August 1, 2008 (Vol. 28, No. 14)

Angelo DePalma Ph.D. Writer GEN

As Scale Gets Larger, the Need for Process Control Understanding Becomes Even More Critical

Bioprocess scale-up is a time for rejoicing that a product has achieved some degree of success, however modest. It is also a time of great challenges, because in process industries, size definitely matters.

“The scale-up process is a continuum in that the next step is often studied as soon as the current production step is in place,” notes Nicholas Markel, principal consultant at Quintiles. That said, there are differences.

Moving from bench- to pilot scale presents the first real need for stringent process-control understanding, a challenge mitigated somewhat by employing platform or plug-and-play process technology. Platforms provide first-generation processes that meet regulatory expectations and are capable of producing material of quantity and quality adequate for clinical testing.

Ramping up from pilot- to production scale presents an even greater need for process understanding. Here the difficulty arises, not necessarily from unit operations, but from financial pressures associated with batch failure, which may lead to material shortages or regulatory delay.

Markel warns of complacency during this stage. “One can easily be lulled into a false sense of security when, in fact, failure can seriously affect a company’s bottom line.” Here is where decisions can lock companies into processes that are difficult and costly to change without substantial investment in validation, product characterization, or additional clinical trials.

Within this context, platform technologies offer the advantage of minimal up-front investment in process development for risky early-stage preclinical and clinical products. Using these tools, processors need only invest heavily in developing the perfect process for molecules that show promise.

One trend noted by Markel is increased reliance on design of experiment (DOE) and other advanced statistical methods for zeroing in on critical input factors and their interactions. Similar techniques have been used for years in high-volume industries like automobiles, electronics, and consumer goods.

A Game of Incomplete Knowledge

Al Dadson, senior director of fermentation development at Xoma, agrees that adoption of platform technologies can greatly facilitate scale-up. Xoma has bacterial and mammalian expression systems and performs contract manufacturing as well as develops its own products. Its fermentation capacity runs to the 500 liter scale, with cell culture up to 2,750 liters.

For upstream development, Xoma carries out initial fermentations at one liter and seven liters, where it identifies critical process parameters thanks to its platform approach and experience. “Platform methods make antibody scale-up more efficient,” says Dadson, “despite the fact that every antibody is unique.”

As many experts have pointed out, downstream operations are becoming the bottleneck in bioprocessing as expression levels soar. Dadson likens upstream improvements to the electronic industry’s Moore’s Law, which predicts a doubling in capability for computer chips every few years.

While fermentors produce multiples of yesterday’s protein titers due to more robust growth and protein expression, filters, chromatography systems, and various tanks supporting purification cannot expand their capacities. A 500 liter tank can only hold 500 liters, after all, and chromatography resins have finite capacities. The high cost of resins and the physical-mechanical limitations on column size mean that processors must cycle chromatography media rapidly to keep up. “The scale at which you can process downstream becomes limiting; consequently, so much effort is directed there these days.” Dadson adds.

Scale-up for contract manufacturers can be challenging in its own right due to incomplete process knowledge and the uniqueness of every product. Customers do not always present a scale-up-ready process, and small-scale CMOs are often wary of sharing their secrets and customers with a larger contractor. “You don’t always get the full story,” Dadson notes.

Incomplete knowledge notwithstanding, some early-stage processes dropped at the door of a CMO are simply not ready for prime time. Each demands an individual assessment. When a low-expressing system is all that is needed to get the sponsor to the next development stage, sometimes it is best left alone. Xoma can often improve yields simply by expressing the protein in its cell and bacterial lines, but that takes additional time and resources which some customers lack, Dadson says.

While tanks and columns hold finite volumes, scheduling their effective use can free up hidden capacities, both upstream and downstream.

Uncover Hidden Capacity

“Scheduling issues during scale-up are not usually associated with main process directly,” points out Charles Siletti, director of planning and scheduling application at engineering and simulation software firm Intelligen. “They most often arise with support systems like buffer preparation downstream and media prep during cell culture expansion.” Preparing, cleaning, filling, and dispensing tanks are operations processors often take for granted, “until you try to fit them into a plant at large scale.” Bioprocessors have a good feel for these operations at small-scale where everything is done by hand, but this changes at large scale, particularly as resources are shared.

