Histopathology 2015, 67, 709–718. DOI: 10.1111/his.12701
Clinicopathological significance of ARID1B in breast invasive ductal carcinoma Fei Shao,1,2 Tiantian Guo,1 Pei Jou Chua,1 Lili Tang,2 Aye Aye Thike,3 Puay-Hoon Tan,3 Boon Huat Bay1 & Gyeong Hun Baeg1 1
Department of Anatomy, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, 2Department of Breast Surgery, Xiangya Hospital, Central South University, Changsha, Hunan, China, and 3Department of Pathology, Singapore General Hospital, Singapore
Date of submission 21 November 2014 Accepted for publication 25 March 2015 Published online Article Accepted 28 March 2015
Shao F, Guo T, Chua P J, Tang L, Thike A A, Tan P-H, Bay B H & Baeg G H (2015) Histopathology 67, 709–718. DOI: 10.1111/his.12701
Clinicopathological significance of ARID1B in breast invasive ductal carcinoma Aims: Identification of prognostic and predictive biomarkers for breast cancer is essential to better stratify patients for treatment and evaluate patient outcome. AT-rich interactive domain-containing protein 1B (ARID1B) is implicated in cell proliferation, but its role in tumorigenesis remains unclear. Methods and results: Immunohistochemical analysis of ARID1B expression using breast cancer tissue microarrays containing 156 breast invasive ductal carcinoma patient samples and subsequent statistical data analysis based on ARID1B immunoreactivity score were performed to examine the correlation between clinicopathological parameters in breast cancer and ARID1B expression. In-vitro assays were also performed to study the role of ARID1B in cell cycle progression. Univariate analysis revealed that high
ARID1B expression is correlated closely with histological grade (P = 0.045) and size (P = 0.043) of invasive breast cancer. These findings were confirmed by multivariate analysis. Notably, increased ARID1B expression was frequently detected in the aggressive triple-negative breast cancer subtypes (P = 0.039) and associated with decreased 5-year disease-free survival rate. Lastly, MDA-MB-231 cells with reduced ARID1B activity displayed a delay in G1 to S phase cell cycle transition and consequently showed a decrease in cell proliferation compared with controls (P < 0.001). Conclusions: ARID1B potentially serves as a valuable prognostic and predictive biomarker as well as a therapeutic target in breast cancer.
Keywords: ARID1B, breast cancer, immunohistochemistry, proliferation, tissue microarray
Introduction Breast cancer is the leading cause of cancer death within the female population.1 Patients diagnosed with early-stage breast cancer have an excellent outcome after adjuvant therapy, but those with advanced-stage or metastatic breast cancer have a high risk of mortality.2 A number of clinical prognostic parameters in breast cancer, including tumour Address for correspondence: G H Baeg, Department of Anatomy, Yong Loo Lin School of Medicine, National University of Singapore, 117594 Singapore, Singapore. e-mail: [email protected] © 2015 John Wiley & Sons Ltd.
histological grade, tumour size and oestrogen receptor (ER)/progesterone receptor (PR)/human epidermal growth factor receptor 2 (HER-2) reactivity pattern, has been established and incorporated into clinical practice.3,4 However, there is still a need for identifying additional prognostic and predictive biomarkers of breast cancer to stratify patients more accurately for treatment and evaluate the patient’s overall outcome. SWItch/sucrose non-fermentable (SWI/SNF) complex is a conserved chromatin-remodelling complex which functions to play crucial roles in cell differentiation, proliferation, DNA repair and many other cellular processes in an ATP-dependent manner.5 SWI/
710
F Shao et al.
SNF consists of two major subclasses, BRG1/BRMassociated factor (BAF) and polybromo-associated BRG1/BRM-associated factor (PBAF) complexes. Both complexes contain evolutionarily well-conserved core subunits such as BRG1 or BRM, which is a main catalytic subunit for ATPase activity.6 While AT-rich interactive domain-containing protein 1A (ARID1A) and 1B (ARID1B) are unique to the BAF complex, BAF200 and BAF180 are specific to the PBAF complexes.7,8 ARID1A and ARID1B are more than 60% identical in their amino acid sequences and exist as two mutually exclusive isoforms in BAF complex.8–10 Importantly, recent studies using whole-genome or exome sequencing revealed that Arid1a is mutated in multiple human cancers such as ovarian and breast cancer, suggesting that ARID1A behaves as a tumour suppressor.11,12 Mutant alleles of Arid1b were also identified in subsets of primary cancer cells and human cancers.12–14 Gene mutations are considered a common mode of the inactivation of tumour suppressor genes, but the possible role of ARID1B in tumorigenesis remains largely elusive. Several in-vitro studies revealed that ARID1B promotes cell proliferation and thus has an opposing role in the tumorigenesis of ARID1A.15–17 In support of this, inactivating mutant alleles of Arid1b responsible for the Coffin– Siris syndrome and non-syndromic intellectual disability were shown to cause a cell cycle delay of G1– S transition.16,18–21 In this study, the association of ARID1B expression with clinicopathological parameters was determined by using breast cancer tissue microarrays (TMAs). We also performed in-vitro assays by modulating ARID1B activity to better stratify the findings of TMA study.
Materials and methods CLINICAL MATERIALS
Tissue samples of 156 patients with breast invasive ductal carcinoma were collected between 2004 and 2007 from the Singapore General Hospital. The paraffin-embedded tissue samples were constructed into TMAs with 1-mm cores. Ethics approval was obtained from the Institutional Review Board. The patients’ ages ranged from 35 to 88 years, with a median of 56 years. IMMUNOHISTOCHEMISTRY
Tissue microarray slides were stained using the Leica BOND-MAXTM System (Leica Microsystems GmbH,
Wetzlar, Germany). An anti-ARID1B antibody (clone 2D2; Sigma-Aldrich, St Louis, MO, USA) was used as a primary antibody (1:300). A negative control was included by omitting the primary antibody to ensure the specificity of the staining. Bond Epitope Retrieval Solution 1 [ethylenediamine tetraacetic acid (EDTA)based pH 9.0 solution] was employed to perform antigen retrieval at 98°C for 20 min. Following that, 10 min of peroxidase blocking was conducted. The samples were then incubated with the ARID1B antibody for 30 min. Subsequently, diaminobenzidine (DAB) staining and haematoxylin counterstaining were performed. The percentage of immunopositive staining in malignant epithelial cells was recorded, and the intensity of staining was classified into four groups: 0 (negative), 1+ (weak), 2+ (moderate) and 3+ (strong). An immunoreactivity score (IRS) was computed by summing all multiplications of the different staining intensity groups with the percentage of cells stained. The scores were then validated independently by two researchers, including a pathologist. CELL CULTURE AND SILENCING OF ARID1B
The breast cancer MDA-MB-231 cell line was cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum. The cells were plated on a six-well plate, and then transfected with human ARID1B ON-TAR GET plus SMARTpool siRNA (UGAUCAACAUGGCGG ACAA, CCGAAUUACAAACGCCAUA, UCUCAAAGCA GACGGCAAA, ACGAGCAUCCAGAGAGAAA). Nontargeting siRNA was used as a negative control. QUANTITATIVE REAL-TIME RT–PCR
