Stereoselective Interactions of Peptide Inhibitors with the β-Amyloid Peptide

Residues 16–20 of the β-amyloid peptide (Aβ) function as a self-recognition element during Aβ assembly into fibers. Peptides containing this motif retain the ability to interact with Aβ and, in some cases, potently inhibit its assembly. Replacing l- with d-amino acids could stabilize such peptides and permit their evaluation as therapeutic agents for Alzheimer's disease. Here we have assessed the effect that such a chiral reversal has on inhibitory potency. d-enantiomers of five peptides, KLVFFA, KKLVFFA, KFVFFA, KIVFFA, and KVVFFA, were unexpectedly more active as inhibitors in an in vitro fibrillogenesis assay. Circular dichroism showed that d-KLVFFA more effectively prevented Aβ adopting the β-sheet secondary structure correlated with fibrillogenesis. Electron microscopy showed that fiber formation was also more strongly inhibited by d-KLVFFA. Heterochiral inhibition was confirmed using d-Aβ, on the principle that enantiomeric proteins exhibit reciprocal chiral biochemical interactions. With d-Aβ, l-KLVFFA was the more potent inhibitor, rather than d-KLVFFA. Most significantly, d-peptides were more potent at reducing the toxicity of both Aβ1–40 and Aβ1–42 toward neuronal cells in culture. This unforeseen heterochiral stereoselectivity of Aβ for d-peptide inhibitors should be considered during future design of peptide-based inhibitors of Aβ neurotoxicity and fibrillogenesis.

␤-Amyloid peptide (A␤) 1 is a 4-kDa peptide that when assembled into amyloid, progressively accumulates in Alzheimer's disease (AD) (1,2). The neurotoxic properties specifically associated with aggregated forms of this peptide provide key causal evidence linking A␤ to the pathology of AD (3,4). This link is supported by A␤ immunization experiments in transgenic mice expressing mutant alleles of human amyloid precursor protein. Such vaccination reduces both the A␤ peptide levels and deposits in the brain (5) and also attenuates the associated cognitive impairment in these mice (6,7). Subsequent studies showed that full assembly of A␤ into mature fibers similar to those found in plaques is not necessary for toxicity. Smaller aggregates with molecular weights corresponding to dimers (8,9), trimers, and tetramers (10) exhibit toxicity in cell culture as well. These findings support and highlight the necessity that A␤ assembly intermediates be included as targets in AD drug development.
While the delineation of which assembly intermediate(s) retain toxicity will aid in the rational design of specific inhibitors, screening assays have already identified a number of antifibrillogenic compounds (11)(12)(13). Studies show that the A␤ 16 -20 region is important for A␤ assembly (14,15), possibly as an A␤ self-recognition motif (16). These findings led to the development of a series of peptides incorporating this sequence, which bind to the homologous region of A␤ and block its oligomerization (17). In another study an eleven residue peptide containing three prolines, which have a low propensity to form ␤-sheets, inhibited A␤ fibril formation (18). Peptides that protect against A␤ neurotoxicity in cell culture have been prepared by combining the A␤-recognition motif with a disruptive element comprised of charged amino acids (19). Other inhibitors of A␤ assembly were developed by modifying peptides containing the A␤-recognition motif with cholyl groups (20) or with N-methyl amino acids (21).
The usefulness of this family of peptidic antifibrillogenic agents may depend on reducing their inactivation by proteolysis. One approach to do this is the replacement of L-with D-amino acids, which creates peptide bonds that are resistant to proteases. This approach was evaluated with proline-containing inhibitor peptides and was found to improve the inhibitor stability without altering the antifibrillogenic activity (18). Similarly a cholyl-modified peptide exhibited activity when prepared with D-amino acids (20). It is fortuitous that the mirror image conformational change in these peptides, in going from L-to D-forms, did not have a profound negative effect on their potency. This might be expected given the effects of stereochemistry on A␤ assembly (22) and other protein/protein interactions (23).
In the present study we investigated this question by quantitatively evaluating the effect of L-to D-amino acid substitutions, on the potency of peptide inhibitors of A␤ assembly. We assessed the antifibrillogenic activity using Thioflavin T fluorescence, circular dichroism and electron microscopy. Unexpectedly, we found that the D-enantiomers of these peptides were considerably more potent at inhibiting A␤ fibrillogenesis than the original L-peptides. To confirm this counterintuitive finding, we showed that fibrillogenesis of the enantiomer of A␤, D-A␤, was conversely more potently inhibited by an L-enantiomer inhibitor peptide. Most significantly, the D-enantiomers more efficiently blocked A␤-induced neurotoxicity in neuronal cell cultures. The design and optimization of antifibrillogenic molecules employing the A␤ 16 -21 recognition motif should therefore take into consideration that its interaction with A␤ displays a heterochiral stereoselectivity.

