Panel of human cell lines with human/mouse artificial chromosomes

Human artificial chromosomes (HACs) and mouse artificial chromosomes (MACs) are non-integrating chromosomal gene delivery vectors for molecular biology research. Recently, microcell-mediated chromosome transfer (MMCT) of HACs/MACs has been achieved in various human cells that include human immortalised mesenchymal stem cells (hiMSCs) and human induced pluripotent stem cells (hiPSCs). However, the conventional strategy of gene introduction with HACs/MACs requires laborious and time-consuming stepwise isolation of clones for gene loading into HACs/MACs in donor cell lines (CHO and A9) and then transferring the HAC/MAC into cells via MMCT. To overcome these limitations and accelerate chromosome vector-based functional assays in human cells, we established various human cell lines (HEK293, HT1080, hiMSCs, and hiPSCs) with HACs/MACs that harbour a gene-loading site via MMCT. Model genes, such as tdTomato, TagBFP2, and ELuc, were introduced into these preprepared HAC/MAC-introduced cell lines via the Cre-loxP system or simultaneous insertion of multiple gene-loading vectors. The model genes on the HACs/MACs were stably expressed and the HACs/MACs were stably maintained in the cell lines. Thus, our strategy using this HAC/MAC-containing cell line panel has dramatically simplified and accelerated gene introduction via HACs/MACs.

Human artificial chromosomes (HACs) and mouse artificial chromosome (MACs) have unique characteristics as vectors for gene delivery, which include stable and independent maintenance without disruption of the host genome and the capacity to carry numerous genes and megabase-sized genomic loci with physiological regulatory elements [1][2][3][4] . HAC/MAC technologies have been used for gene and cell therapies of Duchene muscular dystrophy [5][6][7][8][9][10] and to generate trans-chromosomic (Tc) animals that include a mouse model of Down syndrome 11,12 and humanised drug metabolism [13][14][15][16][17][18] . Furthermore, several types of HACs have been used in cancer research and drug development for cancer therapy [19][20][21] , centromere and telomere function elucidation 22 , a system of quantitatively tracking epigenetic memory in the field of synthetic biology 23 , and protein production 24 .
To accelerate the gene loading of multiple genes into HACs/MACs, we have developed several systems for multiple gene insertions, such as the simultaneous or sequential integration of multiple gene-loading vectors (SIM) system 25,26 via hypoxanthine-guanine phosphoribosyltransferase (HPRT)-deficient cells and drug screening with HAT, a multi-integrase (MI) system [27][28][29][30] , and homologous recombination with clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) 31 . HACs/MACs are transferrable into desired cells by microcell-mediated chromosome transfer (MMCT) 32 . Although MMCT traditionally employs polyethylene glycol 33 , we developed a novel microcell membrane fusion method with the envelope proteins of measles virus (MV) 34,35 , amphotropic virus, and ecotropic virus 36 , which improved the transfer efficiency (1 × 10 −4 -1 × 10 −5 ). However, specialised equipment and a laborious and time-consuming process are required for MMCT of HACs/MACs. This is because, in accordance with each experimental purpose, HACs/MACs with desired genes are constructed in CHO and A9 cells, and individually transferred to a target cell line via MMCT, and then clones are isolated, which contain the desired HACs/MACs (Fig. 1a). Therefore, we have previously employed mouse embryonic stem cells that contain a MAC with the MI system to facilitate the generation of Tc mice 37 . Under such circumstances, preprepared human cell lines that contain HACs/MACs will be a useful Step 1: TransfecƟon Step 2: MMCT hiMSC hiPSC HEK293 HT1080 Human cell line panel ). The human cell lines were directly available for gene loading into HACs/MACs with site-specific recombination. Simultaneous insertion of multiple gene-loading vectors into HEK293 cells, HT1080 cells, and hiPSCs (201B7) that contained HACs/MACs was tested by the SIM system. Each gene-loading vector of the SIM system contained Emerald luciferase (ELuc), a red fluorescent protein (tdTomato), or blue fluorescent protein (TagBFP2-N). hiMSCs that contained 21HAC2 were used to attempt insertion of three types of plasmid vectors that contained a red fluorescent protein (mCherry) driven by a different promoter with the Cre-loxP system. Each mCherry was expressed by a general constitutive promoter, e.g., PGK, CAG, or EF1α.  (Table 1). Then, plasmid vector(s) with a GOI were inserted into the HAC/MAC via the SIM system for HEK293 cells, HT1080 cells, and hiPSCs or the Cre-loxP system for hiMSCs. Fluorescence in situ hybridisation (FISH) analyses revealed that a single additional HAC or MAC was maintained independently from the host chromosome in each somatic/stem cell line. Specifically, HEK293 cells contained 21HAC1 ( Table 1. Regarding hiPSCs (201B7) with MAC6, the Neo resistance gene on MAC6 was disrupted (MAC6-ΔNeoR) for further gene insertion and drug selection. Although promoters for overexpression of transgenes often undergo gene silencing in human pluripotent stem cells 44 , HACs/MACs maintained the desired gene expression level in long-term cell culture, which differed from gene transduction with plasmid DNA via random insertion. Therefore, the insertion of the NeoR gene driven by the PGK promoter on MAC6 would be applicable to obtain a clone with an inserted circular plasmid vector for drug selection of hiPSCs. Thus, the NeoR gene on MAC6 in hiPSCs was knocked out to establish 201B7/MAC6-ΔNeoR cells ( Supplementary Fig. S1e).

