Cell Therapy Manufacturing: Quality By Design

Before they can deliver cell therapy products to large numbers of patients, engineers must find a way to manufacture them safely, reliably, and cost-effectively. Now, a team of bioengineers have a manufacturing paradigm they think can improve that process.

by Kayt Sukel
November 06, 2017

Over the past decade, researchers have made astonishing advances in regenerative medicine. But moving from discovery to the large-scale manufacturing of complex, living cell products typically grown in petri dishes or desktop bioreactors is far more difficult than mass-producing conventional pharmaceutical products.

Before they can deliver cell therapy products to large numbers of patients, engineers must find a way to manufacture them safely, reliably, and cost-effectively. Fortunately, a team of bioengineers at the University of Toronto and the University of British Columbia believe they already have a manufacturing paradigm they can build upon: the 2006 International Conference on Harmonization's Quality by Design (QbD) framework.

As a design philosophy, QbD has already been widely adopted by both traditional manufacturers and small molecule biopharmaceutical producers. Conceptually, QbD is simple: design consistency and reliability into each step of the manufacturing process.

For biological products, this involves generating target profiles of product identity, purity, and potency for each stage of the process, so engineers can monitor and validate product attributes as cells and other materials move down the production line. One misstep could produce unexpected results—a cell that cannot survive once transplanted into a human patient or a product with a genetic or proteomic composition altered enough by processing to render it useless as a treatment.

Despite the added complexities of growing and modifying living cells, Peter Zandstra, director of the University of British Columbia's School of Biomedical Engineering, believes that QbD methodology can provide a guiding framework as bioengineering companies scale-up and scale-out production. They will just have to make a few important tweaks along the way.

"In traditional biopharmaceutical manufacturing, the cell is the source of the product," Zandstra said. "With cell therapy products, the product is the cell itself. That means there are many different unit operations, both in terms of upstream and downstream processing, that we have to consider as we develop the technologies we need for full manufacturing implementation."

That starts with the source and destination of the cells. Autologous cells from and for a single patient would generally involve low-volume production. Allogeneic cells, harvested from human donors or cadavers, involve large batches of live cells mass-produced for a large population of patients.

From there, engineers must determine efficient protocols to isolate cells from specific patients or to generate cells from reprogrammed allogenic stem cells. Given the variability of those inputs, manufacturing multiple batches of stable, effective cell therapy products is not a sure thing.

Then engineers must develop automated, systematic ways to isolate the cell products themselves, as well as ways to preserve and transport cells so they reach the patients they are meant to treat.

Given the complexities of the process, QbD methodology can guide the design and construction of manufacturing process lines while reducing the potential for accidental contamination or errors, Zandstra said.

QbD is all about measuring against a set of profiles. While engineers can adapt some current tools to cells, they must also create new technologies, said Yonatan Lipsitz, a bioprocess engineer trainedat University of Toronto's stem cell bioengineering lab. This is because current methods cannot provide all the data needed to ensure that cells, notorious for responding to their environment by changing in unexpected ways, were not adversely affected by the altered conditions of each stage of the production process.

"We need to adapt the bioreactor technologies currently used in other fields," Lipsitz said. "People are trying to figure out how you can get cells that are quite sensitive to their environment to grow in a large bioreactor format. That's the scale-up challenge.

"The other piece is scaling out, or figuring out how to efficiently run a large number of small bioreactors and get the same outcomes from all of them. You'll also need to see new technologies in the separation of cells, so you can harvest the products in a way that keeps the cells alive, healthy, and doing what you need them to do," he said.

Then there is the need for new analytical technologies. Classic biopharmaceutical metrics can give you read-outs on different parameters of a cell during a process, but they cannot provide the "whole story," Lipsitz said.

"We are going to need different sensors or other analytical technologies that can give us more information about what cells are actually doing and experiencing during the production process," he explained. "That way we can make sure the cells are meeting the critical quality attributes."

Using the QbD framework can help biomedical engineers, both on the bench and the manufacturing sides of the house, to create viable process lines for cell therapy products, Zandstra said. But he cautioned that manufacturing needs and capabilities must be addressed early on in the development pipeline, even before late stage clinical trials, to help better define the key quality metrics that will govern mass production.

That starts when engineers truly understand the biology of the systems they are trying to mass-produce, Lipsitz added.

"This field is still in its infancy," he said. "People are finally solving a lot of the fundamental challenges inherent to cell manufacturing, and it's really exciting.

But a good process design needs to be augmented by biological understanding, biological modeling, and insight from biological data.

“You need to understand that this is a complex product, highly interactive with its environment, and not always so well understood,” he said. “If you don't, you won't be able to really look at the relationship between your critical quality attributes and what may be affecting them during manufacturing to find the right path forward."

Kayt Sukel is an independent technical writer based in Houston, TX.