On the Path to Precision in Alzheimer’s Disease Management
September 21, 2022 Justyna Lisowska
Alzheimer’s disease (AD) is the most prevalent form and/or cause of dementia – a group of age-related neurodegenerative disorders severely impacting cognition, motor function, and social ability due to progressive dysfunction or loss of neurons in the central nervous system. Today, 55 million people worldwide live with AD or other forms of dementia. As the global population and life expectancy rise, this number is expected to reach 78 million by 20301. Interestingly, although the percentage of people with AD increases with age and affects mainly people over 65 years, almost 10% of patients are between their 30s and mid-50s1. This represents a substantial socio-economic burden with the estimated global annual cost of about US$1 trillion2.
In the current absence of effective disease management strategies, a research paradigm shift is needed. Moving away from a traditional, clinical symptom-focused research approach to one that seeks to decipher disease-associated molecular mechanism(s), could encourage breakthroughs in AD diagnosis and drug development. Critical to this transition is the adoption of digital tools like Genedata Profiler®, which simplifies knowledge extraction from a variety of high-dimensional data for informed decision-making while overcoming diverse data complexities.
A Challenging Quest for An Effective Treatment
While a lot of research has been done in recent years to understand the molecular mechanisms underpinning AD, this condition remains incurable as studies continuously fail to translate into effective disease-modifying treatments. Today, most pharmacological interventions available for patients only slightly improve cognitive and behavioral functions. While few competing mechanisms of pathogenesis have been proposed, one theory has dominated the neuroscientific field for years guiding treatment development: amyloid β (Aβ) plaque accumulation. Several investigational drugs targeting Aβ plaque biosynthesis or metabolism—predominately orally administered immunotherapies, or small molecule compounds— entered clinical trials but failed to demonstrate therapeutic benefit in large-scale studies. Even an Aβ-directed monoclonal antibody Aduhelm (aducanumab), available on the market since June last year, received FDA approval under highly controversial circumstances3. While the drug reduces the buildup of the amyloid plaques in the brains of AD patients, trials have shown inconsistent results regarding its efficacy in reducing cognitive decline. In the context of recent findings questioning the veracity of published data4 in the landmark research paper on amyloid plaques, one can wonder whether the efforts have focused on the right mechanisms and targets. Although amyloid plaques have been found in both, early and late-onset patients, their presence may be symptomatic and result from other currently overlooked upstream molecular pathways. Given the complex, multifactorial nature of AD, applying a more comprehensive treatment approach targeting more than a single molecular driver might be beneficial to patients.
The Biological Complexity of Alzheimer’s Disease
The pathophysiology underlying AD is not easy to grasp. As a multifaceted progressive condition, AD develops over years with heterogenous clinical manifestations and symptoms that overlap with other types of dementia or dementia-like diseases. Genome sequencing studies have greatly contributed to elucidating the genetic determinants of the disease accountable for around 70% of all cases2. Yet, while several genes have been identified as key risk factors for both familial early-onset (Amyloid Precursor Protein (APP), Presenilin-1 (PSEN-1), Presenilin-2 (PSEN-2)) and sporadic late-onset of the disease (apolipoprotein E (ApoE)), the genetic landscape of AD is much more complex2. AD originates from the integrated effect of hundreds of genetic variants, of different levels of causality and prevalence, converging on multiple molecular pathways. Also, even though both sporadic and familial forms of AD share clinical symptoms and neuropathological hallmarks of the disease, such as extracellular amyloid plaques and intraneuronal Tau tangles, little is known of any converging pathophysiological processes preceding clinical manifestations. Since other environmental and behavioral factors in addition to age, sex, and other comorbidities contribute to the disease's onset and progression, it is important to understand their relationship to facilitate the development of preventive and therapeutic measures.
Mechanistically Driven, Biomarker-Guided Translational Research Approach
To generate novel, effective interventions for AD, a more comprehensive research approach needs to be undertaken. Genome sequencing used to unravel genomic signatures should be followed by high-throughput molecular profiling and systems biology analyses to identify abnormalities on other levels of biological organization (epigenomics, transcriptomics, proteomics, metabolomics, etc.). This integrative multi-omics approach would allow researchers to mechanistically characterize the modulation of disease-associated gene regulatory programs and decipher signaling pathway convergence points - key disease nodes for targeted treatment development.
Understanding disease mechanisms on multiple molecular levels could also help to identify distinct, actionable disease modules/endotypes. This would provide the basis for the discovery of more objective, clinically relevant disease biomarkers for patient selection and stratification during clinical trials, and improved patient diagnosis. By selecting patients with a particular molecular signature who are more likely to benefit from the trial investigational compound, we can better demonstrate its efficacy. In addition, replacing the use of clinical manifestations with objective measurements of pathophysiological changes in AD diagnostics could help to distinguish AD patients from patients with other comorbidities that share common symptoms. This would allow for more effective and targeted disease management. It could finally enable the diagnosis of patients before clinical onset thus, faster administration of suitable treatment. Enrolling patients in trials at an earlier stage of disease progression, the prodromal/preclinical stage, also provides a better chance for an investigational, disease-modifying drug to reverse the targeted mechanism, increasing the clinical trial success rate.
