a novel gene (hip) activated in human primary liver cancer1 · in colon, brain, kidney, or lung. in...

[CANCER RESEARCH 52, 5089-5095, September 15, 1992]

A Novel Gene (HIP) Activated in Human Primary Liver Cancer1

Chantai Lasserre,2 Laurence Christa, Marie-ThÃ©rÃ¨seSimon, Philippe Vernier, and Christian BrÃ©chotINSERM U75, CHU Necker, 156 rue de Vaugirard, 75742 Paris cedex 15 fC. L., L. C., M-T. S., C. B.]; UnitÃ©d'HÃ©patologie,HÃ´pital LaÃ«nnec,rue de SevrÃ©s,75007Paris [C. B.]; and Laboratoire de Neurobiologie Cellulaire et MolÃ©culaire,CNRS, Gif sur Yvette F-91190 [P. V.], France

ABSTRACT

Differential screening of a human hepatocellular carcinoma complementary DNA library using subtracted probes allowed us to identify anovel gene named HIP whose expression at the transcriptional level waselevated in liver tumors. The protein potentially encoded by the complementary DNA showed 68.5% identity with the bovine pancreaticthread protein and 49% identity with the human reg protein, which hasbeen proposed as a pancreatic islet cell regenerating factor and is identical to the pancreatic stone or pancreatic thread protein. Sequenceanalysis suggests that the bovine pancreatic thread protein encodinggene is, in fact, the bovine homologue of the HIP gene. Furthermore,data base searches revealed a significant similarity of the HIP andpancreatic stone protein/pancreatic thread protein/reg sequences withthe C-type lectin superfamily. The HIP sequence, like pancreatic stone

protein/pancreatic thread protein/reg protein, consists of a single carbohydrate recognition domain linked to a signal peptide which would beinvolved in secretion of the protein.

HIP mRNA was expressed at a high level in the tumors of seven of29 hepatocellular carcinomas. In contrast, HIP mRNA was not detectedin nontumorous adjacent areas or in normal adult and fetal liver, suggesting that HIP could be involved in liver cell proliferation or differentiation. HIP mRNA expression is tissue specific, since it is present inthe normal small intestine and pancreas, while it could not be evidencedin colon, brain, kidney, or lung.

In summary, our results show the existence of a novel family withinthe superfamily of C-type lectin which may be involved in liver, pancre

atic, and intestinal cell proliferation or differentiation.

INTRODUCTION

Although highly differentiated hepatocytes retain replicativepotential. In response to partial hepatectomy, rodent liver regeneration involves almost all hepatocytes, which divide eitheronce or twice before returning to their normal quiescent state(1). Liver cells can also proliferate in response to cell necrosisdue to viral infections or chemicals. Similarly, liver cirrhosis inhumans is characterized by so-called regenerative nodules separated by areas of fibrosis; cell proliferation within these nodules likely accounts in part for the emergence of HCC3 within

the cirrhotic tissue (2).As altered gene expression is a common feature of neoplastic

cells, several authors have sought the identification of geneswhich are abundantly or specifically expressed in tumoroustissues as compared with the corresponding normal tissues(3, 4). It is assumed that such genes might be involved in somesteps in cell transformation. In addition, the steady-state levelof their transcripts may also provide information on the hepa-tocyte differentiation both during carcinogenesis and in fullydeveloped tumors.

In the present report we used a subtractive hybridizationtechnique which allows genes to be selected on the basis of their

Received 5/27/92; accepted 7/3/92.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.

1Work supported by INSERM, ARC, LNC, Fondation de France, EC.2 To whom requests for reprints should be addressed.3 The abbreviations used are: HCC, hepatocellular carcinoma; cDNA, comple

mentary DNA; PSP, pancreatic stone protein; FTP, pancreatic thread protein; polyA+ RNA, polyadenylate-containing RNA; BPTP, bovine pancreatic thread protein;CRD, carbohydrate-recognition domain; HCA, Hydrophobie cluster analysis.

differential expression in tumorous as compared with normalliver tissues. This procedure (5, 6), greatly facilitates the identification of tumor-specific mRNAs. A cDNA library was generated from RNAs extracted from a human primary liver cancer. Differential screening of this library with subtracted cDNAprobes allowed us to identify a novel gene named HIP, whosemRNA level was elevated in liver tumors. HIP mRNA expression in normal tissues was restricted to the small intestine andpancreas and, in particular, was not detected in normal liver.The protein potentially encoded by HIP cDNA shows a significant similarity with the products of three genes which wererecently shown (7) to be identical: PSP (8) or PTP (9) and reg(7, 10). In addition, on the basis of structural similarities, HIPprotein probably belongs to the C-type lectin superfamily (11).

It has been hypothesized that the PSP/PTP/reg proteinmight be involved in cell regeneration in the pancreas (10, 12)and brain (13) and in the prevention of pancreatic calculi formation (14). Therefore, our report identifies a gene belongingto a novel gene family, which may be involved in liver, pancreatic, and intestinal cell differentiation and/or proliferation.

MATERIALS AND METHODS

Tissue Samples. The liver sample used to construct for the cDNAlibrary was an early, well-differentiated HCC obtained at surgery froma young woman (Patient C. H.) uninfected by hepatitis B or C viruses.The HCC developed on the left liver lobe, while a histologically benignadenoma was present on the right lobe. The nontumoral liver washistologically normal. Twenty-nine primary liver cancers were also obtained at surgery; both tumorous and nontumorous liver samples wereavailable in 28 cases. Human pancreas, liver, and intestine were obtained at surgery. Rat tissue samples were obtained from male Sprague-Dawley rats. All tissues were immediately frozen in liquid nitrogen andstored at -80Â°C until use.

RNA Extraction and Northern Blots. Total RNA was extracted using the hot phenol method (15). Poly A+ RNA was purified by means

of oligo(dT) cellulose affinity chromatography. RNAs were separatedon a denaturing formaldehyde gel as previously described (16) and thentransferred onto Hybond-C-extra nitrocellulose membranes (Amer-sham, United Kingdom). The probe used to hybridize the filters was theBamHl-BamHl fragment prepared from HIP cDNA. Mouse /3-actincDNA was used as reference probe to verify the quantity of RNAapplied to the filters. Hybridization and washing were carried out aspreviously described (16).

HCC cDNA Library and Subtracted cDNA Probes: Generation andScreening. Oligo(dT)-primed, double-stranded cDNA was preparedfrom poly A+ RNA isolated from the primary liver cancer of Patient C.

H. (17). Blunt ends were obtained by treatment with T4 polymerase.Â£coRIsites were methylated, and EcoRl linkers were ligated to thecDNA. After Â£coRIdigestion the cDNA was size-fractionated in asucrose gradient. The cDNAs exhibiting a mean size between 1 and 2kilobases were selected and inserted into the EcoRI site of the blue-script

vector (Stratagene). DH 5Â«competent cells (BRL) were transformedwith ligation mix. The library gave rise to 5 x IO4 recombinant clones

that were stored by plating on Hybond C extra nitrocellulose membranes (Amersham, United Kingdom) and screened without amplification.

Oligo(dT)-primed, single-stranded cDNA probes were prepared from3 Mgof poly A+ RNA isolated from (a) Patient C. H. (HCC), (b) Patient

C. H. (adenoma), and (c) normal liver. Probes were labeled during thecDNA synthesis by adding 1.2 Â¡iC\/u\of [a-32P]dCTP (400 pCi/ml;

5089

Research. on January 7, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

A NOVEL GENE (HIP) AND HUMAN PRIMARY LIVER CANCER

Amersham) with cold dCTP (80 MMfinal concentration). Subtraction ofthe cDNA population was performed by liquid hybridization with 60 Â¿tgof normal liver poly A* RNA at 68Â°Cfor 40 h. Subtracted single-

stranded cDNA was separated from cDNA-RNA hybrid molecules using hydroxylapatite (Biorad) chromatography. Subtracted single-stranded DNA was used to hybridize the cDNA library. Recombinantclones were ordered on nitrocellulose filters in order to facilitate theanalyses of differential hybridization patterns obtained with the subtracted cDNA probes. Filters were triplicated, and each was hybridizedwith 10s cpm/ml of subtracted cDNA probes for 14 h. Filters werewashed for 30 min at 65Â°Cin 0.1% sodium dodecyl sulfate:0.3x stan

dard saline citrate buffer and subjected to autoradiography for 10 daysat -80Â°Cwith intensifying screens.

Ileum cDNA Library. Oligo(dT)-primed, double-stranded cDNAwas prepared from poly A+ of human ileum, and a XZAP cDNA library

was generated using the XZAP cDNA synthesis kit (Stratagene).Nucleotide Sequence Analysis. HIP cDNA was digested with

BamHl restriction enzyme, and the inserts obtained were subcloned inthe bluescript vector (Stratagene). Double-stranded DNA was se-

quenced using the dideoxy chain termination method. Data basesearches and sequence analyses were performed with the GCG package(18) and hydrophobic cluster analysis (19, 20) with the MANSERprogram on a VAX computer (Digital).

Primer Extension Experiments. Primer extension assays were performed using 20 Mgof ileum total RNA, 5' end-labeled [7-'2P]ATP, and

murine leukemia virus reverse transcriptase (BRL). The primer was asynthetic 21-mer oligonucleotide corresponding to nucleotides â€”¿�1+20

of the cDNA. The size of extended primer was measured by comparisonwith a sequencing ladder.

RESULTS

Differential Screening of the HCC cDNA Library. A cDNAlibrary was generated from the tumorous tissues of a patientwith a well-differentiated HCC (Patient C. H.). Among the50,000 recombinant clones obtained, 2,500 were analyzed ontriplicate filters by differential screening. Three single-stranded

cDNA probes were synthesized from (a) the HCC tumor liversample used for the cDNA library, (b) the adenoma from thesame patient, and (c) liver samples from several normal controlindividuals and then subtracted with liver poly A+ RNA from

several normal subjects. Among the 2,500 clones tested with thethree probes, 330 showed a differential pattern of hybridization,i.e., the cDNA probe derived from the HCC tissue yielded amarkedly more intense signal than the probes obtained from theadenoma and normal liver. These 330 clones were again submitted to the differential screening procedure in order to enhance the selectivity of the analysis.

As for the first screening, hybridization was performed withthe cDNA probes derived from liver RNA samples of Patient C.H. However, subtraction was achieved with total RNA from ahistologically normal liver sample adjacent to the tumor andthus obtained from the same liver. This procedure shouldindeed avoid selection of genes activated by the surgical resection itself. In the second round of hybridization, 10 recombinant clones strongly hybridizing with the subtracted cDNAprobe derived from the HCC were picked up. These cDNAclones were then used as probes in a Northern blot analysis ofRNAs extracted from the HCC, adenoma, and normal livertissues of Patient C. H. Intense expression of the correspondingmRNAs was detected in the HCC and adenoma, while no signalwas detectable in the nontumorous tissues and normal liver(Fig. 1). One of these 10 clones was selected for further analysis,on the basis of intensity of hybridization.

Nucleotide and Amino Acid Sequence of the HIP Gene. ThiscDNA, which identified a 0.9-kilobase mRNA molecule, was

KB

0.9

Fig. 1. RNA blot analysis of liver RNA from Patient C. H. and from normalliver. The probe was the BamH\-BamH[ fragment purified from HIP cDNA.Lanes I lo 4 (exposure times are in parentheses). Patient C. H. I une /, HCC, 40/ig of total RNA (20 h); Lane 2, nontumorous area of HCC, 40 Mgof total RNA(7 days); Lane 3, adenoma, 40 eg of total RNA (20 h); Lane 4 nontumorous areaof adenoma, 40 eg of total RNA (7 days); Lanes 5 and 6, normal liver from twodifferent subjects, 40 ua of total RNA (7 days).

further analyzed and forms the basis of this report. Its nudootide sequence was determined. As shown in Fig. 2, the cDNAhad one open-reading frame encoding a 175-amino acid protein, assuming that the ATG at nucleotide +1 is the start codon(21) and the TGA at nucleotide 526 is the stop codon. The 3'

untranslated region included the last 222 nucleotides and exhibited the canonical polyadenylation signal (AATAAA) in position 735. The poly(A) tail was included in the cDNA. Aprimer extension assay indicated that cap sites were located 40and 60 nucleotides upstream of the first ATG (data not shown).

