Nature 419, 583 - 587 (2002); doi:10.1038/nature01079 Somatic cell nuclear transfer I. WILMUT*, N. BEAUJEAN*, P. A. DE SOUSA*, A. DINNYES*†, T. J. KING*, L. A . PATERSON*, D. N. WELLS‡ & L. E. YOUNG*† * Roslin Institute, Roslin, Midlothian EH25 9PS, UK ‡ AgResearch Ltd, PB3123, Hamilton, New Zealand † Present addresses: Research Group for Animal Breeding, Hungarian Academy of Sciences & Szent István University, Hungary (A.D.); and School of Human Dev elopment, Division of Reproductive Medicine, Queens Medical Centre, Nottingham N G7 2UH, UK (L.E.Y.). Correspondence and requests for materials should be addressed to I.W. (e-mail: i [email protected]). Cloning by nuclear transfer from adult somatic cells is a remarkable demonstrati on of developmental plasticity. When a nucleus is placed in oocyte cytoplasm, th e changes in chromatin structure that govern differentiation can be reversed, an d the nucleus can be made to control development to term. Cloning by present methods is very inefficient owing to the extraordinary demand s placed on the oocyte cytoplasm in reprogramming a somatic nucleus rather than a sperm nucleus. The cumulative loss observed throughout development is assumed to reflect inappropriate expression of many genes whose harmful effect is exerte d at different stages of development. These fundamental limitations to cloning a re being addressed by analyses of the underlying cellular mechanisms. In time, t his information may be used in the development of treatments to cause cells of o ne phenotype to 'transdifferentiate' to another. Several laboratories have used a variety of somatic cell types to create cloned sheep, cattle, mice, pigs, goats, rabbits and cats (results are summarized in re f. 1). However, those laboratories have failed so far to obtain offspring in oth er species, including the rat, rhesus monkey and dog. Consistent among all publi shed research is that only a small proportion of embryos reconstructed using adu lt or fetal somatic cells developed to become live young, typically between 0 an d 4%. The low overall success rate is the cumulative result of inefficiencies at each stage of the process (Fig. 1). Figure 1 A comparison of survival to term of bovine embryos produced either by in vitro fertilization or nuclear transfer from embryos, fetal or adult fibrobla sts. Full legend High resolution image and legend (20k) Fetal loss and pathologies of cloned offspring In addition to embryonic loss, somatic cell nuclear transfer is also associated with very high rates of fetal, perinatal and neonatal loss, and production of ab normal offspring. Not all of these effects are due solely to nuclear transfer as similar problems are reported after embryo culture2. Typically, at least one-third of the cattle and sheep confirmed pregnant with cl oned embryos lose their fetuses during gestation1. Abnormal development of the p lacenta, including vascular reduction, is a principal contributor to loss partic ularly during early pregnancy in sheep and cattle3. It may also contribute to so me of the defects reported in neonates4. In cattle the rate of loss is also incr eased in the second and third trimesters of pregnancy after nuclear transfer (co mpared with in vitro fertilization (IVF)), with greater losses when adult rather than fetal or embryonic nuclei are used5. The overaccumulation of placental flu id in hydroallantois occurs rarely in natural cattle pregnancies, but can affect up to 2% and 40% of pregnancies established with IVF6 and cloned embryos7, resp ectively. In cloned mice, the placentae are often 2–3-fold heavier than from natural mating8, although a lack of vascularization has not been reported. Many cloned offspring die within the first 24 h of birth (Table 1). Common anoma lies include respiratory distress, increased birth weight and major cardiovascul ar abnormalities that can result in gross distension of the liver and dilated ma jor vessels. Some of these anomalies may be secondary effects of another abnorma lity. The production of oversized offspring after nuclear transfer in ruminants mimics