HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 282, Issue 14, 10311-10324, April 6, 2007

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 282/14/10311    most recent M608207200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Necula, M. Articles by Glabe, C. G. Search for Related Content PubMed PubMed Citation Articles by Necula, M. Articles by Glabe, C. G. Social Bookmarking                      What's this?

Small Molecule Inhibitors of Aggregation Indicate That Amyloid Oligomerization and Fibrillization Pathways Are Independent and Distinct^*

From the Department of Molecular Biology and Biochemistry, University of California, Irvine, California 92697

Received for publication, August 25, 2006 , and in revised form, February 5, 2007.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Alzheimer disease is characterized by the abnormal aggregation of amyloid peptide into extracellular fibrillar deposits known as amyloid plaques. Soluble oligomers have been observed at early time points preceding fibril formation, and these oligomers have been implicated as the primary pathological species rather than the mature fibrils. A significant issue that remains to be resolved is whether amyloid oligomers are an obligate intermediate on the pathway to fibril formation or represent an alternate assembly pathway that may or may not lead to fiber formation. To determine whether amyloid oligomers are obligate intermediates in the fibrillization pathway, we characterized the mechanism of action of amyloid aggregation inhibitors in terms of oligomer and fibril formation. Based on their effects, the small molecules segregated into three distinct classes: compounds that inhibit oligomerization but not fibrillization, compounds that inhibit fibrillization but not oligomerization, and compounds that inhibit both. Several compounds selectively inhibited oligomerization at substoichiometric concentrations relative to amyloid monomer, with some active in the low nanomolar range. These results indicate that oligomers are not an obligate intermediate in the fibril formation pathway. In addition, these data suggest that small molecule inhibitors are useful for clarifying the mechanisms underlying protein aggregation and may represent potential therapeutic agents that target fundamental disease mechanisms.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Protein aggregation into amyloid fibrils is a pathological hallmark of many neurodegenerative diseases, including Alzheimer disease (AD).² AD is characterized, in part, by the aggregation of amyloid protein (A) into fibrillar amyloid plaques in select areas of the brain (1). Compelling evidence indicates that A aggregation is critical for neurodegeneration, suggesting that preventing this process may be an effective therapeutic approach for the treatment of AD (2–4).

A number of small molecules have been reported to inhibit A fibrillogenesis. Since it was initially presumed that toxicity is associated with mature fibers (5–8), the majority of inhibitor screens have been directed toward identifying modulators of A fibrillization. These fibril inhibitor screens have resulted in the discovery of multiple inhibitor molecules (9–24). Some compounds have also been shown to inhibit A-mediated cellular toxicity, and this activity was correlated with modulation of fibrillization (13, 15, 16, 21, 25).

However, A aggregation is a complicated process and appears to involve more than a simple conversion of soluble monomer to fiber. More recent evidence has pointed to the role of soluble amyloid oligomers or prefibrillar aggregation intermediates as the primary toxic species in degenerative amyloid diseases (2, 3). Electron microscopy and atomic force microscopy have identified spherical particles of 3–10 nm that appear at early times of incubation and disappear as mature fibrils appear (26). These spherical oligomers appear to represent intermediates in the pathway of fibril formation, because they are transiently observed at intermediate times of incubation during fibril formation. Although oligomers are kinetic intermediates, it is not yet clear whether they are obligate intermediates in the pathway for fibril formation, whether they coalesce directly to form fibrils (26, 28–33), or whether oligomers populate a different aggregation pathway that is distinct from the classic nucleation-dependent fibril assembly pathway (34–36). This same controversy extends to the aggregation of many other amyloidogenic proteins, since many types of amyloids display the same type of kinetically unstable intermediate, the soluble oligomer. Insulin, immunoglobulin light chain, amylin, -synuclein, transthyretin, prion protein, ₂-microglobulin, -lactoglobulin, phosphoglycerate kinase, and albebetin oligomers have been described as "off pathway" species (37–47). Oligomers of insulin, amylin, huntingtin, and albebetin have been reported to be "on pathway" intermediates or precursors of fibril formation (47–50). Therefore, the published data are equivocal as to whether oligomers are intermediates on the pathway leading to fiber formation or whether they represent "off pathway" aggregates that populate an alternative aggregation pathway (51).

Soluble A oligomers have been referred to by a variety of names, including amorphous aggregates, micelles, protofibrils, prefibrillar aggregates, and ADDLs (26, 29, 52, 53, 55–57). At longer aggregation times, curvilinear fibers form that have a beaded appearance. These structures have also been called "protofibrils," because they appear to be formed by the coalescence of the spherical subunits (26). These prefibrillar soluble oligomers are specifically recognized by a polyclonal antibody, A11,³ that recognizes a generic backbone epitope that is common to the oligomeric state independent of the protein sequence (58). A11 does not recognize A monomer, A dimer, trimer, or tetramer or A fibrils (58). A11 positive oligomers display the same intermediate kinetics as observed for soluble oligomers and protofibrils by electron microscopy and atomic force microscopy and A11 blocks the toxicity of A oligomers, indicating that they represent the primary toxic species. A*56 is a soluble oligomeric form of A that is closely associated with pathogenesis in the Tg2576 mouse model of AD, and A*56 is specifically recognized by A11 anti-oligomer antibody (59). Although ADDLs were originally described as low molecular weight trimeric or tetrameric species, more recent investigations indicate that native masses of ADDLs are the same as previously reported for other A-soluble oligomers (60). A11 and anti-ADDL antibodies identify the same time course of soluble oligomer accumulation in the 3xTg-AD mouse (61). Therefore, A11 antibody recognizes a significant and important class of oligomers associated with AD that are distinct from amyloid fibrils.

Analysis of the mechanism of action of amyloid aggregation inhibitors holds the promise of clarifying the relationship between oligomers and fibrils, because if oligomers are obligate intermediates on the pathway to fibril formation, then all inhibitors that block oligomer formation would be expected to block fibril formation as well. Alternatively, if fibrils and oligomers represent distinct aggregation pathways, then it would be expected that some inhibitors would block oligomerization but not fibril formation.

