We set up a titration testing 11 concentrations each of remdesivir and hydroxychloroquine (HCQ) each on 4 cell lines (HepG2 and PLC/PRF/5 liver lines, HCT116 and HT29 intestinal lines). Our primary goal is to find a CC50 concentration, defined by us as a drug concentration for which there are 50% as many surviving cells after a 4-6 day period in the presence of each drug. Note that this is not necessarily the standard way CC50 is defined, but it best enables a CRISPR screen to determine host genetic dependencies for drug toxicity. To be more specific, this titration would give us the dose needed to perform a genome-wide CRISPR screen where we determine whether the knockout of any gene induces more or less cell survival over a 4-6 drug treatment period. This is the simplest CRISPR screening protocol since it requires relatively little drug feeding (re-feeding every 2-3 days, this means two doses of the drug) and doesn’t require any cell sorting, as we will simply collect genomic DNA from surviving cells after the treatment period.
We note that we have had lengthy discussions about whether our use of cancer cell lines and not primary cells is problematic here. We think it may be to some degree. However, currently we are moving forward as is because (a) we have these cell lines, which allows us to generate some data quickly; (b) primary cell lines are typically not usable for genome-wide CRISPR screens. So, we could use primary cells for RNA-seq, which may well be useful, but having RNA-seq on these cell lines that we will further subject to genome-wide CRISPR screening does provide value in comparing pathways where gene expression is altered and genes influence drug-dependent cell survival. We are still considering whether to perform RNA-seq on primary cells, although we would first have to acquire them.
We will also use this titration to determine an appropriate dose for each drug in each cell line for RNA-seq profiling. We would want a dose that elicits a cellular response so that we can read out differences in the transcriptome without inducing massive apoptosis since apoptosis might induce its own transcriptomic profile that swamps the specific drug-response pathway signal.
We plan on monitoring cell survival and proliferation visually every day for up to 5 days and performing flow cytometric counting of live cells at day 2 and day 4/5 (ensuring that the control well does not grow to 100% confluence).
We have attached the protocol used for this experiment as well as the daily observations.
As a brief note on visual inspection up through day 2 which is reflected in the linked Excel sheets:
· On day 1, remdesivir gave complete toxicity for all cell lines at 400 uM and showed some signs of toxicity at 200 uM. No clear signs of toxicity were noted at lower doses.
· On day 1, HCQ uniformly induced near-complete cell death at and above 400 uM and showed signs of toxicity at and above 150 uM. HepG2 cells showed signs of toxicity at and above 60 uM.
· On day 2, both drugs showed more signs of toxicity upon visual inspection. For remdesivir, concentrations at or above 60-100 uM gave uniform toxicity. The approximate CC50 upon visual inspection was 5-10 uM for HT-29 and HepG2 and 10-40 uM for HCT116 and PLC/PRF/5.
· On day 2, HCQ showed uniform toxicity at and above 150-200 uM. The approximate CC50 upon visual inspection was 60 uM for HT-29, 80-100 uM for HCT116 and PLC/PRF/5, and 20 uM for HepG2.
We performed flow cytometric analysis and will post data analysis soon.