The nucleotide and deduced amino acid sequences werescreened for potential similarities. We found a 49% amino acidhomology with the products of three human genes which wererecently shown (7) to be identical, the PSP (8), PTP (9), and thereg protein (10) encoding genes, as well as with their rat counterparts [PSP and reg rat proteins] ( 10, 22) (Fig. 3). However,the protein potentially encoded by the HIP cDNA sequenceexhibited a better (68.5%) similarity with a BPTP (23). Furthermore, the BPTP sequence is closer to that of HIP than tothat of the PSP/PTP/reg gene (68.5% versus 44%) at the aminoacid level. In addition, the number of amino acids in HIP andBPTP is identical (n = 175); in contrast it is different from thatof PSP/PTP/reg human and rat proteins (166 and 165 aminoacids, respectively). It is therefore very likely that the HIP protein comprises the human homologues bovine BTP, HIP, andBPTP, being more distantly related to the PSP/PTP/reg humanand rat proteins (Fig. 3).

The jV-terminus of the HIP protein sequence is highly hydro-phobic and reminiscent of the signal peptide cleaved in themature PSP/PTP/reg and bovine PTP proteins. It is not yetclear whether the NH2 end of the mature protein is theglutamine in position 23 (which is conserved in HIP/BPTP andPSP/reg) or that in position 25 (which is changed to glutamicacid in HIP/BPTP) (8). The potential trypsin cleavage site at

5090




position 33-34 of PSP/PTP/reg protein is also conserved inHIP/BPTP. Lastly, a potential JV-glycosylation site is situatedat position 136 of the HIP sequence but is not conserved in theBTP or PSP/PTP/reg sequences.

-iagtcgcaqacact

ATG CTG CCT CCC ATO GCC CTG CGC ACT GTA TCT TGG ATGMet Leu Pro Pro Met Ala Leu Pro Ser Val Ser Trp Met

CTG CTT TCC TGC CTC ATG CTG CTG TCT CAG GTT CAÃ•GGTLeu Leu Ser Cys Leu Met Leu Leu Ser Gin Val Gin Gly

GAA GAA CCC CAG AGG GAA CTG CCC TCT GCA CGG ATC CGCGlu Glu Pro Gin Arg Glu Leu Pro Ser Ala Arg Ile Arg

TGT CCC AAA GGC TCC AAG GCC TAT GGC TCC CAC TGC TATCys Pro Lys Gly Ser Lys Ala Tyr Gly Ser His Cys Tyr

GCC TTG TTT TTG TCA CCA AAA TCC TGG ACA GAT GCA GATAla Leu Phe Leu Ser Pro Lys Ser Trp Thr Asp Ala Asp

CTG GCC TGC CAG AAG CGG CCC TCT GGA AAC CTG GTG TCTLeu Ala Cys Gin Lys Arg Pro Ser Gly Asn Leu Val Ser

GTG CTC AGT GGG GCT GAG GGA TCC TTC GTG TCC TCC CTGVal Leu Ser Gly Ala Glu Gly Ser Phe Val Ser Ser Leu

GTG AAG AGC ATT GGT AAC AGC TAC TCA TAC GTC TGG ATTVal Lys Ser Ile Gly Asn Ser Tyr Ser Tyr Val Trp Ile

GGG CTC CAT GAC CCC ACA CAG GGC ACC GAG CCC AAT GGAGly Leu His Asp Pro Thr Gin Gly Thr Glu Pro Asn Gly

GAA GGT TGG GAG TGG AGT AGC AGT GAT GTG ATG AAT TACGlu Gly Trp Glu Trp Ser Ser Ser Asp Val Met Asn Tyr

TTT GCA TGG GAG AGA AAT CCC TCC ACC ATC TCA AGC CCCPhe Ala Trp Glu Arg Asn Pro Ser Thr Ile Ser Ser Pro

GGC CAC TGT GCG AGC CTG TCG AGA AGC ACA GCA TTT CTGGly His Cys Ala Ser Leu Ser Arg Ser Thr Ala Phe Leu

AGG TGG AAA GAT TAT AAC TGT AAT GTG AGG TTA CCC TATArg Trp Lys Asp Tyr Asn Cys Asn Val Arg Leu Pro Tyr

GTC TGC AAG TTC ACT GAC tagtgcaggagggaagtcagcagcctgVal Cys Lys Phe Thr Asp *

tgtttggtgtgcaactcatcatgggcatgagaccagtgtgaggactcaccct 604

ggaagagaatattcgcttaattcccccaacctgaccacctcattcttatctt 656

tcttctgtttcttcctccccgctgtcatttcagtctcttcattttgtcatac 708

ggcctaaggctttaaagagcaataaaatttttagtctgcpolyA 747

Fig. 2. Sequence of human HIP mRNA and deduced sequence of encodedpreprotein. Noncoding sequences are in lowercase letters, and the polyadenylationsite aataaa is underlined. The stop codon is marked by an asterisk.

3913

7826

11739

15652

19565

23478

27391

312104

351117

390130

429143

468156

507169

552175

HIP and PSP/PTP/reg Proteins Belong to the C-Type Lectin

Superfamily. Data base searches revealed another interestingfeature, namely, the significant similarity of the HIP and PSP/PTP/reg sequences with the calcium-dependent, C-type, animal

lectin superfamily. Although this similarity has already beennoticed (24), a systematic structural analysis and its potentialfunctional consequences remained to be done. The C-type lectin

family comprises either multifunctional membrane proteinslike the hepatic asialoglycoprotein receptor or small solubleproteins with no second type of binding site such as mannose-binding proteins and invertebrate lectins. C-type animal lectinsare structurally characterized by a conserved CRD comprising120 to 130 amino acid residues including four invariable cys-

teines which form two intrachain disulfide bridges (11). Theresolution of the three-dimensional structure of CRD inthe crystallized rat mannose binding Protein A at 2.5 A (25)allowed an accurate comparison of the HIP/PSP/PTP/regsequences with the structural regions of C-lectin CRD to bemade. As a first step we used a two-dimensional analysis of the

sequences to facilitate the alignment of CRD in representativemembers of the C-lectin superfamily with the HIP and regsequences. The HCA, an example of which is presented inFig. 4, gave rise to a sequence segmentation which fits, withvery few exceptions, with the three-dimensional structure of theCRD of the mannose-binding Protein A. The two a helices andthe five /3-strands were easily identified in the CRD of the 14sequences of C-lectins we analyzed. It was more difficult todelineate the four loops, and their lengths vary significantlyamong different CRDs. However, the residues, probably involved in the coordination of the two calcium ions required forcarbohydrate binding, are highly conserved in loops LI, L3, andL4 and in strand 04 (Fig. 5). In this respect, HIP/reg sequencesexhibited variations of consensus amino acids belonging to thecalcium binding site 2 in Loop 4. Acidic amino acids arechanged for serine-threonine/serine and glycine-histidine/

tyrosine, respectively, at the beginning and the end of Loop 4,in the sequence of HIP and PSP/PTP/reg proteins.

