SEARCH

SEARCH BY CITATION

Keywords:

  • dementia model;
  • entorhinal cortex;
  • botulinum toxin;
  • learning;
  • memory;
  • BDNF;
  • carnitine;
  • synaptic plasticity

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

A rat dementia model with cognitive deficits was generated by synapse-specific lesions using botulinum neurotoxin (BoNTx) type B in the entorhinal cortex. To detect cognitive deficits, different tasks were needed depending upon the age of the model animals. Impaired learning and memory with lesions were observed in adult rats using the Hebb-Williams maze, AKON-1 maze and a continuous alternation task in T-maze. Cognitive deficits in lesioned aged rats were detected by a continuous alternation and delayed non-matching-to-sample tasks in T-maze. Adenovirus-mediated BDNF gene expression enhanced neuronal plasticity, as revealed by behavioral tests and LTP formation. Chronic administration of carnitine over time pre- and post-lesions seemed to partially ameliorate the cognitive deficits caused by the synaptic lesion. The carnitine-accelerated recovery from synaptic damage was observed by electron microscopy. These results demonstrate that the BoNTx-lesioned rat can be used as a model for dementia and that cognitive deficits can be alleviated in part by BDNF gene transfer or carnitine administration. © 2002 Wiley-Liss, Inc.

Alzheimer's disease is characterized psychiatrically by a progressive cognitive decline and pathologically by an extensive loss of neurons. Although histopathological profiles such as senile plaques and neurofibrillary tangles are considered a prerequisite for the neuropathological diagnosis (Mizutani, 1994), only weak correlations were detected between psychometric indices and plaques or tangles, whereas the synaptic density was shown to correlate strongly with the results of psychological tests (Terry et al., 1991). Thereafter, the concept “synaptic pathology”, in which synaptic loss is the major correlate of cognitive impairment, has been supported by many studies (Lippa et al., 1992; Masliah and Terry, 1993; Scheff and Price, 1993; Heinonen et al., 1995; Blennow et al., 1996; Shimohama et al., 1997; Heffernan et al., 1998; Cambon et al., 2000). The primary events of neuronal perikaryal and synaptic losses were shown to occur in the entorhinal cortex of Alzheimer's patient brains (Mizutani and Kasahara, 1997; Morrison and Hof, 1997; Price et al., 2001). The selective deposition of amyloid-β protein in the entorhinal-dentate projection was observed in a transgenic mouse model of Alzheimer's disease (Su and Ni, 1998). Entorhinal cortex dysfunction was detected in early Alzheimer's disease by positron emission tomography (Eustache et al., 2001).

Based on these backgrounds in the research on Alzheimer's disease, we attempted to generate an animal model of dementia by inducing synaptic-specific impairment in the entorhinal cortex of rats. Botulinum neurotoxin (BoNTx) type B, which specifically degrades synaptobrevin (VAMP) in synapses to quit neurotransmission, was used for this purpose. Bilateral injection of BoNTx produced demented rats whose capacity for learning and memory were assessed by maze tests. Using these rats, we investigated whether brain-derived neurotrophic factor (BDNF) and carnitine could promote functional recovery or neuronal plasticity in the entorhinal-dentate projection. BDNF has been known to enhance neuronal survival and differentiation (Morse et al., 1993; Kirschenbaum and Goldman, 1995; Altschuler et al., 1999; Miyata et al., 2001; Sun et al., 2001). Carnitine derivatives have been shown to protect neuronal derangement after brain ischemia (Calvani and Arrigoni-Martelli, 1999) and to rescue motoneurons from axotomy-induced cell death (Fernandez et al., 1995). Carnitine was shown to improve cholinergic function and learning capacity in aged rats in our previous study (Ando et al., 2001). Thus, we examined the effects of BDNF and carnitine on the entorhinal synaptic damage.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Animals

Male Fischer 344 rats were obtained from an aging colony maintained at the Tokyo Metropolitan Institute of Gerontology. Eight-month-old rats were used for a BDNF gene-transfection experiment (sham-operated control group, NC1; toxin-injected group, TC1; toxin-injected and BDNF-treated group, TB). Twenty-three-month-old rats were used for a carnitine experiment (sham-operated control group, NC2; toxin-injected group, TC2; carnitine-treated and toxin-injected group, TA). Some additional rats (9–26 months old) were used for preliminary experiments in which dose dependency of BoNTx and the procedure for injection of the toxin (unilateral or bilateral injection) were examined. Rats were maintained under 12-hr light/12-hr dark conditions and were fed ad lib autoclaved CRF-1 (Charles River, Atsugi, Japan). Rats of the experimental group (TA) in the carnitine experiment were given drinking water containing acetyl-L-carnitine (ALCAR) so that they received a dose of 100 mg ALCAR/kg body weight per day. ALCAR was given for 2 months before BoNTx injection and 4 months after, until the end of the experiment.

