Antibiotic prophylaxis and treatment of reactive arthritis. Lessons from an animal model

1238 ARTHRITIS & RHEUMATISM Vol. 39, No. 7, July 1996, pp 1238-1243 0 1996, American College of Rheurnatology

ANTIBIOTIC PROPHYLAXIS AND TREATMENT OF REACTIVE ARTHRITIS

Lessons from an Animal Model

YONG ZHANG, CHRISTEL GRIPENBERG-LERCHE, KARL-OVE SODERSTROM, AULI TOIVANEN, and PAAVO TOIVANEN

Objective. To study the effect of antibiotic prophylaxis and treatment of reactive arthritis (ReA), using an experimental model.

Methods. Yersinia enterocolitica 0:8, when injected intravenously into Lewis rats, causes a sterile arthritis closely resembling human ReA in 70% of the animals. Arthritis develops in 1-2 weeks; in some of the animals it remains chronic, and exacerbations occur. This model was applied to study the effect of a 7-day treatment with ciprofloxacin, using 2 different dosages (20 or 100 mg/kg/day) and 4 different schedules for initiation of treatment. The effects were evaluated by determining the daily arthritis score, the number of rats developing arthritis, and fecal excretion of Yersiniu. In addition, weight gain was monitored. At autopsy (35 or 60 days after inoculation with bacteria), samples were obtained for determination of Yersiniu-specific antibodies in the serum. At the same time, samples were collected from mesenteric lymph nodes, lung, spleen, and liver for bacterial cultures, and from the ankle joints for histologic evaluation. In a separate experiment, ciprofloxacin concentrations in samples from serum and mesenteric lymph nodes were analyzed by high performance liquid chromatography.

Results. A 7-day course with 100 mg/kg/day of ciprofloxacin, started on day 3 after bacterial inocula-

Supported by the Academy of Finland, the Sigrid Juselius Foundation, and Bayer AG.

tion, completely prevented the development of ReA and eliminated Yersinia during the 60-day experiment. If a dosage of 20 mglkglday was used, development of acute arthritis was prevented, but some of the animals had positive fecal cultures a t the end of experiment. If antibiotic treatment was started on day 5, the preventive effect was still observed, but was less pronounced. If the treatment was started at the peak of the development of arthritis, no effect on arthritis was observed.

Conclusion. These results indicate that if any effect of antibiotic treatment in Yersirzia-triggered ReA is to be expected, the treatment must be started early and given in sufficient dosage. However, antibiotic treatment has no effect on fully developed arthritis.

Reactive arthritis (ReA) usually develops 1-3 weeks after an infection that has occurred elsewhere in the body. The most common causes are enteric infections by Salmonella, Yersinia, and Campylobacter, as well as urogenital Chlamydia infections, which all lead to ReA predominantly in HLA-B27 positive individuals. Several other infections induce ReA equally in HLA- B27 positive and HLA-B27 negative persons. In inflamed synovial tissue, only inert bacterial structures are usually found, but all the available information indicates that the causative microbe persists somewhere in the body for long periods of time (1,2). Therefore, the question of antibiotic prophylaxis and treatment of ReA is of great importance.

In the treatment of streptococcal tonsillitis, the Yong Zhang, MD, Christel Gripenberg-Lerche, PhQ Karl-

Ove Soderstrom, MD, Auli Toivanen. MD, Paavo Toivanen, MD: Departments of Medical Microbiology, Medicine, and Pathology,

Address reprint r e q m t s to Yong Zhang, MD, Department of Medical Microbiology, Turku University, FIN-20520 Turku. Finland.

