Pathology – Research and Practice 210 (2014) 10–17

Contents lists available at ScienceDirect

Pathology – Research and Practice journal homepage: www.elsevier.com/locate/prp

Original Article

Clinicopathological and molecular significance of Sumolyation marker (ubiquitin conjugating enzyme 9 (UBC9)) expression in breast cancer of black women A.O. J. Agboola a,c,∗ , A.A. Musa b , B.A. Ayoade b , A.A. Banjo c , C.C. Anunobi c , A.M. Deji-Agboola d , E.A. Rakha a , C. Nolan a , I.O. Ellis a , A.R. Green a a

Division of Pathology, School of Molecular Medical Sciences, Nottingham University Hospitals and University of Nottingham, Nottingham, United Kingdom Department of Surgery, Olabisi Onabanjo University, Sagamu, Nigeria c Department of Morbid Anatomy and Histopathology, Olabisi Onabanjo University, Sagamu, Nigeria d Department of Medical Microbiology Parasitology, Olabisi Onabanjo University, Sagamu, Nigeria b

a r t i c l e

i n f o

Article history: Received 1 May 2013 Received in revised form 22 August 2013 Accepted 23 September 2013 Keywords: Breast cancer UBC9 Black women Patient outcome

a b s t r a c t The majority of breast cancers (BC) in Nigerian women are triple negative and show breast cancerassociated gene 1 (BRCA1) deficiency as well as the basal like phenotype, with a high mortality rate. In contrast to the well-defined predictive factors for the hormonal therapy, there is a paucity of information on the BRCA1 deficiency breast tumor biology, particularly among African women. BRCA1 Sumoylation (UBC9) has been speculated to be involved in the ER transcription activity, BRCA1 deficiency and triple negative BC. We therefore hypothesized that UBC9, a SUMOylation marker, may have contributed to the aggressive nature of BRCA1 tumor phenotype observed in Nigerian women. This study investigated the immunoprofiles of UBC9 in tissue microarray (TMA) of 199 Nigerian women and correlated their protein expression with clinical outcome, pathological responses and the expression of other biomarkers to demonstrate the functional significance in Nigerian women. The protein expression of UBC9, as compared with other biomarkers, showed an inverse correlation with steroid hormones (ER, progesterone (PgR)), BRCA1, p27, p21 and MDM4, and a positive correlation with triple negative, basal cytokeratins (CK14 and CK5/6), epidermal growth factor receptor (EGFR), basallike breast cancer phenotype, p53, phosphoinositide-3-kinases (PI3KCA), placental cadherin, (P-cadherin) and BRCA1 regulators (metastasis tumor antigen-1 (MTA1). Survival analysis showed that those tumors positive for UBC9 expression had a significantly poorer breast cancer-specific survival (BCSS) as compared with those showing negative expression. UBC9 remained an independent predictor of outcome for BCSS. This study demonstrates that UBC9 appears to play an important role in the tumor biology of Nigerian women. Therefore, a novel UBC9 targeted therapy in black women with BC could enhance a better patient outcome. © 2013 Elsevier GmbH. All rights reserved.

Introduction Women carrying reduced or aberrant BRCA1 breast tumors have triple negative breast cancer (TNBC) and the basal phenotype with aggressive features. This is particularly common among black women [1,2]. SUMOylation has been identified as a modulator of transcription factor activities involved in numerous cellular signaling pathways, including BRCA1, cell cycle and steroid hormones in breast cancer pathogenesis [3–13]. UBC9 is of one the key enzymes required for SUMOylation. As a member of the Small Ubiquitin-like-modifier (SUMO) substrate

∗ Corresponding author at: Olabisi Onabanjo University, Department of Morbid Anatomy and Histopathology, OACHS, P.M.B 2022, Sagamu, Ogun state, Nigeria. 0344-0338/$ – see front matter © 2013 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.prp.2013.09.011

family, it has been cited in a variety of cellular pathways, including cell growth, chromatin remodeling, proliferation and apoptosis [14]. SUMO ylation-independent functions of UBC9 have also been reported [15]. Experimental evidence supports this notion, where it functions as a regulator of nuclear transport and co-regulator of transcription. Kurtzman and Schechter et al. [16] observed that UBC9 binds to a nuclear localization signal at the N-terminus of the visual system homeobox 1 (Vsx-1) homeodomain that is saddled with responsibility of regulating bipolar cell differentiation of the zebra-fish undergoing retinogenesis [16]. A similar transcription modulation activity was reported with respect to glucocorticoid receptors [17]. Although UBC9 is ubiquitously expressed in normal epithelial cells, it has been frequently reported to be either up-regulated or over-expressed in tumor cells, including ovarian, lung, head and

