250

IN VITRO T LYMPHOCYTE PROLIFERATIVE RESPONSE TO YERSINIA ENTEROCOLITICA IN

REITER’S SYNDROME

Lack of Response in Other HLA-B27 Positive Individuals

MICHAEL B. BRENNER, SHIGETO KOBAYASHI, CRAIG W. WIESENHUTTER, ARTHUR K. HUBERMAN. PETER BALES. and DAVID T. Y. YU

We assessed the in vitro T lymphocyte tritiated thymidine (3HTdr) incorporation response of Reiter’s patients in the United States to a serotype 3 strain of Yersink enterocofiticu. The mean ’HTdr incorporation response to the formalin killed form of this strain was 27,409 f 5,488 counts per minute for 14 HLA-B27 positive Reiter’s patients compared with 5,414 k 3,490 cpm for a control group of 11 HLA-B27 positive normal individuals (P < 0.0005). This high response in Reiter’s patients was observed with the formalin killed form of Y enterocofiticu serotype 3, but not with a heat killed form or a rough mutant form derived from the same bacterial strain. Further, Y enterocofitica of a different serotype (serotype 8) which is not associated with reactive arthri- tis failed to induce the high proliferative response ob- served with the serotype 3 strain. This response indi- cates that T lymphocytes from spontaneous Reiter’s patients are capable of recognizing and proliferating to determinant@) on the formalin killed form of Y entero- cofiticu serotype 3. Since this bacterium is associated with Reiter’s syndrome in Europe but not the United States, these data are consistent with the possibility that our patients have previously encountered these or simi-

~ _ _ _ __ From the Department of Medicine, UCLA School of Medi-

cine and Veterans Administration, Wadsworth Medical Center, Los Angeles, California.

Supported by the Nora Eccles Treadwell Foundation and USPHS grant AM28536-02. D. T. Y . Yu is a Senior Investigator of the Arthritis Foundation; M. B. Brenner is supported by the Veterans Administration Research and Career Development Award; C. W. Wiesenhutter is supported by the C & F Training Grant AI- 07126.

Michael B. Brenner, MD; Shigeto Kobayashi. MD; Craig W. Wiesenhutter, MD; Arthur K. Hubeman, MD; Peter Bales, MD; David T. Y . Yu, MD.

Submitted for publication April II, 1983; accepted in re- vised form September 30, 1983.

lar determinants through unrecognized infection with other microorganisms.

The etiology of Reiter’s syndrome remains un- known in the vast majority of spontaneous cases. However, 2 fundamental observations have contribut- ed to our understanding of the disease. First, a subset of Reiter’s syndrome, sometimes termed reactive ar- thritis, has been linked on epidemiologic evidence with antecedent bacterial infection. Salmonella typhimur- ium and Shigella flexneri have been associated with a number of cases (1,2), while Reiter’s-like reactive arthritis following infection with Yersinia enterocoli- tica has been frequently reported in Finland (3-5).

Second, both patients with spontaneous Rei- ter’s syndrome and those with reactive arthritis associ- ated with defined antecedent infection share a high incidence of the HLA-B27 histocompatibility allo- type. The association has been reported varying from 63-96% for Reiter’s syndrome and up to 90% for post- Yersinia reactive arthritis (6-9).

These two observations have led to the sugges- tion that, at least for those cases of reactive arthritis which follow identified infections, a combination of predisposing host genetic makeup along with an appro- priate environmental trigger result in disease (6,7,10). Here we report that patients with spontaneous Reiter’s syndrome have a heightened in vitro T lymphocyte proliferative response to one of the bacteria associated with reactive arthritis, Yersinia enterocolitica sero- type 3 (Ye3).

