CDC20 expression in oestrogen receptor positive breast cancer predicts poor prognosis and lack of response to endocrine therapy

Endocrine therapy is the standard treatment for oestrogen receptor positive (ER+) breast cancer. Despite its efficacy, around half of patients will develop resistance to this treatment and eventually relapse. Identification of effective and reliable biomarkers to predict the efficacy of endocrine therapy is of crucial importance in the management of ER+ breast cancer. Emerging evidence has revealed that the cell division regulator CDC20 exhibits an oncogenic function and plays important roles in tumourigenesis and progression of solid tumours. In this study, we investigated the prognostic and predictive role of CDC20 in early ER+ breast cancer patients. The biological and clinical impact of CDC20 expression was assessed in large clinical annotated cohort of ER+ breast cancer with long-term follow-up at the mRNA level, using METABRIC and KM-Plotter datasets, and the protein level using immunohistochemistry on patients presenting at Nottingham. CDC20 expression was correlated with clinico-pathological parameters, molecular subtypes, clinical outcome and efficacy of endocrine therapy. High CDC20 mRNA expression was associated with poor clinico-pathological parameters including large tumour size and high tumour grade (P < 0.0001) in patients with ER+ breast cancer. High CDC20 mRNA expression was significantly associated with poor patient outcome (P < 0.0001). Importantly, high CDC20 expression was correlated with poor response to endocrine treatment in patients who treated with hormonal therapy only (P < 0.01). In multivariate analysis, CDC20 mRNA was an independent predictor of poor clinical outcome after treatment with endocrine therapy (P = 0.02). CDC20 is a candidate biomarker for a subgroup of ER+ breast cancer characterised by poor clinical outcome. This study shows that the CDC20 could act as potential predictive biomarker of poor response to endocrine therapy in ER+ breast cancer.


INTRODUCTION
Breast cancer is a heterogeneous disease with various biological subtypes [1] with the most common form, approximately 75%, of breast cancer being oestrogen receptor posi\ve (ER+) [2,3]. Endocrine therapy is the main treatment for ER+ tumours, which has vastly improved survival and has reduced mortality [4]. However, a high propor\on of pa\ents receiving adjuvant endocrine therapy s\ll experience relapse and become resistant to treatment [5,6]. It is therefore highly desirable to predict, at an early stage of treatment, which ER+ pa\ents will and will not benefit from endocrine therapy.
Cell division cycle 20 homolog (CDC20) is a spindle assembly checkpoint molecule that required for the anaphase-promo\ng complex/cyclosome (APC/C) ac\va\on during mitosis, leading to ini\a\on of chroma\d separa\on and entrance of cell cycle into anaphase [7,8]. Defects in CDC20 func\on may therefore terminate mito\c arrest, which lead to tumourgenesis [9,10]. Consistent with the no-\on that CDC20 may func\on as an oncogene, several studies show overexpression of CDC20 in different types of cancers [11,12,9,13]. Indeed, its overexpression is suggested as a biomarker of poor outcome in pancrea\c [14], colon [15], primary non-small cell lung [16] and ovarian cancer [11].
In terms of breast cancer, two reports have demonstrated that CDC20 is a poten\al key player in the progression of breast cancer where it is significantly higher in breast cancer cells and high-grade primary tumour \ssues [17] and indicates an aggressive course of disease risk [18]. We aimed to in-ves\gate the role of CDC20 expression in ER+ tumours. In par\cular, we assessed whether CDC20 had a role in endocrine resistance which could be used to improve therapy predic\on in ER+ breast cancer.

MATERIALS AND METHOD CDC20 mRNA expression
The Molecular Taxonomy of Breast Cancer Interna\onal Consor\um (METABRIC) [19], comprising 1,506 ER+ breast cancer, was used as a discovery cohort to analyse and explore the prognos\c value of CDC20 mRNA and its role as predic\ve biomarker of clinical outcome for pa\ents who treated with endocrine therapy, Table 1.
The Kaplan Meier Plocer-Breast Cancer (KM-Plocer) online dataset [20], was used as a valida\on cohort for the prognos\c and predic\ve value of CDC20 mRNA expression using 2,061 pa\ents with ER+ breast cancer. The prognos\c value of CDC20 mRNA expression and associa\on with clinical outcome and clinico-pathological parameters were further validated using the Breast Cancer Gene-Expression Miner v4.0 (bc-GenExMiner v4.0) database [21]  informa\on on therapy and outcomes and tumour characteris\cs are prospec\vely maintained. The clinico-pathological parameters for the cohort series are summarised in Table 1.

Western bloPng
ImmunobloHng was performed as previously described [22], the specificity of CDC20 was validated on BT474 human breast cancer cells (American Type Culture Collec\on; Rockville, MD, USA) using primary CDC20 an\body (HPA039484, Sigma-Aldrich, 1:2000). This showed a specific band for CDC20 protein at the predicted size of 55 kDa (Fig. 1A).

IHC staining and Evalua4on
Tumour samples were arrayed as previously described [23]. The IHC staining was performed on 4 μm \ssue microarrays (TMAs) sec\ons using Novolink polymer detec\on system (Leica Biosystems, RE7150-K), detailed method was described in previous publica\on [22]. Sec\ons were incubated, overnight at 4 °C, with the primary CDC20 an\body diluted at 1:500. CDC20 immunoreac\vity was assessed using high-resolu\on digital images (Nanozoomer, Hamamatsu Photonics) and viewing soeware (Xplore; Philips, UK). Evalua\on was based on a semiquan\ta\ve assessment using a modified histochemical score (H-score), which includes an assessment of both the intensity and the percentage of stained cells [24]. The staining intensity of invasive tumour cells was scored into four categories 0 (no staining); 1 (weak staining); 2 (moderate staining) and 3 (strong staining). The percentage of each category was es\mated, and the H-score calculated. TMA cores were only assessed if tumour burden was >15%.

Clinical outcome data and events defini4on
Clinical outcomes including breast cancer specific survival (BCSS) was defined as the \me in months from the diagnosis to the date of breast cancer-related death. Recurrence free survival (RFS) was defined as the \me in months from diagnosis un\l developing local or regional recurrence. Distantmetastasis free survival (DMFS) was defined as the \me in months from diagnosis un\l developing distant-metastasis. For the benefit of endocrine therapy, the expression of CDC20 was inves\gated with clinical outcome on the endocrine treated cohort only. Secondary outcomes included associa-\ons with clinico-pathological parameters.

Sta4s4cal analysis
Data analysis was performed using SPSS sta\s\cal soeware (version 25, Chicago, IL, USA). The analysis for this study compared low and high expression of CDC20. The Chi-square test was performed for inter-rela\onships between categorical variables. Spearman's correla\on coefficient was used to examine the associa\on between con\nuous variables. One-way analysis of variance (ANOVA) with the post-hoc Tukey was used for differences between three or more groups. Kaplan-Meier analysis was used to assess the associa\on of CDC20 expression with clinical outcome. Mul\variate Cox Regression analysis with adjustment of covariates was used to iden\fy independent prognos\c biomarkers.
Benjamini-Hochberg procedure for mul\ple test correc\on was performed. P value of ≤ 0.05 was considered significant. The dichotomisa\on of CDC20 mRNA and protein expression into low and high groups was determined using X-Tile (X-Tile Bioinforma\cs Soeware, Yale University, version 3.6.1).

CDC20 expression in ER+ BC
High CDC20 mRNA expression in the METABRIC cohort was observed in 870 cases (58%), where low expression was observed in 636 cases (42%). CDC20 protein expression was localised to the cytoplasm of invasive tumour cells, with expression levels varying from absent to high (H-score range 0-250) ( Fig. 1B and 1C). CDC20 expression was dichotomised into low and high using an H-score of 120 resul\ng in 85 (25%) cases showing high expression and 262 (75%) cases with low expression.
Associa4on of CDC20 expression with clinic-pathological characteris4cs in ER+ breast cancer CDC20 mRNA expression was associated with aggressive clinico-pathological parameters including nega\ve expression of PR, poor NoHngham Prognos\c Index (NPI) and high tumour grade (P<0.0001, Fig. 2A-2C) using the METABRIC dataset. The associa\on with PR and NPI was validated using the bc-GenExMiner v4.0 dataset (P< 0.0001; Supplementary Fig. 1A and 1B). In contrast, CDC20 protein expression showed no sta\s\cal significance associa\on with any of the clinico-pathological parameters.

Clinical significance of CDC20 in pa4ents with ER+ BC
High mRNA CDC20 expression was significantly correlated with poor clinical outcome. Thus, results of METABRIC dataset showed that high expression of CDC20 mRNA was associated with poor RFS (P< ! 5 0.0001), DMFS (P< 0.0001) and high risk of death from breast cancer (P< 0.0001) (Fig. 3A-3C). To further validate our findings we used the KM-Plocer dataset which showed that high mRNA expression of CDC20 was also associated with poor clinical outcomes including RFS (P< 0.0001), DMFS (P< 0.0001) and BCSS (P< 0.0001) (Supplementary Fig. 2A-2C). Furthermore, results from the bc-GenEx-Miner v4.0 datasets showed that cases with low expression of CDC20 mRNA had favourable clinical outcomes compared to the high expression group, which showed poor clinical outcome (P< 0.0001; Supplementary Fig. 2D).

