| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Third Medical Department, University of Leipzig, D-04103 Leipzig, Germany
Address all correspondence and requests for reprints to: R. Paschke, M.D., Third Medical Department, University of Leipzig, Ph. Rosenthal Strasse 27, D-04103 Leipzig, Germany. E-mail: pasr{at}server3.medizin.uni-leipzig.de
Abstract
Autonomously functioning thyroid nodules are characterized by an increased proliferation and function, which is predominantly caused by constitutively activating TSH receptor mutations leading to an activation of cAMP. In contrast to autonomously functioning thyroid nodules, cold thyroid nodules are functionally inactive and less differentiated. Their molecular cause is still unknown. To further investigate the pathophysiological aspects of autonomously functioning thyroid nodules and to elucidate the molecular etiology of cold thyroid nodules, it is essential to identify genes with differential expression in autonomously functioning thyroid nodules and cold thyroid nodules and to compare this expression to that in normal surrounding tissue. The list of possible candidates for differential regulation ranges from growth factors and their receptors to transcription factors or oncogenes. Therefore, we evaluated the potential of cDNA arrays and studied the expression of 588 known genes from 6 different classes of proteins in thyroid nodules characterized for their function. Forty-seven genes showed a differential expression between nodular and surrounding tissue identified by the expression arrays. The differential expression of 15 transcripts was verified by real-time PCR. About 25% of the transcripts determined by LightCycler PCR are considered false positives because data from PCR and array analysis did not agree. This indicates the reliability of cDNA expression arrays to identify differentially expressed genes in thyroid nodules compared with their surrounding tissue. The 15 selected genes were additionally quantified by real-time PCR in 7 additional cold thyroid nodules, autonomously functioning thyroid nodules, and their surrounding tissues. The highest number of differentially expressed genes was in the group of signal transduction proteins (4 of 38 detectable genes) and extracellular cell signaling and communication proteins (2 of 62 detectable genes). In contrast, transcripts of other classes of proteins were unchanged (e.g. DNA-binding molecules and stress responses). Most of the transcripts were down-regulated in autonomously functioning thyroid nodule and cold thyroid nodule compared with the respective surrounding tissue. This finding could be the result of a dominant activation of a signal transduction pathway, with the cAMP pathway being the likely candidate for autonomously functioning thyroid nodules. The qualitatively similar pattern of changes in this limited number of genes in autonomously functioning thyroid nodules and cold thyroid nodules could suggest a similar dominant activation of a specific signaling cascade in cold thyroid nodules as the constitutively activating mutations in autonomously functioning thyroid nodules.
IN ABOUT 50% of the German population an increased thyroid volume is detectable (1), which is mainly caused by an insufficient supply of iodine. Among the most frequent thyroid pathologies are endemic euthyroid goiter, nonautoimmune hyperthyroidism caused by autonomously functioning thyroid nodules (AFTN), and cold thyroid nodules (CTN).
The clinical phenotype of AFTN is characterized by hyperthyroidism and
nodular growth, which is caused by somatic mutations of the TSH
receptor in 50–80% of the AFTN (2, 3). These mutations
constitutively activate the cAMP cascade and cause a stimulation of
growth and function (4) explaining the clinical phenotype.
Less frequently, somatic mutations in the 
-subunit of the
Gs protein are known to cause an aberrant
activation of the cAMP cascade downstream of the TSH receptor
(5).
In contrast to AFTN, cold thyroid nodules are functionally inactive and less differentiated. Their molecular etiology is largely unknown. Mutations were identified in the Ras oncogen (for a review see Ref. 6). However, in a sample of CTN with predominant clonal origin, ras mutations are rather rare (7). To date, there are no indications for mutations in other genes. Furthermore, there are no data about altered signal transduction cascades or changed expression of oncogenes, tumor suppressors, or modulators of the intracellular signal transduction. However, genes with a changed expression in nodular thyroid tissue compared with normal surrounding tissue (ST) could define the molecular etiology of CTN and further explain the pathophysiological aspects of AFTN. As the list of possible candidates for differential regulation ranges from growth factors and growth factor receptors to transcription factors or oncogenes, conventional techniques (e.g. Northern blotting) are too time consuming to address this large number of transcripts. We therefore used cDNA arrays that offer the advantage of a highly parallel analysis of gene expression to analyze changes between AFTN and CTN compared with their ST. This approach also allows us to evaluate which group of genes is most frequently affected in the molecular etiology of thyroid nodules and possibly deduce a molecular defect from the expression pattern. We therefore investigated the differential expression of 588 genes (e.g. oncogenes, stress response-related genes, apoptosis-related genes, transcription factors, receptors, and cell-cell communication-related genes) in thyroid nodules compared with their normal ST using cDNA expression arrays.
