Here we focus on two therapeutic protocols that are in clinical trials -- each involves grafting of cells from other species (xenogeneic) into the human brain: fetal pig neurons to replace dopaminergic neurons lost in Parkinson's disease; and immortalized mouse fibroblasts producing retrovirus vectors for delivery of a suicide gene to brain tumors. These two procedures are very different. In the first, normal cells are obtained from animals specifically bred and monitored for this purpose, with the aim of long-term survival and integration into the brain cytoarchitecture (Fig. 1). Ultimately this can be seen as potentially more secure than using human fetal cells, as the latter cannot be obtained under conditions optimized for cell survival or monitored for pathogens (except through pre-abortion screening of maternal blood). The major caveat to the use of pig cells is that intrinsic pathogens are not yet fully characterized, as made evident in the recent report that pig cells produce C-type retrovirus particles that can infect human cells¹. Moreover, it is not yet clear how effective this procedure will be in the treatment of ongoing neurodegenerative conditions, such as Parkinson's disease and Huntington's disease.

Fig. 1 Treatment of Parkinson's disease by transplantation of pig neurons. a, Cells are obtained from the mesencephalic brain region of fetal pigs under optimized conditions for cell survival. This region contains dopaminergic neurons, as well as other neural and glial cells. A total of 1.2 x 107 cells in 240 µ l (from 15-20 pig fetuses) is injected into the striatum on one side of the Parkinson patient's brain, with the patient conscious and using local anesthesia. b, Parkinson patients suffer from a loss of dopaminergic neurons in the substantia nigra, which sends processes into the striatum. This results in insufficient release of dopamine (DA) in the striatum onto neurons projecting to the globus pallidus or the substantia nigra. The result is loss of movement control, rigidity and tremor. c, Transplantation of pig neurons into the striatum potentially restores dopaminergic synaptic inputs onto neurons projecting out of the striatum and allows resumption of normal movements.

In contrast, in the tumor therapy paradigm, the immortalized mouse fibroblast cell lines being used have been passaged extensively in culture. Although well characterized for pathogens, these cells contain high levels of expressed retrovirus sequences, and have been specifically engineered to express components of the virion that increase infectivity for human cells. Retrovirus vectors are deemed valuable in the treatment of brain tumors as they only deliver genes to dividing tumor cells and not to post-mitotic neural cells (Fig. 2). The short half life and low titers of traditional retrovirus vectors mandate implantation of packaging cells that release vectors over time^2,3, but even with this protocol, gene delivery to tumor cells in the human brain is extremely low⁴. Thus it is not clear how effective this approach to treating brain tumors will be.

Fig. 2 Treatment of brain tumors by transplantation of mouse retrovirus producer cells releasing vector bearing the TK gene followed by drug administration. In a typical paradigm, the bulk of the tumor is removed by surgery and mouse producer cells are injected into the resection cavity. Typically 105 cells are repeatedly injected. A few days later (to allow time for gene delivery to tumor cells) patients receive a two week course of intravenous ganciclovir (red). a, Mouse producer cells contain both vector (X) and packaging (gag/pol and env) retrovirus sequences; express viral antigens (), and HSV-TK (TK); and release vector particles () bearing the HSV-TK gene. Non-dividing neural cells (white) and non-dividing tumor cells (blue) do not integrate the HSV-TK gene and hence do not express TK. Dividing cells, primarily tumor cells (blue), incorporate the transgene and express TK. b, Ganciclovir (GCV) treatment leads to death of dividing tumor cells expressing TK because phosphorylated GCV (GCV-P) acts as a "false" nucleotide and disrupts DNA replication. GCV-P can be transmitted between cells in close contact, through gap junctions, but cannot be passed from one cell to the other through the extracellular space. Antibodies generated against virus and murine antigens lead to destruction of producer cells.

Why cell transplantation for treating degenerative brain disease?

Parkinson's disease is a movement disorder characterized by akinesia, bradykinesia, rigidity and tremor. Approximately 1% of the population above age 65 has Parkinson's disease¹⁰. Symptoms result from the selective loss of midbrain dopaminergic neurons and their synapses, which release dopamine into the caudate and putamen regions of the brain¹¹. Physiological compensatory mechanisms in the brain can suppress development of symptoms until approximately 60-80% of dopaminergic neurons are lost^11,12. The discovery that administration of L-dopa (the chemical precursor of dopamine) reduced the signs of Parkinson's disease provided one of the first rational pharmacological therapies for neurodegenerative disease¹³. However, L-dopa usually only provides temporary relief, and after months or a few years of continued use the patients experience severe drug and dose fluctuations, including the so-called "on-off" phenomenon^14,15. Alternative therapies for patients are neurosurgical treatments, including surgical lesions to the brain, the so-called thalamotomies¹⁶, pallidotomies¹⁷ and high frequency electrical stimulation (to the basal ganglia circuitry)¹⁸. The rationale for the new emerging transplantation treatment for Parkinson's disease is the replacement of dopamine neurons that have died. Neuronal cell transplantation to humans is still in the early phases of technological development, and procedures need to be optimized with respect to surgical protocols, cell sources, cell types, cell preparation, cell delivery, auxillary procedures (such as co-introduction of growth factors) and modulation of immunological parameters.

