Somatic GNAS mutation causes widespread and diffuse pituitary disease in acromegalic patients with McCune-Albright syndrome.

CONTEXT: McCune-Albright syndrome (MAS) is caused by sporadic mutations of the GNAS. Patients exhibit features of acromegaly. In most patients, GH-secreting pituitary adenomas have been held responsible for this presentation. However, surgical adenomectomy rarely eliminates excess GH production.
OBJECTIVE: The aim of this study was to elucidate pituitary pathology in patients with MAS and to explain the basis of failure of adenomectomy to eliminate GH hypersecretion. DESIGN AND
SETTING: We conducted a case series at the National Institutes of Health. INTERVENTION(S): Interventions included medical therapy and transsphenoidal surgery. PATIENTS AND MAIN OUTCOME MEASURES: We studied clinical and imaging features and the histology and molecular features of the pituitary of four acromegalic MAS patients.
RESULTS: We identified widespread and diffuse pituitary gland disease. The primary pathological changes were characterized by hyperplastic and neoplastic change, associated with overrepresentation of somatotroph cells in structurally intact tissue areas. Genetic analysis of multiple microdissected samples of any type of histological area consistently revealed identical GNAS mutations in individual patients. The only patient with remission after surgery received complete hypophysectomy in addition to removal of multiple GH-secreting tumors.
CONCLUSIONS: These findings indicate developmental effects of GNAS mutation on the entire anterior pituitary gland. The pituitary of individual cases contains a spectrum of changes with regions of normal appearing gland, hyperplasia, and areas of fully developed adenoma formation, as well as transitional stages between these entities. The primary change underlying acromegaly in MAS patients is somatotroph hyperplasia involving the entire pituitary gland, with or without development of somatotroph adenoma. Thus, successful clinical management, whether it is medical, surgical, or via irradiation, must target the entire pituitary, not just the adenomas evident on imaging.

Patients with McCune-Albright syndrome (MAS) develop endocrine abnormalities including GH excess and a clinical presentation of gigantism and/or acromegaly (1–5). The pathophysiology of the GH excess in MAS at the cellular and organ level is not clearly understood. Clinical observations suggest that pituitary dysfunction can be observed independent of adenoma formation because only 33–65% of patients with the MAS and acromegaly exhibit imaging evidence of a pituitary tumor, which is substantially less frequent than patients suffering from sporadic acromegaly (99%). Furthermore, selective adenomectomy does not appear to cure hormonal imbalance in these patients (4, 6, 7). Failure to understand the pathophysiology of GH excess in patients with MAS has precluded the development of adequate therapeutic strategies and limited understanding of its basic biological principles.

MAS was first described in 1937 (8, 9) and comprises polyostotic fibrous dysplasia as well as café-au-lait skin pigmentation and a variety of endocrine abnormalities as its major manifestations. Based on clinical observations, Happle (24) predicted genetic effects on embryonic tissue development as an explanation for the scattered asymmetric distribution of bone lesions and the variability of endocrinopathic features. Weinstein et al. (10) detected activating mutations of the GNAS gene coding for the α-subunit of the stimulatory G protein as the primary genetic alteration present in a mosaic population of susceptible cells. Mutational substitutions occur at the Arg position (R201), most commonly with cysteine (R201C) or histidine (R201H) substitutions, although changes in glutamine (227) have also been described. In individual patients, identical mutations have been consistently detected in different pathologically involved endocrine organs and/or bone and skin lesions, and the mosaic pattern of phenotypic manifestations in MAS has been postulated to be a result of postzygotic somatic mutation of GNAS during early embryogenesis, specifically at the inner cell mass stage (10, 11). Furthermore, Weinstein et al. (10) detected the characteristic R201C and R201H mutations in four somatotroph MAS-associated pituitary adenomas, linking the pituitary tumor to the underlying syndrome. Although no cure for acromegaly has been reported in MAS patients after adenomectomy, only limited attention has been paid to possible pathological changes in the nontumorous pituitary gland in MAS. Only two case reports are known to us that report nonneoplastic changes in MAS-associated pituitary tissue (12, 13). To obtain insight into the pituitary basis of acromegaly associated with MAS, we used clinical studies and analysis of anterior pituitary gland and adenomas removed from three patients (obtained after hemi- or panhypophysectomy) and pituitary gland from one patient obtained at autopsy.

