miRNA Profiles as a Predictor of Chemoresponsiveness in Wilms’ Tumor Blastema

The current SIOP treatment protocol for Wilms’ tumor involves pre-operative chemotherapy followed by nephrectomy. Not all patients benefit equally from such chemotherapy. The aim of this study was to generate a miRNA profile of chemo resistant blastemal cells in high risk Wilms’ tumors which might serve as predictive markers of therapeutic response at the pre-treatment biopsy stage. We have shown here that unsupervised hierarchical clustering of genome-wide miRNA expression profiles can clearly separate intermediate risk tumors from high risk tumors. A total of 29 miRNAs were significantly differentially expressed between post-treatment intermediate risk and high risk groups, including miRNAs that have been previously linked to chemo resistance in other cancer types. Furthermore, 7 of these 29 miRNAs were already at the pre-treatment biopsy stage differentially expressed between cases ultimately deemed intermediate risk compared to high risk. These miRNA alterations include down-regulation in high risk cases of miR-193a.5p, miR-27a and the up-regulation of miR-483.5p, miR-628.5p, miR-590.5p, miR-302a and miR-367. The demonstration of such miRNA markers at the pre-treatment biopsy stage could permit stratification of patients to more tailored treatment regimens.


Introduction
Wilms' Tumor (nephroblastoma), the commonest pediatric renal tumor, arises predominantly in children under 5 years old, and is thought to develop from embryonic kidney cells [1,2]. Of all the embryonal tumors, Wilms' represents the best example of organogenesis ''gone awry'', as it contains varying amounts of the triphasic elements of nephrogenesis including epithelial, stromal and blastemal cell types, which are recognisable by light microscopy.
Overall, outcomes for Wilms' tumor are excellent and a current key aim is the minimisation of therapy in these very young patients to avoid long-term sequelae [3]. The current SIOP (Societe Internationale d'Oncologie Paediatrique) protocol for the treatment of Wilms' Tumor in patients over the age of 6 months includes neoadjuvant chemotherapy followed by nephrectomy [4]. The rationale is to down-stage tumors pre-operatively in order to facilitate safer nephrectomy [5,6]. Post-nephrectomy, histological evaluation permits staging and risk-based stratification. According to the CCLG [Children's Cancer and Leukaemia Group] protocol, up-front diagnostic biopsy of the tumor is additionally performed; thus therapy-naïve tissue is evaluable in such cases. There are currently no means at biopsy stage of identifying which cases will be 'blastemal' and so deemed high-risk at nephrectomy. Along with morphologically identified diffusely anaplastic cases [known to be associated with TP53 mutation], blastemal Wilms' represent the chemo-resistant cohort requiring more intensive adjuvant chemotherapy. Studies that consider the molecular mechanisms of chemo-resistance in Wilms' tumor have focused almost exclusively on gene expression profiling with no evidence of signatures that might predict chemotherapeutic response [7]. Evidence is emerging for the role of miRNAs in both renal development [8] and Wilms' tumor progression [9]. Given the lack of clear genomic/genetic pointers, our aim was to identify markers of chemo resistance in Wilms' tumor by miRNA profiling.

Patient Samples
In this study, the term chemo responsiveness is based on histological findings such that intermediate risk cases are considered to have shown better response, while high risk cases showed relative resistance. The patient cohort (Table 1)  miRNA Profiling Using TaqMan Low Density Arrays Total RNA was extracted according to the manufacturer's instructions using RecoverAll Total Nucleic Acid Isolation Kit (Ambion) which recovers the small RNA fraction and allows for the downstream analysis of miRNAs. RNA was reverse transcribed using the miRNA reverse transcription kit (Applied Biosystems). Megaplex primer pool A used in this reaction corresponds to TaqMan Low Density Array (TLDA) A (Applied Biosystems), and allows the simultaneous reverse transcription of 377 miRNAs, 3 endogenous controls and 1 miRNA assay unrelated to mammalian species on TLDA A. Alternatively, for the individual analysis of miRNAs using Taqman miRNA assays, target-specific stem-loop RT primers were used in combination with the miRNA reverse transcription kit. Megaplex RT product was pre-amplified using TaqMan PreAmp Master Mix (26) and PreAmp Primer Pool A (Applied Biosystems). The PreAmp primer pool contains forward primers specific for the miRNAs on TLDA A and a universal reverse primer (Applied Biosystems). Each cDNA sample was analyzed on TLDA A using the 7900HT RT-qPCR system (Applied Biosystems).

