Monitoring Stem Cell Research

Table of Contents

The Biology of Nuclear Cloning and the Potential of Embryonic Stem Cells for Transplantation Therapy

Rudolf Jaenisch, M.D.
Whitehead Institute,
9 Cambridge Center,
Cambridge, MA

SUMMARY

An emerging consensus is that somatic cell nuclear transfer (SCNT) for the purpose of creating a child (also called "reproductive cloning") is not acceptable for both moral and scientific reasons. In contrast, SCNT with the goal of generating an embryonic stem cell line ("therapeutic cloning") remains a controversial issue. Although therapeutic cloning holds the promise of yielding new ways of treating a number of degenerative diseases, it is not acceptable to many because the derivation of an embryonic stem cell line from the cloned embryo (an essential step in this process) necessarily involves the loss of an embryo and hence the destruction of potential human life.

In this article, I will develop two main arguments that are based on the available scientific evidence. 1) In contrast to an embryo derived by in vitro fertilization (IVF), a cloned embryo has little if any potential to ever develop into a normal human being. This is because, by circumventing the normal processes of gametogenesis and fertilization, nuclear cloning prevents the proper reprogramming of the clone's genome, which is a prerequisite for development of an embryo to a normal individual. It is unlikely that these biological barriers to normal development can be solved in the foreseeable future. Therefore, from a biologist's point of view, the cloned human embryo, used for the derivation of an embryonic stem cell and the subsequent therapy of a needy patient, has little if any potential to create a normal human life.   2) Embryonic stem cells developed from a cloned embryo are functionally indistinguishable from those that have been generated from embryos derived by in vitro fertilization (IVF). Both types of embryonic stem cells have an identical potential to serve as a source for therapeutically useful cells.

It is crucial that the ongoing debate on the possible therapeutic application of SNCT is based on biological facts. The goal of this article is to provide such a basis and to contribute to a more rational discussion that is founded on scientific evidence rather than on misconceptions or misrepresentations of the available scientific data.

I. Introduction

It is important to distinguish between "reproductive cloning" and "nuclear transplantation therapy" (also referred to as "SCNT" or "therapeutic cloning"). In reproductive cloning a cloned embryo is generated by transfer of a somatic nucleus into an enucleated egg with the goal to create a cloned individual. In contrast, the purpose of nuclear transplantation therapy is to generate an embryonic stem cell line (referred to as "ntES cells") that is "tailored" to the needs of a patient who served as the nuclear donor. The ntES cells could be used as a source of functional cells that would be suitable for treating an underlying disease by transplantation.

There is now experience from cloning of seven different mammalian species that is relevant for three main questions of public interest: 1) Would a cloned human embryo be "normal"? 2) Could the problems currently seen with cloning be solved in the foreseeable future? 3) Would ES cells derived from a cloned human embryo be "normal" and useful for cell therapy? The arguments advanced in this article are strictly based on molecular and biological evidence that has been obtained largely in the mouse. I will not attempt to review the cloning literature but only refer to selected papers on cloned mice. The relevant literature on cloning of mammals can be found in recent reviews (Byrne and Gurdon, 2002; Gurdon, 1999; Hochedlinger and Jaenisch, 2002b; Oback and Wells, 2002; Rideout et al., 2001; Wilmut, 2001; Young et al., 1998).

II. Most cloned animals die or are born with abnormalities

The majority of cloned mammals derived by nuclear transfer (NT) die during gestation, and those that survive to birth frequently display "Large Offspring Syndrome", a neonatal phenotype characterized by respiratory and metabolic abnormalities and enlarged and dysfunctional placentas (Rideout et al., 2001; Young et al., 1998).  In order for a donor nucleus to support development into a clone, it must be reprogrammed to a state compatible with embryonic development.  The transferred nucleus must properly activate genes important for early embryonic development and also suppress differentiation-associated genes that had been transcribed in the original donor cell.  Inadequate "reprogramming"¹ of the donor nucleus is most likely the principal reason for developmental failure of clones. Since few clones survive to birth, the question remains whether survivors are fully normal or merely the least affected animals carrying through to adulthood despite harboring subtle abnormalities that originate in faulty reprogramming but that are not severe enough to interfere with survival to birth or beyond.

III. Reprogramming of the genome during normal development and after nuclear transfer

The fundamental difference between nuclear cloning and normal fertilization is that the nucleus used in nuclear cloning comes from a somatic (body) cell that has not undergone the developmental events required to produce the egg and sperm.  Nuclear cloning involves the transplantation of a somatic nucleus into the oocyte from which the nucleus has been removed.  However, the genes in the somatic nucleus are not in the same state as those in the fertilized egg because nuclear transplantation short-cuts the complex process of egg and sperm maturation which involves extensive "reprogramming" of the genome, a process that shuts some genes off and leaves others on. Reprogramming during gametogenesis prepares the genome of the two mature gametes with the ability to activate faithfully the genetic program that ensures normal embryonic development when they combine at fertilization (Fig 1a). This reprogramming of the genome begins at gastrulation, when primordial germ cells (PGCs) are formed, and continues during differentiation into mature gametes resulting, in a radically different chromatin configuration of sperm and oocyte (Rideout et al., 2001).

Experiments have shown that uniparental embryos (embryos whose genomes are derived solely from either the maternal or paternal parent) do not develop normally. Uniparental embryos first seem normal; they direct cleavage (early development to the blastocyst stage) despite profound differences in their epigenetic organization (Reik et al., 2001). However, uniparental embryos fail soon after the implantation of the embryo into the wall of the uterus, indicating that both parental genomes are needed and functionally complement each other beginning at this later step of embryogenesis. Presumably, the different epigenetic organization of the two genomes is crucial for achieving normal development. Moreover, it has been well established that the imbalance of imprinted² gene expression represents an important cause of embryonic failure.

In order for cloned embryos to complete development, genes normally expressed during embryogenesis but silent in the somatic donor cell, must be reactivated. This complex process of epigenetic³ remodeling (i.e., the reconfiguration of the genome by turning on and turning off specific genes) that occurs during gametogenesis in normal development ensures that the genome of the zygote can faithfully activate early embryonic gene expression (Fig 1a). In a cloned embryo, reprogramming, which in normal gametogenesis requires months to years to complete, must occur in a cellular context radically different from gametogenesis and within the short interval (probably within hours) between transfer of the donor nucleus into the egg and the time when zygotic transcription becomes necessary for further development. Given these radically different conditions, one can envisage a spectrum of different outcomes to the reprogramming process ranging from (i) no reprogramming of the genome, resulting in immediate death of the NT embryo; through (ii) partial reprogramming, allowing initial survival of the clones, but resulting in an abnormal phenotype and/or lethality at various stages of development; to (iii) faithful reprogramming producing fully normal animals (Fig 1b). The phenotypes observed over the past five years in cloned embryos and newborns suggest that complete reprogramming is the exception, if it occurs at all.

IV. Development of clones depends on the differentiation-state of the donor nucleus

The majority of cloned embryos fail at an early step of embryonic development, soon after implantation in the wall of the uterus, an early step of embryonic development (Hochedlinger and Jaenisch, 2002b; Rideout et al., 2001). Those that live to birth often display common abnormalities irrespective of the donor cell type (Table 1). In addition to symptoms referred to as "Large Offspring Syndrome", neonate clones often suffer from respiratory distress and kidney, liver, heart or brain defects (Cibelli et al., 2002). However, the abnormalities characteristic of cloned animals are not inherited by their offspring (Tamashiro et al., 2002), indicating that epigenetic aberrations (i.e., failure of genome reprogramming) rather than genetic aberrations (changes in the sequences within the DNA) are the cause.

The efficiency of creating cloned animals is strongly influenced by the differentiation-state of the donor nucleus (Table 1). In the mouse, for example, only 1-3% of cloned blastocysts derived from somatic donor nuclei, e.g., those prepared from fibroblasts or cumulus cells, will develop to adult cloned animals (Hochedlinger and Jaenisch, 2002b).  In certain cases, such as those using terminally differentiated B or T cell donor nuclei, the efficiency of cloning is so low as to preclude the direct derivation of cloned animals. In stark contrast to these examples, cloning using donor nuclei prepared from embryonic stem (ES) cells is significantly more efficient (between 15 and 30 %, Table 1). This correlation with differentiation-state suggests that embryonic nuclei require less reprogramming of their genome, ostensibly because the genes essential for embryonic development are already active and need not be reprogrammed. In fact, the nucleus of an embryonic cell such as an ES cell may well have the same high efficiency to generate postnatal mice after nuclear transfer as the nucleus prepared from a recently fertilized egg  (Table 1, compare Fig. 4). Nonetheless, most if not all mice that have been cloned from ES cell donor nuclei, in contrast to mice derived through natural fertilization from the zygote, are abnormal, indicating that the processes of gametogenesis (development of sperm and of egg) and fertilization endows the zygote nucleus with the ability to direct normal development. In summary, these data indicate that the potential of a nucleus to generate a normal embryo is lost progressively with development.

V. Adult cloned animals: how normal are they?

The observation that apparently healthy adult cloned animals have been produced in seven mammalian species (albeit at low efficiency) is being used by some as a justification for attempting to clone humans. In fact, even those that survive to adulthood, such as Dolly, may succumb relatively early in adulthood because of numerous health problems. Insights into the mechanisms responsible for clone failure before and after birth have come from molecular and biological analyses of mouse clones that have reached (i) the blastocyst stage, (ii) the perinatal period and (iii) adulthood.

(i) Most clones fall short of activating key embryonic genes and fail early.

