Multiple endocrine neoplasia type 1 (MEN1) and type 4 (MEN4)

1Multiple endocrine neoplasia (MEN) is characterized by the occurrence of tumors involving two or more endocrine glands within a single patient. The disorder has previously been referred to as multiple endocrine adenopathy or the pluriglandular syndrome. However, glandular hyperplasia and malignancy may also occur in some patients, and the term multiple endocrine neoplasia is now preferred. Four major forms of MEN are recognized and referred to as types 1–4 (MEN1–MEN4), and each form is characterized by the development of tumors within specific endocrine glands (Table 1). All these forms of MEN either may be inherited as autosomal-dominant syndromes or may occur sporadically; that is, without a family history. However, this distinction between sporadic and familial cases may sometimes be difficult, because in some sporadic cases, a family history may be absent because the parent with the disease may have died before symptoms developed. MEN1, which is also referred to as Wermer’s syndrome is characterized by the combined occurrence of tumors of the parathyroid glands, the pancreatic islet cells, and the anterior pituitary. In addition to these tumors, adrenal cortical tumors, carcinoid, facial angiofibromas, collagenomas, and lipomatous tumors have been described in patients with MEN1. However, in MEN2, which is also called Sipple’s syndrome and was previously referred to as MEN2A, medullary thyroid carcinoma (MTC) occurs in association with phaeochromocytoma and parathyroid tumors. In MEN3, which was previously referred to as MEN2B, MTC and pheochromocytoma occur in association with a marfanoid habitus, mucosal neuromas, medullated corneal fibers, and intestinal autonomic ganglion dysfunction leading to megacolon, with rare occurrence of parathyroid tumors. MTC may also occur as the sole manifestation of the MEN2 syndrome, and this variant is referred to as MTC-only. In MEN4, which is also referred to as MENX, patients develop parathyroid and anterior pituitary tumors, which may occur in association with tumors of the reproductive organs, adrenals and kidneys. This review will focus on the clinical features, genetics and epigenetics of the MEN1 and MEN4 syndromes.

2The incidence of MEN1 has been estimated from randomly chosen postmortem studies to be 0.25% and to be 1–18% among patients with primary hyperparathyroidism, 16–38% among patients with gastrinomas, and less than 3% among patients with pituitary tumors. The disorder affects all age groups, with a reported age range of 5–81 years, with clinical and biochemical manifestations of the disorder having developed in 80% and more than 98% of patients, respectively, by the fifth decade. The clinical manifestations of MEN1 are related to the sites of tumors and to their products of secretion. Parathyroid tumors, resulting in primary hyperparathyroidism, are the most common feature of MEN1 and occur in ∼95% of MEN1 patients. Pancreatic islet tumors, which are also referred to as pancreatic neuroendocrine tumors (NETs) consist of gastrinomas, insulinomas, pancreatic polypeptidomas (PPomas), glucagonomas, vaso-active intestinal polypeptidomas (VIPomas) and non-functioning tumors and these occur in ∼40% of MEN1 patients; and anterior pituitary tumors, consisting of prolactinomas, somatotrophinomas, corticotrophinomas, and non-functioning adenomas, occur in ∼30% of patients. In addition, some MEN1 patients may also develop adrenocortical tumors, lipomas, carcinoid tumors, facial angiofibromas and collagenomas. Parathyroid tumors are the first manifestation of MEN1 in more than 85% of patients, and in the remaining fewer than 15% of patients, the first manifestation may be an insulinoma or a prolactinoma. Combinations of these affected glands and their pathologic features (for example, hyperplasia or single or multiple adenomas of the parathyroid glands) have been reported to differ in members of the same family and even between identical twins. MEN1 is inherited as an autosomal-dominant disorder in such families, but a non-familial (i.e., sporadic) form may have developed in 8–14% of patients with MEN1, and molecular genetic studies have confirmed the occurrence of de novo mutations of the MEN1 gene in approximately 10% of patients with MEN1. In the absence of treatment, these tumors have been observed to be associated with an earlier mortality in patients with MEN1. Studies of MEN1-related mortality have demonstrated that 28–46% of deaths are directly related to MEN1, most commonly as a result of malignant pancreatic islet tumors, gastrinomas and foregut carcinoids. In addition, these deaths occur at a significantly younger age than those occurring in unaffected individuals. Furthermore, a recent multicentre study from France and Belgium has suggested that ∼70% individuals with MEN1 currently die of causes directly related to MEN1. In particular, malignant pancreatic islet tumors and thymic carcinoid tumors were associated with a marked increased in risk of death (Hazard ratio >3, P < 0.005).

