Glenn H. Kageyama

Histochemical localization of cytochrome oxidase in the hippocampus: Correlation with specific neuronal types and afferent pathways

Cytochrome oxidase was histochemically localized in the hippocampus and dentate gyrus of various species of mammals. The most intense staining was observed within stratum moleculare of areas CA1-3 and the outer molecular layer of the dentate gyrus, as well as the somatic and basal dendritic layers of CA3. These regions correspond to the synaptic terminal fields of major excitatory afferent pathways to the hippocampus. The somata of CA3 pyramidal cells and various interneurons were more intensely stained than CA1 pyramidal cells and dentate granule cells, and these levels appeared to correlate positively with their reported rates of spontaneous firing. At the electron-microscopic level, the highest concentrations of densely reactive mitochondria were localized within the distal apical dendritic profiles of principal cells (granule and pyramidal) and certain interneurons (pyramidal basket and stratum pyramidale interneurons). The specific layers in which these structures were found are known to receive intense excitatory input from the perforant pathway. High concentrations of reactive mitochondria were also observed within the somata and proximal dendrites of CA3 pyramidal cells and various interneurons, confirming our light-microscopic observations. These results demonstrated that not only can soma and dendrites of the same cell have disparate but distinct levels of cytochrome oxidase activity, but the pattern of reactivity within a neurons apical and basal dendrites, or even within specific dendritic segments of the same dendrite can be quite different. While the levels of somatic reactivity correlate with reported levels of spontaneous and/or synaptic activity, the degree of dendritic and somatic staining appeared to be more closely related to the intensity of convergent and/or pathway-specific excitatory synaptic input.

Primary auditory cortex in the rat: transient expression of acetylcholinesterase activity in developing geniculocortical projections

A characteristic pattern of acetylcholinesterase (AChE) activity is expressed transiently in primary auditory cortex (cortical area 41) of developing laboratory rats during early postnatal life. This AChE activity occurs as a dense plexus in cortical layer IV and the deep part of layer III. This transient band of AChE activity is first detected by histochemical techniques on postnatal day (P) 3, reaches peak intensity at approximately P8-10, and declines to form the adult pattern by P23. The ventral nucleus of the medial geniculate body of the thalamus also displays prominent, and transient, staining for AChE. This intense staining for AChE, found within neuronal somata and neuropil, is detected at the time of birth, reaches peak intensity around P8, and declines to adult levels by P16. The areal and laminar patterns of the transient band of AChE activity in temporal cortex correspond to the patterns of anterograde transneuronal labeling of geniculocortical terminals following injection of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) into the inferior colliculus. Placement of lesions that include the medial geniculate nucleus or the geniculocortical axons results in a marked decrease in AChE staining in thalamorecipient layers of auditory cortex. Placement of lesions that include the medial globus pallidus reduce AChE staining of some axons in temporal cortex of developing rats, but the dense band of AChE in layers III and IV remains. Placement of lesions in the inferior colliculus in newborn animals results in marked decrease in AChE staining in cells of the ipsilateral ventral medial geniculate nucleus and in ipsilateral auditory cortex of developing pups. These data indicate that transiently expressed AChE activity is characteristic of geniculocortical neurons, including their somata in the medial geniculate body and their terminal axons in primary auditory cortex. This AChE activity is expressed early in postnatal development, probably during the time when thalamocortical axons are proliferating in cortical layer IV and forming synaptic contacts with cortical neurons.

