Integrate Upstream/Downstream To Reduce Development Risks And Costs

By Tyler Menichiello, Chief Editor, Bioprocess Online

“The process is the product” is a mantra that describes the modern regulatory understanding of biologics, rooted in federal oversight dating back to the Biologics Control Act of 1902. It implies that any changes made to a manufacturing process can fundamentally alter the end product in a way that may compromise its safety and efficacy.

In the world of biopharmaceuticals, there’s no understating the importance of effective and timely process development. Ineffective process engineering can tank a program’s viability, whether that’s from failing to scale or from moving forward with an expensive and unsustainable process.

Process development is a delicate dance in which upstream and downstream teams must work in tandem to create a manufacturing process that’s as controlled as it is efficient. In practice, however, these teams often operate in isolation from one another, innovating and designing their respective halves of the process in parallel, but not necessarily in coordination.

This isolated development can lead to misalignment between teams, resulting in reactive work or rework that can cause setbacks. And while the separation of upstream and downstream teams is a natural byproduct of the complexity of bioprocessing, their isolation from one another is an avoidable challenge.

So, how can organizations better align upstream and downstream development and avoid this trap of isolated development?

That was the subject of the Bioprocess Online Live event “Optimizing Process Development Through Upstream and Downstream Integration,” a panel discussion featuring Eric Doerr, drug substance technical lead at Sanofi; Mark Fitchmun, president and CEO of Somatek; Doug MacDonald, the director of purification and process development at Immunome; Branden Salinas, Ph.D., VP of process sciences at Umoja Biopharma; and Raj Prabu Vijayakumar Saraswathi, Director of Biologics Process Development at RAPT Therapeutics.

These leaders, with collective experience across key modality types (e.g., mAbs, ADCs, and CGTs), shared their perspectives on how teams can work better together to improve process performance, reduce risk, and minimize costs in development.

Where Do You Draw The Line?

While nearly every bioprocess is divided into upstream and downstream operations, exactly where this dividing line lies is highly variable. As Fitchmun put it, if you ask five people, “you’re going to get seven opinions.” Some organizations draw the line at cell removal, while others treat harvest as part of upstream development.

While this boundary between upstream and downstream isn’t fixed, the panel highlighted several factors that influence it:

• Modality type (e.g., mAbs, AAVs, and ADCs all vary in their manufacturing process)
• Company size (i.e., smaller companies tend to share responsibilities cross-functionally)
• Facility layout or design (e.g., large companies may have physical separation between upstream and downstream teams or have them operating on different floors)

How organizations divide upstream and downstream doesn’t matter nearly as much as making sure that both sides coordinate with one another throughout development.

The Value In Cross-Functional Coordination And Process Locks

Naturally, upstream and downstream teams have different goals in terms of performance metrics and priorities, and these can sometimes come into conflict. According to Fitchmun, it’s not necessarily the misalignment of goals that causes problems for development teams; it’s that they have to do their work at the same time. This puts the onus on leaders to create visibility between teams and set shared goals to keep both teams rowing in the same direction.

“If you don’t have connected goals at the appropriate level, then you can get really disconnected on what you’re trying to achieve,” Salinas told the audience.

Here’s an example of how this conflict might play out: An upstream team is pursuing improvements in titer or cell culture performance without fully considering the downstream impact. While such gains may seem beneficial upstream, they can introduce new purification constraints that force downstream teams into a reactive position. In such cases, it’s challenging for downstream teams to make progress when they’re constantly trying to adapt to upstream changes.

This kind of disconnect between teams can delay development and result in repetitive, avoidable rework. That’s why maintaining coordination between upstream and downstream teams is crucial throughout process development, and it should be intentional from the outset. One way our panelists recommended doing so is by performing temporary process locks to keep teams on the same page.

“Upstream should lock long enough for downstream to understand what’s going on with their own process,” said Fitchmun.

Doerr agreed. “It’s critical that you have those lock-ins, because if you don’t, and you’re continuing to play around on the upstream, you never get a really robust understanding or design space of what’s happening in your downstream,” he said.

Taking An Integrated Approach To Cost Reduction

Coordinating upstream and downstream development through shared goals and process locks is a great strategy to maintain process control, performance, and timelines. It’s also the most practical way to manage COGs, since reducing costs is about optimizing the entire process, not just certain steps.

For one, less time and materials are wasted when upstream and downstream teams coordinate development to avoid setbacks and needless rework. Moreover, decisions made at the earliest stages can impact the entire process and have major cost implications.

Even something like feed media can influence process performance, as our panelists discussed. Changes in media composition, supplier differences, or even switching from powder to liquid formulations can affect cell growth, impurity profiles, and downstream recovery.

As such, upstream and downstream teams should be integrated as early as possible. Saraswathi recommended involving downstream teams as early as clone selection, since clone choice can affect downstream work. (For example, high-performing clones can produce unfavorable glycoforms, show unstable expression, or make purification more difficult.)

Asked And Answered

During every Live event, the audience is encouraged to submit questions for our experts. Unfortunately, not all of these questions get answered on air.

Thankfully, our panelists were kind enough to respond in writing to some of our audience’s unanswered questions. Check them out below and be sure to stay tuned to Bioprocess Online for news and updates on all upcoming Live events!

*The questions and responses below have been edited for clarity.*

How do we move beyond the “knowledge silo” to build a communication thread that gives scientists, engineers, and analysts total visibility into the upstream–downstream continuum?

Doerr: An excellent lever we can use from our friends in the commercial operations world is the establishment of an omnichannel — the coordinated creation, development, and dissemination of process knowledge (i.e., media and documentation) to all stakeholders involved with a process or product. For STEM professionals, what I have seen work best is having a comprehensive, end-to-end process flow knowledge transfer package that thoroughly explains the What, How, and Why in minute detail, supported by regular maintenance and the ongoing publication and dissemination of smaller dossiers or white papers that engage SMEs to continuously improve and define the design space around the bioprocess.

Salinas: When possible, co-locate office and lab space between groups like upstream, downstream, drug product, and analytical. Consider mixed seating, where your neighbors are from other groups. Making rotations and other on-the-job, shadowing-type activities mandatory, particularly where your facilities limit comingling, is another great way to reduce knowledge silos.

MacDonald: I agree with [Salinas]. Sitting in the same areas as your cross-functional counterparts definitely helps improve communication. Additionally, building a high-level CMC "handbook" that details the objectives, responsibilities, and metrics for each function is a helpful tool for the team. Lastly, creating cross-functional metrics enables good discussions while also educating the team.

How successful have strategies like membrane-based purification or phase-based separations been in reducing manufacturing COGs?

Fitchmun: I have only seen phase-based (liquid/liquid) separations used with small molecules or in academic settings and publications. About half of our projects employ membranes or monoliths. However, packed-bed chromatography offers superior resolution compared to membranes, except when working with very dilute feed streams. Packed-bed chromatography also offers greater scaling, fixability, and predictability than either membranes or monoliths. I can’t generalize about process economics because that is highly dependent on the application and manufacturing environment. That said, membrane chromatography is often the best choice if volume reduction of dilute feed streams is the most important thing to accomplish.

MacDonald: Membranes can reduce processing times and buffer consumption in big ways. They also save a lot of time and money during late stage since re-use and lifetime don’t need to be tested, as they are marketed as "single-use". Testing new chromatography resins has been successful, especially for the Protein A capture column, with the implementation of competitors to MabSelect Sure. A couple of these new resins can offer higher binding capacities at a much lower price per liter, which allows for double the savings in manufacturing, since columns can be smaller.

Salinas: Membranes and monoliths are great options when high throughput is needed. This can help reduce COGs for products that are unstable when bound to the chromatography ligand or in the load and elution solution conditions.

Do you have experience using liquid NaOH as an excipient (e.g., in final formulation media), and if so, did you encounter any compliance or handling issues?

Fitchmun: There are two challenges in working with the solid form of NaOH in GMP settings. First, it is deliquescent, which makes it difficult to measure out with precision. Second, the heat of the solution is quite high, so making concentrated solutions can pose problems. So, in most settings, it is a lot easier to work with solutions than pellets. I see both in use, but mostly, I see (and specify) solutions.

Is chromatography commonly used in industrial bioprocessing?

Eric Doerr: Yes, chromatography at the industrial scale is quite common across a host of different bioprocesses. At the full-scale, common chromatography types include ion exchange, affinity, and certain cases of either mixed-mode or hydrophobic interaction types. Size-exclusion chromatography is also possible in particular cases but is generally not preferred due to column lifetime and sensitivity challenges at larger scales.

Doug Macdonald: Yes; various modalities take advantage of product or impurity properties such as charge and hydrophobicity. These include ion-exchange and hydrophobic interaction chromatography (HIC), and mixed modes that take advantage of both are quite common. For certain proteins, such as antibodies, affinity resins are the standard. Chromatography has been the standard for decades, however, membrane filtration has now been successfully used in many instances.

Salinas: Yes, chromatography is used at the largest scales in biotech (e.g. 80,000 L). The most common type is packed-bed. However, membrane and monolith options are becoming more common. There are also older and simpler fluidized-bed types of chromatography still in use.

Saraswathi: In‑house column packing requires highly skilled labor, which is why companies and CDMOs are shifting toward pre‑packed columns. With this shift, the chromatography column diameter becomes a limitation, often requiring multiple cycles to achieve throughput. For very large columns (1 – 2 m), facility footprint and reproducible packing also become significant constraints.