3A diagnosis of MEN1 may be established in an individual by one of three criteria (Fig. 1). Thus, MEN1 may be clinically diagnosed in an individual on the basis of the occurrence of two or more MEN1-associated endocrine tumors (Fig. 1). In addition, a diagnosis of familial MEN1 is established for an individual who has the occurrence of one of the MEN1-associated tumors and is a first degree relative of a patient with a clinical diagnosis of MEN1. Finally, a genetic diagnosis of MEN1 is made on identification of a germline MEN1 mutation in an individual, who may be asymptomatic and has not yet developed any of the serum biochemical or radiological abnormalities indicative of tumor development (Fig. 2).

4The treatment for each type of MEN1 associated endocrine tumor is generally similar to that for the respective tumors occurring in non-MEN1 patients. However, the treatment outcomes of MEN1-associated tumors are not as successful as those in non-MEN1 patients, for several reasons. First, MEN1-associated tumors, with the exception of pituitary NETs, are usually multiple, thereby making it difficult to achieve a successful surgical cure. For example, MEN1 patients often develop multiple submucosal duodenal gastrinomas, thereby reducing surgical cure rates compared with similar sporadic solitary tumors, such that only ∼15% of MEN1 patients, compared to ∼45% of non-MEN1 patients, are free of disease immediately after surgery, and at 5 years this decreased to ∼5% in MEN1 patients, compared to ∼40% in non-MEN1 patients. MEN1 patients also develop multiple parathyroid tumors, and subtotal parathyroidectomy has resulted in persistent or recurrent hypercalcaemia within 10 years in 20–60% of MEN1 patients, as opposed to ∼4% in non-MEN1 patients. Secondly, occult metastatic disease is more prevalent in MEN1 patients with NETs than in patients with sporadic endocrine tumors. For example, metastases are present in up to 50% of patients with MEN1-associated insulinomas, whereas <10% of non-MEN1 insulinomas metastasise. Thirdly, MEN1-associated tumors may be larger, more aggressive, and resistant to treatment. For example, ∼85% of anterior pituitary tumors in MEN1 patients, as opposed to 64% in non-MEN1 patients are macroadenomas at the time of diagnosis; ∼30% of anterior pituitary tumors in MEN1 patients have invaded surrounding tissue (Hardy classification grades III and IV), compared to 10% in non-MEN1 patients; and >45% of anterior pituitary NETs in MEN1 patients had persistent hormonal over-secretion following appropriate medical, surgical and radiotherapy treatment, compared with between 10% and 40% in non-MEN1 patients. Thus, the treatment of MEN1-associated tumors is difficult requires a multi-disciplinary team of experts.

5The MEN1 gene is located on chromosome 11q13 and consists of 10 exons (Fig. 3) that encode a 610-amino acid protein, referred to as menin. Menin is ubiquitously expressed and is predominantly a nuclear protein in non-dividing cells. Mutations of the MEN1 gene (Figs. 2 and 3) have been characterized, and a total of 1336 mutations (1133 germline and 203 somatic mutations) were reported in the first decade following the identification of the gene. The 1133 germ-line mutations of the MEN1 gene, which consist of 459 different mutations are scattered throughout the entire 1.830 bp coding region and splice sites of the MEN1 gene. Approximately 23% are nonsense mutations, around 41% are frameshift deletions or insertions, 6% are in-frame deletions or insertions, 9% are splice-site mutations, and 20% are missense mutations, and 1% are whole or particular gene deletions (Fig. 4). More than 10% of the MEN1 mutations arise de novo and may be transmitted to subsequent generations. It also is important to note that between 5% and 10% of patients with MEN1 may not harbor mutations in the coding region of the MEN1 gene, and that these individuals may have whole gene deletions or mutations in the promoter or untranslated regions, which remain to be investigated. One study showed that approximately 33% of patients who do not have mutations within the coding region have large deletions involving complete exons. Such abnormalities will not be easily detected by DNA sequence analysis.

