Differential Neurotoxicity Related to Tetracycline Transactivator and TDP-43 Expression in Conditional TDP-43 Mouse Model of Frontotemporal Lobar Degeneration.

Frontotemporal lobar degeneration (FTLD) is among the most prevalent dementias of early-onset. Pathologically, FTLD presents with tauopathy or TAR DNA-binding protein 43 (TDP-43) proteinopathy. A biallelic mouse model of FTLD was produced on a mix FVB/129SVE background overexpressing wild-type human TDP-43 (hTDP-43) using tetracycline transactivator (tTA), a system widely used in mouse models of neurological disorders. tTA activates hTDP-43, which is placed downstream of the tetracycline response element. The original study on this transgenic mouse found hippocampal degeneration following hTDP-43 expression, but did not account for independent effects of tTA protein. Here, we initially analyzed the neurotoxic effects of tTA in postweaning age mice of either sex using immunostaining and area measurements of select brain regions. We observed tTA-dependent toxicity selectively in the hippocampus affecting the dentate gyrus significantly more than CA fields, whereas hTDP-43-dependent toxicity in bigenic mice occurred in most other cortical regions. Atrophy was associated with inflammation, activation of caspase-3, and loss of neurons. The atrophy associated with tTA expression was rescuable by the tetracycline analog, doxycycline, in the diet. MRI studies corroborated the patterns of atrophy. tTA-induced degeneration was strain-dependent and was rescued by moving the transgene onto a congenic C57BL/6 background. Despite significant hippocampal atrophy, behavioral tests in bigenic mice revealed no hippocampally mediated memory impairment. Significant atrophy in most cortical areas due solely to TDP-43 expression indicates that this mouse model remains useful for providing critical insight into co-occurrence of TDP-43 pathology, neurodegeneration, and behavioral deficits in FTLD.SIGNIFICANCE STATEMENT The tTA expression system has been widely used in mice to model neurological disorders. The technique allows investigators to reversibly turn on or off disease causing genes. Here, we report on a mouse model that overexpresses human TDP-43 using tTA and attempt to recapitulate features of TDP-43 pathology present in human FTLD. The tTA expression system is problematic, resulting in dramatic degeneration of the hippocampus. Thus, our study adds a note of caution for the use of the tTA system. However, because FTLD is primarily characterized by cortical degeneration and our mouse model shows significant atrophy in most cortical areas due to human TDP-43 overexpression, our animal model remains useful for providing critical insight on this human disease.

Prominent microglial activation in cortical white matter is selectively associated with cortical atrophy in primary progressive aphasia.

AIMS: Primary progressive aphasia (PPA) is a clinical syndrome characterized by selective language impairments associated with focal cortical atrophy favouring the language dominant hemisphere. PPA is associated with Alzheimer's disease (AD), frontotemporal lobar degeneration (FTLD) and significant accumulation of activated microglia. Activated microglia can initiate an inflammatory cascade that may contribute to neurodegeneration, but their quantitative distribution in cortical white matter and their relationship with cortical atrophy remain unknown. We investigated white matter activated microglia and their association with grey matter atrophy in 10 PPA cases with either AD or FTLD-TDP pathology. METHODS: Activated microglia were quantified with optical density measures of HLA-DR immunoreactivity in two regions with peak cortical atrophy, and one nonatrophied region within the language dominant hemisphere of each PPA case. Nonatrophied contralateral homologues of the language dominant regions were examined for hemispheric asymmetry. RESULTS: Qualitatively, greater densities of activated microglia were observed in cortical white matter when compared to grey matter. Quantitative analyses revealed significantly greater densities of activated microglia in the white matter of atrophied regions compared to nonatrophied regions in the language dominant hemisphere (P < 0.05). Atrophied regions of the language dominant hemisphere also showed significantly more activated microglia compared to contralateral homologues (P < 0.05). CONCLUSIONS: White matter activated microglia accumulate more in atrophied regions in the language dominant hemisphere of PPA. While microglial activation may constitute a response to neurodegenerative processes in white matter, the resultant inflammatory processes may also exacerbate disease progression and contribute to cortical atrophy.

