Menin promotes hepatocellular carcinogenesis andepigenetically up-regulates Yap1 transcriptionBin Xua,b,c,1, Shan-Hua Lia,b,c,1, Rong Zhenga, Shu-Bin Gaoa,b,c, Li-Hong Dinga, Zhen-Yu Yinc,d, Xiao Lina,c, Zi-Jie Fenga,Sheng Zhangc,d, Xiao-Min Wangc,d, and Guang-Hui Jina,b,c,2

aDepartment of Basic Medical Sciences of Medical College, bState Key Laboratory of Cellular Stress Biology, and cFujian Provincial Key Laboratory ofChronic Liver Disease and Hepatocellular Carcinoma, Xiamen University, Xiamen 361102, People’s Republic of China; and dDepartment of HepatobiliarySurgery, Affiliated Zhongshan Hospital of Xiamen University, Xiamen 361004, People’s Republic of China

Edited by Tak W. Mak, The Campbell Family Institute for Breast Cancer Research, Ontario Cancer Institute at Princess Margaret Hospital, University HealthNetwork, Toronto, Canada, and approved September 20, 2013 (received for review June 25, 2013)

Menin is a scaffold protein encoded by the multiple endocrineneoplasia type 1 (MEN1) gene in humans, and it interacts witha variety of transcriptional proteins to control active or repressivecellular processes. Here, we show that heterozygous ablation ofMen1 in female mice reduces chemical carcinogen-induced livercarcinogenesis and represses the activation of the inflammationpathway. Using ChIP-on-chip screens and ChIP assays, we find thatmenin occupancy frequently coincides with H3K4me3 at the pro-moter of many liver cancer-related genes, such as Yes-associatedprotein (Yap1). Increased menin and Yap1 expression in humanhepatocellular carcinoma specimens was associated with poorprognosis. Our findings reveal that menin plays an important epi-genetic role in promoting liver tumorigenesis, and support thenotion that H3K4me3, which is regulated by the menin–mixed-lineage leukemia complex, is a potential therapeutic target inhepatocellular carcinoma.

H3K4 methylation | interleukin 6

Menin, the product of the MEN1 gene in humans (Men1 inmice), is responsible for the inherited tumor syndrome,

multiple endocrine neoplasia type 1 (MEN1) (1). The phenotypeof mice with a heterozygous loss of Men1 (Men1+/−) is markedlysimilar to the human MEN1 disorder (2). Both MEN1 alleles areinactivated somatically in human endocrine tumors, which sug-gests that menin is a tumor suppressor (3). Supporting this no-tion, menin interacts with the JUN family transcription factorJUND and inhibits its transcriptional activity (4). Interestingly,as an activator, menin interacts with the trithorax group (trxG)proteins (Drosophila) and the mixed-lineage leukemia (MLL)(humans) histone methyltransferase (HMTase) complex, and italters histone tail modifications and the transcription of targetgenes (5–7). The methylated H3K4 can be recognized by otherspecific binding proteins, including WDR5, RbBP5, and Ash2L,that alter chromatin structure and lead to the activation of genetranscription (5). Menin is a highly specific partner for the MLLcomplex, and it links the MLL complex to target gene promotersin MLL-associated leukemogenesis (5–7). To date, the putativebiological function of menin in the liver is not well defined.Homozygous loss of Men1 (Men1−/−) in mice leads to embryoniclethality at embryonic day (E) 11.5–13.5 and is associated withdefective growth and organogenesis, including the liver (8).Microarray screenings indicate that menin is a key regulator ofthe gene networks activated in fibrogenesis and associated withhepatocellular carcinoma (HCC) (9). The potential link betweenmenin and liver disease offers a fresh opportunity to uncover thefunction of menin in the liver.HCC is the main type of liver cancer, and it is governed by

cumulative genetic and epigenetic alterations (10). Research inthe past several years has greatly advanced our understanding ofthe essential role of key regulators and their roles in HCC. Forexample, the essential functions of the Mst1/2, HNF4a–miRNA,IKKβ–NF-κB, and Ptpn11–Shp2 pathways in liver tumorigenesishave been demonstrated in genetically engineered mice (11–14).

Central to the Hippo pathway is a kinase cascade leading fromthe tumor suppressor Hippo (Mst1/2) to Yes-associated protein(Yap1), an oncogene in various tumors (11). Inactivation ofMst1/2 induces Yap1-mediated activation of various target genesfunctionally involved in control of liver organ size and tumori-genesis (15). In accord with the repressive nature of Mst1/2, thegenome-wide analysis of HCC identified Yap1 as an important“driver oncogene” (16). Although excellent studies have dem-onstrated that Yap1 is a major downstream target of the Hippopathway (11, 15), the transcriptional regulation of Yap1 bycritical effectors in HCC has not been defined to date. It is im-portant to uncover how Yap1 activation is associated with otherwell-defined molecular alterations in HCC.This study is designed to investigate the possible tumor pro-

moter function of menin in the liver.

