Classification Of Human Cancers Using Microrna Expression Profiles

MicroRNAs (miRNAs) are small, RNA molecules encoded in the genomes of plants and animals. These highly conserved, short (~22-mer) RNAs regulate the expression of genes by binding to the 3'-untranslated regions (30-UTR) of their target mRNAs. They were first identified in Caenorhabdetis elegans in 1993 [8] and it took several years before the first human miRNAs were discovered [8]. miRBase is a database that provides an integrated interface to miRNA sequence data, their annotation and predicted targets [9]. In the last 2 years, there has been a flurry of papers published describing miRNAs and their potential applications in drug discovery research. Several independent studies have shown that miRNAs could be key regulators in early development, cell proliferation and cell death [10]. Not surprisingly, there have been many studies of the expression and function of miRNAs in tumors [11].

There have been several lines of evidence suggesting that miRNAs function in a manner similar to that of siRNAs. In plants, miRNAs base pair with mRNAs with 100% complementarity and the target mRNA is directly cleaved and degraded using the RNA interference machinery in the cell [8]. miRNAs found in animals do not display the same level of complementarity to the mRNAs as that found in plants. As a result, suppression of the corresponding protein synthesis occurs via a mechanism that is not yet fully understood. Consequently, algorithms to identify the mRNAs that are the targets of miRNAs are being developed in many research labs. A month prior to the New York Academy of Sciences meeting in Feb 2005, six of the researchers who had developed algorithms to identify the targets of miRNAs were assigned "homework" where they were asked to predict the targets of two animal miRNAs [12]. In this blind test, it was found that there was virtually no overlap between the results of the four different algorithms that were reported. This startling outcome underlines the difficulty of the task of predicting in silico the targets of miRNAs in animals. Nevertheless, the task is one that is particularly well suited to bioinformatics since a major part of the algorithm involves matching miRNAs to the target mRNA at the primary sequence level. miRNAs usually do not bind to long complementary stretches in animals due to the lack of 100% complementarity with the target sequence. Consequently, loop-outs and non-Watson-Crick base pairing occur which is tackled differently at the computational level by different algorithmic approaches. Hence, the lack of overlap between the predictions made by the various algorithms.

The lack of 100% complementarity between a given miRNA sequence and the corresponding target mRNA may also indicate that the miRNA targets more than one mRNA, which potentially infers considerable regulatory functions to miRNAs. There is published evidence to suggest that on average, miRNAs have 100 target sites [13].

At the other end of the spectrum, several groups have been trying to identify new miRNA genes in silico [14]. In this analysis, a comparative genomics approach was taken. By looking at the 3'-UTRs across human, mouse, rat and dog genomes, Xie et al. discovered 106 conserved motifs. These motifs were reminiscent of many miRNA binding sites based on their length, distribution and the fact that they tended to end with the nucleotide "A", which is a characteristic of miRNA sequences. Further analysis of the 106 motifs by searching for complementary sequences in the miRNA database found 72 of the motifs to be complementary to known miRNAs. Additionally, 242 conserved and stable stem-loop sequences were found in the conserved sequences (across the four genomes) that were complementary to the 72 motifs mentioned above. These represented 113 sequences that encode already discovered miRNAs and 129 novel predicted miRNA genes. A representative set of 12 predicted new miRNA genes were tested in a set of pooled 10 adult human tissues (breast, prostate, pancreas, colon, stomach, uterus, lung, brain, liver and kidney). Although this set of 10 tissues is by no means complete, six of the 12 predicted miRNAs were found to be expressed in the pooled tissues. Considering that this analysis was limited to motifs conserved across four genomes and that the conserved motifs were stringently matched for complementarity to the known and predicted miRNAs, there exists the very real potential of discovering additional miRNAs either by relaxing the stringency of the sequence match or by limiting the analysis to only those genomes more closely related to humans (other primates, for example).

On the experimental side, an important publication in June 2005 [11] has described the characterization of miRNA expression profiles in human cancers across a diverse panel of human tissues (colon, kidney, lung, prostate, uterus and breast), both normal and cancerous. The analysis of the expression-profiling data suggests that the expression of the miRNAs distinguishes tumors of different origin, and that most of the tumors had lower levels of expression of the miRNA than the corresponding normal tissue. The data were so distinct that the classification of normal versus tumor could be done with 100% accuracy based on the differences in levels of expression between normal and tumor tissue. Compared to miRNA-expression levels, mRNA expression levels do not provide this level of specificity in distinguishing between normal and tumor tissue. This level of consistency in miRNA-expression level changes between normal and tumor tissue predicts that the monitoring of miRNA expression levels has diagnostic potential in the clinic.

Research in the field of miRNAs has not yet reached a steady state by any means. Several research labs are actively engaged in trying to discover additional miRNAs, in developing robust algorithms to predict the targets of miRNAs and in the understanding of their function and mechanism of action [10]. The multiple predictions of targets for miRNAs, the constant discovery of new miRNAs and the identification of their various functions suggest that we have much more to discover about the role of miRNAs in plant and animal biology.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment