Hyperforin Prevents B-Amyloid Neurotoxicity and Spatial Memory

Home , Curcumin, Hyperforin

Molecular Psychiatry (2006) 11, 1032–1048 & 2006 Nature Publishing Group All rights reserved 1359-4184/06 $30.00 www.nature.com/mp ORIGINAL ARTICLE Hyperforin prevents b-amyloid neurotoxicity and spatial memory impairments by disaggregation of Alzheimer’s amyloid-b-deposits MC Dinamarca1, W Cerpa1, J Garrido2, JL Hancke3 and NC Inestrosa1 1Centro de Regulacioń Celular y Patologıá ‘Joaquıń V. Luco’ (CRCP), MIFAB, Santiago, Chile; 2Departamento de Biologıá Celular y Molecular Facultad de Ciencias Biolo´gicas, Pontificia Universidad Cato´lica de Chile, Santiago, Chile and 3Instituto de Farmacologıá, Universidad Austral, Valdivia, Chile

The major protein constituent of amyloid deposits in Alzheimer’s disease (AD) is the amyloid b-peptide (Ab). In the present work, we have determined the effect of hyperforin an acylphloroglucinol compound isolated from Hypericum perforatum (St John’s Wort), on Ab-induced spatial memory impairments and on Ab neurotoxicity. We report here that hyperforin: (1) decreases amyloid deposit formation in rats injected with amyloid fibrils in the hippocampus; (2) decreases the neuropathological changes and behavioral impairments in a rat model of amyloidosis; (3) prevents Ab-induced neurotoxicity in hippocampal neurons both from amyloid fibrils and Ab oligomers, avoiding the increase in reactive oxidative species associated with amyloid toxicity. Both effects could be explained by the capacity of hyperforin to disaggregate amyloid deposits in a dose and time-dependent manner and to decrease Ab aggregation and amyloid formation. Altogether these evidences suggest that hyperforin may be useful to decrease amyloid burden and toxicity in AD patients, and may be a putative therapeutic agent to fight the disease. Molecular Psychiatry (2006) 11, 1032–1048. doi:10.1038/sj.mp.4001866; published online 25 July 2006 Keywords: amyloid-b-peptide; hyperforin; neurotoxicity; spatial learning; disaggregation

Introduction a1-antichymotrypsin,13 Ginkgo biloba extract,14 type IV collagen15 and b-sheet breaker peptides.16 Never- Alzheimer’s disease (AD) is one of the most common theless, an effective therapeutic approach that inter- neurodegenerative dementias, characterized by a 1 feres directly with the neurodegenerative process in progressive deterioration of cognitive functions. AD is eagerly awaited. The neurotoxicity of the amyloid-b-peptide (Ab)is On the other hand, the progressive deterioration of highly dependent on its conformation, quaternary 2–4 memory and learning causes that AD patients com- structure and morphology of the bundles formed. It monly exhibit symptoms of depression in the early has been proposed that Ab aggregation would require stages of the disease, since they realize that their inherent depolymerization mechanisms in order to cognitive functions are getting worse. Hyperforin explain the morphology and the stable size of the 5 (HYP) is an acylphloroglucinol, a photosensitive senile plaques observed in AD, which indicates that and natural derivative from Hypericum perforatum, amyloidogenesis is a continuous process of polymer- also known as the St John’s Wort. The natural product ization and depolymerization.6 As Ab is toxic to 7 is a complex mixture of compounds comprising neurons, the main targets for therapeutic interven- several natural derivatives and HYP was identified tion of the Ab cascade include the inhibition of Ab as the molecule responsible for the antidepressant production, the inhibition of Ab aggregation and fibril activity by a mechanism involving the inhibition of formation, in addition to the inhibition of the uptake of monoamines and other neurotransmitters.17 consequent inflammatory responses caused by the Besides its antidepressive activity, it has been Ab deposition. In this context, several substances are suggested that HYP possesses memory-enhancing known to inhibit Ab fibrillogenesis in vitro, including properties in rodents.18 These antecedents drove us laminin,6,8 melatonin,9 nordihydroguaiaretic acid,10 11 12 to investigate whether hyperforin derivatives would polyphenols, site-directed monoclonal antibodies, be able to reduce both the b-amyloid deposition and improve spatial learning acquisition. We used a rat Correspondence: Dr NC Inestrosa, CRCP Biomedical Center P., model consisting of an intrahippocampal stereotaxic Catholic University of Chile, PO Box 114-D, Santiago, Chile. bilateral injection of preformed Ab fibrils. The E-mail: [email protected] Received 19 January 2006; revised 1 May 2006; accepted 23 May injections were administered into the hippocampus 2006; published online 25 July 2006 to induce the formation of b-amyloid deposits and the Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1033 spatial learning acquisition was evaluated using the mances of the different groups were recorded. After Morris water maze protocol.16 After the behavioral training, the animals were fixed by intracardiac studies were finished, a histological analysis of the perfusion to carry out the histochemical procedures. hippocampal region was performed, to evaluate amyloid deposition, the reactive astrocytes and Behavioral test microglia around the injection site. All animals were trained in a circular water maze We report here that HYP treatment partially (1.6 m diameter and 75 cm deep, painted black)20 induces b-amyloid hippocampal burden fragmenta- using a two trial per day regimen. The platform tion and decreased of astroglial and microglial (9 cm diameter) is located in the center of northwest reaction and both events can be related to a significant quadrant (hidden platform quadrant). Data were improvement of the spatial learning acquisition. HYP gathered with a video tracking system for water maze prevents both Ab fibrils as well as Ab oligomer (HVS Imagen, Hampton, UK). Briefly, the rats were neurotoxicity, reducing the reactive oxygen species trained with two trials a day, for 5 consecutive days, generated by them. Moreover, we present direct followed by 2 days off, and then trained for five evidence that HYP induces disaggregation of the additional days. Each trial began when the rats were b-amyloid in a dose and time-dependent manner allowed to swim. A trial continued until the animal in vitro, and that fibrils disaggregates into protofibrils, was put onto the platform for 5 s and returned to its which are the intermediate species in this phenom- cage. Upon completion of the trials the rats were enon in vitro. Our results open the possibility that removed from the maze, dried and returned to its HYP could be of potential use as a new therapeutic cage. Water (50 cm deep) was maintained at 19–211C, agent for AD. for details see Chacon et al.16

Perfusion and fixation Materials and methods Animals were anesthetized with Equitesin (3.5 ml/kg Synthetic peptides and reagents i.p.) and injected with heparin (4 USP/kg, i.p.) before

Ab1À40 and Ab1À42 E22G peptides corresponding to the perfusion. Rats were perfused through the heart with human sequence (Bachem Inc., Torrance, CA, USA, saline (0.9% NaCl) followed by fixation with 4% lot no. T-20964amd and Genemed Synthesis Inc., paraformaldehyde in 0.1 M phosphate-buffer saline South San Francisco, CA, USA) were dissolved in (PBS). Brains were removed from their skulls and dimethyl sulphoxide (DMSO) at a concentration post-fixed in the same fixative for 3 h at room of 15 mg/ml and immediately stored in aliquots at temperature, followed by 20% sucrose in PBS at 41C À201C before assaying. INDENA (Milan, Italy) pro- overnight. After fixation, the brains were coded to vided diciclohexyl ammonium salt of Hyperforin ensure unbiased processing and analysis. The brains

C35H51O14C12H24N (HYP), octahydroperforinate were then cut into 30 mm coronal sections with a lithium salt C35H59O4Li (OHP-Li) and octahydroper- cryostat (Leitzx 1900) at À201C, from Bregma 21 forinate of galanthamine C35H51O14C17H22NO3 (OHP- À1.8 mm to Bregma À4.8 mm. Sections from the Gal). The compounds were dissolved in DMSO at same brain were divided in groups for analysis by the 2 mg/ml and kept protected from light at À201C. Anti- following procedures: Nissl staining (0.3% cresyl GFAP polyclonal antibody was obtained from DAKO violet) and immunohistochemical staining of glial (DAKO, Carpinteria, CA, USA). OX-42 a monoclonal fibrillary acidic protein (GFAP). antibody anti-CD11b protein from the histocompatibility complex class 2 was obtained from Serotec, Immunohistochemical staining Oxford, UK. Free-floating immunohistochemical procedure was performed as previously described.22 Washing and Surgical and injection protocol dilution of immunoreagents was performed with 0.01 M phosphate-buffered saline (PBS) with 0.2% Male Sprague–Dawley rats (300 g; 3 months old) were Triton X-100 (PBS-T) throughout the experiments, anesthetized with Equitesin (3.5 ml/kg i.p.) and and two PBS-T washes were performed between each injected bilaterally into the dorsal hippocampus antibody incubation. Sections were pretreated with

