Most stem cell biologists agree that human embryonic stem (hES) cells have tremendous potential for treating human disease and injury. This potential is based upon the observation that these cells can, under appropriate conditions in vitro or in vivo, differentiate into most, if not all, cell types of the adult human body . The path to successful hES cell therapies would therefore seem straightforward: make the desired cell type from hES cells, such as neurons to treat neurodegenerative disease or pancreatic β-cells to treat type I diabetes, and then transfer these cells to the desired site. Embryonic stem (ES)-derived cells could also be used as a source of needed growth factors or signaling molecules. A large body of literature already exists on differentiation and transplantation into mice of mouse ES cells (for example [2, , –5]), and there is a growing literature on the differentiation of hES cells, in vitro and following transplantation (for example [6, , –9]). But a number of concerns, both scientific and ethical, are impeding the progress of this work. Our goal here is to review these problems and consider how we might overcome them.
One obstacle that we do not explicitly address is the reluctance on the part of many scientists to take time to understand the nature of the ethical concerns and begin developing reasoned ethical responses to these concerns. Those engaged in the ethical and political debate about the generation and use of hES cells must understand the science, whether or not they are scientists. Scientists who use “ethically sensitive material” play an important role in informing these debates and will be in the best position to do so when they develop their ethical reasoning skills and their abilities to discuss and defend their work on ethical grounds. In what follows, we illustrate this process of integrating scientific and ethical concerns by analyzing six areas of particular interest: tumorigenicity, animal product contamination, genetic compatibility, funding, cell type for transplantation, and “embryo-friendly” derivation methods.
1. Tumor Formation
Any ES cell-based therapy includes the risk of tumors forming from undifferentiated transplanted cells. Researchers are defining approaches to inhibit this outcome by eliminating tumorigenic cells either prior to or subsequent to transplantation. As with any experimental therapy, care must be taken in determining what would be an acceptable level of risk for the appropriate target population, and recipients will need to be thoroughly educated about possible outcomes.
The early embryonic origin of ES cells, the pluripotential inner cell mass of the blastocyst (Fig. 1A), gives these cells the two hallmark characteristics of stem cells, self-renewal and pluripotency. The ability of ES cells to proliferate readily, however, holds up in vivo, where they can form teratocarcinoma-like tumors in adult mice if injected subcutaneously, intramuscularly, or into the testis [6, 10, 11]. These tumors contain differentiated cells that derive from all three primary germ layers, as well as undifferentiated pluripotent stem cells. This tendency to form tumors has also been observed when ES-derived cells are transplanted into normal or lesioned mice [12, –14], raising the very real concern that hES cell-based therapy may also lead to unwanted tumor formation.
Figure Figure 1.. Making embryonic stem cell lines. The standard approach to making an ES cell line involves isolation of the inner cell mass of a blastocyst-stage embryo (A), which is then placed in culture (B). A pluripotential cell line can also be derived from a single blastomere isolated from an eight-cell-stage embryo by coculturing it with a pre-existing ES cell line and then isolating the blastomere-derived cells after they have adapted to growth in culture. The residual embryo can be implanted (B). ES cells can also be isolated from a blastocyst generated using somatic cell nuclear transfer (SCNT). Here a nucleus from a somatic cell is transferred to an enucleated donor egg, then grown to the blastocyst stage in vitro and the inner cell mass isolated (C). Altered nuclear transfer uses a nucleus that carries a mutation in a gene required for normal development. The example shown uses the cdx2 gene, which is required for implantation. Although ES cells can be isolated from these embryos, the embryo is unable to implant (D). Cell hybrids with the properties of ES cells can be isolated by fusing a somatic cell with an ES cell (E). Abbreviation: ES, embryonic stem.
