Multiple endocrine neoplasia type 1: a chromatin writer’s block

Authors


Prof. Dr H. Th. Marc Timmers, Department of Physiological Chemistry, University Medical Center Utrecht, STR 3.223, PO Box 85060, 3508 AB Utrecht, The Netherlands.
(fax: +31 88 75 68101; e-mail: h.t.m.timmers@umcutrecht.nl).

Abstract.

Multiple endocrine neoplasia type 1 (MEN1) is caused by inactivating germ line mutations of the MEN1 tumour suppressor gene. The MEN1 gene product, menin, participates in many cellular processes, including regulation of gene transcription. As part of a protein complex that writes a trimethyl mark on lysine 4 of histone H3 (H3K4me3), menin is involved in activating gene transcription. Several functions of the menin histone methyltransferase complex have been discovered through protein interaction studies. Menin can interact with nuclear receptors and regulate transcription of hormone responsive target genes. Menin regulates transcription of cyclin-dependent kinase inhibitor and Hox genes via the chromatin-associated factor LEDGF. Aberrant expression of menin target genes in tumours in MEN1 patients suggests that loss of writing of the H3K4me3 mark contributes to MEN1 tumourigenesis. At present, drugs are being developed that target chromatin modifications. The identification of compounds that could restore H3K4me3 on menin target genes would provide new therapeutic strategies for MEN1 patients.

Introduction

Multiple endocrine neoplasia type 1 (MEN1) is an inherited syndrome characterized by the combined occurrence of parathyroid adenomas, gastroenteropancreatic tumours, pituitary tumours and several other tumour types, often at a young age [1]. Most frequently, MEN1-related tumours cause symptoms due to hypersecretion of hormone products, or to mass effect. Gastrinomas and neuro-endocrine tumours have malignant potential. MEN1 is caused by inactivating germ line mutations of the MEN1 tumour suppressor gene [2]. The product of this gene, menin, is a nuclear protein that is ubiquitously expressed [3]. To learn more about the pathogenesis of MEN1, many studies have addressed the function of menin. One of these functions is regulation of transcription of genes into messenger RNA. Several lines of investigation indicate that menin function is tightly connected to chromatin regulation.

Genes in eukaryotic cells are packaged into chromatin fibres, which are subjected to extensive ATP-dependent remodelling and posttranslational modification processes. Menin has been found as an integral part of chromatin modification complexes. These findings provide a model for the transcription regulatory functions of menin. In addition, several menin-binding proteins have been identified over the last years that can recruit menin and its associated modification enzymes to specific target genes involved in the proliferation and differentiation of cells. Presumably, aberrant expression of these menin target genes contributes to endocrine tumour formation.

In this review, we describe recent findings regarding the function of menin as a co-activator of gene transcription. We discuss the relevance of these findings for MEN1 tumourigenesis and future therapeutic strategies for MEN1 patients.

Menin as part of chromatin-modifying protein complexes

In each cell, roughly 2 m of DNA is wrapped around the core histones (H2A, H2B, H3 and H4) and organized into small packages called nucleosomes (Fig. 1a). The network of DNA and histone proteins forms the chromatin fibre. The histone proteins do not only provide a rigid framework for DNA but are also in fact highly dynamic due to extensive post-translational chemical modifications. Acetylation, methylation, ubiquitination and phosphorylation are amongst the chemical marks that can be not only ‘written’ on but also ‘erased’ from histone tails (Fig. 1a,b). The nearly unlimited number of combinations of modifications forms the ‘histone code’. This code has been proposed to constitute a set of instructions embedded in the chromatin complementary to the genetic code of the DNA sequence [4]. Chromatin modifications can lead to conformational changes of the histone proteins which affect DNA accessibility, but can also allow the recruitment of effector proteins, ‘readers’ of the code that can modulate gene transcription (Fig. 1b).

Figure 1.

 An overview of chromatin modifications (a) DNA is wrapped around core histone proteins into nucleosomes. Each nucleosome contains four pairs of histone proteins (H2A, H2B, H3 and H4). Histone proteins and especially their N-terminal tails can be chemically modified. As an example a schematic representation of the N-terminus of histone H3 is shown, with several modifications indicated. Red dots denote methylation, blue blots indicate phosphorylation and green dots indicate acetylation. (b) Methylation of lysine 4 of histone H3 (H3K4) is written by mixed-lineage leukaemia (MLL) family members, it can be read by the chromatin remodelling protein BPTF (bromodomain PHD finger transcription factor) or by the TFIID component TAF3 (TBP-associated factor 3), and it can be erased by the Jarid1 family of demethylases [47].

