Nature Biotechnology
22, 25 - 26 (2004)
doi:10.1038/nbt0104-25
Do cloned mammals skip a reprogramming step?Josef Fulka Jr.1, Norikazu Miyashita2, Takashi Nagai2
& Atsuo Ogura31 J.F. Jr. is in Center for Cell Therapy and Tissue Repair, Prague, DOB1, CS-104 01, Prague 10, Czech Republic 2 N.M. and T.N. are in the National Institute of Agrobiological Sciences, Tsukuba, Kannondai 2-1-2, 305-8602 Japan 3 A.O. is in RIKEN BioResource Center, 3-1-1, Koyadai, ScubaInstitute, Tsukuba, 305-0074, Japan.
Correspondence should be addressed to Takashi Nagai taku@affrc.go.jpIt is widely accepted that at least some populations of cloned animals have an attenuated lifespan compared with their conventionally bred counterparts. This has been attributed both to premature aging or senescence and to accumulation of abnormalities in gene expression in their tissues. Here, we argue that these problems arise because the process of nuclear transfer used to create cloned animals skips one of the two essential, independent steps involved in the reprogramming of cell nuclei.
Senescence
Since the birth of Dolly the sheep in 1996, the 'real biological age' of cloned animals has been a matter of much debate1. It has been argued that Dolly was either 6 years old (on the basis of her date of birth) or 12 years old (on the basis of the age of the donor mammary gland cell used in her creation) when she was euthanased because of serious progressive lung disease.
One proposed means of assessing a clone's age is by measuring the length of its telomeres and the speed of their erosion. Simple measurements of telomere length suggest that cloned animals have telomeres that are similar in length to, or even longer than, telomeres from naturally bred animals2,
3,
4,
5. Telomeres from Dolly6 and from bovine clones7 are shorter than those of age-matched controls, however. In addition, certain cloned animals have even shorter telomeres than those in the somatic donor cells from which they were actually derived. As many cloned large animals reach 4−6 years of age with no signs of premature aging, these variations and errors in telomere restoration do not necessarily seem to lead to premature aging.
A comparison by Clark et al.8 of in vitro culture parameters and characteristics of sheep fibroblast cells used as nuclear donors in cloning and cells derived from corresponding cloned fetuses showed that complete telomere restoration is not necessarily achieved after nuclear transfer; in fact, the proliferation and lifespan of the cloned cells from the fetus are the same as those of the donor cell line. The authors thus concluded that the lifespan of a clone is influenced by the genetically determined speed of telomere erosion.
Because of their short life cycle, mice are an ideal system for studying the longevity of cloned animals. A study by Ogonuki et al.9 showed that cloned mice die significantly earlier than controls. As many of the cloned mice suffer from serious pathologies (e.g., pneumonia and hepatic failure), however, premature aging might not be the primary cause of death.
What is normal?
This raises the question of whether any clones are completely 'normal'10? The expression of several (imprinted and nonimprinted) genes differs substantially in cloned animals compared with conventionally bred counterparts. Of about 10,000 genes analyzed in mouse clones, approximately 400 show abnormal expression patterns, especially in placentas11. Notably, aberrant expression seems to be somewhat tissue-specific, with nonplacental organs having a lesser extent of abnormal gene expression.
A similar analysis of expression of genes in the Oct4 group in mice showed that embryos derived from embryonic stem cells have a normal expression pattern12,
13, whereas blastocysts produced by somatic cell transfer have abnormal expression (additional factors, e.g., culture conditions, may also influence the expression of certain genes14,
15). Thus, we conclude that the premature aging of clones is not the only (or the main) reason why cloned animals die earlier than naturally bred counterparts.
Reprogramming by steps
To elucidate this phenomenon, we must look more closely at the reprogramming of the nucleus after its transfer to the recipient cell. There are essentially two independent natural periods when cell nuclei can be reprogrammed. The first period begins immediately after fertilization when, for example, the paternal chromatin is intensively demethylated. The embryo methylation level reaches its lowest phase at the blastocyst stage (by day 3.5 in the mouse), and the methylation pattern is gradually established thereafter, the exact time depending on the cell line16,
17.
The second reprogramming period occurs in developing germline cells. For example, in mouse primordial germ cells, the imprinting memory established in parental gametes is erased between days 10 and 12 of pregnancy18. On the basis of cloning studies, we may assume that the purpose of this second reprogramming phase is to erase, by as yet unknown mechanisms, all the epigenetic errors that had been accumulated before, and that this reprogramming step enhances the chance of producing error-free gametes.
Thus, at fertilization, both spermatozoa and oocytes should be epigenetically error-free. Certainly, this is not the case for a somatic cell nucleus used for nuclear transfer. Moreover, fertilization has been honed by millions of years of evolution to ensure that sperm (donor) and oocyte (recipient) are uniquely prepared to ensure fidelity of nuclear imprinting.
