pigenetic Inhibition of Nuclear Receptor Small Heterodimer Partner Is ssociated With and Regulates Hepatocellular Carcinoma Growth

B IL IA R Y TR A C T ackground & Aims: Aberrant hypermethylation of romoter regions in cytosine-guanine dinucleotides CpG) islands has been shown to be associated with ranscriptional silencing of tumor-suppressor genes n many cancers. This study evaluated the methyltion profile and the tumor-suppressive function of he small heterodimer partner (SHP, NR0B2) in the evelopment of human hepatocellular carcinoma HCC). Methods: Human HCC pathologic specimens nd cell lines were used as model systems in this tudy. Results: The expression of SHP is diminished n HCC pathologic specimens and cell lines by epigeetic silencing owing to SHP promoter hypermethyltion. In vitro methylation decreased SHP promoter ransactivation and nuclear receptor LRH-1 binding, n event that was reversed by demethylation. Overexression of SHP inhibited HCC foci formation, arested HCC tumor growth in xenografted nude mice, nd increased the sensitivity of HCC cells to apoptotic timuli. Further analysis of a total of 19 normal liver nd 57 HCC specimens showed that down-regulation of HP gene expression may be a common denominator of CC. Conclusions: We propose that SHP functions as novel tumor suppressor in the development of HCC. hese findings provide new insight into the molecular echanisms leading to this common cancer and may ave both diagnostic and therapeutic applications.

Background & Aims: Aberrant hypermethylation of promoter regions in cytosine-guanine dinucleotides (CpG) islands has been shown to be associated with transcriptional silencing of tumor-suppressor genes in many cancers.This study evaluated the methylation profile and the tumor-suppressive function of the small heterodimer partner (SHP, NR0B2) in the development of human hepatocellular carcinoma (HCC).Methods: Human HCC pathologic specimens and cell lines were used as model systems in this study.Results: The expression of SHP is diminished in HCC pathologic specimens and cell lines by epigenetic silencing owing to SHP promoter hypermethylation.In vitro methylation decreased SHP promoter transactivation and nuclear receptor LRH-1 binding, an event that was reversed by demethylation.Overexpression of SHP inhibited HCC foci formation, arrested HCC tumor growth in xenografted nude mice, and increased the sensitivity of HCC cells to apoptotic stimuli.Further analysis of a total of 19 normal liver and 57 HCC specimens showed that down-regulation of SHP gene expression may be a common denominator of HCC.Conclusions: We propose that SHP functions as a novel tumor suppressor in the development of HCC.These findings provide new insight into the molecular mechanisms leading to this common cancer and may have both diagnostic and therapeutic applications.
N uclear receptors have been implicated in a wide variety of cancers, including the initiation and progression of prostate cancer, 1 female reproductive and breast cancers, 2 leukemias, 3 colon cancer, 4 bone cancer, 5 hepatoma, 6,7 and lung cancer. 8Co-activators of nuclear receptors also are involved in the development of cancers, as illustrated by translocations of chromosome 8 in several leukemias that create a fusion protein including the co-activator Tif2. 9Ligand-independent transcription activity for several nuclear receptors may contribute to their complex roles in cancers. 8Because of their important physiologic roles in cancers, nuclear receptors and the critical genes that they regulate are emerging targets for molecular diagnostic tests and cancer therapeutics. 10,11e nuclear receptor small heterodimer partner (SHP) is a critical regulator of several metabolic diseases. 12,13Our recent study showed that SHP functions in cellular growth and survival and that SHP-deletion produced a transformed phenotype of mouse embryonic fibroblasts and was associated with spontaneous hepatoma formation in SHP-knockouts (Zhang et al,  unpublished data).These observations raised the question that SHP might play a role in the development of human hepatocellular carcinoma (HCC).In this report we analyzed expression of the SHP gene and its methylation status in human HCC specimens, characterized the epigenetic mechanism that resulted in SHP gene silencing in HCC, and examined the tumor-suppressive function of SHP in HCC.

