Frank Laukien - The Evolution of Evolutionary Processes in Organismal and Cancer Evolution

You see my screen? We have the slides. Are the slides coming through?

I get a thumbs up. That looks like my slides, thank you, Tamra.

And in full screen mode. I guess I'm in presentation mode; I did not mean to do that. Well, in the interest of time, I'll stick with that right now. So after two incredible days on Wednesday and Thursday, I thought we just couldn't match that possibly on the third day, but I have to say that the morning and early afternoon speakers have just been spectacular. It is really amazing how many new insights we are getting into cancer, how complex it is, but also how much reason there is for hope as we have more insights. We are at the intersection of so many different fields, all the way to bioelectric morphology and information theory as presented by Perry this morning. I'm really pleased that we're moving beyond mutations and methylation; they are very important but only a part of the story. We are focusing on larger genome destabilization, genome chaos, the karyotype, and moving beyond genetics to the molecular phenotype of the cell. The proteomics talks and post-translational modification talks, for instance by Michael Gillette today focusing on phospho sites, have been remarkable. These PTMs are not templated by DNA, nor are the metabolites that Elizabeth has been focusing on. Seeing not only a focus on real-time genome evolution but also on the molecular phenome is something I really appreciate. I'll skip a bunch of slides and go straight to glycomics. If proteomics and post-translational phosphorylation aren't hard enough, we need an even bigger lamppost to look into glycomics because it is also an alphabet or code of life. It uses different logic, is not well understood, and I'll try to convince you that it is also very important in cancer. Beyond looking at cancer cells of the genome and molecular phenome, another great trend is looking at the host. The stroma and extracellular matrix are also changing and have nucleic oscillation levels; they are enabling cancer as a clinical phenomenon. Michael Levin gave a shocking and important talk about interactions between cells, bioelectric interactions, and tissue morphology. The picture gets more complicated, but we are becoming broader in our thinking, pulling out of some partial dead ends. By pulling up a level or two beyond genome into the molecular phenome and looking at the host longitudinally, we have a chance of getting a deeper understanding and therefore know how to detect earlier, treat earlier, and perhaps come to some holy grail insights and therapies like what Patrick presented today. Let me talk briefly about something more conceptual before I get to glycomics. I've been harping on the evolution of evolutionary processes, both in organismal and cancer evolution. Evolvability as a trait per se. Darwinian natural selection in organismal evolution and cancer evolution is really also looking at the stochasticity of the cell, the flexibility, the evolvability. Ken Pienta has published on that. Random mutations theory had to be there at the beginning of evolution, but over time additional processes from cell biology have largely replaced random infinitesimal mutations for adaptive evolution. There are many exceptions with deleterious effects, but in adaptive beneficial evolution, random mutations play a minor role. This changing view of limitations of the modern synthesis is having a big impact on cancer evolution. Cancer evolves not only with somatic mutation theory but also with many other genetic and karyotype rearrangement processes. Bob Weinberg today said that once you have multi-clonal species, you don't have time for selection; it cannot be Darwinian because it proliferates so quickly. As adaptive changes accumulate, they evolve more efficient evolutionary processes and faster evolvability. In real-time multiclonal cancer cell quasi-evolution, we have examples of major genome rearrangements, whole genome doubling, aneuploidy, polyploidy, multinucleated giant cancer cells, and the epigenome. There are amazing papers by Grail showing incredible specificity and tissue of origin characterization, but they weren't that sensitive in early stages. For that, you need to look at the host, the host interactions, and the immune system to get higher sensitivity. The host is the most sensitive detector before we can detect it any other way. The host's immune system is the best drug maker. I will now go to a specific topic: glycosylation. The glycocalyx of cancer cells gets more and more decorated. This decoration is not junk; it is functionally important. It is even more complex than the proteome. We observe empirically that as prostate or breast cancer progresses, the cancer cells and the stroma become more glycosylated. We need to study the dynamics of tumor and immune glycosylation because they will be critical to improving immunotherapy. We have to look at another code, the sugar code, a third or fourth alphabet of life. It requires higher information density, uses branched molecules, fuzzy logic, and is incredibly hard to understand. But it is very important empirically in cancer progression. Since I'm out of time, I'll leave it at that and take a question or two. Thank you very much.

Yes, I mean, I think there's the narrow definition of epigenetics: DNA methylation, histone code, chromatin structure. But if you look at the broad definition of epigenetics, where post-translational modifications like phosphorylation or glycosylation are included, because they are not strictly part of the gene and not templated by the genome, I think it plays a tremendously important role in cancer evolution. I take a broader role of evolution than just genetics or the genome. Any molecules that can store biological information inherently become evolvable. In long-term organismal evolution, epigenetic inheritance exists but only for a few generations, so long-term evolution is at the genome. But in cancer cells, they evolve so fast and proliferate so fast that you have multiple generations, and epigenetic fast inheritance, whether DNA methylation or phosphorylation or glycomics, is very important and plays a role for the next progeny of the next cancer cell generation.

Yes, I think from the detection side, you need sensitivity and specificity. You need to look at the genome and things related to the genome as being much more specific, just like in viral detection. But you also need to look at the host response, which may be more sensitive but inherently less specific. I don't know what that was.