Insulin-like growth factors maintain steady-state metabolism of proteoglycans in bovine articular cartilage explants

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 267, No. 2, December, pp. 416-425,1988

Insulin-like Growth Factors Maintain Steady-State Metabolism Proteoglycans in Bovine Articular Cartilage Explants

of

FRANK P. LUYTEN,*,’ VINCENT C. HASCALL,? S. PETER NISSLEY,* TERESA I. MORALES,? AND A. HARI REDDI* *Bone Cell Biology Section, tBone Research Branch, National Institute of Dental Research, and ~Metabolism Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892 Received April 28,1988, and in revised form August 1,1988

The influences of insulin-like growth factor I (IGF-I) and insulin-like growth factor II (IGF-II) on biosynthesis and catabolism of proteoglycans (PG) in bovine articular cartilage explants were examined to define their potential use in a chemically defined medium. In both short- (10 days) and long-term (40 days) cultures, 10 to 20 rig/ml IGF-I maintained PG synthesis at the same or higher levels than in a medium containing 20% fetal calf serum (FCS). Catabolic rates were slower in IGF-I medium than in medium with only 0.1% albumin, but somewhat faster than for cultures in medium with 20% FCS. In long-term cultures 20 rig/ml IGF-I maintained a steady-state condition; the amounts of glycosaminoglycan and DNA per hydroxyproline content were constant throughout the culture period. The half-maximal dose response for IGF-I on PG synthesis (4.5 rig/ml) was distinctly different from that for the IGF-I effect on PG catabolism (1.5 rig/ml), indicating that these two components of PG metabolism can be experimentally uncoupled. IGF-II was less potent than IGF-I in the same batches of articular cartilage; 100 rig/ml IGF-II increased PG synthesis and decreased PG catabolism relative to 0.1% albumin alone, but the responses were only about 60% of those for 5 rig/ml IGF-I. These results suggest that the chondrocytes regulate PG synthesis primarily via the type I IGF receptor and that the IGF-II response is through the same receptor. Evidence is also provided indicating that the cartilage explants initially contain about 50 ng IGF-I per gram wet weight; this matrix-bound IGF-I diffuses into the medium during culture. The chondrocytes synthesize little or no IGF-I that is released into the medium under the culture conditions used. o 19~3 Academic press, I~~.

The cellular mechanisms involved in the maintenance of equilibrium between synthesis and degradation of proteoglycans (PG)’ in articular cartilage play a major role in the normal function of cartilage in

vivo, and there is a growing realization of the importance of growth and differentiation factors in the regulation of these processes. Thus, understanding their role in cartilage is an important objective. Of the many factors known to influence cartilage metabolism, pituitary growth hormone and insulin-like growth factors (IGFs) have been investigated. Recent studies have implicated the local production of IGFs in the action of growth hormone on growth plate cartilage (1). While there is much interest in the actions of IGFs on chondrocytes in vitro, and growth plate in vivo (2-5), there has been little information on their role in differentiation and me-

i To whom correspondence should be addressed. On leave from the Department of Rheumatology, University Hospital, Ghent, Belgium. Supported by DHHS NIH Fogarty International Center, Fellowship F05 TWO 3772. *Abbreviations used: PG, proteoglycans; IGFs, insulin-like growth factors; FCS, fetal calf serum; DMEM, Dulhecco’s modified Eagle’s medium; GAG, glycosaminoglycan; Hyp, hydroxyproline; RIA, radioimmunoassay. 0003-9861188 $3.00 Copyright All rights

0 1988 by Academic Press, Inc. of reproduction in any form reserved.

416

INSULIN-LIKE

GROWTH

FACTORS

tabolism of hyaline articular cartilage. Previous work has shown that bovine articular cartilage explants maintain steady-state metabolism of proteoglycans in the presence of fetal calf serum (FCS) for up to 3 weeks (6-8). It has also been reported that IGF-I can be substituted for FCS to maintain elevated levels of biosynthesis of PG and their phenotype in shortterm cultures of bovine articular cartilage (9). The influence of IGF-I and IGF-II on biosynthesis and catabolism of proteoglycans was examined in bovine articular cartilage explants, in both short- and longterm cultures. We show that IGF-I at 20 rig/ml is sufficient to maintain steadystate metabolism of proteoglycans for at least 6 weeks in the absence of fetal calf serum. Furthermore, this growth factor has the ability to balance biosynthesis and catabolism to maintain steady-state concentration of PG in the matrix. MATERIALS