Aging renders the brain vulnerable to amyloid beta-protein neurotoxicity.

The formation of fibrillar deposits of amyloid beta protein (Abeta) in the brain is a pathological hallmark of Alzheimer's disease (AD). A central question is whether Abeta plays a direct role in the neurodegenerative process in AD. The involvement of Abeta in the neurodegenerative process is suggested by the neurotoxicity of the fibrillar form of Abeta in vitro. However, mice transgenic for the Abeta precursor protein that develop amyloid deposits in the brain do not show the degree of neuronal loss or tau phosphorylation found in AD. Here we show that microinjection of plaque-equivalent concentrations of fibrillar, but not soluble, Abeta in the aged rhesus monkey cerebral cortex results in profound neuronal loss, tau phosphorylation and microglial proliferation. Fibrillar Abeta at plaque-equivalent concentrations is not toxic in the young adult rhesus brain. Abeta toxicity in vivo is also highly species-specific; toxicity is greater in aged rhesus monkeys than in aged marmoset monkeys, and is not significant in aged rats. These results suggest that Abeta neurotoxicity in vivo is a pathological response of the aging brain, which is most pronounced in higher order primates. Thus, longevity may contribute to the unique susceptibility of humans to Alzheimer's disease by rendering the brain vulnerable to Abeta neurotoxicity.

Amyloid beta peptide-induced cholinergic fibres loss in the cerebral cortex of the rat is modified by diet high in lipids and by age.

The influence of diet and age on the effects of intracerebral injection of beta-amyloid peptide (Abeta1-40) in vehicle phosphate-buffered saline (PBS) and on the effects of vehicle alone on cholinergic fibres of the cerebral cortex was studied in rats. The experiments were carried in two groups of animals: one group of young adult rats and a second group of aged rats. Each group of animals, depending on the diet received, was divided into high-cholesterol, high-fat, and a control diet group. In order to evaluate the interaction of Abeta/PBS-cholesterol and of Abeta/PBS-fat, animals without dietary manipulation receiving Abeta and PBS injection were used as controls. High-cholesterol fed animals showed a statistically significant reduction of 49.62% in the number of cholinergic fibres at the Abeta injection site as compared with that at PBS injection site, while the high-fat and control animals showed a significant reduction of 28.13 and 26.81%, respectively. In all diet groups, the loss of cholinergic fibres caused by Abeta as compared to that caused by PBS injection was significantly greater in aged rats in comparison with that observed in the young animals. Furthermore, the results of a multivariate linear regression model revealed that the greatest reduction in cholinergic fibres was in the high-cholesterol fed animals (35 fibres/mm) as compared with that seen in the high-fat and control animals. A significantly greater reduction was also observed at Abeta injection site (28 fibres/mm) as compared with that caused by PBS injection, and a reduction of 16 cholinergic fibres per mm was found in aged animals as compared to that seen in young adult rats. These results show that high-cholesterol diet enhances the toxicity of Abeta peptide and that this is also age-dependent. Therefore, this study increases the evidences of the role of cholesterol in the pathology of Alzheimer's disease (AD).

Accumulation and age-related elevation of amyloid-ß within basal forebrain cholinergic neurons in the rhesus monkey.

