Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages ]287-]29]
March 31, 1992
ESTROGEN AND OSTEOARTHRITIS: A STUDY OF SYNOVIAL ESTRADIOLAND ESTRADIOL RECEPTOR BINDING IN HUMAN OSTEOARTHRITIC KNEES
C.L. TSAI'*, T.K. LIU1 and T.J. CHEN 2
1Department of Orthopaedic Surgery, School of Medicine, National Taiwan University, Taipei, Taiwan, ROC 2Department of Physiology, School of Medicine, National Taiwan University, Taipei, Taiwan, ROC
Received February 8, 1992
Estrogen appears to be a risk factor in knee osteoarthritis (OA). Results from 21 patients revealed that synovial estradiol level was highly related to the severity of OA. Increased estradiol receptor bindings in the medial compartment of the femoral condylar and tibia plateau cartilages were observed; the increase was significantly higher in the medial than in the lateral compartment (p < 0.05). Although the synovial estradiol level was significantly lower in women with OA than in men with OA (p < 0.01), postmenopausal women were hypothesized to be more susceptible to OA, since the possible existence of synovial testosterone might counteract the high estradiol synovial level in men, which results in a lower incidence of knee OA in men than in postmenopausal women. We suggest that excessive synovial estradiol and higher estradiol receptor bindings may be involved in the development of knee OA, particularly in postmenopausal women. ® 1992AcademicPress, Inc.
Osteoarthritis (OA) is a common cartilage disease of old age. In two large population studies, the incidence of knee OA has been observed to be more prevalent in women than in men, mostly over the age of 55 (1,2). The disease is characterized by cartilage breakdown and erosion. As complicated as the etiopathogenesis of OA, OA can be categorized into primary and secondary with the hypothesized etiologies being related to mechanical wear, failure of chondrocyte to maintain a balance of matrix synthesis and degradation, and even bone remodelling and synovium-mediated events (3). Among these, OA in postmenopausal women has been scrutinized and a role of endocrine imbalance relating to menopause has been suggested. In women with OA, increased rates of previous hysterectomy and oophorectomy were noted when
* To whom correspondence should be addressed. Abbreviations : OA,osteoarthritis; E2,estradiol; HTP,hydroxylapatite;: DES,diethylstilbestrol.
1287
0006-291X/92 $1.50 Copyright © 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
compared to the population control (4), suggesting excessive estrogen is responsible for dysfunctional uterine bleeding in these OA women, and possibly predisposes these women to OA. In partially meniscectomized rabbits, systemic estrogen administration results in the increases in the frequency and severity of knee OA (5). Intraarticular injections of estradiol to ovariectomized rabbits further induced an up-regulation of estrogen receptors in condylar cartilage at early OA stage (6,7) and cartilage degeneration and erosion at the late stage (6). These lead to the hypothesis that higher incidence of knee OA in postmenopausal women might be related to the estradiol level in the synovial fluid. To further understand the importance of estrogen in OA, the relationship between the synovial estradiol level and the estradiol receptor concentrations was investigated in osteoarthritic human knees. MATERIALS AND METHODS Six (3 men, 3 women; age range 58-70) control subjects, with no previous history of arthritis, were aspirated to obtain the synovial fluid. Twenty-one patients were selected (10 men, average age 66.8 -t- 4.4 range 57-70; 11 women, average age 67:3 -I4.3 range 60-74). Of these, 2 women and 3 men experienced lateral plateau or supracondylar fractures 10 to 30 years ago. All women experienced naturally occurring menopause prior to the diagnosis. The severity of OA was graded based on the anteropostedor roentgenography and a score was assigned ranging from 0 (normal) to 4 (severe joint space narrowing, cysts, osteophytes, and sclerosis) for both medial and lateral compartments, according to Kellgren grading system (8). The excised condyle and plateau were obtained at the times of bilateral or unilateral total knee replacements in all patients. The cartilage was then rinsed with sterile saline, minced, and pooled to obtain approximately 1 g for each sample of condyle or plateau. Special care was taken to only sample the cartilage from the regions adjacent to the erosions (9) to avoid the higher levels of degradative enzymes. Synovial fluid (1-8 ml) obtained from each knee joint was stored at -20 ~ until radioimmunoassay of the estradiol level. Briefly, the thawed sample (100 ul) was incubated in estradiol antibody-coated tubes with 1 ml of [,2~ I]estradiol (Diagnostic Products Corp., CA, USA) at room temperature for 3 h. The tube was decanted and drained for 2 to 3 min, and further dried and counted for 1 min in a gamma counter. The estradiol concentration in the samples was estimated by interpolation into the standard calibration curve with concentrations ranged from 0 to 3600 pg/ml. Intraassay coefficient of variation was 11.3%. The sensitivity was 8 pg/ml. The cytosolic receptor binding of estradiol was assayed using HTP-column (10), and expressed as fmol/mg cytosol protein. Samples (1:4, w/v) were homogenized in buffer A30 (1 nM EDTA, 50 nM Tris, 12 mM monothioglycerol, 30% (v/v) glycerol, pH 7.5), then centrifuged (107,000 g, 1 h) to yield the cytosol fraction. For determination of total and nonspecific bindings, cytosol (100 ul) and 500x DES (Sigma, MO, USA) were separately incubated in 3 nM [ 3H] estradiol (150 ul)(New England Nuclear Corp., MA, USA) at 4"Ofor 18 h. The mixture was then pipetted into a 0.5-cm HTP-column, washed with 4 ml of TM buffer (10 mM Tris, 12 mM monothioglycerol, pH 7.5), eluted by ethanol, and counted in a scintillation counter. Protein content in the cytosol was also determined (11).
RESULTS In women, 46% (5 in 11 cases) were diagnosed of bilateral OA. However, only 20% (2 in 10 cases) in men were diagnosed of bilateral OA. It appears in our series that the 1288
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
6
Female
Male
Fig. 1. Synovial levels of E2 in OA patients without fractures. A significant difference between women (n=9) and men (n=7) is noted (*% p < 0.01). All data were expressed as mean +SEM, and analyzed with Student's t- test and Duncan's multiple range test.
incidence of bilateral OA in women is twice that in men. In all cases , all medial compartments showed severe cartilage degeneration with a grade of 3 or 4, but lateral compartments with a grade between 0 and 1. Synovial 152 level was low (0-0.19 pg/ml) in the control subjects. In OA without fractures, men revealed higher levels of synovial E2 than women with OA (Fig. 1)(p < 0.01), and in both sexes the levels were significantly higher than the control subjects (p < 0.05). In contrast, OA with fractures revealed similarly low synovial E2 levels as in control. Significantly higher E2 receptor bindings in the medial compartment were positively correlated to the severity of non-traumatic OA in both sexes (Fig. 2). Moreover, the estrogen receptor level in the medial compartment of femoral condyle in men was significantly higher than that in women (p < 0.05). No detectable E~ receptor binding was found in the 5 0 A cases with previous fractures. DISCUSSION With regard to fat distribution, women are more obese in the thighs than men, thus causing a bowlegged walk and mechanical stress on the medial side of the knees (12). Postmenopausal women are reported to have twice as many bilateral OA than age-matched men (t 2), which was confirmed in the present study. Significantly higher E2 receptor bindings in the medial compartment were positively correlated to the severity of non-traumatic OA in both sexes. While the concentration of E2 is assumed to be uniform, a possible explanation for this observation might be attributed to the age-related changes of matrix synthesis. Shorter proteoglycan chains are produced with increasing age (13), thus resulting in a fragile cartilage less capable of withstanding the normal loading forces. As the cartilage is aged, the forces and 1289
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Female
~"
Male
100"
100
8O
80
o~ eL
~
60
_= "~ .=.
40
40
~
2o
20
~
*
*
60
o
el.