Total RNA was extracted from MDA-MB-231 cells transfected with siRNA for negative control or Arid1b. SuperScript III first-strand synthesis system (Invitrogen, Carlsbad, CA, USA) was used to generate cDNA from the RNA samples, and quantitative RT–PCR (qRT–PCR) was subsequently performed. The difference in gene expression was calculated based on the 2DDCt method. The specific primer sets used were as follows; ARID1B forward: 50 -CCCAATCAGAAGGGA GATCA-30 , reverse: 50 -GTGTCCAAAGCCCACGTACT30 ; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward: 50 -GAAGGTGAAGGTCGGAGTCAACG-30 , reverse: 50 -TGCCATGGGTGGAATCATATTGG-30 ; cyclin D1 forward: 50 -AACAGAAGTGCGAGGAG 0 0 GAG-3 , reverse: 5 -CTGGCATTTTGGAGAGGAAG-30 ; cyclin-dependent kinase 6 (CDK6) forward: 50 TGCACAGTGTCACGAACAGA-30 , reverse: 50 -ACCTC GGAGAAGCTGAAACA-30 . © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
CELL PROLIFERATION ASSAY
AlamarBlue cell viability reagent (Invitrogen) was employed to determine cell proliferation. Cells transfected with siRNA were reseeded onto a 24-well plate. AlamarBlue solution was added to the cells every 24 h, and the resulting fluorescence intensity was measured. CELL CYCLE ANALYSIS
MDA-MB-231 cells were harvested at 96 h after siRNA transfection and washed subsequently with icecold phosphate-buffered saline (PBS) before fixation in 70% ethanol at 4°C O/N. Fixed cells were stained with the solution containing 20 lg/ml propidium iodide, 200 lg/ml RNase A and 0.1% Triton X-100 for 30 min at room temperature in the dark. The cells were applied to BD LSRFortessa TM cell analyzer (BD Biosciences, Franklin Lakes, NJ, USA) for cell cycle analysis. Data obtained were analysed using the FlowJo software version 10.0 (Tree Star Inc., Ashland, OR, USA). STATISTICAL ANALYSIS
Statistical analysis for immunohistochemistry staining was carried out using the PASW Statistics 18 software. A paired-samples t-test was performed to evaluate the different expression of ARID1B in tumour samples and paired adjacent non-tumour samples. Both univariate and multivariate analyses were then carried out to determine the correlation between ARID1B expression level and clinicopathological features. Specifically, univariate analysis was conducted by Fisher’s exact test while binary logistic regression was used for multivariate analysis. Tumour size, histological grade (incorporating the parameters of tubule formation, nuclear pleomorphism and mitotic index) and associated ductal carcinoma in situ (DCIS) grade and extent were considered the outcome of interest, whereas patients’ age, race and ER/PR/HER-2 status were identified as potential confounders. For each of the main outcome of interests, both IRS and potential confounders were included in the model. Backward stepwise regression was then carried out to filter out the non-significant potential confounders with the largest P-value in each round, until only statistically significant confounders were detected.22 The Kaplan–Meier method was applied to calculate the survival curve, and the difference was analysed by log-rank test. Data analysis for in-vitro experiments was performed with the © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
711
unpaired t-test. Findings were considered statistically significant at P < 0.05.
Results CORRELATION BETWEEN ARID1B EXPRESSION AND CLINICOPATHOLOGICAL PARAMETERS IN BREAST CANCER
Details of clinicopathological information retrieved from the patients’ records are described in Table 1. ARID1B staining was detected exclusively in the nuclei of cells in both normal and cancer breast tissues. Different intensities of ARID1B staining in the nuclei are depicted in Figure 1A–C. No staining was observed in a negative control (Figure 1D). The immunohistochemical expression level of ARID1B was determined by IRS, which takes into account the percentage of immunopositive cells and staining intensity. The median IRS of 125 was set as the cut-off point to stratify ARID1B immunostaining into high and low groups. Notably, ARID1B expression levels were found to be significantly high in tumour-bearing tissues compared with those in adjacent non-tumour tissues (Figure 2, P = 0.0018). To study the correlation between ARID1B expression levels and clinicopathological parameters of breast cancer patients, univariate statistical analysis was carried out. High ARID1B expression was correlated significantly with patients’ age (P = 0.003), tumour histological grade (P = 0.045), tumour size (P = 0.043) and nuclear pleomorphism (P = 0.003) (Table 2). Importantly, these observations were confirmed further by multivariate analysis, indicating that immunohistochemical staining of ARID1B can serve as an independent predictor of tumour grade, tumour size and nuclear pleomorphism in breast cancer (Table 3). It is worth noting that increased ARID1B expression is detected in the ER()/PR()/ HER-2() immunophenotype of breast cancer at a high frequency compared with non-triple negative subtypes (P = 0.039), suggesting that ARID1B may participate in the development of this triple-negative group of aggressive and highly invasive breast cancers (Table 2). The correlation between ARID1B expression levels and 5-year disease-free survival was also investigated. Among all 156 patients, 27 of them had recurrence within 5 years. In the recurrence group, patients with high or low ARID1B expression (as determined by IRS) have median disease-free survivals of 19.53 or 37.00 months, respectively (Figure 3). Consistently, the Kaplan–Meier plots showed that patients with high ARID1B
712
F Shao et al.
Table 1. Clinicopathological population
features
of
the
study
Table 1. (Continued) Clinicopathological parameters
Clinicopathological parameters
n
%
Age (years) 80
51.3
>56
76
48.7
129
82.7
Malay
10
6.4
India
7
4.5
10
6.4
1 and 2
78
50.0
3
68
43.6
Unavailable
10
6.4
1 and 2
75
48.1
3
70
44.9
Unavailable
11
7.1
1 and 2
54
34.6
3
91
58.3
Unavailable
11
7.1
1 and 2
88
56.4
3
57
36.5
Unavailable
11
7.1
None/minimal
89
57.1
Extensive
23
14.7
Unavailable
44
28.2
Low/intermediate
45
28.8
High
45
28.8
Unavailable
66
42.3
≤50
131
84.0
>50
24
15.4
Race
Others
% 1
0.6
Negative
57
36.5
Positive
99
63.5
Negative
79
50.6
Positive
77
49.4
Negative
93
59.6
Positive
57
36.5
6
3.8
0
72
46.2
1, 2 and 3
75
48.1
Unavailable
9
5.8
ER and/or PR(+)/HER-2()
71
45.5
ER and/or PR(+)/HER-2(+)
32
20.5
ER()/PR()/HER-2()
23
14.7
ER()/PR()/HER-2(+)
24
15.4
6
3.8
Oestrogen receptor status§
≤56
Chinese
Unavailable
n
Progesterone receptor status§
Histological grade*
Nuclear pleomorphism†
Tubule formation†
Mitotic index†
Associated DCIS extent‡
Associated DCIS grade
Tumour size (mm)
HER-2 status§
Unavailable LN stage¶
Molecular classification
Unavailable
*Histological grade is based on the assessment of nuclear pleomorphism, tubule formation and mitotic index. †Scores 1, 2, 3 for nuclear pleomorphism refer to mild, moderate and nuclear pleomorphism, respectively; scores 1, 2, 3 for tubule formation refer to >75, 10–75 and 125 n (%)
P-value
≤56
55 (68.8)
25 (31.3)
0.003**
>56
34 (44.7)
42 (55.3)
Clinicopathological parameters Age (years)
52 (66.7)
26 (33.3)
3
34 (50.0)
34 (50.0)
0.045*
Nuclear pleomorphism 1 and 2
53 (70.7)
22 (29.3)
3
32 (45.7)
38 (54.3)
0.003**
Tubule formation 1 and 2
33 (61.1)
21 (38.9)
3
52 (57.1)
39 (42.9)
0.728
3
IRS > 125 n (%)
P-value
0
46 (63.9)
26 (36.1)
0.241
1, 2 and 3
40 (53.3)
35 (46.7)
ER and/or PR(+)/ HER-2()
45 (63.4)
26 (36.6)
ER and/or PR(+)/ HER-2(+)
17 (53.1)
15 (46.9)
ER()/PR()/ HER-2()
8 (34.8)
15 (65.2)
ER()/PR()/ HER-2(+)
14 (58.3)
10 (41.7)
LN stage
0.039*
DCIS, Ductal carcinoma in situ; LN, Lymph node. *P < 0.05. **P < 0.01. Table 3. Multivariate analysis
Mitotic index 1 and 2
IRS ≤ 125 n (%)
Clinicopathological parameters
Molecular classification
Histological grade 1 and 2
Table 2. (Continued)
51 (58.0) 34 (59.6)
37 (42.0)
Odds ratio
Standard error
P-value
95% CI
2.484
0.464
0.050*
1.001–6.169
ARID1B
2.171
0.378
0.040*
1.036–4.551
ER
0.182
0.386
0.000***
0.085–0.387
0.865
Parameters
23 (40.4)
Tumour size
Associated DCIS extent None/minimal
52 (58.4)
37 (41.6)
Extensive
13 (56.5)
10 (43.5)
1.000
ARID1B Grade of tumour
Associated DCIS grade Low/intermediate
30 (66.7)
15 (33.3)
High
20 (44.4)
25 (55.6)
0.056
Associated DCIS grade ARID1B
2.533
0.463
0.0458*
1.022–6.279
ER
0.257
0.474
0.006**
0.108–0.696
0.328
0.392
0.005**
0.152–0.708
Tumour size (mm) ≤50
80 (61.1)
51 (38.9)
>50
9 (37.5)
15 (62.5)
0.043*
Tubule formation ER
Oestrogen receptor (ER) status Negative
30 (52.6)
27 (47.4)
Positive
59 (59.6)
40 (40.4)
0.407
Nuclear pleomorphism ARID1B
3.340
0.393
0.002**
1.547–7.209
ER
0.157
0.406
0.000***
0.071–0.348
0.292
0.367
0.001***
0.142–0.599
Progesterone receptor (PR) status Negative
45 (57.0)
34 (43.0)
Positive
44 (57.1)
33 (42.9)
1.000
Mitotic index ER
HER-2 status Negative
53 (57.0)
40 (43.0)
Positive
31 (54.4)
26 (45.6)
0.866
CI, Confidence interval; ARID1B, AT-rich interactive domain-containing protein 1B; ER, Oestrogen receptor; DCIS, Ductal carcinoma in situ. *P < 0.05. **P < 0.01. ***P < 0.001. © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
proliferative function, ARID1B shows a pro-proliferative function.15 For instance, ARID1B was shown to be required in pre-osteoblast cell lines for elevated c-Myc oncoprotein expression, which is observed frequently in a variety of human cancers and is known to be essential in preventing cell cycle arrest in response to growth-inhibitory signals.15,29 Furthermore, mouse embryonic stem cells with bi-allelic inactivation of ARID1B showed a reduced proliferation rate and abnormal cell cycle dynamics,30,31 and Arid1b-deficient human fibroblasts displayed a delayed G1–S phase cell cycle progression.16 Consistent with the possible role of ARID1B in promoting cell proliferation, loss-of-function Arid1b mutations that result in
A 1.5 Relative Arid1b mRNA level
(n = 23) displayed high ARID1B expression. Hence, our study suggests that high ARID1B expression may contribute to the pathogenesis of breast cancer, and serves potentially as a promising prognostic biomarker. The SWI/SNF complex acts to disrupt chromatin structure and generate more accessible nucleosomal sites for DNA binding factors.25,26 Furthermore, the SWI/SNF factors, in particular ARID1A and ARID1B, play essential roles in the cellular resistance to various types of DNA damage.27 Notably, recent studies using integrated cancer-genome sequencing analyses have identified inactivating mutant alleles of Arid1a and/or Arid1b in a wide variety of cancers and a subset of primary cancer cells, suggesting that they may encode products involved in tumorigenesis.12–14 However, functional analyses of those mutant alleles have not yet been carried out to assess whether the mutant alleles can cause malignancies in animal models. Indeed, the exact role of ARID1B in tumorigenesis remains largely elusive. A recent study demonstrated that ARID1B is a specific vulnerability in human cancers harbouring Arid1a mutant alleles, indicating that ARID1B is required for Arid1a mutations to promote tumorigenesis.28 In addition, it was reported that ARID1 family subunits have opposing roles in cell proliferation. While ARID1A has an anti-
715
***
1.0
0.5
si
si
N
AR
eg
at
1.0
ID
ive
1B
0.0
B 6000
0.6
0.4
Fluorescence value
Cum disease-free survival
0.8
ARID1B expression Low expression High expression
0.2
***
siNegative siARID1B
***
4000
*** ***
2000
Low expression-censored High expression-censored
0 0
0.0 0.00
10.00 20.00 30.00 40.00 50.00 Months after initial treatment
4 Time (days)
6
60.00
Figure 3. AT-rich interactive domain-containing protein 1B (ARID1B) expression in relation to the 5-year disease-free survival in 156 breast cancer patients. The Kaplan–Meier plot demonstrates that patients with high ARID1B expression exhibit a significantly worse prognosis than those with low ARID1B expression (log-rank test, P = 0.042). © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
2
Figure 4. Effect of AT-rich interactive domain-containing protein 1B (ARID1B) inhibition on cell proliferation. A, Relative Arid1b mRNA level in MDA-MB-231 breast cancer cells treated with either ARID1B siRNA or control siRNA. Arid1b mRNA level in cells treated with ARID1B siRNA is reduced by more than 75% compared with that in control cells. Values are presented as mean standard error of the mean. ***P < 0.001. B, Inhibition of ARID1B in MDAMB-231 cells significantly decreases cell proliferation. ***P < 0.001.
F Shao et al.
716
siNegative
siARID1B
500
500
400
400 G1
300
Count
Count
A
G2-M
200
G2-M
200
s
100
s
100
0
0 0
B
G1
300
200
50
400 600 800 PE-Texas Red-A
1K
0
200 400 600 800 PE-Texas Red-A
*** ***
siNegative siARID1B
Percentage of cells
40 30 20 10 0 G1 C
S
G2/M
1.5 siNegative
Relative mRNA level
1K
*
*
Cyclin D1
Cdk6
1.0
0.5
0.0
a cell cycle delay are also known to be a predominant cause of neuronal diseases, such as corpus callosum abnormalities and Coffin–Siris syndrome.16,18–21 More importantly, our breast cancer TMA analysis exhibited a strong correlation between high ARID1B expression and clinicopathological parameters such as tumour size and grade in breast cancer. Our in-vitro analyses using MDA-MB-231 cells also showed that inhibition of ARID1B can suppress cell prolifera-
siARID1B
Figure 5. Effect of AT-rich interactive domain-containing protein 1B (ARID1B) inhibition on cell cycle. A, B, Suppression of ARID1B activity in MDAMB-231 cells causes a delay in G1–S phase cell cycle transition. Significantly more G1 but fewer S phase cells are observed in MDA-MB-231 cells transfected with Arid1b siRNA compared with control. ***P < 0.001. C, Disruption of ARID1B activity decreases the mRNA levels of cyclin D1 and cdk6 by approximately 30% and 15%, respectively, compared with control. *P < 0.05. Values are presented as mean standard error of the mean.
tion through decreasing the expression of cyclin D1/ CDK6, subsequently delaying G1–S phase cell cycle progression. All these observations suggest the positive role of ARID1B in cell proliferation. By contrast, it was reported that Arid1b is repressed transcriptionally by promoter hypermethylation in the pancreatic cancer cell line MiaPaCa2 and overexpression of ARID1B significantly inhibits the colony forming activity of the cells in liquid culture and soft agar,32 © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
suggesting that ARID1B acts as a tumour suppressor. Therefore, we still cannot exclude the possibility that ARID1B has dual roles in cell proliferation and acts in a tissue- or cell-type specific manner. The effects of Arid1b mutant alleles on tumorigenesis and tissuespecific ARID1B function in cell growth in both cellular and animal models remain to be addressed. None the less, we showed clearly that high ARID1B expression is associated closely with clinicopathological parameters in breast cancer. Furthermore, we demonstrated that ARID1B expression level is correlated with disease-free survival rate in breast cancer patients. The positive role of ARID1B in tumorigenesis was supported further by the observation that ARID1B is expressed highly in aggressive/invasive triple-negative breast cancer subtypes. Taken together, our study suggests that ARID1B may serve as a prognostic and predictive biomarker of breast cancer, and as a useful therapeutic target to treat breast cancer.
Acknowledgements This work was supported by the NUS start-up grant no. R-181-000-142-133.
References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J. Clin. 2011; 61; 69–90. 2. Carlson RW, Allred DC, Anderson BO et al. Metastatic breast cancer, version 1.2012: featured updates to the nccn guidelines. J. Natl. Comp. Canc. Netw. 2012; 10; 821–829. 3. Bundred NJ. Prognostic and predictive factors in breast cancer. Cancer Treat. Rev. 2001; 27; 137–142. 4. Yang XR, Chang-Claude J, Goode EL et al. Associations of breast cancer risk factors with tumor subtypes: a pooled analysis from the breast cancer association consortium studies. J. Natl Cancer Inst. 2011; 103; 250–263. 5. Reisman D, Glaros S, Thompson EA. The SWI/SNF complex and cancer. Oncogene 2009; 28; 1653–1668. 6. Phelan ML, Sif S, Narlikar GJ, Kingston RE. Reconstitution of a core chromatin remodeling complex from SWI/SNF subunits. Mol. Cell 1999; 3; 247–253. 7. Yan Z, Cui K, Murray DM et al. Pbaf chromatin-remodeling complex requires a novel specificity subunit, BAF200, to regulate expression of selective interferon-responsive genes. Genes Dev. 2005; 19; 1662–1667. 8. Wang W, Xue Y, Zhou S, Kuo A, Cairns BR, Crabtree GR. Diversity and specialization of mammalian SWI/SNF complexes. Genes Dev. 1996; 10; 2117–2130. 9. Inoue H, Furukawa T, Giannakopoulos S, Zhou S, King DS, Tanese N. Largest subunits of the human SWI/SNF chromatinremodeling complex promote transcriptional activation by steroid hormone receptors. J. Biol. Chem. 2002; 277; 41674–41685. 10. Wang X, Nagl NG, Wilsker D et al. Two related arid family proteins are alternative subunits of human SWI/SNF complexes. Biochem. J. 2004; 383; 319–325. © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
717
11. Jones S, Wang TL, Shih Ie M et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science 2010; 330; 228–231. 12. Wu JN, Roberts CW. ARID1A mutations in cancer: another epigenetic tumor suppressor? Cancer Discov. 2013; 3; 35–43. 13. Sausen M, Leary RJ, Jones S et al. Integrated genomic analyses identify ARID1A and ARID1B alterations in the childhood cancer neuroblastoma. Nat. Genet. 2013; 45; 12–17. 14. Stephens PJ, Tarpey PS, Davies H et al. The landscape of cancer genes and mutational processes in breast cancer. Nature 2012; 486; 400–404. 15. Nagl NG Jr, Wang X, Patsialou A, Van Scoy M, Moran E. Distinct mammalian SWI/SNF chromatin remodeling complexes with opposing roles in cell-cycle control. EMBO J. 2007; 26; 752–763. 16. Sim JC, White SM, Fitzpatrick E et al. Expanding the phenotypic spectrum of ARID1B-mediated disorders and identification of altered cell-cycle dynamics due to ARID1B haploinsufficiency. Orphanet J. Rare Dis. 2014; 9; 43. 17. Bentley AM, Normand G, Hoyt J, King RW. Distinct sequence elements of cyclin B1 promote localization to chromatin, centrosomes, and kinetochores during mitosis. Mol. Biol. Cell 2007; 18; 4847–4858. 18. Halgren C, Kjaergaard S, Bak M et al. Corpus callosum abnormalities, intellectual disability, speech impairment, and autism in patients with haploinsufficiency of ARID1B. Clin. Genet. 2012; 82; 248–255. 19. Hoyer J, Ekici AB, Endele S et al. Haploinsufficiency of ARID1B, a member of the SWI/SNF-A chromatin-remodeling complex, is a frequent cause of intellectual disability. Am. J. Hum. Genet. 2012; 90; 565–572. 20. Santen GW, Aten E, Sun Y et al. Mutations in SWI/SNF chromatin remodeling complex gene ARID1B cause Coffin-Siris syndrome. Nat. Genet. 2012; 44; 379–380. 21. Tsurusaki Y, Okamoto N, Ohashi H et al. Mutations affecting components of the SWI/SNF complex cause Coffin-Siris syndrome. Nat. Genet. 2012; 44; 376–378. 22. Lwin ZM, Guo C, Salim A et al. Clinicopathological significance of calreticulin in breast invasive ductal carcinoma. Mod. Pathol. 2010; 23; 1559–1566. 23. Xue C, Wang X, Peng R et al. Distribution, clinicopathologic features and survival of breast cancer subtypes in southern China. Cancer Sci. 2012; 103; 1679–1687. 24. Sullivan HC, Oprea-Ilies G, Adams AL et al. Triple-negative breast carcinoma in African American and caucasian women: clinicopathology, immunomarkers, and outcome. Appl. Immunohistochem. Mol. Morphol. 2014; 22; 17–23. 25. Vignali M, Hassan AH, Neely KE, Workman JL. ATP-dependent chromatin-remodeling complexes. Mol. Cell. Biol. 2000; 20; 1899–1910. 26. Mohrmann L, Verrijzer CP. Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim. Biophys. Acta 2005; 1681; 59–73. 27. Watanabe R, Ui A, Kanno S et al. SWI/SNF factors required for cellular resistance to DNA damage include ARID1A and ARID1B and show interdependent protein stability. Cancer Res. 2014; 74; 2465–2475. 28. Helming KC, Wang X, Wilson BG et al. ARID1B is a specific vulnerability in ARID1A-mutant cancers. Nat. Med. 2014; 20; 251–254. 29. Inghirami G, Chiarle R, Simmons WJ, Piva R, Schlessinger K, Levy DE. New and old functions of STAT3: a pivotal target for individualized treatment of cancer. Cell Cycle 2005; 4; 1131–1133. 30. Flores-Alcantar A, Gonzalez-Sandoval A, Escalante-Alcalde D, Lomeli H. Dynamics of expression of arid1a and arid1b subun-
718
F Shao et al.
its in mouse embryos and in cells during the cell cycle. Cell Tissue Res. 2011; 345; 137–148. 31. Yan Z, Wang Z, Sharova L et al. BAF250B-associated SWI/SNF chromatin-remodeling complex is required to maintain undifferentiated mouse embryonic stem cells. Stem Cells 2008; 26; 1155–1165. 32. Khursheed M, Kolla JN, Kotapalli V et al. ARID1B, a member of the human SWI/SNF chromatin remodeling complex, exhibits tumour-suppressor activities in pancreatic cancer cell lines. Br. J. Cancer 2013; 108; 2056–2062.
Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Details of the oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER-2) antibodies and the cut-off value.
