Periodic Reporting for period 1 - CTS-TEs-ADprogress (Cell type-specific molecular analysis of epigenetic changes and transposable element derepression in Alzheimer's disease progression)
Okres sprawozdawczy: 2020-09-01 do 2022-08-31
Alzheimer’s disease (AD) is a major contributor to disease burden and healthcare costs worldwide. AD is usually diagnosed once symptoms like memory impairment become evident. However, at this point typical AD pathology such as Aβ plaques and cell death is already widespread, suggesting that molecular changes have occurred decades before symptom onset. With an increasingly aging population and no available treatments, it has become imperative to identify the molecular mechanisms underlying onset and progression of AD. Chronic environmental stress and age-associated changes in stress response have been associated as drivers of AD pathology. The epigenome plays a critical role in translating stress signals into a cellular response by influencing gene expression, which can either promote or inhibit cell survival. Several studies have shown that alterations in chromatin structure, including heterochromatin loss, and associated changes in gene expression contribute to neurodegeneration. In addition, neuronal death was also linked to transposable element (TE) dysregulation due to epigenetic changes, which can lead to changes in gene expression and insertional mutations due to transposition. However, our understanding of epigenetic changes at onset and during progression of AD pathology is very
limited, as current studies have two major limitations: 1) lack of cell type resolution due to use of bulk tissue samples and 2) coverage of only few or only one disease stage.
Here, single-cell RNA-seq and ATAC-seq as well as CUT&RUN will be implemented to identify cell type-specific alterations of gene expression and gene regulatory mechanisms during onset and progression of AD pathology in the APPPS1 mouse model. APPPS1 mice are a well-established AD model, which recapitulates many characteristics of preclinical AD in human patients and thereby allows correlating the identified changes with the development of specific pathological hallmarks. Focus of the analysis will be the hippocampus, which is essential for learning and memory and degenerates in AD. Samples will be collected from APPPS1 and wildtype control mice at 6 weeks, 3 months, 9 months and 18 months of age to identify changes across the lifetime of the mice. In addition, the resulting mouse data will be integrated with biomarker and genome-wide association study data from AD patients to identify clinically relevant alterations.
limited, as current studies have two major limitations: 1) lack of cell type resolution due to use of bulk tissue samples and 2) coverage of only few or only one disease stage.
Here, single-cell RNA-seq and ATAC-seq as well as CUT&RUN will be implemented to identify cell type-specific alterations of gene expression and gene regulatory mechanisms during onset and progression of AD pathology in the APPPS1 mouse model. APPPS1 mice are a well-established AD model, which recapitulates many characteristics of preclinical AD in human patients and thereby allows correlating the identified changes with the development of specific pathological hallmarks. Focus of the analysis will be the hippocampus, which is essential for learning and memory and degenerates in AD. Samples will be collected from APPPS1 and wildtype control mice at 6 weeks, 3 months, 9 months and 18 months of age to identify changes across the lifetime of the mice. In addition, the resulting mouse data will be integrated with biomarker and genome-wide association study data from AD patients to identify clinically relevant alterations.
Work performed from the beginning of the project (September 2020) included management of the APPPS1 transgenic mouse colony as at least 12 male mice of each age group (6 weeks, 3, 9 and 18 months) were required to cover the project's objectives. Priority were given of course to the oldest cohort, which was established before the beginning of the project. At the beginning of the project the cohorts were expanded to a 6 months and 12 months cohort to better cover the lifetime of mice for the phenotypic analysis with immunohistochemistry stainings. Due to the Corona Virus Pandemic breeding of mice during 2020 was quite limited and could only expanded to the required levels in 2021. Therefore, collection of the required brain samples from the APPPS1 transgenic and wildtype control mice began in summer of 2021 and commenced in autumn of 2022. Collection of brain samples continued throughout the 1 year break of the project, which allowed me to run at least some of the planned experiments described in the proposal. In total 12-15 frozen cortex and hippocampus samples for all six age groups could be collected until October 2022. In addition, cortex and hippocampus samples from wildtype mice were collected at post-natal day 21 (hippocampus development in mice completed, high level of synapse pruning occurs) and 25 months of age. These samples were included in the single nuclei RNA-seq and ATAC-seq assays to allow for comparison with the APPPS1 mice to determine if we see molecular evidence of early aging or de-differentiation.
I successfully completed sorting of nuclei and ATAC-seq library preparation in autumn of 2022. ATAC-seq libraries were sequenced at the beginning of 2023 and analysis is currently ongoing.
