Ultra-safe cell line

Jef Boeke


video: https://www.youtube.com/watch?v=ZF2wFdMFn9Q&list=PLHpV_30XFQ8R2Kpcc1pwwXsFnJOCQVndt&index=3


So the projects that we probably talked the most about is what we would like to call the ultra-safe human cell line. The basic idea is to design a cellular genome for safety at multiple levels. We envision that such a cell line or cell lines could be used in biotechnology and in stem cell therapies. I should also point out that Isaacs was supposed to give the second half of this talk. I have incorporated an abbreviated version of his talk. This represents both of our work, though you won't be hearing from him.

The genome as you probably know is mostly DNA that does not encode proteins. This tiny slither here is about 2% of the genome roughly, 1-2%. To put it into perspective, that's still the equivalent of 3 to 5 of these yeast genomes which we are still just over half-way towards assembling at this point. An interesting aspect of this is that because it's split up into small pieces, to do this kind of engineering, you don't need the capability to make giant DNAs, you can work with relatively garden-variety DNA that is manipulated in labs every day. This is a project that once certain decisions were made about design, you could literally start on this tomorrow. Now the idea would be to engineer in to such a cell line many features aimed at making those cells safe. Just to reiterate something I said before, we would envision doing this inside of an iPS cell line that had a germline firewall in it to ensure that these would be used for somatic applications only.

So for biotechnology, the products made would be less risky for example as we just heard from George there is at least for bacteria possibility to engineer for multi-virus resistance. We could make a human cell line that would be unable to support the growth of viruses and therefore not be contaminated by viruses.

Another major application would be in stem cell therapy. The idea would be to engineer those cells so that they would have a reduced chance of these cells giving rise to cancer, producing prions or other undesirable traits. We believe that through engineering that this could be accomplished.

Let me give you an example. There are at least 35 well-defined tumor suppressor genes, and many others that are proposed, in our genome, to help keep cancer in check. We can case-harden these genes against mutations in various ways. A famous tumor suppressor gene is the p53 gene, so called guardian of the genome. What you see here is a map of the mutations that arise in cancer. What you can see first of all is that almost all of these mutations are in one part of the gene and they are extremely variable. There are some spots that are off-the-scale here, and these are the hot spots, and then there are others that are much less frequently utated. The impact is huge in that p53 is mutated in 50% of human cancers and therefore by implication one could, if one never had mutations in p53, one would anticipate a major impact on the probability of cancer, and most of these mutations occur in these hot spots. Of those 9 hot spots, 7 fall in codons that contain the CPG... all of those codons could be recoded to other codons that didn't contain CPG, which could then specify the same amino acid, but then would lose their ability to be hot spots for the mutation. This is a somewhat specialized case because p53 is unusual in having many point mutations. Many suppressor genes are lost by deletion and other loss of function changes. The inspiration from nature is that some very long lived animals have extra copies of p53. Potentially adding extra copies of these tumor suppressor genes might be another way to improve the cancer resistancy of these cell lines.

  • Retrotransposons jump around in the genome of stem cells (Klawitter et al., Nat Comm. 2016)

My own career started in studying transposable elements which are parts of almost all genes. The orange in this diagram is the non-transposon components, and the other ones are the mobile DNAs or transposons that have the ability to jump around and insert into genes that cause mutation. These elements have been recently shown to jump in a somewhat uncontrolled way in the genome of stem cells. Since many therapies are in the works using these cells, this is an important thing to be aware of, and potentially something that could be engineered out. Our recent work from George Church's lab shows that it is possible to do this with at least in a limited number of elements and through synthesis I think this could be done in a very comprehensive elegant and precise way. There are about 100 active so-called retrotransposons in the average human genome and these could be precisely removed from a stem cell genome, preventing such instability. This is a direct decedent of one of the components of the synthetic yeast genome where we also removed these active elements.

Other genome modifications could be added for free if you like at the same time. George mentioned the idea of a base-line genome which would contain the ancestral alleles of genes. In this particular pilot, we're talking about the coding region only. So there are something like 10,000 of these on average, non-synonymous variants in every coding genome, and we could infer which of those were ancestral in most cases and thereby make a kind of baseline human genome, if you will. This would then allow a very clean way to assess the impact of any individual coding variant on that background cell line. These could then be systematically tested. Secondly, it would be possible, whether it would be wise is a matter for debate, to include the genome scrambling system which some of us think would be a useful tool for learning about the human genome.

