Tetracycline and its analogues protect Caenorhabditis elegans from β amyloid-induced toxicity by targeting oligomers

Neurobiology of Disease 40 (2010) 424–431

Contents lists available at ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r.com/ locate /ynbd i

Tetracycline and its analogues protect Caenorhabditis elegans from βamyloid-induced toxicity by targeting oligomers

Luisa Diomede a,⁎, Giuseppe Cassata b, Fabio Fiordaliso c, Monica Salio c, Diletta Ami d,e, Antonino Natalello d,Silvia Maria Doglia d, Ada De Luigi a, Mario Salmona a

a Department of Molecular Biochemistry and Pharmacology, Istituto di Ricerche Farmacologiche “Mario Negri’, Via La Masa 19, 20158 Milan, Italyb C. elegans Genetics, IFOM, Fondazione Istituto FIRC di Oncologia Molecolare, Milan, Italyc Unit of Bio-imaging, Department of Cardiovascular Research, Istituto di Ricerche Farmacologiche “Mario Negri”, Via La Masa 19, 20158 Milan, Italyd Department of Biotechnology and Biosciences, University of Milano-Bicocca, Piazza della Scienza 2, 20126 Milan, Italye Fondazione IRCCS Policlinico San Matteo, Viale Camillo Golgi 19, 27100 Pavia, Italy

⁎ Corresponding author. Fax: +39 02 39014744.E-mail address: [email protected] (L. DioAvailable online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.nbd.2010.07.002

a b s t r a c t

Article history:Received 11 June 2010Revised 5 July 2010Accepted 7 July 2010Available online 14 July 2010

Keywords:Alzheimer's diseaseC. elegansTransgenicTetracyclineAmyloidAmyloid beta-peptide

The accumulation and deposition of amyloid beta (Aβ) peptide in extracellular dense plaques in the brain is akey phase in Alzheimer's disease (AD). Small oligomeric forms of Aβ are responsible for the toxicity and theearly cognitive impairment observed in patients before the amyloid plaque deposits appear. It is essential forthe development of an efficient cure for AD to identify compounds that interfere with Aβ aggregation,counteracting the molecular mechanisms involved in conversion of the monomeric amyloid protein intooligomeric and fibrillar forms. Tetracyclines have been proposed for AD therapy, although their effects on theaggregation of Aβ protein, particularly their ability to interact in vivo with the Aβ oligomers and/oraggregates, remain to be understood. Using transgenic Caenorhabditis elegans as a simplified invertebratemodel of AD, we evaluated the ability of tetracyclines to interfere with the sequence of events leading to Aβproteotoxicity. The drugs directly interact with the Aβ assemblies in vivo and reduce Aβ oligomer deposition,protecting C. elegans from oxidative stress and the onset of the paralysis phenotype. These effects werespecific, dose-related and not linked to any antibiotic activity, suggesting that the drugs might offer aneffective therapeutic strategy to target soluble Aβ aggregates.

mede).irect.com).

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Alzheimer's disease (AD) is one of the most frequent proteinmisfolding-linked pathologies and there is an urgent need for anefficient therapeutic strategy, since no effective treatment is available atpresent (Golde, 2006; Wechalekar et al., 2008). AD causes theaccumulation of extracellular dense plaques in the brain, and theirmajor component is the amyloid beta (Aβ) peptide, responsible forprogressive neurodegeneration and toxicity (Kayed et al., 2003).Numerous studies have suggested that soluble, small oligomeric formsof Aβ are responsible for the toxicity and the early cognitive impairmentobserved in patients before the amyloid plaque deposits appear (Kayedet al., 2003; Taylor et al., 2010; Nimmrich and Ebert, 2009). Themechanism underlying the conversion of the monomeric amyloidprotein into crossβ oligomeric andfibrillar forms is complex and closelydependent on ubiquitous constituents of the environment where theamyloid is deposited (Kayed et al., 2003; Taylor et al., 2010; Nimmrichand Ebert, 2009). Compounds able to interfere with the molecularmechanisms involved in the amyloidogic pathway have been proposed

as a possible prophylactic and therapeutic approach (Bartolini et al.,2007; Stains et al., 2007).

Various strategies have been designed to interfere with Aβaggregation, including direct interaction with the misfolded protein,modification of the formation kinetics of amyloid fibrils, orthe facilitation of their re-absorption (Becker and Greig, 2008).Tetracyclines, a well-known class of antibiotic drugs, are emerging asanti-amyloidogenic compounds that attenuate the resistance of Aβfibrils to proteolysis, preventing accumulation of the peptide in vitro(Forloni et al., 2001). Tetracyclines cross the blood brain barrier andare already used in clinical practice for central nervous system injury,offering the advantage of a safe toxicological profile and well-characterized pharmacological properties (Choi et al., 2007; Forloniet al., 2009; Noble et al., 2008). More research is needed to deciphertheir effects on the aggregation of the Aβ protein, particularly theirability to interact in vivo with the Aβ oligomers and/or aggregates.

We examined the tetracyclines' effects on the sequence of eventsleading to Aβ proteotoxicity, using transgenic C. elegans as a simplifiedinvertebrate model of AD (Link, 1995, 2005). To investigate the drugs'ability to counteract the Aβ deposition and toxicity in vivowe employedthe CL4176 transgenic C. elegans strain, engineered to inducibly expresshuman Aβ1–42 peptide in muscle when temperature is raised (Link,2005). These nematodes, when shifted to non-permissive conditions,

425L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

become rapidly paralyzed and it was suggested that the acute inductionof the transgene caused the production of Aβ oligomers which areresponsible for the toxicity (Wuet al., 2006, 2010) offering anoriginal invivo model for investigating the toxicity specifically related to smalloligomeric forms. Another C. elegans strain, CL2006, constitutivelyexpressing cytoplasmic human Aβ3–42 in the body wall muscle cells(Link, 1995, 2005), was considered too. In this strain the age-relatedprogressive reduction of muscle-specific motility correlates with theaccumulation of both fibrils and oligomers of Aβ3–42 (17). Both strainshave already been employed as in vivo models of AD and used todemonstrate the effect ofGingko biloba extract EGb 761, gingkolides andother compounds, in counteracting the Aβ toxicity (Wu et al., 2006;McColl et al., 2009; Luo, 2006; Gutierrez-Zepeda et al., 2005; Moritaet al., 2009).