ResultsMenin Is Up-Regulated in HCC and Is Correlated with Poor Prognosis.To unravel the functional significance of menin in HCC, wemeasured the expression of menin in 89 patients’ primary HCCs.The HCC specimens exhibited robust expression and exclusivenuclear staining of menin (42.7%) compared with adjacent tis-sues (Fig. 1A). Kaplan–Meier survival analysis showed that the3-y overall survival (Fig. 1B, P = 0.000) and the recurrence-freesurvival (Fig. 1C, P = 0.002) were significantly lower in themenin–(+) HCC patients than in the menin–(–) patients. Theserum level of alpha-fetoprotein (AFP) was dramatically elevated

Significance

Epigenetic changes commonly occur in hepatocellular carci-noma (HCC) and are associated with aberrant gene expression.Most studies have focused on epigenetic gene-silencing events;therefore, the mechanism that promotes gene activation inHCC is not well established. We identify an epigenetic activa-tion mechanism whereby menin promotes Yes-associatedprotein (Yap1) transcription, which is associated with a poorprognosis for HCC patients. Substantial overexpression of themenin–mixed-lineage leukemia complex is associated with in-creased histone 3 lysine 4 trimethylation at certain loci of thetumor promoter in HCC. Heterozygous ablation of multipleendocrine neoplasia type 1 (Men1) in mice reduces diethylni-trosamine-induced development of HCC. Our findings revealthat menin plays an important epigenetic role in up-regulatingYap1 transcription, leading to liver tumorigenesis.

Author contributions: G.-H.J. designed research; B.X., S.-H.L., R.Z., S.-B.G., L.-H.D., Z.-J.F.,and S.Z. performed research; Z.-Y.Y., S.Z., and X.-M.W. contributed new reagents/analytictools; B.X., S.-H.L., X.L., and G.-H.J. analyzed data; and G.-H.J. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1B.X. and S.-H.L. contributed equally to this work.2To whom correspondence should be addressed. E-mail: ghjin@xmu.edu.cn.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1312022110/-/DCSupplemental.

17480–17485 | PNAS | October 22, 2013 | vol. 110 | no. 43 www.pnas.org/cgi/doi/10.1073/pnas.1312022110

Dow

nloa

ded

by g

uest

on

June

3, 2

020

in the menin–(+) group compared with the menin–(–) group (Fig.1D, P = 0.001). Further correlation analyses in both cohorts showedthat the robust expression of menin was associated with the moreaggressive phenotype of HCC, including tumor multiplicity andneoplasm staging (Table S1).To explore how menin influences liver cancer cells, the HepG2

cells were stably transfected with either a control (shLuc) orshMEN1 (17). MEN1 shRNA substantially repressed HepG2 cellproliferation and colony forming activity (CFA) (Fig. 2 A–C).Conversely, ectopic expression of menin increased the CFA inHepG2 and PLC/PRF5 HCC cell lines (Fig. 2 D–F). Further-more, reduced expression of menin significantly suppressed tu-mor volume and weight in two independent pairs of MEN1shRNA knockdown (KD) HepG2 xenografts (Fig. 2 G and Hand Fig. S1). Overall, these findings suggest a potential tumor-promoting function of menin in HCC.

The Essential Role of the Menin–MLL Complex in Yap1 Transcriptionand Function. Currently, it is poorly understood whether alter-ations of active histone modifications, including those of H3K4,are associated with liver tumorigenesis. We further identified thegenomic occupancy of menin and two associated transcription-ally active histone modifications, H3K4me3 and H3K79me2 inHepG2, using ChIP-on-chip screens. Our data showed thatmenin occupied the promoter regions of thousands of humangenes (Fig. S2 A and B and Dataset S1). Menin occupancy fre-quently coincides with H3K4me3 and H3K79me2; however,menin can also target promoters independently of H3 mod-ifications (Fig. S2B). Menin occupancy at the HOXA clusters isextensive with broad footprints that spanned intergenic and in-tragenic portions of the HOXA genes in chromosome 7 (Fig.S2A), which is similar to the epigenetic-regulating features ofmenin in leukemia systems (5–7). Hyperexpression of HOXAgenes, including HOXA5, 7, 10, and 13, was observed in a series

of HCC specimens, and deregulation of HOXA13 led to cell-cyclealterations (18). Fig. S2C shows that genes (HOXA7, 9, and 13)with promoters bound by menin were correlated with highermRNA levels in menin-overexpressing PLC/PRF5 cells (Fig.S2C). We previously found that menin represses cell proliferationand transcription of pleiotrophin (PTN) through polycomb group(PcG)-mediated H3K27me3 in lung adenocarcinoma and mela-nomas (19, 20). Nevertheless, the PTNmRNA level is augmentedby menin in HepG2 cells (Fig. S2D), which further implies a tu-mor-promoting function of menin in HCC in a tissue-specificmanner. Unsupervised hierarchical clustering analysis indicatedthat the top 26 genes with menin promoter occupancy coincidewith H3K4me3 and H3K79me2 in ChIP-on-chip, and these genesare closely related with tumor development and progression, in-cluding the proto-oncogene Yap1 (Fig. S2E). These findings in-dicate an interesting epigenetic mechanism by which meninregulates gene expression through H3 modification in HCC.We explored the relationship between menin and Yap1 ex-

pression. The substantial MEN1 siRNAs reduced Yap1 mRNAand protein expression in HepG2 cells (Fig. 3A), whereas theexpression and nuclear localization of Yap1 were increased inmenin-overexpressing HepG2 cells (Fig. 3B and Fig. S3 A andB). Following the increased total Yap1 protein level resultingfrom menin expression, phospho-Yap1 (Ser127) was also slightlyelevated in PLC/PRF5 cells (Fig. S3C). Nevertheless, there areno obvious effects on Yap1 protein expression by menin over-expression or MEN1 siRNA KD in other types of cell lines, in-cluding MCF-7, Wilm’s, and A549 (Fig. S3 D and E), suggestinga liver-specific function of menin in regulating Yap1. Correlated

Fig. 1. Up-regulation of menin in HCC patients correlates with poor prog-nosis. (A) H&E staining and IHC staining of menin in HCC paraffin sections(n = 89). The dotted lines indicate juncture of tumor (T) and adjacent tissue(N). Menin was easily detectable in the nucleus of certain HCC. Originalmagnification, 100× or 400×, respectively. (B and C) Kaplan–Meier curves foroverall and tumor-free survival in menin-hyperexpressing (+) and -under-expressing (–) HCC patients (P = 0.000 and P = 0.002, respectively, log ranktest). (D) Serum AFP level of HCC patients with menin–(+) (n = 37) or (–) (n =50). In each panel, the line indicates the median (P = 0.001).

Fig. 2. Menin promotes HCC cell proliferation and HepG2 xenograftgrowth. (A) The HepG2 cells were stably transfected with either controlshLuc or shRNA against MEN1 retrovirus, and the expression of menin wasconfirmed by Western blot. (B and C) The proliferation and CFA assaysperformed by shLuc and shMEN1 HepG2 cells (n = 3). (D) HepG2 and PLC/PRF5 cells were stably transfected with either empty vector or MEN1-over-expressing pLNCX2–MEN1 retrovirus, and the expression of menin wasconfirmed by Western blot. (E and F) The transfected cells were used toperform CFA assays. The above data are mean ± SD (n = 3). (G) Both shLucand shMEN1 KD HepG2 cells were s.c. transplanted into nude mice (n = 13 or14, respectively), and the tumor volume (mean ± SEM) was measured every2 d at the indicated time points after transplantation (P < 0.05). (H) Thetumor weight and middle tumor pictures (P = 0.000).

Xu et al. PNAS | October 22, 2013 | vol. 110 | no. 43 | 17481

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

June

3, 2

020

with reduced Yap1 expression, the specific targeting of Yap1 bysiRNA dramatically reduced the HepG2 CFA (Fig. 3C and Fig.S3F). Furthermore, the promoted CFA by ectopic expression ofmenin was reduced by the Yap1 KD in HepG2 cells (Fig. 3D andFig. S3G).Yap1 is an important “driver gene” for HCC (16), however the

transcriptional regulation of Yap1 by critical effectors in HCChas not been defined to date. In our ChIP-on-chip analyses, wefound that menin occupancy coincides with the transcriptionalactivation of H3 modification at Yap1 promoter loci (Fig. S2E),which raises an interesting hypothesis of whether menin regu-lates Yap1 transcription through epigenetic mechanisms, such asH3K4me3. First, we assessed the expression of the MLL–HMTasecomplex in HCC using Western blotting. Associated with immu-nohistochemistry (IHC) observation, menin (6/11), Ash2L (9/11),WDR5 (6/11), RbBP5 (5/11), and MLL (5/8) expression wererobustly activated in HCC compared with the surrounding tissues(Fig. 3E). We further performed ChIP assays using three distinctpairs of primers for the Yap1 promoter loci (Fig. 3F). As expected,the ChIP assays clearly showed that menin bound to the Yap1promoter, and MEN1 KD notably reduced binding of menin atYap1 loci but not at the GAPDH locus (Fig. 3F). Further, eitherMEN1 or MLL KD by siRNA reduced the H3K4me3 level atYap1 loci, especially in primer pair 1 (PP1) and PP2 (Fig. 3G).Nevertheless, although menin bound to the promoter of Yap1,there are no obvious effects on H3K4me3 levels at Yap1 loci bymenin overexpression in other types of cell lines, including MCF-7, Wilm’s, and A549 (Fig. S3 H and I), which further suggests theliver-specific function of menin in regulating Yap1 was dependenton H3K4 histone remodeling. In addition, the substantial KD ofmenin and MLL clearly reduced Yap1 expression (Fig. 3A andFig. S3J) and CFA (Fig. 3H and Fig. S3K) in HepG2 cells, re-spectively. It appears that a pivotal biological role of Yap1 in theliver is attributable partly to the menin–MLL complex.