6Most (75%) of the MEN1 mutations are inactivating and are consistent with those expected in a tumor-suppressor gene. The mutations not only are diverse in their types but are also scattered throughout the 1830-bp coding region of the MEN1 gene with no evidence for clustering as observed in MEN2 (Table 1). However, some of the mutations have been observed to occur several times in unrelated families (Fig. 3 and Table 2). Mutations at nine sites in the MEN1 gene accounted for over 20% of all the germ-line mutations (Table 2). Of these nine types of mutations, five are deletional and insertional mutations involving codons 83 and 84 (nt359 del 4), 120 (Lys[K]120 del), 210–211 (nt 738 del 4), and codons 514–516 (nt 1656–7 del or ins C); one is a novel acceptor splice site in intron 4, and three are nonsense mutations (Arg98Stop, Arg415Stop, and Arg460Stop). These mutations at these nine different sites could be considered to represent potential “hot” spots (Table 2). Such deletional and insertional hot spots may be associated with DNA sequence repeats, which may consist of long tracts of either single nucleotides or shorter elements ranging from dinucleotides to octanucleotides. The DNA sequences in the vicinity of codons 83 and 84 in exon 2, and codons 210–211 in exon 3, contain CT and CA dinucleotide repeats, respectively, flanking the 4-bp deletions; this finding would be consistent with a replication-slippage model in which misalignment of the dinucleotide repeat takes place during replication, followed by excision of the 4-bp single-stranded loop. A similar replication-slippage model may also be involved at codons 119 and 120, each of which consists of AAG nucleotides encoding a lysine (K) residue. The deletions and insertions of codon 516 involve a poly(C)7 tract, and a slipped-strand mispairing model also is the most likely mechanism to be associated with this mutational hot spot. Thus, the MEN1 gene appears to contain DNA sequences that may render it susceptible to deletional and insertional mutations.

7Correlations between MEN1 mutations and clinical manifestations of the disorder appear to be absent. For example, a detailed study of five unrelated families with the same 4-bp deletion in codons 210 and 211 (Table 3) revealed a wide range of MEN1-associated tumors; all affected family members had parathyroid tumors, but members of families 1, 3, 4, and 5 had gastrinomas, whereas members of family 2 had insulinomas. In addition, prolactinomas occurred in members of families 2, 3, 4, and 5 but not in family 1, which was affected with carcinoid tumors. Another study of seven unrelated families with the same g → a novel acceptor splice-site mutation in intron 4 revealed a similarly wide range of MEN1-associated tumors and a lack of genotype/phenotype correlation. The apparent lack of genotype/phenotype correlation, which contrasts with the situation in MEN2 (Table 1), together with the wide diversity of mutations in the 1830-bp coding region of the MEN1 gene, makes mutational analysis for diagnostic purposes in MEN1 more difficult than that for MEN2.

8A total of 24 different polymorphisms (12 in the coding region [10 synonymous and 2 non synonymous], 9 in the introns, and 3 in the untranslated regions) of the MEN1 gene have been reported (Fig. 3 and Table 4). It is important to recognize the occurrence of these polymorphisms as they need to be distinguished from mutations when performing analysis for genetic diagnosis.

9More than 90% of tumors from MEN1 patients have loss of heterozygosity (LOH), and this has generally been taken as evidence that the MEN1 gene acts as a tumor-suppressor gene, consistent with Knudson’s two-hit hypothesis. However, this LOH represents only one mechanism by which the second hit may occur, with the other mechanisms being intragenic deletions and point mutations. MEN1 tumors (e.g., parathyroids, insulinoma, and lipoma) that do not have LOH have been shown to harbor different somatic and germ-line mutations of the MEN1 gene (Fig. 4), and this is consistent with the Knudson two-hit hypothesis.

10Parathyroid, pancreatic islet cell, and anterior pituitary tumors may occur either as part of MEN1 or more commonly as sporadic, nonfamilial tumors. Tumors from patients with MEN1 have been observed to harbor the germ-line mutation together with a somatic LOH involving chromosome 11q13, or point mutations, as expected from Knudson’s model and the proposed role of the MEN1 gene as a tumor suppressor. However, LOH involving chromosome 11q13, which is the location of MEN1, has also been observed in 5–50% of sporadic endocrine tumors, thus implicating the MEN1 gene in the etiology of these tumors. Approximatley 200 somatic MEN1 mutations (Fig. 4) have been reported between 1997 and 2007, and these have occurred in several different endocrine tumors. Thus, these have been detected in 18% of sporadic parathyroid tumors (total number, n = 452), 38% of gastrinomas (n = 105), 14% of insulinomas (n = 43), 57% of VIPomas (n = 7), 16% of non-functioning pancreatic tumors (n = 32), 60% of glucagonomas (n = 5), 2.0% of adrenal cortical tumors (n = 83), 35% of bronchial carcinoid tumors (n = 26), 3.5% of anterior pituitary adenomas (n = 167), 10% of angiofibromas (n = 19), and 28% of lipomas (n = 8). These somatic mutations are scattered throughout the 1830-bp coding region (Fig. 3), and 18% are nonsense mutations, 40% are frameshift deletions or insertions, 6% are in-frame deletions or insertions, 7% are splice-site mutations, and 29% are missense mutations (Fig. 4). A comparison of the locations of the somatic and germ-line mutations revealed a higher frequency (39% [somatic] vs. 23% [germline]; P < 0.001) of somatic mutations in exon 2, but the significance of this observation in the context of the Knudson two-hit hypothesis remains to be elucidated. The tumors harboring a somatic MEN1 mutation had chromosome 11q13 LOH as the other genetic abnormality, or “hit”, consistent with Knudson’s hypothesis. These studies indicate that although inactivation of the MEN1 gene may have a role in the etiology of some sporadic endocrine tumors, the involvement of other genes, for example, the GNAS1 gene encoding the G protein-stimulatory α subunit, with major roles in the etiology of such sporadic endocrine tumors, is highly likely.

