A review of multiomics platforms in pituitary adenoma pathogenesis

Pituitary adenomas (PA), or pituitary neuroendocrine tumors (PitNETs), represent 15% of all central nervous system tumors. Classic description of PitNETs solely by hormonal classification has given way to key transcription factors that play a role in the pathology of PitNETs including steroidogenic factor-1 (SF-1), t-box pituitary transcription factor (TPIT), and pituitary transcription factor 1 (PIT-1). Germline mutations in various familial PitNETs are discussed including those in familial isolated pituitary adenoma (FIPA), multiple endocrine neoplasia (MEN), neurofibromatosis 1 (NF1), and Carney complex. Recent advances in next generation sequencing have improved insight into the pathogenesis of PitNETs. A review of key studies in evaluating the genomic analysis of PitNETs was performed. Chromosomal mutations, whole exome sequencing, microRNA genomics, methylomics and transcriptomics were analyzed. Moreover, the multiomic analysis of various genomic panels has helped to better understand PA classification.


Introduction
Pituitary tumors represent nearly 15% of all central nervous system (CNS) tumors and are mostly comprised of pituitary neuroendocrine tumors (PitNETs), formerly called pituitary adenomas in earlier World Health Organization (WHO) classifications [1]. While most often benign, Pit-NETs can lead to a wide range of clinical symptoms through hormone secretion or mass effect on critical structures. The diagnosis and evaluation of PitNETs has evolved from the original histological description, e.g., acidophil, basophil, or chromophobe tumors as determined by hematoxylin and eosin staining. Most recently, however, there has been a concerted effort to develop a more personalized approach to diagnosing PitNETs that relies on the genetic, transcription, and protein biomarkers. In 2014, Raverot et al. [2] demonstrated the prognostic value of using genetic markers in the identification of tumors by proposing a novel clinicopathological classification that consisted of tumor size, hormonal immunochemistry, invasiveness, and markers of proliferation. The proposed classification demonstrated high prognostic value in predicting post-operative outcomes. With each successive iteration to the WHO classification, there has been a pivot away from the nonspecific nomenclature of "hormone-producing pituitary adenomas" and a shift towards the adenohypophyseal cell lineage designation of the tumors. In 2017, the WHO published the 4th edition of the classification of endocrine tumors and updated their diagnostic guidelines for PitNETs, introducing transcription factors into the diagnostic criteria [3]. This has been further solidified in the 2021 WHO guidelines [4].
The advancement of next generation sequencing genomics, including methylomics, transcriptomics, proteomics and even glycomics, have reinvigorated the discovery of disease pathogenesis. The integration of multiple platforms, termed multiomics, can be an effective way to understand disease across different levels of gene regulation. Biomarkers can also be combined with imaging and clinical assessment. Studying the multiomics of Pit-NETs has the tremendous potential of enhancing our understanding of the pathogenesis, invasiveness, recurrence, and prognosis of these tumors. In this review paper, we provide a detailed account of how multiomics is being utilized for optimizing the diagnosis and treatment of PitNETs.

