An Immunochemical Analysis of the Distribution of a Brain-Specific Polypeptide, PEP-19

Immunochemical and immunohistochemical techniques were used to map the tissue distribution and cellular localization of a rat brain-specific polypeptide, termed PEP-19. PEP-19 was found to be abundant in the cerebellum and olfactory bulbs but was present at much lower levels in other gross brain regions. It was undetectable in all nonneural tissues examined but was present in the cerebellum of several vertebrates, including rat, mouse, guinea pig, monkey, and human. Immunohistochemical analysis revealed that PEP-19 was localized to the soma, axon, and dendritic processes of rat cerebellar Purkinje cells with no demonstrable immunoreactivity in nonneuronal cell types. Furthermore, mutant mice showing degeneration of Purkinje cells exhibit markedly decreased levels of PEP-19. Because PEP-19 appears during the final stages of maturation of Purkinje cells, it may be utilized as a probe to monitor the development of these neurons in vivo.     

Increased number of nitric oxide synthase immunoreactive Purkinje cells and dentate nucleus neurons in schizophrenia

There is growing interest in the cerebellum as a site of neuropathological changes in schizophrenia. Reports showing that schizophrenics have higher nitric oxide synthase (NOS) activity and MAPKinase levels in the vermis, point to possible aberrations in the cerebellar signal transduction of schizophrenics. It has been speculated that Ca²⁺-dependent extracellular to intracellular signal transduction may be disrupted in the cerebellum of schizophrenics. We decided to test this hypothesis by studying the nitrergic system and markers of the Ca²⁺-triggered signal cascade in the cerebellum of schizophrenics, depressives and controls. The cellular distribution of two calcium sensor proteins (VILIP-1 and VILIP-3) and of neuronal NOS immunoreactivity was studied morphometrically in the flocculonodulus, the inferior vermis and the dentate nucleus of 9 schizophrenics, 7 depressive patients and 9 matched controls. In comparison to controls and depressed patients there were fewer Nissl-stained neurons in the dentate nucleus of schizophrenics. The number of NOS-expressing Purkinje neurons was however strongly increased. In the flocculonodulus and the vermis no differences between the groups were found with regard to the density of Nissl-stained Purkinje cells. The number of NOS-expressing Purkinje neurons was increased in schizophrenics, however. No differences between schizophrenics, depressives and controls were found in the number of VILIP-1 immunoreactive dentate nucleus neurons and VILIP-3 immunoreactive vermal and flocculonodular Purkinje cells. Our data provide further histochemical evidence in favor of structural abnormalities in discrete cerebellar regions of schizophrenics. They confirm and extend earlier reports of increased cerebellar NOS immunoreactivity in schizophrenia and point to possible neurodevelopmental disturbances. Our failure to show an altered expression of two calcium sensor proteins possibly points to a less important role of calcium signaling in cerebellar pathology of the disease.     

Unequal patterns of development of succinate-dehydrogenase and acetylcholinesterase in Purkinje cell bodies and granule cells isolated in bulk from the cerebellar cortex of the immature rat

Abstract— –A preparative procedure for the isolation in bulk of two cellular populations of the cerebellar cortex of the immature rat, the granule cells and the Purkinje cell bodies, is described. The procedure is used to delineate the developmental pattern of succinate-INT-reduclase (EC 1.3.99.1) and acetylcholinesterase (EC 3.1.1.7) in the crucial period of cerebellar maturation, i.e. between 12 and 19 days postnatally. Although the overall yield of neuronal RNA diminished with age, the proportion of RNA in the Purkinje cell body fraction increased while that in the granule cells decreased and microscopic examination of the fractions confirmed this result. The yields of succinate-INT-reductase and of acetylcholinesterase in the fractions paralleled the yields of RNA. A significant finding was the trend toward diminishing specific activities (units/μg of RNA) with age of both enzymes in the Purkinje cell bodies as against the opposite, upward trend of their specific activities in the granule cells. An additional finding of interest was the different ratio of true acetylcholinesterase/total cholinesterase activity in the two cell types, with the granule cells consistently exhibiting higher true acetylcholinesterase values than the Purkinje cell bodies. The present report thus supplements the histoenzymological data on the developing rat cerebellum in that it reveals specific differences in the enzymatic development of two different cerebellar types, a finding which was greatly facilitated by the availability of the procedure for their bulk isolation.     

