|Home | About | Journals | Submit | Contact Us | Français|
Evidence accumulated over more than two decades has implicated Ca2+ dysregulation in brain aging and Alzheimer's disease (AD), giving rise to the Ca2+ hypothesis of brain aging and dementia. Electrophysiological, imaging, and behavioral studies in hippocampal or cortical neurons of rodents and rabbits have revealed aging-related increases in the slow afterhyperpolarization, Ca2+ spikes and currents, Ca2+ transients, and L-type voltage-gated Ca2+ channel (L-VGCC) activity. Several of these changes have been associated with age-related deficits in learning or memory. Consequently, one version of the Ca2+ hypothesis has been that increased L-VGCC activity drives many of the other Ca2+-related biomarkers of hippocampal aging. In addition, other studies have reported aging- or AD model-related alterations in Ca2+ release from ryanodine receptors (RyR) on intracellular stores. The Ca2+-sensitive RyR channels amplify plasmalemmal Ca2+ influx by the mechanism of Ca2+-induced Ca2+ release (CICR). Considerable evidence indicates that a preferred functional link is present between L-VGCCs and RyRs which operate in series in heart and some brain cells. Here, we review studies implicating RyRs in altered Ca2+ regulation in cell toxicity, aging, and AD. A recent study from our laboratory showed that increased CICR plays a necessary role in the emergence of Ca2+-related biomarkers of aging. Consequently, we propose an expanded L-VGCC/Ca2+ hypothesis, in which aging/pathological changes occur in both L-type Ca2+ channels and RyRs, and interact to abnormally amplify Ca2+ transients. In turn, the increased transients result in dysregulation of multiple Ca2+-dependent processes and, through somewhat different pathways, in accelerated functional decline during aging and AD.
It has been over 20 years since it was initially proposed that altered Ca2+ regulation might play a role in brain aging and Alzheimer's disease (AD) (Landfield, 1983, 1987; Khachaturian, 1984, 1989; Gibson & Peterson, 1987; Disterhoft et al., 1994). In brain neurons from aging rodents and rabbits, as compared to neurons from younger animals, Ca2+ influx associated with action potentials induces a larger Ca2+-dependent afterhyperpolarization (AHP) (Landfield & Pitler, 1984; Kerr et al., 1989; Moyer et al., 1992; Potier et al., 1992; Disterhoft et al., 1996, 2004; Stutzmann et al., 2006) and impairs short-term synaptic plasticity (Landfield et al., 1986; Thibault et al., 2001). Furthermore, these findings have been reinforced by studies showing that pharmacologically isolated Ca2+ action potentials (Pitler & Landfield, 1990; Disterhoft et al., 1996), whole-cell Ca2+ currents (Campbell et al., 1996), and Ca2+ transients during repetitive spike trains also are larger in hippocampal neurons from aged animals (Thibault et al., 2001; Hemond & Jaffe, 2005). Conversely, Ca2+ influx via ligand-gated N-methyl-D-aspartate (NMDA) receptor channels appears reduced in aged animals (Barnes et al., 1997; Magnusson, 1998; Shankar et al., 1998).
Our studies on this general Ca2+ dysregulation hypothesis have focused primarily on apparent excess Ca2+ influx via voltage-gated Ca2+ channels (VGCC) (Landfield, 1996; Thibault et al., 1998). Studies of the L-type VGCC (L-VGCC) antagonist suggested that the aging-related increase in Ca2+-mediated responses might depend on greater activity through L-VGCC (Moyer et al., 1992; Campbell et al., 1996). Increased L-VGCC activity with aging was confirmed directly by single channel recording in partially dissociated hippocampal slices (Thibault & Landfield, 1996). Moreover, changes in L-VGCCs appear to be functionally relevant, as L-VGCC antagonists improve learning and memory in aged animals (Deyo et al., 1989; Disterhoft et al., 2004) and some AD patients (Forette et al., 2002). Furthermore, the increase in L-VGCC density is positively correlated with cognitive impairment in aged animals (Thibault & Landfield, 1996).
In addition to the accumulating evidence of increased Ca2+ influx through L-VGCCs, there is also recent evidence that altered function of intracellular organelles might play a critical role in Ca2+ regulation during aging or AD (Toescu & Verkhratsky, 2003). In particular, changes in intracellular Ca2+ release from the endoplasmic reticulum (ER) appear likely to contribute to brain Ca2+ dyshomeostasis, and have been associated with changes in [Ca2+]i. Therefore, in this review, we summarize several lines of evidence implicating altered release from intracellular stores in aging and AD, and attempt to integrate this evidence with the role of Ca2+ influx in aging-related Ca2+ dysregulation.
Several comprehensive reviews have recently considered mechanisms associated with Ca2+ sequestration and release by the ER in both peripheral cells (Bootman et al., 2001; Berridge, 2002; Carafoli, 2002; Fill & Copello, 2002) and in neurons (Paschen & Mengesdorf, 2005; Verkhratsky, 2005). Accordingly, only the points most relevant to ER function in brain aging are briefly recapitulated here. Two distinct intracellular Ca2+ release channels are present in several types of muscle and brain cells, the inositol 1,4,5-trisphospate receptor (IP3R) and the ryanodine receptor (RyR), each having multiple isoforms in different tissues. These receptor channels function to amplify or trigger Ca2+ rises initiated by either plasmalemmal Ca2+ influx or ligand binding, thereby inducing Ca2+ signaling cascades. Amplification is achieved through either the actions of Ca2+-induced Ca2+ release (CICR), provided by RyR, or actions of IP3-induced Ca2+ release (IICR) through IP3Rs.
Originally described in skeletal and cardiac muscle cells, RyRs in the membrane of the sarcoplasmic reticulum are an integral and essential Ca2+ source for excitation-contraction coupling (Endo, 1977; Fill et al., 1989; Takeshima et al., 1989; Meissner, 1994). Furthermore, an apparent direct physical interaction, which favors alignment between L-VGCCs and RyRs, enables L-VGCCs to function as a preferred source of extracellularly derived Ca2+ in triggering CICR from RyRs and amplifying Ca2+ transients (Lu et al., 1994; Cheng et al., 1996; Wang et al., 2001). In the brain, similar Ca2+ amplification functions of RyRs have been identified, again mediated in part by a close juxtaposition to L-VGCCs (Chavis et al., 1996; Empson & Galione, 1997; Borde et al., 2000; Fagni et al., 2000; Sukhareva et al., 2002).
The other major source of intracellular Ca2+ occurs in response to stimulation of IP3Rs by IP3 generated from activation of a number of metabotropic G-protein-coupled receptors. In some cases IP3Rs can also trigger Ca2+-sensitive K+ channels and hyperpolarize neurons (Sawada et al., 1987; Fink et al., 1988; Furuichi et al., 1989; Zhang et al., 1990; Berridge, 1993; Khodakhah & Ogden, 1995; Irving & Collingridge, 1998; Taylor et al., 1999; Johenning et al., 2002; Rossi & Taylor, 2004). Moreover, IP3Rs are also sensitive to Ca2+ concentrations (Bezprozvanny et al., 1991; Missiaen et al., 1992; Tsukioka et al., 1994; Hagar et al., 1998) and, depending on the cell type studied, it appears that IP3R may also be favorably aligned with L-VGCCs or metabotropic glutamate receptors (mGluR), through interactions with the scaffold protein Homer 1a (Tu et al., 1998; Fagni et al., 2000; Yamamoto et al., 2005).
Release of Ca2+ from these two intracellular channels is regulated in part by the Ca2+ concentration gradient present between luminal ER Ca2+ and cytoplasmic Ca2+ (Alonso et al., 1999; Kiryushko et al., 2002; Solovyova et al., 2002) and is, thus, also dependent on the Ca2+-refilling function of sarcoplasmic/endoplasmic reticulum Ca2+-ATPases (SERCA). Sarcoplasmic/endoplasmic reticulum Ca2+-ATPases maintain the relatively high levels of Ca2+ in the ER (hundreds of µm) that serve CICR, and IICR, and, in the process, contribute to the control and reduction of cytosolic Ca2+ (Thastrup et al., 1990; MacLennan et al., 1997; Mogami et al., 1998; Meldolesi, 2001; Berridge, 2002; Verkhratsky, 2004).
Although cell culture models of Ca2+-dependent cell death are generally not viewed as clear models of brain aging, or even AD, they are often employed in studies of ischemic events. These events increase in frequency with advancing age, and it is also possible that neuronal vulnerability from such events increases with aging. Therefore, examining the role of Ca2+ release from intracellular stores in cell death models may help elucidate implications of aging-related alterations in intracellular release. In particular, delayed toxicity after exposure to high glutamate (GLU) in cell culture (excitotoxicity) is a common model used to mimic a wide range of neurological insults, including anoxia/ischemia, head and spinal cord trauma, and even chronic neurodegenerative diseases such as AD. Dysregulated Ca2+ homeostasis and altered Ca2+ influx through NMDA receptors were identified as primary contributors to neuronal cell death early in the study of excitotoxicity (Rothman & Olney, 1986; Choi et al., 1987; Wahl et al., 1989; Regan & Choi, 1991; Randall & Thayer, 1992; Dubinsky, 1993; Lu et al., 1994; Marks et al., 1996; Tymianski & Tator, 1996; Toescu, 1998; Lee et al., 1999; Limbrick et al., 2001; Lipton, 2004). In excitotoxicity models, Ca2+ dysregulation is frequently manifested as an irreversible Ca2+ rise or slowed Ca2+ clearance, and is ultimately associated with neuronal death.
Several investigations of excitotoxicity have focused on a potential role of the ER in sustained Ca2+ elevations. These studies have found that blocking CICR with high concentrations of ryanodine, which lock RyRs in a low conductance state (Bezprozvanny et al., 1991; Coronado et al., 1994; Humerickhouse et al., 1994), or irreversibly inhibiting SERCA function and passively emptying ER stores with thapsigargin prior to GLU exposure, reduces sustained Ca2+ plateaus, as well as other indices associated with neuronal cell death (e.g. lactate dehydrogenase (LDH) release) (Frandsen & Schousboe, 1991; Segal & Manor, 1992; Leski et al., 1999; Clodfelter et al., 2002). Similar results have been noted in models of stroke and ischemia, particularly in astrocyte preparations (Duffy & MacVicar, 1996; Kuwabara et al., 1996; Verkhratsky et al., 1998; Aley et al., 2006). Somewhat paradoxically, while short-term ER Ca2+ depletion prior to an insult appears protective against necrotic (excitotoxic) cell death, long-term depletion of ER Ca2+ induces apoptosis, as indicated by elevations of apoptotic markers, stress responses and disturbance in protein synthesis, and/or massive cell death (Doutheil et al., 1999; Mengesdorf et al., 2001; Verkhratsky & Petersen, 2002; Paschen, 2003; Verkhratsky & Toescu, 2003; Lindholm et al., 2006).
Thus, excessive release of Ca2+ from the ER may play an important role in excitotoxicity. Moreover, evidence suggests that such excessive release may be dependent on the relative maturity of the cells. It is well established that embryonic cortical/hippocampal neurons become increasingly vulnerable to GLU toxicity after a few weeks in culture (Choi, 1992; Toescu & Verkhratsky, 2000), an age in culture that coincides with the emergence of sustained Ca2+ plateaus following GLU insult (Attucci et al., 2002). Interestingly, ryanodine is particularly effective in reversing the Ca2+ plateau and in providing neuroprotection in older cultures (Fig. 1) (Clodfelter et al., 2002). Moreover, recent evidence suggests that the lethal Ca2+ plateau may be maintained by sustained Ca2+ influx via depolarized NMDA receptors (Norris et al., 2006). Together, these data indicate that the plateau may be sustained by CICR. Although age in culture is clearly not equivalent to brain aging, it is associated with increasing vulnerability and Ca2+ influx, which may model some aspects of normal aging (Porter et al., 1997). Conceivably therefore, if Ca2+ release from ER is altered with aging, this alteration may develop in parallel with altered Ca2+ influx (Clodfelter et al., 2002).
With the increasing development of transgenic (Tg) mouse models of AD, numerous studies testing the view that altered Ca2+ homeostasis might play a role in AD have recently emerged. Initial studies in fibroblasts from AD patients (Gibson et al., 1996) or in cells bearing the human presenilin 1 (PS1) AD mutation (Begley et al., 1999; Guo et al., 1999; Leissring et al., 1999; LaFerla, 2002; Stutzmann, 2005) found evidence of abnormal Ca2+ release through IP3R pathways (Leissring et al., 1999). Interestingly, abnormal IP3-mediated Ca2+ elevations in fibroblasts have also been seen in asymptomatic members of AD families (Etcheberrigaray et al., 1998).
Several studies also have implicated RyRs as being responsible for enhanced intracellular release in PS1 mutated animals (Chan et al., 2000; Mattson et al., 2000; Schneider et al., 2001; Popescu & Ankarcrona, 2004; Stutzmann et al., 2006). Smith and colleagues (2005) examined cultured cortical neurons from mice bearing a transgene containing three AD-related mutations (3×Tg mice), which develop both plaques and tangles, and observed an increase in RyR expression along with greater Ca2+ efflux from the stores in response to caffeine (an agonist at the RyR) (Smith et al., 2005). A recent study combining electrophysiological and Ca2+ imaging methods in cortical slice neurons from Tg mice bearing the PS1 mutation alone, or the 3×Tg transgene, or nontransgenic control animals, assessed the effects of aging vs. those of the PS1 mutation on ER release (Stutzmann et al., 2006). This study found that the PS1 mutation is a critical calciopathic mutation and that increased RyR expression is likely a major factor in the AD mutation-mediated enhancement of ER release. Although photolysis of IP3 was shown to evoke larger Ca2+ transients and Ca2+-dependent hyperpolarizations in Tg mice, the increase in IP3 effects was mediated by CICR from RyRs, triggered in response to IICR. However, some puzzling results also were seen. The enhanced IP3-mediated Ca2+ release and resulting hyperpolarization was larger in Tgs than in non-Tgs at all ages and did not change with aging in any Tg or non-Tg model. Conversely, the AHP induced by trains of spikes and VGCC activation increased with aging in all three model strains but did not differ between Tg and non-Tg mice (Stutzmann et al., 2006).
While little is known regarding underlying mechanisms, it appears that altered CICR, perhaps in combination with IICR, confer some of the phenotypes of disrupted Ca2+ homeostasis in neurons from 3×Tg mice. Still, other sources and mechanisms likely also contribute. The PS1 mutation (which, alone, does not induce amyloid plaques), in combination with amyloid precursor protein (APP) mutations, increases Aβ production (Mullan & Crawford, 1993; Price & Sisodia, 1994; Tanzi et al., 1996; Holcomb et al., 1998; Selkoe, 1998). Some studies have found that Aβ production can exacerbate Ca2+ responses to NMDA or GLU exposure (Mattson, 1997). Furthermore, Aβ toxicity has been attributed, in part, to effects on VGCCs (Davidson et al., 1994; Weiss et al., 1994; Ueda et al., 1997; Ramsden et al., 2002; Bobich et al., 2004; Webster et al., 2006), which could trigger CICR from IP3Rs or RyRs (Koizumi et al., 1998; Ferreiro et al., 2004). However, APP proteolysis (γ-secretase activity) alone does not appear sufficient, because the PS1 mutation (rather than other more amyloidogenic mutations) must be present for the Ca2+ dysregulation to occur (Stutzmann et al., 2006). A possible alternative mechanism suggests that presenilins form Ca2+ leak channels in ER membranes of mouse fibroblasts, independently of γ-secretase activity. Mutations in presenilin interfere with this leak function, and result in greater Ca2+ filling and release from ER (Tu et al., 2006). Furthermore, a gene microarray study conducted in autopsied hippocampal tissue from human AD patients (Blalock et al., 2004) found that multiple genes encoding proteins involved in ER receptor function, or in protein folding and chaperoning, which are also mediated in part by the ER, were down-regulated in incipient AD. These widespread changes may reflect ER membrane/receptor instability in sporadic AD as well.
In addition, it should be noted that effects of PS1 mutations on Ca2+ dysregulation have been observed to occur via other processes, including capacitative Ca2+ entry (Yoo et al., 2000; Smith et al., 2002; Herms et al., 2003; Zatti et al., 2006), changes in mitochondrial potential (Begley et al., 1999; Ankarcrona & Hultenby, 2002; Chan et al., 2002; Behbahani et al., 2006), and L-VGCCs (Cook et al., 2005). Clearly therefore additional work will be needed to resolve the relative contributions of the different sources to the Ca2+dysregulation seen in various models of neurodegenerative diseases.