Subjects and Methods
In all patients thyroid nodules were detected by ultrasound. Hot
and cold thyroid nodules were characterized by scintigraphy. All
preoperatively identified nodules were also identified at surgery and
postoperatively by histology according to the WHO criteria
(8). All patients were euthyroid at surgery. Clonality of
the nodules was previously determined by X chromosome inactivation
(9). Somatic TSH receptor mutations in the hot nodules
were previously determined by denaturing gradient gel electrophoresis
and subsequent direct sequencing of the positive PCR fragments
(10). Molecular, histological, and clinical data of the
patients are given in Table 1
. Informed
consent for the analysis was given by the patients.
|
Total RNA was isolated from ten AFTN (no. 69, 110, 115, 121–123, and 140–143), 10 CTN (no. 32, 46, 50, 51, 53, 60, 61, 65, 67, and 69), and their ST using the Atlas Pure RNA Isolation Kit according to the manufacturer’s instructions (CLONTECH Laboratories, Inc., Palo Alto, CA). Tissues were obtained from male and female patients at surgery and stored at -80 C.
The quality of the total RNAs was examined by denaturing agarose gel electrophoresis and staining with ethidium bromide, and the 18S and 28S RNA bands were visualized by UV illumination. The concentrations of total RNA were quantified spectrophotometrically. The same RNA preparations of AFTN 110, 122, and 123 and CTN 32, 60, and 61 and their corresponding ST were used for both the cDNA expression arrays and the real-time PCR.
Preparation of radioactively labeled cDNA probes for array hybridization
Two micrograms of total RNA for the CTN 32, 60, and 61 and the
AFTN 110, 122, and 123 and their corresponding ST were reverse
transcribed in a 10-µl reaction containing 5 x reaction buffer
[250 mM Tris-HCl (pH 8.3), 375 mM KCl, and 15
mM MgCl2; CLONTECH Laboratories, Inc.]; 0.5 mM each of deoxy (d)-CTP,
dGTP, and dTTP; 1.3 MBq [-32P]dATP (110
TBq/mmol; Amersham Pharmacia Biotech, Braunschweig,
Germany); 5 mM dithiothreitol (CLONTECH Laboratories, Inc.); 1 µl 10 x CDS Primer Mix (CLONTECH Laboratories, Inc.); and 50 U Moloney murine leukemia virus
reverse transcriptase according to the Atlas cDNA Expression Arrays
User Manual (CLONTECH Laboratories, Inc.). The
32P-labeled cDNA was purified from unincorporated
nucleotides and small (<0.1-kb) cDNA fragments by column
chromatography (Chroma Spin-200
diethylpyrocarbonate-H2O column). Afterward, the
incorporation of [-32P]dATP into the probe
was checked by scintillation counting of the fractions. The fractions
that comprised the first peak (purified labeled probe) were pooled and
had a total of 1–10 x 106 cpm. The pooled
fractions were stored at -20 C until hybridization to the arrays.
Hybridization of cDNA probes to the Atlas Arrays
The arrays (Atlas Human cDNA Expression Array, CLONTECH Laboratories, Inc.) were prehybridized with 1.5 mg heat-denatured sheared salmon testes DNA in 10 ml ExpressHyb for 30 min at 68 C. Afterward, the prehybridization solution was replaced by 5 ml hybridization solution containing equivalent amounts of labeled probe derived from nodular tissue and the corresponding ST, respectively. After overnight hybridization of the membranes in roller bottles under continuous agitation, the arrays were washed three times with 100 ml wash solution I (2 x SSC/1% SDS) and twice with 100 ml wash solution II (0.1 x SSC/0.5% SDS) at 68 C. Then the membranes were wrapped in plastic and exposed to phosphorimager screens. The screens were scanned with a BAS Reader 3000 (Fuji Photo Film Co., Ltd., Tokyo, Japan) after 2-d exposure. To allow reuse of the membranes (not more than twice), stripping of the probes was carried out according to the manufacturer’s instructions. The membranes were placed into a boiling 0.5% SDS solution for 15 min. Afterward, the efficiency of stripping was checked with a Geiger hand counter. Then the membranes were rinsed in wash solution I (2 x SSC/1% SDS), wrapped in plastic, and exposed to phosphorimager screens to ensure complete removal of radioactive label before they were stored at -20 C until reuse. Due to limitations in the amount of tissue and RNA available, we investigated three CTN and three AFTN with one array for each sample. Chen et al. (11) repeated their experiments using the Atlas Human cDNA Expression Arrays at least three times and found only low variation in repeating the method. In general, overall variation is more likely to be due to biological variation than to variation within the assay. A membrane hybridized with a cDNA derived from a nodular tissue was hybridized with a cDNA from an ST during the second hybridization, and a membrane hybridized with a cDNA from an ST was hybridized with a cDNA derived from nodular tissue, respectively.
