| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Pathology and Laboratory Medicine (S.L.A.) and Department of Medicine (S.E.), Mount Sinai Hospital, and Department of Laboratory Medicine and Pathobiology (S.L.A.) and Department of Medicine (S.E.), The University of Toronto, Toronto, Ontario, Canada M5G 1X5
 |
Abstract |
|---|
 Top Abstract I. Introduction: Pituitary...II. Pituitary...III. Pathogenetic Mechanisms in...IV. Concluding CommentsReferences |
|---|
|
I. Introduction: Pituitary Adenomas |
|---|
|
TopAbstractI. Introduction: Pituitary...II. Pituitary...III. Pathogenetic Mechanisms in...IV. Concluding CommentsReferences |
|---|
For many years there has been controversy regarding the basis of pituitary tumorigenesis. Two prevailing theories have pitted hormonal stimulation against an intrinsic pituitary defect. Several animal models have provided support for the role of hormonal stimulation in the development of these neoplasms, and there is evidence for adenohypophysial production of hypothalamic hypophysiotropic hormones that may be responsible for excess stimulation. Other growth factors have also been shown to cause pituitary tumors. In contrast, the clonal nature of pituitary adenomas and the lack of associated hyperplasia in most patients with pituitary adenomas argue for a molecular defect as the etiology of these lesions. An integrated approach reconciles the two proposed theories of tumorigenesis by applying the multistep theory of carcinogenesis. It is likely that the majority of pituitary adenomas develop from transformed cells that are, nevertheless, dependent on hormonal and/or growth factor stimulation for tumor progression, which will be discussed below.
A. Definition
Pituitary adenomas are nonmetastasizing neoplasms composed of
adenohypophysial cells. Although they usually arise in the sella
turcica, they may occasionally be ectopic (3). These tumors exhibit a
wide range of hormonal and proliferative behavior.
They may be small lesions with a slow rate of growth. When hormonally inactive, such tumors are not usually detected clinically and therefore represent incidental findings discovered either as radiographic "incidentalomas" or at postmortem examination. When they produce hormones in excess, however, they can give rise to a severe clinical syndrome, such as acromegaly or Cushing’s disease, that can be lethal despite the relative lack of tumor growth.
Some pituitary adenomas are rapidly growing tumors that give rise to symptoms of an intracranial mass. They may cause hypopituitarism and/or visual field disturbances; they may invade locally downward into the paranasal sinuses, laterally into the cavernous sinuses, and upward into the parenchyma of the brain. These more aggressive tumors may be hormonally active, secreting any number of hormones in excess, or may be clinically nonfunctional.
B. Epidemiology
The true incidence of pituitary tumors has not been established
with certainty. With modern methods of imaging and biochemical analysis
of hormonal activity, the most recent data suggest that pituitary
adenomas are common, occurring in approximately 20% of the general
population.
Various studies have examined the incidence of such lesions at autopsy or in routine radiological evaluation of asymptomatic patients, yielding data on the development of incidental, slowly growing tumors that do not give rise to clinical symptoms of either a sellar mass or hormonal excess syndromes. Careful histological assessments have shown prevalences of 22.5% (4) and 27% (5). Using high resolution computed tomography or magnetic resonance imaging, approximately 20% of "normal" pituitary glands harbor an incidental lesion measuring 3 mm or more in diameter (6). The majority of these tumors that are asymptomatic are clinically nonfunctioning tumors that are now recognized to be of gonadotroph differentiation, or prolactinomas that have not caused clinical symptomatology that is recognized by the patient (7, 8). Interestingly, the sex incidence is equal in these studies, and the incidence increases with age in autopsy analyses so that more than 30% of people 50–60 yr of age harbor clinically undetected tumors.
Clinically diagnosed pituitary adenomas have been said to represent 10% of intracranial neoplasms (9). Improvements in radiographic imaging, biochemical detection of hormonal abnormalities, and microsurgical techniques have raised the number of surgical procedures, and in some series pituitary adenomas represent approximately 25% of surgically resected intracranial neoplasms (10); however, this may reflect a bias that reflects the interests of the surgeon or institution. Epidemiological data obtained before 1969 indicated annual incidence rates of up to 1.85 per 100,000 population (11) with geographic and racial variation; again, these figures may be low, and the diagnosis appears to be more frequent today because of increased awareness and improved diagnostic techniques. Prolactinomas are the most common type of adenoma; about one third of pituitary adenomas are not associated with clinical hypersecretory syndromes, but present with symptoms of an intracranial mass, such as headaches, nausea, vomiting, or visual field disturbances. GH- or ACTH-producing adenomas each account for 10–15% of pituitary adenomas, and TSH-producing adenomas are rare (9, 12, 13).
The relative frequency of the various adenoma types in surgical series varies with several factors, including geography and the therapeutic approach of the clinicians involved. For example, in some centers, prolactinomas are rare in surgical material because the endocrinologists prefer a medical approach to management (9, 13, 14, 15). There is usually a female preponderance in tumor occurrence. Women usually present at a younger age and have a higher incidence of PRL-secreting adenomas and ACTH-secreting tumors, whereas men tend to present in middle or older age with clinically nonfunctioning tumors (14).
Pituitary adenomas are infrequent in childhood. Only about 3.5–8.5% of pituitary adenomas are diagnosed before the age of 20 yr (16, 17). Childhood tumors exhibit a female preponderance, and some have suggested that they are smaller, less invasive, and less aggressive than tumors of adults (16). Hormone excess is common, and clinically nonfunctioning tumors that present with mass effects are rare. Tissue destruction results in loss of GH secretion with growth retardation. Patients with GH-secreting adenomas have an almost uniform incidence of PRL production by the tumor, and pure GH-producing adenomas are rare in children (16).
In random autopsies, 0.9% of pituitary adenomas identified were multiple (18). As expected of incidental adenomas encountered at postmortem examination, most were small and clinically silent. In another review of more than 3000 surgically resected pituitary adenomas, 11 were defined as "double adenomas" (19) and in 2 of these cases, hormone excess attributable to both tumors was manifest. In surgical series, multiple adenomas are rarely reported; synchronous detection of more than one tumor has been reported (18, 19, 20), and metachronous double adenomas have occurred in the same patient (21).
C. Classification
Pituitary adenomas have been classified by various groups of
investigators in different ways: A functional classification
of pituitary adenomas defines these tumors based on their hormonal
activity in vivo. This is the common clinical approach used
by endocrinologists. It characterizes the tumors as GH-producing
adenomas associated with acromegaly and/or gigantism, adenomas causing
hyperprolactinemia and its clinical sequela, ACTH-producing adenomas
associated with Cushing’s or Nelson’s syndromes, TSH-producing
tumors, the rare clinically detectable gonadotroph adenomas, and the
large group of clinically nonfunctioning or "endocrinologically
inactive" adenomas.
