Upstream Processing:Vendor Notes: Generating Stable, High-Expressing Cell Lines for Recombinant Protein Manufacture

ABSTRACT

Manufacturing recombinant proteins at industrially relevant levels requires technologies that can engineer stable, high-expressing cell lines rapidly, reproducibly, and with relative ease. Commonly used methods incorporate transfection of mammalian cell lines with plasmid DNA containing the gene of interest. Identifying stable, high-expressing transfectants is normally laborious and time consuming. To improve this process, the ACE System has been developed based on pre-engineered artificial chromosomes with multiple recombination acceptor sites. This system allows targeted integration of single or multiple gene copies and eliminates the need for random integration into native host chromosomes. To illustrate the usefulness of the ACE System in generating stable, high-expressing cell lines, we present several case studies covering CHO cell lines expressing monoclonal antibodies.

Many methods are currently available for producing cell lines that express recombinant proteins. Most of them use plasmid transfection, or viral transduction procedures, to incorporate DNA sequences containing the gene of interest into mammalian cell lines. These processes often result in transfectants with highly variable protein expression due to random integration of the DNA into the host genome. Furthermore, these methods may necessitate time-consuming amplification events, or re-infection, to boost the cell's productivity. As a result, the process of generating and selecting a high-expressing, stable clonal cell line suitable for the clinical and commercial manufacture of biopharmaceuticals can be labor intensive and extremely time consuming.

To increase the speed and efficiency of generating high-expressing, stable cell lines for the manufacture of recombinant proteins, Chromos has developed a novel cell-line engineering system, the ACE System, based on the company's proprietary artificial chromosome technology. The system differs from conventional technologies. It facilitates the targeted and reproducible integration of multiple copies of genes into specific sites on an artificial chromosome that resides in a production cell line without amplification. These artificial chromosomes, or ACEs, contain fully functional centromeres and telomeres, and as a result are as mitotically stable as the host chromosomes. Ultimately, this results in the generation—with minimal screening and reduced timelines—of high-expressing stable clonal cell lines with high levels of gene expression. ACEs also can be purified to homogeneity by flow cytometry and readily transferred to a variety of cell types using commercially available transfection agents. This feature enables the auditioning of alternative cell lines for improved product quality or quantity, thereby providing an option not typically found in conventional mammalian cell line engineering technologies.

THE ACE SYSTEM

Core Components

The ACE System was designed so that one or more genes could be reliably and reproducibly loaded, with relative ease, onto an existing ACE and screened for incorporation and expression. For mammalian cell line engineering, the ACE System consists of four main components (Figure 1):

Figure 1

1. Platform ACE: A pre-engineered artificial chromosome containing 50–70 recombination acceptor sites, which allows the insertion of multiple copies of DNA sequences.

2. Platform ACE Cell Line: A manufacturing quality cell line (CHOK1SV, Lonza) containing the Platform ACE that grows to high cell density under serum-free growth conditions.

3. ACE Targeting Vector (ATV): A plasmid that contains a single recombination donor site for recombination into the acceptor sites on the Platform ACE, selection marker, and the gene(s) of interest along with all genetic elements required for enhanced expression in CHO cells.

4. ACE Integrase: A site-specific DNA recombinase that catalyzes the targeted integration of the ATV onto the Platform ACE residing in the Platform ACE Cell Line.

Targeted Integration

Each recombination acceptor cassette on the Platform ACE consists of a lambda phage attP site flanked by a simian virus 40 (SV40) promoter at the 5' position and an open reading frame sequence encoding the puromycin resistance (puromycin) gene at the 3' position. This confers puromycin resistance to cells carrying the Platform ACE (Figure 1A). All ATVs (Figure 1C) encode a bacterial attB site upstream of a promoterless, secondary drug-selectable marker gene (e.g., zeocin, blasticidin, neomycin, or hygromycin), which becomes activated by the SV40 promoter when the ATV integrates correctly via recombination between the attB site on the ATV and an attP site residing on the Platform ACE (Figure 1D). The ATV also contains the target gene cassette, which consists of the gene(s) of interest flanked by insulators and a 5' upstream CX promoter (chicken Beta-actin promoter and CMV immediate/early enhancer). Multiple copies of the same gene or multiple genes (e.g., heavy and light chains of an antibody) can be placed into the ATV and loaded on to the Platform ACE. The site-specific recombination is mediated by the ACE Integrase, a proprietary version of the lambda phage integrase that has been genetically engineered to function in mammalian cells without bacterial cofactors (Figure 1), resulting in the generation of two new sites, attR and attL. The ACE Integrase reaction is unidirectional and catalyzes only the integration of the ATV onto the Platform ACE because it lacks the bacterial cofactors required for excision. The combination of multiple attP sites on the Platform ACE and the unidirectional ACE Integrase enables multiple loadings (during a single transfection) or sequential loadings (via multiple transfections) with ATVs. Moreover, the ATV itself has a considerable carrying capacity and has been able to carry payloads exceeding 1.25 Mbp.