Analysis of the arrays
To investigate the differential gene expression between CTN,
AFTN, and their corresponding ST, respectively, Tagged Image File
Format files containing autoradiographic images of the arrays
were analyzed using AIDA Array Compare software (Raytest
Isotopenmeßgeräte GmbH, Straubenhardt, Germany). First
the background was subtracted from the hybridization signals, and then
the arrays were normalized to the expression levels of the housekeeping
gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to correct for
errors arising from unequal labeling of probes due to different amounts
of RNA, different efficiencies of the labeling reactions, or
differences in hybridization efficiencies. The expression levels of the
genes were compared in 2 ways: 1) the ratio nodular tissue/ST was
calculated; and 2) the difference in nodular tissue-ST was determined
(an example is shown in Fig. 1). We
performed the evaluation of the expression levels in these 2 ways
because not all expression patterns can be described as a ratio
(e.g. a gene is only expressed in 1 of the 2 compared
tissues). To compare the results of the 2 methods of analysis a ranking
was performed according to the following criteria: 1) genes with ratios
of 5 or more and 0.2 or less = 10 points, genes with ratios of 3
or more and 0.3 or less = 6 points, and genes with ratios of 2 or
more and 0.5 or less = 4 points; and 2) genes with differences of
4800 arbitrary units or more and -4800 arbitrary units or less =
10 points, genes with differences of 3200 arbitrary units or more and
-3200 arbitrary units or less = 6 points, and genes with
differences of 1600 arbitrary units or more and -1600 arbitrary units
or less = 4 points. To better distinguish between strong and weak
differential gene expression we performed the ranking with 10, 6, and 4
points. Furthermore, an error caused by a low signal to noise ratio can
be minimized. The points of a gene reached in the 3 CTN and AFTN were
added, and the sum of points reached according to the 2 ways of
analysis (ratio or difference) were multiplied (compare Fig. 1). The
multiplication of the points favors genes that showed differential
expression in both types of analysis. The array results of the first 15
genes with the highest differential gene expression (product of the
points reached) were evaluated by real-time PCR after RT of total RNA
of 10 CTN and 10 AFTN and their corresponding surrounding tissue to
investigate the reliability of the expression array data.
|
One microgram of total RNA for the CTN, AFTN, and their corresponding ST was reverse transcribed in a 20-µl reaction. The reaction mixture consisted of 5 x first strand buffer [250 mM Tris-HCl (pH 8.3), 375 mM KCl, and 15 mM MgCl2; Life Technologies, Inc., Karlsruhe, Germany), 0.5 mM dNTPs, 5 mM dithiothreitol (Life Technologies, Inc.), 15 U Prime RNase Inhibitor (PeqLab, Erlangen, Germany), 2.5 µM oligo(deoxythymidine)18, and 200 U Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.). After RT the cDNAs were diluted 1:5 in ribonuclease-free water.
Quantification by real-time PCR
The quantification of the 15 genes by real-time PCR was
performed using a LightCycler (Roche, Mannheim, Germany).
Oligonucleotide primers were designed to be intron spanning and were
purchased from BioTeZ (Berlin, Germany) and MWG Biotech
(Ebersberg, Germany). LightCycler hybridization probes were designed
and supplied by O. Landt (TIB MOLBIOL, Berlin, Germany). Sequences were
obtained from the GenBank database. The nucleotide sequences of the two
primers and the two hybridization probes are available on request.
