Manufacturing Strategies: BPI Theater @ Interphex 2015

Bill Hartzel (director of strategic execution for advanced delivery technologies, Catalent Pharma) 1:30–1:55 pm

Reducing the Risk Associated with the Filling of Biologics with Advanced Aseptic Processing 

Hartzel discussed leveraging “blow–fill–seal” (BFS) aseptic processing technology to reduce risks associated with biologics fill and finish. A number of drug recalls have been associated with microbial and particulate contamination of glass vials, problems that could be solved using BFS. Automation eliminates human intervention at this critical stage, driving out associated risks. Hartzel said that BFS plastic (polypropylene) containers have a more inert surface than glass.

In BFS, resin pellets run through an extruder, which forms the containers. They are filled with drug product, then automatically stoppered and sealed. The full, closed process takes 15 seconds in a class A environment. Hartzel’s company fills 8–50 units in each 15-second cycle depending on container design. Containers are open for only a short time, further reducing the risk of particulate contamination. The container–closure is created as part of the process.

Hartzel explained how quality by design (QbD) principles are built into this aseptic filling process. His company has microbially challenged the equipment and machinery to understand its critical control parameters. Replacing glass with plastic provides shatter resistance, lowers shipping weight, and offers design flexibility. However, extractables and leachables are different, and container stability must be evaluated. (Glass is impervious; plastic is not.) In evaluating container–closures, companies evaluate safety and efficacy of the product inside them over time. Biologics pose special challenges regarding thermal stability (heat is involved in forming BFS containers) and gas permeation (e.g., moisture loss, oxygen ingress or egress).

The resin is heated to 180° C, but that temperature dissipates quickly (in a second or two). Studies by the equipment manufacturers have shown that the peak temperature for drug product contact is about 40° C. Gas permeation issues are drug-product specific and dependent on temperature. Finally, Catalent performs extraction testing to provide a baseline benchmark for further leachables studies.

Case Study: Hartzel’s company compared the compatibility of a model MAb formulation with that of a glass container system and its own Advasept plastic containers in a BFS system. The team investigated the effects of vial-formation heat on the MAb itself as well as testing the formulation’s performance over time (nine months to two years) stored in plastic and in glass. The performance of the molecule did not differ between the container types, and all results were comparable for both. Oxidation was higher in glass, however, than in plastic. Leachables were different — but all remained under set threshold levels. Extractables were low as well.

An audience member asked whether Catalent has filled a commercial product using the BFS system. Hartzel said his company has been producing commercial products using this technology for over 30 years (over 20 years for large biomolecules). Another attendee asked whether clients run stability tests (for regulatory purposes) on products that were exposed to the 40° C temperature. Hartzel replied in the affirmative. Catalent can produce small feasibility batches with ≥1 L of material for testing heat effects and stability. Someone asked whether the resin supply is proven pyrogen free. Hartzel said that among the many evaluations his company has performed on incoming material are endotoxin and microbial tests. Specifications for incoming raw materials are designed to allow the company to operate aseptically.

Another audience member asked about prefilled syringes and autoinjection types of container designs because container-design flexibility is considered an advantage with BFS technology. So far Catalent has focused on basic containers, but Hartzel said that customers are showing interest in more elaborate designs. Prefilled syringes, he said “can be done.” Some manufacturers have tried. For many products, delivery devices are integrated into drug development. Hartzel said that even more complexity than prefilled syringes will be possible in the future with BFS. “We can make things that are flat or have multiple ports, we can do things to change that dynamic, and that’s where I think the technology will ultimately go.”

Another questioner brought up the low end of the temperature scale. Hartzel said that polyolefins in general operate fairly well in cold conditions. His company has data showing successful freezing to (and thawing from) –20 °C and is currently looking at –80 °C. Someone pointed out that stopper manufacturers don’t recommend temperatures below –40 or –50 °C. But Hartzel pointed out that with BFS, products are completely enclosed and encapsulated, making stopper integrity less critical.

A final question concerned the maximum ~40 °C exposure. Could the company prove that products would never face warmer temperatures? Hartzel said that the dynamic nature of the process makes true validation difficult. Simply inserting a temperature sensor into a moving system is a logistical challenge. And different container designs can create different short-term conditions (e.g., the ratio of surface to volume). “We need to develop a better method for getting in there and testing it to be able to say that it is validated,” he said.

View the full presentation videos

Frank Nygaard (senior technology partner in global business development, NNE Pharmaplan) 1:00–1:25 pm

Where Does Biotech Flexibility Matter? Case Study of 12 Flexible Facilities

Nygaard described some things his company (a design– engineering firm) has learned through its involvement in 12 different flexible-facility projects. He focused on facility design and manufacturing strategy, single-use technology, production capacity, and technology integration. Cost and flexibility are high on the list of priorities for new facility projects. Companies need to respond with agility to changes in demand while managing risk to the lowest possible levels.

Disposables are a key enabler for flexible manufacturing. Traditional stainless steel systems are fixed in place; single- use systems can be physically moved around a facility, effectively decoupling a bioprocess from the building in which it occurs. This allows companies to start small and then grow big. It also helps them reduce investment costs, although operational costs can increase through consumables expenditures.

In determining whether to implement single-use technology, a company must consider the capacity it needs. Nygaard showed that for an example monoclonal antibody (MAb) produced at 3 g/L in the 100- to 300-kg range, disposables offer worthwhile benefits. For larger production volumes, however, stainless steel might be the right solution. He also brought up the question of manufacturing capacity and related uncertainties, which in a start-up company also favor single-use technologies.

NNE has developed a flexible facility concept called “Bio on Demand,” which groups different functions together (e.g., raw materials, manufacturing, and supports). A traditional plant includes a seed-culture laboratory, drug-substance production cell culture and harvest areas, purification suites, and a final drug-product filling area. The flexible facility concept is the basis of the 12 multiproduct facility designs that Nygaard presented. Not only did mammalian fed-batch cell cultivation involve single-use bioreactors, but so did other processes and systems. His 2011–2015 case study projects are located in Brazil, central Europe, Denmark, Russia, and Asia. Nearly half were optimized for MAb manufacturing. Bioreactor batch sizes were pilot scale (500–1,000 L), launch scale (1,000–2,000 L), and commercial scale (2,000 L). Two of six commercial designs allowed pooling from more than one single-use bioreactor, with up to three reactors inoculated with the same seed material and their outputs processed together as one 6,000-L batch.

Manufacturing strategies were mostly campaign based, but two projects allowed for concurrent manufacturing. NNE has seen a geographical difference in those approaches, with emerging markets more focused on campaign strategies. Even with closed systems, their regulatory authorities generally do not accept the strategy of making different products in the same cell culture suite at the same time. Europe and the United States are more accepting of concurrent manufacturing.

NNE clients wanted facilities designed for optimal capacity use, and campaign manufacturing requires down time for product change-over. Such changes are frequent in pilot plants. Most designs have the same number of upstream and downstream suites. But concurrent production does not translate to concurrent purification, so facilities based on that strategy typically require more than one downstream process area.

All 12 case studies were “stick-built” facilities rather than modular constructions. NNE evaluated both options in two cases and found that, for their locations, the traditional model worked better. Nygaard mentioned the “ballroom” facility concept — with closed-system cell culture and initial purification (including viral safety) running in the same room — but showed only one example of it. That facility was meant for local manufacturing only and would not be making products for Europe or North America. Meanwhile, cleanroom classifications differ according to the location of the plant and its intended market.

Use of disposable bioreactors was a major criterion for inclusion in this study. Nygaard reported that a third of the plants had a hybrid approach to harvest clarification; the rest were fully single use. Cost and scale limitations applied, especially for depth filters and 2,000-L bioreactors (some of which were stainless steel). In media preparation, larger scales also favor the use of stainless steel, so 25% of the facilities had hybrid solutions and 75% were exclusively single use. Most use disposables for storing prepared media; most used them for intermediate product holding downstream, too. Buffer preparation was 75% single-use and 25% hybrid, with the same scaling effects as with media preparation upstream. Two of the 12 cases implemented chromatography and ultrafiltration/diafiltration (UF/DF) skids as well as other purification equipment.