(À3.5 mm AP, 72.0 mm ML and À2.7 mm DV, accord- 0.5% H2O2 for 30 min to reduce endogenous perox- ing to Bregma) stereotaxically with a 10 ml Hamilton idase activity followed by treatment with 3% bovine syringe with 27G stainless-steel. The animals were serum albumin (BSA) at room temperature for 1 h to injected in the hippocampus bilaterally with 3 ml (at avoid non-specific binding. GFAP detection was rate of 0.5 ml/min) of 40 mgAb fibrils formed as performed using rabbit anti-GFAP (1:500) polyclonal described previously.19 Mock animals were injected antibody incubated overnight at 41C. A horseradish with the same volume of vehicle alone (artificial peroxidase (HRP)-conjugated goat anti-rabbit IgG Cerebrospinal Fluid (aCSF)). To observe the effect of second antibody (1:600) was used for GFAP detection, HYP on neurotoxicity and behavior rats were coin- incubated for 1 h at room temperature. The staining jected with amyloid fibrils and 6 mM HYP. At 4 days was developed by incubating for 15 min with 0.6% after the intracerebral injection, rats were trained in diaminobenzidine followed by the addition of H2O2 the Morris Water Maze for 2 weeks and the perfor- (0.01% final concentration) and incubation for 4 min.

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1034 After immunostaining, all sections were mounted on titer plates precoated with (50 mg/ml) polylysine, in gelatin-coated slides, air-dried and dehydrated by 100 ml of Neurobasal medium supplemented with serial rinses in graded ethanol solutions, cleared with B-27 and maintained at 371C with an atmosphere

xylene and cover-slipped with Canada balsam 8% CO2. After 2 h, the culture media was changed. All (Merck, Darmstadt, Germany). experiments were performed on the seventh day of culture using Neurobasal medium without B-27 Immunofluorescence staining supplement. The hippocampal neurons were incu- At 3 weeks after the intrahippocampal injection, the bated for 24 h with Ab in the absence or presence of brain sections were fixed and mounted in slides HYP. Briefly, cells were washed two times with 100 ml washed with PBS and PBS-T, and followed by glycine of PBS and then placed in Neurobasal medium 0.15 M and NaBH4 (10 mg/ml) incubations to diminish (100 ml). To assess cytotoxicity, we used the modified the background. Sections were washed again with 3-[4,5-dimethylthiazol-2-yl]-2,5-dipheniltetrazolium PBS and PBS-T and then followed by treatment with bromide (MTT) assay as an index of mitochondrial 3% BSA at room temperature for 1.5 h to avoid non- activity.27 The MTT assay was carried out by adding specific binding. Activated microglial cells detection 10 ml of a 5 mg/ml MTT solution to each well (100 ml) was performed using OX-42 (1:100) an anti-type and the multiwell plate was incubated for 30 min, 3-complement receptor (CD11B) antibody (Serotec, after which the reaction was stopped with 100 mlofa Oxford, UK) incubated overnight at 41C. After wash- MTT solubilization solution containing 10% Trito´n ing with PBS-T containing 0.5 mg/ml BSA, sections X-100 in acidic isopropanol, 0.1 N HCl (pH 4.8). were incubated for 2 h at room temperature with an Absorption values at 550–650 nm were determined anti-mouse IgG-FITC second antibody (1:500). The using an automatic microtiter plate reader (Uniskan, sections were washed with PBS-T in 0.5% BSA, PBS Helsinki, Finland) and MTT reduction results were and finally water. The slides were cover-slipped with expressed as a percentage of control (untreated) fluorescent mounting medium (DAKO). cells.26

Nissl staining Measurement of intracellular peroxides Mounted sections were defatted in xylene and Relative levels of cellular peroxides were quantified hydrated in ethyl alcohol and water series. Nissl by confocal laser microscope image analysis of staining (cresyl violet) was performed as previously 23 cultured cells loaded with 2,7-dichlorofluorescein described. diacetate (DCF) as detailed previously.28 Briefly, cells Congo red staining grown on glass coverslip pre-coated with polylysine, Mounted sections were defatted in xylene and under the same condition described above. After 7 hydrated in ethyl alcohol and water series. Congo days in culture, the neurons were treated in Neuro- mM b red (CR) staining using ‘alkaline CR methods’ was basal medium without B-27 supplement with 5 A mM performed as previously described.24,25 in the presence or absence of 1 HYP for 12 h. The cells were loaded for 30 min (371C) with 200 mM DCF Image analysis in KRH-glc containing 0.02% pluronic acid. Cover- The ROS fluorescence was imaged with a LSM 5 slips were washed three times and left in KRH- Pascal Zeiss confocal microscope, and analyzed by glucose for 20–40 min until cell fluorescence had Sigman Scan Pro software.26 The fluorescence inten- reach plateau. sity quantification was carried out as the difference between the final fluorescence emitted by the neurons Ab disaggregation assay in the presence of the respective treatment and the Ab peptide (250 mM) in PBS buffer pH 7.2 was stirred background fluorescence of the cells without treat- for 24 h at room temperature to form amyloid fibrils. ment of 10 pictures of three independent experiments. Then, Ab fibrils were incubated with HYP, added at The number of neurons in vivo experiments was final concentrations as indicated in the corresponding calculated with the Image analysis (Image-Pro experiments. Time course studies were designed Express) in Â 40 magnification coronal brain sections to measure relative changes in fibril concentration, stained with Nissl. Digital quantification of positive determined semiquantitatively by (Thioflavine-T) stained areas, sections stained for nitrotyrosine, ThT fluorescence and qualitatively on the transmis- sections stained with CR and Th-S fluorescence sion electron microscope. The samples were stirred positive deposits were measured with Sigma Scan continuously at room temperature for 6 h at Pro software, using three 40 Â pictures of the injection 1300 r.p.m. and sterile aliquots were taken at different site per treatment.23 Results from this analysis time points for analysis. The measurement of the revealed the average of total area of the deposit. relative changes in fibril concentrations was determined at different time intervals by ThT fluorescence Primary neuronal hippocampal cell cultures and cell and transmission electron microscopy. viability assay Hippocampal neurons were dissociated by trypsini- ThT-based fluorometric assay zation as described previously26 and seeded at a Aliquots of Ab peptide at the indicated concentra- density of 120 Â 103 cells per well of 48-well micro- tions were incubated at different times in PBS pH 7.2

