News & Views | Published:


Cellular defences destroyed

A lack of blood flow can kill nerve cells, by causing a massive influx of calcium ions. But what's happened to the cellular mechanisms for coping with excess calcium?

Without calcium, life would simply not be possible: it imparts strength to bones and serves as a messenger in myriad cellular processes. Underlying this ubiquitous signalling role is a 10,000-fold concentration gradient across the cell membrane, with 50–100 nM free calcium ions inside cells, separated by only a lipid bilayer from 1 mM calcium outside. Exquisitely regulated channels permit dollops of calcium to rush in across the membrane, causing localized increases in intracellular calcium levels and the activation of appropriate molecules. Yet under pathological conditions, such as a stroke, calcium can also be a killer, flooding into neurons and inducing lethal derangements. Although much has been learned about how excess calcium gets in, it has been puzzling that the mechanisms normally responsible for calcium extrusion are not better able to cope with the deluge. A paper by Bano et al., just published in Cell1, provides insight into why exactly these mechanisms may fall short, just when we need them the most.

When the brain experiences ischaemia — the blood-flow shortage associated with, for instance, a stroke — the excitatory neurotransmitter glutamate is released from nerve terminals and from other brain cells, and accumulates excessively in the extracellular space. This results in ‘excitotoxic’ overstimulation of the receptors found on most neurons, inducing large quantities of Ca2+ ions to enter these cells (Fig. 1). NMDA-type glutamate receptors, whose membrane channels have a naturally high Ca2+ permeability, contribute prominently to this Ca2+ influx and toxicity2. Moreover, in some neurons subjected to ischaemia, AMPA-type glutamate receptors develop enhanced Ca2+ permeability and become major direct routes for toxic Ca2+ entry3. The problem is compounded by the release of Ca2+ from internal stores and by the entry of Ca2+ through several other types of membrane channel (see, for instance, refs 4, 5).

Figure 1: Glutamate overproduction and calcium flow.

In response to ischaemia (a reduction in blood flow), the neurotransmitter glutamate is released from neurons and other brain cells (astrocytes; not shown). This induces Ca2+ entry (green arrows) into responsive neurons through NMDA-type glutamate receptors and Ca2+-permeable AMPA-type glutamate receptors. Ca2+ also enters through voltage-gated Ca2+ channels and several other types of Ca2+-permeable membrane channel, and is released from intracellular stores. The NCX proteins represent a primary cellular defence against Ca2+ overload, pumping Na+ ions in and Ca2+ ions out. But Bano et al.1 show that in ischaemic conditions NCX is destroyed by the Ca2+-activated protease calpain.

Working on the assumption that increased Ca2+ entry or release might not be the whole story, however, Bano et al.1 turned a spotlight on the membrane proteins that bear the lion's share of the responsibility for maintaining low intracellular Ca2+ concentrations — the Na+/Ca2+ exchanger (NCX) family. Under normal conditions, these proteins transport Na+ ions down their concentration gradient into cells, and harness the resulting energy to transport potentially large amounts of Ca2+ back out. Why, wondered Bano et al., isn't NCX on the job in ischaemic neurons, at least after the restoration of blood flow and Na+ gradients?

The authors found that NCX, in particular the NCX3 isoform, is cleaved in rat brains after a simulated stroke (induced by a 15-minute blockage of the middle cerebral artery, followed by restoration of blood flow). They linked this event back to excitotoxicity by identifying the same cleavage in cultured cerebellar granule neurons that were bathed in glutamate. The authors also identified the Ca2+-dependent protein-cleaving enzyme calpain as the culprit, given both the sensitivity of these events to a pharmacological calpain inhibitor, and the size of cleavage fragments. Recognizing the limited specificity of pharmacological protease inhibitors, Bano et al. also overexpressed a portion of a natural calpain inhibitor, calpastatin, in granule neurons; this elegant molecular probe also blocked glutamate-induced NCX3 cleavage.

Having established an association between excitotoxic or ischaemic conditions and calpain-mediated degradation of NCX, Bano et al. went on to delineate its functional importance. They showed that overexpressing either calpastatin or the calpain-resistant NCX isoform, NCX2, reduced the glutamate-induced increase in both neuronal intracellular Ca2+ concentration and neuronal death. Finally, the authors found that suppressing NCX3 expression, achieved by treatment with small interfering RNAs, had the opposite effect: granule neurons became sensitized to glutamate-induced Ca2+ overload and death.

Bano and colleagues' observations1 provide evidence that the calpain-mediated cleavage of NCX can be a ‘feedforward’ step in excitotoxic neuronal death. It is highly plausible that compromising the cell's main high-capacity mechanism for removing Ca2+ would increase the impact of the initial glutamate-induced burst of Ca2+ entry. More than a decade ago, Khodorov and colleagues6 reported that the sustained increase in the intracellular Ca2+ concentration to which cerebellar granule cells become victim after exposure to high glutamate levels persisted after the removal of both glutamate and extracellular Ca2+. This led them to suspect a disruption of homeostatic mechanisms. Moreover, they found that removing extracellular Na+ — expected to inhibit NCX-mediated Ca2+ extrusion (but also to have other effects) — did not augment cell death, leading them to propose specifically that NCX was suppressed. Bano et al. have now provided solid molecular evidence to support this view.

Further studies will be needed to answer several questions that arise. In particular, given the known heterogeneity in the characteristics of excitotoxicity in different types of neuron, it will be important to see whether the timing and extent of calpain-induced NCX3 (or NCX1) cleavage varies from neuron to neuron, or from one compartment to another within the same neuron. Further, does NCX cleavage depend on the route of Ca2+ entry into the cytoplasm, as do other elements of the excitotoxic cascade7? Does NCX cleavage contribute to the death of other brain cells after ischaemia (for example, the AMPA-receptor-mediated death of oligodendrocytes, or the non-excitotoxic death of astrocytes)? Stepping back, it will be important to test whether inhibiting NCX cleavage reduces brain damage in animal stroke models. If so, this might form the basis for a new class of neuroprotective treatments, useful in strokes and perhaps other conditions.

However, one should not conclude yet that NCX cleavage always enhances neuronal death. Glutamate exposure increases intracellular Na+ as well as Ca2+ levels, and, in Na+-loaded and depolarized cells, NCX can run in the reverse direction, mediating Na+ efflux and Ca2+ influx and thus potentiating excitotoxic neuronal death8. NCX cleavage might be expected to oppose these events. Furthermore, excitotoxic insults, like other insults, can induce neurons to undergo cellular suicide (programmed cell death), and intermediate increases in intracellular Ca2+ concentrations can stave off this death programme9,10. Although NCX has low affinity for Ca2+, its inhibition can prolong the transient large fluctuations in Ca2+ that are evoked by the depolarization of sympathetic neurons11. So, one could envisage a situation in which an initial harmful burst of Ca2+ influx is followed by relative normalization of Ca2+ levels, and calpain-induced NCX cleavage helps to elevate these late Ca2+ levels to the point at which programmed cell death is inhibited.

As Bano et al.1 point out, calpain-induced NCX cleavage has an intriguing parallel in the caspase-induced cleavage of another Ca2+-extrusion pump, PMCA12. Perhaps both of these cleavage events can either promote or inhibit cell death in different circumstances, specifically acting at low injury levels to ensure that a commitment to undergo programmed cell death does not occur casually.


  1. 1

    Bano, D. et al. Cell 120, 275–285 (2005).

  2. 2

    Choi, D. W. et al. J. Neurosci. 8, 185–196 (1988).

  3. 3

    Calderone, A. et al. J. Neurosci. 23, 2112–2121 (2003).

  4. 4

    Xiong, Z. -G. et al. Cell 118, 687–698 (2004).

  5. 5

    Aarts, M. et al. Cell 115, 863–877 (2003).

  6. 6

    Khodorov, B. et al. FEBS Lett. 324, 271–273 (1993).

  7. 7

    Aarts, M. et al. Science 298, 846–850 (2002).

  8. 8

    Kiedrowski, L., Czyz, A., Baranauskas, G., Li, X. -F. & Lytton, J. J. Neurochem. 90, 117 (2004).

  9. 9

    Franklin, J., Sanz-Rodriguez, C., Juhasz, A., Deckwerth, T. L. & Johnson, E. M. J. Neurosci. 15, 643–664 (1995).

  10. 10

    Choi, D. W. Trends Neurosci. 18, 58–60 (1995).

  11. 11

    Wanaverbecq, N., Marsh, S. J., Al-Qatari, M. & Brown, D. A. J. Physiol. (Lond.) 550, 83–101 (2003).

  12. 12

    Schwab, B. L. et al. Cell Death Differ. 9, 818–831 (2002).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.