| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
1 and
2 Mutations in Nonfunctioning Pituitary Tumors1
Department of Medicine, Queen Elizabeth Hospital, Edgbaston, Birmingham, B15 2TH, United Kingdom
Address all correspondence and requests for reprints to: J.A. Franklyn, Department of Medicine, University of Birmingham, Queen Elizabeth Hospital, Edgbaston, Birmingham, B15 2TH, United Kingdom. E-mail: franklja{at}bham.ac.uk
| Abstract |
|---|
|
|
|---|
and
ß isoforms, an observation that may account for abnormalities
of T3-mediated negative regulation of the glycoprotein
hormone common
-subunit. Reduced TR protein was associated with a
parallel reduction in TRß messenger RNA (mRNA), although TR
1 and
2 mRNA levels were similar in nonfunctioning tumors and normal
pituitaries. Because TR
shows aberrant posttranscriptional
processing, and TRß is under ligand-dependent autoregulation, we
hypothesized that aberrant TR expression in nonfunctioning tumors may
reflect mutation in receptor coding and regulatory sequences, and
therefore screened TR
mRNA and TRß T3 response
elements and ligand binding domains for sequence anomalies. Screening
TR
mRNA in 23 tumors and subsequently sequencing candidate fragments
identified one silent change from published sequences and three novel
missense mutations, two in the common TR
region (ser45ile and
lys370asn) and one that was
2 specific (ser377leu). TRß response
elements failed to show any differences from published sequences in 14
nonfunctioning tumors. Sequencing of TRß ligand binding domains were
also identical to wild type in 23 nonfunctioning tumors. The functional
significance of the novel TR
mutations is unknown; definition
of mutant TR action may provide insight into the role of TRs
in the growth control of pituitary cells. | Introduction |
|---|
|
|
|---|
-subunit.
Immunocytochemical staining demonstrates the presence of intact
gonadotropins and/or subunits of these glycoprotein hormones in up to
79% of nonfunctioning tumors (1, 2, 3, 4, 5, 6, 7), whereas
-subunit expression is
reported in between 40 and 70% (7, 8). In approximately 25% of cases,
elevated circulating levels of
-subunit protein have been
demonstrated (5, 7, 8, 9), and similarly, an increase in tumor messenger
RNA (mRNA) expression has been described (1). Such increased
-subunit expression is observed despite normal or even elevated
circulating concentrations of thyroid hormone (T3) and
gonadal steroids (10), which normally act via a negative feedback
mechanism to inhibit
-subunit production, suggesting abnormal
thyroid receptor (TR) and/or steroid receptor function in these tumors.
We have reported recently that nonfunctioning tumors demonstrate
reduced expression of TR
and ß isoforms when compared with normal
pituitaries (11). Despite a reduction in TR
1 and
2 protein as
determined by tumor immunocytochemistry, we found that TR
mRNAs were
present at normal levels, suggesting abnormality in posttranscriptional
processing of these mRNAs. In contrast, we reported a reduction in both
TRß protein and mRNAs in nonfunctioning tumors, findings consistent
with a change in TRß gene transcription. In the present studies we
examined the hypothesis that abnormalities of TR
and ß expression
in nonfunctioning pituitary tumors reflect sequence abnormalities of
the TR genes. To investigate a possible mechanism leading to aberrant
TR
mRNA processing, we examined the frequency of mutations of TR
coding sequences in a series of nonfunctioning tumors. In view of our
findings for TRß gene expression, and because TRß gene
transcription may be autoregulated in a ligand- dependent manner (12),
we also examined both promoter and ligand binding domains of the TRß
gene for sequence abnormalities. | Materials and Methods |
|---|
|
|
|---|
Clinically nonfunctioning pituitary macroadenomas were excised trans-sphenoidally from 23 patients (11 male, 12 female; age at the time of presentation 57 ± 2.2 yr (mean ± SE). Twelve normal human pituitary glands were obtained from postmortem proceedings (carried out within 12 to 24 h after death). All samples were stored at -80 C. Whole blood was also obtained from three subjects treated surgically for pituitary macroadenomas and found to have receptor mutations (see below).
