Non-engineers, myself included, have a simplistic view of scaleup. All that is required is to add more, make it bigger, get a larger work area, and presto. Unfortunately for those who work with complex manufacturing processes, particularly for the production of therapeutic proteins, scaling up is a bit more involved.

Proper planning, with the desired end result in mind, is essential when scaling bioreactors from bench to pilot to production scales. Essential to the process is physical characterization of critical parameters such as volumetric mass transfer coefficient (kLa), mixing, volumetric power input, and heating/cooling.

“The engineering design space, which is determined at lab scale, needs to account for potential limitations at large scale, and process development systems must be characterized as a proper scale-down model of the production scale,” explains Thorsten Adams, Ph.D., director of product management at Sartorius Stedim Biotech.

In addition to its normal bioprocess support products and services, Sartorius is developing a scaling tool incorporating all of the company’s bioreactor characterization data. This tool will help manufacturers define their process design space and calculate automatic scale-up of the dependent parameters such as gas flow, stirrer speeds, and pumping rates. “It will eventually be available as standalone software enabling complete process scale conversations,” Adams says.

Defining a robust design space is therefore essential for successful scale-up...

Defining a robust design space is therefore essential for successful scale-up, as is recognition of the potential impact of process parameters on product quality and yield.

“This requires a large set of experiments, usually several hundreds,” Adams says. For this Sartorius employs its own automated ambr250 high-throughput bioreactor system combined with a design of experiment (DOE) approach, which “has improved data quality and dramatically increased the pace at which the process design space can be established.”

Biocompare’s Bioreactor Search Tool
Find, compare and review bioreactors
from different suppliers Search

ambr bioreactors feature pH, DO, and temperature measurement, plus off-gas analysis, and may be combined with a fully integrated bioprocess analyzer. “Adopting a process analytic approach that includes online biomass determinations or spectroscopy at scale help us to move away from a time-based to an event-based process control, for example for establishing optimal harvest points or monitoring virus production for vaccines.”

In the past, scale-up challenges centered on yield, which was problematic due to the inherent physical limitations in duplicating mixing times and mass transfer. Now attention has turned to optimizing product quality, which is often expressed in a protein’s glycosylation pattern, its aggregation behavior, and heterogeneity.

“These considerations are even more relevant for newer molecule formats like bispecific antibodies,” Adams adds.

Countering inhomogeneities

When scaling up bioreactors the main challenge is the presence of inhomogeneities between scales, which are unavoidable, affect process performance and product quality, yet are difficult to address during development. These inherent differences affect process performance and key product attributes from the perspective of yield (or productivity) and quality of therapeutic proteins.

Zsolt Popse, president of Sysbiotech, notes, for example, how hydrostatic pressure, which is higher at the bottom of the bioreactor, influences gas solubility and that mixing efficiency can generate gas and mass transfer differences. “Furthermore, fluid dynamics vary substantially causing a different hydrodynamic environment. For fragile mammalian cells that hydrodynamic stress is a main concern in larger vessels.”

Using small-scale (scale-down) bioreactors that mimic their production-scale counterparts is one way to come to grips with this phenomenon. However, Popse notes that even small differences in design (and to a lesser extent performance) between scales “may be unavoidable because of technical reasons and equipment availability limitations.”

Process developers, therefore, turn to computational modeling to assist in scaling up bioreactors, employing a variety of scaling criteria such specific power input, impeller tip speed, impeller shear rate, specific impeller pumping rate, and mass transfer coefficient, KLa. Since maintaining all variables constant simultaneously is impossible due to restriction in design and configuration of different bioreactors, constant specific power input is the most often used scaling criterion.

Within this context, any process sensing or monitoring implemented at small scale, and that improves process understanding and control, will facilitate the scale-up path. In pharmaceutical manufacturing, this is achieved by introducing a Process Analytical Technology (PAT) and implementing a Quality by Design (QbD) approach.

“On the other hand,” Popse explains, “computational fluid dynamics (CFD) can be used to identify critical constraint in geometry for scale-up, simulate and optimize mixing, gas hold-up and mass-transfer coefficients, and distribution of gases within bioreactors. CFD provides the potential for simulating the large-scale fermentor’s behavior using small fermentors.”

Nowadays there are a lot of sensors available on the market that are easy to implement on bioreactors, whatever the scale. CFD is a non-intrusive method that can be implemented on a non-expensive small-scale glassware bioreactor.

Popse notes that both quality and yield may be difficult to maintain or improve during scale-up, but if small differences in yield are acceptable then quality between scales should not be compromised.

Replication is key

ABEC has long been a player in the stainless steel bioreactor market, with products ranging in size from tens of liters to tens of thousands of liters of working volume. What is not perhaps as known is that ABEC also produces the largest-volume single-use bioreactor, with up to 4,000 liters of working volume. Single-use or disposable bioreactors have become standard production tools for many development-stage protein therapeutics, and have made significant inroads into production.

ABEC 4,000 L BioreactorWhen designing a large-scale bioreactor to replicate results at smaller scale, ABEC considers three main factors: mixing/blending, mass transfer (e.g. of oxygen and carbon dioxide), and shear forces on cells. “The relative significance of those factors can vary from customer to customer,” says Brady Cole, vice president for commercial operations, ABEC. “It depends on what factors have the biggest influence on product yield and quality.”

During the ten or so years between the introduction and widespread adoption of single-use plastic bioreactors, a common criticism held that scalability was difficult in the transition between plastic and stainless steel. Cole agrees that discrepancies in mixing and aeration exist, leading to poorer performance at larger volumes in single-use than in steel. Abec’s solution is to customize products by matching mixing and aeration designs in the two platforms, thus facilitating scale-up from one to the other.

“For us, the issue of steel or plastic doesn’t matter with respect to performance. We consider those same parameters, at whatever scale, and perform scaling studies. In the end we can provide a product that delivers equivalent performance.”

Moreover technologies exist to enable replication of critical sensing and monitoring between scales and formats. Cole notes that some process characteristics, such as shear or blend times, are typically not measured in real time anyway. ABEC models bioreactors at experimental scale for those parameters that are followed, using CFD and other quantitative methods, “to assure that those characteristics remain relatively constant within pilot and production bioreactors.”

Image: Sysbiotech offers a unified control system C-Bio, which facilitates bioreactor scaleup. One interface serves bench-scale bioreactors, from bench scale up to 20 000 L, with all fermentation data stored in one place for easy access.

Image: ABEC 4,000L CSR Bioreactor