Patients and Methods

Patients

The patients were studied as part of an Institutional Review Board-approved protocol (98-D-0145) at the National Institutes of Health. Endocrine evaluation included, among others, basal measurements of plasma GH, prolactin (PRL), GHRH, and IGF-I (somatomedin-C) and serial measurement of plasma GH during a standard glucose tolerance test (three patients), as previously described (14). Standard commercially available assays were used to measure hormone levels.

The skull and sella turcica were assessed with computed tomography (CT) and magnetic resonance imaging (MRI). Medical therapy, which was not successful in controlling excess IGF-I levels in these patients, consisted of cabergoline, octreotide, and pegvisomant in various combinations at various points during treatment. None of the patients was receiving medical therapy at the time of surgery. Surgery (three patients) was via a sublabial, transnasal, transsphenoidal approach to the sella using intraoperative navigation and removal of a channel of bone from the anterior portion of the nasal cavity to the sella using a drill. This permitted exploration of the sella with selective excision of one or more adenomas and removal of a portion of the abnormal-appearing anterior lobe (two patients) or total hypophysectomy (one patient) when the entire gland appeared abnormal at surgery.

Microscopic evaluation and immunohistochemistry

Serial sections were taken from paraffin-embedded tissue blocks for histological and immunohistochemical examinations. The morphologies of the spectrum of pathological changes were photodocumented and analyzed through use of hematoxylin and eosin- (H&E) and reticulin-stained sections. Immunohistochemistry was performed after antigen retrieval according to a modified protocol that we have previously published (15). Primary antibodies included anti-PRL and anti-GH (Dako, Carpinteria, CA). The presence and intensity of antibody expression were examined in conjunction with serially sectioned H&E sections and reticulin preparations.

Microdissection and mutation analysis

Five-micrometer tissue sections from formalin-fixed, paraffin-embedded tissue blocks were used for microdissection and mutation analysis. Sections for microdissection were directly consecutive to those sections investigated by H&E, reticulin staining, and immunohistochemistry. Microdissection was performed under direct light microscopic visualization using a 30-gauge needle as previously described (16). Areas of neoplasia, hyperplasia, and normal-appearing pituitary tissue were identified on the corresponding reticulin-stained section and selectively microdissected.

A PCR and subsequent enzymatic digest-based technique for mutation detection was used (17). The sense primer, 5′-TTGTTTCAGGACCTGCTTCGCAGC-3′, was designed with a mismatch 3 bases from the 3′ end that, when incorporated into the product, creates a restriction site for PvuII only if a mutation encoding R201C is present. This primer also permits recognition of an NlaIII site created by the mutation encoding R201H. The antisense primer was 5′-AGGTAACAGTTGGCTTACTGGAAG-3′ (GenBank accession no. M21142.1; sense, bases 418–441; antisense, bases 518–495), and amplification resulted in a PCR product of 101 bp. After amplification, the PCR product was incubated with NlaIII or PvuII (New England Biolabs, Inc., Beverly, MA) at 37 C. The full digest volume was loaded on a 2% Tris-borate EDTA agarose gel prestained with ethidium bromide, and bands were visualized by UV transillumination and photodocumented.

Electron microscopy

The tissues were submitted in 2.5% glutaraldehyde (4 C) and postfixed in 0.5% phosphate-buffered osmium tetroxide. The specimens were dehydrated in ethanol series and propylene oxide. Dehydrated specimens were embedded in Spurr's epoxy resin. Semi-thin sections were stained with toluidine blue. Ultra-thin sections were double-stained with uranyl acetate and lead citrate and examined with the electron microscope (Philips CM 10 TEM).

For selective and targeted electron and immunoelectron microscopic analysis of different types of pathology (e.g. areas of hyperplasia vs. areas of neoplasia), pathological areas were identified and characterized using H&E, reticulin, and immunohistochemical stains. Subsequently, corresponding areas were excised from the paraffin blocks and separately processed for electron and immunoelectron microscopy, as described previously (15).

Results

Clinical presentation/surgery

Patients 1–3 were diagnosed with MAS on clinical grounds on the basis of typical café-au-lait spots, polyostotic fibrous dysplasia, with or without additional hyperfunctioning endocrinopathies (Table 1). GH excess was suspected on the basis of either rapid growth in the absence of precocious puberty (patient 3) or the attainment of normal or greater than parental predicted height in the setting of precocious puberty (which should result in short stature) and massive expansion of the craniofacial bones, which occurs in association with GH excess in MAS (14, 18, 19). GH excess was diagnosed on the basis of an elevated random GH, nonsuppressible GH on a standard oral glucose tolerance test, and elevated IGF-I. Pituitary MRI and CT demonstrated single or multiple foci of microadenoma and macroadenoma (Figs. 1 and 2 and Table 1). At surgery the appearance of the pituitary was abnormal, with multiple foci of soft, white and gray-white, poorly demarcated areas of scattered abnormality within the anterior lobe and multiple foci of marginally encapsulated adenomas (patients 2 and 3) or a single larger tumor arising within an abnormal-appearing anterior lobe (patient 1). The only patient in whom the GH and IGF-I levels returned to normal or less after surgery was the patient with complete removal of his tumors and complete hypophysectomy. Patient 4 presented after death in response to his final wishes to have his body studied. In addition to a longstanding history of MAS with fibrous dysplasia, gonadotropin-independent precocious puberty, hyperthyroidism treated with radiation, café-au-lait macules, and typical gonad involvement on autopsy, this patient had signs of acromegaly including distal tufting of the digits, and signs of GH excess superimposed on craniofacial bones involved with fibrous dysplasia with massive, relatively symmetric expansion of frontal and gnathic bones, and a pituitary adenoma at autopsy. He died as a result of pulmonary compromise secondary to advanced scoliosis.

Table 1.

Clinical characteristics of acromegalic MAS patients

Patient no.	Gender, age (yr)	MAS features	Other	OGTT, GH at 30 min (<1 ng/ml)a	Pituitary MRI/CT	Surgical observations, procedure	GH (pre-/post-op) (0–1.0 ng/ml)a	IGF-I (pre-/post-op) (114–492 ng/ml)a	PRL (pre-/post-op) (3.5–31 ng/ml)a
1	M, 19	PFD, CAL	None	10	Macroadenoma	Selective removal of poorly defined macroadenoma; abnormal gland	17/1.0–2.0	3.4/(normal then was 0.34–1.9 U/ml)	26/–
2	F, 29	PFD, PP, CAL	Breast cancer	40.5	Multiple microadenomas	Multiple microadenomas, selective adenomectomy of three poorly defined microadenomas; biopsy of abnormal anterior lobe	21.4/6.1	517/275	26/4.1
3	M, 19	PFD, PP, CAL	Hypogonadism, hypothyroid, unilateral blindness/deafness	127	Macroadenoma with suprasellar and retroclival extension	Multiple poorly defined microadenomas, large macroadenoma; total hypophysectomy	139/0.3	1140/39	17/<1
4	M, 41	PFD, PP, CAL, HT, HP	Fatal scoliosis, autopsy	NA	NA	NA	ND	NA	NA

M, Male; F, female; OGTT, oral glucose tolerance test; PFD, polyostotic fibrous dysplasia; PP, precocious puberty; CAL, café-au-lait; HT, hyperthyroidism; HP, hypophosphatemia; ND, not done; NA, not available.

Normal ranges.

Fig. 1.

Patient 2, a 29-yr-old woman had acromegaly and extensive fibrous dysplasia of the frontal bones and the bones of the skull base with heterogeneous enhancement of the dysplastic bone (A, midline sagittal views; B, serial coronal views) on cranial and pituitary T1 MRI (A, left, midline sagittal view without enhancement; right, afterGd-DTPA administration). A and B, Dysplastic bone fills the sphenoid sinus region (*) and the entire anterior skull base (**); the arrows indicate the pituitary, and the arrowhead in right image indicates a nonenhancing microadenoma. B, Coronal views of pituitary MRI after contrast administration shows at least three nonenhancing microadenomas (arrows) within the anterior lobe on serial slices. C, Graph demonstrates the course of her GH and IGF-I levels over 3 yr. At surgery (arrow), three poorly defined microadenomas were removed with a portion of her surrounding anterior lobe, resulting in a transient drop in the levels of both hormones. Although IGF-I dropped into the normal range, the GH levels did not reach normal levels. After the presence of mammosomatotroph hyperplasia was evident, she was offered additional surgery for removal of the remainder of the gland, while the recently established surgical access through the dysplastic bone of the skull base was available, but she declined. Within 6 months, IGF-I was elevated again, coinciding with a rise in random GH levels. Follow-up MRI have been negative for additional tumor formation.

Fig. 2.

A, Contrast-enhanced sagittal MRI of patient 3, a 19-yr-old male with extensive, heterogeneously enhancing fibrous dysplasia of the bones of the skull base (left panel, *), hypogonadism, hypothyroidism, and acromegaly associated with a multinodular pituitary macroadenoma occupying his sella turcica and extending intracranially into the interpeduncular cistern and behind the clivus (arrow in center panel). The sella was surgically approached in stages; the first stage provided a midline transsphenoidal route to the sella (right panel). At the second stage, performed several days later, the approach was enlarged along the distal ventral edge, permitting exposure of a pituitary that contained multiple microadenomas and a moderately large macroadenoma. The tumors and the patient's pituitary were completely removed. B, Graph demonstrating the course of the patient's GH and IGF-I levels over 2 yr. He responded incompletely to octreotide (30 mg monthly) during the course of serial endocrine assessment (the black closed circles are values obtained while off medical therapy). Removal of his tumors and complete hypophysectomy (arrow) produced a prompt drop of his GH and IGF-I to 0.3 and 39 ng/ml, respectively. Regrettably, he had an intracranial hemorrhage 10 d after surgery from which he did not survive.

Anterior pituitary gland pathology

In all four cases, anterior pituitary gland tissue revealed abundant and widespread structural and cytological changes. Masson trichrome and reticulin stains consistently revealed mild diffuse fibrosis in nontumorous anterior pituitary gland tissue. H&E and reticulin stains revealed areas of regular anterior gland architecture (Fig. 3, A and B), areas of hyperplasia (characterized by marked distension of the reticulin network) (Fig. 3C), and areas of neoplasia (focal mass with breakdown of the reticulin network) in three cases (Fig. 3, D–F). Further characterization of these areas by anti-GH and anti-PRL immunohistochemistry allowed for the identification and characterization of six distinct phenotypic patterns in MAS pituitary gland: 1) areas of regular anterior gland architecture including scattered somatotroph and lactotroph cells (Fig. 3A); 2) areas of regular architecture, but exclusively or nearly exclusively composed of cells with somatotroph differentiation (Fig. 3B); 3) areas of acidophilic hyperplasia, nearly exclusively composed of cells with somatotroph differentiation (Fig. 3C); 4) areas of neoplasia, predominantly composed of tumor cells with somatotroph differentiation (Fig. 3D); 5) areas of neoplasia, predominantly composed of tumor cells with lactotroph differentiation (Fig. 3E); and 6) areas of neoplasia, composed of mammosomatotroph cells, the immunohistochemical profile of which has been defined as “positivity for GH and immunoreactivity for PRL in the same areas on consecutive sections” (12) (Fig. 3F).

Fig. 3.

Structural and cytological changes in anterior pituitary gland tissue in patients with MAS; representative examples for different types of changes are illustrated by H&E stain, reticulin stain, and immunohistochemistry for GH and PRL, which were performed on adjacent sections. A, Mildly fibrotic, but otherwise regular anterior gland architecture with scattered somatotroph and lactotroph cells. B, Mildly fibrotic, but otherwise regular anterior gland architecture with predominantly somatotroph and only a few scattered lactotroph cells. C, Acidophilic hyperplasia, almost exclusively composed of somatotroph cells. D, Neoplasia with dominant somatotroph differentiation. E, Neoplasia with dominant lactrotroph differentiation. F, Neoplasia with predominantly mammosomatotroph differentiation.

Well-characterized areas of regular pituitary gland, areas of hyperplasia, and areas of neoplasia were separately excised from paraffin blocks after analysis of histological and immunohistological stains on adjacent sections and subjected to ultrastructural studies and immunoelectron microscopy (Fig. 4).

Fig. 4.

A, In mammosomatotroph adenomas, mammosomatotroph differentiation was confirmed in tumor cells that demonstrated well-developed rough endoplasmic reticulum and characteristic granule extrusions (arrows). B and C, Double-immunolabeling of thin sections revealed positive signal for GH (large signal) in neuroendocrine granules of cells with somatotroph differentiation and positive granular signal for PRL (small signal) in cells with lactotroph differentiation. B, Mammosomatotroph cells, in contrast, revealed granules with immunoreactivity for both GH and PRL. C, Similarly, double-labeled mammosomatotroph cells were also detected in areas of acidophilic hyperplasia. The inset shows at higher magnification that positive signal for both GH and PRL are detected in neuroendocrine granules within the same cell. D, Rare scattered mammosomatotroph cells were detectable in regular anterior pituitary gland. A mammosomatotroph cell (right) is adjacent to a cell with exclusive immunoreactivity with anti-PRL (left).