The anatomic or radiographic classification obtained by neuroradiological examination is based on tumor size and degree of local invasion. These data are of critical importance to the surgeon when planning an operative approach for tumor resection. The most widely used classification, proposed by Hardy in the 1970s (22), was based primarily on skull x-rays, pneumoencephalography, polytomography, and carotid angiography. It has been validated by the application of computed tomography scanning and magnetic resonance imaging, which are more accurate. This classification places adenomas into one of four grades.
1. Grade I adenomas, or microadenomas, are intrapituitary lesions that measure less than 1 cm in diameter. While these lesions may be detected with sophisticated imaging techniques, by definition they do not cause bony destruction to the sella that can be identified on conventional imaging.
2. Grade II adenomas are larger than 1 cm in diameter but still remain intrasellar or exhibit suprasellar extension without invasion. Sellar enlargement is usually identified but these tumors do not cause bony erosion.
3. Grade III adenomas are small or large locally invasive tumors that may be associated with diffuse sellar enlargement and may have suprasellar extension, but in either case cause bony erosion of the sella turcica.
4. Grade IV adenomas are large invasive tumors that involve extrasellar structures including bone, hypothalamus, and the cavernous sinuses.
A subclassification of grade I, II, and III tumors identifies the degree of suprasellar invasion as small (A), moderate (B), or large (C).
Invasive adenomas are a subject of controversy. Some have suggested that significant local invasion should be considered a sign of malignant potential (23). However, infiltrative pituitary tumors that can invade dura, bone, and the cavernous sinuses are relatively common (23, 24, 25), yet they do not exhibit the ability to metastasize; these are generally classified as benign but aggressive adenomas. Large invasive pituitary adenomas can invade the sphenoid bone and inferiorly, presenting as nasopharyngeal masses (26), or may invade posteriorly to involve or destroy the bony clivus (27).
The incidence of invasion varies depending on whether the lesion is examined grossly or microscopically. Invasive lesions are less frequently identified by imaging techniques or by the neurosurgeon than by the pathologist examining dural biopsies by light microscopy (24).
Invasiveness appears to correlate, to some extent, with tumor type and size. The most commonly invasive groups include thyrotroph adenomas and silent corticotroph adenomas (23, 25); in addition, the unusual plurihormonal silent subtype 3 adenomas are generally invasive (28). Macroadenomas are more often invasive than are microadenomas. Grossly invasive adenomas are recognized by the neurosurgeon and are usually not amenable to complete resection; however, there are no accepted markers to predict invasive behavior and forewarn possible recurrence for smaller lesions. Cytological features are not valid, since they do not differ in recurrent and nonrecurrent tumors. Ploidy analyses have not found aneuploidy to correlate with hormone profile or recurrence (29, 30). Some authors have suggested that the proliferation markers Ki-67, proliferating cell nuclear antigen (PCNA), or p105 (30, 31, 32, 33, 34, 35) may be useful in this regard.
Histological classification of pituitary adenomas before the era of immunohistochemistry and electron microscopy was a frustrating and unsuccessful exercise. These tumors were classified as acidophilic, basophilic, and chromophobic using conventional stains; acidophilic adenomas were said to be associated with acromegaly or gigantism, basophilic adenomas were thought to be the cause of Cushing’s disease, and chromophobic tumors were considered to be nonfunctioning from the endocrine perspective. However, the value of such classification was questioned when it became obvious that some chromophobic adenomas were associated with florid clinical symptomatology of hormone excess, and some acidophilic or basophilic adenomas were clinically hormonally inactive. The application of more sophisticated histochemical stains led to an enhanced classification of these adenomas, but still proved to be relatively insensitive, nonspecific, and therefore unreliable.
The development of immunohistochemical classifications based on the detection of antigens in tissue revolutionized the classification of pituitary adenomas. Since hormones are well recognized as antigenic substances by other species, this technology allowed the development of highly specific antisera to adenohypophysial hormones and precipitated the morphologist’s ability to accurately determine hormone content of tumor cells.
Currently, pituitary adenomas are classified by hormone content. This
functional approach most closely correlates with the clinical
presentation of the patients. The outline for this system is provided
in Table 1
. Other markers of cell
differentiation, such as the transcription factors that regulate
hormone expression and keratins, can also be used to classify and
subclassify pituitary adenomas by immunohistochemistry. Some of these
can obviate the need for ultrastructural examination except in unusual
situations. From the clinical perspective, hormonal activity is the
basis for diagnosis and therapy. Biologically, however, it remains to
be established whether other characteristics, such as proliferation
markers, growth factor, and receptor expression, or oncogene product
expression, will prove to be the most reliable predictors of tumor
behavior, such as invasive growth, recurrence, or metastasis. If these
markers are found to be useful in the guidance of therapeutic
management, the classification of these tumors will undergo a
revolution. Nevertheless, the application of immunohistochemical
staining methods to determine tumor cytogenesis and pathogenesis will
likely remain a mainstay of morphological classification.
|
In the family of glycoprotein-producing adenomas, there has been
controversy concerning the diagnosis of gonadotroph adenomas without
ultrastructural confirmation of cytodifferentiation. Previously,
immunolocalization of glycoprotein hormones was unreliable; there was
significant cross-reactivity, particularly because of the common
-subunit, and fixation led to artefactual negativity in some cases.
These problems have been reduced by the development of more sensitive
and specific antisera and improvements in tissue fixation for antigen
recognition. It now appears that the gonadotropic hormones, as detected
by antisera to ß-FSH and ß-LH, are present in many clinically
nonfunctioning adenomas. Some of these are recognized by electron
microscopy as having gonadotropic differentiation, but some have
characteristics of less well differentiated cells, resembling the
"null" cells that were initially thought to be undifferentiated
precursors of adenohypophysial cells (37). The vast majority of null
cell adenomas and the related group of tumors classified as oncocytomas
express transcription factors and glycoprotein hormone subunits that
allow characterization as tumors of gonadotroph differentiation (38); a
line of transgenic mice expressing simian virus 40 T antigen driven by
the ß-FSH promoter has provided an animal model of these adenomas
(39). The role of electron microscopy in the classification of these
tumors remains a subject of controversy, but since there is currently
usually little clinical impact, the need for this expensive and
time-consuming exercise remains academic.
For unusual plurihormonal adenomas, electron microscopy continues to play an important role in determining cytodifferentiation and structure-function correlations.
The ideal classification of any group of tumors is a
clinicopathological classification that correlates endocrine
manifestations and aggressiveness of pituitary adenomas with specific
morphological phenotypes. Table 2
summarizes a scheme that permits maximal structure-function
identification. Generally, aggressive behavior is a phenomenon of
silent adenomas and unusual plurihormonal adenomas as well as the rare
lactotroph adenoma with GH immunoreactivity, known as the "acidophil
stem cell adenoma." Additional information, such as tumor size,
radiological, gross or microscopic evidence of invasion, and the
proliferative activity of a tumor as identified by flow cytometry or
immunohistochemical proliferation markers, can be incorporated in a
multidisciplinary fashion to determine the optimal therapeutic approach
to management of the individual patient.
|
|
II. Pituitary Cytodifferentiation |
|---|
|
TopAbstractI. Introduction: Pituitary...II. Pituitary...III. Pathogenetic Mechanisms in...IV. Concluding CommentsReferences |
|---|
A. Early pituitary development
Novel transcription factors that play a role in anterior
primordial development are being identified at a rapid pace. Many of
these are implicated in early pituitary organogenesis.