Submitted for publication September 25, 1995; accepted in

value of penicillin prophylaxis ;o prevent the develop-

cepted (3,4). For other forms of ReA, the issue of Of rheumatic fever and synovitis is generally

Turku University, Turku, Finland. antibiotic prophylaxis and treatment has remained un-

( 6 ) observed definite reduc- (5). Bardin et revised form March 5 , 1996. tion in the rate of recurrence of Chlamydia ReA in a

ANTIBIOTIC PROPHYLAXIS AND TREATMENT OF ReA 1239

Greenland Inuit population when the triggering infection was promptly treated. Long-term treatment of Chlamydia ReA with tetracyclines has also yielded promising results, even though it is unclear to what extent the effect is due to the antibiotic versus the antimetabolic effect (73). One report describes an outbreak of Salmonella infection among Swedish radiolo- gists, leading t o several cases of ReA. Some of the doctors had taken antibiotics at the time of the first symptoms, but no effect on the occurrence rate or severity of the arthritis could be demonstrated (9). Likewise, in chronic ReA, mostly caused by Yersinia, only marginal effects of a 3-month treatment with ciprofloxacin were observed (10). Regarding long-term treatment of acute Yersinia ReA, the situation is unclear (7), and at least 2 clinical trials are in progress.

It has become apparent that controlled human studies to investigate the issue of antibiotic prophylaxis and treatment of R e A are not easy to perform. We have addressed this by using a rat model of ReA. When injected intravenously into Lewis o r S H R rats, Yersinia enterocolitica 0:8 causes a synovitis, mostly in the hind limb ankles, that greatly resembles human ReA (11-15). Other serotypes of Y enterocolitica o r Yersinia pseudo- tuberculosis are not effective, nor is bacterial inoculation by a gastric route. Likewise, all other rat strains studied have been resistant t o arthritis development. Approxi- mately one-third of the Lewis and S H R rats show macroscopic diarrhea, and most of them have a few small abscesses o n the skin. Arthritis is reactive by nature, since live Yersinia or Yersinia D N A was observed in the synovial tissue of only 2 of 28 rats o n days 3-14 after bacterial inoculation; on days 21-35, synovial samples from all 29 rats studied were negative (13,15,16). Instead, bacterial antigens were demonstrated by immunofluorescence (15). In some of the animals the arthritis remains chronic up to 8 months, and exacerbations may occur. Ciprofloxacin, a broad-spectrum fluoro- quinolone, was the antibiotic for use in this investigation, due to its high degree of intracellular penetration, including penetration into polymorphonuclear cells and phagocytic monocytes (17,18).

MATERIALS AND METHODS

Experimental animals. Two hundred ten specific pathogen-free male Lewis rats weighing 225-275 gm, pur- chased from Harlan-Sprague-Dawley (Indianapolis, IN), were used. They were kept in Macrolon 111 cages under filter tops in laminar flow hoods, 2-3 animals in each cage, on autoclaved aspen wood bedding. Otherwise they were conventionally

Table 1. Effect of ciprofloxacin on the development of experimental reactive arthritis after intravenous inoculation with Yersiniu enreroco- lirica 0:s”

Days of ciprofloxacin treatment Treatment

P U P 3-9 5-1 1 10-16 13-19

Control 15/21 (72) 21/29 (72) 16/20 (80) 10/16 (62) Ciprofloxacin, 0/14 5/19 (26)T 8/13 (62) 11/16 (68)

20 rngikg/day Ciprofloxacin, 0/14 8/19 (42)$ 10/13 (77) 11/16 (68)

100 mg/kg/day

* The rats were injected with Yrrsinia on day 0. Thereafter, they were treated with ciprofloxacin for 7 consecutive days, starting on day 3, 5 , 10: or 13. Values are the number of rats developing arthritis/total in each group (%). i P = 0.002 versus controls. by Fisher’s exact test. $ P = 0.036 versus controls, by Fisher’s exact test.

maintained by giving free access to standard autoclaved diet and water.

Microbial inoculation. Strain 8081 of Y enterocolitica 0:8, originally a blood isolate from a patient with septicemia, was cultured and handled as described previously ( 1 4 3 ) . The in vitro minimum inhibitory concentration of ciprofloxacin for this strain is 0.012 pg/ml, and the minimum bactericidal concentration (MBC) is 0.1 pg/ml. A single injection of 2 X lo7 (1.97-2.56 X lo7 in different experiments) viable bacteria in 0.1-0.2 ml of physiologic saline was administered in a tail vein.

Antibiotic treatment. Ciprofloxacin powder (849-858 pg/mg) was supplied by Bayer AG (Wuppertal, Germany). It was dissolved and diluted to a concentration of 10 mg/ml or 25 mg/ml in physiologic saline. The drug was administered by subcutaneous injections 12 hours apart for 7 consecutive days. Two different dosages, 20 mg/kg/day or 100 mg/kg/day, and 4 different time schedules for initiation of treatment in relation to the bacterial inoculation were applied. Half of the control animals given the bacterial inoculation were treated with physiologic saline instead of ciprofloxacin, and the other half were left untreated. Since no difference between these 2 control groups was observed in terms of development of arthritis. they are reported as 1 group. The number of rats used in each group is shown in Table 1.

Antibiotic concentrations. In a separate experiment, ciprofloxacin concentrations in samples from serum and mesenteric lymph nodes were analyzed by high performance liquid chromatography (HPLC), with a detection limit of 10 ng/ml(19).

Clinical assessment. The rats were monitored individ- ually by 2 independent observers who assessed changes on the eyes, fur, nose, and scrotum, and potential development of skin abscesses and diarrhea. Body weight was also recorded. To monitor development of arthritis, each limb was assigned a score of 0-4, as follows: 0 = no detectable swelling; 1 = mild swelling; 2 = overt swelling, with contours of the joint maintained; 3 = swelling with obliteration of contours; 4 = gross distortion of the joint. A daily arthritis score for each rat was determined by adding the scores for the 4 limbs. All monitor- ing was performed daily for the 14 first days after the bacterial inoculation, and 3 times a week thereafter. Fecal samples from

1240 ZHANG ET AL

individual rats were cultured once a week for the presence of Yersinia, as described previously (14,15).

Autopsies. In each experiment, half of the rats were killed 35 days after the bacterial inoculation and the rest at 60 days. Samples from mesenteric lymph nodes, lung, spleen, and liver were obtained under aseptic conditions for bacterial cultures. Samples from the ankle joints were obtained for histologic evaluation of arthritis.

Antibody assays. IgM-, IgG-, and IgA-class antibodies against Y eriterocolitica were determined, by enzyme-linked immunosorbent assay (ELISA) (14,15), from serum samples obtained before the bacterial inoculation and 35 and 60 days thereafter.

Statistical analysis. Fisher’s exact test and the Mann- Whitney U test (I-tailed) were utilized for statistical comparisons.

RESULTS

When a 1-week course of ciprofloxacin was started on day 3 after the bacterial inoculation, i.e., before the appearance of any signs of arthritis, development of arthritis was completely prevented by the drug at both dosage levels. When ciprofloxacin treatment was started on day 5 , a significant reduction in the arthritis incidence and in mean arthritis score compared with controls was observed (Table 1 and Figure 1). If rats that showed no signs of arthritis on day 5 are considered separately, only 4 of 29 with ciprofloxacin treatment started on day 5 subsequently developed arthritis, without any difference between the 2 ciprofloxacin dosages. In contrast, some of the animals treated with ciprofloxacin at 20 mg/kg/day showed exacerbation of arthritis 4 weeks after the bacterial inoculation (Figure 1).

When ciprofloxacin treatment was started on day 10, no clear effect on the arthritis incidence was observed. However, the severity of arthritis was significantly reduced in the animals treated with 100 mg/kg/day (P < 0.05 compared with controls at 14 and 21 days; Figure 1). Finally, when the ciprofloxacin treatment was started on day 13, at the peak of the arthritis development, no effect of ciprofloxacin at either dosage was observed.

Yersinia inoculation initially caused a decrease of body weight for a few days. Thereafter, body weight started to increase; this began earlier in the rats with ciprofloxacin treatment started on day 3 or 5 than in the controls or in those with the treatment started later. Histologic examination of synovial samples obtained at 35 days and 60 days showed proliferation of synovial lining cells in the arthritic rats, with mononuclear cell infiltrates within the synovia. Staining with hematoxylin and eosin did not demonstrate any bacterial structures, and no information additional to that obtained by mac-

