The correlation analysis of WWOX expression and cancer related genes in neuroblastoma — a real time RT-PCR study

Neuroblastoma is one of the most common paediatric cancers, described as unpredictable due to diverse patterns of behaviour. WWOX is a tumour suppressor gene whose expression is reduced in many tumour types. Loss of its expression was shown to correlate with more aggressive disease stage and mortality rate. The aim of this study was to investigate the role of the WWOX tumour suppressor gene in neuroblastoma formation. We performed real-time RT-PCR to analyse levels of WWOX expression in 22 neuroblastic tumour samples in correlation with genes involved in cell cycle regulation (CCNE1, CCND1), proliferation (MKI67), apoptosis (BCL2, BIRC5, BAX) and signal transduction (EGFR, ERBB4). We also evaluated two potential mechanisms — promoter methylation (MethylScreen method) and loss of heterozygosity (LOH) status, which could be connected with regulation of WWOX gene expression. We found a positive correlation between WWOX gene and BCL2 and HER4 JM-a and negative with cyclin D1 and E1. Our observations are consistent with previous findings and emphasise the role of WWOX in cell cycle and apoptosis regulation. Moreover, strong positive association with HER4 JM-a in this tumour type may indicate a role for WWOX in neuroblastoma cell differentiation. The presented results indicate that LOH in locus D16S3096 (located in intron 8) may be involved in the regulation of WWOX mRNAexpression. However, no association between methylation status of WWOX promoter and its expression was observed.


INTRODUCTION
Neuroblastoma is one of the most common paediatric cancer (Maris & Matthay, 1999;Heck et al. 2009;Kaatsch, 2010).This malignant tumour is an embryonic cancer consisting of undifferentiated neuroectodermal cells derived from the neural crest (Schwab et al., 2003).In most cases neuroblastoma arises in the adrenal glands, but can also appear in the abdomen, chest or pelvis (Heck et al., 2009).This type of tumour is often characterized as enigmatic or unpredictable due to contrasting patterns of behaviour from life-threatening progression, development to ganglioneuroblastoma or ganglioneuroma to spontaneous regression (Schwab et al., 2003).Neuroblastoma classification is still based on age, histo-logic stage and Shimada pathology (Shimada et al.;1999;Schwab et al., 2003).The most common molecular marker of neuroblastoma pathology is MYCN amplification which is regarded as a prognostic factor in determining treatment.Moreover, chromosomal abnormalities, including chromosome gain or LOH within chromosome 1p, 17q appear quite frequently.Other chromosomes affected in neuroblastoma formation include chromosomes 2q, 3p, 4p, 11q, 14q, 16p and 19q (Schwab et al., 2003).
WWOX is a tumour suppressor gene located on the long arm of chromosome 16 (Bednarek et al., 2000).This area is known as a common fragile site FRA16D, which is affected in many cancers (Chen et al., 1996;Latil et al., 1997;Bednarek et al., 2000).Genomic alteration within the WWOX region and its differential expression has been found in variety of tissues and tumour types (Kuroki et al., 2002;Aqeilan et al., 2004;Guler et al., 2004;Kuroki et al., 2004).The reduction of WWOX expression in many cancers was found to correlate with more aggressive disease stage and higher mortality rate ( breast, gastric, bladder, lung cancer) (Nunez et al., 2005;Pluciennik et al., 2006;Aqeilan et al., 2007;Ramos et al., 2008;Maeda et al., 2010).