Four conserved cysteines can form disulfide bridges betweenal and (35 on one hand and between /33 and /34 on the otherhand. An identical situation has been shown to occur for PSP/PTP/reg protein (26, 27) and bovine-PTP (28) and could

HshipBtptp

HaregHsptpRnptp

Consensus H"S"

X Q T 3 3 YA Q T N 3T R . N K

Consensus Q- - 120

Consensus - Q - 0 Q- - Q 0-13-0 D - Q - - B LM;1J 176

Fig. 3. Comparison of amino acid sequences of HIP (Hship), bovine PTP (Btptp) (23). human reg (Hsreg) (10), human pre-PSP (Hsptp) (8), and rat PTP (Rnptp)(10, 22). Six conserved cysteine residues are marked by arrows. The stop codons are marked by asterisks.

5091




RNMBPA

HSHIP

HSREG

a.2 ÃŸ2L1L2 L3L4ÃŸ3 B4 ÃŸ5^MM ^^m â€”¿� â€”¿�â€” â€”¿� â€”¿�^^ i^HI ___

Fig. 4. HCA plots of the potential CRD of several C-leetins. HCA plot alignments were performed with the Mansek program (18). RNMBPA, rat mannose-binding Protein A (38); HSHIP, human HIP; HSREG, human reg (7); MRLEC2,barnacle lectin Bra-2 (30); RNLECI, rat hepatic lectin 1 (39); MMLHR, mouselymphocyte homing receptor (40); HSPGCC, human cartilage proteoglycan (41).Vertical bars delineate consensus stretches determined by HCA plots. Domainsdefined by the crystallographic analysis of RNMBPA are indicated by horizontalbars upon the sequence (thick bars, a helices; medium bars, ÃŸstrands; thin bars,loops). *, prolines; *, glycines; D, threonines; El, serines. The numbering wasgiven according to the method of Weiss et al. (25).

therefore be predicted for HIP protein too. Two extra cysteines,found at the end terminus of HIP and PSP/PTP/reg proteins(Fig. 3), are surrounded by a sequence stretch conserved in thesea urchin lectin (29) as well as in the acorn barnacle lectinsBra-2 and Bra-3 (30, 31) where they generate, as for PSP/PTP/reg, another intrachain disulfide bridge.

In summary, the HIP sequence, like PSP/PTP/reg protein, ismade of a single CRD of C-type lectin, linked to a signal pep-tide required for the soluble lectin to be secreted.

Northern Blot Analysis of HIP mRNA. We then analyzedthe expression of the 0.9-kilobase HIP mRNA in various normal rat and human tissues. In rats, HIP mRNA was not detected in regenerating rat liver (data not shown), brain, muscle,colon, or adult and fetal liver (Fig. 6A). In contrast, the HIPmRNA was clearly detected in the duodenum, jejunum, andileum (Fig. 6A). In humans, HIP mRNA was found in the smallintestine and pancreas (Fig. 6.-I) but not in the colon, brain,

kidney, or lung (Fig. 6A). We also examined HIP mRNA in apanel of human HCCs and in two cases of human pancreaticcell hyperplasia (hyperinsulinism). The 0.9-kilobase transcriptwas clearly detected in 7 of 29 HCCs examined, while it was notshown in nontumorous areas or in fetal and adult normal liver(Fig. 6Ã„).It was also identified in the two cases of hyperinsulinism tested (Fig. 6B). As this gene was expressed in normal smallintestine and pancreas as well as in HCC, it was designated"HIP" (hepatocellular carcinoma, small intestine, pancreas).

Finally, to confirm that the nucleotide sequence of the HIPcDNAs, obtained from tumor liver, was identical to the normalgene sequence, we synthesized a cDNA library from humanileum poly A+ RNA and screened it with the HIP probe. This

allowed us to clone another HIP cDNA and to show that thecoding sequence in the HIP cDNA obtained from the normaland tumorous tissues was identical (data not shown). Southernblot analysis showed a single band with EcoRl digestion andtwo bands with Hindlll digestion. There was no evidence of agross rearrangement in the tumor (C. H.) used for the cDNAlibrary (data not shown).

DISCUSSION

We report the identification of a novel gene isolated from ahuman HCC cDNA library by means of differential screening.The gene was named HIP since its expression is markedly increased in human primary liver cancer while it is normallyfound in the small intestine and pancreas. Sequence comparison showed that the HIP-encoded protein presents significanthomology with the human and rat reg proteins, the human andrat PSPs, and the human and bovine thread proteins (PTP andBPTP); the human PTP, PSP, and reg proteins are in factidentical products of a single copy gene (7). PTP/reg and PSPonly differ by three amino acids located in the NH2 extremity ofthe protein. Other minimal nucleotide sequences are observedbetween PSP/PTP and reg in the 5' untranslated region of the

mRNA (8) and in the first intron of the gene (7). PTP/reg andPSP sequences thus probably correspond to allelic variants ofthe same gene.

