0021-972x/92/7501-0295$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1992 by The Endocrine Society
Vol. 75, No. 1 Printed in U.S A
Immunoreactivity to Yersinia enterocolitica Patients with Autoimmune Thyroid Disease* PATRICIA ARSCOTT, EVAN D. ROSEN, MICHAEL M. KAPLAN, THOMAS ELLIS, JAMES R. BAKER, JR.
RONALD NORMAN
J. KOENIG, THOMPSON,
Departments of Medicine (P.A., E.D.R., R.J.K., J.R.B.), Surgery (N.T.), and Pathology University of Michigan Medical School, Ann Arbor, Michigan 48109-0666 ABSTRACT
Antigens
in
AND
(P.A., T.E., J.R.B.),
hormone levels. Patients and controls with serological reactivity to YOP also showed reactivity with Yersiniu membranes. In addition to the serological studies, cellular immune responses were determined by peripheral blood mononuclear cell proliferation assays. Cellular reactivity to the release proteins was present in four of five Graves’ and both Hashimoto’s patients tested, but also in two of six nonthyroid illness patients with serological immunity to the release proteins. Intrathyroidal lymphocytes obtained from two Graves’ patients demonstrated marked proliferation in response to the release proteins. These results indicate that there is no unique pattern of serological reactivity against Yersinia membranes or the release proteins in patients with autoimmune thyroid diseases and suggest that any causal relationship between Yersinia infection and Graves’ disease may be related to T-cell immunity. (J Clin Endocrinol Metub 75: 295-300, 1992)
Recent reports have suggested that Yersiniu enterocoliticu proteins encoded by a 72-kilobase virulence plasmid (known as release proteins and now identified as YOP2& are antigens recognized specifically by patients with Graves’ disease and of potential etiological importance in this disorder. To examine this hypothesis, we evaluated immune responses to YOP in patients with autoimmune thyroid disease and in normal controls. Humoral responses to Yersinia were assessed using Western blots of crude Y. enterocolitica membrane proteins, Yersinia release proteins (YOP,.,), and human thyrocyte membranes. Twentyfour of 25 Graves’ and 10 of 18 Hashimoto’s patients showed reactivity with the release proteins, primarily the 67., 46-, 36-, and 25-kilodalton bands. However, 17 of 24 normal subjects also demonstrated serological reactivity to the release proteins, and the pattern of reactivity of these sera was similar to that in the thyroid patients. No correlation was noted between serological reactivity to the release proteins and thyroid
S
EVERAL studies have suggested an association between Graves’ disease and infection with the enteric pathogen Yersinia enferocolifica. Graves’ patients have been shown to have serological reactivity to Yersinia at an increased frequency compared to the general population (l-3). The identification of a Yersinia TSH-binding moiety with affinity and specificity similar to those of the mammalian receptor (4) led to the intriguing hypothesis that Graves’ disease may result from molecular mimicry (5). However, no study has documented a causal relationship between Yersinia infection and the development of Graves’ disease. While there have been scattered reports of the development of Graves’ disease temporally associated with acute Yersinia infections (6), the overall incidence of Graves’ disease is not increased in patients with Yersinia infections (1). Recent efforts to evaluate the relationship between Yersinia and Graves’ disease have focused on the role played by several proteins produced by the pathogen’s 72-kilobase (kb) virulence plasmid (7). These release proteins (YOP2m5), so-
called because they are released from the bacteria under low concentrations of divalent cations, may be important to the association between the bacteria and Graves’ disease for a number of reasons. The release proteins are necessary for the pathogenesis of the organism and are present in Y. enterocolifica serotypes associated with Graves’ disease (8). These proteins reportedly share antigenic cross-reactivity with heat shock proteins and other cellular constituents thought to be autoantigens (9). There is evidence of serological reactivity to YOPzmJ in patients with chronic Yersinia infection and in many patients with post-Yersinia reactive arthritis, another autoimmune syndrome (10). Preliminary investigations have suggested that TSH induces the expression of these proteins (9) and that immunoreactivity to some of the YOP is uniquely associated with specific aspects of Graves’ disease (8). Because of the potential etiological and diagnostic importance of these findings, we examined serological reactivity to Y. enferocolifica release proteins in a population of North American patients with autoimmune thyroid disease and compared it with that in a group of normal controls. Through this, it was hoped to determine whether some aspect of serological reactivity to the Yersinia YOP2-5 was significantly associated with Graves’ disease or alterations in thyroid function in patients with AITD. Since T-cell reactivity is thought to play a major role in the etiology of Graves’ disease (1 l), we also assessed the ability of Yersinia YOPzm5 to stimulate the proliferation of T-lymphocytes from patients with Graves’ disease.