Basal forebrain cholinergic neurons (BFCN) are selectively vulnerable to damage and loss in a number of neurodegenerative disorders that afflict the elderly, particularly Alzheimer's disease. The reasons for this selective vulnerability remain poorly understood. Given that intraneuronal accumulation of the amyloid-ß peptide (Aß) has been shown to exert deleterious effects on neurons, we tested potential accumulation of Aß within BFCN in rhesus monkeys, which like the human display age-related accumulation of this peptide in plaques. The non-isoform-specific Aß antibodies 1282 and 6E10 and the specific antibodies to 1-40 amino acid isoform of Aß (Aß1-40) and 1-42 amino acid isoform of Aß (Aß1-42) species were used in immunohistochemical experiments of basal forebrain in young and aged rhesus monkeys. All four antibodies visualized cortical plaques in the same sections in which BFCN were examined, in aged but not in young animals. The basal forebrain region within which the BFCN are localized was virtually free of plaques. Appreciable Aß immunoreactivity was present within the nucleus basalis of Meynert-cholinergic cell group 4 (nbM-Ch4), the major component of BFCN, with all antibodies used. Quantitation of optical density indicated significant age-related increases in immunoreactivity in nbM-Ch4 neurons with the Aß1-40 (p<0.002) and 1282 (p<0.03) antibodies. Immunoreactivity for 6E10 displayed a small, non-significant age-related increase in nbM-Ch4 neurons (p>0.05). No age-related changes were detected in Aß1-42 immunoreactivity in these neurons. Unlike the BFCN, cortical neurons within the same sections were virtually devoid of Aß immunoreactivity, particularly with isoform-specific antibodies. Both smooth and granular intraneuronal Aß immunoreactivity, reminiscent of endosomal/lysosomal packaged peptide, were observed within nbM-Ch4 neurons. In some nbM-Ch4 neurons, 1282 immunoreactivity had the appearance of large peptide aggregates. Significant accumulation and age-related increase of Aß in BFCN is likely to interfere with the normal functioning of these neurons. It remains to be determined if similar accumulation of Aß occurs in human BFCN.

Relationship between plaques, tangles, and loss of cortical cholinergic fibers in Alzheimer disease.

Recent observations in our laboratory have indicated substantial and systematic regional variations in the loss of cortical cholinergic fibers in Alzheimer disease (AD). Previous attempts to study the relationship between cortical cholinergic loss and the density of cortical pathological lesions have resulted in conflicting findings. Furthermore, most reports have correlated density of plaques and tangles with the residual level of cholinergic innervation rather than its loss. The purpose of the present study was to determine the relationship between loss of cholinergic axons and density of tangles and beta-amyloid (Abeta) deposits in various cortical areas of AD brains. Abeta deposits and tangles were observed throughout the cerebral cortex. Quantitative analysis revealed almost no correlation between loss of cholinergic fibers and the density of Abeta deposits. Qualitative observations revealed similar results when cored and neuritic plaques were considered separately. By contrast, cholinergic fiber loss displayed a significant correlation with the density of tangles (r = 0.52-0.79). However, in a few areas, such as the cingulate cortex, tangle density appeared to be unrelated to the loss of cholinergic fibers. These results indicate that cortical cholinergic denervation in AD is related to cytoskeletal pathology. However, the lack of a perfect relationship with cytoskeletal pathology implicates additional factors in the cholinergic pathology of AD.

Neurofibrillary tangles have no obligatory predilection for acetylcholinesterase-rich neurons.

Parts of the brain that are prone to NFT formation normally contain many neurons that are intensely acetylcholinesterase (AChE)-positive. In this study, we used thioflavin-S immunofluorescence, AChE histochemistry, and AChE immunocytochemistry to investigate the possibility that intense AChE positivity may act as a perikaryal marker for the vulnerability to NFT formation. Our observations in entorhinal and motor cortices and in the subthalamic nucleus demonstrate major mismatches between the distribution of AChE-rich neurons in normal brains and the distribution of NFT in AD. There is therefore no obligatory relationship between intense AChE positivity in the premorbid period and subsequent vulnerability to tangle formation.

Acetylcholinesterase-rich pyramidal neurons in Alzheimer's disease.

The distribution of acetylcholinesterase (AChE)-rich pyramidal neurons was studied in the cortices of 7 Alzheimer's Disease (AD) patients and 4 normal-aged subjects. Both groups showed a characteristic distribution of these neurons with the highest density in motor and premotor areas, moderate density in association cortices, and low density in limbic-paralimbic areas. Three areas (Brodmann areas 6,22, and 24) were chosen for quantitative analysis. The number of pyramidal neurons that display an AChE-rich staining pattern was significantly reduced in AD patients. Nerve cell density was not significantly different in adjacent Nissl-stained sections. The density of AChE-rich (cholinergic) fibers was also decreased in all three cortical areas of the AD patients but was not correlated with the number of AChE-rich neurons. Loss of AChE-rich neurons was more pronounced in areas with high counts of tangles. These findings show that layer 3 and 5 pyramidal neurons in AD display a reduction of AChE activity. This phenomenon can not be attributed to the well known loss of cortical neurons or cholinergic innervation in AD.