Here we utilize the anti-oligomer antibody, A11, as a primary read out for oligomer formation, and we analyze the mechanism of action of small molecules that have been reported to inhibit the aggregation of different amyloidogenic proteins or their toxicity. The results demonstrate that some inhibitors specifically target oligomers, whereas others specifically inhibit fibrillization. These data indicate that soluble oligomers are not an obligate intermediate for fibril formation and that oligomers and fibrils represent separate and distinct aggregation pathways. The results further indicate that screening for aggregation inhibitors using fibril-specific assays, like thioflavin T (ThT) fluorescence, will not necessarily identify inhibitors of oligomerization.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Materials—Synthetic A₄₂ was prepared as previously described (57). 200-mesh formvar/carbon-coated nickel grids were obtained from Electron Microscopy Sciences (Ft. Washington, PA), and 4–20% Tris-HCl gels were from Bio-Rad. 96-well clear, flat bottom microplates were from Nalge Nunc International (Rochester, NY), and 3,3',5,5'-tetramethylbenzidine was from KPL (Gaithersburg, MD). A11 anti-oligomer antibody is available from Invitrogen. Horseradish peroxidase-conjugated anti-rabbit IgG (AR) was purchased from Promega (Madison, WI), 6E10 and 4G8 antibodies were from Signet (Dedham, MA), and the ECL chemiluminescence kit was from Amersham Biosciences. 0.2-µm nitrocellulose membranes were form Schleicher & Schuell. Small molecule compounds and all other reagents were from Sigma or Calbiochem. The following small molecules or analogs of small molecules reported previously to bind amyloid or to modulate protein aggregation and/or toxicity or screened for such activities were tested here separately for their ability to inhibit A₄₂ oligomer and fiber formation: apigenin (62), azure C (62), basic blue 41 (62), (trans, trans)-1-bromo-2,5-bis-(3-hydroxycarbonyl-4-hydroxy)styrylbenzene (BSB) (63), Chicago sky blue 6B (25), Congo red (10, 25, 64–66), -cyclodextrin (67, 68), curcumin (11, 69), daunomycin hydrochloride (15, 20), dimethyl yellow (62), direct red 80 (25), 2,2'-dihydroxybenzophenone (62), hexadecyltrimethylammonium bromide (C16) (9), hemin chloride (62, 70), hematin (62, 70), indomethacin (71), juglone (72), lacmoid (62), meclocycline sulfosalicylate (72), melatonin (18), myricetin (62), 1,2-naphthoquinone (16), nordihydroguaiaretic acid (11), R(–)-norapomorphine hydrobromide (19), orange G (25), o-vanillin (2-hydroxy-3-methoxybenzaldehyde) (24), pherphenazine (62), phthalocyanine (62), rifamycin SV (16, 62, 73), phenol red (74), rolitetracycline (15, 20), quinacrine mustard dihydrochloride (62), thioflavin S (25, 66), ThT (75), and trimethyl(tetradecyl)ammonium bromide (C17) (9). In addition, diallyltartar, eosin Y, fenofibrate, neocuproine, nystalin, octadecylsulfate, and rhodamine B have also been tested.

Inhibition of A₄₂ Oligomerization—A₄₂ stock solutions (2 mM) were obtained by dissolving the lyophilized peptide in 100 mM NaOH followed by water bath sonication for 30 s. The oligomerization reaction was initiated by diluting the stock solution in phosphate-buffered saline (PBS), pH 7.4, 0.02% sodium azide (45 µM final A₄₂ concentration, final pH 7.4). Because A₄₂ forms oligomers that are free of fibers at early time points, we refer to these conditions as "oligomerization conditions." The reactions were incubated at 25 °C for up to 15 days in the absence or presence of small molecule compounds (0.01–3000 µM) dissolved in Me₂SO. Incubation at 25 °C was chosen because the rates of fibril nucleation and elongation are favored by incubation at 37 °C (76, 77). Control reactions were carried out in the presence of 1% Me₂SO vehicle. The oligomerization reactions were assayed by dot blot, ELISA, Western blot, and transmission electron microscopy (TEM), as described below. In addition, A₄₂ oligomerization was conducted in H₂O, pH 2–3, in the absence or presence of select compounds (300 µM) using HFIP-based stock solutions, according to a previously published protocol (58). The effects of select compounds were assessed by the dot blot assay.

Inhibition of A₄₂ Fibrillization—A₄₂ stock solutions (2 mM) were obtained by dissolving the lyophilized peptide in 100 mM NaOH followed by water bath sonication for 30 s. Fibrillization of A₄₂ (45 µM final concentration) was initiated by diluting the stock solution in 10 mM HEPES, 100 mM NaCl, 0.02% sodium azide, pH 7.4. Because A₄₂ forms fibers that are free of oligomers, we refer to these conditions as "fibrillization conditions." The reactions were stirred at room temperature for up to 6 days in the presence or absence of small molecule compounds (30 and 300 µM) dissolved in Me₂SO. Control reactions were carried out in the presence of 1% Me₂SO vehicle. The reactions were assayed by light scattering, ThT fluorescence assay, and TEM, as described below. Inhibition of fibrillization was also monitored under the oligomerization conditions described above.

