Genome-Wide Methylation Analysis and Epigenetic Unmasking Identify Tumor Suppressor Genes in Hepatocellular Carcinoma

GASTROENTEROLOGY 2013;145:1424–1435

Genome-Wide Methylation Analysis and Epigenetic Unmasking Identify Tumor Suppressor Genes in Hepatocellular Carcinoma KATE REVILL,1,2 TIM WANG,1,3 ANJA LACHENMAYER,2,6 KENSUKE KOJIMA,2 ANDREW HARRINGTON,2 JINYU LI,1 YUJIN HOSHIDA,2 JOSEP M. LLOVET,2,4,5,§ and SCOTT POWERS1,§ 1 Cancer Genome Center, Cold Spring Harbor Laboratory, Woodbury, New York; 2Mount Sinai Liver Cancer Program, Division of Liver Diseases, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York; 3Whitehead Institute for Biomedical Research, Massachusetts Institute of Technology, Cambridge, Massachusetts; 4HCC Translational Research Laboratory, Barcelona-Clinic Liver Cancer Group, Institut d’Investigacions Biomediques August Pi I Sunyer (IDIBAPS), CIBERehd, Liver Unit, Hospital Clinic, University of Barcelona, Catalonia, Spain; 5Institucio Catalana de Recerca i Estudis Avancats (ICREA), Catalonia, Spain; and 6Department of General, Visceral, and Pediatric Surgery, University Hospital Düsseldorf, Düsseldorf, Germany

BASIC AND TRANSLATIONAL LIVER

BACKGROUND & AIMS: Epigenetic silencing of tumor suppressor genes contributes to the pathogenesis of hepatocellular carcinoma (HCC). To identify clinically relevant tumor suppressor genes silenced by DNA methylation in HCC, we integrated DNA methylation data from human primary HCC samples with data on up-regulation of gene expression after epigenetic unmasking. METHODS: We performed genome-wide methylation analysis of 71 human HCC samples using the Illumina HumanBeadchip27K array; data were combined with those from microarray analysis of gene re-expression in 4 liver cancer cell lines after their exposure to reagents that reverse DNA methylation (epigenetic unmasking). RESULTS: Based on DNA methylation in primary HCC and gene re-expression in cell lines after epigenetic unmasking, we identified 13 candidate tumor suppressor genes. Subsequent validation led us to focus on functionally characterizing 2 candidates, sphingomyelin phosphodiesterase 3 (SMPD3) and neurofilament, heavy polypeptide (NEFH), which we found to behave as tumor suppressor genes in HCC. Overexpression of SMPD3 and NEFH by stable transfection of inducible constructs into an HCC cell line reduced cell proliferation by 50% and 20%, respectively (SMPD3, P ¼ .003 and NEFH, P ¼ .003). Conversely, knocking down expression of these genes with small hairpin RNA promoted cell invasion and migration in vitro (SMPD3, P ¼ .0001 and NEFH, P ¼ .022), and increased their ability to form tumors after subcutaneous injection or orthotopic transplantation into mice, confirming their role as tumor suppressor genes in HCC. Low levels of SMPD3 were associated with early recurrence of HCC after curative surgery in an independent patient cohort (P ¼ .001; hazard ratio ¼ 3.22; 95% confidence interval: 1.66.5 in multivariate analysis). CONCLUSIONS: Integrative genomic analysis identified SMPD3 and NEFH as tumor suppressor genes in HCC. We provide evidence that SMPD3 is a potent tumor suppressor gene that could affect tumor aggressiveness; a reduced level of SMPD3 is an independent prognostic factor for early recurrence of HCC. Keywords: nSMase2; NEFH; Sphingomyelin Phosphodiesterase; 5-aza-2-deoxycitidine.

A

ge-adjusted incidence rates of hepatocellular carcinoma (HCC) have tripled in the United States from

1975 through 2005,1 and worldwide this cancer is the third leading cause of cancer death.2 Currently, the most effective systemic therapy for patients with HCC is treatment with the tyrosine kinase inhibitor sorafenib, which increases median survival and time to radiologic progression by nearly 3 months.3,4 The mechanisms contributing to hepatocarcinogenesis are unclear, however, it is widely accepted that HCC exhibits numerous genetic abnormalities, such as chromosomal alterations, gene amplifications, and mutations, as well as epigenetic alterations.5,6 A number of locus-specific studies in HCC report promoter DNA methylationassociated silencing of tumor suppressor genes,7–9 and increased DNA methylation levels of tumor suppressor genes have also been reported to correlate positively with HCC development and progression.5,10 Aberrant DNA methylation and other posttranslational histone modifications are frequently found in surrounding tissues5,6 and in patients with chronic hepatitis or cirrhosis, conditions that commonly precede development of this disease.11 In this study, we have performed a systematic and combinatorial approach to identify hypermethylated tumor suppressor genes in HCC. First, we performed arraybased analysis on 27,578 CpG sites in 71 primary HCC samples and 8 nondiseased (ND) normal tissues to identify genes differentially methylated between primary HCC and ND normal liver. Second, we performed microarray analysis of 4 liver cancer cell lines from an epigenetic unmasking experiment to identify genes up-regulated after DNA de-methylation and inhibition of histone deacetylation. Finally, we combined both datasets to provide a list of biologically relevant hypermethylated genes §

Authors share co-senior authorship.

Abbreviations used in this paper: ACTL6B, actin-like 6B; CHTN, cooperative tissue network; CpG, cytosine preceding guanine; DAC, 5-aza20 deoxycitidine; DGKI, diacylglycerol kinase, iota; ELOVL4, ELOVL fatty acid elongase 4; GSTP1, glutathione S-transferase 1; HCC, hepatocellular carcinoma; LDHB, lactate dehydrogenase B; LRAT, lecithin retinol acyltransferase (phosphatidylcholineretinol O-acyltransferase); PRPH, peripherin; ND, nondiseased; NEFH, neurofilament, heavy polypeptide; shRNA, short hairpin RNA; SMPD3, sphingomyelin phosphodiesterase 3; TSA, trichostatin A. © 2013 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2013.08.055

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

for further validation and characterization as tumor suppressor genes and prognostic markers in HCC.

Table 1. Clinical Data of Patients Included in Study HCC Genomic Consortium

Age older than 60 y Male sex Origin Barcelona Milan New York Child-Pugh class A Etiology Hepatitis C virus Hepatitis B virus Alcohol Other Tumor size, cm 2 >2 Degree of differentiationa Well Moderate Poor Satellitesa Absent Present Vascular invasion Absent Present BCLC stage 0 A B C Albumin level <3.5 mg/dL Bilirubin level >1 mg/dL AFP level >100 mg/dL Events Follow-up period, mob Deaths Late recurrence Early recurrence <2 y

Mixed etiologies (n ¼ 164, validation)

56 (71.8) 52 (66.7)

110 (67.1) 113 (68.9)

23 25 29 69

(29.9) (32.1) (37.2) (88.5)

36 58 70 101

(22) (35.4) (42.7) (61.6)

77 (100)

37 39 12 21

(22.6) (23.8) (7.3) (12.8)

10 (12.8) 67 (87.2)

32 (19.5) 119 (72.6)

19 (24.4) 41 (52.6) 16 (20.1)

25 (15.2) 55 (33.5) 16 (10)

25 (29.5) 4 (6.4)

110 (52.4) 30 (11.0)

33 (42.3) 44 (57.1)

96 (58.5) 49 (29.9)

6 59 5 7 12 17 20 52.9 20 10 28

(7.7) (75.6) (6.4) (9) (15.4) (21.8) (25.6) (24-108) (25.6) (12.8) (35.9)

17 122 9 1 7 38 23 85 28 24 39

(10) (74.4) (35.5) (0.6) (4.3) (23.2) (14.0) (60.9-117.2) (17.1) (14.6) (23.8)

NOTE. Data are expressed as median (range, quartile 25quartile 75). AFP, a-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; HCV, hepatitis C virus. a Missing data for degree of differentiation (n ¼ 69) and satellites (n ¼ 72) of 241 cases. b Follow-up period.

For epigenetic unmasking, cells were treated with 500 nm trichostatin A (TSA) (Sigma, St Louis, MO) for 24 hours and 1 mM 5-aza-20 -deoxycitidine (DAC) (Sigma) for 48 hours.14 For tetracycline-induced expression, cells were treated with 5 mg/mL tetracycline (Sigma) for 48 hours and cultured in tetracyclinereduced fetal bovine serum (Invitrogen, Life Sciences Technologies, Santa Clara, CA) containing media.

Methylation Profiling Whole genome methylation analysis was performed using the HumanMethylation27 BeadChip (Illumina, San Diego, CA) on 71 samples from the HCC Genomic Consortium, 24 samples comprising 12 HCC samples with 12 paired adjacent normal from cooperative tissue network (CHTN) and 8 ND normal adult liver samples (Figure 1). Genes were considered to be hypermethylated if the b value for one or more of their corresponding probes showed a fold increase of 1.5 in the mean b of HCC samples above the mean b of normal ND liver, and if this change was statistically significant (P  .05). In addition, we only considered probes passing a minimum cut-off value of 0.3 mean b value of HCC samples (Figure 1). For detailed methods and statistical analysis see Supplementary Material.

Results Patient Characteristics Table 1 summarizes clinical characteristics of the patients analyzed in this study. The HCC Genomic Consortium is split into 2 groups, 1 that is hepatitis Crelated only and 1 with mixed etiology. Patients from the HCC Genomic Consortium were predominantly Barcelona Clinic Liver Cancer stage A and Child-Pugh class A with tumor sizes of >2 cm. Clinical information for samples collected from the CHTN tissue bank were not available and are therefore not listed; however, in contrast to patients within the HCC Genomic Consortium, these patients do not have a background of viral infection. This situation reflects the differences in obtaining samples from these different sites. Although the HCC Genomic Consortium collects samples from clinical sites, Cold Spring Harbor Laboratory utilizes a commercial tissue bank. The differing etiologies and method of collection between both groups allowed us to evaluate how these variables affect the application of our approach to identify novel tumor suppressor genes in a cancer phenotype.

Candidate Tumor Suppressor Gene Discovery A total of 678 genes were identified as being significantly hypermethylated in HCC samples compared with ND normal liver tissues (Figure 1, Supplementary Table 1). Among the list of significantly hypermethylated genes in primary HCC were 37 known tumor suppressor genes, including adenomatous polyposis coli, p16, glutathione S-transferase 1 (GSTP1), retinoblastoma

BASIC AND TRANSLATIONAL LIVER

We analyzed a previously reported genomic dataset of hepatitis Crelated HCC patients with prognostic information (n ¼ 77) as our training cohort, as described previously12 (NCBI Gene Expression Omnibus, accession number GSE9829). For prognostic validation, we analyzed a previously reported genomic dataset of mixed etiology HCC patients from the HCC Genomic Consortium with prognostic information (n ¼ 164), GEO accession GSE19977.13 Clinical demographics are summarized in Table 1. All patients were treated with curative resection. Curativity of resection was confirmed by postsurgical histological and imaging assessment every 3 to 6 months. Tumor recurrence was

Variable

determined based on typical imaging features or histological confirmation.

Cell Culture and Treatment

Materials and Methods Patients

HCV HCC (n ¼ 77, training)

1425

1426 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Figure 1. Study overview. Seventy-one patients from the HCC Genomic Consortium were profiled for methylation status and 4 liver cancer cell lines analyzed for gene reexpression after epigenetic unmasking to identify candidate tumor suppressor genes (TSGs). Candidates were validated for methylation and gene expression in 12 paired HCC samples from CHTN and copy number variation (CNV) in the HCC Genomic Consortium. Functional validation was performed in vitro and in vivo for the strongest candidates and prognostic implications explored. BASIC AND TRANSLATIONAL LIVER

1, and SPINT2, which are known to be silenced by DNA methylation in HCC,15,16 giving us confidence in our methodology. In parallel, we performed an epigenetic unmasking experiment14 on the liver cancer cell lines, SNU-449, HepG2, Hep3B, and JHH2 with DAC treatment, which blocks DNA methylation, and TSA to inhibit histone deacetylation (Figure 1). The optimal treatment conditions were determined by the robust re-expression of GSTP1 and retention of the unmethylated GSTP1 band by combined bisulfite restriction analysis upon treatment with 500 nM DAC for 48 hours and 1 mM TSA for 24 hours, indicating de-methylation of GSTP1 (Supplementary Figure 1). RNA microarray was used to select for genes that were re-expressed after DAC/TSA treatment. We identified genes that were transcriptionally silent during the mock treatment (<1.5 times the median background probe intensity), but were re-expressed after the DAC/TSA combination treatment. A total of 253 genes passed this cut-off (Figure 1, Supplementary Table 2).

As a means to identify biologically relevant candidate tumor suppressor genes silenced by DNA methylation, we filtered the list of genes identified by array analysis as being hypermethylated in primary HCC with genes up-regulated from the epigenetic unmasking procedure. Using this approach, we arrived at a list of the following 13 candidate genes: actin-like 6B (ACTL6B); C19orf30; diacylglycerol kinase (DGKI); DLX1; ELOVL fatty acid elongase 4 (ELOVL4); lactate dehydrogenase B (LDHB); lecithin retinol acyltransferase (phosphatidylcholineretinol O-acyltransferase) (LRAT); MLF1; neurofilament, heavy polypeptide (NEFH); PPM1N; peripherin (PRPH); SLC8A2; and sphingomyelin phosphodiesterase 3 (SMPD3) (Figure 1, Supplementary Table 3). These genes share some ontology by DAVID pathway analysis—SLC8A2, PPM1N, DGKI, and SMPD3 are involved in ion binding; DLX1, MLF1, and SMPD3 are involved in development; and LDHB, LRAT, and SMPD3 are involved in metabolic processes. Enrichr analysis17 reveals that several of these candidates, LRAT, DLX1, NEFH, PRPH, DGKI, ELOVL4, SMPD3, and MLF1

December 2013

are targets of at least 2 members of the PRC2 polycomb repressive complex and DLX1, NEFH, DGKI, ELOVL4, SMPD3, MLF1 are marked by the active histone mark H3K4me1 in the adult liver (Supplementary Table 4). We also wanted to assess the suitability of using a much smaller group of samples for candidate gene selection, as we acknowledge that, in some instances, analysis of large sample numbers is not an option. By performing the same analysis described here on methylation data from samples obtained from CHTN (comprising just 12 HCC samples with 12 paired adjacent liver), we identified 403 significantly hypermethylated genes. When we filter this list using the list of 253 genes up-regulated after epigenetic unmasking, we arrive at the following 10 candidate genes: ACTL6B, DGKI, ELOVL4, PPM1N, G-proteincoupled receptor 63 (GPR63), LDHB, LRAT, NEFH, PRPH, and SMPD3. Nine of which were also identified using the larger dataset described here, validating our approach.

