Electrical synapses in mammalian CNS: Past eras, present focus and future directions.

Gap junctions provide the basis for electrical synapses between neurons. Early studies in well-defined circuits in lower vertebrates laid the foundation for understanding various properties conferred by electrical synaptic transmission. Knowledge surrounding electrical synapses in mammalian systems unfolded first with evidence indicating the presence of gap junctions between neurons in various brain regions, but with little appreciation of their functional roles. Beginning at about the turn of this century, new approaches were applied to scrutinize electrical synapses, revealing the prevalence of neuronal gap junctions, the connexin protein composition of many of those junctions, and the myriad diverse neural systems in which they occur in the mammalian CNS. Subsequent progress indicated that electrical synapses constitute key elements in synaptic circuitry, govern the collective activity of ensembles of electrically coupled neurons, and in part orchestrate the synchronized neuronal network activity and rhythmic oscillations that underlie fundamental integrative processes. This article is part of a Special Issue entitled: Gap Junction Proteins edited by Jean Claude Herve.

KV1 channels identified in rodent myelinated axons, linked to Cx29 in innermost myelin: support for electrically active myelin in mammalian saltatory conduction.

Saltatory conduction in mammalian myelinated axons was thought to be well understood before recent discoveries revealed unexpected subcellular distributions and molecular identities of the K(+)-conductance pathways that provide for rapid axonal repolarization. In this study, we visualize, identify, localize, quantify, and ultrastructurally characterize axonal KV1.1/KV1.2 channels in sciatic nerves of rodents. With the use of light microscopic immunocytochemistry and freeze-fracture replica immunogold labeling electron microscopy, KV1.1/KV1.2 channels are localized to three anatomically and compositionally distinct domains in the internodal axolemmas of large myelinated axons, where they form densely packed "rosettes" of 9-nm intramembrane particles. These axolemmal KV1.1/KV1.2 rosettes are precisely aligned with and ultrastructurally coupled to connexin29 (Cx29) channels, also in matching rosettes, in the surrounding juxtaparanodal myelin collars and along the inner mesaxon. As >98% of transmembrane proteins large enough to represent ion channels in these specialized domains, ∼500,000 KV1.1/KV1.2 channels define the paired juxtaparanodal regions as exclusive membrane domains for the voltage-gated K(+)conductance that underlies rapid axonal repolarization in mammals. The 1:1 molecular linkage of KV1 channels to Cx29 channels in the apposed juxtaparanodal collars, plus their linkage to an additional 250,000-400,000 Cx29 channels along each inner mesaxon in every large-diameter myelinated axon examined, supports previously proposed K(+)conductance directly from juxtaparanodal axoplasm into juxtaparanodal myeloplasm in mammalian axons. With neither Cx29 protein nor myelin rosettes detectable in frog myelinated axons, these data showing axon-to-myelin linkage by abundant KV1/Cx29 channels in rodent axons support renewed consideration of an electrically active role for myelin in increasing both saltatory conduction velocity and maximum propagation frequency in mammalian myelinated axons.

Trafficking of gap junction channels at a vertebrate electrical synapse in vivo.

Trafficking and turnover of transmitter receptors required to maintain and modify the strength of chemical synapses have been characterized extensively. In contrast, little is known regarding trafficking of gap junction components at electrical synapses. By combining ultrastructural and in vivo physiological analysis at identified mixed (electrical and chemical) synapses on the goldfish Mauthner cell, we show here that gap junction hemichannels are added at the edges of GJ plaques where they dock with hemichannels in the apposed membrane to form cell-cell channels and, simultaneously, that intact junctional regions are removed from centers of these plaques into either presynaptic axon or postsynaptic dendrite. Moreover, electrical coupling is readily modified by intradendritic application of peptides that interfere with endocytosis or exocytosis, suggesting that the strength of electrical synapses at these terminals is sustained, at least in part, by fast (in minutes) turnover of gap junction channels. A peptide corresponding to a region of the carboxy terminus that is conserved in Cx36 and its two teleost homologs appears to interfere with formation of new gap junction channels, presumably by reducing insertion of hemichannels on the dendritic side. Thus, our data indicate that electrical synapses are dynamic structures and that their channels are turned over actively, suggesting that regulated trafficking of connexons may contribute to the modification of gap junctional conductance.

Gap junction-mediated electrical transmission: regulatory mechanisms and plasticity.