EXPERIMENTAL PROCEDURES
Synthetic Peptides-A␤ 1-40 and A␤ 1-42 were prepared by American Peptide Company, Inc., D-A␤ 1-40 was prepared by Biosource International, Inc. and peptide inhibitors were prepared by ResGen Invitrogen Corp. Peptides were of Ն95% purity. Peptide inhibitors, with the exception of KKLVFFA, were amidated at the C terminus.
Disaggregation of A␤-A␤ 1-40 and A␤ 1-42 were treated to remove aggregates using a modified version of a previously published procedure (24). A 2.5 mg/ml solution of A␤ 1-40 or A␤ 1-42 in HFIP (Aldrich) was degassed in a round bottom flask with nitrogen then sonicated in a bath sonicator (VWR) for 1 h. The solution was then filtered through an Anotop 25 Plus 20-nm filter (Whatman). Aliquots were stored in polypropylene tubes at Ϫ80°C until use.
Thioflavin T Fibrillogenesis Assay-An aliquot of A␤ 1-40 in HFIP was dried under nitrogen in a polypropylene tube. The A␤ 1-40 was dissolved to 80 M in 0.04 M Tris base, pH 10.5, and sonicated under nitrogen for 15 min in a bath sonicator. An equal volume of 0.3 M NaCl, 0.01% sodium azide solution containing 20 M ThT (Fluka) was added and the pH adjusted to pH 7.4. A 20-l aliquot of A␤ 1-40 /ThT solution was combined with 20 l of TBSA buffer containing the peptides to be tested, in a 384-well solid white polystyrene microplate (Corning). Microplates were sealed with a plastic sheet and incubated in a PerkinElmer, model HTS 7000 Plus microplate reader at 37°C. Plates were shaken for 1 min before fluorescent readings every 15 min using bandpass filters of 430 nm (excitation) and 485 nm (emission). In this procedure ThT fluorescence is measured throughout the assembly process, in a similar way to that previously described for monitoring light chain fibrillogenesis in real-time (25). The presence of ThT during the assembly reaction was previously found not to interfere with the formation of fibrillar A␤ (26). We have also observed using CD that 5 M ThT does not interfere with development of the ␤-sheet secondary structure in A␤ (data not shown).
Circular Dichroism Spectroscopy-The incubation conditions were modified for CD studies to permit measurements down to 190 nm. Buffered solutions (5 mM Tris, pH 7.4, 100 M EDTA) with and without either 10 M L-KLVFFA or D-KLVFFA inhibitor peptide and 20 M disaggregated L-A␤ 1-40 or D-A␤ 1-40 peptide were prepared. In addition, 10-M solutions of D-and L-inhibitor peptides were prepared in the absence of A␤ as controls. All solutions, 100 l, were incubated at 37°C for 48 h in sealed 96-well solid black polystyrene microplates (Corning) as described for the ThT assay. Samples were transferred to 0.1-cm path length quartz cuvettes and CD scans were performed immediately and after 48 h using a Jasco J-715 spectropolarimeter at 37°C between 190 and 250 nm, with a resolution of 0.1 nm and a bandwidth of 1 nm. The secondary structure content was estimated from the CD spectra using the PEPFIT program (27,28), which has been found suitable for use with spectra from peptides such as A␤ (29).
Electron Microscopy-The solutions described above were sampled just before CD spectroscopy for preparation of platinum/carbon replicas. was examined using a JOEL 2000 FX transmission electron microscope (JOEL) at a magnification of ϫ21,000.
A␤-induced Neurotoxicity Assays-For the A␤ 1-42 -induced neurotoxicity assay, primary cultures of rat cortical/hippocampal neurons were established from 18-day-old fetuses. After removal of meninges, cortical/hippocampal tissues were centrifuged, dissociated, and diluted to 0.8 -1.0 ϫ 10 6 cells/ml in Neurobasal B-27 medium (Invitrogen) containing 25 M glutamate (Sigma), 0.5 mM glutamine (Invitrogen), 25 M ␤-mercaptoethanol (Invitrogen), and 2% B-27 supplement (Invitrogen). The cell suspension was plated on poly-L-lysine-coated 96-well plates and incubated at 37°C, 5% CO 2 in a humid atmosphere. On day 4 and 7 of in vitro culture, the medium (50 l and 60 l/well, respectively) was exchanged with fresh pre-warmed glutamate-free Neurobasal B-27 medium (70 l and 80 l/well, respectively). To evaluate the protective effect of inhibitor peptides, culture medium was completely removed from the cells and replaced by 50 l of inhibitor peptide solution (24 or 48 M) followed by 50 l of A␤ 1-42 (24 M), both prepared in glutamatefree Neurobasal B-27 medium. The A␤ 1-42 solution was not preincubated before addition to the cells. The cells were then incubated for a further 72-h period after which viability was assessed using the WST-1 reagent (Roche Diagnostics). Briefly, 10 l of WST-1 reagent was added to each well and following incubation for 20 -40 min the absorbance was read at 450 nm using a microplate reader (PerkinElmer, model HTS 7000 Plus). For the A␤ 1-40 -induced neurotoxicity assay human neuroblastoma cells (SH-SY5Y, American Type Culture Collection) were cultured in a mixture (1:1) of Eagle's minimum essential medium (Sigma) and Ham's F12 medium (Cellgro), with 1% non-essential amino acids and 10% fetal bovine serum (Hyclone). The cells were trypsinized, suspended in the same medium with N2 supplement (Invitrogen), and plated at a density of 5 ϫ 10 4 cells/well in 96-well plates. Cells were incubated to permit adherence for a 2-h period at 37°C in 5% CO 2 . The A␤ 1-40 was preincubated for 24 h at 37°C with or without a 4-fold molar excess of peptide inhibitor and then added to the cells. After incubation for a further 24 h, the cell viability was assessed using the MTT assay (Roche Applied Science). Briefly, 10 l of MTT labeling reagent was added to each well for 4 h followed by addition of 100 l of the solubilization solution. Following incubation for 16 h at 37°C, the absorbance was read at 560 nm using a microplate reader. For both A␤ 1-42 and A␤ 1-40 , the percent cell viability was calculated as being equal to [(A 450 sample Ϫ A 450 blank)/(A 450 control Ϫ A 450 blank)] ϫ 100.

D-Enantiomers of Inhibitor Peptides More Potently Inhibit A␤
Assembly than L-Enantiomers-To evaluate the effect of chirality on the potency of peptide inhibitors of A␤ assembly, two enantiomeric pairs of inhibitor peptides were compared. One of these peptides, KLVFFA, corresponds to the primary sequence of A␤ 16 -21 , an A␤ self-recognition motif (16). A second peptide, with an additional lysine at the N terminus, KKLVFFA, was also tested, as it was previously demonstrated to bind specifically to peptides containing the A␤ 16 -21 KLVFF motif (17). The relative ability of these peptides to inhibit A␤ assembly was assessed using the amyloid specific dye ThT, which increases in fluorescence as A␤ peptides assemble into fibrillar structures (26). Both KLVFFA and KKLVFFA inhibited about 16-fold more potently when in the D-form compared with the L-form. L-KLVFFA exhibited antifibrillogenic activity at 400 M (Fig.  1A), which was similar to that of D-KLVFFA at 25 M (Fig. 1C). L-KKLVFFA at 400 M (Fig. 1D) inhibited slightly less effectively than did D-KKLVFFA at 25 M (Fig. 1F). Three additional related peptide pairs, KIVFFA, KFVFFA, and KVVFFA, were produced by substitutions at the leucine, a residue previously shown to be critical for binding (16). The results of testing these at 20 M (KIVFFA and KFVFFA) or at 100 M (KVVFFA) are presented in Fig. 2. After incubation for 10 h, the D-forms were found to be more inhibitory in each case. We observed that while replacement of the leucine by isoleucine or phenylalanine did not greatly change the inhibitor activity compared with the parent peptide D-KLVFFA, replacement by valine did reduce the inhibition. Replacement by alanine (data not shown) caused a near complete loss of inhibitor activity, consistent with the much reduced binding of 125 I-A␤ 1-40 to solid phase bound D-KAVFF compared with L-KLVFF reported previously (16). It is concluded that inhibitor peptides interacting with the KLVFF motif, inhibit A␤ fibrillogenesis more potently when in the D-form.

