Menin induces apoptosis in murine embryonic fibroblasts.

Multiple endocrine neoplasia type I (MEN1) is a hereditary tumor syndrome characterized by multiple endocrine and occasionally non-endocrine tumors. The tumor suppressor gene Men1, which is frequently mutated in MEN1 patients, encodes the nuclear protein menin. Although many tumor suppressor genes are involved in the regulation of apoptosis, it is unclear whether menin facilitates apoptosis. Here we show that ectopic overexpression of menin via adenoviruses induces apoptosis in murine embryonic fibroblasts. The induction of apoptosis depends on Bax and Bak, two proapoptotic proteins. Moreover, loss of menin expression compromises apoptosis induced by UV irradiation and tumor necrosis factor-alpha (TNF-alpha), whereas complementation of menin-null cells with menin restores sensitivity to UV- and TNF-alpha-induced apoptosis. Interestingly, loss of menin reduces the expression of procaspase 8, a critical protease that is essential for apoptosis induced by death-related receptors, whereas complementation of the menin-null cells up-regulates the expression of procaspase 8. Furthermore, complementation of menin-null cells with menin increases the activation of caspase 8 in response to TNF-alpha treatment. These results suggest a proapoptotic function for menin that may be important in suppressing the development of MEN1.

Multiple endocrine neoplasia type I (MEN1) 1 is a hereditary tumor syndrome that is characterized by tumors in multiple endocrine organs, including the pituitary, parathyroid glands, and the pancreatic islets (1,2). The spectrum of tumors found in MEN1 patients has recently been expanded to include other tumors such as lipomas, angiofibromas, melanomas, and collagenomas (1,3,4). The tumor suppressor gene mutated in MEN1 patients, Men1, was identified by positional cloning (5). Loss of heterozygosity at the Men1 locus is frequently observed in tumors from patients carrying a germ line mutation in Men1 (1,2,6). Thus, mutations in Men1 lead to multiple tumors in endocrine as well as some non-endocrine tissues. Men1 encodes a nuclear protein of 610 amino acid residues (7)(8)(9)(10), and its orthologs are found in mouse, rat, zebra fish, fruit fly, and snail (10 -14). Heterozygous Men1 knock-out mice develop a tumor syndrome that closely mimics human MEN1, indicating an essential role for menin in suppressing the development of the disease (15). However, menin does not display any known significant homology to any known consensus protein motifs. This makes it challenging to identify how menin functions as a tumor suppressor.
Menin interacts with a number of proteins, including the transcription factors JunD, NF-B, Smad3, and Pem (16 -19). The interaction between menin and JunD or NF-B inhibits their activation of gene transcription (16,17), presumably leading to inhibition of cell proliferation. Interaction of menin with Smad3, a signal transducer in the TGF-␤ signaling pathway, may facilitate TGF-␤-induced inhibition of cell proliferation (18). Ectopic expression of menin in Ras-transformed NIH 3T3 cells inhibits cell proliferation (20). Although inhibition of cell proliferation may be one way in which menin suppresses tumorigenesis, many tumor suppressor proteins, including p53 and BRCA1, regulate multiple cellular processes, including inhibition of growth, DNA repair, and apoptosis (21,22).
It is unclear whether menin plays a crucial role in apoptosis, although ectopic expression of menin in an insulinoma cell line increases the number of cells stained by annexin V, one marker for apoptosis (23). The current studies investigate the potential role of menin in promoting apoptosis, which could contribute to suppression of the development of MEN1. Here we demonstrate that overexpression of menin via adenoviruses results in apoptosis that depends on the presence of Bax and Bak, two proapoptotic proteins (24,25). Moreover, we show that targeted deletion of menin causes increased resistance to both UV-and TNF-␣-induced apoptosis. Furthermore, menin induces the expression of caspase 8, an essential component in death-related receptor pathways (26). In addition, cells complemented with menin display increased activation of caspase 8 in response to TNF-␣ treatment, consistent with their greater sensitivity to TNF-␣-mediated apoptosis. These results suggest that menin may suppress the development of MEN1, at least in part, by modulating apoptosis.