AND

METHODS

IGF preparations. IGF-I was a gift from Amgen; rat IGF-II was purified to homogeneity from BRL 3A conditioned medium (10). IGF-I and IGF-II were radioiodinated using lactoperoxidase or chloramine-T by Hazelton Laboratories. Cartilage organ cultures. Articular cartilage was dissected from the metacarpophalangeal joints of newborn to lo-month-old bovines. For each experiment, tissue was harvested from a single animal. The cartilage was minced, washed three times in serumfree Dulbecco’s modified Eagle’s medium (DMEM, GIBCO) with 4.5 g glucose/liter and antibiotics, and maintained in culture at 37°C under 95% air and 5% COZ with daily changes of medium (6). The medium/ tissue ratio was constant throughout the whole culture period; approximately 1.5 ml per 100 mg of wet tissue. For the first 2-5 days, the tissue was cultured in batch in 50-ml vials; then it was redistributed in 100 to 150 mg wet weight portions in 24-well tissue culture plates (Costar, Cambridge, MA). A maximum of 12 wells per plate was used to minimize temperature fluctuation during the labeling procedures. Biosynthesis experiments. Biosynthesis of sulfated macromolecules was measured as previously reported (6). On the day of labeling, each culture was incubated in identical aliquots of 1.5 ml serum-free DMEM with 30 &i [?S]sulfate (New England Nuclear) for 3 h at 37°C. All cultures, regardless of previous treatment, were labeled in aliquots of the same isotope solution. Following labeling, the cultures were washed twice with ice-cold 10 mM EDTA,

AND

CARTILAGE

PROTEOGLYCANS

417

0.1 M sodium phosphate, pH 6.5, and the tissue was digested with 0.05% (w/v) proteinase K (protease type XXVIII, Sigma, St. Louis, MO) overnight at 60°C in the same buffer (11). This treatment resulted in a complete digestion of the tissue. Aliquots of the digests (500 ~1) were eluted from Sephadex G-25 (PD10) columns to determine radioactivity in the macromolecular fraction. In other aliquots, glycosaminoglycan (GAG) concentrations were measured using the metachromatic dimethylmethylene blue assay (In), and hydroxyproline (Hyp) concentrations were determined after acid hydrolysis in 6 N HCI at 100°C for 16 h by the method of Woessner (13). DNA determination. DNA was determined using bisbenzimidazole (Hoechst 33258) (14). The method is based on the enhancement of fluorescence in high salt buffer when this compound binds to double stranded DNA. The excitation maximum is at about 356 nm, the emission maximum at 458 nm. A 50-~1 proteinase K digest was mixed with 1940 ~1 phosphate saline buffer (0.05 M NaHPO,, 2.0 M NaCI, 2 mM EDTA, pH 7.4) and 10 ~1 of a 20 mg% bisbenzimidazole stock solution in HzO. Fluorescence was measured 1 h after mixing. Pilot studies were done to standardize measurements of DNA for articular cartilage samples. Papain and proteinase K (O.l%, 16 h, 60°C) solubilized the samples well, while dispase (O.l%, 90 min, 37OC) followed by collagenase (0.4%) 16 h, 37°C) treatment left some residual particles. DNA values were highest for the proteinase K digests, similar to a procedure described previously (11). While digestion of cartilage tissue was almost complete within 3 h at 6O”C, the digest still contained intact cells. This gives a low DNA value because bisbenzimidazole penetrates slowly into intact cells. Sonication after 3 h incubation increased the value about lo-fold to a level nearly the same as for an overnight proteinase K digestion, which results in total cell lysis. This latter treatment was chosen for routine assays. There was a linear correlation between cell number and DNA content, yielding a value of 8-9 pg DNA per cell, which corresponds well with the estimated amount of DNA per cell for most mammalian cells (11). The DNA measurements on different cartilage batches were always done with the same DNA standard stock solution frozen in aliquots (100 @g/ml) at -70°C and the same bisbenzimidazole stock solution kept at 4°C. Under these conditions the standard curves were very reproducible over a period of at least 6 months. Catabolic rate of radiolabeled proteoglycans. For catabolism, cultures were maintained in their appropriate medium for 5 days prior to labeling, and then labeled in serum-free medium containing 40 &i/ml of [35S]sulfate for 6 h. Immediately after labeling, the cultures were washed twice followed by daily changes for 2 days to remove any residual unincorporated isotope. Cultures were then maintained in a medium ap-

