Curcumin and a hemi-analogue with improved blood–brain barrier permeability protect against amyloid-beta toxicity in Caenorhabditis elegans via SKN-1/Nrf activation

Elaine Hui-Chien Lee, Sherlyn Sheau-Chin Lim, Kah-Hay Yuen and Chong-Yew Lee

School of Pharmaceutical Sciences, Universiti Sains Malaysia, Penang, Malaysia

Keywords

amyloid-beta; blood–brain barrier permeability; Caenorhabditis elegans; curcumin analogues; SKN-1/Nrf

Correspondence

Chong-Yew Lee, School of Pharmaceutical Sciences, Universiti Sains Malaysia, 11800 Penang, Malaysia.

Received June 26, 2018

Accepted September 17, 2018 doi: 10.1111/jphp.13052

Abstract

Objectives This study aims to investigate the blood–brain barrier (BBB) perme- ability of curcumin analogues with shortened linkers and their ability to protect against amyloid-beta toxicity in a whole organism model.

Method Four curcumin analogues were synthesized. These analogues and cur- cumin were evaluated for their BBB permeability in the parallel artificial mem- brane permeability assay. The transgenic Caenorhabditis elegans GMC101 that expresses human Ab_1–42 was treated with the compounds to evaluate their ability to delay Ab-induced paralysis. Expression of skn-1 mRNA was examined on nematodes treated with selected efficacious compounds. In vitro Ab aggregation in the presence of the compounds was performed.

Key findings The four analogues showed improved BBB permeability vs cur- cumin in the PAMPA with the hemi-analogue C4 having the highest permeability coefficient. At 100 lM, analogues C1 and C4 as well as curcumin significantly prolonged the survival of the nematodes protecting against Ab toxicity. However, only curcumin and C4 showed protection at lower concentrations. skn-1 mRNA was significantly elevated in nematodes treated with curcumin and C4 indicating SKN-1/Nrf activation as a possible mode of action.Conclusions Analogue C4 provides a new lead for the development of a cur- cumin-based compound for protection against Ab toxicity with an improved BBB permeability.

Alzheimer’s disease is a common cause of dementia in the ageing population. It is a neurodegenerative condition marked by progressive cognitive impairment leading to overt disability, morbidity and death. At the cellular level, the hallmarks of the disorder are the deposition of toxic amyloid beta (Ab) oligomers, increased oxidative stress, the development of neurofibrillary tangles and glial activation, multiple events that interact to cause neuronal cell dysfunc- tion and death.^[1] Due to the complex interplay between these multiple causative processes, the pathogenesis of Alz- heimer’s disease is not precisely understood. However, the negative or toxic effects of Ab in the intra- and intercellular environments of neurons and glial cells are evident in numerous studies.^[2–4]

Disease-modifying agents that address these multiple fac- tors of Alzheimer’s disease and extend beyond symptomaticcontrol are being intensively studied and sought after. A notable example is curcumin, a naturally occurring pheno- lic diarylheptanoid constituent of the spice turmeric (Cur- cuma longa). Curcumin has demonstrated remarkable properties in vitro such as the ability to inhibit Ab aggrega- tion or cause its disaggregation^[5,6] and to reduce Ab- induced toxicity in various cell culture preparations.^[7–9] Despite these promising attributes, little has been done to address its ability to penetrate the blood–brain barrier (BBB) and be available in efficacious concentrations in the brain. Sensitive multiphoton microscopy has shown evi- dence of curcumin BBB penetration but its presence in the mouse brain was weak and required the accumulation of curcumin via repeated intravenous administrations.^[10] Cheng et al.^[11] using the Mardin-Darby canine kidney cell monolayer expressing human multidrug resistance 1 (MDR1) as their BBB model reported low curcumin perme-ability rate (1.8 9 10^—6 cm/s) even when the compoundwas formulated as nanoparticles. A recent study^[12] reported that curcumin conjugated to a nanoliposome and a transac- tivator of transcription (TAT) peptide showed a higher per- meability coefficient (2.2 9 10^—5 cm/s) across the human cerebral microvascular (hCMEC/D3) cell monolayer but it remains to be tested whether conjugation with such large moieties would retain curcumin bioactivity.