Mutation analysis

From consecutive sections of histologically fully characterized sections, 34 foci were microdissected from three cases, including from hyperplastic and neoplastic areas described above as well as areas of normal-appearing pituitary gland tissue (Fig. 5A).

Fig. 5.

A, Tissue microdissection; representative examples of microdissected normal tissue, hyperplastic tissue, and neoplastic tissue are shown. Before genetic analysis, different areas of normal, hyperplastic, and neoplastic anterior pituitary gland tissue were microdissected from H&E-stained slides; reticulin stains and immunohistochemical preparations from adjacent serial sections were used for identification of areas of interest. B, GNAS mutation analysis of pituitary gland tissue from three cases with MAS. DNA was extracted from microdissected tissue samples, amplified by PCR, and separately digested with NlaIII and PvuII, which can digest the PCR product in case the R201H mutation or the R201C of the GNAS gene is present, respectively. Case 1, Eleven foci of pituitary tissue were microdissected, mutation bands (arrow) are detected in different areas of hyperplasia and neoplasia, but not in three foci of normal-appearing pituitary gland (digestion with NlaIII). Case 2, Fifteen foci of pituitary tissue were microdissected; mutation bands (arrow) are detected in different areas of hyperplasia and neoplasia, but not in five different foci of normal-appearing pituitary gland (digestion with NlaIII). Case 3, Nine foci of pituitary gland tissue were microdissected; mutation bands (arrow) are detected in different areas of pituitary gland tissue with overrepresentation of somatotroph cells and areas of somatotroph hyperplasia (digestion with PvuII).

All types of hyperplastic and neoplastic pathological changes as well as regularly structured areas with excessive numbers of somatotrophs showed evidence of an additional “mutation band” after restriction enzyme digestion (Fig. 5B). Evidence of somatic GNAS mutation was not detectable in most normal pituitary control specimens taken from the same patient tissue (Fig. 5B). In a few samples, histologically normal pituitary tissue of the patients revealed a weak mutation band (data not shown).

Genetic analysis of multiple microdissected samples of any type of pathological area consistently revealed identical GNAS mutations, i.e. either R201H or R201C, in individual patients (Fig. 5B).

Discussion

The morbidities of vision and hearing loss associated with fibrous dysplasia of the craniofacial bones in MAS appears to be more common and more severe in patients with GH excess (14, 19, 20). This highlights the importance of early detection and successful treatment of the pituitary endocrinopathy. Current pharmacological therapy with dopaminergic and somatostatin analogs and GH receptor blockade produces normal serum PRL and IGF-I levels in many patients, but not all (14, 21). In patients who fail medical therapy, selective excision of a pituitary adenoma(s) rarely controls excess GH secretion.

In 1984 Kovacs et al. (12) described mammosomatotroph hyperplasia of the pituitary in a patient with MAS associated with acromegaly. However, in that case they only had access to the lateral wing of the pars distalis because the pituitary tissue was removed during surgery that was primarily being performed for bone reconstructive surgery, during which “normal” pituitary tissue was also removed. The report describes hyperplasia alone; the patient had no adenoma in the tissue examined. Despite that report, now 27 yr old, the only other report describing pituitary hyperplasia in MAS associated with acromegaly was a recent case report of a patient with MAS, acromegaly, and pituitary hyperplasia. In that case, Madsen et al. (13) describe a single focus of hyperplasia adjacent to a mammosomatotroph adenoma (the gland in that sample was included as part of the removal of a recurrent mammosomatotroph adenoma); this was the only previously reported case of histological confirmation of an adenoma occurring in a pituitary that also contained hyperplasia in MAS and acromegaly. The other cases of MAS and acromegaly in which tissue has been analyzed are described as containing adenomas only, usually mammosomatotroph adenomas or somatotroph adenomas; this can be explained, at least in part, because the gland was not included as part of the surgical specimen, since the goal of surgery was selective removal of an adenoma. In the current study, the pituitary gland was included as part of a systematic study; the pituitary gland in its entirety was studied histologically, demonstrating heterogeneous involvement of the entire gland with mammosomatotroph hyperplasia; and multiple adenomas within a single patient were identified clinically and histologically.

Activating mutations in GNAS codon 201 or codon 227 maintain Gsα in a constitutively active stage. Activating mutations of GNAS have been implicated in the formation of a subset of somatotroph adenomas. In patients with MAS, R201C or R201H activating mutation of GNAS can be found in different endocrine and nonendocrine organs that are affected by pathological proliferation of cells (10, 11).

It was noted in the original publication that linked MAS with activating mutations of GNAS that different tissues from a single patient with MAS always have an identical specific activating mutation. Because the activating mutation is a somatic event, Weinstein et al. (10) explained the concordance of mutation types with an activating mutation occurring in a multipotent cell with the capacity of differentiation into all three germ layers (10). Subsequently, MAS pathology in fibrous dysplasia has been well-characterized as an effect of developmental arrest.

Although Weinstein et al. (10) associated activating mutation of GNAS with lesion formation in patients with MAS, the effect of an activating GNAS mutation on the pituitary gland or other endocrine organs remains poorly understood. We demonstrate here in pituitary gland tissue of patients with MAS that the effects of GNAS mutation are ubiquitous and widespread and not restricted to adenoma formation. To the contrary, the pituitary gland tissue revealed an abundance of independent, nonneoplastic extensions of acini that were predominantly or exclusively populated by somatotroph or mammosomatotroph, occasionally also lactotroph, cells (somatotroph/mammosomatotroph hyperplasia). Because the occurrence of hyperplastic foci was far more frequent than the occurrence of frank adenoma, the primary effect of GNAS mutation in MAS is nonneoplastic intraacinar cellular proliferation. Formation of frank adenoma may or may not occur. Of the four cases of widespread pituitary hyperplasia included in this study, one had developed four adenomas, two showed single adenomas, and one was free of tumor.

Most hyperplastic foci analyzed in this study were predominantly composed of somatotroph cells. Ultrastructurally, we also detected mammosomatotroph cells in these areas, as previously demonstrated in the case report of Kovacs et al. (12); in addition, immuno-ultrastructural analysis of our cases revealed mammosomatotroph cells in randomly sampled areas with regular architecture, in contrast to the normal adenohypophysis in which mammosomatotroph cells are uncommon (12).

In each case, the individually sampled hyperplastic foci revealed the same type of GNAS mutation. This concordance of genetic findings in individual foci of nonneoplastic mammosomatotroph proliferation supports Weinstein's developmental concept for the pathogenesis of MAS-associated lesions. Although mutations in genes that encode G proteins can alter pituitary-cell division and hormone production (22) and interruption of cAMP signaling precludes somatotroph development in cAMP-responsive element-binding protein mutant mice (23), it is not known whether an activating GNAS mutation itself is responsible for the developmental arrest of the mammosomatotroph cell. In one case, we investigated and detected GNAS mutation in areas with predominant somatotroph differentiation, suggesting that differentiation into a somatotroph cell can occur in cells with an activating GNAS mutation.

Our findings demonstrate that one of the primary effects of a GNAS activating mutation occurring during early development of the pituitary gland is nonneoplastic intraacinar proliferation of mammosomatotroph/somatotroph cells. It remains to be clarified whether this process merely increases the pool of “eligible” cells for random progression into adenoma or whether additional genetic or environmental molecular events are required to transform a cell with an activating GNAS mutation into a tumor cell.

Finally, these observations also have clinical significance. The pituitary pathology is widespread and nontumorous pathological change diffusely involving the entire gland is often present, although not evident using modern imaging. Thus, resection of what initially appears to be a single adenoma has been consistently met with failure of normalization of GH and/or IGF-I levels in previously reported patients and in the current series (review of the published cases in which surgery was performed for MAS and acromegaly indicates that there are no reported cases in which the patient has been cured by transsphenoidal surgery alone).

Therefore, it is important to treat the entire pituitary gland. Attempts to control excess GH and IGF-I with medical therapy with a somatostatin analog or GH-receptor antagonist and dopaminergic therapy to control excess PRL are generally used before surgery. However, in some patients these are not successful, and surgical treatment may be required. However, surgical elimination of excess GH production requires total hypophysectomy, which occurred in patient 3. Thus, the diffuse and varied pituitary pathology must be considered for successful management of MAS patients with GH excess.