| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles: Hormones and Reproductive Health |
Departments of Endocrinology (S.J.B.A., D.F.W., G.M.B., A.B.G., J.P.M.), Clinical Biochemistry (J.P.W., C.L.D., J.M.B.), and Morbid Anatomy (J.F.G.), St. Bartholomew’s and the Royal London School of Medicine and Dentistry, London EC1A 7BE, United Kingdom
Address all correspondence and requests for reprints to: Dr. S. J. B. Aylwin, Department of Clinical Biochemistry, Molecular Endocrinology, Lab 1.4, St. Bartholomew’s and the Royal London School of Medicine and Dentistry, London EC1A 7BE, United Kingdom. E-mail: saylwin{at}mds.qmw.ac.uk
Abstract
The orphan nuclear receptors, steroidogenic factor 1 (SF-1) and DAX-1,
are involved in gonadotroph differentiation, and SF-1 has been shown to
activate the LH-ß and glycoprotein hormone 
-subunit (GSU) gene
promoters. Pituitary adenomas from 34 patients [13 somatotroph tumors,
4 prolactinomas, and 17 clinically nonfunctioning pituitary adenomas
(NFPAs)] were enzymatically dispersed and cultured in
vitro for 48 h. Tissue culture medium was collected and
assayed for LH, FSH, and GSU; messenger RNA was extracted from
adherent cells, and expression of SF-1 and DAX-1 messenger RNA was
determined by RT-PCR and verified by direct DNA sequencing. The
presence of DAX-1 protein in tumor tissue was confirmed by
immunocytochemistry.
DAX-1 was demonstrated in all NFPAs, 7 of 13 somatotroph tumors and 0
of 4 prolactinomas. SF-1 expression occurred in 8 of 16 NFPAs, 4 of 12
somatotroph tumors, and 1 of 4 prolactinomas. LH secretion in
vitro was greater in NFPAs that were SF-1 positive
(P < 0.05). Neither FSH secretion nor GSU
secretion in vitro were significantly related to the
expression of SF-1 or DAX-1. SF-1-positive somatotroph tumors
immunostained positively for LH-ß and/or FSH-ß and secreted
gonadotropins in vitro.
SF-1 expression is associated with the in vitro secretion of LH by NFPAs. A proportion of somatotroph tumors also express SF-1 and DAX-1 and secrete gonadotropin hormones in vitro.
THE MATURE ANTERIOR pituitary gland contains specific hormone-producing cell types that develop from pluripotential progenitor cells (1). Differentiated cell types emerge in a precise spatial and temporal pattern, with their development being regulated by several tissue- and cell-specific transcription factors (2). For example, a tissue-specific POU domain factor Pit-1 is required for terminal differentiation, growth, and survival of somatotrophs, lactotrophs, and thyrotrophs (3) and also regulates the expression of the cell specific genes, GH, PRL, and TSH-ß-subunit (4).
Two members of the nuclear receptor superfamily, steroidogenic factor 1
(SF-1) and DAX-1, have recently been shown to play important roles in
pituitary gonadotroph differentiation and regulation of
gonadotroph-specific hormone genes (5). SF-1 (also called
Ad4BP) expression was first demonstrated in adrenal and gonadal cells
but was subsequently identified in the pituitary, with localization to
the gonadotroph (6, 7). SF-1 is important for the
development and differentiation of pituitary gonadotrophs
(6) and has been shown to regulate both the glycoprotein
hormone -subunit gene (GSU) (7) and the LH
ß-subunit gene (LH-ß) (8). Mice with disruption of the
gene encoding SF-1 (ftz-f1) exhibit adrenal hypofunction, gonadal
agenesis, and phenotypic XY sex reversal (9). Like SF-1,
DAX-1 also plays a role in adrenal and gonadal development and was
first isolated by positional cloning from subjects with X-linked
adrenal hypoplasia congenita (10). Affected (male)
individuals with DAX-1 gene mutations present with adrenal hypofunction
at birth and gonadotropin deficiency at the time of expected puberty
(11). Conversely, duplication of the DAX-1 gene leads to
male-to-female sex reversal (12). SF-1 and DAX-1 share
spatial and developmental patterns of expression (13), and
their mutual involvement in the development of the adrenal and gonadal
axes suggested that SF-1 and DAX-1 might interact in the expression of
tissue-specific genes. SF-1 has been shown to regulate DAX-1 expression
(5, 14), and DAX-1 may regulate its own expression by
acting to inhibit SF-1-mediated expression (15). DAX-1 is
thought to oppose the transcriptional effects of SF-1 by interacting
directly with SF-1 and recruiting corepressor molecules
(16).
