Gene Therapy (2002) 9, 708–712  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt SESSION IV – CHROMOSOME ENGINEERING The use of chromosome-based vectors for animal transgenesis

Y Kuroiwa1, H Yoshida1, T Ohshima1, T Shinohara2, A Ohguma1, Y Kazuki2, M Oshimura2, I Ishida1 and K Tomizuka1 1Pharmaceutical Research Laboratory, Kirin Brewery Co Ltd, Gunma, Japan; and 2Department of Molecular and Cell Genetics, School of Life Sciences, Faculty of Medicine, Tottori University, Tottori, Japan

This article summarizes our efforts to use chromosome- human artificial chromosomes (HACs) containing a defined based vectors for animal transgenesis, which may have a chromosomal region should prevent the introduction of benefit for overcoming the size constraints of cloned trans- additional genes other than genes of interest and thus genes in conventional techniques. Since the initial trial for enhance the utility of chromosome vector system. Using this introducing naturally occurring human chromosome frag- technique a panel of HACs harboring inserts ranging in size ments (hCFs) with large and complex immunogulobulin (Ig) from 1.5 to 10 Mb from three human chromosomes (hChr2, loci into mice we have obtained several lines of trans-chro- 7, 22) has been constructed. Tc animals containing the mosomic (Tc) mice with transmittable hCFs. As expected the HACs may be valuable not only as a powerful tool for normal tissue-specific expression of introduced human functional genomics but also as an in vivo model to study genes was reproduced in them by inclusion of essential therapeutic gene delivery by HACs. remote regulatory elements. Recent development of ‘chro- Gene Therapy (2002) 9, 708–712. DOI: 10.1038/sj/gt/3301754 mosome cloning’ technique that enable construction of

Keywords: chromosome; vector; transgenic animal; ES cells; Cre-loxP; telomere

Introduction host cells, which may prevent artifacts and gene silenc- ing3 caused by multiple copy insertion of transgenes; (4) The ability of human chromosome-derived natural frag- there is no risk for insertional mutagenesis because they ments (hCFs) as vectors for introducing large stretches of are freely segregating from host chromosomes. 1 human DNA into mice was first demonstrated in 1997. On the other hand, one major problem of hCF vectors Transferred hCFs were stably maintained as an extra has been that they are structurally undefined.4,5 The hCFs chromosome in the somatic cells of mice and the human are usually generated as a consequence of accidental frag- genes included in them were expressed under proper mentation of intact chromosome during the chromosome tissue-specific regulation. In some cases they could be transfer process and run the risk of containing unrelated, transmitted through the mouse germline, resulting in the deleterious genes in addition to the genes of interest. In establishment of novel mouse strains (trans-chromosomic this context, we recently developed a ‘chromosome clon- 1,2 (Tc) mice) containing a heritable hCF. Thus, such an ing’ technique that enabled construction of human arti- approach employing chromosome vectors for animal ficial chromosomes (HACs) including only defined transgenesis has been thought to be useful for overcom- chromosomal regions in the homologous recombination ing the size constraints of cloned transgenes in conven- proficient chicken DT40 cell line.6 The complete human tional techniques and facilitate functional studies of genome sequence which is now available,7 should facili- human genome in the ‘post sequencing’ era. tate the minimization of HACs containing only desired Compared with conventional techniques using cloned chromosomal regions. transgenes, the chromosome vector system have several Here, we outline the current status of our ‘trans- advantages: (1) whole genomic sequences including all chromosomic’ technology and the recent progress on the the introns and essential regulatory elements can be used development of HACs. as transgenes, which results in the correct and controlled expression of transgenes in vivo; (2) very large genes, gene clusters and specific chromosomal regions, which Results and discussion cannot be cloned as contiguous DNA fragments by clon- ing techniques, can be introduced; (3) chromosome vec- Trans-chromosomic mice tors are maintained as a single copy chromosome in the Our procedure to introduce chromosome vectors into mice is illustrated in Figure 1. Microcell-mediated chro- mosome transfer (MMCT) was employed to introduce Correspondence: K Tomizuka, Pharmaceutical Research Laboratory, Kirin chromosome vectors into mouse embryonic stem (ES) Brewery Co Ltd, 3 Miyahara-cho, Takasaki-shi, Gunma, 370-1295, Japan cells.1 Our previous studies indicated that the use of Chromosome-based vectors for animal transgenesis Y Kuroiwa et al 709

