Neural grafting of cholinergic neurons in the hippocampal formation

The basal forebrain cholinergic system as target for cell replacement therapy in Parkinson's disease

In recent years there has been a renewed interest in the basal forebrain cholinergic system as a target for the treatment of cognitive impairments in patients with Parkinson's disease, due in part to the need to explore novel approaches to treat the cognitive symptoms of the disease and in part to the development of more refined imaging tools that have made it possible to monitor the progressive changes in the structure and function of the basal forebrain system as they evolve over time. In parallel, emerging technologies allowing the derivation of authentic basal forebrain cholinergic neurons from human pluripotent stem cells are providing new powerful tools for the exploration of cholinergic neuron replacement in animal models of Parkinson's disease-like cognitive decline. In this review, we discuss the rationale for cholinergic cell replacement as a potential therapeutic strategy in Parkinson's disease and how this approach can be explored in rodent models of Parkinson's disease-like cognitive decline, building on insights gained from the extensive animal experimental work that was performed in rodent and primate models in the 1980s and 90s. Although therapies targeting the cholinergic system have so far been focused mainly on patients with Alzheimer's disease, Parkinson's disease with dementia may be a more relevant condition. In Parkinson's disease with dementia, the basal forebrain system undergoes progressive degeneration and the magnitude of cholinergic cell loss has been shown to correlate with the level of cognitive impairment. Thus, cell therapy aimed to replace the lost basal forebrain cholinergic neurons represents an interesting strategy to combat some of the major cognitive impairments in patients with Parkinson's disease dementia.

Fructus mume extracts alleviate cognitive impairments in 5XFAD transgenic mice

Background

Fructus mume (F. mume) has been used as a traditional treatment for ulcer, cough, and digestive problems for many years in Asian countries. Previous studies have demonstrated that F. mume extracts alleviate cognitive deficits in rats with chronic cerebral hypoperfusion and in mice with scopolamine treatments. The present experiment was conducted to examine the effects of F. mume on cognitive impairments in 5XFAD transgenic mice with five familial Alzheimer’s disease (AD) mutations.

Methods

F. mume was administered daily to 5XFAD mice at 12 weeks of age and continued for 90 days. Cognitive function was evaluated using a spatial memory version of the Morris water maze task, the object/location novelty recognition test, and contextual fear conditioning at 24 weeks of age. To elucidate the possible mechanisms underlying the memory improving effects of F. mume in 5XFAD mice, we examined alterations in hippocampal cholinergic function.

Results

Vehicle-treated 5XFAD mice exhibited hippocampus-dependent memory impairments compared with non-transgenic littermates, which was reversed in F. mume-treated 5XFAD mice. In addition, reduced hippocampal choline acetyltransferase (ChAT) levels in 5XFAD mice were reversed by F. mume treatment, indicating that F. mume enhances the effects of cholinergic neuronal function.

Conclusions

F. mume may have therapeutic effects on cognitive impairments in AD.

Microdialysis measures of functional increases in ACh release in the hippocampus with and without inclusion of acetylcholinesterase inhibitors in the perfusate

Because brain extracellular acetylcholine (ACh) levels are near detection limits in microdialysis samples, an acetylcholinesterase (AChE) inhibitor such as neostigmine is often added to microdialysis perfusates to increase ACh levels in the dialysate, a practice that raises concerns that the inhibitor might alter the results. Two experiments compared functional differences in ACh release with and without neostigmine. In the first experiment, 30-60% increases in extracellular ACh concentrations in the hippocampus were evident during food-rewarded T-maze training with 20-500 nm neostigmine in the perfusate but no increases were seen without neostigmine. In the second experiment, 78% increases in ACh release in the hippocampus were seen after injections of the GABA(A) receptor antagonist, bicuculline, into medial septum only if neostigmine (50 nm) was included in the perfusate. These findings suggest that, in the hippocampus, endogenous brain AChEs are very efficient at removing extracellular ACh, obscuring differences in ACh release in these experiments. Therefore, inclusion of AChE inhibitors in the microdialysis perfusate may be necessary under some conditions for observations of functional changes in release of ACh in the hippocampus.