It was previously shown that ectopically induced WWOX overexpression in different cell lines (breast, lung) promoted apoptosis, suppressed anchorage-independent growth and inhibited colony formation in Matrigel (Bednarek et al., 2001;Kuroki et al., 2004;Fabbri et al., 2005;Qin et al., 2006;Lewandowska et al., 2009;Xiong et al., 2010;Zhou et al., 2010).WWOX as a partner of several transcription factors participates in controlling expression of genes which are responsible for tissue morphogenesis and cell differentiation (Ramos et al., 2008;Aqeilan et al., 2008;Lewandowska et al., 2009).More recently, Gourley and coworkers showed that WWOX restoration or its overexpression in an ovarian cell line resulted in decreased attachment and reduced cell migration on fibronectin, which as an ECM component is associated with peritoneal metastasis (Gourley et al., 2009).Several studies on animal models not only confirmed the role of WWOX in tumourigenesis (Aqeilan et al., 2007;Ludes-Meyers et al., 2007;Aqeilan et al., 2008;Ludes-Meyers et al., 2009), but also revealed its potential role in steroidogenesis and proper gonadal function (Aqeilan et al., 2009).
The aim of our study was to evaluate the role of WWOX in neuroblastoma carcinogenesis.We examined WWOX expression, methylation status and the frequency of LOH within its genomic region in neuroblastoma samples and analyzed the correlation between its expression level and genes involved in cell cycle regulation (CCNE1, CCND1), proliferation (MKI67), apoptosis (BCL2, BIRC5, BAX) and signal transduction (EGFR, ERBB4).The study included 22 neuroblastic tumour samples: 18 neuroblastoma (8 described as poorly differentiated) and 4 ganglioneuroblastoma.The group consisted of 12 males and 10 females.There were 8 patients aged under one year, and the remaining group were patients older than one year old (mean age 2.86).According to clinical records 5 out of 22 tumours were metastatic.MYCN status was also determined, as well as stage, risk of tumour development and prognosis according to INPC.

Patients
More detailed information is shown in Table 1.This study was conducted after receiving patients' family consent and approved by the Institutional Bioethics Committee RNA, DNA isolation and cDNA synthesis.RNA was isolated from neuroblastic tissues stored at -80°C in RNAlater (Ambion), using TRIzol reagent (Invitrogen).10 μg of total RNA was used to cDNA synthesis, at the total volume of 100 μl with ImProm RT-II reverse transcriptase (Promega).The conditions of reverse transcription were as follows: 5 min incubation at 25°C, and 60 min at 42°C, heating at 70°C for 15 min.The synthesised cDNA was diluted with sterile deionised water to 150 μl and 2 μl of cDNA was used in a PCR reaction.
DNA was recovered using back extraction buffer (BEB, 1 M Tris Base, 4 M guanidinium thiocyanate, and 50 mM sodium citrate) from organic remains of TRIzol after RNA isolation and was performed according to the manufacturer's protocol.
Real-time RT-PCR analysis.Real-time RT-PCR was performed with Light Cycler 480 II (Roche).SYBR Green I and qPCR Core Kit for SYBR Green I (Eurogentec) was used to detect PCR products.Reactions were performed in duplicate.We analyzed relative expression of 11 genes (BAX, BCL2, BIRC5, EGFR, MKI67, WWOX, HER4 (isoforms JM-a and JM-b), CCNE1, CCND1, CDH1).The expression level of the studied genes was normalized to two reference genes (RPLP0, RPS17).
As the level of WWOX mRNA was relatively low, we performed a semi-nested RT-PCR for assessing the WWOX expression level.In the first PCR reaction we used primers: 5'-TGCAACATCCTCTTCTC-CAACGAGCTGCAC-3' and 5'-TCCCTGTTGCATG-GACTTGGTGAAAGGC-3' in 50 μl volume.Next, 2 μl of 200-fold diluted PCR product (171 bp) was a template for semi nested PCR.The cycling protocol included 2 min at 94°C, 30 s denaturation at 94°C, 30 s annealing at 63°C, 1 min extension at 72°C, repeated for 77 cycles, and additional extension for 7 min at the same temperature as previous step.
Sequences of used primers and the PCR reaction conditions, as well as the length of received products are presented in Table 2.
Relative expression level was calculated according to Roche algorithm (Pfaffl et al., 2002), which includes the differences in efficiency and crossing point (Cp) of each sample versus calibrator sample and normalize the value to a reference gene.