Incorporation of multiomics in 2017 WHO classification of tumors of the pituitary gland
In 2017, WHO revised the classification of PitNETs to better represent the clinical characteristics of these tumors. A new emphasis was placed on identifying lineagerestricted pituitary transcription factors to determine adenohypophysial cell lineages. Recent studies have revealed that key transcription factors which drive normal adenohypophysial cell differentiation also play key roles in Pit-NET cytodifferentiation and hormone production [5]. The three main adenohypophysial cell lineages and their correlated transcription factors are ( Fig. 1): (1) gonadotrophs driven by expression of steroidogenic factor-1 (SF-1), (2) corticotrophs driven by t-box pituitary transcription factor (TBX19 or TPIT), and (3) somatotrophs, thyrotrophs, lactotrophs, and mammosomatotrophs driven by pituitary tran-scription factor 1 (PIT-1; also known as POU1F1) [6]. Null cell tumors and plurihormonal tumors are considered separately: null cell tumors do not express any transcription factors, immunomarkers, or hormones, whereas plurihormonal tumors produce multiple hormones and are usually associated with PIT-1. The use of transcription factors is more reliable than prior methods that relied on microscopic morphology, immunohistochemistry against hormones, in situ hybridization, reverse hemolytic plaque assays, and electron microscopy [6]. Transcription factor identification is especially useful in tumors with minimal immunopositivity such as silent corticotrophs, somatotrophs, plurihormonal and immunonegative tumors [5]. Approximately 70-80% of non-functional adenomas show staining for gonadotrophins, termed silent gonadotrophin adenomas (SGA) which overlap histologically with null cell adenomas that are more clinically aggressive [7]. Various transcription factors, including SF-1, PIT-1, and TPIT have been able to distinguish SGA from null cell adenomas (NCA). Importantly, while transcriptomics has improved the specificity of adenoma diagnosis, it is only partially descriptive of the tumor and should be used in conjunction with classi-cal methods of diagnosis such as hormone expression [5,8].
The 2017 WHO reclassification also abandoned the term "atypical adenoma", which was previously defined by mitotic index >3%, Ki-67 expression >3%, and positive p53 staining, as its prognostic value was inconsistent [3]. Instead, the authors suggest considering the rate of growth, extent of invasion, and Ki-67 expression to define the aggressiveness of a tumor [3,6].
Several specific PitNET tumors have been identified as especially high-risk. One type, silent corticotroph adenomas are TPIT positive, often large, and tend to invade the cavernous sinus [9,10]. Given their propensity towards more frequent and earlier recurrences, they should be monitored closely postoperatively. Notably, they often exhibit inconsistent Ki-67 and p53 levels and may rarely transform into carcinomas [11]. Crooke's cell adenomas are another aggressive subtype of the corticotroph lineage that require close monitoring following resection given their exceptional tendency to recur [12]. Sparsely granulated somatotoph adenomas are an aggressive subtype of PitNETs which can be identified by their PIT-1 positivity and characteristic sparse granulation on histology [13]. Finally, plurihormonal PIT-1-positive adenomas, formerly known as silent adenoma type III, are large, monomorphous, invasive tumors with very high rates of recurrence [10,14].
The 2021 WHO classification of central nervous system (CNS) follows the 2017 WHO classification, grouping PitNETs by adenohypophyseal cell lineage and expression of pituitary hormones and transcription factors [4]. Both the 2017 and 2021 WHO classifications also recognize a new entity, the pituitary blastoma, which is a rare embryonal neoplasm of infancy composed of primitive blastemal cells, neuroendocrine cells, and Rathke epithelium. Pituitary blastomas are usually seen as part of the DICER1 syndrome and most commonly produce ACTH followed by GH [4].
While the 2017 WHO classification transitioned towards the use of transcription factor use for tumor identification, there has also been a rapidly growing body of literature on the multiomic dysregulations that contribute to the pathogenesis of nonfunctioning, functioning, and familial pituitary adenomas. For each type of tumor, we discuss the genomics, epigenetic, exomic, transcriptomic, and proteomic dysregulations as well as provide their clinical correlates.

Dysregulated signaling pathways in non-functioning PitNETs
Nonfunctioning pituitary adenomas (NF-PitNETs) are typically benign, characterized by their absence of hormonal overproduction, and show specific molecular changes that are hallmarks to their pathogenesis. Using immunohistochemical testing, NF-PitNETs are divided into null cell adenomas and clinically silent gonadotrophs. Based on the 2017 WHO classification, null cell adenomas are defined as a pituitary adenomas without immunohistochemical evidence of cell-type-specific differentiation when using pituitary transcription factors and adenohypophyseal hormones. Silent gonadotrophs are discussed separately in section 4.2.
The MAPK signaling can be dysregulated in NF-PitNETs through alterations in the ERK1/2, JNK, p38 and ERK5 pathways. Through pathway network analysis, Long et al. [17] demonstrated that ERK/MAPK pathways were significantly activated in NF-PA development. It is well reported that ERK/MAPK can regulate proliferation apoptosis, and stress responses. ERK1/2 is upregulated through the cascade reactions in the RAS-Raf-MEK-ERK1/2 activation fashion. Phosphorylation at Ser446 or Ser447 in B-Raf are significantly increased in NF-PitNETs, in turn causing abnormal constitutive activation of the ERK/MAPK pathway. The two final important stress-activated cascades are p38 and JNK; they share synergistic functionality and both play important roles in inflammatory responses and controlling apoptosis. Both p38 and JNK pathways are involved in NF-PitNETs.