The tetracyclines reduced the Aβ-induced paralysis and this effectwas related in vivo to a strong reduction of the soluble oligomeric Aβforms and fibril deposition. In addition, Aβ transgenic C. elegansstrains were confirmed as attractive biological systems to follow Aβaggregation in vivo allowing the accurate identification and investi-gation of promising therapeutic molecules specifically interactingwith oligomeric Aβ species.

Materials and methods

C. elegans strains

The construction and characterization of the CL2006 strain, and ofthe temperature-inducible transgenic nematode strain CL4176 and itscontrol CL802, have been described elsewhere (Link, 1995, 2005). TheCL2006 strain constitutively produces a body wall muscle-specificAβ3–42 (McColl et al., 2009), while the expression of muscle-specificAβ1–42 in CL4176 depends on raising the temperature from 16 to24 °C. The transgenic arrays in the CL2006, CL4176 and CL802 strainsall contain the dominant mutant collagen [rol-6 (su 1066)] asmorphological marker. All nematode strains were obtained from theCaenorhabditis Genetic Center (University of Minnesota, USA) andwere propagated at 16 °C on solid Nematode Growth Medium (NGM)seeded with E. coli (OP50) for food.

To prepare age-synchronized animals, nematodes were trans-ferred to fresh NGM plates on reaching maturity at 3 days of age andallowed to lay eggs overnight. Isolated hatchlings from the synchro-nized eggs (day 1) were cultured on fresh NMG plates at 16 °C.

Paralysis assay

A profile indicating the time points at which the temperaturewas raised, tetracyclines were administered and paralysis was ratedis given in Fig. 1A. CL4176 and CL802 worms (the transgenic controlstrain), after egg synchronization, were placed on fresh NMG plates(35×10 mm culture plates, 100 worms/plate) seeded with tetracy-cline-resistant E. coli for 54 h at 16 °C. To induce the transgeneexpression, the temperature was raised from 16 to 24 °C. Differenttimes before (30 h and 6 h) or after (18 h) the temperature shift,worms were treated with vehicle or drugs (–100 μM, 100 μl/plate)and paralysis was scored at 2 h intervals until all worms wereparalyzed. Worms that did not move or only moved the head whengently touched with a platinum loop were scored as paralyzed.Tetracycline hydrochloride and azithromycin were from Fluka(Switzerland), doxycycline hydrochloride, mynocicline hydrochlo-ride and Congo red were from Sigma Aldrich (Switzerland). Alldrugs were freshly dissolved in water before use. Synchronized eggsfrom CL2006 were placed on fresh NMG plates seeded withtetracycline-resistant E. coli at 16 °C. Seventy-two hours later, theywere treated with vehicle or drugs (25–100 μM, 100 μl/plate) andparalysis was evaluated until 120 h of age.

Western and dot blotting of Aβ species

The Aβ species in the transgenic C. elegans strains were identifiedby immunoblotting using Tris–Tricine gel, and by Western blotting(Wu et al., 2006). At the end of the experiments, the worms werecollected by washing with M9 buffer and homogenized in lysis buffer(25 mM Tris, pH 7.5, 5 mMNaCl, 5 mM EDTA, 1 mM dithiothreitol andprotease inhibitor mixture) using the TeSeE homogenizer (Bio-Rad)and acid-washed glass beads (Sigma). Samples were then heated in asample buffer containing 5% β-mercaptoethanol (1:1 v/v, Bio-Rad),cooled and equal amounts of the total protein (50–70 μg) were loadedin each lane of the gel. For dot blotting, equal amounts of proteinsfrom homogenized samples (3–5 μg) were spotted onto nitrocellulosemembranes. Mouse monoclonal antibody to human Aβ 1–17 (6E10,1:500 dilution) obtained from Signet (Emeryville, CA), rabbitpolyclonal antibody recognizing high molecular weight oligomers(A11, 1:1000 dilution) from Biosource (CA, USA) and monoclonalanti-a smooth muscle actin from Sigma (1:1000 dilution) were used.Anti-mouse IgG peroxidase conjugates or anti-rabbit IgG peroxidaseconjugates from Sigma (1:5000 dilution) were used as secondaryantibodies. The mean density of the Aβ-reactive bands was analyzedusing Progenesis SameSpots software (Nonlinear Dynamics, UK).

Immuno-electron microscopy

Individual nematodes were cut in a transversal plane with a razorblade just below the pharyngeal terminal bulb and fixed overnight in2% paraformaldehyde and 1% glutaraldehyde in 120 mM PIPES, 50 mMHEPES, 8 mM MgCl2 and 20 mM EGTA (PHEM) buffer to preservecellular structures. C. elegans were then quickly washed in a drop of7.5% gelatin (Sigma Aldrich G-2500), placed on a pre-warmed slide andembedded in 0.5 ml tubes at 37 °C in 15% gelatin. The tubes weremoved to an ice-bucket and, when the gelatin has solidified theembedded samples were dissected in 0.5–1 mm3 blocks and processedovernight at 4 °C with 2.3 M sucrose on a rotating wheel. Afterremoving the excess of sucrose, worms were placed on holders andimmediately frozen in liquid nitrogen. The holder with the frozen tissueblock was then mounted on the arm of a Leica EM UC6 ultramicrotomeequipped with a cryochamber (Leica EM UC6) and an antistatic device(Diatome, Switzerland). The sample was trimmed, sectioned to 50 nmthickness, collected on formvar/carbon-coated slot copper grids andincubated overnight with a rabbit anti-oligomer antibody A11 (1:100)at 4 °C, followed by 10-nm colloidal gold-conjugated protein Aincubation (Cell Microscopy Center, Utrecht, The Netherlands) for30 min. After labeling, sections were treated with 1% glutaraldehyde,counterstained with uranyl acetate and examined with an Energy FilterTransmission Electron Microscope (EFTEM, ZEISS LIBRA® 120)equipped with a YAG scintillator slow-scan CCD camera.

