Back to Home page

This section is from the article: Baev et al., Physiology and pathophysiology of cortico - basal ganglia - thalamocortical loops: Theoretical and practical aspects. Prog Neuropsychopharmacol Biol Psychiatry. 2002, 26(4): 771-804.

5.2.4. Transplantation

When degenerative loss of nigrostriatal dopaminergic cells was discovered to be the etiological basis of PD, the possibility of substituting them with transplanted dopamine-producing cells—whether of neural, paraneural, or transfected cell origin (see Marciano and Greene, 1992)—was embraced.  Within the framework of the proposed theory, transplantation should be considered a way to restore damaged function by replacing a whole automatism (organ transplantation) or by implanting new structural elements (tissue or cellular transplantation) in the malfunctioning system with the hope that they will be integrated so that normal function can be restored. In neurobiology, only tissue and cellular transplantation has been studied intensively, especially fetal neurotransplantation. Transplantation of the whole brain or its entire parts remains science fiction.

Neurotransplantation studies and transplantation experimental therapies in neurological disorders have produced modest results. The miraculous results occasionally reported, for example, the successful treatment of Parkinson’s disease by transplanting adrenal cells into striatum, are rarely confirmed and are excluded from this discussion.

Numerous studies of brain stem cells renewed interest in neurotransplantation.  The underlying simplistic assumption is that transplanted tissue or cells will establish correct connections with the surrounding neurons. As a result, the damaged system would be rewired completely or almost completely and lost function would be restored. To better understand the prospects and limitations of neurotransplantation, consider the following discussion of the phylo- and ontogenesis of NOCS. Neurodegenerative disorders are emphasized because these disorders are the major target of experimental neurotransplantation treatments.

            Each animal species has evolved to fit its ecological niche optimally.  Evolutionary solutions to constructing an animal are usually the best or almost the best based on a set of possible automatisms inherent in a given species. Here, the term automatisms is used in its broadest sense to define cellular, molecular, and other types. If evolutionary solutions are not good enough for survival in a changing environment and a species cannot adjust quickly, it becomes extinct. This statement applies to any NOCS that is also built to optimally match the control needs for corresponding objects.

            To accelerate the evolutionary process, nature evolved numerous mechanisms, including aging and death.  More exactly, numerous mechanisms could be responsible for aging. They have evolved so that individual life expectancy optimally matches evolutionary needs. Life expectancy must be optimal for a species to reproduce; the evolutionary value of an individual life is limited.

            The concept of aging is best illustrated by analyzing biological automatisms (i.e., biological optimal control systems and mechanisms that can make these automatisms less efficient with time). Assume that we can create an artificial life form or its computer model and want to accelerate its evolution by applying the concept of death.  Based on biological optimal control systems (Sections 3 and 5), there are numerous means to do so. One strategy, however, seems the most efficient: gradual programmed deterioration of initiating systems at a specific time after birth.  The speed of this deterioration determines the creature’s life expectancy. All other means appear to be less effective and hence not optimal. Aging likely represents an optimal solution. Aging makes initiating systems vulnerable to unfavorable environmental influences. Initiating systems, especially error systems, are used to adjust the controlling system to a new situation. Individuals with less efficient initiating systems will die sooner than those with more efficient initiating systems.

            Similar reasoning is applicable to NOCS. These systems also are affected by aging. The most serious dysfunctions of NOCSs happens when their error distribution system is damaged. PD was conceptualized as a disease of the error distribution system. Another neurological degenerative disorder, Alzheimer’s disease, frequently affects the older population. Recent data suggest that this disease is also a disease of the error distribution system. In Alzheimer’s disease, the neurons of the nucleus basalis undergo substantial degeneration. Together with dopaminergic and norepinephrine neurons, these neurons broadcast prediction errors as global reinforcement or teaching signals to large number of postsynaptic structures (Schultz and Dickinson, 2000).

            How would NOCSs be built during ontogenesis? Genetic automatisms are used to store phylogenetic solutions, and ontogenesis should be viewed as executing these automatisms to build an individual, including individual NOCSs. Ontogenesis is a multistage process. Each automatism is started by a corresponding initiating signal(s). These automatisms have their own hierarchy and should be coordinated perfectly in space and time. This simple conceptual description clarifies an enormous complexity of ontogenetic development and the uniqueness of some initiating signals and automatisms for this process. Most of these stages occur only once during ontogenesis and are never repeated. For example, at early stages of the development of the nervous system, specific neural cells should receive initiating signals from target cells to establish connections with them. At this early stage, these cells are usually located near each other and become much more distant as the animal grows. If these connections fail to become established in the appropriate time window, they will not be established in the future.  Gene knockout experiments and numerous other studies of birth defects support this perspective.

            Building a NOCS by executing genetic automatisms is a process of structural optimization that is followed by learning. The latter is needed to optimize parameters of the neural network. Genetic information does not account for all the environmental variables that an animal can encounter during its development. It is hard to imagine that genetic memory has this capacity. Consequently, learning processes (i.e., optimization of parameters) are important for the development of NOCSs.

Structural optimization of the network works like an initial structural approximation of the network that allows it to compute specific classes of functions. Structural optimization is the most complex problem associated with creating a neural control system and mostly occurs during embryogenesis and some time after birth (before maturity). However, some minor structural optimization changes can occur even in adults. Without structural optimization, NOCSs could not be built so quickly during development. Without genetic information, structural optimization would not occur as rapidly as it does during ontogenesis. Evolutionary time would be needed and the lifespan of an individual would be insufficient to build complex NOCSs. Millions of years were needed to evolve corresponding genetic automatisms.

Optimization of system parameters, theoretically a quicker and easier process to achieve, mostly relies on learning processes. To facilitate discussion, the process of NOCS development is divided into structural optimization and optimization of parameters. In many cases both optimization processes occur simultaneously and are inseparable in time. They can even share the same genetic automatisms. Embryonic motility (Section 3.5) is a good example of how these two types of optimization occur in the nervous system. Embryonic motility starts immediately after the connections between motoneurons and muscles are established and lasts until birth. Embryonic motility is necessary for the motor control system to develop correctly and should be considered as part of the learning process conducted primarily through trial and error. After birth motility persists until death and plays the same role that it played during embryogenesis.

            The genes for regenerating neural tissue are suppressed in higher vertebrates. Lower vertebrates like salamanders can regrow a lost extremity or regenerate a damaged spinal cord. The regenerating capabilities of some simpler animals are even more remarkable. Vertebrates higher on the evolutionary ladder do not possess this capability. Why? And how did the genes for regenerating neural tissue become suppressed? These questions have no answers yet. However, logical conclusions can be based on the organizing principles of NOCSs in higher vertebrates.

            The more evolutionarily complex an animal is, the more it relies on learning while it copes with its environment (i.e., by using acquired automatisms).  In the highest vertebrates, complex models of very complex controlled objects must be built within the highest NOCSs. Their construction usually requires considerable time and is the major reason why the interval between birth and maturity was significantly extended during the course of evolution. Even lower NOCSs become more controlled by higher ones. In humans, for example, biped locomotion or the use of hands is impossible without the involvement of the highest motor control levels, and the learning period is lengthy. Simpler animals can perform simpler locomotions like quadruped locomotion or swimming immediately after birth.

            This extended time is needed for structural and parameter optimization of higher level NOCS. In higher vertebrates, some stages of structural optimization in higher NOCSs can occur only after birth (i.e., after substantial development of lower NOCSs, which are controlled objects for the higher ones). The cerebellum is a typical example of how initial structural approximation can be achieved in the nervous system. Its network is hardwired after birth when lower automatisms have already formed. In the initial stages of cerebellar wiring, several climbing fibers connect with a single Purkinje cell. Later, only one climbing fiber input survives. The other climbing fibers whose signals could not be minimized likely were rejected. Experiments with visual deprivation are another excellent example of the timing of structural optimization in ontogenesis. A brief period of visual deprivation immediately after birth prevents cortical visual detectors from forming, and the animal will remain blind for the rest of its life.

            Consequently, the most likely reason for suppressing regenerating genes of neural tissue in higher vertebrates is the enormous complexity of their highest NOCSs, which undergo complex optimization processes during ontogenesis. Theoretically, it is easy to explain why the complexity of optimization processes becomes a prohibiting factor for the regeneration of the nervous system. If damage in the highest NOCS could be repaired in an adult higher vertebrate animal, the structure of the damaged NOCS and its connections with surrounding NOCSs could be repaired and parameters of these systems could be adjusted properly. To do so, neurons would have to multiply and grow. Each growing neuron must be provided with the correct sequence of initiating signals to establish the correct connections. Damaged and surrounding NOCSs would have to return to the initial stage of structural optimization and repeat the entire developmental process.

All the knowledge accumulated within these NOCSs would be lost. The animal loses the automatisms that these NOCSs subserved. The development of new NOCS and corresponding automatisms can take years. Who would care for such an animal in the wild? Because evolution is usually parsimonious, it is more reasonable to maintain a population of a species through reproduction mechanisms that already include caring for young. If such repairs could be performed in humans as a result of a breakthrough in medical technology, an individual would become a completely different individual with new experiences.

            There are two major reasons for a gene to become suppressed during the course of evolution.  Genes either become ineffective or begin to contradict new, more effective genetic solutions. In higher vertebrates, genes of regeneration of the nervous system became incapable of repairing damage to complex higher NOCSs, which evolved as a result of new more complex genetic solutions. Even if genes for regenerating neural tissue are artificially activated in a vertebrate animal whose highest NOCSs was injured, these genes most likely could not repair these systems. During evolution, these genes were created to repair much simpler systems. Evolutionary circumstances did not favor the development of such genes for complex NOCSs.

            Neurotransplantation differs from the transplantation of entire organs (e.g., heart, liver, or kidney), in which a damaged automatism is replaced with a new one. In contrast, neurotransplantation repairs a damaged neuronal automatism by adding new neuronal elements. It is based on the assumption that these new elements “know” how to establish correct connections with surrounding neuronal structures. The above analysis shows that this assumption is incorrect. These cells do not “know” how to establish correct connections in an adult brain.  During development, new neural cells appear only if the development of surrounding neurons is perfectly coordinated in time so that correct connections can be established. In an adult brain, a transplanted neural or stem cell will not receive the sequence of signals needed to establish the necessary connections. Therefore, a transplanted cell most likely can establish only aberrant connections with surrounding neurons.

Ironically, the establishment of wrong connections with the surrounding neurons is most often mentioned as evidence that transplanted cells have been integrated with the surrounding neural tissue. Experimental facts like growth of the dendritic tree, new synapses, and the limited appearance of new neurons in an adult brain are considered strong support for neurotransplantation. These capabilities are necessary but not sufficient for successful neurotransplantation.  The capability to heal a skin cut on a finger should not be identified with the capability to regrow an entire extremity.

A network must possess complex, specific computational mechanisms to be able to incorporate new neuronal elements. Neurocomputing has shown that scaling up a neural network is not a simple matter. For example, to extend a trained neural network of 200 neurons to 201 neurons, an entire training session would be needed. The more complex the network is, the more complex these mechanisms should be. Genes of regeneration for neural tissue were blocked during the evolution of higher vertebrates. Consequently, new neurons can appear in the adult brain in ontogenesis and become integrated by the surrounding neural tissue only in regions that have solved the above computational problem (i.e., computational problems related to incorporating a new neuron are not very complex). Theoretically, gene engineering transplanted neurons could provide information about how to find target neurons in an adult brain. Possibly, genes of regeneration could be unblocked and improved in higher vertebrates. If these improvements became possible, science would have surpassed the creative achievements of evolution.

Several clinical reports show that transplantation of fetal dopaminergic cells into the striatum improves parkinsonian symptoms. These cells are implanted in the striatum because the axons of dopaminergic neurons implanted in the substantia nigra pars compacta—the correct location for dopaminergic neurons—cannot reach the striatum in adult brains. These cells will not recreate the correct error distribution circuitry and will not reproduce the necessary temporal dynamics of local dopamine release. The question now becomes: How can these observations be explained by the proposed theory? In other words, why do symptoms improve by implanting embryonic neurons that do not make the correct connections with the surrounding neural tissue? Moreover, graft cells are implanted only in some parts of the striatal network. Several explanations are possible.

First, partial damage to the network computing the model flow serves as a functional neurosurgical procedure (Section 5.2.1). A typical transplantation procedure consists of half a dozen tracks, and damage to the striatal network can be substantial. This mechanism definitely accounts for the immediate clinical transplantation effect (i.e., symptoms improve within the first 48 hours of transplantation). This period for rewiring is brief and edema develops in the striatum.   Second, dopaminergic cells liberate dopamine in the intrastriatal intercellular space in the absence of rewiring. Such a result would add noise to the system (Section 5.2.2).  Third, the new circuits that result from sprouting are not connection-specific. This explanation is equivalent to adding noise to the system. Such nonspecific implanted circuits will work like DBS (i.e., like “biological noise generators”). Fourth, grafted cells initiate a process of repairing the entire system by a function-appropriate structural reorganization similar to that of early ontogenesis. However, it is highly unlikely that simple tissue implantation can initiate a process of de-differentiation followed by differentiation.  Fifth, embryonic grafted cells release trophic factors, which can improve the function of the existing elements of the error distribution system through humoral influences. The potential role of growth factors is discussed in the Discussion.  Finally, the positive result may be a placebo effect and reflects a “desired improvement”. The results of implanting dopamine-producing cells of paraneural and transfected cell origin are also discussed in the Discussion.