The bicoid-related pituitary homeobox factor Ptx1 was initially proposed as an activator of POMC gene expression (44). It has subsequently been identified as an early determinant of brain and facial development that precedes pituitary development (45) and is subsequently expressed in all adenohypophysial cell types (46).
Pituitary homeobox factor 2 (Ptx2), structurally related to Ptx1, expresses two alternatively spliced mRNA products that encode two proteins of 271 and 317 amino acids. Ptx2 is expressed in the developing and mature pituitary as well as in eye and brain tissue (47), suggesting that it may play a role in the development and maintenance of these structures.
Two members of the Lhx gene family, a group of LIM homeobox genes, have been implicated in pituitary development but at an earlier stage of development than cell differentiation for hormone production (48). Lhx3 and Lhx4 direct formation of the pituitary at the stage when the oral ectoderm invaginates to form Rathke’s pouch; subsequently, Lhx3 remains expressed throughout the gland whereas Lhx4 develops a pattern of expression restricted to the anterior lobe. Null mutation of either Lhx3 or Lhx4 does not prevent formation of Rathke’s pouch, but animals devoid of both genes develop only a rudimentary pouch. Targeted disruption of Lhx3 alone prevents further differentiation of all adenohypophysial cells; lack of Lhx4 alone results in defective, but not absent, gonadotroph differentiation, suggesting that this transcription factor supports, but is not essential for, the development of those cells.
P-LIM is another LIM homeobox protein transcription factor that is selectively expressed in the pituitary with highest levels at the early stages of Rathke’s pouch development (49). It appears to be expressed in all pituitary cell types, however, and therefore is not a likely candidate for regulation of terminal cytodifferentiation.
Another early marker of pituitary differentiation is the Rathke’s pouch homeobox (Rpx) protein that is identified in the pituitary primordium before the onset of known cell-specific differentiation (50). The expression pattern appears to be more extensive than the area destined to become the adenohypophysis alone; it is likely, therefore, that Rpx is involved in the initial determination of the anterior region of the embryo. This factor is subsequently extinguished in the mesendoderm and ultimately becomes restricted to Rathke’s pouch. Down-regulation of Rpx occurs at the time of onset of other pituitary-specific transcription factors; failure to down-regulate this factor leads to pituitary hypoplasia in Ames dwarf mice (51), suggesting a role for this gene in modulation of early pituitary cell proliferation.
Another factor implicated in early pituitary development is the Prophet of Pit-1 (PROP-1). Inactivating mutations of PROP1, a paired-like homeodomain protein that is necessary for Pit-1 expression, have been identified as the cause of Pit-1 deficiency in Ames dwarf mice (52) and in humans with combined pituitary hormone deficiency (53, 54).
Id, a member of the helix-loop-helix family of transcription factors, is also found early in development and in some pituitary tumor cell lines but is decreased or absent in differentiated cells (55). Its role in pituitary development remains unclear.
B. Pit-1
Pit-1, also known as GHF-1, is a 33-kDa 291-amino acid protein
that belongs to the homeobox family of developmental regulatory
proteins (56, 57). The presence of an additional domain, conserved in
Pit-1 and the proteins OCT-1, OCT-2, and UNC-86, gave rise to the term
POU-domain, which characterizes this family of homeodomain proteins
(58, 59, 60).
This protein is notable for its transcriptional effects and pituitary-restricted expression. It binds the promoter sequences and activates the structurally related GH and PRL genes in rat and human (58, 61, 62, 63, 64). When expressed in heterologous cell types, Pit-1 is capable of activating reporter gene constructs containing the rat GH or PRL gene promoters (56, 65, 66), even when expressed at levels lower than in pituitary cells. Extinction of GH expression in fibroblast-pituitary cell hybrids is accompanied by loss of Pit-1 protein and mRNA expression (67).
The role of Pit-1 in cytodifferentiation was recognized when it was found that pit-1 gene expression in the developing pituitary is closely correlated with the onset of GH and PRL production (68, 69). Unexpectedly, it was noted that Pit-1 expression is also temporally associated with the onset of TSH production in the fetal rat adenohypophysis (69, 70). It was subsequently shown that the gene encoding the ß-subunit of TSH contains unique Pit-1 DNA-binding sites that bind Pit-1 but with lower affinity than sites in the GH or PRL gene 5'-flanking regions (71, 72). Further studies have shown that additional factors are required for TSH gene expression (73) and thyrotroph differentiation (74), and that there may be Pit-1-independent as well as Pit-1-dependent origins of this cell type (75). Several isoforms of Pit-1 result from alternative mRNA splicing; Pit-1ß (76, 77, 78) and Pit-1T (79, 80) may have specific functions in activating GH and TSH gene transcription selectively.
Using in situ hybridization and immunocytochemistry, pit-1 gene expression was identified by day 15 to 16 in the embryonic rat pituitary, preceding PRL and GH gene activation (69). Pit-1 mRNA transcripts were subsequently detected in all five phenotypically distinct pituitary cell types; however, Pit-1 protein was detected only in the nuclei of somatotrophs, lactotrophs, and thyrotrophs (69), suggesting that translational controls may dictate the pattern of rodent Pit-1 expression. However, studies of human pituitary adenomas that represent relatively homogeneous cell populations have shown that transcriptional control dictates selective expression of the pit-1 gene in human adenohypophysial cells responsible for GH, PRL, and ß-TSH synthesis (81, 82, 83). Detection of pit-1 mRNA transcripts in human pituitary cells also correlates with the localization of Pit-1 protein by immunocytochemistry.
These differences could be attributed to species-specific variation in
pit-1 gene expression or decreased mRNA stability in certain
human adenohypophysial cell types (e.g., corticotrophs,
gonadotrophs). Comparison of 5'-flanking (
0.4 kb) and
5'-untranslated regions in human, rat, and mouse pit-1 genes (84)
reveals highly conserved sequences only near the TATAA box,
transcription and translation start sites, and at PB1 and PB2 elements
— two autoregulatory Pit-1-binding sites (85). Interestingly, binding
sites for the cAMP-regulated transcription factor CREB (cAMP response
element-binding protein) are present in rodent but not human
pit-1 promoter sequences, suggesting that
species-dependent differences in regulation of the pit-1
gene may occur. However, many of the other similarities observed in
5'-regulatory sequences of rat and mouse pit-1 genes are
more likely to reflect the recent divergence of these rodent species
rather than functional constraints on pit-1 gene
evolution within mammalian subclasses.