We hypothesized that replacing the heptadienedione lin- ker of curcumin with a single carbonyl cross-conjugated dienone thus shortening the molecule, reducing its molecu- lar weight and increasing its lipophilicity would result in better BBB permeability. Shortening the linker may also improve its stability as this portion of the molecule has been known to contribute to its degradation and low bioavailability.^[13] We were also interested to determine whether such modified curcumin analogues would still retain or improve the protective properties of curcumin against Ab. To this end, four analogues were prepared and evaluated in a transgenic Caenorhabditis elegans that expresses Ab_1–42 and manifests paralysis as a result of Ab toxicity to its bodywall muscle cells.^[14] BBB permeability of the analogues was determined using the parallel artificial membrane permeability assay for blood–brain barrier (PAMPA-BBB).

The protective effect of curcumin has been attributed to its Ab binding property.^[5,6] In another perspective, Michael acceptor motifs (alpha, beta-unsaturated carbonyl) are embedded within the structures of curcumin and the analogues in this study. Michael acceptor moiety-bearing compounds are known to activate the Kelch-like ECH-associated protein 1-Nuclear Factor (erythroid- derived 2)-like 2 (Keap1-Nrf2) pathway, a master regulator of cytoprotective responses to endogenous and exogenous electrophilic and oxidative stressors.^[15,16] Accordingly, to elucidate the possible mechanism(s) underlying the protec- tion against Ab by the compounds, two mechanisms were explored in this study: their Ab aggregation inhibitory potential, and their ability to induce the SKN-1 pathway, the nematode orthologous version of the mammalian Nrf2 pathway.

Materials and Methods

General experimental details for synthesisReagents (synthetic grade or better) were obtained from Sigma-Aldrich Chemical Company Inc. (Singapore, Singa- pore), Acros Organics (Geel, Belgium) and Biobasic Canada Inc. (Toronto, Canada) and used without further purification. ¹H and ¹³C NMR spectra were recorded on a Bruker Avance III 500 MHz spectrometer (Bruker Biospin AG, Fallanden, Switzerland). Chemical shifts (d) were reported in parts per millions (ppm) and the couplingconstants in Hz. Electrospray Ionisation Mass Spectra (ESI- MS) were obtained using a Waters 2695 Separations Mod- ule with Micromass Quattro Micro mass spectrometer (Waters Corporation, Milford, MA, USA) and accurate m/z for molecular ions were reported. Melting points of the compounds were obtained using a digital melting point apparatus Stuart SMP10 (Staffordshire, UK) and were uncorrected. All synthesized compounds were purified by flash column chromatography on silica gel 60 (230–400 mesh). All reactions were monitored by thin layer chro- matography on silica gel 60 F₂₅₄ plates (Merck, Darmstadt, Germany). The purities of compounds C1, C2 and C3 (higher than 95%) were confirmed through HPLC on a Agilent Tech 1120 Compact HPLC system with the use of aPurospher STAR RP-18e column (4.6 9 100 mm, 5 lm) at 1 ml/min on two different solvent systems.

Synthesis of curcumin analogues C1, C2 and C3

To a solution of 4-hydroxybenzaldehyde (2 mol) in glacial acetic acid was added a ketone (1 mol) such as acetone (for C1), cyclopentanone (for C2) and cyclohexanone (for C3). The resultant mixture was saturated with anhydrous HCl and stirred at room temperature for 2 h. After standing for 48 h, the mixture was poured into cold water and the sus- pension was extracted with ethyl acetate (50 ml 9 3). The collected organic extract was dried with anhydrous Na₂SO₄,evaporated in vacuo to give a solid residue and purified by column chromatography on silica gel using hexane: ethyl acetate as eluting solvents. The compounds were recrystal- lized from ethyl acetate or ethanol.,5-bis(4-hydroxy-3-methoxyphenyl)-1,4- pentadien-3-one (C1)

Greenish yellow powder. C₁₉H₁₈O₅. Yield: 54.6%. Melting point: 134–138 °C. ¹H NMR (DMSO-d₆, 500 MHz): d 3.85 (s, 6H, –OCH₃ 9 2), 6.84 (d, J = 8 Hz, 2H, arom), 7.16

(d, J = 15 Hz, 2H, CH=CH–CO 9 2), 7.37 (d, J = 1.5 Hz, 2H, arom), 7.65 (d, J = 15 Hz, 2H, arom-CH = CH 9 2),

9.71 (s, 2H, –OH 9 2). ¹³C NMR (DMSO-d₆, 125 MHz):

d 188.5, 149.9, 148.4, 143.2, 126.8, 123.8, 123.4, 116.1,

111.9, 111.8, 56.17; ESI-MS m/z: 327.24 [M + H]^+. HPLC

(MeOH : H₂O = 55 : 45) t_R (min): 5.593. Purity: 99.84%. HPLC (ACN : H₂O = 40 : 60) t_R (min): 4.697. Purity: 99.36%.