11The role of the MEN1 gene in the etiology of other inherited endocrine disorders, in which either parathyroid or pituitary tumors occur as an isolated endocrinopathy, has been investigated by mutational analysis. MEN1 mutations have been reported in 42 families with isolated hyperparathyroidism (FIHP), and 38% of these are missense mutations, and fewer than 31% are nonsense or frame-shift mutations, which would result in a truncated and likely inactivated protein. This contrasts significantly (P < 0.01) with the situation in MEN1 patients in whom more than 65% of the germ-line mutations are protein-truncation and about 23% are missense mutations (Fig. 4). These observations are consistent with a more likely association between missense mutations and the milder FIHP variant, but it is important to note that the mutations associated with FIHP are also scattered throughout the coding region and not clustered, a situation that is similar to that found for germ-line MEN1 mutations (Fig. 3). Furthermore, the occurrence of protein-truncating mutations in FIHP patients and particularly deletions, such as the 4 bp, involving codons 83–84 (Table 2), which are identical to those observed in MEN1 patients, makes it difficult to establish an unequivocal phenotype–genotype correlation. However, the sole occurrence of parathyroid tumors in these families that harbor MEN1 mutations that are similar to those found in families with MEN1 is remarkable, and the mechanisms that determine the altered phenotypic expressions of these mutations remain to be elucidated. However, nonsense mutations (Tyr312Stop and Arg460Stop) have been detected in MEN1 families with the Burin or prolactinoma variant which is characterized by a high occurrence of prolactinomas and a low occurrence of gastrinomas. Furthermore, a splice mutation (c.446-3c → g) has been detected in an MEN1 kindred from Tasmania, in whom there was an absence of somatotrophinomas. However, some other families with isolated acromegaly do not have abnormalities of the MEN1 gene, even though segregation analysis in families with isloated acromegaly and LOH studies of somatotrophinomas indicated that the gene was likely to be located on chromosome 11q13. Interestingly, mutations of the aryl hydrocarbon receptor interacting protein (AIP) gene, which is also located on chromosome 11q13 and 2.7 Mb telomeric to the MEN1 gene, have been identified in some families with isolated acromegaly. However AIP mutations have not been deteced in patients with MEN1, who do not have MEN1 mutations.

12MEN1 is an uncommon disorder, but because of its autosomal-dominant inheritance, the finding of MEN1 in a patient has important implications for other family members; first-degree relatives have about a 50% risk of development of the disease. Screening for MEN1 in patients involves the detection of tumors and ascertainment of the germ-line genetic state, that is, normal or mutant gene carrier (Fig. 1). Detection of tumors entails clinical, biochemical, and radiologic investigations for MEN1-associated tumors in patients. The characterization of the MEN1 gene has facilitated identification of individuals who have mutations and hence a high risk of acquiring the disease (Figs. 2 and 3).

13MEN1 mutational analysis is helpful in clinical practice in several ways that include: (1) confirmation of the clinical diagnosis; (2) identification of family members who harbor the MEN1 mutation and require screening for tumor detection and early treatment; and (3) identification of the 50% of family members who do not habour the familial germline MEN1 mutation and can therefore be reassured and alleviated of the burden of anxiety of developing tumors. This latter aspect cannot be over-emphasized as it not only helps to reduce the personal cost to the individuals and their children, but also to the health services in not having to undertake unnecessary biochemical and radiological investigations (Table 5). Thus, MEN1 mutational analysis can be useful in clinical practice.

14The current guidelines recommend that MEN1 mutational analysis should be undertaken in: (1) an index case with two or more MEN1-associated endocrine tumors (i.e. parathyroid, pancreatic or pituitary tumors); (2) asymptomatic first degree relatives of a known MEN1 mutation carrier; (3) a first-degree relative of a MEN1 mutation carrier expressing familial MEN1 (i.e. having symptoms, signs, biochemical or radiological evidence for one or more MEN1-associated tumors). In addition, MEN1 mutational analysis should be considered in patients with suspicious or atypical MEN1. This would include individuals with: parathyroid adenomas before the age of 30 years or multigland parathyroid disease; gastrinoma or multiple pancreatic islet cell tumors at any age; or individuals who have two or more MEN1-associated tumors, which are not part of the classical triad of parathyroid, pancreatic islet and anterior pituitary tumors (e.g. parathyroid tumor plus adrenal tumor).

