A Theory of Cortical Neuron-Astrocyte Interaction

In addition to describing the mechanisms by which associative cortex processes and utilizes information, any clinically useful model should also provide insight into the maladies to which the brain is universally vulnerable. A theory fulfilling these criteria departs radically from currently held concepts of brain function and proposes that motile astrocytes control synaptic activity and organize synapses into synchronously firing groups.



With the discovery of calcium waves in astrocytes and the demonstration that calcium waves can be propagated through astrocytic networks(1, 2), another layer of complexity may have been added to associative cortical functioning. The presence of glial signaling in the brain and demonstrated interactions between neurons and glia (3, 4, 5) have served as a warning to neurotheoreticians that neuronal networks alone may not explain this biological computer. Models of associative cortical function must now account for not only the activities of astrocytes, the largest cell population in the brain, but also for the relationships astrocytes and astrocytic signaling networks have with neurons in the encoding and manipulation of sensory information. In addition to describing the mechanisms by which associative cortex processes and utilizes information, any clinically useful model should also provide insight into the maladies to which the brain is universally vulnerable. A theory fulfilling these criteria departs radically from currently held concepts of brain function and proposes that motile astrocytes control synaptic activity and organize synapses into synchronously firing groups.


Dynamic Aspects of the Neuron-Astrocyte Relationship

Astrocytes are not necessarily a stationary cell population (6, 7, 8); astroglia in vitro extend processes with highly motile tips (9) and human fetal astrocytes in tissue culture are in a state of almost constant motion (Antanitus D., unpublished observations). Additional evidence for astrocytic movement has been observed in vitro, where astrocytes change shape by developing filopodial extensions when exposed to neurotransmitter (10). The direction of filopodial growth is vectored up the concentration gradient of the applied neurotransmitter. Such observations, if applicable to the brain in situ, support the theoretical premise that the filamentous cytoplasmic extensions of astrocytes in vivo may be actively motile and neurotransmitter tropic. In addition, evidence shows that astrocytes can swell or shrink depending on activity in nearby neuronal elements (11, 12, 13); this, too, may alter the spatial relationship between the astrocyte and nearby neuronal processes, including pre- and postsynaptic elements.


Figure 1a
Schematic of one synapse enveloped by filopodial extension of one astrocyte.
In vivo, astrocytes envelop synapses so intimately that they are thought to compartmentalize the spaces between cells in associative cortex (14) (Fig. 1a). One astrocyte may envelop some of the synapses of several neurons, and one neuron may have its synapses enveloped by several astrocytes. This theory proposes that the neurotransmitter-tropic filopodial response of astrocytes and/or activity related astrocytic swelling in vitro also occur in vivo where they provide a mechanism that mediates the astrocytic envelopment and segregation of synapses. Just as the extent of filopodial growth in vitro is dependent on neurotransmitter concentration and exposure time and the degree of swelling is dependent on extracellular ionic concentrations, the extent of astrocytic envelopment of synapses in vivo depends on the synapses' firing history. An astrocyte would thus be expected to change its envelopment of individual synapses within the total population of enveloped synapses based on the relative firing frequencies of each synapse. I refer to this in vivo astrocytic movement guided by synaptic activity as infotropism.


Astrocytes Control Neurotransmitter Release

Figure 1b
In absence of astrocytic calcium wave, free Ca2+ available in synaptic cleft.


Figure 1c
Free Ca2+ allows the release of neurotransmitter resulting in both post synaptic activation and infotropic movement of astrocyte toward neurotransmitter within the synaptic cleft.
Nearly 30 years ago, Katz and Miledi (15) demonstrated that the amount of neurotransmitter released from presynaptic terminals in response to a propagated action potential was related to [Ca2+] in the synaptic cleft. This relationship was hypothesized to be nonlinear, with synaptic release proportional to the fourth power of [Ca2+] (16). Thus, small changes in synaptic cleft [Ca2+] may have relatively large effects on synaptic transmission; indeed, this concept has received recent support (17, 18). Astrocytes that envelop synapses, it is proposed, are able to control synaptic cleft [Ca2+] throughmechanisms of the astrocytic calcium wave. Two major events occurring during the astrocytic calcium wave have recently been described (1,2): liberation of intracellular protein bound Ca2+ into the astrocytic cytoplasm and simultaneous uptake of Ca2+ from the astrocytes' extracellular environment. Indirect evidence also suggests the return of free Ca2+ to the extracellular space upon termination of the calcium wave (1). In vivo, astrocytes' ability to rapidly reduce and restore Ca2+ in the synaptic clefts they envelop affords them control of neurotransmitter release and, therefore, control of neurotransmission itself (Fig. 1b,c).