EXPERIMENTAL PROCEDURES
Plasmid Construction-PCR-amplified menin was inserted into the BglII and NotI site of the adenoviral shuttle vector (pAdTrack-CMV) (27) to generate pAdTrack-menin. Plasmids for generating recombinant retroviruses were made by inserting PCR-amplified human menin cDNA into the BamHI/NotI site of the retroviral vector pMX-puro. The constructs were sequenced to verify the fidelity of the sequence.
Production of Recombinant Adenoviruses and Retroviruses-Production of menin-expressing recombinant adenovirus is based on the pAdEasy system, a simplified method for generating recombinant adenovirus (27). Adenoviral titers were determined by plaque assay as well as fluorescence-activated cell scanning (FACS) analysis for percentage of green fluorescent protein (GFP)-positive cells. Retroviruses were packaged using the Bosc23 packaging cell line as described previously (28).
Preparation of Immortalized Murine Embryonic Fibroblasts (MEFs)-Men1 ⌬N3-8/ϩ mice (Men1 ϩ/Ϫ ) heterozygous for the Men1 locus were maintained on a 129s6/SvEvTac background (Taconic, Germantown, NY) (15). Men1 ϩ/Ϫ male and female mice were mated, and 9.5 days after plugging the females were euthanized. The embryos were placed in gelatin-coated 12-well plates and dissociated into single cells by repeated pipetting in trypsin buffer. LXSN16E6E7 retroviruses that express the human papillomaviral E6 and E7 open reading frame (29) were used to infect the primary MEFs to immortalize the cells. The infected MEFs were subjected to selection with 500 g/ml G418. Two pairs of menin-null and menin-expressing MEF cell lines (heterozygous and wild-type), each pair derived from littermates, were established.
Establishment of Menin-complemented MEF Cell Lines-Immortalized Men1 Ϫ/Ϫ cells (Ͻ passage 12 after isolation from embryos) were seeded in 6-well plates on day 0, infected with vector or menin-expressing retroviruses on day 1, and switched to fresh medium on day 2. Cells were subjected to selection with 2 g/ml puromycin on day 5 (72 h after switching to fresh medium).
Annexin V and Trypan Blue Staining-For annexin V staining of adenovirus-infected cells, 5 ϫ 10 4 cells were seeded in a 6-well dish and allowed to attach for 5 h. After attachment, cells were infected with adenoviruses (day 0). On the next day (day 1) the medium with adenoviruses was aspirated, and fresh medium was added. On day 2 (24 h after switching to fresh medium), cells were processed for staining. Briefly, cells were collected, stained with annexin V-Cy5 as instructed by the manufacturer (MBL International Inc., Watertown, MA), and analyzed on a BD Biosciences LSR cytometer. The percentage of annexin V-positive cells was determined from gated GFP-positive cell populations.
To examine apoptosis of the menin-null MEFs infected with either vector retroviruses or retroviruses expressing menin, the cells were passaged for ϳ2 weeks after initial retroviral infection to obtain sufficient numbers of cells for experiments. For trypan blue staining to detect loss of membrane integrity, 8 ϫ 10 4 cells were seeded in a 60-mm dish on day 0. On day 1, cells were treated with or without UV irradiation (100 mJ/m 2 ) by a Spectrolinker XL-1000 (Spectronics Corporation, Westbury, NY) or with or without TNF-␣ (10 ng/ml, R&D Systems, Minneapolis, MN)/cycloheximide (5.0 g/ml, Sigma). On day 2 (24 h after the treatment), the cells were collected, stained with trypan blue (0.2%), and counted. Duplicate samples were examined for each data point.
To provide an independent assessment of apoptosis in response to UV or TNF-␣ treatment, annexin V staining was measured at 16 h after respective treatments. Briefly, 3 ϫ 10 5 cells were seeded/100-mm dish on day 0, treated or not treated on day 1, and harvested 16 h after the treatment on day 2. Cells were collected and stained with annexin V-fluorescein isothiocyanate from the MEBCYTO® apoptosis kit as instructed by the manufacturer (MBL International Inc.).
