clinicopathological significance of human apurinic/apyrimidinic endonuclease 1 (ape1) expression in...

PRECLINICAL STUDY

Clinicopathological significance of human apurinic/apyrimidinicendonuclease 1 (APE1) expression in oestrogen-receptor-positivebreast cancer

Tarek M. A. Abdel-Fatah • Christina Perry •

Paul Moseley • Kerstie Johnson • Arvind Arora •

Stephen Chan • Ian O Ellis • Srinivasan Madhusudan

Received: 19 December 2013 / Accepted: 20 December 2013 / Published online: 1 January 2014

� Springer Science+Business Media New York 2013

Abstract Oestrogen metabolites can induce oxidative

DNA base damage and generate potentially mutagenic

apurinic sites (AP sites) in the genomic DNA. If unre-

paired, mutagenic AP sites could drive breast cancer

pathogenesis and aggressive phenotypes. Human apu-

rinic/apyrimidinic endonuclease 1 (APE1) is a key DNA

base excision repair (BER) protein and essential for

processing AP sites generated either directly by oestro-

gen metabolites or during BER of oxidative base dam-

age. Our hypothesis is that altered APE1 expression may

be associated with aggressive tumour biology and impact

upon clinical outcomes in breast cancer. In the current

study, we have investigated APE1 protein expression in

a large cohort of breast cancers (n = 1285) and corre-

lated to clinicopathological features and survival out-

comes. Low APE1 protein expression was associated

with high histological grade (p \ 0.000001), high mitotic

index (p \ 0.000001), glandular de-differentiation

(p \ 0.000001), pleomorphism (p = 0.003), absence of

hormonal receptors (ER-/PgR-/AR-) (p \ 0.0001) and

presence of triple negative phenotype (p = 0.001). Low

APE1 protein expression was associated with loss of

BRCA1, low XRCC1, low FEN1, low SMUG1 and low

pol b (ps \ 0.0001). High MIB1 (p = 0.048), bcl-2

negativity (p \ 0.0001) and low TOP2A (p \ 0.0001)

were likely in low APE1 tumours. In the ER-positive

sub-group, specifically, low APE1 remains significantly

associated with high histological grade, high mitotic

index, glandular de-differentiation (ps \ 0.00001) and

poor breast cancer specific survival (p = 0.007). In the

ER-positive cohort that received adjuvant endocrine

therapy, low APE1 protein expression is associated with

poor survival (p = 0.006). In multivariate analysis, low

APE1 remains independently associated with poor sur-

vival in ER-positive tumours (p = 0.048). We conclude

that low APE1 expression may have prognostic and

predictive significance in ER-positive breast cancers.

Keywords DNA base excision repair � APE1 � Breast

cancer � Biomarker

Introduction

There is an increased risk of breast cancer in women

with chronic exposure to high levels of oestrogens.

Although oestrogen–oestrogen-receptor-mediated cell

proliferation is considered a major cause for breast car-

cinogenesis, there is growing evidence to support a non-

oestrogen-receptor mediated mechanism promoting breast

cancer [1]. Oestrogen metabolites such as 2,3-quinone

catechols and 3,4-quinone catechols are known to induce

genomic DNA damage [2–5]. The 3,4-quinones catechols

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10549-013-2820-7) contains supplementarymaterial, which is available to authorized users.

T. M. A. Abdel-Fatah � P. Moseley � K. Johnson � S. Chan �S. Madhusudan

Department of Oncology, Nottingham University Hospitals,

Nottingham NG51PB, UK

C. Perry � A. Arora � S. Madhusudan (&)

Division of Oncology, Academic Unit of Oncology, School of

Medicine, University of Nottingham, Nottingham University

Hospitals, Nottingham NG51PB, UK

e-mail: [email protected]

I. O. Ellis

Division of Oncology, Department of Pathology, School of

Medicine, University of Nottingham, Nottingham University

Hospitals, Nottingham NG51PB, UK

123

Breast Cancer Res Treat (2014) 143:411–421

DOI 10.1007/s10549-013-2820-7

http://dx.doi.org/10.1007/s10549-013-2820-7

can interact with adenine and guanine bases to form

unstable depurinating 4-OH-E2/E1-1-N3 adenine and

4-OHE2/E1-1-N7 adducts. Spontaneous depurination of

adducts can generate potentially mutagenic apurinic sites

(also known as AP sites). Interestingly, breast cancer

patients or those at risk of developing breast cancer have

been found to have higher levels of depurinating oest-

rogen–DNA adducts in their urine compared to those at

normal risk for breast cancer. Moreover, oestrogen

metabolites also induce generation of reactive oxygen

species (ROS) that cause oxidative DNA base damage

[2–5]. Taken together the data suggest an important role

for DNA damage in oestrogen-induced breast

carcinogenesis.

Base excision repair (BER) is required for the accu-

rate removal of damaged DNA bases induced by oxi-

dising and alkylating agents. BER is performed by two

major sub-pathways: the short-patch pathway (SP-BER)

and long-patch pathway (LP-BER) [6, 7]. Although the

sub-pathways differ from each other in the length of the

repair patch and in the subsets of enzymes involved, in

both pathways excision of a damaged base by a DNA

glycosylase enzyme leads to the formation of a poten-

tially mutagenic apurinic/apyrimidinic (AP) site inter-

mediate. Human AP endonuclease 1 (APE1), cleaves the

phosphodiester backbone on the 50 side of the AP site to

allow further processing in BER [6, 7]. The DNA repair

function of APE1 is performed by the conserved C-ter-

minal domain of the protein. APE1 is also intimately

involved in the coordination of BER and interacts with

several factors within BER such as DNA glycosylase,

DNA polymerase beta, poly (ADP-ribose) polymerase 1

(PARP1), flap endonuclease 1 (FEN1), X-ray repair

cross-complementing gene 1 (XRCC1) and DNA ligase

(reviewed in [8–11] ). APE1 deficiency leads to accu-

mulation of non-coding AP sites in the genomic DNA

which promotes mutagenicity in cells [12–14]. Moreover,

APE1 variants with differential repair capacity have also

been reported in humans [15]. Polymorphic variants of

APE1 have been implicated in cancer predisposition in

several tumours [8], including in breast cancer [16]. In

human tumours, several groups, including ours, have

demonstrated that altered sub-cellular localisation of

APE1 may have prognostic and/or predictive significance

in lung, prostate, colorectal, ovarian, gastric, pancreatico-

biliary and head & neck cancer patients (reviewed in [8,

17, 18]). In addition to performing DNA repair, APE1 is

also involved in cytoplasmic redox regulation of tran-

scription factors, reducing an oxidized cysteine residue in

the target protein to activate DNA binding and tran-

scriptional activities [8]. Redox regulation is performed

by the N-terminal region of APE1 [8]. APE1 has also

recently been shown to be essential for acetylation-

mediated gene regulation [16] and RNA quality control

[19].

