Proteoglycan depletion rather than fibrillation determines the effects of salicylate and indomethacin on osteoarthritic cartilage.

548
PROTEOGLYCAN DEPLETION, RATHER THAN
FIBRILLATION, DETERMINES THE EFFECTS O F
SALICYLATE AND INDOMETHACIN ON
OSTEOARTHRITIC CARTILAGE
MARSHALL J . PALMOSKI and KENNETH D. BRANDT
The effects of salicylate and indomethacin on
glycosaminoglycan (GAG) synthesis by atrophic and
osteoarthritic (OA) canine cartilage were examined in
vitro by transecting distal femora at the metaphysis and
incubating the knuckle, with its overlying cap of articular cartilage, in medium containing sodium salicylate or
indomethacin, and 35S04. Atrophic cartilage had an
intact articular surface, but its uronic acid content
averaged 65% of the control level, and GAG synthesis
was decreased to 50% of control values. Both salicylate
and indomethacin decreased net GAG synthesis in the
atrophic cartilage by an additional 10%. OA cartilage
showed surface disruption, a uronic acid content 49% of
the control value, and a 49% increase in net GAG
synthesis. Salicylate and indomethacin profoundly decreased GAG synthesis in the OA cartilage. However,
GAG synthesis and uronic acid content of cartilage
which had been lacerated in vitro immediately prior to
culture (to stimulate fibrillation) were normal and not
affected by either drug. The data emphasize the importance of matrix proteoglycan content in protecting the
_____
From the Rheumatology Division, Indiana University
School of Medicine, Indianapolis.
Supported in part by grants from the National Institute of
Arthritis, Metabolism and Digestive Diseases (AM-20582 and
AM-7448) and awards from the Arthritis Foundation and the Grace
M. Showalter Trust.
Marshall J . Palmoski, PhD: Associate Professor of Medicine and Anatomy (current address: Department of Pharmacology,
Bristol-Myers Co., PO Box 4755, Syracuse, NY 13221); Kenneth D.
Brandt, MD: Professor of Medicine and Head, Rheumatology
Division, Indiana University School of Medicine.
Address reprint requests to Kenneth D. Brandt, MD,
Rheumatology Division, Indiana University School of Medicine, 541
Clinical Drive, Indianapolis, IN 46223.
Submitted for publication August 13, 1984; accepted in
revised form November 9, 1984.
Arthritis and Rheumatism, Vol. 28, No. 5 (May 1985)
chondrocyte from the suppressive effects of salicylate
and indomethacin on GAG metabolism, and suggest
that the lower proteoglycan content of OA cartilage may
be more important than fibrillation in rendering it
vulnerable to the metabolic effects of these drugs.
We have previously reported that aspirin, when
fed to dogs in quantities sufficient to achieve a serum
salicylate concentration of 20-25 mg%, aggravated the
degeneration of articular cartilage in osteoarthritic
(OA) joints, although it had no effect on normal
cartilage from the contralateral knees (1). In vitro
studies have shown that the marked augmentation of
glycosaminoglycan (GAG) synthesis in slices of cartilage from osteoarthritic joints was suppressed by
salicylate to a much greater extent than GAG synthesis in normal cartilage, and that this difference was
associated with greater uptake of the drug by fibrillated cartilage than by normal cartilage (2). Other data
are consistent with the concept that the suppressive
effects of salicylate and also of indomethacin on GAG
synthesis in normal cartilage are inversely related to
the proteoglycan content of the extracellular matrix,
and indicate that the proteoglycan content of cartilage
may affect its permeability to these drugs (3,4).
To determine whether the augmented effect of
salicylate on osteoarthritic cartilage was due to decreased proteoglycan content andlor to disruption of
surface integrity, providing increased access of the
drug to the chondrocytes, we examined the effects of
salicylate on GAG synthesis in normal cartilage which
had been lacerated in vitro to simulate fibrillation, as
well as on osteoarthritic cartilage and atrophic cartilage, which had an intact surface but a reduced proteoglycan content.
