Globin Gene in a Hematopoietic Cell Line Using a New Site-Speciﬁc Integrating Non-Viral System

ORIGINAL ARTICLE Long-term and efﬁcient expression of human β-globin gene in a hematopoietic cell line using a new site-speciﬁc integrating non-viral system

K Dormiani1, H Mir Mohammad Sadeghi1, H Sadeghi-Aliabadi1, K Ghaedi2,3, M Forouzanfar2, H Baharvand4,5 and MH Nasr-Esfahani2

Targeted integration of a therapeutic gene at specific loci in safe genomic regions by a non-viral vector can restore the function of the damaged gene. This approach also minimizes the potential genotoxic effects of transferred DNA. In this study, we have developed a non-viral vector that functions according to site-specific recombination (SSR). The vector contained a bacterial backbone and puromycin resistance gene (purr), a β-globin expressing cassette and an attB recombination site. We used phiC31 integrase to insert a copy of the vector into specific genomic locations of a human hematopoietic cell line. Site-specific integration of the vector with one or two copies in the transcriptionally active regions of the genome was confirmed. After genomic integration, we used Cre recombinase to remove the bacterial backbone and purr. This removal was verified by negative selection and genomic PCR screening. Following deletion of these sequences, the stable β-chain expression was continued for several months in the absence of selective pressure. Consequently, this vector may potentially be a powerful tool for ex vivo correction of β-globinopathies such as β-thalassemia through successful genomic integration of a functional copy of the globin gene into the patient’s target cells.

Gene Therapy (2015) 22, 663–674; doi:10.1038/gt.2015.30

INTRODUCTION be genetically corrected and hematopoietic stem cells derived Gene therapy is a technique for treating commonly inherited and from these induced pluripotent stem cells can be transplanted to 9 acquired disorders by delivering functional genetic material into the patient to restore globin chain expression. For this purpose, specific cells or tissues for therapeutic benefit. This approach new promising technologies such as non-viral site-directed either cures a disease or slows its progression.1,2 Although highly integrating methods have the capability to insert the corrective promising, therapeutic gene transfer faces a number of challenges DNA into specific genomic regions in stem or progenitor cells – such as the design of efficient vectors. Vectors designed for without interfering with their proliferation and pluripotency.10 13 curative circumstances should provide tissue-specific, position- In this method, known as site-specific recombination (SSR), independent and appropriately regulated expression of corrective bacteriophage integrases such as phiC31 offer the ability to DNA particularly for an extended period. Such vectors should have mediate insertion of an engineered DNA that contains the attB site minimal risks for genomic toxicity, immunogenicity and signaling (donor plasmid) into a subset of preferred locations in mammalian pathway disruption.3–5 The vectors for gene delivery can be genome, termed pseudo attP sites.14 These specific docking sites administered in vivo or ex vivo. Ex vivo gene therapy consists of are non-identical consensus sequences, however, they have partial transferring the therapeutic gene into patient-derived stem or similarity to the native attP and are mostly located in the progenitor cells, which are able to proliferate and have the intergenic regions compared with other areas of human – capacity for re-implantation into the donor. Thus, there is no need chromosomes.15 17 Following phiC31 integrase activity and for immunosuppression.6 This approach can resolve the limita- generation of two new hybrid junctions (attL and attR), tions of finding matched donors, as well as the risks of graft- a unidirectional recombination occurs that favors stable integra- versus-host disease and graft rejection because of allogeneic tion of the donor plasmid.12 In light of the biosafety aspects of this transplantation in patients with thalassemia major.3 Beyond the approach, Sivalingam et al.18 have concluded that the phiC31 classical gene therapy that uses hematopoietic stem cells of the system is safe for human applications because of the lack of patient to treat β-thalassemia, novel developing technologies tumorigenicity, rare chromosomal aberrations and low influence such as induced pluripotent stem cells may open a new era for on cellular transcription. A number of pseudo attP sites, called safe cell-based gene therapy.7,8 In this technique after generation of harbors, provide secure locations for hosing and stable expression patient-specific induced pluripotent stem cells, globin defects can of a transgene at the level necessary to achieve therapeutic

1Department of Pharmaceutical Biotechnology and Isfahan Pharmaceutical Sciences Research Center, School of Pharmacy and Pharmaceutical Sciences, Isfahan University of Medical Sciences, Isfahan, Iran; 2Department of Molecular Biotechnology at Cell Science Research Center, Royan Institute for Biotechnology, ACECR, Isfahan, Iran; 3Biology Department, School of Sciences, University of Isfahan, Isfahan, Iran; 4Department of Stem Cells and Developmental Biology at Cell Science Research Center, Royan Institute for Stem Cell Biology and Technology, ACECR, Tehran, Iran and 5Department of Developmental Biology, University of Science and Culture, ACECR, Tehran, Iran. Correspondence: Professor H Baharvand, Department of Stem Cells and Developmental Biology at Cell Science Research Center, Royan Institute for Stem Cell Biology and Technology, ACECR, Tehran, 19395-4644, Iran or Professor MH Nasr-Esfahani, Department of Molecular Biotechnology at Cell Science Research Center, Royan Institute for Biotechnology, ACECR, Khorasgan, Isfahan, 81593-58686, Iran. E-mail: [email protected] or [email protected] Received 24 September 2014; revised 7 March 2015; accepted 16 March 2015; accepted article preview online 1 April 2015; advance online publication, 23 April 2015 β-Globin expression by a non-viral integrating DNA K Dormiani et al 664 effect.11,19 An appealing feature of the integrase is that it prefers RESULTS transcriptionally active, open chromatin domains for targeted Design and functionality testing of the β-globin donor vector integration. This quality leads to a more uniform, higher rate of 15,21 We developed a phiC31 integrase-based plasmid to express wild- expression and allows the application of different promoters type β-globin allele using targeted integration. In order to to differentially regulate transgene expression in different 22 construct this non-viral vector, different genetic elements in lineages. A series of studies have highlighted the potential use defined orders and orientations were composed. The vector of phiC31 integrase-based vector systems in vivo by expressing contained three main parts: (i) a bacterial backbone that provided 22,23 therapeutic genes in animal models and ex vivo by propagation of the plasmid with structural stability, along with a transferring corrective DNAs into cultured cells such as stem cassette-encoding puromycin resistance gene (purr) to select 24–27 cells. In general, studies of kinetics and cellular localization, as stable clones, (ii) a tissue-specific β-globin expression unit for well as short half-life and biosafety of the enzyme have reported therapeutic benefit and (iii) an attB fragment to achieve a fair that phiC31 integrase can be considered an efficient, reliable tool frequency of site-specific chromosomal integration. In addition, 10,12,20,28 for gene addition therapy. three elements of human β-globin locus control region named In this study, we have utilized the phiC31 system to develop DNase I hypersensitive sites that included HS2, HS3 and HS4 (HSs) a site-specific integrating non-viral vector that contains the entire were added to the structure of the β-globin expression unit. These human β-globin gene (HBB), promoter and specific enhancers. elements contribute synergistically to produce controlled, tissue- Inclusion of two loxP sites also allows for removal of auxiliary specific globin expression during erythroid differentiation.28 sequences outside the therapeutic globin cassette by the Cre We also inserted two loxP sites that flanked the bacterial backbone recombinase subsequent to genomic integration. We evaluated and puromycin selection gene to delete these sequences by Cre expression of HBB by co-transfecting the constructed vector with recombinase subsequent to genomic integration of the vector the plasmid-encoding phiC31 integrase into K562 cells to monitor (Figure 1). After Escherichia coli DH5α was transformed with the HBB expression. Following isolation of clonal cell lines, by constructed vector, the cells were cultured in Luria Bertani agar continuous culture of the individual clones in the absence of medium contained ampicillin. Amplification of the pHBB/attB10 antibiotic selection, we observed remarkably stable, uniform vector was identified in a number of single colonies by the colony expression of the vector-encoded β-globin. PCR method. All colonies contained the vector as a high-copy

Figure 1. Schematic model of the pHBB/attB10 structure as a donor plasmid. In addition, this diagram illustrates the role of phiC31 integrase and Cre recombinase in site-speciﬁc integration of the therapeutic globin cassette in the target genome in the absence of bacterial backbone and purr. Initially, pHBB/attB10 that conveys the β-globin expression unit enters the target cells. The phiC31 integrase encoded by another plasmid mediates the ﬁrst recombination reaction between the attB site on pHBB/attB10 and pseudo attP site in the target genome that leads to chromosomal integration of the plasmid. After isolation of the recombinant clone, TAT-NLS-Cre recombinase is added to the culture medium. Transduced enzyme catalyzes the second recombination between two loxP sites as substrates. This reaction results in deletion of the bacterial backbone and purras a minicircle DNA from the structure of the integrated vector.

