|
Martins
C, Kedda MA, Kew MC, MRC/CANSA/University
Molecular Hepatology Research Unit, Department
of Medicine, University of the Witwatersrand Medical School, 7 York
Road, Parkto
wn 2193, Johannesburg, South Africa.
Supported in part by grants/bursaries from the University of the
Witwatersrand, and the Foundation for Research Development,
Pretoria, South Africa
Correspondence to: Professor M.C. Kew, Department of Medicine,
Medical School, 7 York Road, Parktown, 2193, Johannesburg, South
Africa.
Telephone:
+27(0)11 488 3626, Fax. +27(0)11 643 4318
Email:014anna@chiron.wits.ac.za
Received:
1999-08-10
Subject
headings: carcinoma,
hepatocellular; southern African b
lacks; cumulative LOH; tumor suppressor genes; microsatellite
genomic instabilit
y; liver neoplasms
Martins
C, Kedda MA, Kew MC. Characterization of six tumor suppressor genes
and microsatellite instability in hepatocellular carcinoma in
southern African blacks. World
J Gastroentero, 1999;5(6):470-476
Abstract
AIM: To analyse cumulative loss
of heterozygosity (LOH) of ch
romosomal regions and tumor suppressor genes in hepatocellular
carcinomas (HCC
s) from 20 southern African blacks.
METHODS: p53, RB1, BRCA1, BRCA2, WT1 and E-cadherin- ge
nes were analysed for LOH, and p53 gene was also analysed for the
codon 249
mutation, in tumor and adjacent non-tumorous liver tissues using
molecular tech
niques and 10 polymorphic microsatellite markers.
RESULTS: p53 codon 249 mutation was found in 25% of the subj
ects, as was expected, because many patients were from Mozambique, a
country wit
h high aflatoxin B1 exposure. LOH was found at the RB1,
BRCA2 and WT1
loci in 20%(4/20) of the HCCs, supporting a possible role of these
genes in HC
C. No LOH was evident in any of the remaining genes. Reports of
mutations of p
53 and RB1 genes in combination, described in other populations,
were not
confirmed in this study. Change in microsatellite repeat number was
noted at 9
/10 microsatellite loci in different HCCs, and changes at two or
more loci w
ere detected in 15%(3/20) of subjects.
CONCLUSION: We propose that microsatellite/genomic
instability
may play a role in the pathogenesis of a subset of HCCs in black
Africans.
INTRODUCTION
The evolution of cancer is thought to occur from the stepwise
accumulation of ge
netic aberrations in the same cell. These include loss of function
of tumor supp
ressor genes, activation of proto-oncogenes, faulty DNA mismatch
repair, and th
e integration of viral DNA[1,2].
Hepatocellular carcinoma (HCC) is a le
ading cause of death in both Africa and the Far East, resulting in
at least 310000 deaths worldwide each year[3].
HCC is multifactorial in aet
iology and its
pathogenesis is complex. The major risk factors involved in the
development of the tumor are chronic HBV and HCV infections,
cirrhosis and aflatoxin B1 (AFB) exposure[4,5].
Heavy
dietary AFB intake is thought to cause a guanine (G) to thymine (T)
transversion at the third base of codon 249 of the p53 gene, and for
this reason clustering of this point mutation occurs in HCCs from
Africa and China[6-8].
In addition, other mutations in, or deletions of, the p53 gene (on
chromosome 17p13.1) are found with relatively high frequency in
human HCCs in other countries. The functional loss of this tumor
suppressor gene, as well as its abnormal expression, have been
proposed to play a significant role in HCC development[9]
or at least in the development
of a subset of HCCs. The majority of p53 alterations reported to
date have loss of one allele accompanied by mutations of the second
allele[10].
Abnormalities of the p53 gene, such as gene mutation, deletion, or
the nuclear accumulation of mutant p53 protein have also been found
to correlate with increased allelic loss at the Breast Cancer
Susceptibility Gene 1 (BRCA1) locus (17q21). This gene is thought to
encode a transcription factor which acts as a tumor suppressor[11].
LOH of the BRCA1 gene in HCC was reported in a Korean study[12].
The Breast Cancer Susceptibility Gene 2 (BRCA2) (13q12-13) product
is thought to be a tumor suppressor[13]
involved in cellular
proliferation and differentiation[14],
and may be involved in the development of HCC[15].
