Development of a Full Process Train, Single-Use Facility

November 2, 2012
Richard Pearce

Volume 2012 Supplement, Issue 2

This article describes best practices for implementing a single-use process train at a bioproduction facility.

The biopharmaceutical industry continues to face enormous pressure to accelerate time to market, improve productivity and efficiency, and reduce costs. To address these challenges, drug developers are actively evaluating and leveraging single-use products and systems. The move to single-use solutions, either alone or in conjunction with stainless steel, is likely to continue growing to address these imperatives.

To date, many discussions on the topic of single-use system implementation have been theoretical in nature rather than describing actual examples of implementation and the rationale guiding the decision-making process. This article provides insights into how a single-use process train was developed at EMD Millipore's Advanced Bioproduction Facility in Martillac, France, how templated processes are being leveraged, and offers some best practices when considering a single-use process train.

The Advanced Bioproduction Facility has been designed as a single-use facility for the manufacture of biopharmaceuticals (see Figure 1). Owned by Serono since 1994, the facility became part of Merck KGaA in 2007, and was transferred to EMD Millipore in 2012. The 37,000-square-foot site, which has been inspected by FDA and Agence Francaise de Securite Sanitaire des Produits Sante (AFSSAPS), can provide support across different stages of production of a therapeutic protein, including clone selection, cell banking, upstream and downstream process development, GMP production, formulation, and fill-finish.

The facility is used to provide biodevelopment and clinical-supply services for biopharmaceutical companies globally. The combination of single-use technology with template processes make it possible to go from clone selection to GMP product in twelve months, whereas the typical timeframe is on the order of 15 to 18 months or longer for a company doing this for the first time.

Figure 1: Single-use, upstream production suite with a 200-L bioreactor.

Adoption of single-use technology is growing, especially in the area of small-scale or clinical-scale production. The drivers for this adoption are numerous and include:

  • Reduction/elimination of cleaning and cleaning validation costs

  • Elimination of carryover

  • Reduction in turnaround time between batches/campaigns

  • Ease of duplication of manufacturing suites in multiple locations

  • Increased flexibility.

From EMD Millipore's perspective, the industry is actively considering what its standard single-use platforms are, particularly around bioreactor systems. Many customers are evaluating or have evaluated use of single-use bioreactors for both feed train and in the GMP environment at least for production of preclinical material. Gaps in the downstream area are now being filled though single-use chromatography systems that are as efficient, as easy to use, and as flexible as traditional systems.

Adding to the interest in single-use systems is the ability for systems and technologies to handle larger batch sizes. Until recently, single-use systems did not have capacity to handle a one kilogram batch of protein, but the newer systems, in particular tangential flow filtration (TFF) and chromatography, offer the necessary capacity.


The main benefit experienced through implementation of EMD Millipore's single-use facility is reduction in changeover time, which delivers increased flexibility within the production facility. The changeover time for our 200-L seed bioreactor has been reduced from greater than 24 hours for our stainless-steel bioreactor to less than three hours. Also, set-up times are significantly reduced; for example, a TFF step that previously took at least 12 hours and required a 12-hour shift can now be done in less than 7 hours for exactly the same process, just by using the single-use equipment.

With single-use technology, capacity can be increased very quickly by adding new bioreactors and, in addition, capitol costs are significantly lower. For example, a new bioreactor can be in installed for under €350,000 ($451,000). With single-use technology, one can increase capacity within two to three months rather than the 6 to 12 months typical for a stainless-steel-based facility.

With a single-use facility, if a bigger filtration system is needed, for example, one that has been validated for use in the facility, it can simply be wheeled in without the need to reengineer the room or utilities. Within a fixed stainless-steel plant there is limited flexibility and it is much more difficult to make modifications.

With increased speed in changing out from one batch to another, multiple small batches can be run through an existing facility, so scheduling within the manufacturing plant becomes less of an issue.

Another benefit of single-use technology is the ability to develop a process in one location and then when a facility is ready, easily move the equipment to that second location. With single-use technology, process development can be conducted in parallel with construction of the new facility. Traditionally, with a stainless-steel facility, everything is done in series, and very little can be done in parallel. In an off-line location, a process can be developed using the equipment and once the production facility is ready, equipment is dropped in, plugged into the wall for the air and the electricity, and effectively it is up and running.

Single-use systems have a limitation of 1000 to 2000 L and scaling beyond that would typically incorporate either traditional stainless steel or a hybrid approach. EMD Millipore has taken the same approach to scaling up the single-use systems as it would for stainless steel. Predesigned systems make this approach easier, where one designs the experiment around the systems rather than designing the systems around the experiment. When designing the Martillac facility and the processes that are run in it, engineers defined a set of operating ranges and targets that were not scale dependent.

