Genetic Mechanisms of Coffee Extract Protection in a Caenorhabditis elegans Model of b-Amyloid Peptide Toxicity

Epidemiological studies have reported that coffee and/or caffeine consumption may reduce Alzheimer’s disease (AD) risk. We found that coffee extracts can similarly protect against b-amyloid peptide (Ab) toxicity in a transgenic Caenorhabditis elegans Alzheimer’s disease model. The primary protective component(s) in this model is not caffeine, although caffeine by itself can show moderate protection. Coffee exposure did not decrease Ab transgene expression and did not need to be present during Ab induction to convey protection, suggesting that coffee exposure protection might act by activating a protective pathway. By screening the effects of coffee on a series of transgenic C. elegans stress reporter strains, we identified activation of the skn-1 (Nrf2 in mammals) transcription factor as a potential mechanism of coffee extract protection. Inactivation of skn-1 genetically or by RNAi strongly blocked the protective effects of coffee extract, indicating that activation of the skn-1 pathway was the primary mechanism of coffee protection. Coffee also protected against toxicity resulting from an aggregating form of green fluorescent protein (GFP) in a skn-1–dependent manner. These results suggest that the reported protective effects of coffee in multiple neurodegenerative diseases may result from a general activation of the Nrf2 phase II detoxification pathway. EPIDEMIOLOGICAL studies have indicated that coffee consumption may be protective in Alzheimer’s (AD) (Barranco Quintana et al. 2007) and Parkinson’s diseases (PD) (Hu et al. 2007). Although a universal consensus on the cause(s) of Alzheimer’s disease has not been reached, the best-supported hypotheses posit that accumulation of the b-amyloid peptide (Ab) is central to the induction of Alzheimer’s disease pathology. The strongest evidence for this claim is the identification of germ line mutations in the gene encoding amyloid precursor protein (APP), (Chartier-Harlin et al. 1991; Murrell et al. 1991), or in the presenilin genes involved in cleaving Ab from APP (reviewed in Cruts et al. 1996) in familial cases of early-onset AD. The pathological similarities between familial and sporadic AD, along with the results of many studies of transgenic mice engineered to overexpress Ab in the brain, argue for a causal role of Ab accumulation in all forms of AD. Intramuscular accumulation of Ab is also observed in inclusion body myositis (IBM), a severe myopathy that can be phenocopied by overexpression of APP in transgenic mice (Sugarman et al. 2002). Models of Ab toxicity can potentially be used to investigate the biological mechanisms underlying the protective effects of coffee consumption in Alzheimer’s disease. In this context, caffeine administration has been reported to reverse cognitive impairment and amyloid b (Ab) levels in transgenic Alzheimer’s disease model mice (Arendash et al. 2009). To examine the protective effects of coffee in a genetically tractable model, we investigated the effects of aqueous coffee extracts in a C. elegans model of Ab toxicity (Link et al. 2003). In this model, temperature upshift induces expression of human Ab42 in body wall muscle. This induction of Ab42 leads to a highly reproducible paralysis phenotype, in which all induced animals become paralyzed within 28 hr of temperature upshift. This model has previously been used to investigate gene expression changes in response to Ab accumulation (Link et al. 2003) and the role of specific genes (Fonte et al. 2008; Hassan et al. 2009) and autophagy (FlorezMcClure et al. 2007) in countering Ab toxicity. Here we employ this model to investigate candidate genes and pathways that might play a significant role in the protective effects of coffee. MATERIALS AND METHODS C. elegans strains used in this study: CB1301 unc-54(e1301) I CB66 unc-22(e66) IV DA465 eat-2(ad465) II CL4176 smg-1(cc546) I; dvIs27 [pAF29 (Pmyo-3TAb42) 1 pRF4] X CL437 ced-7(n1892) III; ced-5 (n1815) IV; mec-4 (u231) X CL2166 dvIs19(Pgst-4TGFP) III Supporting information is available online at http:/ www.genetics.org/ cgi/content/full/genetics.110.120436/DC1. Corresponding author: Institute for Behavioral Genetics, University of Colorado, Campus Box 447, Boulder, CO 80309. E-mail: [email protected] Genetics 186: 857–866 (November 2010) CL2070 dvIs70[pCL25(Phsp-16.2TGFP) 1 pRF4) V SJ4006 zcIs4(Phsp-4TGFP) V TJ356 zIs356(daf-16TGFP) CL685 ldIs001(skn-1TGFP) CL6176 smg-1(cc546) I; dvIs19(Pgst-4TGFP) III; 1/nT1(unc d;let-? IV;V); dvIs27 [pAF29 (Pmyo-3TAb42) 1 pRF4] X CL6180 smg-1(cc546) I; dvIs19(Pgst-4TGFP) III; skn-1(zu67) IV/nT1(unc d;let-? IV;V); dvIs27 [pAF29 (Pmyo-3TAb42) 1 pRF4] X CL6222 smg-1(cc546) I; dvIs19(Pgst-4TGFP) III; skn-1(zu169) IV/nT1(unc d;let? IV;V); dvIs27 [pAF29 (Pmyo-3TAb42) 1 pRF4] X CL2337 smg-1(cc546) I; dvIs38 [pCL60 (Pmyo-3TGFPT degron/long 39 UTR) 1 pRF4]. Coffee extract preparation: We employed an aqueous extraction protocol originally developed in the Pallanck lab (Trinh et al. 2010). Coffee beans (Starbucks House Blend, caffeinated or decaffeinated) were ground for 3 min in a standard coffee grinder, and then an 18.4% (w/v) slurry of grounds in deionized water was boiled for 30 min before removing grounds with a French press. Extracts were filter sterilized by passage through a Nalgene Fast PES Filter Unit, a 75-mm diameter membrane (0.2 mm pore size), and