Restoration of Iodide Uptake in Dedifferentiated
Thyroid Carcinoma: Relationship to Human Na?/I?
Symporter Gene Methylation Status*
GOPALAKRISHNAN M. VENKATARAMAN, MUSTAFA YATIN, REGINA MARCINEK,
AND KENNETH B. AIN
Thyroid Cancer Research Laboratory, Medical Service, Veterans Affairs Medical Center, Lexington,
Kentucky 40511; and the Division of Endocrinology and Molecular Medicine, Department of Internal
Medicine, University of Kentucky Medical Center, Lexington, Kentucky 40536-0084
cannot sufficiently concentrate therapeutic radioiodide, is a terminal
disease without any effective systemic treatment or chemotherapy.
thyroid carcinoma could be consequent to methylation of DNA in
critical regulatory regions and could be reversed with chemical de-
methylation treatment. Analysis of hNIS messenger ribonucleic acid
(mRNA) expression in 23 tumor samples revealed that although loss
of this expression corresponded to loss of clinical radioiodide uptake,
some thyroid carcinomas with hNIS mRNA expression did not con-
centrate iodide, suggesting additional posttranscriptional mecha-
nisms for loss of hNIS function. In addition, analysis of DNA meth-
ylation in CpG-rich regions of the hNIS promoter extending to the
first intron failed to define specific methylation patterns associated
with transcriptional failure in human thyroid tumor samples. In
seven human thyroid carcinoma cell lines lacking hNIS mRNA, treat-
ment with 5-azacytidine or sodium butyrate was able to restore hNIS
mRNA expression in four cell lines and iodide transport in two cell
lines. Investigation of methylation patterns in these cell lines re-
vealed that successful restoration of hNIS transcription was associ-
ated with demethylation of hNIS DNA in the untranslated region
within the first exon. This was also associated with restoration of
expression of thyroid transcription factor-1. These results suggest a
role for DNA methylation in loss of hNIS expression in thyroid car-
cinomas as well as a potential application for chemical demethylation
therapy in restoring responsiveness to therapeutic radioiodide.
(J Clin Endocrinol Metab 84: 2449–2457, 1999)
iodide symporter (NIS) located in the basolateral membrane
of thyroid follicular cells (1). This iodide-concentrating abil-
ity of thyroid follicular cells is exploited for the treatment of
differentiated thyroid epithelial carcinomas, using therapeu-
tic dosages of131I. Loss of iodide-concentrating ability in the
face of distantly metastatic disease results in significant mor-
bidity and mortality for around 10% of patients with differ-
entiated thyroid cancers (2). In addition, anaplastic thyroid
cancers, which are unable to take up radioactive iodide and
do not respond to systemic chemotherapies, are invariably
as well as the exon-intron organization have been revealed
by Smanik et al. (3, 4). We reported the cloning and charac-
terization of a 1.3-kb region of the upstream regulatory re-
gion and defined a minimal essential hNIS promoter that
shows tissue-specific expression in a human thyroid cell line
HE INITIAL step in the synthesis of thyroid hormone is
(5). Other investigators have further evaluated hNIS pro-
moter constructs (6, 7). It is possible that alterations in hNIS
expression, responsible for the loss of iodide-concentrating
ability in human thyroid cancer metastases, may correspond
to changes in hNIS promoter activity. This may be similar to
the loss of E-cadherin expression demonstrated in human
thyroid cancer cell lines, correlating to methylation of CpG
has CpG-rich regions as well as additional CpG islands
downstream from the transcription start site, DNA methyl-
ation may be responsible for alterations in hNIS expression.
Nearly half of all human genes have CpG islands associ-
ated with transcriptional start sites. Unmethylated CpG is-
lands are seen in highly transcribed genes, whereas heavily
methylated CpG islands inhibit transcription (9). Although
overall DNA methylation is often decreased in cancers, CpG
islands in critical gene promoter regions can become hyper-
methylated, resulting in loss of gene expression (10). Such
methylation may be effective in silencing gene expression
despite variable degrees of CpG site methylation from 20–
100% (11). Laboratory and clinical studies have suggested
that chemical agents may demethylate these regions and
restore gene expression. Examples include use of 5-azacyti-
dine to restore expression of O6-methylguanine-DNA meth-
yltransferase in human cervical, brain, and colon carcinomas
in human leukemic cells (14); and sodium butyrate to induce
PRL receptor expression in human breast cancer cells (15).
Received January 11, 1999. Revision received March 16, 1999. Ac-
cepted March 22, 1999.
Address all correspondence and requests for reprints to: Kenneth B.
Ain, M.D., Thyroid Nodule and Oncology Clinical Service, Division of
Endocrinology and Molecular Medicine, Department of Internal Med-
icine, Room MN520, University of Kentucky Medical Center, 800 Rose
Street, Lexington, Kentucky 40536-0084. E-mail: firstname.lastname@example.org.
* This work was supported by NCI Grant CA-58935, V.A. Merit
Review 596-0003, and the Ephraim McDowell Cancer Research Foun-
dation (Lexington, KY).
The Journal of Clinical Endocrinology & Metabolism
Copyright © 1999 by The Endocrine Society
Vol. 84, No. 7
Printed in U.S.A.
In this study we tested the hypothesis that methylation of
the characterized hNIS promoter and potentially regulatory
downstream regions correlate with the loss of hNIS messen-
loss of iodide uptake in samples of thyroid tumor tissues. In
addition, using human thyroid carcinoma cell lines and pu-
loss of hNIS mRNA expression and functional activity mea-
sured as iodide uptake.
