September 1, 2014 (Vol. 34, No. 15)
Angelo DePalma Ph.D. Writer GEN
To boost productivity in the manufacture of therapeutic proteins, one may try improving specific production elements—cell lines, processes, or any of the ancillary tools related to transfection. One may even try improving purification, which not only impacts quality but also contributes to overall productivity.
Producers of non-monoclonal antibody therapeutics are beginning to appreciate the benefits of platform processes that obviate the need to re-invent the wheel constantly. A group at Protein Sciences led by Nikolai Khramtsov, Ph.D., has reported on a universal process for expressing and purifying influenza recombinant hemagglutinins (rHAs), components of the company’s commercial Flublok® vaccine, at different scales without the need to redevelop the process for new rHAs. Apparently the company’s manufacturing approach, which occurs in insect cell expression systems, works for all seasonal and pandemic rHAs.
Access to a platform production method is significant, since the rHAs change every year. By expressing these antigens in insect cells, Protein Sciences avoids the legacy egg-based production method and streamlines manufacturing.
“rHA sequences change all the time, but our process of expression and purification remains the same,” Dr. Khramtsov tells GEN.
Flublok is only the fourth FDA-approved product manufactured in insect cells. The cells are transiently transfected using a proprietary baculovirus expression vector.
At present, Protein Sciences employs a simple batch strategy; however, the company plans to implement its recently patented fed-batch system, which Dr. Khramtsov says will be similar to plasmid-based mammalian fed-batch processes. Scales are typically in the 600–2,500 L range, with titers at mg/L levels. One Protein Sciences customer has scaled the baculovirus process to 20,000 L.
Getting Genes Where They Need to Be
MaxCyte has been successful at using flow electroporation to streamline development of biotherapeutics. Electroporation involves subjecting cells to an electric field that temporarily permeabilizes cell membranes, allowing external materials such as genes to enter cells. The company has used electroporation to express genes transiently in insect, CHO, and approximately 70 other cell lines, providing high yields of antibodies, antibody-like molecules, and vaccines.
The technique easily transfects 10 billion cells—a quantity sufficient to produce up to low gram quantities of recombinant protein. “Traditionally, stable pools or stable clones must be generated to produce comparable quantities,” says James Brady, Ph.D., director of technical applications. “By eliminating the need to generate stable pools or clones, our process significantly shortens development timelines.”
MaxCyte produces and sells three instruments for various cell types and quantities.
Flow electroporation yields much higher levels of transfection efficiency and post-transfection viability relative to other transient transfection methods. The differences are especially pronounced with CHO cells, which are challenging to transfect with conventional technology. Due to the efficiency of transient expression, stable or targeted integration is not required to generate high protein titers with the MaxCyte process.
“We and our customers have obtained titers exceeding 1 g/L following transfection of CHO cells with human IgG1 expression plasmids,” Dr. Brady continues. A single-flow electroporation with MaxCyte’s STX instrument provides enough cells for up to 4 L of high-density cell culture, enabling production of up to 4 g of protein from a single transfection. Tenfold higher levels of protein can be generated from a single electroporation with the company’s VLX transfection system.
Design of Experiment
Hanuman Mallubhotla, Ph.D., research director for biopharmaceutical development at Syngene International, has achieved excellent results in work to improve antibody protein production through a statistical design-of-experiment (DoE) approach in mammalian cell fed-batch processes.
Dr. Mallubhotla applied this strategy using JMP, a statistical software package available from SAS Institute, to evaluate 15 basal and 7 feed media while controlling for feed rate and temperature. He optimized media and feed based on statistically observed interactions and amino acid/metabolite profiles. The result was a titer increase from 0.5 g/L in shake flasks to more than 3.0 g/L in bioreactors. Investigators collected samples and analyzed for glucose, lactate, titer, and amino acid content.
“Optimization of the cell culture processes usually happens as an afterthought,” Dr. Mallubhotla observes. “Most companies are under severe pressure to produce the material under aggressive deadlines. We accomplished a step-by-step DOE methodology to optimize the cell culture process before going into manufacturing.”
Organizations that build this methodology into its standard development plans will be capable of developing a cell culture processes for antibodies and perhaps for other molecules, quickly and predictably, Dr. Mallubhotla adds. “It will be like an assembly line.”
Companies have attempted to reach platform development processes, but each requires revision and rework to accommodate molecular differences. Through this process—perhaps system is a better term—Syngene provides its customers with clinical and commercial material while shortening time-to-market. “We are talking about a principle that can applied to many situations, as opposed to an application platform itself,” Dr. Mallubhotla says.
A good deal has been discussed regarding scaledown methods involving small, parallel bioreactors. In April 2014, contract manufacturer Gallus Pharmaceuticals officially adopted the ambr15™ microbioreactor from TAP Biosystems, a Sartorius Stedim Biotech business unit.
“We use the ambr15 to perform many types of screening and response surface designs,” says Matthew Zustiak, Ph.D., principal scientist, cell culture development, Gallus. Optimizations include media and feed screening, feed strategy optimization, feed timing, feed quantity, and process parameters such as pH and temperature.
According to Dr. Zustiak, ambr’s main strength is its ability to run a 48-vessel, fed-batch study that closely mimics conditions inside a larger-scale bioreactor, and to perform large studies examining multiple process variables in a single device. “The ambr excels at individual control of the feed supplementation or pH control strategy since each vessel has independent control.”