POSTNATAL MIGRATION OF NEURONS AND FORMATION OF LAMINAE IN RAT CEREBRAL CORTEX

Migration of neurons and formation of laminae in the developing neocortex were studied by means of thymidine autoradiography. Timed pregnant rats received a single pulse injection of [3H]thymidine in the morning of embryonic day (E)13, 14, 15, 16, 17, 18 or 19. Pups were killed on postnatal day (P)0, 1, 2, 3, 4, 6, 10, 30, or 60 and brains were processed for autoradiography. Neurons in posterior (visual) cortical areas labeled by [3H]thymidine administration on E13 or E14 were found predominantly in the cortical subplate; cells labeled on El5 in layer VI; cells labeled on E16 in layers VI and V, cells labeled on El7 in layers V and IV; E18 in layers IV and III; and E19 in layers III and II. By the day of birth (PO), neurons labeled from E13-16 injections were already in their mature laminae in cortex. Many of the cells labeled on E17 were still situated within the cell-dense cortical plate (CP) at PO, and within layer V by P1. Cells labeled on E18 were found in the most superficial part of the CP on PO, in the deep part of the CP on P1, and formed layer IV on P2 and P3. At PO, many E19 labeled cells appeared to be in migration to the cortex and were found in the CP on P1, in layer III by P4, and in layer II by P6. Cells in the auditory cortex labeled by [3H]thymidine injections on a particular day were situated more superficially than comparable labeled cells in the visual cortex, indicating a lateral to medial gradient in which the auditory cortex is formed earlier than the visual cortex. Distributions of labeled cells in the somatosensory cortex were similar to those in the visual cortex. These data provide a detailed and comprehensive description of the position of varied populations of cortical neurons during the early postnatal period, as well as a description of the formation of cortical laminae at times when major systems of afferents are growing into the cortex and making synaptic connections with their target cells.

Do subplate neurons comprise a transient population of cells in developing neocortex of rats

Studies were undertaken to determine whether neurons of the subplate layer represent a transient or stable population of cells in developing neocortex of rat. The first set of studies sought to determine the fraction of subplate neurons that is lost during early postnatal development. The optical dissector method was used to analyze fluorescently stained material in animals the age of postnatal day 0 (P0) to P40. These results demonstrate a reduction of slightly less than half of the total number of subplate neurons from P0 to P40. Counts of labeled cells in littermates at varied ages after [3H]thymidine or BRDU treatment on gestational day 14 (G14 ‐ birthdate of occipital subplate neurons) or G18 (birthdate of layers III–IV neurons) demonstrate loss of approximately 50% of neurons in the subplate layer between P0 and P40, somewhat greater than the loss of neurons from cortical layers III–IV. The second set of studies investigated whether subplate neurons display cellular atrophy during postnatal development. Analysis of subplate neurons injected intracellularly with Lucifer yellow in fixed slice preparations indicates no reduction in soma size, number of dendrites, or extent of dendritic fields of subplate neurons taken from animals age P0 to P60. The third set of studies investigated whether functional markers of subplate neurons are reduced during postnatal development. Analysis of tissue stained histochemically for cytochrome oxidase or acetylcholinesterase, or stained immunocytochemically for GABA, somatostatin, or neuropeptide Y, demonstrate a remarkable loss of expression of staining patterns from late gestational ages to P20. These data demonstrate that, although subplate neurons seem not to be a transient population of cells in the usual sense of being eliminated by cell death or structural atrophy, the loss of histochemical and immunocytochemical markers indicates that they may be a functionally transient population of cells. J. Comp. Neurol. 426:632–650, 2000.

The relationship between CNS metabolism and cytoarchitecture: A review of 14C-deoxyglucose studies with correlation to cytochrome oxidase histochemistry

Since the inception of the 14C-deoxyglucose method and its extension to in vivo imaging of regional cerebral glucose metabolism in humans by positron emission tomography, uncertainty has persisted concerning the type of work to which regional metabolism is coupled, as well as the distribution of this work within the neuron. 14C-deoxyglucose studies indicate that functionally-coupled neural metabolism is more apparent in axon terminals and perhaps dendrites than neuronal perikarya. Moreover, it appears that most of the metabolism in axon terminals is accounted for by Na+-K+-ATPase activity. Nevertheless, cytochrome oxidase histochemistry reveals the presence of intensely reactive mitochondria in soma-dendrite regions opposite presynaptic axon terminals, thereby indicating that continuous temporal and spatial summation of postsynaptic graded potentials is associated with increased metabolism. While the situation concerning the relative postsynaptic metabolic prices of EPSPs and IPSPs remains uncertain, the presence of elevated levels of cytochrome oxidase activity within certain classes of presynaptic terminals indicates that active excitation and inhibition is associated with increases in presynaptic metabolism. This observation has been confirmed in 14C-deoxyglucose studies. Nevertheless, studies of neonatal hippocampus indicate that, before metabolic activity shifts to dendritic and telodendritic regions of electrophysiological activity, metabolism is high in somal foci of biosynthesis.

Glutamate-immunoreactivity in the retina and optic tectum of goldfish.

Glutamate was immunohistochemically localized in the goldfish retina and tectum at the light and electron microscopic (E.M.) levels using double affinity purified antisera against glutaraldehyde conjugated L-glutamate. In retina, glutamate-immunoreactivity (Glu+) was observed in cone inner segments, cone pedicles, bipolar cells, a small number of amacrine cells and the majority of cells in the ganglion cell layer. The latter were shown to be ganglion cells by simultaneous retrograde labeling. Centrally, Glu+ was observed in axons in the optic nerve and tract, and in stratum opticum and stratum fibrosum et griseum superficialis (SFGS) of the tectum. The Glu+ in the optic pathway disappeared four days after optic denervation and was restored by regeneration without affecting the Glu+ of intrinsic tectal neurons. In tectum, Glu+ was also observed in torus longitudinalis granule cells, toral terminals in stratum marginale, some pyramidal neurons in the SFGS, multipolar and fusiform neurons in stratum griseum centrale, large multipolar and pyriform projection neurons in stratum album centrale, and many periventricular neurons. Glu+ was also localized within unidentified puncta throughout the tectum and within radially oriented dendrites of periventricular neurons. At the E.M. level, a variety of Glu+ terminals were observed. Glu+ toral terminals formed axospinous synapses with dendritic spines of pyramidal neurons. Ultrastructurally identifiable Glu+ putative optic terminals formed synapses with either Glu+ or Glu- dendritic profiles, and with Glu- vesicle-containing profiles, presumed to be GABAergic. These findings are consistent with the hypothesis that a number of intrinsic and projection neurons in the goldfish retinotectal system, including most ganglion cells, may use glutamate as a neurotransmitter.

LARGE-SCALE SYNAPTIC ERRORS DURING MAP FORMATION BY REGENERATING OPTIC AXONS IN THE GOLDFISH

During the formation of visual maps, growing axons initially form a map by using topographically distributed cues that direct their growth and branching to the appropriate target region. This initial map is typically roughly retinotopic and is subsequently refined through activity‐dependent rearrangement or cell death. Although synaptic connections are thought to be rearranged during the later refinement phase, there is no clear evidence that synapses are being formed during the initial targeting phase of development. Also, because optic fiber growth can be accurately directed during normal development, it is unclear whether regenerative fibers that have more pathway disorder would behave similarly. This issue was addressed by using optic fibers of goldfish that have the capacity to regenerate a retinotopic projection and can reestablish a rough retinotopic order without impulse activity. The optic nerve of goldfish was crushed, and at various times later, a small number of optic fibers in ventronasal retina was labeled with wheatgerm agglutinin–horseradish peroxidase. The tectum was then processed for electron microscopy to look at the distribution of labeled synapses during regeneration. At 3 weeks, synapses were observed at the far anterior end of the tectum and none were yet seen at the correct posterior retinotopic position. At 4–5 weeks, synapses were seen in nearly equal numbers at the incorrect anterior end and at both correct (retinotopic) and incorrect posterior positions. At late stages of regeneration, synapses were restricted to their correct posterior retinotopic position in the tectum, as they were in normal fish. These findings show that the formation of global retinotopic order entails the formation and subsequent elimination of a large number of highly ectopic synapses. Synaptic rearrangement is a major feature of targeting in this system and may be required for the regeneration of a retinotopic projection. J. Comp. Neurol. 409:299–312, 1999.

Relationships between patterns of acetylcholinesterase activity and geniculocortical terminal fields in developing and mature rat visual cortex

Intraocular injections of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) result in anterograde transneuronal labeling of geniculocortical axon terminals in cortical area 17. In area 17 of developing rat pups, transported WGA-HRP occurs primarily in layer I and in a band that includes layer IV and deep layer III; this pattern is virtually identical to the laminar pattern of endogenous acetylcholinesterase (AChE) activity. In adult animals, transported WGA-HRP again is localized in layer I and in deep layer III and layer IV, but the endogenous AChE activity is found most prominently in deep layer IV and layer V. These results indicate that geniculocortical terminal fields are co-extensive with transient patterns of AChE activity in the developing rat, but not with the mature pattern of AChE in the adult.

Dense HRP filling in pre-fixed brain tissue for light and electron microscopy.

The use of neuroanatomical markers in tissues that have been pre-fixed has been virtually ignored, even though this approach could offer certain advantages over in vivo methods, in terms of convenience of application and choice of markers. We have found that HRP can be used on well-fixed brains of cats and goldfish to fill neurons, dendrites, axons, terminals, glial cells, and glial processes for high-resolution light microscopy and electron microscopy. Best results were obtained using brains that were perfusion-fixed with 2.5% depolymerized paraformaldehyde and 1.5% glutaraldehyde. Two methods of HRP application were used: optically guided injections of microliter quantities into various regions of cat brain, and optic nerve fills in goldfish by attaching an HRP-filled polyethylene tube for periods of 1 day to 2 weeks. HRP applied in these ways to pre-fixed tissue was found to fill neurons or glial cells with solid label in the anterograde and retrograde directions.

Specificity of attachment and neurite outgrowth of dissociated basal forebrain cholinergic neurons seeded on to organotypic slice cultures of forebrain

Development and differentiation of basal forebrain-derived cholinergic neurons were studied using a new technique that combines dissociated cell cultures with organotypic slice cultures. Slices of cerebral cortex or entire forebrain hemispheres were taken from early postnatal rat pups and maintained as organotypic cultures on membranes. Dissociated cell suspensions of basal forebrain tissue, taken from rat or mouse fetuses at gestational day 15-17, were seeded on to the slice cultures. Combined cultures were maintained for two to 14 days in vitro. Cultures processed for acetylcholinesterase histochemical staining demonstrated that stained neurons display regional variation in attachment to the slice, with most attachment occurring on cortex and with no detectable attachment on the caudate-putamen. Regional differences in attachment occur between cortical areas, with medial (cingulate) cortex showing much denser cell attachment than lateral (parietal) cortex, and across cortical layers, with layer I and deep layers showing more attachment than middle cortical layers. Similar patterns were observed on slices from rat brain irrespective of whether rat or mouse dissociated cells were used. Tyrosine hydroxylase-stained dissociated cells from ventral midbrain displayed a different pattern of attachment, with prominent attachment to the caudate putamen and less apparent specificity of regional and cortical laminar attachment. Little evidence of neurite outgrowth occurred during the first two days in vitro, but by four days, acetylcholinesterase-positive basal forebrain cells displayed several short and thick neurites that appeared to be dendrites, and one long process that appeared to be an axon. By seven days in vitro, dendrites are well developed and the presumed axon has extended branches over wide areas of cortex. These studies revealed several different types of cell-tissue interaction. The degree of cell growth and differentiation ranged from robust growth when dissociated cells were seeded on to slice cultures of normal target tissue, to apparently no attachment or growth when cells were seeded on to non-target tissue. This combined technique appears to be a useful method for studies of specificity of cell attachment and patterns of neurite outgrowth.