Enantiomeric A␤ (D-A␤ 1-40 ) Exhibits Reciprocal Chiral
Inhibitor Specificity-Enantiomeric proteins exhibit reciprocal chiral specificity with respect to their biochemical interactions (23). A␤ synthesized entirely of D-amino acids (D-A␤) was used to re-assess the effect of L-and D-forms of KLVFFA on A␤ assembly by ThT. As shown in Fig. 3, in the case of D-A␤ assembly, the L-enantiomer of KLVFFA was the more potent inhibitor.
D-Peptide Inhibitors Most Effectively Reduce the Transition of A␤  to ␤-Sheet Secondary Structure-A␤ transitions to a ␤-sheet-rich secondary structure during assembly into amyloid fibrils. We used CD to assess whether the inhibitor peptides could prevent A␤ from adopting a ␤-sheet conformation, and to determine which enantiomer was the most effective. L-A␤ 1-40 before incubation exhibited a CD spectra, which when deconvoluted (27) corresponded to a secondary structure of 60% random coil and 40% ␤-sheet (C of c ϭ 0.96, Fig. 4A). L-A␤  after incubation in the absence of peptide inhibitors, transitioned to a secondary structure 60% ␤-sheet and 40% ␣-helix (C of c ϭ 0.98; Fig. 4A). The presence of 10 M D-KLVFFA during the L-A␤ incubation prevented this random coil to ␤-sheet-rich secondary structure transition, maintaining A␤ close to its original conformation, 56% random coil and 44% ␤-sheet (C of c ϭ 0.92; Fig. 4A). The addition of 10 M L-KLVFFA to the incubation of A␤ 1-40 permitted similar structural transitions to those observed when A␤ was incubated alone, which corre-sponded to 40% ␤-sheet and 60% ␣-helix (C of c ϭ 0.98; Fig. 4A). Identical experiments performed using D-A␤ 1-40 are presented in Fig. 4B. In this case the CD spectra are mirror images of those produced by L-A␤  . The presence of D-KLVFFA had little effect on the CD spectra of D-A␤ after a 48-h incubation. In contrast, the addition of L-KLVFFA maintained the D-A␤ in a structure similar to that observed before incubation. Comparable results were obtained when similar experiments were performed in TBSA buffer, which contains physiological levels of salt (data not shown).
D-Peptide Inhibitors More Potently Reduce Fiber Formation by L-A␤ 1-40 -EM was used to determine the effect of the enantiomeric inhibitor peptides on A␤ 1-40 fiber formation. L-A␤ 1-40 (20 M) incubated alone (Fig. 5A) or in the presence of 20 M L-KLVFFA (Fig. 5B) formed large fibrillar structures with the characteristic morphology of A␤ amyloid. However, when 20 M D-KLVFFA was present during the incubation, very few or no fibrillar structures were observed (Fig. 5C). Conversely, when 20 M D-A␤ 1-40 was incubated either alone (Fig. 5D) or in the presence of 20 M D-KLVFFA (Fig. 5F), numerous densely packed fibrils were observed. In the presence of 20 M L-KLVFFA, D-A␤ 1-40 did not form fibrils (Fig. 5E).
The D-KLVFFA Inhibitor Peptide Effectively Blocks A␤ Assembly Facilitated by Pathological Chaperones-The effect of the D-KLVFFA inhibitory peptide was tested in an A␤ fibrillogenesis assay in the presence of heparin. This glycosaminogly- can acts as a natural pathological chaperone, facilitating the formation of amyloid fibrils in vitro (30) and is believed to play a role in A␤ fibrillogenesis in vivo (13). In our A␤ assembly assay, heparin increased the ThT fluorescence by 1.8-fold. When the D-KLVFFA peptide was added to the incubations containing both A␤ and heparin, no increase in ThT fluorescence was observed, indicating a complete inhibition of chaperone catalyzed A␤ assembly (Fig. 6).
D-Inhibitor Peptides Selectively Protect Against L-A␤ 1-42 and LA␤  Toxicity in Cell Culture-A␤ 1-42 at 12 M reduced the viability of cultured rat primary cortical neuronal cells to 20 and 29% of that observed in control cells incubated without A␤   (Fig. 7A). In the presence of 6 or 12 M D-KKLVFFA, the toxicity of 12 M A␤ 1-42 was significantly attenuated (p Ͻ 0.001). The resulting cell viabilities were 91.9% and 123.1% of control cells respectively, with the effect of 12 M peptide being significantly greater (p Ͻ 0.001) than that of 6 M peptide. In the presence of 6 or 12 M, of L-KKLVFFA the toxicity of 12 M A␤ 1-42 persisted, and the viability remained similar to that seen without addition of the peptide inhibitors, 29.5% and 29.4% of control cells, respectively. A␤ 1-40 at 1 M reduced viability of SH-SY5Y human neuroblastoma cells to 57.3% of control levels (Fig. 7B). The cell viability in the presence of A␤ 1-40 was significantly improved to 75% of control levels (p Ͻ 0.001) by the addition of 4 M D-KKLVFFA, but was not significantly changed, 54%, after the addition of 4 M L-KKLVFFA. A␤ 1-40 at 2.5 M reduced viabil-ity to 54.3% of control. This toxicity was prevented by the addition of 10 M D-KKLVFFA, which resulted in a viability of 75% of control cells (p Ͻ 0.001). This was not seen with the addition of 10 M L-KKLVFFA, which resulted in a viability of 52.7%, not different from that seen with 2.5 M A␤ alone.
Full Potency of D-Peptides Requires All D-Amino Acids-We investigated the antifibrillogenic activity of L-KLVFFA after the systematic replacement of each L-residue by its corresponding D-amino acid (Fig. 8A). Each such chiral switch of a single residue resulted in a loss of antifibrillogenic activity except in the case of the N-terminal lysine residue. The greatest losses in activity were observed when the L-amino acids from the central region of the peptide (residues 17 to 20) were substituted. Similarly, all single D-to L-substitutions in D-KLVFFA decreased the potency of the D-enantiomer (Fig. 8B). Again, the greatest loss of potency occurred when the chiral inversion was made at the center of the peptide.
D-Peptides Containing the A␤ 16 -20 Motif Inhibit A␤ Assembly by Interacting at the Complementary Region of A␤-Self-recognition of the A␤ 16 -20 KLVFF motif is believed to take place with the L-enantiomer inhibitor peptides described thus far (17,19). To establish whether this is also the case with the Denantiomers of peptide inhibitors, a series of overlapping 6 and 7 D-peptides spanning the entire length of A␤ were synthesized and evaluated using the ThT fibrillogenesis assay (Fig. 9). The only peptides found to exhibit inhibitory activity, D-A␤ 16 -22 and D-A␤ 15-20 , contained the entire A␤ 16 -20 sequence.

Inhibition of A␤ Assembly by Peptides Containing the KLVFF A␤-Recognition Motif Is
Heterochirally Selective-Protein/protein interactions can be highly stereospecific. For example, the HIV-1 protease cleaves only the naturally occurring L-form of its substrate peptide, thus exhibiting a homochiral specificity. A synthetically prepared D-enantiomer of this HIV-1 protease maintains the homochiral specificity characteristic of the native enzyme, by acting only on the D-form of its substrate (23). Analogously, the ribonuclease S-protein specifically binds the L-form, but not the D-form of the S-peptide to form an active complex (31). In other cases, protein/protein interactions exhibit no chiral preference. For example, a synthetic HIV-1 GP41 fusion peptide inhibits HIV-1 envelope glycoprotein mediated-cell fusion both in the D-and L-form indicating that the peptide's chirality is not critical for peptide/peptide interactions (32). Similarly the co-chaperone DnaJ-bound D-peptides and L-peptides with comparable affinities (33).
We show here that inhibitor peptides containing the entire KLVFF A␤-recognition motif inhibited A␤ assembly with a heterochiral type of stereospecificity. D-Enantiomers of these inhibitor peptides preferentially inhibited the assembly of the natural L-enantiomer of A␤. Conversely, L-enantiomers of these inhibitor peptides preferentially inhibited the assembly of the synthetic D-enantiomer of A␤. We have demonstrated this unusual heterochiral specificity in A␤ assembly assays using the amyloid specific dye ThT. We have also observed this specificity in the inhibition of the structural transition of A␤ to ␤-sheet employing CD and in the inhibition of A␤ fiber formation employing EM. The hetereochiral specificity of the inhibitor peptides was also evident in their biological activity, as the Dpeptides more effectively blocked A␤-induced neurotoxicity in cultured neurons.
We have considered alternative explanations for our observations. For instance the possible selective contamination of L-inhibitor peptides by aggregation promoting metal ions such as copper (34) would cause an apparent reduction in the potency of this enantiomer in our assays. However such a possibility was excluded based on our results showing that the L-peptides had greater potency than the D-peptides, when inhibiting the assembly of D-A␤  . This demonstrated that the lower potency of the L-peptides with L-A␤ 1-40 , is an inherent characteristic of the interaction, and not attributable to a contaminant.
We have considered the possibility that when an inhibitor peptide of opposite chirality is combined with A␤ the resulting solution fails to develop a ␤-sheet-type spectra because of spectral subtraction rather than because of inhibition of a change in A␤ structure. We do not believe this to be the case here, because we observe a spectrum for the solution of A␤ and inhibitor which exhibits a negative minimum at around 197 nm characteristic of high random coil content. If in the mixture of A␤ and inhibitor of opposite chirality both had converted to a high ␤-sheet secondary structure, the resulting CD spectra would instead have cancelled each other out, at all wavelengths, producing a flat spectrum at zero mdeg. If the spectral cancellation were not complete, we would have expected some residual positive or negative signal at the 218 nm, characteristic of ␤-sheet. The results we observe are more consistent with the peptide inhibitors causing a true inhibition of the conversion of A␤ to a high ␤-structure form.
The A␤ 16 -20 Region Is the Site of D-Peptide Inhibitor Interaction with A␤-A previous study reported that peptides containing the KLVFF A␤-recognition motif interact with A␤ through the homologous primary sequence, in a homochiral-specific manner, when studied using a direct binding assay (17). In view of this report we questioned whether the D-peptide inhibitors function by interacting with A␤ through this site. To establish whether this indeed was the case, we prepared and tested a series of 20 overlapping 6-and 7-mer D-peptides cor- responding to the sequence of A␤ beginning at the N-terminus. From the ThT assay results it was clear that only peptides incorporating the KLVFF motif, namely A␤ 15-20 and A␤ 16 -22 , possessed antifibrillogenic activity. This suggests that the Dpeptide inhibitors do indeed interact with A␤ via the KLVFF motif, just as do the L-peptides. It remains unclear why a homochiral stereospecificity was observed for the direct binding of A␤ to immobilized peptides containing the KLVFF motif (17) while a heterochiral stereospecificity was observed for the antifibrillogenic activity of the inhibitor peptides described in the present report. The replacement of amino acids by their enantiomers causes a local change of secondary structure (35) which we observed resulted in a loss of inhibitor potency (Fig.  8). It is possible that the solid phase attachment of the Dpeptide (17) in a similar manner restricted its conformation to ones that do not interact with A␤.
D-KLVFFA Inhibits A␤ Assembly Promoted by a Pathological Chaperone-Certain non-A␤ components that are associated with AD plaques promote A␤ assembly and are therefore referred to as "pathological chaperones" (36,37). Glycosaminoglycans are important examples of such pathological chaperones that induces both ␤-sheet structure and amyloid fibrils in A␤ (30). For an inhibitor to function effectively in vivo, inhibition of chaperone-mediated fibrillogenesis would be a requirement. D-KLVFFA, when added to incubations promoted by heparin, completely inhibited fibrillogenesis, clearly demonstrating that not only does this peptide inhibit A␤ assembly, it is also effective when a chaperone plays a catalytic role in the process. Further testing would be required to determine whether this inhibitor peptide is active with other pathological chaperones as well.
D-Enantiomer Inhibitor Peptides are More Potent Than Are L-Forms at Countering A␤ Neurotoxicity-We found that the D-enantiomer of KKLVFFA was more potent than the L-form in preventing the neurotoxicity of A␤ in vitro. This neuroprotective activity is consistent with the prevention of A␤ assembly by this inhibitor peptide. As the toxicity of A␤ has been shown to require some type of supramolecular assembly, the ability of D-peptides to prevent such intermolecular interactions by A␤ understandably leads to neuroprotection in cell culture experiments. This effect was demonstrable with A␤ 1-40 and A␤ 1-42 , the two forms of A␤ prominently found in AD. The greater neuroprotective effect of the D-enantiomer was evident both in primary culture of rat neurons and in cultures of neuroblastomal cells. Although the L-enantiomer failed to exhibit neuroprotective effects under the conditions tested here, a higher concentration could have proven effective. This is suggested by the ability of higher concentrations of the L-inhibitors to prevent the ThT fluorescence of L-A␤. Indeed, previous studies have reported modulation of A␤ toxicity by L-enantiomeric derivatives containing the KLVFF motif (38). Showing that Dpeptides are more potent than their L-counterparts at reducing A␤ toxicity indicates that they could constitute useful drug leads for therapeutic interaction in AD.