Allelic Loss and Gain, But Not Genomic Instability,
as the Major Somatic Mutation in Primary
Gang Wang,1,2Yan Zhao,2Xiaoming Liu,2Lu Wang,3Chungen Wu,2Weiling Zhang,2Wanqing Liu,2
Pingping Zhang,1Wenming Cong,4Yuanrong Zhu,5Lisheng Zhang,6Saijuan Chen,3Dafang Wan,1Xintai Zhao,1
Wei Huang,2*and Jianren Gu1*
1National Laboratory for Oncogene & Related Genes, Shanghai Cancer Institute, Shanghai, China
2Chinese National Human Genome Center at Shanghai, Shanghai, China
3Shanghai Institute of Hematology, Rui Jin Hospital, Shanghai Second Medical University, Shanghai, China
4Eastern Hepatobiliary Surgery Hospital, Shanghai, China
5Qidong Liver Cancer Institute, Qidong, China
6Guangxi Cancer Institute, Guangxi, China
To identify genetic abnormalities in primary hepatocellular carcinoma (HCC), we performed microsatellite analysis (MSA) on
60 Chinese HCC specimens. Utilizing a semi-quantitative MSA and 292 highly polymorphic markers spanning all 22 autosomes,
we found that somatic allelic imbalance (AI) occurred frequently in HCC. To evaluate the nature of the AI, comparative
genomic hybridization was performed on 20 HCC specimens. The combined use of these two methods revealed frequent
allelic loss on 17p, 9p21–p23, 4q, 16q21–q23.3, 13q, 8p21–p23, and 6q24–q27, whereas there was frequent allelic gain on 1q,
17q, and 8q24. The region with the highest incidence of genomic imbalance was 17p13 (65%), followed by 9p21–p23 (55%),
4q (35–51%), 16q21–q23.3 (52%), 17p12 (49%), 13q (39–46%), 8p21–p23 (41–45%), 8q24 (41%), and 1q32 (40%). In addition,
aberrations of 19p13.3, 16p13.3, 13q33–q34, 9q13–31, and 7q were reported for the first time. The presence of a close
correlation of 17p13 deletion with abnormalities of some other loci implies that 17p13 could play a crucial role in oncogenesis.
Interestingly, microsatellite instability was rarely seen in our patients, in contrast to that observed in European HCC samples.
© 2001 Wiley-Liss, Inc.
Hepatocellular carcinoma (HCC) is one of the
most prevalent malignancies in Asia and Africa. It
has been proposed that the development of HCC is
a multi-step process resulting from synergism of
both environmental factors, such as hepatitis B
virus, hepatitis C virus, and aflatoxin, and genetic
factors, although its precise molecular pathogenesis
remains to be elucidated (Cho et al., 1994; Khakoo
et al., 1996; Di Bisceglie, 1997; Montesano et al.,
1997; Akriviadis et al., 1998; Schafer and Sorrell,
Over the last few years, the exploration of genes
associated with HCC has been performed either by
direct examination of known oncogenes and tumor
suppressor genes (TSGs), or by screening of chro-
mosome regions with frequent allelic imbalance
such as loss of heterozygosity (LOH), suggestive of
TSGs or genomic instability detected by whole-
genome scanning technologies including microsat-
ellite analysis (MSA) and comparative genomic hy-
bridization (CGH) (Buetow et al., 1989; Nishida et
al., 1992; Fujimoto et al., 1994; Marchio et al., 1997;
Nagai et al., 1997; Piao et al., 1998). In the present
work, we analyzed a large series of HCC specimens
using genome-wide semi-quantitative MSA and
addressed genomic imbalance or instability in
HCC through the combined use of MSA and CGH.
MATERIALS AND METHODS
Patients and Tissue Specimens
All primary HCCs and noncancerous liver or
white blood cell (WBC) specimens for MSA and
CGH were obtained from sixty patients during
surgery with informed consent. These samples
were from three hospitals, namely, the Shanghai
Gang Wang, Yan Zhao, Xiaoming Liu, and Lu Wang contributed
equally to this work.
Supported by: Chinese High Tech Program; Grant number: 863;
Chinese Key Basic Research Project; Grant number: 973; National
Natural Science Foundation of China; Shanghai Commission for
Science and Technology.
*Correspondence to: Dr. Jianren Gu, National Laboratory for
Oncogene & Related Genes, Shanghai Cancer Institute, 25/2200,
Xietu Road, Shanghai 200032, China, or Dr. Wei Huang, Chinese
National Human Genome Center at Shanghai, Shanghai, China.
Received 9 November 2000; Accepted 1 December 2000
Published online 27 April 2001
GENES, CHROMOSOMES & CANCER 31:221–227 (2001)
© 2001 Wiley-Liss, Inc.
Eastern Hepatobiliary Surgery Hospital, Guangxi
Cancer Institute, and Qidong Liver Cancer Insti-
tute, representing three different regions of South-
ern China with a high incidence of HCC, with
exposure to HBV and aflatoxin as important risk
factors. All of the specimens were proven to be
HCC by histopathologic examination. Of the total
of 60 patients, 51 were male and 9 female. The
ages ranged from 27 to 74 years with an average of
52 years. The characterization of patients and spec-
imens is shown in Table 1.
Genomic DNA was prepared by the SDS-pro-
teinase K and phenol-chloroform extraction meth-
ods (Maniatis et al., 1989).
Polymorphic Microsatellite Loci and Allelotyping
All microsatellite (MS) markers came from an
ABI PRISM linkage mapping set version 2 panel.
because solid tumors are usually composed of a
mixture of neoplastic and non-neoplastic cells, the
number of PCR cycles was adjusted so that the
reaction was stopped before reaching the plateau
level. Amplified products were diluted and loaded
onto a 5% polyacrylamide gel containing 7.5 M
urea in the ABI PRISM™ 377 Sequencer (PE Ap-
plied Biosystems, Foster City, CA). The gel image
was collected and analyzed by utilizing the ABI
software GeneScan (version 3.0) and Genotyper
Assessment of Allelic Imbalance and Microsatellite
The allelic status was determined by comparing
PCR product patterns of paired cancerous and non-
cancerous liver/WBC samples. Heterozygous al-
leles were used to compare the difference of allelic
status and termed as “informative.” Microsatellite
instability (MSI) was defined as the presence of
novel fragments in amplified products of tumor
DNAs compared with reference DNAs.
Comparative Genomic Hybridization
CGH was carried out basically according to
methods previously described by Kallioniemi et al.
(1994). Negative and positive controls were in-
cluded for each experiment and analysis. In the
CGH profile, two levels of imbalances were taken
into account. The thresholds for gains and losses
were 1.25 and 0.75, respectively.
Analysis of the relationship between clinicopath-
ologic characteristics and allele imbalance (AI) at
each locus, and the relationship between AI at
different loci, was performed using Fisher’s exact
test. The threshold of statistical significance was
set at P ? 0.01.
Identification of 17p13 as the Region With Most
The clinical and pathologic information on the
60 Chinese HCC patients investigated in this work
is summarized in Table 1. Of all the 11,416 infor-
mative genotyping pairs carried out on these pa-
tients, 1,882 (16.5%) were defined as having AI
(data not shown). The location of each polymor-
phic locus and the corresponding frequency of AI
are shown in Figure 1. Statistical analysis of the
frequencies of all markers demonstrated that loci
with an AI frequency above 30% were beyond the
Gaussian distribution, and 30% was then defined as
the threshold for differentiating between nonran-
dom variations and random ones. Loci with non-
random aberration were concentrated in 13 chro-
TABLE 1. Characteristics of Patients and Specimens
Clinicopathologic characteristics Number
?3 cm, ?5 cm
Tumor differentiation gradea
Serum AFP levelb
Capsule of tumor
aGrade of differentiation was performed according to the method of
Edmondson and Steiner.
cPortal vein invasion of tumor cells, or intrahepatic metastasis, was
observed under the microscope.
WANG ET AL.
was defined according to the comprehensive ISCN. The boxed regions exhibited the most frequent
aberrant chromosomal regions in 60 primary HCC specimens.
