Rubella Virus and Chronic Joint Disease: Is There an Association?

Abstract

Synovial fluid samples and/or biopsies from 79 patients with various chronic inflammatory joint diseases or traumatic joint injury were tested for rubella virus (RV) in order to confirm or refute results from other studies that suggested RV as a cause of chronic inflammatory joint disease. Sixty-eight of the 72 patients tested had RV antibodies. RV RNA was detected by reverse transcription-PCR in the synovial fluid cells from two patients. RV was also isolated by cell culture from the synovial fluid of one of these two patients. This patient was a 42-year-old female with common variable immune deficiency and Mycoplasma hominis arthritis, while the other was a 68-year-old female with rheumatoid arthritis. While these results fail to confirm that RV is associated with chronic inflammatory joint disease, they suggest that RV may persist within a joint and be reactivated when cell-mediated immunity is suppressed.

Joint symptoms are a complication of naturally acquired rubella and have been reported to occur in more than 50% of adult females with rubella but are rare among children and adult males with rubella (20). Symptoms are usually transient and vary in severity from joint pain alone (arthralgia) to frank arthritis. Following vaccination, joint symptoms occur less frequently (in 8 to 40% of vaccinees) and are usually less severe and of shorter duration than those that occur following naturally acquired rubella, although there is some variation, depending on the vaccine strain used (3, 20). Rubella virus (RV) has been isolated from joint aspirates following natural infection and vaccination (reviewed in reference 1).

In view of the widespread use of rubella vaccines, reports that RV was associated with chronic inflammatory joint disease generated considerable public concern. The Institute of Medicine in the United States established an inquiry, which concluded that further well-designed studies were required to determine whether there was an association between rubella and chronic joint disease in adult women (14).

Most previous studies on patients with chronic joint diseases have examined peripheral blood mononuclear cells (PBMCs) for RV; to our knowledge, there have been few published studies in which samples from joints were examined (12, 24). We therefore tested synovial fluid (SF), SF cells (SFCs), and synovial biopsies for RV by using both a sensitive reverse transcription-nested PCR (RT-PCR) (5, 6) and a well-established RV isolation technique (4). SFs and SFCs from adults and children with various chronic inflammatory joint diseases were tested, together with synovial biopsies from patients with osteoarthritis and traumatic joint injury (TJI) to determine if RV was present in the synovia of RV-seropositive patients.

MATERIALS AND METHODS

Study population and specimens.

Seventy-nine patients were recruited from four rheumatology clinics in London and Manchester and an orthopedic day surgery unit in London. Patients were diagnosed as having rheumatoid arthritis (RA), seronegative spondyloarthropathy (SNA), juvenile chronic arthritis (JCA), osteoarthritis (OA), infective arthropathy, gout, unexplained monoarthropathies, and TJI. Specimens were collected from 79 patients, 23 of whom were females (Table 1). Approval was obtained from all relevant ethical committees. Informed consent was obtained from all patients or their parents or guardians.

TABLE 1.

Detection of RV in SF and/or synovial biopsies from 79 patients

Diagnosis	Age range (mean)a	No. of patients			No. of specimens			No. of patients RV positive/no. tested	No. of patients RV IgG positive/no. testede
		Mb	Fc	Total	SFd	SFC	Biopsyd
OA	29–80 (59.1)	18	4	22	16	9	12	0/22	21/21
RA	30–86 (60.9)	6	6	12	12	12	0	1/12f	10/10
SNAg	19–45 (29.7)	8	4	12	12	11	0	0/12h	11/11
JCAi	2–14 (9.8)	4	6	10	8	9	1	0/10h	8/9
Undifferentiated arthritis	29–58 (40)	3	1	4	4	3	1	0/4	3/3
Gout	42–70 (56.5)	2	0	2	2	2	0	0/2	1/1
Infective arthritis	42	0	1	1	1	1	0	1/1j	0/1
TJI	22–49 (30.1)	15	1	16	8	3	16	0/16	14/16
Total no. (%)		56	23	79	63	50	30	2/79 (2.5)	68/72 (94.4)

^aAge is given in years. 

^bM, male. 

^cF, female. 

^dSynovial biopsy. 

^eSerum samples were tested for RV IgG by enzyme immunoassay (4). 

^fRV RNA was detected in SFCs from one patient. 

^gDiagnosis includes: seronegative oligoarthropathies, psoriatic arthritis, and Reiter’s syndrome. 

^hIncludes one sample in which there was evidence of inhibition in the first-round PCR. 

ⁱDiagnosis includes poly- and pauciarticular JCA with and without HLA B27. 

^jRV RNA detected in SF and RV isolated from SF from a patient who had CVID and M. hominis arthritis. 

