NF-κB Activation and iNOS Expression in the Synovial Membrane of Rat Temporomandibular Joints after Induced Synovitis

Abstract

NF-κB plays a pivotal role in pathogenesis in general arthritis. However, the participation of NF-κB in inflammation of the temporomandibular joint (TMJ) is poorly understood. We examined NF-κB expression in rat TMJs with synovitis induced by condyle hypermobility. By immunohistochemistry, NF-κB immunoreactivity was found mainly in the cytoplasm, not the nucleus, of the synovial lining cells of induced-synovitis and control TMJs. Southwestern histochemistry, a new method for detecting transcription factors, showed greater NF-κB expression in the nucleus of the synovial lining cells in the hypertrophic synovium than in control synovium. Increased numbers of the synovial lining cells with immunoreactivity for inducible nitric oxide synthase (iNOS), which is transcriptionally regulated by NF-κB, were also seen in the inflamed synovium. These findings indicate that excess mechanical stress increases NF-κB activation in the TMJ and suggest that active NF-κB is involved in the progression of TMJ inflammation.

INTRODUCTION

The mechanism for the development and progression of TMJ disorders has not been elucidated. Milam et al. (1998) proposed three models to explain the pathogenesis of TMJ disorders: direct mechanical injury, hypoxia/reperfusion, and neurogenic inflammation models. They speculated that mechanical stress causes the initial stage of each model, and presumed that TMJ tissue is damaged directly or indirectly by excessive mechanical stress. They also postulated that the free radicals generated by mechanical stress during jaw movement have significant effects in the initiation of TMJ disorders. Abundant NO production (Takahashi et al., 1996) and intense iNOS expression (Homma et al., 2001) in the synovial membrane are found in patients with inflamed TMJs. It has been suggested that NO generated by iNOS participates in the progress of TMJ inflammation. Recently, we localized iNOS in the synovial lining cells in normal rat TMJs, and postulated that NO also acts as a mediator to maintain the physiological condition of the TMJ (Masuda et al., 2002). Nerve fibers containing substance P, which is thought to modulate neurogenic inflammation, are also found in normal rat TMJ synovium (Kido et al., 1993).

Transcription factors control and induce the expression of several genes. NF-κB plays a pivotal part in regulating inflammation-associated genes (Baeuerle and Hankel, 1994). In the normal state, this factor localizes in the cytoplasm as a complex with inhibitory IκB subunits. IκB masks the nuclear localization signal (NLS) of NF-κB to prevent its nuclear translocation. In stimulated cells, this inhibitor is phosphorylated, polyubiquinated, and rapidly degraded by proteasome. Dissociated NF-κB is the active form; it translocates into the nucleus, where it accelerates transcription by binding to specific elements known as κB sites in the promoter regions of target inflammation-associated genes [interleukin 1 (IL-1), tumor necrosis factor a (TNF-α), iNOS, etc.]. In rheumatoid arthritis (RA), NF-κB activation is seen in hypertrophy of the synovium (Handel et al., 1995; Marok et al., 1996). In streptococcus cell wall (SCW)- or collagen-induced arthritis model rats, NF-κB is also activated in the synovium (Miagkov et al., 1998; Szabo et al., 1998). Furthermore, the development of arthritis induced by SCW is markedly inhibited by intra-articular injection of NF-κB decoys (Miagkov et al., 1998). These findings suggest that NF-κB plays an important role in regulating synovial inflammation.

Milam et al. (1998) proposed that NF-κB, activated by oxidative stress that is induced by excess mechanical stress during functional or pathological jaw movement, participates in the process of TMJ inflammation. However, it is still not clear whether NF-κB is involved in TMJ arthritis. Therefore, besides examining the expression of iNOS in the inflamed TMJ, the aim of this study was to examine the participation of NF-κB in the progression of TMJ arthritis (synovitis) by analyzing the localization of activated NF-κB in the synovial membrane of rat TMJs induced by condylar hypermovement. We used conventional immunohistochemistry and Southwestern histochemistry, a novel method of localizing transcription factors in situ.

MATERIALS & METHODS

Inducing Synovitis in the Rat TMJ

Twenty male Wistar rats (4 wks old) were housed in a specific-pathogen-free, temperature-controlled room on a 12-hour alternating light-dark cycle, and given food and water ad libitum throughout the experimental period. The experimental procedure for inducing synovitis in the rat TMJ was similar to that described previously (Muto et al., 1998a,b) with slight modifications. With each rat under ether anesthesia, its head was fixed in the dorsal position, and condyle hypermobility was forced by maximal mouth opening (approximately 20 mm). This procedure was repeated 20 times once a day for 8 wks. As a control, 8 non-forced, age-matched rats were used. Throughout this study, the animals were treated according to the guidelines for animal care at Kyushu University.