Ideograms showing the location and frequency of imbalance of each locus used. The location
mosomal regions, that is, 17p13 (37/57, 65%),
followed by 9p21–p23 (55%), 16q21–23.3 (52%), 4q
(40–51%), 17p12 (49%), 13q12.1–q12.3 (46%), and
8p21–p23 (41–45%). Moreover, a number of re-
gions not described in previous studies were shown
in this work to be associated with a high rate of AI,
including 17p12 (49%), 13q33–q34 (39%), 19p13.3
(40%), 16p13(40%), 21q21–q22
Correlation Between AI on 17p13 and That on
Analysis of the association between AI in differ-
ent regions revealed a statistically significant corre-
lation between abnormality on 17p13 and that on
9p21-23, 4q26/q35, 16q21-q23.3, 13q, 1q32, and
16p13.3 (Tables 2 and 3). As shown in Table 3,
among 29 samples confirmed to be heterozygous
for both sites, D17S831 (17p13) and D9S288
(9p21–23) shared the common AI in 10, whereas no
AI was observed in 13 (P ? 0.001). In addition,
9p21–p23 exhibited a similar correlation with loci
at 4q26, 16q21–23.3, 13q14, and 16p13.3, although
the P values (0.004–0.009) were lower than those
between 17p13 correlated with other loci. A similar
relationship was found between 4q35 and other
loci. These correlations may therefore reveal an
HCC, suggesting that structural/functional alter-
ation of some genes in these loci might be associ-
ated with a pathway or network underlying hepa-
Because tumor size and some other parameters
(serum level of ?-fetoprotein, metastasis) are con-
sidered an indication of disease stages, it was im-
portant to look for a possible association between
the consequence of these accumulated events and
the clinicopathologic status of HCC patients. It is
of interest that both 17p13 and 9p21–23 were
present as regions with the highest AI among pa-
tients with small HCCs (?3 cm, 6/7). Neverthe-
less, no statistically significant correlation was
noted between loci with AI and disease stage,
probably because of a biased distribution of HCC
samples with different tumor sizes (small, 7; inter-
mediate, 17; and large, 36).
Comparison of MSA/CGH Results in 20 HCC
Two mechanisms of genomic imbalance could
cause AI, allelic loss (LOH), or allelic gain/ampli-
fication, and they cannot be distinguished using
MSA alone. To determine the nature of the de-
tected changes in allele dosage, we performed
CGH on 20 of the 60 HCC cases. Abnormal CGH
profiles were observed in all tumor DNAs exam-
ined. A summary of losses and gains of DNA se-
quences is given in Figure 2. As shown in Figures
1 and 2, frequent loss of DNA copies was mapped
to 17p (12/20), 9p21–pter (7/20), 16q22–q24 (7/20),
4q31.1–qter (6/20), 4q22–q28 (6/20), 8p21–23 (8/
20), 6q22–qter (6/20), 13q14–q22 (5/20), and
9q13–q31 (5/20). In contrast, the most frequent
gains revealed by MSA occurred on 1q(15/20),
8q(13/20), 6p (10/20), 3q11–q24 (8/20), 17q (7/20),
and 3q25–qter (6/20); high-level amplifications de-
fined by CGH were identified in 8q (7/20),
1q21–25 (5/20), 6p (3/20), 13q31–34 (2/20), and
7q21–32 (2/20). Comparison of results obtained
TABLE 2. Observed Correlation of Different
aP-values calculated from Exact Test with the threshold of significance
taken as P ? 0.01. The analysis of correlation was performed between
every two regions exhibiting frequent aberrations. Here, only significant
results are summarized.
TABLE 3. Examples of Correlated Abnormalities on
Different Loci in 60 HCC Specimens*
*ROH, retention of heterozygosity; LOH, loss of heterozygosity.
WANG ET AL.
with MSA and CGH showed an overall good con-
sistency (72% at the level of chromosome subre-
gions or arms, data not shown). Most importantly,
the combined utilization of MSA/CGH allowed us
to determine that, in primary HCC, frequent AI on
17p13, 9p21–22, 4q, 13q12–22, 16q21–23.3, and
8p21–p23 was actually caused by LOH, whereas AI
on 1q and 8q was caused by gain of DNA copies.
Microsatellite Instability Is Rare in HCC
Specimens From Southern China
MSA can reveal both AI and MSI. The presence
of novel fragments in amplified products of cancer
DNAs compared with those of reference DNAs is
a strong indicator of genomic instability. Frequent
MSI has been reported in colorectal carcinoma and
some other solid tumors (Ghimenti et al., 1999).
