Expression of PLAGL2 in bladder urothelial carcinoma and its relationship to lymph node metastasis and survival

The purpose of this study was to investigate PLAGL2 expression associated with pathological features and prognosis and predicted lymph node metastases in the bladder urothelial carcinoma (BUC) tissue. The pathologic specimens and clinical data of 203 patients with bladder urothelial carcinoma after radical resection were collected. The expression of PLAGL2 was detected by immunohistochemically staining. The influence on lymph node metastasis and the prognoses of BUC patients were analyzed. The expression of PLAGL2 in BUC and positive lymph nodes was significantly higher than the normal bladder tissues (89.06% and 76.56% vs 21.88%, P < 0.001). Logistic regression analysis showed that PLAGL2 expression was an independent risk factor for BUC lymph node metastasis (P < 0.05). COX proportional hazards regression model showed that the time to recurrence and overall survival of patients with overexpression of PLAGL2 were significantly lower than those with low expression (P < 0.05). PLAGL2 is highly expressed in the BUC tissue and metastatic lymph node relative to the normal bladder tissue. This expression correlates to tumor size and number, and tumor grade and stage. Overexpression of PLAGL2 can be an independent predictor for lymph node metastasis and patient survival.

Each year nearly 430,000 patients are diagnosed with bladder cancer and more than 165,000 patients die from the disease in world-wide 1 . Bladder cancer has become the number one urinary tract cancer in China. The incidence is around 7.68/100,000 2 . Urothelial carcinoma accounts for more than 90% of the bladder cancer. The primary pathway for metastasis is via lymph node; often the first manifestation of bladder cancer is in the pelvic lymph nodes. Lymph node metastasis has a great impact on the treatment strategy and survival of the bladder cancer patients 3 . At present, diagnosis of pelvic lymph node metastasis is mainly based on pelvic MRI and/or CT images. The polymorphic adenoma gene-like protein 2 (PLAGL2) is located on the chromosome 20ql1.21. It contains the initiation codon ATG and the termination codon TAG but devoid of AATAAA polyadenylation signal. Its open reading frame encodes a protein comprising 496 amino acids. This protein contains six zinc finger structures, in an innate sequences in a carboxyl terminal that is rich in rich proline and serine 4 . Many studies have shown that PLAGL2 plays an important role in the development and progression of tumors [5][6][7] . Yang et al. 6 found that overexpression of PLAGL2 was associated with lung adenocarcinoma and that earlier stages of the disease had a lower expression of PLAGL2. Liu et al. 8 found that the expression of PLAGL2 in the gastric and colorectal cancer tissues was significantly higher than the expression of PLAGL2 in adjacent normal tissues. Moreover, the PLAGL2 expression was positively correlated to the depth of tumor penetration in colorectal cancer; Guo et al. 9 found that PLAGL2 expression in prostate cancer was significantly higher than PLAGL2 expression in benign prostatic hyperplasia tissue. In addition, they found that the extent of PLAGL2 expression in prostate cancer is closely related to the prognoses of patients. PLAGL2 expression can be used as an independent predictor for biochemical recurrence-free survival and OS. All these studies suggested that PLAGL2 plays a role in tumor penetration, metastasis and eventual patient survival. Although the expression and carcinogenic mechanisms of PLAGL2 have been extensively studied, the role of PLAGL2 in the development and progression of bladder

Relationship between PLAGL2 expression and pathologic parameters in patients with BUC.
The overexpression of PLAGL2 was positively correlated with tumor size, multiplicity of tumors, tumor grade, tumor cT, and LNS, (P < 0.05, Table 3). It had no significant relationship to age and sex (P > 0.05, Table 3). Strong staining of PLAGL2 was more common in the larger, multiple, high-grade, high stage, and positive lymph node tumor specimens (Fig. 2).
The effects of PLAGL2 expression and other clinicopathological parameters on pelvic lymph node metastasis of BUC. Based on subgroup analysis of lymph node metastasis, we found that the pelvic lymph node metastasis of BUC was significantly correlated with tumor size, multiplicity of tumors, tumor grade, tumor cT, and the level of PLAGL2 expression. (P < 0.05, Table 4). It was not correlated with the age and gender  of the patients (P > 0.05, Table 4). Multivariate logistic regression analysis suggested that PLAGL2 overexpression was an independent risk factor for lymph node metastasis of BUC (P < 0.05, Table 5).
Factors influencing recurrence and tumor-free survival rate of BUC patients at 5-year post operation. COX singular factor regression analysis showed that tumor size, tumor cT, lymph node metastasis, and the level of PLAGL2 expression were correlated to 5-year RFS of BUC (P < 0.05, Table 6). COX multivariate analysis suggested that lymph node metastasis and PLAGL2 overexpression resulted in a shorter RFS for BUC patients (P < 0.05, Table 6). The 5-year RFS curve of BUC was tested with Kaplan-Meier survival curve (Fig. 3). It showed that PLAGL2 overexpression had a significantly shorter RFS (P < 0.001).
Factors influencing the disease-free OS rate of BUC patients at 5-year post operation. COX single factor regression analysis showed that tumor size, cT, lymph node metastasis, and overexpression of PLAGL2 were correlated with disease-free OS of BUC (P < 0.05, Table 7). COX multivariate analysis showed that a high cT stage, lymph node metastasis, and overexpression of PLAGL2 were predictors of a shorter 5-year survival time (P < 0.05, Table 7). The 5-year OS curve of BUC was tested with the Kaplan-Meier curve (Fig. 4). It showed that PLAGL2 overexpression was associated with a 5-year survival time, which is significantly shorter than the survival time of patients with lower expression (P < 0.001).