SF and serum samples were obtained from patients when they attended the clinics either as new patients or at follow-up of established rheumatological disease. Patients were investigated if they had chronic arthritis (symptoms for 3 months or longer) and suffered from effusion of one or more joints. Effusions were aspirated with the patient’s consent for the indications of pain and uncomfortable limitation of movement or to establish a diagnosis. Synovial biopsies were obtained from 30 patients undergoing diagnostic arthroscopy following trauma or unexplained synovitis. A synovial biopsy was the only specimen tested from 14 patients. A blood sample was obtained simultaneously from 72 of the 79 patients. Samples were transported to the laboratory at 4°C within 24 h of collection and processed immediately.

Processing of SF.

SFCs were isolated by centrifugation if a sufficient volume of SF was received. From 1 to 5 ml of SF was centrifuged at 400 × g, the cell pellet was washed once with maintenance medium (Eagle’s minimum essential medium, 2% fetal calf serum, 2 mM glutamine, 100-IU/ml penicillin, 100-μg/ml streptomycin), and the cells were resuspended in maintenance medium and counted.

Detection of RV. (i) RV isolation in cell culture.

SF and SFC samples (when sufficient amounts were available) were tested for RV by cell culture, but there was insufficient synovial tissue to be tested for RV by this method. SF (100 μl) and SFCs (10⁶) were inoculated into Vero cell cultures and adsorbed for 2 h at room temperature before addition of maintenance medium (4). All specimens were cultured for 7 to 10 days at 37°C and then passaged into further Vero cells. RV RNA was identified in these cell cultures (VE2) by RT-PCR as described below. We have demonstrated this to be at least 100 times more sensitive than an indirect immunofluorescence assay (7).

(ii) RNA extraction.

Total RNA was extracted by using RNAzol-B in accordance with the manufacturer’s instructions. Sixty microliters of SF, 10⁶ SFCs, or 100 μl of VE2 was mixed with 600 μl of RNAzol-B (400 μl for VE2) and incubated on ice for 5 min. Synovial biopsies (10 to 100 mg) were placed directly into 400 to 800 μl of RNAzol-B and homogenized thoroughly. Negative control extractions of equal volumes of fresh maintenance medium were performed in parallel with each batch of specimens. One-fifth volume (120 μl) of chloroform was added, mixed, incubated on ice for 5 min, and then centrifuged at 10,500 × g for 20 min. The aqueous phase was transferred to a new tube. Three microliters of linear acrylamide (25 mg/ml) and an equal volume of isopropanol were added, mixed, and placed at −20°C overnight to precipitate RNA. Samples were then centrifuged at 10,500 × g for 20 min, and the pellet was washed once with 75% (vol/vol) ethanol, vacuum dried, and stored at −70°C until tested. Prior to RT-PCR analysis, the pellets were dissolved in 22 μl of molecular biology grade water.

Detection of RV RNA by RT-PCR.

RV RNA was detected by RT-PCR that amplifies a region of the E1 open reading frame of the RV genome (6). Sterile water reagent blanks, high and low positive controls, and strict precautions to prevent contamination of PCR mixtures were employed (6). This method was shown to be specific for RV RNA and is sufficiently sensitive to detect 0.1 50% tissue culture-infective dose of RV or 14 to 20 genome equivalents. No loss of sensitivity was observed when titrations of RV diluted in SF were tested in parallel with dilutions in maintenance medium. In addition, RV RNA was detected in RV-spiked SF after incubation for 24 and 48 h at both 4°C and room temperature (7).

RNA control—detection of coxsackievirus RNA by RT-PCR.

To control for possible loss of RNA during sample processing or the presence of enzymatic inhibitors in SF, approximately 1 50% tissue culture-infective dose of coxsackievirus B3 (CVB3)-infected cell culture supernatant was added to 60 μl of SF prior to RNA extraction; the amount of CVB3 RNA was 10 to 100 times the detection limit of the enterovirus RT-PCR. The same sample was used simultaneously for RV and enterovirus RT-PCRs. CVB3 RNA was detected by using a well-established enterovirus RT-PCR assay (17). Twenty-seven SF samples were tested in this way.

Serological tests for RV-specific IgG and IgM.

To test for evidence of past or recent RV infection, serum samples were tested for RV-specific immunoglobulin G (IgG) antibodies by enzyme immunoassay (Rubek; Launch Diagnostics) and/or latex agglutination (Rubalex; Orion Diagnostica, Espoo, Finland) and for RV-specific IgM by M antibody capture radioimmunoassay (4).

RESULTS

Detection of RV by RT-PCR and virus isolation in SF, SFCs, and synovial biopsy samples.

Sixty-three SF samples from the 79 patients were analyzed by RT-PCR, and 57 were also tested by virus isolation (Table 1). RV was detected by both RT-PCR and virus isolation in one sample from a 42-year-old female patient with common variable immune deficiency (CVID) and Mycoplasma hominis arthritis. No RV was isolated from the other 56 synovial fluid samples (Table 1). There was evidence of inhibition in the first-round PCR of two samples (from one SNA and one JCA patient); and five samples (from three OA and two TJI patients) were considered to contain inhibitory components, as CVB3 RNA was not detected. RV RNA was not detected in these samples.