Mc
Lc
Mt
Lt
Me
Lc
Mt
Lt
Fig. 2. Specific E~ bindings in cartilages of the medial condyle (Mc), lateral condyle (Lc), medial tibia plateau (Mt), and lateral tibia plateau (Lt) in OA patients without fractures. A significant difference in estradiol receptor bindings is noted between medial and lateral compartments of condyle and plateau in both sexes(*, p < 0.05). All data were expressed as mean +SEM, and analyzed with Student's t- test and Duncan's multiple range test.
stress exerted by the mechanical axis further subject the medial compartment to cartilage lesions and degeneration. A looser matrix might be synthesized in the medial compartment, allowing large molecules in the synovial fluid, such as estrogens, to be uptaken by the chondrocytes,
and therefore resulting in cell death and cartilage
erosion. On the other hand, the estradiol receptor bindings in OA with fractures were extremely low (p < 0.05) , compared to OA without fractures. Thus, low estradiol receptor bindings and the absence of synovial estradiol in trauma-associated OA cases strongly
suggest
that
different
mechanisms
exist for trauma-associated
and
non-trauma-associated (estrogen-associated) OA. While the findings of high synovial E2 level and high estrogen receptor bindings in cartilage from the male patients might suggest a higher risk to OA, the effects of synovial estradiol in men might be also counteracted by endogenous testosterone, thereby resulting in a lower estrogen/androgen ratio or a lower unopposed,
free
estradiol to interact with cartilage. As suggested, estrogen administration to the C57BL mice susceptible to spontaneous OA, attenuates the development of OA in male mice, perhaps partly by counteracting the endogenous testosterone level and partially by its direct action on the cartilage (14). As a result, the net effect of higher synovial estradiol found in men with OA would have been attenuated, as suggested in the animal model (14). On the other hand, progesterone production declines sharply prior to menopause, and therefore the estrogen/progesterone
ratio or the unopposed estrogen in
perimenopausal women would be elevated and involved in the development of OA (4). High relative or unopposed estradiol in postmenopausal women might subject them rather than men to the chondrodestructive action of estradiol. 1290
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The present work demonstrated that the increase in the estrogen receptor bindings in human OA cartilages and higher synovial estradiol level in OA patients were possibly related to the development of knee OA. The increases in the local estradiol level and its receptor in the cartilage might further increase the risks to OA. This risk appears to be independent of sex; however, men would be less susceptible to knee OA than women, if the increases in synovial estradiol level could be counteracted by the possible existence of synovial testosterone. The onset of OA, particularly in women, might be induced pre-menopausally or peri-menopausally by an excess of estrogen, which might result in an up-regulation of receptors in cartilages, impairment of metabolism, and the subsequent osteoarthritic changes.
ACKNOWLEDGMENTS Special thanks are dedicated to our staff members for providing the specimens. The authors are indebted to Yu-Mai Chen and Tina Y. Wong for their excellent assistance. This work is supported by a grant from National Science Council, Republic of China(NSC80-0412-B002-58).
REFERENCES 1. Kellgren, J:H., and Lawrence, J.S. (1958) Ann. Rheum. Dis. 17, 388-397. 2. Maurer, K., and Hyattsville, M.D. (1979) DHEW publication no. (PHS)79-16661. 3. Howell, D.S. (1986) Am. J. Med. 80 (suppl 4B), 24-28. 4. Spector, T.D., and Silman, A.J. (1988) Br. Med. J. 297, 899-890. 5. Rosner, I.A., Goldberg, V.M., Getzy, L., and Moskowitz, R.W.(1979)Arthritis Rheum. 22, 52-58. 6. Tsai, C.L., and Liu, T.K. (1992) Clin. Orthop. (in press). 7. Tsai, C.L., and Liu, T.K. (1992) (in review). 8. Kellgren, J.H. (1957) Ann. Rheum. Dis. 16, 494-501. 9. MarteI-Pelletier, J., Pelletier, J.P, Cloutier, J.M., Howell, D.S., Ghandur-Mnaymneh, L., and Woessner, J.F.Jr. (1984) Arthritis Rheum. 27, 305-312. 10. Chen, T.J., and Jow, G.M. (1987) Chinese J. Physiol. 30, 55-67. 11. Sedmak, J.J., and Grossberg, S.E. (1977)Anal. Biochem. 10, 139'i45. Freeman and Company, New York. 12. Radin, E.L., Paul, I.L., and Rose, R.M. (1972) Lancet 1,519-522. 13. Brandt, K.D., and Fife, R.S. (1986) Clin. Rheum. Dis. 12,117-130. 14. Silberberg, M., and Silberberg, R. (1963) J. Endocrinol. 72, 449-451.
1291
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages ]287-]29]
March 31, 1992
ESTROGEN AND OSTEOARTHRITIS: A STUDY OF SYNOVIAL ESTRADIOLAND ESTRADIOL RECEPTOR BINDING IN HUMAN OSTEOARTHRITIC KNEES
C.L. TSAI'*, T.K. LIU1 and T.J. CHEN 2
1Department of Orthopaedic Surgery, School of Medicine, National Taiwan University, Taipei, Taiwan, ROC 2Department of Physiology, School of Medicine, National Taiwan University, Taipei, Taiwan, ROC
Received February 8, 1992
Estrogen appears to be a risk factor in knee osteoarthritis (OA). Results from 21 patients revealed that synovial estradiol level was highly related to the severity of OA. Increased estradiol receptor bindings in the medial compartment of the femoral condylar and tibia plateau cartilages were observed; the increase was significantly higher in the medial than in the lateral compartment (p < 0.05). Although the synovial estradiol level was significantly lower in women with OA than in men with OA (p < 0.01), postmenopausal women were hypothesized to be more susceptible to OA, since the possible existence of synovial testosterone might counteract the high estradiol synovial level in men, which results in a lower incidence of knee OA in men than in postmenopausal women. We suggest that excessive synovial estradiol and higher estradiol receptor bindings may be involved in the development of knee OA, particularly in postmenopausal women. ® 1992AcademicPress, Inc.
Osteoarthritis (OA) is a common cartilage disease of old age. In two large population studies, the incidence of knee OA has been observed to be more prevalent in women than in men, mostly over the age of 55 (1,2). The disease is characterized by cartilage breakdown and erosion. As complicated as the etiopathogenesis of OA, OA can be categorized into primary and secondary with the hypothesized etiologies being related to mechanical wear, failure of chondrocyte to maintain a balance of matrix synthesis and degradation, and even bone remodelling and synovium-mediated events (3). Among these, OA in postmenopausal women has been scrutinized and a role of endocrine imbalance relating to menopause has been suggested. In women with OA, increased rates of previous hysterectomy and oophorectomy were noted when
* To whom correspondence should be addressed. Abbreviations : OA,osteoarthritis; E2,estradiol; HTP,hydroxylapatite;: DES,diethylstilbestrol.
1287
0006-291X/92 $1.50 Copyright © 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
compared to the population control (4), suggesting excessive estrogen is responsible for dysfunctional uterine bleeding in these OA women, and possibly predisposes these women to OA. In partially meniscectomized rabbits, systemic estrogen administration results in the increases in the frequency and severity of knee OA (5). Intraarticular injections of estradiol to ovariectomized rabbits further induced an up-regulation of estrogen receptors in condylar cartilage at early OA stage (6,7) and cartilage degeneration and erosion at the late stage (6). These lead to the hypothesis that higher incidence of knee OA in postmenopausal women might be related to the estradiol level in the synovial fluid. To further understand the importance of estrogen in OA, the relationship between the synovial estradiol level and the estradiol receptor concentrations was investigated in osteoarthritic human knees. MATERIALS AND METHODS Six (3 men, 3 women; age range 58-70) control subjects, with no previous history of arthritis, were aspirated to obtain the synovial fluid. Twenty-one patients were selected (10 men, average age 66.8 -t- 4.4 range 57-70; 11 women, average age 67:3 -I4.3 range 60-74). Of these, 2 women and 3 men experienced lateral plateau or supracondylar fractures 10 to 30 years ago. All women experienced naturally occurring menopause prior to the diagnosis. The severity of OA was graded based on the anteropostedor roentgenography and a score was assigned ranging from 0 (normal) to 4 (severe joint space narrowing, cysts, osteophytes, and sclerosis) for both medial and lateral compartments, according to Kellgren grading system (8). The excised condyle and plateau were obtained at the times of