Electrophysiological markers of brain aging have been extensively characterized in the hippocampal formation (Landfield & Pitler, 1984; Moyer et al., 1992; Barnes, 1994; Thibault et al., 1998; Norris et al., 1998; Disterhoft et al., 2004; Burke & Barnes, 2006), a region well-established to be important for memory processes and highly vulnerable to deleterious/degenerative changes with aging. Many of the consistent biomarkers of aging, such as the slow AHP (sAHP), are Ca2+-dependent or Ca2+-mediated. However, it is important to assess the degree to which the ER contributes to the established biomarkers of aging. Both CICR and IICR pools exist within the ER of hippocampal CA1 and CA3 pyramidal neurons. The amount of Ca2+ released via CICR and IICR depends on binding of intracellular ligands including Ca2+, cyclic ADP ribose (cADPR), nicotinic acid adenine dinucleotide phosphate (NAADP) or IP3 (Verkhratsky, 2005), and also depends on the Ca2+ sequestering capacity of the ER, which determines ER Ca2+ content ([Ca2+]L) (Verma et al., 1992; Murayama & Ogawa, 1996; Dawson, 1997; Garaschuk et al., 1997). Solovyova and colleagues using a dual indicator loading technique (low affinity indicator for imaging Ca2+ in the ER, and high affinity indicator for imaging Ca2+ in the cytosol) were able to show that the resting [Ca2+]L in sensory neurons is in the range of 200–300 µm, and high concentrations of IP3 or caffeine result in approximately a 40% decrease in luminal Ca2+ (Solovyova et al., 2002). Depolarization induced [Ca2+]L release was less effective, ranging from 5 to 30 µm. Other techniques for imaging Ca2+ within the ER include the use of aequorin or cameleons. However, there are limitations with these techniques, as the Ca2+ reporting proteins must be genetically engineered and selectively targeted to the ER (Miyawaki et al., 1997; Alonso et al., 1998; Solovyova & Verkhratsky, 2002). In addition, they require long incubation times for transfecting and loading and, thus, preclude their use in acute brain slices.
Consequently, there have been only a handful of studies in neurons examining the effects of aging on ER Ca2+ concentration and release, or on RyR expression. Studies focusing on measures of ER Ca2+ content have generally relied on the use of single wavelength indicators to measure changes in [Ca2+]i transients activated by caffeine, and have found varying results, depending on the experimental approach or preparation. In an early study, no net change in ER Ca2+ release with aging was reported in synaptosomes from the whole brain (Martinez-Serrano et al., 1992). More recently, acute dissociation of several brain regions (cerebellar, basal forebrain, and hippocampal neurons) from aged animals found that CICR magnitude was reduced and that Ca2+ transients recovered more slowly (Verkhratsky et al., 1994; Kirischuk & Verkhratsky, 1996; Murchison & Griffith, 1999; Xiong et al., 2002; Alshuaib et al., 2006). In studies focusing on RyR expression, no clear pattern or consistent changes have been seen in neurons of normal aging rats and mice. Two studies reported no change in brain RyR expression during aging (Martini et al., 1994; Stutzmann et al., 2006), although a recent study of peripheral neurons found a transient elevation in protein levels (RyR3) in mid-aged rats (Vanterpool et al., 2006).
Another approach to the investigation of the possible role of the ER in brain aging is to examine the effects of aging on Ca2+-dependent processes that are modulated, in part, by intracellular Ca2+ release. In CA1 neurons, postsynaptic injection of IP3 or of RyR inhibitors prevents the induction of long-term potentiation and attenuates paired-pulse facilitation (Wang & Kelly, 1997). Similarly, bath application of thapsigargin or cyclopiazonic acid (blockers of SERCA) prevents the induction of long-term depression in both single neurons and in field potential measures (Reyes & Stanton, 1996). High concentrations of ryanodine also selectively reduce the sAHP and spike-frequency accommodation (Borde et al., 2000; Shah & Haylett, 2000). While examining the effect of aging on long-term depression induction (Norris et al., 1998), Foster and colleagues recently reported that cyclopiazonic acid, thapsigargin or ryanodine (agents that reduce CICR) all prevented long-term depression in aged neurons (Kumar & Foster, 2005). However, long-term potentiation, which tends to be decreased with aging (Burke & Barnes, 2006), was enhanced by high ryanodine concentrations in aged slices (Kumar & Foster, 2004). Ca2+-dependent processes mediated largely by IICR and mGluRs activation also have been shown to change with aging. Compared to younger animals, type 1 mGluR activation results in a reduced phosphoinositide turnover in aged rats, perhaps mediated by a reduction in phospholipase C activity (Nicolle et al., 1999). Similarly, protein kinase C (PKC) was also reported to show reduced activity in aging neurons (Araki et al., 1994; Pascale et al., 1998).
