Vous êtes sur la page 1sur 5



A Brief History of Perfusion

How High-Concentration Cultures Will
Characterize the Factory of the Future

by John Bonham-Carter and Jerry Shevitz

odays renewed interest in products were realized in the late
perfusion culture is due to an 1980s and early 1990s. In those early
increased awareness of its days of the modern biotechnology
advantages, some general industry, production cell lines were
improvement in equipment reliability, not fully developed, and their product
and a broadening of operational skills expression was very small from a
in the biomanufacturing industry. few micrograms to a few hundred
Some misperceptions persist, however, milligrams per liter in batch or fed-
according to a 2011 review by Eric batch. Attainable cell concentrations
Langer (1). Our view here of the were only a few million per milliliter.
history of perfusion and fed-batch Spin Filters: Perfusion offered a way
processes includes some discussion of to derive more product from such low
technological process improvements producers. It was well known that
and challenges that the bioprocess perfusion could increase cell
industry faces. concentration by as much as an order
A team of authors at Serono in of magnitude (3). The spin filter was
Switzerland wrote in 2003: the most common perfusion device
used; it was the best cell-separating
The major advantage of the
Perfusion performed at CMC Biologics device available at the time, supported
perfusion mode is high cell number (www.refinetech.com)
by reputable equipment
and high productivity in a relatively
small-size bioreactor as compared
alternating tangential-flow filters, Spin filters remain in use at a few
with batch/fed-batch. In order to
vortex-flow filters, acoustic settlers sites but have been largely phased out,
sustain high cell number and
(sonoperfusion), and hydrocyclones. largely because of their limited scale-up
productivity, there are needs to feed
All are described well in the 2003 potential and unreliability: When a
medium during the cell propagation
paper mentioned above. But only a bioreactors volume scales up by the
phase and the production phase. In
few types are reliable at larger scales cube of its radius, the surface area of its
contrast to batch and fed-batch
and scalable enough for bioindustrial spin filter screen scales by the square of
processes, where there is no
use. its radius. An internal spin filter can
metabolites removal, in continuous
Here I compare the ATF System take up a significant portion of
processes medium is perfused at
from Refine Technology with spin production space within a vessel, and
dilution rates exceeding the cellular
filters, cell settlers, and centrifuges. I once its screen fouls, the run is
growth rate. For this, a good
am not including other technologies terminated. An external production
separation device is needed to retain
here because of scalability limitations spin filter may solve this shortcoming,
cells in the bioreactor. (2)
and a lack of proven market but it has drawbacks related to cost,
Many cell retention devices acceptance. maintenance and sterilization
perform well, to a greater or lesser difficulties.
degree, at small scale, including Perfusions Early Potential A more important factor behind
gravity-based cell settlers, spin filters, The advantages of using perfusion for the lackluster acceptance of perfusion
centrifuges, cross-flow filters, enhancing production of cell-derived in those early years was the rapid
24 BioProcess International 9(9) O ctober 2011
evolution of cell biology. New, more particularly in the past few years
productive expression systems and another critical transition in
improved media development biopharmaceutical manufacturing
permitted large increases in culture occurred. Further advancements in
productivity; product concentrations development of cell lines, expression
were increasing from several hundred systems, and media formulations
milligrams to about a gram per liter. resulted in an impressive ability to grow
Production needs could, therefore, be cells to very high concentrations and
achieved with the well-understood achieve product concentrations
fermentation technologies, batch and previously inconceivable. Using fed- Microcarrier-based perfusion in a single-use
fed-batch. Scale up was accomplished batch as a reference, in the mid 1990s ATMI bioreactor (www.refinetech.com)
simply by moving to bigger vessels. attainable cell concentrations were about
The success of batch and, more 5 106 cells/mL, with record product So the face of biomanufacturing
important, fed-batch, not only concentrations of 12 g/L; today those today is very different from that of just
inhibited the wider acceptance of spin are greater than 15 106 cells/mL, a decade ago. Nearly everyone uses
filters, but also of other evolving cell- with product concentrations of up to perfusion in some way from large
separation technologies. The 10 g/L. Although those concentrations biopharmaceutical companies such as
difficulties associated with spin-filter are still not typical, they indicate where Pfizer, Medarex, and Genentech (68)
operations and the undeveloped state the field is heading. Those results are to small biotech and novel vaccine
of new perfusion technologies amplified by the use of perfusion, manufacturers such as CMC Biologics
stigmatized the process. The through which substantially higher and Crucell (9, 10). Outside the
dominance of fed-batch continued cell concentrations and product output established biomanufacturing
well into the next decade. can be achieved (4, 5). infrastructure, biosimilar and other
However, despite the dominance of relatively new biological manufacturers
fed-batch as an industry standard, Perfusion Returns to such as Biocon and A-Bio are also
perfusion continued to be Manufacturing looking favorably on the perfusion
championed. Perfusion offered an A general lack of manufacturing model because of its associated cost
excellent solution for production with capacity forecast at the beginning of efficiency. Perfusion is back.
unstable proteins that could not this century was overcome through Simplicity and reliability have long
remain in the toxic environment of an both biological innovation and been key factors to consider in
ever-deteriorating fed-batch culture. engineering construction. Todays biologics production, especially where
With perfusion, such products could overcapacity places most of the manufacturing involves high-value
be removed rapidly from a vessel and available space in the hands of products in a large-batch environment.
stored appropriately to preserve their relatively few companies. Even as The industry is now being challenged
stability. Many people chose perfusion some large biofacilities are as it moves forward to realize the
to bypass constraints of space and cost mothballed, newer companies build muchtouted factory of the future,
factors. Furthermore, as culture modern facilities based on the latest which will incorporate several
