Utilizing Stable Gene-Edited Knockout Pools for Genetic Screening and Engineering in Chinese Hamster Ovary Cells

Abstract

Chinese hamster ovary (CHO) cells are the primary host for biopharmaceutical production. To meet increasing demands for productivity, quality, and complex molecule expression, genetic engineering, particularly clustered regularly interspaced short palindromic repeats (CRISPR)-mediated gene knockout (KO), is widely used to optimize host cell performance. However, systematic screening of KO targets remains challenging due to the labor-intensive process of generating and evaluating individual clones. In this study, we present a robust, high-throughput CRISPR workflow using stable KO pools in CHO cells. These pools maintain genetic stability for over 6 weeks, including in multiplexed configurations targeting up to seven genes simultaneously. Compared to clonal approaches, KO pools reduce variability caused by clonal heterogeneity and better reflect the host cell population phenotype. We demonstrate the utility of this approach by reproducing the beneficial phenotypic effects of fibronectin 1 (FN1) KO, specifically prolonged culture duration and improved late-stage viability in fed-batch processes. This workflow enables efficient identification and evaluation of promising KO targets without the need to generate and test large numbers of clones. Overall, screening throughput is increased 2.5-fold and timelines are compressed from 9 to 5 weeks. This provides a scalable, efficient alternative to traditional clonal screening, accelerating discovery for CHO cell line engineering for biopharmaceutical development.

Keywords: CHO; clonal heterogeneity; genetic engineering; host cell line engineering; stable CHO pools.

FIGURE 1

Workflow for the generation of host cell protein KO clones and subsequent performance capture in a fed‐batch bioprocess. (Created in BioRender. Marzluf, J. (2025) https://BioRender.com/s43e103). Further improvements in timeline and data integrity can be achieved by bypassing the single cell cloning directly to pooled fed‐batch evaluation. This reduces the workflow by 5–6 weeks. Furthermore, KO‐scores in bulk pools may be increased by employing sequential transfections with the same RNP‐complex. KO, knockout; RNP, ribonucleoprotein.

FIGURE 2

Multiple transfections increase KO‐efficiency and correlate with phenotypic loss of FUT8. (A) InDel percentages of FUT8 KO pools generated using LE and HE sgRNAs across one, two, and three sequential transfections (TR1–TR3), performed at 48‐h intervals. [n = 6; mean ± SD]. (B) Correlation between FUT8 KO‐scores and percentage of FITC‐LCA‐negative cells (FUT8‐deficient phenotype) measured at Days 7 and 14 after the final transfection. Each point represents a KO pool collected at the indicated time point after completing one, two, or three transfections. [n = 6; mean ± SD]. FITC‐LCA, fluorescein isothiocyanate labeled lens culinaris agglutinin; FUT8, α1,6‐fucosyltransferase 8; HE, high efficiency; KO, knockout; LE, low efficiency; SD, standard deviation; TR, transfection.

FIGURE 3

Stability of single and multiplexed KO pools over multiple weeks. (A) KO‐scores in single gene KO pools after one, two, and three sequential RNP‐complex transfections. Each subsequent column pair (black + blue bars) represents two pools after the respective number of transfections in the first week (black bars) and after four weeks (blue bars). [n = 2; mean ± SD]. (B) KO‐score in one 7x KO pool after three multiplexed transfections in the first week and after six weeks. [n = 1]. KO, knockout; RNP, ribonucleoprotein; SD, standard deviation.

FIGURE 4

KO of FN1 increases final fed‐batch titer and prolongs the stationary phase in CHO clones and heterogenous KO pools. (A) Growth curves showing VCC and viability during shake‐flask fed‐batch with three different antibody coding genes. (n = 12 [FN1 KO clones], 3 [KO pools], 3 [WT pools]; mean ± SD). (B) KO scores along the cell line generation process until fed‐batch cultivation. (n = 12 [KO clones], [KO pools]; mean ± SD). (C) IVCC of all FN1 KO fed‐batch experiments. (n = 12 [FN1 KO clones], 3 [KO pools], 3 [WT pools]; mean ± SD). (D) Final titer fold change of all FN1 KO fed‐batch experiments. Titers were normalized against WT samples in their respective experiment and then plotted for the fold change observed. (n = 12 [FN1 KO clones], 3 [KO pools], 3 [WT pools]; mean ± SD). CHO, Chinese hamster ovary; FN1, fibronectin 1; IVCC, integral viable cell concentration; KO, knockout; SD, standard deviation; WT, wild‐type.

FIGURE 5

Empirical variance in fed‐batch bioprocess behavior comparing clones and pools. (A) Performance parameters peak viable cell concentration, final titer, and integral of viable cell concentration were normalized against WT samples in their respective experiment and then plotted for the fold change observed. (n = 175 [clone], 36 [pool]; mean ± SD). (B) Required number of replicates to resolve a 0.3‐fold mean difference between groups with 90% power in peak VCC, final titer, and IVCC. IVCC, increased integral viable cell concentration; SD, standard deviation; VCC, viable cell concentration; WT, wild‐type.