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1035 at room temperature. For coincubation experiments, reduced the reactive astrocyte proliferation induced HYP derivatives were added in different concentra- by Ab, decreased the intensity of GFAP staining in the tions. To quantify amyloid formation, the ThT astrocyte soma and abolished completely the enlarge- fluorescence method was used.29 ThT binds specifi- ment of the astrocyte perikaryon (Table 1). The HYP cally to amyloid, such binding produces a shift in its derivatives, OHP-Gal and OHP-Li, also decrease the emission spectrum, and an increase in the fluorescent astrogliosis reaction but not as well as in the case of signal, which is proportional to the amount of HYP (Figure 1g and i, Table 1). amyloid formed.30 Following incubation, Ab alone, We analyzed whether HYP and HYP derivatives or in the presence of the HYP derivatives, 50 mM have positive effect over the neuronal cell loss sodium phosphate buffer pH 6.0 and 0.1 mM ThT in a generated by Ab injection using a specific neuronal final volume of 2 ml were added. Fluorescence was stain (Nissl stain), which did not stain glial cells.22 monitored at excitation 450 nm and emission 485 nm Sections of the dorsal hippocampal region stained using a Shimadzu spectrofluorometer, as described with Nissl showed that the injection of Ab fibrils previously.31 We confirmed that HYP did not quench produced an important neuronal damage (Figure 1d), ThT fluorescence at the used concentrations. and HYP protect against the loss of granular neurons present mainly in the upper leaf of the dentate gyrus Electron microscopy near the injection site (Figure 1f, see the inset). Fresh aliquots of samples were diluted 1:3 in water Animals injected with Ab fibrils and OHP-Gal and and 5 ml were placed on Parlodion/carbon coated OHP-Li showed some protection but the effect was 300-mesh copper grids for 1 min. Excess sample was clearly smaller than HYP (Figure 1h and i, see the removed and 15 ml of 2% aqueous uranyl acetate was inset). The quantification of neuronal loss appears placed onto the grid for 30 s, followed by removal of similar in animals treated with HYP and aCSF excess staining solution with filter paper and air- (Table 1), however, the neuron levels in animals drying. Observations were carried out using Philips treated with Ab decreased almost 70% respect to the Tecnai 12 electron microscope, as described pre- control and this reduction was specific prevented by viously.6 The quantification of amyloid aggregates the concomitant injection of HYP (Table 1) and only and fibrils was made using Sigma Scan Pro software.16 partially affected by HYP derivatives.

Statistical analysis Data from the image analysis were exported to a HYP avoids the formation of CR-positive amyloid Sigma Plot file for statistical analysis. Results were deposits induced by Ab injection expressed as mean7standard error. Student’s t-test With the aim to detect whether the protection against was performed to analyze the image data. astrogliosis and Ab-neurotoxicity generated by HYP and HYP derivatives correlated with a reduction of amyloid deposits, specific amyloid staining was used. Results As observed in Figure 2a and b, control animals HYP protects from the neuropathological changes injected with Ab fibrils showed the presence of strong induced by Ab in vivo birefringence staining with CR in the upper leaf of the To evaluate the potential neuroprotective effect of dentate gyrus, respectively. In contrast, amyloid HYP derivatives against the Ab-neurotoxicity in vivo, deposits found in rats coinjected with Ab and HYP male Sprague–Dawley rats were injected stereotaxi- were clearly reduced, showing only small fragmented cally into the dorsal hippocampus, with 80 mgofAb amyloid deposits by staining with CR (Figure 2c–d). fibrils in the absence or presence of 6 mM of each The image analysis of coronal brain sections analyz- compound.23 In order to study the histopathological ing the amyloid deposit area and birefringence changes observed after the treatments, the astrocytic stained with CR revealed a 90 and 82% reduction response and neuronal damage at the injection site of amyloid burden in animals injected with HYP was analyzed (Figure 1). The presence of reactive compared to controls (Figure 2i), respectively. We astrocytes is an early event in AD pathology and size, also analyzed the HYP effect on the amyloid quantity and density increase in response to brain burden formation using another marker for Ab, the injury.32 The GFAP immunoreactivity was used to Thioflavin-S dye. These results showed than HYP visualize the astrocyte response after the intrahippo- reduced a 79% of the amyloid burden formation (data campal injection. The animals injected with Ab fibrils not shown) consistent with the results observed with showed an important glial reaction (Figure 1c) in CR staining. The HYP derivative OHP-Gal showed contrast to the basal GFAP levels of the control rats some effect on the prevention of the amyloid deposits injected with aCSF (Figure 1a). Animals injected with formation (Figure 2e and f), decreasing the area of amyloid fibrils plus HYP showed a decrease of the amyloid deposits in a 50% and the birefringence in a GFAP-staining comparable to control (Figure 1e). The 65% (Figure 2i). On the other hand, the OHP-Li had injection of Ab fibrils induced astrocyte proliferation, no effect in prevented the amyloid deposits formation an increase in astrocyte density, soma size and GFAP (Figure 2g and h). Therefore, the decrease of the staining in the astrocytes present around the injection neuronal damage and astroglial response induced by site (Table 1). Contrary, coinjection of HYP and Ab Ab fibrils injection of the rats treated with HYP, and

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1036

Figure 1 HYP protects from hippocampal morphological and astrogliosis reaction generated by amyloid fibril injection. (a–i) GFAP immunodetection on coronal sections of hippocampus hypertrophic astrocytes. (b–j) Nissl staining shows cellular layers of hippocampus. (a, b) aCSF; (c, d) amyloid fibril (e, f) coinjection of amyloid fibril and HYP; (g, h) coinjection of amyloid fibril and OHP-Gal; (i, j) coinjection of amyloid fibril and OHP-Li. GFAP staining shows Â 40 magnification and Nissl staining, panels show Â 4 magnification, bar correspond to 250 mm, inset shows Â 40 magnification, bar correspond to 25 mm.

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1037 Table 1 Astrocytic response and neuronal damage generated by Ab are prevent by HYP

Compound Density of reactive GFAP intensity of Size of the Number of neurons astrocytes (GFAP þ astrocytic soma astrocytic soma in the dentate gyrus cell/4 Â 103 mm2) (arbitrary units) (arbitrary units) (cells/4 Â 103 mm2) aCSF 1275 135712 102719 200720 Ab 4078 260710 311717 6078 HYP 2374* 151720* 10679* 162715** OHP-Gal 2676* 180715* 112713* 10579* OHP-Li 2774* 163713* 11678* 101711*

Fibrillar b-amyloid was injected stereotaxically in the hippocampus alone or coinjected with HYP, OHP-Gal or OHP-Li. The quantification of the GFAP inmunostain and Nissl stain of coronal brain sections around the injection site was realized using the Sigma Scan Pro program. *P < 0.05; **P < 0.01. The bold characters emphasize the hyperforin action over the others natural compounds.

Figure 2 HYP decreases the amyloid deposition in a rat model. (a, b) CR stain and birefringence at polarized light of control rats injected with Ab fibril. (c, d) Coinjection of amyloid fibril and HYP; (e, f) coinjection of amyloid fibril and OHP-Gal; (g, h) coinjection of amyloid fibril and OHP-Li. (i) Graph represents a digital quantification of the birefringence intensity of the amyloid deposits (black bars) and amyloid deposits CR-positives area (gray bars). Each image was quantified with the Sigma Scan Pro software. Panels show Â 40 magnification, bar correspond to 25 mm. Each bar represents the average7s.e.m. **P < 0.005 and *P < 0.05 (t-student test). in a minor order the OHP-Gal, correlates with the behavioral impairment caused by Ab injection, decrease of amyloid burdens by this compound. evaluated by the Morris water maze test20 (Figure 3). The analysis of the behavioral performance showed HYP protects from the spatial memory loss induced by that animals injected with Ab had the highest latency Ab deposits values (black circles), in accordance with the neuro- Based on these observations, we then analyzed toxic effects of Ab.16 The animals injected with Ab whether HYP and HYP derivatives prevent the and HYP present significantly lower escape latency

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1038 values than animals injected with Ab (Figure 3a), case for the animals injected with OHP-Gal or OHP-Li indicating that HYP was able to reduce the cognitive (Figure 3c and e). When we analyzed the effect of both impairment on spatial memory performance induced compounds over the escape latency in the animals by Ab. In contrast, animals injected with aCSF or HYP injected with Ab fibrils, we did not observe positive did not show significant difference in their escape effects, therefore OHP-Gal (Figure 3c) and OHP-Li latency values mainly during the second week of (Figure 3e) did not protect from the behavioral training (Figures 3a and 4a), suggesting that HYP impaired induced by Ab. Spatial acuity is a more neither induces spatial learning impairments nor sensitive parameter to determine spatial learning, enhances the cognitive status of the rat is also the representing the probability to find the rat in a

Figure 3 HYP protects the spatial memory acquisition from amyloid fibril-neurotoxicity. Behavioral performances in Morris water maze evaluated by the escape latency showed along the 2 weeks of training, aCSF (blue circles, n = 6), HYP alone (a), OHP-Gal alone (c) and OHP-Li alone (d) (red squares n = 5), amyloid fibrils (black circles, n = 6), and green squares (n =5) amyloid fibril coinjected with HYP (a), OHP-Gal (c) or OHP-Li (d). (c, d and f) Spatial acuity displayed by rats injected as described in (a, b or c). *P < 0.05 (two-way ANOVA test).

