February 1, 2010 (Vol. 30, No. 3)

Angelo DePalma Ph.D. Writer GEN

Uses Include Cell-Line Stability Testing, Strain or Clone Screening, and Medium Development

With volumes ranging from the low tens of milliliters to nanoliters, microbioreactors (MBRs) have become one of the more interesting engineering stories in biotech. At the high-end, MBRs serve as versatile scaledown systems for process development, whose utility is limited only by control capabilities. Capabilities vary significantly below 30 mL or so, but this is where much of the exciting work is going on.

Bioprocessors use MBRs below about 15 mL for cell-line stability studies, strain or clone screening, cell culture medium development (including raw material testing), limited process development (pH, temperature, and dissolved-oxygen optimization), and process characterization/validation studies.

How closely the scaledown model resembles production-scale bioreactors depends on which parameters are identified as critical, how well they are controlled in the MBR, and generally how scale-sensitive the process is to begin with.

According to data from Dasgip, the number of parameters monitored during cell culture or fermentation has expanded from 12, a decade ago, to more than 40 today. Not every process tracks all these variables, but bioprocessors have come to believe that the tighter the control over critical parameters, the more reliable and robust the process, which will be reflected in product quality and consistency.

Dasgip’s BioLector, one of a small group of highly controllable MBRs, runs between 6 and 96 individualized fermentations or cell cultures in a microtiter plate format. “The 48- and 96-well plates are the workhorses,” says Karl Rix, Ph.D., CEO of Dasgip BioTools. Each well, containing between 100 µL and 1 mL, provides a mass transfer co-efficient (kLa) of as high as 1,000/hr. Wells are covered with a gas-permeable membrane to allow exchange of headspace gas.

The BioLector’s claim to fame is its ability to monitor critical process parameters, for example pH, dissolved oxygen, biomass, fluorescence, and others, through noninvasive optical sensors. “The BioLector provides not just an endpoint measurement of any of these parameters, but monitors each well over the course of a fermentation,” Dr. Rix adds. “It measures how pH and DO change over time and how quickly nutrients are depleted, for both microbial and mammalian processes.”

BioLector monitors wells through fiberoptic probes positioned below each optically transparent well, so any parameter measurable in that fashion is fair game. Turnaround time per measurement for all wells is about two minutes.

Another interesting feature of BioLector is the patented flower-shaped footprint of each well of the 48-well plate, which serves as a baffle to promote mass transfer. “It is this feature,” says Dr. Rix, “that provides results that are comparable to large-scale bioreactors.”

MBRs always involve some trade-off compared with large-scale reactors, and the BioLector is no exception. Automated liquid additions and pH control are impossible due to the membrane stretching across each well, so the device is not 100% comparable with larger reactors. “But that is not the point,” Dr. Rix observes. “Our strength is the ability to screen strains and clones, and to pick ones you’ll use at the next level rapidly and efficiently.”

The BioLector is based on technology that was developed at the RWTH Aachen University and m2p-labs. Dasgip holds marketing and distribution rights for the device for Europe and North America.

To aid the reproducibility and consistency of cell sampling or the addition of inducer substrates, m2p-labs developed RoboLector™, an automated microfermentation platform. Designed to communicate directly with  BioLector software, the RoboLector combines the BioLector platform with a robotic liquid-handling device that can be programmed to manipulate individual cultures at specific time points, cell concentrations, pH, or other parameters.

Seahorse Bioscience’s SimCell™ 200 MBR features working volumes (0.7 mL) similar to those of the BioLector. Seahorse claims the product is suitable for screening, media and process optimization, and late-stage process optimization. The SimCell Micro Bioreactor holds up to 200 MBR “cards” of six wells each and provides agitation, feeds, pH balance, and gas exchange. “There are, of course, sampling challenges with this device,” says Ashraf Amanullah, Ph.D., a director at Genentech, “but this company has run hundreds of tests and shown good scalability and reproducibility with bioreactors of up to a couple of hundred liters.”

MBRs in the next-higher size category include Applikon Bio’s family of devices. The company’s µ-24 Bioreactor, designed to replace shake flasks and microtiter plates for bioprocess development, consists of 24 vials, each with a working volume of 5 mL. Operators have full control over dissolved oxygen, temperature, pH, and gas exchange within the fully closed containers. The system offers three cassette formats and four different closure options.

“As you would customize a stirred-tank bioreactor, you can customize the µ-24 cassette closure to suit your applications,” says David Laidlaw, small scale products manager.

Applikon has another product, the µ-Flask, which works in 24- or 96-well plate formats; it allows reading of pH and DO but does not provide the same level of control as the µ-24.

“There are ways to monitor small cultures, but controls become complicated with really small volumes,” Laidlaw says. Another issue with small working volumes is restricted ability to assay or sample the process in ways that provide meaningful data. “If you can’t sample, it’s very difficult to understand what’s going on in there.”

Dr. Amanullah has used the µ-24 system for optimizing fermentations of Saccharomyces cerevisiae, Escherichia coli, and Pichia pastoris. According to a 2007 study authored by Dr. Amanullah with S. cerevisiae, he observed “high inter-well reproducibility” and runs “demonstrated comparable growth to a 20 L stirred tank bioreactor fermentation by off-line metabolite and biomass analyses.”

