Oligomeric forms of β-amyloid (Aβ) peptides associated with Alzheimer's disease (AD) disrupt cellular Ca2+ regulation by liberating Ca2+ into the cytosol from both extracellular and intracellular sources. We elucidated the actions of intracellular Aβ42 by imaging Ca2+ responses to injections of Aβ oligomers into Xenopus oocytes. Two types of signal were observed: (1) local, “channel-like” transients dependent on extracellular Ca2+ influx, which resembled signals from amlyoid pores formed by extracellular application of oligomers; and (2) local transients and global Ca2+ waves, resembling Ca2+ puffs and waves mediated by inositol trisphosphate (IP3). The latter responses were suppressed by antagonists of the IP3 receptor (caffeine and heparin), pretreatment with the Gi/o-protein inhibitor pertussis toxin, and pretreatment with lithium to deplete membrane inositol lipids. We show that G-protein-mediated stimulation of IP3 production and consequent liberation of Ca2+ from the endoplasmic reticulum by intracellular Aβ oligomers is cytotoxic, potentially representing a novel pathological mechanism in AD which may be further exacerbated by AD-linked mutations in presenilins to promote opening of IP3 receptor/channels.
Alzheimer's disease (AD) is characterized by the abnormal proteolytic processing of amyloid precursor protein, resulting in increased production of a self-aggregating form of β-amyloid (Aβ) (Haass et al., 1992; Small et al., 2010). Strong evidence indicates that soluble Aβ aggregates represent the toxic species in the etiology of AD by promoting uncontrolled elevation of cytosolic Ca2+ levels (Walsh et al., 2002; Kayed et al., 2003; Demuro et al., 2005, 2010,Deshpande et al., 2006; Bezprozvanny and Mattson, 2008; Green and LaFerla, 2008; Berridge, 2010). One source of Ca2+ arises from the action of extracellular Aβ oligomers to disrupt the integrity of the plasma membrane via mechanisms proposed to include destabilization of the membrane lipid structure (Hertel et al., 1997; Mason et al., 1999; Sokolov et al., 2006), activation of endogenous channels (Wang et al., 2000; De Felice et al., 2007; Alberdi et al., 2010), and formation of intrinsic Aβ channels in the cell membrane (Arispe et al., 1993; Pollard et al., 1993; Lin et al., 2001; Quist et al., 2005; Demuro et al., 2011).
Intracellular actions of Aβ are further likely to contribute in the pathogenesis of AD because intracellular Aβ accumulation has been shown to precede extracellular deposits (Gouras et al., 2000), and the endoplasmic reticulum (ER) of neurons has been identified as the specific site of intracellular Aβ production (Hartmann et al., 1997). Importantly, intra-neuronal accumulation of Aβs has been shown to lead to a profound deficit of long-term potentiation and cognitive dysfunction in AD mice models (Oddo et al., 2003; Knobloch et al., 2007). The specific mechanisms underlying the intracellular toxicity of Aβ have not yet been established. However, in addition to promoting influx of extracellular Ca2+, there is also evidence that Aβ oligomers evoke the liberation of Ca2+ from intracellular stores (Ferreiro et al., 2004; Demuro et al., 2005).
Here, we examined the processes underlying Ca2+ mobilization by intracellular Aβ, using the Xenopus oocyte as a model cell system because its large size enables direct microinjection of amyloid oligomers into the cytoplasm. We show that injection of Aβ42 oligomers, but not monomers or fibrils, potently evokes two types of cytosolic Ca2+ signals: (1) local transients that are dependent on extracellular Ca2+ and that resemble the multistep channel-like signals described previously from amyloid pores formed in the plasma membrane by extracellular application of Aβ oligomers (Demuro et al., 2011); and (2) local “puff-like” signals, repetitive global Ca2+ waves, and sustained Ca2+ elevations that closely resemble the hierarchy of events evoked by release of Ca2+ from the ER mediated by inositol 1,4,5-trisphosphate (IP3) (Callamaras et al., 1998). The intracellular Ca2+ release signals are blocked or substantially reduced by antagonists of the IP3 receptor (IP3R) and are abolished by pretreatment with pertussis toxin (PTX) to block G-protein-mediated activation of phospholipase C (PLC) and by blocking the recycling of membrane inositol lipids by lithium, although Ca2+ signals evoked by photorelease of IP3 are unaffected by the latter treatments. Moreover, intracellular injection of Aβ oligomers causes acute cytotoxicity in the absence of extracellular Ca2+, but this effect is abrogated by blocking IP3 production or by chelating cytosolic Ca2+. We thus conclude that the stimulation of IP3 production by intracellular Aβ42 oligomers and consequent IP3-mediated liberation of Ca2+ from ER stores may contribute importantly to Ca2+ signaling disruptions and neurotoxicity in AD.