Gene-Specific Validation of Candidate Tumor Suppressor Genes in Primary HCC

1427

candidates that were identified in both independent analyses (ie, SMPD3, NEFH, LDHB, ACTL6B, PRPH, and DGKI) using the fully quantitative method of methylationspecific pyrosequencing (Supplementary Figure 2). We observed significant hypermethylation in tumor samples compared with their adjacent counterparts for DGKI (P ¼ .002), NEFH (P ¼ .016), LDHB (P ¼ .004), ACTL6B (P ¼ .003), and SMPD3 (P ¼ .002), but not PRPH (P ¼ .686) (Figure 2A). For those genes with significant hypermethylation in tumor samples, DGKI, NEFH, LDHB, ACTL6B, and SMPD3, we analyzed the expression status in 12 samples from paired HCC using quantitative real-time polymerase chain reaction. SMPD3 exhibited significantly reduced expression in all 12 tumors when calculating fold change from paired adjacent (P ¼ .001). Nine of 11 tumors that were assessed for methylation were also hypermethylated, correlating significantly with loss of expression (P ¼ 5.95  106, Fisher’s). NEFH exhibited reduced expression in 7 of 12 tumor samples, which was not significant (P ¼ .601); however, 4 of the 7 tumors with reduced expression were also hypermethylated, and none of the samples expressing NEFH were methylated (P ¼ .029, Fisher’s). LDHB exhibited reduced expression in

BASIC AND TRANSLATIONAL LIVER

Using HCC paired with adjacent liver tissue from CHTN, we confirmed the methylation status of 6 of 9

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

Figure 2. Candidate gene methylation validation. (A) Methylation-specific pyrosequencing analysis of 6 candidate genes—SMPD3, NEFH, DGKI, LDHB, PRPH, ACTL6B—in 12 paired normal HCC samples. Normal tissue is indicated by a black box and whisker, tumor tissue is indicated by a red box and whisker. Values are displayed as the percentage of methylation at each CpG site. (B) Transcript expression analysis of 12 primary HCC samples relative to matched normal for SMPD3, NEFH, DGKI, and ACTL6B exhibiting hypermethylation in tumor vs normal by pyrosequencing analysis. Data from quantified quantitative real-time polymerase chain reaction values were normalized to matched normal tissue and log-2 transformed ( SD, *P  .05, **P  .01, ***P  .001).

1428 REVILL ET AL

BASIC AND TRANSLATIONAL LIVER

6 of 12 samples (only 2 of 6 samples with reduced expression were hypermethylated) and ACTL6B in 6 of 12 samples (only 4 of 6 samples with reduced expression were hypermethylated), neither of these were significant for loss of expression (P ¼ .642, P ¼ .282, respectively). By contrast, the expression levels of DGKI were significantly increased in 11 of 12 HCC samples (P ¼ .0002) (Figure 2B). Microarray analysis of samples obtained from the HCC Genomic Consortium revealed significantly reduced expression in tumors compared with ND normal liver for ACTL6B (P ¼ .01), C19orf30 (P ¼ .015), PPM1N (P ¼ .03), LRAT (P ¼ 2.51  1014), SLC8A2 (P ¼ .027), and SMPD3 (P ¼ 5.03  106) (Supplementary Figure 3). We did not observe significant correlation between gene silencing and hypermethylation obtained from array data for these samples. Methylation is not the only mechanism of gene silencing, and DNA methylation is often observed without accompanying gene silencing (reviewed extensively18 and elsewhere). In addition, other mechanisms to silence expression of a tumor suppressor gene should be expected, such as deletion and inactivating mutations.18 On samples obtained from the HCC Genomic Consortium, we performed copy number analysis within a 0.5-Mb window for those genes where promoter methylation was confirmed in our paired samples (Figure 1). We observed copy number loss of SMPD3 and NEFH in 21% and 13% of samples, respectively (Supplementary Figure 4A). The level of copy number loss of SMPD3 and NEFH in HCC vs adjacent cirrhotic tissues when using the same cut-offs is comparable with cyclin-dependent kinase inhibitor 2A, PTEN, and RB1, which are tumor suppressor genes known to be deleted in HCC (Supplementary Figure 4B). Samples exhibiting copy number loss of SMPD3 or NEFH tended to also be hypermethylated (Supplementary Figure 4C). In samples from the HCC Genomic Consortium, we observed copy number loss of ELOVL4 in 26% of patients and significant loss of expression of LRAT (P ¼ 2.51  1014), and copy number loss of LRAT in 19% of patients (Supplementary Figure 4D). Due to assay limitations, we were unable to confirm methylation status of ELOVL4 and LRAT by pyrosequencing.

Overexpression of SMPD3 and NEFH Reduces Cell Proliferation in a Human HCC Cell Line Due to the presence of methylation and loss of expression in at least half of the tumors surveyed, as well as evidence of copy number loss, we next wanted to test whether overexpression of SMPD3 or NEFH leads to an inhibitory effect on cell proliferation. To achieve this, we created constructs, pDEST SMPD3 or pDEST NEFH, using a tetracycline operator plasmid (TetO) that allows for expression of either SMPD3 or NEFH under the influence of an inducible promoter (Supplementary Figure 5A and B). To correct for treatment with the inducible agent tetracycline and the effects of gene overexpression, we also created an inducible construct pDEST

GASTROENTEROLOGY Vol. 145, No. 6

b-galactosidase, to express the gene for bacterial b-galactosidase, which is neutral (Supplementary Figure 5C). The HCC cell line JHH-7 has low or absent SMPD3 protein expression and NEFH messenger RNA expression (Supplementary Figure 5), and was successfully cloned with a tetracycline repressor plasmid (TetR) necessary for repression in the absence of tetracycline to produce TetR JHH-7 cells (Supplementary Figure 5A and B). This TetR cell line was used to introduce pDEST SMPD3 and pDEST NEFH TetO plasmids, to give JHH-7 pDEST SMPD3 and JHH-7 pDEST NEFH. By inducing SMPD3 expression with tetracycline in JHH-7 pDEST SMPD3 cells, we observed that proliferation is significantly reduced by at least 50% at 96 hours post induction in these cells compared with noninduced cells (P ¼ .003) (Figure 3A). We observe a less striking, but significant, effect on cell proliferation when expression of NEFH is induced in JHH-7 pDEST NEFH cells. Cell proliferation is significantly reduced by 20% (P ¼ .034) in cells induced to express NEFH at 96 hours post induction (Figure 3B). We see no significant effect on cell proliferation in cells induced to overexpress the control gene b-galactosidase, arguing against cell toxicity upon tetracycline treatment and any aberrant effects of gene overexpression in these cells (Figure 3C).

Loss of SMPD3 and NEFH Promotes Tumorigenicity We knocked down expression of Smpd3 and Nefh in an immortalized line of mouse embryonic hepatoblasts lacking tumor protein p53 and overexpressing v-myc myelocytomatosis viral oncogene homolog (avian) (PHMI cells) that are not tumorigenic in vivo,19 these cells provide an appropriate sensitized background where a single lesion can trigger tumorigenesis.20,21 Using these cells, we first assessed the tumorigenic effects of short hairpin RNAs (shRNAs) on migration and invasion in vitro and then assessed their ability to promote tumorigenesis after subcutaneous injection and orthotopic liver transplantation via intrasplenic injection into recipient mice. Successful knockdown of Smpd3 by 3 independent shRNAs was confirmed by Western blotting (Supplementary Figure 6A). shRNA 4 provided the highest level of Smpd3 knockdown as evidenced by a barely detectable protein band. As Nefh encodes a protein product of 220 kD and was difficult to detect by Western blotting, successful knockdown of Nefh by 2 independent shRNAs—shRNA 2 and shRNA 5—was confirmed by quantitative real-time polymerase chain reaction (Supplementary Figure 6B). Both short hairpins provided >90% knockdown of Nefh. Because enforced expression of SMPD3 or NEFH in an HCC cell line reduces cell proliferation, we assessed whether knockdown of these proposed tumor suppressor genes can affect other aspects of tumor biology, namely cell migration and invasion, in PHMI cells in vitro before performing in vivo experiments. Compared with cells transfected with the nontargeting shRNA to Renilla

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

1429

luciferase, cell migration was significantly increased in PHMI cells with shRNA knockdown of Nefh (P ¼ .022) or Smpd3 (P ¼ .001) (Figure 4A and B). Cell invasion was dramatically enhanced in cells with knockdown of Smpd3 (P ¼ .005), but not Nefh (Figure 4C and D). This result indicates that loss of Nefh can promote cell migration but not cell invasion. Loss of Smpd3, on the other hand, can promote both cell migration and invasion, suggesting that the loss of Smpd3 might have more pronounced effects on tumor growth and aggressiveness than the loss of Nefh. We then tested if reduced levels of either of these genes could play a tumorigenic role in vivo. A significant increase in tumor volume was observed in nude mice subcutaneously transplanted with PHMI cells with shRNA-mediated knockdown of either Smpd3 or Nefh vs PHMI cells with shRNA directed to the nontarget control (Figure 5). A significant increase in tumor volume was observed at 5 weeks post injection in animals with shRNAmediated Smpd3 knockdown compared with background from control animals (shRNA 1, P ¼ .035; shRNA 4, P ¼ .01; shRNA 6, P ¼ .007) (Figure 5A). By week 7, the knockdown of Smpd3 by shRNA 4 resulted in a mean tumor volume of around 400 mm3, significantly higher than background (P ¼ .002) and the partial knockdown of Smpd3 by shRNAs 1 and 6 in tumors of approximately 200 mm3 were again significantly

higher than background (P ¼ .001 and P ¼ .05, respectively) (Figure 5A). The level of knockdown achieved by each hairpin was proportional to the tumorigenic effect observed in the subcutaneous tumorigenicity assay, indicating specific knockdown of Smpd3 and arguing against off-target effects (Figure 5A). A significant but less pronounced increase in tumor volume was observed at 5 weeks post injection in animals with shRNA-mediated Nefh knockdown (shRNA 2, P ¼ .0006; shRNA 5, P ¼ .0009) and by week 8, the mean tumor volumes of cells with knockdown of Nefh by shRNA 2 and shRNA 5 are approximately 300 mm3 and 200 mm3, respectively, and significantly higher than background (shRNA 2, P ¼ .002; shRNA 5, P ¼.005) (Figure 5B). Again, off-target effects of these shRNAs are unlikely because shRNA 2 and shRNA 5 knock down Nefh to similar levels and the resulting tumors are within 100 mm3 of each other (Figure 5B). These results strongly suggest that Smpd3 and Nefh have tumor-suppressive functions in the liver and that Smpd3 is a more potent tumor suppressor gene than Nefh. When PHMI cells depleted for Smpd3 or Nefh protein expression using the previously described most potently acting shRNAs were transplanted into the livers of recipient mice, tumors developed within 6 weeks (Figure 5C). Microscopic examination of the resultant in situ liver

BASIC AND TRANSLATIONAL LIVER

Figure 3. Overexpression of SMPD3 and NEFH significantly reduces cell proliferation. JHH-7 pDEST SMPD3 (A), JHH-7 pDEST NEFH (B), and JHH-7 pDEST b-galactosidase (LACZ), and (C) human hepatocellular carcinoma cells were induced to express their respective transgenes with 5 mg/ mL of tetracycline. Cell viability was measured using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay at 48, 72, and 96 hours post induction. Assays were performed in triplicate. Cell viability measurements of JHH7 pDEST SMPD3 and JHH7 pDEST NEFH were normalized by those obtained for the control cells JHH7 pDEST LACZ.

1430 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

BASIC AND TRANSLATIONAL LIVER

Figure 4. Depletion of SMPD3 and NEFH significantly increases cell migration and/or invasion. (A) Effect on cell migration of tumor protein p53 (p53)/; Myc mouse hepatoblasts (PHMI cells) with shRNA directed to either Nefh or Smpd3. Panels are, from left to right: migration of cells with nontargeting knockdown (REN shRNA), migration of cells with knockdown of Nefh, and migration of cells with knockdown of Smpd3. (B) Quantification of cell migration for each cell line. Cell number is averaged over 5 fields of view. (C) Effect on cell invasion of p53/; Myc mouse hepatoblasts (PHMI cells) with shRNA directed to either Nefh or Smpd3. Panels are, from left to right: invasion of cells with nontargeting knockdown, invasion of cells with knockdown of Nefh, and invasion of cells with knockdown of Smpd3. (D) Quantification of cell invasion for each cell line. Cell number is averaged over 5 fields of view.

tumors classified them as aggressive, solid hepatocellular carcinomas. The tumors appeared to be composed of a population of undifferentiated cells growing as a sheet without any histological evidence of gland formation or any other structures. The cells were large with a more

basophilic staining cytoplasm than normal liver. We established that the liver tumors arose from the Smpd3and Nefh-depleted PHMI cells, as the tumors were green fluorescent proteinpositive. These tumors were highly proliferative, indicated by strongly positive

immunohistochemical staining for proliferating cell nuclear antigen (Figure 5C).

SMPD3 Expression Is an Independent Prognostic Factor for Early Recurrence To evaluate the potential association of SMPD3 expression and recurrence of HCC, patients from the HCC Genomic Consortium HCV HCC (n ¼ 77) were divided into 2 groups according to levels of high and low SMPD3 expression of microarray data by performing log-rank testing in the interquartile range to determine the optimal cut-off (Supplementary Figure 7 and Supplementary Table 5). Clinical variables of high and low SMPD3 expression groups are listed in Supplementary Table 6. The clinical variables between both groups are similar and differences are not statistically significant. Kaplan-Meier plots (Figure 5D and E) were used to assess association between recurrence and SMPD3 expression. The “SMPD3-low” group showed significantly higher early recurrence compared with “SMPD3-high” group (P ¼ .018, Figure 5D). We were able to validate this finding in an independent cohort from the HCC Genomic Consortium (n ¼ 164; P ¼ .001; Figure 5E). The association of low SMPD3 expression level with early recurrence remained significant in multivariate analysis performed on the validation cohort (hazard ratio ¼ 7.1; 95% confidence interval: 1.66.5; P ¼ .001), even after adjusting with other significant predictors of early recurrence, Barcelona Clinic Liver Cancer stage B (hazard ratio ¼ 3.31; 95% confidence interval: 2.1223.92; P ¼ .002) and presence of satellites (hazard ratio ¼ 3.17; 95% confidence interval: 1.785.64; P ¼ .0001), indicating an independent prognostic implication of SMPD3 expression (Table 2). Early recurrence is generally due to dissemination of primary tumor cells, within the first 2 years of resection, and is associated with tumor aggressiveness.18,22

Discussion Some success has been achieved in treating solid malignancies using a combination of histone deacetylase and DNA methyltransferase inhibitors.23 Recent advances in HCC treatment strategies, such as those proposed by Lachenmayer et al, in which the preclinical efficacy of histone deacetylase inhibition combined with sorafenib therapy is assessed, holds great promise in the treatment and management of this disease.24 Defining hypermethylated biomarkers to identify patients who would most benefit from this type of combined treatment is an important area of research. Recently, the vertical integration of methylation data with both transcriptomic and genomic data was described as a method to identify tumor suppressor gene candidates in HCC. This study by Neumann et al identified 3 candidates (ie, PER3, PROZ, and IGFLAS) in a single cohort of HCC, 2 of which, PER3 and IGFALS, exert tumorsuppressive activities in vitro.25