Download figure to PowerPoint
Studies using mouse ES cells have suggested a number of means to inhibit tumorigenicity following transplantation. If, instead of undifferentiated ES cells, a more differentiated derivative is used for grafting , the cells are less likely to generate a tumor. For example, ES cell-derived neural stem cells can be used for transplantation to treat neurodegenerative diseases [12, 14, 16]. More differentiated cell types can be isolated from a mixed cell population using either sorting techniques, such as fluorescence-activated cell sorting , or selective approaches . However, ES-derived tissue-specific cells, such as neural stem cells, can also form tumors (neural tumors in the case of ES-derived neural stem cells) following transplantation to the adult mouse . A more terminally differentiated cell, no longer capable of proliferation, such as dopaminergic neurons [15, 19], can also be generated and used for transplant. The limitation of this approach is that it does not provide a pool of stem cells with the capacity to continually produce newly differentiating cells if the transplanted cells die, perhaps due to an existing pathological condition. Additional steps designed to eliminate contamination of transplant material with residual undifferentiated ES cells will likely be essential, and two different approaches have been reported. One uses a negative selection step based upon the ES cell-specific expression, in an engineered cell line, of a compound that is toxic to undifferentiated ES cells under certain culture conditions. This approach has been used successfully to eliminate mouse ES cells from a mixed cell population prior to transplant . A variation on this theme involves engineering a drug-inducible suicide gene that could be expressed in the undifferentiated stem cells following transplantation . The other approach treats the mixed cell population with the ceramide analogue N-oleoyl serinol (S18) to selectively induce apoptosis of ES cells . This treatment prevents subsequent teratocarcinoma formation following transplantation of mixed populations containing both ES cells and ES-derived neural stem cells .
We clearly have in hand today the technology needed to generate a population for transplant that is essentially free of contaminating undifferentiated stem cells. But how certain must we be that we have eliminated any chance that these potential tumor-forming ES cells remain? Most therapeutic treatments, particularly those used to treat terminal conditions, carry some risk. Clinical trials with experimental therapies generally carry higher risks. What constitutes an acceptable risk will vary depending on the condition to be treated and the anticipated benefit of the therapeutic intervention. Before clinical trials for hES therapies begin, efforts must be made to minimize contamination; information about the limitations, success, and failure of such efforts must be communicated; and informed consent from patients must clearly be obtained. Communicating realistic expectations for success is a complicated task, particularly in an area in which patient expectations are high.
There is a particularly strong need to manage public expectations for hES cell therapies and to understand the role that particular patient's medical vulnerabilities may play in obtaining true informed consent. Public understandings of the power and limitations of potential hES cell therapies can be clarified by increased public education campaigns that address both the science and the ethics of stem cell research and therapies. Addressing patient expectations may be more difficult. Although it is relatively easy to understand how financial inducements or familial pressures can constitute coercion or undermine autonomous choice, it is much less clear how desperation, pain, and hope might impact a patient's ability to comprehend risk and confound consent. Although therapeutic treatments are being developed in laboratories, work on clarifying key features of informed consent in the context of hESC therapies should be occurring simultaneously. Important insights have been gained in the area of genetic therapy, where “high hopes” led to “harsh lessons” [23, –25].
2. Contaminating Animal Products
Any cell-based therapeutic agent used in humans must be free of animal contaminants that may contain pathogens or elicit an immune reaction after transfer to a host. All hES cell lines on the NIH registry were derived in the presence of animal products. This limitation has contributed to the development of various state policies that allow for funds to be used to generate new cell lines free of contamination. It has also led to the development of approaches to removing animal products from existing lines. Both the local development of new contamination-free lines and the efforts to eliminate contamination in existing lines raise questions about negative downstream consequences of the federal restriction on funding.
Conventionally, mouse and human cell lines are grown in medium that contains animal serum, a source of nutrients and growth factors. In addition, ES cell lines, mouse and human, are generally grown on mouse-derived feeders, a layer of fibroblasts treated so they no longer proliferate, that provides additional factors that both promote ES cell proliferation and inhibit their differentiation. All of the hES cell lines on the NIH registry approved for use by government-sponsored laboratories were isolated under conditions that included animal serum and mouse feeders. The presence of animal products presents a barrier to the use of these particular cell lines for human therapies, since these products may contain animal pathogens and elicit an immune response.
One example of possible animal product contamination is the demonstration that hES cells grown in the presence of animal products express a nonhuman sialic acid, Neu5Gc . Humans have circulating antibodies directed against this epitope, and its presence could elicit an immune response, resulting in transplant rejection. Culture in human serum reduced the levels of Neu5Gc on the hES cells, suggesting a means of reducing the threat of animal product contamination . Another concern is the transfer of murine viruses from mouse feeder layers to the hES cells, although recent analysis suggests that a number of lines cultured on mouse feeders are not infected with murine leukemia virus .
