Estrogen, a representative steroid hormone, is known to regulate diverse physiological processes of target tissues, including reproductive (Galand et al., 1971), cardiovascular (for review, see Dimitrova et al., 2002), skeletal (Migliaccio et al., 1992), nervous (for review, see Behl, 2002), and connective tissues (Clark et al., 1992) in both sexes. The biological activities of estrogen are initiated by a binding to the specific receptor proteins, namely the estrogen receptors (ERs) (for review, see Pavao and Traish, 2001). Two main isoforms of ER have been identified to date: ERα and ERβ. After the recent discovery of ERβ (Kuiper et al., 1996), the classic, most predominant type was renamed as ERα. Immunocytochemistry using specific antibodies has revealed that ERα is widely distributed in various tissues, whereas the definitive distribution of ERβ protein remains unclear because specific antibodies to ERβ are as yet unavailable (Pavao and Traish, 2001).
The temporomandibular joint (TMJ) is a bilateral diarthrosis between the mandibular condyle and temporal bone. Temporomandibular disorder (TMD) is characterized by a triad of symptoms, including joint sounds, pain, and limited mandibular movement (Kuttila et al., 1998). These symptomatic variants indicate that TMD is caused by a combination of factors, such as occlusion, mental stress, strength, endurance, and hormones; however, the etiology of this disease is not fully understood. Several epidemiological studies have reported that TMD is more prevalent in women than in men (Campbell et al., 1993; LeResche, 1997; Kapila and Xie, 1998), which suggests the involvement of sex hormones, such as estrogen, in the pathogenesis of this disease. Although estrogen is known to play important roles in the etiology of postmenopausal osteoarthrosis or rheumatoid arthritis in systemic joints (Ushiyama et al., 1995; Khalkhali-Ellis et al., 2000), little information has been available regarding the relationship between estrogen and the etiology of TMD (for review, see Warren and Fried, 2001).
The synovial membrane in the TMJ has an important role in joint movement because of the involvement of the synovial lining cells in the synovial fluid metabolism, which effects smooth jaw movement. Many ultrastructural investigations have pointed out that the synovial membrane consists of two kinds of synovial lining cells: macrophage-like type A and fibroblastic type B cells (Barland et al., 1962; Graabæk, 1984). In rodents, type A cells are characterized by numerous vesicles, vacuoles, and lysosomes, while type B cells possess a well-developed rough endoplasmic reticulum and a number of dense secretory granules. In addition, recent immunocytochemical and scanning electron microscopy studies (Nozawa-Inoue et al., 1999; Iwanaga et al., 2000; Andoh et al., 2001; Nio et al., 2002) revealed that type B cells develop characteristic cytoplasmic processes that are easily distinguishable between type A and B cells. Type B cells have thick dendritic processes that extend toward the articular cavity and partially cover the synovial membrane, whereas type A cells develop dense filopodia-like surface folds. Therefore, the morphology of the cytoplasmic processes is a useful marker for type B cells in the TMJ.
Heat shock proteins (Hsps) protect cells against irreversible damage when the cells are exposed to stressful conditions (Lindeman et al., 1998). Hsps generally can be divided into two groups of Hsps (large or small) (Hopkins et al., 1998) according to their molecular weight. Hsp25 (one of the small Hsps) serves to protect cells against various stimuli, such as hyperthermia (Lavoie et al., 1993), oxidative stress (Mehlen et al., 1995), and inflammatory cytokines (Mehlen et al., 1995). In addition, the involvement of Hsp25 in cell differentiation has been observed even under normal conditions (Welsh and Gaestel, 1998). In our previous studies (Nozawa-Inoue et al., 1999; Andoh et al., 2001), various cellular elements of TMJ were shown to be intensely immunoreactive for Hsp25 in normal rats and mice. In particular, type B lining cells in the synovial membrane more intensely expressed Hsp25 protein and mRNA, which suggests that Hsp25 is a useful marker for type B synoviocytes. The fact that Hsp25 was originally discovered as an estrogen-regulated protein indicates a relationship between Hsp25 and estrogen (Edwards et al., 1981). Furthermore, biochemical studies have shown that hsp25 may be directly upregulated by estrogen (Porter et al., 1996) due to the estrogen-responsive element (ERE) in the 5′ promoter region. Taken together, these findings easily lead us to a hypothesis that the TMJ is a target tissue for estrogen. However, little information is available regarding the localization of ERα in the TMJ. In the present study, therefore, we examined the expression of ERα protein and mRNA in the TMJ of adult male rats by immunocytochemistry and in situ hybridization using a specific antiserum and cDNA probe, respectively.
MATERIALS AND METHODS
All experiments were performed under the guidelines of the Niigata University Intramural Animal Use and Care Committee.
Animals and Tissue Preparation
Six Wistar rats (4 weeks old, weighing approximately 100 g) were used in this study. The rats were all male, to avoid any influence from the sexual cycle in female rats on the results. They were anesthetized by an intraperitoneal injection of chloral hydrate (400 mg/kg), and perfused with a fixative containing 4% paraformaldehyde and 0.025% glutaraldehyde in a 0.067 M phosphate buffer (pH 7.4). The TMJs were removed en bloc and immersed in the same fixative for 12 hr. After the tissue blocks were decalcified with 10% ethylenediamineteraacetic acid disodium (EDTA-2Na) solution for 4 weeks at 4°C, they were dehydrated through a series of graded ethanol, cleared in xylene, and embedded in paraffin. Serial paraffin sections were sagittally cut (5 μm thick), mounted onto silane-coated glass slides, and dried overnight. Some paraffin sections were stained with hematoxylin and eosin.
Deparaffinized sections were processed for immunohistochemistry for ERα by the avidin-biotin-complex (ABC) method (Hsu et al., 1981). After the inhibition of endogenous peroxidase with 0.3% H2O2 in absolute methanol for 15 min, the sections were reacted with a polyclonal antiserum against ERα (1:50; Santa Cruz Biotechnology, Inc., Santa Cruz, CA)—which recognizes ERα from mice, rats, and humans—for 24 hr at 4°C. The bound primary antibody was then localized using biotinylated anti-rabbit IgG (Vector Laboratories, Inc., Burlingame, CA) and then an avidin-peroxidase complex (ABC kit; Vector Laboratories, Inc.) for 90 min each at room temperature. For the final visualization, 0.04% 3,3′-diaminobenzidine tetrahydrochloride and 0.002% H2O2 in a 0.05 M Tris-HCl buffer (pH 7.6) were used.
For confocal microscopic observation, deparaffinized sections incubated with the same primary antibody (1:50) were reacted with fluorescein isothiocyanate (FITC)-conjugated anti-rabbit IgG (1:100; Vector Laboratories, Inc.) for 1 hr at room temperature. To visualize the nuclei, the sections were incubated with DNase-free RNase (100 μg/mL; Sigma-Aldrich Co., St. Louis, MO) in 2× standard saline citrate (SSC) for 20 min at 37°C, and counterstained with propidium iodide (1:3,000; Molecular Probes, Inc., Eugene, OR) in 2× SSC for 5 min. The sections were then examined in a confocal laser scanning microscope (LSM 510; Carl Zeiss, Jena, Germany).
For immunoelectron microscopy, ERα in the rat TMJs was visualized by an immunogold labeling method. Decalcified tissue blocks were sectioned at a 500 μm thickness with a Microslicer (Dosaka EM, Kyoto, Japan), dehydrated with N,N-dimethylformamide, and embedded in glycolmethacrylate (GMA; Nisshin EM Co. Ltd., Tokyo, Japan). Ultrathin GMA sections were prepared in a Leica Ultracut R (Leica Microsystems, Wetzlar, Germany) ultramicrotome with a diamond knife, and collected on nickel grids. The sections were soaked with 1% bovine serum albumin for 15 min, and reacted with two consecutive incubations with rat anti-ERα monoclonal antibody (1:500; Stressgen Biotechnologies Corp., Victoria, Canada) and gold conjugated rabbit anti-rat IgG (1:100, 10-nm gold particles; Sigma Chemical Co., St. Louis, MO) for 1 hr each at room temperature. Following a rinsing in 0.01 M phosphate-buffered saline (PBS) and distilled water, they were stained with lead citrate and examined under a Hitachi H-7000 transmission electron microscope (Hitachi Co. Ltd, Tokyo, Japan).
Immunocytochemical controls of the antiserum were made by 1) replacing the primary antibody with nonimmune serum or PBS, and 2) omitting the anti-rabbit IgG or the avidin-peroxidase complex.
A proven ERα probe, complementary to 301-346 bp of ERα cDNA (Koike et al., 1987), was used in this study. The oligonucleotide was labeled with 35S-dATP, using terminal deoxyribonucleotidyl transferase (Promega, Madison, WI) at a specific activity of 5 × 108 d.p.m./μg DNA.
In Situ Hybridization
Three male Wistar rats were decapitated under deep anesthesia, as described above. Immediately afterward, the synovial membranes of the TMJ were removed and frozen in liquid nitrogen. These specimens were kept in a deep freezer at –80°C prior to use. An in situ hybridization procedure was performed according to Mowa and Iwanaga (2001). Briefly, 20 μm cryostat sections obtained from fresh-frozen tissue were mounted onto glass slides precoated with 3-amino-propyltriethoxysilane, fixed in 4% paraformaldehyde for 10 min, and then acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine-HCl (pH 8.0) for 10 min. The prepared sections were prehybridized for 2 hr in a buffer containing 50% formamide, 0.1 M Tris-HCl (pH 7.5), 4 × SSC (1 × SSC; 150 mM NaCl, and 15 mM sodium citrate), 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin, 0.6 M NaCl, 0.25% sodium dodecyl sulfate (SDS), 200 μg/ml tRNA, 1 mM EDTA, and 10% dextran sodium sulfate. Hybridization was performed at 42°C for 10 hr. The slides were washed in 2× SSC containing 0.1% sarkosyl (Nacalai Tesque, Kyoto, Japan) and twice at 55°C in 0.1× SSC containing 0.1% sarkosyl. The sections were dipped in Kodak NTB2 nuclear track emulsion and exposed.
The specificity of the in situ hybridization was checked by the disappearance of signals when an excess dose of corresponding cold oligonucleotides was added to the hybridization fluid. Consistent ERα mRNA signals above background levels were considered positive.
Immunocytochemistry for ERα demonstrated an intense immunoreaction in various cellular elements (the synovial lining cells, vascular endothelial cells, synovial fibroblasts, stromal cells in the articular disc, and chondrocytes) in the rat TMJ.
ERα-Immunoreaction and mRNA in the Synovial Membrane of Rat TMJ
The inner surface of the articular capsule, but not the articular cartilage or the articular disc, was covered with a synovial membrane (Fig. 1). This synovial membrane formed a few folds protruding into the articular cavity, particularly in the postero-superior portion (Fig. 1). Immunocytochemistry showed that the synovial lining layer consisted of one to three layers of ERα-immunopositive and -negative cells (Fig. 2a and c). Under light microscopy, both types of synovial lining cells appeared round in profile. Although the arrangement of the synovial lining cells showed a regional difference (they formed a comparatively thick cell layer at the anterior portion and the tip of the synovial fold (Fig. 2a), and a single flattened cell layer at the other area (Fig. 2c)), synovial lining cells with ERα immunoreactivity were recognized in all portions of the synovial membrane. Intense ERα immunoreactions were found in the nuclei of the synovial lining cells, whereas there were weak reactions in the cytoplasm (Fig. 2c).
Confocal laser microscopic observations with ERα immunostaining and nuclear propidium iodide demonstrated more clearly the immunolocalization of ERα in the synovial lining cells. ERα-positive synovial cells exhibited intense and weak immunoreactions in the nucleus and cytoplasm, respectively; however, a few cells lacked any nuclear immunoreaction despite the positive reaction in their cytoplasm. Furthermore, comparatively thick cell processes and slender cytoplasmic processes lining the articular cavity were also seen to exhibit ERα immunoreactivity (Fig. 2b).
The sublining layer of the synovial membrane also contained ERα-immunopositive cells, and some endothelial cells and fibroblasts showed an immunoreaction in their nuclei (Fig. 2b and c). However, immunocontrol sections did not show any specific immunoreaction for ERα in the synovial membrane (Fig. 2d).
At the electron microscopic level, gold particles indicative of ERα immunoreactivity were scattered in the nucleus, cytoplasm, and plasma membrane of the synovial cells (Fig. 3). These immunopositive cells extended thick, long cytoplasmic processes toward the articular cavity and covered the surface of the synovial membrane (Fig. 3a), which indicates that they were type B cells.
In situ hybridization histochemistry showed intense signals for ERα mRNA in the synovial lining layer and the sublining layer (Fig. 4a). The synovial lining layer contained cells with and without ERα mRNA (Fig. 4a). In control sections incubated with an excess dose of corresponding cold oligonucleotides, no specific signals for ERα mRNA were recognizable in the synovial membrane (Fig. 4b).
ERα-Immunoreactivity in the Articular Disc
The articular disc also contained ERα-positive cells. Intense immunoreactions were particularly noted in the nuclei of the articular stromal cells located in the anterior and posterior bands (Fig. 5a and b). Immunoreactive stromal cells were rare in the central portion (Fig. 5c). The positive cells were round in shape or possessed short cytoplasmic projections, appearing stellate in profile (Fig. 5a and b).
ERα-Immunoreactivity in the Mandibular Condyle
Immunostaining revealed localization of ERα-immunoreactivity in some populations of chondrocytes in the mandibular condyle of the rat TMJ (Fig. 6). Since the terminology for the cartilaginous layer of the rat condyle is not yet well established, we divided the articular cartilage into four layers based on the classification by Bloom and Fawcett (1975): fibrous, proliferative, maturative, and hypertrophic. The cartilaginous layers did not show any uniform ERα immunoreactivity. ERα-positive chondrocytes were more abundant in the maturative and hypertrophic cell layers in the condyle compared to the fibrous and proliferative layers (Fig. 6). The nuclei showed more intense ERα immunoreactions than the cytoplasm in these chondrocytes (Fig. 6). In contrast, a few chondrocytes in the fibrous and proliferative cell layers demonstrated a weak immunoreactivity for ERα (Fig. 6).
A wide distribution of ERα-immunoreactions has been reported in primate dermal fibroblasts (Bentley et al., 1986); human fibroblasts; human endothelial cells (Brandi et al., 1993); mammary gland cells (for review, see Pelletier, 2000); chondrocytes in cows, pigs, and humans (Claassen et al., 2001); human osteoblasts (Eriksen et al., 1988); and human osteoclasts (Pensler et al., 1990), in addition to female reproductive tissues, indicating that estrogen has diverse effects on development, growth, and homeostasis (Clark et al., 1992). Although many epidemiological studies have reported a higher frequency of TMD in females than in males (Campbell et al., 1993; LeResche, 1997; Kapila and Xie, 1998), there has been controversy concerning the presence of ERα in the TMJ. Abubaker et al. (1993) found ERα immunoreactivity in the articular disc of human TMJs; however, Campbell et al. (1993) found no such immunoreactivity in the articular disc of a TMJ obtained from a TMD patient. Furthermore, some researchers have observed sexual dimorphism in ER expression in the baboon TMJ by autoradiography (Aufdemorte et al., 1986; Milam et al., 1987). With the use of immunocytochemistry in the present study, we were able to clearly demonstrate the distribution of ERα in various cellular elements in the rat TMJ, including synoviocytes, stromal cells in the articular disc, and chondrocytes. Furthermore, certain types of synovial lining cells expressed ERα mRNA, as demonstrated by in situ hybridization histochemistry. To our knowledge, this is the first report to demonstrate a precise distribution of ERα in normal male rats, which indicates that TMJ is a target tissue for estrogen.
The lining layer of the synovial membrane contains two kinds of lining cells: macrophage-like type A and fibroblastic type B cells (Barland et al., 1962; Graabæk, 1984). We were unable to observe all organellae in the ERα-positive cells due to the use of GMA resin, which is disadvantageous for detailed observations. However, these two types of lining cells are also distinguished by their cytoplasmic processes (Iwanaga et al., 2000; Nio et al., 2002). The current confocal and immunoelectron microscopic observations indicated that these ERα-immunopositive synovial lining cells can be categorized as type B cells because of the above-mentioned characteristic profiles. This evidence supports the possibility that synovial lining cells with ERα mRNA are fibroblastic type B cells. This notion is compatible with previous reports on the human cruciate ligament (Liu et al., 1996) and cultured synoviocytes obtained from human RA patients (Khalkhali-Ellis et al., 2000).
The functional significance of ERα/estrogen in synovial type B cells remains unclear. In our previous reports (Nozawa-Inoue et al., 1999; Andoh et al., 2001), type B cells in rat and mouse TMJs showed an immunoreactivity for Hsp25, which was originally categorized as an estrogen-related protein (Edwards et al., 1981). Moreover, synovial type B cells have been reported to synthesize and secrete collagens, fibronectin (Matsubara et al., 1983; Mapp and Revel, 1985), and glycosaminoglycans (including hyaluronan) (Roy and Ghadially, 1967). The administration of estrogen has been shown to induce a decrease in the volume and synthesis of collagen in the rat tendon (Fischer, 1973), periodontal tissue (Dyer et al., 1980), and TMJ discs (Abubaker et al., 1996). Therefore, it is reasonable to consider that estrogen might regulate metabolism—including collagen synthesis—via ERα expressed in synovial type B cells and stromal cells in the articular disc.
In the current study, condylar cartilage showed ERα immunoreactivity in chondrocytes in both the maturative and hypertrophic layers, consistent with a previous report that investigated ERα expression in cultured rat mandible (Ng et al., 1999). The effects of estrogen on bone metabolism are well established. In contrast, there have been few studies regarding the function of estrogen in cartilage in the TMJ, or even in other systematic joints (Claassen et al., 2001). Ng et al. (1999) demonstrated that mandibular cartilage cultured with media containing estrogen caused a decrease in the extracellular matrix and thickness of this cartilage. These results suggested that estrogen downregulates cartilage metabolism, as a result of binding with ERα in the maturative and hypertrophic chondrocytes.
It is interesting that ERα-immunoreactions were clearly observed (by confocal laser microscopy and immunoelectron microscopy) in the cytoplasm and plasma membrane as well as in the nucleus. The traditional concept of the biological function of ERα is based on transcriptional activity regulated by the ERE of target genes. In contrast, an alternative pathway for estrogen action—nongenomic action of this molecule—has been predicted under physiological conditions; however, this is still the subject of ongoing discussions. Cytoplasmic and plasma membrane complexes of estrogen/ERα appear to participate in signal transduction, resulting in the regulation of cell growth, survival, and migration (for review, see Levin, 2002; Razandi et al., 2003). The activity of p38, a member of the mitogen-activated protein (MAP) kinase family, is stimulated by estrogen. Upregulation of p38 by estrogen gives rise to activation of the MAPKAPK-2 kinase, and the phosphorylation of Hsp27 in human endothelial cells. Estrogen has been shown to utilize this pathway to protect endothelial cells from any metabolic disruption of actin cytoskeleton or hypoxia-induced cell death, and to stimulate angiogenesis (Razandi et al., 2000). Interestingly, our previous immunocytochemical study for Hsp25, a homologue of human Hsp27 (Gaestel et al., 1993), showed cytosolic expression in synovial type B lining cells, stromal cells in the articular disc, and chondrocytes in the rat TMJ (Nozawa-Inoue et al., 1999). The Hsp25 immunoreaction pattern appears to be comparable to the expression pattern of ERα, as shown in this study. Thus, it is likely that ERα and Hsp25 colocalize in the cells of the rat TMJ, and cooperate with each other on such a signal transduction.
Our own immunocytochemical observations provided a clue as to the putative involvement of ERα in the physiological function of TMJ tissues, at least in male rats. However, Kennedy et al. (1999) reported a stage-specific expression of ERα-immunoreaction in chondrocytes of femoral growth plates. In that study, an intense reaction occurred in immature rats but not in aged ones. This finding indicates that reproductive and skeletal tissues have different controls for ERα expression. Further investigations on age- and sex-related changes in ERα expression are needed to clarify the functional significance of estrogen in the TMJ.
The authors thank Mr. M. Hoshino and K. Takeuchi, Division of Oral Anatomy, Niigata University Graduate School of Medical and Dental Sciences, for their technical assistance.