DNA hypermethylation in non-functioning PitNETs
Genome-wide methylation assays of functioning vs. nonfunctioning adenomas revealed that NF-PitNETs, in general, are defined by a higher frequency of gene hypermethylation [18]. Assessing β-values, with 0 being unmethylated and 1 being totally methylated, between normal pituitary sections and tumor sections reveal more methylation on average in NF-PitNETs. The researchers were able to identify the aberrantly methylated CpG loci, and upon further analysis of the regions of interest they found significantly higher levels of methylation occurring on select promoters stratifin (SFN), signal transducer and activator of transcription 5A (STAT5A), fibroblast growth factor receptor 2 (FGFR2), dual specificity phosphatase 1 (DUSP1), and protein tyrosine phosphatase receptor type E (PTPRE). These promoters have not been extensively studied in the context of pituitary tumors, but the epigenetic silencing of SFN which encodes for the 14-3-3 protein, a known tumor suppressor, is similar to effects in other tumors such as gynecological malignancies. STAT5A expression is inversely correlated with promoter DNA methylation and downregulated in NF-PitNETs. Methylation profiles can also elucidate which downstream pathways are driven by the changes: Kober et al. [19] found that Wnt signaling, p53 signaling, and PI3K/AKT signaling were significantly more enriched in NF-PitNETs compared to normal pituitary sections.

Transcriptomics in non-functioning PitNETs
Principal component analysis (PCA) of the transcriptomics of non-functional compared to functional Pit-NETs revealed that pituitary tumors clustered according to the transcription factor responsible for cytodifferentiation of the pituitary gland. Pituitary tumors, were seen to cluster among 3 subtypes depending on driving transcription factors [20]. Clinically follicle stimulating hormone (FSH)/leutinizing hormone (LH)-PitNETs and NCAs were driven by SF-1, while adrenocorticotropic hormone (ACTH)-PitNETs, and silent corticotroph Pit-NETs were driven by TPIT, and thyroid-stimulating hormone (TSH), prolactin (PRL)-, and growth hormone (GH)-PitNETs were driven by PIT-1. Pathway enrichment analysis found several NF-PitNETs demonstrated significant alterations of genes involved in metabolism resulting in higher upregulation of calcium signaling pathways compared with functional adenomas. Moreover, the transcriptome of NF-PitNETs showing alterations in calcium-related genes implicated a wide-range of disrupted signaling pathways. Distinctly, ACTH-PitNETs showed enrichment of renin-angiotensin system regulation and TSH-/PRL-/GH-PitNETs showed clustering of fatty acid metabolism genes.
Neou et al. [23] evaluated 134 patients with a large variety of functional tumor types with differing aggressiveness via multiomics, specifically assessing chromosomal alterations, miRNomics, methylomics and RNA transcriptomics. They demonstrated that the methylome of PitNETs stratified the tumors into three groups, based on collapsed CpGs and associated with secretion type: met1 correlated with somatotrophs, lactotrophs, and thyrotrophs; met2 correlated with gonadotrophs; met3 correlated with corticotrophs. They noted that PIT-1 mutations may drive expression of tet methylcytosine dioxygenase 2 (TET2), which encodes a demethylating enzyme where PIT-1 derived tumors demonstrate global hypomethylation.
With regards to assessing epigenetic activity Neou et al. [23] investigated the viability of microRNAs (miRNA) as a method for classification. MicroRNA are noncoding RNAs responsible for gene expression/silencing through RNA cleavage. PitNETs were divided into 4 groups based on miRNA clusters and these groups separated PRL (miR1), GH (miR2), ACTH (miR3), and FSH/LH tumors (miR4).
Finally, investigating the transcriptome of PitNETs by RNA sequencing produced arguably the most informative profile to date. To bypass the need for exome enrichment, researchers may opt for RNA sequencing (transcriptome sequencing), analyzing the RNA being tran-scribed in a cell/tissue at that time. A benefit to high throughput RNA sequencing is that it provides insight into which genes are active and quantifies how much they are transcribed, however it does not identify eventual downstream translated proteins. Using their 134 samples to determine mRNA clusters, Neou and colleagues categorized the tumors into 6 distinct groups (t1-t6) that correlated with the WHO 2017 classification guidelines, with a few notable discrepancies. Mixed GH-PRL PitNETs were clustered together with somatotroph PitNETs (t6 cluster), implying that they are GH-predominant in origin. USP8 wild-type corticotrophs (t1 cluster) appeared more aggressive with greater sphenoid sinus invasion, which was further supported by the transcriptome demonstrating increased epithelial-mesenchymal-transition (EMSA) signature. Overall, multifactor analysis that combined transcriptome classification with GNAS/USP8 mutation status better fit the molecular groups than the WHO 2017 histoprognostic classifications.
Neou et al. [23] performed whole exome, single nucleotide polymorphisms (SNP), DNA methylation, mi-croRNA profiling, and mRNA sequencing on 134 PitNETs. Multifactor analysis provided a comprehensive and new molecular classification for PitNETs that, in addition to confirming the histological relevance of WHO 2017 classifications, provided an accurate pangenomic blueprint for predicting a tumor's aggressiveness and responsiveness to treatment.

Gonadotrophs
Gonadotrophs are a great example of the need for genetic testing to determine the PitNET subtype. Gonadotrophs are commonly negative for FSH and LH but can be reliably detected by their nuclear immunoreactivity for SF-1. Given the lack of hormone secretion, they are often clinically silent and therefore act like nonfunctioning adenomas [24]. Among those that stained positive for a pituitary hormone, FSH-positive tumors were significantly more common than LH-positive tumors. Despite being clinically silent, these tumors are distinct from null cell tumors. In addition to SF-1 expression, these tumors show variable expression of nuclear transcription factors including estrogen receptor α (Erα), and GATA2/3 [25]. When comparing to NCAs, SGAs had lower MIB-1 levels (2.49 ± 1.41%) than NCAs (3.43 ± 2.76%; p = 0.044). Further, Hong et al. [26] showed that NCAs more commonly invade the cavernous sinus than gonadotroph adenomas (p = 0.043), which led to a significant difference in extent of gross total resection (p = 0.039). Gonadotrophs tend to be indolent in nature. In a study for 1166 patients with pituitary adenomas, 12.8% (n = 149) were SGAs and 9.2% (n = 107) were NCAs. NCAs were significantly more common in female patients than SGAs (61.7% vs 26.8%, p < 0.001). Gonadotrophs had a significantly lower recurrence (p = 0.021), adjuvant radiation therapy usage (p = 0.002), and postoperative diabetes insipidus (p = 0.028) than NCAs. NCA subtypes were independently associated with tumor recurrence (HR 3.64, 95% CI 1.07-12.30; p = 0.038), cavernous sinus invasion (HR 3.97, 95% CI 1.04-15.14; p = 0.043) and larger anteroposterior dimension of the tumor (HR 2.23, 95% CI 1.09-4.59; p = 0.030).
Gonadotrophs can also present as part of a synchronous multiple PitNET lesion. In a study of 1055 Pit-NETs, 13 (1.23%) tumors had more than one cell lineage [25]. Of the 13 tumors, 8 were gonadotrophs that co-existed with one of the following tumors: lactrotroph, poorly differentiated PIT-1 tumor, or corticotroph.
Epigenetic dysregulation also significantly varies by PitNET subtype and makes up an important part of PitNET multiomics. In a study of 105 PitNETs, 35 (33.3%) were SGAs and were found to have hypermethylation of estrogen receptor 1 (ESR1) and Ras Association Domain Family Member 1 (RASSF1) genes (p = 0.055 and p = 0.050, respectively) [27]. Both ESR1 and RASSF1 are implicated in cancer progression.
Further, a study by Principe et al. [28] discovered that gonadotrophs have an increased number of CD68+ and CD163+ M2-like macrophages. Presence of M2-like macrophages correlated well with increased gonadotroph tumor invasion that is mediated by collagen degradation via the matrix metalloproteinase-9 enzyme (MMP-9). M2-like macrophages exert a protumor role in solid tumors, promoting aggressiveness, proliferation, and invasion.

