Description

Alzheimer’s Disease (AD) is a progressive and incurable neurodegenerative disease. The disease results in the gradual degeneration and eventual death of neurons causing complications with movement and mental function (Gao and Hong 2008). AD is characterized by plaques of amyloid-beta (Aβ) peptides in the cerebral cortex of the brain. Aβ peptides range from 37 to 49 amino acid residues in length (Chen et al. 2017). Aβ 1-40 and Aβ 1-42 are the most common species in the cerebral cortex of AD patient samples, with Aβ 1-40 being the most abundant (Seynnaeve et al. 2018; Spies et al. 2010). Aβ accumulation and aggregation are thought to lead to a series of neurodegenerative events (Novo et al. 2018). However, the factors triggering the aggregation of Aβ proteins have not been completely characterized. Furthermore, exactly how Aβ aggregation leads to neurodegeneration remains unclear (Wolfe and Cyr 2011).

The role of epigenetic mechanisms in the development of AD has become a subject of intense investigation. Epigenetics invokes changes in phenotype resulting from changes on chromatin structure and accessibility occurring without changes in the underlying DNA sequence (Esposito and Sherr 2019). Epigenetic mechanisms, such as histone post-translational modifications (PTMs), may provide insight into the development of AD. We hypothesize that the toxic effect of Aβ aggregation might be related to its association with altered histone marks. In fact, AD animal models and post-mortem brain samples from patients exhibiting a decline in gene expression might be mediated via histone PTMs (Esposito and Sherr 2019).

Saccharomyces cerevisiae models have become valuable tools in neurodegenerative disease research due to the genetic and environmental manipulations possible (Seynnaeve et al. 2018). Furthermore, upon expression of disease-related human proteins, yeast recapitulate many of the features of neurodegenerative diseases and display an easily observable disease phenotype (Bennett et al. 2019; Treusch et al. 2011). In addition, yeast models are affordable and time-efficient (Bennett et al. 2019; Seynnaeve et al. 2018). A powerful Aβ 1-40 yeast model has been previously developed by Treusch et al. where Aβ 1-40 is expressed and targeted to the secretory pathway. Aβ 1-40 peptides are targeted to the endoplasmic reticulum and transit throughout the secretory pathway, thus mimicking Aβ trafficking. Remarkably, a screen for Aβ toxicity modifiers using this model revealed a known risk factor for AD, genes associated with previously discovered AD risk factors as well as novel risk factors (Treusch et al. 2011). Here, we exploit this model to characterize the histone PTM landscapes associated with Aβ aggregation (Treusch et al. 2011).

We constructed the same yeast model, but in a different mating type for consistency with our experiments in other neurodegenerative disease proteinopathies (Chen et al. 2018). S. cerevisiae were transformed with either a control vector or a vector encoding for Aβ 1-40 and grown on galactose (inducing) and glucose (non-inducing) supplemented media. In agreement with previous work (Treusch et al. 2011), expression of Aβ 1-40 modestly impaired cell growth (Figure 1A). We verified expression of the Aβ peptide by Western blotting (Figure 1B).

We analyzed the histone PTM landscape associated with the overexpression of Aβ 1-40 by way of immunoblotting. We did not enrich for any particular genomic regions, but rather looked at global modification changes throughout the whole genome. We characterized various histone methylation, acetylation, and phosphorylation sites. We focused on histone H3 and H4 as these are most abundantly modified. We also focused on modifications conserved between yeast and human. In contrast to our findings in other neurodegenerative disease yeast models (Chen et al. 2018), we did not detect any significant changes in acetylation levels on lysine 14, 18, 27, or 56 of histone H3 or on lysine 12 of histone H4 (Figure 1C). Moreover, we did not detect any significant changes in the tri-methylation levels on H3K4 or in mono-, di-, or tri-methylation levels on H3K79 or in H3S10ph levels (Figure 1C). However, we discovered a modest, but highly reproducible decrease in the levels of H3K9ac in the context of Aβ 1-40 overexpression (Figure 1D). The magnitude of this change parallels the extent of growth suppression elicited by Aβ 1-40 (Figure 1A). Furthermore, these changes detected genome-wide are likely a significant underestimate of the potential greater difference detected at specific chromatin loci. Interestingly, decreases in H3K9ac were found in transgenic AD mouse models (Currais et al. 2019). Furthermore, treatment with HDAC inhibitors in mice reverses the decrease in histone acetylation levels leading to improved memory (Lu et al. 2015). Very recently, increases in H3K9ac have been linked to AD in post-mortem human brains and fly models (Nativio et al. 2020). Despite the differences in direction of change, we interpret the overlap between various model systems and patient tissues to point a true role for H3K9ac and its associated ‘writers’, ‘erasers’, and other binding partners at various stages of AD pathology.

Epigenetic factors regulate gene expression. Histone deacetylation results in a condensed chromatin structure leading to decreased gene expression (Shahid et al. 2021). To explore if the modest changes in H3K9ac observed are sufficient to impact gene expression, we chose a straightforward approach in which we extracted and quantitated total RNA levels in Aβ 1-40 yeast and compared it to control. Increases in total RNA levels suggest an up-regulation in gene expression, while decreases in RNA levels suggest a down-regulation in gene expression. We find a ~30 percent decrease in the RNA levels in the Aβ 1-40 strain compared to control (Figure 1E). Notably, we observed this effect despite protein overexpression. Therefore, we establish that decreases in H3K9ac are accompanied by decreases in global RNA levels.

In summary, we observe a modest, but highly reproducible decrease in the acetylation on lysine 9 of histone H3 in the context of Aβ 1-40 overexpression in S. cerevisiae. Furthermore, we observe a global decrease in RNA levels in the context of Aβ 1-40 suggesting decreased gene transcription. Our findings offer additional evidence supporting a role for H3K9ac in AD pathology and allude to novel diagnostic and therapeutic epigenetic tools for AD and other neurodegenerative diseases.