Misfolded protein oligomers induce an increase of intracellular Ca²⁺ causing an escalation of reactive oxidative species

Abstract

Alzheimer's disease is characterized by the accumulation in the brain of the amyloid β (Aβ) peptide in the form of senile plaques. According to the amyloid hypothesis, the aggregation process of Aβ also generates smaller soluble misfolded oligomers that contribute to disease progression. One of the mechanisms of Aβ oligomer cytotoxicity is the aberrant interaction of these species with the phospholipid bilayer of cell membranes, with a consequent increase in cytosolic Ca²⁺ levels, flowing from the extracellular space, and production of reactive oxygen species (ROS). Here we investigated the relationship between the increase in Ca²⁺ and ROS levels immediately after the exposure to misfolded protein oligomers, asking whether they are simultaneous or instead one precedes the other. Using Aβ₄₂-derived diffusible ligands (ADDLs) and type A HypF-N model oligomers (OAs), we followed the kinetics of ROS production and Ca²⁺ influx in human neuroblastoma SH-SY5Y cells and rat primary cortical neurons in a variety of conditions. In all cases we found a faster increase of intracellular Ca²⁺ than ROS levels, and a lag phase in the latter process. A Ca²⁺-deprived cell medium prevented the increase of intracellular Ca²⁺ ions and abolished ROS production. By contrast, treatment with antioxidant agents prevented ROS formation, did not prevent the initial Ca²⁺ flux, but allowed the cells to react to the initial calcium dyshomeostasis, restoring later the normal levels of the ions. These results reveal a mechanism in which the entry of Ca²⁺ causes the production of ROS in cells challenged by aberrant protein oligomers.

Introduction

Alzheimer’s disease (AD), which is the most common neurodegenerative disease and the most common form of dementia, is characterized by the extracellular deposition in the brain of the amyloid β (Aβ) peptide in the form of senile plaques [1] and by the intraneuronal formation of neurofibrillary tangles of the tau protein [2]. The aggregation process of Aβ generates a large variety of protein aggregates, such as oligomers, protofibrils and fibrils, all characterised by high levels of polymorphism [3]. According to the amyloid hypothesis, the small diffusible oligomers of Aβ are neurotoxic and are thought to contribute to AD development and progression [3,4,5]. Oligomer cytotoxicity appears to result, in its early phases, from the aberrant interactions of such species with a number of molecular targets on neurons, including the lipid bilayer of their cell membranes [1, 3, 5, 6]. This interaction results in the disruption of cell membranes, compromising its ability to maintain cellular homeostasis, and promoting two important early biochemical changes. The first is the uncontrolled increase in cytosolic calcium (Ca²⁺) levels flowing from the extracellular space into the cytosol [7,8,9,10,11,12,13,14], and the second is the accumulation of reactive oxygen species (ROS) [10, 11, 15, 16].

It is known that Aβ oligomers are able to interact and insert into the phospholipid bilayer of the cell membrane causing the passage through it of small molecules and ions, such as free Ca²⁺ ions [6, 7, 11, 17, 18], as well as permitting the activation of ionotropic glutamate receptors functioning as Ca²⁺ channels, including the N-methyl-d-aspartate (NMDA) receptors [9, 14, 15, 18,19,20,21] and the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors [14, 15, 18, 19, 22]. In particular, the rapid oligomer-induced activation of extrasynaptic NMDA/AMPA receptors is a crucial mechanism in the AD pathogenesis. This process takes place through the insertion of the oligomers in the bilayer, which changes the mechanical properties of the membrane that is transmitted down to the receptors that are therefore activated through their mechanosensitivity, without a direct interaction with the oligomers [14]. Other Ca²⁺ channels that seem to be involved in the Aβ-induced flux of Ca²⁺ ions are the transient receptor potential melastatin 2 (TRPM2) [Aβ₄₂ ADDLs oligomers increase intracellular Ca²⁺ levels and ROS production