LOH at the BRCA2 locus has been reported in HCC[15,16]
and it has been suggested that
mutations of the BRCA2 gene may be involved in hepatocarcinogenesis[15].
The retinoblastoma (RB1) gene (13q14.2) product (pRB) functions as a
cell cycle regulator[17],
and its absence leads to unrestricted cell growth. Although there is
no definite evidence that mutations of the RB gene are involved in
HCC, LOH of the RB1 gene has been documented in human HCCs[18,12].
LOH of WT1 and 11p13 have been reported in human HCCs[12,16].
WT1 appears to be involved in proliferation, differentiation and
apoptosis[19,20].
The product of the E-cadherin (Uvomorulin) gene (16q221) is the
primary adhesion molecule in epithelium[21].
Loss of function of E-cadherin may lead to decreased cell-cell
adhesion[22],
cellular phenotypic changes, and the development of invasive
properties[23].
In HCC, multicentric development and the formation of intrahepatic
metastases is common[24].
LOH on chromosome 16q has been previously reported to be important
in the initiation or progression of HCC[25,26].
Polymerase
chain reaction (PCR) amplification of microsatellites (sequences
uniformly distributed throughout the human genome) provides a simple
and effective method of rapidly detecting loss of heterozygosity/microsatellite
instability (LOH/MI)[27].
Microsatellite instability is defined as the loss or gain o
f microsatellite repeats at 2 or more loci and is detected by the
presence of extra bands or band shifts between tumor and non-tumorous
tissue DNA.
In
this study, we examined the G-T transversion at codon 249 of the p53
gene, LOH of the p53, RB1, BRCA1, BRCA2, WT1 and E-cadherin genes,
and micr
osatellite instability at 10 loci flanking these genes, in HCC and
adjacent non
-tumorous liver, from 20 southern African blacks.
MATERIALS AND METHODS
Subjects
The subjects included 20 southern African black men, aged
between 20 and 40 years. HCC tissue and matched non-tumorous liver
were obtained at necropsy or durin
g surgical resection. DNA was extracted from the tissues using a
modified “salt
ing-out” procedure[28].
HBV markers
The HBV status of the subjects was determined previously
using commercially available kits to detect HBV markers in serum
(Abbott Labs, Chicago, IL, USA).
LOH and microsatellite instability
Microsatellite instability (MI) and loss of heterozygosity (LOH)
studies were carried out by PCR and gel electrophoresis using
polymorphic repeat markers (Table 1).
PCR
products of the polymorphic loci p53, D17S846 and RB1.20 were
resolved on 4% composite agarose gels, while radioactively labeled
PCR products of the remaining loci were resolved on polyacrylamide
gels, and viewed by autoradiography. Band mobility shifts between
tumor and matched non-tumorous liver DNA wer
e scored as a change in allele repeat number. LOH was characterized
by the disappearance of one band or a considerable (>80%)
decrease in band intensity in heterozygotes, whilst microsatellite
instability was determined by expansion and/or contraction of
microsatellite sequences.
PCR for LOH and MI
A standard PCR protocol (primers, Table 1) was followed for
the p53, WT1, D1
3S137, RB(1.20), D13S120, D13S127, D17S855, and D17S846 loci. Each
PCR reaction
consisted, at final volume, of 100ng DNA, 1U Taq DNA
polymerase (Prom
ega, Madison, USA), 1×buffer, 1mM each dATP, dTTP, dGTP, 0.1mM d
CTP, 0.025μCi α32 P dCTP, and 50pmol of each pr
imer; in a total volume of 50μL, amplification for 30 cycles
of denaturation at 94℃
for 30s, annealing 55℃
for 30s, extension at 72℃
for 1min, and a final cycle of 72℃
for 10 minutes.
The
PCR reaction for the D16S301 and D16S260 loci (primers, Table 1)
consisted, at final volume, of 100ng DNA, 1U Taq- DNA Polymerase, 1×buffer,
0.1
% gelatin, 1mM each dGTP, dATP, dTTP, 0.1mM dCTP, and 0.025μCi
α 32P
dCTP, 50pmol of each primer; in a total volume
of 25μL, amplification for 25 cycles of denaturation at 94℃
for 1min, annealing at 55℃
for 2min, extension at 72℃
for 2.5min, and a final cycle of 72℃
for 10 minutes.
p53 codon 249 mutation
The p53 codon 249 mutation was detected by PCR-RFLP
using primer sequences
F3 and R3 (Table 1), and confirmed by sequence analysis. The PCR
reaction consis
ted of, at final volume[6],
100ng DNA, 2.5U of Taq-DNA polymerase (Promega), 1×buffer, 1mM MgCl2,
0.8mM each of dCTP, dATP, dGTP, dTTP, and 50pmol of each primer; in
a total volume of 50μL, amplification for 30 cycles of
denaturation at 94℃
for 15s, annealing at 56℃
for 15s, and extension at 72℃
for 30s. The 110bp PCR product was sized on ethidium bromide stained
agarose gels against a 100bp DNA ladder (Promega). A G to T
transversion at the third base of codon 249 was detected by the
presence or absence of a HaeⅢ
restriction site[6].All
samples shown by digestion to have the codon 249 mutation were
sequenced in both directions both upstream and downstream in
separate reactions, to confirm the presence of the mutation.
Table 1 PCR primers
|
Gene/Locus
|
Primer
|
Primer
sequence
|
Amplicon
|
Amplicon
length
|
|
p53
|
p53F3
|
5′GTTGGCTCTGACTGT-ACCAC
|
exon
7 spanning codon 249[6]
|
110bp
|
|
|
p53R3
|
5′CTGGAGTCTTCCAGT-GTGAT
|
|
|
|
p53
|
p53ivs1a
|
5′GCACTTTCCTCAACTCTACA
|
ALU
sequence within intron 1
|
200bp-300bp
|
|
|
p53ivs1b
|
5′AACAGCTCCTTTAATGGCAG
|
of
p53 gene[43]
|
|
|
D13S120
|
1353L
|
5′ATGACCTAGAAATGATACTGGC
|
(AC)73repeat
at D13S120[44]
|
112bp-136bp
|
|
(BRCA2)
|
1353R
|
5′CAGACACCACAACACACATT
|
|
|
|
D17S846
|
FF
|
5′TGCATACCTGTACTACTTCAG
|
(GGAA)25
repeat a
t D17S846[45]
|
250bp-300bp
|
|
(BRCA1)
|
RF
|
5′TCCTTTGTTGCAGATTTCTTC
|
|
|
|
D17S855
|
FS
|
5′GGATGGCCTTTTAGAAAGTGG
|
AC
repeat at D17S855[46]
|
145bp
|
|
(BRCA1)
|
RS
|
5′ACACAGACTTGTCCTACTGCC
|
|
|
|
WT1
|
400
|
5′AATGAGACTTACTGGGTGAGG
|
AC
repeat within 3′
untranslated
|
100bp-200bp
|
|
|
401
|
5′TTACACAGTAATTTCAAGCAACGG
|
sequence
of WT33[47]
|
|
|
RB1
|
B57
|
5′TGTATCGGCTAGCCTATCTC
|
[CTTT(T)]n
(n=14-26) repeat
|
400bp-600bp
|
|
|
B103
|
5′AATTAACAAGGTGTGGTGGT
|
within
intron 20 of RB gene[48,49]
|
|
|
D13S127
|
1341L
|
5′CAGATATGTACTCATGCACATG
|
(AC)35
repea
t at D13S127[44]
|
130bp-142bp
|
|
(BRCA2/RB1)
|
1341R
|
5′AAACAAATGAGTTTGGCTGT
|
|
|
|
D13S137
|
F
|
5′TTTCCTCATTCTTTCCCAATTG
|
(GT)22
repeat at
D13S137[50]
|
±135bp
|
|
(RB1)
|
R
|
5′CAGGAGGGATGGACTCACTTC
|
|
|
|
E-cadherin
|
E-cadF1
|
5′GATCCTAAGGACAAATGTAGATGCTCT
|
D16S301
locus polymorphic
|
146bp
|
|
|
E-cadF1
|
5′AGCCACTTCCCAGAACTTGGCTTCC
|
AC
region[51]
|
|
|
E-cadherin
|
E-cadF2
|
5′GGTTGAGATGCTGACATGC
|
D16S260
locus polymorphic
|
±234bp
|
|
|
E-cadR2
|
5′CAGGGTGGCTGTTATAATG
|
AC repeat region[52]
|
|
Note:
WT1: Wilm′s
tumor gene; RB1: Retinoblastoma gene; BRCA1: Brea
st cancer susceptibility gene 1; BRCA2: Breast cancer susceptibility
gene 2;
bp: base pairs.
Sequencing
All sequencing was carried out using the Sequenase PCR
Product Sequencing Kit (United States Biochemical Corp., Cleveland,
Ohio), according to the manufacture
r′s
instructions.
RESULTS
HBV status
Seven patients were currently infected with HBV (5 of these
were HBsAg-positive
; HBeAg-negative; the HBeAg status of the remaining 2 was unknown),
and 6 were
previously infected (anti-HBc and anti-HBs-positive). The HBV status
of the
remaining patients was not known (Table 2).
LOH/MI analyses
LOH was noted for the WT1 (1/13 subjects), RB (1.20) (1/10
subjects),
D13S120 (1/20 subjects) and D13S127 (2/14 subjects) loci (Table 2).
The D13S137 and D13S127 loci flank
the RB1 gene, while the RB (1.20) repeat sequence is within intron
20 of the same gene. LOH at the D13S127 locus suggests loss of at
least a portion of the RB1 gene as shown in 2/14 informative
subjects. LOH at RB (1.20) indicated loss of the RB1 gene in a
further 1/
10 informative subjects. The RB1 gene was thus lost in 3/18
informativ
e subjects (Table 2). LOH at the D13S120 and D13S127 loci flanking
the BRC
A2- gene was shown in 2/20 informative subjects (Table 2). No LOH
was found for any of the remaining loci (Table 2).
Microsatellite/genomic instability (or a gain/loss of
microsatellite repeats
) was found in 15% (3/20) of subjects.
p53 gene codon 249 analysis
The p53 codon 249 mutation was detected in 25% of the
subjects using PCR-RFLP analysis, and confirmed by sequencing. The
p53 codon 249 mutation was det
ected in the tumor tissue of 3 subjects, in the non-tumorous liver
of 1 subject
, and in both the tumor and non-tumorous liver tissue of 1 subject
(Table 2)
.
Sequencing
gel electrophoresis of the p53 gene product revealed a gel artifact,
in all subjects with wild-type chromosomes, previously described by
Kapelne
r et al(1994).
All
tumors were at an advanced stage. No attempt was made to correlate
the presence of LOH or microsatellite instability with clinical or
other features.
Table 2 LOH, SSCP and sequence analysis
|
Subject
number
|
VNTRs
|
p53
codon 249
|
HBV
status
|
|
p53
|
WT1
|
RB1
|
BRCA2
|
BRCA1
|
E
cadherin
|
|
(ALU)
|
(AC)
|
D13S137
(GT)2
2
|
RB1.20[CTTT(T)]n
|
D13S120
(AC)73
|
D13S12
7(AC)35
|
D17S85
5(AC)
|
D17S846
(GGAA)25
|
D16S301
(AC)
|
D16S260
(AC)
|
T
|
NT
|
|
1
|
NI
|
NI
|
-
|
-
|
-
|
NI
|
-
|
-
|
-
|
NI
|
-/-
|
-/-
|
HBsAg+;
HBeAg-
|
|
2
|
-
|
-
|
-
|
-
|
-
|
NI
|
-
|
-
|
-
|
NI
|
-/-
|
-/-
|
HBsAg+
|
|
3
|
-
|
-
|
-
|
-
|
-
|
-
|
-
|
-
|
NI
|
-
|
+/-
|
-/-
|
HBsAg+;
HBeAg-
|
|
4
|
-
|
-
|
-
|
NI
|
-
|
-
|
?
|
NI
|
-
|
-
|
-/-
|
-/-
|
anti-HBc+;
anti-HBs+
|
|
6
|
-
|
?
|
?
|
?
|
-
|
NI
|
?
|
-
|
?
|
-
|
+/-
|
-/-
|
HBsAg+;
HBeAg-
|
|
7
|
?
|
?
|
?
|
?
|
-
|
-
|
?
|
?
|
| |