Skip to main content

Recommendations from the NIH AD Research Summit 2018

The NIH Alzheimer's Disease (AD) Research Summits are key strategic planning meetings tied to the implementation of the first goal of the National Plan to Address Alzheimer's: to treat and prevent AD by 2025. They bring together a multi-stakeholder community, including government, industry, academia, private foundations, and patient advocacies to formulate an integrated, translational research agenda that will enable the development of effective therapies (disease-modifying and palliative) across the disease continuum for the cognitive as well as neuropsychiatric symptoms of AD.

The 2012 and 2015 AD Research Summits delivered recommendations that served as the basis for developing research implementation milestones detailing specific steps and success criteria for NIH and other stakeholders toward the development of effective treatment for and prevention of AD. The milestones span the entire AD research landscape, including basic, translational, clinical, and health services research, and serve as the basis for the development of the NIH Alzheimer's Disease Bypass Budget.

The goal of the 2018 AD Summit was to feature progress toward achieving the AD research implementation milestones and to continue the development of an integrated multidisciplinary research agenda necessary to enable precision medicine research and accelerate the development of successful therapies for AD.

The program agenda was organized around seven sessions:

  1. Novel Mechanistic Insights into the Complex Biology and Heterogeneity of AD
  2. Enabling Precision Medicine for AD
  3. Translational Tools and Infrastructure for Predictive Drug Development
  4. Emerging Therapeutics
  5. Understanding the Impact of the Environment to Advance Disease Prevention
  6. Advances in Disease Monitoring, Assessment, and Care
  7. Building an Open Science Research Ecosystem to Accelerate AD Therapy Development

More than 80 leading experts on Alzheimer's disease and other complex diseases joined forces as speakers and co-chairs, and, through a multi-step, iterative process, developed a series of new research recommendations that addressed the topics of each Summit Session and the overarching programmatic themes of the Summit: 1) enabling precision medicine research to understand disease heterogeneity; 2) enhancing research rigor, reproducibility, and translatability; and 3) enabling rapid translational learning through open science systems and incentives.

The new recommendations expand the 2012/2015 research framework; they will serve as the basis for updating the AD/ADRD research implementation milestones and help guide both the public and private sectors toward meeting the research goals set forth in the National Plan to Address Alzheimer's Disease, a national strategy aimed at identifying effective interventions to treat and prevent Alzheimer's by 2025.

SESSION ONE: Novel Mechanistic Insights into the Complex Biology and Heterogeneity of AD

  1. Promote studies that integrate epidemiologic, genomic, and basic research to understand the mechanisms underlying individual differences in the risk, age-of-onset, progression, clinical presentation, and therapeutic response. Place increased emphasis on:
    • under-represented and minority populations
    • patients with extreme phenotypes, including extremes of age-of-onset of AD and atypical presentations of AD
       
  2. Accelerate research focused on understanding the mechanisms by which genetic variants (including APOE) discovered through genome-wide association studies and sequencing studies, influence AD risk. To this end, ensure rapid and broad sharing of large-scale genetic/genomic data, similar to the open-science, data-sharing model of the Accelerating Medicines Partnership for AD (AMP-AD) program.
     
  3. Determine and catalogue the effect(s) of AD genetic variants by integrating genomic data with other multi-omics data from brain, peripheral tissues, and induced pluripotent stem cell (iPSC) cellular models from well-phenotyped cohorts.
     
  4. Understand the cell-specific vulnerability to amyloid-beta, tau, and other pathogenic insults by integrating single-cell molecular data/networks with existing tissue-based data/networks ideally obtained from the same individuals.
     
  5. Enhance research on neuron-glia, glia-glia, and other intercellular interactions in the central nervous system (CNS) to understand their role in the development and propagation of AD-related processes (pathologic, electrophysiologic, and molecular).
     
  6. Support the development of integrative models of cellular interactions in the CNS underlying brain aging and transition to AD/neurodegeneration and ensure their rapid and wide dissemination to the research community.
     