Dot Blot Assay—The dot blot assay was performed as previously described (58) to detect A₄₂ oligomer formation. Briefly, 2-µl aliquots of each oligomerization reaction were applied onto nitrocellulose membranes. The membranes were blocked (1 h at room temperature or overnight at 4 °C) with 10% nonfat milk in Tris-buffered saline containing 0.01% Tween 20 (TBS-T). 0.02% sodium azide was added for overnight blocking. The membranes were then washed and incubated with affinity-purified anti-oligomer antibody (A11, 0.8 µg/ml) (58) diluted in TBS-T containing 5% milk for 1 h at room temperature. The membranes were washed again and incubated with horseradish peroxidase-conjugated AR diluted 1:5000 in TBS-T containing 5% milk for 1 h at room temperature. Then the blots were washed and developed with the ECL chemiluminescence kit. Separate membranes served as controls for the presence of A₄₂ and were treated as described above, except that the A11 and AR antibodies were replaced with 6E10/4G8 and horseradish peroxidase-conjugated anti-mouse IgG, respectively. These antibodies were diluted 1:10,000 in TBS-T containing 3% bovine serum albumin. All washes were performed with TBS-T, three times for 5 min, except for the last wash before detection, which was for 10 min.

ELISA—An ELISA was performed as previously described (58). Briefly, A₄₂ was subjected to oligomerization in the absence and presence of select small molecules (0.01–3000 µM) identified as oligomerization inhibitors by the dot blot assay, as described above. Control reactions were supplemented with 1% Me₂SO vehicle. Aliquots of each oligomerization reaction were applied to 96-well clear, flat bottom plates containing 100 µl of coating buffer (0.1 M sodium bicarbonate, pH 9.6). The A₄₂ concentration used was within the linear range of the assay. The plates were incubated 2 h at 37°C, washed, blocked with 10% bovine serum albumin/TBS-T for 2 h at 37°C, and washed again. Then 100 µl of A11 antibody (1:1500 dilution in 3% bovine serum albumin/TBS-T) was added to each well, and the plates were incubated for 1 h at 37 °C. After washing, 100 µl of AR antibody (1:5000 dilution in 3% bovine serum albumin/TBS-T) was added to each well and the plates were incubated for 1 h at 37 °C. The plates were then washed and developed with 3,3',5,5'-tetramethylbenzidine. The reactions were stopped with 100 µl of 1 M HCl and assayed by absorbance at 450 nm. PBS was used for all washes, which were performed three times. Each reaction was performed in triplicate, and the data points were fit to dose-response curves as described before (78), using the Sigmaplot software (Systat Software Inc., Point Richmond, CA). The IC₅₀, defined as the concentration of small molecule required to attain half-maximal absorbance, was obtained from the fit.

Western Blot—10-µl aliquots of each oligomerization reaction, small molecule-treated or control containing the Me₂SO vehicle, were mixed in a 1:1 ratio (v/v) with 2x SDS sample buffer, boiled for 5 min, and loaded onto 4–20% Tris-HCl gels. The gels were transferred to nitrocellulose membranes, which were then subjected to the dot blot assay with the A11 antibody.

ThT Fluorescence Assay—To investigate the time course of A₄₂ fibril formation under fibrillization conditions and in the absence of small molecules, 10-µl aliquots were removed at various time points during the fibrillization reaction and mixed with 120 µl of ThT (3 µM, dissolved in the fibrillization buffer). ThT fluorescence was measured at _ex = 442 nm and _ex = 482 nm until a plateau was reached in a Gemini XPS plate reader (Molecular Devices, Sunnyvale, CA). To investigate the effect of small molecules on A₄₂ fibril formation and on ThT fluorescence, similar measurements were performed, except that all readings were taken after 4 days of incubation and ThT emission was recorded at both 482 nm (in the presence of A₄₂) and 482 and 520 nm (in the absence of A₄₂).

Turbidity Assay—A₄₂ was incubated under both oligomerization and fibrillization conditions, as described above, in the presence and absence of compounds for 7 and 4 days, respectively. Each reaction was then assayed by turbidity at a 400-nm wavelength, to estimate the amount of fibrillar material. The reactions were then centrifuged at 14,000 rpm for 30 min. Part of each supernatant was removed and assayed by the same method at a 400-nm wavelength. The resulting values were then used to correct each of the turbidity readings corresponding to the assembly reactions.

Electron Microscopy—1-µl aliquots of the aggregation reactions were adsorbed onto 200-mesh formvar/carbon-coated nickel grids until dry. The grids were then washed with water, stained with 2% uranyl acetate, and washed again. The grids were allowed to dry between all steps and were viewed in a Phillips CM 12 microscope operated at 65 kV.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Kinetics of A₄₂ Oligomerization—We have recently described an oligomer-specific antibody (A11) that recognizes A oligomers and protofibrils and does not react with monomeric A, fibrillar A, or the amyloid precursor protein (58). We used this antibody to specifically monitor oligomer formation independently from fibril formation, which was measured by light scattering (17, 79) and ThT fluorescence (80). Lyophilized A₄₂ was dissolved in 100 mM NaOH, and the oligomerization reactions were initiated by diluting the resulting stock solution in 1x PBS, pH 7.4, as described under "Experimental Procedures." Aliquots of each reaction were removed at various time points and tested in parallel by immunobloting with A11 antibody and TEM. At time point zero, these aliquots reacted only very weakly with the A11 antibody, had strong reactivity with both 6E10 and 4G8 antibodies (Fig. 1A), and showed little aggregation by TEM (Fig. 1B), consistent with the majority of A₄₂ being in nonaggregated form. Significant A11 immunoreactivity was observed as early as 1 day after the initiation of aggregation (Fig. 1A) and correlated with the appearance of small oligomers by TEM (Fig. 1C). Strong A11 immunoreactivity was observed after 4 days of incubation, indicative of the formation of large amounts of A11-positive oligomers (Fig. 1A). TEM analysis confirmed the presence of oligomers at this time point (Fig. 1D). Longer incubation (i.e. 6 days) resulted in the appearance of A₄₂ fibrils, which appeared to coexist with A11-positive oligomers for extended periods of time (Fig. 1, A and E). The presence of 1% Me₂SO, which is used as vehicle for the small molecule compounds, did not significantly alter the kinetics of aggregation (Fig. 1A).

View larger version (52K): [in this window] [in a new window]   FIGURE 1. Time course of A42 assembly under conditions that facilitate oligomerization. A42 was incubated for up to 15 days in 1x PBS, pH 7.4, at room temperature without stirring (45 µM final concentration), in the presence or absence of 1% Me2SO (DMSO), using stock solutions of freshly dissolved peptide. Aliquots were removed at various time points during aggregation and analyzed in parallel by dot blot (A) and TEM (B–E) assays. A, aliquots were spotted onto nitrocellulose membranes and probed with A11, 6E10, and 4G8 antibodies. Oligomer-specific immunoreactivity formed within 1 day of aggregation, became strong after 4 days of incubation, and was stable for extended periods of time. 6E10 and 4G8 immunoreactivity diminished as aggregation proceeded. B–E, aggregation reactions were also assayed by TEM at 50,000-fold magnification. A42 showed minimal aggregation at time point zero (B) but formed oligomers as small as 3 nm within a day of the initiation of aggregation (C). The oligomers appeared to grow in size to yield a heterogeneous population of oligomers and protofibrils by day 4 (D). These oligomers and protofibrils constituted the predominant population for up to 6 days, after which time they coexisted with fibrils with straight morphology (E). Bar, 150 nm.

Identification of A₄₂ Oligomerization Inhibitors—The ability of small molecules to inhibit A₄₂ oligomer formation was examined using oligomer-specific antibody immunoreactivity in a dot blot assay. Control samples were treated with 1% Me₂SO vehicle, and the small molecule additives were used at 30 and 300 µM, when present. Aggregation was conducted at room temperature without stirring for up to 5 days under oligomerization conditions. The control reactions showed strong A11 immunoreactivity, indicative of oligomer formation (Fig. 2). Reactions incubated with some of the small molecules showed decreased A11 immunoreactivity (Fig. 2), suggesting that these compounds may block A₄₂ oligomer formation. The test compounds did not interfere with the binding of A₄₂ to the nitrocellulose membrane, as determined by 6E10 immunoreactivity at time point zero of aggregation (data not shown). The active compounds are numbered and listed in Table 1. None of the molecules exhibited A11 reactivity in the absence of A₄₂ with the exception of hemin and hematin (data not shown). Therefore, the apparent enhancement of A11 immmunoreactivity of A₄₂ in the presence of hemin (Fig. 2, A3 and A4) and hematin (Fig. 2, H7 and H8) constitutes an artifact that can be attributed to this phenomenon, and this assay cannot determine whether these compounds interfere with oligomer formation. These data indicate that small molecules are capable of inhibiting A₄₂ oligomerization, and some of the molecules that have previously been reported as inhibitors of fibril formation inhibit oligomerization as well.

View this table: [in this window] [in a new window]   TABLE 1 Potency of compounds active for inhibition of A42 oligomerization

View larger version (104K): [in this window] [in a new window]   FIGURE 2. Select small molecules inhibit A42 oligomerization. A42 was incubated under oligomerization conditions without stirring (45 µM, final concentration) for up to 5 days, in the absence of any additive (column 1), in the presence of 1% Me2SO vehicle (column 2), or in the presence of small molecules listed under "Experimental Procedures" (all other columns). The ability of each compound to inhibit oligomerization was tested at two different concentrations (30 and 300 µM) placed adjacent to each other (i.e. compound 1, row A, columns 3 and 4; compound 2, row A, columns 5 and 6). The active compounds are listed in Table 1, and their positioning is as follows: 7 (A5, A6), 12 (A7, A8), 10 (A9, A10), 23 (B3, B4), 25 (B5, B6), 27 (B9, B10), 29 (C3, C4), 21 (C5, C6), 11 (D3, D4), 15 (D5, D6), 18 (D7, D8), 13 (D9, D10), 20 (E9, E10), 4 (F3, F4), 9 (F5, F6), 16 (F7, F8), 5 (F9, F10), 6 (G5, G6), 24 (G7, G8), 2 (H3, H4), 28 (H5, H6), 22 (H9, H10), 17 (I5, I6), 19 (I7, I8), 1 (I9, I10), 3 (J3, J4), 26 (K3, K4). The intensity of each dot represents the amount of oligomer present in each reaction, probed with the oligomer-specific antibody A11. More than half of the small molecules tested were able to inhibit A42 oligomerization. Buffer treated with 1% Me2SO served as negative control (K8, K9), and preformed A42 oligomers were used as positive control (K7, K10).

HFIP-based A₄₂ solutions were subjected to oligomerization in H₂O, as previously described (58) in the absence and presence of 1% Me₂SO vehicle. In the absence of Me₂SO, A₄₂ formed A11-immunoreactive oligomers within 1 day of initiation of aggregation, consistent with previous observations (58) (data not shown). In the presence of Me₂SO, oligomers were not detected even after extended incubation times (data not shown). Reactions containing 1% Me₂SO are the true control for this experiment, because all the compounds are delivered in Me₂SO. Therefore, the ability of the inhibitors of A₄₂ oligomerization (as determined under oligomerization conditions) to inhibit oligomer formation obtained from HFIP-based A₄₂ solutions could not be examined.

View larger version (73K): [in this window] [in a new window]   FIGURE 3. Effect of inhibitors on A42 oligomeric species. A42 was incubated under oligomerization conditions without stirring (45 µM, final concentration) for 5 days in the presence of 1% Me2SO vehicle (lanes O) or 300 µM small molecules: azure C (lanes 1), basic blue 41 (lanes 2), meclocycline sulfosalicylate (lanes 3), R(–)-norapomorphine hydrobromide (lanes 4), Congo red (lanes 5), rolitetracycline (lanes 6), Chicago sky blue 6B (lanes 7), hemin (lanes 8), o-vanillin (lanes 9), daunomycin hydrochloride (lanes 10), C16 (lanes 11), 1,2-naphthoquinone (lanes 12), nordihydroguaiaretic acid (lanes 13), hematin (lanes 14), C17 (lanes 15), neocuproine (lanes 16), lacmoid (lanes 17), rifamycin SV (lanes 18), juglone (lanes 19), myricetin (lanes 20), ThT (lanes 21), 2,2'-dihydroxybenzophenone (lanes 22), rhodamine B (lanes 23), curcumin (lanes 24), phenol red (lanes 25), indomethacin (lanes 26), quinacrine mustard dihydrochloride (lanes 27), pherphenazine (lanes 28), and eosin Y (lanes 29). Each reaction was assessed for immunoreactivity with anti-oligomer-specific antibody A11 (A) or 6E10 (B) by Western blot. The small molecules inhibited oligomer formation in a manner that is generally consistent with their IC50 values.

We also analyzed the same Western blots with 6E10 to visualize the effects of compounds on A₄₂ species that are not detected with A11 (Fig. 3B). In general, we observed the same reciprocal relationship between A11 staining and 6E10 staining for the majority of samples as we had observed previously in the dot blot assay (Fig. 1A). This indicates that the conformation of the 6E10 and A11 epitopes are maintained even in the presence of SDS. The 6E10 staining also revealed differences in the sizes of the products that accumulate in the presence of the inhibitors. Azure C, basic blue 41, meclocycline, and o-vanillin (Fig. 3B, lanes 1–3 and 9) promoted the accumulation of A₄₂ aggregates ranging in size from 50 kDa to material that sticks at the top of the gel. This size range overlaps that of the high molecular weight A₄₂ oligomers stained by A11 in the absence of inhibitors, indicating that conformationally distinct, SDS-resistant aggregates of approximately the same size can be detected by Western blots. Since many of the compounds that cause the accumulation of SDS-resistant, A11-negative, 6E10-positive high molecular weight aggregates appear to actually promote fibril formation (see below), the simplest interpretation is that this material may represent fibrils and small fragments of fibrils that are partially sensitive to dissociation by SDS. Indeed, Western blots of fibril preparations show a similar size distribution of SDS-resistant 6E10 immunoreactivity that is not recognized by A11 (supplemental Fig. 2). Other compounds (Fig. 3B, lanes 4, 6, 12, 17, 18, 20, 21, and 24) promoted the accumulation of both the high molecular weight aggregates and low molecular weight monomer, dimer, and trimer that are A11-negative. A third class of compounds (Fig. 3B, lanes 5, 11, 14, 15, 16, and 19) caused the accumulation of low molecular weight A and/or material that sticks at the top of the gel but very low amounts of intermediate sized aggregates. A few compounds (Fig. 3, lanes 8, 13, and 27) appeared to inhibit the accumulation of both A11- and 6E10-positive bands. The interpretation of this latter group is not yet clear. They may promote the formation of products that do not display either A11 or 6E10 epitope, or they may favor the precipitation of A₄₂ into a form that is not solubilized by the sample buffer.

These data confirm the inhibitory activity of the molecules identified as A₄₂ oligomerization inhibitors by the dot blot assays and ELISAs and suggest that some of the inhibitors may actually promote fibril formation. In addition, the potency qualitatively visualized by Western blots correlates well with the potency quantitatively determined by ELISA.

Inhibition of Fibril Formation—Immunoreactivity with the A11 anti-oligomer antibody indicates that only a subset of the small molecules tested constitute inhibitors of A₄₂ oligomer formation. This suggests that some of the small molecules tested may be specific inhibitors of fibrillization, since many of them were originally reported as inhibitors of fibril formation. To test this hypothesis directly, A₄₂ was incubated under fibrillization conditions, as described under "Experimental Procedures," in the presence of 1% Me₂SO (control reaction) or small molecules (30 and 300 µM). Fibril formation was assayed by ThT fluorescence, light scattering, and TEM analysis. Under these conditions, A₄₂ supplemented with 1% Me₂SO readily assembled into fibers through a nucleation-dependent mechanism (supplemental Fig. 3A). The lag time of assembly was about 13 h, and the ThT fluorescence plateau was reached in 4 days. TEM confirmed that aggregation was minimal at time point zero and that A₄₂ assembled into fibrils with classic morphology after 4 days of incubation (supplemental Fig. 3, B and C). Because A₄₂ fibers formed rapidly under these conditions (e.g. the reaction reached plateau levels within 4 days) and because they do not show A11-immunoreactive oligomer contamination by dot blot assay (data not shown), these assembly conditions allow a reasonably rapid assay of the effects of compounds exclusively on fiber formation.

Although ThT fluorescence is useful for assaying aggregation in control reactions and establishing working conditions, its utility for quantifying fibrillization in samples containing small organic molecules that adsorb significantly at 442 nm (the excitation wavelength of ThT) is limited by absorption and fluorescence artifacts (Table 2). Monitoring ThT fluorescence in the absence of A revealed that most compounds strongly interfered with ThT fluorescence when emission was monitored at both 520 nm (emission maximum of ThT) and 482 nm (emission maximum of ThT in the presence of fibrillar material). Some compounds interfered with ThT absorption, causing a decrease of ThT emission, whereas others increased fluorescence emission due to their intrinsic fluorescent properties (Table 2). The decrease of ThT fluorescence observed in reactions containing A₄₂ and small molecules meclocycline sulfosalicylate, Chicago sky blue 6B, hemin, o-vanillin, C16 (30 µM), hematin, neocuproine, lacmoid, rifamycin SV, rhodamine B, eosin Y, orange G, diallyltartar, direct red, and apigenin (Table 2) paralleled the decrease of A₄₂ turbidity (Fig. 4B) and the decrease of fibrillar content by TEM (supplemental Fig. 4 and data not shown), indicating that these compounds inhibit fibrillization. However, the reliability of ThT as an indictor of fibril inhibition in these cases may be questionable, since most of these molecules significantly altered ThT fluorescence in the absence of A (Table 2).