In the human fetal pituitary, Pit-1 mRNA and protein are identified as early as 6 weeks of gestation; ACTH immunoreactivity is detected 1 week later, and GH immunoreactivity is detectable at 8 weeks (86). Pit-1 is found only in cells containing GH, PRL, and/or TSH throughout human gestation. These results are consistent with those reported in rodents (68, 69) where Pit-1 appears by day 15–16, immediately preceding the onset of PRL and GH mRNA. In contrast, the sequence of cytodifferentiation differs in the human, since PRL, ß-TSH, and the ß-subunits of the gonadotropins only appear 4 weeks later at 12 weeks of gestation (40, 41, 42, 43). These findings support the hypothesis that Pit-1 is insufficient for the cytodifferentation of lactotrophs and thyrotrophs that occurs much later than the onset of Pit-1 expression. The prolonged time span of human adenohypophysial cytodifferentiation allows careful and accurate dissection of the factors that must be required to act in concert with Pit-1 to promote the subsequent expression of PRL and ß-TSH.
In rodents and humans, differentiation and/or maintenance of somatotroph, lactotroph, and thyrotroph phenotypes are dependent on expression of a functional pit-1 gene; mutations in the pit-1 gene result in hypopituitarism (84, 87, 88, 89) and hypoplasia of somatotrophs, lactotrophs, and thyrotrophs (87). An interesting observation is that Pit-1 mRNA and protein are highly expressed during human pituitary development at 17–19 weeks (86) when GH levels are extremely high (42) and near term (86) when there is proliferation of lactotrophs (41). These data suggest that Pit-1 plays an important role not only in the differentiation process, but also in the regulation of hormonal activity and possibly also of cell proliferation. It has also been shown that pit-1 antisense oligonucleotides not only block GH and PRL transcription but also inhibit [3H]thymidine incorporation by somatotroph and lactotroph cell lines, suggesting that Pit-1 may regulate DNA replication and cell proliferation (90). This effect could be direct, similar to that of the POU-domain transcription factor OCT-1, which can stimulate viral DNA replication (91), or indirect, by regulation of mitogen function. For example, the receptor for GHRH, a member of the G protein-coupled receptor family, is coexpressed with Pit-1 and may be regulated by Pit-1 (92, 93). GHRH is known to play a role in the development and proliferation of GH-producing pituitary adenomas (see below). The cell type-specific expression of Pit-1 in human pituitary adenomas suggests a possible role for this transcription factor not only in the determination of cell phenotype and hormonal activity of these neoplasms but also in the regulation of pituitary tumor growth. To date, however, there has been no evidence of a correlation between Pit-1 expression and tumor growth in human pituitary adenomas (81, 82, 83).
C. CREB
CREB binding sites are present in many gene promoters, and the
factors that bind these sites are implicated in the regulation of
numerous hormone genes. In the anterior pituitary, the Pit-1 gene
promoter (94, 95) and the human -subunit gene promoter (96) appear
to be regulated by cAMP via CREs. Transgenic mice that overexpress a
dominant negative CREB exhibit dwarfism with somatotroph hypoplasia
(97). Although the ubiquitous nature of CREB makes it an unlikely
candidate to control cell-specific differentiation, it appears that in
concert with other factors, this transcription element plays an
important and necessary role in somatotroph development.
D. Estrogen receptor (ER)
A number of studies have established that estrogen acting directly
through its receptor regulates PRL gene transcription, synthesis, and
secretion (98, 99, 100, 101, 102, 103, 104, 105, 106, 107). The PRL promoter contains a nonpalindromic
estrogen response element that functions as weak transcription
activator that is enhanced by cooperation with Pit-1 to activate PRL
gene transcription (107). Many studies also demonstrate a role for
estrogen in mediating a positive or negative effect on the expression
of the ß-FSH and ß-LH hormone genes and on levels of secretion of
these gonadotropins (108, 109, 110). There is direct evidence that the
classic ER (ER) binds to the upstream region of the rat ß-LH gene
(111).
The sequence of differentiation of adenohypophysial cells in the human
fetal pituitary, in contrast to the rodent gland, implicates a
transcription activator that is distinct from Pit-1, is common to
lactotrophs and gonadotrophs, and has its onset at or just before 12
weeks of gestation in the human fetal pituitary (40, 41, 42, 43). While
ACTH-containing corticotrophs differentiate at 6–7 weeks, and
GH-containing somatotrophs appear at 8 weeks, PRL is not expressed
until 12 weeks of gestation. At 9 weeks, there are cells that contain
-subunit of the glycoprotein hormones, but the ß-subunits of TSH,
FSH, and LH are also only detectable at 12 weeks. These data suggest
that ER may be implicated in the regulation of hormone production
and cytodifferentiation of mammosomatotrophs/lactotrophs and
gonadotrophs in a cell type-specific fashion.
Studies using [3H]estradiol binding implied ER expression
in 85% of cells in the anterior lobe but not in intermediate or
posterior lobe cells (112); uptake of radiolabeled estrogen was
reportedly found in cells containing immunoreactivity for PRL, ß-FSH,
ß-LH, ß-TSH, and GH (113); however, the specificity of these
reactions was not established. Immunohistochemical studies to localize
ER in the human pituitary and its adenomas were initially
unsuccessful due to the limited sensitivity of the detection method
employed and/or low levels of ER protein expression in this tissue
(114). Using antigen-retrieval methods, however, ER can be localized
by immunocytochemistry in the nontumorous adenohypophysis (115, 116) in
cells containing PRL or gonadotropin ß-subunits. The localization of
ER in thyrotrophs is controversial. GH-immunoreactive cells
containing nuclear positivity for ER may be mammosomatotrophs that
are known to exist in the human pituitary (117). ACTH-containing cells
are reported to be negative for ER (115, 116).
Biochemical analyses have demonstrated that ER is most reliably
localized in PRL-producing adenomas (115, 116, 118, 119). However, this
detection method requires large amounts of protein and has relatively
low sensitivity; comparative studies have shown no correlation between
detection of ER mRNA and the presence or amount of protein detected
by radioactive ligand binding (115). The closest correlations between
hormone production and ER expression have been documented using a
ribonuclease (RNase) protection assay (116) and RT-PCR (115), which are
the most sensitive and specific methods to identify even low levels of
expression. These studies have documented correlation between ER
expression and the production of PRL or gonadotropins (115, 116);
splice variants of ER mRNA are also selectively expressed by certain
types of pituitary adenomas (120). Corticotroph adenomas do not express
ER. Somatotroph adenomas that do not produce PRL as well as GH are
devoid of ER; the lack of ER expression in these cells suggests
that the GH-releasing activity of estrogen (121) is either mediated by
other pathways or involves a selective effect on mammosomatotrophs. As
previously suggested, PRL expression was more consistently found in DG
than in sparsely granulated somatotroph adenomas (122), indicating the
similarity between DG-GH tumors and mammosomatotroph adenomas, and
ER expression had the same pattern.