First an optimal PCR reaction for all investigated genes was
established using the LightCycler-DNA Master SYBR Green I Kit
(Roche) according to the manufacturer’s instructions;
annealing temperatures and MgCl2 concentrations
were optimized to create a one-peak melting curve. Additionally, the
amplicons were checked by agarose gel electrophoresis for a single band
of the expected size. PCRs were processed through 40 cycles of a 3-step
PCR, including 0 sec of denaturation at 95 C, a 7-sec
template-dependent annealing phase, and a template-dependent elongation
at 72 C ranging from 8–25 sec for the different PCR amplicons. The
optimal PCR conditions for each target are shown in Table 2. Afterward, the PCR fragments were
cloned into the pGEM-T vector (Promega Corp., Madison,
WI), and the plasmids were sequenced to certify the specificity of the
PCR reactions. The quantification of the different targets in the CTN,
AFTN, and their corresponding ST was performed using the
LightCycler-DNA Master Hybridization Probes Kit (Roche). A
20-µl reaction consisted of 2 µl LightCycler DNA Master
Hybridization Probes (containing Taq DNA polymerase,
reaction buffer, dNTP mix (with deoxy-UTP instead of dTTP) and 10
mM MgCl2), additional
MgCl2 according to the optimization, 0.5
µM of each primer, 0.15
µM of each hybridization probe (3'-fluorescein
labeled and 5'-LightCycler Red 640 labeled), and 2 µl template.
Dilutions of the plasmids were used to generate calibration curves for
each template. The quantification of each template was performed in
triplicate in 1 PCR run. Furthermore, the corresponding nodules and ST
were measured in the same PCR run. To normalize for differences in the
amount of cDNA added to the reactions, quantification of GAPDH and
ß-actin was performed as an endogenous control. The differential
expression of the investigated genes was calculated as the ratios of
CTN/surrounding tissue and AFTN/surrounding tissue, respectively. The
determined ratios were also normalized to the ratio of the housekeeping
gene GAPDH and ß-actin. Differences between nodular and surrounding
tissue were evaluated by paired sample t tests (the data for
1 patient were paired) after testing for differences in the whole
dataset (PCR quantification in 10 AFTN and 10 CTN, and their respective
ST) with ANOVA. P 
0.05 was characterized as
statistically significant.
|
Atlas arrays
A representative set of cDNA expression arrays is shown in Fig. 2. The visual comparison of the
expression pattern of nodular (AFTN and CTN) and surrounding tissue
shows a high similarity, and some of the differences between nodular
and surrounding tissue are clearly visible. A high percentage of genes
expressed across all investigated tissues were transcription factors
and DNA-binding proteins (
25% of the expressed genes) located at
positions D1a–D7n. However, we found no genes in this group that show
a differential expression between CTN or AFTN and their corresponding
surrounding tissue. Another group that is characterized by a strong
expression in both nodular and surrounding tissue comprises genes
involved in stress response (B7h–7n). Also, these genes show no
differential expression between the nodular and surrounding tissues.
Among the group of intracellular signal transduction modulators and
effectors (20% of the expressed genes) that are located at
positions B1d–B7g, we found five genes that reached the criteria of
differential expression between CTN, AFTN and their surrounding tissue
described in Subjects and Methods. The tyrosine kinase
receptor HEK, TrkB receptor, Tyr phosphatase receptor-
, and
glycoprotein 130 showed a decreased expression in both CTN and AFTN
compared with their surrounding tissue (Table 3). The p97-MAPK showed a slightly
decreased expression in CTN compared with their ST (Table 3).
However, in two of three AFTN the expression was slightly increased
compared with that in their ST (Table 3). A further large group of
expressed genes consists of apoptosis-associated genes at
positions C1a–C5h. One gene of this group, clusterin, is characterized
by a strongly decreased expression in the AFTN and CTN 32 and 60
compared with their surrounding tissue (Table 3). The DNA repair
protein (position C6d) shows an increased expression in two of three
CTN and AFTN compared with their ST (Table 3). The expression of the
TGFß receptor III and c-myc, which belong to the group of
oncogenes and tumor suppressors, is decreased in both CTN and AFTN
compared with their surrounding tissue (Table 3). The cell cycle
control protein CLK-1 also shows a decreased expression in the nodular
tissues compared with the normal surrounding (Table 3). The
platelet-derived growth factor (PDGF) receptor (position E1j) is
characterized by a 3-fold lower expression in AFTN compared with their
ST, and the receptor of the CRF (position E2k) shows a 2- to 6-fold
increased expression in AFTN and a 2- to 3-fold increased expression in
CTN compared with their ST (Table 3). In the group of growth factors,
IGF-II (position F1a) and nerve growth factor (HBNF)-I are decreased up
to 6-fold in nodular tissue compared with the corresponding ST (Table 3). The monocyte chemotactic and activating factor (position F3a),
which belongs to the group of chemokines, shows decreased expression in
AFTN 110 and 122 compared with their ST and increased expression in CTN
32 and 60 compared with their ST (Table 3).
|
|
The 15 genes with the highest differential gene expression
determined according to the described criteria (see analysis for the
array data in Materials and Methods and in Fig. 1
) were
selected for confirmation of their differential expression by real-time
PCR. We have normalized expression of all transcripts to GAPDH and
ß-actin. There was no significant difference in the normalization to
ß-actin or GAPDH with regard to the transcripts with prominent
differential expression. Therefore, Table 4 shows the expression data after
normalization to GAPDH. In some cases the real-time PCR could not
confirm the results of the expression arrays. About 25% of the
transcripts determined by LightCycler PCR are considered false
positives because data from PCR and array analysis did not agree.
Differences changing the tendency of gene expression between nodular
and normal surrounding tissue were found in 19 comparisons. In AFTN 110
the real-time PCR indicated increased expression of tyrosine
phosphatase receptor-, glycoprotein 130, and DNA repair protein
XRCC1 in nodular tissue, in contrast to decreased or unchanged
expression in the expression arrays. The real-time PCR showed increased
expression of tyrosine kinase HEK and p97-MAPK in nodular tissue of
AFTN 122 in contrast to the expression array. Furthermore, unchanged or
decreased expression of tyrosine phosphatase receptor-, clusterin,
DNA repair protein XRCC1, and monocyte chemotactic and activating
factor was detected in AFTN 122 by real-time PCR, in contrast to the
arrays. In AFTN 123 the real-time PCR indicated decreased expression of
DNA-repair protein XRCC1, p97-MAPK, and monocyte chemotactic and
activating factor in contrast to increased expression detected by the
array analysis. In CTN 32 we found opposite results for PDGR receptor
and HBNF-1. TrkB, DNA, repair protein XRCC1, HBGF-8, and c-Myc showed
the opposite expression pattern in CTN 60 by real-time PCR in contrast
to the array data. In the CTN 61 the real-time PCR indicated increased
expression of tyrosine phosphatase receptor- in contrast to the
array analysis, which showed decreased expression. We were not able to
establish RT-PCR amplification to detect the CRF receptor, for which we
found expression in both AFTN and CTN as suggested by array analysis.
We tested seven primer pairs including the primers used by
CLONTECH Laboratories, Inc., to generate the CRF receptor
fragments spotted onto the arrays. Furthermore, we hybridized a human
thyroid cDNA library (RZPD, Berlin, Germany) with radioactively labeled
CRF receptor riboprobe derived from an RT-PCR fragment generated from
human brain cDNA. No significant homology to any spotted clone was
found.
|
In summary, we found a significantly lower expression of 11 genes of 15
in AFTN compared with their ST (Table 4). Furthermore, 9 of 15 genes
had significantly lower expression in CTN compared with their ST (Table 4). Moreover, we compared clinical and mRNA expression data of the 20
patients to determine whether any particular change can be attributed
to treatment of the patients. However, we could not find any
correlation between the patient’s treatment with
T4 or methimazole and the expression of the
investigated genes.
Discussion
The array technology is a powerful tool to assess the expression
of a large number of different genes in a single experiment. Therefore,
we used this technology to investigate gene expression in cold and hot
thyroid nodules compared with that in their surrounding normal tissue
to identify differentially expressed genes that might be involved in
nodular development and classes of proteins most frequently affected by
nodular development. We used low density nylon membranes spotted with
588 different genes in duplicate that represent a broad, but
restricted, range of protein classes. However, we found differences in
the expression pattern between the spotted gene groups. Despite the
relatively low number of spotted genes we found 15 genes in 3 AFTN and
3 CTN that showed a clear differential expression compared with their
surrounding tissue. Among the different genes present on the arrays,
transcripts for signal transduction proteins
(modulators/effectors/intracellular transducers) and extracellular cell
signaling and communication proteins were predominantly affected (Table 5). On the other hand, we found no
differentially expressed genes in the group of DNA synthesis, repair,
and recombination proteins or in the group of transcription factors and
general DNA-binding proteins and only 1 gene in the group of
apoptosis-associated genes (Table 5).