View this table: [in this window] [in a new window]   TABLE 2 Inhibition of A42 fibrillization by small molecules: correlation between ThT fluorescence and turbidity measurements

Therefore, for the purpose of clarifying the effect of small molecules on A₄₂ fibrillization, aggregation was also qualitatively assayed by turbidity at a 400-nm wavelength, as previously described (17, 79). Because small molecules can undergo significant spectral changes in the presence of proteins (78, 83), correcting the turbidity measurements by subtracting buffer blanks containing compounds and lacking the protein is not always adequate. Artifacts related to this phenomenon were avoided by using the supernatant of the fibrillization reactions, obtained after centrifugation, to correct the turbidity readings, as described under "Experimental Procedures." Four days after the initiation of aggregation and in the absence of compounds, A₄₂ had significant turbidity (Fig. 4, A and B) and assembled into fibrils (supplemental Fig. 3C). In the presence of oligomerization inhibitors, the turbidity of most reactions either increased (Fig. 4A) or decreased (Fig. 4B) compared with the fibrillar control. The fact that some compounds increase turbidity compared with controls suggests that they may actually promote fibrillization. The effects on A₄₂ fibrillization were consistent at both compound concentrations examined and are shown here when compounds were tested at the higher concentration, 300 µM (Fig. 4). C16 and C17, however, constitute exceptions, and their effect on aggregation is shown at both concentrations. More than half of the small molecules tested did not inhibit A aggregation into fibrils, as shown by constant or increased levels of turbidity in reactions containing these molecules relative to control (Fig. 4A and data not shown).

The compounds that inhibited A₄₂ oligomerization but either promoted fibrillization or did not inhibit it include azure C, basic blue 41, R(–)-norapomorphine hydrobromide, Congo red, rolitetracycline, daunomycin, C16 (300 µM), 1,2-naphthoquinone, nordihydroguaiaretic acid, C17 (300 µM), juglone, myricetin, ThT, curcumin, indomethacin, quinacrine mustard dihydrochloride, and pherphenazine. All of these oligomerization inhibitors, except for Congo red, rolitetracycline, myricetin, and indomethacin, appear to actually promote fibrillization (Fig. 4A). These specific inhibitors of oligomerization are referred to as Class I compounds (Table 3). The remaining A₄₂ oligomerization inhibitors, meclocycline sulfosalicylate, hemin, o-vanillin, C16 (30 µM), hematin, C17 (30 µM), neocuproine, lacmoid, rifamycin SV, 2,2'-dihydroxybenzophenone, rhodamine B, phenol red, and eosin Y, partially inhibited A₄₂ fibril formation (Fig. 4B). The small molecules that inhibit both A₄₂ oligomerization and fibrillization are referred to as Class II compounds (Table 3). The small molecules apigenin, Chicago sky blue 2B, diallyltartar, direct red, and orange G did not inhibit oligomerization but partly inhibited A₄₂ fibrillization (Fig. 4B). These compounds are referred to as Class III compounds (Table 3). Melatonin, -cyclodextrin, octadecylsulfate, BSB, thioflavin S, phthalocyanine, fenofibrate, dimethyl yellow, and nystalin did not inhibit either oligomerization or fibrillization (data not shown) when assayed by the methods described under "Experimental Procedures."

View this table: [in this window] [in a new window]   TABLE 3 Classes of compounds active for inhibition of A42 aggregation

View larger version (24K): [in this window] [in a new window]   FIGURE 4. Effect of compounds on A42 fibrillization. A42 was incubated under fibrillization conditions with stirring (45 µM final concentration) for 4 days in the presence of 1% Me2SO vehicle or small molecules (30 and 300 µM). Data are shown here for 300 µM compound concentration except for C16 and C17, for which data are plotted at both concentrations. Aliquots of each reaction were assayed by turbidity at 400 nm wavelength. Small molecules tested segregated into compounds that do not inhibit (A), partly inhibit (B), or have biphasic behavior (A and B; C16 and C17) on A42 assembly into fibrils. Each reading was obtained in triplicate and is expressed as a percentage of the control reaction ± S.D. Compounds that inhibited neither oligomerization nor fibrillization are not shown.

Since fibrils ultimately form after extended incubation times under oligomerization conditions, we also tested the effects of the small molecule inhibitors under these conditions where both oligomers and fibrils form TEM. TEM analysis showed that the Class I oligomer-specific inhibitory compounds azure C, basic blue 41, R(–)-norapomorphine hydrobromide, daunomycin, C16, 1,2'-naphtoquinone, nordihydroguaiaretic acid, C17, juglone, ThT, curcumin, quinacrine mustard dihydrochloride, and pherphenazine promote A₄₂ fiber formation after 3 days of incubation, when A₄₂ is still present in oligomeric form in the control reaction (supplemental Fig. 5, A–N). These data show that these compounds promote fiber formation and are consistent with the 6E10 Western blot data (Fig. 3B) and data obtained under fibrillization conditions (Fig. 4A). The remaining Class I compounds Congo red (supplemental Fig. 5O), myricetin, and indomethacin did not induce fiber formation at this time point. This is also consistent with the lack of activity of these compounds on A₄₂ fiber formation observed under fibrillization conditions (Fig. 4A).