productivity increased, and although it technologies. Few organizations would platform technologies. One such
greatly benefited fed-batch processes, now consider building a new, technology is certainly the adoption of
perfusion promised even greater multiple20,000-L bioreactor facility. disposables throughout production
output from a continuous culture. Rising competition in the healthcare facilities.
So the use of perfusion never died; sector, whether through generics/ Perfusion is a broad term, which
in fact, as the use of spin-filters biosimilars or multiple drugs with the many people may still view
declined, other cell separation devices same indication, requires the vast unfavorably. Although many, in fact,
slowly emerged. Those were based on majority of biopharmaceutical use perfusion at some level, not
filtration, gravity settling, and products to be more easily produced in everyone admits to it nor to how
centrifugation. Continued smaller and more flexible plants they do it, nor how often. Companies
development of numerous products even in multiple locations. New are experimenting with perfusion to
that held out the promise of ultrahigh-density cell culture solve challenges or implement novel
commercialization provided the processes such as concentrated solutions at many process stages: high
driving force to experiment with new fed-batch and concentrated perfusion density, large-volume cell banking
culture technologies. Occasionally a are well suited to this new (11); seed expansion (8); n1 perfusion
perfusion process, primarily one based manufacturing environment and (12); and even final production
on using spin-filters, cell settling, and facilitate a shift toward single-use reactors (13). Perfusion has evolved
centrifugation, was scaled to technologies. That helps companies too: It is no longer solely a two- or
commercial production. reduce both risk and capital three-month process, but can be as
High-Concentrations Are a Game- investment, allowing them to delay short as a three-day boost to a
Changer: From the early 2000s and making major facility decisions. standard fed-batch process. Perfusion
26 BioProcess International 9(9) O ctober 2011
Figure 1: The ATF System scale cell-retention device that offers a
Level high degree of confidence for scaling to
Control a commercial manufacturing process.
Scaling up a bioreactor introduces its
Fluid Inlet
own issues, so engineers dont want
perfusion equipment to add further
Pump complications. Several technologies have
Pump been used at large scale, and each system
Filtrate brings its own limitations. For example,
well-known spin-filter technology,
HF or Screen Module previously discussed, uses a two-
dimensional screen to retain the cells.
Housing Limitations of the system (whether
internal or external) arise during scale
Diaphragm up and at elevated cell concentrations
when rapid feed rates are required.
ATF Controller
Pump Consequently, to reduce risks of screen
blockage, the process duration must be
shortened or the culture maintained at
has become a specialist operation. low cell concentration to prevent
Implementation depends on the excessive accumulation of cell debris on
nature of different facilities, cell the screen. The latter is usually what
lines, processes, and products as occurs.
well as each companys own operating Different but familiar problems
philosophy. Success depends on many occur with inclined or gravimetric
factors, not least of which is a settlers. Cells spend significant time
companys choice of perfusion system. in an external, suboptimal
But one challenge that of environment within the settler
producing a reliable cell-retention (particularly) as the size of a system is
device may have been solved to a increased. Additionally, as a system is
great degree by a relatively new increased, when greater perfusion rates
hollow-fiber perfusion device are required, raising recirculation flow
rates can lead to inefficient cell
Case Study separation and significant cell loss,
The ATF System (Figure 1) offers which lowers output and increases
An Xcellerex bioreactor used in an
nearly linear scale-up for simplicity of costs.
intensified perfusion process with
operation and validation. Generally, ATF System equipment (www.refinetech.com) Centrifuges have been scaled up
conventional filtration systems will successfully for several perfusion
fail rapidly when used to separate reversibly through the hollow fibers. processes, often to very high flow
media from a complex suspension of a Flow is controlled by the diaphragm rates. However, the high level of fine-
cell culture with a high bioburden. By moving up and down in the ATF tuning required to maintain the
contrast, this particular system, due to systems pump. This generates a rapid reproduceability of such systems
its flow dynamics, has an inherent low-shear flow between vessel and particularly during scale up as well
self-cleaning ability to allow its range pump, ensuring rapid exchange and as their cost greatly discourage their
of filter materials and pore sizes to prompt return of cells to the reactor use.
perform significantly longer than and minimizing their residence Despite those issues, each cell-
might otherwise be expected. outside the bioreactor. The choice of retention device has a solid following
A standard hollow-fiber module is pore size for the hollow fiber among a number of companies.
used to separate cells and product. determines what elements are retained Skilled and experienced individuals
However, unlike systems that and which ones pass through to the maintain such systems. They assess
recirculate a culture through a filter in permeate (perfusion or concentrated and improve scale-up and scale-down
one direction, the alternating fed-batch operation modes). performance.
tangential-flow action constantly From Research to Manufacturing For companies that require simpler
cleans the fibers every five to 10 the Scale-Up Challenge: For companies systems that can be operated by a
seconds with a backflush action. With requiring increased protein production nonspecialist or that do not want to
only a single connection to the in preclinical work, many perfusion devote years to building those requisite
bioreactor, cells and media enter and technologies can quickly deliver. One skills, the ATF system can provide a
leave the ATF system, flowing common approach is to choose a small- robustly scalable process platform for
28 BioProcess International 9(9) O ctober 2011
most cell lines. Laboratory-scale devices are run as standard to
produce the same conditions and flows that commercial scale
devices will use. Two key parameters to keep constant are the
filtrate flow ratio and the flow through each individual hollow
fiber. Other parameters that would normally require attention
e.g., filter surface area and residence time are factored
into the equipment configuration design to limit variability
potential. Scale-up is therefore straightforward to help teams
build their confidence and experience rapidly. Additionally,
unlike the older systems, a failure in the ATF system does not

