The Journal of Neurophysiology Vol. 79 No. 2 February 1998, pp. 1045-1052
Copyright ©1998 by the American Physiological Society

Spiral Intercellular Calcium Waves in Hippocampal Slice Cultures

Marni E. Harris-White¹, Stephen A. Zanotti², Sally A. Frautschy¹, and Andrew C. Charles²

¹ Department of Medicine, UCLA/Veterans Affairs Medical Center, Sepulveda, CA 91343; and ² Department of Neurology, UCLA School of Medicine, Los Angeles, California 90095

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Harris-White, Marni E., Stephen A. Zanotti, Sally A. Frautschy, and Andrew C. Charles. Spiral intercellular calcium waves in hippocampal slice cultures. J. Neurophysiol. 79: 1045-1052, 1998. Complex patterns of intercellular calcium signaling occur in the CA1 and CA2 regions of hippocampal slice organotypic cultures from neonatal mice. Spontaneous localized intercellular Ca²⁺ waves involving 5-15 cells propagate concentrically from multiple foci in the stratum oriens and s. radiatum. In these same regions, extensive Ca²⁺ waves involving hundreds of cells propagate as curvilinear and spiral wavefronts across broad areas of CA1 and CA2. Ca²⁺ waves travel at rates of 5-10 µm/s, are abolished by thapsigargin, and do not require extracellular Ca²⁺. Staining for astrocytes and neurons indicates that these intercellular waves occur primarily in astrocytes. The frequency and amplitude of Ca²⁺ waves increase in response to bath application of N-methyl-D-aspartate (NMDA) and decrease in response to removal of extracellular Ca²⁺ or application of tetrodotoxin. This novel pattern of intercellular Ca²⁺ signaling is characteristic of the behavior of an excitable medium. Networks of glial cells in the hippocampus may behave as an excitable medium whose spatial and temporal signaling properties are modulated by neuronal activity.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The study of cellular activity in the hippocampus has focused primarily on synaptic neuronal activity. Recently, however, nonsynaptic signaling pathways including glial-glial and bidirectional glial-neuronal signaling have received increasing attention as potential mediators of hippocampal function (Dani et al. 1992; Duffy and MacVicar 1995; Mennerick and Zorumski 1994; Porter and McCarthy 1996; Wu and Barish 1994). These nonsynaptic signaling pathways may modulate traditional synaptic neuronal signaling or play independent roles in the function of the nervous system. Intercellular Ca²⁺ signaling has been observed in hippocampal astrocytes in culture and in hippocampal slice cultures in response to chemical and electrical stimuli (Cornell-Bell et al. 1990; Dani et al. 1992; Duffy and MacVicar 1995). In this study we report novel patterns of organized Ca²⁺ signaling, including spiral waves, involving broad areas of the hippocampus.

Spiral waves are a recurring phenomenon in a wide diversity of natural systems (Block et al. 1996; Davidenko et al. 1992; Siegert and Weijer 1995; Zaritsky et al. 1993). They have been modeled extensively as a characteristic pattern of activity in excitable media (Diks et al. 1995; Steinbock et al. 1993; Winfree 1972). In the Xenopus oocyte, spiral waves of increased [Ca²⁺]_i propagate throughout the cytoplasm; investigation of these waves yielded new insights into the mechanisms of cellular Ca²⁺ signaling (Atri et al. 1993; Lechleiter et al. 1991; Lechleiter and Clapham 1992). We now report intercellular spiral waves in the hippocampus with properties similar to those observed in other systems such as the Xenopus oocyte.

	METHODS

Abstract Introduction Methods Results Discussion References

Hippocampal slice cultures

Hippocampal slice cultures were prepared from 6- to 7-day-old CD-1 mice (Stoppini et al. 1991). Animal care was in accordance with UCLA Animal Research Committee guidelines. Hippocampi were cut into 400-µm sections that were transferred to culture dishes containing Costar membrane inserts (0.4 µm). Slices were then maintained in culture for 2-10 days before experimentation. The culture medium was a mixture of Eagle's minimum essential medium (MEM) + N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES; 5%; Gibco), horse serum (25%; Sigma), Hanks' balanced salt solution (25%; Gibco), glucose (6.5 mg/ml), and penicillin-streptomycin (50 U/ml to 0.05 mg/ml), pH 7.2. Slices were maintained in 5% CO₂ in a humidified incubator at 37°. At the time of the experiments, slices had flattened out to 150-200 µm thick, comprising ~10 cell layers. Serial cryostat sections and staining with glial fibrillary acidic protein (GFAP), neuronal cell surface antigen (NCSA), and microtubule associated protein-5 (MAP-5) (see Immunohistochemistry) showed that slices maintained normal hippocampal cellular architecture throughout all ages in culture. Only the far edges of the slice thinned out to <10 cell layers; these areas were not imaged. The cellular composition was consistent throughout the thickness of the slice; there was no glial cell layer covering the surface of the slice as verified by cryostat sections.

Experimental procedures

After 2-10 days in culture, slices were submersion loaded in 50 µM fura2-AM (Calbiochem) for 1 h at room temperature in a humidified and oxygenated chamber. The membranes with the slices attached were then excised from the plastic insert, glued to a glass coverslip (ROSS super glue gel was applied around the edge of the membrane; glue hardened immediately on contact with liquid and never contacted the slice) to a glass coverslip, and placed on the stage of a Nikon inverted microscope. Slices were perfused at a rate of 2 ml/min at room temperature that was oxygenated at 95% O₂-5% CO₂ with artificial cerebrospinal fluid (ACFS) containing (in mM) 17 NaCl, 4 KCl, 2.5 MgSO₄, 2 CaCl₂, 1.25 KH₂PO₄, 10 glucose, 5 NaHCO₃, and 20 HEPES, pH 7.2. After fura loading, slices were perfused for 30 min before imaging. For experiments in which O₂ was removed from the perfusion buffer (including 95% N₂-5% CO₂ replacement), the pH was adjusted to 7.2. The pH was constant throughout all experiments. All agonists and antagonists were added to the perfusion medium. Glutamate agonists and antagonists were obtained from Tocris Cooksen. Culture media was obtained from Cellgro. Unless otherwise specified, all other reagents were obtained from Sigma.

Measurement of [Ca²⁺]_i

[Ca²⁺]_i was measured with a fluorescence imaging system that was previously described in detail (Charles et al. 1991). Briefly, slices were excited with a mercury lamp through 340- and 380-nm band-pass filters; fluorescence at 510 nm was recorded through a ×20, 0.75 numerical aperture objective on a Nikon Diaphot inverted microscope with a Silicon intensified target camera to an optical memory disk recorder. Images were then digitized and subjected to background subtraction and shading correction, after which [Ca²⁺]_i was calculated on a pixel-by-pixel basis by a frame grabber and image analysis board (Data Translation) as previously described (Charles et al. 1991). Tracings in Figs. 2C, 3C, and 4 were based on fluorescence of a 4 × 4 pixel area located within each cell body. Layers of cell bodies within the slice were clearly visualized with the system described, although the resolution was not sufficient to delineate the entire extent of each cell. Background fluorescence was estimated by measuring the fluorescence of an area of the coverslip adjacent to the slice. Slices not loaded with fura showed no fluorescence under our experimental conditions. The entire thickness of the slice appeared to be loaded with fura, and fluorescence from cell bodies of at least three layers of cells could be distinguished in each imaged field as determined by focusing up and down through the field. However, our calculation of [Ca²⁺]_i did not account for additional out-of-focus and background fluorescence within the slice, and therefore it was possible that actual changes in [Ca²⁺]_i were greater than those reported. We have reported our measurements as "estimated [Ca²⁺]_i."

View larger version (75K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   FIG. 2. Spontaneous Ca2+ puffs and the response to N-methyl-D-aspartate (NMDA). Top left: region of the slice explant that is imaged (box). Top right: baseline fura2 fluorescence in this region. A: spontaneous Ca2+ puffs that originate from multiple foci in the s. oriens. B: same field 10 s following bath application of 20 µM NMDA. A large spiral wave develops and propagates to involve most cells in the field. Scale: 550 µm horizontally × 400 µm vertically. C: tracings of estimated intracellular Ca2+ ([Ca2+]i) vs. time in 3 representative cells in the field shown in A and B. Spontaneous asynchronous Ca2+ puffs occur before addition of NMDA, after which 2 broad intercellular wavefronts pass through all cells. P, Ca2+ puffs; W, spiral Ca2+ waves.

View larger version (89K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   FIG. 3. Spontaneous spiral wave and annihilation of colliding waves. Top left: region of the slice explant that is imaged (box). Top right: baseline fura2 fluorescence of this region. This slice is an example of the ~2% of slices showing spontaneous broad intercellular wavefronts. A: initiation of a spiral intercellular Ca2+ wave in a single cell. B: same field later in the recording, when 2 curvilinear Ca2+ wavefronts collide and annihilate. Arrows in schematic panels to the right of B indicate direction of propagation of the waves. Scale: 55 µm horizontally × 4 µm vertically. C: tracings of estimated [Ca2+]i vs. time in 3 representative cells in the field. The intercellular waves shown in A and B are indicated by the letters above the appropriate peak on the tracing. Cell 1 shows periodic Ca2+ waves, whereas cells 2 and 3 show a less regular pattern of Ca2+ wave occurrence.

View larger version (23K): [in this window] [in a new window]   FIG. 4. Effect of removal of extracellular Ca2+ ([Ca2+]o). Tracings show estimated [Ca2+]i vs. time in 2 representative cells before and after replacement of normal experimental medium with medium containing no added Ca2+ and 1 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N,N-tetraacetic acid. Single cell Ca2+ oscillations in cell 2 are abolished by removal of [Ca2+]o. [Ca2+]i waves continue in cell 1 after removal of [Ca2+]o but their frequency is reduced. O, single-cell Ca2+ oscillations; P, Ca2+ puffs; W, broad intercellular waves.

Neuron cell surface antigen labeling

After fura2 measurements, stage coordinates for the imaged fields were obtained and then slices were removed and treated with antibody to the NCSA (Chemicon; 1/100) for 45 min. Slices were rinsed three times and exposed to a Texas red-tagged secondary antibody for 30 min. The slice was returned to previous image coordinates on the microscope stage and neuronal cell bodies were identified with a rhodamine filter set.

Immunohistochemistry

Slices were fixed in 4% paraformaldehyde fixative for 1 hr and rinsed in phosphate buffered saline. For cryostat sectioning, slices were cryoprotected with sucrose (successive 10, 15, and 20% solutions), sectioned at 10 µm on a Zeiss cryostat, and mounted onto gelatin-coated slides. Slide-mounted whole slices or sections were pretreated with a 0.3% hydrogen peroxide solution (30 min) to block endogenous peroxidase activity and subsequently blocked with a buffer containing bovine serum albumin, Tween20, and a serum appropriate for the particular antibody being used. Slices were incubated with either a GFAP (1:500) or MAP-5 (1:200) antibody for 1 h at room temperature followed by the appropriate biotinylated secondary antibody. Antibody binding was detected with the use of avidin-biotin-peroxidase (ABC Kit, Vector Labs) and the substrate 3,3-diaminobenzidiene.

	RESULTS

Abstract Introduction Methods Results Discussion References

The slice explant cultures used for these studies were prepared from 6- to 7-day-old CD-1 mice. These slice explants maintain the characteristic cellular architecture of the hippocampus for up to 2 wk as indicated by phase-contrast microscopy, loading of cells with fura2, and immunolabeling for neurons with MAP-5 and NCSA and for astrocytes with GFAP (Fig. 1 and other data not shown). Labeling for GFAP reveals a complex network of astrocyte cell bodies and processes throughout the slice that is particularly dense in the s. oriens and s. radiatum and more sparse in the pyramidal neuronal layer (Fig. 1). MAP-5 staining reveals neuronal cell bodies primarily in the pyramidal neuronal layer, with some cell bodies and extensive thin dendritic fibers in the s. oriens and straum radiatum. Because our imaging system uses a standard inverted microscope, fura2 fluorescence is collected from multiple vertical layers of cells in each field. We are able to resolve many cell bodies, but not the individual cellular processes that form a complex network throughout the slice. A disadvantage of this limited resolution is the inability to definitively identify individual cells involved in Ca²⁺ signaling. An advantage, however, is the ability to observe patterns of signaling occurring in large fields of cells in multiple planes of the slice.

View larger version (128K): [in this window] [in a new window]   FIG. 1. Typical hippocampal slice organotypic culture. A: phase contrast image of a typical hippocampal slice organotypic culture after 4 days in vitro. The box outlines a typical imaged field as represented in B and C and in Figs. 2 and 3. Scale: 4.5 × 4.5 mm. B: neuron cell surface antigen antibody labeling of same region outlined in the box in A. Image shows immunofluorescence labeling of pyramidal neurons within a typical imaged field. Scale: 500 µm horizontally × 470 µm vertically. C: glial fibrillary acidic protein (GFAP) antibody labeling of a field of cells in the same region of slice. Image shows DAB staining of astrocytes in the CA1 region. Astrocyte cell bodies and processes are more sparse within the neuronal pyramidal layer (right) but are densely packed in the stratum oriens (left). Scale: 500 µm horizontally × 470 µm vertically. DG, dentate gyrus; SR, s. radiatum; P, pyramidal neuronal layer; SO, s. oriens.

Organotypic slices exhibit multiple patterns of spontaneous Ca²⁺ signaling

Extensive spontaneous Ca²⁺ signaling occurs in these slices without external stimulation. We observe multiple distinct patterns of spontaneous Ca²⁺ signaling in the CA1 and CA2 regions of the hippocampal slice that we characterize as 1) asynchronous, single-cell oscillations; 2) small intercellular waves limited to 5-15 cells (Ca²⁺ "puffs"); and 3) broad intercellular wavefronts simultaneously involving 50-100 cells that propagate in curvilinear and spiral patterns. Each pattern of signaling was observed in >50 different slice preparations.

Individual cells show rapid, low-amplitude Ca²⁺ oscillations with a period of 5-20 s and an estimated peak [Ca²⁺]_i of 75-150 nM. These single-cell oscillations are seen throughout CA1 and CA2, but occur most commonly in or immediately adjacent to the pyramidal neuronal cell layer. Small multicellular Ca²⁺ waves (Ca²⁺ puffs) occur frequently from many different sites of initiation in the s. oriens and s. radiatum in CA1 and CA2. These spontaneous Ca²⁺ puffs originate from a single cell (Fig. 2A) and spread concentrically at rates of 10-16 m/s to a total of 5-15 cells with estimated Ca²⁺ concentrations ranging from 100 to 200 nM. Although it is not possible to determine whether Ca²⁺ puffs originate in a neuron or involve some neurons, immunostaining for a neuronal marker (NCSA) that labels neurons in the pyramidal neuronal layer (Fig. 1) reveals that the majority of cell bodies participating in the Ca²⁺ puffs are not neurons. Parallel immunostaining for GFAP reveals that Ca²⁺ puffs occur in regions containing many astrocytes (Fig. 1).

In these same regions, broad intercellular wavefronts with a width of 5-10 cells and spanning as many as 100 cells and up to 500 m occur as shown in Fig. 3. Broad intercellular wavefronts are seen spontaneously in ~20% of slices. These intercellular waves may originate in a single cell and propagate in curvilinear and spiral patterns (Fig. 3A). The propagation rate of the waves ranges from 5-10 m/s, and the estimated peak [Ca²⁺]_i amplitude of the wavefront (125-200 nM) is consistent throughout the distance of propagation. In some cases the wavefront passes undiminished throughout the field of view, whereas in others the wavefront becomes irregular and breaks up as it propagates. When two curvilinear wavefronts collide, they annihilate at the point of contact (Fig. 3B). After collision, portions of the waves that have not collided merge and continue to propagate tangentially from the site of collision. The minimum apparent refractory period between waves ranges from 20 to 30 s in different regions of each field of cells (Fig. 3C).

Wavefronts most commonly originate in or near the pyramidal neuronal cell layer (CA1-CA2) and travel peripherally through the s. oriens region. However, waves also propagate from the s. radiatum or s. oriens and continue across the pyramidal layer. When these waves encounter the pyramidal neuronal cell layer, there is consistently a change in the temporal and spatial characteristics of propagation. In some cases, Ca²⁺ waves are abolished when they encounter the pyramidal layer, and in other cases there is a delay in propagation of 5-10 s or an interruption in the continuity of the wavefront, after which the wavefront is reestablished on the other side of the pyramidal layer. The alteration of wave propagation at the pyramidal neuronal layer, where there is a decreased density of glial cell bodies, suggests that glial cell-cell contacts are involved in wave propagation.

Wavefronts are initiated periodically from the same 1-5 foci within a 400 × 500 µm field, in contrast to the Ca²⁺ puffs that are initiated nonperiodically from as many as 15 different sites of initiation. As with the Ca²⁺ puffs, it is not possible to determine with certainty whether the spiral waves are initiated by neurons or glia. However, staining for neuronal markers on the same fields reveals that the majority of the cells participating in the waves are not neurons, and parallel staining for GFAP suggests that these cells are astrocytes (Fig. 1).

Ca²⁺ waves do not require extracellular Ca²⁺

Pharmacological manipulation of the slice provides information regarding mechanisms of Ca²⁺ signaling activity. Replacing the experimental medium with medium containing no added Ca²⁺ and 1 mM ethylene glycol-bis (-aminoethyl ether)-N,N,N,N-tetraacetic acid (EGTA) immediately stops most Ca²⁺ oscillations in individual cells (Fig. 4; n = 9 slices). The immediate cessation of Ca²⁺ signaling in these cells indicates that extracellular Ca²⁺ ([Ca²⁺]_o) is rapidly and significantly decreased. Ca²⁺ puffs continue for up to 4 min in the absence of [Ca²⁺]_o, although their frequency is reduced. Broad intercellular wavefronts also continue with reduced frequency for up to 4 min following removal of [Ca²⁺]_o (Fig. 4), after which no further waves are seen. The health of the slices does not appear to be compromised by this treatment, because returning the slices to regular perfusion buffer (containing Ca²⁺) restores the prior signaling patterns and frequency within 1 min.

Ca²⁺ puffs and wavefronts also continue to occur for 4 min in the presence of the nonspecific Ca²⁺ channel blocker Ni²⁺ (n = 3 slices, 100 M NiCl₂; data not shown), showing that voltage-gated Ca²⁺ channels are not necessary for the propagation of puffs or waves. Both U73122 (20 µM), which inhibits phospholipase C, and thapsigargin (5 µM), which depletes [Ca²⁺]_i stores, abolish all puffs and spiral waves, consistent with a mechanism involving formation of 3-inositol-1,4,5-trisphosphate (IP₃) and release of Ca²⁺ from intracellular stores (n = 3 slices each; data not shown).

Ca²⁺ wave behavior is altered by neuronal activity

The glutamate receptor subtype agonist NMDA has a dramatic effect on the characteristics of Ca²⁺ waves (Fig. 2B). Bath application of 20 µM NMDA evokes an immediate, simultaneous increase in [Ca²⁺]_i in the region identified morphologically as the pyramidal cell layer of CA1. This immediate response is followed in 2-5 s by broad intercellular wavefronts in the s. oriens (n = 5 slices) that continue for 1-2 min after application of NMDA. The NMDA-evoked waves travel faster (10-15 µm/s) and have greater estimated peak [Ca²⁺]_i (200-300 nM) than spontaneous waves. Because NMDA receptors are generally confined to neurons and are not present on glial cells, these results suggest that stimulation of neurons with NMDA results in a secondary activation of glial Ca²⁺ waves.

One possible mechanism for activation of glial Ca²⁺ signaling by neurons is activation of non-NMDA glutamate receptors by neuronally released glutamate. Consistent with this mechanism, the metabotropic glutamate receptor (mGlu_r) antagonist, (±)-methyl-4-carboxyphenylglycine (MCPG; 100 µM), abolishes puffs and waves (n = 3 slices; data not shown). Inhibition of neuron-mediated hippocampal astrocytic Ca²⁺ signaling by MCPG was described previously by Porter and McCarthy (1996). However, the mGlu_r agonist trans-(±)-1-amino-1,3-cyclopentanedicarboxylic acid (125 µM) has a variable effect, abolishing activity in some instances and having no effect in others (n = 7 slices; data not shown). 6,7-Dinitroquinoxaline-2,3-dione (70 µM), a kainate/quisqualate antagonist, increases the number of sites of initiation of spiral waves as well as increasing the frequency of the waves, resulting in a more disorganized pattern of activity (n = 7 slices; data not shown).

Tetrodotoxin (TTX; 1 µM), which inhibits voltage-gated Na⁺ channels and thereby action potentials, has an effect similar to 0 Ca²⁺ (n = 10 slices; data not shown). TTX immediately abolishes most single cell Ca²⁺ oscillations, but Ca²⁺ puffs and broad [Ca²⁺]_i waves continue to occur for 4 min with reduced frequency.

Oxygen dependence of Ca²⁺ signaling

Removal of O₂ from the perfusion medium results in the immediate cessation of virtually all spontaneous Ca²⁺ oscillations, puffs, and broad intercellular waves (n = 6 slices; data not shown). Slices perfused with N₂ purged buffer show no Ca²⁺ signaling for 2 min (n = 3 slices). After ~2 min of hypoxia Ca²⁺ signaling returns, consisting of all patterns of signaling described in the previous paragraphs. Longer term hypoxia (15-30 min) results in permanent and complete loss of calcium signaling. Control patterns of Ca²⁺ signaling could be recovered following shorter periods of hypoxia (<15 min).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

These results suggest that Ca²⁺ puffs and wavefronts represent patterns of spontaneous glial intercellular signaling whose frequency, propagation velocity, and amplitude can be altered by neuronal activity. Several characteristics of the Ca²⁺ puffs and waves are consistent with these patterns of signaling occurring primarily in astrocytes. First, the activity occurs with regional specificity in the hippocampus (i.e., it involves cell bodies in the s. radiatum and s. oriens, areas consisting largely of astrocytic cell bodies and processes as demonstrated by GFAP labeling, and fewer neuronal cell bodies and dendrites evidenced by immunabeling with MAP-5 and NCSA). Second, Ca²⁺ waves are delayed or abolished when they encounter the pyramidal neuronal cell layer where there are more neuronal cell bodies and fewer glial cell bodies and processes. Third, the slow rate of propagation of Ca²⁺ puffs and waves (5-15 µm/s) is consistent with the previously described propagation velocity for Ca²⁺ waves in glial cell cultures (Charles et al. 1991; Cornell-Bell et al. 1990). Finally, the propagation of the waves via Ca²⁺ release from intracellular stores is also consistent with previously described mechanisms for glial Ca²⁺ signaling (Charles et al. 1993). By contrast, the asynchronous single-cell Ca²⁺ oscillations that occur primarily in the pyramidal neuronal layer and that are abolished by TTX and removal of [Ca²⁺]_o most likely represent neuronal activity. Similar spontaneous single-cell Ca²⁺ transients were previously described in pyramidal neurons in the immature hippocampal slice (Dailey and Smith 1994).

Spontaneous Ca²⁺ puffs are remarkably similar in their spatial and temporal characteristics to spontaneous glial Ca²⁺ waves observed in mixed glial/neuronal cultures (Charles 1994). However, the broad intercellular wave fronts are significantly different from any previously described patterns of glial Ca²⁺ signaling in culture or slice preparations; these previously reported spontaneous, mechanically-induced, or transmitter-induced glial Ca²⁺ waves (Charles et al. 1991; Charles,1994; Cornell-Bell et al. 1990; Dani et al. 1992; Duffy and MacVicar 1995) do not propagate as wavefronts in curvilinear or spiral patterns.

Potential mechanisms of intercellular Ca²⁺ signaling

The effects of thapsigargin and U73122 on Ca²⁺ puffs and wavefronts are consistent with an Ca²⁺ signaling mechanism involving IP₃-mediated release of Ca²⁺ from intracellular stores. The relatively constant rate of propagation and amplitude of the Ca²⁺ wavefronts over long distances suggest that there is a regenerative mechanism involved in this type of Ca²⁺ signaling. Two possible pathways of propagation of the response from cell-to-cell are gap junctions and the release of a glial extracellular messenger. There is extensive evidence for gap junctional communication as a mechanism for propagation of glial Ca²⁺ waves (Charles et al. 1992; Finkbeiner 1992). In addition, we and others have observed that Ca²⁺ waves in glial cultures can be altered by rapid perfusion of the medium and can cross gaps between cells, suggesting that an extracellular messenger may also be involved (Charles 1997; Hassinger et al. 1996). Parpura et al. (1994) and Jeftinija et al. (1996) provide strong evidence for Ca²⁺-mediated release of transmitters from glial cells. We favor the hypothesis that a combination of gap junctional communication and Ca²⁺-dependent release of an extracellular messenger is involved in the intercellular propagation of Ca²⁺ waves.

The O₂ dependence of the spontaneous Ca²⁺ signaling in the slices is consistent with studies showing that acute slice preparations require O₂ perfusion to maintain normal synaptic activity. Schurr et al. (1989) demonstrated that the amplitude of the population spike in an acute hippocampal slice decreased to 0 mV within 2 min of hypoxia induction. Consistent with our findings, they also report that periods of hypoxia lasting >15 min resulted in poor recovery of function. Organotypic slices as prepared in the present study also have similar long-term O₂ requirements as acute slices as indicated by the observation that slices do not survive in culture unless they are maintained at an air-media interface (Stoppini et al. 1991).

The effects of NMDA, TTX, and 0 [Ca²⁺]_o provide evidence that glial [Ca²⁺]_i signaling is modulated by neuronal activity. Because glial cells do not generally express NMDA receptors, the immediate and simultaneous increase in [Ca²⁺]_i evoked by NMDA in the pyramidal neuronal layer of CA1 most likely represents the neuronal response to NMDA, whereas the delayed and sustained propagation of broad intercellular wavefronts in the s. oriens and s. radiatum is consistent with secondary activation of glial signaling. Activation of glial Ca²⁺ signaling by stimulation of neuronal pathways, including application of NMDA, was described previously in hippocampal slice preparations by Dani et al. (1992) and Porter and McCarthy (1996). Although our observations do not eliminate the possibility that NMDA-evoked Ca²⁺ waves may be traveling through neuronal processes, the delayed and sustained propagation of slow waves through regions of the slice that consist primarily of astrocyte cell bodies is more consistent with neuronal activation of glial signaling. Conversely, the reduced frequency of Ca²⁺ puffs and waves under conditions that inhibit neuronal activity (TTX and 0 [Ca²⁺]_o) indicate that although these patterns of intercellular signaling do not require neuronal activity, they are inhibited when neuronal activity is inhibited, consistent with the previous findings of Porter and McCarthy (1996). The effects of other pharmacological manipulations, particularly the other glutamate receptor agonists and antagonists, are difficult to interpret because we are unable to determine whether the agonists and antagonists are affecting neurons presynaptically or postsynaptically, glially, or all of the above.

Potential functions of intercelluar Ca²⁺ signaling

Although there is significant evidence that neuronal activity modulates glial intercellular signaling in the hippocampal slice, a crucial question is whether the converse is also true, i.e., that intercellular glial Ca²⁺ signaling in turn modulates neuronal activity. Glia-to-neuron signaling has been reported in dissociated culture preparations (Charles 1994; Hassinger et al. 1995; Nedergaard 1994; Parpura et al. 1994). Although we do not as yet have direct evidence for glia-to-neuron signaling in the slice preparation, our observation of broad [Ca²⁺]_i wavefronts that simultaneously cross over a neuronal layer raises the possibility that these Ca²⁺ wavefronts could temporally and spatially coordinate neuronal activity. Preliminary patch-clamp recordings in these same preparations reveal bursts of action potentials occurring in neurons in the same fields of cells that are displaying [Ca²⁺]_i waves. Studies are presently underway to determine the relationship of this neuronal bursting activity with Ca²⁺ waves.

CA1 and CA3 pyramidal neurons may display temporal and spatial coordination of discharges, even when there are sparse collateral connections between these neurons (Buzsaki and Chrobak 1995; Buzsaki et al. 1992; Taylor et al. 1995). Taylor et al. (1995) report synchronized oscillations in CA3 neuronal activity at intervals of 24 s mediated by activation of mGlu_rs. Although our studies do not involve CA3, this temporal, spatial, and pharmacological pattern of neuronal synchronization is consistent with the patterns of intercellular signaling we report here, suggesting that broad glial wavefronts could be involved in neuronal synchronization in the hippocampus. In addition, because broad glial wavefronts continue in the absence of [Ca²⁺]_o, it is possible that glial intercellular signaling plays a role in the nonsynaptic synchronization of neuronal activity in the hippocampus that has been observed in low [Ca²⁺]_o (Hochman et al. 1995; Jefferys and Haas 1982; Taylor and Dudek 1984). Support for a glial role in the coordination of neuronal activity in the hippocampus is provided by the observation of a close relationship between the magnitude of glial depolarizations and simultaneously recorded field potentials in rat hippocampal slices during 4-aminopyridine-induced synchronous activity (Avoli et al. 1993).

Because these studies were performed with immature hippocampal slices, another possibility is that the patterns of Ca²⁺ signaling reported here are involved in the development of the functional cellular architecture of the hippocampus, as has been suggested for multicellular Ca²⁺ signaling observed in cortical slice preparations (Kandler and Katz 1995; Yuste et al. 1995). Ca²⁺ puffs in the hippocampus have similar spatial and pharmacological properties as Ca²⁺ waves (Kandler and Katz 1995; Yuste et al. 1995) in cortical slices of postnatal rats. Aniksztejn et al. (1995) reported current oscillations induced by metabotropic receptors in immature rat CA3 hippocampal neurons; these current oscillations have similar temporal and pharmacological properties to the intercellular signaling we report here.

Comparison with Ca²⁺ signaling in other preparations

Intercellular Ca²⁺ signaling was reported previously in astrocytes in hippocampal slice preparations (Dani et al. 1992; Duffy and MacVicar 1995). However, this signaling does not occur in organized wavefronts with clear periodicity and refractory periods, and spiral waves were not observed. Spiral intercellular waves of depolarization were reported in cardiac muscle (Davidenko et al. 1992). Significant differences between spiral waves in cardiac muscle and those reported here include a much faster rate of propagation in cardiac muscle, the propagation of a depolarization rather than a change in [Ca²⁺]_i, and the relatively uniform cellular composition of the heart compared with the heterogeneous cellular composition of the hippocampus.

Spreading depression (SD) is a slowly propagating depression of neuronal activity and transmembrane ionic gradients arising in vertebrate gray matter (Leao 1944). Spiral waves of SD have been reported in the intact retina (Gorelova and Bures 1983). Waves of SD in the retina have several features in common with spiral calcium waves in hippocampal slice cultures. Waves of SD propagate at rates of 18-60 µm/s (Gorelova and Bures 1983; Nedergaard et al. 1995; Shibata and Bures 1974), which are similar to the rates of 5-15 µm/s we report here. Both retinal SD and hippocampal Ca²⁺ waves display refractory periods and annhilation of waves on collision (Gorelova and Bures 1983), and both occur in a heterogeneous cellular environment involving the participation of neurons and astrocytes (Fernandes de Lima et al. 1993; Nedergaard et al. 1995).

There are several features of Ca²⁺ signaling within Xenopus laevis oocytes that are similar to those seen between cells in hippocampal slice explants. Oocytes also display multiple patterns of increased Ca²⁺ signaling including focal increases in [Ca²⁺]_i, Ca²⁺ oscillations, and propagating planar and spiral wavefronts that annihilate on collision (Lechleiter et al. 1991). In oocytes the rate of propagation of Ca²⁺ waves is 10-30 µm/s, whereas in hippocampal slices it is 5-15 µm/s. In oocytes the refractory period beween Ca²⁺ waves is 4.7 s, whereas in hippocampal slices it is 20-30 s. In both preparations, the Ca²⁺ waves are primarily mediated by release of Ca²⁺ from intracellular stores. The major difference in the pattern of signaling in the two preparations is that the structure of the wavefront is less uniform in the brain slice than in the oocyte. This difference is likely due in large part to the heterogeneous multicellular composition of the slice and variability in the propagation of the response from cell to cell.

Spiral waves have been identified as a characteristic behavior of excitable media (Diks et al. 1995; Steinbock et al. 1993; Winfree 1972). The patterns of intercellular signaling we observe in hippocampal slice explants are similar to patterns of signal propagation observed in a number of other systems that have been modeled as excitable media (Davidenko et al. 1992; Lechleiter et al. 1991; Siegert and Weijer 1995). Modeling of spiral wave behavior in other systems may therefore have applications to intercellular signaling in the brain. Our studies raise the possibility that networks of glial cells behave as an excitable medium that responds to neuronal input and that the behavior of this medium plays a role in temporal and spatial coordination of activity in the hippocampus.

	ACKNOWLEDGEMENTS

  We thank Drs. James Sneyd and Scott Boitano for helpful comments. Data acquisition and analysis software were written by Dr. Michael Sanderson.

  This work was supported by a Turken Foundation Fellowship and a French Foundation Fellowship to M. E. Harris and National Institute of Neurological Disorders and Stroke Grants R29-NS-32283 and P1-NS-2808 to A. C. Charles and R01-NS-3195 to S. A. Frautschy.

	FOOTNOTES

  Address for reprint requests: A. Charles, UCLA Dept. of Neurology, 710 Westwood Plaza, Los Angeles, CA 90095.