Neurotransmitters Evoke Calcium Waves

Astrocytic calcium waves can be evoked in vitro by exposure to both excitatory and inhibitory neurotransmitters applied in suprathreshold concentrations (1,2). Although most of the evidence for calcium waves has come from observations in vitro, there is some evidence for calcium wave production in situ as well (5, 19). This theory proposes that similar astrocytic calcium waves also occur in vivo.

It is unlikely that sufficient neurotransmitter would be released at one astrocytically enveloped synapse to provoke a calcium wave in the enveloping astrocyte. One astrocyte, however, can envelop many synapses in vivo. Thus, if calcium waves occur in vivo within the cortex, coordinated exposure to neurotransmitters at many of the synapses one astrocyte controls could be sufficient to trigger calcium wave production. A reciprocal relationship would then be established between an astrocyte and the synapses it envelops. The astrocyte controls neurotransmitter release through calcium wave modulation of synaptic [Ca2+] and an astrocytic calcium wave is precipitated in turn by coordinated neurotransmitter release at the many synapses that the astrocyte envelops.

Figure 2a
Astrocytic calcium wave reduces free Ca2+ in synaptic cleft.
Figure 2b
Despite arriving action potential and open calcium channels, neurotransmitter release is reduced during astrocytic calcium wave.
These interdependent astrocytic and neuronal events are summarized in Figs. 1b, c and 2. In the absence of either a presynaptic action potential or an astrocytic calcium wave, a high level of Ca2+ is available within the extracellular space of the synaptic cleft (Fig. 1b). An action potential arriving at the presynaptic terminal opens the terminal's calcium channels. Extracellular Ca2+ enters the terminal and facilitates neurotransmitter release into the synaptic cleft (Fig 1c). The released neurotransmitter activates the postsynaptic terminal. Simultaneously, the astrocyte enveloping the synapse is also exposed to the released neurotransmitter resulting in two consequences: 1) the "infotropic" response of the astrocyte is activated, moving the enveloping astrocyte's surface closer to the highest concentration of neurotransmitter within the synaptic cleft; 2) if the [neurotransmitter] threshold of the astrocyte is exceeded, a calcium wave is generated.

During a calcium wave, the synaptic environment changes dramatically (Fig 2a). The astrocytic calcium wave reduces Ca2+ in the cleft. Decreased [Ca2+] in the cleft inhibits further neurotransmitter release, despite the arrival of action potentials (Fig 2b). Only with cessation of the astrocytic calcium wave will Ca2+ return to its original level in the synaptic cleft allowing neurotransmitter release at high levels.


Astrocytes Organize Synapses

Figure 3
Schematic of an astrocyte and its domain of influence (here, eight synapses).
3a. In absence of astrocytic calcium wave, Ca2+ available in synaptic clefts.
3b. Neurotransmitter released at four synapses within the domain.
3c. Astrocytic calcium wave results from exposure to neurotransmitter. Ca2+ is reduced in synaptic clefts of entire domain inhibiting neurotransmission in all eight synapses.
3d. Termination of calcium wave returns previous level of Ca2+ to all synaptic clefts resulting in secondary synchronous neurotransmission throughout the domain.
By simultaneously regulating neurotransmission in all of the synapses an astrocyte has enveloped, the astrocytic calcium wave may coordinate synapses into synchronously firing groups. Thus, all of the synapses enveloped or partially enveloped by an astrocyte may be within that astrocyte's domain of synaptic influence.
Fig. 3a illustrates an astrocyte and many of the synapses it envelops. Initiation of a calcium wave in this astrocyte occurs when a sufficient number of presynaptic terminals within its domain of influence release neurotransmitter simultaneously, exposing the astrocyte to suprathreshold stimulation (Fig 3b). During the astrocyte's calcium wave all the synaptic extracellular space within its domain undergoes a decrease in Ca2+, inhibiting neurotransmitter release and neurotransmission throughout the astrocyte's entire domain of influence (Fig. 3c). Termination of the calcium wave shifts Ca2+ back into all the synaptic clefts simultaneously. Although calcium channels in each of the synapses within the domain may have opened at different times, the sudden, simultaneous availability of increased Ca2+ to all opened channels might be expected to phase lock and synchronize neurotransmitter release. Synchronous neurotransmission throughout the entire domain would be expected as the astrocyte's calcium wave is reinitiated (Fig 3d). Such synchronous neurotransmission, it is hypothesized, might occur at relatively high frequencies in vivo because astrocytes at room temperature in vitro can generate calcium wave oscillations with frequencies of several times per minute (1, 2, 3, 20).


Astrocyte Connections Allow Expanded Control of Synapses

Astrocytes are connected to each other by unidirectional gap junctions that allow the exchange of many biologically important molecules (21, 22). Astrocyte to astrocyte unidirectional transfer of unbound cytoplasmic Ca2+ occurs through gap junctions during calcium wave generation (23). The presence of unbound Ca2+ within astrocytic cytoplasm is known to facilitate calcium wave activation (23). Inositol 1,4,5-triphosphate (InsP3), which is important in the release of Ca2+ from intracellular stores (24), has also been hypothesized to move through gap junctions and play a role in intracellular calcium wave activation and intercellular calcium wave propagation (25). Sequential intercellular propagation of calcium waves may occur as a consequence of the relative combined effects of intercellular Ca2+ and/or InsP3 transmission and coincident neurotransmitter stimulation from overlapping domains of synaptic influence. Varieties of calcium wave propagation supporting these contentions have been observed in monolayer and organotypic culture preparations (2, 19, 23). A group of astrocytes acting as a single cell or syncytium may, through the generation of a pervasive calcium wave, control and synchronize the firing patterns of synapses of a great many neurons. This putative coupling of astrocytic and synaptic function implies that the initiation of a calcium wave by a subgroup of synapses within the domain should inhibit neurotransmitter release throughout the entire domain, and this initiation should effect a secondary neurotransmission synchrony in the domain's larger group of synapses. This final array of synchronously firing synapses would be related only in part to the synapses initiating the calcium wave.

At a rate equal to the generation of the astrocyte's calcium wave, the astrocyte would be expected to infotropically increase its influence on the initiating synapses of the subgroup by drawing closer to them, decreasing their extracellular perisynaptic spaces. These decreases in the perisynaptic spaces would not only increase the astrocyte's ability to control extracellular Ca2+ but would also, by increasing concentration of released neurotransmitter within a now more restricted synaptic cleft, increase the synapses' efficiency in provoking a calcium wave. Such an activity-dependent change in the astrocytic calcium wave response, apparently being potentiated by previous episodes of activity in nearby synapses, has recently been demonstrated in situ (5). This reciprocating relationship might enable an even smaller subgroup of presynaptic terminals within the astrocyte's domain to precipitate a calcium wave, in turn affecting neurotransmission within the entire domain of synapses.

After the pervasive and synchronous neurotransmitter release throughout the astrocyte's domain of influence, a calcium wave is again initiated. A reciprocating relationship between calcium wave activation and synchronous neurotransmitter release would be expected to continue as long as enough presynaptic terminals are releasing sufficient quanta of neurotransmitter to exceed the astrocyte's calcium wave threshold. The frequencies of these astrocytic-synaptic oscillations would be dependent upon and reflective of both the character and duration of the initiating stimuli.

From this vantage, this theory proposes that sensory information is encrypted by associative cortex in patterns of synchronous synaptic firing that are determined by varying patterns of astrocytic syncytia. Astrocytic-synaptic oscillations occur at various frequencies for various periods of time. The spatial patterns, frequencies and duration of these oscillations are dependent on both the sensory input creating them and the influences of pre-established astrocytic domains of influence.


The Input Changes the Hardware

It has traditionally been held that associative cortical integration of sensory input is accomplished purely through utilization of complex networks of neuronal connections. The theory presented here, however, contends that the associative areas of brain include plastic, motile astrocytes that participate in the encoding and manipulation of information by organizing and controlling groups of synapses. Sensory information that began its journey as neuronal action potentials eventually reaches areas of the central nervous system in patterns that activate astrocytic calcium waves. According to the theory, activated astrocytes in turn infotropically encode the information with resulting synchronously firing synaptic domains. The character of the synaptic domains reflects the primary perceived sensory information at any moment in time, but the pools of synchronous synapses are modified by the presentation of novel information. Through the mechanism of infotropic astrocytic movement, metaphorically speaking, `the input changes the hardware.'

The theory predicts that repetitive, identical activation patterns of sensory cortices will be mirrored in repetitive infotropic responses of associative cortical astrocytes and the creation of representational domains of influence. Because each synapse and astrocyte may participate in a host of infotropically encoded patterns which reflect the character, duration and chronology of sensory input, individual synaptic or astrocytic activations have deterministic meaning only in the context of their place in activation sequences.

Applying the Theory of Neuron-Astrocyte Interaction

The theory of neuron - astrocyte interaction may be useful in provoking a reformulation of concepts of some disorders affecting the nervous system. The following sections consider epilepsy, minor head trauma, and the effects of neuron-astrocyte interaction on gap junction blockage and associative memory.


Is Epilepsy an Astrocytic Disorder?

No universally accepted hypothesis has yet been offered that explains the varieties of rhythmic and synchronous neuronal firing that define epileptic seizures. Many investigators have suggested, however, that astroglia play a significant role in epilepsy (26, 27, 28). This theory allows for synchronous neuronal firing involving the domains of many astrocytes acting as a syncytium. A calcium wave produced by millions of astrocytes acting as a syncytium would, at the wave's termination, be expected to result in the synchronous firing of neurons in large regions of the brain. The widespread, coordinated release of neurotransmitter would then precipitate a second equally pervasive calcium wave in the astrocytic syncytium, which again would effect synchronous neuronal firing, and so on, resulting in the synchronous and rhythmic neuronal firing characteristic of the epileptic seizure. Epilepsy, this theory contends, is primarily a disorder of astrocytes caused by pathological, high frequency syncytial calcium waves preceding and modulating a consequential neuronal synchrony. Varieties of seizures, according to this theory, are dependent on the extent and character of the pathologically large astrocytic syncytia producing them.

The partial complex seizures of childhood known as petit mal may provide an example of the astrocytic etiology of epilepsy. Petit mal seizures can often be provoked by hyperventilation (for example, by having a child blow repeatedly on a pinwheel). After several minutes of hyperventilating the pH of the child's blood becomes slightly alkaline as a consequence of the respiratory depletion of carbon dioxide. With this shift toward the alkaline, the petit mal seizure begins as evidenced by a characteristic three-per-second spike-wave pattern on the electroencephalogram. In addition to their many other roles, astrocytes near the capillaries of cerebral cortex envelop these blood vessels with their cytoplasmic extensions. The exposure of these perivascular astrocytes, which also control synaptic domains of influence, to the sudden alkaline shift in the blood may result in primary perivascular calcium wave activation and secondary perivascular synaptic synchrony. Indeed, the three-per-second spike-wave pattern seen on the electroencephalogram may reflect the synaptic (spike) and astrocytic (wave) oscillations the theory proposes, with unusually high frequency calcium waves resulting from the effects of alkalosis combined with secondarily coordinated neurotransmitter release. Of additional interest is the demonstration that many of the medications used in the treatment of epilepsy are astrocytic calcium wave inhibitors (29, 30).


Does Minor Head Trauma Create Nonsense Calcium Waves?

Equally mysterious are the mechanisms producing unconsciousness resulting from minor head trauma insufficient in force to cause any detectable injury. Mechanical perturbation has been shown to precipitate calcium waves in vitro (20). A blow to the head results in a mechanical compression wave traveling through the brain. This mechanical force could be sufficient to produce a pattern of widespread sequential calcium waves that reflect the shape and velocity of the mechanical compression wave. The astrocytic calcium waves so produced would be unrelated to, and for a while unresponsive to the influence of, normal sensory input. Meaningful interactions between astrocytes and synapses could be overwhelmed by the disruptively nonsensical mechanically induced calcium wave patterns. Despite the inability of the mechanical force to produce macro or microscopic injury, the brain -- the person -- would be "knocked out" or temporarily unconscious.