Received August 20, 1991. Address all correspondence and requests for reprints to: James R. Baker, Jr., M.D., 1520 MSRB I, University of Michigan Medical Center, Ann Arbor, Michigan 48109-0666. * This work was supported by grants to J.R.B. from the NIAID (R29AI-30501) and the Boots Corp. MSTP Training Grant GM-07863 (to E.D.R.). Core support was also provided by the Michigan Multipurpose Arthritis Center (P60-AR-20557). This work was presented in abstract form at the National Meeting of the American Federation for Clinical Research, Seattle, WA, May 1991.
295
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
296
ARSCOTT ET AL. Materials and Methods
Patients Sera from 25 Graves’ disease patients were obtained. Each patient was within 6 months of diagnosis at the time of donation. All were hyperthyroid, had the clinical stigmata of Graves’ disease, and were found to have diffuse increased uptake on thyroid scan. Serum TSH concentrations were undetectable at the time of diagnosis in all patients (co.05 mU/L; Magic Lite, Corning Diagnostics, Medfield, MA). Free T4 values for these patients ranged from 33.4-137.7 pmol/L (normal, 8.8 22.7). Sera from 18 Hashimoto’s disease patients also were obtained. All of these patients either had goiters or were hypothyroid at the time of diagnosis and had titers of antimicrosomal antibodies in excess of 1:1600 (agglutination assay, Ames Diagnostics, Ames, IA). However, these patients were euthyroid on replacement therapy at the time of the study. Control sera were obtained from patients seen in a general medicine out-patient clinic who did not have thyroid disease or from laboratory personnel of the same age and sex as the patients.
TSH binding
studies
Iodination grade human TSH (Calbiochem, San Diego, CA) was iodinated with the lactoperoxidase method and purified by Concanavalin-A affinity chromatography (12). Yersinia enterocolitica, serotype 0:3, was a gracious gift of Dr. Allen Cross, Walter Reed Army Institute of Research. Serotype 0:3 was chosen because the reports linking Y. enterocolitica exposure to Graves’ disease have specifically identified this serotype (1, 2, and 3), and this serotype has also been shown to have a saturable TSH-binding site (4). Single colonies of Y. enterocolitica or E. coli DH5a (used as a control) were picked from Lennox plates and grown overnight at 37 C in a shaking incubator in 2 mL Lennox broth. A volume of the broth containing an OD bOOnm of 0.7 of bacteria was spun down in siliconized microfuge tubes. Bacterial pellets were suspended in 1 mL 100 rnM Tris, pH 8.0, and centrifuged. The pellets were then suspended in 1.5 mL lysis buffer (13.3 mg EGTA and 10 mg lysozyme/ 100 mL 100 rnM Tris, pH 8.0) and incubated for 47 min at room temperature. Bacteria were microcentrifuged for 4 min at room temperature, and the pellets were washed with 0.5 mL 10 mM Tris, pH 7.5. The pellets were microcentrifuged again, suspended in 0.3 mL binding mix [lo mM Tris (pH 7.5) and 10,000 cpm [‘?]TSH, with or without 0.25 U nonradiolabeled bovine TSH (Sigma, St. Louis, MO)], and incubated at room temperature for 1 h. After a lo-min centrifugation at 4 C, the tubes were placed on ice, the supernatants were decanted, and the pelleted radioactivity was measured in a y-counter.
Proteinase-K
digestion
of
and isolation
preparations
Membranes were prepared from whole bacteria using differential centrifugation (13). The final membrane pellet was suspended in 10 mM CHAPS in phosphate-buffered saline, and the protein content was estimated by absorbance at 280 nm.
of
YOP,-, (release proteins)
Release proteins were prepared using the method of Heesemann et al. (14). Control preparations were made in an identical manner, except that bacterial cultures were not treated with EGTA.
SDS-PAGE SDS-PAGE was carried out according to the technique of Laemmli (15). Approximately 50 pg total protein were placed in each lane, with polyacrylamide gel gradients ranging from 5-20%. The gels were either stained with Coomassie brilliant blue or electrotransferred to nitrocellulose for immunoblotting studies.
Immunoblotting
studies
Western blots of nitrocellulose transfers were performed using standard techniques (14) with the following modifications, Strips were incubated overnight at 4 C with dilutions of patient serum. Both 1:lO and 1:40 dilutions of serum were used. The strips were then washed three times in PBS containing 0.05% Tween-20 for 10 min before being placed in antihuman IgG (y-chain specific) alkaline phosphatase conjugate (Sigma) for 2-4 h at 27 C. The strips were removed from the secondary antibody, washed as previously described, and placed in BCIP/NBT substrate (Kirkegard Perry, Gaithersburg, MD). Development of strips was stopped by washing with water after 20 min or earlier if significant background color developed.
