Báo cáo khoa học: Docosahexaenoic acid stabilizes soluble amyloid-b protofibrils and sustains amyloid-b-induced neurotoxicity in vitro ppt

Docosahexaenoic acid stabilizes soluble amyloid-b protofibrils and sustains amyloid-b-induced neurotoxicity in vitro Ann-Sofi Johansson 1 , Anita Garlind 2 , Fredrik Berglind-Dehlin 1 ,Go ¨ ran Karlsson 3 , Katarina Edwards 3 ,Pa ¨ r Gellerfors 1 , Frida Ekholm-Pettersson 1 , Jan Palmblad 4 and Lars Lannfelt 1 1 Department of Public Health and Caring Sciences, Uppsala University, Rudbeck Laboratory, Sweden 2 Neurotec Department, Division of Clinical Geriatrics, Karolinska Institutet, Karolinska University Hospital, Huddinge, Stockholm, Sweden 3 Department of Physical and Analytical Chemistry, Uppsala University, Sweden 4 Department of Medicine, Division of Hematology, Karolinska Institutet, Karolinska University Hospital, Huddinge, Stockholm, Sweden Alzheimer’s disease is characterized neuropathologically by two types of protein deposits, extracellular amyloid plaques and intracellular neurofibrillary tangles. Pla- ques consist mainly of fibrillar amyloid-b (Ab) protein, whereas tangles are composed of fibrillar tau protein. The aggregation of Ab is believed to drive the disease process, with extensive neuronal loss as a consequence [1]. Monomeric Ab readily aggregates via a population of soluble intermediates, protofibrils, into amyloid fibrils in vitro [2]. Which one of these Ab species drives the disease process has been debated, as both fibrils and protofibrils have previously been found to be toxic to neurons [3–5] and affect electrophysiologic parame- ters [3,4]. Fatty acids are known to affect the aggregation of various polymerizing proteins, including amyloidogenic proteins [6–8]. Arachidonic acid (AA, 20:4, x6) is, for example, commonly used to induce tau aggregation in vitro. Recently, a diet enriched in docosahexaenoic Keywords amyloid-b; docosahexaenoic acid; micelles; neurotoxicity; protofibrils Correspondence A S. Johansson, Department of Public Health and Caring Sciences, Rudbeck Laboratory, Dag Hammarskjolds vag 20, SE-75185 Uppsala, Sweden Fax: +46 18 4714808 Tel: +46 18 4715030 E-mail: ann-sofi.johansson@pubcare.uu.se (Received 2 November 2006, accepted 13 December 2006) doi:10.1111/j.1742-4658.2007.05647.x Enrichment of diet and culture media with the polyunsaturated fatty acid docosahexaenoic acid has been found to reduce the amyloid burden in mice and lower amyloid-b (Ab) levels in both mice and cultured cells. However, the direct interaction of polyunsaturated fatty acids, such as docosahexaenoic acid, with Ab, and their effect on Ab aggregation has not been explored in detail. Therefore, we have investigated the effect of docosahexaenoic acid, arachidonic acid and the saturated fatty acid arachidic acid on monomer oligomerization into protofibrils and protofibril fibrillization into fibrils in vitro, using size exclusion chromatography. The polyunsaturated fatty acids docosahexaenoic acid and arachidonic acid at micellar concentrations stabilized soluble Ab42 wild-type protofibrils, thereby hindering their con- version to insoluble fibrils. As a consequence, docosahexaenoic acid sus- tained amyloid-b-induced toxicity in PC12 cells over time, whereas Ab without docosahexaenoic acid stabilization resulted in reduced toxicity, as Ab formed fibrils. Arachidic acid had no effect on Ab aggregation, and neither of the fatty acids had any protofibril-stabilizing effect on Ab42 har- boring the Arctic mutation (AbE22G). Consequently, AbArctic-induced toxicity could not be sustained using docosahexaenoic acid. These results provide new insights into the toxicity of different Ab aggregates and how endogenous lipids can affect Ab aggregation. Abbreviations AA, arachidonic acid; AD, Alzheimer’s disease; APP, amyloid precursor protein; Arc, Arctic; CMC, critical micelle concentration; Cryo-TEM, Cryo transmission electron microscopy; CSF, cerebrospinal fluid; DAD, diode array detector; DHA, docosahexaenoic acid; MTT, 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide; PUFAs, polyunsaturated fatty acids; SEC, size exclusion chromatography. 990 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS acid (DHA, 22:6, x3) was found to dramatically reduce the amyloid burden in aged Tg2576 mice, and biochemi- cal studies demonstrated reduced Ab levels and amyloid precursor protein (APP) processing [9]. However, the direct interaction of polyunsaturated fatty acids (PU- FAs) with Ab, and their effect on Ab aggregation, have not been explored in detail. Therefore, we have investi- gated the effect of unsaturated DHA and AA and the saturated fatty acid arachidic acid (20:0) on Ab mono- mer oligomerization into protofibrils and protofibril fi- brillization into fibrils in vitro, using size exclusion chromatography (SEC) [2]. Two Ab peptides were used in this study, Ab42wt, which is implicated in sporadic Alzheimer’s disease (AD), and Ab42 with the Arctic mutation, Ab42Arc (E22G). This mutation was found in a family with hereditary AD in northern Sweden, and has been associated with the accelerated formation of Ab40 protofibrils in vitro [10]. Recently, we have shown that this is the case also for Ab42Arc, which displays accelerated protofibril formation as well as accelerated protofibril fibrillization into fibrils [2]. In addition, the effect of DHA on Ab-induced toxic- ity