In the past few years, intensive research efforts accompanied by advances in investigational technologies (e.g., neuroimaging, Single Molecule Array, Mass Spectrometry, etc.,) have led to the discovery of several AD biological indicators of proven confirmatory value in disease diagnosis. Although these biomarkers have emerged from different sources and various modalities (Cerebral Spinal Fluid (CSF), blood, or Positron Emission Tomography (PET)), they all focus on the predominant disease neuropathological readouts: Aβ plaques and Tau-based neurofibrillary tangles5. Clinically relevant biomarkers leveraging other pathophysiological processes underlying AD (such as neurodegeneration, synaptic dysfunction, neuroinflammation, immune system alteration, metabolic dysfunction, oxidative stress, etc.), that may precede these symptomatic hallmarks, are still in development. Combining pre-clinical data from different disease models— for example, animal, patient pluripotent stem cell-derived neurons or organoids or post-mortem patient brain samples— could help elucidate these overlooked mechanisms and decipher new objective disease pathophysiological measures of diagnostic value. Such biomarkers could then be applied to biorepository samples from failed trials to demonstrate their clinical validity and guide future studies.
Recent progress in sensor technologies and the widespread use of mobile and wearable devices represent an opportunity to develop next-generation, digital biomarkers. Capable of detecting subtle sensory and motor changes occurring several years before clinical manifestations of the disease, these biomarkers can serve as effective early diagnostic tools5. In addition, by continuously collecting objective data in real-life settings, digitally captured biomarkers could also be used to track disease progression and monitor patient response to treatment during a clinical trial and beyond.
This multidimensional approach: in-vivo measurement of AD pathophysiological features through genetic tests, fluid biomarkers, or neuroimaging coupled with digitally enabled assessment of discrete clinical symptoms would increase diagnostic precision and translate into more effective treatment.
Digitalizing Translation Research
To enable the transition from clinically based approaches to mechanistically driven, biomarker-guided, and digitally supported AD management, we need to embrace new digital technology. This is because such an approach leverages huge amounts of heterogenous complex datasets. To derive biomedically relevant, actionable insights from molecular research, clinical, and real-world evidence studies, data needs to be federated, curated and made FAIR (Findable, Accessible, Interoperable, and Reusable).
Genedata Profiler, a Bio-IT innovative practices award-winning data integration and analytics platform has been purposely built to address translational research data challenges. Located in the cloud, the platform serves as a single point of access to the variety of data incorporated from diverse instruments, disparate locations (including internal and external public and/or public databases), and drug development stages. Given the high failure rate of neurological clinical trials, re-using data from previous and/or ongoing studies is crucial to understand their limitations and improve future decision-making. To enhance data discoverability and facilitate retrieval, once gathered under one roof, such data needs to be cataloged and labeled with consistent annotations. This is ensured by the highly advanced data and metadata management functionality of Genedata Profiler.
When handling large and complex datasets from omics or longitudinal clinical/real-world evidence studies, the risk of generating false conclusions is high. Therefore, before any analysis occurs, raw data needs to be processed, normalized and quality controlled. To enable different datasets to be comparatively analyzed, multimodal data also requires integration and harmonization. All these data-handling steps required for subsequent applications can be streamlined and automated in Genedata Profiler with intuitive point-and-click workflows.
Finally, once the data of interest has been found and prepared for its intended purpose, it needs to be connected to high-performance analytics to seamlessly explore different specific research questions and/or validate the hypothesis. Applied to multi-omics datasets, artificial intelligence (AI)-based analytical approaches such as machine learning can be of great assistance to decipher molecular patterns, sub-classify diseases, and identify new biomarkers or therapeutic targets. In addition to molecular and imaging data, AI-driven algorithms can also support the analysis of longitudinal multimodal data generated by sensors and digital devices, allowing the discovery of clinically relevant digital biomarkers. Genedata Profiler provides a wide spectrum of innovative analytical and data visualization capabilities for novel insight generation. This is ensured by its inbuilt statistical and AI-driven analysis tools, as well as through enhanced interconnectivity with external expert analytics and data visualization solutions (such as RStudio, Power BI, or Spotfire).
Finally, while working with digitally captured real-world evidence data as well as clinical trial data, one needs to ensure the protection of sensitive patient information. Fine-grained data access control as well as automatic patient consent revocation ensures data security and privacy in the Genedata Profiler system.
A mechanistically driven translational research approach could be a real game-changer to spur innovation in AD management. Yet, to make it a reality, one needs to overcome the associated data complexities. Bringing disparate, voluminous, and heterogenous data together, and integrating and analyzing the right data with the right tools to extract biomedically valid conclusions is not trivial. This cannot be realized without the support of cloud-based digital solutions providing high-performance data storage and computing environment. Insights unlocked with these tools could improve patient diagnosis and accelerate the development of efficacious AD treatment.