Lymphocyte
proliferation
Lymphocyte proliferations were carried out using freshly isolated peripheral blood mononuclear cells (PBMC). Briefly, 100,000 cells were suspended in each well of 96-well plates using 100 pL/well serum-free media (ABC media, Pan Data Systems, Rockville, MD). One hundred microliters per well of various concentrations of YOP, mitogens, thyroid autoantigens, or medium alone were then added to each well in quadruplicates. Cells were cultured for 5 days, pulsed overnight with [3H] thymidine (0.8 pCi/well), harvested, and counted. Intrathyroidal lymphocytes were obtained by mechanical and enzymatic disruption of thyroidectomy specimens (16), with subsequent primary culture for 2448 h. The nonadherent cells were then removed and isolated using Ficoll-Hypaque gradients. The cells were washed and cultured with stimulants in a manner similar to the PBMC, except that only 20,00050,000 cells/well were used.
Yersinia
Bacteria were grown and lysed as described above. After the microcentrifugation and washing steps, bacterial pellets were suspended in 0.3 mL 10 rnr.4 Tris, pH 7.5, with or without the addition of proteinaseK and incubated at 37 C. Several concentrations of proteinase-K were tested, ranging from 0.05-Z mg/mL. Incubation times from l-4 h were used. After digestion, samples were spun for 4 min in a microfuge and washed three times with 10 mM Tris, pH 7.5. Binding of [iZ51]TSH was assessed as described above, except that the binding mix included 1 rnr+.r phenylmethylsulfonylfluoride and 5 mg/mL leupeptin to inactivate any residual proteinase-K. Supernatants from the binding reactions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), which demonstrated the integrity of the (Y- and P-chains of the radiolabeled TSH.
Membrane
Preparation
JCE & M -1992 Vol75.Nol
Results TSH binding
studies
serotype 0:3 bound 26.5 -I- 2.0% (mean + n = 15) of the total input counts of [‘251]TSHtracer. The addition of 0.25 IU nonradioactive TSH reduced tracer binding to 8.9 + 3.2%. The addition of a lo-fold molar excessof BSA, the carrier protein for nonradioactive TSH, did not alter tracer binding. Surprisingly, treatment of Yersinia lysates with proteinaseK did not inhibit TSH binding, but often caused a modest enhancement of binding (Fig. 1A). Analysis of the proteinaseK-treated bacterial preparation by SDS-PAGE and Coomassie blue staining showed two major protein bands to be resistant to enzymatic digestion (Fig. 1B). Attempts to detect [‘251]TSHbinding to theseproteins by detergent solubilization and gel filtration or protein blotting followed by probing with [‘251]TSHwere not successful(data not shown). Y. enterocolitica
SEM;
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
YERSINIA
ANTIBODIES
IN GRAVES’
A +/-
kDa 200 -
Unlabeled Proteinase Incubation
TSH K (hrs)
58
66-
i
43-*
”
B +/-
C +/-
D +/-
E
F
G
H
+/-
1
FIG. 2. Western blot analysis of serological reactivity of five Graves’ disease patients (A-E) with Yersinia proteins. The immunoblots were performed either with (+) or without (-) preincubation of the serum with an E. coli strain shown to have TSH-binding ability (DH5cr). Patients were reactive with a large number of bands of the Yersinia preparations. Reactivity with the majority of these bands was not removed by preincubation with the E. coli membranes. The bands identified by the Graves’ patients were similar to the ones identified by three normal sera (F-H).
-
36.F 26.6-
Graves’ patients were also recognized by these controls (Fig. 2, lanes F, G, and H). 9
// Characterization
Proteinase Incubation
297
DISEASE
K (hrs)
1
;
+ 2
3’
FIG. 1. A, [9]TSH binding to Y. enterocolitica is unaffected by proteinase-K treatment of the bacterial lysates. Yersinia lysates were incubated with tracer TSH in the absence (bar 1) or presence (bar 2) of nonradiolabeled TSH, and the bacteria-associated radioactivity was determined. The Yersinia lysates used in bars 3-6 were pretreated with 1 mg/mL proteinase-K for l-3 h, and the proteinase-K was inactivated before performing the [rz51]TSH binding study. B, SDS-PAGE of proteinase-K-digested Yersinia lysates. Aliquots of the Yersinia lysates used in A were subjected to SDS-PAGE and Coomassie blue staining. Two major bands (36 and 27 kDa) are resistant to proteolysis. Western blots of Y. enterocolitica
membranes
Crude Y. enterocolitica membranes were analyzed using Western blotting techniques in an attempt to identify the specific antigenic bands recognized by Graves’ diseasepatients. Western blots probed with sera from five Graves’ patients (Fig. 2) demonstrated binding to multiple bands in the bacterial membrane preparations. Preincubation of the serawith E. coli membranes showed that specific TSH binding removed some of the bands; however, the majority of the serological activity to the Y. enterocolitica membranes remained intact. The activity of three normal individuals who had positive serology against Yersinia is also shown in Fig. 2 for comparison. Most of the bands recognized by the
of
Yersinia
release proteins
The YOP2-5 were initially analyzed by SDS-PAGE and protein staining. As expected (14), six prominent protein bands were identified, corresponding to mol wt of 67, 58, 46, 40, 36, and 25 kilodaltons (kDa; data not shown). SDSPAGE analysis of control preparations from bacteria not treated with EGTA revealed no protein bands. Since one of the proteinase-K-resistant Yersinia proteins is 36 kDa, we considered that this may be the same speciesas the 36-kDa release protein. However, this seemedunlikely, since treatment of the YOP2-5preparations with proteinaseK results in complete protein degradation (data not shown). Treatment of Yersinia cultures with EGTA did not lead to a decreasein TSH binding, which might be expected if YOPZ5 were releasedinto the broth. Binding of [‘251]TSHto preparations of YOP2-5 also was not detected when analyzed by gel filtration or protein blotting, followed by probing with [‘251]TSH(data not shown). Thus, it would seemunlikely that any of the release proteins are responsible for the TSHbinding activity of Yersiniu. Serological
studies of YOPz-,
Sera from patients with Graves’ disease,Hashimoto’s thyroiditis, or normal individuals were used to probe Western blots of YOP2-5. Representative data are shown in Figs. 3 and 4. Sera from both of the patient groups and the controls showed reactivity with YOI’Z-5; the most intense blotting
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
298
ARSCOTT Graves’ Patients
JCE & M. 1992 Vol75.Nol
ET AL.
Hashimoto’s Patients
Normals
67
46 40 36
25
3. Western blots of Y. enterocolitica release proteins probed with serum from the two patient groups and normal controls. Graves’ patients (left group) showed reactivity with all of the RP (YOP& While less reactivity was noted in serum from the Hashimoto’s patients (center group) or age- and sex-matched normal subjects (right group), similar bands were recognized in all three groups of sera.
FIG.
al
2 5 2 CT E 8 t a
patient serum. However, individual sera from control patients reactive to the 46-kDa band showed immunostaining of this protein as intense as that seenin the Graves’ patients, suggesting that reactivity to this YOP is not unique. No association was noted between titers of antithyroglobulin or microsomal antibodies and reactivity to a particular release protein in the Hashimoto’s patients, and thyroid hormone levels in the Graves’ patients also showed no correlation with any immunoblotting pattern.
80
60
40
20
Cellular
0 Any
RP
87 Kd
Release =
Graves’
Patients
58
Kd
Protein
46
Kd
40
Kd
Molecular
@% Hashimoto’s
Patients
36 Kd
25 Kd
Weight n
Controls
FIG. 4. Percentage of sera reactive with any specific release protein from the two patient groups and controls. Overall reactivities to release proteins were similar in all three groups. However, Graves’ patients showed evidence of increased reactivitv with the 58- and 46-kDa bands compared with the controls.
occurred with the 46-, 40-, and 36-kDa bands. No band was uniquely recognized by either of the patient groups, and the intensity of the blotting was not greater for any of the patient groups compared to the controls. Overall, 24 of 25 Graves’ patients’ sera showed reactivity with at least one of the YOP compared with 10 of 18 Hashimoto’s patients and 17 of 24 controls. Western blots probed with higher concentrations of serum (1:lO instead of 1:40) showed greater background; however, the same banding pattern was seen (data not shown). We attempted to correlate serological reactivity to specific YOPZm5from Fig. 4 with several aspects of autoimmune thyroid disease. Only two of the YOP were recognized by serum from the Graves’ patients at a higher frequency than the control groups. In particular, the 46-kDa band was recognized by almost 100% of the Graves’ patients serum, but by less than 50% of the Hashimoto’s or the nonthyroid
immune
reactivity
to YOPzm5
PBMC from Graves’ patients and control patients (without thyroid disease) demonstrating serological reactivity with YOP2-5 were evaluated for proliferative responses to the release proteins (Fig. 5). Four of five Graves’ patients and both Hashimoto’s patients assayed showed dose-dependent proliferation in responseto YOP2-5that was not observed in response to similar amounts of control (EGTA-negative) broth precipitates (Fig. 5A). Two of six seropositive nonthyroid controls also showed proliferative responsesto the releaseproteins in their PBMC (Fig. 5B). Stimulation indices of positive responsesranged from 3-10 times the background level (medium). However, the seropositive (nonthyroidal) control patients who demonstrated proliferative responsesto Yersinia showed levels of reactivity equivalent to those in the Graves’ patients. None of three patients with Reiter’s syndrome showed evidence of cellular reactivity to releaseproteins in their PBMCs. However, these three patients also had no serological evidence that Yersinia was responsiblefor their reactive arthritis. We were fortunate to obtain intrathyroidal lymphocytes from two Graves’ patients. These two patients demonstrated proliferative responsesto the releaseproteins in their PBMC, and both demonstrated significant proliferation of their intrathyroidal lymphocytes to the release proteins (15,000 + 500 to 25,000 f 650 cpm with background counts
Vol. 75, No. 1 Printed in U.S A
Immunoreactivity to Yersinia enterocolitica Patients with Autoimmune Thyroid Disease* PATRICIA ARSCOTT, EVAN D. ROSEN, MICHAEL M. KAPLAN, THOMAS ELLIS, JAMES R. BAKER, JR.
RONALD NORMAN
J. KOENIG, THOMPSON,
Departments of Medicine (P.A., E.D.R., R.J.K., J.R.B.), Surgery (N.T.), and Pathology University of Michigan Medical School, Ann Arbor, Michigan 48109-0666 ABSTRACT
Antigens
in
AND
(P.A., T.E., J.R.B.),
hormone levels. Patients and controls with serological reactivity to YOP also showed reactivity with Yersiniu membranes. In addition to the serological studies, cellular immune responses were determined by peripheral blood mononuclear cell proliferation assays. Cellular reactivity to the release proteins was present in four of five Graves’ and both Hashimoto’s patients tested, but also in two of six nonthyroid illness patients with serological immunity to the release proteins. Intrathyroidal lymphocytes obtained from two Graves’ patients demonstrated marked proliferation in response to the release proteins. These results indicate that there is no unique pattern of serological reactivity against Yersinia membranes or the release proteins in patients with autoimmune thyroid diseases and suggest that any causal relationship between Yersinia infection and Graves’ disease may be related to T-cell immunity. (J Clin Endocrinol Metub 75: 295-300, 1992)
Recent reports have suggested that Yersiniu enterocoliticu proteins encoded by a 72-kilobase virulence plasmid (known as release proteins and now identified as YOP2& are antigens recognized specifically by patients with Graves’ disease and of potential etiological importance in this disorder. To examine this hypothesis, we evaluated immune responses to YOP in patients with autoimmune thyroid disease and in normal controls. Humoral responses to Yersinia were assessed using Western blots of crude Y. enterocolitica membrane proteins, Yersinia release proteins (YOP,.,), and human thyrocyte membranes. Twentyfour of 25 Graves’ and 10 of 18 Hashimoto’s patients showed reactivity with the release proteins, primarily the 67., 46-, 36-, and 25-kilodalton bands. However, 17 of 24 normal subjects also demonstrated serological reactivity to the release proteins, and the pattern of reactivity of these sera was similar to that in the thyroid patients. No correlation was noted between serological reactivity to the release proteins and thyroid
S
EVERAL studies have suggested an association between Graves’ disease and infection with the enteric pathogen Yersinia enferocolifica. Graves’ patients have been shown to have serological reactivity to Yersinia at an increased frequency compared to the general population (l-3). The identification of a Yersinia TSH-binding moiety with affinity and specificity similar to those of the mammalian receptor (4) led to the intriguing hypothesis that Graves’ disease may result from molecular mimicry (5). However, no study has documented a causal relationship between Yersinia infection and the development of Graves’ disease. While there have been scattered reports of the development of Graves’ disease temporally associated with acute Yersinia infections (6), the overall incidence of Graves’ disease is not increased in patients with Yersinia infections (1). Recent efforts to evaluate the relationship between Yersinia and Graves’ disease have focused on the role played by several proteins produced by the pathogen’s 72-kilobase (kb) virulence plasmid (7). These release proteins (YOP2m5), so-
called because they are released from the bacteria under low concentrations of divalent cations, may be important to the association between the bacteria and Graves’ disease for a number of reasons. The release proteins are necessary for the pathogenesis of the organism and are present in Y. enterocolifica serotypes associated with Graves’ disease (8). These proteins reportedly share antigenic cross-reactivity with heat shock proteins and other cellular constituents thought to be autoantigens (9). There is evidence of serological reactivity to YOPzmJ in patients with chronic Yersinia infection and in many patients with post-Yersinia reactive arthritis, another autoimmune syndrome (10). Preliminary investigations have suggested that TSH induces the expression of these proteins (9) and that immunoreactivity to some of the YOP is uniquely associated with specific aspects of Graves’ disease (8). Because of the potential etiological and diagnostic importance of these findings, we examined serological reactivity to Y. enferocolifica release proteins in a population of North American patients with autoimmune thyroid disease and compared it with that in a group of normal controls. Through this, it was hoped to determine whether some aspect of serological reactivity to the Yersinia YOP2-5 was significantly associated with Graves’ disease or alterations in thyroid function in patients with AITD. Since T-cell reactivity is thought to play a major role in the etiology of Graves’ disease (1 l), we also assessed the ability of Yersinia YOPzm5 to stimulate the proliferation of T-lymphocytes from patients with Graves’ disease.
Received August 20, 1991. Address all correspondence and requests for reprints to: James R. Baker, Jr., M.D., 1520 MSRB I, University of Michigan Medical Center, Ann Arbor, Michigan 48109-0666. * This work was supported by grants to J.R.B. from the NIAID (R29AI-30501) and the Boots Corp. MSTP Training Grant GM-07863 (to E.D.R.). Core support was also provided by the Michigan Multipurpose Arthritis Center (P60-AR-20557). This work was presented in abstract form at the National Meeting of the American Federation for Clinical Research, Seattle, WA, May 1991.
295
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
296
ARSCOTT ET AL. Materials and Methods
Patients Sera from 25 Graves’ disease patients were obtained. Each patient was within 6 months of diagnosis at the time of donation. All were hyperthyroid, had the clinical stigmata of Graves’ disease, and were found to have diffuse increased uptake on thyroid scan. Serum TSH concentrations were undetectable at the time of diagnosis in all patients (co.05 mU/L; Magic Lite, Corning Diagnostics, Medfield, MA). Free T4 values for these patients ranged from 33.4-137.7 pmol/L (normal, 8.8 22.7). Sera from 18 Hashimoto’s disease patients also were obtained. All of these patients either had goiters or were hypothyroid at the time of diagnosis and had titers of antimicrosomal antibodies in excess of 1:1600 (agglutination assay, Ames Diagnostics, Ames, IA). However, these patients were euthyroid on replacement therapy at the time of the study. Control sera were obtained from patients seen in a general medicine out-patient clinic who did not have thyroid disease or from laboratory personnel of the same age and sex as the patients.
TSH binding
studies
Iodination grade human TSH (Calbiochem, San Diego, CA) was iodinated with the lactoperoxidase method and purified by Concanavalin-A affinity chromatography (12). Yersinia enterocolitica, serotype 0:3, was a gracious gift of Dr. Allen Cross, Walter Reed Army Institute of Research. Serotype 0:3 was chosen because the reports linking Y. enterocolitica exposure to Graves’ disease have specifically identified this serotype (1, 2, and 3), and this serotype has also been shown to have a saturable TSH-binding site (4). Single colonies of Y. enterocolitica or E. coli DH5a (used as a control) were picked from Lennox plates and grown overnight at 37 C in a shaking incubator in 2 mL Lennox broth. A volume of the broth containing an OD bOOnm of 0.7 of bacteria was spun down in siliconized microfuge tubes. Bacterial pellets were suspended in 1 mL 100 rnM Tris, pH 8.0, and centrifuged. The pellets were then suspended in 1.5 mL lysis buffer (13.3 mg EGTA and 10 mg lysozyme/ 100 mL 100 rnM Tris, pH 8.0) and incubated for 47 min at room temperature. Bacteria were microcentrifuged for 4 min at room temperature, and the pellets were washed with 0.5 mL 10 mM Tris, pH 7.5. The pellets were microcentrifuged again, suspended in 0.3 mL binding mix [lo mM Tris (pH 7.5) and 10,000 cpm [‘?]TSH, with or without 0.25 U nonradiolabeled bovine TSH (Sigma, St. Louis, MO)], and incubated at room temperature for 1 h. After a lo-min centrifugation at 4 C, the tubes were placed on ice, the supernatants were decanted, and the pelleted radioactivity was measured in a y-counter.
Proteinase-K
digestion
of
and isolation
preparations
Membranes were prepared from whole bacteria using differential centrifugation (13). The final membrane pellet was suspended in 10 mM CHAPS in phosphate-buffered saline, and the protein content was estimated by absorbance at 280 nm.
of
YOP,-, (release proteins)
Release proteins were prepared using the method of Heesemann et al. (14). Control preparations were made in an identical manner, except that bacterial cultures were not treated with EGTA.
SDS-PAGE SDS-PAGE was carried out according to the technique of Laemmli (15). Approximately 50 pg total protein were placed in each lane, with polyacrylamide gel gradients ranging from 5-20%. The gels were either stained with Coomassie brilliant blue or electrotransferred to nitrocellulose for immunoblotting studies.
Immunoblotting
studies
Western blots of nitrocellulose transfers were performed using standard techniques (14) with the following modifications, Strips were incubated overnight at 4 C with dilutions of patient serum. Both 1:lO and 1:40 dilutions of serum were used. The strips were then washed three times in PBS containing 0.05% Tween-20 for 10 min before being placed in antihuman IgG (y-chain specific) alkaline phosphatase conjugate (Sigma) for 2-4 h at 27 C. The strips were removed from the secondary antibody, washed as previously described, and placed in BCIP/NBT substrate (Kirkegard Perry, Gaithersburg, MD). Development of strips was stopped by washing with water after 20 min or earlier if significant background color developed.
Lymphocyte
proliferation
Lymphocyte proliferations were carried out using freshly isolated peripheral blood mononuclear cells (PBMC). Briefly, 100,000 cells were suspended in each well of 96-well plates using 100 pL/well serum-free media (ABC media, Pan Data Systems, Rockville, MD). One hundred microliters per well of various concentrations of YOP, mitogens, thyroid autoantigens, or medium alone were then added to each well in quadruplicates. Cells were cultured for 5 days, pulsed overnight with [3H] thymidine (0.8 pCi/well), harvested, and counted. Intrathyroidal lymphocytes were obtained by mechanical and enzymatic disruption of thyroidectomy specimens (16), with subsequent primary culture for 2448 h. The nonadherent cells were then removed and isolated using Ficoll-Hypaque gradients. The cells were washed and cultured with stimulants in a manner similar to the PBMC, except that only 20,00050,000 cells/well were used.
Yersinia
Bacteria were grown and lysed as described above. After the microcentrifugation and washing steps, bacterial pellets were suspended in 0.3 mL 10 rnr.4 Tris, pH 7.5, with or without the addition of proteinaseK and incubated at 37 C. Several concentrations of proteinase-K were tested, ranging from 0.05-Z mg/mL. Incubation times from l-4 h were used. After digestion, samples were spun for 4 min in a microfuge and washed three times with 10 mM Tris, pH 7.5. Binding of [iZ51]TSH was assessed as described above, except that the binding mix included 1 rnr+.r phenylmethylsulfonylfluoride and 5 mg/mL leupeptin to inactivate any residual proteinase-K. Supernatants from the binding reactions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), which demonstrated the integrity of the (Y- and P-chains of the radiolabeled TSH.
Membrane
Preparation
JCE & M -1992 Vol75.Nol
Results TSH binding
studies
serotype 0:3 bound 26.5 -I- 2.0% (mean + n = 15) of the total input counts of [‘251]TSHtracer. The addition of 0.25 IU nonradioactive TSH reduced tracer binding to 8.9 + 3.2%. The addition of a lo-fold molar excessof BSA, the carrier protein for nonradioactive TSH, did not alter tracer binding. Surprisingly, treatment of Yersinia lysates with proteinaseK did not inhibit TSH binding, but often caused a modest enhancement of binding (Fig. 1A). Analysis of the proteinaseK-treated bacterial preparation by SDS-PAGE and Coomassie blue staining showed two major protein bands to be resistant to enzymatic digestion (Fig. 1B). Attempts to detect [‘251]TSHbinding to theseproteins by detergent solubilization and gel filtration or protein blotting followed by probing with [‘251]TSHwere not successful(data not shown). Y. enterocolitica
SEM;
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
YERSINIA
ANTIBODIES
IN GRAVES’
A +/-
kDa 200 -
Unlabeled Proteinase Incubation
TSH K (hrs)
58
66-
i
43-*
”
B +/-
C +/-
D +/-
E
F
G
H
+/-
1
FIG. 2. Western blot analysis of serological reactivity of five Graves’ disease patients (A-E) with Yersinia proteins. The immunoblots were performed either with (+) or without (-) preincubation of the serum with an E. coli strain shown to have TSH-binding ability (DH5cr). Patients were reactive with a large number of bands of the Yersinia preparations. Reactivity with the majority of these bands was not removed by preincubation with the E. coli membranes. The bands identified by the Graves’ patients were similar to the ones identified by three normal sera (F-H).
-
36.F 26.6-
Graves’ patients were also recognized by these controls (Fig. 2, lanes F, G, and H). 9
// Characterization
Proteinase Incubation
297
DISEASE
K (hrs)
1
;
+ 2
3’
FIG. 1. A, [9]TSH binding to Y. enterocolitica is unaffected by proteinase-K treatment of the bacterial lysates. Yersinia lysates were incubated with tracer TSH in the absence (bar 1) or presence (bar 2) of nonradiolabeled TSH, and the bacteria-associated radioactivity was determined. The Yersinia lysates used in bars 3-6 were pretreated with 1 mg/mL proteinase-K for l-3 h, and the proteinase-K was inactivated before performing the [rz51]TSH binding study. B, SDS-PAGE of proteinase-K-digested Yersinia lysates. Aliquots of the Yersinia lysates used in A were subjected to SDS-PAGE and Coomassie blue staining. Two major bands (36 and 27 kDa) are resistant to proteolysis. Western blots of Y. enterocolitica
membranes
Crude Y. enterocolitica membranes were analyzed using Western blotting techniques in an attempt to identify the specific antigenic bands recognized by Graves’ diseasepatients. Western blots probed with sera from five Graves’ patients (Fig. 2) demonstrated binding to multiple bands in the bacterial membrane preparations. Preincubation of the serawith E. coli membranes showed that specific TSH binding removed some of the bands; however, the majority of the serological activity to the Y. enterocolitica membranes remained intact. The activity of three normal individuals who had positive serology against Yersinia is also shown in Fig. 2 for comparison. Most of the bands recognized by the
of
Yersinia
release proteins
The YOP2-5 were initially analyzed by SDS-PAGE and protein staining. As expected (14), six prominent protein bands were identified, corresponding to mol wt of 67, 58, 46, 40, 36, and 25 kilodaltons (kDa; data not shown). SDSPAGE analysis of control preparations from bacteria not treated with EGTA revealed no protein bands. Since one of the proteinase-K-resistant Yersinia proteins is 36 kDa, we considered that this may be the same speciesas the 36-kDa release protein. However, this seemedunlikely, since treatment of the YOP2-5preparations with proteinaseK results in complete protein degradation (data not shown). Treatment of Yersinia cultures with EGTA did not lead to a decreasein TSH binding, which might be expected if YOPZ5 were releasedinto the broth. Binding of [‘251]TSHto preparations of YOP2-5 also was not detected when analyzed by gel filtration or protein blotting, followed by probing with [‘251]TSH(data not shown). Thus, it would seemunlikely that any of the release proteins are responsible for the TSHbinding activity of Yersiniu. Serological
studies of YOPz-,
Sera from patients with Graves’ disease,Hashimoto’s thyroiditis, or normal individuals were used to probe Western blots of YOP2-5. Representative data are shown in Figs. 3 and 4. Sera from both of the patient groups and the controls showed reactivity with YOI’Z-5; the most intense blotting
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 22:03 For personal use only. No other uses without permission. . All rights reserved.
298
ARSCOTT Graves’ Patients
JCE & M. 1992 Vol75.Nol
ET AL.
Hashimoto’s Patients
Normals
67
46 40 36
25
3. Western blots of Y. enterocolitica release proteins probed with serum from the two patient groups and normal controls. Graves’ patients (left group) showed reactivity with all of the RP (YOP& While less reactivity was noted in serum from the Hashimoto’s patients (center group) or age- and sex-matched normal subjects (right group), similar bands were recognized in all three groups of sera.
FIG.
al
2 5 2 CT E 8 t a
patient serum. However, individual sera from control patients reactive to the 46-kDa band showed immunostaining of this protein as intense as that seenin the Graves’ patients, suggesting that reactivity to this YOP is not unique. No association was noted between titers of antithyroglobulin or microsomal antibodies and reactivity to a particular release protein in the Hashimoto’s patients, and thyroid hormone levels in the Graves’ patients also showed no correlation with any immunoblotting pattern.
80
60
40
20
Cellular
0 Any
RP
87 Kd
Release =
Graves’
Patients
58
Kd
Protein
46
Kd
40
Kd
Molecular
@% Hashimoto’s
Patients
36 Kd
25 Kd
Weight n
Controls
FIG. 4. Percentage of sera reactive with any specific release protein from the two patient groups and controls. Overall reactivities to release proteins were similar in all three groups. However, Graves’ patients showed evidence of increased reactivitv with the 58- and 46-kDa bands compared with the controls.
occurred with the 46-, 40-, and 36-kDa bands. No band was uniquely recognized by either of the patient groups, and the intensity of the blotting was not greater for any of the patient groups compared to the controls. Overall, 24 of 25 Graves’ patients’ sera showed reactivity with at least one of the YOP compared with 10 of 18 Hashimoto’s patients and 17 of 24 controls. Western blots probed with higher concentrations of serum (1:lO instead of 1:40) showed greater background; however, the same banding pattern was seen (data not shown). We attempted to correlate serological reactivity to specific YOPZm5from Fig. 4 with several aspects of autoimmune thyroid disease. Only two of the YOP were recognized by serum from the Graves’ patients at a higher frequency than the control groups. In particular, the 46-kDa band was recognized by almost 100% of the Graves’ patients serum, but by less than 50% of the Hashimoto’s or the nonthyroid
immune
reactivity
to YOPzm5
PBMC from Graves’ patients and control patients (without thyroid disease) demonstrating serological reactivity with YOP2-5 were evaluated for proliferative responses to the release proteins (Fig. 5). Four of five Graves’ patients and both Hashimoto’s patients assayed showed dose-dependent proliferation in responseto YOP2-5that was not observed in response to similar amounts of control (EGTA-negative) broth precipitates (Fig. 5A). Two of six seropositive nonthyroid controls also showed proliferative responsesto the releaseproteins in their PBMC (Fig. 5B). Stimulation indices of positive responsesranged from 3-10 times the background level (medium). However, the seropositive (nonthyroidal) control patients who demonstrated proliferative responsesto Yersinia showed levels of reactivity equivalent to those in the Graves’ patients. None of three patients with Reiter’s syndrome showed evidence of cellular reactivity to releaseproteins in their PBMCs. However, these three patients also had no serological evidence that Yersinia was responsiblefor their reactive arthritis. We were fortunate to obtain intrathyroidal lymphocytes from two Graves’ patients. These two patients demonstrated proliferative responsesto the releaseproteins in their PBMC, and both demonstrated significant proliferation of their intrathyroidal lymphocytes to the release proteins (15,000 + 500 to 25,000 f 650 cpm with background counts