Menin was first demonstrated to be part of protein complexes that are involved in erasing the acetyl mark from histones [5]. Deacetylation of histones results in repression of gene transcription. Menin had been shown before to be involved in repression of JunD- and NF-κB-dependent transcription [6, 7]. Telomerase is an enzyme that prevents shortening of chromosome ends and can thereby prevent cells from entering senescence. Increased telomerase activity is a hallmark of stem cells and is also found in many types of cancer. Menin was demonstrated to attenuate transcription of the telomerase (hTERT) gene. Consequently, loss of menin could contribute to the development of endocrine tumour cells. Possibly, repression of hTERT transcription is also dependent on recruitment of deacetylation activity, but it is unclear how menin is targeted to the hTERT promoter [8, 9].

More recently, Hughes et al. and Yokoyama et al. found by immunoprecipitation and protein purification experiments that menin could also be part of mixed-lineage leukaemia (MLL) protein-containing complexes through a direct interaction with MLL proteins [10–12]. Chromosomal rearrangements of the MLL genes leading to oncogenic MLL fusion proteins are found in several types of leukaemia [13]. MLL protein family members contain so-called (SET) domains that have specific histone methyltransferase (HMT) activity for protein trimethylation of lysine 4 of histone H3 (H3K4). Thus, MLL is a writer of the H3K4 trimethylation mark (H3K4me3) [14]. This mark is associated with transcription activation, by recruiting effector proteins such as the chromatin remodelling protein BPTF and the TFIID basal transcription initiation complex (Fig. 1b) [15, 16]. As such, the menin HMT complex was found to activate several target genes, including homeobox domain (Hox) genes, cyclin-dependent kinase inhibitor tumour suppressor genes and nuclear hormone receptor responsive genes [10, 11, 17, 18]. Analysis of genome-wide DNA occupancy revealed that menin is located at 5′-regions of many active genes, but also at 3′-regions and intergenic loci [19, 20].

Factors that tether the menin HMT complex to target genes

The homeobox (Hox) gene clusters contain genes that are involved in the regulation of patterns of development. Several homeobox domain (Hox) genes, such as HoxA7, A9, A10, C6 and C8 have been identified as menin target genes [10, 11]. The cyclin-dependent kinase inhibiting (CDKI) genes CDKN2C and CDKN1B encode the p18Ink4c and p27Kip1 proteins. These proteins are tumour suppressors that restrict cellular proliferation through binding to cyclin-dependent kinases CDK4, 6 and CDK2 respectively. CDKI binding either prevents formation of CDK–cyclin complexes or interferes with CDK–cyclin complex activity resulting in cell cycle arrest. The Cdkn2c and Cdkn1b genes were amongst the first reported menin HMT target genes [18]. However, how menin is recruited to these genes was not clear. The analysis of genome-wide DNA occupancy by menin showed frequent co-occurrence of menin occupancy and H3K4me3 [19]. The ubiquitous presence of menin throughout the genome suggests that the menin HMT complex can be tethered to genes in many different ways. Selective menin promoter recruitment may depend on the specificity of menin–protein interactions with factors such as DNA sequence-specific transcription factors or chromatin components (Fig. 2a,b). In addition, promoter association may be stabilized by the low-affinity sequence-independent interaction of menin with DNA, as has been shown for the Caspase 8 gene (Fig. 2c) [21, 22].

Figure 2.

 Menin can be recruited to DNA by (a) interaction with sequence-specific transcription factors such as nuclear receptors (NRs), (b) interaction with the chromatin-associated factor LEDGF, or (c) a direct interaction with DNA, in all cases leading to activation of transcription. Hox, homeoboxdomain; CDKI, cyclin-dependent kinase inhibitor.

Nuclear receptors

Evidence for menin being tethered to well-defined DNA response elements came from the finding that menin is able to interact with the nuclear hormone receptor for oestrogen (ERα) and the peroxisome proliferator-activated receptor gamma (PPARγ) (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data) [17]. The nuclear receptor protein family consists of 48 receptors which include the receptors not only for steroid hormones but also for thyroid hormone and vitamin A and D. Nuclear receptors are DNA-binding transcription factors that can directly translate changes in hormone levels to alterations in gene expression via recruitment of histone-modifying proteins. The pattern of expression of several of the members of this family is highly tissue-specific, as are the manifestations of MEN1. Co-regulators of nuclear receptor-driven transcription can interact with nuclear receptors though domains that contain so-called LXXLL motifs within α-helical secondary structures [23]. Menin contains one such LXXLL motif; LLWLL (amino acids 263–267). Interestingly, mutation of the LLWLL motif to LLWAA does not disrupt menin binding to PPARγ. However, it does interfere with co-activation of PPARγ-mediated transcription (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data). This indicates that the LLWLL motif in menin is in fact a nuclear receptor interaction domain.