Corticotrophs
The second most commonly-identified somatic mutation, a gain-of-function Ubiquitin Specific Peptidase 8 (USP8) mutation, occurs in 30% of corticotroph tumors. Song et al. [29] sequenced 125 adenomas and found recurring mutations of USP8 in corticotrophs. USP8 encodes for a deubiquitinase enzyme that removes conjugated ubiquitin from epidermal growth factor receptor (EGFR), thus preventing EGFR degradation. Gain-of-function mutations in USP8 increase deubiquination of EGFR leading to excess EGFR signaling as well as increased proopiomelanocortin (POMC) levels. Cleavage of POMC leads to the ACTH peptide hormone. Clinically, USP8-positive tumors are smaller, more frequently found in females, produce more ACTH, and have better outcomes when compared to USP8negative corticotrophs. USP8 wild-type PitNETs appeared more aggressive when compared to USP8-mutated PitNETs (p = 0.018) with more common sphenoid sinus invasion (p = 0.007). However, in a study by Albani et al. [30], USP8mutated corticotroph tumors were diagnosed at a younger age (p = 0.028), had higher preoperative 24-hour urinaryfree cortisol levels (p = 0.045), and had a higher recurrence rate (p = 0.026) at the 10-year follow-up than the USP8-wild type tumors. Further, the elevated invasiveness of PitNETs with wild-type USP8 was corroborated by the increased the epithelial-mesenchymal-transition signature in the transcriptome signature of corticotrophs, a genetic marker of tumor aggressiveness. Interestingly, USP8 status did not affect cavernous sinus invasion or MIB1/Ki67 index. Furthermore, somatostatin receptor 5 (SSTR5) expression is increased in USP8-mutated corticotroph PitNETs. The presence of USP8 mutation may reflect more favorable response to somatostatin receptor ligands (e.g., paserotide) and EGFR-directed therapies (e.g., geftinib).
Epigenetic alterations in DNA methylation also play a role in corticotroph PitNET tumorigenesis. Salomon et al. [31] reported that DNA hypomethylation was seen in only some of the PitNET subtypes; more specifically, POMC hypomethylation was noted in ACTH-secreting tumors. It is worth noting that in corticotroph PitNETs, expression of SSTR5 was also higher in USP8-mutated tumors compared with USP8-wild type tumors.

Somatotrophs
The most frequent somatic change in somatotroph Pit-NETs is the heterozygous gain-of-function mutation of the GNAS gene. The mutation to the G-protein subunit negatively affects GTPase activity, increases cAMP production, ultimately resulting in increased cell proliferation. GNAS is an imprinted gene and any somatotroph PitNET with a GNAS mutation will be located on the maternal allele. GNAS mutations are found in 30-40% of somatotroph PitNETs and are clinically relevant because they may increase the sensitivity to somatostatin therapies. Song et al. [29] reported that GNAS-mutated somatotrophs correlated with being less invasive and less resistant to somatostatin analogs. Similarly, GNAS-mutated tumors also had an increased expression of Dopamine Receptor 2 (DRD2); the impact of the DRD2 is well-known from their high prevalence in lactotrophs, where it is thought to explain why lactotrophs are so responsive to dopamine agonists. However, the relationship between increased dopamine receptor expression and responsiveness to dopamine agonists has not been established yet in somatotrophs [32]. Interestingly, the gonadotroph transcription factor SF-1 has been found to be expressed in GNAS-wild type somatotrophs [22]. Somatotrophs have also been shown to have increased SSTR5 and SSTR2 expression in comparison to other SST receptors, which impacts their response to somatostatin analogs (SSA) [33,34]. SSAs mainly target SSTR2 and SSTR5 and have shown promising results. First-generation SSAs, octreotide and lanreotide, preferentially bind SSTR2 and have moderate affinity for SSTR5. The newer SSA, pasireotide, has greater binding affinity for SSTR5 and less for the rest (SSTR2 > SSTR1 > SSTR3). The resistance to SSAs is also influenced by GNAS imprinting [35]. PitNETs with relaxed GNAS imprinting had lower expression of GNAS and SSTR expression, resulting in lower sensitivity to analogs and therefore more aggressive behavior.
DNA hypomethylation of GH1 and SSTR5 is seen in GH-secreting tumors. Song et al. [29] sequenced 125 ade-nomas and found recurring mutations of GNAS in somatotrophs. Also, Song et al. [29] found additional recurring loss-of-function mutations in MEN1 in plurihormonal adenomas producing both GH and PRL from patients without a history of familial MEN1 syndrome or without germline mutations of MEN1. Lastly, alterations of chromosome 1, 2, 11 and 18 positively correlated with tumor recurrence.

Thyrotrophs
Thyrotroph PitNETs account for 0.5-3% of all Pit-NETs [32]. Most of these tumors are sporadic and produce TSH, leading to a clinical presentation consistent with central hyperthyroidism. The most common transcription factors involved in this tumor subtype are the PIT-1, GATA-2/3, β-TSH, and/or a-subunit (a-SU). They also express SSTR1, SSTR2, SSTR3, and SSTR5. Additionally, Luo et al. [32] suggest the involvement of Wnt4 gene in tumorigenesis, as evidenced by the increased Wnt4 gene expression seen in TSH PitNETs.
Thyrotroph tumors are often composed of fascicles of spindle cells and can contain calcification and fibrous connective tissue. Basic fibroblast growth factor (bFGF) is involved in regulation of TSH and mediates the process of fibrosis, which may explain why 40% of thyrotroph tumors are more fibrotic than the rest of the PitNETs subtypes.
TSH-secreting PitNETs also include poorlydifferentiated PIT-1 tumors, previously known as silent subtype 3. These are more aggressive and invasive tumors; they express PIT-1, GATA-2/3, ER-α, a-SU, TSHb, GH, PRL, and/or CAM5.2 and do not have a consistent hormone expression pattern. Furthermore, TSH can be secreted in conjunction with GH and PRL as part of a plurihormonal PitNET. More specifically, co-secretion with GH and PRL is seen in 16% and 10% of TSH PitNETs respectively. Co-secretion with the rest of the hormones is very rare.
While surgery is the first line treatment for thyrotroph tumors, somatostatin analogues can be effective against some TSH-secreting tumors given their SSTR expression. Unfortunately, poorly-differentiated PIT-1 tumors are resistant to medical therapy. Thyrotroph cells also express dopamine receptors on their cell membrane, which is why dopamine receptor agonists, such as bromocriptine and cabergoline can be effective.

Lactotrophs
Lactotroph pituitary adenomas express mainly PRL and arise from PIT-1 lineage of adenohypophyseal cells. They are the most common functioning PitNETs, accounting for 30-50% of all adenomas. In addition to PIT-1, lactotrophs also express ER-α. They are further subdivided into three subtypes: sparsely granulated lactotroph adenomas (SGLA), densely granulated lactotroph adenomas (DGSA), and acidophil stem cell adenomas (ASCA). SGLAs are the most common subtype and have an excellent response to treatment with a dopamine agonist. SGLAs demonstrate Golgi-like PRL expression, while DGLAs and ASCAs show expression diffusely in the cytoplasm. AS-CAs will usually demonstrate scattered GH expression as well diffuse PRL reactivity.
While lactotrophs can appear as part of a familial syndrome such as MEN1, FIPA, or Carney complex, somatic mutations can occur in up to 19.8% of prolactinomas. The most common mutation identified was in the splicing factor 3 subunit B1 (SF3B1), which results in stronger binding of the pituitary-specific positive transcription factor 1, leading to abnormally elevated prolactin levels as well as more rapid tumor growth. The presence of mutated SF3B1 portends worse prognosis for patients. Furthermore, lactotroph PitNETs demonstrated higher expression of Dopamine receptor 2 (DRD2) which explains the effectiveness of dopamine agonists (e.g., cabergoline) as treatment and their adoption as the first-line agent. They also express SSTR5, which is why SSA can influence prolactinomas. Given its affinity for SSTR5, pasireotide seems to have the most effective SSA for this subtype.
Further multiomic research is necessary to elucidate which lactotroph PitNETs are refractory to medical therapy. A study by Delgrange et al. [36] identified lactotroph PitNETs in men as more likely to be treatment resistant and demonstrate histopathological aggressiveness as compared to women, and are thus considered another high-risk variant. As other studies have shown contradicting results [37,38], further investigation utilizing multiomic analysis is thereby needed.

Pathogenesis of familial PitNETs
Although most pituitary tumors are sporadic in nature, 5% of PitNETs are driven by monogenic germline mutations (Fig. 3). The familial subset of PitNETs can be divided into isolated (only pituitary affected) and syndromic, such as multiple endocrine neoplasia 1 and 4 (MEN1/4), neurofibromatosis type 1, and Carney complex. Familial isolated pituitary adenoma (FIPA) and MEN1 account for most familial cases of pituitary adenoma. FIPA is characterized by the presence of PitNETs in two or more family members with no other syndromic features. Twenty percent of FIPA patients are positive for a mutation in the Aryl hydrocarbon receptor Interacting Protein (AIP) gene, which normally regulates inflammatory pathways by stabilizing the aryl hydrocarbon receptor (AHR) in the cytoplasm; a widely distributed tumor suppressor protein expressed in GH cells and PRL-secreting cells of the pituitary. The AIP mutation promotes somatotrophs and lactotrophs in patients in their 20s, with 40% of patients presenting with gigantism, reflective of its impact on younger patients. Another key characteristic is that AIP mutations predispose patients to more aggressive and large tumors with an increased risk for apoplexy, that are poorly responsive to somatostatin analogues.
MEN1 patients typically present with the triad of a Pit-NET, primary hyperparathyroidism, and a pancreatic neu-roendocrine tumor (i.e., gastrinoma) [33,34]. The MEN1 phenotype is inherited in an autosomal dominant fashion with a germline mutation to the MEN1 gene located on chromosome 11, a ubiquitous transcription cofactor that normally regulates G 1 -S cell cycle proteins, in turn serving as a tumor suppressor. It encodes for the MENIN protein that is a scaffold protein and exerts its effect on cell proliferation via histone methylation [33,34]. Patients with MEN1 syndrome that do not present with a mutation to the MEN1 gene often have mutations in CDKN1B; this phenotype is called the MEN4 syndrome. Correspondingly, MEN1 pituitary-related disease usually presents with lactotrophs while MEN4 syndrome patients present with somatotroph PitNETs. Tumors that are MEN1 mutation carriers are more aggressive than MEN1-negative tumors.
Other familial syndromes, such as McCune-Albright syndrome, Carney complex, and X-linked acrogigantism, may result in either PitNETs or, less commonly, pituitary hyperplasia [35]. McCune Albright syndrome is caused by a missense mutation at Gln227 or Arg201 of the GNAS gene during embryogenesis, leading to increased intracellular cAMP and subsequent hormonal dysregulation [39]. Because the mutation is acquired spontaneously during embryogenesis, the cells affected, and myriad symptoms are distributed in a mosaic pattern. The typical triad of symptoms are ostotic fibrous dysplasia, café-au-lait spots, and hyperfunctioning endocrinopathies, including acromegaly in 20% of cases and hyperprolactinemia in 81% [40]. In a review of all 112 patients with MAS and acromegaly, 54% had a PitNET, and many of these tumors were exceptionally resistant to treatment. Surgery, which was attempted in a minority of the patients, rarely cured GH/IGF-1 excess, and medical treatment with somatostatin analogs achieved control of acromegaly in only 30% of patients.
Carney complex is another multiple endocrine neoplasia that is most often inherited in an autosomal dominant fashion [41]. Germline mutation of the PRKAR1A gene on chromosome 17, responsible for 70% of familial cases, leads to dysregulation of protein kinase A (PKA) activity and tumor development in endocrine and non-endocrine organs. PKA is a cAMP-dependent threonine kinase that is involved in numerous intracellular signaling processes and its role in each cell type is dependent on the substates present, its localization within the cell, and the activity of regulatory proteins [41]. Loss of function of PKA regulation, as seen in the PRKAR1A mutation of Carney complex, results in increased mitosis and cell proliferation in specific cells, including the somatotrophs of the anterior pituitary. In a study of 353 patients, this mutation resulted in a GH-secreting Pit-NET in 12% of cases [42]. The PRKAR1A mutation also drives development in 37% of non-familial cases of Carney complex, an even more rare occurrence [41]. Other mutations in the protein kinase A regulatory subunit, such as PRKACB, have been described [43]. Interestingly, the acromegaly that results from Carney complex may be due to a single GH-secreting tumor, multiple tumors, or hypersecretion of normal pituitary gland surrounding a tumor [44]. For this reason, surgeons may need to pursue more aggressive resections, including sacrifice of the pituitary stalk, in Carney complex patients to cure the acromegaly. Evidence from pediatric patients with Cushing's disease also identified CABLES1, a regulator in adrenal-pituitary negative feedback, as a novel pituitary tumor-predisposing gene [45]. Lastly, screening in familial PA determined that 33% of families harbored cadherin-related 23 (CDH23) variants [46].
In X-linked acrogigantism (XLAG), a somatic or germline duplication in the G protein-coupled receptor 1 (GPR101) gene leads to excess release of growth hormone and gigantism in the pediatric population that it affects [47]. Patients can develop either hyperplasia or PitNET, with the most common subtype being a somatotroph/lactotroph mixed adenoma. Excess growth hormone and prolactin result in acceleration of linear growth early in life and hyperprolactinemia, respectively [47,48]. The duplication is inherited in an X-linked dominant fashion, so females who are heterozygous will also develop the pathology. Most often, XLAG arises from a de novo GPR101 duplication and therefore only a single individual is affected; only three familial cases have been described [49]. Approximately 7.8 to 10% of patients with gigantism harbor this mutation [50]. Although PitNETs are often macroadenomas, they are less likely to invade locally or cause pituitary apoplexy than do the AIP-mutated somatotroph/lactotroph mixed macroadenomas [48]. While PitNETs should be treated with surgical resection as first-line therapy, pituitary hyperplasia may be controlled with GH receptor antagonist drugs [48,51]. Neurofibromatosis 1 is a syndrome characterized by changes in skin pigmentation and development of nerve sheath tumors. Its pathology is driven by a mutation in the NF1 gene, which encodes for the tumor suppressor neurofibromin, and is either inherited in an autosomal dominant manner or results from a de novo mutation. Although central nervous system tumors (e.g., optic gliomas) are common, pituitary adenomas are very rarely associated with NF-1 but there does not appear to be a clear link between the two [52]. Similarly, tuberous sclerosis and Lynch syndrome are two entities rarely associated with PitNETs, however, there is not a clear link to PitNET tumorigenesis. Further multiomic investigation is warranted to elucidate these relationships.
The "three PAs" is a new term describing the association between PitNETs, pheochromocytomas, and paragangliomas [53,54]. Although they may co-occur by coincidence, growing evidence suggests classical pheochromocytoma/paraganglioma-predisposing genes may also promote pituitary tumorigenesis. These tumor cells are driven by a mutated form of succinate dehydrogenase, an enzyme critical for cellular energy production, and mutations in any of the four genes encoding the SDH subunits (SDHx; SDHA, SDHB, SDHC, SDHD) are associated with the 3PAs [49,54]. Less frequently, other genetic causes of this association have been implicated, namely the MEN hereditary endocrine syndromes [55][56][57][58]. The PitNETs involved are commonly lactotrophs, somatotrophs, or non-functioning and represent more aggressive and treatment resistant tumors [58]. In a comprehensive review of patients with both PitNETs and pheochromocytoma/paraganglioma, the genetic screening was limited with nearly half of the patients not undergoing any testing, highlighting the need for comprehensive genome-wide analysis to ascertain association/causality [54].

Conclusions
The increased reliance on transcription factor characterization of PitNETs will likely add to current hormonal immunostaining and continue to be refined. Developing a better understanding of PitNET multiomics will be instrumental in improving our management of PitNETs that are refractory to our current gold-standard treatments. The multiomic approach will allow better patient classification, and prediction of prognosis.

Author contributions
NM, RDS, KP researched and wrote the paper. GP, JJE and MK edited the paper. All authors contributed to editorial changes in the paper. All authors read and approved the final paper.

Ethics approval and consent to participate
Not applicable.