Bisulfite Genomic Sequencing Analysis
Briefly, genomic DNA was isolated from normal liver, HCC specimens, or HCC cell lines, EcoRI-digested, denatured in a 0.3-mol/L NaOH solution, and incubated in 2.5 mol/L sodium bisulfite/20 mmol/L hydroquinone at 55°C for 16 hours.The DNA was desalted, desulfonated in 0.3 mol/L NaOH at 37°C for 15 minutes, and then neutralized in 10 mol/L ammonium acetate.The samples were precipitated and resuspended in Tris-ethylenediaminetetraacetic acid.Bisulfite-modified DNA was used for nested polymerase chain reaction (PCR) using primers specific for the modified DNA.Two sets of nested primers were designed with F1/R1 amplifying the promoter region (covering 6 orphaned CpGs) and F2/R2 amplifying exon 1 (covering all of the CpGs in exon 1) (Figure 1B).First, modified DNA was subjected to PCR using F1/R1 or F2/R2 primers.Two microliters of the resultant PCR products were subjected to 35 cycles of PCR under the same conditions using the F1=/R1= or F2=/R2= primers.PCR products were subcloned into pGEMT-Easy vector (Promega, Madison, WI).Four clones from each DNA template were sequenced to determine the methylation status, which is expressed as a percentage of methylated cytosines per total CpG dinucleotides (methylation [%] ϭ methylated cytosines/ total cytosines in CpG dinucleotides ϫ 100).Bisulfite profiles for the DNA sequences assayed can be found in supplementary Table 1 (see supplementary material online at www.gastrojournal.org).The primer sequences can be obtained from supplementary Table 2 (see supplementary material online at www.gastrojournal.org).

In Vitro Methylation of DNAs, Demethylation, and Histone Deacetylation
Reporter constructs (human SHP promoter [hSHP]) were methylated in vitro with 1 U of SssI methylase (New England Biolabs, Ipswich, MA) for each microgram of DNA in the presence of 0.16 mmol/L Sadenosylmethionine at 37°C for 3 hours.For demethylation analysis, HepG2 cells were cultured with medium containing 10% fetal bovine serum, and treated with different doses (1, 3, and 5 mol/L) of the demethylating agent 5=-aza-2=-deoxycytidine (Aza; Sigma, St. Louis, MO; A3656) at different times (1, 2, and 3 days).Cells were fed fresh medium containing freshly prepared Aza every day.For examining changes of SHP messenger RNA (mRNA) by histone deacetylation, HepG2 cells were treated with histone deacetylase inhibitor trichostatin A (TSA; Sigma, T8552) at different concentrations (0.1, 0.5, and 1 mol/L) for 16 hours.For examining the synergistic effect of Aza (1 mol/L) and TSA (0.5 mol/L) on SHP expression, HepG2 cells were treated with Aza for 2 days.TSA was applied alone or in combination with Aza for the last 16 hours of culture.Vehicle (0.005% acetic acid) was added to untreated cells.SHP mRNA was examined by Northern blot analysis.

Cell Culture, Transient Transfection, and Chromatin Immunoprecipitation Assay
Detailed methods of this section (ie, cell culture, transfection, ChIP assay) can be found in our recent publication. 14The human SHP promoter luciferase construct (hSHPp Luc) was engineered in our laboratory.The mutant liver receptor homolog-1 (LRH-1) (DNA binding domain, [DBD] mutated) plasmid was obtained from Dr Timothy Osborne (University of California, Irvine, CA).For luciferase assays, HepG2, Huh7, and Hep3B cells were transfected with Fugene-6 (Roche, Indianapolis, IN).pGL3 or hSHPp Luc (0.3 g) were treated without or with SssI (1 U), transfected together with 50 ng of pcDNA3, LRH-1, or mutant LRH-1 expression vector.For 5-Aza-2=-deoxycytidine (Aza)-treated groups, 2 mol/L of final concentration of Aza was added into the Dulbecco's modified Eagle medium media 2 days after transfection and cultured for 24 hours.Luciferase activities were measured and normalized against ␤-galactosidase activity (Promega).For chromatin immunoprecipitation (ChIP) assays, hSHPp Luc (2 g) was treated without or with SssI and then transfected into Huh7 cells (10 cm) in the presence or absence of Aza (2 mol/L), and cells were chromatin cross-linked, immunoprecipitated with anti-LRH-1 antibodies (20 g), or anti-MeCP2, MBD1, MBD2, HDAC, or sin 3A antibodies, with rabbit normal immunoglobulin G as a negative control (Upstate).Ethanol-extracted DNA was used as the template to amplify the SHP promoter.The primer sequences are listed in supplementary Table 3 (see supplementary material online at www.gastrojournal.org).

