Our DNA is what makes us who we are and the central dogma underpinning genetics is that genes are encoded by DNA, the code is copied into RNA which is then decoded to build proteins. When the genome was sequenced it was discovered that only a small proportion of our DNA (2%) encodes genes and rest is noncoding and was termed “junk” DNA.
We now know that the “junk” DNA is not inert. Some of it encodes regulatory information that signals when and in which tissue a gene should be expressed. Large pieces of noncoding DNA are copied (transcribed) into long bits of noncoding RNA, or long noncoding RNAs (lncRNAs) and these were similarly regarded as mere transcriptional “noise”. However, it soon became apparent that this “noise” has a function in regulating the expression of coding genes, but whether this is universally true is still under intense debate. One reason for this uncertainty has been the lack of experimental methods whereby the functions of lncRNA can be tested. Our study published in Nature Communications has identified one of these lcnRNAs with a dual function- one that monitors the speed that cells go through their cell cycle and other that maintains the shape of cells.
Small interfering RNAs (siRNAs) and a protein called Argonaute
To work out what a gene does we can literally knock it out in the lab by using enzymes to cut it out of the DNA. This ensures that no protein is made and the resulting effect on the cell can be ascribed to the function of the knocked-out gene. A knockout approach for lncRNA poses a problem because researchers can’t be sure that no other regulatory sequences are also being removed.
A better method is to prevent protein synthesis by degrading the RNA, using a small synthetic RNA molecule. The small interfering RNA (siRNA) approach reduces the amount of RNA produced and is called knock-down (as opposed to knock-out). A protein called Argonaute 2 (AGO2) is attracted by siRNAs. It is part of a gang of proteins that chop up and destroy the siRNA-targeted RNA. We therefore explored whether siRNA could be used to identify functions of a lncRNA known as GNG12-AS1.
Frontend targeting by siRNA inhibits its transcription
Under the microscope GNG12-AS1 RNAs were visible at the site of its transcription and also at various other parts of the DNA. To work out what it was doing at these sites we knocked it down with siRNAs designed to target various regions of the transcript. We were pleased to find that these siRNAs were remarkably efficient at knocking down GNG12-AS1. To our surprise we found that siRNAs targeting the frontend of GNG12-AS1 resulted in an increased expression of its neighbouring gene DIRAS3, while siRNAs targeting the middle or tail end had no effect on DIRAS3 expression. Frontend targeting of GNG12-AS1 recruited AGO2, which chopped up the emerging GNG12-AS1 RNA before it got a chance to complete its transcription. In contrast backend targeting of GNG12-AS1 allowed it to be transcribed fully, before being chopped up by AGO2. These results mean that the expression of DIRAS3 gene is normally subdued by its neighbour and when the neighbour is prevented from being transcribed, DIRAS3 can be expressed more readily. When GNG12-AS1 was chopped up after it had been transcribed, DIRAS3 was still suppressed since it was the transcription of GNG12-AS1 that suppressed it rather that the RNA of GNG12-AS1.
Without GNG12-AS1 cells change shape and migrate faster
Regardless of where we targeted the siRNA to GNG12-AS1, when the level of GNG12-AS1 was reduced, cells underwent morphological changes and started migrating. By doing a genome-wide expression profile we identified several coding genes that were upregulated when GNG12-AS1 was knocked down/absent. Bioinformatic analyses predicted that these genes would be involved in tumour metastasis and we confirmed experimentally that the cell migration we observed with GNG12-AS1 knockdown was due to the upregulation of these metastasis-signalling genes.
Coding and noncoding tumour suppressors
DIRAS3 is a well-known tumour suppressor that is often silenced in breast and ovarian cancer. It suppresses the growth of tumours by slowing the progression of cells through the cell cycle and promoting cell death. Too much DIRAS3 will prevent normal cells surviving and GNG12-AS1 helps to modulate DIRAS3 at the right level. In addition to maintaining DIRAS3 levels, GNG12-AS1 has a tumour suppressor role of its own, preventing the expression of a whole network of genes involved in metastasis. Genes within the metastasis network cause cells to change their shapes, break away from their group and start migrating faster to invade a new site. So far, we have found that GNG12-AS1 is often silenced or lost together with DIRAS3 in breast cancer. Next steps will be to see whether this is limited to breast cancer or whether GNG12-AS1 is a tumour suppressor in other cancers as well. More analysis is required to work out exactly how and where GNG12-AS1 suppresses the network of metastasis genes.
In the near future, so-called transcriptional “noise” may well be exploited to develop more successful, better targeted cancer interventionist therapies.