MATERIALS AND METHODS Human subjects. Patients were seen from July 1980

through June 1982 in the Rheumatology Clinics at the UCLA

Arthritis and Rheumatism, Vol. 27, No. 3 (March 1984)

T CELLS IN REITER’S SYNDROME 25 1

Center for the Health Sciences and the Veterans Administra- tion Wadsworth Medical Center, Los Angeles, CA. Thirteen patients had well characterized Reiter’s syndrome with all 3 of the classic features including arthritis, conjunctivitis, and urethritis ( 1 1). Another 5 Reiter’s patients had typical arthri- tis and either conjunctivitis or urethritis. The initial arthritis episode was historically preceded by symptoms of infection of the gastrointestinal tract in only 3 of the 18 Reiter’s patients, but no pathogenic organisms had been isolated. Thirteen patients had definite ankylosing spondylitis (12). Controls were normal healthy individuals without symptoms of arthritis. HLA-B27 positive normal subjects did not have pelvic radiographs to rule out the presence of occult sacro- iliitis.

Typing for HLA-A, B, C, and DR allotypes was performed in the laboratory of Dr. Paul I. Terasaki using standard procedures (13). All but 4 Reiter’s patients and all ankylosing spondylitis patients were HLA-B27 positive. No particular HLA-A, C, or DR allele was predominant in the Reiter’s patients.

Source of bacteria. Yersinia enterocolitica serotype 3, biotype 4, phage type VIII (abbreviated Ye3) and serotype 8. biotype 1, phage type X, (abbreviated Ye81 were kindly provided by Professor Mollaret, Pasteur Institute, Pans, France.

Preparation of bacteria. Bacteria were grown in Trypticase Soy Broth (TSB) (BBL Microbiology Systems, Cockeysville, MD) to mid-logarithmic phase at room tem- perature (23-25°C) or at 37°C.

Formalin killed bacteria were prepared by adding 37% formaldehyde (Mallinckrodt, St. Louis, MO) to broth cultures to a final concentration of 0.3% (14). After 48 hours of incubation, the bacteria were washed 5 times with phos- phate buffered saline (PBS).

Heat killed bacteria were prepared by autoclaving washed bacteria at 121°C for 45 minutes at 115 psi of pressure (14). Unless otherwise stated, all bacteria used to inoculate TSB media were identified as smooth colonies when viewed on MacKonkey agar plates under the dissect- ing microscope.

Rough mutant colonies were derived after 10 days’ growth at room temperature in nonreplenished TSB media. These rough mutant colonies were identified under the dissecting microscope, subcultured, and found to retain their rough morphology after growth in broth media (15). They were subsequently prepared by formalin killing as described above.

Isolation of peripheral blood mononuclear cells and their fractions. Mononuclear cells were isolated from hepa- rinized samples of peripheral venous blood by centrifugation on Ficoll-Hypaque gradients (16). The T-enriched lympho- cyte fractions of peripheral blood mononuclear cells (PBMC) were obtained by pelleting the PBMC with sheep erythro- cytes which had been treated with 2-aminoethylisothio- uronium bromide hydrobromide (Sigma, St. Louis, MO) (17). The non-T cells were those at the interface of the Ficoll-Hypaque gradient.

Characterization of leukocyte fractions. T lympho- cytes were enumerated by sheep erythrocyte rosettes and also by indirect immunofluorescent staining using a mono- clonal mouse anti-human pan-1 antibody (T101) as the first

antibody, and fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG (Antibodies Inc., Davis, CA) as the second antibody. B lymphocytes were enumerated by direct immunofluorescent staining with FITC-conjugated goat anti- human IgM antibodies (Tago, Burlingame, CA). Monoclonal antibodies Leu-2a and Leu-3a (Becton Dickinson, Mountain View, CA) were utilized to phenotypically characterize T lymphocytes. Monocytes were enumerated by a peroxidase staining procedure as previously described (18).

Using these procedures, the T-enriched fraction was shown to contain 93-97% T cells, 1 4 % B cells, and 5-16% monocytes. Non-‘I‘ fractions contained 6-16% T cells, 30- 40% B cells, and 30-50% monocytes.

Leukocyte viability was assessed by dye exclusion using 0.2% trypan blue.

