Accelerated generation of gene‐engineered monoclonal CHO cell lines using FluidFM nanoinjection and CRISPR/Cas9

Chinese hamster ovary (CHO) cells are the commonly used mammalian host system to manufacture recombinant proteins including monoclonal antibodies. However unfavorable non‐human glycoprofile displayed on CHO‐produced monoclonal antibodies have negative impacts on product quality, pharmacokinetics, and therapeutic efficiency. Glycoengineering such as genetic elimination of genes involved in glycosylation pathway in CHO cells is a viable solution but constrained due to longer timeline and laborious workflow. Here, in this proof‐of‐concept (PoC) study, we present a novel approach coined CellEDIT to engineer CHO cells by intranuclear delivery of the CRISPR components to single cells using the FluidFM technology. Co‐injection of CRISPR system targeting BAX, DHFR, and FUT8 directly into the nucleus of single cells, enabled us to generate triple knockout CHO‐K1 cell lines within a short time frame. The proposed technique assures the origin of monoclonality without the requirement of limiting dilution, cell sorting or positive selection. Furthermore, the approach is compatible to develop both single and multiple knockout clones (FUT8, BAX, and DHFR) in CHO cells. Further analyses on single and multiple knockout clones confirmed the targeted genetic disruption and altered protein expression. The knockout CHO‐K1 clones showed the persistence of gene editing during the subsequent passages, compatible with serum free chemically defined media and showed equivalent transgene expression like parental clone.

well studied system (> 35 years) and high tolerance for manufacturing variables in bioreactor (pH, pressure, temperature). [1,2](PMID: 34895638) However, the unfavorable non-human glycoprofile displayed on CHO-produced recombinant proteins, including monoclonal antibodies, have negative impacts on product quality, immunogenicity, pharmacokinetics, and therapeutic efficiency.The presence of glycosylation patterns such as fucose on antibody Fc regions can inhibit antibody-dependent cellular cytotoxicity (ADCC) and antibodydependent cell-mediated phagocytosis (ADCP). [3,4] earlier study has reported that a non-fucosylated version of Avelumab, an FDA-approved anti-PD-L1 antibody, displayed enhanced antitumor activity when compared with its fully fucosylated counterpart in vivo. [5]Therefore, creation of a CHO cell line devoid of genes involved in non-human glycosylation pathways is of very high commercial interest.Moreover, regulatory agencies require CHO cells used in commercial production to be monoclonal to ensure drug product homogeneity. [6]Standard methods to generate monoclonal cell lines requires longer timeline and comprise of laborious workflows. [6]Here, we present CellEDIT, an approach that combines CRISPR/Cas9 and the single-cell intranuclear nanoinjection capacity of FluidFM, to address some of the current shortcomings of current bulk transfection approaches.Nanoinjection of RNP complexes enables unprecedented control over the amount and stoichiometry of cargo delivered to single individual cells, thereby most likely reducing off-target editing and increasing editing efficiency when targeting multiple loci simultaneously.In addition, the force-controlled nature of FluidFM avoids high cell toxicity observed after electroporation and is, in contrast to lipofection, also efficient in so called hard-to-transfect cell lines.In a proof-of-concept study, we generated gene knockout (KO) CHO-K1 cell lines for three well-known targets that eliminate fucosylation (FUT8), increase survival (BAX), and transgene expression (DHFR) with the guarantee of monoclonality origin in an accelerated time frame (2-3 weeks).The proposed approach was amenable to generate a monoclonal cell line with multiple genetic alterations and equivalent levels of transgene expression.

Design of sgRNA
To generate a knockout clone for FUT8, BAX, and DHFR, the sgRNAs targeting the coding region were adapted either from earlier study [7] (FUT8 and BAX; Figure 1A