A.O.J. Agboola et al. / Pathology – Research and Practice 210 (2014) 10–17

neck, melanoma and breast carcinomas with the aid of microarray analysis, real time polymerase chain reaction (PCR) and immunohistochemistry (IHC) [18–22]. In breast tumor cells, it enhances tumor growth and promotes invasion and metastasis [15]. It is also involved in the regulation of ER alpha [23], where UBC9 binds to BRCA1 amino terminal domain for the down regulation of ER alpha activity [24]. Also, Qin [3] reported that UBC9 has a direct effect on the BRCA1 activities, where UBC9 mediates nuclear localization and growth suppression of BRCA1in breast cancer cell line [3]. There are potential links between loss of BRCA1 E3 ubiquitin ligase activity, loss of ER repression and UBC9 binding. In line with this, Xu et al. [24] reported that BRCA1 cancer-predisposing mutation suppresses the ability to bind to UBC9 and thereby inhibits ER activity, therefore suggesting a role for BRCA1 in regulating SUMO modifications of ER [24]. The deregulation of this BRCA1 SUMO modification switch in ER-positive or -negative breast cancers has been reported [24]. UBC9 also plays a vital role in cancer progression and resistance to chemotherapy [20,25,26]. The prognostic relevance of this marker has also been cited in breast cancer, where it was demonstrated to be associated with higher tumor grade, invasive ductal histological type, larger size, lymph node involvement, basal phenotype, resistance to chemotherapy and poor clinical outcome [20,26]. Until now, in contrast to the well defined predictive factors for the hormonal therapy, there is a paucity of information on the BRCA1 deficiency breast tumor biology, particularly among black women. The majority of breast cancers (BC) in Nigerian women are triple negative and show BRCA1 deficiency, as well as the basal like phenotype, with a high mortality rate. We therefore hypothesized that UBC9, a SUMOylation marker, may have contributed to the aggressive nature of BRCA1 tumor phenotype observed in Nigerian women. The aim of this study was to investigate the prognostic significance of UBC9 expression in this Nigerian tumor series using IHC and TMA in order to establish the relationship between the UBC9, pathological features, expression of biomarkers and clinical outcome. Materials and methods Patients The patient cohort comprised formalin-fixed, paraffin embedded (FFPE) breast cases from 199 women presenting at the Olabisi Onabanjo University Teaching Hospital, Sagamu, and Histopathology Specialist laboratory, Idi-Araba Lagos, Nigeria. The clinical history and tumor characteristics were assessed in a standardized manner for all the patients. Out of the 199 cases, 130 (65.3%) and 69 (34.7%) patients were under and above 50 years old respectively. One hundred and thirty-nine (69.9%) women were premenopausal. The tumor characteristics also varied: 3 (1.5%) cases were Grade 1, 125 (62.8%) Grade 2 and 71 (35.7%) were Grade 3. The majority of the tumors 178 (89.5%) were larger than 2 cm, and 174 (87.4%) were invasive ductal/no special histological tumor type (NST), 182 (99.2%) patients were lymph node-positive, and 155 (77.9%) had definite vascular invasion. The immunoreactivity, scoring and categorization of ER, PgR, cytokeratins (CK: CK5/6, CK14,), EGFR, p53, BRCA1, BRCA1 downregulator (MTA1), cell adhesion molecules (␤ catenin, Ecadherin, P-cadherin), basal-like and triple-negative phenotypes were defined in this series as previously described [27,28]. The American society of clinical oncology/College of American pathologists guideline recommendations testing in breast cancer was used for HER2 assessment [29]. Borderline (2+) cases were confirmed by chromogenic in situ hybridization (CISH) as previously described

