I'll post this because I have access:

PNAS | February 19, 2002 | vol. 99 | no. 4 | 1755-1757

Commentary
Will embryonic stem cells be a useful source of dopamine neurons for transplant into patients with Parkinson's disease?
Curt R. Freed*

Division of Clinical Pharmacology and Toxicology, University of Colorado School of Medicine, C237, 4200 East Ninth Avenue, Denver, CO 80220

Human embryonic stem (ES) cells entered the popular lexicon during the political debate of summer 2001 and the decision by President George Bush to allow federal funding for research on the 64 human ES cell lines that had been derived by the day of his speech. Before the policy decision, the National Institutes of Health did a comprehensive review of the field (http://www.nih.gov/news/stemcell/scireport.htm). Human ES cells come from preimplantation embryos, most of which are generated at in vitro fertilization clinics (1). Within days after fertilization, the embryo consists of a hollow sphere, the blastocyst, which contains a cluster of a few hundred identical cells called the inner cell mass that can eventually develop into a fetus. When removed from the blastocyst, these cells can be propagated indefinitely in specialized media (2). When the media are changed to allow differentiation, cells continue to divide and aggregate, forming embroid bodies. Although these cell clusters lack the organization of an embryo, they contain all tissue types including skin, muscle, bone, and neurons. Because ES cells can become any cell in the body, there is hope that they will lead to treatment for diseases like diabetes, Parkinson's, Alzheimer's, and heart failure. The problem is controlling cell growth and differentiation. If large numbers of ES cells are transplanted into an organ like the brain, they grow into every cell type and form tumor-like masses called teratomas, eventually killing their host. How can ES cells be restricted to produce useful cells without overgrowing? In their article in this issue of PNAS, Björklund et al. (3) have transplanted small numbers of partially differentiated mouse ES cells from embroid bodies into a rat model of Parkinson's disease and have shown that at least some of the cells become the dopamine neurons that are needed to reverse the Parkinson condition. The authors suggest that the brain environment may encourage survival of cells appropriate for treating the disease while controlling the tendency to form a tumor mass. The results were far from perfect. Of 25 rats receiving transplants of 1,000-2,000 cells, 56% of animals showed surviving grafts containing dopamine neurons whereas 20% had lethal teratomas and 24% had no cells survive. Although these proportions are not promising as a treatment for humans with Parkinson's disease, the results illustrate the principle that relatively undifferentiated cells can develop into neurons appropriate for a specific brain region without invariably leading to uncontrolled cell growth.


How can embryonic stem cells be restricted to produce useful cells without overgrowing?

The hope is that research on human ES cells may reveal methods for producing an infinite supply of dopamine neurons for transplant into patients. Parkinson's disease is caused by the death of a small number of dopamine neurons located in a nucleus in the midbrain, the substantia nigra (4). Axons project to the forebrain and supply dopamine to the putamen and caudate nucleus. Without dopamine, patients are frozen in place. In one of the most remarkable drug developments of the 20th century, Cotzias and colleagues (5) demonstrated in 1967 that the amino acid L-dopa can be taken by mouth, enter the brain, be converted to dopamine, and improve Parkinson's symptoms. By 1979, a new treatment strategy, replacement of the damaged dopamine neurons by cell transplantation, proved successful in a rat model of Parkinson's (6-8). Dopamine cells survived, grew axons and dendrites, and improved behavior of the rats. Experiments showed that only fetal dopamine cells from the midbrain at a specific developmental stage could survive transplantation---13-15 days after fertilization in the rat and 6-8 weeks after fertilization in the human. At this stage, both rat and human fetus are only 2-3 cm in overall length.

