| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Biology and Genetics |
Institute of Biomedical Sciences, Academia Sinica, Taipei 11529, Taiwan [C. Y. S., Y. L. L., C. H. K., J. C. L., C. W. W.]; Department of Surgery, Tri-Service General Hospital, Taipei 100, Taiwan [J. C. Y.]; Department of Pathology, Shin Kong Wu Ho-Su Memorial Hospital, Taipei 111, Taiwan [C. T. Y.]; National Health Research Institutes, Taipei 11529, Taiwan [Y. S. J., C. W. W.]; and Department of Surgery, National Taiwan University Hospital, Taipei 100, Taiwan [C. S. H.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The greatest insights in our understanding of molecular tumorigenesis
have been obtained from the study of family cancer syndromes
(3–5)
. In the case of breast cancer, an inherited
component has been suspected for many years because of reports of
families with large numbers of affected individuals. In the Li-Fraumeni
syndrome, there is a high incidence of premenopausal breast cancer. As
a result of the identification of the germline mutations in 
50% of
pedigrees, these syndromes were linked to the p53 tumor
suppressor gene (6)
. The documentation of >1000 mutations
of the p53, accounting for about 25–30% of all breast
cancers tested, has clearly established its role in breast cancer
development (7)
. The inheritance of breast cancer
susceptibility in families has also led to the localization of breast
cancer susceptibility genes, including BRCA1 localized to
chromosome arm 17q and BRCA2 to chromosome arm 13q (reviewed
in Ref. 8
). Furthermore, the mapping of homozygous
deletions on human chromosome 10q23 has resulted in the isolation of a
candidate
TSG,2
PTEN/MMAC1, that appears to show a considerable
frequency of mutation in breast cancer and to be responsible for Cowden
disease, an inherited breast and thyroid cancer syndrome (reviewed in
Ref. 9
). In addition to these highly penetrant genes,
other genes, in which mutations might result in a more moderate
increase in breast cancer risk, have been predicted; a well-known
example is the ataxia telangiectasis gene. It has been suggested
that heterozygote carriers of defective forms of the gene predisposing
to ataxia telangiectasia are at higher risk of developing breast cancer
(reviewed in Refs. 10
and 11
). The recent
isolation of this gene, designated ATM, supports this
hypothesis.
To seek clues to the initiation of breast cancer development, we explored the suggested function of these genes linked to inherited breast cancer syndromes. It is intriguing to find that, except PTEN/MMAC1, the products of the other four genes (p53, ATM, BRCA1, and BRCA2) are all directly involved in a common molecular pathway related to cellular responses against DSBs arisen in DNA (reviewed in Refs. 12 and 13 ). ATM is a component of the cell cycle checkpoint machinery that senses DNA damage and is activated after the formation of DSBs induced by ionizing radiation or other DSB-inducing agents (reviewed in Ref. 11 ). p53, acting downstream of ATM, is then phosphorylated and consequently triggered to initiate a protective response either by blocking the cell cycle for DNA repair or by inducing apoptosis in genetically damaged cells (reviewed in Refs. 14 and 15 ). Alternatively, the proteins encoded by BRCA1 or BRCA2 can work in concert with a protein called Rad51, which is known to help repair DSBs in the DNA double helix (reviewed in Refs. 16 and 17 ). Given the central role of DSB-induced chromosome breakage to cause genomic deletion or nonhomologous recombination (18, 19) , we postulated that breast cancer etiology is specifically linked to DSB-initiated CIN, a defining characteristic of mutator phenotype of cancer cells (20) . Because genomic deletion or nonhomologous recombination are reflected in a genetic mechanism called LOH (21) , this genome-wide LOH study was conducted to test our hypothesis. We sought to determine whether breast cancer progression is associated with an increased level of genomic instability characterized by a specific mutator phenotype manifested as DSB-initiated CIN.
LOH is common to all human solid tumors and allows the expressivity of recessive loss-of-function mutations in TSGs (21) . Therefore, the detection of recurrent LOH in a chromosomal region is now considered to be critical evidence for the localization of TSGs. However, two methodological limitations always affect valid assessment of LOH results. Because detection of LOH is based on the comparison of tumor cells and corresponding normal tissues to identify genetic deletion that occurs specifically in tumor tissues, it is essential to be able to obtain pure tumor cells to provide the homogeneous materials required for reliable analysis. However, the infiltrating nature of most invasive solid tumors, including breast cancer, leads to contamination of the surrounding normal tissue and creates technical problems in LOH analysis. To overcome this problem, the labor-intensive, inaccurate, and time-consuming process of tissue microdissection must be used, which originally involved manual or micromanipulator-guided scraping of the area of interest of a thin tissue section. The recently developed technique of LCM has greatly facilitated this procedure and now provides an efficient and reliable one-step method for obtaining pure populations of cells from specific microscopic regions of tissue sections under direct visualization (22) . LCM was used in the present study to obtain pure tumor cells. Secondly, because the magnitude of the genomic instability manifested as LOH varies among different forms of tumors and between the same tumor in different study populations, it would not be precise to define the loci of frequent LOH simply on the basis of the magnitude of an increased frequency used in previous studies. To overcome this problem, the present study used a genome-wide search for LOH and identification of loci showing a high LOH frequency based on comprehensive alteration profiles. On the basis of these improved methods, this study aimed at dissecting out the contribution of putative genomic loci involved in breast cancer progression and elucidating the genetic mechanism underlying breast tumorigenesis.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
LCM.
To ensure that tissue samples assayed consisted of >90% tumor cells,
LCM was performed on routinely immunostained slides using a PixCell
laser capture microscope (Arcturus Engineering, Mountain View, CA) as
described previously (22)
with minor modification.