Fluorescent staining of β-amyloid

Age-synchronized transgenic CL2006 worms, fed or not withtetracyclines or Congo red as described before, were fixed in 4%paraformaldehyde/PBS, pH 7.4, for 24 h at 4 °C. Nematodes werestained with 1 mM 1,4-bis(3-carboxy-hydroxy-phenylethenyl)-ben-zene (X-34) in 10 mM Tris pH 8.0 for 4 h at room temperature (Linket al., 2001), destained, mounted on slides for microscopy andobserved with an inverted fluorescent microscope (IX-71 Olympus);images were acquired using a CDD camera.

Fourier-transform infrared analysis (FTIR)

FTIR absorption spectrawere collected from4000 to 600 cm−1 usinga UMA 500 infrared microscope equipped with a nitrogen-cooledmercury cadmium telluride detector (narrowband, 250 μm), coupled toa FTS-40A spectrometer (both from Digilab, MA, USA). For infrared

Fig. 1. Effect of tetracyclines on Αβ-induced paralysis in CL4176 and CL2006 transgenic C. elegans strains. A) Diagram illustrating the paralysis assay showing the time at which thetemperature was raised in CL4176 and CL802 worms, when the drugs were administered and when the paralysis assay was done. B) Egg-synchronized CL4176 and CL802 wormswere placed for 54 h at 16 °C on fresh NMG plates seeded with tetracycline-resistant E. coli. The temperature was raised from 16 to 24 °C and after 18 h, worms were fed vehicle or50 μM tetracyclines (100 μl/plate). Paralysis was scored at 2 h intervals until all worms were paralyzed. Data are shown as percentages±SD of worms not paralyzed (n=100, threeindependent assays). C) Percentages of paralyzed worms fed or not with tetracyclines or other antibiotics. Egg-synchronized CL4176 and CL802 worms were placed for 54 h at 16 °Con fresh NMG plates seeded with tetracycline-resistant E. coli. The temperature was raised from 16 to 24 °C and, after 18 h, worms were fed vehicle, 50 μM tetracycline, doxycycline,minocycline, azithromycin or ampicillin (100 μl/plate). Paralysis was scored 42 h after the temperature rise. Egg-synchronized CL2006 worms, placed at 16 °C for 72 h on fresh NMGplates seeded with tetracycline-resistant E. coli, were fed vehicle or 50 μM of drugs (100 μl/plate) and paralysis was determined 48 h later. Data are shown as percentages±SD ofparalyzed worms (n=100, three independent assays). *pb0.01 vs. CL4176 worms fed vehicle and †pb0.01 vs. CL2006 fed vehicle (Student's t test). D) Dose–response effect oftetracycline on Aβ-induced paralysis in CL4176 worms. Egg-synchronized CL4176 worms were placed at 16 °C for 54 h on fresh NMG plates seeded with tetracycline-resistant E. coli.Temperature was raised from 16 to 24 °C and, after 18 h, they were fed tetracycline (6.2–500 μM, 100 μl/plate). Paralysis was scored 42 h after the temperature rise. Each value is themean±SD ratio of the percentage of paralyzed worms to untreated ones. E) Representative images of X-34 staining of whole-mount and fixed sections of CL2006 worms fed or notwith tetracycline. The nematodes, synchronized and placed at 16 °C on tetracycline-resistant E. coli for 72 h, were fed vehicle or 50 μM tetracyclines (100 μl/plate) for 48 h and at120 h of age they were stained with X-34 dye. X-34 staining was visualized at short wavelength excitation. Scale bar 20 μm.

426 L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

measurements single intact specimens taken directly froman agar platewere extensivelywashed indistilledwater, deposited on aBaF2windowand dried at room temperature for about 30 min.

FTIR spectra were acquired in transmission mode from thepharynx region selected by the microscope variable diaphragmaperture, adjusted from 60 μm×60 μm to 100 μm×100 μm, depend-ing on the specimen size. Spectra with an excellent signal-to-noiseratio (peak to peak noise 0.5 mA at 2000 cm−1) were obtained withthe following instrument parameters: 4 cm−1 spectral resolution, 512scan co-additions, 20 kHz scan speed and triangular apodization.Three independent experiments were done to verify the spectrareproducibility (see electronic supplementary Figure S1). To resolvethe overlapped infrared bands, second derivative analysis of thespectra was done by the Savitsky–Golay method (3rd gradepolynomial, 5 smoothing points), after binomial 11-point smoothingof the measured spectra, using GRAMS/32 software (GalacticIndustries Corporation, USA).When necessary, spectrawere correctedfor residual water vapor absorption before second derivative analysis.

Superoxide production

Superoxide anions in worms were estimated by the spectropho-tometric determination of the product of the reduction of nitro bluetetrazolium (NBT) (Choi et al., 2006). To this end, the generation of

superoxide anions was measured on 100 μl of sample after addition of1.5 μl of 50 nM phorbol myristate acetate and 50 μl 1.8 mM NBT(Sigma-Aldrich, St Louis, MO, USA) and followed by incubation at37 °C for 30 min. The absorbance was determined at 560 nm againstblank samples which contained no worm homogenate, using theInfinite M200 multifunctional micro-plate reader (Tecan, Austria).Superoxide production was expressed as the absorbance/mg ofprotein, as a percentage of untreated control worms (% NBT). Theprotein content was determined using a Bio-Rad Protein assay (Bio-Rad Laboratories GmbH, Munchen, Germany).

Statistical analysis

Vehicle and drug effects were compared by an independentStudent's t-test, and the IC50 was determined using Prism version 4.0for Windows (GraphPad Software, CA, USA). A p valueb0.05 wasconsidered statistically significant.

