Ask the Expert: Automated, Closed-Loop, Inline Monitoring of CAR T Cells in a Production Process

Jan Van Hauwermeiren (vice president of sales and marketing at Ovizio Imaging Systems) joined BPI for an Ask the Expert webinar on 29 August 2019. He demonstrated how his company’s iLine F microscopy system combines sensitive optical capability with machine learning to monitor chimeric antigen receptor (CAR) T-cell expansion, transduction, and activation. The device captures holographic images of cells as they pass through a single-use conduit running between a microscope and bioreactor. Cells then funnel back into culture without additional handing. Because the device records 1,200 images per day, it precisely tracks individual cells over time. The system is fully automated and enclosed to ensure regulation-compliant analyses that do not inhibit cellular development.

Van Hauwermeiren’s Presentation
Each image captured by the iLine F system is a composite. When CAR T cells travel from a bioreactor through a gamma-irradiated BioConnect single-use channel, they cross a sterile probe that records data regarding the light the cells absorb and deflect. Cells then travel back into the bioreactor with no loss during transfer. Data collected help to construct an intensity image that identifies basic cellular features and a phase-contrast image that visualizes sampled cells’ refractive indices.

Combining those images yields a rich, three-dimensional account of each cell’s shape, size, circularity, volume, and membrane integrity. A refocusing tool ensures data collection for cells that travel below or above the probe’s focus plate. With the device gathering new data points every 30 minutes, researchers can track cell viability, density, and more in real time.

Compliant with CFR 21 Part 11, the OsOne machine learning platform enables the system to make such phenotypic distinctions. It comes with reference data for 70 validated optical properties. Those data “train” iLine F algorithms to distinguish between, e.g., cells that are viable or dying, transduced or nontransduced, and infected or noninfected.

Therefore, the system can mark and follow specific cells reliably without staining or labeling. It also can account for cellular subpopulations, contaminants, and debris. Several Ovizio customers use antibody-coated magnetic beads to activate CAR T cells, for instance. The iLine F system can differentiate activated from nonactivated cells on those beads to determine whether a culture is truly growing or whether debris and contaminants are skewing viability counts. And the iLine F’s rich data set enables researchers to track trends across different donors, bioprocesses, products, and allied companies.

A key feature of the system is its integrability. Validated software communicates seamlessly with different platforms and servers. And because of the BioConnect mechanism, the iLine F microscope can attach to several types of bioreactors without compromising process integrity or sample size. The result is strong, data-driven, regulatory-compliant analytical capability.

Questions and Answers
Does 3D holographic analysis help to determine the cell-cyle phase? We have some experience identifying cell-cycle phases. Our algorithms are incomplete, but that kind of analysis is entirely possible. Please reach out to us for more detailed information.
Because dyes are not used, what imaging factors determine viability? We analyze cellular responses to light. A round, viable cell creates a reflective peak when hit by light, whereas a dead or dying cell losing circularity and membrane integrity will deflect light. Algorithms can determine the difference between those reflection peaks and deflections.

Can the system detect batch contaminations? The magnification level cannot detect bacterial or microbial contamination, but scientists can identify such problems when cells die more rapidly than anticipated.

Can the technology detect a virus-infected cell? Yes. With baculovirus expression vectors and insect cells, for instance, we can see a clear difference between infected and noninfected cells. The same will go for lentivirus and adeno-associated viruses (AAVs), although algorithms are still in development for those cases.

Scale-out is a key challenge in cell therapy, but the iLine F system can monitor only one bioreactor at a time. Are there plans to facilitate higher scales? There are no plans to multiplex. We have focused instead on reducing the device’s footprint. Multiplexing would create challenges because the device uses sensitive optical setups. When multiplexing, you need to move things around, which generates high costs. So a multiplex device would be more expensive than multiple one-on-one devices.

More Online 
The full presentation of this webcast can be found on the BioProcess International website at the link below.

Watch the full webcast now.