The acute neurotoxicity and effects upon cholinergic axons of intracerebrally injected beta-amyloid in the rat brain.

The acute neurotoxicity and effects upon cholinergic axons of an intracerebrally injected synthetic peptide corresponding to the first 1-40 amino acids of beta-amyloid protein (beta AP1-40) was studied in rats. A synthetic peptide with the reverse sequence (beta AP40-1) or the vehicle alone were injected in the contralateral hemisphere as control. The size of the resulting lesions was quantified in serial sections using an image analyzer. Counts of cholinergic and noradrenergic fibers were also obtained around the lesion area. The results revealed that beta AP1-40 was significantly more toxic than both reverse peptide and the vehicle. The latter two, however, also caused considerable neurotoxicity. beta AP1-40 was toxic to both cholinergic and noradrenergic fibers to the same extent, and this toxicity was limited to the immediate vicinity of the lesion. This study confirms and extends the results of previous studies reporting neurotoxic effects of intracerebrally injected beta-amyloid in the rat. Our results also show that beta AP1-40 itself is not the source of the altered acetylcholinesterase enzyme activity that has been described in the plaques and tangles of Alzheimer's disease.

Selective cell loss in Edinger-Westphal in asymptomatic elders and Alzheimer's patients.

Exaggerated pupillary response to a low concentration of cholinergic antagonists has been suggested as an early marker for Alzheimer's Disease (AD). To examine the anatomic basis of this phenomenon, we determined possible neuropathological changes in the Edinger-Westphal (EW) nucleus, a midbrain neural center with a significant functional role in the control of pupil size. Stereologically determined neuronal numbers within the EW were counted in individuals with pathologically confirmed AD, control cases with no AD-type pathology, and subjects with AD pathology not meeting diagnostic criteria for AD. The EW of AD patients displayed a marked and striking neuronal loss when compared with controls. In contrast, the number of neurons in the somatic portion of the nucleus of the third cranial nerve (NCNIII) remained intact. The EW in brains from clinically normal individuals with evidence of early AD-type pathology also displayed a significant and selective loss of neurons. The magnitude of EW neuronal loss in the latter group was smaller than that observed in AD. These findings suggest that pupillary hypersensitivity in AD may be caused by abnormalities in the EW. Neuronal loss and pathology within the EW in a subpopulation of clinically silent controls with pathologic findings consistent with early-stage AD constitutes a possible explanation for the reported exaggerated pupil response in some normal elderly subjects.

Acetylcholinesterase-rich neurons of the human cerebral cortex: cytoarchitectonic and ontogenetic patterns of distribution.

Layers 3 and 5 of the adult human cerebral cortex contain a very large number of pyramidal neurons that express intense acetylcholinesterase (AChE) enzymatic activity and AChE-like immunoreactivity. The density of these neurons is high in motor, premotor, and neocortical association areas but quite low in paralimbic cortex. These AChE-rich neurons are located predominantly within layer 3 in the premotor and association cortex, within layer 5 in the non-isocortical components of the paralimbic cortex, and are equally prominent in layers 3 and 5 in the motor cortex. Almost all Betz cells in the motor cortex and up to 80% of layer 3 pyramidal neurons in some parts of the association neocortex yield an AChE-rich staining pattern. The existence of a specific laminar and cytoarchitectonic distribution suggests that the AChE-rich enzymatic pattern of these neurons is selectively regulated. The AChE-rich enzymatic reactivity of the layer 3 and layer 5 neurons is not detectable during early childhood, becomes fully established during adulthood, and does not show signs of decline during advanced senescence in mentally intact individuals. The AChE activity (or enzyme synthesis) in these neurons is therefore held in check for several years during infancy and childhood and begins to be expressed at a time when the more advanced motor and cognitive skills are also being acquired. The absence of immunostaining with an antibody to choline acetyltransferase suggests that these AChE-rich neurons are not cholinergic. The regional distribution of these AChE-rich neurons does not parallel the regional variations of cortical cholinergic innervation. Whereas the AChE-rich pyramidal neurons of layers 3 and 5 almost certainly represent one subgroup of cholinoceptive cortical neurons, their AChE-rich enzymatic pattern is probably also related to a host of non-cholinergic processes that may include maturational changes and plasticity in the adult brain.