|
Seventy-five percent of the expression data determined by real-time PCR confirmed the ratios of differential gene expression determined by the cDNA expression arrays. The cases for which the ratios determined by the two methods are subject to strong variation (i.e. HBNF-1 and TrkB receptor) demonstrate the importance of a verification of the expression array data by a second method. Errors in the expression arrays might be due to an inhomogeneous hybridization and weak signals of genes with a low expression level. Furthermore, the cDNA probes of the three CTN, the three AFTN, and their ST were only hybridized once to the arrays, in contrast to a determination in triplicate by real-time PCR, since Chen et al. (11) found a good reproducibility of the Atlas Human cDNA Expression Arrays. However, the additional investigation of seven CTN and AFTN by real-time PCR indicates a significantly decreased expression of 11 genes in the nodular tissues, which confirms the expression array data and indicates that the expression arrays are a good choice to get an overview of the differentially expressed genes within a large group of genes.
Expression data for the CRF receptor have only been observed on the expression arrays because we were not able to obtain a RT-PCR product with a number of primer pairs that worked with mRNA from brain tissue. Moreover, hybridization of an in vitro CRF receptor RNA to a thyroid cDNA library did not detect positive clones. It is, therefore, too speculative to discuss the results of the cDNA expression arrays with regard to differences in the expression of the CRF receptor. Only Tg showed increased expression in AFTN compared with their ST. The detection of increased Tg mRNA in AFTN is probably due to increased function of the AFTN and illustrates the reliability of the real-time PCR measurements.
We found significantly decreased expression of 11 of 15 genes verified
by real-time PCR in AFTN and decreased expression of 9 of 15 genes
verified by real-time PCR in CTN compared with their surrounding
tissue. Nine of the 11 genes verified for their differential expression
belong to a group of proteins involved in signal transduction either as
receptors (i.e. PDGF receptor, TGFß receptor III, tyrosine
kinase receptor HEK, tyrosine kinase TrkB receptor, and tyrosine
phosphatase receptor-) or signal transmitting molecules
(e.g. glycoprotein 130, P97-MAPK, IGF-II, and HBNF-1). These
transcripts are predominantly down-regulated in thyroid nodules, which
might be interpreted as a reduction in signaling diversity and partial
loss of the regulative system of epithelial cells. Our observations
support the hypothesis of a restricted diversity of gene expression in
tumor cells as indicated by earlier experiments (12).
Furthermore, the decreased expression of 40 genes of 290 expressed
genes in CTN compared with their surrounding tissue corresponds to
findings of a lower mRNA complexity in undifferentiated tissue compared
with differentiated tissue (13, 14). Although a
predominant down-regulation of signal transduction proteins has
recently been shown for carcinoma of the larynx (15), this
is a new finding for benign epithelial tumors. As altered apoptosis is
a frequent feature of cancer (16) changes in this group of
genes might be markers to differentiate between benign and
dedifferentiated tumors. In this regard it is very interesting to note
that we only found a single gene in this group of genes (clusterin,
related to apoptosis) as differentially expressed in hot and cold
thyroid nodules.
Wegrowski et al. (17) found an enhancement of clusterin protein and mRNA expression by TGFß1 in porcine epithelial cells in primary culture. Our data concerning the clusterin mRNA expression in AFTN and CTN are in line with these findings. We found decreased expression of clusterin in AFTN (P < 0.01) and CTN compared with their surrounding tissue, and in a former investigation we found decreased expression of TGFß1 in AFTN and CTN (18). A number of physiological functions have been proposed for clusterin based on its distribution and in vitro properties (19). According to Koch-Brandt and Morgans (19), the decreased expression of clusterin in AFTN and CTN might indicate an increased number of cells undergoing apoptosis in both AFTN and CTN.
The detected significantly lower expression of c-myc in established hot thyroid nodules compared with their ST is in line with findings of Pirson et al. (20) showing a biphasic expression of c-myc in dog thyroid cells in primary culture; after TSH stimulation they found an increased expression of c-myc followed by a sharp down-regulation of the c-myc mRNA. In contrast to AFTN there are no significant differences in c-myc expression in CTN compared with their ST. These data suggest activation of different signaling cascades in AFTN and CTN.