Although there is controversy regarding the optimal source of cells, it is clear that dopaminergic neurons have, as a biological phenotype, the capacity to regulate dopamine release within nanomolar concentrations in the host striatum. The importance of this is apparent when one considers that excessive or unregulated dopamine release is thought to underlie other neurologic syndromes, such as dyskinesias, dystonias and possibly schizophrenia^19-21. Adult neurons are not a feasible source as they do not survive isolation (or transplantation) because of shearing of long differentiated processes and metabolic failure resulting in cell death (usually within minutes of isolation). Other cell types, such as skin fibroblasts from the same patient, genetically engineered to release dopamine²², do so in a diffuse manner creating a concentration gradient, high at the point of implantation and lower with distance away from it. The brain responds by down-regulating dopamine receptors in cells close to such an implant, while upregulating receptors in more distant cells⁵. Transplants of fetal dopaminergic neurons, in contrast, send out neuronal processes and form synaptic connections with endogenous neurons over a wide range, thereby achieving regulated release and responses that are appropriately modulated^{5, 24}. It is hard to imagine how non-neuronal cells could function better than neuronal cells in this transplantation paradigm. In the future, propagation and manipulation of neural progenitor cells to generate a sufficient source of neurons of the appropriate phenotype might be a potential alternative to fetal neurons. Still, even in these early phases, transplantation of human fetal cells has provided promising results, including in some cases marked symptomatic recovery of Parkinson patients paralleled by restitution of dopamine synapses in the host brain and reduced need for, or more effective action of, L-dopa and other therapies^6,7,24. Problems of transplanting human fetal cells to Parkinson patients include the ethical and practical concerns of using fetal tissue; the inability to obtain a sufficiently large number of appropriate human fetal neurons, and difficulties in rapid and extensive screening for pathogens^6-8,24. For these reasons, other cell sources are being sought.

Xenogeneic neuronal transplant survival and host responses

The major risk to a recipient of a neural xenogeneic cell preparation appears to be loss or rejection of the cells²⁵. Even in fully immune-suppressed rodent hosts, where grafts to the brain usually survive for an extended period, at least 90-95% of the total number of implanted fetal neurons (in both allo- and xenogeneic donor tissue) die. This is probably due in part, to intrinsic developmental signals and damage incurred during cell preparation, inadequate trophic support and oxidative stress^31-33. This primarily non-immunological neuronal cell death can be overcome to some extent by increasing the initial cell dose or protecting the cells with trophic factors and antioxidants. In addition, cell-mediated or antibody-mediated rejection can kill transplanted cells and potentially cause a harmful inflammatory response in the host, in so far as it could damage neurons and cause cerebral edema. The CNS inflammatory response, even during xenogeneic cell rejection in immune competent animals, appears restricted, however, and the resulting death of transplanted cells usually leaves very minor damage to adjacent host tissues³¹. The relatively low immunological reaction in the brain has been attributed to a low level of expression of MHC-I and II on neural tissue^25,34. Neurons normally express almost no MHC-I (major histocompatibility complex class 1), although both glial cells and neurons can be induced to express immunological markers under conditions of inflammation, infection or inactivity^34,35. Moreover, the blood-brain barrier, which consists of an endothelial layer with tight junctions, a thick basal membrane and astrocytic end feet, filters blood entering the brain and may to some extent restrict entry of complement and immune cells. In fetal neural cell xenotransplantation into rodent host brains (in the absence of preformed anti-species specific antibodies), cyclosporine and other immune suppressive regimes, which are regularly used in humans for allogeneic kidney and heart transplantation, are usually sufficient to prevent massive rejection^25,36. During rejection, upregulation of MHC class I and class II antigens is seen, both in donor and host tissue^31,36. This can further trigger immune responses from class I and II stimulated T cells and, by clonal expansion, fully activate the efferent arm of the immune system, unless immunosuppressive regimes are applied^25,37. Xenogeneic donor tissue outside the brain is usually vigorously rejected, even in the presence of immune suppression^38,39. This complement-mediated, hyperacute rejection depends on complement and preformed serum antibodies against xenogeneic antigenic determinants, such as galactose alpha (1,3) galactose-containing epitopes on the donor tissue^38,39. Although this hyperacute rejection is less vigorous in the brain than in other transplantation sites, little is known about its full effect on pig cells in primate and human host brain^9,38-40.

