovine articular cartilage in culture. Arch Biochem
Biophys 224:206-223, 1983
9. Ashton IK, Matheson JA. Changes in response with age of human
articular cartilage to plasma somatomedin activity. Calcif Tissue
Int 2989-94, 1979
10. Denko CW, Boja B, Moskowitz RW: Growth promoting peptides
in osteoarthritis: insulin, insulin-like growth factor-1, growth hormone. J Rheumatol 17:1217-1221, 1990
11. Moskowitz RW, Boja B, Denko CW: the role of growth factors in
degenerative joint disorders. J Rheumatol 18 (suppl 27):147-148,
12. Denko CW, Boja B: Growth factors in asymptomatic osteoarthritis-insulin, insulin-like growth factor-I growth hormone. Inflammopharmacology 2:71-76, 1993
13. Middleton JFS, Tyler J A Upregulation of insulin-like growth
factor I gene expression in the lesions of osteoarthritic human
articular cartilage. Ann Rheum Dis 51:440-447, 1992
14. Shakwijk J, Joosten LAB, van den Berg WB, van de Putte LBA:
Chondrocyte nonresponsiveness to insulin-like growth factor 1 in
experimental arthritis. Arthritis Rheum 32894-900, 1989
15. Dor6 S , Pelletier J-P, DiBattista JA, Tardif G, Brazeau P, MartelPelletier J Human osteoarthritic chondrocytes possess an increased number of insulin-like growth factor 1binding sites but are
unresponsive to its stimulation: possible role of IGF-1 binding
proteins. Arthritis Rheum 37:253-263, 1994
16. Guler HP, Zapf J, Schmid C, Froesch E R Insulin-like growth
factors I and I1 in healthy man: estimation of half lives and
production rates. Acta Endocrinol 21:753-758, 1989
17. Holly JMP, Claffey DCP, m a n - H u g h e s SC, Frost VJ, Yateman
ME: Proteases acting on IGFBPs: their occurrence and physiological significance. Growth Regul 3:88-91, 1993
18. Jones JI, Clemmons D R Insulin-like growth factors and their
binding proteins: biological actions. Endocr Rev 16:3-34, 1995
19. Froger-Gaillard B, Hossenlopp P, Adolphe M, Binoux M: Production of insulin-like growth factors and their binding proteins by
rabbit articular chondrocytes: relationships with cell multiplication. Endocrinology 124:2365-2372, 1989
20. Mohan S : Insulin-like growth factor binding proteins in bone cell
regulation. Growth Regul 3:67-70, 1993
21. Mohan S, Baylink DJ: Characterisation of the IGF regulatory
system in bone. Adv Exp Med Biol343:397-406, 1993
22. Tesch GH, Handley CJ, Cornell HJ, Herington AC: Effects of free
and bound insulin-like growth factors on proteoglycan metabolism
in articular cartilage explants. J Orthop Res 10:14-22, 1992
23. Dean DD, Martel-Pelletier J, Pelletier J-P, Howell DS, Woessner
J F Jr: Evidence for metalloproteinase and metalloproteinase
inhibitor imbalance in human osteoarthritic cartilage. J Clin Invest
84~678-685, 1989
24. Yateman ME, ClaEey DC, Cwyfan-Hughes SC, Frost VJ, Wass
JAH, Holly J M P Cytokines modulate the sensitivity of human
fibroblasts to stimulation with insulin-like growth factor-I by
altering endogenous IGF-binding protein production. J Endocrino1 137:151-159, 1993
25. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS,
Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller
JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of
rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988
26. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
Christy W, Cooke TD, Greenwald R, Hochberg M, Howell D,
Kaplan D, Koopman W, Longley S 111, Mankin H, McShane DJ,
Medsger T Jr, Meenan R, Mikkelsen W, Moskowitz R, Murphy W,
Rothschild B, Segal M, Sokoloff L, Wolfe F: Development of
criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 29:10391049, 1986
27. Bowsher RR, Lee W-H, Apathy JM, O’Brien PJ, Ferguson AL,
Henry DP: Measurement of insulin-like growth factor-I1 in physiological fluids and tissues. I. An improved extraction procedure
and radioimmunoassay for lumen and rat fluids. Endocrinology
128:805-814, 1991
28. Cheetham TD, Taylor A, Holly JMP, Clayton K, Cwyfan-Hughes
SC, Dunger DB: The effects of recombinant human insulin-like
growth factor-I, (IGF-I) administration on the levels of IGF-I,
IGF-I1 and IGF binding proteins in adolescents with insulindependent diabetes mellitus. J Endocrinol 142367-374, 1994
