September 1, 2006 (Vol. 26, No. 15)

Finding the Optimal Window of Operation and Steering Clear of Bottlenecks

Process development and integration of unit operations during the drug development cycle for a typical microbial or mammalian cell culture process is important. Most biopharmaceutical companies have their own in-house process development programs. In addition, several CMOs are involved in process development and the production of material for preclinical, Phase I–III, or even GMP manufacturing.

Despite all of this, there is often a lack of vision, especially from small research companies, for process development, resulting in a nonscalable and highly expensive process. Developing a scalable, robust, and regulatory-compliant process requires time, sophisticated equipment, skilled scientists and engineers, and, above all, money (Figure 1). Process development can only be justified if the developed process is scalable, cost-effective, and can be validated. Integration and selection of unit-operations from the research phase is of prime importance and making the right decisions early on can definitely save time and money in the later stages of product development.

Figure 1: As the product moves toward FDA approval and GMP manufacturing, costs increase exponentially.

Integration of Discovery-Development Phase

The potential to streamline process development by integrating it with earlier discovery/research phases or development is great. As part of upstream integration, researchers and molecular biologists can influence the cultivation process and the productivity of cells by the selection of the host organism, strain, vector, and clone.

Often a molecular biologist incorporates an antibiotic-resistant gene during vector design and selection. Selection of some antibiotics, such as penicillin, ampicillin, and chloramphenicol, can create a problem for process development and GMP manufacturing, especially when the stability of the expression vector and the clone is dependent on the addition of these antibiotics.

The use of antibiotics is a factor for consideration in the development process of a drug substance. Close cooperation between discovery and process development scientists delineates requirements such as shift toward nutritional selection, strains that can minimize the secretion of acetic acid, strains that are needed for protein and/or plasmid stability, and fusion of target proteins.

The choice of host organism is based on factors such as the nature of the product (e.g., essential post-translational modification) and achieved product yields. However, other factors such as costs, in-house experience, and equipment availability are also important. The molecular biologist is often unaware of those factors and therefore the involvement of process engineers or scientists in strain selection is of significance.

Metabolic pathways are another important consideration. It is crucial that the selected strain uses its metabolic pathways efficiently to increase yield of the product of interest simultaneously with reduced production of byproducts that can be inhibitory or could degrade the product of interest.

An important contribution from molecular biologists can be the addition of specific tags to the protein of interest during vector and clone preparations. The use of tags reduces downstream development time because specific, single-step chromatography can be used. This enhances the recovery and yield by reducing the number of necessary steps for product purification. However, it should be understood that the involved step is expensive, and careful consideration must be given to benefits versus costs.

Molecular biologists can also help by selecting the ideal strategy for protein expression. In microbial systems, the product of interest can be found intracellularly or extracellularly depending on different factors such as organism, vector, nature of the product, and complexity. The formation of inclusion bodies has often been exploited because inclusion bodies consist mainly of the product of interest and can easily be isolated by centrifugation.

Additionally, the product is better protected in inclusion bodies compared to secreted proteins. On the other hand an intracellular product requires cell lysis, which has some disadvantage for downstream processing due to its higher load of cell debris, resulting in a more complex process of elimination. In most cases, secreted protein reduces the complexity of downstream processing by reducing the load of host cell debris, DNA, RNA, and other intracellular proteins.

Typically a scientist at lab scale uses entirely different methods than the ones used by a bioprocess engineer at the process development or manufacturing stage. For example, if the protein of interest is expressed intracellularly, a research scientist will prefer to use chemical lysis methods rather than the homogenization typically used at large scale.

Although it is not possible to use the same or similar methods at small-scale, one should always try to mimic large-scale conditions. A good starting point could be the use of a small-scale tangential flow filtration device when the process is moving toward lab-scale instead of centrifugal filter or stirred cells for concentration and ultrafiltration purposes.

Integration of process development operations often improves process efficiency and gives better control over process inconsistency. Most of the time companies try to streamline and improve individual unit operations like cell expression, fermentation, filtration, and chromatography.

Integration of Process Development Operations

Process improvement is usually defined by the single goal of increasing the final yield. This leads to an increase in the fermentation yields and often neglects the importance of interactions among unit operations, making fermentation gains less valuable. This can result in an increase in the fermentation volume to meet the large-scale demands even though a significant improvement in upstream yield was already achieved.

This constant improvement in upstream yield has now led to a situation where gains from fermentation are less important due to severe bottlenecks on downstream operations. Using the integrated process design approach one can find an ideal window of operation where the capacities of the unit-operations are optimal and therefore the gain is maximized.


Fermentation, as the first unit operation in the process, has a major impact on other subsequent unit operations (Figure 2). Fermentation medium is an important parameter that can affect the entire process. Bacterial fermentation can be carried out using either chemically defined or complex media. Selection of the medium impacts the growth rate, productivity, cell viability, foaming, and by-product formation.

These factors correlate with fermentation efficiency and optimization but they also influence downstream operations. For example, one of the direct effects of media selection that can be observed during homogenization is cell rigidity and cell wall composition. Both factors affect cell breakage and thereby performance of the homogenization. Cells grown in complex media typically have rigid cell walls and more effort is required to break the cells open, compared to the cells grown in defined media.

Similarly cells grown in defined media are prone to break open during the centrifugation step due to high shear. This results in an increase in the impurity load for the purification of a secreted product. Antifoam addition to the fermentation broth is also common and often unavoidable. Selection of the right antifoam and the right quantity to be used is critical for downstream processing. The addition of antifoaming agent can rapidly foul the membrane and reduce the flux. It can also affect the expanded chromatography capture step and reduce its efficiency to capture the product or impurities, depending on its objective.

Figure 2: Effect of fermentation on subsequent unit operation.


Time of harvest affects the cells ability to be lysed during the homogenization step. This is due to the cell’s fragility during the exponential phase, where cell division is rapidly happening, causing weaker cell walls compared to the cells in the stationary phase. Low cell viability and extended batch or fed batch results in more cell-related impurities compared to short batch or fed batch harvested during an early stationary phase. Processes using TFF as an initial clarification step are directly affected by such variations due to the inconsistency of feed material.

Homogenization can create variability in feed for centrifuge and TFF depending on the number of passages and pressure applied for cell rupture (Figure 3). Multiple passages with high pressure results in particulate load that has fine cell debris with a complex size and density distribution. Integrating homogenization with centrifugation and finding the ideal window of operation optimizes the efficiency. This includes obtaining the target product recovery with a higher centrifuge flow at a minimum number of passages with an efficient rupture pressure.

In the event that this is not done, variability from homogenizer output can affect not only the centrifuge efficiency but also TFF flux due to the inability of the centrifuge to capture fine particulate impurities present in supernatant.

Another approach would be to integrate the latest technologies available, such as production of bacterial ghosts by protein E-mediated lysis (lysis tunnel formed by the expression of PhiX174 lysis gene). As a result, the expensive, time-consuming, and risk-associated homogenization process could be avoided.

This method of lysis could also ease the primary recovery process as the ghost cells are still intact, and the particulate material is minimal compared to other physical methods of lysis. The contaminants, such as DNA and RNA, could be degraded during the process as the respective enzymes that catalyze the degradation could also release during the process.

For extracellular products, centrifugation becomes a direct target to create a window for subsequent NFF (normal flow filtration) or TFF optimization.

Integrating process development with GMP is advantageous because it reduces the uncertainty of scale-up issues. It can also help by speeding up the validation process because most critical parameters are defined during process development. Speed to market is achieved by meaningful process development. In addition, it eases the tech transfer from the process development lab to the manufacturing facility.

Effective process development needs to keep large-scale processes in mind while developing the process and defining the critical process parameters. One example is the scale-up of process parameters such as dissolved oxygen (DO) control. The process development lab uses pure oxygen or higher agitation/impeller revolution per minute for DO control, which might be necessary to reach higher cell densities and product yield. Such control measures may not be available at large scale.

Another example of the benefit of integrated process development and manufacturing is the avoidance of issues associated with heat transfers at small scale, which could be problematic as the scale increases.

Similarly, for downstream operations, designing filtration process must take into account filter geometry, protein binding, pump requirements, linear scalability, and batch-to-batch variations. When designing concentration and diafiltration steps one must consider the effectiveness and impact at large scale. For example, it might be possible to do several buffer exchanges at a small scale without any volume consideration, but at a large scale this may become nearly impossible. Likewise, when designing a chromatography process several column volumes of equilibration, washing, and elution may be common but this must be minimized during the final stage of process development. The amount of buffer used by chromatography and filtration processes can be high if the process is developed without consideration of scale. It can also impact WFI, buffer preparation area, and buffer storage capacity.

The discovery phase influences the course of process development in different areas such as selecting the host organism, protein expression, and consequently, purification. These are extremely important factors in late phases of biopharmaceutical manufacturing.

Process development is a vital part of GMP manufacturing and must meet the expectations of developing scalable, cost-effective, and robust processes. Methods and techniques used at process development scale to achieve purity and yield must be feasible and practical at large scale.

Integrated process design is of utmost importance. Integration of all stages early on in process development allows companies to steer clear of bottlenecks in the overall manufacturing process. Using integrated process design, an optimal window of operation is found for the process at large scale and thereby efficiencies realized.

Figure 3: In this scenario, the TFF flux is a function of centrifuge efficiency and impurity load from the homogenization feed.

Mahendra Kumar Pallapothu ([email protected]) is senior scientist for product development, Novartis Animal Health Canada, Aqua Health Business. Jignesh Padia ([email protected]) is process development associate and Anke Kayser, Ph.D., [email protected]) is process development scientist at QSV Biologics. Web:

Previous articleAnna Barker, Ph.D.
Next articleAffymetrix Selected to Genotype SHARE Heart Study Samples