Site-specific Integration Ushers in a New Era of Precise CHO Cell Line Engineering

Abstract

Chinese hamster ovary (CHO) cells are widely used for the production of therapeutic proteins. Customarily, CHO production cell lines are established through random integration, which requires laborious screening of many clones to isolate suitable producers. In contrast, site-specific integration (SSI) accelerates cell line development by targeting integration of transgenes to pre-validated genomic loci capable of supporting high and stable expression. To date, a relatively small number of these so called ‘hot spots’ have been identified, mainly through empirical methods. Nevertheless, nuclease-mediated and recombinase-mediated SSI have revolutionized cell line engineering by enabling rational and reproducible transgene targeting.

Graphical abstract

Introduction

Chinese hamster ovary (CHO) cells are the expression host of choice for the commercial production of recombinant glycoproteins. At present, a diverse portfolio of biotherapeutics produced by CHO cells are approved for market, including several monoclonal antibodies (mAbs), as well as other proteins such as interferon-β and tissue plasminogen activator [1]. Since the advent of CHO cell culture production processes more than 30 years ago, dramatic improvements in volumetric productivity, referred to as titer, are a testament to the persistent efforts of the CHO community. In particular, industrial mAb titers of the past were typically well below 1 g/L. However, it is now routine for biomanufacturing teams to achieve mAb titers in the grams per liter range [2]. Innovations across the upstream process have led to this improvement in CHO cell productivity. These innovations include media and bioreactor process optimization as well as cell line engineering for better process compatibility, culture longevity, and waste product reduction.

However, within the cell line generation stage of the upstream process, it remains customary to use untargeted transgene integration methods to generate stably transfected CHO cell pools from which production cell lines are established. Consequently, heterogeneity within cell pools warrants the tedious screening of multiple clones to identify those with suitable production characteristics. In contrast, site-specific integration (SSI), whereby the transgene encoding the recombinant protein of interest is integrated at a predetermined genomic locus, offers the means to generate more consistent clones, reducing cell line development timelines. As such, SSI has emerged as a promising strategy by which CHO cell line development teams can repeatedly target their preferred genomic sites that are capable of highly active and stable expression. The aim of this brief review is to examine the implementation of SSI technology in CHO cell line engineering thus far with a focus on tools, obstacles, and achievements.

Traditional CHO cell line development

A familiarity with the traditional cell line generation workflow (Figure 1) is necessary to appreciate the power and precision of SSI. The following methodologies have been in practice among academics and industrial cell line development teams since the first recombinant CHO product approval. Although instrumentation and hardware have evolved significantly over that period, the underlying principles have remained relatively constant. Recombinant DNA (rDNA) encoding the protein(s) of interest and a selection marker is delivered to naive CHO founder cells through chemical (e.g. calcium phosphate, lipofection) or physical (e.g. electroporation) transfection processes. Within the nucleus, rDNA integrates randomly into the genome, resulting in clones with variable transgene integration site and copy number. Following transient expression, selective pressure is applied by dosing the culture with the appropriate antibiotic to eliminate cells that do not harbor the transgene, thereby establishing stably transfected polyclonal pools. Gene amplification, mediated by either the dihydrofolate reductase (DHFR)-methotrexate (MTX) or glutamine synthetase (GS)-methionine sulfoximine (MSX) system, can be performed at this step to generate cells with higher expression levels by increasing transgene copy number [3,4]. Subcloning is performed by limiting dilution or more high-throughput methods, such as flow cytometry or robotic imaging technology [5]. From single cells, subclonal populations are expanded and screened so that they can be ranked according to productivity and other characteristics (including viability, glycosylation, stability, and so on). Finally, the leaders are advanced for further cell line characterization, including more extensive assessment of long-term recombinant expression stability and determination of product quality attributes.

Figure 1.

The traditional CHO cell line development workflow. First, founder cells are transfected with recombinant DNA (rDNA) carrying the transgene for the desired protein product. Cells are subsequently selected on semi-solid media to ensure random integration has occurred. Gene amplification can be performed during selection steps to increase gene copy number. Next, subcloning methods are used to isolate single clones in a 96-well plate format. Clones are cultured and expanded to facilitate screening of transgene expression levels. Top producers move forward for characterization of their performance and product quality attributes in shake flasks and bioreactors before further process development and optimization.