Reduction of menin mRNA levels or the use of MEN1 knock out mouse cells results in reduced levels of H3K4 trimethylation on hormone responsive genes and decreased transcription of these genes (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data) [17]. Furthermore, menin is required for PPARγ-dependent adipogenesis, indicating that the menin HMT complex is in fact important for the biological function of nuclear receptors (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data).

β-Catenin

Additional support that menin can co-activate gene transcription was shown by Sierra et al. [24] who studied the β-catenin protein. β-Catenin is a critical component of the WNT signalling pathway which is involved in many developmental pathways and in various cancers. β-Catenin co-regulates TCF-mediated transcription of the c-Myc oncogene. Aberrant nuclear localization or function leading to up regulation of c-Myc expression is one of the hallmarks of colorectal carcinoma. The menin HMT complex was shown to be recruited to an enhancer of c-Myc by interacting with the C-terminal activation domain of β-catenin [24]. This means that menin is in fact involved in active transcription of the c-Myc oncogene. Conversely, menin HMT was recently demonstrated to collaborate with β-catenin to inhibit proliferation of pancreatic cells [25].

Lens epithelium-derived growth factor

The recent discovery that menin interacts with lens epithelium-derived growth factor (LEDGF) a chromatin-binding factor shows that menin can also be recruited to DNA via chromatin-interacting proteins. Co-localization of MLL, LEDGF and menin was observed on several menin target genes, including HoxA9, CDKN1B and CDKN2C [26]. By using a fusion protein consisting of the chromatin interaction motif of LEDGF and the MLL-ENL fusion oncoprotein, it was elegantly demonstrated that LEDGF and menin are required for bringing MLL to the HoxA9 gene locus [26].

MEN1 mutations affect menin’s function as a recruiter of HMT activity

To assess the relevance of a biological function of menin for MEN1, the effects of disease-related mutations can be studied. In immunoprecipitation experiments, Hughes et al. [11] showed that several patient-derived mutant forms of menin could not recruit HMT activity. This could be caused not only by disruption of the menin HMT complex but also by loss of the interaction between menin and LEDGF [26]. Re-introduction of menin mutants in MEN1 gene knock out mouse cells could neither upregulate expression of the Cdkn2c and Cdkn1b cyclin-dependent kinase inhibitor tumour suppressor genes nor the Caspase 8 gene to wild type levels [18, 22].

Analysis of the interaction of a large number of menin mutants with ERα showed that several but not all mutations disrupt the menin–ERα interaction. In reporter assays however, none of the tested mutants could co-activate nuclear receptor-dependent transcription, suggesting that all of these MEN1 mutants are defective in bringing H3K4me3 activity to ERα [17].

Loss of menin HMT activity in MEN1-associated tumours

Besides the menin HMT complex, several other protein complexes exist that also possess methyltransferase activity directed at histone H3K4. Although the issue is still under debate, it is unlikely that in the absence of menin a global loss of H3K4me3 occurs in cells. Probably, loss of H3K4me3 on specific target genes leads to alterations in gene expression in MEN1 tumours. The role for H3K4me3 in tumourigenesis in general is not clear; however, the link between H3K4me3 and parathyroid tumourigenesis was further strengthened by the studies of another inherited syndrome in which patients develop parathyroid tumours, the hyperparathyroidism jaw tumour syndrome (HPT-JT). HPT-JT is caused by inactivating mutations of the HRPT2 gene, which encodes the parafibromin protein. This protein is a component of the PAF complex, which is critical to establish proper H3K4me3 patterns in yeast [27]. Disease-related parafibromin mutants displayed defective recruitment of HMT activity.

Compelling evidence for the importance of the cyclin-dependent kinase inhibiting genes in endocrine tumourigenesis comes from mouse models. Mice lacking either the Cdkn2c gene (encoding p18Ink4c) or the Cdkn1b gene (encoding p27Kip1) develop tumour phenotypes that show significant overlap with the manifestations of MEN1 [28]. In addition, inactivating germ-line mutations of the CDKN1B gene have been identified in two patient cases that displayed MEN1-like phenotypes, nowadays referred to as MEN4 [29, 30]. Reduced levels of p27Kip1 have been found in tumours in MEN1 patients [18].