Thus, the evidence on the nature of altered CICR or IICR in neurons of normally aging mammals is somewhat inconsistent, perhaps reflecting the type of preparation, cell or brain region specificity, or the difficulty in imaging Ca2+ and its sources within the intact hippocampal slice (Brown & Jaffe, 1994). Recently therefore we sought to systematically test the contributions of CICR to aging changes in one of the brain regions studied most extensively in relation to aging (hippocampus). Specifically, we tested the key prediction that, if increased CICR plays a major role in normal brain aging, then blocking it with high concentration ryanodine should reduce the aging differences in multiple Ca2+ biomarkers of aging.
More broadly, in fact, several other important tenets of the overall Ca2+ hypothesis have, for some time, required adequate testing. These tenets and predictions include: (i) if a common mechanism of Ca2+ dysregulation underlies many aspects of brain aging, then multiple Ca2+-dependent biomarkers of aging in the hippocampus should emerge at approximately the same age in adulthood; and (ii) if Ca2+ dysregulation is a major factor in cognitive decline then Ca2+ biomarkers should precede or coincide with the earliest age of cognitive impairment, which in some studies of rats has been as early as 12-months old (approximately mid-life). To test these predictions and the involvement of CICR on the emergence of Ca2+-related biomarkers, we recently conducted an extensive age course study combining electrophysiological and Ca2+ imaging techniques in hippocampal slices from male rats. Animals at five age points were used to identify the age of onset for three Ca2+-mediated markers of aging, the sAHP, spike accommodation, and the synaptically activated Ca2+ transient. A subset of hippocampal slices received a high dose of ryanodine to block the contribution of CICR to the overall Ca2+ response. In this study, we also employed the least invasive procedures available (sharp intracellular electrodes instead of patch clamping electrodes, nondissociated slices) to minimize interactions of preparation trauma and age.
Results were consistent with the above predictions. That is, ryanodine essentially eliminated aging differences in the three markers (e.g. the sAHP, Fig. 2), and the three biomarkers were first detectable simultaneously and at 12 months of age (Fig. 2), an age range early enough to account for cognitive decline. The ryanodine-sensitive component of the Ca2+ response (i.e. CICR) during a 20-s train of synaptic spikes appears to be minimal in young neurons compared to aged neurons and, notably, CICR contributed most to the [Ca2+]i elevation during the first few seconds of the train (Fig. 3). This rapid ‘booster’ action of CICR on Ca2+ responses is consistent with its strong effect on the AHP (Fig. 2) (Gant et al., 2006).
Thus, results of this large study provide considerable support for the proposition that in the hippocampus, an aging-related increase in CICR is necessary, from the onset, for the development of aging changes in several Ca2+-related processes. Moreover, the findings may help to resolve some of the contradictions in the literature by elucidating the conditions under which the contributions of CICR are most prominent. However, one apparent paradox is that similar kinds of evidence support a critical role for L-VGCCs in aging-related Ca2+ dysregulation (Thibault et al., 1998; Disterhoft et al., 2004). Nevertheless, these two lines of evidence are not necessarily contradictory, given that L-VGCCs and RyRs appear to operate in series in many cell types. In this view, then, Ca2+ influx via L-VGCCs may be the preferred source for triggering elevated CICR in aging. Together, the data suggest that aging changes in both types of channel may be part of the same pathway of dysregulation, in turn, suggesting the utility of expanding this version of the Ca2+ hypothesis to incorporate the results on Ca2+ release from intracellular stores (Fig. 4).
The work summarized above points to the following basic conclusions:
We thank Dr. Nada Porter for her valuable input and editorial comments on the manuscript. Our research described here was supported by grants R37-AG04542, PO1-AG10836, and T32-AG00242 from the National Institute on Aging and P20-RR15592 from the National Institutes of Health.