4.12.22-Health-Science-Startups-Cellino

MARINNA MADRID: Hi, thank you so much for this opportunity. My name is Marinna Madrid. I'm a co-founder and Chief Product Officer at CELLINO. And we are making personalized cell therapies viable at scale, really for the first time. And we are supported by a fabulous group of investors, including The Engine, which is an MIT-based venture fund.

This is a video of a highly skilled stem cell scientist. And this stem cell scientist is physically removing undesirable cells from an induced pluripotent stem cell culture. This is a very common part of processing iPSCs. iPSCs and their derivatives have a ton of potential to treat a wide range of chronic degenerative diseases.

But, as you can see, the processes for culturing these cells are extremely manual, extremely labor-intensive, and require a high level of operator skill. And so what that means is they are highly prone to variability, and they're also very difficult to scale. Running manual processes like this might be OK for phase 1/2a, where you're treating, let's say, 6 to 20 patients, but for phase 3 and beyond, you really need a more robust, scalable approach.

And so what we're doing is we are automating these cell therapy processes-- both reprogramming and differentiation. And so what we do is we image cells in brightfield throughout these processes, and we use image-guided machine learning algorithms that we've trained to characterize individual cells. And then instead of that physical scraping away process, we have a laser processing technology that we can use to, essentially, zap away any cells that are unwanted or displaying an undesirable phenotype. And what's so unique about this optical approach for characterization and cell manipulation is those optical processes are A, highly amenable to automation, but B, compatible with closed manufacturing.

So this image on the right is a rendering of our clinical-grade system, which we're currently designing and have physical mock-ups of. And for that system we have closed cassettes, and what that means is that multiple samples of cell therapies-- for example, multiple samples of different patient samples for an autologous therapy-- can be processed in the same facility at the same time. And so that brings down the cost of manufacturing significantly.

In addition to designing the clinical-grade system, we have our automated process development system up and running right now, just down the street. If anyone is interested in seeing it, we're happy to give you a tour.

It's well plate based. So we have a robotic arm that moves well plates from the incubator, to the imager, to the laser processing system, to the liquid handling system. And we use this to really take the large volumes of experimental data that are necessary to train our image-guided machine learning algorithms.

The platform is quite application agnostic. Here, right now, we're applying it to iPSC generation-- going from patient blood cells to induced pluripotent stem cells. And I won't go into this in heavy detail. But at a high level, what we're doing is we're replacing the visual decision-making parts of the process with automated, image-guided machine learning characterization, and replacing that physical scraping away of unwanted cells with laser-based removal of unwanted cells.

But I mentioned the platform is quite biology agnostic. We've also applied it to differentiation processes. These are the few of the cell types that we've differentiated from iPSCs-- skeletal muscle; dopaminergic neurons, which we're actively working on and could be used to treat Parkinson's; and retinal pigment epithelial cells, which can be used as a treatment for age-related macular degeneration.

So for example, in the case of iPSC derived retinal pigment epithelial cells, the job of these cells in vivo is to phagocytose pieces of photoreceptors. So we trained image-guided machine learning algorithms to, based on brightfield images, predict which RPE cells would be stronger at phagocytosing and worse at phagocytosing, and then laser-killed the cells that were predicted to be worse at phagocytosing, with the goal of forming a more highly functional sheet of cells, on average.

And in terms of partnerships, we are heavily focused on the regenerative medicine industry in general, and particularly excited about induced pluripotent stem cell-derived cell therapies. So if you are in these industries, please don't hesitate to reach out. As I mentioned, the platform is quite biology agnostic. Right now we're actively working on automating reprogramming processes, we're exploring gene editing applications, and we've also worked and will continue to work on differentiation processes. So if your work falls into any of those categories, or even outside, please don't hesitate to reach out because we'd be happy to discuss.

Thank you so much.