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

1431

In this study, we combined whole genome methylation data using the HumanBeadchip27K array from primary HCC samples with re-expression data after epigenetic unmasking in a panel of liver cancer cell lines. Although similar methods have previously been used individually14 or combined with other methods25 to identify candidate tumor suppressor genes, to our knowledge, we are the first to use this particular combined approach to successfully identify tumor suppressor genes that can be functionally validated in vivo. With this approach, we identified 13 candidate tumor suppressor genes and, when functionally characterized, 2 candidates (NEFH and SMPD3) behave as tumor suppressor genes in vivo. Although it is feasible that more candidates could be revealed by relaxing our strict criteria, an important factor in this study was to obtain a manageable number of genes for downstream validation and functional characterization efforts, both in vitro and in vivo. In addition, by applying the same criteria of candidate gene selection to an independent set of just 12 HCC samples, we identify several candidates that were selected from the HCC Genomic Consortium cohort (comprising 71 HCC samples), indicating that this method of candidate tumor gene discovery can be applied to smaller sample sizes without greatly compromising the identification of a number of potentially relevant candidate genes. That we observe no overlap with candidates reported by Neumann et al25 most probably reflects our differing approaches to candidate gene discovery. A comparison of SMPD3 and NEFH antiproliferative activity with the growth suppressive activity of PER3 and IGFLAS as identified by Neumann et al can be found in Supplementary Table 7. Within the candidate tumor suppressor genes from our combined analysis, several share biological features, as assessed by DAVID or Enrichr analysis. Interestingly, 8 genes, including SMPD and NEFH are targets of at least 2 members of the PRC2 polycomb repressive complex, indicating that these genes are targeted for epigenetic repression during development (reviewed in Simon and Kingston26). Six genes, including SMPD3 and NEFH, are marked by the active histone mark H3K4me1 in adult liver, indicating that these genes are maintained in the active state in the normal liver, suggesting a functional impairment when these genes undergo a silencing event. Several of our candidates have been reported to be silenced in cancer previously. ELOVL4 is methylated with associated loss of expression in pancreatic cancer27 and LRAT expression is reduced in prostate, renal, and breast cancer.28 We observed striking copy number loss for these genes in a subset of HCC samples. In addition, LRAT exhibited a significant reduction of expression, suggesting that either of these genes may play a role in the pathogenesis of HCC and warrant further study. LDHB is silenced by methylation in gastric29 and prostate cancer30 and decreased expression of LDHB enhances claudin-1mediated hepatoma cell invasiveness, suggesting that the loss of this gene may play an important role in HCC.31 In this study, HCC samples exhibit significant

BASIC AND TRANSLATIONAL LIVER

December 2013

1432 REVILL ET AL

DNA methylation of the CpG island associated with the promoter region of LDHB; however, LDHB expression was not significantly down-regulated in either of the independent sets of samples analyzed in this study. In this report, we present novel evidence that both SMPD3 and NEFH function as tumor suppressor genes. The hypermethylation and silencing of SMPD3 and NEFH have previously been linked to various

GASTROENTEROLOGY Vol. 145, No. 6

malignancies,32,33 but not yet HCC, and have not, as yet, been functionally characterized as tumor suppressor genes in any cancer. Overexpression of NEFH disrupts normal cell structure and function34 and its loss is proposed to activate the Akt/ b-catenin pathway and increase glycolysis.33 We demonstrate that NEFH is significantly hypermethylated in a subset of HCC; we also show evidence of copy number loss and reduced

BASIC AND TRANSLATIONAL LIVER

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

1433

Table 2. Univariate and Multivariate Analysis of Clinical Variables and SMPD3 Expression for Early Recurrence Univariate analysis

Multivariate analysis

Clinical variables

n

P value

HR

95% CI

Origin Male sex Age older than 60 y Etiology, hepatitis C Size >2 cm Albumin level <3.5 g/L Bilirubin level >1 mg/dL AFP level >100 mg/dL Vascular invasion Satellites BCLC stage A BCLC stage B Child Pugh A Degree of differentiation, moderate or poor SMPD3 low expression

157 151 151 146 151 140 146 142 145 140 149 149 140 95

.76 .228 .785 .143 .295 .157 .28 .246 .216 .0000012 .632 .003 .122 .2

1.06 0.71 1.84 2.93 1.63 1.72 1.35 1.42 1.4 3.92 1.25 5.67 0.21 1.4

0.731.53 0.411.24 0.611.93 0.6912.31 0.654.09 0.813.64 0.782.35 0.792.55 0.832.39 2.266.79 0.53.16 1.7918 0.031.53 0.842.35

163

.002

2.7

1.445.09

n

P value

HR

95% Cl

138

.0000876

3.17

1.785.64

138

.00156

7.103

2.1223.92

138

.0011

3.222

1.66.5

expression of NEFH. Overexpression of NEFH reduces cell proliferation and knockdown promotes cell migration in vitro and tumor growth in vivo, implicating NEFH as a bona fide tumor suppressor gene, however, we observe a weaker tumor suppressive role for NEFH than we do for SMPD3. SMPD3 is located at 16q22, an area that is frequently deleted in HCC.35 We also observe copy number loss of SMPD3 in 21% of samples from the HCC Genomic Consortium, suggesting that SMPD3 is a tumor suppressor gene that is inactivated by both deletion and/or promoter hypermethylation in HCC. SMPD3 may also behave as a tumor suppressor gene in other cancers because deletion of Smpd3 has been reported in a mouse Trp53-driven osteosarcoma and missense mutations of SMPD3 have been identified in human leukemia.36 SMPD3 encodes a sphingomyelinase that cleaves sphingomyelin to produce ceramide, which, via the phosphoinositide-3-kinase/AKT pathway, is involved in growth arrest, differentiation, and apoptosis.37,38 We confirmed that loss of SMPD3 impairs apoptosis and increases phosphorylation of Akt at Ser473 (data not shown). We show that SMPD3 is significantly hypermethylated and exhibits down-regulated expression in the majority of primary HCCs. Overexpression of SMPD3 reduces cell proliferation and knockdown promotes both tumor migration and invasion in vitro, indicating that loss

of this gene enhances cell migration and, more importantly, promotes cell invasion. In addition, we demonstrate that knockdown of Smpd3 promotes dramatic tumor growth in vivo. Prognostic analysis indicates that SMPD3 appears to be an independent prognostic factor of early recurrence of HCC, the loss of which can influence the aggressiveness of this cancer. Detecting SMPD3 loss might offer a novel and effective therapeutic strategy for HCC patients who could perhaps be treated with inhibitors of the phosphoinositide-3-kinase/AKT pathway. In summary, we believe that our approach to identify biologically meaningful tumor suppressor genes epigenetically silenced in HCC is a powerful and cost-effective method that can be applied to small sample sizes to yield meaningful results. This work has led to the discovery and characterization of 2 genes in HCC that behave as tumor suppressor genes in vivo and provided multiple candidates for future analysis.

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http:// dx.doi.org/10.1053/j.gastro.2013.08.055.

= Figure 5. Depletion of Smpd3 and Nefh promotes liver carcinoma formation. (A) Subcutaneous growth of tumor protein p53 (p53)/; Myc hepatoblasts infected with either nontargeting shRNA, shRNA 4, shRNA 6, or shRNA 1 to Smpd3 (n ¼ 6 injections, asterisks indicate that the indicated tumor group is significantly different than controls, error bars denote SD, *P < .05; **P < .01; ***P < .001). Tumor volumes were determined from 2 to 7 weeks post injection. (B) Subcutaneous growth of p53/; Myc hepatoblasts infected with either nontargeting shRNA or shRNA2 or shRNA 5 to Nefh (n ¼ 6 injections, asterisks indicate that the indicated tumor group is significantly different than controls, error bars denote SD, *P < .05; **P < .01; ***P < .001). Tumor volumes were determined from 5 to 8 weeks post injection. (C) Images of mouse livers and sections taken 6 weeks after transplantation of p53/; Myc mouse hepatoblasts with knockdown of either Smpd3 or Nefh. Panel columns are, from left to right: intact livers; fluorescent imaging of intact liver for green fluorescent protein (GFP)positive transplanted cells; H&E staining of liver tissue sections showing the border between normal liver and carcinoma (arrows); immunohistochemical detection of GFP; and immunohistochemical detection of proliferating cell nuclear antigen (PCNA). The last 3 are from the same tissue block. Scale bars ¼ 100 mm. (D) Association of SMPD3 expression with time to early and late recurrence after surgery early recurrence in patients from HCC Genomic Consortium. SMPD3 high-expressing patients are indicated in blue and SMPD3 low-expressing patients are indicated in red. (E) Early recurrence in patients from validation cohort. SMPD3 high-expressing patients are indicated in blue and SMPD3 low-expressing patients are indicated in red.

BASIC AND TRANSLATIONAL LIVER

NOTE. Bold type indicates independent significance. AFP, a-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; CI, confidence interval; HR, hazard ratio.

1434 REVILL ET AL

References

BASIC AND TRANSLATIONAL LIVER

1. Altekruse SF, McGlynn KA, Reichman ME. Hepatocellular carcinoma incidence, mortality, and survival trends in the United States from 1975 to 2005. J Clin Oncol 2009;27:1485–1491. 2. Ferlay J, Shin HR, Bray F, et al. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893–2917. 3. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 2008;359:378–390. 4. Villanueva A, Hernandez-Gea V, Llovet JM. Medical therapies for hepatocellular carcinoma: a critical view of the evidence. Nat Rev Gastroenterol Hepatol 2013;10:34–42. 5. Calvisi DF, Ladu S, Gorden A, et al. Mechanistic and prognostic significance of aberrant methylation in the molecular pathogenesis of human hepatocellular carcinoma. J Clin Invest 2007;117:2713–2722. 6. Nishida N, Goel A. Genetic and epigenetic signatures in human hepatocellular carcinoma: a systematic review. Curr Genom 2011; 12:130–137. 7. Matsuda Y, Ichida T, Matsuzawa J, et al. p16(INK4) is inactivated by extensive CpG methylation in human hepatocellular carcinoma. Gastroenterology 1999;116:394–400. 8. Lim SO, Gu JM, Kim MS, et al. Epigenetic changes induced by reactive oxygen species in hepatocellular carcinoma: methylation of the E-cadherin promoter. Gastroenterology 2008;135:2128–2140. 2140 e1–8. 9. Harder J, Opitz OG, Brabender J, et al. Quantitative promoter methylation analysis of hepatocellular carcinoma, cirrhotic and normal liver. Int J Cancer 2008;122:2800–2804. 10. Nishida N, Kudo M, Nagasaka T, et al. Characteristic patterns of altered DNA methylation predict emergence of human hepatocellular carcinoma. Hepatology 2012;56:994–1003. 11. Ammerpohl O, Pratschke J, Schafmayer C, et al. Distinct DNA methylation patterns in cirrhotic liver and hepatocellular carcinoma. Int J Cancer 2012;130:1319–1328. 12. Chiang DY, Villanueva A, Hoshida Y, et al. Focal gains of VEGFA and molecular classification of hepatocellular carcinoma. Cancer Res 2008;68:6779–6788. 13. Villanueva A, Hoshida Y, Battiston C, et al. Combining clinical, pathology, and gene expression data to predict recurrence of hepatocellular carcinoma. Gastroenterology 2011;140:1501–1512 e2. 14. Cameron EE, Bachman KE, Myohanen S, et al. Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer. Nat Genet 1999;21:103–107. 15. Nishida N, Nagasaka T, Nishimura T, et al. Aberrant methylation of multiple tumor suppressor genes in aging liver, chronic hepatitis, and hepatocellular carcinoma. Hepatology 2008;47:908–918. 16. Edamoto Y, Hara A, Biernat W, et al. Alterations of RB1, p53 and Wnt pathways in hepatocellular carcinomas associated with hepatitis C, hepatitis B and alcoholic liver cirrhosis. Int J Cancer 2003; 106:334–341. 17. Chen EY, Tan CM, Kou Y, et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinform 2013; 14:128. 18. Gronbaek K, Hother C, Jones PA. Epigenetic changes in cancer. APMIS 2007;115:1039–1059. 19. Zender L, Xue W, Cordon-Cardo C, et al. Generation and analysis of genetically defined liver carcinomas derived from bipotential liver progenitors. Cold Spring Harb Symp Quant Biol 2005;70:251–261. 20. Zender L, Xue W, Zuber J, et al. An oncogenomics-based in vivo RNAi screen identifies tumor suppressors in liver cancer. Cell 2008; 135:852–864. 21. Xue W, Kitzing T, Roessler S, et al. A cluster of cooperating tumorsuppressor gene candidates in chromosomal deletions. Proc Natl Acad Sci U S A 2012;109:8212–8217. 22. Hoshida Y, Toffanin S, Lachenmayer A, et al. Molecular classification and novel targets in hepatocellular carcinoma: recent advancements. Semin Liver Dis 2010;30:35–51.

GASTROENTEROLOGY Vol. 145, No. 6 23. Juergens RA, Wrangle J, Vendetti FP, et al. Combination epigenetic therapy has efficacy in patients with refractory advanced non-small cell lung cancer. Cancer Discov 2011;1:598–607. 24. Lachenmayer A, Toffanin S, Cabellos L, et al. Combination therapy for hepatocellular carcinoma: additive preclinical efficacy of the HDAC inhibitor panobinostat with sorafenib. J Hepatol 2012; 56:1343–1350. 25. Neumann O, Kesselmeier M, Geffers R, et al. Methylome analysis and integrative profiling of human HCCs identify novel protumorigenic factors. Hepatology 2012;56:1817–1827. 26. Simon JA, Kingston RE. Mechanisms of polycomb gene silencing: knowns and unknowns. Nat Rev Mol Cell Biol 2009;10:697–708. 27. Omura N, Li CP, Li A, et al. Genome-wide profiling of methylated promoters in pancreatic adenocarcinoma. Cancer Biol Ther 2008;7:1146–1156. 28. Simmons DP, Peach ML, Friedman JR, et al. Evidence that sequence homologous region in LRAT-like proteins possesses anti-proliferative activity and DNA binding properties: translational implications and mechanism of action. Carcinogenesis 2006;27:693–707. 29. Maekawa M, Taniguchi T, Ishikawa J, et al. Promoter hypermethylation in cancer silences LDHB, eliminating lactate dehydrogenase isoenzymes 1-4. Clin Chem 2003;49:1518–1520. 30. Leiblich A, Cross SS, Catto JW, et al. Lactate dehydrogenase-B is silenced by promoter hypermethylation in human prostate cancer. Oncogene 2006;25:2953–2960. 31. Kim JH, Kim EL, Lee YK, et al. Decreased lactate dehydrogenase B expression enhances claudin 1-mediated hepatoma cell invasiveness via mitochondrial defects. Experimental cell research 2011; 317:1108–1118. 32. Corcoran CA, He Q, Ponnusamy S, et al. Neutral sphingomyelinase-3 is a DNA damage and nongenotoxic stress-regulated gene that is deregulated in human malignancies. Mol Cancer Res 2008;6:795–807. 33. Kim MS, Chang X, LeBron C, et al. Neurofilament heavy polypeptide regulates the Akt-beta-catenin pathway in human esophageal squamous cell carcinoma. PloS One 2010;5:e9003. 34. Szebenyi G, Smith GM, Li P, et al. Overexpression of neurofilament H disrupts normal cell structure and function. J Neurosci Res 2002; 68:185–198. 35. Zhang WD, Hirohashi S, Tsuda H, et al. Frequent loss of heterozygosity on chromosomes 16 and 4 in human hepatocellular carcinoma. Jpn J Cancer Res 1990;81:108–111. 36. Kim WJ, Okimoto RA, Purton LE, et al. Mutations in the neutral sphingomyelinase gene SMPD3 implicate the ceramide pathway in human leukemias. Blood 2008;111:4716–4722. 37. Hannun YA. Functions of ceramide in coordinating cellular responses to stress. Science 1996;274:1855–1859. 38. Oskouian B, Saba JD. Cancer treatment strategies targeting sphingolipid metabolism. Adv Exp Med Biol 2010;688:185–205.