The major thing that we would be doing is try to engineer virus resistance. You already heard George Church mention this. There are some people doing this in ecoli. By eliminating some triplets from the genetic code genome-wide, it should subsequently be possible to eliminate otherwise essential components of the translation machinery that reads virus code. Since all viruses depend in an absolute way on translation, this would be an extremely effective way to eliminate their virulence. This would be a practical engineering of the human genome that I think justifies certainly doing this pilot.

Viruses and cells share the same genetic code. This is something that allows for horizontal gene transfer, which we know from a study of genomes, has been pervasive both among the bacteria and from the bacteria to the eukaryotes, so our mitochondria and chloroplasts of plants were transferred into our genomes or into our cells via these horizontal transfer events. And so, a benefit of this kind of code engineering is that it's a way to block this. We specifically have been talking about virus infection. And the practical impact of this is actually quite significant... I just realized when I was preparing for this, I took the polio vaccine when I was a kid. I didn't ask my mom which one I got. I almost certainly took the vaccine that was contaminated with the monkey virus Sp40 along with millions of other people around the world. Fortunately there's no evidence that Sp40 did us any harm except give us crazy ideas I don't know. But this is a kind of contamination that would not have happened if there was a safe cell line lying around. More recently there have been some big disasters in industry that led to contamination of drugs, made with cells, by VC virus. It shut down an entire plant and led to a lawsuit and so on. We hope that this project might be useful to industry by developing lines that would never have this liability of being infectable by a virus.

And then as I said, there's a larger issue of horizontal gene transfer where it's obviously on people's minds with regards to recombinant DNA in general and applications of it in the field. In addition to engineering virus resistance and generating stable cell lines for manufacturing biologics, there are lots of other interesting ideas that can be applied in terms of new kinds of therapeutics that could be enabled by a modified translation system. This idea of safe-guarding and bio-containing the cells.

Exactly how this is going to be done, details remain to be worked out. It would probably be done by deleting one or two codons from the code and substituting them with synonymous codons and then deleting the tRNAs that recognize those particular codons, and doing this genome-wide. These are the open questions-- which codons should we target? How do we check that the code changes don't cause problems? At what parallelization? What multiplexity?

Thank you.

Q: You talked about getting rid of the... as a safety measure... you could differentiate stem cells into... I think you could get a gamete out of that. In your desire to tease apart the chromosomes, it would be helpful to have lines that have deletions and inversions. And given that, if one takes into consideration the non-disjunction as a mechanism for safety, those inversions and deletions have a safety value as well.

A: Very good point. Yes, we could balance our chromosomes.

Q: Can you elaborate on-- the ancestral orphome I guess. If it doesn't work, I guess you learn from it, but it seems to me it's going to find a sequence that would find the right proteome for a cell to be viable.. We all have variants and alleles, you know, 10,000... could you elaborate on how an oprhanome is going to leave to a viable proteome and all the different combinations.

A: I would be very surprised if by making such an orphanome that we made a big impact on viability of cells. Virtually every variant allele in the human genome has a major allele and a minor allele, and for the most part (although there are exceptions to this) those major and minor alleles are shared across populations. For the vast majority, you can identify the major allele rather easily, and that's what we would be using. I would say from first principles that's likely to work. There are some interesting exceptions specific to non-synonymous variants that are highly enriched in caucasians, so decisions would have to be made there as to are we going to make a caucasian or an african genome, that's an interesting unresolved question that remains to be tackled.. But I think the number of changes of that type would be small enough that it would be possible to make different versions.

Q: Hi Jef.

A: Yes.

Q: For those retrotransposons, do you think they are necessary evil or are they a parasite? Are there any bad consequence of getting rid of them?

A: Very good question. Might those transposons be doing something good for us? This is the question that always comes up at the bar at the transposon meetings. It's been doing so ever since I entered the field. With the yeast genome, we're going to have an answer to that question pretty soon. So far we haven't been able to connect any essential function to transposons. There are many transposons that died out, but still present to service... garbage you throw out, junk you save. These old transposon genomes are saved and they accidentally turn into an enhancer or something what have you. So it will not be wise to remove all copies of transposons, because there would be hardly anything left. We would be targeting just the actively mobile ones which presumably don't have important functions, but if they did then that would be an interesting scientific result to learn about.