the findings following other embryo manipulations when increased birth we ight is the most readily recognized characteristic of the 'large offspring syndr ome'2. Other common characteristics include prolonged gestation, fluid accumulat ion, enlargement of organs, sluggish onset of labour and difficulty in breathing . There has not been an adequate assessment of post-natal development of clones. A bnormalities that have been described include failure of the immune system, stru ctural abnormalities of the brain, digestive dysfunction, enteritis and umbilica l infections. There is a suggestion that the fate of cloned individuals may be i nfluenced by genetic background or donor cell type. In two independent studies i n which the strains of mice used differed, mice cloned from cumulus cells became obese in adult life9 whereas those from Sertoli cells of immature mice died at an unusually early age10. The fact that the obesity was not heritable after sexu al reproduction provides direct evidence that failures associated with cloning a re epigenetic9. By contrast, physiological studies of cloned calves suggest norm ality, at least for the tests examined11. However, the eldest of these animals w as only 4 years at the time of the tests and it is important that studies extend throughout a lifespan, which is potentially more than 15 years. Cellular factors influencing outcome Maintenance of ploidy Normal development of cloned embryos depends on maintainin g normal ploidy by coordinating the cell cycles of the recipient oocyte and dono r nucleus (see ref. 12 for review). Three effective strategies have been defined . When using mammalian oocytes arrested at metaphase II as the recipient cells, normal ploidy is maintained if the donor nucleus is awaiting DNA replication (G0 or G1). Alternatively, advantage may be taken of the fact that ploidy is also c hanged if chromosomes are shed in a polar body. Following transfer of G2/M-phase nuclei (4C) into enucleated metaphase II mouse oocytes13, expulsion of a polar body returns the reconstructed embryo to normal ploidy of 2C, provided that cyto chalasin B is omitted from the culture medium. In the third strategy, an interph asic donor nucleus may be at any stage of the cycle if the metaphase II oocyte i s activated before transfer and M-phase promoting factor (MPF) levels decline. H owever, cytoplasts with basal levels of MPF may not be optimal for reprogramming somatic nuclei14. Although these general principles were established quickly, i t seems probable that there is much to learn of the detailed mechanisms in diffe rent species. Rabbits were successfully cloned only when adjustments were made t o accommodate for the rapid progression through the first cell cycle in that spe cies15. Donor cell cycle stage Although few precise comparisons have been made, there ap pears to be a 'window of opportunity' in the cycle of donor cells, spanning G2/M /G1/G0 phases, when nuclear reprogramming is more effective. Dolly (the first su ccessful clone of an adult animal) was created from a donor cell induced to quie scence by serum deprivation16, and since then many other offspring have been pro duced using cells that were expected to be quiescent. Three studies have now car efully selected cells in G1 and contrasted them with cells in G0 (refs 7, 17, 18 ). All three studies obtained calves from cells in G1, although in one case no c alves were derived with fetal cells using the well-proven techniques in G0 (ref. 17). The other two studies, however, showed a 3–7-fold increase in the production of viable offspring with embryos derived from adult quiescent cells (G0)7, 18. Donor cells at metaphase may be considered optimally compatible with metaphase II oocytes and have resulted in improved rates of blastocyst development using murine embryonic stem (ES) cells, as opposed to interphasic nuclei; however, the post-implantation development was then characterized by a high mortality19. Determining optimal cell cycle stage of the donor cell is of great practical importance and would also focus mechanistic