Discussion
BUC is the most common malignant tumors of the urinary tract and about 30% of BUC are invasive tumors 10 .
Lymphatics is the most important pathway for metastasis. The presence of lymph node metastasis is a predictor   Table 3. Relationship between PLAGL2 expression and pathologic parameters in patients with BUC.     tumor tissues such as lung cancer, gastric cancer, colorectal adenocarcinoma, and prostate cancer 6,8,9 . In our study, the positive rate of PLAGL2 expression in the bladder cancer tissue was 83.3%; of which weak positive rate (+) was 20.2%, intermediate positive rate (++) was 36.0%, and strong positive rate (+++) was 27.1%. On the other hand, the positive rate of PLAGL2 expression was 21.88% in the 64 normal bladder tissues; all of which was either weak positive or intermediate positive. This difference was statistically significant. We have thus confirmed for the first time the overexpression of PLAGL2 in the BUC. We also showed that the overexpression of PLAGL2 is correlated with tumor size, multiplicity of tumors, pathological grade, tumor cT, and lymph node metastasis. COX multivariate analysis suggested that lymph node metastasis and overexpression of PLAGL2 are important factors for diminishing the 5-years postoperative RFS and OS of the BUC patients. Braud et al. 12 found that the survival time of patients with regional metastasis after radical cystectomy and pelvic lymphadenectomy was 20% lower than those without. Karakiewicz et al. 13 reported that tumor grade, tumor cT, and the presence of carcinoma in situ are independent predicators of lymph node metastasis and ultimately the survival of the patients.
The cT, pathological grade, and lymph node metastasis of bladder cancer are important predictors for prognoses 14,15 . Hensen et al. 7 have shown that the overexpression of PLAGL2 could cause tumor formation. PLAGL2 overexpression was correlated with the typical manifestations of malignant transformation. Abdollahi 16 found that the overexpression of PLAGL2 not only caused acute granulocytic leukemia in mice, but was also found in 20% of acute myeloid leukemia patients 16 . Other studies have shown that PLAGL2 was highly expressed in a variety of tumors and promoted tumor progression and metastasis 6,17 . Bladder urothelial carcinoma (BUC) with lymph node metastasis often denotes a poor prognosis. In clinical practice, the existence of lymph node metastasis prior to surgical resection relies on MRI and/or CT studies. However, MRI or CT studies are often equivocal in identifying small lymph node metastases or lymphadenopathy caused by inflammatory conditions. Our study found that the risk of lymph node metastasis in PLAGL2-positive bladder cancer patients was 1.661 times higher than that of PLAGL2-negative bladder cancer patients. Thus PLAGL2 overexpression can be an independent predictor for lymph node metastasis.
The relationships of PLAGL2 expression with lymph node metastasis and patient survival have great clinical significance. Currently pelvic lymphadenectomy is routinely performed with radical cystectomy for high-risk bladder tumors. Some experts advocated extended pelvic lymph node dissection to further improve prognoses 18 . However, extending lymphadenectomy may lead to longer operations and increased morbidities. If there is a reliable preoperative predictor for lymph node metastasis, this predictor could provide guidance for the pelvic lymph node dissection. The results of our study showed that PLAGL2 could be used as an independent predictor for pelvic lymph node metastasis. Furthermore, PLAGL2 could be used as a predictor for recurrence-free and overall 5-year survival of bladder cancer patients.
Cancer biomarkers are indicators that can objectively evaluate and indicate biological processes, pathological processes, and pharmacological responses to interventions, however, clinically useful cancer biomarkers remain rare, because cancer is a complex disease 19 . Thus far, no effective biomarkers have yet been found in bladder cancer. Our results show that PLAGL2 can be used as a cancer hallmarks for bladder cancer. And cancer hallmarks can enable tumor growth and metastasis dissemination, the collection of these hallmarks could help in understanding of the mechanisms by which invading cells give rise to recurrent tumors and the effect of adjuvant therapeutics have on their evolution which will facilitate the development of new strategies to delay or prevent recurrence and malignant tumor progression 20 . Through the detection of cancer hallmarks, it is possible to judge the course of the disease, predict the survival of the patient, as well as the efficacy and prognosis. And the continuous exploration of cancer hallmarks and the gradual realization of individualization of treatment have become the current research hotspots 21 .
The weaknesses of this study include the following. First, it was a retrospective study, so it had inherent bias. Second, the comparison of the tissues were all from patients with BUC rather than normal patients. Comparison with normal bladder tissue from patients without BUC would be more precise. However, this approach would not be practical. The specific mechanism of PLAGL2 is still unknown. Further work is needed to elucidate the role and function of PLAGL2 in tumor genesis.