To test whether the compounds inhibit fibril formation under the same conditions, aliquots of each of these reactions were also removed 7 days after initiation of aggregation and analyzed by turbidity, as explained under "Experimental Procedures." At this time point, control A₄₂ solutions in the absence of test compounds exist as a mixture of oligomers and fibers under oligomerization conditions (Fig. 1). Since pure oligomer solutions have a low turbidity, the signal corresponding to fibers is expected to be the main contributor to the turbidity measurements. Consistent with observations obtained at early time points during aggregation (3 days; supplemental Fig. 4), Class I compounds did not inhibit A₄₂ fibrillization upon prolonged incubation (supplemental Fig. 6A). In contrast, Class II and Class III compounds inhibited fiber formation (supplemental Fig. 6B). Data obtained after 7 days of incubation under oligomerization conditions (supplemental Fig. 6) is also consistent with the effect of compounds on A₄₂ fiber formation assessed under fibrillization conditions (Fig. 4). Taken together, these data indicate that the effects of compounds on fiber formation are consistent regardless of the experimental conditions.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
A₄₂ Assembly Pathways—The first major set of findings of this study is related to mechanisms underlying A assembly. We characterized the mechanism of action of a set of small molecules that have been reported as inhibitors of amyloid aggregation or amyloid toxicity in order to determine which steps in the A₄₂ assembly pathways they inhibit. We screened for oligomer-inhibitory activity using the anti-oligomer-specific antibody A11 (58), and we used the widely employed ThT and light scattering assays for measuring fibril formation (80, 85). Since all assays are sensitive to potential artifacts, we also characterized the effects of the inhibitors by Western blotting and TEM. We found that the inhibitory compounds segregate into three classes based on their activity. Class I compounds inhibit oligomerization without inhibiting fibrillization, Class II compounds inhibit both, and Class III compounds only inhibit fibrillization. The Class I compounds can be further subdivided on the basis that a subset of compounds actually promote fibril formation, whereas the remainder have no effect on fibrillization. The finding that Class I compounds block oligomerization without inhibiting fibrillization indicates that A₄₂ oligomer formation is not an obligatory step on the pathway leading to fiber formation. Rather, oligomer formation constitutes an alternate aggregation pathway (Fig. 5). This is consistent with the observations that oligomer formation is considerably more sensitive to urea treatment than fibril formation and that fibril formation can proceed efficiently under concentrations of urea where oligomers are undetectable by A11 (86). Previous results from ultrastructural analysis have been interpreted as indicating that spherical A oligomers coalesce to form "protofibrils," which then form mature fibrils (36, 87), implying that oligomers are intermediates on the fibril assembly pathway. The finding that oligomers are not obligate intermediates does not necessarily imply that oligomers do not ultimately form fibrils, since there may be more than one pathway leading to A fibril formation. Alternatively, oligomers may represent an "off pathway" assembly state that does not directly convert to fibrils but buffers the concentration of monomer that ultimately assembles into fibrils. Further work will be necessary to unambiguously clarify this issue.

View larger version (8K): [in this window] [in a new window]   FIGURE 5. Hypothetical model for A42 aggregation in vitro. In vitro, freshly dissolved A42 mostly exists as an unfolded monomer (M). Based on their ability to modulate A42 aggregation into oligomers and fibers, the compounds tested here appear to stabilize at least three types of intermediate A42 conformations. Class I compounds stabilize an A42 intermediate conformation (I1) that is assembly-competent and can aggregate directly into fibrils without previous association into oligomers. Class II compounds stabilize an A42 intermediate conformation (I2), that does not favor A42 assembly into either oligomers or fibrils. Class III compounds stabilize an A42 intermediate conformation (I3) that supports A42 oligomer formation but is incompetent for assembly into fibrils. The dotted lines are used to indicate pathways that are not favored but possible. Our data suggest that A42 adopts multiple intermediate conformations that are not prone to aggregation or support selective aggregation into either oligomers or fibers. The presence of oligomers on the classic A42 fibrillization pathway is possible but not obligatory.

An unexpected finding is that A oligomers formed under different conditions display polymorphism in terms of differential epitope accessibility. This type of polymorphism has recently been reported for prion oligomers (93). We found that oligomers prepared by dilution into PBS from NaOH stock solutions are recognized by A11 but display greatly reduced 6E10 and 4G8 immunoreactivity, indicating that these epitopes are either conformationally altered or inaccessible in NaOH oligomers. In contrast, oligomers prepared by dilution of HFIP stock solutions into water at pH 3 are equally immunoreactive with A11, 6E10, and 4G8 (58). Oligomers prepared by the two methods have the same morphology and size distribution by electron microscopy. These results indicate that there are two distinct conformations of A11-positive A oligomers and suggest that sandwich ELISA methods that use 6E10 or 4G8 may not detect all types of A oligomers.

Pharmacological Implications—Recent advances facilitate the specific screening for small molecules that inhibit A₄₂ oligomerization and fibrillization. The development of the anti-oligomer antibody A11 (58) allows the specific identification of inhibitors of this important class of oligomers, whereas careful selection of well established fibrillization assays allows accurate screening for inhibitors of fibrillization. ThT fluorescence has been widely used to measure amyloid fibrillization (80, 85). Although this assay is useful for quantifying fibrillar material in many circumstances (24), its use for drug screening purposes may be limited by small molecule interference with fluorophore adsorption (94) and possible competitive binding to amyloid fibers (78).

We identified 29 small molecules capable of inhibiting oligomer formation. Their potency varies greatly, from compounds that are active at substiochiometric concentrations relative to A₄₂ monomer to compounds that require an 70-fold excess to inhibit oligomer formation by 50%. The substoichiometric inhibitors include azure C, basic blue 41, meclocycline, and R(–)-norapomorphine hydrobromide that are active at nanomolar concentrations. So far, only few compounds have been shown to inhibit A oligomerization. These are curcumin (69), -cyclodextrin derivatives (68), Congo red (64), Ginkgo biloba extracts (95, 96), and benzyl-containing compounds, including o-vanillin (24). Although we confirmed this activity for Congo red, o-vanillin, and curcumin, only the first two were active at low micromolar concentrations. These data suggest that oligomers are amenable to drug treatment by a variety of unrelated compounds, and the low concentrations required for inhibition predict that such an approach is therapeutically feasible. We were unable to assess the effect of these inhibitors on oligomerization of A₄₂ in distilled H₂O at pH 2–3, when the peptide is delivered from HFIP-based stock solutions (58), so it remains unclear if the inhibitors are effective under these conditions.