Activation of Botulinum Neurotoxin (BoNTx)

BoNTx was activated by adding of 1 μg of trypsin-TPCK (Worthington Biochemical Co., NJ) to 40 μg of BoNTx type B (Wako, Tokyo, Japan) according to Sathyamoorthy and DasGupta (1985). The reaction was carried out in 60 μl of 100 mM acetate buffer (pH 6.0) for 30 min at 37°C. A 10-fold excess (w/w) of phenylmethylsulfonyl fluoride (Sigma, St. Louis, MO; 0.6 μl of 100 mM ethanol solution) was added to inactivate trypsin, and the mixture was kept for 15 min at 37°C. The inactivation procedure was repeated once again, and the mixture was neutralized with 20 μl of 25 mM Tris-HCl buffer (pH 9.0). The solution of activated toxin was diluted with a solution containing 500 μg rat albumin/ml in 100 mM NaCl, 75 mM acetate buffer (pH 7.3).

Botulinum Neurotoxin Injection

Botulinum toxin injection was carried out under anesthesia with halothane using an infusion syringe pump (Harvard Apparatus Inc., MA). The anesthetized rat was placed in a stereotaxic apparatus, and each 0.5 μl of toxin was injected over 30 min into the right and left entorhinal cortices with the following coordinates: AP = −8.3 mm from the bregma, DV = −5.0 mm from the dura and ML = ±4.5 mm from the midline. Sham-operated control animals were injected with the vehicle without toxin.

Adenovirus-Mediated Transfection of BDNF Gene

The coding region of mouse brain-derived neurotrophic factor (mBDNF) cDNA (GenBank accession number X55573) was amplified from mouse neuroblastoma cell line Neuro2a by reverse transcription-polymerase chain reaction (RT-PCR). The amplified cDNA was subcloned into pcDNA3.1/Myc-His A vector (Invitrogen, San Diego, CA) to fuse a myc-his tag at the end of the mBDNF. The epitope-tagged mBDNFmyc-his fusion protein cDNA was next subcloned into a pIRES-EGFP vector (CLONTECH, Palo Alto, CA). This construct contained the internal ribosome entry site (IRES) from poliovirus (Dirks et al., 1993) and provided dicistronic transcription of mBDNFmyc-his and enhanced-green fluorescence protein (EGFP) cDNA. The replication-defective recombinant adenovirus carrying mBDNFmyc-his-IRES-EGFP cDNA (AxmBDNFME) (Fig. 1) was constructed essentially as described by Miyake et al. (1996). The cDNA containing mBDNFmyc-his-IRES-EGFP was first inserted into a cassette cosmid pAxCAwt that carried an adenovirus type-5 genome lacking E1A, E1B and E3 regions, to prevent virus replication (Kanegae et al., 1995). Recombinant viruses were produced in human embryonic kidney 293 cells (ATCC, Rockville, MD) by calcium phosphate co-transfection of the cosmid with the appropriately cleaved adenovirus genome lacking the E3 region. As a negative control, the viruses carrying just IRES-EGFP cDNA (AxIRES-EGFP) were propagated. Viruses were purified by centrifuging twice in a cesium chloride density gradient and dialyzed against phosphate-buffered saline containing 10% glycerol (Kanegae et al., 1994). Titration of the recombinant AxmBDNFME and AxIRES-EGFP was determined as infectious titer per cell (MOI, multiplicity of infection) and found to be 1.5 × 1010 and 4.5 × 1010 infectious particles/ml, respectively.

thumbnail image

Figure 1. Recombinant adenovirus AxmBDNFME. The mBDNFME-IRES-EGFP cDNA was cloned downstream of the CAG (cytomegalovirus enhancer-chicken β-actin hybrid) promoter in pAxCAwt that carried an adenovirus type-5 genome lacking E1A, E1B (δE1A · E1B) and E3 (δE3) regions to prevent virus replication. A rabbit β-globin poly (A) sequence was located downstream of the EGFP cDNA.

