Stereological estimates of total neuron numbers in the hippocampus of adult reeler mutant mice: Evidence for an increased survival of Cajal-Retzius cells.

The cytoarchitecture of the brain is disrupted severely in reeler mice. This is caused by a deficiency in the protein, Reelin, which is essential for the normal migration and positioning of neurons during development. Although cell migration is clearly affected by the reeler mutation, it is believed that the total number of neurons is not. Thus, we were surprised to find an unusually large number of calretinin-immunopositive cells, presumably Cajal-Retzius cells, in the molecular layer of the adult reeler hippocampus (Deller et al. [1999]; Exp. Neurol. 156:239-253). This suggested that the reeler mutation affects the number of neurons in the hippocampus. In order to verify this hypothesis, unbiased stereological methods were employed. Calretinin immunostaining was used as a marker for Cajal-Retzius cells in control as well as reeler mice and Nissl staining was used to identify hippocampal principal neurons. Total numbers of calretinin-immunopositive cells, calretinin-immunoreactive Cajal-Retzius cells, and Nissl-stained neurons were estimated in different subfields of the reeler and the control hippocampus. Stereological estimates (P < 0.05) revealed that the total number of calretinin-immunopositive and Cajal-Retzius cells in reeler mice are 1.5 and 2.1 times that of controls, respectively. No significant difference in total neuron number was found in any hippocampal subfield. These data demonstrate that the reeler mutation affects the number of calretinin-immunoreactive Cajal-Retzius cells in the adult hippocampus, probably due to a reduced excitatory innervation by entorhinal terminals in the absence of reelin. However, the reeler mutation does not affect mechanisms that determine total hippocampal neuron number.

Role of afferent innervation and neuronal activity in dendritic development and spine maturation of fascia dentata granule cells.

By using slice cultures of hippocampus as a model, we have studied the development of dendritic spines in fascia dentata granule cells. We raised the question as to what extent spine development is dependent on a major afferent input to these neurons, the fibers from the entorhinal cortex and neuronal activity mediated by these axons. Our results can be summarized as follows: (i) the entorhino-hippocampal projection develops in an organotypic manner in co-cultures of entorhinal cortex and hippocampus. Like in vivo, entorhinal fibers, labeled by anterograde tracing with biocytin, terminate in the outer molecular layer of the fascia dentata. (ii) The layer-specific termination of entorhinal fibers is not altered by the blockade of neuronal activity with tetrodotoxin. Likewise, the differentiation of the dendritic arbor of postsynaptic granule cells does not require neuronal activity. Blockade of neuronal activity did not affect the mean spine number of granule cell dendrites in entorhino-hippocampal co-cultures, but led to a relative increase in thin, long filiform spines that are characteristic of immature neurons. (iii) The maturation of the granule cell dendritic arbor is, however, controlled by the afferent fibers from the entorhinal cortex in an activity-independent manner. In single slice cultures of hippocampus lacking entorhinal input, Golgi-impregnated granule cells have much shorter, less branched dendrites when compared with granule cells in entorhino-hippocampal co-cultures. This reduction in dendritic length in granule cells lacking entorhinal input results in a lower mean total number of spines per neuron, but the mean number of spines per microm is not reduced in the absence of entorhinal innervation. These results indicate that innervation by fibers from the entorhinal cortex, but not neuronal activity mediated via these axons, is essential for the normal development of the granule cell dendritic arbor. Neuronal activity is required, however, for the maturation of dendritic spines.

Blockade of neuronal activity alters spine maturation of dentate granule cells but not their dendritic arborization.

Organotypic co-cultures of the entorhinal cortex and hippocampus were examined to determine the role of the entorhinal fibers in the dendritic development and formation of spines of dentate granule cells. Quantitative analysis of Golgi-impregnated granule cells in single hippocampal cultures and co-cultures with the entorhinal cortex revealed that the presence of entorhinal fibers promoted the elongation and differentiation of the target granule cell dendrites. This was accompanied by an increase in the total number of spines. The contribution of neuronal activity to this afferent-mediated dendritic development was tested by chronic application of the sodium channel blocker tetrodotoxin for 20 days in vitro. Tracing with biocytin showed that the formation of the entorhinohippocampal pathway was unaffected by the blockade of neuronal activity. The dendritic arbor of cultured granule cells and the number of dendritic spines did not differ between tetrodotoxin-treated slices and untreated controls. However, there was a significant increase in the relative number of filiform spines on granule cell dendrites in tetrodotoxin-treated co-cultures. Such filiform spines are a characteristic feature of immature neurons. These results suggest the cooperation of two mechanisms in the dendritic development of dentate granule cells: the specific afferent-mediated dendritic arborization and the activity-dependent maturation of spines.

Different primary target cells are important for fiber lamination in the fascia dentata: a lesson from reeler mutant mice.

The factors determining the lamina-specific termination of entorhinal and commissural afferents to the fascia dentata are poorly understood. Recently it was shown that early generated Cajal-Retzius (CR) cells in the outer molecular layer and reelin, synthesized by CR cells, play a role in the lamina-specific termination of entorhinal fibers which form transient synapses with CR cells before establishing their definite contacts with granule cell dendrites (J. A. del Rio et al., 1997, Nature 385, 70-74). By using anterograde tracing with Phaseolus vulgaris leukoagglutinin we show that the normal, sharply delineated entorhinal projection to the outer molecular layer is retained in reeler mutant mice lacking reelin. This coincides with the regular presence of CR cells, the primary, transient target cells of entorhinal fibers. In contrast, the commissural fibers were found to terminate in an abnormal broad, not clearly defined area. This widespread projection coincides with the distribution of granule cells which in the mutant do not form a dense cell layer but are scattered all over the hilus due to a migration defect. Unlike the entorhinal fibers, the commissural fibers arrive in their target layer late in development, when granule cell dendrites are already there. We hypothesize from these results that the presence of the adequate postsynaptic element at the time of fiber ingrowth, CR cells for the early ingrowing entorhinal fibers and granule cells for the late-arriving commissural fibers, is crucial for the normal formation of these layer-specific projections.

The hippocampus of the reeler mutant mouse: fiber segregation in area CA1 depends on the position of the postsynaptic target cells.

Area CA1 of the rodent hippocampus is characterized by a highly lamina-specific and nonoverlapping termination of afferent fiber tracts. Entorhinal fibers terminate in stratum lacunosum-moleculare and commissural/associational fibers terminate in strata radiatum and oriens. It has been hypothesized that this fiber lamination depends on specific signals for the different afferent fiber tracts that are located on distinct dendritic segments of the postsynaptic neuron. In order to test this hypothesis, entorhinal and commissural/associational afferents to Ammon's horn were investigated in the adult reeler mutant mouse, in which developmental cell migration defects have disrupted the normal array of cells. Golgi staining revealed a deep and a superficial principal cell layer in the mutant. The morphology of the cells located in the deep pyramidal cell layer was considerably abnormal, whereas most cells located in the superficial pyramidal cell layer showed an almost normal cellular and dendritic morphology. Anterograde tracing with Phaseolus vulgaris leukoagglutinin revealed that the duplication and disorganization of the pyramidal cell layer in area CA1 are mirrored by the duplication and disruption of afferent fiber termination zones. In the zone above the abnormal deep pyramidal cell layer, i.e., between the two cell layers, entorhinal fibers as well as commissural/associational fibers terminate and intermingle. In contrast, in the zone above the fairly normal superficial pyramidal cell layer, entorhinal and commissural/associational fibers retain their normal fiber segregation. These data suggest that the normal laminar organization of the murine hippocampus depends on positional cues presented by their target cells.

Developmental distribution of a reeler gene-related antigen in the rat hippocampal formation visualized by CR-50 immunocytochemistry.

During histogenesis of the neocortex, Cajal Retzius cells in the marginal zone express the glycoprotein reelin which is developmentally regulated and involved in the formation of the inside out mode of cortical layering. Cajal Retzius cells are also present in the developing hippocampus. There, inhibition of reelin by blocking with CR-50, an antibody which recognizes the N-terminus of this protein, leads to abnormal development of layer-specific connections. Here we report the developmental distribution pattern of reelin expressing neurons in the rat hippocampal formation using CR-50 immunocytochemistry. Labelled Cajal Retzius cells were located near the hippocampal fissure in neonate rats. Many of these cells were still present in the adult. From postnatal day 4 on, neurons in other layers were stained with the CR-50 antibody. In adult rats immunopositive neurons were found in all hippocampal subfields and in the entorhinal cortex. These observations indicate that in the rat hippocampal formation reelin is expressed in different neuronal types during development and in adulthood. Moreover, Cajal Retzius cells in the marginal zone near the hippocampal fissure are still found in adult animals.

A role for Cajal-Retzius cells and reelin in the development of hippocampal connections.

During development of the nervous system, specific recognition molecules provide the cues necessary for the formation of neural connections. In some regions, guiding cues for axonal pathfinding and target selection are provided by specific cells that exist only transiently during development, such as the floorplate or the cortical subplate. In the hippocampus, distinct groups of fibres innervate different layers. We have tested the hypothesis that transient neurons in the hippocampus provide positional information for the targeting of these fibres. Here we report that ablation of Cajal-Retzius cells in organotypic slice cultures of hippocampus prevented the ingrowth of entorhinal but not of commissural afferents. Experiments inhibiting Reelin (an extracellular matrix protein expressed by Cajal-Retzius cells) and analysis of reeler mutant mice showed dramatic abnormalities in the development of entorhinal afferents. Thus Cajal-Retzius cells and reelin are essential for the formation of layer-specific hippocampal connections.

Spine loss in experimental epilepsy: quantitative light and electron microscopic analysis of intracellularly stained CA3 pyramidal cells in hippocampal slice cultures.

The sequence of neuronal alterations resulting from epileptic activity is poorly understood. In the hippocampus of some epileptic patients, there is a loss of certain neuronal types in the hilar region and in CA3. The neuronal alterations preceding this degeneration probably affect synaptic structures. Here we have estimated the number of dendritic spines, major postsynaptic elements of hippocampal neurons, in defined dendritic segments of identified (intracellularly stained) CA3 pyramidal neurons in "epileptic" slice cultures of hippocampus and in control cultures. Slice cultures were prepared from five- or six-day-old rat pups and maintained in vivo for 23 days before epileptic activity was induced by application of the convulsants bicuculline and picrotoxin for three days. Individual CA3 pyramidal neurons were then intracellularly injected with horseradish peroxidase, and the number of dendritic spines was counted in proximodistal dendritic segments by applying the Sholl method. In addition, the total dendritic length was measured and the branching index evaluated. The number of spines on CA3 pyramidal cell dendrites in the "epileptic" cultures was found to be decreased by 40%. This spine loss affected proximal and peripheral dendritic segments of the CA3 pyramidal neurons to a similar extent. No significant differences were observed between control and "epileptic" cultures in dendritic length or in the branching index. Quantitative electron microscopic analysis did not reveal differences between "epileptic" cultures and control cultures in the spine area of the labelled CA3 pyramidal cells, indicating that there was a real spine loss, not just a reduction in the size of the spines. We conclude that epileptic activity causes morphological alterations in defined postsynaptic compartments of hippocampal pyramidal cells surviving under these conditions.