Exposure of Postnatal Rats to a Static Magnetic Field of 0.14 T Influences Functional Laterality of the Hippocampal High-Affinity Choline Uptake System in Adulthood; In vitro Test with Magnetic Nanoparticles

Our previous experiments indicated an age- and sex-dependent functional lateralization of a high-affinity choline uptake system in hippocampi of Wistar rats. The system is connected with acetylcholine synthesis and also plays a role in spatial navigation. The current study demonstrates that a single in vivo exposure of 7- or 14-day-old males to a static magnetic field of 0.14 T for 60-120 min evokes asymmetric alterations in the activity of carriers in adulthood. Namely, the negative field (antiparallel orientation with a vertical component of the geomagnetic field) mediated a more marked decrease in the right hippocampus. The positive field (parallel orientation) was ineffective. Moreover, differences between the carriers from the right and the left hippocampi were observed on synaptosomes pretreated with superparamagnetic nanoparticles and exposed for 30 min in vitro. The positive field enhanced more markedly the activity of carriers from the right hippocampus, the negative that from the left hippocampus, on the contrary. Our results demonstrate functionally teratogenic risks of the alterations in the orientation of the strong static magnetic field for postnatal brain development and suggest functional specialization of both hippocampi in rats. Choline carriers could be involved as secondary receptors in magnetoreception through direct effects of geomagnetic field on intracellular magnetite crystals and nanoparticles applied in vivo should be a useful tool to evaluate magnetoreception in future research.

Is there a future for neural transplantation?

Traditionally neural transplantation has had as its central tenet the replacement of missing neurons that have been lost because of neurodegenerative processes, as exemplified by diseases such as Parkinson disease (PD). However, the effectiveness and widespread application of this approach clinically has been limited, primarily because of the poor donor supply of human fetal neural tissue and the incomplete neurobiological understanding of the circuit reconstruction required to normalize function in these diseases. So, in PD the progress from promising neural transplantation in animal models to proof-of-principle, open-labeled clinical transplants, to randomized, placebo-controlled studies of neural transplantation has not been straightforward. The emergence of previously undescribed adverse effects and lack of significant functional advantage in recent clinical studies has been disappointing and has served to cast a new, and perhaps more realistic, perspective on this treatment approach. In fact, there have been calls by some involved in neural transplantation to return to the drawing board before pressing on with further clinical trials, and the return to basic experimentation. This therefore precipitates the question - is there a future for neural transplantation? It is important to remember that there are a number of possible explanations for the disappointing results from the recent clinical trials in PD, ranging from the mode of transplantation to patient selection. Nevertheless, almost irrespective of these reasons for the current trial results, there have always been significant practical and ethical problems with using human fetal tissue, and so a number of alternative cell sources have been investigated. These alternative sources include stem cells, which are attractive for cell-based therapies because of their potential ease of isolation, propagation and manipulation, and their ability in some cases to migrate to areas of pathology and differentiate into specific and appropriate cell types. Furthermore, the availability of stem cells derived from non-embryonic sources (e.g. adult stem cells derived from the sub-ventricular zone) has removed some of the ethical limitations associated with the use of embryonic human tissue. These potentially beneficial aspects of stem cells means that there is a future for neural transplantation as a means of treating patients with a range of neurological disorders, although whether this will ever translate into a truly effective, widely available therapy remains unknown.

Septohippocampal acetylcholine: involved in but not necessary for learning and memory?

The neurotransmitter acetylcholine (ACh) has been accorded an important role in supporting learning and memory processes in the hippocampus. Cholinergic activity in the hippocampus is correlated with memory, and restoration of ACh in the hippocampus after disruption of the septohippocampal pathway is sufficient to rescue memory. However, selective ablation of cholinergic septohippocampal projections is largely without effect on hippocampal-dependent learning and memory processes. We consider the evidence underlying each of these statements, and the contradictions they pose for understanding the functional role of hippocampal ACh in memory. We suggest that although hippocampal ACh is involved in memory in the intact brain, it is not necessary for many aspects of hippocampal memory function.

Fetal hippocampal CA3 cell grafts enriched with fibroblast growth factor-2 exhibit enhanced neuronal integration into the lesioned aging rat hippocampus in a kainate model of temporal lobe epilepsy

Aging impairs the conduciveness of the lesioned hippocampus for robust survival of neurons derived from homotopic fetal cell grafts (Zaman and Shetty, Neuroscience 109:537-553, 2002), suggesting a need for graft augmentation in fetal graft-mediated therapeutic strategies for the lesioned aging hippocampus. We hypothesize that pretreatment and grafting of donor hippocampal CA3 cells with fibroblast growth factor-2 (FGF-2) considerably enhances graft neuronal integration into the lesioned CA3 region of the aging hippocampus. We employed the optical fractionator cell counting method and quantified the number of surviving cells and neurons derived from 5'-bromodoxyuridine-labeled embryonic day 19 CA3 cell grafts pre-treated and transplanted with FGF-2 into the lesioned CA3 region of the middle-aged and aged rat hippocampus at 4 days post-lesion. In both middle-aged and aged hippocampus, pre-treatment and transplantation of CA3 cell grafts with FGF-2 resulted in a robust yield of surviving cells (72-80% of injected cells) and neurons (62-69% of injected cells) from grafts. The overall yield was dramatically greater than the yield observed earlier from standard (untreated) fetal CA3 cell grafts into the lesioned aging hippocampus but was highly comparable to that observed for standard fetal CA3 cell grafts into the lesioned young hippocampus (Zaman and Shetty, Neuroscience 109:537-553, 2002). Thus, a robust neuronal integration from fetal CA3 cell grafts can be achieved into the lesioned CA3 region of the aging hippocampus with a simple pre-treatment and grafting of donor fetal CA3 cells with FGF-2. These results have implications toward the development of suitable cell grafting strategies for repair of the lesioned aging hippocampus in neurodegenerative diseases, particularly the temporal lobe epilepsy, stroke, and Alzheimer's disease.

Grafts of fetal septal cells after cholinergic immunotoxic denervation of the hippocampus: a functional dissociation between dorsal and ventral implantation sites

Three-month-old Long-Evans rats were subjected to intraseptal infusions of 0.8 microg of 192 IgG-saporin followed, 2 weeks later, by intrahippocampal suspension grafts containing fetal cells from the medial septum and the diagonal band of Broca. The suspensions were implanted in the dorsal or the ventral hippocampus. Sham-operated and lesion-only rats were used as controls. Between 18 and 32 weeks after grafting, all rats were tested in a water maze (using protocols placing emphasis on reference memory or on working memory) and an eight-arm radial maze. The lesion produced extensive cholinergic denervation of the hippocampus, as evidenced by reduced acetylcholinesterase-positivity and acetylcholine content. Depending upon their implantation site, the grafts restored an acetylcholinesterase-positive reinnervation pattern in either the dorsal or the ventral hippocampus. Nevertheless, the grafts failed to normalize the concentration of acetylcholine in either region. The cholinergic lesion impaired working memory performance in both the water maze and the radial maze. To a limited degree, reference memory was also altered. Grafts placed in the ventral hippocampus had no significant behavioral effect, whereas those placed in the dorsal hippocampus normalized working memory performance in the water maze. Our data show that infusion of 192 IgG-saporin into the septal region deprived the hippocampus of its cholinergic innervation and altered spatial working memory more consistently than spatial reference memory. Although the cholinergic nature of the graft-induced reinnervation remains to be established more clearly, these results further support the idea of a functional dissociation between the dorsal and the ventral hippocampus, the former being preferentially involved in spatial memory.

Fetal hippocampal grafts containing CA3 cells restore host hippocampal glutamate decarboxylase-positive interneuron numbers in a rat model of temporal lobe epilepsy

Degeneration of CA3-pyramidal neurons in hippocampus after intracerebroventricular kainic acid (KA) administration, a model of temporal lobe epilepsy, results in hyperexcitability within both dentate gyrus and the CA1 subfield. It also leads to persistent reductions in hippocampal glutamate decarboxylase (GAD) interneuron numbers without diminution in Nissl-stained interneuron numbers, indicating loss of GAD expression in a majority of interneurons. We hypothesize that enduring loss of GAD expression in hippocampal interneurons after intracerebroventricular KA is attributable to degeneration of their CA3 afferent input; therefore, fetal CA3 grafts can restore GAD interneuron numbers through graft axon reinnervation of the host. We analyzed GAD interneuron density in the adult rat hippocampus at 6 months after KA administration after grafting of fetal mixed hippocampal, CA3 or CA1 cells into the CA3 region at 45 d after lesion, in comparison with "lesion-only" and intact hippocampus. In dentate and CA1 regions of the lesioned hippocampus receiving grafts of either mixed hippocampal or CA3 cells, GAD interneuron density was both significantly greater than lesion-only hippocampus and comparable with the intact hippocampus. In the CA3 region, GAD interneuron density was significantly greater than lesion-only hippocampus but less than the intact hippocampus. Collectively, the overall GAD interneuron density in the lesioned hippocampus receiving either mixed hippocampal or CA3 grafts was restored to that in the intact hippocampus. In contrast, GADinterneuron density in the lesioned hippocampus receiving CA1 grafts remained comparable with lesion-only hippocampus. Thus, grafts containing CA3 cells restore CA3 lesion-induced depletions in hippocampal GAD interneurons, likely by reinnervation of GAD-deficient interneurons. This specific graft-mediated effect is beneficial because reactivation of interneurons could ameliorate both loss of functional inhibition and hyperexcitability in CA3-lesioned hippocampus.