research. Delayed activation After nuclear transfer, an activation stimulus triggers devel opment. It was initially considered that the direct exposure of chromosomes to f actors in the oocyte cytoplasm before activation20 and allowing time for the re- assembly of the nuclear apparatus required for the cell cycle21 might aid develo pment. Although delaying oocyte activation after nuclear transfer can enhance de velopment in cattle and mice, it is not essential in these species14, 22. Stage of development of donor cells In general there is a decrease in the propor tion of embryos becoming offspring when donor nuclei are obtained from later sta ges of development. The change is assumed to reflect the degree of cellular diff erentiation and epigenetic constraints. The proportion of mouse embryos that dev eloped to offspring with transferred blastomere nuclei from 2-, 4- and 8-cell st age embryos decreased from 22%, 14% and 8%, respectively23. Embryonic survival d ecreases further with pluripotent mouse ES cells (2–11%) and somatic cells in G0/G1 (1?C3% (ref. 1)). The reduction in efficiency occurs later in cattle, as the propor tion of transferred bovine cloned embryos that developed into live offspring fro m morulae, fetal and adult fibroblasts was 28%, 13% and 5%, respectively (Fig. 1 ). Differences between species are associated with the later onset of major tran scription from the embryonic genome in ruminants than mice24. This may mean that developmental changes in chromatin structure take place later in ruminant devel opment and so less remodelling is required in cleavage stage embryos. Additional ly, after nuclear transfer more time and cell cycles are available for remodelli ng before transcription begins in ruminants. Recent observations in mice using f etal neural cells as nuclear donors have also suggested a loss of developmental totipotency with neural differentiation25. However, terminal differentiation-ass ociated DNA rearrangements25 may not be incompatible with nuclear totipotency as evidenced by successful cloning from terminally differentiated lymphocytes26. I n mice, differences in the pattern of fetal loss when using hybrid and inbred do nor cells were particularly apparent with ES cells as nuclear donors, and effici ency was also lower when inbred cumulus cells were used27. Molecular mechanisms The unique pattern of mortality during embryonic, fetal and neonatal stages in c lones apparently reflects the inappropriate expression of genes whose effect is exerted at different stages of development and can be fatal. Perturbations in th e expression of imprinted genes associated with phenotypic defects have now been observed in sheep, cattle and mice. Many imprinted genes have critical roles in regulating growth during fetal development; furthermore, the common characteris tics of the large offspring syndrome were recognized as being similar to those c linical and experimental situations in which the expression or effect of H19 and /or insulin-like growth factor II (IGF2) is perturbed2. Expression of IGF2 recep tor (IGF2R) was reduced in the liver, kidney, heart and muscle of unusually larg e sheep fetuses resulting from embryo culture28. The change in expression was as sociated with a loss of methylation at a differentially methylated region involv ed in regulation of gene imprinting. Unexpectedly, there was no change in expres sion of IGF2 or H19. Mouse fetuses derived by nuclear transfer from ES cells display aberrant express ion of H19, Igf2 and other imprinted genes known to have important roles in feta l or placental development. Some mice cloned from ES cells survive until adultho od and appear phenotypically normal despite epigenetic aberrations in at least o ne of the six imprinted genes that were examined29. Evidently 'normal' life is p ossible despite considerable variation in gene expression, and this probably dep ends on the particular gene affected, its degree and nature of perturbation, eff ects of redundancy, and on species-specific variation in susceptibility to any s ingle imprinted gene disruption. However, rather than being the consequence of n uclear transfer, the imprinting status of the cloned mice examined in ref. 29 ma y be due to the epigenetic instability inherent to ES cell lines. When somatic c ells were used as nuclear donors in mice, imprinting errors were not detected in the phenotypically normal fetus; however, the level of expression of several ge nes was elevated and imprinting errors were observed in their overgrown placenta e30. Apart from the improper expression of imprinted genes, the dysregulation of non- imprinted genes also occurs in clones31. However, two measures of the correct re programming of gene expression have been reported elsewhere. In female somatic c ells, reactivation of the inactive X chromosome can occur in early embryos after nuclear transfer, and reprogramming of telomerase activity can allow restoratio n of telomere length in cloned animals (reviewed in ref. 32). Epigenetic regulation of gene expression Errors leading to inappropriate express ion could arise at any of the levels at which regulation of gene expression occu rs, including the organization of the nucleus, chromatin structure and the avail ability of regulatory molecules. Chromatin proteins regulate access of the trans cription machinery to chromosomal DNA and so ultimately determine the pattern of gene expression. This chromosomal infrastructure is far more dynamic than had b een understood previously, adding to the impression that nuclear and chromosome structure are in a continual state of flux33. A large number of structural prote ins and enzymes that modify chromatin structure have been identified and are pri ncipal candidates for regulating early reprogramming events. The linker histone H1 may be involved in the regulation of gene expression in ea rly embryos (reviewed in ref. 34). The somatic form of histone H1 is apparently absent or in very low concentration in oocytes and early embryos, but becomes de tectable when major transcriptional activation of the embryonic genome occurs. S omatic H1 is lost from most mouse nuclei soon after transfer, but the cell cycle stage of donor and recipient cells influences the rate of loss35. Enzymatic mod ifications of histones include phosphorylation, methylation, acetylation and ubi quitination, or the removal of these modifications36. These modifications are re cognized by other structural proteins and enzymes, which together may stabilize the pattern of gene expression. Chromatin remodelling is also achieved by multisubunit ATP-dependent enzymes (re viewed in ref. 36). The helicase activity of these enzymes is able to expose DNA or redistribute nucleosomes in a tissue-specific manner. The loss of a principa l component of the basal transcriptional complex from somatic nuclei incubated i n frog oocyte extract provided the first indication that members of the SWI/SNF family of enzymes may have roles in the development of cloned embryos37. Finally, methylation of CpG dinucleotides in chromosomal DNA provides a genome-w ide means of regulation, usually associated with the inheritance of lineage-spec ific gene silencing between cell generations (reviewed in ref. 38). Widespread d isruptions in DNA methylation have now been described in cloned embryos of mice and cattle (reviewed in ref. 39), suggesting that present methods of cloning are disruptive and that the effects are complex. In cloned cattle embryos, the high ly methylated status of the donor cells was retained through several cell divisi ons. There was great variation between cloned embryos consistent with the variat ion in stage and cause of death of cloned embryos. In contrast, in the pig, chan ges in methylation of short interspersed element sequences were similar to those observed after in vitro fertilization, although at the 4–8-cell stage both were more extensively methylated than in embryos fertilized in vivo (reviewed in ref. 39). Together these observations emphasize the need for further studies of different regions of the genome with different donor cells and methods of nuclear transfer in several species. It is