For example, processors assume that if a piece of equipment is utilized 25% of the time at small scale, its potential utilization at large scale should be 60 or 70%. But due to scheduling conflicts 70% may not be feasible, particularly when several processes or operations share one piece of equipment.

Siletti suggests paying attention, not only to major process operations, but to support systems as well, including details as to when every prep and holding tank needs to be turned around, cleaned, and used.

Intelligen’s SchedulePro scheduling software can help smooth out the kinks in equipment utilization scheduling, Siletti reports. The software, which uses an interface language familiar to bioprocessors, checks for equipment conflicts and overbooking, and allows users to create what-if scenarios to optimize processes.

Companies looking for bench-to-feasibility scale-up services up to early-stage clinical manufacturing have a number of choices including larger contract manufacturers.

An increasingly attractive option is a university-based manufacturing organization like Florida Biologix, which is affiliated with the University of Florida. The nonprofit company, with 35 employees, boasts 16,000 square feet of validated, cGMP, multiproduct manufacturing and testing space, and a 5,000 square foot development laboratory next door. Florida Biologix customers include large biotechs with exploratory projects, to virtual and startup companies. According to director of scientific operations Robert Zwerner, Ph.D., “no students work here.”

Many Florida Biologix clients are new to clinical development, and most have non-GMP processes. The company tries to give them a complete package, which they can transfer to a larger contract manufacturer as their molecules progress, Dr. Zwerner reports. That includes batch records, regulatory support, cell banking, cloning, serum-free adaptations, and upstream/downstream unit operation development. The company also performs extensive downstream process optimization and scale-up, particularly for chromatographic separations.

One current customer taking advantage of the Florida Biologix services had developed its own cell line but found after Phase I that those cells would not serve them through commercialization due to low expression yields.

With the primary scale-up focus on equipment and machinery, media companies will point out the importance of ingredients in the scale-up process.

Media Scalability Overlooked

Choice of raw materials, sourcing, supply chain, and security all play a role in how easily scale-up occurs, notes Bruce Lehr, director of marketing at SAFC Biosciences. “If you’re planning to manufacture something at the 10,000 liter scale, it is critical that the media you design for it is scalable and can be formulated consistently.”

As with unit operations, media also lend themselves well to the platform approach, whereby one base formulation will work fairly well across many cell lines. Optimization is then reduced to an exercise in tweaking the base media and devising feed and supplementation strategies. One of SAFC’s specialties is designing these supplements and formulating them as liquid concentrates or in powder form, a strategy popular for large processes.

Increasingly, says Lehr, companies are looking to eliminate hydrolysates and move toward chemically defined media. “It’s largely a consistency issue but cost may be a factor here as well.” While chemically defined media are more expensive than hydrolysates, he believes that chemically defined products may bring about upstream or downstream efficiencies that are less easy to quantify.

Processors frequently need to tweak media formulations during scale-up, but not all have the time, equipment, and expertise to do so on-site. SAFC offers a service, imMEDIAte Advantage™ (IA), which prepares new formulas, up to 200 liters’ worth, in less than 10 days.

“IA allows customers to test these variations in smaller bioreactors to assist in finding the final formula,” says Lehr. IA delivers media, feeds, and concentrates in both liquid and powder formats. Successful formulae are easily transferred to SAFC’s large-scale manufacturing, which uses the same raw materials as IA, he adds.

Tips for Successful Process Development and Scale-up

‘ Define your process.

‘ Obtain independent, objective reviews.

‘ Try to adopt platform technologies.

‘ Use statistical methods and DOE.

‘ Invest in process development

commensurate with life cycle.

‘ Use disposables where possible.

‘ Use standard, off-the-shelf equipment whenever possible.

‘ Develop and cultivate relationships

with vendors.

‘ Qualify your analytical methods

early on.

‘ Have a plan and know when to stop.

Nicholas Markel of Quintiles

Previous articleBristol-Myers Squibb Bids $4.5B for ImClone
Next articleCary Institute of Ecosystem Studies Is Recipient of $750K EPA Award