Materials and Methods
Oocyte preparation and electrophysiology.
Xenopus laevis were purchased from Nasco International, and oocytes were surgically removed (Demuro et al., 2005) following protocols approved by the University of California, Irvine Institutional Animal Care and Use committee. Stage V–VI oocytes were isolated and treated with collagenase (1 mg/ml collagenase type A1 for 30 min) to remove follicular cell layers. Intracellular microinjections were performed using a Drummond microinjector. Approximately 1 h before imaging, oocytes in Ca2+-free Barth's solution were injected with fluo-4 dextran (low affinity; Kd for Ca2+ ∼3 μm) to a final concentration of 40 μm assuming equal distribution throughout a cytosolic volume of 1 μl. In some experiments, caged d-myo-inositol 1,4,5-trisphosphate P4(5)-[1-(2-nitrophenyl)ethyl]ester was also injected to final concentration of 8 μm. Oocytes were then placed animal hemisphere down in a chamber with a base formed by a fresh, ethanol-washed microscope cover glass and were bathed in a Ca2+-containing Ringer's solution (in mm: 110 NaCl, 2 KCl, 1.8 CaCl2, and 5 HEPES, pH 7.2.) or a Ca2+-free Ringer's solution in which CaCl2 was omitted and replaced by 2 mm EGTA and 5 mm MgCl2. In the experiments of Figure 2, a two-electrode voltage clamp (Geneclamp; Molecular Devices) was used to control the membrane potential; all other experiments were performed at resting membrane potential.
Ca 2+ imaging.
Oocytes were imaged at room temperature by wide-field epifluorescence microscopy using an Olympus inverted microscope (IX 71) equipped with a 60× oil-immersion objective, a 488 nm argon–ion laser for fluorescence excitation, and a CCD camera (Cascade 128+; Roper Scientific) for imaging fluorescence emission (510–600 nm) at frame rates of 30 to 100 s−1 (Fig. 1A). Fluorescence was imaged within a 40 × 40 μm region within the animal hemisphere of the oocyte, and measurements are expressed as a ratio (ΔF/F0) of the mean change in fluorescence at a given region of interest (ΔF) relative to the resting fluorescence at that region before stimulation (F0). Mean values of F0 were obtained by averaging over several frames before stimulation. MetaMorph (Molecular Devices) was used for image processing, and measurements were exported to Microcal Origin version 6.0 (OriginLab) for analysis and graphing.
Preparation of Aβ42 oligomers and microinjection.
Oligomerization of Aβ42 monomers was performed by dissolving 0.5 mg of lyophilized human recombinant Aβ1–42 peptide (rPeptides) in 20 μl of freshly prepared DMSO and quickly diluting with 480 μl of double-distilled water in a siliconized Eppendorf tube. After 10 min sonication, samples were incubated at room temperature for 10 min and then centrifuged for 15 min at 14,000 × g. The supernatant fraction was transferred to a new siliconized tube and stirred at 500 rpm using a Teflon-coated microstir bar for 8–48 h at room temperature (Demuro et al., 2011). Aggregation states were confirmed by Western blotting. The potency of oligomer preparations to induce macroscopic Ca2+ influx was assayed by local bath application (1 μg/ml) to voltage-clamped oocytes. Aβ42 oligomer preparations used here evoked Ca2+-dependent Cl− currents of >1 μA on polarization to −100 mV. Fibrillar Aβ42 preparations were made as described for Aβ42 oligomers but were stirred for ∼7 d at room temperature and were then spun at 14,000 × g for ∼10 min to pull down large insoluble aggregates that otherwise clogged the injection pipette. Aβ42 monomer with a scrambled peptide sequence was obtained from rPeptide and was treated following the same protocol used for oligomerization of the regular Aβ42.
Microinjection of 10 nl of Aβ42 oligomers (1 μg/ml) into oocytes was performed using a Drummond nanoinjector mounted on an hydraulic micromanipulator. A glass pipette (broken to a tip diameter of 8–10 μm) was filled with injection solution and was inserted vertically down through the entire oocyte to a pre-established position with the tip positioned a few micrometers inward from the plasma membrane and centered within the image field.
Lyophilized Aβ42 monomer (catalog #A-1163-1) and Aβ42 scramble (catalog #A-1004-1) were purchased from rPeptide. Aβ42-HilyteFluor555-labeled (catalog #60480-01) was purchased from AnaSpec, fluo-4 and caged IP3 from Invitrogen, caffeine from Sigma, heparin from Thermo Fisher Scientific, and PTX and IP3 from Tocris Bioscience.
Intracellular injection of Aβ42 oligomers evokes increases in cytosolic Ca2+
We chose to investigate the molecular mechanisms of intracellular toxicity of Aβ42, rather than the Aβ40 form, in light of the reported correlation between the early onset of familial AD and the increased level of Aβ42 in the AD brain (Jarrett and Lansbury, 1993; Kim and Hecht, 2005). Oocytes were injected with Aβ42 oligomers or monomer from a micropipette inserted through the cell so that its tip lay a few micrometers inward from the plasma membrane, centered within the 40 × 40 μm imaging field (Fig. 1A). Because our primary objective was to characterize the liberation of Ca2+ from intracellular stores induced by Aβ42 oligomers, experiments (with the sole exception of those in Fig. 2) were performed in a Ca2+-free bathing solution including 2 mm EGTA to abolish any contribution from extracellular Ca2+ influx.
The black trace in Figure 1B shows mean measurements (from six oocytes) of fluorescence ratio averaged across the image field, recorded after injection of 10 nl of Aβ42 oligomers (1 μg/ml) into oocytes. Calcium signals increased over a few tens of seconds after injection of Aβ42 oligomers and reached a maximal, sustained level (mean ΔF/F0 ∼0.5) after ∼60 s. In marked contrast, we observed little or no response after injection of equivalent amounts (10 nl, 1 μg/ml) of Aβ42 monomer, tested within 1 h after initially dissolving into solution (Fig. 1B, red trace). We also failed to observe any detectable Ca2+ signals in response to injections of a fibrillar preparation of Aβ42 (n = 4 oocytes) or after injections of a preparation made from a scrambled Aβ42 peptide sequence (n = 4 oocytes); both 1 μg/ml at a volume of 10 nl.
Spatiotemporal patterns of Ca2+ signals evoked by intracellular Aβ42 oligomers and IP3
Aβ42 oligomer injections evoked distinct spatiotemporal patterns of Ca2+ signals, as illustrated in Figure 1C–E as line-scan (kymograph) images, derived by measuring fluorescence along a single line within the imaging field and displaying its evolution over time as a pseudocolored representation using the kymograph function in MetaMorph.
Figure 1C shows an example in which the Ca2+ fluorescence signal increased slowly and monotonically at several “hotspots” along the line scan. In contrast, Figure 1D illustrates an oocyte in which injection of Aβ42 oligomers evoked repetitive waves that propagated across the image field at intervals of 6–7 s, and Figure 1E illustrates an example of transient, localized Ca2+ signals that arose sporadically at discrete locations. The records in Figure 1C–E were obtained in different oocytes, but in some cases, we observed a progressive transition in patterns of Ca2+ signals over a few minutes after Aβ42 injection from local transients through repetitive waves, to a sustained elevation.
These patterns of Ca2+ signals closely resemble the respective hierarchy of local Ca2+ puffs, propagating repetitive Ca2+ waves, and sustained global Ca2+ elevations generated in oocytes by IP3-mediated Ca2+ liberation (Callamaras et al., 1998). For comparison, Figure 1F shows representative Ca2+ responses evoked by intracellular injection of 10 nl of 100 pm IP3, displaying transient local signals at several locations, followed by a spreading wave. Responses to IP3 injections typically began with shorter latencies than responses to injection of Aβ42 oligomers and invariably terminated much faster, persisting for a mean of only 10–20 s, whereas Aβ responses were still evident even after 10–15 min.