Figure 1 A schematic diagram showing the construction of MH(ES) cells to produce chimeric mice and Tc mice expressing human genes on the transferred hCF. G418r, G418-resistant. small human chromosome fragments (hCFs) may addition to the danger of including unrelated, deleterious increase the probability of successful chimera production genes as mentioned above. We therefore developed a from microcell hybrid ES (MH(ES)) cells and germline ‘chromosome-cloning’ technique where a defined transmission of the hCFs.2 Therefore, the library compris- chromosomal region could be cloned into a stable human ing approximately 700 independent clones of human– minichromosome vector by combination of Cre/loxP- mouse monochromosomal hybrids was screened to mediated chromosome translocation13 and telomere- obtain small hCFs containing the gene of interest.1,2 To directed chromosome truncation8 in homologous recom- date we successfully produced chimeric mice from bination-proficient chicken DT40 cells.14 MH(ES) cells containing the hCF derived from hChr2, 4, In our system (Figure 2), the hCF(SC20) was chosen as 6, 7, 11, 14, 21 or 22.1,2,8,9 In addition, germline trans- a basal minichromorome vector (SC20 vector) because of mission of transferred hCF was observed in chimeras its relatively high mitotic and meiotic stability in mice.2,12 produced from MH(ES) cells containing the hChr2, 7, 14 At first, a loxP sequence was integrated to RNR2 locus or 21-derived hCF.1,2,10 on the SC20 vector in DT40 cells, to which various Functional expression of introduced genes was exten- chromosomal regions could be cloned. As the first model sively investigated in the Tc mice containing the transmit- of HAC construction carrying a defined chromosomal table hChr2-derived hCF (hCF(2-W23)) and hChr14- insert, we attempted to clone the region between HCF2 derived hCF (hCF(SC20)), each of which include and LIF loci containing Ig␭ locus on the hChr22, the 10 immunoglobulin (Ig) kappa light chain (2 Mb) and heavy Mb-sized hChr22-derived insert. In DT40 cells, the chain (1.5 Mb) locus, respectively.1,2 The results of struc- hChr22 was truncated at the LIF locus by integrating tural analysis of human Ig mRNA and Ig protein human telomeric repeats (TTAGGG)n, followed by inte- expression in the sera of Tc mice suggested that these gration of a loxP sequence to the HCF2 locus.6 Next, the large and complex loci were properly expressed to recon- DT40 clone containing the SC20 vector and the DT40 stitute the diverse and functional repertoire of human Igs clone containing the hChr22 fragment (hCF(22)) were in the Tc mice.1 Furthermore, the introduction of fused, resulting in DT40 cell hybrids with both SC20 vec- hCF(SC20) into endogenous IgH-knockout strain, in tor and hCF(22). The hybrids were transfected with a Cre which functional B-lymphocytes and Ig production are recombinase-expression vector to induce Cre/loxP- absent, by mating resulted in the rescue of its defects, mediated chromosomal translocation between the SC20 indicating that the stability and functionality of the vector and hCF(22). Occurrence of the translocation was hCF(SC20) is likely to be sufficient for restoration of B confirmed by nested PCR analysis and then the hybrid cells in adult mice.2 cells containing the possible translocated chromosomal Stability tests using MH(ES) cell lines2,6revealed the fragment were isolated by fluorescence activated cell hCF(SC20) was highly stable (Ͻ0.1% loss/doubling) in sorter (FACS).6 Fluorescence in situ hybridization (FISH) contrast to the hCF(2-W23) (3.2% loss/doubling), hCF(22) analysis in the sorted cells demonstrated successful gen- (>5% loss/doubling) and the hChrY-derived minichro- eration of the human artificial chromosome (HAC) carry- mosome reported by Shen et al.11 Similar results were ing the defined hChr22-derived insert, named ␭HAC obtained in various types of cell lines from different spec- (Figure 3). ies12 and in the somatic and germ cells of Tc mice.2 Thus, To see the functionality of the HAC, we transferred the it was suggested that each hCF may have an intrinsic ␭HAC to mouse embryonic stem (ES) cells where level of mitotic and meiotic stability, which is less affec- endogenous IgH and Ig␬ genes were inactivated. In the ted by the type of host cells. ES cells, whereas hCF(22) was drastically lost after 15- day culture (about 20 doublings) under nonselective con- Construction of HACs dition, ␭HAC carrying this unstable hCF(22)-derived This differential stability of hCFs could be another prob- insert, as well as the SC20 vector itself, was stably main- lem in using the hCFs as gene delivery vectors, in tained at 99.8% retention per cell division throughout the

Gene Therapy Chromosome-based vectors for animal transgenesis Y Kuroiwa et al 710

Figure 2 Schematic diagram of the chromosome-cloning system. DT40/SC20 vector, the DT40 clone with the SC20 vector where a loxP sequence is integrated at the RNR2 locus. DT40/hCF22, the DT40 clone with the hChr22 fragment truncated at the LIF locus where a loxP sequence is integrated at the HCF2 locus. The two DT40 clones were fused, resulting in the hybrid clone containing both SC20 vector and hCF22. In the hybrid clone into which Cre recombinase was introduced, the ␭HAC was generated by recombination between the SC20 vector and hCF22. The ␭HAC is composed of the SC20 vector (red) and the 10 Mb-sized hChr22 insert carrying the Ig␭ locus.