Download figure to PowerPoint

The recombinant viruses were transfected into the entorhinal cortex in the same way that BoNTx was injected.

Learning and Memory Tests

Rats were put on a reduced diet for 1 month before the tests and the average body weight was adjusted to about 300g. Before the tests, the animals were subjected to handling for 2 days and habituated to the apparatus. Each rat was trained to run to the goal and take food pellets.

Non-matching-to-sample T-maze test.

An apparatus of a T-maze was constructed in the same field as used for the Hebb-Williams maze (see below). A stem alley (18 cm × 35 cm) was connected to an arm alley (15 cm × 85 cm) in a T shape. At either entrance of the arms, a Plexiglas barrier was set to force rats to run into one side of the arms. One food cup each was located at both ends of the arms. A rat was first directed at random to the right or left arm in a forced run. After getting a food pellet, the rat was put back into the start box. After a 5-sec or 60-sec delay, the guillotine door was opened, and the rat was allowed to run down the stem and choose either arm, where the Plexiglas barrier was removed. Choosing the arm opposite the side in the forced run was a correct response to get a pellet in the choice run. Because the blocked side of the arm in the forced run was changed at random, this choice was regarded as being based on working memory because of its trial-dependency (Ando and Ohashi, 1991).

Continuous alternation T-maze test.

The apparatus used for the continuous alternation T-maze test was the same as the apparatus used for a non-matching-to-sample T-maze test without food cups. The test was carried out essentially according to the procedure of Gerlai (1998). A rat was first directed to the left arm by blocking the entry to the right arm using a Plexiglas barrier. When the rat returned to the stem after exploring the left arm, the barrier was removed. The rat was allowed to run down the stem and choose either arm. After the rat chose one arm, the opposite arm was blocked by the barrier. Free choices were monitored continuously for 10 min, and alternative rates or correct responses were calculated.

Hebb-Williams maze test and AKON-1 maze test.

Hebb-Williams mazes and AKON-1 maze were constructed in a square field (75 cm × 75 cm × 15 cm high), and tests using those mazes were done as reported previously (Ando et al., 2001). In another study of ours in this volume (Kobayashi et al., 2002), a principal component analysis revealed that four tasks (No. 3, 5, 6, and 12) were controlled by factors that were different from the common major factor for the other eight tasks. Therefore, only the latter eight tasks were employed in this study.

Slice Preparation and Electrophysiological Recording

The method used in the present study was essentially the same as described previously (Furuse et al., 1998). The hippocampus from the rat brain was sectioned rapidly into 400-μm thick slices using a rotary slicer (Dosaka DTY-8700, Kyoto, Japan). The slices were preincubated for at least 1 hr at 30°C under a 95% O2/5% CO2 atmosphere in a standard medium (124 mM NaCl, 5.0 mM KCl, 1.25 mM NaH2PO4, 2.0 mM MgSO4, 2.5 mM CaCl2, 22.0 mM NaHCO3, 10.0 mM glucose). For testing, a slice was placed in a recording chamber and perfused continuously with the standard medium at a rate of 2 ml/min. A bipolar stimulating electrode was placed in the stratum radiatum in the CA1 region to stimulate the Schaffer-collateral/commissural pathways. A recording glass pipette was placed in the pyramidal cell-body layer to monitor the amplitude of the population spike (PS). At the beginning of each experiment, a strength–response curve was established and the stimulus intensity was adjusted so that the amplitude of PS was 50% of maximum. After the stability of the responses to test stimuli given at 20-sec intervals was confirmed, a tetanus stimulus (100 pulses of 100 Hz) was delivered to elicit long-term potentiation (LTP). The test stimuli were reapplied every 20 sec and responses were recorded for 30 min.

Histological Procedure

Two days before being sacrificed, animals to be examined by electron microscopy were injected with 0.3 μl of India ink into the same brain region as the toxin so that the region was cut out correctly. After deep anesthesia, animals were perfused intracardially with Lillie's neutral buffered 10% formalin for light microscopy or with physiological saline followed by Karnovsky's fixative for electron microscopy. The brains were removed and kept in each fixative. For the light microscopic observations, the brains were fixed for more than 7 days, dehydrated with ethanol, and embedded in paraffin. Coronal sections were made from the right halves of brains and horizontal sections were made from the left halves. The sections were stained with hematoxylin and eosin, or cresyl violet.