Briefly, the stained, dehydrated tissue section was overlaid with a
thermoplastic film mounted on an optically transparent cap. The
visually selected areas (tumor cells) were bound to the membrane by
short, low-energy laser pulses, resulting in focal melting of the
polymer. On an average, the LCM shot performed to obtain tumor cells
for this genome-wide LOH for one single tumor was 2,000, which
yielded 20,000 tumor cells. The cells were immersed in 50–100 µl of
digestion buffer, containing 10 mM Tris-HCl (pH 8.0), 1
mM EDTA, 400 µg/ml proteinase K, and 1% Tween 20, and
digested at 55°C overnight. After digestion, the enzyme was heat
inactivated (95°C for 10 min), and the extract was used directly for
allelotyping PCR (LOH detection). Efforts were made to optimize LCM
conditions, including improving capture efficiency by overnight
pretreatment of the slide with xylene and ensuring successful PCR
amplification by avoiding fragmentation of DNA in the captured tissue
by the use of a modified protocol for tissue fixation. Genomic DNA
extracted from the corresponding WBC samples using conventional
proteinase K-phenol/chloroform extraction served as the normal control
for LOH analysis.
Allelotyping PCR.
DNA from the microdissected tumor specimens and blood samples was used
for allelotyping PCR using fluorescence-labeled primers from Applied
Biosystems PRISM Linkage Mapping Set-MD10 (PE Biosystems, Foster City,
CA), comprising 400 polymorphic microsatellite markers, with an average
spacing of 9.2 cM and average heterozygosity of 0.79. The markers were
selected from the Généthon linkage map (29)
,
on the basis of chromosomal location and heterozygosity. These markers
were arranged in 28 panels, facilitating simultaneous amplification of
multiple primer pairs (about four markers/per PCR assay) and
significantly reducing the efforts required to perform the arduous task
of genome-wide screening. PCR amplification was carried out using DNA
(of tumor and blood), 0.4 unit of Taq polymerase, 0.2 mM
deoxynucleotides, and 2.5 mM MgCl2 in
a total reaction volume of 10 µl. The PCR conditions used were 95°C
for 12 min to activate the Taq polymerase, followed by 45 cycles of
denaturation of 94°C for 15 s, annealing at 55°C for 15 s, and extension at 72°C for 30 s. The final elongation was at
72°C for 10 min. PCRs were run in a GeneAmp PCR 9600 thermocycler (PE
Biosystems, Foster City, CA). The reaction conditions sometimes yielded
unsatisfactory amplification for certain markers and were modified, and
if necessary, these particular markers were run singly. PCR products
were electrophoresed on a 377 ABI PRISM sequencer, and the fluorescent
signals from the different sized alleles were recorded and analyzed
using GENOTYPER version 2.1 and GENESCAN version 3.1 software.
Definition of Allelic Loss (LOH).
For a given informative marker, we used our previous definition
(24, 30) that the marker is considered to display LOH when
a 3-fold or greater difference is seen in the relative allele intensity
ratio between the tumor DNA and normal DNA. However, because we used
LCM throughout the study, almost all LOH markers showed a >5-fold
difference in intensity, and no ambiguous cases were found.
Data Analysis.
The statistical methods used in previous genome-wide screening studies
(CGH studies), which involved examination of genomic changes at the
chromosome level, were not applicable to our study. More importantly,
the characteristics of the 400 markers used in the present study,
including a very intense distribution throughout the whole genome and
regular spacing among markers, provided a unique opportunity to
precisely measure genomic instability at the level of individual
genomic loci (markers). This allowed us to develop a novel index to
estimate the level of genomic abnormality of individual tumors. The
index of the PLOH was a measure of the proportion (prevalence) of
acquired allelic loss (LOH) in a breast cancer tumor or at specific
stages/grades of a breast tumor and was obtained by dividing the total
number of markers showing LOH in a tumor by the number of informative
(heterozygous) markers.
To obtain evidence for the role of genomic instability in breast cancer progression, the average PLOHs of tumors of different grades or stages were correlated to tumor grade or tumor stage. The difference in the average PLOH between tumors with different pathological indexes was examined using Student t test. In addition, linear regression was used to evaluate the trend of an increasingly higher average PLOH as tumor progressed to poorer grade or later stage.
It was of particular interest to determine whether tumor progression was specifically associated with LOH at particular genomic loci. Given the fact that most cancers are clonal (31, 32) , implying breast cancer development will generally progress from low grade/early stage to high grade/later stage (reviewed in Refs. 33 and 34 ), we applied a well-known strategy to establish a model for breast tumorigenesis. The strategy used, which has been successfully applied in defining the genetic events underlying the progression of colon cancer (35) , is that a prevalent genetic alteration that is consistently seen, regardless of histological stage or grade, is considered an early event in the tumorigenesis, and an alteration that is associated with a histologically and clinically advanced stage or grade is interpreted as a late event. However, in contrast, it has also been suggested that, although clonal, cancers are highly heterogeneous (31, 36) . More importantly, a recent study argued against the concept of sequential progression and proposed an alternative model, i.e., different breast tumor grades may have distinct molecular origins (37) . Thus, to account for the complexity of these issues, including the possibilities that: (a) either genomic instability per se or allelic loss at specific loci are the driving forces in tumor progression; and (b) molecular heterogeneity of individual cancer cells might be an inevitable consequence of genetic instability (31) , we used two different approaches to select putative markers for the sites of TSGs from the 400 markers used in the evaluation. The first approach was to select 20 markers (the top 5%), the LOH of which showed the highest degree of association with an increase of genomic instability, from the top 50 markers most frequently targeted by LOH. This was determined from the difference in average PLOH in tumors with or without LOH at these particular markers (evaluated by t test). The identification of markers significantly associated with an increase in PLOH may provide clues about the sites of putative TSGs implicated in genomic instability leading to tumor formation. The second approach selected markers showing a significantly high frequency of LOH in tumors of different pathological grades, and these markers were identified if their LOH frequencies were among the top 5% of the 400 markers in tumors of specific grades. Recurrent loss at particular genomic loci detected by markers showing high LOH frequency is always considered to suggest the sites of putative TSGs (21) , but genetic changes that are very prevalent within a particular subgroup may be missed or their significance diluted when all breast tumors are analyzed together (37) . Thus, this approach to select grade-specific markers would avoid this problem, and the loss of the sites identified might either confer cells with selective growth advantage during clonal evolution or lead to genetic divergence or heterogeneity characterizing specific grades of tumor. We did not select markers based on tumor staging, because tumor stage is often a combined clinical and pathological process and is not necessarily related to tumor cell differentiation (27, 28) .
