Shan Liu, PhD | Cell Line Development Scientist | Molecular Devices
Rebecca Kreipke, PhD | Field Applications Scientist | Molecular Devices
The market for biotherapeutics and therapeutic proteins was valued at $93.14 billion in 2018 and is expected to nearly double by the end of 20221. Yet, the selection of high-producing mammalian CHO cell lines continues to represent a major bottleneck in process development for the production of biopharmaceuticals2. Therefore, it is increasingly important to develop new high-throughput methods for the selection of high-expressing CHO cell lines in an efficient and cost-effective manner. Traditional methods using multiple rounds of methotrexate (MTX ) or methionine sulfoximine (MSX) induced gene amplification and limiting dilution are inefficient, time-consuming, labor-intensive, and do not always yield clones with the desired production level. FACS-based screening methods can increase efficiency and throughput, but require high capital investment, intensive operator training, and dedicated personnel. FACS workflows are vulnerable to cross-contamination due to non-disposable components that have direct contact with cells. Furthermore, for cell lines sensitive to fluidics pressure, cells may show lower viability and not recover well after FACs screening.
In this study, the CloneSelect® Single-Cell Printer f. sight™ and a CHO cell line with a fluorescent reporter were used for rapid, early stage identification of clones producing high levels of a therapeutic protein, recombinant human Interleukin 15 (tr-hIL15). The genes encoding the tr-hIL15 and the reporter protein were linked by an internal ribosome entry site (IRES), so that they were transcribed in the same mRNA but were translated independently. Since they came from a same mRNA molecule, the reporter protein’s expression level could be used to predict the relative expression level of the therapeutic protein for each clone.
Materials and methods
Construction of recombinant DNA plasmids
To co-express recombinant tr-hIL15 protein with an enhanced green fluorescent protein (EGFP), the corresponding tr-hIL15 open reading frame was inserted into a mammalian expression vector downstream of Cytomegalovirus (CMV) promoter and upstream of the IRES, which was followed by an EGFP open reading frame and SV40 polyA (Figure 1). To decrease the translation initiation efficiency of the IRES sequence, ATG11 and ATG12 were deleted3–4. The completed recombinant plasmid contained a zeocin resistance gene expression cassette for selection of stable CHO cell lines. All gene synthesis, subsequent cloning, and sequencing were performed by GenScript (Piscataway, NJ).
Cell culture and transfection
Unless otherwise noted, all reagents and labware were from Thermo Fisher Scientific such as the FreeStyle™ CHO-S cell line. Cells were grown in suspension using 125 mL vented cap shake flask at 37°C, 5% CO2, 70–80% humidity, and shaking at 125 rpm. Growth media used was XP CHO Growth A Medium (Molecular Devices), supplemented with 4 mM L-glutamine. Cells were transfected by electroporation using a Neon Transfection System (Thermo Fisher Scientific) with linearized tr-IL15-IRES-EGFP plasmids and controls. ClonaCell™-CHO ACF supplement (STEMCELL) was added to the growth media at 3% concentration to help the freshly transfected cells recover. After a recovery period of 48–72h, transfected cells were reseeded at a viable cell density of 3–5x105 cells/mL in XP CHO Growth A selection medium supplemented with 4 mM L-glutamine and 100 ug/mL Zeocin™ reagent. Stable pools were subsequently subjected to 200 ug/mL and 300 ug/mL Zeocin selections
Fluorescence-aided sorting and clone screening
24–48 hours prior to f. sight sorting, stable pool cells were reseeded at 5x105 viable cells/mL in fresh XP CHO Growth A selection medium with 300 ug/mL Zeocin added. On the day of sorting, 60 µL of single-cell suspension with 1x106 viable cells/mL was screened by the CloneSelect SingleCell Printer f. sight. Cells sorted by their size, roundness, and mean fluorescence intensity were deposited at a density of one cell per well into standard 96-well plates (Corning 3300) prefilled with complete cloning media. The complete cloning media was Zeocin-free EX-CELL® CHO cloning medium (Sigma SAFC) supplemented with 4 mM L-glutamine and 2.5% CHO ACF supplement. All seeded wells were imaged on the CloneSelect® Imager (with transmitted light only) on day 0, 1, 2, 7, and 14 post-sort for verification of single-cell-derived clones. Confirmed single cell clones were sequentially expanded from 96-well plates to 24-well plates, 6-well plates, and shake flasks. Clones were maintained in Zeocin-free XP CHO Growth A media supplemented with 4 mM L-glutamine from this point onwards.
