| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
German Diabetes Research Institute and Department of Endocrinology, University of Dusseldorf (J.S., S.W., M.S., M.L., J.F., W.A.S.), D-40225 Dusseldorf, Germany; and Research Department, B.R.A.H.M.S Diagnostica, Biotechnology Center (N.G.M.), D-16761 Hennigsdorf/Berlin, Germany
Address all correspondence and requests for reprints to: J. Seissler, M.D., German Diabetes Research Institute, University of Dusseldorf, Auf’m Hennekamp 65, D-40225 Dusseldorf, Germany. E-mail: sei{at}dfi
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Since the first report of rNIS at the 10th International Thyroid Conference in 1995 in Toronto, Canada, by Nancy Carrasco, the potential role of NIS as a new autoantigen was soon the "flavor of the month" in thyroid autoimmunity (7) and has been the focus of several groups. The first indirect evidence for an involvement of NIS in thyroid autoimmunity was reported before cloning of the antigen by the demonstration of antibody-mediated inhibition of iodine uptake in primary cultures of thyrocytes in serum from a patient with HT (8). The first study examining a larger group of patients detected autoantibody reactivity against slot-blotted recombinant rNIS in 22 of 26 patients with GD and 3 of 20 patients with HT (9). The same group also reported on a functional bioassay using rNIS-transfected CHO cells, where sera from 4 of 34 patients with HT inhibited iodine uptake (10). Another group showed the binding of IgG from patients’ sera to rNIS peptides in an enzyme-linked immunosorbent assay (ELISA) system (11). To date, only 1 report used hNIS to study autoantibodies. This report describes a bioassay in which sera from patients with autoimmune disease inhibit the uptake of iodine in CHO cells (12). Nevertheless, all of these studies are still being debated, and more data are needed to establish the importance of autoantibodies to NIS.
In this study we evaluate a direct binding assay using recombinant hNIS expressed by in vitro transcription/translation. We examined antibodies to NIS in patients with autoimmune thyroid disease using a radioligand assay similar to those described by us and others for a variety of autoantigens (13, 14, 15, 16). Our results show that NIS may be a target of humoral autoimmunity in only a few patients with HT or GD.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Sera were obtained from 177 patients with GD (female to male ratio, 4:1, median age, 42 yr; range, 10–76 yr), 72 patients with HT (female to male ratio, 3:1; median age, 40 yr; range, 15–72 yr), and 165 healthy controls (female to male ratio, 1:4; median age, 31 yr; range, 18–40 yr). Diagnosis of GD or HT was based on the clinical presentation; measurement of TSH, free T3, and free T4 levels; and detection of TPO antibodies (DYNOtest anti TPO, B.R.A.H.M.S Diagnostica, Hennigsdorf/Berlin, Germany) and autoantibodies to the TSH-R (TBII; DYNOtest TRAK human, B.R.A.H.M.S Diagnostica).
The mouse monoclonal NIS antibody 2.2, raised against the extracellular domain of recombinant hNIS, was a gift from J. C. Morris, Mayo Clinic (Rochester, MN). The rabbit antihuman GLUT-2 antibody was obtained from Research Diagnostics (Flanders, NJ).
cDNA cloning and expression
Ribonucleic acid was isolated from human thyroid tissue using RNA Clean (AGS, Heidelberg, Germany). First strand cDNA was synthesized with AMV reverse transcriptase (Roche Diagnostics, Mannheim, Germany) using random primers. The full-length human Na+/I- symporter was amplified through 30 cycles at 94 C for 30 s, 60 C for 30 s, and 72 C for 60 s in a GeneAmp PCR system 9700 (Perkin-Elmer Corp., Weiterstadt, Germany) using specific primer pairs (5'-GCCGCCACCATGGAGGCCGTGGAGACC-3' and 5'-TGGCCCTGTCCTCAGAGGTT-3') designed from the published NIS cDNA sequence including a Kozak sequence (GCCGCCACC) in the forward primer (GenBank database accession no. U66088) (4). The PCR reaction contained 1 µL cDNA, 1.5 mmol/L MgCl2, 200 µmol/L deoxy-NTPs, 1 µmol/L of each primer, and 0.86 U Expand High Fidelity DNA polymerase (Roche Diagnostics). After reamplification under the same conditions, products were blunted with Klenow and ligated into the HincII site of pGEM 4Z vector (Promega Corp., Madison, WI). Nucleotide sequences of cloned products were determined using an automated sequencing apparatus (Applied Biosystems, Foster City, CA).
Detection of NIS antibodies
The cDNA clone in the pGEM 4Z vector encoding full-length hNIS under the control of the SP6 promoter was used to express antigen by in vitro transcription and translation. One microgram of purified hNIS cDNA was transcribed and translated in the presence of [35S]methionine (Amersham Pharmacia Biotech, Braunschweig, Germany) using the rabbit reticulocyte lysate system (TNT kit, Promega Corp.) according to the manufacturer’s description. Incorporation of radioactivity was determined by precipitation with 10% trichloroacetic acid and liquid scintillation counting. Human glucose transporter-2 (GLUT-2), encoding a protein with 12 membrane-spanning domains of 524 amino acids, was expressed under the same conditions and used as the control protein.
Aliquots of radiolabeled NIS or GLUT-2 (30,000 cpm/sample) were diluted in 100 µL Tris buffer [20 mmol/L Tris, 150 mmol/L NaCl (pH 7.4) with 0.1% BSA, 5 mmol/L methionine, 5 mmol/L benzamidine, 1 mmol/L phenylmethylsulfonylfluoride, 2 mmol/L ethylenediamine tetraacetate, and 1% Triton X-100] and incubated with 20 µL serum or 5 µL monoclonal antibody (anti-hNIS or anti-GLUT-2) overnight at 4 C. After the addition of 150 µL protein A-Sepharose (50%, vol/vol; Pharmacia Biotech, Piscataway, NJ) for 2 h, absorbed immunocomplexes were washed and subjected to SDS-PAGE and autoradiography using the buffer system from Laemmli.
Next, we analyzed NIS antibodies by a radioligand assay to facilitate antibody screening and quantify antibody levels. Ten microliters of serum diluted in 50 µL Tris buffer were incubated with 30,000 cpm radiolabeled hNIS for 12 h at 4 C in 96-well microtiter plates (Greiner, Nurtingen Germany). After the addition of 20 µL protein A-Sepharose (50%, vol/vol) for 2 h, probes were transferred into prewashed 96-well filtration plates (Multiscreen BV, 1.2 µm, Millipore Corp., Bedford, MA) and washed extensively in Tris buffer. After the addition of 20 µL scintillator liquid (Microscint, Canberra-Packard, Downers Grove, IL), the radioactivity of bound immunocomplexes was directly measured in the 96-well plates (Top Count ß counter, Canberra-Packard). In each experiment the same positive (serum P) and negative serum (serum C) was used as internal control to calculate antibody levels in arbitrary units (AU) as follows: (cpm test serum - cpm C)/(cpm P - cpm C) x 100. All sera were analyzed in duplicate.
Statistical analysis
To determine the optimal cut-off for hNIS autoantibody (hNIS-Ab)
positivity, receiver operating characteristic plot analysis was
performed (17), including the data from patients with GD
and HT for sensitivity and healthy controls for specificity. The
difference in the prevalence of hNIS-Ab in the groups was calculated
using the 
2 test. Comparisons between the data
in the different groups were made using the nonparametric Mann-Whitney
U rank sum analysis. Correlations between different assays were made
using the Spearman rank correlation.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The full-length hNIS cDNA clone (nucleotides 348-2279) encoding a
protein with 643 amino acids and a predicted molecular mass of 68.7 kDa
was obtained by DNA amplification of cDNA derived from a thyroid cDNA
library and cloned into pGEM 4Z. The DNA sequence determined for the
cloned human NIS was identical to that previously described
(4). SDS-PAGE analysis of the
[35S]methionine-labeled protein, generated by
in vitro transcription and translation, revealed a major
polypeptide of about 65 kDa (Fig. 1
, lane
1). Immunoprecipitation with the monoclonal antibody 2.2 revealed that
the single major band was hNIS (Fig. 1, lane 8).
|
Binding of hNIS-Ab was initially tested by immunoprecipitation
using 20 sera from TPO antibody and/or TSH receptor antibody-positive
patients with autoimmune thyroid disease and 15 sera from healthy
controls. SDS-PAGE revealed a positive immunoreactivity in sera from 3
patients with HT, 1 patient with GD, and 1 healthy individual. As
illustrated in Fig. 1
, immunoreactive sera strongly bound hNIS, but did
not show any reactivity with the control protein GLUT-2. This confirms
the specificity of the autoantibody reactivity against hNIS.
Subsequently, all sera were tested by radioligand assay. In this assay
data were expressed in AU based on a positive (Fig. 1, lane 4) and a
negative (Fig. 2, lane 7) standard
serum.
|
To detect antibodies to hNIS in a format that allows screening on
a large scale and provides quantitative results, we developed a
radioligand assay. Antibodies were measured in 249 patients with
autoimmune thyroid disease and 165 healthy individuals. Figure 2 shows
the proportion of bound radioactivity by the sera of Fig. 1. The data
were difficult to interpret, because a substantial number of sera from
healthy individuals gave a clear signal in this hNIS antibody assay.
Binding varied from 0.5–17.6% of the added radioactivity in patients
with thyroid disease compared with 0.5–7.9% of that in normal
controls and 18.0% with the monoclonal antibody (dilution, 1:100).
Using the monoclonal hNIS antibody, a positive signal was observed up
to a dilution of 1:8000, indicating a high sensitivity of the assay.
There was no difference in the total IgG concentration in hNIS
antibody-positive and -negative subjects, which could influence
antibody measurement (data not shown). The established method to define
the optimal decision threshold between healthy individuals and patients
suspected of having the parameter in question is receiver operating
characteristic plot analysis (17). As there were no data
concerning the presence or absence of hNIS autoantibodies in healthy
individuals, the cut-off for positivity was arbitrarily set at 20 AU,
corresponding to a specificity of 95.2%. This specificity is similar
to that of established assays detecting TPO autoantibodies, where the
presence of antibodies in otherwise healthy individuals is a well known
observation (18). The inter- and intraassay coefficients
of variation were 10.3% (n = 10) and 14.8% (n = 10),
respectively.
Prevalence of autoantibodies to hNIS
Using the above-described criteria, 19 of 177 patients with
GD (10.7%) and 15 of 72 patients with HT (20.8%) were positive for
hNIS autoantibodies. Among 165 healthy individuals, 8 (4.8%) also had
positive results. The difference between the prevalence in HT and
controls was highly significant (2 =
12.85; P < 0.001), but the difference between GD and
controls did not reach the level of significance
(2 = 3.30; P = 0.06),
nor was there a significant difference between HT and GD. The
prevalence in HT and GD patients was also calculated using more
stringent demands for specificity at 97.0% and 99.4% (Table 1). The sensitivities at these cut-offs were 9.0% and
5.6%, respectively, in the GD group and 15.3% and 6.9%,
respectively, in the HT group. Figure 3
illustrates the distribution of individual sera in patients and
controls. The specificities used for calculations in Table 1 are
indicated by dotted lines. Comparing the hNIS antibody
values in the three groups, there was a significant difference between
the data for HT patients compared with those for patients with GD
(P < 0.001) and healthy controls (P <
0.01). No difference was observed between the group of GD and healthy
controls (P = 0.13; Table 2). There was no
statistical difference in hNIS-Ab levels between males (7.5 ±
11.2 AU; median, 2.3 AU) and females (7.9 ± 11.3 AU; median, 2.3
AU; P = 0.25), but among the positive patients, females
had slightly higher values (40.6 ± 20.4; median, 31.6 AU) than
males (29.7 ± 6.8; median, 31.1 AU).
|
|
|
TBII were present in 90.4% of patients with GD and 12.8% of patients with HT, as determined in the novel coated tube assay using human recombinant TSH-R (2). These data are well in concordance with the percentage reported for GD patients receiving treatment and for HT patients in this new assay. TPO antibodies were present in 100% of patients with HT. There was no correlation between hNIS antibodies and TBII, but there was a weak correlation between hNIS and TPO antibodies (r = 0.35; P < 0.05). All TBII-negative GD patients had hNIS values below 20 U. Therefore, hNIS autoantibody detection did not increase the diagnostic power for GD or HT.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our data are in contrast to previous studies reporting NIS antibodies
in 63–84% of patients with GD and in 12–26% of patients with HT,
using rat NIS (rNIS) in slot blot assays or synthetic rNIS
peptides in ELISA (9, 10, 11). There are major differences
between the antigen preparations and the antibody detection systems in
comparison to our study. Endo and co-workers used rat recombinant NIS
expressed as a histidine-tagged fusion protein, which was purified by
affinity chromatography. As it was difficult to isolate the highly
hydrophobic hNIS protein from Escherichia coli (J. Seissler,
unpublished results), we expressed hNIS in an in vitro
transcription and translation system (TNT system). The TNT system
offers some inherent advantages, including a one-step production of the
recombinant antigen, resulting in a radiolabeled protein without the
need for further purification steps. The detection of autoantibodies by
autoradiography or radioligand assay may be more sensitive and specific
compared with Western blot, slot blot, or ELISA techniques. This has
been demonstrated in several studies for other autoantigens, including
insulin (19), glutamic acid decarboxylase
(20), the tyrosine phosphatase-like protein IA-2
(13, 21), steroid 21-hydroxylase, 17
-hydroxylase
(16), and the TSH-R (14). Some autoantigens
were inactive using prokaryotic expression systems (e.g.
glutamic acid decarboxylase), but preserved their conformation when
produced in the TNT system. It is important to note that we could not
exclude the possibility that expression of hNIS by in vitro
transcription/translation results in incomplete folding of the protein.
This could be due to both the strong hydrophobicity of hNIS and the
expression of the protein in an unglycosylated form. As discussed for
the TSH-R, it could be possible that sugar residues in the
extracellular domains are part of the autoantigenic epitopes
(15). Those antibodies would be missed in our NIS antibody
assay. However, these points do not explain the difference from
previous studies, as they also used unglycosylated proteins, such as
rat NIS purified from E. coli under denaturing conditions (6
mol/L guanidine) for Western blotting (9) or synthetic NIS
peptides for ELISA (11). A positive binding in those
systems indicates the presence of linear epitopes; thus, incomplete
folding or the lack of glycosylation of our hNIS preparation may not
explain the observed discrepancy in the prevalence of hNIS-Ab. It is
intriguing that the comparison with previous studies revealed a similar
frequency of NIS antibodies in patients with HT, but a significantly
lower prevalence in GD. It could be speculated that sera from patients
with GD possess NIS autoantibodies directed to cryptic epitopes that
are not accessible in the full-length hNIS molecule. Alternatively, the
higher binding to denatured NIS may be explained by an as yet unknown
serum factor promoting unspecific binding of autoantibodies toward NIS
or NIS polypeptides. Both hypotheses remain to be clarified in further
studies.
Another area currently under debate is the effect of sera from patients with autoimmune disease on the functional activity of NIS. Several studies in which inhibition of iodine uptake by sera from patients with thyroid diseases was measured in bioassays provided controversial results. Raspe and co-workers reported autoantibody-mediated inhibition of iodine uptake in only 1 patient with HT of 147 sera from patients with autoimmune thyroid disease selected on the basis of high TPO antibodies and hypothyroidism (8). Some reports showed inhibition of iodide uptake in rNIS (10)- or hNIS (12)-transfected CHO cells mediated by sera from patients with autoimmune thyroid disease. However, in the first study a 90% inhibition of iodide uptake was also seen in the sera of controls. This nonspecific effect could be removed by dialyses and was not seen with purified IgG, questioning the use of unfractionated serum in these assays. The more detailed study by Ajjan and co-workers (12) excluded a nonspecific serum effect in their assay, but the total uptake was below 1% of the total 125I activity added. This uptake could be reduced by 50% by the strongest serum from a GD patient. Contradicting data come from a study by Ho and co-workers (22), who tested more than 500 sera from various thyroid patients for iodine uptake-inhibiting activity in transfected COS-7 cells. Only 14 sera showed inhibiting activity, which was lost after dialysis and IgG extraction.
It is interesting to note that all studies reporting a relatively high prevalence of hNIS-ab look at small groups with 20–40 patients and controls. The study by Ho and co-workers and our study are the first to screen large numbers, where selection bias is less likely to occur. Both studies show a similar distribution pattern of hNIS-Ab in autoimmune thyroid patients and controls. Before submission of this paper, Ajjan and colleagues reported on a follow-up study employing a direct binding assay based on the TNT system (23). In contrast to our study, they found a significant difference between 49 GD patients, 29 HT patients, and healthy controls, with 22% of the GD patients and 24% of the HT patients positive. Again, the control group included only 20 individuals, so the differences from our study are most likely based on serum selection, resulting in different cut-off calculations.
The present data indicate that the importance of NIS as an autoantigen in humoral thyroid autoimmunity may be lower than previously reported, at least for patients with GD. The measurement of NIS autoantibodies with currently available assay systems does not offer any additional diagnostic benefit to detect patients with autoimmune thyroid disease. The clinical and pathogenic importance of a potential autoimmune response to NIS remains to be determined.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received June 27, 2000.
Revised August 14, 2000.
Accepted August 31, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Yoshida, I. Hisatome, S. Taniguchi, Y. Shirayoshi, Y. Yamamoto, J. Miake, T. Ohkura, T. Akama, O. Igawa, C. Shigemasa, et al. Pendrin Is a Novel Autoantigen Recognized by Patients with Autoimmune Thyroid Diseases J. Clin. Endocrinol. Metab., February 1, 2009; 94(2): 442 - 448. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Dohan, A. De la Vieja, V. Paroder, C. Riedel, M. Artani, M. Reed, C. S. Ginter, and N. Carrasco The Sodium/Iodide Symporter (NIS): Characterization, Regulation, and Medical Significance Endocr. Rev., February 1, 2003; 24(1): 48 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-K. Chung Sodium Iodide Symporter: Its Role in Nuclear Medicine J. Nucl. Med., September 1, 2002; 43(9): 1188 - 1200. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