Lymphocyte cultures. Cells were cultured in triplicate samples in round-bottom 96-well microculture plates (Titre- tek, Linbro, Hamden, CT) at 200 pI final volume in a 37°C 5% CO? incubator using RPMI 1640 media supplemented with 25 mM HEPES, 100 units/ml penicillin, 100 pg/ml streptomycin, 10 mM L-glutamine, and 10% heat-inactivated pooled human serum. These reagents were purchased from Gibco (Grand Island, NY). Eighteen hours before termina- tion of culture, l pCi of methyl 3H-thymidine (3HTdr) (5 mCilmM, New England Nuclear, Boston, MA) was added to each well. Cultures were harvested onto glass fiber filters (Titertek, Flow Lab. Inglewood, CA) with an automatic cell harvester (Skatron, Lierbyen, Norway). Scintillation count- ing was performed in Biofluor (New England Nuclear).

Bacteria were added to lymphocyte cultures such that the final concentration of bacteria was 9 x lo7 per ml. This was found in preliminary experiments to be the optimal concentration.

Assay of anti-Yersinia antibody by the bacterial agglu- tination test. Assay of anti- Yersinia antibody present in serum samples from patients and normal subjects followed standard bacterial agglutination procedures for heat or formalin killed Yersinia (19). A serum sample from a rabbit immunized with Ye3 served as positive control. No detect- able agglutination titers against Ye3 were found in any of the subjects studied.

Calculation of results. Results of 3HTdr incorpo- ration were expressed as counts per minute of radioactivity. The result of each triplicate was computed into mean and standard error. All standard errors were less than 20% of the means. The mean cpm of all unstimulated cultures (back- ground) were <4,000 cpm, and in each case, were subtracted from the mean cpm of the stimulated cultures to derive the “net” cpm. Statistical significance was assessed by the one- sided t-test assuming unequal variance.

RESULTS Lymphocyte culture conditions. The optimal cul-

ture conditions for lymphocyte proliferation to killed Yersinia described in Materials and Methods were determined by preliminary experiments. Dose re- sponse curves, such as those shown for 3 subjects in Figure 1, revealed that optimal 3HTdr incorporation

252

1 5 0

1

x a 100

n 0 - x - Y Y U t 0 3

L

c I-

n= 50 + Y z

S U B J E C T A

9x1,)’ y X l u 4 9 x 1 0 ~ 9 x 1 0 ~ 9x10’ Bacter id /ml

Figure 1 . Dose response curves. Formalin killed Yersinia enteroco- lirica serotype 3 varying from 9 x 10’ to 9 x lo7 bacteridml were added to the culture wells containing 2 x lo5 T-enriched lympho- cytes. Net tritiated thymidine (3HTdr) incorporation from the sixth day of culture is shown.

for any given subject occurred with the addition of killed Ye3 to a final concentration of 9 x lo7 bacte- ridml in each culture well. Higher concentrations produced either inhibition of the response or no addi- tional response (data not shown).

Lymphocytes were harvested for assay of 3HTdr incorporation from days 2-9 of culture. This is shown in representative experiments in Figure 2. Figure 2A depicts the response for an HLA-B27 positive normal subject while Figure 2B is the re- sponse for an HLA-B27 positive Reiter’s patient. The optimal response in all subjects examined occurred on day 6, 7, or 8. Another point observed in this instance was that the Reiter’s patient (Figure 2B) could be distinguished from the normal subject (Figure 2A) by being a responder rather than a nonresponder to stimulation with formalin killed Ye3. However, non- discriminating responses occurred in both subjects with heat killed Ye3.

For purposes of comparing responses among subjects, data were analyzed separately from the

BRENNER ET AL

sixth, seventh, and eighth days of culture. Data from the sixth day of culture are presented in all subsequent figures (Figures 3-6). When data from the seventh or eighth day of culture were analyzed, the comparative differences in responses were similar to those present- ed from the sixth day of culture.