Author names in bold designate shared co-first authorship. Received March 13, 2013. Accepted August 29, 2013. Reprint requests Address requests for reprints to: Kate Revill, PhD, Mount Sinai Liver Cancer Program, Division of Liver Disease, Icahn School of Medicine at Mount Sinai, 1425 Madison Avenue, Rm 11-76, New York, New York 10029. e-mail: [email protected]. Acknowledgments The authors thank R. Finn at the University of California, Los Angeles for his generous gift of normal liver tissue from patients with no history of liver disease. We gratefully acknowledge the laboratory group of S. Lowe for sharing of engineered murine hepatoblasts. The authors also thank Laia Cabellos (Icahn School of Medicine at Mount Sinai), Lisa Bianco, Michael Cahn, and Raisa Puzis (Cold Spring Harbor Laboratory) for excellent technical assistance, and James Duffy (Cold Spring Harbor Laboratory) for his help with figure formatting.

Conflicts of interest The authors disclose no conflicts. Funding This work was supported by National Institutes of Health grants CA124648 (S.P.) and by a grant from the Starr Cancer Consortium (S.P.). Josep M. Llovet is supported by grants from the National Institute of

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER

1435

Diabetes and Digestive and Kidney Diseases (1R01DK076986), European Commission-FP7 Framework (HEPTROMIC, proposal no. 259744), the Samuel Waxman Cancer Research Foundation, the Spanish National Health Institute (SAF-2010-16055), and the Asociación Española Contra el Cáncer. GEO accession numbers: GSE44909 and GSE44970.

BASIC AND TRANSLATIONAL LIVER

December 2013

1435.e1 REVILL ET AL

Supplementary Material Patients An independent set of primary 12 HCC with paired adjacent adult liver samples was obtained via Cold Spring Harbor from the CHTN (http://www.chtn.nci.nih.gov) with appropriate Institutional Review Board or corresponding committee approval. Seven of these patients are male (58.3%), and mean patient age was 73 years old (range, 5381 years). None of these samples have a background of viral infection. All samples were de-identified so a study using these samples is not considered human subject research under the US Department of Human and Health Services regulations and related guidance (45CFR, Part 46).

Tissue Extraction Tissue was pulverized at 80 C using the BioPulverizor (Biospec, Bartlesville, OK) to ensure a homogenous mix of cells and stored at 80 C. Genomic DNA was extracted using the DNeasy tissue kit (Qiagen, Germantown, MD) and RNA extracted using the RNeasy mini kit (Qiagen).

DNA Methylation Analysis Genomic DNA for methylation profiling was quantified using the Agilent 2100 Bio analyzer. Five hundred nanograms of genomic DNA were bisulfite-converted using the EZ DNA Methylation Kit (Zymo Research, Irvine, CA). The converted DNA was hybridized to a BeadChip according to manufacturer’s instructions. The raw signal intensity for both methylated and unmethylated DNA were measured using a BeadArray Scanner (Illumina). The methylation level, or b value, of each individual cytosine preceding guanine (CpG) is obtained using the following formula by GenomeStudio (Illumina): (M) / (M) þ (U), where M is methylated DNA and U is unmethylated DNA. Extracted DNA samples (3 mg) were mutagenized by treatment with sodium bisulfite solution, as described previously.1 After polymerase chain reaction (PCR) amplification using primers specific for converted DNA, the methylation status of the CpG island associated with exon 1 of GSTP1 was assessed by combined bisulfite restriction analysis, and for SMPD3, NEFH, LDHB, ACTL6B, PRPH, and DGKI, by methylation-sensitive pyrosequencing. Forward, reverse, and sequencing primer sequences for methylation-specific pyrosequencing can be found in Supplementary Table 8. Biotinylated primer is indicated with BTN on either the forward or reverse strand. Forward and reverse primer sequence for combined bisulfite restriction analysis analysis of GSTP1 can also be found in Supplementary Table 8. For combined bisulfite restriction analysis, equal amounts of the 200-bp PCR product for GSTP1 amplified from sodium bisulfate-converted DNA were incubated in the presence and absence of the restriction enzyme, BSTUI (New England Biolabs, Ipswich, MA) at 60 C for 2 hours

GASTROENTEROLOGY Vol. 145, No. 6

as described previously.2 Post digestion of the PCR product and, in the absence of methylation, a 200-bp amplicon was apparent. Methylation, at one of the BSTUI (CGCG) sites within the amplicon, yielded fragments of 163 bp and 37 bp (Supplementary Figure 1B). For pyrosequencing analysis, we used the Pyromark Q24 system and reagents from Qiagen according to manufacturer’s guidelines. Quantitative DNA methylation analysis was carried out on the Pyromark Q24 system with the Pyromark Gold Q96 reagent kit, and results were analyzed using the PyroMark Assay Design Software 2.0 (Qiagen).

Candidate Gene Expression Analysis Single-stranded complementary DNA was synthesized from 1 mg total RNA using random primers and SuperScript II Reverse Transcriptase (Invitrogen) according to manufacturer’s instructions. Real-time PCR analysis was performed using 20-ng single-stranded complementary DNA as well as specific primers and SYBR Green PCR Master Mix (Applied Biosystems, Carlsbad, CA) on the 7900HT Fast Real Time instrument (Applied Biosystems) using the following conditions: 50 C for 2 minutes, 95 C for 10 minutes; 40 cycles: 95 C for 15 seconds, and 60 C for 15 seconds. Primer specificity was checked by melting curve analysis for each primer pair (sequences can be found in Supplementary Table 9), the PCR amplicon length and absence of primer dimers was verified once by gel electrophoresis. All reactions were performed in triplicate. The glyceraldehyde-3-phosphate dehydrogenase or hypoxanthine phosphoribosyltransferase 1 housekeeping gene was used as endogenous control for normalization. The mean was taken for the normal liver expression levels, excluding any outliers, and this value used to calculate fold change from HCC expression. All values were log-2 transformed and plotted. P values were calculated using Student’s t test of log-2 transformed quantities.

Single Nucleotide Polymorphism Arrays Genotypes and hybridization intensities for >238,000 single nucleotide polymorphisms were measured with the StyI chip of the Affymetrix 500K Human Mapping Array set (Affymetrix, Santa Clara, CA). Data were analyzed as described previously.3 For each gene, we used a cut-off of 1.8 for copy number loss based on the range of variation observed in adjacent cirrhotic tissues obtained from the same patients.

Plasmid Construction For cloning of genes under an inducible vector, complementary DNA was obtained from open biosystems and cloned into the pDestination tetracycline operator (TetO) cloning vector pcDNA4/TO using the Gateway T-REX cloning system from Invitrogen to generate pDEST SMPD3, pDEST NEFH, and pDEST LACZ. Cells were first stably transfected with pcDNA6/TR to generate

December 2013

tetracycline repressing (TetR) clones then pDEST TetO constructs containing genes of interest were stably transfected into these cells following manufacturer’s instructions. For cloning of small hairpin RNA, oligonucleotides were designed using the online tool Designer of Small Interfering RNA (http://biodev.extra.cea.fr/DSIR/DSIR.html). Tennanogram template was amplified and 50 Xho and 30 EcoR1 sites added by primers, 50 -TACAATACTCGA GAAGGTATATTGCTGTTGACAGTGAGCG-30 and 50 CTAAAGTAGCCCCTTGAATTCCGAGGCAGTAGGCA-30 . PCR conditions were 94 C for 1 minutes, 30 cycles: 94 C for 40 seconds, 68 C for 40 seconds, 72 C for 1 minute, and final step 72 C for 5 minutes. PCR products were cloned into a XhoI/EcoRI-digested ampicillin/puromycin-resistant, GFPexpressing, retroviral vector.

In Vivo Work All studies utilizing murine hepatoblasts involving shRNAs were approved by Cold Spring Harbor’s Institutional Animal Care and Use Committee. Three mice were used for each shRNA with injections on both flanks. Earlypassage immortalized liver progenitor cells were transduced by retroviruses expressing shRNAs and experiments performed as described previously.4

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e2

containing 5% fetal bovine serum was added to the bottom chamber of the Transwell.

Data Processing and Analysis Methylation data were input using KNNImputer (http://function.princeton.edu/knnimpute/) with default parameters and only probes mapping to autosomal CpG sites were retained. Low-quality methylation arrays were removed from further analysis if the average Pearson correlation of a given array to all other arrays was <0.8. Aggregated gene methylation scores were obtained by averaging the b values of a gene’s CpG islandpositive probes. Hypermethylated genes were then identified using the Significance Analysis of Microarrays (siggenes) R package with a Wilcoxon rank-sum test (http://www. bioconductor.org/packages/2.3/bioc/html/siggenes.html). GeneChip Human Exon data were summarized and normalized using the aroma.affymetrix R package6 (http://www.aroma-project.org/). Significantly unregulated genes were obtained using the Significance Analysis of Microarrays (siggenes) R package. To identify unsilenced genes, this list was further refined by including only genes whose expression exceeded 1.5 times the average background probe intensity upon cell line treatment.

Prognostic Analysis Immunohistochemistry and Immunoblotting Paraffin-embedded liver tumor sections were stained as described previously.4 Antibodies for Western blotting were SMPD3 (Abcam cat# ab85017), b-galactosidase (Abcam cat # ab616), and b-actin (Abcam cat# ab8227).

Biological Assays To measure proliferation of cells the 3-[4,5dimethylthiazol-2-yl]-2,diphenyltetrazoliumbromide assay was performed (Roche, Nutley, NJ). This reaction yields a purple formazan product that is detected using a 96-well plate reader at 570 nm. Cells were plated at 1.5  104 in the presence or absence of 5 mg/mL tetracycline and cultured for 4 days. Tetracycline-induced expression occurs at 48 hours. To measure cell migration, cells were serum starved for 3 hours, then 2.5  105 cells suspended in serum-free medium were seeded in the top chamber of the 8-mm Transwell (BD Biosciences, San Jose, CA) and 2% fetal bovine serum medium was added to the bottom chamber of the well. Assays were performed during a 22-hour period. After removal of nonmigrated cells, cells were fixed with 100% methanol for 2 minutes and stained with 1% Toludine blue for 2 minutes. Five randomized fields were captured in each Transwell under the microscope and counted using the ImageJ software (National Institutes of Health, Bethesda, MD). Invasion assays were performed as stated here, with the exception that the 8-mm Transwells were precoated with Matrigel, and serum

Association of SMPD3 expression level and time to early (<2 years after surgery) or late (2 years) recurrence was assessed by Kaplan-Meier curve, log-rank test, and Cox regression modeling. Cut-off to define SMPD3-high and -low groups was determined by surveying the interquartile range at 5-percentile intervals. We found the smallest log-rank test P value at 65-percentile cut-off in the training set, and applied the same cut-off without making modifications in the validation set in evaluating association with early recurrence (Supplementary Table 5). Clinical prognostic variables significant (P < .05) by univariate analysis for early recurrence were analyzed together with low SMPD3 expression in multivariate Cox regression modeling. All analyses were performed by using R statistical package (www.r-project.org) and GenePattern genomic analysis toolkit (www.broadinstitute.org/ genepattern). Supplementary References 1. Grunau C, Clark SJ, Rosenthal A. Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucleic Acids Res 2001;29:E65–5. 2. Simpson DJ, McNicol AM, Murray DC, et al. Molecular pathology shows p16 methylation in nonadenomatous pituitaries from patients with Cushing’s disease. Clin Cancer Res 2004;10:1780–1788. 3. Chiang DY, Villanueva A, Hoshida Y, et al. Focal gains of VEGFA and molecular classification of hepatocellular carcinoma. Cancer Res 2008;68:6779–6788. 4. Sawey ET, Chanrion M, Cai C, et al. Identification of a therapeutic strategy targeting amplified FGF19 in liver cancer by Oncogenomic screening. Cancer Cell 2011;19:347–358.

1435.e3 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Figure 1. Drug dosage determination. (A) Drug treatment optimization—quantitative realtime PCR of GSTP1 expression. (B) Representation of GSTP1 CGI analyzed by combined bisulfite restriction analysis (COBRA), including BSTU1 restriction sites. (C) Pattern of restriction for COBRA of GSTP1 treated with 500 nM DAC and 1 mM TSA. Arrows denote the pattern of restriction of untreated DNA. Retention of the unmethylated DNA band after restriction when treated with DAC/TSA indicates demethylation.

Supplementary Figure 2. Pyrosequencing results. Pyrosequencing results of 11 paired tumors (solid black line) with normal (dashed gray line). The level of methylation at each CpG site surveyed by pyrosequencing is plotted for each sample. Panels (A)(D), in order, are candidates SMPD3, NEFH, DGKI, LDHB, PRPH, and ACTL6B.

December 2013

Supplementary Figure 3. Microarray expression of candidates. Microarray expression results for each candidate, ACTL6B, C19orf30, DGKI, DLX1, ELOVL4, PPM1N, LDHB, LRAT, MLF1, NEFH, PRPH, SLC8A2, and SMPD3. Normal tissue is indicated by a black box and whisker, tumor tissue with a gray box and whisker (*P < .05; **P < .01; ***P < .001).

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e4

1435.e5 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Figure 4. Copy number loss and methylation of SMPD3 and NEFH. (A) NEFH and SMPD3 copy number via single nucleotide polymorphism (SNP) array. Dotted lines denote copy number variation of adjacent cirrhotic tissues. (B) CDKN2A, DLC1, PTEN, and RB1 copy number via SNP array. Dotted lines denote copy number variation of adjacent cirrhotic tissues. (C) SMPD3 or NEFH is hypermethylated in the majority of samples with low copy number, indicated by red circles. (D) LRAT and ELOVL4 copy number via SNP array. Dotted lines denote copy number variation of adjacent cirrhotic tissues.

December 2013

Supplementary Figure 5. Verification of TetR JHH7 pDEST expression constructs. (A) Western blot of JHH7 pDEST SMPD3 cells 5 mg/mL tetracycline. (B) Expression analysis of JHH7 pDEST NEFH cells by quantitative realtime PCR 5 mg/mL tetracycline. (NEFH encodes a protein product of 220 KDa and was difficult to detect by Western blotting techniques). (C) Western blot of JHH7 pDEST LACZ cells 5 mg/mL tetracycline.

Supplementary Figure 6. Verification of shRNA knockdown. (A) Western blot showing knockdown of SMPD3 with 3 shRNAs. (B) Quantitative real-time PCR of NEFH expression knockdown with 2 shRNAs.

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e6

1435.e7 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Figure 7. SMPD3 expression percentile bar graph. The 65th percentile is marked by a black line. High SMPD3 expressing samples are underlined in blue, low SMPD3 expressing samples in red.