418

LUYTEN

propriate to the experiment as described under Results, with daily changes. Radiolabel in each daily collected medium fraction was measured. After the last day, tissue samples were digested with proteinase K as described above and the percentage of ?S-labeled molecules remaining in the tissue was calculated on a daily basis (15,16). Ctmditianing of the media. Articular cartilage was prepared from the metacarpophalangeal joints of a newborn and a 2-month-old calf. Fresh tissue from both animals was washed in serum-free DMEM with 0.1% bovine albumin, 25 ml/3 g wet tissue, for 3 h at 37°C. This washout medium sample was collected and centrifuged and the supernatant was stored at -20°C. Fresh medium (DMEM with 0.1% albumin) was added and the tissues were cultured batchwise at 37°C. Conditioned media (25 ml/3 g tissue) were collected after 3,7, and 11 days. After centrifugation (30 min, 15OOg), the supernatants were lyophilized and redissolved in 1 ml of 1 M acetic acid. The samples were kept for 1 h at 4°C and centrifuged for 5 min at 11000 rpm (Beckman, Microfuge 12) and the supernatants were applied on a Sephadex G-75 column (50 X 1.0 cm) in 1 M acetic acid. IGF-I was measured in the column fractions (1 ml) using a radioimmunoassay (RIA) (17). IGF-I antiserum was provided by Dr. L. Underwood and Dr. J. Van Wyk to the National Hormone and Pituitary Program, NIDDK. IGF-II was measured in the same column fractions using a radioreceptor assay with detergent-solubilized type II receptor from rat placental membranes (17,18). In an additional experiment, we investigated whether conditioned medium would support PG biosynthesis. Articular cartilage was cultured in controls with daily refeedings of the medium (DMEM with 0.1% albumin) and PG synthesis was measured daily by adding small aliquots of radiolabel. In parallel, a number of cartilage tissue cultures from the same batch were cultured in the same medium without refeedings and labeled directly by adding small aliquots of radiolabel after up to 9 days culture in the same medium. Cycloheximide expe-riment. In order to examine the autocrine production of IGFs by articular cartilage, cycloheximide was used. Cartilage from the metacarpophalangeal joints of a 4-month-old calf was dissected, minced, washed twice in DMEM with 0.1% albumin, and transferred to 24-well tissue culture plates. Each well contained 2 ml DMEM with 0.1% albumin for about 100 mg wet tissue. The medium was changed on Days 1 and 2; 10 wells containing tissue were then incubated for 2 days in a medium with 100 fig/ml cycloheximide (19) and 10 wells in a separate plate with medium alone. Each medium was applied to a Sephadex G-75 column in 1 M acetic acid. IGF-I was measured in the column fractions using the RIA. IGF-I measurements were done on media from an additional 2-day culture period for both cycloheximide and control cultures. The effect of

ET AL.

0.

, 0

10

I 20

8

I 30

8

I 40

-

5’. 20% xs

ng IGF-l/ml

FIG. 1. Dose-dependent stimulation of proteoglycan synthesis by IGF-I. Duplicate cultures of cartilage explants from two different animals were cultured for 5 days in the presence of the indicated concentrations of IGF-I and then labeled with [?S]sulfate for 3 h. The tissue digests were chromatographed on Sephadex G-25 to determine radioactivity in the macromolecular fraction. The bars show the incorporated radiolabel per Hyp content when cultured in the presence of 20% FCS.

cycloheximide on the cartilage PG synthesis was monitored by incubation of a series of samples of the same batch with medium containing 100 pg/ml cycloheximide for up to 3 h, followed by a [35S]sulfate labeling. The [%S]sulfate incorporation into macromolecules in the presence of different doses of IGF-I was used as a positive control. RESULTS

&Feet of IGF-I on PG Synthesis in ShwtTerm Cultures Explant cultures of articular cartilage from two approximately 6-month-old animals were maintained for ‘7 days of culture in DMEM with 0.1% bovine albumin to bring the cultures to a basal level of PG synthesis. They were then supplemented with 0, 5, 10, 20, and 50 rig/ml IGF-I and after 5 days labeled with [35S]sulfate. Figure 1 shows the 35S incorporation per microgram Hyp for the different experimental conditions. Both animals showed the same qualitative response, increasing synthesis to a plateau by lo-20 rig/ml. These levels in each case were nearly the same as the levels measured in medium with 20% FCS.