The ACE System targeted integration increases the efficiency of screening; only cells in which the ATV has correctly integrated into the Platform ACE are selected. Because multiple gene copies are inserted in a single round of loading into a consistent molecular environment, very few colonies (100–200) have to be screened to identify the high-expressing clones.

THE ACE SYSTEM PROCESS

Transfection, or loading of the Platform ACE, is accomplished by cotransfecting the Platform ACE Cell Line with both the ATV and ACE Integrase plasmid, using standard transfection methods. The ACE Integrase is only transiently expressed and not incorporated into the host chromosome. To generate high-expressing, stable, clonal cell lines, a "single-load" or "double-load" process can be used, depending on the required expression level and generation time (Figure 2). ATV construction can take between one and two months, depending on the target gene DNA source. Single-load candidate cell lines can be generated in three to four months from ATV transfection into Platform ACE Cell Line, using standard transfection reagents and methods and employing simple screening methodologies to determine expression and protein yield. Briefly, loading of ATVs onto the Platform ACE in the Platform ACE Cell Line is accomplished by a lipid-mediated cotransfection of ATV and expression plasmid containing the coding region for ACE Integrase under adherent conditions. After transfection of the Platform ACE Cell Line, drug-resistant colonies are switched to a basic serum-free media, selected, and expanded from 96-well and 24-well cultures to shake flask, and serially screened for growth characteristics and productivity by ELISA or HPLC Protein A chromatography. This process takes between one-and-a-half and two months and results in the generation of individual pools of cells (primary transfectants) that can be used to produce material for research programs (Figure 2). Selected primary transfectants are subsequently single-cell subcloned by limiting dilution, expanded to shake flask, and subjected to performance testing in terminal shake-flask cultures. This requires an additional one-and-a-half to two months to complete. Candidate clonal cell lines are selected, based on growth, yields, and stability of expression and take between three and four months to generate from ATV transfection.

Figure 2. The ACE System process overview. Single-load and double-load maximum batch titers with candidate cell-line generation times from transfection of the Platform ACE Cell Line.

Because not all recombination acceptor sites on the Platform ACE are targeted in a single-load process, a second transfection, or double load, can be performed if the user desires higher levels of expression. In the double-load process, single-load primary transfectants are loaded with a second ATV containing the gene(s) of interest and an alternate drug resistant marker. As with the single-load process, individualized pools of drug-resistant transfectants are generated, screened, and single-cell subcloned to identify clonal lines for performance testing in terminal shake-flask cultures. Although the double-load process requires an extra one-and-a-half to two months to generate clonal candidate cell lines, compared with the single load process, it has routinely resulted in titers that are >50 percent higher than those obtained with a single load.

The ACE System has been used to generate a number of CHO cell lines expressing a variety of different recombinant proteins. For monoclonal antibodies, the single-load process routinely produces clonal cell lines with yields of 300 to 700 mg/L in non-fed (unsupplemented CD-CHO medium from Invitrogen), batch terminal, shake-flask cultures, and 500 to 1,000 mg/L for the double-load process. Several cell lines have been subject to growth optimization and scale-up, in which 2–5 fold gains in performance have been noted.

CASE STUDIES

Performance of Candidate Cell Lines After Single-Load and Double-Load Processes

This case study focused on comparing the expression of candidate cell lines under non-fed, batch shake-flask conditions generated by both the single and double load processes. The Platform ACE Cell Line was transfected with an ATV containing the sequences for the heavy and light chains of an IgG1 monoclonal antibody flanked by cHS4 insulator sequences and the hygromycin resistance gene. After two transfections, 350 drug resistant colonies were generated at the 96-well stage and screened for antibody expression by ELISA. Selected primary transfectant colonies with titers >1 μg/mL were expanded to 24-well plates and subjected to terminal over growth assays, where titers were measured by ELISA. The top 20 primary transfectants, as determined by the 24-well overgrowth assay, were expanded to shake flask cultures for assessment of protein yield and growth characteristics. This process took six weeks to complete and resulted in primary transfectants with titers of ~450 mg/L under terminal, non-fed, shake-flask culture (unsupplemented CD-CHO medium from Invitrogen). These primary transfectants were used to generate approximately 5 g of research material in a 10-L batch Wave bag bioreactor.