We then extended the analysis carried out with the model HypF-N OAs to Aβ oligomers, using Aβ₄₂-derived diffusible ligands (Aβ₄₂ ADDLs) [39] at the concentration of 1 μM. In previous works it was shown that Aβ₄₂ ADDLs, similarly to HypF-N OAs, are able to cause a progressive increase of the intracellular Ca²⁺ levels in SH-SY5Y cells by activating rapidly extrasynaptic NMDA and AMPA receptors [14]. We therefore prepared freshly formed Aβ₄₂ ADDLs oligomers and evaluated whether they maintained this effect. The treatment over time of SH-SY5Y cells with Aβ₄₂ ADDLs showed a gradual increase of the intracellular Ca²⁺ levels, which was clearly detectable already after 5 min and reached a plateau after 180 min of treatment (images in Fig. 5a, histograms in Fig. 5b and corresponding kinetic plot in Fig. 5c). When cells were pre-treated with CNQX, with memantine, or with both CNQX and memantine, a slight reduction of the Aβ₄₂ ADDLs-induced cytoplasmic Ca²⁺ increase was observed in the early stages, up to 10 min of treatment (Fig. 5a,b). With Aβ₄₂ ADDLs, a combination of both inhibitors showed kinetics similar to the memantine treatment. This reduction was followed by a gradual increase of the intracellular Ca²⁺ concentration, until normal levels were reached after prolonged treatment (Fig. 5a,b). Overall, these pre-treatments cause a deceleration of the intracellular Ca²⁺ increase at early time points (Fig. 5c).

Fig. 5

Aβ₄₂ ADDLs oligomers increase intracellular Ca²⁺ levels and ROS production in SH-SY5Y cells. a Representative confocal scanning microscopy images of free Ca²⁺ levels in SH-SY5Y cells following the treatment with no inhibitors (first row), 5 µM CNQX (second row), 10 µM memantine (third row), and both inhibitors (fourth row), and analysed after 5, 10, 15, 30, 60, 90, 120 and 180 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. b Semi-quantitative analysis of intracellular free Ca²⁺-derived fluorescence. The value for untreated cells refers to 0 min and did not change with time. c Kinetic plots showing the fluorescence versus time as reported in panel b. d Representative confocal scanning microscopy images of intracellular ROS levels in SH-SY5Y cells following the treatment with no inhibitors (first row), 5 µM CNQX (second row), 10 µM memantine (third row), and both inhibitors (fourth row), and analysed after 5, 10, 15, 30, 60, 90, 120 and 180 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. e Semi-quantitative analysis of intracellular ROS-derived fluorescence. The value for untreated cells refers to 15 min and did not change with time. f Kinetic plots showing the fluorescence versus time as reported in panel e. Three different experiments were carried out, with 10–22 cells each, for each condition. Data are represented as mean ± SEM (n = 3). The single (*), double (**) and triple (***) asterisks refer to p values < 0.05, < 0.01 and < 0.001, respectively, relative to untreated cells. The single (§), double (§§) and triple (§§§) symbols refer to p values < 0.05, < 0.01 and 0.001, respectively, relative to Aβ₄₂ ADDLs oligomers without inhibitors at corresponding time points

The treatment of SH-SY5Y cells with Aβ₄₂ ADDLs under the same conditions also showed a gradual increase of ROS production, which was evident after 15 min up to 180 min, and hence slower than that observed by monitoring Ca²⁺ concentration (Fig. 5d–f). Interestingly, such increase appeared to occur more rapidly than that observed with HypF-N OAs, which can be attributed to the known oxidative potential of Aβ₄₂ ADDLs through Ca²⁺-independent mechanisms [44,45,46]. Cellular pre-treatment with CNQX or memantine, or both inhibitors, determined again a reduction of ROS levels in the early stages, up to 30 min for CNQX and 60 min for memantine and both inhibitors together, followed by a gradual increase, until normal levels were reached after 90 min (Fig. 5d–f). These pre-treatments, therefore, caused a deceleration of the ROS increase mediated by the oligomers (Fig. 5f), which was again more marked than that detected by monitoring intracellular Ca²⁺ levels. All these results confirmed the observation with the HypF-N OAs.

