On the 30th and 31st of May, my colleague David Hodl and I will be attending the Phacilitate Automation SIG meetings in London. We will spend the two days in focused workshop sessions which, in the words of the organizing committee, are “designed to deliver a roadmap to cost-effective, robust and interconnected manufacturing and supply chain models.”

In the lead-up to this event, I thought it valuable to draw out some of the dominant themes from a recent Phacilitate automation webinar, What Must be Automated, and What Can’t be Automated in Cell-Based Therapy Bioprocessing. This webinar, chaired by David Hodl, was impressive for the collective experience in cell and bioprocessing that it brought to the panel, specifically:

  • Dr. Vijay Chiruvolu, Vice President of Process Sciences & Engineering at Kite – A Gilead Company
  • Dr. Nicholas Medcalf, Innovation Lead, Advanced Therapies at Innovate UK
  • Dr. Qasim Rafiq, Associate Professor, Cell & Gene Therapy Bioprocessing, Department of Biochemical Engineering at University College London
  • Dr. Valérie Steenwinckel, Director of Industrialisation at Celyad

Of the dominant themes discussed in the webinar, I thought they could be summarized as:

  • Viability – Would you pay $12 for a loaf of bread?
  • When you said automation, what did you really mean?
  • What is the “rock and a hard place” dilemma facing cell therapy manufacturers?

Viability - would you pay $12 for a loaf of bread?

Dr. Qasim Rafiq set the tone early in the webinar, expressing the excitement felt by many of us that this industry has moved from potential to real with the recent regulatory approval of a number of cell therapy products.

“For me, the major challenge now is how do we take these products which have received regulatory approval, and actually turn them into commercially viable products?” said Dr. Rafiq.

And that seems to me to be the primary motivation for automation.

At present, cell therapy manufacture bears little resemblance to an industrial process. It is a strange hybrid of highly sophisticated science and cottage industry.

For example, it currently takes several highly skilled individuals to make a therapy for a single person. But, to meet the demand, to come close to the potential of saving hundreds of thousands of lives, and to be viable, it should be the other way around; “one individual needs to be able to produce therapies for many people at a time.”

The baker who kneads by hand one loaf at a time better be making very fine bread, and charging a lot for the product!

When you said automation, what did you really mean?

This then brings us to the subject of automation.

In response to the question posed by David Hodl, “what do we mean by automation?” Dr. Nicholas Medcalf drew a very useful distinction between automation and mechanization.

“An automated platform is one which effectively you can charge with the materials, press go according to the desired menu, and it will deliver the API or drug product at the end of the process, or at least a significant way down that path. By contrast, mechanization is an assistant to human activity that reduces the variability that would otherwise occur with a human operator.”

Although Dr. Medcalf has picked an exaggerated example, at least this presents for us some opportunities to open the discussion and seek solutions with useful levels of mechanization or automation.

I would add that a mechanized system, let’s use the example of a pump, becomes an automated system when we free the operator from the task of starting, adjusting and stopping the pump. At a simple level, a system is automated when it can turn itself on and off at a suitable time. Most of us will be familiar with a float valve for a pump in a ship’s bilge. There is no computer needed, but it is still automatic.

To this end, Dr. Medcalf also poses the question, “if we mechanize, instead of fully automate, do we really need completely autonomous systems that can manage the production themselves? Or, [could] mechanization, [providing] some assistance to manufacture based around the cheaper engineering solution be the real answer?”

What is the “rock and a hard place” dilemma facing cell therapy manufacturers?

For her contribution, Dr. Valérie Steenwinckle explored the dilemma of investment timing.

“I think we are all pursuing automation because we would like to reduce the cost of the product, and we would like to be more standardized. But it’s kind of a dream, and it’s really difficult to know when you have to make the big investment because we are not talking about small amounts of money.”

The problem of “when” is further compounded by the progress of clinical trials, and the prohibitive cost of bridging studies that arise if automation is first considered post Phase III.

The problem is very real, but almost certainly it is not permanent. For example, the biopharmaceutical industry has matured to the point where development scientists and engineers have a smorgasbord of devices available for R&D, process development and clinical trials, increasing in scale at each step, that can be further increased in scale and automated for commercial production. Think of chromatography columns, filters, fermentation vessels, centrifuges, bioreactors, pumps, valves, instrumentation sensors, and so on.

It seems to me that the highest priority technology development required for cell therapy companies is not fully integrated and automated (process in a box), but rather individual process operation modules that can be conveniently connected in closed sets. Cell separation and cell selection devices, cell washing/medium exchange devices, genetic modification devices (viral vectors, electroporators, cell wall permeabilization), cell expansion systems, formulation systems and so on. Ideally, developers of these systems will rapidly find common ground on interconnectivity of both fluidics and data.

And this brings us to the perennial challenge – who pays?

In the words of Dr. Rafiq, “who bears the cost of development? It’s one thing saying, ‘The vendors should develop these technologies,’ and so on, but (a) they don’t have as much information or knowledge about the process, and (b), do they really have the incentive to be able to develop bespoke platforms and systems, or systems specific for cell and gene therapy?”

At Invetech, we develop process unit modules as well as fully integrated and automated systems and acknowledge that there are fewer challenges in integrating processes that are already mechanized. I, therefore, wonder if we can envision new ways of sponsoring the development and commercialization of cell and gene therapy process modules? This is my hope for the upcoming Automation SIG. If we can get the mechanization in place, the automation will quickly follow.