The neuronal complement of cells within the central nervous system (CNS) is not capable of significant regeneration. For many neurological, especially neurodegenerative, disorders, effective therapy is either short acting or unavailable.

With the establishment in the mid-1980s of reliable methods of neural tissue transplantation came the rapid realization that this technique could be applied clinically. In a short period of time, many centres worldwide initiated programmes of neural tissue transplantation of either fetal tissue or autologous adrenal medulla into patients suffering from the neurodegenerative disorder, Parkinson's disease. Although these attempts have met with only limited success they have done much to stimulate research in the field of neural transplantation, a long neglected area of transplantation biology. This, together with our increasing knowledge of neurobiology will hopefully allow the full potential of this exciting area of transplantation to be established.

The first published attempts at neural tissue transplantation date back to the late 19th century, but only following the work of the modern day pioneers, published in the early 1970s, has the field of neuronal transplantation progressed. Having defined the conditions required for graft survival and devised methods of graft identification, initially by histochemistry and more recently by immunohistochemistry, the stage was set for a rapid advance in the techniques of neural transplantation, and, to a lesser extent, in the understanding of the immune response to allogenic neural tissue. This was greatly assisted by the brain being considered a site of immunological privilege, a concept missing from a series of experiments using tumour grafts between animals of ill-defined genetic background. The survival of these grafts, together with the apparent lack of major histocompatibility complex antigen expression within the brain and the presence of the blood–brain barrier, meant that the question of rejection of allogenic tissue, one of the major stumbling blocks in all other fields of transplantation, did not have to be addressed, at least in theory.

A large variety of animal models, both functional and structural, have been devised to examine various aspects of neural tissue transplantation. The most widely used has been a movement disorder model of Parkinson's disease first described in 1973, induced by the injection of 6-hydroxydopamine into the substantia nigra of animals. This leads to a selective degeneration of the nigral projections to the striatum, similar to the changes seen in Parkinson's disease. The resultant neuronal degeneration causes a fall in the dopamine concentration in the striatum.

Administration of a dopamine agonist produces a characteristic movement disorder that is quantifiable. Insertion of fetal substantia nigra reduces this movement disorder, and considerable outgrowth of monoamine-containing neurones can be shown histochemically.

As an alternative, and by way of circumventing the putative ethical problems surrounding the use of fetal tissue, grafts of adrenal medulla (a source of dopamine) have also been used and appear to produce a similar improvement in the movement disorder.

Another, and perhaps widely under-used, functional model involves the selective reduction of the cholinergic input to the hippocampus by a lesion in the septohippocampal pathway. This induces a memory and behaviour abnormality that can be improved by the insertion of a septal graft rich in cholinergic neurones. Similar losses of specific populations of neuronal cells are seen in patients suffering from the neurodegenerative disorder, Alzheimer's disease.

Rather than inducing defects, congenital neuronal deficiency and the resultant deficit that this produces has provided a further way of examining the functional influence of a graft. One such model is the Brattleboro rat, which suffers from diabetes insipidus due to a deficiency of vasopressin-secreting neurones. Grafts of fetal anterior hypothalamus placed into the recipient third ventricle partially alleviate this deficiency.

Finally, a recently devised model utilizes the Purkinje cell degeneration mutant mouse, into which fetal cerebellar primordia are placed. The grafts appear to repopulate the deficient molecular layer of the cerebellum, synapses developing between the Purkinje cell-containing grafts and the host cell population.

In all these models, it is far from clear how the grafts exert their effect. Since there does appear to be synapse formation within the graft it is possible that a direct graft–host interaction is involved, but it may be that the grafts are merely acting as an exogenous source of the deficient neurotransmitter, or perhaps the source of a neurotrophic factor that is causing a degree of regeneration or repair of the host cells.

Whatever the exact mode of action, the reported success of grafts in these and many other studies, including autografts of adrenal medulla in a primate model, led to the clinical application of this technique to patients suffering from Parkinson's disease.

Graft survival in any situation requires the implantation of viable tissue to an area of adequate blood supply with the minimum of trauma. For neural tissue transplantation, the use of young (fetal) tissue is essential. Fetal tissue appears to survive a short period of anoxia, while still retaining the capacity for continued growth and differentiation.

Neural tissue transplants may be of solid tissue or cell suspensions and can be placed directly into the host brain, into a preformed cavity or into the ventricle. The use of a cell suspension depends upon dissociation of the required area from the donor brain using trypsin, with subsequent stereotactic implantation into the host. Graft viability depends upon the origin of the tissue, the age, the mechanical trauma applied, and the use of trypsin. A graft containing viable cells is required to obtain good survival rates.

Similar parameters govern the implantation of solid tissue grafts, unless into a preformed cavity, where a delay in graft insertion may improve survival. The ventricular system offers a preformed cavity with a ready blood supply. However, this site only provides graft access to periventricular tissues and will not stop graft migration. Placement into the ventricle can be easily achieved stereotactically with minimal trauma.