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Particular attention was paid to the definition of marker
heterozygosity. Because of the nature of short tandem repeat
(microsatellite) markers on electrophoresis, it is usual to see, in
addition to the main bands, minor "shadow bands," which are thought
to result from slipped stand mispairing at tandem dinucleotide repeats
(38)
. Sometimes, the band of the shorter allele can be
especially intense because it contains both the main band of this
allele and the major shadow band derived from the longer allele. If the
size difference of the two alleles is small, this artifact might become
obvious and might confound the definition of heterozygosity status.
Because all of our study subjects were Taiwanese (Chinese), a
relatively genetically homogenous population as compared with the
Caucasian population (on which the Généthon microsatellite
markers are based), a relatively higher proportion of individuals
possess parental alleles with very similar allelic sizes. Accordingly,
this artifact caused by "shadow band confounding" would be expected
to be more common. To avoid this problem, in addition to using the
usual definition of heterozygosity based on a 
2-bp difference in
allelic length, we also adopted a more conservative criterion that only
a 4-bp difference in allelic length between two alleles would be
considered as heterozygous. On thorough comparison of the results using
either of the two above criteria, only minor differences were found.
Therefore, the study results were presented based on the routinely used
criterion, i.e., a 2-bp-difference between two alleles was
considered to represent heterozygosity (informative), and the
heterozygous markers served as the target to assess LOH status.
|
25% LOH, the highest
values being seen with D17S921 on 17p12 (48.4%),
D16S503 on 16q22.1 (42.0%), D17S831 on 17p13.3
(41.6%), D17S938 on 17p13.2 (37.7%), D11S4175
on 11q22.1 (34.4%), and D1S249 on 1q32.1 (33.9%). Some of
these loci are known to harbor TSGs; these include p53
(17p12–13), E-cadherin (16q22.1), and ATM
(11q22–23), all of which have been suggested to play a critical role
in breast tumorigenesis (reviewed in Refs. 34
and
39
; reported in Ref. 40
). In contrast, 55
markers showed <10% LOH among the informative cases. The frequencies of LOH on individual chromosomal arms can be also estimated by dividing the number of markers showing LOH on a given chromosomal arm by the number of total informative markers on that arm. The highest frequency of LOH for all arms was seen on chromosome 17p (35%), which was mainly attributable to LOH at the locus of p53. The arms with the highest frequency of LOH (>20%) were 1q, 4p, 8p, 8q, 11q, 13q, 16q, 17p, 17q, and 22q. This chromosomal arm deletion profile is similar to that obtained using CGH reported for breast carcinoma in Western populations (reviewed in Ref. 41 ), implying that similar tumorigenic mechanisms might underline the causes of breast cancer in Taiwan, although it is unique in that its incidence is the lowest in the world. Nevertheless, one notable exception is that the loss of 4p, which has rarely been reported in previous genome-wide searches in breast cancer and may suggest the site of ethnic-specific cancer susceptibility genes. It is also possible that the relatively low resolution of the method of CGH, used previously to detect a small deletion at 4p, might be the explanation.
To estimate the magnitude of genomic instability in individual tumors, the PLOH was calculated on the basis of comprehensive deletion profiles of 400 markers distributed throughout the entire genome. The values for the PLOH ranged from 2 to 66.3%, with a mean of 17.6% (SD, 13.0%) and the median of 14.6%. The reason for the higher mean value was attributable to extremely high frequencies of PLOH (>40%) in some tumors. After careful evaluation of their deletion profiles, these tumors were found to show evidence of aneuploidy, in which an extensive loss of all informative markers on particular chromosomal arms or even of the whole chromosome was found. We therefore purposely differentiated the six cases with putative aneuploidy from the other tumors, with LOH profiles consistent with the pattern of interstitial deletion. Although aneuploidy and interstitial deletion are both classified as CIN (20) , the mechanisms underlying these two genetic events are obviously different. Aneuploidy in cancer results from a defective mitotic machinery resulting in chromosome missegregation (reviewed in Refs. 20 and 42 ), whereas interstitial deletion is caused by the formation of DSBs in the DNA (reviewed in Refs. 18–20 ), the latter being the main target of the present study.
Genomic Instability during Breast Cancer Progression.
To test our hypothesis that breast cancer progression is associated
with DSB-initiated CIN and manifests as a mutator phenotype, we
compared the difference in average PLOH among tumors with different
pathological and clinical outcomes. In the interstitial deletion tumors
analyzed, a higher average PLOH was consistently found to be associated
with poorer tumor differentiation, as reflected by a higher tumor grade
or higher individual grade indices (Fig. 2
). An increase in average PLOH was more obvious in tumors displaying
very few (<75%) tubule formation and substantial nuclear pleomorphism
or in tumor cells with higher-than-moderate mitotic counts
(P < 0.05 as compared with tumors displaying
>75% of tubular architecture, little nuclear pleomorphism, or very
few mitotic counts, respectively). Overall, a significantly increasing
trend toward DSB-initiated CIN was seen as IDC progressed to poorer
grades or latter stages (P < 0.05 assessed
by linear regression analysis). In contrast, lobular carcinomas
displayed a genetically stable phenotype, in which the average PLOH was
close to that for well-differentiated IDC (Fig. 2)
.
|
25% and >50% of
heterozygosity), and more importantly, their LOH status showed the most
significant association with an increased level of PLOH, correlating to
an increase of >15% of DSB-initiated CIN (Fig. 3
). LOH in these markers was expected to be either the causes or the
consequences of the increasing level of genomic instability during
tumor progression found in previous section. Thus, it is notable that,
among these four loci, 15q15.1, 2q34, 13q32.3, and 3p24.3, reflected by
LOH at four markers (D15S1007, D2S325, D13S158, and
D3S1266), are the genomic sites in which RAD51, Ku80,
DNA ligase IV, and FA-D are located. These genes are
key players in DSB-related repair or CIN (20, 43)
.
|
). Using 10 microsatellite markers, we examined LOH on nine chromosomal
arms (6q, 8p, 13q, 16q, 17p, 17q, 19p, 20q, and 22q) in DNA
microdissected from both the DCIS and IDC components. At least three of
these tumors (case nos. 467, 541, and 557) supported the clonal
evolution hypothesis, simultaneously showing LOH at the same markers.
However, two tumors (case nos. 467 and 557) revealed a specific LOH
profile, in which LOH at a given locus was only detected in the DCIS
component, but this marker remained intact in the IDC component,
supporting the presence of genetic heterogeneity (Fig. 4)
. These
results suggest that genetic heterogeneity was common in these breast
cancers, and that heterogeneity of breast cancer might occur at the
earliest detectable grade/stage, i.e., DCIS, of progression.
|
). Notably, several of the genomic loci identified, including 17p12–13,
11q22.1, and 16q22, were commonly found in both lower grade tumors and
higher grade tumors, suggesting that the genes located in these loci,
e.g., p53 at 17p12–13, ATM at
11q22.1, E-cadherin at 16q22 (reviewed in Refs.
34
and 39
), might contribute to an early
event during progression. DCIS was not included in this table, because
the limited number of cases of DCIS included in this study might result
in an imprecise estimation of LOH frequency.
|
). The changes in patterns of A1, A2, and A3 are consistent
with the conventional concept of clonal evolution, in which tumor
progression occurs via a sequential pathway driven by genetic
alterations occurring at early (well differentiated; A1), intermediate
(moderately differentiated; A2), and late (poorly differentiated; A3)
stages. In contrast, if LOH of a marker was preferentially seen at
specific grades/stages, putative genes located around this marker may
be involved in grade/stage-specific heterogeneity. Thus, the patterns
of A4, A5, and A3 were considered to represent such genetic alterations
leading to heterogeneity of well-differentiated, moderately
differentiated, and poorly differentiated tumors. Finally, if the
contribution of genomic instability is also taken into consideration,
in which DSB-initiated CIN would become more obvious as the tumor
progresses, the patterns of A1 to A5 would become B1 to B5,
respectively. The timing of LOH of individual markers belonging to
these 10 patterns can be represented in a breast tumorigenic model, as
shown in Fig. 5A
.
|
). Interestingly, the
loci for ATM (11q22) and p53 (17p12–13), the key
checkpoint genes of DSB, were lost at the earliest stage. In addition,
as mentioned earlier, several loci identified (specified in Fig. 5B
) were those possibly harboring genes participating in
DSB-related checkpoint/repair. More interestingly, these loci were
found to be lost at different grades, yielding a clue to suggest that
accumulation of multiple genetic defects associated with DSB-initiated
CIN may be required for the continuous decay of chromosomal integrity.
To test this possibility, we performed a multivariate regression
analysis to correlate the extent of CIN in a tumor (measured by PLOH)
with the LOH status of eight loci that probably harbor CIN-related
genes (including 17q12–13 for p53, 11q22 for
ATM, 16q24.3 for FA-A, 15q14–15 for
RAD51, 12p12 for RAD52, 17q21 for
BRCA1, 14q32 for XRCC3, and 2q34 for
Ku80). A consistently significant association of increased
CIN in tumors with a higher number of LOH at these CIN-related loci was
found, showing, on average, a 7% increase in PLOH for each additional
LOH at these loci (P < 0.0001).
With regard to genetic heterogeneity, it is notable that no DSB-related
genes were identified in well-differentiated tumors, which showed the
lowest level of CIN (Fig. 5B
). Interestingly, two loci
(3p22.1 and 9q22.33) that contributed to the well-differentiated state
were the sites at which hMLH1 and XPA are
located. Defects of these two genes would result in MIN and NIN, which
is consistent with the suggestion that MIN/NIN tumors display a low
level of CIN (20)
.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Using a PCR-based method to detect LOH may be confounded by genomic amplification, which might yield a false-positive loss of markers. To exclude the possibility that the loci defined to show high LOH frequency were attributable to amplification, we examined whether these genomic regions are those known to be frequently amplified and to harbor known breast cancer oncogenes. The oncogenes implicated in breast cancer development are MYC (at 8q24.12–24.13), cyclin D1 (at 11q13–14), and ERBB2 (at 17q11–12; reviewed in Ref. 46 ). As a result, LOH at these loci must be interpreted with caution. However, none of the loci identified to show high LOH frequency in the present study were the "frequently amplified" loci. Thus, possible artifacts attributable to amplification was minor.