RNA and DNA extraction
Total RNA was isolated from pituitary tissue using a single-step acid guanidinium phenol-chloroform extraction technique (13). Briefly, pituitary tissue was homogenized in the presence of RNAzol B (Biotech Laboratories, Houston, TX). Total RNA was extracted after centrifugation with 0.1 vol chloroform. The aqueous phase was mixed with an equal volume of isopropanol, and RNA was allowed to precipitate at -20 C overnight, before centrifugation and washing. DNA was extracted from pituitary tissue and blood leukocytes, using the DNace Clini Pure kit (Bioline, London, UK). Pituitary tissue was minced into approximately 10- to 20-mg pieces before following the manufacturers instructions and storing the resulting DNA at 4 C.
Reverse transcriptase PCR
RT was performed using avian myeloblastosis virus (AMV) reverse
transcriptase (Promega Corp., Madison, WI) in a total
reaction volume of 50 µL, with 2 µg pituitary total RNA, 60 pmol
oligo(dT)15, 10 µL 5x AMV reverse transcriptase buffer
(Promega Corp.), 5 µL deoxynucleotide triphosphate mix
(200 µM each) (Boehringer Mannheim, Germany), 50 units of
ribonuclease inhibitor (RNasin, Promega Corp.) and 15
units of AMV reverse trancriptase (Promega Corp.). Hot
start PCR was carried out in 50 µL volumes using 1 µL of the
resulting RT product, 30 pmol of each primer, 1 µL of deoxynucleotide
triphosphate mix (200 µM each) (Boehringer Mannheim,
Germany), 5 µL 10x PCR reaction buffer and 2 U Taq DNA
polymerase (Boehringer Mannheim). Amplification of the GC-rich region
encompassing the TRß1 promoter direct repeat (DR4) was carried out as
above, but PCR reactions were supplemented with 5% dimethylsulfoxide
and 5% glycerol. Amplification of the TRß1 ligand binding domain was
achieved by nested PCR, with outer primers 4570 and 4571 and inner
primers 607 and 608 (see Table 1
for primer
sequences). Oligonucleotide design was carried out using DNAStar
software (Madison, WI).
|
TR
1 and
2 mRNA from 23 nonfunctioning tumors and 12 normal
pituitaries was screened for mutation using the RNase Mismatch
Technique of Ambion, Inc. (Austin, Texas) (14). Specific
primers were designed to enable the amplification of discrete TR
1
and
2 products of approximately 1 kilobase. Essentially, screening
was accomplished following the methodology of the Ambion, Inc. RNase mismatch kit with minor modifications of the
heterozygote short-cut protocol. Briefly, TR
1 and
2 complementary
DNA (cDNA) was amplified with primers tagged with T7 promoter
sequences, allowing subsequent PCR products to be transcribed in
vitro. Reannealing of resulting single-stranded RNA from mutant
and wild- type alleles resulted in inappropriate hybridization at any
points of anomaly, which could then be recognized by RNase enzyme
cutting at sites of mismatch. This allowed PCR fragments differing from
wild-type sequences to be identified by agarose gel electrophoresis.
TRß1 promoter and ligand binding domains were not subject to mismatch
screening, but were instead sequenced directly.
Sequencing
TR
1 and
2 PCR fragments identified by RNase Mismatch
screening as potentially different from wild type were analyzed on an
ABI 377 automated sequencer, together with directly amplified
TRß1 products. PCR products were purified for sequencing after
excision from TAE agarose gels and elution using QIAQuick gel
purification kits (Quiagen, Hilden, Germany). Fragment
concentration was gauged in relation to a DNA mass ladder
(Pharmacia, Uppsala, Sweden). Fifty to 100 ng of the
resulting pure fragment was subsequently used per sequencing reaction,
with 4 µL AmpliTaq (Perkin Elmer, Foster City, CA) and
3.2 pmol forward or reverse primer. Overlapping PCR fragments were
sequenced in both directions using internal primers to yield products
of 3400 bp. Electropherograms were analyzed using DNASTAR
software. PCR fragments from tumors containing potential mutations were
further analyzed using restriction enzymes whose recognition sequences
were either abolished or created by the nucleotide change in
question.