Do General Anesthetics Work by Blocking Astrocytic Gap Junctions? A similar unconscious state in which the cellular elements of associative cortex cannot respond normally to sensory input could, according to the theory, result from astrocytic gap junction blockade. Without functional gap junctions, astrocytes become unable to transmit their control of synchronous synaptic firing through the cortex. Because astrocytes cannot form syncytia, their control is restricted to only the synapses within their individual domains of influence. Several general anesthetics have been demonstrated to be reversible gap junction blockers (21, 31).


Does Memory Depend on the Neuron-Astrocyte Relationship?

Many authors have speculated on the involvement of astrocytes in memory and CNS plasticity (32, 33). The neuron-astrocyte theory here, however, proposes that the infotropic character of astrocytes is crucial to the cellular basis of associative memory. Also, neuroplasticity may more properly be termed neuronal-astrocytic plasticity, with the astrocytes' capacity to influence groups of synapses compromised only by the degree to which they have already organized these groups. Biological imprinting that occurs within defined critical periods of many species might represent a more permanent infotropically mediated astrocytic organization of synaptic domains. This premise is supported by the work of Muller and Best (34), who reinduced ocular-dominance plasticity in adult animals by supplementing the adult visual cortex with astrocytes cultured from the visual cortex of newborn animals.



The neuron-astrocyte theory challenges the notion that neuronal networks alone, no matter how complex, can explain the many normal and pathological aspects of brain function. This model, in which motile astrocytes organize and modulate synaptic activity, may permit widely divergent aspects of brain function and pathology to be viewed from a single theoretical vantage. Testing key aspects of this theory, i.e., astrocytic movement in vivo and astrocytic control of neurotransmission in vivo, will present significant technical challenges. When these challenges are overcome, however, we may learn significant new lessons about cortical function, its cellular substrates, and the mechanisms by which it is compromised in a variety of disease states.



Acknowledgments I thank R. Sidman, J. Yesinowski, R. Seitz, and K. Fretwell for their helpful comments and K. Fretwell for providing the illustrations.