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1039

Figure 4 HYP protects from the spatial memory impairment induced by amyloid fibril-injection. (a) Escape latency values of representative days of training of rats injected with aCSF (white bars), HYP alone (gray bars), amyloid fibrils (black bars) and HYP plus amyloid fibrils (dash bars). (b) Average time in the four quadrant (platform quadrant) of all days of training. (c) Swimming speed scores of the different groups, showing no motor alterations by the injection procedure. (d) Representative swimming paths at day 8 of training. Each bar represents the average7s.e.m. *P < 0.05 (two-way ANOVA test). specific region around the hidden platform. In Figure and 10 of training, where we appreciated the most 3b, the relationship between spatial acuity and the significant differences in the spatial learning. The escape latency average for the animals of the different Figure 4a shows that the rats injected with Ab already groups treated with HYP is shown. The graph shows showed high escape latency values contrary to the that animals injected with Ab were localized in a animals injected with Ab and HYP which showed an region of the graph with high escape latency values escape latency values similar to the animal controls and low spatial acuity scores, reflecting the impaired injected with aCSF or HYP. Therefore, the presence of spatial memory of these animals. In contrast, rats HYP prevented the spatial learning impairment injected with Ab and HYP showed low escape latency generated for the amyloid deposits. An additional values (similar to control rats) and spatial acuity parameter analyzed was the permanence in the scores close to those of control animals, supporting Quadrant 4 (Q4), an imaginary annulus where the the protecting effect of HYP on memory impair- platform is located (Figure 4b). As observed in the ment generated by Ab neurotoxicity. On the other graph, animals injected with Ab and HYP showed hand, the animals injected with OHP-Gal (Figure 3d) higher permanence time in Q4, similar to control rats. or OHP-Li (Figure 3f) were localized in a region of In contrast, those rats injected with Ab showed lower the graph with high escape latency values and low time of permanence. To rule out that both procedures spatial acuity scores, similar to the rats injected affect the locomotion performance, the swimming with Ab. speed average (cm/s) was recorded (Figure 4c), no With the aim to further study the positive effects of difference among the studied groups was observed. HYP over the spatial memory loss induced by Ab,we Representative navigation paths at day 8 of train- analyzed others behavior parameters. First, we ana- ing show the notorious impaired spatial learning lyzed the escape latency correspond to the days 8, 9 acquisition of animals injected with Ab fibril in

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1040

Figure 5 HYP prevents the nitrotyrosine proteins formation and reduce the inflammatory process associated to Amyloid deposits in the rat hippocampus. (a) aCSF; (b) HYP alone; (c) Amyloid fibrils; (d) HYP plus amyloid fibrils. (e) OX-42 positives microglia cells were quantifications with the Sigma Scan Pro software. Nitrotyrosine immunofluorescence of coronal brain sections of rats injected with: (e) aCSF; (f) HYP alone; (g) Amyloid fibrils; (h) HYP plus amyloid fibrils. (j) anti- nitrotyrosine staining quantifications performed with the Sigma Scan Pro software. Each bar represents the average7s.e.m. *P < 0.05 (two-way ANOVA test).

comparison with those coinjected with HYP, which associated with the brains of AD patients and is display a navigation pattern similar to control particularly concentrated in the areas of senile plaque animals (Figure 4d). formation.34 We performed immunohistochemical procedures using the antibody OX-42 as a microglial HYP prevents the inflammatory process and protein reaction marker in the injection site.35 The injection tyrosine nitration induced by amyloid of aCSF shows the basal fluorescence (Figure 5a). As only HYP reduces significant the astrocyte reac- Injection of HYP alone shows a weak signal, similar tion, the neuronal death and the spatial memory to the basal levels of rat control (Figure 5b). On the impairment triggered by Ab fibril injection, we other hand, amyloid fibrils injection induced a analyzed more deeply its property. We first analyzed sixfold increase in the activated microglia density its effect over the inflammatory process, since one of (Figure 5c). The amyloid fibril and HYP coinjection the faculty described for St John’s wort is its anti- prevents the activation of microglial cells (Figure 5d) inflammatory properties.33 The reactive microglia is to a level close to animals injected with HYP. Thus,

Figure 6 HYP protects from Ab neurotoxicity in vitro. (a) Hippocampal neuron viability assay using MTT reduction method with increasing HYP concentrations. (b) Hippocampal culture treated with 5 mM final product from Ab aggregation assay in the absence or presence of increasing HYP concentrations (0.01–1.0 mM) for 24 h. Hippocampal neurons were treated and then loaded with 2,3-DCF probe during 30 min and observed at confocal microscope in vivo to evaluate the ROS production. (c) Confocal pictures of neurons without treatment showing the background fluorescence. Cells were incubated with 5 mM Ab (d), with 1 mM HYP (e) or with 5 mM Ab and 1 mM HYP (f) for 12 h, (g) Quantification of the fluorescence for the neurons treated with Ab, HYP and Ab with HYP. Each bar represents the fluorescence intensity average 7 the s.e.m. (h) Electron microscope showed the material corresponding to Ab-oligomers. The picture shows the amylospheroids with a diameter of 9–12 nm. (i) Neuronal viability was measured by MTT method. Hippocampal culture treated with HYP and Ab-oligomers. The soluble Ab fraction was added to the neurons at the concentrations of 10 mM.*P < 0.01 and **P < 0.001.

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1041 HYP reduces the appearance of active microglial We next evaluated the HYP effect on the generation induced by Ab, preventing the inflammatory process of radical oxygen species (ROS) induced by the associated to the Ab neurotoxicity. injection of Ab. Evidence described showed that