bilateral or unilateral total knee replacements in all patients. The cartilage was then rinsed with sterile saline, minced, and pooled to obtain approximately 1 g for each sample of condyle or plateau. Special care was taken to only sample the cartilage from the regions adjacent to the erosions (9) to avoid the higher levels of degradative enzymes. Synovial fluid (1-8 ml) obtained from each knee joint was stored at -20 ~ until radioimmunoassay of the estradiol level. Briefly, the thawed sample (100 ul) was incubated in estradiol antibody-coated tubes with 1 ml of [,2~ I]estradiol (Diagnostic Products Corp., CA, USA) at room temperature for 3 h. The tube was decanted and drained for 2 to 3 min, and further dried and counted for 1 min in a gamma counter. The estradiol concentration in the samples was estimated by interpolation into the standard calibration curve with concentrations ranged from 0 to 3600 pg/ml. Intraassay coefficient of variation was 11.3%. The sensitivity was 8 pg/ml. The cytosolic receptor binding of estradiol was assayed using HTP-column (10), and expressed as fmol/mg cytosol protein. Samples (1:4, w/v) were homogenized in buffer A30 (1 nM EDTA, 50 nM Tris, 12 mM monothioglycerol, 30% (v/v) glycerol, pH 7.5), then centrifuged (107,000 g, 1 h) to yield the cytosol fraction. For determination of total and nonspecific bindings, cytosol (100 ul) and 500x DES (Sigma, MO, USA) were separately incubated in 3 nM [ 3H] estradiol (150 ul)(New England Nuclear Corp., MA, USA) at 4"Ofor 18 h. The mixture was then pipetted into a 0.5-cm HTP-column, washed with 4 ml of TM buffer (10 mM Tris, 12 mM monothioglycerol, pH 7.5), eluted by ethanol, and counted in a scintillation counter. Protein content in the cytosol was also determined (11).
RESULTS In women, 46% (5 in 11 cases) were diagnosed of bilateral OA. However, only 20% (2 in 10 cases) in men were diagnosed of bilateral OA. It appears in our series that the 1288
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
6
Female
Male
Fig. 1. Synovial levels of E2 in OA patients without fractures. A significant difference between women (n=9) and men (n=7) is noted (*% p < 0.01). All data were expressed as mean +SEM, and analyzed with Student's t- test and Duncan's multiple range test.
incidence of bilateral OA in women is twice that in men. In all cases , all medial compartments showed severe cartilage degeneration with a grade of 3 or 4, but lateral compartments with a grade between 0 and 1. Synovial 152 level was low (0-0.19 pg/ml) in the control subjects. In OA without fractures, men revealed higher levels of synovial E2 than women with OA (Fig. 1)(p < 0.01), and in both sexes the levels were significantly higher than the control subjects (p < 0.05). In contrast, OA with fractures revealed similarly low synovial E2 levels as in control. Significantly higher E2 receptor bindings in the medial compartment were positively correlated to the severity of non-traumatic OA in both sexes (Fig. 2). Moreover, the estrogen receptor level in the medial compartment of femoral condyle in men was significantly higher than that in women (p < 0.05). No detectable E~ receptor binding was found in the 5 0 A cases with previous fractures. DISCUSSION With regard to fat distribution, women are more obese in the thighs than men, thus causing a bowlegged walk and mechanical stress on the medial side of the knees (12). Postmenopausal women are reported to have twice as many bilateral OA than age-matched men (t 2), which was confirmed in the present study. Significantly higher E2 receptor bindings in the medial compartment were positively correlated to the severity of non-traumatic OA in both sexes. While the concentration of E2 is assumed to be uniform, a possible explanation for this observation might be attributed to the age-related changes of matrix synthesis. Shorter proteoglycan chains are produced with increasing age (13), thus resulting in a fragile cartilage less capable of withstanding the normal loading forces. As the cartilage is aged, the forces and 1289
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Female
~"
Male
100"
100
8O
80
o~ eL
~
60
_= "~ .=.
40
40
~
2o
20
~
*
*
60
o
el.
