Our approach enabled us to investigate almost all currently known human micro-RNAs simultaneously. Quantitative RT-PCR was chosen because of it's recognized precision and superiority over other methods for gene expression analysis [5–7]. Our experiments were run in triplicate and on average revealed a CV <30% covering variability caused by methodological (mean CV <5%, data not shown), intra- and inter-individual sources (table 1). Since this CV is clearly below the definition of a two-fold threshold for differential gene expression, we are convinced that the results of our analysis represent true differences in micro-RNA expression between the cell line pairs.
Of 738 micro-RNA species examined, only 9.8% (72) appeared differentially expressed with considerable overlap in micro-RNA gene expression occurring between the cell line pairs NTERA-2-R/NTERA-2 and NCCIT-R/NCCIT, but considerably less overlap with the cell line pair 2102EP-R/2102EP cells (table 1).
Functionally, the most striking distinguishing feature between the cell lines NTERA-2 and NCCIT on the one hand, and 2102EP on the other hand is a difference in their capacity to differentiate. Whereas the former two lines can differentiate upon stimuli such as exposure to retinoid acid, the latter is nullipotent and is therefore devoid of alterations induced by processes of cellular differentiation . Interestingly, there is increasing evidence for a link between cellular differentiation and resistance to chemotherapy in germ cell tumors, both in cell line models, as well as in primary tumor material, which seems to be linked to two microRNA clusters, namely miR371-373 and miR302 .
In contrast to many features shared by NTERA-2 and NCCIT, extragonadal NCCIT cells show a higher resistance level to cisplatin, compared to NTERA-2 . In this context, the several fold-change differences measured in the micro-RNA cluster covering hsa-miR-371-373 is of interest. Based on our findings, hsa-miR-371-373 expression in NTERA-2-R increased 4.0-4.7-fold relative to the paternal cisplatin sensitive NTERA-2 line, and even more in NCCIT-R, relative to the cisplatin sensitive paternal NCCIT cells, showing an 12.0-16.3-fold upregulation, table 1. Noteworthy, this micro-RNA cluster has been discussed in the context of the presence of wild-type p53 in germ cell tumors, counteracting tumorigenesis by induction of senescence [2, 10, 11]. By up-regulation, the cluster prevents p53-driven cellular senescence via a plethora of target genes, e.g. NEO1 or LATS2 , therefore leading to cell proliferation even in the presence of wild-type p53. Furthermore, this cluster has been described in cells exhibiting stem cell properties . Noteworthy, this cluster has recently been identified by other authors to be the most significant differentially expressed microRNA in human germ cell tumors . Therefore, it appears a promising target for further analysis in germ cell tumors, e.g. by posttranscriptional silencing with siRNA, to examine the functional role of the cluster in altering gene expression and thereby elucidate it's contribution to the development of cisplatin resistance.
As already mentioned, there are other gene changes which have been discussed in the context of cisplatin resistance . The tumor suppressor p21, regulating transition through the cell cycle and acting downstream of p53, has already been associated with hsa-miR520 belonging to the miR-106/302 family . An increased expression of miR-106/302 family members inhibits the tumor suppressor p21 and rescues human mammary epithelial cells from Ras-induced senescence . In our analysis, hsa-miR520c as well as hsa-miR520h were up-regulated (3.9-7.1-fold) in NTERA-2-R/NTERA-2 and NCCIT-R/NCCIT cell line pairs, which could point towards another mechanism counteracting senescence.
In an in vitro model of cisplatin resistance of a squamous cell carcinoma cell line, has-miR-21 was found to be down-regulated in cisplatin resistant cells . In our cell line pairs, significant down-regulation of has-miR-21 was observed in NCCIT-R compared to sensitive NCCIT cells. Furthermore, hsa-miR-146a, a microRNA targeting BRCA1, was the most highly up-regulated microRNA in another in vitro model of cisplatin resistance of the breast cancer cell line MCF7 . Similarly, both NTERA-2-R/NTERA-2 and NCCIT-R/NCCIT showed up-regulation in our analysis, yet to a lesser extent. However, has-miR-221, which was also highly up-regulated in cisplatin resistant MCF7 cells, was down-regulated in NCCIT-R and unaltered in the other two cell lines in our experiments, indicating that cisplatin resistance seems to be cell line or tumor type specific. In a lymphoblastic leucemia cell line overexpression of hsa-miR-125b conferred a survival advantage through inhibition of caspase 3 activation after exposure to a broad spectrum of apoptotic stimuli . In our analysis, however, hsa-miR-125b appeared to be up to 5-fold down-regulated in NCCIT-R/NCCIT and 2102EP-R/2102EP cell line pairs, thus making an inhibition of apoptosis through this micro-RNA species unlikely.
According to our analysis, further micro-RNA species appeared either about 8-fold up-regulated (hsa-miR-512-3p/-515/-517/-518/-525) or about 10-fold down-regulated (hsa-miR-99a/-100/-145) in both NTERA-2-R/NTERA-2 and NCCIT-R/NCCIT cell line pairs. Literature results describing a role of these micro-RNAs in cisplatin resistance are missing, warranting further studies to elucidate their potential role in development of resistance.
Recently, a role of the miR-106b seed family members in cisplatin resistance of testicular cancer has been described by another group . Although detectable, neither miRNA-106b, nor miRNA-106a, miRNA-17-5, miRNA-93, and miRNA-20 were differentially expressed in our model. In our view, this points at the multifactorial nature of cisplatin resistance in germ cell tumors.