As stated above in order for clones to develop, the genes that are normally expressed during embryogenesis, but are silent in the somatic donor cell, must be reactivated (Hochedlinger and Jaenisch, 2002b; Rideout et al., 2001). It is the failure to activate key "embryonic" genes that are required for early development that leads to the demise of most clones just after implantation. Recently, a set of about 70 key embryonic genes termed "Oct-4 like" genes have been identified that are active in early embryos but not in somatic donor cells. Importantly, the failure to faithfully activate this set of genes can be correlated with the frequent death of cloned animals during the immediate post-implantation period (Bortvin et al., 2003). These results define "faulty reprogramming" as the cause of early demise of cloned embryos as the failure to reactivate key embryonic genes that are silent in the donor cell.

(ii) Newborn clones misexpress hundreds of genes.

Clones that survive to birth suffer from serious problems, many of which appear to be due to an abnormal placenta. The most common phenotypes observed in animals cloned from either somatic or ES cell nuclei are fetal growth abnormalities such as increased placental and birth weight. This has suggested that surviving clones had accurately reprogrammed the "Oct-4 like" genes that are essential for the earliest stages of development, i.e. those immediately following implantation of the embryo into the uterus. The abnormal phenotype of those clones that do survive through these early stages and develop to birth indicates that other genes that are important for later stages of development but are not essential for early survival are not correctly reprogrammed.  To assess the extent of abnormal expression of various genes in the cells of clones, global gene expression has been assessed by microarray analysis of RNA prepared from the placentas and livers of neonatal cloned mice, i.e., clones that survived development and were viable at birth; these clones had been derived by nuclear transfer (NT) of nuclei prepared either from cultured ES cells or from freshly isolated cumulus cells (somatic cells that surround the egg) (Humpherys et al., 2002).  Direct comparison of gene expression profiles of over 10,000 genes (of the 30,000 or so in the mammalian genome) showed that for both classes of cloned neonatal mice, approximately 4% of the expressed genes in their placentas differed dramatically in expression levels from those in controls, and that the majority of abnormally expressed genes were common to both types of clones.  When imprinted genes, a class of genes that express only one allele (either from maternal or paternal origin), were analyzed, between 30 and 50% were not correctly activated. These data represent strong molecular evidence that cloned animals, even those that survive to birth, suffer from serious gene expression abnormalities.

(iii) Cloned animals develop serious problems with age

The generation of adult and seemingly healthy adult cloned animals has been taken as evidence that normal cloned animals can be generated by nuclear transfer, albeit with low efficiency. Indeed, a routine physical and clinical laboratory examination of 24 cloned cows of 1 to 4 years of age failed to reveal major abnormalities (Lanza et al., 2001). Cloned mice of a corresponding age as that of the cloned cows (2 - 6 months in mice vs. 1 - 4 years in cows) also appear "normal" by superficial inspection. However, when cloned mice were aged, serious problems, not apparent at younger ages, became manifest. One study found that the great majority of cloned mice died significantly earlier than normal mice, succumbing with immune deficiency and serious pathological alterations in multiple organs (Ogonuki et al., 2002). Another study found that aged cloned mice became overweight with major metabolic disturbances (Tamashiro et al., 2002). Thus, serious abnormalities in cloned animals may often become manifest only when the animals age.        

Firm evidence about aging and "normalcy" of cloned farm animals is incomplete or anecdotal because cloned animals of these species are still comparatively young (relative to their respective normal life span). For example, the premature death of Dolly (Giles and Knight, 2003) is entirely consistent with serious abnormalities in cloned sheep that become manifest only at later ages. Also, two of the analyzed cloned cows developed disease soon after the study on "healthy and normal cattle" (Lanza et al., 2001) had appeared: one animal developed an ovarian tumor and another one suffered brain seizures (J. Cibelli, pers. comm.). While it cannot be ruled out that these are "spontaneous" maladies unconnected with the cloning procedure, a more likely alternative is that these problems were direct consequences of the nuclear transfer procedure.

(iv) Are there any "normal" clones?

It is a key question in the public debate whether it is ever possible to produce a normal individual by nuclear cloning, even if only with low efficiency. The available evidence suggests that it may be difficult if not impossible to produce normal clones for the following reasons: 1) As summarized above, all analyzed clones at birth showed dysregulation of hundreds of genes. The development of clones to birth and beyond despite widespread epigenetic abnormalities suggests that mammalian development can tolerate dysregulation of many genes. 2) Some clones survive to adulthood by compensating for gene dysregulation. Though this "compensation" assures survival, it may not prevent maladies to become manifest at later ages. Therefore, most if not all clones are expected to have at least subtle abnormalities that may not be so severe as to result in an obvious phenotype at birth but will cause serious problems later as seen in aged mice. Clones may just differ in the extent of abnormal gene expression: if the key "Oct-4 like" genes are not activated, clones die immediately after implantation. If those genes are activated, the clone may survive to birth and beyond.

