Clinicopathological and prognostic significance of Ras association and pleckstrin homology domains 1 (RAPH1) in breast cancer

Ras association and pleckstrin homology domains 1 (RAPH1) is involved in cytoskeleton regulation and re-epithelialisation in invasive carcinoma and, therefore, may play a key role in carcinogenesis and metastasis. We, herein, investigated the biological and clinical significance of RAPH1 in breast cancer using large annotated cohorts. The clinicopathological and prognostic significance of RAPH1 was assessed at the genomic and transcriptomic levels using The Cancer Genome Atlas (TCGA) dataset (n = 1039) and the results were validated using the Molecular taxonomy of breast cancer international consortium (METABRIC) cohort (n = 1980). RAPH1 protein expression was evaluated by immunohistochemistry in a large, well-characterised cohort of early-stage breast cancer (n = 1040). In both the TCGA and METABRIC cohorts, RAPH1 mRNA expression and RAPH1 copy number alteration were strongly correlated. RAPH1 mRNA overexpression was significantly correlated with high expression of adhesion and EMT markers including CDH1, TGFβ1 and CD44. RAPH1 mRNA overexpression was a significant predictor of a poor prognosis (Hazard ratio 3.88; p = 0.049). High RAPH1 protein expression was associated with higher grade tumours with high proliferation index, triple negative phenotype and high E-cadherin expression. High RAPH1 protein expression was an independent predictor of shorter survival (Hazard ratio 4.37; p = 0.037). High RAPH1 expression is correlated with aggressive breast cancer phenotypes and provides independent prognostic value in invasive breast cancer.


INTRODUCTION
Recent developments in diagnostic accuracy and drug therapies have improved outcomes in early-stage breast cancer (BC). However, more than 20% of BC patients have poor clinical outcomes related to recurrence and metastasis [1,2]. Various studies have been conducted to identify factors associated with tumour invasion and metastasis to uncover novel candidates for targeted treatment.
Ras association and pleckstrin homology domains 1 (RAPH1), also referred to as lamellipodin (LPD) [3], acts as a downstream effector of the Ras pathway [4], which plays an important role in epithelial-mesenchymal transition (EMT) [5,6]. EMT is thought to play an important role in the development of distant metastasis in BC [7]. Previous studies showed changes in intracellular tight junction characteristics and weakened cell adhesive properties during EMT as underlying mechanisms inducing the loss of epithelial features and the gain of mesenchymal morphology [8,9]. RAPH1 has a role in regulating cell migration and reepithelialisation [10]. Aberrant expression of RAPH1 may also promote cancer cell adhesion at distant, metastatic sites [10]. However, the clinical significance and prognostic role of RAPH1 in BC remains to be defined.
In this study, RAPH1 was investigated in BC at the gene and protein phases using large well-characterised and annotated cohorts of early stage BC with emphasis on the relation between RAPH1 expression and clinicopathological factors including the biomarkers associated with EMT and patients' outcomes.