Demonstration of gene loading with three vectors by simultaneous introduction using the SIM system into HACs/MACs in HEK293 and HT1080 cells.
We demonstrated that somatic/stem cell lines that contained HACs/MACs accepted three plasmid vectors by simultaneous introduction using the SIM system ( Supplementary Fig. S2b). As model genes for demonstration, a luminescent protein, Emerald luciferase (ELuc), and two fluorescent proteins, tdTomato and TagBFP2-N, were selected (hereafter collectively called ElTB). Specifically, we evaluated HEK293 cells that contained 21HAC1, 21HAC2, MAC2, or MAC4, HT1080 cells that contained 21HAC2 or MAC4, and hiPSCs that contained MAC6-ΔNeoR by introducing the three vectors with each model gene using the SIM system. Numerous drug-resistant clones were observed and an arbitrary number of clones were analysed by PCR to detect the correct insertion of the plasmid into the HAC/MAC. A summary of the PCR analysis is shown in Supplementary Table S1. The obtained drug-resistant clones were analysed by fluorescence microscopy, flow cytometry (FCM), and luciferase activity assays. These assays showed fluorescent proteins and luciferase expression in each cell line of representative HEK293 and HT1080 clones as expected. The results of HEK293/21HAC2-ElTB cells are shown in Fig. 3a,b,e, HEK293/MAC4-ElTB cells are shown in Fig. 3c-e, HT1080/21HAC2-ElTB cells are shown in Fig. 3f,g,j, and HT1080/MAC4-ElTB cells are shown in Fig. 3h-j. Representative results of the expression frequency of the introduced fluorescent protein gene and Table 1. List of human cell lines that contained HACs/MACs. Cell lines, HACs/MACs, EGFP marker, and drug resistance are summarised. Regarding drug resistance, R indicates cells resistant to antibiotics, i.e., Hyg hygromycin, Puro puromycin, BS blasticidin S, and S indicates that cells are selectable by drugs, i.e., HAT hypoxanthine-aminopterin-thymidine medium and Ganc ganciclovir. www.nature.com/scientificreports/ luciferase gene expression level were similar for each parental clone used for gene transfer. Therefore, to further expand the cell line panel, clones that show excellent gene expression should be selected. FISH analyses showed stable maintenance of HACs/MACs that contained the transgenes independently from host chromosomes. Fluorescence imaging of HEK293/21HAC1-ElTB and HEK293/MAC2-ElTB cells also showed expression of tdTomato and TagBFP2-N (Supplementary Fig. S3a,b). FISH analyses were also performed in HEK293/21HAC1 and HEK293/MAC2 cells that contained the vectors using the ELuc plasmid probe. The results showed that one HAC/MAC was maintained in each cell and that the transfected plasmid was inserted into the HAC/MAC as expected ( Supplementary Fig. S3c,d). These results showed that multiple gene loadings into the HAC/MAC via the SIM system was achieved in our established human cell line panel ( Fig. 2a-f), which enables seamless application of our HAC/MAC technology for gene function assays in human cells in the future.
