Immunohistochemical analysis of cyclin A expression in Wilms tumor

Study design and patients

This retrospective study was conducted on tumor specimens and clinical data of patients with WT who underwent nephrectomy between January 1996 and October 2010. All specimens were obtained from the archives of Institute of Pathology, School of Medicine, University of Belgrade, Serbia (where the study was performed) and Mother and Child Health Care Institute of Serbia “Dr. Vukan Čupić,” Belgrade, Serbia. Of the available 59 samples, fourteen cases were excluded due to inadequate sampling or incomplete clinical data and two were excluded as bilateral tumors thus, the study was limited to 43 patients.

For each patient included in the study, available clinical–morphological data were collected by reviewing the patients’ medical histories. The data collected included sex, age (months), tumor stage, histological type, and prognostic group (Table 1). The mean age of patients was 52.4 months (range 7–132 months) and there were 28 (65.1%) females and 15 (34.9%) males. The histological diagnoses and classification were performed in accordance with the revised Societé International d’Oncologie Pediatrique (SIOP) Working Classification (Vujanić et al., 2002) criteria for histological subtypes (according to Section A) and strict adherence to the criteria for anaplasia (Vujanić et al., 1999). Based on the histological material, WTs were classified into following risk groups: (a) low risk (cystic partially differentiated WT), (b) intermediate risk (IR; regressive, epithelial, stromal, or mixed), and (c) high risk (blastemal or diffuse anaplasia) (Vujanić et al., 2002). Examination of prognostic significance of histologic components of WT characterized by a different degree of aggressiveness, but also different sensitivity to chemotherapy, provides a more precise classification of tumor subtypes within a group of medium risk and hence the adaptation of therapeutic protocols.

A total of 35 (81.4%) patients were treated according to SIOP protocols that include PrOpChTh for all patients older than 6 months, surgical operation, postoperative chemotherapy, and in some cases radiation therapy, as well as other medical standards and procedures in diagnosing and treating diseases (SIOP 9, SIOP 6, and some according to 93-01, depending on the period when the diagnosis was established). For each primary tumor, the percentage of regressive/necrotic changes in the total tumor mass was determined by a semiquantitative method, based on the macroscopic and microscopic examination. In the remaining viable tumor tissue, the proportion of epithelial, blastemal, and stromal components was determined microscopically. Seven (20%) cases were predominantly blastemal histological type, nine (25.7%) mixed, seven (20%) regressive, five (14.3%) stromal, and two (5.7%) predominantly epithelial. Among anaplastic tumors, three (8.6%) were diffusely anaplastic and two (5.7%) showed focal anaplasia. The revised SIOP staging criteria (Vujanić et al., 2002) was also used to determine tumor stage. Stage 1 was diagnosed in 15 (42.8%) WT cases, stage 2 in eight (22.8%), stage 3 in nine (25.7%), and stage 4 in three (8.6%) cases. According to the prognostic group, 25 (71.4%) cases were classified as intermediate (IR) group and 10 (28.6%) cases as high-risk group (HR).

Eight patients (18.6%) were not treated with PrOpChTh and they were analyzed separately. These cases were classified according to the Children’s Oncology Group criteria (blastemal type–IR, and not the high risk as it would be in SIOP). Five cases (62.5%) were predominantly blastemal histological type, 2 diffusely anaplastic (25%) and 1 (12.5%) was mixed histological type.

The follow-up period was 5 years and the OS time was defined as a period between date of surgery and date of death. Patients still alive at the end of the study were censored. Since according to the State Laws and Legislation patients are not required to receive therapy or to be followed up in the same clinic where they were diagnosed and operated, the event free survival could not be analyzed due to insufficient number of subjects to be included in the analysis.

The study was conducted in accordance with the ethical standards of the 1964 Helsinki Declaration and applicable national regulations. The ethical approval was received from Ethics Committee of School of Medicine, University of Belgrade, (1600/I-38). For this type of study formal consent is not required.

Methods

For each tumor, at least two paraffin-embedded tissue sections, representative of the global histology, were selected for IHC analysis of cyclin A. Tissue samples from formalin-fixed paraffin blocks and polyclonal rabbit antihuman antibody (H-432; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) were used. Deparaffinized samples were treated 15 min with hydrogen peroxide solution (3%) to stop peroxidase activity. Thereafter, samples immersed in 10 mM citrate buffer (pH 6.0) were treated in a microwave oven (20 min, 620 W) for antigen retrieval. Nonspecific staining was prevented with blocking peptide (DAKO, Glostrup, Denmark) the primary antibody for cyclin A (diluted 1:200) was applied and left overnight at 4 °C. Detection was done with streptavidin-biotin technique (Dako REALTM Detection Systems LSAB™+; DAKO, Glostrup, Denmark). Finally, chromogen (diaminobenzidine) was applied for IHC development and Mayer’s hematoxylin for the counterstaining. IHC stained normal lymph node tissue was used as a positive control, whereas negative control was based on the omission of primary antibody. Strong nuclear staining was considered positive for cyclin A1. The IHC expression of cyclin A was scored with semiquantitative technique (Table 1) as previously described (Todorovic et al., 2014). In normal kidney parenchyma adjacent to the tumor, focal expression of cyclin A was found in epithelial cells of distal convoluted tubules (Figs. 1A–1D) while the expression was absent in glomerular cells.

Consequently, moderate and diffuse expression in WT was considered as cyclin A overexpression. For statistical analyses, cases with no or focal expression and those with moderate or diffuse expression were grouped together, respectively.

Statistical analysis

Statistical analyses were performed at IC Faculty of Mechanical Engineering, Belgrade, Serbia, with SPSS 23.0 for Windows (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 4. Normality of distribution was tested with Shapiro–Wilks test. For testing the differences between parameters the Pearson χ² test, Fisher exact test and t-test were used. The OS was calculated by Kaplan–Meier method and groups were compared using Log-rank statistics. Univariate and multivariate Cox regression were used to analyze the independent factors related to OS. The p < 0.05 was considered significant.

Cyclin A expression in all components of WT

The overexpression of cyclin A (Table 2) was almost equal in preoperatively nontreated and PrOpChTh WTs (37.5% vs. 34.3%, respectively, p > 0.05, Fisher test).

In patients with PrOpChTh, tumors with cyclin A overexpression were larger (10.5 ± 4.06 cm) than those with low cyclin A level (7.52 ± 5.57 cm), (p = 0.012, t-test). Cyclin A overexpression was found more often in stages 3 and 4 (77.8%, 66.7%, respectively) than in stages 1 and 2 (13.3% and 12.5%, respectively), (p = 0.004, Pearson χ²), (Table 1). Regarding histological type, cyclin A overexpression was found most often in focal anaplasia, stromal, and diffuse anaplastic WTs (Table 2). No difference was found between diffuse anaplastic and other histological types (p = 0.266, Fisher test). In 8/25 (32%) cases of IR tumors and in 4/10 (40%) cases of HR tumors we detected cyclin A overexpression, but without difference (p = 0.706, Fisher test), (Table 1).

Cyclin A expression in individual components of WT

Epithelial and blastemal components of all WTs examined, overexpressed cyclin A with similar frequency (31.3% and 37.1%, respectively; p = 0.797), and more often than the stromal component (17.7%, p = 0.255 and p = 0.106, respectively; Fisher test), (Table 2). Epithelial component was not present in three cases. Similar to analysis of all components together, in epithelial component, we found no difference in cyclin A overexpression in relation to PrOpChTh (p = 0.225), tumor size (p = 0.088), and prognostic category (p = 1.000, Fisher test). Cyclin A overexpression was detected in 55.6% and 33.3% in stages 3 and 4, respectively and in 20% and 12.5% in stages 1 and 2, respectively (p = 0.187) (Table 1). There was no difference in cyclin A overexpression between diffuse anaplastic and other histological types (p = 0.224, Fisher test). In blastemal component, the cyclin A overexpression was more frequent in stages 3 and 4 (77.8% and 66.7%, respectively) compared to stages 1and 2 (13.3% and 25%, respectively, p = 0.009, Pearson χ²), (Table 1), whereas there was no difference in relation to PrOpChTh (p = 0.691), prognostic category (p = 0.444), and tumor size (p = 0.164, Fisher test). Overexpression of cyclin A related to histological type showed no difference between diffuse anaplastic and other histological types (p = 0.541, Fisher test), (Table 2). The stromal component was not present in one sample. The cyclin A overexpression (Table 1) revealed no difference in relation to stages 3 and 4 (22.2% and 66.7%, respectively) and stages 1 and 2 (6.67% and 12.5%, respectively), (p = 0.085, Pearson χ²). There was no difference in relation to PrOpChTh (p = 0.635), tumor size (p = 0.148), prognostic category (p = 1.000), and between diffuse anaplastic WT and other histological types (p = 0.442), (Table 2).

Cyclin A expression in primary WT and WTs with metastasis

Cyclin A overexpression in all components was more frequent in WTs with metastasis compared to WTs without metastasis (Table 1) but without difference (66.7% vs. 31.3%, p = 0.265, Fisher test). There was also no significant difference in frequency of cyclin A overexpression in individual components between WTs with and WTs without metastasis (Table 1). We did not detect cyclin A overexpression in only one case of primary tumor with metastasis, which was regressive histological type.

Survival analysis

The OS analysis (Fig. 2A) revealed that 97.1%, 71.4%, and 60% of PrOpChTh patients and 75%, 50%, and 25% of patients without therapy were alive after 1, 3, and 5 years, respectively (Log-rank 3.650, p = 0.056). The correlation was found for survival and stage (Log-rank 21.640, p = 0.000), and prognostic group (Log-rank 11.263, p = 0.001), (Figs. 2B–2C). After 5 years of follow-up, patients with stage 1 and 2 tumors had OS of 80% and 62.5%, respectively, while patients with stage 3 and 4 had OS of 44.4% and 0%, respectively. The rates of OS after 1, 3, and 5 years for patients with IR prognostic group were 100%, 84%, and 76%, respectively; patients with HR had survival rates of 90%, 40%, and 20% after 1, 3, and 5 years. In the group of patients with diffuse anaplastic WT, the OS was 0% after 3-year period in comparison to all other histological types (Log-rank 8.759, p = 0.003), (Fig. 2D). There was no difference in OS between patients related to gender (Log-rank 2.763, p = 0.096). In regard to OS and cyclin A expression, there was no correlation in all WT components (total expression), (Log-rank 3.160, p = 0.075), while in individual WT components the correlation was found only in blastemal component (Log-rank 5.035, p = 0.025). Patients who overexpressed cyclin A in blastemal component had the 5-year OS of 38.5%, (Figs. 3A–3D).

Univariate analysis (Table 3) showed that tumor stage (p = 0.001), prognostic group (p = 0.004), as well as cyclin A expression in blastemal component (p = 0.042), were significant prognostic factors for OS. After multivariate analysis, none of these factors remained statistically significant.