When designing for 100–400% capacity expansion, the clients took a number of approaches: 25% equipment swapping (starting with one size and later replacing it equipment with a larger size), 58% fitting out (leaving enough space in the beginning to add more equipment later on), and 17% drop blocks (self-contained units, somewhat modular). All cases were based on fed-batch production, although NNE reports that some clients are interested in continuous processing as well as antibody–drug conjugates (ADCs). Fill and finish suites were also included in many designs.

View the full presentation video

Martin Hofmann (managing director, Biotechflow) 2:00–2:25 pm

Case Study: Overcoming the Design Challenges of Expanded-Bed Adsorption Columns

Hofmann’s company designs chromatography columns and skids, primarily for production scale. He described how it addressed one customer’s expanded-bed adsorption (EBA) chromatography column project. The client’s downstream process begins with whole cells from a 5,000-L or 10,000-L fermentor. Those cells secrete monoclonal antibody (MAb) into their supernatant, and the whole cell broth (with intact, viable cells) is pumped onto the expanded bed of Rhobust tungsten-carbide agarose protein A media from Upfront Chromatography. The tungsten carbide makes these media three times heavier than water so it will expand with 450 cm/h flow.

The customer requested that the maximum overall height of the column be 2 m, but a 1-m piston adjustment was required at pilot scale. Biotechflow had to design a head outlet that prevented vortex creation (which otherwise would cause media to leave the column), and the column could have no dead-space piston seals. Hofmann’s company was engaged for its ultrasound experience because the customer wanted ultrasound sensors (which can be adjustable to monitor flow and expanded-bed position). And the customer needed results in six months.

Hofmann detailed his company’s approach to this column-design project. A standard design would not work. The column had to be made of acrylic rather than steel, but the acrylic could have no affinity for the chromatography beads. So first, the company researched and found the right material of construction (resistant to 40% alcohol), but soon discovered that it would need a specialized autoclave for sterilizing the large tube. The team chose inflatable seals, which had to withstand 6-bar pressure. The large working scale presented its own challenges (e.g., requiring a 2-m lathe for shaping the column).

Next, Biotechflow used a piston to find out what bed height was needed. Vortices caused by a stirrer caused resin beads to leave the column and get stuck in the pipework of the GE chromatography skid. So the challenge was to prevent vortex formation while allowing air to leave freely and the piston to move 1 m at 2.5 mm/sec, with only a system of relief valves as a failsafe. So the company designed a complex, angled piston with two C-shaped, zero–dead-space seals. The purpose of those seals is both to seal and to ensure that the piston is square to the tube and will hold in place without slipping. And the customer needed to be able to clean behind them. After experimenting with storm-drain type outlet designs, the engineers developed a variation that solved the vortex problem. By pushing into the swirling vortices, the valve completely stopped them at 450 cm/h flow rate. Up-flow in this column requires no mesh and no downward piston motion; expanded-mode elution is in a smaller volume by down-flow.

Ultrasound: Six ultrasound transceivers on each column monitor the position of the piston, the height of the clear supernatant, and how high the expanded bed reaches. Material harvested from the fermentor begins with high viscosity, and as the pressure drop increases, it disturbs the expanded bed. Flow rate drops smoothly and then is increased to keep the bed expanded — a process that depends on information from the ultrasound probes in analogue mode. Hofmann mentioned his 2003 paper on monitoring concentrations of MAbs and cells using ultrasound (1). He described it as a linear relationship between ultrasound attenuation and MAb concentration. He called this a “universal detector” that can be used to monitor cells, eluates, buffer, and proteins. Key to its function are zirconia crystals, which have to marry perfectly to the column tube wall, and Hofmann described how the engineers made sure of that. He also highlighted the means by which they ensure that the column itself remains level, even in the case of earthquakes, which is important to separation efficiency.

Biotechflow delivered this column in about six months, and the customer reports antibody titers coming off it at 1.3 g/L (22 g/L media). EBA is providing an 82% yield (improved from 70% with clarification followed by a packed protein A column) in 6.8 hours of processing (rather than 18.5 hours with the traditional column). The supplier is now going into full production with this column design. In its own studies, Biotechflow sees a yield of 1.34 g/L based on a smaller volume, with good removal of DNA and host-cell protein. Currently, these columns are installed at current good manufacturing practice (CGMP) compliant facilities in Canada, Europe, Australia, and in the United States, all producing phase 3 clinical trial materials.