The top primary transfectants were single-cell subcloned by limiting dilution, selected, and expanded to shake-flask cultures, and subjected to performance testing in terminal non-fed, shake-flask cultures. The growth profile and corresponding fluorescent in situ hybridization (FISH) image from the top single-load subcloned cell line are shown in Figure 3A. Cells reached a viable cell density of 7.0 x 10⁶ cells/mL, with titers of 740 mg/L and a specific productivity of 14 pg/cell/day. The corresponding FISH images demonstrated that the ACE was intact and contained the antibody heavy and light chain sequences. Moreover, no integration of the antibody sequences onto the host CHO genome was detected.

Figure 3

The single-load primary transfectants, in addition to being subcloned, were also subjected to a double load with a second ATV containing the heavy and light chain sequences and an alternate secondary drug resistance marker (zeocin). This double-load process resulted in primary transfectants with titers reaching 833 mg/L under terminal, non-fed, shake flask cultures. As with the single-load process, the top double-load primary transfectants were single-cell subcloned, resulting in a number of clonal cell lines reaching yields of approximately 1,000 mg/L in terminal, non-fed shake flask cultures. The growth profile and corresponding FISH image from a top, double-load subcloned cell line are shown in Figure 3B. Cells reached a viable cell density of 7.0 x 10⁶ cells/mL, with titers of 1,140 mg/L and a specific productivity of 45 pg/cell/day. This increase in expression is believed to be from an increase in gene copies resulting from the second round of transfection. Once again, FISH analysis shows an intact ACE loaded with the antibody heavy and light chain sequences with none integrated on the host genome.

The top candidate, double-load, single-cell subcloned cell line was subjected to process development and fed-batch by supplementing the basal CD-CHO medium with glucose and plant hydrolysates. Compared with batch conditions, feeding resulted in a threefold improvement in expression (3 g/L) for a fed-batch 250-mL shake flask and a fourfold improvement in expression (>4 g/L) for a fed-batch 1-L bioreactor (Table 1).

Table 1

Under the ACE System Process, candidate cell lines are also subjected to a stability study as part of their performance testing. The final candidate cell lines are selected, based on their stability over a minimum number of generations, as well as their overall expression. Stability studies consist of maintaining cultures of the candidate cell lines in 125-mL shake flasks and passaging them twice weekly to ~3 x 10⁵ cell/mL. At specific times throughout the study, the cells are subjected to full productivity analysis in 250-mL shake flasks to determine growth profiles, maximum titers, and specific productivities. The stability study showed that single-load and double-load candidate cell lines were stable for over 60 generations. In addition, it was shown that these candidate cell lines were stable for over 96 generations when subjected to fed-batch conditions in 250-mL shake flasks.

Rapid Generation of Candidate Cell Lines

In this case study, the goal was to generate as quickly as possible a cell line expressing a monoclonal antibody to support pharmacological and toxicological testing, as well as early clinical development. It was estimated that a cell line expressing greater than 500 mg/L would be sufficient for these purposes. A single-load process was selected with minimal clone screening and an emphasis on generating a stable cell line in less than four months from transfection. Briefly, the Platform ACE Cell Line was targeted with an ATV containing the sequences for the heavy and light chain of an IgG4 monoclonal antibody flanked by cHS4 insulator sequences and the hygromycin resistance gene. After transfection, only 50 drug-resistant colonies were taken forward to 96-well plates and only 10 primary transfectants expanded to 24-well plates. Finally, two primary transfectants whose antibody titers were greater than 200 mg/L were selected for single-cell subcloning by limiting dilution. The resulting candidate cell line had an antibody titer of approximately 430 mg/L and was generated in under three months from transfection. This candidate cell line was shown to be stable for over 50 generations and that antibiotic selection was not required to maintain their stability. This candidate cell line was then subjected to scale-up and simple fed-batch by supplementing the basal CD-CHO medium with glucose and plant hydrolysates. Antibody titers were doubled (Table 2) in the 1.6-L fed-batch bioreactor with a significant increase in specific productivity and culture time. The antibody titer fell slightly to 660 mg/L when scaled up to the 15-L and 500-L fed-batch bioreactors. After purification, over 140 g of purified antibody was recovered from the 500-L fed-batch bioreactor and used to support the pharmacological and toxicological studies.