Human pituitary adenomas are monoclonal expansions of anterior
pituitary cells and, as such, represent a potentially useful model for
the investigation of the transcriptional control of gene expression. A
number of investigators have examined expression of SF-1 messenger RNA
(mRNA) and protein in human pituitary adenomas (17, 18)
and have concluded that SF-1 expression correlates with the expression
of gonadotropins (17) and is restricted to cells derived
from gonadotroph lineages (18). One group has examined
DAX-1 mRNA expression in pituitary adenomas and has shown that, in the
majority of tumors, DAX-1 is expressed in parallel with SF-1
(19). To date, no studies have examined expression of SF-1
and DAX-1 in relation to tumorous hormone secretion in
vitro. The GSU gene is regulated by SF-1 in gonadotrophs,
whereas the transcription factors responsible for GSU expression in
thyrotrophs are less clearly defined but may include Msx-1 and Pitx-1
(20, 21). However, GSU is secreted from tumors other
than gonadotroph and thyrotroph adenomas and is a frequent finding in
plurihormonal tumors presenting with acromegaly (22, 23, 24, 25);
the mechanisms involved in the apparently aberrant expression of GSU
in these plurihormonal tumors is not defined. In addition to GSU,
plurihormonal tumors have been shown to express gonadotropin and TSH
ß-subunits (25, 26). We were therefore interested in
examining the pattern of SF-1 and DAX-1 expression in relation to the
secretion of GSU and the gonadotropins, LH and FSH, both in adenomas
of gonadotroph and somatotroph origin.
Materials and Methods
Clinical details and patient selection
Pituitary tumors were collected from 56 patients at the time of transsphenoidal adenomectomy. Tissues were divided, at the time of surgery, for diagnostic histological studies and for tissue culture. All subjects gave informed consent, at the time of operation, for surgical specimens to be used for diagnostic and research purposes.
Acromegaly was diagnosed on the basis of persistently measurable GH and inadequate suppression of GH on administration of an oral glucose load. Those patients with PRL-secreting macroadenomas represented a subset of patients with hyperprolactinemia that had not responded to treatment with dopamine agonist therapy and who required debulking surgery before external beam radiotherapy. Patients who presented because of mass effect, without clinical features of anterior pituitary hormone excess and without laboratory evidence of tumor-derived hormone hypersecretion, were classified as having clinically nonfunctioning pituitary adenomas (NFPAs).
After surgery, confirmation of the clinical diagnosis was obtained by light microscopy and immunocytochemistry. Tumors from patients with NFPAs were examined by immunocytochemistry to exclude low-grade GH, PRL, or ACTH secretion. Tumors were excluded from the series if significant normal pituitary tissue was present on microscopy, if microscopy suggested an alternative histological diagnosis, if the tissue seemed necrotic, or if histological data were unavailable.
Morphology and immunocytochemistry
All tumors were examined by standard hematoxylin and eosin,
reticulin, and periodic acid Schiff stains, and routine immunostaining
was performed for GH, PRL, ACTH, TSH-ß, LH-ß, FSH-ß (antibodies
against the whole molecule obtained from BioGenex Laboratories, Inc. Ltd., Berkshire, UK), and GSU (rabbit polyclonal,
UCB Bioproducts Braine-L’alleud, Belgium), using a standard
streptavidin-biotin horseradish peroxidase method and previously
optimized antigen recovery (27). The amount of hormone
positivity in all tumors was initially assessed and quantified by means
of a three-point scale (-, negative or scattered cells representing
less than 10%; +, 10–50% positive; ++, >50% positive).