29. Coulson VJ, Wass JAH, Abdulla AF,Cotterill AM, Holly JMP:
Insulin-like growth factor binding proteins (IGFBPs) in acromegaly. Growth Regul 1:119-124, 1991
30. Cwyfan-Hughes SC, Johnson MR, Heinrich G, Holly J M P Could
abnormalities in insulin-like growth factors and their binding
proteins during pregnancy result in gestational diabetes? J Endocrinol 147517-524, 1995
31. Lamson G, Giudice LC, Rosenfeld RG: A simple assay for
proteolysis of IGFBP-3. J Clin Endocrinol Metab 72:1391-1393,
32. McAlindon TE, Teale JD, Dieppe PA Levels of insulin related
growth factor I in osteoarthritis of the knee. Ann Rheum Dis
52:229-231, 1993
33. Hayden JM, Mohan S , Baylink DJ: The insulin-like growth factor
system and the coupling of formation to resorption. Bone 17
34. Kanzaki S , Hilliker S , Baylink DJ, Mohan S : Evidence that human
bone cells in culture produce insulin-like growth factor-binding
protein-4 and -5 proteases. Endocrinology 134:383-392, 1994
35. Schiltz PM, Mohan S, Baylink DJ: Insulin-like growth factor
binding protein-4 inhibits both basal and IGF mediated chick
pelvic cartilage growth in vitro. J Bone Miner Res 8:391-396,1993
36. Meijers MHM, Bunning RAD, Russell RGG, Billingham MEJ:
Evidence for cathepsin B involvement in subchondral bone
changes during early natural osteoarthritis in the guinea pig
(abstract). Br J Rheumatol 33 (suppl 1):90, 1994
37. Olney RC, Wilson DM, Mohtai M, Fielder PJ, Smith R L Interleukin-I and tumor necrosis factor-alpha increase insulin-like
growth factor-binding protein-3 (IGFBP-3) production and
IGFBP-3 protease activity in human articular chondrocytes. J
Endocrinol 146279-286, 1995
38. Westacott CI, Whicher JT, Barnes IC, Thompson D, Swan AJ,
Dieppe P A Synovial fluid concentrations of five different cytokines in rheumatic diseases. Ann Rheum Dis 49:676-681, 1990
39. Westacott CI, Atkins RM, Dieppe PA, Elson CJ:Tumor necrosis
factor-a receptor expression on chondrocytes isolated from human
articular cartilage. J Rheumatol 21:1710-1715, 1994
Fowlkes JL, Suzuki K, Nagase H, Thrailkill KM: Proteolysis of
insulin-like growth factor-binding protein-3 during rat pregnancy:
a role for matrix metalloproteinases. Endocrinology 13528102813, 1994
Xu S, w a n - H u g h e s SC, van der Stappen JWJ, Sansom J, Burton
JL, Donnelly M, Holly JMP: Altered insulin-like growth factor-I1
level and insulin-like growth factor binding protein-3 (IGFBP-3)
protease activity in interstitial fluid taken from the skin lesions of
psoriasis. J Invest Dermatol 106:109-112, 1996
Hodgkinson SC, Napier JR, Spencer GSG, Bass JJ: Glycosaminoglycan binding characteristics of the insulin-like growth factorbinding proteins. J Mol Endocrinol 13:105-112, 1994
Donnelly MJ, Holly JMP: Does IGFBP-3 regulate IGFBP-4
proteolysis in cultured human dermal fibroblasts. J Endocrinol
149:Rl-R7, 1996
Xu S, Cwyfan-Hughes SC, van der Stappen JWJ, Sansom J, Burton
JK, Donnelly M, Holly JMP: Insulin-like growth factors (IGFs)
and IGF-binding proteins in human skin interstitial fluid. J Clin
Endocrinol Metab 80:2940-2945, 1995
Binoux M, Hossenlopp P: Insulin-like growth factor (IGF) and
IGF-binding proteins: comparison of human serum and lymph.
J Clin Endocrinol Metab 67509-514, 1988
Cwyfan-Hughes SC, Wass JAH, Holly JMP: Two site specific
radioimmunoassays which demonstrate the presence of proteolyticaUy modified insulin-like growth factor-binding protein-3 in the
circulation. J Endocrinol 137:321-328, 1993
Matsumoto T, Gargosky SE, Iwasaki K, Rosenfeld RG: Identification and characterisation of insulin-like growth factors (IGFs),
IGF binding proteins (IGFBPs) and IGFBP proteases in human
synovial fluid. J Clm Endocrinol Metab 81:150-155, 1996
Schneiderman R, Rosenberg N, Hiss J, Lee P, Liu F, Hintz RL,
Maroudas A Concentration and size distribution of insulin-like
growth factor-I in human normal and osteoarthritic synovial fluid
and cartilage. Arch Biochern Biophys 32417S188, 1995
Без категории
Размер файла
912 Кб
like, local, factors, disruption, growth, joint, arthritis, system, insulin
Пожаловаться на содержимое документа