Longitudinal Fragility Phenotyping Predicts Lifespan and Age-Associated Morbidity in C57BL/6 and Diversity Outbred Mice

Alison Luciano,Laura Robinson,Gaven Garland,Bonnie Lyons,Ron Korstanje,Andrea Di Francesco,Gary A. Churchill

DOI: https://doi.org/10.1101/2024.02.06.579096

2024-02-08

Abstract:Aging studies in mammalian models often depend on natural lifespan data as a primary outcome. Tools for lifespan prediction could accelerate these studies and reduce the need for veterinary intervention. Here, we leveraged large-scale longitudinal frailty and lifespan data on two genetically distinct mouse cohorts to evaluate noninvasive strategies to predict life expectancy in mice. We applied a modified frailty assessment, the Fragility Index, derived from existing frailty indices with additional deficits selected by veterinarians. We developed an ensemble machine learning classifier to predict imminent mortality (95% proportion of life lived [95PLL]). Our algorithm represented improvement over previous predictive criteria but fell short of the level of reliability that would be needed to make advanced prediction of lifespan and thus accelerate lifespan studies. Highly sensitive and specific frailty-based predictive endpoint criteria for aged mice remain elusive. While frailty-based prediction falls short as a surrogate for lifespan, it did demonstrate significant predictive power and as such must contain information that could be used to inform the conclusion of aging experiments. We propose a frailty-based measure of healthspan as an alternative target for aging research and demonstrate that lifespan and healthspan criteria reveal distinct aspects of aging in mice.

Bioinformatics

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to develop and validate non - invasive prediction tools in mouse models to predict the lifespan and healthspan of mice in advance, thereby accelerating aging research and reducing the need for late - stage veterinary interventions. Specifically:

1. **Develop prediction tools**: The paper aims to develop a non - invasive method capable of predicting the lifespan and healthspan of mice. By using machine - learning algorithms and longitudinal frailty assessment (such as Fragility Index, FgI), researchers hope to accurately predict the remaining lifespan of mice before their natural death.

2. **Improve prediction accuracy**: Existing prediction tools (such as Temperature × Weight [TxW] and AFRAID clock) have limitations in their performance under different genetic backgrounds and dietary conditions. Therefore, the paper attempts to improve these tools to make them more applicable to mice of different genetic backgrounds and genders and to maintain high prediction accuracy under different experimental conditions.

3. **Evaluate the applicability of existing tools**: The paper also evaluates the performance of existing tools (such as TxW and AFRAID clock) under different genetic backgrounds, genders, and dietary conditions to determine whether they can be widely used in aging research.

4. **Introduce healthspan as an alternative endpoint**: In addition to predicting lifespan, the paper also proposes the possibility of using healthspan as an alternative endpoint. Healthspan is defined as the survival time of mice before the appearance of four or more severe health defects. This helps to more comprehensively evaluate the effects of aging interventions.

### Main research content

- **Data sources**: The study used two mouse cohorts with different genetic backgrounds - C57BL/6J (homozygous line) and Diversity Outbred (diversity outbred line). These mice were longitudinally observed for a long time under different dietary conditions.

- **Evaluation indicators**: The body temperature, body weight, and FgI (frailty index based on multiple clinical symptoms) of mice were evaluated weekly. Through these data, researchers constructed prediction models to evaluate the lifespan and healthspan of mice.

- **Model development and validation**: Researchers used machine - learning algorithms (such as random forest and support vector machine) to construct prediction models and evaluated the performance of the models through cross - validation and test sets. The results showed that the model combining FgI and other clinical indicators performed well in predicting 95% of the life proportion (95PLL).

### Conclusions

- **Challenges in predicting lifespan**: Although the prediction model based on FgI performs well in some cases, it still has certain limitations, especially in the generalization ability under different genetic backgrounds and dietary conditions.

- **Importance of healthspan**: Healthspan, as an alternative endpoint, may more accurately reflect the actual effects of aging interventions than simply predicting lifespan. Therefore, future research should focus more on the evaluation of healthspan.

- **Future directions**: It is necessary to further optimize prediction models, especially by combining more biomarkers and clinical data, to improve the accuracy and reliability of prediction.

In summary, this paper provides an important reference for aging research by developing and evaluating new prediction tools and emphasizes the importance of healthspan in evaluating aging interventions.