• Open Access

Calcium homeostasis and modulation of synaptic plasticity in the aged brain


Thomas C. Foster, PhD, Department of Neuroscience, McKnight Brain Institute, University of Florida, PO Box 100244, Gainesville, FL 32610-0244, USA. Tel.: (352) 392 4359; fax: (352) 392 8347; e-mail: foster@mbi.ufl.edu


The level of intracellular Ca2+ plays a central role in normal and pathological signaling within and between neurons. These processes involve a cascade of events for locally raising and lowering cytosolic Ca2+. As the mechanisms for age-related alteration in Ca2+ dysregulation have been illuminated, hypotheses concerning Ca2+ homeostasis and brain aging have been modified. The idea that senescence is due to pervasive cell loss associated with elevated resting Ca2+ has been replaced by concepts concerning changes in local Ca2+ levels associated with neural activity. This article reviews evidence for a shift in the sources of intracellular Ca2+ characterized by a diminished role for N-methyl-D-aspartate receptors and an increased role for intracellular stores and voltage-dependent Ca2+ channels. Physiological and biological models are outlined, which relate a shift in Ca2+ regulation with changes in cell excitability and synaptic plasticity, resulting in a functional lesion of the hippocampus.

Ca2+ and the balance of cell growth and death

Fluctuations in the level of intracellular Ca2+ affect almost every physiological, chemical, and biological process of the neuron. In particular, there is an exquisite balance between two conflicting processes: mechanisms that permit the entry of Ca2+ in order to initiate growth and mechanisms for limiting excess intracellular Ca2+ in order to stave off degenerative processes (Choi, 1992; Berridge et al., 1998; Holscher, 2005). The relative strengths of these two competing forces determine the vitality of the cell.

At the level of the synapse the magnitude and the duration of a rise in intracellular Ca2+ acts on bidirectional activity-dependent synaptic plasticity, increasing or decreasing synaptic transmission (Cummings et al., 1996; Neveu & Zucker, 1996; Yang et al., 1999). Furthermore, post-synaptic changes in intracellular Ca2+ mediate dendrite growth or retraction (Wong & Ghosh, 2002). Long-term potentiation (LTP) is an increase in synaptic transmission induced by a brief, large magnitude rise in postsynaptic Ca2+ and long-term depression (LTD) is a decrease in synaptic strength induced by a modest and prolonged rise in intracellular Ca2+.

Ca2+ and the aging synapse

Two major principles of the Ca2+ hypothesis of aging synaptic function are that the regulation of Ca2+ is altered with advanced age, and that this disruption in Ca2+ homeostatic processes can result in changes in the transmission of information through brain systems (Foster & Norris, 1997; Foster, 1999). The transmission properties most readily affected include cell excitability and the induction of synaptic plasticity. For example, while LTP magnitude does not change with age, induction impairments are observed using weak stimulation parameters. Moreover, the susceptibility to LTD increases with advanced age. The increased susceptibility to LTD may lead to synaptic loss and a reduction in brain volume (Zhou et al., 2004; Shinoda et al., 2005; Foster, 2006). Overall, studies agree that aging is associated with a shift in the mechanisms that regulate the induction or threshold of Ca2+-dependent synaptic modification rather than a loss of expression mechanisms. In turn, the shift in synaptic plasticity favoring synaptic depression, along with reduced excitability, may constitute a functional lesion preventing the transfer of information through brain structures. Thus, changes in synaptic plasticity represent biological markers of Ca2+ dysregulation during senescence and a hypothesized mechanism for cognitive decline.

Much work has been devoted to determining the nature of Ca2+ dysregulation during aging. While the direction of synaptic plasticity is determined by the level and duration of a rise in intracellular Ca2+, the source of Ca2+ influx can also have a major influence on the susceptibility to synaptic modification. The major sources of intracellular Ca2+ include influx into the cell through N-methyl-d-aspartate (NMDA) receptors or voltage-dependent Ca2+ channels (VDCC) and release of Ca2+ from intercellular Ca2+ stores (Fig. 1). Our work and that of others indicate that, in the hippocampus, aging is associated with a decreased involvement of NMDA receptors, and an increased role for VDCC and internal Ca2+ stores in regulating synaptic modifiability (Foster & Norris, 1997; Foster, 1999; Kumar & Foster, 2005).

Figure 1.

The major sources of intracellular Ca2+ include influx into the cell through N-methyl-d-aspartate (NMDA) receptors or voltage-dependent Ca2+ channels (VDCCs) and release of Ca2+ from intercellular Ca2+ stores. Normally, the level of Ca2+ in the extracellular fluid is approximately 2 mm and resting levels in the neuron are an order of magnitude lower. With neural activity Ca2+ enters the cell through VDCCs and NMDA receptors. Ca2+ can be released from intracellular stores through Ca2+ induced Ca2+ release following activation of the ryanodine receptor (RyR) or metabotropic receptor (e.g. mGluR) activation of the inositol 1,4,5-trisphosphate (IP3) pathway. Depending on the magnitude and duration of a rise in intracellular Ca2+, a number of different signaling cascades and biological processes are modified.

Physiological model for Ca2+ dysregulation in mediating altered synaptic modifiability

Figure 2 illustrates a physiological model linking altered synaptic plasticity to a shift in the source of Ca2+. The NMDA receptor is the major source of Ca2+ for induction of LTP and NMDA receptor activation depends on post-synaptic depolarization to relieve the Mg2+ blockade of the receptor channel. A decrease in NMDA receptor activation may underlie the observation that LTP induction is more difficult and less reliable with advanced age. As noted above, LTP impairments are not observed when strong depolarizing stimuli are employed. Rather, induction deficits are observed for stimulation parameters near the threshold for induction. A major constraint on NMDA receptor activation and induction of LTP arises due to hyperpolarization of the post-synaptic neuron associated with Ca2+ from two other Ca2+ sources, VDCCs, and intracellular Ca2+ stores. In fact, the expanded capacity of these two Ca2+ sources in regulating synaptic plasticity is related to another physiological marker of aging; augmentation in the amplitude and duration of the Ca2+-dependent, K+-mediated afterhyperpolarization (AHP) (Fig. 2). The AHP amplitude is a critical component for models of altered synaptic plasticity in aged animals (Foster & Norris, 1997; Foster & Kumar, 2002) and the contribution of internal Ca2+ stores for regulation of the AHP is of particular importance. Our previous work (Kumar & Foster, 2004) demonstrated that inhibition of Ca2+ release from intracellular stores has a greater effect in reducing the AHP in aged (∼40%), relative to young animals (∼15%). The source of Ca2+ from internal stores was specific to ryanodine receptors (RyR) and Ca2+-induced Ca2+ release, as the AHP was not reduced by metabotropic receptor antagonisms, suggesting that the inositol 1,4,5-trisphosphate pathway was not involved. In addition to reducing the AHP in aged animals, blockade of RyRs enabled the induction of LTP during low frequency (5 Hz) synaptic activation. These results demonstrate that intracellular Ca2+ stores are a major contributor to the larger AHP and to regulating the threshold for synaptic plasticity in older animals.

Figure 2.

Physiological model linking altered synaptic plasticity thresholds to a shift in the source of Ca2+. A modest rise in Ca2+ from voltage-dependent Ca2+ channels (VDCC) and intracellular stores promote induction of long-term depression (LTD). In addition, the rise in Ca2+ enhances the afterhyperpolarization (AHP) (upper left inset), the hyperpolarization inhibits activation of N-methyl-d-aspartate (NMDA) receptors resulting in an increase in the long-term potentiation (LTP) threshold. A shift in the balance of LTD/LTP results in decreased synaptic transmission.

Reduction in the AHP is thought to facilitate the synaptic activity-mediated post-synaptic depolarization required for NMDA receptor channel opening. Our initial finding that intracellular Ca2+ stores are a major determinant of the larger AHP was recently confirmed and extended to middle-aged animals (Gant et al., 2006). As such, the growth of the AHP in middle-aged animals may contribute to the emergence of impaired LTP within this age range (Rex et al., 2005). These mechanisms may explain why treatments such as estrogen, which decrease release of Ca2+ from intracellular stores (Chen et al., 1998), can have a greater effect in reducing the AHP in aged animals (∼40%) relative to young animals (∼20%) when the AHP amplitudes are matched (Kumar & Foster, 2002). Furthermore, the estrogen-mediated decrease in AHP is likely to contribute to a reduction in the threshold for induction of LTP in aged females (Sharrow et al., 2002).

Evidence has been provided for an age-related increase in the number of L-type VDCCs (Thibault & Landfield, 1996). Interestingly, blockade of L-channels reduces the AHP amplitude in a similar manner in young and old animals (Disterhoft et al., 2004). Regardless, the reduction in the AHP by blockade of L-type VDCCs facilitates LTP induction in aged animals (Norris et al., 1998b). Despite the fact that LTP requires a large rise in intracellular Ca2+, these studies demonstrate that LTP induction is facilitated by blocking specific Ca2+ sources (VDCC and intracellular stores) that contribute to the AHP. The idea that the AHP sets the threshold for LTP is supported by work showing that LTP induction is facilitated by reducing the AHP amplitude through blockade of K+ channels (Norris et al., 1998b). Finally, facilitation of LTP following inhibition of intracellular Ca2+ stores can be reversed by application of an L-channel agonist, resulting in a subsequent increase in the AHP (Kumar & Foster, 2004). In this case, a treatment that promotes Ca2+ influx impairs LTP if Ca2+ influx augments the AHP amplitude. The results provide strong support for a model of impaired LTP threshold (Fig. 2) whereby an increase in Ca2+ from L-type channels and intracellular stores contributes to a larger AHP, and the subsequent hyperpolarizing response decreases NMDA receptor involvement in synaptic plasticity with advanced age.

In contrast to impairment in the induction of LTP, the ability to induce LTD is enhanced in aged animals. The alteration is not due to change in the expression of LTD; rather, the increased propensity for LTD observed under standard recording conditions is due to a reduction in the LTD threshold. The dependence of LTD induction on Ca2+ regulation is readily demonstrated by altering the Ca2+ : Mg2+ ratio in the recording medium. Little or no LTD is observed for adult animals under Ca2+ : Mg2+ conditions that mimic cerebral spinal fluid; however, robust LTD can be observed in adults when Ca2+ levels are elevated in the recording media (Norris et al., 1996). Likewise, more effective induction protocols, including paired-pulse low-frequency stimulation can induce robust LTD in younger animals (Kemp & Bashir, 1999). Like the modification in the threshold for LTP, the increased susceptibility to LTD during aging results from a shift in the source of Ca2+. In aged animals, blockade of NMDA receptors does not prevent LTD suggesting that NMDA receptors are not the major contributors for this form of Ca2+-dependent plasticity (Norris et al., 1996). In contrast, LTD in aged animals is inhibited by blockade of L-type channels or RyR antagonists (Norris et al., 1998b; Kumar & Foster, 2005). These same treatments reduce the AHP. Indeed, recent studies indicate that facilitated induction of LTD may depend on hyperpolarization of the cell and suppression of NMDA receptors (Azad et al., 2004; Froemke et al., 2005).

Together, the results paint a picture of a shift in Ca2+ homeostasis and modulation of synaptic plasticity in the aged brain involving reduced Ca2+ influx from NMDA receptors. In addition, VDCCs can interact with intracellular stores (Fagni et al., 2000) resulting in increased Ca2+ release from intracellular stores. The Ca2+ from intracellular stores underlies an increase in the amplitude and duration of the AHP and the AHP acts to limit NMDAR activation, adjusting the threshold for induction of synaptic modification (Kumar & Foster, 2004, 2005).

Biochemical model for Ca2+ dysregulation in mediating altered synaptic function

In addition to a physiological model, we have provided evidence for a biochemical model of altered Ca2+ signaling (Fig. 3), involving a shift in the activity of Ca2+-dependent kinases/phosphatases. These kinases and phosphatases are thought to act on α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptors to mediate the expression of LTP and LTD, respectively. In order to identify mechanisms for age-related changes in kinase/phosphatase activity, it is important to determine whether changes are due to altered expression, localization, or activation of enzymes. In most cases, the expression of a particular kinase or phosphatase is unaltered. However, for several kinases that are important for synaptic function (PKC, PKA, CaMII), the basal activity is reduced, localization is shifted, and stimulation induced activation is impaired (Foster, 2004). In contrast, phosphatase activity may be elevated in normal aging. Evidence for this model comes from a study demonstrating that inhibition of protein kinases selectively decreases synaptic strength in adults, and post-synaptic inhibition of protein phosphatase 1 (PP1), an enzyme critical for LTD (Mulkey et al., 1994), selectively increases synaptic transmission in aged rats (Norris et al., 1998a). Interestingly, genetic inhibition of this LTD-inducing pathway facilitates memory in aged animals (Genoux et al., 2002).

Figure 3.

Biochemical model of altered Ca2+ signaling resulting in biological markers of aging. The modest rise in Ca2+ during neural activity can activate calcineurin (CaN). In turn, CaN dephosphorylates proteins, cAMP response element binding protein (CREB), and BAD involved in gene regulation and cell health. CaN activation of protein phosphatase 1 (PP1) results in dephosphorylation of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptors and a decrease in synaptic transmission.

The connection between PP1 activity and Ca2+ regulation is thought to be mediated by an age-related increase in calcineurin (CaN) activity (Foster et al., 2001). CaN is the only known phosphatase whose activity is directly regulated by the level of intracellular Ca2+. Activation of CaN results in the dephosphorylation of an inhibitory regulator of PP1, permitting increased PP1 activity. The increase in CaN activity in aged animals does not appear to be due to an inability of CaN to associate with membrane binding proteins (Foster et al., 2001). However, a large rise in oxidative stress in very old animals may result in CaN aggregates in the cytosol (Agbas et al., 2005). In this case, aggregate formation is likely to reduce the regulation imposed by targeting proteins (Dell’acqua et al., 2006).

In addition to altered synaptic transmission, the increase in phosphatase activity contributes to changes in the phosphorylation state of CaN substrates including the bcl-2 family member protein BAD, and the cAMP response element binding protein (CREB) (Foster et al., 2001; Monti et al., 2005). Dephosphorylation of BAD and CREB is associated with several characteristics of aging including decreased cell viability, increased susceptibility to neurotoxicity, and impaired memory. In this way, Ca2+ dysregulation is connected to several markers of aging through increased CaN activity.

Intracellular Ca2+ coming from several sources contribute to increased CaN activation. The age-related increase in CaN activity is reduced by L-channel antagonists (Foster et al., 2001). In addition, CaN activity may be preferentially increased by NMDA receptor activity in aged animals (Jouvenceau & Dutar, 2006). Furthermore, mild oxidative stress is thought to increase CaN activity through release of Ca2+ from intracellular stores (Kamsler & Segal, 2004) or a decreased effectiveness of inhibitory proteins (Lin et al., 2003). Thus, Ca2+ dysregulation can be linked to markers of aging including a decrease in synaptic transmission, through a shift in the balance of activity of Ca2+-dependent kinase/phosphatase enzymes that mediate the direction of synaptic plasticity.

Mechanism for altered Ca2+ regulation

A major gap in our knowledge of age-dependent synaptic dysfunction concerns the mechanisms for altered Ca2+ homeostasis to promote L-channels and RyRs over NMDA receptors. Despite considerable research, it is unclear whether the expression of NMDA receptors is altered with age (Foster, 2002), and far less work has been directed at determining age differences in the expression of L-channels or RyRs. Alternatively, post-translational modifications are likely to contribute to the shift in activity. These Ca2+ sources are influenced by phosphorylation state. As noted above, CaN activity is increased with age. CaN dephosphorylation of the NMDA receptor complex decreases NMDA receptor function (Lieberman & Mody, 1994; Wang & Salter, 1994). Furthermore, oxidative stress is thought to increase release of Ca2+ from intracellular stores, which then activates CaN (Kamsler & Segal, 2004). This would suggest that release of Ca2+ from intracellular stores could decrease NMDA receptor function through CaN activation. Indeed, in aged animals, blockade of Ca2+ release from intracellular stores increases NMDA receptor function and this increase in the NMDA receptor response is not observed under conditions of CaN inhibition (Kumar & Foster, 2004).

Normal aging is associated with increased oxidative stress, which can influence synaptic plasticity (Serrano & Klann, 2004). However, reactive oxygen species (ROS) are also essential components of normal cellular processes including synaptic function (Klann et al., 1998; Klann, 1998). The increase in oxidative stress with age is likely due to changes in the balance of the production of ROS and mechanisms for scavenging these damaging molecules. To defend against ROS, cells employ antioxidant molecules and enzymatic mechanisms. Both the antioxidant mechanisms and sources of oxidative stress are highly localized within intracellular and extracellular compartments. Thus, ROS arise from different locations in association with synaptic activity (Lafon-Cazal et al., 1993; Hongpaisan et al., 2004; Tejada-Simon et al., 2005) and mitochondrial oxidative metabolism (Papa & Skulachev, 1997; Sastre et al., 2003; Genova et al., 2004). In turn, superoxide dismutase (SOD), which converts superoxides to the less reactive hydrogen peroxide, is differentially localized to the cytosol (SOD-1), mitochondria (SOD-2), and extracellular space (SOD-3). Interestingly, overexpression of cytosolic and extracellular Cu/ZnSOD (SOD-1, SOD-3) has age-dependent influences on memory and synaptic plasticity (Kamsler & Segal, 2003b; Hu et al., 2006). Moreover, acute application of hydrogen peroxide or antioxidant enzymes differentially influence synaptic plasticity in an age-dependent manner (Kamsler & Segal, 2003a; Watson et al., 2006) suggesting that memory and synaptic plasticity impairments could result from a shift in the management of ROS with age.

The exact mechanism for ROS-mediated regulation of synaptic plasticity is unknown; however, evidence suggests an influence on Ca2+ sources and Ca2+-dependent signaling cascades. For example, an increase in extracellular oxidative stress results in an increase in the influx of Ca2+ from L-type channels in the heart (Campbell et al., 1996; Yamaoka et al., 2000; Sims & Harvey, 2004) and in hippocampal neurons (Lu et al., 2002; Akaishi et al., 2004). Cysteine is the amino acid most susceptible to oxidation and oxidation of cysteines on RyRs increases Ca2+ release from intracellular stores (Hidalgo, 2005). Interestingly, oxidation of extracellular cysteine residues on the NMDA receptor decreases NMDA receptor activity (Lipton et al., 2002). Thus, age-related changes in the contribution of various Ca2+ sources appear to be mimicked by treatments that increase oxidative stress.

Figure 4 illustrates possible mechanisms for age-related alterations in Ca2+ sources. The function of the major Ca2+ sources (NMDA receptor, VDCC, and intracellular stores) is determined by regulatory mechanisms involving phosphorylation state and the generation of reactive oxygen species. An age-related increase in oxidative stress is likely to contribute to both a rise in intracellular Ca2+ originating from L-channels and intracellular stores, as well as a decrease in Ca2+ influx through NMDA receptors. In turn, Ca2+ from L-channels and intracellular stores can act in concert with oxidative stress to promote CaN activity, which further decreases NMDA receptor function. Thus, oxidative stress could initiate a cascade of events to shift the source of Ca2+ resulting in the expression of several markers of brain aging.

Figure 4.

Model of possible mechanisms underlying age-related alterations in Ca2+ sources. The function of the major Ca2+ sources, N-methyl-D-aspartate (NMDA) receptor, voltage-dependent Ca2+ channel (VDCC), and intracellular stores is determined by regulatory mechanisms involving reactive oxygen species (ROS), which can increase Ca2+ flux through VDCCs and ryanodine receptor (RyR) and decrease Ca2+ flux through NMDA receptors. The rise in intracellular Ca2+ increases the afterhyperpolarization (AHP), which can inhibit NMDA receptor activation. In addition, the rise in intracellular Ca2+ activates CaN and dephosphorylation of the NMDA receptor decreases the function of this receptor.

Moreover, a shift in hormonal status with increased release of corticosteroids or reduced estrogen could facilitate brain aging. Elevated corticosteroid levels associated with intense, inescapable, or unpredictable stress enhance the AHP, modify synaptic plasticity, and impair learning on a spatial discrimination task (Foster, 1999; Krugers et al., 2005). Thus, 5 days of behavioral stress produces an increase in the susceptibility to LTD induction lasting weeks to months (Artola et al., 2006). Furthermore, exposure to 8 days of swim stress enhances LTD in cognitively intact aged animals (Lee et al., 2005). Stress-mediated changes appear to involve activation of glucocorticoid receptors resulting in impaired NMDA receptor-dependent plasticity and an enhanced role of VDCCs in synaptic plasticity (Krugers et al., 2005). In contrast, a mild or acute stress, and activation of mineralocorticoid receptors, reduces the AHP amplitude, promotes induction of LTP and facilitates learning (Weiss et al., 2005; Ahmed et al., 2006; Kumar & Foster, 2007). Together the results suggest that a shift in the threshold for synaptic plasticity or alterations in the induction mechanism may be due to age-related impairments in the adaptive response to environmental or behavioral stress.

Finally, the alternative must be considered, that the age-related changes may represent compensatory mechanisms to protect neurons from toxic levels of Ca2+ (Foster, 1999). Thus, the shift in Ca2+ sources and a subsequent increase in the AHP and altered synaptic plasticity, resulting in decreased cell excitability and decreased synaptic transmission, may represent adaptive processes that attempt to limit toxic levels of Ca2+ associated with neurodegenerative diseases. As such, an important line of research will be to determine whether treatments that ameliorate or promote the aging biomarkers will provide benefits or detriments to memory and cell health.

Relationship between normal aging and neurodegenerative diseases

Together, the results indicate that the age-related decrease in synaptic strength is mediated through the Ca2+-dependent enzyme pathways linked to LTP and LTD. Furthermore, the results are consistent with the idea that an enhancement of LTD-induction mechanisms with advanced age is associated with an early characteristic of age-related cognitive decline; forgetting of rapidly acquired hippocampal-dependent memory. The physiological and biochemical models suggest potential treatments that might ameliorate brain aging (Foster, 2006). Possible avenues include adjustment of Ca2+ influx, nootropic drugs that enhance synaptic transmission, and neuromodulators that indirectly influence Ca2+-dependent processes including steroids. Early intervention may be particularly important in decreasing oxidative damage, which may underlie the subsequent cascade of age-related events leading to Ca2+ dysregulation (Foster, 2006). Other avenues include correcting factors that might promote Ca2+ dysregulation such as age-related changes in hormonal function, specifically, decreased gonadal steroids (Foster, 2005) and increased glucocorticoids (Foster & Norris, 1997; Foster, 1999). In contrast, genetic or environmental factors may interact with normal aging to promote neurodegeneration. Thus, β-amyloid or fragments of the protein can disrupt Ca2+ homeostasis, synaptic plasticity, and memory in a manner analogous to normal aging and prior to the emergence of neurodegeneration. Altered amyloid processing associated with Alzheimer's disease may disrupt the ability of neuromodulators to regulate the AHP amplitude (Ohno et al., 2004). Finally, β-amyloid reduces NMDA receptor function and impairs LTP through enhanced CaN activity (Chen et al., 2002). As the disease progresses, increased expression of CaN inhibitory proteins may shift the balance of Ca2+ activated kinases/phosphatases to promote tau phosphorylation, neurodegeneration, and tangle formation (Ermak et al., 2001; Cook et al., 2005).


The level of intracellular Ca2+ plays a central role in normal and pathological signaling within and between neurons. These processes involve a cascade of events for locally raising and lowering cytosolic Ca2+. As the mechanisms for Ca2+ dysregulation are illuminated, hypotheses concerning Ca2+ homeostasis and brain aging have been modified. The idea that senescence is due to pervasive cell loss associated with elevated resting Ca2+ has been replaced by concepts concerning changes in local Ca2+ levels due to neural activity. A shift in the relative contribution of various Ca2+ sources may provide an early marker that correlates with a rise in oxidative stress. Subsequently, changes in physiological activity including cell excitability and synaptic plasticity may be associated with benign cognitive decline such as increased forgetfulness. Over time, the shift in Ca2+ signaling pathways may have an influence on synaptic connectivity. Finally, disease states or injury may interact with age-related Ca2+ dysregulation, increasing the extent of Ca2+ dysregulation, and ultimately contributing to neurodegenerative processes.