Adapting manual cell line development workflows to automation

We have a plate hotel (for ambient storage of source labware including plates and pipette tips), incubator and plate imager integrated.  Other equipment that can be integrated includes analytical devices such as plate readers or the Octet HTX.  Even an automation-friendly centrifuge can be integrated.  Lab space and budget are the main limiting factors to the types and numbers of devices that can be integrated.  The liquid handler itself is enclosed in a HEPA enclosure to ensure sterility during culture manipulations.

We have not had the need to perform any plate manipulations/liquid handling outside of working hours.  We have run plate manipulations/liquid handling through much of a working day.  Cultures are processed by the automation system from plating through scale up, a process that takes up to 8 weeks.

We have not evaluated cell lines transfected with the same plasmid using the manual and automated workflows side by side.

Monoclonality is established using probability based on limited dilution.  We also have the ability to image plates for assurance of monoclonality using a plate imager that is integrated into the automated platform.

Initial dilution for limited dilution cloning and culture scales larger than 6 well plates (such as T flasks and shake flasks) as well as banking are still manual operations.

We chose not to automate some of the most difficult workflow steps, such as those involving capped vessels, so adaptation wasn’t as difficult as it could be.  We initially intended to use 1-well plates to replace T-25 flasks but it was difficult to move the plates on the system without spilling.  We then switched to 6-well plates.  Some scale up steps required pipetting method adaptation, as was expected.  We had to determine the best method to dislodge and aspirate the most cells with the least loss of viability.  This involved moving dispenses.  We learned we didn’t need to tip the plate, as we would in a manual process, to remove a sufficient amount of medium to transfer sufficient cells.

Our first fully automated process began 5 months after installation.  Previous testing and development of the workflow produced a functional workflow, but handling a truly representative number of plates in real world conditions allowed us to learn and improve even more.  As we went along we continued to make more improvements to operations to improve functionality and efficiency.  In a sense we are still making the adaptation, as methods could still see refinement.

The main reason we changed to automation was to allow higher throughput.  The automated system allows one operator to process many more clones than they could with a manual workflow.  This gives us the ability to screen more clones per project or handle more projects with the same number of clones.  Screening more clones could potentially allow us to find better clones.  Looking through more hay could help us find more needles.  Stretching the metaphor, finding more needles increases our chances of finding the best needle.  Initiating more cell line development projects could allow cell line development to begin earlier in the overall project timeline, such as before final molecule selection.  This can take cell line development off the critical path of CMC development, shortening overall project timelines.

Walk-away processing steps also frees up personnel to perform other tasks.  These other tasks could include process improvements to produce more and better clones.

Preparation of the assay plates is automated, but as we chose to not integrate our analytical devices that portion of the workflow is not automated.  We have an integrated plate imager, so assays using that device can be automated.  Results from both offline and integrated assays can be uploaded into the automated system’s database and appended to clone tracking information.

We can collect plate images through the automated system.  This way we can get cell counts, viability and colony confluence at any of the plate-based stages.

We have not compared cost between automated and manual processes.  It is possible the overall material cost is higher simply due to consumables cost.  Of course there is also the capital cost of the system itself.  Those costs can be ultimately at least partially offset by reduction in FTEs required to run a cell line development campaign and the potential increase in throughput, either in numbers of clones processed or number of cell line development campaigns initiated.

Limited dilution to provide high probability of monoclonality or limited dilution with an image to provide supporting evidence of monoclonality are adequate ways of assurance that the cell lines have a high probability of being clonally derived.  We have the capability to use both limited dilution and integrated, automated imaging to support clonality.