Typically, screening efforts are extensive, and many clones must be carried through the cell line development process due to the high level of heterogeneity observed among stably transfected cells. Cell-to-cell variability has been attributed to the preexistence of genetically distinct sub-populations present within the host cell culture prior to transfection as well as the differences in rDNA insertion sites among cells [6,7]. Genomic rearrangements associated with rDNA integration and MTX-induced gene amplification have also been reported [8,9]. The mechanism of random integration is thought to involve homology-independent DNA repair pathways initiated at the site of spontaneous DNA double-strand breaks (DSBs). Specifically, random integration in mammalian cells is the product of DNA polymerase θ-mediated end joining and nonhomologous end joining (NHEJ) [10,11]. Within these repair pathway families, synthesis-dependent microhomology-mediated end joining (SD-MMEJ) was recently reported to be the dominant mechanism driving random integration of rDNA in CHO cells [12•]. Mammalian DNA repair continues to be an active field of research and has been reviewed elsewhere [13].

Genomic hot spots

What makes a good integration site?

Cell line characteristics, such as expression level and stability, directly linked to the site of transgene integration are termed positional effects. The transcriptional activity at a particular site is governed by its chromatin state and corresponding accessibility. Moreover, loss of cell-specific productivity during extended culture of CHO cells has been associated with repressive remodeling of the transgenic chromatin via promoter methylation and histone deacetylation [14,15]. Beyond epigenetic silencing, chromosomal rearrangements and deletions are known to contribute to expression instability by reducing transgene copy numbers [16,17]. Curiously, genomic loci vary with respect to their propensity for genetic instability (reviewed in [18]).

In the context of CHO cell line engineering, chromosomal loci that have been determined to possess the intrinsic ability to confer exceptional stability and enhanced transcriptional activity are referred to as hot spots. For the small collection of CHO hot spots that have been identified thus far (Table 1), the underlying molecular mechanism for their respective properties remains largely speculative. For example, Koduri et al. reported that the hot spot for increased recombinant protein expression (HIRPE) was found to trigger the formation of inverted repeats and contain several matrix attachment region (MAR)-like sequence elements [19]. When included in transfected rDNA sequences, MARs, which tether chromatin to the nuclear matrix, were previously shown to improve transgene expression levels [20].

Table 1.

Confirmed hot spot loci in the CHO genome

Scaffold	Accession	Position or (Region)	Genea	Reference
1970	NW_003613585.1	(1896657.. 1901695)	unannotated	[19]
1924	NW_003614386.1	60148	unannotated	[21••]
1361	NW_003614117.1	1075377	unannotated
1558	NW_003614502.1	414422	unannotated
5035	NW_003613746.1	1864526	unannotated
934	NW_003613582.1	837184	BMP5
2625	NW_003614999.1	237547	SSBP2
156	NW_003613587.1	5618538	TRMT6
424	NW_003614241.1	177811	unannotated
2259	NW_003615469.1	86304	CLCC1
3405	NW_003617368.1	4675	FAM114A1 (NOXP20)
1727	NW_003614039.1	865981	LRBA
1385	NW_003613853.1	898266	unannotated
1485	NW_003614008.1	438214	DCN
2165	NW_003613824.1	931631	CEP128
832	NW_003614624.1	117635	AACS
262	NW_003613881.1	1571782	ALDH5A1
1796	NW_003613796.1	188123	SMAD6
624	NW_003613718.1	2206187	unannotated
1753	NW_003615261.1	117475	unannotated
1149	NW_003616625.1	49532	PTPRQ
2241	NW_003613637.1	132202	unannotated (ROSA26)
208b	NW_003613622.1	1248580	ADGRL4	[9]
243b	NW_003615666.1	191785
4938c	NW_003615627.1	49456	KIAA1551 (C12ORF35)	[22•]
1588	NW_003613932.1	(1120880.. 1127796)	HPRT	[22•,32•]
1228	NW_003613703.1	(2394533..2398130)	CLCN3	[23]
1492	NW_003613833.1	~1756190	FER1L4	[25,28••]
7847	NW_003618207.1	(5356.. 18545)	unannotated	[51]
11799	NW_003617543.1	(1.. 14931)	unannotated

^aAccording to Annotation Release 102 of the CriGri_1.0 RefSeq found on CHOgenome.org accessed on July 9, 2018

^bHot spot formed by the rearrangement of two scaffolds

^cDisruption of this gene was previously shown to increase productivities of transgenes integrated elsewhere [52]

In a landmark study for SSI in CHO cells, Gaidukov et al. identified 21 novel hot spots for high and stable expression [21••]. Their findings reaffirm previous observations that hot spots can exist in both annotated loci and intergenic regions of the genome.