2,5-bis (4 hydroxy-3-methoxybenzylidene) cyclopentanone (C2)

Yellow powder, Yield: 87.0%,  m.p.  196–198 °C;  ¹H NMR (DMSO-d₆): d 3.07 (4H, s, –CH₂CH₂–), 3.84 (s,

CO 9 2),  7.17  (d,  J = 9 Hz,  2H,   arom-CH=CH 9 2), 7.25 (s, 2H, arom), 7.36 (s, 2H, arom), 9.72 (br s, 2H, OH 9 2). ¹³C NMR (DMSO-d₆, 125 MHz):  d 195.3, 149.0,

148.2, 135.2, 133.3, 127.6, 125.2, 116.3, 115.0, 56.0,

26.3; ESI-MS  m/z:  353.29  [M + H]⁺.  HPLC  (MeOH : H₂O = 65 : 35)  t_R  (min):  3.973.  Purity:  99.73%.  HPLC

(ACN : H₂O = 50 : 50) t_R (min): 3.097. Purity: 99.80%.

2,6-bis (4-hydroxy-3-methoxy benzylidene) cyclohexanone (C3)

Yellow powder, Yield: 73.3%, m.p. 174–176 °C;¹H NMR (DMSO-d₆):  d 1.71–1.75 (m, 2H), 2.9 (t,  4H), 3.81 (s, 6H,   OCH₃ 9 2),   6.85   (d,   J = 9 Hz,   2H,   CH=CH–

CO 9 2),  7.04  (d,  J = 9 Hz,  2H,  arom-CH=CH 9 2),

7.12 (s, 2H, arom),  7.57  (s,  2H,  arom),  9.57  (s,  2H, OH 9 2).   ¹³C   NMR   (DMSO-d₆,   125 MHz):   d 189.0,

148.3, 147.9, 136.6, 134.0, 127.4, 124.7, 116.0, 115.2,

56.1, 28.4, 23.03. ESI-MS m/z: 367.27 [M + H]⁺. HPLC (MeOH : H₂O = 65 : 35)  t_R  (min):  4.870.   Purity: 98.84%.  HPLC  (ACN : H₂O = 50 : 50)  t_R  (min):  4.177.

Purity: 98.84%.

Caenorhabditis elegans strain and culture conditions

The transgenic strain GMC101 (dvIs100 (unc-54p::A- beta-1-42::unc-54   3⁰-UTR  + mtl-2p::GFP).   mtl-2p::GFP) and the food source Escherichia coli (E. coli) strain OP50 were obtained from the Caenorhabditis Genetics Center (University of   Minnesota,   Minnesota,   MN, USA). The GMC101 strain constitutively produces a full-length human Ab_1–42 in bodywall muscle cells that aggregates in vivo and upshifting L4 or young adults worms from 16 to 25 °C induces paralysis.^[14] The nematodes were routinely propagated at 16 °C on solid nematode growth medium (NGM) seeded with spots of OP50 as the food source.

Media preparation

Curcumin and C4 were purchased from Sigma-Aldrich Inc. (Singapore, Singapore). The curcumin analogues (C1, C2 and C3) were synthesized as described in the previous sec- tion. Stock solutions of all compounds were made with 100% DMSO. The final concentration of DMSO did not exceed 1% in the media. All compounds were added directly into the molten NGM to a final concentration of 100, 10 and 1 lM; NGM with 1% of DMSO served as vehi- cle control. All media were spotted with 600 ll of OP50 food source.

Ab-induced paralysis assay

Paralysis assays were performed as described previously.^[14] In brief, egg-synchronized population of transgenic GMC101 strain was prepared by allowing gravid adult nematodes to lay eggs on the NGM plate containing either a vehicle or test compound at the desired concentrations for 2 h at 16 °C. Adult nematodes were then removed and plates with eggs were allowed to develop into L3 larvae stage at 16 °C. At 68 h post-egg lay, the test plates were temperature upshifted to 25 °C to induce transgene expres- sion. Scoring for paralysis was performed at 2 h intervals beginning from 24 h post-temperature upshift. Nematodes were scored as paralysed if they failed to respond to a touch-provoked movement with platinum wire. At least three independent experiments were performed for each test compounds. Approximately 55–100 worms were grown on each assay plate. Survival curves were generated using Prism 5 (GraphPad Software Inc., La Jolla, CA, US). The delays in paralysis were assessed by comparing the median survival time at which 50% nematodes were paralysed, rep- resented as median 95% confident interval, between nematodes fed with vehicle or compounds. Significant dif- ferences between groups were assessed using the log-rank (Mantel-Cox) analysis. A P value <0.05 was considered sta- tistically significant.

Parallel artificial membrane permeability assay for blood–brain barrier

Porcine brain lipid (PBL) was purchased from Avanti Polar Lipids (Alabaster, AL, USA). Dodecane, caffeine and donepezil were sourced from Sigma-Aldrich (St. Louis, MO, USA) while DMSO was from BioBasic Canada Inc. The 96-well donor plate microplate (PVDF membrane, pore size 0.45 lM) and the acceptor micro- plate were both from Merck Millipore Bioscience (Bed- ford, MA, USA). The method by Di et al.^[17] was followed. Stock solutions of each reference and test com- pound were prepared in DMSO at 1 mM and then diluted with phosphate buffered saline (PBS) to obtain the donor solution with the final concentration of 10 lM. 300 ll of the compounds were added to the donor wells. The lipid membrane of the PAMPA-BBB model was pre- pared by dissolving 20 mg of PBL in 1 ml dodecane. The filter membrane of the acceptor microplate was coated with 4 ll of the PBL solution and then filled with 300 ll of PBS. Thereafter, acceptor plate was gently placed into the donor plate and incubated at room temperature for 18 h. Upon completion of incubation, the plates were separated and the concentration of the compounds in both donor and acceptor wells were quantified by the UVabsorbances of the compounds (curcumin and C2, k_max 420 nm; C1 and C3, k_max 395 nm; C4, k_max 340 nm). Permeability rates (P_e) and membrane retention (R) were calculated using the following equation:

Real-time quantitative PCR

For quantitative gene expression analysis, total RNA sam- ples, derived from three biological replicates of each treat- ment were pooled. qRT-PCR experiments were performedusing  KAPA  SYBR^® FAST  quantitative  PCR  (qPCR)  kit Biosystems, Wilmington, MA, USA) on EcoTM   Real- Time PCR system (Illumina Inc, San Diego, CA, USA).

primer and 100 ng of cDNA as the template. Each pooled

reverse (5⁰ ? 3⁰): GTCGACGAATCTTGCGAATC; cdc-42:

forward     (5⁰ ? 3⁰):      CTGCTGGACAGGAAGATTACG,

reverse (5⁰ ? 3⁰): CTCGGACATTCTCGAATGAAG. Data

Synthetic full-length Ab_1–40 peptides (Sigma-Aldrich) were pretreated with HFIP (1,1,1,3,3,3-Hexafluoro-2-propanol) to ensure it is in a monomer state, subsequently evaporated under stream of nitrogen gas and redissolved in dimethyl- sulfoxide (DMSO) to keep as stock. The peptide stock solu- tions were later diluted in PBS (pH 7.4) to the final concentration of 10 lM used in experiments. Thioflavin T (ThT) assay was performed as follows: 10 lM of Ab_1–40 peptide incubated with 10 lM test compounds or plain PBS was mixed with 20 lM ThT at 37 °C for 24 h. The ThT flu- orescence intensity was measured on LS 45 Luminescence Spectrometer (Perkin-Elmer, Waltham, MA, USA) with excitation and emission wavelengths at 440 and 484 nm, respectively. Each assay was run in triplicates in three inde- pendent experiments.