As schematically shown in Fig. 2, the two stages when the majority of clones fail are immediately after implantation and at birth. These are two critical stages of development that may be particularly vulnerable to faulty gene expression. Once cloned newborns have progressed through the critical perinatal period,  various compensatory mechanisms may counterbalance abnormal expression of other genes that are not essential for the subsequent postnatal survival. However, the stochastic occurrence of disease and other defects at later age in many or most adult clones implies that such compensatory mechanisms do not guarantee "normalcy" of cloned animals. Rather, the phenotypes of surviving cloned animals may be distributed over a wide spectrum from abnormalities causing sudden demise at later postnatal age or more subtle abnormalities allowing survival to advanced age (Fig. 2). These considerations illustrate the complexity of defining subtle gene expression defects and emphasize the need for more sophisticated test criteria such as environmental stress or behavior tests. However, the available evidence suggests that truly normal clones may be the exception.

It should be emphasized that "abnormality" or "normalcy" is defined here by molecular and biological criteria that distinguish cloned embryos or animals from control animals produced by sexual reproduction. The most informative data for the arguments presented above come from the mouse. There is, however, every reason to believe that these difficulties associated with producing mice and a variety of other mammalian embryos by nuclear transplantation will also afflict the process of human reproductive cloning (Jaenisch and Wilmut, 2001).

(v) Is it possible to overcome the problems inherent in reproductive cloning?

It is often argued that the "technical" problems in producing normal cloned mammals will be solved by scientific progress that will be made in the foreseeable future. The following considerations argue that this may not be so.

A principal biological barrier that prevents clones from being normal is the "epigenetic" difference (such as distinct patterns of DNA methylation⁴) between the chromosomes inherited from mother and from father, i.e. the difference between the "maternal" and the "paternal" genome of an individual. Such methylation of specific DNA sequences is known to be responsible for shutting down the expression of nearby genes.  Parent-specific methylation marks are responsible for the expression of imprinted genes and cause only one copy of an imprinted gene, derived either from sperm or egg, to be active while the other allele is inactive (Ferguson-Smith and Surani, 2001). When sperm and oocyte genomes are combined at fertilization, the parent-specific marks established during oogenesis and spermatogenesis persist in the genome of the zygote (Fig 3A). Of interest for this discussion is that within hours after fertilization, most of the global methylation marks (with the exception of those on imprinted genes) are stripped from the sperm genome whereas the genome of the oocyte is resistant to this active demethylation process (Mayer et al., 2000). This is because the oocyte genome is in a different "oocyte-appropriate" epigenetic state than the sperm genome. The oocyte genome becomes only partially demethylated within the next few days by a passive demethylation process. The result of these post-fertilization changes is that the two parental genomes are epigenetically different (as defined by the patterns of DNA methylation) in the later stage embryo and remain so in the adult in imprinted as well as non-imprinted sequences.

In cloning, the epigenetic differences that are established during gametogenesis may be erased because both parental genomes of the somatic donor cell are introduced into the egg from the outside and are thus exposed equally to the demethylation activity present in the egg cytoplasm (Fig 3B). This predicts that imprinted genes should be particularly vulnerable to inappropriate methylation and associated dysregulation in cloned animals. The results summarized earlier are consistent with this prediction. For cloning to be made safe, the two parental genomes of a somatic donor cell would need to be physically separated and separately treated in an "oocyte-appropriate" and a "sperm-appropriate" way, respectively. At present, it seems that this is the only rational approach to guarantee the creation of the epigenetic differences that are normally established during gametogenesis. Such an approach is beyond our present abilities. These considerations imply that serious biological barriers exist that interfere with faithful reprogramming after nuclear transfer. It is a safe conclusion that these biological barriers represent a major stumbling block to efforts aimed at making nuclear cloning a safe reproductive procedure for the foreseeable future.

It has been argued that the problems in mammalian cloning are similar to those encountered with IVF 30 years ago: Thus, following this argument, the methods of culture and embryo manipulations just would need to be improved to develop reproductive cloning into a safe reproductive technology that is as acceptable as IVF. This argument appears to be fundamentally flawed. It is certainly correct that merely "technical" problems needed to be solved to make IVF efficient and safe.  It is important to distinguish between the perfection of  technical skills to imitate a biological event and the development of wholly new science to overcome the blocks to events that have severe biological restrictions. Nuclear cloning faces serious biological barriers that cannot be addressed by mere adjustments in experimental technique. Indeed, since the birth of Dolly no progress has been made in solving any of the underlying biological issues of faulty gene reprogramming and resulting defective development.    

VI. Therapeutic applications of SCNT

(i) Reproductive cloning vs. therapeutic cloning

In spite of the biological and ethical barriers associated with reproductive cloning, nuclear transfer technology has significant therapeutic potential that is within our grasp. There is an enormous distinction between the goals and the end product of these two technologies. The purpose of reproductive cloning is to generate a cloned embryo that is then implanted in the uterus of a female to give rise to a cloned individual. In contrast, the purpose of nuclear transplantation therapy is to generate an embryonic stem cell line that is derived from a patient (referred to as "ntES cells") and can be used subsequently for tissue replacement.

