Calcif Tissue Int (2008) 83:354364 DOI 10.

1007/s00223-008-9183-9

Estrogen and Testosterone Attenuate Extracellular Matrix Loss in Collagen-Induced Arthritis in Rats
Kalaivani Ganesan Mitali Tiwari Chidambaram Balachandran Bhakthavatsalam Murali Manohar Rengarajulu Puvanakrishnan

Received: 12 June 2008 / Accepted: 11 September 2008 / Published online: 18 October 2008 Springer Science+Business Media, LLC 2008

Abstract Rheumatoid arthritis (RA) is a sexually dimorphic, autoimmune inammatory disorder affecting the joints. Joint disability in RA results primarily from loss of matrix components (collagen and glycosoaminoglycan) in the cartilage and synovium. This study was carried out to understand the effect of physiological levels of testosterone, estrogen, and progesterone on oxidative stress-induced changes in matrix composition in rat synovium in arthritis. Arthritis induction in castrated and ovariectomized rats resulted in enhanced oxidative stress and this was assessed by lipid peroxidation levels and depletion of antioxidants. This, in turn, led to signicantly (p \ 0.01) increased levels of TNF-a and matrix metalloproteinase-2 (MMP-2), subsequently resulting in loss of collagen, elastin, and glycosoaminoglycan (GAG) and disorganization of reticulin as evidenced by biochemical quantitation and also by staining for collagen, reticulin, and elastin. Treatment with physiological doses of dihydrotestosterone (25 mg topically) and estrogen (5 lg/0.1 ml subcutaneously) restored the antioxidant levels signicantly (p \ 0.05) and reduced the levels of TNF-a and MMP-2, with estrogen exhibiting a higher potency. This, in turn, attenuated the damage to reticulin organization as well as the loss of collagen and

GAG in the articular tissues. However, elastin loss could not be attenuated by either treatment. Progesterone (2 mg/ 0.1 ml subcutaneously) was not shown to have any signicance in disease modication, and on the contrary, it inhibited the protective effects of estrogen. However, progesterone contributed to increased collagen levels in the tissues. Keywords Collagen Elastin Reticulin Glycosoaminoglycan Arthritis

K. Ganesan M. Tiwari R. Puvanakrishnan (&) Department of Biotechnology, Central Leather Research Institute, Adyar, Chennai 600 020, India e-mail: puvanakrishnan@yahoo.com C. Balachandran Department of Pathology, Madras Veterinary College, Vepery, Chennai, India B. M. Manohar Centre for Animal Health Studies, Tamil Nadu Veterinary and Animal Sciences University, Madhavaram, Chennai, India

Rheumatoid arthritis (RA) is a chronic, inammatory, autoimmune disease of the connective tissue leading to progressive joint destruction. The brocartilaginous tissues from synovial joints have a signicant amount of extracellular matrix (ECM) consisting of collagen, elastin, glycosoaminoglycan (GAG), and proteoglycan (PG). Matrix metalloproteinases (MMPs) are a family of zinc-dependent endopeptidases responsible for the proteolytic degradation of ECM components, and they are secreted as inactive latent forms in normal tissue [1]. The process of proenzyme activation is a critical event in the regulation of MMP activity and it plays an important role in MMP-mediated ECM degradation in the RA synovium [2]. Several studies have reported that MMP-2 is constitutively secreted by synovial broblasts [3] and tumor necrosis factor (TNF)-a induced during arthritic inammation is shown to induce MMP-2 activation in the synovium [4]. Hence, it can be speculated that downregulation of inammation in RA could attenuate the matrix loss in the tissue. Sex steroids have a signicant impact on the development of autoimmune diseases in humans and rodents [5]. A barrage of clinical [6] and experimental [7, 8] studies have reported the role of sex hormones on inammation in

123

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

355

arthritis. Our earlier studies have shown that testosterone and estrogen possess anti-inammatory effects [9] in collageninduced arthritis (CIA) in rats. In this study, we have sought to understand the inuence of sex steroids on ECM in CIA in rats by administering testosterone (in the form of dihydrotestosterone; DHT) to castrated arthritic rats and estrogen and progesterone (independently or in combination) to ovariectomized arthritic rats. To evaluate the effects of each of the sex hormones on ECM remodeling, collagen and GAG levels were estimated quantitatively, and in addition, changes in collagen reticulin and elastin were studied by histology. Also, the inammatory surge in arthritis is analyzed in terms of oxidative stress as well as changes in TNF-a and MMP-2 levels, and these are correlated with changes in matrix biochemistry under the inuence of the sex steroids.

retracted laterally toward one side and the ovary was exposed through a thin muscle mass just below the dorsal muscle mass. Each incision was of minimum length to allow the extrusion of ovary. Ligation of the upper horn of the uterus, including the arteries, with chromic catgut was carried out. The ovary was excised and the wounds were closed. A high degree of aseptic procedure was maintained throughout the operation. Hormone Treatment Animals were rested for 3 weeks after castration and ovariectomy. Hormones were administered 5 days prior to induction of arthritis. Hormone administration was continued for 4 weeks from the day of induction at the concentrations reported below. The physiological dose of hormone administered was standardized as reported in our earlier study [10]. DHT Treatment DHT gel (2.5%; 25 mg) was applied topically to castrated rats. From the prescription of the gel, three different dermal concentrations (12.5, 25, and 50 mg) were tried, to determine the optimal concentration that led to the attainment of physiological levels of testosterone in the circulation. After application of DHT, blood was collected from the orbital plexus and testosterone levels in serum were assayed by competitive immunoassay. Application of 25 mg led to the attainment of physiological levels of circulating testosterone (120160 ng/dl), and hence that dose was considered for application. Estrogen Based on earlier reports, three different concentrations of estradiol benzoate (2.5, 5, and 10 lg) were administered subcutaneously to determine the optimal concentration for attaining physiological levels of estradiol. After administration, serum estradiol levels were assayed by RIA. Administration of 5 lg/0.1 ml olive oil led to the attainment of physiological levels (4565 pg/ml) of estradiol, and hence that dose was used in the study. Progesterone Progesterone was administered subcutaneously to ovariectomized rats at three different concentrations (1, 2, and 4 mg). After administration, serum progesterone levels were assayed by RIA. Administration of 2 mg/0.1 ml olive oil corresponded to the physiological progesterone levels in normal rats (4560 ng/ml). Hence, this dose was used in the study.

Methods Animals Random-bred male and female Wistar rats (80100 g) were purchased from Tamil Nadu Veterinary and Animal Sciences University, Chennai. The rats were housed in solidbottomed polypropylene cages containing sterilized husk as bedding material and the husk was changed on alternate days. The animals were under strict veterinary supervision and maintained in control rooms with a 12-h light/dark cycle. They were fed commercial rat diet (Poultry Research Station, Chennai, India) and water ad libitum. This study conformed to the guiding principles of Institutional Animal Ethical Committee and Committee for the purpose of control and supervision of experiments on animals and the guide for the care and use of laboratory animals published by the National Institutes of Health (NIH publication No. 85-23). Castration Castration of male rats was performed according to standard surgical procedure under ketamine anesthesia. A single incision was made over the scrotal skin and the testicles were squeezed out with intact tunica vaginalis by gentle pressure between the thumb and the index nger. The transxation ligation was made using 3.0 chromic catgut around the spermatic cord. Testicles were then severed and scrotal incisions were treated as open wounds. Ovariectomy A single longitudinal skin incision was made ventrally at the level of the lower poles of the kidney after shaving the furs and cleansing with 70% alcohol. The skin was

123

356

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

Estradiol Benzoate Plus Progesterone Estradiol benzoate (5 lg/0.1 ml olive oil), followed by progesterone (2 mg/0.1 ml olive oil), was subcutaneously administered to one group of ovariectomized rats, with a 1h interval between the two administrations. Control Control rats were administered 0.1 ml olive oil subcutaneously. Hormonal levels in serum after their exogenous administration were measured by enzyme immunoassays. Induction and Assessment of CIA Arthritis was induced by injecting a 0.1-ml emulsion of type II collagen (1.5 mg/ml) in 0.1 M acetic acid with an equal volume of heat-killed Mycobacterium tuberculosis (4.5 mg/ml) in incomplete Freunds adjuvant into the right hind paw of rats [11]. The control rats received 0.1 ml incomplete Freunds adjuvant. Rats were examined every day for arthritis development by visual observation and by paw volume measurements using a plethysmometer. Experimental Groups Induced rats were assessed for disease severity regularly for 4 weeks. The experimental groups comprised six rats (n = 6) per group, and they were as follows. Group I: Male normal (MN) Group II: Male arthritic (MA) Group III: Castrated control (CSXC) Group IV: Castrated arthritic (CSXA) Group V: Castrated arthritic treated with DHT (CSXA ? T) Group VI: Female normal (FN) Group VII: Female arthritic (FA) Group VIII: Ovariectomized control (OVXC) Group IX: Ovariectomized arthritic (OVXA) Group X: Ovariectomized arthritic treated with estrogen (OVXA ? E) Group XI: Ovariectomized arthritic treated with progesterone (OVXA ? P) Group XII: Ovariectomized arthritic treated with estrogen and progesterone (OVXA ? E ? P) On day 28, rats were sacriced under ketamine anesthesia (44 mg/kg, i.p.). A portion of the synovial joint was dissected, a 10% homogenate was prepared in 0.1 M TrisHCl, pH 7.4, and another portion was preserved at -70C for estimation of collagen and GAG. The remaining joint was preserved in formalin for histology.

Measurement of Paw Swelling Changes in the edema of the right hind paw were measured using a plethysmometer. The hind paw volume was measured by dipping the foot into the mercury bath up to the anatomic hairline, i.e., lateral malleolus, and noting the extent of displacement of mercury as reected by the increase in the level of colored uid. The difference between initial and nal volumes indicated the edema volume. Percentage inhibition of paw volume was determined as: Vc Vt =Vc 100 where Vc is the arthritic-group edema volume, and Vt is the treatment-group edema volume. Lipid Peroxide Level The lipid peroxide level in articular tissues was determined by the method of Okhawa et al. [12] in terms of thiobarbituric acid-reactive substances (TBARS), an adduct of malondialdehyde, having maximum absorption at 532 nm. Antioxidant Assays Levels of reduced glutathione (GSH), superoxide dismutase (SOD), catalytic catalase (cCAT), peroxidative catalase (pCAT), and glutathione peroxidase (GPx) level in bone homogenate were determined according to Moron et al. [13], Marklund and Marklund [14], Caliborne [15], Quesenberry and Lee [16], and Rotruck et al. [17], respectively. Protein was assayed by Lowrys method [18]. The percentage increase in the levels of different antioxidants upon steroid treatment was calculated as follows: At Ac =Ac 100 where At is the level of antioxidant after the specic treatment and Ac is the level of the antioxidant in the respective diseased state. TNF-a Assay TNF-a level in the tissue homogenate was measured by sandwich ELISA according to the manufacturers protocol using the rat TNF-a ELISA development kit (Peprotech Asia). Briey, a microtiter plate was coated with 100 ll of capture antibody (rabbit anti-rat TNF-a, 1 lg/ml) and incubated overnight at room temperature (RT; 2832C). After washing, the plate was again incubated for 1 h at RT using block buffer (1% BSA in PBS). One hundred microliters of sample was added to each well after washing the plate thoroughly, and the plate was incubated for 2 h at RT. The plate was again washed and 100 ll of detection

123

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

357

antibody (biotinylated rabbit anti-rat TNF-a, 0.5 lg/ml) was then added and the plate incubated for 2 h at RT. After washing, 100 ll of avidin peroxidase was added and the plate was again incubated for 30 min at RT. This was followed by a nal wash and the addition of ABTS substrate solution. Absorbance was measured at 405 nm, with 650 nm as reference wavelength, using an ELISA reader (Multiskan Ascent VI.24, version 1.3.1). MMP-2 Estimation MMP-2 was estimated in the bone homogenate by ELISA. Briey, 100 ll of the homogenate was added to the microtiter plates, and the plates kept at 4C overnight. The plates were then blocked with 1% BSA. After washing in PBS-0.04%Tween, mouse anti-MMP-2 (Calbiochem) was added and the plate was incubated for 2 h at RT (28 32C). Washings were done as before and rabbit antimouse IgG conjugated to peroxidase (1:500) was added and incubated for 2 h at RT. Quantitation of the bound enzyme was made by incubation with freshly prepared ophenylenediamine dihydrochloride (0.5 mg/ml) in 0.1 M citrate buffer (pH 5.0) containing 0.012% H2O2 for 45 min at RT, and the reaction was terminated by the addition of 20 ll 8 N H2SO4. Absorbance was read at 490 nm, with 405 nm as reference wavelength, using an ELISA reader (Multiskan Ascent VI.24, version 1.3.1). Estimation of Collagen A portion of lyophilized joint tissue was hydrolyzed using 6 N HCl for 3 h at 130C in a sealed glass ampoule. Absorbance at 555 nm was determined by reaction of the sample with p-dimethyl aminobenzaldehyde [19]. The amount of hydroxyproline in the sample was calculated using pure hydroxyproline as standard. Since the hydroxyproline content of brocartilaginous collagen is 14%, a factor of 7.14 was used to convert hydroxyproline to collagen [20]. Estimation of Total and Sulfated Glycosaminoglycans To the lyophilized sample, papain (1 mg/ml) was added and the mixture incubated for 1 h at 37C. To a 0.1-ml volume of the above sample, 1.2 ml of basic dye solution (1 mg/ml alcian blue dye 8GX in 0.5 M sodium acetate, pH 8.6) was added and the contents were mixed thoroughly. After 15 min, the absorbance was measured at 480 nm against the blank [21] and this formed the total GAG in the tissues. Sulfated GAG content was determined by the addition of 1.2 ml of acid dye solution (1 mg/ml alcian blue dye 8GX in 15% phosphoric acid/2% H2SO4) to 0.1 ml of sample and the absorbance was measured at 480 nm against blank [22].

Histopathology A 10% buffered formalin-xed rat ankle joint was decalcied in 5% nitric acid, processed for parafn embedding, and sectioned at a 5-lm thickness. Sections were deparafnized and rehydrated. They were then stained for (i) collagen using Massons trichrome technique [23], (ii) elastin [24], and (iii) reticulin by silver impregnation [25]. Statistics Data are presented as the mean SE of six animals. Comparison between means of different groups was made by analysis of variance followed by Duncans multiplerange analysis using statistical software package SPSS, version 13.0. In Tables 1 and 2, means bearing different superscripts in a column differed signicantly.

Results Inhibition of Paw Volume Plethysmograph recordings (Table 1) showed severe edema of an ovariectomized arthritic rat which was suppressed signicantly on estrogen administration. Progesterone application, independently or in combination with estrogen, did not reduce the edema. Castrated male arthritic rats also had a higher edema volume, similar to that of female arthritic rats, and it was found to be lowered in DHT-treated rats. The percentage inhibition of edema volume showed that estrogen exhibited a signicantly (p \ 0.01) higher percentage inhibition (78.1%) compared to that of testosterone (67.3%). Progesterone did not show any inhibitory effect, and in combination with estrogen, it exhibited a slight reduction of edema (16.83%). Lipid Peroxidation Figure 1 demonstrated that peroxidation was increased in all groups of arthritic rats, with ovariectomized arthritic rats showing highly signicant levels. Treatment with testosterone (p \ 0.05) and estrogen (p \ 0.01) reduced the lipid peroxide level signicantly. Progesterone, independently or in combination with estrogen, did not inhibit lipid peroxidation. Effects of Steroids on Antioxidants Levels of antioxidants, viz., GSH, GPx, SOD, and catalase, were shown to decrease in the arthritic state for both male and female rats (Table 2). Treatment with testosterone and estrogen reverted the levels of antioxidants back to normal.

123

358

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

Table 1 Effects of steroids on edema volume in CIA in rats Group MN MA CSXC CSXA CSXA ? T Parameter Paw volume Paw volume Edema volume Paw volume Paw volume Edema volume Paw volume Edema volume Inhibition FN FA OVXC OVXA OVXA ? E Paw volume Paw volume Edema volume Paw volume Paw volume Edema volume Paw volume Edema volume Inhibition OVXA ? P OVXA ? E ? P Paw volume Edema volume Paw volume Edema volume Inhibition 0.95 0.01 0.96 0.01 0.96 0.01 0.96 0.01 0.97 0.01 0.96 0.01 0.95 0.01 0.97 0.01 0.97 0.01 0.96 0.01 Day 0 0.95 0.01 0.96 0.01 Day 7 0.97a 0.01 2.01 0.06 1.05 0.05 1.01a 0.02 2.29 0.02 1.32 0.02 1.73 0.03 0.75 0.03 43.18% 0.99 0.02 2.33 0.05 1.36c 0.04 0.98a 0.02 2.49 0.03 1.54d 0.04 1.60 0.04 0.65 0.04 57.8% 2.48a 0.02 1.51e 0.03 2.21 0.05 1.26 0.05 18.18%
c d a b e d a a b c d b c

Day 14 0.99a 0.0.02 1.98 0.06 1.02 0.06 1.04a 0.03 2.24 0.02 1.28 0.02 2.24 0.02 0.80 0.04 37.5% 1.02 0.02 2.31 0.04 1.34cd 0.03 1.0a 0.01 2.49 0.05 1.54e 0.06 1.64 0.03 0.69 0.04 55.2% 2.46a 0.03 1.49f 0.03 2.22 0.03 1.27 0.03 17.53%
c e a b f e a a c c e b d

Day 21 1a 0.02 2.82 0.08 1.86 0.07 1.08a 0.03 3.55 0.04 2.59 0.04 1.80 0.03 0.83 0.04 67.95% 1.04 0.02 3.65 0.07 2.67c 0.07 1.01a 0.01 4.06 0.04 3.10d 0.05 1.66 0.04 0.69 0.04 77.74% 4.1a 0.03 3.14e 0.03 3.56 0.06 2.6 0.05 16.13%
c d a b e d a a b c d b c

Day 28 1.02a 0.02 2.88d 0.07 1.92b 0.07 1.11a 0.04 3.59e 0.04 2.63c 0.04 1.84c 0.04 0.86a 0.04 67.3% 1.06a 0.03 3.67e 0.05 2.69c 0.05 1.03a 0.01 4.1f 0.04 3.15d 0.05 1.66b 0.06 0.69a 0.06 78.1% 4.13a 0.04 3.16f 0.04 3.58e 0.06 2.62c 0.06 16.83%

Note: Please see Methods (Experimental Groups) for group abbreviations. Values are the mean SE for six animals. Paw volume is expressed as milliliters of mercury displaced. Means in the same column bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.01

Table 2 Effects of steroids on antioxidant levels in collagen-induced arthritis in rats Group GSH GPx SOD Catalase Catalytic MN MA CSXC CSXA CSXA ? T FN FA OVXC OVXA OVXA ? E OVXA ? P OVXA ? E ? P 3.71a 0.08 2.85 0.17 3.68 0.12 2.28 0.14 3.06 0.14 3.69a 0.10 1.80e 0.10 3.67 0.09 1.20 0.10 3.21 0.11 1.29f 0.14 1.23 0.08
f b f a bc d a c

Peroxidative 2.82a 0.13 1.70c 0.06 2.70a 0.11 1.33e 0.04 1.89c 0.13 2.74a 0.15 1.12e 0.08 2.61a 0.15 0.58f 0.04 2.25b 0.11 0.70f 0.05 0.69f 0.05

79.81a 3.27 52.69 1.97 75.21 3.22 45.41 0.98 59.91 2.34 77.56a 4.37 35.18f 2.06 73.04 67.12
ab g cd e a d

1.33a 0.06 1.10 0.02 1.31 0.05 0.96 0.05 1.16 0.03 1.30a 0.06 0.82e 0.03 1.26 0.05 0.67 0.04 1.27
ab f a bc d a c

700a 10.2 600 11.3 712 9.6 550 12.5 640 11.8 742a 9.4 530e 12.0 753 8.6 480 9.3 680 11.7 470f 105 483 9.8
f b f a c e a d

3.56 1.31

19.63 2.24
cd

0.02

20.78g 1.84 22.13 1.01

g

0.68f 0.05 0.76 0.03

ef

Note: Please see Methods (Experimental Groups) for group abbreviations. All values are the mean SE for six animals. Units are as follows: GSH (reduced glutathione), mg GSH/g tissue; GPx (glutathione peroxidase), lmol GSH utilized/min/mg protein; SOD (superoxide dismutase), units of enzyme required to inhibit 50% pyrogallol autooxidation/min/mg protein; catalytic catalase, lmol H2O2 decomposed/min/mg protein; peroxidative catalase, lmol HCHO produced/min/mg protein. Means in the same column bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.05

