N-1 Perfusion Platform Development Using a Capacitance Probe for Biomanufacturing

Abstract

Fed-batch process intensification with a significantly shorter culture duration or higher titer for monoclonal antibody (mAb) production by Chinese hamster ovary (CHO) cells can be achieved by implementing perfusion operation at the N-1 stage for biomanufacturing. N-1 perfusion seed with much higher final viable cell density (VCD) than a conventional N-1 batch seed can be used to significantly increase the inoculation VCD for the subsequent fed-batch production (referred as N stage), which results in a shorter cell growth phase, higher peak VCD, or higher titer. In this report, we incorporated a process analytical technology (PAT) tool into our N-1 perfusion platform, using an in-line capacitance probe to automatically adjust the perfusion rate based on real-time VCD measurements. The capacitance measurements correlated linearly with the offline VCD at all cell densities tested (i.e., up to 130 × 10⁶ cells/mL). Online control of the perfusion rate via the cell-specific perfusion rate (CSPR) decreased media usage by approximately 25% when compared with a platform volume-specific perfusion rate approach and did not lead to any detrimental effects on cell growth. This PAT tool was applied to six mAbs, and a platform CSPR of 0.04 nL/cell/day was selected, which enabled rapid growth and maintenance of high viabilities for four of six cell lines. In addition, small-scale capacitance data were used in the scaling-up of N-1 perfusion processes in the pilot plant and in the GMP manufacturing suite. Implementing a platform approach based on capacitance measurements to control perfusion rates led to efficient process development of perfusion N-1 for supporting high-density CHO cell cultures for the fed-batch process intensification.

Keywords: Chinese hamster ovary (CHO); capacitance; cell-specific perfusion rate (CSPR); in-line probe; monoclonal antibody (mAb); perfusion N-1; platform; process analytical technologies (PAT); process intensification; scale-up.

Figure 1

Depiction of online capacitance probe perfusion N-1 control strategy. (A) Capacitance probe. (B) Capacitance/DeltaV controller, perfusion rate calculation, D_perf, is based on CSPR and VCD. (C) Perfusion pump, drives perfusion rate, D_perf. (D) Media inlet pump, controlled by bioreactor scale. (E) Bioreactor scale. (F) ATF controller, controls ATF cycle rate.

Figure 2

Correlation of measured permittivity with offline measured VCD. Individual measurements for mAbs 1–6 (A–F) aggregated from 5 L runs are plotted with simple linear regression and 95% confidence intervals. Data points were aggregated from the following number of 5 L runs for mAbs 1–6 are as follows: 8 (A), 25 (B), 11 (C), 22 (D), 26 (E), 14 (F), respectively.

Figure 3

VCD and viability and cumulative media perfused for volumetric (volume-specific exchange rate) and online CSPR controlled N-1 for mAb1 and mAb5. Performance comparison for two mAbs using either a volume specific exchange rate or online controlled CSPR. For mAb1, VCD and viability are plotted in (A) and cumulative media is plotted in (B). For mAb 5, VCD and viability are plotted in (C) and cumulative media is plotted in (D). For mAb1, n = 1; for mAb5, n = 3. In panel (C,D), data are plotted as average ± standard deviation.

Figure 4

CSPR process development experiment. Comparison of different CSPRs in two mAbs are shown. For mAb3, cumulative media usage for each CSPR tested is plotted in (A), and VCD and viability are plotted in (B) (n = 1 for each value of CSPR). For mAb5, cumulative media usage for each CSPR tested is plotted in (C), and VCD and viability are plotted in (D) (n = 1).

Figure 5

Inoculation VCD process development experiment. Comparison of performance for each inoculation density tested for mAb4 (n = 1 for each inoculation density). VCD (A), cumulative media perfused in liters (B), viability (C), and doubling time (hours) (D) are plotted.

Figure 6

Scale up performance of N-1 perfusion bioreactors controlled with the capacitance probe. Comparison of cell culture performance at different bioreactor scales. VCD and viability are plotted for mAb4 ((A); 5 L: n = 4, 200 L: n = 4, 500 L: n = 3) and mAb5 ((B); 5 L: n = 5, 200 L: n = 1, 500 L: n = 3). Data are plotted as average ± standard deviation.

Figure 7

Optimized offline VSER (volume specific exchange rate) based on capacitance developed N-1 process. Comparison of online capacitance probe-controlled perfusion rate versus an optimized, offline volume-specific exchange rate developed based on historic media usage data from online capacitance probe-controlled studies for three mAbs. VCD and viability are plotted for each mAb (A,C,E), and cumulative media perfused for each mAb (B,D,F). n = 1 for each condition.