Our hypothesis is that altered APE1 expression may be

associated with aggressive tumour biology and impact

upon clinical outcomes in oestrogen-/oestrogen-receptor-

driven breast cancer. In the current study, we have inves-

tigated APE1 expression in primary breast cancers. We

demonstrate that loss of APE1 expression in ER-positive

breast cancer is associated with an aggressive phenotype

and poor clinical outcome in patients. The data presented

here support the hypothesis that APE1 loss may select for

aggressive ER-positive breast cancers.

Methods

Patients

The study was performed in a consecutive series of 1650

patients with primary invasive breast carcinoma who were

diagnosed from 1986 to 1999 and entered into the Not-

tingham Tenovus Primary Breast Carcinoma series. Patient

demographics are summarised in supplementary table S1.

This is a well-characterized series of patients with long-

term follow-up that have been investigated in a wide range

of biomarker studies [20, 21]. All patients were treated in a

uniform way in a single institution with standard surgery

(mastectomy or wide local excision) with radiotherapy.

Patients received standard surgery (mastectomy or wide

local excision) with radiotherapy. Prior to 1989, patients

did not receive systemic adjuvant treatment (AT). After

1989, AT was scheduled based on prognostic and predic-

tive factor status, including Nottingham Prognostic Index

(NPI), oestrogen-receptor-a (ER-a) status, and menopausal

status. Patients with NPI scores of \3.4 (low risk) did not

receive AT. In pre-menopausal patients with NPI scores of

C3.4 (high risk), classical cyclophosphamide, methotrex-

ate, and 5-flurouracil (CMF) chemotherapy was given;

patients with ER-a positive tumours were also offered HT.

Postmenopausal patients with NPI scores of C3.4 and ER-apositivity were offered HT, while ER-a negative patients

received classical CMF chemotherapy.

Median follow up was 111 months (range

1–233 months). Survival data, including overall survival,

disease-free survival (DFS), and development of loco-

regional and distant metastases (DM), was maintained on a

prospective basis. DFS was defined as the number of

months from diagnosis to the occurrence of local recur-

rence, local lymph node (LN) relapse or DM relapse.

Breast cancer-specific survival (BCSS) was defined as the

number of months from diagnosis to the occurrence of BC

related-death. Local recurrence-free survival (LRS) was

defined the number of months from diagnosis to the

412 Breast Cancer Res Treat (2014) 143:411–421

123

occurrence of local recurrence. DM-free-survival was

defined as the number of months from diagnosis to the

occurrence of DM relapse. Survival was censored if the

patient was still alive at the time of analysis, lost to follow-

up, or died from other causes.

Tumour Marker Prognostic Studies (REMARK) criteria,

recommended by McShane et al. [22], were followed

throughout this study. Various clinicopathologic variables

were recorded for the cohort; including patient age at

diagnosis, tumour size, grade and stage and are displayed

in supplementary table 1. Median follow up was

111 months (range 1–233 months). Survival was censored

if the patient was still alive at the time of analysis, lost to

follow-up, or died from other causes. Ethical approval was

obtained from the Nottingham Research Ethics Committee

(C202313).

Construction of tissue microarray (TMA)

TMAs were constructed. Area-specialised histopathologists

identified and marked formalin-fixed paraffin-embedded

tissue blocks containing tumour tissue on haematoxylin

and eosin-stained slides. Two replicate 0.6-mm cores from

the centre and periphery of the tumours were taken and

arrayed into a recipient paraffin block using a tissue

puncher/arrayer (Beecher Instruments, Silver Spring, MD,

USA). Four micron sections of the tissue array block were

cut and placed on Superfrost Plus slides for immunohis-

tochemical staining.

Immunohistochemistry (IHC)

The TMAs were immunohistochemically profiled for

APE1 and other biological antibodies (Supplementary

Table S2) as previously described [23]. A standard strep-

tavidin–biotin complex method was used to achieve tissue

staining. Negative controls were obtained by omitting the

primary antibody in each case. The tissue slides were de-

paraffinised with xylene and then rehydrated through five

decreasing concentrations of alcohol (100, 90, 70, 50 and

30 %) for 2 min each. Endogenous peroxidise activity was

blocked by incubation in a 1 % hydrogen peroxide/meth-

anol buffer. Antigen retrieval was carried out by micro-

wave treatment of the slides in sodium citrate buffer (pH

6.0) for 10 min. The slides were rinsed in phosphate buffer

solution (PBS) and incubated with blocking serum diluted

in PBS to block non-specific staining. The slides were

incubated for 1 h with the primary rabbit polyclonal anti-

APE-1 antibody clone NB100-101 (Novus Biologicals Inc,

Littleton, CO) at a dilution of 1:500. After washing with

PBS, sections were incubated with the secondary antibody

(Vector Labs, Burilingame, CA, USA) for 30 min followed

by the avidin–biotin complex for a further 30 min. 3-30-

Diaminobenzidine tetrahydochloride was used as a chro-

mogen. All sections were counterstained with Gill’s

haematoxylin.

To validate the use of TMAs for immunophenotyping,

full-face sections of 40 cases were stained and protein

expression levels of APE1 antibodies were compared. The

concordance between TMAs and full-face sections was

excellent (k = 0.8). Positive and negative (by omission of

the primary antibody and IgG-matched serum) controls

were included in each run.

Evaluation of immunohistochemical staining

The tumour cores were evaluated by specialist pathologists

and oncologists blinded to the clinico-pathological char-

acteristics of patients. Whole field inspection of the core

was scored and intensities of nuclear staining were grouped

as follows: 0 = no staining, 1 = weak staining,

2 = moderate staining, 3 = strong staining. The percent-

age of each category was estimated (0–100 %). H score

(range 0–300) was calculated by multiplying intensity of

staining and percentage staining. H score in a range of

0–300 was generated. The median H score of 100 was

taken as the cut-off and low APE1 expression was classed

as H score of B100 and[100 was classed as high for APE1

expression. Not all cores within the TMA were suitable for

IHC analysis as some cores were missing or lacked tumour.

HER2 expression was assessed according to the new

ASCO/CAP guidelines using IHC and fluorescence in situ

hybridisation (FISH) [24].

Breast cancer cell lines and Western blot analysis

MCF-7 and MDA-MB-231 were purchased from ATCC

and grown in RPMI medium supplemented with 10 %

FBS, 1 % penicillin/streptomycin. We generated a stable

APE1 knock down (KD) MDA-MB-231 breast cancer cell

line using shRNAs as described previously [25]. MCF-7,

MDA-MB-231 and APE1 KD MDA-MB-231 cells were

investigated for APE1 expression as well as for the spec-

ificity of the APE1 antibody used in the current study.

Statistical analyses

Data analysis was performed using SPSS (SPSS, version 17

Chicago, IL, USA). Where appropriate, Pearson’s Chi

square, Fisher’s exact, v2 for trend, Student’s t and ANO-

VAs one way tests were performed using SPSS software

(SPSS, version 16, Chicago, IL, USA). Cumulative sur-

vival probabilities were estimated using the Kaplan–Meier

method. Differences between survival rates were tested for

significance using the log-rank test. Multivariate analysis

for survival was performed using the Cox hazard model.

Breast Cancer Res Treat (2014) 143:411–421 413

123

The proportional hazards assumption was tested using

standard log–log plots. Each variable was assessed in

univariate analysis as a continuous and categorical variable

and the two models were compared using an appropriate

likelihood ratio test. Hazard ratios (HR) and 95 % confi-

dence intervals (95 % CI) were estimated for each variable.

All tests were two-sided with a 95 % CI. p values for each

test were adjusted with Benjamini and Hochberg multiple

p value adjustment and an adjusted p value of \0.05 was

considered significant.

Results

APE1 expression in breast cancer (whole cohort)

We first investigated specificity of the antibody used in the

current study in a panel of breast cancer cell lines. MCF-7

and MDA-MB-231 breast cancer cells have robust APE1

expression. APE1 KD MDA-MB-231 has more than 95 %

loss of APE1 protein confirming the specificity of the

antibody used in the current study (Fig. 1a).

A total of 1285 tumours were suitable for APE1 protein

expression analysis (Table 1). 649/1285 (50.5 %) had low

APE1 expression compared to 636/1285 (49.5 %) tumours

that had high APE1 expression (Fig. 1b, c). Low APE1

expression was associated with higher tumour size

(p = 0.021), higher tumour grade (p \ 0.000001), higher

mitotic index (p \ 0.000001), tumour de-differentiation

(p \ 0.000001), pleomorphism (p = 0.003), tumour type

(p = 0.0001) and triple negative disease (p = 0.001)

(Table 1). ER-/PR-/AR- tumours were more likely to

have low APE1 expression (p \ 0.0001). Low APE1

expression was also associated with cytokeratin 18 (CK18)

(p = 0.000049), cytokeratin 19 (CK19) (p = 0.043), bcl-2

negativity (p \ 0.0001), low TOP2A (p = 0.048) and high

MIB1 expression (p = 0.048). DNA repair protein levels

associated with low APE1 expression included absent

BRCA1 (p \ 0.00001), XRCC1 (p \ 0.000001), FEN1

(p \ 0.0001), SMUG1 (p \ 0.000001) and pol b(p \ 0.000001) (Table 1).

The data suggests that low APE1 may be associated with

aggressive breast cancer. We then proceed to conduct sub-

group analysis in ER-positive as well as in ER-negative

breast cancers.

APE1 expression in ER-positive breast cancer

A total of 942 tumours were suitable for analysis. 47 % (446/

942) of patients with ER-positive breast cancer were clas-

sified as having low APE1 expression. Univariate associa-

tions between APE1 expression and clinicopathological

variables are shown in Table 2. Low APE1 expression was

associated higher tumour grade (p \ 0.0001), higher mitotic

index (p \ 0.0001). Tumour de-differentiation (p \ 0.0001)

and tumour type (p = 0.001). PR-/AR- tumours were

more likely to have low APE1 expression (p = 0.004 and

\0.0001, respectively). Low APE1 expression was also

associated with bcl-2 negativity (p = 0.008) and low

TOP2A (p \ 0.0001). DNA repair protein levels associated

with low APE1 expression included absent BRCA1

(p \ 0.0001), XRCC1 (p \ 0.0001), FEN1 (p \ 0.0001)

and SMUG1 (p \ 0.0001) (Table 2).

Loss of APE1 expression was associated with worse

BCCS in patients with tumours positive for ER

(p = 0.007) (Fig. 2a). On subgroup analysis of patients

who were ER-positive/high NPI ([3.4) and treated with

actinAPE1

APE1negative APE1positive

A CB

Fig. 1 APE1 protein expression in breast cancer. a Western blot of

cell extract from a panel of breast cancer cell lines (MCF-7, MDA-

MB-231 and APE1 knock down (KD) MDA-MB-231 cells. See text

for details). b Microphotograph of APE1 negative breast cancer

tissue. c Microphotograph of APE1 positive breast cancer tissue


123

Table 1 APE1 expression and breast cancer (whole cohort)

Variable Ape1 protein expression p value

Low High

Pathological parameters

Tumour size 0.021

\1 cm 56 (8.6 %) 76 (11.9 %)

[1–2 cm 314 (48.2 %) 332 (52.1 %)

[2–5 cm 260 (39.9 %) 216 (33.9 %)

[5 cm 22 (3.4 %) 13 (2.0 %)

Tumour stage 0.307

1 404 (61.9 %) 395 (61.9 %)

2 182 (27.9 %) 192 (30.1 %)

3 67 (10.3 %) 51 (8.0 %)

Tumour gradea <0.000001

G1 78 (12.0 %) 137 (21.5 %)

G2 193 (29.6 %) 225 (35.3 %)

G3 381 (58.4 %) 275 (43.2 %)

Mitotic index <0.000001

M1 (low;

mitoses \10)

180 (27.7 %) 264 (41.7 %)

M2 (medium;

mitoses 10–18)

124 (19.1 %) 128 (20.2 %)

M3 (high;

mitosis [18)

345 (53.2 %) 241 (38.1 %)

Tubule Formation <0.000001

1 ([75 % of definite

tubule)