Genomic instability exists at two distinct levels, at the nucleotide sequence level, resulting in base substitutions (NIN) or deletions or insertions of a few nucleotides (MIN), and at the chromosome level, resulting in the loss or gain of whole chromosomes or large portions thereof (CIN; Ref. 20 ). In breast cancer, MIN is seen in only a small subset (<10%) of tumors (34) , and there is little evidence of mutation hotspots to support a significant etiological role of NIN. In contrast, aneuploidy, an abnormal number of chromosomes representing the prototype of CIN, is relatively common in breast cancer. On the basis of genome-wide detection of allelic loss (LOH) and the exclusion of tumors presenting aneuploidy, we specifically defined the other form of CIN, i.e., DSB-initiated CIN, as a contributing determinant to breast cancer progression. Support for our hypothesis comes from the observations that: (a) DSB-initiated CIN, measured by PLOH, significantly increased as the tumors progressed to a poorer grade or later stages; (b) in sequential steps toward CIN, the loci of p53 and ATM, the key checkpoint genes against DSB, were lost at the earliest stage; (c) several loci identified as associated with breast tumorigenesis are genomic sites where the genes involved in DSB-related checkpoint/repair are probably located, and tumors possessing a higher number of LOH loci at these sites displayed a greater level of CIN. Furthermore, on the basis of known interaction between genes implicated in the same tumorigenic mechanism, we are also able to verify this proposed model of tumor progression. Our result shows that sequential loss of the locus for p53 (17p12–13) occurred in well-differentiated tumors, followed by loss of the locus for BRCA1 (17q21.31) in moderately differentiated tumors. A recent model of mice homozygous for mutations in both the p53 and Brca1 genes is concordant with this finding. In this model, p53-/- (a defective checkpoint) is the initial event in tumor progression to acquire an increased growth potential and genomic instability. The subsequent loss of BRCA1 function, which plays a role in multiple repair pathways, would be expected to accelerate the accumulation of additional mutations leading to malignant transformation (47) . Etiologically, this model was considered to be a more plausible mechanism for explaining the relationship between p53 and BRCA1 in breast cancer formation, because it does not envisage a growth-arrest stage attributable to loss of BRCA1 (17, 47, 48) . Our model certainly does not exclude the contribution of aneuploidy to breast tumorigenesis. Rather, it shows that several tumorigenic loci probably harbor the genes involved in either chromosomal stability (e.g., WRN, FA-A, and FA-D) or DSB-related checkpoint/repair. These findings suggest that CIN affecting the integrity of both chromosomal number and chromosomal structure is of particular importance for breast cancer development. Consequently, an increased mutation rate (mutator phenotype) attributable to CIN, probably in combination with selection and clonal expansion, provides the basic requirement for breast tumor progression.
One cautionary note should be raised. We were aware that LOH at a few markers cannot be taken as sufficient evidence for the involvement of the genes in tumorigenesis, and the demonstration of specific mutation in putative genes would be required. However, the causal link between some of these suggested genes, such as p53, ATM, and BRCA1, and breast cancer has been confirmed (reviewed in Refs. 8, 11, and 17 ). The increased risk of developing breast cancer in familial cancer syndromes, such as xeroderma pigmentosum and Muir-Torre syndrome (4) , lends additional support to the contribution to breast cancer development of certain other genes, such as XPA and hMLH1. On the other hand, to the best of our knowledge, with the exception of p53, the probability of finding somatic mutation in these checkpoint/repair (caretaker) genes (e.g., RAD51, RAD52, FA-A, FA-D, XPA, and hMLH1) has been shown to be extremely low. Because LOH only represents the specific indicator of the "one-hit" needed to inactivate TSGs, these genes do not seem to fulfil Knudson’s "two-hit" criterion for a TSG in cancer formation. However, recent evidence, which suggests two nonmutually exclusive possibilities, provides support for a tumorigenic role of these genes defined solely by LOH in breast cancer. The first is that a growing number of common TSGs, including p53 and ATM, have been found to exhibit the haplo-insufficiency phenotype (49, 50) , which implies that homozygous inactivating mutations and complete loss of function are not necessary to cause defective tumor suppressor function. Thus, a half normal level of the gene product, resulting from allelic loss or LOH per se, would be able to produce the phenotypic manifestation (51) . It is still possible that other (epigenetic) mechanisms of inactivation abrogating the function of these genes are implicated in breast tumorigenesis. For example, hypermethylation of the BRCA1 promoter region has been found in breast cancer (52) , supporting the role of BRCA1 in sporadic breast cancer development (53) . Therefore, our identification of multiple LOH loci possibly targeting DSB-related genes should not be considered to merely reflect the sites of putative TSGs. Instead, these findings can be further regarded as clues suggesting that breast cancer progression is driven by the accumulation of multiple genetic alterations to inactivate the genes participating in the common defensive mechanism against DSB. In accordance with our findings, a recent breast cancer study involving the detection of LOH at the loci of five DSB-repairing genes (BRCA1, BRCA2, RAD51, RAD52, and RAD54) showed that the probability of simultaneous LOH at more than two of these loci was much higher than expected (54) and underlines the additional importance of finding LOH at these DSB-related loci. More importantly, these findings, including our own, suggest that these proteins associated with DSB-related checkpoint/repair work together to maintain chromosomal stability. As a result, a decrease in the amount of any one of these DSB-related genes as a result of LOH may alter the stoichiometric relationship between them. This inference can be supported by molecular findings that certain of these proteins are found in the same DNA-repairing complex or pathway (reviewed in Refs. 12, 13, and 16 ). Consequently, disruption of any one component of this DNA-repairing complex or pathway may hinder DNA repair and result in the accumulation of unrepaired DSB, subsequently leading to CIN without cell cycle checkpoint and to tumor formation.
The most inconsistent findings regarding the pathogenesis of breast
cancer are those related to whether breast cancer progression follows a
process of dedifferentiation, from well differentiated (grade I) to
poorly differentiated (grade III) over time. The current grading system
for breast cancer is based on semiquantitative indices that assess the
degree of tubule formation, nuclear pleomorphism, and mitotic activity
(27)
and assumes a sequential change from lower grade to
higher grade tumors. Supporting this concept, epidemiological
observations based on a randomized screening trial involving a large
cohort of women indicate that the natural history of breast carcinoma
follows a progressive disease model, by which regular breast screening
to prevent deterioration of the malignancy grade of tumors is
recommended (33)
. If this dedifferentiation is to happen,
one would see a group of lower grade cells adjacent to a group of
higher grade cells in the same tumor population. However, it is not
unusual for some tumors to always present with a poorly differentiated
phenotype without a well-differentiated counterpart. Furthermore, a
recent study that showed distinct chromosomal abnormality profiles in
well-differentiated and poorly differentiated breast tumors (the most
striking difference being the loss of 16q, seen in 65% of grade I
tumors but only 16% in grade III tumors) concludes that different
breast tumor grades may have distinct molecular origins and
pathogenesis (37)
. The findings of the present study,
which are based on a comprehensive and more precise picture of
genome-wide LOH profiles, only confirmed that grade III tumors are not
directly progressed from grade I tumors. However, the results indicate
that the loss of 16q markers was an early event, because the LOH
frequencies were relatively high in tumors of all grades. We also do
not agree with the suggestion that different breast cancer grades have
totally distinct origins. More comprehensively, our model suggests that
breast cancers are sequentially progressive with regard to CIN.
However, CIN might also cause genetic heterogeneity (31)
,
which was revealed by the findings that LOH at some markers was
observed only in the DSIS component but not in the IDC component of the
same tumors. In addition, some markers were found to preferentially
lose at specific tumor grades, suggesting their contribution to genetic
heterogeneity during tumor development. Therefore, breast cancer
progression follows a clonal evolution model driven by CIN, but
different tumors would present distinct molecular profiles caused by
genetic heterogeneity resulting from CIN. On the basis of this model,
the development of a grade I tumor is the consequence of genetic
heterogeneity initially driven by DSB-initiated CIN, and if the genomic
loci targeted by LOH happen to be those triggering MIN or NIN, tumor
cells would progress via a molecular pathway less affected by CIN and
consequently become well-differentiated tumors, generally presenting a
relatively mild clinical progression. This model is more biologically
plausible and is consistent with colorectal cancers with MIN, which are
always associated with a better prognostic phenotype (reviewed in Ref.
55
). Furthermore, the high number of genetic alterations
found to be associated with genetic heterogeneity in poorly
differentiated tumor cells (i.e., the loci of the patterns
of A3 and B3 in Fig. 5B
) has also led to the suggestion
that, as DSB-initiated CIN reaches a maximum (the highest PLOH), breast
cancer would progress to be a highly heterogeneous disease with diverse
molecular profiles. Consequently, as suggested previously, every breast
tumor has a different set of genomic deletions (40)
.
We were particularly interested in genomic loci showing a significantly high frequency of LOH in well-differentiated tumors, because they suggest sites of putative TSGs playing an initiating role in tumor progression. In addition to those associated with DSB, one of these regions, 16q22.1, has been suggested to reflect the involvement of E-cadherin (reviewed in Refs. 34 and 39 ). In breast cancer, an association has been found between loss of E-cadherin expression and advanced stage, poor tumor grade, or unfavorable prognosis (reviewed in Ref. 56 ). The known function of E-cadherin as a cell-cell adhesion molecule contributing to the suppression of tumor metastasis does not seem to favor its initiating role during tumor progression. However, a recent demonstration that the E-cadherin germline mutation is responsible for the development of familial gastric cancer raises the possibility that disruption of normal cell-cell adhesion could also play a role in the initiation of cell proliferation by allowing escape from growth-control signals (57) . Alternatively, E-cadherin could also be a direct growth suppressor via a mechanism mediated by p27Kip1 (58) . Furthermore, the cytoplasmic domain of E-cadherin can modulate the Wnt signaling pathway by inhibiting the availability of free cytoplasmic ß-catenin and thus the subsequent ß-catenin-mediated transactivation of growth-promoting genes (59) . The possibility that disruption of this molecular pathway could result in breast cancer is supported by the finding that Tcf1-/- (the other effect modifier of the ß-catenin pathway) mice developed adenomas in the mammary glands (60) . Interestingly, the locus (5q31.3) at which Tcf1 is located was also defined as one of the important loci contributing to breast cancer progression in our model.
If the proposed DSB-initiated CIN model correctly describes breast cancer pathogenesis, it would be interesting to ask why breast epithelium cells are particularly susceptible to defective DSB-related checkpoint/repair, because this system is believed to be essential in all cells. This tissue specificity may be explained by the possibility that breast cells that have lost DSB-related checkpoint/repair have a growth advantage with respect to DSB-checkpoint/repair-proficient cells. This may be imposed by a selective micro-environment, and in the breast, we suspect an increased exposure to estrogen may be one cause. Recently, the role of reproductive hormones in breast tumorigenesis has been greatly expanded to suggest that estrogen might be a complete carcinogen directly causing genetic alteration and tumor initiation (61) . We have recently provided epidemiological evidence to suggest that breast cancer can be initiated by estrogen via exposure to estrogen metabolites, including catechol estrogen and catechols estrogen quinones (25) . These metabolites cause oxidative DNA damage and depurination (abasic sites), possibly resulting in clustered sites of base damage in both DNA strands, and attempts to correct such damage by standard base excision repair can further lead to the formation of DSBs (62) . This may shed light on why DSB-related checkpoint/repair is particularly important in breast tumorigenesis.