There were controversies in previous reports with
regard to the incidence of MSI in HCC (Salvucci et
al., 1999; Sheu et al., 1999). In this work, a scrutiny
of PCR fragment patterns in 11,416 tumor/refer-
ence sample genotypes revealed only 22 pairs
(0.19%) with MSI on 20 loci, and the distribution of
these MSIs was almost random (data not shown).
In this study, we describe a relative high-resolu-
tion global view of allelotyping in 60 primary HCC
specimens from Southern China with 292 highly
polymorphic microsatellite markers. Moreover,
combined MSA and CGH were applied for the first
time on 20 HCC specimens and yielded results
that were both consistent and complementary. As a
result, abnormal regions with a frequency above
40% in the examined cases can be separated into
two groups: LOH regions that were mapped to
sequences identified by CGH in 20 patients with
HCC. Gains are shown on the right side of the
chromosome ideograms and losses on the left.
High-level amplifications are shown as thick lines.
Each vertical line represents the affected chromo-
somal region seen in a single tumor specimen.
Summary of losses and gains of DNA
ALLELIC LOSS AND GAIN AS MAJOR SOMATIC MUTATION
17p13/p12, 9p21–p23, 4q, 13q12–22, 16q21–23.3,
and 8p21–p23, and gain/amplification regions on
1q32, 8q24, and 6p. Interestingly, all of these re-
gions except 9p21–p23 were found to be affected
in two recent reports on comprehensive allelotyp-
ing of HCC, one from Korea and the other from
France (Marchio et al., 1997; Piao et al., 1998). The
fact that lesions on chromosomal regions 17p13, 4q,
8p21–23, 13q, 8q24, and 16q22–24 were identified
in our report and in two previous reports, indepen-
dent of the marker panel used, the analytic tech-
niques performed, and the populations examined,
suggests that a common pathway or network may
exist in the molecular pathogenesis of HCC. Allelic
loss of 17p13, 9p21–23, 8p, 13q, and 16q also occurs
in other types of human cancers, indicating that
these regions harbor tumor suppressor loci that are
important in regulating cell proliferation among
different cell types. In this work, 17p12, 19p13.3,
13q33–q34, 16p13.2–13.3, and 2q32.1–q35 are re-
ported for the first time to be frequently implicated
in HCC, possibly due to a more sensitive MSA
applied or as a reflection of different etiologic fac-
tors compared with those of HCC patients from
other geographic areas.
Among all chromosome regions with AI in our
HCC patient series, 17p13 was shown to be the
most highly affected. Interestingly, allelic loss on
chromosome arm 17p was also among the most
common genetic abnormalities in many other hu-
man cancers, and deletion of 17p13 was reported in
various human cancers including colorectal, gastric,
ovarian, breast, and pancreatic carcinoma (Cun-
ningham et al., 1992; Foulkes et al., 1993; Nieder-
acher et al., 1997; Yustein et al., 1999). The TP53
gene on 17p13.1 was thought to be the gene asso-
ciated with the genesis of these cancers. Neverthe-
less, the frequency of mutation in the TP53 gene
varied in different tumors with 17p involvement,
and evidence was obtained to suggest the presence
of other putative tumor suppressor genes distal to
the TP53 gene (White et al., 1996; Chattopadhyay
et al., 1997; Konishi et al., 1998; Liscia et al., 1999).
Indeed, in our study, all three markers on 17p13
exhibited a high frequency of LOH. In particular,
cases 15, 24, and 31 showed allelic loss at the locus
D17S849 (17p13.3), but retained heterozygosity at
D17S938 (17p13.1). The importance of 17p13 in
hepatocarcinogenesis was also reflected by its pres-
ence in 6/7 cases with small HCC. Most interest-
ingly, we found a significant correlation between
LOH at 17p13.3 and that at 9p21–23, 4q26, 4q35,
and 13q14 (Table 2). This correlation raised several
important issues for hepatocarcinogenesis. We
could ask whether certain chromosome loci could
be attacked simultaneously by some specific carci-
nogenic factors responsible for HCC development.