Characterisation of hiMSCs that contain HAC and demonstration of gene loading by the Cre-loxP system. hiMSCs with a transferred 21HAC2 (Fig. 2g), which were clones A03 and D11, stably expressed EGFP (Fig. 4a). Various MSC markers were also analysed by RT-PCR in these clones (Fig. 4b). Expression level of CXCL1 and IL8 is higher in resultant clones than in parental hiMSCs. CD90 is a typical MSC marker and the others including CXCL1 and IL8 are linked to therapeutic functionality. For the recipient hiMSC cell lines employed for HAC introduction, 6-thioguanine (6TG)-resistant clones (#3 and #6) were used to obtain HAT-resistant cell lines by HPRT gene reconstruction. The 21HAC2-carrying cell lines (A03 and D11) showed comparable or higher expression levels of the various MSC markers compared with the original hiMSC cell line. These results indicated that the hiMSC clones that contained 21HAC2 maintained the characteristics of MSCs (Fig. 4b).
To evaluate the HAC retention ratio after long-term cell culture with or without an antibiotic (blasticidin; Bsd), FISH analysis was performed and results showed stable maintenance of the HAC at population doubling levels (PDLs) of 24 and 39, even without Bsd (Fig. 4c). These results indicated that the two hiMSC clones that contained 21HAC2 (hiMSC/21HAC2 A03 and D11) could be used as a platform for gene loading 38 .
We validated whether 21HAC2 functioned as a safe harbour for gene expression in these established clones. As an example of functional analysis, we attempted to evaluate the expression level of the transgene promoted by three types of constitutive promoters (PGK, EF1α, and CAG) on 21HAC2 in hiMSCs with a defined (single in www.nature.com/scientificreports/ this study) copy number. Among the drug-resistant clones obtained by transfection with a plasmid that carried each promoter, 10 clones were picked up in order of the fluorescence intensity of mCherry under a fluorescence microscope and used for subsequent analysis to measure each promoter activity. Fluorescence imaging of mCherry-expressing cells with each promoter indicated that CAG and EF1α promoter activities were higher than the PGK promoter activity (Fig. 4d). qRT-PCR of the mRNA expression level of mCherry also indicated that CAG and EF1 promoter activities were higher at 22.4-fold (CAG) and 40.5-fold (EF1α) compared with the PGK promoter activity (n = 10) (P < 0.01) (Fig. 4e). There was no significant difference between the activities of CAG and EF1α promoters. These results supported a previous study that compared promoter activity in MSCs with a viral vector system for gene expression 45 . Because the Cre-loxP system had the same adaptor as the SIM system in 21HAC2, the SIM system would be applicable to hiMSCs/21HAC2. These results showed that hiMSCs that contained 21HAC2 were applicable to gene loading and functional analyses.
Characterisation of hiPSCs that contained MAC and demonstration of gene loading by the SIM system. hiPSCs (201B7) that contained MAC6-ΔNeoR expressed EGFP (Fig. 5a). We performed longterm cell culture without an antibiotic (G418) and quinacrine-Hoechst karyotyping to evaluate the HAC retention ratio. Karyotyping of hiPSCs showed an ideal karyotype that included a MAC, 47, XX, + MAC (Fig. 5b, left panel). The long-term culture of hiPSCs revealed that the karyotype and MAC were stable at a PDL of 20 (Fig. 5b, right panel). Next, we attempted gene loading with the SIM system into 201B7/MAC6-ΔNeoR cells. The obtained hiPSC clone [(201B7)/MAC6-ΔNeoR-ElTB] expressed tdTomato, TagBFP2-N, and EGFP (Fig. 5c). FISH analysis showed that the transgenes were integrated into the MAC and that the MAC was independently maintained in the hiPSCs (Fig. 5d). The expression level of each fluorescent marker and ELuc in 201B7/MAC6-ΔNeoR-ElTB cells was evaluated by FCM analysis and luciferase assays, respectively (Fig. 5e). Expression of all introduced genes was detectable among the representative two clones. Furthermore, 201B7/MAC6-ΔNeoR-ElTB cells showed a normal karyotype and stable maintenance of the MAC as well as 201B7/MAC6-ΔNeoR.