Table 2

Auditioning of Cells for Expression

Among the key features of the ACE system is the ability to purify ACEs (loaded with the gene[s] of interest or unloaded) using flow sorting and transfer of ACEs to a variety of mammalian cells, primary cells, or cell lines^{1, 2, 3}. This feature enables the quick use of loaded ACEs to audition host cells for improved quality of the product (e.g., desired glycosylation pattern) or enhanced quantity (e.g., improved growth or expression). To demonstrate this feature, a DG44 CHO Platform ACE Cell Line was loaded with monoclonal antibody heavy and light chains² and candidate cell lines were generated according to the single load process. The resultant candidate cell line had a specific productivity of 12 pg/cell/day. The loaded ACE was then isolated by flow cytometry and transferred into the parental DG44 line (without the Platform ACE). The resulting cell lines containing the transferred ACE had comparable levels of antibody expression with an average specific productivity of 11.4 pg/cell/day. This demonstrated that a loaded ACE could successfully be transferred from one cell line to another without any loss of antibody expression. Furthermore, upon transfer of the same loaded ACE to a CHOK1SV cell line, the average specific productivity increased to 55 pg/cell/day. In both transfers, FISH analyses revealed intact single ACEs, demonstrating that the specific productivity differences were due to the host cell. Beyond demonstrating the difference in expression in CHO lines for this monoclonal antibody, it illustrated how a single ACE can be used to audition cell lines for a desired trait without having to separately engineer both lines from the beginning of an engineering process.

Multiple Gene Expression—Potential for Metabolic Engineering

Another unique feature of the ACE System is its ability to integrate several copies of different genes onto the same ACE (essentially a double load with an ATV containing a different gene or genes). To highlight this novel feature,⁴ a study was conducted in which the DG44 CHO Platform ACE Cell Line was loaded with both green fluorescent protein (GFP) and human erythropoietin (EPO). The Platform ACE was initially loaded with an ATV encoding GFP, resulting in a cell line expressing GFP at levels detectable by fluorescence microscopy and flow cytometry. This GFP-expressing cell line was then loaded with a second ATV encoding EPO. The resulting cell lines maintained parental GFP expression levels and also expressed EPO at levels greater than 400 IU/10⁶ cells/day as measured by ELISA. The ability of the ACE System to incorporate and stably express different genes by sequential loading provides the basis for potential metabolic engineering applications. For instance, genes encoding growth factors, anti-apoptotic factors, or factors affecting post-translational modifications or protein secretion could be sequentially loaded onto an ACE expressing a product gene, thereby enhancing the growth characteristics of that cell line for clinical and commercial manufacture in a bioreactor. Alternately, a Platform ACE could be loaded with multiple copies of metabolic factors, sorted and purified, and transferred into a pre-existing production cell line. Cell lines containing multiple ACEs have been generated and preliminary data have demonstrated that both ACEs are quite stable. Metabolic engineering studies like those discussed above are currently ongoing.

CONCLUSIONS

The ACE System is a potent biological engineering system that is being applied to the engineering of mammalian cells for the clinical and commercial manufacture of biopharmaceuticals. Using the ACE System, stable, high-expressing clonal cell lines can be generated in three to six months with minimal screening. Depending on the process selected, candidate cell lines are generated with antibody titers in terminal, non-fed, shake-flask cultures of 300 to 1,000 mg/L and specific productivities in the range 20 to 40 pg/cell/day. These candidate cell lines had a stable gene expression for over 90 generations and were amenable to media and growth optimization and scale-up, to which 2–5 fold gains in yields have been noted. These results show that ACE System candidate cell lines can be generated faster and perform as well, if not better, than those generated by existing technologies. The ACE System allows the downstream genetic engineering of cell lines and the auditioning of alternative cell lines to be performed more readily and routinely with minimal effort. It is clearly an alternative to conventional methods of cell line generation

Malcolm L. Kennard is director of cell line engineering at Chromos, Burnaby, British Columbia, Canada, 604.415.7100, info@chromos.com Danika L. Goosney is an associate for research planning and reporting at the Canadian Institutes of Health Research; and Harry C. Ledebur, Jr., is vice president of scientific affairs at Chromos.

REFERENCES

1. De Jong G, Telenius AH, Telenius H, Perez C, Drayer J, and Hadlaczky G. Mammalian artificial chromosome pilot production facility: large-scale isolation of functional satellite DNA-based artificial chromosomes. Cytometry 1999;35:129-33.

2. Lindenbaum M, Perkins E, Csonka E, Fleming E, Garcia L, Greene A. A mammalian artificial chromosome engineering system (ACE System) applicable to biopharmaceutical protein production, transgenesis and gene-based cell therapy. Nucleic Acids Res. 2004;32:21.

3. Perez CF, Vanderbyl SL, Mills KA, and Ledebur HC. The ACE System: A versatile chromosome engineering technology with applications for gene-based cell therapy. Bioprocessing J. 2004 July/August;61-8.

4. Vanderbyl S, Sullenbarger B, White N, Perez CF, MacDonald GN, Stodola T. Transgene expression after stable transfer of a mammalian artificial chromosome into human hematopoietic cells. Experimental Hematology 2005;2005;33:1470–6.