was investigated in PC12 cells, a cell line widely used as a neuronal cell model. Here, we demonstrate that PUFAs have a profound effect on Ab aggregation in vitro, stabilizing soluble intermediates, i.e. protofibrils, and thereby maintaining Ab-induced toxicity in PC12 cells. Results Unsaturated fatty acids stabilize Ab42wt protofibrils Fatty acids were incubated with Ab42wt in a physio- logic buffer environment at a molar ratio of 1 : 1. The mixture was assayed for monomer and protofibril con- tent, respectively, as a function of incubation time by SEC. Ab species eluting in the void volume of the Superdex 75 column, and not pelleted in the centrifu- gation step, were defined as Ab protofibrils. The gel- included peak eluting at  20 min was defined as Ab monomers but could contain dimers as well [11]. No other protein peaks were detected. Fibril formation was measured indirectly as decline in protofibril peak area, i.e. amount of peptide pelleted in the centrifuga- tion step. We have previously shown that the mass of peptide lost in the pellet corresponds to the reduced peak area, using amino acid analysis to quantitate pelleted material, and that this pelleted material is thioflavin T positive [2]. The PUFAs had a significant effect on Ab42wt in vitro assembly (Fig. 1). Both DHA (Fig. 1A) and AA (Fig. 1B) increased the monomer oligomerization rate. Interestingly, these fatty acids also had an appar- ent effect on protofibril stability. Soluble protofibrils remained stable for at least 25 h, whereas in the pres- Fig. 1. The PUFAs DHA and AA stabilize Ab42wt protofibrils, but the saturated fatty acid arachidic acid does not. SEC was used to assay protofibril and monomer content as a function of Ab42wt incubation time. (A) Ab42wt (50 l M) incubated with 50 lM DHA (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) as a control (filled symbols). DHA stabilizes protofibrils and accelerates monomer oligomerization into protofibrils. (B) Ab42wt (50 l M) incu- bated with 50 l M AA (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) as a control (filled symbols). AA stabilizes protofibrils and accelerates monomer oligomerization into proto fibrils. (C) Ab42wt (50 l M) incubated with 50 lM arachidic acid (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) (filled symbols) as a control. Arachidic acid did not have any effect on Ab aggregation, in contrast to DHA and AA. The error bars represent ± SEM of three independent experiments. A S. Johansson et al. Docosahexaenoic acid stabilizes Ab protofibrils FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 991 ence of fatty acid vehicle only (dimethylsulfoxide), essentially all protofibrils had aggregated into large fibrillar aggregates by that time. In contrast, the satur- ated fatty acid arachidic acid (Fig. 1C) did not have any effect on either Ab42wt monomer oligomerization or protofibril stability. Neither unsaturated nor saturated fatty acids stabilize Ab42Arc protofibrils When Ab42Arc peptide was investigated in the same way as Ab42wt, no stabilizing effect on Ab42Arc pro- tofibrils was observed. On the contrary, DHA acceler- ated protofibril assembly into insoluble fibrils, pelleted in a centrifugation step (Fig. 2A). AA also seemed to accelerate Ab42Arc protofibril fibrillization slightly, even though this effect did not reach statistical signifi- cance (Fig. 2B). In agreement with its effect on Ab42wt, the saturated fatty acid arachidic acid did not result in any change in aggregation kinetics for Ab42Arc (Fig. 2C). In these experiments, Ab42Arc monomers were not detected, owing to the high aggregation rate of this peptide, and the presence of Tween-20 in the elution buffer, further increasing the monomer oligomerization rate (unpublished results). Tween-20 was, however, not present during incubation, but only during analysis. In this way, interference by Tween-20 was kept at a minimum. SDS ⁄ PAGE analysis: soluble Ab42wt is stabilized by DHA In an attempt to determine whether an oligomer of a specific size is stabilized by DHA, SDS ⁄ PAGE analysis was performed. Samples were centrifuged to pellet insol- uble Ab species, and supernatants were analyzed with SDS ⁄ PAGE. Incubation with 50 lm DHA overnight stabilized soluble Ab42wt, as demonstrated by the increased amount of soluble species, ranging from mo- nomers to tetramers (Fig. 3). For Ab42Arc, there was a slight increase in the amount of monomer after incuba- tion with DHA overnight, but this was not as apparent as for Ab42wt. At 0 h of incubation, DHA seemed to slightly accelerate the aggregation of Ab42Arc into insoluble species, in line with the SEC data. No high molecular weight oligomers were observed. Ab42wt protofibril stabilization is dependent on the fatty acid concentration and micelle formation The magnitude of acceleration of Ab42wt monomer oligomerization (Fig. 4A) and increase in protofibril stability (Fig. 4B) induced by DHA was dependent on DHA concentration and micelle formation. DHA did not affect Ab42Arc protofibril stability at any concen- Fig. 2. PUFAs do not stabilize Ab42Arc protofibrils, and DHA actu- ally accelerates Ab42Arc protofibril fibrillization into fibrils. SEC was used to assay protofibril content as a function of Ab42Arc incuba- tion time. No monomers were detected, owing to the high aggre- gation rate. (A) Ab42Arc (50 l M) incubated with 50 l M DHA (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) (filled symbols) as a control. DHA accelerates protofibril fibrillization into fibrils as measured by decline in protofibril area due to pelleted spe- cies in a centrifugation step. (B) Ab42Arc (50 l M) incubated with AA (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) (filled symbols) as a control. (C) Ab42Arc (50 l M) incubated with 50 l M arachidic acid (open symbols) or with fatty acid vehicle (1% dimethylsulfoxide) (filled symbols) as control. The error bars repre- sent ± SEM of three independent experiments, except in (C), where only one experiment is shown. However, this experiment is representative of two independent experiments. Docosahexaenoic acid stabilizes Ab protofibrils A S. Johansson et al. 992 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS tration (Fig. 4C). A concentration of 1 lm DHA, i.e. 50· molar excess of Ab, did not have any effect on Ab42wt aggregation, whereas 10 lm DHA was enough to stabilize Ab42wt protofibrils, although not to the same extent as 50 lm DHA (molar ratio 1 : 1). Fatty acid micelles, detected in the SEC assay as a gel-inclu- ded peak eluting just after the void (Fig. 5), were detected at 10 and 50 lm DHA and AA, but not at 1 lm. Arachidic acid did not stabilize protofibrils or form micelles at any of the concentrations used. Cryo-transmission electron microscopy (cryo-TEM) Ab1–42wt was incubated with DHA and compared with a control with fatty acid vehicle only (dimethyl- sulfoxide). Samples were analyzed with cryo-TEM after 0 and 8 h of incubation at 37 °C. Ab incubated in the presence of DHA for 8 h formed long, soluble, fibril-like structures with somewhat atypical granular features (Fig. 6D). Control samples formed insoluble aggregates that, in most cases, were too large to be visualized by cryo-TEM (the thickness of the vitrified film is limited to 0.5 lm). One of these large aggregates is shown, however, in Fig. 6B. DHA prolongs Ab42wt-mediated cell toxicity To examine the effect of DHA on Ab-induced neuro- toxicity in a neuronal cell model, we set up a kinetic toxicity study, using PC12 cells. Ab42wt and Ab42Arc were incubated with or without 50 lm DHA for var- ious time points up to 25 h at 37 °C, and then added at the same time to the cell cultures. Toxicity was eval- uated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl- tetrazolium bromide (MTT) assay, and data were shown as percentage cell toxicity (% inhibition of MTT reduction) compared to NaCl ⁄ P i alone. The ‘toxicity’ measured by this assay refers to the impaired ability of mitochondria to reduce MTT. This assay reflects cell viability, and makes no distinction between apoptosis and necrosis. Ab42wt and Ab42Arc both Fig. 3. SDS ⁄ PAGE analysis confirms DHA stabilization of soluble Ab42wt. Ab 42wt and Ab42Arc (50 l M) were incubated overnight at 37 °C with 50 l M DHA or with fatty acid vehicle (1% dimethylsulf- oxide) as a control. Corresponding nonincubated samples were also prepared. All samples were centrifuged to pellet insoluble species. Soluble oligomers were separated with SDS ⁄ PAGE and visualized with silver staining. Lane 1: fresh Ab42wt with 1% dimethylsul foxide. Lane 2: 50 l M fresh Ab42wt with 50 lM DHA. Lane 3: 50 l M Ab42wt incubated overnight with 1% dimethylsulfoxide. Lane 4: 50 l M Ab42wt incubated overnight with 50 lM DHA. Lane 5: 50 l M fresh Ab42Arc with 1% dimethylsulfoxide. Lane 6: 50 lM fresh Ab42Arc with 50 lM DHA. Lane 7: 50 lM Ab42Arc incubated overnight with 1% dimethylsulfoxide. Lane 8: 50 l M Ab42Arc incu- bated overnight with 50 l M DHA. Fig. 4. DHA stabilizes Ab42wt protofibrils and accelerates mono- mer oligomerization in a dose-dependent manner. SEC was used to assay protofibril and monomer content as a function of Ab42 incu- bation time. Ab42Arc monomers were not detected, owing to the high aggregation rate. (A) Ab42wt monomer oligomerization as a function of Ab incubation time. Ab42wt (50 l M) was incubated with 1, 10 or 50 l M DHA (open symbols) or fatty acid vehicle (1% dimethylsulfoxide) as a control (filled symbols). (B) Ab42wt protofi- bril fibrillization into fibrils as a function of A b incubation time. Ab42wt (50 l M) was incubated with 1, 10 or 50 lM DHA (open symbols) or 1% dimethylsulfoxide as a control (filled symbols). (C) Ab42Arc protofibril fibrillization into fibrils as a function of Ab incu- bation time. Ab42Arc (50 l M) was incubated with 1, 10 or 50 lM DHA (open symbols) or 1% dimethylsulfoxide (filled symbols). A S. Johansson et al. Docosahexaenoic acid stabilizes Ab protofibrils FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 993 resulted in an initial toxicity of  30%. Ab42wt incu- bated with DHA maintained this toxicity over time, in contrast to Ab42wt incubated with vehicle, when the toxicity was reduced with prolonged Ab incubation time (Fig. 7A). Toxicity induced by Ab42Arc was also reduced with extended Ab incubation time, but for this peptide, DHA had no effect (Fig. 7B). When the pep- tides were incubated overnight with different concen- trations of DHA, a concentration of DHA shown not to stabilize protofibrils (i.e. 1 lm) did not maintain Ab42wt-induced toxicity. Ab42Arc-induced toxicity was not maintained with either a low or a high con- centration of DHA (Fig. 7C). DHA in itself did not result in increased toxicity, as Ab42Arc was equally toxic with or without