It is known that Hox genes are critical for endocrine organ development [31]. However, how dysregulation of these genes can contribute to tumourigenesis is not clear. A translational study on the expression of Hox genes in MEN1 versus non-MEN1-related parathyroid adenomas showed distinct patterns of expression within these groups [32]. This indicates that aberrant expression of Hox genes could also play a role in MEN1 tumourigenesis.

The interaction of menin with nuclear receptors is very interesting as many MEN1-related tumours are known to express nuclear receptors. For example, expression of ERα has been reported in prolactinomas, one of the most frequently occurring pituitary adenomas in MEN1 patients [33]. Nuclear receptors can have both proliferative and anti-proliferative functions. Polymorphisms in the vitamin D receptor (VDR) gene have been associated with increased risk for parathyroid adenomas [34]. Furthermore, VDR levels have been found to be downregulated in sporadic parathyroid adenomas [35]. Menin can directly interact with VDR (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data). We investigated the expression of VDR target genes in MEN1 parathyroid adenomas compared with sporadic adenomas and normal parathyroid tissue. We found that VDR target genes were specifically downregulated in MEN1-related parathyroid adenomas (K. M. A. Dreijerink, C. J. M. Lips, H. Th. M. Timmers, unpublished data).

Taken together, analysis of menin HMT target genes shows that the expression of these genes is affected in MEN1 tumours, which underlines the potential relevance of this mechanism for MEN1 tumourigenesis.

Future directions and clinical implications

Involvement of menin in the ‘writing’ of histone modifications and histone methylation in particular is now well-established. Disease-related MEN1 mutations frequently disrupt menin HMT activity. Generation of mouse models and aberrant expression of menin target genes in MEN1 tumours suggest that disruption of menin-dependent HMT activity may very well be relevant for MEN1 tumourigenesis. In addition, these observations raise the intriguing possibility that several of the previously reported cellular functions of menin, such as its role in co-activating Smad-dependent transcription, but also GTPase activity, DNA replication and DNA repair could also be related to H3K4me3 [36–39].

Several MEN1-related lesions can be treated adequately either by surgical intervention or by treatment with drugs, but gastrinomas and carcinoid tumours can very often not be treated curatively. Furthermore, MEN1 patients often require multiple surgeries during their lifetime, leading to a high disease burden [40]. Therefore, more advanced treatment strategies are needed. Further understanding of the function of menin in histone methylation may lead to the discovery of new therapeutic targets for this cause. Several compounds are known to interfere with histone modification. As loss of menin function leads to reduced H3K4me3 levels on target genes an inhibitor of H3K4 demethylase activity, causing persistence of remaining H3K4trimethylation, could be an effective approach. Several compounds that possess H3K4me3 demethylase inhibiting activity are currently being developed [41]. Histone deacetylase inhibitors (HDACis) are a well-known class of histone modification regulating drugs, with very promising effects in preclinical and clinical studies. The HDACi Vorinostat/Zolinza has recently been approved for the treatment of cutaneous T-cell lymphoma in the United States and many other HDACis are being developed and tested [42]. Interestingly, HDACis also have some indirect demethylase inhibiting activity and may thus hold promise for treatment of MEN1 patients [43, 44].

Strikingly, menin-HMT seems to have a dual role in tumourigenesis. Whereas loss of MEN1 clearly leads to endocrine tumourigenesis, its presence is required for MLL-dependent leukemogenesis and possibly colorectal tumour formation [24, 45]. Furthermore, menin is an important factor in adaptive β-cell proliferation when insulin demand is increased [46]. Because menin inhibits β-cell proliferation, in this role menin may be a pro-diabetic factor. Therefore targeting menin function will have to be tailored to the specific tissue and mechanism aimed to target.

Conclusion

Several recent reports support that menin, the product of the MEN1 gene, is an important factor for the writing of histone H3K4 trimethylation. Menin HMT target gene specificity may be directed by the protein interaction affinity of menin for other transcription and chromatin factors such as nuclear receptors, β-catenin and LEDGF. In MEN1 tumours, writing of H3K4me3 on specific target genes is blocked as shown by the altered expression of these genes in MEN1 patients. Downregulation of Hox, cyclin-dependent kinase inhibitor and nuclear receptor target genes could contribute to MEN1 tumourigenesis. Compounds that interfere with the removal of the histone methylation mark are being developed. As menin-HMT can both be involved in tumour suppression and growth, specific targeting of menin-HMT to treat tumours poses a great challenge for future research.

Conflict of interest statement

No conflict of interest was declared.

Ancillary