We suggest that, at present, a complete reprogramming in cloning is only possible through these two steps (Fig. 1). The first reprogramming step occurs in oocytes to initialize the memory of the differentiated somatic cells. The second reprogramming step occurs as chosen cells (primordial germ cells and their successor cells) in a given clone pass through the germ-cell formation processes. This is also supported by results from obese or otherwise abnormal mouse clones, whose phenotypes are not manifested in their offspring19,
20. Also, telomeres in spermatozoa from cloned bulls are the same length as telomeres in controls, whereas telomeres in their somatic cells are shorter7,
21. This supports the notion that gametes in clones are error-free19,
20. There is, however, no chance of developing cloned animals whose cells pass through the second step22.
Conclusions
During reproduction, reprogramming occurs in two steps. The first reprogramming event results in the initial de-differentiation of the transferred nucleus, making it competent to direct the development of the embryo. The second reprogramming event has at least three roles: first, epigenetic errors are erased (by as yet unknown mechanisms); second, genomic imprinting is erased and reestablished; and third, telomere length is adjusted definitively, following elongation at the first reprogramming and subsequent gametogenesis.
We presume that cloned animals die earlier not because they are biologically too old, but because they accumulate abnormalities in expression of different genes. When single cells are isolated from cloned fetuses or animals, their proliferation and viability are normal8. This has also been recently shown in intestine-derived cloned blastulae from amphibians that were transferred to normal host embryos; after several months, the transferred cells contributed to several host tissues23.
Our conclusions have several implications for biotechnology. First, cells obtained by 'therapeutic cloning' will probably have the same life span as normal cells but may have abnormal gene expression caused by epigenetic errors. Second, the progeny of cloned animals will be normal. This is especially important for the use of cloned animals in xenotransplantation or the production of valuable pharmaceutical proteins in their milk24,
25. Third, and perhaps most important, the problems stated above argue against the application of human reproductive cloning. The incomplete reprogramming of donor nuclei during somatic cell nuclear transfer will probably have such dire effects on gene expression and health that the production of children by such techniques as presently available should be prohibited.
REFERENCES
- Wilmut, I., Schnieke, A.E., McWhir, J., Kind, A. & Campbell, K.H.S. Nature 385, 810–813 (1997).
- Lanza, R.P. et al. Science 288, 665–669 (2000).
- Tian, X.C., Xu, J. & Yang, X. Nat. Genet. 26, 272–273 (2000).
- Wakayama, T. et al. Nature 407, 318–319 (2000).
- Betts, D.H. et al. Proc. Natl. Acad. Sci. USA 96, 1077–1082 (2001).
- Shiels, P.G. et al. Nature 399, 316–317 (1999).
- Miyashita, N. et al. Biol. Reprod. 66, 1649–1655 (2002).
- Clark, A.J. et al. Nat. Cell Biol. 5, 535–538 (2003).
- Ogonuki, N. et al. Nat. Genet. 30, 253–254 (2002).
- Wilmut, I. Nat. Med. 8, 215–216 (2002).
- Humpherys, D. et al. Proc. Natl. Acad. Sci. USA 99, 12889–12894 (2002).
- Boiani, M., Eckardt, S., Scholer, H.R. & McLaughlin, K.J. Genes Dev. 16, 1209–1219 (2002).
- Bortvin, A. et al. Development 130, 1673–1680 (2003).
- Rideout III, W.M., Eggan, K. & Jaenisch, R. Science 293, 1093–1098 (2001).
- Wilmut, I. et al. Nature 419, 583–586 (2002).
- Reik, W. & Walter, J. Nat. Rev. Genet. 2, 21–32 (2001).
- Li, E. Nat. Rev. Genet. 3, 662–673 (2002).
- Lee, J. et al. Development 129, 1807–1817 (2002).
- Tamashiro, K.L.K. et al. Nat. Med. 8, 262–267 (2002).
- Shimozawa, N. et al. Genesis 34, 203–207 (2002).
- Miyashita, N. et al. Theriogenology 59, 1557–1565 (2003).
- Hochedlinger, K. & Jaenisch, R. N. Engl. J. Med. 349, 275–286 (2003).
- Byrne, J.A., Simonsson, S. & Gurdon, J.B. Proc. Natl. Acad. Sci. USA 99, 6059–6063 (2002).
- Prather, R.S., Hawley, R.J., Carter, D.B., Lai, L. & Greenstein, J.L. Theriogenology 59, 115–123 (2003).
- Robl, J.M. et al. Theriogenology 59, 107–113 (2003).
Acknowledgments
J.F.Jr. appreciates support from The Japan Society for the Promotion of Sciences (S03152).
|