Furthermore, they demonstrate that these sites retain their properties when retargeted de novo, at least in CHO cell lines with a common lineage. Still, hot spot activity in one species may not be conserved at the orthologous loci of related species. Between two loci, ROSA26 and GRIK1, with potential hot spot activity based on their properties in mouse and human cells, one (ROSA26) supported stable expression in CHO cells [21••], while the other did not [22•].

Overall, the sequence-specific mechanism of hot spot activity is poorly understood, likely differing from site to site. Indeed, one recent example of an integration site that fit the criteria of a hot spot was found to have resulted from a unique chromosomal rearrangement, preventing de novo retargeting [9]. Even when a targetable hot spot is successfully identified, how far its activity extends to surrounding chromosomal regions remains unclear. At present, the gaps in our understanding of hot spots necessitate empirical strategies for identifying and harnessing these sites. The determination of hot spot hallmarks certainly warrants further investigation among researchers in the CHO community.

The hunt for hot spots

Prior to the sequencing of the CHO genome and the development of genome editing tools, pioneering SSI efforts were strictly limited to low-throughput reverse genetics approaches. In this manner, genomic sequences, including the HIRPE and the expression augmenting sequence element (EASE), flanking rDNA in highly productive clones were isolated for use on transgene expression vectors to exploit their enhancer properties [19,23]. More recently, a CHO genomic library was used to screen for the presence of expression enhancing elements [24]. Library-based approaches are certainly useful for high-throughput screening, but sacrifice information regarding the properties of each fragment in its native chromosomal context.

A more aggressive approach for the empirical identification of genomic hot spots was enabled by site-specific recombinases (SSRs) through the tag-and-exchange approach. Recently, this approach was used to establish high-producing CHO cell lines with expression cassettes integrated at pre-validated hot spots capable of supporting titers of several grams per liter [25,26]. The tag-and-exchange strategy is employed by generating a large pool of tagged cells, each with a stably integrated tagging cassette, referred to as a landing pad. The landing pad carries both a recombinase target site (RTS) and a reporter gene to rapidly evaluate the expression level supported by an integration site. The landing pad can be integrated randomly [9,25], introduced through viral infection [21••], or targeted to recombinase pseudosites [26]. It was previously shown that the genetic elements contained in rDNA and the manner in which rDNA is introduced into the cell can influence the site of integration in CHO cells [12•,21••]; users should keep these potential biases in mind. After screening of tagged cells, top performing clones with the highest and most stable expression of the reporter gene are used as reliable hosts for the production of recombinant proteins of interest via recombinase-mediated SSI.

The SSI toolbox

Recombinase-mediated SSI: landing pads and RMCE

CHO cell line development teams have embraced an assortment of SSR technologies, including Flp-FRT and Cre-loxP as well as the φC31 and Bxb1 integrases. These systems reliably catalyze targeted integration of donor DNA at their respective recognition target given that the donor carries the cognate site. Consequently, prior establishment of tagged cells harboring a genomic RTS is a prerequisite for recombinase-mediated SSI. Despite this shortcoming, recombinase-mediated SSI has several advantages to offer. When surveying the genome for transcriptional hot spots by untargeted integration of a landing pad, a pre-marked and pre-validated locus can be subsequently targeted without explicit determination of the chromosomal coordinates. Additionally, landing pads can be configured with selection markers and strategically placed promoter elements to ensure proper integration of rDNA donors. Thus, recombinase-mediated SSI enables the use of platform cell lines that support highly efficient and reproducible transgene targeting.

Methodologies have evolved with the development of recombinase-mediated cassette exchange (RMCE) systems, which facilitate efficient SSI in a unidirectional manner by exchanging a designated portion of the landing pad with a donor cassette (reviewed extensively in [27]). An RMCE competent cell line is established through integration of a landing pad flanked by an orthogonal RTS pair. When tested against the commonly used Flp-FRT RMCE system, it was determined that the novel Bxb1 RMCE system is more precise for generating CHO cell clones with high fidelity cassette exchange events [28••]. Integration of cumbersome amounts of heterologous DNA poses a challenge due to low efficiencies and potential off-target effects yet is necessary for engineering cell lines capable of expressing multiple transgenes at once. Lately, RMCE efforts in CHO cells have focused on the development of methods to target multiple sites simultaneously. A binary system using orthogonal RMCE cassettes was pioneered for independent co-expression of transgenes at predetermined levels [29]. A more recent study demonstrated targeted integration of up to nine copies of a mAb expression cassette across three pre-validated loci [21••]. Platform cell lines engineered in this manner may be beneficial for hosting production of increasingly complex multi-subunit protein products, such as bispecific and trispecific antibodies.