549
SALICYLATE AND INDOMETHACIN IN OA
MATERIALS AND METHODS
Experimental groups and procedures. Fourteen adult
mongrel dogs, whose weights ranged from 23-30 kg, were
used. Four animals (the Cru-X group) underwent transection
of the right anterior cruciate ligament, as previously described (1). All Cru-X animals bore weight on the operated
limb by the third day after surgery and were maintained in
pens (8 x 10') for 16 weeks, at which time they were killed
with an overdose of sodium pentothal.
The right hind legs of 4 additional dogs (the IK group)
were immobilized in a cast with approximately 90" of hip and
knee flexion, to induce articular cartilage atrophy (5). These
animals were killed after 6 weeks of confinement in pens, as
described above.
A third group of 4 dogs (the Lac group) was killed
with an overdose of sodium pentothal within 5 days after
procurement. At the time of death, both knees of each
animal were opened aseptically and the distal femurs were
removed at the metaphysis with a bone saw. After the
femurs were cleaned of adhering soft tissue, the articular
cartilage of the central weight-bearing regions of the medial
and lateral condyles of the right knee of each dog in the Lac
group was lacerated with a scalpel at approximately 1-mm
intervals, with care taken not to penetrate the subchondral
bone.
Culture technique. The medial and lateral condyles of
both knees of each dog in the above 3 experimental groups
were separated through the intercondylar groove with a bone
saw, and were placed separately in breakers containing 15 ml
of Ham's F-12 nutrient mixture, pH 7.4, supplemented with
10% newborn calf serum, streptomycin (50 pg/ml), penicillin
(50 units/ml) (all obtained from Gibco, Grand Island, NY),
and ascorbic acid (50 pg/ml). In half of the experiments, the
medial condyles were incubated with gentle rocking for 24
hours at 37°C in 5% COz, 95% air in medium containing
sodium salicylate (10-3M) or indomethacin (4 x 10-6M),
with Na2-35S04(10 pCi/ml) added for the final 8 hours of
incubation, while the lateral condyles were cultured in an
identical fashion, but without the drug added. In the remaining experiments, the conditions were reversed, i.e., the
lateral condyles were incubated in medium which contained
drug, while the medial condyles were cultured in the absence
of drug.
Characterization of medial and lateral condyles. The
remaining 2 dogs were killed, and cartilage from both knees
was obtained and was used to provide comparisons between
medial and lateral femoral condyles with respect to uronic
acid content and GAG biosynthesis in the absence of drug.
Determination of net glycosaminoglycan synthesis.
Following incubation of the tissues as described above, the
medium was decanted and the condyles were rinsed with
0.9% saline. The spent medium and rinse were combined and
dialyzed against 200 volumes of 0.05M sodium acetate, pH
6.8, for 48 hours at 4°C in Spectrapor No. 3 dialysis tubing
(approximate molecular weight cutoff = 3,500 daltons; Spectrum Medical Industries, Los Angeles, CA).
The articular cartilage was then shaved with a scalpel
from the central, habitually-loaded surface of each condyle.
Cartilage from each condyle was pooled separately. Portions
(approximately 15 mg) of the shaved cartilage were weighed
and used for determination of uronic acid content. Representative full-thickness samples of the cartilage from the
central weight-bearing regions of each condyle were obtained with a Craig biopsy needle and were placed in 10%
buffered formalin for histologic study, while the remainder
was divided into 3 approximately equal portions (20-30 mg
wet weight each) which were used to assay net GAG
synthesis.
'The triplicate samples were suspended in 3 ml of
0.1M borate buffer, pH 8.0, containing 0.02M calcium chloride, and were digested under toluene for 24 hours at 55°C
with pronase (200 pg/lO mg cartilage wet weight; Calbiochem-Behring, San Diego, CA). The digests were placed in
Spectrapor No. 3 membrane tubing and dialyzed overnight
at 4°C against running tap water, following which aliquots
(0.1 ml) were added to 10 ml of Ready-Solv HP 60 (Beckman
Instruments, Irvine, CA) and counted in a Beckman liquid
scintillation spectrometer. Results (nondialyzable counts per
minute) were adjusted for differences in wet weight of the
cartilage.