Nucleus basalis (Ch4) and cortical cholinergic innervation in the human brain: observations based on the distribution of acetylcholinesterase and choline acetyltransferase.

The nucleus basalis (NB) of the human brain is a large, complex, and highly differentiated structure. Many of its neurons are magnocellular, hyperchromic, isodendritic, acetylcholinesterase-rich, and choline-acetyltransferase-positive. Concurrent histochemical and immunological staining demonstrated that all choline-acetyltransferase-positive NB neurons in the human brain also contain acetylcholinesterase enzyme activity. Only a small minority of acetylcholinesterase-rich magnocellular cell bodies in the NB failed to show choline acetyltransferase immunoreactivity. Sections that were counterstained for Nissl substance showed that 80-90% of all magnocellular neurons in the NB were choline-acetyltransferase-positive and therefore cholinergic. These characteristics, which are very similar to those of the NB in the monkey brain, justified the designation of these cholinergic neurons in the human brain as the Ch4 (or NB-Ch4) complex. On morphological grounds, the compact parts of the human NB-Ch4 complex were divided into distinct sectors which appeared to show a greater level of differentiation than in the monkey brain. In addition to the compact sectors, interstitial elements of NB Ch4 were embedded within adjacent fiber bundles. The putative cortical projections from NB-Ch4 were identified in the form of acetylcholinesterase-rich fibers. These fibers formed a dense plexus in all cortical regions but also displayed laminar and regional variations. Limbic and paralimbic areas had higher concentrations of these fibers than the immediately adjacent neocortical association areas. Alzheimer's disease was associated with a marked depletion of cortical acetylcholinesterase. Two cases of Alzheimer's disease with relatively selective NB-Ch4 cell loss supported the hypothesis that the corticopetal cholinergic pathways in the human brain may have a topographical organization similar to that in the monkey brain.

Cytoarchitecture and neural afferents of orbitofrontal cortex in the brain of the monkey.

The orbitofrontal cortex of the monkey can be subdivided into a caudal agranular sector, a transitional dysgranular sector, and an anterior granular sector. The neural input into these sectors was investigated with the help of large horseradish peroxidase injections that covered the different sectors of orbitofrontal cortex. The distribution of retrograde labeling showed that the majority of the cortical projections to orbitofrontal cortex arises from a restricted set of telencephalic sources, which include prefrontal cortex, lateral, and inferomedial temporal cortex, the temporal pole, cingulate gyrus, insula, entorhinal cortex, hippocampus, amygdala, and claustrum. The posterior portion of the orbitofrontal cortex receives additional input from the piriform cortex and the anterolateral portion from gustatory, somatosensory, and premotor areas. Thalamic projections to the orbitofrontal cortex arise from midline and intralaminar nuclei, from the anteromedial nucleus, the medial dorsal nucleus, and the pulvinar nucleus. Orbitofrontal cortex also receives projections from the hypothalamus, nucleus basalis, ventral tegmental area, the raphe nuclei, the nucleus locus coeruleus, and scattered neurons of the pontomesencephalic tegmentum. The non-isocortical (agranular-dysgranular) sectors of orbitofrontal cortex receive more intense projections from the non-isocortical sectors of paralimbic areas, the hippocampus, amygdala, and midline thalamic nuclei, whereas the isocortical (granular) sector receives more intense projections from the dorsolateral prefrontal area, the granular insula, granular temporopolar cortex, posterolateral temporal cortex, and from the medial dorsal and pulvinar thalamic nuclei. Retrograde labeling within cingulate, entorhinal, and hippocampal cortices was most pronounced when the injection site extended medially into the dysgranular paraolfactory cortex of the gyrus rectus, an area that can be conceptualized as an orbitofrontal extension of the cingulate complex. These observations demonstrate that the orbitofrontal cortex has cytoarchitectonically organized projections and that it provides a convergence zone for afferents from heteromodal association and limbic areas. The diverse connections of orbitofrontal cortex are in keeping with the participation of this region in visceral, gustatory, and olfactory functions and with its importance in memory, motivation, and epileptogenesis.