INSULIN-LIKE

GROWTH

FACTORS

WEEKS

FIG. 2. Proteoglycan biosynthesis in long-term cultures of cartilage explants in the presence of IGF-I. Articular cartilage explants were cultured for up to 5 weeks in 0,5, or 20 rig/ml IGF-I or 20% FCS. Incorporation of [%]sulfate per microgram Hyp was measured every week in duplicate cultures.

They were quantitatively different, however, in terms of their basal and plateau values, reflecting animal to animal variation. Effect of IGF-I on PG Synthesis in LongTerm Cultures Explant cultures of articular cartilage of an animal about 6 months old were kept for 4 days in DMEM with 0.1% albumin and then maintained for up to 5 weeks in 0,5, or 20 rig/ml IGF-I or 20% FCS. Duplicate cultures were labeled every week and digested with proteinase K and the amount of 35S-labeled macromolecules was determined (Fig. 2). PG synthesis in cultures, maintained in 20 rig/ml, was high by 1 week and increased to a plateau value 5 times the basal value by 5 weeks. For 5 ng/ ml, a plateau value about 3 times basal was reached. Cultures with 20% FCS initially were at the same level as for 20 rig/ml IGFI but in this experiment decreased somewhat after 3 weeks to a level similar to that for 5 rig/ml IGF-I. Eflect of IGF-I

on PG Catabolism

The effect of 5 and 20 rig/ml IGF-I on loss of radiolabeled PG from the tissue was

AND

CARTILAGE

PROTEOGLYCANS

419

measured in long-term culture and compared with 20% FCS and 0.1% albumin controls. Both 5 and 20 rig/ml gave the same PG catabolic rate (& - 18 days), which was slower than that in the 0.1% albumin medium alone (tljz - 9.5 days) but somewhat faster than that for the cultures maintained in 20% FCS (& - 23.5 days) (Fig. 3a). Because there was no difference between 5 and 20 rig/ml IGF-I, a second series of catabolism experiments was done with a different batch of cartilage to investigate the effect of lower doses of IGF-I. The tl12 values for this experiment are shown in Fig. 3b. Again no difference could be found between 5 and 20 rig/ml IGF-I. However, there was a dose response between 0.1 and 5 rig/ml with values increasing from about 6.8 to 14 days. The half-maximum response was estimated to be approximately 1.5 ng/ ml IGF-I (see Fig. 9 below). Effect Of IGFWrmz GAG Content The concentration of GAG per microgram Hyp in the tissue remained constant throughout the entire long-term culture period for media containing 20% FCS and 20 rig/ml IGF-I (Fig. 4). The 5 rig/ml IGF-I medium was unable to maintain a constant GAG/Hyp ratio but the loss of PG was much less than that for the 0.1% albumin medium alone. E#ect of IGF-I

on DNA Content

The results of DNA determinations per Hyp content in the long-term cultures are shown in Fig. 5. The DNA content per microgram Hyp was constant throughout the 5-week culture period for the samples cultured in DMEM containing either 0.1% albumin alone or 5 or 20 rig/ml IGF-I. In this experiment, however, there was an increase in DNA content for the samples cultured with 20% FCS. This result was confirmed by histological sections of some tissue pieces. The samples grown in 20% FCS showed cell proliferation at the margins of the tissue after 6 weeks in culture whereas no outgrowth was seen in the IGF-Itreated samples (Fig. 6). The 6-week IGFI-treated sample was indistinguishable by

420

LUYTEN ET AL. a 100 SO

0

4

s

12

16

20

24

Days in Culture b

20

-

0.1% alb.

QA,gU&nl+

Sna/mll

20%

FCS

IGF-I

FIG. 3. Effect of IGF-I on PG catabolism. Cartilage cultures were labeled and then maintained in media with different doses of IGF-I, 20% FCS, or 0.1% albumin alone, with daily refeedings. (a) The percentage of 35S-labeled molecules remaining in the tissue was calculated on a daily basis and plotted as a semi-log curve. (b) The half-lives (ti,a) of radiolabeled PG in different media were derived from the decay curves of the percentage of radiolabel remaining in the matrix each day. Results are the means of two samples, from lo-day cultures.

these morphological criteria from sections prepared from freshly isolated tissues (data not shown).