Determination of uronic acid content. Portions (approximately 15 mg) of cartilage were placed in 2 changes of
acetone for 24 hours, dried to constant weight in vacuo at
8O"C, and digested with pronase as above. The GAGS were
isolated from the pronase digests by precipitation with
9-aminoacridine hydrochloride, and were converted to their
sodium salts with Bio-Rad AG-50 (Na') (6). After the resin
was removed by filtration, the uronic acid content of the
filtrate was determined (7).
Histologic examination. Histologic sections (6p) of
full-thickness samples of the articular cartilage were examined microscopically after staining with Safranin 0-fast
green.
RESULTS
Gross observations. None of the knee joints
contained an effusion. At the time of death, articular
cartilage from the control knees of the Cru-X and IK
dogs, and from both knees of t he Lac dogs, was
grossly normal. In all 4 Cru-X dogs, marginal osteophytes were present on medial and lateral femoral
condyles of the unstable knees and the cartilage surface was dull and rough. In contrast, the unstable
knees of IK dogs had no osteophytes and the cartilage
was smooth, although it appeared thinner than in the
control knees.
Histology and histochemistry. In all control samples, the cartilage surface and tidemark were intact,
and Safranin 0 staining was normal. All samples of
Cartilage from the unstable knees of dogs from the
Cru-X group showed a reduction in Safranin 0 staining
in comparison with the contralateral control cartilage.
Horizontal flaking of the cartilage surface was noted in
all samples, but no vertical fibrillation was seen. T h e
cartilage thickness was comparable with that of the
PALMOSKI AND BRANDT
550
Table 1. Comparison of articular cartilage from medial and lateral
femoral condyles of knees of normal dogs
Dog
Source of
cartilage*
Uronic acid content,
% of tissue dry
weight
Net
glycosaminoglycan
synthesist
RMC
RLC
LMC
LLC
RMC
RLC
LMC
LLC
4.1
4.0
4.2
4.1
3.9
3.8
4.0
3.8
12,432 +. 1,021
12,805 2 858
13,301 t 1,231
13,110 +- 1,410
18,567 2 823
17,492 t 1,210
17,890 t 1,390
18,420 t 1,521
1
2
* RMC = right medial condyle; RLC = right lateral condyle; LMC
= left medial condyle; LLC = left lateral condyle.
t
Nondialyzable "S cpm/lO mg wet weight of cartilage. Values
represent mean & SEM of triplicate samples.
contralateral control knees, and the tidemark was
intact in all cases. No brood capsules were observed.
Cartilage from the scored condyles of dogs from
the Lac group showed cuts extending to varying
depths below the surface. None penetrated the tidemark. Safranin 0 staining adjacent to the cuts and
throughout the matrix was normal. In dogs from the
IK group, the thickness of cartilage from the immobilized leg was decreased to about 70% of that of
cartilage from the contralateral knee, and Safranin 0
staining was reduced. The articular surface, however,
was intact in all samples of atrophic cartilage.
Comparison of cartilage from medial and lateral
femoral condyles of normal dogs. Studies utilizing 2
normal dogs indicated that there was no difference
between the medial and lateral femoral condyles with
respect to uronic acid content or net glycosaminoglycan synthesis (Table 1). In all cases, the uronic acid
content was about 4% of the tissue dry weight, while
net glycosaminoglycan synthesis in the medial condyles averaged 103 f 2% of that in the lateral conTable 2. Uronic acid content of osteoarthritic, lacerated, and
atrophic articular cartilage
Uronic acid content
Group
(no. of dogs)*
Cru-X (4)
Lac (4)
IK (4)
Source of cartilage
Control knee
Experimental knee
Control knee
Experimental knee
Control knee
Experimental knee
% of tissue
dry weight
2 SEM
4.1
2.0
4.3
% of
control
0.2
-
f 0.5
49
2
0.1
-
4.5 +. 0.2
96
65
2
4.0 2 0.1
2.6 t 0.2
* Cru-X = anterior cruciate ligament transection (osteoarthritic);
Lac = cartilage lacerated in vitro; IK = immobilized knee (atrophic); control = cartilage from contralateral knee.