Many scientists recognize the potential of NtES cells for organ transplantation (for recent review see (Hochedlinger and Jaenisch, 2003). This procedure is currently complicated by immune rejection due to immunological incompatibility. Thus, virtually all organ transplants undertaken at present involve the use of donor organs that are recognized as foreign by the immune systems of the recipient and thus are targeted for destruction by these immune systems. To treat this "host versus graft" disease, immunosuppressive drugs are routinely given to transplant recipients in order to suppress this organ rejection. Such immunosuppressive treatment has serious side effects including increased risks of infections and malignancies.  In principle, ES cells can be created from a patient's nuclei using nuclear transfer.  Because ntES cells will be genetically identical to the patient's cells, the risks of immune rejection and the requirement for immunosuppression are eliminated. Moreover, ES cells provide a renewable source of replacement tissue allowing for repeated therapy whenever needed. Finally, if ES cells are derived from a patient carrying a known genetic defect, the mutation in question can be corrected in the ntES cells using standard gene targeting methods before introducing these ES cells (or derived tissue-specific stem cells) back into the patient's body.

(ii) Combining nuclear cloning with gene and cell therapy

In a "proof of principle" experiment, nuclear cloning in combination with gene and cell therapy has been used to treat a mouse genetic disorder that has a human counterpart (Figure 4). To do so, the well-characterized Rag2 mutant mouse was used as "patient" (Rideout et al., 2002). This mutation causes severe combined immune deficiency (SCID), because the enzyme that catalyzes immune receptor rearrangements in lymphocytes is non-functional. Consequently, these mice are devoid of mature B and T cells, a disease resembling human Omenn syndrome (Rideout et al., 2002).

In a first step, somatic (fibroblast) donor cells were isolated from the tails of Rag2-deficient mice and their nuclei were injected into enucleated eggs. The resultant embryos were cultured to the blastocyst stage and isogenic ES cells were isolated. Subsequently, one of the mutant Rag2 alleles was targeted by homologous recombination in ES cells to restore normal Rag2 gene structure and function. In order to obtain somatic cells for treatment, these genetically repaired ES cells were differentiated into embryoid bodies and further into hematopoietic precursors by expressing HoxB4, a transcription factor that is responsible for programming the behavior of the hematopoietic stem cells, i.e., those cells that are able to generate the full range of red and white cells in the blood. Resulting hematopoietic precursors were transplanted into irradiated Rag2-deficient animals in order to treat the disease caused by their Rag2 mutation. Initial attempts to engraft these cells were, however, unsuccessful because of an increased level of natural killer (NK) cells in the Rag mutant host.  ES cell derived hematopoietic cells express low levels of the MHC antigens and thus are a preferred target for NK mediated destruction. Elimination of NK cells by antibody depletion or genetic ablation allowed the ntES cells to efficiently populate the myeloid and to a lesser degree the lymphoid lineages of these mice. Functional B and T cells that had undergone proper rearrangements of their immunoglobulin and T cell receptor alleles as well as serum immunoglobulins were detected in the transplanted mutants.  Hence, important cellular components of the immune system were restored in mice that previously were unable to produce these cells.

This experiment demonstrated that embryonic stem (ES) cells derived by NT from somatic cells of a genetically afflicted individual can be combined with gene therapy to treat the underlying genetic disorder. Because Rag2 deficiency causes an increase in NK activity and necessitated the elimination of NK cells prior to transplantation in the above-described experiments, some have concluded that "The experiment failed to show success with therapeutic cloning" (Coalition and Ethics, 2003) and that "This indicates that the only successful therapy using cloned embryos would be through 'reproductive' cloning, to produce born clones who can serve as tissue donors for patients" (Prentice, 2002). This is a troubling misinterpretation of the data. (i) It has been shown that ES cell-derived hematopoietic cells can successfully engraft and rescue lethally irradiated mice indicating that increased NK activity is a peculiarity of Rag2-deficiency (Kyba et al., 2002).  Therefore, it would seem that for most diseases, no anti-NK treatment would be required to assure engraftment of ES cell-derived somatic cells. (ii) It is correct that treatment of a human patient with Omenn syndrome, which is equivalent to Rag2 deficiency, by SCNT may also require anti-NK treatment to transiently reduce NK activity. This would allow the transplanted cells to engraft as in the mouse experiment.  Once these cells are successfully engrafted, there is every reason to believe that such anti-NK treatment would no longer be necessary.

In conclusion, the mouse experiment indicates that, unlike the situation with reproductive cloning, no biological barriers exist that in principle prevent the use of SCNT to treat human diseases. The technical issues in using SCNT and human stem cells for therapeutic purposes need, however, to be solved, but there are no indications at present that these represent formidable problems that will resist relatively rapid solution.

VII. Faulty reprogramming after nuclear transfer: does it interfere with the therapeutic potential of ES cells derived by SCNT?