4 0

30

20

1 0

0 0

4 0

3 0

2 0

1 0

00

0 10 20 30 40 50 60

Days after yersinia inoculation

Figure 1. Effect of a 7-day course of ciprofloxacin on the development of experimental Yersinia reactive arthritis. Two different dosages and 4 different schedules for treatment initiation (shaded areas) were used. Arthritis onset usually occurred 6-7 days (range 4-13 days) after intravenous inoculation with Yersinia. The antibiotic treatment had no effect on the time of onset of arthritis, only on the severity.

roscopic evaluation was revealed. Synovial samples were not cultured for Yerszniu, since earlier studies, carried out under the same conditions and in the same animal quarters as the present study, have revealed living bacteria in synovia extremely rarely, and even then only soon after the bacterial inoculation (13,15,16), indicating the reactive nature of the arthritis.


m 30 40 M 60

Days after yersinia inoculation

Figure 2. Effect of a 7-day course of ciprofloxacin on fecal excretion of Yerstnta enterocoliticu 0 :8 . Two different dosages and 4 different schedules for treatment initiation (shaded areas) were used.

Both dosage levels of ciprofloxacin had a significant effect on the fecal excretion of Yersinia with all the treatment schedules used, initially reducing the number of the animals with positive fecal cultures to 0 (Figure 2). However, the only group in which the fecal cultures remained negative during the entire 60-day experiment was that in which 100 mg/kg/day of ciprofloxacin was started on day 3. When ciprofloxacin was given at 20 mg/kg/day for 7 days, even if treatment was begun on day 3, return of positive fecal cultures during the latter part of the study was observed in almost one-fifth of the animals. Bacterial cultures from the tissues (lymph nodes, lung, spleen, liver) at the time of autopsy gave positive results in only a few of the samples on day 35; again, the 100 mg/kg/day-treated animals were the only groups in which all cultures were negative (Table 2). Both dosages of ciprofloxacin also induced significant prevention of diarrhea and skin abscesses if the treatment was started on day 3 or day 5 , but not if it was started later. However, occurrence of diarrhea or skin abscesses was not a prerequisite for arthritis development. Antibiotic treatment as such had no clear effect on the development of Yersinia-specific antibodies, determined by ELISA on serum samples obtained on days 35 and 60. However, the arthritic rats had significantly lower IgG and IgA responses than those without arthritis (with P values <0.01 or <0.05 in different groups; data not shown).

The ciprofloxacin concentration was determined by HPLC in sera and mesenteric lymph nodes of rats treated subcutaneously with 20 mg/kg/day given in 2 daily doses for 6Y2 days. The animals were killed 1, 4, 8, or 24 hours after the last dose. Maximal concentrations of 0.69-0.81 pg/ml were measured in the serum at 1 hour, and concentrations of 0.03-0.21 pg/ml at 24 hours after the last drug injection. For lymph nodes, the corresponding concentrations were 2.8-3.1 pg/gm and

Table 2. Effect of a 7-day course of ciprofloxacin on detection of Yersinia in different organs, by culture" ~

Day 357 Day 60t

Positive cultures Positive cultures

Treatment Lymph Total no. Lymph Total no. group node Lung Spleen Liver studied node Lung Spleen Liver studied

Control 0 0 2 5 49 0 1 1 1 37 Ciprofloxacin, 0 2 2 2 36 0 0 0 0 20

Ciprofloxacin, 0 0 0 0 36 0 0 0 0 20 20 mg/kg/day

100 mg/kg/day

* Values are the number of rats with positive culture results. Results from all experiments are combined. i After intravenous inoculation of Yersinia enterocolirica 0 : 8 .

1242 ZHANG ET AL

0.21-0.32 pg/gm, all exceeding the MBC required for the strain used (0.1 pg/ml).

DISCUSSION

Two major conclusions seem apparent on the basis of the results of the present study. First, antibiotic treatment was effective in preventing the development of ReA only when given early enough, i.e., starting on day 3 after bacterial inoculation. Second, only high doses of the drug (100 mg/kg/day of ciprofloxacin) were capable of completely eliminating Yersinia during the 60-day experiment, and only when the treatment was started on day 3. Antibiotic treatment started 2 days later also showed a preventive effect, but 4 of 29 rats which were nonarthritic on day 5 did subsequently develop arthritis. In addition, in the group with this treatment schedule arid the lower dosage (20 mg/kg/day), exacerbation of the arthritis was observed in some of the animals.

At the end of the 60-day experimental period, all bacterial cultures from the mesenteric lymph nodes, lung, spleen, and liver of the rats treated with ciprofloxacin were sterile. However, -15% of the animals treated with the lower dose still continued to excrete Yersinia in the feces. This was at the end of the experiments, when most of the animals appeared healthy, with no signs of arthritis; occurrence of mild arthritic symptoms did not correlate with positive fecal cultures. All this information suggests that yersiniae hide somewhere in the body for long periods of time after the arthritic symptoms have subsided. We have previously presumed that the site of bacterial persistence would be either intestinal mucosa or mesenteric lymph nodes. The present findings provide evidence against the latter possibility and rather suggest that the microbes stay in the depths of intestinal mucosa, occasionally being se- creted in the feces. This is also the clinical experience from patients with Yersinia infection, with or without ReA (20,21). The occult persistence of Yersinia is also exemplified by several cases of Yersinia septicemia fol- lowing transfusion of stored blood obtained from healthy individuals (22).

Before definite recommendations for antibiotic treatment and prophylaxis of human ReA can be made, further studies are needed. However, on the basis of the presently available information, a few lessons are evi- dent. Regarding Chlamydia trachomatis, antibiotic treatment of the initial infection as well as of the subsequent arthritis is apparently beneficial (6-8), whereas even a 3-month course with ciprofloxacin had no clear effect on chronic Yersinia ReA (10). This difference might be

understood, at least to some extent, on the basis of different pathophysiologic mechanisms of joint inflam- mation associated with Chlamydia and Yersinia. The presence of bacterial antigens of both organisms has been demonstrated in the inflamed joint, whereas DNA of only chlamydia has been shown to be present (for review, see ref. 23). All efforts to demonstrate the presence of Yersinia DNA in inflamed human joint tissue have yielded negative results (24,25). These negative findings do not, however, exclude the possibility of the presence of live Yersinia elsewhere in the body.

The present findings demonstrate that antibiotic treatment in Yersinia-triggered ReA must be started early and given in sufficient dosage if there is to be any effect. However, antibiotic treatment of fully developed arthritis is not effective. It is also noteworthy that the dosage of ciprofloxacin that prevented the development of experimental arthritis was significantly higher than the dosages of 10-30 mg/kg/day that are generally used for humans.

1.

2.

3.

4.

5.

6.

7.

8.

REFERENCES

Toivanen A: Reactive arthritis. In, Rheumatology. Edited by J H Klippel, PA Dieppe. London, Mosby, 1993 Toivanen A, Toivanen P: Epidemiologic aspects, clinical features, and management of ankylosing spondylitis and reactive arthritis. Curr Opin Rheumatol 6:354-359, 1994 Mason T, Fisher M, Kujala G: Acute rheumatic fever in West Virginia: not just a disease of children. Arch Intern Med 151:133- 136, 1991 Lue HC, Wu MH, Wang JK, Wu FF, Wu YN: Long-term outcome of patients with rheumatic fever receiving benzathine penicillin G prophylaxis every three weeks versus eveiy four weeks. J Pediatr 125312-816, 1994 Leirisalo-Rep0 M: Treatment of reactive arthritis. Rheumatol Europe 24:20-22, 1995 Bardin T, Enel C, Cornelis F, Salski C, Jorgensen C, Ward R, Lathrop GM: Antibiotic treatment of venereal disease and Reit- er’s syndrome in a Greenland population. Arthritis Rheum 35: 190-194, 1992 Lauhio A, Leirisalo-Rep0 M, Lahdevirta J, Saikku P, Rep0 H: Double-blind, placebo-controlled study of three-month treatment with lymecycline in reactive arthritis, with special reference to Chlamydia arthritis. Arthritis Rheum 34:6-14, 1991 Lauhio A, Sorsa T, Lindy 0, Suomalainen K, Saari H, Golub LM, Konttinen YT: The anticollagenolytic potential of lymecycline in the long-term treatment of reactive arthritis. Arthritis Rheum 35:195-198. 1992

9. Locht H, khlstrom E, Lindstrom FD: Reactive arthritis after Salmonella among medical doctors: study of an outbreak. J Rheumatol 20:845-848, 1993

10. Toivanen A, Yli-Kerttula T, Luukkainen R, Merilahti-Palo R, Granfors K, Seppala J: Effect of antimicrobial treatment on chronic reactive arthritis. Clin Exp Rheumatol 11:301-307, 1993

11. Toivanen A, Merilahti-Palo R, Gripenberg C, Lahesmaa-Rantala R, Soderstrom KO, Jaakkola UM: Yersinia-associated arthritis in the rat: experimental model for human reactive arthritis? Acta Pathol Microbiol Imrnunol Scand (C) 94:261-269, 1986

12. Hill JL, Yu DTY: Development of an experimental animal model


13.

14.

15.

16.

17.

18.

19.

for reactive arthritis by Yersinia enterocolitica infection. Infect Immun 55:721-726, 1987 Merilahti-Palo R, Gripenberg-Lerche C, Soderstrom KO, Toi- vanen P: Long term follow up of SHR rats with experimental yersinia associated arthritis. Ann Rheum Dis 51:91-96, 1992 Gripenberg-Lerche C, Toivanen P: Yersinia associated arthritis in SHR rats: effect of the microbial status of the host. Ann Rheum Dis 52:223-228, 1993 Gripenberg-Lerche C, Skurnlk M, Zhang L, Soderstrom K-0, Toivanen P: Role of YadA in arthritogenicity of Yersinia enterocolitica serotype 0:s : experimental studies with rats. Infect Immun

Gripenberg-Lrrche C, Skurnik M, Toivanen P: Role of YadA- mediated collagen binding in arthritogenicity of Yersinia enterocolitica serotype 0:8: experimental studies with rats. Infect Immun

Easmon CSF, Crane JP: Uptake of ciprofloxacin by macrophages. J Clin Pathol 38:442-444, 1985 Easmon CSF, Crane JP, Blowers A: Effect of,ciproHoxacin on intracellular organisms: in-vitro and in-vivo studies; J Antimicrob Chemother 18 (suppl D):43-48, 1986 Scholl H. Schmidt K, Weber B: Sensitive and selective determination of picogram amounts of ciproHoxacin and its metabolites in biological samples using high-performance liquid chromatography

625568-5575, 1994

63:3222-3226, 1995

and photothennal post-column derivatization. J Chromatogr 416:

20. Toivanen A, Lahesmaa-Rantala R, Vuento R, Granfors K: Asso- ciation of persisting IgA response with Yersinia-triggered reactive arthritis: a study on 104 patients. Ann Rheum Dis 46:898-901, 1987

21. De Koning J, Heesemann J, Hoogkamp-Korstanje JAA, Festen JJM, Houtman PM, van Oijen PLM: Yersinia in intestinal biopsy specimens from patients with seronegative spondyloarthropathy: correlation with specific serum IgA antibodies. J Infect Dis

22. Wagner SJ, Friedman LI, Dodd RY: Transfusion-associated bacterial sepsis. Clin Microbiol Rev 7:290-302, 1994

23. Toivanen A, Toivanen P: Aetiopathogenesis of reactive arthritis. Rheumatol Europe 245-8, 1995

24. Viitanen AM, Arstila TP, Lahesmaa R, Granfors K, Skurnik M, Toivanen P: Application of the polymerase chain reaction and imrnunoHuorescence techniques to the detection of bacteria in Yersinia-triggered reactive arthritis. Arthritis Rheum 34:89-96, 1991

25. Nikkari S, Merilahti-Palo R, Saario R, Soderstrom K-0, Granfors K, Skumik M, Toivanen P: Yersinia-triggered reactive arthritis: use of polymerase chain reaction and immunocytochemical staining in the detection of bacterial components from synovial specimens. Arthritis Rheum 35:682-687, 1992

321-330. 1987

159:109-112, 1989

Antibiotic prophylaxis and treatment of reactive arthritis. Lessons from an animal model

Documents

Transcript of Antibiotic prophylaxis and treatment of reactive arthritis. Lessons from an animal model