Survival of grafted fetal neural cells in kainic acid lesioned CA3 region of adult hippocampus depends upon cell specificity

We hypothesize that the degree of graft cell survival within the damaged CNS correlates with the specificity of donor cells to the region of grafting. We investigated graft cell survival following transplantation of fetal micrografts into the CA3 region of the adult rat hippocampus at a time-point of 4 days after an intracerebroventricular administration of kainic acid (KA). Grafts consisted of 5'-bromodeoxyuridine (BrdU) labeled embryonic day (E) 19 cells from hippocampal fields CA3 and CA1 and E15 and E19 cells from the striatum. Absolute cell survival in these grafts was quantitatively analyzed at 1 month postgrafting, using BrdU immunostaining of serial sections and three-dimensional reconstruction of grafts. Absolute graft cell survival in lesioned CA3 was dramatically greater for cells having hippocampal origin (CA3 cells, 69% cell survival; CA1 cells, 42% cell survival) than those having nonhippocampal origin, such as striatal cells (E15 cells, 12% cell survival; E19 cells, 4% cell survival). This difference is in sharp contrast to survival of these cells in culture, where E19 cells from both hippocampal and nonhippocampal origins exhibited similar survival. Comparison of survival among hippocampal cell types indicated significantly greater survival for cells that are specific to the lesioned area (i.e., CA3 cells) than for those that are nonspecific to the lesioned area (i.e., CA1 cells). Graft cell survival in the intact CA3 region (contralateral to KA administration), however, did not differ either between cells having hippocampal and nonhippocampal origins or between CA3 and CA1 cells (CA3 cells, 26% cell survival; CA1 cells, 33% cell survival; and E15 striatal cells, 20% cell survival). These results underscore the finding that enhanced survival of fetal cell grafts in the lesioned CNS is critically dependent upon the specificity of donor fetal cells to the region of transplantation. Thus, grafting of cells that are specific to the lesioned area is a prerequisite for achieving maximal graft cell survival and integration in the lesioned host CNS.

Intrahippocampal cholinergic-rich transplants restore lead-induced deficits: a preliminary study in rats

In the present study restorative potential of fetal cholinergic rich cell suspensions in ameliorating cognitive deficits in rats perinatally exposed to lead was studied. Lactating dams with 1-day old litters were given 0.2% (w/v) lead acetate in drinking water throughout lactation from postnatal day (PND) 1 to PND21 at the end of which the treatment was stopped and the animals were weaned. On PND42 lead exposed rats were given bilateral, intrahippocampal, cholinergic rich fetal neural transplants (approximately 60,000 cells per site) and subsequently assessed 3 and 6 months posttransplantation. Control animals (Sham operated and transplanted) were also run in parallel. Lead exposed rats exhibited a decreased learning ability and locomotor activity. A significant decrease in the levels of acetylcholinesterase and sodium potassium ATPase Na+,K+-ATPase activity was observed in hippocampal region of lead exposed rats. The levels of lead were increased by fivefold in the hippocampal region of lead exposed rats. Transplantation showed marginal improvement in the above impairments at 3 months which were more marked at 6 months. Lead levels at 6 months were not significantly higher in lead exposed rats as compared with the control. Results confirm previous findings that fetal neural transplants help in restoring the lost functional deficits and demonstrate their restorative potential in case of lead induced deficits.