123

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

359

Fig. 1 Effects of steroids on lipid peroxidation in CIA. All values are the mean SE for six animals. Means bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.05. Superscript bc (OVXA ? E) differs signicantly from superscript f (OVXA, OVXA ? P, OVXA ? E ? P), at p \ 0.01

Fig. 3 Effects of steroids on TNF-a level in CIA. All values are the mean SE for six animals. Means bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.05. Superscript d (MA) differs signicantly from superscript e (FA, CSXA), at p \ 0.01; superscript b (OVXA ? E) differs signicantly from superscripts g (OVXA), fg (OVXA ? P), and f (OVXA ? E ? P), at p \ 0.01

Fig. 2 Percentage increase in antioxidants on steroid treatment in CIA. Percentage increase in activities of different antioxidants upon treatment of castrated arthritic rats with testosterone and of ovariectomized arthritic rats with estrogen

On the contrary, progesterone, independently or in combination with estrogen, did not show any signicant effect. Figure 2 demonstrates that the percentage increase in the levels of antioxidants was higher in ovariectomized arthritic rats treated with estrogen compared to DHT-treated, castrated arthritic rats. TNF-a Levels The level of the proinammatory molecule, TNF-a, was increased in the arthritic groups. Treatment with testosterone and estrogen reduced the levels signicantly (p \ 0.05 and p \ 0.01, respectively), but progesterone did not show any signicant effect (Fig. 3). MMP-2 Levels The level of MMP-2 was increased signicantly in the arthritic groups, but it reverted upon administration of testosterone and estrogen (Fig. 4). On the contrary, progesterone administration was not noted to have any signicance. Matrix Biochemistry The level of collagen was reduced in castrated arthritic and ovariectomized arthritic rat groups (Fig. 5). Estrogen and

Fig. 4 Effects of steroids on MMP-2 level in CIA. All values are the mean SE for six animals. Means bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.05. Superscript b (OVXA ? E, CSXA ? T, MA) differs signicantly from superscript d (OVXA, OVXA ? P, OVXA ? E ? P), at p \ 0.01

Fig. 5 Effects of steroids on collagen level in CIA. All values are the mean SE for six animals. Means bearing different superscripts differ signicantly by Duncans multiple-range analysis, at p \ 0.05. Superscript c (OVXA ? E) differs signicantly from superscript a (OVXA), at p \ 0.01

testosterone restored the level to normal, whereas progesterone, independently or in combination with estrogen, was shown to signicantly (p \ 0.01) increase collagen levels above the control. Total and sulfated glycosaminoglycan levels were signicantly reduced in all arthritic groups (Fig. 6), but they

123

360

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

Fig. 6 Effects of steroids on total and sulfated glycosoaminoglycan levels in CIA. All values are the mean SE for six animals. *p \ 0.05 by Duncans multiple-range analysis

were observed to increase signicantly (p \ 0.05) upon testosterone and estrogen administration. Histopathology Figure 7 showed that Massons trichrome stained collagen bers blue. Castrated and ovariectomized arthritic rats exhibited fragmentation of collagen bers (Fig. 7b, d). Rats administered testosterone (Fig. 7c) and estrogen (Fig. 7e) had collagen levels comparable to that of normal rats (Fig. 7a), while groups administered progesterone showed intense collagen deposition (Fig. 7f, g). Figure 8 demonstrates that silver impregnation stained reticulin bers black. All arthritic-group rats stained positive for reticulin network (Fig. 8bg), with estrogen- and testosterone-treated groups showing a more organized network pattern (Fig. 8c, e, f). Verhoeffs stained elastin bers black. Thin needle-like bers were noted in control rats (Fig. 9a) but were absent in the arthritic groups. Also, it was observed that treatment with testosterone, estrogen, or progesterone did not restore the lost elastin bers (Fig. 9b, c, d).

Discussion Understanding the physiological and pathological role of sex hormones is of considerable importance not only to understand the basis of autoimmune diseases, but also to provide possibilities for new therapeutic strategies based on natures own experiment [26]. This study was carried out to unravel the role of sex hormones in matrix remodeling in CIA in rats. Castrated and ovariectomized arthritic rats were compared for changes in collagen and GAG levels to understand the inuence of sex hormones on connective tissue matrix. Castration as well as ovariectomy resulted in enhanced matrix loss in arthritis. It can be inferred from the results of this study that ovariectomy resulted in more severe inammation than castration did. To understand how each hormone at the physiological level