25 (3.9 %) 49 (7.7 %)

2 (10–75 % definite

tubule)

172 (26.5 %) 241 (38.1 %)

3 (\10 % definite

tubule)

452 (69.6 %) 343 (54.2 %)

Pleomorphism 0.003

1 (small-regular

uniform)

16 (2.5 %) 18 (2.8 %)

2 (Moderate

variation)

220 (34.0 %) 271 (42.9 %)

3 (Marked variation) 412 (63.6 %) 343 (54.3 %)

Tumour type 0.0001

IDC–NST 358 (63.4 %) 298 (54.8 %)

Tubular carcinoma 87 (15.4 %) 143 (26.3 %)

Medullary carcinoma 22 (3.9 %) 5 (0.9 %)

ILC 48 (8.5 %) 48 (8.8 %)

Others 50 (8.8 %) 50 (9.2 %)

Lymphovascular

invasion

0.861

No 423 (65.9 %) 420 (66.4 %)

Yes 219 (34.1 %) 213 (33.6 %)

Aggressive phenotype

Her2 overexpression 0.287

No 561 (87.9 %) 556 (89.8 %)

Yes 77 (12.1 %) 63 (10.2 %)

Table 1 continued


Low High

Triple negative

phenotype

0.001

No 499 (78.3 %) 531 (85.6 %)

Yes 138 (21.7 %) 89 (14.4 %)

Basal-like phenotype 0.106

No 538 (86.6 %) 544 (89.6 %)

Yes 83 (13.4 %) 63 (10.4 %)

Cytokeratin 6 (CK6) 0.129

Negative 463 (83.1 %) 453 (86.5 %)

Positive 94 (16.9 %) 71 (13.5 %)


Negative 478 (87.1 %) 465 (88.9 %)

Positive 71 (12.9 %) 58 (11.1 %)


Negative 73 (14.1 %) 31 (6.3 %)

Positive 445 (85.9 %) 460 (93.7 %)


Negative 45 (8.1 %) 26 (5.0 %)

Positive 512 (91.9 %) 493 (95.0 %)

Hormone receptors

ER 0.000054

Negative 197 (30.6 %) 130 (20.7 %)

Positive 446 (69.4 %) 497 (79.3 %)

PgR 0.000003

Negative 296 (48.1 %) 205 (34.7 %)

Positive 320 (51.9 %) 385 (65.3 %)

AR <0.000001

Negative 239 (46.1 %) 134 (29.6 %)

Positive 279 (53.9 %) 365 (73.1 %)

DNA repair

ATM 0.152

Absent 215 (55.4 %) 193 (50.3 %)

Normal 173 (44.6 %) 191 (49.7 %)

BRCA1 0.000003

Absent 121 (26.8 %) 65 (14.3 %)

Normal 331 (73.2 %) 390 (85.7 %)

XRCC1 <0.000001

Low 109 (23.2 %) 40 (8.7 %)

High 360 (76.8 %) 422 (91.3 %)

FEN1 0.000004

Low 256 (79.5 %) 281 (65.7 %)

High 92 (20.5 %) 147 (34.3 %)

SMUG1 <0.000001

Low 199 (47.2 %) 121 (29.8 %)

High 223 (52.8 %) 285 (70.2 %)


123

adjuvant tamoxifen, low APE1 expression was associated

with a worse BCCS (p = 0.006) (Fig. 2b). No association

was seen between APE1 expression and BCSS in ER-

positive/high NPI patients not treated with tamoxifen

(p = 0.382) in this historical cohort (Fig. 2c; Table 3). On

multivariate Cox regression analysis APE1 was indepen-

dently predictive of BCSS (p = 0.048), alongside tumour

stage (p \ 0.0001), grade (p \ 0.0001) and size

(p = 0.009) in ER-positive breast cancers (Table 4).

APE1 in ER-negative breast cancer

Three hundred and twenty-six patients from the Notting-

ham Tenovus Primary Breast Carcinoma series were clas-

sified as ER-negative; 60.4 % (197/326) had low APE1

expression. In this ER-negative cohort (Table 3), low

APE1 expression was associated with tumour grade

(p = 0.030) and mitotic index (p = 0.016) and also CK18

(p = 0.022), AR (p = 0.006) and XRCC1 (p = 0.002)

expression. On Kaplan–Meier analysis APE1 expression

level was not associated with BCSS (p = 0.992) (Fig. 2d).

No association were seen between APE1 expression and

BCSS in ER-negative/high NPI patients treated or not

Table 1 continued


Low High

Polb <0.000001

Low 310 (50.8 %) 142 (23.3 %)

High 300 (49.2 %) 468 (76.8 %)

Cell cycle/apoptosis regulators

MIB1 0.048

Low 178 (32.2 %) 202 (38.0 %)

High 374 (67.8 %) 330 (62.0 %)

P53 0.067

Low expression 414 (78.1 %) 419 (82.6 %)

High expression 116 (21.9 %) 88 (17.4 %)

Bcl-2 0.000032

Negative 236 (40.8 %) 165 (29.0 %)

Positive 343 (59.2 %) 403 (71.0 %)

TOP2A 0.00034

Low 223 (50.2 %) 174 (38.7 %)

Overexpression 231 (49.8 %) 276 (61.3 %)

Statistically significant p values are shown in bold

BRCA1 breast cancer 1 (early onset), HER2 human epidermal growth

factor receptor 2, ER oestrogen-receptor, PgR progesterone receptor,

CK cytokeratin

Basal-like: ER-, HER2- and positive expression of either CK5/6,

CK14 or EGFR; Triple negative: ER-/PgR-/HER2-a Grade as defined by Nottingham Grading System (NGS)

Table 2 APE1 expression and ER-positive breast cancer

Variable Ape1 protein expression p valve

Low High


Tumour size 0.302

\1 cm 47 (10.5 %) 63 (12.7 %)

[1–2 cm 229 (51.3 %) 266 (53.6 %)

[2–5 cm 157 (35.2 %) 159 (32.1 %)

[5 cm 13 (2.9 %) 8 (1.6 %)

Tumour stage 0.843

1 277 (62.1 %) 307 (61.8 %)

2 131 (29.4 %) 152 (30.6 %)

3 38 (8.5 %) 38 (7.6 %)

Tumour gradea <0.0001

G1 73 (16.4 %) 127 (25.6 %)

G2 176 (39.5 %) 207 (41.7 %)

