Although the rate of cancer incidence is greater in the breast than the ovaries, the latter has a higher relative mortality. This is because ovarian tumours are often diagnosed at an advanced stage. In addition, a majority of tumours that are initially responsive to chemotherapy eventually acquire drug-resistance to these therapeutic agents [5, 6]. Comprehending how this drug resistance occurs is the basis for developing strategies to improve treatment outcomes.
In this study, the five sFRP isoforms were characterized for the first time in a range of chemosensitive and chemoresistant human epithelial ovarian cancer cell lines. Similar to other tumour types, several of the sFRP isoforms were detected in ovarian cancer cells. However, sFRP4 was the dominant isoform in all the cell lines tested. sFRP4 was first isolated from the ovary and it is probable that it may have specificity for this organ ; hence, its dominant expression could be indicative of a specific functional role in the ovary. Similarly, sFRP4 mRNA was the only isoform to be differentially expressed between the different cell lines. Western blots and IHC for sFRP4 protein also confirmed that the chemoresistant cells expressed lower levels of this protein. The differential expression of sFRP4 seen in the chemosensitive and chemoresistant cell lines indicated that sFRP4 could serve as a prognostic marker for ovarian tumours.
IHC showed heterogeneous expression of sFRP4 in the various ovarian cell populations examined, in contrast to the pattern of expression in the normal IOSE cells, which was homogenous. The chemosensitive cell line was found to exhibit a larger sub-population of sFRP4 positive cells than the chemoresistant cells. It is known that ovarian tumours have a heterogeneous population of cells. Therefore it is very likely that the cells have a differential response to chemotherapeutic treatment [3, 34]. Our results indicated that the heterogeneity of sFRP4 expression was correlated with the response of the cell lines to Cisplatin, with sFRP4 expression as a positive marker for chemosensitivity.
Furthermore, IHC of the cancerous cell lines following treatment with Cisplatin revealed that, compared to untreated controls, both the chemosensitive cell line A2780 and the chemoresistant cell lines lost most of their sFRP4 expressing cells; indicating that Cisplatin selectively targets sFRP4 expressing cells within the heterogeneous population.
Increasing the sFRP4 expression of the chemoresistant cell lines using transfection prior to Cisplatin treatment resulted in more cells expressing sFRP4 and, as predicted, increased their sensitivity to treatment. In contrast, knocking down sFRP4 expression of the chemosensitive A2780 cells by 40% was sufficient to confer these cells with partial resistance to subsequent chemotherapeutic treatment when compared to controls. These results are comparable with a study by He et al., (2005), who reported that significant down regulation of sFRP4 expression promoted cell growth and inhibited chemotherapeutic drug-induced apoptosis in mesothelioma cell lines . Our transfection and silencing experiments confirmed that sFRP4 appeared to have a direct influence on the chemo-response of cancer cells. The data generated from this study present a novel finding and indicate a potential avenue for future research on sFRP4.
The heterogeneous expression of sFRP4 we observed using IHC in our cell lines indicated the presence of at least two sub-populations within ovarian cell cultures. The chemosensitive cells had a larger population of sFRP4 positive cells, and this correlated with their greater sensitivity to treatment. Hypermethylation of the sFRP4 gene has been reported in various cancers and is associated with tumour progression and malignancy [28, 29]. We did not examine our ovarian cancer cell lines on whether low sFRP4 expression was associated with hypermethylation of the sFRP4 gene itself, and this would be useful to know.
Our experiments raised some possible explanations for acquired resistance to Cisplatin in human ovarian cancers. From our IHC analysis we know that Cisplatin selectively targets sFRP4 expressing cells, thus continuous treatment would gradually deplete this sub-population of cells over time. Consequently, the resultant population remaining would now be comprised largely of cells that do not express sFRP4 and, hence, be more resistant to subsequent treatment. These data suggest that loss of sFRP4 expression is functionally associated with a more malignant and chemoresistant phenotype, and that treatment itself may select for these cells, thereby resulting in acquired chemoresistance.
Suppression of sFRP4 in the chemosensitive cells resulted in a corresponding increase in β-catenin expression of these cells. This study confirmed the existence of an inverse relationship between sFRP4 and β-catenin expression, and the presence of a signalling interaction between these two proteins in the ovarian cell lines. These findings are consistent with other studies that also reported a similar inverse relationship between these two proteins in endometrium and breast [36, 37], and suggests that sFRP4 may act as a tumour suppressor through its interaction with the Wnt/β-catenin signalling pathway by modulating the cellular cytosolic β-catenin pool.
Mucinous tumours, unlike other ovarian epithelial sub-types, have a well characterised progression from benign to borderline and ultimately to adenocarcinomas [38, 39]. In addition, mucinous adenocarcinomas, compared to other subtypes, respond poorly to chemotherapy and are known to have a poor prognosis . Similar to our data, a trend has been observed by other studies where down regulation of sFRP4 expression was similarly associated with stage and grade of the cancer [32, 40]. These results suggest that the progressive decline of sFRP4 expression in higher grade disease states could be associated with both tumour progression as well as onset of chemoresistance. However, due to our relatively low patient cohort size, this trend needs to be assessed independently in larger patient numbers to further validate the significance of this finding.