Adenovirus Infection
Detailed methods of adenovirus infection were described in our recent publication. 14Briefly, HepG2 or Huh7 cells were cultured under standard conditions.Both the green fluorescent protein (GFP) control adenoviruses and the GFP-SHP adenoviruses were provided by Dr Kazuhiro Oka.HepG2 or Huh7 cells were plated at 2 ϫ 10 6 per 10-cm dish, and infected the next day with viral supernatant for 2 hours (multiplicity of infection, 10).The cells were cultured for an additional 2 days and then used in different experiments as indicated.

Apoptosis Analysis by Flow Cytometry
To induce apoptosis by starvation, Huh7 cells were starved for 72 hours with 0.1% serum, then released into the cell cycle by adding 10% fetal bovine serum containing medium.To examine tumor necrosis factor ␣ (TNF␣)-induced apoptosis, Huh7 cells were treated with 50 mol/L cycloheximide plus 50 ng/mL TNF␣ at the indicated time, or treated with 1 mol/L of synthetic retinoid 6-[3-(1-adamantyl)-4-hydroxyphenyl]-2-naphthalenecarboxylic acid (AHPN) for 24 hours 15 and examined for apoptotic cell death by Annexin V staining.Detached and adherent cells were harvested, then stained either by propidium iodide or by Alexa Fluor 488 annexin V.For propidium iodide staining, cells were fixed with 100% methanol, then centrifuged, resuspended in 50 L of RNase (1 mg/ mL), and incubated for 30 minutes at room temperature, followed by the addition of 1 mL of 0.1 mg/mL of propidium iodide.Cells were analyzed for DNA content by flow cytometry (Becton Dickinson, Franklin Lakes, NJ).Annexin V staining was performed using Vybrant Apoptosis Assay kit #2 (Molecular Probes, Invitrogen, Carlsbad, CA), by measuring the fluorescence emission at 530 and greater than 575 nm.

Foci Formation
HepG2 cells (1 ϫ 10 5 ) were infected with GFPadenovirus (Ad) or SHP-Ad for 2 hours (multiplicity of infection, 10) and grew for 2 days, suspended with 8 mL of 0.4% top agar and 1ϫ Dulbecco's modified Eagle   medium before being poured onto 10-cm tissue culture dishes coated with 10 mL of 0.7% bottom agar.The plates were prepared in triplicate.Fourteen days later, 3 areas per plate were chosen randomly and the number of visible colonies was counted, and the size of the colonies was measured.Green fluorescence was used as a marker to determine whether the cells were infected with GFP or SHP.Only foci formed from GFPor SHP-infected cells were used for statistical analysis.

Tumorigenicity Assay
Exponentially growing Huh7 cells (2 ϫ 10 6 ) were transduced with GFP-Ad or SHP-Ad for 2 days (multiplicity of infection, 10), then injected subcutaneously into the scapular region of 6-week-old anesthetized nude mice (Swiss nu/nu, Harlan, Indianapolis, IN) at 2 injection sites per mouse.In another group, when tumors grew for 10 days, an intratumoral injection of 10 9 pfu SHP-Ad virus or GFP-Ad control virus was performed as previously described. 16Tumor growth was monitored every 3 days by the external measurement in 2 dimensions with calipers as previously described.Tumor volume was determined according to the equation: V ϭ (L ϫ W 2 ) ϫ 0.52, where V is volume, L is length, and W is width (supplementary Table 4; see supplementary material online at www.gastrojournal.org).At 15 days after virus injection, tumors were harvested.

Real-Time Quantitative PCR Analysis
Total RNA extraction from HCC specimens was performed using the RNAeasy Mini Kit (Qiagen).Genomic DNA contaminants were removed by incubating the RNA solution with DNase I (RNase free) at 37°C for 20 minutes.Reverse transcription was performed with the SuperScript II First-Strand Synthesis System for reverse-transcription PCR (Invitrogen).First, the following RNA/primer mixture was prepared in separate tubes: 5 g RNA, 1 L oligo(dT) 18 (0.5 g/L), and 1 L 10 mmol/L deoxynucleoside triphosphate mix.Volumes were adjusted with Diethypyrocarbonate (DEPC)-treated H 2 O to a final volume of 10 L. Next, reactions were incubated at 65°C for 5 minutes and placed on ice for at least 1 minute.To each reaction, 5ϫ RT Buffer (4 L), 0.1 mol/L DTT (2 L), and RNaseOUT (1 L), were added and then incubated at room temperature for 2 minutes.Finally, 200 units of SuperScript II RT was added to each tube, incubated at 42°C for 50 minutes and heat inactivated at 70°C for 10 minutes.Products were stored at Ϫ20°C until use in real-time PCR experiments.

Statistical Analysis
Data are expressed as mean Ϯ SD.Statistical analyses were performed using the Student unpaired t test or 1-way analysis of variance by SPSS (Chicago, IL).A P value of less than .05was considered statistically significant.

Epigenetic Repression of the SHP Gene in HCC Specimens
To determine if diminished SHP mRNA expression occurs in HCC, we determined the SHP expression profile in 10 pairs of normal human liver and HCC pathologic specimens by Northern blot.The analysis revealed that the expression of SHP mRNA was downregulated in 9 of 10 HCC specimens as compared with the normal surrounding liver tissue (Figure 1A, left).We further confirmed a correlated down-regulation of SHP protein levels in HCC specimens using immunohistochemistry analysis (Figure 1A, right, and supplementary Figure 1; see supplementary material online at www. gastrojournal.org).Thus, diminished SHP expression appeared to be a common event in HCC.The data suggest that suppression of SHP expression may be associated with HCC development or progression.
CpG islands, clusters of CpGs, most often exist in promoters and extend downstream into transcribed regions (pro-CpG islands). 17CpG islands also can occur toward the 3= ends of genes 18 or exist within exons (exonic CpG islands).Exonic CpG islands generally are shorter, with lower observed-to-expected CpG ratios and a smaller percentage of guanine-cytosine dinucleotides (GC) than pro-CpG islands. 19The classic definition of these islands has been for sequences greater than 200 bp in length, with a GC content greater than 50% and a observed-to-expected CpG ratio of 0.6 or greater.The definition of a CpG island has been revised recently with more stringent measures.Nontypical CpG islets or orphaned CpGs also have been reported. 20umor-suppressor genes often are inactivated during tumor progression by mechanisms that include point mutations or deletions within genes, and methylation of promoters.We sequenced the coding region of the SHP gene and results showed no sequence differences between normal liver and HCC within the region analyzed.Analysis of the sequence around the SHP gene promoter and the entire SHP gene including 2 exons and 1 intron revealed the characteristics of an exonic CpG island, with an observed/expected CpG ratio of 0.6 and GC content of 59%.This CpG island is located 48 bp downstream of the transcription initiation site of the SHP gene (supplementary Figure 2A and B; see supplementary material online at www.gastrojournal.org).The 5=-flanking region of the SHP gene between Ϫ382 and the transcription start site harbors 6 sparsely distributed CpGs, in conjunction with the exonic CpG island.The presence of a CpG island in this location raised the possibility of epigenetic regulation of SHP expression by CpG methylation.
To study the involvement of CpG methylation in the regulation of SHP, the SHP CpG island, as well as the region encompassing the 6 CpG orphans at -47, -83, -203, -238, -263, and -381 in the promoter of SHP gene, were analyzed by bisulfite sequencing using nested PCR (Figure 1B).Because the nature of bisulfite PCR and sequencing is strand-specific, the possibility of a strandspecific bias in DNA methylation was tested by assaying methylation of the complementary (-) DNA strand.The overall methylation level for the positive strand was matched closely to that of the negative DNA strand, suggesting the absence of strand-specific bias in the DNA methylation profile.Thus, sequencing results from mixed positive and negative strands were used for further analysis.As illustrated in Figure 1B, extensive hypermethylation was observed in the SHP gene of the 10 HCC samples at both the conventional CpG island and nonconventional CpG orphans (supplementary Figure 3; see supplementary material online at www.gastrojournal.org).In contrast, the normal surrounding tissue had few methylated CpGs in this region.The overall methylation of the 6 CpG sites was 93.3% (56 of 60) in HCC compared with 6.6% in surrounding tissue (4 of 60) (supplementary Table 1; see supplementary material online at www. gastrojournal.org).In a similar fashion, the CpGs in the SHP exonic CpG island also were hypermethylated (113 of 140; 80%), in contrast to the undermethylated status in surrounding tissue (6 of 140; 4.2%).The percentage of methylation for most individual CpG sites in the SHP gene was about 80%-100% in HCC, whereas it was less than 20% in the surrounding tissue (Figure 1C).The hypermethylation status of SHP generally agrees with the low expression level of SHP in HCC (Figure 1A).The mRNA level of SHP in HCC sample 9 was not down-regulated, which did not correlate well with the highly methylated exonic CpGs.It is possible that the unmethylated CpG sites in SHP promoter play a role.Thus, methylation of CpGs in the SHP promoter appears to be a more critical determinant of SHP mRNA expression.Overall, these data indicate that CpG methylation is involved in SHP gene silencing in HCC.

