NEWS & EVENTS

Opportunities and Challenges in Establishing a Comprehensive Assessment of Cognitive and Dementia-like Phenotypes Associated with Existing Models of Vascular Disease

September 20 - 21 , 2021
Virtual

Description

Clinical investigations have established that vascular-associated medical conditions are significant risk factors for various kinds of dementia. And yet, we are unable to associate certain types of vascular deficiencies with specific cognitive impairments. The reasons for this are many, not the least of which are that most vascular disorders are multi-factorial and the development of vascular dementia in humans is often a multi-year or multi-decade progression.

In order to better study vascular disease and its underlying causes, the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH) has invested considerable resources in the development of vascular animal models that recapitulate various aspects of human disease. Many of these models (mainly in the mouse) are genetic, frequently using single-gene mutations to examine the role of specific proteins in vascular function. Collectively, these models could serve as useful tools to associate specific vascular signaling pathways with specific neuro-cognitive impairments.

Some of these mouse vascular models have been examined for cognitive effects, but generally it is challenging to compare results across such models because of differences in methodology and, in some cases, lack of experimental rigor and reproducibility. Measurement of cognitive properties in animals involves a complex set of processes and variables that, while well-characterized and well-understood within the neuro-cognitive scientific community, are generally not practiced effectively within the vascular biology research community.

In order to advance the state of the vascular dementia field and improve the linkage between the vascular biology and neuro-cognitive research communities, NHLBI proposed a workshop to discuss the utility of establishing a comprehensive phenotypic cognitive assessment of a selected set of mouse models, representative of the spectrum of vascular disorders, with particular attention focused on age, sex, and rigor and reproducibility. It is hoped that the development of well-characterized models, with validated cognitive outcomes, can be used to link specific vascular signaling pathways with specific cognitive deficits.

Workshop Participants

Over 20 speakers and designated discussants represented various aspects of vascular biology and neuro-cognitive research fields. The initial phases of the workshop were planned by NHLBI extramural staff, with invited contributions from extramural staff of the National Institute on Aging (NIA) and the National Institute of Neurological Disorders and Stroke (NINDS).

The workshop was attended virtually by more than 100 registrants, composed primarily of academic researchers and NHLBI/NIH staff. A list of suggested readings related to topics discussed is included at the end of this report.

Workshop Agenda

Day One Discussions

  • Vascular dementia as manifest in humans
  • Current NIA and NINDS dementia research strategies
  • Examples of vascular disease/disorder mouse models
  • Thoughts on how to move the field forward

Day Two Discussions

  • Measuring cognition/cognitive-like phenotypes in mice
  • Other factors/measurements that should be included
  • How to ensure rigor and reproducibility in the outcomes
  • Final thoughts on the comprehensive assessment proposal

Summary of Workshop Discussions

Vascular dementia in humans

Vascular dementia (VaD) in humans is a spectrum of cognitive disorders based on etiology, pathology, and tempo of cerebrovascular disease. Challenges in defining VaD in human patients include the following: (1) VaD is not a single condition with a unifying pattern of cognitive deficits, (2) Different pathologies can manifest as VaD and different pathologies (vascular and nonvascular) can and often do co-exist, (3) Cognitive deficiency syndromes can be related to a specific acute event (such as stroke) or to repeated smaller insults (such as small vessel disease), and (4) Cognitive syndrome is, for the most part, a gradually progressive disorder that primarily affects older individuals.

Hypertension plays a prominent role in cognitive decline in humans, particularly as higher levels of hypertension significantly increase the risk of cardiovascular disease such as stroke. The recent NHLBI-sponsored Systolic Blood Pressure Intervention Trial (SPRINT) demonstrated a significantly higher cumulative risk for cardiovascular events in patients at least 50 years of age whose systolic blood pressure exceeded 130 compared to those with readings closer to 120. There was also a trend in SPRINT towards reduced cognitive decline with reduced hypertension. As hypertension is associated with almost all incidents of human vascular dementia, more effective control of blood pressure could have a significant impact in reducing the incidence of vascular dementia.