Your partner in
mean failure of the run. The perfusion device can be easily
exchanged with another in a sterile way to continue the
process. Bioreactor issues actually come to the fore: Can a

large-scale bioreactor handle the oxygen demands of a cell
concentration that is about 10 times higher than usual?

innovation A Factory of the (Near) Future

A stable cell line is a prerequisite for a perfusion process if
it is intended to produce a high-quality product for an
extended time. Considering the state of biological
manufacturing today and industry trends of the past two
decades, some features of the factory of the future can be
A Continued Move Toward Single Use: Innovations in
disposable bioreactor designs have moved the industry
toward their increased use. That trend is reflected by the
large number of companies that are currently supplying
single use Bioreactors (SUBs). Innovative SUBs from sub-
one liter to 2,000 L are readily available today. Along with
SUBs, significant improvements have been made in
processing equipment, sensors, and other components, all
with disposability in mind.
A Shortened Bioreactor Train: The ability to generate
high-cell-concentration cultures combined with the ability
to freeze large volumes of such cultures has made it
possible to create high-volume cell banks. A single sample
can be used to inoculate a relatively large bioreactor
directly, eliminating multiple steps, saving time, and
greatly increasing reliability.
Proyield for Biopharma Simplified Product Stream: Generating a filtered product
stream by filtration perfusion can shorten the steps
between vessel and column.
Concentrated Fed-Batch: In a process that can be

Cost effective considered a form of perfusion, the culture is perfused to

generate ultrahigh cell concentration, greater than 108 cells/
mL; and the product is also retained in the vessel. Product
productivity concentrations greater than 25 g/L have been reported.
The trend to higher product concentrations is not abating.