  Received 15 July 1997; accepted in final form 14 October 1997.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ANIKSZTEJN, L., SCIANCALEPORE, M., BEN-ARI, Y., CHERJBINI, E. Persistent current oscilllations produced by activation of metabotropic glutamate receptors in immature rat CA3 hippocampal neurons. J Neurophysiol. 73: 1422-1429, 1995.[Abstract/Free Full Text]
• ATRI, A., AMUNDSON, J., CLAPHAM, D., SNEYD, J. A single-pool model for intracellular calcium oscillations and waves in the Xenopus laevis oocyte. Biophys. J. 65: 1727-1739, 1993.[Abstract/Free Full Text]
• AVOLI, M., PSARROPOULOU, C., TANCREDI, V., FUETA, Y. On the synchronous activity induced by 4-aminopyridine in the CA3 subfield of juvenile rat hippocampus. J. Neurophysiol. 7: 1018-1029, 1993.
• BLOCK, D. L., ELMEGREEN, B. G., WAINSCOTT, R. J. Smooth dark spiral arms in the flocculent galaxy NGC2841. Nature 381: 674-676, 1996.
• BUZSAKI, G., CHROBAK, J. J. Temporal structure in spatially organized neuronal ensembles: a role for interneuronal networks. Curr. Opin. Neurobiol. 5: 504-510, 1995.[Medline]
• BUZSAKI, G., HORVATH, Z., URIOSTE, R., HETKE, J., WISE, K. High-frequency network oscillation in the hippocampus. Science 256: 1025-1027, 1992.[Abstract/Free Full Text]
• CHARLES, A. C. Glia-neuron intercellular calcium signaling. Dev. Neurosci. 16: 196-206, 1994.[Medline]
• CHARLES, A. C. Intercellular calcium waves in glia. Glia. In press.
• CHARLES, A. C., DIRKSEN, E. R., MERRILL, J. E., SANDERSON, M. J. Mechanisms of intercellular calcium signaling in glial cells studied with dantrolene and thapsigargin. Glia 7: 134-145, 1993.[Medline]
• CHARLES, A. C., MERRILL, J. E., DIRKSEN, E. R., SANDERSON, M. J. Intercellular signaling in glial cells: calcium waves and oscillations in response to mechanical stimulation and glutamate. Neuron 6: 983-992, 1991.[Medline]
• CHARLES, A. C., NAUS, C. C., ZHU, D., KIDDER, G. M., DIRKSEN, E. R., SANDERSON, M. J. Intercellular calcium signaling via gap junctions in glioma cells. J. Cell Biol. 118: 195-201, 1992.[Abstract/Free Full Text]
• CORNELL-BELL, A. H., FINKBEINER, S. M., COOPER, M. S., AND SMITH, S. J. Glutamate induces calcium waves in cultured astrocytes: long-range glial signaling. Science 247: 470-473, 199.
• DAILEY, M. E., SMITH, S. J. Spontaneous Ca²⁺ transients in developing hippocampal pyramidal cells. J. Neurobiol. 25: 243-251, 1994.[Medline]
• DANI, J. W., CHERNJAVSKY, A., SMITH, S. J. Neuronal activity triggers calcium waves in hippocampal astrocyte networks. Neuron 8: 429-440, 1992.[Medline]
• DAVIDENKO, J. M., PERTSOV, A. M., SALOMONSZ, R., BAXTER, W. P., AND JALIFE, J. Spatiotemporal irregularities of spiral wave activity in isolated ventricular muscle. J. Electrocardiol. 24, Suppl: 113-122, 1992.
• DIKS, C., HOEKSTRA, B., DEGOEDE, J. Spiral wave dynamics. Chaos Solitons and Fractals 5: 645-660, 1995.
• DUFFY, S., MACVICAR, B. A. Adrenergic calcium signaling in astrocyte networks within the hippocampal slice. J. Neurosci. 15: 5535-5550, 1995.[Abstract]
• FERNANDES DE LIMA, V. M., SCHELLER, D., HANKE, W., SCHLUE, W. R. Self-sustained depressions in the chick retina and short-term neuronal-glial interactions within the gray matter neuropil. Brain Res. 614: 45-51, 1993.[Medline]
• FINKBEINER, S. Calcium waves in astrocytes-filling in the gaps. Neuron 8: 1101-1118, 1992.[Medline]
• GORELOVA, N. A., BURES, J. Spiral waves of spreading depression in the isolated chick retina. J. Neurobiol. 14: 353-363, 1983.[Medline]
• HASSINGER, T. D., ATKINSON, P. B., STRECKER, G. J., WHALEN, L. R., DUDEK, F. E., KOSSEL, A. H., KATER, S. B. Evidence for glutamate-mediated activation of hippocampal neurons by glial calcium waves. J. Neurobiol. 28: 159-170, 1995.[Medline]
• HASSINGER, T. D., GUTHRIE, P. B., ATKINSON, P. B., BENNETT, M. V., KATER, S. B. An extracellular signaling component in propagation of astrocytic calcium waves. Proc. Natl. Acad. Sci. USA 93: 13268-13273, 1996.[Abstract/Free Full Text]
• HOCHMAN, D. W., BARABAN, S. C., OWENS, J. W., SCHWARTZKROIN, P. A. Dissociation of synchronization and excitability in furosemide blockade of epileptiform activity. Science 270: 99-102, 1995.[Abstract/Free Full Text]
• JEFFERYS, J. G., HAAS, H. L. Synchronized bursting of CA1 hippocampal pyramidal cells in the absence of synaptic transmission. Nature 300: 448-450, 1982.[Medline]
• JEFTINIJA, S. D., JEFTINIJA, K. V., STEFANOVIC, G., LIU, F. Neuroligand-evoked calcium-dependent release of excitatory amino acids from cultured astrocytes. J. Neurochem. 66: 676-684, 1996.[Medline]
• KANDLER, K., KATZ, L. C. Neuronal coupling and uncoupling in the developing nervous system. Curr. Opin. Neurobiol. 5: 98-105, 1995.[Medline]
• LEAO, A. Spreading depression of activity in cerebral cortex. J. Neurophysiol. 7: 359-390, 1944.[Free Full Text]
• LECHLEITER, J., GIRARD, S., PERALTA, E., CLAPHAM, D. Spiral calcium wave propagation and annihilation in Xenopus laevis oocytes. Science 252: 123-126, 1991.[Abstract/Free Full Text]
• LECHLEITER, J. D., CLAPHAM, D. E. Molecular mechanisms of intracellular calcium excitability in X. laevis oocytes. Cell 69: 283-294, 1992.[Medline]
• MENNERICK, S., ZORUMSKI, C. F. Glial contributions to excitatory neurotransmission in cultured hippocampal cells. Nature 368: 59-62, 1994.[Medline]
• NEDERGAARD, M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 263: 1768-1771, 1994.[Abstract/Free Full Text]
• NEDERGAARD, M., COOPER, A.J.L., GOLDMAN, S. A. Gap junctions are required for the propagation of spreading depression. J. Neurobiol. 28: 433-444, 1995.[Medline]
• PARPURA, V., BASARSKY, T. A., LIU, F., JEFTINIJA, K., JEFTINIJA, S., HAYDON, P. G. Glutamate-mediated astrocyte-neuron signalling. Nature 369: 744-747, 1994.[Medline]
• PORTER, J. T., MCCARTHY, K. D. Hippocampal astrocytes in situ respond to glutamate released from synaptic terminals. J. Neurosci. 16: 5073-5081, 1996.[Abstract/Free Full Text]
• SCHURR, A., WEST, C. A., RIGOR, B. M. Electrophysiology of energy metabolism and neuronal function in the hippocampal slice preparation. J. Neurosci. Methods 28: 7-13, 1989.[Medline]
• SHIBATA, M., BURES, J. Optimum topographical conditions for reverberating cortical spreading depression in rats. J. Neurobiol. 5: 107-118, 1974.[Medline]
• SIEGERT, F., WEIJER, C. J. Spiral and concentric waves organize multicellular dictyostelium mounds. Curr. Biol. 5: 937-943, 1995.[Medline]
• STEINBOCK, O., ZYKOV, V., MULLER, S. C. Control of spiral wave dynamics in active media by periodic modulation of excitability. Nature 366: 322-324, 1993.
• STOPPINI, L., BUCHS, P. A., MULLER, D. A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods 37: 173-182, 1991.[Medline]
• TAYLOR, C. P., DUDEK, F. E. Synchronization without active chemical synapses during hippocampal afterdischarges. J. Neurophysiol. 52: 143-155, 1984.[Abstract/Free Full Text]
• TAYLOR, G. W., MERLIN, L. R., WONG, R. K. Synchronized oscillations in hippocampal CA3 neurons induced by metabotropic glutamate receptor activation. J. Neurosci. 15: 8039-8052, 1995.[Abstract]
• WINFREE, A. T. Spiral waves of chemical activity. Science 175: 634-636, 1972.[Abstract/Free Full Text]
• WU, R. L., BARISH, M. E. Astroglial modulation of transient potassium current development in cultured mouse hippocampal neurons. J. Neurosci. 14: 1677-1687, 1994.[Abstract]
• YUSTE, R., NELSON, D. A., RUBIN, W. W., KATZ, L. C. Neuronal domains in developing neocortex: mechanisms of coactivation. Neuron 14: 7-17, 1995.[Medline]
• ZARITSKY, D., RIX, H. W., RIEKE, M. Inner spiral stucture of the galaxy M51. Nature 364: 313-315, 1993.

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society