Comparing the Ca²⁺ and ROS kinetics without NMDA/AMPA inhibitors, the ROS time course appears to be slower in the first minutes (Fig. 6a), suggesting that the increase of the intracellular Ca²⁺ levels anticipates ROS production. Moreover, comparing the times courses in the presence of CNQX or memantine, or both, the ROS time courses appear again to be slower than the corresponding time courses of Ca²⁺ (Fig. 6a–c). Also with this type of oligomers we observed a longer delay in ROS level increase following pre-treatment with CNQX (Fig. 6a, orange dotted line) or memantine (Fig. 6b, blue dotted line) or both (Fig. 6c, yellow dotted line), compared to the Ca²⁺ kinetics following the same pre-treatment (Fig. 6a,b, orange, blue and yellow line, respectively), confirming that the reduction of the early Ca²⁺ influx, observed by inhibiting the NMDA and AMPA receptors, allowed the cells to delay the production of ROS.

Fig. 6

Increase of intracellular Ca²⁺ levels anticipates ROS production in SH-SY5Y cells. a–c Kinetic plots showing the fluorescence associated with intracellular Ca²⁺ and ROS versus time after treatment with Aβ₄₂ ADDLs. The time courses refer to Ca.²⁺ levels (solid lines) and ROS levels (dotted line) without inhibitors (grey), with CNQX (orange), with memantine (blue) and with both CNQX and memantine (yellow)

When repeated on primary rat cortical neurons, the Aβ₄₂ ADDLs had a similar effect. After 10 min of treatment, the Aβ₄₂ ADDLs induced an increase of the intracellular Ca²⁺ levels, which further increased after 60 min of treatment (Fig. 7a,b). When the cells were pre-treated with CNQX or memantine, a significant reduction of the Aβ₄₂ ADDLs-induced cytoplasmic Ca²⁺ levels was observed after 10 min of treatment with the oligomers, confirming the involvement of the receptors in the Ca²⁺ influx (Fig. 7a,b). After 60 min of treatment with ADDLs, the levels of Ca²⁺ in the presence of pre-treatment went back to the levels observed in its absence (Fig. 7a,b). Moreover, Aβ₄₂ ADDLs also induced an increase of ROS levels after 10 min and a further increase after 60 min of treatment (Fig. 7c,d), with the former being significantly reduced with CNQX or memantine (Fig. 7c,d).

Fig. 7

Aβ₄₂ ADDLs oligomers increase intracellular Ca²⁺ levels and ROS production in primary rat cortical neurons. a Representative confocal scanning microscopy images of intracellular free Ca²⁺ levels in primary rat cortical neurons treated with no inhibitors (first row), 5 µM CNQX (second row) and 10 µM memantine (third row), and analysed after 10 and 60 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. b Semi-quantitative analysis of intracellular free Ca²⁺-derived fluorescence. c Representative confocal scanning microscopy images of intracellular ROS levels in primary rat cortical neurons treated with no inhibitors (first row), 5 µM CNQX (second row) and 10 µM memantine (third row), and analysed after 10 and 60 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. d Semi-quantitative analysis of intracellular ROS-derived fluorescence. Three different experiments were carried out, with 10–22 cells each, for each condition. Data are represented as mean ± SEM (n = 3). The single (*) and double (**) asterisks refer to p values < 0.05 and < 0.01, respectively, relative to untreated cells. The single (§) and double (§§) symbols refer to p values < 0.05 and < 0.01, respectively, relative to Aβ₄₂ ADDLs oligomers without inhibitors at corresponding time points

Intracellular Ca²⁺ influx and ROS production induced by Aβ₄₂ ADDLs are connected

We then treated SH-SY5Y cells with Aβ₄₂ ADDLs in the presence and absence of a pre-treatment for 1 h with the antioxidant Trolox. In the presence of Trolox, an initial increase of cytosolic Ca²⁺ concentration was observed, particularly after 10–30 min of treatment with the oligomers, followed by a reduction at180 min (Fig. 8a,b). These results confirm that the maintenance of a redox balance allowed the cells to react to the initial Ca²⁺ flux induced by the Aβ₄₂ ADDLs and normalize Ca²⁺ homeostasis, initially lost because of the action of the oligomers.