Cloning of the cDNA for a human homolog of the rat PEP-19 gene and mapping to chromosome 21q22.2–q22.3

Exon trapping was used to identify fragments of genes on human chromosome 21. One trapped sequence, hmc18h10 (GenBank no. X88329), showed homology to a sequence (GenBank no. S65225) that includes the first three codons of the rat PEP-19 gene and 5′ untranslated leader region. We have cloned the corresponding cDNA for a human homolog of the rat PEP-19 gene and mapped it to the region between markers ERG and D21S56 of chromosome 21q22.2–q22.3. Rat PEP-19 is a neuron-specific polypeptide expressed in several regions of the central nervous system. It serves as a cell-specific marker in Purkinje cells and its expression is developmentally regulated in the cerebellum, but its precise function is unknown. It is also presently unknown whether overexpression of the PEP-19 gene is involved in certain phenotypes of Down syndrome. Received: 3 May 1996 / Revised: 2 July 1996     

Cerebellar localization and colocalization of GABA and calcium binding protein-D28K

Immunocytochemical studies using antibodies raised against the inhibitory neurotransmitter, gamma-aminobutyric acid (GABA) and against the 28 Kd vitamin D dependent calcium binding protein (calbindin) in the cerebellum, are reviewed. The GABA immunoreactive neurones found in the cerebellar cortex were the Purkinje cell (PC), the three classes of intrinsic inhibitory interneurones, stellate, basket and Golgi cells and the cells of Lugaro. Some of the neurons of the cerebellar nuclei were also found to be GABA immunoreactive. A part of these could be identified as extrinsic neurones projecting either back to the cerebellar cortex, or to the inferior olive, both these pathways being topographically highly organized but arising from independent parent neurons. The presumed inhibitory function of these two pathways are discussed. Calbindin immunoreactivity in the cerebellum was confined to the PCs, staining concerned the whole cell including soma, branching dendrites, axons and axons terminals. The antibody, which appears to be tightly bound to the PC in vivo, failed to stain some of the PC when cerebellar slices maintained in vitro were studied. The stability of the antigen-antibody binding and the use of calbindin as a marker specific for the PC in the cerebellum, is discussed. Co-localization of GABA with calbindin as well as with other calcium binding proteins are reported to be found in the PCs. While these co-localizations have led to much speculation, conclusive functional roles for them have not been identified at present.     

Impaired calcium release in cerebellar Purkinje neurons maintained in culture

Cerebellar Purkinje neurons demonstrate a form of synaptic plasticity that, in acutely prepared brain slices, has been shown to require calcium release from the intracellular calcium stores through inositol trisphosphate (InsP(3)) receptors. Similar studies performed in cultured Purkinje cells, however, find little evidence for the involvement of InsP(3) receptors. To address this discrepancy, the properties of InsP(3)- and caffeine-evoked calcium release in cultured Purkinje cells were directly examined. Photorelease of InsP(3) (up to 100 microM) from its photolabile caged analogue produced no change in calcium levels in 70% of cultured Purkinje cells. In the few cells where a calcium increase was detected, the response was very small and slow to peak. In contrast, the same concentration of InsP(3) resulted in large and rapidly rising calcium responses in all acutely dissociated Purkinje cells tested. Similar to InsP(3), caffeine also had little effect on calcium levels in cultured Purkinje cells, yet evoked large calcium transients in all acutely dissociated Purkinje cells tested. The results demonstrate that calcium release from intracellular calcium stores is severely impaired in Purkinje cells when they are maintained in culture. Our findings suggest that cultured Purkinje cells are an unfaithful experimental model for the study of the role of calcium release in the induction of cerebellar long term depression.     

Intrinsically disordered PEP-19 confers unique dynamic properties to apo and calcium calmodulin

PEP-19 (Purkinje cell protein 4) is an intrinsically disordered protein with an IQ calmodulin (CaM) binding motif. Expression of PEP-19 was recently shown to protect cells from apoptosis and cell death due to Ca(2+) overload. Our initial studies showed that PEP-19 causes novel and dramatic increases in the rates of association of Ca(2+) with and dissociation of Ca(2+) from the C-domain of CaM. The goal of this work was to study interactions between the C-domain of CaM (C-CaM) and PEP-19 by solution nuclear magnetic resonance (NMR) to identify mechanisms by which PEP-19 regulates binding of Ca(2+) to CaM. Our results show that PEP-19 causes a greater structural change in apo C-CaM than in Ca(2+)-C-CaM, and that the first Ca(2+) binds preferentially to site IV in the presence of PEP-19 with exchange characteristics that are consistent with a decrease in Ca(2+) binding cooperativity. Relatively weak binding of PEP-19 has distinct effects on chemical and conformational exchange on the microsecond to millisecond time scale. In apo C-CaM, PEP-19 binding causes a redistribution of residues that experience conformational exchange, leading to an increase in the number of residues around Ca(2+) binding site IV that undergo conformational exchange on the microsecond to millisecond time scale. This appears to be caused by an allosteric effect because these residues are not localized to the PEP-19 binding site. In contrast, PEP-19 increases the number of residues that exhibit conformational exchange in Ca(2+)-C-CaM. These residues are primarily localized to the PEP-19 binding site but also include Asp93 in site III. These results provide working models for the role of protein dynamics in the regulation of binding of Ca(2+) to CaM by PEP-19.     

Localization of chondroitin sulfate proteoglycan versican in adult brain with special reference to large projection neurons

Versican is a chondroitin sulfate proteoglycan belonging to the lectican family. Versican has two glycosaminoglycan attachment regions, named the GAGα and GAGβ domains, which are both regulated by alternative splicing and yield four protein isoforms. We have investigated the expression and localization of versican in the developing and adult brain by using anti-versican GAGα and GAGβ antibodies. Western analysis revealed that GAGα-reactive isoform was dominant in the adult brain. Immunohistochemical study demonstrated that GAGα immunoreactivity was detectable from neonatal periods to adulthood, whereas GAGβ immunoreactivity completely disappeared within 3 weeks of birth. In the adult brain, GAGα immunoreactivity was seen in the white matter regions and was also localized in the gray matter including somata and dendrites of cortical and hippocampal pyramidal neurons and cerebellar Purkinje cells. In contrast, GAGα immunoreactivity was not localized on parvalbumin-positive interneurons and cerebellar stellate cells. Furthermore, GAGα immunoreactivity was not co-localized with perineuronal net markers such as Wisteria floribunda agglutinin lectin and phosphacan. Thus, versican was localized on large projection neurons rather than small interneurons. To confirm the binding mechanism of versican to neurons, hyaluronan and chondroitin sulfates were enzymatically removed from brain sections before the immunolabeling of versican. These treatments had no effect on the labeling pattern of versican, suggesting that other versican-interactive molecules are involved in the binding of versican to neurons. This study was supported by a Grant-in-Aid for Scientific Research on Priority Areas “Advanced Brain Science Project” from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.     

Identification and localization of ryanodine binding proteins in the avian central nervous system

Ryanodine binding proteins of the CNS have been identified using monoclonal antibodies against avian skeletal muscle ryanodine binding proteins. These proteins were localized to intracellular membranes of the dendrites, perikarya, and axons of cerebellar Purkinje neurons using laser confocal microscopy and immunoelectron microscopy. Ryanodine binding proteins were not found in dendritic spines. Immunoprecipitation and [3H]epiryanodine binding experiments revealed that the cerebellar ryanodine binding proteins have a native molecular weight of approximately 2000 kd and are composed of two high molecular weight (approximately 500 kd) polypeptide subunits. A comparable protein having a single high molecular weight polypeptide subunit was observed in the remainder of the brain. If the ryanodine binding proteins in muscle and nerve are similar in function, then the neuronal proteins may participate in the release of calcium from intracellular stores that are mechanistically and spatially distinct from those gated by inositol trisphosphate receptors.     

NeuN, a neuronal specific nuclear protein in vertebrates.

A battery of monoclonal antibodies (mAbs) against brain cell nuclei has been generated by repeated immunizations. One of these, mAb A60, recognizes a vertebrate nervous system- and neuron-specific nuclear protein that we have named NeuN (Neuronal Nuclei). The expression of NeuN is observed in most neuronal cell types throughout the nervous system of adult mice. However, some major cell types appear devoid of immunoreactivity including cerebellar Purkinje cells, olfactory bulb mitral cells, and retinal photoreceptor cells. NeuN can also be detected in neurons in primary cerebellar cultures and in retinoic acid-stimulated P19 embryonal carcinoma cells. Immunohistochemically detectable NeuN protein first appears at developmental timepoints which correspond with the withdrawal of the neuron from the cell cycle and/or with the initiation of terminal differentiation of the neuron. NeuN is a soluble nuclear protein, appears as 3 bands (46-48 x 10(3) M(r)) on immunoblots, and binds to DNA in vitro. The mAb crossreacts immunohistochemically with nervous tissue from rats, chicks, humans, and salamanders. This mAb and the protein recognized by it serve as an excellent marker for neurons in the central and peripheral nervous systems in both the embryo and adult, and the protein may be important in the determination of neuronal phenotype.     

Abnormal HNK-1 expression in the cerebellum of an N-CAM null mouse

Human natural killer antigen-1 (HNK-1) is a carbohydrate epitope associated with sulfoglucuronylglycolipids and glycoproteins. Biochemical analyses have demonstrated associations between the HNK-1 epitope and isoforms of the neural cell adhesion molecule (N-CAM) family. In the cerebellum, HNK-1 is prominently expressed in Purkinje cell dendrites and Golgi cells. Purkinje cell expression of HNK-1 reveals an array of parasagittal stripes and transverse zones. Interestingly, the parasagittal expression pattern of HNK-1 is different from those reported with several other markers such as zebrin II/aldolase C and the small heat shock protein HSP25. N-CAM null knockout mice were used to explore the possible role of the HNK-1/N-CAM interaction during the topographical organization of the cerebellar cortex. N-CAM null mice have no N-CAM immunoreactivity but otherwise the cerebellum appears morphologically normal. Further, in the N-CAM null HNK-1 immunoreactivity is abolished from Purkinje cell dendrites but is retained on Golgi cells and neurons of the cerebellar nuclei. Despite the absence of N-CAM/HNK-1, parasagittal stripes and transverse zones in the cerebellum as revealed by using zebrin II immunocytochemistry appear normal.     

Benzodiazepine Receptor Binding in Cerebellar Cortex: Observations in Olivopontocerebellar Atrophy

Benzodiazepine receptor binding was measured in cerebellar cortex of 15 patients with dominantly inherited olivopontocerebellar atrophy (OPCA). The majority of these patients had a moderate to marked Purkinje cell loss, as judged by the lowered levels of dentate nucleus gamma-aminobutyric acid (GABA), a marker of Purkinje cells. Despite the reduction in Purkinje cell number cerebellar cortical benzodiazepine receptor density was either normal or slightly elevated in the OPCA patients. These results are in contrast to the findings in a mutant strain of mice deficient in Purkinje cells in which the concentration of benzodiazepine receptors in cerebellum is greatly reduced. Our data indicate that in the human, cerebellar cortical benzodiazepine receptors are either not significantly associated with Purkinje cells or that in OPCA Purkinje cell loss triggers a de novo synthesis of extra benzodiazepine binding sites. It is concluded that, in contrast with the rodent, in the human benzodiazepine receptor binding may not serve as a marker for cerebellar Purkinje cells.     

Nogo-66 receptor at cerebellar cortical glia gap junctions in the rat

Nogo-A is a myelin inhibitor of neurite outgrowth that accounts for the difficulty in fiber regeneration in the central nervous system. Its 66-amino-acid extracellular domain (Nogo-66) contributes to the inhibitory activity of Nogo-A. The Nogo-66 receptor is widely distributed in neurons of the central nervous system, including the cerebellum. In our study on the distribution of Nogo-66 receptor in the cerebellar cortex in the rat, we unexpectedly found Nogo-66 receptor immunoreactivity in the glia cells, particularly abundant beneath the Purkinje cells. The presence of Nogo-66 receptor in glia cells has not been reported before. A detailed study was thus conducted. Immunoelectron microscopic investigation clearly demonstrated that the Nogo-66 receptor immunoreactivity could be ascertained at the gap junction between glia cells, indicating that the Nogo-66 receptor may modulate the communication between glia cells through gap junctions.     

Development of neurons containing acetylcholinesterase and cholinacetyltransferase in dispersed cell culture of rat cerebellum

Summary Cells from one-day-old cerebellum were grown for up to 30 days in dispersed cell culture. The characteristic neurons (deep cerebellar, Golgi and Purkinje cells) maintained their properties. It was found histochemically that some of the large cells display strong AChE activities in the perikaryon and in some processes, while biochemically the specific activities of the marker enzymes of the acetylcholine system, AChE (EC 3.1.1.7) and ChAc (EC 2.3.1.6), were increased and unchanged, respectively. During cultivation, the number of AChE-positive neurons increased. It can be inferred from these studies that, besides the AChE-positive (cholinoceptive) cells, ChAc-active (cholinergic) neurons (possibly Golgi II. type cells and some neurons in the deep cerebellar nuclei) are present in the cerebellum of the rat.     

Chromogranin A in neurons of the rat cerebellum and spinal cord: quantification and sites of expression.

Chromogranin A (CGA) is an abundant protein of dense-cored secretory vesicles in endocrine and neuronal cells. The present study, for the first time, compares CGA of neurons of the central nervous system with the CGA of adrenal origin. By S1 nucleus protection assay, we found that the 3' part of the CGA mRNA between exons 5-8 of the cerebellum and the spinal cord of the rat is homologous to that of the adrenal. In situ hybridization histochemistry revealed that CGA mRNA in the cerebellar cortex is present in cell bodies of Purkinje cells and in neurons of the deep cerebellar nuclei. The perikarya of these cells also exhibit CGA-like immunoreactivity. CGA mRNA and CGA-like immunoreactivity are also present in the motoneurons of the ventral, lateral, and dorsal horns of the rat spinal cord. The amounts of CGA, as determined by radioimmunoassay in cerebellum and spinal cord, were about one tenth of the amounts detected in the adrenal, adenohypophysis, or the olfactory bulb. The sites of CGA expression suggest that CGA may be involved in signal transduction in the motor system.     

Localization of the type 1 corticotropin releasing factor receptor (CRF-R1) in the embryonic mouse cerebellum

Corticotropin releasing factor (CRF) is present in the adult, as well as in the embryonic and postnatal rodent cerebellum. Further, the distribution of the type 1 CRF receptor has been described in adult and postnatal animals. The focus of the present study is to determine the distribution and cellular relationships of the type 1 CRF receptor (CRF-R1) during embryonic development of the cerebellum. Between embryonic day (E)11 and E12, CRF-R1 immunoreactive puncta are uniformly distributed in the ventricular zone, the site of origin of Purkinje cells, nuclear neurons, and GABAergic interneurons, as well as the germinal trigone, the birthplace of the precursors of granule cells. Between E13 and 18, the distribution of immunolabeled puncta decreases in both the ventricular zone and the germinal trigone and increases in the intermediate zone, as well as in the dorsal aspect of the cerebellar plate. Between E14 and 18, antibodies that label specific populations of cerebellar neurons were combined with the antibody for the receptor to determine the cellular elements that expressed CRF-R1. At E14, CRF-R1 immunoreactivity is co-localized in neurons immunolabeled with PAX-2, an antibody that is specific for GABAergic interneurons. These neurons continue to express CRF-R1 as they migrate dorsally toward the cerebellar surface. Between E16 and 18, Purkinje cells, immunolabeled with calbindin, near the dorsal surface of the cerebellum express CRF-R1 in their cell bodies and apical processes. CRF has been shown to have a depolarizing effect on adult and postnatal Purkinje cells. Further, CRF has been shown to contribute to excitability of hippocampal neurons during embryonic development by binding to CRF-R1; depolarization induced excitability appears to be critical for cell survival. The location of the type one CRF receptor and the presence of its primary ligand, CRF, in the germinal zones of the cerebellum and in migrating neurons suggest that this receptor/ligand interaction could be important in the regulation of neuronal survival through cellular mechanisms that lead to depolarization of embryonic cerebellar neurons.     

Development of neurons containing acetylcholinesterase and cholinacetyltransferase in dispersed cell culture of rat cerebellum.

Cells from one-day-old cerebellum were grown for up to 30 days in dispersed cell culture. The characteristic neurons (deep cerebellar, Golgi and Purkinje cells) maintained their properties. It was found histochemically that some of the large cells display strong AChE activities in the perikaryon and in some processes, while biochemically the specific activities of the marker enzymes of the acetylcholine system, AChE (EC 3.1.1.7) and ChAc (EC 2.3.1.6), were increased and unchanged, respectively. During cultivation, the number of AChE-positive neurons increased. It can be inferred from these studies that, besides the AChE-positive (cholinoceptive) cells, ChAc-active (cholinergic) neurons (possibly Golgi II. type cells and some neurons in the deep cerebellar nuclei) are present in the cerebellum of the rat.     

Regional and cellular distribution of neural visinin-like protein immunoreactivities (VILIP-1 and VILIP-3) in human brain

Neural visinin-like proteins (VILIPs) are members of the neuronal subfamily of intracellular EF-hand calcium sensor proteins termed the NCS family, which are thought to play important roles in cellular signal transduction. While numerous studies suggest a wide but uneven distribution of these proteins in rat and chicken brain, their location in, and possible significance for, the human brain, remains to be established. We used specific polyclonal antisera to map the human brain for VILIP-1 and VILIP-3 immunoreactivities. VILIP-1 was detected in cortical pyramidal cells and interneurons, septal, subthalamic and hippocampal neurons (subfields CA1 and CA4 pyramidal cells and especially hilar interneurons) as well as in cerebellar Golgi, basket, granule, stellate and dentate nucleus neurons. Purkinje cells were free of immunoreaction. VILIP-3 was more restricted in its distribution. It was identified in cerebellar Purkinje cells and a subpopulation of granule neurons. Further, neurons belonging to different nuclei of the brain stem and multiple subcortical nerve cells stained for visinin-like protein 3. A weak immunoreaction appeared in cortical and hippocampal neurons. Intracellularly the immunoreactivity appeared in the perikarya, dendrites and some axons. Sometimes, immunostaining was found in the neuropil. Glia did not express visinin-like proteins. Our findings support, from a neuroanatomical viewpoint, the idea that these calcium sensor proteins may be of relevance for neuronal signalling in the human CNS.     

The Reelin receptors Apoer2 and Vldlr coordinate the patterning of Purkinje cell topography in the developing mouse cerebellum

The adult cerebellar cortex is comprised of reproducible arrays of transverse zones and parasagittal stripes of Purkinje cells. Adult stripes are created through the perinatal rostrocaudal dispersion of embryonic Purkinje cell clusters, triggered by signaling through the Reelin pathway. Reelin is secreted by neurons in the external granular layer and deep cerebellar nuclei and binds to two high affinity extracellular receptors on Purkinje cells-the Very low density lipoprotein receptor (Vldlr) and apolipoprotein E receptor 2 (Apoer2). In mice null for either Reelin or double null for Vldlr and Apoer2, Purkinje cell clusters fail to disperse. Here we report that animals null for either Vldlr or Apoer2 individually, exhibit specific and parasagittally-restricted Purkinje cell ectopias. For example, in mice lacking Apoer2 function immunostaining reveals ectopic Purkinje cells that are largely restricted to the zebrin II-immunonegative population of the anterior vermis. In contrast, mice null for Vldlr have a much larger population of ectopic Purkinje cells that includes members from both the zebrin II-immunonegative and -immunopositive phenotypes. HSP25 immunoreactivity reveals that in Vldlr null animals a large portion of zebrin II-immunopositive ectopic cells are probably destined to become stripes in the central zone (lobules VI-VII). A small population of ectopic zebrin II-immunonegative Purkinje cells is also observed in animals heterozygous for both receptors (Apoer2(+/-): Vldlr(+/-)), but no ectopia is present in mice heterozygous for either receptor alone. These results indicate that Apoer2 and Vldlr coordinate the dispersal of distinct, but overlapping subsets of Purkinje cells in the developing cerebellum.     

Suppression of death receptor signaling in cerebellar Purkinje neurons protects neighboring granule neurons from apoptosis via an insulin-like growth factor I-dependent mechanism

Neuronal apoptosis contributes to the progression of neurodegenerative disease. Primary cerebellar granule neurons are an established in vitro model for investigating neuronal death. After removal of serum and depolarizing potassium, granule neurons undergo apoptosis via a mechanism that requires intrinsic (mitochondrial) death signals; however, the role of extrinsic (death receptor-mediated) signals is presently unclear. Here, we investigate involvement of death receptor signaling in granule neuron apoptosis by expressing adenoviral, AU1-tagged, dominant-negative Fas-associated death domain (Ad-AU1-deltaFADD). Ad-AU1-deltaFADD decreased apoptosis of granule neurons from 65 +/- 5 to 27 +/- 2% (n = 7, p < 0.01). Unexpectedly, immunocytochemical staining for AU1 revealed that <5% of granule neurons expressed deltaFADD. In contrast, deltaFADD was expressed in >95% of calbindin-positive Purkinje neurons ( approximately 2% of the cerebellar culture). Granule neurons in proximity to deltaFADD-expressing Purkinje cells demonstrated markedly increased survival. Both granule and Purkinje neurons expressed insulin-like growth factor-I (IGF-I) receptors, and deltaFADD-mediated survival of granule neurons was inhibited by an IGF-I receptor blocking antibody. These results demonstrate that the selective suppression of death receptor signaling in Purkinje neurons is sufficient to rescue neighboring granule neurons that depend on Purkinje cell-derived IGF-I. Thus, the extrinsic death pathway has a profound but indirect effect on the survival of cerebellar granule neurons.