G3 197 (44.2 %) 162 (32.7 %)

Mitotic index <0.0001

M1 (low; mitoses \10) 170 (38.3 %) 246 (49.7 %)

M2 (medium; mitoses 10–18) 102 (23.0 %) 111 (22.4 %)

M3 (high; mitosis [18) 172 (38.7 %) 138 (27.9 %)

Tubule formation <0.0001

1 ([75 % of definite tubule) 23 (5.2 %) 44 (8.9 %)

2 (10–75 % definite tubule) 142 (32.0 %) 215 (43.4 %)

3 (\10 % definite tubule) 279 (62.8 %) 236 (47.7 %)

Pleomorphism 0.348

1 (small-regular uniform) 14 (3.2 %) 16 (3.2 %)

2 (Moderate variation) 205 (46.3 %) 252 (50.9 %)


Tumour type 0.001

IDC–NST 221 (56.2 %) 204 (47.2 %)


Medullary carcinoma 4 (1.0 %) 0 (0 %)

ILC 43 (10.9 %) 46 (10.6 %)

Others 47 (12.0 %) 44 (10.2 %)

Lymphovascular invasion 0.377

No 286 (65.1 %) 336 (67.9 %)

Yes 153 (34.9 %) 159 (32.1 %)


Her 2 overexpression 0.22

No 408 (93.4 %) 451 (93.4 %)

Yes 29 (6.6 %) 32 (6.6 %)


No 446 (100.0 %) 497 (100 %)

Yes 0 (0 %) 0 (0 %)


Negative 363 (94.3 %) 397 (94.1 %)

Positive 22 (5.7 %) 25 (5.9 %)


Negative 348 (92.1 %) 395 (94.0 %)

Positive 30 (7.9 %) 25 (6.0 %)


Negative 17 (4.7 %) 10 (2.5 %)

Positive 347 (95.3 %) 385 (97.5 %)


123

treated with chemotherapy (p = 0.780 and 0.990, respec-

tively) in this historical cohort (Fig. 2e, f). Also, in a

multivariate model of ER-negative breast cancer patients,

APE1 expression (p = 0.686) was not associated with

BCSS; tumour stage (p \ 0.0001) and grade (p = 0.012)

were the only independent predictors of this outcome

(Table 4).

Discussion

Emerging evidence implicates oestrogen/oestrogen

metabolite-induced DNA base damage as an important

cause for breast carcinogenesis [2–5]. Sub-optimal DNA

base excision repair in cells chronically exposed to oest-

rogen would be expected to lead to accumulation of genetic

mutations that eventually lead to a cancerous phenotype.

Human apurinic/apyrimidinic endonuclease 1 (APE1) is

the key enzyme involved in processing AP sites (apurinic

site) generated during spontaneous depurination of unstable

oestrogen DNA adducts as well as AP sites generated

during repair of ROS induced oxidative base damage by

DNA base excision repair (BER) (reviewed in [8, 10–11]).

Our hypothesis is that in oestrogen/oesteogen receptor-

driven breast cancers, impaired base excision repair and

genomic instability may lead to accumulation of mutagenic

lesions and promote a mutator phenotype with more

malignant clones and aggressive behaviour. As APE1 is a

key player we investigated and have provided the first

evidence that APE1 loss in ER-positive breast cancer is

associated with an aggressive phenotype including high

grade, de-differentiation and mitotic index. Interestingly,

the association between loss of APE1 and loss other DNA

repair factors such as BRCA1, XRCC1 and SMUG1 pro-

vides additional evidence for genomic instability in a

proportion of ER-positive breast cancers. APE1 loss was

associated with poor survival in univariate as well in

multivariate analysis. Interestingly, in ER-positive tumours

that received tamoxifen therapy, low APE1 remains asso-

ciated with poor survival implying that APE1 loss may also

predictive significance. In ER-negative breast cancers,

however, there was no significant association with

aggressive phenotype or survival. The data provides strong

evidence that APE1 loss is an important biomarker in ER-

positive breast cancer. Although our study has limitations

in that it is retrospective, the data presented here is con-

sistent with a previous smaller study of 102 breast carci-

nomas where an inverse correlation between APE1

expression, lymph node status and angiogenesis was

demonstrated [26]. However, the data in breast cancer is in

contrast to that observed in other solid tumours. We and

others have consistently demonstrated that over-expression

of APE1 in tumours such as ovarian, gastric, osteosarcoma,

prostate cancer and colorectal cancer is associated with

adverse prognosis (reviewed in [8, 10–11]). These studies

imply that compared to other solid tumours, the role of

APE1 is distinct in breast cancer and that loss of APE1

Table 2 continued


Low High


Negative 21 (5.5 %) 14 (3.4 %)

Positive 364 (94.5 %) 403 (96.6 %)

Hormone receptors

PgR 0.004

Negative 110 (25.8 %) 83 (17.8 %)

Positive 316 (74.2 %) 384 (82.2 %)

AR <0.0001

Negative 115 (31.6 %) 71 (17.7 %)

Positive 249 (68.4 %) 330 (82.3 %)

DNA repair

ATM 0.107

Absent 137 (52.7 %) 139 (45.9 %)

Normal 123 (47.3 %) 164 (54.1 %)

BRCA1 <0.0001

Absent 66 (20.6 %) 38 (10.5 %)

Normal 254 (79.4 %) 323 (89.5 %)

XRCC1 <0.0001

Low 58 (17.7 %) 23 (6.5 %)

High 269 (82.3 %) 333 (93.5 %)

FEN1 <0.0001

Low 246 (79.4 %) 211 (62.6 %)

High 64 (20.6 %) 126 (37.4 %)

SMUG1 <0.0001

Low 118 (41.7 %) 80 (25.1 %)

High 165 (58.3 %) 239 (74.9 %)


MIB1 0.372

Low 155 (40.3 %) 180 (43.4 %)

High 230 (59.7 %) 235 (56.6 %)

P53 0.925

Low expression 327 (89.1 %) 360 (88.9 %)

High expression 40 (10.9 %) 45 (11.1 %)

Bcl-2 0.008

Negative 101 (25.1 %) 79 (17.6 %)

Positive 302 (74.9 %) 370 (82.4 %)

TOP2A <0.0001

Low 170 (54.3 %) 131 (37.4 %)

Overexpression 143 (45.7 %) 219 (62.6 %)


BRCA1 breast cancer 1 (early onset), HER2 human epidermal growth factor

receptor 2, ER oestrogen-receptor, PgR progesterone receptor, CK cytokeratin

Basal-like: ER-, HER2- and positive expression of either CK5/6, CK14 or

EGFR; Triple negative: ER-/PgR-/HER2-a Grade as defined by Nottingham Grading System (NGS)


123

expression could be an important driver of an aggressive

phenotype.