These data suggest that ER may be the factor responsible for the
development of PRL expression in somatotrophs that express Pit-1. This
factor must have its onset after GH expression during gestation, since
two models of disruption by targeting of diphtheria toxin (123) or by
thymidine kinase obliteration (124) in GH-expressing cells prevent
further development along this pathway. Clearly there must also be a
factor responsible for silencing GH expression in the progression from
mammosomatotrophs to mature lactotrophs (42, 125, 126). Regulation of
ER expression could account for the fluctuations in adenohypophysial
cell populations during pregnancy, when there is transition from
somatotrophs to mammosomatotrophs and lactotrophs (127). Preliminary
data suggest that ER expression is initiated in the fetal pituitary
around 12 weeks of gestation (128); if so, it would explain the
development of PRL secretion and the differentiation of gonadotrophs at
that gestational age.
Mice with disrupted ER display gonadal maldevelopment and consequent
elevated gonadotropins; their circulating PRL levels are decreased but
not undetectable (129, 130). Structurally, there is evidence of
lactotroph differentiation (131). A human with an ER mutation has
also been described (132); he too demonstrated similar hormonal
profiles. These data would suggest that ER is not required for
lactotroph or gonadotroph differentiation; however, the description of
the ERß gene and analysis of its distribution in human tissues (133)
indicate the redundancy of this system. Further studies involving
disruption of both ER genes are required to clearly define the role of
ER in pituitary cell differentiation.
Estrogen has been implicated as a lactotroph growth-stimulating factor. Lactotrophs are known to proliferate during pregnancy (134, 135), and a few lactotroph adenomas may grow during gestation (136). Administration of oral contraceptives was associated with a rapid increase in size and secretion of some lactotroph adenomas (137), and estrogen therapy has been implicated in the pathogenesis of a lactotroph adenoma in a male-to-female transsexual (138). Just as Pit-1 has been postulated to regulate DNA replication and cell proliferation (90), the cell type-specific expression of ER in human pituitary adenomas suggests a possible role for this transcription factor not only in the determination of cell phenotype and hormonal activity of these neoplasms but also in the regulation of tumor growth.
E. Thyrotroph embryonic factor (TEF)
A putative thyrotroph-specific factor has been described; TEF is a
trans-acting factor that belongs to the leucine zipper gene
family of transcription factors that is thought to activate the
expression of the human ß-TSH gene (74). TEF is expressed in a
pattern that correlates temporally and spatially with ß-TSH gene
expression in the rodent fetal pituitary (74); subsequently, it is
expressed in several other tissues. The proximal ß-TSH promoter
contains three independent TEF-binding sites, and TEF is able to
activate a reporter gene under the control of that promoter.
These data suggest an intriguing possibility that TEF is the factor required for the onset of TSH production in cells that produce Pit-1. The relationship between thyrotrophs and somatotrophs has been recognized previously in rats with prolonged hypothyroidism; the development of thyrotroph hyperplasia is associated with transdifferentiation of somatotrophs into thyroidectomy cells (139). It is therefore likely that there is a continuum and that the maturation from somatotrophs to differentiated thyrotrophs requires both the onset of TEF expression and the production of a GH silencing factor.
F. Steroidogenic factor-1 (SF-1)
The nuclear receptor SF-1 is a member of the steroid receptor
superfamily that was identified independently in mouse (140) and bovine
steroidogenic tissues (141) and was shown to be a transcription factor
that regulates the expression of the steroidogenic enzymes cytochrome
P450 CYP11A and CYP11B. The factor is also known as Ad4BP since it
binds to the Ad4 site in the 5'-region of the bovine cytochrome P450
CYP11A and CYP11B genes (141, 142, 143). SF-1 is expressed by all zones of
the adrenal cortex, granulosa, and theca cells of the ovary and Leydig
cells of the testis (144, 145). In situ hybridization has
demonstrated that the expression of SF-1 has its onset at day 9 of
mouse gestation in the urogenital ridge (146) with expression in the
adrenal anlage at day 12 when the cytochrome P450 enzymes are initially
expressed in that tissue (145). Expression of SF-1 is sexually
dimorphic in the developing gonad where it may play a role distinct
from the regulation of the steroidogenic enzymes (146). SF-1 also
regulates the Müllerian inhibiting substance (MIS) gene to
determine Müllerian duct regression in the developing embryo
(147). Targeted disruption of this gene shows that it is essential for
adrenal and gonadal development and sexual differentiation (148).
The factors accounting for gonadotroph differentiation remained
unclear until it was recently demonstrated that SF-1 is necessary for
the differentiation of pituitary gonadotrophs (149) as well as for the
formation of the ventromedial nucleus of the hypothalamus (150). SF-1
is expressed in the embryonic mouse forebrain and in the developing
mouse pituitary before the onset of expression of the gonadotropin
ß-subunits (146, 149). SF-1 mRNA transcripts are detected in normal
mouse gonadotrophs and in an immortalized murine pituitary
gonadotroph-derived cell line (T3–1), and the protein interacts
with a regulatory element in the murine gonadotropin -subunit gene
to enhance transcription (149, 151). Disruption of this gene in mice
results in gross impairment of the development of the ventromedial
hypothalamic nucleus (150) and pituitary glands that lack gonadotropin
immunoreactivity (149); although the GnRH hypothalamic neurons are
present in normal numbers and location (150), GnRH receptor is not
expressed in the pituitaries of these animals (149). These data suggest
that SF-1 plays a role in pituitary gonadotroph differentiation,
development, and function.
Studies of human pituitaries indicate that SF-1 is also involved in cell-specific hormone expression in human adenohypophysial cells. In human tissue, there is close correlation between gonadotropin production and SF-1 expression (38). In the nontumorous gland, SF-1 is expressed and is localized in the nuclei of gonadotropin-containing cells but not in other cell types. In the relatively homogeneous populations of tumors, SF-1 expression is characteristic of gonadotropin-producing adenomas, including the classic gonadotroph adenomas and also null cell adenomas and oncocytomas that are known to produce gonadotropins (152).
SF-1 is expressed almost exclusively in cells that produce the
gonadotropin ß-subunits in the human pituitary and in pituitary
adenomas. Interestingly, SF-1 expression does not correlate with
-subunit production in the human gland, since many GH-producing
nontumorous cells and adenomas express -subunit but not SF-1. It
appears that SF-1 expression is initiated in the fetal pituitary at 12
weeks of gestation (128); if confirmed, this finding would explain the
development of ß-subunit gonadotropin secretion and the
differentiation of gonadotrophs at that gestational age.
G. Corticotroph upstream transcription-binding element (CUTE)
Corticotrophs are the first cells to differentiate in the human
fetal pituitary (41, 42). Although expression of the POMC gene is one
of the most promiscuous events in endocrine tumors outside the
pituitary, adenohypophysial corticotroph lineage is one of the most
stable, since expression of POMC is rarely associated with expression
of other adenohypophysial hormones. Nevertheless, the factors
determining this lineage remain unclear.
The POMC promoter contains an E box motif typical of binding sites for the helix-loop-helix (HLH) transcription factors. A protein with characteristics of an HLH factor that binds to the POMC promoter was identified in nuclear extracts of the murine pituitary corticotroph cell line AtT-20 cells and named CUTE. This protein has been identified in various cells expressing POMC, but not in other pituitary-derived cell lines, and has therefore been implicated as an important determinant of cell-specific expression of the POMC gene in the pituitary and other sites (153). Subsequent studies have identified the HLH transcription factor NeuroD1/beta2 in CUTE complexes (154). The role of this factor in determining cell differentiation remains to be elucidated.
H. Other putative factors
Zn-15 is a zinc finger transcription factor with an unusual
DNA-binding domain. It binds the proximal GH promoter; in transient
transfection studies, it stimulates GH expression and shows synergistic
effects with Pit-1 (155). Little is known of its potential role in
pituitary cytogenesis.
The superfamily of Ets transcription factors is also recognized to play significant roles in the control of growth and development. Cotransfection of Ets-1 and Pit-1 results in synergistic activation of the PRL promoter, suggesting that this factor may mediate ras activation of pituitary-specific gene expression (156, 157). Again, however, it remains to be seen whether this factor is involved in lactotroph differentiation.
Glucocorticoid receptors (GCRs), thyroid hormone receptors (THRs), and retinoic acid receptors play essential roles in transcriptional regulation of pituitary hormones, but these are not expressed in cell-specific fashion and, therefore, are not considered to control terminal cell differentiation. Their role in tumor development is discussed below.
I. A model of adenohypophysial cytodifferentiation
The various transcription factors discussed above have been shown
to regulate cell differentiation and hormone production in the
pituitary. They provide the framework for a new model of cell lineage
in the adenohypophysis (Fig. 1).
Information gleaned from analysis of human pituitary tumors and human
fetal pituitary development has clarified the significance of the new
models. Structure-function correlations at the molecular level have
provided a clearer understanding of the hormonal activity of pituitary
adenomas. Old concepts of plurihormonality have taken on new
significance as these factors are shown to account for the patterns of
hormone expression that have long been recognized in human pituitary
adenomas.
|
|
III. Pathogenetic Mechanisms in Pituitary Adenoma Development and Progression |
|---|
|
TopAbstractI. Introduction: Pituitary...II. Pituitary...III. Pathogenetic Mechanisms in...IV. Concluding CommentsReferences |
|---|
A. Hormonal factors
Evidence supporting a hormonal etiology includes 1) paradoxical
pituitary hormone responses to exogenous hormonal stimulation that are
characteristic of pituitary adenomas, 2) the development of pituitary
adenomas in situations of excessive hypothalamic hormone stimulation or
reduced feedback suppression by target gland hormones, and 3) evidence
of hypothalamic hormone production within the anterior pituitary that
suggests a role for local excess stimulation.
Persuasive arguments against a hormonal etiology, however, are the rarity of hyperplastic changes associated with adenomas, the lack of true adenomatous changes in the pituitary even after long and sustained exposure to hypothalamic hormone stimulation in some instances, and the low frequency of recurrence after successful tumor resection. Additionally, some pituitary adenomas have been shown to lack hypothalamic hormone receptor synthesis.
1. Stimulatory hormone excess.
a. GH-releasing hormone (GHRH).
The putative role of
hypothalamic stimulation in pituitary tumor development has received
support from a substantial body of evidence. GHRH can cause somatotroph
proliferation (159), and somatotroph hyperplasia is well documented as
a consequence of chronic stimulation in patients with extrahypothalamic
tumors secreting GHRH (160, 161). In addition, hypothalamic tumors
containing GHRH have been associated with sparsely granulated
somatotroph adenomas (162). A number of studies have indicated that the
pituitary and pituitary adenomas produce GHRH locally (163, 164, 165), and
GHRH may be overexpressed in aggressive tumors (166). In
vitro, human somatotroph adenoma cells are known to respond to
GHRH stimulation (167, 168, 169, 170, 171, 172), indicating the presence of GHRH receptors
on these tumors, but they are known to lack the down-regulation
characteristic of normal somatotrophs (170, 171). Thus, it would appear
that GHRH stimulation may play a role in the development of these
tumors. Moreover, transgenic mice overexpressing GHRH have proven that
prolonged chronic GHRH overstimulation alone can result in tumor
formation (Fig. 2) (173, 174). However,
in situations of GHRH excess, both in mice and in men, pituitary
GH-producing adenomas were associated with hyperplasia of GH-producing
cells, a phenomenon that is distinctly rare in patients with sporadic
pituitary GH-producing pituitary adenomas (175). Moreover, in humans,
continuous overstimulation by ectopic GHRH does not usually result in
true adenoma formation (176).
|
b. CRH.
The postulated etiology of Cushing’s disease has
shown tremendous flux since Cushing’s description in the 1930s of a
primary pituitary disorder (180). In the 1940s, the documentation of
adrenal hyperresponsivness to ACTH and the presence of Crooke’s
hyalinization in the pituitary brought a primary adrenal etiology to
the fore. In contrast, the autopsy documentation of lesions in the
paraventricular nucleus of patients with Cushing’s disease suggested
that the hypothalamus may be the site of primary pathology; this was
supported by reports of Cushing’s disease associated with increased
intracranial pressure due to intracranial tumors with regression of
cushingoid symptomology after tumor removal. In the 1950s, the trend
implicating the pituitary as the site of primary pathology returned,
with the recognition of the therapeutic efficacy of pituitary
irradiation or microsurgical removal of an adenoma. Nevertheless, it
has been recognized in the last two decades that patients with
Cushing’s disease may have other associated neuroendocrine and
electroencephalogram abnormalities. Reports of therapeutic response to
antiserotoninergic or antidopaminergic agents reverted attention to the
hypothalamus (181). Long-term follow-up of patients who have undergone
transsphenoidal resection of microadenomas has indicated recurrence of
disease in some patients. A few patients with pituitary Cushing’s
disease have corticotroph hyperplasia as the cause of the disorder in
the absence of a discrete adenoma (9). These findings have implicated
CRH excess in the pathogenesis of Cushing’s disease (180).
The characterization of CRH in 1981 permitted its identification in a number of extrapituitary tumors associated with a clinical picture resembling Cushing’s disease; some of these patients had corticotroph hyperplasia (182, 183). In one instance, a hypothalamic gangliocytoma producing CRH was associated with corticotroph hyperplasia and Cushing’s disease (184). These experiments of nature suggested that CRH may play a role in the proliferation of corticotrophs. Animal studies using continuous infusion of CRH have confirmed that prolonged exposure to CRH leads to increased numbers of corticotrophs (185, 186, 187), but it is yet to be demonstrated that CRH alone can cause pituitary corticotroph adenoma formation. As is the case with GH-producing tumors, the pathogenesis is probably multifactorial, and CRH may play a role in the promotion of tumor cell proliferation.
In vitro studies show that CRH treatment of pituitary adenomas increases ACTH and POMC mRNA gene expression (188) while dexamethasone inhibits it (189, 190). Additionally, CRH receptor expression appears not only intact in ACTH-producing pituitary adenomas but, unlike in the rat pituitary, is up-regulated in response to CRH treatment (191). There is currently no evidence of constitutive activation of CRH receptors in corticotroph adenomas. The closely related vasopressin V3 receptor is also intact but may be overexpressed in some corticotroph adenomas, and it has been suggested that it may play a role in tumor development (192).
c. TRH.
Patients with longstanding primary hypothyroidism
develop pituitary thyrotroph hyperplasia and associated lactotroph
hyperplasia; this proliferation has been attributed to TRH stimulation.
These patients exhibit a spectrum of hyperplasia and neoplasia (193, 194), suggesting that continuous stimulation by TRH may lead to
thyrotroph adenoma (195). TRH has been reported to be produced locally
by adenohypophysial cells (163, 196, 197) and by pituitary tumors of
several types, including prolactinomas (198, 199), GH-producing
adenomas (200), and nonfunctioning adenomas (201).
TRH signaling appears to be intact in pituitary adenomas as evidenced by normal binding and TSH and PRL release from thyrotrophs and lactotrophs, respectively (200, 201). Competitive PCR analysis reveals variable levels of TRH receptor expression among pituitary adenomas of the same type but generally similar to that of the normal pituitary (202). While the TRH receptor structure is grossly unaltered in functional pituitary adenomas and there is no evidence for an activating mutation even in thyrotroph adenomas (203), this gene is alternatively spliced in some pituitary tumors (202). Deletion of exon 3 results in a truncated product that neither binds TRH nor is activated by it. The relatively higher levels of the truncated forms compared with the full-length forms of the TRH receptor in lactotroph adenomas (202) may explain some of the pathological in vivo responses to TRH administration (204).
d. GnRH.
The occurrence of gonadotroph adenomas in patients
with hypogonadism has suggested that the chronic stimulation resulting
from primary gonadal failure may play a role in the formation and
growth of these adenomas (205, 206). Nevertheless, the majority of
gonadotroph adenomas are not associated with underlying hypogonadism or
evidence of chronic hypothalamic stimulation in the adjacent
nontumorous adenohypophysis (175) and appear to arise spontaneously.
Both GnRH (207, 208) and GnRH receptor expression have been documented
by multiple techniques in the different types of pituitary adenomas
(209, 210). Furthermore, pituitary adenomas with truncated GnRH
receptors have been described, and these appear to fail to respond to
GnRH stimulation by enhancing calcium transport and gonadotropin
release in vitro (Fig. 3)
(211). Activating mutations of the GnRH receptor have thus far not been
identified in pituitary adenomas.
|
2. Loss of inhibitory hormone regulation.
a. Dopamine.
The role of decreased hypothalamic inhibition was
supported by some authors who found vascular changes including
arteriogenesis in lactotroph adenomas. They speculated that the
neovascularization from the systemic circulation, which has negligible
levels of dopamine, allowed lactotrophs to escape dopaminergic tonic
inhibition (213).
Dopamine signal transduction is mediated through D1 receptors that stimulate adenylyl cyclase activity and D2 receptors (D2R) that inhibit this enzyme. The family of dopamine receptors is much more complex in terms of biochemical, physiological, and pharmacological diversity (214, 215, 216). Nevertheless, it appears that the predominant anterior pituitary dopamine receptor is the D2R (215, 217). Activation of the D2R results in altered cAMP production, potassium and calcium channel fluxes, phosphatidyl inositol turnover, and intracellular calcium concentrations (214). Selective elimination of D2R action in D2R knock-out mice results in lactotroph hyperplasia (218) and, subsequently, lactotroph adenoma formation (219). Dopaminergic regulation of thyrotroph adenomas has been shown to be abnormal but is also highly variable (220, 221, 222, 223). While some tumors can be suppressed by dopamine (220, 222), the dopaminergic resistance that is found in some of these tumors may implicate altered or absent dopamine receptors as an etiological factor (221). Thus far, however, investigation of the D2R gene has revealed it to be structurally intact in human lactotroph adenomas as well as in adenomas that secrete GH or TSH (224). More detailed characterization of the D2R and coupling proteins in tumors from patients with variable dopamine sensitivity is required to resolve this matter.
b. Somatostatin (SS).
GH secretion is under dual hypothalamic
influence by GHRH, which stimulates, and SS, which inhibits GH
secretion. Specific receptors for SS (SSTRs) are expressed on
somatotroph adenomas. Earlier studies suggested a relationship between
the density of SS receptors on GH tumors and the secretory response to
this analog both in vitro and in vivo (225, 226).
Binding sites for SS, however, have also been documented by
autoradiography in tumors resistant to the GH-lowering effects of
octreotide (227). These findings are consistent with differential
adenylyl cyclase coupling by the five subtypes of SSTRs and their
heterogenous mRNA expression in pituitary adenomas (228). Additionally,
expression of SS in large invasive GH tumors appears to be reduced
compared with that in the normal pituitary (163, 164, 229, 230). Taken
together, these findings suggest multiple paracrine, autocrine, as well
as endocrine mechanisms for SS-mediated control of somatotroph function
and proliferation.
c. Glucocorticoid hormones.
Patients with Addison’s disease
are known to develop "adrenalectomy" cells and, with prolonged
glucocorticoid deficiency, pituitary corticotroph hyperplasia. Very
extended disease is life threatening and is rarely seen, but in the few
cases examined, there is evidence of early tumor formation (231). A
role for CRH in mediating the cell proliferation cannot be excluded.
Lack of suppressibility of corticotroph adenomas by glucocorticoids was suggested by one study (232) as a mechanism involved in the pathological ACTH secretion in Cushing’s disease and Nelson’s syndrome. This report has not been confirmed by other investigators (233).
The human GCR pre-mRNA is alternatively spliced to generate a GCR
-isoform and the N-terminally closely related ß-isoform (234). The
ß-isoform, however, differs in its 50-amino acid C terminus, which
contains a unique 15-amino acid sequence that hinders glucocorticoid
binding and gene transactivation (234). The functional intermodulatory
relationship between the two GCR isoforms in the pituitary and
pituitary adenomas will undoubtedly be the focus of future studies.
Nevertheless, a molecular basis for glucocorticoid insensitivity has
already been described in association with generalized or selective
loss of function (234). Specific point mutations resulting in
diminished ligand binding in the glucocorticoid hormone-binding domain
are now known in cases of familial glucocorticoid resistance (235), and
a novel germ line mutation of this sort has been reported to result in
pituitary Cushing’s disease (236). Similarly, rare reports of somatic
mutations in the GCR with diminished glucocorticoid inhibition were
noted in Nelson’s syndrome (237) and ectopic Cushing’s syndrome
(238).
d. Thyroid hormones.