Detection of Release of Free Nucleosomal DNA-Cells were harvested 24 -30 h after infection with adenoviruses. The cells were pelleted by centrifugation, and their densities were adjusted to 10 5 /ml using the cell lysis buffer from the cell death detection kit (Roche Applied Science). The cell lysate was diluted 10-fold, and 100 l was used to detect the release of free nucleosomal DNA by enzyme-linked immunosorbent assay based on instructions from the manufacturer.
DNA Microarray and Northern Blotting Analysis-For the microarray, the vector or menin-complemented MEFs were seeded at a density of 2.5 ϫ 10 5 cells/100-mm dish on day 0. The cells were harvested on day 2 for isolation of RNA using the cesium chloride centrifugation method (33). To increase the reproducibility of gene expression profiles, two independent preparations of RNA from each of the cell lines were analyzed. The RNA was processed to generate biotin-labeled RNA probes (34). The probes were hybridized to the Affymetrix murine GeneChip U74A array at the Penn Microarray Core Facility (Philadelphia, PA). Genes whose levels of expression varied 2-fold or more from the control were further analyzed by Northern blotting analysis, which was carried out as described previously (31).
Caspase 8 Enzyme Activity Assay-To compare caspase 8 enzyme activities in vector-complemented and menin-complemented MEFs treated with TNF-␣, 3 ϫ 10 5 cells were seeded in 100-mm dishes on day 0. Starting on day 1, TNF-␣ (10 ng/ml)/cycloheximide (5 g/ml) was added 0, 3, 6, 12, and 24 h prior to harvesting cells for analysis. Cells were harvested using CHAPS lysis buffer, and protein concentration was determined using BCA reagents (Pierce). The reaction mixtures consisted of the following components: 80 l of caspase 8 assay buffer (0.1 M Tris-Cl, pH 8.0, 1 mM EDTA, 1 mM sucrose, 10% glycerol, supplemented with 10 mM dithiothreitol), 100 g of whole cell lysates, 0.5 l of Ac-IETD-FAC (Pharmingen), a highly specific caspase 8 substrate, in a final volume of 160 l. The mixture was incubated for 2 h at 37°C and measured on a fluorometer with an excitation wavelength of 400 nm and an emission wavelength of 480 -520 nm (Bio-Rad). The result from 0 h of TNF-␣ treatment was set to zero and used to normalize the other measurements. All time points were performed in triplicate.
Statistical Analysis and Quantitation-Standard error bars are noted for appropriate experiments, using Microsoft Excel. Quantitation was performed using either NIH Image 1.63 software (National Institutes of Health, Bethesda, MD) or Imagequant software (Amersham Biosciences).

Adenovirus-mediated Expression of Menin in MEFs Results
in Apoptosis-To examine the biological effects of overexpression of menin, we generated menin-expressing adenoviruses (Ad-menin) and control adenoviruses (Ad-GFP). Ad-menin and Ad-GFP, each of which expresses GFP, were used to infect immortalized murine embryonic fibroblasts. Fig. 1A shows that Ad-menin, but not Ad-GFP, causes the majority of infected cells to shrink and detach from the bottom of the plate, characteristic of apoptotic cells. The infection efficiency of Ad-menin and Ad-GFP was similar (Fig. 1A, bottom panels). These results suggest that overexpression of menin results in death of the MEFs.
To examine whether Ad-menin induces apoptosis, the cells were infected with increasing multiplicities of infection (MOIs, from 0 to 20) of either Ad-menin or control Ad-GFP. The infected cells were stained with fluorescent-labeled annexin V, which specifically binds to phosphatidylserine residues that flip from the inner leaflet to the outer leaflet of the plasma membrane during apoptosis (35). The annexin V-stained cells were detected by FACS. Fig. 1B shows that infection of the cells with Ad-menin increases the percentage of the annexin Vstained cells from 7 to 58% (with MOIs from 0 to 20). In contrast, infection of the cells with Ad-GFP only slightly increases this percentage (from 10 to 18%). These results suggest that ectopic expression of menin results in apoptosis in a dosedependent manner.
