Metal Ion-dependent Effects of Clioquinol on the Fibril Growth of an Amyloid β Peptide*

Although metal ions such as Cu2+, Zn2+, and Fe3+ are implicated to play a key role in Alzheimer disease, their role is rather complex, and comprehensive understanding is not yet obtained. We show that Cu2+ and Zn2+ but not Fe3+ renders the amyloid β peptide, Aβ1–40, nonfibrillogenic in nature. However, preformed fibrils of Aβ1–40 were stable when treated with these metal ions. Consequently, fibril growth of Aβ1–40 could be switched on/off by switching the molecule between its apo- and holo-forms. Clioquinol, a potential drug for Alzheimer disease, induced resumption of the Cu2+-suppressed but not the Zn2+-suppressed fibril growth of Aβ1–40. The observed synergistic effect of clioquinol and Zn2+ suggests that Zn2+-clioquinol complex effectively retards fibril growth. Thus, clioquinol has dual effects; although it disaggregates the metal ion-induced aggregates of Aβ1–40 through metal chelation, it further retards the fibril growth along with Zn2+. These results indicate the mechanism of metal ions in suppressing Aβ amyloid formation, as well as providing information toward the use of metal ion chelators, particularly clioquinol, as potential drugs for Alzheimer disease.

Alzheimer disease (AD) 1 is a neurodegenerative disease characterized by cerebral deposits of extracellular amyloid plaques, intracellular tangles, and intravascular or extravascular deposits (1,2). The deposits are comprised of a mixture of 39 -43-amino-acid polypeptides generally designated as amyloid-␤ (A␤) peptide (3,4). A␤ peptides are secreted by proteolytic processing of several membrane proteins, which are alternative splicing products of the gene coding the amyloid precursor protein (APP) (5)(6)(7). Accumulation of the A␤ peptides as insoluble amyloid deposits occurs in AD due to yet unclear mechanism(s).
The role of metal ions, notably Cu 2ϩ , Zn 2ϩ , and Fe 3ϩ , in AD and metal ion chelators as therapeutic agents has been the subject of interest over recent years (8 -10). Supplementing Cu 2ϩ in drinking water leads to development of the disease in rabbit model (11). High concentrations of Cu 2ϩ (0.4 mM) and Zn 2ϩ (1 mM) have been reported in the senile plaques of the diseased brain, extracellular Cu 2ϩ concentration reaching to 15 M (12). In contrast, Cu 2ϩ has been reported to decrease A␤ deposits in APP23 transgenic mice (13). Moreover, overexpression of human A␤ peptides in transgenic mice leads to a decrease in brain Cu 2ϩ (14). APP knock-out mice show an increase in Cu 2ϩ (15). Thus, the role of metal ions is rather complicated; it is not clear whether the accumulation of metal ions is the cause or the consequence of AD or a biological defense response to decrease progression of the disease. The A␤ peptides bind Cu 2ϩ and Zn 2ϩ with high affinity (16 -21). Although histidine residues, especially His-13, are involved in binding Cu 2ϩ and Zn 2ϩ (20 -22), Cu 2ϩ and Zn 2ϩ differ in their coordination to His-13 (22). Cu 2ϩ and Zn 2ϩ trigger aggregation and precipitation of A␤ peptides (19,(23)(24)(25)(26)(27)(28), which are distinct from the amyloid fibrils, especially in the ␤-sheet content (27,28). Toxicity of granular aggregates is a target of recent extensive studies (29,30). In this respect, the use of "metal-protein attenuation compounds" is proposed as a potential strategy for treating AD (31,32). The antibiotic, clioquinol (CQ), which is also known to be a transition metal ion chelator (33,34), has been found to dissolve deposits of the AD brain tissue (8) and to yield promising results in an animal model for AD in decreasing the number of deposits (8,36). CQ has been the focus of attention as a potential drug (8). However, its effect on the amyloid fibril formation, especially in conjunction with Cu 2ϩ or Zn 2ϩ , has not been understood so far. Therefore, it is important to obtain comprehensive understanding of the interactions with metal ions and their impact on the aggregation and amyloid fibril formation of A␤ peptides.
In the present study, we investigated the role of some metal ions on the aggregation and fibrillogenic propensities of A␤  . The results show that Cu 2ϩ and Zn 2ϩ , but not Fe 3ϩ , render A␤ 1-40 less fibrillogenic or non-fibrillogenic under conditions in which no significant metal ion-induced aggregation occurs. We used this system to examine the effects of metal ion chelators, EDTA and CQ, on the fibril growth of A␤  . The results revealed the dual effects of CQ, explaining the underlying mechanism of clioquinol as a potential drug for Alzheimer disease.

EXPERIMENTAL PROCEDURES
Materials-Human A␤ 1-40 was purchased from the Peptide Institute, Inc. (Osaka, Japan). CuCl 2 dihydrate, anhydrous ZnCl 2 , and anhydrous FeCl 3 were purchased in their purest forms from Aldrich. CQ (5-chloro-7-iodo-8-hydroxyquinoline, commonly known as clioquinol) was procured from Calbiochem. "Protein assay Coomassie Brilliant Blue solution" was purchased from Nacalai Tesque Inc. (Kyoto, Japan).  -50 M A␤ 1-40 in the indicated buffers in the presence or absence of indicated substances was incubated at 37°C for 15 min. The final concentrations of Cu 2ϩ or Zn 2ϩ in the samples were 50 M. Stock CQ solution was prepared in methanol, and it was added to a final concentration of 100 M. The samples were centrifuged at 13,000 ϫ g for 15 min. Peptide concentration in the supernatant was determined by the Bradford method using the commercial protein assay Coomassie Brilliant Blue solution and was represented as a percentage of recovery relative to the control without metal ion and CQ. We did not measure the trace contamination of the metal ions in our buffer system or from the peptide preparations. Therefore, the mentioned concentrations of the metal ions represent the added concentrations of the metal ions. Unless otherwise mentioned, the final concentration of methanol in our experiment due to the addition of CQ did not exceed 2%.

Effects of CQ and Buffer Conditions on Metal Ion-induced Aggregation of A␤
Atomic Force Microscopy-Samples of 50 M A␤ 1-40 in 20 mM HEPES buffer, pH 7.4, containing 100 mM NaCl and 50 M metal ions were incubated for 5 min. An aliquot of the sample was placed on a mica surface and incubated for 1 min. Excess sample was drained and washed quickly with water to remove salts. The mica surface was air-dried. Atomic force microscopy imaging was performed using a dynamic force microscope, SPI3700-SPA300, Seiko Instruments. The scanning tip was a silicon microcantilever (SI-DF20, Seiko Instruments, spring constant ϭ 13 newton/m, resonance frequency ϭ 135 kHz). The scan rate was 0.5 Hz.
Amyloid Fibril Growth of A␤ 1-40 -Samples of 50 M (ϳ0.22 mg/ml) A␤  in 50 mM sodium phosphate buffer, pH 7.4, containing 100 mM NaCl (referred to as buffer A henceforth) in the absence and presence of required concentrations of metal ions were incubated at 37°C with 5 g/ml sonicated fibrils (fibril seed). Aliquots (5 l) of the samples were withdrawn at different time points and added to 1 ml of 5 M thioflavin T in 50 mM glycine-NaOH buffer, pH 8.5 (ThT solution). Fluorescence intensity at 485 nm, which is proportional to the extent of fibril-bound ThT (37) and thus the amyloid fibril growth, was measured with an excitation wavelength of 445 nm using a Hitachi F-4500 fluorescence spectrophotometer. To test the effect of EDTA and CQ, the required concentrations of the chelators were added at the indicated time points.
Total Internal Reflection Fluorescence Microscopy (TIRFM)-ThT was added to the samples in buffer A to a final concentration of 5 M, and the sample was placed on a glass slide. The TIRFM system to observe ThT-bound amyloid fibrils was developed based on an inverted microscope (IX70; Olympus, Tokyo, Japan) as described earlier (38,39).
Circular Dichroism-Samples of 0.22 mg/ml A␤ 1-40 peptide alone in buffer A and the samples of the peptide and fibril seeds (5 g/ml) incubated in the absence or presence of required concentrations of metal ions at 37°C for 80 min were prepared. These samples were diluted 1:1 (v/v) with buffer A, and the far UV CD spectra were recorded using a Jasco-600 spectropolarimeter at 37°C with thermo-stated cell holder and a 0.1-cm path length quartz cuvette. The data shown are the mean residue ellipticity.
Sedimentation Velocity Measurements-Sedimentation velocity measurements were performed using an Optima XL-I analytical ultracentrifuge (Beckman Coulter) with An-60 rotor and two-channel charcoal-filled Epon cells. Samples (0.3 ml) of 30 M A␤ 1-40 in buffer A in the absence and in the presence of 50 M metal ions or 100 M CQ were incubated at 37°C for 15 min followed by centrifugation at the same temperature at 224,000 ϫ g. The protein boundary was scanned at 12-min intervals for its absorbance at 225 nm. The boundary curves at every 24-min intervals are shown. The sedimentation data were analyzed using the software Origin 4.1 (OriginLab).

Effects of CQ and Buffer Conditions on Metal Ion-induced
Aggregation of A␤ 1-40 -In HEPES buffer containing 100 mM NaCl (pH 6.6 or 7.4), both 50 M Zn 2ϩ and 50 M Cu 2ϩ largely decreased the percentage of recovery of A␤   (Fig. 1, A and  B). The Cu 2ϩ -induced decrease in the percentage of recovery was more significant at pH 6.6 than at pH 7.4, indicating the pH dependence of the Cu 2ϩ effects. Although 100 M CQ did not itself exhibit notable effect on the aggregation of A␤ 1-40 , it suppressed the metal ion-induced aggregation of the peptide, increasing its recovery (Fig. 1, A and B). Thus, in HEPES buffer, Cu 2ϩ and Zn 2ϩ induce aggregation of A␤ 1-40 , in agreement with the earlier studies (19, 23-28), which can be sup-pressed by CQ. The atomic force microscopy images (Fig. 1D) indicated that metal ion-induced aggregates are amorphous in nature, supporting earlier reports (27,28).
Importantly, metal ion-induced aggregation of A␤ 1-40 depended on buffer conditions. In buffer A, we did not observe metal ion-induced loss of A␤   (Fig. 1C). Furthermore, sedimentation velocity measurements showed that the metal iontreated or untreated A␤ 1-40 do not exhibit significant change in their sedimentation property (Fig. 2). The apparent sedimentation coefficients of A␤ 1-40 in buffer A alone and in the presence of 50 M Zn 2ϩ and 50 M Cu 2ϩ were calculated to be 0.96, 0.86, and 1.05 s, respectively. Thus, our result shows that the metal ions do not promote significant aggregation of A␤ 1-40 in buffer A.
One possible reason for the observation made in buffer A could be that the formation of the phosphate complexes of these metal ions prevents their interaction with the A␤ peptide. However, our following results on the effects of Cu 2ϩ and Zn 2ϩ on the fibril growth and ␤-sheet formation of the peptide indicate that the peptide interacts with these metal ions in buffer A. Moreover, Cu 2ϩ binding to the peptide in buffer A is further supported by the fluorescence quenching studies presented later on in this study. Thus, taking together these evidences, we presume that the observed lack of metal ion-induced aggregation in buffer A is not due to the lack of metal ion binding. Rather, it represents the conditions in which the metal ionbound peptide exhibits less or no aggregation. In this context, it is important to note that two different laboratories have earlier made contradictory observations on the susceptibility of A␤ peptide toward the Zn 2ϩ -induced aggregation (23,25). Huang et al. (26) clarified that the susceptibility of the A␤ peptide to the Zn 2ϩ -induced aggregation is sensitive to "complex factors in buffer milieu that impact upon the peptide's conformation and polymerization state." Therefore, interestingly, our observation that metal ion-induced aggregation is less pronounced in buffer A is useful to understand the relative fibrillogenic propensities of the metal-free (apo) and metalbound (holo) A␤  .
Effect of Cu 2ϩ , Zn 2ϩ , and Fe 3ϩ on the Fibril Growth of A␤ 1-40 -The relative fibrillogenic propensities of the apo-and holo-forms of A␤ peptides have not been understood so far. Since Cu 2ϩ or Zn 2ϩ does not induce detectable aggregation in buffer A, this system is useful to address the issue. We have investigated the effects of Cu 2ϩ , Zn 2ϩ , and Fe 3ϩ on the amyloid fibril growth of A␤  in buffer A (pH 7.4). Even at substoichiometric concentrations, Cu 2ϩ retarded the fibril growth ( Fig. 3A). At a molar ratio less than 0.4 of the metal ion to A␤ 1-40 , Cu 2ϩ increased the time lag of the onset of the fibril growth (despite providing the fibril seed for nucleation). At the molar ratio more than or equal to 0.4, fibril growth was not observed significantly during the observed time period. However, the fibril growth occurred slowly to the extent of the control (without the metal ion) by 2 days of incubation at the molar ratio less than 0.6 (data not shown). At the molar ratio of 1.0, Cu 2ϩ prevented the fibril growth completely; there was no fibril growth even after 5 days (data not shown). Zn 2ϩ also exhibited a similar effect in preventing the amyloid fibril growth of A␤   (Fig. 3B). On the other hand, Fe 3ϩ did not prevent the fibril growth of A␤ 1-40 at concentrations comparable with those of Cu 2ϩ or Zn 2ϩ (Fig. 3B) as well as at its molar ratio as high as 4.0 (data not shown).