| Results |
|---|
|
|
|---|
sequences
Initial RNase Mismatch screens (Fig. 1
) of overlapping fragments of TR
1 and
2 cDNA identified three fragments with potential TR
1 mRNA
mutations and seven for
2. The three TR
1 candidate fragments were
all identified as candidates for
2 mutations, suggesting common
region TR
gene mutations. Subsequent sequencing of TR
1 and
2
coding region cDNA in the seven tumors identified by the screen process
indicated that five of the potential mutations represented false
positives from the screening technique (all of which were
2
specific), and one represented a silent common region difference from
the published sequence (A to G at nucleotide position 168). Sequencing
of the three remaining fragments revealed three heterozygous mutations
(Fig. 1
), two lying in the common TR
gene region, and one lying in
the
2-specific region. The first mutation was a G to T missense
exchange at nucleotide 190, resulting in a ser to ile substitution in
amino acid 45, which lies in the shared A/B domain of the common TR
region, and which would therefore result in transcription of aberrant
TR
1 and
2 mRNAs (Fig. 2
). A second
heterozygous missense mutation was confirmed at nucleotide position
1168, a G to C change resulting in a lys370asn amino acid substitution
in the common
region. Third, a C to T substitution was noted at
nucleotide position 1187, resulting in a ser to leu change in the
2-specific amino acid 377.
|
|
1,
2, and ß1 expression
in the above patients revealed no detectable TR protein. Indeed, as we
previously reported (11), only 3 of the 23 nonfunctioning tumors
assessed revealed detectable TR protein of either isoform.
Three previously published TR
mutations described in hepatocellular
carcinoma cell lines (15) were also investigated and found to be absent
in all 23 nonfunctioning tumors studied. No coding changes were
apparent either in the cDNA of any of the 12 normal pituitaries
screened in this investigation.
TR ß1 sequences
Two thyroid response elements (TREs) [a palindromic (PAL) TRE at -890 to -866 and a direct repeat (DR4) at -186 to -170] in the upstream regulatory region of TRß1, implicated in autoregulation of the ß1 isoform, were also investigation for the presence of mutations in 14 nonfunctioning tumors and 2 normal pituitaries. No changes from published sequences were noted in either TRE in the tumors studied. To assess whether ligand binding domain mutations might affect the ligand-dependent autoregulation of TRß1 transcription, and hence explain the reduction in TRß mRNAs present in nonfunctioning tumors, the region encompassing nucleotides 1150 to 1698 of the TRß1 gene was also sequenced in 23 tumors and 2 normal pituitary glands. All tumors demonstrated wild-type ligand binding domains.
| Discussion |
|---|
|
|
|---|
Nonfunctioning tumors are a subset of pituitary tumors that may
demonstrate aberrant negative regulation of the common
-subunit. We
have previously shown that such tumors have reduced TR
and ß
isoform expression (11). The low level of TR
protein in
nonfunctioning tumors occurs in association with unaltered levels of
TR
mRNA, whereas levels of TRß mRNA are reduced compared with
those found in normal pituitaries. We therefore examined the hypothesis
that this aberrant expression of TR
and ß in nonfunctioning tumors
reflects a mutation in receptor coding and regulatory sequences.
In the present study, we identified three changes from wild-type coding
sequences in TR
mRNA in a series of 23 nonfunctioning tumors
investigated. The prevalence of these abnormalities indicates that
TR
gene mutations do not provide a general explanation for the
demonstrated reduction in TR
gene expression in human nonfunctioning
tumors, but the described mutations may nonetheless provide insight
into abnormal thyroid hormone-dependent regulation in specific
tumors.
The functional significance of the described ser45ile mutation that
occurs in the relatively poorly characterized common A/B domain is
unclear. The A/B domain is believed to have transactivating properties,
and recently, a 10-amino acid region of TR
1 was characterized that
interacts with transcription factor IIB in chicken (23) and rat (24)
models. A similar region exists in the A/B domain of human TR
1, with
the described ser45ile mutation lying 12 amino acids downstream of this
putative region. Any functional significance of this mutation is likely
to stem from its ability to influence transactivation at this site.