Analysis of another marker of apoptosis, release of free nucleosomal DNA (36), confirms the above finding ( Fig. 2A). The cells were infected with two independent clones of Ad-menin or Ad-GFP before analyzing the release of free nucleosomal DNA. Fig. 2A shows that Ad-GFP results in only a low level of release of free nucleosomal DNA (0.2-0.3 units), but Ad-menin leads to a marked enhancement of the release of free nucleosomal DNA (0.7-0.9 units). The expression of menin from Ad-menin in the cells was confirmed by immunoblotting using an antimenin antibody (Fig. 2B). Collectively, these results indicate that overexpression of menin, but not GFP, leads to apoptosis in the infected cells. This is consistent with a recent report showing an increased number of annexin V-stained cells in an insulinoma cell line stably overexpressing menin (23).

Ad-Menin Induces Apoptosis through a Bax/Bak-dependent
Apoptotic Pathway-To identify the apoptotic pathway that menin activates and to rule out that overexpression of menin simply results in a nonspecific toxicity, we examined whether ablation of Bax and Bak, two proapoptotic proteins essential for multiple apoptotic pathways (37), blocks Ad-menin-mediated apoptosis. Immortalized wild-type and Bax Ϫ/Ϫ /Bak Ϫ/Ϫ double knock-out (DKO) MEFs (24) were infected with increasing MOIs of Ad-menin or Ad-GFP. In contrast to the shrinkage and detachment we observed in normal cells (Fig. 1A), we saw no such changes in the DKO MEFs (data not shown). To confirm the lack of apoptosis in the DKO cells, the infected cells were stained with annexin V. Analysis by flow cytometry shows that the background annexin V staining accounts for 2-4% of the control wild-type and DKO cells (Fig. 3A). Infection of the wild-type cells with increasing MOIs of Ad-menin markedly increases the percentage of apoptotic cells (54%), whereas Ad-GFP only slightly increases this percentage (14%, Fig. 3A). In contrast, even at the highest MOI, Ad-menin and Ad-GFP cause similar amounts of apoptosis in DKO cells (13%, Fig. 3A).
In agreement with the above results, infection of wild-type cells with increasing MOIs of Ad-GFP only slightly increases the release of free nucleosomal DNA, from 0.3 to 0.5 units (Fig.  3B, columns 1-4). In contrast, Ad-menin increases the release of free nucleosomal DNA in the wild-type cells from 0.3 to 1.1 units (Fig. 3B, columns 1 and 5-7), indicating a role for menin in inducing apoptosis. Infection of the DKO cells with Ad-GFP does not increase the release of free nucleosomal DNA (Fig. 3B, columns 8 -11). In contrast to the wild-type cells, DKO cells infected with Ad-menin do not show cleavage of nucleosomal DNA (Fig. 3B, columns 8 and 12-14), although both the wildtype and DKO cells express a similar level of menin after infection (Fig. 3C). Collectively, these results indicate that menin-mediated apoptosis is dependent on Bax and Bak (25) and that the Ad-menin-induced cell death is not a nonspecific cytotoxicity induced by overexpression of menin.
Complementation of Menin-null MEFs with Menin Sensitizes the Cells to UV-and TNF-␣-induced Apoptosis-Next, we determined whether loss of menin expression, which is similar to what occurs in tumors, affects cellular response to apoptosis.
We introduced menin into the immortalized menin-null MEFs by retroviral infection, generating stable cell lines. Fig. 4A shows that menin is expressed in the menin-complemented cells but not in vector-complemented cells. The cells were then treated with UV irradiation, a well-known apoptotic signal, and stained with either trypan blue to detect disruption of cell membrane integrity or annexin V to detect phosphatidylserine translocation. Fig. 4B shows that only 11% of the menin-null cells die after UV treatment, whereas 31% of the menin-complemented cells undergo apoptosis under the same conditions. Levels of apoptosis in untreated cells are similar. Consistent with this, only 11% of the vector-complemented cells are stained positive for annexin V after UV irradiation, whereas 26% of the UV-treated menin-complemented cells are annexin V-positive (Fig. 4C). These results suggest that introduction of menin back into the menin-null cells renders the cells more sensitive to UV-induced apoptosis.