Based in part on concern about contaminating animal products, a number of laboratories have been working on adapting hES cell lines, to grow under serum-free defined medium conditions on human cell-derived feeders or under feeder-free conditions, and some have successfully cultured hES cells under serum-free conditions [28, 29]. Recent reports demonstrate the ability of FGF-2 to replace the need for feeder coculture [30, 31]. Another recent study documents the derivation of hES cell lines under animal product-free conditions . This progress suggests that the problem of animal product contamination is being solved. However, the route to eliminating this roadblock may have been more direct and faster if federal funding had been available for generating uncontaminated cell lines. We will discuss some of the ethical issues associated with funding in section 4.
3. Genetic Compatibility
hES-cell transplants will be subject to immune rejection. The generation of hES cell lines genetically identical to the patient using therapeutic cloning appeared to be a viable option until recent events described below revealed no evidence to support the existence of any hES cell lines derived from cloned embryos. For scientific and ethical reasons, hES cell banks may be a feasible alternative.
The problem of tissue compatibility to avoid immune system based rejection applies to all cell-based therapies, including organ transplants. The stem cell therapy approach that has received the most attention for solving this problem is therapeutic or research cloning; the use of somatic cell nuclear transfer (SCNT). SCNT involves transferring the nucleus from an adult somatic cell, such as a skin cell, to a donated enucleated human egg; the resulting zygote is grown in vitro to the blastocyst stage; and an hES cell line isolated that is genetically identical to the nuclear donor or patient (Fig. 1C). (Additional approaches to generating genetically compatible stem cells are discussed in section 6.) Despite initial reports claiming the isolation of hES cell lines from cloned embryos, SCNT has not yet led to the production hES cell lines [33, 34]. The claim that cell lines are genetically identical to a nuclear donor must be substantiated by comparing the DNA fingerprint of the nuclear donor to the fingerprint of the established stem cell line. Although the initial data presented by South Korean group led by Hwang appeared to support the conclusion that the stem cell lines were derived from SCNT embryos, follow-up studies demonstrated that the lines were in fact derived from noncloned human embryos .
Progress in the area of human SCNT has thus been slowed due in part to these fraudulent claims of success in South Korea. There are at least three important lessons to be learned from this scandalous situation. First, enthusiasm for new biotechnological solutions to devastating human problems must be tempered by continuing commitments to the canon of scientific investigation—new advances must be replicated in other laboratories by other investigators. It would be unethical, or at least unwise, to assume that a technique or procedure had been perfected on the basis of uncorroborated work. Second, peer review in a climate of excitement must maintain its rigor, despite pressures that all media feel to varying degrees “to get the story out.” Third, scientists in areas of “sensitive” research should learn how to think about the ethical issues involved. A simple assertion that a procedure or use of oocytes or blastocysts is “ethical” or “unethical” is not a substitute for ethical reasoning. Such assertions are, one hopes, the conclusion of an ethical argument and understanding the argument is what is increasingly important for all those involved in hES cell research and therapies.
Even if scientifically possible, it is not clear that using SCNT to create patient specific hES cell therapies for every patient is practically feasible or ethically justifiable under conditions of scarcity. The cost and labor involved will likely be excessive, and this in turn raises questions about the just distribution of resources as well as questions of equal access. It is more probable that, as for the bone marrow transplant field, banks of hES cell lines reflecting the major histocompatibility groups will be generated and used in combination with immunosuppressive therapy. A recent feasibility analysis using human leukocyte antigen (HLA) and blood group data from 10,000 consecutive cadaveric organ donors as a stand-in for hES cell lines and kidney transplant recipients as a stand-in for the stem cell therapy recipients suggested that a bank of 150 hES cell lines would provide a full match for class I major histocompatibility proteins HLA-A and HLA-B as well as class II major histocompatibility protein HLA-DR for <20% of recipients and a beneficial match (only one HLA-A or HLA-B mismatch) for 25–50% of recipients . The large number of additional cell lines required to achieve a match for a significant additional fraction of the remaining patient population may make SCNT an attractive alternative. Efficiency cannot justify inequities in the availability of potential therapies for all who might benefit from them. As is the case with organ banks, particular attention will need to be paid to seeing that HLA types common to groups with particular ethnic ancestry are not under-represented in stem cell banks.