For electron microscopy, prefixed tissues were postfixed with 2% osmium tetroxide (pH 7.2, 0.1 M cacodylate buffer), dehydrated with a graded ethanol series, and embedded in Epon 812. Ultra-thin sections were cut by a ultramicrotome, picked up on grid meshes, stained with uranyl acetate followed by lead citrate solutions, and observed in a transmission electron microscope (S-7000, Hitachi Co. Ltd.).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

To determine the optimum dose of BoNTx type B for affecting the behavior of rats, three different doses were tested. A high dose (1 × 103 medium lethal dose [MLD] per entorhinal cortex) caused serious clinical signs, such as tremors, convulsions and violent movements. Two of three rats that received the high dose of the toxin died within 1 month. No rats receiving the low dose (40 MLD) suffered from any aberrant behavior or demonstrated seriously impaired learning. On the other hand, rats receiving the medium dose (2 × 102 MLD) did not exhibit much aberrant behavior such as convulsions, but showed some impaired learning in the Hebb-Williams maze test. Based on the preliminary study, the dose of BoNTx was fixed in this study to be 2 × 102 MLD per entorhinal cortex.

To examine the effects of lesioning by the toxin on one or both sides of entorhinal cortices on cognitive function, a medium dose of BoNTx was given to rats unilaterally or bilaterally. A non-matching-to-sample T-maze test revealed that the bilateral lesions severely impaired cognitive function and unilateral lesions only slightly impaired (Fig. 2). Accordingly, bilateral lesions by BoNTx were employed in the following experiments.

thumbnail image

Figure 2. Learning impairment assessed by a non-matching-to-sample T-maze test. Rats (26 months old) were injected with BoNTx unilaterally (•) or bilaterally (▴) into the entorhinal cortices. Open circles indicate the result of sham-operated control rats. Rats were subjected to the T-maze test 1 week after the operation.

Download figure to PowerPoint

BoNTx (2 × 102 MLD) was bilaterally injected into the entorhinal cortices of adult (8-month-old) rats. One week later, recombinant AxmBDNFME was injected bilaterally into the same brain region of 50% of the BoNTx-treated rats. Subsequently, the rats were subjected to cognitive tests at various intervals. A set of two different problems of the Hebb-Williams maze was employed for each test: a set of Problems No. 7 and 8 at 1 week after BDNF gene transfection, a set of Problems No. 4 and 9 at 4 weeks after BDNF, and a set of Problems No. 10 and 11 at 8 weeks after BDNF (Table I). The toxin groups showed impaired learning whether or not BDNF gene was transfected. The continuous alternation T-maze test revealed impaired short-term memory in the toxin groups with or without BDNF gene transfection (Table II). The AKON-1 maze test discriminated between the performance of all three groups. As shown in the 3-D bar graphs (Fig. 3), rats in the toxin group (TC1) that were not transfected with the BDNF gene did not learn to any extent, but some rats in the BDNF-treated group (TB) visited alleys other than Alley A. Three rats in the TB group found Alley F, which connects to the goal, and three rats visited dead-end alleys except Alley A. One rat in the control group found Alley F and others visited dead-end alleys, except Alley A, 12 times in five trials.

Table I. Total Errors in the Hebb-Williams Maze Test
 Time after BDNF gene treatment
1 week4 weeks8 weeks
  • *

    P < 0.05 significantly different from NC1.

  • **

    P < 0.001 significantly different from NC1.

Hebb-Williams problemsNos. 7, 8Nos. 4, 9Nos. 10, 11
 NC1 group46 ± 818 ± 721 ± 8
 TC1 group77 ± 39*58 ± 7**63 ± 17**
 TB group74 ± 28*108 ± 37**69 ± 15**
Table II. Percent Correct Responses in Continuous Alternation T-Maze Task
ExperimentWeeks after BoNTx injection
259
  • *

    P < 0.05 significantly different from NC1.

  • **

    P < 0.01 significantly different from NC1.