dyles. Cartilage from the contralateral (control) knees
of dogs in all 3 experimental groups exhibited normal
uronic acid content (4.1 2 0.1% of the tissue dry
weight) (Table 2).
Uronic acid content and GAG synthesis in the
experimental groups, and drug effects. Cartilage from
dogs in the Cru-X group. The uronic acid content of
osteoarthritic cartilage from the knees of dogs in the
Cru-X group averaged 49% of that of cartilage from the
contralateral knees of the same animals, while glycosaminoglycan synthesiswas increased. In the 4 dogs in
this group, net GAG synthesis in the osteoarthritic
cartilage was 49 & 5% greater than that in the corresponding control knee (P < 0.01). Both salicylate and
indomethacin strikingly decreased glycosaminoglycan
synthesis in OA cartilage (to 9% and 22%, respectively, of that in osteoarthritic samples cultured in the
absence of the drugs) (Tables 3 and 4).
Cartilage from dogs in the Lac group. Uronic
acid content and net GAG synthesis in experimental
knees of the dogs with lacerated cartilage were similar
to that of control knees (Table 2). GAG synthesis was
not affected by either salicylate or indomethacin (Tables 3 and 4).
Cartilage from dogs in the ZK group. The
uronic acid content of the atrophic cartilage in immobilized knees was reduced to 65% of that in the contralateral knees of dogs in the IK group (Table 2). In
contrast to the osteoarthritic samples, in which GAG
synthesis was increased, in the IK cartilage GAG
synthesis was only 50 ? 4% of control values ( P
< 0.01) (Tables 3 and 4). Both salicylate and indomethacin caused a further decrease in GAG synthesis in the
atrophic cartilage, but this decrease was less than that
seen when the osteoarthritic cartilage was incubated
with the drugs (P < 0.01). Thus, in the absence of
salicylate, GAG synthesis in IK cartilage averaged
48% of that in control cartilage, while in the presence
of salicylate it averaged 37% of that in control cartilage
(Table 3). Indomethacin, too, further reduced GAG
synthesis in the atrophic cartilage by an additional 10%
over the reduction seen without the drug (Table 4).
Neither salicylate nor indomethacin affected
net GAG synthesis in control cartilage of the contralatera1 knees of dogs from any of the experimental
groups.
DISCUSSION
The results of this study confirm our previous
finding that the proteoglycan content of the matrix
SALICYLATE AND INDOMETHACIN IN OA
Table 3.
55 1
Effect of salicylate on glycosaminoglycan synthesis by osteoarthritic, lacerated, and atrophic articular cartilage
Net ;glycosaminoglycan synthesist
Dog
3
4
5
6
7
8
Source of
cartilage*
Uronic acid
content, %
of tissue
dry weight
Control knee
Cru-X knee
Control knee
Cru-X knee
Control knee
Lac knee
Control knee
Lac knee
Control knee
IK knee
Control knee
IK knee
4.1
1.5
3.9
1.8
4.2
4.4
4.5
4.7
4. I
2.8
4.0
2.5
No salicylate added
Salicyiate, iO-3M
' 3 cpm/iO mg
wet weight of
cartilage
3JScpmilo mg
% of
control
wet weight of
cartilage
-
16,989 f 1,342
2,209 t 275
11,560 t 748
1,695 2 350
23,576 f 1,032
22,149 f 760
16,783 t 1,213
16,191 t 683
30,047 t 3,083
12,420 2 1,501
29,685 t 2,450
9,836 f 1,021
17,414 & 1,542
26,295 f 2,742
12,630 f 1,421
17,452 f 982
22,560 5 2,404
23,070 ? 962
17,983 5 842
16,366 -C 1,732
28,345 5 2,978
15,196 f 1,543
31,224 5 3,211
13,114 5 942
151
138
102
91
54
42
% of
control
Salicylate
No salicylate
98
8
92
10
105
96
93
99
I06
82
95
75
~~
* Cru-X = anterior cruciate ligament transection (osteoarthritic); Lac = cartilage lacerated in vitro; IK
=
t
=
~~
immobilized knee (atrophic); control
cartilage from contralateral knee.