As summarized above, most if not all cloned animals are abnormal because of faulty reprogramming after nuclear transfer. Does this epigenetic dysregulation affect the potential of ntES cells to generate functional somatic cells that can be used for cell therapy? To address this question, I will first compare the in vivo development of embryos with the in vitro process of ES cell derivation from explanted embryos. This will be followed by discussing the epigenetic state of the ES cell genome. Finally, I will contrast the phenotype of cloned mice derived from ES cell donor nuclei with that of chimeric mice generated by injection of ES cells into blastocysts.

(i) The phenotype of an  embryo is determined by  its donor nucleus

As mentioned repeatedly above, embryos can be derived from the fertilized egg or from a somatic nucleus by SCNT. The potential of the resulting blastocyst, when implanted into the womb, to develop into a fetus and a postnatal animal depends strictly on the nature of the donor nucleus (Fig 5): (i) When derived from the zygote, most embryos develop to birth and generate a normal animal; (ii) Similarly, most blastocysts cloned from an embryonic stem cell donor nucleus develop to birth but, in contrast to the normally fertilized embryo, the great majority of the cloned animals will be abnormal ("Large offspring syndrome") (Eggan et al., 2001; Humpherys et al., 2001); (iii) The great majority of cloned blastocysts derived from somatic donor nuclei such as fibroblasts or cumulus cells will die soon after implantation and only a few clones will survive to birth and these too will be abnormal, suffering once again from the Large offspring syndrome (Wakayama and Yanagimachi, 2001); (iv) Finally, the likelihood of cloned blastocysts derived from another type of somatic donor nuclei - those present in terminally differentiated lymphoid cells - to generate a cloned animal is extremely low and has not been achieved except by using a two step procedure involving the intermediate generation of embryonic stem cells (Hochedlinger and Jaenisch, 2002a). These observations suggest that a blastocyst retains an "epigenetic memory" of its donor nucleus. This memory determines its potential for fetal development: while a fertilized embryo develops normally, any embryo derived by SCNT will be abnormal though the efficiency of a given clone to develop to birth is strongly influenced by the differentiation state of the donor cell (see Table 1). In other words, the cloned embryo after implantation into the womb will be abnormal because the cloned blastocyst retained an epigenetic memory of its donor nucleus and this causes faulty fetal development. This epigenetic memory is erased when a blastocyst, either derived by nuclear cloning or from the fertilized egg, is explanted into tissue culture and grown into an embryonic stem cell. Erasure of the epigenetic memory has major consequences for the "normalcy" of embryonic stem cells.

(ii) The derivation of embryonic stem cells is a highly selective process that erases the "epigenetic memory" of the donor nucleus

Embryonic stem cells, regardless of whether they have been generated from a fertilized egg or by SCNT, are derived from the cells of a blastocyst that have been explanted and propagated in tissue culture. Of the blastocyst cells that are explanted in this way, those that derive from the portion of the blastocyst termed the inner cell mass (ICM) initially express "key" embryonic genes such as Oct-4. However, soon after explantation, most ICM cells extinguish Oct-4 expression and cease proliferating (Buehr et al., 2003). Only one or a few of the ICM-derived cells will eventually re-express Oct-4 and these few Oct-4-positive cells are those that resume rapid proliferation, yielding the cell populations that we designate as "embryonic stem" cells. These cells represent a cell population that has no equivalent in the normal embryo and may be considered a tissue culture artifact, though a useful one (Fig. 6).

The important point for this discussion is that the propagation of blastocyst cells in vitro results in a rare population of surviving cells that have erased the "epigenetic memory" of the donor nucleus. This process results ultimately in ES cells that have, regardless of donor nuclear origin, an identical developmental potential. In other words, ES cells derived from embryos produced by normal fertilization and those produced from cloned embryos are functionally indistinguishable (Hochedlinger and Jaenisch, 2002b; Rideout et al., 2002; Wakayama et al., 2001). Because the ES cells that derive from normally fertilized embryos are able to participate in the generation of all normal embryonic tissues, we can conclude that the ES cells derived from cloned embryos have a similar potential to generate the full range of normal tissues. 

(iii) ES cells, epigenetic instability and therapeutic potential

Epigenetic instability appears to be a consistent characteristic of ES cells. This was shown when individual ES cells were analyzed for expression of imprinted genes: even cells in a recently subcloned ES cell line differed strongly in the expression of genes such as H19 or Igf2. The variable expression was correlated with the DNA methylation status of the genes, which switched from an unmethylated to a methylated state between sister cells (Humpherys et al., 2001). This was a surprising result in view of the known potential of ES cells to generate terminally differentiated cells that function normally after transplantation into an animal. Possible explanations include (i) that epigenetic instability in ES cells is a consequence of propagation of cells in tissue culture or (ii) that epigenetic instability is a prerequisite for cells to be pluripotent, i.e., this instability may be a manifestation of a plasticity in the gene expression program that is required to enable the ES cells to generate a wide variety of differentiated cell lineages.

