Category: Manufacturing and Bioprocessing
Purpose: In this work, we demonstrate a process that accelerates the serum-free/suspension adaptation for mammalian cells using medium optimization by DOE (design of experiment), instead of repetitive subculturing.
Methods: Chinese Hamster Ovarian (CHO) DP-12 cell line (Clone #1934, CRL-12445 from ATCC) was used as model cell. This anchorage-dependent cell was cultured in TC-treated flask at 5% CO2 and 37 ◦C. The complete medium (CM) was Dulbecco’s Modified Essential Medium (DMEM) supplemented with Fetal Bovine Serum (10%), Insulin (2 µg/mL), Cellgro Trace Element A and B (1x), and Methotrexate (200nM). The basal serum free medium (SFM) was CD CHO chemical defined medium. It was supplemented with GlutaMAXTM (8 mM) and anti-clumping agent (0.4%, v/v), along with Methotrexate (200nM). The adaptation study was performed by the gradual serum reduction approach. This approach cultures cells in mixtures of CM and SFM at decreasing CM ratios (75:25, 50:50, 25:75, 10:90 and 0:100). After cells were adapted at a given level, they were cultured at the next level.
The Ca2+ and Mg2+ concentrations were adjusted and optimized based on the cell doubling time at the adherence stage (CM:SFM = 75:25 and 50:50). The doubling time (td) at each Ca2+/Mg2+ level was compared using t-test, with p< 0.05 as significance level. Cell density and viability were obtained by Countess Cell Counter using trypan blue exclusion method.
A definitive screening design (DSD) was conducted for the adaptation from 10% CM to serum free (0%). The design had 15 runs in total with 3 replicate center points. Six factors and three levels were included (DMEM, insulin, calcium, magnesium, Trace Element A and B). Main effects and interaction terms were included in the linear regression models of viable cell density (VCD) and viability. T-test was used to evaluate significance of each term; and ANOVA was used to evaluate the significance of model explained variance.
Results: At high Ca2+/Mg2+ level doubling time was significantly reduced from 72.5 and 89.1 hours, to 23.35 and 29.84 hours respectively at 75% and 50% CM (p< 0.05). This reduced doubling time decreased the total time required for adherence-culture adaptation process to six days.
The CHO cells were cultured in suspension at the 25% and 10% CM. Cells failed to grow when the CM percentage was reduced below 10%. A definitive screening design was conducted to select the non-serum essential nutrients for suspension adaptation.
Significant models for VCD (p < 0.01, and an adjusted R2 of 0.87) and viability (p < 0.01, and an adjusted R2 of 0.81) were calculated. The VCD model is illustrated in Figure 2. High levels of Cellgro Trace Element A (0.2X) and DMEM (20%) significantly improved the VCD and viability; while the Cellgro Trace Element B significantly impeded cell growth and viability. Ca2+ concentration had significant interaction with Trace Element A and DMEM, supporting cell growth and viability. Investigated insulin concentration (0 to 0.4µg/mL) and magnesium ion (0.81 to 1.42mM) concentration were observed to have no significant effects (p >0.05) in either model.
Conclusion: With the medium composition optimization, the adaptation process was accelerated. In total, the adaptation process took as little time as 25 days, in contrast with previously reported 66 days  (Figure 3).
1. Gowtham, Y.K., Understanding the Transcriptional Landscape of CHO Cell Lines using Next Generation Sequencing Technology RNA-seq, under industrially Relevant Conditions. PhD Thesis, Clemson University, 2016.
Adam Rish– Pittsburgh, Pennsylvania
Yuxiang Zhao– Graduate Student, Duquesne University, Pittsburgh, Pennsylvania
James Drennen– Assoc. Dean Research and Graduate Programs, Duquesne University, Pittsburgh, Pennsylvania
Carl Anderson– Pittsburgh, Pennsylvania