Much information about the role of specific genes in fundamental biological processes and the onset and progression of genetic disease has been gleaned by researchers having the ability to selectively alter the genomic composition of individual genes and study the consequences. This approach enables researchers to observe the effects of a specific mutation, single nucleotide polymorphism (SNP) or deletion in combination with the added layers of regulation present within the cell, including post-translational modification, epigenetic changes associated with chromatin structure and transcriptional mechanisms.
While the Nobel prize-winning work of Capecchi, Evans and Smithies introduced the concept of manipulating the genome of mouse embryonic stem cells, the ability to manipulate a broader range of cell types, human cells in particular, remained a significant challenge for some time. The more recent discoveries of nuclease-based targeting technologies like zinc finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs) and clustered regularly interspaced short palindromic repeats (CRISPR) has greatly increased interest in genome editing and provided even more efficient platforms for achieving targeted genome modification. Despite the increases in efficiencies these technologies offer, there are still a wide range of factors that influence success and failure in genome editing.
While the scope of genome editing is very broad and includes whole organisms, this article will focus on the issues faced primarily by scientists attempting to modify the genome of immortalized cell lines. The ability to create isogenic cell lines in which the genome editing event is the sole differentiator between the phenotypes of two cells is a powerful tool. Despite all the recent developments and improvements in targeting platforms, certain challenges remain and the role played by the choice of cell line in order to achieve success cannot be overstated.
Knowing your line
The majority of genome editing today is being undertaken with nuclease-based technologies (e.g., CRISPR, TALENs and ZFNs). A typical gene editing experiment involves choosing a cell line, growing it up in suitable numbers, then introducing the nuclease platform into the cells, followed by a period of recovery, growth and screening to identify successfully edited cells. Understanding some of the specifics of the cell line being targeted, regardless of the platform being used, is essential.
Many immortalized lines contain significant genetic alterations already—and a good majority of them contain more than the expected standard of two alleles. If a homozygous alteration is required, it could be considerably more difficult to achieve this with high allelic copy numbers.
All three of the targeting technologies (CRISPR, TALENs and ZFNs) require exact knowledge of the genomic sequence to be targeted in order to build reagents with sufficient specificity for the target. Even a single base pair change in the targeted region could affect the ability of the nuclease platform to effectively bind and cut. Furthermore, if an experiment to knock in a gene is being undertaken, then a significant region of homology must be utilized in providing a donor template, and each variation between the actual genomic sequence and the supplied donor will contribute to diminishing efficiency. It is therefore strongly recommended that allelic copy number of the target region be determined and sequencing of the region be undertaken and any differences between alleles be taken into account.
While there is a significant body of literature on certain cell lines with regard to copy number variation and even whole genome sequence, our lab often finds that cell lines from different sources can vary significantly and that genetic drift over time can lead to important changes. Therefore we always resort to empirical determination of allelic copy number and full sequencing of the target region. The cost is reasonably low but the importance enormous.
Another very important consideration for choosing a particular cell line has to do with the need to obtain a clonally pure population at the end. None of the editing platforms has efficiencies high enough to guarantee a pure population of cells derived from a pool. In addition, each platform has some degree of off-target potential, and it is generally a good policy to obtain more than one clonally derived population to compare, to ensure any phenotypes observed are due to the targeted change and not some opportunistic off-target effect or other component of genetic drift. What this means from a practical perspective is that the cell line being used needs to be suitable for expansion from a single cell.
While pool dilution and feeder layer strategies can be employed to some degree, they quickly become impractical from a time perspective in most cases. We have observed that many lines which culture well under standard conditions do not grow well from single-cell clones. In our lab we have come to rely upon a screening strategy for new cell lines that tests different plating densities and media conditions to identify the best conditions for obtaining single-cell clones. Sometimes these conditions may only be used transiently, but can be valuable time-savers by allowing a population derived from a single cell to be expanded more rapidly. However, many cell lines simply grow too poorly from a single-cell state and become impractical for use in genome-editing experiments.
One of the common features for all genome-editing technologies is the requirement for delivery of the editing platform (usually in DNA or RNA form) into the cell. Transfection and electroporation are by far the most common methods for delivering nucleic acids encoding the components of the editing machinery. There is a wide range of transfection efficiencies between different cell lines, and knowledge about just how effectively one can deliver DNA into the cell is critical. Viability and recovery period are also important factors to keep in mind given that the cells will need to be expanded following transfection. Even the best technology platform is rarely capable of even 50-percent editing efficiency, so if only a small percentage of cells is being provided with the necessary reagents and those that do receive reagents are incapable of dealing well with the stress, the search for a correctly modified cell (particularly from a single-cell clone) can be daunting.
Some cell lines (many hematopoietic lines for example) are particularly sensitive to DNA transfection, but RNA can sometimes be used instead to overcome this. It can be good to check whether a line responds better to RNA transfection than DNA transfection. But RNA is slightly more challenging to prepare and work with, so this must be taken into consideration by the researcher carrying out the work.
Biology itself can be another major stumbling block. Is there evidence to indicate whether or not the intended genetic change will adversely affect the ability of the cells to grow? It seems obvious that one would not be attempting modifications in human cells that perhaps had been shown to be lethal in mice, but the thinking needs to go beyond this. If the alteration is expected to have even a slight growth effect on the cells then it could make the isolation and growth from a single-cell clone virtually impossible. We generally recommend in the case of knockouts that at least some siRNA experiments be performed first to see if there is any effect on growth. If the pathway being targeted is well understood, it may be possible to supplement the media or otherwise compensate the cells in some way, to help modified cells cope with the change, at least until a pure population has been derived.
As increasing amounts of genomic data become available to researchers worldwide it is more important than ever to utilize this information in a practical manner. The possibility of readily modifying the human genome within a wide range of cell types to probe how genes affect biology is now a reality, thanks to new gene-editing techniques. But in order to avoid costly delays and unsuccessful editing experiments, careful consideration needs to be given to choice and handling of the cell lines.
Careful validation of newly created lines should also be demanded by the scientific community. The use of simple (and relatively inexpensive) assays post-editing, such as SNP6 and STR profiling, to lend confidence to the fact that an engineered line has not grossly deviated from its parental source, should be expected and journals should be looking for this. As the cost of whole genome sequencing and analysis continues to drop we should begin to see this used as well to give full visibility into all genetic changes which might occur during editing.
By developing a broad range of human disease cell models that faithfully recapitulate predisposing or pathogenic genetic variations (SNPs and mutations), we could vastly improve our understanding of human disease. It is perhaps time to call for a “Translational Genome Project” that probes gene function and pathway interactions at the genetic level, a natural successor to the Human Genome Project and ENCODE Project.
Eric Rhodes, senior vice president of R&D and chief technology officer of Horizon Discovery, has a strong background in gene regulation and gene editing. Prior to joining the company, he was director of business development at Sigma-Aldrich Corp. where he helped the company in launching its nuclease-based gene-editing platform, and served for 10 years as vice president of business development and alliance management at Sangamo BioSciences.
Dr. Jon Moore, vice president of oncology and chief scientific officer of Horizon Discovery, joined the company to lead its oncology target validation and drug discovery efforts. Prior to this, he enjoyed 10 years at Vernalis, culminating in a role as head of biology, where he was responsible for both cell biology and assay development activities. He also led the biology side of several innovative drug discovery programs and managed a major research alliance with GSK.