Mutations in the common region of TR
will affect both TR
1 and
2 variants, and may therefore have a significant impact on normal
receptor function in the pituitary, especially if such TRs act in a
dominant negative manner to interfere with normal receptor function
through altered interactions with corepressors at the TRE, in a similar
manner to heterozygous TRß1 mutations in syndromes of resistance to
thyroid hormone (25). This is particularly relevant to the C-terminal
lys370asn mutation, which substitutes an amino acid in the TR
ligand
binding domain. For T3-mediated inhibition of
transcription, unliganded TRs bind coactivators at the negative
TRE, which dissociate in the presence of thyroid hormone,
allowing corepressor association. An altered ability to bind thyroid
hormone is likely therefore to affect the efficiency with which
coactivators can dissociate from TR
1 in the presence of
T3, in turn disrupting transcriptional inhibition of
negatively regulated genes such as the glycoprotein hormone
-subunit. Indeed, three previously reported TR
1 mutations present
in hepatocarcinoma cell lines (glu350lys, val390ala and
pro398ser), which are in proximity to the lys370asn mutation reported
here, abolish T3 binding affinity (15). The C-terminal
TR
2-specific mutation ser377leu is unlikely to affect any
T3-dependent functions of
2 receptors, given that
this isoform is unable to bind ligand, but may alter its ability to
interact with other nuclear receptors and transcription factors, and as
such, interfere with wild-type receptor function. Because the three
coding changes reported here were not found in leukocyte DNA from the
respective patients, it is unlikely that they are merely polymorphic
changes in sequence. In terms of the clinical associations, the
C-terminal mutations lys370asn and ser377leu were associated with
relatively early onset (age 36 and 38 yr, respectively) and tumor
recurrence. The N-terminal substitution ser45ile was not, however,
associated with either early onset or further recurrence, leading us to
speculate that the C terminus of TR
isoforms may have functional
roles in the control of tumor cell growth.
Direct sequencing of the two TREs upstream of the ß1 transcriptional start site failed to reveal any changes from published sequences (12, 26), indicating that the abnormally low levels of TRß mRNA observed in nonfunctioning tumors does not reflect sequence abnormalities in these regions of the TRß promoter. The defective negative feedback regulation apparent in nonfunctioning tumors is unlikely to be associated with an altered ability of TRß receptors to bind ligand, because none of the TRß ligand binding domains sequenced demonstrated any deviation from wild-type sequences. Therefore it would appear that TRß is down-regulated at the level of transcription in nonfunctioning tumors by trans-acting factor or factors unknown, and not by T3-dependent autoregulatory changes in TRß expression.
In summary, the novel TR mutations described here may be of functional significance in terms of TR action, and further definition of their functional properties may provide insight into the role of TRs in growth control in pituitary cells.
| Acknowledgments |
|---|
| Footnotes |
|---|
Received February 18, 1998.
Revised September 11, 1998.
Accepted October 21, 1998.
| References |
|---|
|
|
|---|
1 nuclear receptor with
impairment in homodimerization. Proceedings of the 79th Annual Meeting
of the Endocrine Society, Minneapolis, MN. P250, 297.
This article has been cited by other articles:
![]() |
S. Y Chan, M. H Andrews, R. Lingas, C. J McCabe, J. A Franklyn, M. D Kilby, and S. G Matthews Maternal nutrient deprivation induces sex-specific changes in thyroid hormone receptor and deiodinase expression in the fetal guinea pig brain J. Physiol., July 15, 2005; 566(2): 467 - 480. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Kamiya, M. Puzianowska-Kuznicka, P. McPhie, J. Nauman, S.-y. Cheng, and A. Nauman Expression of mutant thyroid hormone nuclear receptors is associated with human renal clear cell carcinoma Carcinogenesis, January 1, 2002; 23(1): 25 - 33. [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 |