Chinese hamster ovary (CHO) cells are widely used to produce various recombinant therapeutic proteins due to several advantageous attributes such as genetic manipulability, adaptability to serum-free suspension culture, scalability to large culture volumes, and human-compatible post-translational modifications. For the stable expression of therapeutic proteins, traditional CHO cell line development processes start with transfection of host cells with a transgene, which is then followed by pool selection or single-cell cloning, transgene amplification (optional), characterization of well-performing clones, and final clone selection (Hamaker and Lee, 2018, Tejwani et al., 2021). During the selection process, transgenes are randomly incorporated into the host cell genome (O'Brien et al., 2020). This phenomenon, random integration, can impact the transgene expression of individual cells arbitrarily based on the genomic environment of gene integration sites, such as epigenetic accessibility, transcription activity, and the stability of chromosomal locus (Grav et al., 2018, Lee et al., 2019, Lee et al., 2018, O'Brien et al., 2020). As a result, various culture phenotypes, such as cell growth, viability, productivity, and protein quality, may appear between different clones, often described as clonal variability (Grav et al., 2018, Lee et al., 2019). For this reason, regulatory authorities have been requiring cell lines to be derived from a single cell progenitor or clone (ICH Q5D; European Medicines Agency, 1998) to maintain a consistent manufacturing process in the commercial production of recombinant proteins (Bandyopadhyay et al., 2019, Evans et al., 2015, Grav et al., 2018).

Based on the observation that different clones exhibit sufficient randomness in terms of transgene incorporation into chromosomes, identifying the transgene integration sites of clones can be used to establish the cell’s identity, and thereby, to prove the monoclonality (Langsdorf et al., 2021, Robins et al., 1981). Several techniques have been developed to configure transgene integration sites, including fluorescence in situ hybridization (FISH) for single-cell analysis and next generation sequencing (NGS; Illumina sequencing, PacBio sequencing, Nanopore sequencing), a high-throughput method allowing massive DNA sequencing in parallel (Lee et al., 2018, Zhang et al., 2012, Goodwin et al., 2016, Ivancic et al., 2022). While these methods provide comprehensive information, they require high cost (over $1,000 per sample) and are also time-consuming (4-7 weeks) (Van Nimwegen et al., 2016, Meggendorfer et al., 2022). Given that multiple clones (at least top-10 clones) need to be analyzed to establish one final cell line, developing alternative methods with lower cost and short analysis time is desirable.

Polymerase chain reaction (PCR)-based genome walking techniques, which determine the DNA sequences of unknown genomic regions flanking known sequences (Langsdorf et al., 2021, Shapter and Waters, 2014, Hui et al., 1998), can be employed to identify transgene integration sites. While several PCR-based genome walking techniques, including Alu-PCR, asymmetric interlaced PCR, vectorette PCR, and inverse-PCR, have been developed based on various unknown region-capturing strategies, these methods can be less effective due to the contamination with nonspecific PCR products (Hui et al., 1998, McAleer et al., 1996, Stefano et al., 2016). These unintended amplicon products can be generated by mis-priming, non-specific annealing of primers, or end-repair priming by which unligated cohesive ends of the adaptor and non-target fragments are filled in (repaired) and then anneal together to initiate priming, eventually resulting in the amplification of unintended targets (Cavagnaro et al., 2009, Devon et al., 1995, McAleer et al., 1996, Pillai et al., 2008, Stefano et al., 2016, Uren et al., 2009). Splinkerette-PCR (spPCR) was developed to minimize nonspecific amplification by using a ‘splinkerette hairpin loop’ (Devon et al., 1995, Horn et al., 2007, Potter and Luo, 2010, Uren et al., 2009). The adapter has both top and bottom strands, and the 3’ end of the top strand forms a hairpin loop structure during the PCR step, reducing chances of mis-priming and end-repair priming (Fig. 1b). Furthermore, as the adaptor primer is designed to anneal to the complement of bottom strand (and not to the top strand hairpin loop), the specificity can be further increased (Fig. 1, Table 1). While spPCR was also applied to assess monoclonality in CHO cell lines, it was used complementary to the southern blot assay (Langsdorf et al., 2021), suggesting that there is more room for method optimization.

Here, we report the spPCR-based genome walking technique to amplify and to identify transgene integration sites in the CHO genome with substantially lower cost and analysis time. To increase the efficiency of spPCR amplification step, we calculated frequencies of various restriction sites in the CHO genome using Python scripts and selected suitable restriction enzymes to digest the genomic DNA (gDNA). After testing with a plasmid vector, we validated the spPCR-based genome walking technique to locate CHO housekeeping genes whose information about genomic locus is already known. Finally, this technique was applied to the recombinant CHO (rCHO) cells to identify the integration site of the recombinant protein gene.