Dose–response relationship and latencies of Aβ-evoked Ca2+ signals
Unless otherwise noted, we performed experiments injecting a standard amount of Aβ42 oligomers (10 nl of 1 μg/ml), chosen because this intermediate dose was well below saturating the Ca2+ response, evoked sustained responses persisting for many minutes, and facilitated studies of both puffs and global Ca2+ responses. Injections of a lower dose (0.3 μg/ml) evoked no detectable responses in two of three oocytes and puffs at eight sites in one oocyte. At a dose of 1 μg/ml, the same preparation of oligomers evoked puffs in three of four oocytes (mean of 19 sites per imaging field) and gave a global response in the remaining oocyte. In contrast, higher doses evoked global signals, whereas local puffs could not usually be discerned. Figure 1G shows representative records of fluorescence signals evoked by 10 nl injections of Aβ42 oligomers at concentrations of 3, 10, and 30 μg/ml. The amplitudes of the Ca2+ signals increased progressively with increasing dose, and the latencies to onset and peak of the responses shortened. Different from the well-maintained signals evoked by 1 μg/ml oligomers (Fig. 1B), responses to higher doses displayed an increasingly rapid decay, possibly reflecting desensitization of IP3R or depletion of the ER Ca2+ store.
The delayed onset of responses to Aβ compared with those evoked by injections of IP3 cannot be attributed to slow diffusional spread of Aβ from the injection pipette, because injections of oligomers prepared from a fluorescently tagged Aβ peptide showed an almost immediate (<100 ms) and uniform fluorescence distribution throughout the imaging field (n = 3 oocytes examined). The slower development of Aβ-evoked Ca2+ responses may instead reflect the time required for activation of IP3 production and subsequent accumulation of IP3 to levels exceeding the threshold to evoke Ca2+ release, as contrasted with the immediate delivery of IP3 from an injection pipette. Consistent with that interpretation, Ca2+ signals evoked by IP3-mobilizing agonists show dose-dependent latencies that can be as long as several tens of seconds at low concentrations (Miledi and Parker, 1989). However, it remains possible that some additional process introduces an additional delay in the Aβ-evoked responses. Furthermore, the more sustained responses to Aβ imply that the oligomers evoke persistent production of IP3, whereas the local concentration of IP3 dissipates rapidly by metabolism and diffusion after its injection.
Intracellular and extracellular sources of Ca2+
Extracellular application of Aβ42 oligomers to oocytes induces the formation of Ca2+-permeable pores in the plasma membrane, from which we recorded single-channel Ca2+ fluorescence transients (SCCaFTs) (Demuro et al., 2011). We were thus interested to determine whether similar pores form when Aβ42 oligomers are delivered to the cytosolic face of the membrane. For this purpose, we replaced the Ca2+-free Ringer's solution with a solution including 1.8 mm Ca2+ and voltage clamped the oocyte so that the membrane potential could be held at more negative voltages to increase the electrochemical driving force for Ca2+ influx across the plasma membrane. Figure 2A shows simultaneous records from an oocyte clamped at −80 mV in which injection of Aβ42 oligomers evoked localized Ca2+ transients at multiple sites. These signals showed two qualitatively different characteristics. At some sites (Fig. 2A, traces 1, 2), the transients showed an abrupt rise and approximately exponential decay over several tens of milliseconds that, as described in the following section, resembled IP3-evoked Ca2+ puffs. Different from this, other sites in the same imaging field (Fig. 2A, traces 3, 4) showed more prolonged, “square” fluorescence transients, similar to the multilevel SCCaFTs observed after extracellular application of Aβ42 oligomers (Demuro et al., 2011). Consistent with an extracellular Ca2+ source underlying these channel-like signals, their amplitude reduced when the membrane potential was stepped from −80 to −20 mV to reduce the driving force for Ca2+ influx (Fig. 2B,C). Moreover, SCCaFTs were no longer detectable at 0 mV, and they were not observed in Ca2+-free extracellular solution (n = 15 oocytes). In contrast, the amplitudes of puff-like transients were little affected by changes in membrane potential (mean amplitude at −20 mV ΔF/F0 = 0.55 ± 2.3, n = 49 puffs, 4 oocytes; ΔF/F0 = 0.61 ± 1.7, n = 57 puffs, 4 oocytes at −100 mV), and they persisted in Ca2+-free solution (n = 12 oocytes).