BDNF gene transfer experiment   
 Weeks after BDNF gene transfection259
 NC1 group71 ± 976 ± 2167 ± 18
 TC1 group61 ± 7*38 ± 18**55 ± 22
 TB group59 ± 2242 ± 23**51 ± 23
Carnitine experiment   
 NC2 group81 ± 3192 ± 2095 ± 11
 TC2 group63 ± 2459 ± 2668 ± 33
 TA group64 ± 2352 ± 1672 ± 23
thumbnail image

Figure 3. Learning impairment in BoNTx-lesioned rats. Rats (8 months old) were injected bilaterally with BoNTx into the entorhinal cortices and 1 week later, 50% of them were transfected with the BDNF gene A: Sham-operated control rats (NC1). B: BoNTx-lesioned rats (TC1). C: Partial recovery from the impairment in the rats transfected with the BDNF gene (TB). Rats were subjected to the AKON-1 task (Ando et al., 2001) 3 months after the operation.

Download figure to PowerPoint

In the next experiment, the effect of acetyl-L-carnitine (ALCAR) on the recovery from BoNTx lesion was examined. Rats (23 months old) were given ALCAR for 2 months, then BoNTx (2 × 102 MLD per one side) was injected bilaterally into their entorhinal cortices. Thereafter, cognitive function of the rats was monitored over time. In the Hebb-Williams maze test, the error score of the control group was too high to be differentiated from the scores of the two groups with toxin-lesions (data not shown). The continuous alternation T-maze test showed impaired memory function in the toxin groups whether or not ALCAR was given, but it did not discriminate between the two toxin groups based on the percent correct responses (Table II). The non-matching-to-sample T-maze test showed about the same level of memory acquisition in all groups (see correct responses at Day 3 in Fig. 4). When a 60-sec delay was inserted between a forced run and a consecutive test run, the percentage of correct responses declined (on the 4th day in Fig. 4). The degree of decline was larger in the toxin group of rats (TC2) that were not given ALCAR than in controls and the other toxin group of rats (TA) given ALCAR. After the trial in which a 60-sec delay was inserted, the level of the re-acquisition of the task was significantly lower in the TC2 group than in the control and TA groups (from Day 5–8 in Fig. 4).

thumbnail image

Figure 4. Working memory impairment by BoNTx-lesion was restored in the rats given acetyl-L-carnitine (ALCAR). Three groups of rats (○: sham-operated control, NC2; •: BoNTx-lesioned group, TC2; ▴: BoNTx-lesioned group given ALCAR, TA) attained similar levels of acquisition in non-matching-to-sample T-maze tests with a 5-sec delay during the first 3 days. A non-matching-to-sample T-maze test with a 60-sec delay was inserted at the fourth day, and recovery in working memory was monitored in consecutive tests with a 5-sec delay for 4 days. Working memory was recovered in TA, whereas it was retarded in TC2. *Significant difference between TC2 and NC2 or TA (P < 0.05).

Download figure to PowerPoint

LTP formation was examined in the rats of the NC1 and TB groups at the age of 26 months (18 months post surgery) (Fig. 5). LTP formation was reduced in TC1 at about 60% of that in the control. The reduction was shown to be partially recovered in TB.

thumbnail image

Figure 5. Long-term potentiation (LTP) observed in population spikes (PS). Responding to the tetanus stimulation of the Schaffer-collateral/commissural pathways, LTP was observed at CA1 neurons. The amplitude of PS was reduced in the BoNTx-lesioned group (TC1, n = 3, •), and it increased in BoNTx-lesioned group treated with BDNF-gene (TB, n = 5, ▴). Open circles indicate sham-operated control group (NC1, n = 5). Values are means + SEM.

Download figure to PowerPoint

The brains injected with BoNTx were examined by light and electron microscopy. Under the light microscope, no changes were observed with neuronal cell bodies in the entorhinal cortices or hippocampal pyramidal cell layers in any of the toxin groups (data not shown). Moreover, no ultrastructural changes were seen in the neuronal cell bodies, axons or dendrites (data not shown). Under electron microscopy, morphological changes in synapses were observed in the BoNTx-injected entorhinal cortices. Synapses examined 2 days after BoNTx injection were remarkably enlarged and were full of synaptic vesicles, probably due to the cessation of neurotransmission (data not shown). Four months after the toxin injection, the acutely enlarged synapses were replaced with synapses of irregular size (Fig. 6B). The distribution of synaptic vesicles in the synapses was heterogeneous, i.e., synaptic vesicles were sparse or distributed only on the side of the active zones. In contrast, synapses developed well and were full of synaptic vesicles in the toxin-injected brain of rats given ALCAR (Fig. 6C). A prominent characteristic of the brain was that axospinous synapses had abundant perforated synaptic contacts. Transfection of the BDNF gene failed to show any morphological changes of synapses (data not shown).