Fig. 8

Intracellular Ca²⁺ influx and ROS production induced by Aβ₄₂ ADDLs are connected in SH-SY5Y cells. a Representative confocal scanning microscopy images of intracellular free Ca²⁺ levels in SH-SY5Y cells following no treatment (first row), and pre-treatment with 30 µM Trolox (second row), and analysed after 5, 10, 15, 30, 60, 90, 120 and 180 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. b Semi-quantitative analysis of intracellular Ca²⁺-derived fluorescence. The value for untreated cells refers to 0 min and did not change with time. c Representative confocal scanning microscopy images of intracellular ROS levels in SH-SY5Y cells following no treatment (first row), and treatment in a medium without Ca²⁺ (second row), and analysed after 5, 10, 15, 30, 60, 90, 120 and 180 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. d Semi-quantitative analysis of intracellular ROS-derived fluorescence. The value for untreated cells refers to 15 min and did not change with time. Three different experiments were carried out, with 10–22 cells each, for each condition. Data are represented as mean ± SEM (n = 3). The double (**) and triple (***) asterisks refer to p values < 0.01 and < 0.001, respectively, relative to untreated cells. The single (§), double (§§) and triple (§§§) symbols refer to p values < 0.05, < 0.01 and < 0.001, respectively, relative to Aβ₄₂ ADDLs oligomers without treatment with Trolox or Ca²⁺-deprived medium at corresponding time points

With the same purpose, ROS production in SH-SY5Y cells was evaluated after treatment with Aβ₄₂ ADDLs over time, with or without Ca²⁺ in the cell medium. The absence of extracellular Ca²⁺ determined levels of ROS similar to those observed in untreated cells up to 180 min of treatment with the oligomers, without any initial increase at early time points (Fig. 8c,d), indicating that the cells without any Ca²⁺ influx and dyshomeostasis did not undergo any oxidative stress, despite the treatment with toxic Aβ₄₂ ADDLs in the absence of antioxidants (Fig. 8c,d). These results emphasise that while the suppression of the Ca²⁺ influx in the cells suppresses the oxidative stress for the entire length of time of the analysis, the cellular protection by a reducing agent does not suppress the initial oligomer-induced Ca²⁺ influx.

To confirm these results with different probes of intracellular Ca²⁺ and ROS, we repeated the experiments with ADDLs after 10 and 60 min of treatment, with or without Trolox and with or without Ca²⁺ in the cell medium, using the X-Rhod-1 AM and the CellRoxTM Deep Red Reagent to monitor Ca²⁺ and ROS levels, respectively. The results confirm that the presence of the antioxidant allowed the cells to react to and normalise the initial Ca²⁺ influx observed after 10 min of treatment, which reached the levels of untreated cells after 60 min of treatment (Fig. S3a,b), and that ROS levels remained constant and similar to those of untreated cells when the treatment was performed in a medium without Ca²⁺ (Fig. S3c,d).

The effect of Trolox was also tested on primary rat cortical neurons. The cells were treated with Aβ₄₂ ADDLs for 10 or 60 min, with or without the 1 h pre-treatment with Trolox. Similarly to SH-SY5Y cells, the presence of Trolox did not prevent a slight increase of the intracellular Ca²⁺ concentration, but caused lower levels of Ca²⁺ after both 10 and 60 min of treatment with Aβ₄₂ ADDLs relative to cells pre-treated with Trolox (Fig. 9). This suggests that also in this cellular system the oxidative stress reduction allows the cells to counteract the initial Ca²⁺ influx across the membrane and restore the normal levels of Ca²⁺.