To test whether menin might also be critical for maximal response to other apoptotic signals, we examined whether loss of menin affects the cellular response to TNF-␣, which induces apoptosis by binding to death receptors (38). Vector-complemented cells and menin-complemented cells were treated with TNF-␣ and stained with either trypan blue or annexin V. Fig.  4D shows that only 17% of the vector-complemented cells die after TNF-␣ treatment, compared with 49% of the menin-complemented cells. Staining for annexin V further confirms these results, as 16% of the vector-complemented cells undergo TNF-␣-mediated apoptosis versus 61% of the menin-complemented cells (Fig. 4E). These results indicate that introduction of menin into menin-null MEFs sensitizes the cells to TNF-␣induced apoptosis.
Menin Induces the Expression of Procaspase 8 mRNA-As menin is primarily a nuclear protein and is implicated in control of transcription (39,40), we performed DNA microarray analysis on vector-complemented cells and menin-complemented cells to determine whether any menin-regulated genes are involved in apoptosis. Procaspase 8, a critical protease that is activated by various death receptors that bind TNF-␣ or Fas ligand (26,38), was identified as a gene that is up-regulated by menin. Procaspase 8 is essential for apoptosis induced by these death receptors. Upon binding of apoptotic ligands such as TNF-␣, the activated death receptors recruit adaptor proteins such as Fas-associated death domain, which in turn binds procaspase 8, leading to generation of active caspase 8 by proteolytic cleavage (26,38). In addition, several members of the homeobox gene family as well as the insulin growth factorbinding protein family were also found to be differentially regulated by menin in the DNA microarray analysis. However, it is not yet clear whether these members are involved in menin-related apoptosis.
Given the clear role of procaspase 8 in apoptosis, we chose to focus our studies on its regulation by menin. Two pairs of immortalized menin-null cell lines and menin-expressing cell lines, each from a distinct group of littermates, were established from 9.5-day embryos (Fig. 5A). Expression of menin from these cells was detected as expected, based on genotyping (data not shown) and Western blotting analysis (Fig. 5A). Expression of procaspase 8 mRNA is 2.9-fold higher in meninexpressing cells than in menin-null cells from the first group of littermates, whereas expression of procaspase 8 is 1.6-fold higher in menin-expressing cells in the second group of littermates (Fig. 5B, top panel).
To further confirm that menin is responsible for up-regula-tion of procaspase 8, one of the menin-null cell lines was infected with control retroviruses or viruses expressing menin, and the expression of transduced menin in the infected cells was confirmed by Western blotting analysis (Fig. 6A). Fig. 6B indicates that complementation of the menin-null cells with menin enhances the level of the procaspase 8 mRNA 3.0-fold. Thus, these results indicate that targeted deletion of menin results in decreased levels of the procaspase 8 mRNA (Fig. 5B), whereas restoration of menin increases the levels of the procaspase 8 mRNA (Fig. 6B). Menin Induces the Expression of Procaspase 8 Protein as well as Caspase 8 Enzymatic Activity in TNF-␣-treated Cells-To examine whether menin-mediated up-regulation of the procaspase 8 mRNA leads to increased expression of procaspase 8, three independent pairs of control cell lines and menin-complemented cell lines were generated. The levels of the procaspase 8 protein from the three pairs were detected by Western blotting analysis. Fig. 7A indicates that complementation with menin enhances expression of the procaspase 8 protein in all three pairs of cell lines. The enhanced levels range from 2.1-to 2.5-fold. In addition, enhanced levels of procaspase 8 may lead to increased caspase 8 activity in response to TNF-␣. As shown in Fig. 7B, caspase 8 activity indeed significantly increases in a time-dependent manner in menin-complemented cells in response to the treatment with TNF-␣ (p Ͻ 0.004 24 h after treatment with TNF-␣), whereas its activity only slightly increases in the control vector-infected cells. Collectively, these results suggest that menin enhances apoptosis at least