Nuclease-mediated SSI: controlled induction and repair of DSBs

Nuclease-mediated SSI is implemented by harnessing the repair of DNA DSBs to introduce rDNA. With the development of programmable endonucleases came the means to induce targeted DSBs within the genomes of sequenced organisms, including CHO cells. Zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), and clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) are among the currently available programmable endonuclease systems for targeted induction of DNA DSBs [30].

Reports of nuclease-mediated SSI in CHO cells have varied with respect to the particular pairing of programmable endonuclease and the leveraged DNA repair pathway (Figure 2). Cristea et al. demonstrated the first use of engineered endonucleases, namely ZFNs and TALENs, for targeted integration of transgenes in CHO cells, including homology-independent targeted integration (HITI) of an antibody expression cassette at the hamster FUT8 locus [31]. A more recent strategy, referred to as the PITCh (Precise Integration into Target Chromosome) system, harnessed CRISPR/Cas9 and MMEJ using rDNA donors flanked by short (≤ 40 bp) microhomologies to incorporate a construct containing a single-chain Fv-Fc (scFv-Fc) gene at the hamster HPRT locus [32•]. Yet another innovative approach coupled the use of CRISPR/Cas9 with homology-directed repair (HDR) for targeted integration of homologous rDNA donors, encoding fluorescent reporters, at the hamster C1GALT1C1 (COSMC), MGAT1, and LDHA loci [33]. HDR is a DNA repair pathway that uses homologous DNA to precisely repair DSBs and can be exploited through the use of exogenous donor repair templates. Stimulation of HDR via CRISPR/Cas9 has since emerged as the predominant methodology for nuclease-mediated SSI in CHO cells. Advantages of leveraging this workflow include the ease of use associated with implementation of the CRISPR/Cas9 platform, as well as the high fidelity and directionality of HDR-mediated transgene integration.

Figure 2.

Nuclease-mediated approaches for SSI in CHO cells. First, an engineered nuclease is used to induce a double-strand break (DSB) at a predetermined locus in the genome. Currently, these programmable endonucleases include zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs) and clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9). Next, an endogenous DNA repair pathway is exploited for targeted transgene integration at the cut site. Homology-independent targeted integration (HITI) involves the NHEJ repair pathway to integrate donor sequences that lack homology with the target site. Notably, HITI is nondirectional meaning that rDNA (green) may be inserted in the sense or antisense orientation relative to genomic DNA (blue). Integration via homology-directed repair (HDR) or microhomology-mediated end joining (MMEJ) relies on homology arms that flank the donor DNA, which are either long or relatively short (≤ 40 bp), respectively.

Still, an inherent recalcitrance to HDR-mediated integration of large constructs has been noted as a shortcoming of CHO cells [31]. Several recent reports, described briefly here, have documented promising efforts to improve nuclease-mediated SSI through enrichment of targeted integrants or increased knock-in efficiency. A three-fold increase in the number of CHO cells with HDR-mediated targeted integration events was achieved using a fluorescent enrichment strategy, relying on reporters in Cas9 and donor constructs [34•]. The same report showed that chemical treatment with Scr7 or lithium chloride does not increase the efficiency of HDR-mediated targeted integration; however, other small molecules, including RS-1, L755507, and resveratrol, which are capable of significantly increasing HDR events in other cell types [35,36], have yet to be applied to CHO cells. Furthermore, cold-shock treatment following transfection was very recently reported to enhance HDR rates up to ten-fold in human cell lines [37]. In contrast to physical and chemical treatment-based approaches, the rDNA donor can be modified for enhanced activity as a substrate for HDR-mediated integration. The dependence of HDR efficiency on homologous sequence length was characterized in mouse cells, leading authors to recommend the use of donors with homology arms of 1 Kbp or greater [38]. Linearization and release of HDR donors via dual CRISPR/Cas9 cleavage of circular plasmids increased HDR efficiency two- to five-fold relative to uncut circular donors [39]. Comparable improvement was documented when the double cut HDR donor strategy was applied in CHO cells [22•]. Lastly, tethering donor DNA to Cas9 can improve HDR efficiency at the expense of NHEJ by enhancing donor recruitment to the DSB site as demonstrated in human and mouse systems [40].