High response of T lymphocytes from Reiter’s patients to formalin killed Ye3. Experiments using formalin killed Ye3 were performed to determine if the responses of Reiter’s patients were different from other subject groups, as was suggested in Figure 2. Since the majority of the Reiter’s patients were HLA- B27 positive, controls of HLA-B27 positive and HLA-B27 negative normal subjects were analyzed separately. A third study group consisted of HLA-B27 positive patients with ankylosing spondylitis. This group was useful for comparison since these patients with a clinically related seronegative spondylarthro- pathy were all taking nonsteroidal antiinflammatory drugs in doses similar to the Reiter’s patients.

We found that the T lymphocyte response to formalin killed Ye3 in Reiter’s patients (mean 27,409 * 5,488 cpm) was significantly higher than that for HLA-

150

100

50

2 3 4 5 6 7 8 9

D A Y O F C U L T U R E

Figure 2. Time course of lymphocyte response. T-enriched lym- phocytes at 2 X lo5 cells/well were cultured with formalin killed (-) or heat killed (. . . . . .) Yersinia enterocolitica serotype 3 at 9 x lo7 bacteridml. Net tritiated thymidine (’HTdr) incorporation from days 2 to 9 of culture is shown. A. HLA-B27 positive normal subject. B, HLA-B27 positive Reiter’s patient.

T CELLS IN REITER’S SYNDROME 253

a

t

a a

In these experiments, Yersinia enterocolitica used for lymphocyte stimulation were grown at room temperature. Room temperature was selected because it is the temperature optimum for growth of this bacterium in broth media (20). However, since Yer- sinia in vivo would grow at body temperature, we also tested the response of lymphocytes to formalin killed Ye3 which had been grown at 37°C. The mean re- sponse for 4 HLA-B27 positive Reiter’s patients was 22,239 ? 8,981 cpm. This was comparable with 24,617 2 10,689 cpm observed with Ye3 grown at room temperature. The mean response in 4 HLA-B27 posi-

3,409 cpm for Ye3 grown at 37°C and room tempera- ture, respectively. The mean response for 7 HLA-B27 positive ankylosing spondylitis patients was 3,805 2 3,296 and 5,358 ? 3,982 cpm for Ye3 grown at 37°C and room temperature, respectively. Thus, the higher mean responses observed for T lymphocytes of Rei- ter’s patients were apparent whether the bacteria used

37°C or room temperature.

a tive normal subjects was 3,959 ? 2,307 and 9,301 2

a to stimulate the lymphocyte cultures were grown at 0

a - - To be certain that the difference in responses a a

H L A - B 2 7 ( + ) H L A - B 2 7 ( + ) H L A - B 2 7 ( - ) H L A - B 2 7 ( + ) R E I T E R ’ S N O R M A L N O R M A L A N K Y L O S I N G

S P O N D Y L I T I S

Figure 3. Stimulation of T-enriched lymphocytes by formalin killed Yersiniu enterocoliticu serotype 3 (Ye3) in different subject groups. Lymphocytes at 2 x lo5 cells/well were stimulated with formalin killed Ye3 at 9 x lo7 bacteridml. Net tritiated thymidine (’HTdr) incorporation and mean values from the sixth day of culture are shown. Mean values for Reiter’s patients were significantly different from HLA-B27 positive normal subjects (P < 0.0005) or the other subject groups ( P < 0.01).

B27 positive normal subjects (mean 5,414 t 3,490 cpm) ( P < 0.0005) or for the other groups ( P < 0.01) (Figure 3). Most subjects were tested on more than 1 occasion. Although they showed variability in abso- lute responses when tested on multiple occasions, 28 of 3 1 subjects repeatedly examined showed consistent- ly high or low responses. For example, the net 3HTdr incorporation responses for a Reiter’s patient studied on 3 separate occasions were 40,449, 46,709, and 130,727 cpm, while the responses for an HLA-R27 positive normal subject on 3 occasions were 0, 0, and 10,439 cpm.

Only 4 of the 18 Reiter’s patients studied were HLA-B27 negative. Their mean response on the sixth day of culture was 21,282 2 4,866 cpm, which was not significantly different from the mean response of the HLA-B27 positive Reiter’s patients.