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e8

Supplementary Table 1. Genes Significantly Hypermethylated in HCC Probe ID cg00949442 cg13984181 cg05488632 cg08101264 cg13705284 cg13368786 cg13547644 cg08572611 cg16786703 cg12265829 cg13878010 cg01528948 cg25997474 cg21949305 cg11934695 cg21542793 cg10235817 cg14826456 cg02126753 cg24211388 cg26524899 cg13801416 cg19118077 cg03365437 cg03760483 cg11052143 cg00061629 cg07473175 cg07044282 cg04508649 cg08321346 cg19342782 cg15883716 cg09472203 cg16970232 cg07613278 cg02157083 cg26624914 cg09954385 cg03050522 cg18084554 cg08558340 cg20967028 cg06263495 cg01245656 cg02228185 cg09580336 cg04376617 cg08995424 cg05483509 cg18236477 cg12716838 cg16334795 cg20073553 cg12091944 cg17241310 cg00169548 cg21475402 cg25303383 cg21995126 cg17602451 cg14310034 cg03447931 cg07082331 cg18952647 cg17560332

Gene symbol ABCA3 ABHD7 ABHD9 ACOT8 ACOX2 ACSS2 ACTA1 ACTL6B ADAM8 ADCY4 ADCY5 ADH1B ADH1C ADORA2A ADRA1D ADRA2B ADRA2C ADRB1 AEBP1 AIF1 AKAP5 AKR1B1 AKR1C3 ALDH1A2 ALOX12 ALS2CR11 ALX4 AMIGO2 ANGPTL1 ANGPTL7 ANKMY1 ANKRD13C ANKRD45 AP3B2 APC API5 APOA5 AQP3 ARHGAP8 ARHGEF1 ARID3A ARS2 ART4 ASCL2 ASNS ASPA ATP1A1 ATP5A1 ATP5G2 ATP6V0C ATP8A2 B3GALT3 BACE2 BAPX1 BARD1 BARHL2 BAZ1A BCAN BCDO2 BCHE BCL2 BMP4 BMP6 BMP8B BNC1 BOLL

Mean normal b

Mean tumor b

SD (tumors)

P value

0.2 0.15 0.22 0.09 0.19 0.18 0.11 0.11 0.35 0.13 0.2 0.17 0.22 0.35 0.08 0.29 0.28 0.41 0.09 0.35 0.16 0.09 0.14 0.14 0.31 0.42 0.09 0.15 0.2 0.4 0.41 0.19 0.09 0.07 0.41 0.11 0.37 0.26 0.11 0.13 0.4 0.33 0.25 0.28 0.13 0.17 0.25 0.19 0.19 0.26 0.21 0.1 0.19 0.13 0.15 0.14 0.21 0.15 0.21 0.22 0.18 0.15 0.06 0.15 0.09 0.22

0.59 0.49 0.52 0.32 0.35 0.32 0.34 0.46 0.53 0.34 0.48 0.32 0.34 0.63 0.3 0.53 0.51 0.64 0.36 0.56 0.3 0.46 0.32 0.34 0.69 0.65 0.31 0.35 0.34 0.7 0.67 0.34 0.31 0.33 0.64 0.32 0.6 0.44 0.37 0.33 0.66 0.52 0.46 0.56 0.31 0.33 0.42 0.32 0.32 0.44 0.46 0.31 0.33 0.39 0.31 0.4 0.36 0.51 0.35 0.39 0.36 0.42 0.31 0.35 0.31 0.44

0.24 0.24 0.19 0.21 0.11 0.07 0.26 0.26 0.23 0.2 0.22 0.08 0.06 0.2 0.22 0.23 0.19 0.2 0.21 0.17 0.13 0.27 0.08 0.19 0.2 0.16 0.19 0.2 0.11 0.14 0.21 0.09 0.17 0.26 0.25 0.09 0.18 0.17 0.21 0.22 0.19 0.19 0.17 0.25 0.23 0.09 0.12 0.06 0.16 0.18 0.22 0.25 0.08 0.22 0.1 0.21 0.07 0.26 0.07 0.08 0.2 0.22 0.26 0.22 0.23 0.18

.000 .000 .000 .004 .000 .000 .022 .000 .025 .004 .000 .000 .000 .000 .004 .004 .001 .001 .000 .000 .002 .000 .000 .004 .000 .000 .002 .006 .001 .000 .000 .000 .000 .006 .012 .000 .000 .004 .001 .016 .000 .005 .001 .002 .041 .000 .000 .000 .034 .006 .001 .020 .000 .001 .000 .001 .000 .000 .000 .000 .021 .001 .010 .016 .013 .001

1435.e9 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 1. Continued Probe ID cg16778903 cg23211240 cg15227982 cg09730349 cg08516400 cg12281657 cg00239071 cg08085267 cg03016571 cg24300924 cg03996793 cg26728886 cg26090652 cg05538432 cg04369341 cg08713365 cg18919097 cg20337106 cg00463577 cg04600618 cg04999691 cg17939444 cg00755043 cg09491779 cg18503260 cg18674980 cg04532952 cg20256494 cg14009688 cg17714799 cg17453778 cg04759439 cg14564494 cg02849695 cg00891278 cg02620769 cg03724463 cg04590978 cg05436231 cg19591881 cg17108819 cg26850754 cg24975222 cg07290435 cg19475870 cg14988503 cg23548479 cg09099744 cg10210238 cg18137704 cg06268694 cg09736162 cg05052633 cg09425215 cg08080029 cg12422450 cg24130010 cg03853987 cg00995327 cg00840403 cg10313673 cg10978355 cg20924286 cg18403361 cg08752459 cg06432655

Gene symbol BRUNOL6 BTG4 C10orf26 C14orf143 C14orf166 C14orf50 C14orf68 C17orf57 C17orf73 C18orf43 C19orf30 C1orf118 C1QTNF5 C1S C20orf100 C20orf98 C3orf57 C6orf139 C6orf150 C6orf206 C7orf29 C8orf4 C9orf121 C9orf82 CA13 CA3 CA4 CABP7 CALD1 CASP6 CASR CAST1 CBR3 CCDC19 CCDC37 CCDC65 CCK CCNJ CD164L2 CD34 CD8A CD8B1 CDCP1 CDH10 CDH9 CDKL2 CDKN1C CDKN2A CDKN2B CECR6 CELSR1 CELSR3 CENTA2 CHD2 CHD5 CHGA CHODL CHST10 CHST2 CHST4 CILP2 CKMT2 CLDN11 CLEC14A CLEC2B CLIPR-59

Mean normal b

Mean tumor b

SD (tumors)

P value

0.2 0.28 0.37 0.23 0.14 0.15 0.56 0.17 0.28 0.12 0.15 0.24 0.09 0.22 0.18 0.36 0.24 0.21 0.17 0.14 0.3 0.22 0.04 0.1 0.18 0.14 0.1 0.17 0.18 0.18 0.2 0.27 0.26 0.25 0.34 0.13 0.2 0.13 0.25 0.14 0.07 0.14 0.07 0.16 0.24 0.06 0.19 0.07 0.38 0.15 0.26 0.14 0.14 0.13 0.1 0.21 0.19 0.19 0.08 0.16 0.17 0.27 0.18 0.18 0.17 0.36

0.46 0.44 0.56 0.35 0.34 0.4 0.89 0.32 0.44 0.38 0.31 0.41 0.37 0.36 0.43 0.6 0.39 0.42 0.35 0.51 0.48 0.33 0.33 0.3 0.42 0.38 0.34 0.42 0.31 0.33 0.37 0.47 0.51 0.56 0.57 0.39 0.31 0.41 0.54 0.36 0.33 0.37 0.33 0.34 0.35 0.37 0.38 0.52 0.73 0.3 0.67 0.38 0.37 0.31 0.33 0.47 0.4 0.51 0.37 0.3 0.41 0.62 0.35 0.32 0.39 0.61

0.26 0.2 0.16 0.09 0.11 0.19 0.17 0.12 0.16 0.22 0.11 0.14 0.22 0.09 0.21 0.24 0.18 0.09 0.23 0.24 0.19 0.08 0.21 0.11 0.2 0.22 0.25 0.22 0.14 0.11 0.14 0.22 0.16 0.23 0.18 0.25 0.16 0.22 0.24 0.23 0.24 0.24 0.21 0.07 0.1 0.25 0.22 0.28 0.21 0.13 0.21 0.26 0.23 0.08 0.25 0.27 0.2 0.27 0.27 0.13 0.3 0.22 0.11 0.16 0.09 0.17

.008 .028 .001 .001 .000 .000 .000 .001 .005 .001 .000 .001 .001 .000 .001 .005 .031 .000 .037 .000 .007 .000 .000 .000 .001 .003 .012 .002 .012 .000 .001 .013 .000 .000 .000 .005 .049 .000 .001 .010 .003 .010 .001 .000 .001 .001 .018 .000 .000 .001 .000 .013 .010 .000 .015 .011 .003 .001 .003 .003 .029 .000 .000 .012 .000 .000

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e10

Supplementary Table 1. Continued Probe ID

Gene symbol

Mean normal b

Mean tumor b

SD (tumors)

P value

cg01683883 cg23297477 cg00290506 cg16907566 cg11846236 cg09949775 cg15819333 cg19885761 cg02266731 cg16444968 cg21604803 cg14726637 cg20329958 cg04922810 cg01410472 cg03544320 cg13277939 cg17800442 cg26952662 cg02936872 cg08578023 cg01288089 cg22670329 cg18956481 cg04884908 cg20322977 cg05684891 cg04515001 cg01086895 cg00890257 cg01344452 cg06277657 cg26267310 cg04623955 cg17886204 cg24890043 cg15236866 cg16924616 cg13191049 cg23391785 cg26117023 cg07295678 cg12758687 cg06825142 cg09936561 cg11832722 cg19703610 cg13460409 cg15383120 cg09328024 cg24105685 cg10885338 cg20083676 cg10521852 cg22836229 cg20786074 cg24929737 cg10541755 cg25332298 cg13297865 cg09009111 cg14409958 cg17264618 cg00027083 cg22778981 cg03127334

CMTM2 CMTM3 CNIH3 COL14A1 COL7A1 COMP COX4I1 CPLX2 CPM CPNE7 CPT1C CR1 CRH CRHR2 CRISPLD1 CRMP1 CTAGE5 CTF1 CTHRC1 CTSL2 CTSS CXCL10 CXCL6 CYP24A1 CYP26B1 CYP26C1 DAB2IP DCDC2 DCHS1 DEGS2 DGKE DGKI DHRS10 DIO3 DKFZp434I1020 DKFZP566N034 DLX1 DLX5 DMN DNM3 DOK1 DPYSL4 DRD2 DRD4 DRD5 DSC3 DSCAML1 DSCR6 DUSP22 DYRK3 ECM2 ECRG4 EDG3 EDG4 EFCAB1 EFEMP1 EFNB2 EIF5A2 ELAVL3 ELOVL4 EMILIN2 ENPP2 ENTPD3 EPB41L3 ERBB2 ERG

0.24 0.07 0.03 0.23 0.18 0.3 0.2 0.2 0.29 0.17 0.14 0.23 0.11 0.15 0.09 0.36 0.29 0.16 0.1 0.04 0.19 0.17 0.21 0.09 0.17 0.39 0.3 0.22 0.1 0.09 0.17 0.2 0.27 0.14 0.06 0.04 0.18 0.08 0.33 0.18 0.06 0.16 0.23 0.08 0.15 0.14 0.15 0.23 0.22 0.18 0.21 0.48 0.21 0.41 0.33 0.22 0.06 0.16 0.13 0.1 0.13 0.16 0.27 0.21 0.13 0.18

0.42 0.37 0.41 0.48 0.41 0.53 0.34 0.42 0.45 0.31 0.38 0.47 0.36 0.3 0.3 0.67 0.44 0.31 0.44 0.37 0.31 0.32 0.41 0.48 0.31 0.62 0.59 0.58 0.34 0.4 0.42 0.46 0.52 0.39 0.37 0.35 0.32 0.33 0.54 0.66 0.34 0.36 0.44 0.4 0.4 0.51 0.31 0.43 0.35 0.31 0.37 0.74 0.39 0.69 0.53 0.42 0.31 0.46 0.3 0.35 0.35 0.32 0.51 0.38 0.32 0.43

0.22 0.26 0.26 0.19 0.2 0.23 0.06 0.22 0.13 0.19 0.21 0.21 0.14 0.17 0.18 0.22 0.1 0.1 0.27 0.3 0.07 0.07 0.17 0.24 0.17 0.13 0.25 0.27 0.13 0.3 0.32 0.23 0.23 0.21 0.21 0.27 0.12 0.24 0.27 0.28 0.24 0.22 0.17 0.26 0.2 0.23 0.21 0.19 0.16 0.16 0.09 0.16 0.2 0.18 0.23 0.17 0.22 0.25 0.2 0.24 0.23 0.12 0.2 0.23 0.16 0.16

.031 .002 .000 .000 .001 .006 .000 .007 .001 .041 .002 .001 .000 .016 .001 .000 .000 .000 .001 .004 .000 .000 .001 .000 .029 .000 .002 .000 .000 .006 .037 .002 .002 .001 .000 .002 .002 .006 .036 .000 .001 .015 .000 .001 .001 .000 .032 .004 .031 .021 .000 .000 .012 .000 .017 .001 .002 .001 .018 .005 .012 .000 .001 .044 .002 .000

1435.e11 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 1. Continued Probe ID cg01566404 cg27626299 cg11591325 cg24573501 cg08241785 cg01185754 cg04536922 cg06073471 cg12998491 cg19682367 cg16534233 cg16516400 cg18671950 cg27223047 cg02699167 cg20723355 cg18661868 cg15774153 cg24030449 cg19831575 cg27062617 cg19566405 cg09001953 cg03455458 cg19786920 cg06384053 cg02729303 cg07717632 cg03616357 cg25484904 cg02647265 cg22029275 cg04329382 cg19018097 cg00393585 cg02245378 cg14893161 cg21269934 cg25875213 cg25784308 cg02489552 cg03734874 cg25044651 cg07017374 cg00489401 cg15514848 cg22815110 cg18815943 cg10300684 cg14312526 cg16046465 cg20308679 cg25176823 cg25228126 cg01725199 cg05714219 cg24576425 cg21549904 cg22471346 cg06490988 cg06230736 cg06954481 cg07376029 cg08586737 cg02844545 cg22289360

Gene symbol ETNK2 EVX1 F2R F2RL1 F2RL2 F3 FAM13A1 FAM55C FAM78A FAM79B FAM80A FAM89A FBN1 FBN2 FBXL2 FBXO39 FES FGF19 FGF20 FGF4 FHOD1 FLJ10260 FLJ11200 FLJ12505 FLJ14001 FLJ20010 FLJ20152 FLJ20245 FLJ21159 FLJ21511 FLJ22471 FLJ25477 FLJ27365 FLJ30934 FLJ31659 FLJ32447 FLJ32569 FLJ37478 FLJ37549 FLJ38377 FLJ40365 FLJ42486 FLJ90650 FLT3 FLT4 FMO1 FOXD3 FOXE3 FOXG1B FOXL2 FRMD5 FRZB FSCN1 FZD2 GALNT12 GALNT14 GALNT5 GALNT7 GAS7 GATA2 GATA3 GBX2 GC GCC1 GCM2 GCNT1

Mean normal b

Mean tumor b

SD (tumors)

P value

0.31 0.14 0.06 0.18 0.21 0.19 0.15 0.05 0.31 0.15 0.07 0.24 0.08 0.21 0.05 0.23 0.23 0.06 0.13 0.12 0.1 0.21 0.11 0.07 0.26 0.2 0.17 0.31 0.28 0.27 0.22 0.08 0.22 0.09 0.15 0.15 0.34 0.11 0.02 0.16 0.16 0.3 0.2 0.09 0.16 0.25 0.24 0.16 0.09 0.17 0.19 0.19 0.18 0.21 0.05 0.09 0.17 0.06 0.06 0.34 0.26 0.08 0.21 0.2 0.19 0.23

0.62 0.36 0.34 0.3 0.36 0.33 0.38 0.39 0.51 0.32 0.32 0.48 0.4 0.5 0.31 0.45 0.47 0.37 0.31 0.4 0.31 0.41 0.33 0.34 0.49 0.39 0.31 0.53 0.51 0.53 0.37 0.34 0.4 0.32 0.31 0.31 0.53 0.41 0.32 0.31 0.32 0.5 0.5 0.35 0.51 0.38 0.58 0.55 0.31 0.41 0.44 0.36 0.38 0.36 0.37 0.35 0.31 0.32 0.37 0.53 0.45 0.42 0.34 0.3 0.37 0.47

0.25 0.2 0.26 0.11 0.12 0.12 0.16 0.26 0.2 0.09 0.26 0.22 0.24 0.21 0.23 0.26 0.2 0.26 0.17 0.25 0.26 0.15 0.15 0.28 0.22 0.08 0.15 0.22 0.23 0.18 0.08 0.21 0.21 0.21 0.18 0.17 0.21 0.2 0.26 0.17 0.11 0.23 0.23 0.26 0.27 0.1 0.25 0.25 0.18 0.27 0.27 0.18 0.2 0.16 0.23 0.23 0.15 0.29 0.24 0.2 0.23 0.29 0.07 0.13 0.18 0.22

.001 .003 .004 .004 .001 .001 .000 .000 .007 .000 .008 .004 .000 .000 .002 .021 .001 .001 .004 .003 .037 .000 .000 .009 .004 .000 .007 .006 .006 .000 .000 .001 .023 .003 .017 .011 .018 .000 .002 .021 .000 .026 .000 .005 .000 .000 .000 .000 .001 .017 .011 .015 .006 .013 .000 .002 .010 .018 .000 .009 .031 .001 .000 .032 .006 .003

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e12

Supplementary Table 1. Continued Probe ID cg00509616 cg04448487 cg11452221 cg22341104 cg14917512 cg17301902 cg09649610 cg04267184 cg18691434 cg21870884 cg13562542 cg14963371 cg10189695 cg27634151 cg08368934 cg04034767 cg21296230 cg15425280 cg26599006 cg26609631 cg02659086 cg23984434 cg26656452 cg24418574 cg07703401 cg25462291 cg07364841 cg10146929 cg20520725 cg10046620 cg08009328 cg21663122 cg02192855 cg05414338 cg02909790 cg17718302 cg08260959 cg14743462 cg12024906 cg20899053 cg03020951 cg02906939 cg10883303 cg26069745 cg04317399 cg02248486 cg23432345 cg26521404 cg25882366 cg15539420 cg07123069 cg03918304 cg16632715 cg00005847 cg15520279 cg24417499 cg00754253 cg27120999 cg12582959 cg06291867 cg03296929 cg17252960 cg17980508 cg01305421 cg22375192 cg06638433

Gene symbol Gcom1 GDAP1L1 GEFT GFI1 GNA11 GNA14 GNG4 GNRH2 GPC2 GPR25 GPR27 GPR3 GPR78 GPR83 GPR97 GRASP GREM1 GRIA2 GSCL GSH1 GSTP1 GUCY1A2 HABP2 HADHB HBQ1 HEYL HGFAC HIST1H1A HIST1H2AE HIST1H2AI HIST1H2BF HIST1H2BH HIST1H2BI HIST1H3F HIST1H3G HIST1H3J HIST1H4F HIST3H2A HKR1 HLXB9 HMGA1 HNMT HOXA13 HOXA2 HOXA4 HOXA5 HOXA7 HOXA9 HOXB2 HOXB8 HOXC11 HOXD10 HOXD11 HOXD3 HOXD8 HPCA HRASLS5 HSPA2 HSU79303 HTR7 IBSP ID4 IFI44L IGF1 IGF1R IGF2BP1

Mean normal b

Mean tumor b

SD (tumors)

P value

0.21 0.39 0.17 0.09 0.18 0.13 0.21 0.35 0.3 0.38 0.18 0.2 0.14 0.12 0.28 0.17 0.21 0.13 0.09 0.1 0.14 0.11 0.26 0.14 0.18 0.1 0.44 0.21 0.14 0.22 0.18 0.13 0.1 0.05 0.21 0.07 0.14 0.08 0.19 0.1 0.16 0.2 0.1 0.16 0.34 0.25 0.12 0.23 0.28 0.2 0.13 0.22 0.14 0.15 0.14 0.24 0.26 0.25 0.2 0.23 0.21 0.15 0.17 0.14 0.1 0.21

0.42 0.6 0.31 0.31 0.32 0.34 0.46 0.67 0.58 0.61 0.48 0.3 0.34 0.33 0.44 0.52 0.4 0.31 0.41 0.43 0.38 0.31 0.42 0.33 0.52 0.33 0.66 0.32 0.31 0.41 0.32 0.41 0.31 0.33 0.46 0.39 0.42 0.35 0.38 0.33 0.37 0.41 0.38 0.38 0.5 0.51 0.33 0.56 0.45 0.39 0.31 0.35 0.35 0.31 0.44 0.42 0.45 0.67 0.34 0.54 0.38 0.32 0.32 0.32 0.49 0.39

0.24 0.21 0.17 0.2 0.15 0.17 0.25 0.2 0.25 0.21 0.26 0.12 0.22 0.18 0.19 0.28 0.18 0.16 0.24 0.25 0.26 0.21 0.19 0.08 0.22 0.23 0.2 0.15 0.17 0.19 0.06 0.18 0.17 0.24 0.23 0.27 0.2 0.26 0.23 0.29 0.2 0.18 0.31 0.19 0.21 0.2 0.23 0.29 0.16 0.2 0.2 0.13 0.2 0.14 0.24 0.19 0.14 0.22 0.15 0.2 0.07 0.18 0.09 0.1 0.28 0.23

.015 .007 .031 .003 .011 .001 .007 .000 .002 .003 .002 .021 .010 .001 .022 .000 .006 .002 .000 .000 .012 .013 .020 .000 .000 .007 .001 .049 .007 .005 .000 .000 .001 .002 .002 .001 .000 .006 .022 .030 .005 .001 .013 .001 .030 .000 .014 .002 .005 .010 .014 .004 .005 .001 .001 .007 .000 .000 .010 .000 .000 .012 .000 .000 .000 .035

1435.e13 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 1. Continued Probe ID cg27416067 cg11304234 cg04481779 cg07603484 cg16468729 cg25764191 cg07850604 cg20395892 cg15433631 cg03963198 cg05266781 cg14614211 cg11102782 cg01519742 cg08126211 cg00995520 cg26162582 cg27409364 cg05675373 cg03801286 cg25123470 cg01643580 cg19779211 cg05373457 cg20673481 cg01500097 cg17729667 cg16797831 cg17465304 cg15613048 cg04270799 cg07685869 cg23282559 cg04528819 cg13847070 cg11846956 cg11618577 cg07846220 cg23696949 cg08899626 cg03243946 cg12782180 cg19099850 cg06759890 cg22660578 cg07109287 cg20300246 cg00585790 cg06186808 cg01870826 cg00881370 cg15309006 cg27011193 cg11012046 cg07260592 cg15398520 cg20654468 cg23303074 cg23089840 cg06921282 cg08965324 cg23180489 cg15014458 cg01615704 cg07380496 cg00233307

Gene symbol IL12RB2 IL18 IL20RA IL4I1 IL8 INA INSM2 IRAK3 IRX2 IRX4 IRX5 IRXL1 ISYNA1 JAKMIP1 KAAG1 KCNA3 KCNA6 KCNC1 KCNC4 KCNE1 KCNIP2 KCNK3 KCNQ1 KCNS2 KCNS3 KIAA0274 KIAA0980 KIAA1324 KIF12 KIF17 KIF5A KIFC3 KL KLF14 KLHL3 KLK10 KRTCAP3 LAMA1 LAMC2 LDB2 LDHB LEP LGALS3 LHFPL2 LHX1 LHX2 LHX3 LIMS1 LOC161247 LOC389458 LOC55908 LOC63928 LOC654342 LOC89944 LPA LPAL2 LPXN LRAT LRRC3 LRRC56 LTA4H LYNX1 LYPD3 MALL MAP1B MAP4K1

Mean normal b

Mean tumor b

SD (tumors)

P value

0.08 0.19 0.27 0.44 0.12 0.24 0.09 0.21 0.19 0.12 0.28 0.13 0.35 0.09 0.21 0.12 0.09 0.31 0.28 0.31 0.16 0.28 0.11 0.21 0.12 0.19 0.17 0.19 0.22 0.16 0.23 0.14 0.14 0.17 0.19 0.1 0.46 0.07 0.29 0.15 0.2 0.41 0.17 0.2 0.18 0.1 0.22 0.32 0.44 0.25 0.43 0.18 0.21 0.11 0.15 0.28 0.22 0.16 0.25 0.2 0.13 0.36 0.39 0.27 0.07 0.25

0.32 0.31 0.44 0.74 0.31 0.5 0.32 0.4 0.5 0.45 0.48 0.38 0.53 0.32 0.51 0.32 0.3 0.66 0.45 0.53 0.36 0.46 0.45 0.47 0.31 0.35 0.34 0.36 0.37 0.43 0.37 0.32 0.33 0.48 0.38 0.3 0.77 0.39 0.47 0.31 0.4 0.7 0.43 0.36 0.49 0.32 0.37 0.48 0.73 0.56 0.8 0.38 0.37 0.34 0.32 0.43 0.43 0.34 0.42 0.32 0.36 0.65 0.67 0.43 0.33 0.44

0.17 0.09 0.17 0.21 0.08 0.22 0.2 0.19 0.24 0.25 0.2 0.21 0.2 0.27 0.22 0.17 0.19 0.18 0.19 0.18 0.23 0.16 0.21 0.2 0.23 0.07 0.19 0.17 0.17 0.24 0.2 0.08 0.2 0.24 0.14 0.22 0.2 0.27 0.24 0.11 0.23 0.19 0.27 0.14 0.22 0.27 0.19 0.12 0.19 0.25 0.22 0.2 0.1 0.23 0.2 0.11 0.18 0.19 0.14 0.15 0.1 0.19 0.2 0.2 0.23 0.23

.000 .001 .005 .000 .000 .001 .002 .005 .000 .000 .007 .001 .013 .023 .000 .001 .003 .000 .017 .001 .020 .002 .000 .000 .023 .000 .019 .006 .023 .002 .047 .000 .008 .000 .000 .013 .000 .001 .037 .000 .023 .000 .008 .004 .000 .027 .040 .000 .000 .001 .000 .008 .000 .006 .018 .000 .001 .014 .001 .031 .000 .000 .000 .033 .002 .036

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e14

Supplementary Table 1. Continued Probe ID cg03705396 cg22892110 cg21096399 cg27239157 cg18798750 cg04726446 cg11277230 cg05949660 cg15147516 cg20182358 cg07935568 cg07570142 cg03557733 cg23710218 cg24840099 cg16158681 cg24957240 cg17687883 cg24378421 cg06039392 cg07271264 cg14916079 cg18081258 cg15467759 cg07388493 cg23290344 cg02994956 cg04645342 cg12978308 cg02755525 cg13735974 cg22881914 cg00237010 cg15105703 cg09260089 cg08109815 cg13707560 cg09632136 cg21461100 cg09628601 cg22134325 cg14477619 cg05839235 cg12799895 cg05158615 cg12613344 cg03958979 cg25511429 cg22367989 cg14384532 cg08532057 cg00208967 cg16029760 cg14882700 cg15607672 cg20909686 cg11487705 cg06637774 cg05671350 cg23043245 cg15662251 cg25683904 cg19352038 cg19996355 cg20366906 cg12629325

Gene symbol MAP6D1 MAPK15 MCAM MCF2L2 MGC14289 MGC16372 MGC26744 MICAL1 MIXL1 MLF1 MLNR MOXD1 MPP2 MSC MSX1 MT3 MTERFD3 MTHFD2 MTMR9 MTNR1A MYOD1 NANOS1 NDRG2 NDST1 NDUFS5 NEF3 NEFH NEGR1 NEIL1 NETO2 NFYC NID2 NINJ2 NIP NKX6-2 NMBR NME5 NNMT NOVA2 NPAS1 NPAS4 NPC1L1 NPR3 NPTX2 NPY NR1H3 NR2E1 NRN1 NRP2 NTRK3 NUPL1 OLFM2 OSTbeta OTOP1 OTX2 OVOL1 OXCT1 P2RY6 P4HA3 PACSIN1 PAQR7 PARG PAX3 PBX4 PCDH8 PCDHAC1

Mean normal b

Mean tumor b

SD (tumors)

P value

0.32 0.21 0.32 0.18 0.08 0.2 0.21 0.29 0.17 0.18 0.15 0.12 0.24 0.26 0.31 0.17 0.11 0.15 0.26 0.21 0.13 0.27 0.17 0.5 0.27 0.09 0.4 0.09 0.18 0.19 0.09 0.12 0.35 0.17 0.17 0.25 0.28 0.18 0.14 0.11 0.26 0.39 0.1 0.2 0.2 0.36 0.4 0.2 0.08 0.2 0.2 0.27 0.26 0.17 0.17 0.28 0.09 0.3 0.24 0.23 0.2 0.17 0.18 0.24 0.21 0.07

0.5 0.48 0.51 0.33 0.34 0.39 0.4 0.5 0.53 0.43 0.39 0.36 0.41 0.45 0.52 0.42 0.33 0.38 0.39 0.41 0.38 0.46 0.51 0.75 0.41 0.39 0.63 0.3 0.31 0.49 0.3 0.51 0.56 0.33 0.45 0.45 0.47 0.37 0.3 0.3 0.51 0.61 0.34 0.46 0.45 0.67 0.66 0.39 0.33 0.34 0.34 0.42 0.44 0.35 0.38 0.54 0.32 0.51 0.4 0.42 0.31 0.33 0.32 0.45 0.43 0.41

0.2 0.25 0.23 0.19 0.26 0.2 0.22 0.21 0.27 0.21 0.21 0.25 0.23 0.2 0.2 0.21 0.09 0.21 0.18 0.16 0.22 0.15 0.2 0.21 0.16 0.24 0.16 0.23 0.17 0.24 0.1 0.28 0.13 0.2 0.23 0.18 0.21 0.11 0.21 0.15 0.18 0.2 0.21 0.2 0.22 0.15 0.2 0.19 0.22 0.19 0.16 0.16 0.19 0.17 0.16 0.26 0.23 0.21 0.18 0.21 0.13 0.07 0.18 0.23 0.18 0.26

.010 .003 .027 .034 .006 .011 .018 .005 .000 .001 .002 .009 .049 .011 .003 .001 .000 .002 .047 .000 .002 .000 .000 .001 .016 .001 .000 .011 .037 .001 .000 .000 .000 .031 .001 .002 .017 .000 .044 .001 .000 .002 .002 .000 .002 .000 .000 .010 .003 .039 .011 .013 .009 .004 .000 .007 .009 .010 .019 .013 .023 .000 .037 .018 .001 .000

1435.e15 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 1. Continued Probe ID cg11809091 cg11647681 cg23563234 cg15361590 cg17861230 cg11935147 cg23090824 cg06339657 cg14473924 cg18895972 cg04598121 cg08465862 cg02806777 cg14681055 cg11377136 cg02510853 cg18630040 cg04993257 cg06521280 cg15120942 cg01745657 cg12613383 cg05512099 cg15083227 cg26220350 cg14601284 cg09837648 cg02154186 cg16302441 cg21243096 cg20291049 cg08097882 cg24199834 cg00116838 cg07038400 cg12585943 cg12374721 cg01295203 cg24653967 cg14502651 cg13577076 cg09595479 cg24989962 cg06738602 cg13986130 cg25057743 cg08747889 cg06971096 cg12164282 cg19850348 cg08755040 cg04764624 cg26029902 cg17982102 cg05554936 cg02525756 cg05749161 cg16154416 cg09952204 cg13099330 cg04041960 cg21303386 cg24653181 cg19332710 cg00948524 cg25766046