15Family members, including asymptomatic individuals who have been identified to harbor a MEN1 mutation will require biochemical and radiological screening (5). In contrast, those relatives who do not harbor the MEN1 mutation will have their risk of developing MEN1-associated endocrine tumors markedly decreased from 1 in 2 for an autosomal dominant disorder, to 1 in 3000, 1 in ∼100,000 and 1 in 1000 which are the respective risks for parathyroid, pancreatic islet cell and anterior pituitary tumors for the general population; thus, these relatives without the MEN1 mutation will be freed from the requirement of further repeated clinical investigations.

16MEN1 germline mutational analysis should be considered in those presenting at an early age with a single, apparently sporadic MEN1-associated tumor. The occurrence of germline MEN1 mutations in all patients with sporadic, non-familial parathyroid adenomas is 1%, in gastrinomas is 5%, in prolactinoma is 1%, and in foregut carcinoids is 2%. However, the occurrence of MEN1 mutations tumors in patients developing parathyroid tumors below the age of 40 years has been reported to be higher at 5–13%. The occurrence rates of germline MEN1 mutations in individuals presenting with a single apparent sporadic pancreatic NET at similarly younger age, has not been established, and the current guidelines recommend that MEN1 mutational analysis should be considered in those with gastrinoma or multiple pancreatic NETS.

17Mutational analysis in asymptomatic individuals should be undertaken at the earliest opportunity, and where possible in the first decade of life as tumors have developed in some children by the age of 5 years. Appropriate biochemical and radiological investigations aimed at detecting the development of tumors should then be undertaken in these individuals (Tables 1 and 5). The current guidelines recommend that these investigations should commence from the age of 5 years to detect anterior pituitary tumors and insulinomas, the age of 8 years to detect parathyroid tumors, and the age of 20 years to detect gastrinomas and foregut carcinoid tumors (Table 5), and be undertaken thereafter on an annual basis.

18Biochemical screening for the development of MEN1 tumors in asymptomatic members (Fig. 2) of families with MEN1 is of great importance in as much as earlier diagnosis and treatment of these tumors may help reduce morbidity and mortality. The age-related penetrance (i.e., the proportion of gene carriers manifesting symptoms or signs of the disease by a given age) has been ascertained (Fig. 5), and the mutation appears to be nonpenetrant in those younger than 5 years. Thereafter, the mutant MEN1 gene has a high penetrance, being greater than 50% penetrant by 20 years of age and greater than 95% penetrant by 40 years. Screening for MEN1 tumors is difficult because the clinical and biochemical manifestations in members of any one family are not uniformly similar. Attempts to screen for the development of MEN1 tumors in the asymptomatic relatives of an affected individual have depended largely on measuring the serum con centrations of calcium, gastrointestinal (g–i) hormones (e.g., gastrin), and prolactin, as well as on radiologic imaging of the abdomen and pituitary. Parathyroid overactivity causing hypercalcemia is almost invariably the first manifestation of the disorder and has become a useful and easy biochemical screening investigation. In addition, hyperprolactinemia, which may be asymptomatic, may represent the first manifestation in fewer than 10% of patients and may thus also be a useful and easy biochemical screening investigation. Pancreatic involvement in asymptomatic individuals has been detected by estimating the fasting plasma concentrations of gastrin, PP, and glucagon and by abdominal imaging. However, one study has shown that a stimulatory meal test is a better method for detecting pancreatic disease in individuals who have no demonstrable pancreatic tumors by CT. An exaggerated increase in serum gastrin or PP or both proved to be a reliable early indicator for the development of pancreatic tumors in these individuals.

19At present, it is suggested that individuals at high risk for MEN1 (i.e., mutant gene carriers) undergo biochemical screening (Fig. 6) at least once per annum and also have baseline pituitary and abdominal imaging (e.g., MRI or CT), which should then be repeated at 1- to 3-year intervals (Table 5). Screening should commence in early childhood because the disease has developed in some individuals by the age of 5 years, and it should continue for life because the disease has not developed in some individuals until the eighth decade. The screening history and physical examination should be directed toward eliciting the symptoms and signs of hypercalcemia, nephrolithiasis, peptic ulcer disease, neuroglycopenia, hypopituitarism, galactorrhea and amenorrhea in women, acromegaly, Cushing’s disease, and visual field loss and the presence of subcutaneous lipomas, angiofibromas, and collagenomas. Biochemical screening should include estimations of serum calcium, PTH, gastrointestinal hormones (e.g., gastrin, insulin with a fasting glucose, glucagon, VIP, and PP), chromogranin A, prolactin, and IGF-1 in all individuals, and more specific endocrine-function tests should be undertaken in individuals who have symptoms or signs suggestive of a clinical syndrome. Radiologic screening should include an MRI (or CT scanning) of the pancreas, adrenals, and pituitary, initially as a baseline and then every 1–3 years, as well as imaging for foregut carcinoids.