stored a 2 . Media preparation: Small (5.7 cm) Petri plates containing 10 ml nematode growth media agar (Wood 1988) were supplemented with coffee extract or purified caffeine (Sigma) after agar was solidified and allowed to diffuse throughout the agar before addition of bacteria. Caffeine concentrations were based on the caffeine concentration reported for Starbucks caffeinated drip coffee or from direct mass spectroscopy measurements of decaffeinated coffee extracts (Trinh et al. 2010). Plates were spotted with Escherichia coli strain OP50 or, for feeding RNAi, strain HT115 transformed with a skn-1 dsRNA-expressing plasmid or vector-only control (Kamath et al. 2003). In all experiments worms were exposed from hatching to the coffee extracts or caffeine. Quantification of paralysis kinetics: Staged populations of CL4176, CL6176, CL6180, CL6222, or CL2337 transgenic worms were prepared by synchronous egg lay and induced to express Ab as third-stage larvae by upshift from 16 to 25 as previously described (Link et al. 2003). All paralysis plots were done in triplicate with an average of 129 worms per strain and condition. Plots shown in Figures 1 and 4–7 are representative; all experiments were independently replicated (see supporting information, Table S1). Statistical analysis of paralysis curves was performed using a paired log rank survival test (Peto and Peto 1972) implemented in Statistica. Feeding RNAi interference knockdown of skn-1: skn-1 expression was knocked down by feeding RNAi interference using the corresponding E. coli from the Ahringer RNAi feeding library (clone verified by DNA sequencing). Using standard RNAi feeding protocols (Kamath et al. 2003) we observed that one generation of exposure to the skn-1 RNAi clone was insufficient to induce the 100% embryonic lethality phenotype observed for skn-1(zu169) homozygotes. We therefore developed a two-generation RNAi exposure protocol in which third larval stage worms were placed on the skn-1 (or control vector only, VO) RNAi plates and allowed to grow until the second day of adulthood at 16 . These treated adults were then used for a synchronous egg lay on RNAi plates to generate populations for paralysis quantitation after upshift. Under these conditions, control populations that were not upshifted grew to adulthood but laid 100% dead eggs. Microscopy: DIC and epifluorescence images were acquired on a Zeiss Axiophot compound microscope equipped with a computer-controlled Z-drive and software from Intelligent Imaging Innovations. Photoshop software (Adobe) was used to globally adjust brightness and contrast of digital images and to fuse DIC and epifluorescence images (Figures 4C and 7B). Quantification of Pgst-4TGFP expression: Eggs of CL2166 were collected by hypochlorite, hatched overnight at 16 , and first-stage larvae (L1s) were placed onto NGM or 10% decaffeinated coffee plates and allowed to grow to L4 stage at 16 . Worms were harvested 1000 worms/ml in 1X S-Basal, and sorted using COPAS Biosort 250 Worm Sorter (Union Biometrica Inc., Harvard Biosciences). Length, optical absorbance, and integrated fluorescence intensity at 488 nm (GFP) were factors in the worm sorting. A total of 100 worms were sorted per condition. All sorting was done at room temperature. Quantitative RT–PCR: Ab mRNA levels were quantified as previously described (Link et al. 2003) using an ABI Prism 7000 thermocycler using a TM of 55 and the following primers: forward primer 59 CTTTCTGGCACCAGCAGGTAC and reverse primer 59 CTTGCAGACTTCTCGCTGCTAG. Ab mRNA levels were normalized to reference genes cdc-42, pmp-3, and Y45F10D.4 as described in Hoogewijs et al. (2008). Immunoblotting: Worms were harvested in deionized water with protease inhibitor cocktail (Sigma, P 2714) added, and then snap frozen. Worms were boiled in sample buffer (1X protease inhibitor cocktail, 62 mm Tris pH 6.8, 2% SDS, 10% glycerol, 4% BME) for 10 min, put on ice, and then centrifuged 1 min at 14,000 g. Supernatant was quantitated via Bradford assay (Pierce, 23238). Just before loading, samples were boiled for 5 min in sample buffer with dye (62 mm Tris pH 6.8, 2% SDS, 10% glycerol, 4% BME, 0.0005% BPB). Samples were run at 180 Von Nu PAGE 4– 12% Bis-Tris Gel (Invitrogen, NP0321) using MES SDS Running Buffer (Invitrogen NP0002). ECL DualVue Western markers (GE Healthcare, RPN810) were used as size reference. Gel was transferred to 0.45mm supported nitrocellulose (GE Osmonics, WP4HY00010) using 20% methanol, 39 mm glycine, and 48 mm Tris base. Transfer conditions were 21 V, 108 min. Blots were visualized by Pouceau stain and then boiled for 3 min in PBS. Blots were blocked in TBS-Tween 1 5% milk (100 mm Tris 7.5, 150 mm NaCl, 0.1% Tween-20). Ab was detected with 6E10 (Covance, SIG-39320) at 1 mg/ml; secondary anti-mouse IgG peroxidase conjugate (Sigma, A5906). GFP was probed with 11E5 (Quantum Biotechnologies, AFP5001) at 0.5 mg/ml; secondary anti-mouse IgG (same as above). After stripping, blots were probed with CTSF64 (gift of Blumenthal Lab) at 1/7500; secondary antibody anti-rabbit IgG peroxidase conjugate ( Jackson Immunoresearch Laboratories, 211-032171). Secondary HRP-conjugated antibodies were developed in ECL Plus (Amersham, RPN2132). CstF-64 immunoreactivity was used as an internal control for immunoblots because preliminary experiments indicated that accumulation of actin, a more commonly used internal control, was actually altered by loss of skn-1 activity.