contributed to changes in extracellular matrix during the arthritis milieu, testosterone, estrogen, and progesterone were administered individually. Testosterone was physiologically converted to estrogen, and hence, to obviate this problem, DHT was administered, as DHT could not be aromatized and the effects studied could be attributed solely to androgens. The inammatory response in arthritic tissue led to generation of free radicals including reactive oxygen species (ROS). Aerobic cells were endowed with extensive antioxidant defense mechanisms to counteract the damaging effects of ROS [27]. However, when the balance between these reactive species and antioxidants was altered, a state of oxidative stress resulted, leading to permanent cellular damage. The uncontrolled production of free radicals in arthritis led to decreased GSH levels as a consequence of its consumption during oxidative stress and cellular lysis [28]. GPx catalyzed the reduction of a variety of hydroperoxides using GSH and protected mammalian cells against oxidative damage, resulting in a decreased level in this process. The decreased activities of SOD and catalase could be attributed to the enormous production of free radicals, viz., superoxide anion and hydroxyl radicals [29], due to the inammatory surge in arthritis. Administration of estrogen and testosterone caused a signicant elevation in the levels of antioxidant enzymes, indicating that the oxidative damage caused by ROS was attenuated, whereas progesterone administration did not show any signicant effect. This concurred with an earlier report where therapeutic substitution with 17-b-estradiol restored GPx activity in amenorrheic female patients, whereas progesterone had no signicant effect [30]. Testosterone is also known to induce the cellular antioxidant defense system [31]. An important inference that can be drawn from the results of this study is that estrogen is a more potent antioxidant molecule compared to testosterone (Fig. 1). Free radicals induced deleterious effects on various cellular targets, specically on membrane lipids, leading to extensive generation of lipid peroxidation. This study clearly shows that lipid peroxidation as well as reduced levels of endogenous scavengers can be taken as indirect in vivo evidence for the involvement of free radicals in the progression of arthritis and, also, that estrogen and testosterone signicantly inhibited TBARS levels. An increased level of ROS is known to induce proinammatory cytokines such as TNF-a [32]. Estrogen and testosterone could inhibit TNF-a by suppressing the oxidative stress in the tissue. Earlier in vitro studies demonstrated that sex steroids might exert an inhibitory effect on the secretion of proinammatory cytokines such as TNF-a, IL-1, and IL-6 from stromal cells of the synovium [8] or from macrophages [33], and our results concur with these observations. However, progesterone did not

123

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss Fig. 7 Massons trichrome staining showing the effects of steroids on collagen changes in CIA. (a) Control rat showing normal collagen network in the synovium. (b) CSX arthritic rat synovium showing fragmented collagen. (c) CSX arthritic rat treated with DHT showing collagen ber comparable to normal. (d) OVX arthritic rat showing extensively fragmented and damaged collagen bers. (e) OVX arthritic rat treated with estrogen showing normal collagen. (f) OVX arthritic rat treated with progesterone showing collagen deposition. (g) OVX arthritic rat treated with estrogen and progesterone showing collagen deposition

361

inhibit TNF-a levels, while the inhibitory effects of estrogen were blocked by progesterone. A similar observation was reported earlier in human hepatoma cells [34]. Oxygen-derived free radicals were also shown to be one of the possible causes of cartilage destruction through either a direct degradative effect on matrix components or an indirect action via activation of latent MMPs [35]. Synthesis and release of MMPs due to TNF-a activation led to matrix breakdown [36]. Our study evidences increased levels of MMP-2 in response to increased TNF-a

levels, and this was in agreement with earlier reports showing that stimulation with TNF-a augmented secretion of the active form of MMP-2 in RA [4]. MMP-2 rapidly degraded not only denatured collagen (gelatin) but also a number of native collagen types [37]. Our ndings also demonstrate decreased levels of GAG as well as fragmentation and loss of collagen in arthritis, concurring with the elevated levels of MMP-2. MMP-2 is one of the predominant MMPs capable of elastolysis. Elastin is relatively conserved throughout life,

123

362 Fig. 8 Gomoris staining showing the effects of steroids on reticulin changes in CIA. (a) Control rat showing normal reticulin network in the basement membrane. (b) CSX arthritic rat synovium showing reticulin in the joint. (c) CSX arthritic rat treated with DHT showing increased reticulin that is organized. (d) OVX arthritic rat showing reticulin in the joint. (e) OVX arthritic rat treated with estrogen showing organized reticulin network. (f) OVX arthritic rat treated with progesterone showing reticulin network. (g) OVX arthritic rat treated with estrogen and progesterone showing reticulin network

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss

and once damaged or destroyed, it is difcult to replace, since repair of elastin networks often results in malformed and dysfunctional elastin laments [38]. Our study demonstrates that elastin was lost in arthritic inammation and the loss was not attenuated upon alleviation of the inammatory state. Effects of estrogen on collagen and proteoglycan synthesis remain controversial, with a few studies supporting a stimulatory effect [39] while other studies advocate a suppressive action [40]. The results of our study showing a reduction in the loss of collagen and GAG upon administration of estrogen and testosterone could be due to the decreased MMP-2 levels due to the anti-inammatory

effects of these hormones. Earlier observations showed that estrogen inhibited the expression and activity of MMPs and tissue inhibitors of MMPs [41, 42]. Estrogen therapy was shown to inhibit MMP-2 and MMP-9 activity [43]. Progesterone administration led to increased levels of collagen but it did not inuence GAG level or reticulin disorganization and this could be due to the direct stimulation of collagen genes or to an increase in tissue inhibitor of MMP levels. Although this study shows that the benecial effects of estrogen were neutralized by progesterone, the mechanism by which progesterone altered the effects of estrogen remains to be elucidated. However, it was reported earlier

123

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss Fig. 9 Verhoeffs staining showing the effects of steroids on elastin changes in CIA. (a) Control rat showing normal elastin bers. (b) CSX arthritic rat treated with DHT demonstrates no elastin bers. (c) OVX arthritic rat treated with estrogen demonstrates no elastin bers. (d) OVX arthritic rat treated with progesterone demonstrates no elastin bers

363

that progesterone antagonized estrogens effects in the uterus by downregulation of cytoplasmic and nuclear estrogen receptor concentrations [44]. Although this work has demonstrated the protective effects of testosterone and estrogen in attenuating matrix loss in CIA, a quantitative comparison of the interaction of steroids with different collagen types and glycosaminoglycans is needed, and that would pave the way to understanding endocrine inuences on matrix remodeling in greater detail. In conclusion, this study has conrmed that physiological levels of testosterone and estrogen attenuate damage to reticulin organization as well as loss of collagen and GAG in CIA in rats by suppression of MMP-2 activation through their anti-inammatory and antioxidant activity. However, damaged elastin was not replaced by either treatment strategy. Administration of progesterone, independently or in combination with estrogen, did not prevent GAG loss but it did lead to collagen synthesis.
Acknowledgments The authors thank Dr. A. B. Mandal, Acting Director, CLRI, for his permission to publish this work. The award of a CSIR fellowship to Ms. Kalaivani Ganesan and Ms. Mitali Tiwari and the award of the CSIR Emeritus Scientist Project to Dr. R. Puvanakrishnan are gratefully acknowledged.

References
1. Fujisaki K, Tanabe N, Suzuki N, Kawato T, Takeichi O, Tsuzukibashi O, Makimura M, Ito K, Maeno M (2007) Receptor activator of NF-jB ligand induces the expression of carbonic anhydrase II, cathepsin K, and matrix metalloproteinase-9 in osteoclast precursor RAW264.7 cells. Life Sci 80:13111318

2. Morodomi T, Suzuki K, Sasaguri Y, Morimoto M, Nagase H (1992) Comparative studies of a 92 kDa gelatinolytic metalloproteinase from U937 cells and matrix metalloproteinase 2 from human rheumatoid synovial broblasts. Matrix Suppl 1:87 3. Unemori EN, Hibbs MS, Amento EP (1991) Constitutive expression of a 92-kD gelatinase (type IV collagenase) by rheumatoid synovial broblasts and its induction in normal human broblasts by inammatory cytokines. J Clin Invest 88:1656 4. Migita K, Eguchi K, Kawabe Y, Ichinose Y, Tsukada T, Aoyagi T, Nakamura H, Nagataki S (1996) TNF-a-mediated expression of membrane-type matrix metalloproteinase in rheumatoid synovial broblasts. Immunology 89:553557 5. Yoneda T, Ishimaru N, Arakaki R, Kobayashi M, Izawa T, Moriyama K, Hayashi Y (2004) Estrogen deciency accelerates murine autoimmune arthritis associated with receptor activator of nuclear factor-jB ligand-mediated osteoclastogenesis. Endocrinology 145:23842391 6. Forsblad DElia H, Mattsson LA, Ohlsson C, Nordborg E, Carlsten H (2003) Hormone replacement therapy in rheumatoid arthritis is associated with lower serum levels of soluble IL-6 receptor and higher insulin-like growth factor 1. Arthr Res Ther 5:R202R209 7. Antonio J, Wilson JD, George FW (1999) Effect of castration and androgen treatment on androgen-receptor levels in rat skeletal muscles. J Appl Physiol 87:20162019 8. Subramanian S, Tovey M, Afentoulis M, Krogstad A, Vandenbark AA, Offner H (2005) Ethinyl estradiol treats collageninduced arthritis in DBA/1LacJ mice by inhibiting the production of TNF-a and IL-1. Clin Immunol 115:162172 9. Ganesan K, Selvam R, Abhirami R, Narayana Raju KVS, Murali Manohar B, Puvanakrishnan R (2008) Gender differences and protective effects of testosterone in collagen induced arthritis in rats. Rheumatol Int 28:345353 10. Ganesan K, Balachandran C, Murali Manohar B, Puvanakrishnan R (2008) Comparative studies on the interplay of testosterone, estrogen and progesterone in collagen induced arthritis in rats. Bone 43:758765

123

364

K. Ganesan et al.: Estrogen and Testosterone Attenuate Extracellular Matrix Loss 30. Massafra C, Buenocore G, Giola D, Sargentini I, Farina G (1997) Effects of estradiol and medroxyprogesterone-acetate treatment on erythrocyte antioxidant enzyme activities and malondialdehyde plasma levels in amenorrhoic women. J Clin Endocrinol Metab 82:173175 31. Chisu V, Manca P, Lepore G, Gadau S, Zedda M, Farina V (2006) Testosterone induces neuroprotection from oxidative stress. Arch Ital Biol 144:6373 32. Lean JM, Jagger CJ, Kirstein B, Fuller K, Chambers TJ (2005) Hydrogen peroxide is essential for estrogen-deciency bone loss and osteoclast formation. Endocrinology 146:728735 33. Kramer PR, Kramer SF, Guan G (2004) 17 Beta-estradiol regulates cytokine release through modulation of CD16 expression in monocytes and monocyte-derived macrophages. Arthr Rheum 2004:19671975 34. Cheng X, Shimizu I, Yuan Y, Wei M, Shen M, Huang H, Urata M, Sannomiya K, Fukuno H, Hashimoto-Tamaoki T, Ito S (2006) Effects of estradiol and progesterone on TNF-a-induced apoptosis in human hepatoma HuH-7 cells. Life Sci 79:19881994 35. Burkhardt H, Schwingel M, Menninger H (1986) Oxygen radicals as effectors of cartilage destruction. Arth Rheum 29:379387 36. Poole AR, Nelson F, Hollander A, Reiner A, Pidoux I, Ionescu M (1995) Collagen III turnover in joint diseases. Acta Orthop Scand Suppl 266:8891 37. Woessner JF Jr (1991) Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J 5:2145 2154 38. Russell RE, Thorley A, Culpitt SV (2002) Alveolar macrophagemediated elastolysis: roles of matrix metalloproteinases, cysteine, and serine proteases. Am J Physiol Lung Cell Mol Physiol 283:L867L873 39. Corvol MT, Carrascosa A, Tsagris L, Blanchard O, Rappaport R (1987) Evidence for a direct in vitro action of sex steroids on rabbit cartilage cells during skeletal growth: inuence of age and sex. Endocrinology 120:14221429 40. Tsai CI, Liu TK (1993) Estradiol-induced knee osteoarthrosis in ovariectomised rabbits. Clin Orthop 291:295302 41. Lee YJ, Lim WC, Jin YH, Lee SK (2002) Effects of estrogen receptor and estrogen on the chromatin structure in estrogen receptor stable transfectants. Exp Mol Med 34:9599 42. Oestergaard S, Sondergaard BC, Hoegh-Andersen P, Henriksen LB, Karsdal MA (2006) K, Qvist P, Christiansen C, Tanko Effects of ovariectomy and estrogen therapy on type II collagen degradation and structural integrity of articular cartilage in rats: implications of the time of initiation. Arthr Rheum 54:24412451 43. Nielsen RH, Christiansen C, Stolina M, Karsdal MA (2008) Oestrogen exhibits type II collagen protective effects and attenuates collagen-induced arthritis in rats. Clin Exp Immunol 152:2127 44. Clarke CL, Sutherland RL (1990) Progestin regulation of cellular proliferation. Endocr Rev 11:266301

11. Ashok Kumar D, Settu K, Narayana Raju KVS, Kumanan K, Murali Manohar B, Puvanakrishnan R (2006) Inhibition of nitric oxide and caspase 3 mediated apoptosis by a tetrapeptide derivative (PEP 1261) in cultured synovial broblasts from collagen induced arthritis. Mol Cell Biochem 282:125139 12. Okhawa H, Ohishi N, Yagi K (1979) Assay for lipid peroxides in animal tissues by thiobarbituric acid reaction. Anal Biochem 95:351358 13. Moron MS, Depierre JW, Mannervik B (1979) Levels of glutathione, glutathione reductase and glutathione-S-transferase activities in rat lung and liver. Biochim Biophys Acta 582:6778 14. Marklund SL, Marklund G (1974) Involvement of superoxide anion radical in the autooxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur J Biochem 47:469474 15. Caliborne AL (1985) Assay of catalase. In: Greenwals RA (ed) Handbook of oxygen radical research. CRC Press, Baco Raton, FL, pp 203207 16. Quesenberry MS, Lee YC (1996) A simple spectrophotometric method for estimation of peroxidase activity in blood samples. Anal Biochem 234:5055 17. Rotruck JT, Pope AL, Ganther HE, Swanson AB, Hafeman DG, Hekstra WG (1973) Selenium, biochemical role as a component of glutathione peroxidase purication and assay. Science 179:588590 18. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with folin phenol reagent. J Biol Chem 193:265 275 19. Fan L, Sarkar K, Franks DJ, Uhthoff HK (1997) Estimation of total collagen and types I and III collagen in canine rotator cuff tendons. Calcif Tissue Int 61:223229 20. Neumann RE, Logan MA (1950) The determination of collagen and elastin in tissues. J Biol Chem 186:549556 21. Gold EW (1979) A simple spectrometric method for estimating glycosoaminoglycan concentrations. Anal Biochem 99:183188 22. Gold EW (1981) The quantitative spectrophotometric estimation of total sulphated glycosaminoglycan levels. Biochim Biophys Acta 673:408415 23. Masson P (1929) Some histological methods. Trichrome stainings and their preliminary technique. Bull Int Assoc Med 12:75 24. Verhoeff EH (1908) Some new staining methods of wide applicability, including a rapid differential stain fro elastic tissue. JAMA 50:876 25. Gomori G (1937) Silver impregnation of reticulin in parafn sections. Am J Physiol 13:993 26. Jansson L, Holmdahl R (1998) Estrogen-mediated immunosuppression in autoimmune diseases. Inamm Res 47:290301 27. Halliwell B (1994) Free radicals, antioxidants, and human disease: curiosity, cause, or consequence? Lancet 344:702 28. Hassan MQ, Hadi RA, Al-Rawi ZS, Padron VA, Stohs SJ (2001) The glutathione defense system in the pathogenesis of rheumatoid arthritis. J Appl Toxicol 22:6973 29. Imadaya A, Katsutoshi T, Hiroyori T, Meguni O, Kazuo T (1988) Erythrocytic antioxidant enzymes are reduced in patients with rheumatoid arthritis. J Rheumatol 15:1137

123