MATERIALS AND METHODS

Tissue preparations

Synovial tissue specimens were obtained from RA, osteoarthritis (OA), and normal synovia from patients undergoing synovectomy, joint replacement, or trauma surgery. Samples were fixed in 4% neutral buffered formalin for 6–8 hours, dehydrated, and then embedded in paraffin using an automated tissue processor. All patients with RA fulfilled the American College of Rheumatology criteria for the diagnosis of RA.¹³ Ethical approval was granted by the local ethical committee and informed consent obtained from all patients.

Cell culture

Synovial fibroblasts were obtained by enzymatic digestion of synovial tissues from patients with RA, patients with OA, and normal synovial tissue as described.¹⁴ Briefly, the tissues were minced and incubated with 1.5% Dispase (Boehringer Mannheim, Rotkreuz, Switzerland) in phosphate buffered saline (PBS) for one hour at 37°C under continuous agitation. The cells were pelleted, resuspended, and cultured in Dulbecco’s modified Eagle’s medium F12 (DMEM/F12, Gibco Life Technologies, Basel, Switzerland) containing 10% fetal calf serum, glutamine (0.2 mM), HEPES buffer (10 mM), penicillin-streptomycin (50 IU/ml), and amphotericin B (2.5 μg/ml, all Gibco Life Technologies). Isolated synovial cells were indicated with the patient number and the number of passage. Foreskin fibroblasts were received from the department of dermatology of the University Hospital in Zürich in passage three. Cells were passaged when reaching confluence and transferred to chamber slides (Lab-Tek II, Nalge Nunc Int, Napierville, IL, USA) and grown for 24–96 hours before fixation. SV40 transformed RASF, as described by Parak et al,¹⁵ and the SV40 transformed bronchial epithelial cells (BEAS 2B), obtained from G Alink from Wageningen, The Netherlands, were used as positive controls. Epithelial cells were grown in serum-free keratinocyte growth medium (SFM; Gibco Life Technologies, Basel, Switzerland).

Immunohistochemistry for p53 on paraffin embedded tissues

Tissue specimens were cut into 3 μm sections, mounted on 5-aminopropyltriethoxysilane coated slides, and dried at 50°C for at least two hours. The sections were dewaxed in xylol and rehydrated in decreasing concentrations of ethanol, pretreated with microwave heating in citrate buffer (0.01 M, pH 6.0), and kept at 70°C for 30 minutes in a heat incubator for antigen retrieval.

All the following procedures were performed at room temperature. The slides were incubated with blocking solutions A and B (Vector-Laboratories, CA 94010, USA), both for 15 minutes. To decrease unspecific background reactions slides were “blocked” with 2% horse serum in 4% non-fat milk in Tris-HCl (0.1 M, pH 7.6) for 30 minutes. Primary anti-p53 antibodies (clone DO7, DAKO, Glostrup, Denmark) were diluted 1:100 (concentration 4 μg/ml) and incubated for 40 minutes on the slides. DO7 anti-p53 antibodies recognise an epitope at the N-terminus of the human p53 protein (aa 19–26). Antibodies react with both wild-type and mutant p53. Mouse isotype matched IgG sera in adapted concentrations served as negative controls. As secondary antibodies, goat antimouse antibodies labelled with biotin (Jackson Immuno Research, West Grove, PA, USA, diluted 1:200, concentration 5 μg/ml) were incubated for 30 minutes. Then, alkaline phosphatase conjugated streptavidin (DAKO, Glostrup, Denmark) was applied, also for 30 minutes, in a dilution of 1:50 (concentration 5.4 μg/ml). Finally, staining was developed with the BCIP/NBT solution (Boehringer Mannheim, Rotkreuz, Switzerland) for 20–30 minutes. Between all reaction steps extensive washing was performed with Tris-HCl (0.1 M, pH 7.6) and only before the application of the developing solution the slides were washed with Tris-HCl (0.1 M) at pH 9.7. Tissue sections from a p53 positive neurological tumour (glioblastoma multiforme) served as a positive control.

Immunoreactivity was evaluated using a previously described method (Lee et al,⁵ Sugiyama et al⁶). At least 10 surface areas at ×400 magnification were scored, and the percentage of positive cells calculated for each patient. On average, at least 200 cells could be detected for each high power field, and thus more than 2000 cells were analysed for each specimen. For the p53 immunostaining a score of 0 to IV was assigned: 0 = 0% positive cells; I = <1% positive cells; II = 1–5% positive cells; III = >5–20% positive cells; IV = >20% positive cells.

