Utility of an improved model of amyloid-beta (Aβ_1-42) toxicity in Caenorhabditis elegansfor drug screening for Alzheimer’s disease

To engineer a C. elegans strain expressing full length Aß_1-42 we modified the synthetic signal peptide to be cleaved from the transgenic Aß used in the original expression construct [10]. An extra Asp-Ala (DA) was inserted N-terminal into the human-Aß sequence in the expression vector pCL12(unc-54:Aß _1-42 ) [10] (Additional file 1: Figure S1). Predicted processing of the new product, examined (data not shown) via SignalP 3.0 [11], suggested the position of signal peptide cleavage would yield full length Aß_1-42. Based on this prediction an integrated transgenic strain was engineered in C. elegans. Transgene expression was targeted to the bodywall muscle cells, via use of the promoter of unc-54 (which encodes a heavy chain muscle myosin). We then confirmed the molecular identity of the Aß expressed in the C. elegans strain GMC101 via complementary techniques. Using immunocapture and ESI-MS we determined the mono-isotopic molecular weight of the expressed Aß as 4511.2586 Da with an error of 2.5 ppm (Figure 1A). This mass is consistent with full length Aß_1-42 (expected mass 4511.2697 Da). An additional peak, shifted ~16 Da, was observed corresponding to Aß_1-42 plus a single oxygen (observed mass 4527.2710 Da). SELDI-TOF-MS incorporating antibody capture also confirmed the expression of Aß_1-42 (Additional file 2: Figure S2). In addition, bis/Bicine urea PAGE analysis also resolved a single species consistent with Aß_1-42 (Figure 1B). No additional truncated Aß species were detected.

The amount of Aß_1-42 expressed in this new transgenic strain, GMC101, was then compared to that expressing Aß_3-42 (strain CL2120). Total protein was extracted via a urea-based buffer and the relative amount of Aß was compared via immuno-blot analysis densitrometry (Figure 1C-1D). An equivalent amount of Aß was detected between strains CL2120 (Aß_3-42) and GMC101 (Aß_1-42).

The cellular localization of the accumulated transgenic Aß_1-42 was then examined via immuno-histochemistry. Transverse and longitudinal paraffin-embedded 5 μm sections confirmed Aß localization within the bodywall muscle cells (Figure 2), consistent with the promoter activity of the unc-54 promoter. Fluorescence based immuno-histochemistry also confirmed accumulation of Aß within the bodywall muscle of unsectioned individuals (Additional file 2: Figure S2). The expressed Aß_1-42 also appears to be at least partially aggregated as deposits stained positive for the amyloidogenic dye Thioflavin T (ThT) (Figure 3A). Aggregated Aß_1-42 could also be detected in vivo, via staining in live animals using the lipophillic-congo red derivative, X-34 [12] (Figure 3B). X-34 has brighter fluorescence than ThT and detects more Aß aggregates than ThT, as has been previously shown [12]. Representative images of the head are shown in Figure 3A-3B, as this is a region devoid of background fluorescence (including the GFP expressed in the intestine as a co-marker). We noted, however, that dye-binding aggregates are found throughout the entire length of the adult (Additional file 3: Figure S4).

Soluble oligomers of Aß have been proposed to be the toxic form of Aß [13]. To explore this further in our C. elegans model we used size-exclusion chromatography under native conditions to separate soluble proteins on the basis of size. The expressed Aß_1-42 in our C. elegans model elutes, almost exclusively, as high molecular weight material (>100KDa). Only minor amounts of Aß were detected that eluted in fractions consistent with monomer. These data are consistent with the Aß forming soluble oligomers and/or heteromeric complexes with other cellular macromolecules (e.g. proteins etc).

Differences in transgene array copy number and the genomic insertion sites of the trangenic arrays limits the ability to directly compare phenotypes between strains. However, we observed similar yet distinct toxic effects from Aß_1-42 versus Aß_3-42. The expression of Aß_1-42 had no adverse effects on motility if adults were cultured and maintained at 20°C over 4 days post adulthood (Figure 4B). In contrast, C. elegans expressing Aß_3-42 show an age-dependent paralysis phenotype when cultured at 20°C. However, when Aß_1-42 expressing adults, developed at 20°C, were shifted to 25°C we observed severe and fully penetrant paralysis within 48 h. The paralysis at 25°C appeared more severe for those expressing Aß_1-42 compared to Aß_3-42.