Gene Therapy (2015) 663 – 674 © 2015 Macmillan Publishers Limited β-Globin expression by a non-viral integrating DNA K Dormiani et al 665 number plasmid (data not shown). In addition, an endotoxin-free determine the site-specific genomic integration of pHBB/attB10 in vector was used to confirm the functionality of purr with individual clones (Figure 2a). In the first g-PCR as a preliminary test, transfecting CHO cells compared with the untransfected cells amplification of the expected bands (990 bp) confirmed the general – (negative control) in the presence of 8 μgml 1 puromycin for integration of the β-globin donor vector in each examined clone 10 days. Transfected cells proliferated and completely occupied (Figure 2b). The second round of g-PCR was designed to confirm the well, whereas all negative control CHO cells became nonviable SSR of the vector because of the phiC31 integrase activity to after 1 week (Supplementary Information). recombine attB in pHBB/attB10 and pseudo attPs in the host genome. In this step, we attempted to amplify the fragment in the Transfection and isolation of transgenic clones vector that contained the complete attB sequence. The results To test the functionality of the β-globin-expressing cassette, we established that phiC31 catalyzed the targeted homologous chose the K562 cell line whose erythroid properties made it an recombination between attB and pseudo attP elements, as the fi appropriate ex vivo model to monitor expression of the exogenous expected band of approximately 1270 bp was not ampli ed for the human globin genes.29 We used pHBB/attB10 as the donor vector majority of the transgenic clones. In addition, we used the genome along with pCMV/Int for transient expression of the phiC31 of one of the randomly integrated clones (RI), from integrase at 1:10 and 1:20 weight ratios. After approximately the previous attempts, as the positive control. In this clone, the 3 weeks of monitoring for the emergence of resistant clones integrated vector contained an intact attB fragment, therefore fi (Supplementary Information), they were expanded and genomic the 1270-bp band was ampli ed (Figure 2c). Thus, the clones DNA samples were harvested for genomic PCR (g-PCR) analysis of showed both 990 and 1270-bp PCR bands were considered as the SSR. However, the PCR results indicated random integration randomly integrated and clones with only 990-bp PCR band were of the vector because the attB fragment was not used for considered as targeted clones. According to the results of g-PCRs, recombination in any of the selected clones (Supplementary through all experimental cell lines (30 clones) 80% of them Information). Hence, to limit the number of clones that resulted (26 clones) were determined to be site-specific recombinant clones. from random integration we increased the amount of pCMV/Int in attempt to achieve an optimized relative ratio for successful SSR. Identification of sequence-specific insertion sites Finally, we used 62.5 ng pHBB/attB10 together with 2500 ng of Following detection of the site-specific recombinant clones, we pCMV/Int (1:40 w/w ratio) as the plasmid DNA for co-transfection. used a PCR-based rescue method to identify the vector integration fi Flow cytometry analysis indicated 12% transfection ef ciency for sites in different chromosomes.30 In this method, nested inverse fi K562 cells 1 day post-transfection. The speci city of recombination PCR (iPCR), two pairs of primers were designed to capture unknown fi with participation of the attB sequence was veri ed by means of genomic sequences adjacent to a known sequence of the g-PCR in 24 out of 30 analyzed transgenic clones. These results integrated vector using two-stage PCR. We documented indicated that the amount of integrase-encoding plasmid was an fi a number of integration events in stable transgenic K562 clones. important determinant in the ef ciency of SSR and it was In nine individual clonal populations of the transgenic lines, one dependent on the type of cells used. The resultant puromycin- integration site and in one clone, two independent integration sites resistant clones were isolated and expanded for further analysis. were identified (Table 1). The sequencing results of the genome– vector junction corroborated the results of the g-PCRs that led to SSR analysis using g-PCR selecting site-specific recombinant clones for additional analysis. Initially, the extracted transgenic genomes were used to identify 30 Further, a comparing of one recombination region on chromosome site-specific transfectants by g-PCR. We used three primers to 17 (q25.1) in four independent tested clones indicated that

Figure 2. g-PCR analyses allow verifying site-specific integration of pHBB/attB10 at the genome of isolated clones. (a) Schematic representation of the primer attachment locations on the vector for g-PCR. (b) Representative g-PCR1 analysis on independent K562 clones using GF1 and GR1 primers. The expected PCR product size, which confirmed donor plasmid integration, was 990 bp. (c) Representative g-PCR2 analysis on the same clones and one random integrated clone (RI) implementing GF1 and GR2 primers. Site-specific integration of the donor vector via attB site was confirmed when no 1270-bp band was observed. As a positive control, we used the genome of a randomly integrated clone (RI) in this PCR that showed the 1270-bp band. UT, untransfected K562 cells and M, marker (100 bp) plus DNA ladder (Thermo Scientific).

© 2015 Macmillan Publishers Limited Gene Therapy (2015) 663 – 674 β-Globin expression by a non-viral integrating DNA K Dormiani et al 666 recombination reactions between pHBB/attB10 and the pseudo clones using absolute and relative quantification method.31 For attP site were along with small deletions (⩽22 bp) in the host absolute quantification, the resultant standard curve showed a sequence surrounding the integrated site (Figure 3). However, no correlation coefficient of 0.995. PCR efficiency was determined to large deletions or insertions were detected in the studied clones. be 103% (slope = − 3.242) and the PCR product specificity was confirmed by melting curves analysis. With this method, the Estimation of transgene copy number transgene copy number of β-globin in transgenic clones was We used real-time quantitative PCR (qPCR) to determine the measured by comparing with three copies of endogenous gene transgene copy number present in the genome of experimented estimated in untransfected K562 cells. Using data of three

Table 1. Integration sites of pHBB/attB10 identiﬁed in the genome of 10 transfected clones

Chromosome Integration Nearest Clone Chromosome Ideogram (BLAT) band site gene

Repeat A1 chr6 (p21-1) TREML1 region

C2 chr4 (q12) Intergenic CHIC2

chr10 (q22.1) Intergenic PCDB1

C3 chr17 (q23.2) Intergenic TLK2

D6 chr17 (q25.1-1) Intron 4 GRB2

E7 chr17 (q23.2) Intergenic TLK2

D4 chr2 (p21) Intergenic RPS12P4

G9 chr17 (q25.1-2) Intergenic GALK1

C4 chr11 (p15.4) Intergenic RRM1

F6 chr1(p13.2) Intergenic RPL39P8

B3 chr9 (q22.33) Intergenic C9Orf156

Integration sites of the pHBB/attB10 were retrieved by nested iPCR using genomic DNA extracted from individual clonal populations. Integration site sequences were mapped to the reference human genome sequence and their corresponding chromosomal bands with the database at BLAT (http://genome. ucsc.edu). Also for the nearest gene to each insertion site, we analyzed the sequences with BLAST GeneBank (http://www.ncbi.nlm.nih.gov/BLAST/).

Figure 3. Multi-alignment of a pseudo attP site sequence on chromosome 17 (17q25.1) as the crossover points of the pHBB/attB10 vector. The hypothetical perfect recombination and several imperfect recombinations that show the loss of a number of nucleotides because of microdeletions in this locus from both vector and genome sequences in different individual clones.