CDC20 is a predic4ve biomarker of poor response to endocrine therapy
In pa\ents who received endocrine therapy, tumours with high CDC20 mRNA expression were signi- In pa\ents with five years of follow up aeer endocrine treatment, results showed that high CDC20 mRNA expression was associated with high risk of recurrence (P= 0.008, Fig. 4D) and distant metastasis (P= 0.001, Fig. 4E). The significance of high CDC20 mRNA expression on predic\ng high risk of recurrence and distant metastasis on pa\ents who were treated with endocrine therapy alone was validated using KM-Plocer datasets (P< 0.0001; Supplementary Fig. 3D and 3E).
However, CDC20 protein expression showed no prognos\c associa\on with clinical outcome on the whole cohort of pa\ents with ER+ breast cancer, or the impact of endocrine therapy on pa\ent survival, recurrence or distant metastasis (P> 0.05; Fig. 5A-5F).

CDC20 is an independent prognos4c marker in ER+ breast cancer
High CDC20 mRNA expression was independent of tumour size, nodal stage and tumour grade in predic\ng a higher risk of recurrence (P= 0.005), distant metastasis (P= 0.005) and death from breast cancer (P= 0.0005), Table 3. In those pa\ents treated with endocrine therapy only, CDC20 mRNA expression was an independent prognos\c marker of tumour size, grade and nodal stage in predic\ng the risk of BCSS (P= 0.02), Table 4.

DISCUSSION
Breast cancer is a heterogeneous disease with various biological subtypes [1] with the most common form being ER+/luminal tumours [2,3]. This subtype remains heterogeneous in terms of recurrence, mortality rates, disease prognosis and response to treatment [3] despite acempts to biologically split them into luminal A and luminal B. Endocrine therapy, especially tamoxifen, s\ll the main treatment for pa\ents with ER+ breast cancer. Although sustained treatment with tamoxifen can successfully reduce postopera\ve recurrence and mortality rate, 30 to 50% of these pa\ents will develop resistance and later relapse [25]. Therefore, there is s\ll a need for a more precise method for stra\fying pa\ents based on their prognosis and response to endocrine therapy. CDC20 has a key roles in ac\va\ng the APC/C to ini\ate anaphase and late mitosis exit in the cell cycle [26,8]. CDC20 has also been shown to be a promising prognos\c marker for a variety of tumours; including pancrea\c [27,14], colorectal [15], lung [16] and breast cancer [18]. Here, we focused on the role of CDC20 in ER+ breast cancer and especially in pa\ents who were treated with only endocrine therapy. Our findings revealed that CDC20 is highly expressed in the more aggressive and highly prolifera\ve ER+ tumours, and implicated in resistance to endocrine therapy. Indeed, our findings showed a significant associa\on between high CDC20 mRNA expression and the poor pro-gnos\c clinico-pathological features within ER+ breast cancer. Despite our observa\ons that CDC20 protein was not prognos\c, it is likely that CDC20 expression plays a vital role in ER+ breast cancer progression.
Prolifera\on has a key role in the clinical behaviour of breast cancer and correlates strongly with poor clinical outcome and drug resistance. In addi\on to ER and PR, markers of prolifera\on seems to influence biological and clinical behaviour of ER+ breast cancer [19]. In light of this, our findings showed that high CDC20 mRNA expression was posi\vely correlated with prolifera\on-associated genes, including MKI67, CCNB1, CCNA2 and CCND1. This supports the results of previous studies, which reported that knockdown of CDC20 decreased cell prolifera\on and induced G2/M cell cycle arrest in hepatocellular carcinoma cells [28], and pancrea\c tumours [27]. Altogether, these data suggest that CDC20 is implicated in the prolifera\on of ER+ breast cancer which leads to tumourigenesis and aggressiveness phenotype.
Despite recent efforts to develop new breast cancer biomarkers, only ER and PR measurements are used currently both for clinical diagnosis to classify breast cancer pa\ents and as a guide to endocrine therapy [29]. Mul\gene signatures, including Oncotype DX, Mammaprint, Prosigna, Breast Cancer Index and EndoPredict, can be valuable as addi\onal prognos\c tools with regard to recurrence and the stra\fica\on of risk, but so far studies have not validated their value in predic\ng benefit from endocrine therapy [30]. Therefore, the iden\fica\on of predic\ve biomarkers for endocrine therapy efficacy in addi\on to ER and PR status is of urgent need to stra\fy pa\ents with ER+ breast cancer for targeted therapy. For this purpose, a key aim of this study was to assess the predic\ve val-! 8 ue of CDC20 mRNA and protein expression as a clinical marker of benefit from endocrine therapy in ER+ breast cancer. Our clinical data found a significant unfavourable effect of CDC20 mRNA expression in pa\ents treated with endocrine therapy. These findings lead to sugges\on that assessment of CDC20 mRNA expression prior to adjuvant treatment could predict patents who are highly to resist the endocrine therapy and eventually relapse.
Results from recent clinical trials demonstrated that 10 years of endocrine therapy showed improved RFS and overall survival compared with 5 years of endocrine treatment [31,32]. However, this is at the cost of unnecessary side effects that influence the quality of life for pa\ents [33]. Therefore, it is important to iden\fy a subgroup of pa\ents who are at high risk of relapse and who will not benefit from extended endocrine therapy. Our study demonstrates that for pa\ents with ER+ breast cancer treated with endocrine therapy, high expression of CDC20 mRNA remains a predic\ve marker for high risk of relapse and death from breast cancer at 5 years follow up. We suggest that assessment of CDC20 mRNA in clinical prac\ce would be useful to predict pa\ents who would not benefit from endocrine therapy and could spare them these risks and improve quality of life.
Although a previous study has showed the prognos\c value of CDC20 protein expression in pa\ents with triple nega\ve breast cancer [18] we found no associa\on with either clinico-pathological parameters or pa\ent clinical outcome in ER+ breast cancer or those who treated with endocrine therapy alone. This is might be explained by using different methods to evaluate the IHC staining.
The level of agreement in our study between results of CDC20 mRNA and protein expression was poor. This discrepancy could be explained by several possibili\es, including biological and technical explana\ons. Biological reasons include differences in post-transcrip\onal regula\on of CDC20 expression or tumour-specific differences in CDC20 mRNA/protein stability, while technical issues may include nature of an\body used in this study [34][35][36]. Such discrepancies between mRNA and protein levels for different reasons has occurred in mul\ple studies of breast cancer [37,38].
In summary, our data revealed a posi\ve correla\on of high CDC20 expression at the transcriptomic level with poor clinical outcome in pa\ents with ER+ breast cancer. Also, we have provided evidence that CDC20 mRNA expression in ER+ breast cancer is a poten\ally predic\ve for selec\ng pa\ents who might not experience benefits from endocrine therapy. Addi\onal or alterna\ve target therapies could then be given to those who predicted to be resistance to endocrine therapy, this would be a significant shie toward truly individualised medicine. We suggested that CDC20 mRNA expression could be used in clinical either singularly or in combina\on with other genes as mul\gene signature to guide the choice of endocrine treatment. Western bloHng result for CDC20 expression in BT474 breast cancer cell lysates. CDC20 protein expression in invasive breast cancer cores using IHC B) Nega\ve and C) posi\ve CDC20 protein expression. Figure 2: Associa\on of CDC20 mRNA expression with clinico-pathological parameters A) PR, B) NPI and C) grade using METABRIC dataset. CDC20 mRNA correla\on with prilfera\on associated genes including D) MKI67, E) CCNB1, F) CCNA2 and G) CCND1 using bc-GenExMiner v4.0 dataset. Figure 3: CDC20 mRNA and pa\ent outcome in ER+ breast cancer using METABRIC dataset: A) recurrence, B) distant-metastasis and C) BCSS. Figure 4: CDC20 mRNA expression as a predic\ve biomarker for poor clinical outcome in pa\ents with ER+ breast cancer aeer endocrine treatment using METABRIC dataset A) recurrence, B) distantmetastasis and C) BCSS. Kaplan-Meier survival plots for pateints with ER+ breast cancer aeer endocrine treatemnt and only 5 years follow-up D) recurrence and E) distant-metastasis. Figure 5: CDC20 protein and pa\ent outcome in ER+ breast cancer using NoHngham cohort: A) recurrence, B) distant-metastasis and C) BCSS. CDC20 protein expression and clinical outcome in pa-\ents with ER+ breast cancer who were treated with endocrine therapy only using NoHngham cohort A) recurrence, B) distant-metastasis and C) BCSS.