To best parallel the clinical syndrome of vascular dementia in humans, animal models should address executive function as a salient feature of the syndrome as well as other core domains of cognitive assessment, including attention, memory, and visuospatial function; and non-cognitive features, such as gait measures.

Current NIA and NINDS dementia research strategies

The NIA is supporting the Model Organism Development and Evaluation for Late Onset Alzheimer’s Disease (MODEL-AD) initiative, which will create 50 new mouse models using CRISPR technology. Subsequently, an in-depth characterization will be conducted to align mouse and human phenotypes with a focus on prioritizing the most translationally relevant phenotypes. Importantly, robust and reliable behavioral outcome measures will be measured in tandem with biomarkers, neuropathology, -omics, and neuroimaging.

The data from MODEL-AD, including mouse strains, data types, studies, and assays are available to the scientific community on the AD Knowledge Portal (https://adknowledgeportal.org).

The NINDS strategic approach is closely aligned with NIA vision and objectives. Understanding vascular contributions to cognitive impairment and dementia (VCID) is a priority identified in the National Plan to Address Alzheimer’s Disease. NINDS supports mechanism-oriented VCID research such as better understanding of the neurovascular unit, the impact of cerebrovascular and cardiovascular disease insults, and the various effects of proteinopathies, metabolic disease, and immune response on cognitive function.

Examples of vascular disease/disorder mouse models

During the workshop, more than a dozen mouse models of vascular disease or disorders were presented. They included models of inherited small vessel disease in the brain, atherosclerosis, hypertension, cerebral cavernous malformations, and intracerebral hemorrhage. Models of inherited blood disorders (such as sickle cell anemia) and chronic lung vascular disease (chronic obstructive pulmonary disease) were also discussed. The various pros and cons with which these models accurately (or not) reflect vascular disease in humans was also assessed. In general, it was agreed that none of the models fully reflected the spectrum of vascular-induced cognitive impairment in humans, but that each of the models offered specific insights into the basic biological processes that likely govern the development of VCID.

Some of these vascular disease models have been evaluated for cognitive phenotypes but most have not. None have been assessed so as to allow for comparisons with other vascular models. It was suggested that a rigorous cognitive assessment of a set of existing vascular models could provide the VCID field with a template or standard by which to associate certain aspects of vascular regulation with certain aspects of cognition or cognitive-like behavior.

Moving the field forward

Participants identified the need for multiple animal models to advance the field because VCID is multi-faceted disease and each different model gives researchers the opportunity to study some unique aspect of the disease. There was also agreement that some prioritization of models might be useful, perhaps based on the burden of disease caused by a given pathology or on other criteria.

Age is a critically important factor in the biology of VCID and its assessment. The human relevance of specific mouse models can be improved by studying VCID in aging mice. At the same time, the use of aged animals greatly increases the cost of experimentation and when coupled with the need to use both males and females can create a cost-prohibitive situation for many individual laboratories. Furthermore it was recognized that utilization of genetically diverse mice to mimic human genetic diversity, in combination with mouse models that best mimic human disease through similarity of gene/protein alterations taking into consideration age-appropriate disease manifestations would be of benefit.

Other discussion points included:

  • More emphasis on histological assessments and more use of advanced imaging technologies, such as MRI.
  • Enhanced characterization of human brain vasculature from patients in order to better understand the disease pathology.
  • The measurement of phenotypes with no known brain pathology, such as hyperactivity and gait.
  • Increased attention to environment and animal husbandry issues as they relate to animal behavior.
  • Focus on mouse genetic models, but with awareness that different alleles can have different effects with some simple knockout models.

Measuring cognition/cognitive-like phenotypes in mice

There are many challenges, limitations, and confounding factors when attempting to measure cognition or cognitive-like behaviors in mouse models. Multiple variables can affect intra- and inter-lab reproducibility, including task type, task duration, environmental conditions, time of day, and pre- and post-test subject treatment. These many considerations need to be reported to ensure reproducibility. In addition, multiple factors can act as confounders in behavioral assays and result in misinterpretation of data, e.g., aging-dependent visual impairments and hyperactivity. To ensure rigor and reproducibility, scientists should follow the Animals Research: Reporting of In Vivo Experiments (ARRIVE) guidelines (2010) for experimental research, use appropriate control animals, establish a priori inclusion and exclusion criteria, predetermine sample sizes based on power analyses, randomize and counter-balance experimental conditions, and maintain data blinding until the completion of the analysis.