Results

Tetracyclines protect from Aβ-induced paralysis

We first investigated whether tetracyclines protect against theAβ toxicity related to small oligomeric forms. CL4176 and CL802

427L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

(used as control strain) synchronized eggs were placed on atetracycline-resistant E. coli strain for 54 h at 16 °C, then thetemperature was raised to 24 °C to induce transgene expression.According to the treatment schedule set out in Fig. 1A, worms werefed either vehicle or tetracyclines (6–100 μM, 100 μl/plate) starting atdifferent times before (30 h and 6 h, corresponding respectively to24 h and 48 h of age) or after (18 h, corresponding to 72 h of age) therise in temperature, and paralysis was rated. As already reported (Wuet al., 2006), the temperature rise caused time-dependent induction ofparalysis in CL4176 but not in CL802 worms, in which the Aβ1–42

transgene was not expressed (Fig. 1B). When worms were fed 50 μMtetracycline or its analogues doxycycline and minocycline, starting18 h after the temperature rise (at larval stage L3), paralysiswas substantially delayed compared to vehicle-treated worms(Fig. 1B). The PT50 value, i.e. the time at which half the worms wereparalyzed,was significantly lower inworms fed thedrugs (13.3±1.2 forvehicle-treated CL4176 worms and 17.0±0.9, 17.1±1.0 and 17.2±1.2for nematodes fed tetracycline, doxycycline and minocycline, respec-tively; five assays, 50 worms in each group, p=0.01, Student's t test).Feeding CL4176 and CL802 worms with 50 μM or lower doses oftetracyclines (5–0.6 μM), 30 h or 6 h before the Aβ transgene induction,significantly reduced the number of nematodes by about half, indicatingthat the administration of the drug in the larval stages L1 (1 day of age)or L2 (2 days of age) resulted in a toxic effect.

To quantify tetracyclines' protective effect on CL4176, thepercentage of paralyzed worms were scored 42 h after the temper-ature rise (Fig. 1C). Feeding with 50 μM of tetracycline, doxycycline orminocycline had similar protective effects, significantly reducingparalysis by about 51%, 53% and 49% respectively. Interestingly,paralysis was not affected by 50 μM azithromycin or ampicillin(Fig. 1C), indicating that the tetracyclines' action was not related toa generic antibiotic activity. The effect of tetracycline on paralysis,determined 42 h after the temperature rise, was dose-dependent inthe range of concentrations from 6.2 to 500 μM, supporting the viewof a specific action (Fig. 1D). The IC50, i.e. the drug concentration thatinhibited paralysis by 50%, was 39.9±1.5 μM (n=6, mean±SE).

To see whether tetracyclines also protected CL2006 from thetoxicity induced by constitutive Aβ3–42 expression leading to theaccumulation of both fibrils and oligomers (17), we examined themotility of transgenic CL2006 worms fed or not with the drug.Synchronized CL2006 and CL802 worms were placed at 16 °C ontetracycline-resistant E. coli for 72 h, then fed either vehicle ortetracycline (50 μM, 100 μl/plate). Paralysis was scored on adultanimals up to 120 h of age (Fig. 1A). As shown in Fig. 1C, at this time,the percentage of vehicle-fed CL2006 paralyzed worms (27±1%) wassignificantly lower (p=0.01, Student's t test) than vehicle-fed CL4176(56±4%), suggesting that the constitutive expression of Aβ3–42

produced a less toxic phenotype than the expression of Aβ1–42

oligomers. CL2006 worms fed 50 μM tetracycline, doxycycline orminocycline were similarly protected from toxicity, with only 74% ofparalyzed worms (Fig. 1C).

Tetracyclines affect Aβ aggregation in transgenic C. elegans

We then analyzed whether tetracyclines reduced the degree ofamyloidosis of CL2006 transgenic worms by evaluating the Aβ-reactivedeposits in their headregion (Wuetal., 2006). Theworms, synchronizedandplaced at 16 °C on a tetracycline-resistant E. coli strain for 72 h,werefed either vehicle or tetracyclines (50 μM, 100 μl/plate) for 48 h and at120 h of age they were stained with X-34 dye which recognizes β-amyloid aggregates but not oligomers (Link et al., 2001). As illustrated inFig. 1E, tetracycline strongly reduced the X-34 positive spots. Similarresults were obtained in worms fed the same dose of doxycycline,minocycline or 200 μM Congo red, which binds Aβ fibrils (data notshown).

We also did immuno-electron microscopy studies using the anti-oligomer A11 antibody to see whether tetracyclines affected oligomerdeposition. Detection of A11 binding with gold-labeled secondaryantibody indicated immunoreactive oligomeric inclusions between thetransversally cut myofilaments of bodymuscle cells of both CL2006 andCL4176 strains, but not in control CL802 worms (Fig. 2). This indicatedthat CL4176 transgenic animals, in which no X-34 or thioflavine Sdeposits were observed (Link et al., 2001), produced only oligomericspecies of Aβ1–42. In addition, in CL2006 worms, which showed dye-reactive amyloid deposits, only a fraction of theAβpeptide is aggregatedinto fibrils along with oligomers. A11 immunoreactive binding wassignificantly reduced in CL4176 and CL2006 worms fed tetracyclines(Fig. 2), indicating that the drugs reduced the amyloid fibrillardeposition by inhibiting Aβ oligomerization.

To investigate whether the tetracyclines' protective effects onparalysis were related to different Aβ assemblies, transgenic C. elegansfed or not with tetracyclines for 24 h were analyzed by Westernblotting using antibody specific for total Aβ (6E10) or Aβ oligomers(A11). Tetracycline did not affect total Aβ levels (Fig. 3B, 6E10) ineither CL4176 or CL2006 worms, but there was a specific reduction ofseveral Aβ oligomeric species, with molecular weight about 20–28 kDa, and a simultaneous increase of the band around 4 kDa,corresponding to Aβ monomers (Figs. 3A 6E10).