Gene Therapy (2015) 663 – 674 © 2015 Macmillan Publishers Limited β-Globin expression by a non-viral integrating DNA K Dormiani et al 667 experimental clones for final analysis, we estimated one copy of preliminary test, we used immunofluorescent staining to visualize the donor vector per single cell for clones A1 and F6 and two these cells by the emission of a fluorescent light, which would copies for clone C3 (Table 2). Through the relative quantification verify β-globin peptide production. Immunostaining the original approach, the ratio between the copies of HBB as the target gene K562 cells and a single stably transfected clone (A1) with and CAPZA3 as a reference gene was calculated for three a β-globin-fluorescein isothiocyanate antibody showed green experimental clones. The generated standard curve for reference fluorescent light emitted by A1 cells (Figures 4a and b), while no gene showed a correlation coefficient of 0.989. From the slope, an green fluorescence was observed in K562 cells as negative control amplification efficiency of 96% (slope = − 3.432) was obtained for (Figure 4c). This clearly confirmed the β-globin chain production reference gene and the PCR product specificity was confirmed by following vector integration even after deletion of the bacterial analysis of melting curves. Using copy number ratio between backbone and puromycin selection cassette. In addition, we β-globin and the reference gene, the copy numbers of transgenic quantified the transcriptional activity and compared β-globin HBB were calculated for three experimental clones (Table 3). These expression levels in separate clones in terms of their insertion findings corroborated the results of absolute quantification sites. In this respect, reverse transcription qPCR (RT-qPCR) analysis β method. of -globin transcription level was achieved to verify foreign HBB gene activity. Untransfected K562 cells were used as the standard for quantification of human β-globin transcripts. Analysis of the Analysis of β-globin expression results offered apparent evidence that β-globin mRNA was We performed several experiments to show the production of efficiently transcribed in 12 tested clones at different levels β-globin chain in selected transgenic clones that were treated (Figure 4d). This was probably due to the insertion of the HBB with hemin (chemical inducer for globin chain synthesis). In the vector at different sites with different copy numbers. Immunoblot analysis of the cell extracts showed that the expression cassette fi Table 2. Absolute quantification of transgene copy number in K562 was ef cient and provided active expression of the HBB peptide in clones by qPCR the majority of experimental clones (Figure 4e). Based on these results, we chose three clones that had appropriate rates of Line Quantity Ct Transgene/ Total HBB Rounded expression for further analysis. (ng) value ± s.d.a endogene copy per transgenic HBB Besides, the relative β-globin expression in six randomly b ratio cell copy per cell integrated clones was measured using RT-qPCR 1 month post- transfection. In accordance with previous studies, our results K562 200 22.86 ± 0.12 0 3.00 0 40 23.26 ± 0.12 0 3.00 0 showed that random integration of the donor vector caused 8 25.40 ± 0.19 0 3.00 0 a variable expression level in tested clones. In comparison with expression level of site-specific clones (Figure 4d), randomly C3 200 22.18 ± 0.26 1.64 4.91 2 integrated clones showed a clear fluctuation in β-globin expres- 40 22.59 ± 0.09 1.63 4.90 2 ± sion level (Supplementary Information). It may depend on some 8 24.67 0.13 1.69 5.06 2 parameters such as the copy number of integrated plasmid, A1 200 22.33 ± 0.06 1.41 4.23 1 molecular architecture of the integration site and position effect. 40 22.70 ± 0.14 1.40 4.21 1 8 24.51 ± 0.05 1.39 4.19 1 Cre-mediated vector excision F6 200 22.46 ± 0.05 1.35 4.05 1 The ability of Cre-loxP system for site-specific deletion of bacterial 40 23.20 ± 0.16 1.37 4.09 1 backbone and selection markers in the HBB vector was confirmed ± 8 24.59 0.13 1.32 3.95 1 by PCR screening using appropriate primers. We expected that by Abbreviations: Ct, cycle threshold; HBB, β-globin gene; qPCR, utilizing these primers, the fragment between two loxP sites quantitative PCR. aData are the mean of Ct values ± s.d.; all reactions were (3250 bp) was PCR amplified but after TAT-NLS-Cre treatment, the carried out in triplicate. bFor estimation of the absolute transgene copy loxP-flanked sequences were deleted from the vector and hence number per cell for each clone, the copy number of HBB was determined they were not detected (Figures 5a and b). Deletion of these in untransfected K562 by comparison with the normal human genome and sequences from the vector was confirmed in vitro using PCR then it was subtracted from total rounded copies of HBB in each related (Figure 5c). To manipulate the genomic integrated vector with cell line. modified Cre recombinase, we added the enzyme to the culture

Table 3. Relative quantiﬁcation of transgene copy number in K562 clones by qPCR

a a b c Line Quantity (ng) ΔCt cal (CAP-HBB) ΔCt test (CAP-HBB) Ratio Total HBB copy per cell Rounded transgenic HBB copy per cell

C3 150 − 2.24 − 0.88 2.43 4.86 2 50 − 0.99 0.42 2.62 5.24 2 10 − 0.44 0.93 2.59 5.19 2 A1 150 − 0.78 0.158 1.89 3.78 1 50 − 0.97 0.02 1.95 3.89 1 10 − 0.96 0.06 1.99 3.98 1 F6 150 − 1.84 − 0.70 2.09 4.18 1 50 − 1.32 − 0.18 2.15 4.29 1 10 − 1.46 − 0.44 1.96 3.91 1 Abbreviations: Cal, calibrator; CAP, CAPZA3 gene; ΔCt, Ct of reference gene–Ct of target gene; HBB, β-globin gene; qPCR, quantitative PCR. aData are the mean of ΔCt values. bThe ratio was calculated using Pfafﬂ’s equation. cFor estimation of the transgene copy number per cell, the copy number of HBB was determined in untransfected K562 (three copies) by comparison with the normal human genome and then it was subtracted from total copies of HBB in each related cell line.

Figure 4. Evaluation of transgenic β-globin biosynthesis in the isolated transgenic K562 clones. (a) The active expression of β-globin in one of the transfected clones (A1) indicated by direct immunoﬂuorescent staining of K562 cells using ﬂuorescein isothiocyanate (FITC)-conjugated anti β-globin before cellular treatment with Cre recombinase. (b) Staining the same cells after exclusion of the bacterial backbone and purr through Cre recombinase. (c) Untransfected K562 cells. Scale bar = 500 μm. (d) HBB expression assessed in 12 recombinant clones by RT-qPCR compared with untransfected K562 cells (UT) and normalized to endogenous GAPDH expression. Error bars denote s.e.m. for three independent repeats in each experiment. (e) Immunoblot analysis of 30 μg total cell protein extracts from different transgenic clones compared with untransfected K562 cells (UT) for detection of human β-globin peptide. The β-globin band from transgenic clone cells are shown above and GAPDH as the control are shown below the lanes. The results show that the majority of experimented clones actively expressed a marked level of β-globin chain over the UT cells.

medium of one puromycin-resistant clone and determined the (Figure 6a). On the other hand, following deletion of the bacterial sensitivity of the treated cells to puromycin within 21-day culture backbone and puromycin selection cassette, human β-globin (negative selection). Our results showed that the cells of 14 (63%) expressed by the integrated vector was tracked for an additional out of 22 screened clones were killed through the deletion of 4 months with RT-qPCR. These results indicated that pHBB/attB10 loxP-flanked sequences. These findings were further confirmed by provided long-term in vitro expression of HBB. After His-TAT-NLS- g-PCR.32 For this, the genomes of three negatively selected clones Cre transduction, globin expression was continued by integrated were used as templates in addition to the genome of the same cassette for additional 4 months (Figure 6b). The ability of the β clone before Cre treatment (positive control). As depicted in pHBB/attB10 plasmid to stably produce the human -globin fl Figure 5d, the plasmid fragment that encompassed the bacterial peptide was also evaluated by uorescence-activated cell sorting backbone and purr were not amplified in the selected clones, analysis. The cells of a representative recombinant clone were whereas the 3250-bp band was amplified using genomic DNA immunostained over untransfected control cells and expression of before use of Cre recombinase. HBB was monitored during 8 months at different time points in the absence of puromycin (Figure 6c). The first and second histograms depicted HBB expression at 2 and 4 months post- β Long-term expression analysis of -globin transfection. Following deletion of the intervening sequence We chose three independent clones that had high, medium and flanked by loxP sites, HBB expression was assessed at 6 and low expression levels of β-globin to monitor their β-globin 8 months post-transfection (2 and 4 months after exclusion of production over 6-month continuous culture. A uniform and auxiliary sequences). These findings indicated that nearly 98% of sustained expression from pHBB/attB10 was observed in each all the transfected cells maintained active expression of the clone during this time period in the absence of antibiotic selection β-chain in a continuous culture. Changes in HBB expression levels