Genomic and transcriptomic analysis of RAPH1 gene
In this study, the Cancer Genome Atlas (TCGA) [11] dataset (n = 1039) provided by cBioPortal [12] was used for copy number and mRNA analysis of RAPH1 gene. Briefly, RAPH1 copy number alteration (CNA) datasets from Affymetrix SNP 6.0 arrays and RAPH1 mRNA expression data from RNASeqV2 were accessed, and information on several clinicopathological factors and outcomes were also collected.
The Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) dataset [13,14] including 1980 patients was used for validation of the findings based on the TCGA dataset. In this cohort, patients with oestrogen receptor (ER)-positive tumours with negative lymph node metastasis did not receive adjuvant chemotherapy, whereas those with ER-negative or lymph node metastasis-positive tumours received adjuvant chemotherapy. No patients with HER2-positive BC were treated with trastuzumab. Haematoxylin & eosin staining and immunostaining with antibodies against CD34 and D2-40 according to our previously published methods [15] were used for assessment of lymphovascular invasion (LVI) status. Availability and assessment of oestrogen receptor (ER), progesterone receptor (PR), HER2, the proliferation marker Ki67, EMT related genes ! 5 (E-cadherin and N-cadherin), basal markers (CK5/6 and EGFR), and the stem cell markers (CD24 and CD44) were described in previous studies [16][17][18][19][20][21][22]. This study was approved by the Nottingham Research Ethics Committee 2 (Reference title: Development of a molecular genetic classification of breast cancer). This study was conducted in accordance with the Declaration of Helsinki, and all patients participated in the study after informed consent. characteristics. Tumour samples, available as 0.6-mm cores, were arrayed as previously described [23]. For IHC, the Novolink Max polymer detection system (RE7280-k, Leica, Newcastle, UK) was used to visualise the reaction. . Following optimisations, the primary antibody for anti-RAPH1 was diluted 1:300 in Bond primary antibody diluent (Leica, Germany). For the staining control, a polyclonal rabbit anti-human beta-2 microglobulin antibody (1:2000; Dako, Glostrup, Denmark) was used. 3-3' diaminobenzidine tetrahydrochloride (Novolink DAB substrate buffer plus) was treated as the chromogen.

RAPH1 protein analysis expression
Counter-staining was performed using Meyer's haematoxylin for 6 minutes. Immunostained tissue microarray sections were digitally scanned using a Nano Zoomer (Hamamatsu Photonics, Tokyo, Japan). RAPH1 IHC expression was classified into the following groups according to the cytoplasmic staining intensity of the cancer cells: no staining, weak staining, moderate staining and strong staining ( Figure 1). The H-score system, as previously reported [24,25], was used based on intensity scoring (0, negative; 1, weak; 2, moderate and 3, strong) and proportion scoring (0-100) based on the cytoplasmic RAPH1 expression pattern.

Statistical analysis
Statistical analyses were conducted using the SPSS statistical software v24.0 (IBM, Armonk, NY, USA). The relationship of RAPH1 CNA with RAPH1 mRNA expression was examined using analysis of variance test with Bonferroni correction. Pearson correlation test was used to evaluate the relationship of RAPH1 mRNA expression with the mRNA expression of other genes. The chi-squared test was used to assess differences in protein expression level stratified by several clinicopathological factors including tumour size, lymph node status and ER and HER2 expression. To assess the association of RAPH1 with prognosis, Kaplan-Meier survival curves for BC-specific survival (BCSS) were determined, and the log-rank test was conducted. The BCSS in this study was determined as the time from the day of surgery to the day of death due to breast cancer. For univariate and multivariate survival assessment of clinicopathological factors including RAPH1 expression, 95% confidence intervals (CIs) were assessed using the Cox proportional hazards regression model. In these analyses based on mRNA expression, median mRNA value was determined as a cut-off point. For RAPH1 protein, significant cut-off point was calculated as H-score 90 using X-Tile (X-Tile Bioinformatics Software, Yale University, version 3.6.1) based on patient outcome, and samples were divided into high and low expression groups.

RAPH1 copy number alteration and mRNA expression
In the TCGA cohort of 1039 patients, RAPH1 copy number gain and RAPH1 loss were observed in 99 (9.5%) and 219 (21.1%) cases, respectively. In 1980 patients of the METABIRIC cohort, RAPH1 copy number gain and RAPH1 loss were observed in 27 (1.4%) and 33 (1.7%) cases, respectively. In both cohorts, RAPH1 mRNA expression was higher in the tumours with copy number gain than in the neutral cases (p < 0.0001) and was lower in the tumours with copy number loss than in the neutral cases (both p < 0.0001).
High RAPH1 mRNA overexpression was associated with higher CDH1 (TCGA: p <

