Investigation in Liver Tissues and Cell Lines of the Transcription of 13 Genes Mapping to the 16Q24 Region That Are Frequently Deleted in Hepatocellular Carcinoma

3178 Vol. 8, 3178–3186, October 2002 Clinical Cancer Research

Investigation in Liver Tissues and Cell Lines of the Transcription of 13 Genes Mapping to the 16q24 Region That Are Frequently Deleted in Hepatocellular Carcinoma

Philippe Riou, Raphae¨l Saffroy, Je´rome Comoy, INTRODUCTION Marine Gross-Goupil, Jean-Paul Thie´ry, HCC2 is one of the most frequent human cancers world- Jean-François Emile, Daniel Azoulay, wide (1). However, the molecular mechanisms underlying HCC Dominique Piatier-Tonneau, Antoinette Lemoine, tumorigenesis and tumor metastasis are still poorly understood. 1 HCC, like other solid tumors, seems to develop following mul- and Brigitte Debuire tiple genetic events (2), including the functional inactivation of Service de Biochimie et Biologie molećulaire [P. R., R. S., J. C., TSGs and the activation of oncogenes. Previous studies of LOH M. G-G., J-P. T., A. L., B. D.], Service d’Anatomie pathologique have suggested that several genomic regions may be involved in [J-F.E.], and Centre He´patobiliaire Hoˆpital Universitaire Paul Brousse [D. A.], UPRES 1596-Faculte´deMe´decine Paris-Sud, 94804 Villejuif liver carcinogenesis. These regions are mostly located on chro- Cedex, France, and Geńe´tique Molećulaire et Biologie du mosome arms 1p, 4q, 5q, 6q, 8p, 9p, 11p, 13q, 16p, 16q, and De´veloppement, FRE 2376, Centre National de la Recherche 17p, indicating that dysfunctions of diverse tumor or metastasis Scientifique, 94801 Villejuif Cedex, France [D. P-T.] suppressor genes located on these chromosomes are involved in the development of HCC (3–5). ABSTRACT LOH on the long arm of chromosome 16 has been reported to be frequent in several human cancers, including HCC. Anal- Many studies have associated chromosomal deletions in yses of LOH frequencies (6–12), comparative genomic hybrid- the 16q24 region with human cancers, including hepatocel- ization (13–15), and aberrant DNA methylation (16, 17) have lular carcinoma. A more limited region around the micro- suggested that one or more TSGs involved in the development satellite D16S402 has been shown implicated in the meta- of HCC map to chromosome 16q. LOH on chromosome 16q has static spread of hepatocellular carcinoma, prostate cancer, been correlated with clinicopathological characteristics, such as and Wilms’ tumors. It is likely that one or more tumor the degree of differentiation, size, and the occurrence of metas- suppressor genes are located in this 16q24 area. tases, indicating that LOH on chromosome 16 may be involved We used SYBR Green reagents to quantify, by real- in the progression of HCC (6, 12). time quantitative RT-PCR, the production of mRNA for 13 Subsequent molecular analyses in HCC (5, 10–12, 15, genes mapping to 16q24. The locations of these genes were 18–20), prostate (21, 22), ovarian (23–25), breast (26, 27), determined from published human genome sequencing data. bladder (28), and Wilms’ (29–32) tumors have identified the We studied mRNA levels in 10 liver tumor tissues, 10 non- 16q24 region as a major region of LOH, associated with meta- tumor liver tissues, five hepatoma cell lines, and in isolated static and aggressive behavior of the cancer (21, 22, 29, 32–34). hepatocytes. Results were compared with those for loss of Mashimo et al. (33) recently used microcell-mediated chromo- heterozygosity observed in the D16S402 region and recur- some transfer into a highly metastatic rat prostatic cancer cell rence. line to show that microsatellite marker D16S402 was retained in No down-regulation was observed in tumor tissues. microcell hybrids displaying significant reduction in lung me- Two genes were consistently overexpressed: OKL38 and tastasis. We have shown that changes in the D16S402 micro- CDH13. CDH13, which functions in cell-cell adhesion, seems satellite are frequent in European HCC patients and are corre- to be involved in liver carcinogenesis. However, no relation- lated with a higher risk of recurrence (35). LOH studies in ship was observed between the expression of this gene and several other types of tumor have also suggested that there is at changes in the D16S402 microsatellite or tumor recurrence. least one TSG near the D16S402 locus (19, 23, 26, 31, 36). None of the other genes tested seemed to be a good candidate Data from the human genome sequencing program were for a major tumor suppressor gene in liver carcinogenesis. recently published (37), and information concerning the fine Our results suggest that additional unknown genes in- mapping of genes is now available on via the internet.3 We volved in carcinogenesis are located in the 16q24 area. identified 13 genes mapping to the 16q23.3–24.1 chromosomal region between the D16S422 and D16S3037 loci, encompassing a region of approximately 2 cM, including D16S402 (Fig. 1). We selected only known genes and full-length mRNAs corre- Received 10/19/01; revised 4/22/02; accepted 6/4/02. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom requests for reprints should be addressed, at Service de 2 The abbreviations used are: HCC, hepatocellular carcinoma; LOH, Biochimie et Biologie molećulaire, Hoˆpital Universitaire Paul Brousse, loss of heterozygosity; TSG, tumor suppressor gene; NL, normal liver UPRES 1596-Faculte´deMe´decine Paris-Sud, 14 avenue Paul Vaillant RNA; NTL, nontumor liver RNA; TL, tumoral liver RNA. Couturier, 94804 Villejuif Cedex, France. 3 Internet address: www.genome.cse.ucsc.edu.

Fig. 1 Genetic map and known function of the 13 genes mapping to the 16q23.3–24.1 chromosomal region.

sponding to unknown genes for study. The expression of these nation by the pathologist and were then frozen and stored at genes has not previously been studied in patients with HCC. Ϫ80°C for RNA and DNA extraction. Particular attention was We quantified the expression of these genes in tumoral and paid to obtain the “core” part of the tumor to avoid the adjacent nontumoral tissues from the liver and in five hepatoma cell noncancerous tissue, as proved by histopathological examina- lines. We compared the results obtained for gene expression tion. The clinicopathological features of the HCCs are summa- with those concerning LOH specific to D16S402 and surround- rized in Table 1. For each patient, we collected paired peripheral ing microsatellite markers and metastatic spread. blood samples at a time when the patient was not undergoing surgery, to prevent contamination of the blood with hepatocytes, MATERIALS AND METHODS which may occur during surgery (38), or specimens of non- Tissue and Cell Specimens. The Institutional Review hepatic tissue—such as gallbladder, if possible—for use as a Board of the Hoˆpital Universitaire Paul Brousse and the Com- normal control for microsatellite analysis. Ten NL tissue sam- mittee for Research on Human Subjects at the Faculte´ de Me´- ples were obtained from biopsies of donated livers during graft decine Paris-Sud approved this research. Liver tumor tissue and harvesting. These samples were treated in the same way as HCC noncancerous liver tissues were excised during surgery from 10 samples and were used for histological analysis to determine the patients with HCC. The tissues underwent histological exami- normal expression profile of the genes studied. Human liver

Table 1 Characteristics of the 10 patients Patient Age Underlying liver Etiology Recurrencea No. of nodules Diameter of nodulesb Edmonson’s grade 166Cc HCV ϩϾ325II 2 66 C HEMO ϩϾ3 10 III 318NC0 0 Ͼ340II 448NC0 0 Ͼ3 220 II/III 5 64 C HCV ϩϾ3 10 III 6 39 C HBV ϩ HDV 0 Յ320II 7 45 C HBV ϩ Յ310II 8 75 CH HBV ϩ Յ3 60 II/III 9 45 C HCV 0 Յ318II 10 64 NC 0 0 Յ3 130 III a Within 15 months after curative surgery. b Diameter of the greatest nodule (millimeters). c C, cirrhosis; NC, without cirrhosis; CH, chronic hepatitis; HEMO, hemochromatosis; HCV, hepatitis C virus; HBV, hepatitis B virus; HDV, hepatitis D virus.

Table 2 Oligonucleotide primer sequences used Priming site GenBank accession no. Upstream (5Ј–3Ј) Downstream (5Ј–3Ј) Microsatellite markers D16S422 AFM249xc5 CAGTGTAACCTGGGGGC CTTTCGATTAGTTTAGCAGAATGAG D16S3091 AFMb297zc1 GGGAGATAGCCTTAAACTTTCTTAC TGTTGCTAATAACACTAGGCCA D16S402 AFM031xa5 TTTTGTAACCATGTACCCCC ATTTATAGGGCCATGACCAG WFDC1-GT12 AC009123 TGACTGTGTCCGCTAGAGTG TACGCACGCATCCCC WFDC1-GT20 AC009123 CCTGTCTTCGTAAAGGG TCAAATCGTTCATTTGGGAG D16S3061 AFMa304wa9 CTACTGGTGAGGCTGAGGTG ATATCTCGGGATTTGTTGCTTTAC D16S3037 AFMa191ya9 GAGCCAAGATGACACCACT GCACTGGGAACCTAAGGA Target genes ␤-actin M10277 AGCCTCGCCTTTGCCGA GCGCGGCGATATCATCATC AK022605 AK022605 GCAGCTTCTCACGGGACAA GGACTGTACGTCTCTACTGAAGAGAGC LOC083693 AK025626 CACAGGTGCTCGTGGTTGG CCCACACCCAGAGCCATTC CDH13 NM 018110 GGCGCTTCTAGTCGGACAAA GCCATCGCTGTTCACCTTG HSBP1 NM 001537 CGCCAAGCTGGGCATC ATTCTCCCAATGATCTGGTCAG HUMCLPB AK026033 CAAGAGGTCCAAGTTTGCCCT CTGGGCGTCGTAATTGGCT KIAA0190 D80012 TGTGGCACACAGGCTGT TCCAAAGCGAGGGCAG KIAA0703 NM 014861 CGGAGACCGGATCCCTG GATGTTGCTGAGGGTGGTGAG KIAA1609 AB046829 CACAGACAGGAGCTGAGAGGCT GACCAGGGTAGTCAGATCAAGAGAC MLYCD NM 012213 TGCGCTTCCTGGTGCAG GCCAGGTAACCCGTTCTAGGT OKL38 NM 013370 GTGATCCTGAGCCAAGGCC CCAGACCCTTCTTGACCACGT S1P NM 003791 CAGTTGTTGAAAACGTCCCCAT GTGTACTGGAGGAGGGCATCC TAF1C NM 005679 CTGATGCTGGAGAATTTCAAGCT GACGTGCCCAGGATGGG WFDC1 NM 021197 GCTATGAGTGCCACATCCTGAG GTTGTCCCCTTCCAGGTTCTG

tumor cell lines HepG2, PLC/PRF/C, TONG, HA22TNGH and primers were chosen from the human genome database (Table MAHLAVU were obtained from the American Type Culture 2). PCR was performed as described previously (35, 40). The Collection. Cells were grown in Dulbecco F12–RPMI 1640 PCR products were separated by capillary electrophoresis, (Life Technologies, Inc., Cergy Pontoise, France), containing using an ABI PRISM 310 machine (Applied Biosystems, 10% FCS and gentamycine, and harvested at 70% confluence. Foster City, CA). Human hepatocytes were isolated by collagenase digestion from Real-time Quantitative PCR. First-strand cDNA was NLs from three cadaveric multiple organ donors, as reported synthesized in a volume of 80 ␮l, with TaqMan reverse tran- previously (39). scription reagents (Applied Biosystems). We mixed 10ϫ re- ␮ ␮ DNA Extraction. DNA was isolated by a standard pro- verse transcriptase buffer (8 l),5mM MgCl2, 500 M of each teinase K digestion and phenol/chloroform extraction procedure. dNTP, 2.5 ␮M random hexamers, 32 units of RNase inhibitor, RNA Extraction. RNA was extracted from 20 mg of and 100 units of Multiscribe reverse transcriptase with 4 ␮gof tissue or 106 cells with a QIAamp Tissue kit (Qiagen, Courta- total RNA. The cycling conditions were 10 min at 25°C, 30 min boeuf, France) according to the manufacturer’s instructions. A at 48°C, and 5 min at 95°C. Reactions in which the enzyme was final elution volume of 50 ␮l was used. omitted were used as negative controls. Analysis of Microsatellite Alterations. Seven microsat- PCR was performed with an ABI PRISM 7700 Sequence ellite markers (D16S422, D16S3091, D16S402, WFDC1-GT12, Detector and SYBR Green reagents (Applied Biosystems). Spe- WFDC1-GT20, D16S3061, and D16S3037) spanning the chro- cific primers for human ␤-actin, LOC083693, AK022605, mosomal region 16q23.3–24.1 and their corresponding PCR CDH13, HSBP1, HUMCLPB, KIAA0190, KIAA0703, KIAA1609,