Figure 5. TAT-NLS-Cre treatments of transgenic A1 cells for site-specific deletion of the auxiliary sequences between two loxP sites. (a) Schematic diagram of the chromosomal integrated vector that shows loxP sites and the relative locations for attachment of designed primers before Cre-mediated recombination. (b) The chromosomal integrated vector after Cre recombinase reaction to delete the intervening sequence flanked by loxP sites and new generated location of the primers. The position and distance of the β-globin promoter (P) relative to integration borders from both directions (attL and attR) is demonstrated. (c) PCR on pHBB/attB10 before and after in vitro treating with Cre recombinase. Note that the Cre-treated plasmid did not yield any product in PCR. (d) The same PCR reactions carried out on the genome extracted from untransfected K562 cells (UT), A1 and three representative clones (A1-I, -II and -III) isolated following Cre transduction and negative selection. The results have indicated that Cre enzyme excised the fragment between two loxPs in the genomic integrated vector. White asterisks indicate nonspecific bands artificially amplified after treatment of transgenic clones with Cre enzyme. M, marker 1 kb DNA Ladder (Thermo Scientific). during this period were also estimated using antibody mean the efficiency of the phiC31 integrase-based non-viral system in fluorescence intensity values from three different experiments for order to achieve high-level and continuous production of β-globin each sample. As demonstrated in Table 4, at four time points chain in an erythroid lineage cell line. It has been shown that in before and after removal of auxiliary sequences, mean fluores- addition to complementary DNA, other genomic sequences are cence intensity exhibited no significant differences in involved in the control of transgene expression. Therefore, in β-globin peptide expression. On the other hand, in accordance contrast to intronless small transgene or complementary DNA, with the results of RT-qPCR that displayed a gradual decrease in genomic constructs, which contain complete transgene HBB expression levels within the first 4 months of post- sequences, a native promoter, locus control regions and other transfection, mean fluorescence intensity values also showed a elements located on the 5´ or 3´ end of the gene or within its simultaneous decline in transgene expression. These data also introns can lead to position-independent and persistent expres- signified that following the use of Cre recombinase and exclusion sion of the transgene in comparable levels to the endogenous of bacterial backbone and selection sequences, expression level of gene, at appropriate times in a specific tissue.34,35 As a result, the HBB increased in a similar manner to the primary months after β-globin expression cassette used in the pHBB/attB10 vector was transfection (Figure 6b). equipped with an extended human β-globin promoter length, complete HBB with 5´ and 3´ untranslated regions, a specific polyadenylation signal and proper large genomic regulatory DISCUSSION sequences that adapted to erythroid lineage and expression Some of outstanding features of phiC31 integrase makes it an needs. The β-globin promoter, as a tissue-specific promoter attractive tool for application in biomedical fields such as gene ensures basal expression of β-globin and has a low probability of therapy,15,16 in vitro4,7 and in vivo gene delivery methods.8,9 They activating oncogenes in hematopoietic progenitors by virtue of its include precise integration of a single transgene copy that late-stage erythroid specificity.36 One of the most important minimizes off-targeting effects and preferred insertion at active factors that affects transgene expression rate and duration is the transcription units. Furthermore, the system mediates stable, topological characteristics of the promoter in the vector structure, permanent and site-restricted transgene expression in various particularly when it is integrated into the host chromosomes. types of target cells. On the other hand, transgene silencing has Watanabe et al.37 have shown that the direction of expression been reported to occur for certain conventional genes such as cassette relative to the attB site can negatively impact on not only β-globin as the result of negative position effects through local the strength of the vector transcription, but also transgene regulatory sequences around the integration sites.17,33 Accord- silencing because of the position effect. In other words, the higher ingly, the main goal of this study was to construct and evaluate the relative distance of the promoter from the integration site,

© 2015 Macmillan Publishers Limited Gene Therapy (2015) 663 – 674 β-Globin expression by a non-viral integrating DNA K Dormiani et al 670 the higher and more continuous the transgene expression will expression, HSs elements as position-independent enhancers occur in the cells of interest. have been shown to preserve the structure of active chromatin Given the facts in this report, we designed our expression and protect the integration environment from condensation in cassette in which the β-globin promoter was located distal to the order to prevent silencing of the β-globin transgene.39–41 β-Globin attB sequence with an approximate distance of 2400 bp from the is an intron-dependent gene and the presence of introns is critical crossover point via the fragment that contained the HBB. for the stabilization and cytoplasmic translocation of mRNA.42,43 Conversely, following site-restricted vector integration into the Unlike the lentiviral vector in which a part of the second intron of host chromosome and exclusion of the bacterial backbone and the HBB must be deleted to allow stable proviral transmission of purr, we observed a DNA fragment that spanned the complete globin expression,44 we have utilized a complete intact HBB in length of three enhancer elements (HSs) with about 3300 bp pHBB/attB10 that contained two downstream enhancers, intron located between the promoter and integration border (Figure 5b). 2 and the 3´ untranslated region (3´proximal enhancer), both As β-globin locus control region activity is orientation important for the augmentation of HBB production.45 By taking dependent,38 therefore we have subcloned the HSs elements into these points into our construction design, we showed that the pHBB/attB10 in the same order as they appeared in the human vector actively expressed HBB in the 10 selected clones as genome. In addition to their role in enhancing erythroid-speciﬁc conﬁrmed by RT-qPCR and immunoblotting. Our experiments

Figure 6. Monitoring the duration of transgenic HBB expression. (a) Three independent transfected clones that expressed different amounts of β-globin (C3, A1 and F6) were monitored by RT-qPCR for 6 months. Significant expression of β-globin was evident between the tested clones and untransfected K562 cells (UT) at each indicated time point (Po0.001). The expression level of HBB did not significantly change during the 6-month monitoring period in each clone. (b) RT-qPCR performed on total RNA extracted from A1 cells. As shown, expression during the first 4 months gradually decreased, whereas after Cre treatment the cells showed an increment of the β-globin expression levels (an additive and subtractive expression wave). Expression of β-globin was estimated relative to GAPDH as an endogenous control in transfected cells over the untransfected cells. Error bars signify s.e.m.. There were three independent repeats for each experiment. Different letters denote significant difference between expression levels in different months at Po0.05. (c) Representative fluorescence-activated cell sorting (FACS) analysis for the expression of human β-globin in cells of the A1 clone. Horizontal bars indicate the percentage of β-globin- expressing cells (gray histogram, left) over the isotype control (dark histogram, right). The first and second histograms depict HBB expression at 2 and 4 months post-transfection. After exclusion of auxiliary sequences from the integrated vector using Cre recombinase, the HBB expression was assessed at 6 and 8 months post-transfection (2 and 4 months after Cre treatment of A1 cells). The findings showed that most transfected cells, before and after Cre excision of bacterial backbone and purr, maintained the pHBB/attB10 vector that actively expressed the hemoglobin β-chain in a continuous culture without puromycin. W/O, without; BB, bacterial backbone and m, month.

Table 4. Human β-globin levels during 8 months of continuous culture

Flow cytometric analysis post-transfection (month) 2468

MFI 34.74 ± 0.53 29.22 ± 0.64 37.77 ± 0.33 35.89 ± 0.62 Abbreviations: HBB, β-globin gene; MFI, mean ﬂuorescence intensity. Changes in HBB expression levels during continuous culture of A1 clone cells in different time periods were indicated using antibody MFI. Data are mean ± s.e.m.; all reactions were carried out in triplicate.

Table 5. List of primers used for construction of pHBB/attB10 vector

Ampliﬁed segment Expected size (bp) Primer name Sequence (5´ to 3´)