The rapid paralysis from Aß_1-42 expression in this C. elegans strain is well suited for assessing drug effects. To explore the utility of this nematode model of Aß toxicity to identify protective compounds we examined the effect of PBT2. We exposed L4 (the final larval stage) to a range of PBT2 concentrations for 24 h prior to a shift to 25°C (Figure 5A). A dose of 10 μg/ml was found to offer significant protection against the Aß-induced paralysis. A protective effect of PBT2 was observed after one day at 25°C with significantly fewer individuals exhibiting paralysis (p < 0.001). To examine how quickly PBT2 could act we then minimized the exposure time of cultures prior to the temperature shift to 25°C. When young adults (3 days post egg lay) were exposed to PBT2 and at the same time were shifted to 25°C, significant protection was still observed (p < 0.001, Figure 5B). PBT2 effects on Aß toxicity therefore appear to be rapid and are unlikely to be due to indirect effects on development. Treatment for with PBT2 for 24 h did not affect total Aß levels (Figure 5C), or in vivo aggregated Aß as determined by X-34 staining (data not shown).

Reduction of insulin-like signalling (ILS) in C. elegans is known to protect against Aß toxicity [14]. To explore whether PBT2 acts via modulation of the ILS pathway in C. elegans we used a GFP-based reporter strain. Under conditions of lowered ILS or stress (e.g. starvation, heat shock, etc) DAF-16 (a FOXO transcription factor) translocates from the cytoplasm to the nucleus [15]. Exposure to PBT2 did not alter DAF-16 localization (Figure 5D), suggesting its effects are not via modulation of ILS. We also observed that exposure to PBT2 produced no detectable induction of small heat shock proteins (data not shown), as measured by a hsp-16.2:GFP reporter strain [16]. This data are consistent with the protective effects of PBT2 being derived by a mechanism other than a generalized stress response.

The strains N2, wild type; CL2120, dvIs14(pCL12(unc-54: hu-Aß _1–42 ) + pCL26(mtl-2: GFP)), CL2122; dvIs15(mtl-2: GFP) [28], CL2070, dvIs70(pCL25(hsp16.2:: GFP) + pRF4(rol-6(su1006)) [16] and TJ356; zIs356(daf-16:: DAF-16-GFP) + pRF4(rol-6(su1006)) [15] were obtained from the Caenorhabditis Genetics Center. To engineer a full-length Aß _1–42 expressing strain the pCL12 plasmid [10] was modified by the addition of codons (5’-GAC-CGC-3’) for residues ASP-ALA between the signal peptide and the Aß _1–42 ORF via a QuikChange Multi Site-Directed Mutagenesis Kit (Stratagene). The primers used were: 5'-gcaccagcaggtaccgacgcggatgcagaattccga, and 5'-tcggaattctgcatccgcgtcggtacctgctggtgc. The resulting plasmid, called pCL354 (unc-54:DA-Aß _1-42 ) shown in Additional file 1: Figure S1, encodes: MHKVLLALFFIFLAPAGTDA DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVVIA, where the inserted residues are underlined and Aß_1-42 sequence shown in italics. A transgenic strain was generated via gonad micro-injection and a stable integrant derived following γ-irradiation as previously described [28]. This strain was then back crossed to wild type four times to give GMC101, dvIs100 [pCL354(unc-54:DA-Aß _1-42 ) + pCL26(mtl-2: GFP)]. All strains were cultured at 20°C on NGM [30] or 8P media [31] with E. coli (strain OP50) as indicated. On the first day of adulthood (3-days-old), populations were aged at 20 or 25°C as indicated.

To a ~100 mg liquid-N2 frozen pellet of 5-day-old GMC101 adults 2:1 (v/w) of 70% formic acid was added and incubated for 4 hrs at room temperature. The lysate was centrifuged at 16,500 × g for 15 min, the supernatant retained and neutralized with 1:20 (v/v) 1 M Tris pH 8.0 and then diluted again 1:10 (v/v) in H₂0. For immunocapture W0-2 (epitope: Aß5-8) [32] antibody (50 μg) was bound to 2 mg of dynabeads as per manufacturers instructions (Invitrogen). Subsequent immunocapture, washing and elution steps were performed following the manufacturer's instructions. The eluted material was vacuum-centrifuged to dryness then resuspended in 10 μL of 6 M urea and 5% acetic acid at room temperature for 10 min, and then desalted and concentrated through a C₄ ZipTip (Millipore) for mass spectrometry.

MS measurements were performed on a LTQ-Orbitrap (Thermo Scientific) operated in the positive ion mode, with the sample introduced by nano-electrospray from borosilicate capillaries (New Objective). Typical instrumental parameters included; ionisiation spray voltage, 1.5 kV; capillary voltage, 40 V; tube lens voltage, 60 V; capillary temperature; 300°C; maximum injection time, 100 ms; orbitrap mass resolution, 100000 (at m/z 400); acquisition time, 1-2 min. Spectra were deconvoluted and analysed using Qual Browser v.2.0 software (Thermo Scientific).

SELDI-TOF-MS analysis was also performed on TBS soluble material as previously described [9]. Immunocapture was performed using affinity-purified W0-2 (epitope: Aß5-8) [32] antibody coupled to ProteinChip PS10 arrays (Bio-Rad).