Cholinergic innervation of the human thalamus: dual origin and differential nuclear distribution.

The cholinergic innervation of the human thalamus was studied with antibodies against the enzyme choline acetyltransferase (ChAT) and nerve growth factor receptor (NGFr). Acetylcholinesterase histochemistry was used to delineate nuclear boundaries. All thalamic nuclei displayed ChAT-positive axons and varicosities. Only the medial habenula contained ChAT-positive perikarya. Some intralaminar nuclei (central medial, central lateral, and paracentral), the reticular nucleus, midline nuclei (paraventricular and reuniens), some nuclei associated with the limbic system (anterodorsal nucleus and medially situated patches in the mediodorsal nucleus) and the lateral geniculate nucleus displayed the highest density of ChAT-positive axonal varicosities. The remaining sensory relay nuclei and the nuclei interconnected with the motor and association cortex displayed a lower level of innervation. Immunoreactivity for NGFr was observed in cholinergic neurons of the basal forebrain but not in cholinergic neurons of the upper brainstem. The contribution of basal forebrain afferents to the cholinergic innervation of the human thalamus was therefore studied with the aid of NGFr-immunoreactive axonal staining. The anterior intralaminar nuclei, the reticular nucleus, and medially situated patches in the mediodorsal nucleus displayed a substantial number of NGFr-positive varicose axons, presumably originating in the basal forebrain. Rare NGFr-positive axonal profiles were also seen in many of the other thalamic nuclei. These observations suggest that thalamic nuclei affiliated with limbic structures and with the ascending reticular activating system are likely to be under particularly intense cholinergic influence. While the vast majority of thalamic cholinergic input seems to come from the upper brainstem, the intralaminar and reticular nuclei, and especially medially situated patches within the mediodorsal nucleus also appear to receive substantial cholinergic innervation from the basal forebrain.

Human reticular formation: cholinergic neurons of the pedunculopontine and laterodorsal tegmental nuclei and some cytochemical comparisons to forebrain cholinergic neurons.

Choline acetyltransferase immunohistochemistry showed that the human rostral brainstem contained cholinergic neurons in the oculomotor, trochlear, and parabigeminal nuclei as well as within the reticular formation. The cholinergic neurons of the reticular formation were the most numerous and formed two intersecting constellations. One of these, designated Ch5, reached its peak density within the compact pedunculopontine nucleus but also extended into the regions through which the superior cerebellar peduncle and central tegmental tract course. The second constellation, designated Ch6, was centered around the laterodorsal tegmental nucleus and spread into the central gray and medial longitudinal fasciculus. There was considerable transmitter-related heterogeneity within the regions containing Ch5 and Ch6. In particular, Ch6 neurons were intermingled with catecholaminergic neurons belonging to the locus coeruleus complex. The lack of confinement within specifiable cytoarchitectonic boundaries and the transmitter heterogeneity justified the transmitter-specific Ch5 and Ch6 nomenclature for these two groups of cholinergic neurons. The cholinergic neurons in the nucleus basalis (Ch4) and those of the Ch5-Ch6 complex were both characterized by perikaryal heteromorphism and isodendritic arborizations. In addition to choline acetyltransferase, the cell bodies in both complexes also had high levels of acetylcholinesterase activity and nonphosphorylated neurofilament protein. However, there were also marked differences in cytochemical signature. For example, the Ch5-Ch6 neurons had high levels of NADPHd activity, whereas Ch4 neurons did not. On the other hand, the Ch4 neurons had high levels of NGF receptor protein, whereas those of Ch5-Ch6 did not. On the basis of animal experiments, it can be assumed that the Ch5 and Ch6 neurons provide the major cholinergic innervation of the human thalamus and that they participate in the neural circuitry of the reticular activating, limbic, and perhaps also extrapyramidal systems.