Modulation of acetylcholine and 5-hydroxytryptamine release in hippocampal slices of rats with fimbria-fornix lesions and intrahippocampal grafts containing cholinergic and/or serotonergic neurons

Three-month-old Long-Evans female rats sustained aspirative lesions of the dorsal septohippocampal pathways and, 2 weeks later, received intrahippocampal suspension grafts containing fetal cells from the mesencephalic raphe (rich in serotonergic neurons; RAPHE), the medial septum and the diagonal band of Broca (rich in cholinergic neurons; SEPT), or a mixture of both (COTR). Lesion-only (LES) and sham-operated rats (SHAM) were used as controls. Hippocampal slices of these rats (5-9 month after surgery) were preincubated with [3H]choline or [3H]5-HT, superfused continuously (in the presence of hemicholinium-3 or fluvoxamine) and stimulated electrically (360 pulses, 2 ms, 3 Hz, 26-28 mA) in order to study the presynaptic modulation of acetylcholine (ACh) and serotonin (5-HT) release. The accumulation of [3H]choline and the evoked overflow of [3H]ACh were significantly reduced in slices from LES and RAPHE rats, but reached a close-to-normal level in SEPT and COTR rats. As to accumulation and overflow of [3H]5-HT, the lesion-induced reduction was compensated for only in RAPHE and COTR rats. The relative amount of evoked [3H]5-HT release (in % of tissue-3H) was significantly increased in LES and SEPT rats. Only slight differences (group LES) were found in the sensitivity of muscarinic and serotonergic autoreceptors towards oxotremorine and CP 93,129, respectively. Moreover, CP 93,129 induced a significantly weaker inhibition of ACh release in slices of COTR rats than in all other groups. Using the 5-HT1A receptor agonist 8-OH-DPAT and antagonist Way 100,635, no evidence for a modulatory influence of 5-HT1A receptors was found in RAPHE and COTR rats. It is concluded that despite substantial lesion- and graft-induced changes in the amount of ACh and 5-HT released by hippocampal slices of lesion-only or grafted rats, the presynaptic modulation of these transmitters is only slightly affected by changes in the neuronal environment.

Intrahippocampal grafts containing cholinergic and serotonergic fetal neurons ameliorate spatial reference but not working memory in rats with fimbria-fornix/cingular bundle lesions

Three-month-old Long-Evans female rats sustained aspirative lesions of the dorsal septohippocampal pathways and, 2 weeks later, received intrahippocampal suspension grafts containing cells from the mesencephalic raphe, cells from the medial septum and the diagonal band of Broca, or a mixture of both. Lesion-only and sham-operated rats were used as controls. All rats were tested for locomotor activity 1 week, 3 and 5 months after lesion surgery, for spatial working memory in a radial maze from 5 to 9 months, and for reference and working memory in a water tank during the 9th month after lesioning. Determination of hippocampal concentration of acetylcholine, noradrenaline, and serotonin was made after completion of behavioral testing. Compared to sham-operated rats, all rats with lesions, whether grafted or not, exhibited increased levels of locomotor activity and made more errors in the radial maze. The lesioned rats were also impaired in the probe trial (30 first seconds) of the water-tank test made according to a protocol requiring intact reference memory capabilities. While rats with septal or raphe grafts were also impaired, the rats with co-grafts showed performances not significantly different from those of sham-operated rats. With a protocol requiring intact working memory capabilities, all lesioned rats, whether grafted or not, were impaired in the water-tank test. In the dorsal hippocampus of lesion-only rats, the concentration of acetylcholine and serotonin was significantly reduced. In rats with septal grafts or co-grafts, the concentration of acetylcholine was close to normal, as was that of serotonin in rats with raphe grafts or co-grafts. These results confirm previous findings showing that co-grafts enabled the neurochemical properties of single grafts to be combined. Data from the water-tank test suggest that cholinergic and serotonergic hippocampal reinnervations by fetal cell grafts may induce partial recovery of spatial reference, but not working memory capabilities in rats.

Chronic intraventricular infusion with NGF improves LTP in old cognitively-impaired rats

Aged (21 months) cognitively-impaired male Sprague-Dawley rats received intraventricular infusion of nerve growth factor (NGF) or cytochrome C (Cit C) for 14 or 28 days using miniosmotic pumps and were evaluated either 1 week or 3 months after treatment. Groups of untreated young, aged-impaired and aged non-impaired rats were also evaluated. Under narcose recording and stimulating electrodes were stereotactically implanted in the dentate gyrus and the perforant path. The stimulation intensity was individually adjusted to obtain a half-maximal population spike (P) for test stimuli and a quarter-maximal for tetanization. The amplitude and latency of P and the slope (S) of the field EPSP were determined before and at 2, 5, 15, 30 and 60 min after tetanization at 400 Hz. Paired stimuli at 30 ms interval were also applied before and after tetanization. Aged, cognitively impaired rats showed an absent S potentiation and a delayed P potentiation, both in amplitude and latency, while non-impaired rats behaved like the young controls. Paired pulse inhibition showed no difference among groups before or after tetanization suggesting that the impaired potentiation is not due to an increased retroactive inhibition. NGF treatment ameliorates LTP deficits to levels equivalent to non-impaired rats, while Cit C controls showed no improvement. No differences appear among NGF treated groups, but evidence suggest that the animals evaluated 3 months after treatment developed a stronger potentiation.