DHA (Fig. 7B,C). Also, neither DHA incubated with BSA as a control protein nor DHA alone induced toxicity as compared to NaCl ⁄ P i (data not shown). It is worth noting the very high experimental reproducibility with DHA-stabilized Ab42wt peptide, illustrated by the minimal data variation in three independent experi- ments (Fig. 7C, third bar from the left). Discussion PUFAs repesent a significant proportion of the lipids in the brain, with DHA and AA being the most abun- dant [12]. DHA and AA are essential fatty acids, i.e. they cannot be synthesized de novo, but have to be ingested through the diet. Fatty acids are normally bound to different lipid-binding proteins, but they also exist as free fatty acids. PUFAs have proven effects on the aggregation of proteins, both amyloidogenic, such as a-synuclein [7] and tau [8], and nonamyloidogenic polymerizing pro- teins, such as synexin [13]. Previously, Wilson et al. showed that the presence of PUFAs increased Ab40wt and Ab42wt fluorescence to a higher extent than the presence of saturated fatty acids in a thioflavin T-bind- ing assay [8]. However, it has not been determined in which stage of the aggregation process this occurs. We have shown that PUFAs accelerate the early aggrega- tion process of Ab42wt, and stabilize soluble aggre- gates. This only happens when the concentration of fatty acid is over 10 lm, probably due to micelle for- mation at these concentrations. It thus appears that fatty acid micelles interact with the Ab peptide and facilitate nucleation, thereby accelerating the early aggregation phase. Both a-synuclein [14] and tau [15] aggregation have previously been demonstrated to be induced by anionic micellar detergents and fatty acids. The need for micellar structures explains why arachidic acid did not stabilize protofibrils in our study. The critical melting temperature for this fatty acid is 75 °C [16], and no micelle formation can occur below this temperature. Cryo-TEM revealed granular structures, which may represent DHA micelles. These granular structures are located in close proximity to the soluble Ab aggregates, indicating a direct interaction. Possibly, DHA and AA micelles work as a protective agent and prevent intermolecular interactions and consequent aggregation into insoluble fibrils. The concentrations of free DHA and AA in human cerebrospinal fluid (CSF) have been determined to be 185 nm and 86 nm, respectively [17]. In comparison, the levels of Ab1–42 in CSF are in the range 100– 200 pm [18]. This means that DHA and AA are pre- sent in 1000-fold excess compared to Ab1–42, and could thus affect the aggregation situation for this Ab peptide in vivo. However, it is not clear if the low DHA and AA concentrations found in CSF are suffi- Fig. 5. DHA and AA form micelles at concentrations over 10 lM. DHA, AA and arachidic acid (50 l M) were added to 50 lM Ab42wt and immediately analyzed by SEC, assaying protofibril (void) and monomer (elution time  20 min) content as well as micelles (elu- tion time  15 min). (A) Ab42wt (50 l M) with addition of 50 lM arachidic acid. (B) Ab42wt (50 lM) with addition of 50 lM AA. (C) Ab42wt (50 l M) with addition of 50 lM DHA. (D) Only DHA (50 l M). The large peak eluting at 25 min is fatty acid vehicle (di- methylsulfoxide). Docosahexaenoic acid stabilizes Ab protofibrils A S. Johansson et al. 994 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS cient to form micelles. Critical micelle concentrations for the pure fatty acids are substantially higher; a crit- ical micelle concentration (CMC) in the range of 200 lm has been determined for AA [15]. It is in this context important to note, however, that the CMC for surfactants is well known to decrease in the presence of polymers [19]. In line with this, other amyloidogenic proteins [14,15] have been shown to decrease the CMC for fatty acids dramatically. Moreover, DHA and AA can form mixed micelles with other fatty acids, lower- ing the CMC even further ([20] and references therein). In addition, brain trauma and ischemia have been associated with up to six-fold elevated concentrations of free DHA and AA in human CSF [21,22]. Interest- ingly, these conditions are also associated with an increased risk of AD ([23,24] and references therein). DHA did not stabilize soluble Ab in experiments with Ab42 containing the Arctic mutation, where the glutamic acid in position 22 is substituted by glycine. DHA actually slightly accelerated Ab42Arc protofibril assembly into insoluble fibrils, as demonstrated by SEC. One explanation for this could be that the rapid aggregation of Ab42Arc prevents PUFAs from inter- vening quickly or potently enough. Alternatively, the negative charge in glutamic acid could be essential for the fatty acid interaction. This seems unlikely, however, as aggregation of both the negatively charged a-synuclein [14] and the positively charged tau [15] is induced by anionic, not cationic, micelles and vesicles. In fact, an artificial negative surface, such as carboxy- late-modified polystyrene microspheres [25] or negat- ively charged mica [26], is enough to induce tau and IgG light chain aggregation, respectively. The net charge of the protein is thus probably insignificant. More likely, the anionic surface interacts with clustered positive charges; in the case of Ab, the arginine at position 5, and the lysines at positions 16 and 