MLYCD, OKL38, S1P, TAF1C, and WFDC1 were designed to Table 3 Gene expression in TL, NTL, and NL samples work in the same cycling conditions (50°C for 2 min to permit TL NTL NL uracil N-glycosylase cleavage, 95°C for 10 min, followed by 50 a cycles of 95°C for 15 s, and 60°C for 1 min). The specificity of Gene Mean SD Mean SD Mean SD the nucleotide sequences chosen was confirmed by conducting CDH13 4.92 3.60 0.95 1.23 1.00 0.81b basic local alignment search tool searches. To prevent the am- HSBP1 2.25 2.05 0.80 0.4 1.00 0.95 MLYCD 2.13 1.38 1.28 0.62 1.00 0.43 plification of contaminating genomic DNA, the two primers OKL38 5.60 4.53 2.87 2.05 1.00 0.74b,c bound to different exons. TaqMan RNase control reagents (Ap- S1P 0.93 0.49 0.51 0.35 1.00 0.79 plied Biosystems), designed for ribosomal S18 RNA amplifica- LOC083693 4.73 10.0 0.55 1.00 1.00 1.18 tion, were used as a reference to normalize the results. We used TAF1C 1.22 1.28 1.10 0.41 1.00 0.74 ␮ WFDC1 1.88 2.41 0.84 0.47 1.00 0.52 2.5 l of the reverse transcriptase product for PCR in a final KIAA0703 1.36 1.98 1.15 0.63 1.00 0.76 volume of 25 ␮l. For each PCR, a standard curve was produced, KIAA1609 3.00 3.38 1.25 0.75 1.00 0.50 using four 1:10 dilutions of a positive sample. All samples were HUMCLPB 1.84 2.23 0.86 0.43 1.00 0.95 run in triplicate. The relative amounts of mRNA for each tested AK022605 1.73 1.00 0.93 0.53 1.00 0.67 KIAA0190 2.21 1.71 1.00 0.51 1.00 0.92d gene in the samples were calculated by comparison with stand- Control ␤-actin 1.66 0.87 1.01 0.74 1.00 0.59 ard curves. For each sample, results were normalized, using the a S18 RNA value of the calibrator to obtain a final R-gene value. Gene expression was normalized, using the S18 RNA value of the calibrator to obtain a R-gene value. Each final result for the 13 genes and All samples were resolved in a 1.8% agarose gel to confirm the ␤-actin (used as control) expression were homogenized, using the the specificity of the PCR. R-gene mean obtained in NL. CDH13 Sequencing Analysis. PCR was performed in b Significant difference (P Ͻ 0.05) between TL and NL. c Ͻ coding exons (exons 1–14) and their surrounding regions using Significant difference (P 0.05) between NTL and NL. d Significant difference (P Ͻ 0.05) between TL and NTL. oligonucleotides as previously described (41). PCR reactions were carried out in a 50 ␮l volume containing 0.5 ␮M of each primer, 50 ␮M of each dNTP, 2.5 ␮l of formamide, and 2 units of Taq (Q.biogene, Illkirch, France) using a gene Amp PCR observed only between TL and NTL (P Ͻ 0.05). No signif- system 2400 (Applied Biosystems) and using the following icant up- or down-regulation was observed for the other conditions : (a)94°C (4 min); (b) 40 cycles of 92°C (1 min), genes. No significant difference was observed for ␤-actin, 58°C (1 min), 72°C (1 min); and a final extension step at 72°C used as a control. (7 min). Sequencing was performed on both strands using the R-gene values were determined for three independent prep- ABI Prism dichloro-rhodamine terminator cycle sequencing arations of isolated normal human hepatocytes and for five HCC ready reaction kit (Applied Biosystems) after purification of the cell lines (Table 4). Except for the HSBP1 gene, expression PCR products using the Qiaquick PCR purification kit (Qiagen). levels tended to be lower in isolated hepatocytes than in NL The sequences were analyzed on an ABI 310 automated se- samples. We compared mRNA levels in isolated hepatocyte quencer unit (Applied Biosystems). preparations and cell lines and found no significant difference in Statistical Analysis. Statistical analysis was performed expression level (Ͻ10-fold) between any of the cell lines tested, using the SAS software system (SAS Institute Inc., Cary, NC), for ␤-actin, HSBP1, MLYCD, OKL38 and TAF1C. In contrast, and the Student’s t test was used to determine the statistical overexpression (expression level more than 10 times that significance of differences of gene expression data according to in isolated hepatocytes) was observed for one or several of the samples analyzed. the following genes: CDH13, S1P, LOC083693, KIAA0703, KIAA1609, HUMCLPB, AK022605, and KIAA0190. Diverse associations were observed between the overexpressed genes, RESULTS depending on the cell line, with each cell line displaying a Analysis of Gene Expression in Patients’ Samples and specific gene expression pattern. Strong overexpression was Cell Lines. We assessed the expression of genes by evaluating observed for KIAA1609, with mRNA levels Ͼ50 and Ͼ100 RNA levels in 10 healthy liver tissues to calculate R-gene values times higher in HA22T and MAHLAVU cell lines, respectively, for each gene tested, as described in “Materials and Methods.” than normal. The MAHLAVU cell line displayed levels of The cutoff point for changes in gene expression, as determined CDH13 gene expression more than 300 times higher than nor- from RNA levels, was set at 3 SDs above or below the mean mal, whereas this gene was not expressed in the other cell lines. value for the genes in NL tissues. A 10-fold difference in R-gene Similarly, the WFDC1 gene, which was weakly expressed in values between samples was the maximum observed (for isolated hepatocytes and four hepatoma cell lines, was not HSBP1 and CDH13 genes). Mean R-gene values in NL, were expressed in HepG2. then compared with the R-gene values of 10 matched NTLs and Relationship between LOH at 16q23.3–24.1 and Gene TLs (Table 3). Three genes (CDH13, OKL38, and KIAA0190) Transcription. LOH determination for the 16q23.3–24.1 re- displayed significantly higher levels of expression in liver sam- gion is summarized in Fig. 2. Four of 10 patients (patients 4, 5, ples from patients with HCC than in NL samples. Significantly 7, and 8) exhibited LOH in this chomosomal region in TL higher levels of CDH13 (P Ͻ 0.05) and OKL38 (P Ͻ 0.01) samples. This LOH was also present in the NTL of patients 4 expression were observed in TL than in NL samples. For and 5. No tumor sample displayed down-regulation of expres- OKL38, expression levels were also higher in NTL adjacent to sion of the genes investigated. LOH of microsatellite loci within TL (P Ͻ 0.02). For KIAA0190, a significant difference was or close to each gene was not related to variations in gene