Clinical trials using fetal neural pig cells

Mouse cell therapy for brain tumors

For gene therapy protocols aimed at delivering a therapeutic gene to tumor cells on site, one of the most critical parameters is the efficiency of gene delivery. This includes both the number of tumor cells transduced per given dose and the distribution of these cells in the brain. Currently there is little information available about the efficacy of gene delivery to tumor cells in the human brain. In this therapeutic protocol, immortalized mouse cells are engineered in culture to express functional components of retrovirus virions. Genetic components of the virus, including gag, pol and env genes, are inserted into the cell genome at different sites in order to reduce the chance of generating wild-type virus by recombination between inserted and endogenous retrovirus elements^47,48(Fig. 3). These same cells are then transfected with retrovirus vector sequences consisting of long terminal repeat elements (LTRs) for subsequent genomic integration of the transgene and a psi sequence, for packaging of vector sequences into virions. The therapeutic sequence is the thymidine kinase (TK) gene of herpes simplex virus type 1 (HSV) which converts certain nucleotide analogues, such as ganciclovir (which is not metabolized by the host cell thymidine kinase) to toxic nucleotides that disrupt DNA replication and thereby kill dividing cells^49,50. These producer cells releasing the retrovirus vector are grafted into the tumor bed and should transfer the HSV-TK gene to mitotically active tumor cells and their progeny, but not to non-dividing normal brain cells (Fig. 2). Transduced tumor cells become sensitized to subsequent systemic treatment with GCV. Sadly, studies in humans indicate that this procedure is very inefficient at gene delivery to tumor cells and has no marked therapeutic efficacy⁴. The low level of gene delivery probably reflects a number of factors, including the slow rate of retrovirus release from cells; the instability of virions; the fact that many tumor cells are not dividing during the therapeutic window; the fixed position of grafted cells within the tumor, while tumor cells themselves migrate in the brain; and the rapid immune rejection of these foreign cells expressing viral antigens on their surface. In fact, this last point has been viewed by some as possibly the most important therapeutic component of the system^51,52, as in the context of GCV treatment, immune cells are attracted to a region of dying tumor cells and thus may generate a response against tumor antigens. However, if this is the primary means of tumor killing by this mode of therapy, there are probably more controlled ways of effecting this, rather than introducing mouse cells expressing high levels of retrovirus sequences. Ways to improve gene delivery include direct injection of more stable, higher titer vectors that can deliver genes to both dividing and non-dividing cells (including newer versions of retrovirus, HSV, adenovirus, AAV and lentivirus vectors^53,54) and the use of cells with a lower retroviral load and immunologic camouflage, which may therefore survive longer on site and migrate within and beyond the tumor to reach infiltrating cells.

Fig. 3 Retrovirus sequences and functions. The Moloney murine leukemia virus (M-MULV) is used as the backbone for most retrovirus vectors. Translation products of the retroviral genes are gag, structural components of the virus particles; pol, reverse transcriptase; and env , the virus coat protein. a, Mouse and pig cells bear two types of functionally active retrovirus sequences in their genome, endogenous retrovirus and retrotransposons. The former bear all retrovirus genes: gag (G), pol (P) and env (E), and can be packaged in retrovirus particles. These particles are non-infectious (//) for cells expressing the same virion env proteins (termed interference). Retrotransposons lack env sequences and thus cannot make infectious particles62, but do produce retrovirus proteins and RNA. b, Vector producer cells are mouse cells, which express components shown in panel a, and have been transfected with: retrovirus genes encoding packaging functions that lack the y (packaging) sequence, so that their transcripts cannot be packaged in particles and a vector construct bearing the therapeutic gene (X), which can be packaged in virions by virtue of the sequence. These vectors (virions bearing the vector transcripts) bud out from packaging cells continually without killing them.

It would be hard to envision cells more "qualified" to generate a new pathogenic retrovirus than these genetically engineered mouse producer cells. The mouse genome is full of endogenous retroviral sequences, some of which can be activated to produce retrovirus particles; further, these producer cells have been empowered with DNA sequences designed to produce virions capable of infecting human cells. Although producer cell lines are rigorously screened prior to transplantation, endogenous retroviral sequences may be activated following transplantation, by close proximity to tumor cells that release many growth factors and cytokines, or as a result of subsequent treatments, such as radiation and chemotherapy. This interaction between mouse and human cells occurs, at least initially, in the context of the immune-privileged environment of the brain⁵⁵. The relative lack of concern regarding generation of replication competent retrovirus presumably reflects the brief sojourn of the mouse producer cells in the human brain because their high immunogenicity results in their rejection within a few days to weeks.

