Combined Expression of Metastasis Related Markers Naa10p, SNCG and PRL-3 and its Prognostic Value in Breast Cancer Patients

Combinations of multiple biomarkers representing distinct aspects of metastasis may have better prognostic value for breast cancer patients, especially those in late stages. In this study, we evaluated the protein levels of N-α-acetyltransferase 10 protein (Naa10p), synuclein-γ (SNCG), and phosphatase of regenerating liver-3 (PRL-3) in 365 patients with breast cancer by immunohistochemistry. Distinct prognostic subgroups of breast cancer were identified by combination of the three biomarkers. The Naa10p+SNCG-PRL-3- subgroup showed best prognosis with a median distant metastasis-free survival (DMFS) of 140 months, while the Naa10p-SNCG+PRL-3+ subgroup had the worst prognosis with a median DMFS of 60.5 months. Multivariate analysis indicated Naa10p, SNCG, PRL-3, and the TNM classification were all independent prognostic factors for both DMFS and overall survival (OS). The three biomarker combination of Naa10p, SNCG and PRL-3 performed better in patients with lymph node metastasis, especially those with more advanced tumors than other subgroups. In conclusion, the combined expression profile of Naa10p, SNCG and PRL-3, alone or in combination with the TNM classification system, may provide a precise estimate of prognosis of breast cancer patients.


Introduction
Breast cancer is not only the most frequently diagnosed cancer in women but also the leading cause of cancer death among females worldwide (Jemal et al., 2011). Breast cancer accounts for 29% of the total cancer cases and 14% of the cancer deaths (Jemal et al., 2011;Siegel et al., 2012). Recently, most breast cancer of early stages can be cured by radical mastectomy, and nearly 90% breast cancer patients can survive more than 5 years with multimodality treatment (Holmes et al., 2010;Saini et al., 2011). However, in developing countries, death rate of breast cancer is still above 40%, and most of the deaths were caused by metastasis (Jemal et al., 2011;Rahimzadeh et al., 2014). Micrometastasis is hard to predict by all the current prognosis means alone, which makes patients with the same clinical stages always show different survival time (Cox et al., 2011;Jafferbhoy and McWilliams, 2012), suggesting that novel biomarkers and updated staging procedures should be developed to provide precise cancer prognosis. In this study, we proposed that combining markers representing different aspects of breast cancer metastasis could provide a better prognostic value. Li Min 1 , Ruo-Lan Ma 2 , Hua Yuan 3 , Cai-Yun Liu 1 , Bing Dong 4 , Cheng Zhang 1 , Yan Zeng 1 , Li Wang 1 , Jian-Ping Guo 1 , Li-Ke Qu 1 , Cheng-Chao Shou 1 * N-α-acetyltransferase 10 protein (Naa10p), the catalytic subunit of N-acetyltransferase A (NatA), is involved in cell cycle (Lim et al., 2006), proliferation (Lim et al., 2006;Seo et al., 2010), apoptosis (Gromyko et al., 2010), autophage (Kuo et al., 2010), cell motility (Bauer et al., 2009;Hua et al., 2011), neuron development (Ohkawa et al., 2007;Ohkawa et al., 2008) and 28S proteasome activity (Min et al., 2013). Naa10p is overexpressed in various types of cancer Hua et al., 2011), and its overexpression in breast cancer is negatively correlated with metastasis and indicates good prognosis through different molecular pathway (Bauer et al., 2009;Hua et al., 2011;Zeng et al., 2013).
Naa10p, SNCG and PRL-3 are three critical factors involved in the regulation of breast cancer carcinogenesis. According to the recent information, these three proteins function through different molecular mechanisms in the regulation of breast cancer metastasis. Combination of these three biomarkers and an integrated evaluation of them may represent a more comprehensive view of breast caner, which could provide an opportunity to investigate and explore a modified staging system for breast caner patients. In this study, we investigated the possible correlations of these three proteins with clinical outcome, in an effort to identify high metastasis risk patients with breast cancer to make better treatment options for them.

Patients
Breast cancer tissue specimens were obtained from 365 women having breast surgery (radical or modified radical mastectomy) at Peking University Cancer Hospital and Institute between 1996 and 2002. The patients' age ranged from 25 to 81 (with a median of 50 years). The tumors were staged based on Union Internationale Contre Le Cancer (UICC) TNM classification: 141 Stage I tumors,  Table 1.

Immunohistochemistry
Immunohistochemical staining procedure was described in our previous paper (Wang et al., 2008). All paraffin-embedded specimens were cut into 5μm sections. After baking at 60℃ overnight, sections were dewaxed and rehydrated through xylene and an alcohol series. Thereafter, antigen retrieval was performed via microwave cooking in ethylene diamine tetraacetic acid (pH 8.0, Zymed) for 20 min. Endogenous peroxidase activity was blocked by incubation in 3% hydrogen peroxide for 10 min at room temperature. Non-specific binding was blocked with 10 % goat serum. After that, anti-Naa10p, Anti-PRL-3 and anti-SNCG monoclonal antibodies, which were prepared in our laboratory and the quality, specificity, and sensitivity had been determined in Refs. (Zeng et al., 2013), (Wang et al., 2008) and (Guo et al., 2007), respectively, were applied to each slide and incubated at 4℃ overnight. After 3 washes with phosphate-buffered saline with 0.1% Tween-20, the specimens were incubated with second antibody from the EnvisionTM kit (Dako Cytomation, Cambridge, UK) for 45 min at room temperature. The reaction product was visualized with diaminobenzidine (Sigma) for 5 min at room temperature and the sections were counterstained with hematoxylin. Normal mouse IgG was used as a negative control of the primary antibody.