Fig. 9

Intracellular Ca²⁺ influx and ROS production induced by Aβ₄₂ ADDLs are connected in primary rat cortical neurons. a Representative confocal scanning microscopy images of intracellular free Ca²⁺ levels in primary rat cortical neurons with no treatment (first row), and pre-treatment with 30 µM Trolox (second row), and analysed after 10 and 60 min of treatment with 1 µM (monomer equivalents) Aβ₄₂ ADDLs oligomers. b Semi-quantitative analysis of intracellular free Ca²⁺-derived fluorescence. Three different experiments were carried out, with 10–22 cells each, for each condition. Data are represented as mean ± SEM (n = 3). The single (*) and double (**) asterisks refer to p values < 0.05 and < 0.01, respectively, relative to untreated cells. The single (§) symbol refers to p values < 0.05 relative to Aβ₄₂ ADDLs oligomers without pre-treatment with Trolox at corresponding time points

Discussion

Dysregulation of Ca²⁺ signalling and excessive production of intracellular ROS are common early features of neurodegenerative disorders, in particular AD [13, 15, 47, 48]. Several studies have shown that the passage from the extracellular space into the cytosol of small molecules and ions, such as Ca²⁺ ions, is mediated by the interaction of Aβ oligomers, characteristic of AD, with the lipid bilayer [7, 11, 14, 17, 18]. Our results confirm these observations. They also confirm that this early modification is associated with the increase of cytosolic ROS levels. Interestingly, the increase of ROS production appears to occur more rapidly following the treatment with Aβ₄₂ ADDLs than HypF-N OAs, which can be attributed to the known oxidative potential of Aβ₄₂ ADDLs through Ca²⁺-independent mechanisms [44,45,46].

The maintenance of the Ca²⁺ gradient across the cell membrane, where the Ca²⁺ concentration is 50–100 nM inside the cell and 1.1 mM outside, represents a great energetic expense, because the plasma membrane Ca²⁺-ATPase (PMCA) and the sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) need ATP to pump out the ions from the cytosol and restore homeostasis [6, 28, 36, 37]. Therefore, the increased need for ATP caused by the oligomer-induced Ca²⁺ dyshomeostasis activates the Krebs cycle, electron transport chain and oxidative phosphorylation in mitochondria, which determines the mitochondrial generation of ROS through the increased O₂ reduction [28, 37]. ROS can also be produced by extramitochondrial enzymes, such as NADPH oxidase, xanthine oxidase, cytochrome P450, myeloperoxidase, cyclooxygenase, lipoxygenase and uncoupled nitric oxide synthase, all of which are modulated by Ca²⁺ [37]. This explains the association between Ca²⁺ dysregulation and increased ROS production.

The kinetic results presented here show that the delay in ROS production, which is evident as a lag phase and slower overall process in both time courses of ROS production following Aβ₄₂ ADDLs and HypF-N OAs addition, is suggestive, albeit not a demonstration per se, that the ROS increase follows, and is caused by, that in Ca²⁺. To address further the cause-and-effect relationship between these two events, we took into consideration the data obtained with inhibitors of the Ca²⁺ influx and the known relationship between the two processes. Indeed, the extracellular-to-cytosol influx of Ca²⁺ induced by misfolded protein oligomers arises, at least in its early stages, from the passage of the ions through the AMPA and NMDA receptors [9, 15, 18, 19], which are mechanically activated following the modification of the phospholipid bilayer induced by the oligomers [14]. The pharmacological inhibition of the two glutamatergic receptors, with CNQX and memantine, respectively, delayed transiently the Ca²⁺ influx induced by these oligomers, with no significant increase within the first minutes of treatment. The delay mediated by CNQX and memantine, however, did not only involve the Ca²⁺ influx, but also ROS production. We also observed a delay in ROS levels increase following the pre-treatment with the inhibitors and this delay was even larger than that observed for Ca²⁺ levels. These results suggest that the inhibition of AMPA and NMDA receptors, with the consequent reduction of the early Ca²⁺ influx, allowed the cells to postpone ROS production. At later time points, intracellular Ca²⁺ levels increase despite the persistent inactivation of the two receptors, reaching the same levels observed in the absence of any inhibition, because it is caused by the direct passage of the ions through the cell membrane after the interaction of the oligomers with the lipid bilayer and a consequent destabilization and perforation [17, 18]. In addition, the Ca²⁺ pumps are inhibited by ROS, contributing to increase Ca²⁺ levels at later time points (see below). In the same way, ROS production increases, while continuing to maintain this slight delay because of the inactivation of the AMPA/NMDA receptors. The kinetic data, in particular, indicate that the use of either CNQX or memantine, or both, results in a lag time of the increase in Ca²⁺ levels, followed by the extension of the lag phase in ROS production. Other Ca²⁺ channels are probably involved in the oligomer-mediated Ca²⁺ influx, such as TRPM2 [54]. Indeed, upon treatment with the antioxidant agent Trolox in our experiments, which completely inhibits the increase of ROS levels and prevents its damaging effects, it is likely that the cells are able to restore Ca²⁺ homeostasis effectively, as a result of the lack of ROS-mediated oxidation of the PMCA and SERCA, among other cellular factors.