enhancement Concentrated Perfusion: Although 1 g/L/day is routinely

achievable today using concentrated perfusion,
3 g/L/day has been reported, and 5g/L/day can be
regarded as the next step. The volumetric productivity of
concentrated perfusion means that at 5 g/L/day, one 500-L
reactor would produce 2.5 kg of protein every day, and
potentially 500 kg/year.
If these goals are achieved in the foreseeable future, there is
little reason for even a high-dose blockbuster to be
manufactured in anything larger than a 500-L vessel, whereas
most other products could be handled with current laboratory-
7 Alahari A. Implementation of Cost-
scale equipment. The future size of the Reduction Strategies for HuMAb Manufacturing Further Reading
factory, for upstream processes at least, Processes. BioProcess Int. 7(7) 2009: s48s54. Li L, et al. A Single-Use, Scalable
Perfusion Bioreactor System. BioProcess Int.
looks very small indeed. 8 Tezare R. Running Inoculum Cultures
7(6) 2009: 4654.
in Perfusion Mode to Enable Higher Seeding
Densities of Production Cultures. BioProcess Lim J, et al. An Economic Comparison of
References Three Cell Culture Techniques. BioPharm Int.
1 Langer ES. Perfusion Bioreactors Are International Conference and Exhibition, 2010,
Providence, RI. 24(2) 2011: 5460.
Making a Comeback, but Industry
9 Compton B, et al. Use of Perfusion Whitford WG. Single-Use Systems As
Misperceptions Persist. BioProcess J. 9(2) 2011:
Technology on the Rise. Gen. Eng. News 27(17) Principal Components in Bioproduction.
2007. BioProcess Int. 8(11) 2010: 3442.
2 Voisard D, et al. Potential of Cell
10 De Vocht M. Intensification of a PER. Whitford WG, Cadwell JJS. Interest in
Retention Techniques for Large-Scale High-
C6-Based Recombinant Ad35 Manufacturing Hollow-Fiber Perfusion Bioreactors Is
Density Perfusion Culture of Suspended
Process to Prepare for Commercial Scale Growing. BioProcess Int. 7(9) 2009: 5464.
Mammalian Cells. Biotechnol. Bioeng. 82(7)
2003: 751765. Production. Fifth Annual Optimising
3 Shevitz J, et al. Stirred Tank Perfusion Biomanufacturing Processes, Brussels, Belgium, 3 Corresponding author John Bonham-
Reactors for Cell Propogation and Monoclonal December 2008. Carter is vice president of sales and
Antibody Production. Advances in 11 Noe W. Paradigm Changes and business development, and Jerry Shevitz is
Biotechnological Processes, Vol. 11. A. Mizrahi Technology Gaps in the Biopharmaceutical president and director of R&D at Refine
(Ed). Alan R. Liss., Inc.: New York, 1989; p81. Industry. BioProcess International Conference and Technology, 26 Chapin Road, Unit 1107, PO
4 Carstens J. Perfusion! Jeopardy or the Exhibition, October 2009, Raleigh, NC.
Box 691, Pine Brook, NJ 07058; 1-732-993-
Ultimate Advantage? BioProcess Int. webinar, 12 Yuan H. A Platform for Late-Stage 3003, 46-76-256-5114; jbonhamcarter@
2009 Sept. Monoclonal Antibody Process Development:
refinetech.com; jshevitz@refinetech.com.
5 Bonham-Carter J, et al. Which Process Robust High Seeding Density CHO Cell
Option Is Right for Me? Bioresearch Online 18 Culture Process Utilizing Perfusion at N-1
October 2010. Stage. IBC Antibody Development & Production, To order reprints of this article, contact
Bellevue, WA, 1618 March 2011 Rhonda Brown (rhondab@fosterprinting.com)
6 Yuan H. Cell Culture Process Scale-Up,
Technology Transfer, and Assessment of Process 13 Muller, D. ATF Perfusion-Based 1-800-382-0808. Download a low-resolution
Comparability Using Multivariate Analysis: A Production Platforms for Biopharmaceuticals. PDF online at www.bioprocessintl.com.
Case Study. BioProcess International Conference ESACT Conference, Vienna, Austria, 1518
and Exhibition, October 2009, Rayleigh, NC. May 2011.




Just give us your protein sequence and well do the rest. Working closely with you, well put our process development skills and project
management expertise at your service, providing you with a GMP batch of your product highly efcient, professionally and on time.
Team up with the leading innovator in E. coli production technologies and learn about ESETEC, our unique E. coli secretion technology.
Visit www.wacker.com/biologics.

Wacker Biotech GmbH, Jena, Germany, Tel. +49 3641 5348-150, info.biologics@wacker.com