Immunostaining for the presence of transcription factor antigen was
also performed in tumors where sufficient tissue was available, using
an antibody directed against DAX-1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). DAX-1 antiserum was used at a dilution of
1:500, pretreated in a microwave for 10 min in a 0.01-mol/L citrate
buffer at pH6. Normal pituitary tissue from surgical specimens was used
as a positive control; and fresh biopsy specimens from skeletal muscle,
heart, liver, and kidney were used as negative controls. A rabbit
polyclonal IgG, raised against the DNA binding domain of a bacterially
expressed mouse SF-1-GST fusion protein, was used to immunostain for
SF-1 (Upstate Biotechnology, Inc., Lake Placid,
NY). The antibody was stored at -20 C and used at a 1:1000
dilution.
Pituitary tumor cell culture
Pituitary adenoma tissue was transported to the laboratory in DMEM containing 10% (vol/vol) heat-inactivated FCS, 0.06 g/L penicillin, 0.1 g/L streptomycin, and 2.5 g/L fungizone (Life Technologies, Inc., Paisley, Strathclyde, UK), hereafter referred to as culture medium. Tumor tissue was dispersed enzymatically with trypsin, as described previously (27). Dispersed cells were harvested by centrifugation, washed once, and subsequently resuspended in culture medium. Cell viability was assessed using trypan blue exclusion and was more than 90% of cells in all of the tumors studied after cell dispersion. Cell yield from each tumor varied from 1–15 x 106 cells. The cells were plated in 6-well plates at approximately 1 x 106 cells per well in 4 mL medium. Cultures were incubated at 37 C in a humidified atmosphere of 95% air-5% CO2, for 48 h, to allow cell attachment to occur, after which time the medium was collected and assayed for basal hormone secretion as described below. Adherent cells were lysed with a buffered guanidinium thiocyanate solution and stored at -70 C before mRNA extraction (see below).
mRNA analysis
Cell extract from cultured cells was thawed on ice and centrifuged to remove cell debris. mRNA was isolated using the QuickPrep micro mRNA purification kit (Pharmacia Biotech, St. Albans, Herts, UK), which uses oligo (dT) cellulose to extract poly (A) mRNA. After washing in low- and high-salt buffer [10 mmol/L Tris-HCl (pH 7.5), 1 mmol/L EDTA, and NaCl (0.1 mol/L or 0.5 mol/L respectively)], the mRNA was eluted in 10 mmol/L Tris-HCl and 1 mmol/L EDTA and precipitated with glycogen (0.25 g/L), 0.25 mol/L potassium acetate, and 95% ethanol. After centrifugation, the RNA pellet was washed with 75% ethanol and allowed to dry before being resuspended in diethylpyrocarbonate-treated water. Quantitation was performed by optical density measurement at 260nm, with approximately 0.5–1.0 µg mRNA being recovered from 106 cells.
RT-PCR
RT was performed using the First Strand complementary DNA (cDNA) synthesis kit (Pharmacia Biotech), which uses Moloney murine leukemia virus reverse transcriptase and a NotI–d(T)18 bifunctional primer. RT was also performed on each RNA sample in the absence of Moloney murine leukemia virus reverse transcriptase. PCR carried out on this resultant RT reaction generated no PCR product, confirming the absence of genomic DNA contamination.