Cholinergic innervation of the human striatum, globus pallidus, subthalamic nucleus, substantia nigra, and red nucleus.

The anatomical organization of cholinergic markers such as acetylcholinesterase, choline acetyltransferase, and nerve growth factor receptors was investigated in the basal ganglia of the human brain. The distribution of choline acetyltransferase-immunoreactive axons and varicosities and their relationship to regional perikarya showed that the caudate, putamen, nucleus accumbens, olfactory tubercle, globus pallidus, substantia nigra, red nucleus, and subthalamic nucleus of the human brain receive widespread cholinergic innervation. Components of the striatum (i.e., the putamen, caudate, olfactory tubercle, and nucleus accumbens) displayed the highest density of cholinergic varicosities. The next highest density of cholinergic innervation was detected in the red nucleus and subthalamic nucleus. The level of cholinergic innervation was of intermediate density in the globus pallidus and the ventral tegmental area and low in the pars compacta of the substantia nigra. Immunoreactivity for nerve growth factor receptors (NGFr) was confined to the cholinergic neurons of the basal forebrain and their processes. Axonal immunoreactivity for NGFr was therefore used as a marker for cholinergic projections originating from the basal forebrain (Woolf et al., '89: Neuroscience 30:143-152). Although the vast majority of striatal cholinergic innervation was NGFr-negative and, therefore, intrinsic, the striatum also contained NGFr-positive axons, indicating the existence of an additional cholinergic input from the basal forebrain. This basal forebrain cholinergic innervation was more pronounced in the putamen than in the caudate. The distribution of NGFr-positive axons suggested that the basal forebrain may also project to the globus pallidus but probably not to the subthalamic nucleus, substantia nigra, or red nucleus. The great majority of cholinergic innervation to these latter three structures and to parts of the globus pallidus appeared to come from cholinergic neurons outside the basal forebrain, most of which are probably located in the upper brainstem. These observations indicate that cholinergic neurotransmission originating from multiple sources is likely to play an important role in the diverse motor and behavioral affiliations that have been attributed to the human basal ganglia.

Differential cholinergic innervation within functional subdivisions of the human cerebral cortex: a choline acetyltransferase study.

The distribution of cholinergic fibers in the human brain was investigated with choline acetyltransferase immunocytochemistry in 35 cytoarchitectonic subdivisions of the cerebral cortex. All cortical areas and all cell layers contained cholinergic axons. These fibers displayed numerous varicosities and, on occasion, complex preterminal profiles arranged in the form of dense clusters. The density of cholinergic axons tended to be higher in the more superficial layers of the cerebral cortex. Several distinct patterns of lamination were identified. There were also major differences in the overall density of cholinergic axons from one cytoarchitectonic area to another. The cholinergic innervation of primary sensory, unimodal, and heteromodal association areas was lighter than that of paralimbic and limbic areas. Within unimodal association areas, the density of cholinergic axons and varicosities was significantly lower in the upstream (parasensory) sectors than in the downstream sectors. Within paralimbic regions, the non-isocortical sectors had a higher density of cholinergic innervation than the isocortical sectors. The highest density of cholinergic axons was encountered in core limbic structures such as the hippocampus and amygdala. These observations show that the cholinergic innervation of the human cerebral cortex displays regional variations that closely follow the organization of information processing systems.

Cholinergic innervation of the amygdaloid complex in the human brain and its alterations in old age and Alzheimer's disease.

