Autoantibodies to thyroid hormones the role of thyroglobulin

Document Sample
Autoantibodies to thyroid hormones the role of thyroglobulin Powered By Docstoc
					Clin Exp Immunol 1996; 105:140–147

                   Autoantibodies to thyroid hormones: the role of thyroglobulin

      K. ERREGRAGUI, F. CHEILLAN, J. P. DEFOORT, S. PRATO & V. FERT Immunotech, Marseille, France

                                                (Accepted for publication 28 March 1996)

                Autoantibodies against thyroid hormones (THAA) are frequently detected in the sera of patients with
                thyroid disorders together with autoantibodies against thyroglobulin (TGAA). THAA are considered to
                be a subset of TGAA, but alternative possibilities have not been excluded. We hypothesize that if THAA
                arise through an immune response to iodothyronines carried by circulating thyroglobulin (hTg), THAA
                should be found together with autoantibodies against the peptide backbone of hTg (TPAA) close to the
                hormone-forming sites. We measured TPAA in 178 serum samples, obtained from healthy subjects and
                patients with thyroid disorders, using two hormone-forming peptides isolated from hTg. The occurrence
                of TPAA was much lower than that of TGAA. Autoantibodies to the hormone-rich peptide, P3, were
                significantly more common than autoantibodies to the hormone-poor peptide, P1 (111/178 62.3% forˆ
                TGAA versus 21/178 11.8% for anti-P3 TPAA and 7/178 3.9% for anti-P1 TPAA). The presence of
                                       ˆ                                     ˆ
                                                                                     ‡                 ÿ
                autoantibodies to thyroid hormones was investigated in 25 TPAA and 26 TPAA sera. THAA were
                found more frequently in TPAA sera (10/25 40% for TPAA and 4/26 15.3% for TPAA ).
                                                    ‡              ˆ                ‡              ˆ                      ÿ
                Correlation analysis shows that the anti-P3, but not the anti-P1 binding activity, correlates positively
                with the THAA-binding activity (P 0.001 for anti-T4 THAA; P 0.01 for anti-T3 THAA). Specificity
                                                     <                             <
                of anti-P3 TPAA indicates that a subset of the anti-P3 antibodies is directed against the thyroid hormone
                moiety and another subset is directed against the peptide backbone near the hormone-forming peptide,
                according to our hypothesis. These results indicate that the THAA response is an anti-hTg response
                directed, in a significant number of cases, against the hormone-forming site included in the P3 peptide.
                This response seems to be elicited by either native hormone-rich hTg or by hTg fragments.

                Keywords      thyroglobulin    autoantibodies   thyroid hormones   peptide    immunoassay

                      INTRODUCTION                                     processing pathway, do not elicit an immune response by themselves.
                                                  131                  During maturation of the hormone, they are covalently linked to the
Since the demonstration of abnormal binding of I-thyroxine to
                                                                       high molecular weight hTg prohormone (660 000 D) and a defect in
human serum gammaglobulin [1], numerous cases of circulating
                                                                       processing or thyrocyte leakage under pathologic conditions may
autoantibodies with binding activity for thyroid hormones (THAA)
                                                                       result in a high concentration of circulating immuno-stimulating
have been reported (see [2] for review). Albeit THAA are assumed
                                                                       native hTg or hormone-linked hTg peptides. It is well documented
to affect the balance of hormone bound to its natural binding
                                                                       that either hTg or thyroid hormones, chemically conjugated to a
proteins (thyroxine-binding globulin (TBG), thyroxine-binding
                                                                       heterologous carrier protein, induce anti-thyroid hormone antibodies
prealbumin (TBPA), and human serum albumin (HSA)), they
                                                                       when injected into animals with Freund’s adjuvant [14,15]. The latter
rarely exert pathogenic effects [4]. THAA interferes with the
                                                                       observation indicates that hTg is not the unique carrier protein capable
measurement of the thyroid hormones by immunoassay [7], thus
                                                                       of triggering THAA synthesis. Direct evidence for an immunogenic
leading to inappropriate therapy of thyroid diseases [9]. The
                                                                       role of hTg in THAA formation is lacking. The identification of
occurrence of THAA was estimated to be 1.8% in healthy subjects
                                                                       hormone-precursor peptide sequences in the hTg molecule which
[3] and 5% in thyroid diseases [5], with a frequent association with
                                                                       include donor and acceptor tyrosyl sites has provided insight into the
thyro-globulin autoantibodies (TGAA). There is compelling evidence
                                                                       mechanism of hormone synthesis [16]. Thus, it was demonstrated that
that THAA arise through immunization by thyroglobulin (hTg),
                                                                       anti-hormone MoAbs can distinguish structural differences around
released by the thyroid gland under pathologic conditions [11].
                                                                       hormonogenic sites of different species [18] and that some human
Thyroxine and triiodothyronine (776 and 650 D, respectively) are
                                                                       autoantibodies, directed against the N-terminal thyroid hormone-
haptens which, when released into the circulation by the natural
                                                                       forming domain, are sensitive to the changes induced by the hormone
                                                                       synthesis on hTg [19]. In the light of these findings we hypothesize
   Correspondence: V. Fert, Immunotech, 130 Avenue de Lattre de        that if THAA are formed in response to hTg, they should be detected
Tassigny, 13276 Marseille Cedex 9, France.                             in a significant number of THAA+ samples together with autoanti-

140                                                                                                                 #   1996 Blackwell Science
                                          hTg elicits autoantibodies to thyroid hormones                                                   141
bodies against the peptide backbone near the hormone-forming sites.       24 h at 37 C, the measurement of T4 and T3 was carried out by
In this study we purified and characterized hormonogenic fragments         radioimmunoassay (RIA). Sample (100 l) diluted in 120 mmol/l
of hTg and investigated the relationship between the TPAA and             of potassium phosphate pH 7, buffer, containing 150 mmol/l of
THAA response in 178 sera. Some TPAA+ sera were investigated to           NaCl, 2 mmol/l of EDTA, 0.5 mmol/l of 8-anilinonaphthalene-1-
define further the specificity of the antibodies.                           sulfonic acid (ANS), 0.5 g/l of bovine IgG and 10 mmol/l of
                                                                          sodium azide was incubated in avidin-coated tubes with 200 l of  
                                                                          a 125I-labelled anti-thyroxine or anti-triiodothyronine MoAb
                SUBJECTS AND METHODS
                                                                          (Immunotech), diluted in the same buffer, for 1 h at room tem-
Subjects                                                                  perature with shaking. A biotinylated analogue of thyroxine (50 l;   
Serum samples from 178 subjects, including 76 hyperthyroid, 40            166 nmol/l for the T4 immunoassay and 15 nmol/l for the T3
hypothyroid patients (classified by thyroid-stimulating hormone            immunoassay) was then added to each tube and incubation con-
(TSH) and free hormone measurement), 38 patients with diagnosed           tinued for another hour. The tubes were aspirated and bound
thyroid cancer and 24 healthy blood donors were investigated.             radioactivity was determined in a gamma counter calibrated for
                                                                              I. The hormone concentration of the samples was calculated by
Thyroglobulin                                                             interpolation from a standard curve obtained in the same experi-
Pure human thyroglobulin (>96% by polyacrylamide gel electro-             ment with calibrated solutions of thyroxine and triiodothyronine.
phoresis) was obtained from Scipac (Sittingbourne, UK) (hTg I)            The intra-assay coefficients of variation were <5% at concentra-
and from Immunotech AS (Prague, Czechoslovakia) (hTg II).                 tions in the range 5–300 pmol/l for both T4 and T3 immunoassays.
                                                                          The detection limits of the RIAs are 5 pmol/l. Cross-reactivity of
Isolation of hormonogenic tryptic peptides from human                     the antibody is 0.8% for triiodothyronine (anti-T4), 0.15% for
thyroglobulin                                                             thyroxine (anti-T3), 1% and 0.03% for L-3,3 ,5 -triiodothyronine
                                                                                                                        0 0
Human thyroglobulin (hTg I) was digested with trypsin type XIII,          (for anti-T4 and anti-T3, respectively), 0.01% for diiodotyrosine
TPCK-treated (Sigma Chimie, St Quentin, Fallavier, France) at a           and <0.01% for monoiodotyrosine (for both antibodies). No cross-
ratio of 1:16 (weight trypsin:weight hTg) in 50 mmol/l of Tris,           reactivity was observed with L-tyrosine.
15 mmol/l of CaCl2 pH 8, buffer for 2 h at 37 C. The reaction was
terminated by addition of soybean trypsin inhibitor (Biozyme,             Detection of autoantibodies against the isolated tryptic
Blaenavon, UK) (weight inhibitor: weight trypsin 2:1). Thyroglo-          peptides P1 and P3
bulin peptides were fractionated by reverse-phase high perfor-            Serum samples (25 l) were incubated in duplicate with 200 l
mance liquid chromatography (rp-HPLC) on a C4 column                      iodinated P1 or P3 (50 000 ct/min) in PBS containing 0.1% bovine
(7.8 mm 30 cm, 30 nm, 15 m; Nihon Waters, Japan) in a gra-
         2                                                               serum albumin (BSA) and 10 mmol/l of sodium azide. After
dient of trifluoroacetic acid:acetonitrile at a flow rate of 5 ml/min.      overnight incubation at 4 C, the presence of antibody– 125I peptide
Fractions collected were lyophilized and reconstituted in 50 mmol/        complexes was assessed by adding 25 l of normal human plasma,
l of Tris pH 7.2, buffer for the measurement of protein concentra-        and immediately precipitating the complex with 250 l of cold
tion and for the radioimmunoassay of thyroid hormones. Selected           polyethylene glycol 6000 (Merck, Darmstadt, Germany) to a final
fractions (identified as P1 and P3) were pooled for chloramine T           concentration of 10%. The tubes were centrifuged and the radio-
iodination according to the standard procedure. The iodinated             activity of the pellet was determined. Results were expressed as
peptides were further purified on a Superdex 200 column 16/60              the ratio of bound ct/min (minus the non-specific binding) to total
(Pharmacia, Orsay, France) in 8 mmol/l of sodium phosphate,               ct/min.
150 mmol/l of NaCl pH 7.2, buffer (PBS) at a flow rate of 60 ml/h.
                                                                          Detection of anti-thyroglobulin antibodies
Detection of immunoreactive thyroxine in tryptic peptides                 Biotinylated hTg I was added to avidin-coated tubes (Immunotech)
Hormone, linked to the fractionated thyroglobulin tryptic peptides,       at 0.8 g/tube in PBS buffer containing 0.1% BSA. After 48 h of
was measured after lyophilization, using a slight modification of          incubation at 4 C, the tubes were rinsed and 500 l of serum,
the protocol for the free hormone radioimmunoassay kit (cat. no           previously diluted 1:500 in PBS containing 0.1% BSA, 0.5 mol/l
1464; Immunotech SA, Marseille, France). Fifty microlitres of the         NaCl and 10 mmol/l sodium azide, were added. After 2 h of
fractions, redissolved in 50 mmol/l of Tris pH 7.2, buffer, were          incubation with shaking at ambient temperature, the tubes were
incubated with tracer for 2 h at ambient temperature in the coated        rinsed twice and 500 l of 125I protein A, in PBS buffer containing
tubes. The displacement of the 125I tracer from the coated tubes          0.1% BSA, were added and incubation continued for another 2 h.
was a measure of the thyroxine immunoreactivity of the peptides.          After rinsing, bound radioactivity was determined. The sera were
The intra-assay coefficient of variation was <7%.                          assayed in duplicate together with a serial dilution of a sample
                                                                          calibrated against the reference preparation (MRC 65-93). Results
Measurement of hormone linked to hormonogenic peptides                    were expressed in U/ml.
Thyroglobulins from different sources and purified peptides were
digested to the amino acid level according to the procedure               Detection of anti-thyroid hormone autoantibodies
described by Rolland et al. [8]. Briefly, pronase type VI (Sigma           Sera to be tested were stripped of thyroid hormones by the
Chimie) (weight enzyme:weight substrate, 1:5) was added to                following procedure. Serum (50 l) was diluted with 950 l of
1 nmol of hTg I, hTg II or porcine thyroglobulin (pTg) or to the          0.1% trifluoroacetic acid in water. The diluted samples were
same amount of peptide dissolved in 100 mmol/l of Tris pH 8.6,            filtered immediately through a Sep-Pack C18 cartridge (Millipore,
buffer. Digestion was continued after 48 h at 37 C by addition of
                                                   8                      St Quentin Yvelines, France) and neutralized with 200 l of 0.5
leucylaminopeptidase (Sigma Chimie) (weight enzyme:weight                 mol/l of glycine buffer pH 8.9. A suspension of protein A
substrate, 1:5) and 0.1 mol/l 1,4-dithiothreitol. After a further         Sepharose gel (200 l) (Pharmacia) equilibrated in the glycine
#   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
142                                                        K. Erregragui et al.
buffer was incubated with the sample for 2 h at ambient tempera-
ture with agitation. The gel was centrifuged and the pellet rinsed
twice with 1.4 ml PBS containing 0.1% ovalbumin. This procedure
was repeated twice for a total of three rinses. Revelation of
autoantibodies to thyroid hormones was performed by incubating
the gel with 500 l of 125I thyroxine or triiodothyronine (50 000 ct/
min) in PBS, containing 0.1% ovalbumin and 10 mol/l of 8-
anilinophtalene-1-sulfonic acid for 2 h at ambient temperature with
agitation. The gel was then centrifuged and washed with 1.4 ml of
0.15 mol/l NaCl, 0.05% Tween 20. Radioactivity of the pellet was
determined and results expressed as a percentage of gel-bound ct/
min to total ct/min.

Analysis of anti-P3-positive sera
The selected sera (10 l) were incubated with 100 l of 125I P3 and
100 l of buffer (PBS–BSA) or buffer containing thyroxine and
triiodothyronine (5 mol/l of each) or containing native or trypsin-
                                                                         Fig. 1. Typical reverse-phase high performance liquid chromatography (rp-
treated hTg I or II (0.5 mol/l ). After overnight incubation at 4 C,
                                                                 8       HPLC) profile of hTg I pre-digested with trypsin. The pooled fractions
the presence of antibody– 125I peptide complexes was assessed by          which constitute pool P1 and P3 (boxed) were selected and utilized after
adding 25 l of normal human plasma, and immediately precipitat-
                                                                         iodination and purification for the detection of autoantibodies in patient
ing the complex with 250 l of cold polyethylene glycol 6000
(Merck) to a final concentration of 10%. The tubes were centri-
fuged and the radioactivity of the pellet was determined. Results         hormone when compared with hTg I. Porcine thyroglobulin
are expressed as percentage of 125I P3 binding.                           served as control and, as expected, contained significantly more
                                                                          hormone residues than did hTg I.
Processing of results and statistical analysis
                                                                              Figure 2a,b shows the chromatographic exclusion profile
Samples were considered to be positive for values >100 U/ml for
                                                                          of Pool 1 and Pool 3 iodinated peptides. In the experiment of
the TGAA assay and a ratio of >3% of bound versus total tracer
                                                                          Fig. 2a (Pool 1), two iodinated compounds eluted at 30 kD and
activity for TPAA and THAA assays. Correlation between TGAA,
                                                                          15 kD, respectively. The iodinated 15-kD peptide was used for the
TPAA and THAA levels (in percentage B/T) was analysed, by
                                                                          present study and is referred to as P1. The choice was made after
least square regression. The correlation coefficient was determined
                                                                          quantitative purification of the peptide, followed by microsequen-
by the method of Pearson. P < 0.01 was considered significant.
                                                                          cing, and showed that the 15-kD peak of Pool 1 represented a
                                                                          minor hormone-forming domain (to be published). Figure 2b
                                                                          shows that Pool 3 contained one major peptide, eluting at 30 kD;
                                                                          it is referred to as P3. The minor 15-kD peak represented a
Purification and iodination of hormonogenic tryptic hTg                    contamination of the rp-HPLC fractions by the P1 peptide.
Figure 1 depicts a typical rp-HPLC profile of hTg I cleaved by             Assay of autoantibodies to hTg and to peptides P1 and P3
trypsin. The fractions collected were assayed for protein and             in normal subjects and patients
immunoreactive hormone. Fractions containing immunoreactive               Results of immunoassays for TGAA, anti-P1 TPAA and anti-P3
thyroxine elute in seven peaks. Of these peaks, we chose two              TPAA in the 178 sera are summarized in Table 2. TGAA
(fractions boxed in the figure), one with low hormone content (P1)
and one with high hormone content (P3). These two peaks were
obtained with high reproducibility in several purification runs. The       Table 1. Thyroid hormone content of hTg I and II, porcine thyroglobulin
highly immunoreactive peak eluting in fraction 75 was shown to            (pTg) and of the Pool 1 (P1) and Pool 3 (P3) obtained after tryptic digestion
contain at least three different peptides on SDS-gel electrophoresis      of hTg I and reverse-phase high performance liquid chromatography (rp-
(results not shown), therefore this peak was not selected. Tryptic                                   HPLC) fractionation
digestion was complete in 2 h (kinetic gel filtration experiments,
data not shown). This result is comparable to the recent study of
                                                                                                                   Hormone content
Saboori et al. [10]. Hormone immunoreactivity was measured by a
RIA using a high-affinity anti-thyroxine MoAb. Since the fractions         Material                     Thyroxine                    Triiodothyronine
were not digested with pronase before the assay, the measurement
does not reflect free hormone content but the immunoreactivity of          hTg I                           38                              11
the peptide-bound hormone.                                                P1                               0.03                            0.067
    Table 1 depicts the results of the quantitative hormone assay of      P3                               2.2                             0.39
Pool 1 (P1) and Pool 3 (P3). For this measurement we used pronase         hTg II                           0.008                           0.04
digestion, which releases all amino acids [8]. The results are then a     pTg                           >100                              45
measure of the true thyroxine content of the fractions. P1 contains
very few hormone residues, P3 contains about 6% of the hTg I                  Measurement was performed by radioimmunoassay after pronase
thyroxine content (2.2/38 mol) and 3.6% of the hTg I triiodothy-          digestion. The hormone content is expressed in moles of hormone/100
ronine (0.39/11 mol). hTg II contained a very low amount of               moles of protein.
                                                              #   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
                                           hTg elicits autoantibodies to thyroid hormones                                                     143
                                                                           Table 2. Prevalence of anti-hTg (TGAA), anti-P1 and anti-P3 autoanti-
                                                                           bodies in normal control subjects and patients with various thyroid

                                                                           Population     n      TGAA+          P1+           P3+      P1+ and P3+

                                                                           Controls       25      4 (16)      0 (0)       0    (0)       0 (0)
                                                                           Hyper*         75     55 (73.3)    3 (4)      11   (14.6)     1 (1.3)
                                                                           Hypo*          40     39 (97.5)    4 (10)      7   (17.5)     2 (5)
                                                                           Cancer*        38     13 (34.2)    0 (0)       3    (7.9)     0 (0)
                                                                           Total         178    111 (62.3)    7 (3.9)    21   (11.8)     3 (1.7)

                                                                              Values are the number of patients, with the percentage in parentheses.
                                                                              * Diagnosis of the patients. Hyper, Hyperthyroid; hypo, hypothyroid;
                                                                           cancer, thyroid carcinoma.

                                                                           TPAA presence or absence was not significantly related to the
                                                                           TGAA level in U/ml.
                                                                               Data obtained in the five different assays (TGAA, anti-P1 and
                                                                           anti-P3 TPAA, anti-T4 and anti-T3 THAA) of the 51 samples were
                                                                           compared quantitatively. Figure 3 depicts two of the three sig-
                                                                           nificant correlation analyses. The binding potency (in percentage
                                                                           B/T) of anti-T4 and anti-T3 THAA correlated significantly, as
                                                                           expected (graph not shown, r = 0.575, P < 0.001); more interest-
Fig. 2. Exclusion chromatography profiles of Pool P1 (a) and P3 (b) after   ingly, the binding potency of anti-P3 TPAA and anti-T4 or anti-T3
protein iodination. The calibration of the column, with molecular weight   THAA, respectively, correlated significantly (Fig. 3a, r = 0.631,
standards (in kD), is indicated in the upper part of the figures with       P < 0.001, and Fig. 3b, r = 0.369, P < 0.01). No correlation was
arrowheads. The symmetrical arrows indicate fractions utilized for the     found between TGAA and TPAA in the 51 selected sera. This latter
radioimmunoassay of autoantibodies in patient sera.
                                                                           result was also confirmed for the total population (178 sera).

prevalence was, as expected, very high in the sera of hypothyroid          Study of anti-P3 TPAA specificity, in four sera
patients (97.5%), high in hyperthyroid patients (73.3%) and                Inhibitory compounds (hTg I and II; trypsin-treated hTg I and II;
moderate in thyroid carcinoma (34.2%). Anti-P1 TPAA and anti-              thyroid hormones) were incubated with the P3-labelled peptide in
P3 TPAA were far less common than TGAA (25/178 = 14% for all               four sera, positive for anti-P3 TPAA (nos 25, 31, 65 and 133 in
the TPAA compared with 111/178 = 62.4% for TGAA). The                      Table 3). Due to the limited amount of patient sera available, the
antibody response was directed mainly against P3 (21/                      construction of dose–response curves was not feasible. Each hTg
178 = 11.8%) and to a lesser extent against P1 (7/178 = 3.9%).             form was tested (in duplicate) at the highest feasible concentration
The distribution of anti-P1 TPAA+ and anti-P3 TPAA+ sera in                to maximize its inhibitory potency. For the same reason, thyroid
hypo- and hyperthyroid patients was the same as that of TGAA+              hormones (T3 and T4) were mixed together to a final concentration
sera. In sera from patients with thyroid carcinoma, no response            of 5 mol/l. Their respective concentration in the test was thus
against P1 was found by contrast with the response against P3 or           2.5 mol/l and no solvent was required for their solubilization under
hTg (3/38 = 7.9% for anti-P3 TPAA and 13/38 = 34.2% for                    test conditions. The binding profile of the four sera tested shows no
TGAA). Only few patients were positive for both anti-P1 and                marked individual differences (Fig. 4). Pure thyroid hormones dis-
anti-P3 TPAA (3/178 = 1.6%). No TPAA were found in the control             place the binding of 125 I-P3 to TPAA, but about 20% of the tracer
healthy subjects.                                                          remains bound. In contrast, the trypsin-treated hTg I or II displaced
                                                                           the tracer totally. Native hTg I and II showed a dramatic difference in
Assay of autoantibodies to thyroid hormones in patients with               their displacement potency. Whereas the native hormone-rich hTg I
or without TPAA                                                            inhibited nearly 100% of the tracer binding, the hormone-poor hTg
Twenty-five anti-P1 and anti-P3 TPAA+ samples and 26 TPAA              ÿ    II, at the same concentration, did not inhibit significantly the tracer
samples were selected from the 178 sera. THAA determinations               binding. Concentrations of T3 and T4 (free or peptide-bound) present
were performed on these samples. The results of the THAA, TPAA             in the test (calculated from Table 1) are indicated in the lower part of
and TGAA assays, for each serum, are shown in Table 3. The upper           the figure and illustrate the low amount of hormone present in hTg II.
group consists of the anti-P1 and anti-P3 TPAA+ sera, the lower
group of the TPAA sera. Anti-T4 and anti-T3 THAA+ sera were
found more frequently in the upper group (20% and 28%, respec-
tively) compared with their prevalence in the TPAA group (3.8%
                                                        ÿ                  Direct evidence for a role of hTg as the immunogen that triggers
and 11.5%, respectively). Anti-P3 TPAA were frequently asso-               the human THAA response is still lacking [2]. In this study, we
ciated with a THAA response, in contrast with anti-P1 TPAA (9/             postulated that, if hTg or a related structure were the immunogen
21 = 42.8% for anti-P3 TPAA and 1/7 = 14.3% for anti-P1 TPAA).             eliciting THAA, an antibody response to the peptide backbone in
#   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
144                                                           K. Erregragui et al.
Table 3. Results of assays of autoantibodies in the 51 sera selected according to anti-P1 or P3 TPAA presence (upper group) or absence (lower group) Sera
   are sorted using %B/T values of anti-P3 TPAA in decreasing order. Values underlined are above the positivity threshold and considered positive.

                        Sample           Anti-P3 TPAA*            Anti-P1 TPAA*            TGAA†              Anti-T4 THAA‡            Anti-T3 THAA‡
Selection                 no.                (%B/T)                   (%B/T)               (U/ml)                 (%B/T)                   (%B/T)

                          133                  18.0                      0.0                   13                   47.9                     11.4
                           25                  15.0                      2.2                  671                    1.2                      1.0
                           54                  12.8                      0.0                 2306                    2.9                      1.0
                           31                   8.7                      2.7                 4135                    3.2                      1.3
                          183                   8.6                      5.0                  499                    2.1                      0.8
                          256                   7.7                      0.4                    3                    5.7                     10.9
                          106                   6.6                      0.0                 2050                    5.5                      0.1
                           91                   6.5                      0.0                  607                    2.4                      0.6
                           65                   5.7                      0.0                  567                    8.4                      2.6
                          111                   5.6                      0.0                 2994                    3.0                      4.1
                          244                   5.6                      1.2                   18                    1.2                      3.8
TPAA-positive              12                   5.5                      1.1                 1500                    1.9                      0.6
(25 samples)              245                   5.4                      3.0                   73                    2.5                      1.8
                          243                   5.2                      3.8                  477                    1.7                      1.6
                           18                   4.6                      0.0                  621                    1.8                      0.0
                          153                   4.5                      2.0                  300                    2.0                      0.0
                           40                   4.2                      7.1                  120                    0.9                      1.3
                          248                   4.1                      2.1                  121                    1.6                      0.8
                          236                   3.6                      0.0                  347                    1.0                      0.0
                          264                   3.3                      0.0                  340                    2.4                      3.6
                          259                   3.2                      0.0                  224                    0.5                      3.2
                          237                   3.0                      4.0                 3588                    0.1                      0.0
                           36                   3.0                     21.1                  576                    0.7                      1.4
                           77                   0.7                      6.4                  401                    0.5                      0.8
                          275                   0.5                      5.0                 4533                    0.7                      6.9
Total positive §                              21 (84)                   7 (28)              21 (84)                5 (20)                   7 (28)
                          246                    3.0                      0.4                   70                    0.1                     0.0
                          250                    2.9                      2.2                  113                    2.4                     3.0
                           79                    2.8                      0.2               20 000                    0.3                     1.8
                          240                    2.8                      0.4                   11                    0.1                     0.0
                          235                    2.8                      0.6                  771                    0.1                     0.2
                          263                    2.6                      1.4               19 474                    0.6                     5.2
                          273                    2.4                      1.0                1869                     0.1                     3.0
                          260                    2.4                      0.7                   19                    0.8                     6.0
                           98                    2.2                      0.0                    0                    0.1                     0.5
                          262                    2.1                      0.0                1654                     0.4                     2.2
                          251                    2.0                      0.9                  121                    1.6                     3.7
TPAA-negative             276                    1.9                      0.0                   25                    0.1                     1.2
(26 samples)              230                    1.7                      1.1                  847                    0.1                     0.0
                           38                    1.7                      0.0                1195                     0.4                     0.8
                          101                    1.6                      2.0               13 387                    0.4                     0.0
                          231                    1.6                      0.0                   67                    9.0                     0.0
                          219                    1.6                      0.0                   13                    0.1                     0.0
                          233                    1.5                      0.0                  230                    0.1                     0.0
                          189                    1.3                      0.0                  193                    0.1                     0.0
                          217                    1.2                      0.0                    5                    0.1                     0.0
                          242                    1.1                      1.0                   83                    0.1                     0.0
                          113                    1.0                      2.0                1139                     0.1                     0.0
                           30                    0.3                      1.4                5286                     0.4                     0.5
                          265                    0.0                      2.2                  743                    0.1                     2.6
                           35                    0.0                      1.6                  950                    0.3                     1.1
                            3                    0.0                      3.0                1131                     0.4                     0.9
Total positive §                               0 (0)                    0 (0)               17 (65)                1 (3.8)                 3 (11.5)

   * TPAA, Autoantibodies to peptides P1 or P3.
   † TGAA, Autoantibodies to thryoglobulin.
   ‡ THAA, Autoantibodies to thryoid hormone T3 or T4.
   § Values are the number of patients with the percentage in parentheses.

                                                                  #   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
                                           hTg elicits autoantibodies to thyroid hormones                                                         145

                                                                            Fig. 4. Specificity of four anti-P3 autoantibody against thyroid hormones
                                                                            (TPAA)-positive sera. The binding of 125I-P3 to anti-P3 autoantibodies in
                                                                            the presence of the potentially inhibiting compound was investigated
                                                                            (grouped bars) (mean B/Bmax percentage s.e.m.). Inhibiting compounds,
                                                                            co-incubated with samples and tracer, are identified in the upper part of
Fig. 3. This figure compares raw data (B/T percentage) of autoantibodies
                                                                            the figure (tryp, trypsin-digested). Concentrations of added compound,
against thyroid hormones (THAA) and anti-P3 autoantibodies against the
                                                                            together with the corresponding calculated concentration of free or
peptide backbone of hTg (TPAA) measurement in 51 sera. Anti-P3 TPAA
                                                                            peptide-bound thyroid hormone (T4 + T3), are indicated in the lower
and anti-T4 THAA (a), anti-P3 TPAA and anti-T3 THAA (b). Dashed lines
                                                                            part of the figure.
represent the 95% confidence interval. Correlation coefficients and results
of Student’s test are indicated for each comparison.

                                                                            gel exclusion chromatography. They eluted at 15 kD and 30 kD,
                                                                            respectively, and they were sufficiently pure for use as label in a
the vicinity of the hormone-forming sites should be found in                RIA for the detection of circulating autoantibodies.
human sera containing THAA. There are four major identified                       An observation emerging from this study was the low pre-
sites of hormone synthesis and several minor sites in the hTg               valence of autoantibodies against the tryptic hTg peptides P1 and
sequence [16]. Several tyrosine residues are iodinated to give              P3, compared with that of autoantibodies against the native
mono-iodotyrosine (MIT) and di-iodotyrosine (DIT) and are                   molecule. Only 3.9% of all sera were positive for anti-P1 TPAA,
assumed to be involved in hormone synthesis as donor sites                  and they were found in hypo- and hyperthyroid sera and not found
[12,21]. The peptide P1, which has a low but measurable amount              in the sera of thyroid carcinoma patients (0/38). Anti-P3 TPAA,
of hormone, comprises an early iodinated site which has been                although far less frequent than TGAA, were significantly more
located at residue 1466 (19-residue leader peptide included) by             common than anti-P1 TPAA (11.8%). No TPAA were found in
peptide sequence analysis (V. Fert et al., to be published). Lamas          sera from healthy subjects, in contrast with TGAA. These results
and coworkers [16] have shown that this site undergoes significant           agree with current understanding of the anti-hTg autoantibody
iodination to MIT and DIT, but does not form hormone. This                  response: (i) conformational epitopes constitute the target of the
discrepancy from our results may be due to the difference in                anti-hTg response in humans, and tryptic cleavage destroys the
thyroglobulin source (iodinated in vitro in the study of Lamas              major autoantigenic domains of the native molecule [20]; (ii)
et al.) or in the method for measuring the hormones. The cross-             hormonogenic sites share high inter-species homology, including
reactivities of our anti-thyroid hormone MoAbs with DIT or MIT              consensus sequences around the tyrosyl residues undergoing iodi-
are very low (0.01%) but may lead to an overestimation. The P3              nation [16] and are likely to be tolerated by the immune system;
peptide contains a significantly higher amount of hormone than               (iii) it has been shown that a hormone-linked hTg peptide, but not
does P1. Its location is not known, but may correspond to one of the        the peptide devoid of the hormone moiety, triggered a T cell
two hormonogenic domains of hTg at the N-terminal or C-terminal             response in mice [17]. The comparison of the anti-P1 and anti-P3
end of the molecule in the area of Tyr 24 or Tyr 2572 and 2765.             TPAA prevalence provides further evidence that the presence of a
Trypsinization was performed without prior reduction of the hTg             hormone moiety in a conserved peptide sequence is able to
molecule because of the formation of insoluble peptides during              abrogate partly the natural immune tolerance.
proteolysis. According to the published sequence of hTg [6], about               The determination of THAA in 25 TPAA+ and 26-TPAA sera       ÿ
229 different peptides should be released by trypsin cleavage at            shows that the THAA response is associated with the presence of
arginyl and lysyl residues. Conformational structures maintained            autoantibodies to the hormone-rich peptide P3, but not to the
by disulfide bonds are likely to lower this number. After rp-HPLC            hormone-poor peptide P1. The binding potency of anti-P3 TPAA
purification and 125I labelling, P1 and P3 were further purified by           of the 51 sera selected, irrespective of the positive threshold (fixed
#   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
146                                                       K. Erregragui et al.
at a binding of 3% above the non-specific binding), correlated             resulting from the action of thyrocyte proteases. These two
significantly with the binding potency of THAA. These data are in          potential immunogens may be released into the circulation as a
agreement with the hypothesis that THAA and TPAA responses                result of thyroid leakage.
originate from a common stimulus.
    Immune recognition of the P3 peptide by anti-P3 TPAA may
occur in three possible ways: (i) by the hormone moiety itself; (ii)
by the peptide backbone; (iii) by dual recognition of the hormone         We thank Professors P. Carayon and G. Kaphan and Dr F. Roux
and the peptide (shared epitope). It is important to stress first that     (Marseilles, France) for providing us with serum samples and information
certain sera showed relatively high binding capacity for peptide P3       on the thyroid status of the patients. H. Rickenberg is gratefully acknowl-
and weak or no binding capacity for the T3 or T4 hormones, thus           edged for careful reading of the manuscript and comments.
indicating that the anti-P3 TPAA response was not restricted to the
hormone moiety (see sera nos 25, 54 and 183 in Table 3). The                                         REFERENCES
specificity of anti-P3 TPAA investigated on four sera indicates that
the anti-P3 TPAA response is polyclonal and that it targets at least       1 Robbins J, Rall JE, Rawson RW. An unusual instance of thyroxine-
                                                                             binding by human serum gamma-globulin. J Clin Endocrinol Metab
two domains of the P3 peptide: (i) the hormone residue, since T3
                                                                             1956; 16:573–9.
and T4 compete with the antibody-binding sites to inhibit 60–80%           2 Sakata S. Autoimmunity against thyroid hormones. Crit Rev Immunol
of the binding of the radiolabelled P3; (ii) the peptide backbone,           1994; 14:157–91.
since hTg II, treated with trypsin and containing very little              3 Sakata S, Matsuda M, Ogawa T, Takuno H, Matsui I, Sarui H, Yasuda
hormone, competes more efficiently than do thyroid hormones to                K. Prevalence of thyroid hormone autoantibodies in healthy subjects.
inhibit almost 100% of the binding of the tracer. These results              Clin Endocrinol 1994; 41:365–70.
suggest also that the affinity of anti-P3 TPAA for thyroid hormones         4 Karlsson FA, Wibell L, Wide L. Hypothyroidism due to thyroid-
may be increased when the hormone moiety is presented linked to              hormone-binding antibodies. N Engl J Med 1977; 296:1146–98.
the hTg peptide backbone. The experimental data obtained in the            5 Benvega S, Trimarchi F, Robbins J. Circulating thyroid hormone
inhibition experiments do not lead to a precise mapping of the               autoantibodies. J Endocrinol Invest 1987; 10:605–17.
                                                                           6 Malthiery Y, Lissitzky S. Primary structure of human thyroglobulin
antibody combining-sites of anti-P3 TPAA; i.e. is there an anti-
                                                                             deduced from the sequence of its 8448-base complementary DNA. Eur
body population which recognizes both the hormone and the                    J Biochem 1987; 165:491–8.
peptide backbone as described by others, using MoAbs [18]. A               7 Beck-Peccoz P, Romelli PB, Cattaneo MG, Faglia G, White EL, Barlow
comparison of the displacement potency of the hormone-rich hTg I             JW, Stockigt JR. Evaluation of free thyroxine methods in the presence
and of the hormone-poor hTg II indicates that the accessibility of           of iodothyronine-binding autoantibodies. J Clin Endocrinol Metab
the hTg epitopes to the anti-P3 TPAA antibodies is maximal in two            1984; 58:736–9.
cases, i.e. when hTg is mature with respect to hormone synthesis,          8 Rolland RJ, Aquaron R, Litssitzky S. Thyroglobulin iodoamino acids
or when hTg, mature or otherwise, is presented via its tryptic               estimation after digestion with pronase and leucylaminopeptidase. Anal
fragments. This information agrees with previous studies establish-          Biochem 1970; 33:307–12.
ing that hormone synthesis involving iodination and iodo–tyrosyl           9 Pryds O, Hadberg A, Kastrup KW. Circulating autoantibodies to
                                                                             thyroid hormones: a diagnostic pitfall. Acta Pediatr Scand 1987;
coupling induces conformational changes in the hTg molecule that
affect its immunological properties [18,19]. It is worth noting that      10 Saboori AM, Burek CL, Rose NR, Talor M, Kuppers RC. Tryptic
among the TPAA sera (n = 24), four were THAA+. It is likely that
                                                                             peptide of human thyroglobulin. 1. Immunoreactivity with murine
another hormone-forming site, distinct from P3, and not investi-             monoclonal antibodies. Clin Exp Immunol 1994; 98:454–8.
gated here, was responsible for the THAA in these sera.                   11 Sakata S, Nagai K, Tarutani O et al. Two cases of Graves’ disease with
    Several studies point out the lack of correlation between THAA           antithyroid hormone antibodies: implication on the role of thyroglobu-
and TGAA responses [22,23], an observation often used to suggest             lin as an antigen. J Endocrinol Invest 1990; 13:825–30.
that THAA may arise through a pathway that does not involve hTg           12 Ohmiya Y, Hayashi H, Kondo T, Kondo Y. Location of dehydroalanine
as immunogen. In the present study we show that neither THAA                 residues in the amino acid sequence of bovine thyroglobulin. Identifica-
nor TPAA, which are induced by hTg, correlate with TGAA. It is               tion of donor tyrosine sites for hormonogenesis in thyro-
                                                                             globulin. J Biol Chem 1990; 265:9066–71.
likely, as suggested by Sakata et al. using an animal model [24],
                                                                          13 Ochi Y, Shiomi K, Hachiya T, Yoshimura M, Miyazaki T. Immuno-
that the immune response to thyroid hormone, and by extension to             logical analysis of abnormal binding of thyroid hormone in the
the peptide chain near the hormone-forming sites, is controlled by           gammaglobulin. J Clin Endocrinol Metab 1972; 35:743–52.
genes which are distinct from the genes controlling the anti-hTg          14 Chopra IJ. A radioimmunoassay for measurement of thyroxine in
immune response. TGAA, together with anti-thyroperoxidase                    unextracted serum. J Clin Endocrinol Metab 1972; 34:938–47.
autoantibodies, are often used as an indicator of an autoimmune           15 Pierce CJ, Byfield PHG, Edmonds CJ, Himsworth RL. Autoantibodies
thyroid disease in humans. It should be pointed out that anti-P3             to thyroglobulin cross reacting with iodothyronines. Clin Endocrinol
TPAA represent autoantibodies to hTg which may be missed by                  1981; 15:1–10.
TGAA immunoassays that use hormone-poor hTg (see Table 3,                 16 Lamas L, Anderson PC, Fox JW, Dunn JT. Consensus sequences for
sample nos 133, 256, 244 and 245).                                           early iodination and hormonogenesis in human thyroglobulin. J Biol
                                                                             Chem 1989; 264:13541–5.
    The present work confirms that the THAA response is only one
                                                                          17 Hutchings PR, Cooke A, Dawe K, Champion BR, Geysen M, Valerio R,
component of a TGAA response directed against hTg domains                    Roitt Y. A thyroxine-containing peptide can induce murine experi-
involved in thyroid hormone synthesis. This response is mediated             mental autoimmune thyroiditis. J Exp Med 1992; 175:869–72.
by at least autoantibodies to thyroid hormones and of autoanti-           18 Chan CT, Byfield GH. Callus M, Champion B, Himsworth RL.
bodies which recognize the hTg peptide chain near the hormone-               Structural differences around hormonogenic sites on thyroglobulins
forming sites, but not the hormone moiety per se. THAA can be                from different species detected by monoclonal antibodies. Eur J
elicited either by mature, fully iodinated, hTg or by hTg peptides           Biochem 1986; 159:563–7.

                                                              #   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147
                                           hTg elicits autoantibodies to thyroid hormones                                                      147
19 Mallet B, Lejeune PJ, Ruf J, Piechaczyk M, Marriq C, Carayon P.             bodies in Graves’ disease: their changes after antithyroid therapy and
   Tyrosine iodination and iodotyrosyl coupling of the N-terminal thyroid      relationship with the thyroglobulin antibodies. Acta Endocrinol 1990;
   hormone forming site of human thyroglobulin modulate its binding to         122:22–28.
   auto- and monoclonal antibodies. Mol Cell Endocrinol 1992; 88:89–95.     23 Komaki T, Sakata S, Nakamura S, Matsuda M, Kojima N, Miura K.
20 Male DK, Champion BR, Pryce G, Matthews H, Shepherd P. Antigenic            Changes in anti-thyroid hormone and anti-thyroglobulin antibodies
   determinants of human thyroglobulin differentiated using antigen frag-      during thyroid hormone(s) or prednisolone treatment in 3 cases of
   ments. Immunology 1985; 54:419–27.                                          Hashimoto’s thyroiditis. Folia Endocrinol Jpn 1989; 65:615–20.
21 Marriq C, Lejeune PJ, Venot M, Vinet L. Hormone formation in the         24 Sakata S, Okuda K, Aihara Y, Tanaka Y, Nakamura S, Komaki T,
   isolated fragment 1-171 of human thyroglobulin involves the couple          Tarutani O, Miura K. Genetic control of the production of anti-thyroid
   tyrosine 5 and tyrosine 130. Mol Cell Endocrinol 1991; 81:155–63.           hormone antibodies in mice immunized with human thyroglobulin.
22 Wang PW, Huang MJ, Liu RT, Chen CR. Triiodothyronine autoanti-              Immunol Invest 1986; 15:481–92.

#   1996 Blackwell Science Ltd, Clinical and Experimental Immunology, 105:140–147

Shared By: