Growth hormone-releasing hormone has trophic activity in the human pituitary (43) and, in addition to a case report of diffuse somatotroph hyperplasia in a patient with a growth hormone-releasing hormone-producing bronchial carcinoid (44), there are several cases of true somatotroph adenoma formation in patients with growth hormone-releasing hormone-producing hypothalamic gangliocytomas (45,46). Corticotrophin-releasing hormone is trophic to corticotrophs(47,48), and a corticotroph adenoma developing in association with a corticotropin-releasing hormone-containing sellar gangliocytoma (49)
suggests, as do the observations in growth hormone-releasing hormone-producing transgenic mice (50), that duration, pattern, and local level of exposure to hypothalamic hormone may be critical in defining the stability of the trophic response induced.
For abnormal pituitary trophic responses in the presence of normal levels of hypothalamic hormones, no constitutionally active corticotrophin-releasing hormone receptor mutations in corticotroph adenomas (51) or growth hormone-releasing hormone receptor mutations in sporadic (52) or familial isolated acromegaly (53) have been identified. There is no evidence of somatostatin receptor (SSTR2 or SSTR5) mutations even in the presence of somatostatin analog resistance (54,55), and only isolated cases of Cushing's disease preceded by generalized glucocorticoid resistance resulting from a dominant-negative glucocorticoid receptor mutation (56) and of Nelson's syndrome, in which a glucocorticoid receptor frame shift mutation may have modified glucocorticoid sensitivity (57), have been reported.
In practical terms, the rarity with which ectopic hypothalamic hormone secretion has resulted in pituitary adenoma induction is good evidence against a major primary hypothalamic etiology in the majority of pituitary adenomas. When an association is present, it is far from clear how the shift from hyperplasia to stimulus-independent growth in the pituitary is mediated at the molecular level, even though such transitions are familiar in other endocrine tissues. This in no way excludes an important influence of the hypothalamus in tumor propagation.
ENDOGENOUS TROPHIC INFLUENCES Activation of the gsp Oncogene
The strongest pathogenic mechanism yet defined for pituitary adenomas is the association between activation of the gsp oncogene and the development of somatotroph adenomas (58-62). The GSa subunit of the heterotrimeric GTPase binds guanosine triphosphate (GTP) when growth hormone-releasing hormone receptors are occupied by ligand. Bound to GTP, the a subunit dissociates from the p and a subunits and associates with adenylyl cyclase, stimulating production of cAMP (cyclic adenosine 3', 5'-monophosphate) until the a subunit-bound GTP is hydrolysis to GDP. At this point, cAMP production stops and the a subunit reassociates with the p and y subunits attached to the receptor.
One of two single point mutations in the Gs a subunit of GTPase has been identified in approx 40% of somatotrophinomas and results in constitutive activation of adenylyl cyclase by reducing the susceptibility of a subunit-bound GTP to hydrolysis. In tumors in which one of these mutations is present, sella morphology is more often normal than in somatotroph adenomas arising from other causes, suggesting that Gs mutant-containing adenomas tend to be smaller (61,63), but there is little consistent difference in growth hormone (GH) levels in gsp-positive tumors compared to gsp-negative tumors (61,63,64). Similarly, the patient's age and gender, and the duration and clinical features of the disease are also indistinguishable between groups with and without the gsp mutation. Although intuitively the association between the gsp mutation and induction of somatotroph adenomas is sensible, it is worth noting that activating mutations of gsp have also been reported, albeit at lower frequency, in endocrinologically inactive adenomas where the immediate consequences of the metabolic association is more obscure (65,66).
Multiple endocrine neoplasia type 1 (MEN-1) is a dominantly inherited condition characterized by the association of pancreatic (75%), parathyroid (80%) and pituitary (65%) adenomas. The gene responsible on chromosome 11q13 codes for a 610-amino acid predominantly nuclear protein named menin, the structure of which has no apparent similarities to any previously known proteins (67) and the functions of which are still being defined (68-71). Whether loss of heterozygosity at the menin locus or mutations in the menin coding region are responsible for sporadic pituitary adenomas has been addressed in several reports. In 14 studies examining sporadic pituitary adenomas, loss of heterozygosity on 11q13 was identified in 57 out of 415 cases (13.7%), with results varying from 0% (72-74) to 28% (75). Nine studies found inactivating mutations in the coding region of the menin gene to be relatively rare: 5 out of 368 cases (1.4%) ranging from 0% (72,76-78) to 5% (79). Menin mutations have also been examined in 15 cases of isolated familial acromegaly, none of which was found to harbor a mutation (53,80,81). So far there is an association between loss of heterozygosity at this locus and pituitary tumor formation, but until the portfolio of functions of menin are defined and the specificity of the finding clarified, the location of menin deletion in sporadic pituitary adenomas remains unclear.
Successful cell cycle phase transition, such as the progression from G1 to S phase, depends on timed degradation of cyclin-dependent kinases. One of the key proteins in controlling timed degradation is securin, which is expressed in proliferating cells and in several kinds of tumors in a cell cycle-dependent manner (82)—increasing throughout S phase from low levels at the end of G1 to maximum prevalence at the junction of G2 and M phase (83). The same protein (securin) was subsequently isolated by comparing differences in transcriptional activity between normal and neoplastic rat pituitary and called pituitary tumor-transforming gene (PTTG) (3). More than 50% increases in PTTG have been observed in 23 of 30 endocrinologically inactive pituitary tumors, 13 of 13 GH-producing tumors, 9 of 10 prolactinomas, and 1 adrenocorticotropic hormone (ACTH)-secreting tumor, with more than 10-fold increases evident in some tumors (84). Although this may imply a role for PTTG in pituitary tumorigenesis, the findings might also be expected of any marker of cell division measured semiquantitatively in comparison to a control tissue containing low levels (i.e., normal pituitary). Its place in pituitary tumor formation is still being assessed.
As discussed, an oncogene implicated in the pathogenesis of pituitary adenomas would have to allow for these lesions to remain trophically stable for decades or resolve of their own accord. It is not too surprising, therefore, that there is little evidence for classic oncogene activation, except in the less than 1% of pituitary adenomas that also adopt the aggressive behavior typical of malignant tumors seen in other organ systems (85-90). This is also true for the tumor suppressor genes p53, Rb, and DCC, which are strongly associated with human tumor formation (91-96). The p53 gene codes for a protein that under normal circumstances prevents replication until the integrity of the genome has been assured (97,98), or, failing that, activates apoptosis (99). Inactivation of p53 is the most common genetic alteration found in human tumors, yet the 600-bp region within which 98% of all substitution mutations occur in other tumors shows no evidence of mutation in pituitary adenomas (100). Despite the association of pituitary tumors with p27 and retinoblastoma gene inactivation in mice, no consistent point mutations of p21 and p27 have been found in human pituitary tumors (101,102).
Abnormalities of tumor suppressor's on the short arm of chromosome 9 (p16INK4A and p15INK4B genes encoding the cyclin inhibitors p15 and p16), N-ras, H-ras, K-ras, mycL1, N-myc, myc, bcl1, H-stf1, sea, kraS2, p53, and fos are rarely or at least inconsistently implicated (86,100,103-114). In the absence of known proto-oncogene mutations, perplexed investigators have adduced hypermethylation, dose effects (112,115), and co-induction of tumor suppressors, such as nm23 (116) to explain pituitary adenoma behavior.
Was this article helpful?