The term synapse applies to cellular specializations that articulate the processing of information within neural circuits by providing a mechanism for the transfer of information between two different neurons. There are two main modalities of synaptic transmission: chemical and electrical. While most efforts have been dedicated to the understanding of the properties and modifiability of chemical transmission, less is still known regarding the plastic properties of electrical synapses, whose structural correlate is the gap junction. A wealth of data indicates that, rather than passive intercellular channels, electrical synapses are more dynamic and modifiable than was generally perceived. This article will discuss the factors determining the strength of electrical transmission and review current evidence demonstrating its dynamic properties. Like their chemical counterparts, electrical synapses can also be plastic and modifiable. This article is part of a Special Issue entitled: The Communicating junctions, roles and dysfunctions.

Electrical synaptic transmission in developing zebrafish: properties and molecular composition of gap junctions at a central auditory synapse.

In contrast to the knowledge of chemical synapses, little is known regarding the properties of gap junction-mediated electrical synapses in developing zebrafish, which provide a valuable model to study neural function at the systems level. Identifiable "mixed" (electrical and chemical) auditory synaptic contacts known as "club endings" on Mauthner cells (2 large reticulospinal neurons involved in tail-flip escape responses) allow exploration of electrical transmission in fish. Here, we show that paralleling the development of auditory responses, electrical synapses at these contacts become anatomically identifiable at day 3 postfertilization, reaching a number of ∼6 between days 4 and 9. Furthermore, each terminal contains ∼18 gap junctions, representing between 2,000 and 3,000 connexon channels formed by the teleost homologs of mammalian connexin 36. Electrophysiological recordings revealed that gap junctions at each of these contacts are functional and that synaptic transmission has properties that are comparable with those of adult fish. Thus a surprisingly small number of mixed synapses are responsible for the acquisition of auditory responses by the Mauthner cells, and these are likely sufficient to support escape behaviors at early developmental stages.

Connexin composition in apposed gap junction hemiplaques revealed by matched double-replica freeze-fracture replica immunogold labeling.

Despite the combination of light-microscopic immunocytochemistry, histochemical mRNA detection techniques and protein reporter systems, progress in identifying the protein composition of neuronal versus glial gap junctions, determination of the differential localization of their constituent connexin proteins in two apposing membranes and understanding human neurological diseases caused by connexin mutations has been problematic due to ambiguities introduced in the cellular and subcellular assignment of connexins. Misassignments occurred primarily because membranes and their constituent proteins are below the limit of resolution of light microscopic imaging techniques. Currently, only serial thin-section transmission electron microscopy and freeze-fracture replica immunogold labeling have sufficient resolution to assign connexin proteins to either or both sides of gap junction plaques. However, freeze-fracture replica immunogold labeling has been limited because conventional freeze fracturing allows retrieval of only one of the two membrane fracture faces within a gap junction, making it difficult to identify connexin coupling partners in hemiplaques removed by fracturing. We now summarize progress in ascertaining the connexin composition of two coupled hemiplaques using matched double-replicas that are labeled simultaneously for multiple connexins. This approach allows unambiguous identification of connexins and determination of the membrane "sidedness" and the identities of connexin coupling partners in homotypic and heterotypic gap junctions of vertebrate neurons.

The extent and strength of electrical coupling between inferior olivary neurons is heterogeneous.

Gap junctions constitute the only form of synaptic communication between neurons in the inferior olive (IO), which gives rise to the climbing fibers innervating the cerebellar cortex. Although its exact functional role remains undetermined, electrical coupling was shown to be necessary for the transient formation of functional compartments of IO neurons and to underlie the precise timing of climbing fibers required for cerebellar learning. So far, most functional considerations assume the existence of a network of permanently and homogeneously coupled IO neurons. Contrasting this notion, our results indicate that coupling within the IO is highly variable. By combining tracer-coupling analysis and paired electrophysiological recordings, we found that individual IO neurons could be coupled to a highly variable number of neighboring neurons. Furthermore, a given neuron could be coupled at remarkably different strengths with each of its partners. Freeze-fracture analysis of IO glomeruli revealed the close proximity of glutamatergic postsynaptic densities to connexin 36-containing gap junctions, at distances comparable to separations between chemical transmitting domains and gap junctions in goldfish mixed contacts, where electrical coupling was shown to be modulated by the activity of glutamatergic synapses. On the basis of structural and molecular similarities with goldfish mixed synapses, we speculate that, rather than being hardwired, variations in coupling could result from glomerulus-specific long-term modulation of gap junctions. This striking heterogeneity of coupling might act to finely influence the synchronization of IO neurons, adding an unexpected degree of complexity to olivary networks.

Connexin45-containing neuronal gap junctions in rodent retina also contain connexin36 in both apposing hemiplaques, forming bihomotypic gap junctions, with scaffolding contributed by zonula occludens-1.

Mammalian retinas contain abundant neuronal gap junctions, particularly in the inner plexiform layer (IPL), where the two principal neuronal connexin proteins are Cx36 and Cx45. Currently undetermined are coupling relationships between these connexins and whether both are expressed together or separately in a neuronal subtype-specific manner. Although Cx45-expressing neurons strongly couple with Cx36-expressing neurons, possibly via heterotypic gap junctions, Cx45 and Cx36 failed to form functional heterotypic channels in vitro. We now show that Cx36 and Cx45 coexpressed in HeLa cells were colocalized in immunofluorescent puncta between contacting cells, demonstrating targeting/scaffolding competence for both connexins in vitro. However, Cx36 and Cx45 expressed separately did not form immunofluorescent puncta containing both connexins, supporting lack of heterotypic coupling competence. In IPL, 87% of Cx45-immunofluorescent puncta were colocalized with Cx36, supporting either widespread heterotypic coupling or bihomotypic coupling. Ultrastructurally, Cx45 was detected in 9% of IPL gap junction hemiplaques, 90-100% of which also contained Cx36, demonstrating connexin coexpression and cotargeting in virtually all IPL neurons that express Cx45. Moreover, double replicas revealed both connexins in separate domains mirrored on both sides of matched hemiplaques. With previous evidence that Cx36 interacts with PDZ1 domain of zonula occludens-1 (ZO-1), we show that Cx45 interacts with PDZ2 domain of ZO-1, and that Cx36, Cx45, and ZO-1 coimmunoprecipitate, suggesting that ZO-1 provides for coscaffolding of Cx45 with Cx36. These data document that in Cx45-expressing neurons of IPL, Cx45 is almost always accompanied by Cx36, forming "bihomotypic" gap junctions, with Cx45 structurally coupling to Cx45 and Cx36 coupling to Cx36.

On the occurrence and enigmatic functions of mixed (chemical plus electrical) synapses in the mammalian CNS.

Electrical synapses with diverse configurations and functions occur at a variety of interneuronal appositions, thereby significantly expanding the physiological complexity of neuronal circuitry over that provided solely by chemical synapses. Gap junctions between apposed dendritic and somatic plasma membranes form "purely electrical" synapses that allow for electrical communication between coupled neurons. In addition, gap junctions at axon terminals synapsing on dendrites and somata allow for "mixed" (dual chemical+electrical) synaptic transmission. "Dual transmission" was first documented in the autonomic nervous system of birds, followed by its detection in the central nervous systems of fish, amphibia, and reptiles. Subsequently, mixed synapses have been detected in several locations in the mammalian CNS, where their properties and functional roles remain undetermined. Here, we review available evidence for the presence, complex structural composition, and emerging functional properties of mixed synapses in the mammalian CNS.

Oligodendrocytes control potassium accumulation in white matter and seizure susceptibility.

The inwardly rectifying K+ channel Kir4.1 is broadly expressed by CNS glia and deficits in Kir4.1 lead to seizures and myelin vacuolization. However, the role of oligodendrocyte Kir4.1 channels in controlling myelination and K+ clearance in white matter has not been defined. Here, we show that selective deletion of Kir4.1 from oligodendrocyte progenitors (OPCs) or mature oligodendrocytes did not impair their development or disrupt the structure of myelin. However, mice lacking oligodendrocyte Kir4.1 channels exhibited profound functional impairments, including slower clearance of extracellular K+ and delayed recovery of axons from repetitive stimulation in white matter, as well as spontaneous seizures, a lower seizure threshold, and activity-dependent motor deficits. These results indicate that Kir4.1 channels in oligodendrocytes play an important role in extracellular K+ homeostasis in white matter, and that selective loss of this channel from oligodendrocytes is sufficient to impair K+ clearance and promote seizures.

Electrical Synapses: New Rules for Assembling an Old Structure Asymmetrically.

Electrical synapses are emerging as complex structures, consisting of gap junction-forming connexin proteins and also multiple scaffolding and regulatory protein components, which were assumed to be symmetrically organized across these synapses; however, new findings reveal their synaptic asymmetry.

Connexin32-containing gap junctions in Schwann cells at the internodal zone of partial myelin compaction and in Schmidt-Lanterman incisures.

In vertebrate peripheral nerves, the insulating myelin sheath is formed by Schwann cells, which generate flattened membrane processes that spiral around axons and form compact myelin by extrusion of cytoplasm and adhesion of apposed intracellular and extracellular membrane surfaces. Cytoplasm remains within the innermost and outermost tongues, in the paranodal loops bordering nodes of Ranvier and in Schmidt-Lanterman incisures. By immunocytochemistry, connexin32 (Cx32) protein has been demonstrated at paranodal loops and Schmidt-Lanterman incisures, and it is widely assumed that gap junctions are present in these locations, thereby providing a direct radial route for transport of ions and metabolites between cytoplasmic myelin layers. This study used freeze-fracture replica immunogold labeling to detect Cx32 in ultrastructurally defined gap junctions in Schmidt-Lanterman incisures, as well as in a novel location, between the outer two layers of internodal myelin, approximately every micrometer along the entire length of myelin, at the zone between compact myelin and noncompact myelin. Thus, these gap junctions link the partially compacted second layer of myelin to the noncompact outer tongue. Although these gap junctions are unusually small (average, 11 connexon channels), their relative abundance and regular distribution along the zone that is structurally intermediate between compact and noncompact myelin demonstrates the existence of multiple sites for unidirectional or bidirectional transport of water, ions, and small molecules between these two distinct cytoplasmic compartments, possibly to regulate or facilitate myelin compaction or to maintain the transition zone between noncompact and compact myelin.

Connexin29 and connexin32 at oligodendrocyte and astrocyte gap junctions and in myelin of the mouse central nervous system.

The cellular localization, relation to other glial connexins (Cx30, Cx32, and Cx43), and developmental expression of Cx29 were investigated in the mouse central nervous system (CNS) with an anti-Cx29 antibody. Cx29 was enriched in subcellular fractions of myelin, and immunofluorescence for Cx29 was localized to oligodendrocytes and myelinated fibers throughout the brain and spinal cord. Oligodendrocyte somata displayed minute Cx29-immunopositive puncta around their periphery and intracellularly. In developing brain, Cx29 levels increased during the first few postnatal weeks and were highest in the adult brain. Immunofluorescence labeling for Cx29 in oligodendrocyte somata was intense at young ages and was dramatically shifted in localization primarily to myelinated fibers in mature CNS. Labeling for Cx32 also was localized to oligodendrocyte somata and myelin and absent in Cx32 knockout mice. Cx29 and Cx32 were minimally colocalized on oligodendrocytes somata and partly colocalized along myelinated fibers. At gap junctions on oligodendrocyte somata, Cx43/Cx32 and Cx30/Cx32 were strongly associated, but there was minimal association of Cx29 and Cx43. Cx32 was very sparsely associated with astrocytic connexins along myelinated fibers. With Cx26, Cx30, and Cx43 expressed in astrocytes and Cx29, Cx32, and Cx47 expressed in oligodendrocytes, the number of connexins localized to gap junctions of glial cells is increased to six. The results suggested that Cx29 in mature CNS contributes minimally to gap junctional intercellular communication in oligodendrocyte cell bodies but rather is targeted to myelin, where it, with Cx32, may contribute to connexin-mediated communication between adjacent layers of uncompacted myelin.

Astrocyte and oligodendrocyte connexins of the glial syncytium in relation to astrocyte anatomical domains and spatial buffering.

Astroctyes express a set of three connexins (Cx26, Cx30, and Cx43) that are contained in astrocyte-to-astrocyte (A/A) gap junctions; oligodendrocytes express a different set of three connexins (Cx29, Cx32, and Cx47) that are contained in the oligodendrocyte side of necessarily heterotypic astrocyte-to-oligodendrocyte (A/O) gap junctions, and there is little ultrastructural evidence for gap junction formation between individual oligodendrocytes. In addition, primarily Cx29 and Cx32 are contained deeper in myelin sheaths, where they form autologous gap junctions at sites of uncompacted myelin. The presence of six connexins in macroglial cell populations has revealed unprecedented complexity of potential connexin coupling partners, and with restricted deployment of gap junctional intercellular communication (GJIC) within the "pan-glial" syncytium. New implications for the organization and regulation of spatial buffering mediated by glial GJIC are derived from recent observations of the existence of separate astrocyte anatomical domains, with only narrow regions of overlap between astrocyte processes at domain borders. Thus, widespread spatial buffering in the CNS may occur not successively through a multitude of processes arising from different astrocytes, but rather in a more orderly fashion from one astrocyte domain to another via intercellular coupling that occurs only at restricted regions of overlap between astrocyte domains, augmented by autocellular coupling that occurs within each domain.

Abundance of gap junctions at glutamatergic mixed synapses in adult Mosquitofish spinal cord neurons.

"Dye-coupling", whole-mount immunohistochemistry for gap junction channel protein connexin 35 (Cx35), and freeze-fracture replica immunogold labeling (FRIL) reveal an abundance of electrical synapses/gap junctions at glutamatergic mixed synapses in the 14th spinal segment that innervates the adult male gonopodium of Western Mosquitofish, Gambusia affinis (Mosquitofish). To study gap junctions' role in fast motor behavior, we used a minimally-invasive neural-tract-tracing technique to introduce gap junction-permeant or -impermeant dyes into deep muscles controlling the gonopodium of the adult male Mosquitofish, a teleost fish that rapidly transfers (complete in <20 mS) spermatozeugmata into the female reproductive tract. Dye-coupling in the 14th spinal segment controlling the gonopodium reveals coupling between motor neurons and a commissural primary ascending interneuron (CoPA IN) and shows that the 14th segment has an extensive and elaborate dendritic arbor and more gap junctions than do other segments. Whole-mount immunohistochemistry for Cx35 results confirm dye-coupling and show it occurs via gap junctions. Finally, FRIL shows that gap junctions are at mixed synapses and reveals that >50 of the 62 gap junctions at mixed synapses are in the 14th spinal segment. Our results support and extend studies showing gap junctions at mixed synapses in spinal cord segments involved in control of genital reflexes in rodents, and they suggest a link between mixed synapses and fast motor behavior. The findings provide a basis for studies of specific roles of spinal neurons in the generation/regulation of sex-specific behavior and for studies of gap junctions' role in regulating fast motor behavior. Finally, the CoPA IN provides a novel candidate neuron for future studies of gap junctions and neural control of fast motor behaviors.

Molecular and functional asymmetry at a vertebrate electrical synapse.

Electrical synapses are abundant in the vertebrate brain, but their functional and molecular complexities are still poorly understood. We report here that electrical synapses between auditory afferents and goldfish Mauthner cells are constructed by apposition of hemichannels formed by two homologs of mammalian connexin 36 (Cx36) and that, while Cx35 is restricted to presynaptic hemiplaques, Cx34.7 is restricted to postsynaptic hemiplaques, forming heterotypic junctions. This molecular asymmetry is associated with rectification of electrical transmission that may act to promote cooperativity between auditory afferents. Our data suggest that, in similarity to pre- and postsynaptic sites at chemical synapses, one side in electrical synapses should not necessarily be considered the mirror image of the other. While asymmetry based on the presence of two Cx36 homologs is restricted to teleost fish, it might also be based on differences in posttranslational modifications of individual connexins or in the complement of gap junction-associated proteins.

Gap junction assembly: roles for the formation plaque and regulation by the C-terminus of connexin43.

Using an established gap junction (GJ) assembly system with experimentally reaggregated cells, we analyzed "formation plaques" (FPs), apparent sites of GJ assembly. Employing freeze-fracture electron microscopy methods combined with filipin labeling of sterols and immunolabeling for connexin43 (Cx43), we demonstrated that FPs constitute distinct membrane "domains" and that their characteristic 10-nm particles contain connexin43, thus representing precursors (i.e., GJ hemichannels) engaged in assembly. Analysis of FPs in new systems-HeLa and N2A cells-resolved questions surrounding several key but poorly understood steps in assembly, including matching of FP membranes in apposed cells, reduction in the separation between FP membranes during assembly, and the process of particle aggregation. Findings also indicated that "docking" of GJ hemichannels occurs within FP domains and contributes to reduction of intermembrane separation between FPs. Other experiments demonstrated that FPs develop following a major C-terminal truncation of Cx43 (M257), although assembly was delayed. Particle aggregation also occurred at lower densities, and densities of particles within developing GJ aggregates failed to achieve full-length levels. With regard to regulation, inhibition of assembly following protein kinase C activation failed to occur in the M257 truncation mutants, as measured by intercellular dye transfer. However, several C-terminal serine mutations failed to disrupt inhibition.

Mixed Electrical-Chemical Synapses in Adult Rat Hippocampus are Primarily Glutamatergic and Coupled by Connexin-36.

Dendrodendritic electrical signaling via gap junctions is now an accepted feature of neuronal communication in mammalian brain, whereas axodendritic and axosomatic gap junctions have rarely been described. We present ultrastructural, immunocytochemical, and dye-coupling evidence for "mixed" (electrical/chemical) synapses on both principal cells and interneurons in adult rat hippocampus. Thin-section electron microscopic images of small gap junction-like appositions were found at mossy fiber (MF) terminals on thorny excrescences of CA3 pyramidal neurons (CA3pyr), apparently forming glutamatergic mixed synapses. Lucifer Yellow injected into weakly fixed CA3pyr was detected in MF axons that contacted four injected CA3pyr, supporting gap junction-mediated coupling between those two types of principal cells. Freeze-fracture replica immunogold labeling revealed diverse sizes and morphologies of connexin-36-containing gap junctions throughout hippocampus. Of 20 immunogold-labeled gap junctions, seven were large (328-1140 connexons), three of which were consistent with electrical synapses between interneurons; but nine were at axon terminal synapses, three of which were immediately adjacent to distinctive glutamate receptor-containing postsynaptic densities, forming mixed glutamatergic synapses. Four others were adjacent to small clusters of immunogold-labeled 10-nm E-face intramembrane particles, apparently representing extrasynaptic glutamate receptor particles. Gap junctions also were on spines in stratum lucidum, stratum oriens, dentate gyrus, and hilus, on both interneurons and unidentified neurons. In addition, one putative GABAergic mixed synapse was found in thin-section images of a CA3pyr, but none were found by immunogold labeling, suggesting the rarity of GABAergic mixed synapses. Cx36-containing gap junctions throughout hippocampus suggest the possibility of reciprocal modulation of electrical and chemical signals in diverse hippocampal neurons.

Gap junctions on hippocampal mossy fiber axons demonstrated by thin-section electron microscopy and freeze fracture replica immunogold labeling.

Gap junctions have been postulated to exist between the axons of excitatory cortical neurons based on electrophysiological, modeling, and dye-coupling data. Here, we provide ultrastructural evidence for axoaxonic gap junctions in dentate granule cells. Using combined confocal laser scanning microscopy, thin-section transmission electron microscopy, and grid-mapped freeze-fracture replica immunogold labeling, 10 close appositions revealing axoaxonic gap junctions ( approximately 30-70 nm in diameter) were found between pairs of mossy fiber axons ( approximately 100-200 nm in diameter) in the stratum lucidum of the CA3b field of the rat ventral hippocampus, and one axonal gap junction ( approximately 100 connexons) was found on a mossy fiber axon in the CA3c field of the rat dorsal hippocampus. Immunogold labeling with two sizes of gold beads revealed that connexin36 was present in that axonal gap junction. These ultrastructural data support computer modeling and in vitro electrophysiological data suggesting that axoaxonic gap junctions play an important role in the generation of very fast (>70 Hz) network oscillations and in the hypersynchronous electrical activity of epilepsy.

Temporary loss of perivascular aquaporin-4 in neocortex after transient middle cerebral artery occlusion in mice.

The aquaporin-4 (AQP4) pool in the perivascular astrocyte membranes has been shown to be critically involved in the formation and dissolution of brain edema. Cerebral edema is a major cause of morbidity and mortality in stroke. It is therefore essential to know whether the perivascular pool of AQP4 is up- or down-regulated after an ischemic insult, because such changes would determine the time course of edema formation. Here we demonstrate by quantitative immunogold cytochemistry that the ischemic striatum and neocortex show distinct patterns of AQP4 expression in the reperfusion phase after 90 min of middle cerebral artery occlusion. The striatal core displays a loss of perivascular AQP4 at 24 hr of reperfusion with no sign of subsequent recovery. The most affected part of the cortex also exhibits loss of perivascular AQP4. This loss is of magnitude similar to that of the striatal core, but it shows a partial recovery toward 72 hr of reperfusion. By freeze fracture we show that the loss of perivascular AQP4 is associated with the disappearance of the square lattices of particles that normally are distinct features of the perivascular astrocyte membrane. The cortical border zone differs from the central part of the ischemic lesion by showing no loss of perivascular AQP4 at 24 hr of reperfusion but rather a slight increase. These data indicate that the size of the AQP4 pool that controls the exchange of fluid between brain and blood during edema formation and dissolution is subject to large and region-specific changes in the reperfusion phase.