Heterozygous Men1 Ablation Reduced Diethylnitrosamine-Induced HCCDevelopment in Mice.The diethylnitrosamine (DEN)-induced HCCmouse model is widely used to investigate the features of human

HCC (21). Because Men1−/− causes embryonic lethality in mice,we usedMen1+/− mice to address the tumor-promoting functionof menin in development of primary HCC. We treated Men1WT

and Men1+/− mice with DEN at 15 d of age. After 8 mo, theMen1WT and Men1+/− mice developed visible hepatic tumor foci(Fig. 4A), and histological examination revealed that the livertumors were highly aggressive HCCs (Fig. 4B). We did not observetypical pancreatic islet tumors in either Men1WT or Men1+/− miceat 8 mo (Fig. S4A). These observations ruled out the possibilitythat metastatic tumors came from pancreatic islets to the livers. Asshown in Fig. 4C, 47.6% of the Men1WT female mice developedHCC; however, a significantly lower HCC incidence of only 8.33%was observed in Men1+/− female mice (P < 0.05). Furthermore,Men1+/− dramatically reduced the tumor multiplicity and maximaltumor size in female mice (Fig. 4 D and E). Compared withMen1WT mice, the tumor–liver ratio was reduced to almost one-third in Men1+/− female mice, and it was correlated with a slightlyreduced liver–body weight ratio (Fig. 4 F andG).Men1+/− femalesdisplayed significantly less hepatic injury after DEN administra-tion, which was determined by the reduced serum AFP, IL-6, andTGF-β (Fig. 4 H–J).Unexpectedly, all male mice developed HCC with no differ-

ence in tumor incidence (Fig. 4C) betweenMen1WT andMen1+/−.To elucidate the contribution of menin in hepatocarcinogenesisin male mice, 6-wk-old male mice were exposed to DEN. Ad-ministration of DEN resulted in the stimulation of the Akt, alsoknown as protein kinase B (PKB), STAT3, and MAPK pathwaysin the liver; however,Men1+/− reduced their activation (Fig. S4B).Further, the serum IL-6 and TNF-α concentrations up-regulatedby DEN exposure were remarkably reduced in Men1+/− malemice (Fig. S4 C and D), which indicates that menin participatesin the liver injury response of male mice at early stages. Thecarbon tetrachloride (CCl4)–induced mouse model incorporateschronic injury, inflammation, and fibrogenesis and shares severalfeatures with the microenvironment in which the majority ofhuman HCCs arise (22). In this model, we found that expressionof menin, Yap1, pSTAT3, and pAKT was increased in the liversof males and females exposed to CCl4 (Fig. S4E, lanes 1–3 and

Fig. 3. The essential role of the menin–MLL com-plex in Yap1 transcription and function. (A) ThesiLuc and each of the three distinct siRNAs (siMEN1-1, 2, and 3) that specifically target MEN1 weretransfected into HepG2 cells, and the MEN1 andYap1 mRNA and protein expression were de-termined by quantitative reverse transcription(qRT)-PCR and Western blotting, respectively. TheP1 and P2 are two distinct Yap1 primers. (B) Im-munofluorescent detection of Yap1 (green), DAPI(blue), and the merged signals in control andmenin-overexpressing HepG2 cells. (C) The siLuc andeach of the three distinct siRNAs (siYap1-1, 2, and 3)specifically targeting Yap1 were transfected intoHepG2 cells. The generated cells were used to per-form CFA assays (n = 3). (D) Either siLuc or siYap1-1was transfected into control and menin-over-expressing HepG2 cells, and the generated cellswere used to perform CFA assays. (E) Western blotanalysis of 9–11 paired samples of HCC (T) and ad-jacent tissues (N) from the same patients. (F) Aschematic representation of the human Yap1 andPPs used for ChIP assays. ChIP assays were per-formed using the antibody against menin in shLucand shMEN1 HepG2 cells. (G) ChIP assays using an-tibody against H3K4me3 were performed in siRNA-targeting Luc,MEN1, orMLL HepG2 cells. The rabbitIgG served as a negative control for ChIP assays. (H)The siLuc and each of the three distinct siRNAsspecifically targeting either MLL or MEN1 weretransfected into HepG2 cells, and the transfectedcells were used to perform CFA assays.

17482 | www.pnas.org/cgi/doi/10.1073/pnas.1312022110 Xu et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

020

7–9). Nevertheless, the Men1+/− effectively reduced their effectin female mice but not in males (Fig. S4E, lanes 4–6 and 10–12).In addition, administration of DEN significantly increased themenin levels in the livers of C57BL/6 males compared withfemales at 4 mo (Fig. S4 F and G). These results suggest thatalthough the function of menin is involved in inflammation re-sponse at an early stage, the heterozygous ablation of Men1 waseasily compensated in liver injury and inflammation in male micecompared with female mice.To evaluate the potential relationship of the menin–Yap1 axis

and liver tumorigenesis, we collected liver tumors and surroundingtissues. Because the primary tumors resulting from DEN exposure

of Men1+/− female mice were not large enough for Westernblotting analysis, we used tumors from Men1WT female mice.We detected elevated menin, Yap1, MLL, Ash2L, RbBP5, andWDR5 levels in the liver tumors (Fig. 4K), as well as pAKT,pSTAT3, pERK1/2, and p65 (Fig. S4H). In ChIP assays, fol-lowing the increase of menin at the Yap1 loci (Fig. S4I), theH3K4me3 levels increased nearly threefold in tumors comparedwith surrounding tissues (Fig. 4L). The H3 antibody ChIP wasused as a loading control (Fig. S4I).

The Menin–Yap1 Axis Responded to CCl4-Induced Liver Inflammation.HCC promotion depends on the microenvironment, includinginflammation pathways (12, 13). To determine the relationshipbetween menin–Yap1 axis and inflammation pathways, we usedCCl4-induced inflammation mouse model. Exposure to CCl4significantly stimulated the expression of menin, MLL, and Yap1in the livers of male C57BL/6 mice at 24 and 48 h (Fig. 5A). Thisresponse is associated with elevated mRNA levels of IL-6 andTGF-β and serum concentrations of IL-6 (Fig. S5 A and B). Anincreased binding of menin and accumulation of H3K4me3 atthe Yap1 promoter were revealed by ChIP assays (Fig. 5B andFig. S5C). In accord with the reduction of Men1 or Yap1 ex-pression by siRNA, the IL-6 mRNA level was markedly reducedin primary isolated liver Kupffer cells (KCs) (Fig. 5 C and D).Yap1 KD dramatically reduced IL-6 mRNA expression in twoindependent pairs of Yap1 siRNA KD HL-7702 and HepG2 cells(Fig. 5 E and F and Fig. S5 D and E). Furthermore, the up-regulation of IL-6 by ectopic expression of menin was reduced bythe Yap1 KD in HepG2 cells (Fig. 5G), which suggests thatmenin regulates IL-6 expression at least partly through Yap1.Finally, the protein or mRNA expression of menin and Yap1 wasincreased in HL-7702 or HepG2 cells exposed to IL-6 (Fig. 5 Hand I). These results point to an interesting HCC-related positivefeedback loop between menin–Yap1 and IL-6.

Yap1 Is Epigenetically Regulated by Menin and Correlated with PoorPrognosis in Human HCCs. To confirm the clinical significance ofthe menin–Yap1 axis, we sought to address the deregulation ofYap1 in HCCs. The HCC specimens exhibited robust expressionand exclusive nuclear staining of Yap1 compared with the ad-jacent tissues (Fig. 6A). Kaplan–Meier survival analysis showedthat the 3-y overall survival rate of patients with Yap1– (+)HCCs was significantly lower than that of patients with Yap1–(–)HCCs (Fig. 6B, P = 0.000). A significantly lower recurrence-freesurvival rate was found in HCC patients in the Yap1–(+) groupcompared with Yap1–(–) HCC patients (Fig. 6C, P = 0.000).Furthermore, serum AFP levels were dramatically higher in theYap1–(+) group than in the Yap1–(–) group (Fig. 6D, P =0.004). Yap1 hyperexpression was significantly associated withmore aggressive phenotypes of HCC, including tumor multi-plicity, vascular invasion, and neoplasm staging (Table S2). In23.75% of HCCs, menin and Yap1 were collectively up-regu-lated (Fig. 6E, P = 0.014). In HCC tissue ChIP assays, thebinding of menin and accumulation of H3K4me3 at the Yap1promoter were markedly increased in HCC specimens comparedwith adjacent tissues (Fig. 6 F and G and Fig. S6). These resultssupport the mechanistic and clinical significance of the menin–Yap1 axis as an effective biomarker for HCC diagnosis andprognosis evaluation.

DiscussionThe findings of this study advance our knowledge of the epige-netic activation mechanism of H3K4me3 and indicate that meninplays an important, yet previously unappreciated, role in pro-moting the development of HCC. Menin, as a scaffold protein,lacks identifiable functional motifs, and the putative function ofmenin is not well defined. Menin interacts with a variety oftranscription factors and is involved in a variety of cellular pro-cesses, including gene activation and repression (7). In this no-tion, the crystal structure shows that menin contains a deeppocket that interacts similarly with MLL1 and JUND, indicating

Fig. 4. Heterozygous loss of Men1 reduces DEN-induced development ofHCC. Men1WT (n = 42) and Men1+/− (n = 23) mice were injected with DEN (25mg/kg, i.p.) at 15 d of age and killed 8 mo after DEN injection. Shown arerepresentative liver images (A) and H&E staining sections (B). (C) Shown aredifferent tumor incidences in male and female mice with DEN exposure. (D–G) The compared results of tumor multiplicity, maximal tumor diameter,tumor–liver, and liver–body weight ratios between Men1WT and Men1+/−

mice. (H–J) Circulating serum levels of AFP, IL-6, and TGF-βwere measured byELISA. Data of D–J are represented as mean ± SEM. (K) Western blot analysisof menin, Yap1, pYap1 (ser127), Ash2L, RbBP5, and WDR5 in tumor (T) andnormal (N) liver tissues of mice. (L) H3K4me3 ChIP assays performed withindicated PPs of mouse Yap1 promoter in female mice liver.

Xu et al. PNAS | October 22, 2013 | vol. 110 | no. 43 | 17483

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

June

3, 2

020

that the diverse functions may be largely attributed to the crucialrole of menin as a core scaffold molecule (7). MLL, an epige-netic regulator, plays a critical role in acute leukemias, but theputative biological function of MLL in the liver is not well de-fined. Supporting our notion, HGF-MET (hepatocyte growthfactor-mesenchymal epithelial transition factor) signals werereported to promote the expression of matrix metallopeptidases(MMPs), and the development of HCC through the ETS2–MLLcomplex mediated H3K4me3 (23). Strikingly, whole-genomesequencing analysis revealed that the MLL family of methyl-transferases for H3K4 was somatically mutated in HCCs (24),which further supports the biological relevance of MLL in HCC.

In this report, we propose that the unique biological role ofmenin in promoting HCC depends on the oncogenic activity ofMLL. Although we have identified a unique tumor-promotingaction of menin that is essential for liver tumorigenesis, signif-icant work remains to identify the functions of MLL in HCC. Itis clear that better definitions of active histone modification inHCC are important steps in the design of improved therapeuticstrategies. HCC is the sixth most prevalent cancer, and systemicchemotherapy has marginal activity and frequent toxic effects(10). MI-2 (menin inhibitor-2), a small molecule inhibitor thatis based on the menin structure, has been shown to effectivelyinhibit leukemia cell proliferation by disrupting the menin–MLL

Fig. 5. Menin–Yap1 axis is responsive to CCl4-induced liver inflammation. The Men1WT C57BL/6 mice were treated with CCl4 (2 mL/kg, i.p.) at6 wk of age and killed 24 and 48 h after CCl4injection, respectively. (A) The expression ofmenin, MLL, and Yap1 in liver was detected byWestern blot. (B) The ChIP assays of liver tissueswere carried out with antibodies either againstmenin or H3K4me3 (n = 3). (C and D) The primaryKCs were isolated from WT C57BL/6 mice, andeach of the three distinct siRNAs specificallytargeting either Men1 (Men1-1, 2, and 3) orYap1 (siYap1-1, 2, and 3) were delivered to KCs.The mRNA expression of Men1, Yap1, and IL-6 inKCs was determined by qRT-PCR (n = 3). (E and F )Either HL-7702 or HepG2 cells were transfectedwith each of the three distinct siYap1s, and themRNA level of Yap1 and IL-6 was determined byqRT-PCR (n = 3). The relative mRNA levels werenormalized to β-actin. (G) The siLuc or siYap1-1was transfected to control or menin-overexpressingHepG2 cells, respectively, and the mRNA level ofYap1 and IL-6 was quantified by qRT-PCR (n = 3).(H) The HL-7702 cells were exposed to IL-6 (50 ng/mL), and the protein levels of menin and Yap1 were detected by Western blot. (I) The HepG2 cells were exposed to IL-6 (50 ng/mL), and the mRNA levelof MEN1 and Yap1 was determined by qRT-PCR at the indicated times.

Fig. 6. Yap1 is epigenetically regulated by meninin human HCCs, and it is correlated with poorprognosis. (A) IHC staining of Yap1 in HCC paraffinsections (n = 80). (Scale bar, 500 μm and 100 μm,respectively.) (B and C) Kaplan–Meier curves foroverall and tumor-free survival in Yap1–(+) and (–)HCC patients (P = 0.000 and P = 0.000, respectively,log rank test). (D) Serum AFP level of HCC patientswith Yap1–(+) (n = 32) or (–) (n = 46). In each panel,the line indicates the median (P = 0.004). (E )Hyperexpression of menin and Yap1 in overlappingHCC patients (P = 0.014, χ2 test). (F and G) ChIPassays using antibodies against menin or H3K4me3in four cases of matched HCC and adjacent tissuefrom the fresh specimens (n = 3). (H) A model forthe menin–MLL complex up-regulates Yap1 tran-scription through H3K4me3 in HCC.

17484 | www.pnas.org/cgi/doi/10.1073/pnas.1312022110 Xu et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

020

complex, highlighting the potential therapeutic application ofinhibitors targeting the menin–MLL complex in human diseases(25). Similarly, considering the importance of the menin–MLLcomplex in controlling the aggressive nature of HCCs, we pro-pose a unique therapeutic strategy for HCC by targeting themenin–MLL complex. Further studies will be required to eval-uate the therapeutic effect of epigenetic-targeting drugs, such asMI-2, on drug-resistant, aggressive HCC.Site-specific histone modifications are major epigenetic mech-

anisms for maintaining stable gene transcription, which is funda-mental to the development of most cancers (26). Here, we presentthe profile of genomic occupancy of menin and two associatedtranscriptionally active histone methylation markers, H3K4me3and H3K79me2, in liver cancer cells. These methylation markersare associated with many liver cancer-related genes, includingYap1. Notably, Hippo-dependent or -independent mechanisms ofYap1 transcription remain to be identified. Recently, Wu et al.demonstrated that the Ets family member GABP binds to the Yappromoter and activates YAP transcription in liver cancer (27). Ourresults indicate that menin binds to Yap1 promoter loci and up-regulates H3K4me3. The menin–MLL complex and Yap1 areconstitutively activated in primary HCC patients and DEN-induced liver cancer, and they serve as critical HCC tumorpromoters. The aberrant expression of the menin–Yap1 axis isstrongly correlated with poor prognosis, suggesting the clinicalsignificance of the menin–Yap1 axis as a biomarker for HCCdiagnosis and prognosis evaluation.Epidemiological surveys indicate that HCC occurs mainly in

men and that men are about three to five times more likely todevelop HCC than women (28). Similar sex disparity is seen inmice given DEN (29). Estrogen suppresses IL-6 production inliver KCs and reduces liver cancer risk in females but not in males(29). Our results suggest that Men1+/− dramatically repressed thedevelopment of HCC in female mice but not in male mice. Wealso observed that menin participates in the liver injury responseof male mice at early stages, and menin activation was morereadily compensated upon liver injury in males than in females. Itis reported that menin expression, which contributes to gesta-tional diabetes in females, is repressed by increased prolactin andprogesterone (30). So we suppose that the expression of meninwas regulated by estrogen only in female mice. This is likely one

of the reasons for sex biases of menin in our liver cancer animalmodel. It will be interesting to determine whether the homozy-gous deletion of Men1 can repress development of HCC in malesand determine how menin expression is regulated by estrogen inthe liver. Conditional knockout mice that are viable and have theMen1 gene inactivated in their livers would allow further assess-ment of menin in liver disease processes, which would advancethe understanding of the precise mechanism by which menincontributes to HCC. Here, we did not find a significant sex bias ofhyperexpression of menin in HCC specimens. However, we can-not rule out the possibility that the result was limited by an in-sufficient number of female samples.Altogether, we propose that augmented menin activation

during chronic liver injury plays a crucial role in promotinghepatocarcinogenesis (Fig. 6H).

Materials and MethodsHuman HCC Samples. The study was approved by the Xiamen UniversityMedical Ethics Committee. Frozen and paraffin-embeddedprimary HCC tissuesand corresponding adjacent nontumorous liver samples were obtained fromthe Chronic Liver Disease Biological Sample Bank, Department of Hep-atobiliary Surgery, ZhongshanHospital XiamenUniversity. These sampleswerefrommale and female patients histopathologically diagnosed as having stageI–IV HCC. The demographic data and clinicopathological features of the HCCpatients are listed in Tables S1 and S3. The ages of the cases ranged from 25to 75 y. In total, we examined 89 HCC samples together with adjacent normaltissues. Sections from paraffin-embedded samples were stained with affinity-purified antimenin or anti-Yap1 (Cell Signaling) antibodies (Table S4) for IHC.The specificity of the antimenin antibody was verified in menin-null andmenin-expressing cells (19). The method and procedure of IHC are as de-scribed previously (19). The expression of menin and Yap1 in livers was de-termined from the IHC results by three independent pathologists.

ACKNOWLEDGMENTS. We thank Dr. Francis Collins at National HumanGenome Research Institute for providing the heterozygous Men1 locus(Men1+/−) mice and Dr. Xianxin Hua for critical reading of the manuscript.This research was supported by grants from the Natural Science Foundationof China (91229111 and 81272719 to G.-H.J., 81101924 to S.-H.L., and81101763 to S.-B.G.), the Natural Science Foundation of Fujian Province(2011J06016 to G.-H.J.), and the Natural Science Foundation of Xiamen(3502Z20104001 to G.-H.J.).

1. Chandrasekharappa SC, et al. (1997) Positional cloning of the gene for multiple en-docrine neoplasia-type 1. Science 276(5311):404–407.

2. Crabtree JS, et al. (2001) A mouse model of multiple endocrine neoplasia, type 1,develops multiple endocrine tumors. Proc Natl Acad Sci USA 98(3):1118–1123.

3. Lemos MC, Thakker RV (2008) Multiple endocrine neoplasia type 1 (MEN1): Analysis of1336 mutations reported in the first decade following identification of the gene. HumMutat 29(1):22–32.

4. Agarwal SK, et al. (1999) Menin interacts with the AP1 transcription factor JunD andrepresses JunD-activated transcription. Cell 96(1):143–152.

5. Hughes CM, et al. (2004) Menin associates with a trithorax family histone methyl-transferase complex and with the hoxc8 locus. Mol Cell 13(4):587–597.

6. Yokoyama A, et al. (2005) The menin tumor suppressor protein is an essential onco-genic cofactor for MLL-associated leukemogenesis. Cell 123(2):207–218.

7. Huang J, et al. (2012) The same pocket in menin binds both MLL and JUND but hasopposite effects on transcription. Nature 482(7386):542–546.

8. Bertolino P, et al. (2003) Genetic ablation of the tumor suppressor menin causes le-thality at mid-gestation with defects in multiple organs. Mech Dev 120(5):549–560.

9. Zindy PJ, et al. (2006) Upregulation of the tumor suppressor gene menin in hepato-cellular carcinomas and its significance in fibrogenesis. Hepatology 44(5):1296–1307.

10. Forner A, Llovet JM, Bruix J (2012) Hepatocellular carcinoma. Lancet 379(9822):1245–1255.

11. Zhou D, et al. (2009) Mst1 and Mst2 maintain hepatocyte quiescence and suppresshepatocellular carcinoma development through inactivation of the Yap1 oncogene.Cancer Cell 16(5):425–438.

12. Hatziapostolou M, et al. (2011) An HNF4α-miRNA inflammatory feedback circuitregulates hepatocellular oncogenesis. Cell 147(6):1233–1247.

13. He G, et al. (2010) Hepatocyte IKKbeta/NF-kappaB inhibits tumor promotion andprogression by preventing oxidative stress-driven STAT3 activation. Cancer Cell 17(3):286–297.

14. Bard-Chapeau EA, et al. (2011) Ptpn11/Shp2 acts as a tumor suppressor in hepato-cellular carcinogenesis. Cancer Cell 19(5):629–639.

15. Zhao B, Li L, Lei Q, Guan KL (2010) The Hippo-YAP pathway in organ size control andtumorigenesis: An updated version. Genes Dev 24(9):862–874.

16. Zender L, et al. (2006) Identification and validation of oncogenes in liver cancer usingan integrative oncogenomic approach. Cell 125(7):1253–1267.

17. Wu Y, et al. (2012) Interplay between menin and K-Ras in regulating lung adeno-carcinoma. J Biol Chem 287(47):40003–40011.

18. Cillo C, et al. (2011) The HOX gene network in hepatocellular carcinoma. Int J Cancer129(11):2577–2587.

19. Gao SB, et al. (2009) Suppression of lung adenocarcinoma through menin and poly-comb gene-mediated repression of growth factor pleiotrophin. Oncogene 28(46):4095–4104.

20. Gao SB, et al. (2011) Menin represses malignant phenotypes of melanoma throughregulating multiple pathways. J Cell Mol Med 15(11):2353–2363.

21. Maeda S, Kamata H, Luo JL, Leffert H, Karin M (2005) IKKbeta couples hepatocytedeath to cytokine-driven compensatory proliferation that promotes chemical hep-atocarcinogenesis. Cell 121(7):977–990.

22. Iredale JP (2007) Models of liver fibrosis: Exploring the dynamic nature of in-flammation and repair in a solid organ. J Clin Invest 117(3):539–548.

23. Takeda S, et al. (2013) HGF-MET signals via the MLL-ETS2 complex in hepatocellularcarcinoma. J Clin Invest 123(7):3154–3165.

24. Fujimoto A, et al. (2012) Whole-genome sequencing of liver cancers identifies etio-logical influences on mutation patterns and recurrent mutations in chromatin regu-lators. Nat Genet 44(7):760–764.

25. Grembecka J, et al. (2012) Menin-MLL inhibitors reverse oncogenic activity of MLLfusion proteins in leukemia. Nat Chem Biol 8(3):277–284.

26. Esteller M (2008) Epigenetics in cancer. N Engl J Med 358(11):1148–1159.27. Wu H, et al. (2013) The Ets transcription factor GABP is a component of the hippo

pathway essential for growth and antioxidant defense. Cell Rep 3(5):1663–1677.28. El-Serag HB, Rudolph KL (2007) Hepatocellular carcinoma: Epidemiology and molec-

ular carcinogenesis. Gastroenterology 132(7):2557–2576.29. Naugler WE, et al. (2007) Gender disparity in liver cancer due to sex differences in

MyD88-dependent IL-6 production. Science 317(5834):121–124.30. Karnik SK, et al. (2007) Menin controls growth of pancreatic beta-cells in pregnant

mice and promotes gestational diabetes mellitus. Science 318(5851):806–809.

Xu et al. PNAS | October 22, 2013 | vol. 110 | no. 43 | 17485

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

June

3, 2

020