Tissue Preparation

After the last round of hypermovement, the rats were perfused with 4% paraformaldehyde (PFA) and 0.2% picric acid in 0.1 M phosphate buffer (PB) (pH 7.4) at 4°C for immunohistochemistry and histopathological examination (10 experimental rats and 4 controls), or 4% PFA in 0.01 M phosphate-buffered saline (PBS) (pH 7.2) at 4°C for Southwestern histochemistry and histopathological examination (10 experimental rats and 4 controls). Serial 10-μm-thick sagittal cryosections of the central portion of the TMJ were cut and mounted on gelatin- or 3-aminopropyltriethoxysilane-coated glass slides.

Histopathologic Examination and Quantitative Scoring of Synovitis

The right TMJs from experimental (N = 20) and control (N = 8) animals were used. Sections were selected from each TMJ and stained with hematoxylin and eosin. Three sagittal sections, corresponding to the sagittal schematic diagram of the central portion of the rat TMJ (Fig. 1a), were selected, so that we could examine the changes in the thickness of the synovial cell layers in the 4 synovial portions of Fig. 1a . We used the method of Muto et al. (1998a,b) based on the change of the thickness of the synovial cell layers, graded as a marker of synovitis: grade 0, 1-3 layers; grade 1, 4-6 layers; and grade 2, 7 or more layers.

Immunohistochemistry

Sections from the left TMJs of experimental (N = 10) and control (N = 4) rats were stained by means of a Vectorstain ABC kit (Vector Laboratories, Burlingame, CA, USA) (Masuda et al., 2002). Typical sagittal sections were selected in the same manner as above. Rabbit anti-human NF-κB IgG (0.5 μg/mL) or anti-human iNOS IgG (2 μg/mL) were used as the primary antibodies. After being immunostained, the sections were stained with or without hematoxylin. As immunohistochemistry controls, some sections were incubated in the same way, but with non-immune rabbit IgG or with PBS alone instead of the primary antibodies.

Rabbit anti-human NF-κB (sc-372) and anti-human iNOS (sc-651) polyclonal antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA), and raised against peptides that map to the carboxyl or amino terminus of human RelA or human iNOS, respectively. The cross-reaction of these antibodies with rat NF-κB and iNOS was confirmed from the product data sheet.

Southwestern Histochemistry

Southwestern histochemistry was performed as described by Koji et al. (1994). Two complementary oligonucleic acids (ODN) containing the NF-κB binding sequence, the κB site (Sen and Baltimore, 1986), were synthesized: 5′-GGGACTTTCCC-3′, 3′-CCCTGAAAGGG-5′ (Amersham Pharmacia Biotech, Tokyo, Japan). For annealing to the double-stranded (ds) ODN, the acids were denatured for 10 min in 0.1 M Tris (pH 7.4) containing 0.01 M EDTA (TE), and gradually cooled at room temperature. The ds-ODN was labeled with digoxigenin (DIG) by means of a DIG Oligonucleotide Tailing Kit (Boehringer Mannheim, Mannheim, Germany).

The sections from the left TMJs of experimental (N = 10) and control (N = 4) rats were treated for 15 min with 0.3% H₂O₂ in methanol and immersed for 60 min in a pre-incubation buffer (5% non-fat dry milk and 50 mM NaCl in TE). They were then incubated overnight with DIG-labeled ds-ODN probes (5 μg/mL) dissolved in the pre-incubation buffer. After being treated for 60 min with 5% bovine serum albumin (BSA) and 500 μg/mL normal sheep IgG in PBS, they were incubated for 180 min with horseradish peroxidase-conjugated sheep anti-DIG IgG (1:800: Boehringer Mannheim) dissolved in PBS containing 5% BSA and 100 μg/mL salmon sperm DNA. They were visualized for 5 min with 0.05% diaminobenzidine, 0.025% CoCl₂, 0.025% NiSO₄(NH₄)SO₄, and 0.01% H₂O₂ in 0.1 M PB (pH 7.2), and some were counterstained with eosin. As Southwestern histochemistry controls, some sections were incubated with pre-incubation buffer alone instead of the probe, or treated with the probe in the presence of excess non-labeled ds-ODN (50-fold).