Isolation of RNA and cDNA synthesis

Egg synchronized transgenic GMC101 worms were exposed to media containing vehicle, curcumin or C4 (10 and 100 lM) at 16 °C for 68 h and were upshifted to 25 °C for transgene induction. At 15 h post-temperature upshift, about 8000 worms were harvested for total RNA extraction. Total RNA was extracted with Nucleospin^R RNA Plus kit (Macherey-Nagel, Duren, Germany) according to manufac- turer’s instructions. Purity and concentration of total RNA were determined by using Bioanalyser (UV spectropho- tometer Q3000, Quawell). RNA samples with the purity above 1.8 and 2.0 were qualified for further tests. First strand cDNA was synthesized from 2 lg RNA using HiSen- scriptTM   RH(—) cDNA Synthesis kit (Intron Biotechnology Inc., Gyeonggi-do, Korea).

Results

Preparation of curcumin analogues C1, C2, C3 and C4Three analogues were prepared by acid-catalysed aldol condensation of vanillin with various ketones (acetone, cyclopentanone and cyclohexanone; Figure 1) in moder- ate yields and high purity. The NMR, mass spectra and HPLC chromatograms of the compounds are given in the Appendix S1. They were considered analogues with a shortened linker connecting the two aromatic rings. The substituents on the aromatic rings (the 4-hydroxyl and 3-methoxyl) were maintained in these compounds with the focus being variation to the linker. The linker was acyclic as in C1 or cyclic (C2 and C3) imparting the compounds with variation in size and lipophilicity. The commercially obtained C4 was considered a ‘hemi-analo- gue’ of curcumin as it constituted a symmetric half of the curcumin molecule.

In vitro passive BBB permeability of curcumin and analogues

We first evaluated the passive permeability of the com- pounds in the PAMPA-BBB model. PAMPA evaluates the ability of a compound to permeate the BBB by measuring the rate (P_e) at which a compound transverses a brain lipid-coated filter membrane that mimics the BBB from a donor compartment to an acceptor compartment filled

Protective effect of curcumin analogues in Ab-induced paralysis in GMC101 Caenorhabditis elegans

We next evaluated the ability of curcumin and the ana- logues to delay Ab-induced paralysis in the transgenic

C. elegans GMC101 that expresses the Ab_1–42 peptide in its bodywall cells. Initially, the nematodes were treated with each compound at the highest possible concentration (with solubility and toxicity being the limiting factor) of 100 lM. The nematodes were exposed to the compounds from the egg stage and throughout the assay (Figure 2a). The com- pounds-treated nematodes showed normal development to adulthood; thus, the compounds were deemed non-toxic. Upon temperature upshift from 16 to 25 °C, Ab was expressed in the worms and scoring of nematode survival was performed post 24 h when worm paralysis became pro- nounced. Curcumin, C1 and C4 at 100 lM significantly delayed paralysis of the worms (Figure 2b). The median survival times of nematodes treated with these compounds were also increased significantly (P < 0.001, Table 2). In Blood–brain barrier permeability of curcumin analogues

  Experimental protocol for the GMC101 paralysis assay.

(b)  Effects of curcumin and analogues at 100 lM in the Ab-induced paralysis assay. Time refers to hours after temperature up-shift. Ab expression was induced by upshifting temperature from 16 to 25 °C. Worms began to show paralysis 24 h post-temperature upshift and were scored at 2 h intervals. An asterisk (*) indicates significant differ- ences between vehicle and treatment (**P < 0.01; ***P < 0.001, paired log rank survival test). Data are expressed as percentage of non-paralysed worms derived from at least three independent experi- ments of >100 worms in each experiment. The plots shown are repre- sentatives of three experiments.

Efficacious compounds curcumin and C4 were evaluated for their ability to inhibit Ab aggregation when Ab mono- mers were incubated in the presence of the compounds in vitro. Ab aggregation was monitored for 24 h using the fluorescent Thioflavin T as indicator of aggregation levels of the peptide. At the 24th h post-incubation, curcumin at 10 lM showed a slight inhibition (14.8%) but this reduc- tion in Ab aggregation was not significant (Figure 4a). C4 also did not inhibit Ab aggregation significantly.