The mechanism of APE1 dysregulation in breast cancer

is largely unknown. However, emerging evidence suggests

potential links between ER, APE1 and key transcription

factors. TP53 transcription factor has previously been

shown to be negative regulator of APE1 in cell line systems

[27]. However, in the current study we did not observe any

significant associations between APE1 and p53 suggesting

that p53 may not have a significant role in vivo. Sp1

transcription factor has also previously been shown to be

an important transcription factor regulating APE1 expres-

sion in cells [28]. Interestingly, ER/Sp-1 interaction has

also been shown to regulate expression of genes involved

in proliferation and cell cycle progression in breast cancer

cells [29]. Moreover, APE1 was recently shown to be an

essential regulator of oestrogen-receptor activity as well as

oestrogen response gene expression in cells [30]. Taken

together the data suggest a complex network regulating

APE1 expression and ER function in breast cancers.

Whether the DNA repair function or the redox regulatory

function of APE1 is involved in breast cancer pathogenesis

is unknown. However, the clinical data presented in our

study investigating nuclear expression of APE1 does sug-

gest a role for the DNA repair function of APE1 in breast

cancer pathogenesis. We have recently demonstrated that

expression of other BER factors such as XRCC1 is also

frequently down-regulated in breast cancer [23]. Low

expression of XRCC1 was also associated with aggressive

ER-positive breast cancers and poor breast cancer specific

survival in that study [23]. Taken together, the data pro-

vides compelling evidence that BER deficiency could be a

biomarker of aggressive ER-positive breast cancers.

The ability of PARP inhibitors (that block BER) to

induce synthetic lethality in BRCA deficient breast and

ovarian cancer, suggest that this approach may be feasible

in other DNA repair deficient systems [31, 32]. As tumours

deficient in APE1 have impaired BER capacity we specu-

late that such tumours may be sensitive to double strand

break (DSB) repair inhibitors such as those targeting ATM

ER+ (whole cohort)

p=0.007

ER+ , NPI>3.4, tamoxifen

p=0.006

ER+ , NPI>3.4, no tamoxifen

p=0.382

p=0.992

ER - (whole cohort)

p=0.990

ER - , NPI>3.4, no chemotherapy

p=0.780

ER -,NPI>3.4, chemotherapy

A CB

D FE

High

Low

High

High

Low Low

LowLow

Low

HighHigh

High

Fig. 2 APE1 protein expression in breast cancer. a Kaplan–Meier

curves showing breast cancer specific survival (BCSS) in ER?

tumours (whole cohort). b Kaplan–Meier curves showing breast

cancer specific survival (BCSS) in high risk ER? tumours that

received tamoxifen. c Kaplan–Meier curves showing breast cancer

specific survival (BCSS) in high risk ER? tumours that received no

tamoxifen. d Kaplan–Meier curves showing breast cancer specific

survival (BCSS) in ER- tumours (whole cohort). e Kaplan–Meier

curves showing breast cancer specific survival (BCSS) in high risk

ER- tumours that received chemotherapy. f Kaplan–Meier curves

showing breast cancer specific survival (BCSS) in high risk ER-

tumours that received no chemotherapy. (See text for details)


123

Table 3 APE1 expression and ER-negative breast cancer


Low High


Tumour size 0.283

\1 cm 8 (4.1 %) 10 (7.8 %)

[1–2 cm 79 (40.3 %) 59 (45.7 %)

[2–5 cm 100 (51.0 %) 56 (43.3 %)

[5 cm 9 (4.6 %) 4 (3.1 %)

Tumour stage 0.469

1 117 (59.4 %) 80 (62.0 %)

2 51 (25.9 %) 36 (27.9 %)

3 29 (14.7 %) 13 (10.1 %)

Tumour gradea 0.030

G1 1 (0.5 %) 6 (4.7 %)

G2 14 (7.1 %) 12 (9.3 %)

G3 181 (92.3 %) 111 (86.0 %)

Mitotic index 0.016

M1 (low; mitoses \10) 6 (3.1 %) 14 (10.9 %)

M2 (medium; mitoses 10–18) 20 (10.2 %) 12 (9.4 %)

M3 (high; mitosis [18) 170 (86.7 %) 102 (79.7 %)

Tubule formation 0.067

1 ([75 % of definite tubule) 0 (0 %) 3 (2.3 %)

2 (10–75 % definite tubule) 26 (13.3 %) 21 (16.4 %)

3 (\10 % definite tubule) 170 (86.7 %) 104 (81.3 %)

Pleomorphism 0.063

1 (small-regular uniform) 0 (0 %) 1 (0.8 %)

2 (Moderate variation) 11 (5.6 %) 15 (11.7 %)


Tumour type 0.145

IDC–NST 134 (80.7 %) 91 (85.8 %)


Medullary carcinoma 18 (10.8 %) 5 (4.7 %)

ILC 5 (3.0 %) 2 (1.9 %)

Others 2 (1.2 %) 5 (4.7 %)

Lymphovascular invasion 0.241

No 128 (66.0 %) 75 (59.5 %)

Yes 66 (34.0 %) 51 (40.5 %)


Her 2 overexpression 0.989

No 145 (75.1 %) 94 (75.2 %)

Yes 48 (24.9 %) 31 (24.8 %)

Triple negative phenotype 0.917

No 51 (27.0 %) 32 (26.4 %)

Yes 138 (73.0 %) 89 (73.6 %)


No 90 (52.0 %) 46 (42.2 %)

Yes 83 (48.0 %) 63 (57.8 %)


Negative 97 (58.1 %) 55 (55.6 %)

Positive 70 (41.9 %) 44 (44.4 %)


Negative 128 (76.6 %) 68 (68.0 %)

Positive 39 (23.4 %) 32 (32.0 %)

Table 3 continued


Low High


Negative 55 (35.9 %) 21 (22.1 %)

Positive 98 (64.1 %) 74 (77.9 %)


Negative 24 (14.4 %) 12 (12.1 %)

Positive 143 (85.6 %) 87 (87.9 %)