Mc
Lc
Mt
Lt
Me
Lc
Mt
Lt
Fig. 2. Specific E~ bindings in cartilages of the medial condyle (Mc), lateral condyle (Lc), medial tibia plateau (Mt), and lateral tibia plateau (Lt) in OA patients without fractures. A significant difference in estradiol receptor bindings is noted between medial and lateral compartments of condyle and plateau in both sexes(*, p < 0.05). All data were expressed as mean +SEM, and analyzed with Student's t- test and Duncan's multiple range test.
stress exerted by the mechanical axis further subject the medial compartment to cartilage lesions and degeneration. A looser matrix might be synthesized in the medial compartment, allowing large molecules in the synovial fluid, such as estrogens, to be uptaken by the chondrocytes,
and therefore resulting in cell death and cartilage
erosion. On the other hand, the estradiol receptor bindings in OA with fractures were extremely low (p < 0.05) , compared to OA without fractures. Thus, low estradiol receptor bindings and the absence of synovial estradiol in trauma-associated OA cases strongly
suggest
that
different
mechanisms
exist for trauma-associated
and
non-trauma-associated (estrogen-associated) OA. While the findings of high synovial E2 level and high estrogen receptor bindings in cartilage from the male patients might suggest a higher risk to OA, the effects of synovial estradiol in men might be also counteracted by endogenous testosterone, thereby resulting in a lower estrogen/androgen ratio or a lower unopposed,
free
estradiol to interact with cartilage. As suggested, estrogen administration to the C57BL mice susceptible to spontaneous OA, attenuates the development of OA in male mice, perhaps partly by counteracting the endogenous testosterone level and partially by its direct action on the cartilage (14). As a result, the net effect of higher synovial estradiol found in men with OA would have been attenuated, as suggested in the animal model (14). On the other hand, progesterone production declines sharply prior to menopause, and therefore the estrogen/progesterone
ratio or the unopposed estrogen in
perimenopausal women would be elevated and involved in the development of OA (4). High relative or unopposed estradiol in postmenopausal women might subject them rather than men to the chondrodestructive action of estradiol. 1290
Vol. 183, No. 3, 1992
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The present work demonstrated that the increase in the estrogen receptor bindings in human OA cartilages and higher synovial estradiol level in OA patients were possibly related to the development of knee OA. The increases in the local estradiol level and its receptor in the cartilage might further increase the risks to OA. This risk appears to be independent of sex; however, men would be less susceptible to knee OA than women, if the increases in synovial estradiol level could be counteracted by the possible existence of synovial testosterone. The onset of OA, particularly in women, might be induced pre-menopausally or peri-menopausally by an excess of estrogen, which might result in an up-regulation of receptors in cartilages, impairment of metabolism, and the subsequent osteoarthritic changes.
ACKNOWLEDGMENTS Special thanks are dedicated to our staff members for providing the specimens. The authors are indebted to Yu-Mai Chen and Tina Y. Wong for their excellent assistance. This work is supported by a grant from National Science Council, Republic of China(NSC80-0412-B002-58).
REFERENCES 1. Kellgren, J:H., and Lawrence, J.S. (1958) Ann. Rheum. Dis. 17, 388-397. 2. Maurer, K., and Hyattsville, M.D. (1979) DHEW publication no. (PHS)79-16661. 3. Howell, D.S. (1986) Am. J. Med. 80 (suppl 4B), 24-28. 4. Spector, T.D., and Silman, A.J. (1988) Br. Med. J. 297, 899-890. 5. Rosner, I.A., Goldberg, V.M., Getzy, L., and Moskowitz, R.W.(1979)Arthritis Rheum. 22, 52-58. 6. Tsai, C.L., and Liu, T.K. (1992) Clin. Orthop. (in press). 7. Tsai, C.L., and Liu, T.K. (1992) (in review). 8. Kellgren, J.H. (1957) Ann. Rheum. Dis. 16, 494-501. 9. MarteI-Pelletier, J., Pelletier, J.P, Cloutier, J.M., Howell, D.S., Ghandur-Mnaymneh, L., and Woessner, J.F.Jr. (1984) Arthritis Rheum. 27, 305-312. 10. Chen, T.J., and Jow, G.M. (1987) Chinese J. Physiol. 30, 55-67. 11. Sedmak, J.J., and Grossberg, S.E. (1977)Anal. Biochem. 10, 139'i45. Freeman and Company, New York. 12. Radin, E.L., Paul, I.L., and Rose, R.M. (1972) Lancet 1,519-522. 13. Brandt, K.D., and Fife, R.S. (1986) Clin. Rheum. Dis. 12,117-130. 14. Silberberg, M., and Silberberg, R. (1963) J. Endocrinol. 72, 449-451.
1291