PCR was carried out using the equivalent of 0.1 µg mRNA in a total reaction vol of 50 µL, and a so-called hot-start technique was employed. To reduce the variability between samples, PCR components were premixed before addition to individual PCR tubes. The 50 µL reaction mixture contained 0.2 mmol/L deoxynucleotide triphosphate, 5 µL 10-fold concentrated PCR buffer, 5 µL cDNA, and 100 pmol of either Pit-1, SF-1, or DAX-1 primers. Primers used were as follows: Pit-1: 5'-AGTGCTGCCGAGTGTCTACCA-3' (forward), 5'-TTTCTTTTCCTTTCATTTGCT-3' (reverse), generating a fragment of 560 bp (28); SF-1: 5'-GCATCTTGGGCTGCCTGCAG-3' (forward), 5'-CCTTGCCGTGCTGGACCTGG-3' (reverse), generating a fragment of 230 bp (17); DAX-1: 5'-AAGGAGTACGCCTACCTCAA-3' (forward), 5'-TCCATGCTGACTGTGCCGAT-3' (reverse), generating a fragment of 251 bp (13). All primers were set on different exons of each gene so that they spanned an intron. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was used to verify the integrity of RNA from each specimen, using primers from CLONTECH Laboratories, Inc. (Cambridge, UK) in each PCR and generated a 986-bp fragment. The PCR mixture was heated to 94 C for 3 min and cooled before addition of 2.5 U Taq polymerase (Roche Molecular Biochemicals, Lewes, East Sussex, UK), followed by 30 cycles of denaturation at 94 C for 30 sec, annealing at appropriate temperatures (Pit-1, 60 C; SF-1, 60 C; DAX-1, 55 C) for 30 sec, and extension at 72 C for 45 sec. Controls with water replacing template were included in all experiments. PCR products were separated by electrophoresis using 1.6% agarose gels stained with 0.5 µg/mL ethidium bromide and examined under ultraviolet light for the presence of a band of the expected size. HincII-digested DNA was run in parallel as a molecular weight marker. Transcription factor expression was examined in one tumor before dispersal, with the remainder of the tissue being dispersed as described above. The pattern of transcription factor expression was identical in both mRNA samples (data not shown).
PCR products from SF-1- and DAX-1-positive tumor samples were extracted after agarose gel electrophoresis by adsorption to a silica-gel membrane, washing in high-salt buffer, and elution in 10 mmol/L Tris-HCl (QIAGEN, West Sussex, UK). Purified PCR products were further analyzed by automated sequencing (ABI Prism, Perkin-Elmer Corp., Buckinghamshire, UK).
Hormone assays
LH, FSH, and TSH were measured using two-site chemiluminescent
enzyme immunometric assays on the Immulite autoanalyzer (Euro/DPC
Limited, Gwyenedd, UK). The inter- and intraassay coefficients of
variations for all these assays are less than 10% and 6%,
respectively. Concentrations of GSU were measured by a direct
double-antibody RIA using antibodies purchased from UCB Bioproducts
(Brussels, Belgium), chloramine-T-iodinated antigen (National Institute
for Biological Standards and Control reagent 76/508, Potters Bar,
Hertfordshire, UK) and calibrated against the first International
Reference preparation 75/569 (National Institute for Biological
Standards and Controls). Inter- and intraassay coefficients of
variation were less than 11% and less than 6%, respectively.
Cross-reactivities (in ng/ng) with purified LH, FSH, and TSH
were 3.6%, 1.9%, and 1.3%, respectively. The detection limits of the
above assays, defined as the concentration 2 SD above the
response at zero dose, were as follows: LH, 0.4 IU/L; FSH, 0.6 IU/L;
TSH, 0.008 mU/L; and GSU, 0.1 µg/L. All samples from each
individual tumor were analyzed in the same assay, hormone data being
expressed as the amount of hormone secreted per 24 h per
106 cells. After this normalization, the reported
detection limits were as follows: LH, 2.0 mIU; FSH, 3.0 mIU; TSH, 0.1
µU; and GSU, 0.5 ng.