© 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
Clinicopathological significance of ARID1B in breast invasive ductal carcinoma Fei Shao,1,2 Tiantian Guo,1 Pei Jou Chua,1 Lili Tang,2 Aye Aye Thike,3 Puay-Hoon Tan,3 Boon Huat Bay1 & Gyeong Hun Baeg1 1
Department of Anatomy, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, 2Department of Breast Surgery, Xiangya Hospital, Central South University, Changsha, Hunan, China, and 3Department of Pathology, Singapore General Hospital, Singapore
Date of submission 21 November 2014 Accepted for publication 25 March 2015 Published online Article Accepted 28 March 2015
Shao F, Guo T, Chua P J, Tang L, Thike A A, Tan P-H, Bay B H & Baeg G H (2015) Histopathology 67, 709–718. DOI: 10.1111/his.12701
Clinicopathological significance of ARID1B in breast invasive ductal carcinoma Aims: Identification of prognostic and predictive biomarkers for breast cancer is essential to better stratify patients for treatment and evaluate patient outcome. AT-rich interactive domain-containing protein 1B (ARID1B) is implicated in cell proliferation, but its role in tumorigenesis remains unclear. Methods and results: Immunohistochemical analysis of ARID1B expression using breast cancer tissue microarrays containing 156 breast invasive ductal carcinoma patient samples and subsequent statistical data analysis based on ARID1B immunoreactivity score were performed to examine the correlation between clinicopathological parameters in breast cancer and ARID1B expression. In-vitro assays were also performed to study the role of ARID1B in cell cycle progression. Univariate analysis revealed that high
ARID1B expression is correlated closely with histological grade (P = 0.045) and size (P = 0.043) of invasive breast cancer. These findings were confirmed by multivariate analysis. Notably, increased ARID1B expression was frequently detected in the aggressive triple-negative breast cancer subtypes (P = 0.039) and associated with decreased 5-year disease-free survival rate. Lastly, MDA-MB-231 cells with reduced ARID1B activity displayed a delay in G1 to S phase cell cycle transition and consequently showed a decrease in cell proliferation compared with controls (P < 0.001). Conclusions: ARID1B potentially serves as a valuable prognostic and predictive biomarker as well as a therapeutic target in breast cancer.
Keywords: ARID1B, breast cancer, immunohistochemistry, proliferation, tissue microarray
Introduction Breast cancer is the leading cause of cancer death within the female population.1 Patients diagnosed with early-stage breast cancer have an excellent outcome after adjuvant therapy, but those with advanced-stage or metastatic breast cancer have a high risk of mortality.2 A number of clinical prognostic parameters in breast cancer, including tumour Address for correspondence: G H Baeg, Department of Anatomy, Yong Loo Lin School of Medicine, National University of Singapore, 117594 Singapore, Singapore. e-mail: [email protected] © 2015 John Wiley & Sons Ltd.
histological grade, tumour size and oestrogen receptor (ER)/progesterone receptor (PR)/human epidermal growth factor receptor 2 (HER-2) reactivity pattern, has been established and incorporated into clinical practice.3,4 However, there is still a need for identifying additional prognostic and predictive biomarkers of breast cancer to stratify patients more accurately for treatment and evaluate the patient’s overall outcome. SWItch/sucrose non-fermentable (SWI/SNF) complex is a conserved chromatin-remodelling complex which functions to play crucial roles in cell differentiation, proliferation, DNA repair and many other cellular processes in an ATP-dependent manner.5 SWI/
710
F Shao et al.
SNF consists of two major subclasses, BRG1/BRMassociated factor (BAF) and polybromo-associated BRG1/BRM-associated factor (PBAF) complexes. Both complexes contain evolutionarily well-conserved core subunits such as BRG1 or BRM, which is a main catalytic subunit for ATPase activity.6 While AT-rich interactive domain-containing protein 1A (ARID1A) and 1B (ARID1B) are unique to the BAF complex, BAF200 and BAF180 are specific to the PBAF complexes.7,8 ARID1A and ARID1B are more than 60% identical in their amino acid sequences and exist as two mutually exclusive isoforms in BAF complex.8–10 Importantly, recent studies using whole-genome or exome sequencing revealed that Arid1a is mutated in multiple human cancers such as ovarian and breast cancer, suggesting that ARID1A behaves as a tumour suppressor.11,12 Mutant alleles of Arid1b were also identified in subsets of primary cancer cells and human cancers.12–14 Gene mutations are considered a common mode of the inactivation of tumour suppressor genes, but the possible role of ARID1B in tumorigenesis remains largely elusive. Several in-vitro studies revealed that ARID1B promotes cell proliferation and thus has an opposing role in the tumorigenesis of ARID1A.15–17 In support of this, inactivating mutant alleles of Arid1b responsible for the Coffin– Siris syndrome and non-syndromic intellectual disability were shown to cause a cell cycle delay of G1– S transition.16,18–21 In this study, the association of ARID1B expression with clinicopathological parameters was determined by using breast cancer tissue microarrays (TMAs). We also performed in-vitro assays by modulating ARID1B activity to better stratify the findings of TMA study.
Materials and methods CLINICAL MATERIALS
Tissue samples of 156 patients with breast invasive ductal carcinoma were collected between 2004 and 2007 from the Singapore General Hospital. The paraffin-embedded tissue samples were constructed into TMAs with 1-mm cores. Ethics approval was obtained from the Institutional Review Board. The patients’ ages ranged from 35 to 88 years, with a median of 56 years. IMMUNOHISTOCHEMISTRY
Tissue microarray slides were stained using the Leica BOND-MAXTM System (Leica Microsystems GmbH,
Wetzlar, Germany). An anti-ARID1B antibody (clone 2D2; Sigma-Aldrich, St Louis, MO, USA) was used as a primary antibody (1:300). A negative control was included by omitting the primary antibody to ensure the specificity of the staining. Bond Epitope Retrieval Solution 1 [ethylenediamine tetraacetic acid (EDTA)based pH 9.0 solution] was employed to perform antigen retrieval at 98°C for 20 min. Following that, 10 min of peroxidase blocking was conducted. The samples were then incubated with the ARID1B antibody for 30 min. Subsequently, diaminobenzidine (DAB) staining and haematoxylin counterstaining were performed. The percentage of immunopositive staining in malignant epithelial cells was recorded, and the intensity of staining was classified into four groups: 0 (negative), 1+ (weak), 2+ (moderate) and 3+ (strong). An immunoreactivity score (IRS) was computed by summing all multiplications of the different staining intensity groups with the percentage of cells stained. The scores were then validated independently by two researchers, including a pathologist. CELL CULTURE AND SILENCING OF ARID1B
The breast cancer MDA-MB-231 cell line was cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum. The cells were plated on a six-well plate, and then transfected with human ARID1B ON-TAR GET plus SMARTpool siRNA (UGAUCAACAUGGCGG ACAA, CCGAAUUACAAACGCCAUA, UCUCAAAGCA GACGGCAAA, ACGAGCAUCCAGAGAGAAA). Nontargeting siRNA was used as a negative control. QUANTITATIVE REAL-TIME RT–PCR
Total RNA was extracted from MDA-MB-231 cells transfected with siRNA for negative control or Arid1b. SuperScript III first-strand synthesis system (Invitrogen, Carlsbad, CA, USA) was used to generate cDNA from the RNA samples, and quantitative RT–PCR (qRT–PCR) was subsequently performed. The difference in gene expression was calculated based on the 2DDCt method. The specific primer sets used were as follows; ARID1B forward: 50 -CCCAATCAGAAGGGA GATCA-30 , reverse: 50 -GTGTCCAAAGCCCACGTACT30 ; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward: 50 -GAAGGTGAAGGTCGGAGTCAACG-30 , reverse: 50 -TGCCATGGGTGGAATCATATTGG-30 ; cyclin D1 forward: 50 -AACAGAAGTGCGAGGAG 0 0 GAG-3 , reverse: 5 -CTGGCATTTTGGAGAGGAAG-30 ; cyclin-dependent kinase 6 (CDK6) forward: 50 TGCACAGTGTCACGAACAGA-30 , reverse: 50 -ACCTC GGAGAAGCTGAAACA-30 . © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
CELL PROLIFERATION ASSAY
AlamarBlue cell viability reagent (Invitrogen) was employed to determine cell proliferation. Cells transfected with siRNA were reseeded onto a 24-well plate. AlamarBlue solution was added to the cells every 24 h, and the resulting fluorescence intensity was measured. CELL CYCLE ANALYSIS
MDA-MB-231 cells were harvested at 96 h after siRNA transfection and washed subsequently with icecold phosphate-buffered saline (PBS) before fixation in 70% ethanol at 4°C O/N. Fixed cells were stained with the solution containing 20 lg/ml propidium iodide, 200 lg/ml RNase A and 0.1% Triton X-100 for 30 min at room temperature in the dark. The cells were applied to BD LSRFortessa TM cell analyzer (BD Biosciences, Franklin Lakes, NJ, USA) for cell cycle analysis. Data obtained were analysed using the FlowJo software version 10.0 (Tree Star Inc., Ashland, OR, USA). STATISTICAL ANALYSIS
Statistical analysis for immunohistochemistry staining was carried out using the PASW Statistics 18 software. A paired-samples t-test was performed to evaluate the different expression of ARID1B in tumour samples and paired adjacent non-tumour samples. Both univariate and multivariate analyses were then carried out to determine the correlation between ARID1B expression level and clinicopathological features. Specifically, univariate analysis was conducted by Fisher’s exact test while binary logistic regression was used for multivariate analysis. Tumour size, histological grade (incorporating the parameters of tubule formation, nuclear pleomorphism and mitotic index) and associated ductal carcinoma in situ (DCIS) grade and extent were considered the outcome of interest, whereas patients’ age, race and ER/PR/HER-2 status were identified as potential confounders. For each of the main outcome of interests, both IRS and potential confounders were included in the model. Backward stepwise regression was then carried out to filter out the non-significant potential confounders with the largest P-value in each round, until only statistically significant confounders were detected.22 The Kaplan–Meier method was applied to calculate the survival curve, and the difference was analysed by log-rank test. Data analysis for in-vitro experiments was performed with the © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
711
unpaired t-test. Findings were considered statistically significant at P < 0.05.
Results CORRELATION BETWEEN ARID1B EXPRESSION AND CLINICOPATHOLOGICAL PARAMETERS IN BREAST CANCER
Details of clinicopathological information retrieved from the patients’ records are described in Table 1. ARID1B staining was detected exclusively in the nuclei of cells in both normal and cancer breast tissues. Different intensities of ARID1B staining in the nuclei are depicted in Figure 1A–C. No staining was observed in a negative control (Figure 1D). The immunohistochemical expression level of ARID1B was determined by IRS, which takes into account the percentage of immunopositive cells and staining intensity. The median IRS of 125 was set as the cut-off point to stratify ARID1B immunostaining into high and low groups. Notably, ARID1B expression levels were found to be significantly high in tumour-bearing tissues compared with those in adjacent non-tumour tissues (Figure 2, P = 0.0018). To study the correlation between ARID1B expression levels and clinicopathological parameters of breast cancer patients, univariate statistical analysis was carried out. High ARID1B expression was correlated significantly with patients’ age (P = 0.003), tumour histological grade (P = 0.045), tumour size (P = 0.043) and nuclear pleomorphism (P = 0.003) (Table 2). Importantly, these observations were confirmed further by multivariate analysis, indicating that immunohistochemical staining of ARID1B can serve as an independent predictor of tumour grade, tumour size and nuclear pleomorphism in breast cancer (Table 3). It is worth noting that increased ARID1B expression is detected in the ER()/PR()/ HER-2() immunophenotype of breast cancer at a high frequency compared with non-triple negative subtypes (P = 0.039), suggesting that ARID1B may participate in the development of this triple-negative group of aggressive and highly invasive breast cancers (Table 2). The correlation between ARID1B expression levels and 5-year disease-free survival was also investigated. Among all 156 patients, 27 of them had recurrence within 5 years. In the recurrence group, patients with high or low ARID1B expression (as determined by IRS) have median disease-free survivals of 19.53 or 37.00 months, respectively (Figure 3). Consistently, the Kaplan–Meier plots showed that patients with high ARID1B
712
F Shao et al.
Table 1. Clinicopathological population
features
of
the
study
Table 1. (Continued) Clinicopathological parameters
Clinicopathological parameters
n
%
Age (years) 80
51.3
>56
76
48.7
129
82.7
Malay
10
6.4
India
7
4.5
10
6.4
1 and 2
78
50.0
3
68
43.6
Unavailable
10
6.4
1 and 2
75
48.1
3
70
44.9
Unavailable
11
7.1
1 and 2
54
34.6
3
91
58.3
Unavailable
11
7.1
1 and 2
88
56.4
3
57
36.5
Unavailable
11
7.1
None/minimal
89
57.1
Extensive
23
14.7
Unavailable
44
28.2
Low/intermediate
45
28.8
High
45
28.8
Unavailable
66
42.3
≤50
131
84.0
>50
24
15.4
Race
Others
% 1
0.6
Negative
57
36.5
Positive
99
63.5
Negative
79
50.6
Positive
77
49.4
Negative
93
59.6
Positive
57
36.5
6
3.8
0
72
46.2
1, 2 and 3
75
48.1
Unavailable
9
5.8
ER and/or PR(+)/HER-2()
71
45.5
ER and/or PR(+)/HER-2(+)
32
20.5
ER()/PR()/HER-2()
23
14.7
ER()/PR()/HER-2(+)
24
15.4
6
3.8
Oestrogen receptor status§
≤56
Chinese
Unavailable
n
Progesterone receptor status§
Histological grade*
Nuclear pleomorphism†
Tubule formation†
Mitotic index†
Associated DCIS extent‡
Associated DCIS grade
Tumour size (mm)
HER-2 status§
Unavailable LN stage¶
Molecular classification
Unavailable
*Histological grade is based on the assessment of nuclear pleomorphism, tubule formation and mitotic index. †Scores 1, 2, 3 for nuclear pleomorphism refer to mild, moderate and nuclear pleomorphism, respectively; scores 1, 2, 3 for tubule formation refer to >75, 10–75 and 125 n (%)
P-value
≤56
55 (68.8)
25 (31.3)
0.003**
>56
34 (44.7)
42 (55.3)
Clinicopathological parameters Age (years)
52 (66.7)
26 (33.3)
3
34 (50.0)
34 (50.0)
0.045*
Nuclear pleomorphism 1 and 2
53 (70.7)
22 (29.3)
3
32 (45.7)
38 (54.3)
0.003**
Tubule formation 1 and 2
33 (61.1)
21 (38.9)
3
52 (57.1)
39 (42.9)
0.728
3
IRS > 125 n (%)
P-value
0
46 (63.9)
26 (36.1)
0.241
1, 2 and 3
40 (53.3)
35 (46.7)
ER and/or PR(+)/ HER-2()
45 (63.4)
26 (36.6)
ER and/or PR(+)/ HER-2(+)
17 (53.1)
15 (46.9)
ER()/PR()/ HER-2()
8 (34.8)
15 (65.2)
ER()/PR()/ HER-2(+)
14 (58.3)
10 (41.7)
LN stage
0.039*
DCIS, Ductal carcinoma in situ; LN, Lymph node. *P < 0.05. **P < 0.01. Table 3. Multivariate analysis
Mitotic index 1 and 2
IRS ≤ 125 n (%)
Clinicopathological parameters
Molecular classification