  1. Cornell-Bell AH, Finkbeiner SM, Cooper MS, Smith SJ. Glutamate induces calcium waves in cultured astrocytes: long-range glial signaling. Science 1990;247:470-473.
  2. Cornell-Bell AH, Finkbeiner SM. Ca2+ waves in astrocytes. Cell Calcium 1991;12:185-204.
  3. Nedergaard M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 1994;263:1768-1771.
  4. Parpura V, Basarsky TA, Liu F, Jeftinija K, Jeftinija S, Haydon PG. Glutamate-mediated astrocyte-neuron signaling. Nature 1994;369:744-747.
  5. Pasti L, Volterra A, Pozzan T, Carmignoto G. Intracellular calcium oscillations in astrocytes: a highly plastic bidirectional form of communication between neurons and astrocytes in situ. J Neurosci 1997;17:7817-7830.
  6. Watanabe T, Raff MC. Retinal astrocytes are immigrants from the optic nerve. Nature 1988;332:834-837.
  7. Zhou HF, Lund RD. Migration of astrocytes transplanted to the midbrain of neonatal rats. J Comp Neurol 1992;317:145-155.
  8. Okoye GS, Powell EM, Geller HM. Migration of A7 immortalized astrocytic cells grafted into the adult rat striatum. J Comp Neurol 1995;362:524-534.
  9. Mason CA, Edmondson JC, Hatten ME. The extending astroglial process: development of glial shape, the growing tip, and interactions with neurons. J Neurosci 1988;8:3124-3134.
  10. Cornell-Bell AH, Thomas PG, Caffrey JM. Ca2+ and filopodial responses to glutamate in cultured astrocytes and neurons. Can J Physiol Pharmacol 1992;70:S206--S218.
  11. Sykova E, Svoboda J, Simonova Z, Jendelova P. Role of astrocytes in ionic and volume homeostasis in spinal cord during development and injury. Prog Brain Res 1992;94:47-56.
  12. Kimelberg HK, Sankar P, O'Connor ER, Jalonen T, Goderie SK. Functional consequences of astrocytic swelling. Prog Brain Res 1992;94:57-68.
  13. Nicholson C, Rice ME. Use of ion-selective microelectrodes and voltametric microsensors to study brain cell microenvironment. In: Boulton AA, Baker GB, Walz W, editors. Neuromethods: the neuronal microenvironment. New York: Humana; 1988. p. 247-361.
  14. Peters A, Palay SL, Webster HDeF. The neuroglial cells. In: The fine structure of the nervous system: neurons and their supporting cells. London: Oxford University Press; 1991. p. 273-295.
  15. Katz B, Miledi R. The timing of calcium action during neuromuscular transmission. J Physiol (Lond) 1967;189: 535-544.
  16. Dodge Jr FA, Rahamimoff R. Co-operative action of calcium ions in transmitter release at the neuromuscular junction. J Physiol (Lond) 1967;193:419-432.
  17. Bennett MR, Gibson WG, Robinson J. Probabilistic secretion of quanta and the synaptosecretosome hypothesis: evoked release at active zones of varicosities, boutons, and endplates. Biophys J 1997;73:1815-1829.
  18. Ravin R, Spira ME, Parnas H, Parnas I. Simultaneous measurement of intracellular Ca2+ and asynchronous transmitter release from the same crayfish bouton. J Physiol (Lond) 1997;501 (Pt 2):251--262.
  19. Dani JW, Chernjavsky A, Smith SJ. Neuronal activity triggers calcium waves in hippocampal astrocyte networks. Neuron 1992;8:429-440.
  20. Charles AC, Merrill JE, Dirksen ER, Sanderson MJ. Intercellular signaling in glial cells: calcium waves and oscillations in response to mechanical stimulation and glutamate. Neuron 1991;6:983-992.
  21. Robinson SR, Hampson ECGM, Munro MN, Vaney DI. Unidirectional coupling of gap junctions between neuroglia. Science 1993;262:1072-1074.
  22. Rash JE, Duffy HS, Duder FE, Bilhartz BL, Whalen LR, Yasamura T. Grid-mapped freeze-fracture analysis of gap junctions in gray and white matter of adult rat central nervous system, with evidence for a "panglial syncytium" that is not coupled to neurons. J Comp Neurol 1997;388:265-292.
  23. Finkbeiner SM. Calcium waves in astrocytes -- filling in the gaps. Neuron 1992;8:1101-1108.
  24. Berridge MJ. Inositol triphosphate and calcium signaling. Nature 1993;361:315-325.
  25. Sanderson MJ. Intercellular calcium waves mediated by inositol triphosphate. In: Ciba Foundation Symposium. Chichester: Wiley; 1995. p.175-194.
  26. Ransom BR, Goldring S. Slow hyperpolarization in cells presumed to be glia in cerebral cortex of cat. J Neurophysiol 1973;36:879-892.
  27. White HS, Woodbury DM, Chen CF, Kemp JW, Chow SY, Yen-Chow YC. Role of glial cation and anion transport mechanisms in etiology and arrest of seizures. Advances in Neurol 1986;44:695-712.
  28. Heineman U, Eder C, Lab A. Epilepsy. In: Kettenman H, Ransom BR, editors. Neuroglia. New York: Oxford University Press; 1995. p. 936-949.
  29. White HS, Skeen GA, Edwards JA. Pharmacological regulation of astrocytic calcium channels: for the treatment of seizure disorders. Prog Brain Res 1992;94:77-87.
  30. Nilsson M, Hansson E, Ronnback L. Agonist-evoked Ca2+ transients in primary astroglial cultures -- modulatory effects of valproic acid. Glia 1992; 5:201-209.
  31. Mantz J, Cordier J, Giaume C. Effects of general anesthetics on inter-cellular communications mediated by gap junctions between astrocytes in primary culture. Anesthesiology 1993; 78:892-901.
  32. Smith SJ. Do astrocytes process neural information? Prog in Brain Res 1992;94:119-136.
  33. Muller CM. Glial cells and activity-dependent central nervous system plasticity. In: Kettenman H, Ransom BR, editors. Neuroglia. New York: Oxford University Press; 1995. p. 805-814.
  34. Muller CM, Best J. Ocular dominance plasticity in adult cat visual cortex after transplantation of cultured astrocytes. Nature 1989; 342:427-430.

Copyright (c) 1998, Dale Antanitus. Portions copyright (c) Williams & Wilkins Inc.