Whatever the explanation for the observed epigenetic instability of ES cells may be, it supports the view that the process of generating ES cells erases all epigenetic memory of the donor nucleus and, as a consequence of the selection process, generates epigenetic instability in the selected cells. In other words, epigenetic instability appears to be an intrinsic characteristic of ES cells regardless of whether derived by SCNT or from a fertilized egg. This is consistent with the conclusion that both types of ES cells have an equivalent potency to generate functional cells in culture and, in the longer term, fully normal differentiated tissues upon implantation of these cells in vivo. 

(iv) ES cells form normal chimeras but abnormal nuclear clones

As outlined above, faulty reprogramming leads to abnormal phenotypes of cloned mice derived from ES cell donor nuclei. Why is faulty reprogramming and epigenetic instability a problem for reproductive cloning but not for therapeutic applications? The main reason for this seeming paradox is that, in contrast to reproductive cloning, the therapeutic application of NT does not require the formation of a fetus. Therapeutic applications involve the ability of cloned ES cells to form a single tissue or organ, not to recapitulate all of fetal development. For example, normal fetal development requires faithful expression of the imprinted genes. As outlined above, nuclear cloning causes between 30% and 50% of imprinted genes to be dysregulated consistent with the notion that disturbed imprinting is a major contributing factor to clone failure. As most imprinted genes have no known function in the postnatal animal, the dysregulation of imprinting would not be expected to impede functionality of in vitro differentiated ES cells because this process does not require the formation of a fetus. Therefore, the functionality of mature cells derived in culture from ES cells would not depend on the faithful reprogramming of the imprinted genes. Dysregulation of some imprinted genes such as Igf2 are known, however, to cause disease in the adult. Thus, it will be important to test whether dysregulation of such genes has adverse effects on the function of somatic cells derived from ES cells.

When injected into a blastocyst, ES cells form normal chimeras. It appears that the presence of surrounding "normal" cells, i.e. cells that are derived from a fertilized embryo, prevents an abnormal phenotype of the chimera such as the "Large Offspring Syndrome" that is typical for cloned animals. Any therapeutic application creates, of course, a chimeric tissue where cells derived from ntES cells are introduced into a diseased adult individual and interact with surrounding "normal" host cells. Therefore, no phenotypic abnormalities, such as those seen in cloned animals, would be expected in patients transplanted with cells derived from ntES cells.

VIII. SCNT for cell therapy: destruction of potential human life?

A key concern raised against the application of the nuclear transplantation technology for tissue therapy in humans is the argument that the procedure involves the destruction of potential human life. From a biological point of view, life begins with fertilization when the two gametes are combined to generate a new embryo that has a unique combination of genes and has a high potential to develop into a normal baby when implanted into the womb. A critical question for the public debate on SCNT is this one: is the cloned embryo equivalent to the fertilized embryo?

In cloning, the genetic contribution is derived from one individual and not from two. Obviously, the cloned embryo is the product of laboratory-assisted technology, not the product of a natural event. From a biological point of view, nuclear cloning does not constitute the creation of new life, rather the propagation of existing life because no meiosis, genetic exchange and conception are involved. Perhaps more important is, however, the overwhelming evidence obtained from the cloning of seven different mammalian species. As summarized above, the small fraction of cloned animals that survive beyond birth, even if they appear "normal" upon superficial inspection, are likely not so. The important conclusion is that a cloned human embryo would have little if any potential to develop into a normal human being. With other words, the cloned human embryo lacks essential attributes that characterize the beginning of normal human life.

Taking into account the potency of fertilized and cloned embryos, the following scenarios regarding their possible fates can be envisaged (Fig. 7). Fertilized embryos that are "left over" from IVF have three potential fates: disposal, generation of normal embryonic stem cells or generation of a normal baby when implanted into the womb. Similarly, the cloned embryo has three potential fates: it can be destroyed or could be used to generate a normal ntES cell line that has the same potential for therapy as an ES cell derived from a fertilized embryo. In contrast to the fertilized embryo, the cloned embryo has little if any potential to ever generate a normal baby.An embryonic stem cell line derived by nuclear may, however, help sustain existing life when used as a source for cell therapy that is "tailored" to the need of the patient who served as its nuclear donor.

If SCNT were accepted as a valid therapeutic option, a major concern of its implementation as medical procedure would be the problem of how to obtain sufficient numbers of human eggs that could be used as recipients. Commercial interests may pressure women into an unwanted role as egg donors. The recent demonstration that embryonic stem cells can be coaxed into a differentiation pathway that yields oocyte-like cells (Hubner et al., 2003) may offer a solution to this dilemma. If indeed functional oocytes could be generated from a generic human ES cell line, sufficient eggs could be generated in culture and serve as recipients for nuclear transfer without the need of a human egg donor. It seems that technical issues, not fundamental biological barriers, need to be overcome so that transplantation therapy can be carried out without the use of human oocytes.