RESULTS

Histopathological Examination Scoring Synovitis in the Experimental and Control TMJs

After condylar hypermobility, synovitis-like changes were observed in the synovial membrane of the experimental TMJs (N = 20), as reported previously (Muto et al., 1998a,b). These changes included an increase in the thickness of the synovial lining (Figs. 1b, 1c). Multi-layered changes (grades 1 and 2) were recognized in 70-80% of the synovial membranes in each portion of the TMJ (Table ). Fibrous thickening and vascular dilation were also found in the subsynovial connective tissue. However, the synovial membranes of all control TMJs (N = 8) showed non-inflammatory changes (grade 0) (Figs. 1d, 1e ).

Localization of NF-κB and iNOS by Immunohistochemistry

Strong immunoreactivity for NF-κB was seen in the cytoplasm of the superficial cells of the hypertrophic synovium, but not in their nuclei (Fig. 2a ). Some subsynovial cells (probably fibroblasts) were positive for NF-κB. Also, in the control TMJs, NF-κB immunoreactivity was also recognized in the cytoplasm of the synovial lining cells, but not in their nuclei (Fig. 2c ). The immunoreactivity for iNOS in the experimental and control TMJs was also recognized in the same ways as seen in the above (Figs. 2b, 2d ). Both NF-κB and iNOS were observed in all four portions of the synovium in all experimental and control TMJs. The synovial cell layers with immunoreactivity for NF-κB or iNOS in the synovitis TMJs became thicker than those in the control TMJs (Figs, 2a-2d).

In the immunohistochemical controls, no immunoreactivity for NF-κB or iNOS was observed in either experimental or control TMJs (data not shown).

Localization of NF-κB by Southwestern Histochemistry

Using a novel Southwestern histochemistry method, we saw, in the experimental TMJs, intense positive reaction to the ds-ODN probes for NF-κB in the nuclei of superficial cells of the hypertrophic synovial membrane (Figs. 3a, 3c ), compared with the nuclei of the superficial synovial cells of control TMJs (Fig. 3b ). In the Southwestern histochemistry controls, no positive reaction was observed in sections treated with an excess (50-fold) of non-haptenized ds-ODN probes (Fig. 3d ) or no probe (Fig. 3e ). Similar Southwestern histochemical findings were observed in the other portions of the synovium in all experimental and control TMJs.

DISCUSSION

The NF-κB/Rel family has five subunits: c-Rel, NF-κB1 (p50/p105), NF-κB2 (p52/p100), RelA (p65), and RelB (Baldwin, 1996). Each subunit commonly contains the Rel homology domain that participates in DNA binding, dimerization, and NLS masking by IκB. In response to numerous stimuli, the phosphorylation of IκB leads to its ubiquination and subsequent degradation by the proteasome (Mercurio and Manning, 1999). Activated NF-κB is translocated from the cytosol into the nucleus. In our immunohistochemical analysis, NF-κB was not expressed in the nuclei of the synovial lining cells of inflammatory or control TMJs, but was expressed only in the cytoplasm. NF-κB proteins bind to the κB consensus sequence in the promoter region of target genes after translation into the nucleus (Baldwin, 1996). Our antibody to NF-κB recognizes the carboxyl terminus of RelA containing the transcription activation domains. It is thought that masking of the transcription activation domain of RelA protein by RelA binding to the κB site in the nucleus causes the lack of nuclear NF-κB immunostaining. The probe used for Southwestern histochemistry contains the specific κB sequence for the binding of classic NF-κB, p50/RelA. It can detect activated NF-κB (p50/RelA) that is free or not tightly bound to the sequence after it dissociates from I-κB. With Southwestern histochemistry, large amounts of NF-κB were expressed in the nuclei of synovial lining cells in induced-synovitis TMJs, indicating the nuclear translation of activated NF-κB after dissociation in the cytoplasm of the synovial cells. Therefore, the NF-κB (p50/RelA) detected by this new method is regarded as the activated form, and our results suggest that significant amounts of NF-κB are activated in synovial lining cells in TMJs inflamed by excess mechanical stress. In addition, in the Southwestern histochemistry for NF-κB, there was a clear difference in the NF-κB reaction intensity between the nuclei in the synovial cell layers of TMJs that were or were not subjected to mechanical stress (Figs. 3a, 3b). The synovial lining cell layers of all control TMJs were grade 0 (1-3 cell layers). NF-κB activated by excess mechanical stress may also be involved in the proliferation of the synovial lining cells of rat TMJs.

Immunohistochemically, p50 and RelA are found in synovial cells of the rheumatoid synovium (Handel et al., 1995; Yamasaki et al., 2001). Activated NF-κB has been demonstrated in the synovial membrane in induced-arthritis model rats (Tomita et al., 1999). Campbell et al. (2000) indicated that the p50/RelA heterodimer and p50/p50 homodimer are activated in the synovial cells of wild-type and c-rel^−/− arthritic joints, and suggested that p50 plays a significant role in the progression of arthritis. Our study is the first to report that activated NF-κB (p50/RelA) is expressed in the nuclei of synovial lining cells of the inflamed TMJ. Therefore, we consider NF-κB a potential transcription factor in the molecular mechanism for the development of TMJ inflammation induced by excess mechanical stress.