A bovine gene, homologous to PSP/PTP/reg, has recentlybeen isolated from a normal pancreatic cDNA library (23).Interestingly, amino acid sequence comparison showed that theHIP protein is more similar to the bovine PTP (68.5% homology) than to the human and rat PTP/PSP/reg sequences (49%and 42%, respectively). The bovine PTP (BPTP) sequence isalso closer to HIP (68.5%) than to human PTP/PSP/reg protein (45%). HIP- and BPTP-translated proteins also have thesame length (175 amino acids), but are 9 and 10 amino acidslonger than the human and rat PSP/PTP/reg proteins, respectively. It is therefore very likely that the BPTP protein is thebovine homologue of the HIP protein and is more distantlyrelated to the PSP/PTP/reg human proteins. HIP and BPTP,on the one hand, and the human and rat PSP/PTP/reg, on theother hand, therefore define a new protein family which shareswith the C-type lectins all the structural characteristics of theirCRD (11). The recent elucidation of the three dimensionalstructure of the CRD of the crystallized mannose binding Protein A at high resolution greatly facilitates sequence comparisons between the C-type lectins (25). Owing to on intermediatestep of two-dimensional cluster analysis, it is possible to assignwithout ambiguity the various stretches of HIP and the otherrelated sequences to the structural domain delimited by thecrystallographic approach. The most conserved domains correspond either to the hydrophobic core of the protein or the

5092




LI

HshiptfSKAYGSHCHsregttTNAYRSYCAclec

ttWTSFGSNCMrlec2 ttWVTSENKCSplec Â»KALDGREYPmlecMDYEHspgcf

OGWHKFQGQHspgccttGWNKYQGHRnlecl

ffNWVEYEGSHslec2ftNWVEHQGSRnjnbpa

ffGKKSGKKFHsmbptfGKQVGNKFMm

IhrttIHHGTHCWHselamttLIKESGAWYALFLSPKYYFNEDREYRFFAVSL.

FHVPLEKA.LIETELKY.ILFSDETM.CYKYFAHRRCYRHFPDRECYWFSSSVKCYWFSHSGKFVTNHERMFLTNGEIMTYHYSEKPMSYNTSTEAMc;TT8

NKTTpRpTNTWTDADLACQKRPWVDADLYCQNMNWAEGEQFCQSFS

WMVAHGVCA.RLWHQAWHECA.RHYADAGTYCQSRGWDAAERECRLQGWVDAERRCREQQWTEADKYCQLENWAEAEKYCQLENFSKVKALCSEL.FEKVKALCVKF

.WENARKFCKQNYYDEASAYCQQRYS

GNLVSVLS.SGNLVSVLT.VPSRGDIDSIGHLVSIHS.

D. -SRARLASID. .D QQLVTIE. .D MALVSSA..AHLTSILS.SHLSSIVT.AHLWVTS

.AHLWINS.RGTVAIPRNQASVATPRNTDLVAIQN.THLVAIQN.

GAEGSFVSSLVQAEGAFVASLIETEQNFVYHYF

AADQAWEPLSSADKNNAIIDL

MRDSTMVKAILHEEQMFVNRVGHEEQEFVNNNAWEEQRFVQQHMWEEQKFIVQHTAEENKAIQEVAAAENGAKEE.

.KREIEYLENTLKEEIEYLNSILKSIGNSYS.

.KESGTDDF..ETSTKDDTTP

SEVKRWGKSH.

AFTEVKGH ..HDQDGPLNPFKTSPKSPYSYSPS

....

YVWIGLNVWIGLEMWLGFKMWIGLNLWLGG

DYWVGAYQWIGLYQWIGL

NTWIGLNTWIGL.AFLGI

.AFLGIYYWIGIYYWIGIHHNSNDNN

TTT

Tj?RDPTQGTEPNGEGWDPKK

NRR .WDRTT

...EGN ..FYDSAN..DAAV.WDEYSSSRDYGRPF

LQDGAYNFDKMF

EHD.FDRTIEGD.F

DQ NGP .WDSDGS.WOEVT

[email protected]

GKM.WKV.... NNV .W

L2 1.3 L4

HshipEHsregHAclec

QMrlec2ASplecFPmlecLHspgcf

RHspgccRRnlecl

KHslec2KRnmbpa

MHsmbpVMmlhr

THselamVWSSSDVWSS

.GSWTD.GSDDS.HSWSPTGQWND

.GVWTD.GCWSD.GHWVD.GTWVD.GTYVT

.GGDLT.GNWV

..GTWV

..GTMN.

..YFA.LVS

..YKS.PND.

.FTA.S....HRN.AF.

..SFA.SLPTDSDL.STCLQYEN.PMQ.

.FEN.DYETGFKN.DYRHNYKN.RL.T.YSN.RL.

..YTN.NKTLTKEAEQKPLTEEAK

WERN.WGIGA.

WVGSN.WYATQYWSENN.WSPNE.WRPNQ

.WRPNQ.WRPGQ

.WAVTQWKKI>dWNEGENWGAGENWAPGEppppppP

PPPppppSTISSP

GHSSVNPGYDNGSG

EDDDESEDNYKHQ

..EHSNPQSWQLDSFFSAG

EDDNFFAAG ....EDDDWYGHGLGGGED

DNWHGHELGGSEDR1ÃœHGSG

[33NNAGSDEDNNKKSK

EDNNRQKDEDCASLSCVSLTCTQMVCVLIKCVHIWCVQIWCWII

CVVMICAHFT

CVEVQC

VTIVCVLLLCVEIYCVEIYRSTAFL.

.SSTGFQ.

.MGAGL.

..EDQYR.

..DTKPLY..SCKYN.

..WH

... ENGWH ....KGTD.G

PD.G. ...DN

GKNGIKRERDSGIKREKDVGRWKDYNKWKDVPNWIDLPQWIIDYNQWNDNDLLDDVGQWNDVP

EWNDVPHWNDDV

RWNDDFLWlÃ®IÃ®llSQWNDVPKWNDDAMWNDERCNVRCEDKCSSTCNDRCNVKCGGACNYH

CNYHCRRP

CLQVCQASCSTSCHKRCSKKLPY.VCKFTDFSF.VCKFKNRIIYLICKLPLYNF.VCEIVLMGY.