CpG Hypermethylation of the SHP Gene in HCC Cell Lines
To address the question if methylation of SHP promoter CpGs plays a critical role in silencing SHP in HCC, we analyzed the SHP promoter in 3 human HCC cell lines (HepG2, Huh7, and Hep3B) and revealed a similar hypermethylation profile (Figure 2A).This provided the basis for in vitro functional analysis.To determine if hypermethylation contributes to reduced SHP expression, HepG2 cells were treated with demethylating agent Aza and SHP mRNA was quantified by Northern blot.As expected, Aza treatment induced SHP expression in a time-and dose-dependent manner (Figure 2B, a= and b=).Histone deacetylation is another important epigenetic alteration that leads to modified gene expression. 21o determine if histone deacetylation also contributes to SHP gene silencing in HCC, we treated HepG2 cells with the specific histone deacetylase inhibitor TSA.However, we did not observe changes of SHP mRNA expression (Figure 2B, c=).To further determine any synergistic effects of these 2 repression mechanisms, DNA methylation, and histone deacetylation, we treated HepG2 cells with Aza, TSA, or both.TSA treatment did not result in enhanced re-expression of SHP by Aza (Figure 2B, d=).The data suggest that DNA methylation appears to be the primary mechanism repressing SHP expression in HCC.
To determine directly if methylation of the SHP gene represses SHP expression, we constructed a hSHPp Luc reporter containing approximately 0.8 kb of the upstream region.Transient transfections were performed to assay the effects of DNA methylase SssI and the demethylating agent Aza on the transactivation of SHP by LRH-1.LRH-1 is a well-characterized nuclear receptor that can bind to SHP promoter and activate SHP gene expression. 22Treatment of hSHPp Luc with SssI markedly decreased the luciferase activity of the SHP promoter induced by LRH-1 (Figure 2C, lane 2 vs lane 1, black bar), whereas demethylation by Aza blocked the effect of SssI (lane 4 vs lane 2).In contrast, no effects of SssI and Aza were observed with a DBD mutant LRH-1 (supplementary Figure 4A; see supplementary material online at www.gastrojournal.org).Although the efficiency of SssI and Aza on the transactivation of LRH-1 appears to be variable among each HCC cell line, it is evident that methylation inhibits SHP promoter activity.
There exist 3 LRH-1 binding sites within the 6 CpGs in the SHP promoter (Figure 2D, left, gray bar, and supplementary Figure 2A, pink underline; see supplementary material online at www.gastrojournal.org).ChIP assays were used to determine if SssI diminishes, whereas Aza restores the Co-immunoprecipitation (Co-IP) of LRH-1 on the SHP promoter.As shown in Figure 2D, the PCR product produced by F1/R primers (covering 3 LRH-1 sites and 6 CpGs) was clearly detected from the transfectant using nonmethylated SHP genomic DNA.However, a PCR product from methylated DNA was not observed, indicating that methylation of CpG sites in the SHP gene abolished LRH-1 binding.In contrast, an abundant PCR product was observed in Aza-treated cells.On the other hand, SssI did not completely abolish LRH-1 binding to the SHP promoter when assayed by F2/R primers (covering 1 LRH-1 site and 2 CpGs) and Aza treatment also showed a much weaker reversal effect.These results suggest that all 6 CpG sites may play critical roles for efficient LRH-1 binding and activation of SHP promoter.Further analysis showed an enrichment of methyl-CpG binding proteins MeCP2, MBD1, and the co-repressor sin3A, but not MDB2 and the histone deacetylase HDAC, on SHP chromatin by SssI treatment (Figure 2E).

SHP Inhibition of HCC Tumor Formation
To further address the question of the role of SHP in regulating HCC tumor formation, we used adenovirus-BASIC-LIVER, driven expression of SHP and foci formation assay.SHP was overexpressed using adenovirus in HepG2 cells and this markedly suppressed both large and small foci formation in soft agar as compared with controls (Figure 3A).
To further assess the effect of SHP on tumor development and progression, Huh7 cells were infected with GFP-Ad or SHP-Ad and implanted into the dorsa of nude mice.Tumors became apparent in the GFP-Ad treatment group as early as 10 days after implantation, and smaller tumors were observed from the SHP-Ad-treated group (Figure 3B, left, 4 left-most tumors).We performed a single intratumoral virus injection 10 days after implantation in a second group of mice. 16At day 25, a sustained and significant arrest of tumor growth was observed in the SHP-Ad-injected group, which was in sharp contrast to the larger tumors present in animals infected with the GFP-Ad (Figure 3B, left, 6 right-most tumors).The data are consistent with the decreased tumor volume (right panel) and a reduced vascular system in SHP transduced cells (supplementary Figure 4B and C; see supplementary material online at www.gastrojournal.org).
Recently we observed that SHP activates apoptotic signaling in mouse hepatocytes (unpublished data).To examine the SHP involvement of apoptosis signaling in human HCC, we tested the sensitivity of Huh7 cells to 2 inducers of TNF␣ and AHPN, a newly identified potential ligand of SHP. 15 The cells that overexpressed SHP displayed significantly increased basal, TNF␣, AHPN, and starvationinduced apoptosis (Figure 3C), consistent with our finding in mice.The data suggest that SHP is a potent suppressor of HCC tumor formation, at least in part through regulating apoptotic signaling.

Down-Regulation of SHP Expression as a Common Denominator in HCC
To determine if down-regulation of SHP gene expression is a common denominator in human HCC, we collected a relatively large number of HCC specimens and analyzed SHP mRNA levels using real-time PCR.This included 19 matched pairs of HCC with surrounding liver tissue that was not invaded histologically by HCC plus 38 individual HCC pathologic specimens.Among the HCC specimens, 38 were from patients with chronic hepatitis C, 9 were from patients with chronic hepatitis B, and 10 were from patients with other chronic liver diseases, including nonalcoholic steatohepatitis, cryptogenic cirrhosis, and hemochromatosis.In the matched pairs of HCC and adjacent tissue, SHP levels were downregulated in 18 of 19 HCC specimens as compared with adjacent tissue (normal liver vs HCC) (Figure 4A).When comparing the expression of SHP in each normal adjacent tissue with HCC specimens, the overall mRNA levels of SHP were decreased significantly in HCC (P Ͻ .001)(Figure 4B).To determine if the degree of SHP downregulation was dependent on the underlying disease resulting in HCC, we divided data into groups with chronic hepatitis C, hepatitis B, and other causes.The expression of SHP was decreased to a similar extent in each HCC group regardless of underlying disease present in patients (Figure 4C).These data further support our hypothesis that SHP gene silencing is a common denominator during the development of HCC in both animal models and human pathologic specimens.