11

[30]. Nielsen’s method [31] was used for molecular classification, comprising Luminal A (ER, PR positive and HER2 negative), Luminal B (ER, PR HER2 positive), Basal (ER, PR, HER2 negative and CK5/6 and or EGFR positive), HER2 (ER negative and HER2 positive) and an unclassified group (ER, PR, HER2 CK5/6 and EGFR negative). After the exclusion of the missing cores, the following biomarkers were used for further analysis of UBC9 expression: ER 39 (21.1%), 146 (78.9%), PgR 31(19.7%), 126 (80.3%) cytokeratins (CK: CK 14: 66 (45.8%), 78 (54.2%) CK5/6; 87 (51.2%) 83 (48.8%), HER2: 32 (18.7%), 139 (81.3%), EGFR 59: (37.8%), 97 (62.2%) BRCA1:22 (14.1%),134 (85.9%), BRCA1 down regulator (MTA1: 100 (53.5%), 87 (46.5%), Cell adhesion molecules (␤ catenin: 98 (62.8%), 58(37.2%), Ecadherin:47 (31.8%), 101 (68.2%) P-cadherin: 91 (60.7%), 59 (39.3%), and cell cycle marker: MDM4: 33 (24.3%), 103 (75.7%), p21: 28 (16.9%), 138 (83.1%), p27: 48 (27.4%), 127 (72.6%) and p53: 112 (73.7%),40 (26.3%), positive and negative expression respectively, basal-like phenotype: 67 (55.8%) and triple-negative: 99 (64.29%). Patient outcome and treatment data were retrieved from the patients’ records. Patients were followed up for a median time of 60 months (240 weeks). Of the 199 patients, 85 (42.7%) had complete patient outcome records for survival, while the remaining 114 (57.3%) patients were lost to follow up. Thirty-nine (32.8%) of the cases had recurrence. All patients were treated with a combination of classical chemotherapy and hormonal therapy (tamoxifen. Eighty five of the patients (42.7%)) received radiotherapy. The reporting recommendations for tumor marker prognostic studies (REMARK) criteria, recommended by McShane et al. [32], were followed. This study was approved by the medical advisory committee, Olabisi Onabanjo University Teaching Hospital, and by the Nottingham Research Ethics Committee 2 under the title “Development of a molecular genetics classification of breast cancer”. Tissue microarray array construction One hundred and ninety nine samples from the Nigerian cohort were constructed as tissue microarrays (TMA) as previously described [27,28]. Breast tumor cores were taken from each FFPE donor tissue block marked for the most representative points of tumor (both peripherally and centrally). A precision instrument (ALPHELYS MiniCore® ) was used to take representative tissue cores (0.6 mm diameter, 3 mm height) from each sample, which was arrayed into recipient paraffin. Immunohistochemistry method Table 1 shows the different biomarkers used in this study. Antigen retrieval was performed by microwaving the slides at 800 W for 10 min followed by 560 W for 10 min in citrate buffer (1 M sodium citrate at pH of 6.0) and cooling in running water immediately. The primary antibody for the biomarkers was incubated for 60 min at room temperature. Diaminobenzidine tetrahydrochloride (DAB) solution was incubated for 10 min after which copper-sulphate solution (0.5% copper sulphate in 0.8% sodium chloride) was applied to the slides and incubated for 10 min and counterstained with hematoxylin for 2–3 min, followed by rinsing in tap water. Slides were de-hydrated by immersing in three alcohol baths for 10 s and cleared in two xylene baths followed by application of cover slip. Negative and positive controls were performed by omitting the primary antibody and including control tissues as specified by the antibody supplier, respectively. Immunohistochemical scoring The scoring was performed using the percentage of positive cells to produce a final score in the range 0–100. The cases were

12

Table 1 Sources, dilution, distribution, cut-offs point and pre-treatment used for revalidation. Clone

Source

Dilution

Distribution

Scoring System

Cut-offs

Pre-treatment

Positive control

Negative control

␤-Catenin

Ab5647

Abcam

1:1:00

Nuclear

% Of positive cells

Omitting the antibody

Ab-1 (MS110)

Calbiochem

1:150

Nuclear

% Of positive cells

Antigen retrieval Microwave Antigen retrieval Microwave

Human Colon cancer

BRCA1

10% (negative)

UBC9

Ep2938Y

Novus biologicals

1:100

Nuclear

% Of positive cells

>70% (positive)

Antigen retrieval microwave Antigen retrieval Microwave Antigen retrieval Microwave Antigen retrieval microwave Antigen retrieval Microwave Antigen retrieval Microwave Antigen retrieval Microwave Antigen retrieval microwave

Known case of CK14 breast cancer Normal gastric mucosa

Omitting the antibody Omitting the antibody Omitting the antibody

Myoepithelial cells of normal duct in normal mammary gland Known case of erbB2 strong BC expression Normal breast acini

Omitting the antibody

Known case of MDM4 strong BC expression

Omitting the antibody

Human gastric adenocarcinoma Known case of P-cadherin strong BC expression Normal breast acini