The E. coli and P. pastoris runs were conducted with rapidly increasing oxygen uptake rates and high cell densities, which required gas blending for DO and pH control. Dissolved oxygen and pH control were challenged in the case of E. coli batch fermentations but not for P. pastoris. “The equipment may be the same for microbial and cell cultures, but the latter have much higher oxygen demands,” Dr. Amanullah explains.

When it comes to controlled MBRs, users most often purchase off-the-shelf equipment like the µ-24 system, but it is possible to modify or add on to such systems for high-throughput applications. While he was at his previous company, Dr. Amanullah had Hudson Controls modify the µ-24 with liquid-handling capabilities to automate sampling, reagent addition, and pH control. “The other option is to choose components from various vendors and put them together yourself. It depends on how much know-how exists in your company.”

His group at Genentech uses a different, home-brewed system that employs 50 mL conical centrifuge tubes that are manually manipulated, although the liquid handling is automated.

Dr. Amanullah agrees that sampling is a problem with small MBRs used to model larger reactors (vs. screening cells and media). The 50 mL tubes he currently uses model the 2 L bioreactors that are the workhorse volumes at Genentech and permit sampling over 14-day cultures. “You don’t want to just sample frequently, but to use the same equipment that was used to sample larger bioreactors. Removing even half an mL [from a 5 or 10 mL tube] to do cell counts and metabolite analysis is difficult. You have to sacrifice tubes, which is not ideal. That’s why we decided to operate at slightly larger volumes.”


m2p-labs’ RoboLector combines the BioLector platform with a robotic liquid-handling device.

Usage for Assays

Using MBRs for cell-based assays is not a new idea. The premise is that the three-dimensional nature of reactor systems can mimic the scaffolding found in tissues. Stem cell scientists sometimes refer to this as the biological niche.

The University of Oxford has developed TissueFlex®, a gas-permeable, perfused, sterilizable MBR for evaluating drugs and performing toxicology studies. TissueFlex cultures are long-lived, so the effect of culturing does not carry over into studies based on survivability.

“The response you get in 3-D cultures are more realistic, closer to what you can expect in vivo,” says Shang-Tian Yang, Ph.D., a chemical engineering professor at Ohio State University. However, as he points out, monitoring cell responses is difficult enough in two-dimensional cultures, and extremely challenging in three dimensions due to geometric factors. His solution is to exploit auto-fluorescence in cells engineered to express green fluorescent protein.

Even here problems exist, for example, the fluorescent signal in GFP-programmed cells is weak relative to background (the culture medium autofluoresces, too). Dr. Yang overcomes this obstacle by increasing cell density (the 3-D configuration actually helps here) and subtracting out the background signal. Thus, instead of requiring specialized detection systems, he uses a conventional plate reader. “This platform is robust and can give real-time, noninvasive quantification of cell growth, which in turn can be used for drug screening as in a cytotoxicity or proliferation assay.”

On the extreme low end of the size domain, Prof. Eric Gottwald, managing director at the Karlsruhe Institute of Technology, has been working on chip-sized MBR-like devices that resemble microscopic microtiter plates. Within a size of 1–2 cm2, a device holds hundreds of cube-shaped recesses measuring 300 x 300 x 300 µm, or cylindrical wells 300 µm deep and 300 µm across, each capable of holding 6 million cells. The wells’ size and configuration permits electrical stimulation, perfusion culture, and microscopic analysis.

Creating excruciatingly tiny wells in large numbers has only recently been possible, thanks to a production technique known as microthermoforming. The technique, which blow-molds thin plastic film into a patterned template, is much more rapid than the earlier methods of hot injection molding and hot embossing.

Initially interested in highly parallel cell analysis, Dr. Gottwald worked on liver cancer cells, investigating the impact of the three-dimensional environment on proliferation, growth, and differentiation. He has subsequently broadened his scope to embryonic and neuronal stem cells, particularly the design of artificial stem cell niches.

The goal of his stem cell work, a fascinating combination of mechanics and biology, is to simulate the extracellular matrix of stem cells in their natural niches through introduction of topographic cues, which are important for maintaining or inducing specific stem cell fates. The chips are held within a container that perfuses or “superfuses” cells with medium and growth factors.

His first success was differentiation of neuronal stem cells into photoreceptors. “This is of interest because, compared with the reference culture system—roller bottles—we do not observe apoptosis that normally occurs a few weeks after differentiation.” Cell death in conventional culture makes the photoreceptor cells unusable. Dr. Gottwald believes the technique can be applied to other stem cell types, which he is now testing through several collaborations.

What’s really interesting about this work is that the cells do not need to be removed from the MBR matrix, which is composed of a biodegradable polymer. At some point, patients could be implanted with part or all of the chip after cells have differentiated.

The Karlsruhe MBRs cavities hold enough cells (up to 10,000) to permit screening cells and clones, testing culture media and feeds, etc., but the problem is their size. “The best robot available pipettes 50 nL,” Dr. Gottwald says, “but the microcavity volume is just 27 nL.” The cavities could be made larger, but only at the expense of creating a less-than-ideal niche for cell growth. “300 microns mimics the typical distance between capillaries. If the wells get any larger, we lose some of the characteristics of living tissues.”


A perfusion microbioreactor that is being worked on at Ohio State University

Previous articleU87MG Glioblastoma Cell Line Sequenced in Under a Month
Next articleRoche Takes a Stab at Getting mAbs Under the Skin