Statistical and Bioinformatic Analysis
Ct values and Relative quantities were calculated using RQ Manager 1.2.1 with the application of automatic baseline settings and a threshold of 0.2. Prior to analysis samples were first standardized (mean centred) and the normalized relative expression (NRQ) values of miRNA were calculated with reference to the Ct max (maximum Ct value of a miRNA across samples) using the equation NRQ = 2(Ct max 2Ct). MiRNAs that were expressed in at least 10 samples were retained for analysis.
The combined hierarchical clustering was carried out using Spearman's rank correlation as an object similarity metric and complete linkage was implemented to ascertain cluster similarity using the hclust function. The associated heatmap analysis was implemented using the heatmap.plus function. Colors were generated in a standardized manner using the rank of miRNA expression values across samples.
The non-parametric Wilcoxon's rank sum was used to measure the significance of association of miRNA expression to the various selected sample classes. P-values were corrected for multiple comparisons using the Bonferroni method. Potential targeting of differentially expressed miRNA and diametrically, differentially expressed mRNA was examined by assessing the number of predicted targets present. The significance of this value was calculated by comparing this with that expected by chance using Chi-squared statistic. This analysis was carried out using an algorithm implemented Java v1.5 and the TargetScan version 5.1 database.

Results
In this study, the intermediate risk cases were defined, as patients whose tumors were histologically regressive, epithelial, stromal or mixed, post-chemotherapy. Of these, 69% were of mixed histology, 15% stromal, 8% epithelial and 8% regressive. The high risk group comprised blastemal cases only. Anaplastic cases were deliberately excluded from this analysis, on the basis that these tumors are driven, at least in part, by TP53 mutation and are also readily recognisable microscopically. This pathological stratification parallels the clinical outcomes. Within our high risk blastemal group, 28.6% of patients relapsed and a 78.6% 2 year survival is evident, whereas within the intermediate risk group, there is a 100% 2 year survival and only 15.4% of patients relapsed. The average ages of patients within high and intermediate risk groups are 46 months and 45 months respectively.
In order to assess the miRNA profiles of intermediate and high risk groups, genome-wide miRNA expression analysis was performed using TaqMan Low Density Array A, which includes 377 miRNAs and of these, 295 miRNAs were retained for analysis. Unsupervised hierarchical clustering performed on these 295 miRNAs separates tumors into two clusters that clearly discriminate high risk  appears most similar to blastema originating from pre-treatment high risk blastemal cases [Pre-HR]. This is illustrated as supporting information in Figure S1.
The principal aim of this study was to identify miRNAs that are predictive of chemotherapeutic response at the biopsy stage, and which might then be used to stratify patients prior to the administration of pre-nephrectomy chemotherapy. In order to address this, we compared miRNA profiles of the non-blastemal We investigated for representation of these miRNAs at the biopsy stage, through comparison of pre-treatment biopsies originating from intermediate risk cases [Pre-IR], to pre-treatment biopsies from high risk cases [Pre-HR]. Interestingly, 7 miRNAs from the post-treatment analysis were also differentially expressed at the biopsy stage (Table 4). Although the pre-treatment analysis did not yield any significant p-values, perhaps due to the small sample size of the pre-treatment high risk group, a total of 54 miRNAs approached significance (p = 0.19).