By the late 1980s, fetal dopamine cell transplantation was being performed in humans with Parkinson's (9-17). All of the principles developed in the rat have been validated in human subjects, and we have found that neurons survive for at least 8 years after transplant (Fig. 1). Neurotransplantation with fetal dopamine neurons is now a proven strategy for treatment of patients with advanced Parkinson's disease. In our double-blind, placebo-controlled surgical trial of fetal tissue implants, we found that transplants survived in 85% of patients regardless of age and without immunosuppression and improved signs of Parkinson's disease in patients under age 60 and in older patients who still had a good response to the drug L-dopa (17). Sham surgery patients had no change in their symptoms. About 15% of patients who reduced or discontinued all L-dopa had the same kind of excess movements (dyskinesias) that had been caused by their L-dopa treatment. This study has proven that neurotransplants have significant benefits and also have some of the same side effects as the drug L-dopa. Transplant methods will continue to evolve, regardless of the source of dopamine neurons used for transplant.



Fig. 1. Human fetal dopamine neuron in putamen of Parkinson's patient dying 8 years after cell transplantation. The cell is immunostained for tyrosine hydroxylase and has multiple processes growing out from the cell body to specifically reinnervate the surrounding brain region. Other staining methods used on other transplanted cells in the same patient revealed the neuromelanin pigment particles typical of mature substantia nigra dopamine neurons. Immunosuppression was used for the first year after transplant, but not for the subsequent 7 years. If human ES cells can be differentiated to a phenotype identical to fetal substantia nigra dopamine neurons, those cells should have the same stable survival seen in this patient's transplant. (Magnification: ×400.)

Two problems limit the wide applicability of fetal dopamine cell transplants. First is the difficulty in recovering fragments of fetal brain tissue from elective abortions. The second problem is poor survival of dopamine neurons. About 90-95% of transplanted neurons fail to survive with most dying by apoptotic programmed cell death (15, 17, 18). An unlimited source of dopamine cells produced in tissue culture would solve both of these problems. Because fetal neurons usually are not rejected when transplanted into an adult host brain from the same species, it is likely that cells from a small number of individuals representing different HLA types would provide tissue immunologically acceptable for most everyone. Apoptotic losses could be managed by simply increasing the number of cells transplanted. If survival of laboratory-generated cells was predictably 20%, then transplants of a 5-fold excess would deal with the cell death problem.

Dopamine neurons generated in tissue culture should be identical to dopamine neurons native to the substantia nigra. They must synthesize, release, take up, and catabolize dopamine. Because many in vitro manipulations can induce temporary expression of the dopamine synthetic enzyme tyrosine hydroxylase, candidate cells should be evaluated with a panel of markers such as the dopamine transporter, ptx-3, nurr-I, and aldehyde dehydrogenase. They also should have the ability to project axons to target nuclei like the putamen. While the molecular mechanisms for target recognition by dopamine neurons are unknown and because in vitro conditions do not mimic conditions in the brain, cells must be transplanted into the dopamine-depleted striatum of experimental animals to see whether phenotype is preserved and whether neurites can connect with appropriate target cells. Björklund et al. (3) met these criteria, at least in the 56% of animals that showed surviving dopamine neurons without teratomas.

Will ES cells provide an infinite and safe supply of dopamine neurons? It is too early to say. To reduce the risk of teratomas, cell differentiation should take place in vitro and not after transplant into brain. At the National Institutes of Health, Lee et al. (19) have used a progressive expansion, selection, and differentiation strategy to convert mouse ES cells to a mixed population of mature neurons in tissue culture with up to 30% having the characteristics of dopamine cells. Lee et al., like Björklund et al., started by predifferentiating ES cells to embroid bodies. To increase the fraction of cells that were neural, they continued the differentiation process in neuron-selective media. The combination of expansion and selection of ES-derived cells will likely be fundamental to the safe production of selected phenotypes.