Reference
1 Hofmann M. Use of Ultrasound to Monitor the Packing of Large-Scale Columns, the Monitoring of Media Compression, and the Passage of Molecules, Such as Monoclonal Antibodies, Through the Column Bed During Chromatography. J. Chromatogr. A 989, 2003: 79–94; www.biotechflow.com/resources/jchrom_paper1.pdf.

View the full presentation video

Lorenz Fischer (R&D, PreSens Precision Sensing GmbH) 2:30–2:55 pm

Your Shake Flask Is Not a Black Box Anymore! Online Measurements of Metabolic Activity to Accelerate Your Bioprocess Development

Fischer described a device for measuring online oxygen, pH, and biomass in shake-flask cultures to monitor metabolic productivity. After a brief description of his company, he introduced its optical sensor portfolio for measuring dissolved oxygen (DO), pH, carbon dioxide, and biomass in cell cultures through noninvasive technology. Only a sensor probe (a small patch of fluorescent membrane on the inside wall of the vessel) comes into contact with the culture. Because a light needs to shine through the vessel wall, the technology does not work in black or stainless steel vessels but does for all disposables as well as glass — even amber-colored or semi-opaque. Measured solutions can be colored or have their own background fluorescence because the probes are optically insulated against such influences. Optical measurement eliminates the need to interrupt processes for drawing samples and provides real-time monitoring of culture parameters.

Shake flasks are familiar in biopharmaceutical development laboratories and production seed-train work. This established technology provides less shear than stirred systems and compares favorably with microplates, tubes, and cultivation bags. But without sensors, their culture conditions can seem like a “black box,” especially compared with bioreactors that have integrated electrodes and electrochemical probes. PreSens SFR vario sensors measure pH and DO into shake-flask cultures, providing valuable information with no need to draw samples or modify existing equipment.

The device fits in all commercially available shakers and comes with software that can control up to four devices at a time. Flasks with sensor spots (one for pH, one for DO) are aligned atop an optical device that shines a pulse of excitation light, to which the sensor spots react by emitting fluorescent light depending on analyte concentration. A change in pH or O₂ concentration changes either fluorescence intensity or period, which the detector measures, taking a reading in about 200 ms, allowing one reading per second.

Biomass is a light-scattering measurement at 560 nm. Fluid height varies as flasks are shaken, so an integrated G-force sensor ensures that the optical density reading is always triggered at the same fluid height. It is unaffected by changes in shaking speed.

The battery-powered SFR vario system is wireless, with Bluetooth data communication. For very long measurements, users can switch out rechargeable batteries without stopping measurements. PreSens offers disposable shake flasks with integrated sensors and an autoclavable version with reusable oxygen sensors.

Fischer provided examples of some animal-cell and yeast culture applications. In one experiment, a 1-L shake flask was used with different culture volumes and a constant shaking speed. At a relatively high volume (500 mL), the DO concentration fell after about 40 hours of cultivation because of low aeration. Lower filling volumes allow for better aeration. Results showed spikes that represented severe effects of stopping the cultures to draw samples. At a German university, another team used the SFR vario system to monitor pH of a Chinese hamster ovary (CHO) cell culture in a CO₂ incubator. Every time someone opened the incubator door, the loss of atmosphere affected the culture pH. But online measurements allowed monitoring of the cultures without opening that door to take samples.

A team at Roche used the system with a CHO culture to monitor and maintain DO concentrations and pH values within specified ranges. This team’s control methods were simply adjusting the shaking speed or CO₂ concentration. Higher shaking speeds provide better aeration and thus higher DO concentrations. And of course, pH changes with the partial pressure of CO₂ in an incubator. The team never draw samples in this experiment.

A biomass-monitoring example involved dioxide metabolism of Escherichia coli with glucose and lactose monohydrates. The bacteria digested glucose first, and after that was exhausted, the biomass increased as they changed their metabolism from glucose to lactose. Again, drawing samples caused spikes in the DO curve. With a yeast culture, glucose metabolism produced exponential growth, which slowed to linear growth when the microbes switched over to lactose metabolism. DO values decreased as biomass increased, then when no feed was left, growth came to a full stop.

View the full presentation video