Retinoblastoma and p53 tumor suppressor ... - The FASEB Journal

RESEARCH COMMUNICATIONS

Retinoblastoma hepatoma

and p53 tumor

suppressor

genes in human

cell lines KATHERINE

GALVIN,T

FREDERIC

TROALEN,*

BRIGITTE

BRZSSAC,t,I 14XICIER,

AND 5Laboratory

Hospital Villejuif,

of Molecular Oncology Centre Leon Berard, Lyon, France; tCancer Center, Massachusetts General East, Charlestown, Massachusetts 02129, USA; Service d’Immunologie Moleculaire, Institut Gustave RDussy, France, SLiver Unit, Hadassah Medical Center, Jerusalem, Israel

We analyzed the status of retinoblastoma and p53 genes in 10 human hepatoma cell lines. Polyclonal anti-peptide antibodies generated against peptides homologous to COOH-terminal and leucine-zipper domains of the retinoblastoma protein allowed us to identify two cell lines (Hep 3B and FOCUS) with abnormal expression. The same cell lines have both lacked p53 expression. In contrast to the retinoblastoma gene, the expression of the p53 gene was abnormal in six additional cell lines. Indeed, only the Hep G2 hepatoblastoma cell line (and its derivative Hep G2/2215) appeared to have normal pSS and retinoblastoma gene expression. Our studies indicate that p53 abnormalities are common but retinoblastoma gene aberrations are rare in human hepatoma cell lines.-Puisieux, A., Calvin, K., Troalen, F., Bressac, B., Marcais, C., Galun, E., Ponchel, F., Yakicier, C., Ji, J., Ozturk, M. Retinoblastoma and p53 tumor suppressor genes in human hepatoma cell lines. FASEB J. 7: 1407-1413; 1993. ABSTRACT

Key Words:

studies

hepatocellular

anti-RB

carcinoma

protein antibodies

. .

RB gene p53

immunoc/zemical

CARCINOMA (HCC),2 WHICH is a rare tumor and in the New World, is one of the most common cancers in the Far East and southern Africa (1). Eight chromosomal arms (4q, 5p, 5q, Bq, l3q, l6p, l6q, l7p) have been shown to display frequent deletions in HCC (1-3), which suggests that at least eight tumor suppressor genes are involved in the development of these tumors. The retinoblastoma (RB) gene, which is located on the long arm of chromosome 13, and the p53 gene located on the short arm of chromosome 17 are two tumor suppressor genes known to be deleted frequently and mutated in various human malignancies (4-16). p53 tumor suppressor gene is known to be commonly inactivated in human tumors (reviewed in refs 17, 18). Compared with p53, there has been a limited number of studies of the status of RB gene in various human cancers. It appears, however, that this tumor suppressor gene is inactivated only in a subset of human tumors such as retinoblastomas, osteosarcomas, and carcinomas of the breast, lung, bladder, and prostate (8, 10-12, 16, 19, 20). We and others previously reported on the abnormal expression of the p53 gene in several hepatocellular carcinomaderived cell lines (21, 22). No available data exists on the status of RB gene in HCC cells. This study compares the status of the p53 and RB tumor suppressor genes on 10 HCC cell lines commonly used for studies related to hepatocytespecific functions and HCC-associated cellular and molecu-

HEPATOCELLULAR

in Europe

0892-6638/93/0007-1407/$01.50.

© FASEB

lar changes in humans. We also describe production and characterization of rabbit polyclonal antibodies to the carboxyl-terminal and leucine zipper domains of the RB gene product.

MATERIAL Production

AND

METHODS

of antibodies

to RB protein

Anti-RB protein antibodies were raised in rabbits by injecting synthetic peptides homologous to the carboxyl-terminal region (TSTRTRMQKQKMN DSMDTSNKEEK) or to the leucine-zipper motif (LNTLCERLLSEH. PELEHIIWT) of human RB protein. Peptides corresponding to regions 662-683 (leucine zipper motif) and 793-816 (carboxyl end) of the human RB protein (14), respectively named PA 214 and PA 193, were synthesized using the solid phase method (Applied Biosystems Model 430A peptide synthesizer). The peptide PA 193 was modified with additional residues Gly and Tyr at the NH2-terminal portion to permit a specific coupling reaction with carrier protein. Synthesis of each peptide was performed under conditions that achieved a greater than 99.5% coupling efficiency in the monitoring of each amino acid. The completed peptides were purified by gel filtration, and the purity and identity of each peptide were checked by amino acid composition analysis, high-performance liquid chromatography, and sequencing (Applied Biosystems Model 477A protein sequencer). Peptides PA 193 and PA 214 were conjugated to keyhole limpet hemocyanin (KLH) with, respectively, benzidine and glutaraldehyde as coupling agents. New Zealand rabbits were initially immunized with 100 to 400 cg KLH-conjugated peptides, followed by boosts after 5, 6, and 7 wk. Sera obtained from immunized animals were used directly for immunochemical studies.

Cell lines The following 10 human hepatoma cell lines were used in these studies: Focus (23), HepG2(24), HepG2/22I5 (24, 25), Mahlavu (see ref 26), Hep3B (24), PLC/PRFI5 (27), Huh7 (28), HA22T/VGH(29), TongIHCC (30), and SK-Hep-l (31). Human fibroblast cell lines 2918 (kindly provided by S. Friend), and Detroit 551 (A1CC; CCL 110) and human osteosarcoma cell line U-2 OS (ATCC; HTB 96) were also used for control experiments. All cell lines were maintained in Dulbecco’s modified Eagle’s medium sup. plemented with 10% fetal calf serum, glutamine (2 mM), nonessential amino acids, penicillin, and streptomycin (21). The cells were screened for mycoplasma contamination and found to be negative except for PLC/PRF/5 and Mahlavu.

Immunoprecipitations Subconfluent cells were grown for I h in methionine-free ing 10% dialyzed fetal calf serum. Cells were labeled L55Slmethionine

(Amersham,

10-cm culture

dishes. Metabolic

specific activity labeling

>

medium containwith 0.2 mCi ml’

1000 mCi mol’)

with “P was performed

for 4 h in by incuba-

‘To whom correspondence should be addressed, at: Laboratory of Molecular Oncology, Centre Leon Berard, 28 rue Laennec, Lyon 69008, France. 2Abbreviations: HCC, hepatocellular carcinoma; RB, retinoblastoma; KLH, keyhole limpet hemocyanin.

1407

RESEARCH COMMUNICATIONS tion of cells in a phosphate-free medium in the presence of 0.4 mCi ml’ [‘2Pjorthophosphate for 3 h. Cells were then washed in PBS and proteins were extracted in I ml lysis buffer (50 mM Tris-HCI; pH 7.4, 0.25 M NaCI, 0.1% [v/v] Triton X-100, 5 mM EDTA, 50 mM NaF). All solutions contained a cocktail of protease inhibitors (50 mg 1-i PMSF, 10 mg 1’ soybean trypsine inhibitor, I mg 1’ aprotinin), and Na-vanadate (0.1 mM). Immunoprecipitation experiments were performed as described (21). Briefly, samples were precleared with 40 sl of 50% (v/v) protein-A Sepharose suspension (Pharmacia, Piscataway, N.J.) for 1-16 h at 4#{176}C. Two l of Mab Pab122 ascitis or 5 sl anti-RB rabbit serum was added. After 2 h incubation on a rocker at 4#{176}C and a 5-mm centrifugation at 13,000 x g supernatants were added to Eppendorf tubes containing 40 l of protein-A Sepharose beads and analyzed as described (21).

Southern

and Northern

of 17 p alleles

Total genomic DNA extracted from different cell lines was analyzed for the pYNZ22.l-deflned variable number of tandem repeats localized to the short arm of chromosome 17 (33). A pair of sense (5-CGAAGAG1tAAG’.R3CA CAGG-3’) and anti-sense (5-CACAGTC1 FATTtT1tAGCG-3’) primers was used for PCR amplification of total genomic DNA, as described by Horn et al. (34). Amplified fragments were separated on 2% agarose gel in TBE buffer, transferred to Nytran filters, and probed. The probe used was an oligonucleotide (5’-TI3TGAGACCCTCCCTTACAGAAGCAATGAG-3) that hybridizes to the 70-bp tandem repeats of pYNZ22 locus (34). The oligonucleotide probe was prepared by end-labeling with “P g ATP and T4 polynucleotide kinase (32).

Single-strand polymorphism sequencing of p53 gene

zipper and COOH-terminal peptides of the RB protein. Immunoprecipitation studies performed with and without competitive inhibition by immunizing peptides demonstrated that both antisera specifically recognize multiple forms of the RB protein (110-116 kDa) from this cell line (Fig. 1). When tested for 32P-labeled proteins, the antileucine zipper antisera recognized two additional bands of 60-69 kDa. As these two polypeptides have not been completely competed with the immunizing peptide, their specific immunorecognition remains to be demonstrated. Abnormal

blot analysis

Poly A(+) RNA was isolated from cells by a proteinase K/oligo (dT) cellulose extraction procedure, and l0-g aliquots of each preparation were subjected to electrophoresis under denaturing conditions (32). DNA was isolated using the proteinase K/phenol extraction protocol as described (32). DNA digested with Hind!II was separated by electrophoresis on agarose gels. Both RNA and DNA were transferred to Nytran filters (Schleicher and Schull, Keene, N.H.) and hybridized to “P probes as described (32). Randomly primed [‘5PIRB cDNA probe (14) was prepared as described (32). p53 RNA Northern blot analysis was done as described (21). Analysis

to leucine

analysis

and nucleic

RB protein in two HCC cell lines

HCC cell lines were initially screened for the phosphorylated forms of the RB protein using cell lysates prepared after 32p orthophosphate labeling. Figure 2 shows RB proteins immunoprecipitated with the anti-COOH-terminal antibody from nine HCC cell lines. A llO-kDa phosphoprotein was detected in all cell lines tested except Hep 3B and FOCUS. Both Hep 3B and FOCUS cell lines were further studied by immunoprecipitation using a mixture of both antisera after metabolic labeling with [35S]methionine. A 110-kDa polypeptide band was immunoprecipitated from all cell lines tested including FOCUS cells, with the exception of Hep 3B cells (data not shown). These studies suggested that Hep 3B cells were not able to produce the RB protein. In contrast,

Anti L-Zipper .Ab

35s

acid

II

32r

Anti C-Terminal Ab I

35s +

Genomic DNA was separately amplified with three sets of primers to generate PCR products encompassing, respectively, exons 5 + 6, exon 7, and exons 8 + 9 of p53 gene (sequence and experimental data available at request, detailed protocol to be published elsewhere).’ Labeled PCR products were diluted (1/20) in formamide loading buffer (32), denatured 10 mn at 95#{176}C, and loaded on 10% polyacrylamide gels with and without 5%

32 II -

+

glycerol.

RESULTS of the retinoblastoma

Studies

gene

-

pRB

Previous studies have shown that RB abnormalities are most often detected at the protein level (12, 19). Therefore, we began our investigation by producing antibodies to two different regions of the RB protein. After initial characterization for the specificity, both antibodies were used for characterization of the RB protein in HCC cell lines by immunoprecipitation after metabolic labeling with [35Sjmethionine and [32Pj phosphate. We also studied the structure of the RB gene as well as its expression by RNA Northern blotting. Production

and characterization

of antibodies against RB protein

As immunogens for the production of anti-RB protein antibodies, we used synthetic peptides corresponding to the carboxyl-terminal region, which has been shown to be highly immunogenic, and to the leucine zipper region, which is involved in protein-protein interactions. For initial studies the U-2OS osteosarcoma cell line, which is known to express normal RB protein, was used (15). After metabolic labeling with [35S]methionine or [52P]orthophosphate, cell lysates were used for separate immunoprecipitations with antisera 5Unsal,

1408

H.,

Vol. 7

et al., Proc. Nail.

November

1993

Acad.

Sci. USA.

(in press).

Figure

1. Immunoprecipitations from [32P]orthophosphate and [35S]methionine labeled U-2 OS osteosarcoma cells, using antibodies to the carboxyl-terminal and leucine zipper regions of RB protein. Comparison of immunoprecipitations in the presence and absence of competing immunogenic peptide shows that both antibodies specifIcally recognize the RB protein.

The FASEB Journal

PUISIEUX

ET AL.

RESEARCH COMMUNICATIONS =

u)

>#{176}T = c

c’.J (‘.1 Q. Q)

0 u

t

c’.j.2

a,

0

c’J

z

=

0

=

=

o_

=

I-

C/)

CJ

N-

-

0.

-

RB gene in most HCC cell lines

We investigated the status of the RB gene in cell lines with RB abnormalities by Southern blot hybridization of genomic DNA cut with HindIII, using an RB cDNA probe. Hep 3B and FOCUS cells showed no gross structural rearrangement of the RB gene compared with Mahlavu and Hep G2 cells, which both express apparently normal RB proteins (data not shown). In summary, studies of the RB gene identified two cell lines showing abnormal expression. Hep 3B cells did not express detectable RNA transcripts nor detectable p110 protein. Another cell line (FOCUS) displayed normal-sized RB transcripts and p110 protein. However, the protein product was not phosphorylated and it had an abnormally short halflife (less than 30 mm). These studies showed that the RB gene was abnormal in at least 2 out of 10 HCC cell lines. Both cell lines (Hep 3B and FOCUS) with defective RB protein had integrated HBV in their genome (23, 24).

-‘

>

0. )

Normal C-)

-

I

kDa

200-

92,5-

30’ 24d 69-

Det 550

Figure 2. Retinoblastoma protein (phosphorylated form) in human hepatoma cells as compared with a fibroblast cell line (2918). Antipeptide rabbit antisera generated against the carboxyl-terminal region of the retinoblastoma protein were used for immunoprecipitation experiments on cells labeled with [32P]orthophosphate as described in the text.

Hep3B Focus

FOCUS cells produced a normal-sized RB protein that was not phosphorylated under the experimental conditions used. Normal RB protein is known to display a long half-life (4-6 h) whereas the half-lives of the phosphorylated forms are lower (35). Therefore, we tested whether the half-life of the RB protein varied in different HCC cell lines.

Short half-4fe

Huh-?

of RB protein in FOCUS cells

The half-life of the RB protein was studied by pulse/chase metabolic labeling with [35S]methionine. No RB protein was detected in Hep 3B cells at any time. RB protein was detected in FOCUS cells, but it had an abnormally short halflife of less than 30 mm. In five other hepatoma cell lines and the human fibroblast cell line Detroit 551, the half-life of the RB protein was greater than 240 mm (Fig. 3).

Mahiavu Tong/HCC

Absence of RB transcripts in Hep 3B cells We decided

of RB in cell lines that by analyzing the RB transcripts. A Northern blot of poly A(+) RNA from FOCUS and Hep 3B cells, when compared with five other hepatoma cell lines, showed normal-sized transcripts in six cell lines including FOCUS (Fig. 4a and Fig. 4b). In contrast, there were no detectable RB transcripts in Hep 3B cells, which explains why this cell line lacked detectable RB protein. The same blot probed with a control 13-actin cDNA probe showed that the absence of RB transcripts in Hep 3B cells was not caused by an error of loading. RB transcripts showed

were

to further

abnormalities

consistently

by additional of poly A(+)

RETINOBLASTOMA

study

the status

at the

undetectable

experiments RNA (Fig.

AND

protein

level

in Hep

performed 4b).

p53 PROTEINS

3B

using

cells

when

higher

IN HEPATOMA

tested

amounts

CELLS

H422T/VGH SK-Hep-I Figure

3.

Retinoblastoma

protein

half-life

in human

hepatoma

cells. Cells were labeled with [35S]methionine, followed by with 1 mM methionine for 30 to 240 mm. Cell lysates were immunoprecipitation of retinoblastoma protein by rabbit as described in the text. Det 550 is the fIbroblast cell line 551 from AThC.

a chase used for antisera

Detroit

1409

RESEARCH COMMUNICATIONS

POLY 4(+) RNA

C\JcD (I, :3

C-)

W

o

a)

0

[

-JEc1)

cDc#{252} (i

0

28S-1 18S

Q)

ci)

II

-

-

/3 acti n

a

-RB mRNA

Figure 4. The expression of retinoblastoma gene in hepatoma cells as studied by RNA Northern blotting. Poly A( +) RNAs were separated by denaturing electrophoresis, transferred to Nytran membrane, and hybridized to “P-labeled retinoblastoma (top) or a 13-actin probe used as control (bottom).

Studies

of the p53

gene

cDNA

gene

To further characterize the status of the p53 gene in HCC cell lines, we studied the expression of the p53 gene by RNA Northern blotting as well as biochemical characteristics of p53 protein in all of the 10 cell lines used for RB studies. Several cell lines were also analyzed for the allelic status of the p53 locus and for the presence of mutations in exons 5 to 9 by SSCP and nucleic acid sequencing.

-/3-act in

b

p53 transcripts in HCC cell lines Figure 5 shows RNA Northern blot analysis of p53 transcripts in nine cell lines. Normal-sized transcripts were detectable in five cell lines. Three cell lines (Hep 3B, FOCUS, and Tong/HCC) did not express detectable p53 transcripts, whereas a forth cell line (HA22T/VGH) displayed sensitively shortened transcripts as previously noticed by Hosono et al. (22). All three cell lines showing no p53 expression were derived from patients infected with the hepatitis B virus and had integrated viral sequences. PLC/PRF/5, a cell line not shown in Fig. 5, also had integrated HBV sequences but expressed detectable albeit low levels of p53 transcripts (21). Similarly, the 2215 cell line, an HBV-expressing clone of Hep G2 cells obtained by in vitro DNA transfection (25), also had slightly decreased levels of p53 transcripts. These RNA studies indicated that at least four hepatoma cell lines (Hep 3B, FOCUS, Tong/HCC, HA22T/VGH) had abnormal expression of the p53 gene. The remaining six cell lines (Hep G2, HepG2/22l5, Huh 7, SK-Hep-1, Mahlavu, PLC/PRF/5), which displayed normal-sized p53 transcripts, were further analyzed for the presence of minor changes (point muta-

1410

Vol. 7

November

1993

tions, small deletions, or 5, 6, 7, 8, and 9 of the analyzed separately by cause more than 90% of man tumors have been

insertions) at the DNA level. Exons p53 gene were PCR amplified and SSCP. This region was chosen bep53 gene mutations identified in hulocalized into these exons (4).

p53 mutations in HCC cell lines By SSCP

analysis,

no gel shift

was observed

in exons

5, 6,

7, 8, and 9 of the p53 gene in three cell lines (Hep G2, Huh 7, and SK-Hep-1), suggesting the absence of noticeable mutation. In contrast, Mahlavu and PLC/PRF/5 cells displayed band shifts at exon 7, indicating the presence of a mutation (data not shown). p53 mutation present in these cell lines, both derived from African patients, was localized to the third nucleotide of codon 249 as shown in Fig. 6. Both cell lines displayed guanmne-to-thymine transversion leading to an arginine-to-serine change at this codon 249.

The FASEB Journal

PUISIEUX

Er

AL.

RESEARCH COMMUNICATIONS Increased

HuH7

Hep 3B

+

Hep G2 +

2215 FOCUS

+ +

TONG HA22T

Maihavu

I

SKHep 1 cJ’

-

Half-life

As shown previously, not all p53 mutations lead to an increased half-life of the p53 protein (36, 37), but an increased half-life of the protein often indicates the presence of a mutation at the p53 gene (see ref 6). We therefore analyzed cells metabolically labeled with either [35Sjmethionine or [32P]orthophosphate and demonstrated the expression of normalsized p53 proteins in cell lines Hep G2, HepG2/22l5, Huh 7, SK-Hep-1, Mahlavu, and PLC/PRF/5 (ref 21). As expected, there was no detectable p53 protein in FOCUS, Hep 3B, HA22T/VGH and Tong/HCC cells (data not shown). We compared the half-life of p53 in 6 HCC cell lines by pulse/chase labeling with [35S]methionine, and immunoprecipitations with the anti-p53 Mab Pab122. The half-life of pS3 in Mahiavu, SK Hep 1, and Huh 7 cells was greater than 240 mm. In other cell lines tested (PLCIPRF/5, Hep G2, and Hep G2/2215), p53 protein half-life was short and varied between 15 and 30 mm (data not shown, see ref 21 for Hep G2 and Huh 7 cell lines). Taken together, these studies indicated that all hepatoma cell lines with the exception of Hep G2 (and its derivative Hep G2/2215 obtained by transfection of HBV) displayed an abnormal pattern of p53 expression. As shown previously, mutational inactivation of p53 gene is often associated with an allelic loss at the pYNZ22 locus on the short arm of chromosome 17 (38). We therefore analyzed allelic polymorphism in different HCC cell lines.

VNTR CD

Figure

5. RNA Northern blot analysis of p53 expression in human hepatoma cells. Total RNAs were tested as described in the text with p53 and tubulin cDNA probes. Cells with integrated hepatitis B virus sequences (HBV) were shown on the bottom.

Control ATC

analysis

AIC

of

chromosome

1 7p in HCC cell lines

We studied pYNZ22 alleles in HCC cell lines. This probe recognizes a tandem repeated sequence present at a highly variable copy number in the vicinity of the p53 gene. Because 86% percent of a randomly selected population has been shown to be heterozygous at this locus, this probe has been useful for studies of allelic losses on chromosome l7p (refs 33, 34). Although SK-Hep-1 and Hep G2 cells have two alleles at this locus, the majority of the cell lines tested (FOCUS, Hep 3B, Mahlavu, Huh-7, and PLC/PRF/5) showed a single band. It is unlikely that these cell lines have two alleles with identical numbers of tandem repeats because of the extreme polymorphism of this locus.

Mahiovu G

of p53 in several HCC cell lines

PLCIPRF/5

G

AIC

G

,4

#{149}.4

IG

IG

Figure 6. Two hepatoma cell lines, both derived from African patients, display nucleotide of codon 249 of the p53 gene. See text for experimental conditions.

RETINOBLASTOMA

AND

p53 PROTEINS IN HEPATOMA CELLS

the same

guanine-to-thymine

transversion

at the third

1411

RESEARCH COMMUNICATIONS TABLE

1. Status

of retinoblastoma

and

p53 genes

in hepatitis

B virus-positive

and negative

Stat us of retinoblastoma gene expression

Cell line Hep 3B FOCUS Tong/HCC HA22T/VGH PLC/PRF/5 Huh-i SK-Hep-1 Mahlavu HepG2 HepG2/2215

HBV DNA

RB mRNA

RB protein

Yes Yes Yes No Yes No No No No Yes

Absent Present Present nt. Present Present Present Present Present Present

Absent Present Present Present Present Present Present Present Present Present

/zepatoma

cell lines Status

Protein 1/2 life -

Decreased Normal Normal Normal Normal Normal Normal Normal Normal

l7p alleles

p53 Gene

1 I nt. nt. 1 I 2 1 2 nt.

Deletion N nt. nt. N N N + R N N nt

Point mutation5 nt. nt. nt. nt. 249(G-’T) N.D.’ ND. 249(G-’T) N.D. N.D.

of p53 gene

mRNA

Protein

Absent Absent Absent Truncated Traces Present Present Present Present Present

Absent Absent Absent Absent Present Present Present Present Present Present

Protein 1/2 life -

240 >240 >240