4. Funding Issues
The guidelines of the administration of President George W. Bush, which limit the use of federal funds to work with the approximately 20 lines isolated prior to August 2001, has restricted hES cell research in the United States [37, 38]. These approved lines have limited potential, both in the laboratory and as clinical reagents, and the small number of lines available have limited research to improve the efficiency and ease of hES cell culture. In addition these lines were isolated in the presence of animal products and do not reflect genetic backgrounds necessary to avoid rejection following transplantation . An investigator seeking private or state funds to work on the hES cell lines generated subsequent to August 2001 must take care to separate these studies from other federally funded projects. This may entail extremely careful bookkeeping or an entirely separate infrastructure that does not rely on any federal money. These restrictions and the prospects of nonfederal funding for hES cell work raise a number of ethical and political issues. A thorough assessment of all of them is beyond the scope of the current discussion, but a number of ethical issues should nonetheless be noted.
To date, four states have committed to funding stem cell initiatives (California, Connecticut, New Jersey, and Maryland), and a number of other states are pursuing various types of stem cell initiatives . Although the promise of much-anticipated therapies from state-funded research is high, recent events in California and Connecticut raise worries that must be addressed early if confidence in stem cell research is to be preserved. California, having committed approximately $295 million a year, and Connecticut, having committed approximately $10 million a year, have provided opportunities for new scientists to enter the field but may also have inadvertently created hES cell “bandwagons” for biomedical entrepreneurs and scientists who may not have stem cell experience and expertise. This puts greater burdens on review committees in an already highly charged political climate not only to develop thoughtful methods of ethical oversight for research but also to identify financial and other conflicts of interest. The scientific wings of review committees will also have unusual burdens. More than other areas of research, the promise and hope associated with hES cell research creates pressures that shine a spotlight on the objectivity of review. There is limited oversight guidance here; the guidelines recommended by the National Academy of Science provide a useful starting point, but much more work needs to be done at both the institutional and state level to support the highest standards of oversight .
Additional concerns have been raised at the state and federal level about funding priorities. Taxpayers rightly ask whether these expensive stem cell initiatives should take priority over other public health issues. Consider spinal cord injuries, one of the many types of disease and injury that hES cell therapies hope to remedy. Every year, approximately 7,800 people in the U.S. have their lives completely and permanently altered after a spinal cord injury . Reversing or minimizing the damage of such injuries, getting people off of ventilators, out of wheelchairs, and back to fully independent and productive members of their community is an exciting prospect for those who suffer from such injuries and for their family and friends. In 1993, an estimated six million U.S. preschool children and 400,000 fetuses had blood lead concentrations at dangerous levels, levels that cause learning disabilities, brain damage, and other ailments. Although efforts to minimize exposure have been highly successful, in 2002 the numbers were still high, particularly for black and poor children, who are especially vulnerable to having their lives permanently altered by high-level exposure . These are just two examples of the difficulty public health administrators face in terms of prioritizing funding. The high profile of hES cell research illustrates the need for public officials to be able to carefully identify the principles that establish and justify funding priorities.
There are also ethical considerations raised by policies governing payment for oocytes and blastocysts used in hES cell research and eventual therapies. Virtually all policies and guidelines, including international policies, prohibit payment in exchange for research acquisition of oocytes and blastocysts. Restricting payment for blastocysts has a clear, albeit not definitive, ethical justification. There are important ethical concerns about the welfare of children under conditions in which it is permissible to put embryos and/or children on the market. Some have argued that the relationship between parents and children will be undermined and diminished if the relationship were akin to other market relationships. Furthermore, the dignity and autonomy of human beings is thought to be degraded if human beings are commodified, that is, put on the market for exchange. The buying and selling of blastocysts is similar enough to the buying and selling of children that a reasonable argument could be made to err on the side of caution in the context of payment for blastocysts for hES cell research . Paying women for the use of oocytes in hES cell research and therapy is a different matter, however.