Fig 1:Reprogramming in normal development and nuclear cloning.

a. The genome of primordial germ cells (PGCs) is hypomethylated ("reset", white boxes). Reprogramming and establishment of parent specific epigenetic marks occurs over the course of gametogenesis so that the genome of sperm and egg are competent to express the genes that need to be activated in early embryonic (box with wavy lines) and later (hatched box) development. During cleavage and early postimplantation development "embryonic" genes, such as Oct 3/4, become activated (black box) and are repressed at later stages (stippled boxes) when tissue specific genes (hatched boxes) are activated in adult tissues (A, B, C). Epigenetic reprogramming of imprinted and non-imprinted genes occurs during gametogenesis in contrast to X inactivation and the readjustment of telomere length which take place postzygotically.

b. Reprogramming of a somatic nucleus following nuclear transfer may result in (i) no activation of "embryonic" genes and early lethality, (ii) faulty activation of embryonic genes and an abnormal phenotype, or (iii) in faithful activation of "embryonic" and "adult" genes and normal development of the clone. The latter outcome is the exception if it occurs at all.

Fig. 2: The phenotypes are distributed over a wide range of abnormalities. Most clones fail at two defined developmental stages, implantation and birth. More subtle gene expression abnormalities result in disease and death at later ages.

A: The  genomes of oocyte and sperm are differentially methylated during gametogenesis and are different in the zygote when combined at fertilization. Immediately after fertilization the paternal genome (derived from the sperm) is actively demethylated whereas the maternal genome is only partially demethylated during the next few days of cleavage.  This is because the oocyte genome is in a different chromatin configuration and is resistant to the active demethylation process imposed on the sperm genome by the egg cytoplasm. Thus, the methylation of two parental genomes is different at the end of cleavage and in the adult. Methylated sequences are depicted as filled lollipops and unmethylated sequences as empty lollipops.

3B: In cloning a somatic nucleus is transferred into the enucleated egg and both parental genomes are exposed to the active demethylating activity of the egg cytoplasm. Therefore, the parent specific epigenetic differences are equalized.

Fig. 4:Scheme for therapeutic cloning combined with gene and cell therapy.

A piece of tail from a mouse homozygous for the recombination activating gene 2 (Rag2) mutation was removed and cultured. After fibroblast-like cells grew out, they were used as donors for nuclear transfer by direct injection into enucleated MII oocytes using a Piezoelectric driven micromanipulator. Embryonic stem (ES) cells isolated from the NT-derived blastocysts were genetically repaired by homologous recombination. After repair, the ntES cells were differentiated in vitro into embryoid bodies (EBs), infected with the HoxB4iGFP retrovirus, expanded, and injected into the tail vein of irradiated, Rag2-deficient mice (after (Rideout et al., 2002)).

Fig. 5: Blastocysts retain epigenetic memory of donor nucleus

Blastocysts can be derived from the fertilized egg or by nuclear transfer. After implantation development of the embryo strictly depends on the donor nucleus:  Blastocysts derived from a fertilized egg  will develop with high efficiency to normal animals; blastocysts derived by NT from an ES cell donor will develop with high efficiency to abnormal animals; blastocysts derived by NT from a fibroblast or cumulus cell donor will develop with low efficiency to abnormal animals; blastocysts derived by NT from B or T donor cells will not develop to newborns by direct transfer into the womb (only by a 2 step procedure, compare (Hochedlinger and Jaenisch, 2002a).

Fig. 6: The establishment of ES cells from blastocysts erases epigenetic memory of donor nucleus

Most cells of the inner cell mass turn off Oct-4 like genes and die after explantation of blastocysts into tissue culture. Only one or a few cells turn on the Oct-4 like genes and proliferate. The surviving cells will give rise to ES cells. During this highly selective outgrowth of the surviving cells all epigenetic memory of the donor nucleus is erased. Therefore, regardless of donor nucleus (fertilized egg or somatic nucleus in cloned embryos), all ES cells have an equivalent potency to generate functional differentiated cells.

Fig. 7: Normal and cloned embryos have three possible fates

Embryos derived by IVF ("left over embryos") have three fates: they can be disposed, create normal babies if implanted or can generate ES cells if explanted into tissue culture. Cloned embryos have also three fates: they can be disposed, can generate abnormal babies if any when implanted or can generate ES cells when explanted. The ES cells derived from an IVF embryo or a cloned embryo are indistinguishable (same potency, see figure 6)

Table 1

Development of normal embryos and embryos cloned from ES cell and somatic donor nuclei. Note that normal and ES cell derived blastocysts have a similar potency to develop to term if calculated from the fraction of transplanted blastocysts.

1: (Eggan et al., 2001; Eggan et al., 2002; Rideout et al., 2000); 2 (Wakayama et al., 1998; Wakayama and Yanagimachi, 1999); 3 (Hochedlinger and Jaenisch, 2002a).

_________________

Acknowledgements

I thank my colleagues Bob Weinberg, Gerry Fink, George Daley and Andy Chess for critical and constructive comments on this manuscript.