  7. Establish infrastructure to develop standardized and deeply phenotyped in vitro model resources, including iPSC-based and primary cells, brain slice, and organoid models.
    • establish the translational validity of these in vitro models to recapitulate the molecular/network perturbations identified in the individual (human or animal) from which the in vitro model was generated
    • ensure rapid and broad distribution of the cell-based and organoid research models, data, and analytical methods for use in basic research and therapy development, similar to the open-science/open source principles of the Model Organism Development and Evaluation for Late-Onset Alzheimer's Disease (MODEL-AD) consortium
       
  8. Continue and expand efforts to:
    • discover new pathways leading to synaptic and neural damage and exploit these to develop novel targets for disease treatment
    • determine structural variations of pathogenic peptides collected from well-phenotyped, diverse cohorts to inform the development of structure-specific imaging agents and inhibitors with therapeutic potential
    • integrate research on neurodegeneration with research on fundamental mechanisms of aging to understand the mechanism(s) of vulnerability and resilience in AD across all levels of biologic complexity (from cellular to population level)
    • understand the interaction between peripheral organ systems function (vascular, immune, metabolic, microbiome, etc.) and the CNS in the context of aging by integrating human and animal model research
    • identify the roles of the different arms of the CNS and peripheral immunity in AD to distinguish the immune and inflammatory processes that are detrimental versus beneficial
    • discover and understand disease mechanisms that are common between AD and other neurodegenerative diseases and leverage these for therapy development
       
  9. Create mechanisms, incentives, and policies for independent replication of basic research findings to increase their translatability.
     
  10. Increase transparency in reporting and reproducibility of research findings by enforcing rapid sharing of raw and processed data, analytical methods, and details of experimental design.

SESSION TWO: Enabling Precision Medicine for AD

  1. Ensure that epidemiologic studies represent the current and future projected population trends.
     
  2. Continue to establish new cohorts that include participants across diverse socioeconomic backgrounds at increased or decreased risk of dementia and incorporate collection of novel clinical data (imaging, personal wearables, and sensors for in-home monitoring).
     
  3. Enable precision medicine research for AD by accelerating deep and longitudinal molecular phenotyping across diverse cohorts and special populations. Prioritize molecular profiling in cohorts from special populations and patients with atypical AD and ADRD (Down syndrome, early- and late-onset familial AD, early-onset non-autosomal dominant AD), ethnic minorities, and other under-represented groups.
     
  4. Accelerate the generation of high-quality, multi-omic data as a community resource and maximize data accessibility and usability for downstream analyses. This requires longitudinal as well as postmortem collection of biosamples from brain and peripheral tissues.
     
  5. Support more robust incorporation of systems immunology into AD by augmenting current and future human cohorts with more advanced immunological profiling (e.g., CyTOF).
     
  6. Develop and disseminate best practices for the collection, processing, and multi-omics molecular profiling of human and animal model biosamples. Ensure adoption of best practices across different research groups and institutions to enable harmonized analyses of data collected at the national and international level.
     
  7. Enable access to electronic health records data and their integration with clinical and molecular data to build person-specific predictive models of disease and wellness.
     
  8. Support the mining and integrative analyses of existing and newly generated large-scale molecular clinical data including medical records data and other patient-relevant data to define disease subtypes and to identify biomarkers and molecular signatures for these subtypes. Use the metabolome as a functional readout for other "omics" data to delineate pathways implicated in disease initiation and progression and to identify disease subtypes.
     
  9. Expand biorepository infrastructure to enable storage and wide distribution of clinical data (such as imaging, data from wearables, environmental exposure data, etc.) and biosamples for deep molecular phenotyping, collected from diverse populations including iPSCs and other cells, e.g., skin, fibroblasts, myoblasts, etc. These resources should include biorepositories for samples and clinical data from ongoing clinical trials to support multi-omic data generation.
     
  10. Enable a system biology approach to decipher the complex role of the microbiome in brain aging and AD/ADRD. Robust and rigorous study of the microbiome will require that relevant biosamples, high-quality molecular data, and analytical tools are made available as a community resource.
     
  11. Support deep single-cell molecular profiling, using cells from participants from diverse cohorts to complement whole brain and peripheral tissue molecular profiling conducted in the same participants to aid novel target and biomarker discovery for precision medicine.
     
  12. Create large human single- and multi-gene perturbation datasets that can be used to evaluate computational predictions of key genes and molecular network structures and their links to AD endophenotypes.
     
  13. Support the development of co-culture systems utilizing 3D organoid-like spheres to recapitulate complex interactions in a dish and develop novel ex vivo models of "cognition in a dish" and "ancestry in a dish" for precision medicine research.
     
  14. Support studies that align blood and brain "omics" from longitudinal mouse and other preclinical models with human blood and brain "omics" to enable cross-species dynamic modeling of the disease trajectory.
     
  15. Enable comprehensive analyses of "clinical" and deep "omics" profiles in panels of genetically diverse mice and flies, which better model human genomes to develop preclinical resources with higher predictive validity than standard strains and to accelerate the discovery of translationally relevant disease mechanisms and treatment strategies.
     
  16. Use drugs as probes to reveal disease heterogeneity. To this end:
    • enhance current and future clinical trials by molecular profiling of minimally invasive specimens (serum, peripheral monocytes, microbiome, etc.) and use these molecular signatures as indicators of responsiveness within subgroups of patients
    • re-evaluate legacy clinical trials by making biosamples available for molecular profiling
       
  17. Democratize the use of large-scale molecular and clinical data by building computational infrastructure for sustainable data storage and processing that will be accessible to researchers worldwide. This should include support for Web-based tools and interfaces to allow data mining and analyses of summary results.
     
  18. Liberate data that are necessary for precision medicine research, such as large-scale genetics data, by bringing them within an open-science ecosystem such as the one created by AMP-AD and associated consortia.

SESSION THREE: Translational Tools and Infrastructure for Predictive Drug Development

  1. Maximize the translational potential of genetics research by:
    • expanding genetics efforts in AD/ADRD as open-science programs, similar to the Psychiatric Genomics Consortium and type 2 diabetes (T2D) consortia. This will require a large-scale effort to genotype and sequence AD and ADRD cohorts
    • developing an open-access aggregated genetic data portal for Alzheimer's and related disorders (akin to the AMP-T2D portal) to facilitate genetic data access and analyses by the wide research community and to encourage commercial entities to share genetic data
    • developing a complete understanding of the gene-basis and directionality for genetic variant association so gene networks can be annotated with causality as tools for drug discovery
       
  2. Continue and enhance existing large-scale, open-science molecular profiling efforts by:
    • expanding the bandwidth for the generation and analysis of high-quality molecular data (genetics and epigenetics, transcriptomics, proteomics, metabolomics, lipidomics, glycomics, exposome, etc.) from human tissues and cell lines
    • investing in the development of methods for multiscale modeling across "omics" data types
    • ensuring that diverse cohorts and special populations are prioritized in these efforts
       
  3. Support standardized, single-cell molecular profiling (genomic, proteomic, metabolomic) and open-access data infrastructure to develop a single-cell atlas for the aging brain and for AD/ADRD that can be queried by data scientists as well as basic and clinical researchers. The selection of samples for single-cell profiling needs to be optimized to allow integration of molecular profiles with existing clinical and whole-tissue molecular data (from brain and peripheral tissues) generated on the same individuals.
     
  4. Create a repository of fully characterized, quality controlled, and fully sequenced reference iPSC lines accessible to the wide research community that can serve as:
    • a standard control for healthy genotypes and for generating edited-isogenic lines carrying specific AD risk variants
    • a source of cell lines from individuals from diverse cohorts with AD-related phenotypes and specific naturally occurring mutations/risk alleles
    • a resource to test and provide reference data on protocols to generate all CNS cell types
       
  5. Invest in the development of high-throughput systems of AD-relevant cells and organoids driven by robotics with digital readouts (such as high content imaging) to leverage reinforcement learning techniques for more data-driven target discovery/screening.
     
  6. Expand the existing open-source/open-science translational infrastructure for next-generation AD animal models development by:
    • supporting the generation of large longitudinal, multi-omic data on existing and newly developed transgenic mouse models necessary to build "molecular network maps" that connect molecular attributes of AD across mouse models and humans
    • developing new transgenic mouse models with humanized immune systems expressing AD-risk-related human genes for use in preclinical drug development and to gain a better understanding of the role of the immune system in risk and progression of AD/ADRD
       
  7. Invest in exploring the comparative biology and integrated physiology of non-murine models to better understand pathophysiology (plaques, tangles, and other features) and mechanisms of risk and resilience.
     
  8. Support the development of genome-scale metabolic models to capture metabolic complexity and the heterogeneity of transitions from healthy to pathologic brain aging.
     