Modulation of hippocampal acetylcholine release after fimbria-fornix lesions and septal transplantation in rats

Female Long-Evans rats sustained electrolytic lesions of the fimbria and the dorsal fornix causing a partial lesion of the septohippocampal pathway. Two weeks later, the rats received intra-hippocampal grafts of fetal septal cell suspensions. Nine to twelve months later, the release of acetylcholine (ACh) in the hippocampus of sham-operated, lesion-only and grafted rats was measured by microdialysis. The extent of cholinergic (re)innervation was determined by acetylcholinesterase (AChE) staining and densitometry. In both lesion-only and grafted rats, the ratio of ACh release to AChE staining intensity was increased as compared to sham-operated rats, indicating a loss of endogenous inhibitory mechanisms. Scopolamine (0.5 mg/kg i.p.), a muscarinic antagonist, increased ACh release in all treatment groups. 8-OH-DPAT (0.5 mg/kg s.c.), an agonist at serotonergic 5HT1A-receptors, induced an increase of hippocampal ACh release in sham-operated rats. This effect was lost in lesion-only rats, but was fully restored by neuronal grafting. As 8-OH-DPAT influences hippocampal ACh release by a postsynaptic action, this finding indicates that the host brain exerts a serotonergic influence on the grafted cholinergic neurons.

Comparison of the behavioral and morphological effects of colchicine- or neutral fluid-induced destruction of granule cells in the dentate gyrus of the rat

Virtually complete destruction of dentate gyrus granule cells by colchicine injections produced a persistent incapacity to solve spatial problems in rats. A topographically more selective but only locally complete destruction of granule cells using injections of neutral fluid (NFL) impaired acquisition during the initial stages of Morris water maze testing, thus indicating that limited degeneration of granule cells may weakly but significantly alter spatial learning capabilities. A subamnestic dose (0.08 mg/kg ip) of the NMDA antagonist MK-801 worsened radial maze performance only in NFL-treated rats, suggesting that there may be a synergistic interaction between NMDA blockade and limited granule cell degeneration.

The fimbria-fornix/cingular bundle pathways: a review of neurochemical and behavioural approaches using lesions and transplantation techniques

Extensive lesions of the fimbria-fornix pathways and the cingular bundle deprive the hippocampus of a substantial part of its cholinergic, noradrenergic and serotonergic afferents and, among several other behavioural alterations, induce lasting impairment of spatial learning and memory capabilities. After a brief presentation of the neuroanatomical organization of the hippocampus and the connections relevant to the topic of this article, studies which have contributed to characterize the neurochemical and behavioural aspects of the fimbria-fornix lesion "syndrome" with lesion techniques differing by the extent, the location or the specificity of the damage produced, are reviewed. Furthermore, several compensatory changes that may occur as a reaction to hippocampal denervation (sprouting changes in receptor sensitivity and modifications of neurotransmitter turnover in spared fibres) are described and discussed in relation with their capacity (or incapacity) to foster recovery from the lesion-induced deficits. According to this background, experiments using intrahippocampal or "parahippocampal" grafts to substitute for missing cholinergic, noradrenergic or serotonergic afferents are considered according to whether the reported findings concern neurochemical and/or behavioural effects. Taken together, these experiments suggest that appropriately chosen fetal neurons (or other cells such as for instance, genetically-modified fibroblasts) implanted into or close to the denervated hippocampus may substitute, at least partially, for missing hippocampal afferents with a neurochemical specificity that closely depends on the neurochemical identity of the grafted neurons. Thereby, such grafts are able not only to restore some functions as they can be detected locally, namely within the hippocampus, but also to attenuate some of the behavioural (and other types of) disturbances resulting from the lesions. In some respects, also these graft-induced behavioural effects might be considered as occurring with a neurochemically-defined specificity. Nevertheless, if a graft-induced recovery of neurochemical markers in the hippocampus seems to be a prerequisite for also behavioural recovery to be observed, this neurochemical recovery is neither the one and only condition for behavioural effects to be expressed, nor is it the one and only mechanism to account for the latter effects.