Retrovirus risks and xenogeneic transplantation

One of the risks of xenotransplants is the large number of animal pathogens (with more likely to be discovered) that can cause persistent infections in humans. Prior to transplantation, it would be necessary to screen for all such pathogens⁵⁶. These zoonotic agents include bacteria, prion-like molecules, parasites and a number of viruses such as HSV, papilloma virus, adenovirus, hepatitis virus, measles virus and retroviruses⁵⁷. The chance of cross-species infection is presumably increased by such factors as: the time and closeness of in vivo contact between cells; the relative susceptibility of human cells; and the competence of the immune system to detect and block infectious agents. For several reasons, there are special concerns about the possible risk of retroviruses⁵⁸. First, mammalian genomes are invested with many retrovirally derived sequences, varying from virtually fossilized to those capable of producing replication competent virus, depending on the species. Second, the retroviral genome is especially adept at mutational and recombinational events such that it can change tropism and incorporate toxic viral components and cellular genes. Third, although for the most part retroviral infections are benign, they can cause a variety of diseases, including malignant and degenerative conditions. Notably, situations have been described in which retroviral sequences of one animal species are believed to have led to disease conditions in another species^58,59. For example, C-type retrovirus from Asian mice appear to have been passed as infectious agents to gibbons⁶⁰. Finally, there is a long latent period (many years) between initial exposure and manifestation of infectious virus. In evaluating the relative merit of xenotransplants, therefore, concerns arise both for potential hazards to the patient and for the possibility of horizontal or vertical transmission of infectious retroviral genomes within the population.

The three species considered in the above xenotransplantation paradigms (human as the recipient, and pig and mouse as donors) differ dramatically in the retroviral load of their genome⁵⁸. Although about 1% of the human genome consists of relatively conserved retroviral sequences⁶¹, these are not normally expressed -- exceptions being, for example, in the placenta and some tumors, and after treatment with agents that demethylate DNA. However, even when activated, these human sequences do not yield virion particles or infectious virus. This loss of function reflects their introduction into the genome millions of years ago with consequent mutational inactivation of most retroviral genes⁶². However, the margin of safety appears thin when one considers that a related primate, the baboon, generates infectious retrovirus that can be transmitted horizontally and vertically, and that can cause leukemia⁵⁹. Only recently has the retroviral load of pig cells been evaluated. The pig genome contains approximately 50 proviral related sequences (most of which are probably defective) and porcine cell lines produce C-type particles that infect human cells in culture¹. Mitogenic activation of normal pig peripheral blood mononuclear cells can also lead to the production of C-type particles infectious for both pig and human cell lines (C. Wilson, FDA, personal. communication.). In contrast, the mouse genome is extremely rich in retroviral sequences, many of which appear to have been introduced more recently in evolutionary terms⁵⁸. Some mouse strains produce infectious particles including B-type particles, related to murine mammary tumor virus and C-type particles related to murine leukemia virus. These endogenous retroviruses are not normally pathogenic in mice, with the exception of murine mammary tumor virus, which causes T-cell lymphomas in some strains⁶³. These murine particles can also infect human cells. For example, human tumor cells passaged in nude mice frequently become infected⁵⁸, and a mouse producer cell line contaminated with replication competent recombinant retrovirus and used to transduce bone marrow cells in culture, produced T-cell lymphomas following re-introduction of transduced cells into rhesus monkeys⁶⁴. Still, considering all the retroviral sequences to which humans are exposed from internal and external sources, murine-derived C-type particles have not yet been associated with pathogenic or epidemic episodes in humans, suggesting that the frequency of disease-generating events is exceedingly rare. That said, HIV is a member of the larger retrovirus family.

Since retrovirus sequences seem to be an unavoidable component of the mammalian genome, it is critical to understand how they can cause disease and what conditions favor this outcome. The pathogenic effects can be divided into two broad categories -- malignant transformation and degeneration. Retroviruses are unique among animal viruses in that they are able to incorporate active oncogenes from the host cell genome into their own genome⁵⁸. Such replication-defective viral genomes can be propagated in vivo in concert with helper retrovirus, resulting in rapid and extensive transformation of cells that can cause a broad range of cancers, including lymphomas and leukemias. Retroviruses also mediate primary transforming events by insertional activation of proto-oncogenes in the host cell genome⁶⁵or by production of modified viral proteins that trigger division of host cells^66,67. In the second pathogenic mode, retroviruses can cause cell degeneration or toxicity. For instance, mutation of the envelope protein can change retroviral specificity such that these viruses can repeatedly infect cells^68,69. The resulting viral burden of genomic mutations caused by multiple viral integrations can produce metabolic compromise and elicit an immune reaction against infected cells. Insertional integration of retroviral genomes into cellular genes has also been associated with hereditary deficiency states in mice^70,71, and several strains of retrovirus cause neuronal degeneration. In newborn mice this process involves susceptibility of neural cells during developmental proliferation and production of a toxic env protein in the CNS^72,73. Human T-cell lymphotropic virus (HTLV) type I, in addition to causing malignant transformation of lymphoid cells, produce a neurodegenerative condition in adults, termed tropic spastic paraparesis (in Jamaica), or HTLV-I associated myelopathy (in Japan)⁷⁴. Other retroviruses can cause immune deficiency. For instance, a fusion protein generated by a recombination between murine leukemia virus gag and pol genes results in an immunodeficiency syndrome in mice⁷⁵and feline leukemia virus causes a wasting and immune deficiency syndrome by repeated infections in cats⁷⁶.

There are, however, a number of natural biological limitations to the spread of retroviral infection⁵⁸. For C-type particles, these include not only short half life and infection limited to dividing cells, but also restriction to cells that express the appropriate receptor for the particular virion envelope protein. Further, once infected, cells develop resistance to later infections by virions of the same envelope type. Transfer among individuals is largely limited to transplacental and tissue exchange, or sexual contact. Although the generation of recombinant replication competent retrovirus is small (seemingly smaller in the case of normal pig cells than in the case of mouse cells producing vectors), a worst case scenario would be the generation of a novel infectious virus with pathologic and infectious potential among humans. The normal adult brain may be a particularly safe place to transplant xenogenic cells, since it has relatively few dividing cells as compared to other tissues and hence a reduced chance of retrovirus infection, although this is not the case in the context of a brain tumor where dividing tumor cells would be permissive to retrovirus infection. Infection of human cells with C-type porcine or murine particles would result in the generation of new particles lacking galactose alpha (1-3) galactose residues and hence resistant to complement-mediated lysis⁷⁷ (see Commentary, page 944). The immune suppression used in transplant recipients, might further limit any elimination of such particles. Infectious events could cause disease, for example by insertional activation of an oncogene in a proliferative cell type, mutational generation of a toxic viral protein or recombination with proviral sequences expressed in human cells, which could in turn increase the infectivity and replication competence of the particles.

Careful monitoring of blood samples by PCR and serology is warranted in all xenogeneic cell transplant recipients to assess the possible generation of infectious retrovirus. Indeed such monitoring is supposed to be a component of clinical trials currently under way⁴¹. Further, such recipients should not be of childbearing age and should not donate their blood or organs to others. Other precautionary measures might include immunizing recipients against porcine or murine retroviruses.

Serious concerns remain about the value of transplanting murine cells for treatment of human brain tumors. In contrast, given the obvious advantages of a domesticated animal like the pig, as a source of tissue for transplantation in human neurodegenerative diseases, pharmaceutical measures, transgenic technology and breeding should be investigated to develop and optimize any possible medical benefits and to reduce concerns for the safety of such xenogeneic procedures.

Acknowledgements

References

2. Short, M.P. et al. Gene delivery to glioma cells in rat brain by grafting of a retrovirus packaging cell line. J. Neurosci. Res. 27, 427-439 (1990).

3. Culver, K.W. et al. In vivo gene transfer with retroviral vector producer cells for treatment of experimental brain tumors. Science 256, 1550-1552 (1992).

4. Blaese, R.M. et al. In situ delivery of suicide genes in cancer therapy. XIIIth Europ. Assoc. Cancer Research Meeting. Berlin, Germany (1994).

5. Björklund, A. Dopaminergic transplants in experimental parkinsonism: cellular mechanisms of graft-induced functional recovery. Curr. Opin. Neurobiol. 2, 683-689 (1992).

6. Wenning, G.K. et al. Short- and long-term survival and function of unilateral striatal dopaminergic grafts in Parkinson's disease. Ann. Neurol. 42, 95-107 (1997).

7. Kordower, J.H. et al. Neuropathological evidence of graft survival and striatal reinnervation after the transplantation of fetal mesencephalic tissue in a patient with Parkinson's disease. N. Engl. J. Med. 332, 1118-1124 (1995).

8. Philpott, L.M. et al. Neuropsychological functioning: following fetal striatal transplantation in Huntington's Chorea: three case presentations. Cell Transpl. 6, 203-212 (1996).

9. Deacon, T. et al. Histological evidence of fetal pig neural cell survival after transplantation into a patient with Parkinson's disease. Nature Med. 3, 350-353 (1997).

10. Quinn, N. The clinical application of cell grafting techniques in patients with Parkinson's disease. Prog. Brain Res. 82, 619-625 (1990).

11. Pakkenberg, B., Moller, A., Gundersen, H.J., Mauritzen-Darn, A. & Pakkenberg, H. The absolute number of nerve cells in substantia nigra in normal subjects and in patients with Parkinson's disease estimated with an unbiased stereological method. J. Neurol. Neurosurg. Psychiat. 54, 30-33 (1991).

