Less Than 5 Percent Human Effort: Thoughts on the Role of Automation on Cell and Gene Therapy

“Automation” means different things to different people. To me, this word primarily means “tasks completed with minimal human interaction” – less than 5 percent, to put a figure on it. Under automation, more than 95 percent of the tasks at hand are completed by a non-human system, especially tasks spanning over multiple days such as cell culture. An automated device should be able to carry out an assigned protocol repeatedly without requiring human assistance.

There is no need to explain all the benefits that automation can offer to modern industrialized economies, nor is there any need to discuss the basic benefits that automation can offer to cell and gene therapy – numerous iterations of those articles already exist.

However, it is absolutely worthwhile to take stock of some of the most important recent advances in cell and gene therapy automation.

First and foremost, I am excited about my company’s new cell expansion platform, Quantum Flex. But this year, a new release from Invetech also caught my eye – a device for cell washing and concentration. Both innovations received significant attention – but not as much attention as the advances in the science of cell and gene therapies overall. Approvals of CAR T therapies, for example, always receive fanfare far louder than the release of any automated wonder-tool.

Is this a problem? I don’t think so. Technology and its providers are the enablers of science. The scientists and physicians working on the therapies will always lead the way. I see no problem with that aspect of the status quo. Cell and gene therapy has shown high response rates in the clinic, and so as a technology provider I am quite content to be one of the people who paves the road for scientists and physicians to progress their discoveries towards commercialization.

The past is gone
 

In the past, cell and gene therapy companies borrowed technologies from the blood and transplant spaces, and began implementing them in their own field to meet new unmet medical needs. The goal was simply for the cell and gene field to function. Now, technologies are being specifically designed for the cell and gene therapy market. As the field matures, the transition from borrowed to “native” automated technologies is one aspect of the new era.

However, the big barrier to fully entering that era is cost. It’s fairly common knowledge that labor constitutes a major proportion of the cost weighing down cell and gene therapy, and it’s true that automation has been driving down costs by cutting labor out of the equation since the 18th century. The less obvious point I’d like to make is that automation is not the only means of reducing labor costs in cell and gene therapy. If we keep our focus on the field’s non-human entities – on machines and the buildings that house them – then we can look to a one-off investment in an efficient manufacturing system as one way to bring down running costs. The initial construction may not be cheap, but over time it will pay for itself.

If we also look at the humans who turn the wheels of cell and gene therapy, we need to consider their career ambitions and the economic dynamics of the jobs market that they are navigating. The supply of highly skilled and educated scientists and technicians is far below demand, which has produced a highly competitive market. Individual workers have ample opportunities to boost their salaries by jumping ship, and companies are highly incentivized to catch them before they even hit the water, so to speak. Everyone is taking from everyone else, which increases running costs. If better training and workforce development can draw more talent to the field and strengthen the incentives that connect employees and employers, perhaps those costs can be reduced over time.

Humans vs machines: a false dilemma
 

Using the right automated devices can boost the production and scale of the production process too, thus reducing the cost burden without necessarily shrinking or growing the workforce. In the case of CAR T-cell therapy manufacture, the possibilities are particularly exciting. If you want modified CAR T cells to grow, you have to support their special needs: the right cell culture environment and the right cytokines. You can meet these needs in bags, in flasks, and in containers – but I would recommend growing them in a hollow fiber.

Let’s say you need to grow 2 billion cells. If you choose to do this inside a bag, you will need to add a high volume of expensive cytokines and nutrients, because there is no membrane-based separation inside that bag. Hollow fiber bioreactors create a dual chamber system – in simple terms, think of it as a straw made up of a semi-permeable membrane that allows only a specified size of molecules through its pores. In essence, it allows you to keep the cells inside the straw and provide them with what they need. The membrane allows small molecules, such as glucose, lactate, oxygen, and CO2 to pass freely, allowing the control of the cell culture environment. The large molecules – such as the expensive cytokines and media components – can be kept in the cell compartment, intra-capillary (IC), side. One practical upshot of this is that the total volume of complete media you need for CAR T production may decrease, along with the cost. 

Another practical benefit of applying the hollow fiber approach concerns time. The hollow fiber system demands a far lower number of seed cells than does a bag culture. This can save precious time that would otherwise be spent on jumping hurdles during the extraction of sufficient cell samples from the patients.

The future is normal
 

I would like to point out that hollow fiber devices are not new technology. When they were first released around 2011, they offered a very futuristic approach. But now – in a sense – that future is here and normalized. Numerous studies have demonstrated the ability of hollow fiber technology in the expansion of a multitude of cell types. Looking ahead to the next ten years, we can get a sense of what new normalities the next wave of automated cell and gene technology could introduce.

Right now, we are seeing a great deal of work on biosensing tools for monitoring and altering the cell culture environment. While biosensors are currently adopted at many cell manufacturing levels, we are just beginning to understand the applications of machine learning and feedback circuits in the field of cell and gene therapy. At present, we leave cells in incubators, and we monitor their environment, but we don’t try to change that environment. As practices around cell culturing evolve, we should see the emergence of means to more closely analyze the key markers – such as glucose, lactate, oxygen, CO2, and pH. With enough data, machine learning models may be able to train the systems via feedback loops to optimize the cell cultures. Right now, we don’t have enough data – and even in five years we may still be discussing the possibilities rather than enacting them. But I do believe this is an area to keep a keen eye on – an area from which great things are certain to emerge. In that sense, it has a great deal in common with the story of cell and gene therapy so far.