Values represent mean 2 SEM of triplicate samples.
affects the susceptibility of articular cartilage to the
effects of nonsteroidal antiinflammatory drugs on
GAG biosynthesis (3,4). In addition, our present findings suggest strongly that disruption of the cartilage
surface (fibrillation) in osteoarthritis is less important
than loss of proteoglycans in determining the effects of
salicylate and indomethacin on cartilage GAG metabolism. Thus, GAG synthesis in cartilage which had been
lacerated in vitro to simulate fibrillation, but whose
uronic acid content was normal, was not affected by
either salicylate or indomethacin. In contrast, GAG
synthesis in atrophic cartilage, whose uronic acid
content was diminished even though the articular
surface was intact, was decreased by both drugs.
We recognize that laceration of normal cartilage
with a scalpel only crudely simulates osteoarthritic
fibrillation. However, both conditions cause an increase in the surface area:volume ratio of the carti-
Table 4. Effect of indomethacin on glycosaminoglycan synthesis by osteoarthritic, lacerated, and atrophic articular cartilage
Net glycosaminoglycan synthesist
10
I1
12
13
14
Source of
cartilage*
Uronic acid
content, %
of tissue
dry weight
Control knee
Cru-X knee
Control knee
Cru-X knee
Control knee
Lac knee
Control knee
Lac knee
Control knee
IK knee
Control knee
IK knee
4.3
2.6
4.2
2.0
4. I
4.3
4.3
4.5
3.8
2.1
4.2
2.8
No indomethacin added
Indomethacin, 4 x 10-6M
' 3 cpm/lO mg
'3c p d 1 0 mg
wet weight of
cartilage
% of
control
wet weight of
cartilage
% of
control
8,064 t 951
1 1,802 2 1,432
146
8,231 t 1,021
2,351 t 121
16,253 5 2,340
6,401 t 324
8,519 5 449
8,494 f 782
20,622 2 1,604
20,289 t 1,747
28,788 5 505
10,086 t 1,492
33,230 t 3,141
15,490 2 1,238
29
39
100
98
35
17,458 5 1,234
28,457 f 3,213
8,609 2 897
9,162 f 749
20,285 5 1,317
21,348 t 1,027
27,672 2 1,725
12,452 2 1,432
34,525 2 3,021
20,066 f 1,901
-
162
-
106
-
105
-
45
-
58
* Cru-X = anterior cruciate ligament transection (osteoarthritic); Lac = cartilage lacerated in vitro; IK
from contralateral knee.
t Values represent mean 2 SEM of triplicate samples.
= cartilage
-
47
=
Indomethacin
No indomethacin
I02
20
93
23
99
93
102
95
I04
81
96
71
immobilized knee (atrophic); control
PALMOSKI AND BRANDT
552
Figure 1. Canine knee cartilage, stained with Safranin 0-fast green. A , 16 weeks after anterior cruciate
ligament transection. B, Articular surface lacerated at time of killing. C , Ipsilateral leg irnmmobilized
for 6 weeks. D, Representative control knee.
lage, which may affect ,uptake of drug from the culture
medium. It is particularly notable, therefore, that the
above differences in GAG synthesis occurred despite the
fact that this ratio was increased to a much greater extent
by laceration than by cruciate ligament transection, which
produced only mild surface disruption (Figure I).
The data in this study are consistent with our
previous findings that aspirin, when fed to dogs in
quantities sufficient to achieve a serum salicylate
concentration of approximately 10-3M, aggravated the
degeneration of articular cartilage from habitually
loaded areas of immobilized (8) or unstable joints ( I ) ,
although it had no effect on normal articular cartilage
from analogous sites in the contralateral knees of the
same animal. It should be pointed out that the culture
system employed herein, using whole femoral condyles, more closely resembles the in vivo conditions
than do tissue slices in organ culture, since salicylate
has been reported to suppress GAG synthesis in vitro
in slices of cartilage from the weight-bearing surfaces
of normal canine joints (9).
The in vitro laceration technique used here is
similar to that described by Meachim (lo), in which
multiple superficial slices are made across the articular
surface at close intervals. Notably, the laceration
procedure itself had no effect on the uronic acid
content of the cartilage or on GAG synthesis (Tables
2-4). Mankin and Boyle ( 1 1 ) reported an enhanced
rate of GAG synthesis 1 day after scarification of
rabbit femoral cartilage, with a return to control levels
of synthesis within another 24 hours. Their study,
however, utilized immature rabbits, and GAG synthesis was measured in tissue slices, which may account
for the difference between their results and ours.
In this study, salicylate and indomethacin
caused significantly more marked suppression of GAG
synthesis in OA cartilage than in atrophic cartilage
(Tables 3 and 4). It seems unlikely that this was due
only to the small difference in the uronic acid content
of the tissues (2.0 and 2.6% of the tissue dry weight,
for osteoarthritic and atrophic cartilage, respectively).
The marked effect of these drugs on osteoarthritic
SALICYLATE AND INDOMETHACIN IN OA
cartilage thus suggests that the OA chondrocyte itself
may be more vulnerable to the effects of these agents
than the normal chondrocyte.
ACKNOWLEDGMENTS
We are grateful to James Bean for technical support
and to Roberta Fehrman for secretarial assistance.
REFERENCES
Palmoski MJ, Brandt KD: In vivo effect of aspirin on
canine osteoarthritic cartilage. Arthritis Rheum 26:9941001, 1983
Palmoski MJ, Colyer RA, Brandt KD: Marked suppression by salicylate of the augmented proteoglycan synthesis in osteoarthritic cartilage. Arthritis Rheum 23:83-91,
I980
Palmoski MJ, Brandt KD: Relationship between matrix
proteoglycan content and the effects of salicylate and
indomethacin on articular cartilage. Arthritis Rheum
26528-53 I , 1983
Palmoski MJ, Brandt KD: Effects of salicylate and
indomethacin on glycosaminoglycan and prostaglandin
553
E2 synthesis in intact canine knee cartilage ex vivo.
Arthritis Rheum 27:398-403, 1984
5. Palrnoski MJ, Perricone E, Brandt KD: Development
and reversal of a proteoglycan aggregation defect in
normal canine knee cartilage after immobilization. Arthritis Rheum 22:508-517, 1979
6. Tsiganos CP, Muir H: Studies on protein-polysaccharides from pig laryngeal cartilage: extraction and purification. Biochem J 113:879-884, 1969
7. Bitter T, Muir H: A modified uronic acid carbazole
reaction. Anal Biochem 4:330-334, 1962
8. Palrnoski MJ, Brandt KD: Aspirin aggravates the degeneration of canine joint cartilage caused by immobilization. Arthritis Rheum 25: 1333-1342, 1982
9. Palmoski MJ, Brandt KD: Effects of salicylate on proteoglycan metabolism in normal canine articular cartilage: in vitro. Arthritis Rheum 22:746-754, 1979
10. Meachim G: Effect of scarification on articular cartilage
in the rabbit. J Bone Joint Surg 45B:150-161, 1963
11. Mankin HJ, Boyle CJ: The acute effects of lacerative
injury on DNA and protein synthesis in articular cartilage, Cartilage, Degradation and Repair. Edited by CAC
Bassett. National Research Council, National Academy
of Sciences, 1967, pp 185-199