Upregulation of skn-1 by curcumin and C4

Experiments using GMC101 were repeated using the same protocol (Figure 1b) but without scoring for paralysed worms. The RNA of the nematodes under curcumin and C4 treatment (10 and 100 lM) was extracted and probed for skn-1 mRNA levels at the 15th h post-temperature upshift when Ab was expressed and the number of nema- todes experiencing paralysis was pronounced. RT-PCR analysis showed skn-1 levels were significantly increased in worms treated with curcumin at 10 lM (P < 0.01) and C4 at 10 lM (P < 0.05) and 100 lM (P < 0.001) compared to the vehicle control (Figure 4b).

Discussion

In vitro, curcumin has consistently shown protection against Ab-induced toxicity in various cell lines as well as the ability to rescue long-term potentiation in rat hip- pocampus impaired by Ab.^[6,8,20,21] Its in vivo performance however has been hampered by inherently suboptimal physicochemical properties such as low chemical and meta- bolic stability, and inconsistent BBB permeability.^[22] In this study, we explored the possibility of improving cur- cumin BBB permeability via curcumin analogues with shortened linkers while retaining curcumin protective activity. Such dibenzylidene ketone analogues as C1 and C2 have been shown to retain the Ab binding property of cur-Median survival time is the period at which 50% worms survive. Com-parisons were done between vehicle and treatment. ^aMedian survival time (50% non-paralysed worms). ^bP < 0.001 compared to vehicle. ^cP < 0.01 compared to vehicle. ^dOverall log-rank test.contrast, cyclic linker-bearing analogues C2 and C3 failed to show protective effect at 100 lM. At lower concentra- tions (1 and 10 lM), curcumin and C4 retained their pro- tective effect in a concentration-dependent manner while C1 did not show protection (Figure 3). Thus, curcumin and C4 were deemed as the more efficacious compounds.in a physiological context as in the whole organism C. ele- gans has hitherto not been tested.

Using the PAMPA model, we demonstrated for the first time that the curcumin analogues C1, C2, C3 and C4 showed improved passive BBB permeability compared to curcumin. It would seem that the higher lipophilicities (vs curcumin) of C1, C2 and C3 may have contributed to their improved BBB permeability with the exception of the smal- ler hemi-analogue C4 which possessed the lowest lipophilicity (cLogP 1.23) showing the highest permeability coefficient of the four compounds. This indicates that Effects of curcumin (a), C1 (b) and C4 (c) at 1–100 lM on the Ab-induced paralysis in GMC101. An asterisk (*) indicates significant dif- ferences between vehicle and treatment (***P < 0.001, paired log rank survival test). Data are expressed as percentage of non-paralysed worms from at least three independent assays of >100 worms in each experiment. The plots shown are representative of three experiments.molecular size in the present instance played a larger role towards BBB permeability compared to lipophilicity. A caveat, however, in the present study is that these curcumin analogues merely showed better passive permeability which does not take into account active transport protein- mediated permeation by which a cell-based BBB model would offer.^[25,26] In light of recent findings of interactions of curcumin derivatives with efflux proteins,^[27,28] the pos- sibility that the BBB permeation of these compounds may also involve active transporters requires further confirma- tion with a cell-based BBB assay.

Having found these analogues to have improved BBB permeability, we evaluated their ability to confer protection against Ab in the GMC101 nematode that manifests paraly- sis as a result of Ab toxicity. At the highest screening con- centration (100 lM), curcumin showed protective effect in delaying the onset of Ab-induced paralysis and its effect was dose-dependent thus confirming on previous tissue/ cell-based studies of curcumin protective activity. The abil- ity of curcumin to interact with Ab and inhibit its aggrega- tion has been linked to its ability to attenuate Ab toxicity. However, analogues C2 and C3 which were reported asstrong Ab binders comparable to curcumin^[23,24] failed to show any protection in delaying paralysis onset of the worms. Similarly, C1 which showed protection at 100 lM, was inefficacious at lower concentrations. This draws paral- lel to a study by Park et al.^[8] which showed that the 7-car- bon linker of curcumin was necessary for its protection against Ab-induced death in PC-12 cells and that shorter linkers (as that of C1, C2 and C3) would result in inactivity. Reinke and Gestwicki^[29] proposed that an Ab-binding compound requires (i) a linker of optimal length and flexi- bility and (ii) two aromatic rings attached to both ends of the linker. In clear contrast, C4 (being half the structure of curcumin) lacked such structural requirements but showed significant and dose-responsive paralysis delay and a better median survival time than curcumin. Taken together, this suggests that the protective effect shown by the curcumin and C4 may not be attributed to Ab binding mechanism or necessitating the full structural motif of curcumin.

To confirm on this conjecture, curcumin and C4 were evaluated for their ability to inhibit Ab aggregation in vitro. Curcumin did not lower Ab aggregation significantly despite its reputed ability to bind Ab possibly because the concentration (10 lM) used was insufficient for this prop- erty to be apparent. The hemi-analogue C4 failed to inhibit Ab aggregation as expected for its lack of the structural requirements of an Ab-binding compound. However, working on the premise of curcumin and C4 having a Michael acceptor motif within their structures and that Michael acceptor compounds are potent Nrf2 activators, the skn-1 (the nematode ortholog of Nrf2) gene was mea- sured on the compound-treated worms. We showed that indeed skn-1 mRNA levels of worms treated with curcumin and C4 were elevated significantly. Thus, this indicates that the ability of the compounds to modulate oxidative stress response mechanism via SKN-1 may likely be their mode of protection against Ab toxicity. A means of Ab toxicity is the ability of the peptide to induce oxidative stress and modify cellular redox state.^[4] For example, the Met-35 resi- due of Ab_1–42 peptides located in the lipid bilayer has beenshown to undergo one-electron oxidation to generate sulfu- ranyl free radical leading to oxidative modification of cellu- lar proteins.^[30] Induction of SKN-1 by C4 or curcumin would thus lead to upregulation of downstream effectors such as glutathione-S-transferases to remove the oxidative stressor caused by the presence of Ab. Similar studies but without the presence of Ab showed the ability of plumba- gin^[31] and allyl isothiocyanate^[32] (both electrophilic Nrf2 inducers) to prolong the lifespan and increase oxidative stress resistance of wild type C. elegans via SKN-1 activa- tion. The notion that curcumin and C4 alleviate Ab toxicity through oxidative stress modulation in the present finding requires further investigation into the possible oxidative modification(s) within the nematode when exposed to Ab and how these compounds may be able to modify these changes.

In conclusion, a simplified curcumin analogue C4 has been identified in the present study with improved passive BBB permeability and the ability to protect against Ab toxi- city. C4 represents a potential lead for the further optimiza- tion of similar curcumin-like compounds with therapeutic potentials and promising BBB targeting property in central neurodegenerative diseases such as Alzheimer’s disease.

Declarations

Conflict of interest

The Authors declare that they have no conflict of interest to disclose.

Acknowledgements

The authors gratefully acknowledged support by the Universiti Sains Malaysia Research University (RU) grant (1001/PFARMASI/811266) provided by Universiti Sains Malaysia. EHCL was supported by the MyBrain Scholarship from the Ministry of Education of Malaysia.

1. Crews L, Masliah E. The molecular mechanisms of Curcumin analog C1 neurodegeneration in Alzheimer’s disease. Hum Mol Genet 2010; 19: R12–R20.
2. Benilova I et al. The toxic Ab oligo-

mer and Alzheimer’s disease: an emperor in need of new clothes. Nat Neurosci 2012; 15: 349–357.

1. Doens D, Fernandez PL. Microglia receptors and their implications in ponse to amyloid b for Alzheimer’s disease. J Neuroinflammation 2014; 11:48.

1. Butterfield DA et al. Amyloid b-pep- tide (1-42)-induced oxidative stress in Alzheimer disease: importance in dis- ease pathogenesis and progression. Antioxid Redox Signal 2013; 19: 823– 835.
2. Fang Y et al. Curcumin inhibits for- mation of amyloid b oligomers and fibrils, binds plaques, and reducesamyloid in vivo. J Biol Chem 2005; 280: 5892–5901.

1. Thapa A et al. Curcumin attenuates amyloid-b aggregate toxicity and modulates amyloid-b aggregation pathway. ACS Chem Neurosci 2016; 7: 56–68.
2. Huang HC et al. Curcumin-mediated neuroprotection against amyloid-b- induced mitochondrial dysfunction involves the inhibition of GSK-3b. J Alzheimers Dis 2012; 32: 981–996.