28. Interestingly, the formation of fibrils by association of protofilaments has been proposed to involve a salt bridge between lysine 16 and the glutamic acid at posi- tion 22 [27]. Possibly, DHA micelles can hinder this interaction, and thereby prevent fibrillization. This hypothesis could explain why Ab42Arc aggregation is not affected by DHA, as the aggregation of Ab42Arc Fig. 6. DHA prevents the formation of large insoluble aggregates. Cryo-TEM images of 50 l M Ab42wt incubated with 50 lM DHA (lower panel) or fatty acid vehicle (1% dimethylsulfoxide) (upper panel) for 0 h (left panel) and 8 h (right panel). (A) Ab42wt (50 l M) incubated for 0 h with 1% dimethylsulfoxide. (B) Ab42wt (50 l M) incubated for 8 h with 1% dimethylsulfoxide. (C) Ab42wt (50 lM) incubated for 0 h with 50 l M DHA. (D) Ab42wt (50 lM) incubated for 8 h with 50 lM DHA. Samples were not centrifuged before analysis. Scale bar ¼ 100 nm. A S. Johansson et al. Docosahexaenoic acid stabilizes Ab protofibrils FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 995 should be independent of this salt bridge because of the lack of negative charge in residue 22 (glycine). Possibly, the interaction of anionic micelles, such as DHA micelles, with Ab could be pH-dependent, as the three histidines present in Ab are positively charged at pH < 6, but uncharged at physiologic pH. As protofibrils and other oligomeric species of Ab have been found to be toxic [3–5,28,29], and DHA treatment stabilized soluble protofibrils, we examined what effects the interaction of DHA with Ab have on Ab-induced cell toxicity. PC12 cells, a tumor cell line derived from rat adrenal gland, were chosen as a cell model, as these cells work very well with the MTT assay and are sensitive to Ab. In a kinetic setting, we found that DHA sustained Ab42wt toxicity, in agree- ment with the stabilization of toxic protofibrils observed in the in vitro aggregation studies. Toxicity induced by nonstabilized Ab was reduced, as Ab aggregated into insoluble fibrils, demonstrating the increased toxicity of soluble aggregates in comparison to insoluble aggregates. In contrast, the toxicity induced by Ab42Arc incubated with DHA was indis- tinguishable from that induced by vehicle, again in agreement with the SEC data. Also, in this case, fresh peptide was more toxic than old peptide. In a recent study, a DHA-enriched diet was found to reduce amyloid burden but not soluble Ab in Tg2576 mice [9], indicating a shift from insoluble A b to soluble species. The present study demonstrates that DHA and AA stabilize soluble protofibrils, thus provi- ding in vitro data in agreement with the in vivo data from Tg2576 mice. As there is evidence, both epidemiologic and in ani- mal models, for DHA being beneficial for cognition and reducing the risk for AD, the stabilization of toxic protofibrils by DHA might seem contradictory. There could be several explanations for this. First, the experi- Fig. 7. DHA maintains Ab42wt-mediated toxicity, but not Ab42Arc- mediated toxicity in PC12 cells. Ab-mediated toxicity was evaluated over time in PC12 cells by MTT assay. Toxicity equals inhibition of MTT reduction. Noncentrifuged Ab aggregated for different time periods was incubated with the cells for 4 h at a final concentration of 5 l M Ab. Zero per cent toxicity corresponds to cultures to which only NaCl ⁄ P i was added. All samples were added in triplicate. The graphs in (A) and (B) are representative of two independent experi- ments. (A) Ab42wt (50 l M) was incubated with 50 lM DHA (open symbols) or fatty acid vehicle (1% dimethylsulfoxide) (filled sym- bols) for different time periods. All samples were then added to PC12 cells, and cell toxicity was determined. (B) Ab42Arc (50 l M) was incubated with 50 l M DHA (open symbols) or 1% dimethyl- sulfoxide (filled symbols) for different time periods and then added to PC12 cells, and cell toxicity was determined. Values represent the mean of one sample analyzed in triplicate. (C) Filled bars: 50 l M Ab42wt was incubated overnight at 37 °C in the presence of 1% dimethylsulfoxide, 50 l M DHA or 1 lM DHA, respectively, and then added to cells. Open bars: Ab42Arc (50 l M) was incubated over- night at 37 °C in the presence of 1% dimethylsulfoxide, 50 l M DHA or 1 lM DHA, respectively, and then added to cells. Error bars represent ‚± SEM of three or four independent experiments with triplicates in each. Statistical significance was determined using an unpaired t-test (n ¼ 3 or 4). Welsh correction was used in cases where the standard deviation was not equal between the groups. Docosahexaenoic acid stabilizes Ab protofibrils A S. Johansson et al. 996 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS mental setup in our toxicity experiment was designed to evaluate the effect of DHA interaction with Ab on Ab-induced toxicity, and not to examine the effect of increased DHA concentration in the cellular plasma membrane, in which case DHA naturally should have been added to the culture media. Recently, DHA enrichment of plasma membranes was demonstrated to protect neurons from Ab-induced apoptosis [30]. The levels of PUFAs in biological membranes are import- ant for membrane fluidity, which in turn is important for the interaction of membrane proteins, receptor function, and ion channels [31]. In two recent studies, DHA supplementation of cultured cells resulted in lowered Ab