Discussion
HCC often is diagnosed at an advanced stage when therapies have a limited efficacy. 23The fact that HCC is resistant to conventional chemotherapy leaves advanced disease with few therapeutic options and dismal prognosis.The challenges with early diagnosis of HCC emphasize the need for better biomarkers for early

BASIC-LIVER,
stage HCC.HCC tumors are phenotypically and genetically heterogeneous tumors and usually emerge in the presence of chronic liver disease, including hepatitis B or C, and metabolic liver diseases. 24Recent insights into the biology of HCC suggest that certain molecular alterations are likely to play essential roles in HCC by promoting hepatocyte proliferation, migration, and survival. 25The malignant transformation of hepatocytes is a multistep process associated with preneoplastic changes in gene expression, resulting from altered DNA methylation, the actions of hepatitis B and C viruses, and point mutations or loss of heterozygosity.These observations suggest that there are multiple, perhaps redundant, negative growth regulatory pathways that protect cells against transformation.The identification of such mechanisms may open new avenues for the early diagnosis, prevention, and treatment of HCC.
Aberrations in the DNA methylation patterns are recognized as a hallmark of human cancer.The hypermethylation of promoter regions is a well-categorized epigenetic change that occurs in virtually every type of human neoplasm including HCC and is associated with the inappropriate transcriptional silencing of genes. 26The target genes are distributed in all cellular pathways (apoptosis, DNA repair, cell cycle, and cell adherence), which include classic tumor-suppressor genes and putative new tumor-suppressor genes. 27,28In this study, we identified the nuclear receptor SHP as a novel tumor suppressor in HCC, and showed that CpG hypermethylation is an important mechanism in SHP gene silencing and a critical epigenetic event in liver cancer development.Methylation occurred not only in the exonic CpG island of the SHP gene, but occurred more frequently in the 6 orphaned CpGs of the SHP proximal promoter in both HCC spec- imens and HCC cells.It appeared that these 6 CpGs were critical for maintaining normal SHP promoter activity because methylation of them in vitro blocked the binding of LRH-1 and transactivation of SHP by recruiting methyl-CpG binding proteins and co-repressors.It is presumed that methylation of the 6 CpGs plays a major role in silencing SHP during HCC progression.In addition, 4 additional CpG sites were observed in the upstream region of the SHP promoter.It would be interesting to determine their methylation profile in HCC in future studies.Although the SHP gene promoter has been well characterized to be regulated by several nuclear receptors, essentially all those studies were performed in nontransformed cell systems. 29Our study elucidates an epigenetic mechanism that controls SHP gene function under disease conditions that are associated with an increased risk for the formation of HCC.The hypermethylation profile of the SHP gene promoter was established in 10 matched pairs of HCC specimens.These results were consistent with the decreased SHP mRNA found in an additional 57 HCC specimens as compared with normal liver tissues.These results provide convincing evidence that epigenetic inhibition of the SHP gene function is a critical component in the progression of HCC.A correlation between the stage of disease and the initiation of SHP hypermethylation would be established in future studies.
Consistent with the observation of SHP gene silencing in HCC, overexpression of SHP showed potent activity in inhibiting HCC tumor formation in vitro and arresting HCC tumor growth in vivo.Several potential mechanisms exist that may explain the tumor-suppressor function of SHP, including our observation of SHP inhibiting HCC cell proliferation and activation of HCC cell apoptosis.Studies currently are under way to explore the detailed molecular basis for these effects of SHP.
In conclusion, our study identifies the SHP gene as a potent tumor suppressor and provides several lines of evidence for SHP gene silencing playing a major role in the development of HCC.A tumor-specific hypermethylation profile of the SHP gene CpG island that was identified in HCC may allow these aberrantly hypermethylated loci to be used as a new diagnostic marker in the care of patients with liver masses.

Figure 1 .
Figure 1.SHP promoter hypermethylation in HCC.(A) Left, Northern blot of 10 matched pairs of normal surrounding tissues (S) and HCC specimens (H) probed with hSHP cDNA.Right, immunohistochemistry analysis of hSHP protein in HCC.Blue, DAPI; red, SHP staining.(B) SHP promoter hypermethylation profiling in HCC.The top diagram shows the SHP gene and the location of nested primers used for bisulfite genomic sequencing.Exon 1 is shown as a dashed line, the transcription start site is shown as an arrow, and CpG sites are shown as tick marks.F1/R1 are used to amplify the promoter region containing 6 orphaned CpGs, whereas the F2/R2 are used to amplify the entire region of exon 1.All 20 CpG sites in the predicted CpG island region of the SHP gene were analyzed by bisulfite genomic sequencing for their methylation status.White and black circles denote unmethylated and methylated CpG sites, respectively.Four cloned PCR products were sequenced to determine the percentage of methylation of the CpG sites in the regions analyzed.(C) Summary of 5-methycytosine levels for each CpG site obtained by bisulfite genomic sequencing of the SHP gene in 10 matched pairs of normal and HCC specimens.