Gene symbol PCDHB7 PCDHGA12 PCDHGB7 PCDHGC4 PDE4C PDE4DIP PDE7B PDE8B PDZRN3 PELO PENK PFKP PGLYRP1 PITX3 PKDREJ PKMYT1 PLA2G7 PLAC2 PLAU PLCD1 PLCXD2 PLD5 PLEKHF1 PLEKHQ1 PLTP PLXDC1 PLXNB1 PNMA2 POMC POU3F1 POU3F3 POU4F1 POU4F2 PPM1N PPP2R3A PPT2 PRAC PRDM14 PRH1 PRH2 PRKAR1B PRPH PTGDR PTGER2 PTGS2 PTHR2 PTK7 PTPRN PXDN PYGO1 QSCN6 RAB11FIP4 RAB22A RAB31 RAB3D RAB42 RASA4 RASGRF1 RASGRF2 RBP1 RGS10 RGS7 RHCG RIMS4 RNF135 ROR2

Mean normal b

Mean tumor b

SD (tumors)

P value

0.2 0.22 0.18 0.27 0.3 0.23 0.14 0.29 0.14 0.19 0.2 0.06 0.4 0.13 0.35 0.44 0.06 0.36 0.16 0.32 0.2 0.18 0.15 0.23 0.11 0.15 0.54 0.24 0.25 0.15 0.19 0.13 0.16 0.15 0.33 0.39 0.07 0.19 0.41 0.26 0.2 0.14 0.11 0.1 0.07 0.24 0.13 0.14 0.21 0.13 0.12 0.2 0.32 0.15 0.14 0.21 0.13 0.22 0.1 0.16 0.24 0.18 0.18 0.03 0.27 0.21

0.31 0.42 0.45 0.46 0.47 0.49 0.34 0.45 0.33 0.35 0.42 0.31 0.64 0.39 0.69 0.68 0.42 0.56 0.34 0.5 0.32 0.32 0.32 0.34 0.36 0.33 0.82 0.52 0.44 0.4 0.53 0.48 0.46 0.37 0.53 0.59 0.33 0.48 0.68 0.5 0.54 0.44 0.39 0.37 0.31 0.44 0.35 0.35 0.41 0.36 0.32 0.32 0.52 0.42 0.41 0.39 0.3 0.4 0.37 0.44 0.5 0.35 0.37 0.36 0.42 0.37

0.08 0.2 0.2 0.22 0.21 0.18 0.11 0.15 0.19 0.08 0.19 0.21 0.16 0.25 0.18 0.15 0.28 0.18 0.18 0.17 0.07 0.17 0.12 0.14 0.25 0.18 0.14 0.23 0.17 0.24 0.22 0.25 0.22 0.18 0.17 0.17 0.21 0.27 0.14 0.18 0.26 0.25 0.23 0.22 0.19 0.2 0.23 0.16 0.26 0.25 0.2 0.16 0.22 0.2 0.25 0.19 0.24 0.18 0.23 0.26 0.23 0.17 0.24 0.3 0.21 0.2

.000 .006 .000 .027 .028 .000 .000 .004 .006 .000 .001 .001 .000 .005 .000 .000 .000 .002 .009 .003 .000 .022 .000 .024 .007 .007 .000 .001 .002 .005 .000 .000 .000 .002 .001 .001 .001 .004 .000 .000 .000 .001 .001 .001 .001 .005 .010 .000 .033 .011 .007 .049 .014 .000 .004 .009 .048 .010 .002 .004 .002 .007 .030 .003 .048 .026

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e16

Supplementary Table 1. Continued Probe ID cg26417554 cg06377278 cg15191648 cg18808261 cg24438217 cg01668465 cg16954341 cg20466180 cg15513137 cg15457899 cg01857260 cg15250797 cg00347904 cg02787991 cg24019851 cg12078929 cg09061733 cg04809136 cg22418909 cg15787039 cg21825027 cg19362572 cg17191178 cg12865837 cg19283196 cg11879514 cg20439022 cg03835296 cg14008883 cg27292431 cg25152631 cg09492887 cg05521696 cg25437385 cg03064067 cg21291896 cg22123464 cg15201635 cg04797323 cg06675478 cg02919422 cg19063972 cg21530890 cg20377955 cg25802093 cg04786857 cg15375239 cg19502744 cg02164046 cg25725843 cg20339230 cg19751300 cg11271605 cg06782692 cg08291098 cg10334928 cg15415545 cg05472874 cg15873301 cg16076328 cg16080552 cg19797376 cg25481253 cg25608041 cg20209009 cg18536148

Gene symbol RPUSD3 RUNX3 SALL3 SATB1 SCC-112 SCG2 SCGN SCHIP1 SCN2A2 SCN3B SCRL SCTR SCUBE3 SECTM1 SEMA3B SERHL SERPING1 SF3B14 SFRP1 SGNE1 SH3YL1 SHC4 SHOX2 SIM1 SLC10A4 SLC16A6 SLC16A8 SLC17A1 SLC18A3 SLC22A1 SLC25A36 SLC26A5 SLC2A14 SLC35F3 SLC6A15 SLC6A6 SLC8A2 SMPD3 SOCS2 SOX1 SOX17 SOX21 SOX8 SP8 SPAG6 SPDY1 SPINT2 SRD5A2 SST ST6GAL2 ST8SIA2 ST8SIA5 STEAP4 STK32C STMN3 STON2 SULT1A3 SULT4A1 SYN2 TACSTD1 TACSTD2 TAL1 TAS2R38 TBC1D1 TBX21 TBX4

Mean normal b

Mean tumor b

SD (tumors)

P value

0.22 0.12 0.1 0.2 0.24 0.13 0.17 0.18 0.14 0.2 0.18 0.16 0.17 0.42 0.46 0.36 0.19 0.37 0.16 0.28 0.18 0.14 0.06 0.25 0.24 0.07 0.4 0.24 0.19 0.37 0.05 0.19 0.22 0.1 0.1 0.11 0.23 0.13 0.12 0.21 0.2 0.16 0.12 0.17 0.23 0.22 0.26 0.2 0.13 0.18 0.12 0.17 0.1 0.08 0.15 0.36 0.12 0.09 0.09 0.47 0.18 0.21 0.18 0.46 0.18 0.17

0.36 0.43 0.42 0.39 0.37 0.3 0.35 0.32 0.36 0.4 0.35 0.33 0.47 0.64 0.7 0.59 0.3 0.58 0.32 0.56 0.47 0.32 0.32 0.38 0.4 0.31 0.61 0.39 0.37 0.66 0.37 0.39 0.4 0.35 0.32 0.33 0.36 0.32 0.46 0.37 0.5 0.34 0.34 0.32 0.58 0.51 0.5 0.38 0.46 0.33 0.33 0.37 0.32 0.33 0.39 0.61 0.34 0.35 0.35 0.75 0.46 0.41 0.32 0.7 0.36 0.53

0.14 0.28 0.23 0.09 0.07 0.09 0.18 0.08 0.08 0.23 0.17 0.21 0.23 0.22 0.17 0.24 0.11 0.25 0.19 0.16 0.24 0.11 0.19 0.09 0.18 0.18 0.13 0.19 0.18 0.22 0.26 0.22 0.2 0.26 0.15 0.24 0.17 0.2 0.29 0.18 0.2 0.21 0.23 0.14 0.24 0.22 0.22 0.23 0.25 0.16 0.27 0.21 0.2 0.24 0.2 0.15 0.22 0.23 0.22 0.21 0.2 0.17 0.08 0.16 0.25 0.28

.007 .002 .000 .000 .000 .000 .007 .000 .000 .018 .006 .025 .000 .007 .000 .010 .007 .025 .026 .000 .001 .000 .000 .000 .015 .000 .000 .040 .006 .000 .001 .012 .011 .010 .000 .013 .031 .012 .001 .019 .000 .022 .012 .004 .000 .000 .004 .028 .000 .007 .036 .009 .003 .004 .001 .000 .006 .002 .001 .000 .000 .001 .000 .000 .040 .000

1435.e17 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 1. Continued Probe ID cg24110050 cg08121954 cg12277666 cg02545192 cg06750167 cg25226247 cg26372517 cg02906238 cg15339605 cg00563926 cg18338296 cg07965823 cg25720804 cg14696396 cg20001829 cg23780947 cg06179485 cg27090216 cg11164347 cg10693071 cg03882305 cg07533148 cg06493386 cg24794531 cg16832407 cg07758904 cg08831594 cg03614513 cg12965512 cg18735146 cg14802310 cg07307078 cg04527918 cg14433673 cg07084163 cg25589890 cg10574499 cg09082921 cg17162024 cg22995176 cg09053680 cg20881054 cg24662718 cg03160135 cg12874092 cg18349835 cg12457773 cg00852964 cg26128092 cg20616414 cg01830294 cg00625653 cg09945801 cg18342279 cg25101936 cg14440934 cg12680609 cg14456683 cg06088032 cg24832140 cg08849574 cg16638540 cg21790626 cg09643544 cg10604333 cg01657207

Gene symbol TCTEX1D1 TDO2 TDRD5 TERT TESC TFAP2B TFAP2E TFCP2L1 TFEC TGFBR3 THRSP THSD3 TLX3 TM6SF1 TMEM25 TMEM48 TMEM54 TNFRSF10C TRAPPC3 TRIM36 TRIM50C TRIM58 TRPA1 TRPC1 TRPM3 TSCOT TSPAN15 TSPAN2 TSPAN8 TTC10 TUBA3 TUBB6 UCN UCP2 UGT3A2 ULBP1 UNQ2446 UNQ3045 UNQ9433 UPK3B UTF1 VASH1 VAV3 VGLL2 VIM VIPR2 VMP VNN1 WDR8 WNK2 WNT2 WNT7A WRN ZAR1 ZBTB16 ZDHHC1 ZFP41 ZIC1 ZMYND10 ZNF114 ZNF134 ZNF135 ZNF154 ZNF177 ZNF222 ZNF35

Mean normal b

Mean tumor b

SD (tumors)

P value

0.27 0.34 0.16 0.1 0.21 0.16 0.25 0.15 0.21 0.19 0.47 0.17 0.09 0.16 0.15 0.15 0.17 0.13 0.09 0.26 0.43 0.05 0.15 0.09 0.15 0.09 0.14 0.11 0.2 0.2 0.15 0.14 0.5 0.06 0.16 0.09 0.25 0.17 0.1 0.15 0.08 0.16 0.17 0.21 0.06 0.03 0.16 0.18 0.1 0.26 0.32 0.08 0.2 0.24 0.37 0.15 0.14 0.29 0.17 0.14 0.1 0.18 0.06 0.22 0.15 0.22

0.42 0.52 0.4 0.3 0.32 0.36 0.44 0.4 0.31 0.35 0.75 0.35 0.57 0.38 0.4 0.31 0.3 0.35 0.3 0.44 0.73 0.33 0.31 0.31 0.4 0.35 0.37 0.35 0.37 0.31 0.35 0.36 0.79 0.38 0.32 0.32 0.59 0.33 0.35 0.34 0.57 0.35 0.33 0.36 0.31 0.3 0.34 0.31 0.41 0.53 0.54 0.3 0.36 0.42 0.56 0.36 0.57 0.47 0.42 0.34 0.33 0.48 0.5 0.45 0.31 0.4

0.18 0.14 0.21 0.19 0.11 0.1 0.23 0.21 0.08 0.15 0.22 0.25 0.28 0.2 0.18 0.11 0.14 0.26 0.09 0.14 0.19 0.23 0.19 0.22 0.18 0.24 0.25 0.22 0.1 0.1 0.1 0.22 0.18 0.25 0.14 0.21 0.21 0.13 0.23 0.14 0.27 0.19 0.22 0.19 0.26 0.26 0.18 0.09 0.23 0.22 0.22 0.23 0.17 0.2 0.21 0.1 0.26 0.15 0.29 0.19 0.22 0.23 0.29 0.24 0.19 0.16

.019 .000 .002 .004 .004 .000 .028 .002 .000 .004 .000 .047 .000 .004 .000 .000 .007 .026 .000 .000 .000 .001 .024 .006 .000 .004 .014 .003 .000 .003 .000 .006 .000 .001 .001 .004 .000 .000 .003 .000 .000 .008 .049 .033 .009 .005 .005 .000 .000 .001 .004 .009 .012 .018 .015 .000 .000 .000 .023 .004 .004 .000 .000 .008 .022 .002

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e18

Supplementary Table 1. Continued Probe ID cg21168622 cg23037403 cg08341667 cg01184522 cg03975694 cg25886284 cg06458239 cg16014085 cg05221167 cg16731240 cg12259537 cg17892556 cg18267381 cg22598028 cg05508084 cg18440048 cg02440177 cg06656924

Gene symbol ZNF350 ZNF454 ZNF488 ZNF496 ZNF540 ZNF545 ZNF549 ZNF553 ZNF560 ZNF577 ZNF606 ZNF625 ZNF659 ZNF660 ZNF667 ZNF70 ZNF702 ZNF76

Mean normal b

Mean tumor b

SD (tumors)

P value

0.14 0.16 0.14 0.12 0.21 0.03 0.24 0.34 0.07 0.14 0.19 0.09 0.2 0.1 0.17 0.11 0.09 0.21

0.34 0.33 0.31 0.34 0.54 0.31 0.38 0.58 0.34 0.44 0.32 0.37 0.4 0.44 0.37 0.33 0.49 0.37

0.09 0.19 0.08 0.28 0.21 0.25 0.13 0.15 0.22 0.24 0.17 0.29 0.19 0.25 0.19 0.18 0.25 0.1

.000 .013 .000 .031 .000 .002 .006 .000 .001 .001 .031 .010 .003 .000 .003 .000 .000 .000

1435.e19 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 2. Genes Significantly Up-Regulated After Epigenetic Unmasking Gene ABCD2 ABCG4 ACBD7 ACER2 ACHE ACN9 ACSM3 ACTBL2 ACTL6B ADC AGAP2 AKAP6 ANKRD42 APBA1 APOD ARL6 ARMC2 ARRDC4 ASPHD2 ATCAY ATP1A3 ATP6V1C2 ATP8A1 BBS12 BCAS1 BCO2 BEX5 BHMT BIK C10orf78 C11orf71 C12orf60 C12orf72 C14orf174 C14orf37 C14orf79 C16orf45 C16orf93 C19orf30 C1orf38 C2orf51 C3orf25 C3orf34 C4orf33 C5orf27 C6orf225 C7orf57 C7orf58 C9orf43 C9orf72 CACNA1G CALB1 CCDC103 CCDC112 CCDC113 CCDC148 CCDC87 CCNA1 CCR4 CCR7 CDH15 CGREF1 CHST6 CKMT1A

Treated

Control

Raw P value

Q value

6.55 6.62 8.50 7.07 7.12 7.13 7.36 8.57 6.59 6.88 6.63 6.87 7.47 6.74 6.69 7.74 6.67 8.99 6.67 6.87 8.77 7.78 6.83 6.84 7.58 6.86 6.77 8.44 6.90 7.06 6.54 6.62 7.51 6.94 8.20 6.51 7.34 6.88 9.19 7.05 7.08 7.16 7.71 6.87 7.58 7.21 6.80 6.96 7.00 6.93 7.10 6.85 6.70 7.67 6.49 6.74 7.26 7.83 7.13 6.52 6.74 8.02 7.13 7.81

4.99 6.37 6.30 6.16 6.43 6.46 5.62 5.41 6.17 6.25 6.31 5.79 6.35 6.25 5.88 6.43 5.78 5.64 5.86 6.08 6.07 6.12 5.34 5.85 6.19 5.66 5.45 5.71 6.45 6.35 5.90 5.85 6.07 6.05 5.76 6.03 6.35 6.20 5.84 5.88 5.79 5.74 6.23 6.03 5.94 6.34 5.77 5.26 5.83 5.97 6.39 5.13 6.06 6.14 5.80 5.25 6.32 5.71 5.17 6.29 6.36 6.42 6.46 6.15

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00

.00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .01 .01 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .01 .01 .00 .01 .01 .00 .00 .00

Supplementary Table 2. Continued Gene CKMT1B CPEB3 CRISPLD2 CRYM CYP4F2 DEPDC7 DGKI DHDH DHRS2 DLL3 DLX1 DLX2 DNAJC12 DOC2A DPF1 DPY19L2P2 DSEL EAF2 EIF4E3 ELMO1 ELOVL3 ELOVL4 EPHB3 FAM154B FAM46C FAM49A FAR1 FBXO27 FCHO1 FLJ30679 FRG2 FRG2B FYN GABRB3 GALNT6 GDPD1 GFPT2 GMPR GOLGA6B GOLGA6C GOLGA6D GPR4 GPR50 GPR63 GUCY1B3 HAP1 HERC5 HEY1 HEY2 HIST1H1T HIST3H2BB HPCAL4 HPSE HTR7P HUS1B IQCD ITGA7 ITIH4 ITIH5 JAM3 JPH1 KAL1 KCNN1 KIAA0319 KIF1A KIF5C

Treated

Control

Raw P value

Q value

7.81 6.77 8.72 7.38 6.63 7.14 6.67 6.59 9.83 6.93 6.81 7.13 6.98 6.81 7.07 7.47 7.59 6.84 7.21 6.50 7.34 8.03 6.79 7.39 7.98 6.96 6.53 6.93 7.52 7.35 6.98 6.98 8.29 6.68 6.63 7.08 8.48 6.79 7.03 7.03 7.03 6.99 7.83 6.78 7.65 7.47 7.52 6.75 6.88 6.71 7.20 6.97 8.12 6.89 6.67 6.61 7.08 6.94 7.21 6.81 6.91 6.53 6.84 7.35 7.49 8.78

6.15 6.02 5.64 6.31 5.68 6.35 5.61 6.25 6.09 6.25 6.31 6.25 6.05 6.42 6.25 6.24 5.91 5.83 6.39 5.54 5.68 5.53 6.19 5.92 6.31 5.92 5.22 6.43 6.40 5.93 5.35 5.35 6.25 6.37 6.19 5.97 6.32 5.56 6.04 6.04 6.04 6.39 5.70 5.63 6.29 6.14 5.61 5.97 5.53 5.14 6.15 5.55 6.25 6.47 6.36 5.83 6.21 6.49 6.25 5.78 5.98 4.76 6.42 6.23 6.35 5.92

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .01 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e20

Supplementary Table 2. Continued

Supplementary Table 2. Continued

Gene

Treated

Control

Raw P value

Q value

Gene

Treated

Control

Raw P value

Q value

KRT20 LAMP3 LBH LDHB LETM2 LGI3 LOC147727 LOC392196 LOC645752 LPPR4 LRAT LRRC46 LRRC50 LYPD5 LZTFL1 MAK MAP1A MAP2 MAPRE3 MAPT MEIS3P1 MGC42105 MLF1 MPP7 NAP1L3 NAT1 NCAN NCRNA00087 NCRNA00094 NCRNA00174 NEFH NEFL NEFM NELL2 NKX2-3 NMNAT2 NPHP1 NR4A2 NR4A3 NRIP3 ODZ1 PAIP2B PCDH10 PDLIM3 PGBD5 PHOSPHO1 PLAC1 PLCG2 POLR3GL POTEC POTEG POTEH PP1MN PPFIA3 PPM1E PRAME PREX1 PRKAR2B PRKCG PRND PRPH PRSS35 PTGS1 RASGRP2 RCAN2 REPS2

7.70 8.76 6.85 7.09 6.95 6.87 7.12 8.49 7.15 6.80 7.50 6.88 6.76 6.65 7.16 6.56 7.66 8.26 8.20 7.29 6.93 6.52 7.09 7.42 7.21 6.66 6.67 6.84 6.75 6.87 8.69 7.13 7.30 6.83 6.59 6.73 6.64 7.42 6.71 8.33 7.97 8.47 7.57 6.68 6.50 7.02 6.75 7.99 7.16 6.52 7.22 7.06 6.95 6.71 6.60 6.91 7.16 8.20 6.88 6.84 6.89 6.76 6.58 6.68 7.15 7.24

5.74 6.37 5.98 5.44 6.15 6.30 6.26 5.96 5.97 5.70 5.79 5.93 5.89 6.24 6.49 4.93 6.30 5.70 6.46 6.16 6.44 5.87 5.17 5.41 5.22 5.49 6.36 6.33 6.30 6.28 5.61 6.03 5.16 5.24 6.45 5.89 5.60 6.38 5.73 6.16 5.73 6.43 5.96 5.98 5.81 6.15 5.77 6.29 6.37 5.29 5.91 5.48 6.08 6.35 5.26 5.95 6.20 5.52 6.15 5.85 6.29 5.60 5.64 6.45 6.19 6.17

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00

.00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .01 .00 .01 .00 .00 .01 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .01 .00 .01

RFPL4A RFPL4B RNF150 ROPN1L RTBDN RUNDC2A RUNDC2C RUNDC3A RUNDC3B SAMD5 SCAND3 SH3BP5 SHD SILV SLC2A3 SLC45A1 SLC46A3 SLC4A8 SLC8A2 SMPD3 SNORA14A ST6GALNAC3 STARD5 STAT4 SYP SYT11 SYTL3 TEC TEX19 TEX9 TMEFF2 TMEM86A TMOD2 TNFRSF1B TNNI3 TP63 TRAM1L1 TRDMT1 ULBP2 ULK4 UNC13A USP17 USP17L2 USP17L6P USP2 UTS2D VWA5B2 WDR31 WDR65 WFDC2 ZC3H10 ZCCHC12 ZNF18 ZNF347 ZNF571 ZNF699 ZNF860 ZSCAN5B

9.63 7.43 6.79 7.66 6.96 7.49 7.12 7.16 7.48 6.95 7.07 6.93 6.58 8.97 8.39 6.64 7.06 6.68 6.78 6.90 6.98 7.07 6.79 7.07 7.39 9.32 6.77 7.13 6.70 7.25 7.24 6.90 7.67 6.72 7.38 7.29 6.72 6.86 7.70 7.21 6.78 7.87 7.83 7.93 7.14 7.52 6.72 7.05 6.53 6.57 6.96 8.06 6.76 7.18 6.67 6.93 7.14 7.16

5.05 5.38 6.24 5.66 6.10 6.47 6.32 5.95 5.69 6.11 6.45 6.37 6.32 5.36 5.97 6.14 5.73 4.88 6.33 6.01 6.39 5.77 5.86 5.80 6.11 6.02 5.59 5.88 6.01 5.70 5.34 6.07 6.01 6.30 6.47 5.48 5.77 6.11 6.17 6.25 5.57 6.13 6.14 6.14 6.33 4.27 6.26 6.20 5.38 6.40 6.29 5.09 6.38 6.40 6.14 6.00 5.90 4.67

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00

.00 .01 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .01 .00 .00 .00 .01 .01 .00 .00 .01

1435.e21 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 3. Candidate Gene Selection Methylation Mean Probe ID

Gene symbol

Normal (b)

Tumor (b)

SD

cg08572611 cg03996793 cg06277657 cg15236866 cg13297865 cg03243946 cg23303074 cg20182358 cg02994956 cg00116838 cg09595479 cg22123464 cg15201635

ACTL6B C19orf30 DGKI DLX1 ELOVL4 LDHB LRAT MLF1 NEFH PPM1N (FLJ40125) PRPH SLC8A2 SMPD3

0.11 0.15 0.2 0.18 0.1 0.2 0.16 0.18 0.4 0.15 0.14 0.23 0.13

0.46 0.31 0.46 0.32 0.35 0.4 0.34 0.43 0.63 0.37 0.44 0.36 0.32

0.26 0.11 0.23 0.12 0.24 0.23 0.19 0.21 0.16 0.18 0.25 0.17 0.2

Treated

Control

Raw P value

6.59 9.19 6.67 6.81 8.03 7.09 7.5 7.09 8.69 6.95 6.89 6.78 6.9

6.17 5.84 5.61 6.31 5.53 5.44 5.79 5.17 5.61 6.08 6.29 6.33 6.01

P value 0 0 0 0 0 .02 .01 0 0 0 0 .03 .01

Expression (collapsed probe) Entrez gene ID ACTL6B C19orf30 DGKI DLX1 ELOVL4 LDHB LRAT MLF1 NEFH PPM1N (FLJ40125) PRPH SLC8A2 SMPD3

0 0 0 0 0 0 0 0 0 0 0 0 0

Q value .01 0 0 0 0 0 .01 0 0 0 0 0 0

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e22

Supplementary Table 4. Pathway Analysis of Candidate Tumor Suppressor Genes DAVID analysis GO term

Ion binding SLC8A2 PPM1N DGKI SMPD3 Amyptrophic lateral sclerosis susceptibility NEFH PRPH Intermediate filament NEFH PRPH

Disease

Function

Gene ontology

Secondary metabolic process LDHB LRAT

Transmembrane region LRAT SLC8A2 ELOVL4 SMPD3

Developmental DLX1 MLF1 SMPD3 Regulation of metabolic processes LRAT SMPD3

Enrichr analysis ChEA Name

P value

TRIM28-17542650 (human) SUZ12-18692474 (mouse) RNF2-16625203 (mouse) EED-16625203 (mouse) MTF2-20144788 (mouse) JARID2-20075857 (mouse)

.2858 6.28E-05 .000857 .002422 6.28E-05 .000857

Z score 39.12 1.66 2.12 2.2 1.17 1.6

Combined score 49 16.04 15 13.23 11.32 11.31

Genes LDHB, DLX1, DGKI LRAT, DLX1, NEFH, LRAT, DLX1, NEFH, LRAT, DLX1, NEFH, LRAT, DLX1, NEFH, LRAT, DLX1, PRPH,

DGKI, ELOVL4, SMPD3, MLF1 ELOVL4, MLF1 ELOVL4 PRPH, DGKI, ELOVL4, SMPD3, MLF1 DGKI, ELOVL4

Histone modifications ChIP seq Name H3K4me1 adult liver

Genes .006418

1.55

7.81

DLX1, NEFH, DGKI, ELOVL4, SMPD3, MLF1

1435.e23 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 5. Interquartile Range Log Rank Test on Training Set From HCC Genomic Consortium P (log rank)

Percentile 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75

.094 .094 .012 .012 .113 .113 .019 .019 .011 .07 .013

Supplementary Table 6. Clinical Variables of High and Low SMPD3 Expression N ¼ 241 BCLC stage 0 A B C Tumor size, cm 2 >2 Degree of differentiationa Well Moderate Poor Satellitesa Absent Present Vascular invasion Absent Present

High-expressing SMPD3 n ¼ 83 (%) 10 62 7 1

(12.5) (77.5) (8.75) (1.25)

Low-expressing SMPD3 n ¼ 158 (%) (9) (81.5) (4.8) (4.8)

.267

18 (21.7) 63 (76)

24 (15.2) 123 (77.8)

.288

11 (21.8) 33 (62.3) 9 (17)

33 (27.7) 63 (53) 23 (19.3)

.524

51 (77.3) 11 (16.7)

84 (73) 23 (20)

.691

47 (60.3) 31 (39.7)

82 (5) 62 (43)

.671

BCLC, Barcelona Clinic Liver Cancer a Missing data >10% for degree of differentiation and satellites.

13 119 7 7

P value

December 2013

NOVEL TUMOR SUPPRESSOR GENES IN LIVER CANCER 1435.e24

Supplementary Table 7. Comparison Between Neumann and Revill Candidates Neumann et al25 Functional analysis Transient re-expression in Huh7 cells PER3 Clonogenicity Cell viability Apoptosis IGFALS Clonogenicity Cell viability Apoptosis Migration Transient re-expression in PLC/PRF/5 cells PROZ Clonogenicity Cell viability Apoptosis Migration siRNA knockdown in SNU387 cells PER3 Cell viability Apoptosis

Fold change from mock transfected 0.52 0.42 2.39

P value <.01 <.01 <.05 <.01

0.58 0.67 2.6 1.14

>.05

1.31 1.28 0.81 0.87

>.01 >.01 <.05 >.05

1.27 0.41

<.01 <.05

Revill et al Stable inducible expression in JHH-7 cells Cell proliferation in tetracycline induced vs uninduced cells SMPD3 48 h 72 h 96 h NEFH 48 h 72 h 96 h LACZ 48 h 72 h 96 h shRNA knockdown in PHMI cells and in vivo tumorigenesis SMPD3 Migration Invasion NEFH Migration Invasion

Fold change untreated cells 0.737525004 0.727368598 0.482848233

.695365626 .055475765 .00132306

0.730318519 0.844919786 0.810004675

.010105321 .060659146 .029617225

0.881141389 0.966688348 0.972716487

.157796293 .59171098 .567495153

Fold change from nontargeting shRNA 1.795389049 11.63157895 1.322766571 0.394736842

8.15369E-05 .001067676 .001935858 .360375333

1435.e25 REVILL ET AL

GASTROENTEROLOGY Vol. 145, No. 6

Supplementary Table 8. Methylation-Specific Pyrosequencing Primers SMPD3 NEFH DGKI LDHB PRPH ACTL6B GSTP1a a

Forward

Reverse

Sequencing

50 -GGGTTTTGAGGTAGAAATTAAGGT-30 50 -AAGTTTATTATGGTTTGAGTAGG-30 50 -TTGGGGTTGAGGAGGATAGTTAAATTT-30 50 -[BTN]TGGGGGAGGGAGTGTGTATA-30 50 -TTTTGGAGGGTGGGGTTAAAT-30 50 -AGTTGTTTTAAAGGGGAAGATATATT-30 50 -TTTGGGAAAGAGGGAAAGGT-30

50 -ACCCCCCCAAAAATCTACCCATTAA[BTN]-30 50 -CCTAATAACTACAATCTTCCATTACTC[BTN]-3’ 50 -CACCTTAATCCTAAACCCAACT[BTN]-30 50 -CAACTACTACCCTCTACCTTCTACT-30 50 -CCCTAACTAACCAACTCCTCAATAT[BTN]-30 50 -ACCTCATCAATCCTCCTACAAAATCCTAAT[BTN]-30 50 -TACTAAAAACTCTAAACCCCATC-30

50 -GTTTTTAGTGTATTTTGGGA-30 50 -ATGGTTTGAGTAGGTG-30 50 -GTTAAATTTTGTAGATATTATAAGT-30 50 -AACTCTAAAAACCTCTATAAC-30 50 -AGGGAGAGATTAATGG-30 50 -TTTAAAGGGGAAGATATATTTTATT-30

Combined bisulfite restriction analysis primers.

Supplementary Table 9. Real-Time Polymerase Chain Reaction Primers Forward DGKI NEFH LDHB ACTL6B SMPD3 GSTP1

0

Reverse 0

5 -ATTTACCATGGCCTCTTTGG-3 50 -GAGGAGTGGTTCCGAGTGAG-30 50 -AGCTGCCATGGATGGATTTT-30 50 -AGAATGGCATGATCGAGGAC-30 50 -CCATCGGTACTCTGCTGGA-30 50 -CGCACCCTTGGGCTCTAT-30

0

5 -GGACAGACTGGGGATCATTG-30 50 -TTCCTGGTAGGAGGCAATGT-30 50 -GGCACTTTCAACCACCATCT-30 50 -CGAACATCAGCTCTGTCAGC-30 50 -ACCTCGGCCTTCCAGTCT-30 50 -CCCGCCTCATAGTTGGTGTA-30