Hübner: "Genetic basis"

Thank you. Yeah, thank you Tito's and thank you Professor Bagga and the scientific organizing committee for inviting me to be here. It's really a pleasure. Now what has not been a pleasure is to fill out these things, and it was really, really complicated, so I hope I did everything correctly. So I have been told to show this. So what I would like to talk about is a specific aspect of gene expression that we have not been probing into very deeply. So as you know, if we talk about gene expression, especially in cardiac diseases but also in heart failure, we mostly look at changes in the mRNA and in transcriptional changes and how the genes are regulated. But of course, this is only half of the story, because the RNA is then translated into protein, and this is a post-translational and post-transcriptional process that is being moderated by many, many, many factors. And we have been using a certain method to actually close that gap from RNA to protein a bit better, using a technique that is called ribosome profiling, where we actually look at a snapshot of actively translated RNAs. So basically, what this is, is that if we look at the transcription of a gene, we look at the entire coding sequence including the five prime and three prime untranslated regions. But if we look at the gene body and where the ribosome sits, we can actually see which portion of the genome is actually actively translated into proteins. And we can then map those ribosome-protected reads to the specific genes, and we can show which genes are translationally regulated and not only transcriptionally regulated. This is just to remind you that the ribosome is basically hopping over the RNA in a periodic movement, the three-codon periodic periodicity. And we cannot only see if a gene is occupied by ribosomes, but we can also see whether it hops along the RNA with the periodicity that we would be expecting. And this is just a summary of all ribosome-protected fragments in the genome of a particular subject in the heart. And what you can see is that most of these occupancies are actually moving in a periodic way, so we can actually infer that this is really actively translating the gene of interest. So we have done this in a number of heart failure patients, in dilated cardiomyopathy patients that have been transplanted, and controls, and we have used those 80 cardiac samples to do exactly what I've just described: look at the translationally active genes. And I would like to show you some interesting aspects and perhaps novel aspects that are also important and of clinical relevance. So what we can do with this data is we can actually look whether transcription on the RNA level and translation on the ribosome sequencing level, so on the actively translated genes, differs between patients and controls. We can use this information of actually translating events to look at genetic variants that we are believing to truncate a particular gene and protein of interest, a stop codon for example as a mutation. And we can look at novel translational events within the genome, for example if we can detect novel proteins in the genome that previously have not been annotated. I would actually like to concentrate on protein-truncating variants and would like to showcase this on a couple of examples. As you know, titin is the most commonly affected genetically affected gene within dilated cardiomyopathy patients, and about 15 to 20 percent of DCM patients carry a titin truncating variant, so that there is a stop codon introduced and the reading frame is actually disrupted. When we look at transcriptional regulation of the titin molecule and on the translational effect of the titin molecule, we see that some of the perceived differences from the RNA are actually ameliorated, and we don't see actually a very strong effect on different titin isoforms that we have previously detected from the RNA. So the effect is actually blended if we look at the protein level. However, if we look at mutations of titin truncating a titin truncated alleles in human hearts and DCM hearts and in normal hearts, we see in our data set that 13 out of the 65 DCM patients have a titin truncating variant, so the reading frame is disrupted and the full protein should not be made. Normally, what happens or what we perceive actually is that the RNA is then degraded, and we call this process nonsense-mediated decay. If we look at titin truncating variants and compare titin truncating variants between those that carry a titin truncating variant and those that do not, the wild type, the healthy subjects, we see that there is basically no expression difference and no differences in nonsense-mediated decay. That leaves us with the question: what happens actually to the RNA and to the protein? So when we have a truncated protein because of a stop codon, we would expect this to be unstable. The question is: is it incorporated? The remaining part of the protein incorporated into the sarcomere? And is there a rescue of the deleterious effect from the wild-type allele? And could there be a translational rescue as well? Could there be that the stop codon actually is not functioning fully as a stop codon? Is there a reinitiation of the ribosome downstream of the stop codon? And actually, is there a possibility to synthesize the full-length titin protein? And this is actually quite surprising within the data set, and that makes also the interpretation of genetic variants in our view more complicated than we just perceive as from a simple view that a stop codon always results in a truncated protein. That certainly happens that it comes to a full stop, and this is depicted here on the left side. You see the RNA plotted before and after the stop codon of the titin truncating variant. And since one of the alleles is unaffected and the other allele is affected in a carrier of a titin truncating variant of a DCM patient, we would expect that the RNA before and after is actually the same, and we see this exactly here. But if we look at ribosome occupancy, we would expect that the RNA is actually dropping off after the stop and the protein is not being made after the stop codon, and this is what we exactly see in some of the patients, some of the 13 patients that we have looked at. This is depicted here in a different graph. So we can see the occupancy of the RNA from both alleles of an individual that carries a stop, which is depicted here, and we can see protein that is being made, but after this stop, the allele that is truncated, no protein is being made, and it only comes from the healthy allele. So this is all in order and as we would expect. However, if we look at other patients with a stop mutation, we see a scenario where we do have residual protein being made after the stop codon, and this is obviously raising a lot of questions. And this is depicted here as well, the residual protein that is being made from a couple of other patients. So what actually can explain the translation of downstream titin truncating variants? And what you can actually appreciate from this slide is where we have plotted the ribosomal occupancy around the stop codon of interest between wild-type and those alleles that carry a stop codon, and what you can see is that the ribosome actually is shifting across the stop codon, and we believe that this is not resulting in a full stop, but it is a slippery stop codon, and the residual protein is actually being made also from the disease allele. So there are different scenarios that we could actually appreciate, and we could say: okay, there is a frameshift mutation and a reinitiation after the frameshift mutation, after this stop mutation, and that could result in reinitiation of the protein beyond the stop, and then we would have the production of an N-terminally truncated titin truncating variant. So we have protein that is being made after the stop, and this could be even more deleterious than if this protein would not have been made. And we have shown that there is a read-through in titin truncating variants that we have engineered in a rat model of the human disease, and where we have engineered is this mutation a full stop, and there is also a read-through of this stop mutation in this rat model. So we think this could be a slippery stop codon and it could be a read-through of what we perceive as a full stop and canonical stop. Now we have then gone on and introduced a certain type of truncating variants that we have detected in our human DCM patients into pluripotent induced stem cells, and we have differentiated those iPS cells into cardiomyocytes. And what you can see here is we have engineered the titin truncating variants together with John and quick excitement? This iPS cell line has been made in their lab, and we have then looked at the RNA synthesis of the titin gene in these cell lines, and what we can see is that the RNA is made across the entire spectrum before and after the stop codon that has been engineered, and as we would expect, and then protein is being made from both alleles before the mutation, and then we have a reinitiation of the protein synthesis after many, many thousand amino acids later. So there is an additional protein being made after the stop codon, and this of course is basically supporting our theory that there may be reinitiation into molecules that carry a full-stop mutation. So in conclusion, and I think this raises important questions also for those of you that are active in genetic counseling, because we see that as I've shown you for a particular disease gene, in this case titin, that the isoform production is under extensive translational control. It's not only the RNA; the truncated alleles are not subjected to nonsense-mediated decay. That raises the question: what happens to the RNA and to translational events on the RNA, especially in such a large molecule as titin? And titin truncating variants do not always terminate translation efficiently, and the translation can renew downstream of the titin truncating variants. And the consequences of these translational events are currently not known, and we believe strongly that they can contribute to a variable expressivity of the titin molecule with additional peptides coming from the molecule that interfere with the disease severity of the disease. Do I still have time? Good. Then I would like to go on and tell you a little bit about the point of detecting another cardiac microproteins. And as I've told you at the beginning, we can use ribosome profiling also as a measure of novel translational events. And we have selected a number of known genes that formerly have not been annotated to be protein coding, and they are called long non-coding RNAs. However, in the past, sporadically there have been evidence generated that those long non-coding RNAs could actually encode a protein, and this work has been pioneered by Eric Olson and Dallas. We have done this now on a genome-wide scale and have looked at those long non-coding RNAs and looked for other protein coding events on those long non-coding RNAs or lncRNAs, and we can actually see many, many hundreds of lncRNAs occupied by small open reading frames. Now we have then gone on and characterized those microproteins in in vitro systems to show that they really can be expressed, and we have also used methods like mass spectrometry to actually identify those peptides in vivo from tissue from human hearts. So we cannot only say they are occupied, there are translation events based on ribosome sequencing, but we can also detect these peptides by mass spectrometry in human hearts. So this is just a summary of the different areas we are going into with those novel translational events, and we name them microproteins. So we can actually look at the differential expression of those novel translational events, of those novel microproteins, in heart failure and in controls, and we see a number of those that are strongly regulated, perhaps indicating that they can interfere or are linked to the disease process. And we will be able by novel technologies, for example single-cell technologies, to pinpoint some of those microproteins to distinct cell types. So we have an atlas of those translational events, whether they are specific to a certain cell type or whether they are ubiquitously expressed in the heart. And we can then look at gene-gene correlations, and we actually have identified that many of those microproteins are actually interfering with mitochondrial processes, so they may play an important role for the energy homeostasis in the heart. And of course, it is interesting whether those novel translational events are actually conserved across species, and they are very primate-specific, and that is the reason why they have been so long not been recognized and can only be detected with experimental methods. Last but not least, some of these microproteins are super interesting from their predicted features because they are predicted to be either signal peptides or transmembrane domains, and some of them are also predicted to be released from the cell and may represent factors that are important for cell-cell communication. So with that, I would like to summarize what I've been telling you. First of all, we can use translational and transcriptional events to better explain what gene is dysregulated between heart failure and health. We can look at truncating variants and can show that they often inefficiently terminate translation and may be contributing to the variability in disease phenotype. And we have catalogued a previously undetected number of proteins with potential functions in cardiac energy metabolism and cell-cell communications. And with that, I would like to close and thank not only the people from my lab but also our collaborators from around the world, but also within Berlin. Thank you very much. Thank you very much. I think this was kind of a foeman on start of our...