RESEARCH ARTICLE

Mitochondrial DNA Mutations and MitochondrialDNA Depletion in Breast Cancer

Ling-Ming Tseng,1 Pen-Hui Yin,2,3 Chin-Wen Chi,2,4 Chih-Yi Hsu,1 Chew-Wun Wu,1 Liang-Ming Lee,5

Yau-Huei Wei,3 and Hsin-Chen Lee4,6*

1Departmentof Surgery,Taipei Veterans General Hospital, and National Yang-Ming University,Taiwan,Republic of China2Departmentof Medical Research and Education,Taipei Veterans General Hospital,Taiwan,Republic of China3Departmentof Biochemistry and Molecular Biology,Schoolof Medicine,National Yang-Ming University,Taiwan,Republic of China4Department and Institute of Pharmacology,School of Medicine,National Yang-Ming University,Taiwan,Republic of China5Departmentof Urology,Taipei Medical University-Aff|liated Taipei Municipal Wan-Fang Hospital,Taiwan,Republic of China6Departmentof Education and Research,Taipei City Hospital,Taipei,Taiwan,Republic of China

Somatic mutations in mitochondrial DNA (mtDNA) have been demonstrated in various tumors, including breast cancer. How-

ever, it still remains unclear whether the alterations in mtDNA are related to the clinicopathological features and/or the progno-

sis in the breast cancer. We analyzed somatic mutations in the D-loop region, the common 4,977-bp deletion, and the copy num-

ber of mtDNA in breast cancer and paired nontumorous breast tissues from 60 Taiwanese patients. We found that 18 of the 60

(30%) breast cancers displayed somatic mutations in mtDNA D-loop region. The incidence of the 4,977-bp deletion in nontumo-

rous breast tissues (47%) was much higher than that in breast cancers (5%). The copy number of mtDNA was significantly

decreased in 38 of the 60 (63%) breast cancers as compared to their corresponding nontumorous breast tissues (P ¼ 0.0008).

The occurrence of D-loop mutations was associated with an older onset age (�50 years old, P ¼ 0.042), and tumors that lacked

expressions of estrogen receptor and progesterone receptor (P¼ 0.024). Patients with mtDNA D-loop mutation and breast can-

cer had significantly poorer disease-free survival than those without mutation, when assessed by Kaplan–Meier curves and log-

rank test (P ¼ 0.005). Multivariate Cox regression analysis indicated that a D-loop mutation is a significant marker that is inde-

pendent of other clinical variables and that it can be used to assess the prognosis of patients. Our findings suggest that somatic

mutations in mtDNA D-loop can be used as a new molecular prognostic indicator in breast cancer. VVC 2006 Wiley-Liss, Inc.

INTRODUCTION

Breast cancer is the fourth leading cause of can-

cer death among Taiwanese women. Its incidence

has rapidly increased over the past decade both in

Taiwan and in other areas of Asia (Seow et al.,

1996; Chen et al., 2002). According to epidemiolog-

ical studies, prolonged exposure to estrogen,

including a reduced fertility rate, an earlier men-

arche, and prolonged reproductive stimulation dur-

ing lifetime, is significantly associated with an

increased risk of female breast cancer (Pike et al.,

1993; Shen et al., 2005). The oxidized metabolites

of estrogen, E2,3,4-semi-quinones and E2,3,4-qui-

nones, have been shown to bind to DNA to form

adducts, which may lead to genetic damage (Yager,

2000). In addition, the generation of reactive oxy-

gen species (ROS) during the conversion of

E2,3,4-semi-quinones to E2,3,4-quinones can also

lead to oxidative DNA damage. Estrogen metabo-

lites-induced oxidative stress has thus been

thought to play an important role in the initiation

of breast carcinogenesis (Yager, 2000).

Mitochondria are cytoplasmic organelles and

have a variety of important roles to play, including

the generation of ATP through oxidative phospho-

rylation (OXPHOS), the production of ROS, and

the initiation of apoptosis (Wallace, 1999). Human

mitochondrial DNA (mtDNA) is a 16.6-kb double-

stranded circular DNA molecule, and multiple

copies of mtDNA are present in each mitochond-

rion. The human mitochondrial genome encodes

13 polypeptides that form part of the respiratory

chain, together with 22 transfer RNAs and two ri-

bosomal RNAs that are required for protein syn-

thesis (Anderson et al., 1981). mtDNA is more sus-

ceptible to oxidative damage and has a higher

mutation rate than nuclear DNA due to a lack of

protective histone proteins, limited DNA repair

*Correspondence to: Dr. Hsin-Chen Lee, Department and Insti-tute of Pharmacology, School of Medicine, National Yang-Ming Uni-versity, Taipei, Taiwan 112, Republic of China.E-mail: hclee2@ym.edu.tw

Supported by: Taipei Veterans General Hospital, Grant numbers:V93-234 and V94-248; National Science Council, Taiwan, Republicof China, Grant numbers: NSC 93-2320-B-010-058, NSC 94-2314-B-075-035, and NSC 94-2320-B-010-063; Chen Shuyi Cancer Foundation.

Received 21 December 2005; Accepted 17 February 2006

DOI 10.1002/gcc.20326

Published online 27 March 2006 inWiley InterScience (www.interscience.wiley.com).

VVC 2006 Wiley-Liss, Inc.