Forty-eight SFC samples were analyzed by RT-PCR, and 32 were analyzed by virus isolation. RV RNA was detected by RT-PCR in samples from two patients, including the patient with CVID and a second, 68-year-old woman with RA. RV was also isolated from the SFC of the patient with CVID but not from the other 31 samples.

RV RNA was not detected by RT-PCR in any of 30 synovial biopsy samples (Table 1). Biopsies were not obtained from the two patients in whom RV RNA was detected.

Detection of RV-specific IgG and IgM antibodies.

RV-specific IgG was detected in serum samples from 68 of 72 patients, indicating past infection (Table 1). RV-specific IgM was not detected in any of these patients; thus, none had evidence of recent or current infection. No detectable serum immunoglobulins were detected in the serum sample obtained from the patient with CVID.

DISCUSSION

We failed to show an association between rubella and chronic inflammatory joint disease when we examined specimens from 63 patients with various chronic joint conditions. In addition, there was no evidence of RV in 30 synovial biopsies. RV was detected, however, by both RT-PCR and virus isolation in the SF and SFC samples from a patient with CVID and infective arthritis. This 42-year-old female, who gave no history of rubella immunization, presented with a severe, erosive, symmetrical, initially undiagnosed arthritis with a 4-month duration. A diagnosis of CVID was made when it was demonstrated that the patient had no detectable serum immunoglobulins. M. hominis was subsequently cultured from her SF. The arthritis and her general condition improved following antibiotic therapy. She had received immunosuppressive therapy for concurrent severe cutaneous vasculitis prior to the drawing of the SF sample and the diagnosis of CVID. RV RNA was also detected in the SFCs from a 68-year-old RV-seropositive female with a 10-year history of rheumatoid arthritis and a history of bronchiectasis. However, RV was not isolated from either the SF or the SFCs, and we were unable to obtain a second specimen for confirmation. Rheumatoid factor was present at diagnosis but had since disappeared. She had secondary degenerative arthritis and an effusion in the knee from which no bacteria were cultured. She was not on any concurrent medication and had not received any immunosuppressive agents for 3 years prior to the study.

The number of patients with seronegative spondyloarthropathy in our study was relatively small. It was difficult to obtain samples from these patients, as joints are now infrequently aspirated due to improved treatment, seronegative arthritis is relatively less frequent in the populations studied, and effusions are often too small to aspirate. We tested SFs for evidence of inhibition in RT-PCR in the latter part of our study by adding CVB3 RNA. There was evidence of inhibition in a few specimens, and it is therefore theoretically possible that RV RNA remained undetected in such specimens. As RV is fairly stable in SF, loss of RV during transport of specimens at 4°C prior to testing is unlikely.

Our studies were initiated following the considerable public concern generated by data which suggested that persistent RV infection may play a role in the induction of chronic inflammatory joint disease. Isolation of RV from the PBMCs of 12 of 14 women with chronic joint symptoms with a duration of 1.5 to 7 years following naturally acquired rubella or vaccination was reported (8, 9, 21, 23). RV was also isolated from PBMCs or SF mononuclear cells from seven children with chronic rheumatic disease but not from controls (10). Sequencing of these RV genomes was not reported. Two subsequent studies have failed to confirm these results (13, 24), although Dougherty et al. (12) detected RV RNA in SFCs from 4 (3 with RA, 1 with psoriatic arthritis) of 56 patients with chronic inflammatory arthritis but in none of 27 patients with other arthropathies. RV RNA sequences were also detected in PBMCs from one of three patients with chronic arthropathy or persistent malaise following rubella vaccination but not in samples from 23 individuals with acute rubella vaccine-induced arthropathy (12). Two large controlled studies of rubella immunization of adult women also failed to show that rubella vaccination is associated with chronic joint symptoms (18, 19). A third study (22) showed a marginally significant association, but the results of this study of women immunized postpartum are surprising in that acute joint symptoms were reported in 20% of the women who received a placebo, as well as in 30% of the vaccine recipients.

Our results suggest that RV is not associated with chronic inflammatory joint disease but may occasionally persist within joints. Persistent RV infection may be established in synovial cell cultures (11, 15) and RV antibody titers in the SF of some patients with chronic arthritis are higher than those in serum, suggesting local synthesis of specific antibody (16). Persistence of RV is also suggested by our previous studies, which demonstrated that RV-specific IgM may persist for up to 4 years following both natural RV infection and vaccination with HPV77.DE5 (2). The presence of RV RNA in the SFCs from our two patients may be due to reactivation from cells within the joint as a result of severe immunodeficiency or declining immunity in old age.