EACR 2012

The ceRNA Hypothesis and the non-coding revolution in cancer research and therapy

Professor Pier Paolo Pandolfi – Harvard Medical School, Boston, USA




We are extremely excited and the excitement is due to the fact that, as you might know, our genome is heavily transcribed and yet we have been able to functionalise only a minute percentage of it, the 2% that encodes for proteins. So the genome is transcribed up to 50% and so far we have concentrated our attention on 2% of the transcribed genome that is encoding for proteins. Yet we know that these transcripts are there for a good reason. Actually, I think that this idea of a functionalising, attempting to de-convolute the RNA world that doesn’t encode for proteins, would be extremely important, not only for cancer but also for other diseases. I really think that we are what we are as multicellular organisms because we have that dimension, we haven’t acquired it for nothing. Again for the lay audience, the important discovery that came throughout the analysis of how other organisms utilise their genome is that the lower organisms or single cellular organisms do not have this complexity at the RNA level. So this complexity has evolved, has expanded and it is, for sure, functional. So the important contribution and the important discovery was a way by which we can de-convolute by reading the information which is in the messenger RNA itself. This language, as we call it, this code, allows us to predict which RNA talks to which RNA and hence allows us to predict which non-coding RNA regulates coding proteins, coding RNAs. As I always say, people say what now, based on this new theory, the protein coding genes vis-à-vis the protein non-coding genes? I always give this metaphor, it’s like if protein coding genes would be encoding for the hammer, for the nail, for the utensils whereby you build a house or whereby you paint a painting – you have the same colours, which by the way is not entirely true, but painters throughout the years, throughout the centuries have used the same tools - you have a canvas, you have colours, you have a brush, it’s the way by which you utilise these utensils that changes. So the utilisation of the protein coding genes is completely impacted by the non-coding dimension. This is the way by which you get the subtleties by which we… which are not subtleties, these are profound differences, whereby we are not mice or we are not zebra fish, we are mammals and we are human beings. One eye-opening aspect of all this, and clearly this was again another discovery for me, is that I have been trained to postulate that if a genetic entity is not conserved it’s not important. Here what I am saying, or I found myself saying, is that this dimension which we have acquired and so, therefore, is not shared between mice and elephants and mammals, is the one that dictates speciation which is to say that this dimension, which is not conserved, is as important. Again, this is another thing that for me as a reasonably accomplished scientist that was eye-opening because I was trained to say if it’s not conserved, it’s not important. The other important aspect is that we have used mice and mouse models and I am one of the most established mouse modellers of cancer but we have to be extremely careful because if we give weight to this non-coding dimension and if this non-coding dimension is not conserved, of course the mouse will be very informative to study protein encoding genes but clearly a bit less informative to study the non-coding space which is important if you believe that non-coding genes are as important in disease and therapy response.

Can you tell us about your work with chromosomal translocation?

There are plenty of examples by now of chromosomal translocations whereby the product is a messenger, it doesn’t encode for a protein. Again we have either dismissed those or not even paid attention to them in the past and now, on the basis of this new code which we can apply to any transcript, we could certainly attempt to give function to these ghosts that we have neglected for way too long. As I discussed in the lecture, the implication is that when you generate a mouse model, you have to take into account this dimension so you cannot simply model the genetic event by inserting in the germline of a mouse the protein coding portion. You have to pay attention to the whole transcript which is also extremely interesting. Of course, the whole transcript can give regulatory subtleties to the fusion proteins but could also be functional. So the take home of my lecture is that any messenger can have a function which is independent from the protein it encodes.

What does this mean for the disease?

It means that there is a lot to do and many people have said this is disheartening because you are implying that we missed a big dimension. I never see the half empty glass, I always see it as an opportunity. We can study now this pace and actually we are finding new disease genes, we will find new nodes for therapy and, importantly, we’ll find new means of diagnosis and certification. This is important not only for cancer but for other diseases. We have counted… this is something again, to me mind-boggling, we have counted 400 new disease loci which are mapping close to a transcribed but not translated unit. People say what is the relevance of this for cancer susceptibility, for diabetes susceptibility? The relevance is that this non-coding space was neglected and was neglected for a good reason because we didn’t know what to do with the information. But we already knew that many of the load score association between a disease and a certain familiarity would map the genetic entity in regions which were nearby transcribed elements and, again, we neglected them. I don’t know if you know that many people said for a good number of years that micro-RNAs would be irrelevant because how come micro-RNA had never been discovered in any genetic screening? The answer was very simple – because they are too small first, so people wouldn’t see them, and secondly because they were not encoding for proteins. So the beautiful story that Dr Croce gave probably yesterday, but that he gives, is that they were looking for a gene for CLL for way too many years and then there was no protein encoding gene, there was this small creature that was RNA-made and yet devastating in its outcome.