GENES, CHROMOSOMES & CANCER 45:629–638 (2006)

mechanisms, and a high rate of generation of ROS

in mitochondria (Croteau and Bohr, 1997). Somatic

alterations in mtDNA can result in impairment of

OXPHOS and enhanced ROS production, which

in turn accelerates the rate of DNA mutation. This

scenario has been proposed to contribute to the

early stages of carcinogenesis (Penta et al., 2001).

In the past few years, somatic mtDNA mutations

have been reported in several types of cancers

(Polyak et al., 1998; Fliss et al., 2000; Penta et al.,

2001), including breast cancer (Parrella et al., 2001;

Tan et al., 2002; Zhu et al., 2005). Most of the

mutations occur in the D-loop region, the major

control site for mtDNA replication and trans-

cription (Sanchez-Cespedes et al., 2001; Rosson

and Keshgegian, 2004). Mitochondrial D-loop

DNA mutations have been shown to be correlated

with less-differentiated hepatocellular carcinomas

(Tamori et al., 2004) and with stage progression

and prognosis in non-small cell lung cancers (Mat-

suyama et al., 2003). In addition, mtDNA deple-

tion has also been demonstrated in various cancers

(Simonnet et al., 2002; Lee et al., 2004, 2005; Yin

et al., 2004; Wu et al., 2005). mtDNA depletion

was found to be associated with tumor aggressive-

ness in renal cell carcinoma (Simonnet et al., 2002)

and with less-differentiated gastric carcinoma (Wu

et al., 2005). Although mitochondrial genome insta-

bility and somatic alterations have been demon-

strated in breast cancers, the correlation between

the mtDNA mutations and the clinicopathological

parameters of breast cancer has remained unclear.

To address these issues, we sequenced the

mtDNA D-loop region, quantified the mtDNA

content, and searched for the common 4,977-bp

deletion among breast cancers and corresponding

nontumorous breast tissues of 60 Taiwanese female

patients. The relationship between the mtDNA

alterations, the clinicopathological parameters, and

the prognosis of breast cancer was then analyzed.

MATERIALS ANDMETHODS

Collection of Human Breast Cancer Tissues and

DNA Extraction

Sixty breast cancer samples and their adjacent

nontumorous breast tissues were obtained with

consent from female patients at Taipei Veterans

General Hospital. Among the 60 breast cancer

patients, 32 were under 50 years (the young age

group) and 28 were �50 years old (the older age

group). Twenty-six were postmenopausal patients.

All of the tissues were stored in liquid nitrogen im-

mediately after surgical resection according to a

protocol approved by the medical ethics committee

for conducting human research at the hospital. Total

cellular DNA from the tissues was extracted using

the QIAamp DNA Mini kit (Qiagen, Hilden,

Germany) according to the instructions of the manu-

facturer. The final DNA pellet was dissolved in

double distilled water and frozen at�308C until use.

Direct Sequencing for Screening of Somatic

Mutation in mtDNA D-Loop

Somatic mutations in the D-loop region of

mtDNA were analyzed by direct sequencing of the

products of polymerase chain reaction (PCR) as

described previously (Lee et al., 2004; Yin et al.,

2004). The primer pairs L16190 (np 16,190–

16,209, 50-CCCCATGCTTACAAGCAAGT-30) and

H602 (np 602–583, 50-GCTTTGAGGAGGTAA-

GCTAC-30) were used for the amplification of a

982-bp DNA fragment from the D-loop region

of the mtDNA. PCR was performed in an ABI

GeneAmp PCR System 9700 DNA thermal cycler

(Applied Biosystems, Foster City, CA). The reac-

tions were carried out for 30 cycles in a 50 ll reac-tion mixture containing 100 ng DNA, 200 lM of

each dNTP, 20 pmol of each primer, 2.5 U of

PfuUltra high-fidelity DNA polymerase (Strata-

gene, La Jolla, CA), and 13 PfuUltra HF reaction

buffer. The PCR cycles consisted of 15 sec denatu-

ration at 948C, 15 sec annealing at 588C, and 90 sec

primer extension at 728C. All of the PCR products

were subjected to nucleotide sequencing on an

ABI PRISM1 3100 Genetic Analyzer (Applied Bio-

systems) according to the instructions of the manu-

facturer. All D-loop sequences were interpreted by

the same investigator, and the person did not know

the identity of the patients, or the clinicopathologi-

cal and outcome data. In addition, the primer

L76 (np 76–100, 50-CACGCGATAGCATTGC-

GAGACGCTG-30) was used for re-sequencing to

confirm the mutations in np 303–309 poly-C tract.

Detection of the 4,977-bp Deletion of mtDNA

The 4,977-bp deletion of mtDNA was detected

by using the primers L8150 (50-CCGGGGGTA-

TACTACGGTCA-30) and H13650 (50-GGGGAA-

GCGAGGTTG ACCTG-30) (Lee et al., 2001,

2004; Yin et al., 2004; Wu et al., 2005). PCR was

performed in an ABI GeneAmp PCR System 9700

DNA thermal cycler. The reactions were carried

out for 35 cycles in a 50 ll reaction mixture con-

taining 200 ng DNA, 200 lM of each dNTP,

20 pmol of each primer, 1.0 U of Taq DNA poly-

merase, and 13 reaction buffer. The PCR cycles

consisted of 15 sec denaturation at 948C, 15 sec

Genes, Chromosomes & Cancer DOI 10.1002/gcc

630 TSENG ETAL.

annealing at 588C, and 40 sec primer extension at

728C. A 524-bp PCR product amplified from

mtDNA with 4,977-bp deletion was detected by

electrophoresis on a 1.5% agarose gel at 100 V for

40 min and under UV transillumination after ethi-

dium bromide staining. This method is sensitive

enough to detect the presence of as low as 0.01%

of mtDNA molecules with the deletion (Lee et al.,

2001).

Determination of mtDNAContent

For the determination of mtDNA content rela-

tive to nuclear DNA, the forward primer 50-ACCCACACTGTGCCCATCTAC-30 and the

reverse primer 50-TCGGTGAGGATCTTCATG-

AGGTA-30 (complementary to the sequences of

the b-actin gene) were used to amplify a 107-bp

product (Kim et al., 2004). The PCR was per-

formed in a Roche Light Cycler apparatus, using

the Faststart DNA master SYBR Green kit (Roche

Manheim, Germany). DNA (100 ng) was mixed

with a buffer containing 5 mM MgCl2, 0.2 mM

dNTPs, 20 pmol of forward and reverse primers,

SYBR green I dye, and 0.25 U Hot Start Taq DNA

polymerase in a final volume of 20 ll. The reac-

tions were performed as follows: initial 300 sec

denaturation at 958C followed by 40 cycles of 1 sec

at 958C, 6 sec at 588C, and 18 sec at 728C. For theanalysis of mtDNA, the forward primer 50-CACC-CAAGAACAGGGTTTGT-30 and the reverse

primer 50-TGGCCATGGGTATGTTGTTAA-30,which are complementary to the sequence of the

ND1 gene, were used to amplify a 108-bp PCR

product. The threshold cycle number (Ct) values

of the b-actin gene and the mitochondrial ND1

gene were determined for each individual quanti-

tative PCR run. The –ddCt (mtDNA to b-actingene) represents the mtDNA content in a cell.

Each measurement was carried out at least three

times and normalized in each experiment against a

serial dilution series of a control DNA sample.

Statistical Analysis

Qualitative and quantitative changes in mtDNA

were analyzed using the Statistical Program for

Social Sciences program package. Fisher’s exact

test was used to compare mtDNA alterations and

clinicopathological parameters. The overall sur-

vival (OS) and disease-free survival (DFS) rates of

patients with and without mtDNA alterations were

analyzed by Kaplan–Meier estimates and com-

pared by the log-rank test. OS and DFS were cal-

culated from the date of tumor diagnosis. Cox pro-

portional hazards regression methods were used to

investigate the relationship between survival, clini-

copathological variables, and mtDNA alterations

using both univariate and multivariate models.

Hazard ratios are presented with their 95% confi-

dence intervals (95% CI). All statistical tests were

two-sided. The difference between groups was

considered statistically significant when the Pvalue was smaller than 0.05.

RESULTS

Somatic Mutations in the D-Loop of mtDNA in

Breast Cancer

mtDNA from 60 pairs of tumor and matched ad-

jacent nontumorous breast tissues were analyzed

by direct DNA sequencing. After nucleotide

sequencing, nucleotide changes in breast cancers

were determined as somatic mutations by compari-

son with the sequence in the noncancerous part of

the same patient. Eighteen of 60 (30%) tumors had

somatic mutations in the mtDNA D-loop, and 13

of them (72%) had mutations in np 303–309 poly-C

tract (Table 1). Four tumors had 2 mutations, and

the remaining 14 tumors had 1 mutation (Table 1).

Among the 22 somatic mutations of mtDNA, 19

mutations were heteroplasmic in the tumor tissue

and 10 mutations were heteroplasmic in the

matched nontumorous tissue. The results indicate

that somatic mutations occurred more frequently

in the mtDNA D-loop region of breast cancer dur-

ing carcinogenesis.

The 4,977-bp Deletion of mtDNA in Breast Cancer

The common 4,977-bp deletion of mtDNA was

detected in 28 (47%) nontumorous breast tissues of

the 60 patients with breast cancer (Fig. 1 and Table 1).

Only three (5%) of the breast cancer samples were

found to carry the mtDNA deletion.

mtDNA Depletion in Breast Cancer

The mtDNA contents of the 60 pairs of tumor

and corresponding nontumorous breast tissues

were measured by quantitative real-time PCR.

The gel in Figure 2A shows only single bands that

were specifically amplified from mtDNA and b-actin gene (nuclear DNA). The PCR fragments

were sequenced to confirm the specificity of the

primer pairs (data not shown). The efficiency of

the real-time PCR amplification was examined

using plasmid containing inserted mtDNA and b-actin gene fragments and the results indicated

good efficiency (r2 ¼ 0.984 for mtDNA and 0.999

for b-actin) for our quantitative method (Fig. 2B).

We found that the mean mtDNA content of breast

Genes, Chromosomes & Cancer DOI 10.1002/gcc

631MITOCHONDRIAL DNA ALTERATIONS IN BREAST CANCER

TABLE 1. Summary of the mtDNA Mutations Found in 60 Primary Breast Cancers

Patient number

D-loop mutation

mtDNA depletion

4,977-bpdeletion

np Mutation N Tu

257 þ � �314 � þ þ324 303 8C?8C/9C þ � �

514 5CA?4CA340 303 8C/9C?8C/9Ca þ þ �349 þ þ �350 � � �371 þ þ �375 � þ �377 � � �381 þ � �383 � þ �385 þ � �386 � � �387 303 8C/9C?7C/8C/9C þ � �416 � � �420 150 T?C þ � �425 16390 G?A/G þ � �

303 8C?8C/9C430 303 8C/9C/10C?7C/8C/9C � þ �432 þ þ �438 þ þ �440 � þ �446 303 8C/9C?8C/9Ca þ þ �447 203 G?A � þ �451 þ þ �464 � þ �489 þ � �498 � � �501 � � �510 188 A?G/A þ þ �

303 9C?8C/9C511 þ þ �529 � � �546 þ þ �621 16290 C?T/C þ þ �651 þ þ �692 303 8C?7C/8C þ þ �717 � � �732 þ � �739 188 A?G/A � � �

303 8C/9C/10C?8C/9C760 þ � �765 303 8C/9C/10C?7C/8C/9C/10C þ � �776 303 8C/9C/10C?8C/9C/10C/11C þ � �777 þ þ þ779 þ � �780 þ � �781 þ � �797 � þ �801 þ þ þ806 152 C/T?T/C þ � �807 303 7C/8C/9C?8C/9C þ þ �871 þ � �900 þ þ �954 þ � �955 � þ �

(Continued)

Genes, Chromosomes & Cancer DOI 10.1002/gcc

632 TSENG ETAL.

cancers was significantly lower than that of the cor-

responding nontumorous breast tissues (paired Stu-

dent’s t test, P ¼ 0.0008). Thirty-eight of the 60

(63%) tumors had an obviously lower mtDNA con-

tent compared to their corresponding nontumorous

breast tissue (Table 1). The results were further

confirmed by competitive PCR using different

primer pairs for the mtDNA and the b-actin gene

(data not shown). The observations indicate that

depletion of mtDNA has occurred in most of breast

cancers.

mtDNA Alterations in Relation to

Clinicopathological Parameters

The relationship between mtDNA alterations

and the clinicopathological parameters are sum-

marized in Table 2. Somatic mutations in the

mtDNA D-loop were more frequently detected

(12/28) in the breast cancer of the older onset age

group (�50 years old) than in (6/32) the young age

group (<50 years old, P ¼ 0.042), but mtDNA

depletion and the 4,977-bp deletion were not sig-

nificantly different between the two groups.

Moreover, the occurrence of a somatic mutation

in the mtDNA D-loop was significantly correlated

with breast tumors that did not show expression of

the estrogen receptor (P ¼ 0.029), and marginally

correlated with the tumors that did not show

expression of the progesterone receptor (P ¼0.055). Breast tumors lacking expression of both

estrogen and progesterone receptors had a signifi-

cantly higher frequency of D-loop mutation com-

pared to the tumors that expressed either the estro-

gen receptor and/or the progesterone receptor (P ¼0.024). However, there was no significant correla-

tion between mtDNA depletion or the 4,977-bp

deletion, and the clinicopathological parameters.

Association of mtDNA D-Loop Mutation

with Disease-Free Patient Survival

To assess the prognostic significance of the

mtDNA alterations, we analyzed the OS and DFS

by Kaplan–Meier curves and log-rank test. We

found that patients with a D-loop mutation had a

poorer OS (P ¼ 0.066), but this was not significant

when compared to those without a mutation.

There was also no significant difference in OS

between the patients with and without mtDNA

depletion (P ¼ 0.235). The DFS of the patients

with a D-loop mutation (mean 6 SD: 27.3 6 4.3

months, 95% CI: 18.8–35.7) was significantly lower

than the one of the patients without a mutation

(mean 6 SD: 41.3 6 2.4 months, 95% CI: 36.6–

46.0, P ¼ 0.005; Fig. 3A). The patients with

mtDNA depletion, however, showed no significant

difference in DSF compared to those without

mtDNA depletion (P ¼ 0.975; Fig. 3B). In addi-

tion, univariate Cox regression analysis revealed

that the presence of axillary lymph node involve-

ment was significantly related to a shorter DFS (P¼ 0.031; Table 3). Moreover, the expression status

of the estrogen receptor (P ¼ 0.057) had only a

marginal relationship with DFS (Table 3). Multi-

variate Cox regression analysis under D-loop muta-

tion clustering provided independent information

TABLE 1. Summary of the mtDNA Mutations Found in 60 Primary Breast Cancers (Continued)

Patient number

D-loop mutation

mtDNA depletion

4,977-bp deletion

np Mutation N Tu

958 310 T loss þ þ �961 16304 C?T � þ �1099 � � �1113 þ � �1133 þ � �1154 � þ �1215 � � �Tu, tumor portion; N, nontumorous portion.aThese tumors showed a quantitative alteration in the C-tract pattern.

Figure 1. The mtDNA 4,977-bp deletion in tumor and nontumo-rous breast tissue. The 4,977-bp deletion of mtDNA was detected byPCR as described in Materials and Methods. The primers L8150 andH13650 were used for the amplification of a 524-bp PCR product fromthe 4,977-bp deleted mtDNA in tumor (T) and nontumorous breast tis-sue (N). M: DNA 100-bp ladder.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

633MITOCHONDRIAL DNA ALTERATIONS IN BREAST CANCER

with respect to the prognosis for disease recurrence

among patients (Table 4).

DISCUSSION

The results obtained in this study indicate that

most (70%) of the 60 breast cancers harbored

mtDNA that contained a variety of alterations,

including point mutations in mtDNA D-loop (n ¼18), mtDNA depletion (n ¼ 38), and the common

4,977-bp deletion (n ¼ 3). The mtDNA content

was significantly decreased in most of the breast

cancers tested compared with the corresponding

nontumorous breast tissues (P ¼ 0.0008). The inci-

dence of the common 4,977-bp deletion in non-

tumorous breast tissues was higher than in breast

cancer. We report for the first time that the occur-

rence of somatic mutations in mtDNA D-loop

could be linked to the older onset age group, and

that these tumors lacked expression of the estrogen

receptor and/or the progesterone receptor in the

breast cancer tissue (Table 2). The patients with

an mtDNA D-loop mutation in the breast tumor

had a lower DFS than those without the change.