We examined by TIRFM the samples of ThT-bound fibril seed, the A␤ amyloid fibrils, and the samples of the A␤ peptide and the seed incubated along with Cu 2ϩ or Fe 3ϩ (Fig. 3C). Fibril growth of A␤ 1-40 occurred to the order of a few micrometers in buffer A alone (Fig. 3C, image 2) and even in the presence of 0.2 mM Fe 3ϩ (Fig. 3C, image 3). On the other hand, we could see only fibril seeds in Cu 2ϩ -treated (Fig. 3C, image 4) or Zn 2ϩ -treated (data not shown) samples. Thus, Cu 2ϩ or Zn 2ϩ but not Fe 3ϩ prevents the fibril growth of A␤  .
Cu 2ϩ and Zn 2ϩ Prevent the ␤-Sheet Formation-Amyloid fibrils exhibit characteristic well ordered cross-␤-sheet structure (40,41). A␤ 1-40 freshly dissolved in buffer A exhibited random coil conformation as revealed in its far UV-CD spectrum (Fig. 4, curve 1). Incubation of the peptide in the presence of fibril seeds for 80 min generated a characteristic far-UV CD spectrum for ␤-sheet structure with a minimum around 218 nm (Fig. 4, curve 2). Such an association-induced generation of ␤-sheet structure also occurred in the presence of Fe 3ϩ (Fig. 4,  curve 3). However, induction of ␤-sheet structure did not occur in the presence of Cu 2ϩ or Zn 2ϩ (Fig. 4, curves 4 and 5). Thus, Cu 2ϩ or Zn 2ϩ but not Fe 3ϩ prevents the generation of ␤-sheet structure, consistent with the effects on fibril growth monitored by ThT fluorescence (Fig. 3).
Copper Binding to A␤  in Buffer A-Prevention of the fibril growth and the association-induced ␤-sheet formation shows that the metal ions, Cu 2ϩ and Zn 2ϩ , interact with the A␤ peptide in buffer A, where phosphate complexation of the metal ions is also favored. We further investigated the Cu 2ϩ binding by fluorescence spectroscopy.
The sequence of A␤ 1-40 does not contain any tryptophan residue but a single tyrosine residue at position 10 in addition to two phenylalanine residues at positions 19 and 20. The intrinsic fluorescence of the peptide thus predominantly represents the fluorescence of the sole tyrosine residue upon exciting at 280 nm. The intrinsic fluorescence of A␤ 1-28 is known to be quenched upon binding to Cu 2ϩ (20). We have investigated the effect of the metal ions on the intrinsic fluorescence of A␤ 1-40 peptide. The fluorescence intensity of the peptide progressively decreased upon the addition of the paramagnetic ion, Cu 2ϩ (Fig. 5A). This observed quenching of the fluorescence is not due to the general collisional quenching of the fluorescence by the paramagnetic ion, Cu 2ϩ , as the fluorescence of N-acetyl tyrosine amide was not significantly quenched under the concentration range of the metal ion used (Fig. 5B). Thus, the specific quenching of the intrinsic fluorescence of A␤ 1-40 observed in buffer A is due to the specific binding of Cu 2ϩ , leading to proximal interactions between the quencher (Cu 2ϩ ) and the fluorophore (the tyrosine residue of the peptide). Using the fluorescence quenching data (Fig. 5C), the dissociation constant (K d ) of copper binding to A␤ 1-40 was calculated to be ϳ8 ϫ 10 6 M. The addition of the diamagnetic ion, Zn 2ϩ , however, did not lead to quenching of the fluorescence of A␤ 1-40 (Fig. 5C). Thus, the fluorescence experiment shows that Cu 2ϩ binds to the A␤ peptide in buffer A, under the condition in which the metal ion-induced prevention of fibril growth but not the metal ion-induced aggregation was observed. We presume that Zn 2ϩ also interacts with or binds to A␤ 1-40 in buffer A since Zn 2ϩ behaved similarly to Cu 2ϩ on the fibril growth of A␤   (Fig. 3B) Switching On/Off the Fibril Growth-One of the reasons for our observation that fibril growth is prevented by the metal ions is that the metal ions destabilize the fibrils. However, incubation of the fibrils of the A␤ peptide with the metal ions at 37°C did not result in significant dissociation of the fibrils as judged by ThT binding (data not shown), thus ruling out the possibility of fibril instability as one of the reasons for the observed metal-induced prevention of the fibril growth. Since Cu 2ϩ -treated fibrils are stable but Cu 2ϩ prevents the fibril growth, it was possible to switch on/off the fibril growth by switching A␤ 1-40 to its fibrillogenic apo-and non-fibrillogenic holo-forms (Fig. 6). The results show that one of the key factors determining the A␤ fibril growth is the environmental concentrations of metal ions.