Eflect of IGF-II on PG Biosynthesis and Catabolism Explant cultures of articular cartilage of an approximately g-month-old calf were preincubated for 5 days in basal medium and then maintained for 5 days in the presence of 0.5,2,5,20, and 50 rig/ml IGF-I or with 2,5,20, and 100 rig/ml IGF-II. The tissues were then labeled with [35S]sulfate and the rate of incorporation into macromolecules was determined (Fig. 7). IGF-II was much less potent than IGF-I; while 100

rig/ml IGF-II increased PG synthesis significantly, the level was only about 45% of the difference between the basal value and the plateau value for 20 rig/ml IGF-I. IGF-II also increased t112values for PG catabolism relative to 0.1% albumin alone for cultures from this animal. However, much higher concentrations were required than for IGF-I; for example 100 rig/ml IGF-II showed a tIj2 value of about 10 days compared with 5 days for basal medium and 14 days for saturating amounts of IGF-I (data not shown).

IGF Levels in Conditioned Media Media were collected from large batch cultures of articular cartilage of a newborn

INSULIN-LIKE

0

1

2

GROWTH

3

4

FACTORS

5

AND

5

WEEKS

FIG. 4. Glycosaminoglycan (GAG) content in longterm cultures. GAG and hydroxyproline contents were measured in aliquots of the tissue digests. The GAG/Hyp ratio remained constant for cultures grown in 20% FCS and 20 rig/ml IGF-I.

and a 2-month-old calf. After a 3-h preincubation wash, the explants were cultured and the supernatants were collected on Days 3,7, and 11. These media were eluted from Sephadex G-75 to separate binding protein from free IGF-I (17). Radiolabeled recombinant IGF-I eluted as a peak at Kav = 0.45 (data not shown). Most of the immunoreactive IGF-I eluted in this peak for all samples. The amounts of IGF-I were calculated for the washout and the O-3,3-7, and 7-11 day conditioned media for both animals. The data are shown in Table I. The preincubation wash contained about 70 ng IGF-I total for both animals. A decrease to about 55 ng was observed for the first 3 days, with a rapid decline for subsequent times. Since bovine IGF-I is identical to human IGF-I in amino acid sequence (20), the IGF-I levels measured by the RIA using a human IGF-I standard should accurately reflect bovine IGF-I levels. IGF-II was also determined in the same column fractions. In the 3-h preincubation supernatant, levels of about 30 ng were found which is less than 50% of the corresponding IGF-I levels. This was followed by a rapid drop to undetectable levels after 7 and 12 days in culture.

CARTILAGE

421

PROTEOGLYCANS

These data provided indirect evidence that even if the tissue synthesizes IGF-I the amounts are not sufficient to support high levels of PG biosynthesis. For example, the final concentration of IGF-I in the O-3 day supernatant was only 0.5 rig/ml, not sufficient to increase PG synthesis levels as shown above. This was investigated in an additional experiment. Articular cartilage cultured in basal medium with daily refeedings showed PG synthesis to be the highest at Days 1 and 2, followed by a progressive drop to reach a plateau value around Day 7 (Fig. 8). The same phenomenon was found for the cartilage tissue cultured in the same medium without replacement for up to 4 days. After 5 days the levels of PG synthesis dropped somewhat lower than that for the replacement protocol probably because of nutritional deficiencies. The plateau level of synthesis for Day 6 and later in the replacement protocol could be due to autocrine, low level production of IGF-I. The possible autocrine production of IGF-I by cartilage explants was investigated with cycloheximide pretreatment. IGF-I was measured in the conditioned media of cycloheximide-treated explant cultures versus controls as described under Materials and Methods. No decrease in 10 ZO%FCS

1 Oi+----- 1

2

3

‘I

5

WEEKS

FIG. 5. Cell proliferation in long-term cultures. DNA contents per microgram Hyp were measured in the tissue digests using bisbenzimidazole. The DNA/ Hyp was constant throughout the culture period for the samples in 0,5, and 20 rig/ml IGF-I. The cartilage grown in 20% FCS (open triangles) showed an increase.

