BMSIP Projects 2022
2022
Building a comprehensive soil metagenome database top
PI: Jenny Bhatnagar, Department of Biology
Intern: Daniel Golden
Analyses of brain gene expression and DNA methylation in Alzheimer’s disease model mice – effects of perinatal choline nutrition top
PI: Jan Krzysztof Blusztajn PhD, Department of Pathology and Laboratory Medicine
Intern: Navin Ramanan
Alzheimer’s Disease (AD) is a progressive neurodegenerative disease and the most common form of dementia. Recent estimates suggest that almost 6 million people suffer from AD in the US. Although there has been a keen interest in finding treatments for AD, there is a tremendous lack in disease-altering treatment options. Thus, prevention of AD has become an interesting topic. Supplementation with the essential nutrient, choline, has been found to improve cholinergic functioning, enhance long-term potentiation, and improve spatial memory in rodents, and also reduce amyloid beta pathology in mouse models of AD. Additionally, choline is a methyl donor group and altering early-life methyl donor group levels has been shown to alter DNA methylation. Interestingly, changes in DNA methylation levels naturally occur with aging. These DNA methylation changes appear in a typified pattern that lead to what is known as the “methylation clock” that enables the identification of the methylation – or biological – age of a sample. Furthermore, people with AD show an acceleration of these DNA methylation changes and thus methylation age. Since we know that choline can act as a methyl donor group and choline supplementation protects against age-related cognitive decline and attenuates AD-like pathology in animal models, we hypothesize that these beneficial outcomes of choline supplementation may be mediated through changes in DNA methylation and altered patterns of brain gene expression. In order to elucidate this, we have gathered bulk RNA sequencing and DNA methylation data from several brain regions of wildtype and AD-model mice across 4 timepoints (3, 6, 9, and 12 months of age) with either choline-supplemented or control perinatal diets. RNA-sequencing has produced a robust dataset of 192 FASTQ files that will be analyzed using standard bioinformatic analysis to trim, align, normalize, count, and identify differentially expressed genes across age, diet, genotype, sex, and brain region. Pathway analysis and gene set enrichment analysis (GSEA) will also be conducted using these differentially expressed gene sets to identify biologically relevant systems that are altered. Furthermore, DNA methylation analyses will be performed on the same samples using reduced representation bisulfite sequencing and differentially methylated CpGs and regions will be identified using similar methods to those for differentially-expressed transcripts and methylation age will be generated for each mouse using methods of Stubbs et al., 2017 (PMID 28399939). Finally, correlations between differentially methylated DNA patterns and gene expression will be identified to elucidate any possible cis-acting regulatory sites in relation to age, diet, genotype, sex and brain region.
Graph database representations of clinical and ‘omics data top
PI: Adam Labadorf/Taylor Falk, Department of Neurology, BUSM and Bioinformatics Program
Intern: Merai Dandouch
Graph databases encode data as sets of entities and relationships between them. ‘omics datasets, including genomic, transcriptomic, methylomic, proteomic, etc quantify many different types of biological entities like genes, proteins, and metabolites. The entities captured by each of these data modalities partially overlap. Furthermore, ‘omics data generated from human subjects comprise a dataset where each individual has data points for each of these entities for each modality as well as clinical information. Integrating these large, complex, and diverse data is a substantial challenge. This project explores strategies for encoding ‘omics data features and clinical variables for a set of individuals who donated their brains to a local brain bank.