Injection of Aβ42 oligomers evokes local Ca2+ transients that mimic IP3-evoked puffs
As noted, the Aβ42-evoked transients that persist in the absence of extracellular Ca2+ resemble the IP3-evoked Ca2+ puffs that we described in oocytes (Yao et al., 1995). To further examine commonality among these signals, we compared responses evoked by injection of Aβ42 oligomers with those evoked directly by IP3, using photorelease from caged IP3 to achieve more controlled and reproducible responses than possible with microinjection of IP3. The top trace in Figure 3A shows representative fluorescence records measured from individual sites showing events evoked by weak photorelease of IP3 by a short ultraviolet light flash, and that in Figure 3D shows a corresponding record of local events evoked after injection of Aβ oligomers. In both panels, the images depict the spatial spread of fluorescence signal at the peak of two selected events (indicated by gray shading in the top fluorescence records), and the bottom traces show the time course of those events on an expanded timescale.
The amplitudes of IP3-evoked puffs followed an approximately Gaussian distribution (Fig. 3B) with a mean of ΔF/F0 = 0.41 ± 0.017 (n = 137 puffs; 6 oocytes). Conversely, Aβ42-evoked transients showed a bimodal amplitude distribution (Fig. 3E), with one population of events closely matching the distribution of IP3-evoked puffs but with an additional “tail” of larger events giving an overall mean amplitude of ΔF/F0 = 0.76 ± 0.05 (n = 297 puffs; 9 oocytes). The distributions of event durations (duration at half-maximal amplitude) of IP3-evoked (Fig. 3C) and Aβ42-evoked (Fig. 3F) events were more closely matched, but a greater number of very short (≤50 ms) events were observed with Aβ. Respective mean durations were 369 ± 16 and 213 ± 8.6 ms. The mean spatial spreads of Ca2+ fluorescence signals, measured at half-maximal amplitude at the time of the peak signal, were 3.0 ± 0.4 μm for IP3 and 2.4 ± 0.2 μm for Aβ (no significant difference at p < 0.05).
Ca2+ signals evoked by intracellular Aβ42 oligomers involve IP3Rs
The similarity between Aβ-evoked signals and IP3-evoked Ca2+ puffs and waves strongly suggested that intracellular Aβ42 oligomers promote the liberation of Ca2+ from ER stores via opening of IP3Rs. To further confirm this hypothesis, we examined the action of known antagonists of IP3R function.
Caffeine acts as a reversible, membrane-permeant competitive antagonist at the IP3R (Parker and Ivorra, 1991; Bezprozvanny et al., 1994) and can be used in the Xenopus oocyte without complications from Ca2+ liberation through ryanodine receptors because these are lacking in the oocyte (Parys et al., 1992). Figure 4A shows control records demonstrating almost complete block of transient global Ca2+ signals evoked by strong photorelease of IP3 during bath application of 10 mm caffeine. Figure 4B shows superimposed traces of puff-like signals evoked at several representative sites by injection of Aβ42 oligomers (top) and at all sites within the same imaging field in which any activity was observed during subsequent bath application of 10 mm caffeine (bottom). Caffeine almost completely abolished activity, as demonstrated by a reduction in puff amplitudes and frequency (Fig. 4B) and by a strong and reversible reduction in number of local sites at which events could be detected [Fig. 4C: mean number of responding sites within the 40 × 40 μm image field 71 ± 13.9 before addition of caffeine, 7 ± 2.8 in the presence of 10 mm caffeine (n = 4 oocytes), and 55 ± 16.5 after washing (n = 3 oocytes)].