422

LUYTEN

ET AL.

FIG. 6. Photomicrographs of the explants after 40 days in culture. Tissues were fixed and embt :dded in plastic, and the sections were stained with toluidine blue. Photomicrographs of the explants in 20 rig/ml IGF-I (A) and in 20% FCS (B) are shown. Note cellular outgrowth at the margins (arrov vs) in cultures maintained in serum but not for those maintained in IGF-I.

INSULIN-LIKE

GROWTH

FACTORS

FIG. 7. Comparison of IGF-I and IGF-II on proteoglycan biosynthesis. Explant cultures in duplicate were maintained for 5 days in the presence of the indicated concentrations of IGF-I or IGF-II. The tissues were labeled with [35S]sulfate and the incorporated radiolabel was determined per microgram Hyp. The bar represents the synthesis in 20% FCS.

IGF-I release could be found when the sues were treated with cycloheximide. ter 2 days the controls showed about ng IGF-I release per 100 mg wet tissue

tisAf0.15 per

AND

CARTILAGE

PROTEOGLYCANS

423

FIG. 8. PG biosynthesis in conditioned medium. Cartilage explants cultured in duplicate with daily refeedings were labeled daily by adding small amounts of [?S]sulfate, and the amount of incorporated radiolabel was determined (open squares). Cartilage cultured continuously in the same medium for up to 9 days was also radiolabeled as described above and the rates of synthesis were determined (filled squares).

24 h, while the cycloheximide-treated cultures showed a value of 0.25 ng. After 4 days the levels were 0.08 ng/lOO mg/24 h for both. The incorporation of [35S]sulfate was inhibited by more than 90% by this concentration of cycloheximide after 3 h of pretreatment (data not shown). DISCUSSION

TABLE

I

CONCENTRATION OF IGF-I IN THE CONDITIONED MEDIUM OF CULTURES OF ARTICULAR CARTILAGE Animal

1

Total

[cl

3h O-3 days 3-7 days 7-11 days

65 ng 50 12.5 10

20 rig/g 15 3.8 3.3

Accumulation

137.5 ng

Media

42.1 rig/g

Animal Total 75 ng 60 12.0 8.0 155 ng

2 [c]

25 rig/g 20 4.0 2.6 51.6 rig/g

Note. The total signifies the amount of IGF-I determined in the conditioned medium at designated time periods. The calculated concentration [c] in the tissue is expressed as nanograms per gram wet weight of tissue. The expected initial tissue level of IGF-I, based on the total concentration in the medium, is in the range 42-52 rig/g wet weight in the bovine articular cartilage.

The insulin-like growth factors are well known to have important effects on chondrocytes. They were originally identified as sulfation factors, or somatomedins, on the basis of stimulating [35S]sulfate incorporation into cartilage (21,22). IGF-I and IGFII are able to stimulate clonal growth of chondrocytes; IGF-I is more potent in adult chondrocytes, while IGF-II is more potent in fetal cartilage cells (3,4). Recent data (9) showed evidence that IGF-I stimulates PG synthesis in adult bovine cartilage explants to levels comparable with a medium containing 20% FCS. Insulin stimulated proteoglycan synthesis by cartilage only at pharmacological concentrations (9). Further, it has been reported (23, 24) that bovine articular chondrocytes possess receptors for IGF-I. In the present study, the effect of IGF-I on PG biosynthesis of bovine articular car-

424

LUYTEN 600

14

,500

-400

G: -300

s 2

-200

max.conc.

112

- 100

r

f I’I’I’I’I~I’I’I’I’

0 0

12

3

4

nglml

0 5

6

7

6

9

10

IGF-I

FIG. 9. IGF-I uncouples synthesis and breakdown of proteoglycan. Dose-dependent effect of IGF-I on halflife (ti,a) or degradation of radiolabeled PG compared with synthesis of PG as assessed by [%]sulfate incorporation. The half-maximal response for catabolism is at about 1.5 rig/ml IGF-I whereas for PG synthesis it is higher (4.5 rig/ml).