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1042 HYP have antioxidant properties,36 so to determine when the cells were coincubated with HYP. These whether the neuroprotective effect of HYP against Ab results suggest that HYP prevents the Ab-neurotoxi- peptide neurotoxicity is mediated by an antioxidant city in vitro through the reduction of the oxidative pathway, we performed a nitrotyrosine immunod damage generated by Ab and also in vivo as we etection as an oxidative marker in the injection site showed in Figure 6h. of the animals, because it is well known that protein Recent studies suggest that the toxicity of Ab and tyrosine nitration is an important marker of oxidative other amyloidogenic proteins lies not in the insoluble stress.37 The anti-nitrotyrosine staining was carried fibrils that accumulate but rather in the soluble out 3 weeks after the intrahippocampal injection. oligomeric intermediates.42 These soluble oligomers The basal level of fluorescence is showed in the include spherical structures 3–20 nm in diameter that Figure 5e. Injection of HYP alone showed a weak appear to represent strings of the spherical particles43 signal (Figure 5f). On the other hand, amyloid fibril Ab is produced at discrete sites within cells as a injection induced a clear signal for protein tyrosine monomer,44 and it rapidly enters into equilibrium nitration (Figure 5g). The section of animals with with dimers and trimers,45 a process similar to that coinjection of amyloid and HYP presents low levels of described for synthetic Ab.46 It appears that a fraction nitrotyrosine-positive signal, similar to the control of these oligomers is highly stable (via either strong levels (Figure 5h). These data shows that the forma- hydrophobic interactions or covalent cross-links), and tion of nitrated-proteins is a consequence of the Ab a portion of these is subsequently secreted.45 These injection, which is abolished by the presence of the secreted oligomers of Ab can interact with neurons, HYP (Figure 5h). altering their electrical activity and normal physio- logy.45 In view of these considerations, we generated 47 HYP protected from Ab-neurotoxicity in vitro Ab oligomers from the arctic mutant Ab1À42 E22G To study the more detailing the HYP effects, we used (Figure 6h). Then, the toxicity in hippocampal the hippocampal neuron cultures. The hippocampal neurons was measured. Ab oligomers (10 mM) pro- cells were treated with increasing HYP concen- duced almost 80% of neuronal cell death in 24 h, trations to evaluate the possible drug toxicity (Figure however, HYP was able to prevent its toxicity to 6a), using the MTT reduction assay, an indicator of values similar to the control neurons (Figure 6i), neuronal death. At low HYP concentrations the suggesting that HYP not only protects from Ab fibrils, neuronal viability was not altered (0.01–1.0 mM), but it was also capable to prevent neurotoxicity however, at very high concentrations (100 mM) neuro- induced by the Ab oligomers. nal cell death is clearly detected. Then, the neurons were treated with increasing HYP concentrations in HYP disassembles the preformed amyloid fibrils the presence of Ab fibrils (Figure 6b). Previously, we To understand the effect of HYP on the amyloidogenic had determined that 5 mM Ab generated around 50% process, we designed experiments in which amyloid of neuronal death after 24 h incubation.38 Under these fibrils were incubated with increasing concentrations conditions, the amyloid aggregates caused around of HYP and the amount of remaining amyloid was 47% of the neuronal cell death (Figure 6b), however, measured at defined intervals of time. A concentra- the Ab-neurotoxicity in the presence of increasing tion- and time-dependence effect of HYP on the HYP concentrations was significantly lower than Ab disassembly of the fibers followed by the ThT assay in the absence of the drug, therefore, HYP protected was observed (Figure 7a). Previous to the experi- against the Ab-neurotoxicity in hippocampal neurons ments, the amyloid fibrils were formed at room (Figure 6b). In addition, the reactive oxygen species temperature under continuous stirring for 24 h. We (ROS) have been implicated in many aspects of aging confirmed the presence of fibrils in the aggregates by and neurodegenerative diseases such as AD.39–41 The electron microscopy before the addition of HYP. The HYP effect on the intracellular production of ROS fluorescence of ThT was almost unchanged during the induced by the Ab on hippocampal neurons was incubation of 250 mM Ab fibrils at room temperature evaluated using the fluorescence of the 2,3-DCF and continuous stirring for 6 h. Higher HYP concen- probe.40 Neurons were treated for 12 h with Ab, HYP trations were able to decrease the fluorescence and and Ab in the presence of HYP, and then neurons they disassemble the amyloid fibrils almost instanta- were observed under the confocal microscope (Figure neously. The final remnant percentage of amyloid 6c–f). Figure 6c shows the basal levels of fluorescence aggregated was only 20% (Figure 7a). Lower HYP in the neurons without treatment. The neurons concentrations showed a gradual decrease in fluores- treated with Ab, had an important increase in the cence. The ThT fluorescence decreased immediately ROS fluorescence levels (Figure 6d) respect to the after addition of HYP in a concentration-dependent control neurons. Neurons treated with HYP alone, manner. only show a slight increase in the fluorescence Electron microscopic analyses of the time course (Figure 6e). However, HYP reduces severely the experiments present above were carried out. Aliquots intracellular ROS levels induced by Ab treatment of 250 mM fibrils incubated alone (Figure 7b–e) or in (Figure 6f). The fluorescence intensity levels were the presence of HYP (46 nM) (Figure 7f–i) were quantified in the Figure 6g, the graph showed a processed for electron microscopy at 0, 1, 3 and 6 h significantly decreases of peroxides generated by Ab after the addition of HYP. The starting material was

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1043

Figure 7 Effect of HYP on the depolymerization of amyloid fibrils. Kinetics of 250 mM Ab fibril disaggregation was followed by ThT fluorescence under constant stirring (a); symbols represent the different HYP concentration points in the assay: 0 (K); 0.046 (J); 0.46 (.); 4.6 (,); 46 (’); and 100 mM (&). (a) Electron microscopy. Aliquots took at different times (0, 1, 3 and 6 h). Pictures (b–e) correspond to 250 mM amyloid fibrils alone and (f–i) to amyloid fibrils with 46 nM HYP. The arrowheads indicate the amorphous material. The arrows in (i) indicate the protofibrils (bar = 225 nm). (j) Shows a fourfold increase in the magnification of (i), the arrows indicate the protofibrillar material (bar represents 900 nm). The histogram (k) shows the quantification of the fibrillar material area from samples obtained in three separate experiments. Each point represents the mean values 7s.e.m. with *P < 0.005 and **P < 0.001. characterized by large dense masses formed by throughout the experiment in the absence of HYP numerous sheared amyloid fibrils as depicted in (Figure 7b–e). When the starting material was mixed Figure 7b and f; they did not change consistently with HYP (46 nM) (Figure 7f–i) the large aggregates of

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1044 fibrils were reduced markedly and small amorphous in the upper leaf of the dentate gyrus, which was aggregates (see arrowheads in Figure 7h) were prevented by HYP, these results correlated with the occasionally observed along with the appearance of protection on the spatial memory loss induced by Ab smaller and denser aggregates of 5 nm diameter (see also16). The neuronal cell loss produced by Ab protofibrils (see arrows in Figure 7i and j).48,49 The injection correlates with the behavioural impairment. relative abundance of amyloid fibrils was quantified In this study, we have established that the coinjection (Figure 7k). The black bars show the amyloid fibrils of Ab and HYP resulted in an almost complete without HYP and the gray bars show the amyloid protection from behavioural disturbances, evaluated fibrils plus HYP at 0, 1, 3 and 6 h of incubation. HYP by the Morris water maze test, contrary to what was cause a significant decrease of approximately 20% in observed in animals injected with OHP-Gal or OHP- the amyloid fibrils (Figure 7k). Our data clearly Li. Also, when we analyzed the amyloid deposits suggests that under the conditions tested, HYP formed by injection of Ab and HYP or HYP deriva- induces a significant reversion of the amyloid tives, we observed that only animals coinjected with aggregation process as shown by a decrease in the Ab and HYP showed a reduction of amyloid deposits amount of amyloid fibrils and a replacement of by CR stain (Figure 2c) and ThS fluorescence (data not amyloid fibrils by amorphous material and proto- shown). Our observations indicated only a partial fibrils. decrease of hippocampal amyloid deposits induced by Ab injection, suggesting that the amyloid deposits Discussion are not the only cause of the behavioral disturbances observed here, similar results were obtained pre- Our data shows that HYP protects from behavioural viously using b-sheet breaker peptides.16 and morphological disturbances generated by Ab in We do not know the precise mechanism of action of vivo. We also demonstrate that two others HYP- HYP, but the data showed in this work strongly derivatives had no anti-amyloidogenic effect, but suggest that a reduction of amyloid deposition showed positive effects over the astrogliosis and observed after HYP treatment could be the result of neuronal damage induced by Ab injection, suggesting amyloid fibril depolymerization. Other natural deri- that chemical modifications of HYP can be made with vatives, like the Curcumin of the yellow Curry spice, the aim to improve its stability and bioavailability for reported anti-flammatory and antioxidant effects.51,52 its therapeutic use. In this context, we decided to In addition, Curcumin was able to reverse the further study the effect of HYP in order to elucidate a amyloid-induced neuropathology, along with its possible mechanism. capacity to block amyloid polymerization.53 Ginkgo One of the features present in AD brain analyzed in biloba another natural product is largely used as a the present study is the presence of reactive astrocytes positive memory modulator54 and it has been reported associated with senile plaques, which apparently that shows prevention of age-related spatial memory plays a role in the neurodegenerative events of the deficits in a transgenic mouse model of AD.55 Also, disease.49 In fact, neuronal processes are attracted to certain ingredients of green tea modulate APP plaques in response to trophic signals emanating from cleavage sufficiently to reduce the Ab levels and reactive plaque-associated astrocytes.50 In this con- amyloid plaques in AD mouse models.56 Recently, it text, we analyzed the astroglial cells present after has been reported than Resveratrol, a polyphenol HYP and HYP derivatives injection. We showed that from grapes and red wine, reduce the intracellular HYP clearly diminished the astrogliosis around the and secreted amyloid levels.57 The precise mecha- amyloid deposit, this effect was lower in animals nisms by which all these compounds exert their effect injected with OHP-Gal or OHP-Li. Different degrees of remain to be elucidated. astrocyte reactivity were observed by GFAP immuno- HYP protects from the neurotoxicity induced by Ab staining, after the animals were injected with Ab in hippocampal neuron cultures, moreover, we fibrils alone, they presented a strong glial reaction showed that HYP protects against oligomer-mediated with the presence of classical appearance of hyper- neurotoxicity. This result is important, since Ab trophic astrocytes.16 Thus, at least until 18 days after oligomers (including stable dimers and trimers) can coinjection of HYP plus Ab, the decrease of the interfere with LTP and thus synaptic plasticity45,47 reactive astrocytes was correlated with the protection generating the neurotoxicity observed in triple trans- of the granular cell layer loss. On the other hand, the genic mice,58 and ROS are involved in Ab neurotoxi- microglial activation was analyzed, with the idea to city.26,41,59 Therefore, our results suggest that the associate it to a proinflammatory process.50 We found protection from the Ab neurotoxicity was partially that HYP was able to reduce the activation of mediated by the ROS reduction. microglial cells that surround the Ab deposits, and It has been observed that protofibrils seem to be an this is a clear beneficial effect considering that important intermediate in Ab fibrilogenesis.60 Our reactive microglia are present around amyloid pla- results support the hypothesis that HYP causes the ques in AD brains.34,35 Also, we determined that HYP disassembly of fibrils to amorphous material and prevented the morphological damage caused by protofibrils. Therefore, we assumed that amyloid amyloid fibrils in vivo. The Nissl staining showed deposits are unstable in the presence of HYP. The that the amyloid deposits triggered neuronal cell loss HYP exponential concentration dependence for fibril

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1045 depolymerization suggests that there is a rather width In this context, the antidepressive activity of HYP has concentration range requirement and the critic con- been related to the inhibition of the uptake of centration for its effect on amyloid depolymerization neurotransmitters in a nonselective manner by inter- is lower than 46 nM. Of practical point of view is the ference with intracellular sodium homeostasis.17 observation that this HYP concentration is lower than Also, it was observed that HYP has a facilitator action the plasma levels of the drug when it is administrated on hippocampus acetylcholine released.62 Moreover, orally to humans.61 The disassembly of preformed there is evidence that HYP increase the soluble fibrils, followed by ThT fluorescence was shown to be cleavage of APP, reducing the secreted Ab levels.63 a time- and HYP concentration-dependent process, There are numerous natural derivatives used to suggesting that the HYP anti-amyloidogenic action prevent several pathologies. In the case of AD, may be of therapeutic value for AD patients. curcumin and G. biloba extract EGb 761 are the more Previous studies in behavioral paradigms suggest studied natural compounds. Table 2 shows a sum- that H. perforatum extracts show an antidepressive mary and a comparison of the neurological and the activity, where HYP plays a key role in those effects. anti-amyloidogenic properties between these com-

Table 2 Effects of natural derivatives on AD neurodegenerative and amyloidogenic changes

Hyperforin Curcumin Ginkgo

St John’s wort is an NA NA antidepressant approved by the commission E of the German Federal Institute for Drugs and Medical Devices64

Antidepressant by inhibition of Antidepressant by serotonin ND serotonin uptake by free and noradrenaline increase intracellular Na þ elevation in levels in rat depression mouse brain synaptosomes17 model65

General neurological Weak inhibitor of monoamine Inhibition of monoamine Inhibition of monoamine effects oxidase-A and -B in vitro66 oxidase activity in mouse oxidase activity in mouse brain53 corpus striatum67

Stimulation of Ach release in ND Antagonists of GABAAR, in vivo62 Xenopus leavis oocytes68

Inhibition of ROS production Inhibition of ROS production Inhibition of ROS production in cell culture and in vivo69a in human cell line culture52 in neuroblastoma cell line culture70

Anti inflammatory effect71a Anti-inflammatory effect51 ND

Improve amyloid cognitive ND Prevent spatial memory deficits in vivoa deficits in Alzheimertransgenic model in vivo55

Decrease of b-amyloid Decrease of b-amyloid in vivo53 Decrease of b-amyloid in vivo55 depositions in vivoa

Anti-amyloidogenic Enhancement of secretory APP ND Inhibition of b-amyloid effets proteolytic precessing in PC12 production in aging rats72 cell line culture63

Ab protection in hippocampal Ab protection in Ab protection in hippocampal cell culturea neuroblastoma cell culture53 cell culture73

Inhibition of Ab Inhibition of Ab Inhibition of Ab aggregation in polymerization and polymerization and neuroblastoma cell line74 disaggregates Ab in vivoa disaggregates Ab in vivo53

ND: non determined. NA: non applicable. aPresent manuscript.

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1046 pounds and hyperforin. The neurological effects from the St John Wort, may present advantages for evaluated are related to AD: antidepressant proper- people in risk of AD, and eventually for its treatment. ties, effect over neurotransmission and anti-oxidant and anti-inflammatory properties. On the other hand, we also evaluated the effects of these compounds over Acknowledgments features related to AD: the cognitive alterations We thank Joselyn Mauna and Rocio Artigas for their produced by Ab, the Ab burden stability, the amyloid help with the glial and microglial studies. This production, the cell-toxicity generated by Ab, the research was supported by grants from FONDAP amyloid polymerization and the stability of amyloid (No 13980001) and the Millennium Institute for fibrils. Fundamental and Basic Biology (MIFAB). Concerning the neurological effects of hyperforin, curcumin and ginkgo, it has been described that St John’s wort may be effective for depression. In Europe References this herb is widely prescribed and a number of studies 1 Salmon DP, Bondi MW. The neuropsychology of Alzheimer’s have been conducted that support the treatment disease. In: Terry RD, Katzman R and Bick KL (eds). Alzheimer’s efficacy of certain St John’s wort extracts. An over- Disease, 2nd edn. Raven Press: New York, 1999, pp 39–56. view of 23 clinical studies in Europe indicated that 2 Morgan C, Colombres M, Nunez MT, Inestrosa NC. Structure and this herb might be useful in cases of mild to moderate function of amyloid in Alzheimer’s disease. Prog Neurobiol 2004; 74: 323–349. depression. The Commission E of the German Federal 3 Yankner BA. Mechanisms of neuronal degeneration in Alzheimer’s Institute for Drugs and Medical Devices has approved disease. Neuron 1996; 16: 921–932. St John’s wort as a nonprescription medicine for 4 Mattson MP. Pathways towards and away from Alzheimer’s ‘psycho-vegetative instability, depressive moods, disease. Nature 2004; 430: 631–639. 64 5 Cruz L, Urbanc B, Buldyrev SV, Christie R, Gomez-Isla T, Havlin S anxiety and/or nervous disorder’. On the other hand, et al. Aggregation and disaggregation of senile plaques in neither curcumin nor ginkgo have been approved for Alzheimer disease. Proc Natl Acad Sci USA 1997; 94: 7612–7616. antidepressant treatment, nevertheless in a rat de- 6 Morgan C, Buguen˜ o M, Garrido J, Inestrosa NC. Laminin affects pression model, curcumin showed antidepressant polymerization, depolymerization and neurotoxicity of Ab effects, increasing the serotonin and noradrenaline peptide. Peptides 2002; 23: 1229–1240. 7 Selkoe DJ. Alzheimer’s disease: genes, proteins, and therapy. levels in vivo. Another important issue mentioned in Physiol Rev 2001; 81: 741–766. Table 2 is the inflammatory properties of these 8 Bronfman FC, Garrido J, Alvarez A, Morgan C, Inestrosa NC. compounds. Hyperforin and curcumin have anti- Laminin inhibits amyloid-b-peptide fibrillation. Neurosci Lett inflammatory effects, such activity has not been 1996; 218: 201–203. 9 Pappolla M, Bozner P, Soto C, Shao H, Robakis NK, Zagorski M described for ginkgo. Also, the three compounds et al. Inhibition of Alzheimer b-fibrillogenesis by melatonin. J Biol had been described as anti-oxidant, because they Chem 1998; 273: 7185–7188. reduce the production of ROS. 10 Ono K, Hasegawa K, Yoshiike Y, Takashima A, Yamada M, Naiki Table 2 also showed an analysis of the specific anti- H. Nordihydroguaiaretic acid potently breaks down pre-formed amyloidogenic properties, where the principal differ- Alzheimer’s b-amyloid fibrils in vitro. J Neurochem 2002; 81: 434–440. ences were found between hyperforin and curcumin. 11 Ono K, Yoshiike Y, Takashima A, Hasegawa K, Naiki H, Yamada The paper of Yang et al., showed that curcumin M. Potent anti-amyloidogenic and fibril-destabilizing effects inhibits the amyloid b-fibrils polymerization and of polyphenols in vitro: implications for the prevention reduces the amyloid in vivo, no study on the effects and therapeutics of Alzheimer’s disease. J Neurochem 2003; 87: 53 172–181. of curcumin on cognitive effects was reported. In the 12 Solomon B, Koppel R, Frankel D, Hanan-Aharon E. Disaggregation present work, we demonstrate that HYP inhibits the of Alzheimer b-amyloid by site-directed mAb. Proc Natl Acad Sci amyloid b-fibrils polymerization and reduces the USA 1997; 94: 4109–4112. amyloid deposits in vivo, similar to curcumin, 13 Fraser PE, Nguyen JT, McLachlan DR, Abraham CR, Kirschner DA. however, we also showed the beneficial HYP effects 1-antichymotrypsin binding to Alzheimer Ab peptides is sequence specific and induces fibril disaggregation in vitro. J Neurochem over the spatial memory loss induced by Ab accumu- 1993; 61: 298–305. lation. On the other hand, hyperforin, as ginkgo, has 14 Luo Y, Smith JV, Paramasivam V, Burdick A, Curry KJ, Buford JP an effect over the APP processing. Ginkgo treatment et al. Inhibition of amyloid-b-aggregation and caspase-3 activation by lowering circulating free cholesterol decreased by the Ginkgo biloba extract EGb761. Proc Natl Acad Sci USA b 72 2002; 99: 12197–12202. APP processing and inhibits the production of A . 15 Kiuchi Y, Isobe Y, Fukushima K. Type IV collagen prevents Hyperforin is an activator of secretory processing of amyloid b-protein fibril formation. Life Sci 2002; 70: 1555–1564. APP in part by a mechanism of action that affects 16 Chacon MA, Barria MI, Soto C, Inestrosa NC. sheet breaker peptide intracellular pH.63 In conclusion, HYP: (1) decreases prevents Abeta-induced spatial memory impairments with partial amyloid deposit formation, Ab-induced neuropatho- reduction of amyloid deposits. Mol Psychiatry 2004; 9: 953–961. 17 Singer A, Wonnemann M, Muller WE. Hyperforin, a major logical changes and behavioral impairments in a rat antidepressant constituent of St John’s Wort, inhibits serotonin model of amyloidosis; (2) prevents Ab-induced uptake by elevating free intracellular Na þ . J Pharmacol Exp Ther neurotoxicity in hippocampal neurons, avoiding the 1999; 290: 1363–1368. increase in reactive oxidative species associated with 18 Klusa V, Germane S, Noldner M, Chatterjee SS. Hypericum extract and Hyperforin: memory-enhancing properties in rodents. amyloid toxicity; (3) disaggregates amyloid fibrils in a Pharmacopsychiatry 2001; 34: S61–S69. dose and time-dependent manner, suggesting that in 19 De Ferrari GV, Chacon MA, Barria MI, Garrido JL, Godoy JA, addition to its anti-deppressive action, HYP obtained Olivares G et al. Activation of Wnt signaling rescues neuro-

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1047 degeneration and behavioral impairments induced by b-amyloid 41 Miranda S, Opazo C, Larrondo LF, Munoz FJ, Ruiz F et al. The role fibrils. Mol Psychiatry 2003; 8: 195–208. of oxidative stress in the toxicity induced by amyloid b-peptide in 20 Morris RGM. Developments of a water maze procedure for Alzheimer’s disease. Prog Neurobiol 2000; 62: 633–648. studying spatial learning in the rat. J Neurosci Methods 1984; 11: 42 Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimer’s 47–60. disease: progress and problems on the road to therapeutics. 21 Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates, Science 2002; 297: 353–356. 2nd edn. Academic Press: New York, 1986. 43 Hoshi M, Sato M, Matsumoto S, Noguchi A, Yasutake K et al. 22 Coˆte´ SL, Ribeiro-Da-Silva A, Cuello AC. Immunocytochemistry II. Spherical aggregates of b-amyloid (amylospheroid) show high John Wiley and Sons: New York, NY, 1993. neurotoxicity and activate tau protein kinase I/glycogen synthase 23 Reyes AE, Chacon MA, Dinamarca MC, Cerpa W, Morgan C, kinase-3b. Proc Nat Acad Sci 2003; 100: 6370–6375. Inestrosa NC. Acetylcholinesterase-Ab complexes are more toxic 44 Selkoe DJ. The cell biology of beta-amyloid precursor protein than Ab fibrils in rat hippocampus: effect on rat b-amyloid and presenilin in Alzheimer’s disease. Trends Cell Biol 1998; 8: aggregation, laminin expression, reactive astrocytosis, and neuro- 447–453. nal cell loss. Am J Pathol 2004; 164: 2163–2174. 45 Walsh D, Klyubin I, Fadeeva J, Cullen W, Anwyl R, Wolfe M et al. 24 Puchtler H, Sweat F, LeVine M. On the binding of Congo red by Naturally secreted oligomers of amyloid b protein potently amyloid. J Histochem Cytochem 1961; 10: 355–364. inhibit hippocampal long-term potentiation in vivo. Nature 2002; 25 Elghetany MT, Saleem A. Methods for staining amyloid in tissues: 416: 53–539. a review. Stain Technol 1988; 63: 201–212. 46 Nilsberth C, Westlind-Danielsson A, Eckman CB, Condron MM, 26 Quintanilla RA, Munoz FJ, Metcalfe MJ, Hitschfeld M, Olivares G, Axelman K, Forsell C et al. The ‘Arctic’ APP mutation (E693G) Godoy JA et al. Trolox and 17b-estradiol protect against amyloid causes Alzheimer’s disease by enhanced Ab protofibril formation. b-peptide neurotoxicity by a mechanism that involves modulation Nat Neurosci 2001; 4: 887–893. of the Wnt signaling pathway. JBiolChem2005; 280: 11615–11625. 47 Wang HW, Pasternak JF, Kuo H, Ristic H, Lambert MP et al. Soluble 27 Mun˜oz FJ, Opazo C, Gil-Go´mez G, Tapia G, Ferna´ndez V, Valverde oligomers of b amyloid (1–42) inhibit long-term potentiation but MA et al. Vitamin E but not 17b-Estradiol protect against vascular not long-term depression in rat dentate gyrus. Brain Res 1999; 924: toxicity induced by b-amyloid wild-type and the Dutch amyloid 133–140. variant. J Neurosci 2002; 22: 3081–3089. 48 Harper JD, Wong SS, Lieber CM, Lansbury Jr PT. Assembly of Ab 28 Pong K, Doctrow SR, Huffman K, Adinolfi CA, Baudry M. amyloid protofibrils: an in vitro model for a possible early event in Attenuation of staurosporine-induced apoptosis, oxidative stress, Alzheimer’s disease. Biochemistry 1999; 38: 8972–8980. and mitochondrial dysfunction by synthetic superoxide dismutase 49 Nielsen EH, Nybo M, Svehag SE. Electron microscopy of and catalase mimetics, in cultured cortical neurons. Exp Neurol prefibrillar structures and amyloid fibrils. Methods Enzymol 2003; 17: 84–97. 1999; 309: 491–496. 29 LeVine H. Thioflavin T interaction with synthetic Alzheimer’s 50 Cummings BJ, Su JH, Cotman CW. Neuritic involvement within disease b-amyloid peptides: detection of amyloid aggregation in bFGF immunopositive plaques of Alzheimer’s disease. Exp Neurol solution. Protein Sci 1993; 2: 404–410. 1993; 124: 315–325. 30 Naiki H, Higuchi K, Nakakuki K, Takeda T. Kinetic analysis of 51 Satoskar RR, Shah SJ, Shenoy SG. Evaluation of anti-inflammatory amyloid fibril polymerization in vitro. Lab Invest 1991; 65: 104– property of curcumin (diferuloyl methane) in patients with 110. postoperative inflammation. Int J Clin Pharmacol Ther Toxicol 31 Inestrosa NC, Alvarez A, Perez CA, Moreno RD, Vicente M, Linker 1986; 24: 651–654. C et al. Acetylcholinesterase accelerates assembly of amyloid-b- 52 Fujisawa S, Atsumi T, Ishihara M, Kadoma Y. Cytotoxicity, ROS- peptides into Alzheimer’s fibrils: possible role of the peripheral generation activity and radical-scavenging activity of curcumin site of the enzyme. Neuron 1996; 16: 881–891. and related compounds. Anticancer Res 2004; 24: 563–569. 32 Pike CJ, Cummings BJ, Cotman CW. Early association of reactive 53 Yang F, Lim GP, Begum AN, Ubeda OJ, Simmons MR, Ambegaokar astrocytes with senile plaques in Alzheimer’s disease. Exp Neurol SS et al. Curcumin inhibits formation of amyloid-b oligomers and 1995; 132: 172–179. fibrils, binds plaques, and reduces amyloid in vivo. J Biol Chem 33 Zhou C, Tabb MM, Sadatrafiei A, Grun F, Sun A, Blumberg B. 2005; 280: 5892–5901. Hyperforin, the active component of St John’s wort, induces IL-8 54 van Dongen M, van Rossum E, Kessels A, Sielhorst H, Knipschild expression in human intestinal epithelial cells via a MAPK- P. Ginkgo for elderly people with dementia and age-associated dependent, NF-kappaB-independent pathway. J Clin Immunol memory impairment: a randomized clinical trial. J Clin Epidemiol 2004; 24: 623–636. 2003; 56: 367–376. 34 McGeer PL, Itagaki S, Tago H, McGeer EG. Reactive microglia in 55 Stackman RW, Eckenstein F, Frei B, Kulhanek D, Nowlin J, Quinn patients with senile dementia of the Alzheimer type are positive JF. Prevention of age-related spatial memory deficits in a for the histocompatibility glycoprotein HLA-DR. Neurosci Lett transgenic mouse model of Alzheimer’s disease by chronic Ginkgo 1987; 79: 195–200. biloba treatment. Exp Neurol 2003; 184: 510–520. 35 Rozemuller JM, Eikelenboom P, Pals S, Stam FC. Microglialcells 56 Rezai-Zadeh K, Shytle D, Sun N, Mori T, Hou H, Jeanniton D around amyloid plaques in Alzheimer’s disease express leucocyte et al. Green tea epigallocatechin-3-gallate (EGCG) modulates adhesion molecules of the LFA-1 family. Neurosci Lett 1989; 101: amyloid precursor protein cleavage and reduces cerebral amyloi- 282–292. dosis in Alzheimer transgenic mice. J Neurosci 2005; 25: 36 Hunt EJ, Lester CE, Lester EA, Tackett RL. Effect of St John’s Wort 8807–8814. on free radical production. Life Sci 2001; 69: 181–190. 57 Marambaud P, Zhao H, Davies P. Resveratrol promotes clearance of 37 Beckman JS, Koppenol WH. Nitric oxide, superoxide, and Alzheimer’s disease amyloid-b peptides. J Biol Chem 2005; 280: peroxynitrite: the good, the bad, and ugly. Am J Physiol 1996; 37377–37382. 271: C1424–C1437. 58 Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, Kayed R 38 Alvarez AR, Godoy JA, Mullendorff K, Olivares GH, Bronfman M, et al. Triple-transgenic model of Alzheimer’s disease with plaques Inestrosa NC. Wnt-3a overcomes b-amyloid toxicity in rat and tangles: intracellular Ab and synaptic dysfunction. Neuron hippocampal neurons. Exp Cell Res 2004; 297: 186–196. 2003; 39: 409–421. 39 Barger SW, Horster D, Furukawa K, Gogdman Y, Krieglstein J, 59 Behl C, Davis JB, Lesley R, Schubert D. Hydrogen peroxide Mattson MP. Tumor necrosis factors a and b protect neurons mediates amyloid-b-protein toxicity. Cell 1994; 77: 817–827. against amyloid b-peptide toxicity: evidence for involvement of a 60 Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB. kappa B-binding factor and attenuation of peroxide and Ca2 þ Amyloid b-protein fibrillogenesis. Detection of a protofibrillar accumulation. Proc Natl Acad Sci USA 1995; 92: 9328–9332. intermediate. J Biol Chem 1997; 272: 22364–22372. 40 Butterfield DA. Amyloid b-peptide [1–42]-associated free radical- 61 Cui Y, Gurley B, Ang C, Leakey J. Determination of hyperforin in induced oxidative stress and neurodegeneration in Alzheimer’s human plasma using solid-phase extraction and high-performance disease brain: mechanisms and consequences. Curr Med Chem liquid chromatography with ultraviolet detection. J Chromatogr B 2003; 10: 2651–2659. 2002; 780: 129–135.

Molecular Psychiatry Hyperforin affects Ab deposits in vitro and protects from its neurotoxicity in vivo MC Dinamarca et al 1048 62 Kiewert C, Buchholzer ML, Hartmann J, Chatterjee SS, at recombinant a1b2g2L GABAA receptors. Eur J Pharmacol 2004; Klein J. Stimulation of hippocampal acetylcholine release by 494: 131–138. hyperforin, a constituent of St John’s Wort. Neurosci Lett 2004; 69 Albert D, Zundorf I, Dingermann T, Muller WE, Steinhilber D, 364: 195–198. Werz O. Hyperforin is a dual inhibitor of cyclooxygenase-1 and 63 Froestl B, Steiner B, Muller WE. Enhancement of proteolytic 5-lipoxygenase. Biochem Pharmacol 2002; 64: 1767–1775. processing of the b-amyloid precursor protein by hyperforin. 70 Smith JV, Luo Y. Elevation of oxidative free radicals in Alzheimer’s Biochem Pharmacol 2003; 66: 2177–2184. disease models can be attenuated by Ginkgo biloba extract EGb 64 Cott JM, Rosenthal N, Blumenthal M. St John’s wort and major 761. J Alzheimers Dis 2003; 5: 287–300. depression. J Am Med Assoc 2001; 286: 42–45. 71 Feisst C, Werz O. Suppression of receptor-mediated Ca2 þ 65 Xu Y, Ku BS, Yao HY, Lin YH, Ma X et al. Antidepressant effects of mobilization and functional leukocyte responses by hyperforin. curcumin in the forced swim test and olfactory bulbectomy Biochem Pharmacol 2004; 67: 1531–1539. models of depression in rats. Pharmacol Biochem Behav 2005; 72 Yao ZX, Han Z, Drieu K, Papadopoulos V. Ginkgo biloba 82: 200–206. extract (Egb 761) inhibits b-amyloid production by lowering free 66 Butterweck V. Mechanism of action of St John’s wort in cholesterol levels. J Nutr Biochem 2004; 15: 749–756. depression: what is known? CNS Drugs 2003; 17: 539–562. 73 Bastianetto S, Ramassamy C, Dore S, Christen Y, Poirier J, Quirion 67 Rojas P, Rojas C, Ebadi M, Montes S, Monroy-Noyola A, R. The Ginkgo biloba extract (EGb 761) protects hippocampal Serrano-Garcia N. EGb761 pretreatment reduces monoamine neurons against cell death induced by b-amyloid. Eur J Neurosci oxidase activity in mouse corpus striatum during 1-methyl-4- 2000; 12: 1882–1890. phenylpyridinium neurotoxicity. Neurochem Res 2004; 29: 74 Luo Y, Smith JV, Paramasivam V, Burdick A, Curry KJ et al. 1417–1423. Inhibition of amyloid-b aggregation and caspase-3 activation by 68 Huang SH, Duke RK, Chebib M, Sasaki K, Wada K, Johnston GA. the Ginkgo biloba extract EGb761. Proc Natl Acad Sci USA 2002; Ginkgolides, diterpene trilactones of Ginkgo biloba, as antagonists 99: 12197–12202.

Molecular Psychiatry