FluidFM nanoinjection to generate monoclonal knockout CHO-K1 cells
Detailed method is presented in supporting information.In brief, on day 0, single CHO-K1 cells were seeded to the center of a well and monoclonality was documented (Figure 2A).On Day 1, gRNA/Cas9 RNP complexes targeting FUT8, BAX, and DHFR and GFP mRNA (20 ng µl −1 ) was injected into the nucleus of single cells using the FluidFM Nanosyringe (Figure 2B).12.5% of the gRNA was labelled with Atto550 to monitor injection efficiency.Twenty-four hours post-injection, the cells were imaged using the FluidFM OMNIUM Platform for injection efficiency and survival by imaging for GFP fluorescence (Figure 2C).
The injected cells were incubated at 37 • C, 5% CO 2 (v/v) in humidified atmosphere and monitored for growth and clonal expansion by scanning the well using a Spark Multimode Microplate Reader (Figure 2D).After 14 days, the clones were collected, and biobanks were established (Figure 2E & 2F).In parallel the clones were analyzed by Sanger sequencing of PCR fragments amplified from the targeted loci.Effectiveness of editing was verified at protein level by western blot (WB).
Detailed description of editing analysis can be found in the supporting information.

RESULTS AND DISCUSSION
In this study, we selected three well established genetic targets (FUT8, BAX, and DHFR) for gene engineering using CRISPR/Cas9 system which have direct relevance to the improvement of CHO expression systems. [7,8]Functions of each target in CHO cells and advantages of generating a KO version of these targets were outlined in S2.The editing strategy and the sequences of the used gRNA are illustrated in In earlier reports, we have demonstrated that the CRISPR/Cas9 system is a powerful tool to induce desired gene modification in primary and curated cells compared to other gene editing technologies. [9]To into CHO-K1 cells and evaluated the frequency of indels by ICE analysis.Our data displayed a high level of gene editing where mean indel frequencies for the targeted loci were 94 ± 1.7% (FUT8), 96 ± 0.5% (BAX), and 93 ± 0.5% (DHFR; Figure 1C and S4).Since the sgRNAs were targeted on coding region, high levels of the knockout score were achieved for BAX (96%), DHFR (92%), and FUT8 (79%; Figure 1D).In the pooled CHO-K1 cells, CRISPR based gene editing showed an increased on-target efficacy for three independent targets (> 93%).
Next, we aim to generate monoclonal CHO-K1 cell line for single KO with FUT8, BAX, and DHFR or multiple KO with all the three specified targets.To ensure recombinant biologics homogeneity, regulatory authorities insist on employing CHO cells derived from a single clone for manufacturing. [6]Development of stable monoclonal cell lines are majorly carried out by laborious workflow such as limiting dilution, Clonepix, cell sorting, or positive selection.Moreover, the process requires a longer timeline (> 10 weeks), intensive clone screening and exhibit reduced success rates. [10,11]To overcome these limitations, we hypothesized that CRISPR/Cas9 combined with the FluidFM nanoinjection technology can be harnessed to accelerate the generation of gene-engineered monoclonal CHO cell lines. [12]The CellEDIT approach, starting from single cells, can deliver CRISPR components directly into the nucleus of single CHO-K1 cell (Figure 2).Proposed key benefits from this approach compared to traditional transfection methods include gentle delivery with unprecedented control of the amount and stoichiometry of RNP delivered to cells.This in turn will maximize on-target editing and enable multiplex editing.Furthermore, by starting with a single cell, the generated genome engineered cell lines are monoclonal by design.
As a proof-of-concept, we injected gRNA/Cas9 RNP complex targeting FUT8, BAX, and DHFR into single cells to generate single KO or multiple KO using the FluidFM technology.Monoclonality was verified by well scanning and visual confirmation (Figure 2A).Delivery success was monitored by detection of Atto550 labelled gRNA and GFP fluorescence immediately and 24 h after injection, respectively (Figure 2B and 2C).Expansion of single cells into colonies was monitored by well scanning (Figure 2D).Genetic analysis on Day-14 after injection revealed that by using the CellEDIT workflow, we obtained three clones with KO of all three targeted genes as well as 7, 2, and Next, we repeated the CHO Triple-KO experiment with three independent injection rounds and assessed the cell survival after 24 h.
Our data showed a survival rate of 81% +/−11% after 24 h indicating that the FluidFM nanoinjection per se is not overall toxic.We also observed that around 50% of the injected cells resulted in successful clones at day-7 with > 30% success rate with KO genotype (S8).We also assessed the industry relevance of our approach by exposing Flu- shows that genome engineering using FluidFM does not impair general cell characteristics and functionality of the edited CHO-K1 cells.Moreover, the CellEDIT technology could be expanded to other cell lines to manufacture other drug modalities such as vaccines. [13]