The Universal Human Reference RNA (Stratagene) was used as a calibrator for each reaction.
To avoid amplification of genomic DNA, all primers were designed to be intron spanning.Detection temperature was designated above the non-specific/primer-dimer melting temperature.
LOH analysis.Loss of heterozygosity was analysed with high resolution melting (HRM) of Light Cycler 480 (Roche).We used two microsatellite markers D16S518 and D16S3096 located on chromosome 16; on intron 1 and intron 8 of WWOX gene, respectively.The primer sequences were obtained from the Genome database.PCR conditions were as follows: initial denaturation 95°C for 10 min; 35 cycles of repeated denaturation at 94°C for 30 s, annealing at 56°C (for D16S3096) or 55°C (for D16S518) for 30 s, elongation at 72°C for 60 s.
Analysis of WWOX methylation status.We performed a MethylScreen assay which is based on a set of restriction digestions and subsequently performed Realtime PCR according to (Holemon et al., 2007).We analysed two fragments which comprise the promoter and the first exon of the WWOX gene.
For the first analysed promoter fragment of the WWOX gene, the PCR mix included 2.5 µl of SYBR-Green I, qPCR Core kit for SYBR Green I reagents (Eurogentec), 10 nM of each primer, 4 µl of digested DNA template.Real-time PCR was conducted at the following conditions: 95°C for 5 min, followed by 50 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 90 s, and 77°C for 15 s (additional temperature for reading only specific amplification product size 384 bp).The PCR mix for the second WWOX promoter fragment consisted of 4 µl of digested DNA template, 10 nM of each primer, 2.5 µl of SYBR Green I, qPCR Core kit for SYBR Green I reagents, 70 mM betaine.Real-time PCR were performed at the following conditions: 95°C for 5 min, 50 cycles at 94°C for 30 s, 50°C for 30 s, and 72°C for 90 s, and 83°C for 15 s (additional temperature for reading only specific amplification product size 413 bp ).All reactions were performed in duplicate.
Statistical analysis.Non-parametric Spearman Rank Correlation test was used to analyze correlations between expression of WWOX and other genes.Student's/ Aspin-Welsch t-test was used to estimate differences between WWOX expression in relation to LOH, methylation status and clinical factors.All results were assumed as statistically significant at the confidence level >95% (p<0.05).

RESULTS
Methylation of the WWOX promoter (on both fragments) was observed only in 7 out of 22 neuroblastoma samples and had no influence on WWOX gene expression in our sample population.
Loss of heterozygosity for locus D16S3096 was observed in 50% (11 samples) and for microsatellite D16S518 in 16.6% (3 samples) of neuroblastoma samples.The information about observed LOH and comparison with populational homozygosity according to Genome databases are presented in Table 3.Moreover, an analysis of influence of LOH on WWOX gene mRNA level revealed a tendency to reduction of its expression via this mechanism (p>0.05).In D16S3096, mean WWOX gene mRNA level was 16.6 and 13.3 for heterozygous and homozygous samples, respectively (p>0.05).Due to the low number of homozygous samples in the second examined locus D16S518 the comparison between mean expression level had no statistical legitimacy.
We found a positive, statistically significant correlation between the WWOX expression level and antiapoptotic BCL2 gene (R s =0.6838, p=0.0005) and HER4 isoform JM-a mRNA (R s =0.7165, p=0.0002).Negative significant correlations were observed for WWOX mRNA and cyclins CCND1 (R s =-0.4671, p=0.0284), and CCNE1 (R s =-0.4884, p=0.0211).We did not find any statistically significant association between the WWOX expression level and the expression of proapoptotic BAX gene, nor with the BCL2/BAX ratio.Moreover, we did not find any correlation between the examined suppressor gene and proliferation marker MKI67, as well as with the other investigated genes.More detailed information on correlation levels is presented in Table 4 and correlation plots are presented on Fig. 1.There were no significant correlations between expression of investigated WWOX and clinical factors such as age, stage and histology (favourable/unfavourable) or MYCN status.However, we noticed elevated mean WWOX expression in association with INPC prognostic category and in accordance with tumour type.Detailed information is presented in Table 5.

