Anti-amyloid-β oligomer activity of human phenyl-γ-valerolactone metabolites and the use of yeast model systems to study amyloid neurodegenerative pathologies

Abstract

Age-related neurodegenerative diseases (NDs) include a number of sporadic, and less frequently genetic, pathological conditions affecting the central or the peripheral nervous system, causing cognitive decline and motor symptoms. As the lifespan globally increases, NDs are becoming one of the most pressing medical and societal challenge worldwide, but no effective disease-modifying therapy is currently available. Despite their differences, NDs are associated with a decline in cellular proteostasis and the accumulation of insoluble deposits of misfolded proteins, which aggregate in highly ordered cross-β fibrils named amyloids. Even if amyloid deposits are the main histopathological hallmark of NDs, a series of evidences points to the soluble, prefibrillar amyloid oligomers as the ultimate neurotoxic species. Oligomers can act both extracellularly and intracellularly and induce various cytotoxic effects, including oxidative stress, mitochondrial damage, alterations in endoplasmic reticulum, vesicular trafficking and apoptotic signalling, eventually leading to neuroinflammation and cognitive impairment. The aim of this thesis work is using Saccharomyces cerevisiae, an organism that has been extensively validated to study human diseases, as system to model amyloid oligomer toxicity, in order to identify molecules with anti-oligomer activity and investigate molecular pathways involved in amyloid neurodegenerative pathologies. This PhD thesis consists of two parts. Part I introduces the development of a new yeast model of Alzheimer’s disease (AD) oligomeropathy. Since yeast proved to be quite tolerant to the expression of the human amyloid peptide (Aβ42), whose oligomerization is thought to be causal in AD, I resorted to an artificial polypeptide (β23) that forms amyloid-like oligomers and shows a dose-dependent toxicity in yeast cells. Recently, flavonoids and flavan-3-ols, polyphenolic compounds deriving from dietary sources associated with a reduced incidence of NDs, are gaining attention for their neuroprotective properties, but they are characterized by a generally low bioavailability, as they undergo a broad host- and gut microbiota-assisted metabolism that complicates the identification of the most relevant bioactive species. I thus addressed this issue by examining the ability of phenyl-γ-valerolactones (PVLs), the main circulating flavan-3-ol metabolites in humans, to prevent amyloid-oligomer toxicity. Several PVLs, and particularly the monohydroxylated 5-(4’-hydroxyphenyl)-γ-valerolactone metabolite [(4’-OH)-PVL], relived β23 oligomer-induced cell death in the ‘AD-like’ yeast model and in a human cell line. (4’-OH)-PVL also interfered with Aβ42 oligomer formation in vitro and remodelled preformed AβOs into amorphous, non-toxic aggregates. Conversely, it did not show any significant effect on Aβ fibrillization. Importantly, treatment of AβOs with (4’-OH)-PVL prior to brain injection reduced recognition memory deterioration as well as neuroinflammation in a mouse model of AβO-induced memory impairment. In Part II, a comparative transcriptomic analysis is carried out in the AD-like model and in the yeast models for Parkinson’s disease (PD), Huntington’s disease (HD) and amyotrophic lateral sclerosis (ALS), which are based on the inducible expression of the human amyloidogenic proteins underlying the pathogenesis of these disorders. I evaluated the transcriptional response at different time points, in order to identify both early and late changes associated with amyloid aggregate formation or aimed at their detoxification. RNA-sequencing revealed a fast and differentiated response of S. cerevisiae to the formation of different amyloid structures and the AD-like and PD models, along with a more toxic phenotype, presented the highest number of differentially expressed genes. In all strains, closely connected pathways associated with protein aggregation and proteotoxic stress resulted activated, which included responses to unfolded protein, oxidative stress and deficiency of essential metal ions like iron, corroborating previous findings about the role of these pathways in amyloid toxicity. Mitochondrial dysfunctions were evidenced in both the AD-like and PD models, but with an almost opposite gene modulation, and different alterations in mitochondrial morphology pointed to different implication of the mitochondrial fission and mitophagy processes. A strong decrease in transcripts for proteins involved in translation was characteristic of the PD model, while the downregulation of genes participating to mitotic cell division was a signature of the AD-like model. Even more, the activation of meiotic pathways that could induce haploid cells to sporulate were index of a profound damage of cellular division and DNA replication processes. Lastly, preliminary observation suggested the relevance of protein disordered regions in a context of cellular stress induced by amyloid-aggregate formation.