in part by up-regulation of procaspase 8 protein and, accordingly, by the caspase 8 activity after activation of the death receptors. DISCUSSION Menin, a potent tumor suppressor that suppresses the development of MEN1, interacts with various transcription factors and inhibits the proliferation of oncogenic Ras-transformed cells (20,39). Although the role of menin in the regulation of cell proliferation has been studied, its role in apoptosis has remained unexplored. The current studies suggest a potentially critical role for menin in promoting apoptosis. Transient overexpression of menin in MEFs induces apoptosis that is dependent on the presence of Bax and Bak, two crucial proapoptotic proteins (Figs. 1-3). Complementation of meninnull MEFs with menin, but not vector, renders the cells more sensitive to both UV-and TNF-␣-induced apoptosis (Fig. 4). These findings indicate that extremely high levels of menin induced by recombinant adenoviruses (10ϫ higher than that mediated by retroviral infection; data not shown) can directly cause apoptosis. In contrast, the more physiologically relevant levels of menin induced by retroviruses modulate apoptosis in the presence of apoptotic signals, namely UV and TNF-␣. Perhaps transient, extremely high levels of menin resulting from adenovirus infection trigger downstream apoptotic pathways even in the absence of apoptotic stimuli, whereas physiologically relevant levels of menin potentiate apoptosis induced by various apoptotic signals. Similarly, adenovirus-mediated expression of the tumor suppressor PTEN (phosphatase and tensin homolog deleted on chromosome ten) directly results in increased apoptosis in certain cancer cells, whereas endogenous levels of PTEN do not trigger apoptosis (42,43).
In addition to promoting increased resistance to apoptotic stimuli, loss of the Men1 gene leads to decreased expression of procaspase 8 mRNA, a critical protease in death receptormediated apoptosis (Fig. 5). Moreover, complementation of menin-null MEFs with menin leads to increased caspase 8 mRNA and protein expression (Figs. 6 and 7). Complementation with menin also leads to increased caspase 8 activation in cells treated with TNF-␣, correlating with increased sensitivity to TNF-␣. Given the interaction of menin with other transcription factors including JunD, NF-B, Smad3, and Pem, menin may directly act as a transcriptional co-regulator to modulate the expression of procaspase 8. Alternatively, it may be involved in modifying the promoter of procaspase 8 or the chromatin structure surrounding the promoter. For instance, the promoter of human procaspase 8 has been shown to be methylated in neuroblastoma cells, leading to its decreased expression (45,46). It is also possible that menin may affect other proteins to indirectly regulate expression of procaspase 8.
The ability of menin to up-regulate caspase 8 expression and its enzymatic activity could help to explain the proapoptotic role of menin in cells treated with TNF-␣. TNF-␣ triggers apoptosis by binding to the TNF-␣ receptor, which subsequently recruits adapter proteins that lead to the activation of caspase 8 and the initiation of the apoptotic cascade (38). Menin may therefore potentiate the TNF-␣-mediated apoptotic signal by increasing the amount of procaspase 8 and its activity. However, it is likely that the modulation of caspase 8 is not the only way in which menin promotes apoptosis. Menin may also modulate the expression of other proapoptotic or antiapoptotic genes to regulate apoptosis. Nevertheless, in the DNA microarray analysis, no other genes related to the caspase 8 pathway were differentially regulated by menin.
In conclusion, these studies indicate a role for menin in promoting apoptosis and the expression of procaspase 8. They suggest apoptosis as another critical function of menin, in addition to inhibition of cell proliferation (20,23,47,48) and involvement in DNA repair (30, 47, 49 -51). Attenuation of normal apoptosis in endocrine cells, in combination with increased proliferation and compromised DNA repair capabilities in menin-null cells, may allow cells to accumulate oncogenic mutations, facilitating the development of tumors. Collectively, the current studies suggest a critical role for menin in apoptosis and shed light on the pathogenesis of MEN1.