20Approximately 5–25% of patients with MEN1 may not have mutations of the MEN1 gene. This variability in detecting MEN1 mutations may partly be attributable to differences in methods used to identify the mutations; for example most studies do not systematically examine for large gene deletions, which may be found in up to 33% of patients who do not have coding region mutations. In addition, this variability may be due to phenotype ascertainment, as some studies have included non-familial (i.e. sporadic) patients who may have developed only two (or fewer) endocrine tumors, and the detection rate for MEN1 mutations in these patients was found to be <5%. Such patients with MEN1-associated tumors but without MEN1 mutations may represent phenocopies, or have mutations involving other genes. Phenocopy refers to the development of disease manifestations usually associated with mutations of a particular gene, but instead are due to another etiology, and the occurrence of phenocopies, has been reported in 5–10% of MEN1 kindreds. These phenocopies occurred in two settings. Firstly, in the context of familial MEN1 (Fig. 1), in which a patient with one MEN1-associated tumor, e.g. a prolactinoma, did not have the familial mutation; and secondly, in the context of clinical MEN1, in which patients with two MEN1-associated tumors, who did not have an MEN1 mutation, were demonstrated to have involvement of other genes. These genes may include: CDC73, which encodes parafibromin, whose mutations result in the Hyperparathyroid–jaw tumor (HPT–JT) syndrome; the calcium sensing receptor (CaSR), whose mutations result in familial benign hypercalciuric hypercalcemia (FBHH); and the aryl hydrocarbon receptor-interacting protein (AIP), a tumor suppressor located on chromosome 11q13 whose mutations are associated with familial isolated pituitary adenomas (FIPA). The occurrence of MEN1 phenocopies may confound the diagnosis of MEN1 (Fig. 1) in an individual, and it therefore appears advisable to offer genetic testing to determine the MEN1 mutation status to symptomatic family members within a MEN1 kindred, as well as to all index cases (i.e. patient) with two or more endocrine tumors. If a MEN1 mutation is not identified in the index case with two or more endocrine tumors, then clinical and genetic tests for other disorders such as HPT–JT, FBHH, or FIPA should be considered, as these patients may represent phenocopies for MEN1.

21Initial analysis of the predicted amino acid sequence of menin did not reveal homologies to any other proteins, sequence motifs, signal peptides, or consensus nuclear localization signal, and thus the putative function of menin could not be deduced. However, studies based on immunofluorescence, Western blotting of subcellular fractions, and epitope tagging with enhanced green fluorescent protein revealed that menin was located primarily in the nucleus. Menin has been shown to have at least three nuclear localization signals (NLSs) that are located in the C-terminal quarter of the protein (Fig. 3). Interestingly, the truncated MEN1 proteins that would result from the nonsense and frameshift mutations, if expressed, would lack at least one of these nuclear localization signals (Fig. 3). Menin is predominantly a nuclear protein in non-dividing cells, but in dividing cells, it is found in the cytoplasm. The function of menin still remains to be elucidated, but it has been shown to interact with a number of proteins that are involved in transcriptional regulation, genome stability, cell division and proliferation (Fig. 3 and Table 6). Thus, in transcriptional regulation, menin has been shown to interact with: the activating protein-1 (AP-1), transcription factors JunD and C-Jun, and to suppress Jun-mediated transcriptional activation; members (e.g., p50, p52, and p65) of the NF-КB family of transcriptional regulators to repress NF-КB-mediated transcriptional activation; members of the Smad family, Smad3 and the Smad 1/5 complex, to inhibit the transforming growth factor-β (TGF-β) and the bone morphogenetic protein-2 (BMP-2) signaling pathways, respectively; Runx2, also called cbfa1, which is a common target of TGF-β and BMP-2 in differentiating osteoblasts; and the mouse placental embryonic (pem) expression gene, which encodes a homeobox-containing protein. Recently, the forkhead transcription factor CHES1 has been shown to be a component of this transcriptional repressor complex and to interact with menin in an S-phase checkpoint pathway related to DNA damage response. Menin uncouples ELK-1, JunD, and c-Jun phosphorylation from mitogen-activated protein kinase (MAPK) activation and suppresses insulin-induced c-Jun-mediated transactivation in CHO-1R cells.