Statistical analysis
Comparisons between groups of tumors (see Figs. 2
and 3) were
made using the Kruskal-Wallis test. Where the overall P-value was
significant (P < 0.05), individual comparisons were
made using Dunn’s multiple-comparisons test. All statistical analysis
was performed using GraphPad Prism software
(GraphPad Software, Inc., San Diego, CA). A value of
P < 0.05 was taken as significant.
|
|
Tumor tissue was initially collected from 56 patients, and adequate mRNA was obtained from 45, as judged by positive expression of GAPDH (18 somatotroph tumors, 7 prolactinomas, and 20 NFPAs). Four tumors were subsequently excluded because normal pituitary tissue was present on light microscopy or the tissue appeared necrotic; and in two cases, no tissue was available for histological examination. Two somatotroph tumors secreted PRL, but not GH, in vitro and were removed from subsequent analysis. Three NFPAs stained positively for GH and/or showed evidence of Pit-1 expression; these tumors were thought to represent silent somatotroph adenomas and were also excluded. Results are therefore presented on 34 tumors [13 somatotroph tumors (S1–S13), 4 prolactinomas (P1–P4), and 17 nonfunctioning adenomas (N1–N17)].
Immunocytochemistry and in vitro hormone secretion
Table 1 shows both the
immunocytochemical data and the in vitro secretion of
gonadotropin hormones by 17 clinically nonfunctioning tumors. Within
this group, 6 tumors stained positively for LH-ß and/or FSH-ß
(tumors N1, N2, N3, N9, N10, and N12). Insufficient material was
present from tumor N5 for immunocytochemical analysis of gonadotropin
subunits. Tumors N4 and N11 were additionally diagnosed preoperatively
as gonadotropin-secreting adenomas, on the basis of an elevated serum
FSH (data not shown), although these tumors were negative, by
immunocytochemistry, for gonadotropin ß-subunits. These tumors
(N1–N5 and N9–N12) secreted LH and/or FSH in vitro in
variable proportions. Tumors N6–N8 and N13–N17 were negative (by
immunocytochemistry), showed no more than trace amounts of hormone
secretion in vitro, and were considered to represent null
cell adenomas.
|
RT-PCR
Figure 1 shows the PCR products
generated using Pit-1, DAX-1, and SF-1 primers from a representative
selection of tumors: S1, S2, S7, S8, S9, P1, N1, N2, N12, and N13.
Positive expression of Pit-1, SF-1, and DAX-1 was demonstrated in
cadaveric human pituitary, although the expression of DAX-1 and Pit-1
was weaker than observed in adenoma tissue. DAX-1 and SF-1 PCR products
were subsequently gel-extracted and sequenced, in each case confirming
the GenBank cDNA sequences (S74720 and U76388, respectively). Pit-1
expression was present in 12 of 13 somatotroph tumors tested, and was
absent in all of the nonfunctioning tumors reported, further evidence
against the presence of contaminating nontumorous elements. Tables 1
and 2 show the results of PCR for DAX-1
and SF-1. DAX-1 was present in 16 of 16 NFPAs, whereas SF-1 was present
in 8 of 16 tumors. DAX-1 expression was found in 7 of 13 somatotroph
tumors, whereas SF-1 mRNA was present in 4 of 12 somatotroph tumors.
DAX-1 was present in none of the 4 prolactinomas studied, and SF-1 was
present in only 1 of 4. Interestingly, DAX-1 and SF-1 expression tended
to occur together in somatotroph adenomas (
-square = 4.3,
P < 0.05).
|
|
Figure 2 shows the in
vitro secretion of LH, FSH, and GSU by NFPAs and somatotroph
adenomas, in relation to SF-1 expression, also demonstrated in Tables 1 and 2. LH secretion was significantly greater in SF-1-positive NFPAs
(median, 7.4 mIU/106 cells/24 h), compared with
SF-1-negative tumors (median, <2.0 mIU/106
cells/24 h; P < 0.05). In contrast, FSH secretion was
seen from NFPAs that were both SF-1-positive (median, 8.6
mIU/106 cells/24 h) and SF-1-negative (median,
3.1 mIU/106 cells/24 h). GSU was also secreted
equally in both SF-1-positive NFPAs (median, 1.4
ng/106 cells/24 h) and SF-1-negative tumors
(median, 1.4 ng/106 cells/24 h). LH secretion was
also noted in three somatotroph adenomas, all of which were
SF-1-positive, whereas GSU secretion by somatotroph adenomas
occurred in both SF-1-positive and -negative tumors.