On the down side, Dr. Zustiak notes, there are “clear limitations” around sampling volumes for determining product quality. For example, it is often necessary to use larger samples so that reliable product quality data can be generated along the length of the run. Another limitation is that only four temperature-shift scenarios are available. “Being able to test both the day and degree of a temperature shift in a response surface design would be nice,” Dr. Zustiak admits.
In a recent study, Dr. Zustiak’s group used ambr to develop a process for the production of a biosimilar antibody in CHO cells. They began with media and feed screening, moved onto feed strategy optimization, and continued with pH and temperature control strategy optimization. They also examined the effect of individual medium component supplementation.
“Throughout this development project, we were not only concerned with growth and titer optimization, but also in finding conditions that allowed matching product quality attributes as closely as possible to those of the innovator drug,” explains Dr. Zustiak. “We were specifically interested in the distribution of charged variants and glycoforms.”
As of the writing of this report, Dr. Zustiak has identified specific parameters that affected quality attributes, as well as a medium that provided both high productivity and reasonable quality matches to the originator protein.
The Challenge of Nonconformity
The proliferation of engineered antibodies and antibody-like proteins creates unique problems for manufacturing, particularly in how to express and purify these molecules. Purification by protein A normally requires a functioning Fc region, while activity is based on the variable regions.
As Michael Dyson, Ph.D., group leader at Iontas, has demonstrated, it is possible to screen for activity using only the variable regions, without the need to create an intact antibody. Afterwards, it is possible to express the intact molecule, including the Fc region, for ease of purification while retaining the unique affinity properties of the variable regions.
Iontas concentrates on single-chain variable fragment (scFv) regions. Each scFv is a fusion protein created by chemically connecting variable regions from light and heavy chains with between 10 and 25 amino acid chains.
The scFV regions are expressed in HEK293, CHO, or stem cells through phage display. Then they are selected for high yield and functional properties.
“The main benefit of phage display is the ability to rapidly select for high-affinity, fully human antibodies within weeks, which considerably shortens the lead time to a preclinical development candidate,” Dr. Dyson explains. Lead time is significantly longer with transgenic mice, which may also generate unwanted antibodies. Antibody phage display also generates IgGs to an unlimited range of antigens, and may be controlled to select antigens that require particular buffer conditions, or in the presence of agonists or antagonists—of particular importance when selecting antibodies to GPCRs or proteases.
With respect to larger-scale production, Iontas scientists have developed a method to convert whole populations of antibodies from scFv format to IgG or Fab format, while maintaining the link between the variable heavy and light chains in a single cloning step. The company has methods for expressing these antibodies in 96-well format, which allows screening as IgG or Fab at a very early stage of the screening process, in high throughput, thus shortening the timelines of antibody drug discovery.
Conventionally, variable light and heavy chains have been cloned separately through a laborious process, which limits clone screening to a small number of IgGs. But as Dr. Dyson notes, “Properties such as affinity and activity in cell-based functional assays can change for single clones after conversion from scFv to IgG format. It is therefore critical to perform this conversion early in a discovery project, and screen as many clones in IgG format as possible.”
In addition, Iontas has developed a new antibody discovery platform where antibodies are displayed on the surface of mammalian cells in IgG format instead of scFv display on filamentous bacteriophages. Here, the company uses flow cytometry to identify high-affinity antibodies that display superior manufacturability.
Novel Cell Line
Although recombinant, protein-based pharmaceuticals are often produced in nonhuman mammalian cells, employing human cells could provide proteins with more authentic post-translational modifications.
“Nonhuman mammalian cells do not possess the ‘machinery’ required for human-type glycosylation such as specific glycosidases and glycosyltransferases,” says Jong-Mook Kim, Ph.D., director, cell science team, R&D division, Celltrion.
Dr. Kim and colleagues engineered a new human hybrid cell line they hope will provide an improved alternative for expression of therapeutic proteins. “We somatically fused two human cell lines, 293 cells and Namalwa cells, and isolated a specific clone, F2N78, that shows high-level, stable expression of recombinant proteins.”
The team chose 293 because it is a human embryonic kidney cell line often used for production of recombinant proteins and viral vectors, and Namalwa, a B-cell line derived from Burkitt’s lymphoma. The genome of Epstein-Barr virus (EBV) is inserted into the chromosomes of Namalwa.
“The fusion of these two cell types maintains the stability of the EBV genome but doesn’t produce virus. The EBV genome provides for expression of EBV specific nuclear antigen 1 (EBNA1) protein. The latter is an indispensable condition for the use of cells for the production of monoclonal antibodies and other therapeutics under transient conditions using oriP-based expression vectors.
“We found the constitutive production of EBNA1 and human-specific glycosylation enzymes [GnTIII and a(2,6)ST] was stable for more than one year under serum-free suspension cultures,” explained Dr. Kim. “Further, monoclonal antibody production at ~100 micrograms per milliliter was obtained six days after transfection and stable production of monoclonal antibody using an MTX amplification system was comparable to that of CHO cells.”
The company expects to use its F2N78 host cell line in the future for production of recombinant proteins and vaccines.
“Currently, we are continuing to improve the F2N78 cell line, which has already cleared adventitious viral agent and tumorigenicity issues, to generate a commercially feasible cell line for the production of recombinant proteins and vaccines,” noted Dr. Kim.