Metal Ion-dependent Effects of CQ on the Fibril Growth-The antibiotic CQ, which can bind both Cu 2ϩ and Zn 2ϩ with the same stoichiometry but with different coordination geometry (33), has been proposed to serve as a potential drug for AD (8,31,32). It would be pertinent to study the effects of CQ on the A␤ amyloid fibril growth as well as on the metal ion-induced suppression of the fibril growth.
CQ alone did not affect the fibril growth of A␤ 1-40 significantly (Fig. 7A). The addition of CQ or EDTA to the sample of Cu 2ϩ -treated A␤ 1-40 led to resumption of the Cu 2ϩ -suppresed fibril growth (Fig. 7B). Interestingly, the addition of EDTA, but not CQ, led to resumption of the fibril growth suppressed by Zn 2ϩ (Fig. 7C). Since 8-hydroxyquinoline derivatives are well known transition metal ion chelators (33,34) and CQ forms complexes with both Zn 2ϩ and Cu 2ϩ with a similar stoichiometry of 2:1 (33), it is unlikely that the lack of resumption of the Zn 2ϩ -suppressed fibril growth is because CQ does not chelate Zn 2ϩ . Since complexes of 8-hydroxyquinoline derivatives with Zn 2ϩ are relatively less stable when compared with those with Cu 2ϩ (33,34), a second possibility could be that the possible lesser stability of the Zn 2ϩ -CQ complex may lead to shuttling of Zn 2ϩ between CQ and the A␤ peptide, which may transiently increase the population of holo-form of A␤ 1-40 that is less competent for fibril growth. The third possibility is that the Zn 2ϩ -CQ complex exhibits novel retardation effect on fibril growth. We, therefore, attempted to verify these possibilities as follows.
At 10 M Zn 2ϩ , the fibril growth of the peptide was prevented only partially (Fig. 8A). Since CQ itself does not exhibit significant effect on the fibril growth, the addition of CQ to the sample should not have any effect in the presence of Zn 2ϩ if the first possibility of the lack of metal chelation by CQ is valid. If the second possibility is the primary reason, then in the presence of Zn 2ϩ and CQ, the fibril growth profile would be expected to be in between that of the control and that of the sample in the presence of Zn 2ϩ alone. Surprisingly, the addition of CQ led to further retardation of the fibril growth, indicating a synergistic effect of CQ and Zn 2ϩ on the fibril growth of A␤   (Fig. 8A). On the other hand, we did not observe such synergistic effect of CQ in the presence of Cu 2ϩ ; CQ completely eliminates the suppressive effect of Cu 2ϩ on fibril growth (Fig. 8B).
From the observed synergistic effect of CQ and Zn 2ϩ , it appears that the third possibility is responsible for the observed phenomenon of distinct effects of CQ on the Cu 2ϩ -and Zn 2ϩ -suppressed fibril growth; Zn 2ϩ -CQ complex, but not Cu 2ϩ -CQ complex, retards the fibril growth. Sedimentation velocity experiments of the A␤ peptide in the absence and presence of metal ions and CQ revealed no significant changes (Fig. 9), suggesting that metal ion-CQ complexes do not promote other non-fibrillar aggregates not detected by ThT fluorescence. Thus, our results demonstrate that the effect of CQ on the fibril growth process is metal ion-dependent. DISCUSSION There has been an increasing interest in the role of transition metal ions in AD (8 -10) as A␤ peptides exhibit high affinity binding to Cu 2ϩ and Zn 2ϩ (16 -22). Our study provides two important findings toward the understanding of the role of metal ions and metal ion chelators in AD.