Hybrid SSI: fusing the power of nucleases and recombinases

The full potential of current SSI technology is exemplified by Inniss et al., presenting a hybrid approach that uses both nuclease-mediated and recombinase-mediated SSI for rational and precise cell line engineering of CHO cells [28••]. First, nuclease-mediated SSI is employed to target a landing pad to a pre-validated hot spot, such as the FER1L4 locus, using CRISPR/Cas9 and HDR (Figure 3A). The landing pad can be equipped with a reporter to support fluorescent enrichment, a negative selection marker, and the highly efficient Bxb1 RMCE system. Once a master cell line has been established, RMCE is used to incorporate a donor cassette within the landing pad (Figure 3B), thereby delivering gene(s) encoding the recombinant product of interest to the hot spot. This methodology combines the best features of current SSI tools to produce cell pools that are reproducible, stable, and high-producing. Moreover, cell-to-cell homogeneity within these cell pools ensures less variation in product quality attributes compared to cell pools derived from random integration methods. Screening of SSI clones remains a necessary step to confirm expected production traits, albeit the process is less extensive than that of random integrants due to minimization of positional effects.

Figure 3.

A hybrid SSI strategy to establish a platform for rapid generation of production CHO cell lines. (A) Nuclease-mediated SSI: an engineered nuclease, such as CRISPR/Cas9, is first used to cut the genomic DNA (black) at a known hot spot, stimulating HDR-mediated integration of an RMCE landing pad donor (purple) with flanking homology arms. Landing pads can be designed with a variety of useful features: a negative selection marker to aid in the eventual replacement of the landing pad as well as a fluorescent maker to enable cell sorting and enrichment strategies. An ‘RMCE landing pad’ is one that is flanked by an orthologous RTS pair (attP and attP’ in this example). (B) Recombinase-mediated SSI: an established cell line harboring the landing pad undergoes RMCE when transfected with an RMCE donor cassette (orange) along with the corresponding recombinase. The system shown here uses the highly efficient Bxb1 RMCE system (via attPxattB and attP’xattB’ recombination). Note the placement of the landing pad promoter (P1), which lies outside the boundaries of the exchanged cassettes. It is used to ‘trap’ the promoterless positive selection marker on the incoming donor. Following RMCE, an expression cassette for the gene of interest (GOI) is planted in fertile ground for effective protein production.

CHO-specific bioinformatics resources

Sequencing the genome of the widely used CHO-K1 cell line was a landmark achievement for the advancement of CHO cell line engineering [41]. To aid the CHO community, CHOgenome.org (accessible at http://www.chogenome.org) was established to collect and distribute CHO genomic information as well as provide bioinformatics tools, including a genome viewer and CHO-specific BLAST server [42,43]. In recent years, many more CHO genome sequences have become available. Not surprisingly, striking genetic diversity was detected among commonly used CHO host cell lines from different lineages, supporting the need for the assembly of a gold-standard Chinese hamster (CH) reference genome [44]. Both the CHO and CH genomic sequences are an invaluable resource for effective implementation of SSI technology.

Among the functionalities of the CHO genome website that are pertinent to SSI execution, genes can be searched to retrieve their scaffold and transcript accessions from the current CHO-K1 and CH reference sequences [42]. If targeting a particular annotated locus is desired, this search tool may be especially useful for designing a suitable nuclease-mediated SSI strategy. Likewise, another available web-based tool called CRISPy was developed to help researchers find unique CRISPR/Cas9 target sites in coding sequences of CHO genes [45]. To aid researchers in the design of plasmids needed for the employment of MMEJ-mediated SSI, a web-based design software was developed by the inventors of the PITCh system [46].