Conclusion
PLAGL2 expression can be used as an independent predictor for pelvic lymph node metastasis in BUC. Overexpression of PLAGL2 was found to correlate with tumor size, number, grade, and stage. PLAGL2 expression can serve as predictor for patients' 5-year RFS and OS after radical cystectomy with pelvic lymphadenectomy.

Materials and Methods
Tissue specimen and data collection. 203 consecutive surgical specimens from bladder cancer patients after radical cystectomy with standard lymph node dissections were collected from the Zhuzhou Central Hospital from January 2007 to January 2011, (The extent of standard lymph node dissection is templated as following: lateral to femerogenital nerve, medial to obturator nerve, superior extent at iliac vessel bifurcation, inferior extent at the distal iliac vein.) and all patients were informed consent for study participation. Specimens were fixed with 4% formaldehyde and embedded in paraffin. Pathological examination confirmed that tumors were BUC. Patients who received radiotherapy and chemotherapy before surgery were excluded. In addition, 64 normal bladder sections and 64 positive lymph nodes were selected as comparison groups. Clinical data collected included the age, gender, tumor size, multiplicity of tumors, tumor grade, tumor clinical stage (cT), and lymph node status (LNS) ( Table 1). Construction of tissue microarray. The paraffin-embedded specimens of BUC, positive lymph nodes, and normal bladder tissues were sectioned for hematoxylin and eosin (H&E) staining. Diagnosis was confirmed and representative areas of the tissues were identified for microarray processing. Two 2 × 2 mm tissue cores from each specimen were taken from the corresponding paraffin blocks. Total 5 × 8 tissue cores were placed into each of the receptacle blocks. The microarray blocks were then sliced to a thickness of 4 μm.
Immunohistochemistry. These sections were conventionally de-waxed and rehydrated. Antigen retrieval was performed using a high-pressure cooker at 95 °C in 0.1 M citric acid buffer (pH 5.0 Fuzhou Maixin Biotech. Co., Ltd., Fuzhou, China). The tissue microarrays were then incubated with 3% H 2 O 2 for 10 minutes to block endogenous peroxidase activity, followed by normal goat serum at room temperature. Next, the tissue microarrays were incubated with a rabbit monoclonal anti-PLAGL2 antibody (primary antibody, Abcam Company, UK) at a dilution of 1:800 at 4 °C overnight followed by incubation with DAKO ChemMate EnVision (secondary antibody, DAKO Company, USA) at 37 °C for 30 minutes. The sections were subjected to staining with 3,30-diaminobenzidine (DAB) solution and briefly counterstained with hematoxylin. Last, the sections were dehydrated with ethanol, mounted with a cover slip, and examined with a microscope.
Judgment of the results. The immunologically stained microarray sections were blindly scored by two senior pathologists. Positive expression of PLAGL2 was identified by the appearance of brown granules in cytoplasm and nucleus. Five representative high-power fields from each slide were examined. 100 cells in each field were evaluated. The proportion of positive stained cells and staining intensity were scored. The following criteria were used: ① Staining intensity (I) was recorded as: no staining, 0; weak staining intensity (light yellow or only individual cells were yellow to brown stained), 1; medium staining intensity (between the first two), 2; and strong staining intensity (yellow to brown stained), 3. ② The proportion (P) of positive cells was recorded as 0, <5%; 1, 5% to 25%; 2, 26% to 50%; 3, 51% to 75%; and 4, >75%. For total points, ① + ②, 0 is negative (−), 1-2 is weak positive (+), 3-5 is positive (++), 6-7 is strong positive (+++) 22 . Any discrepancy in the scoring was resolved by discussion between the pathologists. And all methods were performed in accordance with the relevant guidelines and regulations.
Statistical methods. Statistical analysis was performed using the SPSS 22.0 statistical software package.
Qualitative data were analyzed by independent sample chi-square test or by the Fisher exact test. The predictive value of lymph node metastasis was investigated using Logistic regression analysis. The relationships of PLAGL2 expression with the 5-year recurrence-free survival (RFS) and the overall survival (OS) of BUC were estimated by Kaplan-Meier survival curve and Log-rank statistical method. The singular and multivariate analyses of the 5-year RFS and the OS of BUC were evaluated using the COX proportional hazards regression model. P < 0.05 was considered to be statistically significant.