12. Guttman, M. et al. (C-11)RTI-32 PET studies of the dopamine transporter in early dopa-naive Parkinson's disease - implications for the symptomatic threshold. Neurology 48, 1578-1583 (1997).

13. Birkmayer & Hornykiewicz, O. Wien Klin. Wochenschr. 73, 787-788 (1961).

14. Bravi, D. et al. Wearing-off fluctuations in Parkinson's disease. Ann. Neurol. 36, 27-31 (1994).

15. Marsden, C.D. Problems with long-term levodopa therapy for Parkinson's disease. Clin. Neuropharmacol. 17 (Suppl 2), S32-S44 (1994).

16. Benabid, A.L., Pollak, P., Louveau, A., Henry, S. & De Rougement, J. Combined (thalamotomy and stimulation) stereotactic surgery of the VIM thalamic nucleus for bilateral Parkinson disease. Appl. Neurophysiol. 50, 344-346 (1987).

17. Iacono, R.P. et al. The results, indications, and physiology of posteroventral pallidotomy for patients with Parkinson's disease. Neurosurgery 36, 1118-1124 (1995).

18. Benabid, A.L., Pollak, P., Hoffmann, D., Le Bas, J.F. & Gao, D.M. Chronic high-frequency thalamic stimulation in Parkinson's disease. In Therapy of Parkinson's disease, Second Ed. (eds. Koller, W.C., Paulson, G.) 381-402 (Marcel Dekker Inc., New York, 1995).

24. Widner, H. et al. Bilateral fetal mesencephalic grafting in two patients with parkinsonism induced by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). N. Engl. J. Med. 327, 1556-1563 (1992).

19. Nagatsu, T. & Ichinose, H. GTP cyclohydrolase I gene, tetrahydrobiopterin, and tyrosine hydroxylase gene: their relations to dystonia and parkinsonism. Neurochem. Res. 21, 245-250 (1996).

20. Tam, S.Y. & Roth, R.H. Mesoprefrontal dopaminergic neurons: can tyrosine availability influence their functions? Biochem. Pharmacol. 53, 441-453 (1997).

21. Hurtig, H.I. Problems with current pharmacologic treatment of Parkinson's disease. Exp. Neurol. 144, 10-16 (1997).

22. Fisher, L.J., Jinnah, H.A., Kale, L.C., Higgins, G.A. & Gage, F.H. Survival and function of intrastriatally grafted primary fibroblasts genetically modified to produce L-dopa. Neuron 6, 371-380 (1991).

23. Strecker, R.E. et al. Autoregulation of dopamine release and metabolism by intrastriatal nigral grafts as revealed by intracerebral dialysis. Neuroscience 22, 169-178 (1987).

25. Pakzaban, P. & Isacson, O. Neural xenotransplantation: reconstruction of neuronal circuitry across species barriers. Neuroscience 62, 989-1001 (1994).

26. Klassen, H. & Lund, R.D. Anatomical and behavioral correlates of a xenograft-mediated pupillary reflex. Exp. Neurol. 102, 102-8 (1988).

27. Klassen, H. & Lund, R.D. Parameters of retinal graft-mediated responses are related to underlying target innervation. Brain Res. 533, 181-91 (1990).

28. Honey, C.R., Charlton, H.M. & Wood, K.J. Rat brain xenografts reverse hypogonadism in mice immunosuppressed with anti-CD4 monoclonal antibody. Exp. Brain Res. 85, 149-152 (1991).

29. Saitoh, Y., Matsui, Y., Nihonmatsu, I. & Kawamura, H. Cross-species transplantation of the suprachiasmatic nuclei from rats to Siberian chipmunks (Eutamias sibiricus) with suprachiasmatic lesions. Neurosci. Lett. 123, 77-81 (1991).

30. Hantraye, P., Riche, D., Maziere, M. & Isacson, O. Intrastriatal transplantation of cross-species fetal striatal cells reduces abnormal movements in a primate model of Huntington disease. Proc. Natl. Acad. Sci. U.S.A. 89, 4187-4191 (1992).

31. Galpern, W.G., Burns, L.H., Deacon, T.W., Dinsmore, J. & Isacson, O. Xenotransplantation of porcine fetal ventral mesencephalon in a rat model of Parkinson's disease: functional recovery and graft morphology. Exp. Neurol. 140, 1-13 (1996).

32. Mayer, E., Dunnett, S.B., Pellitteri, R. & Fawcett, J.W. Basic fibroblast growth factor promotes the survival of embryonic ventral mesencephalic dopaminergic neurons-I. Effects in vitro. Neuroscience 56, 379-388 (1993).

33. Nakao, N., et al. Overexpressing Cu/Zn superoxide dismutase enhances survival of transplanted neurons in a rat model of Parkinson's disease. Nature Med. 1, 226-231 (1995).

34. Lampson, L.A. Molecular bases of the immune response to neural antigens. Trends Neurosci. 10, 211-216 (1987).

35. Neumann, H., Cavalie, A., Jenne, D.E. & Wekerle, H. Induction of MHC class I genes in neurons. Science 269, 549-552 (1995).

36. Pedersen, E.B., Poulson, F.R., Zimmer, J. & Finsen, B. Prevention of mouse-rat xenograft rejection by a combination therapy of cyclosporin A, prednisolone and azathioprine. Exp. Brain Res. 106, 181-186 (1995).

37. Wood, M.J.A., Sloan, D.J., Wood, K.J. & Charlton, H.M. Indefinite survival of neural xenografts induced with anti-CD4 monoclonal antibodies. Neuroscience 70, 775-789 (1996).

38. Sachs, D.M. & Bach, F.H. Immunology of xenograft rejection. Hum. Immunol. 28, 245-251 (1990).

39. Steele, D.J. & Auchincloss, H.J. Xenotransplantation. Annu. Rev. Med. 46, 345-360 (1995).

40. Hammer, C., Suckfull, M. & Saumweber, D. Evolutionary and immunological aspects of xenotransplantation. Transplant. Proc. 24, 2397-2400 (1992).

41. Public Health Service, H.H.S. Guideline on Infectious Disease Issues in Xenotransplantation. Federal Register 61, 49920-49927 (1996).

42. Ezzeddine, C.D., et al. Selective killing of glioma cells in culture and in vivo by retrovirus transfer of herpes simplex virus thymidine kinase gene. New Biol. 3, 608-614 (1991).

43. Takamiya, Y., et al. An experimental model of retrovirus gene therapy for malignant tumours. J. Neurosurg. 79, 104-110 (1993).

44. Oldfield, E.H., Ram, Z., Culver, K.W., Blaese, R.M., DeVroom, H.L. & Anderson, W.F. Gene therapy for the treatment of brain tumors using intra-tumoral transduction with the thymidine kinase gene and intravenous ganciclovir. Hum. Gene Ther. 4, 39-69 (1993).

45. Mahaley, M.S., Jr., Mettlin, C., Natarajan, N., Laws, E.R., Jr. & Peace, B.B. National survey of patterns of care for brain tumor patients. J. Neurosurg. 71, 826-836 (1989).

46. Loeffler, J.S., et al. Radiosurgery as part of the initial management of patients with malignant gliomas. J. Clin. Oncol. 10, 1379-1385 (1992).

47. Miller, A.D. Retrovirus packaging cells. Hum. Gene. Ther. 1, 5-14 (1990).

48. Danos, O. & Mulligan, R.C. Safe and efficient generation of recombinant retrovirus with amphotropic and ecotropic host ranges. Proc. Natl. Acad. Sci. USA 85, 6460-6464 (1988).

49. Elion, G.B. The biochemistry and mechanism of action of acyclovir. J. Antimicr. Chemother. 12, 9-17 (1983).

50. Moolten, F.L. Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes: a paradigm for a prospective cancer control strategy. Cancer Res. 56, 5276-5281 (1986).

51. Barba, D., Hardin, J., Sadelain, M. & Gage, F.H. Development of anti-tumor immunity following thymidine kinase-mediated killing of experimental brain tumors. Proc. Natl. Acad. Sci. USA 91, 4348-4352 (1994).

52. Tapscott, S.J., Miller, A.D., Olson, J.M., Berger, M.S., Groudine, M. & Spence, A.M. Gene therapy of rat 9L gliosarcoma tumors by transduction with selectable genes does not require drug selection. Proc. Natl. Acad. Sci. USA 91, 8185-8189 (1994).

53. Kramm, C.M., et al. Gene therapy for brain tumors. Brain Pathol. 5, 345-381 (1995).

54. Naldini, L., Blomer, U., Gage, F.H., Trono, D. & Verma, I.M. Efficient transfer, integration, and long-term sustained expression of the transgene in adult rat brains injected with a lentiviral vector. Proc. Natl. Acad. Sci. USA 93, 11382-11388 (1996).

55. Tada, M. & de Tribolet, N. Recent advances in immunobiology of brain tumors. J. Neuro-Oncol. 17, 261-271 (1993).

56. Michaels, M.G. & Simmons, R.L. Xenotransplant-associated zoonoses. Strategies for prevention. Transplantation 57, 1-7 (1994).

57. Ahmed, R., Morrison, L.A. & Knipe, D.M. Persistence of viruses. In Fields Virology, Third Ed. (eds. B.N. Fields, D.M.K., P.M. Howley et al.) 221-241 (Lippincott-Raven Publishers, Philadephia, PA, 1996).

58. Coffin, J.M. Retroviridae: the viruses and their replication. In Fields Virology, Third Ed. (eds. B.N. Fields, D.M.K., P.M. Howley et al.) 1767-1847 (Lippincott-Raven Publishers, Philadelphia, PA, 1996).

59. Kawakami, T.G., Sun, L. & McDowell, T.S. Natural transmission of gibbon leukemia virus. J. Natl. Cancer Inst. 61, 1113-1115 (1978).

60. Lieber, M.M., Sherr, C.J., Todaro, G.J., Benveniste, R.E., Callahan, R. & Coon, H.G. Isolation from the asian mouse Mus caroli of an endogenous type C virus related to infectious primate type C viruses. Proc. Natl. Acad. Sci. USA 72, 2315-2319 (1975).

61. Temin, H.M. Reverse transcription in the eukaryotic genome: retroviruses, pararetroviruses, retrotransposons and retrotranscripts. Mol. Biol. Evol. 6, 455-468 (1985).

62. Krieg, A.M. Human Endogenous Retroviruses. Science & Medicine 4, 34-43 (1997).

63. Stoye, J.P., Moroni, C. & Coffin, J. Virological events leading to spontaneous AKR thymomas. J. Virol. 65, 1273-1285 (1991).

64. Vanin, E.F., Kaloss, M., Broscius, C. & Nienhuis, A.W. Characterization of replication-competent retroviruses from nonhuman primates with virus-induced T-cell lymphomas and observations regarding the mechanism of oncogenesis. J. Virol. 68, 4241-4250 (1994).

65. Kung, H.-J., Boerkoel, C. & Carter, T.H. Retroviral mutagenesis of cellular oncogenes: a review with insights into the mechanisms of insertional activation. Curr. Top. Micro. Immun. 171, 1-25 (1991).

66. Li, J.-P. & Baltimore, D. Mechanism of leukemogenesis induced by mink cell focus-forming murine leukemia viruses. J. Virol. 65, 2408-2414 (1991).

67. Wolff, L. & Ruscetti, S. The spleen focus-forming virus (SFFV) envelope gene, when introduced into mice in the absence of other SFFV genes induces acute erythroleukemia. J. Virol. 62, 2158-2163 (1988).

68. Weller, S.K. & Temin, H.M. Cell killing by avian leukosis viruses. J. Virol. 39, 713-721 (1981).

69. Kristal, B.S., Reinhart, T.A., Hoover, B.A. & Mullins, J.I. Interference with superinfection with cell killing and determination of host range and growth kinetics mediated by feline leukemia virus surface glycoproteins. J. Virol. 67, 4142-4153 (1993).

70. Stoye, J.P., Kenner, S., Greenoats, G.E., Moran, C. & Coff, J.M. Role of endogenous retroviruses as mutagens: the hairless mutation of mice. Cell 54, 383-391 (1988).

71. Barker, D.D., Wu, H., Hartnung, S., M, B. & Jaenish, R. Retrovirus-induced insertional mutagenesis mechanism of collagen mutation in Mov13 mice. Mol. Cell. Biol. 11, 5154-5163 (1991).

72. Lynch, W.P., Czub, S., McAtee, J., Hayes, S.F. & Portis, J.L. Murine retrovirus-induced spongiform encephalopathy: productive infection of microglia and cerebellar neurons in accelerated CNS disease. Neuron 7, 365-379 (1991).

73. Wiley, C.A. & Gardner, M. The pathogenesis of murine retroviral infection of the central nervous system. Brain Pathol. 3, 123-128 (1993).

74. Cann, A.J. & Chen, I.S.Y. Human T-cell leukemia virus types I and II. In Fields Virology, 3 Ed. (eds. Fields, B.N., Knipe, D.M. & Howley, P.M.) 1849-1880 (Lippincott-Raven Publishers, Philadelphia, PA, 1996).

75. Aziz, D.C., Hanna, Z. & Jolicoeur, P. Severe immunodeficiency disease induced by a defective murine leukaemia virus. Nature 338, 505-508 (1989).

76. Mullins, J.I., Chen, C.S. & Hoover, E.A. Disease-specific and tissue-specific prodcution of unintegrated feline leukemia virus variant DNA in feline AIDS. Nature 319, 333-336 (1986).

77. Rother, R.P., et al. A novel mechanism of retrovirus inactivation in human serum mediated by anti-alpha-galactosyl natural antibody. J. Exp. Med. 182, 1345-1355 (1995).

Author's Addresses

1. Neuroregeneration Laboratory, McLean Hospital, Harvard Medical School, Belmont, Massachusetts 02178, USA

2. Program in Neuroscience, Harvard Medical School, Boston, Massachusetts 02115

3. Department of Neurology, Massachusetts General Hospital, Boston, Massachusetts 02114, USA

4. Molecular Neurogenetics Laboratory, Massachusetts General Hospital, Charlestown, Massachusetts 02139, USA