Firstly, our study demonstrates that the soluble holo-forms (Cu 2ϩ -or Zn 2ϩ -bound) of A␤ 1-40 are less fibrillogenic or nonfibrillogenic in nature, which was not understood earlier. Al- though Fe 3ϩ is also found at high concentration in AD brain (12), it does not exhibit a significant effect on the fibrillogenic propensity of A␤  . It appears that high affinity binding of Cu 2ϩ and Zn 2ϩ may actually be beneficial as it does not favor fibril formation. On the other hand, high concentrations of metal ion under certain buffer conditions, as shown in Fig. 1, lead to aggregation and precipitation of the peptide, which may be toxic (29,30).
The second important finding is that CQ exhibits unique metal ion-dependent effects on the fibril growth of the A␤ 1-40 . Chelation of the metal ions such as Cu 2ϩ and Zn 2ϩ by EDTA leads to conditions favorable for amyloid fibril growth. Animal experiments with CQ have provided encouraging results as a potential drug (8,31,32,36). Our in vitro study shows that the addition of CQ leads to resumption of Cu 2ϩ -suppressed but not Zn 2ϩ -suppressed amyloid fibril growth of A␤  . CQ and Zn 2ϩ together exhibit synergistic retardation of the fibril growth, suggesting that retardation of fibril growth by the Zn 2ϩ -CQ complex is responsible for the observed distinct effects of CQ on the Cu 2ϩ -and Zn 2ϩ -suppressed fibril growth. However, we cannot completely rule out the possibility of shuttling of Zn 2ϩ between CQ and A␤ 1-40 due to the possible lesser stability of Zn 2ϩ complexes of 8-hydroxyquinoline derivatives (33,34), leading to the transient formation of the less fibrillogenic holoform. Both of these mechanisms should prove advantageous for the therapeutic use of CQ in AD. Monomeric A␤ is not toxic to the cells (29,30,35). Toxicity has been reported to be associated with ␤-sheet generation and fibril formation (35). On the other hand, granular aggregates, formed before fibril formation, were found be more toxic than the fibrils, although the toxicity depends on the size of the aggregates (29,30). In this respect, CQ can disaggregate the metal-induced aggregates of A␤ 1-40 under certain buffer conditions (Fig. 1) and, along with Zn 2ϩ , it can retard fibril growth (Figs. 5 and 6).
Although our results suggest that the Zn 2ϩ -complex, but not the Cu 2ϩ -complex of CQ, retards the fibril growth of A␤ 1-40 , how the complexes distinctly affect fibril propagation is not clear. We consider that the difference in the observed coordination geometry of the complexes of Cu 2ϩ and Zn 2ϩ (33) may have some role to play. The planar nature of the complex of Cu 2ϩ may lead to stacking of the aromatic ring, promoting self-association of the complex. Such stacking is not favorable due to the fifth coordination involving a water molecule as well as due to puckered configuration of the Zn 2ϩ complex of CQ (33), which can increase the availability of the complex molecule to interact or interfere in the amyloid fibril growth of the A␤ peptide, probably through hydrophobic interactions.
We conclude that Cu 2ϩ and Zn 2ϩ , but not Fe 3ϩ , significantly prevent the amyloid fibril growth of A␤ 1-40 . Our study demonstrates for the first time that the soluble holo-form of the peptide is non-fibrillogenic in nature and that the fibril growth of the peptide can be switched on/off by switching the molecule between its apo-and holo-forms. CQ exhibits distinct effects on the Cu 2ϩ -and Zn 2ϩ -induced suppression of the fibril growth. Particularly, complexation of Zn 2ϩ with CQ may have dual beneficial effects in not only disaggregating the metal ioninduced aggregates but also in retarding the fibril growth of A␤  . Our study should prove useful in understanding the role of the transition metal ions in amyloid fibril propagation of A␤ peptides in AD and provide information concerning the use of metal chelators, particularly CQ, as potential drugs for AD.