Materials and Methods
Cell lines and human tissues
The human thyroid cell lines used in this study were MRO87 and
WRO82 (both follicular carcinomas, provided by G. J. F. Juillard, Uni-
versity of California-Los Angeles School of Medicine), NPA’87 (papil-
lary carcinoma, from Juillard), KAT-5 and KAT-10 [both papillary car-
cinomas, from our laboratory (16)], KAK-1 [benign follicular adenoma,
from our laboratory (17)], and KAT-7 (benign follicular hyperplasia,
from our laboratory). Cultures were previously treated with medium
containing d-valine (18) and cis-4-hydroxy-l-proline (19) to ensure the
absence of fibroblasts. Human thyroid tissues were obtained from fresh
surgical samples (approved by the University of Kentucky institutional
review board). Some tumor samples were supplied by the Cooperative
Human Tissue Network (Philadelphia, PA), and some were obtained
from surgical samples at the Clinical Center, NIH (Bethesda, MD; under
an approved protocol).
Cell lines for evaluation of iodide uptake and hNIS expression were
grown in phenol red-free RPMI 1640 with 5% FBS, 100 nmol/L sodium
selenite, and 0.1 nmol/L bovine TSH (basal medium) (20). They were
plated at a density of 3–5 ? 104cells/9.4 cm2in triplicate in basal
medium and grown for 2–3 days at 37 C in 5% CO2. They were treated
with dimethylsulfoxide (25 ?mol/L daily for 3 days), sodium butyrate
(0.5 or 1.0 mmol/L), phenylacetate (pH 7.0; 5 or 10 mmol/L), or 5-aza-
cytidine (0.5 or 1.0 ?mol/L in 25 ?mol/L dimethylsulfoxide daily for 3
days) until control cells were 80% confluent (3–4 days), then changed to
fresh basal medium and grown for an additional 24 h.
Analysis of CpG content in the hNIS gene sequence
The hNIS gene sequence of the 5?-flanking region (5) and the con-
tiguous transcribed region extending up to the first intron (3, 4) were
analyzed using WINDOW and STATPLOT computer programs (Ge-
netics Computer Group, Madison, WI) to denote CpG dinucleotide
Nucleic acid isolation and amplification
Total RNA and genomic DNA from normal human thyroid, thyroid
tumors, and cell lines (treated with the agents described above) were
isolated by the acid-guanidinium-phenol-chloroform method (21). All
surgical samples were snap-frozen and stored at ?80 C until processed
by homogenization in Trizol reagent (Life Technologies, Gaithersburg,
MD) while still frozen. Complementary DNA (cDNA) was synthesized
from 1.0 ?g total RNA using Moloney murine leukemia virus reverse
Inc., Palo Alto, CA). Each 50 ?L PCR vessel contained 60 mmol/L
Tris-HCl (pH 9.0), 15 mmol/L ammonium sulfate, 3.5 mmol/L MgCl2
[1.5 mmol/L for human thyroid transcription factor-1 (hTTF-1)], 250
?mol/L deoxy-NTPs (Boehringer Mannheim, Indianapolis, IN), 0.2
?mol/L of each primer pair, 1 U AmpliTaq DNA polymerase (Perkin
Elmer, Norwalk, CT), 0.2 ?g TaqStart Antibody (CLONTECH Labora-
tories, Inc.), and 3% cDNA. ?-Actin amplification primers (Stratagene,
La Jolla, CA) confirmed cDNA integrity, purity, and template equiva-
lence for semiquantification. PCR primers (upstream 5? to 3?, down-
stream 5? to 3?, in all cases) used for amplification were CTGCCCCA-
GACCAGTACATGCC/TGACGGTGAAGGAGCCCTGAAG for hNIS
(5) [to amplify a coding region spanning four introns (4) yielding a
303-bp product from cDNA] and AAGTCCAGCATTGCGGCACA/
GAGGGAAGTGCTTATGGTCC for PAX-8 (22) (to amplify a 329-bp
product). Amplification conditions for hNIS and PAX-8 were denatur-
ation at 95 C for 5 min, 40 cycles of 20 s at 95 C and 60 s at 68 C, followed
by extension at 72 C for 3 min. The hTTF-1 product was amplified with
intron-spanning primers, GCCGTACCAGGACACCATGAG/CAGG-
ditions were 95 C for 5 min; 45 cycles of 95 C for 20 s, 60 C for 60 s, and
72 C for 30 s; followed by extension at 72 C for 3 min. The RT-PCR
products were resolved on 2% agarose gels and visualized by ethidium
Methylation-specific PCR analysis
This method uses PCR primer pairs to distinguish methylated from
unmethylated DNA in bisulfite-modified target DNA, in which bisulfite
converts unmethylated cytosines to uracil (23, 24). Genomic DNAs from
normal and tumoral human thyroid tissues and cell lines were isolated
by standard techniques (21), and 1.0-?g aliquots were denatured by
3.0 mol/L sodium bisulfite (pH 5.0 under mineral oil for 16 h at 50 C).
