Comparative Gene Expression Studies in AD and Other Conditions

More than 98% of the human genome is composed of noncoding DNA. It is postulated that much of the genetic variation that influences risk of common and complex human diseases such as Alzheimer's disease (AD) and other neurodegenerative diseases will reside in noncoding though functional, regulatory regions of the genome and will alter disease risk by affecting gene expression.

If this hypothesis is correct, many functional, regulatory genetic variants or their proxies will associate with both disease risk and gene expression levels. Further, gene transcripts, which are regulated by functional disease risk variants, may show differential levels in subjects that have the disease either clinically or prodromally in comparison with control subjects.

In support of this hypothesis, the lab has determined that genetic variants near many genetic risk loci for late-onset Alzheimer's disease (LOAD) associate with brain levels of nearby genes. Additionally, genetic variants that have strong association with brain gene expression levels are enriched for human disease-associated variants, including those for LOAD.

The lab has several ongoing projects that are aimed at harnessing this dual information on gene expression levels and disease risk. The goal is to identify and characterize novel genes, transcripts and genetic risk variants for LOAD. These projects include:

1. Target pathway discovery in AD using transcriptomics

The team has already determined that the strongest genetic risk variants for AD — detected in genomic screens — also associate with the brain expression levels of genes located close to the strong risk variants. These findings suggest that genetic variants that confer risk of AD can be found by modifying the expression levels of genes in the brain.

Many genes exist in different forms, called isoforms, which result in different protein products with diverse functions. Identifying the precise changes in brain levels of these gene isoforms enhances the understanding of risk mechanisms in AD and enables future novel therapeutic approaches.

In this project, the lab aims to uncover specific isoform level changes for the strongest Alzheimer's risk genes using deceased-donor brains of people who had AD. Additionally, researchers will identify the biological consequences of altering specific isoform levels for some of the risk genes by decreasing or increasing their levels in cellular models, as well as investigating functional outcomes such as cell survival.

The goal is to uncover changes to the gene isoform level that implicates novel pathways in AD that may guide future drug discovery efforts and revolutionize therapeutic approaches.

2. A systems approach to targeting innate immunity in AD

As part of a multi-institutional project co-led by Nilüfer Ertekin-Taner, M.D., Ph.D.; Todd Golde, M.D., Ph.D., at the University of Florida; and Nathan Price, Ph.D., at the Institute for Systems Biology, the laboratory is leading projects to identify potential therapeutic targets and novel biomarkers through using multiomics and computational systems biology approaches.

Using whole-genome sequencing, next-generation RNA sequencing and genome-wide genotypes, the team has identified:

• Novel pathways in Alzheimer's disease
• Potential regulatory variants in innate immune genes
• Distinct transcriptional networks implicated with cell-specific pathologies and characterized disease risk association
• Transcriptional profiles of genes in innate immune pathways

The vast amount of multiomics data has been shared with the research community and utilized to nominate novel therapeutic targets for Alzheimer's disease.

In the second five years of this work, the Mayo Clinic team aims to identify precise therapeutic targets, their potential mechanism of action and beneficial therapeutic direction through the addition of epigenetic profiling (H3K27ac and ATACseq) for existing brain samples (700 brain samples from 350 subjects) and performing multi-scale computational analyses followed by collaborative functional validations in model systems.

Related publications

• Allen M, Wang X, Burgess JD, Watzlawik J, Serie DJ, Younkin CS, Nguyen T, Malphrus KG, Lincoln S, Carrasquillo MM, Ho C, Chakrabarty P, Strickland S, Murray ME, Swarup V, Geschwind DH, Seyfried NT, Dammer EB, Lah JJ, Levey AI, Golde TE, Funk C, Li H, Price ND, Petersen RC, Graff-Radford NR, Younkin SG, Dickson DW, Crook JR, Asmann YW, Ertekin-Taner N. Conserved brain myelination networks are altered in Alzheimer's and other neurodegenerative diseases. Alzheimer's & Dementia. 2018;14:352.
• Carrasquillo MM, Allen M, Burgess JD, Wang X, Strickland SL, Aryal S, Siuda J, Kachadoorian ML, Medway C, Younkin CS, Nair A, Wang C, Chanana P, Serie D, Nguyen T, Lincoln S, Malphrus KG, Morgan K, Golde TE, Price ND, White CC, De Jager PL, Bennett DA, Asmann YW, Crook JE, Petersen RC, Graff-Radford NR, Dickson DW, Younkin SG, Ertekin-Taner N. A candidate regulatory variant at the TREM gene cluster associates with decreased Alzheimer's disease risk and increased TREML1 and TREM2 brain gene expression. Alzheimer's & Dementia. 2017;13:663.
• Allen M, Wang X, Serie DJ, Strickland SL, Burgess JD, Koga S, Younkin CS, Nguyen TT, Malphrus KG, Lincoln SJ, Alamprese M, Zhu K, Chang R, Carrasquillo MM, Kouri N, Murray ME, Reddy JS, Funk C, Price ND, Golde TE, Younkin SG, Asmann YW, Crook JE, Dickson DW, Ertekin-Taner N. Divergent brain gene expression patterns associate with distinct cell-specific tau neuropathology traits in progressive supranuclear palsy. Acta Neuropathologica. 2018;136:709.
• Conway O, Carrasquillo MM, Wang X, Bredenberg JM, Reddy JS, Strickland SL, Younkin CS, Burgess JD, Allen M, Lincoln SJ, Nguyen T, Malphrus KG, Soto AI, Walton RL, Boeve BF, Petersen RC, Lucas JA, Ferman TJ, Cheshire WP, van Gerpen JA, Uitti RJ, Wszolek ZK, Ross OA, Dickson DW, Graff-Radford NR, Ertekin-Taner N. ABI3 and PLCG2 missense variants as risk factors for neurodegenerative diseases in Caucasians and African Americans. Molecular Neurodegeneration. 2018;13:53.
• Van der Lee SJ, Conway OJ, Jansen I, Carasquillo MM, Kleineidam L, van den Akker E, Hernández I, van Eijk KR, Stringa N, Chen JA, Zettergren A, Andlauer TFM, Diez-Fairen M, Simon-Sanchez J, Lleó A, et al. A nonsynonymous mutation in PLCG2 reduces the risk of Alzheimer's disease, dementia with Lewy bodies and frontotemporal dementia, and increases the likelihood of longevity. Acta Neuropathologica. 2019;138:237.
• Allen M, Carrasquillo MM, Funk C, Heavner BD, Zou F, Younkin CS, Burgess JD, Chai HS, Crook J, Eddy JA, Li H, Logsdon B, Peters MA, Dang KK, Wang X, Serie D, Wang C, Nguyen T, Lincoln S, Malphrus K, Bisceglio G, Li M, Golde TE, Mangravite LM, Asmann Y, Price ND, Petersen RC, Graff-Radford NR, Dickson DW, Younkin SG, Ertekin-Taner N. Human whole genome genotype and transcriptome data for Alzheimer's and other neurodegenerative diseases. Scientific Data. 2016;3:160089.

3. Multiomics approaches for biomarker discovery in AD

The overarching goal of these studies is to identify genetically driven premorbid blood transcript biomarkers for AD. Researchers in Dr. Ertekin-Taner's laboratory postulate that transcript level changes that occur due to regulatory genetic variants, which also influence disease risk, can be detected prior to the development of clinical AD. The underlying premise is that since disease risk variants are expected to be more frequent in the high-risk preclinical AD population compared with controls, the downstream regulatory effects of transcript level changes also can be detected preclinically.

Indeed, the strongest known genetic risk factor for LOAD, apolipoprotein ε4 (APOE ε4), has higher frequency in subjects with mild cognitive impairment (MCI). This is a high-risk state considered to be a prodrome for AD, especially for amnestic MCI (aMCI). Although APOE ε4 reflects coding and not regulatory polymorphisms, if these observations apply to other disease risk variants, then regulatory variants that influence AD risk and also associate with levels of gene transcripts will result in transcript level differences. These include study participants with AD versus controls and high-risk, preclinical subjects (such as aMCI) versus controls.

If this hypothetical model is correct, it provides a strong rationale for the use of transcript levels as genetically driven biomarkers in preclinical AD. This hypothesis is being tested in a project being conducted by Dr. Ertekin-Taner's laboratory within the umbrella of the Accelerating Medicines Partnership — Alzheimer's Disease (AMP-AD) consortium. In this project, presented at the 2019 National Institute on Aging (NIA) Alzheimer's Association International Conference (AAIC) in Los Angeles, California, the team is measuring whole transcriptome RNA and microRNA levels collected from plasma of elderly participants who have a diagnosis of either incident MCI or incident AD, or who are cognitively normal. Plasma samples collected longitudinally from each patient are being analyzed to identify three things:

• Preclinical biomarkers for AD
• Biomarkers of rate of decline
• Biomarkers of clinical impairment

4. Integrative translational discovery of vascular risk factors in aging and dementia

Dr. Ertekin-Taner's laboratory is leading studies focused on using multiomics measures and systems biology approaches to identify novel genes and pathways associated with vascular risk factors for AD. This project is part of the Molecular Mechanisms of the Vascular Etiology of Alzheimer's Disease (M2OVE-AD) consortium.

Genome-wide genotypes, RNA sequencing transcriptome measures and reduced representation bisulfite sequencing (RRBS)-based DNA methylation are being collected from brain tissue and blood samples in two cohorts: a postmortem cohort of AD cases scored for cerebral amyloid angiopathy (CAA), and a prospective antemortem cohort with brain neuroimaging measures of microbleeds and infarcts, along with rich cognitive and clinical data. Histone acetylation measures (H3K27ac) are also being collected for the postmortem cohort. Using genome-wide genotypes, we have identified association of variants with CAA in the postmortem cohort (platform presentation, 2019 Alzheimer's Association International Conference) and microbleeds and infarcts in the antemortem cohort (2018 Alzheimer's Association International Conference). The team aims to integrate these initial genetic findings with other omics measures collected in these samples to further characterize observed associations. Gene expression and epigenomics measures will also be used independently to identify additional genes and pathways that are associated with measures of vascular pathology.

Brain tissue is composed of multiple cell types: neurons, astrocytes, microglia, endothelia and oligodendroglia, and sub-types of these. The team is conducting a study within this broader project, where we aim to generate gene expression profiles from these different cell types, using neurosurgical brain tissue. These cell types are being further characterized at the single-cell level using 10X genomics technology. The overall goals of this work are to determine which genes are predominantly expressed in certain cell types in the brain, identify and characterize cellular subtypes, and use this information to guide and improve interpretation of bulk tissue expression profiling studies. Initial results from this work were presented as a platform presentation at the 2019 Alzheimer's Association International Conference.

Related publications

• Carrasquillo MM, Reddy JS, Allen M, Reyes DA, McDonnel SK, Graff-Radford J, Kosel ML, Gregory JD, Nguyen T, Lincoln SJ, Malphrus KG, Crook JE, Jack Jr CR, Lowe VJ, Knopman DS, Petersen RC, Kantarci K, Bu G, Ertekin-Taner N. Genome-wide association study of cerebral microbleeds and infarction identifies candidate genes for vascular risk in dementia. Alzheimer's & Dementia. 2018;14:P1118.

5. Harnessing molecular networks of resilience for therapeutic discoveries in AD

This project is led by Dr. Nilüfer Ertekin-Taner in collaboration with Hu Li, Ph.D., in the the Department of Molecular Pharmacology and Experimental Therapeutics at Mayo Clinic's campus in Minnesota.

Although protection from and resilience to Alzheimer's disease (AD) constitute a fundamental aspect to understanding AD pathophysiology, this is a relatively understudied area and the molecular basis of resilience to AD is largely unknown. We aim to fill this knowledge gap by studying cohorts resilient to AD pathology and cognitive decline and leveraging existing and generating novel molecular data to identify biological pathways that confer resilience to AD, validating these pathways in experimental model systems, and identifying potential AD therapeutics that promote resilience.

Dr. Ertekin-Taner's laboratory is leading these studies by integrating whole-genome sequence and whole-transcriptome gene expression data from blood (antemortem cohort) and brain (postmortem cohort) to identify novel genes and pathways associated with resilience to the development of AD. Candidate resilience genes and pathways identified in the antemortem and postmortem cohorts will be validated in induced pluripotent stem cells (iPSC)-based models by Dr. Bu's laboratory, and evaluated as potential drug targets utilizing state-of-the-art, systems-based pharmacogenomics approaches by Dr. Hu Li's group, in close collaboration with Dr. Ertekin-Taner's laboratory. These studies are expected to uncover networks and molecules of resilience and to validate findings of other large-scale efforts to translate these discoveries to viable candidate drugs for AD.