In addition, we found that 18 compounds partially inhibited A₄₂ fiber formation. Of these, hemin, hematin, lacmoid, rifamycin SV, 2,2-dihydroxybenzophenone, and apigenin have been previously shown to inhibit A and tau fibrillization (16, 62, 70, 73). Meclocycline has been reported as an inhibitor of huntingtin aggregation (72), and phenol red was identified as an amylin fibrillization inhibitor (74), suggesting that these compounds may constitute general inhibitors of fiber formation. o-vanillin inhibited A fibrillization (24), and Chicago sky blue and direct red diminished cytotoxicity, an activity correlated with inhibition of A fibrillization (25). Although the concentration required for inhibitory activity differs, the biphasic modulation of A fibrillization observed here with C16 and C17 was previously reported (9). Besides identifying new inhibitors of A fibrillization, our data confirm such inhibitory activity for the above compounds. However, we could not confirm the previously reported inhibitory activity on fiber formation for azure C, basic blue 41 (62), R(–)-norapomorphine hydrobromide (19), rolitetracycline, daunomycin (15), nordihydroguaiaretic acid (11), myricetin (62), curcumin (11, 69), quinacrine mustard dihydrochloride (62), pherphenazine (62), melatonin (18), dimethyl yellow, and phtalocyanine (62). One explanation for the discrepancy between the results presented here and previous observations may raise from the choice of assays. Because of reasons explained above, we did not rely solely on the ThT assay used previously to test many of these compounds. Small molecules might have concentration-dependent multiphasic behavior on modulating protein aggregation (this work; see Refs. 9, 94, 97–99); therefore, this discrepancy could also be attributed to the choice of compound concentrations used to test activity. Juglone did not inhibit A fibrillization, but it appears to inhibit huntingtin aggregation (72), suggesting that this molecule is a specific inhibitor of fibrillization (17, 100). The compounds were active inhibitors of fibrillization under both of the experimental conditions used.

Taken together, these data suggest that selective inhibition of either A₄₂ oligomerization or fibrillization is possible, which allows the separate targeting of either species. This is consistent with previous observations that calmidazolium chloride promotes the conversion of A monomer into clusters of protofibrils, indicating that selective targeting of specific A species is possible (22). The fact that oligomer and fibril formation can be inhibited independently also indicates that screening for fibril inhibitors will only identify a subset of potential oligomer inhibitors.

Mechanism of Inhibition—Most of the Class I compounds, in addition to selectively inhibiting oligomerization, appear to promote fiber formation. We speculate that these molecules stabilize A₄₂ conformations that do not support oligomerization but rather favor fiber formation (Fig. 5, I₁). The fact that some of them function at a concentration 100-fold lower than that of A suggests that these inhibitors preferentially interact with a rate-limiting intermediate, such as a seed or nucleus, that is present at a low concentration. One possible explanation is that these compounds inhibit oligomerization by promoting the formation of fibril seeds, which is consistent with the idea that oligomerization and fibrillization are competing alternative pathways. Since A toxicity is dependent on aggregation state and oligomers and fibers appear to be at opposite ends of the toxicity range, the discovery of molecules that promote fiber formation at the expense of oligomers may be therapeutically useful.

Class I compounds Congo red, rolitetracycline, myricetin, and indomethacin do not significantly increase the amount of fibrillar material at concentrations tested here but still favor fibrillization over oligomerization and may act by a similar mechanism. Class II compounds inhibit A₄₂ assembly into both oligomers and fibrils. We therefore speculate that these compounds stabilize conformations (Fig. 5, I₂) that do not support aggregation. Benzyl-containing compounds were previously reported to belong to this class (24). Here we confirm such action for one of these compounds, o-vanillin. Because, under conditions tested here, curcumin did not inhibit fibril formation, we suggest that this molecule belongs to Class I compounds instead of Class II as previously reported (69). Class III compounds inhibit fibrillization but do not inhibit oligomer formation. This observation predicts that compounds in this class stabilize a conformation (Fig. 5, I₃) that supports oligomerization but does not favor fiber formation. Although numerous small molecules have been previously reported to inhibit fiber formation, few have been tested for their effect on oligomerization. The exceptions include studies that identified naphthalene sulfonates and catecholamines as members of this class of compounds (101, 102).

The inhibitory effects observed here can be attributed solely to the stabilization of compound-bound A₄₂ conformations (Fig. 5, I₁, I₂, and I₃) that either do not support assembly or selectively favor aggregation into either oligomers or fibrils. Binding of these molecules to A₄₂ either depletes the assembly-competent A species, therefore inhibiting aggregation, or promotes aggregation through alternate pathways via stabilization of conformationally different intermediates. Inhibitors of aggregation reported previously to trap assembly-competent species into incompetent conformations include tau assembly inhibitor N744 (78, 100) and -synuclein fibrillization inhibitors baicalein and dopamine (103, 104). Inhibitors or enhancers of aggregation that act by stabilization of early assembly-competent intermediates, as we propose here for compounds pertaining to Classes I and III, include anionic surfactants, planar aromatic dyes, and urea (66, 105, 106). Alternatively, it is also possible that inhibition/enhancement of aggregation results from direct binding of the small molecules to oligomeric/fibrillar A₄₂, similar to Congo red (10, 54). A more complicated inhibitory action that includes both these mechanisms is also possible.

The data presented here suggest that selective inhibition of specific aggregated A species is feasible and useful both for unraveling mechanisms underlying protein fibrillization and for therapeutic testing in models of neurodegeneration.

	FOOTNOTES

 
^* This work was supported by National Institutes of Health Grant NS-38298 and the Larry L. Hillblom Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. 1–6.

¹ To whom correspondence should be addressed: Dept. of Molecular Biology and Biochemistry, 3438 McGaugh Hall, Irvine, CA 92697. Tel.: 949-824-6081; Fax: 949-824-8551; E-mail: cglabe{at}uci.edu.

² The abbreviations used are: AD, Alzheimer disease; A, amyloid protein; A₄₂, amyloid protein containing 42 amino acids; BSB, (trans, trans)-1-bromo-2,5-bis-(