Figure 2 .
Figure 2. Correlation of methylation in the promoter region with silencing of the SHP gene in HCC cells.(A) Methylation status of the 6 CpGs of the SHP gene promoter in HCC cell lines as determined by bisulfite sequencing.Four different clones from each line are presented and methylated CpG sites are represented by black circles.(B) Northern blot analysis of SHP mRNA after treatment of HepG2 cells with demethylating agent Aza or histone deacetylase inhibitor TSA.(C) The hSHP reporter (hSHPp Luc) was methylated in vitro with 1 U of SssI methylase, transfected with the human LRH-1 (100 ng) expression vector in Aza (2 mol/L) untreated or treated cells.Luciferase activity was determined and normalized by ␤-gal activity.For the Aza group, cells were treated with Aza 1 day before the transfection, and were fed fresh medium containing Aza every day.x, -LRH-1; y, ϩLRH-1.(D) ChIP assays.The hSHPp Luc was methylated by SssI, and Huh7 cells were transfected with the hSHPp Luc (nonmethylated or methylated) in the absence or presence of Aza.N.S., nonspecific primers located 4 k upstream of the SHP promoter.Input serves as control.Left, positions of primers used for ChIP analysis.Gray bars indicate 3 LRH-1 binding sites within the 6 CpG sites.(E) ChIP analysis of enrichment of MeCP2, MBD1, and sin3A on SHP chromatin by SssI treatment.The results are expressed as the percentage of immunoprecipitate (IP) over total input DNA used.x, Con; y, SssI.

Figure 3 .
Figure 3. SHP inhibition of tumor formation.(A) HepG2 cells were infected with either GFP-control virus (GFP-Ad) or SHP-adenovirus (SHP-Ad), and foci formation was quantitated in soft agar.Green fluorescence was used to identify GFP-Ad-or SHP-Ad-infected cells.Green arrows in the SHP-Ad group represented foci formed from cells not infected with SHP-Ad (no green fluorescence detected from these cells).Only foci formed GFP-(x) or SHP (y)-infected cells were counted and used for statistical analysis (right panel).(B) Huh7 cells (2 ϫ 10 6 ) were infected with either GFP-Ad (top) or SHP-Ad (bottom), grafted subcutaneously in the dorsa of athymic mice, and tumor growth was monitored at day 10 as indicated (left panel).A second group of mice received an intratumoral injection of 10 9 pfu of GFP-Ad or SHP-Ad at 10 days after tumor implantation and were analyzed at day 25.Tumor volume (right panel) was monitored.OE, GFP; , SHP. (C) Huh7 cells were infected with either GFP-Ad or SHP-Ad, then treated with TNF␣ or AHPN, or starved for 72 hours.The number of apoptotic cells was quantified by Annexin V staining.EtOH and dimethyl sulfoxide (DMSO) are solvents for cycloheximide and AHPN, respectively.x, GFP; y, SHP.

Figure 4 .
Figure 4. Real-time PCR analysis of the SHP gene expression in HCC.(A) Total RNA was isolated from a total of 19 pairs of normal surrounding tissue (x, SURR.TISS) and the corresponding HCC (y) and SHP mRNA levels were analyzed by real-time PCR.(B) Real-time PCR results of the mRNA levels of SHP in 19 normal surrounding liver tissues compared with that of 57 HCC pathologic specimens.(C) The expression pattern of SHP in HCC specimens was compared between disease groups including hepatitis C (HEPC) or hepatitis B (HEPB)."Other" represents patients with cryptogenic cirrhosis, hemochromatosis, or other disorders leading to an increased risk for HCC.