Evaluation of immunohistochemical staining
The results were evaluated under light microscopy (APPLIED IMAGING at 200×) independently by two experienced pathologists without prior knowledge of the clinical information. The discrepancies (<5%) were resolved by simultaneous re-evaluation. We assessed both the percentage of positive cells and the intensity of staining in 10 randomly chosen microscopic field. A semiquantitative scoring system in tumor cells was graded according to a 4-value classification scale as follows: area of staining as <10% of all cancer cells stained within the section was graded as none (0), staining intensity (>10% of all cancer cells stained within the section) was graded as weak (1), moderate (2) or strong (3). The immunohistochemical evaluation of Naa10p, SNCG and PRL-3 were presented as either "negative" or "positive", in which score of ≥2 was defined as "positive".

Statistical analysis
Correlations between protein expression levels and patient clinicopathologic characteristics were performed using Pearson χ 2 test. The Kaplan-Meier method was used to estimate DMFS and OS rates, and the survival differences were tested by log-rank method. The Cox proportional hazard model was used for multivariate analysis to investigate the independence of the risk factors identified as significant in the univariate analysis. Hazard ratios (HR) and 95% confidence interval (CI) >1.0 indicate that positive expression is associated with a poor prognosis while <1.0 indicate that positive expression is associated with a good prognosis. Survival analysis stratified by TNM classification and lymph node metastasis status were also conducted. All statistical analyses were 2-sided, and comparisons made in which p values less than 0.05 were considered statistically significant. All statistical analyses were performed using SPSS for Windows Software (version 13.0).
The associations between these factors and distant metastasis were also analyzed. As expected, clinicopathologic features, including TNM stage (III/IV, II versus I; p<0.001), lymph node metastasis (p<0.001) and tumor size (p<0.001) were significantly related with distant metastasis of breast cancer, whereas age, ER, PR and HER2 status did not affect distant metastasis (p>0.05; Table 1). Among the 3 molecular markers, Naa10p significantly negatively correlated with the presence of distant metastases (p=0.004), and SNCG significantly positively correlated with the presence of distant metastases (p<0.001). PRL-3 positive subgroup showed a HR of 1.557 compared to the PRL-3 negative subgroup, but the p value was not statistically significant (p=0.051). The correlation with three markers with metastasis was reasonable considering their molecular function in cancer cell motility (Hibi et al., 2009;Peng et al., 2009;Zeng et al., 2013). Consistent with the above, multimarker phenotype of Naa10p/SNCG (p<0.001) and of Naa10p/SNCG/PRL-3 (p<0.001) also correlated with the presence of distant metastases (Table 1).

Prognostic Value of Multimarker System for Patients With Breast Cancer
The prognostic value of the Naa10p, SNCG and PRL-3 alone and multimarker combination was evaluated.
For the evaluation of OS, Kaplan-Meier curves and log-rank test were also performed and the results are very similar to those of DMFS (Figure 2). Naa10p positive patients showed a significantly longer OS time than the Naa10p negative group (Figure 2A; log rank χ 2 =10.05; p=0.002), while SNCG positive patients had a significantly shorter DMFS time than those with SNCG negative tumors ( Figure 2B; log rank χ 2 =26.23; p<0.001). PRL-3 positive subgroup also displayed a shorter OS than PRL-3 negative subgroup ( Figure 2C; log rank χ 2 =5.79; p=0.016). When combined Naa10p with SNCG, Naa10p+SNCG-subgroup showed the best prognosis while Naa10p-SNCG-showed the worst ( Figure 2D; log rank χ 2 =50.09; p<0.001). Combining all three markers, Naa10p+SNCG-PRL-3subgroup displayed the longest median OS ( Figure 2E; log rank χ 2 =29.47; p<0.001). The new three markers classification system performed better than the traditional biomarker classification system (ER/PR, HER2 and TN; Figure 2F; log rank χ 2 =0.66; p=0.719) in OS prediction. Univariate Cox analysis was performed and its results were showed in Table 2 (Table 3, middle column). Combined Naa10p/SNCG and combined Naa10p/SNCG/PRL-3 were both independent indicators only in univariate analysis, but not in multivariate analysis. Similar result was obtained in the analysis of OS (Table 3, right column).

Prognostic Value of Multimarker System for Patients with Breast Cancer Stratified by Lymph Node Metastasis
Subsequently, the prognostic value of the Naa10p/ SNCG/PRL-3 three biomarkers classification system was evaluated in the tumors stratified into subgroups with (LN+) or without (LN-) lymph node metastasis.