ICEPNHRRV.ICE.KELTY.TCKKGT

LPF.TCKKGTYRW.VCETEL

YRW.VCEKRRHTA.VCEFPAHLA.VCEFPIKAA

.LCYTASKLA.LCYTAA

[1]

WE*H*

FRETYDQÂ»LDD*

VACGQPPÂ»VACGEPPft

GKAN*NATGEVA*

CQPGSCNÂ»CTNTSCSÂ»

[1]

[2] [21] [21] [22]

Fig. 5. Comparison of sequences of the potential CRD. Hship. HIP; Hsreg, PSP/PTP/reg (7); Aclec, sea urchin lectin (39); Mrlec2, acorn barnacle lectin Bra 2 (30);Splec, flesh fly lectin (35); Pmlec, tunciate lectin (34); Hspgcf, human fibroblast proteoglycan core protein (42); Hspgcc, human cartilage proteoglycan core protein (41);Rnlecl, rat hepatic lectin (39); Hslec2, human asialoglycoprotein receptor (43); Rnmbpa, rat mannose-binding Protein A (38); Hsmbp, human mannose-binding protein(32); Mmlhr. mouse lymphocyte homing receptor (40); Hselam, human leukocyte adhesion molecule (44); *. stop codon; Â».sites where sequence alignments wereinterrupted; black boxes, proposed calcium-binding residues in the mannose-binding protein: [/) or (2), potential consensus for calcium binding in the C-lectin family.Precise residue is nevertheless difficult to assign in some proteins. Sequence stretches corresponding to defined domains of Rnmbpa are boxed, a helices and .Â¡strandsnumbering is that proposed by Weiss el al. (25).

crucial residues which bind the calcium ions characterizing thislectin family. In this respect, it is interesting to note the aminoacid divergence in the second calcium binding site of HIP andPSP/PTP/reg proteins relative to most of the other CRDs. Asimilar variation is, however, found in some of the eight CDRrepeats of the mannose receptor and at the first calcium bindingsite in Loop 1 for several selectins (32). It might have someconsequences on the affinity, the specificity, or the pH dependence of the carbohydrate binding.

On the basis of sequence identities, functional properties, andmultidomain arrangements of the proteins and the exon-intronorganization of the genes, Drickamer et al. proposed a classification of the C-lectins into four families, i.e., I, proteoglycans;II, endocytosis receptors for serum glycoprotein (type II receptors); III, CRDs connected to collageneous domains; and IV,cell adhesion molecules with a lectin domain (33). HIP andPSP/PTP/reg proteins have structural characteristics whichmake them difficult to classify in the four C-lectin familiesdefined above. Contrary to the other C-lectin families, HIP- andreg-related proteins are built of a single CRD linked to a signalpeptide, likely to be cleaved in the mature protein when secretedinto the extracellular environment. In that respect, they areclose to invertebrate lectins like sea-urchin lectin (29), acorn

barnacle lectins Bra-2 and Bra-3 (30,31), tunicate lectin (34), orthe flesh-fly lectin (35) (although this latter has an extra domain). In addition, the divergence between the HIP/PSP/PTP/reg and each of the four C-type lectin subclasses is similar tothat between the four subclones of lectins. Moreover, althoughthe structure of the HIP gene is not yet known, the humanPSP/PTP/reg CRD is encoded by four exons instead of eitherone or three for the other CRD of C-lectins, and has a verydifferent position of exon-intron junctions (7). It is thereforeplausible that HIP/PSP/PTP/reg proteins belong to a fifth classof C-lectins, more closely related to the invertebrate lectins thanto the other types of animal C-lectins. Although the role of HIPis unknown, the sequence homology provides working hypotheses which have now to be tested.

With regard to its expression, the HIP cDNA was obtainedby differential screening of a cDNA library generated from ahuman HCC. It was expressed at high levels in the tumor tissueof 7 of 29 human HCCs. In contrast, HIP mRNA was notdetected in normal fetal or adult liver nor in regenerating ratliver; these observations raise the issue of its role in liver celldifferentiation and proliferation. In this respect, it is interestingto note that the reg protein has been proposed to be a pancreaticislet cell regenerating factor; it was initially identified in the rat

5093




123456 7 â€¢¿� Â» 10 11 11 IJ 14

KB

0.9

KB

B.B tÃŽ

BFig. 6. A, RNA from various normal tissues. The probe was the BamHl-

BamHl fragment purified from HIP cDNA. Lanes I to 6, human tissues; Lane I,colon, 40 ng of total RNA (2 days); Lane 2, ileum, 40 jig of total RNA (2 days);Lane 3, brain, 2.5 Mgof polyA* RNA (4 days); Lane 4, kidney, 40 jig of total RNA

(4 days); Lane 5, lung, 40 ng of total RNA (4 days); Lane 6, pancreas, 40 ng oftotal RNA (1 day). Lanes 7 to 14, rat tissues; Lane 7, duodenum. 40 jig of totalRNA (7 days); Lane 8, colon. 40 ng of RNA (7 days); Lane 9, ileum, 40 ng of totalRNA (7 days); Lane 10, jejunum. 40 Mgof total RNA (7 days); Lane 11, brain, 40Â»igof total RNA (7 days); Lane 12. fetal liver, 40 jig of RNA (7 days); Lane 13,adult liver, 40 <*gof total RNA (7 days); Lane 14, muscle, 40 wg of total RNA (7days). B, Lanes I and 2, Patient K.; Lane 1, HCC tumorous area, 40 ng of totalRNA (2 days); Lane 2, nontumorous area of HCC. 40 ng of total RNA (2 days);Lanes 3 and 4, Patient A.; Lane 3, HCC tumorous area, 40 Â»igof total RNA (5days); Lane 4, nontumorous area of HCC, total RNA (5 days); Lane 5, normaladult liver, 40 *igof total RNA (5 days); Lane 6, normal fetal adult liver. 40 uv,oftotal RNA (5 days); Lane 7, hyperinsulinic pancreas, 40 ng of total RNA (2 days).(Exposure times are in parentheses.)

by differential screening of a cDNA library derived from a regenerating pancreatic islets-derived cDNA library (10). The human reg gene was then isolated and found to be expressed inpancreatic cells and, to a lesser extent, in the kidney and thegastric mucosa (7). In rats, expression of the reg protein occursin normal acinar pancreatic cells, whereas its location in /3-cellsecretory granules is only observed in regenerating islets (12).Reg expression has also been shown to increase in some humancolon and rectal cancers (10). On the other hand, expression ofthe PSP or FTP protein is decreased in human chronic calcifying pancreatitis and is an inhibitor of CaCO3 crystal growthin vitro, suggesting that it could normally control CaCO3 crys-tallorea in pancreatic juice (8, 14). Finally, this pancreatic protein has also been shown to accumulate in the developing human brain and Alzheimer's disease (13).