In the U.S., payment for participation in research is widely practiced. In addition, many college newspapers run advertisements offering sums of money for egg “donation” to infertile couples. Most sperm banks pay for donations. The procedures necessary for oocyte retrieval are invasive, time-consuming, inconvenient, disruptive, and carry some small risks. Why then is payment for oocytes for hES cell research generally prohibited? The justification often given is that if there were financial gain associated with oocyte donation, women, particularly poorer women, would be induced or unduly influenced to donate. There are at least two ethical problems with this justification. First, to assume that women, or poor women, are less capable of making decisions about how they earn money than other people is a faulty assumption that also appears prejudicial and potentially condescending. There are many activities and professions that carry great risks, often with very modest financial gain, such as mining. People should be fully informed about the risks they willingly assume through their choices, and the highest standards of safety should be followed to minimize risks, but prohibiting remuneration for the sake of the participant is not justifiable unless it is consistently and equally applied. This raises a second problem, namely, that there are a variety of inconsistencies generated by policies that prohibit payment for oocytes. Investigators pay considerable amounts of money for various pieces of equipment, for research assistance, for laboratory space—research costs money. It would be very odd if one of the most valuable aspects of hES cell research, oocytes, were obtained virtually free. Currently, the National Academy of Science recommendations suggest that no cash or in-kind payment be provided for donating oocytes for hES cell research, and they attempt to avoid one inconsistency by also suggesting that no payments should be made for sperm donation for research purposes either. Of course, oocyte donation is much more complex than sperm donation, and it is likely that many more oocytes will be needed in the development of hES cell therapies. The NAS recommends that this policy be reviewed and reconsidered. Institutions and states creating hES cell policies should be mindful of the controversy and seek to develop more consistent principles to govern payment for oocytes [45, –47].
5. Selecting and Generating the Right Cell Type for Transplantation
What is the ideal stage in the progression from undifferentiated ES cell to restricted progenitor to differentiated cell at which to transplant for therapeutic purposes? Pros and cons for specific cell types are discussed, although the solution will likely be disease-specific.
Generating the right cell type for transplantation would seem to be a simple matter: replace the dead, dying, or damaged cells in-kind with ES-derived cells. In general, this means transplanting terminally differentiated cells that have lost all capacity for proliferation. Conditions have been established for mouse and hES cells that promote their differentiation into many adult differentiated cell types, including specific subtypes of neurons [15, 19, 48, 49], glia [50, 51], cardiac muscle [52, 53], vascular cells [54, 55], and blood cells [3, 56]. It has proven more difficult to generate endoderm derivatives, but the recent generation of definitive endoderm from both mouse and hES cells suggests that this goal will be accomplished as well [4, 57]. These protocols generally include a sequence of steps that expose the cells to varying cell-cell interaction conditions and growth factors.
It has not been possible, however, to generate every cell type from mouse ES cells, and in fact, we still do not understand the lineage of all cell types in vivo. For example, it has recently been observed that pancreatic β-cells in adults do not differentiate from a stem cell precursor but rather from existing pancreatic β-cells .
The two best examples of successful application of ES cells using rodent models of human disease involve transplant of differentiated derivatives. 6-Hydroxydopamine treatment of animals results in the loss of midbrain dopaminergic neurons, creating a model for Parkinson disease [15, 19, 59]. These mice manifest a behavioral defect, amphetamine-stimulated rotations. Transplantation of mouse or hES-derived dopaminergic neurons can successfully populate the striatum and differentiate into tyrosine hydroxylase producing neurons with appropriate electrophysiological properties. Most importantly, these cells provide some rescue from the behavioral defect. In the second example, ES cell-derived oligodendrocytes can myelinate the CNS of myelin-deficient rats [50, 60]. hES-derived oligodendrocytes have even been demonstrated to myelinate and repair a lesion following spinal cord injury in adult rats, restoring some locomotion .
A single dose of terminally differentiated cells may work for treating conditions or diseases in which the host environment is favorable (for example, following spinal cord injury) and would also avoid the tumor concern that transplanting a proliferative population presents. If, however, the patient is suffering from a degenerative condition that is destroying his or her own cells, the transplanted ES-derived cells may suffer the same fate. Therefore, transplanting a population that includes a proliferating progenitor would provide a renewable source of differentiating cells, provided that in vivo conditions can support the transition from stem to differentiated cell. Progenitors may also be desirable material when the targeted cell type is short-lived. These transplants, however, may run the risk of tumor formation.