Immunohistochemical expression of RAPH1 protein
Among full-face tissue sections, the RAPH1 expression grades and patterns in carcinoma cells were apparently different from those of normal mammary glands, as shown in Figure 3. RAPH1 expression in normal glandular epithelium was uniformly weak. The RAPH1 immunoreactivity of myoepithelial cells was lower than those of glandular cells. In contrast, RAPH1 immunopositivity was observed in the cytoplasm of invasive cancer cells, which was stronger compared to normal epithelial cells. In intraductal carcinoma cells, the ! 8 IHC expression pattern of RAPH1 was similar to that of invasive cancer cells, whereas the RAPH1 immunoreactivity was stronger than that of normal epithelial cells. The RAPH1 expression of myoepithelial cells around ductal components was weak, similar to that observed in normal myoepithelial cells. In both the invasive and intraductal cancer cells, the nuclei of a few cancer cells showed weak to moderate immunoreactivity concomitantly with cytoplasmic immunostaining.
The BCSS in BC patients with high RAPH1 expression was significantly different from that of those with low RAPH1 expression (HR 4.37, p = 0.037; Figure 2b Table 3).

DISCUSSION
Cancer cell migration and cancer cell adhesion at distant metastatic sites have key roles in BC metastasis. EMT is involved in initial transformation of the primary invasive carcinoma to a metastatic phenotype, whereas cell adhesion has its functional effect during the establishment of cancer metastasis at distant sites [4,26,27]. In present study, although RAPH1 protein expression was not associated with the biomarkers related to EMT, RAPH1 overexpression was significantly associated with high tumour proliferation and E-cadherin positivity. Moreover, high RAPH1 mRNA expression was associated with the high mRNA expression levels of CDH1, which encodes E-cadherin. Interestingly, there is evidence that RAPH1 interacts with E-cadherin [28]. Analysis of RAPH1 protein expression by IHC also revealed that its immunoreactivity in normal epithelium, albeit uniformly weak, was higher than that in myoepithelial cells. E-cadherin, a member of a transmembrane glycoprotein family, is involved in cell-to-cell adhesion [29,30]. The extracellular domain of E-cadherin is involved in cell-cell interactions, whereas the intracellular domain regulates the assembly of the actin cytoskeleton via β-catenin, which has an important role in cell adhesion and signal transduction [31][32][33]. Following invasion, migrating BC cells keep or recover epithelial characteristic to adhere metastatic sites [34,35]. Previous studies reported that RAPH1, also referred to as Lamellipodin/LPD, regulates formation of lamellipodia on cells and regulated cell migration and adhesion [10,36,37] by inducing the activity of Ena/VASP and Scar/ WAVE, which are involved in the construction of actin cytoskeleton at the cellular membrane to promote invasion and metastasis of BC [3,[36][37][38][39]. High expression and membrane localisation of the actin regulator RAPH1 was shown to be associated with poor outcomes in BC patients [3]. This result suggested that RAPH1 might be associated with cancer cell adhesion at distant metastatic sites after LVI. In the present study, a strong RAPH1 ! 10 expression was observed in the cytoplasm of invasive BC cells; therefore, further investigation is necessary to elucidate the role of intracellular localisation of RAPH1 on its protein activity.
Previous studies indicated that RAPH1 gene abnormalities were not associated with outcomes in BC [38]. Although the number of cases with RAPH1 CNA was limited in the current study, with less than 10% of the cases in the TCGA cohort exhibiting copy number gain, RAPH1 mRNA expression was strongly associated with RAPH1 CNA. Additionally, the association between RAPH1 mRNA expression and outcome was significant. RAPH1 mRNA expression was significantly associated with the biomarkers related to cancer stem cells and basal type BC. However, there was no association between RAPH1, cancer stem cells and basal type BC at the protein expression level. Recent studies suggest that microRNA-203, which is considered to be associated with tumour invasion and transformation [40,41],

Disclosure of potential conflicts of interest
Takaaki Fujii has received research funding from Eisai Co, Ltd.
Any of the other authors declare that they have no conflict of interest.