For separation of Aß based on peptide hydrophobicity [33] bis/bicine urea-PAGE analysis was performed as previously described [9] but modified for a 20 × 20 cm Protean II xi system (BioRad). Affinity purified 4G8 (epitope: Aß18–22, Signet Laboratories) primary antibody was used at 1 μg/ml.

For comparison of Aß levels ~1000 adults were collected in S-basal [34] in triplicate, then frozen in liquid-N₂. Samples were then extracted in 3 volumes of urea buffer (7 M urea, 2 M Thiourea, 4% w/v CHAPS, 1.5% w/v dithiothreitol and 50 mM Tris pH 8.0) disrupted via sonication, and then centrifuged at 16,500 × g for 10 min. A 10 μl sample of the supernatant was added to 3 μl of loading buffer (10% v/v glycerol, 250 mM Tris pH 8.5, 2% w/v SDS, 0.5 mM EDTA and 0.2 mM Orange-G) and reduced with 1/10 volumes of 0.5 M dithiothreitol. Samples were then heated at 70°C for 10 min, mixed and centrifuged at 13,000 g for 1 min. Material was resolved via Tricine-SDS-PAGE (16%/6M Urea) [35] then transferred to nitrocellulose membranes, boiled for 3 min (via a microwave oven) in PBS pH 7.4 and blocked for 1 h at room temperature in 0.5% (w/v) skim milk. Membranes were probed overnight at 4°C with or 6E10 (epitope: Aß4–9, Sigma) at 1 μg/ml as indicated. Blots were re-probed with anti-α-tubulin (Sigma T6074, 1:10000) to standardize total protein loading. Standard enhanced chemiluminescence was then performed [9].

C. elegans were washed in S-basal [34], fixed overnight in 10% (v/v) Neutral Buffered Formalin (NBF) at 4°C, embedded in agar (2% w/v in phosphate buffered saline) blocks and then fixed again in 10% NBF overnight. Following processing of the agar blocks into paraffin, 5 μm sections were prepared, deparaffinised and treated with 90% formic acid (FA) prior to Aß immunohistochemistry with a 1:200 dilution of 1E8 mouse monoclonal (SmithKline Beecham) antibody (epitope: Aß18–22). Antibody binding sites were detected with a peroxidase labelled streptavidin biotin system (Dako K0675) with a 3,3’-diaminobenzidine tetrahydrochloride (DAB) chromogen (Dako) resulting in a brown reaction product. Samples were counter stained with Harris Haematoxylin solution (Amber Scientific).

A modified procedure of Styren et al[12] was used, where a mixture of 5-formylsalicylic acid (2.2 mmol), p-xylylenediphosphonic acid tetraethylester (1 mmol) and potassium tert-butoxide (10 mmol) in anhydrous dimethylformamide (10 mL) was stirred at 40°C for 16 h. The reaction mixture was cooled to room temperature and poured into ice-water to give yellow precipitate which was isolated by filtration. The yellow solid was further washed with diethyl ether and dried to give 265 mg of the required product. ¹H NMR (500 MHz , d₆-DMSO): d 7.97 (d, J = 2 Hz, 2H), 7.79 (dd, J = 8.5, 2 Hz, 2H), 7.56 (s, 4H), 7.25 (d, J = 16 Hz, 2H), 7.11 (d, J = 16 Hz, 2H), 6.96 (d, J = 8.5 Hz, 2H). MS/EI: 403 (M + 1).

The differential changes in X-34 fluorescence between freshly refolded and fibrillar Aß_1-42 was confirmed by obtaining emission (excitation: 350 nm) and excitation (emission: 490 nm) spectra with a Flexstation 3 plate reader (Molecular Devices) equipped with monochromators, using an average of 15 reads and an integration time of 2 seconds (Additional file 4: Figure S3).

Thioflavin T (ThT) staining of 10% Neutral Buffer Formalin-fixed samples [28] and X-34 in vivo staining in live C. elegans samples [36] were performed as previously described, using a Leica DM2500. Immunohistochemistry on whole C. elegans was performed using 6E10 (epitope: Aß4–9) antibody and counter staining of nuclei using 4',6-diamidino-2-phenylindole (DAPI) was performed using standard protocols [10].

All populations were cultured at 20°C and developmentally synchronized from a 4 h egg-lay. At 64-72 h post egg-lay (time zero) individuals were shifted to 20°C or 25°C, and body movement assessed over time as indicated. Nematodes were scored as paralysed if they failed to complete full body movement (i.e a point of inflection traversing the entire body length) either spontaneously or touch-provoked. Proportion of individuals not paralysed were calculated and confidence intervals determined without a correction for continuity [37]. Comparisons of proportions were made using a 2-tailed Z-test. Experiments were replicated as indicated.