Cytocentric Visionaries: Alan Blanchard
Cell Culture Automation for Reproducibility: People Never Do the Same Thing Twice
Alan Blanchard is the Chief Scientific Officer and Co-founder, along with Thomas Forest Farb, of Thrive Bioscience. They are working to provide a fully automated system for cell culture. Thrive systems use programmable image capture, analysis, and fluid handling, all under incubator conditions.
Here Alicia Henn, CSO of BioSpherix, interviews Dr. Blanchard about how automation technology might help scientific reproducibility. The transcript was edited for length and clarity.
We see you as a Cytocentric Visionary because you are working to bring automated cell handling into a safer and more physiologically relevant environment for cells. So why automate? What’s wrong with how scientists have always done things? It’s worked so far.
Cell culture processes have been stagnant for sixty years. It’s still done manually, even the mundane task of changing the media. In this day and age, doing anything manually needs to be updated, especially with the increasing importance of cell culture.
Look at the cell culturist’s problem. You can buy a microscope that fits in an incubator or you can buy a fluid handling system, but you still need a person to get in there to do manipulations. Unless you automate the entire process from beginning to end, you haven’t really helped the cell culturists. So we see a real need to automate.
What can automation really do to help cell culturists?
There are four advantages of automation over manual labor; reproducibility, sterility, scalability, and documentation.
One of the big problems these days in biological experiments is the lack of reproducibility. People don’t do cell culture the same way twice. They get tired, it’s not particularly interesting work, and not everyone has the aptitude for it. Automation gives you a degree of reproducibility you would never get from manual procedures.
One of the drawbacks of manual techniques is that you have this nice incubator, but the minute you want to do anything with the cells, you take them out. This changes the temperature and the gases inside the incubator, not only for the cells that you’re taking out but for the cells that are still inside. It is more reproducible if the cells never leave the incubator. Everything you need to do for cell passaging is inside the controlled environment.
The second advantage of automation is sterility. People are frankly quite dirty. They’re always shedding microbes, so any manual procedure is a potential contamination event. Stem cell experiments can last for 90 to 100 days. If the experiment gets contaminated on Day 85, you’ve lost three months of work.
Then there’s scalability. With the field of the stem cell culture growing exponentially, people need to do a lot more experiments. Scaling up these experiments is very difficult when you need trained technicians for manual steps. It’s not only expensive in terms of paying people, but finding and training them is expensive, too. You still have the inevitable failures that result from manual procedures. It’s a daunting obstacle when you’re trying to scale up by a factor of 10.
The fourth advantage is documentation. If a technician gets a phone call and cells sit in a reagent for five more minutes, it is never written down. Full automation can include documentation for a complete record of what went on; visual, environmental, pH, any sort of measurement. So, if something goes wrong, you can look back to see what happened.
But you can’t really replace the decision making of a person, can you?
The first decision cell culturists have to make is whether or not the cells are ready to passage. We did a little experiment at a trade show where we showed people images of cells and asked them to estimate the confluence. People were all over the map. We would show the same person the same picture rotated by ninety degrees and they’d give us completely different estimates.
So there is one source of irreproducibility right there. No two people have the same idea what 80% confluence means. Even the same person doesn’t have a consistent idea.
The same is true in the process of trypsinization. During this step, you’re treading a fine line between trypsinizing enough to release the cells but not so much that you damage the cell walls. A person taps on the side of the plate and looks at whether rafts of cells are forming or not. This is a very subjective measure which is not reproducible.
These decisions can be automated by incorporating phase contrast microscopy into the incubator. Advanced image analysis computes an objective measure of confluence in order to decide whether or not it’s time to passage. The software can look at the morphology of the cells and give you an objective measure of the degree of trypsinization of the cells. Cell counts can be automated, too.
What do you say to all the cell culture “moms” and “dads” out there that are uncomfortable with giving up decision-making to an algorithm?
We don’t require that you actually give up the decision making. Our instrument is networked and the operator has the capability at any time to look at the images over the internet. You can be at home sitting on the couch with your iPad and make the decisions yourself. The instrument can send you an email saying “I think the cells are at 80% confluence. Here’s a picture. Is it time to passage?” We don’t want to take away any control from the operator; we want to offer the operator a tool in order to maximize their productivity.
I’ve heard a lot of complaining from stem cell researchers about changing medium on the weekends. Would automation solve that problem?
In the time it takes to induce pluripotency, expand the cells, and differentiate them into the cell type of interest, you’re at 90 days. Every one of those days you have to go in and change the media. Of course, doing this manually means you’re taking them out of the incubator and they’re cooling down, which is stressful for cells. There’s a chance of contamination, and it’s not interesting work, so it’s hard to get people who will do this meticulously for months at a time. Machines do it the same way every time, under constant conditions, and at any time of day.
By measuring the color of the pH indicator phenol red with an electronic spectrometer, we can accurately record the pH and even use it to gauge when to change the media. Cell culturists that are trying to mimic the physiology of the body don’t measure the pH at all. By the time it’s yellow, it is way outside the physiological limits for pH.
In physioxic conditions, cells using glycolytic metabolism generate a lot more lactic acid and those cultures are going to get yellow much faster.
Yes, we could measure that increase in acid, and if the media has to be changed every 16 hours instead of every 24 hours, the instrument could pick up on that and do it. It would not mind that the sixteen hours falls at 2:30am on a Sunday.
Could you see automation used for in vitro toxicology applications or drug testing where there's a high demand right now for better reproducibility?
Yes, with these types of tests, you want to be studying the effects of your chemical or drug and not inadvertently studying the effects of whether your technician got enough sleep last night. Automation gives you the scalability to do a lot more tests more cheaply and more reproducibly and the documentation aspect is always of importance to the regulatory authorities. That way you aren’t relying on some technician having done something correctly with no documentary evidence that it actually happened.
And also distributed manufacturing of a cell product?
If you're doing autologous therapies, you really don’t want to ship a patient’s cells from one place to another because transporting cells has its own issues with environmental controls. You really want to have the whole production facility there in the clinic. Automating it is the most precise and reproducible way of making sure everything is done properly. The regulatory agencies clearly are going to want documentation of every step of the way as well. Our machine would supply that automatically with environmental and protocol data.
In Part Two, we continue our discussion with Dr. Blanchard, talking about mimicking the human cell culturist with robotics and using automation in the basic research laboratory.
About the Author
Alicia D Henn, PhD, MBA
Alicia Henn has been the Chief Scientific Officer of BioSpherix, Ltd for two years. Previously, she was a researcher at the Center for Biodefense Immune Modeling in Rochester, NY. Alicia obtained her PhD in molecular pharmacology and cancer therapeutics from Roswell Park Cancer Institute in Buffalo, NY and her MBA from the Simon School at University of Rochester in Rochester, NY.