Ampr 1893 KpnI F-loxp/amp GGTACCATAACTTCGTATAGCATACATTATACGAAGTTATTGTGCGCGGAACCCCTATTTG XhoI R-amp CTCGAGTCACTCAAAGGCGGTAATACG Purr 1362 PvuI F-pur CGATCGTTCACTCGAGTGTGTCAGTTAG NheI R-pur GGATCCCGAGGCAGTGAAAAAAATGC attB 332 NheIF-attB GCTAGCCGGCTTATGTAGGTCACGGTCTCG XbaIR-attB/loxp GGATCCATAACTTCGTATAATGTATGCTATACGAAGTTATGCTTGCCCGCCGTGACCGTCAAGAAC HBB 3112 BglII F-HBB AGATCTAGACATAATTTATTAGCATGCATGAGC NheI R-HBB GCTAGCACTCATATTTTATTTCCAGAATCTAGCATCTAC HS4 1112 KpnI F-HS4 GGTACCAGGATCCCTTGAGCTCAGGAG XbaI R-HS4 ATCTAGACTGTCTAGTGTATGTGCAGTGAGC HS3 1312 XbaI F-HS3 ATCTAGACTGGATCCACTTGCCCAGTGTTCTTC NdeI R-HS3 CATATGGCTTTCATTAAAAAAAGT CTAACCAGCTGC HS2 918 NdeI F-HS2 CATATGAGCACTTTGGGAGGCTGAGGTG SalI BglII R-HS2 GTCGACAGATCTCATTCTATGGCAATTGATAACAAC Abbreviations: bp, base pair; F, forward reverse; R, reverse primers. Specific restriction site(s) in each primer are underlined. showed remarkable, long-term expression of HBB in three recombination of the lox-modified substrate, but especially for the different transgenic clones with low-frequency variations during cytotoxic effect of the enzyme on transduced cells. The optimum 6 months continuous culture of erythroid clones without antibiotic concentration is dependent on the cell type. For K562 clones, we selection. The distance between the HBB promoter and integra- used the maximum safe concentration of 2.5 μM for 16 h. By tion points from both directions, the optimized combination of increasing the concentrations and duration of incubation, an distal and proximal transcriptional control elements, and the apparent cytotoxic effect was detected. Notably after deletion of ability of phiC31 integrase to insert donor vector at pseudo attP the mentioned sequences, β-globin expression level was quanti- sites in transcriptionally active areas in the host genome resulted fied in one of the Cre recombined clones for an additional in continuous, high-level expression of β-globin. The low copies of 4 months. The obtained results showed that expression of β- integrated transgene per cell as estimated in our three globin in this clone was stable and comparable with the first experimented clones also considered an advantage as it reduced 4 months post-transfection. the chances of insertional mutagenesis and paved the way for safe These data provide reasonable evidences that pHBB/attB10 as a integration. These qualities could result in therapeutic hemoglobin non-viral phiC31-mediated vector, when specifically integrated production in the target tissues. According to previous reports, into the human genome through phiC31 integrase, can support nucleotide changes in the site of recombination by phiC31 efficient and prolonged tissue-specific expression of HBB with low integrase were usual events and not restricted to special sites or variations in K562 cells as an erythroid model. Consequently, the chromosomes.46,47 Based on the majority of reports, these promising results of β-globin expression by pHBB/attB10 in K562 changes were minor such as simple deletions of few numbers of cells encourage us to evaluate the capability of this integrating nucleotides. On the other hand, other researchers documented vector for ex vivo-targeted gene delivery of a working copy of the that approximately 10% of phiC31-mediated integration events in HBB to stably correct globin deficiency in hematopoietic stem cells cultured cells may be aberrant contain chromosome rearrange- or induced pluripotent stem cells derived from patients with ments or large deletions.47,48 However, such events have not been β-globinopathies in our future studies. documented in in vivo studies with phiC31 integrase, nor were any tumors or other adverse events documented in any type of organism treated with the integrase.15 Although sequencing the MATERIALS AND METHODS genome–vector junction areas in our experimental clones resulted Donor vector construction and functionality evaluation in no major deletions, we detected some short deletions. Previous We amplified different fragments to create the donor plasmid with studies have proven that the bacterial backbone and selection DreamTaq DNA polymerase (Thermo Scientific, Waltham, MA, USA) using units consist of various harmful consequences: (i) methylation of specific primers (Table 5). All restriction enzymes were purchased from neighboring bacterial backbone results in transcriptional silencing Thermo Scientific. For vector construction, initially the SV40 promoter and or misexpression level of eukaryotic cassette (such as β-globin downstream purr were amplified from pTRE2pur (Clontech Laboratories, cassette),49–51 (ii) transfer of antibiotic resistance genes into the Mountain View, CA, USA) and cloned into pBudCE4.1 (Invitrogen, Carlsbad, cells for transplantation may lead to entry of these genes to CA, USA) using PvuI and BamHI enzymes to produce pBud/pur. Next, the first loxP sequence, pUC origin of replication and bla promoter along with bacteria in the environment and spread antibiotic-resistant r the ampicillin-resistant gene (amp ) were amplified from pcDNA3.1/Zeo (+) pathogens, (iii) recognition of prokaryotic plasmid backbone fi 52 (Invitrogen) and cloned into the pTZ57RT vector (Thermo Scienti c) by T/A by Toll-like receptor 9 stimulates innate immunity and (iv) cloning. These fragments were subcloned into pBud/pur using XhoI and r expression of pur enzyme as a foreign protein in the recipient KpnI enzymes to generate pBud/pur/Amp/loxP. The vector was trans- could stimulate immunogenic responses.53 formed into E. coli One Shot DH5α-T1 (Invitrogen) in the presence of Accordingly, these auxiliary fragments were deleted from the ampicillin (100 μgml–1). Three bacterial colonies were selected and the integrated vector with recombinant TAT-NLS-Cre.54 The use of this modified plasmid was extracted and confirmed by restriction digestion type of cell permeant Cre recombinase ensured that ectopic and sequencing. On the other hand, the complete sequence of the human β-globin promoter, the HBB and its 5´ and 3´ untranslated regions were chromosomal integration of plasmid-encoding Cre as foreign DNA fi would not occur.4 Therefore, the use of the TAT-NLS-Cre ampli ed using blood-extracted human genome of a healthy volunteer via the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). This large recombinase protein appears to be a powerful tool for rapid fi fl 3100-bp fragment was cloned into the pTZ57RT vector and subsequently and ef cient site deletion of the inserted sequence anked by subcloned into the pBud/pur/Amp/loxP vector where the first loxP loxP sites in the target cell. However, according to our experiments sequence was located upstream of the β-globin promoter (pBud/pur/ the concentration and incubation time with Cre were Amp/loxP/HBB). Next, three β-globin enhancer sequences, HS2 (900 bp), important variables not only for the successful, highly efficient HS3 (1300 bp) and HS4 (1100 bp) were amplified from a normal human

© 2015 Macmillan Publishers Limited Gene Therapy (2015) 663 – 674 β-Globin expression by a non-viral integrating DNA K Dormiani et al 672 genome and separately cloned into the pTZ57RT vector. The HS4 fragment was purified and eluted in 25 μl of elution buffer using the QIAquick PCR was subcloned upstream of pTZ/HS3, which produced pTZ/HS4/HS3 by Purification Kit (Qiagen). For self-circularization, 250 ng of the digested using KpnI and XhoI enzymes. HS2 was subcloned downstream of HS3 in genomic DNA was used for ligation using 10 U of T4 DNA ligase (Thermo pTZ/HS4/HS3 by NdeI and SalI digestion, which produced pTZ/HS4/HS3/ Scientific) in 100 μl of 1X T4 ligase buffer overnight at 16 °C. Following heat HS2. The HS432 fragment was double digested by KpnI and NheI enzymes inactivation of the ligase, the ligation products were ethanol precipitated and subcloned upstream of the β-globin promoter in the pBud/pur/Amp/ and resuspended in 10 μl of Tris buffer (pH 8.2). Using eluted DNA as the loxP/HBB plasmid. The attB fragment was amplified from pDB2 with a template, the first iPCR was performed in a 25 μl reaction volume using reverse primer that contained loxP at its 5´ end. After T/A cloning into recombinant Taq DNA polymerase (Thermo Scientific) with IF1 (5′-TCACGG pTZ57RT, the attB/loxP fragment was subcloned downstream of the HBB TGAGCACGGGAC-3′) and IR1 (5′-CACCTCGACCCGTTCATCATG-3′) as the where this loxP was unidirectional with the first loxP sequence and located first primer pairs. After 3-min denaturation at 95 °C, the DNA was amplified r after pur (Figure 1). We verified all fragments by sequencing. The final for 25 cycles at 95 °C for 40 s, 58 °C for 40 s, 72 °C for 3 min, with a final constructed vector, pHBB/attB10, was transformed into competent E. coli extension for 10 min at 72 °C. For the second nested iPCR, 0.5 μl of the first α fi DH5 and puri ed by the EndoFree Plasmid Maxi Kit (Qiagen). To elucidate iPCR product served as the template. The reaction was carried out for 35 the functionality of the puromycin resistance open reading frame, we cycles using the appropriate nested primer pairs (IF2: 5′-AGCGTCAGCGGGT transfected this vector into CHO-K1 cells (CCL-61 ATCC, Manassas, VA, USA) TCTTG-3′ and IR2: 5′-CAGATGGGTGAGGTGGAG-3′) with the same cycle cultured in a six-well plate using Lipofectamine 2000 (Invitrogen) based on parameters in final 25 μl volumes, with the exception of the annealing ’ the manufacturer s protocol. At 2 days post-transfection, puromycin was temperature, which was 62 °C for 40 s. We separated all suspected DNA added to the wells that contained transfected and untransfected control bands that were amplified using the genome of individual transfected cells at a final concentration of 8 μgml–1. The medium was replenished – clones but not with untransfected parental cells. These DNA samples were every 2 3 days and the cytotoxic effect of the antibiotic was investigated purified by a QIAquick Gel Extraction Kit (Qiagen), T/A cloned into the in both transfected and untransfected cells. pTZ57RT vector and sequenced with suitable universal primers. To map chromosomal integration sites of the donor vector, DNA sequences of each Human erythroid cell culture and transfection band were aligned with the reference human genome (University of The human chronic myeloid leukemia cell line K562 (ATCC CCL-243) was California Santa Cruz, UCSC hg18) using the database at BLAT (http:// grown and maintained in RPMI 1640 medium, GlutaMAX (1X) supplemen- genome.ucsc.edu). Also the sequences were analyzed with BLAST ted with 10% fetal bovine serum, 100 units ml–1 penicillin and 100 μgml–1 GeneBank (http://www.ncbi.nlm.nih.gov/BLAST/) to find the nearest gene streptomycin sulfate (all from Gibco, Rockville, MD, USA) at 37 °C and 6% to each insertion site. CO2. After culturing for 24 h in a six-well plate, K562 cells were stably co- transfected with plasmid DNAs that included pHBB/attB10 as the donor Determination of transgene copy number using qPCR vector and an integrase-expressing plasmid, pCMV/Int. Transfection was β carried out at different weight ratios using Lipofectamine LTX (Invitrogen) The absolute copy number of -globin transgene in puromycin-resistant ’ fi clones was assessed using the standard curve-based quantification according to the manufacturer s instructions. Transfection ef ciency of fi K562 cells was measured by fluorescence-activated cell sorting analysis method. qPCR was performed with speci c primers for human HBB (AcF: ′ ′ ′ 24 h post-transfection (according to method mentioned in Immunostain- 5 -CTTATGGTGCTTCTGGCTCTG-3 and AcR: 5 -ACTCATATTTTATTTCCAGAAT ′ ing and flow cytometric analysis). Transfected cells were split into three CTAGC-3 ) at the following parameters: 95 °C for 30 s, followed by 40 cycles groups in 60 mm plates that contained 3 ml of complete growth medium at 95 °C for 10 s, 60 °C for 30 s and 72 °C for 30 s. All reactions were carried and then they were incubated for an additional 48 h. The cells were out in triplicate in a StepOnePlus Real-Time PCR System (Applied subsequently chosen in the medium that contained puromycin (Sigma- Biosystems, Foster City, CA, USA) using a SYBR Green Master Mix (Takara) Aldrich, Steinheim, Germany) at a final concentration of 3 μgml–1. After a in a 10 μl final volume. Typically, the linearized pHBB/attB10 vector was 3-week incubation period in the presence of puromycin, antibiotic- used as the template in a fivefold serial dilution series at six points –1 resistant clones emerged. Individual transgenic colonies were isolated (10–0.0032 ng μl ) to generate the external standard curve. The mean of using pipette tips and transferred into a 96-well plate. To expand these the resultant Ct values for each input of template was plotted as a function clones, each surviving clone was transferred into one well of a 12-well of the log quantity. To determine the endogenous β-globin copy number plate and subsequently expanded in T-25 flasks. In this stage, the antibiotic in untransfected cells, its genome compared with the normal human was removed from the culture medium and 30 individual clones were genome as the standard. It was shown that three copies of HBB are present selected for further analysis. in untransfected cells over the two copies in the normal diploid genome. To estimate transgene copy number, two series of PCR reactions were set using three dilutions (200, 40 and 8 ng μl–1) of the genomic DNA as the Genomic DNA analysis template. One set of the reactions was performed to amplify the 6 Approximately 1 × 10 cells of each puromycin-resistant clone were endogenous and transgenic HBBs in each transgenic cell line and the isolated to extract genomic DNA using the DNeasy Blood and Tissue Kit. other was carried out to amplify the endogenous HBB in untransfected We performed two rounds of PCR to verify SSR of the pHBB/attB10 vector cells. The DNA quantity of these samples was calculated by interpolation fi in the genome of the target cells. The rst round was conducted to prove using the mean of the resultant Ct values; the total copy number in each successful integration of the vector into the genome of each selected sample was calculated using an online calculator (http://cels.uri.edu/gsc/ ′ clone. We used GF1 as the forward primer (5 -CTTGAGCATCTGGATTCTGC cndna.html). The absolute copy number of transgenic HBB per one cell of ′ ′ -3 ) annealed to the HBB gene sequence and GR1 as the reverse primer (5 - experimental clones was estimated by subtracting three copies per ′ GCTTGCCCGCCGTGACCGTC-3 ) attached to the attB sequence upstream of genome of untransfected K562 cells from total copies of β-globin in each the integration site. In the second PCR round, the same forward primer and fi ′ ′ relevant cell line. To further con rm our result, we also used the relative GR2 reverse primer (5 -CTTCGAGACCGTGACCTAC-3 ) that attached down- copy number method to assess the β-globin transgene. For this purpose, stream to attB, after the integration site were used to prove contribution human CAPZA3 (NC_000012.12) was selected as the reference gene and by the attB fragment in SSR by the phiC31 integrase. In both PCRs, 150 ng an internal standard curve was plotted using fivefold serial dilutions of of genomic DNA was used as the template along with Ex Taq polymerase – normal human genome at six points (250–0.08 ng μl 1). qPCR was (Takara, Tokyo, Japan) in a 25 μl final volume. Amplification was performed performed with specific primers for the reference gene (RcF: 5′- CTGCCTGC using the following cycles: 95 °C for 2 min, 30 cycles at 95 °C for 1 min, TTTCATGGATAGAC-3′ and RcR: 5′- TACTCTTTCCTTGTCCTTCCTG-3′) at the 62 °C for 45 s and 72 °C for 2 min, followed by incubation at 72 °C μ same condition. Two series of PCR reactions were set using three dilutions for 10 min. We analyzed 5 l of each PCR amplicon by agarose gel μ –1 electrophoresis. (150, 50 and 10 ng l ) of the genomic DNA samples as the template. One set of the reactions was performed to amplify the β-globin and reference gene in transgenic genomes (test samples) and the other was carried out Rescue and analysis of integration sites to amplify the reference and endogenous HBB using the genomic DNA of After identification of site-specific recombinant clones, we attempted to normal human fibroblasts (calibrator sample). Thereafter, the relative copy ascertain the precise genomic locations of the donor plasmid by nested number of HBB to the copy number of reference gene was calculated in iPCR. Genomic DNAs were extracted from the transfected cells and 3 μgof each experimental clone. As two copies of CAPZA3 are present in normal the purified genomic DNA was digested with NheI, SpeI and XbaI enzymes human genome, total copies of HBB were estimated. By subtracting three in a final volume of 80 μl at 37 °C overnight. The digested genomic DNAs copies of endogenous β-globin per genome of untransfected K562 cells

Gene Therapy (2015) 663 – 674 © 2015 Macmillan Publishers Limited β-Globin expression by a non-viral integrating DNA K Dormiani et al 673 from estimated total copies, the transgene copy number in each clone was functionality of TAT-NLS-Cre (Excellgen, Rockville, MD, USA) on our plasmid calculated. before using the enzyme in the culture medium. We added 1 μg of pHBB/ attB10, 1 μl of Cre reaction buffer (500 mM Tris.HCl, 330 mM NaCl and μ –1 RNA extraction and quantification of β-globin expression 100 mM MgCl2 at pH 7.5) and 2 l of the enzyme (5 mg ml )toa microcentrifuge tube. The total volume of the reaction mixture was After expansion of each recombinant clone, we collected approximately adjusted to 10 μl using deionized water. The content of the tube was 1×106 cells by centrifugation. Cells were lysed in 750 μl of TRI reagent ’ mixed gently and the tube incubated at 37 °C for 30 min. Next, the vector (Sigma-Aldrich) according to the manufacturer s instructions. Total RNA DNA was purified using an ethanol precipitation. To establish the loss of was extracted from approximately one million cells of each individual the bacterial backbone and puromycin resistance open reading frame from clone on day 30 post-transfection, quantified and stored at –70 °C. For RT- the vector structure, we used 120 ng of DNA from the intact vector and qPCR, complementary DNAs were synthesized using the RevertAid 120 ng of the purified DNA vector after treating with TAT-NLS-Cre as a Premium First Strand cDNA Synthesis Kit (Thermo Scientific) with random template in a PCR reaction. A forward primer (FHS: 5′‐CGTGA hexamer primers. RT-qPCR was performed with primers specific for the β ′‐ ′ GGCAAGGTTTCACTCTG-3′) was designed to attach the upstream of the human -globin transcripts (F-HBB: 5 CACCTTTGCCACACTGAG-3 and ′‐ R-HBB: 5′‐GCCACCACTTTCTGATAGG-3′) designed by Beacon Designer 7.9 HS4 enhancer and the reverse primer (R-pur: 5 CGAGGCAGTG ′ r software (Premier Biosoft International, Palo Alto, CA, USA). All reactions AAAAAAATGC-3 ) attached to the 3´ end of pur . PCR was performed in a μ fi were carried out in triplicate in a StepOnePlus Real-Time PCR System volume of 25 l using Ex Taq polymerase. To assist with ampli cation of the (Applied Biosystems) using the SYBR Green Master Mix (Takara). Vector- fragment of pHBB/attB10 that encompassed the bacterial backbone and r μ μ derived β-globin expression level was calculated by the comparative Ct pur , 0.5 l DMSO and 1 l betaine were added to the reaction mixture. PCR method using glyceraldehydes-3-phosphate dehydrogenase (GAPDH)as cycles were as follows: after 3 min of denaturation at 95 °C, the DNA was the reference gene for normalization. amplified by 35 cycles at 95 °C for 30 s, 55 °C for 30 s, 72 °C for 3 min, with a final extension for 10 min at 72 °C. PCR amplicons were analyzed by r fl agarose gel electrophoresis. One primer is attached to the 3´ end of pur Immunostaining and ow cytometric analysis and the second one was designed to detect the end of the HS4 sequence. Human β-globin expression was visualized by intracellular staining of 6 approximately 1.5 × 10 transfected K562 cells from each clone, which Ex vivo Cre transduction into a recombinant K562 clone. Approximately were treated with 30 μM hemin (Sigma-Aldrich) for 48 h. The cells were 1×105 cells from one of the puromycin-resistant clones (A1) were washed with phosphate-buffered saline (PBS) at 1800 r.p.m. for 5 min and suspended in 0.5 ml of RPMI 1640 medium GlutaMAX using a 24-well resuspended in 300 μl of 4% paraformaldehyde solution for 20 min at 4 °C plate supplemented with 5% fetal bovine serum. To transduce TAT-NLS- μ in the dark. Fixed cells were washed three times in 500 l of PBS buffer that Cre into the cells, 12 μl of TAT-NLS Cre with a total concentration of contained 0.05% Tween 20 (PBS/Tween). For permeabilization, cells were 5mgml–1 was added to the culture medium to give a final concentration μ incubated with 300 l of 0.2% Triton X-100 in PBS for 10 min at room of 2.5 μM and the plate was incubated for 16 h (transduction period) at 37 °C temperature. After washing three times with the PBS/Tween solution, in a humidified incubator and 6% CO . Next, Cre-transduced cells were fi μ 2 nonspeci c binding sites were blocked by treating K562 cells with 300 lof washed with PBS, cultured for 2 days in normal growth medium with 10% PBS buffer that contained 1% bovine serum albumin (1% bovine serum fetal bovine serum and then cloned by the limiting dilution method. albumin/PBS). The cells were then collected by centrifugation and Following expansion of each individual clone, the cells were divided into resuspended in 100 μlofβ-globin-fluorescein isothiocyanate:sc-21757 for two parts. One part was used to corroborate deletion of bacterial backbone transfected clones and normal mouse IgG1:sc-2866 for untransfected cells and purr by negative selection. In this method, puromycin was added to as the isotype control (both from Santa Cruz Biotechnology, Santa Cruz, TX, – the culture medium of each clone at a final concentration of 3 μgml 1 and USA) diluted 1:10 in a 1% bovine serum albumin/PBS solution. Cell the cell morphology was monitored for 21 days. The cells that lost the suspensions were incubated for 1 h at 37 °C and protected from light throughout staining. After incubation, stained cells were washed three bacterial backbone and puromycin resistance open reading frame times with PBS/Tween and resuspended in 30 μl of PBS. Approximately between two loxPs became nonviable in the presence of puromycin. For one-third of the cell suspension was placed on a chamber slide previously the second part, the cells were used for genomic analysis as an additional fi treated with 0.01% poly-l-lysine (Sigma-Aldrich), drained, fixed with 4% support to con rm deletion of these sequences. For this purpose, genomic formalin (Sigma-Aldrich), and mounted for examination. The slides were DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen) and a analyzed with an Olympus BX51 fluorescence microscope and the images PCR reaction was performed with the same primers as the in vitro acquired with an Olympus DP70 digital camera (Olympus, Tokyo, Japan). experiment under similar conditions except for the annealing temperature, The remaining cells were transferred to 500 μl PBS plus 2% fetal calf serum which was 58 °C for 30 s. The PCR products were analyzed by agarose gel as staining buffer. Data were acquired in a BD FACSCalibur and analyzed electrophoresis. with Cell Quest Pro software (Becton Dickinson, Mountain View, CA, USA). Statistical analysis Immunoblot analysis The obtained means ± s.e.m. from different groups of data were analyzed Cells were isolated by centrifugation at 1800 r.p.m. for 5 min. These cells by SPSS (Version 21, Chicago, IL, USA). Statistical significance between the were subsequently lysed with TRI reagent (Sigma-Aldrich) according to the groups was analyzed by the independent t-test and one-way analysis of manufacturer’s protocol. A solubilized protein fraction of each sample variance, which were considered to be significant at Po0.05. (30 μg) was subjected to 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and eletrophoretically transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA) using a Mini-PROTEAN CONFLICT OF INTEREST Tetra cell transfer system (Bio-Rad). After blocking overnight with 10% The authors declare no conflict of interest. (w/v) non-fat dried milk (Merck, Darmstadt, Germany) in PBS buffer, we labeled the membranes with primary mouse monoclonal fluorescein isothiocyanate-conjugated anti-β-globin antibody (1:1000 dilution, Santa ACKNOWLEDGEMENTS Cruz Biotechnology) and primary mouse anti-GAPDH antibody (1:5000, We express our sincerest gratitude to Professor Michelle Calos for providing the pDB2 Millipore, Bedford, MA, USA). The secondary horseradish peroxidase- conjugated goat anti-mouse IgG (DakoCytomation, Kyoto, Japan) was used and pCMV/Int vectors. This study was funded by grants provided by Royan Institute. at a dilution of 1:7500. The protein bands were visualized using an Amersham ECL Advance Western Blotting Detection Kit (GE Healthcare, Freiburg, Germany). REFERENCES 1 Kay MA. State-of-the-art gene-based therapies: the road ahead. Nat Rev Genet 2001; 12:316–328. Bacterial backbone and eukaryotic selection marker removal by 2 O'Connor TP, Crystal RG. Genetic medicines: treatment strategies for hereditary modified Cre recombinase disorders. Nat Rev Genet 2006; 7:261–276. In vitro Cre-mediated vector engineering. In order to delete the bacterial 3 Moi P, Sadelain M. Towards the genetic treatment of beta-thalassemia: new dis- backbone and eukaryotic selection marker, we tested the in vitro ease models, new vectors, new cells. Haematologica 2008; 93: 325–330.

© 2015 Macmillan Publishers Limited Gene Therapy (2015) 663 – 674 β-Globin expression by a non-viral integrating DNA K Dormiani et al 674 4 Quek L, Thein SL. Molecular therapies in beta-thalassaemia. Br J Haematol 2007; gene copy number changes. In: Hilario E, Mackay J (eds), Methods in Molecular 136:353–365. Biology. Humana Press: Totowa, NJ, 2007, pp 205–226. 5 Dong A, Rivella S, Breda L. Gene therapy for hemoglobinopathies: progress and 32 Branda CS, Dymecki SM. Talking about a revolution: the impact of site-specific challenges. Transl Res 2013; 161:293–306. recombinases on genetic analyses in mice. Dev Cell 2004; 6:7–28. 6 Naldini L. Ex vivo gene transfer and correction for cell-based therapies. Nat Rev 33 Kollias G, Wrighton N, Hurst J, Grosveld F. Regulated expression of human A Genet 2011; 12: 301–315. gamma-, beta-, and hybrid gamma beta-globin genes in transgenic mice: 7 Rund D, Rachmilewitz E. β-Thalassemia. N Engl J Med 2005; 355: 1135–1146. manipulation of the developmental expression patterns. Cell 1986; 46:89–94. 8 Ye L, Chang JC, Lin C, Sun X, Yu J, Kan YW. Induced pluripotent stem cells offer 34 Howden SE, Voullaire L, Wardan H, Williamson R, Vadolas J. Site-specific Rep- new approach to therapy in thalassemia and sickle cell anemia and option in mediated integration of the intact beta-globin locus in the human ery- prenatal diagnosis in genetic diseases. Proc Natl Acad Sci USA 2009; 106: throleukaemic cell line K562. Gene Therapy 2008; 15: 1372–1383. 9826–9830. 35 Kotzamanis G, Abdulrazzak H, Kotsinas A, Gorgoulis VG. Non-viral gene therapy. 9 Papapetrou EP, Zoumbos NC, Athanassiadou A. Genetic modification of In: Yuan X (ed), Non-Viral Gene Therapy. InTech: Croatia 2011, pp 1–26. hematopoietic stem cells with non-viral systems: past progress and future 36 Sadelain M, Lisowski L, Samakoglu S, Rivella S, May C, Riviere I. Progress toward prospects. Gene Therapy 2005; 12: S118–S130. the genetic treatment of the beta-thalassemias. Ann N Y Acad Sci 2005; 1054: 10 Karow M, Calos MP. The therapeutic potential of phiC31 integrase as a gene 78–91. therapy system. Expert Opin Biol Ther 2011; 11: 1287–1296. 37 Watanabe S, Nakamura S, Sakurai T, Akasaka K, Sato M. Improvement of a phiC31 11 Hackett PB. Integrating DNA vectors for gene therapy. Mol Ther 2007; 15:10–12. integrase-based gene delivery system that confers high and continuous trans- 12 Christopher L, Chavez CL, Calos MP. Therapeutic applications of the phiC31 gene expression. N Biotechnol 2011; 28:312–319. integrase system. Curr Gene Ther 2011; 11: 1287–1296. 38 Tanimoto K, Liu Q, Bungert J, Engel JD. Effects of altered gene order or orientation 13 Liu Y, Lakshmipathy U, Ozgenc A, Thyagarajan B, Lieu P, Fontes A et al. hESC of the locus control region on human β-globin gene expression in mice. Nature engineering by integrase-mediated chromosomal targeting. Methods Mol Biol 1999; 398:344–348. 2010; 594: 229–268. 39 Fraser P, Hurst J, Collis P, Grosveld F. DNaseI hypersensitive sites 1, 2 and 3 of the 14 Glover DJ, Lipps HJ, Jans DA. Towards safe, non-viral therapeutic gene expression human beta-globin dominant control region direct position-independent in humans. Nat Rev Genet 2005; 6:299–310. expression. Nucleic Acids Res 1990; 18: 3503–3508. 15 Calos MP. The phiC31 integrase system for gene therapy. Curr Gene Ther 2006; 6: 40 Noordermeer D, Branco MR, Splinter E, Klous P, Van IW, Swagemakers S et al. 633–645. Transcription and chromatin organization of a housekeeping gene cluster con- 16 Groth AC, Olivares EC, Thyagarajan B, Calos MP. A phage integrase directs efficient taining an integrated beta-globin locus control region. PLoS Genet 2009; 4: site-specific integration in human cells. Proc Natl Acad Sci USA 2000; 97: e1000016. 5995–6000. 41 Bungert J, Davé U, Lim KC, Lieuw KH, Shavit JA, Liu Q et al. Synergistic regulation 17 Thyagarajan B, Olivares EC, Hollis RP, Ginsburg DS, Calos MP. Site-specific geno- of human beta-globin gene switching by locus control region elements HS3 mic integration in mammalian cells mediated by phage phiC31 integrase. Mol Cell and HS4. Genes Dev 1995; 9: 3083–3096. Biol 2001; 21: 3926–3934. 42 Zhu J, Kren BT, Park CW, Bilgim R, Wong PY, Steer CJ. Erythroid-specific expression 18 Sivalingam J, Krishnan S, NG WH, Lee SS, Phan TT, Kon OL. Biosafety assessment of of beta-globin by the sleeping beauty transposon for Sickle cell disease. Bio- site-directed transgene integration in human umbilical cord-lining cells. Mol Ther chemistry 2007; 46: 6844–6858. 2010; 18: 1346–1356. 43 Antoniou M, Geraghty F, Hurst J, Grosveld F. Efficient 3'-end formation of human 19 Giudice A, Trounson A. Genetic modification of human embryonic stem cells for beta-globin mRNA in vivo requires sequences within the last intron but occurs derivation of target cells. Cell Stem Cell 2008; 2:422–433. independently of the splicing reaction. Nucleic Acids Res 1998; 26:721–729. 20 Thyagarajan B, Calos MP. Site-specific integration for high-level protein produc- 44 Leboulch P, Huang GM, Humphries RK, Oh YH, Eaves CJ, Tuan DY et al. Muta- tion in mammalian cells. In: Smales CM, James DC (eds), Therapeutic Proteins: genesis of retroviral vectors transducing human beta-globin gene and beta- Methods and Protocol. Humana Press: Totowa, NJ, 2005, pp 99–106. globin locus control region derivatives results in stable transmission of an active 21 Thyagarajan B, Liu Y, Shin S, Lakshmipathy U, Scheyhing K, Xue H et al. Creation of transcriptional structure. EMBO J 1994; 13:3065–3076. engineered human embryonic stem cell lines using phiC31 integrase. Stem Cells 45 Rubin JE, Pasceri P, Wu X, Leboulch P, Ellis J. Locus control region activity by 5'HS3 2008; 26: 119–126. requires a functional interaction with beta-globin gene regulatory elements: 22 Bertoni C, Jarrahian S, Wheeler TM, Li Y, Olivares EC, Calos MP et al. Enhancement expression of novel beta/gamma-globin hybrid transgenes. Blood 2000; 95: of plasmid-mediated gene therapy for muscular dystrophy by directed plasmid 3242–3249. integration. Proc Natl Acad Sci USA 2006; 103: 419–424. 46 Nishiumi F, Sone T, Kishine H, Thyagarajan B, Kogure T, Miyawaki A et al. Simul- 23 Keravala A, Chavez CL, Hu G, Woodard LE, Monahan PE, Calos MP. Long-term taneous single cell stable expression of 2-4 cDNAs in HeLaS3 using phiC31 phenotypic correction in factor IX knockout mice by using phiC31 integrase- integrase system. Cell Struct Funct 2009; 34:47–59. mediated gene therapy. Gene Therapy 2011; 18:842–848. 47 Ehrhardt A, Engler JA, Xu H, Cherry AM, Kay MA. Molecular analysis of chromo- 24 Ishikawa Y, Tanaka N, Murakami K, Uchiyama T, Kumaki S, Tsuchiya S et al. somal rearrangements in mammalian cells after phiC31-mediated integration. Phage phiC31 integrase-mediated genomic integration of the common Hum Gene Ther 2006; 17: 1077–1094. cytokine receptor gamma chain in human T-cell lines. J Gene Medicine 2006; 8: 48 Chalberg TW, Portlock JL, Olivares EC, Thyagarajan B, Kirby PJ, Hillman RT et al. 646–653. Integration specificity of phage phiC31 integrase in the human genome. J Mol Biol 25 Quenneville SP, Chapdelaine P, Rousseau J, Beaulieu J, Caron NJ, Skuk D et al. 2006; 357:28–48. Nucleofection of muscle-derived stem cells and myoblasts with phiC31 integrase: 49 Chen ZY, Riu E, He CY, Xu H, Kay MA. Silencing of episomal transgene expression stable expression of a full-length-dystrophin fusion gene by human myoblasts. in liver by plasmid bacterial backbone DNA is independent of CpG methylation. Mol Ther 2004; 10:679–687. Mol Ther 2008; 16:548–556. 26 Ortiz-Urda S, Keene D, Lin Q, Calos MP, Khavari P. PhiC31 integrase-mediated non- 50 Pham CT, MacIvor DM, Hug BA, Heusel JW, Ley TJ. Long-range disruption of gene viral genetic correction of junctional epidermolysis bullosa. Hum Gene Ther 2003a; expression by a selectable marker cassette. Proc Natl Acad Sci USA 1996; 93: 14: 923–928. 13090–13095. 27 Keravala A, Ormerod BK, Palmer TD, Calos MP. Long-term transgene expression in 51 Davis RP, Costa M, Grandela C, Holland AM, Hatzistavrou T, Micallef SJ et al. A mouse neural progenitor cells modified with phiC31 integrase. J Neurosci Methods protocol for removal of antibiotic resistance cassettes from human embryonic 2008; 173: 299–305. stem cells genetically modified by homologous recombination or transgenesis. 28 Ellis J, Pannell D. The beta-globin locus control region versus gene therapy vec- Nat Protoc 2008; 3: 1550–1558. tors: a struggle for expression. Clin Genet 2001; 59:17–24. 52 Hemmi H, Takeuchi O, Kawai T, Kaisho T, Sato S, Sanjo H et al. A Toll-like receptor 29 Young K, Donovan-Peluso M, Bloom K, Allan M, Paul J, Bank A. Stable transfer and recognizes bacterial DNA. Nature 2000; 408:740–745. expression of exogenous human globin genes in human erythroleukemia 53 EMA. Guideline on the Non-Clinical Studies Required Before First Clinical Use of Gene (K562) cells. Proc Natl Acad Sci USA 1984; 81: 5315–5319. Therapy Medicinal Products. European Medicines Agency: London, UK, 2008. 30 Takagi S, Kimura M, Katsuki M. A rapid and efficient protocol of the inverted PCR 54 Peitz M, Pfannkuche K, Rajewsky K, Edenhofer F. Ability of the hydrophobic FGF using two primer pairs. Biotechniques 1992; 13: 176–178. and basic TAT peptides to promote cellular uptake of recombinant Cre recom- 31 Hoebeeck J, Speleman F, Vandesompele J. Real-time quantitative PCR as an binase: a tool for efficient genetic engineering of mammalian genomes. Proc Natl alternative to southern blot or fluorescence in situ hybridization for detection of Acad Sci USA 2002; 99: 4489–4494.

Supplementary Information accompanies this paper on Gene Therapy website (http://www.nature.com/gt)