Immunofluorescence of p53

RASF in passages 2–12, OASF, and foreskin fibroblasts (FSFB) in passages 4–6 or normal (N) SF (passages 8 or 9) cultured on chamber slides (Labtec II, Nunc) were washed twice with PBS and then fixed with methanol/acetone (1:1) for 10 minutes at −20°C. Thereafter, the slides were dried at room temperature for 10 minutes and immediately processed or stored at −20°C. Rehydrated in Tris-HCl (0.1 M, pH 7.6), the slides were blocked with 2% horse serum and 4% non-fat milk in Tris-HCl for 30 minutes. After washing with PBS-Tween (0.05%), the slides were incubated with primary anti-p53 antibodies in a dilution of 1:1000 in PBS-Tween (DO7, concentration 0.4 μg/ml) for 30 minutes at room temperature. As negative controls mouse isotype matched IgGs were used. After several washing steps Cy3 conjugated sheep antimouse sera (DIANOVA, distributed by Milan Analytica Ag, La Roche, Switzerland) diluted 1:1000 (concentration 0.14 μg/ml) in PBS-Tween were applied as secondary antibodies for 30 minutes at room temperature. Rinsed in PBS, the slides were finally mounted in fluorescent mounting medium (DAKO) and analysed with a Zeiss immunofluorescence microscope at a wavelength of 515–545 nm. The immunostaining was evaluated by counting. At least 10 high power fields at ×400 magnification with a total number of 90–200 evaluated cells were scored.

Irradiation of cultured fibroblasts with x rays

RASF in passages 3 to 11, FSFB in passage 5, and NSF in passage 9 were grown on chamber slides, and irradiation (10 Gy) was carried out at room temperature with a Pantak Therapax 300 kV x ray unit at 0.7 Gy/min. After 0, 2.5, 5, or 20 hours cells were fixed as described above. Appropriate non-irradiated controls were used at 0 and 20 hours. Expression of p53 was determined by immunofluorescence, and the percentage of positive cells was evaluated by counting.

SCID mouse experiments

SCID mice (n = 28) were obtained from Charles Rivers GmbH (Sulzfeld, Germany) and kept permanently under sterile conditions. Implantation of RASF and NSF together with normal human cartilage was performed as described.¹⁴ Additionally in one SCID mouse SV40 transformed synoviocytes were engrafted to serve as positive control. Cells were trypsinised and dissolved in an inert collagen sponge (Gelfoam; Pharmacia, Upjohn, Dübendorf, Switzerland). Sponges were co-implanted with pieces of human articular cartilage of 1 mm³ size under the renal capsule of SCID mice. Mice were anaesthetised by intraperitoneal injection of Xylocain (lidocain hydrochloride, 0.014 mg/g of body weight; Astra Pharmaceutica, Dietikon, Switzerland) and Ketalar (ketamine hydrochloride, 0.09 mg/g of body weight; Parke-Davis, Baar, Switzerland) in an isotonic solution. An incision was made on the left flank and the left kidney was carefully exteriorised. A small incision of the renal capsule was performed with a forceps and sponge and cartilage were co-engrafted adjacent to the kidney. Peritoneum and skin were closed using 5–0 Prolene suture material. After 60 days, mice were killed, and the implants were fixed in 4% buffered formalin and embedded in paraffin according to standard procedures. Paraffin embedded sections were stained with haematoxylin and eosin, or analysed for the expression of p53 with anti-p53 antibodies as described above. The immunohistochemistry of the SCID mice sections was repeated twice. Invasion into the cartilage was quantified according to a semiquantitative score ranging from 0 to IV referring to the number of invading cells and the number of affected cartilage sites (0, no invasion; 0.5, invasion of 1–2 cell layers; I, invasion of 3–5 cell layers; I.5, invasion of 3–5 cell layers at three independent sites of the cartilage; II, invasion of 6–10 cell layers; II.5, invasion of 6–10 cell layers at three independent sites; III, invasion of >10 cell layers; III.5, invasion of >10 cell layers at two independent sites; IV, invasion of >10 cell layer at three or more sites of the cartilage). Expression of p53 was determined by counting the number of p53 positive invading cells in comparison with all invading cells.

Statistical analysis

For statistical analysis, the Mann-Whitney U test was used. Values of p<0.05 were considered to be significant. Data were expressed as the mean (SD).

RESULTS

Expression of p53 in untreated or irradiated RASF

The in vitro expression of p53 in untreated RASF was low in comparison with the SV40 transformed cells. In average, 3.7% of RASF expressed p53 (table 1) and this was slightly higher than in OASF, FSFB, and NSF, but did not reach statistical significance, even when the data of the control fibroblasts were pooled (p>0.05). p53 was transiently inducible by ionising irradiation. Immediately after irradiation, the expression of p53 in all cells was about 4% and this did not differ from non-irradiated controls fixed at 0 or 20 hours. Two and a half hours after irradiation the percentage of p53 expressing cells increased remarkably about 10–15 times, reaching a maximum of 33–54% (figs 1 and 2). After 20 hours, only 10–20% of the cells still expressed p53 in the nucleus. In the dose response and in the time course of the p53 induction after irradiation no major differences were detected between RASF, NSF, and FSFB cell cultures. Non-irradiated, SV40 transformed RASF and BEAS 2B cells, with a strong nuclear staining for p53, served as positive controls. Appropriate IgG controls were negative (not shown).