critical not to assume that mammalian pre-implantatio n events (and therefore reprogramming of a somatic nucleus) are the same in all species. The rapid demethylation of the paternal genome that is seen in several species does not occur in the sheep (Fig. 2: N.B., I.W., R. R. Meehan and L.E.Y. , unpublished data). Figure 2 5-Methyl cytosine immunostaining of pronuclear DNA in mouse and sheep zygotes. Full legend High resolution image and legend (8k) The future As the fate of cloned embryos is determined by molecular events within hours of nuclear transfer, it is disappointing that so little is known about these events during the early development of cloned embryos. During somatic cell nuclear tra nsfer a great deal is asked of the molecular mechanisms that have evolved to reg ulate fertilization and pregnancy. Viewed in this light, it is still surprising that somatic cell cloning ever produces viable offspring. Although some improvem ents in efficiency are to be expected from optimization of present procedures, g reater benefits might be expected from intervention to assist reprogramming of t he transferred nucleus. At present the means to enhance the success of nuclear t ransfer are not known, but may involve the use of remodelling complexes and fact ors that remove somatic epigenetic modifications before transfer. In addition to application of this information in nuclear transfer, new understanding of mecha nisms that regulate developmental plasticity will lead to methods to change cell s of one phenotype to another as a means of providing histocompatible cells for treatment of degenerative diseases. A new era of developmental biology and regen erative medicine awaits. References 1. Wilmut, I. & Peterson, L. A. Somatic cell nuclear transfer (clonin g) efficiency [online] http://www.roslin.ac.uk/public/webtablesGR.pdf (2002). 2. Young, L. E. & Fairburn, H. R. Improving the safety of embryo technologies: p ossible role of genomic imprinting. Theriogenology 53, 627-648 (2000) | Article | PubMed | 3. Hill, J. R. et al. Evidence for placental abnormality as the major cause of m ortality in first-trimester somatic cell cloned bovine fetuses. Biol. Reprod. 63 , 1787-1794 (2000) | PubMed | 4. Hill, J. R. et al. Clinical and pathologic features of cloned transgenic calv es and fetuses (13 case studies). Theriogenology 51, 1451-1465 (1999) | Article | PubMed | 5. Heyman, Y. et al. Frequency and occurrence of late-gestation losses from catt le cloned embryos. Biol. Reprod. 66, 6-13 (2002) | PubMed | 6. Hasler, J. F. et al. Production, freezing and transfer of bovine IVF embryos and subsequent calving results. Theriogenology 43, 141-152 (1995) | Article | 7. Wells, D. et al. Proc. 32nd Ann. Conf. Soc. Reprod. Biol. Abstract No. 25 (20 01) 8. Tanaka, S. et al. Placentomegaly in cloned mouse concepti caused by expansion of the spongiotrophoblast layer. Biol. Reprod. 65, 1813-1821 (2001) | PubMed | 9. Tamashiro, K. L. et al. Cloned mice have an obese phenotype not transmitted t o their offspring. Nature Med. 8, 262-267 (2002) | Article | PubMed | 10. Ogonuki, N. et al. Early death of mice cloned from somatic cells. Nature Gen et. 30, 253-254 (2002) | Article | PubMed | 11. Lanza, R. P. et al. Cloned cattle can be healthy and normal. Science 294, 18 93-1894 (2001) | Article | PubMed | 12. Campbell, K. H., Loi, P., Otaegui, P. J. & Wilmut, I. Cell cycle co-ordinati on in embryo cloning by nuclear transfer. Rev. Reprod. 1, 40-46 (1996) | PubMed | 13. Ono, Y., Shimozawa, N., Ito, M. & Kono, T. Cloned mice from fetal fibroblast cells arrested at metaphase by a serial nuclear transfer. Biol. Reprod. 64, 44- 50 (2001) | PubMed | 14. Wakayama, T. & Yanagimachi, R. Effect of cytokinesis inhibitors, DMSO and th e timing of oocyte activation on mouse cloning using cumulus cell nuclei. Reprod uction 122, 49-60 (2001) | PubMed | 15. Chesne, P. et al. Cloned rabbits produced by nuclear transfer from adult som atic cells. Nature Biotechnol. 