thumbnail image

Figure 6. Electron micrographs of synapses in the second layer of the entorhinal cortices of 27-month-old rats. A: Sham-operated control shows that synapses are uniform in size and synaptic vesicles are sparse in number. B: Toxin-injected group TC2 shows that synaptic vesicles are sparse or heterogeneously distributed in synapses. C: Carnitine-treated and toxin-injected group TA shows that there are large synapses full of synaptic vesicles and many perforated synaptic contacts.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

According to the concept “synaptic pathology,” which is based on the observation that synaptic loss is correlated with the results of psychological tests in dementia patients (Terry et al., 1991), synapse-specific lesions were employed to produce an animal model for dementia (see review by Yamada and Nabeshima, 2000). Based on different ideas, various demented animals have been generated by many researchers. Infusion of neurotoxic β-amyloid protein (Nitta et al., 1994) and transgenic mice expressing human amyloid precursor protein (Moran et al., 1995; Price et al., 1998; Su and Ni, 1998) were approaches investigating the pathogenesis of Alzheimer's disease. A unique genetic model lacking L-isoaspartyl methyltransferase, which is responsible for converting the isomerization of aspartic residues in β-amyloid protein, was generated to elucidate the pathogenesis of sporadic senile dementia of Alzheimer's type (Ikegaya et al., 2001). Excitotoxic reagents, cholinergic neurotoxin AF64A (Walsh et al., 1984) and glutamatergic toxin ibotenic acid (Hortnagl and Hanin, 1994), and cholinergic immunotoxin 192IgG-saporin (Roßner, 1997) were used to destroy specific neurons. Transection of nerve connections and electrolytic lesions of brain tissues have been used as devastating surgical methods to destroy brain functions. Among the dementia models reported in the literature, synapse-specific lesions have not been employed to develop dementia. The dementia models caused by synapse-restricted damage in this study are potentially useful for testing the recovery of neuronal functions, because the injured neuronal cells may survive and regenerate novel synaptic contacts.

Botulinum neurotoxin (BoNTx) has been used as a pharmacological denervation agent, and since 1989 it has been used as a therapeutic tool for alleviating abnormal muscle contractions in dystonia, because local injection of the toxin produces sustained blockade of neuromuscular transmission for months (Jankovic, 1998; Munchau and Bhatia, 2000). Because BoNTx type B was found to block neurotransmitter release by proteolytic cleavage of synaptobrevin (Schiavo et al., 1992) and synaptobrevin has been shown to be involved in the release mechanism (Almeida et al., 1997), BoNTx is recognized as a new tool in neurobiology, with a well-understood molecular mechanism, for the specific impairment of synaptic transmission. Although tetanus toxin has been used to produce epileptic models in some studies (George and Mellanby, 1982; Lee et al., 1995), this kind of neurotoxin had not been used to produce dementia models. We have attempted to develop an impaired cognitive model by injecting BoNTx-B into a specific region of the brain.

The entorhinal cortex is known as a major source of neocortical and subcortical input to the hippocampus, and lesions of the key cortex are associated with cognitive deficits. The primary events of neuronal and synaptic losses were shown to occur in the entorhinal cortex of brains from Alzheimer's patients (Mizutani and Kasahara, 1997; Morrison and Hof, 1997; Eustache et al., 2001; Price et al., 2001). The entorhinal cortex was shown also to be the region of the brain that is first affected by neurofibrillary tangles in the patient brains (Delacourte et al., 1999). A PET study revealed that the cerebral metabolic rate of glucose, a marker of the functional activity of synapses, declined in the left entorhinal cortex in early Alzheimer's disease and that the metabolic decline of synapses might underlie memory impairment (Eustache et al., 2001). Thus, we chose the entorhinal cortex as the region to be injected with BoNTx to mimic the pathology of Alzheimer's disease.