100

h

z a v 80

m 0 7

x

60 v

W Y

c g 20

0

&

a

L

0

s. 0

1 8 0 0

a i t

H L A - B 2 7 ( + ) H L A - B 2 7 ( + ) H L A - B 2 7 ( + ) R E I T E R ’ S NORMAL A N K Y L O S I N G

S P O N D Y L I T I S

Figure 4. Stimulation of T-enriched lymphocytes by heat killed Yersiniu enterocoliticu serotype 3 (Ye3) in different subject groups. Lymphocytes at 2 x 10’ cells/well were stimulated with heat killed Ye3 at 9 x lo7 bacteria/ml. Net tritiated thymidine (’HTdr) incorpo- ration and mean values from the sixth day of culture are shown. Mean values were not significantly different.

254 BRENNER ET AL

Ye8 R O U G H Ye3 Ye 3

Figure 5. Stimulation of T-enriched lymphocytes from Reiter’s patients by 3 formalin killed strains of Yersinia enterocolitica serotype 3 (Ye3). Ye3 is the wild type, smooth strain, while rough Ye3 is a mutant derived from the wild type strain (see text). Ye8 is Y enterocolitica serotype 8 strain. Lymphocytes at 2 x 10’ cells/well were all stimulated with these bacterial strains at 9 x lo7 bacte- ridml. Net tritiated thymidine (3HTdr) incorporation and mean values from the sixth day of culture are shown. Mean values were significantly different (see text).

noted was not accounted for by a difference in lympho- cyte viability, we determined the percent of viable cells remaining after 6 days of culture. The mean percent for 7 Reiter’s patients was 82.2 ? 5.5% compared with 88.9 2 9.3% for 10 HLA-B27 positive normal subjects. These 2 values were not significantly different.

Response to modified Y e3: heat denaturation. To examine the specificity of Ye3 as a T lymphocyte stimulator, we heat denatured the bacterial proteins by autoclaving the bacteria instead of formalin killing. Results shown in Figure 4 reveal that the responses to heat killed Ye3 were similar in all subject groups tested. The mean responses for HLA-B27 positive Reiter’s, normal, and ankylosing spondylitis subjects were 21,888 r 5,546, 15,383 ? 4,076, and 29,018 2 8,479 cpm, respectively, and were not significantly

different. For each subject, the responses to heat killed Ye3 (Figure 4) and formalin killed Ye3 (Figure 3) were performed on the same day from the same blood sample.

Response to Ye3 mutant: rough strain. Mutation can spontaneously occur when bacteria are allowed to overgrow in exhausted bacteriologic media. A pheno- typically different mutant strain derived from the origi- nal wild type, smooth strain can be identified because of the altered colony morphology. These rough mu- tants are reported to have major changes in the lipo- polysaccharides of their cell walls as well as minor changes in their outer membrane proteins (15,2 1). Such a rough mutant of Ye3 was subcultured as described in Materials and Methods. The T lympho- cyte response to a formalin killed rough mutant was examined in 6 Reiter’s patients. The mean response to the wild type, smooth strain was 31,424 * 8,570 cpm compared with 3,214 & 1,491 cpm when stimulation was with an equivalent amount of the rough mutant ( P < 0.001) (Figure 5 ) . Thus, the ability to stimulate T lymphocytes was lost when the wild type, smooth form mutated to a rough form.

Response to another Yersinia serotype not associ- ated with reactive arthritis: Ye8. A European serotype 3 Yersinia enterocolitica strain described thus far had been chosen for study because it is one of the sero- types predominantly associated with reactive arthritis. A Y enterocolitica serotype 8 strain (Ye@, thus far not associated with reactive arthritis, was chosen for comparison. The response to formalin killed Ye8 in 6 Reiter’s patients (mean 4,076 ? 2,032 cpm) was signifi- cantly lower than the response to formalin killed Ye3 (mean 31,424 k 8,570 cpm) (P < 0.001) (Figure 5).

Response of other lymphocyte fractions to for- malin killed Ye3: lack of specific response in non-T and unfractionated PBMC. Parallel experiments were car- ried out to determine whether the high responses noted in the T cell preparations from Reiter’s patients would also be observed with non-T cells and PBMC. Figure 6A shows the results of non-T cells stimulated by formalin killed Ye3 in HLA-B27 positive Reiter’s patients and normal subjects. The responses were comparable with means of 6,033 4 2,045 and 4,743 r 2,883 cpm, respectively. In addition, Figure 6B shows that the responses in Reiter’s patients were similar to controls when the unfractionated PBMC preparations were examined. The mean for Reiter’s patients was 28,531 ? 7,689 cpm, and the mean for control subjects was 29,943 2 7,205 cpm, which were not significantly different. Thus the higher responses noted in the T-

T CELLS IN REITER’S SYNDROME 255

DISCUSSION

20 30[

I a

10 1 :

a

a

H L A - B 2 7 ( + ) H L A - B 2 7 ( + ) R E I T E R ’ S NORMAL

0

80

A n .-

20

a

a i 0

a - L a

a 0

R 0 I a

0

H L A - B 2 7 ( + ) H L A - B 2 7 ( + ) R E I T E R ’ S NORMAL

Figure 6. Stimulation on non-T cells and unfractionated peripheral blood mononuclear cells (PBMC) by formalin killed Yersinio entero- rolitica serotype 3 (Ye3) in different subject groups. Lymphocytes of each type were cultured at 2 x lo5 cells/well and were stimulated with formalin killed Ye3 at 9 x lo7 bacteridrnl. Net tritiated thymidine (3HTdr) incorporation and mean values for each group from the sixth day of culture are shown. Mean values were not significantly different. A , Responses for non-T cells. B, Responses for unfractionated PBMC.

enriched lymphocyte preparations from Reiter’s pa- tients (Figure 3) were not observed when non-T cells or unfractionated PBMC from these same individuals were examined. For each subject, these experiments with T cells, non-T cells, or PBMC preparations were performed on the same day from the same blood sample.

We have shown that patients with spontaneous Reiter’s syndrome have a heightened in vitro T lym- phocyte proliferative response to formalin killed Ye3. The mean 3HTdr incorporation for 14 HLA-B27 posi- tive Reiter’s patients was 27,409 t 5,488 cpm versus 5,414 2 3,490 cpm for a control group of 1 1 HLA-B27 positive normal individuals ( P < 0.0005). It is known that heat killing of bacteria will denature bacterial proteins and eliminate certain antigenic determinants which can be serologically detected in formalin killed bacteria of the same strain (14,19). It was therefore not surprising that lymphocyte responses to these 2 prepa- rations were different. The higher responses to formal- in killed Ye3 were not observed with heat killed Ye3 (Figures 2-4). Elimination of the discriminating re- sponse also occurred when Ye3 mutated to the rough form (Figure 5 ) . In summary, these observations sug- gest that the discriminating, high response in Reiter’s patients requires the preservation of determinants expressed by the formalin killed form of the wild type or smooth Ye3. Other reported studies examining peripheral blood lymphocyte responses to Yersinia enterocolitica in Reiter’s patients which failed to iden- tify a discriminating response utilized only heat killed bacteria (22,23). Clearly, analysis of both heat and formalin killed forms of this strain of Yersinia associat- ed with reactive arthritis was essential to the success of the present study.

The role of HLA-B27 in the spondylarthropath- ies remains controversial. Whether HLA-B27 is in linkage disequilibrium with a disease associated gene or whether the HLA-B27 molecule itself is the impor- tant factor in disease susceptibility is not known. It is clear from the lack of response in HLA-B27 positive normal individuals and ankylosing spondylitis patients that the response described in this report is not HLA- B27 associated. Further, the 4 HLA-B27 negative Reiter’s patients present among the 18 studied here had a mean response similar to the HLA-B27 positive Reiter’s patients.

The high responses in Reiter’s patients’ T lym- phocyte preparations were not observed when the non-T cell or unfractionated PBMC were examined. Nondiscriminating responses in the non-T cells (Fig- ure 6A) were not surprising since little evidence for B cell hyperactivity or autoantibodies exists in this sero- negative disease (24). However, since PBMC were typically composed of approximately 70% T cells, it was noteworthy that a discriminating response similar

256 HRENNER ET AL

t o that for T-enriched lymphocytes was not observed. The fact that for each individual, the experiments with PBMC and T cells were carried out on the same blood sample on the same day suggests that this was not because of technical variations from sample to sample o r from day to day . One other possibility was that ou r rosetting procedure had depleted or enriched the Leu- 2a o r Leu-3a positive populations of lymphocytes. However. a phenotypic analysis using these monoclo- nal antibodies failed to identify abnormal o r altered proportions of these cells in ou r T-enriched o r PBMC lymphocyte preparations (data not shown). The per- centage of monocytes in culture has also been shown to affect the level of proliferative responses to mitogen and soluble antigens (25). We therefore determined monocyte numbers and found no significant differ- ences in the T-enriched fraction of lymphocytes from Reiter’s patients and control groups (data not shown). Thus, we are unable to explain the lack of discriminat- ing responses in unseparated PBMC. The analysis of several major lymphocyte fractions was, however, important in identifying a high response in Reiter’s patients’ T lymphocytes.

To compare lymphocyte responses with anoth- e r serotype of Y enterocoliticu. we chose Ye8 because this serotype has not been associated with reactive arthritis. The high response to Ye3 (mean 31,424 C

8,570 cpm) was not observed with Ye8 (mean 4,076 2 2,032 cpm, P < 0.001) a s shown in Figure 5 . Epidemio- logic studies have shown that Keiter’s syndrome or reactive arthritis has followed epidemics of Shigelln jlcxrzrri but not Shigelfcr sorlnei (26). Those authors have suggested that only certain strains of bacteria may be arthritogenic. Our in vitro data would be consistent with this conclusion. since Ye3 has been associated with reactive arthritis but Ye8 has not. However, analysis of a larger series of bacterial strains including examples o f t hose associated and not associ- ated epidemiologically with reactive arthritis will be necessary to confirm this point.

Why should spontaneous Reiter’s patients in the United States have a high response to a bacterium associated with reactive arthritis in Europe? It is unlikely that our patients have actually had antecedent infection with Ye3. because this serotype of Yersinia has only very rarely been isolatcd in the United States, even when specifically sought (27-29). Further, the bacterial agglutination test w e utilized revealed no measurable serum titer recognizing Ye3 in any of our subjects. If these patients have not previously encoun- tered Ye3, perhaps they have encountered similar

antigenic determinants through unrecognized infec- tion. Although out da ta suggest this possibility, direct proof such as the identification of antibody from spontaneous Reiter’s syndrome patients which recog- nize putative bacteria remains the goal of future work.

REFERENCES

I . Vartiainen NJ, Hurri L: Arthritis due to Salmonella typhimurium: report of 12 cases of migratory arthritis in association with Salmonella typhimurium infection. Acta Med Scand 175:771-776. 1964

2. Noer HK: An “experimental” epidemic of Reiter’s syndrome. JAMA 197593-698. 1966

3 . Ahvonen P: Human yersiniosis in Finland. I . Bacteriolo- gy and serology. Ann Clin Res 4:30-38, 1972

4. Ahvonen P: Human yersiniosis in Finland. 11. Clinical features. Ann Clin Kes 4:39-48, 1972

5 . Laitinen 0, Tuuhea J. Ahvonen P: Polyarthritis associ- ated with Yersinia enterocolitica infection. Ann Rheum Dis 31:34-39, 1972

6. Aho K, Ahvonen P, Lassus A, Sievers K. Tiilikainen A: HI,-A 27 in reactive arthritis: a study of Yersinia arthri- tis and Reiter’s disease. Arthritis Rheum 17:521-526, 1974

7. Laitinen 0, Leirisalo M. Skylv G : Relation between HLA-B27 and clinical features in patients with Yersinia arthritis. Arthritis Rheum 2O:l 121-1 124, 1977