A number of best practices can be applied when considering implementation of a single-use process train. To begin with, it is necessary to think differently about how the facility will be run and what the facility must do. The traditional approaches of engineering are not necessarily applied. The designer must consider how to get flexibility, how to move equipment in and out of rooms and maintaining the integrity of those operations—as an example. In designing the facility it is some of the simple things that can make a big difference. At Martillac, the same utility panels (i.e., gases, power, water) are installed in the upstream and downstream suites to reduce maintenance and provide manufacturing flexibility. The sizes of doors and corridors and systems are aligned to enable bioreactors to be moved between the suites.

Another consideration is how the manufacturing process will be scaled up: by leveraging larger bioreactors or replicating the manufacturing process by adding more production lines? And then, will that process be run in one location or in multiple locations?


In addition to implementing a full single-use process train, EMD Millipore leveraged a templated downstream process for monoclonal antibody (mAb) production at the facility. This approach reduced process development time, reduced the time and effort required for equipment specification and procurement, and most importantly, enabled these activities to be conducted in parallel, thus shortening overall project timelines. The process template was designed for both laboratory and production scale. A scaled-down version of the system was developed to confirm that the processes will scale from 3 L to 50, to 200, and 1000 L; that model enables parallel process development and a rapid move into the production environment.

The templated process helps optimize integration of the purification technologies. For example, there must be a good fit between clarification and the capture protein A step. The right processing conditions then ensure that when the protein A is eluted off, it is done at the right concentration and pH to enable for viral activation, and then with minimal adjustment in processing can move straight into the ion exchange purification, for example.

The systems must also be flexible to be sized appropriately. Typically at this stage, the preclinical and phase one materials have expression level ranges from less than 1 g/L to up to 3 g/L. The single-use systems need to be connected in such a way that they are flexible enough to change the size of the chromatography column or the size of the filtration system depending on the output of the bioreactor.

The advantages of applying a template approach to the downstream processing of mAbs are well known and include standardization of unit operations, buffers, certain operating parameters, and equipment. This standardization allows streamlining and minimizing process development, rapid scale-up, and facilitation of technology transfer.

In the development of clinical supply, speed is more important than creation of a fully optimized process. Clearly, the process must deliver material of acceptable quality, yield, and purity but optimization for a commercial process can be left until the molecule shows promise in the clinic. As such, a pragmatic approach to minimize time and effort makes sense.

Such an approach can be facilitated through a combination of single-use technologies and a template approach to downstream processing in which operating parameters can be preselected to reduce process development efforts. This approach is expected to be of value for smaller, emerging drug manufacturers who are unlikely to have the extensive experience needed to establish a downstream processing template. Furthermore, use of prepackaged devices, systems, and ancillaries can reduce the time and effort required for specification, procurement, and installation.

In the approach described here, devices and operating parameters were preselected based on experience. In doing so, the number of devices and parameters to screen was effectively reduced. Preconfigured standard systems reduce the specification, delivery, and implementation time and effort. The risk element is considerably reduced through use of well-established protocols, data collection, analyses, and scale-up tools.


To demonstrate proof of principle, processes were developed for two different mAbs from two different cell lines. One of the processes was then scaled to the 100-L level using predetermined sizing and selection tools. These tools allowed process development scientists to determine not only the membrane area or resin volume required for a particular bioreactor volume and titer, but also the specific device and system size, and associated bag and mixer sizes.

The process delivered effective clearance of impurities in two different production runs for two different monoclonal antibodies; impurity levels were below targeted levels at the end of the process.

We then scaled up the process in a 200-L bioreactor for one of the antibodies. Overall, yield at both scales was approximately 85% which is satisfactory for a clinical-scale process. Product quality attributes were also assessed. There were no differences in the charge variant distribution pattern between scales or between unit operations. Distribution of change variants was the same in final pool from both scales.

The overall process required less than 12 months from start to finish. Manufacturing clone selection required three months; process development and fitting was completed in two months, during which all assays and a confirmatory run of the final process at bench scale were completed. The preparation time for pilot scale required about one month and included a water run which was found to be exceedingly helpful to confirm the workflow and ensure everything fit properly together before conducting the three 200-L scale runs.

This project demonstrates the value of combining a process template with a preconfigured, single-use process train. Benefits included reduced time and effort for process development as well as equipment specification and procurement. Most importantly, this approach allowed both of these activities to be conducted in parallel, thus shortening the overall project timeline.

Single-use technology has come a long way in recent years. And while there is a need for standardization, there is also a need for continued innovation in this area. If the industry tries to standardize too many things, too early, there is a risk that it may actually end up limiting itself in terms of finding innovative approaches.

RICHARD PEARCE is director of strategy and business development, BioPharmaceutical Process Solutions, EMD Millipore, Bedford, MA,