Gonadotropin secretion, in relation to DAX-1 expression, is shown in
Fig. 3. All the NFPAs were positive for
DAX-1. In vitro LH secretion (three tumors) and FSH
secretion (two tumors) were observed only in somatotroph adenomas that
were positive for DAX-1. GSU secretion did not differ significantly
between DAX-1-positive (median, 9.5 ng/106
cells/24 h) and -negative (median, 5.1 ng/106
cells/24 h) groups.
Immunocytochemistry for DAX-1
To confirm the RT-PCR data, immunocytochemistry for DAX-1 was
performed on 12 tumors, where sufficient tissue was available. The
DAX-1 antibody gave clear nuclear staining on control sections of
normal pituitary and in pituitary adenomas (see Fig. 4). Strongly positive staining was
observed in 5 of 6 NFPAs tested (N8, N10, N11, N12, and N15 positive;
N16 negative), all of which were shown to express DAX-1 mRNA.
DAX-1-positive staining was also observed in 4 somatotroph adenomas
(S3, S5, S6, and S7) that were DAX-1 mRNA-positive, and in one adenoma
where DAX-1 mRNA was negative (S11). This discrepancy may indicate
tumor heterogeneity, and it is worth noting that the sample for
immunocytochemistry was positive for FSH-ß, whereas in
vitro FSH secretion was not detected in the sample used for tissue
culture and mRNA extraction. Somatotroph adenoma S12 was negative for
both DAX-1 mRNA and protein. Taken together, the combined
immunocytochemical and RT-PCR data suggest that the expression of DAX-1
is preferentially expressed in, but not restricted to, adenomas of
gonadotroph origin.
|
Discussion
In this study, we have examined a large series of pituitary
adenomas for the presence of the gonadotroph-specific orphan nuclear
receptors, SF-1 and DAX-1, and related this to tumor type and to the
in vitro expression of pituitary hormones. In
vitro secretion of LH was restricted to tumors that were positive
for SF-1, whereas secretion of neither FSH nor GSU was related to
the presence of SF-1. DAX-1 expression was demonstrated in all NFPAs
but was also present in 7 of 17 GH- and/or PRL-secreting tumors and was
not related to hormone secretion. By determining in vitro
tumor hormone secretion, we have been able to gain a better insight
into the protein synthetic capacity of adenomatous cells without the
influence of hypothalamic factors, gonadal steroids, and the
paracrine effects of surrounding pituitary tissue. Furthermore, we
measured in vitro secretion and subsequently extracted mRNA
from the same populations of dispersed tumor cells to avoid our results
being confounded by heterogeneity within adenoma tissue.
Three previous studies have examined the expression of SF-1 in human
pituitary adenomas. Ikuyama et al. (18)
demonstrated SF-1 mRNA by Northern blotting in 5 of 13 NFPAs, and the
authors noted an association between the expression of SF-1 and both
FSH-ß and GSU, although they did not report on LH-ß expression
in that study. A subsequent paper by the same group demonstrated SF-1
positivity, by RT-PCR, in 12 of 18 NFPAs in association with LH-ß and
FSH-ß expression (19). Asa et al.
(17) also used RT-PCR and identified SF-1 in 8 of 8
gonadotroph/null cell/oncocytomas. Our findings in NFPAs are in line
with these investigators, but we have extended these observations to
show that in vitro secretion of LH occurred predominantly in
the NFPAs that were SF-1-positive. Whereas SF-1 is well established as
a marker of gonadotroph differentiation (5), our data
provide additional support to previous studies in cell lines and
transgenic mice that have demonstrated that SF-1 is necessary for the
expression of the LH-ß gene (8, 29). Interestingly,
although the FSH-ß promoter is less well characterized, there is
evidence that FSH-ß gene expression may not be regulated by SF-1
(30), and indeed we failed to demonstrate an association
between SF-1 expression and in vitro FSH secretion.
All of the gonadotroph and null cell adenomas in our series expressed DAX-1 mRNA. One previous study has addressed the expression of DAX-1 in pituitary adenomas, the authors demonstrating DAX-1 expression in 11 of 18 NFPAs, although they included a number of Pit-1-positive tumors that might be better considered as silent somatotroph adenomas (19). DAX-1 was expressed in 7 of 8 tumors that they considered to be of gonadotroph lineage, where it was associated with the expression of gonadotroph-specific genes and with SF-1.
Human subjects with DAX-1 mutations and SF-1 (ftz-f1) disrupted mice exhibit abnormalities of gonadotropin synthesis and release (6, 31). However, although SF-1 is a transcriptionally active protein in gonadotroph, adrenal, and gonadal tissues (32), DAX-1 has an inhibitory effect on SF-1-mediated transcription (15, 16). At present, it seems likely that both SF-1 and DAX-1 are necessary for the initial differentiation of gonadotroph cells in man; whereas, in the mature cell, DAX-1 opposes the transcriptional effects of SF-1 (33). Our data demonstrate that DAX-1 expression is not confined to pituitary adenomas of gonadotroph origin, and we were unable to demonstrate a clear relationship between DAX-1 expression and gonadotropin secretion.
In our series, SF-1 and DAX-1 mRNA were detectable in 33% and 54% of somatotroph tumors, respectively, frequently in association with in vitro gonadotropin secretion, gonadotropin immunoreactivity, and/or serum elevation of gonadotropins. The presence of DAX-1 protein in both NFPAs and somatotroph tumors was confirmed by immunocytochemistry. Previous studies have also demonstrated SF-1 in tumors of nongonadotroph origin. Asa et al. (17) documented SF-1 mRNA in a minority of corticotroph, lactotroph, and somatotroph adenomas. In one of the corticotroph adenomas, the expression of SF-1 was diffuse and associated with LH-ß immunostaining. More frequently, adenomas exhibited foci of SF-1 positivity in association with gonadotropin hormones, and the authors were inclined to view these as nests of normal cells within the adenoma. This explanation is hard to reconcile with earlier work that has repeatedly demonstrated monoclonality in the majority of pituitary tumors (34). Ikuyama et al. (18) were able to demonstrate SF-1 mRNA in sufficient quantities to be demonstrated by Northern blot analysis in one of seven somatotroph tumors; this tumor also expressed FSH-ß. The one previous study looking at DAX-1 expression did not identify DAX-1 mRNA in somatotroph adenomas, but the authors highlighted three of the clinically nonfunctioning tumors studied that expressed both Pit-1 and DAX-1 mRNA (19). Electron microscopy of these tumors was not reported, but it is possible that they were silent subtype III (plurihormonal) adenomas (35).
Our findings and those of previous investigators strongly suggest that a minority of tumors from patients presenting with acromegaly express hormones and transcription factors that are specific to gonadotroph cells in the normal pituitary. It is likely that the expression of SF-1 and DAX-1 and the in vitro secretion of gonadotropin hormones indicate a process of tumor dedifferentiation into pluripotent cells.
GSU secretion in vitro was not significantly greater in
SF-1-positive NFPAs or somatotroph tumors, compared with SF-1-negative
tumors. Ikuyama et al. (18) and Asa et
al. (17) also demonstrated GSU expression in
somatotroph adenomas in the absence of SF-1. GSU is expressed in the
fetal pituitary in advance of SF-1 expression and gonadotroph
determination and is present, albeit in reduced quantities, in the
FTZ-F1-disrupted (SF-1 knockout) mouse (6). The GSU
gene is expressed in thyrotroph and placental cells, as well as in
gonadotroph cells, and GSU-secreting murine thyrotroph and human
choriocarcinoma cell lines do not express SF-1; GSU promoter
activity is determined by cis-elements other than the SF-1
binding site in these models (36). It is evident that
there are a number of different mechanisms that can lead to
tissue-specific expression of the GSU gene.
GSU expression may arise in SF-1-negative tumors because of loss of
expression of a repressor protein or through altered regulation of the
GSU promoter; indeed, a recent report has suggested that the GSU
promoter contains a cis-element that can act as a repressor
in the GH3 somatotroph cell line (37). However, we have
previously demonstrated that GSU secretion is strongly correlated
with that of TSH in somatotroph adenomas (25), and it may
be that thyrotroph-specific factors account for the presence of GSU
in SF-1-negative GSU-secreting cells in somatotroph tumors. It is of
interest that GSU secretion in the somatotroph adenomas that were
SF-1 negative frequently occurred in association with in
vitro TSH secretion, as shown in Table 2.
In summary, our studies have demonstrated that in pituitary adenomas of gonadotroph origin, DAX-1 was present in all cases, but the expression of SF-1 was found in only 50%, where it was positively associated with LH secretion. In somatotroph adenomas, SF-1 and DAX-1 expression were demonstrated in a smaller proportion; but in these tumors, there was an association with the in vitro secretion of LH. Previous studies have suggested that SF-1 and DAX-1 expression is restricted to cells of gonadotroph lineage (17, 18, 19). The expression of these factors in pituitary adenomas from patients presenting with acromegaly supports the hypothesis that some adenomas may dedifferentiate into, or arise from, a common gonadotroph/somatotroph precursor.
Acknowledgments
We thank our surgical colleagues at Addenbrooke’s Hospital (Cambridge, UK) and the Royal Hospitals NHS Trust (London, UK) for providing surgical specimens. We also acknowledge the work of Carole Nickols for her help with immunocytochemistry.
Footnotes
1 This work was supported in part by the award of a Wellcome Trust
Vacation Scholarship (to J.P.W.). 
Received September 13, 2000.
Revised November 30, 2000.
Accepted December 7, 2000.
References
-subunit gene in pituitary gonadotropes. Mol Endocrinol. 8:878–885.-subunit promoter. Mol Endocrinol. 11:1782–1794.-subunit gene promoter
in thyrotropes. Mol Cell Endocrinol. 146:77–86.[CrossRef][Medline]
-subunit production and plurihormonality in
human pituitary corticotroph tumours— an in vitro and
immunohistochemical study. Eur J Endocrinol. 133:25–32.This article has been cited by other articles:
 |
|
 |
|
  E. De Menis, F. Roncaroli, V. Calvari, V. Chiarini, P. Pauletto, G. Camerino, and N. Cremonini Corticotroph adenoma of the pituitary in a patient with X-linked adrenal hypoplasia congenita due to a novel mutation of the DAX-1 gene Eur. J. Endocrinol., August 1, 2005; 153(2): 211 - 215. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Kumar, A. E. Gururaj, R. K. Vadlamudi, and S. K. Rayala The Clinical Relevance of Steroid Hormone Receptor Corepressors Clin. Cancer Res., April 15, 2005; 11(8): 2822 - 2831. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P L Hanson, S J B Aylwin, J P Monson, and J M Burrin FSH secretion predominates in vivo and in vitro in patients with non-functioning pituitary adenomas Eur. J. Endocrinol., March 1, 2005; 152(3): 363 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Tajima, A. Dantes, Z. Yao, K. Sorokina, F. Kotsuji, R. Seger, and A. Amsterdam Down-Regulation of Steroidogenic Response to Gonadotropins in Human and Rat Preovulatory Granulosa Cells Involves Mitogen-Activated Protein Kinase Activation and Modulation of DAX-1 and Steroidogenic Factor-1 J. Clin. Endocrinol. Metab., May 1, 2003; 88(5): 2288 - 2299. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Marx and L. K. Nieman Aggressive Pituitary Tumors in MEN1: Do They Refute the Two-Hit Model of Tumorigenesis? J. Clin. Endocrinol. Metab., February 1, 2002; 87(2): 453 - 456. [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 |
DNA size marker; R, human cadaveric
pituitary mRNA; W, water; GAPDH: human GAPDH positive control.