The Application of MicroRNAs in Cancer Diagnostics - RNA Interference from Biology to Therapeutics - page 281

The non-coding RNA Growth Arrest Specific 5 (GAS5) is known to bind as a
decoy to the glucocorticoid receptor (GR), thereby preventing its interaction with
the glucocorticoid response elements in genomic DNA. GR target genes are involved
in repression of apoptosis and are normally upregulated upon GR binding. Thus,
interaction of GAS5 with the GR may allow or even promote apoptosis (reviewed
in [ 235 ]). In agreement with this model, GAS5 transcript levels have been found to
be significantly downregulated in breast cancer cells compared to normal breast
epithelium [ 260 ]. Thus, GAS5 may have potential as a future cancer biomarker or
even as an anticancer therapeutic agent.

13.6 Future Perspectives

Although a large number of molecular profiling studies have identified several miR-
NAs with potential as future cancer biomarkers, no miRNA-based diagnostic test
has yet been developed and approved for use in routine clinical practice. Lack of
sufficient independent validation and inconsistencies between results from different
miRNA profiling studies of cancer tissue samples remain major limitations for the
clinical translation of novel miRNA biomarker candidates. Such inconsistencies
may be explained, at least in part, by the use of different microarray platforms and
RT-qPCR methodologies in studies of the same tumor type [ 140 ] . Standardized pro-
cedures for sample preparation and miRNA quantification could accelerate the dis-
covery and validation of new miRNA cancer biomarkers, a prerequisite for their
translation into clinical usage. Improved and more detailed clinical annotation of
patient samples included in miRNA expression profiling studies may also help push
this fi eld forward.
Furthermore, miRNA profiling studies conducted so far have often been under-
powered, i.e., insufficient sample numbers in the discovery and/or validation phase.
Consequently, the results obtained from these studies may not be representative for
a given cancer type and, hence, less likely to lead to the development of clinically
useful biomarkers. Recent studies focusing on long non-coding RNAs tend to suffer
from similar limitations in sample numbers. As an additional complicating factor,
most tumors are characterized by marked histologic and genetic heterogeneity [ 261,
262 ]. Accordingly, the exact cancer tissue specimen(s) used for molecular profiling
analyses may not be representative for the entire tumor. To address this issue, some
studies have used laser-microdissection for isolation of specific cell fractions from
tumor tissue samples prior to miRNA profiling. Although this is a laborious process
not compatible with routine daily clinical procedures, such studies can contribute
invaluable new and more detailed insights into tumor biology and mechanisms of
carcinogenesis [ 263 ] .
Identification of circulating tumor biomarkers could be an alternative solution to
this problem. Indeed, the expression of specific circulating miRNAs seems to be a
good surrogate of tumor miRNA expression and, thus, has initiated a new paradigm
useful for noninvasive testing for early diagnosis, prognosis, and/or therapeutic