Adult rats that were injected with BoNTx showed impaired learning in both the Hebb-Williams maze and AKON-1 maze tests (Table I, Fig. 3) and showed impaired short-term memory in the continuous alternation T-maze test (Table II). Aged rats that were given BoNTx showed memory impairment in the continuous alternation T-maze test (Table II), but the results were not significantly different from those of the aged control in the Hebb-Williams maze test. The failure to find a difference might be due to the poor maze performance of the aged rats. The same failure was anticipated for the AKON-1 maze test. Because AKON-1 was too difficult for naive aged rats to solve the problems, it was not used for the experiments with aged rats. Impaired cognitive function produced by BoNTx was observed for more than 4 months in both the adult and aged groups. This seems to be compatible with the observation that the blockade of muscle contraction by BoNTx was sustained for months (Jankovic, 1998; Munchau and Bhatia, 2000).

Although the entorhinal area is shown to be affected early and severely by Alzheimer's disease (Mizutani and Kasahara, 1997; Morrison and Hof, 1997; Eustache et al., 2001; Price et al., 2001), neuronal plasticity or synaptogenesis is reported to be greater in the entorhinal cortex than in other areas of the cortex (Scheff et al., 1993). In experimental pathology, entorhinal cortex lesion in rats is a well-known model of synaptic plasticity (Cotman and Nieto-Sampedro, 1984). Post-lesion neuronal plasticity is as likely to occur in aged rats as in young adult rats (Kugler et al., 1993). Therefore, we attempted to repair the BoNTx-induced damage to the entorhinal cortex by means of BDNF or carnitine. BDNF is known to promote the survival of neurons (Kirschenbaum and Goldman, 1995) and to prevent post-lesion neuronal degeneration (Morse et al., 1993; Kiprianova et al., 1999). The present study showed that post-lesion gene transfer of BDNF was partially effective for restoring the deficits in learning capacity as evidenced by performance in the AKON-1 maze test, and synaptic plasticity as LTP.

The protective action of carnitine for neuronal damage has been demonstrated in brain ischemia animal models (Shuaib et al., 1995; Calvani and Arrigoni-Martelli, 1999). The clinical effects of carnitine on Alzheimer's disease were also examined. Normalization of high-energy phosphate levels was observed in carnitine-treated patients (Pettegrew et al., 1995). Carnitine treatment for early-onset Alzheimer's patients was reported to reduce deterioration as measured by the mini-mental state examination (Thal et al., 2000). In the present entorhinal lesion model, carnitine was shown to partially ameliorate the cognitive deficit in aged rats, as evidenced by the non-matching-to-sample T-maze test. The ameliorating effect of carnitine is in accordance with our previous observation (Ando et al., 2001). It has been demonstrated that delayed non-matching-to-sample T-maze task performance is affected by entorhinal lesions (Wiig and Bilkey, 1994; Baxter and Murray, 2001). A detailed study on the role of the entorhinal cortex indicated that entorhinal lesions did not affect standard spatial reference memory in a water maze but disrupted non-matching-to-sample performance, probably due to impaired attentional mechanisms (Bannerman et al., 2001).

Electron microscopic examination revealed that synapses were well developed in the entorhinal cortex of carnitine-treated and toxin-injected rats. Enlarged synapses full of synaptic vesicles and perforated synaptic contacts were observed frequently (Fig. 6C). The appearance of perforated synapses may indicate that the synapses are well developed (Geinisman, 1993; Desmond and Weinberg, 1998; Neuhoff et al., 1999). This finding seems to be consistent with the observation that chronic administration of acetyl-L-carnitine increased both the numerical density and surface density of synapses (Bertoni-Freddari et al., 1994). Thus, both BDNF and carnitine seem to enhance plasticity responses against neuronal damages. BDNF seems to be applicable to the treatment of degenerative disease, and carnitine can be used to enhance neuronal plasticity and thereby delay the onset of degenerative diseases. Because carnitine and its derivatives have been tested for anti-aging effects in many studies (Sershen et al., 1991; Bertoni-Freddari et al., 1994; Taglialatela et al., 1994; Caprioli et al., 1995; Katz et al., 1997; Aureli et al., 2000; Ando et al., 2001; Kaur et al., 2001; Lohninger et al., 2001; Liu et al., 2002), it is expected that carnitine will have an expanding role in preventive medicine.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

This work was supported by the fund for Comprehensive Research on Senile Dementia from the Tokyo Metropolitan Government (1999–2001).

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES