Our next speaker is Dr. Lawrence Chin. The title of his talk is Unique Nascent RNA Sequences That Distinguish Normal Brain and Clinical Neurons. I want to thank all you tumor diehards for sticking around and listening to the talks this afternoon. So, I want to tell you a little bit about the work, no disclosures, that has been done with our tumor lab in conjunction with Cornell University on a novel technique invented at Cornell by John Liss and Charles Danko called ProSeq and CroSeq. But first, I'm just going to give you a little bit of background here. As you recall, the human genome is primarily non-coding, so less than 2% of the genome is actually expressing gene product, and those are the exons. There's a huge amount of repetitive DNA, a large amount of unique non-coding DNA, and also introns, and then an area that we're interested in, which are the regulatory sequences. Now, the gene consists of what you commonly know as the gene itself, which contains exons and introns, but there are large areas of regulatory sequences, both 5' and 3' that contain enhancers and promoters that are important for how the decision is made for whether or not a gene is transcribed. So part of the reason that these elements are so widely separated from the actual area of expression of a gene is that one of the mechanisms that's used to regulate gene expression is the binding of transcription factors to the areas to enhancers, which then provide a conformational change that allows the binding of RNA polymerase II to the promoter, and it's the RNA polymerase II is what transcribes the RNA. And so you can see there is some method to the madness of why these regulatory elements are so distinct geographically from the area of the gene. So the technique that was developed at Cornell is one in which RNA polymerase II, which is bound very tightly to the DNA, and in fact stays bound to the DNA even in tissues that have been frozen for many, many years, actually can be revived in these frozen tissues and reanimated, so to speak, so that you can have samples in which all of the RNA is degraded, but the DNA still retains the bound RNA polymerase II, and this RNA polymerase II can, under the right conditions, be restimulated to start transcribing the DNA again, and you can then sequence the little portions of RNA, hence the term nascent RNA, and get a very complete description of the DNA. So this is an example of what can be obtained in the sequencing of these genes with this technique. And so for comparison, down here, this is RNA-seq data. These are areas that are actually in which RNA sequencing shows results. This is the gene up here. So you can see on this segment of DNA, here's a gene here, and the little spikes are the exons, so most of it is introns of regulatory sequences. Here are microRNA sequences. And this is the cro-seq data. So you can see there's really essentially complete coverage of the entire genome here in comparison to what you get with RNA-seq. And importantly, you get complete coverage of all the areas of the regulatory elements as well as the introns. And so it's from this data that you, and also you can get expression of microRNAs which don't show up in the RNA-seq data. So you can see that there is very complete data that can come out of this technique. And processing this technique allows you to mirror some other ways of evaluating genome. For instance, histone acetylation and DNase areas that are sensitive to DNase. These are areas that are indicative of DNA transcription because it indicates relaxation of the chromatin, which is needed for the DNA to be transcribed. So this technique really allows you to very completely evaluate the genome. So for instance, this is a GBM sample in the region of the EGFR gene. Those are exons. These are the regulatory sequences. And you can see here's CroSeq data on this particular tumor sample. This is non-tumor brain. So you can see, first of all, that there's a tremendous amount of amplification that can be seen. And these spikes here represent a way of analyzing the data using machine learning techniques that can pick out key areas of what would be regulatory sequences that correlate with other epigenetic data. So confirming that you can use this technique to pick out regulatory sequences. This is a summary of really a huge series of experiments and data processing, but 20 total RNA, sorry, 20 total brain tumor samples and three PDXs or patient derived xenografts were analyzed using this. And if you look carefully, you'll see that the oldest specimens are from 1988. So some of these samples are 30-year-old GBM samples. And you can get expression maps that can then be mapped into the traditional, what we think of as pro-neuro, neuro-classical, and mesenchymal GBM types. And then you can determine level of expression and you can also determine how many fold increase over a control non-tumor specimen. So in each of these samples, you can see, for instance, the classical EGFR is highly expressed and also highly differentially expressed in tumors over normal brain. This is a heat map that illustrates in each of the tumor samples, as well as the PDXs here, the degree of similarity of the regulatory elements with normal brain and cultured brain. And there's a high amount of similarity with normal brain and very little similarity with cultured brain and also with traditional brain tumor cell lines. And so this really indicates that when you put something into culture or you have a brain tumor cell line, this is something that really is not at all like a brain tumor. And it's really a different beast. And also shows that when you create patient-derived xenografts, that these are much more like real tumors and that they share regulatory elements with normal brain. So where this might come into play is if you look at all of the different brain tumors and divide them into similarity to a differentiated tumor types or cell types, stem cell-like cell types or immune cell types, that there are very specific regulatory elements of transcription factors, for instance, stem cell, that are highly expressed in the brain tumors using this technique. So, for instance, SOX2 or HOX or POW. And you can see that you can actually drill down on specific regulatory elements that can be picked out. And then if you look at an entire genome map, you can see that there are many, many of these elements that are differentially expressed in brain tumors over normal brain. So what has been demonstrated thus far is that this technique, Cro-Seq, is effective in a wide variety of tissues, including very old frozen specimens, and that you can get very high-resolution maps of regulatory elements that really correlate well with known epigenetic methods. This is a much more sensitive technique than RNA-Seq, and that you can develop distinctive regulatory fingerprints of tumors that indicate that stemness, or the importance of the stem cells in these brain tumors, is quite significant. And I think that this is going to be a valuable tool for understanding regulation of tumor genomes. So, obviously, many thanks to the people at Cornell, and especially Dr. Danko, who developed the Cro-Seq technique in his lab at Cornell. Thank you very much. Applause

nascent RNA sequences

gene expression

CroSeq technique

regulatory sequences

tumor genome regulation