In order to focus on potential markers of chemo resistance in Wilms' tumor we considered further the 7 miRNAs that were found in this study to be significantly altered between intermediate risk and high risk groups post-treatment, and also already differentially expressed in the pre-treatment analyses. Our assumption is that these miRNAs may be particularly important indicators of the tumor's potential to respond to pre-operative chemotherapy. These miRNA alterations included the downregulation of miR-193a.5p and miR-27a, and the up-regulation of miR-483.5p, miR-628.5p, miR-590.5p, miR-302a and miR-367 in high risk cases (Table 4). Although miR-20a expression was altered in both pre-and post-treatment analyses the directional changes were not consistent and therefore this miRNA was not considered further.

Discussion
According to the SIOP protocol, standard management of renal tumors is pre-operative chemotherapy followed by nephrectomy. Although in most cases of Wilms' tumor such therapy induces a considerable degree of tumor shrinkage with evidence of necrosis and cellular differentiation, a proportion of cases show lesser volume reduction and retain an abundance of blastemal tissue post-chemotherapy and these are stratified as high-risk blastemal cases and have a poorer prognosis [4].
To date, there is a very limited understanding of the molecular effects of therapy on Wilms' tumor and the factors that produce chemo resistance in this tumor. This is partly due to challenges in the acquisition of paired pre-and post-treatment cases, as the National Wilms tumor study group (NWTS) in North America currently recommends primary surgical resection of Wilms' tumors and even within the SIOP protocol, only a few countries offer biopsy-based diagnosis prior to treatment.
Previous studies have focused on the assessment of histological features in pre-versus post-treatment Wilms' tumor cases [10,11]. Barroca (2010), found that the cytologic features of pre-treatment fine needle biopsies did not have any bearing on the type and extent of post-treatment regressive changes, however, all predominantly blastemal (high risk) post-treatment tumors corresponded to pre-treatment cases with considerable blastemal components and little necrosis or apoptosis. In order to explain this phenomenon, Barroca suggested that two types of blastema may occur in Wilms' tumor and that each of these types exhibits different chemotherapeutic sensitivities, proliferative properties and abilities to undergo apoptosis or necrosis [11]. Although no molecular analysis has been conducted to date that confirms this theory.
In this study we performed a direct comparison of miRNA profiles between post-treatment blastema, isolated from intermediate risk cases and the predominant blastemal component in high risk cases and identified 10 miRNAs to be significantly upregulated in the high risk cases (Table 2). Of particular interest are the up-regulation of miRNA 106b and miR-25 in the high risk blastema. These miRNAs are members of the miR-106b,25 cluster located within intron 13 of the Mcm7 gene that has been found to directly target a number of important tumor suppressor pathways and may therefore be involved in the promotion of cellular proliferation in high risk blastemal regions [12]. For   [13], as well as cyclin-dependent kinase inhibitor p21/CDKN1A, [14] and pro-apoptotic Bcl-2 family member BCL2L11 [Bim] [12]. The suppression of even one of these pathways by the over-expression of miR-106b identified here, has the potential to lead to a highly proliferative blastema with reduced apoptosis and could thereby contribute to the persistent and voluminous nature of this blastemal component in high risk cases. A number of recent publications suggest a significant role for miRNAs in the molecular response to chemotherapy and a selection of miRNAs has now been associated with chemo resistant phenotypes in a range of cancer types [15,16]. Varying levels of chemo sensitivity may be inferred by the influence of miRNA up-or down-regulation on multiple mRNA targets and subsequent protein expression. This may impact the cellular response to anticancer agents by the dysregulation of survival pathways, apoptotic capabilities, specific drug targets, DNA repair systems or drug transport/metabolism [17,18]. However, it has also been shown that the role of miRNAs in chemo resistance is complex and often miRNAs are seen to have the opposite effects towards similar anticancer agents in different cancer types [19]. A number of differentially expressed miRNAs identified in this study has been previously associated with chemo resistance. For example, miR-15b, which is down-regulated in pre-treatment high risk cases here, has previously been shown to play a role in the development of multi-drug resistance in gastric cancer cells through up-regulation of the Bcl2 protein and consequent modulation of apoptotic pathways [20]. Alternatively, up-regulation of miR-222 is also known to confer Tamoxifen drug- resistance in breast cancer cells through the negative regulation of ERa and cell cycle inhibitor p27Kip1 [21,22]. In this study, in order to establish a potential miRNA marker that may be predictive of chemotherapy response at biopsy stage, we selected miRNAs where expression was altered between intermediate and high risk cases not only after, but also already before chemotherapy. The assumption here is that any marker that might indicate a resistant tumor type prior to treatment would remain detectable in high risk cases, as it is precisely the tumor cells expressing this marker that would be chemo resistant.
Here we have highlighted a signature of 7 differentially expressed miRNAs within pre-treatment diagnostic biopsy specimens that may be predictive of chemo resistance in Wilms' tumor. These include the down-regulation of miR-193a.5p and miR-27a, and the up-regulation of miR-483.5p, miR-628.5p, miR-590.5p, miR-302a and miR-367. Of particular interest are the protein targets of miR-27a, miR-302 and miR-483.5p. The downregulation of miR-27a has been linked to p-glycoprotein expression and has previously been reported to correlate with an upregulation of this protein and to contribute to chemo resistance and relapse in leukemia patients [23]. The principal role of pglycoprotein is to mediate the elimination of toxic substances from the cell. It is normally expressed in kidney cells and its expression has been shown to be enhanced by Vincristine and Dactinomycin chemotherapy in Wilms' tumor patients [24].
It is plausible that the over-expression of miR-302 could promote an embryonic stem cell-like cell cycle profile through the suppression of cyclin D1 and Cdk4 [25]. This is particularly applicable here as Wilms' tumor likely arises from multipotent embryonic kidney precursor cells and the expression of residual multipotent properties may remain evident within regions of persistent blastema that are a morphological feature of chemoresistant histology. In addition, miR-367 is part of the miR-302 cluster, and the miR-302/367 promoter is known to be transcriptionally regulated by embryonic stem cell-specific transcription factors [26]. These miRNAs are expressed at high levels in post-treatment high risk cases in this study and may therefore confer multipotent properties on these blastemal cells [27]. miR-483.5p upregulated in our high risk cases, is located within intron 2 of the IGF2 gene on chromosome 11p15 [28] and is thought to be co-expressed with its IGF2 host gene [29]. As the expression of IGF2 has a particularly significant role in Wilms' tumorigenesis and is often increased due to associated loss of imprinting (LOI) events [30], the expression of this miRNA may therefore also be frequently affected in tandem with IGF2 in Wilms' tumor, although this has not been studied to-date. miR-483-5p has been shown to target the suppressor of cytokine signalling 3 (SOCS3) [31]. This results in the decreased expression of the SOCS3 protein leading to sustained signaling of the Interleukin-6 (IL-6)/Signal transducers and activators of transcription 3 (STAT3) cytokine pathways which stimulates growth and proliferation and is potentially anti-apoptotic [32].
It is possible that one of the differentially expressed miRNAs identified in this study may alone be the key component in determining response to chemotherapy in Wilms' tumor patients. But, perhaps these miRNAs in fact work in tandem in regulating the complex pathways of chemotherapeutic response and may need to be considered as a miRNA signature in future studies. In order to validate the miRNAs highlighted here, a prospective study would need to be undertaken wherein the expression of these miRNAs is assessed in patient biopsy samples. The suggestion here is that miRNA markers identifiable in pre-treatment biopsy specimens could allow for early prediction of potential drug resistance and may be an important step towards customizing current regimens in non-responsive Wilms' tumor patients.