In conclusion, this study demonstrates that DSB-initiated CIN is the major driving force of breast cancer progression, and this mutator phenotype also contributes to tumor heterogeneity. Because the issue of the involvement of CIN in the development of breast cancer has seldom been addressed so extensively, these results, based on genome-wide survey, will serve as an essential foundation for a better understanding of the molecular mechanisms that occur during breast tumorigenesis.
|
FOOTNOTES |
|---|
1 To whom requests for reprints should be
addressed, at Institute of Biomedical Sciences, Academia Sinica, Taipei
11529, Taiwan. Fax: 886-2-2782-3047; E-mail: bmcys{at}ccvax.sinica.edu.tw 
2 The abbreviations used are: TSG, tumor
suppressor gene; DSB, double-strand break; CIN, chromosome instability;
LOH, loss of heterozygosity; LCM, laser capture microdissection; DCIS,
ductal carcinoma in situ; IDC, invasive ductal
carcinoma; CGH, comparative genomic hybridization; MIN, microsatellite
instability; NIN, nucleotide instability; PLOH, the proportion of loci
showing LOH in a tumor.
Received 12/17/99. Accepted 5/16/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J.-C. Yu, S.-l. Ding, C.-H. Chang, S.-H. Kuo, S.-T. Chen, G.-C. Hsu, H.-M. Hsu, M.-F. Hou, L. Y. Jung, C.-W. Cheng, et al. Genetic susceptibility to the development and progression of breast cancer associated with polymorphism of cell cycle and ubiquitin ligase genes Carcinogenesis, September 1, 2009; 30(9): 1562 - 1570. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. M. Kuerer, C. T. Albarracin, W. T. Yang, R. D. Cardiff, A. M. Brewster, W. F. Symmans, N. M. Hylton, L. P. Middleton, S. Krishnamurthy, G. H. Perkins, et al. Ductal Carcinoma in Situ: State of the Science and Roadmap to Advance the Field J. Clin. Oncol., January 10, 2009; 27(2): 279 - 288. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-l. Ding, J.-C. Yu, S.-T. Chen, G.-C. Hsu, S.-J. Kuo, Y. H. Lin, P.-E. Wu, and C.-Y. Shen Genetic variants of BLM interact with RAD51 to increase breast cancer susceptibility Carcinogenesis, January 1, 2009; 30(1): 43 - 49. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-M. Hsu, H.-C. Wang, S.-T. Chen, G.-C. Hsu, C.-Y. Shen, and J.-C. Yu Breast Cancer Risk Is Associated with the Genes Encoding the DNA Double-Strand Break Repair Mre11/Rad50/Nbs1 Complex Cancer Epidemiol. Biomarkers Prev., October 1, 2007; 16(10): 2024 - 2032. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Maeda, T. Chiba, S. Kawashiri, T. Satoh, and K. Imai Epigenetic Inactivation of I{kappa}B Kinase-{alpha} in Oral Carcinomas and Tumor Progression Clin. Cancer Res., September 1, 2007; 13(17): 5041 - 5047. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-T. Bau, Y.-C. Mau, S.-l. Ding, P.-E. Wu, and C.-Y. Shen DNA double-strand break repair capacity and risk of breast cancer Carcinogenesis, August 1, 2007; 28(8): 1726 - 1730. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-L. Lo, J.-C. Yu, S.-T. Chen, G.-C. Hsu, Y.-C. Mau, S.-L. Yang, P.-E. Wu, and C.-Y. Shen Breast cancer risk associated with genotypic polymorphism of the mitotic checkpoint genes: a multigenic study on cancer susceptibility Carcinogenesis, May 1, 2007; 28(5): 1079 - 1086. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-l. Ding, J.-C. Yu, S.-T. Chen, G.-C. Hsu, and C.-Y. Shen Genetic Variation in the Premature Aging Gene WRN: A Case-Control Study on Breast Cancer Susceptibility Cancer Epidemiol. Biomarkers Prev., February 1, 2007; 16(2): 263 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Shin, E. Hyjek, E. Early, and D. M. Knowles Intratumoral Heterogeneity of HER-2/neu in Invasive Mammary Carcinomas Using Fluorescence In-Situ Hybridization and Tissue Microarray International Journal of Surgical Pathology, October 1, 2006; 14(4): 279 - 284. [Abstract] [PDF]
|
|
|
|
 |
|
 M.-F. Tsai, C.-C. Wang, G.-C. Chang, C.-Y. Chen, H.-Y. Chen, C.-L. Cheng, Y.-P. Yang, C.-Y. Wu, F.-Y. Shih, C.-C. Liu, et al. A new tumor suppressor DnaJ-like heat shock protein, HLJ1, and survival of patients with non-small-cell lung carcinoma. J Natl Cancer Inst, June 21, 2006; 98(12): 825 - 838. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Elstrodt, A. Hollestelle, J. H.A. Nagel, M. Gorin, M. Wasielewski, A. van den Ouweland, S. D. Merajver, S. P. Ethier, and M. Schutte BRCA1 Mutation Analysis of 41 Human Breast Cancer Cell Lines Reveals Three New Deleterious Mutants Cancer Res., January 1, 2006; 66(1): 41 - 45. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. E. Ellsworth, D. L. Ellsworth, D. M. Neatrour, B. Deyarmin, S. M. Lubert, M. J. Sarachine, P. Brown, J. A. Hooke, and C. D. Shriver Allelic Imbalance in Primary Breast Carcinomas and Metastatic Tumors of the Axillary Lymph Nodes Mol. Cancer Res., February 1, 2005; 3(2): 71 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Nessling, K. Richter, C. Schwaenen, P. Roerig, G. Wrobel, S. Wessendorf, B. Fritz, M. Bentz, H.-P. Sinn, B. Radlwimmer, et al. Candidate Genes in Breast Cancer Revealed by Microarray-Based Comparative Genomic Hybridization of Archived Tissue Cancer Res., January 15, 2005; 65(2): 439 - 447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Lacroix, R-A Toillon, and G Leclercq Stable 'portrait' of breast tumors during progression: data from biology, pathology and genetics Endocr. Relat. Cancer, September 1, 2004; 11(3): 497 - 522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-T. Bau, Y.