The selective oxidation and inactivation of the Ca²⁺ regulatory proteins mediated by ROS may represent an adaptive response to the oxidative stress, because it down-regulates ATP production through the mitochondrial electron transport chain and the inevitable generation of ROS associated with it [28].

Further evidence of the importance of Ca²⁺ influx in ROS production occurs in the treatment with the oligomers in a medium without Ca²⁺ (to inhibit Ca²⁺ influx) and with an antioxidant agent (to inhibit ROS production). The absence of Ca²⁺ in the extracellular medium fully inhibits the increase of the intracellular Ca²⁺ levels that normally flow from the extracellular space, but at the same time fully inhibits ROS production, confirming that ROS result from the need to restore Ca²⁺ homeostasis. By contrast, treatment with the antioxidant agent Trolox leads the restoration of Ca²⁺ homeostasis, but only at prolonged time points. This latter analysis showed that Ca²⁺ ions enter the cells in the first minutes, because the antioxidant agent inhibits only ROS production and is not able to inhibit the oligomer-mediated activation of AMPA and NMDA receptors that occurs within the first minutes of interaction of the oligomers with the cell membrane. This rapid increase of intracellular Ca²⁺ is followed by a decrease, suggesting that the cells are able to pump out Ca²⁺ and restore homeostasis as they benefit from an effective antioxidant capacity induced by Trolox and absence of any direct ROS-induced inhibition of the PMCA, SERCA, other pumps and possibly other cellular factors involved in these processes.

Conclusions

Vicious cycles, or positive feedback loops, exist between Ca²⁺ signalling and ROS production [28, 37], and even between Aβ production and Ca²⁺ signalling [55] and between Aβ production and ROS production [55], where the various events sustain each other. However, a precise cause-and-effect relationship between increased levels of intracellular Ca²⁺ and cytosolic ROS production at the very early stages of the overall dysregulation induced by misfolded protein oligomers emerges from our results by three distinct lines of evidence, namely: (i) a lag time observed in the time course of oligomer-induced ROS production (but not in Ca²⁺ increase), (ii) an ability of AMPA/NMDA receptor inhibitors to retard ROS production even more effectively than Ca²⁺ influx and (iii) an inability of antioxidant agents to inhibit the early Ca²⁺ influx, while fully maintaining the redox status of the cells, whereas a Ca²⁺ deprived medium inhibits fully and effectively both Ca²⁺ influx and ROS production. Hence, the oligomers cause the entry of Ca²⁺ ions in the cells, determining the formation of ROS due to the increased demand of ROS-generating ATP production by mitochondria; ROS in turn prevent the cells from pum** back Ca²⁺ ions into the extracellular space and from restoring the normal Ca²⁺ homeostasis, indicating a positive feedback on Ca²⁺ dyshomeostasis on the longer time scale.