Modified DNA was purified on a resin column (QIAGEN, Chatsworth,
CA) and further treated with 0.3 n NaOH for 5 min before ethanol
precipitation. The PCR mixture contained 16.6 mmol/L ammonium
sulfate, 67 mmol/L Tris-HCl (pH 8.8), 6.7 mmol/L MgCl2, 10 mmol/L
?-mercaptoethanol, 1.25 mmol/L deoxy-NTPs, 0.2 ?L TaqStart anti-
body, 1 U AmpliTaq DNA polymerase, 10 pmol each of sense and
antisense methylation-specific primers, and 50 ng bisulfite-modified
DNA target. Primers used for analysis of the hNIS promoter CpG island
methylation were selected for cytosine-rich regions containing CpG
dinucleotides near the 3?-end of the primers, hNIS-MET-P (sense, 5? to
3?, TTAGGTTTGGAGGCGGAGTCGC; antisense, 5? to 3?, ACCGAC-
TATCTATCCCTCTCCCTAAACG) for a 143-bp product from methyl-
ated DNA and hNIS-UNMET-P (sense, 5? to 3?, TTGTTTTTAGGTTT-
GGAGGTGGAGTTGT; antisense, 5? to 3?, CAACCAACTATCTA-
TCCCTCTCCCTAAACA) for a 151-bp product from unmethylated
genomic DNA. Additional sets of primers were similarly designed to
analyze further downstream elements. They were hNIS-MET-L (sense,
AAAACGAATACG) for a 265-bp product, hNIS-UNMET-L (sense,
TAGGATAGATAGATAGTAGGGGTGGAT; antisense, CTCCACAAC-
CTCCATAAAAACAAATACA) for a 275-bp product, hNIS-MET-C
(sense, AGGTCGTGGAGATCGGGGAAC; antisense, ACGATAAAC-
CTCCGACGACACG) for a 242-bp product, and hNIS-UNMET-C
(sense, TTATGGAGGTTGTGGAGATTGGGGAAT; antisense, CATAA-
CAATAAACCTCCAACAACACA) for a 252-bp product. The amplifi-
cation conditions were Taq polymerase activation at 95 C for 5 min and
40 cycles of denaturation at 94 C for 20 s, annealing at 60 C for 30 s, and
polymerization at 72 C for 30 s. Methylation-specific PCR products were
resolved by agarose gel electrophoresis and visualized by ethidium
bromide staining and UV transillumination.
Iodide uptake assay
Cell lines treated with differentiation agents and control cultures
were washed with 2 mL buffer containing 10 mmol/L HEPES (pH 8.3),
5.5 mmol/L glucose, 5.4 mmol/L KCl, 1.3 mmol/L CaCl2, 0.4 mmol/L
Na2HPO4, 0.44 mmol/L KH2PO4, and either 137 mmol/L NaCl (buffer
A) or 100 mmol/L choline chloride (buffer B). After a 60-min incubation
in the same buffer supplemented with Na[125I] (1.0 ?Ci/2 mL) and 1.0
NaOH and ?-counted (5). A parallel set of dishes, similarly plated and
treated, was used for normalization of uptake activity, using the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (25) as an
index of cell viability. ? counts of cells incubated in buffer B were
subtracted from counts of cells incubated in buffer A under correspond-
ing conditions to account for nonspecific binding of radioiodide.
Clinical radioiodine uptake
Assessment of radioiodine uptake in clinical tumor samples was
based upon the results of131I whole body scans (using 5 mCi131I tracer
doses) performed 6–8 weeks after excision of the primary tumor during
2450 VENKATARAMAN ET AL.
JCE & M • 1999
Vol 84 • No 7
surgical thyroidectomy. The presence of radioiodine uptake in meta-
static tumor deposits was presumed to be indicative of positive radio-
iodine uptake in the primary tumor sample. This is based upon the
iodide uptake is far less common than tumor dedifferentiation. The
absence of radioiodine uptake in palpable or radiologically discernible
tumor metastases was presumed to reflect loss of radioiodine uptake in
the primary tumor sample. This designation is probably correct; how-
ever, it is possible that metastases may have less functionality than their
parent tumors. In the absence of persistent tumor metastases, the as-
sessment of radioiodine uptake was not possible. Some tumor samples
were obtained from recurrent tumors that had been documented to lack
radioiodine uptake on the basis of previous whole body scanning.
CpG dinucleotide distribution in the context of the
regions up to the first intron were analyzed for the presence
of CpG islands. The frequency plot (Fig. 1) shows a region of
the promoter surrounding the transcription start site (5) and
extending upstream for about 100 bp to be rich in CpG
dinucleotides (region P). This was the only upstream region
in the characterized promoter that was CpG rich. Sequence
comparison revealed that this region shared significant ho-
mology to the rat NIS promoter region (26). This region was
selected for analysis of methylation status in clinical tumors
and cell lines. Additional CpG-rich sequences are present
downstream from this region, extending to the first intron.
Regions L and C, selected for methylation analysis, corre-
sponded to CpG-rich sequences in the hNIS leader and cod-
ing regions, respectively, within the first exon.
hNIS mRNA expression and clinical iodide uptake in
Primary thyroid tumors were analyzed by RT-PCR for the
expression of hNIS mRNA (Fig. 2 and Table 1). Messenger
RNA for hNIS was poorly expressed in all 6 tall cell papillary
carcinomas, ranging from undetectable in 4 tumors to mod-
erately positive in 2 tumors (5 cases shown in Fig. 2). In
contrast, hNIS mRNA expression was clearly detectable in
both follicular carcinomas, 9 of 10 typical papillary carcino-
mas (variable levels of expression), and both anaplastic thy-
roid carcinomas. Two of the 3 Hurthle cell carcinomas were
negative for hNIS mRNA expression. Among the 19 tumor
samples that were able to be assessed for clinical radioiodide
uptake, 13 cases exhibited concordance of hNIS mRNA ex-
pression with whole body scanning (7 with concordant pos-
itive findings and 6 with concordant negative findings). In 6
cases (dispersed between all of the tumor histologies except
follicular carcinoma) there was no detectable radioiodine
uptake on whole body scanning despite detectable hNIS
mRNA in the tumor sample. Analysis of thyroid transcrip-
tion factor mRNA expression in these discordant cases re-
vealed that all expressed PAX-8, and only 2 of the 6 cases
expressed TTF-1. As only 1 of 7 tumor samples, with con-
cordant positive radioiodine uptake and hNIS mRNA ex-
pression, lacked TTF-1 mRNA expression, loss of this factor
may contribute to the loss of hNIS function, but is not totally
Methylation status and hNIS mRNA expression in
The NIS promoter was only faintly methylated in normal
human thyroid tissues and pooled human white blood cells.