Prognostic Value of Multimarker System for Patients With Breast Cancer Stratified by and TNM classification
To explore whether our new classification system has the same prognosis value in patients with different TNM classification, all 365 patients were stratified in to Stage I subgroup (n=141), Stage II subgroup (n=128) and Stage For patients of stage I and stage II, patients with Naa10p+SNCG-PRL-3-tumors had a trend of longer DMFS compared with other subgroups, but did not reach a statistical difference ( Figure 4A, Stage I, log rank χ 2 =0.696; p=0.706; Figure 4B, Stage II, log rank χ 2 =4.885; p=0.087). And for Stage III/IV patients, Naa10p+SNCG-PRL-3-subgroup exhibited a significant longer DMFS than others ( Figure 4C, long rank χ 2 =15.034; p=0.001). In regard of OS, our new classification system also performed better in Stage III/IV subgroup than in Stage II and Stage I subgroups ( Figure 4D, Stage I, long rank χ 2 =2.021, p=0.364; Figure 4E, Stage II, long rank χ 2 =2.186, p=0.335; Figure 4F, long rank χ 2 =14.367, p=0.001

Discussion
Metastasis is the main cause of death in patients of breast cancer, and micrometastasis, the initial stage of metastasis has no reflection on patient's symptoms (Cox et al., 2011). Recently, micrometastasis is still hard to predict by any cancer biomarkers alone, resulting that patients at the same clinical stages often have different outcome (Jafferbhoy and McWilliams, 2012). In the current study, we have analyzed three molecules regulating cancer metastasis by different mechanisms, aiming to provide a more precise prediction of metastasis in breast cancer patients than the traditional molecular classification system consisting of ER, PR and HER2.
The components of our three biomarkers classification system are all involved in cell adhesion, motility and   DOI:http://dx.doi.org/10.7314/APJCP.2015.16.7.2819 Prognostic Value of Combined Expression of 3 Markers in Breast Cancer migration (Wang, 2006;Guo et al., 2007;Zeng et al., 2013). As showed in Figure 5, many of the molecules associated with Naa10p, SNCG and PRL-3 are metastasis-related proteins. Several key molecules, such as HIF-1α, integrin α1, integrin β1, RAC1 and myocilin, were identified to be associated with the three biomarkers by STIRNG analysis (Franceschini et al., 2012). Even though there are some joint protein associated with 2 or 3 of Naa10p, SNCG and PRL-3, most of key molecules identified are just related to only one of the three biomarkers, meaning that these three biomarkers may represent different aspects of cancer metastasis mechanism.
Detailed information of published materials also supported this opinion. Though still under dispute, Naa10p was considered to have a very close relationship with HIF-1α and involved in angiogenesis (Arnesen et al., 2005;Bilton, 2005;Jeong, 2002;Murrayrust et al., 2006). Naa10p also interacts with MLCK to repress cell motility (Bauer et al., 2009), interacts with PIX-β to inhibit RAC1 pathway and to suppress metastasis of breast cancer cell (Hua et al., 2011). MMP2 and MMP9 were regulated by Naa10p (Zeng et al., 2013). Taken together, Naa10p had a negative role in breast cancer metastasis through several molecular pathways. For SNCG, which is involved in cell adhesion, invasion and metastases (Hibi et al., 2009;Pan et al., 2006), the molecular mechanism remained obscure. SNCG was also identified as an interacting protein of myocilin, which is believed to have a role in cytoskeletal function (Surgucheva et al., 2005). Therefore, it is not surprising that increased SNCG levels correlated with the presence of distant metastasis and unfavorable outcome. PRL-3 was found to promote cancer cell motility, invasion, and metastasis through integrin β1-ERK1/2 and-MMP2 signaling (Ming et al., 2009;Peng et al., 2009;Vairaktaris et al., 2008). Based on these findings, we concluded that evaluation of all the three biomarkers expression level may provide a comprehensive view of different aspects of cancer metastasis mechanism, and may have a potential value in prognosis of breast cancer metastasis.
In this study, we revealed that both Naa10p, SNCG, PRL-3 alone and multimarker phenotype of Naa10p/ SNCG and of Naa10p/SNCG/PRL-3 were correlated with the presence of distant metastases, which also reflected in the postoperative survival time (DMFS and OS). Additionally, according to our results, molecular staging with three biomarkers (Naa10p, SNCG and PRL-3) identifies patient prognosis more accurately than the traditional clinical staging, particularly for patients with lymph node metastasis. The three biomarkers classification system also explained differences in the outcome of breast cancer patients belonging to the same TNM group, especially patients of Stage III and Stage IV. The reason why our new classification system performed better in patients with more advanced tumor were probably that all the three biomarkers were mainly conducive in the process of metastasis, which almost only occurred in cancer of late stage.
In conclusion, our results suggested that combined expression profile of Naa10p, SNCG and PRL-3, alone or in combination with TNM classification system, has better prognostic value for patients with breast cancer than the traditional molecular typing system, especially in patients with lymph node metastasis. The new multimarker system seemed to be promising in discriminating good from poor prognostic patients, with the potential to provide information for adjuvant therapy choice and, possibly, in establishing a more personalized prognosis model for each breast cancer patient.