Omitting the antibody

Omitting the antibody Omitting the antibody

Omitting the antibody

Omitting the antibody

Human breast carcinoma Normal breast acini

Omitting the antibody Omitting the antibody

Normal breast acini

Omitting the antibody

Normal breast acini

Omitting the antibody

Human brain tissue

Omitting the antibody

A.O.J. Agboola et al. / Pathology – Research and Practice 210 (2014) 10–17

Antibody

A.O.J. Agboola et al. / Pathology – Research and Practice 210 (2014) 10–17

13

Fig. 1. (A) and (B) showing positive and negative UBC9 immunochemistry expression. Mag. ×20.

scored without knowledge of the clinicopathological parameters or patient outcome. TMAs were scored independently twice by one observer (Johnson Agboola, JA). The mean of the scores was calculated to reach a final score. A proportion of these was counterscored by an observer (Andrew Green, AG) to ensure reproducibility. Statistical analysis Statistical analysis was performed using SPSS 16.0 statistical software. Chi-squared analyses were used for inter-relationships between the UBC9 expression, clinicopathological parameters and other biomarkers. The Kaplan–Meier survival method and the logrank test were used for survival curves. Multivariate analyses using Cox proportional hazard regression models were performed and from the model, both the risk factor and 95% confidence intervals were generated. A two-sided p-value of 72% high expressions were used as cut-off points. A hundred and eighteen (59.3%) Nigerian breast cancer cases were considered positive and 81 (40.7%) negative, respectively. The expression of SUMOlyation markers in relation to clinicopathological variables is summarized in Table 2. There was no positive correlation between UBC9 and clinicopathological variables, except for nuclear pleomorphism (p = 0.03). A significant proportion of tumors expressing UBC9 are inversely associated with BRCA1 (p = 0.04), steroid hormones; ER, PgR (p < 0.001), p27 (p < 0.001), p21 (p = 0.001) and MDM4 (p = 0.02) expression. In addition, UBC9-positive tumors were positively associated with the BRCA1 regulator marker MTA1 (p < 0.001), EGFR (p < 0.001), basal cytokeratin: (CK5/6 (p < 0.001). CK14 (p = 0.01)), p53 (p < 0.001), P-cadherin (p = 0.02), ␤-catenin (p = 0.007), PI3KCA (p < 0.001), triple negative tumors (p < 0.001) and basal-like breast cancer (p < 0.001) using Nielsen’s classification. There was no significant correlation with HER-2 and E-cadherin expression (Table 3). Outcome Univariate survival analysis showed that tumors positive for UBC9 expression had a significantly poorer BCSS as compared with those showing negative expression (p < 0.001) and DF1 (p = 0.03) (Fig. 2). In Cox multivariate analysis, UBC9 remains a prognostic factor for BCSS, independent of tumor grade, size, and lymph node involvement (p = 0.004) (Table 4).

Discussion Recent advances in cancer studies suggest that the protein modification involving SUMO has wider functions in the transcriptional and cellular regulations, including proliferation, invasion and metastasis, particularly in breast cancer [15,24,33]. Identifying the roles of UBC9, a SUMOylation protein in a series of breast cancers that are mainly triple negative and basal-like, will enhance the understanding of the involvement of BRCA1 dysfunction observed in breast cancer of Black women. The expression of this biomarker was observed in the nucleus, which is in line with previous studies of carcinomas that demonstrated co-localization of the SUMO proteins in the nucleus of cells and involved in the regulation of the several cellular functions [14,34]. The intensities of the biomarker staining were in agreement with a study investigated in Chinese women, where Chen et al. observed high UBC9 expression in Chinese women with BC [26]. Recently, UBC9 has been identified as having a role in E2induced ER alpha and BRCA1 degradation via ubiquitin-proteasome pathway [3,24]. In keeping with this notion, the majority of UBC9 tumors lacked ER expression. UBC9 has been speculated to downregulate BRCA1 expression [11,24], which is consistent with the current observations where a large proportion of tumors expressing UBC9 did not express BRCA1 and also had a positive association with BRCA1 down-regulators. UBC9 was previously observed to upregulate MTA1 via interaction with RNA polymerase 11 at the SUMO promoter region [13]. The result of this study might reflect its interaction with ER, basal like and triple-negative tumors. In one of the studies, we have reported a higher percentage of triple negative, basal-like breast cancer and BRCA1 deficiency in indigenous African women [28]. Thus, UBC9 expression may have explained in part the reasons behind the higher incidence of ER negativity, BRCA1 deficient, basal phenotype and TNBC observed in this study and in other black women with BC. Therefore, UBC9 may be useful for screening for the basal like phenotype, BRCA1 and triple negative patients in the indigenous African women. PI3KCA, E- and P-cadherin and ␤-catenin have been implicated in cell-cell adhesion [35]. Over-expression of P-cadherin and ␤-catenin and also loss of E-cadherin are associated with aggressive tumors, including high-grade, ER-negative and basal phenotype-related tumors [36–38]. High PI3KCA expression in the tumor predisposes cells toward epitheal–mesenchymal transition, necessary for cell invasion and metastasis [39–41]. In a related development, UBC9 was found to promote breast cell invasion and metastasis in a SUMOylation-independent manner [15]. Therefore, the link between the UBC9 expression and EMT observed in Nigerian BC might probably explain the reason for the high degree of metastasis at diagnosis. Cell cycle regulatory genes have been shown to be good predictors of poor outcome in BC [42]. SUMOylation of the p53 has been reported in a previous study [43]. UBC9 involvement in the cell cycle

14

A.O.J. Agboola et al. / Pathology – Research and Practice 210 (2014) 10–17

Table 2 Relationship between UBC9 expression and clinicopathological variables in breast cancer. Variables

UBC9 expression Negative (%)

Positive (%)

2 Value

p Value

Age (years) ≤50 >50

54 (66.7) 27 (33.3)

76 (64.4) 42 (35.6)

0.10

0.74

Lymph node involvement Negative Positive

9 (11.1) 72 (88.9)

8 (6.8) 110 (93.2)

1.15

0.28

Menopausal Pre Post

51 (63.0) 30 (37.0)

88 (74.6) 30 (25.4)

3.07

0.07

Mitotic figure Low Medium High

58 (71.6) 12 (14.8) 11 (13.6)

68 (57.6) 28 (23.7) 22 (18.6)

4.12

0.12

Nuclear pleomorphism Small uniform cells Moderate increase in size Marked variation

4.96

0.03

35 (43.2) 46 (56.8)

33 (28.0) 85 (72.0)

Tubule formation >75% 10–75% 2.0

7 (8.6) 74 (91.4)

14 (11.9) 104 (88.1)

0.52

0.46

Tumor type Typical medullary Atypical medullary Tubular Lobular Ductal NST Mucinous Tubulolobular Lobular mixed Tubular mixed Mixed NST Others

1 (1.2) 1 (1.2) 2 (2.5) 0 (0.0) 69 (85.2) 2 (2.5) 1 (1.2) 2 (2.5) 3 (3.7) 0 (0.0) 0 (0.0)

0 (0.0) 1 (0.8) 1 (0.8) 3 (2.5) 105 (89.0) 1 (0.8) 0 (0.0) 1 (0.8) 6 (5.1) 0 (0.0) 0 (0.0)

7.84

0.44

Vascular invasion Negative Positive

66 (81.5) 15 (18.5)

89 (75.4) 29 (24.6)

1.02

0.31

Others: spindle, metaplastic and alveolar lobular tumor histological type.

regulation was reported by Park et al., where UBC9 negatively regulates BRCA1 via p21, p27 and Gadd45 promoters [44]. The association of UBC9 with abnormal expression of p53 and p21 in Nigerian tumors implied that these tumors would have had disorganized

cell cycle control and thereby encouraged rapid abnormal cell divisions, which is a hallmark of tumor aggressiveness. The results of UBC9 expressions in relation to patient outcome observed in this study is in agreement with previous studies, where

Fig. 2. UBC9 expression in relation to (A) BCSS and (B) disease free interval (DF1) demonstrating survival and recurrence differences, respectively, between positive and negative expression in Nigerian BC.

A.O.J. Agboola et al. / Pathology – Research and Practice 210 (2014) 10–17

15

Table 3 UBC9 expression compared with other biomarkers protein expression. Variables

UBC9 expression Negative (%)

Positive (%)

2 Value

p Value

BRCA1 Negative Positive

43 (75.4) 14 (24.6)

91 (91.9) 8 (8.1)

8.11

0.04

Basal cytokeratins CK14 Negative Positive

34 (68.0) 16 (32.0)

44 (46.8) 50 (53.2)

5.90

0.01

CK5/6 Negative Positive

51 (82.3) 11 (17.7)

32 (29.6) 76 (70.4)

43.66