Histological grade 1 and 2
Table 2. (Continued)
51 (58.0) 34 (59.6)
37 (42.0)
Odds ratio
Standard error
P-value
95% CI
2.484
0.464
0.050*
1.001–6.169
ARID1B
2.171
0.378
0.040*
1.036–4.551
ER
0.182
0.386
0.000***
0.085–0.387
0.865
Parameters
23 (40.4)
Tumour size
Associated DCIS extent None/minimal
52 (58.4)
37 (41.6)
Extensive
13 (56.5)
10 (43.5)
1.000
ARID1B Grade of tumour
Associated DCIS grade Low/intermediate
30 (66.7)
15 (33.3)
High
20 (44.4)
25 (55.6)
0.056
Associated DCIS grade ARID1B
2.533
0.463
0.0458*
1.022–6.279
ER
0.257
0.474
0.006**
0.108–0.696
0.328
0.392
0.005**
0.152–0.708
Tumour size (mm) ≤50
80 (61.1)
51 (38.9)
>50
9 (37.5)
15 (62.5)
0.043*
Tubule formation ER
Oestrogen receptor (ER) status Negative
30 (52.6)
27 (47.4)
Positive
59 (59.6)
40 (40.4)
0.407
Nuclear pleomorphism ARID1B
3.340
0.393
0.002**
1.547–7.209
ER
0.157
0.406
0.000***
0.071–0.348
0.292
0.367
0.001***
0.142–0.599
Progesterone receptor (PR) status Negative
45 (57.0)
34 (43.0)
Positive
44 (57.1)
33 (42.9)
1.000
Mitotic index ER
HER-2 status Negative
53 (57.0)
40 (43.0)
Positive
31 (54.4)
26 (45.6)
0.866
CI, Confidence interval; ARID1B, AT-rich interactive domain-containing protein 1B; ER, Oestrogen receptor; DCIS, Ductal carcinoma in situ. *P < 0.05. **P < 0.01. ***P < 0.001. © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
proliferative function, ARID1B shows a pro-proliferative function.15 For instance, ARID1B was shown to be required in pre-osteoblast cell lines for elevated c-Myc oncoprotein expression, which is observed frequently in a variety of human cancers and is known to be essential in preventing cell cycle arrest in response to growth-inhibitory signals.15,29 Furthermore, mouse embryonic stem cells with bi-allelic inactivation of ARID1B showed a reduced proliferation rate and abnormal cell cycle dynamics,30,31 and Arid1b-deficient human fibroblasts displayed a delayed G1–S phase cell cycle progression.16 Consistent with the possible role of ARID1B in promoting cell proliferation, loss-of-function Arid1b mutations that result in
A 1.5 Relative Arid1b mRNA level
(n = 23) displayed high ARID1B expression. Hence, our study suggests that high ARID1B expression may contribute to the pathogenesis of breast cancer, and serves potentially as a promising prognostic biomarker. The SWI/SNF complex acts to disrupt chromatin structure and generate more accessible nucleosomal sites for DNA binding factors.25,26 Furthermore, the SWI/SNF factors, in particular ARID1A and ARID1B, play essential roles in the cellular resistance to various types of DNA damage.27 Notably, recent studies using integrated cancer-genome sequencing analyses have identified inactivating mutant alleles of Arid1a and/or Arid1b in a wide variety of cancers and a subset of primary cancer cells, suggesting that they may encode products involved in tumorigenesis.12–14 However, functional analyses of those mutant alleles have not yet been carried out to assess whether the mutant alleles can cause malignancies in animal models. Indeed, the exact role of ARID1B in tumorigenesis remains largely elusive. A recent study demonstrated that ARID1B is a specific vulnerability in human cancers harbouring Arid1a mutant alleles, indicating that ARID1B is required for Arid1a mutations to promote tumorigenesis.28 In addition, it was reported that ARID1 family subunits have opposing roles in cell proliferation. While ARID1A has an anti-
715
***
1.0
0.5
si
si
N
AR
eg
at
1.0
ID
ive
1B
0.0
B 6000
0.6
0.4
Fluorescence value
Cum disease-free survival
0.8
ARID1B expression Low expression High expression
0.2
***
siNegative siARID1B
***
4000
*** ***
2000
Low expression-censored High expression-censored
0 0
0.0 0.00
10.00 20.00 30.00 40.00 50.00 Months after initial treatment
4 Time (days)
6
60.00
Figure 3. AT-rich interactive domain-containing protein 1B (ARID1B) expression in relation to the 5-year disease-free survival in 156 breast cancer patients. The Kaplan–Meier plot demonstrates that patients with high ARID1B expression exhibit a significantly worse prognosis than those with low ARID1B expression (log-rank test, P = 0.042). © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
2
Figure 4. Effect of AT-rich interactive domain-containing protein 1B (ARID1B) inhibition on cell proliferation. A, Relative Arid1b mRNA level in MDA-MB-231 breast cancer cells treated with either ARID1B siRNA or control siRNA. Arid1b mRNA level in cells treated with ARID1B siRNA is reduced by more than 75% compared with that in control cells. Values are presented as mean standard error of the mean. ***P < 0.001. B, Inhibition of ARID1B in MDAMB-231 cells significantly decreases cell proliferation. ***P < 0.001.
F Shao et al.
716
siNegative
siARID1B
500
500
400
400 G1
300
Count
Count
A
G2-M
200
G2-M
200
s
100
s
100
0
0 0
B
G1
300
200
50
400 600 800 PE-Texas Red-A
1K
0
200 400 600 800 PE-Texas Red-A
*** ***
siNegative siARID1B
Percentage of cells
40 30 20 10 0 G1 C
S
G2/M
1.5 siNegative
Relative mRNA level
1K
*
*
Cyclin D1
Cdk6
1.0
0.5
0.0
a cell cycle delay are also known to be a predominant cause of neuronal diseases, such as corpus callosum abnormalities and Coffin–Siris syndrome.16,18–21 More importantly, our breast cancer TMA analysis exhibited a strong correlation between high ARID1B expression and clinicopathological parameters such as tumour size and grade in breast cancer. Our in-vitro analyses using MDA-MB-231 cells also showed that inhibition of ARID1B can suppress cell prolifera-
siARID1B
Figure 5. Effect of AT-rich interactive domain-containing protein 1B (ARID1B) inhibition on cell cycle. A, B, Suppression of ARID1B activity in MDAMB-231 cells causes a delay in G1–S phase cell cycle transition. Significantly more G1 but fewer S phase cells are observed in MDA-MB-231 cells transfected with Arid1b siRNA compared with control. ***P < 0.001. C, Disruption of ARID1B activity decreases the mRNA levels of cyclin D1 and cdk6 by approximately 30% and 15%, respectively, compared with control. *P < 0.05. Values are presented as mean standard error of the mean.
tion through decreasing the expression of cyclin D1/ CDK6, subsequently delaying G1–S phase cell cycle progression. All these observations suggest the positive role of ARID1B in cell proliferation. By contrast, it was reported that Arid1b is repressed transcriptionally by promoter hypermethylation in the pancreatic cancer cell line MiaPaCa2 and overexpression of ARID1B significantly inhibits the colony forming activity of the cells in liquid culture and soft agar,32 © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
ARID1B and breast cancer
suggesting that ARID1B acts as a tumour suppressor. Therefore, we still cannot exclude the possibility that ARID1B has dual roles in cell proliferation and acts in a tissue- or cell-type specific manner. The effects of Arid1b mutant alleles on tumorigenesis and tissuespecific ARID1B function in cell growth in both cellular and animal models remain to be addressed. None the less, we showed clearly that high ARID1B expression is associated closely with clinicopathological parameters in breast cancer. Furthermore, we demonstrated that ARID1B expression level is correlated with disease-free survival rate in breast cancer patients. The positive role of ARID1B in tumorigenesis was supported further by the observation that ARID1B is expressed highly in aggressive/invasive triple-negative breast cancer subtypes. Taken together, our study suggests that ARID1B may serve as a prognostic and predictive biomarker of breast cancer, and as a useful therapeutic target to treat breast cancer.