levels [32,33]. There are no studies, to our knowledge, on how DHA supplementation affects APP processing and trafficking in cultured cells. Most likely, such experiments would be very informative regarding the beneficial effects of PUFAs in AD. Second, the acceleration of monomer oligomeriza- tion into protofibrils induced by DHA and AA might reduce steady-state levels of smaller oligomers, e.g. Ab*56 [34] or Ab globulomers [35], which are possibly more toxic than protofibrils. However, in our SDS ⁄ PAGE analysis, no oligomers of that size ( 56– 60 kDa) could be observed, with or without DHA pre- sent. The analysis did, however, confirm the results from the SEC experiments, as soluble Ab42wt was sta- bilized by DHA and not pelleted in the previous cen- trifugation step, as was the case with Ab42wt without DHA present. Most likely, protofibrils are dissociated into monomers, dimers, trimers and tetramers under these conditions. This is probably the case also for oligomers of intermediate sizes, such as Ab*56 and Ab globulomers. Therefore, it cannot be excluded that an oligomer of intermediate size is stabilized by DHA, and subsequently disrupted by the SDS. However, using an SEC column with a higher exclusion limit ( 600 kDa), we still did not observe stabilization of a gel-included species. Most of the DHA-stabilized Ab still eluted in the void volume (data not shown). These data argue against stabilization of an intermediate smaller than 600 kDa. Also, the DHA-stabilized struc- tures observed with cryo-TEM are very large. The molecular size of protofibrils is not clear, but as these Ab species elute in the void volume of the Superdex 75 column, they should be larger than 100 kDa. Probably, this void volume peak represents a soluble population of Ab oligomers of various sizes, ranging from 100 kDa up to as high as 1000 kDa. Third, PUFAs exert a number of biological effects on cells, including activation of transcriptional factors and signal transduction systems implicated in AD. As mentioned before, a DHA-enriched diet affects APP processing and trafficking, lowering Ab levels in AD mice. This would, of course, be beneficial, and can out- weigh the presumed detrimental effect of protofibril stabilization. Experimental procedures Synthetic peptides Synthetic Ab42wt and Ab42Arc were purchased from Poly- Peptide Laboratories GmbH (Wolfenbuttel, Germany), and recombinant Ab1–42wt was purchased from rPeptide (Ath- ens, GA, USA). The lyophilized peptides were stored desicca- ted in glass vials. For all peptide solutions, siliconized tubes were used (Sigma-Aldrich, St Louis, MO, USA). All peptides were dissolved to the desired concentration using their true peptide weight, as determined by the manufacturer. Ab aggregation kinetics in vitro Aliquots of arachidic acid (20:0), AA (20:4, x6) and DHA (22:6, x3) were stored at a concentration of 5 mm in 100% dimethylsulfoxide in Eppendorf tubes at ) 80 °C. Air was evacuated with nitrogen gas prior to sealing the tubes with parafilm, in order to prevent oxidation. Ab was weighed on a Mettler Toledo AX26 Delta Range balance (Mettler Toledo, Stockholm, Sweden), dissolved in 10 mm NaOH to a concentration of 100 lm, and diluted 1 : 1 with 2 · NaCl ⁄ P i (100 mm sodium phosphate, 200 mm NaCl, pH 7.4). To the sample group, fatty acid solution (5 mm in 100% dimethylsulfoxide) was added to a final concentration of 50 lm with 1% dimethylsulfoxide. To the control group, 100% dimethylsulfoxide was added to a final concentration of 1% (dimethylsulfoxide control). SEC Samples were analyzed on a Merck Hitachi D-7000 HPLC LaChrom system with a diode array detector (VWR, Stock- holm, Sweden). Before each SEC analysis, 1.8% Tween-20 was added to the sample to a final concentration of 0.6%, resulting in a final peptide concentration of  35 lm. Prior to injection, the sample was centrifuged for 5 min, 17 900 g at 16 °C with an Eppendorf S417R centrifuge with fixed angle rotor, to remove insoluble fibrillar material. Ten microliters of supernatant was analyzed on a Superdex 75 PC3.2 ⁄ 30 column (GE Healthcare Biosciences, Uppsala, Sweden). Samples were eluted with NaCl ⁄ P i ⁄ Tween-20 (50 mm sodium phosphate, 0.15 m NaCl, pH 7.4, 0.6% Tween-20) at a flow rate of 0.08 mLÆmin )1 at ambient tem- perature. All injected samples were subjected to wavelength scan between 200 and 400 nm, and data were collected at 214 nm. Peak areas were integrated using Merck Hitachi model D-7000 chromatography data station software. A S. Johansson et al. Docosahexaenoic acid stabilizes Ab protofibrils FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 997 SDS/PAGE Samples were centrifuged at 17 900 g for 5 min with an Eppendorf S417R centrifuge with fixed angle rotor, and pro- teins in the supernatant were separated with SDS ⁄ PAGE on a Tris ⁄ Tricine 10–20% gel (Bio-Rad, Richmond, CA, USA). Bands were visualized with silver staining, and veri- fied with western blot, using the Ab-specific primary anti- body 6E10 (Signet Laboratories, Dedham, MA, USA). Cryo-TEM Peptide preparation Peptides were dissolved in 10 mm NaOH to a final concentra- tion of 100 lm, and this was followed by vortexing for 2 min. The peptide solution was then neutralized with 2 · NaCl ⁄ P i , and this was followed by vortexing for 1 min. DHA or dimethylsulfoxide was added to a final concentration of 50 lm DHA or 1% dimethylsulfoxide. The peptide solution was incubated at 37 °C for 44 h, and analyzed by cryo-TEM. Sample preparation for cryo-TEM and image recording The method consisted, in short, of the following (a more detailed description is available in Almgren et al. [36]). The samples were equilibrated at 25 °C and approximately 99% relative humidity in a climate chamber. A small drop ( 1 lL) of sample was deposited on a copper grid coated with a perforated polymer film that had been covered with a thin carbon layer on both sides. Excess liquid was then removed by means of blotting with a filter paper, leaving a thin film of the solution on the grid. Immediately after blot- ting, the sample was vitrified in liquid ethane, held just above its freezing point. Samples were kept below ) 165 °C and protected against atmospheric conditions during both transfer to the transmission electron microscope and exam- ination. The cryo-TEM investigations were performed with a Zeiss EM 902 A Transmission Electron Microscope (Carl Zeiss NTS, Oberkochen, Germany). The instrument was operated at 80 kV and in zero loss bright-field mode. Digi- tal images were recorded under low-dose conditions with a BioVision Pro-SM Slow Scan CCD camera (Proscan GmbH, Scheuring, Germany) and analysis software (Soft Imaging System, GmbH, Mu ¨ nster, Germany). In order to visualize as many details as possible, an underfocus of 1– 2 lm was used to enhance the image contrast. MTT assay PC12 cells were plated at a density of 10 000 cells per well in a 96-well plate (cell+; Sarstedts, Landskrona, Sweden) and incubated in treatment media (RPMI 1640 supplemented with 10% dialyzed fetal bovine serum) for 48–72 h. Ab at a concentration of 50 lm, incubated with or without 50 lm DHA, was added in triplicate to a final concentration of 5 lm, and incubated for 4 h at 37 °C. MTT was added to each well to a final concentration 0.5 mgÆ mL )1 , and incuba- ted for 4 h. Solubilization of formazan product was achieved by addition of 100 lL of solubilization buffer (20% SDS in dimethylformamide, pH 4.8) to each culture, and incubation for 24 h at 37 ° C. The formazan product was quantified by absorbance measurement at 570 nm (Spectramax 190). Acknowledgements This work was supported by grants from EU Consorti- ums Diadem and Apopis, Gun och Bertil Stohnes Stiftelse, Stiftelsen fo ¨ r Gamla tja ¨ narinnor, Swedish League against Rheumatism, Mary, A ˚ ke och Hans La ¨ ndells stiftelse, Alzheimerfonden, Hja ¨ rnfonden, the Swedish Research Council (project nos. 2003-5546, 71X-05991, 71BI-14589 and 2005-5581), the Research and Developmental Department, Stockholm County Council, and Bertil Ha ˚ llstens forskningsstiftelse. We thank Anne-Lie Svensson for supplying the PC-12 cells and introducing the first author to culturing of PC-12 cells and the MTT assay. References 1 Selkoe DJ (2001) Alzheimer’s disease: genes, proteins, and therapy. Physiol Rev 81, 741–766. 2 Johansson AS, Berglind-Dehlin F, Karlsson G, Edwards K, Gellerfors P & Lannfelt L (2006) Physiochemical characterization of the Alzheimer’s disease-related pep- tides A beta 1-42Arctic and A beta 1-42wt. FEBS J 273, 2618–2630. 3 Hartley DM, Walsh DM, Ye CP, Diehl T, Vasquez S, Vassilev PM, Teplow DB & Selkoe DJ (1999) Protofi- brillar intermediates of amyloid beta-protein induce acute electrophysiological changes and progressive neu- rotoxicity in cortical neurons. J Neurosci 19, 8876–8884. 4 Walsh DM, Hartley DM, Kusumoto Y, Fezoui Y, Condron MM, Lomakin A, Benedek GB, Selkoe DJ & Teplow DB (1999) Amyloid beta-protein fibrillogenesis. Structure and biological activity of protofibrillar inter- mediates. J Biol Chem 274, 25945–25952. 5 Ward RV, Jennings KH, Jepras R, Neville W, Owen DE, Hawkins J, Christie G, Davis JB, George A, Kar- ran EH et al. (2000) Fractionation and characterization of oligomeric, protofibrillar and fibrillar forms of beta- amyloid peptide. Biochem J 348 (Part 1), 137–144. 6 Ma Z & Westermark GT (2002) Effects of free fatty acid on polymerization of islet amyloid polypeptide (IAPP) in vitro and on amyloid fibril formation in culti- vated isolated islets of transgenic mice overexpressing human IAPP. Mol Med 8, 863–868. Docosahexaenoic acid stabilizes Ab protofibrils A S. Johansson et al. 998 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 7 Sharon R, Bar-Joseph I, Frosch MP, Walsh DM, Ham- ilton JA & Selkoe DJ (2003) The formation of highly soluble oligomers of alpha-synuclein is regulated by fatty acids and enhanced in Parkinson’s disease. Neuron 37, 583–595. 8 Wilson DM & Binder LI (1997) Free fatty acids stimulate the polymerization of tau and amyloid beta peptides. In vitro evidence for a common effector of pathogenesis in Alzheimer’s disease. Am J Pathol 150, 2181–2195. 9 Lim GP, Calon F, Morihara T, Yang F, Teter B, Ubeda O, Salem N Jr, Frautschy SA & Cole GM (2005) A diet enriched with the omega-3 fatty acid doc- osahexaenoic acid reduces amyloid burden in an aged Alzheimer mouse model. J Neurosci 25, 3032–3040. 10 Nilsberth C, Westlind-Danielsson A, Eckman CB, Con- dron MM, Axelman K, Forsell C, Stenh C, Luthman J, Teplow DB, Younkin SG et al. (2001) The ‘Arctic’ APP mutation (E693G) causes Alzheimer’s disease by enhanced Abeta protofibril formation. Nat Neurosci 4, 887–893. 11 Walsh DM, Lomakin A, Benedek GB, Condron MM & Teplow DB (1997) Amyloid beta-protein fibrillogenesis. Detection of a protofibrillar intermediate. J Biol Chem 272, 22364–22372. 12 Lopez GH, Ilincheta de Boschero MG, Castagnet PI & Giusto NM (1995) Age-associated changes in the con- tent and fatty acid composition of brain glyceropho- spholipids. Comp Biochem Physiol B Biochem Mol Biol 112, 331–343. 13 Sterner DC, Zaks WJ & Creutz CE (1985) Stimulation of the Ca2+-dependent polymerization of synexin by cis-unsaturated fatty acids. Biochem Biophys Res Com- mun 132, 505–512. 14 Necula M, Chirita CN & Kuret J (2003) Rapid anionic micelle-mediated alpha-synuclein fibrillization in vitro. J Biol Chem 278, 46674–46680. 15 Chirita CN, Necula M & Kuret J (2003) Anionic micelles and vesicles induce tau fibrillization in vitro. J Biol Chem 278, 25644–25650. 16 Dawson R, Elliott D, Elliott W & Jones K (1986) Data for Biochemical Research, 3rd edn. Clarendon Press, Oxford. 17 Pilitsis JG, Diaz FG, Wellwood JM, Oregan MH, Fair- fax MR, Phillis JW & Coplin WM (2001) Quantification of free fatty acids in human cerebrospinal fluid. Neuro- chem Res 26, 1265–1270. 18 Olsson A, Vanderstichele H, Andreasen N, De Meyer G, Wallin A, Holmberg B, Rosengren L, Vanmechelen E & Blennow K (2005) Simultaneous measurement of beta-amyloid (1-42), total tau, and phosphorylated tau (Thr181) in cerebrospinal fluid by the xMAP technol- ogy. Clin Chem 51, 336–345. 19 Holmberg K, Jo ¨ nsson B, Kronberg B & Lindman B (2003) Surfactants and Polymers in Aqueous Solution, 2nd edn. John Wiley & Sons Ltd, Chichester. 20 Holland P & Rubingh D (1983) Nonideal multicompo- nent mixed micelle model. J Phys Chem 87, 1984–1990. 21 Pilitsis JG, Coplin WM, O’Regan MH, Wellwood JM, Diaz FG, Fairfax MR, Michael DB & Phillis JW (2003) Free fatty acids in cerebrospinal fluids from patients with traumatic brain injury. Neurosci Lett 349, 136–138. 22 Pilitsis JG, Coplin WM, O’Regan MH, Wellwood JM, Diaz FG, Fairfax MR, Michael DB & Phillis JW (2003) Measurement of free fatty acids in cerebrospinal fluid from patients with hemorrhagic and ischemic stroke. Brain Res 985, 198–201. 23 Fleminger S, Oliver DL, Lovestone S, Rabe-Hesketh S & Giora A (2003) Head injury as a risk factor for Alz- heimer’s disease: the evidence 10 years on; a partial rep- lication. J Neurol Neurosurg Psychiatry 74, 857–862. 24 Kalaria RN (2000) The role of cerebral ischemia in Alz- heimer’s disease. Neurobiol Aging 21, 321–330. 25 Chirita CN & Kuret J (2004) Evidence for an intermedi- ate in tau filament formation. Biochemistry 43, 1704– 1714. 26 Zhu M, Souillac PO, Ionescu-Zanetti C, Carter SA & Fink AL (2002) Surface-catalyzed amyloid fibril forma- tion. J Biol Chem 277, 50914–50922. 27 Petkova AT, Ishii Y, Balbach JJ, Antzutkin ON, Leap- man RD, Delaglio F & Tycko R (2002) A structural model for Alzheimer’s beta-amyloid fibrils based on experimental constraints from solid state NMR. Proc Natl Acad Sci USA 99, 16742–16747. 28 Whalen BM, Selkoe DJ & Hartley DM (2005) Small non-fibrillar assemblies of amyloid beta-protein bearing the Arctic mutation induce rapid neuritic degeneration. Neurobiol Dis 20, 254–266. 29 Klein WL, Stine WB Jr & Teplow DB (2004) Small assemblies of unmodified amyloid beta-protein are the proximate neurotoxin in Alzheimer’s disease. Neurobiol Aging 25, 569–580. 30 Florent S, Malaplate-Armand C, Youssef I, Kriem B, Koziel V, Escanye MC, Fifre A, Sponne I, Leininger- Muller B, Olivier JL et al. (2006) Docosahexaenoic acid prevents neuronal apoptosis induced by soluble amy- loid-beta oligomers. J Neurochem 96, 385–395. 31 Youdim KA, Martin A & Joseph JA (2000) Essential fatty acids and the brain: possible health implications. Int J Dev Neurosci 18, 383–399. 32 Lukiw WJ, Cui JG, Marcheselli VL, Bodker M, Botkj- aer A, Gotlinger K, Serhan CN & Bazan NG (2005) A role for docosahexaenoic acid-derived neuroprotectin D1 in neural cell survival and Alzheimer disease. J Clin Invest 115, 2774–2783. 33 Oksman M, Iivonen H, Hogyes E, Amtul Z, Penke B, Leenders I, Broersen L, Lutjohann D, Hartmann T & Tanila H (2006) Impact of different saturated fatty acid, polyunsaturated fatty acid and cholesterol containing diets on beta-amyloid accumulation in APP ⁄ PS1 trans- genic mice. Neurobiol Dis 23, 563–572. A S. Johansson et al. Docosahexaenoic acid stabilizes Ab protofibrils FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal compilation ª 2007 FEBS 999 [...].. .Docosahexaenoic acid stabilizes Ab protofibrils A.-S Johansson et al 34 Lesne S, Koh MT, Kotilinek L, Kayed R, Glabe CG, Yang A, Gallagher M & Ashe KH (2006) A specific amyloid-beta protein assembly in the brain impairs memory Nature 440, 352–357 35 Barghorn S, Nimmrich V, Striebinger A, Krantz C, Keller P, Janson B, Bahr M, Schmidt M, Bitner... Harlan J et al (2005) Globular amyloid beta-peptide oligomer 1000 ) a homogenous and stable neuropathological protein in Alzheimer’s disease J Neurochem 95, 834–847 36 Almgren M, Edwards K & Karlsson G (2000) Cryo transmission electron microscopy of liposomes and related structures Colloids Surfaces A: Physicochem Engineering Aspects 174, 3–21 FEBS Journal 274 (2007) 990–1000 ª 2007 The Authors Journal . Docosahexaenoic acid stabilizes soluble amyloid-b protofibrils and sustains amyloid-b- induced neurotoxicity in vitro Ann-Sofi Johansson 1 , Anita Garlind 2 ,. aggregation in vitro, stabilizing soluble intermediates, i.e. protofibrils, and thereby maintaining Ab-induced toxicity in PC12 cells. Results Unsaturated fatty acids