Here are a few of the problems I've heard suggested that might make cloning of genetic engineered cells more difficult:
1) Cell lines are very sensative and may sometimes loose parts or all of some chromosomes, but continue to live.
Cell lines may not have all the DNA of the original organism. Primary cells lines are those that are taken straight from an organism, and so should have as complete a copy as possible. The ability of a cell line to grow indefinately may actually depend upon errors, such as missing chromosomes. Fewer chromosomes to replicate might speed up cell division, making the damaged cells even more fit in cell cultures than intact cells. I will place links to web pages devoted to the details of mammalian cell cultures here as I find them.
2) Gene Targeting is difficult. Each new gene needs a resistance type gene associated with it such as resistance to 6-Thioguanine.
Gene targeting works by creating a strand of DNA with regions homologous with the DNA in the organism on each side of the DNA you wish to place in the organism. A region is also added, such as resistance to 6-Thioguanine, so that only cells with the corrected targeted gene can survive in it's presence. Correcting multiple genes might require multiple additional genes that help the cell survive multiple adverse conditions. Selecting for DNA in bacterial cells involves anti-biotic resistance. Eukaryotic cells in culture must have similar resistances for selection. This requirement for multiple selections may place an upper limit on the number of genes targeted. Plus, the resistance gene would be present in all future organisms derived from the cell line, making further change even more difficult.
I've been considering a gene construction that could overcome this problem. I thought of this on March 26, 1997. Here it is:
You have first a long region of homologous DNA 5' to the targeted gene, then the replacement gene for the targeted gene, then a small string of homologous DNA 3' of the gene, just as a buffer region. Follow this by the 5' sequence of a transposable element, then the resistance for 6-Thioguanine, then a heat shock promoter (or some other activatable promoter) and the transposase gene specific for the transposable element. You follow all this by the 3' sequence for the transposable element, then a good long region of homologous DNA after the gene.
This, in theory, would give you the resistance, but a heat shock could cause the transposable element, with the resistance gene, to pop out. Keeping the excision precise and not getting another transposable element in the same spot would probaly be another problem. Also, there are many copies of mammalian transposable elements in a mammalian cell (maybe 50,000). You definately don't want them all hopping out, only your one transposable element. It's possible that the transposable element and transposase gene from some other species might work in a mammalian cell.
Even if this approach worked, the most studied transposable element I know of is the P-element of Drosophila. It is not very good at "Popping Out" unless it has a good chromosome (without the P-element) to use for repair. If both chromosomes have the P-element, it just copies the P-element on the other chromosome when excision occurs. The lab where I work, we tried to excise a P-element by putting a balancer chromosome as the other chromosome, without the P-element. We wanted a bad excision so that the adjacent DNA would be yanked out too. It excised precisely every time because the balancer chromosome had a homologous region that the DNA used for repair.
With a P-element on both chromosomes, it's likely that the P-element would not even be excised at all, because the other chromosome with a P-element would be used for repair. Perhaps electroporating plasmid DNA, with the sequence of the repair site you want produced, into the cell just prior to activating the transposase gene could greatly increase the precise excision of the P-element by serving as a template for repair.
Transposase also tends to pop the transposable element back into a chromosome at another point when it's excised. Transposases have actually been altered so that they can excise, but re-introduction function has been knocked out. You would need such a knocked out transposase to prevent additional DNA damage.
This theorizing is getting too far. If you'd like to help me do this research, first an experiment to see if Drosophila P-elements can be activated in yeast might be worth a try. If successful, trying to generate precise excisions without a chromosome to use for repair, by electroporating the desired sequence into the cells, would be the next step. Then try the same thing with a targeted gene replacement adjacent to the P-element. Then move up to mouse ES cells instead of yeast.
This is the course I would take, constructing some of the genetic tools needed for targeting and "cleanly" replacing genes in a mammalian cell line, without leaving lots of extraneous DNA behind in the process.
Arthur Kerschen
Molecular and Cellular Biology Department
University of Arizona