Suitable scale-down perfusion systems generally have been unavailable for process development (PD) activities. Some commercially available systems require daily media exchanges. No such system performs in a way that accurately represents large-scale perfusion, and none maintains sufficiently high cell densities. Kevin Lee (cofounder of Erbi Biosystems) joined BPI on 4 May 2021 to explain how his company’s Breez bioreactor system integrates all the functions of a stirred-tank reactor (STR) into a compact format that can facilitate PD, enabling one person to run three to four times as many experiments as with an STR.

Lee’s Presentation
With a 2-mL working volume, the Breez bioreactor is one of only a few milliliter-scale perfusion systems. It consists of a single-use cassette bearing rigid microfluidic channels covered by a silicone membrane. Requisite bottles, pumps, mixers, tubes, filters, and sensors are integrated into the cassette such that the bioreactor is packaged as a presterilized, precalibrated “pod.” Each pod can be programmed to perform sophisticated control strategies automatically, including n – 1 perfusion, cell concentration, and media exchanges. But a Breez reactor occupies ~8.3% of the footprint needed for a 3-ft² STR and requires far less media and fewer reagents, substantially driving down costs for raw materials and associated cleanroom and laboratory overhead. A standard Breez system comes with 4–12 pods that can operate independently and be inoculated asynchronously, facilitating study of different culture parameters.

In a typical process, fluid is introduced from one of four bottle inputs into the cassette, then pumped into a growth chamber comprising three compartments with connecting channels. The channels bear cavities of specially designed shapes and sizes. Fluid expulsion from one compartment generates flow, mixing, and gas transfer in the other two. The base of the growth chamber holds a cell-retention device that leverages fluid movement within the system to perform continuous cleaning. Automated valving enables direction of liquid to cell-free filtrate, waste, or a port for cell sampling.

A highly developed manufacturing process helps to ensure repeatability across cassettes, ensuring robust results. And because silicone is highly permeable to O₂ and CO₂, gas concentration can be controlled reliably across a larger range than expected.

The bioreactor is equipped with pairs of pH and dissolved oxygen (DO) sensors that ensure against photobleaching during long-term runs and two optical density (OD) sensors that are designed not to drift over time. Because the process harvests material continuously, the reactor — despite its small size — can provide sufficient samples for all common analytical instruments, including those for protein characterization.

Case Studies: Lee provided several examples of the Breez reactor’s applicability for PD. Erbi collaborated with Sanofi to compare the system’s performance with that of an STR. In one experiment, Sanofi performed four replicate runs, expressing a MAb) using a CHO cell line. Cultures from both systems were consistent across all four runs. Breez reactors averaged MAb titers of 1 g/L, with cell densities of >140 × 10⁶ and >90% cell viability. The team took few off-line samples because it was able to use system OD readings as a proxy for cell growth rates.

Erbi performed similar studies with MilliporeSigma, comparing MAb production in Breez and 3-L glass bioreactors using two cell lines. Breez system cultures exhibited less stress and higher viability than those from the glass reactors, and during further testing, the systems performed the same way under varied control conditions. MilliporeSigma scientists praised the Breez reactor’s integrated cell-retention device, automated cell bleeding, and pH and DO controls.

With Sweden’s KTH Royal Institute of Technology, Erbi has investigated high-density inoculation to accelerate time to steady-state perfusion. Seeding cultures with >20 × 10⁶ cells/mL, the team reduced the growth phase from one week to less than four days. That capability could facilitate media optimization studies significantly. Lee also described how a cell therapy company is applying the Breez system to study how different media affect donor-cell growth rates and phenotypes. Because the system uses small volumes (1–2 million cells), the same donor material and media can be used across multiple reactors with consistent results.

Overall, Breez users comment most about how the system enables more experiments using the same resources, saving time and money without incurring significant capex costs and without the complexity of some automated solutions.

Questions and Answers
What cell types can be cultured in a Breez reactor? Customers have cultured Escherichia, Rhodococcus, Clostridium, and Actinomyces bacterial strains. Mammalian hosts have included CHO (up to 250 million cells/mL) systems, HC-20 hybridomas, HEK cells, and human T cells (>100 million cells/mL). Erbi is eager to expand to other cell types.

Learn more about the Breez bioreactor system at https://erbi-bio.com.

Watch the complete presentation now.