Standardization with well-delineated standard operating procedures (SOPs) is critical for high-throughput cognitive assessments and comparison across multiple models. Validated training of experimental personnel is essential. Newer methodologies, such as Touchscreen Cognitive Testing (TCT), may provide a more effective alternative to traditional assays. TCT enables high-throughput testing with standardization and reproducibility through computer automation, reduced human error, reduced stress on animals, and tests similar to those used with humans to increase translatability. It is also amenable to open science and data sharing.

Given that no single animal model can recreate all human VCID pathologies, including biological processes (microinfarcts and/or microhemorrhages, arteriosclerosis, lacunes/strokes), imaging and biomarkers (white matter hyperintensities, enlarged perivascular spaces, and diminished cerebrovascular reactivity), and clinical presentations (impaired executive function, diminished verbal fluency, attention, and processing speed), it is increasingly important to evaluate multiple vascular mouse models using a common experimental framework and to facilitate comparison of results across distinct models.

Other factors/measurements that should be included

Other possible factors to consider include genetics (background, rare vs. common disease, transgenic vs. KO, etc.), aging, sex as a biological variable, environmental factors (feed, microbiome, light, sound, handling, etc.), and cohorting (batching). Discussants indicated all are important factors that need to be fully reported.

The workshop participants agreed that some type of brain imaging modality should also be incorporated into any set of measurements that seeks to have a comprehensive assessment of cognition or cognitive-like phenotypes associated with vascular mouse models. In vivo imaging provides significant opportunities to measure cerebrovascular functioning that can be accomplished uniquely in the mouse. MRI was the most commonly mentioned preferred imaging modality, but again, affordable access to all laboratories can be problematic.

How to ensure rigor and reproducibility in the outcomes

Rigor and reproducibility in biomedical research are essential for discoveries to be translated into improved human health. Several recent reports, including the NIH Advisory Committee to the Director Report on Enhancing Rigor, Transparency, and Translatability in Animal Research, as well as the findings in a Nature report on factors that can improve reproducibility, or reduce it.

Biased reporting and time pressure have been cited as two significant causes of irreproducibility. Other factors include failure to conduct adequate statistical sample size calculations and subsequent statistical analysis. Increasingly, for experiments that seek to measure complex phenotypes and behaviors (such as cognition), it is essential to have a robust statistically justified sample size, well-defined experimental endpoints, well-chosen experimental controls, methods for randomization and counter-balancing, and double-blinded assessments.

Assay proficiency metrics and cross-lab training are important elements of rigor. There is also a need to establish reproducibly measured positive controls under the experimental conditions being used and appropriate assays for what is being measured. The use of dedicated centers to ensure standardized evaluations of complex phenotypes by fully trained personnel with well-calibrated equipment should also be considered.

Final thoughts on the comprehensive assessment proposal

This workshop identified a need for multiple vascular models to dissect the multifaceted aspects of Vascular Cognitive Impairment. In addition, the measurement of complex phenotypes, such as cognition and cognitive-like behaviors, requires careful considerations and execution by trained personnel. In order to achieve reliable comparisons of one model to another, the assessments are probably most effectively done at dedicated centers that can provide the standardization and rigor needed to produce reproducible outcomes.

Models most closely aligned with known Vascular Cognitive Impairment risk factors should be considered high priority. Likewise, measures of VCID and executive function in humans, including processing speed, attention, and working memory, should be prioritized. The use of newer cognitive measurement technologies, such as touchscreen cognitive testing, may offer significant advantages over older cognitive measures. The inclusion of imaging and histological analysis also was strongly encouraged.

The establishment of a rigorous and reproducible assessment of cognitive impairments associated with existing vascular disease models offers the prospective linkage of specific vascular signaling pathways with specific cognitive deficits. Such a linkage may create a template or standard for a comprehensive assessment of VCID outcomes in vascular disease models, as well as an expedited conduit to the development of new therapeutic modalities.