In vitro, the expression of pluripotency markers was verified. The iPS clones 201B7/MAC6-ΔNeoR-ElTB showed activity of alkaline phosphatase and expression of OCT 3/4 as well as 201B7 (Fig. 5f,g). Quantitative RT-PCR analysis confirmed expression of OCT 3/4 in each clone (Fig. 5h). To verify the pluripotency of these iPSC lines, we tested their ability to differentiate into all three germ layers in vivo using the teratoma method. 201B7/ MAC6-ΔNeoR and 201B7/MAC6-ΔNeoR-ElTB cells showed maintenance of pluripotency to differentiate into the three germ layers after gene loading and cloning (Fig. 5i,j). These results showed that 201B7/MAC6-ΔNeoR cells were applicable to gene loading with MAC technology.

Conclusion
We generated a novel human cell line panel with HEK293 cell, HT1080 cell, hiMSC, and hiPSC (201B7) lines that contained HACs/MACs, which enabled rapid and precise insertion of GOIs at a defined site on HACs/MACs by a simple transfection method. The GOIs were stably expressed in each cell line, which indicated that the integration site can act as a "safe harbour" to support transgene expression. Thus, our new preprepared cell panel with HACs/MACs may dramatically simplify the construction of HACs/MACs with desired genes and constructed HACs/MACs can be used immediately and directly for functional analyses of genes in desired cell lines.    35 . Twelve flasks of CHO cells were prepared and micronuclei were induced by treatment with 0.1 µg/mL colcemid. The detailed MMCT protocol has been described previously 26 . The collected microcells were cocultured and fused with 2 × 10 6 cells of each recipient cell line for 24 h in a 6-cm dish (Corning, Corning, NY, USA). Then, the fused recipient cells were subcultured into three 10-cm dishes. Drug selection was started with optimal selectable antibiotics after a further 24 h of incubation. After 14-21 days, drug-resistant colonies were picked up and expanded for the following analyses.
Gene knockout by CRISPR/Cas9. Gene knockout of HPRT1 was performed in HEK293 cells, hiMSCs, and hiPSCs with the multiplex CRISPR/FokI-dCas9 vector system 50,51 . The multiplex CRISPR/FokI-dCas9 vec- FISH analysis. Cells were treated with colcemid to induce metaphase arrest, treated with 0.075 M KCl, and then fixed with methanol/acetate (3:1) (FUJIFILM Wako). FISH was performed with the p11-4 alpha satellite probe 52 to stain the alpha satellite of hChr.13, 21 and HAC, and mouse Cot-1 DNA to stain the MAC. The probes were labelled with digoxigenin (Roche, Basel, Schweiz) and the inserted plasmid vector targeted to the chromosome fragment was labelled with biotin (Roche). The DNA probes were labelled with a nick translation kit (Roche), following the manufacturer's instructions. The detailed protocol has been described previously 26 .
Teratoma formation and histological analysis. Mice were maintained under specific pathogen-free conditions with a 12-h light-dark cycle. Human iPSCs (1 × 10 6 ) were subcutaneously transplanted into a testis of anaesthetised severe combined immunodeficiency mice (Charles River, Yokohama, Japan). A mixed anaesthetic agent prepared with 0.3 mg/kg medetomidine hydrochloride, 4 mg/kg midazolam, and 5 mg/kg butorphanol tartrate was administered intraperitoneally to the mice. Teratomas appeared after ~ 8 weeks. The anaesthetised mice were sacrificed and the teratomas were collected. Then, the teratomas were fixed with 20% neutral formalin/PBS and processed for paraffin sectioning. The sections were stained with haematoxylin and eosin. www.nature.com/scientificreports/ AGG AAC CAT CTC AC -3ʹ and R 5′-ATT TGG GGT GGA AAG GTT TG -3′, and CCL2 F 5′-GCA GCA AGT GTC CCA AAG AA -3′ and R 5′-AAC AGG GTG TCT GGG GAA AG -3ʹ.
Alkaline phosphatase staining. The iPS cells were prepared in confluently and fixed with 4% paraformaldehyde/PBS (Sigma Aldrich) in a freezer overnight. The staining was performed with Alkaline Phosphatase Staining Kit (Cosmo Bio Co., LTD., Tokyo, Japan). The image were obtained by microscopy for cell culture (OLYMPUS CORPORATION, Tokyo, Japan) and the attached CCD camera DP22 and Software.

Data availability
The datasets of vector sequences and cell lines generated during and/or analysed during the current study are available from the corresponding author on reasonable request.