Using a somewhat different approach, Kawasaki et al. (20) have been able to generate dopamine neurons from mouse ES cells without embroid body formation by exposing the cells to factors generated by PA6 stromal cells. ES cells grown on a layer of either live or paraformaldehyde-fixed PA6 cells led to a mixed population of differentiated cells of which 16% were tyrosine hydroxylase-positive. The cell cultures produced dopamine. Importantly, the tyrosine hydroxylase phenotype was preserved after transplant into rats. To prevent tumor formation from residual ES cells, Kawasaki et al. treated their cell preparation with mitotic inhibitors. Although the differentiation factor or factors secreted by PA6 cells are unknown, the neural fate could be blocked by treatment with BMP4, an antineuralizing agent. The reports from Lee et al. and Kawasaki et al. indicate that a number of strategies will likely be successful for converting ES cells to dopamine neurons.

What about producing dopamine neurons from stem cells found in skin, blood, bone marrow, or umbilical cord blood? Recent evidence suggests that some stem cells from each of these sources have the potential to become neurons (21-23). Producing specific cell types from a patient's own stem cells would have immunologic advantages. It remains to be seen whether specific neuronal phenotypes can be generated efficiently and economically.

Immortalization of cells during normal development offers another route for producing useful dopaminergic neurons. Using the simian virus 40 large T antigen, we have immortalized fetal rat dopamine neurons. These cells express phenotypic markers of dopamine neurons and have behavioral effects after transplant into the rat model of Parkinson's (24). While the cells divide continuously in vitro, factors in brain inhibit tumor formation after transplant (25). Cells with a dopaminergic phenotype have been produced from a v-myc immortalized mouse stem cell line in which the nurr-I orphan receptor gene has been expressed and after cells have been exposed to astrocytes derived from ventral mesencephalon (26).

How will human ES cells fit into cell replacement strategies? Diseases treatable with ES cells will likely fall into two categories. First is replacement of specific cell types, like dopamine neurons for Parkinson's disease. There is no evidence that dopamine cells can arise spontaneously from a stem cell population in the adult brain. Thus, creation of dopamine neurons from ES cells would offer a therapeutic strategy not possible even with normally functioning intrinsic stem cells. The second category of treatment is replacement of failed host stem cells. A common example is bone marrow transplantation for aplastic anemia. Novel stem cell populations are still being revealed. Cardiac myocytes, previously thought to be fixed in number in the adult, have been shown recently to be replaceable by host tissue in the setting of cardiac transplantation, indicating that myocardial stem cells exist (27). Progenitor cells capable of generating neurons have been detected in adult human hippocampus as well as in the subventricular zone (28, 29). We have shown that human neural stem cells can be integrated into the germinal zone of the developing monkey brain in utero and can join in region-appropriate neurogenesis (30). If diseases like Alzheimer's prove to be the result of neural stem cell failure in some patients, transplants of neural stem cells may be therapeutic. For human ES cells to be useful for stem cell replacement, they should be partially differentiated into lineage-restricted stem cells to eliminate the pluripotency responsible for teratomas. This goal is being achieved (31-34).

In summary, there are a number of strategies for producing dopamine neurons in tissue culture that may lead to cells appropriate for transplantation into humans with Parkinson's disease. Björklund et al., Lee et al., and Kawasaki et al. have shown that generation of dopamine neurons from mouse ES cells is possible. Hopefully, a similar outcome will be demonstrated for human ES cells. Alternatively, stem cells derived from brain, blood, bone marrow, or skin may be convertible to dopamine neurons. Immortalized dopamine neurons may be useful if cell replication after transplant can be controlled. The risk of tumor formation is present in all ES and immortalized cell sources. Eliminating the possibility of uncontrolled proliferation is fundamental to the development of this field. The winners in this lottery are unknown, but the characteristics of the successful cell preparation can be specified. The cells should be identical to authentic fetal human midbrain dopamine neurons capable of reinnervating the dopamine-denervated striatum. Must the cell preparation be uniformly dopaminergic? Probably not. Clinically effective transplants of human fetal mesencephalon contain many cell types, with only a small fraction being dopamine cells. Whether transplants require the presence of cell types other than dopamine neurons is unknown. The successful generation of an unlimited supply of dopamine neurons will make neurotransplantation widely available for patients with Parkinson's disease. Embryonic stem cells are opening an exciting era in human therapeutics.