To confirm that tetracyclines affected the oligomeric Aβ species,samples were analyzed by dot blotting using the A11 antibody. Asshown in Figs. 3B–C, A11 immunoreactivity in wild type CL802wormswas not affected by tetracyclines, probably because this antibodyrecognizes oligomeric structures other than Aβ. Tetracyclines signif-icantly inhibited Aβ oligomerization in CL4176 worms (untreated100%, treated 58±10%, p=0.001, n=3) and – though less – inCL2006 worms (untreated 100%, treated 72±11% , p=0.05, n=3)suggesting that the shift from larger oligomers to monomers isinvolved in the protective effect against Aβ toxicity.

Changes in the secondary structure of Aβ protein, from solublenon-amyloidogenic form to β-sheet amyloidogenic structures, whichresulted in oligomerization and insoluble fibril formation, are criticalin AD progression. FTIR microspectroscopy proved a powerfultechnique to obtain information on the secondary structure ofproteins in complex biological systems (Thumanu et al., 2009; Amiet al., 2008; Choo et al., 1996), including nematodes (Ami et al., 2004).

We monitored the aggregation kinetics of Aβ1–42 in CL4176 atdifferent times after the temperature rise by recording the infraredabsorption spectra of the pharynx region of single intact specimens.The second derivative absorption spectra in the Amide I region (1700–1600 cm−1) was calculated to identify the secondary structure and theresponse of protein aggregates. As shown in Fig. 4A, at time0, just beforethe temperature rise, the spectrum of CL4176 showed two main bandsat 1656 cm−1 and at 1636 cm−1, due to the α-helix and β-sheetsecondary structures of the total protein content, similar to those in thespectra from CL802 worms (Fig. 4B). Induction of Aβ1–42 expressionresulted in a time-dependent shift of the band at 1636 cm−1 towardslower wave-numbers and in a new component around 1623 cm−1,which canbe assigned to the intermolecularβ-sheet structure of proteinaggregates (Kneipp et al., 2003). Fourty-six hours after the temperatureinduction, when the majority of worms were paralysed, the 1623 cm−1

component increases (Figs. 4A and B), with intensity varying in thedifferent specimens, reflectingpossible differences inAβ1–42 expression.The 1623 cm−1 β-aggregate component was also observed in CL2006worms at 120 h of age (Fig. 4C),whenAβ-reactive depositswere visible.As already reported (Kneipp et al., 2003; Doglia et al., 2008) the infraredresponse of protein aggregates in situ often consists of a shoulder thatoccurs at lower wave-numbers of the band due to native β-sheetstructures. This shoulder makes it possible to monitor easily thepresence of aggregates in a reproducible manner (Doglia et al., 2008).

The effect of tetracyclines in situ on the secondary structures of theAβ protein was then investigated. Feeding CL4176 and CL2006 worms

Fig. 2. Immuno-electron microscopy of transgenic C. elegans strains fed tetracycline. Transmission electron microscopy images of the contractile apparatus in the body wall of CL802,CL4176 and CL2006 transgenic C. elegans. Myofilaments are oriented longitudinally and perpendicular to the surface. Dense bodies (Db) anchor the thin (actin) filaments and limiteach sarcomere (Src). Each muscle cell is attached basally to the underlying basal lamina (Bl) and hypodermis (Hyp). The animal's outer surface is covered by a collagenous cuticleapproximately 0.5 μm thick. The noncontractile body (muscle belly) of the muscle cell contain the cytoplasm with mitochondria (Mit). A11-positive oligomers revealed by theinteraction with the 10 nm gold-conjugated protein A accumulated only between the trasversally cut myofilaments of sarcomeres in CL4176 and CL2006 worms. Tetracyclinedrastically reduced oligomer deposition in both transgenic strains. Scale bar 500 nm.

428 L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

with tetracyclines (50–100 μM) reverted the protein aggregation, thecomponentaround1623 cm−1beingappreciablyreducedwithspectrasimilar to the control CL802 (Fig. 4B). The same behaviour was alsoobserved in the infrared spectra of worms fed with doxycycline andminocyclinebutnotwithazithromycin.Thisresultindicatesthat invivotetracyclines reversed the β-sheet amyloidogenic structures of Aβpeptideoligomersandfibrilstoanon-amyloidogenicformandthatthiseffect isnot related to their antibioticactivity.

Tetracyclines reduced superoxide production

Oxidative stress occurs in transgenic C. elegans strains expres-sing human Aβ1–42 peptide and an increase of superoxide anionslevels precedes the paralysis (Luo, 2006; Gutierrez-Zepeda et al.,

2005). We investigated whether an antioxidant action, besides ananti-oligomeric and anti-fibrillogenic effect, contributed to theprotective effect of tetracyclines against Aβ toxicity. On the basis ofthe kinetics of the superoxide anions generation before paralysis(Luo, 2006; Gutierrez-Zepeda et al., 2005) we determined super-oxide production in CL4176 worms 33 h after the temperature riseand in CL2006 worms at 96 h of age. As expected, superoxide levelsrose significantly in both the Aβ expressing transgenic strainscompared to the control strain CL802 (Fig. 5). Feeding the CL4176worms with 50 μM tetracycline, starting 18 h after the temperaturerise, inhibited the production of oxygen free radicals, whereas noreal reduction was observed in CL802 worms, indicating that thetetracyclines' effect on superoxide production was specificallylinked to Aβ expression (Fig. 5). There was also a significant

Fig. 3. Aβ species in transgenic C. elegans strains. A) Representative western blot of Aβ species in CL802, CL4176 and CL2006 worms fed vehicle or tetracycline. CL4176 and CL802worms were maintained for 54 h at 16 °C, then the temperature was raised from 16 to 24 °C. Eighteen hours later, worms were fed vehicle or 50 μM tetracyclines (100 μl/plate) for24 h. CL2006, maintained at 16 °C for 72 h, were treated with vehicle or 50 μM tetracyclines (100 μl/plate) for 24 h. The worms were then collected and equal amounts of proteinwere loaded on each gel lane and immunoblotted with anti-Aβ antibody (6E10) or actin. Arrows indicate the Aβ oligomers (20 kDa) and monomers (4 kDa). B) Representative dotblot of total Aβ (6Ε10) and oligomers of Aβ (Α11) in worms before and after tetracycline feeding, as described above. C) Quantification of 6Ε10 immunoreactive total Aβ species andD) of A11 oligomers in CL802, CL4176 and CL2006 worms fed either vehicle or 50 μM tetracycline. Data are the mean volume of the immunoreactive band/the volume of totalponceau protein±SD from three independent experiments. *pb0.05 and **pb0.01 vs. CL802 untreated worms and ¶¶pb0.01 vs. CL4176 untreated worms, †pb0.05 vs. untreatedCL2006 and ††pb0.01 vs. untreated CL4176 worms (Student's t test).