8. Leirisalo M. Skylv G , Kousa M , Voipio-Pulkki L-M, Suoranta H . Nissila M. Hvidman L. Nielsen ED. Svej- gaard ‘4, Tiilikainen A. IAi’nen 0: Followup study on patients with Keiter’s disease and reactive arthritis, with special reference to HLA-B27. Arthritis Rheum 25:249- 259, 1982

Y. Arnett C: The implications of HL-A W27. Ann Intern

10. Nicholls A: Reiter’s disease and HL-A27. Ann Rheum Dis (suppl) 34:27-28, 1975

I I . Ford DK: Keiter’s syndrome. Hull Rheum Dis 20:588- 591, 1970

12. Kellgren JH: Diagnostic criteria for population studies. Bull Rheum Dis \3:291-292, 1962

13. Terasaki PI, Bernoco D. Park MS, Ozturk G , Iwaki Y: Microdroplet testing for HLA-A. -B , -C and -D antigens. Am J Clin Pathol 69: 103-120, 1978

14. Garvey JS, Cremer NE, Sussdorf DH: Methods in Immunology. Boston, W. A. Benjamin, Inc., 1972, pp

IS. Luderitz 0, Westphal 0: The significance of enterobac- terial mutants for the chemical investigation of their cell- wall polysaccharides. Angewandte Chemie Chem Int Edit 5:198-210, 1966

Med 84194-95, 1976

149- I52

T CELLS IN REITER’S SYNDROME 257

16. Boyum A: Isolation of mononuclear cells and granulo- cytes from human blood. Scand J Clin Lab Invest (suppl

17. Saxon A, Feldhaus JL, Robbins RA: Single step separa- tion of human ’r and B cells using AET treated SRBC rosettes. J Immunol Methods 12:285-288, 1976

18. Fan PT, Yu DTY, Pearson CM, Bluestone R: Human monocyte-lymphocyte interaction: a new technique. J Immunol 119: 156-161, 1977

19. Winblad S: Immune response to Yersinia and Pasteur- ella, Manual of Clinical Immunology. Second edition. Edited by NR Rose, H Friedman. Washington. DC. American Society for Microbiology, 1980, pp 474-488

20. Bottone EJ: Yersinia enterocolitica: a panoramic view of a charismatic organism. CRC Crit Rev Microbiol 5:211-241. 1977

21. Nikaido H: Permeability of the outer membrane of bacteria. Angewandte Chemie 16:337-350, 1979

22. Ford DK, Doreen MR. Prafulla S: Cell-mediated im- mune responses of synovial mononuclear cells to sexual- ly transmitted, enteric and mumps antigens in patients with Reiter’s syndrome, rheumatoid arthritis and anky- losing spondylitis. J Rheumatol 8:220-232, 198 1

23. Enlow RW, Bias WB, Bluestone R, Arnett FC: Human

97) 21:77-89, 1968

lymphocyte responses to selected infectious agents in Keiter’s syndrome and ankylosing spondylitis. Rheuma- to1 Int 1:171-175, 1982

24. Bluestone R, Goldberg LS, Weisbart RH, Morris RI , Holborow EJ: Aberrant immunity in (HL-AIW27-posi- tive rheumatic disease. Ann Rheum Dis (suppl) 34:46- 48, 1975

25. Rosentreich DL. Mizel SB: The participation of macro- phages and macrophage cell lines in the activation of T lymphocytes by mitogens. lmmunol Rev 40: 102-135, 1978

26. Simon DG. Kaslow RA, Rosenbaum J , Kaye RL, Calin A: Reiter’s syndrome following epidemic Shigellosis. J Rheumatol 8:969-973, 1981

27. Snyder JD, Christenson E, Feldman RA: Human Yer- sinia enterocolitica infections in Wisconsin: clinical, laboratory and epidemiological features. Am J Med

28. Quan TJ: Biotypic and serotypic profiles of 367 Yersinia enterocolitica cultures of human and environmental ori- gin in the United States. Contrib Microbiol Immunol

29. Bissett ML: Yersinia enterocolitica isolates from hu- mans in California. J Clin Microbiol 4:137-144, 1976

72:768-774. 1982

5:83-87, 1979