-P. Fu, S.-T. Chen, T.-C. Cheng, J.-C. Yu, P.-E. Wu, and C.-Y. Shen Breast Cancer Risk and the DNA Double-Strand Break End-Joining Capacity of Nonhomologous End-Joining Genes Are Affected by BRCA1 Cancer Res., July 15, 2004; 64(14): 5013 - 5019. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. Schlechter, Q. Yang, P. S. Larson, A. Golubeva, R. A. Blanchard, A. de Las Morenas, and C. L. Rosenberg Quantitative DNA Fingerprinting May Distinguish New Primary Breast Cancer From Disease Recurrence J. Clin. Oncol., May 15, 2004; 22(10): 1830 - 1838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. C. Wang, M. Lin, L.-J. Wei, C. Li, A. Miron, G. Lodeiro, L. Harris, S. Ramaswamy, D. M. Tanenbaum, M. Meyerson, et al. Loss of Heterozygosity and Its Correlation with Expression Profiles in Subclasses of Invasive Breast Cancers Cancer Res., January 1, 2004; 64(1): 64 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Lerebours, P. Bertheau, I. Bieche, L.-F. Plassa, M.-H. Champeme, K. Hacene, C. Toulas, M. Espie, M. Marty, and R. Lidereau Two Prognostic Groups of Inflammatory Breast Cancer Have Distinct Genotypes Clin. Cancer Res., September 15, 2003; 9(11): 4184 - 4189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. E. Ellsworth, D. L. Ellsworth, S. M. Lubert, J. Hooke, R. I. Somiari, and C. D. Shriver High-Throughput Loss of Heterozygosity Mapping in 26 Commonly Deleted Regions in Breast Cancer Cancer Epidemiol. Biomarkers Prev., September 1, 2003; 12(9): 915 - 919. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Zhu, W. Qin, H. Ehya, J. Lininger, and E. Sauter Microsatellite Changes in Nipple Aspirate Fluid and Breast Tissue from Women with Breast Carcinoma or Its Precursors Clin. Cancer Res., August 1, 2003; 9(8): 3029 - 3033. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-P. Fu, J.-C. Yu, T.-C. Cheng, M. A. Lou, G.-C. Hsu, C.-Y. Wu, S.-T. Chen, H.-S. Wu, P.-E. Wu, and C.-Y. Shen Breast Cancer Risk Associated with Genotypic Polymorphism of the Nonhomologous End-Joining Genes: A Multigenic Study on Cancer Susceptibility Cancer Res., May 15, 2003; 63(10): 2440 - 2446. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Holst, G. J. Nuovo, M. Esteller, K. Chew, S. B. Baylin, J. G. Herman, and T. D. Tlsty Methylation of p16INK4a Promoters Occurs in Vivo in Histologically Normal Human Mammary Epithelia Cancer Res., April 1, 2003; 63(7): 1596 - 1601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Chang, C. M. Khoo, M. L. Naylor, R. S. Maser, and R. A. DePinho Telomere-based crisis: functional differences between telomerase activation and ALT in tumor progression Genes & Dev., January 1, 2003; 17(1): 88 - 100. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. V. Schmidt Genes Involved in Breast Cancer Progression: Analysis of Global Changes in Gene Expression or Retroviral Tagging? Am. J. Pathol., December 1, 2002; 161(6): 1973 - 1977. [Full Text] [PDF]
|
|
|
|
|
|
H. Rubin and G. R. Anderson Correspondence re: G. Anderson et al., Intrachromosomal Genomic Instability in Human Sporadic Colorectal Cancer Measured by Genome-Wide Allelotyping and Inter-(Simple Sequence Repeat) PCR. Cancer Res., 61: 8274-8283, 2001. Cancer Res., November 1, 2002; 62(21): 6350 - 6351. [Full Text] [PDF]
|
|
|
|
|
|
Z. Li, Z. H. Meng, A. Sayeed, R. Shalaby, B.-M. Ljung, and S. H. Dairkee Genome-wide Allelotyping of a New in Vitro Model System Reveals Early Events in Breast Cancer Progression Cancer Res., October 15, 2002; 62(20): 5980 - 5987. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Novak, E. Oppliger Leibundgut, J. Hager, D. Muhlematter, M. Jotterand, C. Besse, N. Leupin, D. Ratschiller, J. Papp, G. Kearsey, et al. A high-resolution allelotype of B-cell chronic lymphocytic leukemia (B-CLL) Blood, August 13, 2002; 100(5): 1787 - 1794. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Franco, S.-F. Picard, F. Mege, L. Arnould, and S. Lizard-Nacol Absence of Genetic Abnormalities in Fibroadenomas of the Breast Determined at p53 Gene Mutations and Microsatellite Alterations Cancer Res., November 1, 2001; 61(21): 7955 - 7958. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Anderson, B. M. Brenner, H. Swede, N. Chen, W. M. Henry, J. M. Conroy, M. J. Karpenko, J.-P. Issa, J. D. Bartos, J. K. Brunelle, et al. Intrachromosomal Genomic Instability in Human Sporadic Colorectal Cancer Measured by Genome-Wide Allelotyping and Inter-(Simple Sequence Repeat) PCR Cancer Res., November 1, 2001; 61(22): 8274 - 8283. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Maitra, I. I. Wistuba, C. Washington, A. K. Virmani, R. Ashfaq, S. Milchgrub, A. F. Gazdar, and J. D. Minna High-Resolution Chromosome 3p Allelotyping of Breast Carcinomas and Precursor Lesions Demonstrates Frequent Loss of Heterozygosity and a Discontinuous Pattern of Allele Loss Am. J. Pathol., July 1, 2001; 159(1): 119 - 130. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Loeb A Mutator Phenotype in Cancer Cancer Res., April 1, 2001; 61(8): 3230 - 3239. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, no LOH; ni, noninformative.