With regard to the tissue specificity of HIP gene expression,HIP RNA expression in adults was only detected in the smallintestine and pancreas. We will investigate in detail which celltypes express the gene. In this view it should be remembered

that the liver, pancreas, and small intestine have a commonembryonic origin. In addition, the so-called "oval" liver cells,

presumably of ductal origin, proliferate in rodents in responseto various chemical carcinogens and are candidates for generating HCC (1). They can differentiate into both pancreatic cellsand hepatocytes (1) and, in certain conditions, give rise to intestinal metaplasia (36). Conversely, rat pancreatic ductularcells can differentiate into hepatocytes following a copper depletion-repletion regimen (37). Given the expression pattern ofHIP, as well as the potential role of lectins in differentiationand development (34), it will be interesting to explore its possible expression in candidate liver progenitor cells. In conclusion, we have identified a new gene whose sequence homologywith the PSP/PTP/reg gene shows the existence of a novelfamily of genes with different tissue specificities, possibly involved in liver, pancreatic, and intestinal differentiation and/orcell proliferation.

ACKNOWLEDGMENTS

We thank A. Lamouroux, M. Darmon, and J. Mallet for their interest and contribution to this work, D. Houssin and D. Franco for theliver samples, J-P. Bonnefont and A. Lemoine for pancreatic samples,and I. de Waziers for intestinal samples. We also thank J. Weimann forproviding some rat RNAs.

REFERENCES

1. Fausto, N. Hepatocyte differentiation and liver progenitor cells. Curr. Opin.Cell Biol., 2: 1036-1042, 1990.

2. Bulinaseli. P., Enzmann, H., Hacker, H. J., Weber, E., and Zerban, H.Comparative pathobiology of hepatic pre-neoplasia. In: P. Bannasch, D.Keppler, and G. Weber (eds.), Liver Cell Carcinoma, Falk Symposium 51,pp. 55-76. Dordrecht, the Netherlands: Kuver Academic Publishers, 1989.

3. Scott, M. R., Westphal, K. H., and Rigby, P. W. J. Activation of mouse genesin transformed cells. Cell, 34: 557-567, 1983.

4. Yamamoto, M., Machara, Y., Takahashi, K., and Endo, H. Cloning of sequences expressed specifically in tumors of rat. Proc. Nati. Acad. Sci. USA,80: 7524-7527. 1983.

5. Rhyner, T. A., Faucon Biguet, N., Berrard, S., Borbely, A. A., and Mallet, J.An efficient approach for the selective isolation of specific transcripts fromcomplex brain mRNA populations. J. Neurosci. Res., 16: 167-184, 1986.

6. Sargent, T. D. Isolation of differentially expressed genes. Methods Enzymol.,152: 423-432, 1987.

7. Watanabe. T., Yonekura, H., Terazono, K., Yamamoto, H., and Okamoto,H. Complete nucleotide sequence of human reg gene and its expression innormal and tumoral tissues. J. Biol. Chem., 265: 7432-7439, 1990.

8. Giorgi, D., Bernard, J-P., Rouquier. S., lovanna, J., Sarles, H., and Dagorn,J-C. Secretory pancreatic stone protein messenger RNA, nucleotide sequence, and expression in chronic calcifying pancreatitis. J. Clin. Invest., 84:100-106, 1989.

9. Gross, J., Carlson, R. I., Braver, A. W., Margolies, M. N., Warshaw, A. C,and Wands, J. R. Isolation, characterization, and distribution of an unusualpancreatic human secretory protein. J. Clin. Invest., 76: 2115-2126, 1985.

10. Terazono, K., Yamamoto, H., Takasawa, S., Shiga, K., Yonemura, Y.,Tochino, Y., and Okamoto, H. A novel gene activated in regenerating islets.J. Biol. Chem., 263: 2111-2114, 1988.

11. Drickamer, K. Two distinct classes of carbohydrate-recognition domains inanimals' lectins. J. Biol. Chem., 263: 9557-9560, 1988.

12. Terazono, K., Uchiyama, Y., Ide, M., Watanabe, T., Yonekura, H., Yamamoto, H., and Okamoto, H. Expression of reg protein in rat regeneratingislets and its co-localization with insulin in the beta cell secretory granules.Diabetologia, 33: 250-252, 1990.

13. Ozturk, M., De La Monte, S. M., Gross, J., and Wands, J. R. Elevated levelsof an exocrine pancreatic secretory protein in Alzheimer disease brain. Proc.Nati. Acad. Sci. USA, 86: 419-423, 1989.

14. Multigner, L., Sarles, H., Lombardo, D., and de Caro, A. Pancreatic stoneprotein. II. Implication in stone formation during the course of chroniccalcifying pancreatitis. Gastroenterology, 89: 387-391, 1985.

15. MÃ¶rÃ¶y,T., Etiemble, J., Trepo, C., Tiollais, P., and Buendia, M. A. Transcription of woochuck hepatitis virus in the chronically infected liver. EMBOJ., 4: 1507-1514, 1985.

16. Cariani, E., Lasserre, C., Seurin, D., Hamelin, B., Kemeny, F., Franco, D.,Czech, M. P., Ullrich, A., and Brecho!, C. Differential expression of insulin-like growth factor II mRNA in human primary liver cancers, benign livertumors, and liver cirrhosis. Cancer Res., 48: 6844-6849, 1988.

17. Watson, C. J., and Jackson, J. F. An alternative procedure for the synthesisof double-stranded cDNA for cloning in phage and plasmid vectors. In: DNACloning, a Practical Approach, Vol. I, pp. 79-88. City: Publishers, 1985.

5094




18. Devereux, J., Hacberli, P., and Smithies, O. A comprehensive set of sequenceanalysis for the VAX. Nucleic Acids Res., 12: 387-395, 1984.

19. Gaboriaud, C, Bissery, V., Benchetrit, T., and Mornon, X. Hydrophobiecluster analysis: an efficient new way to compare and analyse amino acidsequences. FEBS Lett., 224: 149-155, 1987.

20. Lemesle-Varloot, L., Henrissat, B., Gaboriaud, C., Bissery, V., Morgat, A.,and Mornon, J-P. Hydrophobie cluster analysis: procedure to derive structural and functional information from 2-D-representation of protein sequences. Biochimie (Paris), 72: 555-574, 1990.

21. Kozak, M. Compilation and analysis of sequences upstream from the translational start site in eukaryiotic mRNAs. Nucleic Acids Res., 12: 857-872,1984.

22. Rouquier, S., Verdier, J-M., lovanna, J., Dagorn, J-C., and Giorgi, D. Ratpancreatic stone protein messenger RNA, abundant expression in matureexocrine cells, regulation by food content, and sequence identity with theendocrine reg transcript. J. Biol. Chem., 266: 786-791, 1991.

