Within STEMCARDIORISK it was key to initially establish how sensitive the hPSC models were for investigating genotype-phenotype relationships. This is achieved by using genetically-matched cell lines in which the cardiac disease-linked genetic variant is the sole modified variable. In our first iteration we used CRISPR/Cas9 to introduce specific heterozygous ion channel mutations into a wild-type hPSC line. In patients it is known that the region of the ion channel containing the mutation influences the risk of an arrhythmic event occurring. In our genetically matched hPSC models we also detected such changes in the cardiomyocyte’s electrophysiology.
We subsequently expanded on this and determined that we could also use these hPSC disease models that we generated to detect differences in disease severity caused by the presence of commonly occurring modifying variants. We believe that by performing these investigations in very tight genetically controlled conditions we can address some of the discordant results in literature of whether some of these variants exacerbate or reduce the severity of the disease in patients.
To be able to address these questions we have needed to reduce heterogeneity by refining our cardiomyocyte differentiation strategy to improve yield and purity. This has also involved establishing a panel of cardiac markers to ensure that the resulting populations for comparison between the different lines are equivalent, and developing cryopreservation protocols so that the same stock of cardiomyocytes can be used over multiple assays. These developments have clearly reduced the level of variability we observe in our measurements thereby also improving the sensitivity of our models to detect subtle functional differences.
Concurrently we also developed a novel allele replacement strategy that enabled the rapid and simultaneous generation of cell lines containing different individual genetic mutations while maintaining the correct genomic context. Where once it would have taken us more than one year to generate 12 lines of different targeted genetic mutations, we can now do this in less than 2 months. Not only can we use this to generate a range of different models of known mutations with different phenotypes, but this approach can also potentially be used to classify the pathogenicity of variants identified in patients for which their clinical relevance is currently unknown. We are now investigating further how this could be employed diagnostically.
Finally, we have also demonstrated that we can use the hPSC disease models we have developed to detect mutation-specific drug response differences. Overall, the methods and models we have developed within STEMCARDIORISK will be of important value in the areas of drug safety and personalised medicine, particularly with regards to improved individual risk stratification and patient-specific pharmacotherapy.
The project has also resulted in 14 publications to date, along with ~16 poster and (invited) oral presentations at ~15 conferences, and a submitted patent. Members of the group employed on this project have subsequently found employment within the same field in both academia and industry, as well as successfully obtaining their own independent funding. Results from STEMCARDIORISK have also formed the basis for several new proposal for funding, including successful applications to Health~Holland and the Dutch Research Council.