Hormone receptors

PgR 0.254

Negative 184 (98.9 %) 120 (100 %)

Positive 2 (1.1 %) 0 (0.0 %)

AR 0.006

Negative 123 (80.4 %) 63 (64.9 %)

Positive 30 (19.6 %) 34 (35.1 %)

DNA repair

ATM 0.482

Absent 78 (61.4 %) 53 (66.3 %)

Normal 49 (38.6 %) 27 (33.8 %)

BRCA1 0.052

Absent 56 (41.7 %) 27 (29.0 %)

Normal 77 (58.3 %) 66 (71.0 %)

XRCC1 0.002

Low 49 (36.0 %) 17 (17.3 %)

High 87 (64.0 %) 81 (82.7 %)

FEN1 0.938

Low 103 (78.6 %) 68 (79.1 %)

High 28 (21.4 %) 18 (20.9 %)

SMUG1 0.187

Low 79 (59.8 %) 41 (50.6 %)

High 53 (40.2 %) 40 (49.4 %)


MIB1 0.504

Low 22 (13.6 %) 18 (16.5 %)

High 140 (86.4 %) 91 (83.5 %)

P53 0.507

Low expression 85 (52.8 %) 57 (57.0 %)

High expression 76 (47.2 %) 43 (43.0 %)

Bcl-2 0.554

Negative 133 (78.7 %) 84 (75.7 %)

Positive 36 (21.3 %) 27 (24.3 %)

TOP2A 0.478

Low 60 (41.1 %) 43 (45.7 %)

Overexpression 86 (58.9 %) 51 (54.3 %)


BRCA1 breast cancer 1 (early onset), HER2 human epidermal growth factor

receptor 2, ER oestrogen-receptor, PgR progesterone receptor, CK cytokeratin

Basal-like: ER-, HER2- and positive expression of either CK5/6, CK14 or

EGFR; Triple negative: ER-/PgR-/HER2-a Grade as defined by Nottingham Grading System (NGS)


123

kinase, a key DSB damage signalling factor. Pre-clinically,

we have recently demonstrated that this approach is fea-

sible in APE1 deficient cells. ATM kinase inhibitor was

synthetically lethal in APE1 deficient cells in that study

[25]. Similarly, we have recently shown that XRCC1

deficient breast cancers can also be targeted by synthetic

lethality using DSB repair inhibitors targeting ATM or

DNA-PK [23]. Together the data provides evidence that

BER deficiency could be exploited by a synthetic lethality

strategy targeting DSB repair.

In conclusion, we provide the first evidence that loss of

APE1 is associated with poor prognosis in ER-positive

breast cancer. Our clinical data provides additional support

for the emerging role of DNA repair in oestrogen/oestrogen

metabolite-induced breast cancer pathogenesis.

Conflict of interest The authors declare that they have no conflict

of interest

References

1. Yue W, Yager JD, Wang JP, Jupe ER, Santen RJ (2013) Estrogen

receptor-dependent and independent mechanisms of breast cancer

carcinogenesis. Steroids 78(2):161–170. doi:10.1016/j.steroids.

2012.11.001

2. Cavalieri E, Chakravarti D, Guttenplan J, Hart E, Ingle J, Jan-

kowiak R et al (2006) Catechol estrogen quinones as initiators of

breast and other human cancers: implications for biomarkers of

susceptibility and cancer prevention. Biochim Biophys Acta

1766(1):63–78. doi:10.1016/j.bbcan.2006.03.001

3. Cavalieri EL, Rogan EG (2011) Unbalanced metabolism of

endogenous estrogens in the etiology and prevention of human

cancer. J Steroid Biochem Mol Biol 125(3–5):169–180. doi:10.

1016/j.jsbmb.2011.03.008

4. Gaikwad NW, Yang L, Muti P, Meza JL, Pruthi S, Ingle JN et al (2008)

The molecular etiology of breast cancer: evidence from biomarkers of

risk. Int J Cancer 122(9):1949–1957. doi:10.1002/ijc.23329

5. Lavigne JA, Goodman JE, Fonong T, Odwin S, He P, Roberts DW

et al (2001) The effects of catechol-O-methyltransferase inhibition

on estrogen metabolite and oxidative DNA damage levels in

estradiol-treated MCF-7 cells. Cancer Res 61(20):7488–7494

6. Kim YJ, Wilson DM 3rd (2012) Overview of base excision repair

biochemistry. Curr Mol Pharmacol 5(1):3–13

7. Svilar D, Goellner EM, Almeida KH, Sobol RW (2011) Base

excision repair and lesion-dependent subpathways for repair of

oxidative DNA damage. Antioxid Redox Signal 14(12):

2491–2507

8. Abbotts R, Madhusudan S (2010) Human AP endonuclease 1

(APE1): from mechanistic insights to druggable target in cancer.

Cancer Treat Rev 36(5):425–435. doi:10.1016/j.ctrv.2009.12.006

9. Tell G, Quadrifoglio F, Tiribelli C, Kelley MR (2008) The many

functions of APE1/Ref-1: not only a DNA repair enzyme. Anti-

oxid Redox Signal 11(3):601–620. doi:10.1089/ars.2008.2194

10. Busso CS, Iwakuma T, Izumi T (2009) Ubiquitination of mam-

malian AP endonuclease (APE1) regulated by the p53-MDM2

signaling pathway. Oncogene 28(13):1616–1625

11. Demple B, Sung JS (2005) Molecular and biological roles of

Ape1 protein in mammalian base excision repair. DNA Repair

(Amst) 4(12):1442–1449

12. Wilson DM 3rd, Barsky D (2001) The major human abasic

endonuclease: formation, consequences and repair of abasic

lesions in DNA. Mutat Res 485(4):283–307

13. Tang M, Pham P, Shen X, Taylor JS, O’Donnell M, Woodgate R

et al (2000) Roles of E. coli DNA polymerases IV and V in

lesion-targeted and untargeted SOS mutagenesis. Nature

404(6781):1014–1018

14. Yu SL, Lee SK, Johnson RE, Prakash L, Prakash S (2003) The

stalling of transcription at abasic sites is highly mutagenic. Mol

Cell Biol 23(1):382–388

15. Hadi MZ, Wilson DM 3rd (2000) Second human protein with

homology to the Escherichia coli abasic endonuclease exonu-

clease III. Environ Mol Mutagen 36(4):312–324

16. Bhakat KK, Mantha AK, Mitra S (2009) Transcriptional regula-

tory functions of mammalian AP-endonuclease (APE1/Ref-1), an

essential multifunctional protein. Antioxid Redox Signal 11(3):

621–638. doi:10.1089/ARS.2008.2198

17. Fishel ML, Kelley MR (2007) The DNA base excision repair

protein Ape1/Ref-1 as a therapeutic and chemopreventive target.