The development of pituitary
thy-rotroph adenomas in patients with prolonged primary
hypothyroidism has been interpreted as evidence of the
antitumorigenic role of feedback hormones in the pituitary
(193, 194, 195). However, as indicated above, the associated lactotroph
hyperplasia has provided evidence for the role of TRH stimulation in
the development of these adenomas.
Thyroid hormones mediate their actions via nuclear THRs that bind to
specific regulatory hormone response elements (239, 240). There are two
major classes of THRs, designated as and ß, which undergo
alternative splicing to generate 1 and 2 and ß1 and ß2
isoforms (239, 240). With the exception of the ß2 form, which
predominates in the hypothalamic-pituitary axis, these receptor
isoforms are ubiquitously expressed. Of interest in the pituitary, the
ß1 and ß2 isoforms appear to be expressed to a lesser extent in
endocrinologically inactive adenomas compared with the normal gland
(239, 240). These findings, although preliminary in nature, raise the
possibility that diminished negative feedback inhibition resulting from
reduced THR expression may play a role in the inappropriate peptide
release and cell growth associated with endocrinologically inactive
pituitary adenomas. The putative differential hormone-regulatory and
mitogenic effects of the different THR isoforms in the pituitary remain
to be clarified.
e. Gonadal hormones.
The development of pituitary gonadotroph
adenomas in patients with prolonged primary hypogonadism suggests that
lack of feedback hormone suppression may cause tumors in the pituitary
(205, 206); however, again, the role of GnRH stimulation cannot be
distinguished from that of gonadal hormone inhibition in the
development of these adenomas.
B. Molecular events
Evidence in favor of intrinsic pituitary cell defects accounting
for the development of these lesions is based primarily on the
monoclonal nature of these tumors. Although pituitary adenomas are
monoclonal, somatic mutations that have been identified in other
malignancies are usually absent, and the molecular events leading to
pituitary tumorigenesis remain unknown. Mutations involving ras,
p53, protein kinase C (PKC), c-erbB2 (neu), and
retinoblastoma (Rb) genes are rare or absent in these neoplasms. Only a
small fraction of adenomas have activating mutations of the Gs. An
obvious candidate gene is provided in multiple endocrine neoplasia type
1 (MEN-1), which is characterized by the development of pituitary
adenomas. Loss of heterozygosity (LOH) at the MEN-1 gene locus is rare
in sporadic adenomas; the recent cloning of the MEN-1 gene has allowed
more specific analysis of the structure of this putative tumor
suppressor in pituitary tumors; although germ-line mutations and LOH
are frequently encountered in familial adenomas, sporadic tumors
exhibit a low frequency of mutations and LOH, and
menin mRNA expression appears to be intact in most sporadic
adenomas.
1. Clonality. The technique of clonality assessment using X
chromosome inactivation patterns has evolved from the Lyon hypothesis,
which states that only one X chromosome is active in any mature female
somatic cell; the inactivation occurs early in embryogenesis and
persists throughout the lifespan of the cell and its progeny. Several
studies using techniques based on X chromosome inactivation have shown
that pituitary adenomas exhibit a pattern of monoclonality (Fig. 4) (241, 242). Most lesions found to
display a polyclonal pattern were found to be contaminated with normal
pituitary tissue; the small size of adenomas associated with Cushing’s
syndrome, in particular, has confounded the interpretation of the
clonal status of these tumors (242, 243, 244).
|
-subunit dissociates from the ß- and 
-subunits of the
stimulatory protein Gs when GTP displaces its bound GDP, stimulating
adenylyl cylase to produce cAMP from ATP (245). cAMP, in turn,
activates cAMP-dependent protein kinases, increases intracellular
calcium transport, and may potentiate the effect of activated inositol
phospholipid-dependent protein kinases. The weak intrinsic GTPase
activity of Gs and the action of GTPase-activating peptides
dissociates GTP from Gs and terminates the response. Additionally,
the multiple structural and functional isoforms of adenylyl cyclase
underscore the complexity of this redundant system of signal
transduction coupling and provide some insight into the array of
potential somatic mutations that could alter both pituitary cell
division and hormone production. At least three subunits (Gs, Gi, Gq)
are now known to be involved in cell signaling (246). The stimulatory
Gs is involved in the GHRH pathway, the inhibitory Gi in the SS
pathway, and Gq in the TRH- and GnRH-signaling pathways.
One of the earliest and most exciting molecular defects to be described
in endocrine oncology involved single-point mutations in two critical
domains of the Gs: codon 201 where Arg is switched to a Cys or codon
227 where Gln is replaced with Arg. Substitutions at these codons
activate adenylyl cyclase by inhibiting the hydrolysis of GTP, thereby
maintaining Gs in a constitutively activated state. These mutated G
proteins, also known as gsps, were first described in a
subset of somatotroph adenomas (245, 247). Subsequent studies, however,
have identified this mutation in nonfunctional pituitary adenomas (248, 249) and in other functional pituitary neoplasms (250).
Interestingly, no correlation has yet been found between the presence
of the gsp mutation with patient age, sex, tumor size, or
circulating levels of GH or IGF-I. Some investigators have reported
that acromegalic patients with tumors that exhibit gsp
mutations have higher circulating GH levels than patients whose tumors
lack such mutations (251), but this has not been a consistent finding
(252). Furthermore, the presence of this mutation appears to correlate
with a DG ultrastructural morphology of tumorous somatotrophs (253) and
possibly with greater GH responsiveness to inhibition by the SS analog
octreotide (254). To gain more insight into the functional consequences
of the gsp mutation, other intracellular cAMP targets have
been investigated in adenomas with and without the gsp
mutation. Tumors with gsp mutations are associated with
higher circulating levels of the free glyco-protein -subunit,
and its production by tumor cells in vitro was also found to
be significantly higher by tumors with the mutations than by those
without gsp mutations (252). Additionally, the
cAMP-responsive transcription factor CREB has been found, by Western
blotting, to be elevated in its phosphorylated (activated) form in
somatotroph adenomas with gsp mutations compared with
nonfunctional adenomas (Fig. 5) (255). It
remains to be shown if detection of these intracellular targets will
serve as consistent ancillary markers for the presence of the
gsp mutation.
|
Based on the findings that mutations of Gq result in constitutive
activation of phospholipase C and possess transforming potential (258),
pituitary adenomas of the various types were screened for mutations in
this G protein. No mutations were identified in the conserved
GTP-binding and hydrolysis domains of Gq or the highly similar
G11 (203, 259).
In contrast to the stimulatory effects of the Gs mutations,
inactivating mutations in the -subunit of the inhibitory
Gi2-coupling protein gip2 have been identified. A substitution of
glutamine for arginine at codon 205 has been described in
endocrinologically inactive pituitary adenomas (249).
b. Ras.
A family of three related ras
protooncogenes (H-ras, K-ras, and
N-ras) each encode a 21-kDa protein with intrinsic GTPase
activity (260). Ras proteins share common structural and
functional properties with membrane-anchored G protei