Analysis of protein production
For the cell growth and clone stability studies, cells were reseeded at 1x105 viable cells/mL every 3–4 days. For all other shake flask analysis, cells were seeded at 3x105 viable cells/mL in 20 mL of XP CHO Growth A media supplemented with 4 mM L-glutamine. Conditioned media were harvested at day three post-seeding. All media samples were centrifuged to remove cells and debris and then stored at -80°C. The tr-hIL15 protein titers were determined by ELISA (ab218266, Abcam), and total cell numbers were measured at the beginning of seeding and the final sample collection time point for the calculation of specific productivity rate (SPR). The SPR measured in picograms of specific protein per cell per day (pg cell-1day-1:pcd) counts both growth rate and productivity5, which provides a more comprehensive way to evaluate and compare cell lines.
Prediction of clonal therapeutic protein titers by the EGFP-based f. sight fluorescence-aided sorting
A reporter-based CloneSelect Single-Cell Printer f. sight selective cloning process was developed to isolate single cells producing high levels of therapeutic proteins. In this system, the reporter expression levels were used to predict the relative expression levels of the therapeutic protein on a per cell basis. Cells from stably transfected pools with different fluorescence intensities were seeded into 96-well plates to obtain single cell clones and therapeutic protein secreted from these clones were measured by ELISA three weeks post-seeding (Figure 2a). The mean fluorescence intensity of each cell seeded was quantified in relative fluorescence units (RFU) by the CloneSelect Single-Cell Printer f. sight. Cells were divided into four groups (A, B, C, D) based on their RFU level (EGFP expression level). There is a 2.9–4.4 fold of increase in average level of relative tr- hIL15 protein expression observed in group B–D with medium-high EGFP expression, comparing to that of group A with low EGFP expression. That suggests by selectively isolating cells with higher fluorescence intensity (medium-high EGFP expression), we can eliminate unwanted cells with low levels of expression at single cell-sorting phase, and prioritize resources on cells with higher protein expression levels at early cloning stage.
To assess the correlation between cells’ protein production in 96-well plate (early stage) and their titers in shake flasks (late stage), we expanded 12 clones from group C into shake flasks. Conditioned media from these shake flasks were harvested at day three post-seeding with a seeding density of 3x105 viable cells per mL and the tr-hIL15 protein expression was quantified by ELISA (Figure 2b). The 21-day protein production of single cell clones in 96-well plate correlated well (r= 0.89) with 3-day protein production of these clones in 125 mL shake flasks. However, we were unable to identify the same pattern from cells classified in group D (data not shown). Cells from group D with the highest EGFP expression and the highest average protein expression level didn’t scale up well. Four out of the twelve clones died during the expansion process.
Assessment of clone stability
Assessment of clone stability is a crucial step for cell line development when developing therapeutic proteins. The ICH Q5D guidelines6 require biopharmaceutical manufacturers to conduct stability trials to ensure consistent product quality, based on an assumption that a consistent cell should yield a consistent product. The average timeline for cell expansion from cell bank to production bioreactor for a production run is about 60 generations, so the clone stability trial is typically designed to run for the same amount of time. Here, we picked clone 2G11, one of the clones with highest titers in shake flask study for the stability test. In Figure 3a, clone 2G11 showed stable cell growth during continuous passaging. In Figure 3b, SPRs were repeatedly measured, resulting from 0.201 to 0.157 (pg cell-1day-1) with a loss of -0.044 (-22%) during the trial. Typically, clones with productivity loss (measured by SPR here) less than 30% during trials would be considered stable. As revealed in Figure 3a-b, clone 2G11 showed a desired profile with stable cell growth and stable protein production, which suggests this clone could be a good candidate to be brought forward for further characterizations.
This study established an effective process for generating recombinant CHO cell lines producing high levels of therapeutic proteins. Additionally, the ability to rapidly and selectively screen clones in a 96-well plate allows for the elimination of unwanted clones at an early cloning stage, which increases the efficiency of the process. After selective cloning, the average specific product rate of selected clones with high EGFP expression was improved to 0.174 pg cell-1day-1, which is about 5.5 fold higher compared to that of the unsorted parent stable pools (0.032 pg cell-1day-1)
Furthermore, because this method does not rely on the availability of an antibody specific for the therapeutic protein or reporter protein being expressed, it can be easily implemented into any cell line development process.
- Browne SM, Al-Rubeai M. Selection methods for high-producing mammalian cell lines. Trends Biotechnol. 2007 Sep;25(9):425–32.
- CloneTech Laboratories, Inc.; pIRES Vector Information. Protocol No. PT3266-5, Version No. PR59976, 2005
- Davies MV, Kaufman RJ. The sequence context of the initiation codon in the encephalomyocarditis virus leader modulates efficiency of internal translation initiation. J Virol.1992 Apr;66(4):1924–32.
- Brezinsky SCG, Chiang GG, Szilvasi A, et al. A simple method for enriching populations of transfected CHO cells for cells of higher specific productivity. J Immuno Meth. 2003 Jan;277:141–155
- ICH Q5D Derivation and characterization of cell substrates used for production of biotechnological/biological products