20, 366-369 (2002) | Article | PubMed | 16. Wilmut, I., Schnieke, A. E., McWhir, J., Kind, A. J. & Campbell, K. H. S. Vi able offspring derived from fetal and adult mammalian cells. Nature 385, 810-813 (1997) | PubMed | 17. Kasinathan, P., Knott, J. G., Wang, Z., Jerry, D. J. & Robl, J. M. Productio n of calves from G1 fibroblasts. Nature Biotechnol. 19, 1176-1178 (2001) | Artic le | PubMed | 18. Wells, D. et al. Proc. Curr. Status Perspec. Cloning Stud. 73 (National Inst itute of Agrobiological Sciences, Tsukuba, Japan, 2001) 19. Zhou, Q., Jouneau, A., Brochard, V., Adenot, P. & Renard, J. P. Developmenta l potential of mouse embryos reconstructed from metaphase embryonic stem cell nu clei. Biol. Reprod. 65, 412-419 (2001) | PubMed | 20. Campbell, K. H. S., McWhir, J., Ritchie, W. A. & Wilmut, I. Sheep cloned by nuclear transfer from a cultured cell line. Nature 380, 64-66 (1996) | PubMed | 21. Wakayama, T., Perry, A. C. F., Zuccotti, M., Johnson, K. R. & Yanagimachi, R . Full-term development of mice from enucleated oocytes injected with cumulus ce ll nuclei. Nature 394, 369-374 (1998) | Article | PubMed | 22. Kubota, C. et al. Six cloned calves produced from adult fibroblast cells aft er long-term culture. Proc. Natl Acad. Sci. USA 97, 990-995 (2000) | Article | P ubMed | 23. Cheong, H. T., Takahashi, Y. & Kanagawa, H. Birth of mice after transplantat ion of early cell-cycle-stage embryonic nuclei into enucleated oocytes. Biol. Re prod. 48, 958-963 (1993) | PubMed | 24. Thompson, E. M. Chromatin structure and gene expression in the preimplantati on mammalian embryo. Reprod. Nutr. Dev. 36, 619-635 (1996) | PubMed | 25. Yamazaki, Y. et al. Assessment of the developmental totipotency of neural ce lls in the cerebral cortex of mouse embryo by nuclear transfer. Proc. Natl Acad. Sci. USA 98, 14022-14026 (2001) | Article | PubMed | 26. Hochedlinger, K. & Jaenisch, R. Monoclonal mice generated by nuclear transfe r from mature B and T cells. Nature 415, 967-969 (2002) | Article | PubMed | 27. Wakayama, T. & Yanagimachi, R. Mouse cloning with nucleus donor cells of dif ferent age and type. Mol. Reprod. Dev. 58, 376-383 (2001) | Article | PubMed | 28. Young, L. E. et al. Epigenetic change in IGF2R is associated with fetal over growth after sheep embryo culture. Nature Genet. 27, 153-154 (2001) | Article | PubMed | 29. Humpherys, D. et al. Epigenetic instability in ES cells and cloned mice. Sci ence 293, 95-97 (2001) | Article | PubMed | 30. Inoue, K. et al. Faithful expression of imprinted genes in cloned mice. Scie nce 295, 297 (2002) | Article | PubMed | 31. Daniels, R., Hall, V. & Trounson, A. O. Analysis of gene transcription in bo vine nuclear transfer embryos reconstructed with granulose cell nuclei. Biol. Re prod. 63, 1034-1040 (2000) | PubMed | 32. Rideout, W. M., Eggan, K. & Jaenisch, R. Nuclear cloning and epigentic repro gramming of the genome. Science 293, 1093-1098 (2001) | Article | PubMed | 33. Misteli, T. Protein dynamics: implications for nuclear architecture and gene expression. Science 291, 843-847 (2001) | Article | PubMed | 34. Clarke, H. J., McLay, D. W. & Mohamed, O. A. Linker histone transitions duri ng mammalian oogenesis and embryogenesis. Dev. Genet. 22, 17-30 (1998) | Article | PubMed | 35. Bordignon, V., Clarke, H. J. & Smith, L. C. Factors controlling the loss of immunoreactive somatic histone H1 from blastomere nuclei in oocyte cytoplasm: a potential marker of nuclear reprogramming. Dev. Biol. 233, 192-203 (2001) | Arti cle | PubMed | 36. Nakao, M. Epigenetics: interaction of DNA methylation and chromatin. Gene 27 8, 25-31 (2001) | Article | PubMed | 37. Kikyo, N., Wade, P. A., Guschin, D., Ge, H. & Wolffe, A. P. Active remodelin g of somatic nuclei in egg cytoplasm by the nucleosomal ATPase ISWI. Science 289 , 2360-2362 (2000) | Article | PubMed | 38. Robertson, K. D. & Wolffe, A. P. DNA methylation in health and disease. Natu re Rev. Genet. 1, 11-19 (2000) | Article | PubMed | 39. Fairburn, H. R., Young, L. E. & Hendrich, B. D. Epigenetic reprogramming: ho w now, cloned cow. Curr. Biol. 12, R68-R70 (2002) | Article | PubMed | Acknowledgements. We each acknowledge discussions with colleagues during the dev elopment of th