CONCLUSION
In conclusion, the present proof-of-concept study shows that genome engineering using a novel workflow based on FluidFM mediated intranuclear injection of CRISPR/Cas9 components, can efficiently generate monoclonal single and triple KO cell lines.The CellEDIT process provides the fundaments to generate monoclonal CHO-K1 cell lines with improved properties to produce protein-based therapeutics.
In future studies, we aim to test whether the gentleness and the precise nature of the CellEDIT workflow can be applied to insert transgenes coding for proteins of interest into defined genomic loci.Furthermore, we hypothesize that the proposed strategy can be applied to any mammalian cell (e.g.,: cancer cells, primary cells, induced pluripotent stem cells) to create cellular models for fundamental research, disease modeling, and drug discovery.Of note, the CellEDIT workflow is available as service.

FigureF I G U R E 2
Figure1Aand 1B and described in detail in the supporting information.

F I G U R E 3
generate a CRISPR/Cas9 mediated single KO of FUT8, BAX, and DHFR in CHO-K1 cell line, we first optimized the electroporation conditions using DsRed mRNA.The following electroporation settings were found optimal for transfecting CHO-K1 cells (1650 V, 10 ms, and three pulses) with high transfection efficiency and viability (> 90%; S3).To evaluate the ability of CRISPR/Cas9 to generate on-target gene editing on FUT8, BAX, and DHFR, we electroporated sgRNA/Cas9 RNP complex Development of monoclonal CHO-K1 cell line with FUT8, BAX, and DHFR knockout.(A).Representative images of Sanger sequencing alignment of the wild type control (Wt) and gene edited CHO-K1 cell line with single knockout (KO) to the genomic sequence of FUT8 (top), BAX (middle), and DHFR (low).(B) Images of Sanger sequencing alignment of the wild type control (Wt) and gene edited CHO-K1 cell line (ID: 0386945383) with multiple knockout (KO) to the genomic sequence of FUT8 (top), BAX (middle), and DHFR (low).(C) Western blot showing detection of FUT8 (left), BAX (center), and DHFR (right) for selected clones.GAPDH is used as a loading control.(D) DsRed expression analysis in parental wild type CHO-K1 and multiple KO CHO-K1 monoclonal cell line in comparison with untransfected wild type CHO-K1 cells.

1
clone(s) with deletion of BAX, DHFR, and FUT8 respectively.Of note, not all obtained clones were analyzed (S7).Editing outcomes were analyzed by alignment of Sanger traces to the genomic target site (Figure 3A, single KO and 3B 3KO experiments).Sanger trace files were subsequently deconvolved into individual alleles as described in the supporting information and results are presented in S6.Subsequently, we performed protein analysis by western blot for the selected clones to verify the KO at the protein level and found that majority of the clones showed coherence with genetic data Figure 3C.Of note, one BAX KO clones showed full expression (clone ID: 0338319599) while other clones showed monoallelic expression (one BAX single KO clone and one DHFR clone for multiple KO; Figure 3C).These data demonstrate the functionality of the edited cell lines generated by CellEDIT.Importantly, all the generated gene KO clones (single or multiple) demonstrated similar levels of cell growth, microscopical morphology like WT CHO-K1 cells except the endogenous target gene (FUT8, BAX, and DHFR) expression (data not shown).To ensure that the genome engineered cells were still suitable as expression hosts, 3x KO CHO-K1 monoclonal cell line were transfected with DsRed mRNA.Subsequent comparison of DsRed expression levels of the transfected cells by flowcytometry with parental wild type CHO-K1 revealed no observable difference Figure 3D.Last, the knockout CHO-K1 monoclonal cell lines developed by our approach were passaged 2-5 times and persistence of gene editing was confirmed by Sanger sequencing/ICE analysis (S7).
idFM generated FUT8 KO, BAX KO and Triple KO CHO-K1 cells to serum free chemically defined media (CDM) with step-wise reduction of serum status.Both parental CHO cells and gene-edited cells were unaffected by the serum-free condition and demonstrated the industrial suitability of our strategy.We also tested an expression level of CHO-WT, CHO-FUT8 KO, CHO-BAX KO, and CHO-Triple KO cells for mRNA transgene expression in CDM.We noticed both CHO-WT and all the three different KO cells created with nanoinjection displayed same level of expression (95%-98% of DsRed + cells; S9).Overall, this