We observed a significantly decreased expression of IGF-II in AFTN and CTN compared with their ST (P < 0.001). IGF-II is known to bind to the IGF-I receptor with a 10-fold weaker affinity than IGF-I (21). In a former study we found a slightly decreased expression of IGF-I in AFTN and CTN compared with their ST (18). In fact, growth factors such as IGF-I might play an important role during early clonal expansion. According to the hypothesis suggested by Dawson and Wynford-Thomas (22), aberrant growth factor expression is an early requirement for tumor promotion. Our results show a lack of difference (IGF-I) or a significantly lower concentration (IGF-II) of these growth factors in AFTN compared with their ST. This suggests that intrinsic signaling through a somatic mutation (e.g. TSHR for AFTN) might be the dominant factor that confers the selective growth advantage independent of a secondary mitogen (e.g. growth factors) in established hot thyroid nodules that are the cause of hyperthyroidism in these patients.
In both AFTN and CTN we found a significantly lower expression of TGFß receptor III (P < 0.01 and P < 0.001, respectively). The ligand of this receptor, TGFß1, has been shown to bind to thyroid epithelial cells (23) and inhibit growth by TSH receptor down-regulation (23, 24). Furthermore, a stimulation of differentiation by TGFß1 could be shown (25). In a previous study we found decreased expression of TGFß1 in CTN and AFTN compared with their surrounding tissue (18). TGFß1 transduces signals through two different types of receptors, type I and type II (26). The type III receptor might play a more indirect role, as it delivers the ligand to the signaling receptors (26), increases TGFß binding to the signaling receptors and enhances cell responsiveness to the ligand (27). The reduction of TGFß1, shown in a previous study, and the decreased expression of TGFß receptor III could intensify the higher proliferation in AFTN caused by increased cAMP, because TGFß1 is a negative regulator of thyroid cell proliferation.
In both AFTN and CTN we found significantly decreased expression of
PDGF receptor (P < 0.001). Data concerning the
expression of PDGF receptors in the thyroid are rare. Heldin et
al. (28) found active PDGF receptors in the human
anaplastic thyroid carcinoma cell line HTh 74. However, the
significance of PDGF receptor expression on thyroid epithelial cells is
not clear. Our observation of a decreased expression of PDGF
receptor- in nodular tissue is most likely an indication for a low
significance of signal transduction via PDGF in late stages of nodular
development.
The expression of the tyrosine kinase receptor HEK in thyroid epithelial cells and especially the significantly decreased expression in AFTN and CTN compared with that in their ST differs from findings of Wicks et al. (29). They found the tyrosine kinase receptor HEK to be expressed by some pre-B and thymic T cell lines, but to normally be undetectable in adult human tissues (29). Furthermore, the expression of HEK is supposed to be developmentally regulated, and inappropriate expression may contribute to oncogenesis (29). However, our data showing a decreased expression of HEK in AFTN and CTN do not suggest such a mechanism in thyroid tissue. Further studies should investigate whether the decreased expression of the HEK or the PDGF receptor indicates a cross-talk between the G protein-coupled TSHR and the PDGF receptor or the tyrosine kinase receptor HEK, as previously reported for other G protein and tyrosine kinase receptors (30, 31). In this case, the decreased expression of both the PDGF receptor and HEK could be a response to limit synergy with the cascade that primarily causes the aberrant proliferation (TSHR signaling in AFTN).
In AFTN we found a significantly decreased expression of the tyrosine kinase TrkB, the receptor for brain-derived neurotropic factor and neurotropin-4. McGregor et al. (32) showed a distinct immunoreactivity of TrkB within a subset of normal C cells. In contrast, no distinct receptor staining was observed in non-C cell areas of the thyroid (32). Therefore, it is likely that the expression data that we determined are due to the expression of TrkB in the C cells of the thyroid. The increased proliferation of the thyroid epithelial cells in AFTN (33) and most likely also in CTN may increase the ratio between thyroid epithelial cells and C cells, which might be the cause of a relative decrease in TrkB expression in AFTN.
In conclusion, the down-regulation of several signal transducing components seems to reflect the disturbed signaling system in AFTN and CTN, which is probably caused by dominance of a specific signaling pathway (e.g. cAMP signaling in case of AFTN). Although a similar effect is seen for CTN, its primary cause is currently unknown. None of the nodules harbor a Ras or a TSHR mutation (7). Moreover, for CTN the pattern of changes for this limited number of genes present on the arrays is not qualitatively different from that for AFTN, and therefore, it does not suggest a disturbance of a specific candidate signaling pathway. However, the similarity of the gene expression pattern of AFTN and CTN might be attributed to a common property of AFTN and CTN, i.e. increased proliferation. As G protein-coupled receptor signaling also involves the use of signaling molecules of other signaling pathways (30, 31), down-regulation of tyrosine kinase signaling (e.g. PDGF receptor and Tyr kinase receptor HEK) could be a response to limit synergy with the cascade that primarily causes aberrant proliferation. Moreover, the qualitatively similar pattern of changes in this limited number of genes in AFTN and CTN could suggest a dominant activation of a specific signaling cascade in CTN similar to the constitutively activating mutations in AFTN.