23. De La Monte, S. M., Ozturk, M., and Wands, J. R. Enhanced expression ofan exocrine pancreatic protein in Alzheimer disease and the developing human brain. J. Clin. Invest., 86: 1004-1013, 1990.

24. Patthy, L. Homology of human pancreatic stone protein with animal lectins.B. J. Lett., 253: 309-311, 1988.

25. Weiss, W. L, Kahn, R., Fourme, R., Drickamer, K., and Hendrickson, A.Structure of the calcium-dependent lectin domain from a rat mannose-bind-ing protein determined by MAD phasing. Science (Washington DC), 254:1608-1615, 1991.

26. De Caro, A., Bonicel, J. J., Rouimi, P., De Caro, J. D., Sarles, H., andRovery, M. Complete amino acid sequence of an immunoreactive form ofhuman pancreatic stone protein isolated from pancreatic juice. Eur. J. Bio-chem., 168: 201-207, 1987.

27. Rouimi, P., de Caro, J., Bonicel, J., Rovery, M., and de Caro, A. The disulfidebridges of the immunoreactive forms of human pancreatic stone proteinisolated from pancreatic juice. FEBS Lett., 229: 171-174, 1988.

28. Cai, L., Harris, W. R., Marshak, D. R., Gross, J., and Crabb, J. W. Structuralanalysis of bovine pancreatic thread protein. J. Prat. Chem., 9: 623-632,1990.

29. Giga, Y., Ikai, A., and Takahashi, K. The complete amino acid sequence ofEchinoidin, a lectin from the coelomic fluid of the sea urchin Anthocidariscrassispina, homologies with mammalian and insect lectins. J. Biol. Chem.,202:6197-6203, 1987.

30. Muramoto, K., and Kamiya, H. The amino-acid sequence of multiple lectinsof the acorn barnacle Megalabanus rosa and its homology with animal lectins.Diodi iin. Biophys. Acta, 1039: 42-51, 1990.

31. Muramoto, K., and Kamiya, H. The position of the disulfide bonds and theglycosylation site in a lectin of the acorn barnacle Megalabanus rosa. Bio-chim. Biophys. Acta, 1039: 52-60, 1990.

32. Taylor, M. E., Conary, J. T., Lennartz, M. R., Stahl, P. D., and Drickamer,

K. Primary structure of the mannose receptor contains multiple motifs resembling carbohydrate recognition domains. J. Biol. Chem., 265: 12156-12162, 1990.

33. Bezouska, K., Crichlow, G. V., Rose, J. M., Taylor, M. E., and Drickamer,K. Evolutionary conservation of intron position in a subfamily of genesencoding carbohydrate-recognition domains. J. Biol. Chem., 266: 11604-11609, 1991.

34. Suzuki, T., Takagi, T., Furukohri, T., Kawamura, K., and Mitsuaki, N. Acalcium-dependent galactose-binding lectin from the junicate Pollyandro-carpa misakiensis, isolation, characterization, and amino-acid sequence. J.Biol. Chem., 2Â«5:1274-1281, 1990.

35. Takahashi, H., Komano. H., Kawagushi, N., Kitamura, N., Nakanishi, S.,and Natori, S. Cloning and sequencing of cDNA of Sarcophaga peregrinahumoral lectin induced on injury of the body wall. J. Biol. Chem., 260:12228-12233, 1985.

36. Elmore, L. W., and Sirica, A. E. Phenotopic characterization of metasplasicintestinal glands and ductular hepatocytes in cholangiofibrotic lesions rapidlyinduced in the caudate liver lobe of rats treated with furan. Cancer Res., 51:5752-5759, 1991.

37. Reddy, J. K., Rao, M. S., Yeldandi, A. V., Tan, X., and Dwidedi, R. S.Pancreatic hepatocytes. An in vivo model for cell lineage in pancreas of adultrat. Dig. Dis. Sci., 36: 502-509, 1991.

38. Drickamer, K., Dordal, M. S., and Reynolds, L. Mannose-binding proteinsisolated from rat liver contain carbohydrate-recognition domains linked tocollageneous tails, complete primary structures, and homology with pulmonary surfactant apoprotein. J. Biol. Chem., 26/: 6878-6887, 1986.

39. Drickamer, K., MamÃ³n, J. F., Binns, G., and Leung, J. O. Primary structureof the rat liver asialoglycoprotein receptor, structural evidence for multiplepolypeptide species. J. Biol. Chem., 259: 770-778, 1984.

40. Dowbenko, D. J., Diep, A., Taylor, B. A., Lusis, A. J., and Lasky, L. A.Characterization of the homing receptor gene reveals correspondence between protein domains and coding exons. Genomics, 9: 270-277, 1991.

41. Doege, K. J., Sasaki, M., Kimura, T., and Yamada, Y. Complete codingsequence and deduced primary structure of the human cartilage large aggregating proteoglycan, aggrecan, human-specific repeats, and additional, alternatively spliced forms. J. Biol. Chem., 266: 894-902, 1991.

42. Zimmermann, D. Râ€žand Ruoslahti, E. Multiple domains of the large fibro-blast proteoglycan, versican. EMBO J., 8: 2975-2981, 1989.

43. Spiess, M., and Lodish, H. F. Sequence of a second human asialoglycoproteinreceptor: conservation of two receptors genes during evolution. Proc. Nati.Acad. Sci. USA, 82: 6465-6469, 1985.

44. Ord, D. C., Eras, T. J., Zhou, L-J., Rambaldi, A., Spertini. O., Griffin, J.,and Tedder, T. F. Structure of the gene encoding the human leukocyte adhesion molecule-l(TQl,Leu-8) of lymphocytes and neutrophils. J. Biol.Chem., 265: 7760-7767, 1990.

5095



1992;52:5089-5095. Cancer Res Chantal Lasserre, Laurence Christa, Marie-Thérèse Simon, et al. A Novel Gene (HIP) Activated in Human Primary Liver Cancer

Updated version

http://cancerres.aacrjournals.org/content/52/18/5089

Access the most recent version of this article at:

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

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/52/18/5089To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



a novel gene (hip) activated in human primary liver cancer1 · in colon, brain, kidney, or lung. in...

Documents