Although caffeine inhibits Ca2+ release through IP3R, it also acts on targets including cyclic nucleotide phosphodiesterases and PLC (Toescu et al., 1992; Taylor and Broad, 1998). We therefore also examined the effect of intracellular injection of heparin, which has been shown to inhibit IP3-evoked responses in the oocyte (Yao and Parker, 1993). Figure 5A shows representative records of local Ca2+ fluorescence signals obtained from three different regions 5 min after intracellular injection of Aβ42 oligomers. The micropipette was withdrawn immediately after injecting Aβ42, refilled with heparin, and reinserted into the oocyte. The traces in Figure 5B show corresponding recordings from the same regions 2 min after intracellular injection of 10 nl of heparin (10 μg/ml), revealing a complete cessation of activity. Moreover, we failed to observe events at any other sites in the imaging field, although puff activity would have continued unabated at this time in the absence of heparin. Similar results were obtained in five oocytes (Fig. 5C). The mean number of responding sites within the 40 × 40 μm imaging field before injection of heparin was 29 ± 5 and only 0.6 ± 0.4 after injection. Figure 5D shows a similar suppression of ongoing Ca2+ waves after heparin injection. We further attempted to use xestospongin as a more specific IP3R inhibitor (Gafni et al., 1997) but found this ineffective because, in our hands, extracellular application of 1 μm xestospongin failed to appreciably inhibit control responses to photoreleased IP3.
Preincubation of oocytes with the Gi/o-protein inhibitor PTX prevents generation of Aβ42-evoked calcium signals
Together, the preceding results strongly imply that intracellular Aβ oligomers evoke liberation of Ca2+ through IP3R channels. In principle, this could arise through a direct action of Aβ on the IP3R or indirectly through stimulation of PLC to promote IP3 production. To discriminate between these possibilities, we incubated oocytes with PTX (2 μg/ml) for 24 h before recording to inhibit G-protein (Gi/o)-mediated activation of PLC (Moriarty et al., 1989). Pretreatment with PTX had little effect on Ca2+ signals evoked by strong photorelease of IP3 (Fig. 6A: control ΔF/F0, 1.20 ± 0.14, n = 5 oocytes; PTX ΔF/F0, 1.27 ± 0.04, n = 5 oocytes), whereas responses after injections of Aβ42 oligomers were greatly attenuated (Fig. 6B) and activity was detected at very few sites (Fig. 6C).
Aβ42-evoked calcium signals are suppressed after treatment with lithium to deplete membrane inositol lipids
As an additional means to block production of IP3, we used lithium to block the inositol monophosphatase enzyme involved in recycling and de novo synthesis of inositol and hence deplete the membrane inositol phospholipids from which IP3 is generated (Berridge et al., 1989; Harwood, 2005) Oocytes were incubated overnight in Ca2+-free solution including 100 μm Li+ together with rabbit serum (1:500 dilution) to stimulate IP3 production (Tigyi et al., 1990) and were then washed for 2 h in solution with Li+ but without serum. When tested at this time, Ca2+ signals evoked by strong photorelease of IP3 were of similar amplitude to controls (Li+-treated, mean ΔF/F0, 2.70 ± 0.29, n = 5 oocytes; control ΔF/F0, 3.06 ± 0.27, n = 4), but responses to intracellular injections of Aβ42 oligomers were almost completely abolished (Fig. 6D).
Cytotoxicity of intracellular Aβ
We hypothesized that the sustained liberation of Ca2+ through IP3R induced by Aβ would exert cytotoxic effects. To test this, and to further confirm the mechanism of action of Aβ oligomers, we examined the long-term viability of oocytes after intracellular injections of Aβ under various experimental conditions. Unless otherwise noted, oocytes were bathed in a solution containing no added Ca2+ so as to selectively examine the cytotoxicity of intracellular Ca2+ release in the absence of Ca2+ influx through Aβ plasmalemmal pores. We assessed viability by visual inspection of breakdown of the black pigmentation in the animal hemisphere (Fig. 7A,B), which typically precedes cell death. Figure 7C shows the percentages of oocytes scored as viable at times up to 42 h after intracellular injections of Aβ oligomers, monomer, and a scrambled Aβ peptide sequence. No oocytes injected with Aβ oligomers remained viable at the 37 h time point. In marked contrast, oocytes injected with a scrambled Aβ peptide sequence prepared in the same way as used to generate Aβ oligomers showed little difference in survival from control, vehicle-injected oocytes. Moreover, oocytes injected with an equivalent amount of Aβ monomer showed only a slightly lower viability than control, and this small difference may reflect spontaneous oligomerization before and after injection. The blue triangles in Figure 7C further show data from oocytes incubated in 2 mm Ca2+ solution after injection of Aβ oligomers. Little difference is apparent from corresponding oocytes bathed in Ca2+-free solution, suggesting that influx of extracellular Ca2+ did not appreciably accelerate the onset of toxicity.