As described in the previous section, hNIS mRNA was un-
detectable in 4 of 6 tall cell papillary carcinomas and was low
FIG. 1. CpGdinucleotidefrequencyinthehNISpromoterregion.TheDNAsequenceofthehNISpromoteranditscontiguoustranscribedregion
(up to the first intron) were assessed by computer analysis. Nucleotide positions are in reference to the adenosine residue of the ATG translation
start site. The bold arrow indicates the position of a TATA box-like element. The shaded box (P) denotes the region of the hNIS promoter chosen
for methylation analysis. The open box (L) and the solid box (C) denote the leader and coding regions, respectively, of the first exon that were
analyzed for methylation status.
FIG. 2. hNIS mRNA expression in tall cell papillary thyroid carci-
noma. RT-PCR products were resolved on a 2% agarose gel and
visualized by ethidium bromide staining. PCR substrates are: lane 1,
no cDNA (negative control); lane 2, normal thyroid (positive control);
lanes 3–7, tall cell papillary thyroid carcinomas (samples 11–15, Ta-
ble 1); and lane 8, Life Technologies 1-kb Plus DNA ladder.
IODIDE UPTAKE RESTORATION IN THYROID CANCER 2451
methylated in all but 1 case, but displayed lower signal
intensities for the methylated amplification product. A CpG-
rich segment of the coding region (region C) displayed het-
erogeneous methylation among tall cell tumors without any
particular correlation to hNIS mRNA expression. However,
of the 10 papillary thyroid tumors, there was no apparent
association of methylation, in regions P, L, or C, with loss of
hNIS mRNA expression. Likewise, although both follicular
carcinomas expressed hNIS mRNA, they each showed dif-
ferent methylation patterns between the regions. All 3 cases
of Hurthle cell carcinoma had unmethylated hNIS promoter
regions and variably methylated L and C regions, but only
1 of them expressed hNIS mRNA.
Treatment of thyroid carcinoma cell lines to restore
expression of hNIS mRNA and effect on iodide uptake
Seven human thyroid neoplastic cell lines, devoid of hNIS
mRNA expression under basal monolayer conditions, were
treated with putative chemical demethylation agents in an
attempt to restore hNIS expression. These cell lines were
derived from three papillary carcinomas (NPA’87, KAT-5,
and KAT-10), two follicular carcinomas (WRO82 and
MRO87), and two benign follicular neoplasms (KAK-1 and
KAT-10) (5). Three different demethylation or redifferenti-
ation agents (viz. sodium butyrate, phenylacetate, and 5-
in one of the benign follicular adenomas under at least one
the hNIS mRNA reexpression in cell lines KAK-1 and
To investigate whether reexpression of NIS mRNA is suf-
ficient to restore hNIS function, i.e. iodide uptake, we treated
responding cells under the same conditions as those used to
restore mRNA expression and analyzed125I uptake activity.
Of the four responding cell lines tested, there was a greater
TABLE 1. Thyroid tissue analysis
Tumor histologySample no.
Methylation status by region
Typical papillary carcinoma1
Tall cell variant papillary carcinoma
Hurthle cell carcinoma
Normal thyroid tissue
Methylation: ??, distinctly positive; ?, moderately positive; faint, slightly positive; ?, negative. mRNA expression: ??, comparable to
normal thyroid; ?, moderate level; ?, negative. Clinical radioiodine uptake: ?, positive tumor uptake; ?*, assumed (patient euthyroid); ?, no
tumor uptake; NA, results not available.
FIG. 3. MethylationanalysisofthehNISpromoterinproximitytothe
TATA box (region P). Products of methylation-specific PCR analysis
of sodium bisulfite-modified genomic DNA from thyroid tumors using
a methylation-specific primer pair (MET) and a nonmethylated spe-
cific primer pair (UNMET) were electrophoresed on an agarose gel in
adjacent lanes. Lanes 1 and 22, Life Technologies 1-kb Plus DNA
ladder; lanes 2–21 (even-numbered lanes contain the 151-bp UNMET
product, and odd-numbered lanes contain the 143-bp MET product).
Lane pairs starting with 2–12 represent the reaction pairs of tall cell
papillary cancer samples 11–16, respectively (Table 1). Lane pairs
starting with 14–20 represent the reaction pairs for anaplastic car-
cinoma (Table 1, sample 22), negative control (no template DNA),
normal thyroid, and pooled human leukocyte DNA, respectively.
2452VENKATARAMAN ET AL.
JCE & M • 1999
Vol 84 • No 7
than 2-fold increase in uptake in KAK-1 cells (derived from
5-azacytidine compared to that in untreated cells (Fig. 5a).
However, no enhancement of uptake was seen using 0.5
?mol/L 5-azacytidine, even though reexpression of hNIS
mRNA was comparable under the two different concentra-
tions of 5-azacytidine (Fig. 4a). The iodide uptake activity in
NPA’87 cells (derived from a papillary carcinoma) was
slightly increased with 1.0 mmol/L sodium butyrate,
whereas 1.0 ?mol/L 5-azacytidine treatment resulted in
more than 15-fold increased uptake (Fig. 5b). As in the other
cell line, the differences in iodide uptake were noted despite
similar expression of hNIS mRNA (Fig. 4b), suggesting the
possible contribution of some other inducible factor.
