FEBS Letters 585 (2011) 2424–2430

journal homepage: www.FEBSLetters .org

Natural polyphenols inhibit different steps of the process of transthyretin(TTR) amyloid fibril formation

Nelson Ferreira, Maria João Saraiva, Maria Rosário Almeida ⇑Grupo de Neurobiologia Molecular, IBMC – Instituto de Biologia Molecular e Celular, Rua do Campo Alegre, 823, 4150-180 Porto, PortugalDepartamento de Biologia Molecular, ICBAS – Instituto de Ciências Biomédicas Abel Salazar, Universidade do Porto, Largo Prof. Abel Salazar, 2, 4099-003 Porto, Portugal

a r t i c l e i n f o a b s t r a c t

Article history:Received 27 May 2011Accepted 24 June 2011Available online 2 July 2011

Edited by Jesus Avila

Keywords:TransthyretinAmyloid inhibitorFibril formation(�)-Epigallocatechin-3-gallateCurcumin

0014-5793/$36.00 � 2011 Federation of European Biodoi:10.1016/j.febslet.2011.06.030

⇑ Corresponding author at: Grupo de Neurobiologiade Biologia Molecular e Celular, Rua do Campo APortugal. Fax: +351 22 6099157.

E-mail addresses: ngomes@ibmc.up.pt (N. Fer(M.J. Saraiva), ralmeida@ibmc.up.pt (M.R. Almeida).

Several natural polyphenols with potent inhibitory effects on amyloid fibril formation have beenreported. Herein, we studied modulation of transthyretin (TTR) fibrillogenesis by selected polyphe-nols. We demonstrate that both curcumin and nordihydroguaiaretic acid (NDGA) bind to TTR andstabilize the TTR tetramer. However, while NDGA slightly reduced TTR aggregation, curcuminstrongly suppressed TTR amyloid fibril formation by generating small ‘‘off-pathway’’ oligomersand EGCG maintained most of the protein in a non-aggregated soluble form. This indicates alterna-tive mechanisms of action supported by the occurrence of different non-toxic intermediates. More-over, EGCG and curcumin efficiently disaggregated pre-formed TTR amyloid fibrils. Our studies,together with the safe toxicological profile of these phytochemicals may guide a novel pharmaco-therapy for TTR-related amyloidosis targeting different steps in fibrillogenesis.

Structured summary of protein interactions:TTR binds to TTR by comigration in gel electrophoresis (View interaction)TTR binds to TTR by dynamic light scattering (View interaction)TTR binds to TTR by electron microscopy (View interaction)

� 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction monomer polypeptide chain induces subtle structural alterations

Transthyretin (TTR) variants are related with different forms ofhereditary amyloidoses and in particular with Familial AmyloidoticPolyneuropathy (FAP). Among more than one hundred single ami-no acid variants reported, the most frequent one is TTR V30M, inwhich valine (V) 30 is substituted by a methionine (M) [1]. Thisvariant presents high tendency to aggregate and to deposit in par-ticular in peripheral nerves, but also in heart, eye and in fewercases, carpal tunnel or leptomeninges revealing clinical and molec-ular heterogeneity [2]. TTR L55P, a variant with a substitution of aproline for leucine, is highly amyloidogenic and is associated withthe most aggressive form of the disease [3]. Likewise, TTR Y78F is avery aggregate-prone mutant, namely under physiological condi-tions [4], while TTR T119M, with a substitution of methionine forthreonine, is a non-amyloidogenic and non-pathogenic TTR variantwith protective properties [5].

The TTR molecule is a tetramer of identical subunits of 127 aminoacid residues each. The occurrence of amino acid substitutions in the

chemical Societies. Published by E

Molecular, IBMC – Institutolegre, 823, 4150-180 Porto,

reira), mjsaraiv@ibmc.up.pt

that decrease the stability of the tetramer and increase the amyloi-dogenic potential of the protein. Thus, it is widely accepted thatmolecular mechanisms underlying TTR amyloid fibril formation in-volve destabilization of the TTR tetramer that dissociates into non-native monomers with low conformational stability and aggregatesto form amyloid fibrils [6,7].

Considering this mechanism, several small compounds havebeen suggested for TTR amyloidosis therapy, acting as TTR stabiliz-ers, impairing amyloid fibril formation or disrupting TTR fibrils.Among those, are non-steroidal anti-inflammatory drugs (NSAIDs)and derivatives or natural compounds [8,9]. However some of thepreviously suggested compounds present low binding affinity orlow selectivity of binding to TTR ex vivo and, in some cases, severeadverse side effects associated with their administration in vivo[10–12]. Recently, our group reported that (�)-epigallocatechingallate (EGCG), the most abundant catechin in green tea, inhibitsTTR aggregation in vitro [13] raising the possibility of using thispolyphenol or analogs to treat FAP.

Curcumin, nordihydroguaiaretic acid (NDGA), rosmarinic acid,caffeic acid and scyllo-inositol are natural occurring polyphenolspresenting structural similarities with thyroxine (T4), a physiologicTTR ligand. Most of them have been suggested for the preventionand treatment of neurodegenerative disorders such as Alzheimer’s

lsevier B.V. All rights reserved.

N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430 2425

and Parkinson’s diseases through inhibition of aggregation or dis-ruption of pre-existing fibrils of amyloid-b (Ab) or a-synuclein(aS) [14,15]. In the present work, we evaluated the ability of thesephytochemicals to bind TTR and modulate different steps of theprocess of TTR amyloid fibril formation, in comparison with EGCG.

2. Materials and methods

2.1. Reagents

(�)-Epigallocatechin gallate (EGCG) and curcumin were pur-chased from Cayman Chemicals (Ann Arbor, MI, USA) and nord-ihydroguaiaretic acid (NDGA), rosmarinic acid and caffeic acidfrom Sigma–Aldrich (St. Louis, MO, USA).

2.2. Samples

Whole blood from five heterozygotic carriers of TTR V30M andfrom five control individuals was collected in the presence of EDTAand centrifuged for plasma separation. All donors gave informedconsent.

Recombinant wild-type TTR (TTR WT) and TTR variants, namelyTTR V30M, TTR L55P and TTR Y78F, were produced in a bacterialexpression system and purified as previously described [12].

2.3. T4 competition assays

Displacement of T4 from TTR, either from plasma or recombi-nant, was assayed qualitatively by incubation of whole plasma(5 ll) or recombinant protein (10 lg) with [125I]T4 (specific radio-activity 1250 lCi/lg; Perkin–Elmer, MA, USA) in the presence ofdifferent phytochemicals (10� molar excess relative to TTR tetra-mer) solubilized in DMSO. Subsequently, proteins were separatedby PAGE, revealed by phosphor imaging and quantified using theImageQuant program version 5.1. [12]. Competition of polyphenolswith T4 for the binding to TTR was quantitatively assayed by gel fil-tration as previously described [12]. Briefly, TTR WT (30 nM) inTris/NaCl/EDTA buffer was incubated with a trace amount of[125I]T4 and with increasing concentrations (0–10 lM) of competi-tors. [125I]T4 bound to TTR was separated from free T4 through gelfiltration column. All samples were run in duplicate and the datawas analyzed using GraphPad Prism 2.0.

2.4. Isoelectric focusing (IEF) in semi-denaturing conditions

Thirty microliters of human plasma from five TTR V30M hetero-zygotic carriers and five control individuals were incubated in thepresence of the phytochemicals (5 ll of a 10 mM solution) at 37 �Cfor 1 h. The samples were subjected to PAGE, and the TTR gel bandwas excised and applied to a semi-denaturing (4 M urea) pH 4–6.5gradient IEF gel [12]. Proteins were stained with Coomassie Blue.The gels were scanned and subjected to densitometry analysisusing the ImageQuant program.

2.5. Dynamic light scattering (DLS) and transmission electronmicroscopy (TEM)

A solution of TTR Y78F (2 mg/ml) in PBS with or without 10�molar excess of different polyphenols was filtered through 0.2 lmAnotop syringe filters (Whatman, England) and incubated at 37 �Cfor 4, 6, 12 and 16 days. DLS measurements were performed at25 �C in a Malvern Zetasizer Nano ZS (Malvern, Worcestershire,UK). Each sample was measured tree times; average distributionsare presented.

Aliquots from the samples studied by DLS were also analyzed byTEM as previously reported [13,16]. Sample aliquots (5 ll) were

adsorbed to carbon-coated collodion film supported on 200-meshcopper grids. The grids were negatively stained with 1% uranyl ace-tate and visualized with a Zeiss microscope at 60 kV.

2.6. Cell culture assays

To test the different polyphenols as TTR aggregation inhibitorswe used a rat Schwannoma (RN22) (American Type Cell Collection)cell line stably transfected with TTR L55P cDNA [16]. Cells weregrown till 80% cell confluence with the compounds at 1 lM con-centration in cell medium (�5 days). Then, cells were incubatedfor further 24 h still in the presence of compounds but in serum-free media. TTR in the medium was quantified by ELISA, and equalamounts of TTR (500 ng) were doted onto a 0.2 lm pore celluloseacetate membrane – dot-blot filter assay. TTR aggregates retainedin the membrane were immunodetected using rabbit anti-humanTTR (DAKO) (1:500) followed by anti-rabbit HRP antibody(1:1500) and ECL visualization (GE Healthcare). Dot blots werequantified using the ImageQuant program. Experiments were re-peated at least three times in duplicate. All values are expressedas means ± S.D. Comparison between groups was made using theStudent’s t-test, a P value of less than 0.05 was considered statisti-cally significant.

2.7. Cell toxicity assays

Rat Schwannoma cells (RN22) were propagated and maintainedas described previously [16]. Briefly, 80% confluent cells in Dul-becco’s minimal essential medium with 1% fetal bovine serumwere exposed for 24 h to 2 lM of TTR Y78F oligomers. These olig-omers were obtained by incubation of soluble TTR Y78F either inthe absence or presence of a 10� molar excess of EGCG, curcuminor NDGA at 37 �C for 6 days. Then, cells were trypsinized and celllysates were used for determination of caspase-3 activation withthe CaspACE fluorimetric 96-well plate assay system (Sigma). Pro-tein concentration in lysates was determined with the Bio-Rad pro-tein assay kit.

2.8. Disruption of TTR fibrils

To study the effect of the different polyphenols on TTR fibrilsdisruption, we used TTR pre-formed fibrils prepared by incubationof a filtered (0.2 lm filters) solution of TTR L55P (2 mg/ml in PBS)for 15 days at 37 �C. Subsequently, the samples were incubatedeither in the absence (control) or presence of a 10� molar excessof the polyphenols for 4 days at 37 �C. The disruption effect wasevaluated by TEM and DLS, as described above.

3. Results

3.1. Competition for T4 binding

To determine the ability of polyphenols to compete with T4 forthe binding to TTR the compounds were incubated with isolatedTTR, either recombinant WT (data not shown) or V30M (Fig. 1A),and [125I]T4 and the samples were analyzed by PAGE. Similarex vivo assays were performed with TTR in whole plasma from het-erozygote TTR V30M carriers or control individuals (data notshown). Samples incubated with curcumin and NDGA, presentedless intense TTR bands, indicating that both phytochemicals com-peted with [125I]T4, being NDGA the strongest competitor.Rosmarinic acid, caffeic acid and scyllo-inositol did not competewith [125I]T4 for its binding to TTR.

The relative inhibition potency of binding, expressed as the ra-tio of EC50 T4/EC50 competitor, was determined by competition of

Fig. 1. TTR binding and resistance to dissociation studies. (A) PAGE analysis of recombinant TTR V30M incubated with [125I] T4, in the absence or presence of the compounds.(B) Curves of T4 displacement from TTR WT by polyphenols, obtained by gel filtration assay. (C) IEF analysis of plasma TTR stability after treatment with polyphenols. Plasmasfrom normal and TTR V30M heterozygotic carriers were treated with EGCG, curcumin, NDGA. Different molecular species visualized in the IEF gel after Coomassie Bluestaining are indicated. These gels are representative of others run in parallel. (D) The histogram shows TTR tetramer/total TTR bands ratio obtained after densitometry of IEFgels corresponding to the analysis of 5 plasma samples (normal + V30M carriers) (⁄P 6 0.05; ⁄⁄P 6 0.01).

2426 N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430

the polyphenols with T4 for the binding to recombinant TTR WT as-sessed by gel filtration. The displacement curves obtained (Fig. 1B)demonstrate that NDGA is a more potent T4 competitor for thebinding to TTR than curcumin, with relative inhibition potenciesof 0.1 and 0.02, respectively. These results were consistent withthe in vitro and ex vivo data from the PAGE binding assay.

3.2. Effect of natural polyphenols on TTR tetramer dissociation

The effect of the natural polyphenols on TTR tetramer resistanceto dissociation was analyzed ex vivo by IEF under semi-denaturingconditions. EGCG, formerly described as being a strong TTR tetra-mer stabilizer, was used as Ref. [13]. In the absence of any com-pound, plasma TTR presented a characteristic band pattern,composed of monomer, an oxidized monomer and several lowerpI bands corresponding to different forms of tetramers. The ratiotetramer/total TTR (tetramer + monomers) is normally higher(0.63) for plasma from normal individuals than for the heterozyg-otic TTR V30M carriers plasma (0.39) [12]. The results obtained forplasma incubated with the different polyphenols showed that ros-marinic acid, caffeic acid and scyllo-inositol do not stabilize TTR inwhole plasma (data not shown). On the contrary, curcumin incu-bated samples displayed 25% increase of the tetramer/total TTR ra-tio, in both normal and TTR V30M heterozygotic, whereas NDGAhad lower stabilizing effect (14%). EGCG, presented the strongeststabilizing effect (30%), and also altered the migration of TTR tetra-mers in the IEF gel as previously referred [13] (Fig. 1C and D).

3.3. Effect of polyphenols on TTR aggregation in vitro

We next tested curcumin and NDGA for their effect on TTR Y78Faggregation, by TEM and DLS analysis. This TTR variant forms fi-brils under physiologic conditions [4] and its aggregation pathwayhas already been detailed by TEM [13]. In brief, soluble proteinevolves to oligomeric species and short twisted fibrils within2 days, forms large protein aggregates at 4 days, and after 6 daysof incubation the sample is mostly composed of fibrils as shownin Fig. 2A, control for 0, 4 and 6 days. TTR Y78F was incubated withcurcumin, NDGA and EGCG as reference inhibitor, and analyzed atdifferent time points, up to 16 days.

In accordance with previous results, EGCG powerfully impairedpre-incubated TTR Y78F samples (Fig. 2A, EGCG for 4 and 6 days),thus only soluble protein and small ‘‘off pathway’’ oligomers, butno large aggregates or fibrils, could be distinguished in the prepa-rations for all time points analyzed. Similarly, in curcumin treatedsamples, only small protein oligomers, but no larger protein aggre-gates or fibrils (at 4 and 6 days incubation), were detectedthroughout the period investigated (16 days) (data not shown).In contrast, NDGA pre-incubated samples displayed higher proteinaggregation in comparison with samples incubated with otherpolyphenols, indicating that NDGA was not as effective ascurcumin or EGCG in preventing TTR Y78F aggregation (Fig. 2A,NDGA at 4 and 6 days).

In parallel, the same samples were analyzed by DLS for addi-tional characterization of the protein species formed. Initially,

Fig. 2. (A) TEM analysis of TTR Y78F aggregation, incubated at 37 �C under stagnant conditions, in the absence (control) and presence of EGCG, curcumin or NDGA for 4 and6 days. Scale bar = 200 nm. (B) DLS analysis of TTR Y78F aggregation for 0, 4 and 12 days in the absence (black line) and presence of EGCG (green line), curcumin (red line) andNDGA (blue line).

N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430 2427

(0 days) the sample was composed by particles of 8 nm hydrody-namic diameter (dH) (Fig. 2B) which is consistent with previous re-port for the soluble TTR molecule [17]. After 4 days incubation, thecontrol sample (black line) exhibited a monodisperse particle pop-ulation of dH around 2000 nm, corresponding to large proteinaggregates and fibrils. Extended incubation, for 6, 12 and 16 days,resulted in a homogenous population with dH around 4000 nm,corresponding to TTR fibrils and bulky aggregates. In the presenceof the different polyphenols, TTR Y78F aggregation followed di-verse protein aggregation pathways. After 4 days, EGCG pre-incu-bated samples (green line) displayed a major particle populationcorresponding to soluble protein, indicating that this polyphenolstrongly inhibits TTR Y78F aggregation. This inhibitory effect wasmaintained at 12 days and up to 16 days incubation. Curcuminpre-treated samples (red line) presented a homogeneous popula-tion of 80 nm dH, corresponding to small TTR aggregates, after4 days incubation. This characteristic particle size homogeneitywas stably sustained during the period tested (16 days). Four daysafter incubation, samples with NGDA (blue line) presented 2 mainparticle populations corresponding to soluble protein and largeaggregates (dH � 1000 nm). Although NDGA did not effectively im-pair TTR fibril formation, particle size (dH � 1000 nm) was lowerthan control (dH � 4000 nm) after 12 days incubation period.These results were consistent with those obtained from TEM.Hence, curcumin is an efficient inhibitor of TTR fibril formation,whereas NDGA is not. However, curcumin does not prevent fibrilformation as efficiently as EGCG does and originates differentintermediates.

3.4. Effect of polyphenols on TTR aggregation in a cell culture system

The effect of curcumin and NDGA on TTR aggregation was alsoinvestigated in a rat Schwannoma cell line transfected with TTRL55P that secretes the TTR variant to the medium, where it aggre-gates [16]. Conditioned medium was filtered through cellulose ace-

tate membrane and the retained aggregates are immunodetectedby anti-TTR antibody. In this assay, we compared the media ob-tained by incubation of cells with curcumin, NDGA or EGCG tothe media of non-treated cells. Dot-blot of conditioned mediumobtained in the presence of curcumin (Fig. 3A) revealed almostcomplete inhibition of TTR aggregation (90%). This effect was com-parable to that of EGCG (Fig. 3B). NDGA reduced protein aggregateformation but still its potency was approximately 4� lower thanthat of curcumin.

3.5. Effect of polyphenols on TTR fibril disruption

Comparison of crystallographic structure of TTR L55P with TTRWT revealed key structural differences that might be representa-tive of amyloid precursor species [18]. Thus, TTR L55P is an impor-tant tool to study TTR fibrillogenesis and amyloid formationin vitro, under physiological conditions.

TTR L55P pre-formed fibrils were incubated with the polyphe-nols for 4 days at 37 �C. At the end of this incubation period, TEM(Fig. 4A) and DLS (Fig. 4B) analysis evidenced that EGCG, curcuminand NDGA clearly reduced the amount and size of TTR fibrils(dH � 5000 nm). As expected, EGCG powerfully disrupted TTRpre-formed fibrils into a polydisperse size particle population ofsmall amorphous protein aggregates (Fig. 4). Fibril disruption bycurcumin resulted in a major particle population of dH 100 nm(Fig. 4B), corresponding to small spherical oligomers (Fig. 4A)while NDGA induced moderate TTR fibril disaggregation as stilllarger aggregates and fibrils were occasionally observed in NDGAtreated samples (Fig. 4A).

3.6. Cell toxicity assay

Having characterized the inhibitory properties of EGCG, curcu-min and NDGA on TTR fibril formation we next evaluated the cyto-toxicity of the oligomeric species induced by the polyphenols.

Fig. 4. Effect of EGCG, curcumin and NDGA on TTR L55F pre-formed fibrils (t = 0). (A) TTR preparations were incubated in the absence (control) or presence of the differentpolyphenols and analyzed by TEM 4 days after incubation. Scale bar = 200 nm. (B) The same samples were analyzed by DLS in the absence (control) and presence of EGCG,curcumin and NDGA as indicated.

Fig. 3. (A) Immunodetection of TTR aggregates after dot-blot filter assay of the medium from TTR L55P transfected RN22 cells grown in the absence (control) or in thepresence of EGCG, curcumin and NDGA. (B) Representation of the percentage of aggregation inhibition based on quantification of the dots obtained in A (⁄⁄P 6 0.01;⁄⁄⁄P 6 0.005).

2428 N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430

N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430 2429

RN22 cells were incubated in the absence or presence of TTR Y78Foligomers alone or pre-treated with polyphenols for 24 h. Cell ly-sates were used for determination of Caspase-3 activation, andthe results showed that, while extracellular addition of non-treated TTR oligomers considerably increased intracellular levelsof Caspase-3, EGCG and curcumin induced oligomers significantlyalleviate extracellular induced toxicity (77% and 24% reduction,respectively). NDGA incubated oligomers presented only marginalreduction (2%) of Caspase-3 activation.

4. Discussion

Recent studies on Ab amyloid formation indicate that severalsmall aromatic compounds interfere with aggregation pathwayspossibly by remodeling the amyloid intermediates through differ-ent mechanisms or interactions [19].

Searching for efficient inhibitors of TTR amyloidosis and consid-ering that different proteins might share similarities in the misfold-ing pathway we investigated the effect of natural polyphenols onTTR amyloid formation and compared to the effect of EGCG. Curcu-min has been previously reported to bind to TTR WT [20]. However,no further characterization of the interaction has been conducted.Herein, we demonstrate that both curcumin and NDGA bind toTTR WT or TTR V30M in vitro and ex vivo at the T4 binding site.While NDGA presents the highest T4 displacement potency, curcu-min revealed to be a stronger tetrameric stabilizer, comparable toEGCG, indicating that T4 competition potency is not directly relatedto TTR stabilization efficiency and that curcumin and NDGA maypresent different modes of interaction at the T4 binding channelor may bind to diverse TTR regions. Concerning EGCG, it does notcompete with T4 for the binding to TTR [13] therefore the stabiliza-tion effect of EGCG on TTR is independent from the stabilization ef-fect of T4 and additionally strengthens the effect of T4 [21].Furthermore, the characterization of the EGCG-TTR V30M complexrevealed that EGCG binds at the surface of TTR molecule, in a regioninvolving amino acid residues at the interface of both dimmers, pro-moting tetramer stabilization [21].

The effect of phytochemicals on TTR aggregation inhibitionin vitro was assessed by TEM and DLS analysis using soluble TTRY78F that avoids acidification to form amyloid in vitro [4]. Curcu-min presented a potency of aggregation inhibition similar to EGCGwhile NDGA presented lower potency. In addition, a detailed anal-ysis of TTR aggregation dynamics in the presence of each of thethree polyphenols evidenced different modes of action for differentcompounds. EGCG treated samples maintained a major particlepopulation corresponding to soluble TTR while curcumin inducedTTR oligomerization into a monodisperse particle population, cor-responding to small spherical aggregates. In contrast, NDGA exhib-ited lesser inhibition potency by presenting larger particles insolution. We further investigated the polyphenol inhibition prop-erties on TTR aggregation in cultured cells expressing TTR L55P.The results obtained corroborated the in vitro studies revealing sig-nificant TTR aggregation impairment in the cell media by curcumin(90%) and to a less extent by NDGA (23%).

It is known that early non-fibrillar TTR aggregates are the pri-mary cytotoxic species, whereas the soluble counterpart and matureamyloid TTR fibrils are not toxic [11,22]. On the other hand, previousstudies have shown that curcumin is able to block Ab 42 and aS oli-gomer toxicity and increase cell viability in SH-SY5 human neuro-blastoma cells [14,23]. Likewise, it has been reported that EGCG-induced Ab 42 and aS oligomers were non-toxic in cell based assays[24]. In accordance, EGCG and curcumin significantly decreasedextracellular TTR oligomeric induced toxicity, indicating that bothproduced non-toxic ‘‘off-pathway’’ aggregation species though dif-ferent size aggregation intermediates were generated in each case.

Besides inhibition of aggregation, curcumin and NDGA havealso been reported to destabilize Ab and aS pre-formed fibrilsin vitro and in vivo [15,25]. The results demonstrated that thesephytochemicals disrupted pre-formed amyloid fibrils, though withdifferent potency. EGCG powerfully disrupted mature TTR fibrilsinto a very polydisperse population of protein aggregates support-ing results obtained for other amyloidogenic variants [13]. Curcu-min markedly disaggregated amyloid fibrils, originating aparticular population of small amorphous aggregates, morpholog-ically similar to previously reported curcumin destabilized Ab fi-brils [15]. NDGA decreased the levels and size of TTR fibrils, butonly to a minor extent.

Several human clinical trials indicate that curcumin is well tol-erated and non-toxic [26] although a few in vivo studies report ad-verse effects at very high doses [27]. Therefore, while consideringin vivo testing in mouse model for FAP, different doses of curcuminor optimized derivatives should be assayed.

In conclusion, the present study demonstrated that curcumin,NDGA and EGCG, all natural occurring polyphenols, impaired TTRaggregation through different modes of action: EGCG strongly sup-pressed TTR tetramer dissociation and mostly sustained TTR assoluble protein; curcumin induced TTR oligomerization into a char-acteristic homogeneous population of small ‘‘off-pathway’’ non-toxic aggregates and NDGA moderately reduced the amount ofTTR aggregation.

Acknowledgements

This work was supported by Fundação para a Ciência e Tecnolo-gia (FCT) through a grant [POCI/SAU-MMO/57321/2004] and fel-lowship to Nelson Ferreira [SFRH/BD/28566/2006]. The authorsacknowledge Rui Fernandes from Advanced Tissue Analysis (ATAF)from Instituto de Biologia Molecular e Celular (IBMC) for excellenttechnical assistance with Electron Microscopy.

References

[1] Saraiva, M.J. (2001) Transthyretin mutations in hyperthyroxinemia andamyloid diseases. Hum. Mutat. 17, 493–503.

[2] Benson, M.D. and Kincaid, J.C. (2007) The molecular biology and clinicalfeatures of amyloid neuropathy. Muscle Nerve 36, 411–423.

[3] Jacobson, D.R., McFarlin, D.E., Kane, I. and Buxbaum, J.N. (1992) TransthyretinPro55, a variant associated with early-onset, aggressive, diffuse amyloidosiswith cardiac and neurologic involvement. Hum. Genet. 89, 353–356.

[4] Redondo, C., Damas, A.M., Olofsson, A., Lundgren, E. and Saraiva, M.J. (2000)Search for intermediate structures in transthyretin fibrillogenesis: solubletetrameric Tyr78Phe TTR expresses a specific epitope present only in amyloidfibrils. J. Mol. Biol. 304, 461–470.

[5] Alves, I.L., Divino, C.M., Schussler, G.C., Altland, K., Almeida, M.R., Palha, J.A.,Coelho, T., Costa, P.P. and Saraiva, M.J. (1993) Thyroxine binding in a TTR Met119 kindred. J. Clin. Endocrinol. Metab. 77, 484–488.

[6] Colon, W. and Kelly, J.W. (1992) Partial denaturation of transthyretin issufficient for amyloid fibril formation in vitro. Biochemistry 31, 8654–8660.

[7] Quintas, A., Saraiva, M.J. and Brito, R.M. (1999) The tetrameric proteintransthyretin dissociates to a non-native monomer in solution. A novelmodel for amyloidogenesis. J. Biol. Chem. 274, 32943–32949.

[8] Almeida, M.R., Gales, L., Damas, A.M., Cardoso, I. and Saraiva, M.J. (2005) Smalltransthyretin (TTR) ligands as possible therapeutic agents in TTR amyloidoses.Curr. Drug Targets: CNS Neurol. Disord. 4 (5), 587–596.

[9] Connelly, S., Choi, S., Johnson, S.M., Kelly, J.W. and Wilson, I.A. (2010)Structure-based design of kinetic stabilizers that ameliorate thetransthyretin amyloidoses. Curr. Opin. Struct. Biol. 20 (1), 54–62.

[10] Purkey, H.E., Dorrell, M.I. and Kelly, J.W. (2001) Evaluating the bindingselectivity of transthyretin amyloid fibril inhibitors in blood plasma. Proc.Natl. Acad. Sci. 98, 5566–5571.

[11] Cardoso, I., Merlini, G. and Saraiva, M.J. (2003) 40-Iodo-40-deoxydoxorubicinand tetracyclines disrupt transthyretin amyloid fibrils in vitro producingnoncytotoxic species: screening for TTR fibril disrupters. FASEB J. 17, 803–809.

[12] Almeida, M.R., Macedo, B., Cardoso, I., Alves, I., Valencia, G., Arsequell, G., Planas,A. and Saraiva, M.J. (2004) Selective binding to transthyretin and tetramerstabilization in serum from patients with familial amyloidotic polyneuropathyby an iodinated diflunisal derivative. Biochem. J. 381, 351–356.

[13] Ferreira, N., Cardoso, I., Domingues, M.R., Vitorino, R., Bastos, M., Bai, G.,Saraiva, M.J. and Almeida, M.R. (2009) Binding of epigallocatechin-3-gallate totransthyretin modulates its amyloidogenicity. FEBS Lett. 583, 3569–3576.

2430 N. Ferreira et al. / FEBS Letters 585 (2011) 2424–2430

[14] Yang, F., Lim, G.P., Begum, A.N., Ubeda, O.J., Simmons, M.R., Ambegaokar, S.S.,Chen, P.P., Kayed, R., Glabe, C.G., Frautschy, S.A. and Cole, G.M. (2005)Curcumin inhibits formation of amyloid beta oligomers and fibrils, bindsplaques, and reduces amyloid in vivo. J. Biol. Chem. 280, 5892–5901.

[15] Ono, K. and Yamada, M. (2006) Antioxidant compounds have potent anti-fibrillogenic and fibril-destabilizing effects for alpha-synuclein fibrils in vitro.J. Neurochem. 97, 105–115.

[16] Cardoso, I., Almeida, M.R., Ferreira, N., Arsequell, G., Valencia, G. and Saraiva,M.J. (2007) Comparative in vitro and ex vivo activities of selected inhibitors oftransthyretin aggregation: relevance in drug design. Biochem. J. 408, 131–138.

[17] Hou, X., Parkington, H.C., Coleman, H.A., Mechler, A., Martin, L.L., Aguilar, M.I.and Small, D.H. (2007) Transthyretin oligomers induce calcium influx viavoltage-gated calcium channels. J. Neurochem. 100, 446–457.

[18] Sebastião, M.P., Saraiva, M.J. and Damas, A.M. (1998) The crystal structure ofamyloidogenic Leu55Pro transthyretin variant reveals a possible pathway fortransthyretin polymerization into amyloid fibrils. J. Biol. Chem. 273, 24715–24722.

[19] Ladiwala, A.R., Dordick, J.S. and Tessier, P.M. (2011) Aromatic small moleculesremodel toxic soluble oligomers of amyloid beta through three independentpathways. J. Biol. Chem. 286, 3209–3218.

[20] Pullakhandam, R., Srinivas, P.N., Nair, M.K. and Reddy, G.B. (2008) Binding andstabilization of transthyretin by curcumin. Arch. Biochem. Biophys. 485, 115–119.

[21] Miyata, M., Sato, T., Kugimiya, M., Sho, M., Nakamura, T., Ikemizu, S., et al.(2010) The crystal structure of the green tea polyphenol (�)-epigallocatechingallate–transthyretin complex reveals a novel binding site distinct from thethyroxine binding site. Biochemistry 49, 6104–6114.

[22] Sousa, M.M., Cardoso, I., Fernandes, R., Guimarães, A. and Saraiva, M.J. (2001)Deposition of transthyretin in early stages of familial amyloidoticpolyneuropathy: evidence for toxicity of nonfibrillar aggregates. Am. J.Pathol. 159, 1993–2000.

[23] Wang, M.S., Boddapati, S., Emadi, S. and Sierks, M.R. (2010) Curcumin reducesalpha-synuclein induced cytotoxicity in Parkinson’s disease cell model. BMCNeurosci. 11, 57.

[24] Ehrnhoefer, D.E., Bieschke, J., Boeddrich, A., Herbst, M., Masino, L., Lurz, R.,Engemann, S., Pastore, A. and Wanker, E.E. (2008) EGCG redirectsamyloidogenic polypeptides into unstructured, off-pathway oligomers. Nat.Struct. Mol. Biol. 15, 558–566.

[25] Garcia-Alloza, M., Borrelli, L.A., Rozkalne, A., Hyman, B.T. and Bacskai, B.J.(2007) Curcumin labels amyloid pathology in vivo, disrupts existing plaques,and partially restores distorted neurites in an Alzheimer mouse model. J.Neurochem. 102, 1095–1104.

[26] Goel, A., Kunnumakkara, A.B. and Aggarwal, B.B. (2008) Curcumin as‘‘Curecumin’’: from kitchen to clinic. Biochem. Pharmacol. 75 (4), 787–809.

[27] Burgos-Morón, E., Calderón-Montaño, J.M., Salvador, J., Robles, A. and López-Lázaro, M. (2010) The dark side of curcumin. Int. J. Cancer 126, 1771–1775.