In contrast to the growing number of tools available for SSI execution, fewer resources are available for the analysis of transgene integration sites. These resources include transcriptomic data compiled on the CHO genome database that can be used to determine local transcriptional activity of native genes in the vicinity of an integration site [43]. Additionally, in a noteworthy study, Feichtinger et al. characterized the epigenetic landscape of several cell lines derived from CHO-K1, providing data tracks that reveal methylation patterns as well as histone mark profiles, which indicate chromatin state, throughout the duration of a batch culture [47•]. Still, very little is known regarding the location and function of noncoding regulatory elements present within the CHO genome due in part to the relatively poor quality of the CHO-K1 genome assembly. However, the CH genome was recently reassembled using third-generation sequencing technology for considerably improved contiguity [48•]. Hopefully, this new genome will provide useful insights regarding the broader chromosomal context of an integration site based on cross-analysis of regions that align between the two genomes.

Future directions and challenges

Despite its many advantages, one major weakness of current SSI technology is that it fails to generate clones with comparable expression levels to those obtained through high copy number random integration methods [49]. Indeed, the expression level achieved from hot spots was found to be strictly limited by the number of transgene copies integrated at those sites, showing a direct linear relationship between cumulative expression level and copy number [21••]. An accumulative SSI strategy was previously developed in CHO cells that uses successive rounds of transfection, fluorescent enrichment, and Cre recombinase-mediated SSI at single loxP targets to precisely control the accumulation of scFv-Fc expression cassettes up to six copies [50•]. While valuable as a proof of concept, the copy numbers are still relatively low. Future SSI advancements will certainly focus on developing robust methods for rapidly increasing transgene copy number in a targeted and more precise manner relative to random integration or gene amplification. Perhaps such a strategy will involve controlled concatemerization of expression cassettes prior to transfection.

Additionally, rational targeting of chromosomal sites is currently limited by our inability to predict the location of hot spots. Current strategies rely heavily on pre-validated loci identified through empirical methods. As the number of identified hot spots increases, patterns may be detected that explain the mechanism by which these sites confer such desirable qualities. Moreover, access to a collection of loci with characteristic levels of transcriptional activity that are of relatively low or medium strength may be of particular usefulness and even necessary for the implementation of complex cell line engineering projects, such as metabolic pathway engineering and the expression of multi-subunit protein complexes or non-coding RNAs. Certain desirable expression properties, including temporal control with respect to growth phase, may be possible by harnessing native genetic regulatory networks. Rational selection of targets may also be extended to SSI strategies that intentionally disrupt the expression of problematic host cell proteins during targeted integration to ease downstream processing burdens. Targeted disruption of native genes raises intriguing questions, such as the impact of rDNA integration on local chromatin regulation and transcriptional architecture.

Despite the clear advantages of SSI, it does not solve every issue that can impede cell line development, such as the inherent difficulties associated with expressing certain product molecules and spontaneous mutations that impair the native cellular machinery. Fortunately, SSI is just one of many approaches available to cell line development teams. In their current form, SSI technologies (Table 2) are quite effective, enabling robust targeting of user-determined loci and reproducible expression levels. Whereas traditional cell line development approaches that use random integration can take six months or more to establish suitable production cell lines, SSI platforms have been commercially implemented for accelerated timelines [49]. For reference, mAb yields in the grams per liter range can be generated in as little as 5 weeks from transfection of an RMCE competent master cell line to the end of fed-batch production [26]. For now, approaches that employ the strengths of both nuclease-mediated and recombinase-mediated SSI tools, namely CRISPR/Cas9 and RMCE systems, are the state of the art for SSI in CHO cells. Moving forward, SSI technology will continue to be a powerful tool for precise cell line engineering in the CHO community.

Table 2.

Site-specific integration methods

SSI Method	Advantages	Disadvantages
Recombinase-mediated	Useful for empirical identification of hot spots Amenable to negative selection and promoter trap Generates retargetable master cell line platform	Requires prior integration of RTS into genome Targets landing pad that was integrated randomly (i.e. chromosomal coordinates are unknown)
Nuclease-mediated	Targets predetermined chromosomal coordinates with known expression characteristics	Suffers from low efficiency Not amenable to negative selection or promoter trap
Hybrid (nuclease- and recombinase-mediated)	Targets predetermined chromosomal coordinates Amenable to negative selection and promoter trap Generates retargetable master cell line platform	Requires two-step cell line generation process to obtain producing cell line from naive founder cells

Abbreviations: RTS, recombinase target site.

Highlights.

• A catalog of confirmed hot spots in the CHO genome is compiled

• Emerging strategies for more efficient nuclease-mediated SSI are highlighted

• Progress toward multi-site and accumulative SSI for controlled gene copy number is discussed