scRNAseq analysis of mouse and human neurons to understand early Alzheimer’s disease related pathogenesis in the entorhinal cortex top
PI: Jean-Pierre Roussarie, Anatomy & Neurobiology at BUSM
Intern: Manas Dhanuka
Preclinical stages of Alzheimer’s disease (AD) affect only specific neurons. Neurons from layer II of the entorhinal cortex (ECII), key to new memory formation, are the first ones to form neurofibrillary tangles (NFTs) and to degenerate. A subset of these, in the transentorhinal cortex (TEC) strikingly form NFTs before the rest of EC. The TEC is much more developed in humans than in mice, and virtually nothing is known about it other than its vulnerability to the disease. AD itself is a very human specific disease: while certain features of the disease appear spontaneously in some species and can be modeled in transgenic mice, the specific degeneration of EC/TEC cells only occurs in humans. We have previously studied extensively the molecular profile of ECII neurons in the mouse, and have used functional genomics to model these neurons in silico. We identified genes that could drive their vulnerability. We also used single-nucleus RNAseq in the mouse to study subpopulations of ECII neurons and potentially identify the elusive mouse homolog of TEC neurons. Lastly we recently generated a human single-nucleus RNAseq data of human EC. We think that a cross species comparison might help us identify genes that could be key to drive the pathology in human ECII neurons. We are also very interested in the possibility that TEC neurons could have represented a specific human evolutionary innovation. In this project, we will integrate our mouse and human single-nucleus RNAseq data using softwares like LIGER or Seurat to help identify in the human dataset the homologs of the mouse subpopulations.
RNA sequencing and proteomics analysis to identify molecular pathways implicated in the development of aortic aneurysms top
PI: Dr. Francesca Seta, Vascular Biology Section at the Boston University School of Medicine
Intern: Kyra Griffin-Mitchell
Cardiovascular diseases remain the leading cause of morbidity and mortality in USA. Dr Seta’s research focuses on cellular and molecular mechanisms of vascular diseases, with the goal of identifying novel therapeutic targets. Ongoing studies are examining the role of the transcription factor Bcl11b and the enzyme sirtuin-1 on the development of aortic aneurysms, abnormal aortic dilations for which there is no targeted therapies. We found that both proteins are protective against aortic aneurysms, as demonstrated in mice with Bcl11b or Sirtuin-1 deletion, which develop aortic aneurysms or ruptures when treated with the hypertensive agent angiotensin II (see Valisno et al Circ Research 2021 and Fry et al, JAHA 2015). However, the molecular mechanisms by which these proteins exert their beneficial effects is elusive. To address this, we ran both RNA sequencing and proteomics analysis on aortas from angII-treated controls and mice with either Bcl11b or sirtuin1 deletions to identify downstream transcriptomics and proteomics pathways. We would like to analyze these datasets and possibly integrate them to identify the molecular pathways involved in the formation of aortic aneurysms.
Characterizing epigenetic changes to study the mechanisms of stem cell driven epithelial tissue regeneration using RNAseq, ATACseq, HiC top
PI: Dr. Andrey Sharov, Dermatology at the Boston University School of Medicine
Intern: Go Ogata
To study stem-cell driven epithelial tissue regeneration, we need to understand how keratinocytes establish distinct gene expression programs during differentiation, and why these programs are altered in pathological conditions associated with stem cell exhausting/loss (aging) or uncontrolled expansion (cancer). Epigenetic mechanisms regulate covalent DNA/histone modifications and higher-order chromatin remodeling. Terminal cell differentiation program is maintained via interaction of the promoters of actively transcribed genes with their distal enhancer elements, which provide functional and structural frameworks for cell-specific transcription controlled by lineage-specific transcription factors. Although most genetic and epigenetic studies focus on protein-coding genes, regulation of the non-coding genome is emerging as a new critical component in skin biology. Transposable elements (TEs) constitute a large portion (>44%-55%) of the entire mouse or human genomes, respectively. We use mouse skin as primary research model. We hypothesized that distinct classes of TEs differentially contribute to the control of gene expression in epidermal keratinocytes, mediated by epigenetic regulators LSH and SETDB1 that serve as critical determinants mediating the TE silencing and preventing pro-inflammatory responses in the epidermis. The short-term goal of the summer research program is to characterize epigenetic changes in mouse keratinocytes isolated from Lsh and Setdb1 knockout mice. We are planning to perform bioinformatic analysis of our own NGS available data using RNA-seq, ATAC-seq, WGBS, ChIP-seq, Chromatin Conformation Capture (HiC) pipelines.