Inflammatory cytokines, such as IL-1 and TNF-α, promote expression of the iNOS gene via NF-κB activation (Xie et al., 1994; Taylor et al., 1998). The overproduction of NO generated by iNOS causes tissue damage during inflammation (Moncada et al., 1991; Mayer and Hemmens, 1997). Peroxynitrite formed by the reaction of NO with superoxide reacts with several biological molecules (van der Vliet et al., 1995; Dalton et al., 1999). Peroxynitrite is also markedly increased in many diseases (Greenacre and Ischiropoulos, 2001), and is thought to be a key mediator of NO-mediated tissue injury (Beckman and Koppenol, 1996). In addition, peroxynitrite production is markedly accelerated in RA patients and experimentally-induced-arthritis animals (Kaur and Halliwell, 1994; Szabo et al., 1998). We also demonstrated that nitrotyrosine, which is a peroxynitrite product (Greenacre and Ischiropoulos, 2001), is localized in synovial lining cells of the induced-synovitis TMJ in rats (unpublished data). NO production and iNOS expression are enhanced in inflammatory articular disease (Farrell et al., 1992; Sakurai et al., 1995), including TMJ disease (Takahashi et al., 1996; Homma et al., 2001). In this study, synovial cell layers with the iNOS immunoreactivity became thicker in the synovitis TMJ, although iNOS is also found in the synovium of the control rat TMJ (Masuda et al., 2002). This indicates that mechanical stress of the TMJ promotes iNOS expression via NF-κB activation. It also suggests that the NO produced by iNOS plays a significant role as a mediator in the progression of TMJ inflammation.

In conclusion, this study revealed NF-κB activation and intense iNOS expression in synovial lining cells in the rat TMJs with synovitis induced by condylar hypermovement. We postulate that the progression of TMJ disease is triggered by the promotion of NF-κB activation in response to excessive mechanical stress.

Table.

Scoring the Thickness of the Synovial Cell Layers in the Experimental Rat TMJs (N = 20)

	Number of Joints
Portion of synovium	A^a	B	C	D
^a A, B, C, and D in the Table correspond to the boxed letters in Fig. 1a . The change in thickness of the cell layers is graded: grade 0, 1-3 layers; grade 1, 4-6 layers; grade 2, 7 or more layers.
Grade 0	4	5	5	4
Grade 1	13	13	13	13
Grade 2	3	2	2	3

Figure 1.

A schematic illustration and light micrographs of the rat TMJ. (a) A schematic diagram of sagittal sections of the central portion of the rat TMJ. The rat TMJ is divided into upper and lower compartments (UC, LC) by an articular disc, and the synovial membrane lines these compartments. The boxes indicate the observed portions of each section. Boxed numbers correspond to the respective Fig. numbers, and boxed letters (A to D) correspond to the portions observed in the histopathological analysis. Light micrographs of 10-μm sagittal sections of the (b,d) anterior and (c,e) posterior portions in induced-synovitis and control rat TMJs. In the induced-synovitis TMJ (b,c), the synovium consisting of 7 cell layers or more (grade 2, G2) is seen in the anterior (b) and posterior (c) portions. Vascular dilation, fibrin deposition (asterisk), and fibrous thickening are also observed. C, mandibular condyle; V, blood vessel. In the control TMJ (d,e), 1-3 layers (grade 0, G0) face the upper and lower joint compartments. No inflammatory change is found in the control synovium. Hematoxylin and eosin staining (b-e). Original magnification: x400 (b-e). Bar = 25 μm (b-e).

Figure 2.

Light micrographs of serial sagittal sections of the anterior portion in induced-synovitis (a,b) and control (c,d) rat TMJs immunostained with either anti-NF-κB (a,c) or anti-iNOS (b,d) antibody. NF-κB immunoreactivity is increasingly expressed in the cytoplasm of the synovial lining cells of the 7 or more layers (grade 2, G2) in the induced-synovitis TMJ (a) compared with that of the 1 or 3 layers (grade 0, G0) in the control TMJ (c), but is not shown in the nuclei of the synovial cells in the experimental or control TMJs. C, mandibular condyle; LC, lower joint compartment; V, vein. The immunoreaction for iNOS is also increasingly seen in the synovial lining cells of the 1-3 layers in the experimental TMJ (b) compared with the control TMJ (d). Original magnification: x100 (a-d). Bar = 100 μm (a-d).