Traditionally, the most significant prognostic

indicator for patients with breast cancer is the

presence or absence of axillary lymph node in-

volvement (Cianfrocca and Goldstein, 2004).

Breast cancer lacking expression of estrogen and

progesterone receptors has also been related to

poor prognosis (Cianfrocca and Goldstein, 2004).

In our sample set, positive axillary lymph node was

correlated with a poor DFS (P ¼ 0.031). Patients

with breast cancer lacking expression of the estro-

gen or progesterone receptors also showed a trend

toward a shorter DFS, but this was not significant

(P ¼ 0.057 and 0.084, respectively). We found that

patients with an mtDNA D-loop mutation showed

a significant relationship to poor DFS (Tables 3

and 4), although the occurrence of the mutation

did not show an association with axillary lymph

node status (Table 2). Moreover, a D-loop muta-

tion provided independent prognostic information

on disease recurrence (Table 4). Our findings thus

suggest that somatic mutations in the D-loop

region of mtDNA in breast cancer can be used as a

new molecular prognostic biomarker.

In this study, some of the mtDNA D-loop muta-

tions detected in the tumors or surrounding non-

tumorous tissues were heteroplasmic. Heteroplas-

mic mtDNA mutations in nontumorous tissue have

also been reported in other breast cancer studies

(Tan et al., 2002; Rosson and Keshgegian, 2004).

This is possibly because the DNA was not derived

from a microdissected tumor; the heteroplasmic

mutation could hence be attributable to the con-

tamination from the surrounding non-neoplastic

cells. In the past few years, similar heteroplasmic

mutation patterns of the mtDNA D-loop have

been demonstrated in various nontumorous tissues

from elderly subjects (Coskun et al., 2003, 2004).

Our results reveal that the occurrence of D-loop

mutations is associated with an older onset age

(P ¼ 0.042), suggesting that aging-related DNA

damage may contribute to the mtDNA D-loop

mutations found in breast cancer. It is noteworthy

that heteroplasmic mutation patterns of the mtDNA

were also observed in the surrounding nontumorous

tissues. These results suggest that an mtDNA D-

loop mutation may occur before morphological

changes found at the early stages of breast tumori-

genesis.

Figure 2. mtDNA depletion in breast cancer. mtDNA content wasdetermined by the quantitative real-time PCR as described in Materialsand Methods. Panel A: Only a single band for each lane is shown andthis is the PCR product specifically amplified from (1) the b-actin geneand (2) the mtDNA. M: DNA 100-bp ladder. Panel B: The efficiency ofthe real-time PCR amplification was examined using 10�4 to 1 ng of

plasmids containing either an mtDNA insert (open circles) or a b-actingene insert (filled circles) and the threshold cycle number (Ct) valueswere plotted as a function of the logarithm of amount of plasmid. Thecorrelation coefficient r2 is 0.999 for b-actin gene and 0.984 formtDNA.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

634 TSENG ETAL.

Reduced mtDNA content was detected in most

breast cancers. This finding was consistent with a

recent study of mtDNA content, which showed

that a reduction occurred in most examined breast

tumors, but this proportion was greater for papil-

lary thyroid carcinomas, suggesting that changes

in mtDNA content during carcinogenesis may be

regulated in a tumor-specific manner (Mambo et

al., 2005). In addition, it has been reported that

somatic mutations in mtDNA D-loop are associ-

ated with mtDNA depletion in hepatocellular car-

cinomas (Lee et al., 2004). We also found that

72% of the breast cancers with mtDNA D-loop

mutations had a reduced mtDNA copy number.

There is increasing evidence that mtDNA muta-

tion- and depletion-induced OXPHOS dysfunc-

tion is associated with an increased tumorigenicity

(Petros et al., 2005) and an invasive phenotype

(Amuthan et al., 2001). Thus, the decrease in the

copy number of mtDNA in breast cancer may

result in mitochondrial dysfunction, and contrib-

ute to altered energy metabolism, increased ROS,

and an attenuated apoptotic response to anti-

cancer drugs; these might, in turn, promote

TABLE 2. Clinicopathological Features of the Breast Cancer Patients With and Without mtDNA Alterations

n

D-loop mutation

P-value

mtDNA depletion

P-valueNegative(n ¼ 42)

Positive(n ¼ 18)

Negative(n ¼ 22)

Positive(n ¼ 38)

Age (years)<50 32 26 6 13 19�50 28 16 12 0.042* 9 19 0.496

Menopausal statusPremenopausal 34 25 9 14 20Postmenopausal 26 17 9 0.575 8 18 0.430

TNM stageI 10 10 0 4 6II 26 16 10 7 19III 19 12 7 9 10IV 5 4 1 0.084 2 3 0.532

Tumor size (cm)<2.0 7 6 1 2 52.0–5.0 47 33 14 17 30>5.0 6 3 3 0.406 3 3 0.797

Axillary lymph node status0 30 23 7 9 211–3 8 4 4 4 4>3 22 15 7 0.334 9 13 0.461

Histologic grade1 6 6 0 4 22 34 25 9 11 233 20 11 9 0.081 7 13 0.298

Lymphatic and vascular invasionPositive 18 11 7 8 10Negative 42 31 11 0.325 14 28 0.413