Using scRNAseq to investigate the use of mesenchymal stromal-cell derived extracellular vesicles in mitigating pathogenic signaling in human oligocortical spheroids top
PI: Dr. Ella Zeldich, Anatomy & Neurobiology at Boston University School of Medicine
Intern: Raghad Yamani
Extracellular vesicles (EVs) are released by nearly every cell type and are an important structure in inter-cellular communication. Abnormal EV signaling is found in many conditions including ischemia, Alzheimer’s Disease (AD) and Down Syndrome (DS). However, EVs from stem cells from healthy animals have recently emerged as a possible therapeutic intervention to address a variety of neurological conditions. Mesenchymal stromal cell-derived (MSCs) extracellular vesicles from the bone marrow of young healthy monkeys contain microRNAs and proteins and previous studies have shown that MSC-EV treatment mitigates inflammation and oxidative stress, promotes myelination, and improves functional recovery in a rhesus monkey model of cortical injury. EVs have also been shown to reduce AD pathology in mouse models by promoting anti-inflammatory processes and slowing the progression of AD. While AD currently effects over 6 million people in the United States, individuals with DS are disproportionately affected by early onset AD. Therefore, investigating the efficacy of MSC-EVs as a potential therapeutic to mitigate AD like pathology in DS is critical. Accordingly, the current study aims to explore the application of EVs on 3D human brain models of DS. We generated human oligocortical spheroids (OLS) containing neurons, astrocytes and oligodendrocytes allowing investigation of the effects of the EVs in human, physiologically relevant conditions. We used DS-derived OLS generated from isogenic induced pluripotent stem cell (iPSCs) lines to evaluate the efficacy of EV treatment in DS. Trisomic OLS display significantly higher levels of amyloid beta (Aβ40 and Aβ42) depositions in both the soluble and insoluble fractions. Additionally, trisomic OLS are consistently smaller than their euploid counterparts, and have elevated levels of cleaved-caspase 3 ( CC3) detection indicating more cell death. When treated with EVs, trisomic OLS demonstrated greater preserved cortical volume, significantly decreased levels of Aβ40 and Aβ42 in both fractions, and significant reduction in cell death compared to the untreated trisomic OLS. These results suggest that EVs alleviated the AD-related pathology in DS-derived OLS. Evaluation of the markers of cortical layer neurons demonstrated significantly higher counts of neurons expressing deep and superficial layer markers, suggesting that EVs contributed to greater preserved cortical volume of trisomic OLS by promoting neurogenesis and alleviating trisomy-induced deficits. Our studies show for the first time the efficacy of MSC-EVs in mitigating DS and AD-related cellular phenotypes and pathological depositions in human OLS. However, the molecular networks and signaling pathways mediating these events are still unknown. Therefore, we attempted to perform more mechanistic studies using single cell RNA sequencing (scRNA-seq) experiment. We have submitted 6 samples: 2 euploid samples, 2 trisomic samples not treated with EVs and 2 trisomic samples treated with EVs for scRNA-seq using 10xGenomics platform. The OLS were successfully dissociated, and submitted for a submitted for a single cell capture using 10X Genomics Chromium® single cell (v3 Chemistry) platform, the libraries were prepared and sequences. The project will include bioinformatics analysis of the data generated data through scRNA-seq. The project will include but not limited to read alignment, dimension reduction and clustering, principal component analysis, differential gene expression analyses using Seurat pipeline, potential pseudotime analysis.
Long non-coding RNA expression in post mortem brains in alcohol use disorder top
PI: Huiping Zhang, Departments of Psychiatry and Medicine, Section of Biomedical Genetics
Intern: Janvee Patel
Chronic alcohol consumption can result in alcohol use disorder (AUD), but the underlying mechanisms are not well understood. We hypothesized that long-term alcohol use could alter the expression of coding genes (or mRNAs) and noncoding genes (such as long noncoding RNAs) in the brain, thus promoting neuroadaptation to alcohol. The project aims to identify differentially expressed long noncoding RNAs (lncRNAs) in specific brain regions of subjects with AUD. In addition, the interactive effect of mRNAs and lncRNAs on the risk of AUD will be studied. Ribo-depleted RNA-Seq data from 192 AUD and matched healthy control post mortem human brain tissue samples spanning 8 brain regions will be analyzed.