Table 4 Gene expression in isolated hepatocytes and cell lines Genea Isolated hepatocytes HA22TNGH HepG2 PLC/PRF/C MAHLAVU TONG CDH13 0.05 0 0 0 16.6 0 HSBP1 1.1 6.6 3.4 2.8 7.6 2.4 MLYCD 0.5 0.1 0.4 0.4 0.9 0.2 OKL38 0.7 0.3 1.2 0.3 0.9 0.4 S1P 0.1 1.8 1.2 0.7 1.6 0.6 LOC083693 0.1 2.2 1.7 0.4 3 0.4 TAF1C 0.2 0.9 0.7 0.5 1.1 0.4 WFDC1 0.09 0.01 0 0.02 0.09 0.01 KIAA0703 0.2 2 1.1 5.2 2.1 5.3 KIAA1609 0.2 27.4 3.1 2.8 10.6 4.1 HUMCLPB 0.1 0.2 1.2 0.7 0.4 1.8 AK022605 0.1 3 1.3 0.7 1.6 6.8 KIAA0190 0.4 5.1 4.4 1.6 4.2 2.1 Control ␤-actin 0.6 5.8 0.9 1.3 2.5 2.9 a Gene expression was normalized, using the S18 RNA value of the calibrator to obtain a R-gene value. Each final result for the 13 genes and the ␤-actin (used as control) expression were homogenized, using the R-gene mean obtained in NL.

Fig. 2 Gene expression levels and LOH analysis for the 10 patients studied. ؉, gene expression Ͼ3 SD; n, gene expression Յ3 SD; Ⅺ, retention of heterozygosity; u, not informative; f, LOH.

expression. The OKL38 and CDH13 genes were overexpressed We also checked for LOH in 16q23.3–24.1 microsatellite in four of six patients without LOH. Three of the four patients markers in the five cell lines (Fig. 3). The presence of one allele with LOH displayed OKL38 overexpression whereas only one in a cell line may be indicative of this allele being uninformative displayed CDH13 overexpression. In the nontumor sample from or of LOH, but the presence of both alleles eliminates the patient 4 that displayed LOH, CDH13 and OKL38 were over- possibility of LOH. PLC/PRF/C, TONG, HA22TNGH, and expressed (data not shown). MAHLAVU cell lines displayed one allele for all of the mic-

DISCUSSION Many studies have shown that chromosomal deletions at the 16q24 region are associated with human cancers, including HCC. A more limited region around the microsatellite D16S402 has been implicated in the metastatic spread of HCC (19, 35), prostate cancer (33), and Wilms’ tumors (31). There is probably at least one TSG in this 16q24 area. However, no TSG involved in liver carcinogenesis has been identified in this chromosomal region. The inactivation of a TSG requires complete loss or inactivation of both alleles. This occurs generally by mutation of one allele and deletion of the other. This mechanism, initially proposed by Knudson (42) for the retinoblastoma gene, has been confirmed for a large variety of genes and was usually associated with the lack of or down-expression of the gene. The aim Fig. 3 Allele numbers in the cell lines. of this study was to use quantitative RT-PCR on total RNA isolated from human liver tissues to identify known sequences mapping to this chromosomal region that were down-regulated. To assess the respective contributions of tumor and stromal cells rosatellite markers tested, which is in favor of the existence of to this phenomenon, gene expression was also studied on total LOH at the 16q23.3–24.1 region, taking into account that the RNA isolated from isolated normal hepatocytes and from hep- seven microsatellite markers tested have a 50–80% probability atocellular carcinoma cell lines. of being informative. However, the MAHLAVU cell line We studied 13 genes mapping to 16q23.3–24.1. The loca- showed high levels of CDH13 mRNA. In contrast, although two tions of these genes were determined from published human alleles were observed for the D16S402 and for the WFDC1 genome sequencing data (37). The patterns of expression of microsatellite markers in the HepG2 cell line, this latter did not several of these genes have never been studied in human tissues. express CDH13 and WFDC1 at the mRNA level. We quantified mRNA levels by real-time quantitative RT- Correlation between mRNA Levels and Clinical and PCR with SYBR Green reagents. This technique can be used to Pathological Features. The clinical and pathological features study a larger number of samples simultaneously than can be or histological characteristics of tumors, such as the number and studied by Northern blotting. It also provides more accurate and diameter of nodules, recurrences, and extrahepatic metastases, reproducible RNA quantification and requires smaller quantities were correlated with LOH and mRNA gene expression. Despite of tumor tissue. We used two controls to normalize the results: the small number of subjects, recurrence seems associated with S18 RNase and the ␤-actin gene. The ␤-actin gene tended to be LOH at the D16S402 locus. Seventy-five percent of patients overexpressed (doubling of expression) in tumor samples. This exhibiting LOH had recurrence, whereas only 33% of patients phenomenon has already been reported in hepatoma and other exempt of LOH recurred. This was in agreement with our types of tumor (43–45). We, therefore, decided to normalize the previous results on a larger series of subjects (35), although here final results with S18 RNA. Nevertheless, identical conclusion no statistical significance can be drawn. Overexpression seems could be drawn from the results obtained with ␤-actin as the not correlated with recurrence or other histological features for reference (data not shown). any of the genes tested. In contrast, three of the five patients Some genes, such as TAF1C, KIAA0703, and HSBP1, were who did not display CDH13 overexpression, four of whom expressed similarly in NL, TL, and adjacent NTL samples, as displayed LOH at this locus, suffered metastatic recurrence well as in isolated hepatocytes and hepatoma cell lines. This within 15 months of surgery. Surprisingly, all of the patients suggests that these genes are not involved in liver carcinogen- (n ϭ 3) without cirrhosis or underlying liver disease displayed esis. Considerable heterogeneous variation was observed in the overexpression of at least four genes whereas six of the seven expression of the other genes, and their role in liver carcino- patients with cirrhosis or hepatitis displayed no overexpression genesis is unclear. However, the liver carcinogenesis seemed to or the overexpression of only one or two genes. CDH13 was be most closely associated with the expression of two genes: overexpressed in two of seven patients with cirrhosis or hepatitis OKL38 and CDH13. and in three of three patients without liver disease. In contrast, The expression of OKL38, described as a pregnancy- overexpression of OKL38 seemed to be independent of under- induced growth inhibitor gene, was similar in cell lines and lying liver disease (three of three versus four of seven). isolated normal hepatocytes, but was much stronger in tumors CDH13 Mutation Analysis in Cell Lines. Except for the and in samples taken from tissue adjacent to tumor than in MAHLAVU cell line, which displays high levels of CDH13 normal healthy liver tissues. These observations are not consist- mRNA, a lack of expression was observed in the other cell lines. ent with a role for this gene as a TSG in liver carcinogenesis, In the aim to explore whether point mutations in the CDH13 and the biological significance of the observed changes is coding sequence might give rise to down-expression of this unclear. gene, exons 1–14 and their surrounding regions were sequenced. CDH13 was the only gene to display significantly higher No base change altering the amino acid sequence was detected expression in tumor samples than in NL and NTL samples. In in the five cell lines analyzed. contrast, little or no expression of this gene was detected in four

of five cell lines tested in this work. CDH13 encodes a protein patients and for the five cell lines studied. No mutation of the belonging to the cadherin family of cell surface glycoproteins CDH13 gene was found in cell lines exhibiting LOH. However, responsible for selective cell recognition and adhesion (46). the frequent correlation observed between changes in the Down-regulation of the CDH13 gene in HepG2 cells, as ob- D16S402 microsatellite in HCC and the higher risk of recur- served here, was also reported in a previous study (47). No rence are more likely to be related to a TSG other than CDH13. mutation of the CDH13 gene was found in the cell lines ana- In conclusion, we studied the expression patterns, in human lyzed. Mechanisms such as changes in the trans-acting elements liver tissue samples and hepatoma cell lines, of 13 genes map- specific for the CDH13 gene promoter or methylation of the ping to a chromosomal region likely to harbor one or several promoter region leading to gene silencing may be involved in TSG. Abnormalities of CDH13, a gene involved in cell-cell the down-regulation of expression observed in cell lines. It is adhesion, have been described in several human cancers and known that the cell-cell adhesion of cadherin is mediated by seem to play a role in HCC. However, no clear relationship has cytoplasmic proteins such as catenins. Mutation of the ␤-catenin been observed between expression of this gene and changes in gene occurs in 15–20% of HCCs (48–50), which can contribute the D16S402 microsatellite or tumor recurrence. None of the to cancer invasion and metastasis. Some experimental and clin- other genes tested is a good candidate for a major TSG in liver ical reports have previously suggested a role for CDH13 in carcinogenesis. Our results suggest that additional unknown breast and lung carcinogenesis and metastasis progression (41, genes may be involved in carcinogenesis and located in the 47, 51–53). In particular, decreases in CDH13 production have 16q23.3–24.1 area. frequently been observed in breast cancer (47, 52, 54). The transfection of a breast carcinoma cell line displaying low levels of CDH13 gene expression with a cDNA encoding H-cadherin REFERENCES resulted in significant inhibition of tumor growth (47). In con- 1. WHO. 1994 World Health Statistics Annual. Geneva: WHO, 1995. trast, the study carried out by Mashimo et al. (33) on microcell- 2. Sugimura, T. Multistep carcinogenesis: a 1992 prospective. Science mediated chromosome 16 transfer in a highly metastatic rat (Wash DC), 258: 603–607, 1992. prostatic cancer cell line suggested that CDH13 is not involved 3. Boige, V., Laurent-Puig, P., Fouchet, P., Flejou, J. F., Monges, G., Bedossa, P., Bioulac-Sage, P., Capron, F., Schmitz, A., Olschwang, S., in the metastasis of this tumor type. Wyder et al. (55) recently and Thomas, G. Concerted nonsyntenic allelic losses in hyperploid showed that mouse H-cadherin is overexpressed on the mem- hepatocellular carcinoma as determined by a high-resolution allelotype. branes of endothelial cells of tumor-penetrating blood vessels, Cancer Res., 57: 1986–1990, 1997. whereas it is present on only a subset of endothelial cells in 4. Nagai, H., Pineau, P., Tiollais, P., Buendia, M. A., and Dejean, A. healthy organs, suggesting a role for H-cadherin in the mecha- Comprehensive allelotyping of human hepatocellular carcinoma. Onco- gene, 14: 2927–2933, 1997. nism of angiogenesis. This finding is consistent with our results. 5. Sheu, J. C., Lin, Y. W., Chou, H. C., Huang, G. T., Lee, H. S., Lin, We found that the level of CDH13 expression was 20 times Y. H., Huang, S. Y., Chen, C. H., Wang, J. T., Lee, P. H., Lin, J. T., Lu, higher in NL than in isolated hepatocytes, suggesting that cells F. J., and Chen, D. S. Loss of heterozygosity and microsatellite insta- other than hepatocytes play an important role in CDH13 expres- bility in hepatocellular carcinoma in Taiwan. Br. J. Cancer, 80: 468– sion in the liver. This is the first report of CDH13 expression in 476, 1999. HCC and hepatoma cell lines. Five of 10 tumor samples and one 6. Tsuda, H., Zhang, W. D., Shimosato, Y., Yokota, J., Terada, M., of five cell lines displayed overexpression of CDH13. A similar Sugimura, T., Miyamura, T., and Hirohashi, S. Allele loss on chromosome 16 associated with progression of human hepatocellular carci- discrepancy has already been observed for the maspin gene, noma. Proc. Natl. Acad. Sci. USA, 87: 6791–6794, 1990. originally described as a TSG, which affects cell motility and 7. Sakai, K., Nagahara, H., Abe, K., and Obata, H. Loss of heterozy- invasion (56). Maass et al. (57) detected maspin staining in 23 gosity on chromosome 16 in hepatocellular carcinoma. J. Gastroenterol. of 24 pancreatic cancer tissues, which suggests that maspin Hepatol., 7: 288–292, 1992. expression is a common event in pancreatic cancer cells. Al- 8. Nishida, N., Fukuda, Y., Kokuryu, H., Sadamoto, T., Isowa, G., though the molecular and biological mechanisms underlying the Honda, K., Yamaoka, Y., Ikenaga, M., Imura, H., and Ishizaki, K. Accumulation of allelic loss on arms of chromosomes 13q, 16q and 17p functions of maspin and CDH13 are unknown, several authors in the advanced stages of human hepatocellular carcinoma. Int. J. have suggested that these proteins act by blocking tumor cell Cancer, 51: 862–868, 1992. migration and proliferation, thereby preventing invasion and 9. Chou, Y. H., Chung, K. C., Jeng, L. B., Chen, T. C., and Liaw, Y. F. metastasis (48, 58, 59). Maass’ findings for pancreatic cancer, Frequent allelic loss on chromosomes 4q and 16q associated with human together with our findings for HCC, provide new information hepatocellular carcinoma in Taiwan. Cancer Lett., 123: 1–6, 1998. about the factors regulating cell development. 10. Piao, Z., Park, C., Kim, J. J., and Kim, H. Deletion mapping of One of the most frequent types of genetic alteration in HCC chromosome 16q in hepatocellular carcinoma. Br. J. Cancer, 80: 850– 854, 1999. is somatic LOH; this results in the inactivation of TSGs, which 11. Gramantieri, L., Trere, D., Pession, A., Piscaglia, F., Masi, L., normally regulate cell growth and prevent abnormal cell prolif- Gaiani, S., Mazziotti, A., and Bolondi, L. Allelic imbalance on 16q in eration. In our study, four of the tumor samples tested displayed small, unifocal hepatocellular carcinoma: correlation with HBV and 16q23.3–24.1 LOH. All these samples displayed LOH at the HCV infections and cellular proliferation rate. Dig. Dis. Sci., 45: 306– D16S402 locus. In three of these cases, recurrence was observed 311, 2000. within 15 months of surgery. These results confirm the prog- 12. Bando, K., Nagai, H., Matsumoto, S., Koyama, M., Kawamura, N., Tajiri, T., Onda, M., and Emi, M. Identification of a 1-Mb common nostic value of changes in the D16S402 microsatellite for HCC region at 16q24.1–24.2 deleted in hepatocellular carcinoma. Genes metastasis that we previously reported (35). No correlation was Chromosomes Cancer, 28: 38–44, 2000. observed between the pattern of gene expression and LOH 13. Wong, N., Lai, P., Pang, E., Fung, L., Sheng, Z., Wong, V., Wang, status at loci in the vicinity of the investigated genes for the 10 W., Hayashi, Y., Perlman, E., Yuna, S., Lau, J., and Johnson, P.

Genomic aberrations in human hepatocellular carcinomas of differing 29. Grundy, R. G., Pritchard, J., Scambler, P., and Cowell, J. K. Loss of etiologies. Clin. Cancer Res., 6: 4000–4009, 2000. heterozygosity on chromosome 16 in sporadic Wilms’ tumour. Br. J. 14. Guan, X. Y., Fang, Y., Sham, J. S., Kwong, D. L., Zhang, Y., Liang, Cancer, 78: 1181–1187, 1998. Q., Li, H., Zhou, H., and Trent, J. M. Recurrent chromosome alterations 30. Shearer, P. D., Valentine, M. B., Grundy, P., DeCou, J. M., Banavali, in hepatocellular carcinoma detected by comparative genomic hybrid- S. D., Komuro, H., Green, D. M., Beckwith, J. B., and Look, A. T. ization. Genes Chromosomes Cancer, 29: 110–116, 2000. Hemizygous deletions of chromosome band 16q24 in Wilms tumor: detec- 15. Balsara, B. R., Pei, J., De Rienzo, A., Simon, D., Tosolini, A., Lu, tion by fluorescence in situ hybridization. Cancer Genet. Cytogenet., 115: Y. Y., Shen, F. M., Fan, X., Lin, W. Y., Buetow, K. H., London, W. T., 100–105, 1999. and Testa, J. R. Human hepatocellular carcinoma is characterized by a 31. Mason, J. E., Goodfellow, P. J., Grundy, P. E., and Skinner, M. A. highly consistent pattern of genomic imbalances, including frequent loss 16q loss of heterozygosity and microsatellite instability in Wilms’ of 16q23.1–24.1. Genes Chromosomes Cancer, 30: 245–253, 2001. tumor. J. Pediatr. Surg., 35: 891–897, 2000. 16. Kanai, Y., Ushijima, S., Tsuda, H., Sakamoto, M., Sugimura, T., 32. Skotnicka-Klonowicz, G., Rieske, P., Bartkowiak, J., Szymik- and Hirohashi, S. Aberrant DNA methylation on chromosome 16 is an Kantorowicz, S., Daszkiewicz, P., and Debiec-Rychter, M. 16q hetero- early event in hepatocarcinogenesis. Jpn. J. Cancer Res., 87: 1210– zygosity loss in Wilms’ tumour in children and its clinical importance. 1217, 1996. Eur. J. Surg. Oncol., 26: 61–66, 2000. 17. Kanai, Y., Ushijima, S., Tsuda, H., Sakamoto, M., and Hirohashi, S. 33. Mashimo, T., Watabe, M., Cuthbert, A. P., Newbold, R. F., Rinker- Aberrant DNA methylation precedes loss of heterozygosity on chromo- Schaeffer, C. W., Helfer, E., and Watabe, K. Human chromosome 16 some 16 in chronic hepatitis and liver cirrhosis. Cancer Lett., 148: suppresses metastasis but not tumorigenesis in rat prostatic tumor cells. 73–80, 2000. Cancer Res., 58: 4572–4576, 1998. 18. Yumoto, Y., Hanafusa, T., Hada, H., Morita, T., Ooguchi, S., 34. Caligo, M. A., Polidoro, L., Ghimenti, C., Campani, D., Cecchetti, Shinji, N., Mitani, T., Hamaya, K., Koide, N., and Tsuji, T. Loss of D., and Bevilacqua, G. A region on the long arm of chromosome 16 is heterozygosity and analysis of mutation of p53 in hepatocellular carci- frequently deleted in metastatic node-negative breast cancer. Int. J. noma. J. Gastroenterol. Hepatol., 10: 179–185, 1995. Oncol., 13: 177–182, 1998. 19. Piao, Z., Kim, H., Malkhosyan, S., and Park, C. Frequent chromo- 35. Salvucci, M., Lemoine, A., Saffroy, R., Azoulay, D., Lepere, B., somal instability but no microsatellite instability in hepatocellular car- Gaillard, S., Bismuth, H., Reynes, M., and Debuire, B. Microsatellite cinomas. Int. J. Oncol., 17: 507–512, 2000. instability in European hepatocellular carcinoma. Oncogene, 18: 181– 20. Lin, Y. W., Lee, I. N., Chen, C. H., Huang, G. T., Lee, H. S., Lee, 187, 1999. P. H., Lu, F. J., and Sheu, J. C. Deletion mapping of chromosome 16q24 36. Kashiwaba, M., Tamura, G., Suzuki, Y., Maesawa, C., Ogasawara, in hepatocellular carcinoma in Taiwan and mutational analysis of the S., Sakata, K., and Satodate, R. Epithelial-cadherin gene is not mutated 17-beta-HSD gene localized to the region. Int. J. Cancer, 93: 74–79, in ductal carcinomas of the breast. Jpn. J. Cancer Res., 86: 1054–1059, 2001. 1995. 21. Elo, J. P., Harkonen, P., Kyllonen, A. P., Lukkarinen, O., Poutanen, 37. The Genome International Sequencing Consortium. The human M., Vihko, R., and Vihko, P. Loss of heterozygosity at 16q24.1-q24.2 is genome. Nature (Lond.), 409: 860–921, 2001. significantly associated with metastatic and aggressive behavior of 38. Lemoine, A., Le Bricon, T., Salvucci, M., Azoulay, D., Pham, P., prostate cancer. Cancer Res., 57: 3356–3359, 1997. Raccuia, J., Bismuth, H., and Debuire, B. Prospective evaluation of 22. Pan, Y., Matsuyama, H., Wang, N., Yoshihiro, S., Haggarth, L., Li, circulating hepatocytes by ␣-fetoprotein mRNA in humans during liver C., Tribukait, B., Ekman, P., and Bergerheim, U. S. Chromosome 16q24 surgery. Ann. Surg., 226: 43–50, 1997. deletion and decreased E-cadherin expression: possible association with 39. Pichard, L., Fabre, I., Daujat, M., Domergue, J., Joyeux, H., and metastatic potential in prostate cancer. Prostate, 36: 31–38, 1998. Maurel, P. Effect of corticosteroids on the expression of cytochromes 23. Kawakami, M., Staub, J., Cliby, W., Hartmann, L., Smith, D. I., and P450 and on cyclosporin A oxidase activity in primary cultures of Shridhar, V. Involvement of H-cadherin (CDH13) on 16q in the region human hepatocytes. Mol. Pharmacol., 41: 1047–1055, 1992. of frequent deletion in ovarian cancer. Int. J. Oncol., 15: 715–720, 1999. 40. Saffroy, R., Riou, P., Soler, G., Azoulay, D., Emile, J. F., Debuire, 24. Suzuki, S., Moore, D. H., Ginzinger, D. G., Godfrey, T. E., Barclay, B., and Lemoine, A. Analysis of alterations of WFDC1, a new putative J., Powell, B., Pinkel, D., Zaloudek, C., Lu, K., Mills, G., Berchuck, A., tumour suppressor gene, in hepatocellular carcinoma, Eur. J. Hum. and Gray, J. W. An approach to analysis of large-scale correlations Genet., 10: 239–244, 2002. between genome changes and clinical endpoints in ovarian cancer. 41. Sato, M., Mori, Y., Sakurada, A., Fujimura, S., and Horii, A. The Cancer Res., 60: 5382–5385, 2000. H-cadherin (CDH13) gene is inactivated in human lung cancer. Hum. 25. Launonen, V., Mannermaa, A., Stenback, F., Kosma, V. M., Puistola, Genet., 103: 96–101, 1998. U., Huusko, P., Anttila, M., Bloigu, R., Saarikoski, S., Kauppila, A., and 42. Knudson, A. G., Jr. Mutation and cancer: statistical study of reti- Winqvist, R. Loss of heterozygosity at chromosomes 3, 6, 8, 11, 16, and 17 noblastoma. Proc. Natl. Acad. Sci. USA, 68: 820–823, 1971. in ovarian cancer: correlation to clinicopathological variables. Cancer 43. Finnegan, M. C., Goepel, J. R., Hancock, B. W., and Goyns, M. H. Genet. Cytogenet., 122: 49–54, 2000. Investigation of the expression of housekeeping genes in non-Hodgkin’s 26. Chen, T., Sahin, A., and Aldaz, C. M. Deletion map of chromosome lymphoma. Leuk. Lymphoma, 10: 387–393, 1993. 16q in ductal carcinoma in situ of the breast: refining a putative tumor 44. Chang, T. J., Juan, C. C., Yin, P. H., Chi, C. W., and Tsay, H. J. suppressor gene region. Cancer Res., 56: 5605–5609, 1996. Up-regulation of ␤-actin, cyclophilin and GAPDH in N1S1 rat hepa- 27. Cleton-Jansen, A. M., Callen, D. F., Seshadri, R., Goldup, S., toma. Oncol. Rep., 5: 469–471, 1998. Mccallum, B., Crawford, J., Powell, J. A., Settasatian, C., van Beeren- 45. Goidin, D., Mamessier, A., Staquet, M. J., Schmitt, D., and donk, H., Moerland, E. W., Smit, V. T., Harris, W. H., Millis, R., Berthier-Vergnes, O. Ribosomal 18S RNA prevails over glyceralde- Morgan, N. V., Barnes, D., Mathew, C. G., and Cornelisse, C. J. Loss of hyde-3-phosphate dehydrogenase and ␤-actin genes as internal standard heterozygosity mapping at chromosome arm 16q in 712 breast tumors for quantitative comparison of mRNA levels in invasive and noninva- reveals factors that influence delineation of candidate regions. Cancer sive human melanoma cell subpopulations. Anal. Biochem., 295: 17– Res., 61: 1171–1177, 2001. 21, 2001. 28. Yoon, D. S., Li, L., Zhang, R. D., Kram, A., Ro, J. Y., Johnston, D., 46. Takeichi, M. Cadherin cell adhesion receptors as a morphogenetic Grossman, H. B., Scherer S., and Czerniak, B. Genetic mapping and regulator. Science (Wash DC), 251: 1451–1455, 1991. DNA sequence-based analysis of deleted regions on chromosome 16 47. Lee, S. H-cadherin, a novel cadherin with growth inhibitory func- involved in progression of bladder cancer from occult preneoplastic tions and diminished expression in human breast cancer. Nat. Med., 2: conditions to invasive disease. Oncogene, 20: 5005–5014, 2001. 776–782, 1996.

48. Miyoshi, Y., Iwao, K., Nagasawa, Y., Aihara, T., Sasaki, Y., 54. Miki, Y., Katagiri, T., and Nakamura, Y. Infrequent mutation of the Imaoka, S., Murata, M., Shimano, T., and Nakamura, Y. Activation of H-cadherin gene on chromosome 16q24 in human breast cancers. Jpn. J. the ␤-catenin gene in primary hepatocellular carcinomas by somatic Cancer Res., 88: 701–704, 1997. alterations involving exon 3. Cancer Res., 58: 2524–2527, 1998. 55. Wyder, L., Vitaliti, A., Schneider, H., Hebbard, L. W., Moritz, 49. de La Coste, A., Romagnolo, B., Billuart, P., Renard, C. A., Buen- D. R., Wittmer, M., Ajmo, M., and Klemenz, R. Increased expression of dia, M. A., Soubrane, O., Fabre, M., Chelly, J., Beldjord, C., Kahn, A., H/T-cadherin in tumor-penetrating blood vessels. Cancer Res., 60: ␤ and Perret, C. Somatic mutations of the -catenin gene are frequent in 4682–4688, 2000. mouse and human hepatocellular carcinomas. Proc. Natl. Acad. Sci. USA, 95: 8847–8851, 1998. 56. Zou, Z., Anisowicz, A., Hendrix, M. J., Thor, A., Neveu, M., Sheng, S., Rafidi, K., Seftor, E., and Sager, R. Maspin, a serpin with tumor- 50. Terris B. Pineau, P., Bregeaud, L., Valla, D., Belghiti, J., Tiollais, suppressing activity in human mammary epithelial cells. Science (Wash P., Degott, C., and Dejean, A. Close correlation between ␤-catenin gene alterations and nuclear accumulation of the protein in human hepato- DC), 263: 526–529, 1994. cellular carcinomas. Oncogene, 18: 6583–6588, 1999. 57. Maass, N., Hojo, T., Ueding, M., Luttges, J., Kloppel, G., Jonat, W., 51. Lee, S. W., Reimer, C. L., Campbell, D. B., Cheresh, P., Duda, R. B., and Nagasaki, K. Expression of the tumor suppressor gene Maspin in and Kocher, O. H-cadherin expression inhibits in vitro invasiveness and human pancreatic cancers. Clin. Cancer Res., 7: 812–817, 2001. tumor formation in vivo. Carcinogenesis (Lond.), 19: 1157–1159, 1998. 58. Sheng, S., Carey, J., Seftor, E. A., Dias, L., Hendrix, M. J., and 52. Toyooka, K., Toyooka, S., Virmani, A., Sathyanarayana, U., Euhus, Sager, R. Maspin acts at the cell membrane to inhibit invasion and D., Gilcrease, M., Minna, J., and Gazdar, A. Loss of expression and motility of mammary and prostatic cancer cells. Proc. Natl. Acad. Sci. aberrant methylation of the CDH13 (H-cadherin) gene in breast and lung USA, 93: 11669–11674, 1996. carcinomas. Cancer Res., 61: 4556–4560, 2001. 59. Sheng, S., Truong, B., Fredrickson, D., Wu, R., Pardee, A., and 53. Zhong, Y., Delgado, Y., Gomez, J., Lee, S. W., and Perez-Soler, R. Sager, R. Tissue-type plasminogen activator is a target of the tumor Loss of H-cadherin protein expression in human non-small cell lung cancer suppressor gene maspin. Proc. Natl. Acad. Sci. USA, 95: 499–504, is associated with tumorigenicity. Clin. Cancer Res., 7: 1683–1687, 2001. 1998.

Downloaded from clincancerres.aacrjournals.org on September 23, 2021. © 2002 American Association for Cancer Research. Investigation in Liver Tissues and Cell Lines of the Transcription of 13 Genes Mapping to the 16q24 Region That Are Frequently Deleted in Hepatocellular Carcinoma

Philippe Riou, Raphaël Saffroy, Jérome Comoy, et al.

Clin Cancer Res 2002;8:3178-3186.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/8/10/3178

Cited articles This article cites 58 articles, 21 of which you can access for free at: http://clincancerres.aacrjournals.org/content/8/10/3178.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/8/10/3178.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/8/10/3178. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.