In the spring of 2022 (after re-start of the project) I focused my efforts on improving the scRNA-seq/SPLiT-seq as described in Objective 1 to allow assessment of transposable element expression in the SPLiT-seq dataset. I switched the analysis pipeline to use the STAR aligner STARsolo options for scRNA-seq for barcode deconvolution and determination of read counts, which allows quantification based on exons only and whole genes (exons and introns) as well as on spliced and unspliced transcripts, which allows analysis of RNA velocity. RNA velocity can be used for the analysis of time-resolved phenomena such as embryogenesis and as in my case neurodegeneration. I implemented the new analysis pipeline using published SPLiT-seq data of 100 mouse brain nuclei to have a limited dataset that can be easily monitored. Next, I applied the new analysis pipeline to small SPLiT-seq dataset of mouse hippocampus, which I had generated previously. The main idea behind this approach was to determine if I can map a SPLiT-seq-based snRNA-seq dataset onto a reference dataset created by integrating different 10X-based snRNA-seq datasets of the mouse hippocampus. To create the reference dataset, I used two datasets: 1) dentate gyrus development and 2) adult mouse hippocampus. Next I mapped my own SPLiT-seq dataset onto this reference, which confirmed previous marker-based cell type annotations (see attached Figure).
For completion of Aim 1.1 we decided to switch to the commercial SPLiT-seq kit offered by Parse Biosciences, which offers many advantages to the home-brewed version and many groups working on single cell or single nuclei RNA-seq are switching to from using the more costly and limited 10X Genomics sequencing kits. In this case I selected hippocampi (3 biological replicates each) from the following sample groups: 1) P21 C57BL/6J wildtype mice, 2) 3 months old C57BL/6J wildtype mice, 3) 9 months old C57BL/6J wildtype mice, 4) 3 months old APP/PS1 mice, 5) 6 months old APP/PS1 mice, 6) 9 months old APP/PS1 mice, 7) 18 months old APP/PS1 mice and 8) 25 months old C57BL/6J wildtype mice. Nuclei from these samples were isolated as recommended, fixed and single nuclei RNA-seq libraries were prepared according to Parse Biosciences. The eight final libraries, which should provide data from about 100,000 nuclei, were finally sequenced this September and analysis will commence as soon as possible, but after the funding period had ended.
I successfully completed sorting of nuclei and ATAC-seq library preparation in autumn of 2022. ATAC-seq libraries were sequenced at the beginning of 2023 and analysis is currently ongoing.
In the spring of 2022 (after re-start of the project) I focused my efforts on improving the scRNA-seq/SPLiT-seq as described in Objective 1 to allow assessment of transposable element expression in the SPLiT-seq dataset. I switched the analysis pipeline to use the STAR aligner STARsolo options for scRNA-seq for barcode deconvolution and determination of read counts, which allows quantification based on exons only and whole genes (exons and introns) as well as on spliced and unspliced transcripts, which allows analysis of RNA velocity. RNA velocity can be used for the analysis of time-resolved phenomena such as embryogenesis and as in my case neurodegeneration. I implemented the new analysis pipeline using published SPLiT-seq data of 100 mouse brain nuclei to have a limited dataset that can be easily monitored. Next, I applied the new analysis pipeline to small SPLiT-seq dataset of mouse hippocampus, which I had generated previously. The main idea behind this approach was to determine if I can map a SPLiT-seq-based snRNA-seq dataset onto a reference dataset created by integrating different 10X-based snRNA-seq datasets of the mouse hippocampus. To create the reference dataset, I used two datasets: 1) dentate gyrus development and 2) adult mouse hippocampus. Next I mapped my own SPLiT-seq dataset onto this reference, which confirmed previous marker-based cell type annotations (see attached Figure).
For completion of Aim 1.1 we decided to switch to the commercial SPLiT-seq kit offered by Parse Biosciences, which offers many advantages to the home-brewed version and many groups working on single cell or single nuclei RNA-seq are switching to from using the more costly and limited 10X Genomics sequencing kits. In this case I selected hippocampi (3 biological replicates each) from the following sample groups: 1) P21 C57BL/6J wildtype mice, 2) 3 months old C57BL/6J wildtype mice, 3) 9 months old C57BL/6J wildtype mice, 4) 3 months old APP/PS1 mice, 5) 6 months old APP/PS1 mice, 6) 9 months old APP/PS1 mice, 7) 18 months old APP/PS1 mice and 8) 25 months old C57BL/6J wildtype mice. Nuclei from these samples were isolated as recommended, fixed and single nuclei RNA-seq libraries were prepared according to Parse Biosciences. The eight final libraries, which should provide data from about 100,000 nuclei, were finally sequenced this September and analysis will commence as soon as possible, but after the funding period had ended.
I have build a vast collection of hippocampus and cortex samples (both frozen and fixed) throughout the lifetime of wildtype control mice and the Alzheimer's disease mouse model APPPS1, which allows systematic studies of cell type-specific changes during onset and progression of Alzheimer's disease pathology (here mainly Abeta plaques). Studies like this are usually not possible using human samples as they mostly are available from end stage disease only. While we had hoped to already have run and completed analysis of such studies, especially single nuclei RNA-seq assays as well as ATAC-seq in sorted nuclei, this was not possible due to the pandemic, which would have been of great value for the community affected by Alzheimer's disease. However, the brain collection will act as a great resource for upcoming projects of several PhD students.