Acknowledgements This work was supported by the NUS start-up grant no. R-181-000-142-133.
References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J. Clin. 2011; 61; 69–90. 2. Carlson RW, Allred DC, Anderson BO et al. Metastatic breast cancer, version 1.2012: featured updates to the nccn guidelines. J. Natl. Comp. Canc. Netw. 2012; 10; 821–829. 3. Bundred NJ. Prognostic and predictive factors in breast cancer. Cancer Treat. Rev. 2001; 27; 137–142. 4. Yang XR, Chang-Claude J, Goode EL et al. Associations of breast cancer risk factors with tumor subtypes: a pooled analysis from the breast cancer association consortium studies. J. Natl Cancer Inst. 2011; 103; 250–263. 5. Reisman D, Glaros S, Thompson EA. The SWI/SNF complex and cancer. Oncogene 2009; 28; 1653–1668. 6. Phelan ML, Sif S, Narlikar GJ, Kingston RE. Reconstitution of a core chromatin remodeling complex from SWI/SNF subunits. Mol. Cell 1999; 3; 247–253. 7. Yan Z, Cui K, Murray DM et al. Pbaf chromatin-remodeling complex requires a novel specificity subunit, BAF200, to regulate expression of selective interferon-responsive genes. Genes Dev. 2005; 19; 1662–1667. 8. Wang W, Xue Y, Zhou S, Kuo A, Cairns BR, Crabtree GR. Diversity and specialization of mammalian SWI/SNF complexes. Genes Dev. 1996; 10; 2117–2130. 9. Inoue H, Furukawa T, Giannakopoulos S, Zhou S, King DS, Tanese N. Largest subunits of the human SWI/SNF chromatinremodeling complex promote transcriptional activation by steroid hormone receptors. J. Biol. Chem. 2002; 277; 41674–41685. 10. Wang X, Nagl NG, Wilsker D et al. Two related arid family proteins are alternative subunits of human SWI/SNF complexes. Biochem. J. 2004; 383; 319–325. © 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
717
11. Jones S, Wang TL, Shih Ie M et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science 2010; 330; 228–231. 12. Wu JN, Roberts CW. ARID1A mutations in cancer: another epigenetic tumor suppressor? Cancer Discov. 2013; 3; 35–43. 13. Sausen M, Leary RJ, Jones S et al. Integrated genomic analyses identify ARID1A and ARID1B alterations in the childhood cancer neuroblastoma. Nat. Genet. 2013; 45; 12–17. 14. Stephens PJ, Tarpey PS, Davies H et al. The landscape of cancer genes and mutational processes in breast cancer. Nature 2012; 486; 400–404. 15. Nagl NG Jr, Wang X, Patsialou A, Van Scoy M, Moran E. Distinct mammalian SWI/SNF chromatin remodeling complexes with opposing roles in cell-cycle control. EMBO J. 2007; 26; 752–763. 16. Sim JC, White SM, Fitzpatrick E et al. Expanding the phenotypic spectrum of ARID1B-mediated disorders and identification of altered cell-cycle dynamics due to ARID1B haploinsufficiency. Orphanet J. Rare Dis. 2014; 9; 43. 17. Bentley AM, Normand G, Hoyt J, King RW. Distinct sequence elements of cyclin B1 promote localization to chromatin, centrosomes, and kinetochores during mitosis. Mol. Biol. Cell 2007; 18; 4847–4858. 18. Halgren C, Kjaergaard S, Bak M et al. Corpus callosum abnormalities, intellectual disability, speech impairment, and autism in patients with haploinsufficiency of ARID1B. Clin. Genet. 2012; 82; 248–255. 19. Hoyer J, Ekici AB, Endele S et al. Haploinsufficiency of ARID1B, a member of the SWI/SNF-A chromatin-remodeling complex, is a frequent cause of intellectual disability. Am. J. Hum. Genet. 2012; 90; 565–572. 20. Santen GW, Aten E, Sun Y et al. Mutations in SWI/SNF chromatin remodeling complex gene ARID1B cause Coffin-Siris syndrome. Nat. Genet. 2012; 44; 379–380. 21. Tsurusaki Y, Okamoto N, Ohashi H et al. Mutations affecting components of the SWI/SNF complex cause Coffin-Siris syndrome. Nat. Genet. 2012; 44; 376–378. 22. Lwin ZM, Guo C, Salim A et al. Clinicopathological significance of calreticulin in breast invasive ductal carcinoma. Mod. Pathol. 2010; 23; 1559–1566. 23. Xue C, Wang X, Peng R et al. Distribution, clinicopathologic features and survival of breast cancer subtypes in southern China. Cancer Sci. 2012; 103; 1679–1687. 24. Sullivan HC, Oprea-Ilies G, Adams AL et al. Triple-negative breast carcinoma in African American and caucasian women: clinicopathology, immunomarkers, and outcome. Appl. Immunohistochem. Mol. Morphol. 2014; 22; 17–23. 25. Vignali M, Hassan AH, Neely KE, Workman JL. ATP-dependent chromatin-remodeling complexes. Mol. Cell. Biol. 2000; 20; 1899–1910. 26. Mohrmann L, Verrijzer CP. Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim. Biophys. Acta 2005; 1681; 59–73. 27. Watanabe R, Ui A, Kanno S et al. SWI/SNF factors required for cellular resistance to DNA damage include ARID1A and ARID1B and show interdependent protein stability. Cancer Res. 2014; 74; 2465–2475. 28. Helming KC, Wang X, Wilson BG et al. ARID1B is a specific vulnerability in ARID1A-mutant cancers. Nat. Med. 2014; 20; 251–254. 29. Inghirami G, Chiarle R, Simmons WJ, Piva R, Schlessinger K, Levy DE. New and old functions of STAT3: a pivotal target for individualized treatment of cancer. Cell Cycle 2005; 4; 1131–1133. 30. Flores-Alcantar A, Gonzalez-Sandoval A, Escalante-Alcalde D, Lomeli H. Dynamics of expression of arid1a and arid1b subun-
718
F Shao et al.
its in mouse embryos and in cells during the cell cycle. Cell Tissue Res. 2011; 345; 137–148. 31. Yan Z, Wang Z, Sharova L et al. BAF250B-associated SWI/SNF chromatin-remodeling complex is required to maintain undifferentiated mouse embryonic stem cells. Stem Cells 2008; 26; 1155–1165. 32. Khursheed M, Kolla JN, Kotapalli V et al. ARID1B, a member of the human SWI/SNF chromatin remodeling complex, exhibits tumour-suppressor activities in pancreatic cancer cell lines. Br. J. Cancer 2013; 108; 2056–2062.
Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Details of the oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER-2) antibodies and the cut-off value.
© 2015 John Wiley & Sons Ltd, Histopathology, 67, 709–718.