429L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

reduction of oxygen free radicals generation in CL2006 worms,synchronized and placed at 16 °C on a tetracycline-resistant E. colistrain for 72 h, then fed 50 μM of tetracyclines (Fig. 5). No

Fig. 4. FTIR spectra of intact C. elegans expressing Aβ peptides. A) Second derivative FTIR spe48 h) after the temperature induction. B) Second derivative FTIR spectra of CL4176 worms, 4or with vehicle (continous line). The dotted line indicates the second derivative spectrum ofat 120 h of age, fed for 24 h with 100 μM tetracycline (dashed line) or with vehicle (continuopoint to higher aggregate levels. All the spectra were collected from the worm pharynx regithe total protein content.

differences were observed in animals fed with azithromycin (data notshown) indicating that the antioxidant properties of tetracyclines tooare not related to their antibiotic effect.

ctra of CL4176 transgenic worms taken just before (time 0) and at different times (24–6 h after the temperature induction, fed for 18 h with 100 μM tetracycline (dashed line)the control strain CL802. C) Second derivative FTIR spectra of CL2006 transgenic wormsus line). The control strain CL802 is reported with the dotted line. In B and C the arrowson and were are normalized to the tyrosine band around 1515 cm−1 to compensate for

Fig. 5. Effect of tetracyclines on superoxide anions production in transgenic C. elegansstrains. Age-synchronized CL4176 or CL802 worms, fed vehicle or 50 μM tetracyclines,were collected 33 h after the temperature rise in 1.6 ml 1% Tween 20 in PBS.Synchronized CL2006 worms placed at 16 °C for 72 h were fed 50 μM tetracyclines for24 h, then collected the same way. The NBT assay was done as described in Materialsand Methods. Error bars indicate SD. Results are the absorbance/mg of protein as apercentage of untreated control worms (% NBT). *pb0.05 vs. vehicle-treated CL2006and CL4176 worms and †pb0.05 vs. vehicle-treated CL802 worms (Student's t test).

430 L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

Discussion

Tetracyclines, a well-known class of antibiotics including doxycy-cline and minocycline, inhibit the aggregation and disrupt pre-formedAβ amyloid fibrils in vitro (Forloni et al., 2001). Their effects are notexclusively against Αβ since they disaggregate othermisfolded proteinstoo (Cardoso et al., 2003;Malmoet al., 2006; Thomas et al., 2003). In vivothey exert beneficial effects as anti-prion drugs, reducing or abolishingprion infectivity byboosting theprion protein's sensitivity to proteolyticdegradation (Forloni et al., 2009;De Luigi et al., 2008). The tetracyclines'anti-fibrillogenic efficacy in models of Parkinson's and Huntington'sdiseases was accompanied by the inhibition of caspase-1, caspase-3,inducible nitric oxide synthase expression and nitric oxide-mediatedtoxicity (Thomas et al., 2003). Their pharmacological effects maytherefore involve more than one pathway but this remains to beclarified. However, based on in vitro studies and animal models, at leastthree clinical trials have been undertaken to assess possible clinicalbenefits of tetracyclines in the treatment of prion and Alzheimer'sdiseases and transthyretin amyloidosis.

To establish whether tetracyclines' effects are linked to their anti-amyloidogenic activity we correlated the pharmacological actionswith in vivo amyloid aggregation and toxicity, employing twotransgenic C. elegans strains (Link, 1995, 2005) in which the Aβpeptide accumulates intracellularly in muscle cells, resulting in anage-dependent paralysis phenotype. The tetracyclines delayed theAβ-induced paralysis and the effects were specific, dose-related andnot linked to any antibiotic activity.

To decipher the details of the mechanism of action of these drugs,we looked at their ability to interfere in vivo with the soluble Aβoligomeric aggregates. The tetracycline-induced changes in thesecondary structure of Aβ protein, from soluble form to β-sheet-richstructures corresponding to the presence of oligomers and fibrils, wasinvestigated here in situ for the first time by collecting the FTIRabsorption spectra from the pharynx of single specimens of C. eleganstransgenic strains. The expression of the Aβ peptides in CL4176 andCL2006 strains, was accompanied by the formation of intermolecularβ-sheet structures. Feeding the worms tetracyclines significantlyreduced the component at 1623 cm−1, indicating that these drugsdisassembled both Aβ oligomeric and fibrillar β-sheet assemblies,restoring their non-amyloidogenic structures. These effects resultedin a reduction of oligomeric Aβ assemblies in the body muscle of bothCL4176 and CL2006 transgenic worms.

We show here directly for the first time that oligomers of humanAβ1–42, already designated as responsible for the toxic phenotype(Wu et al., 2010), specifically accumulated in the muscle cells ofparalyzed CL4176 worms and that tetracyclines by reducing theoligomeric species, protected C. elegans from the Aβ insult. Oligomericdeposition was also observed in adult CL2006 worms in which theamyloid produced progressively aggregates to form fibrillar deposits(Link et al., 2001). In this strain, tetracyclines reduced both oligomersand amyloid aggregates, protecting against Aβ induced paralysis. Asindicated by the protein expression profile, the tetracyclines' effectswere not accompanied by a reduction of total Aβ produced which canaffect amyloid formation and deposition, but counteract oligomerformation and therefore amyloid deposition.

It was recently reported that CL2006worms produced human Aβ3–42,which is found in senile plaque and constitutes about half of the total Aβaccumulated inADbrains (McColl et al., 2009). Sinceno transgenic animalmodels expressing Aβ3–42 are currently available, this C. elegans strainoffers an innovative approach to investigate therapeutic strategies tocounteract the toxicity of amino truncated Aβ. Our findings indicate forthe first time that tetracyclines counteract the toxicity induced by Aβ3–42

aswell, so theymay help to counteract the oligomerization of various Aβ-truncated products. Moreover, as these nematode models may also berelevant to inclusion bodymyositis, a debilitatingmyopathy caused byAβaccumulation in muscles (Rebolledo et al., 2008), the findings suggest apotential for tetracyclines for this disease too.

According to the “amyloid cascade” hypothesis, generation ofoxidative stress by Aβ may be part of the neurodegenerative processof AD (Butterfield, 1997). We report that tetracyclines, whoseantioxidant properties have already been described (Kraus et al.,2005), reduced superoxide production in the two transgenic C. elegansstrains employed in this study. This effect is linked to Aβ productionbut seems unrelated to the paralysis phenotype. Superoxide produc-tion was reduced by L-ascorbic acid too, but it had no protective effecton the paralysis phenotype or on Aβ oligomerization (Wu et al.,2006). The Ginkgo biloba extract EGb 761, which inhibited superoxideproduction to the same extent as L-ascorbic acid but which alsoreduced Aβ oligomerization, protected against Aβ-induced paralysis(Wu et al., 2006). Thus, superoxide production in transgenic Aβ-expressing strains may be a consequence of the general perturbationfollowing amyloid accumulation, although the toxicity is primarilydue to the paralysis induced by oligomer formation. Tetracyclinesprevented oxidation in both CL4176 and CL2006 worms, and thiscontributed to protecting against the Aβ-induced damage.

In conclusion, we found that tetracyclines are multifunctionalmolecules that can interact directly in vivowith different Aβ assembliesand reduce Aβ oligomer deposition, protecting transgenic C. elegansstrains from the paralysis phenotype. Physico-chemical studies haveindicated thatwhen tetracyclinewas co-dissolvedwithAβ1–42orAβ1–40, supramolecular complexes are immediately formed, preventing theformation of fibrillar aggregates. In particular, tetracycline led to theformation of colloidal particles that specifically sequester oligomers, sopreventing the progression of the amyloid cascade (unpublishedobservations).

Tetracyclines can therefore be proposed as a therapeutic strategytargeting soluble Aβ aggregates, and possibly also for other forms ofamyloidosis. Aβ-expressing worms, particularly those producing theoligomeric assemblies, whose degeneration phenotype is specificallylinked to the production of toxic protein, offer an attractive tool forscreening of candidate drugs for the treatment of AD and can providesignificant insight into the mechanisms of the drug–proteininteractions.

Acknowledgments

The authors would like to thank Dr. Christopher D. Link for helpfulfeedback during this work and for a critical reading of the manuscript.

431L. Diomede et al. / Neurobiology of Disease 40 (2010) 424–431

Role of funding source.This work was partly supported by the Cariplo Foundation, Project

n. 2009-2543, by the Project Nobel Guard and by the Banca Intesa SanPaolo.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at 10.1016/j.nbd.2010.07.002.

References

Ami, D., Natalello, A., Zullini, A., 2004. Doglia SM Fourier transform infraredmicrospectroscopy as new tool for nematode studies. FEBS Lett. 576, 297–300.

Ami, D., Neri, T., Natalello, A., Mereghetti, P., Doglia, S.M., Zanoni, M., Zuccotti, M.,Garagna, S., Redi, C.A., 2008. Embryonic stem cell differentiation studied by FT-IRspectroscopy. Biochim. Biophys. Acta 1783, 98–106.

Bartolini, M., Bertucci, C., Bolognesi, M.L., Cavalli, A., Melchiorre, C., Andrisano, V., 2007.Insight into the kinetic of amyloid β(1–42) peptide self-aggregation: elucidation ofinhibitor's mechanism of action. Chembiochem 8, 2152–2161.

Becker, R.E., Greig, N.H., 2008. Alzheimer's disease drug development in 2008 andbeyond: problems and opportunities. Curr. Alzheimer Res. 5, 346–357.

Butterfield, D.A., 1997. Beta-amyloid-associated free radical oxidative stress andneurotoxicity: implications for Alzheimer's disease. Chem. Res. Toxicol. 10, 495–506.

Cardoso, I., Merlini, G., Sariva, M.J., 2003. 4′-iodo-4′-deoxydoxorubicin and tetra-cyclines disrupt transthyretin amyloid fibrils in vitro producing noncytotoxicspecies: screening for TTR fibril disrupters. FASEB J. 17, 803–809.

Choi, H.S., Kim, J.W., Cha, Y.N., Kim, C., 2006. A quantitative nitroblue tetrazolium assayfor determining intracellular superoxide anion production in phagocytic cells. J.Immunoassay Immunochem. 27, 31–44.

Choi, Y., Kim, H.S., Shin, K.Y., Kim, E.M., Kim, M., Kim, H.S., Park, C.H., Jeong, Y.H., Yoo, J.,Lee, J.P., Chang, K.A., Kim, S., Suh, Y.H., 2007. Minocycline attenuates neuronal celldeath and improves cognitive impairment in Alzheimer's disease models.Neuropsychopharmacology 32, 2393–2404.

Choo, L.P., Wetzel, D.L., Halliday, W.C., Jackson, M., LeVine, S.M., Mantsch, H.H., 1996. Insitu characterization of beta-amyloid in Alzheimer's diseased tissue by synchrotronFourier transform infrared microspectroscopy. Biophys. J. 71, 1672–1679.

De Luigi, A., Colombo, L., Diomede, L., Capobianco, R., Mangieri, M., Miccolo, C., Limido,L., Forloni, G., Tagliavini, F., Salmona, M., 2008. The efficacy of tetracyclines inperipheral and intracerebral prion infection. PLoS ONE 3, e1888.

Doglia, S.M., Ami, D., Natalello, A., Gatti-Lafranconi, P., Lotti, M., 2008. Fourier transforminfrared spectroscopy analysis of the conformational quality of recombinantproteins within inclusion bodies. Biotechnol. J. 3, 193–201.

Forloni, G., Colombo, L., Girola, L., Tagliavini, F., Salmona, M., 2001. Anti-amyloidogenicactivity of tetracyclines: studies in vitro. FEBS Lett. 407, 404–407.

Forloni, G., Salmona, M., Marcon, G., Tagliavini, F., 2009. Tetracyclines and prioninfectivity. Infect. Disord. Drug Targets 9, 23–30.

Golde, T.E., 2006. Disease modifying therapy for AD? J. Neurochem. 99, 689–707.Gutierrez-Zepeda, A., Santell, R.,Wu, Z., Brown,M.,Wu, Y., Khan, I., Link, C.D., Zhao, B., Luo,

Y., 2005. Soy isoflavone glycitein protects against beta amyloid-induced toxicity andoxidative stress in transgenic Caenorhabditis elegans. BMC Neurosci. 6, 54–62.

Kayed, R., Head, E., Thompson, J.L., McIntire, T.M., Milton, S.C., Cotman, C.W., Glabe, C.G.,2003. Common structure of soluble amyloid oligomers implies common mecha-nism of pathogenesis. Science 300, 486–489.

Kneipp, J., Miller, L.M., Joncic, M., Kittel, M., Lasch, P., Beekes, M., Naumann, D., 2003. Insitu identification of protein structural changes in prion-infected tissue. Biochim.Biophys. Acta 1639, 152–158.

Kraus, R.L., Pasieczny, R., Lariosa-Willingham, K., Turner, M.S., Jiang, A., Trauger, J.W., 2005. Antioxidant properties of minocycline: neuroprotection in anoxidative stress assay and direct radical-scavenging activity. J. Neurochem.94, 819–827.

Link, C.D., 1995. Expression of human b-amyloid peptide in transgenic Caenorhabditiselegans. Proc. Natl. Acad. Sci. USA 92, 9368–9372.

Link, C.D., 2005. Invertebrate models of Alzheimer's disease. Genes Brain Behav. 4,147–156.

Link, C.D., Johnson, C.J., Fonte, V., Paupard, M.C., Hall, D.H., Styren, S., Mathis, C.A., Klunk,W.E., 2001. Visualization of fibrillar amyloid deposits in living, transgenicCaenorhabditis elegans animals using the sensitive amyloid dye, X-34. Neurobiol.Aging 22, 217–226.

Luo, Y., 2006. Alzheimer's disease, the nematode Caenorhabditis elegans, and ginkobiloba leaf extract. Life Sci. 78, 2066–2072.

Malmo, C., Vilasi, S., Iannuzzi, C., Tacchi, S., Cametti, C., Irace, G., Sirangelo, I.,2006. Tetracycline inhibits W7FW14F apomyoglobin fibrilar extension andkeeps the amyloid protein in a prefibrillar, highly cytotoxic state. FASEB J. 20,346–347.

McColl, G., Roberts, B.R., Gunn, A.P., Perez, K.A., Tew, D.J., Masters, C.L., Barnham, K.J.,Cherny, R.A., Bush, A.I., 2009. The Caenorhabditis elegans A beta 1–42 model ofAlzheimer disease predominantly expresses A beta 3–42. J. Biol. Chem. 284,22697–22702.

Morita, M., Osoda, K., Yamazaki, M., Shirai, F., Matsuoka, N., Arakawa, H., Nishimura, S.,2009. Effects of non-steroidal anti-inflammatory drugs on Aβ deposition in Aβ1–42

transgenic C. elegans. Brain Res. 1295, 186–191.Nimmrich, V., Ebert, U., 2009. Is Alzheimer's disease a result of presynaptic failure?

Synaptic dysfunctions induced by oligomeric beta-amyloid. Rev. Neurosci. 20,1–12.

Noble, W., Garwood, C., Stephenson, J., Kinsey, A.M., Hanger, D.P., Anderton, B., 2008.Minoycline reduces the deveoplment of abnormal tau species in models ofAlzheimer's disease. FASEB J. 23, 1–12.

Rebolledo, D.L., Minniti, A.N., Grez, P.M., Fadic, R., Kohn, R., Inestrosa, N.C., 2008.Inclusion body myositis: a view from the Caenorhabditis elegans muscle. Mol.Neurobiol. 38, 178–198.

Stains, C.I., Mondal, K., Ghosh, I., 2007. Molecules that target beta-amyloid.ChemMedChem 2, 1674–1692.

Taylor, M., Moore, S., Mayes, J., Parkin, E., Beeg, M., Canovi, M., Gobbi, M., Mann, D.M.,Allsop, D., 2010. Development of a proteolytically stable retro-inverso peptideinhibitor of beta-amyloid oligomerization as a potential novel treatment forAlzheimer's disease. Biochemistry 49, 3261–3271.

Thomas, M., Le, W.D., Jankovic, J., 2003. Minocycline and other tetracycline derivatives:a neuroprotective strategy in Parkinson's disease and Huntington's disease. Clin.Neuropharmacol. 26, 18–23.

Thumanu, K., Tanthanuch, W., Lorthongpanich, C., Heraud, P., Parnpai, R., 2009. FTIRmicrospectroscopic imaging as a new tool to distinguish chemical composition ofmouse blastocyst. J. Mol. Struct. 933 (1), 104–111.

Wechalekar, A.D., Hawkins, P.N., Gillmore, J.D., 2008. Perspectives in treatment of ALamyloidosis. Br. J. Haematol. 140, 365–377.

Wu, Y., Wu, Z., Butko, P., Christen, Y., Lambert, M.P., Klein, W.L., Link, C.D., Luo, Y., 2006.Amyloid-beta-induced pathological behaviors are suppressed by Ginkgo bilobaextract EGb 761 and ginkgolides in transgenic Caenorhabditis elegans. J. Neurosci.26, 13102–13113.

Wu, Y., Cao, Z., Klein, W.L., Luo, Y., 2010. Heat shock treatment reduces beta amyloidtoxicity in vivo by diminishing oligomers. Neurobiol. Aging 31, 1055–1058.