SCID mice experiments

In agreement with previous investigations,³ RASF from five different patients invaded the co-engrafted cartilage significantly stronger (p<0.01) than normal synovial fibroblasts (invasion scores 2.7–3.5 v 1.1–1.5, respectively). SCID mice sections from these experiments were stained with DO7 anti-p53 antibodies. NSF showed very limited invasion and negligible staining for p53. On the other hand, RASF showed a strong cartilage invasion and 20% of the cells invading the cartilage were found to express p53 in the nucleus (figs 3 and 4). Of note, the cells not invading the articular cartilage and located outside the invasive zones were completely negative for p53.

Immunohistochemistry of p53 in synovial tissues

The expression of p53 in synovial tissues from RA, OA, and normal synovia was low in comparison with the neurological tumour (glioblastoma multiforme) as determined by immunohistochemistry using DO7 anti-p53 antibodies. In RA tissues 0–7% of the cells stained positive for p53 with a mean of 1.9% (table 2, fig 5). 13/14 RA samples showed an expression of p53 in <5% of the cells. Positive signals were mainly located in the sublining. The expression of p53 in OA synovial tissues was not significantly different from that in RA synovial tissues. None of the normal synovia demonstrated any signal for p53, and this difference was significant for the expression of p53 in RA tissues (p = 0.032). Sections incubated with mouse control serum were negative.

DISCUSSION

This study was designed to determine whether expression of p53 in RASF could be associated with the invasive behaviour. Therefore, we investigated the functional p53 response in RASF grown in vitro and the induction of p53 in RASF engrafted into the SCID mouse co-implantation model of RA.

In vitro, the expression of p53 as determined by immunofluorescence in untreated RASF was <5%. This is in line with previous results⁴ reporting p53 staining in 10% for unstimulated fibroblast-like synoviocytes, while dermal fibroblasts showed a p53 expression in about 5% of the cells. As shown in figs 1 and 2, p53 was transiently inducible in RASF, NSF, and FSFB by x ray irradiation, which did not differ between all the types of fibroblasts examined. The transient induction suggests the presence of wild-type p53 in the majority of synovial fibroblasts. This is in line with studies showing either no p53 mutations or only in a small cell population of RASF.^9–11 In this context, it is interesting that Migita et al¹⁶ have demonstrated a functional p53 response in RASF by treatment with the proteasome inhibitor MG-132. This resulted in both an accumulation of p53 and a p21 mediated G1 arrest of the cell cycle.

Most interestingly, our study shows that in in vivo experiments using SCID mice p53 was induced in RASF at sites of cartilage invasion, although the vast majority of RASF were, in vitro, p53 negative. The number of p53 positive cells was 20%, but not all cells at sites of invasion stained positively for p53. Moreover, RASF located outside the invasive zones were p53 negative. An alternative explanation to the induction of p53 might be that cells expressing p53 have a selective advantage when invading the cartilage explant. In that case, cells expressing mutant p53 would potentially be more invasive. However, this appears not to be likely because the cells were grown in vitro over several passages and cells carrying functional relevant p53 mutations should become the predominant cell population, which was clearly not the case.

Our results are consistent with investigations in rat adjuvant arthritis¹⁷ as well as collagen induced arthritis.¹⁸ In these models the expression of p53 occurs in established arthritis, suggesting also that the expression of p53 is a consequence and not the cause of the inflammatory process. These results suggest that p53 induction is a normal consequence of inflammation. In addition, the present study indicates that p53 can be induced during the invasive process in RASF even in the absence of inflammation.

The finding that expression of p53 is a consequence or “adaptive” response to certain stimuli is consistent with the finding of wild-type p53 expression in various inflammatory diseases such as OA, reactive arthritis, inflammatory bowel disease, psoriasis, Helicobacter pylori infection, and many others.¹⁹ Just recently, Yamanishi and coworkers presented evidence for a functional suppressive role of p53 in collagen induced arthritis, because p53 −/− mice showed a more pronounced inflammation.¹⁸ The authors concluded that p53 is induced in the rheumatoid joint with protective functions such as cell cycle arrest, apoptosis, or suppression of matrix metalloproteinases. Similarly, we suggest that an adaptive p53 response occurs also in the non-inflammatory SCID mouse co-implantation model of RA.

In summary, we presented in vitro and in vivo evidence for a functional p53 response in RASF. These data suggest that p53 at sites of cartilage invasion could be induced during the destructive process driven by activated RASF.