Potential markers of disease behavior in acromegaly and gigantism

Abstract

Introduction:

Acromegaly and gigantism entail increased morbidity and mortality if left untreated, due to the systemic effects of chronic GH and IGF-1 excess. Guidelines for the diagnosis and treatment of patients with GH excess are well established; however, the presentation, clinical behavior and response to treatment greatly vary among patients. Numerous markers of disease behavior are routinely used in medical practice, but additional biomarkers have been recently identified as a result of basic and clinical research studies.

Areas covered:

This review focuses on genetic, molecular and genomic features of patients with GH excess that have recently been linked to disease progression and response to treatment. A PubMed search was conducted to identify markers of disease behavior in acromegaly and gigantism. Markers already considered in clinical care guidelines were excluded. Literature search was expanded for each marker identified. Novel markers not included or only partially covered in previously published reviews on the subject were prioritized.

Expert opinion:

Recognizing the most relevant markers of disease behavior may help the medical team tailoring the strategies for approaching each case of acromegaly and gigantism. This customized plan should make the evaluation, treatment and follow up process more efficient, greatly improving the patients’ outcomes.

1. Introduction

Acromegaly is a clinical entity characterized by manifestations affecting multiple organs due to the systemic effects of GH and IGF-1 excess. In the vast majority of cases, acromegaly is caused by a benign GH-secreting type of pituitary neuroendocrine tumor (PitNET), termed somatotropinoma, although it can infrequently be due to GH or GHRH-secreting ectopic tumors, or somatotroph hyperplasia of the anterior pituitary [1]. Acromegaly has traditionally been considered as a rare disease: a recent metanalysis reported a prevalence of acromegaly of 2.8–13.7 cases per 100,000 inhabitants, with an annual incidence of 0.2–1.1 cases per 100,000 inhabitants [2]. The variability in disease incidence and prevalence among studies could be explained by the changes in diagnostic criteria, the improved sensitivity of biochemical tests and an increased disease awareness over the years [3]. Most of the cases are diagnosed during adulthood (the median age at diagnosis is in the fifth decade of life) and a delay between disease onset and diagnosis, currently estimated to be of 4.5–5 years, is considered typical, due to the indolent disease progression [2].

Less frequently, GH excess starts in childhood or adolescence, before the closure of growth plates, and therefore the disease manifests as gigantism. Although acromegaly and gigantism appear to be clinical variants of a single entity, specific genetic alterations might determine the development of one or another presentation. Both acromegaly and gigantism entail a potential for high morbidity due to the consequences of the chronic exposure to GH overproduction and, therefore, IGF-1 excess. Indeed, chronic IGF-1 excess is associated with increased mortality rate, mainly due to cardiovascular complications, while biochemical control can normalize the mortality rate to that of the general population [4].

Clinical guidelines for the diagnosis and treatment of GH excess have been established and are frequently updated [5]. Nevertheless, the medical teams taking care of patients with acromegaly and gigantism must be aware that disease presentation and behavior greatly vary among patients. Specific genetic and molecular features might account for differences in the clinical phenotype and on the response to treatment, although many of such disease biomarkers are not yet fully validated and, thus are not taken into consideration in the guidelines. Familiarizing with markers of disease behavior encountered at different stages of the patients’ caretaking may be helpful for designing the best strategy to approach the disease in each case, especially in patients with an atypical presentation or suboptimal response to standard therapeutic strategies [6]. This way, the medical team will be able to determine what to expect in terms of disease aggressiveness and complications, to decide whether to establish a more aggressive or expectant therapeutic approach, as well as to perform additional studies in each patient.

1.1. Literature search

A literature search was done in PubMed using the terms “aggressive pituitary adenoma”, “aggressive pituitary neuroendocrine tumor”, and “molecular markers of acromegaly” to identify markers of disease behavior in acromegaly and gigantism; the literature search was expanded for each marker identified. On this note, it is important to emphasize that the term “pituitary neuroendocrine tumor, PitNET” is not yet fully accepted as a replacement to the term “pituitary adenoma” and is not yet of routine clinical use. This term, however, has been widely used in the scientific literature in the last couple of years, and it has therefore been adopted for this review. Excellent reviews on markers of disease behavior were found in the literature, and an effort was done to avoid repetition of biomarkers that have been previously been extensively reviewed [6–8]. Novel markers not included in previously published reviews or that required an updated review were prioritized, since the aim of this review is to analyze recent genetic and molecular discoveries that could potentially be translated into the clinic as disease biomarkers.

2. Germline genetic defects

Around 5–7% of all PitNETs occur in a familial setting, most of them in families with multiple endocrine neoplasia type 1 (MEN1) or familial isolated pituitary adenoma (FIPA) [9]. The proportion of familial cases is higher among young patients, and up to 20% of the hormone-secreting PitNETs diagnosed in children are due to a germline genetic defect [10,11]. Regarding GH excess cases, it is known that 50% of the patients with gigantism carry known congenital genetic alterations (germline AIP mutations in 30–40% of cases and germline or mosaic GPR101 gene amplification in 8–10%), but there are no data available to accurately estimate the prevalence of cases of acromegaly caused by germline genetic defects [12–14]. While taking the clinical history of the patients, it is necessary to specifically enquire about any personal or family history of pituitary tumors and other endocrine and non-endocrine neoplasms, as well as to thoroughly explore the patient for clinical manifestations that could be indicative of a PitNET syndrome. Most patients with familial forms of acromegaly and gigantism display particular features explained by the underlying genetic defects, which distinguish them from sporadic patients; such characteristics have implications for the diagnostic and therapeutic approaches.

2.1. Familial isolated pituitary adenoma

The most common cause of familial acromegaly and gigantism is FIPA, an autosomal dominant syndrome characterized by the presence of PitNETs in two or more members of the same family in the absence of other clinical features, such as those encountered in the syndromes of multiple endocrine neoplasia [15]. Any type of PitNET can be found in this setting, but three quarters of the patients have somatotropinomas or prolactinomas, and these tumors are diagnosed earlier than their non-familial counterparts [16–19]. One-fifth of FIPA cases are due to loss-of-function germline variants in the aryl hydrocarbon receptor interacting protein gene (AIP) and a small number of cases are known to be due to other gene defects, but the genetic cause remains unknown in most FIPA patients [19–23].

AIP mutations cause 6–8% of all the sporadic pituitary adenomas and 12% of the macroadenomas diagnosed before the fourth decade of life, one-fifth of the pediatric functional pituitary adenomas and one-third of the cases of gigantism, as well as 3–4% of the sporadic PitNETs in the general population and 8% of the cases of acromegaly resistant to the treatment with somatostatin analogs (SSAs). These patients may present sporadically, because AIP mutation-associated FIPA displays incomplete penetrance (15–33%) [24]. In most patients, the disease arises during the second and third decades of life, and two-thirds of cases are diagnosed at age ≤30 years [25]. Ninety-three percent of the AIP mutation positive patients have macroadenomas, with GH excess in 80% of the cases; and 30–40% of patients develop gigantism [17,25]. Most tumors are somatotropinomas or mammosomatotroph adenomas, almost always of the sparsely granulated type, but some patients develop plurihormonal tumors [25,26]. Patients with AIP mutations have an increased risk of tumor apoplexy, this complication occurring in around 8% of the patients, even with familial presentation in rare cases [25–27]. Patients with GH excess and AIP mutations often display impaired response to the treatment with first-generation SSAs and dopamine agonists (DAs), usually requiring multimodal treatment [17,26,28,29]. It has been shown, however, that some of the patients resistant to first-generation SSAs can respond to treatment with pasireotide [30]. Besides, a less frequent presentation is recognized, consisting of apparently indolent PitNETs detected only by targeted screening test in mutation carriers [25,26,31].

2.2. X-linked acrogigantism

Gene copy gains affecting the of the GPR101 gene, encoding an orphan G protein-coupled receptor, cause a particular type of gigantism with onset in early infancy, known as X-linked acrogigantism (X-LAG) [32]. Height and weight are normal at birth in most patients, but there is an abrupt growth acceleration thereafter, always before the age of 4–5 years. Indeed, X-LAG patients are significantly younger and taller at diagnosis than patients with gigantism due to other causes [12,33]. These patients often develop marked acral growth and facial coarsening with hypertelorism and other manifestations that resemble the features of adult acromegaly patients; some of them also present increased appetite [13,32,34]. Most of the documented cases have presented sporadically, and vertical transmission has been demonstrated in only three families (out of 33 X-LAG cases reported so far) [32,35–37]. Around three-quarters of X-LAG patients are females with germline GPR101 gene amplification. In males, the genetic defect has been found at the germline level in familial cases, but sporadic patients present mosaic gene amplification [13,38]. Most X-LAG patients present with markedly elevated GH and IGF-1 with hyperprolactinemia, while GHRH is normal or only slightly elevated; they usually develop pituitary macroadenomas, but a few patients have pituitary hyperplasia and coexistence of both components has been observed [13,33,34,39]. Most patients do not achieve control when treated with first-generation SSAs or DA and, although surgical resection or radiotherapy were curative in a few patients, extensive resection or radiation was required, at the expense of hypopituitarism [36]. More recently, pegvisomant has shown to effectively control IGF-1 levels and growth velocity (while preserving pituitary function) in patients with extensive pituitary involvement; it has therefore been suggested that this drug should be considered as the first choice of treatment in X-LAG [13,34,40].

2.3. Multiple endocrine neoplasia type 1 and type 4

MEN1 is an autosomal dominant condition characterized by the development of primary hyperparathyroidism, gastroenteropancreatic neuroendocrine tumors and PitNETs, although some patients also develop other endocrine tumors and associated non-endocrine features such as lipomas, facial angiofibromas, collagenomas and ependymomas [41]. Ninety percent of MEN1 cases are due to loss-of-function germline mutations in the MEN1 gene, 85% have familial presentation and 15% present as simplex patients and the penetrance in almost complete by the fifth decade of life [42–45]. Parathyroid tumors are the most constant feature of MEN1, arising in 95% of the patients, and they are usually the first manifestation of the syndrome [44]. Approximately 30–40% of MEN1 patients develop PitNETs: these tumors frequently appear at a young age and are the first disease manifestation in 17–29% of patients [46,47]. Only 9% of the MEN1-associated PitNETs are clinically somatotropinomas, while two-thirds are prolactinomas, but histologically, 10– 39% of tumors secrete more than one hormone [46,48]. PitNETs in MEN1 patients are significantly larger and more invasive than those occurring in non-MEN1 sporadic patients [48]. Suboptimal response of MEN1-related functional PitNETs to medical and/or surgical treatment has been reported, with normalization of hormone secretion achieved in less than half of the patients [46].

About 2% of the MEN1 mutation-negative MEN cases are due to loss-of-function variants in the cyclin dependent kinase inhibitor 1B gene (CDKN1B); this condition is termed as MEN4 [49,50]. The phenotype in MEN4 patients is more variable than in those with MEN1 gene defects: PitNETs, neuroendocrine tumors and various types of benign and malignant neoplasms may arise alongside with primary hyperparathyroidism, which is the most common component of the syndrome, but MEN4 might also present as FIPA [51]. Only seven cases of GH-excess have been described in patients with CDKN1B mutations, out of 14 patients with PitNETs reported in this setting; four cases occurred in two families with a FIPA phenotype [52]. An MEN1-like phenotype has been found to be due to variants in genes encoding other cyclin dependent kinase inhibitors (CDKN1A, CDKN2B and CDKN2C) in a small number of patients, although none of such individuals presented with GH excess [50].

2.4. Three P association

The coexistence of pheochromocytomas/paragangliomas and PitNETs in a single individual or in a single family has been recently termed as “Three P Association” (3PAs). This uncommon phenotype is usually caused by germline loss-of-function mutations in the genes encoding the succinate dehydrogenase subunits and its assembly factor (SDHA, SDHAF2, SDHB, SDHC, SDHD, and SDHAF2, collectively known as SDHx genes), but a few cases caused by alterations in other genes (RET, MAX, MEN1, VHL) have also been reported [21–23,53,54]. Besides their association with 3PAs, germline SDHx mutations have been associated with an apparent FIPA phenotype in a couple of families, but not with sporadic PitNETs [21–23]. Two-thirds of the patients develop prolactinomas and the rest have non-functioning PitNETs or somatotropinomas; therefore, this syndrome is a very rare cause of GH excess [54–56].

2.5. McCune-Albright syndrome

The McCune-Albright (MAS) syndrome is a rare disorder characterized by the association of monostotic or polyostotic fibrous dysplasia, café-au-lait skin spots and one or more manifestations of endocrine hyperfunction (precocious puberty, hyperthyroidism, Cushing’s syndrome and PitNETs, among others) [57]. MAS is caused by mosaic activating mutations in the codon 201 of the GNAS gene, which encodes the protein G stimulatory alpha subunit (Gsα) [58]. One-fifth of MAS patients develop GH excess, often with hyperprolactinemia, but pituitary disease only occurs in patients with an affected maternal allele, due to the GNAS gene imprinting pattern [59]. Most patients develop extensive pituitary involvement: somatotroph hyperplasia is the predominant pattern, somatotroph neoplasia, lactotroph neoplasia and mammosomatotroph neoplasia have also been described [60,61]. Surgery is usually not curative, and MAS patients respond only partially to first-generation SSAs or DAs and might require combined treatment with radiotherapy and pegvisomant, although radiotherapy is contraindicated in patients with craniofacial fibrous dysplasia, due to the possibility of sarcomatous transformation of the bone lesions [59,62].

2.6. Carney complex

Carney complex (CNC) is an infrequent autosomal dominant syndrome of multiple endocrine neoplasia (adrenal, testicular, pituitary and thyroid tumors) and cardiocutaneous manifestations (lentiginosis, myxomas, nevi), with full penetrance [63,64]. Most cases are caused by de novo, loss-of-function mutations or deletions in the PRKAR1A gene; defects in PRKACB and 2p16 have also been described [65–69]. Up to 80% of CNC patients display subclinical GH excess, but clinically evident acromegaly is rare (12% of patients), and two-thirds of CNC patients have co-existent mild hyperprolactinemia [63,70]. Acromegaly in the context of CNC usually has an insidious evolution and is characterized by the development of multiple GH or GH and prolactin-secreting microadenomas surrounded by hyperplastic tissue [71].

2.7. Neurofibromatosis type 1

Neurofibromatosis type 1 (NF1) is a relatively common autosomal dominant disease characterized by the presence of café-au-lait skin spots, skinfold freckling and cutaneous neurofibromas, and is due to mutations (de novo in 42% of patients) affecting the NF1 gene [72]. There are a number of case reports in the literature of GH excess occurring in NF1 patients, usually in the presence of optic gliomas. In a recently reported cohort, 11% of children with neurofibromatosis type 1 and optic gliomas presented GH excess; in all cases the gliomas were located within the optic chiasm [73]. Rarely, a co-existing pituitary mass has been described, which could be coincidental [74,75]. The pathophysiological mechanism explaining the occurrence of GH excess in the setting of NF1 mutations is uncertain [76].

2.8. Recommendations for genetic screening of GH excess patients

When the findings of the anamnesis and clinical examination suggest the possibility of one of the abovementioned conditions, genetic screening should be performed. Particularly in young patients, a negative family history does not preclude the existence of a germline genetic defect, due to the possibility of low disease penetrance or de novo genetic defects. The presence of additional features at the physical examination or a personal or family history of associated neoplasms that would be indicative of a specific genetic syndrome, should be taken into consideration [24]. Gigantism in patients <5 years-old is highly indicative of X-LAG. The main predictors for the presence of an AIP mutation in patients with acromegaly or gigantism are age at onset 0–18 years, positive family history and large tumor size [77]. Unless there is a high level of suspicion for a specific condition, next-generation sequencing-based techniques (custom gene panels or whole-exome sequencing, WES) and/or copy number variation analysis (multiplex ligation-dependent probe amplification or comparative genomic hybridization, CGH) are preferred over single-gene screening. Patients with manifestations of MAS or male patients in whom X-LAG is suspected might require the examination of DNA samples from various tissues to detect mosaicism. The results of genetic screening will serve to determine the need for clinical screening for other syndrome components in the same patient and for genetic testing in other family members.

3. Histopathology and dysregulated gene expression at tumor level

3.1. Granulation pattern

Somatotropinomas originate from PIT-1 positive cells and express mainly GH, but acromegaly and gigantism can also be caused by tumors co-secreting GH and prolactin, including mammosomatotroph adenomas, mixed somatotroph and lactotroph adenomas and plurihormonal adenomas, or by somatotroph or mammosomatotroph hyperplasia [78]. Based on their density of GH-containing secretory granules and their low molecular weight cytokeratin (most frequently represented by CAM5.2) staining patterns, somatotropinomas can be divided into densely granulated and sparsely granulated tumors. Densely granulated somatotropinomas represent the most common subtype; these tumors contain numerous large secretory granules (400–600 nm), most cells are positive for PIT1, alpha subunit and GH by immunohistochemistry, and present a characteristic perinuclear pattern of cytokeratin staining [79,80]. These tumors respond well to surgery and SSAs and are hypointense or isointense on T2 magnetic resonance images (MRI) [78,80]. On the contrary, sparsely granulated adenomas display sparse secretory granules (100–250 nm) and variable positivity for GH. Cytokeratin immunostaining evidences characteristic paranuclear dense aggregates of cytokeratin and endoplasmic reticulum known as fibrous bodies [78,79]. Sparsely granulated somatotropinomas are diagnosed in younger patients, are larger and more invasive, display T2 hyperintensity on MRI and have a poor response to surgery and SSAs, but respond better to pasireotide compared with the densely granulated subtype [78–85]. A third type of somatotropinoma is recognized which has a mixed granulation pattern and is termed “intermediate granulated somatotropinoma”; this subtype behaves clinically as the densely granulated adenomas [80].

Multiple mechanisms might explain the variable clinical presentation and response to treatment among these two morphological subtypes of somatotropinomas. For instance, sparsely granulated tumors display reduced expression of the somatostatin receptor type 2 (SST₂) and CDKN1B at the mRNA and protein levels [86–88]. These tumors also present reduced expression (mRNA and protein) of E-cadherin, a protein implicated in the formation of the adherens junction [88,89]. Indeed, in somatotropinomas, E-cadherin expression levels positively correlate with serum GH and IGF-1 levels and with the response to treatment with SSAs, and negatively with tumor size and invasiveness [89,90]. Loss of E-cadherin expression is a marker of epithelial-mesenchymal transition (EMT), a process by which polarized epithelial cells develop increased migratory capacity, invasiveness, resistance to apoptosis and increased production of extracellular matrix, acquiring a mesenchymal cell phenotype. In somatotropinomas, loss of E-cadherin expression and EMT might be triggered by loss of ESRP1 expression [90].

3.2. Somatostatin receptors

Resistance to the treatment with first-generation SSAs is not uncommon among patients with GH excess [3]. The main subtypes of somatostatin receptors (SSTs) expressed in somatotropinomas are SST₅ and SST₂, which mediate most of the effects of the clinically available SSAs (SST₂ for octreotide and lanreotide and both receptors for pasireotide), but they also express, at lower levels, SST₃, SST₁ and SST₄ [91–93]. Results regarding the expression of SSTs in somatotropinomas vary among studies due to the use of quantitative PCR or immunohistochemistry with polyclonal or monoclonal antibodies. In general, studies using monoclonal anti-SST₂ antibodies (the gold-standard is clone UMB-1, which detects the SST_2A subtype) have associated a high tumoral (predominantly membranous) immunoreactivity with a favorable response to first-generation SSAs, while negative immunohistochemistry is found in unresponsive tumors [81,94–98]. Other studies have found that SST₂ gene (SSTR2) mRNA levels positively correlate with the biochemical and tumoral responses to octreotide, while SST₅ gene (SSTR5) expression levels negatively correlate with the biochemical response; patients with a good response to octreotide have a SSTR2/SSTR5 ratio ≥1.3 [92,93].

In contrast, studies evaluating the response to pasireotide have shown mixed results. This second-generation SSA binds SST_1–3 and SST_5, with preferential affinity for SST₅ over SST₂ [99]. In concordance with this, it has been observed that high SST₅ immunoreactivity might predict a good response to pasireotide, and somatotropinomas with lower SSTR2 mRNA expression and lower SSTR2/SSTR5 ratio displayed better response to pasireotide in vitro [81,100]. On the contrary, other studies suggested that the effects of pasireotide on somatotropinomas are actually driven primarily by SST₂, since the expression of this receptor at the protein and mRNA levels positively correlates with the response to pasireotide, while SST₅ shows no correlation [101,102]. Further studies in larger cohorts are required to determine if the expression of SST₂, SST₅ or both receptors is necessary for the response to pasireotide.

A truncated form of SST₅ with only four transmembrane domains, sst5TMD4, is expressed in most somatotropinomas, and its mRNA levels positively correlate with resistance to the treatment with octreotide and with features of aggressiveness [103,104]. A single SSTR5 mutation has been described in a unique somatotropinoma resistant to octreotide, and loss of one copy of the SSTR5 gene has been detected in about 10% of somatotropinomas resistant to SSAs [105,106]. Therefore, genetic defects in the genes encoding SSTs do not seem to be a common cause of resistance to SSAs.

3.3. Filamin A

Occasionally, somatotropinomas exhibit resistance to the treatment with first-generation SSAs despite high SST₂ expression, indicating that the variable response to these drugs might also be mediated by alterations in postreceptor signaling pathways. Along these lines, the scaffold cytoskeletal protein filamin A (FLNA) is involved in the stabilization and signal transduction of SST₂ in somatotroph cells in vitro [107]. FLNA expression at the mRNA level directly correlates with SST2, SST5 and dopamine 2 receptor gene (DRD2) expression in somatotropinomas, although its expression has not been directly linked with any clinical parameters [108,109].

3.4. Beta-arrestin

Internalization of SSTs is mediated by G protein‐coupled receptor kinase (GRK2)-phosphorylated beta-arrestins, leading to receptor endocytosis [108]. A couple of studies demonstrated that low expression of the beta-arrestin 1 and 2 genes (ARRB1 and ARRB2) at the mRNA level correlated with good response to first-generation SSAs [110,111]. Nevertheless, in a different study, ARRB1 and ARRB2 levels were not associated with SSA response, tumor invasiveness or SSTR2, SSTR5 or DRD2 expression [108]. Although interesting from the physiological point of view, the role of beta-arrestins as an indicator of the therapeutic response to SSAs is yet to be clarified.

3.5. ZAC1

One of the multiple intracellular effects of octreotide includes the upregulation of the tumor suppressor gene ZAC1, which is highly expressed in the normal pituitary [112]. Interestingly, the treatment with SSAs before pituitary surgery increases the expression of ZAC1 in somatotropinomas, and the silencing of this tumor suppressor blocks the antiproliferative effect of octreotide in pituitary cells in vitro [112,113]. ZAC1 immunoreactivity in somatotropinomas positively correlates with the biochemical and tumoral response to the treatment with SSAs [112].

3.6. AIP immunostaining

Aside of the role of AIP mutations in familial GH excess, AIP protein expression is reduced in invasive somatotropinomas and in those with poor response to SSAs, although AIP immunoreactivity is not indicative of AIP mutations [29,81,114–116]. Tumors with low AIP expression display reduced SST₂ immunoreactivity; however, the expression of AIP in somatotropinomas does not depend on the expression of SST₂, indicating that each of these markers might be an independent predictor of the response to SSAs [81,117,118]. Indeed, it has been demonstrated that AIP immunohistochemistry is an excellent marker of tumor invasion and response to first-generation SSAs in somatotropinomas, although it does not predict the response to treatment with pasireotide [81,115–117]. Similar to ZAC1, AIP immunoreactivity increases after treatment with SSAs, and AIP upregulates ZAC1 expression in vitro, but the exact mechanism underlying this effect is unclear [118].

3.7. Aberrant membrane receptors

Aberrant GH secretion in response to stimulation with glucose-dependent insulinotropic polypeptide (GIP) has been described in a subgroup of patients with acromegaly [119,120]. Overexpression of the GIP receptor (GIPR) has been observed in a subset of gsp negative (see next section) somatotropinomas; the overexpressed receptor drives the paradoxical GH increase during the oral glucose tolerance test (OGTT) observed in such patients [120,121].

3.9. Recommendations for histopathological studies in GH excess patients

Although immunohistochemistry for the abovementioned markers is not currently available on a routine basis in most centers, it can be requested via research studies. Further validation of such markers in larger cohorts is warranted, so that they can be standardized and therefore applied in the routine clinical care of patients with GH excess. Immunostaining for CAM5.2, however, is available in most specialized Pathology units, and this staining per se can be an extremely valuable tool as a prognostic marker of disease behavior and response to treatment.

4. Somatic genetic, genomic and epigenomic defects

PitNETs are usually considered monoclonal tumors. According to this concept, the accumulation of multiple somatic genetic mutations and/or epimutations conferring growth advantage to a single somatic pituitary cell leads to clonal expansion and, eventually, to a PitNET [122]. Not surprisingly, alterations in the expression of genes involved in cell cycle and apoptosis regulation, as well as growth factors and their receptors, are common findings in PitNETs [123]. Despite their prominent role in cancer, genetic defects affecting classic tumor suppressor genes or oncogenes, are rarely implicated in the pathogenesis of PitNETs, and are only found in cases of atypical adenomas and pituitary carcinomas [124–126]. Somatic mutations affecting genes involved in inheritable causes of somatotropinomas are not frequently detected either.

4.1. Somatic GNAS activating variants

Somatic activating pathogenic variants of the GNAS gene account for the most common somatic genetic defect in somatotropinomas. Such mutations are always located at the codons 201 (as those causing MAS) or 227, and their functional consequence is the creation of a constitutively active Gsα subunit, commonly known as the gsp oncogene [127,128]. GNAS mutations have been detected in 4.4–59% of sporadic somatotropinomas, according to different studies [123]. The gsp oncogene originates from the maternal allele, which is normally monoallelically expressed in the pituitary due to the tissue-specific genetic imprinting pattern of GNAS [129,130]. Somatotropinomas harboring the gsp oncogene are smaller, respond better to the treatment with first generation SSAs, and are more often of the densely granulated subtype, although these results vary among studies [79,131–133]. A recent metanalysis found that the gsp oncogene is related with a higher response to the acute octreotide suppression test in patients with acromegaly, suggesting that this genetic defect could be used as a prognostic factor for response with SSAs [134]. Besides gsp oncogene, no other recurrent somatic mutations have been identified in somatotropinomas [135–137].

4.2. Genomic and epigenomic alterations in somatotropinomas

A number of recent studies have highlighted the role of genomic and epigenomic alterations in somatotropinomas. Recurrent chromosome losses affecting chromosomes 1, 6, 13, 14, 15, 16, 18, and 22 were identified by means of whole-genome sequencing and single-nucleotide polymorphism (SNP) array analysis in one study including 12 somatotropinomas, Chromosomal gains, albeit less frequent, were identified in chromosomes 3, 7, 12, 20 and X [135]. In a different study, the same group, using whole-genome sequencing, SNP array, methylation analysis and transcriptome analysis, divided somatotropinomas in two categories: one with recurrent aneuploidy, composed of only gsp negative tumors, and a second one, with chromosomal stability, including a majority of gsp positive tumors. Tumors with chromosomal stability displayed activation of the PKA signaling pathway, while tumors with aneuploidy displayed activation of genes related to cell cycle progression [138]. Similar findings were reported by a different group in a study using tumor array-CGH and transcriptome analysis, which detected recurrent arm-level losses of chromosomes 1, 15q and 16 and extensive gains of chromosomes 5, 7, 10, 19, 20 and X. This study also identified two groups of somatotropinomas based on the proportion of genome disruption; all gsp positive tumors belonged to the low copy number group. High copy number somatotropinomas displayed a high degree of genomic disruption; however, these tumors were more frequently densely granulated. Interestingly, no differences were seen between groups in terms of clinical and biochemical features, as well as tumor size and invasiveness, except that patients with a paradoxical GH response during OGTT had high copy number tumors. Copy number profiles indicating chromothripsis were identified in two patients, one with a germline AIP mutation and one with gigantism [137]. The finding of recurrent large genomic alterations in somatotropinomas in multiple studies was somehow unexpected, considering that PitNETs most often behave as benign tumors.

Another recent multiplatform analysis identified chromosomal losses at 12q21.1 occurring only in somatotropinomas, but not in other subtypes of PitNETs. Somatotropinomas displayed an abundance of hypomethylated regions and from this finding it was concluded that the elevated expression of SST5, GH1 and GH2 in somatotropinomas results from the hypomethylation of the promoter regions of these genes. Higher expression of PD-L1, a target for widely used anti-cancer drugs, was found in somatotropinomas compared with other tumor types, which coincided with the presence of immune infiltrates. Treatment with SSAs resulted in reduced expression of MKI67, which encodes the well-known tumor proliferation marker Ki-67, as well as increased expression of MUC1 and CD40; the latter could be indicative of increased apoptosis [139].

From these studies, we can conclude that chromosomal losses and gains are frequent in somatotropinomas. Despite the finding of recurrent alterations, further studies should be performed to determine whether specific genomic alterations correlate with the disease phenotype.

5. Expert opinion

As it has been proven in the recent years, an adequate treatment of patients with acromegaly or gigantism prevents long-term complications derived from the chronic exposure to IGF-1 excess, thereby normalizing the mortality rate of such individuals. When approaching a patient with any form of GH excess, it should be taken into consideration that guidelines are evidence-based general recommendations meant to facilitate the care of patients, and not rigid rules that cannot be modified. The clinical and biochemical presentation, the disease behavior and the response to treatment are quite variable among patients with GH excess. Disease biomarkers are therefore useful tools to assist the clinical teams when deciding the best therapeutic strategies for each case. Attempts have been done to incorporate multiple disease markers into clinical care. For instance, a risk score has recently been proposed, which should still be validated in clinical studies, taking into consideration the following markers: GH and IGF-1 values at diagnosis, tumor volume, growth, presence of hyperintensity in T2 MRI, a sparsely granulated pattern and Ki-67 staining [8].

Within the last years, clinical studies and translational basic research have identified multiple biomarkers (summarized in Table 1) that should still be adequately validated in large clinical studies and incorporated into the care of patients. A good clinical history and a careful clinical examination can identify cases requiring genetic testing. Patients with specific genetic disorders might require additional clinical tests and a modified therapeutic approach because they might not respond to standard therapies equally as good as sporadic patients. In addition, once a germline genetic defect has been confirmed, genetic testing and counselling should be offered to the rest of the family. In patients who have undergone a surgery, multiple immunohistochemical markers such as granulation pattern, SST₂, AIP and ZAC1 might be useful predictors of the response to SSAs. The gsp oncogene is currently the only well-known recurrent somatic genetic alteration in somatotropinomas, but recent studies have also highlighted the importance of large genomic alterations in such tumors as possible predictors of clinical behavior.

Table 1.

Potential markers of disease behavior in acromegaly and gigantism: Effects on presentation, response to treatment and additional features.

	Marker	Clinical presentation	Response to treatment	Other features	Evidence
Germline genetic defects.	AIP LOF variants	Young onset (usually adolescence). FIPA. Gigantism. Large tumor size. Increased risk for apoplexy.	Poor response to first-generation SSAs.		Analysis of clinical features and response to treatment in two large cohorts.
	GPR101 gene amplification	Very young onset gigantism (early childhood). Sporadic or FIPA presentation.	Poor response to first-generation SSAs. Surgery or radiotherapy usually not effective. Consider pegvisomant as first choice of treatment.	Might require tissue sampling for genetic diagnosis (in cases with mosaicism).	Analysis of all cases published in the literature; no large cohorts of patients available.
	MEN1 LOF variants	Young onset. Usually PHPT in another family member. Screening for additional features required.	Poor response to surgery and SSAs.	Plurihormonal tumors are frequent.	Analysis of clinical features and response to treatment in multiple cohorts.
	CDKN1B LOF variants	Young onset. Screening for additional features required. MEN1-like or FIPA presentation.			Analysis of all cases published in the literature; no large cohorts of patients available.
	SDHx LOF variants	Pheochromocytoma or paraganglioma in the same patient or another family member.		Variants in multiple different genes can cause the same phenotype.	Analysis of all cases published in the literature; no large cohorts of patients available.
	McCune-Albright syndrome	Young onset. Gigantism. Sporadic presentation (mutation not inherited). Extensive pituitary involvement. Screening for additional features required.	Poor respose to first-generation SSAs. Radiotherapy might be contraindicated.		Analysis of multiple cohorts.
	Carney complex	Young onset. Subclinical presentation possible. Screening for additional feaures required. Extensive pituitary involvement.			Analysis of small case series.
	Neurofibromatosis type 1	Young onset. Optic chiasm glioma.			Analysis of small case series.
Histopathology and dysregulated expression at tumor level.	Sparsely granulated pattern	Young onset. Large tumor size. Invasive tumor.	Poor response to first-generation SSAs.		Well-validated in multiple studies and included in the WHO classification of pituitary tumors.
	High E-cadherin expression	Small, not invasive tumor.	Good response to first-generation SSAs.		Small number of patients analyzed, requires further validation.
	High SST2 expression		Good response to first-generation SSAs. Probably also good response to pasireotide.		Well-validated in multiple cohorts, with consistent results for first-generation SSAs. Should become part of routine histopathological studies in the short term.
	High SST5 expression		Poor response to first-generation SSAs if higher than SST2. Probably, good response to pasireotide.		Requires further validation, inconsistent resuts among studies.
	High sst5TMD4 expression		Poor response to SSAs.		Small number of patients analyzed, requires standardization and further validation.
	Filamin A expression		Directly correlates with SSTR2, SSTR5 and DRD2 expression at the mRNA level.		Small number of patients analyzed so far, unclear role as a marker of clinical response.
	Low beta-arrestin expression		Probably, good response to first-generation SSAs.		Small number of patients analyzed so far, requires standardization and further validation. Conflicting results among studies.
	High ZAC1 expression		Good response to SSAs.		Small number of patients analyzed, requires standardization and further validation.
	Low AIP expression	Invasive tumor.	Poor response to SSAs.		Multiple studies, but small number of patients analyzed. Requires standardization and further validation.
	GIPR overexpression	Paradoxical GH increase during OGTT.			Analysis in one study, requires further validation.
Somatic genetic and genomic defects.	gsp oncogene	Small, not invasive tumor.	Good response to first-generation SSAs.		Analyzed in multiple studies, but the results vary among cohorts.
	High copy number alterations			gsp negative	Recently described, requires further investigations to determine its significance.
	Low copy number alterations			gsp positive	Recently described, requires further investigations to determine its significance.

FIPA, familial isolated pituitary adenoma; LOF, loss-of-function; NGS, next-generation sequencing; OGTT, oral glucose tolerance test; PHPT, primary hyperparathyroidism; SSAs, somatostatine analogues; WHO, World Health Organization. See references in text.

Keeping in mind that there are genetic, histological and molecular characteristics that can modify the course of disease and the response to standard treatments in each patient with GH excess will help the medical team recognizing the cases that require special, non-routine tests. Many of the markers reviewed here are not yet widely available and it would not be feasible to implement them in every center, but the patients and their treating physicians can access such tests by means of research studies. Hopefully, once standardized, these novel disease biomarkers will be integrated together with currently applied clinical, biochemical and radiological markers into risk score systems, with the aim of customizing the clinical care of each patient.

Basic research is looking into novel genetic causes, molecular interactions and therapeutic targets that should hopefully be translated into novel recommendations, to change the approach to disease. A number of disease markers have been identified in the recent years, but most of them still require standardization and validation in larger populations before they can be applied in a routine basis; such validation studies should be available in the short term. The cost of genetic studies has dropped dramatically within the last few years, and the trend continues. It is therefore expected that within the next few years many more individuals will have undergone WES or genome sequencing as part of their medical case, which will facilitate the search for specific genetic defects, once a diagnosis of GH excess has been established.

Figure 1.

Simplified scheme of intracellular signaling pathways regulating the response to GHRH and somatostatin in somatotroph cells. Proteins described in the text as potential biomarkers are marked with a red star. See text for detailed description.

Article highlights.

• An appropriate treatment of patients with acromegaly and gigantism greatly prevents complications derived from GH excess and normalizes the mortality rate of these patients to that of the general population.

• The disease course and the response to standard therapeutic approaches varies among patients with GH excess, and the medical teams should acknowledge the existence of markers of disease behavior for tailoring the clinical approach and treatment in each patient.

• A careful clinical history and examination are essential in every case and might be able to identify individuals with disease-associated congenital genetic defects.

• The existence of a germline genetic defect cannot be ruled out solely by a negative family history, due to the possibility of low disease penetrance or de novo genetic defects.

• Clinical features at the physical examination or a personal or familial history of associated neoplasms that would be indicative of a specific genetic syndrome, should be taken into consideration.

• Unless there is a high level of suspicion for a specific condition, next-generation sequencing-based techniques and/or copy number variation analysis are preferred over single-gene screening.

• Low molecular weight cytokeratin immunostaining, available in most centers, differentiates between densely granulated and sparsely granulated somatotropinomas. Sparsely granulated adenomas require special attention because they are diagnosed in younger patients, are larger and more invasive, and have a poor response to surgery and first-generation SSAs.

• A high immunoreactivity to SST₂ is usually a good predictor of a favorable response to SSAs. In addition, proteins involved in the post-receptor effects of SSAs, such as ZAC1 and AIP, can also contribute to predict the therapeutic response.

• Somatic activating mutations of the GNAS gene constitute the only well characterized recurrent mutation so far identified in somatotropinomas. Tumors harboring this genetic defect are smaller, respond better to the treatment with first-generation SSAs, and are more often of the densely granulated subtype.

• Recurrent chromosome losses in somatotropinomas have recently been identified, although their frequency is still uncertain. In general, tumors negative for GNAS mutations display more features of genomic instability, compared with GNAS positive tumors.

• Most of the markers reviewed here are not yet widely available, but currently the patients and their treating physicians can access such tests by means of research studies.