The cholinergic innervation of the human amygdaloid complex was studied immunohistochemically with a choline acetyltransferase (ChAT) antibody in eight brains: five control and three with Alzheimer's disease (AD). All amygdaloid nuclei displayed ChAT-immunopositive axons and varicosities. The density of these axons reached levels that were higher than in any other part of the forebrain except for the striatum. The highest level of ChAT-immunopositive profiles was seen in the basolateral nucleus and the second highest in the lateral part of the central nucleus. The basomedial, accessory basal, and cortical nuclei, the amygdalohippocampal and cortico-amygdaloid transition areas, as well as the anterior amygdaloid area, showed a moderate density of ChAT-positive varicosities and fibers. The lateral nucleus displayed a relatively low density of cholinergic innervation, and there were only rare ChAT-positive fibers in the medial nucleus. Although the level of cholinergic innervation in the lateral nucleus was relatively lower than in many of the other amygdaloid nuclei, it was approximately equivalent to that of entorhinal cortex, a region that receives one of the heaviest cholinergic inputs in the cerebral cortex. The distribution of the cholinergic fibers as studied by ChAT immunohistochemistry was nearly identical to that observed with AChE histochemistry. Quantitative densitometry in control specimens showed that there was no decline of amygdaloid cholinergic input when middle-aged subjects were compared with senescent subjects. In AD there was a severe and regionally selective depletion of this innervation in the amygdaloid complex. The cortical, accessory basal, and lateral nuclei displayed the most severe loss of ChAT-positive profiles, whereas the basolateral, and especially the central, nuclei displayed relatively little change. There was no consistent relationship between the loss of cholinergic fibers and the density of amyloid plaques and neurofibrillary tangles in amygdaloid nuclei.

Architecture of connectivity within a cingulo-fronto-parietal neurocognitive network for directed attention.

The spatial distribution of directed attention is coordinated by a large-scale neural network. The three principal cortical components of this network are located in the region of the frontal eye fields, posterior parietal cortex, and the cingulate cortex. We injected a retrogradely transported fluorescent dye into the frontal eye fields and another into the posterior parietal cortex of the monkey brain. Large numbers of neurons in the cingulate cortex were retrogradely labeled with each of the two fluorescent dyes. The two types of retrogradely labeled neurons were extensively intermingled, but neurons labeled with both tracers constituted less than 1% of retrogradely labeled cingulate neurons. Other cortical areas that contained retrograde neuronal labeling included the premotor, lateral neuronal labeling included the premotor, lateral prefrontal, orbitofrontal, opercular, posterior parietal, lateral temporal, inferior temporal, parahippocampal, and insular regions. These areas contained neurons labeled with each of the two dyes but virtually no neurons labeled with both. In the thalamus, retrogradely labeled nuclei failed to display evidence of double labeling. The overlap between the two populations of retrogradely labeled neurons was far more extensive at the cortical than at the thalamic level. These observations show that cortical and thalamic projections to the frontal eye fields and posterior parietal cortex do not represent axonal collaterals of single neurons but originate from two distinct and partially overlapping populations of neurons.

Neuropsychological patterns and language deficits in 20 consecutive cases of autopsy-confirmed Alzheimer's disease.

A retrospective chart review of clinical symptoms was done for 20 consecutive patients in whom postmortem examination had revealed senile plaques and neurofibrillary tangles in a distribution consistent with Alzheimer's disease. All patients had met clinical diagnostic criteria for probable Alzheimer's disease. On initial examination, 1 to 14 years beyond putative onset of the dementia, all patients displayed at least some memory impairment. In 16 patients, disturbances of attention or recent memory were among the most salient features. In two patients, language disturbances, and in two others, visuospatial deficits, were more prominent than difficulties with memory and attention. On initial examination, 17 of the 20 patients displayed word-finding difficulties, characteristically in the context of a fluent, anomic aphasia. All of the 12 reexamined patients demonstrated progressive, although variable, deterioration. In general, the initial salient deficit remained salient during much of the disease course. Language comprehension was spared in the earlier stages but eventually deteriorated. Severe deficits emerged in all major cognitive domains as the disease reached the terminal stages. Nonfluent aphasias (eg, Broca's aphasia) were not observed even in the advanced stages of the disease.