Figure 3 Verification of generation of the ␭HAC by FISH analysis. FISH analysis of the FACS-sorted cells using hChr14 (red) and hChr22 (green)- specific probes is shown.

45-day nonselective culture (about 70 doublings) (Figure the MH(ES) cells using the homozygous IgH-knockout 4). The fact that we could stabilize mitotically unstable strain as a host for chimera production. In the sera of the hCF(22)-derived insert by cloning it into the SC20 vector resultant ␭HAC chimeras, human Ig␮ and Ig␭ proteins suggests that the centromere of the SC20 vector domi- were detected by enzyme linked immunosorbent assays nantly contributed to mitotic stability of the ␭HAC and (ELISAs) at 0.2–1 mg/ml. When immunized with human that the SC20 vector may be used as a stable cloning vec- G-CSF (granulocyte colony-stimulating factor), the ␭HAC tor for various chromosomal inserts. chimera showed immune response of G-CSF-specific To examine whether the ES cells containing the ␭HAC human Ig␭ antibodies in the sera. These data suggest that have pluripotency to differentiate and whether the Ig␭ ␭HAC was stably maintained during development of B gene on the chromosome insert of the ␭HAC can be func- cells and that Ig␭ locus on the ␭HAC was fully functional. tionally expressed in vivo, we created chimeric mice from In addition to B cells, the ␭HAC was also observed in

Gene Therapy Chromosome-based vectors for animal transgenesis Y Kuroiwa et al 711 region around Ig␬ locus defined by yWHZ30-4 and CD8A loci (about 5 Mb) and the hChr7 region around CYP3A locus defined by COL1A2 and AF0006754 loci (about 3 Mb) were cloned into the SC20 vector, respectively. The ⌬HAC and ⌬⌬HAC carry the 2.5 Mb-sized hChr22 insert defined by HCF2 and AP000344 loci and the 1.5 Mb-sized hChr22 insert defined by AP000553 and AP000344 loci, respectively. In all the cases GFP-positive DT40 hybrid cells, in which the chromosomal insert was successfully translocated to the SC20 vector, were obtained at a fre- quency of 1–10 × 10Ϫ7 per Cre transfected DT40 hybrid cells. The ⌬HAC, ⌬⌬HAC and #7HAC was introduced into mouse ES cells, respectively (see Figure 1) and the resultant MH(ES) cells containing each HAC were used for the production of chimeras. In the sera of ⌬HAC and ⌬⌬HAC chimeras we could detect human Ig ␮ and ␭ pro- teins by ELISA assays. Transcript of cytochrome P450 3A4 (CYP3A4) gene was also detected specifically in liver and small intestine of #7HAC chimeras by RT-PCR analy- sis, which coincides well with the tissue specificity observed in humans. Furthermore, we recently found that the ⌬HAC was germline-transmittable. Retention of the ⌬HAC was confirmed by PCR marker analysis and ELISAs for the detection of Ig ␮ and Ig ␭. FISH analyses also revealed that the percentage of the metaphase spre- ads containing the ⌬HAC averaged 70% in the tail fibro- blasts prepared from PCR-positive F1 offspring, which Figure 4 Mitotic stability of the ␭HAC in mouse ES cells. Mitotic stab- ility was determined by FISH using human COT1 DNA probe. is comparable to the previous data in the Tc mice with hCF(SC20).2 Breeding tests for the transmission of other HACs are now underway. tail fibroblasts of chimeras, indicating that the ␭HAC was stably maintained in somatic cells during development Potential applications in vivo. We described the chromosome-cloning system by which To determine whether this chromosome-cloning sys- various hChr regions could be cloned into the SC20 min- tem works for other hChr regions, we similarly con- ichromosome vector, irrespective of its size and intro- structed further four HACs, called #2HAC, #7HAC, duced into mice stably, allowing human genes of interest ⌬HAC and ⌬⌬HAC, using the SC20 vector as a basal vec- to be functionally expressed in vivo. In our HACs, the tor (Figure 5). In the #2HAC and #7HAC, the hChr2 chromosome inserts flanked by the loxP-integration and

Figure 5 Construction of several HACs by means of the chromosome-cloning system. The defined regions of hChr22, hChr2 and hChr7 were cloned into the SC20 minichromorome vector. The #2HAC and #7HAC carry the hChr2 insert defined by cosyWHZ30-4 and CD8A loci and the hChr7 insert defined by COL1A2 and AF006752 loci, respectively. The ⌬HAC and ⌬⌬HAC carry the 2.5 Mb-sized hChr22 insert defined by HCF2 and AP000344 loci and 1.5 Mb-sized hChr22 insert defined by AP000553 and AP000344 loci, respectively.

Gene Therapy Chromosome-based vectors for animal transgenesis Y Kuroiwa et al 712 telomere-truncation sites are structurally determined by 2 Tomizuka K et al. Double trans-chromosomic mice: maintenance the information of the recently published human genome of two individual human chromosome fragments containing Ig sequence.7 The availability of structurally defined HAC heavy and kappa loci and expression of fully human antibodies. vectors would be of great value in the construction of Proc Natl Acad Sci USA 2000; 97: 722–727. 3 Garrick D, Fiering S, Martin DI, Whitelaw E. Repeat-induced animals carrying human genetic elements to model spe- gene silencing in mammals. Nature Genet 1998; 18:56–59. cific diseases or the production of various therapeutic 4 Grimes B, Cooke H. Engineering mammalian chromosomes. products for industrial purposes. Hum Mol Genet 1998; 7: 1635–1640. Another potential application of the HAC vector may 5 Brown WRA, Mee PJ, Shen MH. Artificial chromosomes: ideal be the use in the treatment of human genetic diseases. vectors? Trends Biochem 2000; 18: 218–223. Indeed, the advantages of chromosome vector system in 6 Kuroiwa Y et al. Manipulation of human minichromosomes to animal transgenesis, mentioned above, are also desirable carry greater than megabase-sized chromosome inserts. Nature characteristics for therapeutic vectors to overcome vari- Biotechnol 2000; 18: 1086–1090. ous problems in existing viral and non-viral vector sys- 7 International Human Genome Sequencing Consortium A physi- tems. Although the inefficient delivery of a large chromo- cal map of the human genome. Nature 2001; 409: 860–921. 8 Kuroiwa Y et al.Efficient modification of a human chromosome some molecule into the cells remains to be overcome our by telomere-directed truncation in high homologous recombi- ‘trans-chromosomic’ mice may provide a useful model nation-proficient chicken DT40 cells. Nucleic Acids Res 1998; 26: system for in vivo assessment of the function and 3447–3448. behavior of chromosome-based therapeutic vectors. As 9 Shinohara T et al. Mice containing a human chromosome 21 for the hChr14-derived hCF(SC20) minichromosome vec- model behavioral impairment and cardiac anomalies of Down’s tor, we are in the process of minimizing it, so that it will syndrome. Hum Mol Genet 2000; 10: 1163–1175. no longer contain extra-human genes, aiming at using it 10 Kazuki Y et al. Germline transmission of a transferred human as a vector for human gene therapy. For example, dystro- chromosome 21 fragment in transchromosomal mice. JHum phin is the largest gene (2.4 Mb) in humans and is respon- Genet 2001; 46: 600–603. sible for Duchenne muscular dystrophy.15 Because of its 11 Shen MH et al. Human mini-chromosomes in mouse embryonal stem cells. Hum Mol Genet 1997; 6: 1375–1382. large size, the entire dystrophin locus has never been 12 Shinohara T et al. Stability of transferred human chromosome cloned even using YAC vectors. However, by using our fragments in cultured cells and in mice. Chromosome Res 2000; HAC system, it will be able to be cloned and the HAC 8: 713–725. carrying the entire dystrophin locus could be evaluated 13 Smith AJH et al. A site-directed chromosomal translocation for its usefulness as a therapeutic vector for Duchenne induced in embryonic stem cells by Cre-loxP recombination. muscular dystrophy in a model animal, such as mdx Nature Genet 1995; 9: 376–384. mice.16 14 Dieken ES et al.Efficient modification of human chromosomal alleles using recombination-proficient chicken/human microcell hybrids. Nature Genet 1996; 12: 174–182. Acknowledgements 15 Dunnen JTD et al. Recombination of the 2.4 Mb human DMD- gene by homologous YAC recombination. Hum Mol Genet 1992; We wish to thank S Tanaka, S Igami, T Ishihara, M 1:19–28. Shionoya for excellent technical assistance. We also thank 16 Ryder-Cook AS et al. Localization of the mdx mutation within K Hanaoka and M Hayasaka for technical advice and the mouse dystrophin gene. EMBO J 1988; 7: 3017–3021. valuable discussions.

References 1 Tomizuka K et al. Functional expression and germline trans- mission of a human chromosome fragment in chimaeric mice. Nature Genet 1997; 16: 133–143.

Gene Therapy