tilage explants was pronounced; 20 rig/ml IGF-I increased PG synthesis to levels greater than or equal to those obtained with 20% FCS even in long-term cultures (up to 40 days). The increase was about twofold over basal levels for very young animals (~6 months) and up to fivefold for the older animals (-10 months). In addition an effect of IGF-I on the loss of radiolabeled PG from the tissue was demonstrated. IGF-I slows down catabolism in a dose-dependent manner approaching, but not quite achieving, the low catabolic rate observed for 20% FCS medium. Significant decreases in rates of PG catabolism started from 0.1 rig/ml IGF-I and plateau values were observed at about 5 rig/ml. This was in contrast with the effects of IGF-I on biosynthesis which were first apparent only at 2 rig/ml and reached a plateau only at 10 to 20 rig/ml (Fig. 9). These findings, observed in three separate experiments, indicate that the cellular pathways regulating PG synthesis and catabolism can be differentially modulated by the same factor. The absence of cellular proliferation in cultures grown in the presence of IGF-I for at least 40 days, as evaluated by DNA mea-

ET AL.

surements and histology, is noteworthy. When grown in a medium with 20% FCS, the DNA content was doubled by 6 weeks in the experiment shown. This experiment was done with cartilage from a lo-monthold animal where contamination with other cell types of underlying or adjacent tissues is minimal. Thus, in the presence of serum, cellular proliferation can occur in these explant cultures after a lag phase of 1 to 2 weeks, a finding that may be critical in evaluating long-term cultures. In saturating amounts of IGF-I, higher synthetic and slower catabolic rates are achieved, such that a constant amount of PG is maintained in the matrix. Thus, this factor alone is sufficient for the articular chondrocytes to support a steady state for at least 6 weeks, a reasonable time period to study proteoglycan metabolism and its control by resident chondrocytes. In shortterm cultures (5 days) the chemical characteristics of the PGs are not changed by IGF-I treatment (9). We are presently conducting more extensive analysis of the PGs synthesized in long-term (5-6 weeks) cultures maintained only with IGF-I to verify the stability of the PG structure in addition to the metabolic parameters described above. While active on metabolic parameters, IGF-II was less potent than IGF-I, requiring concentrations 50-100 times as high to achieve similar effects. This suggests that IGF-II, similar in this respect to insulin (9), acts in this system predominantly via the type I IGF receptor, as has been reported in other chondrocyte culture systems (25,26). Is there any IGF-I present in the cartilage matrix and is there endogenous production of IGF-I? Based on the results presented in this study, it can be concluded that IGF-I and IGF-II are present at high concentrations in the cartilage of young animals. IGF-I was detectable in fresh extracts of bovine articular cartilage (3-h wash). The total amount of IGF-I in articular cartilage batches of two animals was about 150 ng in 3 g wet weight of tissue (Table I) or a concentration of about 50 rig/g in fresh tissue of young animals which is above saturation levels for the effect of IGF-I on PG metabolic parameters. How-

INSULIN-LIKE

GROWTH

FACTORS

ever, our experiments to determine if the chondrocytes actually synthesize IGF-I indicated that they do so, if at all, at levels too low for maintenance of a steady-state condition. Thus it is likely that most of the IGF-I present in the tissue is derived from external sources. This would not be surprising since IGF-I is a cationic, low-molecular-weight substance and may well be bound and perhaps concentrated in the anionic PG matrix (27). In conclusion, 20 rig/ml IGF-I maintains steady-state metabolism of PG in bovine articular cartilage explants for at least 6 weeks in culture. This defined medium should be useful for investigation of other factors modulating articular cartilage metabolism. The presence of this factor in the tissue poses the question of whether it may be a key endogenous regulator of proteoglycan homeostasis. REFERENCES 1. NILSSON, A., ISGAARD, J., LINDAHL, A., DAHLSTROM, A., SKOTTNER, A.,AND IsAK~SON,O. G. P.(1986)Science233,571-574. 2. VETTER,~., HELBRING,G.,HEIT, W.,PIRSIG, W., STERZIG,K.,AND HEINZE,E.(~~~~) Growth49, 229-245. 3. VETTER, U., ZAPF, J., HEIT, W., HELBRING, G., HEINZE, E., FROESCH, E. R., AND TELLER, W.M.(1986)J. Clin. Invest. 77,1903-1908. 4. SCHOENLE, E., ZAPF, J., HUMBEL, R. E., AND FROESCH, E. R. (1982) Nature (London) 296, 252-253. 5. SCHOENLE, E., ZAPF, J., HAURI, C., STEINER, T., AND FROESCH, E. R. (1985) Acta Eno!ocrinol. 108,167-174. 6. HASCALL, V. C., HANDLEY, C. J., MCQUILLAN, D. J., HASCALL, G. K., ROBINSON, H. C., AND LOWTHER, D. A. (1983) Arch. B&hem. Bip phys. 224,206-223. 7. HASCALL,V.C.,MORALES,T.I.,HASCALL,G.K., HANDLEY, C. J., AND MCQUILLAN, D. J. (1983) J. Rheumatol. 11,45-52. 8. MCQUILLAN,D.J.,HANDLEY,C.J.,ANDROBINSON, H. C. (1986) B&hem. .I 237,741-747.

AND

CARTILAGE

PROTEOGLYCANS

425

9. MCQUILLAN, D. J., HANDLEY, C. J., CAMPBELL, M. A.,BoLIs, S.,MILWAY,V. E., AND HERINGTON, A. C. (1986) B&hem. J. 240,423-430. 10. GREENSTEIN, L. A., GAYNES, L. A., ROMANUS, J. A., LEE, L., RECHLER, M. M., AND NISSLEY, S.P.(1987)inMethodsinEnzymology(Barnes, D.,and Sirbasku,D.A.,Eds.),Vol.146,pp.259269,AcademicPress,San Diego. 11. OEGEMA,T.R.,CARPENTER,B.J.,ANDTHOMPSON, P.c.,J~.(1984)J. O&hop. Res. 4,345-351. 12. FARNDALE, R. W., SAYER, C. S., AND BARRETT, A. J. (1982) Connect. Tissue Res. 9,247-248. 13. WOESSNER, J. F. (1961) Arch. Biochem. Biophys. 93,440-447. 14. LABARCA, A., AND PAIGEN, K. (1980)Anal. Birr them 102,344-352. 15. CAMPBELL,M.A.,HANDLEY,C.J.,HASCALL,V.C., CAMPBELL,R. A., AND LOWTHER,D. A.(1984) Arch. Biochem. Biophys. 234,275-289. HASCALL,V.C. 16. MoRALEs,T.I.,WAHL,L.M.,AND (1984)J. Biol Chem. 11,6720-6729. 17. DAUGHADAY,W.H.,PARKER,K. A.,BOROWSKY, S.,TRIVEDI,B.,AND KAPADIA,M.(~~~~) End@ crinology 110,575-581. 18. GELATO, M. C., KIESS, W.,LEE,L.,MALOZOWSKI, S.,RECHLER, M. M., AND NISSLEY,~. P. (1988) J. C&n. Endocrinol. Metab. 67,669-675. 19. KIMURA,J.H.,CAPUTO,C.B.,ANDHASCALL,V.C. (1981) J. Biol. Chem. 9,4368-4376. Biol. 20. HONEGGER,A.,AND HuMBEL,R.E.(~~~~)J. Chem. 261,569-575. 21. SALMON,W.D.,ANDDALIGHADAY,W.H.(~~~~)J. Lab. Clin. Med. 49,825-836. 22. DAUGHADAY, W. H., HALL, K., RABEN, M. S., SALMON, W.D.,VANDEN BRANDE,J.L.,AND VANWYK,J.J.(~~~~) Nature(London)235,107. WYK,J.J.,FOSTER,M.B.,AND 23. TRIPPEL,S.B.,VAN SVOBODA, M. E. (1983) Endocrinology 112, 2128-2136. 24. WATANABE, N., ROSENFELD, R. G., HINTZ, R.L., DoLLAR,L.A.,ANDSMITH,R.H. (1985) J.Endccrinol. 107,275-283. G.L.,BAx25. READ,L.C.,BALLARD,F.J.,FRANCIS, TER, R. G. B., BAGLEY, C. J., AND WALLACE, J. C. (1986) Biochem. J. 233,215-221. 26. SESSIONS, C. M., EMLER, C. A., AND SCHALCH, D.S.(1987)EncEoctinology5,2108-2116. 27. VANLENT,P.L.,VANDENBERG,W.,SCHALKWIJK, J.,VANDENPUTTE,B.A.,ANDVANDENBERSSELAAR, L. (1987) J. Rheumatol. 4,798-805.