Alternatively, we could ask whether the alterations
of different chromosomes might be sequential
events during the development and progression of
HCC. These mechanisms await further investiga-
Over the last few years, besides LOH and gain/
amplification, attention has also been paid to MSI,
an indication of genomic instability in human can-
cers such as colorectal carcinoma, which implies a
defect in genes involved in the DNA mismatch
repair system. Previous studies on MSI in HCC
resulted in quite different data. A group from Tai-
wan reported that MSI might play a small role in
HCC, but others stated that MSI was among com-
mon types of mutation in European HCC samples
or liver specimens with hepatitis B viral cirrhosis
(Salvucci et al., 1996, 1999; Sheu et al., 1999). The
high frequency of MSI, in addition, has been asso-
ciated with two DNA mismatch repair genes,
hMSH1 and hMSH2, located, respectively, on chro-
mosome 2 and on 3p21.3–p23 in one report (Mac-
donald et al., 1998). In this study, we observed
random MSI of very low frequency (0.19% in
11,416 genotype pairs) in HCC from Southern
China. No obvious AI was noted in the chromo-
some regions harboring hMSH1 and hMSH2.
Therefore, our data strongly support the idea that,
in contrast to the finding in HCC from Europe,
MSI is not a major pattern of somatic mutation, and
therefore defective DNA mismatch repair seems
unlikely to play an essential role, in HCC in China.
Understanding the mechanisms responsible for this
difference should provide new insights into the
further exploration of the etiologic factors and mo-
lecular pathogenesis of HCC in China.
Gratitude is extended to Prof. Zhu Chen for
reviewing this manuscript and for his constructive
Akriviadis EA, Llovet JM, Efremidis SC, Shouval D, Canelo R,
Ringe B, Meyers WC. 1998. Hepatocellular carcinoma. Review.
Br J Surg 85:1319–1331.
Buetow KH, Murray JC, Israel JL, London WT, Smith M, Kew M,
Blanquet V, Brechot C, Redeker A, Govindarajah S. 1989. Loss of
heterozygosity suggests tumor suppressor gene responsible for
primary hepatocellular carcinoma. Proc Natl Acad Sci USA 86:
Chattopadhyay P, Rathore A, Mathur M, Sarkar C, Mahapatra AK,
Sinha S. 1997. Loss of heterozygosity of a locus on 17p13.3,
independent of p53, is associated with higher grades of astrocytic
tumours. Oncogene 15:871–874.
WANG ET AL.
Cho Y, Gorina S, Jeffrey PD, Pavletich NP. 1994. Crystal structure
of a p53 tumor suppressor-DNA complex: understanding tumor-
igenic mutations. Science 265:346–355.
Cunningham J, Lust JA, Schaid DJ, Bren GD, Carpenter HA, Rizza
E, Kovach JS, Thibodeau SN. 1992. Expression of p53 and 17p
allelic loss in colorectal carcinoma. Cancer Res 52:1974–1980.
Di Bisceglie AM. 1997. Hepatitis C and hepatocellular carcinoma.
Review. Hepatology 26:34S–38S.
Foulkes WD, Black DM, Stamp GW, Solomon E, Trowsdale J.
1993. Very frequent loss of heterozygosity throughout chromo-
some 17 in sporadic ovarian carcinoma. Int J Cancer 54:220–225.
Fujimoto Y, Hampton LL, Wirth PJ, Wang NJ, Xie JP, Thorgeirsson
SS. 1994. Alterations of tumor suppressor genes and allelic losses
in human hepatocellular carcinomas in China. Cancer Res 54:281–
Ghimenti C, Tannergard P, Wahlberg S, Liu T, Giulianotti PG,
Mosca F, Fornaciari G, Bevilacqua G, Lindblom A, Caligo MA.
1999. Microsatellite instability and mismatch repair gene inacti-
vation in sporadic pancreatic and colon tumours. Br J Cancer
Kallioniemi OP, Kallioniemi A, Piper J, Isola J, Waldman FM, Gray
JW, Pinkel D. 1994. Optimizing comparative genomic hybridiza-
tion for analysis of DNA sequence copy number changes in solid
tumors. Review. Genes Chromosomes Cancer 10:231–243.
Khakoo SI, Grellier LF, Soni PN, Bhattacharya S, Dusheiko GM.
1996. Etiology, screening, and treatment of hepatocellular carci-
noma. Review. Med Clin North Am 80:1121–1145.
Konishi H, Takahashi T, Kozaki K, Yatabe Y, Mitsudomi T, Fujii Y,
Sugiura T, Matsuda H, Takahashi T, Takahashi T. 1998. De-
tailed deletion mapping suggests the involvement of a tumor
suppressor gene at 17p13.3, distal to p53, in the pathogenesis of
lung cancers. Oncogene 17:2095–2100.
Liscia DS, Morizio R, Venesio T, Palenzona C, Donadio M, Calla-
han R. 1999. Prognostic significance of loss of heterozygosity at
loci on chromosome 17p13.3–ter in sporadic breast cancer is evi-
dence for a putative tumour suppressor gene. Br J Cancer 80:821–
Macdonald GA, Greenson JK, Saito K, Cherian SP, Appelman HD,
Boland CR. 1998. Microsatellite instability and loss of heterozy-
gosity at DNA mismatch repair gene loci occurs during hepatic
carcinogenesis. Hepatology 28:90–97.
Maniatis T, Fritsch J, Sambrook J. 1989. Molecular cloning: a labo-
ratory manual (2nd edition). Cold Spring Harbor: Cold Spring
Harbor Laboratory Press. p 9.37–9.57.
Marchio A, Meddeb M, Pineau P, Danglot G, Tiollais P, Bernheim
A, Dejean A. 1997. Recurrent chromosomal abnormalities in hep-
atocellular carcinoma detected by comparative genomic hybrid-
ization. Genes Chromosomes Cancer 18:59–65.
Montesano R, Hainaut P, Wild CP. 1997. Hepatocellular carcinoma:
from gene to public health. Review. J Natl Cancer Inst 89:1844–
Nagai H, Pineau P, Tiollais P, Buendia MA, Dejean A. 1997.
Comprehensive allelotyping of human hepatocellular carcinoma.
Niederacher D, Picard F, van Roeyen C, An HX, Bender HG,
Beckmann MW. 1997. Patterns of allelic loss on chromosome 17
in sporadic breast carcinomas detected by fluorescent-labeled
microsatellite analysis. Genes Chromosomes Cancer 18:181–192.
Nishida N, Fukuda Y, Kokuryu H, Sadamoto T, Isowa G, Honda K,
Yamaoka Y, Ikenaga M, Imura H, Ishizaki K. 1992. Accumulation
of allelic loss on arms of chromosomes 13q, 16q and 17p in the
advanced stages of human hepatocellular carcinoma. Int J Cancer
Piao Z, Park C, Park JH, Kim H. 1998. Allelotype analysis of
hepatocellular carcinoma. Int J Cancer 75:29–33.
Salvucci M, Lemoine A, Azoulay D, Sebagh M, Bismuth H, Reyns
M, May E, Debuire B. 1996. Frequent microsatellite instability in
post hepatitis B viral cirrhosis. Oncogene 13:2681–2685.
Salvucci M, Lemoine A, Saffroy R, Azoulay D, Lepere B, Gaillard S,
Bismuth H, Reynes M, Debuire B. 1999. Microsatellite instability
in European hepatocellular carcinoma. Oncogene 18:181–187.
Schafer DF, Sorrell MF. 1999. Hepatocellular carcinoma. Review.
Sheu JC, Lin YW, Chou HC, Huang GT, Lee HS, Lin YH, Huang
SY, Chen CH, Wang JT, Lee PH, Lin JT, Lu FJ, Chen DS. 1999.
Loss of heterozygosity and microsatellite instability in hepatocel-
lular carcinoma in Taiwan. Br J Cancer 80:468–476.
Simonetti RG, Camma C, Fiorello F, Politi F, D’Amico G, Pagliaro
L. 1991. Hepatocellular carcinoma. A worldwide problem and the
major risk factors. Review. Dig Dis Sci 36:962–972.
White GR, Stack M, Santibanez-Koref M, Liscia DS, Venesio T,
Wang JC, Helms C, Donis-Keller H, Betticher DC, Altermatt HJ,
Hoban PR, Heighway J. 1996. High levels of loss at the 17p
telomere suggest the close proximity of a tumour suppressor. Br J
Yustein AS, Harper JC, Petroni GR, Cummings OW, Moskaluk CA,
Powell SM. 1999. Allelotype of gastric adenocarcinoma. Cancer
ALLELIC LOSS AND GAIN AS MAJOR SOMATIC MUTATION