DISCUSSION
WWOX is a tumour suppressor gene with decreased expression in many tumour types (Yendamuri et al.,   2003; Aqeilan et al., 2004;Nunez et al., 2005;Nunez et al., 2005;Dias et al., 2007;Ramos et al., 2008;Maeda et al., 2010).One of the possible mechanisms inactivating its expression is loss of heterozygosity and methylation of CpG islands within the promoter region.LOH of WWOX locus has been found in few tumour types like Wilm's tumours (Skotnicka-Klonowicz et al., 2000), prostate (Carter et al., 1990), breast (Bednarek et al., 2000), gastric (Aqeilan et al., 2004), oesophageal (Kuroki et al., 2002) and pancreatic (Kuroki et al., 2004) cancer.In our study, we observed LOH on both examined microsatellite markers.However, the observed hemizygosity in intron 1 was the same as populational homozygosity.As for intron 8, hemizygosity was relatively frequent.Similar frequency for the same marker was observed in glioblastoma multiforme tumour samples (Kosla et al., 2011) and intronic deletions spanning several hundred base pairs of intron 8 were reported in many other cancers (Paige et al., 2000).Loss of heterozygosity is an abnormality frequently appearing in neuroblastoma which mainly affects other chromosomes than 16.However, LOH at this chromosome but within region 16p12-p13 was previously found to be connected with both sporadic and familial neuroblastoma (FNB locus) (Furuta et al., 2000;Weiss et al., 2000).In our study we did not observe any correlation of LOH with prognostic variables.However, such a frequent LOH at the WWOX gene region in our neuroblastoma population, seems to require further studies.
The second above mentioned inactivating mechanism i.e. methylation of the WWOX region appears to be an insignificant mechanism in neuroblastoma tumours.
In our study we also assessed the correlation of WWOX expression with the expression of genes with different cell functions i.e. apoptosis, cell cycle, adhesion, proliferation and signal transduction.
The observed positive correlation with antiapoptotic BCL2 is consistent with previous findings in glioblastoma and colon cancer (Kosla et al., 2011;Zelazowski et al., 2011).The influence of WWOX on apoptosis pathways is not well defined.In vitro studies conducted on MDA-MB231 cells transiently transfected with adenovirus harbouring WWOX cDNA resulted in increased cell death (Iliopoulos et al., 2007).On the other hand, the same cell line but stably transfected with WWOX cDNA had upregulated BCL2 expression and was characterized by increased invasion through basal membranes, suppressed anchorage independent growth and higher migration in Matrigel (Lewandowska et al., 2009).Gourley et al. found similar discrepancies after ovarian cancer cell lines transfection.Adherent cells modified with WWOX cDNA did not have higher apoptotic potential in contrast to those grown in suspension (Gourley et al., 2009).The highest rate of spontaneous regression is characteristic for neuroblastoma tumours (Oue et al., 1996).The molecular process underlying and explaining this phenomenon is still not well defined, although it is believed that programmed cell death and apoptosis may regulate its appearance.BCL2 as the apoptosis pathway regulator does not have a well defined role in neuroblastoma.Some research has shown that its overexpression correlated with favourable histology whereas others found that increased expression of this gene leads to tumour progression (Abel et al., 2005).Overall, there seems to be no correlation between this gene and prognostic variables (Ramani et al., 1994;Maris & Matthay, 1999).
In our study we did not observe any correlation between BCL2 expression and prognostic variables, thus the association found between this gene and the WWOX only confirms the correlation of their expressions also observed in other tumours.
Moreover, we observed a significant negative correlation of expression between WWOX and both cyclins (CCND1 and CCNE1).Similar correlation of WWOX and cyclin expression has been recently reported in Wilm 's tumours and colon cancer (Zelazowski et al., 2011;Pluciennik et al., 2012).
Both cyclins are phase G1/S specific and their increased expression was reported in neuroblastoma tumour and was associated with unfavourable histology (Hiyama et al., 2004).Cyclin E1 has been also associated in this tumour with stage 4 and poor prognosis in neuroblastoma patients (Mao et al., 2012).The second cyclin, CCND1 was shown to be overexpressed in unfavourable, malignant neuroblastomas (Hiyama et al., 2004;Molenaar et al., 2008) and was also connected with an undifferentiated phenotype (Molenaar et al., 2008).Thus, acquired negative correlation between WWOX tumour suppressor gene and both cyclins may indicate that this gene negatively regulates the cell cycle in tumour cells.
During our study we also examined differences in the expression of signal transduction genes, i.e.EGFR and both isoforms of HER4.We only observed a strong positive correlation between WWOX and HER4 JM-a, which is consistent with previous findings.Aqelian and coworkers established the competition between WWOX and YAP protein for interaction with HER4, this interaction modulates signal transduction of the HER pathway (Aqeilan et al., 2005).The sequestration of JM-a in cytoplasm as result of WWOX-HER4 interaction has been confirmed in breast cancer (Aqeilan et al., 2007), Wilm's tumours (Pluciennik et al., 2012), glioblastoma (Kosla et al., 2011) and colon cancer (Zelazowski et al., 2011).The HER family receptors are involved in embryonic development of the sympathetic system (Britsch et al., 1998;Casalini et al., 2004), however, their role in neuroblastoma is not well understood (Izycka-Swieszewska et al., 2011).There have been conflicting reports in terms of HER4 expression alterations in cell lines and primary tumours.Low expression of HER4 has been established in some neuroblastoma cell lines, and primary tumours, indicating no correlation with tumour stage (Ho et al., 2005).However, Richards et al. consistently detected expression of HER 4 in 7 out of 9 examined neuroblastoma cell lines and in all examined 20 tumour samples, indicating its potential role in promotion of neuroblastoma growth (Richards et al., 2010).More recently, Izycka-Swieszewska and coworkers (2011) evaluated via immunohistochemistry the expression of HER1-4 receptors in 103 NT and their prognostic significance and clinicopathological correlations.HER4 expression was found in 87% of all tumours, however, its low expression was more frequent in poorly differentiated neuroblastomas, but it did not correlate with the histological risk groups.Moreover, HER4 inversely correlated with the MKI index, and high expression was more frequent in children older than 18 months.On the other hand, expression of this receptor was characteristic for high risk group tumours and for metastatic rather than localized stage of disease (Izycka-Swieszewska et al., 2011).
Nevertheless, taking together the above mentioned publications describing the role of HER4 in neuroblastoma and observed in our study the strong correlation between expression of WWOX and HER4 we may suggest a potential role of this tumour suppressor gene in neuroblastoma cell differentiation regulation.We noticed decreased WWOX mean expression in poorly differentiated neuroblastoma samples and in samples which were assigned to the unfavourable prognostic category.
In conclusion, our study conducted on neuroblastoma samples revealed the potential role of LOH in altering WWOX expression.Moreover, the negative correlation observed with both cyclins and the positive correlation with BCL2, suggest its role in the regulation of the cell cycle and apoptosis of neuroblastoma cells.This results are consistent with previous findings.
Our observations for WWOX in neuroblastoma samples also outlined its potential role in the regulation of cell differentiation, however, further studies ought to be conducted.

.
Tissue samples were obtained from children treated in the Department of Paediatric Oncology and Hemtology, Medical University of Lodz, Poland, and in the Department of Bone Marrow Transplantation, Pediatric Oncology, and Haematology, Medical University, Wroclaw, Poland.

Figure 1 .
Figure 1.Correlation analysis of WWOX and cancer related genes expression in neuroblastic tumours.Both axes represent the relative expression level of investigated genes.