  9. Generate a neurodegenerative disease analyses common to promote analyses of large-scale molecular data and to stimulate novel methods development.
     
  10. Expand support for quantitative systems pharmacology approaches that couple biological network and pathway analyses with mechanistic systems models and integrate data from disparate sources (e.g., preclinical and clinical; in vitro, ex vivo, and in vivo; acute and chronic intervention) to enable predictive drug development. These efforts should ensure full transparency of data and analytical methods development and encourage precompetitive academic-industry collaborations.
     
  11. Support cross-disciplinary training in all aspects of quantitative systems pharmacology, spanning experimental and clinical work to various types of modeling and simulation.
     
  12. Provide support for high-cost capital equipment/core facilities and staff training to make high-throughput technologies such as molecular profiling, cryo-electron microscopy, advanced human brain imaging, and other emerging technological capabilities available for wide use.
     
  13. De-risk novel candidate targets by supporting the development of high-quality, target-enabling packages (TEPs) that can serve as starting points for drug discovery campaigns or as research tools to understand the biology of "dark targets." These enabling tools should include: purified target proteins/crystal structures, biochemical assays suitable for functional characterization and for compound screens, validated antibodies, cell-based assays for target modulation, and potent and specific small molecules or biologics. To maximize the utility and translational impact of these tools, all data and reagents should be made available to the wide community of researchers with no restriction on use.
     
  14. Incorporate emerging generative type artificial intelligence cheminformatics models to enable inference of novel chemical compound classes into AD drug discovery data sets/projects.
     
  15. Accelerate the development of the next-generation CNS imaging ligands and biofluid molecular signatures targeting a variety of disease processes (neuroinflammation, bioenergetic/metabolic compromise, oxidative stress, synaptic pathology) that can be used as research tools or developed into diagnostic, prognostic, theragnostic, or target engagement biomarkers. These reagents should be made available as open-source research tools for target validation and as enabling tools for predictive drug development.

SESSION FOUR: Emerging Therapeutics

  1. Continue to support and enable access to integrative in silico approaches for use in target prioritization, drug discovery and development, drug repurposing, and combination therapy development.
     
  2. Develop an AD connectivity map (CMap), whereby genes, drugs, and disease states are connected by common gene and other "omics" expression signatures in disease-relevant cell types (iPSC neurons, microglia, astrocytes, mixed cell cultures, organoids).
     
  3. Develop an academic/industry partnership where industry partners can submit failed Phase II/III compounds for molecular profiling to enable computational drug repositioning analysis for AD.
     
  4. Improve the regulatory environment for repurposing of drugs and combination therapies to obtain longer periods of exclusivity similar to those for orphan indications.
     
  5. Invest in the expansion of therapeutic modalities including natural products, gene therapy, antisense oligonucleotides, cell therapy, etc.
     
  6. Provide support and expertise to enable robust clinical development plans for novel therapeutics that use precision medicine principles:
    • mechanistic biomarker-informed dosing paradigms for subpopulations
    • biomarkers/outcome measures including telemedicine and digital/wearable technologies
    • the inclusion of a companion diagnostic
       
  7. Support the development and access to novel statistical approaches (e.g., Bayesian methods, modeling, and simulations) for more efficient trial design.
     
  8. Evaluate the utility of intermediate, prognostic endpoints and platform trials such as umbrella, basket, etc., to enable faster and cheaper proof-of-concept trials.
     
  9. Encourage industry utilization of NIH funding to increase the number of therapeutic agents and therapeutic mechanisms tested in Phase II proof-of-mechanism trials.
     
  10. Enable access to data and associated biosamples and biomarkers from completed, federally and privately funded clinical trials to advance understanding the heterogeneity of disease mechanisms and treatment response.
     
  11. Enable wide access to electronic medical records and invest in better electronic phenotyping of AD through the application of machine learning methods that can help define AD electronic phenotypes in a data-driven manner.
     
  12. Increase investment in clinical trials that robustly test a variety of lifestyle interventions by incorporating deep molecular profiling and digital/wearable technologies for tracking responsiveness.
     
  13. Create synergies between research programs and translational infrastructure for AD and other neurodegenerative diseases as well as rare diseases that share pathophysiology, pathology, or symptoms with AD to garner mechanistic insights that can be leveraged therapeutically.
     
  14. Expand existing cross-disciplinary training modules to develop the next generation of drug developers for AD/ADRD.
     
  15. Institute an expedited review track for applications focused on drug discovery, preclinical, and clinical drug development for Alzheimer's disease to mitigate erosion of patent life and to allow a viable commercialization strategy for NIH-supported molecules.
     
  16. Ensure that review panels for translational research applications are staffed with adequate cross-disciplinary expertise to evaluate all aspects of drug discovery and development for a variety of therapeutic targets and therapeutic modalities.
     
  17. Continue and expand existing efforts to educate patients, caregivers, physicians, and other stakeholders on the importance of patient enrollment into natural history and non-interventional studies and trials that could yield biomarkers for precision medicine research.
     
  18. Enable public-private partnerships to foster the use of existing cohorts and to develop new cohorts that can inform drug development for the preclinical stage of AD.

SESSION FIVE: Understanding the Impact of the Environment to Advance Disease Prevention

  1. Quantify the exposome in existing and new AD cohorts to gain a more precise measure of environmental exposure factors and their relationship to AD risk and trajectories of disease progression. These efforts should employ:
    • a life-course approach across diverse populations (e.g., race, ethnicity, immigration status, geographical region, education, age, gender)
    • methods aimed at understanding how ancestry, race/ethnicity, and socioeconomic disparities interact with exposome factors to modulate AD risk
       
  2. Incorporate new environmental and behavioral sensors and cognitive assessment technologies in clinical research designed to detect the impact of gene-by-environment (e.g., poverty, discrimination, toxic and non-adverse exposure, education) and gene-by-behavior (e.g., physical activity, diet) interactions on AD risk and resilience.
     
  3. Support the inclusion of measures of AD-related phenotypes and environmental exposures in non-AD cohorts to enable new discovery research and to accelerate cross-validation of discoveries made in AD cohorts. These efforts should include the development of novel, open-source computational methods for data mining and data integration.
     
  4. Support gut and oral microbiome molecular profiling (metagenome, meta-transcriptome, metaproteome, and metabolome) across diverse cohorts and in clinical trials to better understand disease heterogeneity, gene-environment interactions, and differential responsiveness to treatment. These efforts should include in silico approaches to study the gut-microbiome-brain axis in AD and ADRD.
     
  5. Expand efforts to understand the mechanistic links between sleep/circadian disruption and AD and related dementias at multiple levels (epigenetic, gene expression, proteomic, neuronal, network, systems) to identify new targets and approaches for AD prevention.
     
  6. Continue and expand research on the role of social and psychosocial factors, on AD risk and resilience to risk from a life-course perspective, to inform intervention strategies and health policy across diverse ancestral groups and socioeconomic backgrounds. Prioritize research that can reveal the heterogeneous mechanistic pathways of disparities in health burden of AD and can test whether causal pathways to AD differ across disparities populations.
     
  7. Harness new high-throughput analytical techniques in research on model organisms and in pseudo-randomized experiments with human subjects to investigate the causal role of the exposome in the etiology of AD.
     
  8. Enhance lifestyle interventions by incorporating:
    • digital health devices to track adherence to treatment and treatment response
    • deep, longitudinal multi-omics profiling to understand mechanisms of action and person-specific trajectories of response
       
  9. Support studies that focus on multi-modal interventions that combine both pharmaceutical and lifestyle interventions that are personalized to the individual.
     
  10. Identify new risk reduction strategies by remediating negative environmental exposures (e.g., lack of neighborhood safety, discrimination, poverty, lack of access to medical care, poor diet, low physical activity, social isolation, toxicants, inadequate air/light quality, poor sleep); these interventions should incorporate deep phenotyping to understand determinants of response and inform precision medicine for disease prevention.
     
  11. Support research and commercialization efforts to develop wearables that can increase rigor in measuring environmental exposures as well as intervention dose.
     
  12. Develop culturally sensitive guidelines for social and behavioral interventions for AD and ADRD prevention to inform study design, cohort selection, intervention delivery and dosing, and outcomes assessments. Incentivize efforts to increase harmonization of protocols across studies to enhance research rigor and reproducibility.
     
  13. Develop cross-disciplinary training in AD, aging, epidemiology, neuropsychology, environmental health, genomics, and data science to enhance the workforce needed for research on gene-environment interactions in AD and AD health disparities.

SESSION SIX: Advances in Disease Monitoring, Assessment, and Care

  1. Build end-to-end secure, high-frequency data-capture platforms to enable continuous monitoring of research participants across the disease trajectory; these capabilities should include remote methods for consenting and collection of multiple health indices.
     
  2. Develop and deploy citizen science methods to engage diverse and under-represented populations including the full spectrum of age, race/ethnicity, and technological sophistication/access in optimizing approaches to disease monitoring and data sharing.
     
  3. Test innovative digital data collection platforms that include validation methods of objective health indices and existing disease biomarkers and generation of digital biomarkers for health and disease.
     
  4. Develop wearable/sensor data standards and support the development of automated scripts that convert incoming data into a standardized format to enable their integration with existing data.
     
  5. Maintain an open-access national repository of evidence-based tools for cognitive assessment that have been validated in diverse participants that are iteratively updated and refined with use. This resource should include modifiable data collection forms and analytical tools.
     
  6. Incentivize academic-industry partnerships to build and make publicly available novel data harmonization and analytical solutions.
     
  7. Develop and disseminate new pervasive computing assessment methods that can be embedded in existing and new clinical research studies.
     
  8. Provide support to existing and new clinical research studies for data storage, data curation, and data management to enable rapid and broad, high-fidelity data sharing.
     
  9. Ensure that studies generating rich molecular and digital datasets on well-phenotyped cohorts make all traditional, derived, and raw data and all data coding files associated with any published studies available for secondary use in discovery and replication research.
     
  10. Support digital health workshops and training programs for researchers from academia and industry.
     
  11. Alter institutional practices that impede team science and data sharing in academia by:
    • defining and quantifying team science activities as new metrics of scientific contribution for career advancement
    • implementing pilot programs to test new incentive structures that promote and reward collaborative research and data-sharing efforts
    • showcasing existing open-science programs
       
  12. Engage regulatory agencies to address unique issues related to use of digital technologies for biomarker discovery and validation.
     
  13. Develop criteria and methods for screening, testing, and validating new technologies for disease monitoring.
     
  14. Convene a national, multi-sector committee to survey, inform, and troubleshoot barriers to data access and sharing.
     
  15. Ensure that review panels for applications using/developing digital technologies have the adequate technical expertise and up-to-date knowledge of digital platforms.

SESSION SEVEN: Building an Open Science Research Ecosystem to Accelerate AD Therapy Development

  1. Provide support to modernize the data management/data governance and data infrastructure of existing and legacy cohorts to maximize data accessibility and usability.
     
  2. Acquire and share biomarker and other patient-level data and biological samples from all early- and late-stage drug trials and lifestyle intervention trials for AD/ADRD to clarify their predictive and theragnostic value, find surrogate endpoints, and elucidate the therapeutic mechanism of action.
     
  3. Develop standard consenting language, simplified data, and material transfer agreements, and support data and sample sharing platforms to ensure the rapid, widespread, appropriate, and productive use of data and samples.
     
  4. Explicitly mandate the open distribution of publicly and philanthropically funded computational and experimental research tools for use/reuse across the research community to accelerate independent evaluation of research findings.
     
  5. Establish and maintain cloud-based resources for data storage, data sharing, and compute to enable reproducible computational analyses; support cloud-based compute costs so that researchers are not limited by their access to local compute services.
     
  6. Establish mechanisms to harmonize methods and outcomes across all research domains including standard operating procedures and best practices for all stages of analysis so that they can be reused, evaluated, and expanded.
     
  7. Promote and support early sharing of research observations through preprint servers and other methods and data resources.
     
  8. Promote interaction among researchers across different disciplines and disease fields to clarify disease mechanisms and therapeutic targets, to advance the discovery of repurposed and investigational drug treatments and to maximize their relevance to the treatment and prevention of AD.
     
  9. Incentivize investigators to independently evaluate high-value therapeutic hypotheses through directed funding and through unrestricted access to reagents and tools.
     
  10. Maximize the comparability, interoperability, and synergy across complementary NIH research programs (All of Us, AMP-AD, AMP-Parkinson's Disease, BRAIN Commons, etc.) to accelerate the development of precision medicine for AD/ADRD.

NOTE: The recommendations were considered and adopted by the National Advisory Council on Aging on May 22 and 23, 2018.