The cytotoxicity of Aβ oligomers was almost completely abrogated in oocytes that were pretreated with PTX to block G-protein-mediated production of IP3 (Fig. 7D, green triangles) or with Li+ to block recycling of membrane inositol lipids (Fig. 7D, blue triangles). Importantly, strong protection was also observed in oocytes that were injected with EGTA (final intracellular concentration of ∼1 mm) to chelate cytosolic free Ca2+ (Fig. 7D, red circles). Together with the Ca2+ imaging data, these results demonstrate that intracellular Aβ oligomers induce cytotoxicity as a result of elevated cytosolic Ca2+ levels evoked by the IP3 pathway.
Our results demonstrate that intracellular injection of oligomeric forms of the AD-related peptide Aβ42 into Xenopus oocytes potently elevates cytosolic Ca2+ levels, resulting in cytotoxicity. The Ca2+ arises from at least two sources: (1) entry across the plasma membrane and (2) liberation from intracellular stores. In particular, we show that the intracellular Ca2+ liberation evoked by Aβ42 oligomers involves opening of IP3Rs as a result of stimulated production of IP3 via G-protein-mediated activation of PLC. This process evokes local and global cytosolic Ca2+ signals resembling those evoked by IP3 itself but that are more persistent and result in acute cytotoxicity.
There is already considerable evidence that extracellular Aβ oligomers disrupt the integrity of the plasma membrane to allow Ca2+ influx down its electrochemical gradient (Demuro et al., 2005), and we recorded channel-like Ca2+ signals supporting the notion that Aβ42 oligomers insert into the membrane in which they aggregate to form intrinsic Ca2+-permeable pores (Demuro et al., 2011). The observation that intracellular injections of Aβ42 oligomers evoke qualitatively similar channel-like signals that are dependent on extracellular Ca2+ and vary in size with membrane potential indicates that pore formation can occur bidirectionally, with oligomers inserting into the plasma membrane from either the extracellular or cytosolic faces of the membrane. In light of this result, we had expected that a similar process would occur in the membranes of intracellular organelles, leading to a leak of Ca2+ from ER and other stores through intrinsic Aβ pores. However, the strong suppression of Aβ42-evoked Ca2+ signals in the absence of extracellular Ca2+ by IP3R antagonists and by blocking IP3 production indicates that the major action of Aβ42 to mobilize intracellular Ca2+ is through IP3R. Any additional contribution from intrinsic Aβ pores in intracellular membranes appears to be relatively slight, although in a few instances we could barely resolve small “channel-like” transients in oocytes bathed in Ca2+-free solution after suppressing IP3-mediated signals with caffeine or heparin. Possible factors that may account for the small magnitude of Aβ pore-mediated Ca2+ release from intracellular stores compared with influx across the plasma membrane include the lower electrochemical gradient for Ca2+ ions across the ER membrane relative to the plasma membrane at hyperpolarized voltages, and differences in lipid compositions between internal and plasma membranes (van Meer and de Kroon, 2011) that may affect the binding of Aβ and its aggregation into ion channels (Simakova and Arispe, 2011).
Previous studies have shown that extracellular application of Aβ oligomers can liberate calcium from intracellular stores (Demuro et al., 2005) and have implicated IP3Rs in this process (Ferreiro et al., 2004). However, to our knowledge, our results provide the first demonstration that intracellular Aβ evokes Ca2+ liberation through IP3Rs in an IP3-dependent manner. We show that injections of increasing concentrations of Aβ oligomers evoke progressive patterns of Ca2+ signals from local Ca2+ puffs through repetitive waves to sustained elevations that closely resemble the hierarchy of signals evoked by increasing concentrations of IP3 itself (Callamaras et al., 1998). Moreover, responses to Aβ are blocked by the IP3R antagonists caffeine and heparin, excluding a direct action of Aβ oligomers acting as agonists to promote opening of the IP3R channel. Instead, the suppression of responses to Aβ by pretreatment with PTX to inhibit PLC-mediated IP3 generation and by pretreatment with lithium to deplete the supply of inositol required to maintain the membrane inositol lipids used to generate IP3 strongly implicate the production of IP3 itself in the generation of the Ca2+ signals. Although we cannot entirely exclude the possibility that Aβ oligomers may act on the IP3R to increase its sensitivity to a low, subthreshold basal level of cytosolic IP3 generated by constitutive G-protein-mediated stimulation of PLC, we consider this unlikely. Resting oocytes rarely showed any evidence of ongoing calcium signals, yet Aβ42 oligomer injections were able to evoke potent responses. Moreover, elevations of cytosolic [Ca2+] in otherwise unstimulated oocytes typically fail to evoke intracellular Ca2+ release (Yao and Parker, 1993; Yamasaki-Mann and Parker, 2011), although [Ca2+] elevations enormously increases the open probability of the IP3R channel at low [IP3] (Foskett, 2010). Instead, the most parsimonious explanation is that Aβ42 oligomers stimulate the production of endogenous IP3 by PLC in a G-protein-dependent manner. Moreover, the persistence of the IP3-mediated Ca2+ signals and resulting cytotoxicity imply that intracellular Aβ oligomers remain present and actively stimulate IP3 production for many minutes or hours, because IP3 is itself metabolized within tens of seconds (Sims and Allbritton, 1998).