Acknowledgments
Footnotes
This work was supported by a grant from the Deutsche Forschungsgemeinschaft (DFG/Pa423/10-1) and the BMBF Interdisciplinary Center for Clinical Research at University of Leipzig (01KS9504, Projects B10 and B14).
Abbreviations: AFTN, Autonomously functioning thyroid nodules; CTN, cold thyroid nodules; dCTP, deoxy-CTP; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HBNF-1, nerve growth factor-1; PDGF, platelet-derived growth factor; ST, surrounding tissues.
Received March 14, 2001.
Accepted June 26, 2001.
References
) mutations
in 75 toxic thyroid nodules: detection by denaturing gradient gel
electrophoresis. J Mol Med, in press
This article has been cited by other articles:
 |
|
 |
|
  K. Krause, B. Jessnitzer, and D. Fuhrer Proteomics in Thyroid Tumor Research J. Clin. Endocrinol. Metab., August 1, 2009; 94(8): 2717 - 2724. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Fujarewicz, M. Jarzab, M. Eszlinger, K. Krohn, R. Paschke, M. Oczko-Wojciechowska, M. Wiench, A. Kukulska, B. Jarzab, and A. Swierniak A multi-gene approach to differentiate papillary thyroid carcinoma from benign lesions: gene selection using support vector machines with bootstrapping Endocr. Relat. Cancer, September 1, 2007; 14(3): 809 - 826. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Eszlinger, K. Krohn, A. Kukulska, B. Jarzab, and R. Paschke Perspectives and Limitations of Microarray-Based Gene Expression Profiling of Thyroid Tumors Endocr. Rev., May 1, 2007; 28(3): 322 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. L. Griffith, A. Melck, S. J.M. Jones, and S. M. Wiseman Meta-Analysis and Meta-Review of Thyroid Cancer Gene Expression Profiling Studies Identifies Important Diagnostic Biomarkers J. Clin. Oncol., November 1, 2006; 24(31): 5043 - 5051. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.-C. Gerard, M. Boucquey, M.-F. van den Hove, and I. M. Colin Expression of TPO and ThOXs in human thyrocytes is downregulated by IL-1{alpha}/IFN-{gamma}, an effect partially mediated by nitric oxide Am J Physiol Endocrinol Metab, August 1, 2006; 291(2): E242 - E253. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Maier, H. van Steeg, C. van Oostrom, S. Karger, R. Paschke, and K. Krohn Deoxyribonucleic Acid Damage and Spontaneous Mutagenesis in the Thyroid Gland of Rats and Mice Endocrinology, July 1, 2006; 147(7): 3391 - 3397. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Krohn, M. Eszlinger, R. Paschke, I. Roeder, and E. Schuster Increased power of microarray analysis by use of an algorithm based on a multivariate procedure Bioinformatics, September 1, 2005; 21(17): 3530 - 3534. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Krohn, D. Fuhrer, Y. Bayer, M. Eszlinger, V. Brauer, S. Neumann, and R. Paschke Molecular Pathogenesis of Euthyroid and Toxic Multinodular Goiter Endocr. Rev., June 1, 2005; 26(4): 504 - 524. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Jarzab, M. Wiench, K. Fujarewicz, K. Simek, M. Jarzab, M. Oczko-Wojciechowska, J. Wloch, A. Czarniecka, E. Chmielik, D. Lange, et al. Gene Expression Profile of Papillary Thyroid Cancer: Sources of Variability and Diagnostic Implications Cancer Res., February 15, 2005; 65(4): 1587 - 1597. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Eszlinger, K. Krohn, K. Berger, J. Lauter, S. Kropf, M. Beck, D. Fuhrer, and R. Paschke Gene Expression Analysis Reveals Evidence for Increased Expression of Cell Cycle-Associated Genes and Gq-Protein-Protein Kinase C Signaling in Cold Thyroid Nodules J. Clin. Endocrinol. Metab., February 1, 2005; 90(2): 1163 - 1170. [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 |