22Menin has also been reported to directly bind to doubled-stranded DNA and this is mediated by the positively charged residues in the nuclear localization signals (NLSs) in the carboxyl terminus of menin. The NLSs appear to be necessary for menin to repress the expression of the insulin-like growth factor binding protein-2 (IGFBP-2) gene by binding to the IGFBP-2 promoter. In addition, each of the NLSs has also been reported to be involved in menin-mediated induction of caspase 8 expression. The NLSs may therefore have roles in controlling gene transcription as well as targeting menin into the nucleus. Furthermore, gene expression profile studies, using pituitary and pancreatic islet tumors obtained from men1 mouse models, have revealed altered expression of genes involved in transcription, cell cycle and chromatin remodeling.

23A role for menin in controlling genome stability has been proposed because of its interactions with: a subunit of replication protein (RPA2), which is a heterotrimeric protein required for DNA replication, recombination, and repair; and the FANCD2 protein, that is involved in DNA repair and mutations of which result in the inherited cancer-prone syndrome of Fanconi’s anemia. Menin also has a role in regulating cell division as it interacts with the nonmuscle myosin II-A heavy chain (NMHC II-A), which participates in mediating alterations in cytokinesis and cell shape during cell division and the glial fibrillary acidic protein (GFAP) and vimentin, which is involved in the intermediate filament network. Menin also has a role in cell cycle control as it interacts with: the tumor metastases suppressor NM23H1/nucleoside diphosphate kinase, which induces guanosine triphosphate (GTP)ase activity; and the activator of S-phase kinase (ASK), which is a component of the Cdc7/ASK kinase complex that is crucial for cell proliferation. Indeed, menin has been shown to completely repress ASK-induced cell proliferation.

24The functional role of menin as a tumor suppressor also has been investigated, and studies in human fibroblasts have revealed that menin acts as a repressor of telomerase activity via hTERT (a protein component of telomerase). Furthermore, overexpression of menin in the human endocrine pancreatic tumor cell line (BONI) resulted in an inhibition of cell growth that was accompanied by upregulation of JunD expression but downregulation of δ-like protein 1/preadipocyte factor-1, proliferating cell nuclear antigen, and QM/Jif-1, which is a negative regulator of C-Jun. These findings of growth suppression by menin also are observed in other cell types. Thus, expression of menin in RAS-transformed NIH3T3 cells partially suppressed the RAS-mediated tumor phenotype in vitro and in vivo, and overexpression of menin in CHO-IR cells also suppressed insulin-induced AP-1 transactivation, and this was accompanied by an inhibition of C-Fos induction at the transcriptional level. Furthermore, menin expression in Men1-deficient mouse Leydig tumor cell lines induced cell cycle arrest and apoptosis. In contrast, depletion of menin in human fibroblasts resulted in their immortalisation. Thus, menin appears to have a large number of functions through interactions with proteins, and these mediate alterations in cell proliferation.