In 1985, the results following the implantation of autografts of adrenal medulla in two patients suffering from Parkinson's disease were published. The initial symptomatic improvement noticed in these patients was very short lived, the grafts having been stereotactically placed into the head of the caudate nucleus. The publication of this work was met with initial scepticism, especially in view of the poor results. However, a subsequent report, published in 1987 from Mexico, described enormous success using adrenal medulla autografts in two young patients. This sparked off the clinical application of this technique in many centres worldwide.

Patients suffering from Parkinson's disease seemed the ideal starting place for studies on human neural tissue transplantation. Advantages include a specific neurochemical defect, together with the ability to use not only fetal cells, which show both good survival and extensive neurone outgrowth, but also autologous adrenal medulla as the graft material.

The preliminary rodent work utilizing these systems was applied directly to the human situation, and now in excess of 100 patients have received adrenal medulla to caudate autografts. Sporadic claims of success generated further interest. In general the results were disappointing, not only in terms of morbidity and mortality, but also due to a lack of symptomatic improvement. In 1988, questions were asked about the validity of the technique and a plea was made for extensive basic research before clinical trials proceeded any further. As an alternative to adrenal medulla, fetal tissue containing substantia nigra had also been used as a source of dopaminergic neurones. In the 1-methyl-4-phenyl-1,2,5,6-tetrahydropyridine-induced Parkinson's disease model in both rodents and primates, considerable success in amelioration of many motor deficits has been seen following such treatment. This naturally resulted in clinical trials using fetal mesencephalic tissue. Initial results, where such grafts have been placed into the putamen of the basal ganglia, rather than the caudate nucleus, appear encouraging.

The success of this work shows that human fetal dopaminergic-containing neurones can survive, grow, and continue to synthesize and store dopamine following allotransplantation into the putamen of immunosuppressed hosts. There are however, still many unanswered questions surrounding neural transplantation. Graft–host interaction is certainly not fully understood, although implied by the improvement of clinical signs. How long will the graft function, and if it fails, is this due to immune rejection or the underlying disease process? Is it that the graft is acting merely as a source of growth factor?

Future success and the ability to develop a robust system depends upon answers being established to these difficult questions.

Already, work is being directed towards the immune response to allogenic neural tissue. This is an area of major importance, for the criteria upon which the brain has for so long been considered a site of immunological privilege can no longer be substantiated. There are now several reports showing that neuroepithelium is capable of major histocompatibility complex antigen expression both in vitro and in vivo, and that neural tissue allografts are immunologically rejected. How this response is generated is uncertain, but it appears that it is only the glial population of cells that is capable of such antigen expression.

The question of how to overcome this response is therefore foremost. This may be achieved by the use of immunosuppressive drugs, but treatment with monoclonal antibodies directed against T-cell subsets involved in rejection also appears to enhance graft survival. The fact that the neuroepithelium can be dissociated into a single cell suspension and that alloreactivity appears to be related to only a small subpopulation of the cells may indicate that grafts devoid of those reactive cells will provide an alternative way to circumvent the rejection response. Initial studies using this technique suggest that allograft survival can be improved in this way.

The use of genetically engineered cells as a source of either growth factors or neurotransmitters offers a method of graft production that is not dependent upon tissue availability and would enable the development of grafts containing large populations of cells producing precisely what is required. Obviously, graft–host interaction would be limited, but this may be ideal in situations where only a source of growth factor or transmitter is required.

The biology of growth factors is expanding rapidly and the use of growth factors in the central nervous system has already been seen as a way of ameliorating many and varied neurodegenerative disorders. Neuronal cell degeneration appears to be arrested by direct injection of nerve growth factor. Cholinergic neurones, the cell type affected in patients suffering from Alzheimer's disease, appear to be especially sensitive to the effects of this factor. Careful optimism surrounds the future application of this finding and preliminary studies indicate that infusions of nerve growth factor actually reverse the atrophy of these neurones in the brains of aged rats.

Important structural studies, at both light and electron microscopic levels using more specific markers, continue in an attempt to detail the precise nature of the interaction between the graft and the host tissues.

That neural tissue can be successfully transplanted, and that the basic techniques are available has now been established. How to apply these techniques to best advantage, however, remains unclear. Future work must be directed towards more basic research, with judicious clinical application as a result of that research. In that way, the full potential therapeutic benefits may be realized and hopefully neural tissue transplantation will become widely available for the treatment of a variety of neurological disorders.

Borges LF. Historical development of neural transplantation. Appl Neurophysiol 1988; 51: 265–77.

Fuchs HE, Bullard DE. Immunology of transplantation in the central nervous system. Appl Neurophysiol 1988; 51: 278–96.

Backlund E-O, et al. Transplantation of adrenal medullary tissue to the striatum in Parkinsonism. J Neurosurg 1885; 62: 169–73.

Sladek JR Jr, Shoulson I. Neural transplantation: a call for patience rather than patients. Science 1988; 240: 1386–8.

Lindvall O, et al. Grafts of fetal dopamine neurons survive and improve motor function in Parkinson's disease. Science 1990; 247: 574–7.

Bartlett PF, Rosenfeld J, Bailey KA, Cheesman H, Harvey AR, Kerr RSC. Allograft rejection overcome by immunoselection of neuronal precursor cells. Prog Brain Res 1990; 82: 153–60.