Mol Aspects Med 28(3–4):375–395

18. Tell G, Damante G, Caldwell D, Kelley MR (2005) The intra-

cellular localization of APE1/Ref-1: more than a passive phe-

nomenon? Antioxid Redox Signal 7(3–4):367–384

19. Berquist BR, McNeill DR, Wilson DM 3rd (2008) Character-

ization of abasic endonuclease activity of human Ape1 on alter-

native substrates, as well as effects of ATP and sequence context

on AP site incision. J Mol Biol 379(1):17–27

20. Ellis IO, Galea M, Broughton N, Locker A, Blamey RW, Elston

CW (1992) Pathological prognostic factors in breast cancer. II.

Histological type. Relationship with survival in a large study with

long-term follow-up. Histopathology 20(6):479–489

21. Elston CW, Ellis IO (1991) Pathological prognostic factors in

breast cancer. I. The value of histological grade in breast cancer:

experience from a large study with long-term follow-up. Histo-

pathology 19(5):403–410

22. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M,

Clark GM (2005) Reporting recommendations for tumor marker

Table 4 Multivariate analysis of breast cancer specific survival in ER?ve and ER-ve breast cancers

Variables Multivariate Cox regression analysis

ER-positive ER-negative

Exp (B) CI 95 % p value Exp (B) CI 95 % p value

APE1 0.901 0.813–0.999 0.048 1.038 0.865–1.247 0.686

Tumour stage 1.640 1.400–1.922 <0.0001 2.000 1.560–2.565 <0.0001

Tumour grade 1.375 1.187–1.594 <0.0001 2.643 1.239–5.641 0.012

Tumour size 1.119 1.028–1.217 0.009 1.114 0.961–1.291 0.151


123

http://dx.doi.org/10.1016/j.steroids.2012.11.001

http://dx.doi.org/10.1016/j.steroids.2012.11.001

http://dx.doi.org/10.1016/j.bbcan.2006.03.001

http://dx.doi.org/10.1016/j.jsbmb.2011.03.008

http://dx.doi.org/10.1016/j.jsbmb.2011.03.008

http://dx.doi.org/10.1002/ijc.23329

http://dx.doi.org/10.1016/j.ctrv.2009.12.006

http://dx.doi.org/10.1089/ars.2008.2194

http://dx.doi.org/10.1089/ARS.2008.2198

prognostic studies (REMARK). J Natl Cancer Inst 97(16):

1180–1184

23. Sultana R, Abdel-Fatah T, Abbotts R, Hawkes C, Albarakati N,

Seedhouse C et al (2013) Targeting XRCC1 deficiency in breast

cancer for personalized therapy. Cancer Res 73(5):1621–1634.

doi:10.1158/0008-5472.CAN-12-2929

24. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC,

Cote RJ et al (2007) American Society of Clinical Oncology/

College of American Pathologists guideline recommendations for

human epidermal growth factor receptor 2 testing in breast can-

cer. J Clin Oncol 25(1):118–145

25. Sultana R, McNeill DR, Abbotts R, Mohammed MZ, Zdzienicka

MZ, Qutob H et al (2012) Synthetic lethal targeting of DNA

double-strand break repair deficient cells by human apurinic/

apyrimidinic endonuclease inhibitors. Int J Cancer 131(10):2433–

2444. doi:10.1002/ijc.27512

26. Kakolyris S, Kaklamanis L, Engels K, Fox SB, Taylor M,

Hickson ID et al (1998) Human AP endonuclease 1 (HAP1)

protein expression in breast cancer correlates with lymph node

status and angiogenesis. Br J Cancer 77(7):1169–1173

27. Zaky A, Busso C, Izumi T, Chattopadhyay R, Bassiouny A, Mitra

S et al (2008) Regulation of the human AP-endonuclease (APE1/

Ref-1) expression by the tumor suppressor p53 in response to

DNA damage. Nucleic Acids Res 36(5):1555–1566. doi:10.1093/

nar/gkm1173

28. Fung H, Bennett RA, Demple B (2001) Key role of a downstream

specificity protein 1 site in cell cycle-regulated transcription of

the AP endonuclease gene APE1/APEX in NIH3T3 cells. J Biol

Chem 276(45):42011–42017. doi:10.1074/jbc.M106423200

29. Kim K, Barhoumi R, Burghardt R, Safe S (2005) Analysis of

estrogen receptor alpha-Sp1 interactions in breast cancer cells by

fluorescence resonance energy transfer. Mol Endocrinol

19(4):843–854. doi:10.1210/me.2004-0326

30. Curtis CD, Thorngren DL, Ziegler YS, Sarkeshik A, Yates JR,

Nardulli AM (2009) Apurinic/apyrimidinic endonuclease 1 alters

estrogen receptor activity and estrogen-responsive gene expres-

sion. Mol Endocrinol 23(9):1346–1359. doi:10.1210/me.2009-

0093

31. Dedes KJ, Wilkerson PM, Wetterskog D, Weigelt B, Ashworth

A, Reis-Filho JS (2011) Synthetic lethality of PARP inhibition in

cancers lacking BRCA1 and BRCA2 mutations. Cell Cycle

10(8):1192–1199

32. Irshad S, Ashworth A, Tutt A (2011) Therapeutic potential of

PARP inhibitors for metastatic breast cancer. Expert Rev Anti-

cancer Ther 11(8):1243–1251. doi:10.1586/era.11.52


123

http://dx.doi.org/10.1158/0008-5472.CAN-12-2929

http://dx.doi.org/10.1002/ijc.27512

http://dx.doi.org/10.1093/nar/gkm1173

http://dx.doi.org/10.1093/nar/gkm1173

http://dx.doi.org/10.1074/jbc.M106423200

http://dx.doi.org/10.1210/me.2004-0326

http://dx.doi.org/10.1210/me.2009-0093

http://dx.doi.org/10.1210/me.2009-0093

http://dx.doi.org/10.1586/era.11.52

clinicopathological significance of human apurinic/apyrimidinic endonuclease 1 (ape1) expression in...

Documents