Restoration of iodide uptake and demethylation of
The cell lines KAK-1, KAT-5, KAT-10, and NPA’87, in
which hNIS expression was restored, were grown under
basal and reexpression conditions, and the DNA were ana-
lyzed for their methylation status at the same three gene
revealed that the P region was unmethylated under all con-
ditions, basal or otherwise. Methylation of the L and C re-
gions under basal conditions was clearly evident in all four
cell lines (Fig. 6). The PCR product specific for unmethylated
DNA in the L and C regions was undetectable or merely
faintly present in the same cell lines, suggesting that the cell
populations were homogeneously methylated in these re-
gions. Treatment with 5-azacytidine was associated with de-
creased methylation at the L and C regions in all four cell
specific PCR products and de novo or increased expression of
the corresponding unmethylated PCR products to equal or
greater intensity than the methylated product bands. The
susceptibility of KAT-5 and KAT-10 cells to the demethyl-
ation effects of 5-azacytidine in the C region appeared less
than that in the L region. Although sodium butyrate treat-
ment was associated with reexpression of hNIS mRNA in
effect in KAT-5, analysis of methylation patterns of the
NPA’87 and KAT-10 responses to sodium butyrate failed to
in all three regions.
In the three cell lines that failed to express hNIS mRNA
despite treatment with 5-azacytidine, sodium butyrate, or
phenylacetate (MRO87, WRO82, and KAT-7), baseline meth-
ylation pattern analysis revealed that region P was not meth-
ylated, whereas regions L and C were homogeneously meth-
ylated. Treatment with 5-azacytidine did not affect the
baseline demethylated status in the P region of these cell
lines; however, the results in the L and C regions were dif-
ferent from those in the four responding cell lines. Region C
appeared methylated under basal conditions in the three cell
TABLE 2. Human thyroid cell line analysis
Cell line Culture additive
Methylation status by region
Lethal additive concentration for this cell line
Culture additives: AzaC, 5-azacytidine (low, 0.5 ?ol/L; high, 1.0 ?ol/L); NaB, sodium butyrate (low, 0.5 mol/L; high, 1.0 mmol/L); PhAc,
phenylacetate (low, 5.0 mol/L; high, 10 mol/L). Specific125I uptake (monolayer cultures): ?, positive; ?, negative. Methylation: ?, positive; ?,
negative. mRNA expression (hNIS, TTF-1, Pax-8): ?, positive; ?, negative; ND, not done.
IODIDE UPTAKE RESTORATION IN THYROID CANCER 2453
lines (analysis of WRO82 failed to reveal either methylated
or unmethylated products) and did not become demethyl-
ated in response to 5-azacytidine, except for minimal detec-
tion of an unmethylated product for WRO82 cells (a meth-
ylated product becomes clearly visible). The demethylation
response to 5-azacytidine in the L region was similar in
responsive and nonresponsive cell lines. The failure of 5-
azacytidine to effectively demethylate the C region distin-
guished cell lines that failed to reexpress hNIS mRNA from
those that regained such expression.
Comparison of hNIS reexpression to expression patterns of
TTF-1 and PAX-8
may be consequent to reexpression of one or more transcrip-
tion factor(s), we performed RT-PCR analysis for thyroid-
specific transcription factors, TTF-1 and PAX-8 (Table 2).
PAX-8 mRNA was expressed under all conditions tested in
all of the four cell lines that were able to reexpress hNIS
mRNA, whereas TTF-1 mRNA expression was found even
under basal conditions in cell lines NPA’87 and KAK-1 (data
not shown). Basal TTF-1 expression was undetectable in cell
lines KAT-5 and KAT-10, although TTF-1 mRNA expression
was induced by 5-azacytidine treatment. Likewise, phenyl-
acetate treatment induced TTF-1 mRNA expression only in
the KAT-5 cell line; however, sodium butyrate did not have
such an effect in either of the cell lines.
Treatment of metastatic thyroid carcinoma requires effec-
tive systemic agents. Due to the absence of applicable che-
motherapeutics, radioiodine therapy is the only efficacious
modality. The failure to respond to radioiodine portends
grave consequences and is an appropriate target for correc-
FIG. 4. Reexpression of hNIS mRNA in thyroid cell lines. Follicular
adenoma cell line, KAK1 (a). KAK-1 cells were treated in triplicate
with 5-azacytidine as described. The RT-PCR products were resolved
on a 2% agarose gel and visualized by ethidium bromide staining.
Lane 1, No cDNA; lanes 2–4, untreated; lanes 5–7, 0.5 ?mol/L 5-
azacytidine for 3 days (added each day); lanes 8–10, 1.0 ?mol/L
5-azacytidine for 3 days (added each day); lane 11, Life Technologies
1-kb Plus DNA ladder. b, Papillary carcinoma cell line, NPA’87.
NPA’87 cells were treated in triplicate with sodium butyrate or 5-aza-
cytidine as described. The RT-PCR products were resolved on a 2%
agarose gel and visualized by ethidium bromide staining. Lane 1, Life
Technologies 1-kb Plus DNA ladder; lane 2, normal human thyroid;
lanes 3–5, untreated; lanes 6–8, 1.0 mmol/L sodium butyrate for 3
days; lanes 9–11, 1.0 ?mol/L 5-azacytidine for 3 days (added each
FIG. 5. Restoration of iodide uptake in neoplastic thyroid cell lines.
The uptake values are normalized for cell viability, as determined by
the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide as-
say in a parallel set of plates. a, Follicular adenoma cell line, KAK-1.
The KAK-1 cells were treated with 5-azacytidine, and the iodide
uptake was measured in quadruplicate as described. b, Papillary
carcinoma cell line, NPA’87. The NPA’87 cells were treated with
quadruplicate as described.