At present, it is not clear whether Aβ42 oligomers interact only with the G-protein/PLC complex or may also involve endogenous cell-surface receptors that normally regulate this pathway. Xenopus oocytes lack the metabotropic glutamate receptors that are widely found in neurons and only rarely and sparsely express muscarinic receptors (Dascal and Landau, 1980). Conversely, oocytes do express a high density of lipid receptors, which can be activated by lysophosphatidic acid (LPA) and sphingosine-1-phosphate (S1P) to evoke strong IP3-mediated responses (Noh et al., 1998). S1P-induced signaling is mediated via the PTX-sensitive Gq/Gi1 pathway linked to PLC-xβ, whereas LPA-induced signaling uses PTX-insensitive Gq/G11 (Noh et al., 1998). Our finding of almost total inhibition of the Aβ42-induced cytosolic Ca2+ elevation by PTX and protection from cytotoxicity thus suggests that these effects may be mediated by Gq/Gi1.
Since the original formulation of the calcium hypothesis of AD (Khachaturian, 1989, 1994), strong evidence has accumulated that disruption of cellular Ca2+ homeostasis and signaling is a fundamental nexus underlying the neuronal degeneration and disease pathophysiology (Bezprozvanny, 2009; Demuro et al., 2010). The action of intracellular Aβ oligomers to promote IP3-mediated liberation of Ca2+ from ER stores adds an additional mechanism that will not only summate with, but will have a multiplicative effect on, Ca2+ elevations induced by other AD-linked disruptions. Opening of IP3R channels is gated by Ca2+ and IP3 acting as coagonists so that, in the presence of IP3, the IP3R acts as a Ca2+-induced Ca2+ release channel (Yao and Parker, 1993). Hence, entry of extracellular Ca2+ through plasmalemmal pores formed by Aβ42 oligomers would be amplified by triggering intracellular liberation from the ER. Moreover, mutations in presenilin genes associated with familial early-onset AD enhance IP3-mediated Ca2+ liberation (Stutzmann et al., 2004) and are thus expected to potentiate Aβ-evoked Ca2+ liberation through IP3Rs. Finally, depletion of ER Ca2+ stores by Aβ oligomers may turn on store operated Ca2+ entry, leading to a yet additional increase in cytosolic Ca2+ via the STIM/Orai pathway (Cahalan, 2009).
In summary, we show that intracellular Ca 2+ liberation induced by injections of Aβ oligomers into oocytes results in acute cytotoxic effects within 1 d. During the progression of AD, it is likely that much lower concentrations of Aβ would initially cause subtle disruptions in Ca2+ signaling processes regulating synaptic transmission and neuronal excitability (Bezprozvanny and Mattson, 2008; Demuro et al., 2010) that may later progress to overt neurotoxicity as intracellular levels of Aβ oligomers gradually increase to produce both a sustained elevation of cytosolic Ca2+ (Berridge, 1998) and depletion of ER Ca2+ stores (Verkhratsky, 2005).
This work was supported by National Institutes of Health Grants R37-GM48071 (I.P.) and P50-AG16573 (A.D.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Angelo Demuro, 2205 McGaugh Hall, University of California, Irvine, Irvine, CA 92697-4550.