25Menin which acts as a scaffold protein, may increase or decrease gene expression by epigenetic regulation of gene expression via histone methylation or acetylation. For examaple, menin has been shown to be an integral component of histone methyltransferase complexes that contain members from the mixed-lineage leukemia (MLL) and trithorax protein family. These can methylate the lysine 4 residue of histone H3 (H3K4) and H3K4 trimethylation is linked to activation of transcription. Menin, by serving as a molecular adaptor, physically links the MLL histone methyl transferase with the lens epithelium-derived growth factor (LEDGF), which is a chromatin-associated protein implicated in the eitology of leukemia, autoimmunity and human immuno-defeciency virus-1 (HIV-1) disease. Thus, menin has a critical function in the MLL complex and its regulation of multiple transcriptional pathways, for example, menin, as a component of this MLL complex, regulates the expression of genes such as the Hox homeobox genes and the genes for cyclin-dependent kinase inhibitors, p27 and p18. Interestingly, specific deregulation of a subset of 23 Hox genes has been reported in parathyroid tumors from patients with MEN1. In addition, menin has been shown to directly interact with protein arginine methyltransferase 5 (PRMT5), and to recruit PRMT5 to the promoter of the gene encoding the growth arrest-specific protein 1 (GAS1), which is an important factor for binidng of Sonic Hedgehog (Shh) ligand to its cell surface receptor patched (PTCH1) with subsequent activation of the Hedgehog signaling pathway, which increases repressive histone arginine symmetric dimethylation (H4R3M2S) and suppresses GAS1 expression. GAS1 blocks entry to S phase and prevents cycling of normal and transformed cells, hereby acting as a tumor suppressor. Furthermore, menin has been reported to interact with the suppressor of varigation 3–9 homolog family protein (SUV39H1) to mediate H3K mehylation, and thereby silence transcrpitional activty of target genes. Menin has been shown to directly interact with the nuclear receptor for estrogen (ERα) and to act as a coactivator for ERα–mediated transcription, linking the activated estrogen receptor to histone H3K4 trimethylation. Additional studies have also shown that the interaction of menin with JunD may be mediated by a histone deacetylase-dependent mechanism, via recruitment of an mSin3A-histone deacetylase (HDAC) complex to repress JunD transcriptional activity. Thus, these studies reveal that menin can act both as an activator and repressor of transcription by epigenetic mechanisms that involve integration of different chromatin modification complexes. For example, menin may mediate repression of genes targeted by the transcription factor JunD through the mSin3A-histone deacetylse (HDAC) complex, and the growth factor pleiotrophin gene through polycomb group complex-mediated histone H3K27 methylation. Menin may also promote expression of genes, such as those encoding the cyclase-dependent-kinase (CDK) inhibitors, p27Kip1 and P18INK4c, via interaction with the MLL-HMT complex that mediates histone H3K4 methylation (Fig. 7). Targeting of these epigenetic mechanisms by use of small molecules is providing important insights for the development of therapeutic compounds. For example, the crystal structures of human menin in its free form and in complexes with MLL1 have shown that menin contains a deep pocket that binds to a short peptide of MLL1. High-throughput screening studies to identify compounds that target menin and suppress its interaction with MLL have identified that thienopyrimidine has such functions. Furthermore, their structure activity analyses led to generation of some thienopyrimidine analogs, which bind to wild-type menin but not menin mutants that involve the interaction site with MLL, and others which are capable of inhibiting the MLL-fusion protein-mediated leukemogenic transformation. Thus, understanding the role of menin in epigenetic mechanisms has provided a rational and molecular basis for the development of menin-MLL inhibitors.

26Approximately 5–10% of patients with MEN1 do not have mutations of the MEN1 gene, and these patients may have mutations involving other genes. One of these genes is the CDNK1B, which encodes the 196 amino acid cyclin-dependent kinase inhibitor (CK1) p27kip1 and was identified to be involved by investigations of a recessive MEN-like syndrome in a naturally occuring rat model, referred to as MENX. Rats with MENX were observed to develop parathyroid adenomas, pancreatic islet-cell hyperplasia, thyroid C-cell hyperplasia, bilateral phaeochromocytomas, paragangliomas and cataracts. The disorder was inherited as an autosomal recessive trait. Genetic mapping studies localized MENX to the distal part of rat chromosome 4, a region that also contained the putative tumor suppressor CK1P27kip1, which is also refered to as p27. Mutational analysis of the CDNK1B gene identified, in rats with MENX, a homozygous frameshifting insertion of 8 bp, at codon 177, that resulted in a missense peptide and termination at codon 218. This CDNK1B mutation resulted in an absence of p27 protein in the tumor cells. These findings prompted studies in patients with MEN1, who did not harbor MEN1 mutations, for abnormalities of CDKN1B which in man is located on chromosome 12p13. This revealed that approximatley 3% of these patients with MEN1-associated tumors, such as parathyroid adenomas, pituitary adenomas and pancreatic NETs in association with gonadal, adrenal, renal and thyroid tumors have CDNK1B mutations. To date, 8 different heterozygous loss-of-function CDNK1B mutations have been identified in patients with MEN1-like tumors, and this indicates that MEN4 in man is an autosomal dominant-disorder, unlike MENX in rats which is autosomal recessive. In addition, germline CDNK1B mutations may also, rarely, be found in patients with sporadic (i.e. non-familial) forms of primary hyperparathyroidism.

27Combined clinical and laboratory investigations of MEN1 have resulted in an increased understanding of this disorder, which may be inherited as an autosomal-dominant condition. Defining the features of each disease manifestation in MEN1 has improved patient management and treatment and has also facilitated a screening protocol to be instituted. Application of the techniques of molecular biology has enabled identification of the gene causing MEN1 and detection of mutations in patients. In addition, these recent advances have facilitated the identification of mutant MEN1 gene carriers who are at high risk for development of this disorder and thus require regular and biochemical and radiologic screening to detect the development of endocrine tumors. The function of the protein encoded by the MEN1 gene, menin, has been shown to involve regulation of transcription, which involves genetic mechanisms, genome stability, and cell division, but much still remains to be elucidated. Furthermore, inhibitors of the menin-MLL interaction hold promise for provoking new treatments.