2454 VENKATARAMAN ET AL.
JCE & M • 1999
Vol 84 • No 7
tion. Loss of hNIS gene expression appeared a likely cause
for the loss of iodide-concentrating ability; however, we
demonstrate that some thyroid cancers maintain expression
of hNIS mRNA despite loss of function, suggesting diverse
pathophysiology. This was of particular surprise for ana-
iodide uptake (5). In those tumors in which loss of hNIS
mRNA was observed, we attempted to explore potential
mechanisms, focusing on reversible etiologies.
Some investigators have attempted to restore iodide up-
take using retinoids. A nominal increase in iodide uptake
activity was reported in a thyroid follicular carcinoma cell
line, UCLA RO 82 W-1 (WRO82), treated with 13-cis-retinoic
acid (cRA) (27). Direct evidence of the effects of cRA on
reestablishing iodide uptake in dedifferentiated follicular
and papillary thyroid cancers was first reported by Simon et
al. (28). The latest details of their study revealed that only 14
patients (of 20 study patients) did not concentrate any ra-
dioiodine in metastatic tumors at baseline, with only 1 such
patient reestablishing distinct iodide uptake after cRA treat-
ment (an additional 3 patients gained weak uptake) (29). A
case report suggests a positive response to similar treatment
in a single patient (30). Alternatively, a minimal enhance-
ment of iodide uptake with interferon-? was suggested in
several human thyroid cancer cell lines in vitro (31). The
mechanism for effects on iodide transport is unknown for
both of those agents, although they suggest that loss of hNIS
activity may be a reversible phenomenon.
In view of the multiplicity of mechanisms causing loss of
iodide transport, we chose to further evaluate the subset of
tumors with apparent hNIS transcriptional failure and the
relationship to CpG island methylation in the region of the
hNIS promoter. This was of particular importance in tall cell
variant papillary thyroid cancers, because nearly half of such
patients lose clinical iodide transport (32, 33), and we show
this to be a consequence of hNIS transcriptional failure. The
ability of 5-azacytidine to induce hNIS mRNA as well as
iodide uptake in thyroid carcinoma cell lines devoid of basal
hNIS mRNA expression further implicated methylation as a
likely mechanism. In these cell lines, reversal of basal meth-
ylation of the L and C regions appeared to be associated with
of expression of hNIS mRNA in tall cell variant papillary
carcinoma tumors was not able to be assuredly explained by
such methylation patterns. Part of the reason may relate to
the heterogeneity of cell methylation patterns between cells
in the same culture. This may relate to the heterogeneity of
hNIS protein expression demonstrated in normal and ma-
lignant thyroid tissues (34, 35). A similar mechanism has
inherently heterogeneous as mixtures of tumor cells, fibro-
blasts, endothelial cells, smooth muscle cells, and infiltrating
host immune cells. It is also possible that the specific sites of
methylation responsible for loss of hNIS transcription, in or
analyzed in this study.
Alternative explanations for the loss of hNIS mRNA ex-
pression may relate to methylation of thyroid-specific tran-
scription factor genes causing loss of transcription factor
was suggested by the KAT-5 and KAT-10 responses to 5-
azacytidine treatment with acquisition of parallel TTF-1 and
hNIS mRNA expression. Failure to express sufficient TTF-1
and PAX-8 can result in decreased activity of the thyroglob-
ulin gene promoter in human thyroid carcinoma cells (37), a
likely feature of the hNIS gene. Much of this remains spec-
ulative, considering that additional thyroid-specific tran-
scription factors, such as TTF-2 (38, 39) and other poorly
characterized factors (40), have not been similarly analyzed.
FIG. 6. Methylation analysis of hNIS gene regions in cell lines reex-
pressing hNIS mRNA. Products of methylation-specific PCR analysis
of sodium bisulfite-modified genomic DNA from thyroid cell lines,
using two methylation-specific primer pairs (MET for regions L and
C) and two corresponding nonmethylated specific primer pair (UN-
MET for regions L and C) were electrophoresed on an agarose gel in
adjacent lanes. In all gels: lanes 1 and 22, Life Technologies 1-kb Plus
DNA ladder; lanes 2–7, triplicate pairs of cell lines under basal con-
ditions; lanes 8–19, triplicate pairs of cell lines in two different treat-
ment conditions; lanes 20 and 21, negative controls without template
DNA (all even-numbered lanes contain the respective UNMET prod-
ucts, and odd-numbered lanes contain the corresponding MET prod-
ucts). a, The cell line KAK-1 studied with primer pairs specific for
5-azacytidine at 0.5 and 1.0 ?mol/L, respectively. b, The cell line
identical to those in a. c, The cell line NPA’87 studied with primer
pairs specific for region L. Treatment conditions in lanes 8–13 and
at 1.0 ?mol/L, respectively. d, The cell line NPA’87 studied with
primer pairs specific for region C, with conditions identical to those
IODIDE UPTAKE RESTORATION IN THYROID CANCER2455
As additional, possibly complex, processes may affect post-
transcriptional hNIS function, there are multiple opportuni-
ties for gene methylation to reduce iodide transport. In this
way, a response to 5-azacytidine may suggest a role for
methylation in the absence of demonstration of the specific
There are several examples of DNA methylation altering
expression of thyroid-specific genes. In transgenic mice car-
rying the chloramphenicol acetyltransferase (CAT) gene un-
der control of a bovine thyroglobulin promoter, CAT ex-
pression was limited to the thyroid glands and was related
to thyroid-specific demethylation of the bovine thyroglob-
ulin promoter (41). In another example, the transformed rat
thyroid cell line, FRT, is unable to express its native TSH
Avvedimento et al. (43, 44) have shown that transformation
of a rat thyroid cell line, which activated the ras oncogene,
as well as loss of expression of a thyroid-specific trans-acting
factor (presumably TTF-1). Treatment with 5-azacytidine re-
stored both TTF-1 expression and thyroglobulin promoter
activity. Such cases provide evidence that thyroidal tissues
use methylation as a regulatory mechanism for gene expres-
sion, particularly in transformed phenotypes.