ERPositive 36 29 7 14 22Negative 24 13 11 0.029* 8 16 0.662

PRPositive 28 23 5 13 15Negative 32 19 13 0.055 9 23 0.142

ER/PRERþ/PRþ 26 21 5 11 15ERþ/PR� 10 8 2 1.000 3 7 0.706ER�/PRþ 2 2 0 1.000 2 0 0.206ER�/PR� 22 11 11 0.024* 6 16 0.278

Her-2/neu receptorPositive 14 10 4 7 7Negative 46 32 14 1.000 15 31 0.237

ER, estrogen receptor; PR, progesterone receptor.*Difference between the groups is statistically significant.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

635MITOCHONDRIAL DNA ALTERATIONS IN BREAST CANCER

unconstrained proliferation and invasion (Gate-

nby and Gillies, 2004).

We detected the common 4,977-bp deletion of

mtDNA in only three breast cancers but in 47% of

the nontumorous breast tissues. The finding that

there is a lower incidence of the deletion in tumor-

ous tissue than that in nontumorous tissue is con-

sistent with finding in various other cancers (Lee

et al., 2001; Yin et al., 2004; Wu et al., 2005). The

evidence from a previous study of 17 breast cancer

patients by microdissection of the tumor tissue and

in situ PCR provides additional support (Dani et

al., 2004). The common 4,977-bp deletion has

been demonstrated to accumulate with age, pri-

marily in postmitotic tissues (Lee and Wei, 2001).

Moreover, the common mtDNA deletion causes a

loss of five tRNA genes and seven genes encoding

subunits of cytochrome oxidase, Complex I and

ATPase, and may have a strong functional disad-

vantage, possibly thereby repressing the growth of

cancer cells harboring deleted mtDNA. It was

recently demonstrated in transmitochondrial cy-

brid cells that the common mtDNA deletion sensi-

tized the cells to apoptosis at low heteroplasmy

levels (Schoeler et al., 2005). Thus, a decrease in

the incidence of the common deletion in tumors

suggests that there is either a dilution effect due to

rapid cytoplasmic division or an active selection

mechanism that eliminates cancer cells harboring

the mtDNA deletion.

Because the oxidative metabolites of estrogen,

in particular the catechol estrogens, can generate

ROS and form direct adducts with DNA and gluta-

thione (Yager, 2000), breast cancer is considered to

be an estrogen-inducible cancer. It has been

reported that the concentrations of coenzyme Q10,

an important antioxidant and an essential compo-

nent in the electron transfer chain in mitochondrial

Figure 3. mtDNA alterations and patient’s survival. The Kaplan–Meiersurvival curve and log-rank test were used to analyze the disease-free sur-vival (DFS) of the patients with a D-loop mutation (A) or mtDNA deple-tion (B) as compared to those without mutation or depletion.

TABLE 3. Cox Proportional Hazards Univariate Analysisof Disease-Free Survival

Variables HR (95% CI) P-value

D-loop mutationNegative 1.00Positive 3.35 (1.35–8.32) 0.009

mtDNA depletionNegative 1.00Positive 1.02 (0.40–2.58) 0.975

Age (years)<50 1.00�50 1.22 (0.49–2.99) 0.671

Menopausal statusPremenopausal 1.00Postmenopausal 1.13 (0.46–2.77) 0.797

TNM stageI 1.00I vs. II 25,439 (0.00–1.37 3 1093) 0.922I vs. III 61,256 (0.00–3.30 3 1093) 0.916I vs. IV 132,028 (0.00–7.12 3 1093) 0.910

Tumor size (cm)<2.0 1.00<2.0 vs. 2.0–5.0 5.80 (0.65–52.07) 0.117<2.0 vs. >5.0 2.00 (0.26–15.25) 0.504

Axillary lymph node statusNegative 1.00Positive 3.09 (1.11–8.58) 0.031

Histologic grade1 1.001 vs. 2 1.68 (0.21–13.45) 0.6271 vs. 3 4.44 (0.57–34.78) 0.156

Lymphatic and vascular invasionNegative 1.00Positive 1.10 (0.42–2.88) 0.854

ERNegative 1.00Positive 0.41 (0.17–1.03) 0.057

PRNegative 1.00Positive 0.43 (0.16–1.12) 0.084

Bold type indicates a significance of P < 0.05. HR, hazard ratio; ER,

estrogen receptor; PR, progesterone receptor.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

636 TSENG ETAL.

inner membrane, is significantly decreased in

breast cancers compared to the corresponding non-

tumorous tissues (Portakal et al., 2000). In the pres-

ent study, most of the detected somatic mutations

in mtDNA D-loop were in the np 303–309 poly-C

tract, which has been demonstrated to be more sus-

ceptible to oxidative damage than other regions of

mtDNA (Mambo et al., 2003). Moreover, it has

been recently shown that loss of the tumor sup-

pressor TP53 in cancer cells results in an increased

mtDNA vulnerability to damage induced by exog-

enous and endogenous oxidative stress (Achanta et

al., 2005). Therefore, the D-loop mutations and

depletion of mtDNA in breast cancer may result

from enhanced ROS produced by estrogen metab-

olism during carcinogenesis.

In conclusion, our study reveals that several

types of mtDNA alterations occur in breast can-

cers. mtDNA depletion was frequently detected in

the tumors, but the mtDNA 4,977-bp deletion was

found to accumulate in nontumorous tissues rather

than in tumor tissue. Somatic mutations in mtDNA

D-loop region were associated with tumors lacking

expression of the estrogen and progesterone recep-

tors, as well as a poor DFS rate. Our findings sug-

gest that the increased oxidative stress associated

with aging and estrogen exposure may be involved

in mtDNA alterations during tumorigenesis and

that somatic mutations in the D-loop of mtDNA

can be used as a new molecular prognostic indica-

tor in breast cancer.

ACKNOWLEDGMENTS

The authors thank Ms. Shu-Hui Li for excellent

technical assistance.

REFERENCES

Achanta G, Sasaki R, Feng L, Carew JS, Lu W, Pelicano H,Keating MJ, Huang P. 2005. Novel role of p53 in maintainingmitochondrial genetic stability through interaction with DNA Polgamma. EMBO J 24:3482–3492.

Amuthan G, Biswas G, Zhang SY, Klein-Szanto A, Vijayasarathy C,Avadhani NG. 2001. Mitochondria-to-nucleus stress signalinginduces phenotypic changes, tumor progression and cell invasion.EMBO J 20:1910–1920.

Anderson S, Bankier AT, Barrell BG, De Bruijn MHL, Coulson AR,Drouin J, Eperon IC, Nierlich DP, Roe BA, Sanger F, Schreier PH,Smith AJH, Staden R, Young IG. 1981. Sequence and organizationof the human mitochondrial genome. Nature 290:457–465.

Chen CJ, You SL, Lin LH, Hsu WL, Yang YW. 2002. Cancer epide-miology and control in Taiwan: A brief review. Jpn J Clin Oncol32:S66–S81.

Cianfrocca M, Goldstein LJ. 2004. Prognostic and predictive factorsin early-stage breast cancer. Oncologist 9:606–616.

Coskun PE, Beal MF, Wallace DC. 2004. Alzheimer’s brains harborsomatic mtDNA control-region mutations that suppress mito-chondrial transcription and replication. Proc Natl Acad Sci USA101:10726–10731.

Coskun PE, Ruiz-Pesini E, Wallace DC. 2003. Control regionmtDNA variants: Longevity, climatic adaptation, and a forensicconundrum. Proc Natl Acad Sci USA 100:2174–2176.

Croteau DL, Bohr VA. 1997. Repair of oxidative damage to nuclearand mitochondrial DNA in mammalian cells. J Biol Chem 272:25409–25412.

Dani MA, Dani SU, Lima SP, Martinez A, Rossi BM, Soares F,Zago MA, Simpson AJ. 2004. Less Delta mtDNA4977 than nor-mal in various types of tumors suggests that cancer cells are essen-tially free of this mutation. Genet Mol Res 3:395–409.

Fliss MS, Usadel H, Caballero OL, Buta MR, Eleff SM, Jen J,Sidransky D. 2000. Facile detection of mitochondrial DNA muta-tions in tumors and bodily fluids. Science 287:2017–2019.

Gatenby RA, Gillies RJ. 2004. Why do cancers have high aerobicglycolysis? Nat Rev Cancer 4:891–899.

KimMM, Clinger JD, Masayesva BG, Ha PK, Zahurak ML, Westra WH,Califano JA. 2004. Mitochondrial DNA quantity increases with his-topathologic grade in premalignant and malignant head and necklesions. Clin Cancer Res 10:8512–8515.

Lee HC, Li SH, Lin JC, Wu CC, Yeh DC, Wei YH. 2004. Somaticmutations in the D-loop and decrease in the copy number of mi-tochondrial DNA in human hepatocellular carcinoma. Mutat Res547:71–87.

Lee HC, Wei YH. 2001. Mitochondrial alterations, cellular responseto oxidative stress and defective degradation of proteins in aging.Biogerontology 2:231–244.

Lee HC, Yin PH, Lin JC, Wu CC, Chen CY, Wu CW, Chi CW,Tam TN, Wei YH. 2005. Mitochondrial genome instability andmtDNA depletion in human cancers. Ann NY Acad Sci 1042:109–122.

Lee HC, Yin PH, Yu TN, Chang YD, Hsu WC, Kao SY, Chi CW,Liu TY, Wei YH. 2001. Accumulation of mitochondrial DNAdeletions in human oral tissues—Effects of betel quid chewingand oral cancer. Mutat Res 493:67–74.

Mambo E, Chatterjee A, Xing M, Tallini G, Haugen BR, Yeung SC,Sukumar S, Sidransky D. 2005. Tumor-specific changes inmtDNA content in human cancer. Int J Cancer 116:920–924.

Mambo E, Gao X, Cohen Y, Guo Z, Talalay P, Sidransky D. 2003.Electrophile and oxidant damage of mitochondrial DNA leadingto rapid evolution of homoplasmic mutations. Proc Natl Acad SciUSA 100:1838–1843.

Matsuyama W, Nakagawa M, Wakimoto J, Hirotsu Y, Kawabata M,Osame M. 2003. Mitochondrial DNA mutation correlates withstage progression and prognosis in non-small cell lung cancer.Hum Mutat 21:441–443.

TABLE 4. Cox Proportional Hazards Multivariate Analysisof Disease-Free Survival

Variables HR (95% CI) P-value

Axillary lymph node status/D-loop mutationAxillary lymph node status

Negative 1Positive 2.62 (0.93–7.36) 0.070

D-loop mutationNegative 1Positive 2.88 (1.14–7.29) 0.026

ER/D-loop mutationER

Negative 1Positive 0.54 (0.21–1.38) 0.197

D-loop mutationNegative 1Positive 2.82 (1.10–7.25) 0.031

PR/D-loop mutationPR

Negative 1Positive 0.56 (0.20–1.52) 0.252

D-loop mutationNegative 1Positive 2.83 (1.11–7.33) 0.029

Bold type indicates a significance of P < 0.05. HR, hazard ratio; ER,

estrogen receptor; PR, progesterone receptor.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

637MITOCHONDRIAL DNA ALTERATIONS IN BREAST CANCER

Parrella P, Xiao Y, Fliss M, Sanchez-Cespedes M, Mazzarelli P,Rinaldi M, Nicol T, Gabrielson E, Cuomo C, Cohen D, Pandit S,Spencer M, Rabitti C, Fazio VM, Sidransky D. 2001. Detection ofmitochondrial DNA mutations in primary breast cancer and fine-needle aspirates. Cancer Res 61:7623–7626.

Penta JS, Johnson FM, Wachsman JT, Copeland WC. 2001.Mitochondrial DNA in human malignancy. Mutat Res 488:119–133.

Petros JA, Baumann AK, Ruiz-Pesini E, Amin MB, Sun CQ, Hall J,Lim S, Issa MM, Flanders WD, Hosseini SH, Marshall FF, Wal-lace DC. 2005. mtDNA mutations increase tumorigenicity inprostate cancer. Proc Natl Acad Sci USA 102:719–724.

Pike MC, Spicer DV, Dahmoush L, Press MF. 1993. Estrogens, pro-gestogens, normal breast cell proliferation, and breast cancer risk.Epidemiol Rev 15:17–35.

Polyak K, Li Y, Zhu H, Lengauer C, Willson JK, Markowitz SD,Trush MA, Kinzler KW, Vogelstein B. 1998. Somatic mutations ofthe mitochondrial genome in human colorectal tumours. NatGenet 20:291–293.

Portakal O, Ozkaya O, Erden Inal M, Bozan B, Kosan M, Sayek I.2000. Coenzyme Q10 concentrations and antioxidant status in tis-sues of breast cancer patients. Clin Biochem 33:279–284.

Rosson D, Keshgegian AA. 2004. Frequent mutations in the mito-chondrial control region DNA in breast tissue. Cancer Lett215:89–94.

Sanchez-Cespedes M, Parrella P, Nomoto S, Cohen D, Xiao Y,Esteller M, Jeronimo C, Jordan RC, Nicol T, Koch WM,Schoenberg M, Mazzarelli P, Fazio VM, Sidransky D. 2001. Iden-tification of a mononucleotide repeat as a major target for mito-chondrial DNA alterations in human tumors. Cancer Res 61:7015–7019.

Schoeler S, Szibor R, Gellerich FN, Wartmann T, Mawrin C, Die-tzmann K, Kirches E. 2005. Mitochondrial DNA deletions sensi-tize cells to apoptosis at low heteroplasmy levels. Biochem Bio-phys Res Commun 332:43–49.

Seow A, Duffy SW, McGee MA, Lee J, Lee HP. 1996. Breast cancerin Singapore: Trends in incidence 1968–1992. Int J Epidemiol25:40–45.

Shen YC, Chang CJ, Hsu C, Cheng CC, Chiu CF, Cheng AL. 2005.Significant difference in the trends of female breast cancer inci-dence between Taiwanese and Caucasian Americans: Implica-tions from age-period-cohort analysis. Cancer Epidemiol BiomarkPrev 14:1986–1990.

Simonnet H, Alazard N, Pfeiffer K, Gallou C, Beroud C, Demont J,Bouvier R, Schagger H, Godinot C. 2002. Low mitochondrial re-spiratory chain content correlates with tumor aggressiveness in re-nal cell carcinoma. Carcinogenesis 23:759–768.

Tamori A, Nishiguchi S, Nishikawa M, Kubo S, Koh N, HirohashiK, Shiomi S, Inoue M. 2004. Correlation between clinical charac-teristics and mitochondrial D-loop DNA mutations in hepatocel-lular carcinoma. J Gastroenterol 39:1063–1068.

Tan DJ, Bai RK, Wong LJ. 2002. Comprehensive scanning of so-matic mitochondrial DNA mutations in breast cancer. Cancer Res62:972–976.

Wallace DC. 1999. Mitochondrial diseases in man and mouse. Sci-ence 283:1482–1488.

Wu CW, Yin PH, Hung WY, Li AFY, Li SH, Chi CW, Wei YH, LeeHC. 2005. Mitochondrial DNA mutations and mitochondrialDNA depletion in gastric cancer. Genes Chromosomes Cancer44:19–28.

Yager JD. 2000. Chapter 3: Endogenous estrogens as carcinogensthrough metabolic activation. J Natl Cancer Inst Monogr 27:67–73.

Yin PH, Lee HC, Chau GY, Wu YT, Li SH, Lui WY, Wei YH, LiuTY, Chi CW. 2004. Alteration of the copy number and deletion ofmitochondrial DNA in human hepatocellular carcinoma. Br JCancer 90:2390–2396.

Zhu W, Qin W, Bradley P, Wessel A, Puckett CL, Sauter ER. 2005.Mitochondrial DNA mutations in breast cancer tissue and inmatched nipple aspirate fluid. Carcinogenesis 26:145–152.

Genes, Chromosomes & Cancer DOI 10.1002/gcc

638 TSENG ETAL.