Identification of miRNAs contributing to neuroblastoma chemoresistance

Abstract

Background: The emergence of the role of microRNAs (miRNAs) in exacerbating drug resistance of tumours is recently being highlighted as a crucial research field for future clinical management of drug resistant tumours. The purpose of this study was to identify dys-regulations in expression of individual and/or networks of miRNAs that may have direct effect on neuroblastoma (NB) drug resistance.

Methods: Individual subcultures of chemosensitive SH-SY5Y and UKF-NB-3 cells were rendered chemoresistant to doxorubicin (SH-SY5Y, UKF-NB-3) or etoposide (SH-SY5Y). In each validated chemoresistance model, the parental and subcultured cell lines were analysed for miRNA expression profiling, using a high-throughput quantitative polymerase chain reaction (RT-qPCR) miRNA profiling platform for a total of 668 miRNAs.

Results: A unique expression signature of miRNAs was found to be differentially expressed (higher than 2-fold change) within all three NB chemoresistance models. Four miRNAs were upregulated in the subcultured chemoresistant cell line. Three miRNAs were found to be downregulated in the chemoresistant cell lines for all models.

Conclusions: Based on the initial miRNA findings, this study elucidates the dys-regulation of four miRNAs in three separate NB chemoresistant cell line models, spanning two cell lines (SH-SY5Y and UKF-NB-3) and two chemotherapeutic agents (doxorubicin and etoposide). These miRNAs may thus be possibly linked to chemoresistance induction in NB. Such miRNAs are good candidates to be novel drug targets for future miRNA based therapies against aggressive tumours that are not responding to conventional chemotherapy.

Keywords: Chemoresistance; Drug; Neuroblastoma; Resistance; miRNA.

Fig. 1

Results for cell viability analysis of Kelly NB chemoresistance models (n = 1, three technical replicates/data point, SEM not illustrated). In all three assays the chemosensitivity status for each component NB cell line was not suitable for inclusion in the miRNA profiling study (black — chemosensitive parental NB cell line; red — chemoresistant NB cell line).

Fig. 2

Results for cell viability analysis of UKF-NB-3 NB chemoresistance models (two separate runs, n = 1, three technical replicates/data point, SEM not illustrated). The doxorubicin chemoresistance model was included in the miRNA profiling study due to a 10-fold linear change in the IC50 dose for both constitutive cell lines. However, at elevated doxorubicin doses a reproducible, artificial increase in cell viability was noticed within this NB chemoresistance model. The UKF-NB-3/vincristine NB chemoresistance model was not included in the study due to incongruent chemosensitivity profiles for each constituent NB cell line (black — chemosensitive parental NB cell line; red — chemoresistant NB cell line).

Fig. 3

Results for cell viability analysis of SH-SY5Y NB chemoresistance models (n = 1, three technical replicates/data point, SEM not illustrated). The cisplatin chemoresistance NB model was not included in the study due to lack of a suitable chemosensitivity profile by the predicted chemoresistant NB cell line component for this model.

Fig. 4

Cell viability analysis results (2nd run, n = 1, three technical replicates/data point, SEM not illustrated) for re-confirmation of chemosensitivity profiles for each constituent NB cell line chemoresistance model.

Fig. 5

Scatterplot of miRNA expression for SH-SY5Y/DOXO cell line model constituents. This highlights only the marginal variation in the percentage of miRNAs that are actually expressed within the individual cell lines.

Fig. 6

Scatterplot of miRNA expression for SH-SY5Y/ETOPO cell line model constituents. This highlights only the marginal variation in the percentage of miRNAs that are actually expressed within the individual cell lines.

Fig. 7

Scatterplot of miRNA expression for UKF-NB-3/DOXO cell line model constituents. This highlights insignificant variation in the percentage of miRNAs that are actually expressed within the individual cell lines.

Fig. 8

Graphs demonstrating log-fold change in miRNA expression for all three selected NB chemoresistance models. This miRNA signature of seven miRNAs was identified as dysregulated in the chemoresistant NB cell line components of all three selected NB chemoresistance models. miRNAs highlighted in green denote dysregulations in which both constituent cell lines had a raw Cq value below the 35 cutoff value. This specific Cq threshold value was utilised in order to eliminate false positive expression data inclusion due to quantification noise.

Fig. 9

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-125b-1# following normalisation with reference miRNAs. Expected down-regulation of putative miRNA was confirmed in UKF-NB-3/DOXO and SH-SY5Y/DOXO models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.

Fig. 10

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-150 following normalisation with reference miRNAs. Expected up-regulation of putative miRNA was not confirmed in all three chemoresistance models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.

Fig. 11

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-188-5p following normalisation with reference miRNAs. Expected up-regulation of putative miRNA was confirmed in all three NB chemoresistance models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for õpost-run data normalisation and analysis.

Fig. 12

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-204 following normalisation with reference miRNAs. Expected down-regulation of putative miRNA was solely confirmed in UKF-NB-3/DOXO NB model. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.

Fig. 13

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-501-5p following normalisation with reference miRNAs. Expected up-regulation of putative miRNA was confirmed in UKF-NB-3/DOXO and SH-SY5Y/ETOPO models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.

Fig. 14

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-519b-3p following normalisation with reference miRNAs. Expected up-regulation of putative miRNA was not confirmed for all three chemoresistance models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.

Fig. 15

Multiplex RT-qPCR assay results (log-fold change) for hsa-miR-769-3p following normalisation with reference miRNAs. Expected down-regulation of putative miRNA was not confirmed in all three chemoresistance models. Error bars represent standard deviation. Assays for hsa-mir-99b, hsa-mir-125a and hsa-mir-425 were also prepared as reference miRNAs for post-run data normalisation and analysis.