The potential to restore iodide transport in dedifferenti-
deliver tumoricidal radioiodide is not clear. Normal thyroid
tissue, stimulated by TSH, concentrates radioiodide at 1% of
the administered dose per g tissue. Differentiated thyroid
cancer metastases typically concentrate radioiodide at 0.06–
0.3% of the administered radioiodide dose/g tumor (45).
Calculations of the degree of radioiodide uptake and the
biological residence time needed for sufficient therapy of
thyroid cancer suggest that (employing an effective half-life
of at least 4.5 days) tumor destruction can be achieved de-
spite an uptake of only 0.1%, using administered activities of
300 mCi (46). The use of radioiodide dosimetric analysis to
verify upper safety margins of administered doses may per-
mit therapeutic doses exceeding 600 mCi (47), so that tumors
with less than 0.05% uptake may respond to treatment. For
this reason, restoration of hNIS activity sufficient to treat
thyroid cancer does not require hNIS expression to the levels
seen in normal human thyroid follicular cells.
Effective radioiodide therapy requires more than a func-
tional hNIS gene. There should be sufficient expression of
TSH receptors and downstream signal transduction machin-
ery to amplify hNIS expression when TSH levels rise. In
addition, failure to organify radioiodide compromises131I
residence time in thyroid carcinoma cells, permitting radio-
iodide efflux and insufficient radiation delivery. This was
rat thyroid cells, lacking endogenous NIS expression, with
levels of131I uptake in xenografts of these cells, they were
unable to obtain tumoricidal effects due to rapid radioiodide
efflux from lack of effective organification. It is possible that
demethylation therapy may be able to restore additional
critical functions, such as organification, downstream from
iodide transport. Further investigations should delineate
which aspects of radioiodide therapeutics are responsive to
We are grateful for the technical advice of Drs. J. R. Graff and J. G.
1. Carrasco N. 1993 Iodide transport in the thyroid gland. Biochim Biophys Acta.
2. Robbins J, Merino MJ, Boice Jr JD, et al. 1991 Thyroid cancer: a lethal
neoplasm. Ann Intern Med. 115:133–147.
3. Smanik PA, Liu Q, Furminger TL, et al. 1996 Cloning of the human sodium
iodide symporter. Biochem Biophys Res Commun. 226:339–345.
4. Smanik PA, Ryu K-Y, Theil KS, Mazzaferri EL, Jhiang SM. 1997 Expression,
exon-intron organization, and chromosome mapping of the human sodium
iodide symporter. Endocrinology. 138:3555–3558.
5. Venkataraman GM, Yatin M, Ain KB. 1998 Cloning of the human sodium-
iodide symporter promoter and characterization in a differentiated human
thyroid cell line, KAT-50. Thyroid. 8:63–69.
6. Behr M, Schmitt TL, Espinoza CR, Loos U. 1998 Cloning of a functional
promoter of the human sodium/iodide-symporter gene. Biochem J.
7. Ryu K-Y, Tong Q, Jhiang SM. 1998 Promoter characterization of the human
Na?/I?symporter. J Clin Endocrinol Metab. 83:3247–3251.
8. Graff JR, Greenberg VE, Herman JG, et al. 1998 Distinct patterns of E-
cadherin CpG island methylation in papillary, follicular, Hurthle’s cell, and
poorly differentiated thyroid carcinoma. Cancer Res. 58:2063–2066.
9. Baylin SB, Herman JG, Graff JR, Vertino PM, Issa J-P. 1998 Alterations in
DNA methylation: A fundamental aspect of neoplasia. Adv Cancer Res.
10. Baylin SB. 1997 Tying it all together: epigenetics, genetics, cell cycle, and
cancer. Science. 277:1948–1949.
11. Stirzaker C, Millar DS, Paul CL, et al. 1997 Extensive DNA methylation
silencing of the O6-methylguanine-DNA methyltransferase gene. Cancer Res.
13. von Wronski MA, Brent TP. 1994 Effect of 5-azacytidine on expression of
human DNA repair enzyme O6-methylguanine-DNA methyltransferase. Car-
14. Samid D, Yeh A, Prasanna P. 1992 Induction of erythroid differentiation and
fetal hemoglobin production in human leukemic cells treated with phenyl-
acetate. Blood. 80:1576–1581.
15. Ormandy CJ, de Fazio A, Kelly PA, Sutherland RL. 1992 Transcriptional
regulation of prolactin receptor gene expression by sodium butyrate in MCF-7
human breast cancer cells. Endocrinology. 131:982–984.
16. Ain KB, Taylor KD, Tofiq S, Venkataraman G. 1997 Somatostatin receptor
subtype expression in human thyroid and thyroid carcinoma cell lines. J Clin
Endocrinol Metab. 82:1857–1862.
17. Ain KB, Taylor KD. 1994 Somatostatin analogs affect proliferation of human
thyroid carcinoma cell lines in vitro. J Clin Endocrinol Metab. 78:1097–1102.
18. Gilbert SF, Migeon BR. 1975 D-Valine as a selective agent for normal human
and rodent epithelial cells in culture. Cell. 5:11–17.
19. Kao WW-Y, Prockop DJ. 1977 Proline analogue removes fibroblasts from
cultured mixed cell populations. Nature. 266:63–64.
follicular hyperplasia cell line, KAT-50. Thyroid. 7:S99.
21. Chomczynski P, Sacchi N. 1987 Single step method of RNA isolation by acid
22. Poleev A, Fickenscher H, Mundlos S, et al. 1992 Pax8, a human paired box
gene: isolation and expression in developing thyroid, kidney and Wilms’
tumors. Development. 116:611–623.
23. Wang RY-W, Gehrke CW, Ehrlich M. 1980 Comparison of bisulfite modifi-
cation of 5-methyldeoxycytidine and deoxycytidine residues. Nucleic Acids
24. HermanJG,GraffJR,Myo ¨ha ¨nenS,NelkinBD,BaylinSB.1996Methylation-
specific PCR: a novel PCR assay for methylation status of CpG islands. Proc
Natl Acad Sci USA. 93:9821–9826.
25. Mosmann T. 1983 Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxic assays. J Immunol Methods.
26. Tong Q, Ryu KY, Jhiang SM. 1997 Promoter characterization of the rat
Na?/I?symporter gene. Biochem Biophys Res Commun. 239:34–41.
27. Van Herle AJ, Agatep ML, Padua III DN, et al. 1990 Effects of 13 cis-retinoic
2456 VENKATARAMAN ET AL.
JCE & M • 1999
Vol 84 • No 7
acid on growth and differentiation of human follicular carcinoma cells (UCLA Download full-text
RO 82 W-1) in vitro. J Clin Endocrinol Metab. 71:755–763.
28. Simon D, Ko ¨hrle J, Schmutzler C, Mainz K, Reiners C, Ro ¨her H-D. 1996
Redifferentiation therapy of differentiated thyroid carcinoma with retinoic
acid: basics and first clinical results. Exp Clin Endocrinol Diabetes. 104:13–15.
29. Simon D, Koehrle J, Reiners C, et al. 1998 Redifferentiation therapy with
retinoids: therapeutic options for advanced follicular and papillary thyroid
carcinoma. World J Surg. 22:569–574.
30. Gru ¨nwald F, Pakos E, Bender H, et al. 1998 Redifferentiation therapy with
retinoic acid in follicular thyroid cancer. J Nucl Med. 39:1555–1558.
31. Misaki T, Miyamoto S, Alam MS, Kasagi K, Konishi J. 1996 Tumoricidal
cytokines enhance radioiodine uptake in cultured thyroid cancer cells. J Nucl
32. Ain KB. 1995 Papillary thyroid carcinoma: etiology, assessment, and therapy.
Endocrinol Metab Clin North Am. 24:711–760.
33. Ain KB. 1998 Rare forms of thyroid cancer. In: Fagin JA, ed. Thyroid Cancer.
Boston: Kluwer; 319–340.
34. Caillou B, Troalen F, Baudin E, et al. 1998 Na?/I?symporter distribution in
human thyroid tissues: an immunohistochemical study. J Clin Endocrinol
35. Jhiang SM, Cho J-Y, Ryu K-Y, et al. 1998 An immunohistochemical study of
Na?/I?symporter in human thyroid tissues and salivary gland tissues. En-
36. Elisei R, Shiohara M, Koeffler HP, Fagin JA. 1998 Genetic and epigenetic
alterations of the cyclin-dependent kinase inhibitors p15INK4b and p16INK4a
in human thyroid carcinoma cell lines and primary thyroid carcinomas. Can-
37. Chun YS, Saji M, Zeiger MA. 1998 Overexpression of TTF-1 and PAX-8
restores thyroglobulin gene promoter activity in ARO and WRO cell lines.
38. Zannini M, Avantaggiato V, Biffali E, et al. 1997 TTF-2, a new forkhead
protein, shows a temporal expression in the developing thyroid which is
consistent with a role in controlling the onset of differentiation. EMBO J.
39. Chadwick BP, Obermayr F, Frischauf AM. 1997 FKHL15, a new human
member of the forkhead gene family located on chromosome 9q22. Genomics.
40. Ohmori M, Endo T, Harii N, Onaya T. 1998 A novel thyroid transcription
factor is essential for thyrotropin-induced up-regulation of Na?/I?symporter
gene expression. Mol Endocrinol. 12:727–736.
41. Ledent C, Parmentier M, Vassart G. 1990 Tissue-specific expression and
methylation of a thyroglobulin-chloramphenicol acetyltransferase fusion gene
in transgenic mice. Proc Natl Acad Sci USA. 87:6176–6180.
42. Ikuyama S, Niller HH, Shimura H, Akamizu T, Kohn LD. 1992 Character-
ization of the 5?-flanking region of the rat thyrotropin receptor gene. Mol
43. Avvedimento EV, Obici S, Sanchez M, Gallo A, Musti A, Gottesman ME.
1989 Reactivation of thyroglobulin gene expression in transformed thyroid
cells by 5-azacytidine. Cell. 58:1135–1142.
44. Avvedimento VE, Musti AM, Ueffing M, et al. 1991 Reversible inhibition of
a thyroid-specific trans-acting factor by Ras. Genes Dev. 5:22–28.
45. Pochin EE. 1971 Radioiodine therapy of thyroid cancer. Semin Nucl Med.
46. Schlesinger T, Flower MA, McCready VR. 1989 Radiation dose assessments
in radioiodine (131I) therapy. I. The necessity for in vivo quantitation and
dosimetry in the treatment of carcinoma of the thyroid. Radiother Oncol.
47. Benua RS, Leeper RD. 1986 A method and rationale for treating metastatic
thyroid carcinoma with the largest safe dose of131I. In: Medeiros-Neto G,
Gaitan E, eds. Frontiers in thyroidology. New York: Plenum Medical;
48. Shimura H, Haraguchi K, Miyazaki A, Endo T, Onaya T. 1997 Iodide uptake
and experimental131I therapy in transplanted undifferentiated thyroid cancer
cells expressing the Na?/I?symporter gene. Endocrinology. 138:4493–4496.
IODIDE UPTAKE RESTORATION IN THYROID CANCER2457